Palauan Reference Grammar

PalauanReferenceGrammar
PALI LANGUAGE TEXTS: MICRONESIA

Social Sciences and Linguistics Institute

University of Hawaii

Donald M. Topping
Editor

ii
 Palauan Reference
Grammar
LEWIS S. JOSEPHS
With the assistance of

MASA-AKI EMESIOCHEL
MASAHARU TMODRANG
HELEN WILSON

The University Press of Hawaii

HONOLULU
 Open Access edition funded by the National
Endowment for the Humanities / Andrew W.
Mellon Foundation Humanities Open Book
Program.

Licensed under the terms of Creative Commons

Attribution-NonCommercial-NoDerivatives 4.0 In-
ternational (CC BY-NC-ND 4.0), which permits readers to freely
download and share the work in print or electronic format for
non-commercial purposes, so long as credit is given to the
author. Derivative works and commercial uses require per-
mission from the publisher. For details, see
https://creativecommons.org/licenses/by-nc-nd/4.0/. The Cre-
ative Commons license described above does not apply to any
material that is separately copyrighted.

Open Access ISBNs:

9780824879068 (PDF)
9780824879075 (EPUB)
This version created: 20 May, 2019

Please visit www.hawaiiopen.org for more Open Access works

from University of Hawai‘i Press.

The research reported herein was partially supported by Grant

No. GS-33639 from the National Science Foundation and by the
Government of the Trust Territory of the Pacific. Publication of
the finished book was financed by the Trust Territory Govern-
ment. Financial support for Masa-aki Emesiochel, Masaharu
Tmodrang, and Dr. Helen Wilson was made possible through the
abovementioned National Science Foundation Grant, the Trust
Territory Government, and the Culture Learning Institute of the
East-West Center at the University of Hawaii.

Copyright© 1975 by The University Press of Hawaii

All rights reserved
To my father
And to Palau for the Palauans
 Contents

Contents
Dedication viii
Preface xix

1 Sounds and Spelling of Palauan 1

1.1. Sound System of Palauan 1
1.2. Significant Sounds (Phonemes) of Palauan 2
1.3. The Consonants of Palauan 3
1.3.1. Stops 4
1.3.2. Fricative 12
1.3.3. Nasals 12
1.3.4. Liquids 16
1.3.4.1. The Sequences ll and rr 17
1.3.5. Syllabic Consonants 17
1.4. The Vowels of Palauan 18
1.4.1. High Vowels i and u 19
1.4.2. Mid Vowels e, ȩ, and o 20
1.4.3. Low Vowel a 22
1.4.4. The Vowel ȩ and the Process of Vowel
Reduction 22
1.4.5. Other Occurrences of Schwa 24
1.4.6. Long Vowels 25
1.4.7. Vowel Clusters 26
1.5. Further Rules of Palauan Spelling 29
1.6. Suggestions for Further Study 36
2 Palauan Nouns 38
2.1. Grammatical System of Palauan 38
2.2. Identification of Palauan Nouns 39
2.3. Distribution of Palauan Nouns 40
2.4. Pronouns 44
2.5. Plurals 47
2.6. The Palauan Word a 50
2.7. Specific vs. Non-specific Object Nouns 52
3 Noun Possession 59
3.1. Possessor Suffixes and Possessed Nouns 59

ix
 Contents

3.2. Person and Number 60

3.3. Four Sets of Possessor Suffixes 61
3.4. Vowel Reduction 63
3.4.1. Vowel Deletion 65
3.4.2. Shortening of Long Vowels 66
3.4.3. Shortening of Vowel Clusters 67
3.4.3.1. Technical Discussion of Vowel Cluster
Shortening 69
3.4.4. Possessed Nouns With -ng- 70
3.4.5. Alternation Between -ng and -l- 70
3.5. Obligatorily Possessed Nouns 71
3.6. Palauan Noun Phrases 73
3.7. Noun Phrases of Possession vs. Noun Phrases
of Characterization 77
3.8. Unpossessible Nouns 78
3.9. Noun Phrases of Characterization: Further
Examples 80
3.10. Appositional Structures 82
3.11. Some Special Possessed Nouns 83
3.12. Summary of Palauan Noun Phrases 84
4 Palauan Pronouns 86
4.1. Independent Pronouns and Affix Pronouns 86
4.2. Non-emphatic vs. Emphatic Subjects 87
4.3. Use of Emphatic Pronouns Following ȩr 93
4.4. Emphatic Pronouns in Coordinate Noun
Phrases 94
4.5. Emphatic Pronouns Preceded by di 96
4.6. Other Uses of Emphatic Pronouns 96
4.7. Non-emphatic Pronouns as Pronominal
Traces 97
4.8. Pronominalization 102
4.9. Object Pronouns and Perfective Verbs 105
4.9.1. The Zero (Ø) Object Pronoun 107
4.9.2. Further Examples of Perfective Verb Forms 108
4.9.3. Verbs with Limited Perfective Forms 110
4.9.4. Variant Forms of the Object Pronouns 112
4.10. Hypothetical Pronouns 115

x
 Contents

4.10.1. Hypothetical Pronouns with Imperfective

Verbs 116
4.10.2. Hypothetical Pronouns as Agents 117
4.10.3. Hypothetical Pronouns with the Past Tense 118
4.10.4. Reduced Variants of the Hypothetical
Pronouns 119
4.10.5. Distribution of the Reduced Variants 120
4.10.6. Hypothetical Forms of Complex Verb
Phrases 122
4.10.7. Imperative Verb Forms 123
4.10.8. Propositive Verb Forms 124
4.10.9. Summary of Hypothetical Pronouns 125
5 Palauan Verbs 126
5.1. Identification of Palauan Verbs 126
5.1.1. Action Verbs: Transitive and Intransitive 126
5.1.2. State Verbs 130
5.1.3. Further Differences Between Action Verbs and
State Verbs 132
5.2. Distribution of Palauan Verbs 135
5.3. Tense 139
5.3.1. Present Tense 139
5.3.2. Past Tense 140
5.3.2.1. The Auxiliary mla 146
5.3.3. Future Tense 147
5.4. The Verb Marker and Ergative Verb Forms 148
5.5. Imperfective vs. Perfective Verb Forms 155
5.6. Active and Passive Sentences 161
5.7. Further Types of Verbs 162
5.8. Summary of Palauan Verbs 164
6 The Verb Marker and Perfective Verb Forms 165
6.1. The Verb Marker Prefixes 165
6.1.1. The Prefix ou- 167
6.2. The Infixed Verb Marker and Metathesis 169
6.2.1. Absence of the Verb Marker in Hypothetical Verb
Forms 172
6.3. Selected List of Perfective Verb Forms 174
6.3.1. Metathesis in Perfective Verb Forms 176
6.3.2. The Metathesized Verb Marker and Vowel
Blending 179
6.3.3. Deletion of the Metathesized Verb Marker 183

xi
 Contents

6.4. Vowel Reduction and Vowel Deletion in

Perfective Verb Forms 185
6.5. Additional Remarks on Perfective Verb
Forms 187
6.6. Perfective Forms of Transitive Verbs in o- 189
6.7. Hypothetical Forms of Perfective Verbs 191
7 State Verbs 193
7.1. Internal Structure of State Verbs 193
7.2. State Verbs with Plural Subjects 195
7.3. Verbs with mle and -il- in the Past 197
7.4. Transitive State Verbs 198
7.5. State Verbs with bȩ- 200
7.6. The Prefixes bȩkȩ- and sȩkȩ- 201
7.6.1. State Verbs with bȩkȩ- Denoting Smells 203
7.7. Resulting State Verbs 204
7.8. Anticipating State Verbs 206
7.8.1. The Anticipating State Suffix -all 210
7.8.2. Resulting and Anticipating State Verbs as
Nouns 210
7.9. Transitive Verbs Derived from State Verbs 211
8 Noun Derivation 213
8.1. Internal Structure of Palauan Nouns 213
8.2. State Verbs Functioning as Nouns 214
8.3. Nouns Derived with -(ȩ)l - from Intransitive
Verbs 217
8.4. Nouns Derived from Reciprocal Verbs 219
8.5. Abstract Nouns Derived with kl(ȩ)- 220
8.6. Instrument and Action Nouns with o- 222
8.7. Nouns Derived with ul(ȩ)- 224
9 Causative Verbs 227
9.1. Meaning and Use of Causative Verbs 227
9.2. Forms of the Causative Prefix 229
9.2.1. The Prefix omȩ(k)- 230
9.2.1.1. Technical Discussion of the Prefix omȩ(k)- 231
9.2.1.2. Additional Types of Causative Verbs with
omȩ(k)- 234
9.2.1.3. Sample Sentences with omȩ(k)-Causatives 235
9.2.2. The Prefix ol(ȩ)- 236

xii
 Contents

9.2.2.1. Technical Discussion of the Prefix ol(ȩ)- 238

9.2.2.2. Additional Types of Causative Verbs with ol(ȩ)-
239
9.2.2.3. Sample Sentences with ol(ȩ)- Causatives 240
9.3. Verb Stems Allowing Two Causatives 240
9.4. Perfective Forms of Causative Verbs 242
9.5. Ergative Forms of Causative Verbs 245
9.6. Hypothetical Forms of Causative Verbs 246
10 Reciprocal Verbs 248
10.1. Meaning and Use of Reciprocal Verbs 248
10.2. Forms of the Reciprocal Prefix 250
10.2.1. Reciprocal Verbs Related to Causative Verbs 250
10.2.2. Reciprocal Verbs Related to Verbs in ou- 252
10.2.3. Reciprocal Verbs Derived from Transitive Verb
Stems 253
10.2.4. The Reciprocal Prefix cha- 255
10.3. Sample Sentences with Reciprocal Verbs 256
10.4. Extended Functions of the Reciprocal
Prefix 258
11 Reduplication and Further Verb Affixation 261
11.1. Types of Verb Derivation 261
11.1.1. Previous Examples of Reduplication 262
11.2. Reduplication of Simple State Verbs 262
11.2.1. Reduplication of Possessed Nouns 263
11.3. Reduplication of State Verbs in mȩ- 264
11.4. Reduplication of Derived State Verbs 265
11.5. Complex Patterns of Reduplication 266
11.6. Reduplication of Intransitive Action Verbs 268
11.7. Reduplication of Transitive Action Verbs 269
11.8. Reduplication of Verbs in ou- and o- 275
11.9. Reduplication of Causative Verbs 276
11.10. Reduplication of Reciprocal Verbs 277
11.11. “Fossilized” Reduplication 277
11.12. The Predictive and Inchoative Suffixes 278
11.12.1. Predictive and Inchoative Forms of Intransitive
Action Verbs 279
11.12.2. Predictive and Inchoative Forms of Ergative
Verbs 281

xiii
 Contents

11.12.3. Predictive and Inchoative Forms of

Imperfective Verbs 282
11.12.4. Inchoative Forms of State Verbs 283
11.12.5. Predictive and Inchoative Forms of mo 285
11.12.6. The Predictive Word ku 286
12 Imperfective vs. Perfective Verbs 288
12.1. Meaning of Imperfective vs. Perfective Verb
Forms 288
12.2. Imperfective vs. Perfective Verb Forms in
Various Tenses 290
12.3. Further Examples of Contrast Between
Imperfective and Perfective Verb Forms 294
12.4. Imperfective vs. Perfective Verb Forms and
Specific Objects 296
12.5. Contextual Restrictions on the Use of
Imperfective vs. Perfective Verb Forms 297
12.6. The Transitive Verb omes 300
12.7. The Transitive Verb orrengȩs 302
13 Directional Verbs 304
13.1. The Three Directional Verbs of Palauan 304
13.2. Meaning and Use of the Directional Verbs 304
13.3. Directional Verbs Followed by Action Verbs 308
13.4. The Directional Verbs and Future Time 309
13.5. The Directional Verb mo and Change of State
311
13.5.1. Change of State Expressions with Noun
Phrases 314
13.6. Special Verbal Expressions with mo 314
14 Relational Phrases 316
14.1. Distribution and Function of Relational
Phrases 316
14.2. Locational Phrases 318
14.2.1. Locational Phrases with Nouns Describing
Spatial Relationships 320
14.2.2. Additional Examples of Nouns Describing
Spatial Relationships 325
14.3. Directional Phrases 326
14.3.1. Further Types of Directional Phrases 329
14.4. Source Phrases 330

xiv
 Contents

14.4.1. Further Types of Source Phrases 331

14.5. Cause Phrases 332
14.6. Temporal Phrases 333
14.7. Sentences with More than One Relational
Phrase 338
14.8. Relational Phrases Used to Express
Comparison 340
14.9. Further Types of Palauan Relational
Phrases 341
15 Dependent Clauses 344
15.1. Characteristics of Dependent Clauses 344
15.2. Purpose Clauses 346
15.2.1. Purpose Clauses Containing Directional
Verbs 347
15.2.2. Purpose Clauses Following Nouns 349
15.3. Instrument Clauses 350
15.4. Purpose and Instrument Clauses with Past
Tense Verb Forms 351
15.5. Means of Transportation Clauses 353
15.6. Accompaniment Clauses 355
15.7. Specifying Clauses 357
15.7.1. Specifying Clauses Containing Directional
Verbs 359
15.7.2. Specifying Clauses Following lmuut 361
15.7.3. Specifying Clauses Following dmak 362
15.7.4. Specifying Clauses Containing the Perfective
Forms of mȩrkui 364
15.7.4.1. Regional Variation in the Use of ȩl rokir and
ȩl rokui 367
15.7.5. Specifying Clauses Designating Periods of
Time 368
15.7.6. Specifying Clauses in Sentences Designating
Manner 369
15.7.7. Specifying Clauses Following Special Verbs 371
16 Object Clauses 373
16.1. Structure and Function of Object Clauses 373
16.2. Object Clauses Following omuchȩl and mo
mȩrek 375
16.3. Object Clauses Following Transitive State
Verbs 379

xv
 Contents

16.4. Further Examples of Object Clauses 380

16.5. Sentences with Two Objects 381
17 Processes of Sentence Formation: Subject Shifting and
Preposing of Possessor 384
17.1. Four Special Possessed Nouns 384
17.2. Subject Shifting 386
17.3. Preposing of Possessor 390
17.4. Subject Shifting and Preposing of Possessor
with Expressions Containing reng 394
17.5. Preposing in Reciprocal Sentences 397
17.6. Summary of Processes Affecting Palauan Word
Order 400
17.7. Dependent Clauses Resulting from Subject
Shifting 402
17.8. Subject Shifting and Derived Action Nouns 406
17.8.1. Possessed Forms of Derived Action Nouns 409
17.9. Soal and chȩtil Followed by Hypothetical Verb
Forms 410
17.10. Further Discussion of the Four Special
Possessed Nouns 413
18 Negation 419
18.1. Affirmative vs. Negative Sentences 419
18.2. Affirmative vs. Negative Expressions of
Existence 420
18.2.1. Subject Shifting in Affirmative and Negative
Expressions of Existence 424
18.3. The Negative Verb diak 427
18.4. Hypothetical Verb Forms Following diak 430
18.5. Further Examples of Negative Sentences 434
18.6. Equational Sentences: Affirmative and
Negative 437
18.7. The Negative Expression di kea 440
18.8. Negative Verbs as Answers to Questions 441
19 Uses of Hypothetical Verb Forms 445
19.1. Conditional Sentences 445
19.1.1. Conditions in the Present or Future 446
19.1.2. Conditions in the Past 447
19.1.3. Conditional Clauses Containing diak 448

xvi
 Contents

19.2. Permutation of Conditional and Consequential

Clauses 449
19.3. Further Types of Conditional Clauses 450
19.4. Additional Examples of Conditional
Clauses 455
19.5. Imperative Verb Forms 458
19.6. Propositive Verb Forms 464
19.7. Passive Sentences 465
19.7.1. Further Examples of Passive Sentences 470
19.7.2. Passive Sentences Containing Complex Verb
Phrases 472
19.7.3. Passive Sentences and Relational Phrases 473
19.7.4. Negative and Conditional Passive Sentences 474
20 Questions 475
20.1. Yes-No Questions 475
20.2. The Question Word tȩcha 480
20.2.1. Further Types of Questions with tȩcha 484
20.3. The Question Word ngara 486
20.3.1. Further Types of Questions with ngara 490
20.4. The Question Word tela 493
20.5. The Question Word ker 495
20.6. The Question Word oingara 497
20.7. The Special Question Word mȩkȩra 498
20.8. Sentences with Two Question Words 499
21 Direct and Indirect Quotation 500
21.1. Direct vs. Indirect Quotation 500
21.2. Indirect Quotation and Verbs of
Communication and Mental Activity 502
21.3. Direct and Indirect Quotation of
Commands 504
21.4. Direct and Indirect Quotation of Questions 506
21.5. The Expressions ȩl kmo and ȩl ua se 510
21.6. Indirect Quotation Following Nouns of
Communication 511
22 Reason Clauses, Result Clauses, and Time Clauses 512
22.1. Reason and Result Clauses 512
22.1.1. Special Types of Result Clauses 515
22.2. Time Clauses 520

xvii
 Contents

22.2.1. Before and After 523

22.2.2. Movement of Time Clauses and Time Words 524
23 Relative Clauses 527
23.1. Function of Relative Clauses 527
23.2. Derivation of Relative Clauses 529
23.3. Passive Sentences as Relative Clauses 530
23.4. Equational Sentences as Relative Clauses 535
23.5. Relative Clauses Containing State Verbs 537
23.6. Relative Clauses Following chad and klalo 538
24 Modifiers 540
24.1. State Verbs as Modifiers 540
24.2. Other Verbs and Nouns as Modifiers 543
24.3. Demonstratives 545
24.4. Numbers 551
24.4.1. Numbers Above 20 555
24.4.2. Minor Number Sets 557
24.4.3. Use of Number Words in Sentences 558
24.4.4. Ordinal Numbers 559
24.4.4.1. Ordinal Numbers Followed by Specifying
Clauses 560
24.5. Expressions with Two or More Modifiers 562
24.6. Qualifying Words 563
25 The Connecting Words Mȩ and E 565
25.1. Review of Complex Sentences with mȩ and e
565
25.2. Further Uses of the Connecting Word mȩ 568
25.2.1. The Expression mȩ a lȩchub 569
25.3. Further Uses of the Connecting Word e 570
25.3.1. The Expression e ng di 574
25.4. Coordinate Noun Phrases 575
25.4.1. Coordinate Noun Phrases with mȩ a lȩchub 579
Notes 580
Appendix: Guide to Phonetic Symbols Used in This
Textbook 635
Glossary of Terms 637
Bibliography 660

xviii
 Preface

Preface
The main purpose of this book is to offer a comprehensive
description of the Palauan language which will be of use to
speakers of Palauan and to linguists alike. I have attempted to
make my explanations as simple and clear as possible, in spite
of the fact that the data to be analyzed are often discourag-
ingly complex. To achieve this, I have taken special pains to
define difficult concepts at length, and to illustrate these con-
cepts with copious examples. I have presented the material in
such a way that there is a gradual build-up in complexity, with
later analyses depending on, or assuming an understanding of,
earlier ones. If the general reader masters the material as he
goes along, he should have little difficulty in grasping some
of the later, more difficult analyses. Such analyses are, un-
fortunately, necessary because the structures to be explained
are themselves so complex; in many cases, if I had chosen
to present an oversimplified, watered-down description of the
data, I would not have been able to capture some of the es-
sential phonological and grammatical principles that uniquely
characterize the Palauan language.
After an introduction to the sounds and spelling of Palauan
in chapter 1, I go on to describe some of the basic Palauan parts
of speech (nouns and verbs) in chapters 2-5. In chapters 6-11, I
describe the salient features of Palauan morphology (i.e., word
formation) and explain the many complex rules which interact
with each other when verbs and nouns are derived. Finally, in
chapters 12-25, I talk about the major syntactic constructions
and grammatical processes of Palauan. Although the great ma-
jority of discussions are intended for the general reader, in a few
places I have included discussions or remarks of a highly tech-
nical nature which are meant for readers with some degree of
linguistic training. When these discussions appear as a section
of the text or as a note, the number introducing the section
or the note has been preceded with an asterisk (*). General
readers can disregard such materials without losing the train of
the discussion. As they proceed through the text, all readers will
find the extensive cross-referencing of considerable assistance;

xix
 Preface

as further aids in using the book, a list of phonetic and ortho-

graphic symbols, a glossary of linguistic terms, a brief bibliog-
raphy, and an index have also been appended.
I am grateful to Dr. Helen Wilson, who offered me invaluable
assistance in conducting interviews and writing preliminary
versions of some of the chapters. If it had not been for the un-
flagging interest of my two principal informants, Masa-aki Eme-
siochel and Masaharu Tmodrang, in this research and for their
strong determination to make a description of the Palauan lan-
guage available to the Palauan people, this book would never
have been possible. I thank them profoundly for their patience
and cooperation. I am also indebted to the many people in
Palau, in particular to the teachers and administrators of the
Palau High School, who helped me in numerous ways to bring
this research to completion. Finally, I would like to express my
sincerest gratitude to Dr. Donald Topping, Director of the Social
Sciences and Linguistics Institute of the University of Hawaii,
who provided the crucial initial inspiration for this volume and
who painstakingly reviewed the manuscripts, offering innu-
merable suggestions for improving organization, style, and
content; and to Dawn Reid, who gave so much of her time typing
the original drafts.

xx
 1 Sounds and Spelling of
Palauan

1SoundsandSpellingofPalauan
1.1. SOUND SYSTEM OF PALAUAN
When a linguist attempts to present a description of a language,
he usually finds it desirable to treat the sound system of
the language separately from the grammatical system. Even
though this division may seem arbitrary or artificial (for how
can we have the words, phrases, and sentences of a language
without sound, and vice versa?), we are nevertheless going to
examine the sound system of Palauan before proceeding to a de-
scription of the grammatical system.
Every language of the world uses its own particular set of
sounds to construct words. No two languages have precisely
the same set of sounds or the same number of sounds. Even
though the human vocal apparatus is capable of producing an
extremely large number and variety of sounds, speakers of dif-
ferent languages actually use only a relatively small number
of such possible sounds when communicating in speech. The
Hawaiian language, for example, uses only thirteen distinctive
sounds, perhaps the smallest number among world languages;
closer to the average are Chamorro, with twenty-five distinctive
sounds, and English, with approximately thirty. When a human
being learns his native language, he becomes so used to the
particular sounds of it that he may have great difficulty in pro-
nouncing the sounds of some other language correctly. As a
result of such imperfect pronunciation, he is said to speak the
second language with a “foreign accent.”
In addition to having a limited number of sounds, every lan-
guage organizes its sounds into a system which is unique to that
language. This systematic organization involves such things as
the positions and combinations in which the various sounds can
occur, the frequency with which certain sounds occur, the varia-
tions of pronunciation which particular sounds undergo, and so
forth. In this chapter we will take up some of the more out-
standing features of the sound system of Palauan.

1
 1 Sounds and Spelling of Palauan

1.2. SIGNIFICANT SOUNDS (PHONEMES) OF

PALAUAN
When a linguist studies the sound system of a language, his
primary concern is to discover the significant sounds, or
phonemes. In order to determine what the significant sounds
of the language are, the linguist tries to find cases in which the
substitution of one sound for another results in a different word
and, consequently, a change of meaning. For example, if we take
the Palauan word tub ‘spit’ and substitute a d for the first sound
t, we will get dub, which is a completely different word meaning
‘dynamite’. On the basis of this pair of words—tub ‘spit’ and dub
‘dynamite’—we can conclude that t and d are significant sounds,
or phonemes, of Palauan. To use another term common among
linguists, we can say that t and d are contrastive sounds in
Palauan because they contrast with each other (or are in oppo-
sition to each other) in otherwise identical environments. Lin-
guists use the term ‘environment’ in a technical sense to refer
to the sound or sounds which are adjacent to or close to some
other sound. For the pair of words under discussion, both t and
d contrast in the same environment, because both of them are
followed by ub.
The pair of words tub ‘spit’ and dub ‘dynamite’ can be
called a minimal pair because the difference between them is
minimal—i.e., determined by the substitution of a single sound.
Some other minimal pairs in Palauan include

(1) blai ‘house’ — mlai ‘canoe’

chad ‘person’ — char ‘price’
kar ‘medicine’ — ker ‘question’
tet ‘purse’ — tut ‘breast’

What are the contrastive sounds in each of the above minimal

pairs? Why don’t the pairs of words brer ‘raft’—brak ‘taro’ or
daob ‘ocean’—taod ‘fork’ qualify as minimal pairs?
There are also differences in the individual sounds of a lan-
guage which are non-contrastive. Even though the linguist
can identify and describe such differences, the speakers of the
language are often unaware of them because they are auto-
matic or predictable varieties of the same sound. For example,
in Palauan the letter d is used to represent four phonetically dif-
ferent sounds, as in the following words. (Note that the phonetic

2
 Palauan Reference Grammar

transcription of the words is given in square brackets [ ]; the

unfamiliar phonetic symbols will be explained below in the dis-
cussions of d and other consonants).

(2) Palauan Spelling Phonetic Transcription English Gloss

dub [dup, ðup] ‘dynamite’
dmak [tmakh] ‘together’
dngod [θŋoð] ‘tattoo needle’

To most native speakers of Palauan, the sounds represented

by the letter d in the above three words probably all sound
alike. To the linguist-phonetician, however, they are very dif-
ferent sounds: that is why the linguist uses four different pho-
netic symbols [d, t, θ, ð] to represent what speakers of Palauan
think of as only one sound and what is spelled with the letter
d. The differences among [d, t, θ, ð] which the linguist-pho-
netician hears are non-contrastive or non-significant; such non-
contrastive sounds are called allo phones of a particular sound
(or phoneme). By studying the distribution of sounds in a lan-
guage (i.e., where the sounds do and do not occur), the linguist
can predict which allophones of a particular phoneme will occur
in a given environment. While contrastive sounds (or phonemes)
provide us with minimal pairs, as in the case of tub ‘spit’—dub
‘dynamite’ mentioned above, non-contrastive sounds (or allo-
phones of a particular phoneme) never give us minimal pairs, as
will be illustrated below.
In the following discussion of the sound system of Palauan,
we will examine the significant sounds (phonemes) of the lan-
guage, as well as some of the non-significant variations (allo-
phones) of these sounds. In addition, some discussion of the
distribution of these sounds will be given: In the discussion it
will be necessary to introduce some technical linguistic terms
and concepts which are essential to our understanding of how
the Palauan sound system works.

1.3. THE CONSONANTS OF PALAUAN

The consonants of Palauan can be seen at a glance in the fol-
lowing chart. Notice that the chart includes labels arranged
horizontally across the top and vertically along the left side. It
will be worthwhile to learn these new terms and to associate
them with the facial diagram (Fig. 1) which shows the important

3
 1 Sounds and Spelling of Palauan

articulators and points of articulation which are used in

making the sounds of Palauan. The sounds of speech are pro-
duced when a particular articulator—e.g., the tip, blade, or
back of the tongue—moves and touches some point of articu-
lation—e.g., the teeth. In the production of certain sounds, pairs
of speech organs serve simultaneously as articulator and point
of articulation; this happens, for example, when the lips are
brought together or when the vocal cords are closed against
each other. A thorough understanding of how the sounds are
produced and their relationships to each other will help in un-
derstanding the many complex changes that Palauan sounds un-
dergo when different words or parts of words come together
during conversation.

CHART OF PALAUAN CONSONANTS

Points of Articulation

Manner of bilabial dental alveolar velar glottal

articulation
Stops voiceless t k ch
voiced b d
Fricative s
Nasals m ng
Liquids r, l

In the above chart, the five terms along the top—bilabial,

dental, alveolar, velar, and glottal—represent the different
points of articulation at which consonants are pronounced,
while the four terms at the left—stops (voiceless or voiced),
fricative, nasals, and liquids—describe different manners (or
ways) of articulation.

1.3.1. STOPS
The largest subtype of Palauan consonants consists of the stops
b, t, d, k, and ch. These sounds are called stops because in their
production the outward flow of air which originated in the lungs
is completely stopped at some point in its passage through the
throat or mouth; this stoppage is achieved by placing some
articulator against some point of articulation. Two Palauan
stops, b and d, are labelled voiced, because when either of
them occurs at the beginning of a word before a vowel (i.e.,

4
 Palauan Reference Grammar

before a, i, e, o, or u—see 1.4 below), they are pronounced with

a simultaneous vibration of the vocal cords. This vibration is
caused by air passing through the vocal cords, thus producing a
“buzzing” sound called voicing. The term voiceless is used to
describe the stops t, k, and ch; in the production of such sounds,
the vocal cords do not vibrate but remain at rest, and the air is
allowed to pass quietly between them. (In addition to the three
voiceless stops t, k, and ch, Palauan has one other voiceless
sound—s, which will be examined in some detail below.)

DIAGRAM OF THE HUMAN SPEECH ORGANS

Drawing by Vanna Condax

Figure 1

Bilabial Stop b. The consonant represented by the letter b is

called a bilabial stop because the outward air flow is stopped
completely by closing the two lips. When it occurs at the be-
ginning of a word before an l or a vowel, it is voiced (phonetic
symbol [b]), as in the following examples:

(3) Palauan Phonetic English Gloss

Spelling Transcription

5
 1 Sounds and Spelling of Palauan

blai [blay] ‘house’

bloes [bloεs] ‘shot’
bai [bay] ‘community
house’
bung [buŋ] ‘flower’
beot [bεoth] ‘easy’
bilis [bilis] ‘dog’

When the consonant b occurs in certain environments, its

pronunciation changes. For example, b is pronounced without
voicing when it immediately precedes or follows another conso-
nant (except l) in the same word. In other words, in such
environments, b becomes a voiceless bilabial stop, which is
represented by the phonetic symbol [p]. The different pronunci-
ations of the consonant b as voiced [b] or voiceless [p] are non-
significant; they are predictable from the kind of environment
in which b appears. We can therefore say that the different
pronunciations of b as [b] and [p] are allophones of b. As we
mentioned at the end of 1.2 above, non-contrastive sounds (or
allophones of a given phoneme) never provide us with minimal
pairs. Since [b] and [p] are non-contrastive sounds, we never
find Palauan minimal pairs like *pung— bung. (The asterisk
mark * is used to identify items which do not occur in the
Palauan language.) Some examples showing the voiceless allo-
phone of b preceding or following another consonant are given
here:

(4) btuch [ptuɂ] ‘star’

bsibs [psipsǝ] ‘drill’
bngak [pŋakh] ‘my flower’
brer [prεr] ‘raft’
tbak h
[tpak ] ‘my spit’
kbokb [kpokpǝ] ‘wall’

We can conclude that the Palauan consonant b is a phoneme

which has two allophones—[b] and [p]. The voiced allophone [b]
occurs at the beginning of words (i.e., word-initially) before
vowels and the consonant l, and between two vowels, as in oba
[oba] ‘have, carry’ and rubak [rubakh] ‘old man’. On the other
hand, the voiceless allophone [p] occurs before or after conso-
nants other than l and at the end of words (i.e., word-finally),
as in tub [tup] ‘spit’ and bob [bap] ‘above’.

6
 Palauan Reference Grammar

Dental Stops t and d. The consonants t and d are voiceless and

voiced, respectively. Like the consonant b, they are stops, be-
cause they involve complete stoppage of the outward air flow.
But while the closure for b is bilabial, the closure for t and d is
dental. This dental closure is produced by placing the tongue
tip (the articulator) against the back of the upper teeth (the
point of articulation). In comparing b on the one hand with t
and d on the other, we can say that the three sounds are the
same with respect to manner of articulation, since they are
all stops; but b differs from t and d in point of articulation,
since the former is produced at the lips, while the latter are
articulated in back of the teeth. Recall how this is shown in
the chart of Palauan consonants given in 1.3, where the terms
across the top represent the five points of articulation at which,
consonants are found, while the terms listed at the left describe
the four possible manners of articulation. Note further that t
and d are identical in both point of articulation and manner
of articulation—i.e., they are both dental stops; what differen-
tiates them is the presence or absence of voicing.
Both t and d have allophones which are determined by the
environment in which each of them occurs. The consonant t
has two different pronunciations. When t occurs at the end of
a word it is released quite strongly with an audible puff of air.
This type of articulation is known as aspiration; the phonetic
symbol for such an aspirated t is [th]. The aspiration (or puff of
air) that accompanies a word-final t in Palauan can be heard in
the following words:

(5) liluut [liluwth] ‘returned’

dakt [ðakth] ‘fear’
1
chȩlat [ɂǝlath] ‘smoked (fish)’

The other pronunciation of t involves no aspiration; this

unaspirated t is represented by the phonetic symbol [t]. This
allophone of t occurs at the beginning of words (when either a
consonant or vowel follows) and within words (or word-inter-
nally) when it is preceded by another consonant and followed
by a vowel. The following examples illustrate these environ-
ments:

7
 1 Sounds and Spelling of Palauan

(6) tbak [tpakh] ‘my spit’

tmuu [tmuw] ‘enter’
tkul [tkul] ‘its edge’
tȩruich [tǝruyɂǝ] ‘ten’
tuu [tuw] ‘banana’
tȩchang [tǝɂaŋ] ‘who?’
rȩktel [rǝktεl] ‘his sickness’

We can therefore say that the Palauan phoneme t has two al-
lophones whose distribution is predictable as specified below:

[th] (aspirated)—word-finally
[t] (unaspirated)—elsewhere (i.e., word-initially and
word-internally).

The consonant d has four possible pronunciations, de-

pending on the environment in which it occurs. To illustrate
these different pronunciations, it will be necessary to introduce
two new phonetic symbols. They are [θ], which sounds like the
first sound of the English word thin, and [ð], which sounds like
the first sound of the English word then. Both of these sounds
are pronounced by putting the tongue tip against the back of
the upper teeth without complete closure; thus, they are not
stops, but fricatives. Fricatives involve partial closure or con-
striction between the articulator and point of articulation; their
characteristic quality is one of audible friction. While both [θ]
and [ð] are therefore dental frica tives, the former is voiceless
while the latter is voiced. They are not listed in the chart in 1.3
because they are not separate phonemes, but merely allophones
of the phoneme d.
The distribution of the four possible pronunciations of d may
be summarized as follows:

a. Word-initially before a vowel, the pronunciation of d

ranges from [d] to [ð]; in this environment, [d] tends to appear
in rapid, casual speech, while [ð] is heard in more careful, con-
trolled speech. Words which show this alternate pronunciation
include

(7) dub [dup, ðup] ‘dynamite’

deel [dεyl, ðεyl] ‘nail’

8
 Palauan Reference Grammar

b. Word-initially before a consonant, the pronunciation of d

varies between [t] and [θ]. The allophone [t] tends to appear
before b and m and in rapid speech in general, while [θ] occurs
before k and ng and in careful pronunciation. Some words
illustrating these allophones are listed below:

(8) dmak [tmakh] ‘together’

dbak [tpakh] ‘my dynamite’
dngod [θŋoð] ‘tattoo needle’

You may have noticed that tbak ‘my spit’ (cf. 6 above) and
dbak ‘my dynamite’ are pronounced in the same way, even
though they are spelled differently. This identical pronunciation
is reflected in identical phonetic transcriptions—namely, [tpakh]
for both words—and is due to the fact that t has the voiceless
allophone [t] word-initially (before any consonant) and d also
has the voiceless allophone [t] word-initially before a b. If we
heard the word [tpakh] spoken in isolation, we would therefore
be unable to tell whether the utterance meant ‘my spit’ or
‘my dynamite’; however, looking at the written forms poses no
difficulty, since tbak ‘my spit’ is spelled with the same word-
initial consonant as tub [tup] ‘spit’, while dbak ‘my dynamite’ is
spelled with the same initial consonant as dub [dup, ðup] ‘dy-
namite’.
c. When the consonant d occurs between vowels or at the
end of a word, it is pronounced with the allophone [ð], as in the
following examples:

(9) mȩdal [mǝðal] ‘his face’

kȩdeb [kǝðεp] ‘short’
chȩdil [ɂǝðil] ‘mother’
bad [bað] ‘stone’
kid [kið] ‘we’
eangȩd [yaŋǝð] ‘sky’

Velar Stop k. The consonant represented by the letter k is called

a velar stop because it is articulated by raising the back of the
tongue (the articulator) against the velum (the point of artic-
ulation) to form a complete closure. This consonant has three
principal allophones, whose distribution is described below:

9
 1 Sounds and Spelling of Palauan

a. When k occurs word-finally, it is pronounced with aspi-

ration; the phonetic symbol for this aspirated allophone of k is
[kh]. Observe the words below:

(10) brak [prakh] ‘taro’

2
chȩrmek [ɂǝrmεkh] ‘my animal’
dȩrumk [ðǝrumkh] ‘thunder’

What other Palauan consonant which we have studied thus far

has an aspirated allophone with the same kind of distribution?
b. Word-initially (before a consonant or vowel), k is pro-
nounced with the unaspirated allophone [k], as the following ex-
amples show:

(11) klou [klow] ‘big’

kmarȩd [kmarǝð] ‘light’
kid [kið] ‘we’
ker [kεr] ‘question’

This allophone also appears word-internally when k is next to

any consonant except l, as in lotkii [lotkiy] ‘remembers it’, skuul
[skuwl] ‘school’, and kbokb [kpokpǝ] ‘wall’.
c. Between vowels, k is pronounced with the voiced allo-
phone [g], as in the words below:

(12) olȩkiis [olǝgiys] ‘wake up’

mȩkeald [mǝgεalðǝ] ‘warm’
rȩkas [rǝgas] ‘mosquito’

Do you recall any other Palauan consonant which has a voiced

allophone between vowels?

Glottal Stop ch. The consonant ch is formed by closing the vocal

cords tightly against each other to impede the outward flow
of air. Since the space between the vocal cords is called the
glottis, the stop sound described here is known as a glottal
stop (phonetic symbol [ɂ]). Although a sequence of two
letters—ch—is used in the Palauan spelling system to write the
glottal stop, it is just a single consonant sound like b, t, d, or k.
The English sound system does not have a glottal stop phoneme,

10
 Palauan Reference Grammar

but speakers of American English frequently use [ɂ] in certain

words. For example, the negative expression uh uh is normally
pronounced with a glottal stop at the beginning of each of its
syllables.
The Palauan glottal stop phoneme shows no allophonic vari-
ation, and is pronounced as [ɂ] in all environments. A sampling
of words containing this sound is given here:

(13) charm [ɂarm] ‘animal’

chisel [ɂisεl] ‘news of him’
mȩched [mǝɂεð] ‘shallow’
dȩngchokl [ðǝŋɂoklǝ] ‘sit’
taoch [taoɂ] ‘channel’
tȩruich [tǝruyɂǝ] ‘ten’

Some special comment needs to be made about when and

when not to spell words with an initial ch. There are some
Palauan words which are pronounced with an initial glottal stop
under all circumstances—that is, regardless of whether they are
spoken in isolation or spoken following another word within a
sentence. For example, words like chad ‘person’ or chull ‘rain’
are pronounced as [ɂað] and [ɂul:ǝ], respectively, both when
spoken alone and when preceded by another word in simple
sentences like

(14) a. Ng chad ȩr a Siabal. [ŋɂaðǝrasyabal] 3

‘He’s Japanese.’

b. Ng chull. [ŋɂul:ǝ]
‘It’s raining.’

On the other hand, there are some words which have an

initial glottal stop when spoken in isolation, but lose this glottal
stop when preceded by another word. For example, words like
oles ‘knife’ and omes ‘see’ are [ɂolεs] and [ɂomεs], respectively,
when uttered in isolation. Note, however, that the initial [ɂ] dis-
appears in simple sentences like

(15) a. Ng oles. [ŋolεs]

‘It’s a knife.’

b. Ak mla omes ȩr ngii. [akmlaomεsǝrŋiy]

‘I’ve seen him.’

11
 1 Sounds and Spelling of Palauan

As the Palauan spelling of the words under discussion in-

dicates, a word is always spelled with initial ch if the [ɂ]
pronunciation is maintained within sentences; on the other
hand, if no initial [ɂ] is pronounced when a word appears in
a sentence, then no initial ch is ever included in the spelling.
When you are in doubt about whether or not to spell a word with
an initial ch, you can easily test it by using it in sentences like
14–15.

1.3.2. FRICATIVE
We have seen in 1.3.1. above that the stop consonants of
Palauan are characterized by complete stoppage or interruption
of the outward air flow; this is achieved by placing some ar-
ticulator tightly against some point of articulation. It is also
possible to produce consonants by forming a partial closure
or constriction between articulator and point of articulation.
When the outward flow of air is forced through such a narrow
passage, audible friction is heard. For this reason, such sounds
are called fricatives. Palauan has only one fricative, the
phoneme s. This sound, which can be identified by a strong
“hissing” quality, is produced by touching the sides of the blade
of the tongue against the teeth and part of the alveolar ridge,
which is the bony protrusion above the teeth. Because the
alveolar ridge is involved in its pronunciation, s is classified as
an alveolar fricative in the chart in 1.3. Palauan s is always
voiceless and usually sounds as if it is somewhere between
English sh (as in she) and s (as in see). It does not show any allo-
phonic variation and is pronounced as [s] in all positions. Some
words containing s are now given:

(16) sils [sils] ‘sun’

sers [sεrs] ‘garden’
mȩsilȩk [mǝsilǝkh] ‘wash’
mȩngiis [mǝŋiys] ‘dig’

1.3.3. NASALS
In Fig. 1 you will notice that the outward air flow from the lungs
can escape either through the mouth or the nasal passage. In
most languages, either one or the other of these “escape routes”
is closed off during the production of consonant sounds. All of
the Palauan stops and the fricative s, for example, involve air

12
 Palauan Reference Grammar

passing through the mouth only; during the pronunciation of

these sounds, air is prevented from entering the nasal passage
by raising the velum against the back wall of the throat (see
Fig.1). On the other hand, Palauan nasal sounds are made
by forming a closure somewhere in the mouth and leaving the
velum at rest so that the air flow can pass freely through the
nasal passage.

Bilabial nasal m. The nasal sound m is produced simply by

holding the lips tightly closed and letting the air escape through
the nose. Because the two lips are used to make the closure,
this nasal is identified as bilabial. (What is the other bilabial
consonant of Palauan?) The bilabial nasal m has no allophonic
variants and is pronounced [m] in all of its occurrences:

(17) mad [mað] ‘die’

omoes [omoεs] ‘shoot’
blim [blim] ‘your house’

Velar nasal ng. In articulating the nasal sound ng, the speaker
blocks off the passage of air through the mouth with a closure
between the back of the tongue (articulator) and the velum
(point of articulation), while leaving the nasal passage open.
Since the air is prevented from entering the mouth at the velum,
this nasal is classified as velar. (What is the other velar con-
sonant of Palauan?)
Though represented in Palauan spelling with a sequence of
two letters, the velar nasal ng is one single sound. It has two
principal allophones, whose distribution is specified as follows:

a. Before t, d, s, and r, ng is pronounced as a dental nasal

(phonetic symbol [n]). Because this allophone is a nasal, the out-
ward air flow passes through the nose, but the closure in the
mouth is made by placing the tongue tip against the back of the
upper teeth (cf. the articulation of t and d). Some examples con-
taining the [n] allophone of ng are now given:

(18) iungs [yuns] ‘island’

mȩrangd [mǝranðǝ] ‘(a kind of) coral’
sȩngsongd [sǝnsonðǝ] ‘stick’
ngduul [ṇduwl] ‘clam’

13
 1 Sounds and Spelling of Palauan

ngriil [ṇriyl] ‘place near beach’

ng til [ṇtil] ‘it’s her purse’

The dot in the phonetic representation [ṇ] means that the dental
nasal allophone of ng is pronounced as a separate syllable—see
1.3.5. below.

b. In all environments distinct from those described in the

preceding paragraph, the consonant ng is pronounced as a velar
nasal (the phonetic symbol for this sound, which is found at
the end of English words like sing, is [ŋ]). In other words, the
allophone [ŋ] appears before vowels, in word-final position, and
before consonants other than t, d, s, and r. Some examples are
listed below:

(19) ngau [ŋaw] ‘fire’

ngor [ŋor] ‘mouth’
reng [rεŋ] ‘heart, spirit’
bung [buŋ] ‘flower’
ngklem [ŋklεm] ‘your name’
nglim [ŋlim] ‘drunk (up)’
ng chȩtik [ŋɂǝtikh] ‘I don’t like it.’

Is there any good explanation we can give for the distri-

bution of the allophones of ng? Notice that the dental nasal
allophone [n] occurs only before sounds which are dental or
alveolar; in other words, this allophone precedes sounds whose
point of articulation (dental or alveolar) is close to its own. On
the other hand, the velar nasal allophone [ŋ] has a less re-
stricted distribution, since it occurs before all vowels, in word-
final position, and before consonants such as k, ch, and l. If we
assume that because of this less restricted distribution the al-
lophone [ŋ] is somehow more “basic” than [n], we can say that
[ŋ] changes to [n] before t, d, s, and r because speakers move
the point of articulation of the nasal forward (from a velar to a
dental position) in anticipation of the pronunciation of the fol-
lowing consonant. This process, which is very common in lan-
guages, is called assimilation. In the case under discussion, we
say that [ŋ] has assimilated to (or has become similar in pro-
nunciation to) a following t, d, s, or r, thereby becoming [n].

14
 Palauan Reference Grammar

The only exceptions to the above-mentioned distribution of

the allophones of ng are found among words which have been
borrowed into Palauan from Japanese and English. In such
words the allophone [n] appears in environments other than
before t, d, s, and r. Some examples are nas [nas] ‘eggplant’,
niziu [niǰuw] ‘twenty (used often when counting change)’, John
[ǰan], etc. In spelling these words, Palauans use the single letter
n rather than the letter sequence ng. Note, further, that in
spelling a word of Japanese origin such as sensei ‘teacher’ the
single letter n is used instead of ng, even before the dental con-
sonant s.
Special mention needs to be made about when and when not
to spell ng at the end of one and the same word. Quite a few
Palauan words end in a, o, or u when pronounced within a sen-
tence, but they have a word-final ng when spoken in isolation or
at the end of a sentence. This common rule of Palauan pronun-
ciation is illustrated in the sentences below, where the words
mȩnga ‘eat’ and mo ‘go’ are spelled in two different ways:

(20) a. Ak mo mȩngang.
‘I’m going to eat (it).’

b. Ak mo mȩnga ȩr a ngikȩl.
‘I’m going to eat the fish.’

(21) a. Ng mong.
‘He’s going.’

b. Ng mo ȩr a skuul.
‘He’s going to school.’

The rule for spelling words of this kind is simple to remember: if

word-final ng is pronounced and heard, as in 20a and 21a above,
we also spell ng; if, however, no ng is pronounced or heard, as
in 20b and 21b, it is omitted from the spelling. When words like
mȩnga ‘eat’ and mo ‘go’ are cited for discussion in this text, they
will be cited in the shorter form.
In addition to the above, there are many Palauan words
which are always pronounced with a final ng, even within sen-
tences. Words of this type, which of course are always spelled
with word-final ng, include native Palauan words like bung

15
 1 Sounds and Spelling of Palauan

‘flower’, bang ‘goatfish’, ding ‘ear’, reng ‘heart’, and chȩdeng

‘shark’, and borrowed words like hong ‘book’ and blatong
‘plate’.

1.3.4. LIQUIDS
The Palauan consonants which we have already discussed ex-
hibit three different types of articulation. The non-nasal con-
sonants involve either complete closure (the stops b, t, d, k,
and ch) or narrowing (the fricative s) of the speech tract. The
nasal con sonants m and ŋ are characterized by closure in one
part of the speech tract (i.e., the mouth) and free passage in the
other (i.e., the nose). In this section we will examine a fourth
type of Palauan consonant—the liquids l and r. Both of these
consonants are articulated by making a partial closure in the
mouth.

Liquid 1. The consonant l is made by touching the upper surface

of the blade of the tongue against the top teeth and alveolar
ridge and by allowing some air to escape over the sides of the
tongue. It has no significant allophonic variations and appears
as [l] in all environments. The following are some examples con-
taining l:

(22) lius [lius] ‘coconut’

luut [luwt ] h ‘return’
mȩlai [mǝlay] ‘take’
rael [raεl] ‘road’

Liquid r. The consonant r is called a tapped r because it is made

with a quick tapping movement of the tongue tip against the
alveolar ridge above the upper teeth. Though there is a special
phonetic symbol for this tapped r, it will be adequate for our
purposes to use [r]. This sound appears in all environments, as
in the words below:

(23) rakt [rakth] ‘sickness’

rȩkas [rǝgas] ‘mosquito’
bȩras [bǝras] ‘rice’
kar [kar] ‘medicine’

16
 Palauan Reference Grammar

1.3.4.1. The Sequences ll and rr

In some Palauan words, two identical liquid consonants occur
next to each other. The sequence ll differs from l in that it is
held about twice as long as the single consonant. The phonetic
representation for this long l is [l:], where the colon [:] indi-
cates the extra length. The sequence rr differs from r in that
it is pronounced as a trilled r rather than a tapped r. A trilled
r (phonetic symbol [r̄]) is composed of two or three tapped r’s
pronounced in rapid succession. The words below illustrate the
Palauan sequences ll and rr in various positions:

(24) llel [l:εl] ‘its leaf’

kall [kal:ǝ] ‘food’
rrom [r̄om] ‘liquor’
kȩrrȩkar [kǝr̄ǝgar] ‘tree’
rruul 4 [r̄uwl] ‘made, done’

1.3.5. SYLLABIC CONSONANTS

The words of Palauan can consist of different numbers of syl
lables, or pulses of air. It is fairly easy to count syllables: for
example, kar ‘medicine’ has one, elii ‘yesterday’ has two, mȩ
dȩngȩltȩrir ‘knows them’ has five, and so on. When any con-
sonant occurs before a vowel, it is pronounced along with that
vowel as part of the same syllable. In bilek ‘my clothing’, for
instance, b is part of the first syllable and l is part of the
second. When certain types of consonants—specifically, nasals
and liquids—occur before other consonants in word-initial po-
sition, they become syllabic—that is, they are pronounced as
separate syllables. To indicate this syllabic quality in the pho-
netic transcription, a dot is placed under the regular phonetic
symbol for the nasal or liquid—i.e., [ṃ], [ŋ], [ḷ], and [ṛ]. These
syllabic consonants appear in cases like the following:

(25) ng boes [ṃboεs] 5 ‘it’s a gun’

mchiiȩlak [ṃɂiyǝlak ] h ‘wait for me!’
Ngchesar [ŋɂεsar] (village name)
nglim [ŋlim] ‘drunk (up)’
ngduul [ṇduwl] ‘clam’
lmangȩl [ḷmaŋǝl] ‘cry’
ltel [ḷtεl] ‘his return’
rsȩchek [ṛsǝɂεkh] ‘my blood’

17
 1 Sounds and Spelling of Palauan

rtangȩl [ṛtaŋǝl] ‘is to be pounded’

Syllabic r ([ṛ]) is pronounced as a trilled r by some speakers

and with considerable friction by others. The only exception to
the analysis given above concerns the sequence ml, as in mlai
[mlay] ‘canoe’. Here, the m is not syllabic but is pronounced
along with the other sounds in the word as a single syllable.

1.4. THE VOWELS OF PALAUAN

The vowels of Palauan are summarized in the following chart:

CHART OF PALAUAN VOWELS

Tongue advancement

Tongue height front central back

high i u
mid e ȩ o
low a

Along the left side of the chart, three levels of tongue height
are shown, while along the top, three degrees of tongue ad-
vancement are indicated. These terms will be explained in
detail below. Unlike consonants, which involve closure or nar-
rowing of the speech tract, vowels allow relatively free, unre-
stricted passage of the outward air flow. The different vowel
sounds (or vowel qualities) are produced by changing the
shape of the mouth cavity; this is accomplished by holding the
tongue in various positions, each of which can be described in
terms of tongue height and tongue advancement. All Palauan
vowels are automatically voiced (i.e., the vocal cords vibrate
during their production) and are pronounced with the velum
raised to shut off the nasal passage (i.e., they do not have a
“nasal” quality).
Palauan vowels distinguish three degrees of tongue height—
high, mid, and low. High vowels are pronounced with the
tongue raised high in the mouth and very close to the palate;
low vowels are articulated with the tongue low in the mouth,
relatively distant from the palate; and mid vowels are pro-
nounced somewhere in between. To get some idea of the “dis-
tance” between high vowels and low vowels (which are at the

18
 Palauan Reference Grammar

“extreme” ends of the series), simply watch what happens to

your mouth during the pronunciation of pairs of vowels like i-
a and u-a. When you move from the high vowels i and u to
the low vowel a, your mouth opens widely; here, the jaw is
lowered in order to get the tongue into a low position. Now try
to pronounce the vowel sequence i-e-a; you should be able to
recognize three different positions of vowel height as your jaw
moves progressively lower. For further practice, move in the op-
posite direction from low to high—i.e., a-e-i.

1.4.1. HIGH VOWELS i AND u

Differences in tongue height are not sufficient to distinguish all
of the Palauan vowels from each other. For example, the chart
in 1.4 above shows that Palauan has two high vowels—i and u.
While both of these vowels are articulated with the tongue in a
relatively high position, they differ from each other with respect
to tongue advancement. In pronouncing i, the blade (or front) of
the tongue is advanced and raised towards the alveolar ridge
and the front portion of the palate. In pronouncing u, however,
the tongue is retracted and the back of the tongue is raised
towards the back portion of the palate and the velum. Because
the tongue is advanced towards the front of the mouth for i,
this vowel is identified as a high front vowel; and because
u involves a retraction of the tongue towards the back of the
mouth, it is labelled as a high back vowel. It is not all that
easy to observe or feel the difference in tongue advancement
between i and u. However, if you try to repeat these vowels in
succession (i-u, i-u, etc.) you may be able to feel the tongue re-
tract as you move from i to u. One further difference between
these two vowels is easier to recognize. Notice that when you
pronounce u, your lips become rounded as if you are going to
whistle; this rounding is absent for i, where your lips remain
spread apart, as if you are beginning to smile. Thus, we say that
u is a rounded vowel while i is an unrounded vowel.
Although vowels, like consonants, can have allophones, the
vowels of Palauan in general show little allophonic variation.
Therefore, as the words below illustrate, i is pronounced [i]
(similar to the vowel sound in English heat) and u is pronounced
[u] (similar to the u in English rude) under all circumstances:

(26) sils [sils] ‘sun’

kim [kim] ‘large clam’

19
 1 Sounds and Spelling of Palauan

chisel [ɂisεl] ‘news of him’

mȩtik [mǝtik ] h ‘find’
btuch [ptuɂ] ‘star’
bung [buŋ] ‘flower’
kȩruk [kǝrukh] ‘my medicine’
subȩlek [subǝlεkh] ‘my homework’

1.4.2. MID VOWELS e, ȩ, AND o

While the high vowels show two degrees of tongue ad-
vancement, the mid vowels show three. In addition to the mid
front vowel e and the mid back vowel o, we have the mid
central vowel ȩ . In the pronunciation of this vowel, the tongue
is neither advanced (as for e) nor retracted (as for o); rather,
the tongue remains flat and at rest. The mid central vowel ȩ
and the mid front vowel e are quite different in pronunciation:
ȩ sounds something like the weak vowel “uh” in English words
like “about” and “again”, while e sounds like the vowel in
English “bed.” The phonetic symbol for the mid central vowel
ȩ is [ǝ] (commonly referred to as schwa), and that for the mid
front vowel e is [ε]. As you know, the Palauan spelling system
uses only one letter—namely, e—to spell both of the sounds [ε]
and [ǝ]. Although speakers of Palauan will not find this con-
fusing, non-native speakers may have trouble deciding when to
pronounce the letter e as [ε] and when to pronounce it as [ǝ]. To
assist non-native speakers, in this text we shall use the special
symbol ȩ (e with a comma under it) to represent [ǝ]. As you will
see later, it will be handy to have the two symbols e and ȩ in
order to make certain discussions clear.
In the list below, you will find some common Palauan words
containing the mid vowels e and o. (Further discussion of ȩ will
continue below.)

(27) sers [sεrs] ‘garden’

ngklel [ŋklεl] ‘his name’
elii [εliy] ‘yesterday’
ngor [ŋor] ‘mouth’
oles [olεs] ‘knife’
mȩlȩcholb [mǝlǝɂolbǝ] ‘bathe’

20
 Palauan Reference Grammar

There is one important difference between the pronunciations

of e and o which we have not yet mentioned: o is a rounded
vowel, while e is not. Recall that among the high vowels, u
is rounded, while i is not. What similarity of pattern can you
identify?
The Palauan vowel ȩ has a very restricted distribution, since
it occurs only in unstressed syllables. Every Palauan word of
two or more syllables has just one stressed syllable, with the
remaining syllables unstressed. It is usually not too difficult to
identify the stressed syllable in such words, since this syllable
tends to be louder and stronger than the nearby syllables. For
practice, compare the stressed syllable with the unstressed
syllables in words like klúkuk ‘tomorrow’, mȩngȩlébȩd ‘hit’,
ngklém ‘your name’, and chillȩbȩdák ‘hit me’. To identify the
stressed syllable, a stress mark (') has been placed over the
vowel which is found in it; this stress mark, however, is not used
in the Palauan spelling system. Until now, we have not identified
the stressed syllables in our phonetic transcriptions of Palauan
multisyllabic words, although a completely specified phonetic
transcription would have to take account of them.
The list below contains words of two or more syllables which
have already appeared in this chapter. The stressed syllable has
been identified with a stress mark. Notice that every ȩ (schwa
= [ǝ]) which occurs is found in an unstressed syllable.

(28) chȩlat [ɂǝláth] ‘smoked (fish)’

tȩchang [tǝɂáŋ] ‘who?’
rȩktel [rǝktέl] ‘his sickness’
eangȩd [yáŋǝð] ‘sky’
kmarȩd [kmárǝð] ‘light’
olȩkiis [olǝgíys] ‘wake up’
mȩsilȩk [mǝsílǝkh] ‘wash’
lmangȩl [ḷmáŋǝl] ‘cry’
mȩngȩlebȩd [mǝŋǝlέbǝð] ‘hit’

Note further that since the great majority of Palauan one-syl-

lable words are stressed, there are almost no Palauan one-syl-
lable words containing ȩ. 6

21
 1 Sounds and Spelling of Palauan

1.4.3. LOW VOWEL a

The only low vowel in Palauan is a (phonetic symbol [a]), which
is classified as a low central vowel. There is no contrast be-
tween front and back low vowels in Palauan. Several words con-
taining this vowel are listed below:

(29) chad [ɂað] ‘person’

mȩlat [mǝlát ] h ‘tear, rip’
ngak [ŋakh] ‘I, me’

1.4.4. THE VOWEL ȩ AND THE PROCESS OF VOWEL REDUCTION

The six vowels listed in the chart in 1.4 above are among the
significant sounds (or phonemes) of Palauan. It is possible to
find minimal pairs which show how the various vowels con-
trast with each other in otherwise identical environments. Note,
for example, the minimal pairs kar ‘medicine’—ker ‘question’,
char ‘price’— chur ‘laughter’, kid ‘we’—ked ‘hill’, and delék
‘my nail’—dȩlék ‘my abdomen’. In the last minimal pair cited,
the vowels e and ȩ contrast in an unstressed syllable. Since ȩ
occurs only in unstressed syllables, it is extremely difficult to
find minimal pairs which show ȩ to be contrastive with other
vowels. For this and another reason to be explained below, many
linguists would not recognize the mid central vowel ȩ as a sep-
arate phoneme of Palauan, but would consider it an allophone
of some other vowel phoneme or phonemes.
When we compare related forms of certain words, we can
see a close relationship between ȩ and various other vowels. In
the list below, the left column gives the simple form of a word,
while the right column gives one of its “possessed” forms:

(30) Simple form Possessed form

bsibs ‘drill’ bsȩbsék ‘my drill’

chur ‘laughter’ chȩrík ‘my laughter’
sers ‘garden’ sȩrsék ‘my garden’
ngor ‘mouth’ ngȩrék ‘my mouth’
kar ‘medicine’ kȩrúk ‘my medicine’

The words in the left column contain instances of the vowels i,

u, e, o, and a. In each of the corresponding possessed forms, the
mid central vowel ȩ appears where we would expect the vowels

22
 Palauan Reference Grammar

i, u, e, o, or a. The possessed forms have two syllables (you will

notice that one of the endings -ek, -ik, or -uk has been added
to each of them), of which the first is unstressed and second is
stressed. Since the vowel ȩ appears in an unstressed syllable
where we would expect some other vowel, it seems as if i, u,
e, o, and a have changed to ȩ in this environment. This kind of
process, which is called vowel reduction, is observed in many
languages of the world: commonly, certain full vowels reduce
to the “weaker” or more “neutral” mid central vowel schwa
under certain conditions. We shall now explain this statement
further.
If we rewrite the chart of vowels given in 1.4 as a kind of
“vowel triangle”—namely,

—we can see that ȩ is more or less in the middle while the other
vowels are at the edges or extremes. From the point of view of
tongue height and tongue advancement, the mid central vowel
ȩ is least extreme or deviant in its articulation: it is neither high
nor low, nor is it front or back. For this reason, the mid central
vowel ȩ may be described as a neutral vowel, while i, u, e,
o, and a are referred to as full vowels. In Palauan, then, as in
many languages, the full vowels lose their basic qualities (i.e.,
no longer sound like [i], [u], etc.) and reduce to a neutral vowel
(i.e., ȩ [a]) in unstressed syllables. Because ȩ therefore results
from (or is derived from) any of the full vowels, some linguists
would argue that it is not a separate phoneme but merely one of
the allophones of each of the full vowels. The process of vowel
reduction introduced by the examples of 30 above will be pre-
sented in greater detail in 3.4.

23
 1 Sounds and Spelling of Palauan

1.4.5. OTHER OCCURRENCES OF SCHWA

You may have noticed that our phonetic transcriptions for
certain words show word-final schwas which are not reflected in
the Palauan spelling. A few of these words are repeated, along
with new examples, in the list below:

(31) bsibs [psípsǝ] ‘drill’

kbokb [kpókpǝ] ‘wall’
mȩkeald [mǝkεálðǝ] ‘warm’
dȩngchokl [ðǝŋɂóklǝ] ‘sit’
mȩrangd [mǝránðǝ] ‘(a kind of) coral’
mȩlȩcholb [mǝlǝɂólbǝ] ‘bathe’
ralm [rálmǝ] ‘water’
diall [ðiál:ǝ] ‘ship’

The appearance of word-final [ǝ] in the examples of 31 illus-

trates a very general rule of Palauan pronunciation: whenever
a word ends in a sequence of two consonants, this cluster
of consonants is followed by a schwa release (which is of
course unstressed). Because the schwa release is predictable
and speakers always pronounce it automatically, it does not
need to be indicated in the spelling.
In some instances, a word-final u preceded by a consonant
also results in a schwa release; here, too, the schwa is not re-
flected in the spelling. Some words which fall into this category
include the following:

(32) Palauan Phonetic English gloss

spelling 7 transcription
omdasu [omðáswǝ] ‘think’
ochadu [oɂáðwǝ] ‘something to cut with,
tongs’
kuoku [kwókwǝ] ‘skin which is shed’

Sometimes a schwa is predictably added to break up a parti-

cular cluster of consonants. For example, sequences of the form
dental consonant + l (i.e., tl, dl, and sl) never occur in Palauan.
It is also impossible for ch to be directly preceded or followed
by another consonant. When such “impossible” combinations
result from certain types of word formation, a schwa must be
inserted to separate the consonants. The words below, for ex-

24
 Palauan Reference Grammar

ample, are formed by inserting an l after the first consonant;

the resulting consonant cluster must be broken up with an in-
tervening schwa, which is also indicated in the Palauan spelling.

8
(33) Palauan spelling Phonetic transcription English gloss
tȩlub [tǝlúb] ‘spat’
dȩlangȩb [ðǝláǝb] ‘covered’
sȩlesȩb [sǝlέsǝb] ‘burned’
chȩlat [ɂǝláth] ‘smoked’

Schwa often occurs next to or between vowels which are not

stressed. In such cases, it is indicated in the spelling, as the fol-
lowing examples illustrate:

9
(34) Palauan spelling Phonetic transcription English gloss
chuiȩuíi [ɂuyǝwíy] ‘reads it’
kiiȩsíi [kiyǝsíy] ‘digs it’
kiuȩtíi [kiwǝtíy] ‘cuts it’
siuȩsíi [siwǝsíy] ‘cures it’

1.4.6. LONG VOWELS

All the full vowels of Palauan except a can occur long. These
long vowels are spelled simply by doubling the letter—i.e., ii,
uu, ee, and oo. Phonetically, Palauan long vowels are indeed
greater in length (i.e., time it takes to say them) than the corre-
sponding short vowels, but they also have some additional fea-
tures. All of the long vowels contain a gliding articulation. The
front vowels i and e are followed by a y-glide when long, while
the back vowels u and o are followed by a w-glide. The glide
sounds y and w involve movement of the tongue towards a high
front or a high back position, respectively. In forming ee, for ex-
ample, the tongue makes a smooth transition from the mid front
position of e to a high front position; similarly, in articulating oo,
the tongue begins at the mid back position for o and then moves
towards a high back position. In this text, long vowels are in-
dicated in the phonetic transcriptions as sequences of vowel +
glide, and we will use the phonetic symbols [y] and [w] to stand
for these glides. The list below contains words with long vowels:

(35) diil [ðiyl] ‘abdomen’

ngii [ŋiy] ‘he, she, it’
buuch [buwɂǝ] ‘betel nut’

25
 1 Sounds and Spelling of Palauan

ngduul [ṇduwl] ‘clam’

deel [dεyl] ‘nail’
kmeed [kmεyð] ‘near’
dȩkool [dǝgowl] ‘cigarette’
sȩkool [sǝgowl] ‘playful’

The difference between short vowels and long vowels in

Palauan is phonemic because it can serve to distinguish be-
tween otherwise identical words. This phonemic difference
means that we can find minimal pairs in which a short vowel
contrasts with a long vowel in exactly the same envi-
ronment—e.g., buch ‘spouse’—buuch ‘betel nut’ and chis ‘de-
pression in the sea floor’— chiis ‘escape’.

1.4.7. VOWEL CLUSTERS

As many of the words in this chapter illustrate, Palauan con-
sonants can occur in different combinations or clusters; such
clusters are found in words like mlai ‘canoe’, tkul ‘its edge’,
brer ‘raft’, and rakt ‘sickness’. It is also possible for the full
vowels of Palauan to appear in various kinds of clusters. Of the
two adjacent vowels, one or the other may be stressed, or—less
frequently—neither may be stressed. Given only the Palauan
spelling of words containing vowel clusters, it is very difficult to
predict the correct pronunciation. This is because some of the
(spelled) vowels are pronounced in different ways, depending
on whether or not they are stressed and whether they precede
or follow the adjacent vowel. Before formulating some of these
very complicated rules of pronunciation, let us list a represen-
tative number of Palauan words which contain vowel clusters.
Stress marks are indicated in the Palauan spellings and in the
phonetic transcriptions because we will need to refer to stress
in the rules of pronunciation.

(36) 1. eángȩd [yáŋǝð] ‘sky’

2. eólt [yóltǝ] ‘wind’
3. iédȩl [yέðǝl] ‘mango’
4. iúngs [yúns] ‘island’
5. eungél [εuŋέl] ‘under it’
6. oách [wáɂ] ‘leg’
7. uél [wέl] ‘turtle’
8. uingȩl [wiŋǝl] ‘tooth’
9. soál [soál] ‘his wish’

26
 Palauan Reference Grammar

10. cháus [ɂáws] ‘lime’

11. ngáu [ŋáw] ‘fire’
12. kléu [klέw] ‘young coconut’
13. udóud [uðówð] ‘money’
14. klóu [klów] ‘big’
15. suélȩb [swέlǝb] ‘noon’
16. suóbȩl [swóbǝl] ‘study, homework’
17. báil [báyl] ‘clothing’
18. róis [róys] ‘mountain’
19. búil [búyl] ‘moon’
20. chúi [ɂúy] ‘hair’
21. blái [bláy] ‘house’
22. tȩkói [tǝgóy] ‘word’
23. sȩchȩléi [sǝɂǝlέy] ‘friend’
24. díak [ðíakh] ‘isn’t’
25. líus [líus] ‘coconut’
26. diáll [ðiál:ǝ] ‘ship’
27. ráel [ráεl] ‘road’
28. bóes [bóεs] ‘gun’
29. táoch [táoɂ] ‘channel’
30. díong [ðíoŋ] ‘bathing place’

In order to account for the phonetic transcriptions of 36, we

need to formulate rules of pronunciation like the following:

a. Word-initially before any stressed vowel, the unstressed

front vowels e and i are pronounced as the glide [y] (see items
1–4 in the list above). If word-initial unstressed e precedes an-
other unstressed vowel, however, as in item 5, it is pronounced
[ε].

b. Word-initially before any stressed vowel, the unstressed

back vowels o and u are pronounced as the glide [w] (see
items 6–8). Word-internally before a stressed vowel, however,
unstressed o is pronounced [o], as in item 9.

c. Word-internal or word-final unstressed u’s are pro-

nounced as the glide [w], regardless of whether a stressed
vowel precedes them (as in items 10–14) or follows them (as in
items 15–16).

27
 1 Sounds and Spelling of Palauan

d. Following a stressed vowel word-internally and word-fi-

nally, i is pronounced as the glide [y] (see items 17–23). Pre-
ceding a vowel word-internally, however, i is pronounced [i],
whether it is stressed (as in items 24–25) or unstressed (as in
item 26).

e. Word-internally following stressed vowels, e is pro-

nounced [ε], as in items 27–28, and o is pronounced [o], as in
items 29–30.

We must give special attention to the problem of when and

when not to spell a word with a final vowel cluster ei. A good
number of Palauan words which end in e when pronounced
within a sentence take a word-final i when spoken in isolation
or at the end of a sentence. The sentences below, which contain
me ‘come’ and che ‘fishing’, illustrate this variation in pronunci-
ation:

(37) a. A Droteo a mei.

‘Droteo is coming.’

b. A Droteo a me ȩr a blik.
‘Droteo is coming to my house.’

(38) a. Ng mo ȩr a chei.
‘He is going fishing.’

b. Ng mo ȩr a che ȩr a klukuk.
‘He is going fishing tomorrow.’

The rule of spelling involved here is similar to that observed for

word-final ng at the end of 1.3.3 above: in other words, if word-
final i is pronounced and heard, as in 37a and 38a, it is also
spelled; if, on the other hand, no i is pronounced or heard, as in
37b and 38b, it is not included in the spelling. When cited for
discussion in this text, words like me ‘come’ and che ‘fishing’
will be cited in the shorter form.

28
 Palauan Reference Grammar

1.5. FURTHER RULES OF PALAUAN SPELLING

In this section we will list various spelling rules of Palauan
which have not been covered in the sections above. Most of the
rules below concern the proper spelling of individual words and
phrases. Often, the decision to spell something as a separate
word is based on a grammatical analysis of the item in question.
Such analysis allows us to identify or isolate one and the same
word as it appears in different, though related, constructions.
Our understanding of many aspects of Palauan grammatical
structure will be facilitated if we spell a particular word in the
same way in all of its occurrences, even though there might
be some differences in pronunciation from one occurrence to
another. In the discussion which follows, we will try to keep
references to grammatical terms and concepts at a minimum;
however, any terms or concepts which do need to be introduced
for purposes of identification will be given thorough treatment
elsewhere in the text.
a. The relational word ȩ r, which has a wide range of
English equivalents such as ‘in, at, to, from, out of, of, because
of’, etc., is always spelled as a separate word. Furthermore, the
word a, which precedes all verbs and nouns (but not pronouns
or demon stratives), should be spelled as a separate word. Ob-
serve the following examples:

(39) a. Ak mo ȩr a skuul.
‘I’m going to school.’

b. Kȩ mo ȩr ker?
‘Where are you going?’

c. Ak mȩsuub a tȩkoi ȩr a Merikel.

‘I am studying English.’

d. Ak milsuub ȩr a blik.
‘I was studying at home.’

e. Ak mȩrael ȩr a klukuk.
‘I am leaving tomorrow.’

f. Ng hong 10 ȩr a Droteo.
‘It’s Droteo’s book.’

29
 1 Sounds and Spelling of Palauan

g. Ng hong ȩr ngii.
‘It’s his book.’

h. A Droteo a milil ȩr tiang.

‘Droteo is playing here.’

The word ȩr is not pronounced identically in all the examples

of 39. If the preceding word is vowel-final, as in 39a–c, the ȩ of
ȩ r is dropped; thus, for example, the three words mo ȩr a of
39a are pronounced [mora]. On the other hand, if the preceding
word is consonant-final, as in 39d–h, the vowel of ȩr is retained,
giving [ǝr].
In certain cases, we know from the grammatical structure
that we have a sequence of the form ȩr + a + noun, even though
the a is not pronounced. The following expressions fall into this
category:

(40) er a elii [εrεlíy] ‘yesterday’

er a elȩchang [εrέlǝɂaŋ] ‘now, today’
er a Belau [εrbέlaw] ‘in/of Palau’

In the examples of 40, the vowel of ȩr is usually not reduced and

is therefore pronounced as [ε].
The word ȩr is spelled as a separate word in the following
special expressions:

(41) ngar ȩr ngii [ŋarŋíy] ‘there is’

mla ȩr ngii [mlarŋíy] ‘there was’
mo ȩr ngii [morŋíy] ‘there will be’
mochu ȩr ngii [moɂurŋíy] ‘there is about to be’

In the first two expressions of 41, we find the very common

Palauan verbs ngar ‘is (located)’ and mla ‘was (located)’. These
verbs also appear in sentences like the following:

(42) a. A Droteo a ngar ȩr a stoang.

‘Droteo is at the store.’

b. A John a mla ȩr a Guam.

‘John was in Guam.’

c. A Droteo ng ngar ȩr ker?

‘Where is Droteo?’

30
 Palauan Reference Grammar

d. A Toki ng mla ȩr ker?

‘Where was Toki?’

b. In order to show that a noun refers to more than one

human being, we attach rȩ to the beginning of that noun. For
example, while chad ‘person’ refers to one human being, rȩchad
‘people’ refers to two or more. Other examples of this contrast
include ngalȩk ‘child’—rȩngalȩk ‘children’, sensei ‘teacher’—
rȩsensei ‘teachers’, sȩchȩlik ‘my friend’—rȩsȩchȩlik ‘my
friends’, and ekȩbil ‘girl’—rekȩbil ‘girls’. As the last example
shows, if the word in question begins with a vowel, then we
simply attach r instead of rȩ. Some sentences containing plural
words like rȩchad ‘people’ are now given:

(43) a. Ak ulȩmes ȩr a rȩngalȩk.

‘I was watching the children.’

b. Ng delmȩrab ȩr a rȩsensei.
‘It’s the teachers’ room.’

c. A rȩlluich ȩl chad a mlad.

‘Twenty people died.’

d. Ak milstȩrir a rua Toki.

‘I saw Toki and her friends.’

c. The word ȩl is used in many kinds of constructions to

relate one word to another. Some of the most common usages
are illustrated below.

1. With demonstratives, which are used to point out

people, animals, or things:

(44) tia ȩl klalo ‘this thing’

se ȩl hong ‘that book’
tirka ȩl chad ‘these people’
aika ȩl charm ‘these animals’

In all of the examples of 44, we do not pronounce the ȩ of ȩl

because a vowel-final word precedes. Where have we seen a
similar rule of pronunciation?

31
 1 Sounds and Spelling of Palauan

2. With numbers, which occur in many different series, de-

pending on what is being counted:

(45) ta ȩl chad ‘one person’

chimo ȩl kluk ‘one dollar’
eru ȩl klok ‘two o’clock’
tȩluo ȩl oluchȩs ‘one pencil’
tȩruich mȩ a ta ȩl chad ‘eleven people’
dart ȩl kluk ‘one hundred dollars’
euid ȩl klok ‘seven o’clock’
tȩruich ȩl oluchȩs ‘ten pencils’

In some of the examples of 45, the ȩ of ȩl is not pronounced,

while in others it is. What factors determine this rule of pronun-
ciation?

3. With various kinds of modifiers, which describe or qualify

some other word:

(46) elȩcha ȩl sils ‘today’s weather’

mȩkȩlȩkolt ȩl ralm ‘cold water’
bek ȩl tutau ‘every morning’
mȩkngit ȩl chad ‘bad person’
ungil ȩl chad ‘good person’

In the expressions of 46, the modifying or qualifying word pre-

cedes the modified word. As we might expect, the ȩ of ȩl is not
pronounced when it follows a vowel-final word; the ȩ is pro-
nounced, however, after a consonant-final word. If the word pre-
ceding ȩl ends in an l, as in the case of ungil ȩ l chad ‘good
person’, then ȩl is completely omitted from the pronunciation:
[uŋil?að].
In addition to the expressions of 46, it is possible to have ex-
pressions in which the modifying or qualifying word follows the
word modified, as in

(47) chad ȩl mȩngitakl ‘person who sings’

soal ȩl mo ȩr a chei ‘his desire to go fishing’
sensei ȩl ungil ‘teacher who is good’
Droteo ȩl sensei ‘Droteo, who is a teacher’

32
 Palauan Reference Grammar

You should have no difficulty predicting how the word ȩl is pro-

nounced in the examples of 47.

4. With various types of complex constructions:

(48) dirrek ȩl sensei ‘is also a teacher’

di tȩlkib ȩl kukau ‘only a little taro’
mo mȩrek ȩl mȩsuub ‘finish studying’
omȩngur ȩl oba a taod ‘eat with a fork’
blȩchoel ȩl mȩsuub ‘always studies’
mo ȩl ngar ȩr a mlai ‘go in a car’
omuchȩl ȩl mȩsuub ‘begin studying’
mo ȩl obȩngkel a Toki ‘go with Toki’

Again, the correct pronunciation of ȩl in the examples of 48 can

be easily predicted.

d. Palauan has several sets of pronouns, which are short

words referring to various persons such as ‘I’, ‘you’, ‘he’, ‘they’,
etc. Some Palauan pronouns are spelled as separate words,
while others are not. A brief summary is given below.

1. The non-emphatic subject pronouns ak ‘I’, ng ‘he/she/

it,’ kȩ ‘you’, etc., are spelled as separate words, as in

(49) a. Kȩ mo ȩr ker?
‘Where are you going?’

b. Ak mȩnguiu ȩr a hong.
‘I’m reading the book.’

c. Tȩ di mililil.
‘They just fooled around.’

d. Ng kmal ungil.
‘It’s very good.’

2. The pronouns ngak ‘I’, ngii ‘he/she/it’, kau ‘you’, etc.,

which are used after the relational word ȩ r or as emphatic
subjects, are also spelled as separate words, as in

33
 1 Sounds and Spelling of Palauan

(50) a. Ng hong ȩr ngii.

‘It’s his book.’

b. Ak ulȩmes ȩr kau.
‘I saw you.’

c. Ngak a sensei.
‘I’m a teacher.’

d. Ngii a lilȩchȩsii a babier.

‘He wrote the letter.’

3. The object pronouns -ak ‘I’, -ii ‘him/her/it’, -au ‘you’,

etc., are written as part of the action verb which accompanies
them. The hyphen in our notations -ak, -ii, -au, etc. indicates
that these pronouns are not independent words but must be at-
tached to the end of other words.
Some examples containing these pronouns (italicized) are
given below:

(51) chillȩbȩdak ‘hit me’

chillȩbȩdii ‘hit him/her/it’
chillȩbȩdau ‘hit you’
sosȩbii ‘burns it’
milstȩrir ‘saw them’

4. The possessor pronouns are added to nouns to indicate

the owner or possessor of something. These possessor pronouns
have many forms, even for the same person, of which only a few
are illustrated below. As you can see, these pronouns (italicized)
are always attached to the word indicating the thing possessed.

(52) bilek ‘my clothing’

blik ‘my house’
ngȩrem ‘your mouth’
mlirir ‘their car(s)’
sȩbȩchel ‘his ability’
soam ‘your desire’

5. The hypothetical pronouns ku- ‘I’, lo- ‘he/she/it’, chomo-

‘you’, etc., are attached to the beginning of verbs in a large va-
riety of complicated constructions. A few sample sentences con-
taining these hypothetical pronouns (italicized) are given below:

34
 Palauan Reference Grammar

(53) a. Ng diak kusuub.

‘I’m not studying.’

b. A John a diak loilil.

‘John isn’t playing’.

c. A hong a longuiu ȩr ngii a John.

‘The book is being read by John.’

d. Kȩ mȩkȩra a chomoruul a mlai?

‘What do you do to make a canoe?’

e. Ng chȩtik a Droteo a lolamȩch.

‘I don’t like Droteo to chew (betel nut).’

f. Ng soal a kbo kusuub.

‘He wants me to study.’

e. Palauan has many word sequences which function to ex-

press a single meaning or idea and which often have single
English words as their equivalents. Even though we might want
to spell these sequences as single words, detailed grammatical
analysis leads us to conclude that they actually involve more
than one word and should therefore be spelled as in the ex-
amples below:

(54) e le ak [εlεkh] ‘because I…’

e le ng [εlεŋ] ‘because he…’
mȩ a [ma] ‘and’
mȩ ak [makh] ‘so I…’
mȩ ng [mǝŋ] ‘so he…’
e ng di [εndi, ǝndi] ‘but’
er se ȩr a [εrsεra, ǝrsεra] ‘when’
el kmo [εlkmo, ǝlkmo] ‘[say] that…’
el ua se [εlwasε, ǝlwasε] ‘[say] that…’
ng diak [ṇdiakh] ‘isn’t’
ng di kea [ṇdigεa] ‘no longer is’
di mle ngii [dimlεŋiy] ‘by himself’
ko ȩr a [kora] ‘kind of, like’
a lȩko (ak) [alǝgo(k)] ‘(I) intended to…’
a lsȩkum [alsǝgum] ‘if, when’

35
 1 Sounds and Spelling of Palauan

f. Numerous exceptions to all of the Palauan spelling rules

explained above can be found in two classes of words, personal
names and words of foreign origin. First of all, many Palauan
personal names—e.g. Francisco, Polycarp, Hermana, etc.—are
borrowed from other languages and therefore contain certain
sounds which are not native to Palauan. In order to spell these
sounds, it is necessary to use non-Palauan letters such as p, f,
j, y, and w. Furthermore, the spelling of many native Palauan
names, including personal names like Yaoch, Polloi, etc. as well
as place names like Peleliu, Kay angel, etc., does not conform
exactly to the rules presented above. We should not be sur-
prised or disturbed that Palauan personal names in particular
show so much deviation from the “standard” rules of Palauan
spelling. This is perhaps as it should be, since names are very
individual things.
Second, words of foreign origin—especially technical
terms—often appear in Palauan speech. Since these words
contain sounds which are not native to Palauan, it becomes nec-
essary in some cases to use non-Palauan letters. If the word
is borrowed from Japanese and contains no sounds strange
to Palauan, it should be possible to spell this word only with
Palauan letters, as in basio ‘place’ and iasai ‘vegetables’. If,
however, the Japanese word contains sounds which do not occur
in Palauan, then it is necessary to use non-Palauan letters such
as z and h. 11 Words of this type include daiziob ‘all right’, skozio
‘airport’, benzio ‘toilet’, hutsu ‘common, usual’, kohi ‘coffee’,
and keizai ‘economies’. If the borrowed word comes from
English—e.g. government, post office, party, etc.—it is usually
spelled as in English, unless a native Palauan spelling has
become commonplace, as in the case of skuul ‘school’.

1.6. SUGGESTIONS FOR FURTHER STUDY

Although we have examined the most important features of the
Palauan sound system, there are many details which we have
had to omit. Furthermore, our descriptions of the articulation
of Palauan sounds and our phonetic transcriptions of Palauan
words have been rather rough and oversimplified. For these
reasons, this chapter should not be thought of as a guide to
learning the correct pronunciation of Palauan. Such a goal can
only be achieved with the assistance of a native speaker. If you
are interested in looking at the sound system of Palauan in

36
 Palauan Reference Grammar

greater detail, you can consult such technical works as Carlson

1968, Flora 1969, and Wilson 1972, which are mentioned in the
bibliography.

37
 2 Palauan Nouns

2PalauanNouns
2.1. GRAMMATICAL SYSTEM OF PALAUAN
In the preceding chapter we described the sound system of
Palauan in terms of its significant sounds (phonemes) and their
variants (allophones). In our analysis, we grouped these
phonemes into various classes—e.g., stop consonants, nasal
consonants, high vowels, mid vowels, etc.—whose members
share common features of pronunciation and show similarities
in distribution, which refers to the way in which sounds
combine with each other in the formation of words. For ex-
ample, the class of nasal consonants m and ng is distinguished
from the other types of Palauan consonants by having a nasal
articulation, in which the outward air flow escapes through the
nose, but not the mouth; furthermore, the nasals m and ng
share the distributional feature of appearing as syllabic conso-
nants ([ṃ] and [ŋ] or [ṇ]—cf. 1.3.5) before other consonants in
word-initial position. To take another example, the liquid con-
sonants l and r are distinguished phonetically from the other
groups of consonants because they alone are produced with a
partial closure in the mouth; moreover, l and r have the unique
distributional feature of appearing as “double” consonants ll
and rr, which no other consonants can do.
We have seen, then, that in order to describe the sound
system of Palauan, we grouped the sounds into classes and
specified the distributional characteristics of these classes. In
describing the grammatical system of Palauan, our approach
will be very similar, though in this case the “building blocks” of
our analysis will be words rather than sounds. Thus, in order
to get a clear picture of the grammatical system of Palauan,
we will classify the words of the language into various groups
called parts of speech and describe the distributional charac-
teristics common to the members of each group. Just as there
are restrictions in all languages on the way sounds can combine
with each other—for example, a word beginning with sr would
be unthinkable in Palauan, as it would be in English—so are
there restrictions on the way words can combine with each
other. For example, the order of words in a simple expression

38
 Palauan Reference Grammar

like chad ȩr a omȩnged ‘fisherman’ is rigidly fixed, and any

other combination of these words—e.g. *omȩnged ȩ r a chad,
*chad omȩnged ȩ r a, 1 etc.—is impossible and meaningless.

2.2. IDENTIFICATION OF PALAUAN NOUNS

Nouns constitute one of the most important parts of speech
in Palauan. As a very rough preliminary definition, let us say
that nouns name or make reference to many different types
of things or living beings. Nouns are like identification labels
which point to the various persons and things which we deal
with in daily life and talk about in everyday conversation. We
can divide nouns into several subgroups, depending on what the
noun refers to.
The easiest nouns to identify are those whose referents can
be perceived by one or more of our five senses. Nouns of this
type, which identify things we can see, hear, touch, taste, or
smell, are called concrete nouns. It is convenient to divide
the concrete nouns of Palauan into two categories—human and
nonhuman. Why this division is basic will be explained below.
As you might expect, human nouns make reference to human
beings; some examples of human nouns are given in the list
below:

Human nouns

(1) chad ‘person’ sensei ‘teacher’

sȩchȩlei ‘friend’ ngalȩk ‘child’
buch ‘spouse’ buik ‘boy’
chȩdil ‘mother’ Droteo ‘Droteo’
dȩmal ‘his father’ Toki ‘Toki’

By contrast, non-human nouns refer to anything which is

not human, such as animals and living or non-living things.
In the list below, the words in the left column are non-human
nouns designating animals, while those in the right columns are
non-human nouns referring to living or non-living things.

Non-human nouns

Animals Living or non-living things

(2) bilis ‘dog’ kȩrrȩkar ‘tree’ blai ‘house’

39
 2 Palauan Nouns

ngikȩl ‘fish’ bung ‘flower’ babier ‘paper, letter’

malk ‘chicken’ daob ‘ocean’ mlik ‘my car’
babii ‘pig’ omoachȩl ‘river’ kall ‘food’

Given the above distinctions, how would you classify the follow-
ing words: mlai ‘canoe’, btuch ‘star’, tbak ‘my spit’, kȩrȩbou
‘cow’, tuu ‘banana’, rȩdil ‘woman’, iis ‘nose’, sers ‘garden’,
tolȩchoi ‘baby’, ngduul ‘clam’, and subȩlek ‘my homework’?
As opposed to concrete nouns, abstract nouns have ref-
erents which cannot be perceived by any of the five senses. This
is because abstract nouns refer to many different kinds of con-
cepts, ideas, or emotions which can “exist” only in our minds
but not in the everyday world where we can see them, touch
them, etc. The abstract nouns listed below, then, refer to things
over which we can have only conceptual (i.e., mental), but not
perceptual control:

Abstract nouns

(3) reng ‘heart, spirit’ blȩkeu ‘bravery’

dakt ‘fear’ klȩchad ‘human life’
soal ‘his desire’ klȩmȩra ‘truth’
kirek ‘my obligation’ klȩngit ‘sin’
sȩbȩchem ‘your ability’ klausȩchȩlei ‘friendship’

The distinction between concrete vs. abstract nouns is not

always as clear-cut as we have implied above. It is often very
difficult to decide how to classify a particular noun in terms of
this distinction. How would you deal, for example, with the fol-
lowing nouns: tȩkoi ‘word, language’, char ‘price’, chais ‘news’,
ker ‘question’, and ngakl ‘name’?

2.3. DISTRIBUTION OF PALAUAN NOUNS

Now that we have identified the major classes of Palauan nouns
and seen the range of meanings which they can cover, let us see
how we can identify nouns in terms of their distribution—that is,
in terms of how they combine with other words in the formation
of sentences. We shall first discuss nouns in their function as
sentence subject and sentence object.
Observe the following sentences:

40
 Palauan Reference Grammar

(4) a. A ngalȩk a mȩnga a ngikȩl.

‘The child is eating fish.’

b. A Droteo a chillȩbȩdii a bilis.

‘Droteo hit the dog.’

What both of these examples have in common is that they de-

scribe the occurrence of an action: in 4a, the action of eating,
represented by the word mȩnga ‘eat’, is taking place at the
present time (i.e., at the time someone is saying the sentence),
and in 4b, the action of hitting, designated by the word
chillȩbȩdii ‘hit’, occurred at some time in the past. In both 4a
and 4b, the action words mȩnga and chllȩbȩdii serve to relate
two nouns—the one doing the action and the one affected by
the action. In 4b, for instance, the noun Droteo tells us who per-
formed the action of hitting, while the noun bilis ‘dog’ identifies
what received the effect of this action. See if you are able to in-
terpret 4a in a parallel way.
Nouns like ngalȩk ‘child’ and Droteo of 4a–b, which refer
to the person who performs, carries out, or causes the action
of the sentence, function as sentence subjects and are called
subject nouns. On the other hand, nouns like ngikȩl ‘fish’ and
bilis ‘dog’ of 4a–b, which tell us what is affected by the action
of the sentence, function as sentence objects and are termed
object nouns. Our definitions of these terms now need to be
expanded.
In denning subject noun, we implied falsely that the subject
of an action sentence must always be human. Although in fact
most subjects in action sentences usually are human, occa-
sionally we can find a non-human subject. In 5a below, for ex-
ample, the subject noun is an animal, and in 5b, it is something
non-living (an act of nature):

(5) a. A malk a killii a bȩras.

‘The chicken ate up the rice.’

b. A dȩrumk a ulȩkȩrngii a ngalȩk.

‘The thunder woke up the child.’

What words designate the actions in 5a–b above, and what

nouns identify the objects?

41
 2 Palauan Nouns

Furthermore, in defining object noun, we did not make it

clear that any type of noun can function as sentence object. In
5b above, for instance, the object noun ngalȩk ‘child’ refers to
a human being, and in the following example, the object noun
designates an abstract concept:

(6) A John a rirȩllii a klȩngit.

‘John committed a sin.’

The examples in 4–6 above allow us to describe some of the

distributional features of Palauan nouns. To summarize what we
have discovered so far, we can say that in action sentences,
nouns can appear either before the action word (in which case
we speak of subject nouns), or after the action word (in which
case we speak of object nouns). In other words, Palauan action
sentences show the basic pattern subject noun + action word
+ object noun; in such sentences, the position of the noun (pre-
ceding or following the action word) tells us whether we in-
terpret it as sentence subject or sentence object. You may have
noticed that so far we have chosen to omit discussion of the
word a, which appears before every noun and action word in
4–6 above. An explanation of this word will be provided in 2.6
below.
Unlike the examples of 4–6 above, there are many Palauan
action sentences which have only a subject noun, but no object
noun. Observe the following examples:

(7) a. A Droteo a mililil.

‘Droteo was playing.’

b. A ngȩlȩkek a rȩmurt.
‘My child is running.’

Clearly, the words mililil ‘was playing’ and rȩmurt ‘is running’
refer to actions. But these actions are of quite a different nature
from the actions of eating, hitting, etc. seen in the examples
of 4–6. While eating, hitting etc., are types of actions which
naturally have an effect on something else (i.e., we eat some-
thing, we hit someone, etc.), playing and running are not ac-
tions which we direct at someone or something else, but actions
in which the doer involves only himself. For this reason, the sen-
tences of 7 contain no object nouns.

42
 Palauan Reference Grammar

One more type of Palauan sentence has only a subject noun,

but no object noun. Rather than designating an action (as in
4–7 above), this type of sentence describes the subject noun
in some way. Most commonly, this description involves a state
or condi tion which the subject noun is in, as the following ex-
amples illustrate:

(8) a. A bȩchik a smechȩr.

‘My wife is sick.’

b. A ralm a mȩkȩlȩkolt.
‘The water is cold.’

c. A John a mȩtongakl.
‘John is tall.’

d. A mubi a ungil.
‘The movie is good.’

In 8a–b, the states involved are temporary (that is, they will
eventually change), while in 8c–d the states are relatively per-
manent (that is, unchanging).
Another sentence type involving description of the subject
noun is one which identifies the subject noun in terms of some
profession, nationality, or other feature. Observe the sentences
below:

(9) a. A Droteo a sensei.

‘Droteo is a teacher.’

b. A sȩchȩlik a chad ȩr a Siabal.

‘My friend is Japanese.

c. A Francisco a rubak.
‘Francisco is an old man.’

Yet another sentence type characterized by description of

the subject noun specifies the location of the subject noun, as in
the examples below:

(10) a. A rȩngalȩk a ngar ȩr a sers.

‘The children are in the garden.’

43
 2 Palauan Nouns

b. A oluchȩs a ngar ȩr a chȩlsel a skidas.

‘The pencil is inside the drawer.’

In the examples of 10, the word ngar ‘be (located)’ introduces

the word sequences ȩr a sers ‘in the garden’ and ȩr a chȩlsel
a skidas ‘inside the drawer’, which tell us where the subject
nouns are located. Word sequences of this kind, which are
called locational phrases, are described in detail in 14.2.
In studying the distribution of Palauan nouns, we have so
far focused our attention on the occurrence of nouns as sen-
tence subject and sentence object. Let us briefly examine one
more “environment” or position in which nouns are observed to
occur. In example 10a above, we looked at the word sequence
ȩr a sers ‘in the garden’, which designates a place or a location.
This sequence of words consists of ȩ r, corresponding to English
‘in’, and the noun sers ‘garden’ (which is preceded by the word
a). Because the word ȩr in 10a relates the subject noun rȩn-
galȩk ‘children’ to the noun sers ‘garden’ by telling where the
children are located, we call ȩr a relational word (see chap.
14). The relational word ȩr, which can designate many types of
relationships such as ‘in, at, on, to, from, out of, because of’,
etc., is always followed by a noun. Some of these uses of ȩ r are
illustrated in the sentences below:

(11) a. A John a mo ȩr a stoang.

‘John is going to the store.’

b. A beab a tilobȩd ȩr a blsibs.

‘The mouse came out of the hole.’

c. A Toki a smechȩr ȩr a tȩretȩr.

‘Toki is sick with a cold.’

We can see, then, that another distributional feature of Palauan

nouns is that they appear following the relational word ȩr.

2.4. PRONOUNS
In the sentences of 3–11 above, it is possible to replace the
subject nouns with shorter words which refer to the same
person or thing. For example, with 4a, repeated here as 12,
compare sentence 13:

44
 Palauan Reference Grammar

(12) A ngalȩk a mȩnga a ngikȩl.

‘The child is eating fish.’

(13) Ng mȩnga a ngikȩl.

‘He/she is eating fish.’

In 13, the word ng has substituted for ngalȩk ‘child’ of 12 as the

sentence subject; such substitute words are called pronouns. A
sentence with a pronoun subject like 13 can only be spoken if it
is clear to whom the pronoun refers. In other words, 13 would
make little sense as the very first sentence in a conversation,
but it is perfectly acceptable in the following dialog:

(14) A: A ngalȩk ng mȩnga a ngarang?

‘What is the child eating?’

B: Ng mȩnga a ngikȩl.
‘He/she is eating fish.’

In the above dialog, 13 can appear as B’s response to A’s

question because it is clear that the pronoun ng refers to ngalȩk
‘child’, which was introduced into the conversation by A.
Now compare 10a, repeated here as 15, with sentence 16:

(15) A rȩngalȩk a ngar ȩr a sers.

‘The children are in the garden.’

(16) Tȩ ngar ȩr a sers.

‘They are in the garden.’

As you can see, the pronoun which substitutes for rȩngalȩk

‘children’ is tȩ ‘they’, but not ng ‘he, she’. Which pronoun
is chosen depends on whether the replaced word is singular
(ngalȩk ‘child’) or plural (rȩngalȩk ‘children’). Whereas
Palauan singular nouns refer to one single person, plural
nouns refer to two or more persons. In 2.5 below, we will
discuss the formation of Palauan plural words in greater detail.
The pronouns ng ‘he, she’ and tȩ ‘they’ introduced in 13
and 16 above refer to someone other than the speaker of the
sentence or the person spoken to. That is, they refer to some
third party whom the speaker is interested in talking about. Be-
cause pronouns like ng ‘he, she’ and tȩ ‘they’ make reference

45
 2 Palauan Nouns

to some third party, it is not surprising that they are called

third person pronouns. Now, as you might imagine, it is also
possible to have pronoun subjects which refer to the speaker
(or first person) of a sentence and the person spoken to (or
second person). These two possibilities are seen in the fol-
lowing sentences:

(17) Ak mo ȩr a stoang.
‘I’m going to the store.’

(18) Kȩ mȩruul a ngarang?

‘What are you making?’

Here we observe the first person singular pronoun ak ‘I’ and

the second person singular pronoun kȩ ‘you’.
When we look at how the third person subject pronouns
of Palauan are used, we can see why the distinction between
human and non-human is so important. The pronouns ng and
tȩ are the only pronouns available in Palauan to make reference
to some third party. While the majority of speakers can use tȩ
only to refer to two or more human beings (as in 16 above),
ng has a much wider range of use, since it can refer not only to
anything singular (whether human beings, animals, or living or
non-living things) but also to plural things, as long as they are
not human. For this reason, ng can be translated as ‘he’, ‘she’,
or ‘it’ when singular and as ‘they’ when non-human plural, as in
13 above and examples like the following:

(19) A: A bilis ng ngar ȩr ker?

‘Where {is the dog / are the dogs}?’

B: Ng mȩchiuaiu ȩr a eungel a tebȩl. 2

‘{It is / They are} sleeping under the table.’

(20) A: A bȩlochȩl ng silebȩk ȩr ker?

‘Where did the {pigeon / pigeons} fly out from?’

B: Ng silebȩk ȩr a kȩrrȩkar.
‘{It / They} flew out of the tree.’

(21) A: A oluchȩs ng ngar ȩr ker?

‘Where {is the pencil / are the pencils}?’

46
 Palauan Reference Grammar

B: Ng ngar ȩr a chȩlsel a skidas.

{It is / They are} inside the drawer.’

When reference is being made to something plural, we see that

tȩ is confined to human beings, while ng can substitute for an-
imals (as in 19 and 20) or things (as in 21). In other words,
we can describe how the pronouns tȩ and ng refer to plurals
in terms of the basic distinction human noun vs. non-human
noun: tȩ substitutes for the former and ng for the latter. 3
Before leaving our discussion of pronouns (which we will
take up in greater detail in chap. 4), we need to make the point
that in terms of distribution, pronouns are really a subtype of
nouns, since they can occur in all of the environments in which
nouns are observed to occur. So far we have seen that pro-
nouns, like nouns, can function as sentence subjects. Now, let us
confirm that pronouns, like nouns, can also function as sentence
objects and can appear after the relational word ȩr. Observe the
following pairs of sentences:

(22) a. A John a mȩluchȩs ȩr a babier.

‘John is writing the letter.’

b. A John a mȩluchȩs ȩr ngii.

‘John is writing it.’

(23) a. A beab a tilobȩd ȩr a blsibs.

‘The mouse came out of the hole.’

b. A beab a tilobȩd ȩr ngii.

‘The mouse came out of it.’

In 22a, the sentence object is the noun babier ‘letter’ (for the
function of ȩr before a babier, see 2.7 below), which is replaced
in 22b by the pronoun ngii ‘it’. And in 23a, the noun blsibs ‘hole’
appears after the relational word ȩr ‘out of’; it likewise is re-
placed by ngii ‘it’ in 23b.

2.5. PLURALS
As we have seen before, in order to form a plural noun in
Palauan, we simply attach rȩ- to the beginning of the noun.
(If the noun begins with a vowel, then rȩ- will be shortened

47
 2 Palauan Nouns

to r-.) Because rȩ- precedes the noun to which it is attached,

it is called a prefix; the hyphen in our notation rȩ- indicates
that something must follow the prefix to form a whole word.
We have also noted that rȩ- can only be added to human
nouns: in other words, it is only in the category of human nouns
that Palauan speakers make a distinction between singular (re-
ferring to just one person) and plural (referring to two or more
persons). This fact is another justification for making the basic
distinction human vs. non-human in Palauan, since this dis-
tinction explains the distribution of the prefix rȩ-. Thus, we can
have singular-plural pairs like sȩchȩlik ‘my friend’—rȩsȩchȩlik
‘my friends’, chad ‘person’—rȩchad ‘people’, and kangkodang
‘tourist’—rȩkangkodang ‘tourists’, but never pairs like bilis
‘dog’—*rȩbilis ‘dogs’ or babier ‘letter’—*rȩbabier ‘letters’. Here
are some sentences containing plural human nouns:

(24) a. Ak milstȩrir a rȩsȩchȩlim ȩr a party.

‘I saw your friends at the party.’

b. A rȩchad ȩr a Merikel a mȩkekȩmangȩt.

‘Americans are tall.’

c. Ng delmȩrab ȩr a rȩsensei.
‘It’s the teachers’ room.’

d. Tȩ di rȩngalȩk ȩr a skuul.
‘They’re just students.’

The plural prefix rȩ- can also attach to number words (see
24.4) when they refer to human beings. In such cases, however,
the presence of rȩ- is optional—that is, rȩ- may or may not be
attached, with no apparent difference in meaning. To indicate
the optionality of rȩ- before number words, we enclose rȩ- in pa-
rentheses, as in the examples below:

(25) a. A (rȩ)tȩlolȩm ȩl chad a mle ȩr a party.

‘Six people came to the party.’

b. A (rȩ)tȩruich ȩl ngalȩk a mlad.

‘Ten children died.’

c. A (rȩ)dart ȩl chad a mle sengkio.

‘One hundred persons voted.’

48
 Palauan Reference Grammar

d. Ak milstȩrir a (rȩ)tȩde ȩl sensei.

‘I saw three teachers.’

e. Tȩ milkodȩtȩrir a (rȩ)lluich mȩ a teua ȩl chad.

‘They killed twenty-tour people.’

We have the option of omitting the plural prefix rȩ- from the
examples of 25 because rȩ- does not seem to add anything to
the meaning: in each case, the presence of the number word
(tȩlolȩm ‘six (people)’, tȩruich ‘ten’, etc.) makes it clear that two
or more human beings are being talked about.
The plural prefix rȩ- can also attach to certain other
words—e.g., ua ‘like, as’ and bek ‘each, every’—when these
words are associated with a human noun. Observe the following
examples:

(26)a. Tirke ȩl teru ȩl chad tȩ rua tȩchang?

‘Who are those two people (i.e., what are their names)?’

b. Ak milstȩrir a rua Toki ȩr a stoang.

‘I saw Toki and her friends at the store.’

c. A rȩbek ȩl ngalȩk a kirir ȩl mo ȩr a skuul.

‘Every child must attend school.’

d. Ak milstȩrir a rȩbek ȩl chad a present.

‘I gave each person a present.’

In 26a, the word ua appears before tȩchang ‘who?’, which is a

question word referring to human beings, while in 26b, ua pre-
cedes the name of a person (Toki). In 26c-d, bek ‘each, every’ is
linked to the following human noun by the word ȩl.
The plural prefix rȩ- also can attach to certain words which
name states or conditions (or, less frequently, actions) to form
(or derive) a noun referring to the group of people charac-
terized by the particular state or condition. For example, from
meteet ‘rich’, we can form rȩmeteet ‘those whose are rich, rich
people’, or from mȩsaik ‘lazy’, we can form rȩmȩsaik ‘those who

49
 2 Palauan Nouns

are lazy, lazy people.’ Derived plural nouns of this type are il-
lustrated in the sentences below. Note that their distribution is
the same as that of any other noun.

(27)a. A irȩchar, e a rȩmeteet a ulȩngȩseu ȩr a rȩmechȩbuul.

‘In earlier times, the rich helped the poor.’

b. A rȩdȩngȩrengȩr a mo ȩr a bȩluu ȩr a ngau, e a

rȩmȩkȩdu a mo ȩr a babȩluadȩs.
‘Those who are poorly-behaved will go to hell, while
those who are well-behaved will go to heaven.’

c. A rȩmȩsaik a blȩchoel ȩl diak a kȩlir.

‘Those who are lazy (and don’t work) often don’t have
food.’

d. A rȩmȩkekȩdeb a mo ȩr a uchei, e a rȩmȩkekȩmangȩt a

mo ȩr a uriul.
‘(Let’s have) the short ones go to the front and the tall
ones go to the back.’

e. A rȩmȩruul a kall a chȩdal a rȩdil, e a rȩmȩngoit a

udoud a chȩdal a sȩchal.
‘Those who prepare the food are the relatives of the
woman, and those who contribute the money are the
relatives of the man.’ 4

As we will see in chaps. 5 and 7, the words to which rȩ- attaches

in 27a-d are called state verbs, while those to which rȩ- at-
taches in 27e are called action verbs.

2.6. THE PALAUAN WORD A

In the sentences of 4–27 above, we have seen many instances
of the Palauan word a. Although we can describe the distrib-
ution of this word fairly accurately, we will have trouble saying
exactly what it means. It seems that the major function of a
is simply to “introduce” certain Palauan parts of speech when
they occur in a sentence: for example, sentences 4–27 show that
every Palauan noun (unless it is a pronoun) must be directly
preceded by a, regardless of whether the noun functions as sen-
tence subject or object, or follows the relational word ȩr. Fur-

50
 Palauan Reference Grammar

thermore, a always introduces the verb of the sentence, which

directly follows the subject noun and names an action (as in 4–7)
or a state or condition (as in 8). Though we will define the term
‘verb’ with much greater care in chap. 5, we will nevertheless
begin to use it now, since we will need to make reference to
verbs in many of our discussions. A few of the action verbs and
state verbs which have appeared in the sentences above are
listed here:

(28) Action verbs State verbs

mȩnga ‘eat’ smechȩr ‘sick’

killii ‘ate it up’ mȩkȩlȩkolt ‘cold’
mȩruul ‘make, do’ ungil ‘good’
mililil ‘was playing’ ngar ‘be (located)’
rȩmurt ‘is running’
tilobȩd ‘came out of’

As sentences like 13, 16, 17, 22b, and 23b show, Palauan
pronouns are never introduced by the word a, whether they
function as sentence subject (as in 29a–b below) or object (as in
29c– d), or follow the relational word ȩr (as in 29e):

(29) a. Ak mȩluchȩs a babier.

‘I’m writing a letter.’

b. Tȩ mo ȩr a skuul.
‘They’re going to school.’

c. A John a mȩruul ȩr ngii.

‘John is making it.’

d. A Toki a ulȩmes ȩr ngak.

‘Toki saw me.’

e. A beab a tilobȩd ȩr ngii.

‘The mouse came out of it.’

Examples 29a–b require us to qualify the general statement

made above that the verb of a sentence is always introduced by
a, since a does not precede the verb when the sentence subject
is a pronoun.

51
 2 Palauan Nouns

There is another group of Palauan words which are not

usually introduced by a. This group includes words like tia ‘this
thing, this place/here’, se ‘that thing, that place/there’, ngika
‘this person’, ngike ‘that person’, etc., which are called demon-
stratives (see 24.3) because they point out persons or things
or specify where someone or something is located. Some typical
examples including demonstratives are now given:

(30) a. Tia a oluchȩs.

‘This is a pencil.’

b. A Toki a milil ȩr sei.

‘Toki is playing there.’

c. Ngka 5 ȩl chad a sensei.

‘This person is a teacher.’

d. Ngke5 ȩl chad ng mȩkȩrang?

‘What’s that person doing?’

Though the above explanation of the distribution of a is

incomplete and very oversimplified, it should serve as a nec-
essary introduction to a word which will turn up in almost every
Palauan sentence we examine. Until we read some of the other
chapters in this book, we will not have enough knowledge of
Palauan grammar to understand why the above treatment of a
is inadequate. We will see later, for example, that a does not ac-
tually introduce single nouns or verbs, but rather certain groups
of associated words called noun phrases and verb phrases
(see 3.6 and 5.2). We will also see that there are a few further
Palauan words which, like pronouns and demonstratives, are
never introduced by a, and that there are other conditions
under which a does not appear when it would otherwise be ex-
pected.

2.7. SPECIFIC VS. NON-SPECIFIC OBJECT NOUNS

In this section, we will examine an important contrast which
is found only among nouns functioning as sentence objects. In
order to get a clear grasp of this contrast, we will at first deal
only with sentences whose object nouns are non-human. Ob-
serve the following pairs of examples:

52
 Palauan Reference Grammar

(31) a. A nȩglȩkek a mȩdakt a dȩrumk.

‘My child is afraid of thunder.’

b. A ngȩlȩkek a mȩdakt ȩr a dȩrumk.

‘My child is afraid of the thunder.’

(32) a. Ng soak ȩl mȩnga a ngikȩl.

‘I like to eat fish.’

b. Ng soak ȩl mȩnga ȩ r a ngikȩl.

‘I want/would like to eat the fish.’

(33) a. Ak ousbech a biskang ȩl mo ȩr a chei.

‘I use a spear to go fishing with.’

b. Ak ousbech ȩr a biskang ȩl mo ȩr a chei.

‘I’m using the spear to go fishing with.’

You will notice that the only difference in form between the a-
and b-sentences of each pair is that the latter contain the word
ȩr (italicized) before the object nouns dȩrumk ‘thunder’, ngikȩl
‘fish’, and biskang ‘spear’ (which are of course preceded by
a). As the English translations indicate, the meanings of the a-
and b-sentences of each pair are very different from each other.
Since the only difference in form (or formal difference) be-
tween the members of each pair is the presence of ȩr in the b-
sentences, we can conclude that ȩ r contributes to the meaning
difference in a very important way.
How can we characterize the meaning difference observed
in the sentence pairs above? In each of the a- sentences, the
speaker is making a general statement about something: in
other words, he is saying that in general, or on many different
occasions, such and such is the case. In the b-sentences, on the
other hand, the speaker is making a specific statement about
some particular single occasion.
In 31a, for example,

(31a) A ngȩlȩkek a mȩdakt a dȩrumk.

‘My child is afraid of thunder.’

53
 2 Palauan Nouns

the speaker asserts that his child has a fear of thunder which
is observed on many different occasions, whenever there is
thunder, while in 31b

(31b) A ngȩlȩkek a mȩdakt ȩr a dȩrumk.

‘My child is afraid of the thunder.’

the speaker is saying that his child is afraid of the particular

rumblings of thunder which can be heard at the present
moment. Thus, a dȩrumk of 31a refers to thunder in general,
whereas ȩr a dȩrumk of 31b refers to a specific instance of
thunder. Because of this distinction, a speaker can use 31b only
when it is actually thundering, while 31a could be uttered at any
time whatsoever. Example 31b might appear, for instance, in a
conversation like the following, where it is B’s response to A’s
question:

(34) A: Ngara mȩ a ngȩlȩkem a lmangȩl?

‘Why is your child crying?’

B: Ng mȩdakt ȩr a dȩrumk.
‘He’s afraid of the thunder.’

The function of ȩr in a sentence like 31b, therefore, is to in-

dicate that the object noun is specific; if ȩr is absent, as in 31a,
the sequence a plus following noun is interpreted in a general
(i.e., non-specific) sense. We shall call a word sequence like ȩr
a dȩrumk ‘the thunder’ of 31b a specific object, while labelling
a dȩrumk ‘thunder’ of 31a a non-specific object. We can now
explain the pairs of sentences in 32 and 33 in greater detail.
Since the object noun following mȩnga ‘eat’ of 32a

(32a) Ng soak ȩl mȩnga a ngikȩl.

‘I like to eat fish.’

is not preceded by ȩr, it is non-specific, and the whole sentence

is therefore interpreted as a general statement expressing the
speaker’s liking for fish. By contrast, since the object noun of
32b

(32b) Ng soak ȩl mȩnga ȩr a ngikȩl.

‘I want/would like to eat the fish.’

54
 Palauan Reference Grammar

is specific, this sentence would be uttered on a single occasion

when some particular fish is involved. For example, 32b might
occur in the following conversation, where it appears as B’s
answer to A’s question:

(35) A: Ngara a soam ȩl mȩnga ȩr ngii?

‘What would you like to eat?

Ng ngar ȩr ngii a babii mȩ a ngikȩl mȩ a chȩmang.

‘There’s pork, fish, and crab.’

B: Ng soak ȩl mȩnga ȩr a ngikȩl.

‘I’d like to eat the fish.’

In B’s response, ȩr a ngikȩl ‘the fish’ makes specific reference

to the fish which A has prepared on a particular occasion and
which A is offering to B as part of a choice of foods. In example
33a,

(33a) Ak ousbech a biskang ȩl mo ȩr a chei.

‘I use a spear to go fishing with.’

the object noun biskang ‘spear’ is not preceded by ȩr and is

therefore non-specific. As expected, the sequence a biskang
does not designate some specific spear but simply refers to the
idea of ‘spear’ in general. For this reason, 33a is a general
statement expressing a habit: the speaker says that he uses
a spear whenever he goes fishing. On the other hand, the se-
quence ȩr a biskang of 33b

(33b) Ak ousbech ȩr a biskang ȩl mo ȩr a chei.

‘I’m using the spear to go fishing with.

refers to a particular spear which the speaker is using on some

given occasion.
In 2.5 above, we saw that the prefix rȩ- serves to identify or
mark plurality with human nouns only. Since rȩ- is restricted to
occurrence with human nouns, how is the difference between
singular and plural indicated for non-human nouns? It is dif-
ficult to give a good answer to this question because the lin-
guistic facts are very complicated. First of all, it is impossible

55
 2 Palauan Nouns

to distinguish between singular and plural for non-human sen-

tence subjects; therefore, the following examples have two pos-
sible meanings, as indicated (cf. 19A, 20A, and 21A):

(36) a. A bilis a mȩchiuaiu ȩr sei.

‘The {dog is / dogs are} sleeping there.’

b. A oluchȩs a ngar ȩr a bebul a tebȩl. 6

‘The {pencil is / pencils are} on the table.’

With non-human sentence objects, the presence or absence

of the “specifying” word ȩr results in a partial distinction be-
tween singular and plural. Observe the examples below:

(37) a. Ak ousbech ȩr a bilas ȩr a klukuk.

‘I need the boat tomorrow.’

b. Ak ousbech a bilas ȩr a klukuk.

‘I need {a boat / the boats} tomorrow.’

(38) a. Ak ousbech ȩr a mlim ȩl mo ȩr a ochȩraol.

‘I need your car to go to the money-raising party.’

b. Ak ousbech a mlim ȩl mo ȩr a ochȩraol.

‘I need your cars to go to the money-raising party.’

When ȩr precedes the non-human object noun, as in 37a and

38a above, it invariably refers to a specific singular object.
If ȩr is omitted, however, as in 37b and 38b, the sequence a
plus following noun can sometimes be interpreted in more than
one way: for example, a bilas of 37b can refer to a non-spe-
cific singular object (‘a boat’) or to a specific plural object
(‘the boats’). As we can see, then, the presence or absence of ȩ
r does not result in an absolutely clear-cut distinction between
singular and plural, since some object nouns which are not pre-
ceded by ȩr (e.g. a bilas of 37b) can also be interpreted as sin-
gular.
As we have seen above, non-human object nouns may or may
not be preceded by the specifying word ȩr, with a significant
difference in meaning. With human object nouns, however, the
use of ȩ r seems to be required (or obligatory): that is, human
object nouns, whether singular or plural, must be interpreted as
specific. Note the following examples:

56
 Palauan Reference Grammar

(39) a. Ak ulȩmes ȩr a Toki ȩr a party.

‘I saw Toki at the party.’

b. Lak mongȩlebȩd ȩr a ngalȩk!

‘Don’t hit the child!’

c. Ak ulȩmes ȩr a rȩsȩchȩlim ȩr a party.

‘I saw your friends at the party.’

d. A Droteo a milȩngȩlebȩd ȩr a rȩngalȩk.

‘Droteo was hitting the children.’

All names of people, such as Toki of 39a, refer uniquely to

particular persons and are therefore automatically specific.
Everything we have said above about the occurrence of ȩr
before an object noun will be valid only when the verb of the
sentence is of a particular type—namely, imperfective. The
difference in meaning between imperfective and perfective
verb forms is not very easy to explain and will be postponed
until chap. 12. For our present purposes, it will be sufficient
to know how to distinguish imperfective and perfective verb
forms in terms of an obvious formal difference. Perfective verb
forms always have a special pronoun ending which identifies the
object, whereas imperfective verb forms do not. Compare the
imperfective and perfective forms for a few Palauan verbs:

(40) Imperfective Perfective

mȩngȩlebȩd ‘hit’ cholȩbȩdak ‘hits me’

mȩnga ‘eat’ kolii ‘eats it’
mȩruul ‘make, do’ rullii ‘makes it’
omes ‘see’ mȩsa ‘sees him/her/it’

In the list above, the italicized portions of the perfective verbs

refer to the object: -ak ‘me’, -ii (or, rarely -a) ‘him/her/it’, etc. In
this chapter, perfective verbs have occurred in sentences such
as 4b, 5a–b, 6, 24a, and 25d–e.
When third person object nouns follow perfective verb
forms, they can never be preceded by the specifying word ȩr. In
other words, while 41a and 42a below are correct, 41b and 42b
are not:

57
 2 Palauan Nouns

(41)a. A Toki a chillȩbȩdii a ngalȩk. ‘Toki hit the child.’

b. *A Toki a chillȩbȩdii ȩr a
ngalȩk.

(42)a. A dȩrumk a ulȩkȩrngii a ‘The thunder woke up the

bilis. dog.’
b. *A dȩrumk a ulȩkȩrngii ȩr a
bilis.

Why should it be that the specifying word ȩr never occurs

before an object noun if the preceding verb is perfective? An
answer can be given if we carefully analyze the structure of
perfective verb forms: as we will see in 4.9, all perfective verb
forms include pronoun endings which refer to specific objects.
The endings -ii and -ngii in 41 and 42 are objects of this kind.
Because perfective verb forms always imply a specific object,
there is no further need to indicate specificity with a word like
ȩr. In other words, use of ȩr after perfective verb forms would
be redundant (i.e., add nothing to the meaning) and therefore
unnecessary.
As you can see, the specifying word ȩr discussed in this
section is responsible for some very important meaning distinc-
tions among nouns which are used as sentence objects. The
function of ȩ r observed here—to identify sentence objects as
specific—should be carefully distinguished from that of the re-
lational word ȩr, which is best considered as a different word.
This latter word expresses various types of relationships (pri-
marily spatial and temporal) and corresponds to English ‘to,
at, in, out of, because, etc.’

58
 3 Noun Possession

3NounPossession
3.1. POSSESSOR SUFFIXES AND POSSESSED
NOUNS
In chap. 2, we learned how to identify nouns in terms of their
meaning and distribution. We did not pay much attention, how-
ever, to the internal structure of nouns—that is, to the way
certain parts or units are put together to form words which
function as nouns. Most of the nouns cited in chap. 2 consist
of only one meaning-bearing unit (or morpheme); in other
words, nouns like ngalȩk ‘child’, ngikȩl ‘fish’, bilis ‘dog’, sers
‘garden’, chad ‘person’, etc. cannot be broken down into two or
more parts which individually have meaning. On the other hand,
the plural words discussed in 2.5 do consist of more than one
meaning-bearing unit: for example, the single word rȩngalȩk is
formed from the prefix rȩ-, which means or designates ‘plural’,
and the noun stem ngalȩk ‘child’. We use the term stem to
refer to a noun when it appears in combination with something
else—in this case, the plural prefix rȩ-. Can you identify the
noun stems in the words rȩsensei ‘teachers’, rȩchad ‘people’,
and rȩkangkodang ‘tourists’?
Nouns also function as stems in words like sȩrsek ‘my
garden’ and chimal ‘his hand’. In these words the nouns sers
‘garden’ and chim ‘hand’ are followed by -ek and -al, which
identify the person who owns or possesses the noun in question.
The endings -ek and -al, and many others to be discussed below,
are called suffixes because they follow the noun stem. How do
suffixes differ from prefixes?
Before classifying the suffixes which attach to Palauan
nouns, we should have a clear idea of their function. Basically,
they serve to express a relationship of possession between
the noun stem (the thing possessed) and the person or thing
to which the suffix refers (the possessor). This relationship
of possession can be of several different types. It may involve
actual ownership of something —e.g., mlik ‘my car’, bilsȩngel
‘his boat’, blim ‘your house’. Or it may express a whole-part
relationship: that is, the noun stem is something which is an in-
herent part of the person or thing referred to by the suffix, as in

59
 3 Noun Possession

chimak ‘my hand’, ngȩrel ‘his mouth’, chȩlsel ‘its inside’, or tkul
‘its edge’. In addition, it may involve bonds of kinship, marriage,
or friendship, as in dȩmak ‘my father’, bȩchim ‘your spouse’, or
sȩchȩlil ‘his friend’, or other kinds of association, as in beluak
‘my country’, dȩchal ‘his ability’, etc. Besides their primary
function of expressing possession, the suffixes under discussion
here have the secondary function of expressing characteri-
zation; this latter function will be explained in detail in 3.7
below.
In the discussion to follow, we will call endings like -ek ‘my’,
-al ‘his’, etc. possessor suffixes, since they follow the noun
stem and designate the possessor. 1 Any noun which consists of
a noun stem followed by a possessor suffix will be called a pos-
sessed noun. Further terms will be introduced as needed.

3.2. PERSON AND NUMBER

The possessor suffixes of Palauan can be classified into several
sets. 2 Each set contains seven different suffixes which corre-
spond to the seven pronouns of Palauan (see chap. 4) and which
can be distinguished from each other in terms of various com-
binations of person and number. As we saw in 2.4, a first
person pronoun is one which refers to the speaker, a second
person pronoun is one which refers to the person addressed,
and a third person pronoun has to do with some third party
who is being talked about. Number involves a distinction be-
tween singular and plural: the former applies to one person or
thing, while the latter applies to two or more persons or things.
The chart below gives the seven possessor suffixes which make
up the most commonly-occurring set; the corresponding (em-
phatic subject) pronouns are also given for purposes of ref-
erence:

(1) Singular Plural

Possessor Corresponding Possessor Corresponding

suffix pronoun suffix pronoun
1st pers. -ek ‘my’ ngak ‘I’ -ed ‘our kid ‘we (incl.)’
(incl.)’
-am ‘our kȩmam ‘we
(excl.)’ (excl.)’

60
 Palauan Reference Grammar

2nd pers. -em kau ‘you’ -iu ‘your’ kȩmiu ‘you’

‘your’
3rd pers. -el ‘his, ngii ‘he, she, -ir ‘their’ tir ‘they’
her, its’ it’

In the first person plural, there is a distinction between

inclusive and exclusive possessor suffixes. The first person
plural inclusive suffix -ed not only makes reference to the
speaker but also includes reference to the person spoken to;
in addition, any number of other persons associated with the
speaker or hearer may be included. On the other hand, the first
person plural ex clusive suffix -am excludes reference to the
hearer and involves only the speaker and at least one other
person associated with him. Thus, the possessed noun ududed
means ‘our money—belonging to me (or us) and you’, while
ududam means ‘our money—belonging only to us, but not you’.
Although -el and -ir are listed as third person singular and
plural possessor suffixes, respectively, this classification does
not correspond exactly to their use. Because -ir can refer only
to human plural possessors, -el must be used if the plural pos-
sessor is non-human. Therefore, -el not only can mean ‘his’,
‘her’, or ‘its’ but also can be equivalent to ‘their’ if the pos-
sessors are animals or things (cf. the discussion of tȩ ‘they
(human)’ and ng ‘he, she, it, they (non-human)’ in 2.4). Note,
further, that Palauan differs from English in using one and
the same possessor ending -el regardless of whether the third
person singular possessor is male (‘his’) or female (‘her’), or
designates an animal or thing (‘its’). Another difference be-
tween Palauan and English can be seen among the second
person pronouns and possessor suffixes: while English has the
same forms (you, your) for both singular and plural, the Palauan
forms are distinct (kau, -em for singular and kȩmiu, -iu for
plural).

3.3. FOUR SETS OF POSSESSOR SUFFIXES

The possessor suffixes of Palauan can be grouped into four
major sets, depending on the vowel which appears in all the sin-
gular suffixes and in the first person plural inclusive suffix. The
most common set of suffixes has already been presented in 1
above; this is called the e set because all the singular suffixes
(-ek, -em, and -el) and the first person plural inclusive suffix (-

61
 3 Noun Possession

ed) have the vowel e. A typical noun which takes the e set of
possessor suffixes is charm ‘animal’, whose possessed forms are
given below:

(2)chȩrmek ‘my animal’ chȩrmed ‘our (incl.) animal’

chȩrmam ‘our (excl.) animal’
chȩrmem ‘your (sg.) animal’ chȩrmiu ‘your (pl.) animal’
chȩrmel ‘his animal’, etc. chȩrmir ‘their animal’

You will notice that charm ‘animal’ has the full vowel a in its
independent (or unpossessed) form, but shows the neutral
vowel ȩ (schwa) in all of its possessed forms. In other words,
the morpheme charm is pronounced differently (as chȩrm-)
when it functions as the stem of a possessed noun. This dif-
ference in pronunciation, you will recall, is always reflected
in the Palauan spelling. The alternation between charm and
chȩrmek, etc. can be explained ’n terms of the process of vowel
reduction (cf. 1.4.4), which is very common in Palauan: full
vowels usually reduce to the neutral vowel ȩ in unstressed
syllables. Because all of the Palauan possessor suffixes are
stressed, the full vowels of independent forms like charm wind
up in unstressed syllables in possessed forms like chȩrmék,
chȩrmém, etc., thereby reducing to ȩ. This process of vowel re-
duction will be illustrated in considerable detail in 3.4 below.
If you need to refresh your memory on such terms as neutral
vowel, vowel reduction, stressed vs. unstressed syllable,
etc., refer back to 1.4.2 and 1.4.4, or see the glossary.
In addition to the e set of possessor suffixes, we also have
the u set, the i set, and the a set. 3 Each of these sets, which
occur less frequently than the e set, is illustrated below:

(3) Possessor u set: reng i set: buch a set: char

suffix ‘heart, spirit’ ‘spouse’ ‘price’

1st pers. sg. rȩngúk bȩchík chȩrák

2nd pers. sg. rȩngúm bȩchím chȩrám
3rd pers. sg. rȩngúl bȩchíl chȩrál
1st pers. pl. incl. rȩngúd bȩchíd chȩrád
1st pers. pl. excl. rȩngmám bȩchȩmám chȩrȩmám
2nd pers. pl. rȩngmíu bȩchȩmíu chȩrȩmíu
3rd pers. pl. rȩngír bȩchȩrír chȩrrír

62
 Palauan Reference Grammar

Notice, again, that although the unpossessed forms of the nouns

in 3 have the full (stressed) vowels e, u, and a, all of these
reduce to ȩ in the possessed forms because they occur in un-
stressed syllables. In addition to the obvious difference in the
vowel of the suffix— u, i, or a vs. e—the three sets of possessor
suffixes given above are different from the e set in the following
way: for certain suffixes which are vowel-initial in the e set,
the other sets have consonant initial suffixes. These differ-
ences are shown in the chart below:

(4) Possessor suffix e set u set, i set, a set

1st pers. pl. excl. -am -(ȩ)mam

2nd pers. pl. -iu -(ȩ)miu
3rd pers. pl. -ir -(ȩ)rir

The parenthesized ȩ indicated with the suffixes for the u set, i

set, and a set is predictable and appears in order to break up
an otherwise unpronounceable cluster of consonants (cf. 1.4.5.)
The ȩ is required, for example, between ch and a following con-
sonant in forms like bȩchȩmam ‘our (excl.) spouse’ and bȩchȩrir
‘their spouse’. 4 However, it is not needed between ng and a
following consonant in rȩngmam ‘our (excl.) hearts’ or rȩngrir
‘their hearts’.

3.4. VOWEL REDUCTION

In 3.3 above, we saw that the process of vowel reduction ac-
counts for the fact that many Palauan nouns show an alternation
between a stressed full vowel in the independent form and
an unstressed neutral vowel ȩ in the possessed forms. Before
talking about further types of vowel reduction, we shall present
some additional examples in which each of the five full vowels
of Palauan reduces to ȩ in the possessed forms of the noun.
For each noun, only the third person singular possessed form
will be cited, since all the other possessed forms can be pre-
dicted from it: this is because the vowel of the third person sin-
gular possessor suffix (e, u, i, or a) appears in the other singular
forms and in the first person plural inclusive, and because the
remaining plural suffixes will contain the extra consonants m or
r (cf. 3.3 above) if the vowel of the possessor suffix is u, i, or a.
Observe the following:

63
 3 Noun Possession

(5) Independent Noun Possessed Form:

‘his/her/its—’

a reduces
to ȩ: kar ‘medicine’ kȩrul
bad ‘stone’ bȩdul
bar ‘blanket’ bȩrul
malk ‘chicken’ mȩlkel
ngalȩk ‘child’ ngȩlȩkel
kall ‘food’ kȩlel 5

e reduces
to ȩ: ker ‘question’ kȩril
sers ‘garden’ sȩrsel
reng ‘heart, spirit’ rȩngul
sechȩr ‘sickness’ sȩchȩrel

i reduces
to ȩ: bsibs ‘drill’ bsȩbsel
kadikm ‘right hand’ kadȩkmel
chȩdil ‘mother’ dȩlal 6

o reduces
to ȩ: ngor ‘mouth’ ngȩrel
kbokb ‘wall’ kbȩkbel
sȩngsongd ‘stick’ sȩngsȩngdel
bȩlochȩl ‘pigeon’ bȩlȩchȩlel
sȩlokȩl ‘laundry’ sȩlȩkȩlel

u reduces
to ȩ: bȩdul ‘head’ bdȩlul
chur ‘laughter’ chȩril
duch ‘ability’ dȩchal

As the examples of 5 show, it is impossible to predict from

the independent form of the noun whether the possessor suf-
fixes will belong to the e set, the u set, the i set, or the a set.
For example, though the nouns bȩdul ‘head’, chur ‘laughter’,
and duch ‘ability’ all have the full vowel u in their independent
forms, each one takes a different set of possessor suffixes.
Furthermore, we find a few cases in which different words
are pronounced the same (i.e., are homonyms) in their in-
dependent forms but have possessed forms with different suf-

64
 Palauan Reference Grammar

fixes: with chur ‘laughter’—chȩril, for instance, compare chur

‘tongue’—chural and chur ‘rib of coconut frond’—churul. 7 Be-
cause of this unpredictability, the learner of Palauan has no
alternative but to memorize the possessed forms of a noun, to-
gether with the corresponding independent form.
There are a few words in which a full vowel remains intact
and does not reduce to ȩ when a possessor suffix is added. Most
examples of this kind, which are to be considered as excep-
tional or irregular, involve the vowel i, and/or refer to body
parts, as in the list below:

(6) Independent Noun Possessed Form

chim ‘hand, arm’ chimal

chikl ‘neck’ chiklel
ding ‘ear’ dingal
biskang ‘spear’ biskȩlȩngel 8
chur ‘tongue’ chural

3.4.1. VOWEL DELETION

In the possessed forms of certain Palauan nouns, an “extreme”
form of vowel reduction takes place. This extreme form of vowel
reduction is actually vowel deletion, or complete loss of the
vowel. Full vowels following ng, l, or (less frequently) r in the
independent form of a noun are often deleted in the possessed
forms, where the stress is shifted to the possessor suffixes. In
the possessed forms of such words, any word-initial ng, l, or
r becomes syllabic—i.e., is pronounced as a whole syllable (cf.
1.3.5). Some typical examples are listed below:

(7) Independent Noun Possessed Form

ngakl ‘name’ ngklel [ŋklέl]

ngikȩl ‘fish’ ngkȩlel [ŋkǝlέl]
rasȩch ‘blood’ rsȩchel [ṛsǝɂέl]
lusȩch ‘luck’ lsȩchel [ḷsǝɂέl]
ralm ‘water’ lmel 9 [ḷmέl]
klȩngit ‘sin’ kngtil 10 [kŋtíl]
chorus ‘horn’ chorsul [ɂorsúl]
bilas ‘boat’ bilsȩngel 11 [bilsǝŋέl]

65
 3 Noun Possession

In some nouns, a vowel u of the independent form is deleted

in the possessed forms, even when it follows consonants other
than ng, l , or r, as in the following:

(8) Independent Noun Possessed Form

kud ‘louse’ kdul

bung ‘flower’ bngal
busȩch ‘feather, body hair’ bsȩchel
tub ‘spit’ tbal
dub ‘dynamite’ dbal

When the independent form of a noun consists of two identi-

cal consonants with a vowel in between, the vowel and the
second consonant are usually dropped in all the singular pos-
sessed forms and in the first person plural inclusive possessed
form. As a result, the possessed forms appear to be quite dif-
ferent from the independent form, as in the examples below:

(9) Independent Noun Possessed Form

tet ‘handbag’ til 12

tut ‘breast’ tul12
bub ‘fishtrap’ bngel

An exception to the pattern observed in 9 is bab ‘space above/on

top of’, whose possessed form is bebul, etc.

3.4.2. SHORTENING OF LONG VOWELS

So far we have seen how the single vowels i, u, e, o, and a
observed in the independent forms of nouns either reduce to
ȩ or delete completely in the various possessed forms. In this
and the following section, we will examine what happens to se-
quences of vowels under similar circumstances. Let us first
take up the possessed forms of nouns whose independent forms
have long vowels (cf. 1.4.6). As the examples below illustrate,
the long vowels ii, uu, ee, and oo shorten to the corresponding
single vowel in the possessed forms of nouns, where they come
to appear in unstressed syllables:

(10) Independent Noun Possessed Form

66
 Palauan Reference Grammar

oriik ‘broom’ orikel

biich ‘sieve’ bichel
buuch ‘betel nut’ buchel
tuu ‘banana’ tual
beluu ‘country’ belual
luuk ‘nest’ lukel
deel ‘nail’ delel
dȩkool ‘cigarette’ dȩkolel

There are a few exceptions which do not follow the pattern of

10: these include diil ‘abdomen’—dȩlel, in which a long vowel
reduces to ȩ , and luut ‘return’—ltel [ḷtέl], in which a long vowel
disappears completely.

3.4.3. SHORTENING OF VOWEL CLUSTERS

In 1.4.7 we illustrated the many types of vowel clusters which
can be found in Palauan words. If a particular noun has such
a cluster in its independent form, this cluster always reduces
or shortens in the possessed forms of the noun. Recall that, in
any given vowel cluster, either the first or the second vowel may
be stressed. In the list below, which shows the major pattern of
vowel cluster shortening, it is always the stressed vowel of the
original cluster which is deleted in the possessed forms:

(11) Independent Noun Possessed Form

oák ‘anchor’ okúl

oách ‘leg’ ochíl
eólt ‘wind’ eltél
iúngs ‘island’ ingsél
uíngȩl ‘tooth’ ungȩlél
diáll ‘ship’ dillél
suóbȩl ‘study, homework’ subȩlél
cháis ‘news’ chisél
udóud ‘money’ ududél
táem ‘time’ temél
uláol ‘floor’ ulotél
klungiáol ‘benefit, goodness’ klungiolél
ngáu ‘fire’ nguíl
táut ‘aim’ tutél
búil ‘moon, month’ bilél
tȩkói ‘word, speech’ tȩkingél

67
 3 Noun Possession

chúi ‘hair’ chiúl

The pattern of vowel cluster shortening seen in 11 above

does not apply to certain nouns. In the exceptional cases below,
for example, the pattern is exactly opposite from that observed
in 11, since it is the unstressed rather than the stressed vowel
of the original cluster which is deleted in the possessed forms:

(12) Independent Noun Possessed Form

líus ‘coconut’ lisél

bóes ‘gun’ bosél
kléu ‘young coconut’ klengél
téu ‘width’ tengél
buík ‘boy’ bikél

In a few instances, the vowel cluster of the independent

form of a noun shortens in the possessed forms to ȩ or to a com-
pletely new vowel, as in the cases below:

(13) Independent Noun Possessed Form

láok ‘fat’ lȩkél

dáob ‘ocean’ dȩbél
chokdȩmáol ‘maternal uncle’ chokdȩmȩlél
ráel ‘road’ rolél
dȩledáes ‘free time’ dȩledȩsél

A few nouns whose independent forms end in vowel clusters

take only the consonant of the possessor suffix in the singular
possessed forms and in the first person plural inclusive pos-
sessed form. In the examples below, the general vowel cluster
shortening rule illustrated in 11 above is also applicable: in the
possessed forms, the stressed vowel of the independent form
deletes.

(14) Independent Noun Possessed Form

mlái ‘canoe, car’ mlíl

blái ‘house’ blíl
báu ‘smell’ búl
sȩchȩléi ‘friend’ sȩchȩlíl

68
 Palauan Reference Grammar

*3.4.3.1. Technical Discussion of Vowel Cluster

Shortening
The patterns of vowel cluster shortening described in 3.4.3
above can be explained in a more unified manner if we allow
the glides y and w to appear in the underlying forms of Palauan
words. 13 Under this analysis, the independent nouns listed in
11 would have underlying forms like wák ‘anchor’, yólt ‘wind’,
cháys ‘news’, udówd ‘money’, etc., where y and w represent
those vowels which are unstressed in the original cluster and
which are retained in the possessed forms. Since y and w are
consonants rather than vowels, we would expect them to
behave phonologically in certain well-defined ways: first of all,
they could never carry stress, and second, they could not reduce
or delete. If the independent nouns of 11 have underlying forms
with y and w as proposed, then our overall analysis is simplified
considerably, since the examples of 11 (like those of 7 and 8)
would merely involve the deletion of a single vowel. For ex-
ample, from buyl ‘moon’, we would derive the possessed form
bilel in approximately the following manner. The underlying
possessed form *buyl-el would first become *byl-el by deletion
of the only vowel of the noun stem—namely, u. Then, the in-
terconsonantal y would vocalize to i, giving the correct surface
form bilel. By exactly the same process we would have *swobȩl-
el ‘his homework’ → *swbȩl-el (by vowel deletion) → subȩlel (by
vocalization of w).
If we consider the above analysis to be correct in principle,
we no longer need to view the examples of 12 as exceptional.
The alternations illustrated in 12 could easily be explained if
the independent nouns indeed had underlying forms containing
true vowel clusters—e.g. lius ‘coconut’, teu ‘width’, etc.—rather
than sequences of a vowel and a glide. The rule of reduction
would simply be as follows: if the independent form of a noun
has a true vowel cluster of the form V́V or VV,́ the unstressed
vowel of the cluster deletes in the possessed forms. In addition,
the words of 14 would have underlying forms with a word-final
glide—e.g. mlay ‘canoe, car’, baw ‘smell’, etc.; for nouns of this
type, only the consonants of the possessor suffixes would be re-
quired for the singular possessors and for the first person plural
inclusive possessor.

69
 3 Noun Possession

3.4.4. POSSESSED NOUNS WITH -ng-.

Certain Palauan nouns require an -ng- as a “buffer” between the
noun stem and the possessor suffix. 14 This -ng- has no meaning,
and its appearance is not predictable. As the following examples
show, this -ng- is always followed by the e set of possessor suf-
fixes. Often, a ȩ appears between the ng and the preceding con-
sonant.

(15) Independent Noun Possessed Form

a. bas ‘charcoal’ bȩsȩngel

b. kȩrrȩker ‘earnings of money’ kȩrrȩkȩrngel
c. billum ‘wrapped tapioca’ billȩmȩngel
d. bȩchos ‘thumb’ bȩchȩsȩngel
e. btuch ‘star, starfish’ btȩchȩngel
f. bilas ‘boat’ bilsȩngel
g. dȩleb ‘ghost’ dȩlbȩngel
h. btuu ‘ball’ btungel
i. katuu ‘cat’ katungel 15
j. uum ‘kitchen’ umȩngel
k. teu ‘width’ tengel

Looking at the words of 15, we should have no difficulty in

identifying the various patterns of reduction which affect single
vowels and vowel sequences. In 15a–e, full vowels reduce to
ȩ in unstressed syllables, while in 15f–g full vowels disappear
completely. And in 15h–j the long vowel uu shortens to a single
u, while in 15k the vowel cluster eu is shortened. All of these
patterns are manifestations or indications of a very general
tendency in the sound system of Palauan, which can be summa-
rized as follows: Palauan vowels, whether single or in se-
quence, undergo various types of reduction in unstressed
syllables.

3.4.5. ALTERNATION BETWEEN -ng AND -l-

A very small number of Palauan nouns exhibit an alternation be-
tween word-final -ng in the independent form and -l- in the pos-
sessed forms—e.g. olȩkang ‘pot’—olȩkolel. Some of these also
add -ng- before the possessor suffixes, as in diokang ‘tapioca’—
diokȩlȩngel and biskang ‘spear’—biskȩlȩngel.

70
 Palauan Reference Grammar

3.5. OBLIGATORILY POSSESSED NOUNS

There are quite a few Palauan nouns for which it is impossible
to compare the possessed forms with an independent (or un-
possessed) form simply because the word must always occur
with one of the possessor suffixes. Such nouns are called oblig-
atorily possessed nouns; they contrast with optionally pos-
sessed nouns, which have both independent and possessed
forms. All of the nouns examined in the sections above were
of the optionally possessed type. In the list below you will find
some of the more commonly used obligatorily possessed nouns
of Palauan. As you can see, these fall into several categories
according to meaning. Since these nouns have no independent
forms, they are listed with a third person singular possessor
suffix.

(16) Obligatorily Possessed Nouns

a. Parts of the body budel ‘his/her/its skin’

(human or animal)
ulul ‘his/her/its chest’
kekul ‘his/her nail, its claw’
ulkel ‘his/her/its back’

b. Kinship relations: obȩkul ‘his older brother’

chudȩlel ‘her older sister’
ochȩdal ‘his sister, her
brother’

c. Relationship of part to rsel ‘its end’

whole:
tkul ‘its edge’
bkul ‘its corner’
uchul ‘its trunk (of a tree),
its reason’
llel ‘its leaf (of a tree)’

d. Other relationships: ultil ‘its imprint’

bȩlkul ‘its meaning, his/her/
its function’
ultutȩlel ‘its meaning, his/her/
its function’
ullel ‘behind him/her’

71
 3 Noun Possession

Although the obligatorily possessed nouns of 16 are listed

according to several broad categories of meaning such as ‘parts
of the body’, ‘kinship relations’, etc., we must not misinterpret
this list as implying, for example, that all nouns referring to
parts of the body are necessarily of the obligatorily possessed
type. This is of course not true, since there are many nouns
referring to body parts which have both an independent form
and possessed forms—e.g., chim ‘hand, arm’—chimal, ding
‘ear’—dingal, etc.
While the nouns listed in 16a and 16b can take any of the
possessor suffixes (e.g., budel ‘his/her skin’, budek ‘my skin’,
etc. or ochȩdal ‘his sister’, ochȩdam ‘your sister’, etc.), most
of those listed in 16c and 16d normally occur with a third
person singular possessor suffix. Furthermore, the obligatorily
possessed nouns of 16c and 16d are usually followed by another
noun that specifies or identifies the thing to which the third
person singular possessor suffix refers. For example, the group
of words tkul a daob ‘edge of the ocean’ consists of the oblig-
atorily possessed noun tkul followed by the noun daob ‘ocean’,
which indicates the “possessor” of tkul; like all nouns, daob is
preceded by the word a (cf. 2.6). A sequence of words like tkul
a daob is called a noun phrase of possession. This term will
be explained in greater detail at the end of 3.6 below.
Some typical noun phrases of possession containing the
obligatorily possessed nouns of 16c and 16d are given below:

(17)rsel a kȩrrȩkar ‘top of the tree’

rsel a kȩrreel ‘(other) end of the string’
rsel a chȩldȩchȩduch ‘end of the meeting’
tkul a tebȩl ‘edge of the table’
tkul a chȩldukl ‘edge of the dock’
bkul a chim ‘elbow’ (literally, ‘corner of the
arm’)
bkul a oach ‘knee’ (literally, ‘corner of the leg’)
uchul a kȩrrȩkar ‘trunk of the tree’
uchul a ‘reason for the meeting’
chȩldȩchȩduch
llel a kȩrrȩkar ‘leaves of the tree’
ultil a oach ‘footprint’
bȩlkul a tȩkoi ‘meaning of the word’

72
 Palauan Reference Grammar

For some of the noun phrases of possession in 17, it is possible

to have a possessor noun which is itself possessed: for example,
in addition to bkul a chim ‘elbow’, we can also have bkul a
chimak ‘my elbow’, etc., or in addition to bkul a oach ‘knee’, it is
possible to have an expression like bkul a ochȩrir ‘their knees’,
etc. Similar to these examples are the following, which involve
an obligatorily possessed noun followed by ngor ‘mouth’ and
chim ‘hand’:

(18) bȩrdel 16 a ngor ‘lips’

bȩrdel a ngȩrek ‘my lips’
kekul a chim ‘fingernail’
kekul a chimak ‘my fingernail’

The obligatorily possessed noun uchul ‘its trunk (of a tree),

its reason’ has apparently been reinterpreted as an independent
noun which itself has possessed forms. The resulting
form—uchȩlel—has a related, though different, meaning and
corresponds to English ‘beginning, origin, ancestors’. The pos-
sessed forms of uchul occur in noun phrases of possession like
the following:

(19) uchȩlel a chȩldȩchȩduch

‘beginning of the meeting’ 17

uchȩlir a rȩchad ȩr a Belau

‘{origin / ancestors} of the Palauan people’

3.6. PALAUAN NOUN PHRASES

The distribution of Palauan possessed nouns is exactly the same
as that of unpossessed nouns. In other words, possessed nouns
and unpossessed nouns occur in identical environments and can
function as sentence subject or object, or appear after the re-
lational word ȩ r (cf. 2.3). This identity of distribution is illus-
trated in the following pairs of sentences:

(20) a. A blai a milsesȩb.

‘The house burned down.’

b. A blik a milsesȩb.
‘My house burned down.’

73
 3 Noun Possession

(21) a. A John a chilitii a oluchȩs.

‘John threw away the pencil.’

b. A John a chilitii a olȩchȩsek.

‘John threw away my pencil.’

(22) a. Ng tȩcha a dȩngchokl ȩr a mlai?

‘Who is sitting in the car?’

b. Ng tȩcha a dȩngchokl ȩr a mlim?

‘Who is sitting in your car?’

In our discussion of obligatorily possessed nouns in 3.5

above, we mentioned expressions like tkul a daob ‘edge of the
ocean’, llel a kȩrrȩkar ‘leaf of the tree’, etc. in which an oblig-
atorily possessed noun with a third person possessor suffix is
followed by another noun which identifies the possessor more
specifically. There are many other Palauan expressions which
conform to this general pattern, as in the following:

(23) kekul a katuu ‘cat’s claw’

chimal a Toki ‘Toki’s hand’
mȩdal a oluchȩs ‘pencil point’
bȩlkul a John ‘John’s function’
blil a sȩchȩlik ‘my friend’s house’
blirir a rȩsȩchȩlik ‘my friends’ house’
rȩngrir a rȩchad ‘the people’s hearts/feelings’
bȩchil a Droteo ‘Droteo’s wife’

As the examples above show, the possessed noun at the be-

ginning of the expression may be of the obligatorily possessed
or optionally possessed variety; the third person possessor may
be singular or plural; and the noun identifying the possessor
may be human or non-human. The distribution of the word se-
quences of 23 is identical to that of any single possessed or un-
possessed noun: in other words, word sequences like those of
23 can also occur in the three different environments of 20–22
above. Thus, we can replace either of the subjects blai ‘house’
or blik ‘my house’ of 20a or 20b with blil a Droteo ‘Droteo’s
house’, to give

(24) A blil a Droteo a milsesȩb.

74
 Palauan Reference Grammar

‘Droteo’s house burned down.’

Similarly, the objects in 21 could be replaced by an expression

like olȩchȩsel a Toki ‘Toki’s pencil’, giving the sentence:

(25) A John a chilitii a olȩchȩsel a Toki.

‘John threw away Toki’s pencil.’

And, finally, the nouns following the relational word ȩr in 22

could be exchanged for an expression like mlil a sȩchȩlim ‘your
friend’s car’, resulting in

(24) Ng tȩcha a dȩngchokl ȩr a mlil a sȩchȩlim?

‘Who is sitting in your friend’s car?’

In order to describe the grammar of Palauan adequately, we

must take account of the fact that certain words and groups
of words can substitute for each other in various positions.
We have seen, for example, that blai ‘house’, blik ‘my house’,
and blil a Droteo ‘Droteo’s house’ can all occur as the subject
of a sentence like 20. Though these expressions are different
from each other in structure (blai is a single unpossessed noun,
blik is a single possessed noun, and blil a Droteo consists of
a possessed noun followed by a specific possessor), they do
share identical features of distribution. In other words, they
all can occur in exactly the same environments. In order to
“capture” this fact in our grammatical description, we introduce
the term noun phrase to refer to all those single words or
groups of words which can appear in a particular environment
and can substitute for each other in that environment. Thus,
noun phrases in Palauan can occur as subject or object of the
sentence, or can follow the relational word ȩr. Recall that in dis-
cussing the distribution of Palauan nouns in 2.3, we used the
terms subject noun and object noun, and we said that the re-
lational word ȩr is always followed by a noun. As we can now
see, this earlier discussion was deliberately oversimplified. It
is more correct to speak of subject noun phrases and object
noun phrases and to say that the relational word ȩ r is always
followed by a noun phrase rather than a noun. In the same
way, our discussion of the Palauan word a in 2.6 was much over-
simplified: it should now be clear that what a actually intro-
duces is a noun phrase rather than just a noun.

75
 3 Noun Possession

As we have seen, Palauan noun phrases include the fol-

lowing three types: single unpossessed nouns (e.g., blai ‘house’,
oluchȩs ‘pencil’), single possessed nouns (e.g. blik ‘my house’,
olȩchȩsel ‘his pencil’), and sequences consisting of a possessed
noun followed by a specific possessor (e.g. blil a Droteo
‘Droteo’s house’, mlirir a rȩsȩchȩlik ‘my friends’ car’). It will be
helpful to subclassify these types in the following way. The sim-
plest of the three types is nothing more than the independent
(or unpossessed) form of a noun and will be called a simple
noun phrase. The other two types, however, involve possession
and are more complicated in structure; these will be referred to
as noun phrases of possession.
The three types of noun phrases we have just mentioned are
not the only kinds of noun phrases found in Palauan. In the sen-
tences below, the italicized word groups are also examples of
noun phrases:

(27) a. A sidosia ȩr a Toki a klou.

‘Toki’s car is big.’

b. A chȩrmek ȩ l bilis a mlad er a elii.

‘My dog died yesterday.’

c. A rȩlluich ȩ l chad a mlei.

‘Twenty people came.’

d. Ng soam a bȩches ȩ l blik?

‘Do you like my new house?’

e. A John a kie ȩr se ȩ l blai.

‘John lives in that house.’

f. A buik ȩl mȩngitakl a Droteo.

‘The boy who is singing is Droteo.’

The types of noun phrases found in 27a and 27b will be dis-
cussed in 3.8 and 3.11 below, while those of 27c-f will be ex-
plained in subsequent chapters.

76
 Palauan Reference Grammar

3.7. NOUN PHRASES OF POSSESSION VS. NOUN

PHRASES OF CHARACTERIZATION
In Palauan we can find large numbers of noun phrases which
are formally identical to the noun phrases of possession il-
lustrated in 17 and 23 above but which are interpreted in a
very different way. First of all, let us review how the noun
phrases of possession in 17 and 23 are understood. In each
case, there is a clear-cut relationship of possession (cf. 3.1
above) which involves actual ownership of something, a whole-
part relationship, a tie of kinship, marriage, or friendship, or
some other kind of close association. Thus, in blil a sȩchȩlik
‘my friend’s house’ of 23, my friend owns the house, which is
his property; in chimal a Toki ‘Toki’s hand’ of 23 and uchul a
kȩrrȩkar ‘the trunk of the tree’ of 17, the possessed items are
actual parts of their (human or non-human) possessors; and in
bȩchil a Droteo ‘Droteo’s wife’ of 23 the possessed noun is re-
lated to the possessor by a marriage bond.
Now observe the following noun phrases and compare them
with those of 17 and 23:

(28)rȩktel a bilis chisel a Toki ‘news about Toki’

‘rabies’
blil a blȩngur kȩrul a sokȩl ‘medicine for
‘house for eating in’ ringworm’
blil a dongu ‘tool urerir a {rȩdil / ‘work (meant) for
shed’ rȩsȩchal} {women / men}’

You will notice immediately that the noun phrases of 28, though
identical in structure to those of 17 and 23, cannot at all be
interpreted as noun phrases of possession. In rȩktel a bills
‘rabies’, for example, we cannot sensibly say that we are talking
about a disease possessed by some particular dog, nor can we
say that the tools possess the shed in blil a dongu ‘tool shed’.
Furthermore, in chisel a Toki ‘news about Toki’, we are cer-
tainly not dealing with news which Toki has or possesses. How,
then, can we describe the meaning relationship which is found
between possessed noun and “possessor” in the examples of
28? It seems clear that the “possessor” noun in the examples
of 28 serves to characterize the possessed noun in various
ways—in other words, it gives some information about the pos-
sessed noun’s function, purpose, con tent, etc. For example,
blȩngur ‘eating’ in blil a blȩngur ‘house for eating in’ specifies

77
 3 Noun Possession

the function of the house; Toki in chisel a Toki ‘news about Toki’
constitutes the content or subject matter of the news; and sokȩl
‘ringworm’ of kȩrul a sokȩl ‘medicine for ringworm’ focuses on
the purpose of the medicine by telling what it is designed to
treat. Since the examples of 28 involve characterization of one
sort or another, we will call them noun phrases of character-
ization in order to distinguish them from the noun phrases of
possession observed in 17 and 23. Since noun phrases of pos-
session and noun phrases of characterization are structurally
or formally identical, these two terms are introduced primarily
to reflect the difference in meaning or interpretation described
above.

3.8. UNPOSSESSIBLE NOUNS

In 3.5 above, we clarified the distinction between optionally
possessed nouns and obligatorily possessed nouns. As opposed
to these types of nouns, which either can or must take possessor
suffixes, there are other Palauan nouns which cannot take pos-
sessor suffixes under any circumstances. Such unpossessible
nouns primarily include words borrowed into Palauan from
English, Japanese, and a few other foreign languages (e.g. stoa
‘store’, sidosia ‘car’, hong ‘book’, rrat ‘bicycle’, etc.) as well as
native Palauan words which designate animals, plants, or parts
of the natural environment (e.g. lbolb ‘wolf’, ngas ‘pine tree’,
bang ‘goatfish’, chȩlȩchol ‘beach’, etc.). It is possible, however,
for such nouns to be associated with a possessor by using a
noun phrase of possession containing the relational word ȩr. Ob-
serve the following examples:

(29) delmȩrab ȩr a sȩchȩlik ‘my friend’s room’

sidosia ȩr a Toki ‘Toki’s car’
kombalii ȩr a Droteo ‘Droteo’s company’
restorangd ȩr a John ‘John’s restaurant’
hong ȩr tir ‘their book’
mondai ȩr kau ‘your problem’
sensei ȩr ngak ‘my teacher’
office ȩr ngii ‘his/her office’

In the examples of 29, the relational word ȩr is used to link the

thing (or person) possessed with the possessor. The term rela-
tional word is assigned to ȩ r precisely because ȩr functions

78
 Palauan Reference Grammar

to express various kinds of relationships between words; in the

case under discussion, the relationship which ȩ r designates
is one of possession. Any sequence consisting of the relational
word ȩr followed by a noun (or pronoun) is called a relational
phrase. Sequences like ȩ r a sȩchȩlik, ȩ r a Toki, ȩ r tir, etc. of
29 represent a special type of relational phrase which identifies
the possessor of the immediately preceding noun; such phrases
will be called possessor phrases. 18 The examples in 29 are
therefore noun phrases of possession which contain the subtype
of relational phrase known as possessor phrase. Notice that the
possessor following ȩr is introduced by a if it is a noun, but is
not preceded by a if it is a pronoun (cf. 2.6).
A few nouns of foreign origin have been so “assimilated”
into Palauan that they have come to be used with possessor
suffixes. These include taem ‘time’—temek, temel, etc., tebȩl
‘table’—tebȩlek, tebȩlel, etc., katuu ‘cat’—katungek ‘my sweet-
heart’, katungel ‘his/her sweetheart’, etc., and babier ‘paper,
letter’—babilngek, babilngel 19 etc.
Now observe the following examples and compare them
with those of 29:

(30) sidosia ȩr a Siabal ‘Japanese car’

sensei ȩr a ochur ‘teacher of mathematics’
hong ȩr a iakiu ‘book about baseball’
delmȩrab ȩr a omȩsuub ‘room tor studying’

Since the examples of 30 consist of an unpossessible noun fol-

lowed by a relational phrase, they are formally identical to those
of 29. They are interpreted in a very different way, however.
While the examples of 29 are noun phrases of possession, those
of 30 are noun phrases of characterization (cf. the contrast be-
tween the examples of 17 and 23 vs. 28 discussed in 3.7 above).
In other words, the noun following the relational word ȩr in the
examples of 30 is not actually the possessor of the preceding
noun, but instead characterizes this noun in some way. In si-
dosia ȩr a Siabal ‘Japanese car’, for example, the country of
Japan (Siabal) is not really the possessor or owner of the car;
rather, the car is characterized by having been manufactured
in Japan. Or, in sensei ȩr a ochur ‘teacher of mathematics’, we
cannot conceive of how a non-living, abstract thing like math-
ematics could possess a teacher; instead, ochur ‘mathematics’
characterizes or identifies the teacher by telling us what subject

79
 3 Noun Possession

he teaches. Finally, in hong ȩ r a iakiu ‘book about baseball’, the

noun following ȩr characterizes hong ‘book’ by identifying its
content, while in delmȩrab ȩ r a omȩsuub ‘room for studying’,
the noun following ȩr characterizes delmȩrab ‘room’ by speci-
fying its function or purpose. Just as we labelled the relational
phrases in the examples of 29 pos sessor phrases, we shall
call the relational phrases in the examples of 30 character-
izational phrases. The examples of 30 are therefore noun
phrases of characterization which contain the subtype of rela-
tional phrase known as characterizational phrase.

3.9. NOUN PHRASES OF CHARACTERIZATION:

FURTHER EXAMPLES
It is possible for optionally possessed nouns to occur in noun
phrases of characterization such as the following:

(31) mlai ȩr a Siabal ‘Japanese car’

kall ȩr a Sina ‘Chinese food’
tȩkoi ȩr a Belau ‘Palauan (language)’
tȩkoi ȩr a blai ‘private matters’
tȩkoi ȩr a rael 20 ‘public knowledge’
techall ȩr a ureor ‘opportunity for work’
chad ȩr a omȩnged ‘fisherman’
chad ȩr a Merikel ‘American (person)’
ilumȩl ȩr a party ‘party drinks’
ngalȩk ȩr a skuul ‘student, pupil’
beluu ȩr a ngau ‘hell’
beluu ȩr a eangȩd 21 ‘heaven’

In each noun phrase of characterization above, the independent

(or unpossessed) form of an optionally possessed noun is fol-
lowed by a characterizational phrase introduced by ȩr. Can you
see clearly that the examples of 31 involve characterization
rather than actual possession?
Optionally possessed nouns can of course also occur in noun
phrases of possession such as mlil a Droteo ‘Droteo’s car’, kȩlel
a Toki ‘Toki’s food’, etc.; in such cases, the noun appears in its
possessed form and is followed by another noun which identifies
the possessor. We therefore have a contrast between a noun

80
 Palauan Reference Grammar

phrase of possession like mlil a Droteo ‘Droteo’s car’ and a noun

phrase of characterization like mlai ȩr a Siabal ‘Japanese car’.
Similar contrasts are given in the list below:

(32)Noun Phrase of Noun Phrase of Characterization

Possession

(Possessed Form + a (Independent Form +

+ specific possessor) characterizational phrase
introduced by ȩr)
kȩlel a Toki ‘Toki’s kall ȩr a Siabal ‘Japanese food’
food’
tȩkingel a Masaharu tȩkoi ȩr a Belau ‘Palauan
‘Masaharu’s words’ (language)’
techȩllel a Droteo techall ȩr a ureor ‘opportunity for
‘Droteo’s opportunity’ work’
imȩlel 22 a John ‘John’s ilumȩl ȩr a party ‘party drinks’
drink’
ngȩlȩkel a sȩchȩlik ngalȩk ȩr a skuul ‘student, pupil’
‘my friend’s child’

Some noun phrases of characterization can be associated

with specific possessors, as the following list illustrates:

(33)Noun Phrase of Noun Phrase of Characterization

Charac terization With Specific Possessor

bail ȩr a skuul bilel a Droteo ȩr a skuul ‘Droteo’s

‘school clothes’ school clothes’
bilas ȩr a omȩnged bilsȩngek ȩr a omȩnged ‘my fishing
‘fishing boat’ boat’
mlai ȩr a omilil 23 mlim ȩr a omilil ‘your car for leisure
‘car for leisure use’ use’
kall ȩr a mȩkȩmad kȩliu ȩr a mȩkȩmad ‘your C rations’
24
‘C rations’
diall ȩr a mȩkȩmad dillir a rȩchad ȩr a Siabal ȩr a
‘warship’ mȩkȩmad ‘warships of the Japanese’
hong ȩr a ochur hong ȩr a Droteo ȩr a ochur ‘Droteo’s
‘mathematics book’ mathematics book’

As you can see, the characterizational phrases on the right are

formed from those on the left simply by adding information
about the possessor. If the first noun is possessible, as in the top
five examples, the appropriate possessor suffix is added to it; if

81
 3 Noun Possession

the first noun is unpossessible, however, as in the last example,

the possessor is indicated in a relational phrase introduced by
ȩr.

3.10. APPOSITIONAL STRUCTURES

In order to specify the category of objects to which something
belongs, Palauan makes use of special expressions having the
structure noun phrase of possession + ȩl + noun. A typical ex-
ample is imȩlek ȩ l biang ‘my (drink of) beer’, in which the pos-
sessed noun imȩlek ‘my drink’ (from ilumȩl ‘drink’) specifies the
function which is served by biang ‘beer’ on a particular oc-
casion. While biang ‘beer’ names a specific kind of drink, imȩlek
‘my drink’ refers to a more general category, to any kind of bev-
erage I might drink; therefore, we can also have expressions
like imȩlek ȩl ralm ‘my (drink of) water’, imȩlek ȩ l kohi ‘my
coffee’, etc. In these expressions, the noun phrase of possession
(which happens to be a single possessed noun—cf. 3.6 above)
is linked with the following more specific noun by the word ȩl,
which, as we will see in chap. 23, identifies a relative clause
construction. 25 Structures like imȩlek ȩl biang, in which one
thing is linked with or equated with another, are termed appo-
sitional; other appositional structures of Palauan, such as John
ȩl sensei ‘John the teacher’, will be examined in 23.4.
Appositional structures like imȩlek ȩ l biang refer primarily
to categories of food and drink. The category word can take any
of the possessor suffixes—e.g. imȩlel ȩ l biang ‘his/her (drink
of) beer’, imȩlem ȩl biang ‘your (drink of) beer’, etc.—and a
third person possessor may be specified—e.g. imȩlel a Droteo ȩ
l biang ‘Droteo’s drink of beer’. All appositional structures can
be classified as noun phrases on the basis of their distribution
(cf. 3.6 above). In the list below, the most commonly used cat-
egory words are presented in their independent forms, together
with an example of their use in an appositional structure:

(34)Independent Form of Category Appositional Structure

Word (with 1 st person sg.
possessor)

kall ‘food’ kȩlek ȩl udong ‘my

noodles’

82
 Palauan Reference Grammar

chȩled ‘fish or animal which chȩldik ȩl ngikȩl ‘my fish’

has been caught but not
cooked’
odoim ‘meat or fish which is odimek ȩl babii ‘my pork’
ready to eat’
ongraol ‘starchy food’ ongulek 26 ȩl kukau ‘my
taro’
kliou ‘dessert’ kliungek ȩl tuu ‘my banana
(for dessert)’ 27
charm ‘animal’ chȩrmek ȩl babii ‘my pig’
chȩmachȩl ‘something to chȩmȩlek 28 ȩl buuch ‘my
chew’ betel nut’
udoud ‘money’ ududek ȩl lluich ȩl kluk ‘my
twenty dollars’
dȩllomȩl ‘something planted’ dȩllȩmȩlek ȩl diokang ‘my
tapioca plant’

As the contrasting appositional structures odimek ȩl babii

‘my pork’ and chȩrmek ȩl babii ‘my pig’ illustrate, it is possible
for the same thing (in this case, babii ‘pig’) to be associated with
different types of categories. Another example of this type of
contrast is seen in the expressions dȩllȩmȩlek ȩ l tuu vs. kȩlek
ȩ l tuu: while the former expression categorizes the bananas as
something growing and is best translated as ‘my banana plants’,
the latter expression views the bananas as ready-to-eat food and
has the English equivalent ‘my bananas (for eating)’.

3.11. SOME SPECIAL POSSESSED NOUNS

A small group of nouns referring to liking, disliking, ability, and
obligation always occur with possessor suffixes. The possessed
forms of these nouns are shown in the simple sentences below:

(35) Ng soak a biang.

‘I like beer.’

(36) Ng chȩtil a rrom.

‘He dislikes liquor.’

(37) Ng sȩbȩchir ȩl mong?

‘Can they go?’

83
 3 Noun Possession

(38) Ng kirem ȩl mong?

‘Do you have to go?’

Though the English equivalents contain verbs or ‘helping’ words

like can and have (to), the Palauan sentences use possessed
nouns to express the same concepts: 35, for example, contains
the possesssed noun soak ‘my liking’ and means, literally, ‘My
liking is beer’. For a complete analysis of sentences like 35–38,
see 17.1 and 17.3.

3.12. SUMMARY OF PALAUAN NOUN PHRASES

The three major types of noun phrases discussed in this chapter
are the following:

a. simple noun phrases

b. noun phrases of possession
c. noun phrases of characterization

Simple noun phrases merely consist of a single noun without

a possessor suffix: this single noun may be an unpossessible
noun (e.g. delmȩrab ‘room’, sensei ‘teacher’) or the unpos-
sessed (or independent) form of an optionally possessed
noun (e.g. blai ‘house’, charm ‘animal’).
Noun phrases of possession include many different types,
which are summarized in figure 2.
Noun phrases of characterization also have several sub-
types, as indicated in figure 3.

84
 Palauan Reference Grammar

Figure 2

Figure 3

85
 4 Palauan Pronouns

4PalauanPronouns
4.1. INDEPENDENT PRONOUNS AND AFFIX
PRONOUNS
At various points in the preceding three chapters we have had
occasion to mention something about the pronouns of Palauan.
We found it necessary to introduce many different facts about
Palauan pronouns simply because the use of pronouns is so
closely tied up with problems of the spelling and the grammar.
Thus, in 1.5.d we discussed how the different sets of Palauan
pronouns are to be spelled; in 2.4 we examined some of the dis-
tributional characteristics which pronouns share with nouns, in
addition to showing why the distinction human vs. non-human is
important to pronoun usage; and, finally, in 3.2–3 we analyzed
the possessor pronouns (or suffixes) extensively. There are still
many facts about Palauan pronouns which we have not pre-
sented; this will be done in the present chapter.
As we saw in 1.5.d, some Palauan pronouns are spelled as
separate words, while others are spelled as part of another
word. We shall call the former type independent pronouns,
whereas the latter type will be referred to as affix pronouns. 1
Affix pronouns never occur alone but must always be attached
to the beginning or end of a word; in the former case they are
prefixes, while in the latter case they are suffixes. These two
terms have already appeared in the text and should be familiar
to you: in 2.5, for example, we discussed the human plural
prefix rȩ- (as in rȩngalȩk ‘children’), and in 3.3 we described
the different sets of possessor suffixes (as in bilek ‘my clothes’,
kȩruk ‘my medicine’, etc.).
As we have already seen in the preceding chapters, pro-
nouns have the same distributional features as nouns and refer
to various persons like ‘I, you, he, they, etc’ First person pro-
nouns and second person pronouns make reference to the
participants in a conversation and identify the speaker(s) (‘I,
we’) and the hearer(s) (‘you’), respectively. Third person pro-
nouns do not refer to any of the participants in a conversation
but identify the person(s) or thing(s) being talked about (‘he,
she, it, they’). First, second and third person pronouns can

86
 Palauan Reference Grammar

be either singular or plural, though the third person plural

pronoun refers to human nouns only. In addition, there are two
first person plural pronouns—inclusive and exclusive (cf. 3.2).
The distinctions among Palauan pronouns can therefore be sum-
marized as follows:

(1) 1st pers. sg. 1st pers. pl. incl.

1st pers. pl. excl.
2nd pers. sg. 2nd pers. pl.
3rd pers. sg. 3rd pers. pl. (human only)

Though some special comments will be necessary in a few

cases, we will be able to use the above seven categories to de-
scribe all of the pronoun sets of Palauan. To save space, we will
also adopt the abbreviations given above.

4.2. NON-EMPHATIC VS. EMPHATIC SUBJECTS

There are two sets of independent pronouns which can occur in
the position of sentence subject. One is called non-emphatic
and the other emphatic. Before explaining the difference be-
tween these types, let us list the forms occurring in each:

(2) Non-emphatic Pronouns Emphatic Pronouns

1st pers sg ak ngak
2nd pers sg kȩ kau
3rd pers sg ng ngii
1st pers pl incl kȩdȩ kid
1st pers pl excl aki kȩmam
2nd pers pl kom kȩmiu
3rd pers pl tȩ tir

Though we can see some similarities (especially in the con-

sonants) between corresponding non-emphatic and emphatic
forms, there is no predictable way of deriving one set from
the other. While the emphatic pronouns are always stressed
(cf. 1.4.2) and are pronounced as they are spelled, the non-em-
phatic ones are never stressed and sometimes show a slight
variation in pronunciation. For example, when kȩ ‘you (sg.)’,
kȩdȩ ‘we (incl.)’, and tȩ ‘they’ appear before vowel-initial verbs,
their final ȩ is lost in the pronunciation, as in kȩ ulȩmes
[kulǝmεs] ‘you saw’, kȩdȩ ousbech [kǝðouspεɂ] ‘we need’, and

87
 4 Palauan Pronouns

tȩ omȩngur [tomǝŋur] ‘they are eating’. The unpronounced ȩ

is always spelled, however. Furthermore, when ak ‘I’ and aki
‘we (excl.)’ follow words ending in a full vowel, the a is not
pronounced, as in e le ak [εlεkh] ‘because I…’ and a lȩ ko
ak [alǝgokh] ‘I intended to…’ If, however, the preceding word
has a final ȩ, this schwa is deleted and the following a is pro-
nounced, as in mȩ ak [makh] ‘so I…’. Finally, we know from ex-
pressions like kom osiik [komosiykh] ‘you are looking for’, kom
ua ngarang? [komwaŋaraŋ] ‘how are you?’ and kom smechȩr
[komsmεɂǝr] ‘you are sick’ that the correct form of the second
person plural non-emphatic pronoun is kom; if the word fol-
lowing kom begins with an m, however, only a single m is ac-
tually pronounced, as in kom mȩrredȩl [komǝr̄εðǝl] ‘you are the
leaders’.
When they appear as sentence subject, the non-emphatic
and emphatic pronouns are used in very different ways. Observe
the following brief dialogs:

(3) A: Kȩ mo ȩr ker?
‘Where are you going?’

B: Ak mo ȩr a stoang.
‘I’m going to the store.’

(4) A: Ng tȩcha a mo ȩr a stoang?

‘Who is going to the store?’

B: Ngak a mo ȩr a stoang.
‘Í’m going to the store.’

As we can see, A’s questions in 3 and 4 above are very different

and therefore elicit different types of responses from B. In 4, A
assumes (or knows) that from among a certain group of people,
one person is going to the store; his question to B asks for the
identity of that one person. B uses the emphatic pronoun ngak
‘I’ in his answer to emphasize the fact that it is he himself (and
not Droteo, or Toki, or anyone else who might be around) who
is going to the store. In other words, the use of an emphatic
pronoun as subject implies a contrast (or opposition) between
the person who actually perform some activity and any other
persons who might be available to perform it but who for some
reason do not. This kind of contrastive emphasis is indicated

88
 Palauan Reference Grammar

in the pronunciation of the English equivalent of 4B by an espe-

cially strong stress on the subject pronoun (the accent mark is
used to point this out). Another situation in which 4B would be
appropriate is the following:

(5) A: A Droteo ng mo ȩr a stoang?

‘Is Droteo going to the store?’

B: Ng diak. Ngak a mo ȩr a stoang.

‘No. Í’m going to the store.’

In the dialog above, B uses the emphatic pronoun in his re-

sponse in order to express an opposition between himself and
Droteo as persons who might be available to go to the store.
It is also possible for B’s response in 4 to be shortened to the
emphatic pronoun itself or to a short sentence of the form ng +
emphatic pronoun, meaning ‘it’s—.’ Note the following dialog:

(6) A: Ng tȩcha a mo ȩr a stoang?

‘Who is going to the store?’

B: (Ng) ngak.
‘(It’s) me.’

B need not mention the activity of going to the store in his re-
sponse, since this activity is clearly stated in A’s question. An-
other similar dialog is the following:

(7) A: A Toki a milȩnga a ngikȩl.

‘Toki ate fish.’

B: E kau?
‘And (how about) you?’

In the dialog above, B can ask A what A ate simply by using the
emphatic pronoun kau ‘you’; since A has already talked about
someone eating something, it is clear that B is likewise referring
to the activity of eating.
The third person emphatic pronouns ngii ‘he, she’ and tir
‘they’ can be used in B’s responses in dialogs like 4–7 above if
both A and B know whom the pronoun refers to. Observe the
conversations below:

89
 4 Palauan Pronouns

(8) A: Ng tȩcha a silsȩbii a blai?

‘Who burned down the house?’

B: (Ng) ngii.
‘(It’s) him.’

(9) A: Tȩ rua tȩcha a silsȩbii a blai?

‘Who burned down the house?’

B: (Ng) tir.
‘(It’s) them.’

Now that we have seen how the emphatic pronouns are used
as sentence subject, let us return to a discussion of 3, which
contains a non-emphatic pronoun as subject in B’s response:

(3) A: Kȩ mo ȩr ker?
‘Where are you going?’

B: Ak mo ȩr a stoang.
‘I’m going to the store.’

While the emphatic pronoun ngak ‘I’ of 4B involves contrastive

emphasis, the non-emphatic pronoun ak ‘I’ of 3B does not. In 3,
A is not asking B to single out the particular person who, from
among a certain group of people, is performing some activity;
therefore, B’s answer does not provide this kind of information.
Instead, A simply wants to know where B is going: automati-
cally assuming that B is the only person involved in the activity,
A concentrates on finding out the place B is going to. In his
response, B uses the non-emphatic pronoun ak ‘I’ because he
does not need to provide A with any new information about the
subject of the sentence; ak ‘I’ is merely old information, and the
new information of 3B is supplied by the relational phrase ȩr a
stoang ‘to the store’.
We can therefore characterize the difference between non-
emphatic vs. emphatic (subject) pronouns as follows. The infor-
mation supplied by a non-emphatic pronoun is old—that is, both
speaker and hearer know about it, and it is no longer of interest
to them. On the other hand, the information provided by an em-
phatic pronoun is new—something previously unknown or un-
expected.

90
 Palauan Reference Grammar

Because non-emphatic pronouns express old information,

they can only be used as sentence subjects in situations where
their referents are clear. This is the case in the following short
dialogs, where the non-emphatic pronouns under consideration
have been italicized:

(10) A: A Droteo a mlo ȩr a Guam ȩr a elii.

‘Droteo went to Guam yesterday.’

B: Ng mo lmuut ȩl mè ȩr oingarang?
‘When is he coming back?’

(11) A: A John mȩ a Mary a me ȩr a blik ȩr a klukuk.

‘John and Mary are coming to my house tomorrow.’

B: Tȩ chad ȩr a Merikel?
‘Are they Americans?’

(12) A: A Toki ng oureor ȩr ker?

‘Where does Toki work?’

B: Ng oureor ȩr a bangk.
‘She works at the bank.’

While the sentences spoken by B in 10–12 above are perfectly

natural as part of a conversational exchange, they would sound
very strange if spoken in isolation or “out of context”.
As we noted in 2.6, all Palauan nouns (or, more properly,
noun phrases—cf. 3.6) must be introduced by the word a, unless
they consist of pronouns. Thus, the independent pronouns (non-
emphatic or emphatic) discussed in this section are never pre-
ceded by a. We noted also in 2.6 that Palauan verbs (or, more
properly, verb phrases—see 5.2) are always introduced by a
unless the sentence subject is a pronoun. As the examples in
this section show, we must now qualify this statement to read as
follows: the verb phrase of a sentence is always preceded by a
except when the sentence subject is a non-emphatic pronoun.
If the sentence subject is an emphatic pronoun, however, as in
4B and 5B above, the verb phrase is indeed introduced by a. The
occurrence or non-occurrence of a before the verb phrase of a
sentence is summarized below:

(13)Sentence Type Example

91
 4 Palauan Pronouns

a. non-pronominal subject + A sȩchȩlik a smechȩr. ‘My

verb phrase. friend is sick.’
A Satsko a smechȩr.
‘Satsko is sick.’

b. emphatic pronoun subject Ngak a smechȩr. ‘Í’m

+verb phrase. sick.’
Ngii a smechȩr. ‘Shé’s
sick.’

c. non-emphatic pronoun Ak smechȩr. ‘I’m sick.’

subject +verb phrase.
Ng smechȩr. ‘She’s sick.’

Notice that the emphatic pronouns illustrated in 13b behave

like the independent nouns sȩchȩlik ‘my friend’ and Satsko of
13a in requiring that the following verb phrase be introduced
by a. Because of this fact, many linguists would argue that the
emphatic pronouns are truly independent words, while the non-
emphatic pronouns of 13c are actually prefixes. If the non-em-
phatic pronouns were indeed prefixes, then the nonoccurrence
of a before the verb phrases of 13c could be explained. As-
suming that this analysis were correct, it would appear more
proper to spell the non-emphatic pronouns as part of the fol-
lowing verb—e.g. akmong ‘I’m going’, ngsmechȩr ‘she’s sick’,
tȩme ‘they are coming’, etc. The 1972 Palauan Orthography
Committee decided against spelling the non-emphatic pronouns
as part of the following word, however. The present system of
spelling the non-emphatic pronouns as separate words never-
theless has some merits: in sentences like the following, for
example, it is easy to analyze the structure and identify the mor-
phemes because they are spelled as separate words:

(14) a. Ak ko ȩr a sesmechȩr.
‘I’m rather sick.’

b. Tȩ, di mililil ȩr a Guam.

‘They only fooled around in Guam.’

c. A John a dimlak lȩbo e le ng mle smechȩr.

‘John didn’t go because he was sick.’

92
 Palauan Reference Grammar

4.3. USE OF EMPHATIC PRONOUNS FOLLOWING

ȨR
In 4.2 above we saw how the emphatic pronouns of Palauan can
function as sentence subjects. In this and the following sections
we will describe further environments in which the emphatic
pronouns are observed to occur. Perhaps the most common of
these is after the word ȩr. Recall that in 2.7 we distinguished
between the specifying word ȩr, which identifies a specific
object, and the relational word ȩ r, which expresses certain
types of relationships (mostly spatial and temporal) and has a
large range of English equivalents, including ‘to, at, in, out of,
because of, etc.’ We find that the emphatic pronouns of Palauan
can occur following both of these ȩr’s.
In the sentences below, an emphatic pronoun functions as
sentence object and follows the specifying word ȩr:

(15) a. Kȩ mȩlasȩm ȩr ngak?

‘Are you challenging me?’

b. Ng tȩcha a milȩngȩlebȩd ȩr kau?

‘Who was hitting you?’

c. Ak ulȩmes ȩr {ngii / tir} ȩr a party.

‘I saw {him/her / them} at the party.’

d. Ak dirk mȩluchȩs ȩr ngii.

‘I’m still writing it.’

Since Palauan pronouns always refer to specific persons or

things, it is not surprising that the specifying word ȩr should
precede the object pronouns in the examples of 15. Note that
the sentences of 15, especially those with third person emphatic
pronouns as objects, would sound very strange when uttered in
isolation. It is not difficult, however, to find contexts in which
they would be completely natural; 15d, for example, is appro-
priate as part of B’s response in the dialog below:

(16) A: Kȩ mla mo mȩrek ȩr a babier?

‘Have you finished (writing) the letter?’

B: Ng diak. Ak dirk mȩluchȩs ȩr ngii.

‘No. I’m still writing it.’

93
 4 Palauan Pronouns

You may have noticed that although we have been speaking

of the emphatic pronouns in object position, the specific
pronoun objects of 15 (ȩr ngak ‘me’, ȩ r ngii ‘him/her/it’, etc.) do
not seem to have the implication of contrastive emphasis which
we observed for emphatic pronouns in subject position (cf. 4.2
above). Unfortunately, we cannot explain why this should be the
case. Even though the “emphatic” pronouns do not imply con-
trastive emphasis in all environments, we will continue to use
the term as a convenient way of identifying the pronoun set
ngak ‘I’, kau ‘you’, ngii ‘he/she/it’, etc.
In the following sentences, an emphatic pronoun follows the
relational word ȩr:

(17) a. A sensei ȩr ngak a me ȩr a party.

‘My teacher is coming to the party.’

b. A delmȩrab ȩr tir a kmal kikiongȩl.

‘Their room is very dirty.’

c. A beab a tilobȩd ȩr ngii.

‘The mouse came out of it.’

In 17a-b, the sequences ȩr ngak ‘my’ and ȩr tir ‘their’ are pos-
sessor phrases (cf. 3.8), while in 17c the relational phrase ȩr
ngii ‘out of it’ refers to the place (hole, box, etc.) from which
the mouse emerged. Like the examples of 15, these sentences
do not have any implication of contrastive emphasis.

4.4. EMPHATIC PRONOUNS IN COORDINATE

NOUN PHRASES
Another environment in which the emphatic pronouns are used
(but without any implication of contrastive emphasis) is in co-
ordinate noun phrases (see 25.4). A coordinate noun phrase
is one which consists of two (or more) nouns joined by the word
mȩ ‘and’. Coordinate noun phrases have the same distribution
as other noun phrases and can therefore occur as sentence
subject or object, or can follow the relational word ȩr. These
three uses are illustrated in the sentences below, in which the
coordinate noun phrases consist of sequences of human or non-
human nouns joined by mȩ ‘and’ 2 :

94
 Palauan Reference Grammar

(18) a. A Toki mȩ. a Droteo a mlo ȩr a stoang.

‘Toki and Droteo went to the store.’

b. Tȩ mȩnga a ngikȩl mȩ a kukau mȩ a diokang.

‘They eat fish and taro and tapioca.’

c. Ng sidosia ȩr a Toki mȩ a Droteo.

‘It’s Toki and Droteo’s car.’

Sentences with coordinate noun phrases represent a shortened

or condensed way of expressing information: for example,
Palauan speakers would use 18a rather than something like the
following, in which mlo ȩr a stoang ‘went to the store’ is re-
peated twice:

(19) A Toki a mlo ȩr a stoang, mȩ a Droteo a mlo ȩr a stoang.

‘Toki went to the store, and Droteo went to the store.’

The relationship between sentences like 18a and 19 will be ex-

plained more carefully in chap. 25, where a complete analysis
of the word mȩ ‘and’ is given.
If a coordinate noun phrase contains one or more pronouns,
the members of the emphatic pronoun set are always used. Ob-
serve the following sentences:

(20)a. Ngak mȩ a Helen a kausȩchȩlei.

‘Helen and I are friends.’

b. Tir mȩ a rȩchad ȩr a Merikel a blȩchoel ȩl kauchȩraro.

‘They and the Americans always treat each other as
enemies.’

c. A Droteo a kautokȩtok ngii mȩ a Toki.

‘Droteo and Toki are arguing.’

d. Kȩ mlo ȩr a party kau mȩ a tȩchang? 3

‘With whom did you go to the party?’

As the examples of 20 show, a coordinate noun phrase is not

preceded by a if the first word in it is a pronoun. For an analysis
of sentences 20c-d, which have coordinate noun phrases (ngii
m ȩ a Toki ‘he and Toki’, kau mȩ a t ȩchang ‘you and who?’) in
sentence-final position, see 17.5. Note that the English equiva-

95
 4 Palauan Pronouns

lents given for 20c-d do not reflect the structure of the Palauan
sentences, which, if translated literally, would correspond to
‘Droteo is arguing—he and Toki’ and ‘You went to the
party—you and who?’

4.5. EMPHATIC PRONOUNS PRECEDED BY DI

The emphatic pronouns occur in special expressions of the form
di + (mle +) emphatic pronoun to indicate that a given person
is or was the exclusive participant in some activity. In other
words, these expressions imply that one and only one individual
is involved in the action in question; by excluding the possibility
that other persons may be involved, these expressions produce
a kind of contrastive emphasis. They are very similar in meaning
to English expressions like ‘(by) myself, (by) himself, etc’ Note
that di means ‘only’ or ‘just’ (see 24.6, ex. 46) and that mle ‘was’
precedes the emphatic pronoun if the action of the sentence
took place at some time in the past. Some typical sentences con-
taining these expressions (italicized) are given below:

(21)a. A ngȩlȩkek a sȩbȩchel ȩl mo ȩr a che ȩl di ngii.

‘My child is able to go fishing (all) by himself/on his
own.’

b. A John a di mle ngii ȩl mȩsuub ȩl oukita.

‘John studied the guitar by himself.’

c. Ng di mle ngak ȩl mȩruul ȩr a blik.

‘I built my house (all) by myself.’

d. A ta ȩr tir a di mle ngii ȩl mȩkodir ȩl oba a dub.

‘One of them killed himself with a grenade.’

The grammatical structures found in the examples of 21 are

rather complicated and will not be explained until chap. 15.

4.6. OTHER USES OF EMPHATIC PRONOUNS

In this section we will look at yet further ways in which Palauan
emphatic pronouns can be used. Since the grammatical struc-
tures involved are very complex, we will postpone detailed dis-

96
 Palauan Reference Grammar

cussion of them until later chapters. Observe, for example, the

following sentences, in which the emphatic pronouns have been
italicized:

(22) a. Ng dimlak lȩngak a silsȩbii a blai.

‘It wasn’t me who burned down the house.’

b. Ng dimlak ltir a milkodir a bilis.

‘It wasn’t them who killed the dog.’

c. Ngii di ȩl chad a sȩbȩchel ȩl mo ȩr a skuul.

‘Any person has the right to go to school.’

d. A hong a longuiu ȩr ngii a Droteo.

‘The book is being read by Droteo.’

Examples 22a and 22b are emphatic sentences which deny that
some person or persons were connected with a particular event;
both of these sentences contain dimlak ‘wasn’t, weren’t’, which
is the past form of the negative verb diak ‘isn’t’ (see chap. 18).
In sentence 22c, the subject noun phrase ngii di ȩl chad is used
to refer to people in general and has English equivalents like
‘any person, anybody, anyone (at all), etc’ Finally, in 22d, ȩr ngii
is a special type of phrase which adds nothing to the meaning
but which is required when the subject of a passive sentence
(see 19.7) is singular.

4.7. NON-EMPHATIC PRONOUNS AS

PRONOMINAL TRACES
There are several situations in which Palauan third person non-
emphatic pronouns appear in sentences as a result of certain
grammatical processes. Observe, for example, the following
pairs of sentences:

(23) a. A Droteo a mla mei. ‘Droteo has come.’

b. Ng mla me a Droteo.

(24) a. A ralm a mȩkȩlȩkolt. ‘The water is cold.’

b. Ng mȩkȩlȩkolt a ralm.

97
 4 Palauan Pronouns

While the a-sentences have their subjects at the beginning (as

we would expect), in the b-sentences the subjects have been
shifted to the right of the verb phrases (mla me ‘has come’
in 23 and mȩkȩlȩkolt ‘cold’ in 24). If we consider the a-sen-
tences to be basic, we can simply say that the b-sentences are
derived from them by a process of subject shifting. When a
sentence subject is shifted in this way, a trace of it remains
in its original position in the form of a non-emphatic pronoun.
In the b-sentences above, the sentenceinitial third person sin-
gular non-emphatic pronoun ng stands for the singular subjects
Droteo and ralm ‘water’, which have been shifted out of their
original position. If the shifted subject is human plural, then the
third person plural non-emphatic pronoun tȩ ‘they’ appears as
a trace of it in sentence-initial position. This is illustrated in the
following pair of examples:

(25) a. A rȩsȩchȩlim a mla mei. ‘Your friends have come.’

b. Tȩ mla me a rȩsȩchȩlim.

When the non-emphatic pronouns ng and tȩ appear as the result

of subject shifting, we will refer to them as pronominal traces.
In the b-sentences of 23–25 above, the verb phrase is not intro-
duced by a precisely because it is preceded by a nonemphatic
pronoun (cf. the discussion at the end of 4.2 above).
It is very difficult to describe the meaning difference be-
tween the a- and b-sentences of 23–25. Often, the members of
pairs like 23–25 seem to be interchangeable, but sometimes the
following subtle distinction is observed. While the a-sentence
conveys totally new and unexpected information, the b-sentence
merely confirms that something expected has happened. Thus,
for example, both members of 23 say that Droteo has arrived,
but their implications are different: in 23a, the speaker had no
idea that Droteo would come, while in 23b he was waiting for or
expecting Droteo’s arrival.
The sentences below are also formed by the process of sub-
ject shifting and therefore show a pronominal trace in initial po-
sition:

(26) a. Ng mȩringȩl a chimal a Droteo.

‘Droteo’s hand hurts.’

b. Ng klou a ultutȩlel a babier.

98
 Palauan Reference Grammar

‘The letter is (very) important.’

c. Ng lluich mȩ a etiu a rȩkil a Toki.

‘Toki is twenty-nine years old.’

d. Ng suebȩk a rȩngul a Droteo. 4

‘Droteo is worried.’

e. Ng ngar ȩr ngii a party ȩr a blil a Toki.

‘There’s a party at Toki’s house.’

f. Ng diak a urerek ȩr a elȩchang.

‘I don’t have any work now.’

g. Ng soak a Toki.
‘I like Toki.’

h. Ng chȩtil a kohi.
‘He dislikes coffee.’

In sentences 26a–d the shifted subjects chimal a Droteo

‘Droteo’s hand’, ultutȩlel a babier ‘importance of the letter’,
rȩkil a Toki ‘Toki’s age’, and rȩngul a Droteo ‘Droteo’s heart’
are all noun phrases of possession which designate a specific
third person possessor (cf. 3.6). For 26a–c, the corresponding
sentences with (non-shifted) sentence-initial subjects are also
correct and acceptable, though some speakers seem to prefer
using sentences with shifted subjects. Thus, with 26a, for ex-
ample, compare the following, which is identical in meaning:

(27) A chimal a Droteo a mȩringȩl.

‘Droteo’s hand hurts.’

Can you supply the sentences with (non-shifted) sentence-initial

subjects from which 26b–c are derived?
Another very common Palauan sentence type is derived from
sentences like 26a–d. With 26d, for example, compare the
following, which has the same meaning:

(28) A Droteo a suebȩk a rȩngul.

‘Droteo is worried.’

99
 4 Palauan Pronouns

This sentence has been derived from 26d by taking the specific
possessor Droteo, which was originally shifted to the right of
the verb phrase as part of the shifted subject rȩngul a Droteo
‘Droteo’s heart’, and moving it back to sentence-initial position,
where it replaces the pronominal trace ng. This process, which
we will call preposing of possessor, is widespread in Palauan
and will be dealt with fully in 17.3.
For sentences 26d–h, which have the shifted subjects rȩngul
a Droteo ‘Droteo’s heart’, party, urerek ‘my work’, Toki, and
kohi ‘coffee’, Palauan speakers do not regularly use the corre-
sponding sentences with (non-shifted) sentence-initial subjects.
In other words, subject shifting seems to be required (or oblig-
atory) in certain types of sentences: these include sentences
like 26d, which contain special expressions with reng ‘heart’,
sentences like 26e–f, which have the affirmative and negative
expressions of existence ngar ȩr ngii ‘there is/are’ and diak
‘there isn’t/aren’t’ (see 18.2– 2.1), and sentences like 26g–h,
which contain possessed forms of the nouns of liking—soak,
soal, etc.—and disliking—chȩtik, chȩtil, etc. (cf. 3.11 and see
17.1). In 26e–f, notice that the shifted subjects are followed by
the relational phrases ȩr a blil a Toki ‘at Toki’s house’ and ȩr
a elȩchang ‘now’. These sentences show that we were correct
in stating that the sentence subject is shifted to the right of
the verb phrase rather than to the very end of the sentence.
It just so happens that in sentences like 23b, 24b, 25b, 26a–d,
and 26g–h, shifting the sentence subject to the right of the verb
phrase also placed it in sentence-final position, since these sen-
tences have no relational phrases.
The process of subject shifting is also used to form certain
types of Palauan questions. In the pairs of examples below, the
a-sentence is a statement of fact, while the b-sentence is a yes-
no question. Yes-no questions ask whether such-and-such is
the case and can be answered by ‘yes’ or ‘no’.

(29) a. A Droteo a mla mei.

‘Droteo has come.’

b. Ng mla me a Droteo?
‘Has Droteo come?’

(30) a. A rȩsȩchȩlim a mla mei.

‘Your friends have come.’

100
 Palauan Reference Grammar

b. Tȩ mla me a rȩsȩchȩlim?
‘Have your friends come?’

As you can see, the yes-no questions of 29b and 30b have
been derived from the corresponding statements of 29a and
30a by shifting the subject to the right of the verb phrase; this
process leaves a pronominal trace in sentence-initial position.
You will notice that although the order of words in 29b and 30b
is exactly the same as that in 23b and 25b, respectively, the
over-all pronunciation of these sentences shows an important
difference: in the yes-no questions of 29b and 30b, the pitch (or
intonation) of the voice rises sharply at the end of the sentence,
while in the statements of 23b and 25b, no such rise in pitch is
heard, but rather a slight lowering. The following yes-no ques-
tions are similar in structure to 29b and 30b:

(31) a. Ng klȩbokȩl a bȩchil a Droteo?

‘Is Droteo’s wife pretty?’

b. Ng lmuut ȩl me a Cisco ȩr a klukuk?

‘Is Cisco returning tomorrow?’

c. Ng mlo dȩkimȩs a ears?

‘Did the sail get wet?’

What are the statement sentences which correspond to the yes-

no questions of 31?
In addition to the yes-no questions of 29b, 30b, and 31, there
are also yes-no questions of the following type in Palauan:

(32) A Droteo ng mla mei?

‘Has Droteo come?’

(33) A rȩsȩchȩlim tȩ mla mei?

‘Have your friends come?’

(34) A bȩchil a Droteo ng klȩbokȩl?

‘Is Droteo’s wife pretty?’

Although 32, 33, and 34 do not differ in meaning from 29b, 30b,
and 31a, respectively, they do show a difference in the order of
words. If we consider the latter sentences as basic, it seems as if

101
 4 Palauan Pronouns

the questions of 32–34 are derived by moving the shifted subject

back to sentence-initial position, but preceding the pronominal
trace ng or tȩ This and other processes of question formation
will be examined in much greater detail in chap. 20.

4.8. PRONOMINALIZATION
As we have seen in 4.7 above, the third person non-emphatic
pronouns ng ‘he, she, it’ and tȩ ‘they’ can appear as pronominal
traces for subjects which have been shifted from their original
position. In this section, we will examine another way in which
these non-emphatic pronouns can come to appear in sentences.
Observe the examples below:

(35)a. A Droteo a mlo ȩr a hospital e le ng smechȩr.

‘Droteo went to the hospital because he’s sick.’

b. A rȩsȩchȩlik a mlo ȩr a hospital e le tȩ smechȩr.

‘My friends went to the hospital because they’re sick.’

The sentences of 35 each contain two parts (or clauses—see

chap. 22): in the first clause the speaker describes an event
which occurred, and in the second clause (introduced by e le
‘because’), he gives the reason for the event. Each of the two
clauses in 35a and 35b has its own subject and verb phrase:
for example, in the first clause of 35a, the subject Droteo is fol-
lowed by the verb phrase mlo ‘went’, while in the second clause,
the subject ng ‘he’ is followed by the verb phrase smechȩr
‘sick’. The third person singular non-emphatic pronoun ng ‘he’
of the second clause can only refer to Droteo, the subject of
the first clause. Example 35b is identical in structure, except
that the third person plural non-emphatic pronoun tȩ ‘they’ of
the second clause refers to the plural subject rȩsȩchȩlik ‘my
friends’ of the first clause.
In the examples of 35, a pronominal subject is used in the
second clause in order to avoid having to repeat the full subject
of the first clause. With 35a, for instance, compare the following
example, in which Droteo is repeated as subject of the second
clause:

(36) A Droteo a mlo ȩr a hospital e le a Droteo a smechȩr.

‘Droteo went to the hospital because Droteo is sick.’

102
 Palauan Reference Grammar

Most Palauans would never use a sentence like 36 because

repetition of Droteo as subject of the second clause would seem
clumsy or awkward, or even childish. The English equivalent of
36 would not be used for similar reasons. We can see, then, that
if a Palauan sentence has two adjacent clauses whose subjects
refer to exactly the same person (or thing), then the subject
of the second clause must be a pronoun. If we assume a sen-
tence like 36 to be basic because the full subject is specified
in the second clause, we can say that 35a is derived from it
simply by pronominalizing the second occurrence of Droteo.
In other words, 35a is formed by the process of pronominal-
ization, which replaces the second occurrence of some fully-
specified subject by the appropriate third person non-emphatic
pronoun. When we say “appropriate”, we mean that the non-em-
phatic pronoun must agree in number with the fully-specified
subject which it replaces: thus, ng ‘he’ of 35a replaces the sin-
gular subject Droteo, while tȩ ‘they’ of 35b substitutes for the
plural subject rȩsȩchȩlik ‘my friends’.
Some other sentences in which the second occurrence of a
subject has been pronominalized are given below:

(37) a. A John a mlo smechȩr mȩ ng mlo ȩr a hospital.

‘John got sick, so he went to the hospital.’

b. A Toki a dilu ȩl kmo ng mong. 5

‘Toki said she is going.’

c. A lsȩkum a Droteo a me e ng me kie ȩr a blik.

‘If Droteo comes, he will stay at my house.’

d. A lak losuub a rȩsȩchȩlik e tȩ mo fail ȩr a test.

‘If my friends don’t study, they will fail the test.’

The various grammatical structures found in the examples of 37

are beyond the scope of our present discussion and will be dealt
with in later chapters. Before leaving these examples, however,
we should note that pronominalization can only apply to the
second (or rightmost) occurrence of the subject, but not to the
first (or leftmost) occurrence. If we try to apply pronominal-
ization “leftwards” rather than “rightwards”, we do not get a
correct sentence with the intended meaning. With 37a, for ex-
ample, compare the following:

103
 4 Palauan Pronouns

(38)Ng mlo smechȩr mȩ a John a mlo ȩr a hospital.

‘He got sick, so John went to the hospital [to get him some
medicine, etc.].’

Example 38 is different in meaning from 37a and can only make

sense if ng ‘he’ of the first clause and John of the second clause
are interpreted as referring to different persons; also, 38 would
only be acceptable if both speaker and hearer knew whom the
pronominal subject ng referred to.
So far we have only seen sentences in which pronominaliza-
tion applies to the subject of the second clause. It is also
possible for pronominalization to apply to an object or to a pos-
sessor following ȩr, as in the following examples:

(39)a. Ak lilȩchȩsii a babier e a Droteo a mlo send ȩr ngii.

‘I wrote the letter and Droteo mailed it.’

b. A lsȩkum a Droteo a mo ȩr a party, e ak mo omes ȩr

ngii.
‘If Droteo goes to the party, then I’ll be seeing him
(there).’

c. A lsȩkum ak mȩsa a Droteo ȩr a klukuk, e ak longir a

hong ȩr ngii.
‘If I see Droteo tomorrow, I’ll borrow his book.’

As these sentences show, an emphatic pronoun (rather than a

non-emphatic pronoun) appears after the specifying word ȩr
or the relational word ȩr when an object or a possessor is
pronominalized. Similar to the above sentences is the dialog of
16, which we repeat here for convenience as 40:

(40) A: Kȩ, mla mo mȩrek ȩr a babier?

‘Have you finished (writing) the letter?’

B: Ng diak. Ak dirk mȩluchȩs ȩr ngii.

‘No. I’m still writing it.’

In the above dialog, we can say that the ngii in B’s response re-
sults from applying pronominalization to a second occurrence of
babier ‘letter’; the first occurrence of babier is to be found, of
course, in A’s question. Thus we can see that pronominalization

104
 Palauan Reference Grammar

can apply between two separate sentences, as in the dialog of

40, or between clauses of one and the same sentence, as in 35,
37, and 39 above.

4.9. OBJECT PRONOUNS AND PERFECTIVE VERBS

In the remaining sections of this chapter we will examine two
types of Palauan affix pronouns in detail. A third type—the pos-
sessor suffixes—was discussed extensively in 3.1–3 and needs
no further consideration.
Palauan has a set of affix pronouns which appear as suffixes
on verbs. Since these pronouns always identify the object of the
verli—i.e., the person or thing affected by the action which the
verb designates—they will be called object pronouns. As we
might expect, the object pronouns show person-number distinc-
tions identical to those found among the independent pronouns.
In the list below, the most frequently-occurring form of each
object pronoun is given; the emphatic pronouns (cf. 4.2 above)
are also listed for purposes of comparison:

(41) Object Pronouns Emphatic Pronouns

1st pers sg -ak ngak
2nd pers sg -au kau
3rd pers sg -ii ngii
1st pers pl incl -id kid
1st pers pl excl -ȩmam kȩmam
2nd pers pl -ȩmiu kȩmiu
3rd pers pl -tȩrir tir

As you can see, the object pronouns and the emphatic pronouns
are very closely related in form: except for the third person (hu-
man) plural, all of the object pronouns can be derived from the
corresponding emphatic pronoun simply by removing the initial
consonant (ng or k) from the latter. Further similarities in form
will be noted below.
For purposes of identification, we will use the term per-
fective verb to refer to any verb form which contains a suffixed
object pronoun. The term “perfective” reflects the meaning of
such forms, since any verb form with a suffixed object pronoun
designates an action which is brought to completion or per-
fection. In this section, we will not be particularly concerned

105
 4 Palauan Pronouns

with the meaning of perfective verbs, nor with the contrast be-
tween perfective and imperfective verbs; these topics will be
taken up in chap. 12. Rather, we will focus most of our attention
on the way in which object pronouns attach to verb stems to
form perfective verbs.
In the list below the perfective forms of the Palauan verb
mȩngȩlebȩd ‘hit’ are given. Each perfective form consists of the
verb stem cholȩbȩd- followed by one of the object pronouns
listed in 41. The object pronouns are always stressed (cf. 1.4.2).

(42) cholȩbȩdák ‘hits me’

cholȩbȩdáu ‘hits you (sg.)’
cholȩbȩdii ‘hits him/her/it’
cholȩbȩdíd ‘hits us (incl.)’
cholȩbȩdȩmám ‘hits us (excl.)’
cholȩbȩdȩmíu ‘hits you (pl.)’
cholȩbȩdȩtȩrír ‘hits them’

You will notice that the verb stem cholȩbȩd- which occurs in the
perfective forms of 42 is phonetically quite different from the
related (imperfective) verb mȩngȩlebȩd ‘hit’. Some of the pho-
netic changes observed between these two forms—for example,
the appearance of the vowel o in the first syllable of the per-
fective verbs, or the alternation between the consonants ch and
ng—require lengthy explanation and will not be discussed until
5.5 and 6.3.1–2. Other changes, however, should not be difficult
to understand because they involve the principle of vowel re-
duction explained in 3.4.
Recall that all of the full vowels of Palauan can reduce to
ȩ ([ǝ]=schwa) in unstressed syllables. This is precisely what
happens to the vowel between l and b in the forms under dis-
cussion: in mȩngȩlébȩd [mǝŋǝlέbǝð], we have the stressed full
vowel e [ε], but in cholȩbȩdíi [ɂolǝbǝðíy] this very same vowel
has reduced to ȩ [ǝ] because it is no longer in a stressed syl-
lable. The only remaining phonetic detail in the forms of 42
which we need to mention is the extra ȩ between the verb stem
and the third person plural object pronoun -tȩrir: this ȩ is added
to break up an unpronounceable sequence of consonants (cf.
1.4.5).

106
 Palauan Reference Grammar

4.9.1. THE ZERO (Ø) OBJECT PRONOUN

The perfective verb forms listed in 42 are used in sentences like
the following:

(43)a. A sensei a ‘(Watch out!) The teacher is going to hit

cholȩbȩdau! you (by accident)!’/ ‘The teacher will hit
you (if you misbehave).’

b. A John a ‘John is going to hit Toki!’/ ‘John will hit

cholȩbȩdii a Toki (if she misbehaves).’
Toki!

c. A Droteo a ‘Droteo is going to hit the children!’/

cholȩbȩdȩtȩrir a ‘Droteo will hit the children (if they
rȩngalȩk! misbehave).’

As 43b and 43c show, perfective verbs with a third person object
pronoun suffix can be followed by a noun phrase which specif-
ically identifies the object. A principle of “agreement” is in-
volved, since -ii must be followed by a specific object which is
singular (e.g. Toki), while -tȩrir requires a specific object which
is plural (e.g. rȩngalȩk ‘children’). Thus, in 43b and 43c the
object of the sentence is indicated in two places—as a suffix
on the perfective verb and as a noun phrase following the per-
fective verb.
Now, with 43b and 43c, compare the following sentence,
which is similar in pattern:

(44) A ngalȩk a cholebȩd a bilis!

‘The child is going to hit the dogs!’

The form cholebȩd in 44 is also a perfective verb form, but one

which does not seem to have any object pronoun suffix. For this
reason, we did not list this form in 42. In this form, the absence
of any object pronoun suffix results in a particular meaning:
cholebȩ can only be used if the object is non-human plural, as
in 44. Because the form cholebȩd forms a set (or paradigm)
with the other perfective verb forms listed in 42, many linguists
would attempt to assign cholebȩd a structure similar to that of
the other perfective forms, which consist of verb stem + object
pronoun. If cholebȩd is to have this very same structure, it will
be necessary to propose a kind of “phantom” object pronoun for

107
 4 Palauan Pronouns

non-human plural objects, an object pronoun which is not re-

alized as any actual sounds but which nevertheless has “struc-
tural significance” because it contrasts with the other object
pronouns. Using this analysis, linguists would say that there is a
zero (symbol: Ø) object pronoun which is suffixed to perfective
verbs when the object is non-human plural. Thus, all the per-
fective forms of a given verb would be identical in pattern, as
indicated in the abbreviated list below:

(45) Verb stem + Object pronoun

cholȩbȩd + ak ‘hits me’

cholebȩd + Ø 6 ‘hits them (e.g., the dogs)’
cholȩbȩd + ȩ + tȩrir ‘hits them (e.g., the people’)

From the above discussion, we can see that the third person
object pronouns work somewhat differently from the third
person pronouns of other pronoun sets. In 2.4, we observed that
there is only a two-way distinction among third person non-em-
phatic pronouns: tȩ ‘they’ refers exclusively to human plural,
while ng ‘he, she, it, they’ refers to all singulars as well as
to non-human plural. By contrast, the third person object pro-
nouns under consideration in this section show a three-way dis-
tinction: -tȩrir is used exclusively for human plural objects,-Ø
is used for non-human plural objects, and -ii is used for all sin-
gular objects. Note that both the object pronouns and the non-
emphatic pronouns highlight the significance of the distinction
human vs. non-human, since in each set different pronouns
are used for human vs. non-human plural (tȩ vs. ng and -tȩrir vs.
-Ø).

4.9.2. FURTHER EXAMPLES OF PERFECTIVE VERB FORMS

As we will see in 5.3, all Palauan verbs show three major distinc-
tions of tense. Verbs in the present tense usually describe
something that is going on at the time the speaker utters the
sentence, but certain present tense verb forms can also refer to
an imminent event, one which is just about to happen. Verbs
in the past tense describe something which took place in the
past, at a point in time which precedes that of the utterance. Fi-
nally, verbs in the future tense refer to something which will
occur in the future, at some point in time which follows that of
the utterance.

108
 Palauan Reference Grammar

Perfective verb forms of course show the above-mentioned

distinctions of tense. The forms choiȩbȩdak ‘hits me’,
cholȩbȩdau ‘hits you’, etc. listed in 42 and illustrated in the sen-
tences of 43 and 44 are all in the present tense. As the English
equivalents of 43 and 44 indicate, Palauan perfective verbs in
the present tense express events which are imminent or which
appear likely to occur; for this reason, the examples in 43 and
44 are interpreted as warnings or as precautionary sugges-
tions. The past tense forms of the perfective verbs of 42 are
chillȩbȩdak ‘hit me’, chillȩbȩdau ‘hit you’, etc. These simply
refer to past events and have no connotation of warning or pre-
caution, as in the examples below:

(46) a. A sensei a chillȩbȩdak.

‘The teacher hit me.’

b. A ngalȩk a chillebȩd a bilis.

‘The child hit the dogs.’

There are many Palauan verbs whose perfective forms end

in the object pronouns listed in 41. As a further example, let
us observe the (future) perfective forms of olȩkiis ‘wake up
someone/ something (naturally or intentionally)’, 7 which are il-
lustrated in the set of sentences below:

(47) a. A Droteo a mo okisák.

‘Droteo will wake me up.’

b. A Droteo a mo okisáu.
‘Droteo will wake you (sg.) up.’

c. A Droteo a mo okisíi a Toki.

‘Droteo will wake up Toki.’

d. A Droteo a mo okisíd.
‘Droteo will wake us (incl.) up.’

e. A Droteo a mo okisȩmám.
‘Droteo will wake us (excl.) up.’

f. A Droteo a mo okisȩmíu.
‘Droteo will wake you (pl.) up.’

109
 4 Palauan Pronouns

g. A Droteo a mo okistȩrír a rȩngalȩk.

‘Droteo will wake up the children.’

h. A Droteo a mo okíis a bilis.

‘Droteo will wake up the dogs.’

The future perfective forms of olȩkiis given above consist of the

future “marker” mo (which is actually the Palauan verb meaning
‘go’) followed by the present perfective forms okisak, okisau,
etc. Notice that the verb stem oki(i)s- appearing in the per-
fective forms of 47 is phonetically quite different from the re-
lated (imperfective) verb olȩkiis ‘wake up someone/something
(naturally or intentionally).’ In all the perfective forms, the se-
quence -lȩ- of olȩkiis has disappeared; furthermore, we rec-
ognize an alternation between ii and i. This alternation can be
explained by the very same principle of vowel reduction which
we cited in 3.4.2 to account for the alternation between long
and short vowels in nouns and their possessed forms such as
oríik ‘broom’ and orikék ‘my broom’. In other words, the long
vowel ii appears in stressed syllables in olȩkíis (the imperfective
verb) and okíis (the perfective form for third person plural non-
human objects), but reduces to short i when it appears in un-
stressed syllables in okisák, okisáu, etc.
In the discussion above we have seen how the object pro-
nouns of 41 appear as suffixes in the perfective forms of verbs
such as mȩngȩlebȩd ‘hit’ (see 42) and olȩkiis ‘wake up’ (see 47).
Several other verbs whose perfective forms end in the object
pronouns of 41 are mȩsebȩk ‘kick’ (perfective forms: sobȩkak
‘kicks me’, sobȩ kau ‘kicks you’, silȩbȩkii ‘kicked him/her/it’,
silȩbȩkȩtȩrir ‘kicked them’, etc.), mȩngimd ‘cut (someone’s
hair)’ (perfective forms: kimdak ‘cuts my hair’, kimdau ‘cuts
your hair’, kilȩmdii ‘cut his/her hair’, kilȩmdȩtȩrir ‘cut their
hair’, etc.), and olȩchiis ‘chase’ (perfective forms: ochisak
‘chases me’, ochisau ‘chases you’, ulȩchisii ‘chased him/her/it’,
ulȩchistȩrir ‘chased them’, etc.).

4.9.3. VERBS WITH LIMITED PERFECTIVE FORMS

There are many Palauan verbs whose meaning determines that
they can take only inanimate objects—i.e., ones that are non-
human and non-living. For example, the actions described by
verbs like mȩluchȩs ‘write’, mȩngitakl ‘sing’, mȩlasȩch ‘carve’,
mȩngiis ‘dig’, mȩnguiu ‘read’, etc. can only be directed at

110
 Palauan Reference Grammar

things: in other words, human beings (and animals) cannot be

‘written’, ‘sung’, ‘carved’, ‘dug’, or ‘read’. For this reason, the
perfective forms which such verbs can have are limited to those
with the third person singular object pronoun -ii (in which case
reference is made to an inanimate singular object) and to those
with the third person plural non-human object pronoun -Ø (in
which case reference is made to an inanimate plural object).
Thus, the only possible (past) perfective forms for mȩluchȩs
‘write’ and mȩngitakl ‘sing’ are illustrated in the sentences
below:

(48) a. A Droteo a lilȩchȩsii a babier.

‘Droteo wrote the letter.’

b. A Droteo a liluchȩs a babier.

‘Droteo wrote the letters.’

(49) a. A Droteo a chilȩtȩklii a chȩlitakl.

‘Droteo sang the song.’

b. A Droteo a chilitakl a chȩlitakl.

‘Droteo sang the songs.’

Forms like *lilȩchȩsak (‘wrote me’), *chilȩtȩklau (‘sang you’),

etc. do not occur because they do not make any sense.
Some verbs which ordinarily take inanimate objects are oc-
casionally observed to occur with human objects. For example,
the (past) perfective forms of mȩleng ‘borrow’ are usually
limited to those illustrated in the examples below:

(50) a. Ak lilȩngir 8 a mlil a Cisco.

‘I borrowed Cisco’s car.’

b. Ak lileng a mlil a Cisco.

‘I borrowed Cisco’s cars.’

But we sometimes see perfective forms of mȩleng ‘borrow’ with

a suffixed third person plural human object pronoun, as in the
following:

(51) Ng sȩbȩchek ȩl longȩtȩrir a bebil ȩr a rȩchȩdam?

‘Could I borrow (the services of) a few of your men?’

111
 4 Palauan Pronouns

4.9.4. VARIANT FORMS OF THE OBJECT PRONOUNS

In the perfective forms of certain verbs we observe object pro-
noun suffixes which differ from those listed in 41. These variant
forms, which occur much less frequently than those of 41, are
best regarded as exceptional or irregular since there is no ap-
parent way of predicting which verb stems they will attach to.
For this reason, the foreigner learning Palauan has no choice
but to memorize the correct distribution of these variant forms.
The perfective forms of a few verbs take the object pronouns
given in 41, except that a consonant—ng or k—is inserted be-
tween the verb stem and the object pronoun suffixes. Note,
for example, the (past) perfective forms of olȩkar ‘wake up
someone/something (by creating a disturbance) 9 :

(52) ulȩkȩrngák ‘woke me up’

ulȩkȩrngáu ‘woke you (sg.) up’
ulȩkȩrngíi ‘woke him/her/it up’
ulȩkȩrngíd ‘woke us (incl.) up’
ulȩkȩrngȩmám ‘woke us (excl.) up’
ulȩkȩrngȩmíu ‘woke you (pl.) up’
ulȩkȩrngȩtȩrír ‘woke them (hum.) up’
ulȩkár ‘woke them (non-hum.) up’

Can you explain why olȩkar (the imperfective form) and ulȩkar
(the past perfective form for third person plural non-human
object) have the full vowel a between k and r, while all the other
perfective forms show ȩ in this position? Another verb whose
perfective forms have an ng inserted between the verb stem
and an object pronoun suffix is mȩles ‘slice’. Because this verb
requires non-human objects, the only occurring (present) per-
fective forms are dosȩngii ‘slices it’ and dmes ‘slices them’.
The perfective forms of omes ‘see’, mȩdȩnge ‘know’, oba
‘carry, take’, and mȩlai ‘bring, take’ have a k between the verb
stem and most of the object pronouns. In the list below, the per-
fective forms of these verbs are given in the present tense:

(53) Person and Number omes ‘see’ mȩdȩnge ‘know’

of Object Pronoun

1st pers sg mȩsȩkak mȩdȩngȩlkak

2nd pers sg mȩsȩkau mȩdȩngȩlkau
3rd pers sg mȩsa mȩdȩngȩlii

112
 Palauan Reference Grammar

1st pers pl incl mȩsȩkid mȩdȩngȩlkid

1st pers pl excl mȩsȩkȩmam mȩdȩngȩlkȩmam
2nd pers pl mȩsȩkȩmiu mȩdȩngȩlkȩmiu
3rd pers pl hum mȩs(ȩ)tȩrir mȩdȩngȩltȩrir
3rd pers pl non-hum mes mȩdȩnge

oba ‘carry, take’ mȩlai ‘bring, take’

1st pers sg obȩkak ngoikak

2nd pers sg obȩkau ngoikau
3rd pers sg oba nguu
1st pers pl incl obȩkid ngoikid
1st pers pl excl obȩkȩmam ngoikȩmam
2nd pers pl obȩkȩmiu ngoikȩmiu
3rd pers pl hum obȩ(ti)tȩrir ngoititȩrir
3rd pers pl non-hum olab ngmai

The above paradigms for omes ‘see’, mȩdȩnge ‘know’, oba

‘carry, take’, and mȩlai ‘bring, take’ exhibit the following un-
usual features:

(a) In all paradigms, no k is found in those perfective forms

having a suffixed third person object pronoun.

(b) In the paradigm of omes ‘see’, a ȩ sound is inserted

between the final s of the verb stem and the initial k of the
object pronoun. Insertion of this ȩ is optional before -tȩrir. 10

(c) The third person singular perfective forms mȩsa ‘sees

him/ her/it’ and oba ‘carries him/her/it’ appear to have a
pronoun suffix of the form -a. This form is extremely rare
and is found only in these two verbs and in msa ‘give’ (cf.
note 10).

(d) All of the perfective forms of mȩdȩnge ‘know’ have a final

l added to the verb stem. 11

(e) The third person singular perfective form nguu ‘brings/

takes it’ is difficult to analyze in terms of verb stem and
suffixed object pronoun.

113
 4 Palauan Pronouns

(f) The third person plural human perfective forms obȩ(ti)

tȩrir ‘carries/takes them’ and ngoititȩrir ‘brings/takes them’,
have the extra syllable -ti-. 12 This extra syllable is optional in
the former word, but obligatory in the latter.

The perfective forms of certain verbs have third person

singular object pronouns with the structure vowel + r. Some
commonly-used verbs which take this type of suffix are given in
the list below:

(54)Verb in Perfective Form (in Present Tense) with

Imperfective Third Person Singular Object Pronoun
Form

mȩleng longir ‘borrows it’

‘borrow’
mȩrkui ‘finish rokir ‘finishes it completely’
completely’
olȩngȩseu ngosuir ‘helps him/her’
‘help’
oker ‘ask’ korir ‘asks him/her’
omȩkoad ‘kill’ mȩkodir ‘kills him/her/it’
omech mȩchir ‘connects it’
‘connect’
mȩluk ‘cut into tukur ‘cuts it into pieces’
pieces’
mȩngam ‘break chumur ‘breaks it’
(long object)’
mȩngȩsa chosȩngur ‘makes him busy’
‘occupy, make
busy’
mȩlul ‘burn, durur ‘barbeques it’
barbeque’
omȩkdȩchor mekȩdȩchȩrur ‘builds it’
‘build’
mȩsuk ‘put in’ sukur ‘puts it in’
mȩsib ‘plow’ sibur ‘plows it’
mȩngut ‘wear chutur ‘wears it out’
out, make old’
mȩchar ‘pay mȩchȩrar ‘pays for it’
for, buy’

114
 Palauan Reference Grammar

4.10. HYPOTHETICAL PRONOUNS

Palauan has a special set of pronouns which appear as prefixes
on verbs in a large variety of complicated grammatical con-
structions. Because many of these constructions express hy-
pothetical events or situations—i.e., ones which are not real,
but which are supposed, assumed, or imagined—the term hy-
pothetical has come to be used as an identifying label for the
pronoun prefixes as well as the verb forms to which they are at-
tached. Observe the examples below:

(55) a. A kusuub e ak mo pass ȩr a test.

‘If I study, I’ll pass the test.’

b. A losuub e ng mo pass ȩr a test.

‘If he studies, he’ll pass the test.’

c. A dosuub e kȩdȩ mo pass ȩr a test.

‘If we (incl.) study, we’ll pass the test.’

The sentences of 55 do not describe facts or real happenings

but instead mention hypothetical or imagined situations. In
other words, rather than claiming that anyone is actually
studying now (or that anyone will definitely study in the future),
these sentences propose that if someone studied, then he would
be able to pass the test. In the examples of 55, the italicized
words kusuub ‘(if) I study’, losuub ‘(if) he/she studies’, and
dosuub ‘(if) we (incl.) study’ are all hypothetical verb forms;
they consist of a prefixed hypothetical pronoun (ku- ‘I’, lo-
‘he/she’, and do- ‘we (incl.)’) followed by the verb stem (-suub
‘study’).
Each of the hypothetical pronouns has several variant forms
whose distribution depends on the type of verb to which they
are prefixed. The main purpose of the following sections will
be to summarize these variant forms and describe their distri-
bution; the task of analyzing the many constructions in which
they occur will be postponed until later chapters.

115
 4 Palauan Pronouns

4.10.1. HYPOTHETICAL PRONOUNS WITH IMPERFECTIVE VERBS

In the list below, the present tense hypothetical forms of the
verb mȩlim ‘drink’ are given. The series of hypothetical pro-
nouns (italicized) occurring in these forms is regularly attached
to imperfective verbs.

(56)Person and Number of Hypothetical Form of mȩlim

Hypothetical Pronoun ‘drink’

1st pers sg kulim ‘(if) I drink’

2nd pers sg/pl chomolim ‘(if) you drink’
3rd pers sg/pl lolim ‘(if) he/she/it/
they drink(s)’
1st pers pl incl dolim ‘(if) we (incl.)
drink’
1st pers pl excl kimolim ‘(if) we (excl.)
drink’

As you can immediately see, the hypothetical pronouns show

fewer person-number distinctions than any of the other pronoun
sets. The most striking feature is that there is no singular vs.
plural distinction for the second and third person pronouns:
chomo- ‘you’ can refer to one person or several persons, and lo-
‘he/she/it/they’ serves for any third person whatsoever, whether
singular or plural, human or non-human.
The hypothetical forms of mȩlim ‘drink’ given in 56 above
are derived simply by replacing the verb marker prefix mȩ-
(see 5.4 and chap. 6) by various hypothetical pronouns. In ex-
actly the same way we can derive the hypothetical forms of any
imperfective verb which begins with mȩ-. Note, for example, the
following partial list:

(57)Imperfective Verb in Hypothetical Forms

mȩ-

mȩsilȩk ‘wash’ kusilȩk, chomosilȩk, losilȩk, etc.

mȩnguiu ‘read’ kunguiu, chomonguiu, longuiu, etc.
mȩluchȩs ‘write’ kuluchȩs, chomoluchȩs, loluchȩs,
etc.
mȩngiis ‘dig’ kungiis, chomongiis, longiis, etc.

116
 Palauan Reference Grammar

For those imperfective verbs which begin with o- (which is actu-

ally a variant of the verb marker mȩ- —see 6.1), the initial o- is
dropped before adding the hypothetical pronouns of 56. The hy-
pothetical forms of such verbs are illustrated in the partial list
below:

(58)Imperfective Verb in Hypothetical Forms

o-

omes ‘see’ kumes, chomomes, lomes, etc.

orrengȩs ‘hear’ kurrengȩs, chomorrengȩs, lorrengȩs,
etc.
osiik ‘look for’ kusiik, chomosiik, losiik, etc.

4.10.2. HYPOTHETICAL PRONOUNS AS AGENTS

In all hypothetical verb forms, the prefixed hypothetical
pronoun designates the doer (or agent) of some action (or, as
we will see in 4.10.5 below, the person or thing characterized
by a particular state or condition). Observe the use of the hypo-
thetical verb forms in the examples below:

(59) a. Ng soak a ngȩlȩkek a losuub.

‘I want my child to study.’

b. Ng chȩtik a chomolamȩch a dȩkool.

‘I don’t like you to smoke cigarettes.’

c. A Toki a longȩlebȩd ȩr ngii a Droteo.

‘Toki is being hit by Droteo.’

d. A babier a kuluchȩs ȩr ngii.

‘The letter is being written by me.’

In all of the sentences of 59, the italicized hypothetical pro-

nouns either refer to or identify the agent. In 59a, the prefix lo-
of losuub tells us that it is some third person who is expected
to study, and the specific noun ngȩlȩkek ‘my child’ establishes
this person’s identity. And in 59b, the prefix chomo- of chomo-
lamȩch makes it clear that it is the person being spoken to
who is getting scolded for smoking. Sentences 59c and 59d are
passive sentences (see 5.6 and 19.7) in which the person or
thing being affected by the action (Toki in 59c and babier ‘letter’

117
 4 Palauan Pronouns

in 59d) appears in sentence-initial position and is followed by a

hypothetical verb form which refers to or identifies the agent. In
other words, in 59c lo- of longȩlebȩd tells us that it is some third
person who is hitting Toki, and the specific noun Droteo gives
the person’s exact identity. And in 59d, ku- of kuluchȩs makes
it clear that it is the speaker who is engaging in the activity of
writing the letter.

4.10.3. HYPOTHETICAL PRONOUNS WITH THE PAST TENSE

The hypothetical forms of mȩlim ‘drink’ listed in 56 are in the
present tense. If we compare the past tense hypothetical forms
of this verb, we notice some changes in the prefixed hypo-
thetical pronouns:

(60) 1st pers sg kullim ‘was drunk by me’

2nd pers sg/pl (cho)mullim ‘was drunk by you’
3rd pers sg/pl lullim ‘was drunk by him/her/it’
1st pers pl incl dullim ‘was drunk by us (incl.)’
1st pers pl excl kimullim ‘was drunk by us (excl.)’

The hypothetical verb forms in 60 consist of a prefixed

hypothetical pronoun, followed by the past tense marker -l-
(see 5.3.2), followed by the verb stem -lim. 13 Notice that the
hypothetical pronouns which are o- final in 56 are u- final in
60. 14 In addition, the first syllable of the second person hy-
pothetical pronoun chomu- is often dropped, resulting in mu-
. A similar shortening of chomo- of 56 to mo- is found among
many speakers, especially if particular constructions are in-
volved. Thus, while most speakers use the “full” form chomo- if
the hypothetical verb form is in a question such as

(61) a. Ngara ȩl tȩkoi a chomosuub ȩr a elȩchang?

‘What language are you studying now?’

many of these same speakers prefer the shortened form mo-

if the hypothetical verb form is part of a relative clause (see
chap. 23):

(62) a. A babier ȩl moluchȩs ȩr ngii a mo ȩr a Droteo. 15

‘The letter you’re writing goes to Droteo.’

b. A blai ȩl moruul ȩr ngii a kmal klou.

118
 Palauan Reference Grammar

‘The house you’re building is very large.’

4.10.4. REDUCED VARIANTS OF THE HYPOTHETICAL PRONOUNS

The hypothetical forms of mo ‘go’ and me ‘come’ exhibit yet an-
other set of hypothetical pronoun prefixes, as in the following:

(63) mo ‘go’ me ‘come’

1st pers sg kbo kȩme

2nd pers sg/pl chobo chome
3rd pers sg/pl lȩbo lȩme
1st pers pl incl dȩbo dȩme
1st pers pl excl kibo kime

Notice that the m of mo ‘go’ changes to b in all of the hy-

pothetical forms. 16 A similar change is observed in the hypo-
thetical forms of certain perfective verbs such as msa ‘give’
(kbsa ‘(if) I give (it to) him/her’, lȩbȩskak ‘(if) he gives (it to)
me’, etc.) and mosii ‘shoot him/her/it’ (kbosii ‘(if) I shoot him/
her/it’, lȩbosii ‘(if) he shoots him/her/it’, etc.). The italicized hy-
pothetical pronouns of 63 appear to be reduced forms of those
listed in 56. In other words, the full vowels in ku-, lo- and do- of
56 have reduced to the neutral vowel ȩ (schwa) to give kȩ-, lȩ -,
and dȩ- of 63; and in kbo ‘(if) I go’, the prefix kȩ- has further re-
duced to k- before the verb stem bo ‘go’. In addition, the second
syllable in chomo- and kimo- of 56 has been lost to give cho- and
ki- of 63.
Several sentences containing the hypothetical forms of mo
‘go’ and me ‘come’ are now given (the hypothetical pronouns
are italicized):

(64) a. A kbo ȩr a Guam, e ak mo kie ȩr a blil a Toki.

‘If I go to Guam, I’ll stay at Toki’s house.’

b. A lȩme a Droteo, e ng me kie ȩr a blik.

‘If Droteo comes, he’ll stay at my house.’

c. Ng diak lȩbo ȩr a party.

‘He’s not going to the party.’

d. Ng diak chome ȩr a blik?

‘Aren’t you coming to my house?’

119
 4 Palauan Pronouns

In the pronunciation of 64d, the initial ch of cho- ‘you’ is usually

deleted following the final k of diak ‘isn’t’, which is then pro-
nounced as [g] because of its intervocalic position (cf. 1.3.1).
Therefore, diak chome is pronounced as [ðiagomε].

4.10.5. DISTRIBUTION OF THE REDUCED VARIANTS

The reduced variants of the hypothetical pronouns given in 63
above have a very wide distribution. In the paragraphs below,
we will discuss and illustrate the different environments in
which they are observed to occur.
As opposed to the hypothetical forms of imperfective verbs,
which take the longer variants of the hypothetical pronouns
listed in 56 and 60, the hypothetical forms of perfective verbs
(cf. 4.9 and 4.9.2 above) usually contain the reduced variants
observed in 63. In the examples which follow, the italicized hy-
pothetical pronouns are prefixed to perfective verb forms:

(65)a. A hong a kbilsa a sȩchȩlik.

‘The book was given by me to my friend.’

b. Ng dimlak kbosii a bilis.

‘I didn’t shoot the dog’.

c. A blai ȩl lȩsilsȩbii a buik a blil a Toki.

‘The house which was burned down by the boy is Toki’s
house.’

d. A Droteo a dimlak lȩngai a ilumȩl.

‘Droteo didn’t bring the drinks.’

e. A present a lȩbilskak a Droteo.

‘The present was given to me by Droteo.’

f. A Satsko a dimlak lȩngȩsuir a Toki ȩl rȩmuul a subȩlel.

‘Satsko didn’t help Toki do her homework.’

g. A kall ȩl lȩkila a Droteo a mle bȩdȩrȩchuis.

‘The food which Droteo ate was spoiled.’

h. A hong a chobilskak.
‘The book was given to me by you.’

120
 Palauan Reference Grammar

i. Ng tela ȩl biang a chomȩngilim?

‘How many beers did you drink?’

j. Ng tela ȩl ngikȩl a chomȩkilang?

‘How many fish did you eat?’

In all of the above sentences except 65i and 65j, the italicized
pronouns are reduced variants with which we are already fa-
miliar. In 65i and 65j we encounter still another variant of the
second person hypothetical pronoun: in these examples, chomȩ-
appears to be derived from chomo- by reduction of the full
vowel o to ȩ.
So far we have only looked at the hypothetical forms of verbs
like mȩsuub ‘study’, mȩlim ‘drink’, mȩluchȩs ‘write’, msa ‘give’,
mo ‘go’, me ‘come’, etc., all of which designate actions or ac-
tivities. Verbs which describe states or conditions rather than
actions also have hypothetical forms, as the examples below il-
lustrate:

(66) a. Ng diak ksechȩr.

‘I’m not sick.’

b. A mlik a diak lȩklou.

‘My car isn’t that big/big enough.’

c. A mubi a dimlak lȩmȩkngit.

‘The movie wasn’t bad.’

d. A lȩngar ȩr ngii a ududek, e ak mo ȩr a Guam.

‘If I had the money, I’d go to Guam.’

We can see from the above examples that the reduced variants
of the hypothetical pronouns are used if the verb describes a
state or condition. The italicized hypothetical pronouns in 66
either identify or refer to the person or thing characterized by
the particular state or condition. In 66a, for instance, the prefix
k- makes it clear that it is the speaker of the sentence who is
sick; and in 66c, the prefix lȩ- refers to the third person inan-
imate subject mubi ‘movie.’
It is even possible for nouns to take prefixed hypothetical
pronouns in certain constructions. As the following examples
show, the reduced variants are used in such cases:

121
 4 Palauan Pronouns

(67)a. Ng diak ksensei.

‘I’m not a teacher.’

b. Ak mo olȩngull se ȩl kbo krubak.

‘I’m going to take things easy when I get to be an old
man.’

c. Tȩ me ȩr a blik a {lȩtutau / lȩsuelȩb}.

‘They come to my house in the {morning / afternoon}.’

In 67a and 67b, the hypothetical pronoun k- is prefixed to the

nouns sensei ‘teacher’ and rubak ‘old man’, which identify cate-
gories or types of persons. And in 67c, the hypothetical pronoun
lȩ- precedes the time words tutau ‘morning’ and suelȩb ‘af-
ternoon’. The grammatical constructions found in the sentences
of 67 will be dealt with elsewhere in the text.

4.10.6. HYPOTHETICAL FORMS OF COMPLEX VERB PHRASES

As we will see in 5.3.2.1, 5.3.3, and 19.7.2, Palauan has several
types of complex verb phrases which consist of two, or pos-
sibly three, separate words. Some typical examples of such verb
phrases include mo omes ‘go see, will see’, mo ungil ‘get better’,
me mȩngȩt mokl ‘come (in order to) clean’, mla mo mȩrek
‘has finished’, etc. When a complex verb phrase occurs in a
construction which requires a hypothetical verb form, a hypo-
thetical pronoun is often prefixed to each of its parts. Observe
the following sentences:

(68)a. A mubi ȩl kbo kumes ȩr ngii a mubi ȩr a Dois.

‘The movie which I am going to see is a German movie.’

b. Ngara ȩl mubi a chobo (cho)momes ȩr ngii?

‘What kind of movie are you going to see?’

c. Ng dirkak kbo kmȩrek ȩr a subȩlek.

‘I haven’t finished my homework yet.’

d. A ngȩlȩkek a dirkak lȩbo lungil 17 ȩl smechȩr.

‘My child hasn’t gotten well yet.’

e. A lȩbo lsechȩr 18 a Droteo, e ng diak lȩbo ȩr a skuul.

‘If Droteo gets sick, he won’t go to school.’

122
 Palauan Reference Grammar

f. A babier ȩl lebla lȩbo lȩmȩrek ȩr ngii a Droteo a mo ȩr a

Toki.
‘The letter which Droteo has just finished (writing) goes
to Toki.’

A similar case of multiple occurrence of hypothetical pronouns

is found in 67b, where both the verb bo ‘become’ and the fol-
lowing noun (rubak ‘old man’) are prefixed with k-.
In the hypothetical forms of complex verb phrases such
as those illustrated in 68 above, Palauan speakers often omit
the first occurrence of the hypothetical pronoun. This occurs
most often in rapid, casual speech and when the third person
hypothetical pronoun is involved. Thus, the following sentence
is perfectly acceptable:

(69) A babier a bo loluchȩs ȩr ngii a Satsko.

‘The letter will be written by Satsko.’

Can you speculate why it would make no difference to omit the

prefix lȩ- from bo in the example above?

4.10.7. IMPERATIVE VERB FORMS

As we will see in 19.5, imperative verb forms are used to
express orders or commands. Since commands are always di-
rected at the person or persons being spoken to, they neces-
sarily entail second person pronouns. Therefore, it should not
be surprising that Palauan imperative verb forms are nothing
more than hypothetical verb forms which have a prefixed
second person hypothetical pronoun.
In 4.10.3 above, we saw that the second person hypothetical
pronoun chomo- is shortened to mo- in certain constructions.
This shortened form mo- also appears in the hypothetical forms
of imperfective verbs when they are used as imperatives. Ob-
serve the following sentences:

(70) a. Mosilȩk ȩr a bilem!

‘Wash your clothes!’

b. Monga ȩr a ngikȩl!
‘Eat the fish!’

123
 4 Palauan Pronouns

c. Monguiu ȩr a hong!
‘Read the book!’

In the examples of 70, the prefixed hypothetical pronoun mo-

identifies the agent—i.e., the person who is expected to carry
out the action in question.
If the imperative verb form is perfective, the second person
hypothetical pronoun appears as m-, as in the examples below:

(71) a. Mngilmii a imȩlem!

‘Finish up your drink!’

b. Mlȩchȩsii a babier!
‘Write the letter!’

c. Mkȩtmokl a delmȩrab!
‘Straighten up the rooms!’

d. Mtȩchȩlbȩtȩrir a rȩngalȩk!
‘Bathe the children!’

e. Mchȩlebȩd a bilis!
‘Hit the dogs!’

In all of the imperative verb forms of 71, the prefixed second

person hypothetical pronoun is pronounced as a separate sylla-
ble—i.e., as [ṃ] (cf. 1.3.5).

4.10.8. PROPOSITIVE VERB FORMS

As we will see in 19.6, propositive verb forms are used when
the speaker proposes or suggests that he and the person
spoken to do some action together. As the examples below il-
lustrate, Palauan propositive verb forms are nothing more than
hypothetical verb forms with first person plural inclusive hypo-
thetical pronouns:

(72)a. Dorael!
‘Let’s go!’

b. Domȩngur ȩr tiang!
‘Let’s eat here!’

124
 Palauan Reference Grammar

c. Dȩbo dolim a biang!

‘Let’s go drink a beer!’

d. Dȩkiiȩsii a kliokl e dolȩngull kung.

‘Let’s dig the hole (now) and rest
later.’

Can you explain the distribution of do- vs. dȩ- in the sentences
of 72?

4.10.9. SUMMARY OF HYPOTHETICAL PRONOUNS

The list below summarizes the many variant forms of the hypo-
thetical pronouns which we have had occasion to mention in the
sections above:

(73)Person and Number of Variant Forms

Hypothetical Pronoun

1st pers sg ku-, kȩ-, k-

2nd pers sg/pl chomo-, mo-, chomu-, mu-,
cho-, chomȩ-, m-
3rd pers sg/pl lo-, lu-, lȩ-, l-
1st pers pl incl do-, du-, dȩ-
1st pers pl excl kimo-, kimu-, ki-

125
 5 Palauan Verbs

5PalauanVerbs
5.1. IDENTIFICATION OF PALAUAN VERBS
In 2.2–3 we began the task of identifying the parts of speech
of Palauan by denning nouns in terms of their meaning and
distribution. We observed, among other things, that Palauan
nouns name or make reference to a great variety of persons,
animals, or things and share certain characteristics of distri-
bution such as occurrence in sentence subject and sentence
object positions. In chap.3 we learned more about nouns by ex-
amining the ways in which nouns combine with possessor suf-
fixes to form possessed nouns. In this chapter we will attempt
to define another part of speech— verbs—in a similar manner:
that is, first we will characterize verbs in terms of their meaning
and distribution, and then we will look at the many ways in
which verbs can be formed through the combination of verb
stems and various affixes. The latter type of study, in which we
describe how different morphemes (or meaning-bearing units)
combine with each other in the formation of particular classes
of words, is known as morphology. Because the morphology
of Palauan verbs is extremely complicated, we cannot possibly
provide a complete picture of it in the present chapter; some
of the topics touched upon below will therefore be given fuller
treatment later in the text.

5.1.1. ACTION VERBS: TRANSITIVE AND INTRANSITIVE

Whereas nouns make reference to human beings, concrete

things, abstract ideas, and the like, verbs name actions or
states which involve nouns in various ways. Action verbs de-
scribe actions, activities, or events and can be transitive or in-
transitive. Transitive action verbs name actions which char-
acteristically have a doer and a receiver. As we saw in 2.3, the
doer (or person who performs, carries out, or causes the action
of the sentence) appears before the verb as sentence subject,
while the receiver (the person, animal, or thing affected by the
action of the sentence) follows the verb as sentence object. Be-
cause the action is “transferred”, so to speak, from doer to re-

126
 Palauan Reference Grammar

ceiver, the term ‘transitive’ (derived from a Latin word meaning

‘cross over’) has come to be used. The italicized verbs in the
sentences below are typical transitive action verbs:

(1) a. A ngalȩk a mȩnga ȩr a ngikȩl.

‘The child is eating the fish.’

b. A John a milȩngȩlebȩd ȩr a katuu.

‘John was hitting the cat.’

c. A Toki a mȩngȩtmokl ȩr a delmȩrab.

‘Toki is straightening up the room.’

d. A chad ȩr a Merikel a milkodir a mȩchas.

‘An American killed the old lady.’

e. Ak rirȩngȩsii a chisel a Toki.

‘I heard the news about Toki.’

f. Tȩ chilsbrebȩr a blai.
‘They painted the houses.’

g. A Droteo a milȩngimd ȩr ngak.

‘Droteo was cutting my hair.’

Can you identify the nouns (or pronouns) which function as sen-
tence subject and sentence object in the examples above?
As the examples in 1 show, any action verb is transitive
if it names an action which is done to or directed at some
person, animal, or thing. Some additional Palauan transitive
action verbs are cited (in the imperfective form) in the list
below:

(2) mȩlim ‘drink’ mȩlasȩch ‘carve’

mȩngitakl ‘sing’ mȩruul ‘make, repair’
mȩlamȩch ‘smoke, chew’ mȩleng ‘borrow’
mȩluchȩs ‘write’ mȩnguiu ‘read’
mȩngiis ‘dig’ mȩlesȩb ‘burn’
omes ‘see’ orrengȩs ‘hear’
osiik ‘look for’ orrebȩt ‘drop’

127
 5 Palauan Verbs

As opposed to transitive action verbs, intransitive action

verbs involve only a doer, but no receiver. In other words, these
verbs do not describe actions which can be directed at someone
or something else, but rather actions in which only a doer par-
ticipates. For this reason, sentences with intransitive verbs like
the following never contain object nouns:

(3) a. A ngalȩk a milil ȩr a sers.

‘The child is playing in the garden.’

b. A Droteo a milȩngȩdub ȩr a diong.

‘Droteo was swimming in the stream.’

c. A dȩmak a oureor ȩr a bangk.

‘My father works at the bank.’

d. A Toki a mo ȩr a skuul.
‘Toki is going to school.’

e. A sȩchȩlik a me ȩr a blik.
‘My friend is coming to my house.’

Though the italicized verbs in the examples of 3 obviously refer

to actions or activities, it is not difficult to see that they are ba-
sically very different from those illustrated in the examples of 1.
To repeat what we have said above, transitive action verbs take
an object (i.e., we eat something, we hit someone, we hear
someone or something, etc.), but intransitive action verbs
do not. Thus, we cannot conceive of “swimming someone or
something”, “going someone or something”, and so forth.
Notice that the nouns which follow the intransitive action verbs
of 3 (i.e., sers ‘garden’, diong ‘stream’, etc.) are not objects, but
instead identify the location of an activity (as in 3a-c) or the
goal of some movement (as in 3d-e). In either case we have a re-
lational phrase (cf. 3.8 and see chap. 14) in which the relational
word ȩr ‘in, at, to’ is followed by a noun designating a place or
location.
Whereas intransitive action verbs never take objects, we will
often see sentences containing transitive action verbs in which
the object has been omitted (or deleted). Observe the pairs of
sentences below:

(4) a. Ak milsuub a tȩkoi ȩr a Siabal.

128
 Palauan Reference Grammar

‘I was studying Japanese.’

b. Ak milsuub.
‘I was studying.’

(5) a. A John a mȩngitakl a chȩlitakl ȩr a Ruk.

‘John is singing a Trukese song.’

b. A John a mȩngitakl.
‘John is singing.’

In 4a and 5a above, the presence of the sentence objects tȩkoi ȩ

r a Siabal ‘Japanese (language)’ and chȩlitakl ȩr a Ruk ‘Trukese
song’ shows that mȩsuub ‘study’ and mȩngitakl ‘sing’ are tran-
sitive action verbs. In 4b and 5b, however, these verbs are not
followed by an object; rather, it is merely understood that the
subject of the sentence was studying something or singing
something. Apparently, sentences like 4b and 5b are ac-
ceptable because verbs like mȩsuub ‘study’ and mȩngitakl ‘sing’
designate activities whose possible objects can usually be pre-
dicted (at least in a general way) even if the object noun is
absent. In other words, the object of mȩsuub ‘study’ must be
something which can be studied or learned, and the object of
mȩngitakl ‘sing’ must be something which can be sung. By con-
trast, if a verb designates an activity which can take a large
number of objects, then the object of the sentence cannot be
omitted. This is true for verbs like mȩleng ‘borrow’ and mȩruul
‘make, prepare’, as in the examples below:

(6) a. Ak mo mȩleng ȩr a mlil a Cisco.

‘I’m going to borrow Cisco’s car.’

b. *Ak mo mȩleng.
(? ‘I’m going to borrow.’)

(7) a. A Maria a mȩruul a kall.

‘Maria is making the food.’

b. *A Maria a mȩruul.
(? ‘Maria is making.’)

129
 5 Palauan Verbs

To Palauan speakers, 6b and 7b above are unacceptable be-

cause they sound incomplete—that is, they do not allow us to
predict anything about the omitted object.

5.1.2. STATE VERBS

Whereas action verbs describe actions, activities, or events,
state verbs specify states, conditions, or qualities which tem-
porarily or permanently characterize persons, animals, or
things. In the great majority of cases, state verbs describe the
sentence subject in some way. Because only a sentence subject
is involved, these state verbs are to be classified as intransitive.
Observe the sentences below, in which the intransitive state
verbs are italicized:

(8) a. A ngȩlȩkek a smechȩr.

‘My child is sick.’

b. A eangȩd a mȩkngit.
‘The weather is poor.’

c. Tia ȩl delmȩrab a mȩkeald.

‘This room is warm.’

d. A rȩchad ȩr a Merikel a mȩtongakl.

‘Americans are tall.’

e. A blil a Toki a klou.

‘Toki’s house is big.’

In 8a–c, the states which describe the sentence subjects are

temporary and therefore subject to change, while those in 8d–e
are relatively permanent or unchanging.
Some additional Palauan state verbs are given in the list
below. See if you can decide whether the state involved is tem-
porary or permanent.

(9) ungil ‘good’ klȩbokȩl ‘pretty’

songȩrengȩr ‘hungry’ kekȩdeb ‘short’
kikiongȩl ‘dirty’ meteet ‘rich’
mȩsaul ‘tired’ milkolk ‘dark’
bibrurȩk ‘yellow’ bȩches ‘new’
bȩcheleleu ‘white’ bȩkȩtȩkoi ‘talkative’

130
 Palauan Reference Grammar

ngar ‘be (located)’ mla ‘was (located)’

The verbs ngar ‘be (located)’ and mla ‘was (located)’ are special
state verbs which assert, respectively, the present or past ex-
istence of the subject in a particular location. As the following
sentences illustrate, these intransitive state verbs must be fol-
lowed by a relational phrase which identifies the location of the
subject:

(10) a. A Toki a ngar ȩr a bab.

‘Toki is upstairs.’

b. A udoud a ngar ȩr a chȩlsel a skidas.

‘The money is inside the drawer.’

c. Ak mla ȩr a M-Dock.
‘I was/have been at M-Dock.’

The existential state verbs ngar ‘be (located)’ and mla ‘was
(located)’ will be examined in detail in 18.2.
As we mentioned above, most Palauan state verbs are in-
transitive. There is a very small class of state verbs, however,
which do take objects and must therefore be classified as tran-
sitive. Such transitive state verbs refer to certain types of
mental states or abilities and include items like mȩdȩnge
‘know’, mȩduch ‘know how (to), be skilled at’, mȩtitur ‘not know
how (to), not be capable of’, and mȩdakt ‘be afraid of, fear’. Can
you identify the subjects and objects of the (italicized) transitive
state verbs in the sentences below?

(11) a. A Droteo a mȩdȩnge a tȩkoi ȩr a Sina.

‘Droteo knows Chinese.’

b. A ngȩlȩkek a mȩdakt ȩr a sensei.

‘My child is afraid of the teacher.’

c. Ak mȩduch ȩr a ochur.
‘I know (how to do) math.’

131
 5 Palauan Verbs

5.1.3. FURTHER DIFFERENCES BETWEEN ACTION VERBS AND STATE

VERBS
In 5.1.1–2 above we attempted to define action verbs and state
verbs in terms of a fundamental meaning difference. Thus, we
said that while action verbs designate actions, activities, or
events, state verbs specify states, conditions, or qualities. In ad-
dition, we saw that both action verbs and state verbs can be
transitive or intransitive, resulting in the following four possibil-
ities:

(12)Transitive action mȩnga ‘eat’, mȩngȩlebȩd ‘hit’, etc.

verb:
Intransitive action milil ‘play’, oureor ‘work’, mo ‘go’, etc.
verb:
Transitive state mȩdȩnge ‘know’, mȩdakt ‘be afraid of,
verb: fear’, etc.
Intransitive state ungil ‘good’, smechȩr ‘sick’, ngar ‘is
verb: (located)’, etc.

When we look at some of the grammatical properties of

action verbs and state verbs, we find further support (or evi-
dence) for distinguishing the two types. Action verbs and state
verbs “behave” differently in at least two important ways. First
of all, the past tense forms of action verbs and state verbs are
derived differently. The past tense forms of action verbs involve
the addition of an affix of some kind. In 4.1 we saw that af-
fixes are morphemes which cannot occur as independent words
but which must always be attached to some other word. The
two types of affixes we mentioned were prefixes and suffixes,
which are added to the beginning or end of a word, respectively.
In discussing the past tense forms of action verbs, we need to
speak of a third type of affix—namely, infixes. Infixes are mor-
phemes which are inserted into a word. Thus, for any transitive
or intransitive action verb which begins with the verb marker
prefix mȩ- or m- (see 5.4 below and chap. 6), the past tense is
derived by infixing the past tense marker -il- or -l- after the m
of the verb marker. In the past tense forms given below, the in-
fixed past tense marker has been italicized:

(13) Action Verb—Present Tense Past Tense Form

mȩnga ‘eat’ milȩnga ‘ate’

mȩngȩlebȩd ‘hit’ milȩngȩlebȩd ‘hit’

132
 Palauan Reference Grammar

mȩlim ‘drink’ millim ‘drank’

mȩluchȩs ‘write’ milluchȩs ‘wrote’
milil ‘play’ mililil ‘played’
mȩngȩdub ‘swim’ milȩngȩdub ‘swam’
mo ‘go’ mlo ‘went’
me ‘come’ mle ‘came’

While action verbs which begin with the verb marker prefix
mȩ- or m- take the infix -il- or -l- for the past tense, those which
begin with the verb marker prefix o- derive the past tense by
replacing this initial o- with ul- or ulȩ-. Thus, we have pairs like
omes ‘see’—ulȩmes ‘saw’, osiik ‘look for’—ulsiik ‘looked for’,
oker ‘ask’—ulȩker ‘asked’, oureor ‘work’—ulureor ‘worked’, etc.
As we have seen above, the past tense forms of action verbs
are derived by using an infix or a prefix. By contrast, the past
tense forms of state verbs (whether transitive or intransitive)
are derived with the auxiliary (or “helping”) word mle ‘was,
were’. 1 This auxiliary word is simply placed directly before the
state verb, as in the following:

(14)State Verb Past Tense Form

mȩkngit ‘bad’ mle mȩkngit ‘was/were bad’

mȩtongakl ‘tall’ mle mȩtongakl ‘was/were tall’
bȩches ‘new’ mle bȩches ‘was/were new’
mȩdȩnge ‘know’ mle mȩdȩnge ‘knew’
mȩdakt ‘be afraid of, mle mȩdakt ‘was/were afraid of,
fear’ feared’

As the examples of 13 and 14 show, a major difference in the

way past tenses of verbs are formed can be explained according
to the basic distinction between action verbs and state verbs.
In other words, the validity of this distinction is upheld because
from it we can predict a grammatical phenomenon which would
otherwise be inexplicable—namely, the fact that certain verbs
have infixed -il- or -l- for the past tense while others must be
preceded by the auxiliary word mle ‘was, were’.
A second way in which action verbs and state verbs behave
differently is with respect to the use of mo ‘go’ as an auxiliary.
Both types of verbs can be preceded by mo ‘go’ (or its past

133
 5 Palauan Verbs

tense form mlo ‘went’), but the resulting meanings are totally
different. In 15 below, mo ‘go’ precedes action verbs, while in
16, it precedes state verbs:

(15) a. Aki mlo milil ȩr a blil a Toki.

‘We went to play at Toki’s house.’

b. A Toki a mlo mȩngȩtmokl ȩr a blik.

‘Toki went to clean my house.’

c. Ak mo mȩsuub ȩr a klukuk.
‘I’m going to study tomorrow.’

(16) a. A Toki a mlo smechȩr.

‘Toki got sick.’

b. A delmȩrab a mlo kikiongȩl.

‘The room got dirty.’

c. A John a mla mo mȩduch ȩr a ochur.

‘John has gotten proficient in math.’

As we will see in chap. 13, mo can be used with action verbs

to express two types of meanings: in 15a–b, the sequences mlo
milil ‘went to play’ and mlo mȩngȩtmokl ‘went to clean’ simply
express the fact that the subject went somewhere to perform
a particular activity, while in 15c, mo mȩsuub ‘will study’ de-
scribes a future event. On the other hand, when mo is used
with state verbs, neither of the two abovementioned meanings
applies; instead, the resulting sequences describe a change of
state. Thus, mlo smechȩr ‘got sick’ and mlo kikiongȩl ‘got dirty’
of 16a–b describe states or conditions which developed or came
into existence and which represent a change from the previous
circumstances (i.e., Toki’s being in good health, and the room’s
being clean). And in 16c mla mo mȩduch ‘has become skilled
at’ refers to a recent change of state and implies a contrast be-
tween the present state (John’s being good at math) and some
earlier, opposing state (John’s being poor in math).
Though we will examine sentences like 15–16 more carefully
in later chapters, we can understand enough about them to see
that, in this case, too, the distinction between action verbs and
state verbs is essential to our description of Palauan grammar.

134
 Palauan Reference Grammar

In other words, it is only on the basis of this distinction that we

can predict the different meanings given to the auxiliary word
mo in the examples of 15 vs. 16.

5.2. DISTRIBUTION OF PALAUAN VERBS

In a preliminary and much oversimplified discussion of the Pa-
lauan word a in 2.6, we observed that the major function of this
word is merely to “introduce” (i.e., precede) certain Palauan
parts of speech—specifically, nouns and verbs. Later (cf. 3.6),
we modified our analysis of a by showing that this word actually
introduces noun phrases, which can consist of single nouns as
well as groups of associated words such as blil a Toki ‘Toki’s
house’, John ȩl sensei ‘John the teacher’, etc. In this section, we
will take a similar approach in modifying our analysis of a as it
relates to verbs: in other words, we will claim that a does not
really introduce single verbs, but rather certain groups of asso-
ciated words called verb phrases.
Although verb phrases can consist of single verb forms such
as mȩnga ‘eat’, mililil ‘played’, etc., they often involve a se-
quence of words such as mle ungil ‘was good’, mlo smechȩr ‘got
sick’, mo mȩruul ‘will make/prepare’, mla mo mȩrek ‘has fin-
ished’, etc. In these examples, various kinds of auxiliary words
(cf. 5.1.3 above)—mle, mlo, mo, mla mo—precede state verbs
or action verbs and provide, among other things, information
about the tense (see 5.3 below). Other types of verb phrases
consist of a quali fying word (see 24.6) followed by a verb
form, as in kmal ungil ‘very good’, di mililil ‘just played’, dirk
smechȩr ‘is still sick’, etc.
The sentences of Palauan, like those of all other languages,
are not just “strings” of single words that have been tacked on
to each other in random fashion. Rather, they consist of certain
groups of associated words (i.e., phrases) which are related
to each other in well-defined ways. The three major types of
phrases in Palauan have already been introduced: they include
noun phrases, verb phrases, and relational phrases. These
three kinds of phrases combine in various ways to give the main
sentence types of Palauan. For example, any Palauan sentence
containing a transi tive verb (cf. 5.1.1 above) followed by an
object has the structure subject noun phrase + verb phrase +
object noun phrase, as in the following:

135
 5 Palauan Verbs

(17) Subject Verb Object

Noun Phrase Noun
Phrase Phrase

a. A ngalȩk a a blai. ‘The child burned down

silsȩbii the house.’

b. A ngȩlȩkel a mla a blil a ‘Toki’s child has burned

a Toki sosȩbii Droteo. down Droteo’s house.’

As 17a shows, noun phrases and verb phrases may consist

of single nouns (ngalȩk ‘child’, blai ‘house’) or single verbs (sil-
sȩbii ‘burned it down’), respectively. But they may also consist
of sequences of closely associated words, as in 17b: here,
ngȩlȩkel a Toki ‘Toki’s child’ and blil a Droteo ‘Droteo’s house’
are noun phrases of possession (cf. 3.6), and mla sosȩbii ‘has
burned it down’ is a verb phrase consisting of the auxiliary word
mla (which designates a recent past event—see 5.3.2.1 below)
followed by the perfective verb form sosȩbii ‘bum it down’.
Notice that the subject noun phrase, the verb phrase, and the
object noun phrase in 17a– b are all introduced by the word a.
The sentences of 17 illustrate an important fact about the
distribution of Palauan verbs—or, more properly, verb phrases.
These sentences show that verb phrases containing a transitive
verb follow the subject noun phrase and precede the object
noun phrase (if there is one—cf. the discussion of examples 4–7
in 5.1.1 above). If the verb phrase contains an intransitive verb
(whether an action verb or a state verb), it takes a subject noun
phrase but no object noun phrase, as in the sentences below:

(18) a. A buik a rȩmurt.

‘The boy is running.’

b. A sensei a mei.
‘The teacher is coming.’

c. A bȩchil a Droteo a smechȩr.

‘Droteo’s wife is sick.’

The sentences of 18 can be expanded by placing various

kinds of relational phrases in sentence-final position following
the verb phrase. Compare the following examples:

136
 Palauan Reference Grammar

(19) a. A buik a rȩmurt ȩr a sȩrsel a Droteo.

‘The boy is running in Droteo’s garden.’

b. A sensei a me er a elȩchang.
‘The teacher is coming now.’

c. A bȩchil a Droteo a smechȩr ȩr a tȩretȩr.

‘Droteo’s wife is sick with a cold.’

In the sentences of 19, the italicized relational phrases consist

of the relational word ȩr followed by a noun phrase. As we will
see in chap. 14, relational phrases serve many different func-
tions: in 19a ȩr a sȩrsel a Droteo ‘in Droteo’s garden’ tells us
the location of the activity; in 19b er a elȩchang ‘now, today’
specifies the time of the event; and in 19c ȩr a tȩretȩr ‘with/be-
cause of a cold’ identifies the cause of the subject’s ill health.
Relational phrases can also be added to sentences like 17a–
b, in which a verb phrase containing a transitive verb is fol-
lowed by an object noun phrase; in such cases, the relational
phrase is placed in sentence-final position, after the object noun
phrase. Thus, with 17a, compare the following sentences:

(20) a. A ngalȩk a silsȩbii a blai er a elii.

‘The child burned down the house yesterday.’

b. A Toki a mȩruul a kall ȩr a uum.

‘Toki is making the food in the kitchen.’

c. Ak milleng a udoud ȩr a sȩchȩlik.

‘I borrowed some money from my friend.’

In the sentences of 20, the italicized relational phrases identify

the location, the source, or the time. Can you tell which type
of relational phrase appears in each sentence?
We can summarize the distributional features of Palauan
verb phrases in terms of the following “formulas” for sentence
types:

A. subject noun phrase + transitive verb phrase (+ object

noun phrase) (+ relational phrase).

137
 5 Palauan Verbs

Here, parentheses are used to indicate those elements whose

appearance is optional. Using this formula, we can derive vari-
eties of sentences in which (i) neither of the optional elements
appears, as in 4b and 5b, (ii) both of the optional elements
appear, as in 20a–c, or (iii) one or the other of the optional ele-
ments appears, as in la-g and the following:

(21)A John a mȩsuub er a elȩchang.

‘John is studying now.’

B. subject noun phrase + intransitive verb phrase (+ rela

tional phrase).

If the optional relational phrase is chosen, we have sentences

like 3a–e or 19a–c, but if it is omitted, we simply get sentences
like 8a–e or 18a–c.
In some Palauan sentences we find that the distribution of
the verb phrase with respect to the surrounding noun phrases
does not correspond to that given in the two formulas above.
As we mentioned in 4.7, many Palauan sentences are formed
by a process of subject shifting, in which the subject noun
phrase is moved to the right of the verb phrase; as a result of
this process, a pro nominal trace is left in the original subject
position in the form of a non-emphatic pronoun. Therefore, in
sentences like those below, the verb phrase is followed by the
(shifted) subject noun phrase (italicized) and preceded by a
pronominal trace:

(22) a. Tȩ mla me a rȩsȩchȩlim.

‘Your friends have come.’

b. Ng mȩringȩl a chimal a Toki.

‘Toki’s hand hurts.’

If a sentence like 22b is further modified by proposing the

possessor (cf. 4.7, ex. 28 and see 17.3), then we have a sen-
tence like the following:

(23) A Toki a mȩringȩl a chimal.

‘Toki’s hand hurts.’

138
 Palauan Reference Grammar

Here, the verb phrase mȩringȩl ‘hurts’ is preceded by a noun

phrase indicating the possessor (Toki) and followed by a noun
phrase indicating the thing possessed (chimal ‘her hand’).

5.3. TENSE
As we have seen in many of the examples above, Palauan verb
forms show differences of tense. The tense of a verb specifies
the time of the action or state which the verb designates. There
are three major distinctions of tense in Palauan—present, past,
and future. Each of these will be taken up separately below.

5.3.1. PRESENT TENSE

The major function of present tense verb forms is to describe
actions or states which are in progress at the time the sentence
is uttered. Often, sentences of this kind will contain a temporal
phrase (see 14.6) like er a elȩchang ‘now’, which designates
the present moment. Observe the sentences below:

(24) a. A Toki a mȩsuub er a elȩchang.

‘Toki is studying now.’

b. A ngȩlȩkek a mȩchiuaiu.
‘My child is sleeping.’

c. A sensei ȩr ngak a smechȩr.

‘My teacher is sick.’

d. A tolȩchoi a mȩdakt ȩr a chȩrrodȩch.

‘The baby is afraid of the noise.’

You will notice that the action verbs and state verbs of 24a-d do
not contain any special affixes except the verb marker, which
appears as a prefix mȩ- in mȩsuub ‘study’, mȩchiuaiu ‘sleep’,
and mȩdakt ‘afraid of’ and as an infix -m- in smechȩr ‘sick’ (see
5.4 below and chap. 6). In other words, Palauan has no special
affix to mark the present tense; rather, it is the absence of
such a marker which identifies present tense verb forms. Thus,
Palauan present tense verb forms are unmarked, while the
other tense forms are marked with various affixes or auxiliary
words, as we shall see below.

139
 5 Palauan Verbs

Present tense verb forms are also used in general state-

ments, which express broad generalizations or commonly-ac-
cepted facts about the subject, and in habitual statements,
which express habits or repeated actions which the subject en-
gages in. A few typical examples are given below:

(25) a. A rȩchad ȩr a Sina a mȩnga a bȩras.

‘The Chinese eat rice.’

b. A rȩchad ȩr a Merikel a mȩtongakl.

‘Americans are tall.’

c. A Toki a mo ȩr a skuul ȩl ngar ȩr a mlai.

‘Toki goes to school in a car.’

d. A Satsko a mȩngȩtmokl ȩr a blik ȩr a bek ȩl tutau.

‘Satsko cleans up my house every morning.’

As we saw in 4.9.1–2, Palauan perfective verb forms in

the present tense have a special connotation of warning or
precaution. Additional examples illustrating this point will be
given in 12.2. A similar connotation is observed among the
present tense forms of ergative verbs, which will be discussed
in 5.4 below.

5.3.2. PAST TENSE

Palauan past tense verb forms always describe an event or
state which was in progress in the past, at some point in time
preceding the time of the utterance. As we saw in 5.1.3 above,
the past tense forms of action verbs and state verbs are derived
in an entirely different manner. State verbs in the past tense are
simply preceded by the auxiliary word mle ‘was, were’, as in the
following sentences:

(26) a. Ak mle smechȩr er a elii.

‘I was sick yesterday.’

b. A eolt a kmal mle mȩses ȩr a kȩsus.

‘The wind was very strong last night.’

c. A kall a mle bȩdȩrȩchuis.

‘The food was spoiled.’

140
 Palauan Reference Grammar

d. Ak mle mȩdȩnȩlii a Toki ȩr se ȩr a lȩngalȩk.

‘I knew Toki when she was a child.’

e. Aki mle kaudȩnge ȩr a Guam.

‘We knew each other in Guam.’

As sentences like 26a–b show, verbs in the past tense are often
accompanied by temporal phrases (see 14.6) like er a elii
‘yesterday’, ȩr a kȩsus ‘last night’, etc. which designate some
time point or time period in the past.
The auxiliary word mle is also used with certain action verbs
which have been borrowed into Palauan from foreign languages
such as Japanese or English. Observe the sentences below:

(27) a. A dart ȩl chad a mle sengkio.

‘One hundred people voted.’

b. A sensei a mle harau a blals.

‘The teacher paid the fine.’

c. A bilsȩngek a mle kosio.

‘My boat went out of order.’

d. A Kiyosi a mle {fail / otsir} ȩr a test.

‘Kiyosi failed the test.’

Since borrowed action verbs like sengkio ‘vote’, fail, etc. are to-
tally different in structure from native Palauan action verbs like
mȩnga ‘eat’, mȩsuub ‘study’, etc., it is not at all surprising that
the former do not follow the pattern of the latter (i.e., infixing
-il- or -l-) in deriving the past tense forms (see below).
As we saw in 5.1.3 above, we derive the past tense forms
of transitive and intransitive action verbs by infixing the past
tense marker -il- or -l- after the m- of the verb marker. For most
action verbs, we have two past tense forms, one containing -il-
and the other containing -l-, as in the following:

(28)Action Verb— Present Tense Past Tense Forms

mȩnga ‘eat’ milȩnga, mlȩnga ‘ate’

mȩsuub ‘study’ milsuub, mlsuub ‘studied’

141
 5 Palauan Verbs

mȩchiuaiu ‘sleep’ milȩchiuaiu, mlȩchiuaiu ‘slept’

mȩkȩra ‘do what?’ milȩkȩra, mlȩkȩra ‘did what?’

It is very difficult to determine the exact difference between the

two past tense forms shown for each verb in 28, since many
speakers seem to use the two forms interchangeably. For those
speakers who do use the two forms differently, there is a fairly
subtle distinction in meaning, which we will now attempt to ex-
plain.
Past tense forms with -il- focus on a past action while it was
in progress; often, the particular action is portrayed as going on
or continuing at the moment when some other action or event
occurred. Note the following sentences:

(29) a. Ak milsuub er se ȩr a lȩmad a dengki.

‘I was studying when the electricity went out’.

b. A Toki a milȩchiuaiu er se ȩr a lȩme a Droteo.

‘Toki was sleeping when Droteo came.’

c. A Droteo a milȩngȩdub er se ȩr a kbong.

‘Droteo was swimming when I arrived.’

In the sentences of 29, the activities of studying, sleeping,

and swimming were going on when some other (possibly in-
terrupting) event took place; this latter event is introduced by
er se(ȩr a) ‘when’ (which is followed by a hypothetical verb
form—see 22.2).
By contrast, past tense forms with -l- do not focus on an
event as it was in progress; rather, they seem to view a com-
pleted action or event as something which more or less fully
occupied a particular period of time. For example, if someone
asked

(30) Kȩ mlȩkȩra er a elii?

‘What did you do yesterday?’

it would be appropriate to give answers like the following:

(31) a. Ak mlsuub.
‘I studied.’

142
 Palauan Reference Grammar

b. Ak mlȩchiuaiu e le ak mle smechȩr.

‘I stayed in bed because I was sick.’

In 31a–b, the past tense forms with -l- imply that studying or
staying in bed was essentially all that the subject (ak ‘I’) did
during the particular period of time designated by er a elii ‘yes-
terday’.
Because -il- and -l- look at past events from different view-
points, they cannot substitute for each other in certain environ-
ments. Thus, the sentences of 29 sound extremely strange if
we replace milsuub ‘was studying’ by mlsuub ‘studied’, etc. In
other words, a sentence like the following (cf. 29a):

(32) *Ak mlsuub er se ȩr a lȩmad a dengki.

(? ‘I studied when the electricity went out.’)

makes no sense because it would be impossible for the subject

(ak ‘I’) to have spent a lot of time studying (which is what
mlsuub implies) at the very moment the electricity went out!
So far, we have only looked at the past tense forms of state
verbs and of action verbs which contain the verb marker prefix
mȩ-. There are, of course, many other classes of verbs, and most
of these have past tense forms, as described in the paragraphs
below.
a. As we saw in 5.1.3 above, some Palauan action verbs
begin with o-, which is one form of the verb marker prefix (see
5.4 below and chap. 6). For such verbs, the past tense forms are
derived simply by replacing the initial o- with ul(ȩ)- 2 , as in the
examples below:

(33) Action Verb—Present Tense Past Tense Form

osiik ‘look for’ ulsiik ‘looked for’

oker ‘ask’ ulȩker ‘asked’
omes ‘see’ ulȩmes ‘saw’
orrengȩs ‘hear’ ulȩrrengȩs ‘heard’
omuchȩl ‘begin’ ulȩmuchȩl ‘began’
okiu ‘go by way of’ ulȩkiu ‘went by way of’
obes ‘forget’ ulȩbes ‘forgot’
ousbech ‘need’ ulusbech ‘needed’
oureor ‘work’ ulureor ‘worked’

143
 5 Palauan Verbs

Causative verbs (see chap. 9) also begin with o-, but this o-
is part of the causative prefixes omȩ(k)-, ol(ȩ)- and or-. 3 As in
the examples of 33, the past tense forms of causative verbs are
derived by replacing the initial o- with ul(ȩ)-:

(34)Causative Verb (in imperfective Past Tense Form

form)
omȩkdȩchor ‘make…stand’ ulȩmȩkdȩchor
‘made…stand’
omȩngamȩch ‘make…smoke’ ulȩmȩngamȩch
‘made…smoke’
omȩngim ‘give drink to’ ulȩmȩngim ‘gave drink to’
olȩkar ‘wake up’ ullȩkar ‘woke up’
olȩchiis ‘chase’ ullȩchiis ‘chased’
ollangȩl ‘make…cry’ ulȩllanȩl ‘made…cry’
orrebȩt ‘drop’ ulȩrrebȩt ‘dropped’

In addition to the present and past imperfective forms given

in 34, causative verbs also have present and past perfective
forms. Because the structure of these forms is very complicated,
we will postpone further discussion until 9.4.

b. As we will see in 6.2, there are many Palauan intransitive

action verbs which do not begin with the verb marker mȩ- but
instead contain an infixed verb marker of the form -(ȩ)m- or -u-.
To derive the past tense forms of such verbs, we simply replace
the infixed verb marker (italicized in the examples below) with
the past tense marker -il- or -ir-:

(35) Intransitive Action Verb— Past Tense Form

Present Tense

lmuut ‘return’ liluut ‘returned’

suebȩk ‘fly’ silebȩk ‘flew’
tuobȩd ‘come out’ tilobȩd ‘came out’
chȩmiis ‘escape’ chiliis ‘escaped’
ruebȩt ‘fall’ rirebȩt ‘fell’
rȩmurt ‘run’ rirurt ‘ran’
rȩmos ‘drown’ riros ‘drowned’

144
 Palauan Reference Grammar

In the last three examples of 35, the past tense marker appears
as -ir- instead of -il-. If we consider -il- to be the basic form of
the past tense marker, we can explain the change of l to r as
an instance of complete assimilation: in other words, the l of
-il- assimilates completely to (i.e., becomes identical with) the
preceding word-initial r.
In discussing the past tense forms of transitive action verbs,
we have so far only considered the imperfective forms. Now
we will look briefly at the perfective forms of such verbs, which
can also appear in the past tense. Observe the following ex-
amples:

(36) Transitive Action Verb— Past Tense Form

Perfective Form,
Present Tense

sosȩbii ‘burn it’ silsȩbii ‘burned it’

suesȩb ‘burn them’ silesȩb ‘burned them’
kolii ‘eat it up’ killii ‘ate it up’
kma ‘eat them up’ kila ‘ate them up’
rullii ‘make it’ rirȩllii ‘made it’
rȩmuul ‘make them’ riruul ‘made them’

As you can see, the examples of 36 involve exactly the same

processes which we observed in 35 above. Thus, in the past
tense forms, the past tense marker -il- or -ir- has replaced the
(italicized) infixed verb marker of the present tense forms. Fur-
thermore, the appearance of -ir- in the last two forms is the
result of assimilation. Because the morphology of Palauan per-
fective verb forms is extremely complex, we cannot pursue it
further at this point. A more detailed discussion will be given in
6.3, 6.3.1–3, and 6.4–7.
The above-mentioned assimilation of l to r is also observed
in the past tense forms of action verbs in which the verb marker
prefix mȩ- is followed by a verb stem (see 5.4 below) which
begins with r. Thus, we have forms like mȩruul ‘make’—mirruul
‘made’, mȩrasm ‘sew’—mirrasm ‘sewed’, and mȩrael
‘leave’—mirrael ‘left’.

145
 5 Palauan Verbs

c. The past tense forms of ergative verbs (see 5.4 below)

are derived by infixing the past tense markers -il- or -l- after the
m of the verb marker, as in the following:

(37)Ergative Verb Past Tense Form

Form—Present Tense

mȩchuiu ‘get read’ milȩchuiu, mlȩchuiu ‘got read’

mȩchamȩch ‘get chewed milȩchamȩch, mlȩchamȩch ‘got
up’ chewed up’
mȩngim ‘get drunk up’ milȩngim, mlȩngim ‘got drunk
up’

Unlike the examples of 28, the two past tense forms given for
the ergative verbs of 37 show no differences in meaning or use
and are completely interchangeable.

d. As we saw in 4.10.3, the past tense hypothetical forms

of imperfective verbs consist of a prefixed hypothetical pronoun
which ends in u (e.g., lu- ‘he, she, it’, du- ‘we (incl.)’ etc.), fol-
lowed by the past tense marker -l- and the verb stem. 4 Forms of
this kind are given below in the right-hand column:

(38)Hypothetical Verb Past Tense Form

Form—Present Tense

lolim ‘is drunk by him/her/it’ lullim ‘was drunk by him/

her/it’
dosilȩk ‘is washed by us (incl.)’ dulsilȩk ‘was washed by us
(incl.)’
mongȩlebȩd ‘is hit by you’ mulȩngȩlebȩd ‘was hit by
you’

5.3.2.1. The Auxiliary mla.

Verb phrases which consist of the auxiliary word mla 5 followed
by the present tense form of any action verb are used to express
two special types of past time. First, in the sentences below, mla
+ action verb refers to an event which happened in the recent
past—that is, at a time point which is not too distant from that
of the utterance itself:

146
 Palauan Reference Grammar

(39) a. A John a mla mȩsuub.

‘John has studied/been studying.’

b. A Droteo a mla mei.

‘Droteo has come’

c. Ng tȩcha a mla kolii a kall?

‘Who has eaten up the food?’

d. Ak mla mo mȩrek ȩr a subȩlek.

‘I’ve finished my homework.’

e. A chȩmȩlek a mla mȩchamȩch.

‘My betel nut has (all) been chewed up.’

If we replace the instances of mla + action verb in 39 with the

corresponding past tense verb forms containing infixed -il- or -l-
(i.e., milsuub/ mlsuub ‘studied’, mlei ‘came’, killii ‘ate it up’, mlo
mȩrek ‘finished’, and milȩchamȩch/ mlȩchamȩch ‘got chewed’),
we get sentences which designate relatively remote past time.
Second, mla + action verb can also refer to past expe-
rience; in such cases, it is used most commonly in questions
which ask whether someone has ever had the experience of
doing something. The following examples are typical:

(40) a. Kȩ mla mo ȩr a chelȩbachȩb?

‘Have you ever gone to the rock islands?’

b. Kȩ mla mȩsuub a tȩkoi ȩr a Siabal?

‘Have you ever studied Japanese?’

c. Kȩ mla mȩlasȩm ȩl mȩnga a sasimi?

‘Have you ever tried eating sashimi?

5.3.3. FUTURE TENSE

The function of the future tense is to designate an action or
a state which will take place in the future, at some point in
time following the time of the utterance. In order to express the
future tense, we use the verb mo ‘go’ as an auxiliary word pre-
ceding any present tense verb form. Verb phrases of the form
mo + verb differ in meaning depending on whether the verb fol-
lowing mo is an action verb or a state verb. Sequences of the

147
 5 Palauan Verbs

form mo + action verb simply designate actions or events which

are expected to take place in the future, as in the sentences
below:

(41) a. Ak mo omes ȩr a John ȩr a klukuk.

‘I’m going to see John tomorrow.’

b. A Toki a mo mȩsuub ȩr a Merikel er tia ȩl me ȩl rak.

‘Toki will study in America next year.’

c. Aki mo mini ȩr a blil a Droteo.

‘We’re going to play at Droteo’s house.’

As sentences like 41a-b show, verb phrases designating the

future tense often are accompanied by temporal phrases (see
14.6) such as ȩr a klukuk ‘tomorrow’ or er tia ȩl me ȩ l rak
‘next year’, which refer to some time point or time period in the
future.
Sequences of the form mo + state verb have a future
meaning but also imply a change of state (cf. our discussion
of the examples of 16 in 5.1.3 above). Note, therefore, the fol-
lowing examples:

(42)a. Ng mo mȩkngit a eangȩd ȩr a klukuk.

‘The weather will get worse tomorrow.’

b. A toktang a dilu ȩl kmo a ngȩlȩkek a mo ungil ȩr a

klukuk.
‘The doctor said my child will get better tomorrow.’

c. A tangk a mo mui ȩr a kȩbȩsȩngei.

‘The tank will get full (by) this evening.’

5.4. THE VERB MARKER AND ERGATIVE VERB

FORMS
In the sections above we have already had occasion to refer to
the Palauan verb marker, which can be found in one form or
another in the great majority of Palauan verbs. It is very difficult
to define or specify the meaning of the verb marker; rather, the
best we can do is to say that the verb marker simply functions
to mark or identify a particular word as a verb.

148
 Palauan Reference Grammar

Though we will examine the verb marker in greater detail in

the next chapter, let us briefly review some of its variant forms
and their distribution. By far, the verb marker appears most
commonly as a prefix mȩ- (sometimes reduced to m-); this prefix
is found on all types of verbs, as shown in the list below:

(43)Transitive mȩnga ‘eat’, mȩlim ‘drink’, mȩsilȩk ‘wash’,

action verb: mȩles ‘cut’, etc.
Intransitive mȩrael ‘walk, travel’, mȩngȩdub ‘swim’,
action verb: milil ‘play’, etc.
Transitive mȩdȩnge ‘know’, mȩduch ‘know how to’,
state verb: mȩtitur ‘not know how to’, etc.
Intransitive mȩkngit ‘bad’, mȩsisiich ‘strong’, mȩsaul
state verb: ‘tired’, mȩsaik ‘lazy’, etc.

Less frequently, the verb marker takes the form of a prefix o-, as
in verbs like osiik ‘look for’, oker ‘ask’, oklukl ‘cough’, and okiu
‘go by way of’
In certain classes of verb forms, the verb marker appears as
an infixed element of the form -(ȩ)m-, -u-, or -o-. Many intran-
sitive verbs are of this type: these may be action verbs, as in
35 above, or state verbs like smechȩr ‘sick’, dmak ‘together’, or
kmeed ‘near’. In addition, the present tense perfective forms of
most transitive action verbs have an infixed verb marker; some
typical examples were given in 36 above.
The simplest Palauan verbs are those which consist of just a
single morpheme or meaning-bearing unit; these include a rela-
tively small number of state verbs such as klou ‘big’, dȩngchokl
‘sitting’, ungil ‘good’, cheisȩch ‘stained’, and ngar ‘is (located)’.
The overwhelming majority of Palauan verbs, however, are more
complex in structure than the state verbs we just mentioned and
contain anywhere from two to four morphemes. In this and the
following sections we will survey the structure of verb forms
containing two or more morphemes, beginning with the easier
types and moving on to the more complex ones.
We will first examine a group of verb forms which consist
only of the verb marker prefix and a following verb stem. A
verb stem is a morpheme to which one or more affixes are
added in the process of forming a verb. In the examples below,
the verb stem which follows the verb marker prefix mȩ- or o- is
actually a noun which can occur as an independent word:

149
 5 Palauan Verbs

(44)Derived Verb Related Noun

mȩchat ‘be/get smoked (of fish)’ chat ‘smoke’

mȩchȩlebȩd ‘be/get hit’ chȩlebȩd ‘spanking,
whip’
mȩchȩsimȩr ‘be/get closed’ chȩsimȩr ‘door’
mȩchȩsbrebȩr ‘be/get painted’ chȩsbrebȩr ‘paint’
mȩchas ‘be/get burned’ chas ‘soot’
mȩdub ‘be/get poisoned, dub ‘dynamite, poison’
bombed’
mȩdangȩb ‘be/get covered’ dangȩb ‘lid’
mȩtȩkoi ‘be/get talked to’ tȩkoi ‘language, word’
oboes ‘be/get shot’ boes ‘gun’
obail ‘be/get clothed’ bail ‘article of clothing’
obȩkall ‘be/get driven’ bȩkall ‘sail, driving’

The following verbs also consist of the verb marker prefix

followed by a verb stem, but unlike the examples of 44, the verb
stem cannot appear as a separate word:

(45) mȩchuiu ‘be/get read’ mȩdobȩch ‘be/get cut’

mȩka ‘be/get eaten’ mȩluchȩs ‘be/get written’
mȩchitakl ‘be/get sung’ mȩsebȩk ‘be/get kicked’

Verb stems like -chuiu ‘read’, -ka ‘eat’, -luchȩs ‘write’, etc. are
called bound forms because they never occur alone as inde-
pendent words but must always be connected to some other
morpheme (usually an affix of some kind). All affixes, too, are
necessarily bound forms; thus, we will never find affixes like mȩ-
or o- (verb markers) or -ek ‘my’ or -em ‘your’ (possessor suf-
fixes) occurring as separate words.
Before discussing the meaning and use of the verb forms
in 44 and 45, we need to introduce some other verb forms for
comparison. As we will see in 5.5 below, all of the verbs in 44
and 45 have corresponding imperfective forms, which in most
cases can be easily identified because they show a character-
istic change in the verb-stem-initial consonant. A few examples
are given below:

(46) Verb of 44 or 45 Imperfective Form

mȩchat ‘be/get smoked (of fish)’ mȩngat ‘smoke (fish)’

mȩchuiu ‘be/get read’ mȩnguiu ‘read’

150
 Palauan Reference Grammar

mȩka ‘be/get eaten’ mȩnga ‘eat’

mȩdobȩch ‘be/get cut’ mȩlobech ‘cut’
mȩtȩkoi ‘be/get talked to’ mȩlȩkoi ‘talk, speak’
oboes ‘be/get shot’ omoes ‘shoot’
mȩluchȩs ‘be/get written’ mȩluchȩs ‘write’

All of the imperfective verb forms in 46 are transitive

verbs—that is, they take object noun phrases (cf. 5.1.1 above).
By contrast, the corresponding verb forms of 44 and 45 are not
transitive, but instead have a rather unique function: they take
as their subject what would be the object of the related tran-
sitive verb. Observe the following examples:

(47) a. A Droteo a mla mȩngat a ngikȩl.

‘Droteo has smoked the fish.’

b. A ngikȩl a mla mȩchat.

‘The fish has been smoked.’

(48) a. A Toki a mla mȩnga a kall.

‘Toki has eaten the food.’

b. A kall a mla mȩkang.

‘The food has been eaten.’

In comparing the a- and b-sentences of 47 and 48 above, you

can see that the object noun phrases (italicized) of the transitive
a-sentences have become the subject noun phrases of the b-sen-
tences. Furthermore, the subjects of the b-sentences are viewed
as having undergone the effect of the actions designated by mla
mȩchat ‘has been smoked’ and mla mȩkang ‘has been eaten’.
Linguists use the technical term ergative verb to identify the
verb forms listed in 44 and 45 and illustrated in 47b and 48b.
As we have seen above, ergative verb forms differ from the
corresponding imperfective verb forms in meaning and use
and—in most cases—pronunciation.
In English, too, we have pairs of sentences similar to those
of 47 and 48, as the following example illustrates:

(49) a. This key opens my office door.

b. My office door opens with this key.

151
 5 Palauan Verbs

While 49a is a transitive sentence in which my office door is

the object of the verb open, in 49b my office door appears as
the subject of open and designates the thing which undergoes
the action of opening. Thus, 49b appears to be rather similar
to the Palauan ergative sentences given in 47b and 48b. Note,
however, that while the same verb form—opens—is found in
both of the English sentences, the Palauan sentences in 47 and
48 show different but related verb forms—i.e., ergative mȩchat
‘be/get smoked’ vs. imperfective (transitive) mȩngat ‘smoke’,
etc. Though sentences like 49b are not very common in English,
Palauan sentences like 47b and 48b are found quite frequently.
Ergative verb forms in the present tense have a special
interpretation: they are used as warnings or as suggestions to
take precautions against some expected future event. (Do you
recall any other Palauan verb forms which are used in the same
way?). Note, therefore, the following sentences, in which the
present tense ergative forms have been italicized:

(50) a. Alii, a chimam a{mȩdobȩch / oburȩch}!

‘Watch out, your arm will get {cut / speared}!’

b. Alii, kȩdȩ mȩdul ȩr a ngau!

‘Watch out, we’ll get burned by the fire!’

c. Bart a chȩmȩlek e ng mo ȩl mȩchamȩch.

‘Hide my betel nut, or else it’ll get chewed.’

d. Bart a kȩlek e ng mo ȩl mȩka ȩr a bilis.

‘Hide my food, or else it’ll get eaten by the dog.’

While 50a-b imply imminent danger, 50c-d are less urgent

in tone and suggest that precautions be taken to forestall some
future event which is thought likely to occur. In warning or
precaution sentences like 50a-d (and in ergative sentences in
general, as we will see below), Palauan speakers often omit the
relational phrase which designates the cause or agent respon-
sible for the particular event; and some Palauan speakers even
find that the presence of relational phrases like ȩr a ngau ‘by
the fire’ in 50b and ȩr a bilis ‘by the dog’ in 50d results in rather
awkward sentences.

152
 Palauan Reference Grammar

The above-mentioned facts point to one of the major fea-

tures which distinguish ergative sentences from passive sen-
tences (see 5.6 below): while the agent responsible for the
event is normally expressed in passive sentences, most ergative
sentences do not mention the cause or agent. For this reason,
ergative sentences are used in cases where the cause or agent
is thought to be irrelevant or unimportant. This point is illus-
trated clearly in the dialogs below, in which B’s responses to A’s
questions contain ergative verbs in various past tenses:

(51) A: Ng dirk ngar ȩr ngii a biang?

‘Is there any beer left?’

B: Ng diak. Ng milȩngim er a elii.

‘No. It got drunk up yesterday.’

(52) A: Ng dirk ngar ȩr ngii a kall?

‘Is there any food left?’

B: Ng diak. Ng mla mȩkang.

‘No. It’s been eaten up.’

(53) A: Ng dirk ngar ȩr ngii a hong ȩl kirel ȩl donguiu?

‘Are there still some books we have to read?’

B: Ng diak. Ng mla mȩchuiu ȩl rokui.

‘No. They’ve all been read.’

(54) A: Kȩ mo mȩngȩsbrebȩr ȩr a mlim er oingarang?

‘When are you going to paint your canoe?’

B: Ng diak. 6 Ng mla mȩchȩsbrebȩr.

‘(I don’t need to.) It’s already been painted.’

In the dialogs above, B’s responses do not contain any rela-

tional phrases which identify the cause or agent of the action.
Such phrases are unnecessary because the main purpose of B’s
responses is simply to focus on the past event as it affects the
present situation: in 52, for example, what is important is the
fact that the food was eaten and there is none left now; who
ate the food does not matter. In some cases—e.g., 53—it is clear
from A’s question who the agent of the action is; therefore, there
is no need for B to supply this information in his answer. Notice,

153
 5 Palauan Verbs

further, that while the present tense ergative verb forms in 50

have a connotation of warning or precaution, the various past
tense ergative forms in 51–54 lack this connotation.
As the sentences below illustrate, ergative verbs can be pre-
ceded by the auxiliary mo to indicate future tense:

(55)a. A stoa ȩr a Droteo a mo mȩngai a chȩsmȩrel ȩr a euid ȩl

klok.
‘Droteo’s store will open at seven.’

b. A delmȩrab ȩr ngak a mo mȩchȩsbrebȩr ȩr a klukuk.

‘My room is going to be painted tomorrow.’

c. A bilek a mo mȩsilȩk ȩr a suelȩb.

‘My clothes will get washed this afternoon.’

Unlike the examples of 50, the sentences of 55 (though still

ergative) do not have any sense of warning or precaution. These
sentences simply express a future event with emphasis on the
event itself rather than on the person responsible for the event:
in 55b, for example, our attention is drawn to the fact that the
room will get painted, but it is not important to know by whom.
Because of this difference of emphasis, we cannot at all say that
55a-c and the following “corresponding” transitive sentences
are equivalent in meaning or function:

(56)a. A Droteo a mo mȩlai ȩr a chȩsmȩrel a stoa ȩr ngii ȩr a

euid ȩl klok.
‘Droteo is going to open his store at seven.’

b. A sȩchȩlik a mo mȩngȩsbrebȩr ȩr a delmȩrab ȩr ngak ȩr

a klukuk.
‘A friend of mine is going to paint my room tomorrow.’

c. A Toki a mo mȩsilȩk a bilek ȩr a suelȩb.

‘Toki is going to wash my clothes this afternoon.’

154
 Palauan Reference Grammar

5.5. IMPERFECTIVE VS. PERFECTIVE VERB

FORMS
In 5.4 above, we remarked that every Palauan ergative verb
has a corresponding imperfective form, which can usually be
identified by a characteristic change in the verb-stem-initial
consonant. As we will see below, a detailed analysis of the
structure of imperfective verb forms will require us to modify
this statement considerably. In the following paragraphs we will
explain the structure of both imperfective and perfective verb
forms in some detail; then we will sketch briefly how these two
verb types differ from each other in meaning and use.
All Palauan imperfective verb forms consist of three
morphemes and have the structure verb marker + imperfective
marker + verb stem. As you can see, this structure is more
complex than that of ergative verb forms, which consist of the
two-morpheme sequence verb marker + verb stem. Ergative
verb forms are more “basic” than imperfective verb forms in
the sense that the latter can be derived from the former simply
by putting the imperfective marker between the verb marker
and the verb stem (see the example below). The imperfective
marker has several variants—i.e., l, ng, or m—depending on
the initial consonant of the following verb stem. Furthermore,
once the correct variant of the imperfective marker has been
determined, the initial consonant of the following verb stem is
deleted. To take a simple example, we have the ergative verb
form mȩchuiu ‘be/get read’, which consists of the verb marker
mȩ- and the (bound) verb stem -chuiu ‘read’. To derive the
corresponding imperfective verb form, we need to “fill in” the
formula mȩ + imperfective marker + chuiu with the appro-
priate form of the imperfective marker. As we will see below, the
imperfective marker appears as ng if the following verb-stem-
initial consonant is ch. Thus, we get the sequence mȩ + ng +
chuiu, which, after deletion of the verb-stem-initial consonant
gives us the correct imperfective form mȩnguiu ‘read’. 7
In the light of the above discussion, we can see that it was
incorrect to say that a given ergative verb form and its corre-
sponding imperfective verb from differ with respect to the verb-
stem-initial consonant. In other words, imperfective mȩnguiu
‘read’ is not really derived from ergative mȩchuiu ‘be/get read’
simply by replacing ch with ng; what actually happens is that ng
(a variant of the imperfective marker) is added before the verb

155
 5 Palauan Verbs

stem, whose initial consonant is then deleted. Thus, in mȩnguiu

‘read’, the ng is not part of the verb stem; rather, it is a separate
morpheme—namely, the imperfective marker.
Let us now examine the distribution of the three variants of
the imperfective marker:
a. The imperfective marker appears as -l- before verb stems
which begin with the dental stops t or d, the alveolar fricative s,
the liquid l, or the velar nasal ng. Some imperfective verb forms
containing this variant of the imperfective marker are derived
as follows:

(57)Verb +Imperfective+Verb Delete Verb- Imperfective

Marker Marker Stem Stem -Initial Form
Consonant→

mȩ + 1 +tiud ‘way → mȩliud ‘cut

of (round
cutting’ object)’
mȩ + 1 +tȩkoi → mȩlȩkoi
‘word, ‘talk, speak’
language’
mȩ + 1 +dasȩch → mȩlasȩch
‘carving’ ‘carve’
mȩ + 1 +deel ‘nail’→ mȩleel ‘nail’
mȩ + 1 +sesȩb → mȩlesȩb
‘fire’ ‘burn’
mȩ + 1 +sȩkosȩk → mȩlȩkosȩk
‘cut (meat)’
mȩ + 1 +leng → mȩleng
‘borrow’
mȩ + 1 +luchȩs → mȩluchȩs
‘write’
mȩ + 1 +ngatȩch → mȩlatȩch
‘way of ‘clean’
cleaning’
mȩ + 1 +ngim → mȩlim
‘drink’
mȩ + 1 +ngukȩd → mȩlukȩd
‘fine’ ‘pay a fine’

While most of the verb stems in 57 can be used independently

as nouns, as the English glosses indicate, some stems are
bound—e.g. -sȩkosȩk, -leng, etc. Can you describe the meaning

156
 Palauan Reference Grammar

relationship between pairs of words such as tȩkoi ‘word, lan-

guage’— mȩlȩkoi ‘talk, speak’, ngukȩd ‘fine’—mȩlukȩd ‘pay a
fine’, etc.?
Because of the deletion of the verb-stem-initial consonant,
most of the imperfective verb forms of 57 are different from the
corresponding ergative forms—thus, we have mȩliud ‘cut’ vs.
mȩtiud ‘be/get cut’, mȩlesȩb ‘burn’ vs. mȩsesȩb ‘be/get burned’,
etc. If the verb-stem-initial consonant is l, however, as in the
case of mȩleng ‘borrow’ and mȩluchȩs ‘write’, then the corre-
sponding ergative form is identical. Can you explain why this is
so?
There are a few s-initial verb stems in which the s is not
deleted following the imperfective marker -l-; instead, -l- is
deleted and s remains. This small group of exceptions includes
mȩsilȩk ‘wash’, mȩsebȩk ‘kick’, and mȩsuub ‘study’. Can you
see why for these verbs, too, the imperfective and ergative
forms are identical?

b. The imperfective marker occurs as -ng- before verb stems

which begin with the velar stop k or the glottal stop ch. Some
imperfective verb forms containing this variant of the imperfec-
tive marker are now derived as in 57 above:

(58)Verb +Imperfective+Verb Delete Verb- Imperfective

Marker Marker Stem Stem -Initial Form
Consonant→

mȩ +ng +ka → mȩnga ‘eat’

mȩ +ng +kȩreel → mȩngȩreel
‘line for ‘catch fish
fishing’ (with a line)’
mȩ +ng +kiis → mȩngiis ‘dig’
mȩ +ng +chat → mȩngat
‘smoke’ ‘smoke
(fish)’
mȩ +ng +chȩsimȩr→ mȩngȩsimȩr
‘door’ ‘close’
mȩ +ng +chitakl → mȩngitakl
‘sing’

157
 5 Palauan Verbs

mȩ +ng +chas → mȩngas

‘soot, ‘paint
ash’ (someone)
with ashes’
mȩ +ng +chaus → mȩngaus
‘lime’ ‘put lime on
(betel nut)’

c. The imperfective marker appears as -m- before verb stems

which begin with the bilabial stop b. With verb stems of this
type, the verb marker prefix takes the form o-, as indicated in
the derivations below:

(59)Verb +Imperfective+Verb Delete Verb- Imperfective

Marker Marker Stem Stem -Initial Form
Consonant→

o +m +boes → omoes
‘gun’ ‘shoot’
o +m +bail → omail
‘article of ‘clothe’
clothing’
o +m +btar → omtar
‘swing’ ‘swing’
o +m +burȩch → omurȩch
‘(action ‘spear’
of)
spearing’

d. Following 54 above, we listed some exceptional cases

in which the imperfective marker is deleted before certain s-
initial verb stems. This same phenomenon is found among two
further groups of verbs. First, we do not find any trace of the
imperfective marker in imperfective verb forms if the verb stem
is r-initial, as in mȩruul ‘make, prepare’ and mȩrasm ‘sew’.
Second, the imperfective marker appears to have been deleted
in the imperfective forms of verbs like oker ‘ask’, in which
the verb marker is o- and the verb stem (ker ‘question’) is an
independently-occurring noun.
Practically all Palauan imperfective verb forms have cor-
responding perfective verb forms, which differ in several im-
portant respects. Let us compare the structures of the two verb
types by using the following formulas:

158
 Palauan Reference Grammar

(60)a. imperfective verb verb marker + imperfective marker +

form: verb stem
b. perfective verb verb marker + verb stem + object
form: pronoun

Since the imperfective marker simply functions to identify a

verb form as being imperfective, it is only natural that this
marker would not be part of the structure of perfective verb
forms. In addition, as we observed in 4.9 and 4.9.1–4, perfective
verb forms always take an object pronoun suffix, whereas im-
perfective verb forms do not. A third difference between the two
types of verbs is that although both contain the verb marker
(indicated as the first element in the formulas above), in per-
fective verb forms this verb marker always shifts to a position
following the initial consonant of the verb stem. As a result,
most perfective verb forms show an infixed verb marker of the
form -(ȩ)m-, -u-, or -o-.
The following examples illustrate some of the differences
discussed above:

(61)Imperfective Perfective Verb Form (with 3rd pers. sg.

Verb Form Object pronoun suffix)

mȩlasȩch dosȩchii ‘carve it’

‘carve’
mȩleel ‘nail’ dmelii ‘nail it’
mȩlesȩb ‘burn’ sosȩbii ‘burn it’
mȩlatȩch ngotȩchii ‘clean it’
‘clean’
mȩnga ‘eat’ kolii ‘eat it’

Whereas the imperfective verb forms all have the prefixed verb
marker mȩ-, the corresponding perfective verb forms show the
verb marker (italicized) following the verb-stem-initial con-
sonant. Like the ergative verb forms given in 44 and 45 above,
the perfective verb forms of 61 preserve the verb-stem-initial
consonant (d, s, ng, k, etc.), while the imperfective verb forms
do not. This is due to the presence or absence of the imper-
fective marker, which, as we have seen, causes (or conditions)
the loss of a following verb-stem-initial consonant. Thus, in im-
perfective verb forms, which contain this marker, the verb-stem-
initial consonant is dropped; in ergative and perfective verb
forms, however, where this marker is absent, the consonant in

159
 5 Palauan Verbs

question is retained. For this reason, we should look at ergative

verb forms or perfective verb forms if we wish to determine the
initial consonant of a verb stem.
The formation of Palauan perfective verbs involves a great
many complexities and will not be pursued further until the next
chapter. Equally complex is the difference in meaning and use
between perfective verbs and imperfective verbs. Though we
will deal with this topic extensively in chap. 12, we will make
a few preliminary observations here. The distinction between
perfec tive vs. imperfective, which is found only among tran-
sitive verbs, is essentially the following: whereas perfective
verb forms designate a totally completed (or “perfected”)
action, imperfective verb forms do not indicate completion but
rather focus on the action as it is (or was) in progress. This dif-
ference can be observed in the pairs of sentences below:

(62) a. A Droteo a milȩnguiu ȩr a hong er a elii.

‘Droteo was reading the book yesterday.’

b. A Droteo a chiliuii a hong er a elii.

‘Droteo read the book yesterday.’

(63) a. Ak milȩngiis ȩr a kliokl.

‘I was digging the hole.’

b. Ak kilisii a kliokl.
‘I (completely) dug the hole.’

In 62a and 63a above, the imperfective verb forms

milȩnguiu ‘was reading’ and milȩngiis ‘was digging’ describe
an action which went on for a period of time but was not com-
pleted: in other words, 62a is understood to mean that Droteo
still has some of the book to read, and 63a implies that the
subject must still do some digging. On the other hand, the per-
fective verb forms chiliuii ‘(completely) read’ and kilisii ‘(com-
pletely) dug’ in 62b and 63b designate an action which was
brought to completion: thus, in 62b Droteo finished reading the
book, and in 63b the subject finished digging the hole. As we
saw in 2.7, and as 62 and 63 above illustrate, any specific
singular object following an imperfective verb form must be
introduced by the specifying word ȩr; occurrence of ȩr is im-
possible, however, after all perfective verb forms.

160
 Palauan Reference Grammar

5.6. ACTIVE AND PASSIVE SENTENCES

All Palauan transitive verbs, whether imperfective or perfective,
can occur in active or passive sentences. The transitive verbs
we have looked at so far have all appeared in active sentences
such as the following:

(64) a. A ngalȩk a menga ȩr a ngikȩl.

‘The child is eating the fish.’

b. A John a milȩngȩlebȩd ȩr a katuu.

‘John was hitting the cat.’

c. A ngalȩk a silsȩbii a blai.

‘The child burned down the house.’

In active sentences, the subject noun phrase corresponds to

the agent—i.e., to the person who carries out or performs the
action designated by the transitive verb. Can you identify the
object noun phrases in 64a-c?
Now, with 64a-c above, compare the following passive sen-
tences:

(65) a. A ngikȩl a longa ȩr ngii a ngalȩk.

‘The fish is being eaten by the child.’

b. A katuu a lulȩngȩlebȩd ȩr ngii a John.

‘The cat was being hit by John.’

c. A blai a lȩsilsȩbii a ngalȩk.

‘The house was burned down by the child.’

If we think of the sentences of 65 as being derived from those

of 64, we can see that in the passive sentences of 65, the object
noun phrases of 64 have come to appear in sentence subject
position; furthermore, in 65, the agent, which was the subject
noun phrase in 64, has moved to sentence-final position. Two
further significant changes can be observed in the passive sen-
tences: first, the transitive verb appears in the hypothetical
form (cf. 4.10 and 4.10.1–9); and second, if the verb is imper-
fective, as in 65 a-b, ȩ r ngii is added after it if the object is sin-
gular.

161
 5 Palauan Verbs

An explanation of the above phenomena is beyond the scope

of our present discussion and will be postponed until 19.7.
Nevertheless, we can see that passive sentences are” charac-
terized by the fact that the noun phrase found in subject po-
sition actually refers to the object of the sentence—i.e., the
person, animal, or thing affected by the action designated by
the transitive verb. As the English translations for 64 and 65
indicate, active and passive sentences do not differ in meaning
from the point of view of what information they convey. For
example, 64c and 65c provide us with exactly the same infor-
mation—namely, that an act of burning occurred at some time
in the past, that the agent (or person responsible) was a child,
and that the thing affected was a house. The only difference
between these two sentences—and between active and passive
sentences in general—is a matter of focus or viewpoint: the
active sentence 64c would be used when the speaker is focusing
his attention on the agent (ngalȩk ‘child’) and the agent’s activ-
ities (or behavior, etc.), while the passive sentence 65c would be
used in a situation where the speaker is more interested in the
object (blai ‘house’) than in the agent.
In discussing ergative sentences in 5.4, we noted that while
ergative sentences containing an agent tend to be avoided by
certain Palauan speakers, passive sentences containing an
agent are perfectly acceptable. This is because passive sen-
tences are used when the speaker feels it necessary to mention
both agent and object, although, as mentioned above, he is fo-
cusing his attention on the latter. On the other hand, ergative
sentences are used when identification of the agent is con-
sidered to be irrelevant or unimportant (cf. 51–55 above). Thus,
with 65c, for example, compare the corresponding ergative sen-
tence below:

(66) A blai a milsesȩb.

‘The house (was) burned down.’

5.7. FURTHER TYPES OF VERBS

In the sections above, we have tried to gain an understanding
of Palauan verbs by examining such major distinctions as tran-
sitive vs. intransitive, active vs. passive vs. ergative, and the
like. However, there still remain many features of Palauan verbs

162
 Palauan Reference Grammar

which need to be explained but which are of such a scope that

they will require separate chapters. These remaining topics are
summarized below:

a. In 4.10 and 4.10.1–9, we looked at hypothetical verb forms

from the point of view of the prefixed hypothetical pronouns,
which designate the agent. Hypothetical verb forms have the
structure hypothetical pronoun + imperfective marker + verb
stem (note the lack of the verb marker in these forms) and are
used in a large variety of grammatical constructions. Such con-
structions will be discussed extensively in chaps. 18 and 19.

b. There are several distinct classes of Palauan state verbs

which require special attention. As we will see in chap. 7, it is
possible to form resulting state verbs and anticipating state
verbs, as well as state verbs with the prefixes bȩkȩ-, sȩkȩ-,
and bȩ-. And in chap. 10, we will consider reciprocal verbs, a
subtype of state verbs which have the reciprocal prefixes kai-,
ka-, cha-, etc. and which designate actions which two persons
do to each other mutually.

c. As we saw in 5.3.2 above, causative verbs are formed with

the causative prefixes omȩ(k)-, ol(ȩ)-, etc. They are transitive
verbs and involve actions in which one person causes or forces
someone else to do something or be in a particular state.
Causative verbs will be dealt with extensively in chap. 9.

d. Palauan verbs can be modified in form and meaning by a

large number of reduplicative processes. Special meanings
can also be obtained by adding the predictive or inchoative
suffixes. See 11.12 and 11.12.1–5 for further details.

163
 5 Palauan Verbs

5.8. SUMMARY OF PALAUAN VERBS

The many types of Palauan verbs introduced in this chapter are
summarized in the rather rough diagram below (fig. 4). Note
that this diagram does not take account of hypothetical verb
forms, certain subvarieties of state verbs, and the different verb
forms occurring in active vs. passive sentences (cf. 5.6 above).

Figure 4

164
 6 The Verb Marker and
Perfective Verb Forms

6TheVerbMarkerandPerfectiveVerbForms
6.1. THE VERB MARKER PREFIXES
In 5.4 we noted that the sole function of the Palauan verb
marker is to identify certain words (or classes of words) as
verbs. We also took a preliminary look at the different forms of
the verb marker and described their distribution. We shall now
examine the distributional characteristics of the verb marker
more thoroughly.
By far, the most commonly-occurring variant of the verb
marker is the prefix mȩ- (which is reduced to m- if the following
verb stem begins with a vowel). Less frequently, the verb
marker occurs as the prefix o-, which appears to be an alternate
form of mȩ- (see below). The prefixed verb markers mȩ- and
o- are found in many types of verbs, as we saw in 5.4. Thus,
ergative verb forms like mȩchat ‘be/get smoked’ and oboes ‘be/
get shot’ consist of the verb marker followed by the verb stem
(e.g. mȩ + chat, o + boes), while imperfective verb forms like
mȩngat ‘smoke (fish)’ and omoes ‘shoot’ are more complex in
structure, since the im perfective marker (cf. 5.5) comes be-
tween the verb marker and the verb stem (e.g. mȩ + ng + chat,
o + m + boes).
Many intransitive action and state verbs also contain the
verb marker prefix mȩ-, as the lists below indicate:

(1) Intransitive Action Verbs Intransitive State Verbs

mȩrael ‘walk, travel’ mȩkngit ‘bad’

mȩngȩdub ‘swim’ mȩsaul ‘tired’
mȩlȩcho ‘bathe’ mȩsisiich ‘strong’
mȩchiuaiu ‘sleep’ mȩkeald ‘warm’
mȩngȩdȩchȩduch ‘talk, converse’ mȩkȩlȩkolt ‘cold’
mȩtengȩl ‘come down, land’ mȩkȩsai ‘few (in number)’
mȩrdȩkekl ‘jump’ mȩched ‘shallow’
milil ‘play’ mȩiusȩch ‘calm’
mȩkiis ‘wake up, get up’ mȩtȩtkakl ‘careless’

165
 6 The Verb Marker and Perfective Verb Forms

mȩliod ‘drip’ mȩrechȩd ‘early, fast’

Though most of the verb stems in the examples above are

bound (i.e., they never occur without a prefix), some of them
can occur independently as nouns: thus, we have mȩrael ‘walk,
travel’— rael ‘road’, mȩsaul ‘tired’—saul ‘tiredness, trouble’,
mȩkȩsai ‘few (in number)’—kȩsai ‘insufficient quantity’, and
mȩched ‘shallow’— ched ‘low tide’.
The verb marker prefix always appears as o- if the following
verb stem begins with b; this is true whether or not the
imperfective marker comes between the prefix and the verb
stem. Thus, verbs like the following take o- in both the ergative
and imperfective forms:

(2) Ergative Form Imperfective Related Noun (= Verb

Form Stem)

oboes ‘be/get omoes ‘shoot’ boes ‘gun’

shot’
obail ‘be/get omail ‘clothe’ bail ‘article of clothing’
clothed’
obȩkall ‘be/get omȩkall ‘drive, bekall ‘sail’
driven’ steer’

The verb marker prefix o- observed in the examples of 2

may be a phonetically-altered form of mȩ - which results from
a phonetic process called dissimilation. As we saw in 1.3.3,
assimilation is a phonetic process in which one of two unlike
sounds becomes similar to or identical with the other. Dis-
similation involves exactly the opposite process: here, one of
two similar or identical sounds becomes different from the
other in some way. Now let us see how the process of dissimi-
lation works in the examples of 2.
If we assume mȩ- to be the basic form of the verb marker
(i.e., the form from which the other variants of the verb marker
are derived), then we can represent the basic structure of a
word like oboes ‘be/get shot’ as mȩ + boes. Notice that mȩ +
boes contains two bilabial consonants—m, which is a bilabial
nasal, and b, which is a bilabial stop. There seems to be a rule
in the sound system of Palauan which prevents the m of the verb
marker from being too close to another bilabial consonant; thus,
in the basic structure mȩ + boes, the m dissimilates from the

166
 Palauan Reference Grammar

following consonant b and becomes the vowel o (note that the ȩ

of the verb marker is also lost). Although it is a vowel, this re-
sulting o retains one of the important phonetic features of the
m from which it is derived: the use of both lips in pronouncing
m is carried over to o, which is a mid back rounded vowel (cf.
1.4.2). A similar process of dissimilation accounts for some of
the infixed variants of the verb marker, as we will see in 6.2–3
below.
As we have just seen, the verb marker regularly appears as
o- if it precedes a bilabial consonant. There is a small number of
verbs, however, in which the verb marker is o- even though no
bilabial consonant follows. Such exceptional verbs include the
following:

(3) Verb (Transitive or ntransitive) Related Noun (if any)

oker ‘ask’ ker ‘question’

oklukl ‘cough’ klukl ‘cough’
okoad ‘fight’ koad ‘technique of fighting’
osus ‘greet’ sus ‘act of greeting’
ochȩrchur 1 ‘laugh’ chur ‘laughter’
osiik ‘look for’ siik ‘search’
okor ‘refuse’ ---
okiu ‘go by way of’ ---

You may recall that some of the verbs of 3 are irregular in an-
other respect: as we noted in 5.5, the imperfective forms of
transitive verbs like oker ‘ask’, osus ‘greet’, osiik ‘look for’, etc.
do not seem to contain the imperfective marker at all.

6.1.1. THE PREFIX ou-

There are many Palauan verbs which are formed by adding the
prefix ou- to some other word, usually a noun. This prefix ap-
pears to be a combination of the verb marker o- and some other
element, but its structure is not clear. Verbs formed with ou-
often designate ownership of or control over whatever the stem
noun refers to, as in the following examples:

(4) Derived Verb in ou- Related Noun (= Verb Stem)

oublai ‘own a house’ blai ‘house’

oustoa ‘own/run a store’ stoa ‘store’

167
 6 The Verb Marker and Perfective Verb Forms

ousidosia ‘own a car’ sidosia ‘car’

oucharm ‘keep a pet’ charm ‘animal’
ousers ‘have a garden, farm’ sers ‘garden’

As you can see, the prefix ou- attaches to native Palauan nouns
as well as to nouns of foreign origin. Can you tell which is which
in the examples above?
In other cases, the derived verb in ou- refers to some kind of
activity or relationship which would normally be associated with
the meaning of the stem noun. As the examples below show, it
is difficult to predict exactly what the derived verb will mean:

(5) Derived Verb in ou- Related Noun (= Verb

Stem)

ousibai ‘enslave, make a slave of’ sibai ‘slave’

ouskuul ‘teach, tutor’ skuul ‘school’
oukita ‘play the guitar’ kita ‘guitar’
oureng ‘wish for, hope for’ reng ‘heart, spirit’
ouchais ‘tell (someone) news (about chais ‘news’
something)’
ousȩchȩlei ‘be friends with’ sȩchȩlei ‘friend’

Much less frequently, the stem from which a verb in ou- is

derived is a state verb, as in oumȩra ‘believe’—mȩra ‘true’, or a
bound element, as in ousbech ‘need’. The use of some of these
derived verbs is illustrated in the sentences below:

(6) a. A Droteo a oublai ȩr se ȩl blai.

‘Droteo owns that house.’

b. Kȩ oustoa er a Belau?
‘Do you run a store in Palau?’

c. A rȩchad er a Belau a oucharm a bilis.

‘Palauans keep dogs.’

d. A Droteo a ousers ȩr a ked.

‘Droteo is farming the hillside.’

e. Ak ouskuul ȩr ngii ȩr a ochur.

‘I’m tutoring him in math.’

168
 Palauan Reference Grammar

f. Ng sȩbȩchem ȩl oukita?
‘Can you play the guitar?’

g. Ak ousbech ȩr a bilsȩngem.
‘I need your boat.’

In 5.1.3 we saw that the past tense forms of verbs containing

the verb marker prefix o- are derived simply by replacing this
o- with ul(ȩ)-. The past tense forms of verbs beginning in ou-
are derived in exactly the same way: thus, we have ouskuul
‘teach’— uluskuul ‘taught’, ousbech ‘need’—ulusbech ‘needed’,
etc.

6.2. THE INFIXED VERB MARKER AND

METATHESIS
In 5.4 we noted that many intransitive (action or state) verbs
contain the verb marker in the form of an infix which appears
after the initial consonant of the verb stem. This infix has the
three variants -(ȩ)m-, -u-, and -o-. In the examples below, the -
(ȩ)m- variant occurs; notice that the ȩ is inserted if the initial
consonant of the stem is r or ch:

(7) Intransitive Verb Related Noun (if any)

smechȩr ‘sick’ sechȩr ‘sickness’

kmeed ‘near’ ---
lmuut ‘return, come luut ‘return’
back’
lmangȩl ‘cry’ langȩl ‘crying’
chȩmiis ‘escape, run chiis ‘escape’
away’
rȩmurt ‘run’ rurt ‘ruuning, race’
rȩmos ‘drown’ ---
rȩme ‘go/come ---
home’
tmorȩch ‘slip, slide’ torȩch ‘action of slipping’
ngmasȩch ‘climb, rise ngasȩch ‘rising, ceremony for
(sun)’ mother and newborn
child’

169
 6 The Verb Marker and Perfective Verb Forms

ngmelt ‘sink (in soft ngelt ‘setting’

ground), set
(sun)’

If we look at pairs like smechȩr ‘sick’ and sechȩr ‘sickness’, we

can easily see that the intransitive verb is derived by placing the
verb marker inside the stem (which then appears in two pieces,
s- and -echȩr); for this reason we have been speaking of the in-
fixed verb marker.
The infixed verb marker has the form -u- or, rarely, -o- in
certain intransitive verbs such as suebȩk ‘fly’, ruebȩt ‘fall’,
chuarm ‘suffer’, tuobȩd ‘come out’, and soisȩb ‘go into’. Since
all of these verbs have the bilabial consonants b or m in their
stems, the appearance of the verb marker as -u- or -o- rather
than -(ȩ)m- seems to be due to a process of dissimilation which
is identical to that described in 6.1 above. 2 Thus, if we assume
that -(ȩ)m- is the basic form of the infixed verb marker, then
we can say that the infixed variants -u- and -o- are due to dis-
similation of -(ȩ)m- from a following b or m. Note that the re-
sulting vowels u and o are both rounded; thus, they preserve
a distinguishing feature (namely, use of the two lips) of the m
from which they are derived.
Because the Palauan verb marker can either be a prefix or
an infix, the statement of its distribution will be rather compli-
cated. In an attempt to make a more general statement about
the distribution of the Palauan verb marker, many linguists
would propose an analysis in which the verb marker always ba-
sically occurred as a prefix. Thus, the intransitive verbs listed
in both 1 and 7 above would all have the basic structure verb
marker + verb stem, as follows:

(8) Verb Marker + Verb Stem

a. Verbs of (1): mȩ + rael ‘walk, travel’

mȩ + kngit ‘bad’
mȩ + ched ‘shallow’
b. Verbs of (7): mȩ + sechȩr ‘sick’
mȩ + langȩl ‘cry’
mȩ + ros ‘drown’

170
 Palauan Reference Grammar

In order to derive the correct forms of the verbs of 7, we would

need a special rule in which the verb marker and the initial
consonant of the verb stem exchange positions. In other words,
a sequence like mȩ + sechȩr would become s + mȩ + echȩr,
and then the ȩ of the verb marker would be deleted before the
following vowel, giving the correct form smechȩr. The technical
term metathesis (from a Greek word meaning ‘transpose’) is
used to refer to the process of sound-switching observed here;
as a result of such metathesis, the verb marker appears to have
been infixed into the verb stem.
Metathesis is a fairly widespread phenomenon in Palauan.
As we will see in 6.3 below, the verb marker metathesizes with
the initial consonant of the verb stem in all Palauan perfective
verb forms. And as we have seen above, many intransitive verbs
require this metathesis “rule” to be applied; we cannot predict,
however, which intransitive verbs will be of this type and which
will not. Once the verb marker has metathesized, dissimilation
will occur if the verb stem has a bilabial consonant; thus, a verb
like suebȩk ‘fly’ is derived by the following steps:

(9) mȩ + sebȩk (basic form) →

s + mȩ + ebȩk (by metathesis) →
s + m + ebȩk (by deletion of ȩ) →
s + u + ebȩk (by dissimilation)

Notice that we begin with the basic form mȩ + sebȩk and

successively apply rules of metathesis, deletion, and dissimi-
lation to derive the correct form suebȩk ‘fly’.
In 5.3.2 we said that the past tense forms of intransitive
action verbs like lmuut ‘return’, rȩmurt ‘run’, suebȩk ‘fly’, etc.
are derived by replacing the (metathesized) verb marker by the
past tense markers -il- or -ir-. Thus, in past tense forms like
liluut ‘returned’, rirurt ‘ran’, and silebȩk ‘flew’, we cannot find
any trace whatsoever of the verb marker. The disappearance of
the verb marker in these forms seems to be due to a complex
series of phonetic rules whose end result is that the metathe-
sized verb marker is deleted before the vowel i in an un-
stressed syllable. These conditions are met in forms like lilúut
‘returned’, rir úrt ‘ran’, and silébȩk ‘flew’, where the verb
marker is metathesized to a position before the i of the infixed

171
 6 The Verb Marker and Perfective Verb Forms

past tense markers -il- or -ir-, which are never stressed. A

similar situation will be discussed in greater detail in 6.3.3
below.

6.2.1. ABSENCE OF THE VERB MARKER IN HYPOTHETICAL VERB

FORMS
In 5.7.a we noted that Palauan hypothetical verb forms
characteristically lack the verb marker. Thus, while any imper-
fective verb has the basic structure verb marker + imperfective
marker + verb stem (e.g., mȩlim ‘drink’), its hypothetical forms
have the structure hypothetical pronoun + imperfective marker
+ verb stem (e.g., kulim ‘(if) I drink’, lolim ‘(if) he drinks’, etc.).
Intransitive verbs with the metathesized verb markers -(ȩ)m-, -
u-, or -o- likewise have hypothetical forms which lack the verb
marker, as indicated below:

(10) Intransitive Hypothetical Form (with 3rd pers.

Verb hypothetical pronoun)

smechȩr ‘sick’ lsechȩr ‘(if) he’s sick’

rȩme ‘go/come lȩre ‘(if) he goes/comes home’
home’
lmangȩl ‘cry’ llangȩl ‘(if) he cries’
suebȩk ‘fly’ lsebȩk ‘(if) it flies’
ruebȩt ‘fall’ lȩrebȩt ‘(if) he falls’

Following the analysis given in 8b above, the non-hypothetical

forms in the lefthand column have the basic structure verb
marker + verb stem; metathesis of the verb marker must occur,
of course, to yield the correct forms. By contrast, the hypo-
thetical forms in the righthand column have the basic structure
hypotheti cal pronoun + verb stem. None of the verb forms in 10
contains the imperfective marker, since the distinction between
imperfective vs. perfective verb forms only applies to transitive
verbs.
Now let us observe the hypothetical forms of two commonly-
occurring Palauan intransitive verbs:

(11) Intransitive Hypothetical Form (with 3rd pers.

Verb hypothetical pronoun)

mo ‘go’ lȩbo ‘(if) he goes’

172
 Palauan Reference Grammar

me ‘come’ lȩbe 3 ‘(if) he comes’

If we analyze the verbs of 11 in exactly the same manner as

we analyzed those of 10, we come to the following conclusions.
Since the hypothetical forms in 11 have the basic structure hypo
thetical pronoun + verb stem, then the verb stems under dis-
cussion are bo ‘go’ and be ‘come’. These verb stems also occur
in the non-hypothetical forms, which have the basic structure
verb marker + verb stem—namely, mȩ + bo and mȩ + be. In
order to account for the actually-occurring forms mo ‘go’ and
me ‘come’, the verb marker must metathesize with the initial
consonant of the verb stem, just as in the examples of 8b. But
an additional phonetic rule applies: since the Palauan sound
system does not allow the consonant cluster bm, the b of the
verb stem, which comes to appear before the m of the metathe-
sized verb marker, must be deleted. Thus, even though mo ‘go’
and me ‘come’ would appear to be very simple verb forms,
they are in fact derived from more complex basic structures by
several steps, as in the following:

(12) mȩ + bo (basic form) →

b + mȩ + o (by metathesis) →
b+m+o (by deletion of ȩ) →
m+o (by deletion of b before m)

There are a few other Palauan intransitive verbs which, like

the verbs of 11, show an alternation between m in the nonhypo-
thetical form and b in the hypothetical forms. These can be ex-
plained in terms of the analysis presented above and include
verbs like mȩkiis ‘wake up’—lȩbȩkiis ‘(if) he wakes up’ and mla
‘was (located)’—lebla ‘(if) he was (located)’. 4
At the beginning of this section we made the claim that
Palauan hypothetical verb forms characteristically lack the verb
marker. While this statement is true for many classes of verbs,
we do find some notable exceptions. For example, intransitive
state verbs which have the prefixed verb marker mȩ- (see
1 above) retain this verb marker in their hypothetical forms,
as in mȩkngit ‘bad’—lȩmȩkngit ‘(if) it’s bad’, mȩched
‘shallow’—lȩmȩched ‘(if) it’s shallow’, etc. Furthermore, while
the great majority of intransi tive action verbs beginning with
mȩ- lose this verb marker in their hypothetical forms (mȩrael
‘walk’—lorael ‘(if) he walks’, mȩngȩdub ‘swim’—longȩdub ‘(if)
he swims’, etc.), a very small number retain it, as in mȩchiuaiu

173
 6 The Verb Marker and Perfective Verb Forms

‘sleep’—lȩmȩchiuaiu ‘(if) he sleeps’. Finally, the hypothetical

forms of ergative verbs (cf. 5.4) retain the verb marker, as
in mȩngim ‘be/get drunk’—lȩmȩngim ‘(if) it is/gets drunk’,
mȩchuiu ‘be/get read’—lȩmȩchuiu ‘(if) it is/gets read’, etc.

6.3. SELECTED LIST OF PERFECTIVE VERB

FORMS
In previous chapters we have already had occasion to discuss
some of the major structural features of Palauan perfective verb
forms. Thus, in 4.9 and 4.9.1–4 we considered the form and
distribution of the object pronouns which are suffixed to per-
fective verbs. Then, in 5.5 we wrote “formulas” for the con-
trasting structures of perfective vs. imperfective verb forms
and noted that in the former type, the verb marker always ap-
pears as an infix. Not surprisingly, the infixed verb marker in
perfective verb forms results from the very same process of
metathesis described in 6.2 above to account for intransitive
verbs such as smechȩr ‘sick’, ruebȩt ‘fall’, etc. In perfective verb
forms, however, the metathesized verb marker is affected by
additional phonetic processes which are very complex. Before
clarifying these processes, we need to be familiar with the many
phonetic variations observed in the perfective verb forms them-
selves. Therefore, we will provide a large number of these forms
below.
The four major perfective forms of a representative sample
of Palauan transitive verbs will now be listed according to the
following format. First, the transitive verb will be given in its
imperfective form, along with an English gloss; then, the four
perfective forms will be given in the following order:

(13) Present tense, 3rd Past tense, 3rd pers.

pers. sg. object sg. object
Present tense, 3rd Past tense, 3rd pers.
pers. pl. non-human pl. non-human object
object

It will not be necessary to provide English glosses for the

perfective forms since their meanings can easily be determined
from the chart in 13; thus, the perfective forms of mȩlalȩm
‘plant’, for example, would be translated as follows:

174
 Palauan Reference Grammar

(14) ‘plant it’ ‘planted it’

‘plant them’ ‘planted them’

Observe the perfective forms below:

(15) a. mȩlalȩm ‘plant’: b. mȩlamk ‘shave, scrape’:

dolȩmii dillȩmii tomkii tilȩmkii
dualȩm dilalȩm tuamk tilamk

c. mȩlasȩch ‘carve (canoe)’: d. mȩlamȩch ‘smoke,

chew’:
dosȩchii dilsȩchii chomȩchii chilmȩchii
dmasȩch dilasȩch chuamȩch/ chilamȩch/
nguamȩch ngilamȩch

e. mȩlatȩch ‘clean’: f. mȩlȩcholb ‘bathe,

baptize’:
ngotȩchii ngiltȩchii tochȩlbii tilchȩlbii
ngmatȩch ngilatȩch tocholb tilcholb

g. mȩleel ‘nail’: h. mȩlȩkosȩk ‘cut (meat)’:

dmelii dilelii sokȩsȩkii silȩkȩsȩkii
dmeel dileel sokosȩk silȩkosȩk

i. mȩleng ‘borrow’: j. mȩlȩngoes ‘cook (in

water) (fish, rice, meat)’:
longir lilȩngir songosii silȩngosii
lmeng lileng songoes silȩngoes

k. mȩles ‘slice (fish, meat)’: l. mȩlesȩb ‘burn’:

dosȩngii dilsȩngii sosȩbii silsȩbii
dmes diles suese,b/ silesȩb
smese.

m. mȩliich ‘remove meat n. mȩliud ‘cut (round

(from a coconut)’: object)’:
liiȩchii lilichii tiudii tiludii
lmiich liliich tmiud tiliud

o. mȩluchȩs ‘write’: p. mȩlul ‘burn, barbeque’:

luchȩsu lilȩchȩsii durur dirrur
lmuchȩs liluchȩs dmul dilul

175
 6 The Verb Marker and Perfective Verb Forms

q. mȩnga ‘eat’: r. mȩngȩlebȩd ‘hit, spank’:

kolii killii cholȩbȩdii chillȩbȩdii
kma kila cholebȩd chillebȩd

s. mȩngȩreel ‘catch fish t. mȩngȩtmokl ‘clean,

(with a line)’: straighten up’:
korelii kirrelii kutmȩklii kiltmȩklii
koreel kirreel kutmokl kiltmokl

u. mȩngȩsbrebȩr ‘paint’: v. mȩngib ‘pick (fruit)’:

chosbȩrbȩrii chilsbȩrbȩrii chibngii chilȩbngii
chosbrebȩr chilsbrebȩr chuib chilib

w. mȩngiis ‘dig’: x. mȩngiis ‘open’:

kiiȩsii kilisii kiisii kilisii
kmiis kiliis kmiis kiliis

y. mȩngimd ‘cut (hair), z. mȩngitakl ‘sing’:

trim’:
kimdii kilȩmdii chitȩklii chilȩtȩklii
kuimd kilimd choitakl chilitakl

aa.mȩruul ‘make, prepare’: bb.mȩnguiu ‘read’:

rullii rirȩllii chuiȩuii chiliuii
rȩmuul riruul chȩmuiu chiluiu

6.3.1. METATHESIS IN PERFECTIVE VERB FORMS

When we look at the perfective verb forms of 15, we notice quite
a few familiar things. First of all, we should have no trouble
identifying the object pronoun suffixes: as we saw in 4.9 and
4.9.1–4, the third person singular object pronoun is -ii, or less
frequently, -ngii or vowel + r(ir, ur, etc.), while the third person
plural non-human object pronoun is -ø. The phonetic effect of
these two object pronoun suffixes is quite different and of great
importance to a correct analysis of the phonetic alternations in
perfective verb forms. Since the third person singular object
pronoun is always stressed, all of the syllables of the verb
stem are invariably unstressed in perfective forms like dolȩmíi
‘plant it’, longir ‘borrow it’, etc. On the other hand, since the
third person plural non-human object pronoun ø has no “pho-
netic content”, it of course can never be stressed; therefore, in

176
 Palauan Reference Grammar

perfective forms like duálȩm ‘plant them’, lméng ‘borrow them’,

etc., the stress falls on one of the syllables of the verb stem. The
consequences of this difference in the placement of stress will
become obvious below.
In 5.5 we formulated the basic structures of imperfective vs.
perfective verb forms as follows:

(16) Imperfective verb verb marker + imperfective marker +

form: verb stem
Perfective verb verb marker + verb stem + object
form: pronoun

Recall that in imperfective verb forms, the initial consonant of

the verb stem usually gets deleted following the imperfective
marker (-l-, -ng-, or -m-). On the other hand, the initial con-
sonant of the verb stem is preserved in perfective verb forms,
since there is no preceding imperfective marker. For this
reason, the imperfective and perfective forms of one and the
same verb often look rather different: for example, in the im-
perfective forms m ȩlal ȩm ‘plant’ and m ȩngimd ‘cut (hair)’,
there is no trace of the verb-stem-initial consonants d and k
found, respectively, in the perfective forms dol ȩmii ‘plant it’ and
kimdii ‘cut his hair’. The verb-stem-initial consonant found in
perfective forms also appears in ergative forms (cf. 5.4). Since
these forms lack the imperfective marker and simply have the
structure verb marker + verb stem, the verb-stem-initial con-
sonant is not deleted. Thus, for the verbs under discussion, we
have the ergative forms m ȩdal ȩm ‘be/get planted’ (cf. dol ȩmii
‘plant it’) and m ȩkimd ‘get one’s hair cut’ (cf. kimdii ‘cut his
hair’).
Whereas the basic structures of both imperfective and
perfective verb forms contain the verb marker as their first el-
ement (cf. 16 above and the discussion in 5.5), perfective verb
forms are characterized by the fact that the verb marker must
metathesize with the initial consonant of the following verb
stem in order to produce the form which is actually spoken. In
some cases the metathesized verb marker appears as -m-, while
in others it changes to -u- or -o-, or even disappears completely.
We will now examine the phonetic factors which account for the
various forms of the metathesized verb marker.

177
 6 The Verb Marker and Perfective Verb Forms

As we saw in 6.2 above, the appearance of the metathesized

verb marker as -u- or, rarely, -o- in certain intransitive verbs
seems to be the result of dissimilation: in other words, in
intransitive verbs like suéb ȩk ‘fly’, chuárm ‘suffer’, etc., the
metathesized verb marker changes from the bilabial consonant
m to the vowel u (or o) if another bilabial consonant (b or m)
occurs in the verb stem and if the adjacent vowel is stressed.
The very same principle of dissimilation will account for a good
number of the 3rd pers. pl. non-human object present perfective
forms found in 15. Thus, in the words below, the metathesized
verb marker (italicized) appears as -u- because there is a bil-
abial consonant in the stem and the adjacent vowel is stressed:

(17) duálȩm ‘plant them’

tuámk ‘shave them’
chuámȩch/nguámȩch ‘smoke them’
suésȩb 5 ‘burn them’
chuíb ‘pick them’
kuímd ‘cut them’

In the forms of 17, the ȩ of the metathesized verb marker is

of course deleted before the following vowel. Thus, a form like
duálȩm ‘plant them’ is derived by the following steps:

(18) mȩ + dálȩm + Ø 6 (basic form) →

d + mȩ + álȩm + Ø (by metathesis) →
d + m + álȩm + Ø (by deletion of ȩ) →
d + u + álȩm + Ø (by dissimilation) →
d + u + álȩm (by deletion of Ø object pronoun)

Now let us examine the 3rd pers. pl. non-human object

present perfective forms of verbs whose stems do not contain
a bilabial consonant. Since the conditions which cause dissimi-
lation of the metathesized verb marker are absent in such cases,
the verb marker simply appears as -m-, as we would expect.
Note, therefore, the following forms from 15, in which the verb
marker has been italicized:

(19) dmásȩch ‘carve them’ lmúchȩs ‘write

them’
ngmátȩch ‘clean them’ dmúl ‘burn
them’
dméel ‘nail them’ kmá ‘eat them’

178
 Palauan Reference Grammar

lméng ‘borrow them’ kmíis ‘dig them’

dmés ‘slice them’ kmíis ‘open
them’
lmíich ‘remove meat (from rȩmúul ‘make
them)’ them’
tmíud ‘cut them’ chȩmúiu ‘read
them’

Just as in the forms of 17, the ȩ of the metathesized verb marker

is deleted in the examples of 19 before a following vowel. Thus,
to derive a form like dmásȩch ‘carve them’, we have the fol-
lowing steps:

(20) mȩ + dásȩch + Ø (basic form)

d + mȩ + ásȩch + Ø (by metathesis)
d + m + ásȩch + Ø (by deletion of ȩ)
d + m + ásȩch (by deletion of Ø object pronoun)

All of the forms of 19 are derived in a similar manner except

that in rȩmuul ‘make them’ and chȩmuiu ‘read them’, an ȩ is
inserted to break up the otherwise unpronounceable consonant
clusters rm and chm (cf. 1.4.5).
The remaining 3rd pers. pl. non-human object present per-
fective forms—i.e., those not listed in 17 or 19—all contain -
o- or, much more rarely, -u- as the metathesized verb marker.
Since these variants of the verb marker actually result from a
“blending” of two vowels, they will be explained (in the next
section) together with those 3rd pers. sg. object present per-
fective forms which show a similar pattern.

6.3.2. THE METATHESIZED VERB MARKER AND VOWEL BLENDING

In all 3rd pers. sg. object present perfective forms, the metathe-
sized verb marker comes to appear before a vowel in an un-
stressed syllable. In such an environment, the metathesized
verb marker undergoes several phonetic changes, including
one—vowel blend ing—which we have not yet described. In
order to understand these phonetic changes, we shall look in
detail at the derivation of the 3rd pers. sg. object present per-
fective form of mȩlatȩch ‘clean’—namely, ngotȩchii ‘clean
it’—whose basic form is represented as mȩ + ngatȩch + íi (i.e.,
verb marker + verb stem + object pronoun).

179
 6 The Verb Marker and Perfective Verb Forms

First, the verb marker metathesizes and the ȩ of the verb

marker is deleted before the following vowel, giving ng + m +
atȩch + íi. Then, the m of the verb marker changes to u, giving
ng + u + atȩch + í i. This change of the verb marker to u is not
due to dissimilation (as in the examples of 17 above) but to an
entirely different set of circumstances: here, the verb marker
becomes u before another vowel in an unstressed syllable. In
ng + u + atȩch + íi, the two adjacent unstressed vowels u and
a blend together into a single vowel o, giving the correct per-
fective form ngotȩchíi. We use the term vowel blending to de-
scribe this process because the resulting mid vowel o is a kind
of “compromise” between the high vowel u and the low vowel
a. This “in-between” quality of the resulting o can be easily seen
from the following vowel triangle (cf. 1.4.2):

The steps in the above-mentioned derivation of ngotȩchii

‘clean it’ are summarized below:

(21) mȩ + ngatȩch (basic form) →

+ íi
ng + mȩ + (by metathesis) →
atȩch + íi
ng + m + atȩch (by deletion of ȩ) →
+ íi
ng + u + atȩch (by change of verb marker to u in
+ íi unstressed syllable) →
ng + otȩch + íi (by vowel blending)

By applying the phonetic rules of 21 in the order shown, we

can correctly derive the 3rd pers. sg. object present perfective
forms of all verbs in 15 whose stems have a as the first vowel.
Thus, we can account for dolȩmii ‘plant it’ (stem: dalȩm), tomkii

180
 Palauan Reference Grammar

‘shave it’ (stem: tamk), dosȩchii ‘carve it’ (stem: dasȩch),

chomȩchii ‘smoke it’ (stem: chamȩch), and kolii ‘eat it’ (stem:
kal).
We have seen above that the phonetic process of vowel
blending accounts for the derivation of o from u+a in certain
3rd pers. sg. object present perfective forms. Other occurrences
of o in such forms are also due to vowel blending, but a different
combination of vowels is involved. For example, in the following
derivation of sosȩbii ‘burn it’ (stem: sesȩb), the vowels u and e
blend to give o in an unstressed syllable:

(22) mȩ + sesȩb + (basic form) →

íi
s + mȩ + (by metathesis) →
esȩb + íi
s + m + esȩb (by deletion of ȩ) →
+ íi
s + u + esȩb (by change of verb marker to u in
+ íi unstressed syllable) →
s + osȩb + íi (by vowel blending)

Notice that the above derivation uses the same phonetic rules
as 21 and applies them in the same order. The only difference
between 21 and 22 is that the vowel blending rule affects u+a
in 21, while it applies to u + e in 22.
If we look again at the vowel triangle above, we can see that
the o which results from the blending of u and e preserves one
major phonetic feature from each of its two “source” vowels:
in other words, the mid back vowel o retains the “midness” of
the mid front vowel e and the “backness” of the high back
vowel u. Exactly parallel to the above derivation of sosȩbii are
the derivations for the 3rd pers. sg. object present perfective
forms longir ‘borrow it’ (stem: leng) and dosȩngii ‘cut it’ (stem:
des), whose stems have e as the first vowel. Note that the pho-
netic changes under discussion do not apply in dmelii ‘nail it’,
apparently because the verb stem (deel) has a long vowel ee.
The appearance of o in certain perfective forms results from
the blending of yet another combination of vowels—namely, u
+ ȩ. This phonetic change affects the perfective forms of those
verbs of 15 whose stems have ȩ as the first vowel; 7 furthermore,

181
 6 The Verb Marker and Perfective Verb Forms

it applies regardless of whether the object pronoun suffix is sin-

gular or plural. Thus, the perfective forms cholȩbȩdii ‘hit him’
and cholebȩd ‘hit them’ (stem: chȩlebȩd) are derived as follows:

(23) Derivation of cholȩbȩdii, 3rd pers. sg. object present per-

fective form of mȩngȩlebȩd ‘hit’:

mȩ + chȩlebȩd (basic form) →

+ íi
ch + mȩ + (by metathesis) →
ȩlebȩd + íi
ch + m + ȩlebȩd (by deletion of ȩ) →
+ íi
ch + u + ȩlebȩd (by change of verb marker to u in
+ íi unstressed syllable) →
ch + olebȩd + íi (by vowel blending) →
ch + olȩbȩd + íi (by reduction of e to ȩ in unstressed
syllable 8 )

(24) Derivation of choleb ȩd , 3rd pers. pl. non-human object

present perfective form of m ȩng ȩleb ȩd ‘hit’:

mȩ + chȩlȩbȩd (basic form) →

+Ø
ch + mȩ + (by metathesis) →
ȩlébȩd + Ø
ch + m + ȩlébȩd (by deletion of ȩ) →
+Ø
ch + u + ȩlébȩd (by change of verb marker to u in
+Ø unstressed syllable) →
ch + olébȩd + Ø (by vowel blending) →
ch + olébȩd (by deletion of Ø object pronoun)

In both of the derivations above, the metathesized verb marker

changes to u, and the vowels u + ȩ blend into o in an un-
stressed syllable. Again, the change of u + ȩ to o is an example
of vowel blending because the mid back vowel o keeps the
“midness” of the mid central vowel ȩ and the “backness” of
the high back vowel u. By using derivations like 23 and 24,
we can account for nearly all the present perfective forms of
15 whose verb stems have ȩ as their first vowel—i.e., toch
ȩlbii ‘bathe him’/tocholb ‘bathe them’ (stem: t ȩcholb), sok ȩs

182
 Palauan Reference Grammar

ȩkii ‘cut it’/sokos ȩk ‘cut them’ (stenv: s ȩkco s ȩk), songosii

‘cook it’/songoes ‘cook them’ (stem: s ȩngoes), korelii ‘catch
it’/koreel ‘catch them’(stem: k ȩreel ), and chosb ȩrb ȩrii ‘paint
it’/ chosbreb ȩr ‘paint them’ (stem: ch ȩsbreb ȩr).
One striking exception to the above pattern is found in the
perfective forms kutm ȩklii ‘clean it’/ kutmokl ‘clean them’,
which contain u instead of the expected o. Even though the stem
k ȩtmokl has ȩ as its first vowel, for some unknown reason vowel
blending of the metathesized verb marker -u- and the following
ȩ does not seem to apply, and the ȩ is deleted instead.
To summarize what we have presented above, we can see
that the principle of vowel blending will account for the correct
pronunciation of many of the perfective verb forms listed in 15.
In all cases of vowel blending, the metathesized verb marker -u-
combines with some other vowel (a, e, or ȩ) in an unstressed syl-
lable to give o. This resulting o can be viewed as a compromise
or halfway point between u + a, u + e, or u + ȩ , which are its
three possible sources.

6.3.3. DELETION OF THE METATHESIZED VERB MARKER

As we have seen in the section above, the metathesized verb
marker -u- blends with a low or mid vowel (a, e, or ȩ) in an un-
stressed syllable to give o. If a high vowel (i or u) follows the
metathesized verb marker in an unstressed syllable, however,
vowel blending does not occur, and instead the metathesized
verb marker is deleted. This simple phonetic rule will account
for all of the 3rd pers. sg. object present perfective forms in 15
whose verb stems have i or u as their first vowel. Observe, for
example, the following derivations:

(25) Derivation of kimdii, 3rd pers. sg. object present perfective

form of m ȩngimd ‘cut (hair)’:

mȩ + kimd + (basic form) →

íi
k + mȩ + imd (by metathesis) →
+ íi
k + m + imd (by deletion of ȩ) →
+ íi

183
 6 The Verb Marker and Perfective Verb Forms

k + u + imd + (by change of verb marker to u in unstressed

íi syllable) →
k + imd + íi (by deletion of verb marker)

(26) Derivation of luch ȩsii , 3rd pers. sg. object present per-
fective form of m ȩluch ȩs ‘write’:

mȩ + luchȩs + (basic form) →

íi
l + mȩ + uchȩs (by metathesis) →
+ íi
l + m + uchȩs (by deletion of ȩ) →
+ íi
l + u + uchȩs (by change of verb marker to u in unstressed
+ íi syllable) →
l + uchȩs + íi (by deletion of verb marker)

Derivations identical to 25 and 26 will explain the following

3rd pers. sg. object present perfective forms of 15, which show
no trace whatsoever of the verb marker: lii ȩchii ‘remove meat
(from it)’, tiudii ‘cut it’, chibngii ‘pick it’, kii ȩsii ‘dig it’, kiisii
‘open it’, chitȩklii ‘sing it’, durur ‘burn it’, rullii ‘make it’, and
chuiȩuii ‘read it’. In each case, the first vowel of the verb stem
is i or u.
Just as in the examples above, no trace of the metathesized
verb marker can be found in any of the past perfective forms
listed in 15. This absence is due to the fact that in these forms
the verb marker metathesizes to a position directly preceding
the past tense marker -il- or -ir-, which is never stressed. Be-
cause the past tense marker begins with i, deletion of the verb
marker takes place according to the rule mentioned above. The
following derivation is typical:

(27) Derivation of lilȩngir, 3rd pers. sg. object past perfective

form of mȩleng ‘borrow’:

mȩ + l + il + (basic form, including infixed past tense

eng + ír marker -il-) →
l + mȩ + il + (by metathesis) →
eng + ír
l + m + il + eng (by deletion of ȩ) →
+ ír

184
 Palauan Reference Grammar

l + u + il + eng (by change of verb marker to u in

+ ír unstressed syllable) →
l + il + eng + ír (by deletion of verb marker) →
1 + il + ȩng + ír (by reduction of e to ȩ in unstressed
syllable)

The only exception to the rule under discussion is the 3rd

pers. pl. non-human object present perfective form choitakl
‘sing them’. Here, the metathesized verb marker has not been
deleted before i in an unstressed syllable, but instead remains
as -o-.
In 6.2 above, we considered a small group of intransitive
verbs like suebȩk ‘fly’, ruebȩt ‘fall’, etc. in which the metathe-
sized verb marker appears as -u- because the verb stem con-
tains a bilabial consonant. We noted that in the past tense
forms of these verbs—e.g. silébȩk ‘flew’, rirébȩt ‘fell’, etc.—the
verb marker disappears. We can now see that the loss of the
verb marker in these cases is due to the very same phonetic
rule which accounts for derivations like 25–27 above. In other
words, in a past tense form like silébȩk ‘flew’, the metathesized
verb marker -u- is deleted because it appears before i (the i of
the past tense marker -il-) in an unstressed syllable.
In a very general sense, there is an important similarity be-
tween the phonetic process of vowel blending examined in 6.3.2
above and the process of verb marker deletion discussed in this
section. Both processes convert a vowel cluster in the basic
form of a verb into a single vowel in the actually-spoken form.
Thus, vowel blending and verb marker deletion supply forms
which satisfy a nearly exceptionless principle of the Palauan
sound system—namely, that vowel clusters do not occur in
unstressed sylla bles. The same principle was illustrated in
1.4.7 and 3.4.2–3, where the shortening or reduction of vowel
clusters in the possessed forms of nouns was analyzed in detail.

6.4. VOWEL REDUCTION AND VOWEL DELETION

IN PERFECTIVE VERB FORMS
In 3.4 and 3.4.1–3 we accounted for many of the phonetic dif-
ferences between independent nouns and their possessed forms
in terms of phonetic rules which reduce or delete single con-
sonants or consonant clusters in unstressed syllables. The very

185
 6 The Verb Marker and Perfective Verb Forms

same phonetic rules can account for all of the vowel alterna-
tions observed in the perfective verb forms of 15 above, as we
will now see.
In the examples below, a full vowel which is stressed in the
imperfective form and in the 3rd pers. pl. non-human object per-
fective forms reduces to ȩ when unstressed in certain 3rd pers.
sg. object perfective forms:

(28) Imperfective Form 3rd pers. sg. Object Past

Containing Stressed Full Perfective Form Containing
Vowel Reduced Vowel (ȩ)

mȩlámk ‘shave, tilȩmkíi ‘shaved it’

scrape’
mȩlȩchólb ‘bathe’ tilchȩlbíi ‘bathed him’
mȩlȩkósȩk ‘cut (meat)’ silȩkȩsekíi ‘cut it’
mȩlúchȩs ‘write’ lilȩchȩsíi ‘wrote it’
mȩngȩlébȩd ‘hit’ chillȩbȩdíi ‘hit it’
mȩngitákl ‘sing’ chilȩtȩklíi ‘sang it’
mȩngȩtmókl ‘clean’ kiltmȩklíi ‘cleaned it’
mȩléng ‘borrow’ lilȩngír ‘borrowed it’
mȩngíb ‘pick (fruit)’ chilȩbngíi ‘picked it’

In other cases, a full vowel which is stressed in certain forms

of a verb is deleted altogether when unstressed, as in the follow-
ing examples:

(29) Imperfective Form 3rd pers. sg. Object Past

Containing Stressed Full Perfective Form (Note Loss of
Vowel Vowel)

mȩlálȩm ‘plant’ dillȩmíi ‘planted it’

mȩlátȩch ‘clean’ ngiltȩchíi ‘cleaned it’
mȩlásȩch ‘carve’ dilsȩchíi ‘carved it’
mȩlés ‘slice’ dilsȩngíi ‘sliced it’
mȩlésȩb ‘burn’ silsȩbíi ‘burned it’

Long vowels or vowel clusters which are stressed in certain

forms of a verb are usually reduced to a single vowel when un-
stressed, as the examples below indicate:

186
 Palauan Reference Grammar

(30) Imperfective Form 3rd pers. sg. Object Past

Containing Stressed Long Perfective Form (Note
Vowel or Vowel Cluster Reduction of Long Vowel or
Vowel Cluster)

mȩléel ‘nail’ dilelíi ‘nailed it’

mȩlíich ‘remove meat lilichíi ‘removed meat
(from a (from it)’
coconut)’
mȩngȩréel ‘catch fish (with kirrelíi ‘caught it (with a
a line)’ line)’
mȩngíis ‘dig’ kilisíi ‘dug it’
mȩlȩngóes ‘cook’ silȩngosíi ‘cooked it’
mȩlíud ‘cut (round tiludíi ‘cut it’
object)’
mȩngúiu ‘read’ chiliuíi ‘read it’

In rare cases, a stressed long vowel reduces to ȩ when un-

stressed, as in mȩrúul ‘make’—rirȩllíi ‘made it’.

*6.5. ADDITIONAL REMARKS ON PERFECTIVE

VERB FORMS
Some of the perfective verb forms in 15 exhibit certain unusual
or exceptional features which we have not yet discussed. We
will point these out in the paragraphs below.

a. Though the verb mȩ lul ‘burn, barbeque’ has the stem dul, the
3rd pers. sg. object perfective forms are durur and dirrur. The
final l of the stem dul assimilates to the r of the object pronoun
suffix -ur in the present perfective form, which is derived as fol-
lows:

(31) mȩ + dul + (basic form) →

úr
d + mȩ + ul (by metathesis) →
+ úr
d + m + ul + (by deletion of ȩ) →
úr
d+ u +ul +úr (by change of verb marker to u in
unstressed syllable) →

187
 6 The Verb Marker and Perfective Verb Forms

d + ul + úr (by deletion of verb marker) →

d + ur + úr (by assimilation of l to r)

The past perfective form dirrur is apparently derived by the

following steps:

(32) mȩ + d + il + (basic form, including infixed past tense

ul + úr marker -il-) →
d + mȩ + il + (by metathesis) →
ul + úr
d + m + il + ul (by deletion of ȩ) →
+ úr
d + u + il + ul (by change of verb marker to u in
+ úr unstressed syllable) →
d + il + ul + úr(by deletion of verb marker) →
d + il + l + úr (by deletion of unstressed u—cf. 6.4
above) →
d + il + r + úr (by assimilation of l to r) →
d + ir + r + úr (by assimilation of l to r)

b. The 3rd pers. sg. object perfective forms of mȩnga ‘eat’ are
kolii and killii. The appearance of l in these forms indicates that
the basic form of the verb stem is kal. It seems as if the l of the
verb stem must be deleted in word-final position after the vowel
a, as in mȩnga ‘eat’, kma ‘eat them’, and kila ‘ate them’. The
stem-final l is also retained in kall ‘food’, where it is followed by
the anticipat ing state suffix -l (see 7.8).

c. Mȩngiis ‘dig’ (a native Palauan verb) and mȩngiis ‘open’

(a verb formed from the verb marker mȩ + the imperfective
marker ng + the English word keys) have identical forms for
many speakers. As indicated in 15, however, some speakers dis-
tinguish the 3rd pers. sg. object present perfective forms of
these verbs. For mȩngiis ‘dig’, we have kiiȩsii, with an extra ȩ
pronounced between the long vowel ii and the following con-
sonant; this extra ȩ turns up under similar circumstances in
other native verb forms such as liiȩchii ‘remove meat (from a
coconut)’ and chuiȩuii ‘read it’. By contrast, mȩngiis ‘open’ has
kiisii, with no extra ȩ appearing between the long vowel and the
following consonant; this irregularity is undoubtedly due to the
non-native origin of this verb.

188
 Palauan Reference Grammar

d. For the verb mȩlamȩch ‘smoke, chew’, most speakers use

perfective forms which begin with ch-, but there is some vari-
ation observed for the 3rd pers. pl. non-human object perfective
forms, which for certain speakers begin with ng-. A verb-stem-
initial ng is also seen in the related causative verb omȩngamȩch
‘make (someone) chew/smoke’ (see 9.2.1.2). The ergative form
of this verb, however, shows verb-stem-initial ch: mȩchamȩch
‘be/get chewed/smoked’. Because of this alternation between
verb-stem-initial ng and ch, it is difficult to decide which con-
sonant appears in the basic form of the verb stem. The choice
of ng seems preferable, however: if the verb stem in question
is ngamȩch, then we can correctly predict the imperfective
form mȩlmaȩch from mȩ + l + ngamȩch, where l is chosen
as the imperfective marker because of the following verb-stem-
initial ng. We might speculate that those verb forms showing
a ch-initial verb stem—i.e., perfective forms such as chomȩchii
‘chew/smoke it’ or chuamȩch ‘chew/smoke them’ and the
ergative form mȩchamȩch ‘be/get chewed/smoked’—are the
result of an assimilation rule: that is, the initial ng of ngamȩch
completely assimilates to (i.e., becomes identical with) the ch at
the end of the verb stem.

e. The 3rd pers. sg. perfective forms of mȩruul ‘make, prepare’

show a doubling of the l—i.e., we have rullii ‘make it’ and rirȩllii
‘made it’. It is possible that the stem for this verb is ruull and
that the ll shortens in word-final position, while remaining intact
before a suffix. Further evidence for this analysis may be found
by comparing the derived noun rruul ‘something which is made’
(which is actually a resulting state verb in form—see 8.2) with
its possessed forms rrȩllek ‘the thing I have made’, rrȩllem ‘the
thing you have made’, etc.

6.6. PERFECTIVE FORMS OF TRANSITIVE VERBS

IN O-
In 6.1 above we listed a small number of transitive verbs whose
imperfective forms have the prefix o- as the verb marker. The
perfective forms of a few of these verbs are listed below ac-
cording to the same format presented in 15 (q.v.):

(33) a. oker ‘ask’: b. omail ‘clothe’:

189
 6 The Verb Marker and Perfective Verb Forms

korir kirrir milii mililii

kmer kiler mail milail

c. omoes ‘shoot’: d. osiik ‘look for, find’:

mosii milosii siiȩkii silikii
moes miloes smiik siliik

All of the perfective forms given above can be explained in

terms of phonetic rules with which we are already familiar. A
form like korir ‘ask him’, for example, results from vowel blend-
ing and other rules, as in the following derivation:

(34) mȩ + ker + (basic form) →

ír
k + mȩ + er (by metathesis) →
+ ír
k + m + er + (by deletion of ȩ) →
ír
k + u + er + (by change of verb marker to u in unstressed
ír syllable) →
k + or + ír (by vowel blending)

The m-initial perfective forms of omail ‘clothe’ and omoes

‘shoot’ all involve the phonetic rule deleting b before m which
we introduced in 6.2.1 above (cf. the derivation of mo ‘go’ in
12). Thus, a form like milii ‘clothe him’ is derived by the fol-
lowing steps:

(35) mȩ + bail + íi (basic form) →

b + mȩ + ail + íi (by metathesis) →
b + m + ail + íi (by deletion of ȩ) →
m + ail + íi (by deletion of b before m) →
m + il + íi (by vowel cluster reduction)

In the derivation above, the metathesized verb marker does not

change to u, even though it occurs in an unstressed syllable;
this is probably due to the fact that, after the deletion of b,
the m of the (metathesized) verb marker once again comes to
appear in word-initial position. Note, further, that the stressed
vowel cluster ai found in the imperfective form omáil ‘clothe’
and in the 3rd pers. pl. non-human object perfective forms máil
‘clothe them’ and miláil ‘clothed them’ is reduced to i when un-
stressed in the 3rd pers. sg. object perfective forms milíi ‘clothe

190
 Palauan Reference Grammar

him’ and mililíi ‘clothed him’. The same kind of vowel cluster re-
duction is observed in omóes ‘shoot’ vs. mosíi ‘shoot him’, etc.,
and the long vowel ii of osíik ‘look for’ shortens to i when un-
stressed in silikíi ‘found him’.
In past perfective forms like kirrir ‘asked him’, silikii ‘found
him’, etc., there is no trace of the metathesized verb marker,
which, as expected, is deleted before the high vowel i of the
past tense infix -il- or -ir-. Furthermore, in kirrir ‘asked him’, the
past tense marker appears as -ir- due to assimilation with the
following r of the verb stem ker.

6.7. HYPOTHETICAL FORMS OF PERFECTIVE

VERBS
In 6.2.1 above we noted that most hypothetical verb forms
characteristically lack the verb marker. This statement applies
to the hypothetical forms of perfective verbs as well. Observe
the following examples:

(36) Perfective Verb Hypothetical Form (with 3rd pers.

hypothetical pronoun)

ngotȩchii ‘clean it’ lȩngȩtȩchii ‘(if) he cleans it’

ngmatȩch ‘clean lȩngatȩch ‘(if) he cleans them’
them’
longir ‘borrow llȩngir ‘(if) he borrows it’
it’
lmeng ‘borrow lleng ‘(if) he borrows them’
them’
rȩmuul ‘make lȩruul ‘(if) he makes them’
them’

While the perfective verb forms on the left all contain the
metathesized verb marker (italicized), their corresponding hy-
pothetical forms on the right do not. As we saw in 6.3.1 above,
the perfective verb forms given in 36 have the basic structure
verb marker + verb stem + object pronoun; by contrast, their
corresponding hypothetical forms have the basic structure hy-
pothetical pronoun + verb stem + object pronoun. As we might
expect, the prefixed hypothetical pronoun has different forms
depending on the initial sound of the following verb stem (cf.
4.10.9): for example, in the hypothetical forms of 36 above,

191
 6 The Verb Marker and Perfective Verb Forms

the 3rd pers. hypothetical pronoun appears as l- before an l-

initial verb stem (as in llȩngir and lleng), but is pronounced
as lȩ- otherwise. Some hypothetical forms of perfective verbs
which contain first or second person hypothetical pronouns in-
clude kngȩtȩchii ‘(if) I clean it’, kruul ‘(if) I make them’, chom-
lȩngir ‘(if) you borrow it’, etc.
For those perfective verb forms in which the metathesized
verb marker has been deleted before a high vowel in an un-
stressed syllable, the corresponding hypothetical form is de-
rived simply by prefixing the appropriate hypothetical pronoun.
Thus, we have pairs of forms like silsȩbii ‘burned it’—lȩsilsȩbii
‘(if) he burned it’, silesȩb ‘burned them’—ksilesȩb ‘(if) I burned
them’, chiliuii ‘read it’—chomchiliuii ‘(if) you read it’, rullii
‘make it’—krȩllii ‘(if) I make it’, etc.
The verb stems for omail ‘clothe’ and omoes ‘shoot’ of 33 are
bail ‘article of clothing’ and boes ‘gun’, respectively. Since the
basic structure for hypothetical perfective forms is hypothetical
pronoun + verb stem + object pronoun, as we have seen above,
a verb like omoes ‘shoot’ has the hypothetical perfective forms
lȩbosii ‘(if) he shoots him’, lȩboes ‘(if) he shoots them’, lȩbilosii
‘(if) he shot him’, lȩbiloes ‘(if) he shot them’, etc.

192
 7 State Verbs

7StateVerbs
7.1. INTERNAL STRUCTURE OF STATE VERBS
In previous chapters we have already provided much infor-
mation about the meaning and use of Palauan state verbs.
Thus, in 5.1.1–3 we emphasized the opposition between state
verbs, which describe states, qualities, or conditions that tem-
porarily or permanently characterize someone or something,
and action verbs, which designate actions, activities, or events
in which someone (the doer or agent) participates. We also saw
that both state verbs and action verbs can be either transitive
or intransitive, although the class of transitive state verbs is
rather small. Finally, we noted that the two types of verbs can
be distinguished from each other according to the way in which
their past tense forms are derived: while state verbs use the
auxiliary word mle ‘was, were’, action verbs infix the past tense
marker -il- or -l-.
Whereas we have a fairly clear picture of the meaning and
use of Palauan state verbs, we have not yet paid much attention
to their internal structure. In other words, we still need to
examine the various ways in which morphemes—verb stems
and affixes—combine with each other in the formation of state
verbs. So far, we have only dealt with simple state verbs and
with state verbs containing the verb marker and a verb stem
(cf. 6.1). After reviewing these two types briefly, we will look at
various classes of state verbs whose internal structure is more
complex.
Simple state verbs are those which consist of a single mor-
pheme, or meaning-bearing unit, as in the following examples:

(1) ungil ‘good’ songȩrengȩr ‘hungry’

klȩbokȩl ‘beautiful’ chȩtngaid ‘thin’
dȩchudȩch ‘dirty’ cheisȩch ‘stained’
dibus ‘away, absent’ chȩrodȩch ‘noisy’
sȩkool ‘playful’ ngodȩch ‘strange, different’
chuodȩl ‘old’ ngeiasȩk ‘young’

193
 7 State Verbs

State verbs consisting of the verb marker and a verb stem

can be classified into two types, depending on whether the verb
stem is bound or whether it occurs independently as a noun.
In the examples below, all the verb stems are bound—i.e., they
cannot be used as separate words:

(2) mȩkngit ‘bad’ mȩkeald ‘warm’

mȩrur ‘ashamed, kmeed ‘near’
shy’
mȩdai ‘poor’ chuarm ‘suffer’
mȩrau ‘rich’ dmak ‘together’
mȩkreous ‘precious’ dmik ‘move out, be thrown
out’

By contrast, the verb stems in the following examples are not

bound, but can occur independently as nouns. By prefixing mȩ-
to each of the nouns below, we derive a state verb which des-
ignates a state or condition characterized by the presence of
whatever the corresponding noun refers to.

(3) State Verb Related Noun

mȩched ‘shallow’ ched ‘low tide’

milkolk ‘dark’ ilkolk ‘darkness’
mȩdakt ‘afraid of, fear’ dakt ‘fear’
mȩses ‘industrious, strong’ ses ‘industriousness’
mȩsaul 1 ‘tired’ saul ‘tiredness’
mȩchuu ‘shady’ chuu ‘shadow’
mȩkȩrior ‘unfortunate’ kȩrior ‘misfortune’
mȩduch ‘able to, skilled at’ duch ‘ability’
smechȩr ‘sick’ sechȩr ‘sickness’
smau ‘comfortable with, used sau ‘liking’
to’

In the state verbs of 2 and 3 above, the verb marker appears

in several different forms (cf. 6.1–2). In most cases, it is a prefix
(mȩ- or m-); but with certain verb stems it metathesizes to a
position following the initial consonant of the stem and appears
as -(ȩ)m- or -u-.
The state verbs of 1 through 3 can be used in sentences such
as the following:

(4) a. A mubi a ungil.

194
 Palauan Reference Grammar

‘The movie is good.’

b. A sensei a mle dibus.

‘The teacher was out.’

c. Ak kmal mle {songȩrengȩr / mȩsaul}.

‘I was very {hungry / tired}.’

d. A bilek a mla mo cheisȩch.

‘My clothes have gotten stained.’

e. A bȩchik a smechȩr ȩr a tȩretȩr.

‘My wife is sick with a cold.’

f. A Toki a chuarm ȩr a dȩlȩngchȩklel.

‘Toki is suffering because of her living conditions.’

While 4a, 4e, and 4f describe present states or conditions, 4b

and 4c, which contain the past tense auxiliary mle ‘was, were’,
describe past states or conditions. In 4d, the use of mo ‘go’
before the state verb designates a change of state (cf. 5.1.3).
In 4e and 4f, the state verbs are accompanied by the relational
phrases ȩr a tȩretȩr ‘because of a cold’ and ȩ r a dȩlȩngchȩklel
‘because of her living conditions’, which explain the cause of
the state (see 14.5).

7.2. STATE VERBS WITH PLURAL SUBJECTS

Several simple state verbs which refer to size or dimension
must take the prefix mȩ- if their subject is plural. This mȩ- is
probably an instance of the verb marker prefix, but one which
has the unique function of indicating the plurality of the subject.
Observe the following pairs of sentences:

(5) a. A mlai a klou.

‘The canoe is big.’

b. A mlai a mȩklou.
‘The canoes are big.’

(6) a. Tia ȩl oluchȩs a chȩtngaid.

‘This pencil is thin.’

195
 7 State Verbs

b. Aika ȩl oluchȩs a mȩchȩtngaid.

‘These pencils are thin.’

In 5a and 6a above, the (unprefixed) simple state verbs klou

‘big’ and chȩtngaid ‘thin’ are associated with singular subjects,
while in 5b and 6b the state verbs with mȩ- are associated with
plural subjects. A state verb referring to size or dimension must
take the prefix mȩ- if its subject is overtly plural, but must
remain unprefixed if its subject is overtly singular. The subject
of 6a— tia ȩl oluchȩs ‘this pencil’—is overtly singular because it
contains the demonstrative (or pointing) word tia ‘this’, which
automatically refers to one single thing. By contrast, the subject
of 6b—aika ȩl oluchȩs ‘these pencils’—is overtly plural because
the demonstrative word aika ‘these’ always refers to two or
more things. Thus, in 6a the singular state verb chȩtngaid ‘(is)
thin’ agrees (or corresponds) with the singular subject tia ȩl
oluchȩs ‘this pencil’, while in 6b the plural state verb mȩchȩt-
ngaid ‘(are) thin’ agrees with the plural subject aika ȩl oluchȩs
‘these pencils’. In 5a-b, the situation is somewhat different,
since the subject mlai ‘canoe’ contains no demonstrative words
and is therefore neither overtly singular nor overtly plural. In
such cases, the speaker must add mȩ- to the state verb if he is
talking about a plural subject; thus, in 5a-b, the difference in
meaning is determined solely by the presence or absence of mȩ-
.
Other state verbs which must take mȩ- to identify a plural
subject are illustrated in the following sentences:

(7) a. A rȩchad ȩr a Merikel a mȩ(ke)kȩmangȩt.

‘Americans are tall.’

b. A chiul a Toki a mȩ(ke)kȩmangȩt.

‘Toki’s hair is long.’

c. A kall ȩr a uum a kmal mȩkekȩrei.

‘The quantities of food at the cafeteria are very
small.’

The state verbs given in 7 are (ke)kȩmangȩt ‘tall, long’ and

kekȩre ‘small’. In both of these verbs, the initial syllable ke-
results from repeating (or reduplicating—see chap. 11) part
of the verb stem. This repeated initial syllable is required in

196
 Palauan Reference Grammar

kekȩre ‘small’, but is optional in (ke)kȩmangȩt ‘tall, long’

(hence, our use of parentheses). The alternate forms kȩmangȩt
and kekȩmangȩt do not show any difference in meaning. 2 In
the sentences above, mȩ- is prefixed to the state verb either
because the subject is overtly plural or because the speaker is
referring to something plural. In 7a, the subject noun phrase
rȩchad ȩ r a Merikel ‘Americans’ is overtly plural because it
contains the plural noun rȩchad ‘people’, which consists of the
plural prefix rȩ- (cf. 2.5) and the human noun chad ‘man’. And
in 7b and 7c, the speaker is talking about plural subjects: chiul
a Toki ‘Toki’s hair’ does not refer to just one strand or piece of
hair, but to many; and kall ‘food’ refers to the different kinds of
food available at the cafeteria.
Though it does not refer to size or dimension, the state
verb kikiongȩl ‘dirty’ (which also appears to contain a repeated
portion ki-) is sometimes observed to take mȩ- if a plural subject
is intended. Thus, with 5a-b, for example, compare the following
pair of sentences:

(8) a. A blatingel a Toki a kikiongȩl.

‘Toki’s plate is dirty.’

b. A blatingel a Toki a mȩkikiongȩl.

‘Toki’s plates are dirty.’

7.3. VERBS WITH MLE AND -IL- IN THE PAST

A small number of Palauan verbs have two past tense forms,
one with the auxiliary word mle ‘was, were’ and the other with
the infixed past tense marker -il-. Since these two different ways
of deriving the past tense characterize state verbs vs. action
verbs, respectively (cf. 5.1.3), we must conclude that the verbs
in question can function as either type. In some cases, use of
one or the other of the past tense forms results in a very clear-
cut difference in meaning, as in the examples below:

(9) a. A ngalȩk a mle mȩkar.

‘The child was awake.’

b. A ngalȩk a milkar ȩr a chȩrrodȩch.

‘The child woke up from the noise.’

197
 7 State Verbs

(10) a. A Droteo a mle dmik ȩr a blil a Toki.

‘Droteo was living away from Toki’s house.’

b. A Droteo a dilik ȩr a blil a Toki.

‘Droteo moved out of/got thrown out of Toki’s
house.’

In 9a and 10a above, the auxiliary mle ‘was, were’ makes

it clear that the verbs mȩkar ‘be awake, wake up’, and dmik
‘live away from, move out of’ are being used as state verbs:
therefore, mle mȩkar ‘was awake’ and mle dmik ‘was living
away from’ describe past states. By contrast, the infixed past
tense marker -il- in 9b and 10b tells us that mȩkar and dmik
are functioning as action verbs: here, milkar ‘woke up’ and dilik
‘moved out of’ denote actions or events rather than states.
There are several other Palauan verbs which have two past
tense forms, but unlike the cases above, the use of one or the
other form usually has no effect on the meaning. Thus, most
speakers use the sentences in each of the following pairs inter-
changeably:

(11)a. A John a mle mȩchiuaiu ȩr a ‘John was sleeping on the

ulaol. floor.’
b. A John a milȩchiuaiu ȩr
aulaol.

(12)a. A Hermana a mle dȩngchokl ‘Hermana was sitting

er tiang. here.’ 3
b. A Hermana a dilȩngchokl er
tiang.

(13)a. A Droteo a mle kie ȩr a ‘Droteo was living/lived in

Guam. Guam.’
b. A Droteo a kilie ȩr a Guam.

7.4. TRANSITIVE STATE VERBS

As we saw in 5.1.2, the class of Palauan transitive state verbs
is quite small. Transitive state verbs can be identified by the fol-
lowing two features: first, they take objects because they are

198
 Palauan Reference Grammar

transitive; and second, their past tense forms are derived with
mle because they are state verbs. Both of these identifying fea-
tures are observed in the sentences below:

(14) a. A John a mle mȩdȩnge a tȩkoi ȩr a Siabal.

‘John used to know Japanese.’

b. A Satsko a kmal mle mȩduch ȩr a ochur.

‘Satsko really used to know (how to do) math.’

c. Ak mle mȩtitur a tȩkoi ȩr a Merikel er se ȩr a taem ȩr a

mȩkȩmad.
‘I didn’t know English during the war.’

d. A Droteo a mle mȩrur ȩr a sȩchȩlil.

‘Droteo was ashamed of his friend.’

e. A sȩchȩlik a mle mȩdakt a bisȩbusȩch.

‘My friend used to be afraid of lightning.’

All of the transitive state verbs given above—mȩdȩnge ‘know’,

mȩduch ‘know how (to), be skilled at’, mȩtitur ‘not know how
(to), not be capable of’, mȩrur ‘ashamed of’, and mȩdakt ‘afraid
of’—involve mental states (knowledge of something, fear of
something, etc.) or abilities.
The transitive state verb mȩdȩnge ‘know’ has a full set of
perfective forms—i.e., mȩdȩngȩlkak ‘know me’, mȩdȩngȩlkau
‘know you’, mȩdȩngȩlii ‘know him/her/it’, etc. (cf. 4.9.4, ex. 53).
The use of these perfective forms is illustrated in the sentences
below:

(15) a. Ak mle mȩdȩngȩlii a Toki er se ȩr a kngar ȩr a Guam.

‘I knew Toki when I was in Guam.’

b. Ak mle mȩdȩngȩltȩrir a rȩsȩchȩlil a Droteo er se ȩr a

kngar ȩr a Hawaii.
‘I knew Droteo’s friends when I was in Hawaii.’

As 15a-b and 14a above show, the object noun phrase following
mȩdȩnge ‘know’ can be either human (e.g. Toki in 15a) or inan-
imate (e.g. tȩkoi ȩr a Siabal ‘Japanese (language)’ in 14a). 4

199
 7 State Verbs

7.5. STATE VERBS WITH BȨ-

A few state verbs can be formed by adding the prefix bȩ- to a
verb stem. In most cases, the verb stem can be used as an inde-
pendent noun, as in the following examples:

(16) State Verb Related Noun

bȩchachas ‘sooty’ chas ‘soot, ash’

bȩchochod ‘fragrant’ chochod (type of tree from
which incense is made)
bȩralm ‘watery, ralm ‘water’
fiat-tasting’ 5
bȩsokȩl ‘infected with sokȩl ‘ringworm’
ringworm’

As you can see, the derived state verbs in 16 describe states

or conditions characterized by the presence of whatever the
corresponding noun refers to (cf. the state verbs of 3 above): for
instance, bȩralm ‘watery, flat-tasting’ describes a condition re-
sulting from the presence of too much ralm ‘water’ in food, etc.
Notice that the noun chas ‘soot, ash’ has to be partially repeated
(or reduplicated—see chap. 11) before the prefix bȩ- can be
added.
Certain state verbs in bȩ- are derived from other state verbs,
but unlike the examples in 16 above, the meaning of the derived
state verb cannot be predicted in any consistent way. Note the
following examples:

(17) State Verb Related State Verb

bȩcheleleu ‘white’ cheleleu ‘pale’

bȩchachau ‘empty’ 6 chachau ‘stunted, empty (of nuts)’

A few state verbs in bȩ- do not seem to be related to any

noun or to any other state verb. Therefore, bȩ- is followed by
bound verb stems in words like bȩtimȩl ‘slow’ and bȩlils ‘high,
piercing’.

200
 Palauan Reference Grammar

7.6. THE PREFIXES BȨKȨ- AND SEKȨ-

Although Palauan has a fairly large number of state verbs de-
rived with the prefixes bȩkȩ- and sȩkȩ-, the use of such words
appears to be declining, especially among younger speakers.
These prefixes may be related to each other (note that they
share the syllable -kȩ-), but their development is not clear; fur-
thermore, bȩkȩ- may contain the prefix bȩ- discussed in 7.5
above. State verbs formed with bȩkȩ- and sȩkȩ- are very difficult
to analyze because many Palauan speakers have different
opinions about their acceptability and their meaning.
In some cases, both bȩkȩ- and sȩkȩ- can be prefixed to the
same verb stem, resulting in state verbs which are distinct from
each other in meaning. Observe, for example, the sentences
below, which contain state verbs derived from the stem rurt
‘running, race’: 7

(18) a. A ngalȩk a bȩkȩrurt.

‘The child is a good runner.’

b. A ngalȩk a sȩkȩrurt.
‘The child runs a lot.’

In 18a, the prefix bȩkȩ- derives a state verb which expresses the
subject’s ability or skill in doing the activity referred to by the
verb stem: thus, bȩkȩrurt means ‘good at running’. In 18b, how-
ever, the prefix sȩkȩ- derives a state verb with quite a different
meaning: here, sȩkȩ- implies that the activity designated by the
verb stem is something which the subject does often or likes
to do— hence, the English equivalent ‘run a lot’. A parallel dis-
tinction is found in the following pair of sentences, which have
state verbs derived from the bound stem tungȩl 8 :

(19) a. A Toki a bȩkȩtungȩl.

‘Toki has a keen sense of smell.’

b. A Toki a sȩkȩtungȩl.
‘Toki likes to smell things.’

Examples of contrast between bȩkȩ- and sȩkȩ- such as those

given in 18 and 19 above are relatively uncommon. More often,
we find that either one or the other of these prefixes occurs with
a particular verb stem. In the majority of such cases, the de-

201
 7 State Verbs

rived state verb refers to the subject’s frequent pursuit of an

activity (‘do…a lot’) rather than his skill in doing it. In the list
below, some of the most commonly-used state verbs with bȩkȩ-
and sȩkȩ- are given, together with the related verb stem; if the
verb stem occurs independently as a noun, it is provided with
an English gloss.

(20) State Verb Related

Stem

bȩkȩtȩkoi ‘talkative, talk a lot’ tȩkoi ‘word,

language’
bȩkȩsius ‘swear a lot, talk vividly’ sius
‘swearing’
sȩkȩrael ‘travel a lot, go from place to rael ‘road’
place, can’t settle down’
bȩkureor 9 ‘work a lot, hard-working’ ureor ‘work’
bȩkȩtaut ‘good at shooting’ taut ‘aim’
bȩkȩsȩchȩlei ‘have many friends, friendly’ sȩchȩlei
‘friend’
sȩkȩboes ‘go shooting a lot’ boes ‘gun’
sȩkȩbuachȩl ‘boast a lot about having a buachȩl
girlfriend or boyfriend’
sȩkȩngim ‘drink a lot (of liquor)’ ngim
sȩchiuaiu 10 ‘sleep a lot, sleep late’ chiuaiu

Since the derived state verbs of 18–20 simply have the struc-
ture bȩkȩ/sȩkȩ + verb stem, they automatically preserve the
initial consonant of the verb stem. For example, the t of tȩkoi
‘word, language’ is retained in bȩkȩtȩkoi ‘talkative’, but deleted
in mȩ- lȩkoi ‘talk’, the corresponding imperfective verb. This
verb has the basic structure mȩ + l + tȩkoi—i.e., verb marker +
im perfective marker + verb stem. As we saw in 5.5, the imper-
fective marker appears as l before verb stems like tȩkoi ‘word,
language’, which begin with t; once the correct form of the im-
perfective marker has been chosen, the initial consonant of the
verb stem is deleted. Therefore, the imperfective verb mȩlȩkoi
‘talk’ shows no trace of the initial consonant t of the verb stem
tȩkoi. For similar reasons, the initial consonants of verb stems
like boes ‘gun’ and ngim (a bound stem meaning ‘drink’) are
preserved in derived state verbs like sȩkȩboes ‘go shooting a
lot’ and sȩkȩngim ‘drink a lot (of liquor)’, but are lost in the cor-
responding imperfective verbs omoes ‘shoot’ and mȩlim ‘drink’.

202
 Palauan Reference Grammar

In a small number of cases, state verbs with bȩkȩ- and

sȩkȩ- can only be derived if all or part of the verb stem is re-
peated (or reduplicated—see chap. 11). Observe the following
examples:

(21) State Verb Related Stem

bȩkȩlilangȩl ‘cry a lot’ langȩl ‘crying’

bȩkȩsechȩsechȩr ‘get sick a sechȩr ‘sickness’
lot’
bȩkȩbȩsbes ‘forgetful’ bes ‘forgetfulness’
sȩkȩrker 11 ‘ask ker ‘question’
questions a
lot’

As a native speaker of Palauan, you have undoubtedly found

that some of the derived state verbs listed above are unac-
ceptable to you, or that the meanings provided by the English
glosses do not match your own interpretations. This is because
the use of the prefixes bȩkȩ- and sȩkȩ- is gradually dying
out—or, as linguists say, becoming less productive. For this
reason, many speakers are no longer sure of the correct form
and meaning of state verbs derived with these prefixes.

7.6.1. STATE VERBS WITH bȩkȩ- DENOTING SMELLS

The prefix bȩkȩ- has another function unrelated to that de-
scribed in 7.6 above. This prefix can also be added to nouns
to derive state verbs which refer to various kinds of smells
which are considered to be unpleasant or disagreeable. Some
examples, together with the related nouns, are given below:

(22) State Verb Related Stem

bȩkȩbau ‘smell of rotten meat or bau ‘smell’

fish’
bȩkȩriamȩl ‘smell of football fruit riamȩl ‘football
(i.e., sweaty)’ fruit’
bȩkȩkatuu ‘smell of a cat’ katuu ‘cat’
bȩkȩuel ‘smell of turtle (after uel ‘turtle’
eating turtle)’
bȩkȩchȩluch ‘smell of coconut oil’ chȩluch ‘coconut
oil’

203
 7 State Verbs

bȩkȩngikȩl ‘smell of fish’ ngikȩl ‘fish’

In addition to the above, a small number of state verbs denoting

smell contain bound stems; a typical example is bȩkȩsȩngorȩch
‘smell of a male pig’.

7.7. RESULTING STATE VERBS

In the sections above we have examined certain classes of
Palauan state verbs which are derived with the prefixes mȩ-,
bȩ-, bȩkȩ-, and sȩkȩ-. In this section we will discuss resulting
state verbs, which involve the infix -(ȩ)l-. The meaning of
Palauan resulting state verbs will be easy to understand if we
compare the following sentences:

(23) a. A Toki a mȩlatȩch ȩr a ulaol.

‘Toki is cleaning the floor.’

b. A ulaol a nglatȩch.
‘The floor is clean(ed).’

While 23a describes an action (mȩlatȩch ‘clean’) which is di-

rected at an object (ulaol ‘floor’), 23b describes the state which
the object is in as a result of this very same action. In other
words, nglatȩch ‘cleaned’ of 23b tells us that the floor has un-
dergone the action of cleaning and is now clean. Because forms
like nglatȩch ‘cleaned’ focus on the state resulting from some
completed action, they are called resulting state verbs.
Before discussing how resulting state verbs are derived, let
us look at some further pairs of sentences which parallel 23a–b
above:

(24) a. A Droteo a mȩluchȩs ȩr a babier.

‘Droteo is writing the letter.’

b. A babier a lluchȩs.
‘The letter is written.’

(25) a. A John a mȩngat a ngikȩl.

‘John is smoking the fish.’

b. A ngikȩl a chȩlat.

204
 Palauan Reference Grammar

‘The fish are smoked.’

(26) a. A Droteo a ulȩmoes a bȩlochȩl.

‘Droteo was shooting pigeons.’

b. A bȩlochȩl a mle bloes.

‘The pigeons were (injured from being)
shot.’

As 23–26 show, resulting state verbs can only be formed from

transitive action verbs (cf. 5.1.1) like mȩlatȩch ‘clean’,
mȩluchȩs ‘write’, mȩngat ‘smoke (fish)’, omoes ‘shoot’, etc. Fur-
thermore, while the transitive a-sentences have the agent (or
doer) as subject, no mention of the agent can be made in the
b-sentences, which contain resulting state verbs. This is simply
due to the fact that the b-sentences focus our attention solely
on the resulting state; the person who brought about this state
is irrelevant and need not be mentioned.
Resulting state verbs are derived simply by infixing
-(ȩ)l- after the initial consonant of the verb stem. The ȩ must
be included if the preceding consonant is ch, s, t, or d; oth-
erwise, an unpronounceable consonant cluster would result (cf.
1.4.5). In the list below, a representative sample of resulting
state verbs is given; for purposes of comparison, the related
transitive action verb (in the imperfective form) is also pro-
vided.

(27) Resulting State Verb Related Transitive Verb (in

imperfective form)

klimd ‘cut’ mȩngimd ‘cut (hair)’

chȩlsbrebȩr ‘painted’ mȩngȩsbrebȩr ‘paint’
lleng ‘borrowed’ mȩleng ‘borrow’
sȩlesȩb ‘burned’ mȩlesȩb ‘burn’
tȩlub ‘spat’ mȩlub ‘spit’
dȩles ‘sliced’ mȩles ‘slice’
rruul ‘made, done, mȩruul ‘make, do, fix’
fixed’
blurȩch ‘speared’ omurȩch ‘spear’

Since the resulting state verbs in 27 consist only of the re-

sulting state infix (italicized) and the verb stem, they pre-
serve the stem-initial consonants k, ch, s, t, b, etc. These stem-

205
 7 State Verbs

initial consonants disappear, however, in the corresponding im-

perfective verbs because they are deleted after the imper-
fective marker (cf. the discussion following ex. 20 in 7.6
above). In rruul ‘made, done, fixed’ the resulting state infix ap-
pears as r due to assimilation with the preceding verb-stem-
initial r.

7.8. ANTICIPATING STATE VERBS

Another type of state verb which can only be formed from
transi tive action verbs is the anticipating state verb. An-
ticipating state verbs are derived by adding a suffix of the form
-(ȩ)l or long vowel + l to the verb stem; in addition, they involve
some complicated phonetic changes, as we will see below. In
order to understand the meaning of anticipating state verbs, let
us compare the following two sentences:

(28) a. A Toki a mȩlatȩch ȩr a ulaol.

‘Toki is cleaning the floor.’

b. A ulaol a ngȩtachȩl.
‘The floor is to be cleaned.’

Sentence 28a (which we discussed above as 23a) describes an

action (mȩlatȩch ‘clean’) which is being directed at an object
(ulaol ‘floor’). By contrast, 28b does not describe an action
which is actually being performed at the present moment, but
instead focuses on the object as something which is expected
to undergo (or should undergo) the action at some future time.
In other words, ngȩtachȩl ‘is to be cleaned’ of 28b tells us that
the floor needs cleaning or should be cleaned—i.e., that it is
“waiting” to undergo the effect of the action of cleaning. Be-
cause forms like ngȩtachȩl ‘is to be cleaned’ refer to states
which the speaker ex pects or anticipates, we will call them
anticipating state verbs.
The use of Palauan anticipating state verbs is illustrated fur-
ther in the sentences below:

(29) a. A mlik a tȩlȩmall mȩ ng kirel ȩl ruoll.

‘My car isn’t working, so it needs to be
fixed.’

206
 Palauan Reference Grammar

b. A blim ng ruoll ȩr ker?

‘Where is your house to be built?’

c. Aika ȩl bilek a kirel ȩl mo sȩlokȩl.

‘These clothes of mine need to be washed.’

d. A chiuk a mle kmudȩl.

‘My hair was to be cut.’

e. Ngika a beakl ȩl babii.

‘This is the pig which is to be shot.’

As the sentences of 29 show, the italicized anticipating state

verbs are more or less equivalent to English expressions with
‘need to be…’ or ‘is to be…’ Since these sentences are used pri-
marily to focus our attention on the fact that something needs to
(or should) undergo the effect of some action, knowledge of the
agent is considered irrelevant. Therefore, sentences with antic-
ipating state verbs, like those with resulting state verbs (cf. 7.7
above), cannot include any mention of the agent.
As we mentioned above, anticipating state verbs are derived
by suffixing -(ȩ)l or long vowel + l to the verb stem. In the
list below, some typical anticipating state verbs are given; for
purposes of comparison, the related transitive action verb (in
the imperfective form) is also provided. Stressed syllables are
marked because they will be of importance in the subsequent
discussion.

(30) Anticipating State Verb Related Transitive Verb (in

imperfective form)

kmúdȩl ‘is to be cut’ mȩngímd ‘cut (hair)’

lȩngíil ‘is to be mȩléng ‘borrow’
borrowed’
sȩlókȩl ‘is to be washed’ mȩsílȩk ‘wash’
sȩsóbȩl ‘is to be burned’ mȩlésȩb ‘burn’
chȩtúul ‘is to be smoked’ mȩngát ‘smoke (fish)’
lȩchúkl ‘is to be written’ mȩlúchȩs ‘write’
ruóll ‘is to be made, mȩrúul ‘make, do, fix’
done, fixed’
beákl ‘is to be shot’ omóes ‘shoot’
brúchȩl ‘is to be speared’ omúrȩch ‘spear’

207
 7 State Verbs

Because the anticipating state verbs in 30 have the basic

structure verb stem + anticipating state suffix, they of course
preserve the stem-initial consonants k, ch, s, b, etc. Can you ex-
plain why these stem-initial consonants have disappeared in the
corresponding imperfective verbs?
There is no way of predicting whether the anticipating state
suffix (italicized in 30) will occur as -(ȩ)l or as a long vowel (e.g.
ii or uu) followed by l. 12 If the suffix appears as -(ȩ)l, it is never
stressed; therefore, in anticipating state forms like kmúdȩl ‘is
to be cut’, sȩlók ȩl ‘is to be washed’, s ȩsób ȩl ‘is to be burned’,
etc., the stress falls on the vowel in the syllable preceding the
anticipating state suffix -(ȩ)l. If we compare forms like sȩlókȩl
‘is to be washed’ and mȩsílȩk ‘wash’, we notice two cases of
vowel alternation. First, in the anticipating state form, the full
stressed vowel o appears between l and k, while in the imper-
fective form, a ȩ appears between these same consonants in an
unstressed syllable. Second, in the imperfective form, the full
stressed vowel i is found between s and l, while in the antici-
pating state form, a ȩ appears between these two consonants in
an unstressed syllable.
We can explain the above vowel alternations if we propose
that the verb stem for ‘wash’ has the basic stem silok. The
basic stem silok does not occur in actual pronunciation or
writing but represents a kind of “abstraction” or “phonetic
formula” from which we can predict those full vowels which
will appear in stressed syllables in the related forms of a verb.
13
In the example under consideration, the imperfective form
mȩsílȩk is stressed on the next-to-the-last syllable. Here, the i
of the basic stem silok appears as the full vowel i in the actual
pronunciation because it occurs in the stressed syllable; on the
other hand, the o of the basic stem silok appears as the reduced
vowel ȩ because it is found in an unstressed syllable. In the an-
ticipating state form sȩlókȩl, however, the situation is exactly
the reverse. In this word, it is the o of the basic stem silok which
appears in the stressed syllable and therefore gets pronounced
as the full vowel o, while the i of silok is reduced to ȩ in an un-
stressed syllable.
As the above discussion shows, we can explain the compli-
cated vowel alternations in forms such as mȩsílȩk ‘wash’ and
sȩlók ȩl ‘is to be washed’ in terms of the process of vowel
reduction (cf. 1.4.4 and 3.4), which is perhaps the most im-
portant phonetic process in Palauan. Just as we set up the

208
 Palauan Reference Grammar

abstract basic stem silok to explain the vowel alternations in

mȩsílȩk and sȩlókȩl, so can we set up abstract basic stems to ac-
count for the other pairs of words listed in 30. A few examples
are given below:

(31) Basic Stem Anticipating State Verb Imperfective Verb

sesob sȩsóbȩl ‘is to be burned’ mȩlésȩb ‘burn’

luchus lȩchúkl ‘is to be written’ mȩlúchȩs ‘write’
kimud kmúdȩl ‘is to be cut’ mȩngímd ‘cut (hair)’

The basic stems sesob and luchus given above are just like
silok in their behavior: in the anticipating state verb, the second
vowel of the basic stem appears as a full vowel under stress,
while the first vowel of the basic stem reduces to ȩ in an
unstressed syllable; and in the imperfective verb, exactly the
opposite situation occurs. The basic stem kimud given above be-
haves somewhat differently, since one or the other of its vowels
disappears completely when unstressed, rather than reducing
to ȩ. Thus, in the anticipating state verb kmudȩl ‘is to be cut’,
the i of the basic stem kimud is deleted when unstressed; and
in the imperfective verb mȩngímd ‘cut (hair)’, the u of the
basic stem is similarly deleted. Since vowel deletion can be con-
sidered an “extreme” form of vowel reduction (cf. 3.4.1), the be-
havior of kimud is not really very unusual.
There are several other phonetic changes which occur in the
anticipating state verbs of 30 that should be mentioned. If we
compare the forms mȩlúch ȩs ‘write’—l ȩchúkl ‘is to be written’
and omóes ‘shoot’—beákl ‘is to be shot’, we observe an alterna-
tion between a final s in the imperfective forms and a k in the
anticipating state forms. This alternation is due to a rather un-
usual phonetic rule of Palauan which changes s to k before l.
Thus, a form like lȩchúkl is derived by the following steps:

(32) luchús + 1 (basic form = basic stem + anticipating state

suffix) →
lȩchús + 1 (vowel reduction) →
lȩchúk + 1 (change of s to k before l).

This same phonetic change is observed between the noun sils

‘day’ and its possessed form klsel ‘his holiday’: in the possessed
form, the vowel i is deleted, resulting in the consonant cluster
sl, which then changes to kl.

209
 7 State Verbs

In some of the anticipating state verbs of 30, we observe

vowel clusters whose origin is difficult to explain. Such clusters
are found in ruóll ‘is to be made, done, fixed’ and béakl ‘is to be
shot’.

7.8.1. THE ANTICIPATING STATE SUFFIX -all

The anticipating state verbs derived with -(ȩ)l or long vowel
+ l which we discussed above are still commonly heard, al-
though another pattern of formation has come into fairly wide
use, especially among younger speakers. This pattern involves
adding the suffix -all to the verb stem; since this suffix is always
stressed, anticipating state verbs which are derived with it
always show ȩ for the full vowels of a basic verb stem, since
these come to appear in unstressed syllables. In the list below,
some commonly used anticipating state verbs in -all are given;
for purposes of comparison the alternate anticipating state verb
is also provided.

(33) Anticipating State Verb in Anticipating State Verb in

-all -(ȩ)l

sȩsȩbáll sȩsóbȩl ‘is to be burned’

lȩchȩsáll lȩchúkl ‘is to be written’
ngȩtȩcháll ngȩtáchȩl ‘is to be cleaned’
kȩmȩdáll kmúdȩl ‘is to be cut’
lȩchȩtáll lȩchótȩl ‘is to be tied’

For some verbs, the only existing anticipating state form is

the one in -áll. Some examples include bridáll ‘is to be scattered’
(cf. imperfective omriid ‘scatter’) and didáll ‘is to be followed’
(cf. imperfective omdid ‘follow’).

7.8.2. RESULTING AND ANTICIPATING STATE VERBS AS NOUNS

Many of the resulting and anticipating state verbs discussed
above can function as nouns—e.g. chȩlat ‘smoked fish’ (cf.
mȩngat ‘smoke (fish)’), ilumȩl ‘beverage’ (cf. mȩlim ‘drink’), kall
‘food’ (cf. mȩnga ‘eat’). In addition, certain state verbs can be
formed by adding both the resulting state infix -(ȩ)l- and the
anticipating state suffix -(ȩ)l to a stem simultaneously. State
verbs derived in this way appear to be identical in meaning to
the corresponding resulting state verb: for example, ngȩltachȩl

210
 Palauan Reference Grammar

‘cleaned’ means the same as nglatȩch, klȩmudȩl ‘cut (of hair)’

means the same as klimd, chȩltuul ‘smoked (of fish)’ means the
same as chȩlat, and so on. Many of these state verbs have come
to be used as nouns with special meanings. The use of resulting
and anticipating state verbs as nouns will be discussed in detail
in the following chapter.

7.9. TRANSITIVE VERBS DERIVED FROM STATE

VERBS
A rather large number of Palauan transitive verbs are derived
from simple state verbs merely by adding the verb marker and
the imperfective marker. Such derived verbs usually have a
causative meaning: that is, they describe actions in which the
subject of the sentence causes someone or something to be in
the particular state designated by the related state verb. The
following derived transitive verbs are used commonly:

(34) Transitive Verb (in imperfective Related Simple State

form) Verb

mȩngikiongȩl‘make dirty’ kikiongȩl ‘dirty’

mȩlȩchudȩch ‘make muddy’ dȩchudȩch ‘muddy’
mȩlȩkimȩs ‘make wet’ dȩkimȩs ‘wet’
mȩngeisȩch ‘make stained’ cheisȩch ‘stained’
mȩngȩtom ‘make (a knife, kȩtom ‘blunt’
etc.) blunt’
mȩngȩdeb ‘shorten’ kȩdeb ‘short’
mȩngȩmangȩt ‘lengthen’ kȩmangȩt ‘long’
mȩlodȩch ‘change’ ngodȩch ‘different,
strange’
mȩngȩdidai ‘make higher, pile kȩdidai ‘high’
up’

In a few cases, verb stems which combine with the verb

marker to form state verbs (cf. 2–3 above) also can combine
with the verb marker and the imperfective marker to form tran-
sitive verbs. Observe the examples below:

(35) Transitive Verb (in imperfective form) Related State Verb

mȩngeald ‘make warm, heat up’ mȩkeald ‘warm’

211
 7 State Verbs

mȩngȩlȩkolt ‘make cold, cool down’ mȩkȩlȩkolt ‘cold’

mȩngȩsa ‘make busy, occupy’ mȩchȩsa ‘busy’

212
 8 Noun Derivation

8NounDerivation
8.1. INTERNAL STRUCTURE OF PALAUAN NOUNS
In chaps. 5–7 we examined the internal structure of various
kinds of Palauan verbs by showing how verb stems can
combine with many different affixes such as the verb marker
prefix, the resulting state infix, the anticipating state suffix, etc.
We also observed that while the internal structure of certain
Palauan verbs is quite simple, that of others is extremely
complex. Thus, in 7.1 we discussed simple state verbs such
as klou ‘big’, ungil ‘good’, etc., which consist of just a single
morpheme, or meaning-bearing unit. Most of our time, however,
was spent in explaining the internal structure of numerous
types of complex verb forms, which contain two or more mor-
phemes. For example, in 5.4–5 we saw that ergative verb forms
such as mȩchuiu ‘be/get read’ are made up of two mor-
phemes—the verb marker (mȩ-) and the verb stem (-chuiu
‘read’)—while imperfective verb forms like mȩnguiu ‘read’ have
a basic structure which involves three morphemes—the verb
marker (mȩ-), the imperfective marker (-ng-), and the verb stem
(-chuiu ‘read’).
The internal structure of Palauan nouns, like that of Palauan
verbs, can be simple or complex. As we might expect, simple
nouns consist of just a single morpheme or meaning-bearing
unit and include words such as the following:

(1) mlai ‘canoe, car’ babier ‘paper, letter’

ngau ‘fire’ elȩcha ‘now, today’
chad ‘man, person’ ngikȩl ‘fish’
daob ‘ocean’ dȩrumk ‘thunder’
malk ‘chicken’ kȩrrȩkar ‘tree’

In contrast with the simple nouns listed above, complex

nouns are derived by adding certain types of affixes to verb
stems or to other nouns. Their internal structure therefore in-
volves at least two (and sometimes more than two) morphemes.
In the sections below, we will examine the main groups of

213
 8 Noun Derivation

Palauan complex nouns; some of these are derived with affixes

that are already familiar to us, while others involve entirely new
affixes.

8.2. STATE VERBS FUNCTIONING AS NOUNS

In 7.8.2 we mentioned that many Palauan resulting state
verbs and anticipating state verbs can function as nouns.
Thus, the nouns listed below actually have the form of antic-
ipating state verbs, which, as we saw in 7.8, are derived by
adding the anticipating state suffix -(ȩ)l (or long vowel + l)
to the verb stems of transitive action verbs. For purposes
of comparison, the related transitive action verb (in its imper-
fective form) is also given:

(2) Derived Noun Related Transitive Action Verb (in

imperfective form)

kall ‘food’ mȩnga ‘eat’

ilumȩl ‘drink, mȩlim ‘drink’
beverage’
sȩlokȩl ‘laundry’ mȩsilȩk ‘wash’
suobȩl ‘homework, mȩsuub ‘study’
study’

Whereas words like kall ‘food’, ilumȩl ‘drink, beverage’, and

suobȩl ‘homework, study’ have come to be used exclusively as
nouns (that is, they are no longer used as anticipating state
verbs), sȩlokȩl can function either as an anticipating state verb
meaning ‘is/needs to be washed’ or as a noun meaning
‘laundry’. If we assume that all of the derived nouns in 2 in
fact functioned as anticipating state verbs during some earlier
stage of the Palauan language, we can explain their present-
day (obligatory or optional) use as nouns in terms of a shift in
meaning. In other words, a noun like kall ‘food’ was originally
an anticipating state verb meaning ‘is/needs to be eaten’; and as
such, it was undoubtedly used to describe something that was
expected to undergo (or had to undergo) the action of eating.
For some unknown reason, however, this usage gradually died
out, and instead kall came to stand for the thing which is to be
eaten—namely, food. In the case of sȩlokȩl, the older usage as

214
 Palauan Reference Grammar

an anticipating state verb meaning ‘is/needs to be washed’ sur-

vived even after sȩlokȩl began to be used as a noun referring to
the thing which is to be washed—namely, laundry. 1
Since the words in 2 function as nouns, they can take any of
the possessor suffixes (cf. 3.1). Therefore, we have forms like
kall ‘food’—kȩlek 2 ‘my food’, ilumȩl ‘drink, beverage’—imȩlem
3
‘your drink’, sȩlokȩl ‘laundry’—sȩlȩkȩlel ‘his/her laundry’,
suobȩl ‘homework, study’—subȩlir ‘their homework’, etc. Words
like kȩlek ‘my food’, imȩlem ‘your drink’, etc. are necessarily
complex nouns because they contain at least two mor-
phemes—the noun stem and the possessor suffix; this is also
true, of course, for all of the possessed nouns given in chap.3.
The nouns listed below are formally identical with resulting
state verbs, which, as we saw in 7.7, are derived by placing the
resulting state infix -(ȩ)l- after the initial consonant of verb
stems of transitive action verbs. The words in 3a are used
exclusively as nouns, while those in 3b can function either as
nouns or as resulting state verbs. The related transitive action
verb (in its imperfective form) is also provided.

(3) Derived Noun Related Transitive Action

Verb (in imperfective form)

a. klȩbkab ‘chain’ mȩngȩbkab ‘chain’

chȩlitakl ‘song’ mȩngitakl ‘sing’
klȩngoes ‘meat or fish mȩlȩngoes ‘cook (meat or
4
stew’ fish)’
b. bloes ‘injury from being omoes ‘shoot’
shot’
dȩlobȩch ‘injury from being mȩlobȩch ‘cut (with a
cut’ knife, etc.)’
blalȩch ‘wound from a omalȩch ‘hit (with a
slingshot’ slingshot)’
bletȩch ‘injury from a ometȩch ‘throw (a stone,
stone’ etc.)’
blurȩch ‘wound from a omurȩch ‘spear’
spear’
chȩlat ‘thing which is mȩngat ‘smoke (fish)’
smoked, smoked
fish’
sȩlesȩb ‘thing/place mȩlesȩb ‘burn’
which is burnt’

215
 8 Noun Derivation

rruul 5 ‘thing which is mȩruul ‘make, do’

made/done’

Rather than describing the particular state resulting from

actions such as cooking, shooting, burning, etc., words like
klȩngoes, bloes, sȩlesȩb, etc. have (obligatorily or optionally)
come to stand for the thing (or place) which is cooked, shot,
burnt, etc. This latter usage is illustrated in the sentences
below:

(4) a. Ng ngar ȩ ker a {bloes / blalȩch} ȩr kau?

‘Where {the place you’re shot / your slingshot
wound}?’

b. Ng soak ȩ1 mȩnga a chȩlat.

‘I want to eat some smoked fish.’

c. Se ȩ1 sȩlesȩb ng sȩrsel a tȩchang?

‘That burned area there—whose garden is it?’

d. Ng kmal ungil a rrȩllem. 6

‘The thing you’ve made is very nice.’

A small number of nouns with specialized meanings can be

derived from verb stems by simultaneously adding both the
resulting state infix and the anticipating state suffix. In the
nouns below, the resulting state infix is italicized and the an-
ticipating state suffix is in bold type; and for purposes of com-
parison, the related transitive action verb (in its imperfective
form) is also given.

(5) Derived Noun Related Transitive Action Verb (in

imperfective form)

chȩltuul ‘smoked fish’ mȩngat ‘smoke (fish)’

llȩchukl ‘handwriting, mȩluchȩs ‘write, draw’
7 drawing’
7 ‘hole’ mȩngiis ‘dig’
kliokl
bliull ‘wrapped omail ‘clothe, wrap’
tapioca’

216
 Palauan Reference Grammar

8.3. NOUNS DERIVED WITH -(Ȩ)L- FROM

INTRANSITIVE VERBS
In 3 above, we listed some Palauan nouns which are derived by
combining the resulting state infix -(ȩ)l- with the verb stems of
transitive action verbs. As we will see in this section, it is also
possible to derive nouns by combining the infix -(ȩ)l- with the
stems of intransitive verbs (mostly state verbs). These cases
probably represent an expansion in the use of the resulting
state infix, but since the derived nouns designate abstract qual-
ities (cf. 2.2) and give no indication of the original resulting
state meaning, it will be easier if we simply consider the -(ȩ)l-
in question to be a “grammatical device” for deriving abstract
nouns from intransitive verb stems.
In the examples below, an abstract noun is derived by in-
fixing -(e)l- after the initial consonant of the verb stem, which
occurs independently as a simple state verb:

(6)Derived Noun Related Simple State

Verb

blȩkeu ‘bravery’ bȩkeu ‘brave’

kldung ‘good behavior’ kȩdung ‘well-behaved’
dȩlȩngȩrengȩr ‘poor behavior’ dȩngȩrengȩr ‘naughty’
dȩlȩngchokl ‘way of life, living dȩngchokl ‘sitting,
conditions’ seated’
chȩldȩlȩkelȩk ‘blackness’ chȩdȩlȩkelȩk ‘black’
klȩkool 8 ‘game’ sȩkool ‘playful’
blulak ‘lie’ bulak ‘deceitful, lying’
chȩrrodȩch 9 ‘noise’ chȩrodȩch ‘noisy’

The following abstract nouns are also formed by infixing

-(ȩ)l- after the initial consonant of the verb stem, which must
be preceded by the verb marker mȩ- in the corresponding state
verb (cf. 6.1 and 7.1):

(7) Derived Noun Related State Verb

klȩngit ‘sin’ mȩkngit ‘bad’
klisiich8 ‘strength’ mȩsisiich ‘strong’

klȩngakȩd8 ‘thinness’ mȩsȩngakȩd ‘thin’

kleald ‘heat’ mȩkeald ‘warm, hot’
chȩliuaiu ‘sleep’ mȩchiuaiu 10 ‘sleep, be asleep’

217
 8 Noun Derivation

chȩlsa ‘business, matter’ mȩchȩsa ‘busy’

The use of some of the derived nouns in 6 and 7 is illustrated

in the sentences below:

(8) a. Ng kmal mle klou a {blȩkeu / kldung} ȩr a irȩchar.

‘In earlier times a great many {brave / good} deeds were
done.’

b. A dȩlȩngchokl ȩr a elȩcha ȩl taem a kmal mȩringȩl.

‘Life these days is very difficult.’

c. Ng dirk ngar ȩr ngii a klisichel a Droteo.

‘Droteo is still strong.’

d. Ng kmal klou a klȩngakȩd ȩr a chȩlsel a Sina.

‘There are a lot of poorly-nourished people in China.’ 11

There are quite a few Palauan state verbs referring to size

or dimension which begin with k. This initial k- might have been
some kind of meaning-bearing prefix at an earlier stage of the
Palauan language, but now it no longer has any identifiable
function. At any rate, state verbs referring to size or dimension
can also be made into abstract nouns by infixing -(ȩ)l- after the
initial k-, as in the following:

12
(9) Derived Noun Related State Verb

kllou ‘size, thickness’ klou ‘big’

kldeb ‘shortness’ kȩdeb ‘short’
kldidai ‘height’ kȩdidai ‘high’
klȩmangȩt ‘length, height’ kȩmangȩt ‘long, tall’

The sentences below illustrate the use of some of the de-

rived nouns in 9:

(10) a. Ng tela a klungel 13 a kȩrrȩkar?

‘How thick is the tree?’

b. Ng tela a klȩmȩngȩtem?
‘How tall are you?’

c. Ng tela a klȩmȩngȩtel a kȩrrȩkar?

‘How long is the board?’

218
 Palauan Reference Grammar

d. Ng ua ngara a kldidiul 14 a kȩrrȩkar?

‘How high is the tree?’

8.4. NOUNS DERIVED FROM RECIPROCAL VERBS

As we will see in chap. 10, there is a special class of Palauan
verbs known as reciprocal verbs. These verbs, which are
formed with the reciprocal prefixes kai-, ka-, or kau-, des-
ignate actions which two or more people direct at each other
simultaneously. It is possible to derive nouns from reciprocal
verbs merely by infixing -l- after the initial k- of the reciprocal
prefix. This -l- is undoubtedly the shorter variant of the infix
-(ȩ)l- discussed in 8.3 above: in other words, it is just a gram-
matical device for deriving abstract nouns from reciprocal
verbs. Some typical derived nouns, together with the corre-
sponding reciprocal verb, are given below:

(11) Derived Noun Related Reciprocal Verb

klasoes ‘seeing/being with each kasoes ‘see each other’

other, relationship’
klaingȩseu ‘helping each other’ kaingȩseu ‘help each
other’
klakoad ‘fighting, battle’ kakoad ‘fight with each
other’
klaodȩnge ‘mutual knowledge’ kaodȩnge ‘know each
other’
klaidȩsachȩl ‘race, competition’ kaidȩsachȩl ‘race,
compete’
klausȩchȩlei ‘friendship’ kausȩchȩlei ‘be friends’
klaubuch ‘marriage’ kaubuch ‘be married’
klauchad ‘blood relationship’ kauchad ‘be related’
klaungalȩk ‘parent and child kaungalȩk ‘be related as
relationship’ parent and child’

The examples below show how some of the derived “reci-

procal” nouns of 11 can be used in sentences:

(12) a. Tia ȩl klasoes ȩr kid a di me er a elȩchang e mȩrkong.

‘Our relationship is now over (having reached this
point).’

219
 8 Noun Derivation

b. Ng kmal ungil a klaingȩseu ȩr a rȩchad ȩr a Modȩkngei.

‘The way people in Modekngei help each other is really
nice.’

c. Ng mla ȩr ngii a klakoad ȩr a Peleliu Club ȩr a kȩsus.

‘There was a fight at the Peleliu Club last night.’

d. A klaodȩnge ȩr a Toki mȩ a Droteo a kmal ungil.

‘Toki and Droteo know a lot about each other.’

e. Ng kmal ungil a klaubuch ȩr tir.

‘They have a very good marriage.’

As expressions like klasoes ȩr kid ‘our relationship’, klaubuch

ȩr tir ‘their marriage’, etc. show, derived reciprocal nouns are
un possessible—that is, they cannot take possessor suffixes
(cf. 3.8). Therefore, if a possessor is to be mentioned, it must be
introduced with a noun phrase of possession containing the
relational word ȩr (e.g. ȩr kid ‘of us’, ȩr tir ‘of them’, etc.).

8.5. ABSTRACT NOUNS DERIVED WITH KL(Ȩ)-

The word-initial consonant cluster kl- observed in all of the de-
rived nouns of 9 and 11 and in some of the derived nouns of 6
and 7 should be distinguished from what is clearly a separate
prefix kl(ȩ)-, which is also used to form abstract nouns from
state verbs. The prefix kl(ȩ)- may originally have consisted of
two parts (an initial element k- and the resulting state infix -l-
), but now it functions as a single unit. Some abstract nouns
derived with kl(ȩ)-, together with the corresponding state verb,
are listed below:

(13) Derived Noun Related State Verb

klȩkekȩre ‘smallness’ kekȩre ‘small’

klȩkakȩrous ‘difference’ kakȩrous ‘different’
klungiaol 15 ‘benefit, goodness’ ungil ‘good’
kldachȩlbai ‘skill’ dachȩlbai ‘skillful’
klȩmȩdȩnge ‘knowledge (from study, mȩdȩnge ‘know’
etc.)’
kldiull 16 ‘pregnancy’ dioll ‘pregnant’

220
 Palauan Reference Grammar

klȩmȩra ‘truth’ mȩra ‘true’

klȩngȩltȩngat ‘good fortune’ ngȩltȩngat
‘fortunate’
klȩngar ‘existence’ ngar ‘exist, be
(located)’

As the examples in 13 illustrate, kl(ȩ)- is usually attached to

simple state verbs, but in a few cases it is prefixed to complex
state verbs like mȩdȩnge ‘know’ and mȩra ‘true’, which consist
of the verb marker prefix mȩ- and a bound verb stem.
The derived nouns listed in 13 are typically used in sen-
tences like the following:

(14) a. A omȩsuub ȩl tȩkoi ȩr a Merikel a klungiolek.

‘Studying English is to my benefit.’

b. Ng ua ngara a klȩmȩdȩnge 17 ȩr kau ȩr a tȩkoi ȩr a

Siabal?
‘How much Japanese do you know?’

c. A klȩngȩltȩngȩtel a Droteo, e ng mlo sȩbȩchel ȩl mo ȩr

a Guam.
‘Droteo has had the good fortune of becoming able to
go to Guam.’

The prefix kl(ȩ)- can also be added to certain nouns (usually

human) to form another noun with a more abstract meaning, as
in the examples below:

(15) Derived Noun Related Noun

klȩchȩlid ‘religion’ chȩlid ‘god’

klsȩchal ‘manhood’ sȩchal ‘man’
klȩchad ‘human life, way of life’ chad ‘person’
klȩngalȩk ‘childhood’ ngalȩk ‘child’
klsensei ‘being a teacher’ sensei ‘teacher’
kltoktang ‘being a doctor’ toktang ‘doctor’
klobak 18 ‘chiefs of a community’ rubak ‘old man’

klodam ‘relationship between male odam ‘brothers’

relatives’
klodos ‘relationship between female odos ‘sisters’
relatives’

221
 8 Noun Derivation

Some sample sentences containing the nouns of 15 are

given below:

(16) a. A klȩchad er a elȩcha ȩl taem a diak lua ngar ȩr a mong.

‘Life these days isn’t like what it was a few years ago.’

b. A klsensei a diak lȩkirek ȩl ureor.

‘Being a teacher isn’t meant for me.’/‘I’m not suited to
being a teacher.’

8.6. INSTRUMENT AND ACTION NOUNS WITH O-

The prefix o- can be attached to verbs in two different ways,
thus deriving two classes of nouns, one concrete in meaning
and the other abstract (cf. 2.2). The o- to be discussed in this
section should not be confused with another prefix o-, which,
as we saw in 5.4 and chap. 6, is a variant of the verb marker.
Even though the two prefixes o- are homonymous—i.e., iden-
tical in sound (or form)—they are distinct in function, since one
(the verb marker) derives verbs and the other derives nouns.
Palauan instrument nouns are derived from the imper-
fective forms of transitive verbs simply by replacing the verb
marker prefix mȩ- by the prefix o-. Instrument nouns are con-
crete nouns which designate the tool, implement, or utensil
used in performing a particular action. For example, from the
imperfective verb mȩles ‘cut, slice’, we derive the noun oles
‘knife’, which is an instrument used for cutting. Some
commonly-used Palauan instrument nouns, together with the
corresponding imperfective verb form, are provided below:

(17) Derived Instrument Related Transitive Verb (in

Noun imperfective form)

olaml ‘grass cutter, mȩlaml ‘cut (grass)’

machete’
oluchȩs ‘pencil’ mȩluchȩs ‘write’
oriik ‘broom’ mȩriik ‘sweep’
olamk ‘razor’ mȩlamk ‘shave’
onges ‘grater’ mȩnges ‘scrape’
ongimd ‘something to mȩngimd ‘cut (hair)’
cut with’
olasȩch ‘axe, adze’ mȩlasȩch ‘chop, carve’

222
 Palauan Reference Grammar

osib ‘plow’ mȩsib ‘plow’

osongd ‘comb’ mȩsongd ‘comb’
osaur ‘something to mȩsaur ‘tie’
tie with’

Since the instrument nouns above are derived from the

imperfective forms of transitive verbs 19 by replacing the verb
marker mȩ- with o-, they have the basic structure (prefix) o- +
imperfective marker + verb stem.
Palauan action nouns are derived simply by adding the
prefix o- to transitive or intransitive action verbs. These nouns
designate actions or activities as abstract concepts and are used
in certain grammatical constructions where nouns are required.
In 18 below, the action nouns are derived from transitive action
verbs, while those in 19 are derived from intransitive action
verbs:

(18) Derived Action Noun Related Transitive Verb (in

imperfective form)

omȩluchȩs ‘writing’ mȩluchȩs ‘write’

omȩlim ‘drinking’ mȩlim ‘drink’

omȩnga ‘eating’ mȩnga ‘eat’
omȩsuub ‘studying’ mȩsuub ‘study’

(19) Derived Action Noun Related Intransitive Verb

omilil ‘playing’ milil ‘play’

omȩrael ‘travelling, mȩrael ‘walk, travel’
trip’
omȩngȩdub ‘swimming’ mȩngȩdub ‘swim’

In 19, the prefix o- has been added to intransitive action verbs

which have the verb marker prefix mȩ- (or m-). If an intransitive
verb contains the metathesized verb marker -(ȩ)m- or -o- (cf.
6.2), it cannot be used to derive an action noun with o-. Thus,
from rȩmurt ‘run’ or rȩmos ‘drown’, we cannot derive nouns
such as *orȩmurt ‘running’ or *orȩmos ‘drowning’, etc.
The derived action nouns of 18 and 19 are used in sentences
like the following:

223
 8 Noun Derivation

(20) a. A omȩluchȩs ȩl tȩkoi ȩr a Siabal a kmal mȩringȩl.

‘Writing Japanese is very difficult.’

b. Ng soam a omȩlim ȩl rrom?

‘Do you like to drink liquor?’

c. Tia a delmȩrab ȩr a omȩsuub.

‘This is a room for studying.’

d. A omȩrolek ȩl mo ȩr a Siabal a kmal mle ungil.

‘My trip to Japan was great.’

In 20a, 20b, and 20d, the derived action nouns omȩluchȩs

‘writing’, omȩlim ‘drinking’, and omȩrolek ‘my trip’ (a pos-
sessed form of omȩrael ‘trip’) are part of the subject noun
phrases omȩluchȩs ȩl tȩkoi ȩr a Siabal ‘writing Japanese’,
omȩlim ȩl rrom ‘drinking liquor’, and omȩrolek ȩl mo ȩ r a
Siabal ‘my trip to Japan’. For more information about sentences
like 20b, where the subject noun phrase has been shifted to
the end of the sentence, consult 17.8. In 20c, the action noun
omȩsuub ‘studying’ is part of the characterizational phrase
ȩr a omȩsuub ‘of studying’ (cf. 3.8–9), which is a kind of rela-
tional phrase that characterizes the preceding noun (delmȩrab
‘room’) by describing its function.

8.7. NOUNS DERIVED WITH UL(Ȩ)-

The prefix ul(ȩ)-, which might be related to the past tense prefix
ul(ȩ)- discussed in 5.3.2, is used to derive nouns that refer to
what is left over from the activity specified by the related verb,
generally waste products of some kind. Derived nouns of this
type also require the imperfective marker and—for some un-
known reason—the anticipating state suffix -(ȩ)l (or long vowel
+ l); furthermore, they involve several kinds of complex pho-
netic changes. As the examples below indicate, these nouns can
only be derived from transitive verb stems:

(21) Derived Noun Related Transitive

Verb (in imperfective
form)

ulȩngmúdȩl ‘hair that is cut off’ mȩngimd ‘cut (hair)’

224
 Palauan Reference Grammar

ullȩmáchȩl ‘what is left over from mȩlamȩch ‘chew’

chewing’
ulȩllúmȩl ‘disposable container mȩlim ‘drink’
after contents have
been drunk’
ullȩsóngȩl ‘bones, etc. left over mȩles ‘slice’
from fish or meat’
ulȩnguótȩl ‘debris from clearing mȩngiut ‘clear
ground’ (ground)’
ullȩbákȩl ‘wood shavings’ mȩlabȩk ‘smooth
(wood)’
ulȩngáll ‘inedible parts of food’ mȩnga ‘eat’
urrȩdíil ‘remaining stalk’ mȩrad ‘pick
(flowers)’

All of the derived nouns in 21 are stressed on the syllable

preceding the anticipating state suffix. As we saw in 7.8, certain
full vowels found in the basic stem of a verb characteristically
turn up in anticipating state verbs, where they come to appear
in stressed syllables. Exactly the same phenomenon is observed
in the forms above: thus, if we propose a basic stem like kimud
(cf. 7.8, ex.31) for ‘cut’, we can explain forms like ulȩngmúdȩl
‘hair that is cut off’ vs. mȩngímd ‘cut (hair)’ in the following
way. In ulȩngmúdȩl, the i of the basic stem kimud is deleted in
an unstressed syllable, while the u is maintained in a stressed
syllable. In other words, ulȩngmúdȩl is derived by the following
steps:

(22) ulȩ + ng (basic form = prefix ul(ȩ)- + imperfective

+ kimúd marker ng + verb stem kimud + anticipating
+ ȩl state suffix -(ȩ)l) →
ulȩ + ng (by deletion of verb-stem-initial consonant
+ imúd + following imperfective marker) →
ȩl
ulȩ + ng (by deletion of i in unstressed syllable)
+ múd +
ȩl

By contrast, (imperfective) mȩngímd ‘cut (hair)’ involves exactly

the reverse situation: in this word, the i of the basic stem kimud
is retained because it is stressed, while the u is lost completely
in an unstressed syllable. The derivation of mȩngímd is shown
below:

225
 8 Noun Derivation

(23) mȩ+ ng (basic form = verb marker prefix mȩ- +

+ kímud imperfective marker ng + verb stem kimud) →
mȩ + ng (by deletion of verb-stem-initial consonant
+ ímud following imperfective marker) →
mȩ + ng (by deletion of u in unstressed syllable).
+ ímd

Some of the other phonetic changes observed in the forms

of 21 should also be familiar to us: for example, the full vowel
e of the verb stem des ‘cut’ is reduced to ȩ (schwa) when un-
stressed in a form like ullȩsóngȩl ‘leftover bones’. Or, in urrȩdíil
‘remaining stalk’, the consonant cluster rr is due to assimilation.
There are, however, certain phonetic alternations in 21 which
we cannot explain and which seem to be irregularities: thus, the
appearance of the full vowel o in ullȩsóngȩl ‘leftover bones’ is
quite unexpected, since the related verb stem des ‘cut’ has no
final vowel; and the alternation between the vowel clusters uo
and iu in ulȩnguótȩl ‘debris from clearing ground’ vs. mȩngiút
‘clear (ground)’ is very unusual.

226
 9 Causative Verbs

9CausativeVerbs
9.1. MEANING AND USE OF CAUSATIVE VERBS
As we saw in 5.1.1, all Palauan action verbs are either tran-
sitive or intransitive. Transitive action verbs name actions
which are done to or directed at someone or something; the
person who performs or brings about the action appears as the
sentence subject, while the person, animal, or thing which re-
ceives the effect of the action appears as the sentence object. By
way of review, observe the following sentences with transitive
action verbs; you should have no difficulty identifying the sen-
tence subject and sentence object:

(1) a. A Toki a mirruul a kall.

‘Toki prepared the food.’

b. A sensei a chillȩbȩdii a bilis.

‘The teacher hit the dog.’

c. Ak mo omes ȩr a Tony.
‘I’m going to see Tony.’

Intransitive action verbs, by contrast, involve just a doer, but

no receiver; in other words, they describe actions which by
their very nature cannot be directed at someone or something,
but which only the doer himself can pursue. Therefore, the
following sentences with intransitive action verbs contain sub-
jects, but no objects:

(2) a. A ngalȩk a lilangȩl.

‘The child was crying.’

b. Ak mo oureor ȩr a klukuk.
‘I’m going to work tomorrow.’

There is a special class of transitive action verbs known as

causative verbs. As we will see below, these verbs can be iden-
tified by the presence of the causative prefix, which has quite
a few different forms (omeȩ(k)-, ol(ȩ)-, etc.). These verbs are

227
 9 Causative Verbs

called causative because they involve actions in which the doer

(or subject) causes or forces someone or something to perform
a particular action or be in a particular state. In order to under-
stand this characteristic meaning of causative verbs, compare
the following two sentences:

(3) a. A bilis a chȩmiis.

‘The dog is running away.’

b. A Droteo a olȩchiis ȩr a bilis.

‘Droteo is {chasing the dog away./making the dog run
away.}’

Example 3a is a simple intransitive sentence containing the in-

transitive action verb chȩmiis ‘run away, escape’; this verb con-
tains the metathesized verb marker -(ȩ)m- (cf. 6.2), which is
infixed after the initial consonant (ch) of the verb stem chiis
‘escape’. Example 3b, on the other hand, is a transitive sentence
containing the (transitive) causative verb olȩchiis, which con-
sists of the causative prefix olȩ- followed by the verb stem chiis.
The intransitive sentence 3a has a subject only (bilis ‘dog’),
while the transitive sentence 3b has both a subject (Droteo)
and an object (bilis ‘dog’). Notice that the subject of the intran-
sitive sentence has become the object of the transitive sentence.
Common to the meaning of both 3a and 3b is the information
that the dog is running away, but distinguishing the two sen-
tences from each other is the extra information given in 3b:
here, the presence of the causative verb olȩchiis tells us that
some person (Droteo) is making the dog run away, or causing
him to run away.
The following pair of sentences can be analyzed in exactly
the same way:

(4) a. A ngalȩk a mȩkar.

‘The child is awake.’

b. A rȩdil a olȩkar ȩr a ngalȩk.

‘The woman is waking up the child.’

Example 4a is a simple intransitive sentence containing the (in-

transitive) state verb mȩkar ‘be awake’; the subject of the sen-
tence (ngalȩk ‘child’) is described as being in the particular
state designated by the verb. Example 4b, however, is a tran-

228
 Palauan Reference Grammar

sitive sentence which names an action: this action is denoted

by olȩkar ‘wake up (someone), cause (someone) to be awake’,
which is a causative verb formed from the causative prefix ol(ȩ)-
and the verb stem kar. In this sentence, the subject of olȩkar
(rȩdil ‘woman’) is doing something to the object (ngalȩk ‘child’)
in order to make him wake up—that is, she is causing the child
to be in the particular state designated by the corresponding
state verb mȩkar ‘be awake’. Again, the subject of the intran-
sitive sentence 4a has become the object of the transitive sen-
tence 4b.

9.2. FORMS OF THE CAUSATIVE PREFIX

Since causative verbs are a subtype of transitive action verbs,
they exhibit the same kinds of distinctions observed among
transitive action verbs. In other words, causative verbs have
both imperfective and perfective forms (cf. 5.5 and see 9.4
below), as well as ergative forms (cf. 5.4 and see 9.5 below)
and hypothetical forms (cf. 4.10 and 4.10.1–9 and see 9.6
below). In discussing the variants of the Palauan causative
prefix, we will first concentrate on the imperfective forms of
causative verbs.
In order to derive the imperfective forms of causative verbs,
one of the two causative prefixes omȩ(k)- or ol(ȩ)- is added to
a verb stem. These prefixes are added primarily to the stems
of intransitive verbs according to the following general rule:
omȩ(k)- is prefixed to the stems of intransitive state verbs,
while ol(ȩ)- is prefixed to the stems of intransitive action verbs.
In addition, the prefix omȩ(k)- can occur with the stems of a
few transitive action verbs (see 9.2.1.2 below). Regardless of
whether the verb stem following the causative prefix is tran-
sitive or intransitive, the derived causative verb in omȩ(k)- or
ol(ȩ)- is always transitive. Further, as we will see in 9.3 below,
a small number of rather exceptional verb stems can occur
with both prefixes, sometimes resulting in a slight difference in
meaning.

229
 9 Causative Verbs

9.2.1. THE PREFIX omȩ(k)-

In the list below we can see some typical causative verbs whose
imperfective forms are derived by prefixing omȩ(k)- to the stems
of intransitive state verbs. In the right hand column, the re-
lated state verb is given for purposes of comparison:

(5) Causative Verb in omȩ(k)- Related State Verb

omȩkdȩchor 1 ‘make…stand’ dȩchor ‘stand,

standing’
omȩkungil ‘heal, make… better’ ungil ‘good’
omȩkbȩches ‘renovate, repair, bȩches ‘new’
make…new’
omȩkdȩkimȩs ‘make…wet’ dȩkimȩs ‘wet’
omȩkȩsiu 2 ‘compare, imitate’ osiu ‘joining’
omȩkoad 3 ‘kill’ mad ‘dead’
omȩkikiongȩl ‘make…dirty’ kikiongȩl ‘dirty’
omȩkarȩd ‘light, turn on’ kmarȩd ‘lighted, on
fire’
omȩkdakt ‘frighten’ mȩdakt ‘afraid’
omȩkdirt ‘dry out’ mȩdirt ‘dry’
omȩkdingȩs ‘satisfy, make… full’ mȩdingȩs ‘full’
omȩkringȩl ‘hurt, make… difficult’ mȩringȩl ‘painful,
difficult’

As causative verbs like omȩkikiongȩl ‘make…dirty’ and

omȩkarȩd ‘light, turn on’ show, the k of the causative prefix
omȩ(k)- is deleted if it is followed by a k- initial verb stem. As we
will see in some later examples, the k of omȩ(k)- is also lost if
the following verb stem begins with ng. Thus, we can formulate
the following general phonetic rule: the velar stop k (cf. 1.3.1)
of the causative prefix is automatically deleted before another
velar consonant (k or ng).
You will notice that the related state verbs given in 5 above
are of several different types. Verbs like dȩchor ‘stand,
standing’, ungil ‘good’, etc. are simple state verbs which do
not exhibit the verb marker (cf. 7.1). Verbs like mȩdakt ‘afraid’,
mȩdirt ‘dry’, and kmarȩd ‘lighted, on fire’, however, are more
complex in structure, since they consist of the verb stem and
the verb marker, which appears as a prefix mȩ- or as a metathe-
sized element -(ȩ)m- (cf. 6.2).

230
 Palauan Reference Grammar

Interestingly enough, the verb markers mȩ- and -(ȩ)m- found

in state verbs like mȩdakt, mȩdirt, and kmarȩd do not appear
in this form in the corresponding causative verbs omȩkdakt,
omȩkdirt, and omȩkarȩd: in other words, causative verb forms
like *omȩk mȩdakt, *omȩkmȩdirt, and *omȩkmarȩd are impos-
sible. The nonexistence of such forms is probably due to the fact
that omȩ(k)-, which we have been treating as a single unitary
prefix, is actually a combination of several prefixes. More specif-
ically, if the o- of omȩ(k)- is really one of the variants of the
verb marker, 4 as suggested in 9.2.1.1 below, then forms like
*omȩkmȩdakt, *omȩkmarȩd, etc. would be prevented because
they would contain a second unnecessary instance of the verb
marker.

*9.2.1.1. Technical Discussion of the Prefix

omȩ(k)-
Although we have been dealing with the causative prefix
omȩ(k)- as a single unit (or morpheme), there are several facts
which lead us to believe that it is really a combination of three
elements. Before discussing these facts, we first need to ex-
amine the basic structure of omȩ(k)-, which is represented by
the following formula:

(6) Verb marker + imperfective marker + causative marker

o + m + bȩk

As you can see, the basic structure of omȩ(k)- consists of two

elements with which we are already familiar—the verb marker
and the imperfective marker—and one element which is new
to us—the causative marker bȩk. As discussed in 5.5, the im-
perfective marker has several variants—l, ng, or m—depending
on the initial consonant of the following verb stem. Even
though the causative marker bȩk is not a verb stem, its initial
consonant b nevertheless determines the correct variant of the
directly preceding imperfective marker, which is therefore rep-
resented as m in 6 above. The verb marker which is part of
the structure of 6 appears as o- (rather than mȩ-) as a result of
dissimilation (cf. 6.1); here, the dissimilation is caused by the
presence of the bilabial consonant b in the following causative
marker. The actually-pronounced form omȩ(k)- is derived from
the basic structure o + m + bȩk by deletion of the initial con-
sonant b of the causative marker: this is exactly the same phe-

231
 9 Causative Verbs

nomenon which we observed in the derivation of imperfective

verb forms (cf. 5.5), where the initial consonant of a verb stem
is characteristically deleted following the imperfective marker
(e.g. mȩ + l + dasȩch → mȩlasȩch ‘carve’, mȩ + ng + chuiu →
mȩnguiu ‘read’, etc).
There are several facts which we can bring up as evidence
that the formula given in 6 correctly represents the basic
structure of omȩ(k)-. First, we have already mentioned above
that the non-occurrence of causative forms like *omȩkmȩdakt
or *omȩ kmarȩd (instead of omȩkdakt ‘frighten’ and omȩkarȩd
‘light, turn on’, respectively) could be easily explained if omȩ(k)-
itself already contained the verb marker, as in 6: in other words,
the non-occurrent causative forms would be incorrect because
they would contain two instances of the verb marker—i.e.,
word-initial o- and word-internal -mȩ- or -m-.
Second, the formula in 6 allows us to account for certain
ergative causative verb forms in a regular and consistent way.
As we will see in 9.5 below, Palauan causative verbs in omȩ(k)-
derive their ergative forms according to two different patterns.
One of these patterns, which is close to becoming archaic (i.e.,
has nearly disappeared from standard use), results in ergative
forms which begin with obȩ(k)-. For example, imperfective
omȩkdakt ‘frighten’ has ergative obȩkdakt ‘be/get frightened’,
imperfective omȩkarȩd ‘light, turn on’ has ergative obȩkarȩd
‘be/get lighted/turned on’, etc. As discussed in 5.4, one defining
feature of Palauan ergative verb forms is that they lack the im-
perfective marker and consist merely of the verb marker fol-
lowed by the verb stem (e.g. mȩ + dasȩch ‘be/get carved’, mȩ
+ chuiu ‘be/get read’, etc.). We should therefore expect that the
ergative forms of causative verbs would also lack the imper-
fective marker, and this is precisely the case: in other words,
the obȩk- of obȩk dakt ‘be/get frightened’, etc. consists of the
following simple sequence:

(7) Verb marker + causative marker

o + bȩk

As the formulas in 6 and 7 clearly show, the only difference be-

tween the imperfective and ergative forms of causative verbs is
that the former contain the imperfective marker, while the latter
do not.

232
 Palauan Reference Grammar

Third, as we will see in 9.4 below, the perfective forms of

causative verbs in omȩ(k)- normally do not show an initial o-.
For example, the 3rd pers. sg. object present perfective form
(cf. 6.3) of omȩkdakt ‘frighten’ is mȩkdȩktii ‘frighten him/her/it’.
The derivation of this perfective form can be explained in a plau-
sible way if we assign it a basic form whose elements are con-
sistent with those given in 6 and 7. Recall (cf. 5.5) that Palauan
perfective verb forms of course do not contain the imperfective
marker; therefore, a perfective form like sosȩbii ‘burn it’ has
the basic structure verb marker + verb stem + object pronoun.
The basic structure of the perfective forms of causative verbs is
similar, except that the causative marker is added following the
verb marker. Thus, the basic form of mȩkdȩktii ‘frighten him/
her/it’ is represented as follows:

(8) Verb + causative + verb + object

marker marker stem pronoun
o + bȩk + dakt + íi

By applying a sequence of phonetic rules to the basic form o +

bȩk + dakt + í i, we can derive the actually-spoken form mȩk
dȩktíi, as explained below.
In 6.3.1 we noted that Palauan perfective verb forms are
characterized by the fact that the verb marker metathesizes
to a position following the initial consonant of the verb stem. It
seems reasonable to assume that such metathesis also occurs
in the perfective forms of causative verbs. In the example under
discussion, application of verb marker metathesis to the basic
form given in 8 will yield something like b + m + ȩk + dakt +
íi; here, the metathesized verb marker has been moved to a po-
sition following the initial consonant b of the causative marker,
and for reasons which will become clear when we discuss the
next phonetic rule, we assume that the metathesized verb
marker appears as -m-. Next, we apply a rule which we have
already encountered in 6.2.1: the b of the causative marker,
which has come to appear before the m of the metathesized
verb marker, must be deleted, resulting in m + ȩk + dakt + íi.
After one more rule is applied—namely, the rule reducing un-
stressed a to ȩ (cf. 6.4)—the actually-spoken form mȩkdȩktíi is
produced. The phonetic changes described above are summa-
rized in the following step-by-step derivation:

233
 9 Causative Verbs

(9) o + bȩk + (basic form = verb marker + causative

dakt + íi marker + verb stem + object pronoun)
b + m + ȩk (by metathesis of verb marker)
+ dakt + íi
m + ȩk + (by deletion of b before m)
dakt + íi
m + ȩk + (by reduction of a to ȩ in unstressed syllable)
dȩkt + íi

9.2.1.2. Additional Types of Causative Verbs

with omȩ(k)-
In 9.2.1 above, we listed a large number of causative verbs
which are derived by prefixing omȩ(k)- to the stems of intran-
sitive state verbs. In this section, we will look at two further
patterns of derivation for causative verbs in omȩ(k)-.
A small number of causative verbs can be derived by pre-
fixing omȩ(k)- to the stems of transitive action verbs. The
most common examples are listed below, together with the re-
lated transitive verb:

(10) Causative Verb Related Transitive Verb (in

imperfective form)

omȩka ‘feed, make…eat’ mȩnga ‘eat’

omȩngim ‘make…drink’ mȩlim ‘drink’
omȩngamȩch 5 mȩlamȩch ‘chew, smoke’
‘make…chew, make…
smoke’

In each of the causative verbs of 10, the velar stop k of omȩ(k)-

is deleted because the following verb stem (kal 6 ‘eat’, ngim
‘drink’, and ngamȩch ‘chew, smoke’) begins with another velar
consonant (k or ng).
Since a sentence with a transitive verb has both a subject
noun phrase and an object noun phrase, a sentence with the
corresponding causative verb will contain, in addition to these
noun phrases, a third noun phrase which identifies the person
who is causing the action to occur. In the following transitive
sentence, for example,

(11) A rȩngalȩk a mȩnga ȩr a kukau.

234
 Palauan Reference Grammar

‘The children are eating the taro.’

rȩngalȩk ‘children’ and kukau ‘taro’ are the subject and object,
respectively, of the transitive verb mȩnga ‘eat’. Now observe
what happens to these noun phrases in the corresponding
causative sentence:

(12) A Romana a omȩka ȩr a rȩngalȩk ȩr a kukau.

‘Romana is feeding the children the taro.’

Here, both rȩngalȩk ‘children’ and kukau ‘taro’ have come to

appear as objects following the causative verb omȩka ‘feed’;
and the new subject is Romana, the person who is doing the
feeding. It is also possible to have sentences like 12 in which the
second object is not mentioned—namely,

(13) A Romana a omȩka ȩr a rȩngalȩk.

‘Romana is feeding the children.’

This sentence says that Romana is giving the children some-

thing to eat, but it does not specify what it is they are eating.
A few causative verbs can also be derived from nouns by
prefixing omȩ(k)-. The resulting causative verb, which is of
course transitive, designates some action or activity which in-
volves the related noun. For example, from bar ‘blanket’, we
can derive causative omȩkbar ‘cover…with a blanket’, and from
buch ‘spouse’, we can form causative omȩkbuch ‘marry (i.e.,
join as husband and wife), marry off, mate (animals)’.

9.2.1.3. Sample Sentences with omȩ(k)-

Causatives
Although we have already presented some sentences in 9.1
above to illustrate the meaning and use of causative verbs,
perhaps some further examples are desirable. Thus, the sen-
tences below illustrate the use of some of the causative verbs in
omȩ(k)- listed in 5:

(14) a. A Toki a omȩkdȩchor ȩr a ngalȩk ȩr a bebul a

tebȩl.
‘Toki is making the child stand on the table.’

b. A rȩchad a mo omȩkoad a ngikȩl.

235
 9 Causative Verbs

‘People are going to kill the fish.’

c. A Toki a blȩchoel ȩl mukdakt 7 ȩr a dȩrumk.

‘Toki always gets frightened by thunder.’

d. A toktang a mo omȩkungil ȩr kau.

‘The doctor will get you better.’

e. A rȩkangkodang a omȩkikiongȩl ȩr a kȩdȩrang.

‘The tourists are messing up the beach.’

f. A Droteo a omȩkarȩd ȩr a olbidȩl.

‘Droteo is lighting the lamp.’

g. A kall a mo omȩkdingȩs ȩr kau.

‘The food will fill you up.’

9.2.2. THE PREFIX ol(ȩ)-

In the list below we find some representative causative verbs
whose imperfective forms are derived by prefixing ol(ȩ)- to the
stems of intransitive action verbs. In the column at the right,
the related action verb is provided for purposes of comparison:

(15) Causative Verb in ol(ȩ)- Related Action Verb

ollangȩl ‘make…cry’ lmangȩl ‘cry’

olluut ‘give back, make… lmuut ‘return, come back’
return’
oltobȩd ‘take out’ tuobȩd ‘go out, emerge’
oltengȩl ‘take down (from mȩtengȩl ‘land, come down’
above)’
oltuu ‘put into, make… enter’ tmuu ‘enter’
olsisȩb ‘put into, make… soisȩb ‘enter’
enter’ 8
olsebȩk ‘make…fly’ suebȩk ‘fly’
olsobȩl ‘save, take care of’ suobȩl ‘survive, be saved’
oldik ‘banish’ dmik ‘go into exile’
oliuul ‘transfer’ imuul ‘go from one location
to another’
olȩchiis ‘chase, make…run chȩmiis ‘run away, escape’
away’
olȩkerȩd ‘unload, let off’ kmerȩd ‘get off, get out’

236
 Palauan Reference Grammar

olȩngelt ‘(make…) sink (in ngmelt ‘sink (in soft

soft ground)’ ground), set (sun)’
olȩngasȩch ‘raise, lift up’ ngmasȩch ‘climb, rise (sun)’
orrebȩt ‘drop’ ruebȩt ‘fall’
orros ‘(make…) drown’ rȩmos ‘drown’
orrurt ‘make… run’ rȩmurt ‘run’
orrȩchorȩch ‘(make…) sink ruchorȩch ‘sink (in water)’
(in water)’

As causative verbs like olȩchiis ‘chase, make…run away’,

olȩkerȩd ‘unload, let off’, and olȩngasȩch ‘raise, lift up’ show,
the causative prefix must include ȩ if the following verb stem
begins with a velar consonant (k, ng) or a glottal consonant
(ch). Otherwise, the causative prefix appears as ol-, except in
forms like orrebȩt ‘drop’, orros ‘(make…) drown’, etc., where
the l of the prefix completely assimilates to (i.e., becomes iden-
tical with) the initial r of the following stem.
As we noted in 6.2, many Palauan intransitive verbs contain
the metathesized verb marker, which appears as an infix
of the form -(ȩ)m-, -u-, or -o-. All of the intransitive action
verbs listed in 15 above, with the exception of mȩtengȩl ‘land,
come down’, exhibit the metathesized verb marker. The variants
of the metathesized verb marker found in intransitive action
verbs such as dmik ‘go into exile’, suebȩk ‘fly’, and soisȩb
‘enter’ do not occur as such in the corresponding causative
verbs oldik ‘banish’, olsebȩk ‘make…fly’, and olsisȩb ‘put into,
make….enter’: that is, it is impossible to have causative verb
forms like *oldmik, *olsuebȩk, or *olsoisȩb. The phenomenon
being described here, you will recall, is exactly the same as that
discussed in 9.2.1 above with reference to the intransitive state
verbs of 5 and their corresponding causative verbs. Our expla-
nation for the nonexistence of causative verb forms like *oldmik,
*olsuebȩk, and *olsoisȩb will also parallel that given in 9.2.1: we
propose that ol(ȩ)-, which we have so far assumed to be a single,
unitary prefix, is really a combination of two prefixes, one of
which is the verb marker. Therefore, the nonexistent causative
verb forms under discussion are precluded because they would
contain an extra, unnecessary occurrence of the verb marker.

237
 9 Causative Verbs

*9.2.2.1. Technical Discussion of the Prefix

ol(ȩ)-
Though we have been treating the causative prefix ol(ȩ)- as a
single morpheme, in reality it is probably a combination of two
elements. Some question exists, however, as to which elements
are actually involved, and several different analyses could be
proposed, each presenting its own difficulties and irregularities.
Therefore, the explanations of ol(ȩ)- given below are very tenta-
tive and reflect our inadequate knowledge of the structure of
this prefix.
In order to explain the non-occurrence of causative verb
forms like *oldmik, etc., we suggested at the end of 9.2.2 above
that ol(ȩ)- actually contains the verb marker. Thus, ol(ȩ)- ap-
pears to consist of the variant o- of the verb marker and the
variant -l-of the imperfective marker. Some evidence that -l- is
the imperfective marker is found in the fact that this consonant
is missing in those verb forms whose basic structure character-
istically lacks the imperfective marker. For example, the per-
fective forms of causative verbs in ol(ȩ)- would not be expected
to contain the imperfective marker, and in fact such forms do
not show -l-: thus, we have perfective ochisii ‘chase him/her/
it’ from olȩchiis ‘chase’, perfective osȩbȩkii ‘make it fly’ from
olsebȩk ‘make… fly’, and so forth. In addition, the ergative
forms of causative verbs in ol(ȩ)-would not contain the imper-
fective marker, and these, too, lack -l-: thus, motuu ‘be made to
enter’ is the ergative form of oltuu ‘put into, make…enter’, etc.
If we assume that ol(ȩ)- consists of the verb marker o- and
the imperfective marker -l-, as described above, we run into
certain problems that cannot be explained. First, we have seen
in 6.1 and in 9.2.1.1 above that the verb marker appears as o-
only when a b appears in the following verb stem or causative
marker. In many of the causative verbs listed in 15, there is no
b at all, yet the verb marker invariably appears as o-. Why this
should be so remains a mystery. Second, we noted in 5.5.a that
the imperfective marker takes the variant -l- only when it pre-
cedes verb stems which begin with the consonants t, d, s, l, or
ng. This rule does not hold for causative forms, however, since
-l- can precede k-initial or ch-initial verb stems in words like
olȩkerȩd ‘unload, let off’ and olȩchiis ‘chase’. Even more strik-
ingly, the verb-stem-initial consonant is not deleted after the im-
perfective marker in causative verb forms; such deletion would

238
 Palauan Reference Grammar

be expected, since, as we saw in 5.5, it is one of the general pho-

netic rules which accounts for the correct form of imperfective
verbs. 9
From the discussion above, we can see that the two ele-
ments which make up the prefix ol(ȩ)- involve certain hard-
to-explain irregularities. Perhaps the most puzzling feature of
ol(ȩ)-, however, is its very structure: even though ol(ȩ)- contains
no special causative morpheme like the bȩk of omȩk- (cf. 9.2.1.1
above), but seems to consist merely of the sequence verb
marker + imperfective marker, it nevertheless has a causative
meaning. There is no way of accounting for this unusual fact
unless we modify our analysis of ol(ȩ)- in a significant way, as
suggested briefly below.
It might also be claimed that the structure of ol(ȩ)- involves
a causative marker o- followed by the imperfective marker -l-.
10
Under such an analysis, causative verbs in ol(ȩ)- would, rather
surprisingly, not contain any verb marker at all. In addition, we
would have difficulty understanding why the Palauan language
would have two phonetically divergent causative markers (o-
and -bȩk-). This analysis has a few advantages, however. First of
all, if the o- of ol(ȩ)- were a causative morpheme, then we could
explain the fact that ol(ȩ)- has a causative meaning. Second, we
would have a plausible way of explaining why initial o- remains
in perfective causative forms such as ochisii ‘chase him/her/
it’ (from olȩchiis ‘chase’), etc. If the o- of ol(ȩ)- were the verb
marker, then we would expect it to metathesize in perfective
verb forms (cf. 6.3.1); but, as forms like ochisii (as opposed to,
say, *choisii) show, it obviously does not. This difficulty would be
avoided if the o- of ol(ȩ)- were indeed not the verb marker, but
a causative marker, as suggested here. Third, if the o- of ol(ȩ)-
were the verb marker, ergative causative forms like motuu ‘be
made to enter’ (from oltuu ‘put into, make…enter’) would have
the odd feature of containing two instances of the verb marker
(m- and -o-). The structure of such forms would be more plau-
sible, however, if we analyzed the -o- as a causative marker.

9.2.2.2. Additional Types of Causative Verbs

with ol(ȩ)-
In 9.2.2 above, we saw that most causative verbs in ol(ȩ)- are
derived from the stems of intransitive action verbs. A small
number of causative verbs can be derived, however, by prefixing

239
 9 Causative Verbs

ol(ȩ)- to the stems of intransitive state verbs. A couple of ex-

amples of this type include olȩkeed ‘bring…near’ from kmeed
‘near’ and oldak ‘put together, unify’ from dmak ‘together’.
A few verbs are formed with the prefix ol(ȩ)- even though
they do not appear to have a causative meaning. Originally, they
were probably derived by combining the causative prefix ol(ȩ)-
with a following verb stem, but over a long period of time their
meanings have evolved so as to obscure their causative origin.
Some commonly-used verbs in this category include orrengȩs
‘hear, listen to’, oldingȩl ‘visit’, oltoir ‘chase’, olȩker ‘call’, and
olȩngȩseu ‘help’.

9.2.2.3. Sample Sentences with ol(ȩ)-

Causatives
In this section we will list several sentences illustrating the use
of some of the causative verbs in ol(ȩ)- given in 15 or mentioned
in 9.2.2.2:

(16) a. A rȩngalȩk a olsebȩk a kȩdam.

‘The children are flying kites.’

b. Ak mo olluut ȩr a Droteo mȩ ng diak lȩbo ȩr a chei.

‘I’m going to make Droteo come back from going
fishing.’

c. A ngalȩk a orrebȩt a blatong.

‘The child is dropping plates.’

d. A Cisco a olȩkeed ȩr a bilas.

‘Cisco is bringing the boat near.’

9.3. VERB STEMS ALLOWING TWO CAUSATIVES

A small number of intransitive verb stems allow the formation
of two different causative verbs, one in omȩ(k)- and the other
in ol(ȩ)-. The resulting causative verbs have slightly different
meanings, at least for some Palauan speakers. Because there
is much variation from person to person with regard to the
meaning, or even acceptability, of such causative verbs, the dis-
cussion below may not apply to all speakers.

240
 Palauan Reference Grammar

The contrast between ol(ȩ)- and omȩ(k)- often involves

whether or not the act of causation is intentional. In other
words, it is a question of whether or not the subject of the
causative verb means or intends the action or state to come
about. Note the use of the causative forms of suebȩk a rȩngul
‘worried’ 11 in the following sentences:

(17) a. A sensei a olsebȩk ȩr a rȩngul a Droteo.

‘The teacher is worrying Droteo (without meaning to).’

b. A sensei a omȩksebȩk ȩr a rȩngul a Droteo.

‘The teacher is worrying Droteo (deliberately).’

As the parenthesized parts of the English equivalents show,

ol(ȩ)- implies non-intentional causation, while omȩ(k)- in-
volves intentional causation.
A similar distinction in meaning is found between the two
causative verbs derived from songȩrengȩr ‘hungry’, as illus-
trated in the examples below:

(18) a. Ak olsȩngȩrengȩr ȩr a Toki e le ng dimlak a temek ȩl

mȩruul a kall.
‘I’m letting Toki go hungry because I didn’t have time to
prepare any food.’

b. Ak omȩksȩngȩrengȩr ȩr a Toki e le ng dȩngȩrengȩr.

‘I’m making Toki go hungry because she’s naughty.’

In 18a, the causative verb formed with ol(ȩ)- has a meaning of

non-intentional causation: here, the subject (ak ‘I’) is not pur-
posefully making Toki go without food; rather, there is no food
available, and this situation is beyond the subject’s control. By
contrast, the causative verb formed with omȩ(k)- in 18b has a
meaning of intentional causation and implies that the subject is
deliberately withholding food from Toki as a punishment.
The causative verbs orrael and omȩkrael, derived from the
stem rael ‘road’ (cf. intransitive mȩrael ‘walk, travel’), show a
somewhat different kind of meaning distinction, as illustrated in
the sentences below:

(19) a. A Droteo a orrael ȩr a kȩrȩbou.

‘Droteo is leading the water buffalo.’

241
 9 Causative Verbs

b. A Droteo a omȩkrael ȩr a ngȩlȩkel ȩl mo ȩr a ungil ȩl

klȩchad.
‘Droteo is guiding his child towards a good way of life.’

Though some speakers can use the two causatives inter-

changeably in sentences like 19a-b, the majority use orrael for
‘lead (an animal), drive (a boat, etc.)’ and omȩkrael for ‘guide’.

9.4. PERFECTIVE FORMS OF CAUSATIVE VERBS

In 9.2.1.1 and 9.2.2.1 above, we have already had occasion to
mention the perfective forms of causative verbs, although we
did not give a complete discussion of how they are derived. In
this section, therefore, we will provide the perfective forms of
a representative sample of causative verbs and explain the pho-
netic processes which account for their derivation. Our format
will follow that which we adopted in 6.3 and 6.3.1–3 for pre-
senting the perfective forms of “regular” (i.e., non-causative)
transitive verbs. Thus, in lists 22 and 24 below, the causative
verb is first given in its imperfective form, together with an
English gloss; then, four representative perfective forms will be
listed in the following order:

(20) Present tense, Past tense,

3rd pers. sg. object 3rd pers. sg. object
Present tense, Past tense,

3rd pers. pl. 3rd pers. pl.

non-human object non-human object

We will not provide English glosses for each perfective form,

since their meanings can easily be determined from the chart in
20; thus, the perfective forms omȩkdakt ‘frighten’, for instance,
would have the following English equivalents:

(21) ‘frighten him/her/it’ ‘frightened him/her/it’

‘frighten them’ ‘frightened them’

242
 Palauan Reference Grammar

In the list below, we observe the perfective forms of

causative verbs in omȩ(k)-:

(22)
a. omȩka ‘feed’: b. omȩkcharm
‘make…suffer’:
mȩkȩlii milȩkȩlii 12 mȩkchȩrmii
milȩkchȩrmii
mȩka milȩka mȩkcharm milȩkcharm

c. omȩkdakt ‘frighten’: d. omȩkdȩchor

‘make…stand’:
mȩkdȩktii milȩkdȩktii mȩkȩdȩchȩrur
milkȩdȩchȩrur
mȩkdakt milȩkdakt mȩkȩdȩchor
milkȩdȩchor

e. omȩkikiongȩl ‘make…dirty’: f. omȩkoad ‘kill’:

mȩkikingȩlii milkikingȩlii mȩkodir milkodir
mȩkikiongȩl milkikiongȩl mȩkoad milkoad

g. omȩngamȩch ‘make…chew/ h. omȩngim ‘give drink to’:

smoke’:
mȩngȩmȩchii mȩngȩlmii milȩngȩlmii
13
milȩngȩmȩchii
mȩngamȩch milȩngamȩch mȩngim milȩngim

You will notice that all of the perfective forms in 22 appear

to have lost the initial o- of omȩ(k)-. For practical purposes, we
can simply say that this o- must be deleted as part of the correct
derivation of the perfective forms under discussion; a more
technical analysis such as that given in 9.2.1.1 above, however,
more accurately reflects the complicated phonetic processes
which are at work here. Much less complicated is the derivation
of the past perfective forms of 22: we merely infix the past
tense marker -il- (cf. 5.3.2) after the initial m- of the perfective
verb form. We should also mention in passing how the past
tense forms of imperfective causative verbs in omȩ(k)- are
derived: as shown in 5.3.2.a, we replace word-initial o- with
ul(ȩ)-. Thus, we have, for example, omȩka ‘feed’—ulȩmȩka ‘fed’,
omȩkdakt ‘frighten’— ulȩmȩkdakt ‘frightened’, etc.

243
 9 Causative Verbs

The perfective forms in 22 show the effects of vowel re-

duction or vowel cluster reduction (cf. 1.4.4 and 6.4). Full
vowels or vowel clusters which appear in stressed syllables
in the imperfective causative verb are reduced, respectively, to
the neutral vowel ȩ (schwa) or to a single vowel when they
come to appear in un stressed syllables in certain perfective
causative forms. In the list below, compare the imperfective
causative forms in the left column with the corresponding 3rd
pers. sg. object present perfective forms in the right column;
in the latter forms the stress always appears on the object
pronoun suffix (cf. 4.9, 4.9.1–4, and 6.3.1), which has the form
-íi ‘him/her/it’ or, less frequently, vowel + r (e.g. -ír, -úr, etc.):

(23) Imperfective Form of 3rd pers. sg. object Present

Causative Verb Perfective Form

omȩká ‘feed’ mȩkȩlíi ‘feed him/her/it’

omȩkchárm mȩkchȩrmíi ‘make him/her/it
‘make…suffer’ suffer’
omȩkdȩchór mȩkȩdȩcȩrúr ‘make him/her/it
‘make…stand’ stand’
omȩkikióngȩl mȩkikingȩlíi ‘make it dirty’
‘make…dirty’
omȩkoád ‘kill’ mȩkodír ‘kill him/her/it’

By looking at the italicized vowels or vowel clusters in each of

the above pairs of verb forms, we can clearly see the processes
of reduction and how they are related to the position of the
stress in the word. In contrast to the perfective forms of 23, full
vowels and vowel clusters remain unchanged in the 3rd pers. pl.
object perfective forms given in 22. This is due to the fact that in
such forms, the object pronoun suffix is zero (Ø) (cf. 4.9.1) and
therefore the full vowels or vowel clusters in question remain
stressed (e.g. mȩká + Ø ‘feed them’, mȩkchárm + Ø ‘make them
suffer’, etc.).
In the list below, we observe the perfective forms of
causative verbs in ol(ȩ)-:

(24) a. olȩchiis ‘chase’: b. olȩkiis ‘wake up’:

ochisii ulȩchisii okisii ulȩkisii
ochiis ulȩchiis okiis ulȩkiis

244
 Palauan Reference Grammar

c. ollangȩl d. olluut ‘give back, make..

‘make..cry’: return’:
olȩngȩlii ullȩngȩlii olutii ullutii
olangȩl ullangȩl oluut ulluut

e. olsebȩk ‘make..fly’: f. orrebȩt ‘drop’:

osȩbȩkii ulsȩbȩkii orȩbȩtii urrȩbȩtii
osebȩk ulsebȩk orebȩt urrebȩt

In all of the perfective forms of 24, the -l- of ol(ȩ)- has been
lost; as discussed in 9.2.2.1 above, this is undoubtedly due to
the fact that this -l- is the imperfective marker, which of course
does not occur in perfective verb forms. In order to derive the
past perfective forms of causative verbs in ol(ȩ)-, we simply re-
place word-initial o- with ul(ȩ)- (cf. 5.3.2.a). 14 The past tense
forms of imperfective causative verbs in ol(ȩ)- are also derived
in this way: thus we have, for example, olȩchiis ‘chase’—ul-
lȩchiis ‘chased’, ollangȩl ‘make..cry’—ulȩllangȩl ‘made..cry’,
olsebȩk ‘make.. fly’—ulȩlsebȩk ‘made..fly’, etc. If the verb stem
begins with r, any immediately preceding l assimilates to the r,
as in orrebȩt ‘drop’, urrȩbȩtii ‘dropped it’, etc.
The perfective forms of 24, like those of 22, show various
types of vowel and vowel cluster reduction. Can you identify the
types of reduction involved in the pairs of verbs below?

(25) Imperfective Form of 3rd pers. sg. object Present

Causative Verb Perfective Form

olȩchíis ‘chase’ ochisíi ‘chase him/her/it’

ollúut ‘give back, olutíi ‘give it back’
make..return’
ollángȩl ‘make..cry’ olȩngȩlíi ‘make him/her/it cry’
orrébȩt ‘drop’ orȩbȩtíi ‘drop it’

9.5. ERGATIVE FORMS OF CAUSATIVE VERBS

Though the ergative forms (cf. 5.4) of causative verbs are not
used very frequently, they are nevertheless derived according to
regular patterns. Causative verbs in omȩ(k)- show two different
types of ergative forms. In one type, the causative prefix omȩ(k)-
is replaced by muk-, as in the following examples:

245
 9 Causative Verbs

(26) Ergative Form Corresponding Causative Verb in

omȩ(k)-

mukdakt ‘get frightened’ omȩkdakt ‘frighten’

mukdȩchor ‘be made to omȩkdȩchor ‘make… stand’
stand’
mukringȩl ‘be/get hurt/ omȩkringȩl ‘hurt, make..
harmed’ difficult’

In another type of ergative formation, the causative prefix

omȩ(k)- is changed to obȩ(k)-; a detailed analysis of this phe-
nomenon was given in 9.2.1.1 above. Ergative forms in obȩ(k)-
, however, have almost totally fallen out of use in present-day
Palauan; in other words, they have become archaic.
Causative verbs in ol(ȩ)- have ergative forms with initial mo-
, as in the following:

(27) Ergative Form Corresponding Causative Verb in

ol(ȩ)-

mokiis ‘be/get olȩkiis ‘wake up’

awakened’
motuu ‘be made to oltuu ‘put into, make.. enter’
enter’
modik ‘be banished’ oldik ‘banish’

Some discussion of the structure of mo- was presented in

9.2.2.1 above.

9.6. HYPOTHETICAL FORMS OF CAUSATIVE

VERBS
In order to derive the hypothetical forms of causative verbs,
we prefix the various hypothetical pronouns (cf. 4.10 and
4.10.1–9). If the causative verb is imperfective, we derive the
hypothetical forms by replacing the o- of the causative prefix
with a hypothetical pronoun: thus, the hypothetical forms of
omȩka ‘feed’ are kumȩka ‘(if) I feed’, lomȩka ‘(if) he/she feeds’,
etc. If the causative verb is perfective, however, the hypo-
thetical form is derived by substituting a hypothetical pronoun

246
 Palauan Reference Grammar

for the word-initial mȩ- or o-: thus, corresponding to mȩkȩlii

‘feed him/her/it’, we have kukȩlii ‘(if) I feed him/her/it’, lokȩlii
‘(if) he/she feeds him/her/it’, etc.

247
 10 Reciprocal Verbs

10ReciprocalVerbs
10.1. MEANING AND USE OF RECIPROCAL VERBS
There is a special class of Palauan verbs which are used to ex-
press reciprocal actions or states. Such reciprocal verbs de-
scribe situations in which two or more persons or things are
affecting each other in some way. They may be doing the same
action to each other (e.g., hitting each other, washing each
other, talking to each other, etc.), or they may be in the same
state with reference to each other (e.g., afraid of each other,
close to each other, etc.). 1 Because reciprocal actions or states
cannot occur unless at least two persons or things are involved,
the subject of a reciprocal verb must always be plural.
The following is a typical sentence containing a reciprocal
verb:

(1) A Droteo mȩ a Toki a kaingȩseu.

‘Droteo and Toki are helping each other.’

Here, the reciprocal verb kaingȩseu ‘help each other’, which is

formed by adding the reciprocal prefix kai- (see 10.2 below)
to the verb stem ngȩseu ‘help’, implies that the two persons
mentioned in the subject noun phrase—namely, Droteo and
Toki—are performing the action of helping with a reciprocal or
mutual effect. In other words, Droteo is doing something to
help Toki, and in turn Toki is doing something to help Droteo.
In example 1, the two nouns Droteo and Toki (each preceded
by the word a—cf. 2.6) are joined by the connecting word mȩ
‘and’ (see 25.4) to form a plural subject. Not all subject noun
phrases, however, need to have this form in order to be plural.
Consider, for example, the (italicized) subject noun phrases in
the reciprocal sentences below:

(2) a. A rȩngalȩk a kachȩlebȩd.

‘The children are hitting each other.’

b. A rȩchad ȩr a Modȩkngei a kaingȩseu.

‘The people in Modekngei are helping each other.’

248
 Palauan Reference Grammar

c. Tȩ kaiuȩtoir ȩr a sers.
‘They are chasing each other in the garden.’

In each of the sentences above, we have no trouble under-

standing that the subject noun phrase is plural: in 2a–b, the
plural prefix rȩ- (cf. 2.5) has been added to the human nouns
ngalȩk ‘child’ and chad ‘person’, and in 2c the 3rd pers. pl.
human pronoun tȩ ‘they’ (cf. 4.2) automatically refers to two
or more persons. Now, with example 1 above, repeated here for
convenience,

(1) A Droteo mȩ a Toki a kaingȩseu.

‘Droteo and Toki are helping each other.’

compare the following sentences:

(3) a. A Droteo a olȩngȩseu ȩr a Toki.

‘Droteo is helping Toki.’

b. A Toki a olȩngȩseu ȩr a Droteo.

‘Toki is helping Droteo.’

Examples 3a-b are the non-reciprocal sentences corre-

sponding to 1; they are normal transitive sentences in which
a transitive verb—the causative verb olȩngȩseu ‘help’ (cf.
9.2.2.2)—is preceded by a subject noun phrase and followed by
an object noun phrase. While Droteo and Toki are connected
by mȩ ‘and’ to form the “joint” subject of kaingȩseu ‘help each
other’ in 1, in 3a-b these two nouns have changed their function,
since one occurs as the subject of olȩngȩseu ‘help’ and the other
as object. The non-reciprocal sentences 3a-b and the reciprocal
sentence 1 are of course interpreted differently: in 3a, for ex-
ample, the action of helping only goes in one direction—Droteo
is helping Toki, but not vice versa—while in 1, this action is per-
formed in both directions, with a mutual effect on Droteo and
Toki.
Reciprocal verbs behave like state verbs in that their past
tense forms are derived simply by placing the auxiliary word
mle ‘was, were’ (cf. 5.1.3 and 5.3.2) before the verb. Thus, sen-
tences 1 and 2a above have the following past tense equiva-
lents:

249
 10 Reciprocal Verbs

(4) a. A Droteo mȩ a Toki a mle kaingȩseu.

‘Droteo and Toki were helping each other.’

b. A rȩngalȩk a mle kachȩlebȩd.

‘The children were hitting each other.’

10.2. FORMS OF THE RECIPROCAL PREFIX

Reciprocal verbs can be identified by the presence of the reci-
procal prefix, which has quite a few different forms: kai-, kau-,
kaiuȩ-, kȩ-, ka-, and cha-. It is very difficult to formulate general
rules to predict which of these six variants will be used in de-
riving a particular reciprocal verb, although, as we will see in
10.2.4 below, a phonetic rule can account for the occurrence of
cha-, and some of the other variants tend to be used frequently
with certain types of verb stems. The variants of the reciprocal
prefix are attached primarily to the stems of transitive action
verbs, but in a few cases they can be added to the stems of
transitive or intransitive state verbs. Reciprocal verbs do not
contain the verb marker (cf. chap.6) or the imperfective marker
(cf. 5.5) and simply have the structure reciprocal prefix + verb
stem.

10.2.1. RECIPROCAL VERBS RELATED TO CAUSATIVE VERBS

Many verb stems which allow the formation of causative verbs
(cf. chap.9) also permit the derivation of reciprocal verbs. If the
causative verb is derived with the causative prefix ol(ȩ)- (cf.
9.2.2), then the corresponding reciprocal verb usually takes kai-
, kaiuȩ-, or kau-, as in the following list:

(5) Reciprocal Verb Related Causative

Verb

kaingȩseu ‘help each other’ olȩngȩseu ‘help’

kaisiuekl ‘meet each other’ olsiuekl ‘meet’
kaiuȩtoir ‘chase each other’ oltoir ‘chase’
kaiuȩdingȩl ‘visit each other’ oldingȩl ‘visit’
kaudurokl ‘send (something) to oldurokl ‘send’
each other’

250
 Palauan Reference Grammar

All of the causative verbs in ol(ȩ)- given above are somewhat

special because, as their English glosses indicate, they do not
seem to have much of a causative meaning. As mentioned in
9.2.2.2, verbs of this kind probably had a causative meaning
at some earlier stage of the Palauan language, but over a long
period of time the meaning changed and their causative origin
became obscured. Except for ngȩseu, which can occur as an in-
dependent noun meaning ‘help, assistance’, the verb stems ob-
served in the examples of 5 are all bound: that is, the verb
stems siuekl, toir, dingȩl, and durokl do not occur as separate
words. 2
If a causative verb is derived with the causative prefix
omȩ(k)- (cf. 9.2.1), then the corresponding reciprocal verb is
formed by replacing omȩ- with the reciprocal prefix kau-. Inter-
estingly enough, the k of omȩ(k)- remains in the reciprocal verb,
which therefore seems to have been prefixed with kauk-. Note
the following examples:

(6) Reciprocal Verb Related Causative

Verb

kauklatk ‘remind each other’ omȩklatk ‘remind’

kaukdakt ‘frighten each other’ omȩkdakt ‘frighten’
kaukrael ‘guide/advise each omȩkrael ‘guide,
other’ advise’
kaukȩrreu ‘take care of each omȩkȩrreu ‘take care
other’ of’

Some of the verb stems in the examples of 6 occur indepen-

dently as nouns—e.g. dakt ‘fear’ and rael ‘road’—while others
are bound—latk and kȩrreu.
The examples below illustrate how some of the reciprocal
verbs of 5 and 6 are used in sentences:

(7) a. A mlai a mle kaisiuekl.

‘The cars crashed.’ 3

b. A rȩngalȩk er a Belau a sorir ȩl kaukdakt a

lȩklȩbȩsei.
‘Palauan children like to frighten each other at
night.’

251
 10 Reciprocal Verbs

c. Aki kmal kaukȩrreu.

‘We take good care of each other.’

10.2.2. RECIPROCAL VERBS RELATED TO VERBS IN ou-

In 6.1.1 we saw that many Palauan verbs can be derived by
prefixing ou- to a noun or, much less commonly, to a bound
verb stem. Such verbs in ou- —e.g. oublai ‘own a house’ (from
blai ‘house’), ouskuul ‘teach, tutor’ (from skuul ‘school’),
etc.—either designate ownership or control of something or
refer to some kind of activity associated with the noun from
which they are derived. Many verbs in ou- are transitive and are
normally followed by object noun phrases. Observe, for ex-
ample, the sentences below, in which the objects of verbs in ou-
have been italicized:

(8) a. Ak ousȩchȩlei ȩr a Droteo.

‘I’m a friend of Droteo’s.’

b. Ak oungalȩk ȩr a sȩchal.
‘I have/am the parent of a boy.’

c. A rȩchad ȩr a Siabal a ulumȩkȩmad 4 ȩr a rȩchad ȩr a

Merikel.
‘The Japanese made war on the Americans.’

d. A Droteo a oungȩlakȩl ȩr a bȩchil.

‘Droteo is joking with his wife.’

e. Ak ouchad ȩr kȩmiu e ak mo chȩbuul.

‘Having you as my relatives will reduce me to poverty.’

Verbs in ou- which can be used transitively, such as those

illustrated in 8 above, have related reciprocal verbs derived
with the prefix kau-. Some typical pairs of such verbs are listed
below; the related noun, if any, is also provided:

(9) Reciprocal Verb Related Verb in ou- Related

Noun

kaubuch ‘be married to oubuch ‘be married to buch

each other’ (someone)’ ‘spouse’

252
 Palauan Reference Grammar

kausȩchȩlei ‘be friends with ousȩchȩlei ‘be a sȩchȩlei

each other’ friend of (someone)’ ‘friend’
kaungalȩk ‘be related as oungalȩk ‘have (as) a ngalȩk
parent and child’ child’ ‘child’
kauchad ‘be related to each ouchad ‘have (as) a chad
other’ relative’ ‘person’
kaureng ‘long for each oureng ‘wish for’ reng
other’ ‘heart,
spirit’
kaumȩkȩmad ‘fight with oumȩkȩmad ‘make mȩkȩmad
each other (in a war)’ war on’ ‘war’
kauchais ‘tell each other ouchais ‘tell chais
news’ (someone) news’ ‘news’
kausisbech 5 ‘need/ help ousbech ‘need’ -----
each other’
kaungȩlakȩl ‘joke with each oungȩlakȩl ‘joke with’ ngȩllakȩl 6
other’ ‘joke’

The derived reciprocal verbs of 9 above, which describe

mutual relationships between people or denote activities which
people direct at each other, are used in sentences like the fol-
lowing:

(10) a. Ngak mȩ Droteo a {kausȩchȩlei 7 /kauchad}.

‘Droteo and I are {friends/relatives}.’

b. A Toki mȩ, a Droteo tȩ mo kaubuch ȩr oingarang?

‘When are Toki and Droteo getting married?’

c. Aki mle kausisbech ȩl mȩruul a subȩlam.

‘We helped each other do our homework.’

d. Aki kauchais a chisel a belumam.

‘We’re exchanging news about our home villages.’

10.2.3. RECIPROCAL VERBS DERIVED FROM TRANSITIVE VERB

STEMS
The stems of many transitive action verbs can combine with
the prefixes ka- or kȩ- to derive reciprocal verbs. In the list
below, some commonly-used reciprocal verbs of this type are
given, together with the related transitive verb in its imper-
fective form:

253
 10 Reciprocal Verbs

(11) Reciprocal Verb Related Transitive Verb (in

Imperfective Form)

katȩkoi ‘speak to each mȩlȩkoi ‘speak’

other’
kakimd ‘cut each other’s mȩngimd ‘cut (hair)’
hair’
kachȩlebȩd 8 ‘hit each mȩngȩlebȩd ‘hit’
other’
kakeald ‘keep each mȩngeald ‘heat, warm up’
other warm’
kȩluchȩs ‘write to each mȩluchȩs ‘write’
other’
kȩlatk ‘remember each mȩlatk ‘remember’
other’
kȩtub ‘spit on each mȩlub ‘spit’
other’
kȩsilȩk ‘wash each mȩsilȩk ‘wash’
other’s clothes’

Because they do not contain the imperfective marker, the

reciprocal verbs above show stem-initial consonants which have
been deleted after the imperfective marker in the related imper-
fective transitive verbs. As we saw in 5.5, the stem of a verb
like mȩlȩkoi ‘speak’ is the independently-occurring noun tȩkoi
‘word, language’. In the derivation of imperfective mȩlȩkoi from
the sequence mȩ (verb marker) + l (imperfective marker) +
tȩkoi (verb stem), the initial consonant t of tȩkoi is deleted.
This t remains intact, however, in those verb forms which lack
the imperfective marker. Thus, the t is preserved in reciprocal
katȩkoi ‘speak to each other’, as well as in ergative mȩtȩkoi
‘be/get talked to’ (cf. 5.4). To take another example, imper-
fective mȩngeald ‘heat, warm up’ is derived from the sequence
mȩ (verb marker) + ng (imperfective marker) + keald (verb
stem) simply by dropping the initial k of the (bound) verb stem
keald. The k of keald is preserved, however, in the reciprocal
verb kakeald ‘keep each other warm’, in the intransitive state
verb mȩkeald ‘warm’, and in the derived noun kleald ‘heat’ (cf.
8.3, ex.7), all of which do not contain the imperfective marker.
In some cases, the stem of a transitive action verb is par-
tially or completely repeated (or reduplicated—see chap. 11)
when the reciprocal prefix ka- is added. For some verb stems,
this process is obligatory, while for others it is optional, with

254
 Palauan Reference Grammar

considerable variation from speaker to speaker. Some typical

examples are given below; in the reciprocal verbs, the redupli-
cated syllable has been italicized:

(12) Reciprocal Verb Related Transitive Verb (in

Imperfective Form)

kabeboes ‘shoot each omoes ‘shoot’

other’
kakȩrker ‘ask each oker ‘ask’
other’
kasoes ‘see/meet omes ‘see’
each other’
kasusuub ‘imitate each mȩsuub ‘study, imitate’
other’

As mentioned above, the process of reduplication is merely op-

tional for some of the reciprocal verbs in 12. Thus, kȩsuub
means the same as kasusuub ‘imitate each other’, and kaker
may be used instead of kakȩrker ‘ask each other’.
Very rarely, the stem of a transitive state verb can be
combined with a reciprocal prefix to form a reciprocal verb. A
couple of examples include kaodȩnge 9 ‘know each other’ (cf.
mȩdȩnge ‘know’) and kȩdakt ‘afraid of each other’ (cf. mȩdakt
‘afraid’).
We should also note in passing that the stems of a few in-
transitive state verbs can also be used to form reciprocal
verbs. Such verbs designate relationships of distance (near vs.
far) or similarity (similar vs. different) and include kakeed/
kaiuȩkeed ‘near each other’ (cf. kmeed ‘near’), chachȩroid 10
‘far from each other’ (cf. chȩroid ‘far’), kaisisiu ‘similar to/the
same as each other’ (cf. osisiu ‘similar, the same’), kakngodȩch
11
‘different from each other’ (cf. ngodȩch ‘different, strange’),
etc.

10.2.4. THE RECIPROCAL PREFIX cha-

The reciprocal prefix has the variant cha- when the following
verb stem begins with ch. It is clear that the initial consonant
of this prefix has totally assimilated to—i.e., become identical
with—the ch of the following verb stem. Thus, we can claim that
the variant cha- is really derived from ka- by a phonetic rule

255
 10 Reciprocal Verbs

of assimilation. If this is true, then we have simplified our de-

scription of the reciprocal prefixes somewhat, since we can now
say that they all basically begin with k. Some reciprocal verbs
pronounced with initial cha- are given below, together with the
related imperfective transitive verb or the related state verb:

(13) Reciprocal Verb Related Transitive Verb (in

Imperfective Form) orRelated
State Verb

chachȩlebȩd ‘hit each mȩngȩlebȩd ‘hit’

other’
chachuiu ‘look closely mȩnguiu ‘read, look closely at’
at each other’
chachȩdȩchȩduch mȩngȩdȩchȩduch ‘converse’
‘converse with each
other’
chachȩroid ‘far from chȩroid ‘far’
each other’
chacheche ‘jealous of mȩcheche ‘jealous’
each other’

For some Palauan speakers, such as those living in the north of

Babeldaob, the assimilation rule resulting in cha- is not applied,
and therefore ka- is used as the reciprocal prefix in the verbs of
13.

10.3. SAMPLE SENTENCES WITH RECIPROCAL

VERBS
In this section we will look at a few more examples of how
Palauan reciprocal verbs are used in sentences. Most of the rec-
iprocal verbs are taken from the lists or discussions in 10.2.3–4
above.

(14) a. Aki mle katȩkoi ȩr se ȩl mondai er a elii.

‘We talked with each other about that problem
yesterday.’

b. A Droteo mȩ a Hirosi a millim a rrom e mlo

{chachȩlebȩd / kachȩlebȩd}.

256
 Palauan Reference Grammar

‘Droteo and Hirosi drank (a lot of) liquor and started

hitting each other.’

c. A Satsko mȩ a Toki a kȩsilȩk a bilir.

‘Satsko and Toki are washing each other’s clothes.’

d. A Droteo mȩ a Helen a mle kasoes ȩr a kȩsus.

‘Droteo and Helen saw/met each other last night.’

e. Tȩ mle kaodȩnge er se ȩr a taem ȩr a mȩkȩmad.

‘They knew each other during the war.’

In 14c above, we have a reciprocal sentence which has been

made transitive by the addition of the object noun phrase
bilir ‘their clothes’. This object noun phrase is a possessed
noun (cf. 3.1) in which the 3rd pers. pl. human possessor
suffix -ir refers to the same persons as the subject of the sen-
tence (Satsko mȩ a Toki ‘Satsko and Toki’). Transitive recip-
rocal sentences like 14c are interpreted rather differently from
“simple” (nontransitive) reciprocal sentences like 1, which we
repeat here for convenience:

(1) A Droteo mȩ a Toki a kaingȩseu.

‘Droteo and Toki are helping each other.’

In 1, both persons are affected directly by the reciprocal action

of helping: Droteo is being helped by Toki and Toki is being
helped by Droteo. In 14c, however, the effect of the reciprocal
action is indirect. In this sentence Satsko is not being washed
by Toki, nor vice versa; rather, both persons are receiving the
beneficial effect of similar actions which are directed at some-
thing external—namely, each other’s clothing (bilir).
Other transitive reciprocal sentences whose structure and
interpretation are similar to 14c are given below. For ease of re-
ference, the object noun phrase has been italicized:

(15) a. A Tony mȩ a Cisco a chachuiu a hong ȩr tir. 12

‘Tony and Cisco are reading each other’s books.’

b. A Kiosi mȩ a Droteo a kȩluchȩs a babier.

‘Kiosi and Droteo are writing each other letters.’

c. A Toki mȩ a Droteo a blȩchoel ȩl kȩrengȩs 13 a tȩkingir.

257
 10 Reciprocal Verbs

‘Toki and Droteo always listen to/take advice from each

other.’

d. Aki mle kasoes a chuungam.

‘We could see each other’s shadows.’

e. Tȩ kȩka a kȩlir.
‘They’re eating each other’s food.’

In 15c-e, the possessor suffix on the object noun agrees with

the subject of the sentence. Thus, the -am of chuungam ‘our
shadows’ agrees with the 1st pers. pl. excl. pronoun aki ‘we
(excl.)’, and -ir of tȩkingir ‘their voices’ and kȩlir ‘their food’
agrees with the 3rd pers. pl. human subjects Toki mȩ a Droteo
‘Toki and Droteo’ and tȩ ‘they’.

10.4. EXTENDED FUNCTIONS OF THE

RECIPROCAL PREFIX
In the sections above, we have seen that the primary function
of the Palauan reciprocal prefix is to derive verbs denoting ac-
tions or states which have a direct or indirect mutual effect
on two or more persons. In this section, we will examine two in-
teresting cases in which the abovementioned primary function
of the reciprocal prefix seems to have been extended or mod-
ified.
A rather natural extension of the use of the reciprocal prefix
is found in the derivation of reciprocal verbs from the stems of
intransitive action verbs. Since intransitive action verbs do
not take an object, but rather denote activities in which only
the subject himself can be involved, reciprocal verbs related to
them cannot possibly imply mutual effect. Instead, reciprocal
verbs of this type denote activities which two or more persons
do together or in a group. Observe the sentences below:

(16) a. Aki mle chachiuaiu.

‘We were all asleep.’

b. A rȩngalȩk a kaidȩrurt.
‘The children are running together.’

c. A rubak a mle kadȩrȩborb ȩr a chȩlsel a bai.

258
 Palauan Reference Grammar

‘The old men were sitting together inside the men’s

house.’

The reciprocal verbs in 16a-c are related to the stems of in-

transitive action verbs as follows: chachiuaiu ‘sleep with each
other (in the same room, etc.)’—mȩchiuaiu ‘sleep’, kaidȩrurt 14
‘run with each other’—rȩmurt ‘run’, and kadȩrȩborb 14 ‘sit to-
gether’—rȩborb ‘sit’. These verbs all imply that the subjects
are doing some activity jointly—e.g. sleeping in the same place,
sitting in the same place, etc.—although no reciprocity is in-
volved. However, some Palauan speakers feel that sentences
like 16a-c have a special connotation of “group spirit” or “to-
getherness”. For example, kadȩrȩborb ‘sit together’ of 16c
seems to include the idea of some special comradeship existing
between the men who were sitting together.
Another extended function of the reciprocal prefix is to
weaken certain commands or suggestions, or make them more
polite. In this usage, the reciprocal prefix ka- is added directly
to the hypothetical forms of verbs (cf. 4.10 and 4.10.1–9). Its
presence suggests that the speaker and the person addressed
are somehow mutually involved in the decision at hand; often
there is a strong sense of the speaker’s concern or sympathy.
Note the contrast between the following pairs of sentences:

(17) a. Kurael ȩl mo ȩr a blik.

‘I’d better go home.’

b. Kakurael ȩl mo ȩr a blik.
‘Why don’t I go home.’

(18) a. Molim a kȩrum!

‘Take your medicine!’

b. Kamolim a kȩrum.
‘Let’s take your medicine.’

(19) a. Domoes a bȩlochȩl.

‘We’re to shoot pigeons.’

b. Kadomoes a bȩlochȩl.
‘Let’s shoot pigeons.’

259
 10 Reciprocal Verbs

In the a- sentences of each pair, there is a strong imperative

connotation—that is, the action is something which the speaker
feels must be done (either by himself or someone else). The
addition of ka- in the b-sentences, however, changes these im-
perative sentences into mild or polite suggestions. The only hy-
pothetical form which does not permit ka- is the 1st pers. pl.
exclusive; in other words, we do not have forms like *kaki-
morael or *kaki molim. This is because the function of ka- under
discussion here is to involve the person spoken to in the de-
cision, whereas the use of 1st pers. pl. exclusive forms automat-
ically excludes that person.
Some additional examples of this use of ka- are given in the
sentences below:

(20) a. Kamonguiu ȩr tia ȩl hong, e dȩsȩkau ȩl kmo kȩ mo pass

ȩr a test ng diak.
‘Read this book and let’s see whether or not you pass
the test.’

b. Kalȩbo lȩmȩchiuaiu a Toki, e dȩsa ȩl kmo ng mo ungil

ng diak.
‘Let’s have Toki go to bed and we’ll see if she gets
better.’

c. Kabȩskak tilȩcha ȩl oluchȩs.

‘How about giving me that pencil?’

260
 11 Reduplication and Further
Verb Affixation

11ReduplicationandFurtherVerbAffixation
11.1. TYPES OF VERB DERIVATION
In several of the preceding chapters we have attempted to des-
cribe in detail how the many different types of Palauan verbs
are derived. Thus, in chaps. 5 and 6, we showed how certain
important grammatical affixes such as the verb marker, the
imperfective marker, the past tense marker, hypothetical
pronouns, object pronouns, etc. combine with verb stems
to derive action verbs, state verbs, imperfective and per-
fective verb forms, ergative verb forms, hypothetical verb
forms, and the like. In chap. 7, we specifically treated the many
subclasses of derived state verbs, and in chaps. 9 and 10 we
examined the derivational patterns for causative verbs and
reciprocal verbs.
Although we have already covered many aspects of Palauan
verb derivation, as summarized above, our discussion of it is
not yet complete. Therefore, in this chapter we will concentrate
on the three major derivational processes which remain to be
described. We will focus most of our attention on the process
of reduplica tion, which involves repeating or reduplicating
part (or, less frequently, all) of a verb stem. The patterns of
reduplication are complex and often highly irregular, and we
will therefore examine only the most commonly used (or pro-
ductive) processes, together with the different types of
meaning change which result from reduplication. In this
chapter we will also discuss two suffixes which are important
in verb derivation—the predictive suffix -u, which expresses
the speaker’s prediction that the action of the verb is about to
happen, and the inchoative suffix -a, which denotes an action
or state which has just come about—i.e., which is new or unex-
pected.

261
 11 Reduplication and Further Verb Affixation

11.1.1. PREVIOUS EXAMPLES OF REDUPLICATION

At several points in the text we have already made passing re-
ference to reduplicated verb forms, though we did not analyze
them in detail. In 7.2, for instance, we noted that certain state
verbs referring to size or dimension show a reduplicated part.
Thus, kekȩre ‘small’ obligatorily contains the reduplicated
initial syllable ke-, while the addition of this same syllable is
optional in (ke)kȩmangȩt ‘tall, long’ and (ke)kȩdeb ‘short’. 1
Further, in 7.5–6 we saw that some verbs derived with the
prefixes bȩ- and bȩkȩ- require the verb stem to be partially
reduplicated, as in bȩchachas ‘sooty’ (from chas ‘soot, ash’),
bȩkȩlilangȩl ‘cry a lot’ (from langȩl ‘crying’), bȩkȩsechȩsechȩr
‘get sick often’ (from sechȩr ‘sickness’), etc. Finally, in 10.2.3
we observed reciprocal verbs like kakȩrker ‘ask each other’
(from ker ‘question’) and kasusuub ‘imitate each other’ (cf.
mȩsuub ‘study, imitate’), where all or part of the verb stem is
(optionally) reduplicated before adding the reciprocal prefix
ka-.
In the examples given above, it seems that the reduplication
of all or part of a verb stem has no particular meaning or
function; rather, this reduplication is a kind of “grammatical
device” 2 which is applied—quite unpredictably—when we
derive state verbs, reciprocal verbs, and the like from certain
verb stems. In the majority of instances, however, the redu-
plication of a verb stem does result in a significant change of
meaning. Practically all types of verbs can undergo the process
of reduplication, and speakers often do not agree with each
other on the correct form or meaning. Therefore, some of the
reduplicated forms discussed in the sections below may not be
part of your own “active” speech, although you may understand
them and hear other speakers use them.

11.2. REDUPLICATION OF SIMPLE STATE VERBS

Simple state verbs (i.e., those consisting of a single morpheme
or meaning-bearing unit—cf. 7.1) are normally reduplicated by
taking the initial consonant of the stem, adding the vowel e (pro-
nounced as the full vowel [ε]), and prefixing this extra syllable to
the whole verb stem. If we represent the stem-initial consonant
as C 1 (where C stands for “consonant” and the subscript nu-
meral 1 identifies the first or initial consonant of the stem), then

262
 Palauan Reference Grammar

we can say that the reduplicated forms of simple state verbs are
derived by prefixing the sequence C 1 e to the verb stem. Ob-
serve the examples below:

(1) Reduplicated Form Related Simple State Verb

bebeot ‘fairly easy/cheap’ beot ‘easy/cheap’

dedȩkimȩs ‘kind of wet’ dȩkimȩs ‘wet’
sesȩkool ‘rather playful’ sȩkool ‘playful’
kekȩdung ‘fairly well-behaved’ kȩdung ‘well-behaved’
bebubong ‘somewhat senile’ bubong ‘senile’
sesongȩrengȩr ‘kind of hungry’ songȩrengȩr ‘hungry’

As the English equivalents for the reduplicated forms of 1

show, the function of reduplication with state verbs is to qualify
or weaken the meaning of the verb or give it a more tentative
or less definite quality. Therefore, the English translations
usually contain qualifying words such as ‘kind of’, ‘somewhat’,
‘fairly’, ‘rather’, etc. The use of the reduplicated state verbs of 1
is illustrated in the sentences below. Very often, the expression
ko ȩr a ‘kind of, somewhat’ precedes the reduplicated verb and
gives an added connotation of tentativeness:

(2) a. Ng ko ȩr a bebeot a chȩral a klok ȩr a Hong Kong. 3

‘The prices for watches are pretty low in Hong Kong.’

b. Ng dedȩkimȩs a bilek e le ak killii a chull. 4

‘My clothes are kind of wet because I got caught in the
rain.’

c. Ak ko ȩr a sesongȩrengȩr e ng di chȩtik ȩl omȩngur.

‘I’m kind of hungry but I don’t want to eat.’

11.2.1. REDUPLICATION OF POSSESSED NOUNS

The possessed nouns soak ‘my liking’, soal ‘his liking’, etc. and
chȩtik ‘my disliking’, chȩtil ‘his disliking’ etc. (see 17.1) can
be reduplicated with a similar weakening of meaning. While
chȩtik— chechȩtik follows the pattern for simple state verbs ob-
served in 11.2 above, soak—sosoak involves a repetition of the
first syllable of the possessed noun. The use of these words is
illustrated in the following sentences: 5

263
 11 Reduplication and Further Verb Affixation

(3) a. A Droteo a ko ȩr a sosoal ȩl mo ȩr a

chei.
‘Droteo would sort of like to go
fishing.’

b. Ng chechȩtik ȩl mȩrael.
‘I don’t feel much like leaving.’

Even though words like soak ‘my liking’, chȩtik ‘my dis-
liking’, etc. are nouns, they permit the process of reduplication,
which is otherwise restricted to verbs. This unusual phe-
nomenon may be due to the fact that these nouns are rather like
state verbs in meaning—that is, liking or disliking something is
a kind of “state of mind”. Interestingly enough, the possessed
nouns under discussion translate into English most naturally as
verbs—i.e., soak ‘I like…’, chȩtil ‘he dislikes…’, etc.

11.3. REDUPLICATION OF STATE VERBS IN MȨ-

State verbs consisting of the verb marker mȩ- and a verb stem
(cf. 7.1) are normally reduplicated by adding a syllable of the
form C 1 e. This involves the same pattern as that observed in
11.2 above, except that the verb marker prefix remains at the
beginning of the word. A weakening of meaning is also involved,
as in the following examples:

(4) Reduplicated Form Related State Verb in mȩ-

mȩtetongakl ‘fairly tall’ mȩtongakl ‘tall’

mȩkekekad ‘kind of itchy’ mȩkekad ‘itchy’
mȩchechȩsa ‘kind of busy’ mȩchȩsa ‘busy’
mȩsesulaul ‘kind of sleepy’ mȩsulaul ‘sleepy’
mȩchechuu ‘fairly shady’ mȩchuu ‘shady’
mȩleliliut ‘fairly thin’ mȩliliut ‘thin’
mȩdedȩnge ‘be somewhat mȩdȩnge ‘know’
familiar with’
mȩtetitur ‘be rather mȩtitur ‘not know how (to), not
unfamiliar with’ be capable of’

A few intransitive state or action verbs which contain the

metathesizcd verb marker after the initial consonant of the
stem (cf. 6.2) follow the same rule of adding a reduplicated syl-

264
 Palauan Reference Grammar

lable of the form C 1 e. In these cases, however, the metathe-

sized verb marker remains within the stem, as in the examples
below:

(5) Reduplicated Form Related State Verb With Metathesized

Verb Marker

sesmechȩr ‘kind of smechȩr ‘sick’

sick’
sesmau ‘sort of smau ‘used to’
used to’
chechuarm ‘sort of chuarm ‘suffer’
suffer’

The sentences below illustrate the use of the reduplicated

forms of 4 and 5:

(6) a. A Droteo a ko ȩr a mȩchechȩsa mȩ ng diak lsȩbȩchel ȩl

mong.
‘Droteo is rather busy, so he can’t go.’

b. Ak ko ȩr a mȩsesulaul mȩ ak mochu 6 mȩchiuaiu.

‘I’m kind of sleepy, so I’ll be going to bed.’

c. A dȩlal a di sesmechȩr mȩ ng diak loureor ȩr a sers.

‘His mother is kind of sick, so she’s not working in the
garden.’

11.4. REDUPLICATION OF DERIVED STATE VERBS

Some derived state verbs containing both the resulting state
infix -(ȩ)l- and the anticipating state suffix -(ȩ)l (cf. 7.8.2) may
be reduplicated, although the resulting forms are not particu-
larly common. Again, a reduplicated syllable of the form C 1 e is
prefixed to the derived state verb, as in the following:

(7) Reduplicated Form Related Derived State Verb

7

ngengȩltachȩl ‘not well ngȩltachȩl ‘cleaned’

cleaned’

265
 11 Reduplication and Further Verb Affixation

ngenglȩmodȩl ‘not well nglȩmodȩl ‘mopped’

mopped’
ngenglȩmull ‘not well cut’ nglȩmull ‘cut (grass)’

As the English equivalents show, the reduplicated forms of 7 ex-

hibit a rather unusual change of meaning: they describe states
resulting from the incomplete or inadequate performance of
some task.
Some state verbs derived with the prefix bȩkȩ- (cf. 7.6 and
7.6.1) are reduplicated by inserting the syllable ke between
the two syllables of bȩ kȩ-. The reduplicated forms exhibit a
weakening of meaning, as in the following:

(8) Reduplicated Form Related State Verb in

bȩkȩ-

bȩkekȩsius ‘rather vulgar in bȩkȩsius ‘vulgar in

speech’ speech’
bȩkekȩtȩkoi ‘rather talkative’ bȩkȩtȩkoi ‘talkative’
bȩkekȩsȩngorȩch ‘smell rather bȩkȩsȩngorȩch ‘smell like
like a pig’s house’ a pig’s house’

11.5. COMPLEX PATTERNS OF REDUPLICATION

As we have seen in the sections above, the relatively simple re-
duplicative pattern C 1 e accounts for the reduplicated forms
of quite a few Palauan state verbs. There are some state verbs,
however, whose reduplicated forms follow a much more compli-
cated pattern, which can be symbolized as C 1 eC 1 V(C 2).
This formula indicates that the reduplicated sequence is com-
posed of two syllables: the first syllable—C 1 e—is equal to the
simpler pattern of reduplication with which we are already fa-
miliar, while the second syllable—C 1 V(C 2)—represents a new
element. The second syllable consists of the initial consonant
of the verb stem (C1), followed by some vowel (hence, our use
of the cover symbol V for “vowel”), followed optionally by the
second consonant of the verb stem (C2). The vowel of C 1 V(C 2)
appears as ȩ (schwa) if the verb stem has a single full vowel; if
the verb stem has a vowel cluster, however, the vowel of C 1 V(C
2) will normally be one of the vowels found in this cluster. Thus,

266
 Palauan Reference Grammar

the reduplicated syllable C 1 V(C 2) shows the effects of vowel

reduction (cf. 1.4.4 and 3.4) and vowel cluster reduction (cf.
3.4.3).
The examples below, which are the reduplicated forms of
some state verbs in mȩ-, follow this complex but quite pro-
ductive pattern. Note that the reduplicated sequence is placed
after the verb marker and before the full verb stem:

(9) Reduplicated Form Related State Verb in mȩ-

mȩsesusaul ‘kind of tired’ mȩsaul ‘tired’

mȩdedȩkdakt ‘kind of afraid’ mȩdakt ‘afraid’
mȩsesisaik ‘kind of lazy’ mȩsaik ‘lazy’
mȩrerȩngȩringȩl ‘kind of difficult’ mȩringȩl ‘difficult’
mȩkekȩrkar ‘half awake’ 8 mȩkar ‘awake’

Let us now isolate the second syllable—C 1 V(C 2)—of the

reduplicated sequence and compare it with the verb stem:

9
(10) C 1 V(C 2) Verb Stem

-su- saul ‘tiredness’

-dȩk- dakt ‘fear’
-si- saik ‘laziness’
-rȩng- ringȩl ‘pain, difficulty’
-kȩr- kar ‘(state of) being awake’

We can see clearly that a single full vowel in the verb stem is re-
duced to ȩ in C 1 V(C 2), while a vowel cluster in the verb stem
is reduced to a single full vowel in C 1 V(C 2).
The reduplicated forms of 9, which exhibit a weakening of
meaning, are used in sentences like the following:

(11) a. Ak ko ȩr a mȩsesusaul mȩ ak mochu rȩmei.

‘I’m rather tired, so (I guess) I’ll be going home.’

b. A ngȩlȩkek a mȩdedȩkdakt a dȩrumk.

‘My child is rather afraid of thunder,’

267
 11 Reduplication and Further Verb Affixation

11.6. REDUPLICATION OF INTRANSITIVE ACTION

VERBS
Some intransitive action verbs have reduplicated forms
which follow the patterns C 1 e or C 1 eC 1 V(C 2) observed
for state verbs in 11.2–5 above. If the intransitive action verb
happens to have the metathesized verb marker (cf. 6.2), then its
reduplicated form shows this marker as a prefix mȩ-. Observe
the following examples:

(12) Reduplicated Form Related Intransitive

Action Verb

mȩchechiis ‘keep avoiding/ chȩmiis ‘run away’

sneaking out’
mȩrerȩbȩrebȩt ‘fall one by one’ ruebȩt ‘fall’
mȩleluluut ‘keep returning’ lmuut ‘return’
mȩrerorael 10 ‘walk aimlessly’ mȩrael ‘walk’
mȩsesilil 11 ‘fool around’ milil ‘play’

While mȩchechiis ‘keep avoiding/sneaking out’ is reduplicated

with C 1 e, the other forms above are reduplicated with C 1 eC
1 V (C 2). In mȩrerȩbȩrebȩt ‘fall one by one’, the second syllable
of the reduplicated sequence shows C 2 as b, and for purposes
of pronunciation a ȩ is inserted between this b and the r of the
following verb stem.
As the English equivalents for the reduplicated forms of
12 show, the function of reduplication with intransitive action
verbs is different from what we have already encountered for
state verbs. In general, reduplicated intransitive action verbs
involve actions which are continued or repeated without con-
scious intention in a habitual, absent-minded, or even com-
pulsive way. It is very difficult to find suitable English transla-
tions for reduplicated intransitive action verbs when they are
used in sentences; therefore, the translations given for the sen-
tences below are rather free:

(13) a. A rdȩchel a lius a mȩrerȩbȩrebȩt.

‘The coconuts are falling one by one.’

b. A Moses a di blȩchoel ȩl mȩrerorael a lȩklȩbȩsei.

‘Moses is always wandering around at night.’

268
 Palauan Reference Grammar

c. Ak mȩrael e ko ȩr a di mȩleluluut a rȩnguk.

‘I’m leaving but I don’t know if I really want to (lit., my
mind keeps returning).’

d. A Toki a di mȩsesilil e a blai a di diak lȩkȩltmokl.

‘Toki just fools around and the house doesn’t get
cleaned.’

In 7.3 we saw that a few intransitive verbs like mȩkar ‘be

awake, wake up’, dȩngchokl ‘sit, sitting’, etc. can function both
as state verbs and action verbs since they have past tense
forms with the auxiliary mle or with the infix -il-, respectively.
Further evidence for the dual role of such intransitive verbs
can be found in the fact that some of them exhibit one redupli-
cated form with the qualifying or weakening meaning charac-
teristic of redup licated state verbs and another reduplicated
form with the repetitive or habitual interpretation character-
istic of reduplicated in transitive action verbs. A typical ex-
ample is the intransitive verb mȩkar ‘be awake, wake up’, which
has the reduplicated form mȩkekȩrkar ‘half awake’ (listed in 9
above), as well as the reduplicated form mȩkȩrkar ‘keep waking
up’, which contains only C 1 V(C 2) as the reduplicated syl-
lable. Some intransitive verbs have only a single reduplicated
form, but one which can be interpreted in two ways. For ex-
ample, rȩborb ‘sit (like a man)’ has the reduplicated form mȩd-
edȩrȩborb 12 , which can mean either ‘squat (i.e., sit more or
less the way men do)’ or ‘sit around’.

11.7. REDUPLICATION OF TRANSITIVE ACTION

VERBS
So far, we have examined the various forms and meanings which
result from reduplicating the stems of (intransitive) state verbs
and intransitive action verbs. In this section, we will look at
the reduplicative patterns relevant to transitive action verbs,
dealing first with their ergative forms and later with their im-
perfective forms.
The ergative forms of transitive verbs (cf. 5.4) can be
reduplicated according to three different patterns: the first
two—C 1 e and C 1 eC 1 V(C 2)—are familiar to us from above,
while the third— C 1 V(C 2)—is merely the second pattern minus

269
 11 Reduplication and Further Verb Affixation

the first syllable. It is impossible to predict which of these

three reduplicative patterns will apply to a particular verb stem,
and there is some disagreement among speakers as to the cor-
rectness of forms. The reduplicated forms of ergative verbs all
have the special meaning ‘easy to…’. In the examples below, the
reduplicated syllable has the form C 1 e:

(14) Reduplicated Form Related Ergative Verb Form

mȩchechȩsimȩr ‘easy to close’ mȩchȩsimȩr ‘be/get closed’

mȩchechȩlebȩd ‘easy to hit’ mȩchȩlebȩd ‘be/get hit’
mȩtetȩkoi ‘easy to talk to’ mȩtȩkoi ‘be/get talked to’
mȩsesesȩb ‘flammable’ mȩsesȩb ‘be/get burned’

In the following examples, the reduplicated form is derived

with the pattern C 1 eC 1 V(C 2):

(15) Reduplicated Form Related Ergative Verb

Form

mȩdedȩngȩdangȩb ‘easy to mȩdangȩb ‘be/get

cover’ covered’
obebibuid ‘easy to glue’ obuid ‘be/get glued’
mȩkekikiut ‘easy to clear’ mȩkiut ‘be/get cleared’
mȩlelȩchȩluchȩs 13 ‘easy to write mȩluchȩs ‘be/get
on’ written’
mȩlelȩchȩlechȩt ‘easy to tie’ mȩlechȩt ‘be/get tied’

Notice that, for pronunciation purposes, an extra ȩ must be

added between C 2 of the reduplicated sequence and the fol-
lowing verb-stem-initial consonant in words like mȩdedȩngȩ-
dangȩb ‘easy to cover’, mȩlelȩchȩlechȩt ‘easy to tie’, etc.
The examples below show ergative verb forms reduplicated
with C 1 V(C 2):

(16) Reduplicated Form Related Ergative Verb

Form

mȩngingiokl ‘easy to cook’ mȩngiokl ‘be/get cooked

(starch)’
mȩsusuub ‘easy to study’ mȩsuub ‘be/get studied’
mȩchichuiu ‘easy to read’ mȩchuiu ‘be/get read’

270
 Palauan Reference Grammar

mȩrȩmram ‘easy to mix’ mȩram ‘be/get mixed’

obȩlȩbalȩch ‘easy to shoot obalȩch 14 ‘be/get shot
with a slingshot’ with a slingshot’
obebeu ‘breakable’ obeu ‘be/get broken’

The reduplicated ergative verb forms listed in 14–16 above

are used in sentences such as the following:

(17) a. Ng kmal mȩchechȩlebȩd a otȩchel a Droteo.

‘Droteo’s pitches/throws of the ball are easy to hit.’

b. Tia ȩl sers a mȩkekikiut e le ng mla ȩr ngii a chull.

‘This garden is easy to clear because there’s been some
rain.’

c. A kiuid a obȩlȩbalȩch.
‘Blackbirds are easy to shoot with a slingshot.’

d. Alii. Tilȩcha ȩl butilia a obebeu!

‘Watch out—that bottle is breakable!’

The three patterns of reduplication observed above can also

be applied to the imperfective forms of transitive verbs, but
the resulting forms exhibit much variation from speaker to
speaker. The meaning of reduplicated imperfective verbs is
similar to that observed in 11.6 above for intransitive action
verbs: the action is repeated or continued absent-mindedly or
without any conscious intent.
Many imperfective verbs, when reduplicated, essentially
follow the C 1 V(C 2) pattern, except that the imperfective
marker (cf. 5.5) appears as -l-, -ng-, or -m- directly before the
reduplicated sequence. Thus, in the reduplicated forms below,
the imperfective marker is found preceding the reduplicated
syllable but not preceding the full verb stem:

(18) Reduplicated Form Related Imperfective

Verb Form

mȩlȩbtub ‘keep spitting’ mȩlub ‘spit’

mȩngȩmkimd ‘keep trimming’ mȩngimd ‘cut (hair),
trim’
mȩngikiis ‘keep digging’ mȩngiis ‘dig’
mȩngȩlka 15 ‘keep eating’ mȩnga ‘eat’

271
 11 Reduplication and Further Verb Affixation

omeboes ‘shoot indiscriminately’ omoes ‘shoot’

omebeu ‘keep breaking’ omeu ‘break’
omȩlȩbalȩch ‘play around with a omalȩch ‘shoot with a
slingshot’ sling-shot’

Whereas an imperfective verb form like mȩlub ‘spit’ has the

basic structure

(19) verb marker + imperfective marker + verb stem

mȩ + 1 + tub

the corresponding reduplicated form mȩlȩbtub ‘keep spitting’

has approximately the following basic structure:

(20) verb marker + imperfective marker + C 1 VC 2 + verb stem

mȩ + 1 + tȩb + tub

In 19, the initial consonant t of the verb stem tub ‘spit’ is

deleted following the imperfective marker, giving mȩlub. In 20,
however, the imperfective marker appears directly before the
reduplicated sequence C 1 VC 2 and not the verb stem, and
therefore it is the initial consonant t of the reduplicated se-
quence tȩb which gets deleted, while that of the verb stem re-
mains intact. The different position of the imperfective marker
therefore accounts for the phonetic form of mȩlȩbtub ‘keep
spitting’. The other reduplicated forms of 18 can be explained
in exactly the same way.
Some verbs in which the imperfective marker appears as -l-
have reduplicated forms derived according to the C 1 eC 1 V(C
2) pattern, except that the imperfective marker -l- is taken as C
1 . The reduplicated forms of such verbs will therefore contain
three occurrences of the consonant l, as in the examples below:

(21)Reduplicated Form Related mperfective Verb

Form

mȩlelȩmȩlamȩch ‘chew mȩlamȩch ‘chew’

constantly’
mȩlelȩmȩlimȩt ‘keep bailing’ mȩlimȩt ‘bail’

272
 Palauan Reference Grammar

mȩleltȩlatȩch 16 ‘clean mȩlatȩch ‘clean’

compulsively’
mȩlelȩbȩlobȩch ‘keep mȩlobȩch ‘chop’
chopping’

In the reduplicated forms of 21, l appears first in the syllable C

1 e, then in the syllable C 1 V(C 2), and finally before the verb
stem itself (whose initial consonant has been deleted).
Another pattern used for deriving the reduplicated forms
of imperfective verbs essentially involves adding a syllable of
the form C 1 e. As in the examples of 18, the imperfective
marker is removed from its position before the full verb stem
and shifted to a position directly preceding the reduplicated
syllable. Unlike the examples of 18, however, the initial con-
sonant of the reduplicated syllable is not deleted, even though
the imperfective marker precedes. This rather unexpected phe-
nomenon is observed in the examples below:

(22)Reduplicated Form Related Imperfective Verb

Form

mȩngchechȩlebȩd ‘keep hitting’ mȩngȩlebȩd ‘hit’

mȩngchechuiu ‘keep reading, mȩnguiu ‘read’
read a lot’
ombebȩkall ‘sail/drive around’ omȩkall ‘sail, drive’
ombibtar 17 ‘keep swinging’ omtar ‘swing’

While an imperfective verb form like mȩngȩlebȩd ‘hit’ has

the basic structure

(23) verb marker + imperfective marker + verb stem

mȩ + ng + chȩlebȩd

the corresponding reduplicated form mȩngchechȩlebȩd ‘keep

hitting’ is structured as follows:

(24) verb marker + imperfective marker + C 1 e +verb stem

mȩ + ng + che +chȩlebȩd

273
 11 Reduplication and Further Verb Affixation

In 23, the verb-stem-initial consonant ch is deleted following the

imperfective marker, giving mȩngȩlebȩd ‘hit’. In 24, however,
this deletion rule for some reason does not apply to the initial
consonant ch of the reduplicated syllable, and we simply get
mȩngche chȩlebȩd ‘keep hitting’.
The reduplicated forms of some imperfective verbs are de-
rived like those of 22 above, except that the reduplicated syl-
lable C 1 V(C 2) is also added. A couple of examples, including an
alternate reduplicated form of mȩnguiu ‘read’, are given below:

(25) Reduplicated Form Related Imperfective

Verb Form

mȩngchechichuiu ‘keep reading, mȩnguiu ‘read’

read a lot’
mȩngkekikiut ‘keep clearing’ mȩngiut ‘clear’
ombebibail ‘wrap, clothe’ omail ‘clothe’

The reduplicated imperfective verbs listed in 18, 21, 22, and

25 above can be used in sentences like the following:

(14) a. A Hermana a mȩrȩchȩrachȩd a rȩngul mȩ ng di

mȩlȩbtub.
‘Hermana feels rather nauseous, so she keeps spitting.’

b. Ng diak a blatong e le a Toki a di omebeu ȩr a bek ȩl

sils.
‘There aren’t any plates because Toki keeps breaking
them every day.’

c. Ng di soal ȩl mȩlelȩmȩlamȩch e diak loureor.

‘He just likes to chew (betel nut) and doesn’t do any
work.’

d. Ngara mȩ kȩ di mȩngchechȩlebȩd ȩr a rȩngalȩk?

‘Why are you always hitting the children?’

274
 Palauan Reference Grammar

11.8. REDUPLICATION OF VERBS IN OU- AND O-

Verbs which contain the prefix ou- (cf. 6.1.1) and verbs derived
with the verb marker o- which irregularly lack the imperfective
marker (cf. 6.1, ex. 3) have reduplicated forms, but a large
variety of patterns is observed. Note, for example, the forms
below:

(27) Reduplicated Form Related Verb in ou- or o-

ousesȩchȩlei ‘be sort of friends ousȩchȩlei ‘be friends

with’ with’
oungengȩroel ‘keep scolding oungȩroel ‘scold’
(mildly)’
ourureng ‘miss, be nostalgic for’ oureng ‘wish for’
okiklukl ‘keep coughing’ oklukl ‘cough’
okȩrker ‘ask around’ oker ‘ask’
osisiu ‘same’ osiu ‘joined’

Whatever the form of the reduplicated syllable in the examples

of 27, it is always placed after the prefix ou- or o-. In ous-
esȩchȩlei ‘be sort of friends with’ and oungengȩroel ‘keep
scolding (mildly)’, the reduplicated syllable is C 1 e. In ourureng
‘miss, be nostalgic for’ and okiklukl ‘keep coughing’, however,
the reduplicated syllables ru and ki contain totally unexpected
vowels (cf. note 17 above). Finally, okȩrker ‘ask around’ and
osisiu ‘same’ involve the C 1 V(C2 ) pattern of reduplication.
The meanings of the reduplicated forms of 27 are rather
difficult to predict. While ousesȩchȩlei ‘be sort of friends with’
shows the weakening of meaning observed for state verbs (cf.
11.2–5 above), oungengȩroel ‘keep scolding (mildly)’, ourureng
‘miss, be nostalgic for’, okiklukl ‘keep coughing’, and okȩrker
‘ask around’ involve the connotation of continued or repeated
action observed for action verbs (cf. 11.6–7 above). Note,
further, that the meaning of reduplicated osisiu ‘same’ is not
easily predictable from that of non-reduplicated osiu ‘joined’.
The following sentences illustrate the use of the verbs of 27:

(28) a. Ak okiklukl e le ak smechȩr ȩr a tȩretȩr.

‘I keep coughing because I’m sick with a cold.’

275
 11 Reduplication and Further Verb Affixation

b. Ng di ousesȩchȩlei ȩr a Toki mȩ ng sȩbȩchel ȩl kie ȩr a

blil.
‘He’s sort of close to Toki, so he can stay at her house.’

c. Ak di ourureng ȩr a taem ȩr a Siabal.

‘I feel nostalgic about the Japanese times.’

11.9. REDUPLICATION OF CAUSATIVE VERBS

Causative verbs (cf. chap. 9) can also undergo reduplication.
Those causative verbs derived with the prefix omȩ(k)- (cf. 9.2.1)
show an unusual pattern of reduplication: an extra syllable—ke
or ki—is inserted between the ȩ (which may then be deleted)
and the k of the prefix. A few examples are given below:

(29) Reduplicated Form Related Causative Verb in

omȩ(k)-

omȩkekȩsiu ‘roughly copy’ omȩkȩsiu ‘compare, copy’

omȩkikdakt ‘frighten…a omȩkdakt ‘frighten’
little’
omkiksau ‘sort of make… omȩksau ‘make…used to’
used to’

Causative verbs derived with the prefix ol(ȩ)- (cf. 9.2.2) have
reduplicated forms which follow the familiar patterns C 1 e or
C 1 eC 1 V(C 2). The reduplicated portion is added after the
causative prefix, as in the following examples:

(30) Reduplicated Form Related Causative Verb

in ol(ȩ)-

oltetȩrau ‘sell a little at a time’ oltȩrau ‘sell’

oltetȩbȩtobȩd ‘keep taking out’ oltobȩd ‘take out’
olȩkekȩrkar ‘keep trying to wake olȩkar ‘wake up’
up (gently)’
oltetȩmȩtom ‘keep poking out’ oltom ‘poke out’

As the English equivalents for the reduplicated forms in 29

and 30 show, the function of reduplication seems to differ ac-
cording to whether the related causative verb is prefixed with
omȩ(k)- or ol(ȩ)-. In the former case, the reduplicated form in-

276
 Palauan Reference Grammar

volves a weaker or more tentative connotation, while in the

latter case, the reduplicated form involves a repetitive meaning.
The following sentences illustrate the use of the above redupli-
cated causative verbs:

(31) a. A Toki a omȩkekȩsiu ȩr a bilel ȩr a bilel a Hermana.

‘Toki is making her dress roughly like Hermana’s.’

b. Ak di oltetȩrau a iasai e mȩchȩrar a mlik.

‘I’m just selling vegetables a little at a time and then I’ll
(be able to) buy my car.’

c. A ngikȩl a oltetȩmȩtom ȩr a mȩdal ȩr a bad.

‘The fish keeps poking his head out of the coral.’

11.10. REDUPLICATION OF RECIPROCAL VERBS

As we saw in 10.2.3, ex. 12, some reciprocal verbs exhibit an
optional reduplicated syllable. The majority of speakers cannot
recognize any difference in meaning between the reduplicated
and non-reduplicated forms of reciprocal verbs; therefore, only
a single English gloss is provided for the examples below:

(32) Reduplicated Related Reciprocal Verb

Form

kadȩkdakt kȩdakt ‘fear each other’

kasusuub kȩsuub ‘imitate each other’
kasisiik kȩsiik ‘look for each other’
karuruul kȩruul ‘make..for each other, protect
each other’
kakȩrker kȩker, kaker ‘ask each other’

In the examples of 32, the reduplicated syllable C 1 V(C 2) is

added after the reciprocal prefix ka-.

11.11. “FOSSILIZED” REDUPLICATION

There are many Palauan words—mostly state verbs—whose
phonetic form leads us to suspect that they were once derived
by processes of reduplication. At earlier stages of the Palauan

277
 11 Reduplication and Further Verb Affixation

language, reduplication was probably even more widespread

than it is today. Over a long period of time, many words which
had originally been derived by reduplication gradually changed
in form and meaning. As a result, we have quite a few words
today which contain “fossilized” remains of reduplication which
many Palauan speakers do not even recognize.
Among simple state verbs, the following seem to contain fos-
silized reduplicated portions, which we have italicized:

(33) kikiongȩl ‘dirty’ kȩdidai ‘high’

cheleleu ‘pale’ dȩchudȩch ‘muddy’
chachau ‘stunted, empty (of nuts)’

A couple of state verbs in mȩ- whose stems probably have redu-

plicated portions include the following:

(34) mȩsisiich ‘strong’

mȩkȩlȩkolt ‘cold’
mȩrȩchȩrachȩd ‘nauseous’

In 4.9.4 we mentioned some of the variant forms of the

Palauan object pronouns. We saw that the 3rd pers. pl. human
object pronoun, which is usually -tȩrir, turns up as -titȩrir oblig-
atorily in ngoititȩrir ‘take them’ and optionally in obȩ(ti)tȩrir
‘carry them’. The most plausible explanation for the additional
syllable ti is that it is a fossilized trace of reduplication.

11.12. THE PREDICTIVE AND INCHOATIVE

SUFFIXES
In this and the following sections, we will examine two different
suffixes which when added to verb stems bring about important
changes in meaning. The predictive suffix -u is used to des-
ignate an action which is about to happen—i.e., one which the
speaker judges to be imminent. The inchoative suffix -a makes
reference to a new or unexpected action or state, one which
has just begun or is in its beginning stages. 18 Both of these suf-
fixes are always stressed, with the result that various kinds of
vowel reduction (cf. 1.4.4, 3.4, 3.4.1–3, and 6.4) and vowel
blending (cf. 6.3.2) are observed in the verb forms to which
they are attached. Further, both suffixes require an additional
-ng when they occur in sentence-final position (cf. 1.3.3). The

278
 Palauan Reference Grammar

predictive suffix -u can attach only to intransitive action verbs

and to the ergative and imper fective forms of some tran-
sitive action verbs, while the inchoative suffix -a can attach to
these verb types and to state verbs as well. Therefore, any verb
form which takes -u can take -a, but not necessarily vice versa.

11.12.1 PREDICTIVE AND INCHOATIVE FORMS OF INTRANSITIVE

ACTION VERBS
Intransitive action verbs can take both the predictive and
inchoative suffixes, as in the examples below:

(35) Predictive Form Inchoative Form Related Intransitive

Action Verb
sobȩku ‘about to sobȩka ‘starting to suebȩk ‘fly’
fly’ fly’
rurtu ‘about to rurta ‘starting to rȩmurt ‘run’
run’ run’
mȩrolu ‘about to mȩrola ‘starting to mȩrael ‘leave’
leave’ leave’
robȩtu ‘about to robȩta ‘starting to ruebȩt ‘fall’
fall’ fall’
tobȩdu ‘about to tobȩda ‘starting to tuobȩd ‘go out’
go out’ go out’
longȩlu ‘about to longȩla ‘starting to lmangȩl ‘cry’
cry’ cry’

In suébȩk ‘fly’ and ruébȩt ‘fall’, we observe the metathesized

verb marker -u- (cf. 6.2) adjacent to e in a stressed syllable.
In the predictive and inchoative forms sobȩkú ‘about to fly’ and
sobȩká ‘starting to fly’, however, the vowel cluster ue comes
to appear in an unstressed syllable. Here, just as in the case of
the 3rd pers. sg. object present perfective forms cited in 6.3.2,
the metathesized verb marker -u- blends with the following e
in an unstressed syllable to give the vowel o. (Recall that the
vowel triangle shows o to be phonetically halfway between u
and e). The same phenomenon accounts for the predictive and
inchoative forms of lmangȩl ‘cry’, except that the metathesized
verb marker -m- first changes to -u- in an unstressed syllable (cf.
6.3.2). Thus, we have the following derivation for longȩlú ‘about
to cry’:

279
 11 Reduplication and Further Verb Affixation

(36) mȩ + langȩl (basic form = verb marker + verb stem +

+ú predictive suffix)
l + mȩ + (by metathesis of verb marker)
angȩl + ú
l + m + angȩl (by deletion of ȩ)
+ú
l + u + angȩl (by change of verb marker to u in
+ú unstressed syllable)
l + ongȩl + ú (by vowel blending)

The remaining predictive and inchoative forms of 35 above

show no unfamiliar phonetic changes. In tobȩdú and tobȩdá, for
example, the vowel cluster uo of tuóbȩd ‘go out’ has reduced to
a single vowel in an unstressed syllable. And in rurtú and rurtá,
the metathesized verb marker has been deleted before the high
back vowel u (cf. 6.3.3); thus, we have the following derivation
for rurtú ‘about to run’:

(37) mȩ + rurt + (basic form = verb marker + verb stem +

ú predictive suffix)
r + mȩ + urt (by metathesis of verb marker)
+ú
r + m + urt (by deletion of ȩ)
+ú
r + u + urt (by change of verb marker to u in unstressed
+ú syllable)
r + urt + ú (by deletion of verb marker)

Finally, in mȩrolú and mȩrolá, we notice that the ae of the stem

ráel ‘road’ has reduced to the single vowel o in an unstressed
syllable (cf. note 10 above).
In the pairs of sentences below, we illustrate how the pre-
dictive and inchoative forms of 35 are used. Recall that these
forms are spelled and pronounced with a final -ng when they
occur at the end of a sentence.

(38) a. A skoki a sobȩkung.

‘The plane is about to take off.’

b. A skoki a sobȩkang.
‘The plane is taking off/starting to fly.’

(39) a. Ak mȩrolu er a elȩchang.

280
 Palauan Reference Grammar

‘I’m about to leave now.’

b. A Toki a mȩrolang.
‘Toki is beginning to leave/is just leaving.’

(40) a. A ngalȩk a longȩlung.

‘The child is about to cry.’

b. A Droteo a milȩkekui a ngalȩk mȩ ng longȩlang.

‘Droteo teased the child, so he’s begun to cry.’

11.12.2. PREDICTIVE AND INCHOATIVE FORMS OF ERGATIVE VERBS

When the predictive and inchoative suffixes are added to the
ergative forms of transitive verbs, the resulting words have the
expected interpretations ‘about to be/get…’ and ‘has begun to
be/get…’, respectively, as in the examples below:

(41) Predictive Form Inchoative Form Related

Ergative Verb
obosu ‘about to be/ obosa ‘has begun to oboes ‘be/get
get shot’ be/get shot’ shot’
mȩchȩlȩbȩdu ‘about mȩchȩlȩbȩda ‘has mȩchȩlebȩd
to be/get hit’ begun to be/get hit’ ‘be/get hit’
mȩrȩsmu ‘about to mȩrȩsma ‘has begun mȩrasm ‘be/
be/get sewn’ to be/get sewn’ get sewn’
mȩtȩmȩllu ‘about to mȩtȩmȩlla ‘has begun mȩtȩmall
break down’ to break down’ ‘break down’

You should have no difficulty identifying the kinds of vowel re-

duction and vowel cluster reduction which have taken place in
the predictive and inchoative forms of 41.
The following pairs of sentences show the use of the pre-
dictive and inchoative forms given in 41:

(42) a. A bȩlochȩl a obosu ȩr a Droteo.

‘The pigeons are about to be/get shot by
Droteo.’

b. A bȩlochȩl a obosa ȩr a Droteo.

‘The pigeons have begun to be/get shot by
Droteo.’

281
 11 Reduplication and Further Verb Affixation

(43) a. Ng mȩchȩlȩbȩdu a ngalȩk ȩr a dȩmal.

‘The child is about to get hit by his father.’

b. Ng mȩchȩlȩbȩda a ngalȩk ȩr a dȩmal.

‘The child has begun to be/get hit by his
father.’

11.12.3. PREDICTIVE AND INCHOATIVE FORMS OF IMPERFECTIVE

VERBS
Imperfective verbs which take both predictive -u and in-
choative -a include the following:

(44) Predictive Form Inchoative Form Related

Imperfective
Verb
mȩlȩkingu 19 ‘about mȩlȩkinga19 mȩlȩkoi ‘talk’
to talk’ ‘starting to talk’
omȩkȩllu ‘about to omȩkȩlla ‘starting to omȩkall ‘drive,
drive/sail’ drive/sail’ sail’
mȩngȩsmȩru mȩngȩsmȩra mȩngȩsimȩr
‘about to shut’ ‘starting to shut’ ‘shut’
omrȩchu ‘about to omrȩcha ‘starting to omurȩch ‘spear’
spear’ spear’
mȩlȩchȩlbu ‘about mȩlȩchȩlba ‘starting mȩlȩcholb ‘wash’
to wash’ to wash’

In addition to rules of vowel reduction and vowel cluster re-

duction, a rule of vowel deletion (cf. 3.4.1) affects the pre-
dictive and inchoative forms in 44. Can you tell which rule
applies where?
Some sentence pairs containing the predictive and incho-
ative forms of 44 are given below:

(45) a. A Droteo a mȩlȩkingu ȩr a dȩmal a Toki ȩl kirel a

chȩbȩchiiȩlir.
‘Droteo is about to talk to Toki’s father about their
marriage.’

b. A Droteo a mȩlȩkinga ȩr a dȩmal a Toki ȩl kirel a

chȩbȩ-chiiȩlir.
‘Droteo has started talking to Toki’s father about their
marriage.’

282
 Palauan Reference Grammar

(46) a. A Toki a mȩlȩchȩlbu ȩr a rȩngalȩk.

‘Toki is about to bathe the children.’

b. A Toki a mȩlȩchȩlba ȩr a rȩngalȩk.

‘Toki has begun to bathe the children.’

The sentences below contain further examples of the in-

choative forms of imperfective verbs. The addition of ko ȩl
20
‘just’ reinforces the connotation that a new or unexpected
action has taken place in the very recent past.

(47) a. Ak ko ȩl mȩsuba e le ng mla mȩrael a Droteo.

‘I’ve finally gotten to study because Droteo has
left.’

b. Ak ko ȩl rongȩsa a chais.
‘I’ve just heard the news.’

c. A Droteo a ko ȩl omȩchȩla ȩl mȩsuub.

‘Droteo has just begun to study.’

11.12.4. INCHOATIVE FORMS OF STATE VERBS

State verbs can generally take the inchoative suffix -a, al-
though they cannot take the predictive suffix -u. The function
of -a with state verbs is to show that a change of state is in
progress; in most cases, the state in question is unwanted and
therefore unexpected. Observe the following examples:

(48) Inchoative Form Related State Verb

mȩkȩlȩkȩlta ‘getting cold’ mȩkȩlȩkolt ‘cold’

mȩkelda ‘getting warm’ mȩkeald ‘warm’
songȩrȩngȩra ‘getting hungry’ songȩrengȩr ‘hungry’
kikȩngȩla ‘getting dirty’ kikiongȩl ‘dirty’
mȩrȩngȩla ‘getting sore’ mȩringȩl ‘sore’
mȩchȩrȩchȩra ‘getting salty’ mȩchȩrochȩr ‘salty’
sochȩra ‘getting sick’ smechȩr ‘sick’
bȩrȩlma ‘getting watery/ bȩralm ‘watery,
flat-tasting’ flat-tasting’
mȩdȩkta ‘becoming frightened’ mȩdakt ‘afraid’
mȩrka ‘getting ripe’ marȩk ‘ripe’

283
 11 Reduplication and Further Verb Affixation

mȩkȩrȩnga 21 ‘waking up’ mȩkar ‘awake’

klunga 22 ‘getting big’ klou ‘big’
ungia 23 ‘becoming good, ungil ‘good’
improving’
chȩlla ‘getting rainy’ chull ‘rainy’

Some state verbs take -o or -e as an inchoative suffix instead

of -a. This appears to be an unpredictable property of verbs
such as the following: 24

(49) Inchoative Form Related State Verb

mȩde ‘starting to die’ mad ‘dead’

dȩchȩro ‘standing up (process)’ dȩchor ‘stand’
mȩchȩde ‘getting shallow’ mȩched ‘shallow’
doknge ‘getting together’ dmak ‘together’
mȩkngte ‘getting bad, worsening’ mȩkngit ‘bad’

The following sentences illustrate the use of the inchoative

verbs listed in 48 and 49:

(50) a. Tia ȩl delmȩrab a kmal mȩkeldang.

‘This room is getting very warm.’

b. A chimak a kikȩngȩla mȩ ng kirek ȩl mo mȩlȩbal.

‘My hands are getting dirty, so I’ve got to wash
them.’

c. A ngalȩk a mȩkȩrȩnga ȩr a chȩrrodȩch.

‘The child is waking up because of the noise.’

d. A bdȩluk a mȩrȩngȩlang.
‘I’m getting a headache.’

e. A kall a {mȩchȩrȩchȩrang/bȩrȩlmang}.
‘The food’s getting (too) {salty/flat}.’

f. A eangȩd a ungia mȩ dorael.

‘The weather’s improving, so let’s go.’

g. Ng mȩde a ngau.
‘The fire is (just) dying out.’

h. A rȩngalȩk a doknge a rȩngrir.

284
 Palauan Reference Grammar

‘The children are starting to get along with each

other.’

i. Ng chȩlla mȩ lak dorael.

‘It’s getting rainy, so let’s not go.’

11.12.5. PREDICTIVE AND INCHOATIVE FORMS OF mo

The verb mo ‘go’ has mochu as its predictive form and mocha
as its inchoative form; the appearance of -ch- before the suffixes
-u and -a is unpredictable. 25 Some sentences containing these
forms are given below:

(51) a. Ak mochu ȩr a mubi.

‘I’m about to go to the movie.’

b. Ng mochu ȩr ngii a chull.

‘It’s about to rain.’

c. A tangk a mochu mui.

‘The tank is about to get full.’

d. Ak mochu mȩrek ȩl mȩsuub.

‘I’m about to finish studying.’

e. A Toki a mochu omȩngur.

‘Toki is about to eat.’

f. A Droteo a ko ȩl mocha ȩr a Guam.

‘Droteo should just be arriving in Guam.’

g. A Droteo a ulȩrrimȩl ȩr a Helen mȩ ng ko ȩl

mocha mȩlasȩm ȩl mȩnga a ngikȩl.
‘Droteo persuaded Helen to finally try to eat fish.’

Reciprocal verbs, causative verbs, and verbs formed with

the prefix ou- do not take either the predictive suffix -u or
the inchoative suffix -a. To express predictive or inchoative
meanings with these types of verbs, mochu and mocha, respec-
tively, are required as auxiliaries or “helping” words. Observe
the following examples:

(52) a. A rȩngalȩk a mochu kaiuȩkako.

285
 11 Reduplication and Further Verb Affixation

‘The children are about to start teasing

each other.’

b. A Toki a mochu omȩka ȩr a rȩngalȩk.

‘Toki is about to feed the children.’

c. A Romana a mocha oureor.

‘Romana is starting to work.’

d. Tȩ mocha kaingȩseu.
‘They’re starting to help each other.’

11.12.6. THE PREDICTIVE WORD ku

Identical in function to the predictive suffix -u is the inde-
pendent predictive word ku (kung when in sentence-final po-
sition), which can immediately follow any type of verb. Since ku
does not cause any kind of vowel reduction or deletion in the
verb stem which precedes it, we analyze it as a separate word
rather than as a suffix. The use of ku is illustrated in the sen-
tences below; like the predictive suffix -u, its closest English
equivalent is ‘about to’.

(53) a. Ak mȩnguiu ku ȩr a hong.

‘I’m about to read the book.’

b. A Toki a olȩkar ku ȩr a rȩngalȩk.

‘Toki is about to wake up the children.’

c. Ak mȩlim ku e le ng mȩchȩde a rȩnguk.

‘I’ll have something to drink (now) because I’m getting
thirsty.’

d. Kȩ mȩkȩra kung?
‘What are you about to do?’

e. Alii. A stoa a mȩchȩsimȩr kung.

‘Hey! The store’s about to close.’

f. Ak mȩsuub ku e le ng ngar ȩr ngii a skeng ȩr a klukuk.

‘I’ll be studying now because there’s a test tomorrow.’

g. Ak mȩrolu kung.

286
 Palauan Reference Grammar

‘I’m just about to leave.’

h. A. Toki a mochu omȩngur kung.

‘Toki is about to eat.’

As examples like 53g–h show, it is possible to have both the pre-

dictive suffix -u and the independent predictive word ku in the
same sentence. Note further that 53h and 51e are identical in
meaning.

287
 12 Imperfective vs. Perfective
Verbs

12Imperfectivevs.PerfectiveVerbs
12.1. MEANING OF IMPERFECTIVE VS.
PERFECTIVE VERB FORMS
In previous chapters we have already given extensive consider-
ation to the internal structure of imperfective and perfective
verb forms. Thus, in 5.5 we saw that these two types of verbs
have different basic structures, as given below:

(1) a. imperfective verb verb marker + imperfective marker

forms: + verb stem

b. perfective verb verb marker + verb stem + object

forms: pronoun

We also discussed in detail each of the elements found in the

basic structures of 1: in chap. 6 we examined the verb marker
and showed how it metathcsizes in perfective verb forms; in
5.5 we saw how the variants of the imperfective marker can
be determined from the following verb stem; and in 4.9 and
4.9.4 we classified the object pronouns which are characteris-
tically suffixed to perfective verb forms.
The abovementioned discussions, together with the many
lists and step-by-step derivations presented in chaps. 5 and 6,
have shown us how to distinguish between imperfective and
perfective verbs from a formal point of view. In this chapter,
then, we will examine how the two types of verbs contrast with
each other in terms of meaning and use. As native speakers
of Palauan, you may find the basic distinctions to be described
below rather difficult to grasp. This is because you have learned
to make them so automatically (or unconsciously) that you do
not think in terms of analyzing them. The same is true, of
course, for the native speakers of any language; thus, without
any assistance from a linguist or teacher, most English
speakers, for example, would be totally unaware of the ex-
tremely complex way in which they use the English tense
system.

288
 Palauan Reference Grammar

The distinction between imperfective vs. perfective verb

forms is found only among transitive verbs (cf. 5.1.1); with
one or two exceptions, these are all transitive action verbs.
As their names imply, perfective verb forms describe actions
which the speaker views as completed (or perfected), while
imperfective verb forms do not. Instead, imperfective verb
forms denote actions which the speaker considers as incom-
plete—i.e., actions which are still in progress 1 and have not
yet reached their point of completion or termination. A par-
ticular action can be viewed as complete or incomplete re-
gardless of when it actually occurs; therefore, we will find
that both perfective and imperfective verbs can occur in all
tenses—past, present, and future.
By way of introduction, let us review the following two sen-
tences, which we already presented in 5.5 (examples 59a–b):

(2) a. A Droteo a milȩnguiu ȩr a hong er a

elii.
‘Droteo was reading the book
yesterday.’

b. A Droteo a chiliuii a hong er a elii.

‘Droteo read the book yesterday.’

In the sentences above, we observe the past imperfective form

milȩnguiu and the past perfective form chiliuii. Because both
milȩnguiu and chiliuii are past tense forms, they necessarily
have a common element of meaning: that is, they both indicate
that the subject (Droteo) performed the action of reading at
some time point in the past (elii ‘yesterday’). But the similarity
ends here, since the two verb forms involve a very basic dif-
ference in the speaker’s viewpoint. Thus, in 2a the speaker uses
imperfective milȩnguiu ‘was reading’ to focus on the past action
as it was going on or in progress. The action is described as
having continued for some period of time, but no claim is made
that it was completed. In other words, it is entirely possible
that Droteo still has some of the book to read. By contrast,
the speaker’s use of perfective chiliuii ‘read (completely)’ in 2b
views Droteo’s reading of the book as a completely finished past
action. It is implied that Droteo has no more of the book to read.
Some further pairs of contrasting sentences are given
below:

289
 12 Imperfective vs. Perfective Verbs

(3) a. A ngalȩk a milȩnga a ngikȩl.

‘The child was eating the fish.’

b. A ngalȩk a killii a ngikȩl.

‘The child ate up the fish.’

(4) a. A Cisco a millim a rrom.

‘Cisco was drinking liquor.’

b. A Cisco a ngilȩlmii a rrom.

‘Cisco drank up the liquor.’

Notice that the past imperfective forms in the a-sentences

are best translated with English past progressive forms (was/
were… -ing), while the past perfective forms in the b-sen-
tences have English equivalents like ate up and drank up, in
which a word like up (called a particle or intensifier by
English grammarians) imparts a special connotation of com-
pletion to the actions eat and drink. In the a-sentences, use of
an imperfective verb implies that the action in question did not
totally exhaust or consume the object: in other words, in 3a,
the speaker assumes that there was still some fish left after the
child was eating, and in 4a, the subject (Cisco) did not drink all
of the liquor. By contrast, use of a perfective verb in the b-sen-
tences implies that the objects (ngikȩl ‘fish’ and rrom ‘liquor’)
were totally used up.

12.2. IMPERFECTIVE VS. PERFECTIVE VERB

FORMS IN VARIOUS TENSES
As we have seen in examples 2–4 above, the use of imperfective
vs. perfective forms in the past tense results in a very clear
contrast in meaning. Similar, or related, contrasts in meaning
are found when we observe imperfective and perfective forms
in other tenses. For example, in the pair of sentences below, the
auxiliary word mla (cf. 5.3.2.1) precedes the imperfective and
perfective forms of the action verb mȩngiis ‘dig’ and refers to
an event or activity in the very recent past:

(5) a. Ak mla mȩngiis ȩr a kliokl.

‘I’ve been digging the hole.’

290
 Palauan Reference Grammar

b. Ak mla kiiȩsii a kliokl.

‘I’ve (completely) dug the hole.’

In 5a, the speaker’s use of imperfective mȩngiis implies that al-

though he has been engaged in the activity of digging the hole,
the task is not yet complete; on the other hand, use of per-
fective kiiȩsii in 5b is a clear assertion that the job of digging
the hole has been completely finished. A similar distinction
of meaning is found in the sentences below, which contain
the imperfective and perfective forms of the causative verb
omȩkikiongȩl ‘make… dirty’ (cf. 9.2.1.3 and 9.4, ex. 22e):

(6) a. A rȩkangkodang a mla omȩkikiongȩl ȩr a

kȩdȩrang.
‘The tourists have been making the beach dirty.’

b. A rȩkangkodang a mla mȩkikingȩlii a kȩdȩrang.

‘The tourists have (totally) messed up the beach.’

Example 6b, with perfective mȩkikingȩlii, has a much stronger

connotation of “finality” than 6a, where imperfective omȩkikio
ngȩl leaves open the possibility that the process of dirtying
up the beach might not reach its ultimate (unpleasant or irre-
versible) conclusion.
As we saw in 5.3.2.1, the auxiliary mla can also be used
to express past experience—i.e., to make a statement or ask
a question about whether someone has had the experience of
doing something. With transitive verbs, this connotation of past
experience is conveyed by using the imperfective form after
mla, as in the a-sentences below. These sentences contrast, of
course, with the b-sentences, which contain a perfective form
following mla:

(7) a. Kȩ mla mȩruul a kall ȩr a Sina?

‘Have you ever made Chinese food?’

b. Kȩ mla rullii a kall ȩr a Sina?

‘Have you finished making the Chinese
food?’

(8) a. Kȩ mla mȩnguiu ȩr tia ȩl hong?

‘Have you ever read this book?’

291
 12 Imperfective vs. Perfective Verbs

b. Kȩ mla chuiȩuii tia ȩl hong?

‘Have you finished reading this book?’

While the a-sentences above are interpreted as general ques-

tions about one’s past experience, the b-sentences are questions
about whether some activity was completed on a single, spe-
cific occasion.
Imperfective and perfective verb forms can be preceded by
the auxiliary mo to designate actions in the future (cf. 5.3.3).
There is nothing unusual about the meaning difference between
mo + imperfective verb and mo + perfective verb, as the fol-
lowing pair of sentences shows:

(9) a. Ak mo mȩngiis ȩr a kliokl ȩr a klukuk.

‘I’ll be digging the hole tomorrow.’

b. Ak mo kiiȩsii a kliokl ȩr a klukuk.

‘I’ll (completely) dig the hole tomorrow.’

In 9a, mo + imperfective mȩngiis implies that the activity of

digging will be going on tomorrow, but nothing is said about
whether or not the task will be completed. By contrast, in 9b mo
+ perfective kiiȩsii expresses the speaker’s conviction that the
job can be finished tomorrow. Another pair of sentences similar
to 9a–b is the following:

(10) a. Aki mo mȩnga a kall ȩl ngar ȩr a icebox.

‘We’ll eat some of the food in the icebox.’

b. Aki mo kma a kall ȩl ngar ȩr a icebox.

‘We’ll eat up the food in the icebox.’

While mo + imperfective mȩnga in 10a implies that only some

of the food will be eaten—i.e., the object kall ‘food’ will not be
totally consumed by the action of eating—the sequence mo +
perfective kma in 10b makes it clear that the object will be com-
pletely used up (cf. our discussion of examples 3–4 above).
The present tense forms of imperfective and perfective
verbs also show an important contrast in meaning. Imperfective
verb forms in the present tense, as we have seen in 5.3.1,
have two possible functions. Their primary function is to des-

292
 Palauan Reference Grammar

ignate an action which is going on or in progress at the present

moment—i.e., at the time when the speaker utters the sentence.
This function is observed in the following sentences:

(11) a. A John a mȩngȩsbrebȩr ȩr a blik.

‘John is painting my house.’

b. Ak mȩlasȩch ȩr a mlik er a elȩchang.

‘I’m carving my canoe now.’

A secondary function of imperfective verb forms in the present

tense is to express general statements or habitual state-
ments, as in the examples below:

(12) a. A John a mȩngȩsbrebȩr a blai.

‘John paints houses (as a
profession).’

b. Ak mȩlasȩch a mlai.
‘I carve canoes (as a profession).’

Unlike the sentences of 11, which refer to specific occasions,

the examples of 12 look at certain actions in a general way.
Thus, while 11a would only be true if the subject (John) were
actually painting the speaker’s house at the moment of the
speaker’s utterance, 12a could be uttered at any time because
it is simply a statement about what the subject does habitually
(as a profession, etc.). Notice, further, that the specific occa-
sions designated in 11a–b require specific objects; therefore,
the possessed nouns blik ‘my house’ of 11a and mlik ‘my
canoe’ of 11b must be preceded by the specifying word ȩr (cf.
2.7). By contrast, the general statements expressed in 12 can
only involve objects of the most general type. Here, the unpos-
sessed nouns blai ‘house’ and mlai ‘canoe’ are not introduced
by the specifying word ȩr and merely refer to houses or canoes
in general—i.e., ‘any house at all’, ‘any canoe at all’, etc.
As opposed to the above, Palauan perfective verb forms in
the present tense have a very different, and rather special, func-
tion. To repeat what we said in 4.9.2, perfective verb forms in
the present tense are used to denote actions or events which the
speaker considers imminent—i.e., actions or events which are
just about to occur or which are likely to occur in the very near
future. Compare, for example, the following two sentences:

293
 12 Imperfective vs. Perfective Verbs

(13) a. Ak mȩluchȩs ȩr a babier er a elȩchang.

‘I’m writing the letter now.’

b. Ak luchȩsii a babier er a elȩchang (mȩ kȩ

mȩsang).
‘I’ll write the letter right now (so you can see me
do it).’

In 13a, imperfective mȩluchȩs simply denotes that the action of

writing is taking place (i.e., in progress) at the present moment,
while in 13b perfective luchȩsii implies that the speaker is just
about to start writing the letter.
Because present perfective verb forms have the above-men-
tioned connotation of imminency, they are often found in sen-
tences which are used as warnings or precautionary sugges-
tions. A few typical examples are given below:

(14) a. Alii. A ngalȩk a chubȩlii a milk!

‘Watch out! The child’s about to spill the milk!’

b. Alii. A malk a kolii a bȩras!

‘Hey! The chicken’s about to eat the rice!’

c. Alii. A sensei a cholȩbȩdau!

‘Watch out! The teacher is going to hit you (by
accident)!’/
‘The teacher will hit you (if you misbehave).’

12.3. FURTHER EXAMPLES OF CONTRAST

BETWEEN IMPERFECTIVE AND PERFECTIVE
VERB FORMS
In this section we will look at additional pairs of sentences
that involve constructions in which the use of imperfective vs.
perfective forms results in a significant difference in meaning.
Observe, for example, the sentences below:

(15) a. Ng sȩbȩchem ȩl mȩnga a ngikȩl?

‘Are you able to eat fish?’

b. Ng sȩbȩchem ȩl kolii a ngikȩl?

‘Can you eat up the fish?’

294
 Palauan Reference Grammar

In the sentences of 15, the possessed noun sȩbȩchem ‘your

ability’ is followed by a group of words (mȩnga a ngikȩl or
kolii a ngikȩl) which gives us specific information about the
type of ability involved. This group of words is preceded by ȩl ,
which, as we will see in 17.7, relates it to the possessed noun
sȩbȩchem. While the use of imperfective mȩnga in 15a results
in a general question, the use of perfective kolii in 15b makes
reference to a specific instance or occasion. In other words,
in 15a, some person X asks another person Y whether or not Y
is an eater of fish: perhaps X is concerned that Y might not be
used to eating fish regularly, or perhaps X is just inquiring in a
polite way whether or not Y likes fish. In 15b, on the other hand,
X is asking Y whether Y can eat up some particular fish. While
15b can only be uttered on some particular occasion when the
fish is actually there on the table, etc., 15a is of course not re-
stricted in this way. The following sentences show the same type
of contrast:

(16) a. Ng sȩbȩchem ȩl mȩlim a biang?

‘Are you able to drink beer?’

b. Ng sȩbȩchem ȩl ngilmii a imȩlem ȩl biang?

‘Can you drink up all your beer?’

Note that the object in 16a is general (biang ‘beer’), while that
in 16b is very specific (imȩlem ȩl biang ‘your (drink of) beer’).
In the sentences below, we observe the possessed nouns
soak ‘my liking’ and kirek ‘my obligation’, which are in the same
class as sȩbȩchem ‘your ability’ of 15 and 16 above (see chap.
17). These possessed nouns are related by the word ȩl to groups
of words which provide specific information about one’s liking
or one’s obligation:

(17) a. Ng soak ȩl mȩnguiu ȩr a hong.

‘I want to read (some of) the book.’

b. Ng soak ȩl chuiȩuii a hong.

‘I want to (completely) read the
book.’

(18) a. Ng kirek ȩl mȩngiis ȩr a kliokl.

‘I’ve got to be digging the hole.’

295
 12 Imperfective vs. Perfective Verbs

b. Ng kirek ȩl kiiȩsii a kliokl.

‘I’ve got to (completely) dig the
hole.’

The meaning contrast found in the sentence pairs of 17–18

should already be familiar to us from examples 3–6 above. Thus,
in 17a and 18a the action is viewed as being in progress, but not
necessarily completed, while in 17b and 18b it is seen as being
completed.

12.4. IMPERFECTIVE VS. PERFECTIVE VERB

FORMS AND SPECIFIC OBJECTS
In 2.7 we noted that the specifying word ȩr must always
precede specific singular objects. The presence or absence
of the specifying word ȩr results in an important difference in
meaning, as the sentences below illustrate:

(19) a. Ak ousbech ȩr a bilas ȩr a klukuk.

‘I need the boat tomorrow.’

b. Ak ousbech a bilas ȩr a klukuk.

‘I need {a boat/the boats}
tomorrow.’

While objects marked with ȩr (e.g. ȩr a bilas ‘the boat’ of

19a) are automatically interpreted as specific singular objects,
those not marked with ȩr (e.g. a bilas ‘a boat, the boats’ of 19b)
allow for a wider range of interpretation, as indicated.
The use of the specifying word ȩr is restricted in two inter-
esting ways. First, it never occurs with sentence subjects, but
only with sentence objects; and second, it can only precede
the objects of imperfective verbs. With the objects of per-
fective verbs, use of the specifying word ȩr is prevented, as the
ungrammaticality of 20b shows:

(20) a. Ak chillȩbȩdii a bilis.

‘I hit the dog.’

b. *Ak chillȩbȩdii ȩr a bilis.

296
 Palauan Reference Grammar

If we review all of the example sentences so far given in this

chapter, we will see that they meet the abovementioned restric-
tions on the distribution of the specifying word ȩr.
It is possible to speculate on the reasons why the specifying
word ȩr cannot precede the objects of perfective verbs. Every
perfective verb form, you will recall, contains an object
pronoun suffix. Since all Palauan pronouns refer to specific
persons (‘I, me’, ‘he, him’, etc.), it seems as if perfective verb
forms, which include a pronoun in the form of a suffix, focus on
the completion of an action with reference to some specific
object. In other words, Palauan perfective verb forms in and
of themselves imply a specific object, and use of ȩr to further
indicate the specificity of the object would be redundant and
therefore unnecessary. Thus, ȩr would be prevented in a sen-
tence like 20b since it would add nothing new to the meaning
which was not already supplied by the perfective verb form
itself.

12.5. CONTEXTUAL RESTRICTIONS ON THE USE

OF IMPERFECTIVE VS. PERFECTIVE VERB FORMS
As we have seen in the sections above, imperfective and per-
fective verb forms contrast with each other in terms of certain
fundamental differences of function and meaning. Our analysis
of the nature of these differences is confirmed when we look at
the contextual (or environmental) restrictions on the occur-
rence of the two types of forms. In other words, we can find or
construct contexts in which one type of verb form (imperfective
or perfective) is completely natural while the other is strange or
contradictory. The differences of acceptability are undoubtedly
due to the compatibility—or lack of compatibility—between the
meaning of the imperfective or perfective form in question and
the meaning of the other elements in the context.
To take our first example, observe that there is nothing at all
unusual about 21a below, but 21b seems to make no sense:

(21) a. Ak mla mȩnguiu ȩr a hong e ng di dirkak kbo kmȩrek.

‘I have been reading the book, but I haven’t finished it
yet.’

b. ??Ak mla chuiȩuii a hong e ng di dirkak kbo kmȩrek.

297
 12 Imperfective vs. Perfective Verbs

The key to the difference in acceptability between 21a and 21b

lies, of course, in the fact that 21a has an imperfective verb
form while 21b has a perfective verb form. In the sentences
above, each of these verb forms has been placed in the context
of the expression e ng di dirkak kbo kmȩrek ‘but I haven’t
finished yet’, which clearly states that the subject (ak ‘I’) has
not yet completed reading the particular book. Because imper-
fective mla mȩnguiu of 21a implies that the subject has so far
read some of the book, but not all of it, e ng di dirkak kbo
kmȩrek is “logical” and makes sense. By contrast, since per-
fective mla chuiȩuii of 21b explicitly states that the subject has
finished reading the book, it is contradictory, and therefore un-
acceptable, to add e ng di dirkak kbo kmȩrek, which implies ex-
actly the opposite. 2
The very same phenomenon accounts for the following ex-
amples, in which the imperfective and perfective forms of
mȩngiis ‘dig’ occur in the future tense and the content of the
expression introduced by e ng di ‘but’ is somewhat different:

(22) a. Ak mo mȩngiis ȩr a kliokl ȩr a klukuk e ng di diak

kudȩnge ȩl kmo ng sȩbȩchek ȩl rokir ng diak.
‘I’ll be digging the hole tomorrow, but I don’t know
whether or not I can finish it.’

b. ??Ak mo kiiȩsii a kliokl ȩr a klukuk e ng di diak kudȩnge

ȩl kmo ng sȩbȩchek ȩl rokir ng diak.

Example 22a makes perfect sense, but 22b involves a contradic-

tion and is therefore unacceptable. The contradiction in 22b can
be explained as follows: on the one hand, mo kiiȩsii implies the
speaker’s intention to complete the task of digging the hole at
some future date (klukuk ‘tomorrow’), while on the other hand,
the content of the expression e ng di diak kudȩnge ȩl kmo ng
sȩbȩchek ȩl rokir ng diak ‘but I don’t know whether or not I can
finish it’ expresses the speaker’s doubt about finishing this very
same task.
If the context mentions a span of time with a specific
beginning point and end point, only sentences containing im-
perfective verb forms are acceptable, as the following example
shows:

298
 Palauan Reference Grammar

(23) a. A Droteo a milȩngiis ȩr a kliokl ȩr a euid ȩl klok ȩl mo

eai ȩl klok ȩr a tutau.
‘Droteo was digging the hole from 7 o’clock to 8 o’clock
this morning.’

b. ??A Droteo a kilisii a kliokl ȩr a euid ȩl klok ȩl mo eai ȩl

klok ȩr a tutau.

In the sentences of 23 we have the time span expression ȩr

a euid ȩl klok ȩl mo eai ȩl klok ‘from seven o’clock to eight
o’clock’. 3 Since this expression designates a one-hour stretch of
time, it can occur together with imperfective milȩngiis in 23a,
which views the activity of digging as it was in progress—i.e.,
as having a duration. By contrast, this same time span ex-
pression cannot occur together with perfective kilisii in 23b
without sounding strange. This is because kilisii focuses our at-
tention on the very moment when the activity of digging was
completed, and therefore a sentence like 23b would seem to
imply that the moment of completion lasted a whole hour, which
is of course impossible.
Not all contexts are like those of 21–23 above in restricting
the use of imperfective and perfective verb forms. In the sen-
tences below, for example, which contain a time clause intro-
duced by er se ȩr a ‘when’ (see 22.2), both the imperfective and
perfective forms of mȩnguiu ‘read’ can occur, with a difference
in interpretation:

(24) a. Ak milȩnguiu ȩr a hong er se ȩr a lȩme a Toki.

‘I was reading the book when Toki arrived.’

b. Ak chiliuii a hong er se ȩr a lȩme a Toki.

‘I had finished reading the book when Toki arrived.’

In 24a, the speaker was in the process of reading the book (but
had not completed reading it) when Toki arrived, while in 24b
the speaker had already completed reading the book by the time
of Toki’s arrival.

299
 12 Imperfective vs. Perfective Verbs

12.6. THE TRANSITIVE VERB OMES

The Palauan verb omes ‘see’ needs special consideration be-
cause the meanings of its imperfective and perfective forms
differ from each other in a way which is not entirely predictable
from our discussion above. Before examining this difference in
meaning, let us review a few of the representative perfective
forms of omes: 4

(25) Person and Number of Present Tense Past Tense

Object Pronoun

1st pers sg mȩsȩkak ‘see milsȩkak ‘saw

me’ me’

3rd pers sg mȩsa ‘see him/ milsa ‘saw

her/it’ him/ her/it’

3rd pers pl hum mȩs(e)tȩrir milstȩrir ‘saw

‘see them’ them’

3rd pers pl non-hum mes ‘see them’ miles ‘saw

them’

The imperfective vs. perfective forms of omes convey the fol-

lowing differences of meaning. Use of imperfective omes (past:
ulȩmes) generally implies purposeful or intentional seeing on
the part of the subject. Therefore, the most appropriate English
equivalents are words or expressions like ‘meet’, ‘get together
with’, ‘visit’, ‘meet and talk to’, ‘look at’, ‘watch’, etc. By con-
trast, use of the perfective forms of omes connotes uninten-
tional, unplanned, or casual seeing by the subject. The best
English equivalents in this case would perhaps be expressions
like ‘happen to see’ or ‘get a glimpse of’.
The difference under discussion here is not really incompati-
ble with our previous analysis of the difference between imper-
fective and perfective verb forms (cf. 12.1–2 above). Thus, it
does not seem unusual that the imperfective forms of omes
should be used to describe “serious” actions of seeing whose
durational or progressive quality the speaker would be likely
to focus upon. Similarly, we would expect the perfective forms
of omes to be chosen for describing casual or chance actions of
seeing which are brief in duration and which the speaker would

300
 Palauan Reference Grammar

be likely to view as quickly completed. In the following sentence

pairs, the rather free English equivalents are designed to make
the above-mentioned distinction clear; in each pair, the a-sen-
tence contains an imperfective form of omes, while the b-sen-
tence contains a perfective form:

(26) a. Ak ulȩmes ȩr a John er a elii.

‘I met and talked to John yesterday.’

b. Ak milsa a John er a elii.

‘I happened to see John yesterday.’

(27) a. Ng soak ȩl omes ȩr a Droteo.

‘I want to see (and talk to) Droteo.’

b. Ng soak ȩl mȩsa a Droteo.

‘I want to get a look at Droteo.’

(28) a. Ak mlo ȩr a kȩdȩra er a elii ȩl mo omes a bilas ȩl mle

kaidȩsachȩl.
‘Yesterday I went to the beach in order to watch the
boats racing.’

b. Ak mlo ȩr a kȩdȩra er a elii e miles a bȩtok ȩl bilas ȩl

mle kaidȩsachȩl.
‘Yesterday I went to the beach and happened to see a
lot of boats racing.’

The sentences below further illustrate the use of the imper-

fective vs. perfective forms of omes ‘see’.

(29) a. Ng dimlak a techȩllek ȩl omes ȩr a rȩsȩchȩlik.

‘I didn’t have an opportunity to get together with my
friends.’

b. Ak mle ȩl me omes ȩr a John.

‘I came in order to see John.’

c. Ng dimlak a techȩllek ȩl mes a charm.

‘I didn’t have a chance to get a look at any animals.’

d. Ak miles a bȩtok ȩl mȩtongakl ȩl blai ȩr a Hawaii.

‘I saw a lot of tall buildings in Hawaii.’

301
 12 Imperfective vs. Perfective Verbs

12.7. THE TRANSITIVE VERB ORRENGȨS

The imperfective vs. perfective forms of the verb orrengȩs
‘hear’ often exhibit a distinction of meaning similar to that
described for omes in 12.6 above. Before discussing this dis-
tinction, we shall list some of the perfective forms of orrengȩs:

(30) Person and Number of Present Tense Past Tense

Object Pronoun

1st pers sg rongȩsak ‘hear rirȩngȩsak

me’ ‘heard me’

3rd pers sg rongȩsii ‘hear rirȩngȩsii ‘heard

him/ her/it’ him/ her/it’

3rd pers pl hum rongȩstȩrir rirȩngȩstȩrir

‘hear them’ ‘heard them’

3rd pers pl non-hum rȩmengȩs ‘hear rirengȩs ‘heard

them’ them’

Use of imperfective orrengȩs (past: ulȩrrengȩs) implies

inten tional hearing by the subject over a relatively long du-
ration of time: therefore, the most suitable English equivalent is
‘listen to’. By contrast, use of the perfective forms of orrengȩs
connotes unexpected or casual hearing by the subject which
is completed in a comparatively short duration of time: the
best English equivalent is ‘(happen to) hear’. This distinction in
meaning is found in the pair of sentences below:

(31) a. Ak ulȩrrengȩs ȩr a Toki ȩl oukita.

‘I was listening to Toki play the guitar.’

b. Ak rirȩngȩsii a Toki ȩl oukita.

‘I heard/happened to hear Toki play the
guitar.’

The sentences below further illustrate the use of imper-

fective vs. perfective forms of orrengȩs ‘hear’:

(32) a. Kȩ ulȩrrengȩs a radio er a elii?

‘Did you listen to the radio yesterday?’

302
 Palauan Reference Grammar

b. Ak rongȩstȩrir a rȩchad ȩl mȩngȩrodȩch.

‘I (can) hear the people making noise.’

c. Ng soak ȩl rongȩsii a chisel a Toki er se ȩr a lȩbo ȩr a

Merikel.
‘I want to hear about what Toki did when she went to
America.’

d. Kȩ rirȩngȩsii a dȩrumk ȩr a kȩsus?

‘Did you (happen to) hear the thunder last night?’

303
 13 Directional Verbs

13DirectionalVerbs
13.1. THE THREE DIRECTIONAL VERBS OF
PALAUAN
In this chapter we will examine a special subgroup of Palauan
intransitive action verbs—namely, the directional verbs mo
‘go’, eko ‘go’, and me ‘come’. As we saw in 5.1.1, Palauan
intransitive action verbs denote actions which involve only a
doer, but no receiver. In other words, intransitive action verbs
do not take objects because they name actions in which only
the doer can participate rather than describing actions which
are directed at someone or something else. The intransitive
action verbs mo ‘go’, eko ‘go’, and me ‘come’ are called direc-
tional verbs because they all indicate movement in a particular
direction. This common element of meaning undoubtedly ac-
counts for some of the grammatical features which they share.
Before examining these grammatical features in detail, let us
first discuss how the three directional verbs differ from each
other in meaning.

13.2. MEANING AND USE OF THE DIRECTIONAL

VERBS
The directional verb me corresponds to English ‘come’ and
refers to movement towards the place where the speaker is,
was, or will be. This verb contrasts with both mo and eko, which
refer to movement away from the place where the speaker is,
was, or will be and therefore correspond most closely to English
‘go’. The difference in meaning between mo and eko depends
in an important way on the location of the person addressed
or spoken to (i.e., the hearer). The directional verb eko has the
very specific function of designating movement by the speaker
or some third party towards the supposed or real present,
past, or future location of the person addressed. By contrast,
the directional verb mo is much broader in scope, since it de-
scribes movement by anyone—speaker, hearer, or some third
party—away from the locations of the speaker and hearer.

304
 Palauan Reference Grammar

In order to illustrate the above-mentioned differences, let us

observe how the (italicized) directional verbs are used in the di-
alogs below:

(1) A: Ng sȩbȩchem ȩl me ȩr a blik ȩr a klukuk?

‘Can you come to my house tomorrow?’

B: Chochoi. Ng sȩbȩchek ȩl eko ȩr a blim ȩr a klukuk.

‘Yes, I can come to your house tomorrow.’

(2) A: Ng sȩbȩchem ȩl mo ȩr a blikȩr a klukuk?

‘Can you come/go to my house tomorrow?’

B: Chochoi. Ng sȩbȩchek ȩl mo ȩr a blim ȩr a klukuk.

‘Yes, I can come/go to your house tomorrow.’

Note, first of all, that the sentences in the dialogs 1 and 2

involve some future event, since they contain the temporal
phrase ȩr a klukuk ‘tomorrow’ (see 14.6). Though the English
equivalents for the sentences of 1 and 2 do not reflect it very
clearly, most Palauan speakers feel that the two dialogs de-
scribe very different situations. In dialog 1, A’s use of me ‘come’
implies his intention to be at home at the time of B’s (hoped-
for) arrival, and B’s use of eko ‘go’ indicates B’s belief or con-
viction that A will indeed be there to greet him. In dialog 2,
however, A’s use of mo ‘go’ clearly implies that he does not
intend to be home at the time of B’s arrival: it is possible, for
example, that A is requesting B to go to his (A’s) house on some
type of business or errand, even though A himself will not be
there. B’s use of mo in his response to A reaffirms the impli-
cation that A will not be or is not expected to be present. 1
When eko ‘go’ is used, the past, present, or future location
of the person addressed is usually his own house, as in 1B, but
it is also possible for some other location to be involved, as the
following sentences illustrate:

(3) a. Bo ȩr a bita e a Droteo a ekong.

‘Go next door, and Droteo will be along
(soon).’

b. Bo ȩr a blil a Toki e ak ekong.

‘Go to Toki’s house, and I’ll be along (soon).’

305
 13 Directional Verbs

c. Ak eko mȩ mchiiȩlak.
‘I’ll be there, so wait for me.’

Note, further, that sentences like the following

(4) a. Ng sȩbȩchek ȩl eko ȩr a party ȩr a klukuk.

‘I can come to your party tomorrow.’

b. A Droteo a dilu ȩl kmo ng sȩbȩchel ȩl eko ȩr a party ȩr a

klukuk.
‘Droteo said that he can come to your party tomorrow.’

imply nothing more than that the person addressed will be (or
is expected to be) at the party; though the party might indeed
be held at his house, other locations are certainly possible—in
other words, the person addressed may be hosting a party at
the community hall, etc.
The past tense forms of the directional verbs are provided
below:

(5) Present Tense Past Tense

me ‘come’ mle 2 ‘came’

eko ‘go’ ilȩko ‘went’
mo ‘go’ mlo ‘went’

The past tense forms of me ‘come’ and mo ‘go’ are derived

simply by infixing the past tense marker -l- (cf. 5.3.2) after
word-initial m, which is the verb marker. 3 The past tense form
of eko shows the past tense marker il; though this morpheme is
normally infixed after the initial consonant of a verb form, it is
prefixed in the case of eko, which begins with a vowel. Note,
further, that the initial vowel e of eko is reduced to a ȩ in the
past tense form ilȩko.
The basic distinctions of meaning among the three direc-
tional verbs are of course retained in the past tense forms as
well. Observe, therefore, the following pairs of contrasting sen-
tences:

306
 Palauan Reference Grammar

(6) a. A Droteo a mle er a Belau.

‘Droteo came to Palau.’

b. A Droteo a mlo er a Belau.

‘Droteo went to Palau.’

(7) a. Ak ilȩko ȩr a blim er a elii e kȩ mle 4 dibus.

‘I went to your house (to see you) yesterday and
(discovered) you were out.’

b. Ak mlo ȩr a blim er a elii e kȩ mle dibus.

‘I happened to go to your house yesterday and
(discovered) you were out.’

The meaning difference between 6a and 6b is rather straightfor-

ward and needs little comment: mle ‘came’ designates
movement towards the speaker’s location (Belau ‘Palau’), while
mlo ‘went’ implies movement away from the speaker’s location.
The distinction between 7a and 7b, however, is more subtle. In
7a, with ilȩko, the speaker expected the person addressed to
be home yesterday, but discovered he wasn’t, while in 7b, with
mlo, the speaker merely happened to drop by at the house of
the person addressed, not necessarily assuming that he would
be there.
The sentences below further illustrate the use of me, mo,
and eko. In each sentence, we have italicized the directional
phrase (see 14.3), which is a type of relational phrase that
designates the goal or termination point of the movement in-
volved. As you can see, a directional phrase consists of the rela-
tional word ȩr (in this case, more or less equivalent to English
‘to’) followed by a noun phrase naming the place (or event such
as party, etc.) which serves as goal or termination point.

(8) a. Kȩ me ȩr a Merikel er oingarang?

‘When are you coming to America?’

b. A Droteo a soal ȩl eko ȩr a ochȩraol.

‘Droteo wants to come to your money-raising
party.’

c. A sȩchȩlim ng mlo ȩr a skuul er a elȩchang?

307
 13 Directional Verbs

‘Did your friend go to school today?’

d. Kȩ mo ȩ r a che 5 ȩr a klukuk?
‘Are you going fishing tomorrow?’

13.3. DIRECTIONAL VERBS FOLLOWED BY

ACTION VERBS
The directional verbs me, mo, and eko frequently occur directly
followed by a transitive or intransitive action verb in expres-
sions like me mȩngȩtmokl ‘come and clean’, mo mȩngȩdub ‘go
and swim, go swimming’, and the like. Expressions of this kind
simply mean that at some present, past, or future time, someone
comes or goes to a particular place and performs the desig-
nated activity there. The distinctive meanings of the three direc-
tional verbs are of course preserved in such expressions, whose
structure we can summarize with the formula directional verb
+ action verb. Observe the following examples:

(9) a. A Toki a mle mȩngȩtmokl ȩr a blik ȩr a tutau.

‘Toki came and cleaned my house this morning.

b. Ng sȩbȩchem ȩl me mȩngȩdub ȩr a blik?

‘Can you come (and) swim at my house?’

c. A Toki ng ilȩko mȩruul a kall er a elii?

‘Did Toki come to your place and prepare food
yesterday?’

d. A rȩsȩchal a mlo mȩlasȩch a mlai er a elȩchang.

‘The men went making canoes today.’

e. Ak mo mȩngȩdub ȩr a diong.
‘I’m going swimming at the stream.

If a past action is involved, as in 9a, 9c, and 9d, then only the
directional verb, but not the directly following action verb, is
marked for the past tense. We will postpone further discussion
of this phenomenon until 15.1, where we will observe similar
restrictions against the occurrence of the past tense in various
types of Palauan dependent clauses.

308
 Palauan Reference Grammar

13.4. THE DIRECTIONAL VERBS AND FUTURE

TIME
The present tense forms of the directional verbs me, mo, and
eko can often be interpreted as referring to future time, es-
pecially when they are accompanied by temporal phrases (cf.
14.6) denoting future time such as ȩr a klukuk ‘tomorrow’, er
tia ȩl me ȩl rak ‘next year’, etc. Compare, for example, the fol-
lowing two sentences:

(10) a. A sȩchȩlik a me er a elȩchang.

‘My friend is coming now.’

b. A sȩchȩlik a me ȩr a klukuk.
‘My friend is coming/will come tomorrow.’

In 10a the present tense form me ‘come’ indeed refers to an

action which is happening at the present moment (i.e., at the
time when the speaker utters the sentence); this is made clear
by the accompanying temporal phrase er a elȩchang ‘now’. In
10b, however, the very same present tense form me ‘come’ is
used to refer to an action which is expected to happen in the
future, at the time point designated by the temporal phrase ȩr
a klukuk ‘tomorrow’. Like 10b, the sentences below involve a
future action or event, except that they contain the directional
verbs mo and eko:

(11) a. A Droteo a mo ȩr a Guam er tia ȩl me ȩl rak.

‘Droteo is going/will go to Guam next year.’

b. Ak eko ȩr a blim ȩr a kȩbȩsȩngei.

‘I am coming/will come to your house tonight.’

As the translations for 10b, 11a, and 11b show, the English
directional verbs come and go resemble Palauan directional
verbs in an important way since their present progressive
forms—e.g., am/is/ are coming, am/is/are going, etc.—are com-
monly used to convey a future meaning.
Since the present tense forms of the directional verbs me,
mo, and eko can be used to designate future time, as we have
just seen above, it is not surprising to observe a similar function
for sequences consisting of the present tense form of a direc-

309
 13 Directional Verbs

tional verb followed by an action verb (cf. 13.3 above). Thus, in

the examples below, me / mo / eko + action verb refers to an
action in the future:

(12) a. A Toki a me milil ȩr a klukuk.

‘Toki will come play tomorrow.’

b. Ak eko mȩsuub ȩr a klukuk.

‘I’ll come study at your place tomorrow.’

c. A Toki ng eko mȩngȩtmokl ȩr a blim ȩr a tela ȩl klok?

‘At what time will Toki come clean your house?’

d. Kȩ mo mȩsuub ȩr a Merikel er tia ȩl me ȩl rak?

‘Are you going to study in America next year?’

e. Ak mo omes ȩr a John ȩr a kȩbȩsȩngei.

‘I’m going to see John this evening.’

f. Kȩ, mo chuiȩuii a hong ȩr a klukuk?

‘Are you going to (completely) read the book tomorrow?’

In the sentences above, the three directional verbs are used as

auxiliary (or “helping”) words because they provide important
information about the tense of the immediately following action
verb. Other auxiliary words which we have seen so far include
mle ‘was, were’, which is used to indicate the past tense of state
verbs (cf. 5.1.3), and mla, which indicates a recent past event
or past experience (cf. 5.3.2.1).
How can we explain the fact that the present tense forms
of Palauan directional verbs can function to express future ac-
tions or events? It appears as if the basic meaning of these
verbs, which is to describe movement across physical space
from one location to another, has been extended to cover
“movement across time” from one “location in time” to an-
other—namely, from present to future. This kind of meaning
change, in which directional verbs shift their reference from
space to time, is found quite commonly in the languages of the
world. Thus, in English, the directional verb go indicates future
time in sentences like I’m going to study tomorrow.

310
 Palauan Reference Grammar

13.5. THE DIRECTIONAL VERB MO AND CHANGE

OF STATE
In 13.4 above we saw that the present tense form of the direc-
tional verb mo ‘go’ can be used with a following action verb
to describe a future action or event. In this section, we will ex-
amine how mo (in any tense) can be used together with a state
verb (cf. 5.1.2–3 and chap.7) to designate a change of state.
Palauan change of state expressions have the structure mo +
state verb and usually correspond to English expressions such
as ‘become…’, ‘get…’, etc. The change of state can occur at dif-
ferent points in time, as the following sentences illustrate:

(13) a. A lsȩkum ak omȩngur ȩr a uum, e ak mo smechȩr.

‘Whenever I eat at the cafeteria, I get sick.’

b. A eangȩd a soal ȩl mo mȩkngit.

‘It’s likely the weather will get/take a turn for the
worse.’

c. A ngȩlȩkem ng mo ungil ȩl smechȩr er oingarang?

‘When will your child get better (from his sickness)?’

d. Aki mlo kaodȩnge ȩr a Guam.

‘We got acquainted with each other in Guam.’

e. A bȩchik a mlo smechȩr er a elii.

‘My wife got sick yesterday.’

f. A odoim a mla mo bȩkȩbau.

‘The food has become smelly (spoiled).’

g. A ududek a mla mo diak.

‘My money has run out.’

In 13a, the change of state is happening repeatedly, while in

13b–c it will take place in the future; thus, the present tense
of mo followed by a state verb can designate either a recurrent
change of state or a future change of state. In 13d–g, the change
of state occurred at some time point in the past, with mlo +
state verb in 13d–e and mla mo + state verb in 13f–g dis-
tinguishing relatively remote vs. relatively recent past time
points, respectively. 6

311
 13 Directional Verbs

All of the sentences of 13 draw our attention to a resulting

state—that is, to a state which has come about as the result of
some process of change. This resulting state is explicitly (or
overtly) expressed by the particular state verb which follows
mo. By contrast, the initial state—i.e., the state which existed
when the process of change began—is not overtly expressed,
but merely implied in the sentences of 13. For example, in 13f,
the state verb bȩkȩbau ‘smelly’ describes the present state of
the food (resulting, say, from the food’s having been left in a
warm place) and is contrasted with an implied initial (or earlier)
state in which the opposite was the case (namely, the food was
fresh and edible). In the same way, mla mo diak ‘(lit) has become
nonexistent’ of 13g describes the present unfortunate state of
the speaker’s finances; this state represents the end result of a
process of change which began with the opposite state—namely,
one in which the speaker had adequate financial resources.
As the above discussion shows, Palauan change of state ex-
pressions of the form mo + state verb necessarily involve two
time points—one is the time point of the implied initial (or
earlier) state and the other is the time point of the overtly-ex-
pressed resulting (or later) state. Here, too, we can see how the
original function of the directional verb mo, which is to describe
movement across physical space from one place to another,
has been expanded to include “movement across time” from one
“location in time” (the initial or earlier state) to another (the
resulting or later state) (cf. our discussion at the end of 13.4
above).
In order to describe a change of state which is in the process
of happening at the present moment (i.e., at the time the
speaker utters the sentence) we use the noun omȩrael ‘process’
in one of its possessed forms (e.g., omerolek, omȩrolel, etc.),
followed by ȩl and a change of state expression. The word
omȩrael, which also means ‘trip’, is an action noun derived by
prefixing o- to the intransitive action verb mȩrael ‘walk, travel’
(cf. 8.6). This verb is in turn derived by prefixing the verb
marker mȩ- to the verb stem rael, an independently-occurring
noun meaning ‘road’ (cf. 6.1). In the sentences below, we illus-
trate the use of omȩrael to describe a change of state which is
occurring at the present moment:

(14) a. Ak omȩrolek ȩl mo ungil ȩl smechȩr ȩr a

tȩretȩr.

312
 Palauan Reference Grammar

‘I’m in the process of getting better from

my cold.’

b. A bȩluu ȩr a Hawaii a omȩrolel ȩl mo

kikiongȩl.
‘The land in Hawaii is getting dirty/
polluted.’

c. A skuul a omȩrolel ȩl mo mȩsekȩd.

‘The school is getting (more and more)
crowded.’

d. A tangk a omȩrolel ȩl mo mui.

‘The tank is getting filled.’

e. A sils a omȩrolel ȩl mo mȩringȩl.

‘The sun is getting hotter (and hotter).’

In the examples of 14, the possessed forms of omȩrael ‘process’

must agree with the sentence subject. Thus, in 14a, the 1st pers.
sg. subject ak requires omȩrolek, while in 14b–e the various 3rd
pers. sg. subjects require omȩrolel. For a detailed discussion
of ȩl, which links the possessed forms of omȩrael with the fol-
lowing change of state expression, see chap. 15.
The abovementioned use of omȩrael ‘process’ followed by
a change of state expression does not represent the only way
of describing a current change of state. As we saw in 11.12.4,
the inchoative forms of state verbs—e.g. mȩkȩlȩkȩlta ‘getting
cold’ (cf. mȩkȩlȩkolt ‘cold’), mȩkngte ‘getting bad, worsening’
(cf. mȩkngit ‘bad’), etc.—are also used in this way in sentences
like the following:

(15 a. Tia ȩl delmȩrab a kmal mȩkȩlȩkȩltang.

‘This room is getting very cold.’

b. A eangȩd a mȩkngte mȩ lak dorael.

‘The weather’s getting worse, so let’s not
go.’

313
 13 Directional Verbs

13.5.1. CHANGE OF STATE EXPRESSIONS WITH NOUN PHRASES

It is also possible to have change of state expressions in which
the directional verb mo ‘go’ is followed by a noun phrase that
identifies a profession or nationality, or classifies individuals
or things into a particular category. Some typical change of
state expressions of this type are observed in the sentences
below:

(16) a. A Masaharu a soal ȩl mo {sensei/toktang}.

‘Masaharu wants to become {teacher/
doctor}.’

b. A Toki a mla mo chad ȩr a Merikel.

‘Toki has become an American citizen.’

c. A Oreor a mla mo matsi er a Belau.

‘Koror has become the capital of Palau.’

13.6. SPECIAL VERBAL EXPRESSIONS WITH MO

There are a few Palauan words which must always occur pre-
ceded by a directional verb. The resulting expressions, of which
mo mȩrek ‘finish’ and mo/me rȩme ‘go/come home’ are perhaps
the most common, function as verbs. Observe, for example,
the following sentences, which contain instances of mo mȩrek
‘finish’:

(17) a. Kȩ mo mȩrek ȩr a subȩlem er oingarang?

‘When are you going to finish your homework?’

b. Ak mlo mȩrek ȩr a urerek er a elii.

‘I finished my work yesterday.’

In the sentences above, mo mȩrek is used as an imperfective

transi tive verb: in other words, it is followed by a specific
object which is marked with the specifying word ȩr (cf. 2.7).
The word mȩrek, which is in some way related to the imper-
fective transitive verb mȩrkui ‘finish’ 7 , never changes its form.
Therefore, in order to indicate the past tense with mo mȩrek,
it is the directional verb mo which must take the past tense
marker -l-, as in 17b.

314
 Palauan Reference Grammar

The use of mo/me rȩme ‘go/come home’ is illustrated in the

sentences below:

(18) a. Ak mo rȩme ȩr a beluak.

‘I’m going to return to my country.’

b. Ak ulȩmȩngur e me rȩmei.
‘I ate dinner and came back home.’

c. A Droteo a mlo rȩmei.

‘Droteo went home.’

d. Bo mrei!
‘Go home!’

e. Be mrei!
‘Come home!’

The expressions mo/me rȩme are used in the examples of 18

as intransitive action verbs. The word rȩme, even though
it cannot occur independently as a verb, nevertheless has one
of the important characteristics of Palauan intransitive action
verbs, since it appears to contain the metathesized verb
marker -(ȩ)m - (cf. 6.2). As 18c shows, the form of rȩme does
not change in the past tense, but instead the past tense marker
-l- is infixed into the directional verb mo. In 18d–e, we observe
the imperative (or command) forms of mo/ me rȩme; the
structure of these forms will be explained in 19.5.

315
 14 Relational Phrases

14RelationalPhrases
14.1. DISTRIBUTION AND FUNCTION OF
RELATIONAL PHRASES
The speaker of Palauan, like the speaker of any other language,
always has a choice as to how much information he will convey
in the sentences which he produces. For example, in answer
to a question like ‘What were you doing yesterday?’, a Palauan
speaker might simply say

(1) Ak milsuub.
‘I was studying.’

But if the speaker wanted to provide more information about

what he did than is expressed in 1, he might give an answer like

(2) Ak milsuub ȩr a skuul.

‘I was studying at school.

In 2, the addition of the sequence ȩr a skuul ‘at school’ gives

the hearer a more detailed picture of the activity involved by
indicating the place (or location) where the studying was done.
In a similar way, 3b below supplies more information than 3a be-
cause it contains er a elii ‘yesterday’, which specifies the time
when the subject (Droteo) read the book:

(3) a. A Droteo a chiliuii a hong.

‘Droteo (completely) read the book.’

b. A Droteo a chiliuii a hong er a elii.

‘Droteo (completely) read the book
yesterday.’

Although the expressions ȩr a skuul ‘at school’ and er a elii ‘yes-

terday’ of 2 and 3b provide the hearer with different pieces of
information (place where vs. time when, respectively), they
nevertheless have the common characteristic of putting some
action (such as studying, reading a book, etc.) into perspective
by relating it to a place (e.g. skuul ‘school’) or time (e.g. elii

316
 Palauan Reference Grammar

‘yesterday’) of occurrence. For this reason, expressions like ȩr a

skuul and er a elii are called relational phrases. All relational
phrases consist of the relational word ȩ r followed by the ap-
propriate words—a single noun or a noun phrase—to indicate
such pieces of information as place where, time when, and the
like.
Because relational phrases occur so commonly in Palauan,
we have already made mention of them at various points in the
text, and even analyzed them to some extent. In 5.2, for ex-
ample, we saw that two of the main sentence types of Palauan
can be summarized according to the following formulas:

(4) a. subject noun phrase + transitive verb phrase (+ object

noun phrase) ( + relational phrase)

b. subject noun phrase + intransitive verb phrase (+

relational phrase)

In the formulas above, which show the relative order of the

major parts of Palauan sentences, we have placed the term
“relational phrase” in parentheses because the occurrence of
this element is optional. In other words, Palauan sentences do
not necessarily have to contain a relational phrase in order to
be complete; rather, as we mentioned at the beginning of this
section, relational phrases can be added freely to sentences de-
pending on how much extra information the speaker wishes to
supply. Since more than one relational phrase can occur in a
given sentence, as we will see in 14.7 below, the formulas we
presented in 4 will require some modification.
In discussing examples 2 and 3b above, we said that the
relational phrases found in those sentences serve to put some
action into perspective by relating it to a particular place of
occurrence, time of occurrence, or the like. As it turns out,
our analysis was purposefully oversimplified, since relational
phrases can also put states into perspective by providing
certain types of information. Thus, in the sentences below, we
observe relational phrases occurring together with different
types of state verbs:

(5) a. A sȩchȩlik a mle dibus er a elii.

‘My friend was out/away from home
yesterday.’

317
 14 Relational Phrases

b. A Droteo a ngar ȩr a sers.

‘Droteo is in the garden.’

c. A sensei a smechȩr ȩr a tȩretȩr.

‘The teacher is sick with a cold.’

In the sentences above, the italicized relational phrases provide

us with different types of information about the several states
involved. In 5a er a elii ‘yesterday’ tells us when the past state
mle dibus ‘was out/away from home’ occurred, in 5b ȩr a sers ‘in
the garden’ describes where the subject is located (for further
discussion of the existential verb ngar ‘exist, be (located)’ see
18.2), and in 5c ȩr a tȩretȩr ‘with/because of a cold’ gives the
cause of the state designated by smechȩr ‘sick’.
The italicized expressions in the examples of 5 represent
only three out of many types of relational phrases found in
Palauan. In the sections below, we will classify Palauan rela-
tional phrases according to the many different kinds of infor-
mation they convey. As our discussion proceeds, we will see that
the single Palauan relational word ȩr corresponds to a large va-
riety of English relational words, including ‘in’, ‘at’, ‘on’, ‘to’,
‘from’, ‘of’, ‘out of’, ‘because of’, ‘for’, etc.; and sometimes—as
in the case of er a elii ‘yesterday’—the English equivalent con-
tains no relational word at all.

14.2. LOCATIONAL PHRASES

Any relational phrase which indicates the location of some
action or state is a locational phrase. In each of the sentences
below, the italicized locational phrase narrows down the scope
of an action or activity by indicating its place of occurrence:

(6) a. A Toki a oureor ȩr a bangk.

‘Toki is working at the bank.’

b. A rȩngalȩk a mililil ȩr a kurangd.

‘The children were playing at the playground.’

c. A Toki a mȩsilȩk a bilel ȩr a daob.

‘Toki is washing her clothes in the ocean.’

318
 Palauan Reference Grammar

d. Ng sȩbȩchem ȩl okȩrdak ȩr tiang?

‘Can you let me off here?’

e. A John a mlad ȩr a Merikel.

‘John died in America.’

In the relational phrases of 6, the noun following the relational

word ȩr is introduced by the word a, unless it is a demon-
strative word like tia ‘this place, here’ of 6d (cf. 2.6 and see
24.3) or a pronoun.
The following sentences contain locational phrases (itali-
cized) which supply us with further information about particular
states:

(7) a. A John a mle smechȩr ȩr a Hawaii.

‘John was sick in Hawaii.’

b. A ngalȩk a mle mȩchiuaiu ȩr a delmȩrab.

‘The child was sleeping in the room.’

c. A Droteo a mle dȩngchokl ȩr a kingall.

‘Droteo was sitting on the chair.’

d. Ak mla ȩr a blil a Toki er a elii.

‘I was at Toki’s house yesterday.’

e. A taod a kirel ȩl ngar ȩr a katur, mȩ a oliich a kirel ȩl

ngar ȩr a kadikm.
‘The fork must be on the left, and the spoon must be on
the right.’

f. A sensei a ngar ȩr a uum.

‘The teacher is in the kitchen.’

Sentences 7d-f contain the existential verb ngar ‘exist, be (lo-

cated)’ (past tense: mla ‘existed, was (located)’), which is almost
always followed by a locational phrase. 1 In more complex sen-
tences, the sequence ngar + locational phrase, when preceded
by ȩ l, specifies the means of transportation used to move
from one place to another. This is observed in sentences like the
following, which we will examine more carefully in 15.5:

319
 14 Relational Phrases

(8) Ak mlo ȩr a stoa ȩl ngar ȩr a

sidosia.
‘I went to the store by car.’

In order to ask a question about the location of some action

or state, we use the locational phrase ȩr ker, which consists
of the relational word ȩr followed by the question word ker
‘where?’. 2 Thus, examples 6a, 6c, 7d, and 7f above would be
possible answers to the following questions containing ȩr ker
‘where?’:

(6a’) A Toki ng oureor ȩr ker?

‘Where does Toki work?’

(6c’) A Toki ng mȩsilȩk a bilel ȩr ker?

‘Where is Toki washing her
clothes?’

(7d’) Kȩ mla ȩr ker er a elii?

‘Where were you yesterday?’

(7f’) A sensei ng ngar ȩr ker?

‘Where is the teacher?’

The structure of the above question sentences will be explained

in detail in 20.5.

14.2.1. LOCATIONAL PHRASES WITH NOUNS DESCRIBING SPATIAL

RELATIONSHIPS.
Palauan has a fairly large class of nouns which are used to de-
scribe spatial relationships such as above/on top of vs. below/
under, in front of vs. in back of, and the like. Two nouns of this
type are used in the sentences below:

(9) a. A Droteo a ngar ȩr a {bab/eou}.

‘Droteo is {up there/down
there}.’

b. A Droteo a kie ȩr a {bab/eou}.

‘Droteo lives {upstairs/
downstairs}.’

320
 Palauan Reference Grammar

In the italicized locational phrases of 9a-b we find the nouns

bab ‘area/space above’ and eou ‘area/space below’. These nouns
do not name a specific place (such as skuul ‘school’ or blik ‘my
house’) but instead designate an area of space which is “de-
fined” in relation to something else. Thus, in 9a-b, bab and eou
refer to areas of space above and below the spot where the
speaker of the sentence is located. 3
Now, with 9a-b above, let us compare sentences like the fol-
lowing:

(10) a. A katuu a mȩchiuaiu ȩr a bebul a tebȩl.

‘The cat is sleeping on (top of) the table.’

b. A bilis a mȩchiuaiu ȩr a eungel a tebȩl.

‘The dog is sleeping under(neath) the
table.’

The italicized parts of 10a-b are also locational phrases, but

ones in which the nouns bab ‘area/space above’ and eou ‘area/
space below’ occur in their possessed forms (cf. chap. 3).
In other words, the possessed nouns bebul and eungel have
a 3rd pers. sg. possessor suffix (-ul or -el) which agrees with
the following “possessor” tebȩl ‘table’. Therefore, the expres-
sions bebul a tebȩl and eungel a tebȩl mean, respectively, ‘the
top of the table’ 4 and ‘the space underneath the table’. It is
important to recognize that although expressions like eungel
a tebȩl ‘under(neath) the table’ and chimal a Droteo ‘Droteo’s
hand’ are identical in structure (i.e., they both consist of a pos-
sessed noun followed by another noun), their meanings are ac-
tually quite different. As we saw in 3.6–7, an expression like
chimal a Droteo is a noun phrase of possession in which
Droteo is the actual possessor of something (chimal ‘his hand’)
which is an inseparable part of him. In the case of eungel a
tebȩl, however, eungel ‘the space underneath it’ is obviously not
a part of the table but merely an area of space defined in re-
lation to the table.
In the following sentences, we observe locational phrases in
which the relational word ȩr is followed by bebuk ‘space above
me’ and eungek ‘space below me’, which have the 1st pers. sg.
possessor suffixes -uk and -ek, or by bebmam ‘space above us’
and eungam ‘space below us’, which have the 1st pers. pl. excl.
possessor suffixes -mam and -am:

321
 14 Relational Phrases

(11) a. A klok a ngar ȩr a bebuk.

‘The clock is up there above me.’

b. A katuu a mȩchiuaiu ȩr a eungek.

‘The cat is sleeping underneath me (i.e., under
my chair).’

c. A Satsko a kie ȩr a {bebmam/eungam}.

‘Satsko lives {upstairs/downstairs} from us.’

In order to express the relationships in front of vs. in back

of, Palauan makes use of several different words, including mad
‘front’ (which is probably an extended use of mad ‘eye, face’)
and rȩbai ‘area/space in back of (a building)’ (which seems to be
related to the nouns blai ‘house, building’ and bai ‘community
house’). These words are used in sentences like the following,
in which the locational phrases have been italicized:

(12) a. Ng ngar ȩr ngii a kȩrrȩkar ȩr a mȩdal a

blik.
‘There are trees (right) in front of my
house.’

b. A mlim a ngar ȩr a rȩbai ȩr a blik.

‘Your car is in back of my house.’

Like bab (bebuk, bebul, etc.) and eou (eungek, eungel, etc.)
above, mad ‘front’ can take possessor suffixes: thus, in 12a we
have the possessed noun mȩdal ‘its front’ 5 followed by the “pos-
sessor” blik ‘my house’. By contrast, the noun rȩbai ‘area/space
in back of (a building)’ cannot take any possessor suffixes—i.e.,
it is un possessible (cf. 3.8). In order to express a “possessor”
with rȩbai, it is therefore necessary to use a relational phrase
introduced by ȩr such as ȩr a blik of 12b.
With 12a above, contrast the following sentence:

(13) Ng ngar ȩr ngii a kȩrrȩkar ȩr a ngȩlo ȩr a blik.

‘There are trees (in the general area) in front of my
house.’

It is difficult to find suitable English translations to express the

difference in meaning between 12a and 13. While use of mȩdal
in 12a implies that the trees are very close to the house (pos-

322
 Palauan Reference Grammar

sibly providing it with shade), ngȩlo of 13 implies that the trees

are relatively far away (perhaps across the road). Note further
that ngȩlo ‘area/space in front of (a building) (relatively far
away)’, like rȩbai ‘area/space in back of (a building)’, is an un-
possessible noun.
In the following pairs of sentences, we notice a contrast in
meaning similar to that observed between 12a and 13 above:

(14) a. A Toki a mirrael ȩr a mȩdal a Droteo.

‘Toki was walking right in front of Droteo.’

b. A Toki a mirrael ȩr a uche ȩr a Droteo.

‘Toki was walking somewhere in front of
Droteo.’

(15) a. A Toki a dilȩngchokl ȩr a ullel 6 a Droteo.

‘Toki sat right in back of Droteo.’

b. A Toki a dilȩngchokl ȩr a uriul ȩr a Droteo.

‘Toki sat somewhere behind Droteo.’

In the a-sentences above, mȩdal and ullel designate locations di-

rectly in front of or in back of Droteo, while in the b-sentences
the (unpossessible) nouns uche and uriul are less precise and
refer, in a more general way, to locations anywhere ahead of
or behind Droteo. Thus, it is possible for 14b, for example, to
refer to a situation in which Toki was walking along a trail ahead
of Droteo, but far enough away to be out of his sight. Another
common use of uche and uriul is observed in the following ex-
ample:

(16) A Droteo a mȩlȩchoib ȩr a {uche/uriul} ȩr a

omoachȩl.
‘Droteo is bathing {upstream/downstream}.’ 7

In addition to describing the spatial relationships in front of vs.

in back of, the nouns uche and uriul have also come to be used
in a temporal sense to express the ideas of before vs. after,
respectively. This extension of meaning is observed in the fol-
lowing simple sentences:

(17) a. Ak mo ȩr a uchei.
‘I’ll go first (i.e., before anyone else).’

323
 14 Relational Phrases

b. Ak mo omȩngur ȩr a uriul.
‘I’ll eat later (i.e., after doing some other
things).’

Since the italicized relational phrases of 17a–b refer to points in

time, they are technically temporal phrases (see 14.6 below).
Because uche and uriul can be used in a temporal as well
as a spatial sense, the sentences of 14b and 15b above are ac-
tually ambiguous—that is, they can be interpreted in two dif-
ferent ways. Thus, 14b can also mean ‘Toki left before Droteo
did’, and 15b has the additional interpretation ‘Toki sat down
after Droteo did.’ For further discussion of the temporal uses of
uche and uriul, see 22.2.1. 8
In the sentences below, the italicized locational phrases con-
tain the Palauan nouns chȩlsel ‘inside’ and ikrel ‘outside’:

(18) a. Ng ngar ȩr ngii a oluchȩs ȩr a chȩlsel a

skidas.
‘There is a pencil inside the drawer.’

b. Ng ngar ȩr ngii a rȩkung ȩ r a chȩlsel a

blsibs.
‘There is a crab inside the hole.’

c. A bilas a ngar ȩ r a chȩlsel a

{taoch./omoachȩl}. 9
‘The boat is in the {channel/river}.’

d. Ng ngar ȩr ngii a restorangd ȩ r a chȩlsel a

ii.
‘There is a restaurant inside the cave.’

e. A ngikȩl a rȩmurt ȩ r a chȩlsel a bad.

‘The fish are swimming in the coral.’

f. Ng ngar ȩr ngii a chȩdib ȩr a chȩlsel a kall.

‘There are black ants in the food.’

g. A Toki a dȩngchokl ȩr a ikrel a blil.

‘Toki is sitting outside her house.’

324
 Palauan Reference Grammar

While ikrel of 18g is a possessed form of the independently-

occurring noun iikr ‘area/space outside’, 10 chȩlsel ‘inside’ of
18a-f is an obligatorily possessed noun (cf. 3.5).

14.2.2. ADDITIONAL EXAMPLES OF NOUNS DESCRIBING SPATIAL

RELATIONSHIPS
The italicized locational phrases in the sentences below illus-
trate a few more Palauan nouns which describe spatial relation-
ships:

(19) a. A blik a ngar ȩr a bita ȩ r a skuul.

‘My house is {next door to/across street from} 11 the
school.

b. A Toki a mle dȩngchokl ȩr a bita ȩr a dȩmal.

‘Toki was sitting next to her father.’

c. A blik a ngar ȩr a dȩlongȩlel 12 a blil a Toki mȩ a Droteo.

‘My house is between Toki’s house and Droteo’s house.’

13
d. A Babȩldaob a ngar ȩr a diluchȩs ȩ r a Oreor.
‘Babeldaob is north of Koror.’

e. Ng ngar ȩr ngii a kȩrrȩkar ȩr a bȩlngel 14 a sȩrsek.

‘There is a tree in the middle of my garden.’

f. A Toki a mle dȩngchokl ȩr a bȩlngel a blai.

‘Toki was sitting in the middle of the house.’

g. Ng ngar ȩr ngii a bilas ȩr a bȩlngel a omoachȩl.

‘There is a boat in the middle of the river.’

h. Ak milȩngȩdub ȩr a tkul 15 a daob.

‘I was swimming at the ocean’s edge.’

i. Ng ngar ȩr ngii a bȩtok ȩl komi ȩr a tkul a rael.

‘There’s a lot of trash at the edge/side of the road.’

j. A skuul a dȩchor ȩr a mobȩdul 16 a kȩdȩrang.

‘The school is situated in the general direction of the
beach.’

325
 14 Relational Phrases

In the examples of 10, 12–16, and 18–19 above, we have seen

how expressions like bebul a tebȩl ‘top of the table’, rȩbai ȩ r
a blik ‘area/space in back of my house’, and the like are used
in locational phrases to give specific information about various
kinds of spatial relationships. Since they are noun phrases, it
is not surprising that the expressions under discussion can also
be used as sentence subjects or objects. A few examples are
given below:

(20) a. A bebul a tebȩl a mla mo dȩkimȩs.

‘The top of the table has gotten wet.’

b. A chȩlsel a skidas a kirel ȩl mo ngȩtachȩl.

‘The inside of the drawer must be cleaned.’

c. A dȩlongȩlel a blil a Toki mȩ a Droteo a kikiongȩl.

‘The area between Toki’s house and Droteo’s house is
dirty.’

d. A bilas a ulȩkiu a eungel a did.

‘The boat passed under the bridge.’

e. A bilas a ulȩkiu a dȩlongȩlel a chelȩbachȩb.

‘The boat passed between the Rock Islands.’

In 20a-c, the italicized noun phrases are sentence subjects,

while those in 20d–e are sentence objects.

14.3. DIRECTIONAL PHRASES

Any relational phrase which indicates the goal or termination
point of some action involving movement is called a direc-
tional phrase. Directional phrases commonly follow the direc-
tional verbs me ‘come’, mo ‘go’, and eko ‘go’ (cf. chap. 13)
and simply identify the place to which a person comes or goes.
Directional phrases can also occur with other verbs or expres-
sions involving movement such as tmuu ‘enter’, soisȩb ‘enter’,
mȩrael ȩl mo… ‘walk to’, rȩmurt ȩl mo… ‘run to’, and the
like. 17 Because both locational phrases and directional phrases
identify places, all of the locational phrases discussed in 14.2
and 14.2.1–2 above have the potential of being used as direc-
tional phrases, given the proper context.

326
 Palauan Reference Grammar

In the sentences below, the italicized relational phrases

function as directional phrases because they tell us the goal or
termination point of movement:

(21) a. A Toki a mle ȩr a blik er a elii.

‘Toki came to my house yesterday.’

b. A John a mo ȩr a Guam ȩr a klukuk.

‘John is going to Guam tomorrow.’

c. A beab a {tiluu/silisȩb} 18 ȩr a blsibs.

‘The mouse went into the hole.’

d. A Toki a {rirurt/mirrael} ȩl mo ȩr a blil a

Droteo.
‘Toki {ran/walked}to Droteo’s house.

e. Ak mo ȩr a {bebul/eungel} a rois.
‘I’m going to the {top/foot} of the
mountain.’

f. Bo mtuu ȩ r a chȩlsel a blai.

‘Go inside the house.’

g. A rȩkung a tiluu ȩr a dȩlongȩlel a bad.

‘The crab went between the stones.’

h. A Droteo a rirȩdȩkekl ȩr a daob.

‘Droteo jumped into the water.’

i. A Toki a ngilelt ȩr a mȩsei.

‘Toki sank down into the taro patch.’

In 21e-g, the directional phrases contain some of the nouns de-

scribing spatial relationships which we examined in 14.2.1–2
above.
In the examples of 21, the verbs or expressions of movement
which precede the italicized directional phrases are intran-
sitive because they do not involve any objects. It is also pos-
sible to have sentences in which directional phrases occur with
transitive verbs or expressions of movement—i.e., ones which
take objects. Usually, verbs or expressions of this kind involve
putting something somewhere, or transferring something from

327
 14 Relational Phrases

one place to another. Some common examples include mȩlȩcha

‘put’, omȩche ‘leave’, 19 the causative verbs oltuu ‘put into’ and
olsisȩb ‘put into’ (cf. 9.2.2), and expressions like olab a ilumȩl ȩl
mo… ‘bring drinks to’, nguu a bȩchik ȩl mo… ‘take my wife to’,
etc. 20
In the examples below, the directional phrases are italicized,
and the objects of the transitive verbs are in bold type:

(22) a. A Toki a lilia a komibako ȩ r a eungel a

tebȩl.
‘Toki put the wastebasket under the table.’

b. Lak molȩcha a omotsia ȩr a bebul a tebȩl.

‘Don’t put the toys on the table.’

c. A rȩngalȩk a milȩche a bȩtok ȩl komi ȩr a

sers.
‘The children left lots of trash in the
garden.’

d. A John a milȩchire a hong ȩ r a chȩlsel a

skidas.
‘John left the book in the drawer.’

e. Aki {ultungii/usisȩbii} 21 a John ȩr a

chȩlsel a mlai.
‘We pushed John into the car.’

f. A ngalȩk a ulȩngelt a chimal ȩr a

merikengko.
‘The child sank his hands into the flour.’

In the following examples, directional phrases follow more

complex expressions whose structure we will not explain until
15.7.1; again, the directional phrase has been italicized and the
object has been printed in bold type:

(23) a. Ak olab a ilumȩl ȩl mo ȩr a party.

‘I’m bringing drinks to the party.’

b. A John a nguu a bȩchil ȩl mo ȩr a ochȩraol.

‘John is taking his wife to the money-raising
party.’

328
 Palauan Reference Grammar

In order to ask a question about the goal or termination

point associated with some verb or expression involving
movement, we use the directional phrase ȩr ker ‘where? (=to
what place?)’, which is identical in structure to the locational
phrase ȩr ker ‘where? (= in/at what place?)’ mentioned at the
end of 14.2 above. Thus, examples 21b, 21d, 22a, and 23b
would be possible answers to the following questions containing
ȩr ker ‘where? (= to what place?)’:

(21b’) A John ng mo ȩr ker ȩr a klukuk?

‘Where is John going tomorrow?’

(21d’) A Toki ng {rirurt/mirrael} ȩl mo ȩr ker?

‘Where did Toki {run/walk} to?’

(22a’) A Toki ng lilia a komibako ȩr ker?

‘Where did Toki put the waste basket?’

(23b’) A John ng nguu a bȩchil ȩl mo ȩr ker?

‘Where is John taking his wife (to)?’

14.3.1. FURTHER TYPES OF DIRECTIONAL PHRASES

All of the examples given in 14.3 above contain directional
phrases which follow (transitive or intransitive) verbs or ex-
pressions which designate movement in a particular direction.
There appears to be a special subtype of directional phrase that
occurs after verbs which, rather than designating movement,
name certain kinds of actions or states which are characteris-
tically directed at some particular goal. Often, the goal itself
seems to be the reason for the action or state. Directional
phrases of this kind, which point out the goal (or, possibly,
reason), are italicized in the sentences below:

(24) a. A ngalȩk a lmangȩl ȩr a {dȩmal/kiande}.

‘The child is crying for {his father/candy}.’

b. Ak mȩdȩues ȩ r a ngikȩl.
‘I have an appetite/taste for fish.’

c. A John a ulȩchȩrchur ȩr a oltobȩdȩchur.

‘John laughed at the joke.’

329
 14 Relational Phrases

In 24a the child is crying in order to get something he

wants—in other words, the goal of his crying is to obtain the
presence of his father, or candy. In 24b the subject’s appetite
is directed towards a particular desired object or goal—namely,
fish—and in 24c the subject’s laughter was directed at (and
caused by) the joke. Several other examples containing the
special subtype of directional phrase under discussion are given
below. Notice that the English equivalents usually include ‘for’,
‘at’, or ‘about’:

(25) a. A rȩchad a mle mȩkngit a rȩngrir ȩ r a katsudo.

‘The people were sad about the movie.’

b. A rȩchad a mle kakoad ȩr a kall.

‘The people were fighting with each other over/
for food.’

c. Aki mle chachȩdȩchȩduch ȩr a bȩtok ȩ l mondai.

‘We talked about lots of problems.’

14.4. SOURCE PHRASES

Any relational phrase which indicates the point of origin of
some action involving movement is a source phrase. Source
phrases, like locational and directional phrases, identify places;
therefore, most of the locational and directional phrases treated
in the sections above can also function as source phrases, pro-
vided there is a suitable context. Such contexts are usually pro-
vided by intransitive action verbs like tuobȩd ‘come out’, suebȩk
‘fly (out)’, ruebȩt ‘fall (from)’, and the like, which focus our at-
tention on the point of origin or source (rather than termination
point) of the action. Observe the sentences below, in which
the italicized source phrases follow various intransitive action
verbs:

(26) a. A rȩkung a tilobȩd ȩ r a blsibs.

‘The crab came out of the hole.’

b. A bȩlochȩl a mla suebȩk ȩr a kȩrrȩkar.

‘The pigeon has flown out of the tree.’

c. A Toki a rirebȩt ȩr a chȩldukl.

330
 Palauan Reference Grammar

‘Toki fell off the dock.’

d. A ngalȩk a miltengȩl ȩr a kȩrrȩkar.

‘The child climbed down from the tree.’

e. A katuu a rirȩdȩkekl ȩr a chȩdȩuel a blik ȩl mo ȩr a

kȩrrȩkar.
‘The cat jumped from the roof of my house into the
tree.’

f. A John a rirurt ȩr a kȩdȩrang ȩl mo ȩr a stoang. 22

‘John ran from the beach to the store.’

In the sentences below, the italicized relational phrases are

also source phrases, except that the preceding verb is tran-
sitive. The object of the transitive verb is given in bold type:

(27) a. Aki ultȩbȩdii 23 a John ȩ r a chȩlsel a mlai.

‘We pulled John out of the car.’

b. Ak milȩngaus ȩr ngak ȩr a til a Toki.

‘I got myself some betel nut out of Toki’s
purse.

14.4.1. FURTHER TYPES OF SOURCE PHRASES

There is a subtype of source phrase where the noun phrase
following the relational word ȩ r refers to a human being.
Source phrases of this kind identify the person from whom
something is received, the person from whom some information
is heard, the person from whom some skill is learned, etc. Ob-
serve, for example, the italicized source phrases in the sen-
tences below:

(28) a. Ak milleng a udoud ȩr a Droteo.

‘I borrowed some money from Droteo.’

b. Ak rirengȩs a chisel a Toki ȩr a sȩchȩlik.

‘I heard the news about Toki from my friend.’

c. Ak milsuub a omȩluchȩs ȩl tȩkoi ȩr a Siabal ȩr a chad ȩr

a Siabal.
‘I learned Japanese writing from a Japanese person.’

331
 14 Relational Phrases

d. Ak milsuub a omȩruul ȩl mlai ȩr a dȩmak.

‘I learned canoe-making from my father.’

In order to ask a question about the source, we use a source

phrase consisting of the relational word ȩr followed by the
question word her ‘where?’ (or sometimes ngara ‘what?’) if the
source is a place, or tȩcha ‘who?’ if the source is a human being.
Thus, examples 26a and 28b above would be possible answers
to the following questions:

(26a’) A rȩkung ng tilobȩd ȩr {ker/a ngarang}?

‘Where did the crab emerge from?’

(28b’) Kȩ rirengȩs a chisel a Toki er tȩchang?

‘From whom did you hear the news about Toki?’

14.5. CAUSE PHRASES

Relational phrases are often used to indicate the cause of some
state or condition (or, less frequently, action). Such cause
phrases consist of the relational word ȩ r followed by a word
or expression which could be an appropriate cause for the
particular state or condition. As the following sentences show,
many Palauan cause phrases correspond to English phrases in-
troduced by ‘with’ or ‘from’:

(29) a. Ak smechȩr ȩr a tȩretȩr.

‘I’m sick with a cold/I’ve got a cold.’

b. A dȩmal a Droteo a mlad ȩr a kiubio.

‘Droteo’s father died of a heart attack.’

c. Ak {mȩsaul/skareter}ȩr a klȩdoraib.
‘I’m tired from (so much) driving around.’

d. Ng kekȩre a rȩnguk ȩr a kleald.

‘I’m very uncomfortable from the heat.’

e. A bilek a mlo cheisȩch ȩr a {chȩmachȩl/tuu}.

‘My clothes got stained with {betel nut (juice)/banana
juice}.’

332
 Palauan Reference Grammar

f. A ochik a mȩkekad ȩ r a udȩl.

‘My foot is itchy from the grass.’

g. A Toki a chuarm ȩr a dȩlȩngchȩklel.

‘Toki is suffering because of her living conditions.’

h. A bilsȩngek a rirȩchorȩch ȩr a eolt.

‘My boat sank in the storm.’

i. A bilek a mla mo kikiongȩl ȩr a {dȩchudȩch/rasȩch}. 24

‘My clothes have gotten soiled with {mud/blood}.’

j. A Droteo a mȩchȩsa ȩr a subȩlel.

‘Droteo is busy with his studies.’

k. A ngalȩk a milkar ȩr a chȩrrodȩch.

‘The child woke up from the noise.’

Since cause phrases name things rather than places or

persons, questions about the cause of some state or condition
are formed by using the question word ngara ‘what?’ following
the relational word ȩr. Sentences 29c and 29j, for example,
would be possible responses to questions like the following:

(29c’) Kȩ mȩsaul ȩr a ngarang?

‘What are you tired from?’

(29j’) A Droteo ng mȩchȩsa ȩr a ngarang?

‘What is Droteo busy with?’

14.6. TEMPORAL PHRASES

Relational phrases which name the time of an action or state
are called temporal phrases. In temporal phrases, the noun
phrase following the relational word ȩr can be a single word, as
in ȩr a klukuk ‘tomorrow’, er a elii ‘yesterday’, etc., or a group
of words, as in ȩr tia ȩ l me ȩ l rak ‘next year’, ȩr a ta ȩ l klok
‘at one o’clock’, etc. As we saw in 5.3 and 5.3.1–3, Palauan tem-
poral phrases can refer to specific time points in the present,
past, or future. In the sentences below, for example, the itali-
cized temporal phrases designate time points in the future:

333
 14 Relational Phrases

(30) a. A John a mo mȩrael ȩ r a klukuk.

‘John is going to leave tomorrow.’

b. Ng mo ȩr ngii a party ȩr a kȩbȩsȩngei.

‘There will be a party this evening.’

c. A sȩchȩlik a mo ȩr a Merikel ȩr tia ȩ l me ȩl

rak.
‘My friend is going to America next year.’

In 30a–b, the future time points represented by the italicized

temporal phrases are relatively close to the present moment
(i.e., the moment when the speaker utters the sentence), while
the future time point in 30c is relatively distant or remote.
The sentences below have temporal phrases denoting past
time points:

(31) a. A ngȩlȩkek a chilitii a skuul er a elii.

‘My child skipped school yesterday.’

b. A bȩchik a mle smechȩr ȩr a kȩsus.

‘My wife was sick last night.’

c. A Tadasi a mlo bȩchiil ȩr tia ȩl mlo mȩrek

ȩl rak.
‘Tadasi got married last year.’

The past time points in 31a–b are relatively recent (i.e., close to
the present moment), while the past time point in 31c is in the
relatively distant past.
The temporal phrase er a elȩcha ‘now, today’ can refer to the
present moment or to the “general” present in sentences like
the following:

(32) a. A Toki a mȩchiuaiu er a elȩchang.

‘Toki is sleeping now.’

b. Tȩ ngar ȩr ngii a rȩbȩtok ȩl chad ȩr a Siabal ȩl me er a

Belau er a elȩchang.
‘There are a lot of Japanese coming to Palau now/these
days.’

334
 Palauan Reference Grammar

This temporal phrase can also designate time points before and
after the moment when the sentence is uttered, but on the same
day. Note, therefore, the examples below:

(33) a. A sȩchȩlim a mla me er a elȩchang.

‘Your friend has come just now.’

b. Ak milsa a sensei er a elȩchang.

‘I saw the teacher today.’

c. Kȩ, mo ȩr a che er a elȩchang?

‘Are you going fishing today?’

Below is a partial listing of some of the more common Palauan

temporal phrases which are used to identify various points in
time:

(34) a. today and its parts:

er a elȩcha ‘now, today’

er a (elȩcha ȩl) tutau ‘this morning’
ȩr a suelȩb ‘this noon, this
afternoon’
ȩr a kȩbȩsȩnge ‘this evening’
ȩr a klȩbȩse ‘tonight’

b. time points in the past:

er a elii ‘yesterday’
ȩr a kȩsus ‘last night’
ȩr a tutau er a elii ‘yesterday morning’
ȩr a idȩlisȩb ‘the day before
yesterday’
ȩr a idȩlsȩbel 25 ‘three days ago’
ȩr a uche ȩr a idȩlsȩbel 26 ‘four or more days ago’
ȩr tia ȩl mlo mȩrek ȩl sande/ ‘last week/month/year’
buil/rak 27
ȩr se ȩl mlo mȩrȩk ȩl sande/ ‘the week/month/year
buil/rak 28 before last’
ȩr se ȩr a (taem ȩr a) mȩkȩmad ‘during the war’
ȩrse ȩr a taem ȩr a Siabal 29 ‘during the Japanese
times’
ȩr a irȩchar ‘in earlier times (at
least ten years ago)’

335
 14 Relational Phrases

ȩr a ititiumd ‘in ancient times’

c. time points in the future:

ȩr a klukuk ‘tomorrow’
ȩr a tutau ȩr a klukuk ‘tomorrow morning’
ȩr a kȩbȩsȩnge ȩr a klu kuk 30 ‘tomorrow evening’
ȩr a ngiaos ‘the day after
tomorrow’
ȩr a ngiosel 31 ‘three days from now’
ȩr a ikrel ngiosel 32 ‘four or more days from
now’
ȩr tia ȩl me ȩl sande/buil/rak 33 ‘next week/month/year’
ȩr se ȩl me ȩl sande/buil/rak 34 ‘the week/month/year
after next’

d. hours of the day:

ȩr a ta ȩl klok ‘at one o’clock’

ȩr a eru ȩl klok ‘at two o’clock’
ȩr a ede ȩl klok mȩ a tȩdobȩch ‘at three-thirty’
ȩr a tȩruich ȩl klok ȩr a tutau/ ‘at ten o’clock in the
klȩbȩse morning/evening’
ȩr a suelȩb ‘at noon’
ȩr a suelȩb ȩr a klȩbȩse ‘at midnight’
ȩr a tȩruich mȩ a eru ȩl klok ȩr ‘at twelve midnight’
a klȩbȩse

e. days of the week:

ȩr a kot ȩl ureor ‘on Monday’

ȩr a ongeru ȩl ureor 35 ‘on Tuesday’
ȩr a sabado ‘on Saturday’
ȩr a sande ‘on Sunday’

f. months of the year:

ȩr a kot ȩl buil ‘in January’

ȩr a ongeru ȩl buil ‘in February’
ȩr a ongede ȩl buil ‘in March’

g. days of the month:

336
 Palauan Reference Grammar

ȩr a euid ȩl kȩbȩsȩngil (a ‘on the seventh (of

ongeuid ȩl buil) 36 July)’
ȩr a tȩruich mȩ a eua ȩl ‘on the fourteenth (of
kȩbȩsȩngil (a kot ȩl buil) January)’
ȩr a lluich mȩ a elolȩm ȩl ‘on the twenty-sixth (of
kȩbȩsȩngil (a elȩcha ȩl buil) this month)’

h. frequency of occurrence:

ȩra bek ȩl tutau ‘every morning’

ȩr a bek ȩl sils ‘every day, daily’
ȩr a bek ȩl buil ‘every month, monthly’
ȩr a rokui ȩl taem ‘all the time’
ȩr a bebil ȩr a taem ‘sometimes’

Several sentences containing the temporal phrases of 34h,

which designate how frequently a particular action occurs, are
presented below: 37

(35) a. A Toki a me mȩngȩtmokl ȩr a blik ȩr a bek ȩl

tutau.
‘Toki comes and cleans my house every morning.’

b. Ak mo ȩr a Guam ȩr a bek ȩl buil.

‘I go to Guam every month.’

c. Ak mo ȩr a katsudo ȩr a bebil ȩr a taem.

‘I sometimes go to the movies.’

Temporal phrases designating specific time points such as

hours of the day, days of the month, months of the year, etc. can
be used to indicate the beginning point of some limited period
or span of time during which a given activity is pursued. In
such cases, the temporal phrase is always complemented by an
expression of the form ȩl mo + temporal phrase, which corre-
sponds to English ‘until…’ and indicates the time point at which
the activity ends. Observe the following examples:

(36) a. A Toki a mȩngȩtmokl ȩr a blai ȩr a eai ȩl klok ȩl mo (ȩr

a) etiu ȩl klok.
‘Toki cleans the house from eight o’clock until nine
o’clock.’

337
 14 Relational Phrases

b. Ak mo kie ȩr a Siabal ȩr a ongeim ȩl buil ȩl mo (ȩr a)

ongeuid ȩl buil.
‘I’ll be living in Japan from May until July.’

In 36a-b, the expressions ȩ l mo (ȩr a) etiu ȩl klok ‘until nine

o’clock’ and ȩl mo (ȩr a) ongeuid ȩl buil ‘until July’ are de-
pendent clauses (see 15.7) which specify the end-points of the
activities involved. The presence of ȩr a after mo ‘go’ is optional,
as indicated by the parentheses.
In order to ask a question about the time of an action or
state, we use a temporal phrase consisting of the relational
word ȩ r followed by the question word oingara ‘when?’ 38 Thus,
examples 30c, 31c, and 33b above would be possible answers to
the following questions:

(30c’) A sȩchȩlim ng mo ȩr a Merikel er

oingarang?
‘When is your friend going to America?’

(31c’) A Tadasi ng mlo bȩchiil er oingarang?

‘When did Tadasi get married?’

(33b’) Kȩ milsa a sensei er oingarang?

‘When did you see the teacher?’

14.7. SENTENCES WITH MORE THAN ONE

RELATIONAL PHRASE
As you may have noticed in examples like 7d, 21a, 21b, and 30c
above, many Palauan sentences contain two (or possibly even
more than two) relational phrases, depending on how much in-
formation the speaker wishes to supply. Thus, in 7d,

(7d) Ak mla ȩr a blil a Toki er a elii.

‘I was at Toki’s house yesterday.’

the first relational phrase is a locational phrase, while the

second is a temporal phrase. Similarly, in 21b

(21b) A John a mo ȩr a Guam ȩr a klukuk.

‘John is going to Guam tomorrow.’

338
 Palauan Reference Grammar

we observe a directional phrase followed by a temporal

phrase.
Examples 7d and 21b above show that if a temporal phrase
and some other kind of relational phrase occur together in the
same sentence, the temporal phrase always comes second. This
principle is further illustrated in the examples below, where a
temporal phrase (in italics) is preceded by some other kind of
relational phrase (in bold type):

(37) a. directional phrase designating characteristic goal +

temporal phrase (cf.25b):
A rȩchad a mle kakoad ȩr a kall ȩr se ȩ r a taem ȩ r a
mȩkȩ mad.
‘People fought over food during the war.’

b. source phrase indicating place + temporal phrase

(cf.26c):
A Toki a rirebȩt ȩr a chȩldukl ȩr a kȩsus e riros.
‘Toki fell off of the dock last night and drowned.’

c. source phrase indicating human being + temporal

phrase (cf.28a):
Ak milleng a udoud ȩr a Droteo er a elii.
‘I borrowed some money from Droteo yesterday.’

d. cause phrase + temporal phrase:

A rȩngȩlȩkek a mlo smechȩr ȩr a rȩktir a rȩchad ȩr a
Dois er a elii.
‘My children came down with the German measles
yesterday.’

It is also possible for two kinds of temporal phrases to

appear in the same sentence. In such cases, speakers prefer to
place the temporal phrase indicating the frequency of occur-
rence, or the greater time span, in final position. Observe the
examples below:

(38) a. A rubak a me ȩr a blik ȩr a etiu ȩl klok ȩr a bek ȩl tutau.

‘The old man comes to my house at nine o’clock every
morning.’

b. A John a mle ȩr a euid ȩl klok ȩr a kȩsus.

‘John arrived at seven o’clock last evening.’

339
 14 Relational Phrases

Since Palauan sentences can contain more than one rela-

tional phrase, as we have just seen, the formulas for Palauan
sentence types presented in 4 above must be modified, since
they incorrectly imply that at most, only one relational phrase
can appear in a given Palauan sentence. Thus, we will need to
replace the formulas of 4 with something like the following:

(39) a. subject noun phrase + transitive verb phrase (+ object

noun phrase) (+ relational phrase)n

b. subject noun phrase + intransitive verb phrase ( +

relational phrase)n

In the above formulas, the “raised n” notation (borrowed from

mathematics) will be interpreted to mean that relational
phrases can occur in a given sentence any “reasonable” number
of times. Thus, sentences with one or two relational phrases
are common; those with three, such as 38a above, occur occa-
sionally; and those with four (or more) are quite rare.

14.8. RELATIONAL PHRASES USED TO EXPRESS

COMPARISON
Another major function of Palauan relational phrases is to ex-
press comparison between two persons or things. Observe
how the italicized relational phrases are used in the examples
below:

(40) a. A Droteo a mȩsisiich ȩr a Toki.

‘Droteo is stronger than Toki.’

b. A rȩchad ȩr a Merikel a mȩkekȩmangȩt ȩ r a rȩchad ȩr

a Siabal.
‘Americans are taller than Japanese.’

c. A Toyota a ungil ȩl mlai ȩr a Datsun.

‘A Toyota is a better car than a Datsun.’

d. A kall ȩr a uum a mȩkngit ȩ r a kall ȩ r a restorangd.

‘The food at the cafeteria is worse than the food at the
restaurant.’

340
 Palauan Reference Grammar

e. A ududek a mȩkȩsai ȩ r a ududem.

‘I have less money than you.’

The noun phrases following ȩr in the italicized relational

phrases of 40 are being compared to the sentence subject with
respect to a particular quality—strength, height, etc. In such
a relationship of comparison, it is always the sentence subject
which is interpreted as possessing the particular quality in a
greater or higher degree.

14.9. FURTHER TYPES OF PALAUAN RELATIONAL

PHRASES
Although we have analyzed a large variety of Palauan relational
phrases in the sections above, there are still quite a few types
which we have not yet mentioned. Some of these types of rela-
tional phrases have already been examined in detail in earlier
chapters. For example, in 3.8–9, we looked at the structure of
possessor phrases and characterizational phrases. These
are relational phrases which occur directly following nouns in
examples like the following;

(41) delmȩrab ȩr ngak ‘my room’

sensei ȩr a Toki ‘Toki’s teacher’

chad ȩ r a Siabal ‘Japanese’

hong ȩr a sȩchȩlik ‘my friend’s
book’

ngalȩk ȩr a skuul ‘student’

kall ȩr a Sina ‘Chinese food’

Can you tell which of the italicized relational phrases of

41 are possessor phrases, and which are characterizational
phrases?
In addition, we saw in 5.4 that a relational phrase can be
used following an ergative verb form to identify the cause
or agent responsible for the event designated by the ergative
verb. Relational phrases of this kind are italicized in the
ergative sentences below:

(42) a. Alii, kȩ, mȩdul ȩ r a ngau!

341
 14 Relational Phrases

‘Watch out, you’ll get burned by the fire!’

b. A kall a mla mȩka ȩ r a katuu.

‘The food has been eaten by the cat.’

Some Palauan speakers find the examples of 42 rather awkward

and prefer to omit the relational phrase.
From time to time, you will undoubtedly observe sentences
in which relational phrases are used in such a way that they do
not fit neatly into any of the categories—e.g. locational phrase,
cause phrase, temporal phrase, etc.—discussed in this chapter
and elsewhere. The sentences below contain relational phrases
of this kind, which are difficult to classify and do not occur very
frequently: 39

(43) a. Ak ulȩmȩkȩdo ȩr a Toki ȩr a dengua.

‘I called Toki on the phone.’

b. Ak mo omȩchar ȩr ngii ȩr a kleim ȩl kluk.

‘I’ll buy it for $5.’

c. A delmȩrab ȩr ngak a chȩlsbrebȩr ȩr a bȩcheleleu.

‘My room is painted white.’

d. A blai a rruul ȩr a kȩrrȩkar.

‘The house is made of wood.’

e. A blik a{kmeed / chȩroid} ȩr a blil a Toki.

‘My house is {near / far from} Toki’s house.’

f. A tȩde ȩr tir a chiliis.

‘Three of them escaped.’

g. Ng chȩtik a omȩlmil a Droteo ȩ r a rrom.

‘I dislike Droteo’s drinking of liquor./I dislike the way
Droteo drinks (so much) liquor.’

h. Ng chȩtik a omȩngȩlir a rȩngalȩk ȩr a kiande. 40

‘I dislike the way the children are eating (so much)
candy.’

i. A tȩkoi ȩr a Merikel a kmal {mȩringȩl / beot} ȩr a

Droteo.

342
 Palauan Reference Grammar

‘English is very {difficult / easy} for Droteo.

j. Tia ȩl blai a {ungil / mȩkngit} ȩr ngak e le ng chȩroid

ȩr a rȩchad.
‘This house is {good / bad} for me because it’s far from
people.’

k. Ng ngar ȩr a mong.
‘It’s further over there.’

l. Ng ngar ȩr a mei.
‘It’s closer this way.’

m. Ng ngar ȩr a me mong.
‘It’s further in back of me.’

The sentences above show the following further uses of re-

lational phrases introduced by the relational word ȩr: means
of communication (43a), price (43b), medium or material
(43c-d), relationship of distance (43e), relationship of
whole to part (43f), object (43g-h), and person affected
(43i–j). Examples 43k-m appear to contain locational phrases,
but they are extremely unusual in that verb forms (the direc-
tional verbs mo or me—cf. chap. 13) follow ȩ r instead of the ex-
pected noun phrases.

343
 15 Dependent Clauses

15DependentClauses
15.1. CHARACTERISTICS OF DEPENDENT
CLAUSES
Palauan, like every other language, has many different gram-
matical patterns for combining relatively simple sentences into
more complex ones. Such patterns of complex sentence for-
mation allow speakers to express various types of relationships
between what would otherwise be isolated or disconnected
thoughts. Notice, for example, that the simple sentences of 1
and 2 below can be combined in two possible ways to yield the
more complex sentences 3a and 3b:

(1) A Droteo a ulȩba a oluchȩs.

‘Droteo had/was using a pencil.’

(2) A Droteo a milluchȩs a babier.

‘Droteo was writing a letter.’

(3) a. A Droteo a ulȩba a oluchȩs ȩl mȩluchȩs a

babier.
‘Droteo was using a pencil to write a letter.’

b. A Droteo a milluchȩs a babier ȩl oba a

oluchȩs.
‘Droteo was writing a letter with a pencil.’

The formal difference between 3a and 3b results, of course,

from combining the simple sentences 1 and 2 in opposite
orders. This formal difference in turn accounts for an important
difference in meaning, which we will briefly explain here (more
discussion will be given below). In 3a, the second part of the
sentence—namely, ȩ l mȩluchȩs a babier ‘(in order) to write
a letter’—explains the purpose for which the subject (Droteo)
used a pencil, which is the activity described in the first part of
the sentence—a Droteo a ulȩba a oluchȩs ‘Droteo was using a
pencil’. For this reason, the sequence ȩl mȩluchȩs a babier of
3a will be called a purpose clause. In 3b, however, the second
part of the sentence—namely, ȩ l oba a oluchȩs ‘using a pencil,

344
 Palauan Reference Grammar

with a pencil’—does not explain the purpose of the activity men-

tioned in the first part of the sentence. Instead, ȩl oba a oluchȩs
specifies the instrument (oluchȩs ‘pencil’) used in carrying out
the activity expressed by a Droteo a milluchȩs a babier ‘Droteo
was writing a letter’. For this reason, we will call sequences like
ȩl oba a oluchȩs of 3b instrument clauses.
Although we will discuss and illustrate Palauan purpose
clauses and instrument clauses in greater detail below, we now
know enough about sentences 3a-b above to make the following
observations. Even though 3a and 3b differ from each other
in that the former contains a purpose clause and the latter
contains an instrument clause, the two sentences nevertheless
share some important features. First, the word ȩl appears be-
tween the two parts of each sentence. Next, in the second part
of each sentence—i.e., in both the purpose clause and the in-
strument clause—there is no overtly-expressed subject: that is,
ȩ l is followed immediately by a verb form. 1 And, finally, the verb
following ȩl appears in the present tense (mȩluchȩs ‘write’
or oba ‘use, have’), even though sentences 3a and 3b describe
past actions.
In spite of the fact that the purpose and instrument clauses
of 3a-b lack subjects as well as any marker for the past tense,
the speaker of Palauan nevertheless knows exactly how to in-
terpret the meaning of these sentences. In interpreting a sen-
tence like 3b, for example, the speaker of Palauan automatically
understands the subject of the instrument clause ȩl oba a
oluchȩs ‘using a pencil, with a pencil’ to be identical to that of
the first part of the sentence—namely, Droteo. In other words,
it goes without saying that in 3b one and the same
person—Droteo—was the writer of the letter and the user of the
pencil. 2 Furthermore, the speaker knows that the time when the
instrument (oluchȩs ‘pencil’) was used is identical to that of the
verb milluchȩs ‘wrote’ in the first part of the sentence—namely,
past.
To summarize our discussion in the paragraph above, we
can see that even though the purpose and instrument clauses in
3a-b have no subject and no past tense marker, speakers never-
theless understand them as if they indeed did contain these two
elements. This phenomenon is due to the fact that the speaker’s
correct interpretation of the purpose or instrument clause de-
pends on (or is derived from) information found elsewhere in
the sentence—i.e., in the preceding clause. In other words, the

345
 15 Dependent Clauses

first part of the sentence tells the speaker what the (under-
stood) subject and tense of the purpose or instrument clause
should be. Because the correct interpretation of purpose and
instrument clauses therefore depends on information found in
another part of the sentence, purpose and instrument clauses
will be classified under the general term dependent clause.
The word ȩl, which appears at the beginning of every dependent
clause, will be called a dependent clause introducer.
Examples 1-3 above therefore show that two simple Palauan
sentences like 1 and 2 can be combined in such a way as to
form more complex sentences like 3a-b. In the formation of
complex sentences of this kind, one of the simple sentences 1
or 2 becomes a dependent clause, while the other becomes an
independent (or main) clause. In sentences like 3a-b, the in-
dependent clause comes first and the dependent clause follows.
The term “independent” is appropriate for the first clause in
3a-b since this clause, which contains a subject and a verb fully
specified for tense, could occur alone—or independently—as a
separate sentence of Palauan, as examples 1-2 prove. By con-
trast, dependent clauses cannot occur alone as separate sen-
tences because they are incomplete or deficient, as we have
seen above. Purpose clauses and instrument clauses are only
two of many different types of Palauan dependent clauses, as we
will see in the sections below.

15.2. PURPOSE CLAUSES

Any dependent clause which explains the purpose for which
some action is performed is a purpose clause. In the (ital-
icized) purpose clauses of the sentences below, there is no
overtly-expressed subject, and the verb appears in the present
tense even when the sentence designates a past activity or
event:

(4) a. A Droteo a{ulȩba/ulasbech}a biskang ȩ l omurȩch ȩ r a

ngikȩl.
‘Droteo {was using/needed} a spear to spear the fish.’

b. Ak ousbech a udoud ȩl mo mȩchar a bail.

‘I need money to (go) buy clothes.’

c. Ak ulusbech ȩr a Droteo ȩl mȩruul ȩ r a subȩlek.

346
 Palauan Reference Grammar

‘I needed Droteo to (help me) do my homework.’

d. A Droteo a ngiluu a mlai ȩl mo mȩchar a biang.

‘Droteo took the car in order to go buy beer.’

e. Kȩ okiu ker ȩ l mo ȩ r a kȩdȩrang?

‘What route do you take to get to the beach?’

f. Ng ngar ȩr ngii a bilem ȩ l mo ȩ r a party?

‘Do you have clothes to go to the party with?’

g. Ng ngar ȩr ngii a bento ȩr kau ȩl mo ȩ r a chei?

‘Do you have your lunch for when you go fishing?

In the sentences above, it is not difficult to understand why

we classify the italicized dependent clauses as purpose clauses.
In 4b, for instance, the dependent clause ȩl mo mȩchar a bail
‘(in order) to (go) buy clothes’ explains the purpose for which
the subject (ak ‘I’) needs the money. And in 4d, ȩl mo mȩchar a
biang ‘(in order) to go buy beer’ expresses the subject’s purpose
in using the car. Sentences 4f-g ask questions about the ex-
istence or availability of things which will serve a particular
purpose or function: in 4f, the speaker wants to know if the
hearer has proper clothes for the party, and in 4g, he wants to
know if the hearer has brought along a lunch which he can eat
while fishing.

15.2.1. PURPOSE CLAUSES CONTAINING DIRECTIONAL VERBS

Palauan sentences with purpose clauses often exhibit the fol-
lowing pattern: the independent clause contains one of the
three directional verbs me ‘come’, mo ‘go’, or eko ‘go’ (cf. chap.
13), and the (dependent) purpose clause contains an expression
of the form directional verb + action verb (cf. 13.3) in which
the directional verb matches that of the independent clause. In
the following sentences, which explain the subject’s purpose in
coming or going to a particular place, the two occurrences of a
given directional verb have been italicized:

(5) a. A sȩchȩlik a me ȩr a blik ȩl me mȩsuub.

‘My friend is coming over to my house to study.’

b. Ng sȩbȩchek ȩl eko ȩr a blim ȩl eko rȩmuul a kall.

347
 15 Dependent Clauses

‘I can come to your house to prepare the food.’

c. Ak mlo ȩr a kȩdȩra ȩl mo mȩngȩdub.

‘I went to the beach in order to go swimming.’

d. A Droteo a mlo ȩr a stoa ȩl mo omȩchar a kall.

‘Droteo went to the store to buy food.’

e. Ak mo ȩr a Siabal ȩl mo mȩsuub a tȩkoi ȩr a

Siabal.
‘I’m going to Japan in order to study Japanese.’

Even though the second occurrence of a given directional verb

does not seem to add anything to the meaning of the purpose
clauses in 5, it is nevertheless required if the sentence is to be
grammatical. Thus, in sentences with the structure independent
clause + purpose clause, any directional verb occurring in the
independent clause must be repeated in the purpose clause. The
repeated directional verb found in the purpose clause appears
in the present tense form, regardless of the time of the event
which the sentence designates.
In view of the above discussion and material presented in
13.4, we can easily explain the difference in meaning between
the following contrasting sentences:

(6) a. Ak mo mȩruul a kall.

‘I’m going to prepare the food.’

b. Ak mo (ȩr a blil a Toki) ȩl mo mȩruul a kall.

‘I’m going (to Toki’s house) in order to prepare the
food.’

Since the directional verb mo ‘go’ is being used in 6a to des-

ignate future time, 6a is merely an assertion by the speaker
that he plans (or intends) to prepare food in the relatively near
future. In 6b, however, the activity of preparing the food is ex-
pressed in the purpose clause ȩl mo mȩruul a kall ‘(in order)
to prepare the food’; therefore, 6b is a sentence in which the
speaker explains his pur pose in going somewhere (e.g. Toki’s
house). Because of this meaning difference, 6a might be an
answer to a question like ‘What are you going to do now?’
whereas 6b would be a possible response to the question ‘Why
are you going (to Toki’s house)?’

348
 Palauan Reference Grammar

15.2.2. PURPOSE CLAUSES FOLLOWING NOUNS

As the sentences in 4-5 above show, Palauan purpose clauses
normally occur in complex sentences following an independent
clause. It is also possible for purpose clauses to appear fol-
lowing noun phrases of possession (cf. 3.7) containing the
possessed forms of the nouns techall ‘opportunity, chance’ and
taem ‘time’. The sentences below contain possessed nouns like
techȩllek ‘my opportunity’, techȩllem ‘your opportunity’, temek
‘my time’, temel ‘his/her time’, etc.; here, the possessor suf-
fixes -ek, -el, -em, etc. identify the person who has the time or
opportunity to do something:

(7) a. Ng ngar ȩr ngii a temem ȩ l mo milil ȩ r a kȩbȩsȩngei?

‘Do you have time to play this evening?’

b. Ng diak a temek ȩl mo ȩr a party.

‘I don’t have time to go to the party.’

c. Ng mo ȩr ngii a techȩllem ȩl mo ȩ r a Siabal?

‘Will you have a chance to go to Japan?’

d. Ng dimlak a techȩllel a Droteo ȩl omes ȩr a rȩsȩchȩlil.

‘Droteo didn’t get an opportunity to get together with his
friends.’

e. Ng di kea a techȩllek ȩl mo ȩ r a skuul.

‘I no longer have the opportunity to go to school.

In the examples above, we interpret the italicized portions

as purpose clauses because they explain the purpose to which
some period of time is to be directed or devoted. Thus, in 7b, for
instance, the speaker says he has no time which he can devote
to going to the party; and in 7e, the speaker claims to have no
opportunity (or free time) which he can direct towards the ac-
tivity of going to school. As expected, the purpose clauses in
7 above always contain present tense verb forms, and they do
not have any expressed subjects. Here, the understood subject
of the purpose clause is interpreted as being identical to the
person referred to by the particular possessor suffix attached
to the nouns techall or taem. For example, in 7e, the person
who no longer has the opportunity to go to school—namely, the

349
 15 Dependent Clauses

speaker of the sentence (‘I’)—is the same as the person iden-

tified by the 1st pers. sg. possessor suffix -ek in the possessed
noun techȩllek ‘my opportunity/chance’. 3

15.3. INSTRUMENT CLAUSES

Any dependent clause which identifies the instrument used to
carry out a particular action or activity is called an instrument
clause. Instrument clauses always contain the verb oba ‘use/
have/hold it’ 4 followed by a noun phrase object which names
the instrument used. Thus, in 3b above, repeated here for con-
venience,

(3b) A Droteo a milluchȩs a babier ȩl oba a oluchȩs.

‘Droteo was writing a letter with a pencil.’

the independent clause a Droteo a milluchȩs a babier ‘Droteo

was writing a letter’ is followed by the (dependent) instrument
clause ȩl oba a oluchȩs ‘using a pencil, with a pencil’, in which
oba ‘use/have/hold it’ happens to take the instrument noun
oluchȩs ‘pencil’ (cf. 8.6) as its object. Because the independent
clause and the instrument clause in 3b both have their own
verbs (milluchȩs ‘was writing’ and oba ‘use/have/hold it’, re-
spectively), the word-for-word equivalent for a sentence like 3b
would be something like ‘Droteo was writing a letter using/
holding a pencil’. The most suitable English translation,
however, uses a phrase introduced by (the preposition) with.
Recall that we have already used the term “instrument”
in our discussion of a particular type of Palauan noun deri-
vation: thus, in 8.6 we saw that Palauan instrument nouns
are derived by replacing the verb marker of an imperfective
verb form with the prefix o-, as in mȩluchȩs ‘write’—oluchȩs
‘pencil’, mȩles ‘cut’— oles ‘knife’, etc. In our present discussion,
we are using the term “instrument” in a somewhat less tech-
nical sense, since Palauan instrument clauses do not necessarily
have to contain an instrument noun in o- as does the instrument
clause in 3b. Thus, in the sentences below, the italicized in-
strument clauses contain words like taod ‘fork’, biskang ‘spear’,
etc., which are not instrument nouns in the technical sense de-
fined above but which nevertheless identify the “instrument”
(tool, utensil, etc.) with which a particular activity or task is per-
formed:

350
 Palauan Reference Grammar

(8) a. A Droteo a mȩnga a ngikȩl ȩl oba a {taod / chimal}.

‘Droteo is eating the fish with {a fork/his hand}.

b. Ak {milkodir / ulȩmurȩch} a ngikȩl ȩl oba a biskang.

‘I {killed/speared} a fish with a spear.’

c. Ak milȩngiis ȩr a kliokl ȩl oba a sebȩl.

‘I was digging the hole with a shovel.’

d. A Droteo a chillebȩd a bilis ȩl oba a kȩrrȩkar.

‘Droteo hit the dogs with a stick.’

e. A Droteo ng millȩkosȩk ȩr a tech ȩl oba a ngarang?

‘What was Droteo cutting the meat with?’

f. A rȩchad ȩr a Sina a omȩngur ȩl olab a hasi.

‘The Chinese eat with chopsticks.’

g. A rȩngalȩk ȩr a skuul a mȩruul a subȩlir ȩl olab a

manneng.
‘The students do their homework with fountain pens.’

As the above examples show, the instrument involved is nor-

mally something inanimate (i.e., non-living) such as a spear, a
shovel, etc.; but once in a while, some part of the body such
as the hand (chimal of 8a) can function as the instrument.
Example 8e is a question about what instrument the subject
(Droteo) was using to cut the meat—hence, the occurrence of
the question word ngara ‘what?’ as object of oba in the in-
strument clause. The instrument clauses of 8 of course show
the two major defining characteristics of Palauan dependent
clauses: first, they have no overtly-expressed subject; and
second, they have present tense verb forms, even when the sen-
tence describes a past action or activity, as in 8b-e.

15.4. PURPOSE AND INSTRUMENT CLAUSES

WITH PAST TENSE VERB FORMS
In the sections above, we emphasized the fact that Palauan
purpose and instrument clauses normally have present tense
verb forms, even when the verb of the preceding independent
clause is in the past tense and therefore indicates a past event.

351
 15 Dependent Clauses

A considerable number of Palauan speakers, however, can use

sentences in which the verb of both the independent clause
and the following purpose or instrument clause appears in the
past tense. As a typical example, observe the pair of sentences
below, which have a purpose clause:

(9) a. Ak ulȩba a sebȩl ȩl mȩngiis ȩr a kliokl.

‘I was using a shovel to dig the hole.’

b. Ak ulȩba a sebȩl ȩl milȩngiis ȩr a

kliokl.
‘I used a shovel to dig the hole.’

The only formal difference between 9a and 9b above is that the

tense of the verb in the dependent clause is present vs. past,
respectively. While some speakers think that 9a and 9b are iden-
tical in meaning, others feel there is a rather subtle difference
between them, which we will attempt to explain below.
In a sentence like 9a, the speaker uses the present tense
form of mȩngiis ‘dig’ in the purpose clause if he considers the
action involved to have been in progress when some other past
event occurred. In 9a, this interrupting or coinciding past event
is merely implied, but not expressed; if we wish to express such
an event, we can put 9a in the context of a time clause (see
22.2) like er se ȩr a lȩme a Droteo ‘when Droteo came’, as in
the following:

(9a’) Ak ulȩba a sebȩl ȩl mȩngiis ȩr a kliokl er se ȩr a lȩme a

Droteo.
‘I was using a shovel to dig the hole when Droteo came
by.’

As opposed to 9a, example 9b contains the past tense form

of mȩngiis ‘dig’ in the purpose clause. Here, the speaker views
the action involved as a total, completed experience, and no ref-
erence is made to some other event which may have interrupted
it or coincided with it. Thus, while 9a would be an answer to a
question like ‘What were you using to dig the hole when Droteo
saw you yesterday?’ 9b is simply an answer to the question
‘What did you use to dig the hole yesterday?’ Furthermore,
since 9b seems to focus on a total, completed event, it would
be strange to place it in the context of a time clause like er se
ȩr a leme a Droteo ‘when Droteo came’, which draws our at-

352
 Palauan Reference Grammar

tention to a past action as it was in progress. Thus, while 9a’

above is a natural sentence, the following is contradictory and
unacceptable (hence, no English equivalent is given): 5

(9b’) ??Ak ulȩba a sebȩl ȩl milȩngiis ȩr a kliokl er se ȩr a lȩme

a Droteo.

15.5. MEANS OF TRANSPORTATION CLAUSES

Any dependent clause which specifies the means of trans-
portation used to move from one location to another can be
called a means of transportation clause. Dependent clauses
of this type always contain the existential verb ngar ‘exist,
be (located)’ (see 18.2), followed by a locational phrase (cf.
14.2). The locational phrase after ngar consists of the relational
word ȩr followed by a noun phrase which names any kind of ve-
hicle—e.g., car, train, boat, etc. The following sentences contain
typical means of transportation clauses:

(10) a. Ak mlo ȩr a Siabal ȩl ngar ȩr a skoki.

‘I went to Japan by plane.’

b. A Droteo a blȩchoel ȩl mo ȩr a skuul ȩl ngar ȩr a

sidosia.
‘Droteo always goes to school by car.’

c. A Masaharu a mlo ȩr a Merikel ȩl ngar ȩr a diall.

‘Masaharu went/has gone to America by ship.’

Because the independent clause and the means of transpor-

tation clause in the sentences above both have their own verbs,
a sentence like 10a would have a word-for-word equivalent
something like ‘I went to Japan being in a plane.’ Such a sen-
tence is of course unacceptable in English, which must use
a phrase introduced with (the preposition) by to express the
means of transportation.
Notice that the order of clauses can be reversed in the sen-
tences of 10 above. Thus, in addition to 10a, we can have the
following sentence, which is identical in meaning:

(11) Ak mla ȩr a skoki ȩl mo ȩr a Siabal.

353
 15 Dependent Clauses

‘(lit.) I was in a plane going to Japan.’ = ‘I went to Japan

by plane.’

The dependent clause ȩl mo ȩ r a Siabal ‘going to Japan’ in 11 is

best classified as a specifying clause (see 15.7 below).
Like the purpose and instrument clauses discussed in
15.2–3 above, the means of transportation clauses in 10 have
no overtly-expressed subject. As expected, the subject of the
means of transportation clause is understood to be identical to
that of the preceding independent clause. Furthermore, the ex-
istential verb ngar ‘exist, be (located)’ appears in the present
tense in examples like 10a and 10c, even though the event in
question occurred in the past. Some Palauan speakers can also
use means of transportation clauses containing the past tense
form of the existential verb ngar; thus, with 10c, compare the
following sentence:

(12) A Masaharu a mlo ȩr a Merikel ȩl mla ȩr a

diall.
‘Masaharu went to America by ship.’

For those speakers who feel there is a meaning distinction be-

tween 10c and 12, the difference in interpretation resembles
that mentioned in 15.4 above with reference to purpose and in-
strument clauses. Thus, in 10c, use of the present tense form
ngar ‘exist, be (located)’ in the dependent clause refers to some
state which is still in progress and not yet finished. Therefore,
10c implies that the subject (Masaharu) is still on the ship trav-
eling to America. In 12, however, the past tense form mla ‘ex-
isted, was (located)’ indicates that the speaker views the state
of being on the ship as a total, completed experience. For this
reason, 12 implies that Masaharu is no longer on the ship, but
has already reached his destination.
The following sentence asks a question about the means of
transportation used; therefore, the question word ngara ‘what?’
appears in the locational phrase following ngar:

(13) Ng mlo ȩr a skuul ȩl ngar ȩr a ngarang?

‘How did he go to school?’

354
 Palauan Reference Grammar

15.6. ACCOMPANIMENT CLAUSES

Any dependent clause which identifies the person with whom
some action or activity is being done is termed an accom-
paniment clause. Palauan accompaniment clauses always
contain the very special word obȩngkel ‘be together with’,
which expresses the idea of accompaniment. The term “ac-
companiment” refers to a relationship between human beings
(or, sometimes, animals) such that the individuals involved are
doing something together—i.e., in the same place and at the
same time. The word obȩngkel is unique in Palauan in that it
has the characteristics of both a verb and a noun. Just like
any verb in o-, obȩngkel has a past tense form in ul(ȩ)—namely,
ulȩbȩngkel ‘was together with’ (cf. 5.3.2); furthermore, its dis-
tribution is like that of verbs, since it directly follows the subject
of the sentence (see 14 below). At the same time, however,
obȩngkel behaves like an obligatorily possessed noun (cf.
3.5) because it must always contain a possessor suffix; this
possessor suffix agrees in person and number with the person
who is accompanying the subject in the pursuit of a particular
activity. The characteristics of obȩngkel described above are il-
lustrated in the following simple sentences:

(14) a. Ak ulȩbȩngkel a Toki er a elii.

‘I was with Toki yesterday.’

b. Ak ulȩbȩngktȩrir 6 a rȩsȩchȩlik er a
elii.
‘I was with my friends yesterday.’

In both 14a and 14b, the “accompaniment” word obȩngkel ap-

pears in its past tense form. In each sentence, however, the pos-
sessor suffix on obȩngkel is different because it must agree in
person and number with whoever accompanied the subject (ak
‘I’) in performing the activity. Thus, in 14a, the possessor suffix -
el agrees with the 3rd pers. sg. noun Toki, while in 14b, the pos-
sessor suffix -tȩrir agrees with the 3rd pers. human pl. noun rȩ
sȩchȩlik ‘my friends’. Except for the fact that the o- of obȩngkel
changes to ul(ȩ)- in the past tense, phrases like ulȩbȩngkel a
Toki ‘together with Toki, in Toki’s presence’ and ulȩbȩngktȩrir a
rȩsȩchȩlik ‘together with my friends, in my friends’ presence’ of
14 seem to be no different from the noun phrases of charac-
terization discussed in 3.7.

355
 15 Dependent Clauses

In the sentences below, the special word obȩngkel occurs in

the italicized accompaniment clause introduced by ȩl:

(15) a. Ak mlo ȩr a kȩdȩra ȩ l obȩngkel a Droteo.

‘I went to the beach with Droteo.’

b. Ng sȩbȩchem ȩl mo ȩl obȩngkek?
‘Can you go with me?’

c. Ak mililil ȩl obȩngktȩrir a rȩsȩchȩlik.

‘I was playing with my friends.’

d. Ng soak ȩl mo mȩngȩdub ȩ l obȩngkem.

‘I want to go swimming with you.’

The order of clauses can be switched in the examples of 15

above. Thus, in addition to 15a, we can have the following sen-
tence, which has the same meaning:

(16) Ak ulȩbȩngkel a Droteo ȩl mo ȩr a kȩdȩrang.

‘(lit.) I was together with Droteo going to the beach.’ = ‘I
went to the beach with Droteo.’

As we will see in 15.7 below, the dependent clause in 16 is a

type of specifying clause.
Like the other kinds of dependent clauses described in the
sections above, accompaniment clauses never have an overtly-
expressed subject. In addition, the accompaniment word
obȩngkel can appear in the present tense, even when the whole
sentence involves a past action or event, as in 15a and 15c
above. For some speakers, there is a contrast between the
use of the present vs. past tense form of obȩngkel in the de-
pendent clause. Thus, with 15c, for instance, compare the sen-
tence below:

(17) Ak mililil ȩl ulȩbȩngktȩrir a rȩsȩchȩlik.

‘I played with my friends.’

The difference in meaning between 15c and 17 should now be

familiar to us: while 15c implies that the subject (ak ‘I’) was
playing with his friends when some other interrupting or co-
inciding event occurred, 17 describes a past event as a to-
tality without reference to any other intervening event. For this

356
 Palauan Reference Grammar

reason, 15c could be an answer to a question like ‘What were

you doing when Droteo came?’ and could occur in the following
context:

(15c’) Ak mililil ȩl obȩngktȩrir a rȩsȩchȩlik er se ȩr a lȩme a

Droteo.
‘I was playing with my friends when Droteo came.’

By contrast, 17 would be an answer to a question like ‘What did

you do (yesterday)?’. Therefore, putting it in the context sup-
plied by 15c’ would result in an “illogical” sentence:

(17’) ?? Ak mililil ȩl ulȩbȩngktȩrir a rȩsȩchȩlik er se ȩr a lȩme a

Droteo.

In order to ask a question about who accompanied someone

in doing something, we use an accompaniment clause con-
taining the question word tȩcha ‘who?’, as in the following:

(18) Kȩ mlo ȩr a party ȩl obȩngkel tȩchang?

‘Whom did you go to the party with?’

15.7. SPECIFYING CLAUSES

So far we have classified Palauan dependent clauses into four
types (purpose, instrument, means of transportation, and ac-
companiment) on the basis of the particular kinds of meanings
they convey. This convenient classification should not prevent
us from recognizing that the four types of clauses are basically
very similar in function. All of them have the common function
of specifying, qualifying, or giving further information
about the action, activity, or event of the preceding independent
clause. Thus, in 8b, for example, the instrument clause ȩl oba
a biskang ‘using a spear, with a spear’ provides us with further
information about the subject’s killing of the fish by telling us
what he killed it with; and in 15a the accompaniment clause ȩ
l obȩngkel a Droteo ‘together with Droteo’ specifies or narrows
down the scope of the action of the independent clause—the
subject’s going to the beach—by telling us who accompanied
the subject to the beach. Thus, the four types of dependent
clauses discussed in the sections above could all be identified
by the broader “cover” term specifying clause. We will restrict

357
 15 Dependent Clauses

our use of this term, however, to any dependent clause which,

while fulfilling the rather general function of specifying or nar-
rowing down the scope of the action or state of the preceding in-
dependent clause, neverthless cannot be conveniently assigned
to any of the four categories of dependent clauses described
above. The use of such specifying clauses is illustrated in the
following examples:

(19) a. A Droteo a ulureor ȩl mȩngȩsbrebȩr a blai.

‘Droteo was working/used to work painting houses.’

b. A sȩchȩlik a mla ȩr a diall ȩl me ȩ r a Belau.

‘(lit.) My friend was in a ship coming to Palau.’ = ‘My
friend came to Palau by ship.’

c. A Toki a ulȩbȩngkel a Droteo ȩl mo ȩ r a party.

‘(lit.) Toki was together with Droteo going to the party.’
= ‘Toki went to the party with Droteo.’

d. A Toki a mle dȩngchokl ȩl kmeed ȩ r a Droteo.

‘Toki was seated near Droteo.’

In 19a, the italicized specifying clause narrows down or

limits the scope of the action of working expressed in the
independent clause: in other words, ȩl mȩngȩsbrebȩr a blai
‘painting houses’ describes the particular kind of work involved.
In 19b- d, the specifying clauses all serve to narrow down the
scope of various states described by the independent clauses.
Thus, in 19b, ȩl me ȩr a Belau ‘coming to Palau’ explains the
circumstances under which the subject (sȩchȩlik ‘my friend’)
was on the ship, while in 19c ȩl mo ȩ r a party ‘going to the
party’ clarifies the circumstances under which the subject (Toki)
was with Droteo. Finally, in 19d, ȩl kmeed ȩ r a Droteo ‘near
Droteo’ gives us specific information about where the subject
(Toki) was seated. The grammatical characteristics of the spec-
ifying clauses in 19 above are identical to those of the other
dependent clauses so far examined; in addition, some speakers
can use specifying clauses containing past tense verb forms,
with the expected difference in meaning. Since we have men-
tioned these phenomena so many times in the sections above,
no further discussion will be necessary here. 7

358
 Palauan Reference Grammar

15.7.1. SPECIFYING CLAUSES CONTAINING DIRECTIONAL VERBS

There are many Palauan sentences in which an independent
clause containing a verb indicating movement is followed by a
specifying clause containing one of the three directional verbs
me ‘come’, mo ‘go’, or eko ‘go’ (cf. chap. 13). In such sen-
tences, the specifying clause narrows down the scope of the
action designated in the preceding independent clause by indi-
cating the direction of move ment. As we saw in 14.3, verbs
involving movement are of two types: intransitive movement
verbs like mȩrael ‘walk, travel’, rȩmurt ‘run’, etc. simply de-
scribe how someone or something moves from one location to
another, while transitive movement verbs like oba ‘carry, bring,
take’ and mȩlai ‘bring, take’ 8 describe how some person or
thing is transferred from one location to another.
In the examples below, the specifying clauses containing di-
rectional verbs follow independent clauses which have intran-
sitive movement verbs:

(20) a. Ak mirrael ȩl mo ȩr a Belau.

‘I travelled to Palau.’

b. A Droteo a mȩrael ȩl mo ȩr a stoang.

‘Droteo is walking to the store.’

c. A Toki a rirurt ȩl me ȩr a blik.

‘Toki ran to my house.’

d. Ng sȩbȩchem ȩl ridȩkekl ȩl mo ȩr a bitang?

‘Can you jump to the other side?’

The sentences below are similar to those of 20 except that they

also contain a source phrase (cf. 14.4), which tells us the place
where the movement began or originated:

(21) a. A katuu a rirȩdȩkekl ȩr a chȩdȩuel a blik ȩl mo ȩr a

kȩrrȩkar.
‘The cat jumped from the roof of my house into the
tree.’

b. Ak mirrael ȩr a katsudokang ȩl mo ȩr a kȩdȩrang.

‘I walked from the theater to the beach.’

359
 15 Dependent Clauses

Since the examples of 20–21 contain two clauses and there-

fore two verbs, their word-for-word translations are rather dif-
ferent from the English equivalents given. Thus, a sentence
like 20c means, literally, ‘Toki ran coming to my house’. In the
English equivalent, the directional verb me ‘come’ of the speci-
fying clause is not translated, and the (prepositional) phrase to
my house suffices to convey the meaning. Note, further, that the
specifying clause structure is essential to the meaning of the
sentences in 20. Thus, if we remove ȩl me from 20c, we get the
following sentence, which is quite different in meaning:

(22) A Toki a rirurt ȩr a ‘Toki was running at my

blik. house.’

Here, the relational phrase ȩr a blik ‘at my house’ tells where

the action of running took place and therefore should be clas-
sified as a locational phrase (cf. our discussion in chap. 14,
note 22).
The following sentences resemble those of 20 except that
the independent clause contains a transitive movement verb.
For ease of reference, the object of the transitive movement
verb has been italicized:

(23) a. Ak ullab a ilumȩl ȩl mo ȩr a party.

‘I brought drinks to the party (as my share/
contribution).’

b. Ng tȩcha a ulȩba a John ȩl eko ȩr a blim?

‘Who took John to your house?’

c. A Droteo a ngilai a ilumȩl ȩl me ȩr a party.

‘Droteo brought/carried the drinks to the party.’

d. Ak nguu a bȩchik ȩl mo ȩr a ochȩraol.

‘I’m taking my wife to the money-raising party.’

e. Ak ngoititȩrir a rȩsȩchȩlik ȩl mo ȩr a ochȩraol.

‘I’m taking my friends to the money-raising party.’

f. Ng soak ȩl oldurokl ȩr a ngȩlȩkek ȩl mo ȩr a

Merikel.
‘I want to send my child to America.’

360
 Palauan Reference Grammar

g. A sensei a ulduruklii a dempo ȩl mo ȩr a Saibal.

‘The teacher sent the telegram to Saipan.’

As the above examples show, the transitive movement verbs oba

‘carry, bring, take’, mȩlai ‘bring, take’, and oldurokl ‘send’ can
take either human or non-human objects. The specifying clause
structure is required in the sentences of 23 to express the de-
sired meaning of transferring someone or something from one
location to another. Thus, if we attempt to remove ȩl me from a
sentence like 23c, we get the following, whose meaning is very
different:

(24) A Droteo a ngilai a ilumȩl ȩr a party.

‘Droteo brought party-drinks.’

Here, the object of ngilai ‘brought’ is the noun phrase of char-

acteri zation ilumȩl ȩr a party ‘party-drinks’, which consists
of the noun ilumȩl ‘drink’ followed by the characterizational
phrase ȩr a party ‘for the party’ (cf. 3.7–9).

15.7.2. SPECIFYING CLAUSES FOLLOWING lmuut

A common Palauan sentence type involves a specifying clause
preceded by an independent clause containing the intransitive
action verb lmuut ‘return’. In the sentences below, the speci-
fying clause has a directional verb:

(25) a. A Droteo a lmuut ȩl mo ȩr a Belau ȩr a

klukuk.
‘Droteo is going back to Palau tomorrow.’

b. A John a liluut ȩl me ȩr a Hawaii er a elii.

‘John came back to Hawaii yesterday.’

In the examples of 25, the specifying clause simply clarifies

whether the place to which the subject is returning is in a
direction away from (as in 25a) or towards (as in 25b) the
speaker’s location.
It is also possible to have examples like 25 above in which
any action verb at all occurs in the specifying clause. In such
cases, the verb lmuut conveys the more general meaning ‘do
(something) again/a second time’. Some typical sentences are
now given:

361
 15 Dependent Clauses

(26) a. Ak liluut ȩl mȩnguiu ȩr a hong.

‘I reread the book.’

b. A delmȩrab ȩr ngak a kikiongȩl mȩ a Toki a lmuut ȩl me

mȩngȩtmokl ȩr ngii.
‘My room’s dirty, so Toki will come clean it up again.’

c. Ak liluut ȩl dmu ȩr a Droteo ȩl ua se ng diak lsȩbȩchek

ȩl mong.
‘I repeated to Droteo that I cannot go.’

15.7.3. SPECIFYING CLAUSES FOLLOWING dmak

In 15.6 above, we examined sentences like the following, which
express accompaniment:

(27) A Toki a mlo ȩr a kȩdȩra ȩl obȩngkel a

Droteo.
‘Toki went to the beach with Droteo.’

Recall that the italicized accompaniment clause in 27 con-

tains the unusual word obȩngkel ‘be together with’. Now, with
27, compare the following sentence, which is quite similar in
meaning:

(28) A Toki mȩ a Droteo a dilak ȩl mo ȩ r a

kȩdȩrang.
‘Toki and Droteo went to the beach together.’

In 28, the italicized specifying clause is preceded by an inde-

pendent clause containing the intransitive verb dmak ‘be to-
gether’ (past: dilak). Both 27 and 28 imply that two persons
(Toki and Droteo) carried out the same activity (going to the
beach) in each other’s company—i.e., at the same time and in
the same place. The only difference between the two sentences
involves where the speaker’s attention or interest is focused.
Thus, in 27, the speaker is interested mainly is Toki’s activ-
ities—that is, in what happened from Toki’s point of view; in 28,
however, the speaker is focusing more or less equal attention on
the two individuals involved. The sentences below are similar
in structure and interpretation to 28; the italicized specifying
clause tells us what activity the subjects did in each other’s
company:

362
 Palauan Reference Grammar

(29) a. A Toki mȩ a Droteo a dmak ȩl mȩsuub ȩ r a

library.
‘Toki and Droteo are studying together at the
library.’

b. A rȩsȩchȩlik a dilak ȩl mȩngiis ȩ r a kliokl.

‘My friends were digging the hole together.’

As the sentences of 28 and 29 show, the subject of dmak ‘be

together’ must always be plural: thus, in 28 and 29a, the two
nouns Toki and Droteo are joined by the connecting word mȩ
‘and’ (see 25.4) to form a plural subject, and in 29b, the plural
prefix rȩ- (cf. 2.5) in rȩsȩchȩlik ‘my friends’ indicates that this
word refers to two (or more) individuals. A similar restriction
concerning the subjects of reciprocal verbs was observed in
10.1.
Different in meaning from 29a above is the following sen-
tence:

(30) A Toki mȩ a Droteo a mȩsuub ȩr a library ȩl terung.

‘Toki and Droteo are both studying at the library.’

While 29 implies that Toki and Droteo are studying at the library
in each other’s company (sitting side-by-side, etc.), 30 simply
states that both persons are at the library studying, not neces-
sarily in each other’s company. In 30, the implication that both
Toki and Droteo are studying at the library individually is sup-
plied by the specifying clause ȩ l teru ‘(lit.) as two (people)’, in
which the number word teru refers to two human beings (see
24.4, ex. 27).
The contrast in meaning between 29a and 30 is observed in
the following pairs of sentences as well:

(31) a. A rȩngalȩk a dilak ȩl rȩmurt.

‘The children were running together.’

b. A rȩngalȩk a rirurt ȩl terung.

‘The children were both running.’

(32) a. Aki dmak ȩl mȩlȩkoi a tȩkoi ȩr a Merikel.

‘We speak English simultaneously/all together
(during a language drill, etc.).’

363
 15 Dependent Clauses

b. Ng sȩbȩcham ȩl mȩlȩkoi a tȩkoi ȩr a Merikel ȩl

tedei.
‘The three of us can speak English.’

While 31a means that the children were running together (side-
by-side), 31b simply says that each of the two children was
running (possibly in different directions). Similarly, while 32a
implies that the several persons represented by the subject aki
‘we’ repeat words or sentences of English in unison (as part of
a language drill, etc.), 32b has no connotation of simultaneous
performance of the same action and merely says that the three
persons are able to speak English.
The following sentences further illustrate the use of spec-
ifying clauses containing number words referring to human
beings:

(33) a. A Toki mȩ a Droteo a smechȩr ȩl

terung.
‘Both Toki and Droteo are sick.’

b. Ak mlo ȩr a kȩdȩra ȩl di ngak ȩl

tang.
‘I went to the beach alone.’

In 33a, the specifying clause ȩl teru ‘(lit.) as two (people)’

simply states that each of the subjects (Toki and Droteo) is sick.
And in 33b, the succession of specifying clauses ȩl di ngak ‘(lit.)
as only myself’ and ȩl ta ‘(lit.) as one (person)’ implies the par-
ticipation of only one person—namely, ak ‘I’—in the activity of
going to the beach. 9

15.7.4. SPECIFYING CLAUSES CONTAINING THE PERFECTIVE FORMS

OF mȩrkui

In 4.9 and 4.9.1–2 we saw that Palauan perfective verb forms

show different object pronoun suffixes depending on the
person and number of the object. Observe, therefore, the con-
trasting (past) perfective forms of mȩlim ‘drink’ in the following
sentences:

(34) a. Ak ngilȩlmii a biang.

‘I drank up the bottle of beer.’

364
 Palauan Reference Grammar

b. Ak ngilim a biang.
‘I drank up the bottles of beer.’

Since both 34a and 34b contain perfective verb forms, they
designate a completed past event involving the total con-
sumption of the object biang ‘beer’ (cf. 12.1). What differen-
tiates them is whether the object is interpreted as singular or
plural: thus, in 34a, the object pronoun suffix -ii on ngilȩlmii
‘drank it up’ refers to a singular object (‘bottle of beer’), while
in 34b, the Ø (zero) object pronoun suffix on ngilim ‘drank them
up’ implies a plural object (‘bottles of beer’).
Now, with 34a-b compare the following sentences:

(35) a. Ak ngilȩlmii a biang ȩ l rokir.

‘I drank up the whole bottle of beer.’

b. Ak ngilim a biang ȩl rokui.

‘I drank up all the bottles of beer.’

In 35a-b, the addition of the specifying clauses ȩl rokir and

ȩl rokui places special emphasis on the fact that the action of
drinking the beer expressed in the independent clause was per-
formed in an exhaustive manner—that is, in a manner resulting
in the total consumption of all the beer (one bottle or several
bottles) that was available. Most Palauan speakers feel there is
no difference in meaning, except for emphasis, between 34a and
35a, and between 34b and 35b.
As the pair of examples in 35 shows, the specifying clause ȩl
rokir must be used when a singular object is involved, while ȩl
rokui is required with a plural object. Therefore, ȩl rokir always
follows independent clauses containing 3rd pers. sg. object per-
fective verb forms (e.g. ngilȩlmii ‘drank it up’ of 35a), while
ȩl rokui follows independent clauses with 3rd pers. pl. (non-
human) object perfective verb forms (e.g. ngilim ‘drank them
up’ of 35b). Further pairs like 35a-b are listed below:

(36) a. Kȩ mla chuiȩuii a hong ȩr kau ȩl rokir?

‘Have you read all your book?’

b. Kȩ mla chȩmuiu a hong ȩr kau ȩl rokui?

‘Have you read all your books?’

365
 15 Dependent Clauses

(37) a. A Droteo a lilȩchȩsii a babier ȩl rokir.

‘Droteo wrote the letter completely.’

b. A Droteo a liluchȩs a babier ȩl rokui.

‘Droteo wrote all the letters completely.’

In examples 35–37 above, the specifying clauses ȩl rokir and

ȩl rokui refer to the object of the perfective verb in the in-
dependent clause. Thus, in 37a-b, for instance, the specifying
clauses tell us about one or more letters (babier) which were
written completely. It is also possible for the specifying clauses
ȩl rokir and ȩl rokui to refer to the subject of the verb in the
independent clause, as illustrated below:

(38) a. A bilek a mla mo dȩkimȩs ȩl rokir.

‘This piece of clothing of mine has gotten totally
soaked.’

b. A bilek a mla mo dȩkimȩs ȩl rokui.

‘All of my clothes have gotten completely wet.’

(39) a. A kȩlek a mla mȩka ȩl rokir.

‘My food has been totally eaten.’

b. A kȩlek a mla mȩka ȩl rokui.

‘All of my (various kinds of) food has been eaten
up.’

(40) A rȩchad ȩl mla ȩr a skoki ȩl rirebȩt a di mlad ȩl

rokui.
‘All of the people who were in the plane that
crashed died.’

The verb forms rokir and rokui observed in the specifying

clauses of the examples above are related structurally to the
present perfective forms of the verb mȩrkui ‘finish (completely,
entirely)’ (past: mirrȩkui). The principal perfective forms of
mȩrkui are listed below:

(41) Present Past

3rd pers. sg. object rokir rirȩkir

366
 Palauan Reference Grammar

3rd pers. pl. (non-human) rukui rirȩkui

object

The use of these forms is illustrated in the following sentences:

(42) a. Ak rirȩkir a urerek er a elii.

‘I completed my work/job yesterday.’

b. Ak rirȩkui a urerek er a elii.

‘I completed my (various) tasks yesterday.’

(43) a. Kȩ mla rokir a termpaper ȩr kau?

‘Have you finished your term paper?’

b. Kȩ mla rukui a termpaper ȩr kau?

‘Have your finished your term papers?’

(44) Ak mla rukui a klde ȩl hong.

‘I’ve completely finished (reading) three
books. 10

It appears that the specifying clauses ȩl rokir and ȩl rokui

observed in examples 35–40 above contain, respectively, the
3rd pers. sg. and pl. object present perfective forms of mȩrkui
‘finish (completely, entirely)’. Notice, however, that the 3rd
pers. pl. object present perfective form of mȩrkui is rukui, while
the corresponding form in the specifying clause is rokui. The
change of vowel in the first syllable of these forms is not clearly
understood.

15.7.4.1. Regional Variation in the Use of ȩl

rokir and ȩl rokui
Some regional variation is observed in the use of the specifying
clauses ȩl rokir and ȩl rokui. In Angaur and Peleliu (islands
south of Koror), speakers use these specifying clauses in a
manner exactly opposite from that described above: ȩl rokir
follows 3rd person plural object perfective verb forms, while ȩl
rokui follows 3rd person singular object perfective verb forms.
The following are typical examples of this regional variation:

(45) a. KOROR, BABELDAOB

367
 15 Dependent Clauses

Aki mla kma a kall ȩl rokui.

‘We’ve eaten up all the (various kinds of)
food.’

b. ANGAUR, PELELIU

Aki mla kma a kall ȩl rokir.

‘We’ve eaten up all the (various kinds of)
food.’

(46) a. KOROR, BABELDAOB

Ak mla kolii a blauang ȩl rokir.

‘I’ve eaten up the whole loaf of bread.’

b. ANGAUR, PELELIU

Ak mla kolii a blauang ȩl rokui.

‘I’ve eaten up the whole loaf of bread.’

15.7.5. SPECIFYING CLAUSES DESIGNATING PERIODS OF TIME

A specifying clause containing an expression denoting a period
(or duration) of time can be used to indicate how long the ac-
tivity or state of the preceding independent clause continues.
Some typical examples are given below:

(47) a. A Droteo a mlo ȩr a Guam ȩl {ede ȩl klȩbȩsei/ta ȩl buil}.

‘Droteo went to Guam for {three days/one month}

b. A Toki a mla ȩr a Merikel ȩ ta ȩl rak.

‘Toki was in America for one year.’

c. Ak milsuub ȩr a Siabal ȩl eru ȩl buil.

‘I studied in Japan for two months.’

d. A Toki a milȩchiuaiu ȩl ta ȩl sils.

‘Toki slept the whole day.’

368
 Palauan Reference Grammar

If a period of time is expressed in terms of a beginning point

and an end (or termination) point, we use a temporal phrase
to indicate the former (cf. 14.6) and a specifying clause of the
form ȩl + mo + temporal phrase ‘until’ to indicate the latter. Ob-
serve the sentences below:

(48) a. Ak mla ȩr a beluak ȩ r a kot ȩ l ureor ȩl mo (ȩr a)

sabadong.
‘I was in my village from Monday until Saturday.’

b. A Toki a mȩngȩtmokl ȩr a blik ȩ r a etiu ȩl klok ȩl mo (ȩr

a) tȩruich ȩl klok.
‘Toki cleans my house from nine o’clock until ten
o’clock.’

In the sentences above, the italicized temporal phrases des-

ignate the “source” in time—i.e., the time point when some ac-
tivity or state begins. In the directly following specifying clause,
the directional verb mo ‘go’ is used to indicate “movement
across time” from an earlier time point to a later one (cf. our
discussion at the end of 13.4), and the ȩ r a portion of the tem-
poral phrase can be optionally deleted.
When the speaker is interested only in the termination point
of some activity or state, he can simply use a specifying clause
of the form ȩl + mo + temporal phrase ‘until’, as in the sen-
tences below:

(49) a. Ak mo kie ȩr tia ȩl mo (ȩr a) sabadong.

‘I’ll be (living) here until Saturday.’

b. Kȩ milȩkar ȩl mo (ȩr a) tela ȩl klok ȩr a

kȩsus?
‘Until what time were you up last night?’

c. Kau a chȩrrȩuek ȩl mo ȩr a kodall.

‘You’re my enemy until death.’

15.7.6. SPECIFYING CLAUSES IN SENTENCES DESIGNATING MANNER

Specifying clauses are also frequently used in sentences which
describe the manner or way in which some activity is done.
Usually, the independent clause contains a state verb desig-

369
 15 Dependent Clauses

nating a particular quality, and the following specifying clause

identifies the action or activity which is characterized by this
quality. Observe the following examples:

(50) a. A Santos a dachȩlbai ȩl mȩlasȩch a mlai.

‘Santos is skilled in carving canoes.’

b. A Droteo a {bȩlȩrurt / bȩkȩtimȩl} ȩl mȩngikai.

‘Droteo is a {fast / slow} swimmer.’

c. A Droteo a mle {mȩrechȩd / meoud} ȩl mo ȩr a blil a

Toki.
‘Droteo arrived {early / late} at Toki’s house.’

d. A Droteo a mle {klou / kekȩre} a ngȩrel ȩ l mȩngitakl.

‘Droteo was singing {loudly / softly}.’

e. A Toki a kmal mle mȩringȩl ȩ l oureor.

‘Toki was working very hard.’

f. A Toki a mle ungil ȩl ngotȩchii a mlai.

‘Toki did a good job cleaning the car.’

In 50a, the state verb dachȩlbai ‘skilled’ names a quality of the

subject (Santos), and the following italicized specifying clause
designates the activity to which that quality applies. Similarly,
in 50d, the expressions mle klou a ngȩrel ‘(lit.) his voice was
big’ and mle kekȩre a ngȩrel ‘(lit.) his voice was small’ describe
qualities which characterize the action of singing denoted by
the specifying clause. All of the remaining examples of 50 can
be explained in a parallel way.
Often, speakers can switch the order of clauses in sentences
like those of 50, giving the following, whose meanings are
essentially the same:

(50c’) A Droteo a milȩngitakl ȩl (mle) {klou/kekȩre} a

ngȩrel.
‘Droteo was singing {loudly/softly}.’

(50d’) A Toki a ulureor ȩl kmal (mle) mȩringȩl.

‘Toki was working very hard.’

370
 Palauan Reference Grammar

Notice the optionality of the past tense auxiliary word mle pre-
ceding the state verb of the specifying clause.

15.7.7. SPECIFYING CLAUSES FOLLOWING SPECIAL VERBS

There are a small number of special verbs in Palauan which
always must be followed by a specifying clause. These verbs are
unusual in that they denote rather abstract concepts which we
would not ordinarily expect to be expressed by verb forms. Ob-
serve, for example, the sentence below:

(51) Ak blȩchoel ȩl mȩruul a kȩlel a Droteo.

‘I always prepare Droteo’s food.’

Even though it translates as English ‘always’, the word blȩchoel

of 51 seems to be a verb because it is preceded by a subject (ak
‘I’) and followed by a specifying clause. 11 This specifying clause
(italicized) identifies the activity which the subject is always
pursuing. Further evidence that blȩchoel is a verb)—more
specifically, a state verb—is seen in the fact that it takes the
auxiliary word mle in the past tense (cf. 5.1.3), as the following
examples show:

(52) a. Ak mle blȩchoel ȩl mo ȩr a che ȩl obȩngkel a

dȩmak.
‘I always used to go fishing with my father.’

b. A Droteo a mle blȩchoel ȩl mȩleng a udoud.

‘Droteo always used to borrow money.’

Some further sentences containing blȩchoel are given below:

(53) a. Ng blȩchoel ȩl ngmasȩch a rȩngum.

‘You’re always getting angry.’

b. A Droteo mȩ a Toki a blȩchoel ȩl kaungȩroel.

‘Droteo and Toki are always quarreling with each
other.’

c. A rȩmȩsaik a blȩchoel ȩl diak a kȩlir.

‘Those who are lazy are always running out of
food.’

371
 15 Dependent Clauses

d. A John a blȩchoel ȩl diak losuub.

‘John never studies.’

In the examples below, we illustrate the use of several other

unusual verbs which resemble blȩchoel ‘always’:

(54) (di) tȩlkib ‘(only) a little’:

a. Ak di tȩlkib ȩl sȩbȩchek ȩl
mȩngȩdȩchȩduch.
‘I can only talk a little.’

b. Ak mle tȩlkib ȩl smechȩr er a elii.

‘I was a bit sick yesterday.’

(55) dirrek ‘also’:

a. Ak dirrek ȩl mȩsuub a tȩkoi ȩr a

Siabal.
‘I’m also studying Japanese.’

b. A Droteo a dirrek ȩl sensei.

‘Droteo is also a teacher.’

c. A Toki ng dirrek ȩl mong?

‘Is Toki going too?’

(56) ko ‘just’:

a. A Toki a ko ȩl rongȩsa a chais.

‘Toki has just heard the news.’

b. Ak ko ȩl mȩsubang.
‘I’ve just gotten (a chance) to study.’

In 56a-b, the specifying clause following ko ‘just’ contains an in-

choative verb form (cf. 11.12.3).

372
 16 Object Clauses

16ObjectClauses
16.1. STRUCTURE AND FUNCTION OF OBJECT
CLAUSES
In 5.1.1 we pointed out that Palauan transitive verbs name ac-
tions which involve a doer and a receiver. The receiver—the
person, animal, or thing affected by the action in
question—appears as the sentence object following the tran-
sitive verb. You should have no trouble identifying the noun
phrases which function as sentence objects in the examples
below:

(1) a. Ak milsa a Droteo ȩr a party.

‘I saw Droteo at the party.’

b. A buik a mo omȩka ȩr a bilis.

‘The boy will feed the dog.’

c. Ak lilengir a mlil a Droteo.

‘I borrowed Droteo’s car.’

d. A John a milȩngitakl a chȩlitakl ȩr a Siabal.

‘John was singing Japanese songs.’

The transitive verbs in 1 can only take concrete nouns as ob-

jects; thus, all of the objects in the examples above can be easily
perceived by one or more of our five senses (sight, hearing,
touch, taste, and smell).
There is an important class of Palauan transitive verbs
which can take both concrete objects and objects describing ac-
tions or activities. Compare the a- and b-sentences in the ex-
amples below:

(2) a. Ak mla mo mȩrek ȩr a subȩlek.

‘I have finished my homework.’

b. Ak mla mo mȩrek ȩl mȩruul a kall.

‘I have finished preparing the food.’

373
 16 Object Clauses

(3) a. A Droteo a milsuub a tȩkoi ȩr a Merikel.

‘Droteo was studying English.’

b. A Droteo a milsuub ȩl mȩruul a mlai ȩr a

demal.
‘Droteo learned how to make canoes from
his father.’

While the objects of mo mȩrek ‘finish, stop’ and mȩsuub ‘study,

learn’ in the a-sentences above are concrete (subȩlek ‘my home-
work’ and tȩkoi ȩr a Merikel ‘English’), the objects of these
same verbs in the b-sentences designate certain actions or ac-
tivities—preparing food, making canoes—rather than things. In
the a-sentences, the objects of mo mȩrek ‘finish, stop’ and
mȩsuub ‘study, learn’ are simply noun phrases, preceded by
the word a (cf. 2.6), and, if appropriate, by the specifying word
ȩr (cf. 2.7). By contrast, the italicized objects in the b-sentences
have the structure of dependent clauses, as we will see below.
In discussing the grammatical characteristics of Palauan de-
pendent clauses, we mentioned in 15.1 that dependent clauses
are incomplete or deficient in two important respects. First
of all, dependent clauses never contain an overtly-expressed
subject; and second, the verb of the dependent clause can be
in the present tense, even when the verb of the preceding in-
dependent clause is in some past tense and the whole sentence
therefore designates a past action or event. Both of these fea-
tures are observed in the italicized portions of 2b and 3b above:
there is no subject following ȩl where we would expect one,
and the verb directly following ȩl is in the present tense, even
though both sentences describe past activities. In spite of the
fact that the italicized portions of 2b and 3b therefore lack sub-
jects and marking for the past tense, speakers of Palauan nev-
ertheless have no difficulty interpreting them. Thus, just as in
the case of the various dependent clauses described in chap.
15, Palauan speakers automatically know that the understood
subject of mȩruul ‘make, prepare’ in the italicized portions of
2b and 3b is identical to that of the preceding clause. In 2b,
for example, it goes without saying that the person who pre-
pared the food (i.e., the understood subject of mȩruul a kall
‘prepare the food’) is the same person who recently finished
that very same activity—namely, the subject of the first clause
ak ‘I’. Furthermore, speakers know that the activities described
by the present tense forms of mȩruul ‘make, prepare’ in the

374
 Palauan Reference Grammar

italicized portions of 2b and 3b really took place in the past, at

time points identical to those of mla mo mȩrek ‘has finished’ and
milsuub ‘learned’.
From the discussion above, we can see that the italicized
portions of 2b and 3b should be analyzed as dependent
clauses since their complete interpretation depends on in-
formation about the subject and the tense which is found in
the preceding independent clause. In order to distinguish them
from the many other types of dependent clauses examined in
chap. 15, we will identify the dependent clauses of 2b and 3b as
object clauses. An object clause is therefore a subtype of de-
pendent clause which is used in “object position” immediately
following certain transitive verbs like mo mȩrek ‘finish, stop’
and mȩsuub ‘study, learn’. If we compare the a- and b-sentences
of 2-3 above, we can see that an object clause designating an
action or activity can substitute for or replace a “simple” noun
phrase object denoting something concrete.
One further grammatical characteristic of object clauses
should be mentioned here. As we saw in chap. 15, many Palauan
speakers can use past tense verb forms in purpose clauses,
instrument clauses, and the like. By contrast, object clauses
cannot contain past tense verb forms under any circumstances.
Therefore, a sentence like the following is impossible (cf. 2b):

(4) *Ak mla mo mȩrek ȩl mirruul a kall.

16.2. OBJECT CLAUSES FOLLOWING OMUCHȨL

AND MO MȨREK
In this and the following sections, we will examine some of the
most commonly used verbs which can be followed by object
clauses. Most of these verbs also take concrete objects naming
persons or things. In this section we will show how object
clauses are used following the transitive verbs omuchȩl ‘begin’
and mo mȩrek ‘finish, stop’. Sentences containing the sequence
omuchȩl/ mo mȩrek + object clause simply tell us that the
subject begins or finishes a particular activity; the activity
which is begun or finished is of course expressed by the words
of the object clause.
In the examples below, the object clause following omuchȩl
‘begin’ has been italicized:

375
 16 Object Clauses

(5) a. Ak mo omuchȩl ȩ l mȩngiis ȩr a klioklȩr a klukuk.

‘I’m going to begin digging the hole tomorrow.’

b. A Droteo a ulȩmuchȩl ȩl mȩsuub a tȩkoi ȩr a Merikel er

se ȩr a mȩkȩmad.
‘Droteo began to study English during the war.’

As the examples above show, an object clause can be followed

by a relational phrase: thus, the temporal phrases ȩr a klukuk
‘tomorrow’ and er se ȩ r a mȩkȩmad ‘during the war’ (cf. 14.6)
designate the time at which some activity will begin or did
begin.
The transitive verb omuchȩl ‘begin’ can also have a concrete
noun phrase as object, as in the sentences below:

(6) a. Kȩdȩ mo omuchȩl ȩr a blai ȩr a klukuk.

‘We’ll begin (to build) the house tomorrow.’

b. Kȩ mo omuchȩl ȩr a urerem ȩr oingarang?

‘When are you going to begin your work?’

When omuchȩl is used as a transitive verb, as in the examples

of 5 and 6 above, its subject must always be animate (or
living)—usually a human being. It is also possible for omuchȩl to
be used as an intransitive verb, in which case its subject will
be inanimate (or non-living). The intransitive use of omuchȩl is
observed in the sentences below:

(7) a. A meeting a mo omuchȩl ȩr a euid ȩl klok.

‘The meeting will begin at seven o’clock.’

b. A kȩrrȩkȩriil ȩr a Droteo a ulȩmuchȩl er a

elii.
‘Droteo’s trial began yesterday.’

Since omuchȩl is an intransitive verb in the sentences of 7,

these sentences of course do not contain any object noun
phrases.
As we saw in 13.6, the special verbal expression mo mȩrek
‘finish, stop’ consists of the directional verb mo ‘go’ and mȩrek,
a difficult-to-analyze form of the verb mȩrkui ‘finish.’ The se-

376
 Palauan Reference Grammar

quence mo mȩrek always appears as a unit, and it functions as a

transitive verb. In the sentences below, mo mȩrek ‘finish, stop’
is followed by an object clause, which has been italicized:

(8) a. Ak mlo mȩrek ȩl rȩmurt ȩr a eai ȩl klok.

‘I stopped running at eight o’clock.’

b. Ak mlo mȩrek ȩl mȩsuub a tȩkoi ȩr a Merikel ȩr tia ȩl mlo

mȩrek ȩl rak. 1
‘I finished studying English last year.’

c. Ak mla mo mȩrek ȩl mȩruul ȩr a blai.

‘I’ve finished working on the house (for today).’

d. A Droteo a mla mo mȩrek ȩl bȩchiil.

‘Droteo is no longer married.’ 2

e. Kȩ mla mo mȩrek ȩl omȩngur?

‘Have you finished eating?’

f. Kȩ mla mo mȩrek ȩl mȩlamȩch a dȩkool?

‘Have you finished smoking your cigarette?’

g. A Droteo a mo mȩrek ȩl mȩngiis ȩr a klioklȩr a klukuk.

‘Droteo will finish digging the hole tomorrow.’

The examples in 8 show us many interesting facts about the

form and meaning of mo mȩrek ‘finish, stop’. As we saw in 13.6,
the form of mȩrek does not change when this verbal expression
is put in the past tense. Instead, the past tense marker -l- is
infixed into the directional verb mo to give mlo mȩrek, as ob-
served in 8a–b. Examples 8c–f show that the auxiliary word
mla (cf. 5.3.2.1) is simply placed before mo mȩrek to indicate
recent past time; and in 8g we see that mo mȩrek, which con-
tains the present tense form of the directional verb mo, can be
used to denote a future event (cf. 13.4), as evidenced by the
accompanying temporal phrase ȩr a klukuk ‘tomorrow’.
As the English equivalents for the sentences of 8 indicate,
the sequence mo mȩrek + object clause always refers to the
stopping or finishing of something on a particular instance or
occasion. In 8f, for example, the speaker is asking someone on

377
 16 Object Clauses

a particular occasion (say, before a class is to begin) whether or

not he has finished smoking. Such a question is quite different
in meaning from the following:

(9) Kȩ mla choitii a omȩlamȩch ȩl dȩkool?

‘Have you quit/given up smoking
cigarettes?’

As opposed to 8f, 9 asks the person addressed whether he has

quit or stopped the habitual action of smoking. In 9, the object
of mla choitii ‘has quit’ (cf. imperfective mȩngoit ‘quit, throw
away’) is a noun phrase containing the derived action noun
omȩ lamȩch ‘(action of) smoking’ (cf. 8.6); the structure of sen-
tences like 9 will be examined further in 17.8. 3
The verbal expression mo mȩrek ‘finish, stop’ need not be
followed by an object clause as in the examples of 8 above; thus,
as the examples below illustrate, mo mȩrek can also take a con-
crete object (cf. 2a above):

(10) a. Kȩ mla mo mȩrek a kall?

‘Have you finished (preparing) the food?’

b. Ke mla mo mȩrek ȩr a skuul?

‘Have you finished your schooling?’

c. Ak mla mo mȩrek ȩr a urerek.

‘I’ve finished my work/job (for the day).’

When mo mȩrek is used as a transitive verb, as in 8 and 10

above, its subject must be animate (usually human). But when
mo mȩrek is used intransitively, as in the following sentences,
its subject will be inanimate:

(11) a. A meeting a mo mȩrek ȩr a etiu ȩl klok.

‘The meeting will end at nine o’clock.’

b. A chull a mla mo mȩrek.

‘The rainy season has ended.’

378
 Palauan Reference Grammar

16.3. OBJECT CLAUSES FOLLOWING TRANSITIVE

STATE VERBS
In 7.4 we noted that the small class of Palauan transitive
state verbs can be identified by the following two features:
first, like all other transitive verbs, they can take objects; and
second, like all other state verbs, they have past tense forms
derived with the auxiliary mle ‘was, were’. Two transitive state
verbs—mȩduch ‘know how (to), be skilled at’ and mȩtitur ‘not
know how (to), not be capable of’—can be followed by concrete
objects as well as object clauses. In the sentences below, these
verbs appear followed by concrete objects:

(12) a. Ak mȩduch ȩr a ochur.

‘I’m good at math.’

b. Ak mȩtitur ȩr a misil ȩr a sidosia.

‘I don’t know anything about (fixing) car
motors.’

When mȩduch ‘know how (to), be skilled at’ and mȩtitur ‘not
know how (to), not be capable of’ take object clauses, they ex-
press the subject’s ability or lack of ability, respectively, to do
some activity or task. The following sentences illustrate these
verbs in both the present tense and past tense followed by
object clauses:

(13) a. Ak mȩduch ȩl omȩkall a sidosia.

‘I know how to drive a car.’

b. A Droteo a mle mȩduch ȩl mȩlȩkoi a tȩkoi ȩr a Siabal e

ng di ng mla obes.
‘Droteo used to know how to speak Japanese, but he has
forgotten.’

(14) a. A John a mȩtitur ȩl mȩngikai.

‘John doesn’t know how to swim.’

b. Ak mle mȩtitur ȩl mȩlȩkoi a tȩkoi ȩr a Siabal er se ȩr a

taem ȩr a mȩkȩmad.
‘I didn’t know how to speak Japanese at the time of the
war.’

379
 16 Object Clauses

In 13.5 we saw that the directional verb mo ‘go’ can be

used as an auxiliary word preceding state verbs to designate a
change of state. Since mȩduch ‘know how (to), be skilled at’
and mȩtitur ‘not know how (to), not be capable of’ are state
verbs, they can also occur with mo to denote a change of state.
Note the following examples:

(15) a. Ak mle mȩtitur ȩl mȩngikai e ng di ak mla mo mȩduch.

‘I used to be unable to swim, but (now) I’ve learned
how.’

b. A rubak a mle mȩduch ȩl mȩlȩkoi a tȩkoi ȩr a Ruk e ng

di ng mla mo mȩtitur.
‘The old man used to know how to speak Trukese, but
(now) he’s no longer able to.’

16.4. FURTHER EXAMPLES OF OBJECT CLAUSES

In this section, we will look at two more verbs which allow con-
crete objects as well as object clauses. Since there is nothing
unusual about the grammatical behavior of these verbs, we will
simply list examples without comment.
In 16 below, the transitive verb mȩlasȩm ‘try’ is followed by
a concrete object, while in 17 it takes an object clause:

(16) a. Kȩ mla mȩlasȩm ȩr a sasimi?

‘Have you tried out/tasted the sashimi?’

b. Kȩ mȩlasȩm ȩr ngak?
‘Are you challenging me?’

c. Kȩ mla mȩlasȩm ȩr a Droteo?

‘Have you tried (to ask/consult) Droteo?’

d. Kȩ mla chesȩmii 4 a mlim?

‘Have you tried out/checked your car?’

(17) a. Ak millasȩm ȩl mȩnga ȩr a ngikȩl.

‘I tried to eat the fish.’

b. Kȩ mla mȩlasȩm ȩl omȩkall a sidosia?

‘Have you ever tried driving a car?’

380
 Palauan Reference Grammar

c. Ak mla mȩlasȩm ȩl mȩruul ȩr a subȩlek, e ng di ng kmal

mȩringȩl.
‘I’ve tried to do my homework, but it’s very difficult.’

d. Kȩ mla mȩlasȩm ȩl mȩlȩkoi ȩr a John?

‘Have you tried talking to John?’

The transitive verb mȩsuub ‘study, learn’ is used with a con-

crete object in 18 below, and with an object clause in 19 (cf. 3a
vs. 3b above):

(18) a. Ak mo mȩsuub ȩr a reksi er a Belau.

‘I’m going to study Palauan history.’

b. A Satsko ng milsuub a ngarang?

‘What was Satsko studying?’

(19) a. Ak milsuub ȩl mȩluchȩs a tȩkoi ȩr a Siabal er se ȩr a

mȩkȩmad.
‘I learned how to write Japanese during the war.’

b. Kȩ milsuub ȩl omȩkall a sidosia er oingarang?

‘When did you learn to drive a car?’

16.5. SENTENCES WITH TWO OBJECTS

So far we have only seen object clauses whose unexpressed
subject is understood to be identical to that of preceding tran-
sitive verbs such as omuchȩl ‘begin’, mo mȩrek ‘finish, stop’,
mȩduch ‘know how (to), be skilled at’, mȩtitur ‘not know how
(to), not be capable of’, mȩlasȩm ‘try’, and mȩsuub ‘study,
learn’. There are some cases, however, in which the unex-
pressed subject of the object clause is understood differently, as
in the example below:

(20) A Toki a milȩngȩtakl ȩr a Helen ȩl mo ȩr a

bulis.
‘Toki persuaded Helen to go to the police with
her.’

381
 16 Object Clauses

In 20, the subject (Toki) influenced some other person (Helen)

to do a particular action (i.e., go to the police). In other words,
the action of going to the police was carried out by the person
persuaded (Helen) rather than the person doing the persuading
(Toki). Therefore, the unexpressed subject of the object clause
ȩl mo ȩr a bulis ‘go to the police’ is not identical to Toki, the
subject of milȩngȩtakl ‘persuaded’, but rather to Helen, the
noun immediately following milȩngȩtakl. The verb milȩngȩtakl
in 20 thus appears to be followed by a succession of two objects,
the first one naming the person influenced or affected by the
persuasion and the second one—an object clause—describing
the action pursued as a result of the persuasion.
In the sentences below, we observe an over-all structure
identical to that of 20 above. This “two object” structure is
typical with the verbs olȩngȩseu ‘help’, olisȩchakl ‘teach’, and
oldurȩch ‘tell, ask’. In other words, these verbs not only involve
a person who is helped, taught, or asked (the first object), but
also some kind of an activity which the person is helped, taught,
or asked to do (the second object). For purposes of clarity, we
have italicized the first object in the examples below; the person
referred to by this object is, of course, the understood subject
of the following object clause. Note, in addition, that the first
object will be expressed by an object pronoun suffix (cf. 4.9) if
a perfective form of olȩngȩseu, olisȩchakl, or oldurȩch is used:

(21) a. Ak ullȩngȩseu ȩr a Toki ȩl mȩruul a subȩlel.

‘I helped Toki do her homework.’

b. Ng sȩbȩchem ȩl ngosukak 5 ȩl mo chosbȩrbȩrii a blik?

‘Can you help me paint my house?’

c. Ak ngilsutȩrir a rȩsȩchȩlikȩl mȩngȩtmokl ȩr a blai.

‘I helped my friends clean the house.’

d. A rubak a ullisȩchakl ȩr a Droteo ȩl mȩlasȩch a mlai.

‘The old man taught Droteo how to carve canoes.’

e. A rȩchad ȩr a Arabia a uldȩrchȩtȩrir 6 a rȩchad ȩ r a

Siabal ȩl mȩkodȩtȩrir a rȩchad ȩr a skojo.
‘The Arabs told the Japanese to kill the people at the
airport.’

382
 Palauan Reference Grammar

f. Ak uldȩrchii a John ȩl mȩkodir a Toki.

‘I told John to kill Toki.’

g. A sensei a uldȩrchak ȩl mo ȩr a Guam.

‘The teacher told me to go to Guam.’

383
 17 Processes of Sentence
Formation: Subject Shifting
and Preposing of Possessor

17ProcessesofSentenceFormation:
17.1. FOUR SPECIAL POSSESSED NOUNS
We have already had occasion to mention a special group of
obligatorily possessed nouns (cf. 3.5) which refer to certain
basic ideas such as liking, disliking, ability, and obligation.
These nouns are used in sentences like the following, which we
repeat from 3.11:

(1) a. Ng soak a biang.

‘I like beer.’

b. Ng chȩtil a rrom.
‘He dislikes liquor.’

c. Ng sȩbȩchir ȩl mong?
‘Can they go?’

d. Ng kirem ȩl mong?
‘Do you have to go?’

Even though the English equivalents for the above examples

contain verbs (like, dislike) or verbal expressions (can go, have
to go), we should not be misled into thinking that the corre-
sponding Palauan sentences necessarily contain verbs to ex-
press the same ideas. For example, sentences la and 1b do not
contain any verbs at all, but instead have the possessed nouns
soak ‘my liking’ and chȩtil ‘his disliking’ followed by concrete
nouns such as biang ‘beer’ and rrom ‘liquor’. Thus, these sen-
tences seem to mean something like ‘My liking is beer’ and ‘His
disliking is liquor’, respectively. 1
We know that words like soak ‘my liking’, chȩtil ‘his dis-
liking’, sȩbȩchir ‘their ability’, and kirem ‘your obligation’ of 1
above must be nouns because their form varies according to
whose liking, disliking, ability, or obligation is involved. These

384
 Palauan Reference Grammar

four words are to be classified as obligatorily possessed

nouns because they each must take one of the sets of pos-
sessor suffixes described in 3.3. Since these words occur so
frequently in Palauan sentences, we shall list their possessed
forms below:

(2) Possessor Suffix Noun of Liking Noun of Disliking

1st pers sg soak chȩtik

2nd pers sg soam chȩtim
3rd pers sg soal chȩtil
1st pers pl incl soad chȩtid
1st pers pl excl somam chȩtimam
2nd pers pl somiu chȩtimiu
3rd pers (hum) pl sorir chȩtirir

Noun of Ability Noun of Obligation

1st pers sg sȩbȩchek kirek

2nd pers sg sȩbȩchem kirem
3rd pers sg sȩbȩchel kirel
1st pers pl incl sȩbȩched kired
1st pers pl excl sȩbȩcham kiram
2nd pers pl sȩbȩchiu kiriu
3rd pers (hum) pl sȩbȩchir kirir

The forms given above show that the noun of ability and the
noun of obligation have possessor suffixes belonging to the e-
set, while the noun of liking and the noun of disliking take pos-
sessor suffixes belonging to the a-set and the i-set, respectively.
The only irregularity we observe is in certain forms of the noun
of disliking: in the “plural possessor” forms chȩtimam, chȩtimiu,
and chȩtirir, we note the unexpected insertion of i before the
consonant-initial suffixes -mam, -miu and -rir (cf. 3.3, ex. 4).
Before we can adequately explain the grammatical structure
of the sentences in 1, we must take a preliminary look at an
important Palauan sentence type—namely, the equational sen
tence. As we will see in 18.6, an equational sentence is one in
which two noun phrases are equated with each other. In the
present tense, Palauan equational sentences contain no verb at
all, but merely consist of a subject noun phrase followed by an-
other noun phrase. Note the examples below:

385
 17 Processes of Sentence Formation:

(3) a. Ak ngalȩk ȩr a skuul.

‘I’m a student.’

b. A Droteo a sensei.
‘Droteo’s a teacher.’

The sentences of 3 simply make a statement of the form “A is B”:

in other words, in 3a the subject noun phrase ak ‘I’ is asserted
to belong to the class of individuals designated by the second
noun phrase ngalȩk ȩr a skuul 2 ‘student’, and in 3b the subject
noun phrase Droteo is said to belong to the class of individuals
referred to by the second noun phrase sensei ‘teacher’. In the
past tense, Palauan equational sentences contain the auxiliary
word mle ‘was, were’, which joins the two noun phrases. Thus,
with 3a–b, compare the following examples:

(4) a. Ak mle ngalȩk ȩr a skuul.

‘I was/used to be a student.’

b. A Droteo a mle sensei.

‘Droteo was/used to be a
teacher.’

17.2. SUBJECT SHIFTING

In order to understand the structure of sentences 1a–b above,
we shall first propose that they are basically equational sen-
tences. In other words, sentences 1a–b are derived by a certain
grammatical process from sentences like the following:

(1’) a. A biang a soak.

‘Beer is what I like.’

b. A rrom a chȩtil.
‘Liquor is what he dislikes.’

The sentences in 1’ are equational sentences containing two

noun phrases; in each case, the subject noun phrase (biang
‘beer’ or rrom ‘liquor’) is being equated with one of the special
possessed nouns discussed in 17.1 above. Therefore, the in-
terpretation of 1’a–b follows the “A is B” pattern, and the two

386
 Palauan Reference Grammar

sentences literally mean something like ‘Beer is my liking’ and

‘Liquor is his disliking’ (which of course are not good English
sentences).
Now, the equational sentences in 1’ are used relatively
rarely by Palauan speakers. Such sentences would only be ut-
tered if the speaker wishes to emphasize or single out a par-
ticular thing as the thing he likes or dislikes. Thus, 1’a implies,
for example, that it is beer and only beer (out of a choice of
several different beverages) that the speaker has developed
a taste for. For some reason, the equational sentences in 1’
must usually undergo a modification in form before they can be
spoken as natural, fully acceptable sentences of Palauan. This
modification in form is brought about by the very important
process of subject shifting, which we have already touched
upon in 4.7. In order to see how subject shifting works, let us
look at sentences 1a–b together with their respective “source”
sentences 1’a–b, as in the following scheme:

(5) Source Sentence Resulting Sentence

a. A biang a soak. → Ng soak a biang.
‘I like beer.’

b. A rrom a chȩtil. → Ng chȩtil a rrom.

‘He dislikes liquor.’

Depending on whether we are looking at the process of change

in 5a–b from the viewpoint of the source sentence or the re-
sulting sentence, our interpretation of the arrow notation will
be expressed differently. Thus, if we are focusing our attention
on the source sentence, the arrow means that the source sen-
tence is “changed into” or “transformed into” the resulting sen-
tence. On the other hand, if our discussion is from the viewpoint
of the resulting sentence, the arrow is interpreted to mean that
the resulting sentence “results from” or “is derived from” the
source sentence.
In 5 above, the resulting sentences are derived from the
source sentences by the process of subject shifting. This
process shifts the subject noun phrase of an (equational) source
sentence to the right of the second noun phrase. Thus, in the
resulting sentences of 5, biang ‘beer’ and rrom ‘liquor’—the
subject noun phrases in the source sentences—have come to
appear to the right of the possessed nouns soak ‘my liking’ and

387
 17 Processes of Sentence Formation:

chȩtil ‘his disliking’. When a sentence subject gets shifted in

this way, a trace of it remains in its original position in the
form of a non-emphatic pronoun (cf. 4.2 and 4.7). Because
the shifted noun phrase subjects biang ‘beer’ and rrom ‘liquor’
in the examples of 5 are 3rd pers. sg. nouns, the pronominal
trace which they leave behind is the 3rd pers. sg. non-emphatic
pronoun ng. Thus, the pronominal traces which occupy the
original subject position in the resulting sentences of 5 agree
with the shifted noun phrase subjects biang ‘beer’ and rrom
‘liquor’.
The analysis given above may seem unconvincing and un-
necessary until we see that the process of subject shifting is
quite widespread in Palauan and must therefore be included as
part of any adequate description of the language. To illustrate
this point, let us repeat our discussion of examples 23–24 in 4.7,
which are given as 6–7 below:

(6) a. A Droteo a mla mei. ‘Droteo has come.’

b. Ng mla me a Droteo.

(7) a. A ralm a mȩkȩlȩkolt. ‘The water is cold.’

b. Ng mȩkȩlȩkolt a ralm.

To Palauan speakers, the a- and b-sentences in 6–7 above are

equally acceptable and natural. While the a-sentences have
their subjects at the beginning, as we would expect, in the b-
sentences these same subjects have been shifted to the right of
the verb phrases (mla me ‘has come’ and mȩkȩlȩkolt ‘cold’). In
other words, the b-sentences are derived from the a-sentences
by the process of subject shifting, and since the shifted sub-
jects (Droteo in 6 and ralm ‘water’ in 7) are 3rd pers. sg. nouns,
they leave the pronominal trace ng in their original position. If
the shifted subject is human plural, then the pronominal trace
which it leaves behind is the 3rd pers. human pl. non-emphatic
pronoun tȩ. Thus, with 6, compare the examples below:

(8) a. A rȩsȩchȩlim a mla ‘Your friends have

mei. come.’
b. Tȩ mla me a
rȩsȩchȩlim.

388
 Palauan Reference Grammar

In the b-sentences of 6–8, the verb phrase is not introduced

by the word a because it has come to be preceded by a non-
emphatic pronoun—namely, the pronominal traces ng or tȩ (cf.
4.2).
Practically every Palauan sentence of the form subject noun
phrase + intransitive verb phrase (+ relational phrase) or
subject noun phrase + transitive verb phrase (+ object noun
phrase) (+ relational phrase) (cf. 5.2) can be transformed by
the process of subject shifting. This change in form does not
result in any change of meaning for many speakers; therefore,
we will find that the a-and b-sentences of 6–8 above are often
used interchangeably. Some speakers, however, use the a-vs.
b-sentences in rather different situations. Thus, though both
members of 8 say that the hearer’s friends have arrived, their
implications are different: 8a, with “normal” word order, is used
to express new or unexpected information and therefore im-
plies that the speaker had no advance knowledge that the vis-
itors would come, while 8b, with “shifted” word order, seems
to confirm an event which the speaker was waiting for or ex-
pecting.
To summarize what we have said above, the process of sub-
ject shifting will account for the following two types of deriva-
tions:

(9) Source Sentence Resulting Sentence

a. A biang a soak. → Ng soak a biang.

‘I like beer.’

b. A Droteo a mla mei. → Ng mla me a Droteo.

‘Droteo has come.’

The processes of change in 9a and 9b are identical in that

the subject of the source sentence is shifted and a pronominal
trace is left in its place. What differentiates the two examples,
however, is the following. First, in 9a, which is an equational
sentence, the subject is shifted to the right of the second noun
phrase, while in 9b, which is an intransitive sentence, the
subject is moved to the right of the verb phrase. Second, in 9a,
subject shifting is nearly obliga tory, since the source sentence
is used only rarely, while the resulting sentence (with shifted
subject) is the usual way of expressing the idea involved. By

389
 17 Processes of Sentence Formation:

contrast, in 9b, both the source sentence and the resulting sen-
tence are common and acceptable; therefore, the application of
subject shifting is just optional. Why subject shifting is nearly
obligatory in 9a but optional in 9b seems to be a matter of style
which we cannot predict or explain. Rather, all we can do is
describe the situation by saying that equational sentences con-
taining possessed nouns like soak ‘my liking’ and chȩtil ‘his dis-
liking’ must usually undergo the process of subject shifting. This
accounts for the fact that such sentences nearly always have ng
(a pronominal trace) in initial position and another noun (the
shifted subject) following the possessed noun. As we will see
below and in later chapters, there are other Palauan sentence
types which resemble sentences containing soak, chȩtil, etc. in
that subject shifting is either obligatory or, at least, preferred.

17.3. PREPOSING OF POSSESSOR

In the examples below, we observe some further instances of
subject shifting. In each case, the shifted subject (italicized)
happens to be a noun phrase of possession (cf. 3.7) with a
specifically-mentioned third person possessor:

(10) Source Sentence Resulting Sentence

a. A chimal a Droteo a → Ng mȩringȩl a chimal a

mȩringȩl. Droteo.
‘Droteo’s hand hurts.’

b. A ochil a mlik a → Ng tȩlȩmall a ochil a mlik.

tȩlȩmall.
‘The wheel of my car is
broken.’

c. A rȩkil a Toki a lluich → Ng lluich mȩ a etiu a rȩkil a

mȩ a etiu. Toki.
‘Toki’s age is 29.’/‘Toki is 29
years old.’

d. A ultutȩlel a babier a → Ng klou a ultutȩlel a babier.

klou.
‘The letter is (very)
important.’

390
 Palauan Reference Grammar

Since both the source sentences and the resulting sentences of

10 are natural and acceptable, the process of subject shifting is
optional in the examples above. Some speakers, however, tend
to prefer the sentences with shifted subjects.
Now, with the resulting sentences of 10, compare the follow-
ing sentences, which are identical in meaning:

(11) a. A Droteo a mȩringȩl a chimal.

‘Droteo’s hand hurts.’

b. A mlik a tȩlȩmall a ochil.

‘The wheel of my car is
broken.’

c. A Toki a lluich mȩ a etiu a

rekil.
‘Toki’s age is 29.’

d. A babier a klou a ultutelel.

‘The letter is (very) important.’

The examples of 11 represent a very common sentence type

in Palauan whose structure is best understood if we assume
that the examples of 11 are derived from the resulting sen-
tences of 10 by a special grammatical process called preposing
of possessor. In comparing the two sets of sentences, we see
that the specific possessors Droteo, mlik ‘my car’, Toki, and
babier ‘letter’, which were originally shifted to the right in 10 as
part of the italicized shifted subjects chimal a Droteo ‘Droteo’s
hand’, ochil a mlik ‘wheel of my car’, etc., are optionally moved
back to sentence-initial position in 11, where they substitute
for the pronominal trace ng; at the same time, the possessed
nouns chimal ‘his hand’, ochil ‘its wheel’, etc. come to appear
in sentence-final position. This process is called “preposing of
possessor” because a specific possessor is removed from the
(shifted) noun phrase of possession of which it is a part and pre-
posed (or moved forward) to the beginning of the sentence. The
type of preposing under discussion here is only possible if the
preposed noun phrase is a possessor. Thus, if we try to prepose
a noun phrase which is not a possessor, as in the example below,
we get a completely ungrammatical sentence:

(12) A John a chillȩbȩdii a bilis.

391
 17 Processes of Sentence Formation:

‘John hit the dog.’ →

*A bilis a John a chillȩbȩdii.

If we accept the validity of the grammatical processes of sub

ject shifting and preposing of possessor, we can easily ex-
plain the unusual order of words in the sentences of 11, as well
as the fact that the preposed noun phrases (which at first glance
look like subjects) are actually interpreted as the possessors of
the various possessed nouns found in sentence-final position.
Thus, a sentence like 11a is ultimately derived from the source
sentence of 10a by the following steps:

(11a’) A chimal a Droteo a mȩringȩl. (source sentence)

Ng mȩringȩl a chimal a (by subject shifting)
Droteo.
A Droteo a mȩringȩ1 a (by preposing of
chimal. possessor)

In derivations such as 11a’, it is very important that the subject

shifting and preposing of possessor rules be applied in the order
indicated: in other words, the possessor is preposed after it has
been moved to the right (together with the rest of the noun
phrase of possession of which it is a part) by the subject shifting
rule.
In the additional examples below, the order of words in the
first (or independent) clause is due to the application of the
subject shifting and preposing of possessor rules. The preposed
possessor has been italicized, and the possessed noun which
has been left “isolated” at the end of the clause has been given
in bold type. The second clause, which is introduced by mȩ ‘and
so’ (see chap. 25), is added to make the sentences sound more
complete; this clause describes some action or state which re-
sults from (or is a consequence of) the action or state of the
preceding clause:

(13) a. A Droteo a mlo ȩr a Guam a bȩchil mȩ ak mo mȩruul a

kȩlir.
‘Droteo’s wife went to Guam, so I’m going to prepare
their (i.e., the family’s) food.’

b. A sȩchȩlik a smechȩr a dȩmal mȩ ak mo omes ȩr tir.

‘My friend’s father is sick, so I’m going to visit them.’

392
 Palauan Reference Grammar

c. A Toki a milsesȩb a blil, mȩ ng kie ȩr a blik.

‘Toki’s house burned down, so she’s staying at my
place.’

d. A Satsko a mlo ȩr a skuul a rȩngȩlȩkel mȩ ng diak a

chad ȩr a blil.
‘Satsko’s children have gone to school, so there’s no
one at home.’

e. A ngȩlȩkek a tȩlȩmall a rrat ȩr ngii mȩ ng kirel ȩl di

mȩrael ȩl mo ȩr a skuul.
‘My child’s bicycle is broken, so he’s got to walk to
school.’

In 13e, the original noun phrase of possession is rrat ȩr a

ngȩlȩkek ‘my child’s bicycle’, in which the possessor is ex-
pressed in a possessor phrase introduced by the relational
word ȩr because the noun rrat ‘bicycle’ is unpossessible (cf.
3.8). When the possessor is preposed in such cases, a trace of
it must be left in the form of a pronoun following the relational
word ȩr. Since the relational word ȩr can only be followed by
emphatic pronouns (cf. 4.3), the pronominal trace in 13e must
be the 3rd pers. sg. emphatic pronoun ngii.
Let us now return to some sentences containing the pos-
sessed nouns soal ‘his/her liking’ and chȩtil ‘his/her disliking’.
If a specific third person possessor is mentioned, we get sen-
tences like the following:

(14) a. Ng soal a Droteo a biang. ‘Droteo likes beer.’

b. A Droteo a soal a biang.

(15) a. Ng chȩtil a Toki a sasimi. ‘Toki dislikes Sashimi.’

A Toki a chȩtil a sasimi.

The sentences above can be easily explained in terms of the

rules of subject shifting and preposing of possessor. Thus, the
derivation of 14a and 15a follows the pattern given in 5 above,
since these sentences show the shifted subjects biang ‘beer’
and sasimi ‘sashimi’. In other words, 14a and 15a are derived
by subject shifting from equational sentences in which the first
noun phrase is biang or sasimi and the second noun phrase

393
 17 Processes of Sentence Formation:

contains soal or chȩtil followed by a specific possessor. These

equational sentences are the source sentences in the scheme
below:

(16) Source Sentence Resulting Sentence

a. A biang a soal a → Ng soal a Droteo a biang. ‘Droteo

Droteo. likes beer.’

b. A sasimi a chȩtil → Ng chȩtil a Toki a sasimi. ‘Toki

a Toki. dislikes sashimi.’

Just as in the examples of 5, the application of the subject

shifting rule to the source sentences of 16 is nearly obligatory,
since the occurrence of these source sentences is quite rare.
The b-sentences of 14–15 are derived from the a-sentences
by preposing the possessor, but the conditions under which this
rule applies are somewhat different from what we described
earlier. Thus, in deriving 14b from 14a, for example, we note
that the possessor Droteo can be preposed even though it Was
never moved to the right as part of a shifted subject. In order
to account for the correct ordering of words in 14b, however,
we must still assume that the preposing of possessor rule ap-
plies after the subject shifting rule. The step-by-step derivation
of 14b is therefore as follows:

(17) A biang a soal a Droteo. (source sentence)

Ng soal a Droteo a biang. (by subject shifting)
A Droteo a soal a biang. (by preposing of possessor)

17.4. SUBJECT SHIFTING AND PREPOSING OF

POSSESSOR WITH EXPRESSIONS CONTAINING
RENG
In this section we will examine another common Palauan sen-
tence type in which the processes of subject shifting and
preposing of possessor play an important role. Palauan has a
very large number of expressions consisting of a possessed
form of the abstract noun reng ‘heart, spirit’ followed by an
intransitive verb—usually a state verb. These expressions are
used to express emotional states, feelings, personality traits,
and the like. Often, it is difficult or impossible to predict the

394
 Palauan Reference Grammar

exact meaning of these expressions from the meaning of the

independently-occurring intransitive verb, as the following ex-
amples indicate:

(18) a. Ng ungil a rȩnguk.

‘I’m happy.’

b. Ng klou a rȩngul a sensei.

‘The teacher is patient.’

c. Ng smechȩr a rȩngmam.
‘We’re homesick.’

In the sentences above, we can see the connection between the

two meanings of ungil ‘good’—‘happy’, klou ‘big’—‘patient’, and
smechȩr ‘sick’—‘homesick’, but we have no consistent way of
predicting how the meaning will change when the particular
intransitive (state) verb is associated with the abstract noun
reng ‘heart, spirit’.
Before discussing the grammatical structure of the sen-
tences of 18, we shall list some of the most commonly-used
expressions with reng. The gloss given in parentheses is the
meaning which the intransitive verb has when it occurs inde-
pendently (i.e., without reng):

(19) ngmasȩch a ‘angry’ (‘rise’)

rȩngul
kȩsib a rȩngul ‘angry’ (‘perspiring’)
mȩched a ‘thirsty’ (‘shallow’)
rȩngul
mȩkngit a ‘be in a bad/sad mood’ (‘bad’)
rȩngul
beot a rȩngul ‘easygoing, nonchalant, (‘easy’)
unmotivated’
mȩsisiich a ‘hardworking, (‘strong’)
rȩngul well-motivated’
bȩralm a ‘lazy, unmotivated’ (‘watery’)
rȩngul
mȩchitȩchut a ‘easily discouraged’ (‘weak’)
rȩngul
kekȩdeb a ‘short-tempered’ (‘short’)
rȩngul
kekȩre a rȩngul‘uncomfortable’ (‘small’)

395
 17 Processes of Sentence Formation:

mȩsaul a ‘not feel like’ (‘tired’)

rȩngul
suebȩk a ‘worried’ (‘fly’)
rȩngul
songȩrengȩr a ‘have a strong desire for’ (‘hungry’)
rȩngul
diak a rȩngul ‘inconsiderate, impolite’ (‘isn’t, doesn’t
exist’)

It is easy to see that the examples of 18 are derived by

the process of subject shifting. In other words, we propose that
18a–c have the following source sentences:

(18’) a. A rȩnguk a ungil.

(‘I’m happy.’)

b. A rȩngul a sensei a klou.

(‘The teacher is patient.’)

c. A rȩngmam a smechȩr.
(‘We’re homesick.’)

The sentences in 18’ are not acceptable to any speakers,

however, and we must therefore conclude that subject shifting
is obligatory if the sentence contains a special expression with
reng ‘heart, spirit’. If we follow the subject shifting analysis pro-
posed here, we can easily explain what would otherwise be two
rather unusual facts about sentences like 18. First of all, such
sentences always end in a noun phrase of possession which con-
tains a possessed form of reng; this is of course due to the fact
that the sentence subject is always obligatorily shifted. Second,
sentences of this kind always begin with ng; this ng is the
pronominal trace which appears in the spot originally occupied
by the shifted 3rd pers. sg. noun phrase subject.
If we compare a sentence like 18b with the following, which
is identical in meaning,

(20) A sensei a klou a rȩngul.

‘The teacher is patient.’

we see immediately that the process of proposing of pos-

sessor (cf. 17.3 above) also applies to sentences containing
noun phrases of possession with reng. Since both 18b and 20

396
 Palauan Reference Grammar

are perfectly acceptable sentences, the rule which preposes the

possessor is of course optional rather than obligatory. A couple
of examples parallel to 18b and 20 are given below; the pre-
posed possessor has been italicized:

(21) a. Ng suebȩk a rȩngul a Droteo. ‘Droteo is worried.’

b. A Droteo a suebȩk a rȩngul.

(22) a. Ng mȩched a rȩngrir a ‘My friends are

rȩsȩchȩlik. thirsty.’
b. A rȩsȩchȩlik a mȩched a rȩngrir.

17.5. PREPOSING IN RECIPROCAL SENTENCES

In 10.1 we noted that reciprocal verbs must always have
plural subject noun phrases. One type of plural subject consists
of two single nouns or two noun phrases joined by the con-
necting word mȩ ‘and’ (cf. 25.4), as shown in the examples
below:

(23) a. A Droteo mȩ a Toki a kausȩchȩlei.

‘Droteo and Toki are friends.’

b. A tȩkoi ȩr a Ruk mȩ a tȩkoi er a Belau a kakȩrous.

‘Trukese and Palauan are different.’

c. A rȩchad ȩr a Merikel mȩ a rȩchad ȩr a Sina a mle

kauchȩraro.
‘The Americans and the Chinese used to be enemies.’

Just like any other subject noun phrase, the plural subject noun
phrases in 23 can be moved to the right of the verb by the
process of subject shifting explained in 17.2 above. Thus, when
we apply subject shifting to 23a–c, we get the following sen-
tences, which are equivalent in meaning:

(24) a. Tȩ kausȩchȩlei a Droteo mȩ a Toki.

‘Droteo and Toki are friends.’

b. Ng kakȩrous a tȩkoi ȩr a Ruk mȩ a tȩkoi er a Belau.

‘Trukese and Palauan are different.’

397
 17 Processes of Sentence Formation:

c. Tȩ mle kauchȩraro a rȩchad ȩr a Merikel mȩ a rȩchad

ȩr a. Sina.
‘The Americans and the Chinese used to be enemies.’

Because the shifted subjects in 24a and 24c designate human

beings, the 3rd pers. human pl. non-emphatic pronoun tȩ is
used as a pronominal trace. In 24b, however, the pronominal
trace is ng because this pronoun substitutes for the non-
human plural noun phrase tȩkoi ȩ r a Ruk mȩ a tȩkoi er a Belau
‘Trukese and Palauan’ (cf. 2.4). 3
The sentences of 24 can be further changed (or trans-
formed) by a process of preposing, but one which differs
somewhat from that already discussed in the sections above.
Before explaining what is involved, let us compare the sen-
tences below with those of 24:

(25) a. A Droteo a kausȩchȩlei ngii mȩ a Toki.

‘Droteo is friends with Toki.’

b. A tȩkoi ȩr a Ruk a kakȩrous ngii mȩ a tȩkoi er a Belau.

‘Trukese is different from Palauan.’

c. A rȩchad ȩr a Merikel a mle kauchȩraro tir mȩ a rȩchad

ȩr a Sina.
‘The Americans used to be enemies of the Chinese.’

In the sentences of 25 we note that the first noun or noun

phrase of the shifted subjects of 24 has been moved back to
sentence-initial position, where it replaces the pronominal
traces ng or tȩ. This type of preposing is different from the
process of preposing of possessor in two important respects.
First, the preposed noun or noun phrase is, of course, not a pos-
sessor, but rather the first member of a plural noun phrase of
the form A mȩ B. Second, when the first member of (shifted)
A mȩ B is preposed, a trace of it must remain in the form of
an emphatic pronoun. This accounts for the occurrence of ngii
in 25a–b, where the preposed noun phrase is singular, and for
the occurrence of tir in 25c, where the preposed noun phrase is
plural.

398
 Palauan Reference Grammar

To summarize what we have said above, the sentences of 24

are derived from those of 23 by subject shifting, while those of
25 are in turn derived from those of 24 by the special process
of preposing just described. Note the sample derivation below,
which relates 23a, 24a, and 25a:

(26) A Droteo mȩ a Toki a (source sentence) →

kausȩchȩlei.
Tȩ kausȩchȩlei a Droteo mȩ (by subject shifting) →
a Toki.
A Droteo a kausȩchȩlei ngii (by preposing of first member
mȩ a Toki. of A mȩ B)

As the English equivalents for 25 show, the sentences with

pre-posed subjects involve a change of viewpoint. Thus, in 23a
and 24a, for example, the speaker is paying more-or-less equal
attention to both members of the plural subject (Droteo and
Toki), while in 25a the speaker’s attention is focused more on
the person designated by the preposed noun (Droteo).
In some reciprocal sentences, the plural subject is a noun
phrase of possession in which the possessor is a sequence of the
form A mȩ B. Thus, in the sentence below,

(27) A blȩkȩrdȩlir a Droteo mȩ a Toki a kakngodȩch.

‘The personalities of Droteo and Toki are different from
each other.’

the possessed noun blȩkȩrdȩlir ‘their personalities’ has the 3rd

pers. human pl. possessor suffix -ir, which agrees with the
following plural possessor Droteo mȩ a Toki ‘Droteo and Toki’.
Now, the italicized subject noun phrase of 27 can be moved to
the right of the verb by the process of subject shifting, resulting
in the sentence below:

(28) Ng kakngodȩch a blȩkȩrdȩlir a Droteo mȩ a Toki.

‘The personalities of Droteo and Toki are different from
each other.’

Since the sequence Droteo mȩ a Toki of 28 is a possessor, it

can be moved back to sentence-initial position by the process of
preposing of possessor which we described in 17.3 above. We
therefore obtain the following sentence:

399
 17 Processes of Sentence Formation:

(29) A Droteo mȩ a Toki a kakngodȩch a blȩkȩrdȩlir.

‘Droteo and Toki are different in personality.’

17.6. SUMMARY OF PROCESSES AFFECTING

PALAUAN WORD ORDER
In 17.2–5 above, we have seen how two major grammatical
processes—subject shifting and preposing—bring about
striking changes in the word order of Palauan sentences. These
processes are of wide applicability in Palauan and can account
for many types of sentences other than those dealt with above.
Thus, as we will see in later chapters, the processes of subject
shifting and preposing not only play an important role in the for-
mation of questions and passive sentences, but they also affect
sentences containing negative verbs, existential verbs, and
time clauses.
In addition to the major processes of subject shifting and
preposing, there are a few relatively minor processes which
affect the word order of Palauan sentences. Perhaps the most
obvious of these is found in sentences containing the verb msa
‘give’, which characteristically takes two objects. 4 Thus, in the
examples below, the first object—the person who receives what
is given—is italicized, while the second object—the thing which
is given—is printed in bold type:

(30) a. Ak milsa a Helen a omiange.

‘I gave Helen a souvenir.’

b. Ak milstȩrir a rȩsȩchȩlik a
hong.
‘I gave my friends a book.’

c. Ak mo mȩskau a udoud.
‘I’m going to give you some
money.’

d. A Droteo a milskak a present.

‘Droteo gave me a present.’

As the examples above show, the verb msa ‘give’ is unusual

in the following respects. First of all, this verb seems to have
only perfective forms, but no imperfective forms. Second, the

400
 Palauan Reference Grammar

various object pronoun suffixes found in these perfective forms

agree in person and number with the first object—namely, the
person receiving what is given—rather than with the second
object, which identifies the thing given. Thus, in 30a, the form
milsa ‘gave (it to) him/her’ has the 3rd pers. sg. object pronoun
suffix -a, 5 which agrees with the following specific 3rd pers. sg.
object Helen. By contrast, milstȩrir ‘gave (it to) them’ of 30b has
the 3rd pers. human pl. object pronoun suffix -tȩrir because the
following noun rȩsȩchȩlik ‘my friends’ is human plural. Finally,
the second object can never be preceded by the specifying
word ȩr (cf. 2.7); therefore, there is no overt way of marking
this object as specific vs. non specific or singular vs. plural.
For this reason, the object hong in 30b could also be interpreted
to mean ‘(some) books’, ‘the book’, or ‘the books’.
In the sentences of 30, the order of the two objects can be
reversed. The process which brings about this change in word
order is of relatively minor significance in that its application is
limited to sentences containing the verb msa ‘give’. If the first
object is singular, as in 30a, reversing the order of the two ob-
jects usually results in a sentence which is identical in meaning
and equal in acceptability—namely,

(31) Ak milsa a omiange a Helen.

‘I gave a souvenir to Helen.’ 6

If, however, the first object is plural, as in 30b, interchanging

the two objects results in a rather awkward sentence which
some Palauan speakers accept but others reject—i.e.,

(32) ?Ak milstȩrir a hong a

rȩsȩchȩlik.
‘I gave a book to my friends.’

Perhaps 32 is of questionable acceptability because the 3rd

pers. human pl. object pronoun suffix -tȩrir has come to appear
next to a singular, non-human noun (hong ‘book’). 7

401
 17 Processes of Sentence Formation:

17.7. DEPENDENT CLAUSES RESULTING FROM

SUBJECT SHIFTING
As we have seen above, the process of subject shifting accounts
for sentences in which the nouns of liking and disliking (soal
and chȩtil) are directly followed by concrete nouns or noun
phrases. Now, with a sentence like la, repeated here for conve-
nience,

(1a) Ng soak a biang.

‘I like beer.’

compare the following:

(33) Ng soak ȩl mȩlim a biang.

‘I want to drink some beer.’

While soak is followed by the concrete noun biang ‘beer’ in 1a,

the sequence following soak in 33 has some of the character-
istics of dependent clauses mentioned in chap. 15. First of
all, this sequence is introduced by ȩl, and second, it does not
contain any overtly-expressed subject. We nevertheless know
that the understood subject of mȩlim ‘drink’ is the same as the
person identified by the possessor suffix on soak—namely, the
speaker (ak ‘I’). If we put a sentence like 33 into the past tense,
we have further evidence that the sequence introduced by ȩl is
a dependent clause. Thus, in the sentence below,

(34) Ng mle soak ȩl mȩlim a biang.

‘I wanted to drink some beer.’

the verb form mȩlim remains in the present tense even though
the whole sentence designates a past situation.
Just as we derived example la by subject shifting according
to the following scheme,

(5a) Source Sentence Resulting Sentence

A biang a soak. → Ng soak a biang.

‘I like beer.’

402
 Palauan Reference Grammar

we propose that subject shifting is also responsible for examples

like 33, except that the structure of the source sentence is more
complex. Observe, therefore, the following:

(35) Source Sentence Resulting Sentence

[Ak mȩlim a biang] a soak. → Ng soak ȩl mȩlim a biang.

‘I want to drink some beer.’

The source sentence of 35 is of course not spoken in Palauan,

but must always be transformed into the resulting sentence.
Let us now explain the process of derivation schematized in 35.
While the source sentence of 5a has a single noun (biang ‘beer’)
as its subject, we propose that the “real” subject of the source
sentence of 35 is the bracketed sentence ak mȩlim a biang ‘I
drink beer.’ In other words, the subject of soak does not nec-
essarily have to be a concrete noun, as it is in 5a, but it can
also be a whole activity which involves a subject (or doer) and
a verb phrase. The bracketed sentence in subject position in
35 symbolizes the fact that in this example the subject of soak,
is the whole activity ak mȩlim a biang ‘I drink beer.’ Thus, the
source sentences in both 5a and 35 are equational sentences
in which the subject is being equated with the noun phrase
soak; the only difference is that in 35 an (abstract) activity
rather than a concrete thing is being asserted as the speaker’s
desire.
If we formulate the source sentence of 35 as described
above, we can easily account for the resulting sentence of 35 in
terms of processes and principles with which we are already fa-
miliar. Furthermore, we can see that the derivation of sentences
like la and 33 (= the resulting sentences of 5a and 35) is really
very similar in that the process of subject shifting is involved in
both cases. Now, let us look at the derivation of 35 in detail. In
this example, subject shifting applies to the whole bracketed
sentence ak mȩlim a biang, since this sequence is the subject of
soak. The application of subject shifting results in the following
structure:

(35’) [Ak mȩlim a biang] a soak. → Ng soak [ak mȩlim a biang].

The resulting sentence of 35’ is still not a spoken sentence of

Palauan, but it is “halfway there” in the sense that it contains
sentence-initial ng, which of course is a pronominal trace left

403
 17 Processes of Sentence Formation:

behind by the shifted subject. Two further changes are nec-

essary to transform the resulting sentence of 35’ into a full-
fledged grammatical sentence—namely, into 33 (= the resulting
sentence of 35). These changes turn the shifted subject ak
mȩlim a biang into a dependent clause—i.e.,

(36) Ng soak [ak mȩlim a biang]. → Ng soak ȩl mȩlim a biang.

‘I want to drink some beer.’

In 36, the subject ak ‘I’ of the shifted bracketed sentence is

deleted because the possessor suffix on soak already makes it
clear that the drinker of the beer will be the speaker. In other
words, the subject ak ‘I’ of the bracketed sentence is deleted
under identity with the pronominal possessor of the preceding
possessed noun soak. In addition, the word ȩl is inserted to
introduce the shifted sequence. The resulting sentence of 36
has therefore come to contain the dependent clause ȩl mȩlim a
biang.
In the discussion above, we have seen that the process of
subject shifting accounts for both of the sentences 1a and 33,
repeated here as 37a–b:

(37) a. Ng soak a biang.

‘I like beer.’/‘I’d like some
beer.’

b. Ng soak ȩl mȩlim a biang.

‘I want to drink some beer.’

What differentiates 37a from 37b is that in the former example,

a concrete noun phrase has been shifted, while in the latter ex-
ample, an (abstract) bracketed sentence has been shifted. This
difference in shifted subject correlates with the following con-
sistent difference in meaning. Example 37a, with a concrete
noun phrase following soak, can be either a general statement
(‘I like beer.’) or a statement of the speaker’s desire on a spe-
cific occasion (‘I’d like some beer.’) By contrast, example 37b,
with a dependent clause following soak, can only be a statement
about a specific occasion. A similar contrast in interpretation is
found in sentences containing the noun of disliking chȩtil. Thus,
with 37a–b compare the following pair of sentences:

(38) a. Ng chȩtirir a sasimi.

404
 Palauan Reference Grammar

‘They dislike sashimi.’ /‘They don’t want any sashimi.’

b. Ng chȩtirir ȩl mȩnga a sasimi.

‘They don’t want to eat any sashimi.’

As we have seen above, the possessed forms of soal ‘his/her

liking’ and chȩtil ‘his/her disliking’ can be followed by shifted
subjects which are either concrete noun phrases or (abstract)
bracketed sentences. In other words, soal and chȩtil occur in
equational source sentences of the form noun phrase + soal/
chȩtil or [sentence] + soal/chȩtil. As opposed to soal and chȩtil,
the obligatorily possessed nouns sȩbȩchel ‘his/her ability’ and
kirel ‘his/her obligation’ (cf. 17.1 above) cannot occur in an
equational source sentence whose subject is a concrete noun
phrase; therefore, source sentences containing these two
special nouns can only be of the form [sentence] + sȩbȩchel
/ kirel. 8 For this reason, the possessed forms of sȩbȩchel and
kirel are always followed by shifted subjects which are (ab-
stract) bracketed sentences that take the form of dependent
clauses. This is the case in 1c–d above and in examples like the
following:

(39) a. Ng sȩbȩchek ȩl eko ȩr a blim ȩr a klukuk.

‘I can come to your house tomorrow.’

b. A Droteo a sȩbȩchel ȩl ousbech ȩr a mlik.

‘Droteo can use my car.’

c. Ng kirek ȩl mȩnguiu ȩr tia ȩl hong.

‘I have to read this book.’

d. A Toki a kirel ȩl mȩsuub er a elȩchang.

‘Toki has to study today.’

In deriving 39b and 39d, the processes of subject shifting and

preposing of possessor must both be applied. The step-by-step
derivation of 39b is therefore as follows:

(40) a. [A Droteo a ousbech ȩr a mlik] a sȩbȩchel a Droteo.

(source sentence) →

b. Ng sȩbȩchel a Droteo [a Droteo a ousbech ȩr a mlik].

(by subject shifting applied to bracketed sentence) →

405
 17 Processes of Sentence Formation:

c. Ng sȩbȩchel a Droteo ȩl ousbech ȩr a mlik.

(by dependent clause formation) →

d. A Droteo a sȩbȩchel ȩl ousbech ȩr a mlik.

(by preposing of possessor).

In step c, we use the term “dependent clause formation” to refer

to the rules which introduce ȩl and delete the subject (Droteo)
of the bracketed sentence under identity with the preceding
occurrence of Droteo as possessor of sȩbȩchel. Since the sen-
tence in 40cis an acceptable Palauan sentence, application of
the preposing of possessor rule to derive 40d is merely optional.
9

17.8. SUBJECT SHIFTING AND DERIVED ACTION

NOUNS
At the end of the preceding section, we noted that the possessed
forms of soal ‘his/her liking’ and chȩtil ‘his/her disliking’ can
be followed by shifted subjects which are either concrete noun
phrases or (abstract) bracketed sentences. In this section, we
will examine a third type of sequence which can follow the pos-
sessed forms of soal and chȩtil—namely, abstract noun phrases
containing derived action nouns in o-. These action nouns, as
we saw in 8.6, are derived simply by prefixing o- to transitive
or intransitive action verbs—e.g., we have omȩluchȩs ‘writing’
from mȩluchȩs ‘write’, omilil ‘playing’ from milil ‘play’, and so
on. Derived action nouns designate actions or activities as ab-
stract or general concepts and are used in examples like the
following:

(41) a. A omȩruul ȩl kall a urerir a rȩdil.

‘Preparing food is women’s work.’

b. A omȩnguiu ȩl tȩkoi ȩ r a Sina a kmal mȩringȩl.

‘Reading Chinese is very difficult.’

c. Ak chilitii a omȩlamȩch ȩ l dȩkool.

‘I gave up/quit smoking cigarettes.’

406
 Palauan Reference Grammar

In 41a–b, the action nouns omȩruul ‘preparing’ and omȩnguiu

‘reading’ are part of the italicized subject noun phrases, while
in 41c the action noun omȩlamȩch ‘smoking’ is found in the
italicized object noun phrase. Since omȩruul, omȩnguiu, and
omȩlamȩch are derived from transitive verbs (mȩruul ‘make,
prepare’, mȩnguiu ‘read’ and mȩlamȩch ‘smoke, chew’) they
can be associated with objects. These objects are always intro-
duced by ȩl, which therefore precedes kall ‘food’, tȩkoi ȩ r a
Sina ‘Chinese’, and dȩkool ‘cigarettes’ in 41a–c above. 10
In the examples below, possessed forms of soal and chȩtil
are followed by shifted subjects which contain derived action
nouns in o-.

(42) a. Ng soak a omȩlim ȩl biang.

‘I like drinking beer.’

b. Ng soam a {omȩngȩdub/omȩsub/omȩrael}?
‘Do you like {swimming/studying/traveling}?’

c. A Toki a chȩtil a omȩruul ȩl kall.

‘Toki dislikes preparing food.’

d. Ng chȩtik a omȩlamȩch ȩl buuch.

‘I dislike chewing betel nut.’

e. A sensei a chȩtil a omȩngȩrodȩch ȩr a klas.

‘The teacher doesn’t like people making noise in class.’

As expected, the source sentences for 42a–e are equational sen-

tences in which the subject noun phrase contains a derived
action noun. Thus, 42c, for example, is derived in the following
manner:

(43) a. A omȩruul ȩl kall a chȩtil a Toki.

(source sentence) →

b. Ng chȩtil a Toki a omȩruul ȩl kall.

(by subject shifting) →

c. A Toki a chȩtil a omȩruul ȩl kall.

(by preposing of possessor).

407
 17 Processes of Sentence Formation:

Because derived action nouns in o- designate actions or ac-

tivities as abstract or general concepts, as mentioned above,
it is no surprise that the examples of 42 are interpreted as
general statements (or questions) rather than as statements
(or questions) about specific occasions. Because 42a, for ex-
ample, is a general statement, it contrasts in meaning with 33,
which refers to a specific occasion. Both of these examples are
now repeated here for purposes of comparison:

(44) a. Ng soak a omȩlim ȩl biang.

‘I like drinking beer.’

b. Ng soak ȩl mȩlim a biang.

‘I want to drink some beer.’

As the English equivalents show, the sentences of 44 are quite

different from each other in meaning. In 44b, which has a
dependent clause following soak, the action of drinking beer
refers to a specific occasion. Therefore, this sentence would be
used by the speaker at the very moment when he has a desire
to drink beer. By contrast, example 44a, which has a derived
action noun following soak, views the action of drinking beer
in a general (or perhaps, habitual) sense. For this reason, 44a
could be spoken at any time as an expression of the speaker’s
habit or preference, but would not be appropriate to express the
speaker’s momentary desire to drink beer. In other words, 44b
would be a suitable answer to the question ‘What would you like
to drink?’, while 44a would not.
The contrast in meaning between 44a–b is paralleled in the
pairs of sentences below, which contain possessed forms of
chȩtil ‘his/her disliking’ and sȩbȩchel ‘his/her ability’:

(45) a. A Toki a chȩtil a omȩruul ȩl kall.

‘Toki dislikes preparing food.’

b. A Toki a chȩtil ȩl mȩruul a kall.

‘Toki doesn’t want to make the food.’

(46) a. Ng sȩbȩchem a omȩlim ȩl rrom?

‘Are you capable of drinking liquor?’

b. Ng sȩbȩchem ȩl mȩlim a rrom?

‘Can you have a drink of liquor?’

408
 Palauan Reference Grammar

Example 46a is a rather challenging question in which the

hearer is being asked whether he has the ability—i.e., strength
or maturity—to drink liquor. By contrast, 46b is simply an in-
vitation to drink liquor and implies nothing about the hearer’s
“prowess” as a drinker.

17.8.1. POSSESSED FORMS OF DERIVED ACTION NOUNS

Just like any other nouns, the derived action nouns in o- dis-
cussed above can take the various possessor suffixes. The re-
sulting possessed forms can be used in sentences with soal and
chȩtil, as follows:

(47) a. A sensei a soal a omȩsubek ȩr a tȩkoi ȩr a

Merikel.
‘The teacher likes the way I’m studying English.’

b. Ng soam a omȩlmȩchel a Droteo ȩr a dȩkool?

‘Do you like Droteo’s smoking (so many)
cigarettes?’

c. A dȩmak a chȩtil a omȩrȩllek ȩr a party. 11

‘My father dislikes my having (so many) parties.’

d. Ng chȩtik a omȩlmil a Cisco ȩr a rrom.

‘I dislike the way Cisco drinks (so much) liquor.’

e. Ng chȩtik a omililel a Droteo ȩr a klȩbȩsei.

‘I don’t like Droteo’s fooling around (so much) at
night.’

In the possessed forms omȩsubek ‘my studying’ (cf. mȩsuub

‘study’), omȩlmȩchel ‘his smoking’ (cf. mȩlamȩch ‘smoke’), omȩ
rȩllek ‘my preparing’ (cf. mȩruul ‘make, prepare’), and omȩlmil
‘his drinking’ (cf. mȩlim ‘drink’), you should be able to recognize
certain patterns of vowel reduction and vowel deletion (cf.
3.4 and 3.4.1–3). Note, further, that the objects following these
possessed nouns must be expressed by a relational phrase (cf.
14.9). Thus, the objects tȩkoi ȩ r a Merikel ‘English’ of 47a,
dȩkool ‘cigarettes’ of 47b, party of 47c, and rrom ‘liquor’ of 47d
are all preceded by the relational word ȩr.

409
 17 Processes of Sentence Formation:

As the English equivalents for the sentences in 47 are de-

signed to show, the possessed forms of action nouns in o- always
imply that the habitual action in question is a fact—i.e., that it is
being pursued regularly by the person referred to by the posses-
sor suffix. Thus, in 47b, for instance, the speaker assumes (or
presupposes) it is a fact that Droteo smokes a lot of cigarettes
and then asks the hearer whether he approves of this fact. Sim-
ilarly, in 47e the speaker recognizes the fact that Droteo does
a lot of fooling around and then offers his (negative) opinion or
judgment about this fact.
In 47b, 47d, and 47e, where a specific 3rd person possessor
is mentioned, this possessor always identifies the agent—i.e.,
the person who is doing the action denoted by the possessed
action noun in o-. Thus, in omȩlmȩchel a Droteo ‘Droteo’s
smoking’ of 47b, the “possessor” Droteo is the one who is pur-
suing the activity of smoking. Occasionally, we will observe ex-
pressions of the form possessed action noun in o- + specific
possessor which are interpreted differently from those of 47b,
47d, and 47e. Thus, in the examples below, the italicized “pos-
sessors” actually designate the objects of the actions denoted
by the possessed action nouns in o-:

(48) a. A omȩrȩllel a mlai a kmal mȩringȩl.

‘(The method of) making canoes is very difficult.’

b. Ak mla mȩlasȩm ȩr a omȩrȩllel a kall ȩr a Sina, e ng di

ng diak lsȩbȩchek.
‘I’ve tried (the method of) preparing Chinese food, but
I’m not good at it.’

17.9. SOAL AND CHȨTIL FOLLOWED BY

HYPOTHETICAL VERB FORMS
The possessed forms of soal and chȩtil can be followed by hy-
pothe tical verb forms (cf. 4.10) to convey the idea “X wants/
does not want Y to do something”. Observe the following sen-
tences:

(49) a. A sensei a soal a kusuub.

‘The teacher wants me to study.’

b. A dȩmak a chȩtil a kuruul a party.

410
 Palauan Reference Grammar

‘My father doesn’t want me to have

parties.’

c. Ng soak a rȩngalȩk a lomȩngur.

‘I want the children to eat.’

d. Ng chȩtik a ngȩlȩkek a lolamȩch a dȩkool.

‘I don’t want my child smoking cigarettes.’

e. A rȩsȩchȩlik a sorir a chobong.

‘My friends want you to go.’

In each of the sentences above, one person (or group of

persons) X wants or doesn’t want another person (or group of
persons) Y to do something. While X is identified by the pos-
sessor suffix on soal or chȩtil, Y is expressed by the hypo-
thetical pronoun which is prefixed to the hypothetical verb
form. Furthermore, if X or Y is a third person, then a specific
noun may be mentioned. In 49a, for example, X—the person
desiring something—is identified by the specific noun sensei
‘teacher’ and the 3rd pers. sg. possessor suffix on soal, while
Y—the person who is expected to do something—is identified by
the 1st pers. sg. hypothetical pronoun prefix ku- ‘I’ on kusuub.
Similarly, in 49e, X is identified by the specific plural noun
rȩsȩchȩlik ‘my friends’ and the 3rd pers. human pl. possessor
suffix on sorir, while Y corresponds to the 2nd pers. (sg. or pl.)
hypothetical pronoun prefix cho- ‘you’ on chobong.
It appears that the italicized portions of 49a–e are actually
instances of conditional clauses. As we will see in 19.1,
Palauan conditional clauses express events as possible occur-
rences rather than as real facts; this important feature of their
interpretation is observed in the following example:

(50) A kbo ȩ r a Guam, e ak mo omes ȩr a Toki.

‘If I go to Guam, then I’ll see Toki.’

In 50, the event “I go to Guam” is not a real occurrence but

instead one which is hypothesized or put forth as a pos-
sibility. For this reason, the italicized conditional clause of
50 corresponds to English sequences introduced by ‘if’. As
the examples of 49 and 50 show, Palauan conditional clauses
contain hypothetical verb forms (which characteristically des-
ignate unreal or hypothesized events—hence, the term hy-

411
 17 Processes of Sentence Formation:

pothetical) and are introduced by the conditional clause

marker a ‘if’. If we are correct in assuming that the italicized
portions of 49a–e are conditional clauses, then we should rec-
ognize that the English equivalents for these examples are
rather free. For instance, we have translated 49a as ‘The
teacher wants me to study’, but a word-for-word translation
would be something like ‘The teacher would like it if I studied.’
Both the free translation and the more literal translation are
really equivalent, however, since they both imply that the
speaker has not been studying and that the teacher would like
the situation to change.
Since the sentences of 49 contain conditional clauses and
therefore express possible events, their meaning is quite dif-
ferent from the examples of 47, in which the possessed forms of
action nouns in o- clearly refer to real events (i.e., facts). This
contrast is observed in the following pairs of sentences (= 47a
vs. 49a and 47c vs. 49b):

(51) a. A sensei a soal a omȩsubek.

‘The teacher likes my studying (so hard).’

b. A sensei a soal a kusuub.

‘The teacher wants me to study.’

(52) a. A dȩmak a chȩtil a omȩrȩllek ȩr a party.

‘My father dislikes my having (so many)
parties.’

b. A dȩmak a chȩtil a kuruul a party.

‘My father doesn’t want me to have
parties.’

In the a-sentences above, the possessed forms omȩsubek ‘my

studying’ and omȩrȩllek ‘my making (a party)’ describe activ-
ities or events which are actual facts, and the possessed forms
soal and chȩtil express some third person’s opinion about these
facts. By contrast, the activities described by kusuub ‘(if) I
study’ and kuruul ‘(if) I make (a party)’ in the b-sentences are
not real facts at the present moment, but are events which some
third party would react favorably or unfavorably to if they oc-
curred.

412
 Palauan Reference Grammar

17.10. FURTHER DISCUSSION OF THE FOUR

SPECIAL POSSESSED NOUNS
In the sections above, we have examined the most important as-
pects of the meaning and use of the four possessed nouns soal,
chȩtil, sȩbȩchel, and kirel. In this section we will mention some
further details about each of these words.

a. soal and chȩtil.

As we have seen above, soal and chȩtil are opposite in mean-

ing. Therefore, 53a below usually has 53b as its opposite:

(53) a. Ng soak ȩl mong.

‘I want to go.’

b. Ng chȩtik ȩl mong.
‘I don’t want to go.’

It is also possible to derive an opposite of 53a by adding the

negative verb diak ‘isn’t, doesn’t exist’ (see chap. 18); thus, we
have

(54) Ng diak lsoak 12 ȩl mong.

‘I (really) don’t want to go.’

Though 53b and 54 are interchangeable in many contexts, some

speakers feel that 54 is more emphatic, blunter, or less polite
than 53b.
The 3rd pers. sg. possessed form of soal can also be used in
the meaning ‘look as if’, as shown in the examples below:

(55) a. A eangȩd a soal ȩl mo ungil ȩr a klukuk.

‘The weather looks as if it might be good tomorrow.’

b. A chull a soal ȩl mo ȩr ngii ȩr a kȩbȩsengei.

‘It looks as if it’s going to rain tonight.’

c. A ngais a soal ȩl ruebȩt mȩ bo mungil ȩl orrekȩd.

‘The eggs look as if they’ll fall out, so hold on to them
carefully.’

413
 17 Processes of Sentence Formation:

In a related meaning, the 1st pers. sg. possessed form soak

corresponds to ‘feel as if’, as in the following:

(56) Ak kmal mȩdingȩs mȩ ng soak ȩl mo smechȩr.

‘I’m very full, so I feel as if I’ll be sick.’

The uses of soal and soak described here imply that the speaker
has evidence—through observation or direct personal expe-
rience—that some event is going to take place. Thus, 55a–b, for
example, are predictions based on the speaker’s observation of
some natural phenomenon—e.g., the condition of the sky.

b. sȩbȩchel.

In our discussion above, we referred to sȩbȩchel as a noun of

“ability”: in other words, sȩbȩchel expresses the fact that some-
one is able to do something because he has the time (or oppor-
tunity) to do it, or has the physical capacity to perform the task
involved. In addition to this meaning, sȩbȩchel can also express
the fact that someone has permission to do something: in this
case, someone is able to do something in the sense that no one
else is preventing or forbidding his doing it. Often, it is only the
context or situation which tells us whether sȩbȩchel refers to
ability or permission. Therefore, the following sentences are
ambiguous when examined in isolation (the same is true for 1c,
39a–b, and 46b above):

(57) a. Ng sȩbȩchek ȩl mo er a mubi ȩr a klukuk.

‘I can go to the movies tomorrow.’ (= ‘I have time to go
to the movies tomorrow.’/‘I have permission to go to the
movies tomorrow.’)

b. Ng diak lsȩbȩchem ȩl mo ȩr a che er a elȩchang?

‘Can’t you go fishing now?’ (= ‘Aren’t you able to go
fishing now?’/‘Aren’t you allowed to go fishing now?’)

As 57b shows, the possessed forms of sȩbȩchel remain am-

biguous when they are preceded by the negative verb diak
‘isn’t, doesn’t exist’. It is also interesting to note that the best
English equivalent for sȩbȩchel—namely, can—is ambiguous in
the same way.

c. kirel.

414
 Palauan Reference Grammar

In the examples 39c–d above, we saw that the possessed

forms of kirel, when followed by a dependent clause, convey
the idea of obligation or necessity. The possessed forms of
kirel can also be associated with derived action nouns in o-, in
which case they imply that someone is suited to performing a
particular activity. This usage is normally found in negative sen-
tences like the following:

(58) a. A omȩsuub ȩl ochur a diak lȩkirel a Droteo.

‘Studying math is not something Droteo is suited for.’

b. A omȩlim ȩl biang a diak lȩkirir a rȩngalȩk ȩr a skuul.

‘Drinking beer isn’t meant for students.’

Instead of a derived action noun in o-, a derived abstract noun

in klȩ- (cf. 8.5) may be associated with kirel, as in the example
below:

(59) A klsensei a diak lȩkirek ȩl ureor.

‘Being a teacher isn’t meant for me.’

When sentences like 39c–d are turned into negative sen-

tences by adding diak ‘isn’t, doesn’t exist’, the resulting
meaning is either ‘must not’ or ‘doesn’t have to’. Note the fol-
lowing examples:

(60) a. Ng diak lȩkirek ȩl melim a biang.

‘I must not drink beer.’

b. Ng diak lȩkirem ȩl mo sensei. Ng kirem ȩl mo toktang.

‘You must not become a teacher; you must become a
doctor.’

c. Ng diak lȩkirir ȩl mong.

‘They don’t have to go.’

A noun phrase of possession containing a possessed form of

kirel can be used as a kind of specifying clause (cf. 15.7) to
identify the person who benefits from some activity or the thing
which is the cause or purpose of some activity. Such specifying
clauses are italicized in the sentences below:

(61) a. Ak mȩngȩtmokl ȩr a blai ȩl kirel a Toki.

415
 17 Processes of Sentence Formation:

‘I’m cleaning the house for Toki.’

b. Ak mȩruul aika ȩl kiriu.

‘I’m doing these things for you.’

c. A rȩngalȩk ȩr a skuul a mȩsuub ȩ l kirel a test.

‘The students are studying for the test.’

d. Ak mo mȩruul a kall ȩl kirel a party.

‘I’m going to make food for the party.’

e. A Droteo a mle suebȩk a rȩngul ȩl kirel a test.

‘Droteo was worried about the test.’

f. Aki milȩngȩtmokl ȩr a bȩluu ȩl kirel a eisei.

‘We were cleaning up our village for the sake of proper
sanitation.’

d. The Four Possessed Nouns and Various Tenses.

As we saw in 4a–b above, Palauan equational sentences in

the past tense contain the auxiliary word mle ‘was, were’. This
auxiliary is also used to indicate the past tense in sentences con-
taining the possessed forms of soal, chȩtil, sȩbȩchel, and kirel,
since such sentences are basically of the equational type (cf. our
discussion in 17.2). Note, therefore, the examples below:

(62) a. Ng mle soak ȩl mo ȩr a chei.

‘I wanted to go fishing.’

b. A Toki a mle chȩtil ȩl mȩruul a kall.

‘Toki didn’t want to prepare the food.’

c. Ng mle sȩbȩchem ȩl mo milil er a elii?

‘Were you able/allowed to go out and play yesterday?’

d. Ng mle kiram ȩl mȩsuub.

‘We had to/were supposed to study.’

e. A Droteo a mle kirel ȩl oureor ȩr a Guam, e ng di ng mlo

ȩr a Hawaii.
‘Droteo was supposed to work in Guam, but he went to
Hawaii (instead).’

416
 Palauan Reference Grammar

In order to indicate the future tense in an equational sen-

tence, we use the directional verb mo ‘go’ as an auxiliary. In
such cases, use of mo not only designates a future event but also
implies a change of state (cf. 13.5.1). Observe the following
equational sentences in the future tense:

(63) a. A ngȩlȩkek a mo sensei.

‘My child is going to become a teacher.’

b. A Toki ng mo chad ȩr a Merikel er

oingarang?
‘When is Toki becoming an American
citizen?’

Since sentences with soal, chȩtil, sȩbȩchel, and kirel are equa-
tional sentences, they too use mo ‘go’ as an auxiliary to indicate
future tense. The following examples are typical:

(64) a. Ng mo sȩbȩchem ȩl me ȩr a blik ȩr a

klukuk?
‘Will you be able to come to my house
tomorrow?’

b. Ng mo soam ȩl mȩsuub ȩr a klukuk?

‘Will you be wanting to study tomorrow?’

To express a past change of state in equational sentences,

we use mlo or mla mo (for relatively remote vs. recent past, re-
spectively), as in the sentences below:

(65) a. Kȩ mlo sensei er oingarang?

‘When did you become a teacher?’

b. A ngȩlȩkel a Toki a mla mo padre.

‘Toki’s child has become a priest.’

Equational sentences containing the four possessed nouns

under discussion also use mlo and mla mo to designate changes
of state in the past, as illustrated below:

(66) a. Ng mlo soak a sasimi er se ȩr a kngar ȩr a

Siabal.
‘I got to like sashimi when I was in Japan.’

417
 17 Processes of Sentence Formation:

b. A Toki a mla mo chȩtil ȩl mȩsuub a ochur.

‘Toki has gotten to dislike studying math.’

c. A Satsko a mlo sȩbȩchel ȩl mo ȩr a Guam.

‘Satsko had the opportunity to go to
Guam.’

d. Ng mla mo kirek ȩl mo rȩmei.

‘It’s gotten to the point where I have to go
home.’

418
 18 Negation

18Negation
18.1. AFFIRMATIVE VS. NEGATIVE SENTENCES
The sentences of Palauan, like those of every language, can be
classified into affirmative and negative types. While an affir-
mative sentence asserts (or affirms) the occurrence of some
action, event, state, condition, etc., a negative sentence denies
such occurrence. In other words, a negative sentence makes a
statement of the sort ‘someone is not doing something’, ‘such
and such is not the case’, ‘X is not Y’, ‘there isn’t/aren’t any Z’,
etc. Mostly all Palauan negative sentences contain some form of
the negative verb diak ‘isn’t, doesn’t exist’, which we will ex-
amine in detail in 18.3 below.
To familiarize ourselves with the idea of negation, let us
compare a few affirmative sentences with their negative
counterparts:

(1) a. A Toki a mȩnguiu ȩr a hong.

‘Toki is reading the book.’

b. A Toki a diak longuiu ȩr a hong.

‘Toki isn’t reading the book.’

(2) a. A ngȩlȩkek a smechȩr.

‘My child is sick.’

b. A ngȩlȩkek a diak lsechȩr.

‘My child isn’t sick.’

(3) a. A Droteo a sensei.

‘Droteo is a teacher.’

b. A Droteo a diak lsensei..

‘Droteo isn’t a teacher.’

It is easy to see that the b-sentences above are the denials (or,
in a sense, opposites) of the a-sentences. Thus, while la is a
transitive sentence which asserts that a particular agent or
doer (Toki) is performing a certain activity (reading the book),

419
 18 Negation

1b denies that the agent is engaged in that same activity. In

a parallel way, 2a is an intransitive sentence in which the
subject (ngȩlȩkek ‘my child’) is asserted to be in a particular
state (smechȩr ‘sick’), while 2b denies that the subject is in this
state. Finally, 3a is an equational sentence (see 18.6 below)
which asserts a relationship of equivalency between two noun
phrases (Droteo and sensei ‘teacher’), while 3b denies this rela-
tionship.
In each of the b-sentences above, the verb or noun directly
following the negative verb diak must be prefixed with a hypo
thetical pronoun (cf. 4.10) which agrees in person and
number with the agent or doer (if the sentence is transitive,
as in 1) or the subject (if the sentence is intransitive or equa-
tional, as in 2 and 3). Thus, in 1b and 2b we observe the hy-
pothetical verb forms longuiu and lsechȩr; longuiu is derived
by replacing the verb marker mȩ- of the corresponding imper-
fective transitive verb mȩnguiu ‘read’ with the 3rd pers. sg.
hypothetical pronoun prefix lo-, and lsechȩr is formed by pre-
fixing the reduced variant l- of the 3rd pers. sg. hypothetical
pronoun (cf. 4.10.4) to sechȩr, which is the stem of the intran-
sitive state verb smechȩr ‘sick’. In 3b, the noun directly fol-
lowing diak—namely, sensei ‘teacher’—also shows the reduced
variant l- of the 3rd pers. sg. hypothetical pronoun. The appear-
ance of the hypothetical pronouns after diak in the b-sentences
above can be explained in a very straightforward way, as we will
see in 18.4 below.

18.2. AFFIRMATIVE VS. NEGATIVE EXPRESSIONS

OF EXISTENCE
When a speaker of Palauan wishes to introduce a piece of infor-
mation into a conversation for the first time, he will often use
sentences of the following kind:

(4) a. Ng ngar ȩr ngii a oles ȩr a chȩlsel a skidas.

‘There’s a knife inside the drawer.’

b. Ng ngar ȩr ngii a mlik.

‘I have a car.’

c. Ng mla ȩr ngii a ududel a Toki.

‘Toki had money.’

420
 Palauan Reference Grammar

d. Ng mla ȩr ngii a blai ȩr tiang.

‘There used to be a house here.’

e. Ng mla ȩr ngii a ilumȩl ȩr a party.

‘There were drinks at the party.’

The sentences above are used when the speaker wants to assert
the existence of something which he believes represents new
information for the hearer. In other words, the italicized noun
phrases of 4 introduce the hearer for the first time to (the ex-
istence of) the items in question. Thus, if a speaker utters 4a,
for example, his hearer is presumably finding out for the first
time about the existence of a knife (oles) in the drawer. For this
reason, 4a would appear naturally in a dialog such as the fol-
lowing:

(5) A: Ngara a ngar ȩr ngii ȩr a chȩlsel a skidas?

‘What is there inside the drawer?’

B: Ng ngar ȩr ngii a oles (ȩr a chȩlsel a skidas).

‘There’s a knife (inside the drawer).’

A’s question implies that he does not know what is inside the
drawer; therefore, it is obvious that oles ‘knife’ in B’s response
constitutes a new piece of information. Because A’s question has
already specified the location involved, the parenthesized lo-
cation al phrase ȩr a chȩlsel a skidas ‘inside the drawer’ (cf.
14.2.1–2) may be omitted in B’s response.
Now, sentence 4a should be distinguished from the fol-
lowing:

(6) A oles a ngar ȩr a chȩlsel a skidas.

‘The knife is inside the drawer.’

While oles ‘knife’ is new information for the hearer in 4a, in 6

this same noun represents old information to which the hearer
has already been introduced. In other words, 6 can only be used
when the identity of oles ‘knife’ is clear—i.e., when both speaker
and hearer know what particular knife they are talking about.
For this reason, 6 would be used in a dialog such as the fol-
lowing:

421
 18 Negation

(7) A: A oles ng ngar ȩr ker?

‘Where is the knife?’

B: A oles 1 a ngar ȩr a chȩlsel a skidas.

‘The knife is inside the drawer.’

Since oles in 7B is old information, the new information which

this sentence conveys must be represented by the locational
phrase ȩr a chȩlsel a skidas ‘inside the drawer’. This is in fact
the case, since 7A is a question asking for information about the
location of the knife.
Sentences like 4a–e, which assert the existence of some-
thing or introduce something into a conversation as new infor-
mation, are called affirmative expressions of existence. Such
sentences always contain some form of the special sequence
ngar ȩr ngii ‘there is/are’. As examples 4c–e show, ngar ȩr ngii
‘there is/are’ changes to mla ȩr ngii ‘there was/were’ in the
past tense. These sequences are rather difficult to analyze be-
cause their meaning cannot be readily explained in terms of
their form. They appear to be a combination of the existential
verb ngar ‘exist, be (located)’ (past: mla ‘existed, was (located)’
2
) and the relational phrase ȩr ngii. As we will see below, this
relational phrase, which consists of the relational word ȩr fol-
lowed by the 3rd pers. sg. emphatic pronoun ngii, is probably
a kind of locational phrase (cf. 14.2). Therefore, the literal
meaning of ngar ȩr ngii and mla ȩr ngii seems to be something
like ‘exists in it’ and ‘existed in it’, respectively. Because of this
difficulty of analysis, it is perhaps better to think of ngar ȩr ngii
and mla ȩr ngii as single, indivisible units with the meaning
‘there is/are’ or ‘there was/were’. 3 A similar problem is ob-
served for the future tense form of ngar ȩr ngii ‘there is/are’,
which is mo ȩr ngii ‘there will be’. Though best dealt with as a
single unit, this sequence most likely consists of the future tense
auxiliary mo ‘go’ and the relational phrase ȩr ngii.
The state verb ngar which is found in the sequences ngar
ȩr ngii and mla ȩr ngii is normally used as an existential
verb which denotes the existence of someone or something in
a particular location. As in 6 above, the existential verb ngar
‘exist, be (located)’ always occurs in sentences followed by a
locational phrase, which simply tells us where someone or
something is. This locational phrase has been italicized in the
examples below:

422
 Palauan Reference Grammar

(8) a. A mlik a ngar ȩ r a mȩdal a blai.

‘My car is in front of the house.’

b. A Ngchesar a ngar ȩr a Babȩldaob.

‘Ngchesar is located on Babeldaob.’

c. A Helen a mla ȩr a bitang.

‘Helen was next door.’

Because ngar is always followed by a locational phrase, 4 we

tentatively proposed above that the relational phrase in ngar ȩr
ngii and mla ȩ r ngii is of the locational type. The existential
verb ngar is unusual in that it has the irregular past tense form
mla ‘existed, was (located)’. This form is probably closely re-
lated to the auxiliary word mla, which is used to denote recent
past time or past experience (cf. 5.3.2.1).
The following sentences are additional examples of affir-
mative expressions of existence:

(9) a. Ng ngar ȩr ngii a {hong ȩr ngak/uldȩsuek}.

‘I have {a book/an idea}.’

b. Ng ngar ȩr ngii a kȩrim?

‘Do you have a question?’

c. Ng mo ȩr ngii a ochȩraol ȩr a klukuk.

‘There will be a money-raising party
tomorrow.’

d. Ng mo ȩr ngii a sukal ȩr a imȩlem?

‘Will you take sugar in your drink?’

e. Ng mochu ȩr ngii a chull.

‘It’s about to rain.’

f. Ng mla ȩr ngii a temel a Toki ȩl mo ȩr a

party.
‘Toki had time to go to the party.’

In 9e, mochu is the predictive form of mo ‘go’ (cf. 11.12.5).

Therefore, the sequence mochu ȩr ngii means something like
‘there is about to be’.

423
 18 Negation

As we might expect, all of the affirmative expressions of

existence discussed above can be transformed into negative
ex pressions of existence by using some form of the negative
verb diak ‘isn’t, doesn’t exist’. Like their affirmative counter-
parts, negative expressions of existence introduce something
into a conversation as new information, but at the same time
they deny the existence of this particular thing. In the list
below, we give the negative expressions of existence which cor-
respond to some of the sentences of 4 and 9 above:

(4b’) Ng diak a mlik.

‘I don’t have a car.’

(4c’) Ng dimlak a ududel a Toki.

‘Toki didn’t have any money.’

(4d’) Ng dimlak a blai ȩr tiang.

‘There wasn’t any house here.’

(4e’) Ng dimlak a ilumȩl ȩr a party.

‘There weren’t any drinks at the party.’

(9b’) Ng diak a kȩrim?

‘Don’t you have any questions?’

(9f’) Ng dimlak a temel a Toki ȩl mo ȩr a party.

‘Toki didn’t have any time to go to the party.’

As examples 4c’, 4d’, 4e’, and 9f’ show, the past tense form of
the negative verb diak is dimlak ‘wasn’t, didn’t exist’. Before
analyzing dimlak and other related forms of the negative verb,
we will first concern ourselves with the grammatical structure
of the various affirmative and negative expressions of existence
observed in this section.

18.2.1. SUBJECT SHIFTING IN AFFIRMATIVE AND NEGATIVE

EXPRESSIONS OF EXISTENCE
All of the affirmative and negative expressions of existence pre-
sented in 18.2 above are identical in over-all structure: first,
all of these sentences begin with the 3rd pers. sg. nonemphatic
pronoun ng; second, this ng is immediately followed by verb
phrases containing ngar (ȩ r ngii) or diak; and finally, the

424
 Palauan Reference Grammar

subject of the sentence—namely, the thing whose existence is

being asserted or denied—appears directly after the verb
phrase. These are precisely the three major features which
identify sentences that have been derived by the process of
subject shifting (cf. 17.2). Thus, we propose that all affir-
mative and negative expressions of existence are derived by
subject shifting from source sentences of the form

(10) subject noun phrase + {ngar ȩr ngii/diak}

Therefore, sentences 4b, 4c, 4b’, and 9f’ would be derived

according to the following scheme:

(11) Source Sentence Resulting Sentence

a. A mlik a ngar ȩr ngii. → Ng ngar ȩr ngii a mlik.

b. A ududel a Toki a mla ȩr → Ng mla ȩr ngii a ududel a

ngii. Toki.
‘Toki had money.’

c. A mlik a diak. → Ng diak a mlik.

‘I don’t have a car.’

A temel a Toki ȩl mo ȩr a → Ng dimlak a temel a Toki ȩl

party a dimlak. mo ȩr a party.
‘Toki didn’t have any time
to go to the party.’

When the italicized (3rd person) subjects are moved to the right
of the verb phrase by the subject shifting rule, the pronominal
trace ng appears in their place. Because the source sentences
of 11 are not acceptable (or, at best, very awkward) to Palauan
speakers, we conclude that application of the subject shifting
rule is obligatory: in other words, the source sentences of 11
must be transformed into the resulting sentences of 11 in order
to become acceptable utterances of Palauan. 5
The resulting sentences of 11b and 11d can be further trans-
formed by the preposing of possessor rule (cf. 17.3). Thus,
the possessor Toki of both examples, which was moved to the
right of the verb phrase as part of the shifted subjects ududel
a Toki ‘Toki’s money’ and temel a Toki ȩl mo ȩr a party ‘Toki’s

425
 18 Negation

time to go to the party’, can be preposed to sentence-initial po-

sition, where it replaces the pronominal trace ng. We therefore
get the following sentences, which are identical in meaning:

(12) a. A Toki a mla ȩr ngii a ududel.

‘Toki had money.’

b. A Toki a dimlak a temel ȩl mo ȩr a party.

‘Toki didn’t have any time to go to the party.’

Unless we accept the validity of the subject shifting and

preposing of possessor rules, we have no reasonable way of
accounting for the order of words observed in sentences like
12a-b.
In our discussion above, we implied that any source sen-
tence of the form subject noun phrase + diak must undergo
obligatory application of the subject shifting rule. One notable
exception to this claim is found among sentences containing mo
+ diak, in which the auxiliary mo designates a change of state
(cf. 13.5). Observe the following examples, in which the source
sentences and resulting sentences have the same meaning:

(13) Source Sentence Resulting Sentence

a. A ududek a mla mo diak. → Ng mla mo diak a ududek.

‘My money has run out.’

b. A chull a mla mo diak. → Ng mla mo diak a chull.

‘The rain has stopped.’

c. A urerel a rubak a mlo → Ng mlo diak a urerel a

diak. rubak. 6
‘The old man lost his job.’

The sentences of 13 describe changes of state in which the ital-

icized subject noun phrases ududek ‘my money’, chull ‘rain’,
and urerel ‘his work’ became non-existent—i.e., stopped, dis-
appeared, became used up, etc. Thus, in spite of their English
equivalents, these examples really mean something like ‘My
money has become non-existent’, ‘The rain has become non-ex-
istent’, and ‘The old man’s work became non-existent’.

426
 Palauan Reference Grammar

Interestingly enough, the source sentences of 13 are per-

fectly acceptable to Palauan speakers, and therefore we must
conclude that the subject shifting rule is only optional in cases
like this. It is not very clear why subject shifting should be op-
tional in 13 but obligatory in 11. However, we can speculate
that subject shifting is obligatory only with affirmative and neg-
ative expressions of existence, as in 11. Therefore, one possible
reason why this rule is not obligatory in 13 would be that the
sentences of 13 somehow do not qualify as negative expressions
of existence. Now, recall that negative expressions of existence
introduce something into a conversation as new information
and at the same time deny its existence. The sentences of 13,
however, do not do this since in order to say that something has
become non-existent (has run out, disappeared, etc.), it is nec-
essary to assume that this very same thing existed in the first
place. Therefore, the italicized subjects of 13 do not introduce
new information, but designate things which were already pre-
sumed to be part of the hearer’s knowledge. Instead, the new
information in the sentences of 13 is actually conveyed by the
change of state expressions themselves. Thus, in 13a, for ex-
ample, the fact that the speaker has money (ududek) is not new
information, but the fact that his money ran out (mla mo diak)
is. In this way, then, the sentences of 13 probably do not qualify
as negative expressions of existence, and therefore the subject
shifting rule need not apply to them obligatorily.

18.3. THE NEGATIVE VERB DIAK

Now that we have seen how the negative verb diak is used in
negative expressions of existence, let us examine the various
forms which it can take. The negative verb diak is best classified
as an intransitive state verb. Because it is a state verb, it can
be used together with the auxiliary mo ‘go’ to denote a change
of state, as we saw at the end of the preceding section.
The past tense form of diak ‘isn’t, doesn’t exist’ is dimlak
‘wasn’t, didn’t exist’. At first glance, the form dimlak seems
very difficult to explain, since it appears to contain an unusual
infixed variant of the past tense marker—namely, -ml-. But if
we assume that diak is a state verb, at least some of the ap-
parent irregularity is resolved. Now, since diak is a state verb,
we would expect that its past tense would be formed like that
of all other state verbs—i.e., by using the auxiliary mle ‘was,

427
 18 Negation

were’ (cf. 5.1.3). Thus, we would expect the past of diak to be

mle diak. Even though mle diak does not occur, it nevertheless
seems to be a plausible source for dimlak: here, a special type
of metathesis (cf. 6.2) takes place in which the whole word mle
exchanges positions with the first syllable di of diak. The re-
sulting sequence is di-mle-ak, which then becomes dimlak after
deletion of the e. This appears to be the only possible way of ac-
counting for the mysterious -ml- in dimlak.
There is further evidence that the negative verb diak com-
bines with other words (or morphemes) in strange ways. Thus,
we also observe the negative word dirkak ‘not yet, not ever’,
which is used to express the fact that some action or event has
so far not taken place. The following pair of sentences gives us
a clue about the structure of dirkak:

(14) a. Ak dirk mȩnguiu ȩr a hong.

‘I’m still reading the book.’

b. Ng dirkak kunguiu ȩr a hong.

‘I haven’t read the book yet.’

In 14a, the qualifying word dirk ‘still’ (see 24.6) affirms that
the subject (ak ‘I’) has been reading the book over some period
of time and is continuing to read it at the present moment. In
14b, however, dirkak denies that the subject has ever read the
book (and of course implies that he is not reading it now). In
other words, dirkak describes a kind of state characterized by
the fact that the subject has still not gotten around to reading
the book. When viewed this way, 14b seems to be a denial (or
opposite) of 14a and as such should contain the negative verb
diak. We propose that 14b indeed does contain diak, but as part
of the word dirkak. In other words, the source of dirkak in 14b
seems to be the sequence dirk + diak, which phonetically be-
comes dirkak by deletion of the second occurrence of the syl-
lable di. Notice that the negative verb dirkak ‘not yet, not ever’
in 14b is followed by the hypothetical verb form kunguiu, in
which the 1st pers. sg. hypothetical pronoun prefix ku- (corre-
sponding to the agent—i.e., the person reading the book) has
replaced the verb marker mȩ- of imperfective mȩnguiu ‘read’.
We should distinguish carefully between the meanings of
dimlak ‘wasn’t, didn’t exist’ and dirkak ‘not yet, not ever’, which
are contrasted in the pairs of sentences below:

428
 Palauan Reference Grammar

(15) a. Ng dimlak kbo ȩr a Guam.

‘I didn’t go to Guam.’

b. Ng dirkak kbo ȩr a Guam.

‘I haven’t ever gone to Guam.’

(16) a. A Toki a dimlak loruul a kall ȩr a

Sina.
‘Toki didn’t make Chinese food.’

b. A Toki a dirkak loruul a kall ȩr a

Sina.
‘Toki hasn’t ever made Chinese
food.’

In the a-sentences above, use of dimlak—the past tense form of

diak—refers to something which did not happen on a single,
specific occasion. By contrast, use of dirkak in the b-sentences
implies that something failed to take place repeatedly or on
many occasions. For this reason, the b-sentences are inter-
preted to mean that someone has never had the experience
of doing something. Therefore, sentences like 15b and 16b are
common answers to questions about past experience, which
contain the auxiliary mla (cf. 5.3.2.1). Note the following di-
alogs:

(17) A: Kȩ mla mo ȩr a Guam?

‘Have you ever gone to Guam?’

B: Ng diak. Ng dirkak kbong.

‘No, I haven’t.’

(18) A: A Toki ng mla mȩruul a kall ȩr a Sina?

‘Has Toki ever made Chinese food?’

B: Ng diak. Ng dirkak loruul.

‘No, she hasn’t.’

As in 14b, the negative verb dirkak can also refer to some

event which as of the present moment has not yet occurred. In
such cases, dirkak corresponds to ‘not yet’, as in the sentences
below:

429
 18 Negation

(19) a. A ngalȩk a dirkak lȩbo lȩmȩchiuaiu.

‘The child hasn’t gone to sleep yet.’

b. Ng dirkak kbo kmȩrek ȩr a subȩlek.

‘I haven’t finished my homework
yet.’

As we will see in 19.1.3, the negative verb diak can itself

have a hypothetical form—namely, lak. This form, too, exhibits
the unusual phonetic nature of diak, since we have every reason
to believe that lak has its source in lȩ + diak, which consists of
the 3rd pers. sg. hypothetical pronoun lȩ- followed by the neg-
ative verb. In deriving lak, we delete the first syllable di of diak.
A similar deletion is observed in the past tense hypothetical
form lȩmlak, which is clearly derived from lȩ + dimlak.

18.4. HYPOTHETICAL VERB FORMS FOLLOWING

DIAK
At the beginning of 18.1 above, we observed negative sentences
like the following (= 1b and 2b):

(20) a. A Toki a diak longuiu ȩr a hong.

‘Toki isn’t reading the book.’

b. A ngȩlȩkek a diak lsechȩr.

‘My child isn’t sick.’

As mentioned in 18.1, the negative verb diak is followed by

hypothetical verb forms in the examples of 20. In the tran-
sitive sentence 20a, the lo- of longuiu refers to the doer or agent
(Toki), and in the intransitive sentence 20b, the l- of lsechȩr
refers to the subject (ngȩlȩkek ‘my child’). Now, with 20a–b
compare the following sentences, in which the doer or subject
corresponds to the speaker:

(21) a. Ng diak kunguiu ȩr a hong.

‘I’m not reading the book.’

b. Ng diak ksechȩr.
‘I’m not sick.’

430
 Palauan Reference Grammar

The hypothetical verb forms of 21 have 1st pers. sg. hypo-

thetical pronoun prefixes, while those of 20 have 3rd pers. sg.
hypothetical pronoun prefixes. In addition, the examples of 20
show the specific noun phrases Toki and ngȩlȩkek ‘my child’ in
sentence-initial position.
The structure of the negative sentences in 20–21 can be
easily explained in terms of the subject shifting rule and
two rules which are peculiar to sentences containing diak. In
18.2.1 above we saw that sentences which constitute negative
expressions of existence are derived simply by applying the
subject shifting rule (obligatorily) to source sentences of the
form subject noun phrase + diak. We propose that the negative
sentences of 20–21 are derived in exactly the same way, except
that in their source sentences the subject noun phrase is ac-
tually a bracketed sentence. Our method of analysis here is
identical to that used in 17.7, where we proposed that the pos-
sessed nouns soal ‘his liking’ and chȩtil ‘his disliking’ can occur
in two types of source sentences—namely, subject noun phrase
+ soal/chȩtil or [sentence] + soal/chȩtil. Recall that when the
subject of soal or chȩtil is a bracketed sentence, subject shifting
results in a dependent clause construction, as in the following
derivation:

(22)Source Sentence Resulting Sentence

[Ak mȩlim a biang] a → Ng soak ȩl mȩlim a

soak. biang.
‘I want to drink some
beer.’

Let us first analyze the sentences of 21, which we propose

have the following source sentences:

(23) a. [Ak mȩnguiu ȩr a hong] a diak.

(‘I’m not reading the book.’)

b. [Ak smechȩr] a diak.

(‘I’m not sick.’)

In the source sentences of 23, the negative verb diak is pre-

ceded by a subject noun phrase consisting of a complete
bracketed sentence; the bracketed sentence names an action

431
 18 Negation

or state whose occurrence is being denied. The subject shifting

rule applies obligatorily to 23a-b, giving the following struc-
tures:

(24) a. Ng diak [ak mȩnguiu ȩr a hong].

(‘I’m not reading the book.’)

b. Ng diak [ak smechȩr].

(‘I’m not sick.’)

When the bracketed sentence is moved to the right of the verb

phrase by the subject shifting rule, the pronominal trace ng
automatically appears in its place in sentence-initial position.
Now in order to transform 24a–b into the actually-spoken sen-
tences 21a–b, we need to apply a rule which derives the correct
hypo thetical verb forms in the shifted bracketed sentences.
This rule involves transforming the non-emphatic pronoun ak ‘I’
(the agent or subject of the bracketed sentence) into the cor-
responding hypothetical pronoun ku- or k-, and prefixing it to
the directly following verb, thereby deriving a hypothetical verb
form. As part of this process, the verb following ak changes,
too (cf. 6.2.1): thus, in 24a the verb marker mȩ- of imperfective
mȩnguiu is lost and replaced by ku-, and in 24b the metathe-
sized verb marker -m- of the state verb smechȩr is lost when k-
is prefixed. As a result of these changes, the structures of 24a–b
are transformed into the acceptable sentences of 21a–b.
The appearance of hypothetical verb forms in the shifted
bracketed sentences following diak can be easily understood
if we consider that these forms are used in a large variety
of grammatical constructions to express hypothetical events
or situations—i.e., ones which do not really occur but which
are supposed, assumed, imagined, wished for, etc. Because the
negative verb diak denies the occurrence of something, any
event or situation described in a sentence with diak would be
unreal in the sense that it did not occur. For this reason, the
appearance of hypothetical verb forms after diak seems rather
natural and “logical”.
Now let us return to the negative sentences of 20a–b, in
which a specific noun phrase (Toki in 20a and ngȩlȩkek ‘my
child’ in 20b) is found in sentence-initial position. We propose
that the source sentences for 20a–b are as follows:

432
 Palauan Reference Grammar

(25) a. [A Toki a mȩnguiu ȩr a hong] a diak.

(‘Toki isn’t reading the book.’)

b. [A ngȩlȩkek a smechȩr] a diak.

(‘My child isn’t sick.’)

Applying the subject shifting rule to 25a–b, we get the following

structures:

(26) a. Ng diak [a Toki a mȩnguiu ȩr a

hong].
(‘Toki isn’t reading the book.’)

b. Ng diak [a ngȩlȩkek a smechȩr].

(‘My child isn’t sick.’)

Because the shifted bracketed sentences of 26a–b contain a

specific 3rd pers. agent (Toki in 26a) or subject (ngȩlȩkek ‘my
child’ in 26b), they must be transformed by a special rule which
is based on the principle that the specific 3rd pers. agent or
subject cannot remain in the initial position of the bracketed
sentence. In other words, the noun phrases Toki and ngȩlȩkek
must be either shifted to sentence-final position or preposed
to sentence-initial position. At the same time, the verb form of
the bracketed shifted sentence must become hypothetical; the
resulting hypothetical verb form takes the 3rd pers. sg. hypo-
thetical pronoun prefixes lo- or l-, which agree with the agent
or subject of the bracketed sentence.
If the specific 3rd pers. agent or subject is shifted to
sentence-final position, 26a–b will be transformed into the fol-
lowing sentences:

(27) a. Ng diak longuiu ȩr a hong a Toki.

‘Toki isn’t reading the book.’

b. Ng diak lsechȩr a ngȩlȩkek.

‘My child isn’t sick.’

Though grammatical, examples 27a–b are used less frequently

than 20a–b above, in which the specific 3rd pers. agent or
subject has been preposed to sentence-initial position, where it
replaces the pronominal trace ng. These examples are repeated
as 28a–b below:

433
 18 Negation

(28) a. A Toki a diak longuiu ȩr a hong.

‘Toki isn’t reading the book.’

b. A ngȩlȩkek a diak lsechȩr.

‘My child isn’t sick.’

18.5. FURTHER EXAMPLES OF NEGATIVE

SENTENCES
In the negative sentences of 20 and 21 above, we observed two
different types of hypothetical verb forms following diak. Thus,
in 20a and 21a longuiu and kunguiu are hypothetical forms of
the imperfective (transitive) verb mȩnguiu ‘read’, while in
20b and 21b, lsechȩr and ksechȩr are hypothetical forms of
the (intransitive) state verb smechȩr ‘sick’. Before looking at
negative sentences containing other types of hypothetical verb
forms, let us examine some further sentences containing the
two types we have just mentioned:

(29) diak followed by hypothetical forms of imperfective

(transitive) verbs:

a. A Droteo a dimlak lolim a biang.

‘Droteo didn’t drink any beer.’

b. Ng diak molamȩch a dȩkool?

‘Don’t you smoke cigarettes?’

c. Ng dirkak kimoruul a kall ȩr a Siabal.

‘We’ve never made Japanese food.’

d. A rȩsȩchȩlik a dimlak longiis ȩr a kliokl.

‘My friends didn’t dig the hole.’

(30) diak followed by hypothetical forms of (intransitive)

state verbs:

a. A Toki a diak lsȩngȩrengȩr. 7

‘Toki isn’t hungry.’

b. A mlid a diak lȩklou.

‘Our car isn’t that big/big enough.’

434
 Palauan Reference Grammar

c. A mubi a dimlak lȩmȩkngit. 8

‘The movie wasn’t bad.’

d. Ng dimlak lȩmȩched a chei.

‘The tide wasn’t low.’

In the examples above, the tense of the whole sentence is de-

termined by the form of the negative verb (diak vs. dimlak
vs. dirkak). Therefore, the hypothetical verb form directly fol-
lowing diak always appears in the present tense, even though
the whole sentence may refer to past time (as in 29a, 29c–d, and
30c–d).
Since nearly all types of verbs can have hypothetical forms,
Palauan negative sentences are by no means confined to the
ones listed in 29–30 above. Thus, in the groups of examples
below, we observe diak followed by further types of hypothetical
verb forms:

(31) diak followed by hypothetical forms of intransitive

action verbs (including directional verbs 9 ):

a. A rȩngalȩk a diak loilil ȩr a sers.

‘The children aren’t playing in the garden.’

b. Ngara mȩ ng dimlak mlangȩl?

‘Why didn’t you cry?’

c. Ng diak chome ȩr a party?

‘Aren’t you coming to the party?’

d. Ng dimlak kbo ȩr a skuul er a elii.

‘I didn’t go to school yesterday.’

e. Ng dirkak kibo ȩr a Guam.

‘We’ve never gone to Guam.’

(32) diak followed by hypothetical forms of perfective

(transitive) verbs 10 :

a. Ng dimlak kbosii a babii.

‘I didn’t shoot the pig.’

435
 18 Negation

b. A Tony a dimlak lȩngȩsuir a Satsko ȩl mȩngȩtmokl ȩr a

blai.
‘Tony didn’t help Satsko clean the house.’

c. A Droteo a dimlak lleng a hong.

‘Droteo didn’t borrow the books.’

d. Ng dimlak kkȩrir a sensei ȩr a teng ȩr ngak.

‘I didn’t ask the teacher about my grade.’

(33) diak followed by hypothetical forms of the existential

state verb ngar:

a. A ngȩlȩkem a diak lȩngar ȩr a skuul.

‘Your child isn’t at school.’

b. A sensei a dimlak lȩngar ȩr tiang.

‘The teacher wasn’t here.’

(34) diak followed by hypothetical forms of ergative

verbs 11 :

a. A kall a dirkak lȩmȩruul.

‘The food hasn’t been made yet.’

b. A biang a dimlak lȩmȩngim.

‘The beer wasn’t drunk up.’

(35)diak followed by hypothetical forms of complex verb

phrases 12 :

a. A Toki a dirkak lȩbo lȩmȩrek ȩr a urerel.

‘Toki hasn’t finished her work yet.’

b. A ngȩlȩkek a diak lȩbo lungil ȩl smechȩr.

‘My child isn’t getting any better.’

c. Ng diak kbo kuruul a kall.

‘I’m not going to make the food.’

d. Ngara mȩ ng dimlak chobo mrei?

‘Why didn’t you go home?’

436
 Palauan Reference Grammar

18.6. EQUATIONAL SENTENCES: AFFIRMATIVE

AND NEGATIVE
The structure of Palauan equational sentences is relatively
simple because they merely consist of a subject noun phrase fol-
lowed by another noun phrase. The term equational sentence
is used because the two noun phrases involved are always un-
derstood as being equal or equivalent to each other. In other
words, if the two noun phrases in an equational sentence are
A and B, the equational sentence simply asserts that “A is
B”: the subject noun phrase is equated with or included in
the category of individuals or things designated by the second
noun phrase. Observe the following equational sentences in the
present tense:

(36) a. Ak ngalȩk ȩr a skuul.

‘I’m a student.’

b. Kȩdȩ chad ȩr a omȩnged.

‘We’re fishermen.’

c. A dȩmak a sensei.
‘My father’s a teacher.’

d. Tia a mlil a Toki.

‘This is Toki’s car.’

e. Tilȩcha a blai.
‘That’s a house.’

f. Ng mlik.
‘It’s my car.’

In order to derive sentences which deny that “A is B”, we

simply apply the rules of subject shifting and preposing (cf.
18.4) to source sentences of the form [equational sentence]
+ diak. Thus, in deriving the negative counterpart of 36a, we
begin with the following source sentence:

(37) [Ak ngalȩk ȩr a skuul] a diak.

(‘I’m not a student.’)

437
 18 Negation

The subject shifting rule applies to the whole bracketed sen-

tence, which is moved to the right of diak to give

(38) Ng diak [ak ngalȩk ȩr a skuul].

(‘I’m not a student.’)

To obtain a grammatical sentence, the subject ak ‘I’ of the

shifted bracketed sentence must be changed into a hypothetical
pronoun and prefixed to the following noun phrase. Thus, we
have

(39) Ng diak kngalȩk ȩr a skuul.

‘I’m not a student.’

If the subject of the shifted equational sentence is a specific

third person noun phrase, then it cannot remain in the initial
position of the equational sentence (cf. the discussion following
26a–b above). For example, when the source sentence for the
negative counterpart of 36c—namely,

(40) [A demak a sensei] a diak.

(‘My father’s not a teacher.’)

undergoes the subject shifting rule, we get the following struc-

ture:

(41) Ng diak [a dȩmak a sensei].

(‘My father’s not a teacher.’)

Now, the subject of the shifted bracketed sentence—dȩmak ‘my

father’—must either be moved to sentence-final position or pre-
posed to sentence-initial position. Applying one or the other
of these processes yields the following two grammatical sen-
tences:

(42) a. Ng diak lsensei a dȩmak.

‘My father isn’t a teacher.’

b. A dȩmak a diak lsensei.

‘My father isn’t a teacher.’

Most speakers prefer to use 42b, in which the noun phrase

dȩmak ‘my father’ has been preposed.

438
 Palauan Reference Grammar

The negative counterparts of the other examples of 36 are

derived according to the analysis presented above. Note, there-
fore, the following sentences:

(43) a. Ng diak dȩchad ȩr a omȩnged.

‘We’re not fishermen.’

b. Tia a diak lȩmlil a Toki.

‘This isn’t Toki’s car.’

c. Tilȩcha a diak lȩblai.

‘That’s not a house.’

d. Ng diak lȩmlik.
‘It’s not my car.’

Notice that the negative equational sentence 43d is different

in meaning and structure from the negative expression of ex-
istence 4b’, which we repeat here for convenience:

(44) Ng diak a mlik.

‘I don’t have a car.’

While 44 has a source sentence of the following form (cf. 11c),

(45) A mlik a diak.

(‘I don’t have a car.’)

the source sentence for 43d is different—namely,

(46) [Ng mlik] a diak.

(‘It’s not my car.’)

Equational sentences in the past tense use mle ‘was, were’

between the two noun phrases, while those in the future tense
use the auxiliary mo. The auxiliary mo can also designate a
change of state in equational sentences. Several examples are
given below:

(47) a. A John a mle sensei.

‘John was a teacher.’

b. Ak mo toktang.

439
 18 Negation

‘I’m going to be a doctor.’

c. A Toki a mlo chad ȩr a Merikel.

‘Toki became an American citizen.’

The negative counterparts of 47a–c are as follows:

(48) a. A John a dimlak lsensei.

‘John wasn’t a teacher.’

b. Ng diak kbo ktoktang.

‘I’m not going to be a doctor.’

c. A Toki a dimlak lȩbo lȩchad ȩr a Merikel.

‘Toki didn’t become an American citizen.’

Notice that in 48b–c, a hypothetical pronoun is prefixed both to

the auxiliary mo and to the noun phrase directly following it.

18.7. THE NEGATIVE EXPRESSION DI KEA

The negative expression di kea, which probably consists of the
word di ‘only, just’ followed by kea, corresponds to the English
expression ‘no longer’. In other words, di kea implies that some-
thing which was once the case is no longer the case. The fol-
lowing sentences with di kea are presented without explanation,
since their derivation parallels that of the various negative sen-
tences analyzed in 18.2.1 and 18.6 above:

(49) a. A Toki a di kea {lȩngalȩk ȩr a skuul/lȩkatungek}.

‘Toki is no longer {a student/my girlfriend}.’

b. A John a di kea lȩchad ȩr a Merikel.

‘John is no longer an American citizen.’

c. A blik a di kea lȩngar ȩr sei.

‘My house is no longer located there.’

d. Ng di kea a techȩllek ȩl mo ȩr a skuul.

‘I no longer have the opportunity to go to school.’

e. Ng di kea a ngikȩl.

440
 Palauan Reference Grammar

‘There’s no more fish.’

f. Ng di kea kureor ȩr a bangk.

‘I’m no longer working at the bank.’

The negative expression di kea can also express the idea that
some expected event failed to take place. In this usage, di
kea corresponds to English ‘not… after all’, as in the examples
below:

(50) a. Ng di kea kbo ȩr a Guam.

‘(It turns out that) I’m not going to Guam after all.”

b. Ng di kea kbo kureor ȩr a skuul.

‘(It turns out that) I’m not going to work at the school
after all.’

c. A Droteo a di mle kea lȩbo ȩr a mubi.

‘(It turned out that) Droteo didn’t go to the movies after
all.’

As example 50c shows, di kea becomes di mle kea in the past

tense. The addition of mle seems to indicate that kea functions
as a state verb, but this still does not give us any clues about
the (original or current) meaning of kea. 13

18.8. NEGATIVE VERBS AS ANSWERS TO

QUESTIONS
The Palauan negative verbs diak, dimlak, and dirkak are com-
monly used together with the 3rd pers. sg. non-emphatic
pronoun ng as answers to questions. This phenomenon is ob-
served in the dialogs below:

(51) A: A Droteo ng mȩsuub a tȩkoi ȩr a Merikel?

‘Is Droteo studying English?’

B: Ng diak.
‘No. (He’s not).’

(52) A: Kȩ mlo ȩr a party ȩr a kȩsus?

‘Did you go to the party last night?’

441
 18 Negation

B: Ng dimlak.
‘No. (I didn’t).’

(53) A: Ng ngar ȩr ngii a kȩrim?

‘Do you have any questions?’

B: Ng diak.
‘No. (I don’t).’

(54) A: Kȩ mla mȩnga a kall ȩr a Firiping?

‘Have you ever eaten Filipino food?’

B: Ng dirkak.
‘No. (I haven’t).’

The negative responses given by B in the dialogs above appear

to be short sentences which mean something like ‘it isn’t the
case’ (for ng diak), ‘it wasn’t the case’ (for ng dimlak), and ‘it
hasn’t (yet) been the case’ (for ng dirkak). They contrast, of
course, with the word chochoi ‘yes’, which is used as an affir-
mative re sponse.
Questions containing the negative verb diak are difficult to
analyze because they can be interpreted in at least two different
ways. Often, diak is simply used to add a degree of politeness
to questions which function as offers or invitations, as in the ex-
amples below:

(55) a. Ng diak monga a bobai?

‘Won’t you eat some papaya?’

b. Ng diak chome ȩr a blik?

‘Won’t you come to my house?’

Just like their English equivalents, the Palauan negative ques-

tions above are merely polite (or indirect) substitutes for the
blunter questions ‘Will you eat some papaya?’ and ‘Will you
come to my house?’ Because the examples of 55 are therefore
equivalent in meaning to affirmative questions, speakers would
respond to them as if they indeed were affirmative questions.
Thus, 55a, for instance, might occur in dialogs like the fol-
lowing:

442
 Palauan Reference Grammar

(56) A: Ng diak monga a bobai?

‘Won’t you eat some papaya?’ =
‘Will you eat some papaya?’

B: Chochoi.
‘Yes, I will.’

(57) A: Ng diak monga a bobai?

‘Won’t you eat some papaya?’ =
‘Will you eat some papaya?’

B: Ng diak. Ng chȩtik
‘No (thanks). I don’t like it.’

Now, it is also possible for an example like 55a to be inter-

preted as a general question rather than as an offer or an
invitation. Under such circumstances, A is really asking B a
question of the form ‘Is it the case that you don’t eat papayas?’
Because this question is interpreted in this way, B’s responses
to it are different in meaning from those observed in 56–7
above, as the following dialogs illustrate:

(58) A: Ng diak monga a bobai?

‘Don’t you eat papayas?’ =
‘Is it the case that you don’t eat papayas?’

B: Chochoi. (Ng diak kungang.)

‘No, I don’t. (I don’t eat them.)’

(59) A: Ng diak monga a bobai?

‘Don’t you eat papayas?’ =
‘Is it the case that you don’t eat papayas?’

B: Ng diak. (Ak mȩnga ȩr a bebil ȩr a taem.)

‘Yes, I do. (I eat them from time to time.)’

If we think of A’s question as really meaning ‘Is it the case that

you don’t eat papayas?’, as mentioned above, we can easily ex-
plain how B’s responses are interpreted. Thus, when B answers
chochoi in 58, he actually means ‘Yes, it is the case that I don’t
eat papayas’. Similarly, B’s use of ng diak in 59 corresponds to

443
 18 Negation

‘No, it is not the case that I don’t eat papayas—i.e., I do eat

them’. As the English equivalents show, the way of answering
such negative questions in English is quite different.

444
 19 Uses of Hypothetical Verb
Forms

19UsesofHypotheticalVerbForms
19.1. CONDITIONAL SENTENCES
As we saw in 18.4 , Palauan hypothetical verb forms are re-
quired following the negative verb diak ‘isn’t, doesn’t exist’ in
certain types of sentences. Hypothetical verb forms are not con-
fined to sentences with diak, however, but appear in a large va-
riety of grammatical constructions. Perhaps the most important
of these is the conditional sentence, which we will describe in
detail in this and the following sections.
Palauan conditional sentences consist basically of two
parts—a condition and a consequent—and express the idea
that if some event, action, state, etc., occurs, then something
else will happen. The event, action, state, etc. whose occurrence
is suggested or put forth as a possibility is the condition, while
the event which it would bring about or which would result from
it is called the consequent. The meaning of these two terms
will become clear from the following example:

(1) A lȩngar ȩr ngii a ududek, e ak mo ȩr a Guam.

‘If I had money, (then) I’d go to Guam.’

In the above conditional sentence, the condition a lȩngar ȩr

ngii a ududek ‘if I had money’ is followed by the consequent e
ak mo ȩr a Guam ‘then I’d go to Guam’. The consequent desig-
nates an event (going to Guam) which could take place only if
the preceding condition were met or “satisfied”—i.e., if the sit-
uation designated by a lȩngar ȩr ngii a ududek ‘if I had money’
were to become an actual fact. At the moment of utterance,
however, this condition has not been satisfied, and the speaker
is merely viewing the idea of having money as a possibility
which will hopefully become true.
In example 1 above, both the condition and the consequent
are expressed by clauses which contain their own subject and
verb. In the conditional clause a lȩngar ȩr ngii a ududek ‘if I
had money’, which is introduced by the word a ‘if’, 1 the noun

445
 19 Uses of Hypothetical Verb Forms

phrase ududek ‘my money’ is the subject of the hypothetical

verb form lȩngar ȩr ngii ‘if there existed’. 2 Following the condi-
tional clause is the consequential clause e ak mo ȩr a Guam
‘then I’d go to Guam’, where ak ‘I’ is the subject of the di-
rectional verb mo ‘go’. This clause is introduced by the word
e ‘(and) then’ (see 25.1), and unlike the preceding conditional
clause, its verb is not in the hypothetical form.
Since conditional clauses describe hypothetical or pos-
sible events or states—i.e., ones which are not real but which
instead are supposed, imagined, hoped for, etc.—it is quite un-
derstandable why they should contain hypothetical verb forms,
since such forms commonly designate unreal events or states.
Thus, the reason for using hypothetical forms in conditional
clauses is the same as that for using hypothetical forms in sen-
tences containing the negative verb diak (cf. 18.4), since in
the latter case, too, we are dealing with unreal—i.e., non-oc-
curring—events or states.

19.1.1. CONDITIONS IN THE PRESENT OR FUTURE

If the hypothetical verb form in a conditional clause is in the
present tense, then the events or states designated by the con-
dition and the consequent are imagined as occurring either at
the present moment or at some time point in the future. The
verb of the consequential clause may be in the present or future
tense, as the following examples show (cf. 1 above):

(2) a. A kudȩnge a tȩkoi ȩr a Siabal, e ak mȩrael ȩl mo ex a

Siabal.
‘If I knew Japanese, (then) I’d travel to Japan.’

b. A kbo ȩr a Guam er tia ȩl me ȩl rak, e ak mo kie ȩr a blil

a sȩchȩlik.
‘If I go to Guam next year, (then) I’ll stay at my friend’s
house.’

c. A kisa a John ȩr a klukuk, e ak dmu ȩr ngii.

‘If I see John tomorrow, (then) I’ll tell him.’

d. A lȩme a Droteo ȩr a klukuk, e ng me kie ȩr a blik.

‘If Droteo comes tomorrow, (then) he’ll stay at my house.’

e. A lȩbo lsechȩr a ngȩlȩkek, e ng diak lȩbo ȩr a skuul.

446
 Palauan Reference Grammar

‘If my child gets sick, (then) he won’t go to school.’

f. A lȩme a chull, e kȩ ngmai a sȩlȩkȩlek.

‘If it rains, (then) please bring in my laundry.’

g. A lȩbȩskak a udoud a dȩmak, e ak rullii a party.

‘If my father gave me money, (then) I’d have a party.’

If the subject (or agent) in a conditional clause is a specific

third person noun phrase, as in 2d–g, then it must occur to the
right of the (hypothetical) verb phrase. Thus, the position of the
specific third person noun phrases Droteo, ngȩlȩkek ‘my child’,
chull ‘rain’, and dȩmak ‘my father’ in 2d–g is due to a rule which
is rather similar to the subject shifting rule discussed in 17.2.
In the case of conditional clauses, the subject (or agent) must
be shifted obligatorily, since a sentence like the following (cf. 2f)
is ungrammatical:

(2f’) *A chull a lȩme, e kȩ ngmai a sȩlȩkȩlek.

19.1.2. CONDITIONS IN THE PAST

If the hypothetical verb form of the conditional clause and the
(non-hypothetical) verb form of the consequential clause are
both in the past tense, then the condition and the consequent
are imagined as having occurred at some time point in the past.
Thus, with 1, 2a, and 2g above, compare the following sen-
tences:

(3) a. A lȩbla ȩr ngii a ududek, e ak mlo ȩr a Guam.

‘If I had had money, (then) I would have gone to Guam.’

b. A kble kudȩnge a tȩkoi ȩr a Siabal, e ak mirrael ȩl mo ȩr

a Siabal.
‘If I had known Japanese, (then) I would have travelled to
Japan.’

c. A lȩbilskak a udoud a dȩmak, e ak rirȩllii a party.

‘If my father had given me money, (then) I would have
had a party.’

447
 19 Uses of Hypothetical Verb Forms

In 3a we see that mla—the past tense form of the existential

verb ngar (cf. 18.2)—appears as bla when a hypothetical
pronoun is prefixed; and in 3b the auxiliary mle (which functions
to mark the past tense with a state verb like mȩdȩnge
‘know’—cf. 5.1.2) has likewise changed to ble before the ad-
dition of the hypothetical pronoun. The alternation between m
and b observed here was discussed at length in 6.2.1.

19.1.3. CONDITIONAL CLAUSES CONTAINING diak

When the negative verb diak is used in a conditional clause, the
resulting sentence will mean something like ‘if such-and-such is
not/had not been the case, then…’ Observe the examples below,
in which a present or future condition is involved:

(4) a. A lak lȩbo a Droteo, e ng diak kbong.

‘If Droteo doesn’t go, (then) I won’t go.’

b. A lak losuub a Toki, e ng mo otsir ȩr a test.

‘If Toki doesn’t study, (then) she’ll fail the test.’

c. A lak a ududem, e ng diak chobo ȩr a mubi. 3

‘If you don’t have any money, (then) you won’t go to the
movies.’

Since the verb fololwing a ‘if’ in a conditional clause must

always be in the hypothetical form, we can conclude that lak is
the hypothetical form of the negative verb diak. It is likely that
lak is a shortened (or contracted) version of lȩ + diak, which
would be the expected hypothetical form (i.e., hypothetical
pronoun + verb stem) in the present tense. Can you explain why
lak is itself followed by hypothetical verb forms in examples 4a
and 4b?
Now consider the sentences below, which designate a past
condition:

(5) a. A lȩmlak a ududel a Droteo, e ng dimlak lȩbo ȩr a Guam.

‘If Droteo hadn’t had the money, (then) he wouldn’t have
gone to Guam.’

b. A lȩmlak lȩbo ȩr a bita mȩ llȩngir a oles a Toki, e ng

dimlak lsȩbȩchek ȩl rȩmuul a ngikȩl.

448
 Palauan Reference Grammar

‘If Toki hadn’t gone next door and borrowed a knife,

(then) I wouldn’t have been able to prepare the fish.’

In the conditional clauses of 5 we would expect the hypothetical

verb form to consist of the 3rd pers. sg. hypothetical pronoun
lȩ- followed by dimlak, the past tense form of the negative verb.
But instead of lȩ + dimlak, we observe the hypothetical form
lȩmlak. We conclude that in this case, too, a shortening or con-
traction has occurred, or—stated differently—the syllable di of
diak has been deleted.

19.2. PERMUTATION OF CONDITIONAL AND

CONSEQUENTIAL CLAUSES
Although the order conditional clause + consequential clause
is much preferred in the conditional sentences given above,
Palauan speakers sometimes permute the two clauses—i.e.,
they put the consequential clause first and the conditional
clause last. Thus, with 4c and 5a above, compare the following
examples, whose meaning is identical:

(4c’) Ng diak chobo ȩr a mubi a lak a ududem.

‘You won’t go to the movies if you don’t have any money.’

(5a’) A Droteo a dimlak lȩbo ȩr a Guam a lȩmlak a ududel.

‘Droteo wouldn’t have gone to Guam if he hadn’t had the
money.’

As the examples above show, a consequential clause which has

been moved to sentence-initial position is no longer introduced
by e ‘(and) then’.
There are certain types of conditional sentences in which
permutation of the conditional and consequential clauses is pre-
ferred, or even required. Observe, for example, the sentences
below, which are general questions about the way of doing
something:

(6) a. A rȩchad er a Belau tȩ mȩkȩra a loruul a bȩkai?

‘How do Palauans make pottery?’

b. Kȩ mȩkȩra a chomoruul a ilaot?

‘What do you do to make coconut syrup?’

449
 19 Uses of Hypothetical Verb Forms

General questions like 6a–b are usually phrased in the following

way: a consequential clause containing the special question
word mȩkȩra ‘do what?’ (see 20.7) precedes a conditional
clause which describes a particular activity. Thus, 6a–b actually
mean something like ‘What do Palauans do if they make
pottery?’ and ‘What do you do if you make coconut syrup?’ 4
Other types of sentences in which permutation of the condi-
tional and consequential clauses takes place will be discussed
in 19.3–4 below.

19.3. FURTHER TYPES OF CONDITIONAL

CLAUSES
The Palauan conditional clauses we have so far examined are
characterized by the following features: (i) the verb following
a ‘if’ must be in the hypothetical form, and (ii) a specific third
person subject must be shifted to the right of the verb phrase.
In this section, we will discuss three types of clauses which,
though conditional in meaning, do not exhibit the abovemen-
tioned features.
In order to express a future condition, it is possible to use a
conditional clause introduced by a lsȩkum ‘if’. Observe, for ex-
ample, the sentences below:

(7) a. A lsȩkum ak mo ȩr a Guam, e ak mo kie ȩr a blil a Tony.

‘If I go to Guam, (then) I’ll stay at Tony’s house.’

b. A lsȩkum a Droteo a mo ȩr a skuul ȩr a klukuk, e ng mo

omes ȩr a sensei.
‘If Droteo goes to school tomorrow, (then) he’ll see the
teacher.’

c. A lsȩkum ng diak a ududem, e ng diak lsȩbȩchem ȩl mo

ȩr a mubi.
‘If you don’t have money, (then) you can’t go to the
movies.’

d. A lsȩkum ng ungil a che, e tȩ mo ȩr a chei.

‘If the tide is good, (then) they’ll go fishing.’

450
 Palauan Reference Grammar

In the conditional clauses introduced by a lsȩkum ‘if’, the verb

occurs in its “normal” (i.e., non-hypothetical) form. Fur-
thermore, a specific third person subject need not be shifted to
the right of the verb phrase, as the position of Droteo in 7b illus-
trates. Although it is very difficult to analyze conditional clauses
with a lsȩkum ‘if’, we can speculate that the sequence a ls ȩkum
is itself a combination of a ‘if’ and a “fossilized” hypothetical
verb form lsȩkum (in which the l- appears to be the 3rd pers. sg.
hypothetical pronoun prefix). If lsȩkum is indeed a hypothetical
verb form, then this might explain why no further hypothetical
verb forms are required in the conditional clause.
For some speakers, conditional clauses with a lsȩkum are
interchangeable with those that contain a ‘if’ followed by a
hypothetical verb form (and a shifted third person subject, if
any). For others, however, the two types involve a rather fine
difference in meaning, which we will illustrate with the pairs
below:

(8) a. A kisa a John ȩr a klukuk, e ak subȩdii.

‘If/when I see John tomorrow, I’ll tell him.’

b. A lsȩkum ak mȩsa a John ȩr a klukuk, e ak subȩdii.

‘If I should possibly see John tomorrow, (then) I’ll tell
him.’

(9) a. A lȩme a Droteo, e ng me kie ȩr a blik.

‘If/when Droteo comes, he’ll stay at my house.’

b. A lsȩkum a Droteo a me, e ng me kie ȩr a blik.

‘If Droteo should possibly come, (then) he’ll stay at my
house.’

Though perhaps somewhat exaggerated, the English equiva-

lents in 8–9 above are designed to reflect the following dif-
ference in meaning between the a- and b-sentences. In the
b-sentences with a lsȩkum, the speaker is rather doubtful that
the condition and its consequent will become actual facts, while
in the a-sentences with a ‘if’ and a hypothetical verb form, the
speaker feels somewhat more confident that the condition and
its consequent will become true. For this reason, the a-sen-
tences can sometimes be translated with English ‘when’, which
implies that the future event is expected to occur.

451
 19 Uses of Hypothetical Verb Forms

Just like conditional clauses with a ‘if’ and a hypothetical

verb form, conditional clauses containing a lsȩkum can be per-
muted with a following consequential clause. Thus, the sen-
tence below is equivalent to 7d:

(10) Tȩ mo ȩr a che a lsȩkum ng ungil a chei.

‘They’ll go fishing if the tide is good.’

In order to express a present (or, sometimes, future) condi-

tion, Palauan speakers also make use of conditional clauses in-
troduced by ulȩkum ‘if (only)’. This word, which is probably
related in some way to the lsȩkum of a l sȩkum, is used when
the speaker wishes to emphasize how strongly he desires a
particular condition and its consequent to become true. When
ulȩkum is used with this connotation, the consequential clause
following it is introduced by mȩ ‘(and) so’ (see 25.1):

(11) a. Ulȩkum a sensei ȩr kȩmam a mo ȩr a Guam, mȩ ng mo

diak a klas.
‘If only our teacher would go to Guam, then we wouldn’t
have any class.’

b. Ulȩkum ng ngar ȩr ngii a ududek, mȩ ng mo sȩbȩchek

ȩl mo ȩr a Merikel.
‘If only I had some money, then I could go to America.’

As we can see, conditional clauses with ulȩkum are not intro-

duced by a. Furthermore, the verb form in such clauses is not
hypothetical, nor is the subject shifted. Most Palauan speakers
can use e instead of mȩ in the sentences of 11; for some, no dif-
ference in meaning results, while for others the connotation of
strong desire is lost. For the latter group of speakers, the sen-
tences of 11 with ulȩkum… e would be equivalent to sentences
with a lsȩkum… e.
Another commonly-used sentence type with ulȩkum is illus-
trated below:

(12) a. Ulȩkum ak kau, e ak mo ȩr a Merikel.

‘If only I were you, then I’d go to America.’

b. Ulȩkum ak sensei, e ak olisȩchakl ȩr a ochur.

‘If I could only be a teacher, then I’d teach math.’

452
 Palauan Reference Grammar

In the conditional clauses of 12, ulȩkum is followed by the se-

quences ak kau ‘I—you’ and ak sensei ‘I—teacher’, which are
actually equational sentences (cf. 18.6).
Palauan has yet another type of conditional clause which in
certain cases contrasts in meaning with the conditional clauses
already described. In the sentences below, we observe condi-
tional clauses introduced by a kmu ‘if’ 5 ; again, the verb in this
clause is not hypothetical, nor is the subject shifted. Clauses
with a kmu can refer to present, past, or future conditions:

(13) a. A kmu ak 6 mȩdȩnge a tȩkoi ȩr a Siabal, e ak mȩrael ȩl

mo ȩr a Siabal.
‘If I knew Japanese, (then) I’d travel to Japan.’

b. A kmu ng ngar ȩr ngii a ududek, e ak mȩchȩrar a

bȩchȩs ȩl mlai.
‘If I had money, (then) I’d buy a new car.’

c. A kmu a Droteo a mo ȩr a Guam ȩr a klukuk, e ng nguu

a Toki.
‘If Droteo were to go to Guam tomorrow, (then) he’d
take Toki.’

d. A kmu ak mle mȩdȩnge a tȩkoi ȩr a Siabal, e ak mirrael

ȩl mo ȩr a Siabal.
‘If I had known Japanese (at that time), I would have
travelled to Japan.’

e. A kmu ak mle kau, e ak mlong.

‘If I had been you, I would have gone.’

In 13a–c above, which designate present or future condi-

tions, the conditional clause with a kmu expresses a strong
belief or conviction on the speaker’s part that the condition and
its consequent will not become true. Thus, in 13c, for example, a
kmu a Droteo a mo ȩr a Guam ‘if Droteo were to go to Guam’ im-
plies that Droteo is not really expected to go to Guam, but nev-
ertheless the speaker is speculating what would happen if he
did. Because of this implication, conditional clauses with a kmu
differ subtly in meaning from conditional clauses with a lsȩkum
or a followed by a hypothetical verb form. Let us therefore
repeat 9a–b above as 14a–b and compare them with 14c:

453
 19 Uses of Hypothetical Verb Forms

(14) a. A lȩme a Droteo, e ng me kie ȩr a blik.

‘If/when Droteo comes, he’ll stay at my house.’

b. A lsȩkum a Droteo a me, e ng me kie ȩr a blik.

‘If Droteo should possibly come, (then) he’ll stay at my
house.’

c. A kmu a Droteo a me, e ng me kie ȩr a blik.

‘If Droteo were to come, he’d stay at my house.’

In the examples of 14, the speaker shows successively in-

creasing doubt about whether the condition and its consequent
have any chance of becoming true. Thus, the speaker uses a fol-
lowed by a hypothetical verb form (14a) if he believes there is
some reasonable possibility that Droteo will come. If he thinks
the possibility of Droteo’s coming is relatively small, however,
he will use a lsȩkum (14b). And if he thinks it is very unlikely
that Droteo will come, he will choose a kmu (14c), as mentioned
above.
In sentences designating past conditions, conditional
clauses with a kmu vs. those with a followed by a hypothetical
verb form result in different implications. Thus, compare 3b and
13d, which are both repeated below:

(15) a. A kble kudȩnge a tȩkoi ȩr a Siabal, e ak mirrael ȩl mo ȩr

a Siabal.
‘If I had known Japanese, (then) I would have travelled
to Japan.’

b. A kmu ak mle mȩdȩnge a tȩkoi ȩr a Siabal, e ak mirrael

ȩl mo ȩr a Siabal.
‘If I had known Japanese (at that time), I would have
travelled to Japan.’

According to some speakers, 15b implies that the present situa-

tion is different from that described in the past conditional
clause, while this is not necessarily the case in 15a. In other
words, 15b implies that the speaker in fact knows how to speak
Japanese now, whereas in 15a it is possible that the speaker still
does not know how to speak Japanese.

454
 Palauan Reference Grammar

19.4. ADDITIONAL EXAMPLES OF CONDITIONAL

CLAUSES
A few types of Palauan conditional sentences require special
mention because their English equivalents lack words like ‘if…
then’ and therefore tend to obscure the fact that the corre-
sponding Palauan sentences are really conditional. As we will
see below, this problem arises with certain conditional clauses
consisting of a and a following hypothetical form.
In 17.9 we observed that the possessed forms of soal ‘his
liking’ and chȩtil ‘his disliking’ can be followed by hypothetical
verb forms to convey the idea ‘X wants/does not want Y to do
something’. In such sentences, X is expressed by the possessor
suffix on soal or chȩtil and Y is identified by the pronominal
prefix on the hypothetical verb form. In addition, a specific noun
phrase may be mentioned if X or Y is a third person. Sentences
of this type are illustrated by the following:

(16) a. Ng soak a Droteo a longȩtmokl ȩr a delmȩrab.

‘I want Droteo to straighten up the room.’

b. Ng somam a chobo mrei.

‘We want you to go home.’

c. Ng soam a kungȩsbrebȩr ȩr a kbokb?

‘Do you want me to paint the wall?’

d. A Toki a chȩtil a rȩngalȩk a loilil ȩr a uum.

‘Toki doesn’t want the children to play in the kitchen.’

e. Ng chȩtik a chobo ȩr a party.

‘I don’t want you to go to the party.’

In spite of their English equivalents, the Palauan sentences in

16 are most likely conditional sentences in which the (itali-
cized) conditional clause has been permuted with the conse-
quential clause. Since conditional clauses express unreal ac-
tions, events, states, etc. rather than actual facts, we can see
why their use is appropriate in the examples above. Thus, in
sentences with soal, the conditional clause designates an action
or event which is desired or hoped for but which is not yet
real. In 16a, for instance, the event of the conditional
clause—namely, Droteo’s straightening up the room—has not

455
 19 Uses of Hypothetical Verb Forms

yet happened; therefore, the speaker is viewing Droteo’s

straightening up of the room as a possible event and saying
that he would be pleased if this possibility became an actual
fact. Sentences with chȩtil involve exactly the opposite, since
the conditional clause designates an action or event which is
not desired or hoped for—i.e., one which hopefully will not
become an actual fact. Thus, in 16d, the event of the conditional
clause—namely, the children’s playing in the kitchen—has pre-
sented itself as a possibility, but Toki does not want it to become
an actual fact (because it might inconvenience her, etc.).
Because the examples of 16 are really conditional sentences,
their word-for-word translation would be something like ‘We
would like it if you go home’ (for 16b) or ‘I won’t like it if you
go to the party’ (for 16e). The only unusual feature about the
italicized conditional clauses of 16 is that a specific third person
subject does not need to be shifted. Thus, in 16a and 16d Droteo
and rȩngalȩk ‘children’ appear at the beginning of the condi-
tional clause, directly following a ‘if’. 7
The transitive state verb mȩdakt ‘be afraid (of)’ can be used
with a conditional clause to express the idea ‘X is afraid that Y
will…’ Observe the examples below:

(17) a. A Droteo a mȩdakt ȩr a Toki a lȩbo lsechȩr.

‘Droteo is afraid that Toki will get sick.’

b. Ak kmal mle mȩdakt a kbo kotsir ȩr a test.

‘I was very much afraid that I would fail the test.’

c. A toktang a mȩdakt ȩr a John a lȩmad.

‘The doctor is afraid John will die.’

d. Ak mȩdakt a lȩbo lak 8 a ududek ȩl mo ȩr a Merikel.

‘I’m afraid I won’t have any money to go to America.’

Again, the examples of 17 are conditional sentences in which

the conditional clause containing a ‘if’ followed by a hypo-
thetical verb form has been permuted with the consequential
clause containing mȩdakt ‘be afraid (of)’. If a specific third
person noun phrase is mentioned—Toki of 17a or John of
17c—then it must appear as the object of mȩdakt. If we try

456
 Palauan Reference Grammar

to translate the sentences of 17 literally, we would get very

awkward and unacceptable English sentences such as (for 17a)
‘Droteo has fears about Toki, (wondering) if she will get sick.’
Time words like tutau ‘morning’, suelȩb ‘afternoon’, etc.,
can be used in conditional clauses to express the frequent or
habitual occurrence of a particular action. Consider the ex-
amples below:

(18) a. A bȩchik a oureor ȩr a sers a lȩtutau.

‘My wife works in the garden in the
morning.’

b. A Satsko a soal ȩl mȩsuub a lȩ suelȩb.

‘Satsko likes to study in the afternoon.’

c. Ak mȩruul a kall a lȩklȩbȩsei.

‘I prepare food in the evening.’

The italicized conditional clauses of 18 consist of a ‘if’ followed

by time words to which the 3rd pers. sg. hypothetical pronoun
lȩ- has been prefixed. Since lȩ - is the hypothetical pronoun
corresponding to ng, we can conclude that the conditional
clauses of 18 really consist of a followed by an equational sen-
tence (cf. 18.6). Therefore, these conditional clauses literally
mean something like ‘if it is morning’, ‘if it is afternoon’, etc.
By using a conditional clause like a lȩ tutau ‘in the morning’,
the speaker states that a particular event usually or normally
occurs at a designated time. The routine involved, however, is
not as regular or fixed as that implied by sentences exhibiting
a temporal phrase containing bek ‘each, every’ (cf. 14.6,
ex.34h). Thus, 18a is different in meaning from the example
below:

(19) A bȩchik a oureor ȩr a sers ȩr a bek ȩl tutau.

‘My wife works in the garden every morning.’

For further examples containing temporal phrases with bek

‘each, every’, cf. 14.6, ex. 35.

457
 19 Uses of Hypothetical Verb Forms

19.5. IMPERATIVE VERB FORMS

Palauan imperative verb forms are used to express orders or
commands. Because commands are ordinarily directed at the
person addressed—the “you” of the conversation—it is not sur-
prising that imperative verb forms involve second person pro
nouns. As we will see below, Palauan imperative verb forms
are actually nothing more than hypothetical verb forms prefixed
with the mo- or m- variants of the second person hypothetical
pronoun. We will use the separate term “imperative”, however,
as a convenient way of distinguishing the special usage under
discussion here.
Both the imperfective and perfective forms of transitive
action verbs can occur in commands. To derive the imperative
forms of imperfective verbs, we simply substitute the second
person hypothetical pronoun prefix mo- for the verb marker mȩ-
of the corresponding imperfective verb. A few such forms are
illustrated in the sentences below, which function as orders or
commands:

(20) a. Molim a kȩrum!

‘Drink your medicine!’

b. Mosilȩk ȩr a bilem!
‘Wash your clothes!’

c. Mongiis ȩr a kliokl!
‘Dig the hole!’

d. Monguiu ȩr tia ȩl hong!

‘Read this book!’

Because there is no distinction between singular and plural for

the second person hypothetical pronoun (cf. 4.10.1), the ex-
amples of 20 are ambiguous in that the speaker may be di-
recting the order either to just one person or to a group of two
or more persons.
The imperative forms of perfective verbs are derived with
the variant m- of the second person hypothetical pronoun. This
m- is always pronounced as a separate syllable—namely, [ṃ]
(cf. 1.3.5). All imperative forms of perfective verbs have the
structure hypothetical pronoun m-+ verb stem + object pro

458
 Palauan Reference Grammar

noun. Some typical examples are given in 21 below; in the

lefthand column, a 3rd pers. sg. object pronoun (-ii, -ir, etc.)
has been suffixed to the imperative verb form, while in the
righthand column the imperative verb form has the 3rd pers.
pl. non-hum. object pronoun (ø):

(21) 3rd pers. sg. non-hum, object 3rd pers. pl. non-hum,
object

a. Mngilmii a imȩlem! Mngim a imȩlem!

‘Have your drink!’ ‘Have your drinks!’

b. Msilȩkii a bail! Msilȩk a bail!

‘Wash the (piece of) ‘Wash the clothes!’
clothing!’

c. Mkiiȩsii a kliokl! Mkiis a kliokl!

‘Dig the hole (completely)!’ ‘Dig the holes (completely)!’

d. Mchiȩuii a hong! Mchuiu a hong!

‘Read the book ‘Read the books
(completely)!’ (completely)!’

e. Mlȩchȩsii a babier! Mluchȩs a babier!

‘Write the letter ‘Write the letters
(completely)!’ (completely)!’

f. Mkȩlii a ngikȩl! Mka a ngikȩl!

‘Eat up the fish!’ ‘Eat up (all) the fish!’

g. Mngȩtȩchii a mlai! Mngatȩch a mlai!

‘Clean up the car!’ ‘Clean up the cars!’

h. Mlȩngir a sebȩl! Mleng a sebȩl!

‘Borrow the shovel!’ ‘Borrow the shovels!’

i. Mdȩrur a ngikȩl! Mdul a ngikȩl!

‘Barbeque the fish!’ ‘Barbeque (all) the fish!’

The examples below are similar, except that the object pro-
noun suffixes refer to human beings:

459
 19 Uses of Hypothetical Verb Forms

(22) 1st or 3rd pers. sg. hum. 3rd pers. pl. hum. object
object

a. Mchȩlȩbȩdii a ngalȩk! Mchȩlȩbȩdȩtȩrir a

rȩngalȩk!
‘Hit the child!’ ‘Hit the children!’

b. Mkimdak! Mkimdȩtȩrir a rȩngalȩk!

‘Cut my hair!’ ‘Cut the children’s hair!’

c. Mtȩchȩlbii a Toki! Mtȩchȩlbȩtȩrir a rȩngalȩk!

‘Bathe Toki!’ ‘Bathe the children!’

d. Msiiȩkii a Satsko! Msiikȩtȩrir a rȩsȩchȩlim!’

‘Look for Satsko!’ ‘Look for your friends!’

Because the perfective imperative forms illustrated in 21–22

above are hypothetical verb forms, they of course do not contain
the verb marker in any of its several variants (cf, 6.2.1). In this
respect, they contrast with the (non-hypothetical) perfective
forms listed in 6.3, in which the metathesized verb marker
turns up as -(ȩ)m-, -o-, or -u-. A few of these contrasting forms
are pointed out in the list below:

(23) Perfective Imperative Form of Corresponding

21–22 ( = hypothetical pro Non-Hypotheti cal Perfective
noun m-+ verb stem + object Form (with metathesized
pronoun) verb marker italicized)

mchiȩuii chuiȩuii
‘Read it (completely)!’ ‘reads it (completely)’

mchuiu chȩmuiu
‘Read them (completely)!’ ‘reads them (completely)’

mkȩlii kolii
‘Eat it up!’ ‘eats it up’

mka kma
‘Eat them up!’ ‘eats them up’

mlȩngir longir
‘Borrow it!’ ‘borrows it’

460
 Palauan Reference Grammar

mleng lmneng
‘Borrow them!’ ‘borrows them’

mchȩlȩbȩdii cholȩbȩdii
‘Hit him!’ ‘hits him’

In order to express commands with intransitive action

verbs, a sequence of the form directional verb mo ‘go’ + in-
transitive action verb is frequently used. Observe the examples
below:

(24) a. Bo momȩngur!
‘Have your meal!’

b. Bo mdȩngchokl!
‘Sit down!’

c. Bo mdȩchor!
‘Stand/get up!’

d. Bo mrei!
‘Go home/get out!’

e. Bo mȩchiuaiu!
‘Go to sleep!’

f. Bo mngasȩch ȩr a bilas!
‘Get in the boat!’

g. Bo mkerd ȩr tiang 9 !
‘Get out here!’

h. Bo mtobȩd ȩr tiang!
‘Get out of here!’

Interestingly enough, the imperative form of mo ‘go’ is simply

the verb stem bo (cf. 6.2.1, exs. 11–12) rather than the ex-
pected *mbo—i.e., hypothetical pronoun + verb stem. The basic
form of imperative bo ‘go!’ may indeed be m + bo, but a pho-
netic rule deletes the initial m before a following b. 10 A similar

461
 19 Uses of Hypothetical Verb Forms

phenomenon is found among the imperative forms of various

transitive and intransitive verbs whose stems are b- initial, as
the following examples illustrate:

(25) a. Bosii a bȩlochȩl!

‘Shoot the pigeon!’

b. Bilii a ngalȩk!
‘Dress the child!’

c. Brȩchii a ngikȩl!
‘Spear the fish!’

d. Bȩskak a ududem!
‘Give me your money!’

e. Bȩkiis!
‘Get up/wake up!’

The perfective imperative forms of 25a–c are related to the

imperfective (transitive) verbs omoes ‘shoot’, omail ‘dress,
wrap’, and omurȩch ‘spear’, while bȩskak ‘give (it to) me’ of 25d
is related to perfective msa ‘give’. The non-hypothetical form
for bȩkiis ‘get up/wake up!’ of 25e is intransitive mȩkiis ‘get up/
wake up’.
Hypothetical verb forms with first or third person prefixes
are sometimes used with an imperative connotation when the
speaker feels something must be done by himself or someone
else. This connotation is observed in examples like the fol-
lowing:

(26) a. Kurael ȩl mo ȩr a blik.

‘I’d better go home.’

b. Bilii a ngalȩk e lorael.

‘Dress the child and have/let him go.’

c. Lȩbo ȩr a bita a Droteo mȩ lȩngai a kȩbui.

‘Have Droteo go next door and get some leaves for betel
nut chewing.’

d. Domȩngur ȩr tiang.
‘Let’s eat here.’

462
 Palauan Reference Grammar

Examples like 26d, in which the 1st pers. pl. inclusive hypo-
thetical pronoun do- (or dȩ-) is prefixed to a verb form, are com-
monly used to express the idea ‘let’s (do something)’. Further
examples will be provided in 19.6 below.
Palauan negative commands are formed simply by using
lak, the hypothetical form of diak (cf. 18.3), followed itself by
a hypothetical verb form with a prefixed second person hypo-
thetical pronoun (mo-). The following examples are typical:

(27) a. Lak molim a biang!

‘Don’t drink beer!’

b. Lak monga a kall!

‘Don’t eat the food!’

c. Lak mongȩrodȩch!
‘Stop making noise!’

d. Lak molȩkar ȩr a ngalȩk!

‘Don’t wake up the child!’

e. Lak mobes ȩl subȩdii a

Droteo!
‘Don’t forget to tell Droteo!’

It is interesting to note that in negative commands such as those

above, any transitive verb following lak can only appear in its
imperfective, but not perfective, form. In other words, a sen-
tence like the following is ungrammatical (cf. 27b):

(28) *Lak mkȩlii a kall!

(‘Don’t eat the food!’)

The reason why perfective forms are prevented after lak in sen-
tences like 27a–e seems to be the following: since a negative
command orders someone not to begin or continue a par-
ticular activity, it would be redundant and unnecessary to
mention finish ing that same activity, which is what a per-
fective verb form would imply.
Another way of expressing a negative command is to use a
statement introduced by ng diak, as in the examples below:

463
 19 Uses of Hypothetical Verb Forms

(29) a. Ng diak mongȩrodȩch!

‘Don’t make noise!’

b. Ng diak molȩkar ȩr a ngalȩk!

‘Don’t wake up the child!’

For some speakers, the examples of 29 differ in meaning from

the corresponding commands with lak. Thus, 27c might be a
command directed at children who have already begun to be
noisy, while 29a is a kind of warning which would be uttered
even before any noise has started. Similarly, 27d would be di-
rected at someone who has already begun to wake up the child,
whereas 29b would be a “precautionary” command spoken
while the child is still fully asleep.
If we recall that hypothetical verb forms are characteristi-
cally used to refer to unreal actions, events, etc., we can imme-
diately see why they are appropriate in imperative sentences.
Thus, when an order or command is given that something be
done, the action or event in question has not yet occurred
and is therefore unreal; indeed, the order or command is given
precisely so that the particular action or event will become an
actual fact. We can therefore see that the use of hypothetical
verb forms as imperatives has much in common with the use
of hypothetical verb forms in conditional sentences (cf. 19.1
above) and in negative sentences (cf. chap. 18).

19.6. PROPOSITIVE VERB FORMS

Palauan propositive verb forms are used when the speaker
wishes to propose or suggest that he and the hearer(s)
perform some action or activity together. Propositive verb forms
are actually hypothetical verb forms prefixed with the 1st pers.
pl. inclusive hypothetical pronoun do- or dȩ-. This pronoun is
used because the proposed or suggested action includes both
speaker and hearer(s). Palauan propositive verb forms have
English equivalents of the form ‘let’s (do something)’, as the ex-
amples below illustrate:

(30) a. Dorael!
‘Let’s go!’

b. Doilil ȩr tiang!

464
 Palauan Reference Grammar

‘Let’s play here!’

c. Dȩbo dolim a biang!

‘Let’s go drink a beer!’

d. Doluchȩs!
‘Let’s write it!’

e. Dȩlȩchȩsii!
‘Let’s write it (completely)!’

If the verb in the propositive sentence is transitive, as in 30c–e,

the hypothetical (propositive) form can be imperfective (as in
30c–d) or perfective (as in 30e). Propositive sentences can also
be negative, in which case they include lak:

(31) a. Lak dongȩrodȩch.

‘Let’s not make noise.’

b. Lak dosuub er a elȩchang.

‘Let’s not study now.’

c. Mȩrkong. Lak doilil.

‘Let’s not play any more.’

Because propositive sentences involve actions which are pro-

posed or suggested but have not yet taken place, it is not
surprising that they contain hypothetical verb forms. In other
words, at the moment when a speaker utters a propositive sen-
tence, the action in question is still unreal (though it may occur
in the very immediate future); therefore, the use of hypothetical
verb forms is appropriate.

19.7. PASSIVE SENTENCES

As we mentioned in 5.1.1, all Palauan transitive sentences in-
volve a doer (or agent) and a receiver (or object). While the
agent is the person who performs or carries out a particular
action, the object is the person, animal, or thing which receives
the effect of that action. In Palauan transitive sentences, the
agent is normally expressed by the subject noun phrase,
which precedes the transitive verb, while the object is ex-

465
 19 Uses of Hypothetical Verb Forms

pressed by the object noun phrase, which follows the tran-

sitive verb. This is shown in the sentences below, which contain
both imperfective and perfective forms of transitive verbs:

(32) a. A ngalȩk a mȩnga ȩr a ngikȩl.

‘The child is eating the fish.’

b. A sensei a mȩngȩlebȩd ȩr a
rȩngalȩk.
‘The teacher is hitting the children.’

c. A John a milȩngȩlebȩd a bilis.

‘John was hitting the dogs.’

d. A sȩchȩlik a silsȩbii a blai.

‘My friend burned down the house.’

e. A Toki a chiloit a babier.

‘Toki threw away the letters.’

In any Palauan sentence, it is the subject noun phrase which

is the speaker’s focus of interest or attention. In other words,
the speaker will try to structure a sentence in such a way
that its subject position will be occupied by the noun phrase
which he wishes to emphasize or from whose viewpoint he is
regarding a particular action or event. In the great majority of
cases, the speaker tends to describe an action or event from
the viewpoint of the doer or agent, as in the examples of 32.
Such sentences, in which the subject noun phrase identifies the
agent, are called active sentences because they focus upon the
agent as actively pursuing an activity which is directed at a par-
ticular object.
In some cases, however, the speaker wishes to describe a
situation from the viewpoint of a particular noun phrase which
does not function as agent. Thus, in the examples below it
is the object—i.e., the thing which receives the effect of the
action—which appears in sentence subject position and is
therefore focused upon:

(33) a. A ngikȩl a longa ȩr ngii a ngalȩk.

‘The fish is being eaten by the child.’

b. A rȩngalȩk a longȩlebȩd ȩr tir a sensei.

466
 Palauan Reference Grammar

‘The children are being hit by the teacher.’

c. A bilis a lulȩngȩlebȩd a John.

‘The dogs were being hit by John.’

d. A blai a lȩsilsȩbii a sȩchȩlik.

‘The house was burned down by my friend.’

e. A babier a lȩchiloit a Toki.

‘The letters were thrown away by Toki.’

The sentences of 33, in which the subject noun phrase identifies

the object or receiver, are called passive sentences because
they focus upon the object as passively undergoing the action
designated by the verb phrase (and performed by the agent).
The active and passive sentences of 32 and 33 convey ex-
actly the same amount of information, except that—as men-
tioned above—there is a difference in emphasis or point of view.
Thus, both 32a and 33a tell us that an act of eating is going
on at the present moment, and that the agent and object are
a child (ngalȩk) and a fish (ngikȩl), respectively. All that is dif-
ferent is the point of view: thus, the speaker would use 32a if
he were mainly interested in the child and what the child was
doing, while he would use 33a if for some reason he was partic-
ularly concerned about the fish and what was happening to it.
11

If we compare the passive sentences of 33 with the corre-

sponding active sentences of 32, we observe the following dif-
ferences. Roughly speaking, the noun phrases in subject and
object positions have switched places: thus, the object noun
phrases of 32 have come to appear in subject position in the
passive sentences of 33, while the subject noun phrases of 32
(which designate the agent) have moved to sentence-final po-
sition in 33. At the same time, the non-hypothetical verb forms
of the active sentences have been replaced by hypothetical verb
forms in the passive sentences. In addition, the sequences ȩr
ngii and ȩ r tir appear in the passive sentences 33a–b.
We shall now examine in greater detail the complex
combination of changes which relate the active and passive sen-
tences of 32–33. Our explanation will be simpler if we assume
that the passive sentences, which are more complicated in
structure, are derived from the corresponding active sentences.

467
 19 Uses of Hypothetical Verb Forms

The “exchange” of object and agent mentioned above actually

involves two processes: the agent (= subject noun phrase) of
32 is shifted to sentence-final position in 33, while the object
(= object noun phrase) of 32 is preposed to sentence-initial
position in 33. When the object noun phrase is preposed, a
pronominal trace of it must remain in its original position.
This phenomenon only occurs, however, when the object noun
phrase of the active sentence is marked with the specifying
word ȩr (cf. 2.7). Recall that ȩr can mark a noun phrase as spe-
cific only when the preceding transitive verb is imperfective
and when the object noun phrase is of a particular type. Thus, ȩr
can be used to mark all sinuglar noun phrases (whether human
or non-human) and any plural noun phrase which is human;
if, however, a non-human noun phrase is interpreted as plural,
then it cannot be preceded by ȩr. Now, in the active sentences
32a–c, we can explain the occurrence or non-occurrence of the
specifying word ȩr according to the above principle. Thus, in
32a–b, ȩr appears before the singular noun phrase ngikȩl ‘fish’
or the human plural noun phrase rȩngalȩk ‘children’, but it
cannot occur before non-human bilis ‘dogs’ of 32c if this noun
phrase is to be interpreted as plural. Because the specifying
word ȩr therefore occurs only in 32a–b, it is in these sentences
that a pronominal trace must remain when the object noun
phrase is preposed. Thus, in the passive sentences 33a–b, we
observe the pronominal traces ngii and tir following the speci-
fying word ȩr. These emphatic pronouns (cf. 4.3) agree with
the preposed noun phrase which they replace: in 33a, the 3rd
pers. sg. emphatic pronoun ngii refers to ngikȩl ‘fish’, and in
33b, the 3rd pers. human pl. emphatic pronoun tir refers to
rȩngalȩk ‘children’. In the passive sentences 33d–e, we do not
see any occurrence of specifying word ȩr + emphatic pronoun
at all, simply because the specifying word ȩr is prevented fol-
lowing the perfective verb forms of the corresponding active
sentences 32d–e.
The presence vs. absence of ȩr ngii in otherwise identical
passive sentences results in an important difference of
meaning. Observe, therefore, the following pairs of sentences:

(34) a. A Droteo a mȩnguiu ȩr a hong.

‘Droteo is reading the book.’

b. A Droteo a mȩnguiu a hong.

‘Droteo is reading the books.’

468
 Palauan Reference Grammar

(35) a. A hong a longuiu ȩr ngii a Droteo.

‘The book is being read by Droteo.’

b. A hong a longuiu a Droteo.

‘The books are being read by
Droteo.’

As we saw in 2.7, exs. 37–38, the presence vs. absence of ȩr

in the active sentences of 34 signals whether the object (hong
‘book’) is (specific) singular or plural, respectively. Because
the passive sentences of 35 are derived from the corresponding
active sentences, the presence vs. absence of ȩr ngii tells us in
exactly the same way whether the subject (hong ‘book’) is inter-
preted as singular or plural. Thus, when ȩr ngii follows the verb,
we know that hong is singular, but if no sequence of specifying
word ȩr + emphatic pronoun is found in the sentence, then hong
is interpreted as plural.
In all of the passive sentences given so far (33 and 35),
the hypothetical verb form is prefixed with some variant of
the 3rd pers. (sg.) hypothetical pronoun (lo-, lu-, lȩ-, etc.). This
is because the prefixed pronoun in hypothetical verb forms
always refers to the agent (cf. 4.10.2), and up to now we
have only examined passive sentences with third person agents
(ngalȩk ‘child’ in 33a, sensei ‘teacher’ in 33 b, etc.). It is of
course also possible to have first and second person agents in
passive sentences, and these will be indicated by the appro-
priate pronoun prefixes on the hypothetical verb forms. Thus,
with 35b, compare the sentences below:

(36) a. A hong a kunguiu.

‘The books are being read by me.’

b. A hong a donguiu.
‘The books are being read by us
(incl.).’

c. A hong a monguiu.
‘The books are being read by you.’

If the agent is a first or second person, as in 36, it is “marked”

only in the hypothetical verb form—that is, no specific noun
phrase designating the agent occurs in sentence-final position.

469
 19 Uses of Hypothetical Verb Forms

However, if the agent is a third person, then a specific noun

phrase designating the agent occurs optionally, as in the
passive sentences of 33 and 35. In other words, a specific noun
phrase identifying the agent need not be included if the speaker
and hearer know who the agent is. Thus, with 35b, compare the
following sentence:

(37) A hong a longuiu.

‘The books are being read by him/them.’

Given our previous discussions of the general function of hy-

pothetical verb forms—namely, to express unreal rather than
actual events, states, etc.—it is indeed very difficult to un-
derstand why hypothetical verb forms should be required in
passive sentences. The only speculation we are able to make
is that, in some sense, passive sentences are “less real” than
active sentences because they view a given event or situation
in a less-than-usual way. As we mentioned at the beginning of
this section, the speaker usually describes a given action or
event from the viewpoint of the agent. Therefore, active sen-
tences are highly favored because in them, the agent appears
in sentence subject position, which is reserved for the noun
phrase being given special attention or “prominence”. Now,
when a speaker uses a passive sentence, he is choosing to de-
scribe an action or situation from a somewhat less common
or normal viewpoint—namely, that of the object or receiver,
which therefore appears in the subject position of passive sen-
tences. If we agree that the less common viewpoint of passive
sentences is somehow “less real” than the normal viewpoint of
active sentences, then perhaps we can understand why passive
sentences require hypothetical verb forms.

19.7.1. FURTHER EXAMPLES OF PASSIVE SENTENCES

In the present section we will look at a large variety of Palauan
passive sentences. Since we have already analyzed the form and
meaning of such sentences, very little additional discussion will
be necessary.
In the passive sentences below, we observe the hypothetical
forms of imperfective transitive verbs (cf. 33a–c) and in one
case the hypothetical form of a transitive state verb (mȩdakt ‘be
afraid of’ in 38g):

470
 Palauan Reference Grammar

(38) a. A bȩlochȩl a lulȩmes ȩr ngii a buik.

‘The pigeon was being watched by the boy.’

b. A oles a lousbech ȩr ngii a Droteo.

‘The knife is being used by Droteo.’

c. A Toki a blȩchoel ȩl lolȩngȩseu ȩr ngii a Droteo ȩl

mȩruul a subȩlel.
‘Toki is always being helped by Droteo to do her
homework.’

d. Tia ȩl chȩlitakl a blȩchoel ȩl dongitakl ȩr ngii ȩr a

Christmas.
‘This song is always sung (by us) at Christmas.’

e. A babier a kulluchȩs.
‘The letters were being written by me.’

f. A bilis a lomȩkcharm a Droteo.

‘The dogs are being hurt by Droteo.’

g. A dȩrumk a lȩmȩdakt ȩr ngii a ngalȩk.

‘(lit.) The thunder is being feared by the child.’ = ‘The
child is afraid of the thunder.’

The passive sentences below illustrate how the hypothetical

forms of perfective transitive verbs are used (cf. 33d–e):

(39) a. A buik a lulsa a Tony.

‘The boy was seen by Tony.’

b. A ngalȩk a ksilȩbȩkii.
‘The child was kicked by me.’

c. A tolȩchoi a lulȩkȩrngii a chȩrrodȩch.

‘The baby was awakened by the noise.’

d. A ngikȩl a lȩkila a bilis.

‘The fish were eaten up by the dog.’

e. Ngak a lulȩkȩrngak a Toki.

‘I was awakened by Toki.’

f. A kliokl a lȩkilisii a rȩsȩchȩlik.

471
 19 Uses of Hypothetical Verb Forms

‘The hole was (completely) dug by my friends.’

g. A present a lȩbilskak a Droteo.

‘A present was given to me by Droteo.’

h. A hong a kbilstȩrir a rȩsȩchȩlik.

‘A book was given by me to my friends.’

i. A bȩras a lȩkȩlii a malk.

‘The rice is going to get eaten up by the chicken!’

j. A chȩmȩlem a lȩchȩmȩchii a ngalȩk.

‘Your betel nut is going to get chewed up by the child!’

In 39i–j, the present perfective (passive) forms lȩkȩlii and

lȩchȩmȩ chii are used to express warnings (cf. 12.2, ex. 14).

19.7.2. PASSIVE SENTENCES CONTAINING COMPLEX VERB PHRASES

If the verb phrase of a passive sentence is complex—i.e., if it
includes auxiliary words like mo ‘go, become’, mla (marker for
recent past tense), etc.—then the appropriate hypothetical pro-
noun is normally prefixed to each of its parts. A few typical ex-
amples are given below:

(40) a. A ureor a lȩbla lȩbo lȩmȩrek ȩr ngii a

Droteo.
‘The work has been finished by Droteo.’

b. A mubi a lȩbo lomes ȩr ngii a rȩngalȩk.

‘The movie is going to be seen by the
children.’

c. Tia ȩl babier a kbo kuluchȩs ȩr ngii.

‘This letter will be written by me.’

As we saw in 4.10.6, Palauan speakers often omit the first

occurrence of the hypothetical pronoun with complex verb
phrases. This happens most frequently in rapid, informal speech
and when the third person hypothetical pronoun is involved.
Thus, with 40a–b compare the following acceptable sentences:

(41) a. A ureor a bla lȩbo lȩmȩrek ȩr ngii a Droteo.

472
 Palauan Reference Grammar

‘The work has been finished by Droteo.’

b. A mubi a bo lomes ȩr ngii a rȩngalȩk.

‘The movie is going to be seen by the
children.’

The first occurrence of the hypothetical pronoun can be omitted

in examples like 41a–b because it is redundant: in other words,
the very same information is supplied by the (identical) hypothe-
tical pronouns which are attached to the other parts of the com-
plex verb phrase.

19.7.3. PASSIVE SENTENCES AND RELATIONAL PHRASES

The processes by which Palauan passive sentences are formed
apply more broadly than we indicated in 19.7 above, where
we only examined passive sentences in which the subject noun
phrase corresponds to the object (or receiver) of the related
active sentence. As we will see below, Palauan has passive sen-
tences in which the subject corresponds to a noun phrase which
follows the relational word ȩr in the associated active sen-
tence. In the following pairs of examples, a noun phrase ap-
pearing in a re lational phrase of the active sentence has
become the subject of the passive sentence:

(42) a. A Droteo a oureor ȩr a stoang.

‘Droteo works at the store.’

b. A stoa a loureor ȩr ngii a Droteo.

‘(lit.) The store is worked at by Droteo.’

(43) a. A Toki a riros ȩr tia ȩl diong.

‘Toki drowned in this river.’

b. Tia ȩl diong a lȩriros ȩr ngii a Toki.

‘(lit.) This river was drowned in by Toki.’

(44) a. A ngȩlȩkek a smechȩr ȩr a tȩretȩr.

‘My child is sick with a cold.’

b. A tȩretȩr a lsechȩr ȩr ngii a ngȩlȩkek.

‘It’s a cold that my child is sick with.’

473
 19 Uses of Hypothetical Verb Forms

(45) a. A bȩlochȩl a silebȩk ȩr a kȩrrȩkar.

‘The pigeon flew out of the tree.’

b. A kȩrrȩkar a lȩsilebȩk ȩr ngii a bȩlochȩl.

‘(lit.) The tree was flown out of by the
pigeon.’

In the examples above, we note that the subject noun phrases

of the passive sentences have several different functions in the
relational phrases of the corresponding active sentences. Thus,
in 42 and 43 stoa ‘store’ and tia ȩl diong ‘this river’ identify the
location, in 44 tȩretȩr ‘cold’ designates the cause, and in 45
kȩrrȩ kar ‘tree’ refers to the source. The passive sentences of
42–45 are unusual in that they all contain intransitive rather
than transitive verbs; furthermore, the pronominal trace ngii
appears after the relational word ȩr rather than the speci-
fying word ȩr.

19.7.4. NEGATIVE AND CONDITIONAL PASSIVE SENTENCES

When passive sentences, which contain hypothetical verb
forms, appear with constructions which themselves require hy-
pothetical verb forms, only a single hypothetical verb form is
used. Thus, in examples 46a–c we observe negative passive
sentences with diak, and in 46d we have a conditional passive
sentence:

(46) a. A tȩkoi er a Belau a diak losuub ȩr ngii a John.

‘Palauan isn’t being studied by John.’

b. A Droteo a dimlak longȩlebȩd ȩr ngii a Tony.

‘Droteo wasn’t hit by Tony.’

c. A kȩdȩra a dimlak lȩmad ȩr ngii a Toki.

‘The beach wasn’t where Toki died.’

d. A biang a lak lolim ȩr ngii a sȩchȩlim, e mnguu e loia ȩr

a icebox.
‘If the beer isn’t drunk by your friend, then take it and
put it in the icebox.’

474
 20 Questions

20Questions
20.1. YES-NO QUESTIONS
In every language of the world, there is a distinction between
sentences which make statements and sentences which ask
ques tions. While statements provide or supply information
by describing events, actions, states, etc., questions ask for or
demand information of one type or another. There are many
different ways of asking questions in Palauan, and these will be
taken up in detail below. In general, our study of the structure of
Palauan questions should pose few difficulties, since we are al-
ready familiar with most of the grammatical processes involved.
All languages have a basic distinction between yes-no ques
tions vs. questions which ask for the specific identity of a
person, place, thing, etc. When the speaker uses a yes-no
question, he simply wants to know whether or not such-and-
such is the case; he expects the hearer to answer with ‘yes’
(chochoi) or ‘no’ (ng diak, ng dimlak, or ng dirkak—cf. 18.8).
The following is a typical yes-no question:

(1) A Droteo ng mlo ȩr a skuul?

‘Did Droteo go to school?’

If the speaker, however, wishes to know the specific identity of

someone or something involved in a particular event, state, etc.,
he will use a question containing a specific question word like
tȩcha ‘who?’ or ker ‘where?’ Thus, with 1, contrast the following
questions:

(2) a. Ng tȩcha a mlo ȩr a skuul?

‘Who went to school?’

b. A Droteo ng mlo ȩr ker?

‘Where did Droteo go?’

Before examining questions which contain question words, we

will first analyze questions of the yes-no type.

475
 20 Questions

If the subject of a yes-no question is a non-emphatic

pronoun (cf. 4.2), then the word order of the yes-no question is
identical to that of the corresponding statement. The question,
however, is distinguished from the statement by the intonation
(or pitch) of the voice: while the statement is uttered with a
fairly low, even pitch, falling slightly at the end, the question is
spoken with a steadily rising intonation, which remains high at
the end. Observe the pairs of sentences below:

(3) a. Kȩ, mle smechȩr.

‘You were sick.’

b. Kȩ mle smechȩr?
‘Were you sick?’

(4) a. Ng milil ȩr a sers.

‘He/she is playing in the garden.’

b. Ng milil ȩr a sers?
‘Is he/she playing in the garden?’

(5) a. Tȩ mlo ȩr a che er a elȩchang.

‘They went fishing today.’

b. Tȩ mlo ȩr a che er a elȩchang?

‘Did they go fishing today?’

(6) a. Tȩ chad ȩr a Siabal.

‘They’re Japanese.’

b. Tȩ chad ȩr a Siabal?
‘Are they Japanese?’

In the cases below, too, the word order of the question sentence
and the corresponding statement sentence is identical:

(7) a. Ng soal a biang.

‘He wants some beer.’

b. Ng soal a biang?
‘Does he want any beer?’

(8) a. Ng suebȩk a rȩngrir.

476
 Palauan Reference Grammar

‘They’re worried.’

b. Ng suebȩk a rȩngrir?
‘Are they worried?’

(9) a. Ng ngar ȩr ngii a kȩlem.

‘There’s food for you.’

b. Ng ngar ȩr ngii a kȩlem?

‘Is there any food for you?’

(10) a. Ng diak a chisel a Toki.

‘There isn’t any news about Toki.’

b. Ng diak a chisel a Toki?

‘Isn’t there any news about Toki?’

(11) a. Ng dimlak longuiu ȩr a hong.

‘He didn’t read the book.’

b. Ng dimlak longuiu ȩr a hong?

‘Didn’t he read the book?’

In the examples of 7–11, the ng in subject position is a

pronomin al trace of a subject which has been shifted to the
right of the verb phrase during the normal derivation of the
statement sentence (cf. 17.2 and 17.7). Again, the intonation
pattern is the only factor which differentiates the questions
from the statements in these examples.
If the subject of a yes-no question is a specific noun phrase
rather than a non-emphatic pronoun, then the question can be
expressed in two ways. First, observe the pairs of examples
below:

(12) a. A Droteo a mla mei.

‘Droteo has arrived.’

b. Ng mla me a Droteo?
‘Has Droteo arrived?’

(13) a. A rȩsȩchȩlim a mlo milil ȩr a kȩderang.

‘Your friends went to play at the beach.’

477
 20 Questions

b. Tȩ mlo milil a rȩsȩchȩlim ȩr a kȩdȩrang?

‘Did your friends go to play at the beach?’

Comparison of the yes-no questions of the b-sentences with the

corresponding statements of the a-sentences leads us to the
simple conclusion that the yes-no questions are derived from
the statements by the process of subject shifting (cf. 17.2).
Since the shifted subject Droteo of 12b is singular, the 3rd
pers. sg. non-emphatic pronoun ng appears as a pronominal
trace. In 13b, however, the shifted plural subject rȩsȩchȩlim
‘your friends’ leaves the 3rd pers. (human) pl. pronominal trace
tȩ. As expected, the yes-no questions of 12b and 13b are spoken
with a rising intonation; this feature of their pronunciation dif-
ferentiates them from statement sentences involving subject
shifting, which are pronounced with a low, even pitch. Thus,
12b, for example, contrasts with the following:

(12b’) Ng mla me a Droteo.

‘Droteo has arrived.’

Now, with the yes-no questions of 12b and 13b, repeated

here for convenience, contrast the c-sentences, which are iden-
tical in meaning:

(12) b. Ng mla me a Droteo?

‘Has Droteo arrived?’

c. A Droteo ng mla mei?

(13) b. Tȩ mlo milil a rȩsȩchȩlim ȩr a kȩdȩrang?

‘Did your friends go to play at the beach?’

c. A rȩsȩchȩlim tȩ mlo milil ȩr a kȩdȩrang?

The c-sentences are derived from the b-sentences by optionally

preposing the shifted subject (Droteo in 12 and rȩsȩchȩlim
‘your friends’ in 13) to sentence-initial position. Unlike the in-
stances of preposing observed in chaps. 17 and 18, however, the
preposed subject in 12c and 13c simply precedes, but does not
replace, the pronominal traces ng and tȩ. The examples below
exactly parallel those of 12a–c and 13a–c: the a-sentence is a
statement, the b-sentence is a yes-no question derived from

478
 Palauan Reference Grammar

the a-sentence by subject shifting, and the c-sentence is an-

other yes-no question derived by preposing the shifted subject
of the b-sentence:

(14) a. A dort a mȩduch 1 ȩl kȩrrȩkar.‘Ironwood is a strong

tree.’

b. Ng mȩduch ȩl kȩrrȩkar a ‘Is ironwood a strong

dort? tree?’

c. A dort ng mȩduch ȩl
kȩrrȩkar?

(15) a. A ears a mlo dȩkimȩs. ‘The sail got wet.’

b. Ng mlo dȩkimȩs a ears? ‘Did the sail get wet?’

c. A ears ng mlo dȩkimȩs?

If the subject of a statement sentence is a noun phrase of pos-

session, then three acceptable yes-no questions can be formed
from it. Thus, in the pair of sentences below,

(16) a. A bȩchil a Droteo a klȩbokȩl.

‘Droteo’s wife is pretty.’

b. Ng klȩbokȩl a bȩchil a Droteo?

‘Is Droteo’s wife pretty?’

the yes-no question of 16b is derived from the statement of 16a

by shifting the subject bȩchil a Droteo ‘Droteo’s wife’, which
is a noun phrase of possession. Now, from 16b, we can form
either of the following yes-no questions, which have the same
meaning:

(16) c. A bȩchil a Droteo ng klȩbokȩl? ‘Is Droteo’s wife pretty?’

d. A Droteo ng klȩbokȩl a bȩchil?

The difference between the derivations of 16c–d is as follows: in

16c, the whole shifted subject bȩchil a Droteo ‘Droteo’s wife’ is
preposed, while in 16d only the possessor Droteo is preposed,

479
 20 Questions

leaving the possessed noun bȩchil ‘his wife’ in sentence-final po-

sition (cf. our discussion of preposing of possessor in 17.3).
Another example parallel to 16a–d is given below:

(17) a. A chȩral a bȩras a ‘The price of rice is high.’/‘Rice

mȩringȩl. is expensive.’

b. Ng mȩringȩl a chȩral ‘Is rice expensive?’

a bȩras?

c. A chȩral a bȩras ng
mȩringȩl?

d. A bȩras ng mȩringȩl a
chȩral?

20.2. THE QUESTION WORD TȨCHA

The question word tȩcha ‘who?’ is used when the speaker
wishes to know the identity of one or more persons involved in
a particular event, state, etc. Observe the examples below:

(18) a. Ng tȩcha a sensei ȩr kau?

‘Who is your teacher?’

b. Ng tȩcha a lilȩchȩsii tia ȩl babier?

‘Who wrote this letter?’

c. Ng tȩcha a mlo ȩr a party?

‘Who went to the party?’

d. Ng tȩcha a milosii a bȩlochȩl?

‘Who shot the pigeon?’

If we try to explain the sentences of 18 in a superficial, non-

technical way, we might say that tȩcha ‘who?’ must be preceded
by ng in sentence-initial position when a question is being asked
about the identity of the sentence subject. Though this simple
explanation will allow us to form question sentences like 18a–d
correctly, it nevertheless does not reflect the actual derivation
of 18a–d, which we will now examine in detail.

480
 Palauan Reference Grammar

The examples of 18 are derived by the process of subject

shifting from equational source sentences of the form subject
noun phrase + tȩcha ‘who?’. The derivation of 18a is therefore
represented according to the following scheme:

(19) Source Sentence Resulting Sentence

A sensei ȩr kau a tȩchang. → Ng tȩcha a sensei ȩr kau?

‘Who is your teacher?’

The equational source sentence of 19 must undergo the process

of subject shifting in order to become grammatical. As a conse-
quence of this process, the subject noun phrase sensei ȩr kau
‘your teacher’ has come to appear in final position in the re-
sulting sentence of 19, and the pronominal trace ng occupies
the original subject position. The remaining question sentences
of 18 are derived in exactly the same way, except that their
source sentences have subjects of a rather special kind, as we
will see below.
An important subtype of Palauan equational sentence is il-
lustrated by the examples below:

(20) a. A mlad a Droteo.

‘The one who died is Droteo.’

b. A chillȩbȩdii a Toki a John.

‘The person who hit Toki is John.’

c. A olisȩchakl a tȩkoi ȩr a Merikel a Masaharu.

‘The one who teaches English is Masaharu.’

d. A mle ȩr a blik er a elii a Cisco.

‘The person who came to my house yesterday is Cisco.’

e. A soal ȩl mo ȩr a Siabal a Maria.

‘The one who wants to go to Japan is Maria.’

As the English equivalents indicate, the examples of 20 are used

when the speaker wishes to exhaustively identify a particular
person as the one who is characterized by a particular event,
action, state, etc. Thus, 20b emphasizes that it was John and
only John (from among the possible persons involved) who hit
Toki, and 20e singles out Maria as the person who wants to go

481
 20 Questions

to Japan. Because the examples of 20 have the abovementioned

connotation of exhaustiveness, they differ in meaning from non-
equational sentences containing a subject noun phrase and a
verb phrase. Thus, compare 20b with the following:

(21) A John a chillȩbȩdii a Toki.

‘John hit Mary.’

While 21 simply tells us what John did—namely, hit Mary—and

leaves open the possibility that other persons might have done
the same thing, 20b asserts that John alone was the one who
performed this action.
In the equational sentences of 20, the italicized subject noun
phrases do not contain any noun corresponding to English ‘one’
or ‘person’. In other words, these sentences show that Palauan
verb phrases (together with any object noun phrase or rela-
tional phrase associated with them) can actually function as
noon phrases with the meaning ‘the one who…’ or ‘the person
who…’ Since the second noun phrase of 20a–e (Droteo, John,
etc.) refers to a human being, and since the sentences are equa-
tional, it is clear that the italicized subject noun phrases must
also refer to human beings. 2 The phenomenon under discussion
here is fairly widespread in Palauan; thus, in 2.5 we saw that
verb phrases consisting of state verbs (or, occasionally, action
verbs) can be prefixed with the plural marker rȩ- to function as
noun phrases with the meaning ‘anyone who is…, those who
are…’ One example of this type is repeated here:

(22) A irȩchar, e a rȩmeteet a ulȩngȩseu ȩr a rȩmechȩbuul.

‘In earlier times, the rich helped the poor.’

In addition, we saw in 8.2 that many state verbs derived with

the resulting or anticipating state affixes have come to be
used as nouns meaning ‘(something which is…’, as in the ex-
amples below:

(23) a. Ng soak ȩl mȩnga a chȩlat.

‘I’d like to eat a smoked one (= fish).’

b. Ng mle bȩtok a {sȩlesȩb/tȩlȩmall} er se ȩr a taem ȩr a

mȩkȩmad.
‘There were lots of things {burned/destroyed} during
the war.’

482
 Palauan Reference Grammar

Returning to the question sentences of 18b–d, we can now

see that they are derived by subject shifting from equational
source sentences whose subjects, like those of 20, are actually
verb phrases being used as noun phrases. Therefore, the deri-
vation of 18b–d is schematized as follows:

(24) Source Sentence Resulting Sentence

a. A lilȩchȩsii tia ȩl babier a Ng tȩcha a lilȩchȩsii tia ȩl

tȩchang. babier?
→ ‘Who wrote this letter?’

b. A mlo ȩr a party a → Ng tȩcha a mlo ȩr a

tȩchang. party?
‘Who went to the party?’

c. A milosii a bȩlochȩl a Ng tȩcha a milosii a

tȩchang. bȩlochȩl?
→ ‘Who shot the pigeon?’

In the examples of 24, the italicized subject of the source sen-

tence is obligatorily shifted to the right of tȩcha ‘who?’ in the re-
sulting sentence, and ng remains before tȩcha as a pronominal
trace.
When we think more carefully about the meaning of ques-
tions like 24a–c, we find further evidence in support of the claim
that they are derived from equational source sentences which
have a connotation of exhaustiveness (cf. our discussion fol-
lowing 20 above). In other words, since a question containing
tȩcha ‘who?’ asks for the exhaustive identity of the person or
persons involved in a particular event, action, etc., then it is
only natural that it should be derived from a source sentence
which has this very connotation. Thus, given the structure of
their source sentences, the questions of 24a–c should really be
given English equivalents such as (for 24a) ‘Who is the one who
wrote this letter?’, etc. 3

483
 20 Questions

20.2.1. FURTHER TYPES OF QUESTIONS WITH tȩcha

The question sentences of 18a–d, which are derived by subject
shifting, can be further transformed by a rule which once again
preposes the shifted subject to sentence-initial position (cf.
12–15 above). Thus, with 18a, repeated below for convenience
as 25a, compare 25b:

(25) a. Ng tȩcha a sensei ȩr kau? ‘Who is your teacher?’

b. A sensei ȩr kau ng tȩchang?

In 25b, which is identical in meaning, the shifted subject sensei

ȩr kau ‘your teacher’ of 25a has been moved back to sentence-
initial position, where it precedes the pronominal trace ng. The
very same process can apply to shifted subjects which are ac-
tually verb phrases functioning as noun phrases (cf. our dis-
cussion of 20a–e above). Thus, with 18c, repeated below as 26a,
compare 26b:

(26) a. Ng tȩcha a mlo ȩr a party? ‘Who went to the party?’

b. A mlo ȩr a party ng tȩchang?

So far, we have only examined question sentences with

tȩcha ‘who?’ in which the speaker wishes to know the identity
of the agent or subject (in the non-technical sense mentioned
following 18). If the speaker wants to know the identity of the
object—i.e., of the person receiving the effect of a particular
action—then he can formulate questions with tȩcha according
to two different grammatical patterns. The simpler of the two
patterns is illustrated below; here, the question word tȩcha
‘who?’ merely occupies the position normally occupied by sen-
tence objects (i.e., following the transitive verb phrase):

(27) a. Kȩ milsa a tȩcha ȩr a party?

‘Whom did you see at the party?’

b. Kȩ mȩngiil ȩr tȩchang? 4
‘Whom are you waiting for?’

c. A Droteo ng ulȩba a tȩcha ȩl mo ȩr a ochȩraol?

‘Whom did Droteo take to the money-raising
party?’

484
 Palauan Reference Grammar

Because tȩcha ‘who?’ asks for the identity of a specific person,

it must always be preceded by the specifying word ȩr (cf. 2.7)
when the transitive verb is imperfective, as in 27b.
The second pattern used for asking questions about the
identity of a human object is shown in the examples below:

(28) a. Ng tȩcha a chomilsa ȩr a party?

‘Whom did you see at the party?’

b. Ng tȩcha a lulȩkodir a rubak?

‘Whom did the old man kill?’

The questions of 28, like those of 24, are derived by shifting

subjects which are actually verb phrases being used as noun
phrases; the only difference is that the verb phrase contains a
passive verb form (cf. 19.7). Thus, the questions of 28 are de-
rived as follows:

(29) Source Sentence Resulting Sentence

a. A chomilsa ȩr a party a Ng tȩcha a chomilsa ȩr a

tȩchang. → party?
‘Whom did you see at the
party?’

b. A lulȩkodir a rubak a Ng tȩcha a lulȩkodir a

tȩchang. → rubak?
‘Whom did the old man
kill?’

Because the italicized subjects of the source sentences of 29

contain passive verb forms, the resulting sentences really mean
something like ‘Who is the person who was seen by you at the
party?’ and ‘Who is the person who was killed by the old man?’
In addition to its use in questions which ask for the identity of
the sentence subject or object, the question word tȩcha ‘who?’
can function in other environments where noun phrases nor-
mally occur. For example, in the sentences below, tȩcha follows
the relational word ȩr:

(30) a. Kȩ milluchȩs ȩr a babier ȩl mo ȩr tȩchang?

‘To whom were you writing the letter?’

485
 20 Questions

b. Kȩ oba a hong ȩr tȩchang? 5

‘Whose book do you have?’

And in the following questions, tȩcha appears as the possessor

in a noun phrase of possession:

(31) a. Se ng mlil tȩchang? 6

‘Whose car is that?’

b. Tia ng kȩlel tȩchang?

‘Whose food is this?’

Some question sentences showing additional uses of tȩcha

‘who?’ are provided below:

(32) a. Kȩ mlo ȩr a party kau mȩ tȩchang?

‘With whom did you go the party?’

b. Kȩ tȩchang?
‘Who are you?’

c. Ngka ng tȩchang?
‘Who is this person?’

d. Ng tȩcha a ngklel a sȩchȩlim? 7

‘What’s your friend’s name?’

e. Tia ȩl babier ng tȩcha a milluchȩs ȩr ngii? 8

‘Who was writing this letter?’/ ‘This letter—who was
writing it?’

f. A blai ng tȩcha a silsȩbii?

‘Who burned down the house?’/ ‘The house—who
burned it down?’

20.3. THE QUESTION WORD NGARA

The question word ngara ‘what?’ is used when the speaker
wants to know the identity of a particular thing (whether con-
crete or abstract). Question sentences with ngara ‘what?’ ex-

486
 Palauan Reference Grammar

hibit many different patterns, and the derivation of some of

them is quite complex. It will be easiest, of course, to begin with
the simplest pattern, which is illustrated in the examples below:

(33) a. Kȩ milȩchȩrar a ngara ȩr a stoang?

‘What did you buy at the store?’

b. Tȩ mȩsuub a ngarang?
‘What are they studying?’

c. Ng mo oba a ngarang?
‘What is he going to bring?’

In 33a–c, which are questions about the identity of the sentence

object, the question word ngara ‘what?’ simply occurs in the
normal position occupied by an object noun phrase—namely, di-
rectly following the transitive verb phrase.
If a specific third person subject is mentioned in question
sentences like 33a–c, we have sentences like the following:

(34) a. A Droteo ng mirruul a ngarang?

‘What did Droteo do/make?’

b. A rȩsȩchȩlim tȩ ulȩba a ngarang?

‘What did your friends bring?’

We can easily account for the word order of questions like

34a–b if we propose that they are derived by the already-familar
subject shifting and preposing rules and that they have source
sentences whose structure parallels that of 33a–c. Thus, 34a is
ultimately derived from the following source sentence,

(35) A Droteo a mirruul a ngarang.

(‘What did Droteo do/make?’)

which shows the basic order subject noun phrase (Droteo) +

transitive verb phrase (mirruul) + object noun phrase
(ngarang). As it stands, 35 is not an acceptable sentence;
therefore, it must be further transformed by (obligatory) appli-
cation of the subject shifting rule. Applying subject shifting to
35 gives us the following grammatical sentence:

(36) Ng mirruul a ngara a Droteo?

487
 20 Questions

‘What did Droteo do/make?’

Though 36 is grammatical, many speakers prefer to change it

further by moving the shifted subject Droteo back to sentence-
initial position. When this type of preposing takes place, the
preposed noun phrase merely precedes, but does not replace,
the pronominal trace ng (cf. our discussion of examples 12–15
in 20.1 above), thus deriving 34a. The analysis described here
is summarized in the following step-by-step derivation:

(37) A Droteo a mirruul a (source sentence) →

ngarang.
Ng mirruul a ngara a (by subject shifting) →
Droteo?
A Droteo ng mirruul a (by preposing of shifted
ngarang? subject)

The derivation of 34b is exactly parallel, except that the prono-

minal trace is tȩ because of the shifted human plural noun
phrase rȩsȩchȩlim ‘your friends’.
The question sentences of 33 can undergo a special rule
which preposes the question word ngara; this rule seems to be
applicable only when ngara functions as sentence object (cf.
note 8 above). Thus, with 33a–b, repeated here for convenience,
compare the sentences with preposed ngara 9 , which are iden-
tical in meaning:

(38) a. Kȩ milȩchȩrar a ngara ȩr a stoang?

‘What did you buy at the store?’

b. Ngara kȩ milȩchȩrar ȩr a stoang?

(39) a. Tȩ mȩsuub a ngarang?

‘What are they studying?’

b. Ngara tȩ mȩsuub?

The question sentences of 34, which contain a specific third

person subject, can also be affected by the rule which preposes
ngara. Thus, with 34a, compare the following grammatical sen-
tence, which has the same meaning:

(40) Ngara ng mirruul a Droteo?

488
 Palauan Reference Grammar

‘What did Droteo do/make?’

Example 40 is derived in the following manner. First of all, the

source sentence

(35) A Droteo a mirruul a ngarang.

(‘What did Droteo do/make?’)

is transformed by subject shifting into

(36) Ng mirruul a ngara a

Droteo?
‘What did Droteo do/
make?’

At this point, the rule preposing ngara is applied, giving 40

above. Preposed ngara of 40 precedes, but does not replace,
the pronominal trace ng. A similar phenomenon was observed
in 38b and 39b, where preposed ngara precedes, but does not
replace, the non-emphatic pronouns kȩ and tȩ.
Some further examples in which ngara ‘what?’ refers to the
sentence object are provided below. The various rules men-
tioned above can account for the different patterns observed:

(41) a. A Droteo ng ulȩba a ngara ȩl mȩruul ȩr a blai? 10

‘What did Droteo use to build the house?’

b. Ng ulȩba a ngara a Toki ȩl mȩlȩkosȩk ȩr a tech?10

‘What did Toki use to cut the meat?’

c. Ngara kȩ mirruul er a elii?

‘What did you do yesterday?’

It is also possible for ngara ‘what?’ to appear as the subject

of a passive sentence (cf. 19.7), as in the examples below:

(42) a. Ngara a chomulȩchȩrar ȩr a stoang?

‘What did you buy at the store?’

b. Ngara a lurruul ȩr ngii a Droteo?

‘What did Droteo do/make?’

489
 20 Questions

The passive sentences of 42 are related to active sentences

such as the following ( = 33a and 34a):

(43) a. Kȩ milȩchȩrar a ngara ȩr a stoang?

‘What did you buy at the store?’

b. A Droteo ng mirruul a ngarang?

‘What did Droteo do/make?’

When we compare the passive sentences of 42 with the active

sentences of 43, we see that the agent and the object have ex-
changed positions. This accounts for the fact that ngara ‘what?’
has come to appear in subject position in the passive sentences
of 42, even though it really designates the object of the actions
involved. The following characteristics of passive sentences are
also observed in 42a–b: the verb form following ngara is hypo-
thetical (the prefixes chomu- and lu- identify the person and
number of the agent), and ȩr ngii follows the verb if the subject
of the passive sentence is singular (as in 42b).

20.3.1. FURTHER TYPES OF QUESTIONS WITH ngara

Though ngara ‘what?’ is used most frequently as a sentence
object, as illustrated by the examples in 20.3 above, it can also
fulfil other functions. Thus, the questions below are equational
sen tences in which the subject noun phrase or the second
noun phrase is ngara:

(44) a. Ngara a soam? 11

‘What do you want?’/‘What would you like?’

b. Ngara a ngklel a ‘rrat’ ȩl tȩkoi ȩr a Siabal?

‘What is the word for “bicycle” in
Japanese?’

c. Tia a ngarang?
‘What’s this?’

d. Se a ngarang?
‘What’s that?’

In the sentences below, the question word ngara appears in

a relational phrase:

490
 Palauan Reference Grammar

(45) a. A blim ng rruul ȩr a ngarang?

‘What’s your house made out of?’

b. A rȩchad tȩ mle kakoad ȩr a ngarang?

‘What were the people fighting over?’

c. A beab ng tilobȩd ȩr a ngarang?

‘What (place) did the mouse emerge from?’

d. Ng mlad ȩr a ngarang?
‘What did he die from?’

Can you distinguish among the several types of relational

phrases represented in 45a–d?
The question word ngara can be linked to a following noun
by the word ȩl, as in the examples below:

(46) a. Ngara ȩl tȩkoi a chomosuub er a elȩchang?

‘What language are you studying now?’

b. Ngara ȩl mubi a chobo momes ȩr ngii? 12

‘What kind of movie are you going to see?’

c. A bȩlochȩl ng silebȩk ȩr a ngara ȩl kȩrrȩkar?

‘Which tree did the pigeon fly out of?’

d. Ng mo ngara ȩl blai a blim?

‘What kind of house will yours be? (i.e., what will it be
made out of?)’

When ngara modifies a following noun in this way (see 24.2,

ex. 12) the resulting meaning is ‘which/what/what kind of____?’
In one interesting case, the question word ngara can be used
to refer to people. Note the contrast in meaning between the
following sentences:

(47) a. Tirke ȩl teru ȩl chad tȩ ngarang?

‘What are those two people? (i.e., what is their
profession?)’

b. Tirke ȩl teru ȩl chad tȩ rua tȩchang?

‘Who are those two people?’

491
 20 Questions

In 47a, ngara ‘what?’ asks for information about the profession

of the two people, while in 47b tȩcha ‘who’?’ is a request to have
them identified by name.
When followed by the connecting word mȩ ‘and (so)’ (see
22.1), ngara asks a question about the reason for something
and therefore corresponds closely to English ‘why?’ Observe
the examples below, in which the clause introduced by mȩ des-
ignates the action or state for whose occurrence a reason is
sought:

(48) a. Ngara mȩ a Droteo a dimlak lȩpass ȩr a

test?
‘Why didn’t Droteo pass the test?’

b. Ngara mȩ kȩ mlo ȩr a Saibal?

‘Why did you go to Saipan?’

c. Ngara mȩ ng mle kȩsib a rȩngum?

‘Why were you angry?’

d. Ngara mȩ a rȩsȩchȩlim a silesȩb a blai?

‘Why did your friends burn down the
houses?’

If the clause introduced by mȩ has a specific third person

subject, then this subject can be preposed to sentence-initial
position, leaving behind a pronominal trace. Thus, with 48d,
compare the sentence below, whose meaning is the same:

(49) A rȩsȩchȩlim ngara mȩ tȩ silesȩb a blai?

‘Why did your friends burn down the houses?’

Another way of asking ‘why?’ in Palauan is to use ngara followed

by the obligatorily possessed noun uchul ‘its reason’ (cf. chap.
3, note 17), which is in turn followed by a clause introduced by
mȩ. Questions with ngara uchul mȩ… tend to be more serious
than those with ngara mȩ…—that is, they ask for a detailed ex-
planation of the real reason behind something. Note the sen-
tences below:

(45) a. Ngara uchul mȩ kȩ mȩrmang?

‘What’s the (real) reason you’re
coming?’

492
 Palauan Reference Grammar

b. Ngara uchul mȩ ng mlo soam ȩl

mȩrael?
‘Why have you decided to leave?’

20.4. THE QUESTION WORD TELA

In order to ask a question about the quantity or size of some-
thing, we use the question word tela ‘how much, how many?’
The derivation of questions containing tela involves the very
same processes of subject shifting and preposing discussed in
connection with tȩcha ‘who?’ and ngara ‘what?’ above. Observe
the following examples:

(51) a. Ng tela a klȩmȩngȩtem? ‘How tall are you?’

b. A klȩmȩngȩtem ng telang?

(52) a. Ng tela a chȩrmem ȩl ‘How many dogs do you

bilis? have?’

b. A chȩrmem ȩl bilis ng
telang?

(53) a. Tȩ tela a rȩsȩchȩlim? ‘How many friends do you

have?’

b. A rȩsȩchȩlim tȩ telang?

It is easy to see that the a-sentences above are derived by

subject shifting from equational source sentences of the form
subject noun phrase + tela ‘how much, how many?’ The b-sen-
tences are in turn derived from the a-sentences by optionally
preposing the shifted subject. The step-by-step derivation of 53,
for instance, is as follows:

(54) A rȩsȩchȩlim a telang. (source sentence) →

Tȩ tela a rȩsȩchȩlim? (by subject shifting) →
A rȩsȩchȩlim tȩ telang? (by preposing of shifted subject).

493
 20 Questions

Since the source sentence of 54 is not grammatical as it stands,

subject shifting must be applied to it obligatorily. The resulting
sentence, in which the shifted subject rȩsȩchȩlim ‘your friends’
leaves the pronominal trace tȩ, is perfectly acceptable.
Therefore, application of the preposing rule in the last step of
54 is merely optional.
If the shifted subject in sentences with tela is a noun phrase
of possession, then the process of preposing can apply either to
the entire shifted subject or to the possessor alone (cf. 16c–d
above). Consider the examples below:

(55) a. Ng tela a chȩral a bȩras? ‘How much does the rice

cost?’

b. A chȩral a bȩras ng telang?

c. A bȩras ng tela a chȩral?

(56) a. Ng mle tela a rȩkil a ngalȩk? ‘How old was the child?’

b. A rȩkil a ngalȩk ng mle

telang?

c. A ngalȩk ng mle tela a rȩkil?

(57) a. Ng mle tela a teng ȩ r a ‘What was Toki’s grade?’

Toki?

b. A teng ȩr a Toki ng mle

telang?

c. A Toki ng mle tela a teng ȩr

ngii?

The b-sentences above are derived by preposing the entire

shifted subject (italicized) of the a-sentence, while the c-sen-
tences are formed by preposing only the possessor. In 57c, a
pronominal trace of the preposed possessor Toki remains in
the form of an emphatic pronoun (ngii) following the relational
word ȩr.
Just like ngara ‘what?’, the question word tela ‘how much,
how many?’ can be linked to a following noun by the word ȩl.
The examples below are typical:

494
 Palauan Reference Grammar

(58) a. Ng tela ȩl klok er a elȩchang?

‘What time is it now?’

b. Kȩ me ȩr a tela ȩl klok ȩr a klukuk?

‘At what time are you coming tomorrow?’

c. Ng tela ȩl ududem a ngar ȩr a bangk?

‘How much of your money do you have in the
bank?’

d. Tȩ mle tela ȩl chad a ilȩko ȩr a party?

‘How many people went to your party?’

e. Kȩ ngilim a tela ȩl biang?

‘How much (of the) beer did you drink?’

f. Ng tela ȩl ngikȩl a chomȩkilang?

‘How many fish did you eat?’

20.5. THE QUESTION WORD KER

The question word ker ‘where?’ is used when the speaker wants
to find out the location of some action or state, or the goal or
source of an action involving movement. This question word,
which cannot be introduced by a, always appears in a rela-
tional phrase in sentence-final position. Observe the examples
below, which have non-emphatic pronouns as subjects:

(59) a. Kȩ milsa a Satsko ȩr ker?

‘Where did you see Satsko?’

b. Ng mȩruul ȩr ngii ȩr ker?

‘Where is he making it?’

c. Tȩ mlo ȩr ker er a elii?

‘Where did they go yesterday?’

In 59a–b, ȩr ker ‘where? = at what place?’ is being used as a

locational phrase (cf. 14.2), while in 59c ȩr ker ‘where? = to
what place?’ functions as a directional phrase (cf. 14.3).

495
 20 Questions

If a question sentence with ker has a specific third person

subject, then two patterns are possible, as the following ex-
amples show:

(60) a. Ng ngar ȩr ker a tik? ‘Where’s my purse?’

b. A tik ng ngar ȩr ker?

(61) a. Tȩ mla ȩr ker a rȩngalȩk? ‘Where were the children?’

b. A rȩngalȩk tȩ mla ȩr ker?

The a-sentences above are derived by the obligatory application

of subject shifting to source sentences of the form

(60a’) A tik a ngar ȩr ker.

(‘Where’s my purse?’)

(61a’) A rȩngalȩk a mla ȩr ker.

(‘Where were the children?’)

From the a-sentences we can in turn derive the b-sentences

by optionally preposing the shifted subject. Several pairs of
question sentences similar to 60–61 are now given:

(62) a. Ng chad ȩr ker a John? ‘Where is John from?’

b. A John ng chad ȩr ker?

(63) a. Ng ruoll ȩr ker a blim? ‘Where is your house to be

built?’

b. A blim ng ruoll ȩr ker?

(64) a. Ng tilobȩd ȩr ker a ‘Where did the crab emerge

rȩkung? from?’

b. A rȩkung ng tilobȩd ȩr
ker?

496
 Palauan Reference Grammar

In sentence-final position, and following short verb forms

like mo ‘go’, mla ‘was/were (located)’, etc., the relational phrase
ȩr ker is often contracted and pronounced as if it were a single r
at the end of the preceding word. This phenomenon is observed
in the sentences below:

(65) a. Kȩ ulȩmȩngur e mo ȩr ker? [kulǝmǝŋurεmor]

‘Where did you go after eating?’

b. Kȩ mla ȩr ker? [kǝmlar]

‘Where have you been?’

20.6. THE QUESTION WORD OINGARA

The question word oingara ‘when?’, which is never introduced
by a, is used to ask questions about the time of an event, action,
state, etc. This question word usually appears in sentence-final
position as part of the relational phrase er oingara ‘when?’,
which is classified as a temporal phrase (cf. 14.6). In the ex-
amples below, the sentence subject is a non-emphatic pronoun:

(66) a. Kȩ me er oingarang?
‘When are you coming?’

b. Tȩ mo ȩr a Guam er oingarang?
‘When are they going to Guam?’

c. Kȩ milsa a sȩchȩlik er oingarang?

‘When did you see my friend?’

If a question sentence with oingara has a specific third person

subject, we get sentences such as the following:

(67) a. A Helen ng mirrael er oingarang?

‘When did Helen leave?’

b. A rȩsȩchȩlim tȩ me mȩngȩtmokl ȩr a blik er oingarang?

‘When are your friends coming to clean my house?’

Can you explain how the question sentences of 67 have been de-
rived?

497
 20 Questions

20.7. THE SPECIAL QUESTION WORD MȨKȨRA

Palauan has a special verb mȩkȩra ‘do what?’ which can only
be used in question sentences. Since mȩkȩra is a verb, it can
occur in various tenses: thus, we have milȩkȩra or mlȩkȩra in
the past tense (cf. 5.3.2) and mo mȩkȩra in the future tense.
Some typical sentences containing mȩkȩra ‘do what?’ are given
below:

(68) a. Kȩdȩ mȩkȩrang?

‘What shall we do (now)?’

b. Kȩ mȩkȩrang?
‘What are you doing?’

c. Kȩ milȩkȩra er se ȩr a lȩme a Toki?

‘What were you doing when Toki came?’

d. Kȩ mlȩkȩra er a elii?
‘What did you do yesterday?’

e. Kȩ mȩkȩra kung 13 ?
‘What are you about to do?’

f. Ng milȩkȩra a buik e ruebȩt?

‘How did the boy fall?’

g. Kȩ mlȩkȩra mȩ ke mle otsir ȩr a test?

‘How did you fail the test?’

h. Kȩ mo ȩr a Hawaii ȩl mo mȩkȩrang?
‘What are you going to go to Hawaii for?’

i. A rȩchad er a Belau tȩ mȩkȩra a loruul a bȩkai? 14

‘What do the Palauans do in making pottery?’

498
 Palauan Reference Grammar

20.8. SENTENCES WITH TWO QUESTION WORDS

When the speaker wishes to know the identity of two or more
persons, things, places, etc., he can formulate a question in
which two occurrences of the same question word are joined by
the connecting word mȩ (see. 25.4 ). Some questions of this
type are listed below:

(69) a. Kȩ milsa a tȩcha mȩ a tȩcha ȩr a party?

‘Who (pl.) did you see at the party?’

b. Ng tȩcha mȩ a tȩcha a ulȩbȩngkem ȩl mo ȩr a

chelȩbachȩb?
‘Who (pl.) went with you to the Rock Islands?’

c. Ngara mȩ a ngara a chomoruul ȩl kirel a party?

‘What things are you making for the party?’

d. A Droteo ng mlo ȩr ker mȩ ker?

‘What places did Droteo go to?’

e. Kȩ mlo ȩr a Siabal er oingara mȩ oingarang?

‘On what occasions/at what times did you go to Japan?’

499
 21 Direct and Indirect
Quotation

21DirectandIndirectQuotation
21.1. DIRECT VS. INDIRECT QUOTATION
During the course of a conversation, the speaker often wishes
to report the statements or assertions of some third party. In
so doing, he can make a choice between direct quotation and
indirect quotation. When using direct quotation, the speaker
repeats someone else’s statement or assertion word for
word—i.e., exactly as the other person stated or asserted it.
When using indirect quotation, however, the speaker does not
repeat someone else’s statement in its original form but instead
summarizes or rephrases this statement from his own point
of view. The pair of sentences below will give us a preliminary
idea of the difference between the two types of quotation:

(1) a. A Droteo a dilu ȩr ngak ȩl kmo “Ng soak ȩl eko ȩr a

blim.”
‘Droteo said to me, “I want to come to your house.”’

b. A Droteo a dilu ȩr ngak ȩl kmo ng soal ȩl me ȩr a blik.

‘Droteo told me that he wants to come to my house.’

In example la, which involves direct quotation, the

speaker—who for ease of discussion we shall call John—is re-
peating exactly what Droteo said to him; here, Droteo’s word-
for-word statement is set off by double quotation marks (“…”).
In 1b, however, John does not quote Droteo directly but instead
reports from his own viewpoint on Droteo’s statement. This dif-
ference in viewpoint becomes obvious when we compare the
words following ȩl kmo in 1a vs. 1b. Since the words set off
by double quotation marks in la represent a sentence which
Droteo actually addressed to John, this sentence of course ex-
hibits Droteo’s point of view. Thus, in talking to John, Droteo
naturally used soak ‘my desire’ to refer to his own wish, blim
‘your house’ to refer to John’s house, and eko ‘go’ to refer to his
own action of going to John’s house (recall that the directional
verb eko is specifically used to indicate movement towards the

500
 Palauan Reference Grammar

location of the person addressed). The three words under dis-

cussion must change in 1b, however, because in this sentence
the statement which Droteo made to John (the speaker) is now
regarded from John’s point of view. For this reason, John must
use soal ‘his desire’ to refer to Droteo’s wish, blik ‘my house’ to
refer to his own house, and me ‘come’ to refer to Droteo’s action
(recall that the directional verb me indicates movement towards
the location of the speaker).
As 1a–b show, the sequence ȩl kmo is used to introduce
direct as well as indirect quotation. This sequence characteris-
tically follows verbs of saying like dmu ‘say’ of 1 and—as we will
see below—a large variety of verbs involving thinking, belief,
and the like. It is perhaps easiest to consider ȩl kmo as a single
unit which serves to introduce clauses that describe someone’s
statement, thought, belief, etc. From a more technical view-
point, however, ȩl kmo and whatever follows it seem to con-
stitute a type of speci fying clause (cf. 15.7). Thus, in 1a–b,
the sequences introduced by ȩl kmo give further information
about the verb dilu ‘said’ of the preceding independent clause
by specifying (through direct or indirect quotation) the content
of the statement made. While the ȩl of ȩl kmo is the dependent
clause introducer (cf. 15.1), the exact status of kmo is un-
clear; this word appears to be a special verb which is directly
followed by a clause that has the structure of a full sentence and
describes someone’s statement, thought, belief, etc. 1
In order to simplify our analysis of direct vs. indirect quota-
tion, we implied above that this distinction is relevant only when
the speaker wishes to report what some third party has said.
As the examples below show, the abovementioned restriction is
not at all necessary, since it is also possible for a speaker to
report on some statement which he himself (or even the hearer)
made at some earlier time:

(2) a. Ak dilu ȩr a Droteo ȩl kmo “Ng soak ȩl eko ȩr a blim.” 2

‘I said to Droteo, “I want to come to your house.”’

b. Ak dilu ȩr a Droteo ȩl kmo ng soak ȩl mo ȩr a blil.

‘I told Droteo that I wanted to go to his house.’

(3) a. Kau a dilu ȩr ngak ȩl kmo “Ng soak a biang.”

‘You said to me, “I want some beer.”’

501
 21 Direct and Indirect Quotation

b. Kau a dilu ȩr ngak ȩl kmo ng soam a biang.

‘You told me you wanted some beer.’

In comparing the direct quotation of 2a with the indirect quota-

tion of 2b, can you explain why eko must change to mo and blim
must change to blil? Can you also explain why soak of 3a be-
comes soam in 3b?

21.2. INDIRECT QUOTATION AND VERBS OF

COMMUNICATION AND MENTAL ACTIVITY
Two major groups of Palauan verbs are commonly followed
by indirect quotations introduced by ȩl kmo. The first group
involves verbs of communication—i.e., verbs which denote
the transmission or reception of information. Some typical
examples include dmu ‘say, tell’ (cf. note 1 above), mȩsubȩd
‘notify, tell’, omuachȩl ‘tell a tale, boast’, omulak ‘deceive, lie’,
outingaol ‘lie’, and orrengȩs ‘hear’. These verbs are used in
sentences like the following (cf. the b-sentences of 1–3 above),
where the clause following ȩl kmo represents the speaker’s
summary or rephrasing of what was said, heard, etc.:

(4) a. A Droteo a dilu ȩr ngak ȩl kmo a Toki a ungil ȩl sensei.

‘Droteo told me that Toki is a good teacher.’

b. A chad ȩr a kansok a dilu ȩl kmo ng diak lȩbo ȩr ngii a

chull ȩr a klukuk.
‘The weatherman said that it won’t rain tomorrow.’

c. A Droteo a silȩbȩdak ȩl kmo a dȩmal a mla mad.

‘Droteo informed me that his father died.’

d. Ng tȩcha a silȩbȩdau ȩl kmo a Toki a mlo ȩr a Merikel?

‘Who told you that Toki went to America?’

e. A Toki a omuachȩl ȩl kmo a Droteo a sȩchȩlil.

‘Toki’s telling us some story about Droteo being her
boyfriend.’

f. A Satsko a ulumulak ȩr a Toki ȩl kmo a John a ngar ȩr

ngii a bȩtok ȩl ududel.

502
 Palauan Reference Grammar

‘Satsko told Toki a lie (saying) that John has lots of

money.’

g. A ngalȩk a ulutingaol ȩr a Toki ȩl kmo a sȩchȩlil a Toki a

mlad.
‘The child lied to Toki (saying) that her friend had died.’

h. Ak rirengȩs ȩl kmo a Cisco a mo ȩr a Bulabe.

‘I heard that Cisco is going to Ponape.’

The second group of Palauan verbs which can be followed

by ȩl kmo and an indirect quotation includes verbs describing
certain types of mental activity such as thinking, believing,
knowing, and the like. The most common verbs in this class are
omdasu ‘think, believe’, oumȩra ‘believe’, mȩlȩbȩdebȩk ‘think’,
obes ‘forget’, and mȩdȩnge ‘know’. Their use is illustrated in the
examples below:

(5) a. Ak omdasu ȩl kmo a Toki a mo ȩr a Guam ȩr a klukuk.

‘I think/believe that Toki is going to Guam tomorrow.’

b. A sȩchȩlik a omdasu ȩl kmo a Droteo a ungil ȩl sensei ȩr

a Tony.
‘My friend thinks that Droteo is a better teacher than
Tony.’

c. Ak ulȩmdasu ȩl kmo ak ngoikau.

‘I was thinking I’d take you.’

d. A John a oumȩra ȩl kmo a sensei a milkodir a mȩchas.

‘John believes that the teacher killed the old woman.’

e. A dȩmal a Laurentino a dimlak loumȩra ȩl kmo ng chilitii

a skuul.
‘Laurentino’s father didn’t believe that he skipped
school.’

f. A Toki a diak loumȩra ȩl kmo a Satsko a oumlai.

‘Toki doesn’t believe that Satsko owns a car.’

g. Ak millȩbȩdebȩk 3 ȩl kmo a Toki a me ȩr a klas e ng di

dimlak lȩmei.

503
 21 Direct and Indirect Quotation

‘I was thinking/expecting that Toki would come to class,

but she didn’t come.’

h. Kȩ mla obes ȩl kmo a Droteo a me ȩr a party?

‘Have you forgotten that Droteo is coming to the party?’

i. Kȩ {mȩdȩnge 4 /mala mo mȩdȩnge} ȩl kmo a Ibȩdul a

kmal smechȩr?
‘{Do you know/Have you found out} that Ibedul is very
ill?’

j. Ng dimlak kudȩnge ȩl kmo a ngȩlȩkek a mle otsir ȩr a

skeng.
‘I didn’t know that my child had failed the test.’

k. Ak mle mȩdȩnge ȩl kmo a Droteo a chilitii a Maria.

‘I knew that Droteo had divorced Maria.’

Sentences containing verbs of communication or mental ac-

tivity followed by ȩl kmo and an indirect quotation can some-
times be transformed by moving the subject of the clause after
ȩl kmo into the position of object of the verb of communication
or mental activity. When the subject after ȩl kmo is moved in
this way, it must leave behind a pronominal trace. Thus, with
4e and 5b above, compare the following sentences, which are
essentially equivalent in meaning:

(4e’) A Toki a omuachȩl ȩr a Droteo ȩl kmo ng sȩchȩlil.

‘Toki’s telling us some story about Droteo being her
boyfriend.’

(5b’) A sȩchȩlik a omdasu ȩr a Droteo ȩl kmo ng ungil ȩl sensei

ȩr a Tony.
‘My friend thinks/considers Droteo to be a better teacher
than Tony.’

21.3. DIRECT AND INDIRECT QUOTATION OF

COMMANDS
In 19.5 we saw that Palauan imperative verb forms are used
to express orders or commands. Since commands are nor-
mally directed at the person addressed, it is not surprising that

504
 Palauan Reference Grammar

imperative verb forms are really hypothetical verb forms with

a prefixed second person hypothetical pronoun. Sentence 6a
below is an example of an affirmative command, while 6b is a
negative command:

(6) a. Bo moilil!
‘Go play!’

b. Lak mongȩrodȩch!
‘Don’t make noise!’

If the speaker wishes to report on an order or command

given by himself or someone else, he can use either direct or in-
direct quotation. If he chooses direct quotation, he repeats the
particular command word for word, exactly as it was uttered.
Sentences involving the direct quotation of a command simply
contain ȩl kmo followed by an imperative verb form. Thus, ob-
serve how the imperative verb forms of 6a–b are used as direct
quotations in the examples below:

(7) a. A rubak a dilu ȩr a rȩngalȩk ȩl kmo “Bo moilil!”

‘The old man said to the children, “Go play!”’

b. A sensei a dilu ȩr tir ȩl kmo “Lak mongȩrodȩch!”

‘The teacher said to them, “Don’t make noise!”’

If the subject of the sentence is the speaker himself (i.e., ak

‘I’), and the verb of communication (usually, dmu ‘say’) is in the
present tense, then the sentence takes on an especially threat-
ening tone, as in the following:

(8) Ak dmu ȩr kau ȩl kmo “Bo mrei!”

‘I’m telling you (again), “Go home!”’

Sentence 8 has the connotation of a threat because it would

be used by the speaker to reiterate or repeat an order which
he gave at least once before. This original order would simply
consist of the affirmative command below:

(9) Bo mrei!
‘Go home!’

505
 21 Direct and Indirect Quotation

If the speaker chooses indirect quotation to report on a

command given by himself or someone else, he merely sum-
marizes the content of the command rather than supplying the
exact words spoken. The structure of sentences involving the
indirect quotation of a command will be rather new to us: in-
stead of ȩl kmo, we have mȩ ‘(and) so’ followed by a clause de-
scribing the content of the command. As we will see in 22.1,
the word mȩ ‘(and) so’ is used to introduce an action or event
which takes place as a result of some other action, state, etc. In
the examples below, use of mȩ is appropriate because the clause
following it describes the action, event, or situation which oc-
curred (or was expected to occur) as a result of the command’s
being given:

(10) a. Ng tȩcha a dilu ȩr kau mȩ kȩ me ȩr tiang?

‘Who told you to come here?’

b. A sȩchȩlim a dilu ȩr ngak mȩ ak mei.

‘Your friend told me to come.’

c. A sensei a dilu ȩr kȩmam mȩ aki mȩsuub a tȩkoi ȩr a

Merikel.
‘The teacher told us to study English.’

d. Ak dilu ȩr a rȩngalȩk mȩ ng diak loilil ȩr a sers.

‘I told the children not to play in the garden.’

e. Ak dilu ȩr a Droteo mȩ ng diak longȩrodȩch.

‘I told Droteo not to make noise.’

21.4. DIRECT AND INDIRECT QUOTATION OF

QUESTIONS
Just as a speaker can use either direct or indirect quotation to
report on a statement or a command uttered by himself or
some other party, so can he choose between these two types of
quotation when reporting on a question which he or someone
else asks. As expected, sentences involving the direct quotation
of a question contain ȩl kmo followed by any question that could
occur as an independent sentence (cf. chap. 20). The examples
below are typical:

506
 Palauan Reference Grammar

(11) a. A sensei a ulȩker ȩr a rȩngalȩk ȩl kmo “Kom mla rȩmuul

a subȩliu?”
‘The teacher asked the children, “Have you finished
doing your homework?”’

b. A Droteo a ulȩker ȩr ngak ȩl kmo “Ng soam a biang?” 5

‘Droteo asked me, “Would you like some beer?”’

c. A sensei a ulȩker ȩr ngak ȩl kmo “Kȩ mo mȩkȩrang?”

‘The teacher asked me, “What are you going to do?”’

d. A Droteo a ulȩker ȩr ngak ȩl kmo “Ng tela a

klȩmȩngȩtem?”
‘Droteo asked me, “How tall are you?”’

As the examples above show, sentences involving the direct quo-

tation of a question use the verb oker ‘ask’ (past: ulȩker). The
directly-quoted questions of 11a–b are of the yes-no type (cf.
20.1), while those of 11c–d contain the question words mȩkȩra
‘do what?’ (cf. 20.7) and tela ‘how much, how many?’ (cf. 20.4).
In sentences involving the indirect quotation of a question,
the question is summarized or re-expressed from the speaker’s
point of view. Thus, with the directly-quoted questions of 11b–d,
compare the following indirectly-quoted questions, in which the
(independent or affix) pronouns have changed:

(12) a. A Droteo a ulȩker ȩr ngak ȩl kmo ng soak a biang.

‘Droteo asked me if I wanted a beer.’

b. A sensei a ulȩker ȩr ngak ȩl kmo ak mo mȩkȩrang.

“The teacher asked me what I was going to do.’

c. A Droteo a ulȩker ȩr ngak ȩl kmo ng tela a

klȩmȩngȩtek.
‘Droteo asked me how tall I am.’

In the examples below, we observe further cases of

indirectly-quoted questions of the yes-no type. With indirectly-
quoted yes-no questions, the negative expression ng diak is
often added in sentence-final position; the resulting sequence
corresponds closely to English ‘if/whether…or not’:

507
 21 Direct and Indirect Quotation

(13) a. Ak ulȩker ȩr ngii ȩl kmo ng sȩbȩchel ȩl me ȩr a klas ng

diak.
‘I asked him whether or not he could come to class.’

b. A John a ulȩker ȩl kmo a Mary ng mlo bȩchiil ng diak.

‘John asked whether Mary got married or not.’

c. Ng soak ȩl mo mȩdȩnge ȩl kmo a Droteo ng mlo ȩr a

skuul er a elii ng diak.
‘I’d like to know whether Droteo went to school
yesterday or not.’

d. Kȩ mȩdȩnge ȩl kmo a Toki ng mla mo ungil ȩl smechȩr

ng diak?
‘Do you know whether or not Toki has gotten better?’

e. Ng diak kudȩnge ȩl kmo a Hermana ng mȩduch ȩl

omȩkall a sidosia ng diak.
‘I don’t know whether or not Hermana knows how to
drive a car.’

f. Kȩ mla mȩlasȩm ȩl mȩlȩkoi ȩr a Droteo ȩl kmo ng

sȩbȩchel ȩl mȩskau a udoud ng diak?
‘Have you tried talking to Droteo to see if he can lend
you some money or not?’

As the examples of 13 show, the structure of the indirectly-

quoted question introduced by ȩl kmo is identical to that of any
independently-occurring yes-no question. Thus, when the yes-
no question has a specific third person subject, as in 13b–e, this
subject normally occurs in question-initial position before the
appropriate pronominal trace, which is ng in all of the examples
under consideration here. Note, further, that the indirectly-
quoted questions of 13 can be preceded by a verb of asking like
oker ‘ask’ or by sequences like ng soak ȩl mo mȩdȩnge ‘I’d like
to know (if…)’, ng diak kudȩnge ‘I don’t know (if…)’, etc.
Occasionally, the subject of an indirectly-quoted yes-no
question is moved to the left of ȩ l kmo, where it comes to
function as the object of the preceding verb. Two examples il-
lustrating this variety of sentence are given below:

(14) a. Bo ȩr a bita mȩ mȩsa a Droteo ȩl kmo ng mla me ng

diak.

508
 Palauan Reference Grammar

‘Go next door and see whether Droteo has arrived or

not.’

b. Ak omdasu 6 ȩr a Toki ȩl kmo ng mo ȩr a Guam ng diak.

‘I wonder whether Toki is going to Guam or not.’

In the sentences below, we observe further instances of

indirectly-quoted questions which contain question words of
various types (cf. 12b–c). With indirectly-quoted questions of
this kind, we cannot add ng diak to sentence-final position as we
did in the examples of 13 and 14. For convenience, the question
word in each sentence has been italicized.

(15) a. Ng soak ȩl mo mȩdȩnge ȩ1 kmo ng tela a dengua ȩr

kau.
‘I’d like to know what your phone number is.’
b. Ak ulȩker ȩr a chad ȩr a stoa ȩl kmo{ng tela a chȩral a
bȩras/a chȩral a bȩras ng telang/a bȩras ng tela a
chȩral}.
‘I asked the clerk how much the rice cost.’
c. Ng dimlak lsbȩdak ȩl kmo ng tȩcha a mlo ȩr a party.
‘He didn’t tell me who went to the party.’

d. Ak ulȩker ȩr a Toki ȩl kmo a Droteo ng me er oingarang.

‘I asked Toki when Droteo was coming.’

e. A sȩchȩlik a soal ȩl mo mȩdȩnge ȩl kmo ngara kȩ

milȩchȩrar.
‘My friend wants to know what you bought.’

f. A Toki a ulȩker ȩr tir ȩl kmo{ng mla ȩr ker a Satsko/a

Satsko ng mla ȩr ker}.
‘Toki asked them where Satsko was.’

g. Kȩ mȩdȩnge ȩl kmo ngara uchul mȩ a Toki a chillȩbȩdii

a ngalȩk?
‘Do you know why Toki hit the child?’

509
 21 Direct and Indirect Quotation

21.5. THE EXPRESSIONS Ȩ L KMO AND ȨL UA SE

With few exceptions, all of the occurrences of ȩl kmo found in
the examples of 21.1–4 above can be replaced by ȩl ua se. This
expression consists of the dependent clause introducer ȩ l fol-
lowed by ua ‘like, as’ and se ‘that’. Thus, in a sentence like the
following

(16) A Droteo a dilu ȩl ua se ng smechȩr.

‘Droteo said that he was sick.’

the specifying clause introduced by ȩl ua se gives us further

information about the verb dilu ‘said’ by describing the content
of the statement made.
For many Palauan speakers, ȩl kmo and ȩl ua se can be used
with complete interchangeability, regardless of whether direct
or indirect quotation is involved. Certain speakers, however,
seem to use ȩl ua se primarily for direct quotation. For other
speakers, the use of ȩl kmo vs. ȩl ua se involves a difference of
implication, as illustrated by the following two examples of in-
direct quotation:

(17) a. Ak rirengȩs ȩl kmo a sensei a mlad er a elii.

‘I heard (about the fact) that the teacher died
yesterday.’

b. Ak rirengȩs ȩl ua se a sensei a mlad er a elii.

‘I heard (something to the effect) that the teacher died
yesterday (but is it true?)’

As the rather free English equivalents are designed to show,

ȩl kmo of 17a implies that the speaker is fairly sure about the
factuality of what he has heard, while ȩl ua se of 17b implies
that the speaker has doubts about the truth of the event desig-
nated by the indirect quotation. Because Palauan speakers show
so much variation in the way they use ȩl kmo and ȩl ua se,
it is difficult to formulate any general principles; therefore, no
further discussion will be pursued here.

510
 Palauan Reference Grammar

21.6. INDIRECT QUOTATION FOLLOWING NOUNS

OF COMMUNICATION
Certain nouns of communication such as chais ‘news’, tȩkoi
‘word, story’, and subȩd ‘news, notice’ can be followed by in-
direct quotations introduced by ȩl kmo or ȩl ua se. The indirect
quotation supplies the content of the news or story, as in the ex-
amples below:

(18) a. Ak rirengȩs a chais ȩl ua se a sensei a mlad e ng di diak

kumȩrang.
‘I’ve heard some news (to the effect) that the teacher
died, but I don’t believe it.’

b. Kȩ rirȩngȩsii a subȩd ȩl kmo ng mla ȩr ngii a tsunami?

‘Have you heard the news that there’s been a tidal
wave?’

c. Ng mȩra ȩl tȩkoi ȩl kmo a Maria a mlo bȩchiil?

‘Is it a true story that Maria got married?’

511
 22 Reason Clauses, Result
Clauses, and Time Clauses

22ReasonClauses,ResultClauses,andTimeClauses
22.1. REASON AND RESULT CLAUSES
In 15.1 we mentioned that Palauan has many different
grammatical patterns for combining simple sentences into more
complex ones. More specifically, we saw how two simple sen-
tences like the following

(1) a. A Droteo a ulȩba a oluchȩs.

‘Droteo had/was using a pencil.’

b. A Droteo a milluchȩs a babier.

‘Droteo was writing a letter.’

can be combined into a single longer sentence in which either

la or 1b is used as a dependent clause. Thus, by combining
1a and 1b in different orders, we derive the following two sen-
tences:

(2) a. A Droteo a ulȩba a oluchȩs ȩl mȩluchȩs a babier.

‘Droteo was using a pencil to write a letter.’

b. A Droteo a milluchȩs a babier ȩl oba a oluchȩs.

‘Droteo was writing a letter with a pencil.’

In the complex sentences of 2, the italicized portions are types

of dependent clauses: in 2a, we have a purpose clause (cf.
15.2), and in 2b we find an instrument clause (cf. 15.3). The
dependent clauses of 2, you will recall, have the following char-
acteristics: (i) they are introduced by ȩl; (ii) they do not have
any overtly-expressed subject; and (iii) they normally have a
verb in the present tense, even when the sentence as a whole
designates a past action or event.
Now, combining two simple sentences into a single complex
sentence containing a dependent clause is not the only way of
deriving complex sentences in Palauan. Another way of forming
complex sentences is to join two simple sentences by words like

512
 Palauan Reference Grammar

e le ‘because’ and mȩ ‘(and) so’, which express a particular kind

of relationship between two actions, events, states, etc. Thus,
the two simple sentences below

(3) a. Ng dimlak kbo ȩr a skuul.

‘I didn’t go to school.’

b. Ak mle smechȩr.
‘I was sick.’

can be combined in two different orders, giving the following

complex sentences:

(4) a. Ng dimlak kbo ȩr a skuul e le ak mle smechȩr.

‘I didn’t go to school because I was sick.’

b. Ak mle smechȩr mȩ ng dimlak kbo ȩr a skuul.

‘I was sick, so I didn’t go to school.’

In the examples of 4, the two simple sentences of 3 have been

combined in such a way that they are related in terms of cause
and effect. Thus, in 4a, e le ‘because’ introduces a clause which
explains the cause or reason for the event or situation de-
scribed in the preceding clause, while in 4b mȩ ‘(and) so’ in-
troduces a clause which explains what happened as a result
(or consequence) of the state described in the preceding clause.
The reason clause introduced by e le in 4a and the result
clause introduced by mȩ in 4b are rather different from the
dependent clauses (e.g. purpose clauses, instrument clauses,
etc.) which we reviewed above. As 4a–b show—and as we will
see below—reason and result clauses always have overtly-ex-
pressed subjects, and there are no restrictions on the tense of
the verb which they contain.
Since Palauan reason clauses are not difficult to understand,
we will simply give a few additional examples here:

(5) a. Ng diak lsȩbȩchek ȩl mo ȩr a party e le ak kmal

mȩchȩsang.
‘I can’t go to the party because I’m very busy.’

b. A ngȩlȩkek a diak lsȩbȩchel ȩl mo milil er a elȩcha e le

ng kirel ȩl omȩngur.

513
 22 Reason Clauses, Result Clauses, and Time Clauses

‘My child can’t go play now because he has to have

dinner.’

c. A rȩchad ȩr a omȩnged a dimlak lȩbo ȩr a che e le ng

kmal mle mȩses a eolt.
‘The fishermen didn’t go fishing because the wind was
very strong.’

Though e le ‘because’ probably consists of the connecting

word e (see 25.1) and some other element le, it will be easiest
to consider e le as a single unit. Note that when the vowel-initial
pronoun ak follows e le, as in 5a, the a of ak is lost; thus, e le
ak ‘because I…’ is pronounced [εlεkh] (cf. 1.5.d.5, ex. 54).
Before examining several special types of result clauses, we
shall first familiarize ourselves with a few relatively straight-
forward examples:

(6) a. Ak di mililil mȩ ak mle otsir ȩr a test.

‘I just fooled around, so I failed the test.’

b. A Droteo a smechȩr mȩ a dȩmal a mo omȩkȩdo ȩr a

toktang.
‘Droteo’s sick, so his father’s going to call the doctor.’

c. Ng mȩkngit a eangȩd mȩ a rȩsȩchȩlik a diak lȩbo ȩr a

chelȩbachȩb.
‘The weather’s poor, so my friends aren’t going to go to
the Rock Islands.’

d. A rȩngalȩk a diak a subȩlir er a elȩcha mȩ tȩ mo milil ȩr

a kȩdȩrang.
‘The children don’t have any homework now, so they’re
going to the beach to play.’

e. Ak kmal songȩrengȩr mȩ ng soak ȩl omȩngur.

‘I’m very hungry, so I’d like to eat.’

f. Ak mlo ungil ȩl smechȩr ȩr a tȩretȩr mȩ ng mle sȩbȩchek

ȩl mo ȩr a party.
‘I got better from my cold, so I was able to go to the
party.’

514
 Palauan Reference Grammar

g. Ng ngar ȩr ngii a bȩtok ȩl subȩlam mȩ ng di kea a

techȩllam ȩl mo milil ȩl obȩngkem.
‘We have so much work that we no longer have any
chance to go out with you.’

If mȩ is followed by the vowel a, as in 6a–c, then the ȩ of mȩ

is deleted. Thus, m ȩ ak ‘(and) so I…’ is pronounced [makh],
and mȩ a (i.e., mȩ followed by a, which precedes noun phrases)
is pronounced [ma] (cf. 1.5.d.5, ex.54). Otherwise, the ȩ of mȩ
is retained, as in mȩ ng ‘(and) so he/she/it…’ [mǝŋ] and mȩ tȩ
‘(and) so they…’ [mǝtǝ].

22.1.1. SPECIAL TYPES OF RESULT CLAUSES

The use of mȩ ‘(and) so…’ to introduce result clauses is quite
widespread in Palauan. In this section we will single out several
instances of result clauses which might be difficult to recognize
as such because their English equivalents are not of the form
‘X, so Y’. For example, we have already seen in 20.3.1 that
the question word ngara ‘what?’ (or the sequence ngara uchul
‘what is the reason that…?’) is followed by a clause introduced
by mȩ ‘(and) so’ to ask a question about the cause or reason (i.e.
‘why?’). Note the examples below:

(7) a. Ngara mȩ a Droteo a mle otsir ȩr a test?

‘Why did Droteo fail the test?’

b. Ngara uchul mȩ kȩ mlo ȩr a Guam?

‘For what reason did you go to Guam?’

In 7a–b the clause introduced by mȩ is really a result clause

because it is viewed as designating an event that took place as
the result of something else. Indeed, the speaker’s purpose in
asking the question is precisely to find out what that “something
else” is—i.e., to find out the cause or reason for the event of
the result clause. In 7a, for example, the speaker assumes that
the event of Droteo’s failing the test resulted from something
else which Droteo did or did not do, and it is this “something
else”—the cause—which he wishes to know.
The question word mȩkȩra ‘do what?’ (cf. 20.7) can also be
followed by result clauses, as the sentences below illustrate:

(8) a. Kȩ mlȩkȩra mȩ kȩ mle otsir ȩr a test?

515
 22 Reason Clauses, Result Clauses, and Time Clauses

‘How/why did you fail the test?’

b. Ng mlȩkȩra a buik mȩ ng {rirebȩt/riros}?’

‘How did the boy {fall/drown}?’

The examples of 8, like those of 7, are really questions about the

cause or reason for the event designated in the result clause
introduced by mȩ. The connotation of result can be readily seen
if we give 8a–b literal translations such as ‘What did you do so
that/with the result that you failed the test?’ and ‘What did the
boy do so that/with the result that he fell/drowned?’
The word klsakl, which appears to be a resulting state
verb (cf. 7.7) related to mȩkȩsakl ‘go wrong’, is used in simple
questions like the following:

(9) a. Kȩ klsakl?
‘What’s wrong with you?’

b. Ng klsakl a chimam?
‘What’s wrong with your hand?’

Speakers use klsakl followed by a result clause to ask about the

cause or reason for an unfavorable or undesirable event. This
connotation is clear in the sentences below:

(10) a. Kȩ klsakl mȩ ng diak momȩngur?

‘Why aren’t you eating?’

b. Kȩ mle klsakl mȩ ng dimlak chobo ȩr a party?

‘Why didn’t you go to the party?’

Translated literally, the examples of 10 mean something like

‘What’s wrong with you so that/with the result that you’re not
eating?’ and ‘What was wrong with you so that/with the result
that you didn’t go to the party?’
In 21.3 we observed that the indirect quotation of a
command involves a clause introduced by mȩ ‘(and) so’, as in
the examples below:

(11) a. A sensei a dilu ȩr ngak mȩ ak mo ȩr a Guam.

‘The teacher told me to go to Guam.’

516
 Palauan Reference Grammar

b. Ak dilu ȩr tir mȩ ng diak lomȩkikiongȩl ȩr a delmȩrab.

‘I told them not to get the room dirty.’

Because the clause introduced by mȩ in 11a–b describes an

event or situation which took place (or was expected to take
place) as a direct result of the command’s being uttered, it is to
be analyzed as a result clause.
Similar in structure to the examples of 11 are sentences con-
taining the verb of permission konge ‘permit, allow’. Observe
the sentences below:

(12) a. A Droteo a kilȩnge ȩr a Toki mȩ ng mo mȩngȩdub.

‘Droteo allowed Toki to go swimming.’

b. A dȩmal a Satsko a kilȩnge mȩ a Satsko a mo ȩr a

Hawaii ȩl mo ȩr a skuul.
‘Satsko’s father gave her permission to go to Hawaii in
order to study.’

The verb of permission konge is followed by a result clause

in 12a–b because the sequence introduced by mȩ describes an
event which occurred (or whose occurrence was facilitated) as
a result of the permission being granted.
When the verb of permission konge ‘permit, allow’ is
negated, we get sentences involving the denial of permission. In
the sentences below, the verb konge appears in a hypothetical
form following the negative verb diak ‘isn’t, doesn’t exist’, and
the verb of the result clause is also hypothetical:

(13) a. A dȩlal a Toki a dimlak lȩkȩnge ȩr a Toki mȩ lousbech ȩr

a mlai.
‘Toki’s mother didn’t permit her to use the car.’

b. A sensei a dimlak lȩkȩnge ȩr a rȩngalȩk mȩ loilil ȩr a

obis.
‘The teacher forbade the children to play in the office.’

c. Ng diak kkȩnge ȩr a ngȩlȩkek mȩ lolim a mȩringȩl.

‘I don’t let my child drink hard liquor.’

517
 22 Reason Clauses, Result Clauses, and Time Clauses

We can understand why the verb of the result clause must be

hypothetical in the examples of 13 if we recall (cf. 18.4) that
hypothetical verb forms are characteristically used to designate
unreal (or non-occurring) events. In other words, the events
of the result clauses in 13 are clearly unreal in the sense that
their occurrence was (or is) prevented or precluded by the act
of forbidding represented by the negative form of konge in the
first clause. The negative form of dmu ‘say, tell’ is sometimes
used in a similar way, as the following example illustrates:

(14) A Droteo a dimlak lȩdu ȩr a ngalȩk mȩ lȩbo loilil ȩr a sers.

‘Droteo didn’t give the child permission to play in the
garden.’

Similar in structure to the examples of 11 and 12 are the fol-

lowing sentences, which contain further verbs which are com-
monly followed by result clauses:

(15) a. A Droteo a ulȩrrimȩl ȩr a Maria mȩ ng ko ȩl mocha ȩr a

party.
‘Droteo forcefully persuaded Maria to finally go to the
party.’

b. Ak urrȩmȩlii a Toki mȩ ng mȩngȩtmokl ȩr a blai.

‘I forced Toki to straighten up the house.’

c. Kȩ mo tsiui ȩr kau mȩ ng diak di molius.

‘Watch yourself so you don’t swear.’

d. Kȩ mo kȩrȩkikl mȩ ng diak chomrebȩt.

‘Be careful not to fall.’

Sentences like 15c–d, which contain expressions of precaution

(kȩ mo tsiui ‘watch out’ or kȩ mo kȩrȩkikl ‘be careful’) followed
by a result clause with a negative verb, are used as rather mild,
indirect commands.
The imperative perfective forms (cf. 19.5) of the verb
omȩche ‘leave, let (someone) (do something)’ (cf. chap. 14, note
19) are commonly followed by a result clause containing a hy-
pothetical verb form. Observe the sentences below, which are
requests that someone be permitted to do something:

(16) a. Bȩchire a ngalȩk mȩ lȩbo loilil.

518
 Palauan Reference Grammar

‘Let the child go play.’

b. Bȩchikak mȩ kbo kmȩchiuaiu.

‘Let me go to sleep.’

c. Bȩchititȩrir mȩ lȩbo lousbech ȩr a sidosia.

‘Let them use the car.’

The use of hypothetical verb forms in the result clauses of 16 is

not difficult to understand when we realize that the events de-
scribed in the result clauses are as yet unreal at the time when
the speaker utters the request. In 16a, for instance, the child
has not yet begun to play at the moment when the speaker asks
someone else to allow the child to do so; indeed, the speaker’s
very purpose in uttering such a sentence is to make the event of
the result clause become an actual fact.
The imperative (or propositive) perfective forms (cf.
19.5–6) of the verb mȩngiil ‘wait (for)’ occur with the rather
unusual combination of mȩ‘(and) so’ + conditional clause (cf.
19.1). Observe the examples below:

(17) a. Bo ȩr a blil a Toki e mchiiȩlii a Droteo mȩ a lȩkong, e bo

ȩr a stoang.
‘Go to Toki’s house and wait for Droteo to come; then go
to the store (with him).’

b. Mchiiȩlii a Droteo mȩ a lȩbo lȩmȩrek ȩr a urerel, e

mdak ȩl mȩrael.
‘Wait for Droteo to finish his work, and then leave
together.’

c. Dȩchiiȩlii a Satsko mȩ a lomȩkȩdo, e dȩbong.

‘Let’s wait for Satsko to call, and then let’s go.’

d. Bo ȩr a bita e mchiiȩlak mȩ a kekong.

‘Go next door and wait for me to come there.’

If the subject of the conditional clause after mȩ is a specific

third person noun phrase, as in 17a–c, it must be removed from
the conditional clause and placed in object position following
the perfective form of mȩngiil. It is not clear why mȩngiil ‘wait
(for)’ should require the unusual construction observed here.

519
 22 Reason Clauses, Result Clauses, and Time Clauses

As we have seen in this and the preceding section, the Pa-

lauan word mȩ ‘(and) so’ serves to introduce result clauses. The
word mȩ also has other functions, such as joining two or more
noun phrases (e.g. Toki mȩ a Droteo ‘Toki and Droteo’) and con-
necting two sentences which are parallel in structure. These
functions will be explained in detail when we examine the con
necting word mȩ in chap. 25.

22.2. TIME CLAUSES

Palauan has a variety of time clauses which are used to ex-
press a temporal relationship between two events, actions,
states, etc. Since we have already seen many examples of time
clauses introduced by er se ȩr a ‘when’, we will consider this
type first. In their most common usage, time clauses with er
se ȩr a ‘when’ designate a single past event (or, sometimes,
state) which took place while some other action or state was in
progress. Often, the event of the time clause is interpreted as
having interrupted (or intruded upon) this action or state, which
is described in the preceding independent (or main) clause
(cf. 15.1). This is true in the examples below:

(18) a. Ak milsuub er a elii er se ȩr a lȩme a Droteo.

‘I was studying yesterday when Droteo arrived.’

b. Ak mle dibus er se ȩr a lȩmad a dengki.

‘I was away from home when the electricity went out.’

c. A Toki a milȩngȩtmokl ȩr a blai er se ȩr a kbong.

‘Toki was cleaning the house when I arrived.’

Though it is easiest to think of er se ȩr a ‘when’ as a single

unit which introduces a type of time clause, clearly its structure
is much more complex. We speculate that er se ȩr a consists of
a relational phrase er se ‘there, at that time’ followed by an-
other rather unusual relational phrase in which the relational
word ȩr is followed by a sequence having the structure of a
conditional clause (cf. 19.1). Though similar in structure to
conditional clauses—note that the specific third person subjects
Droteo and dengki ‘electricity’ of 18a–b appear obligatorily in
clause-final position—the sequences following er se ȩr a in
18 are indeed difficult to classify. From a practical (rather than

520
 Palauan Reference Grammar

technical) viewpoint, we can say that er se ȩr a ‘when’ requires

a following clause which contains a hypothetical verb form. The
reason for this phenomenon is obscure: since hypothetical verb
forms normally designate unreal events, as we have seen in so
many previous cases, it is totally mysterious why they should
be required in time clauses introduced by er se ȩr a ‘when’,
which refer to actual (or real) events in the past. As we will
see below, all types of Palauan time clauses must—for some un-
known reason—contain hypothetical verb forms.
Time clauses introduced by er se ȩr a ‘when’ involve other
types of temporal relationships than that illustrated in 18. In the
examples below, er se ȩr a introduces a past event or situation
which designates the broad framework within which some other
event occurred. Observe the sentences below:

(19) a. Ak milȩchȩrar a hong er se ȩr a kbo ȩr a stoang.

‘I bought a book when I went to the store.’

b. A Toki a mlo suebȩk a rȩngul er se ȩr a lak lȩme a

Droteo.
‘Toki got worried when Droteo didn’t come.’

c. A Satsko a chiliis er se ȩr a lesa a dȩleb.

‘Satsko ran away when she saw the ghost.’

In 19a, the event of going to the store described in the time

clause with er se ȩr a represents the framework or “context”
for the action of buying the book mentioned in the preceding
clause. And in both 19b–c, the time clause denotes a “back-
ground” event which prompted or caused the event or state de-
scribed in the clause which precedes.
In yet another usage, time clauses with er se ȩr a designate
an event or state which is simultaneous with—i.e., occurs
more or less during the same period of time as—another event
or state. This meaning is illustrated in the sentences below:

(20) a. Ak mle mȩdȩngȩlii a John er se ȩr a kngar ȩr a New

York.
‘I knew John when I was in New York.’

b. Ak millamȩch a dȩkool er se ȩr a kuruul a kall.

521
 22 Reason Clauses, Result Clauses, and Time Clauses

‘I was smoking cigarettes while/when I was preparing

the food.’

c. Ak milȩnguiu ȩr a simbung er se ȩr a longȩtmokl ȩr a

blai a Toki.
‘I was reading the newspaper while/when Toki was
cleaning the house.’

Recall that the expression er se ȩr a can be followed by

certain noun phrases to form temporal phrases (cf. 14.6, ex.
34b) designating time points in the past. A few temporal
phrases of this kind include er se ȩr a sabado ‘last Saturday’, er
se ȩr a (taem ȩr a) mȩkȩmad ‘during the war’, and er se ȩr a
taem ȩr a Siabal ‘during the Japanese times’.
Palauan time clauses introduced by se ȩ l designate an event
or state in the future which will coincide with some other event
or state, as the examples below illustrate:

(21) a. Ak mo olȩngull se ȩl kbo kmȩchas.

‘I’m going to take time off (from work) when I’m an old
woman.’

b. A Droteo a diak lȩbo loureor se ȩl lȩbo lȩchuodȩl.

‘Droteo’s not going to work when he gets old.’

Time clauses introduced by se ȩl can also denote an event which

habitually or regularly coincides with another event, as in the
following sentences:

(22) a. A eangȩd a blȩchoel ȩl mo mȩkngit se ȩl lȩbo ȩl

ngȩbarȩd a eolt.
‘The weather always gets bad when(ever) the wind
becomes westerly.’

b. A Droteo a mȩlamȩch a dȩkool se ȩl losuub.

‘Droteo smokes cigarettes when(ever) he studies.’

It is easiest to consider se ȩl of 21–22 as a single unit meaning

‘when’ or ‘whenever’, even though it probably has a more com-
plicated structure consisting of se ‘there, at that time’ followed
by ȩl, whose function is not clear. In the clause following se

522
 Palauan Reference Grammar

ȩl, hypothetical verb forms are required, and any specific third
person subject (eolt ‘wind’ of 22a) must be moved to clause-final
position.

22.2.1. BEFORE AND AFTER

In order to indicate that a particular action or event occurred
before or after some other action or event, we use time clauses
introduced by ȩr a uche ȩr a ‘before’ and ȩr a uriul ȩr a ‘after’,
respectively. For practical purposes, we can think of these as
single units which must be followed by clauses whose structure
resembles that of conditional clauses (cf. our discussion of er
se ȩr a ‘when’ in 22.2 above). It is clear that they are more
complex, however: they consist of the relational phrases ȩr a
uche and ȩr a uriul, in which the nouns of spatial relationship
uche ‘in front of’ and uriul ‘in back of’ (cf. 14.2.1) are used in
a temporal sense, followed by a rather unusual type of rela-
tional phrase in which the relational word ȩr introduces a clause
rather than a noun phrase.
In 23 below, we illustrate the use of time clauses with ȩr a
uche ȩr a ‘before’, while in 24 we give examples of time clauses
with ȩr a uriul ȩ r a ‘after’:

(23) a. Ak ulȩmuchȩl ȩl mȩsuub ȩr a uche ȩr a kumȩngur.

‘I began to study before (I had) dinner.’

b. A sȩchȩlik a mirrael ȩr a uche ȩr a kbo kmȩrek ȩr a

urerek.
‘My friend left before I finished my work.’

c. A skoki a rirebȩt ȩr a uche ȩr a lȩbo lȩmȩtengȩl ȩr a

skojio.
‘The plane crashed before landing at the airport.’

d. Ng ngar ȩr ngii a bȩtok ȩl tȩkoi ȩl kirek ȩl mȩruul ȩr a

uche ȩr a kbo ȩr a katsudo.
‘There are lots of things I’ve got to do before I go to the
movies.’

e. Kȩ mȩtik a kȩrrȩkar ȩr a uche ȩr a chobo ȩr a blik.

‘You’ll find a tree on the way to/before arriving at my
house.’

523
 22 Reason Clauses, Result Clauses, and Time Clauses

(24) a. Ak mlo mȩchiuaiu ȩr a uriul ȩr a lorael a Toki.

‘I went to sleep after Toki left.’

b. Ak ulȩmuchȩl ȩl mȩsuub ȩr a uriul ȩr a kbo kmerȩk ȩl

omȩngur.
‘I began to study after I finished having dinner.’

c. A skoki a rirebȩt e mȩsesȩb ȩr a uriul ȩr a ltobȩd ȩr a

skojio.
‘The plane crashed and burned after leaving the
airport.’

d. A daob a mlo mȩringȩl ȩr a uriul ȩr a lȩbo ȩl ngȩbarȩd a

eolt.
‘The ocean got rough after the wind became westerly.’

Similar in usage to the “before” and “after” time clauses of

23–4 are temporal phrases (cf. 14.6) in which ȩr a uche and ȩ
r a uriul are followed by a noun phrase designating an event. A
typical example is given below:

(25) Aki milsaod a tȩkoi er a Belau ȩr a{uche/uriul} ȩr a

chȩldȩchȩduch 1 ȩr kȩmam.
‘We discussed the Palauan language{before/after} our
meeting.’

22.2.2. MOVEMENT OF TIME CLAUSES AND TIME WORDS

All of the time clauses discussed above can exchange positions
with the preceding independent (or main) clause. As a result of
this process, the time clause comes to appear in sentence-initial
position, as in the examples below:

(26) a. Se ȩr a kbo ȩr a Guam e ak kilie ȩr a blil a Tony.

‘When I went to Guam, I lived at Tony’s place.’

b. Se ȩr a kisa a John, e ak dilu ȩr ngii.

‘When I saw John, I told him.’

c. Se ȩl lȩbȩkiis, e tȩ mo ȩr a chei.
‘When they get up, they go fishing.’

d. Se ȩl losuub a Droteo, e ng mȩlamȩch a dȩkool.

524
 Palauan Reference Grammar

‘Whenever Droteo studies, he smokes cigarettes.’

e. A uche ȩr a kbo kmȩrek ȩr a subȩlek, e a Droteo a

mirrael ȩl mo ȩr a blil.
‘Before I finished my homework, Droteo went home.’

f. A uriul ȩr a loureor ȩr a sers a Toki, e ng tilȩllib a

chimal.
‘After Toki worked in the garden, she washed her
hands.’

As the above sentences show, two structural changes take place

when the time clause and the independent clause exchange
positions. First, if the relational word ȩ r is the first word of
the time clause introducer (as in er se ȩr a ‘when’, ȩr a uche
ȩr a ‘before’, and ȩr a uriul ȩr a ‘after’), then it is deleted when
the time clause is moved to sentence-initial position. Second,
the shifted independent clause must be introduced by the con-
necting word e ‘and (then)’ (see 25.1). The exchange of time
clause and independent clause discussed here is, of course, op-
tional, and has no effect on the meaning.
Time words (or expressions) such as klukuk ‘tomorrow’,
irȩchar ‘earlier times’, tȩruich ȩl klok ‘ten o’clock’, etc., which
normally occur in temporal phrases introduced by ȩr, can also
be moved to sentence-initial position. Just as in the case of time
clauses, the relational word ȩr is deleted when a time word
(or expression) is moved to the beginning of the sentence, and
the following clause must be introduced by e. Some typical ex-
amples are now given:

(27) a. A klukuk e ak mo ȩr a chei.

‘Tomorrow I’m going fishing.’

b. A irȩchar e a rȩmeteet a ulȩngȩseu ȩr a rȩmechȩbuul.

‘In earlier times, the rich helped the poor.’

c. Tia ȩl mlo mȩrek ȩl rak, e ak mlo ȩr a Guam.

‘Last year I went to Guam.’

d. A tȩruich ȩl klok e tȩ mirrael.

‘At ten o’clock they departed.’

e. A ongeru ȩl ureor, e tȩ mle ȩr a blik.

525
 22 Reason Clauses, Result Clauses, and Time Clauses

‘On Tuesday they came to my house.’

The examples above seem to put special emphasis on the time

of a particular action or event.

526
 23 Relative Clauses

23RelativeClauses
23.1. FUNCTION OF RELATIVE CLAUSES
In Palauan, as in all other languages, the speaker always has
a choice as to how much information he will supply in the sen-
tences which he utters. One of the most common ways of in-
corporating information into a sentence is to expand a noun
phrase (cf. 3.6) by adding to it a sequence called a relative
clause. Before discussing the grammatical characteristics of
relative clauses, we will first examine their function by com-
paring the sentences in the pairs below:

(1) a. A rȩdil a mlo ȩr a kȩlȩbus.

‘The woman went to jail.’

b. A rȩdil ȩl silsȩbii a blai a mlo ȩr a kȩlȩbus.

‘The woman who burned down the house went to jail.’

(2) a. Kȩ mȩdȩngȩlii a ngalȩk?

‘Do you know the child?’

b. Kȩ mȩdȩngȩlii a ngalȩk ȩl dȩngchokl ȩr a bita ȩr a

Toki?
‘Do you know the child who is sitting next to Toki?’

(3) a. Kȩ mla chuiȩuii a hong?

‘Have you read the book?’

b. Kȩ mla chuiȩuii a hong ȩl ngar ȩr a bebul a tebȩl?

‘Have you read the book which is on the table?’

In the a-sentences above, the italicized subject or object

noun phrase (cf. 2.3 and 3.6) consists of a single noun, while
in the b-sentences this noun phrase has been expanded by
placing a relative clause (in bold type) right after the noun.
The parts of the b-sentences in bold type are properly termed
“relative clauses” because they relate a particular piece of in-
formation to the preceding noun. Thus, in 1b, for example,

527
 23 Relative Clauses

(1b) A rȩdil ȩ l silsȩbii a blai a mlo ȩr a kȩlȩbus.

‘The woman who burned down the house went to jail.’

the relative clause expresses a fact or event—namely, silsȩbii a

blai ‘burned down the house’—which is being related to or as
sociated with the preceding noun rȩdil ‘woman’. By relating
the fact ‘burned down the house’ to the noun ‘woman’, the rel-
ative clause of 1b serves to modify, specify, or narrow down
the identity of this noun. In other words, the subject noun
phrase rȩdil ‘woman’ of la is much vaguer (or less exact) in its
reference than the subject noun phrase rȩdil ȩl silsȩbii a blai
‘the woman who burned down the house’ of 1b, where the rel-
ative clause introduced by ȩl gives detailed information about
what the woman did and thereby narrows down her identity. In
the b-sentences of 2–3, the relative clauses which are part of the
object noun phrases likewise narrow down the identity of the
preceding nouns by describing specific states which are asso-
ciated with them.
In the b-sentences of 1–3, the relative clauses have the
structure of ordinary sentences except that they are introduced
by ȩ l and do not contain any overtly-expressed subject noun
phrases. Furthermore, the missing subject noun phrase of the
relative clause is understood as being identical to the noun
phrase which precedes (or heads) the relative clause. In these
respects, relative clauses closely resemble the various types of
dependent clauses which we classified in chap. 15. Two im-
portant features distinguish relative clauses from dependent
clauses, however. First, relative clauses must always be pre-
ceded by a noun (called the head noun), while dependent
clauses are typically preceded by another clause (the inde-
pendent or main clause—cf. 15.1). Second, there are no re-
strictions on the tense of verbs in relative clauses, whereas the
verbs of dependent clauses tend to be in the present tense,
even when the sentence as a whole designates an event or state
in the past.
To summarize what we have said in this section, Palauan
has “expanded” noun phrases with the structure head noun +
relative clause. The distribution 1 of such noun phrases is of
course identical to that of simpler noun phrases: thus, in 1 the
sequence rȩdil ȩl silsȩbii a blai ‘the woman who burned down
the house’ can substitute for the single noun rȩdil ‘woman’ as
subject noun phrase, and in 2 the sequence ngalȩk ȩl dȩng-

528
 Palauan Reference Grammar

chokl ȩr a bita ȩ r a Toki ‘the child who is sitting next to

Toki’ can replace the single noun ngalȩk ‘child’ as object noun
phrase.

23.2. DERIVATION OF RELATIVE CLAUSES

In order to account correctly for the structure and meaning
of Palauan relative clauses, we propose that “expanded” noun
phrases with the structure head noun + relative clause are
simply derived from sequences in which the head noun is im-
mediately followed by a whole sentence. In other words, the
(subject) noun phrase rȩdil ȩl silsȩbii a blai ‘the woman who
burned down the house’ of 1b has its source in the following
structure:

(4) rȩdil [a rȩdil a silsȩbii a blai]

(‘the woman who burned down the house’)

In order to change 4 into the actually-spoken sequence rȩdil

ȩl silsȩbii a blai, we delete the subject rȩdil ‘woman’ of the
bracketed sentence 2 because it is identical to the preceding
head noun, and we insert ȩl as the relative clause introducer.
3
A sequence of the form head noun + bracketed sentence such
as that in 4 cannot be transformed into a grammatical noun
phrase of the form head noun + relative clause unless the
subject noun phrase of the bracketed sentence is identical to
the preceding head noun. If this “identity condition” is not sat-
isfied, then it is impossible to correctly derive a noun phrase of
the form head noun + relative clause. Thus, a source sequence
like the following (cf. 4)

(4’) *rȩdil [a ngalȩk a silsȩbii a blai]

(*‘the woman who the child burned down the house’)

can never be transformed into a grammatical Palauan structure

because the subject (ngalȩk ‘child’) of the bracketed sentence
is different from the preceding head noun (rȩdil ‘woman’). A
similar condition holds in English, as the ungrammatical
English “equivalent” indicates.
In the sentences below we observe further examples of
Palauan relative clauses (italicized):

529
 23 Relative Clauses

(5) a. A buik ȩl mle ȩr a blik er a elii a Droteo.

‘The boy who came to my house yesterday was Droteo.’

b. A rȩsȩchal ȩl millatȩch ȩr a mlai a mȩsaul.

‘The men who were cleaning the canoe are tired.’

c. Ngke ȩl ngalȩk ȩl mȩnguiu ȩr a simbung ng tȩchang?

‘Who’s that child who’s reading the newspaper?’

d. Ak rirȩngȩsii a ngalȩk ȩl lmangȩl.

‘I heard a child (who was) crying.’

e. Ak mildȩchȩmii a buik ȩl mȩlesȩb ȩr a blai.

‘I caught a boy (who was) setting fire to the house.’

f. Ng mȩkngit a rȩnguk ȩr a sȩchȩlik ȩl mlad.

‘I’m sad about my friend who died.’ 4

In 5a–c the sequence head noun + relative clause functions as

subject noun phrase, in 5d–e this same sequence functions as
object noun phrase, and in 5f it appears in a cause phrase
(cf. 14.5) introduced by the relational word ȩ r. The relative
clauses in 5 can all be derived according to the analysis given
above: because the deleted subject of the relative clause would
have been identical to the preceding head noun in the source
structures of 5a–f, the italicized relative clauses of 5 are auto-
matically understood as having subjects identical to this head
noun.

23.3. PASSIVE SENTENCES AS RELATIVE

CLAUSES
In all of the examples given so far, the head noun preceding
a relative clause identifies the doer or agent of the relative
clause if the latter is derived from a transitive sentence. Thus,
in 1b, for instance,

(1b) A rȩdil ȩl silsȩbii a blai a mlo ȩr a kȩlȩbus.

‘The woman who burned down the house went to jail.’

530
 Palauan Reference Grammar

the head noun rȩdil ‘woman’ corresponds to what would be the

agent in the transitive sentence from which the relative clause
is derived—namely,

(6) A rȩdil a silsȩbii a blai.

‘The woman burned down the house.’

In 6 the agent rȩdil ‘woman’ appears as the subject of the tran-

sitive verb silsȩbii ‘burned it down’, while the thing receiving
the effect of the action—namely, blai ‘house’—appears as the
object of silsȩbii. As we saw in 19.7, a transitive sentence like 6
can be transformed in such a way that the agent and the object
exchange positions. In the resulting passive sentence,

(7) A blai a lȩsilsȩbii a rȩdil.

‘The house was burned down by the woman.’

the object of 6—blai ‘house’—has come to appear in subject po-

sition, and the agent—rȩdil ‘woman’—has been moved to the
end of the sentence. Furthermore, the verb of the sentence has
become hypothetical.
Now, it is possible to derive relative clauses that are pre-
ceded by head nouns which identify the object of the relative
clause rather than the agent. All we need to do is make sure that
the relative clause of the sequence head noun + relative clause
corresponds to a passive sentence and that the abovemen-
tioned identity condition is met. Indeed, it is precisely because
of this identity condition that head nouns referring to the object
of a relative clause must be followed by relative clauses derived
from passive sentences. In other words, only in a passive sen-
tence like 7 would the subject noun phrase (which must be iden-
tical to the preceding head noun in order to form a grammatical
relative clause) actually refer to the person or thing receiving
the effect of the action—i.e., the object of the corresponding
active sentence 6. Observe, therefore, the sentence below, in
which the head noun blai ‘house’ refers to the object of the rel-
ative clause:

(8) A blai ȩl lȩsilsȩbii a rȩdil a blil a Toki.

‘The house which the woman burned down/which was
burned down by the woman was Toki’s house.’

531
 23 Relative Clauses

The sequence head noun + relative clause of 8 is derived just

like that of 4, except that the bracketed sentence is passive in
form ( = 7):

(9) blai [a blai a lȩsilsȩbii a rȩdil]

(‘the house which the woman burned down/which was
burned down by the woman’)

Just as in the case of 4, the subject blai ‘house’ of the bracketed

sentence is deleted because it is identical to the preceding head
noun, and ȩl is inserted as the relative clause introducer.
The sentences below illustrate additional instances of rel-
ative clauses that are derived from passive sentences:

(10) a. A blai ȩl lurruul ȩr ngii a Droteo a mle klou.

‘The house which Droteo built was big.’

b. A buik ȩl kulsa er a elii a Tony.

‘The boy whom I saw yesterday was Tony.’

c. A babier ȩl moluchȩs ȩr ngii a mo ȩr a sensei.

‘The letter you’re writing goes to the teacher.’

d. A biang ȩl lȩngilim a Droteo a mle bȩtok.

‘The amount of beer which Droteo drank was
considerable.’

e. A kall ȩl mrirȩllii ng ngar ȩr ker?

‘Where’s the food you made?’

f. Ng ngar ȩr ngii a chisel a Toki ȩl kudȩngei.

‘I’ve got some news to tell you about Toki.’ 5

g. A subȩlek ȩl kbla kbo kmȩrek ȩr ngii a kmal mle

mȩringȩl.
‘The homework which I’ve just finished was very
difficult.’

h. A tȩkoi ȩl lȩbo losuub ȩr ngii a rȩngalȩk a tȩkoi ȩr a

Sina.
‘The language that the children are going to study is
Chinese.’

532
 Palauan Reference Grammar

In 19.7.3 we saw that Palauan passive sentences can also be

derived by moving into subject position a noun phrase which
follows the relational word ȩr in the corresponding active sen-
tence. Thus, the subject of the following passive sentence

(11) A delmȩrab a losuub ȩr ngii a Droteo.

‘(lit.) The room is (being) studied in by Droteo.’

appears in a locational phrase following ȩr in the corre-

sponding active sentence

(12) A Droteo a mȩsuub ȩr a delmȩrab.

‘Droteo studies/is studying in the room.’

As we might expect, it is also possible to have sequences of the

form head noun + relative clause in which the relative clause
is derived from a passive sentence like 11 and the head noun
therefore identifies the location, source, etc., of the action or
event designated by the relative clause. This phenomenon is ob-
served in the sentences below:

(13) a. Tia a delmȩrab ȩl losuub ȩr ngii a Droteo.

‘This is the room that Droteo studies/is studying in.’

b. Tia a basio ȩl lȩbilosii a bȩlochȩl ȩr ngii a Moses.

‘This is the place where Moses shot the pigeon.’

c. Tia a kȩdȩra ȩl lȩmlad ȩr ngii a John er se ȩr a

mȩkȩmad.
‘This is the beach where John died during the war.’

d. A blai ȩl lȩkie ȩr ngii a rȩchad ȩr a Sina ng ngar ȩr ker?

‘Where’s the house that the Chinese live in?’

e. Tia kid a blsibs ȩl lȩtilobȩd ȩr ngii a beab.

‘Here’s the hole that the mouse came out of.’

A further type of passive sentence can be formed by moving

to sentence-initial position a noun phrase which is part of a de-
pendent clause (cf. chap. 15), an object clause (cf. chap. 16),
or the like. Thus, an active sentence like the following,

(14) A Droteo a millasȩm ȩl mȩnga ȩr a ngikȩl.

533
 23 Relative Clauses

‘Droteo tried to eat the fish.’

which contains an object clause following millasȩm ‘tried’ (cf.

16.4) can be transformed into the passive sentence below:

(15) A ngikȩl a lullasȩm ȩl mȩnga ȩr ngii a Droteo.

‘(lit.) The fish was tried to be eaten by Droteo.’ = ‘Droteo
tried to eat the fish.’

The subject of 15—ngikȩl ‘fish’—was originally an object in 14,

where it follows the (imperfective) transitive verb mȩnga ‘eat’ in
the object clause introduced by ȩl. When a singular noun phrase
object in a dependent clause or object clause is passivized in
this way, it must leave behind a pronominal trace (cf. 19.7), as
the presence of ȩr ngii in 15 indicates. Notice that a sentence
like 15 has no acceptable word-for-word equivalent in English.
Now, a passive sentence like 15 can be used as a relative clause
following the head noun ngikȩl ‘fish’ in a source sentence such
as the following:

(16) A ngikȩl [a ngikȩl a lullasȩm ȩl mȩnga ȩr ngii a Droteo] a

mle bȩkȩbau.
(‘The fish which Droteo tried to eat was spoiled.’)

By “processing” the relative clause of 16, we get the following

grammatical Palauan sentence:

(17) A ngikȩl ȩl lullasȩm ȩl mȩnga ȩr ngii a Droteo a mle

bȩkȩbau.
‘The fish which Droteo tried to eat was spoiled.’

Relative clauses whose derivation follows the pattern of 16–17

are given in the additional examples below:

(18) a. A tȩkoi ȩl kmȩduch ȩl mȩlȩkoi ȩr ngii a tȩkoi ȩr a Siabal.

‘The language which I know how to speak is Japanese.’

b. Ng kmal soak ȩl mo omes ȩr a hong ȩl Droteo a dilu ȩl

kmo a Toki a milȩnguiu ȩr ngii.
‘I’d really like to see the book which Droteo said Toki
was reading.’

c. Ng ngar ȩr ngii a ududem ȩl sȩbȩchem ȩl mȩskak?

534
 Palauan Reference Grammar

‘Do you have any money you can give me?’

23.4. EQUATIONAL SENTENCES AS RELATIVE

CLAUSES
As we saw in 18.6, an equational sentence is one which
has the basic structure noun phrase + noun phrase. In such
sentences, the subject noun phrase is equated with or in-
cluded in the category of individuals or things designated by
the second noun phrase. Equational sentences can serve as rel-
ative clauses, as the following types of expressions indicate:

(19) John ȩl sensei

‘John the teacher’

JAL ȩl kombalii ȩr a Siabal

‘JAL, a Japanese company’

New Delhi ȩl kingall ȩr a government ȩr a India

‘New Delhi, the seat of government of India’

blik ȩl {smengt/kȩrrȩkar}
‘my house, which is (made of) {cement/wood},

bilsȩngek ȩl chert
‘my boat, which is an outboard’

ngȩlȩkek ȩl {rȩdil/sȩchal}
‘my {daughter/son}’ (lit. ‘my child who is a {girl/boy})’

The expressions of 19 are all derived from sequences of the

form head noun + relative clause, where the relative clause is
an equational sentence. Thus, John ȩl sensei ‘John the teacher’
has its source in

(20) John [a John a sensei]

(‘John the teacher’)

The bracketed equational sentence A John a sensei ‘John is a

teacher’ of 20 is transformed into a relative clause by deleting
the subject John under identity with the preceding head noun

535
 23 Relative Clauses

and inserting the relative clause introducer ȩl. For convenience

of discussion, we will apply the term appositional structure to
a sequence like John ȩl sensei ‘John the teacher’, in which one
noun phrase is linked to another by the word ȩl (the relative
clause introducer).
By far, the most commonly-used type of appositional structure
in Palauan consists of a noun phrase of possession (usually
a single possessed noun) followed by ȩl and another noun (cf.
3.10). Such appositional structures are used to specify the
function which something serves on a particular occasion and
refer primarily to categories of food and drink. Several ex-
amples, similar to those given in 3.10, are listed below:

(21) imȩlek ȩl biang

‘my (drink of) beer’

kȩlem ȩl udong
‘your noodles’

odimel ȩl babii
‘his pork’

onguled ȩl kukau
‘our taro’

chȩrmek ȩl bilis
‘my dog’

The appositional structures of 21 are derived exactly like those

of 19. Thus, imȩlek ȩ l biang ‘my (drink of) beer’ has a source of
the form head noun + relative clause—namely,

(22) imȩlek [a imȩlek a biang]

(‘my (drink of) beer’)

The bracketed equational sentence A imȩlek a biang ‘My drink

is beer’ of 22 is transformed into a relative clause by processes
which are already familiar to us.

536
 Palauan Reference Grammar

23.5. RELATIVE CLAUSES CONTAINING STATE

VERBS
In the examples below, the relative clauses introduced by ȩl
contain various types of state verbs (cf. chap. 7):

(23) a. A Toki a silsȩbii a blil a Droteo ȩl bȩches.

‘Toki burned down Droteo’s house, which was new.’

b. Ak milȩngȩtakl ȩr a bilsȩngel a Hirosi ȩl tȩlȩmall.

‘I towed Hirosi’s boat, which was broken.’

c. A Toki a milȩngȩtmokl ȩr a delmȩrab ȩr a Droteo ȩl

kikiongȩl.
‘Toki was straightening up Droteo’s room, which was
dirty.’

There is nothing unusual about the derivation of the sequences

head noun + relative clause of 23 from structures of the form
head noun + bracketed sentence.
Contrasting in structure with 23a–c above are sentences in
which a state verb (or an expression containing a state verb)
precedes a particular noun and is linked to it by ȩl. Sequences
of the form state verb + ȩl will be considered a type of modifier
because they describe, modify, or give further information
about the immediately following noun; other types of modi-
fiers will be discussed in detail in chap. 24. For many Palauan
speakers, there is no difference in meaning between sequences
of the form state verb + ȩ l + noun vs. those of the form
head noun + ȩl + state verb (i.e., relative clauses). For some
speakers, however, the sequence state verb + ȩl + noun has a
different interpretation, as indicated in the examples below (cf.
23a–c):

(24) a. A Toki a silsȩbii a bȩches ȩl blil a Droteo.

‘Toki burned down Droteo’s new house (not his old
one).’

b. Ak milȩagȩtakl ȩr a tȩlȩmall ȩl bilsȩngel a Hirosi.

‘I towed the boat of Hirosi’s that was broken.’

c. A Toki a milȩngȩtmokl ȩr a kikiongȩl ȩl delmȩrab ȩr a

Droteo.

537
 23 Relative Clauses

‘Toki was straightening up the room of Droteo’s that

was dirty.’

While the relative clauses of 23 simply provide additional, non-

essential information about the head nouns which they follow,
the modifiers of 24 supply essential identifying information to
distinguish the modified noun from other items with which it
is implied to be in contrast. Thus, comparing 23a and 24a, we
see that the relative clause ȩl bȩches of 23a provides us with
a certain piece of information about Droteo’s house almost as
an after thought (i.e., the house which Toki burned down just
happened to be new), while the modifier bȩches ȩl of 24a iden-
tifies or singles out Droteo’s new house (as opposed to any other
houses he may own) as the one which Toki burned down. In a
similar way, the modifier tȩlȩmall ȩl of 24b makes it clear that it
was Hirosi’s broken boat that the speaker was towing, and not
some other boat of Hirosi’s; and in 24c kikiongȩl ȩl implies that
Toki was cleaning the particular room of Droteo’s that was dirty,
but not any other of his rooms.

23.6. RELATIVE CLAUSES FOLLOWING CHAD AND

KLALO
Since Palauan has no series of “indefinite” words corresponding
to English someone/anyone, something/anything, etc., it simply
makes use of the nouns chad ‘person, man’ and klalo ‘thing’ to
express these concepts. Thus, in the examples below, chad and
klalo are used to refer to a person or thing whose identity is not
known:

(25) a. Ng ngar ȩr ngii a chad ȩr tiang.

‘Somebody’s here.’

b. Ng mlo ȩr a stoa ȩl mo omȩchar a klalo.

‘He went to the store to buy something.’

The nouns chad and klalo are commonly followed by relative

clauses, in which case we have expressions corresponding to
‘someone/anyone who…’ and ‘something/anything which…’ Ob-
serve the examples below:

(26) a. Ng ngar ȩr ngii a chad ȩl osiik ȩr kau.

538
 Palauan Reference Grammar

‘There’s someone (who’s) looking for you.’

b. Ak rirȩngȩsii a chad ȩl mȩngitakl.

‘I heard someone singing.’

c. A chad ȩl diak lȩmȩduch ȩl mȩngikai a mo rȩmos.

‘Anyone who doesn’t know how to swim will drown.’

d. Ng ngar ȩr ngii a klalo ȩl dibus.

‘There’s something missing.’

e. Ng mla ȩr ngii a klalo ȩl mla mȩrȩchorȩch?

‘Was there something stolen?’

When the third person singular emphatic pronoun ngii is fol-

lowed by di ‘only, just’ and a relative clause, we get a rather
forceful expression corresponding to ‘any…at all’. The following
examples are typical:

(27) a. Ngii di ȩl chad a sȩbȩchel ȩl rullii tia ȩl ureor.

‘Anybody at all could do this work.’

b. Ngii di ȩl chad er a Belau a sȩbȩchel ȩl me ȩr tia ȩl klab.

‘Any Palauan at all can come to this club.’

539
 24 Modifiers

24Modifiers
24.1. STATE VERBS AS MODIFIERS
In 23.1 we observed that relative clauses serve to modify,
specify, or provide identifying information about the head
nouns which they follow. Since relative clauses are ultimately
derived from whole sentences (cf. our discussion in 23.2), they
reflect major Palauan sentence types such as transitive sen-
tence, intransitive sentence, passive sentence, and the like.
For instance, in the following sequence of the form head noun +
relative clause

(1) buik ȩl milkodir a bilis

‘the boy who killed the dog’

the italicized relative clause has been derived from the tran-
sitive sentence

(2) A buik a milkodir a bilis.

‘The boy killed the dog.’

The sequence head noun + relative clause of 1 of course has its

source in the structure below, where the transitive sentence 2
appears as a bracketed sentence 1 following the head noun
buik ‘boy’:

(3) buik [a buik a milkodir a bilis]

(‘the boy who killed the dog’)

Can you explain how the abstract structure of 3 is transformed

into the actually-spoken sequence of 1?
Just as a relative clause can have its source in a transitive
sentence, as 1–3 above illustrate, so can a relative clause
be derived from an intransitive sentence containing a state
verb. Thus, as we saw in 23.5, we can have sequences like the
following:

(4) a. delmȩrab ȩl kikiongȩl

‘room which is dirty’

540
 Palauan Reference Grammar

b. blai ȩl bȩches
‘house which is new’

c. chad ȩl smechȩr ȩr a tȩretȩr

‘person who is sick with a cold’

d. sensei ȩl mȩringȩl a tȩkingel

‘teacher who is strict’

Since the italicized relative clauses of 4 contain intransitive

state verbs, the bracketed sentences from which they are de-
rived must necessarily be intransitive sentences containing
these very same state verbs. Thus, 4c, for example, is derived
from the source structure

(4c’) chad [a chad a smechȩr ȩr a tȩretȩr]

(‘person who is sick with a cold’)

by deleting the identical subject noun phrase chad ‘person’ of

the bracketed sentence and inserting the relative clause in-
troducer ȩl .
Now, in addition to the relative clause structure of 4,
Palauan has sequences in which a state verb (or an expression
containing a state verb) precedes a given noun and is linked
to it by ȩl. Thus, with 4a–b, which have the structure noun + ȩl
+ state verb, compare 5a–b, in which the word order has been
switched to state verb + ȩl + noun:

(4) a. delmȩrab ȩl kikiongȩl

‘room which is dirty’

b. blai ȩl bȩches
‘house which is new’

(5) a. kikiongȩl ȩl delmȩrab

‘dirty room’

b. bȩches ȩl blai
‘new house’

541
 24 Modifiers

Though some Palauan speakers feel that the opposing word

orders of 4–5 result in a significant meaning difference, 2 many
others think that the two word orders merely represent a dif-
ference in style which has no effect on the meaning.
We will introduce the term modifier to distinguish the se-
quences of state verb + ȩl in 5 from the relative clauses (ȩl
+ state verb) in 4. As we will see throughout this chapter, it
will be convenient to have a separate term such as modifier
to refer to a number of grammatical constructions in which a
word or group of words is linked by ȩl to a following noun. The
term “modifier” is used because the constructions in question
function to modify, specify, or give further information
about the noun which they precede. Although relative
clauses perform the same function of modifying or specifying a
noun (cf. 23.1), we will not classify them as modifiers because
they follow rather than precede the modified noun. In other
words, we are using modifier primarily as a cover term for
any construction which precedes a noun and is linked to it
by ȩl. So far, we have seen how the word ȩl joins or relates
dependent clauses (cf. chap. 15), object clauses (cf. chap.
16), and relative clauses (cf. chap. 23) to a preceding inde-
pendent clause or head noun. Now, as we examine Palauan
modifiers, we will see that ȩl can also be used to relate a word or
group of words to a following noun (also called a head noun).
The type of modifier shown in 5a–b—namely, a single intran-
sitive state verb followed by ȩl—is very common in Palauan.
Some further examples are listed below:

(6) ungil ȩl eangȩd

‘good weather’

mȩkngit ȩl tȩkoi
‘bad/dirty word’

klou ȩl bilas
‘big boat’

ngodȩch ȩl chad 3
‘strange person’

bȩkȩsius ȩl chad
‘person who swears a lot’

542
 Palauan Reference Grammar

mȩduch ȩl kȩrrȩkar
‘strong tree/wood’

kekȩre ȩl blai
‘small house; (slang) toilet’

kekȩmangȩt ȩl mlai
‘long canoe’

mȩses ȩl buik
‘diligent boy’

Certain common expressions (i.e., groups of words)

containing intransitive state verbs can also be used as modi-
fiers. Note, for example, how the italicized expressions in the
sentences below

(7) a. A sensei a {mȩringȩl/beot} a tȩkingel.

‘The teacher is {strict/easy-going}.’

b. A ngalȩk a bȩralm a rȩngul.

‘The child is lazy.’

are used as modifiers in the following examples:

(8) a. {mȩringȩl/beot} a tȩkingel ȩl sensei

‘{strict/easy-going} teacher’

b. bȩralm a rȩngul ȩl ngalȩk

‘lazy child’

24.2. OTHER VERBS AND NOUNS AS MODIFIERS

As we saw in 24.1 above, many Palauan modifiers consist of a
state verb, or an expression containing a state verb, followed by
the “linking” word ȩ l. Not all modifiers containing verbs require
state verbs, however. For example, a few intransitive action
verbs such as me ‘come’ and mo mȩrek ‘end, stop’ are com-
monly used as modifiers (italicized) in cases like the following:

(9) tia ȩl me ȩl rak

543
 24 Modifiers

‘next year’

tia ȩl mlo mȩrek ȩl buil

‘last month’

The sequences in 9, you will recall, are used with the relational
word ȩr to form temporal phrases (cf. 14.6). In addition, we
occasionally find Palauan speakers using sequences like the
following,

(10) mȩngitakl ȩl chad

‘man who is singing’

mlȩchȩlebȩd ȩl buik
‘boy who was hit/beaten’

in which a transitive action verb in its imperfective form

(mȩngitakl ‘sing’) or ergative form (mlȩchȩlebȩd ‘was/got hit/
beaten’) is part of the modifier construction.
In 24.3–4 below, we will examine two major types of nouns—
demonstratives and number words—which are regularly
used as modifiers. In addition to these two types, certain other
relatively minor groups of nouns can appear in modifier con-
structions. Thus, in the sequences below, a word or expression
designating a quantity is used to modify the following head
noun:

(11) bȩtok ȩl kall

‘lots of food’

bȩtok ȩl chad
‘many people’

kȩsai ȩl malk
‘a bit of chicken’

di tȩlkib ȩl kall
‘a little food’

rokui ȩl taem
‘all the time’

544
 Palauan Reference Grammar

And in the examples below, the question words ngara ‘what

(kind of)?’ (cf. 20.3.1, ex. 46) and tela ‘how much/many?’ (cf.
20.4, ex. 58) function as modifiers:

(12) ngara ȩl tȩkoi

‘what language?’

ngara ȩl hong
‘what kind of book?’

tela ȩl klok
‘what time?’

tela ȩl udoud
‘how much money?’

All of the sequences given in 5–6 and 8–12 above are “ex-
panded” noun phrases with the structure modifier + head noun
(cf. our discussion of head noun + relative clause sequences
at the end of 23.1). Because they are noun phrases, sequences
of the form modifier + head noun have distributional features
identical to those of simpler noun phrases: that is, they occur
as sentence sub ject or object, or following ȩr in relational
phrases. The sentences below illustrate these three distribu-
tional “slots”:

(13) a. A ngodȩch ȩl chad a dȩngchokl ȩr a bita ȩr a Toki.

‘A strange person is sitting next to Toki.’

b. A Satsko a mirruul a bȩtok ȩl kall ȩl kirel a party.

‘Satsko was preparing lots of food for the party.’

c. Kȩ mo ȩr a che ȩr a tela ȩl klok?

‘At what time are you going fishing?’

24.3. DEMONSTRATIVES
Palauan has a special group of words which speakers use when
they wish to point out or draw attention to a particular
person, animal, or thing. Such demonstrative words (or
demonstratives for short) are used in simple (equational) sen-
tences like the following:

545
 24 Modifiers

(14) a. Tia a olȩchȩsek.

‘This is my pencil.’

b. Se a blil a Droteo.
‘That is Droteo’s house.’

c. Tirke tȩ rua tȩchang?

‘Who are those people?’

d. Tia kid a hong.

‘Here is a book.’

e. Ngka kid a Droteo. 4

‘Here is Droteo.’

As the examples of 14 show, Palauan demonstratives (e.g. tia

‘this’, se ‘that’, ngka ‘this person’, tirke ‘those people’) are
never preceded by the word a, which introduces most other
Palauan nouns and noun phrases (cf. 2.6).
In order to use demonstratives correctly, the Palauan
speaker must take three factors into account. First, he must
choose a different set of demonstratives depending on whether
what he is referring to (i.e., the referent) is a person, animal,
or thing. Second, he must use different forms for singular vs.
plural referents. And, finally, he must consider the relative
distance of the referent from himself and the hearer.
When making reference to non-living things, the Palauan
speaker uses the set of demonstrative words given below. In
considering the relative distance of the referent from himself
and the hearer, the speaker must make a three-way distinction.
Thus, different demonstratives must be used to designate (i)
something near both the speaker and hearer, (ii) something
near the hearer but not the speaker, and (iii) something far from
both the speaker and hearer. These three categories of distance
are listed across the top of the chart below: 5

(15) Demonstratives Referring to Things

Location near near hearer but far from

of speaker and far from speaker speaker and
referent: hearer hearer

546
 Palauan Reference Grammar

singular tia ‘this’ tilȩcha ‘that’ se ‘that (over

there)’

plural aika ‘these’ ailȩcha ‘those’ aike ‘those

(over there)’

Demonstrative words referring to human beings can be

classified in a parallel way, as the following chart shows:

(16) Demonstratives Referring to Human Beings

Location near near hearer but far from

of speaker and far from speaker and
referent: hearer speaker hearer

singular ng(i)ka ‘this ngilȩcha ‘that ng(i)ke ‘that

person’ person’ person (over
there)’

plural tirka ‘these tirilȩcha ‘those tirke ‘those

people’ people’ people (over
there)’

If we compare the demonstratives of 15–16 with each other,

we immediately notice some significant structural patterns.
Most of these words are composed of two morphemes (or
meaning-bearing units), the first one specifying the type of ref-
erent (person or thing, singular or plural), and the second spec-
ifying the relative distance of the referent from the speaker
and hearer. Looking at the demonstratives found in the second
column—namely,

(17) tilȩcha
‘that (thing)’

ailȩcha
‘those (things)’

ngilȩcha
‘that person’

tirilȩcha
‘those people’

547
 24 Modifiers

—we may conclude that the shared part -lȩcha (or perhaps -
ilȩcha) is a distinct morpheme which conveys the meaning ‘lo-
cated near hearer but far from speaker’ common to these four
words. The difference in meaning among the demonstratives
of 17 must therefore be due to the fact that different mor-
phemes (ti-, ai-, ngi-, and tiri-) are prefixed to the shared mor-
pheme -lȩcha. These four morphemes have distinct meanings,
as follows:

(18) ti- ‘non-living thing—singular’

ai- ‘non-living things—plural’
ngi- ‘human being—singular’
tiri- ‘human beings—plural’

The different morphemes of 18 combine with the morpheme

-lȩcha to form the contrasting demonstrative words of 17.
The fact that Palauan demonstrative words are organized
according to a “logical” system becomes clear to us when we
see that the four morphemes of 18 (which specify whether
the referent is a person or thing, singular or plural) recur
just where we would expect them. While ti- ‘non-living
thing—singular’ is somewhat exceptional (see below), the three
remaining morphemes occur consistently as follows:

a. ai- is found in all demonstrative words referring to plural

nonliving things—i.e., aika ‘these (things)’, ailȩcha ‘those
(things)’, and aike ‘those (things) (over there)’,

b. ngi- is found in all demonstrative words referring to a

singular human being—i.e., ng(i)ka ‘this person’, ngilȩcha
‘that person’, and ng(i)ke ‘that person (over there)’. 6

c. tiri- is found in all demonstrative words referring to plural

human beings—i.e., tirka ‘these people’, tirilȩcha ‘those
people’, and tirke ‘those people (over there)’. 7

We can now see that in addition to the morpheme -lȩcha ‘located

near hearer but far from speaker’, Palauan demonstrative words
also contain the distinct morphemes -ka ‘located near speaker
and hearer’ and ke- ‘located far from speaker and hearer’.

548
 Palauan Reference Grammar

The morpheme ti- ‘non-living thing—singular’ is associated

with some irregularities. Thus, where we would expect the
demonstrative word *tika for ‘this (thing)’, we instead have tia,
in which the k of -ka ‘located near speaker and hearer’ has
been lost. In addition, we do not have any demonstrative word
like *tike for ‘that (thing) (over there)’, but the totally unre-
lated word se, which consists of one morpheme only. The three
demonstratives referring to singular non-living things are dis-
tinguished from other demonstratives in that they commonly
appear in locational phrases (cf. 14.2) with ȩr, as illustrated
below:

(19) ȩr tia
‘here, in this place’

ȩr tilȩcha
‘there, in that place’

ȩr se
‘over there, in that place
(over there)’

Although Palauan demonstrative words referring to an-

imals do not involve any new morphemes or morpheme combi-
nations, they represent an unusual (and unexplainable) mixture
of terms from the two sets already given (15 and 16). The
demonstrative words referring to animals are listed below:

(20) Demonstratives Referring to Animals

Location near speaker near hearer but far from speaker

of and hearer far from speaker and hearer
referent:

singular ng(i)ka ‘this ngilȩcha ‘that ng(i)ke ‘that

animal’ animal’ animal (over
there)’

plural aika ‘these ailȩcha ‘those aike ‘those

animals’ animals’ animals (over
there)’

549
 24 Modifiers

As the chart in 20 shows, demonstratives referring to singular

human beings (see 16) are also used to refer to singular an-
imals, while demonstratives denoting plural things (see 15)
also serve to denote plural animals.
One of the most commonly-occurring modifier construc-
tions (cf. 24.2 above) in Palauan consists of a demonstrative
word linked by ȩl to a following head noun. Sequences of this
kind are used when the speaker wishes to identify a person,
animal, or thing in terms of where it is located with relation to
himself and the hearer. The following examples are typical:

(21) tia ȩl hong

‘this book’

se ȩl kȩrrȩkar
‘that tree (over there)’

aika ȩl kahol
‘these boxes’

ng(i)ke ȩl ngalȩk
‘that child (over there)’

tirilȩcha ȩl sensei 8
‘those teachers’

aika ȩl charm
‘these animals’

Can you explain why tia (but not ng(i)ka) is required before the
head noun hong ‘book’, why ng(i)ke (but not se) is required be-
fore the head noun ngalȩk ‘child’, and so on?
Because the sequences of 21 are “expanded” noun phrases
with the structure modifier + head noun, they have the distri-
butional features common to all noun phrases. Thus, in the sen-
tences below, the sequences of 21 are used as sentence subject
or object or as part of a relational phrase with ȩr:

(22) a. Aika ȩl charm a babii.

‘These animals are pigs.’

b. Ng(i)ke ȩl ngalȩk a milosii a malk.

550
 Palauan Reference Grammar

‘That child (over there) shot the chicken.’

c. Ng tȩcha a lilȩchȩsii tia ȩl hong?

‘Who wrote this book?’

d. A bȩlochȩl a silebȩk ȩr se ȩl kȩrrȩkar.

24.4. NUMBERS
Palauan numbers are used, of course, to count or specify the
number of persons, things, etc. being considered. They are
much more complicated than English numbers because dif-
ferent sets must be chosen according to the type of thing being
counted. In this respect, Palauan numbers resemble Palauan
demonstratives, which—as we saw in 24.3 above—occur in
three sets depending on whether the referent is a person,
animal, or thing.
The numbers of Palauan are found in three major sets,
which most speakers use frequently and in a uniform manner,
and several relatively minor sets, which show considerable
variation from one speaker to another and which certain
speakers (especially younger ones) no longer even use. Like the
demonstratives examined above, Palauan numbers by and large
form a system in which various morphemes consistently recur.
As we will see below, most number words contain a morpheme
specifying a number plus another morpheme (a prefix or suffix)
identifying the class or category of what is being counted.
One major set of Palauan numbers is used for counting
various units of time such as hours of the clock, days, years,
and the like. The numbers in this set (which we will identify as
Set I) are listed below:

(23) Set I

1 ta 11 tȩruich mȩ a ta
2 eru 12 tȩruich mȩ a eru
3 ede 13 tȩruich mȩ a ede
4 eua 14 tȩruich mȩ a eua
5 eim 15 tȩruich mȩ a eim
6 elolȩm 16 tȩruich mȩ a elolȩm
7 euid 17 tȩruich mȩ a euid

551
 24 Modifiers

8 eai 18 tȩruich mȩ a eai

9 etiu 19 tȩruich mȩ a etiu
10 tȩruich 20 lluich

Looking at the numbers from 1 to 10 in Set I, we see that except

for ta ‘one’ and tȩruich ‘ten’, which appear to be single mor-
phemes, all of the numbers consist of the prefix e- followed by
some other morpheme. The prefix e- is a separate morpheme
which specifically identifies the category of things being
counted—namely, units of time. This prefix combines with the
various number morphemes—i.e.,

(24) -ru ‘two’ -lolȩm ‘six’

-de ‘three’ -uid ‘seven’
-ua ‘four’ -ai ‘eight’
-im ‘five’ -tiu ‘nine’

—to form number words referring to units of time.

The number words between 11 and 20 in Set I are not par-
ticularly difficult to analyze. Though lluich ‘twenty’ is best ana-
lyzed as a single morpheme, it may in some way be related to
tȩruich ‘ten’ (note the common sequence -uich). The numbers
between 11 and 19 are merely expressions of the form ‘ten and
one’, ‘ten and two’, etc., in which tȩruich ‘ten’ is connected by
mȩ ‘and’ (see 25.4) to the following smaller number. 9
Palauan number words are commonly used as modifiers (cf.
24.1 above). Thus, in the “expanded” noun phrases below, a
number word from Set I is linked by ȩl to a following head
noun. Since the number words of Set I are used to count units
of time, the head nouns in the sequences below are time words
such as sils ‘day’, rak ‘year’, etc.:

(25) ta ȩl sikang ‘one hour’

eru ȩl sils ‘two days’
ede ȩl klȩbȩse ‘three nights’
eua ȩl kȩbȩsȩngil 10 ‘the fourth (of some month)’
eim ȩl buil ‘five months’
elolȩm ȩl rak ‘six years’
euid ȩl klok ‘seven o’clock’
tȩruich mȩ a ta ȩl klok 11 ‘eleven o’clock’

552
 Palauan Reference Grammar

Whereas all Palauan speakers use the number words of Set

I for counting units of time, only certain speakers (mostly
of the older generation) use these same words for counting
flat, square objects such as sheets of paper, books, etc. Such
speakers will therefore use noun phrases like the following:

(26) eru ȩl babier ‘two letters/sheets of

paper’
ede ȩl hong ‘three books’
eua ȩl siasing ‘four photos’

For counting flat, square objects the majority of Palauan

speakers prefer the number words chimo ‘one’, tȩblo ‘two’, klde
‘three’, etc., which can refer to a wide range of non-living things
(see Set III in 29 below).
A second major set of Palauan number words is used exclu-
sively for counting human beings. Thus, with Set I compare
the following, which we will designate as Set II:

(27) Set II

1 ta 11 tȩruich mȩ a ta
2 teru 12 tȩruich mȩ a teru
3 tede 13 tȩruich mȩ a tede
4 teua 14 tȩruich mȩ a teua
5 teim 15 tȩruich mȩ a teim
6 telolȩm 16 tȩruich mȩ a telolȩm
7 teuid 17 tȩruich mȩ a teuid
8 teai 18 tȩruich mȩ a teai
9 tetiu 19 tȩruich mȩ a tetiu
10 tȩruich 20 lluich

When we compare the number words of Set I and Set II, we can
make the following observations:

a. In both sets, the number words for ‘one’ (ta), ‘ten’ (tȩruich),
and ‘twenty’ (lluich) are identical.

b. In both sets, the number words from 2 to 9 consist of two

morphemes—a prefix identifying the class or category of
what is being counted plus a number morpheme. In Set I, the
prefix is e-, while in Set II it is te-. Both of these prefixes

553
 24 Modifiers

combine regularly with the number morphemes of 24 to form

the various number words. While the prefix e- refers primarily
to units of time, the prefix te- designates human beings.

c. In both sets, the number words from 11 to 19 are merely

expressions of the form ‘ten and one’, ‘ten and two’, etc.,
where tȩruich ‘ten’ is connected to another number by mȩ
‘and’. Thus, for example, tȩruich mȩ a ede ‘thirteen (days,
months, etc.)’ and tȩruich mȩ a tede ‘thirteen people’ are
exactly parallel in structure.

Because the number words of Set II refer exclusively to

human beings, they can only occur preceding human head
nouns, as in the following:

(28) teru ȩl chad ‘two people’

teim ȩl sensei ‘five teachers’
tetiu ȩl buik ‘nine boys’
tȩruich mȩ a tede ȩl chad 12 ‘thirteen people’

A third major set of Palauan number words is used in count-

ing animals and a large variety of non-living things. Observe the
words below, which constitute Set III:

(29) Set III

1 chimo 11 tachȩr mȩ a chimo

2 tȩblo 12 tachȩr mȩ a tȩblo
3 klde 13 tachȩr mȩ a klde
4 kloa 14 tachȩr mȩ a kloa
5 kleim 15 tachȩr mȩ a kleim
6 kllolȩm 16 tachȩr mȩ a kllolȩm
7 klȩuid 17 tachȩr mȩ a klȩuid
8 kleai 18 tachȩr mȩ a kleai
9 kltiu 19 tachȩr mȩ a kltiu
10 tachȩr 20 lluich

Comparing the number words of Set III with those of Sets I and
II (cf. 23 and 27 above), we can make the following statements:

554
 Palauan Reference Grammar

a. The number words chimo ‘one’, tȩblo ‘two’, and tachȩr ‘ten’
in Set III are completely different from the comparable number
words in the other two sets. The word lluich ‘twenty’, however,
is the same.

b. The number words from 3 to 9 in Set III consist of the

already familiar structure prefix + number morpheme. Here,
the prefix kl- (lengthened in several cases by the vowel e or ȩ)
is added to the various number morphemes shown in 24. The
number morpheme -ua ‘four’ irregularly changes to -oa when
kl- is prefixed.

c. The number words from 11 to 19 in Set III resemble those of

Sets I and II in that they follow the pattern ‘ten and one’, ‘ten
and two’, etc.
Since the Set III number words refer to animals as well as
non-living things, they typically occur as modifiers in
“expanded” noun phrases like the following:

(30) chimo ȩl malk ‘one chicken’

tȩblo ȩl blai ‘two houses’
klde ȩl hong ‘three books’
kloa ȩl ringo ‘four apples’
kleim ȩl kluk ‘five dollars’
kllolȩm ȩl lius ‘six coconuts’
klȩuid ȩl kahol ‘seven boxes’
tachȩr ȩl uel ‘ten turtles’

24.4.1. NUMBERS ABOVE 20

So far we have only discussed Palauan numbers between 1
and 20. The numbers above 20 follow familiar patterns and are
therefore not difficult to analyze. First we shall examine those
numbers which are multiples of 10—namely,

(31) 30 okede 70 okeuid

40 okoua 80 okai
50 okeim 90 oketiu
60 okolȩm 100 dart

Except for dart ‘one hundred’, which is a single morpheme,

the number words listed above consist of two morphemes—the
prefix ok- (lengthened in most cases by e or o) and one of the

555
 24 Modifiers

number morphemes listed in 24. In okolȩm ‘sixty’ the first l

of the number morpheme -lolȩm ‘six’ is lost when ok- is pre-
fixed. Like lluich ‘twenty’, the number words in ok- (and dart
‘one hundred’ as well) are used in all three major number sets
(cf. 23, 27, and 29 above). The prefix ok- indicates that the ac-
companying number (morpheme) is to be multiplied by 10.
Just as the numbers between 11 and 19 have the structure
‘ten and one’, ‘ten and two’, etc., the numbers between 21 and
29, 31 and 39, etc. take the form ‘twenty and one’, ‘thirty and
one’, and so on. Such numbers are used as modifiers in the ex-
amples below:

(32) a. lluich mȩ a teru ȩl chad

‘twenty-two people’

lluich mȩ a tȩblo ȩl blai

‘twenty-two buildings’

b. okede mȩ a eim ȩl sils

‘thirty-five days’

okede mȩ a teim ȩl sensei

‘thirty-five teachers’

c. okeuid mȩ a etiu ȩl rak

‘seventy-nine years’

okeuid mȩ a kltiu ȩl kluk

‘seventy-nine dollars’

Can you explain why the number 22 must have the form lluich
mȩ a teru before the head noun chad ‘person’, but takes the
form lluich mȩ a tȩblo before the head noun blai ‘house’? Does
the same principle hold for the other pairs?
Numbers which are multiples of 100 are formed by using
a number word from Set I as a modifier of dart ‘one hundred’.
Thus, we have eru ȩl dart ‘two hundred’, ede ȩ l dart ‘three
hundred’, eua ȩ l dart ‘four hundred’, etc. The number 1000
and its multiples are formed in a similar way—i.e., ta ȩl telael
‘one thousand’, eru ȩl telael ‘two thousand’, ede ȩl telael ‘three
thousand’, and so on.

556
 Palauan Reference Grammar

24.4.2. MINOR NUMBER SETS

In this section we will list several Palauan number sets that are
of relatively minor importance because their use is infrequent
or restricted in some way. Our comments on these number sets
will accordingly be very brief.
When counting off one number after another—i.e., when
saying “one—two—three” etc. in sequence—Palauan speakers
use the group of number words below (Set IV):

(33) Set IV

1 ta 7 uid
2 oru 8 iai
3 ode 9 itiu
4 oua 10 machod
5 oim 11 machod mȩ a ta
6 malo 12 machod mȩ a oru, etc.

While ta ‘one’, malo ‘six’, and machod ‘ten’ are single mor-
phemes, the number words from 2 to 5 and from 7 to 9 in Set IV
consist of a prefix (o-, u-, or i-) followed by a number morpheme
(cf. 24). Set IV number words higher than 10 (i.e., machod mȩ a
ta ‘eleven’, machod mȩ a oru ‘twelve’, etc.) are very infrequently
used.
In order to count long things such as pencils, fish, canoes,
bananas, and the like, most Palauan speakers use the number
words in Set III. However, certain speakers (mostly in the older
generation) use the following special number words (Set V) for
counting between one and five long objects:

(34) Set V

1 teluo
2 eruo
3 edeuo
4 euaiuo
5 eimuo

The number words of Set V all contain the suffix -uo. In the
numbers from 2 to 5 this suffix is added to the corresponding
number words of Set I, with minor phonetic changes. Thus,
the final u of eru ‘two’ is deleted when -uo is suffixed to form

557
 24 Modifiers

eruo ‘two (long objects)’; and an extra i is inserted when eua

‘four’ combines with -uo to give euaiuo ‘four (long objects)’. The
number teluo ‘one (long object)’ is formed by adding -uo to what
appears to be a special number morpheme tel- ‘one’.
Another number set which involves a suffix (rather than a
prefix) is the following (Set VI), which is used occasionally to
count bunches of bananas:

(35) Set VI
1 teliud 6 elolȩmiud
2 ereiud 7 euidiud
3 edeiud 8 eaiud
4 euaiud 9 etiuiud
5 eimiud 10 tȩruich ȩl iud

The number words of Set VI show a general pattern already fa-

miliar to us. Thus, in the numbers from 2 to 9 a suffix of the
form -iud is added to the corresponding number word of Set
I, with certain small phonetic changes. (Can you identify these
changes in the number words ereiud ‘two bunches’ and eaiud
‘eight bunches’?) In teliud ‘one bunch’, the suffix -iud has been
added to a special number morpheme tel- (cf. the tel- of teluo
‘one (long object)’ in Set V). 13 Finally, in tȩruich ȩ l iud ‘ten
bunches’ iud is used as a separate word—i.e., as a head noun
preceded by the modifier tȩruich ‘ten’.

24.4.3. USE OF NUMBER WORDS IN SENTENCES

In 25, 26, 28, 30, and 32 above we gave many examples of how
number words can be used as modifiers of a following head
noun. Such sequences, which have the structure number word
+ ȩl + head noun, function as noun phrases and therefore
occur regularly as sentence subject or object, or as part of a re-
lational phrase introduced by ȩr. A few sentences showing the
distribution of these “expanded” noun phrases (italicized) are
given below:

(36) a. A Droteo a mlo ȩr a che ȩr a euid ȩl klok.

‘Droteo went fishing at seven o’clock.’

b. A dart ȩl chad a mle ȩr a ochȩraol.

‘One hundred people came to the money-raising party.’

558
 Palauan Reference Grammar

c. A rȩngalȩk a milkoad a tachȩr ȩl uel.

‘The children killed ten turtles.’

d. Ng tȩcha a milskau a kleim ȩl kluk?

‘Who gave you five dollars?’

As the examples of 36 illustrate, “expanded” noun phrases con-

taining number words as modifiers must always be introduced
by the word a (cf. 2.6).

24.4.4. ORDINAL NUMBERS

Palauan ordinal numbers, which correspond to English ‘first’,
‘second’, ‘third’, etc., are used to indicate the order or rank of
something. The ordinal numbers from 1 to 10 are listed below:

(37) kot ‘first’

ongeru ‘second’
ongede ‘third’
ongeua ‘fourth’
ongeim ‘fifth’
ongelolȩm ‘sixth’
ongeuid ‘seventh’
ongeai ‘eighth’
ongetiu ‘ninth’
ongetȩruich ‘tenth’

Except for the special word kot ‘first’, which is a single mor-
pheme, the ordinal numbers of 37 all consist of at least two
morphemes. It is difficult to decide between the following two
analyses: either a prefix ong- is added to the number words of
Set I, or a prefix onge- combines with the number morphemes
of 24.
We have already seen how ordinal numbers are used in
certain types of time expressions. Thus, in 14.6, ex. 34e we
examined temporal phrases containing the following expres-
sions for the first five days of the week:

(38) kot ȩl ureor ‘Monday’

ongeru ȩl ureor ‘Tuesday’
ongede ȩl ureor ‘Wednesday’
ongeua ȩl ureor ‘Thursday’
ongeim ȩl ureor ‘Friday’

559
 24 Modifiers

The expressions of 38 are “expanded” noun phrases in which

an ordinal number serves as a modifier of the head noun ureor
‘work’. Therefore, the literal meaning of kot ȩl ureor, ongeru ȩl
ureor, etc. is something like ‘the first (day of) work’, ‘the second
(day of) work’, etc.
We also saw in 14.6, ex. 34f that the months of the year
are expressed in Palauan as follows:

(39) kot ȩl buil ‘January’

ongeru ȩl buil ‘February’
ongede ȩl buil ‘March’
ongeua ȩl buil ‘April’
ongeim ȩl buil ‘May’
ongelolȩm ȩl buil ‘June’
ongeuid ȩl buil ‘July’
ongeai ȩl buil ‘August’
ongetiu ȩl buil ‘September’
ongetȩruich ȩl buil ‘October’
ongetȩruich mȩ a ta ȩl buil ‘November’
ongetȩruich mȩ a ongeru ȩl buil
‘December’

As 39 shows, all of the expressions for the months of the year

contain an ordinal number serving as a modifier of the head
noun buil ‘month’. These expressions therefore have the literal
meanings ‘the first month’, ‘the second month’, and so on.
The expressions below further illustrate the use of ordinal
numbers as modifiers:

(40) kot ȩl chad ‘first man’

ongeru ȩl sils ‘second day’
ongede ȩl babii ‘third pig’
ongetȩruich ȩl ngalȩk ‘tenth child’

24.4.4.1. Ordinal Numbers Followed by

Specifying Clauses
As we saw in 15.7.7, Palauan has a small number of special
verbs which must always be followed by a specifying clause
(cf. 15.7) introduced by ȩl. Such verbs—e.g. blȩchoel ‘always (
= invariably do)’, dirrek ‘also’, and ko ‘just’—are somewhat dif-

560
 Palauan Reference Grammar

ficult to understand because their closest English equivalents

do not happen to be verbs. These special verbs (italicized) are
used in sentences like the following:

(41) a. Ak blȩchoel ȩl mȩruul a kȩlir.

‘I always prepare their food.’

b. A sȩchȩlik a dirrek ȩl mong.

‘My friend is also going.’

c. A Droteo a ko ȩl mȩsubang.
‘Droteo’s just gotten (a chance) to study.’

Like blȩchoel ‘always’, dirrek ‘also’, and ko ‘just’ of 41, the

ordinal number kot ‘first’ can be followed by a specifying clause.
Thus, in the sentences below, kot corresponds to ‘do first’ or ‘do
(something) before/ahead of someone else’, and the specifying
clause introduced by ȩl designates the activity involved:

(42) a. Kȩ kot ȩl mo omȩngur.

‘You go ahead and eat first.’

b. Kȩ ma 14 kot ȩl mo ȩr a skuul, e ngak

ekong.
‘You go on ahead to school, and then I’ll
follow.’

c. Ak kot ȩl rȩmurt. 15
‘I’ll run first (in the race, etc.).’

When kot ‘first’ is followed by a specifying clause containing

a state verb, the resulting meaning corresponds to English -
est in words like biggest, fastest, etc. In other words, kot + ȩ l
+ state verb indicates that someone or something possesses a
certain quality in a higher degree than anyone or anything else
being considered. The following examples are typical:

(43) a. A John a kot ȩl kebȩlu ȩl ngalȩk ȩr a skuul.

‘John is the stupidest student.’

b. Ng tȩcha a kot ȩl bȩkȩrurt?

‘Who runs the fastest?’

561
 24 Modifiers

c. A Oreor a kot ȩl klou ȩl beluu er a Belau.

‘Koror is the biggest town in Palau.’

24.5. EXPRESSIONS WITH TWO OR MORE

MODIFIERS
In the sections above we have seen that Palauan demonstra-
tives, numbers, state verbs, and even action verbs can be used
as modifiers. For purposes of simplicity, we only gave examples
in which a particular head noun is preceded by a single modifier
(e.g. tia ȩl hong ‘this book’, klou ȩl mlai ‘large car’, etc.). As the
following examples indicate, however, it is often the case that
a head noun is preceded by a sequence of two (or more) modi-
fiers:

(44) tia ȩl me ȩl buil

‘next month’

tia ȩl mlo mȩrek ȩl rak 16

‘last year’

tirka ȩl teru ȩl chad

‘these two men’

tirke ȩl tede ȩl ungil ȩl sensei

‘those three good teachers’

ngka ȩl kekȩre ȩl babii

‘this small pig’

aike ȩl kloa ȩl charm

‘those four animals’

klde ȩl mȩchȩtngaid ȩl oluchȩs

‘three thin pencils’

Observing a three-modifier expression like tirke ȩl tede ȩl ungil

ȩl sensei ‘those three good teachers’, we find that the normal
order of modifiers is demonstratives—numbers—state (or
action) verbs.

562
 Palauan Reference Grammar

24.6. QUALIFYING WORDS

So far we have examined a large number of expressions with
the structure modifier + ȩl + head noun, in which a modifier
(a demonstrative, number, state verb, etc.) modifies or narrows
down the identity of the head noun to which it is linked by
ȩl. In this section we will briefly look at another kind of mod-
ification—one in which a qualifying word like kmal ‘very’, di
‘only’, etc. qualifies or limits the meaning of a directly-fol-
lowing verb. Thus, in the examples below, we list the com-
monest qualifying words, together with examples showing how
they are used in “expanded” verb phrases of the form qualifying
word + verb:

(45) kmal ‘very, often’

a. Ng kmal ungil a rrȩllem.

‘What you’ve made is very good.’

b. Ng kmal mle mȩkngit a eangȩd.

‘The weather was very poor.’

c. A Droteo a kmal diak losuub.

‘Droteo hardly ever studies.’

(46) di ‘only, just’ 17

a. A dengua a di osisiu.
‘The phone (number) is just the same (as before).’

b. A sȩchȩlik a mlo ȩr a Guam ȩl di mo milil.

‘My friend went to Guam just to fool around.’

c. A ngȩlȩkek a di diak losuub.

‘My child just doesn’t ever study.’

d. Ak di milsuub a tȩkoi ȩr a Merikel er a elii.

‘All I did was study English yesterday.’

e. Ng di ngar ȩr kau.
‘It’s up to you.’

(47) dirk ‘still’

563
 24 Modifiers

a. Ng dirk ngar ȩr ngii a kall?

‘Is there still any food left?’

b. A ngȩlȩkem a dirk mȩchiuaiu.

‘Your child is still sleeping.’

(48) locha ‘perhaps’

a. Ak locha mo ȩr a skuul ȩr a klukuk.

‘Perhaps I’ll go to school tomorrow.’

b. Tȩ locha mla mo smechȩr.

‘Perhaps they’ve gotten sick.’

(49) kilo ‘almost, nearly, what if…?’ 18

a. A ngȩlȩkek a kilo {mad/rȩmos}.

‘My child almost {died/drowned}.’

b. A bilsȩngek a kilo mo ȩr a uche ȩr a klaidȩsachȩl.

‘My boat almost won the race.’

c. Kilo mo a Droteo?
‘What if Droteo went?’

564
 25 The Connecting Words mȩ
and e

25TheConnectingWordsmȩande
25.1. REVIEW OF COMPLEX SENTENCES WITH
MȨ AND E
In preceding chapters we have observed numerous examples
of complex sentences which are formed by joining two simpler
sentences with the words mȩ ‘and (so)’ or e ‘and (then)’. These
words are called connecting words because they connect two
simple sentences into one and relate the ideas which they rep-
resent. Thus, in 22.1 we saw that two simple sentences such as
those in 1a–b below can be joined by the connecting word mȩ
‘and (so)’ to form the more complicated sentence of 2:

(1) a. A bȩchik a mle smechȩr.

‘My wife was sick.’

b. Ng dimlak kbo ȩr a party.

‘I didn’t go to the party.’

(2) A bȩchik a mle smechȩr mȩ ng dimlak kbo ȩr a party.

‘My wife was sick, so I didn’t go to the party.’

When the two independently-occurring sentences 1a–b are

joined by the connecting word mȩ ‘and (so)’ to form the more
complex sentence 2, they become clauses of this more complex
sentence. In other words, 1a has become the independent
(or main) clause of 2, while 1b has become a result clause.
The connecting word mȩ ‘and (so)’ not only serves to join both
clauses of 2 but also functions to introduce the result clause,
which explains what happened as a result (or consequence) of
the state described in the preceding independent clause.
The following are additional examples of complex sentences
in which mȩ ‘and (so)’ connects an independent clause with a
following result clause:

(3) a. Ng mla mo mȩkngit a eangȩd mȩ ng diak lsoak ȩl mo ȩr a

Peleliu.

565
 25 The Connecting Words mȩ and e

‘The weather’s become bad, so I don’t want to go to

Peleliu.’

b. Kȩ mlȩkȩra mȩ kȩ rirebȩt ȩr a chȩldukl?

‘How did you fall off the dock?’

c. A sensei a dilu ȩr ngak mȩ ak olȩngȩseu ȩr a Toki.

‘The teacher told me to help Toki.’ 1

As we saw in 22.1, the simple sentences 1a–b can also be

combined to form a complex sentence of the following type:

(4) Ng dimlak kbo ȩr a party e le a bȩchik a mle smechȩr.

‘I didn’t go to the party because my wife was sick.’

In 4, the independently-occurring sentences of 1 have been

combined in the opposite order from that observed in 2, and
they have been joined instead by the connecting word e (fol-
lowed by le). In other words, 1b has become the independent
clause of 4, while la has become a reason clause. The con-
necting word e and the element le are best considered as a
single unit which joins the two clauses of 4 and at the same
time introduces the reason clause, which explains the cause or
reason for the event or situation described in the preceding in-
dependent clause.
Use of the connecting word e ‘and (then)’ is of course not
confined to the expression e le ‘because’. Thus, as discussed
in 19.1 and 19.3, e is characteristically used in Palauan con-
ditional sentences to join the conditional clause and the
consequential clause. Observe, therefore, the conditional sen-
tences below:

(5) a. A lȩngar ȩr ngii a ududek, e ak mo ȩr a

Guam.
‘If I had money, (then) I’d go to Guam.’

b. A lsȩkum ng ungil a eangȩd, e tȩ mo ȩr a

chei.
‘If the weather’s good, (then) they’ll go
fishing.’

566
 Palauan Reference Grammar

In 5a–b, e serves to introduce a consequential clause, which

describes an event (or state) which can take place only if the
event (or state) of the preceding conditional clause becomes an
actual fact. In 5a, for example, the consequential clause e ak
mo ȩr a Guam ‘(then) I’d go to Guam’ describes an event which
would result from or be a consequent of the realization of
the event of the conditional clause a lȩngar ȩr ngii a ududek ‘if
I had money’.
When a time clause (cf. 22.2 and 22.2.2) exchanges posi-
tions with an independent clause and thereby comes to appear
at the beginning of a sentence, the following independent
clause must be introduced by the connecting word e ‘and
(then)’. Thus, compare the a-sentences with the b-sentences in
the pairs below:

(6) a. Ak kilie ȩr a blil a Tony er se ȩr a kbo ȩr a Guam.

‘I lived at Tony’s place when I went to Guam.’

b. Se ȩr a kbo ȩr a Guam, e ak kilie ȩr a blil a Tony.

‘When I went to Guam, I lived at Tony’s place.’

(7) a. A Droteo a mȩlamȩch a dȩkool se ȩl losuub.

‘Droteo smokes cigarettes whenever he studies.’

b. Se ȩl losuub a Droteo, e ng mȩlamȩch a dȩkool.

‘Whenever Droteo studies, he smokes cigarettes.’

Whereas the a-sentences of 6–7 have an independent clause

followed by a time clause, the b-sentences have a structure in
which a time clause in sentence-initial position is joined to the
following independent clause by the connecting word e. The
connecting word e must also be inserted when a time word
(or expression) such as klukuk ‘tomorrow’ or eim ȩl klok ‘five
o’clock’ has been (optionally) shifted to sentence-initial position,
as in the following:

(8) a. A klukuk e ng me ȩr a blik.

‘Tomorrow he’s coming to my house.’

b. A eim ȩl klok e a rȩsȩchȩlim a mirrael.

‘At five o’clock your friends departed.’

567
 25 The Connecting Words mȩ and e

25.2. FURTHER USES OF THE CONNECTING

WORD MȨ
We have already seen that the connecting word mȩ is used to
relate a result clause to a preceding independent clause
and that in such cases mȩ corresponds to English ‘so’ or ‘and
so’. The connecting word mȩ can also be used to join two in-
dependent clauses which are parallel in structure and which
present information of more-or-less equal significance. In such
cases, mȩ simply establishes a rather loose connection between
the events, states, etc. designated by the independent clauses
and therefore corresponds to English ‘and’ rather than ‘so’ or
‘and so’. Observe the examples below:

(9) a. A Merikel a klou ȩl beluu, mȩ a Belau a kekȩre ȩl beluu.

‘America is a big country, and Palau is a small country.’

b. A bilek a bȩcheleliu, mȩ a bilel a Droteo a bȩkȩrȩkarȩd.

‘My shirt is white, and Droteo’s shirt is red.’

c. A Droteo a ngalȩk ȩr a skuul, mȩ a Toki a sensei.

‘Droteo is a student, and Toki is a teacher.’

d. A Toki a mo ȩr a sers, mȩ a Droteo a mo ȩr a chei.

‘Toki is going to (work in) the garden, and Droteo is
going fishing.’

e. Ak mȩriik ȩr a mȩkȩsokȩs, mȩ a Toki a mȩlemȩd ȩr a

ulaol.
‘I’m sweeping the yard, and Toki’s mopping the floor.’

It is not difficult to recognize that the independent clauses

connected by mȩ ‘and’ in the examples above are parallel in
structure and convey parallel kinds of information. The two
independent clauses of 9b, for instance, consist of a subject
noun phrase (bilek ‘my shirt’ and bilel a Droteo ‘Droteo’s
shirt’) followed by an intransitive state verb (bȩcheleliu
‘white’ and bȩkȩrȩkarȩd ‘red’), and both of them describe the
color of someone’s clothes. In a similar way, the independent
clauses of 9e each contain a subject noun phrase (ak ‘I’ and
Toki), a transitive action verb (mȩriik ‘sweep’ and mȩlemȩd
‘mop’), and an object noun phrase (mȩkȩ sokȩs ‘yard’ and
ulaol ‘floor’), and both of them describe household activities

568
 Palauan Reference Grammar

which their respective subjects are pursuing. Can you see that
9a, 9c, and 9d also contain clauses which are parallel in
structure?
The connecting word mȩ ‘and’ also occurs in imperative
sentences, which are used to give orders or commands (cf.
19.5). In the examples below, each of the clauses connected by
mȩ contains an imperative verb form:

(10) a. Bo ȩr a bita mȩ mlȩngir a oles.

‘Go next door and borrow a knife.’

b. Bo ȩr a blil a Toki mȩ mchȩtȩklii a ngalȩk.

‘Go to Toki’s house and carry the child
back.’

c. Bo ȩr a blim mȩ bo bad. 2
‘Go home and go to sleep.’

25.2.1. THE EXPRESSION mȩ a lȩ chub

The connecting word mȩ occurs with the words a lȩchub to form
mȩ a lȩ chub, an expression corresponding to English ‘or’. Be-
cause the origin of a lȩ chub is obscure, it is easiest to con-
sider mȩ a lȩ chub as a single unit. When mȩ a lȩchub joins two
clauses, the second clause is often introduced by the connecting
word e. 3 Note the following examples:

(11) a. A rȩsȩchal a mo ȩr a che mȩ a lȩchub e tȩ mo mȩliich a

lius.
‘The men (either) go fishing, or they go make copra.’

b. Kȩ mo ȩr a katsudo mȩ a lȩchub e kȩ mo ȩr a party?

‘Are you going to the movies, or are you going to the
party?’

c. Kȩ mȩrael mȩ a lȩchub e kȩ di kiei?

‘Are you leaving, or will you stay?’

As the above examples show, mȩ a lȩ chub is used to connect

two alternative courses of action which the speaker is de-
scribing or asking about. In the question sentences 11b–c the
connecting word mȩ can be omitted, resulting in a slight change
of meaning for certain Palauan speakers. Whereas 11b–c with

569
 25 The Connecting Words mȩ and e

mȩ seem to imply that the person asking the question prefers

the second of the two alternatives mentioned, this connotation
is lost when mȩ is omitted. Thus, someone uttering 11c would
really prefer the person addressed to stay rather than go; ex-
ample 11c without mȩ, however, would imply that the speaker
has no particular preference about which of the two alternatives
the hearer should choose.

25.3. FURTHER USES OF THE CONNECTING

WORD E.
As reviewed in 25.1 above, the connecting word e is used
in conditional sentences to introduce the consequential
clause. Because the consequential clause names an event,
state, etc. which can come about only if the event, state, etc.
of the preceding conditional clause has become a reality, the
consequential clause necessarily follows the conditional clause
in time. For this reason, the connecting word e which introduces
consequential clauses takes on a temporal meaning and corre-
sponds closely to English ‘and then’.
The abovementioned temporal meaning of e ‘and then’ is
reflected in the complex sentences below, where the second
clause (introduced by e) names an event, state, etc. which
follows (or is expected to follow) that of the first clause:

(12) a. A Toki a me e mȩngȩtmokl ȩr a blai.

‘Toki comes and (then) cleans the house.’

b. Tȩ mȩruul ȩr a kall e mȩrael.

‘They make the food and then (they) leave.’

c. Ak luchȩsii a babier e mo send ȩr ngii.

‘I’ll write the letter and then send it off.’

Although there is no overtly-expressed subject in the second

clause of each of the examples above, speakers automatically
interpret this clause as having a subject identical to that of the
preceding clause. Thus, it seems as if a sentence like 12b has its
source in the following:

(12b’) Tȩ mȩruul ȩr a kall e tȩ mȩrael.

‘They make the food, and then they leave.’

570
 Palauan Reference Grammar

Most Palauan speakers would transform the rather awkward

sentence of 12b’ into 12b by deleting the subject tȩ ‘they’ of
the second clause. Such deletion is possible, of course, only be-
cause the subject tȩ ‘they’ of the second clause is identical to
that of the first clause. Thus, in sentences like 12a–c a single
subject (that of the first clause) is sufficient for the proper inter-
pretation.
In the sentences below, which involve past time, the event of
the second clause (introduced by e) is asserted to have followed
that of the first clause:

(13) a. Ak milles ȩr a ngikȩl e milȩngat ȩr ngii.

‘I cut the fish and (then) smoked it.’

b. Ak dilsȩchii a mlai e chilsbȩrbȩrii.

‘I carved the canoe and (then) painted it.’

c. A Toki a ulȩmȩngur e mirrael.

‘Toki ate and (then) left.’

d. Ng mirrael a Droteo e ak mlo mȩchiuaiu.

‘Droteo left, and (then) I went to sleep.’

In 13a–c the subject of the second clause has been deleted be-
cause it is identical with that of the first. However, in 13d each
clause has a different subject (Droteo in the first clause 4 and ak
‘I’ in the second), and therefore the subject of the second clause
cannot be deleted.
In the examples of 13, each of the clauses connected by e
contains a verb in the past tense. It is also possible to have sen-
tences such as the following, where the verb of the first clause
is in the past tense, but the verb of the clause introduced by e
is in the present tense:

(14) a. Ak milles ȩr a ngikȩl e mȩngat ȩr ngii.

‘I cut the fish and (then) smoked it.’

b. Ak dilsȩchii a mlai e chosbȩrbȩrii.

‘I carved the canoe and (then) painted it.’

c. A Toki a ulȩmȩngur e mȩrael.

‘Toki ate and (then) left.’

571
 25 The Connecting Words mȩ and e

d. Ng mirrael a Droteo, e ak mo mȩchiuaiu.

‘Droteo left, and (then) I went to sleep.’

The difference between the sentences of 13 vs. 14 is very dif-

ficult to pinpoint because the opinions of speakers vary so
widely. Some speakers feel that use of the past vs. present tense
in the second clause has no effect whatsoever on the meaning;
for these speakers, the past tense morpheme (-il- or -l-) can pre-
sumably be omitted from the verb of the second clause in 14a–d
because the past tense verb form in the first clause already pro-
vides enough information about the time of the events or actions
in question.
Other speakers feel that the sentences of 13 differ from
those of 14 in terms of whether the speaker is emphasizing or
focusing on the event of the first clause or that of the second
clause. Thus, the tense sequence past—past of 13 seems to em-
phasize the importance of the action in the second clause, while
the tense sequence past—present of 14 puts focus on the action
in the first clause. For this reason, 13b and 14b would be re-
sponses to different questions, as the following dialogs illus-
trate:

(15) A: Kȩ chilsbȩrbȩrii a mlai?

‘Did you paint the canoe?’

B: (cf. 13b) Chochoi. Ak dilsȩchii (a mlai) e chilsbȩrbȩrii.

‘Yes. I carved it and (then) painted it.’

(16) A: Kȩ dilsȩchii a mlai?

‘Did you carve the canoe?’

B: (cf. 14b) Chochoi. Ak dilsȩchii (a mlai) e chosbȩrbȩrii.

‘Yes. I carved it and (then) painted it.’

In 15, A’s question to B shows that A is interested in whether

or not B painted the canoe; therefore, it is appropriate for B to
use the tense sequence past—past in his response, since this se-
quence places emphasis on the event which A is interested in.
The situation in 16 is just the opposite: here, A wants to know
whether or not B carved the canoe, and the tense sequence
past—present in B’s response focuses more attention on this
event than on the event of painting the canoe.

572
 Palauan Reference Grammar

Certain Palauan speakers distinguish between the sentences

of 13 and 14 in yet another way. For these speakers, the tense
sequence past—past of 13 involves a succession of two com-
pleted events in the past, while the tense sequence
past—present of 14 involves a completed past event followed by
an incomplete (or ongoing) present event. Thus, the meanings
of 13b vs. 14b, repeated here for convenience, would differ as
indicated in the English translations:

(13b) Ak dilsȩchii a mlai e chilsbȩrbȩrii.

‘I carved the canoe and (then) painted it.’

(14b) Ak dilsȩchii a mlai e chosbȩrbȩrii.

‘I carved the canoe and now I’m painting it.’

In 12–14 above, we gave sentences in which the connecting

word e ‘and then’ establishes a sequential time relationship be-
tween the two clauses which it joins. Thus, in all of those sen-
tences, the clause following e designates an event, state, etc.
which takes place (or took place) after the event, state, etc.
of the clause preceding e. As we shall now see, the connecting
word e is not restricted to such cases of sequential time rela-
tionship. For example, we observe that e establishes a simul-
taneous time relationship in the sentences below: here, the
two clauses joined by e designate events, states, etc. which are
happening at the same time. In such cases, e corresponds to
English ‘while’ rather than ‘and then’:

(17) a. A Droteo a milȩnguiu a hong, e a Toki a milȩchiuaiu ȩr a

ulaol.
‘Droteo was reading books {and/while} Toki was
sleeping on the floor.’

b. Ak milluchȩs a babier, e a sȩchȩlik a mirruul a kall.

‘I was writing letters {and/while} my friend was
preparing food.’

c. Ak milȩnguiu ȩr a simbung e omȩngur.

‘I was reading the newspaper while eating.’

d. Ngara mȩ kȩ di dȩchor e omȩngur?

‘Why are you standing up while eating?’

573
 25 The Connecting Words mȩ and e

e. Kȩ omȩngur e mȩngȩdȩchȩduch?
‘(Why) are you eating and talking at the same time?’

f. A Droteo a chad ȩr a omȩnged e chad ȩr a sers.

‘Droteo is both a fisherman and a farmer.’

g. Ng kmal smechȩr e mȩtkung.

‘He’s very sick and about to die.’

In 17a–b the clauses joined by e have different (overtly-ex-

pressed) subjects, while in 17c–g the subject of the second
clause has been deleted because it is identical to that of the first
clause.
The connecting word e can also be used to join two clauses
which are put into relatively strong contrast with each other. A
few typical examples are given below:

(18) a. A malk a beot a chȩral, e a ngais a mȩringȩl a chȩral.

‘Chickens are cheap, but eggs are expensive.’

b. A Droteo a mȩtongakl, e a Toki a kekȩdeb.

‘Droteo is tall, but Toki is short.’

c. A sils a ngmasȩch ȩr a chongos, e mo ngmelt ȩr a

ngȩbard.
‘The sun rises in the east {and/but} sets in the west.’

The sentences of 18 are similar to those of 9 above in that the

clauses joined by the connecting word are parallel in structure.

25.3.1. THE EXPRESSION e ng di

The connecting word e occurs with the words ng di to form
e ng di, an expression corresponding to English ‘but’. Though
written as three words and having a literal meaning something
like ‘and (then) it’s just that…’, the expression e ng di is best
considered a single unit which functions to connect two clauses
which are in contrast or opposition with each other. The use of
this expression is illustrated in sentences like the following:

(19) a. A Toki a mle soal ȩl mo ȩr a Guam, e ng di ng mla mo

diak a ududel.
‘Toki wanted to go to Guam, but her money ran out.’

574
 Palauan Reference Grammar

b. Ak mlo ȩr a party e ng di a Droteo a dimlak lsȩbȩchel ȩl

mong.
‘I went to the party, but Droteo couldn’t go.’

c. Ak ilȩko ȩr a blim e ng di kȩ mle dibus.

‘I came to your house, but you were out.’

25.4. COORDINATE NOUN PHRASES

When two or more nouns (or noun phrases) are joined by the
connecting word mȩ, we have a coordinate noun phrase.
Since coordinate means ‘equal in status or rank’, it is appro-
priate as an identifying term for the type of noun phrase under
discussion, simply because each of the nouns (or noun phrases)
in a coordinate noun phrase functions equally in the sentence.
Thus, in the sentence below, where the coordinate noun phrase
Droteo mȩ a Toki ‘Droteo and Toki’ occurs in sentence subject
position, both of the nouns joined by me (Droteo and Toki) serve
equally as subjects of the verb mirruul ‘prepared’:

(20) A Droteo mȩ a Toki a mirruul a kall.

‘Droteo and Toki prepared the food.’

Because both Droteo and Toki are interpreted equally as

subjects of mirruul ‘prepared’ in 20, a good number of linguists
would propose that 20 is derived from a source sentence
containing two parallel clauses joined by mȩ (cf. 25.2 above),
one clause having Droteo as the subject of mirruul and the other
having Toki as the subject of mirruul. This source sentence is
given below:

(21) A Droteo a mirruul a kall, mȩ a Toki a mirruul a kall.

‘Droteo prepared the food, and Toki prepared the food.’

Although grammatical, the source sentence 21 is somewhat

awkward because each of the clauses repeats the sequence
mirruul a kall ‘prepared the food’. For this reason, Palauan
speakers normally transform 21 into the shortened or con-
densed version 20. When this transformation takes place, the
subjects Droteo and Toki, which are the only dissimilar elements
in the clauses of 21, are combined into a coordinate noun

575
 25 The Connecting Words mȩ and e

phrase (Droteo mȩ a Toki) functioning as sentence subject. In

addition, only a single occurrence of the sequence mirruul a kall
‘prepared the food’, which is common to both clauses of 21, ap-
pears after the coordinate noun phrase subject of 20.
Because coordinate noun phrases are a type of noun phrase,
they of course have the same distributional characteristics as
other noun phrases. Thus, in the example below, we observe a
coordinate noun phrase functioning as sentence object (rather
than sentence subject, as in 20):

(22) A Toki a ousbech a babier mȩ a oluchȩs.

‘Toki needs some paper and a pencil.’

In this example, each member (babier ‘paper’ and oluchȩs

‘pencil’) of the coordinate noun phrase babier mȩ a oluchȩs
‘paper and pencil’ functions as the object of ousbech ‘need’.
Thus, it is likely that 22 is derived from a source sentence
containing two parallel clauses connected by mȩ, one clause
having babier ‘paper’ as the object of ousbech and the other
having oluchȩs ‘pencil’ as the object of this same verb. This
source sentence is represented below:

(23) A Toki a ousbech a babier, mȩ a Toki a ousbech a oluchȩs.

‘Toki needs some paper, and Toki needs a pencil.’

The source sentence 23 is extremely awkward to say, and

therefore all Palauan speakers would automatically transform
it into 22. By this process of transformation, the objects babier
‘paper’ and oluchȩs ‘pencil’, which are the only different ele-
ments in the clauses of 23, are condensed into a coordinate
noun phrase (babier mȩ a oluchȩs ‘paper and pencil’) serving as
sentence object. Since the same subject-verb sequence (a Toki
a ousbech ‘Toki needs’) occurs in each clause of 23, it appears
only once in the shortened sentence 22.
The examples below further illustrate the use of coordinate
noun phrases (italicized) as sentence subject:

(24) a. A rȩngalȩk ȩr a skuul mȩ a rȩsensei ȩr tir a mlo ȩr a

Guam.
‘The students and their teachers went to Guam.’

b. A blil a Toki mȩ a blil a Satsko a milsesȩb.

576
 Palauan Reference Grammar

‘Toki’s house and Satsko’s house burned down.’

c. Kau mȩ ngak a mo ȩr a stoang.

‘You and I will go to the store.’

d. A Droteo mȩ ngak a mlo ȩr a party.

‘Droteo and I went to the party.’

e. Ng tȩcha mȩ a tȩcha a ulȩbȩngkem ȩl mo ȩr a Hawaii?

‘Who (pl.) went with you to Hawaii?’

f. Ngara mȩ a ngara a chomoruul ȩl kirel a party?

‘What things are you making for the party?’

In 24c–d the coordinate noun phrase contains one or two em-

phatic pronouns (cf. 4.4), and in 24e–f two occurrences of the
same question word (t ȩcha ‘who?’ or ngara ‘what?’) are joined
by mȩ to form a coordinate noun phrase (cf. 20.8).
In the sentences below, we further illustrate the use of coor-
dinate noun phrases (italicized) as sentence object:

(25) a. Ak mla mȩnga a diokang mȩ a ngikȩl mȩ a chȩmang.

‘I’ve eaten tapioca, fish, and crab.’

b. Ak milsa a Droteo mȩ a Toki mȩ a Helen er a elii.

‘I saw Droteo, Toki, and Helen yesterday.’

c. Ak milstȩrir a rȩngalȩk ȩr a skuul mȩ a rȩsensei ȩr tir.

‘I saw the students and their teachers.’

Even though the coordinate noun phrase objects of 25b–c both

involve groups of individuals and are therefore plural, the form
of the preceding perfective verb (milsa vs. milstȩrir) is deter-
mined by whether the directly following noun phrase (i.e., the
first member of the coordinate noun phrase object) is itself sin-
gular or plural. Thus, milsa ‘saw him/her/it’ is required in 25b
because the immediately following noun phrase (Droteo) is sin-
gular, whereas milstȩrir ‘saw them’ must occur in 25c because
it directly precedes the plural noun phrase rȩngalȩk ȩr a skuul
‘students’.

577
 25 The Connecting Words mȩ and e

The sentences below show how coordinate noun phrases

(italicized) can occur in further distributional “slots” commonly
filled by noun phrases:

(26) a. Tia a delmȩrab ȩr a Droteo mȩ a Toki.

‘This is Droteo and Toki’s room.’

b. Ak mlo shopping ȩr a Ala Moana mȩ a Waikiki.

‘I went shopping at Ala Moana and Waikiki.’

c. Kȩ mlo ȩr ker mȩ ker?

‘What places did you go to?’

d. Ng mlo ȩr a Merikel er oingara mȩ oingarang?

‘On what occasions did he go to America?’

In all of the examples above, a coordinate noun phrase is part

of a relational phrase introduced by ȩr (cf. chap. 14). Can you
identify the type of relational phrase involved in each of the sen-
tences of 26?
In 17.2 we saw that the subject noun phrase of any Palauan
sentence can be shifted to the right of the verb phrase, leaving
the appropriate pronominal trace in the original subject po-
sition. Coordinate noun phrases in sentence subject position can
also be shifted in this manner, as the following pairs illustrate:

(27) a. A Droteo mȩ a Toki a ‘Droteo and Toki are

kausȩchȩlei. friends.’

b. Tȩ kausȩchȩlei a Droteo
mȩ a Toki.

(28) a. A Helen mȩ a sȩchȩlil a ‘Helen and her friend went

mlo ȩr a mubi. to the movies.’

b. Tȩ mlo ȩr a mubi a Helen

mȩ a sȩchȩlil.

In 27b and 28b above, which involve subject shifting, the 3rd
pers. (human) pl. non-emphatic pronoun tȩ ‘they’ remains as
a pronominal trace because the shifted coordinate noun phrase
subjects are (human) plural.

578
 Palauan Reference Grammar

Now, sentences 27b and 28b can be further transformed by

taking the first member of the shifted coordinate noun phrase
and moving it back to sentence-initial position, where it re-
places the pronominal trace tȩ. When this process of preposing
takes place (cf. 17.5), a pronominal trace of the preposed noun
phrase—this time in the form of an emphatic pronoun—must
appear in the position of the first member of the shifted co-
ordinate noun phrase. These changes are seen clearly in the
following sentences, which are derived from 27b and 28b, re-
spectively:

(27) c. A Droteo a kausȩchȩlei ngii mȩ a Toki.

‘Droteo is friends with Toki.’

(28) c. A Helen a mlo ȩr a mubi ngii mȩ a sȩchȩlil.

‘Helen went to the movies with her friend.’

Because the preposed subjects of 27c and 28c (Droteo and

Helen) are singular, the 3rd pers. sg. emphatic pronoun ngii ‘he,
she, it’ appears in the position from which these subjects have
been removed.

25.4.1. COORDINATE NOUN PHRASES WITH mȩ a lȩ chub

In 25.2.1 above we saw that the expression mȩ a lȩchub ‘or’
can be used to connect two clauses. This expression can also
join two nouns (or noun phrases), resulting in a coordinate noun
phrase of the form A mȩ a lȩchub B ‘A or B’. Observe the ex-
amples below:

(29) a. A Cisco mȩ a lȩchub a Tony a me ȩr a party.

‘Either Cisco or Tony is coming to the party.’

b. Ng tȩcha a ungil ȩl sensei? Ng Toki mȩ a lȩchub a

Droteo?
‘Who’s a better teacher—(is it) Toki or Droteo?’

579
 Notes

Notes
1 SOUNDS AND SPELLING OF PALAUAN

1. The symbol ȩ is used in this text to identify one of the

ways in which the Palauan letter e is pronounced. See 1.4.2.
below for a complete discussion.

*2. Some speakers pronounce final k as a voiceless velar

fricative [x] if a vowel precedes and if the syllable is
stressed—e.g. chȩrmek [ɂərmέx] ‘my animal’. This phe-
nomenon appears to be more common among younger
speakers.

3. The phonetic symbol [ŋ] represents a velar nasal pro-

nounced as a separate syllable. See 1.3.5 below.

*4. In this word, the second r actually comes from the infix -l-,
which forms resulting state verbs (see 7.7 for a complete
explanation). The form rruul ‘made, done’ is derived from r-
l-uul by a process of total assimilation: when l follows r,
it changes to r, resulting in the sequence rr [r̄]. Resulting
state verbs in which the infix -l- does not change include kla
‘eaten’ and nglim ‘drunk (up)’.

5. In this example, which is actually a two-word sentence, the

pronunciation of the word ng ‘it’ changes to (syllabic) [ṃ]: in
other words, the pronunciation of ng ‘it’ assimilates to that
of the initial bilabial consonant b [b] of the following word
(cf. ng bilis [ṃbilis] ‘it’s a dog’). Another case in which the

580
 Notes

pronunciation of ng ‘it’ assimilates to that of the initial con-

sonant of the following word was observed in 18—namely,
ng til [ṇtil] ‘it’s her purse’.

6. A few one-syllable words which always or nearly always

occur unstressed are found to contain ȩ—e.g., tȩ ‘they’, and
ȩr ‘of, to, at, in.’

*7. If we were to spell the words under discussion as omdasw,

ochadw, and kwokw, it would no longer be necessary to treat
them as a special category, since they would be similar to
the items of 31 in having a word-final consonant cluster.
Thus, the appearance of the schwa release would be regular
rather than exceptional. Because the Palauan spelling
system uses only u and i (but not w or y), it tends to obscure
the generality of certain phonological processes. Another
case in point is tȩruich ‘ten’, which is pronounced with
a schwa release: [tǝruyɂə]. If this word were spelled with
a final consonant cluster— tȩruych—then the appearance
of the schwa release would make more sense. The author
agrees with Wilson 1972 that in a generative-phonological
analysis of Palauan it is necessary to set up underlying forms
containing w and y and that, ideally, the orthography should
reflect these forms.

*8. The words under discussion are resulting state verbs

(see 7.7 and cf. note 4 above). They are formed by infixing
-l- after the first consonant of (verb) stems like tub ‘spit’,
dangȩb ‘cover’, sesȩb ‘burn’, and chat ‘smoke’. While the re-
sulting state verbs given in 33 require a schwa to intervene
between the t, d, s, or ch of the stem and the following l,
this is not necessary in other resulting state forms like kla
‘eaten’, nglim ‘drunk (up)’, and bloes ‘shot’.

*9. The actual pronunciation of these examples could be ex-

plained more clearly if the Palauan spelling system used w
and y in addition to u and i. These words would then be

581
 Notes

spelled as follows: chuywiy ‘reads it’, kiysiy ‘digs it’, kiwtiy

‘cuts it’, and siwsiy ‘cures it’. The appearance of [ǝ] would
be due to the fact that consonant clusters like yw or y/w+
C could not occur and would have to be broken up by an in-
tervening schwa (cf. the nonoccurrence of tl, dl, sl, etc. il-
lustrated in 33). To take a further example, note how the
current Palauan spelling of mȩchiuaiu ‘sleep’ gives us little
clue to the actual pronunciation of the word—namely, [mǝɂi-
wáyǝwǝ]. This pronunciation could easily be predicted if the
spelling were mechiwayw: the final schwa release would be
due to the (underlying) final cluster -yw, while the schwa be-
tween y and w would serve to break up a non-permissible
consonant cluster.

10. The letter h, as in hong ‘book’, is found almost exclusively in

words borrowed from Japanese or English (see 1.5.f below
and note 11).

11. The letter h is used only very rarely in the spelling of native
Palauan words. The most obvious examples are found in
the unusual pair of words hngong (exclamation to draw at-
tention to a pleasant smell) and hngob (exclamation to draw
attention to an unpleasant smell). Another possible example
is hal ‘stop!’, but this word might be borrowed from German
Halt ‘stop, halt’.

2 PALAUAN NOUNS

1. Recall that the asterisk mark * is used in this text to identify

any words, expressions, or sentences which cannot or do not
appear in the Palauan language.

582
 Notes

2. When referring to certain common household animals such

as dogs, pigs, etc., some Palauan speakers can use the plural
pronoun tȩ, which, as mentioned above, normally implies a
human plural subject. Thus, these speakers accept a dialog
like the following, in which tȩ refers to two or more dogs:

A: A bilis tȩ ‘What are the dogs doing?’

mȩkȩrang?
B: Tȩ kaiuȩtoir ȩr a ‘They’re chasing each other in the
sers. garden.’

3. In another group of pronouns, too, we can see the impor-

tance of the distinction between human and non-human. As
we will see in 4.9, there is a set of object pronouns which
appear as endings on action words (verbs). The various pro-
nouns in this set have many different forms, of which several
are illustrated in bold type in the words chillȩb ȩd ak ‘hit
me’, chillȩbȩd au ‘hit you’, mȩdȩngȩl ii ‘knows him/her/it’,
and mils tȩrir ‘saw them’. Among this set of object pro-
nouns, the ending tȩrir is used only if the object referred
to is human plural; for non-human plural (i.e., animals
and living or non-living things), no ending is added to the
verb. Observe, therefore, the following sentences, in which
we have forms of omes ‘see’:

a.Ak milstȩrir a rȩsȩchȩlim. ‘I saw your friends.’

b.Ak miles a bȩtok ȩl ius. ‘I saw lots of crocodiles.’
c. Ak miles a bȩtok ȩl bilas. ‘I saw lots of boats.’

The use of tȩrir vs. no ending when referring to something

plural is exactly parallel to the above-mentioned use of
the subject pronouns tȩ and ng.

4. This sentence describes the distribution of duties for a

Palauan money-raising party (ochȩraol).

583
 Notes

5. As these forms show, the words ngika ‘this person’ and

ngike ‘that person’ are often shortened by omitting the
vowel i.

6. Actually, these sentences can be interpreted in yet other

ways, since the distinction between specific vs. non-specific
is also not found among subject nouns. The translations
given for 36a–b involve specific noun subjects, but we can
also interpret these sentences as having non-specific noun
subjects. Thus, 36a, for example, can also mean ‘A dog is
sleeping there.’ or ‘Some dogs are sleeping there.’

3 NOUN POSSESSION

1. In 1.5.d.4 we simply referred to these suffixes as possessor

pronouns in order to keep the presentation relatively un-
complicated.

2. Much of the information in this and the following sections is

given in Mancill and Woods 1969:35–46, where many useful
lists of words and forms are provided.

*3. Wilson 1972, §§ 3.3–4 suggests that the u, i, and a of

these suffixes are really part of the underlying forms of noun
stems. In other words, there are underlying vowel-final
noun stems like rengu, buchi, and chara, which simply add -
k, -m, etc. as possessor suffixes and in which the final under-
lying vowel deletes in the independent forms. Contrasting
with this type are underlying consonant-final noun stems
like charm ‘animal’, udoud ‘money’, etc., which add -ek, -em,
etc. as possessor suffixes.

4. As the English equivalents imply, the forms bȩchȩmam and

bȩchȩrir refer to one single spouse who has two or more
husbands (or wives). By contrast, the words rȩbȩchȩmam

584
 Notes

‘our (excl.) spouses’ and rȩbȩ chȩrir ‘their spouses’, which

have the plural prefix rȩ-, make reference to more than one
spouse.

5. Note the loss of one of the l’s in the possessed forms of kall
‘food’.

6. Note the complete loss of the syllable chȩ- in the possessed

form of this word. The same phenomenon is observed in
chȩdam ‘father’— dȩmal ‘his father’.

7. Another example of this type is dui ‘title’—dial vs. dui ‘co-

conut frond (for burning)’—diul.

8. See 3.4.4 below for a discussion of the phonetic changes

observed in biskȩlȩngel.

*9. The possessed form of ralm ‘water’ is derived approxi-

mately as follows. The underlying form is *ralm + é l; first,
the a following r is deleted in unstressed syllable, giving
*rlm + él; then, the initial cluster of three consonants is sim-
plified to give the surface form lmel.

10. In the possessed form of klȩngit ‘sin’, the sequence -lȩ- has
been lost completely. Furthermore, the ng of the possessed
form becomes syllabic between two consonants, giving
[kŋtíl].

11. See 3.4.4 below for a discussion of the ng which appears in

the possessed forms of bilas ‘boat’. Some of the possessed
nouns given in subsequent word lists will also show this ng.

12. Oddly enough, both tet ‘handbag’ and tut ‘breast’ have iden-
tical irregular forms for the first person plural exclusive,
second person plural, and third person plural. These are
dȩtmam, dȩtmiu, and dȩtȩrir, respectively. Note the unpre-
dictable alternation between initial t and d.

13. This discussion is based on Wilson 1972, §§ 2.4.3.3 and 3.6,

where a much more detailed analysis is provided.

585
 Notes

14. The hyphens on both sides of the consonant in the notation

-ng- imply that other parts of the word must follow and
precede.

15. The possessed forms of katuu ‘cat’ have a special meaning—

katungel ‘his/her sweetheart’, etc.

16. This word also appears in an expression like bȩrdel a kim,

where it refers to the inside surface of a clam shell.

17. Note that this expression contrasts in meaning with uchul a

chȩldȩ chȩduch ‘reason for the meeting’.

18. Many other kinds of Palauan relational phrases exist. These

will be discussed in detail in chap. 14.

19. This noun is not as fully assimilated as the others, since

we can also hear babier ȩr ngak ‘my letter’, babier ȩr ngii
‘his letter’, etc. Note also the rather unusual alternation
between r of the independent form and l of the possessed
forms.

20. If we were to translate this and the above expression lit-

erally (i.e., word-for-word), we would have ‘matters/things
of the house’ for tȩkoi ȩr a blai, and ‘matters/things of the
road’ for tȩkoi ȩr a rael.

21. This and the above expression are used primarily in a

religious context and mean, literally, ‘country of fire’ and
‘country of the sky’. In normal conversation, the word
babȩluadȩs is used for ‘heaven’.

22. This possessed form is very irregular, since the -lu- of the
independent form ilumȩl ‘drink’ is lost completely.

23. A word-for-word translation of this expression would be

something like ‘car for playing’.

24. This expression means, literally, ‘food of war’.

586
 Notes

*25. Expressions like imȩlek ȩl biang consist of a head noun

(imȩlek) followed by a relative clause (ȩl biang) whose
source is an equational sentence (A imȩlek a biang. ‘My
drink is (a) beer.’) The literal translation of imȩlek ȩl biang
is therefore ‘my drink which is (a) beer’. See 23.4 for justifi-
cation and explanation of this analysis.

26. Notice the unusual phonetic changes in the possessed

forms of ongraol: r is deleted and the vowel cluster ao
changes to u.

27. This expression seems to be used only on formal occasions;

it is more common to say kȩlek ȩl tuu.

28. Notice the loss of -chȩ- in the possessed form of chȩmachȩl.

4 PALAUAN PRONOUNS

1. These two terms are taken from Mancill and Woods

1969:46.

2. Recall that mȩ and a following a are pronounced [ma].

3. Many speakers pronounce the coordinate noun phrase in

this sentence with [mǝ] between the two nouns. In other
words, it appears as if the question noun tȩcha ‘who?’ is op-
tionally preceded by a in this construction. A similar phe-
nomenon is found in noun phrases of possession containing
tȩcha, as illustrated in the following sentence:

Tia a mlil (a) tȩchang? ‘Whose car is this?’

4. When the possessed form of reng ‘heart’ is used with the

verb suebȩk ‘fly’, we obtain the special meaning ‘(someone)
is worried’. For a discussion of other special expressions
with reng, see 17.4.

587
 Notes

5. In this sentence, ng could refer to someone other than

Toki if that person’s identity were clear from the context.
Therefore, 37b might also mean ‘Toki said that he is going.’

6. Because there is a Ø-suffix in cholébȩd [ɂolέbǝð], the stress

remains on the second syllable; therefore, the vowel e [ε] of
this syllable does not reduce to ȩ [ǝ].

7. Olȩkiis is a causative verb related to the intransitive

action verb mȩkiis ‘get up, wake up (naturally); stand up’.
See 5.1.1 and 9.2.2 for further details.

8. Notice that the third person singular object pronoun ap-

pears as -ir in this perfective form. See 4.9.4 below for more
details.

*9. Olȩkar is a causative verb related to the intransitive

action or state verb mȩkar ‘wake up, be awake’. See 7.3
and 9.2.2 for further details.

10. The inserted ȩ under discussion here is the only sound

which differentiates certain perfective forms of omes ‘see’
from those of omes ‘give’. The perfective forms of the latter
verb are mȩskak ‘gives (it to) me’, mȩskau ‘gives (it to) you’,
msa ‘gives (it to) him/her’, mȩstȩrir ‘gives (it to) them’, etc.

11. Another verb in which l appears in the perfective forms is

mȩnga ‘eat’—kolii ‘eats it’.

12. This added syllable seems to be a kind of reduplication

(see 11.11).

*13. Strictly speaking, -lim is not a verb stem but rather a se-
quence consisting of the -l- allomorph of the imperfective
marker (see 5.5) and a reduced form of the verb stem ŋim
‘drink’.

*14. For a possible explanation of this change from o to u, see

chap.5, note 4.

588
 Notes

15. For certain speakers, replacing mo- by chomo- in a sentence

like this results in a kind of contrastive emphasis (cf. 4.2
above). Thus, we also have the following:
A babier ȩl chomoluchȩs ȩr ngii a mo ȩr a Droteo.
‘The letter yóu’re writing goes to Droteo.’

16. For a technical explanation of this alternation between m

and b, see 6.2.1. Note, further, that Palauan speakers also
change m to b in the hypothetical forms of me ‘come’ when
certain types of grammatical constructions are involved. In
the sentences below, for example, we have relative clauses
(see chap. 23):

a. A kall ȩl kbe kuruul ng ‘For whom is the food

kȩlel a tȩchang? which I’m coming to
prepare?’
b. A blai ȩl lȩ be ‘The house which Toki is
longȩt-mokl ȩr ngii a coming to clean is
Toki a blil a Droteo. Droteo’s house.’

In the sentences above, the verbs following kbe and lȩbe are
also hypothetical verb forms. Further examples of this
kind, in which two adjacent verbs each have prefixed hy-
pothetical pronouns, will be considered in 4.10.6 below.

17. Notice that the third person hypothetical pronoun lȩ-

further reduces to l- before the vowel-initial (state) verb
ungil ‘good’

18. Reduction of lȩ- to l- is also observed in this word.

5 PALAUAN VERBS

1. Although use of mle ‘was, were’ as an auxiliary word with

state verbs probably developed from mle ‘came’, the past
tense form of the intransitive action verb me ‘come’, it will
be preferable to regard these items as separate words.

589
 Notes

*2. The surface form ul(ȩ)- is probably derived from underlying

o-il- (i.e., the verb marker prefix o- followed by the past
tense marker -il-). The mid back vowel o and the high front
vowel l assimilate to each other to yield the single high back
vowel u.

3. Though difficult to prove, it is possible that the causative

prefixes listed here actually consist of the verb marker
prefix o- followed by some other (causative) morpheme. See
9.2.1.1 and 9.2.2.1 for further discussion.

*4. The surface forms of hypothetical pronouns like lu-, du-,

etc. are probably derived from underlying sequences of the
form lo-il-, do-il-, etc., which consist of the basic forms of
the hypothetical pronoun prefixes followed by the past tense
marker -il-. The u in lu-, du-, etc. is derived by “mutual” as-
similation of the mid back vowel o and the high front vowel
i, resulting in the single high back vowel u. Recall that the
same process may account for the derivation of the surface
form ul(ȩ)- from underlying o-il- (cf. note 2 above).

5. Though possibly related, the auxiliary word mla and the ex-
istential state verb mla ‘was/were (located)’ (see chap. 18)
are best considered as separate words.

6. The use of ng diak ‘no’ in B’s response is quite interesting.

Notice that, as in the case of 51–53, ng diak is normally
the negative response to yes-no questions (see 20.1). In
54, however, B’s use of ng diak serves to tell A that the as-
sumption of A’s question—namely, that B has yet to paint his
canoe—is not correct.

7. Much of the analysis presented here and in subsequent

paragraphs is based on Wilson 1972: 120–128.

590
 Notes

6 THE VERB MARKER AND PERFECTIVE VERB

FORMS

1. Note that the verb stem chur ‘laughter’ must be repeated

(or re duplicated—see chap. 11) in order to form the intran-
sitive verb ochȩrchur ‘laugh’.

2. The basis for this observation is found in Wilson 1972:107.

3. The hypothetical form of me ‘come’ also appears as lȩme.

Cf. chap. 4, note 16.

4. Notice that the auxiliary word mla (cf. 5.3.2.1), which is

probably related to mla ‘was (located)’, also has hypothetical
forms with b—e.g. kbla, lȩbla, etc.

5. Rather exceptionally, some speakers use smésȩb instead of

sués ȩb—in other words, these speakers fail to apply the rule
which normally changes the metathesized verb marker to -u-
if there is a bilabial consonant in the stem.

6. Even though the Ø object pronoun suffix has no phonetic

content (i.e., does not correspond to any actual sounds), we
include it in 18 and similar derivations in order to show
that the basic structure of perfective forms like duálȩm is
identical to that of all other perfective forms—namely, verb
marker + verb stem + object pronoun. In the final step of 18,
which results in the actually-pronounced form, the Ø object
pronoun is deleted.

*7. The ȩ in question is always found in an unstressed syllable

and therefore appears to be the reduced form of one of
the full vowels (cf. 1.4.4). Which full vowel is actually the
source of the ȩ cannot be determined, however. See Wilson
1972:24–29 for further discussion.

8. For further discussion of this type of phonetic rule, see 6.4

below.

591
 Notes

7 STATE VERBS

1. The state verb mȩsaul ‘tired’ is found in the commonly-

used expression kȩ kmal mȩsaul. This sequence, which lit-
erally means ‘You are very tired’, is used as an equivalent of
English ‘Thank you’.

2. The same is true for kȩdeb and kekȩdeb, both of which

mean ‘short’.

3. Some speakers feel that 12a and 12b differ in meaning

as follows. In 12a, mle dȩngchokl denotes a completed
past state: in other words, it is implied that the subject
(Hermana) was sitting here at some time in the past but is
no longer sitting here at the present moment. By contrast,
dilȩngchokl designates a past action whose effect is still con-
tinuing: that is, the subject sat down here and is still sitting
here. For some speakers, the following two sentences differ
in a parallel way:

a. Ak mle smechȩr. ‘I was sick (but no longer am).’

b. Ak silechȩr. ‘I’ve been sick (and still am).’

4. For further information on the grammatical properties of

certain Palauan transitive state verbs, see 16.3 (mȩduch and
mȩtitur followed by object clauses) and 21.2 (mȩdȩnge fol-
lowed by various structures).

5. When associated with the word reng ‘heart, spirit’, the state
verb bȩralm describes someone’s personality, as in the sen-
tence
Ng bȩralm a rȩngul a sȩchȩlim.
‘Your friend is lazy/unmotivated.’
See 17.4 for more details.

592
 Notes

6. When applied to a person, the state verb bȩchachau can

mean ‘stupid’, as in the sentence
Ng bȩchachau a bdȩlul a John. ‘John is stupid/empty-
headed.’

7. The related intransitive action verb is rȩmurt ‘run’, in

which the verb marker has metathesized and appears as the
infix -ȩm- (cf. 6.2).

8. The related transitive verb, in its imperfective form, is

mȩlungȩl ‘smell’.

9. The second ȩ of bȩkȩ- is deleted before the initial vowel u of

the stem ureor ‘work’.

10. In this word, sȩkȩ- has unexpectedly shortened to sȩ-.

11. If we consider the basic form of this word to be sȩkȩ + kȩr +

ker, where the stem ker ‘question’ has been totally repeated,
then the actually-pronounced form sȩkȩrker is derived by
dropping one of the two identical sequences -kȩ-.

*12. Wilson 1972: 89–90 proposes that in cases such as lȩngiil

‘is to be borrowed’ and chȩtuul ‘is to be smoked’ the basic
forms of the verb stems are vowel-final—i.e. lengi and
chatu. In forms which do not contain a suffix, such as the im-
perfective verbs mȩleng ‘borrow’ and mȩngat ‘smoke (fish)’,
the stem-final vowel is deleted. If there is a suffix, however,
as in the anticipating state forms, the stem-final vowel re-
mains and becomes long.

*13. This analysis is based on Dyen 1971:248, where the concept

of “morphophonemic formula” is introduced.

593
 Notes

8 NOUN DERIVATION

1. The two uses of sȩlokȩl are illustrated in the sentences

below. In sentence a, sȩlokȩl is an anticipating state verb
meaning ‘is/needs to be washed’, while in b, a possessed
form of sȩlokȩl is used as a noun meaning ‘your laundry’.

a. A bilek a kirel ȩl ‘My clothes need to be

sȩlokȩl. washed.’
b. Ng bȩtok a sȩlȩkȩlem? ‘Do you have a lot of
laundry?’

2. In the possessed forms of kall ‘food’—i.e., kȩlek, kȩlel,

kȩlem, etc.—the sequence ll is shortened to a single l.

3. In the possessed forms of ilumȩl ‘drink’—i.e., imȩlek, imȩlel,

imȩlem, etc.—the whole syllable -lu- has been dropped. Such
deletion of a complete syllable is very rare.

4. As perfective forms like songosii ‘cook it’ (cf. 6.3.j) show,

the stem of this verb has an initial s—i.e., sȩngoes. The k in
klȩngoes ‘meat or fish stew’ is due to a rather unusual pho-
netic rule which changes s to k before l (cf. 7.8, ex.32).

5. Recall (cf. 7.7) that the second r in rruul is due to assim-

ilation of the resulting state infix l to the preceding verb-
stem-initial r.

6. In the possessed forms of rruul ‘thing which is made/

done’—i.e., rrȩllek, rrȩllel, rrȩllem, etc.—the long uu has
rather exceptionally reduced to a ȩ (cf. the examples at the
end of 3.4.2). In addition, the l has doubled (cf. 6.5.e).

7. Notice that the verb-stem-final s observed in mȩluchȩs

‘write, draw’ and mȩngiis ‘dig’ has changed to k before the
l of the anticipating state suffix. Cf. 7.8, ex.32 and note 4
above.

594
 Notes

8. Another instance of the rule changing s to k before l is ob-

served in this word. Cf. notes 4 and 7 above.

9. The rr in this word is due to assimilation. Cf. note 5 above.

10. Recall that mȩchiuaiu can be either a state verb or an action

verb (cf. 7.3).

11. This sentence means, literally, ‘The thinness (from poverty)

in China is very great.’

12. Some of the derived nouns listed here and in 6 and 7 above
are due to Mancill and Woods 1969:33–4.

13. In klungel and the other possessed forms of kllou ‘size,

thickness’, one of the l’s is lost, the vowel cluster ou reduces
to u, and ng is inserted before the possessor suffix (cf. 3.4.4).

14. While klȩmangȩt ‘length, height’ can refer to human beings

(as in 10b) or to things (as in 10c), kldidai ‘height’ is re-
stricted to things.

15. Notice the unusual insertion of the vowel cluster -ao- into
this derived noun.

16. Notice that the o of the state verb changes to u in the de-
rived noun.

17. While klȩmȩdȩnge refers to someone’s knowledge of things

(e.g., language), klaodȩnge of 12d has to do with two (or
more) persons’ knowledge of each other.

18. It is not very clear whether klobak and rubak are actually
related words.

19. A rare case in which o- substitutes for the verb marker in

an intransitive verb to derive an instrument noun is ob-
served in osebȩk ‘wing’—suebȩk ‘fly’. In this example, the
metathesized verb marker -u- of suebȩk (cf. 6.2) is missing
in the derived instrument noun because it has been replaced
by the prefix o-.

595
 Notes

9 CAUSATIVE VERBS

1. Many speakers also use this form to express the related

meaning ‘build’. Some speakers, however, distinguish be-
tween omȩkdȩchor ‘make…stand’ and omȩkȩdȩchor ‘build’.
The source and function of the additional ȩ in the latter form
is a mystery.

2. The full vowel o of osiu ‘joining’ reduces to ȩ in the

causative verb omȩkȩsiu ‘compare, imitate’.

3. The initial m of mad ‘dead’ changes to o when the prefix

omȩk- is added. This change of m to o may be due to a pho-
netic rule of dissimi lation similar to that observed in 6.1
and 6.2. In other words, the m of mad dissimilates from the
m in the causative prefix omȩk- and becomes o.

4. In order to avoid possible confusion in the discussion to

follow, we should point out here that the sequence -mȩ-
found in omȩ(k)- is not the same as the mȩ - which we
have dealt with extensively as the major variant of the verb
marker (cf. 6.1). The -mȩ- of omȩ(k)- does not represent a
single unit or morpheme which has structural significance;
as we will see in 9.2.1.1, it actually consists of one mor-
pheme (the imperfective marker -m-) and part of another
(the -ȩk of the causative marker bȩk).

5. Some speakers delete the ȩ of the prefix between the con-

sonants m and ng. Thus, we sometimes have omngim
‘make…drink’ and omngamȩch ‘make…chew,
make…smoke’.

6. For a discussion of the l in this verb stem, cf. 6.5.b.

7. Mukdakt is one of the ergative forms of causative

omȩkdakt ‘frighten’. See 9.5 below.

596
 Notes

8. Though translated identically, oltuu and olsisȩb are

somewhat different in meaning; this meaning difference is
parallel to that found between tmuu and soisȩb, the action
verbs to which they are related. While soisȩb implies dif-
ficulty in entering (perhaps because the entrance is too
narrow, etc.), tmuu is more neutral in connotation.
Therefore, olsisȩb means ‘make (someone) enter with diffi-
culty, force/push into’, while oltuu lacks the connotation of
force and simply means ‘put into, make…enter’.

9. Most of the observations in this paragraph are due to Jo Ann

Flora (personal communication). See also Wilson 1972:153.

10. This is essentially the position taken in Wilson

1972:150–155.

11. Suebȩk a rȩngul, which means, literally, ‘his heart flies’, is

one of many expressions consisting of an intransitive (action
or state) verb and a possessed form of reng ‘heart, spirit’.
Expressions of this kind are used commonly in Palauan to
denote feelings, emotions, or character traits. See 17.4 for
more details.

12. In the perfective forms mȩkȩlii and milȩkȩlii, we note the

presence of the final l of the verb stem kal ‘eat’. For further
discussion, cf. 6.5.b.

13. The appearance of -l- in the perfective forms mȩngȩlmii and

milȩ ngȩlmii indicates that the verb stem for ‘drink’ contains
an l. This l also turns up in the related noun ilumȩl ‘drink,
beverage’.

*14. The sequence ul(ȩ)- is probably derived from o- (the verb

marker prefix) followed by -il- (the past tense marker). Cf.
chap.5, note 2.

597
 Notes

10 RECIPROCAL VERBS

*1. Our explanation of “reciprocity” here is quite oversim-

plified. For a discussion of the logical properties of Palauan
reciprocal sentences, see Wilson 1972:180–200.

2. The independently-occurring noun ngȩseu ‘help, assistance’

is used in sentences like the following:

a. Ak ulȩngit a ngȩseu ȩr a ‘I need a helping hand.’

chim.
b. Ak ulȩngit a ngȩseu ȩr a ‘I need some financial
udoud. help.’

The bound verb stems toir and dingȩl can be prefixed with o-
to derive the nouns otoir ‘(action of) chasing’ and odingȩl
‘visit’. The exact identity of this o- prefix is not clear,
though it is probably related to the o- which is used to
derive instrument and action nouns (cf. 8.6). The de-
rived nouns otoir and odingȩl are shown in the following
sentences, where they are used in a possessed form (cf.
chap.3):

c. A otirel a babii a mȩringȩl ‘Catching a pig is a

ȩl tȩkoi. difficult task.’
d. A odngȩlel a Surech ȩl me ‘Surech’s visits to Palau
er a Belau a bek ȩl buil. take place every month.’

3. This sentence means, literally, ‘The cars met each other.’

4. This is the past tense form of oumȩkȩmad ‘make war on’.

Recall (cf. 6.1.1) that the past tense forms of verbs in ou- are
derived by replacing the o of ou- with ul-.

5. The (bound) verb stem sbech needs to be partially repeated

(note the extra syllable -si-) before the reciprocal prefix kau-
can be added. For further discussion of this and other types
of reduplication, see chap. 11.

598
 Notes

6. The form of this noun indicates that it is derived by simulta-

neously adding the resulting state infix -(ȩ)l- and the an-
ticipating state suffix -(ȩ)l to a verb stem (cf. 8.2, ex. 5).
Notice that the -ll- of ngȩllakȩl ‘joke’ is shortened to a single
l in the derived verbs oungȩlakȩl ‘joke with’ and kaungȩlakȩl
‘joke with each other’.

7. It is also possible to have the reciprocal verb kausȩsȩchȩlei,

in which the extra syllable -sȩ- represents a reduplicated
portion of the stem sȩchȩlei ‘friend’ (see chap. 11 and cf.
note 5 above). The word kausȩ sȩchȩlei implies a weaker,
more casual friendship than kausȩchȩlei, as the English
equivalent for the following indicates:

Ngak mȩ a Droteo a ‘Droteo and I are sort of

kausȩsȩchȩlei. friends.’

8. The reciprocal verb chachȩlebȩd ‘hit each other’ is also pos-

sible. See 10.2.4. below.

9. The -o- in this verb appears to be a reduced variant of the

verb marker mȩ- of mȩdȩnge ‘know’ (cf. 6.1).

10. This verb is also discussed in 10.2.4 below.

11. The additional k following the reciprocal prefix in kakn-

godȩch is unpredictable. Some speakers also pronounce this
word with a ȩ between the k and the ng—i.e., kakȩngodȩch.

12. Here, the possessor of hong ‘book’ is expressed with a pos-

sessor phrase consisting of the relational word ȩr and
the pronoun tir ‘they’. A possessor phrase of this kind is re-
quired because hong ‘book’ is an unpossessible noun (cf.
3.8).

13. This reciprocal verb is related to the causative verb or-

rengȩs ‘hear’. For some speakers there is a contrast be-
tween kȩrengȩs ‘listen to/take advice from each other’ and
kaiuȩrengȩs ‘hear each other’.

599
 Notes

14. Note the unusual appearance of the extra syllable -dȩ- in

this reciprocal verb.

11 REDUPLICATION AND FURTHER VERB

AFFIXATION

1. In other words, for ‘tall, long’ we have either kekȩmangȩt or

kȩman gȩt, and for ‘short’ we have either kekȩdeb or kȩdeb.
Addition of the reduplicated initial syllable does not change
the meaning in any way.

2. We used the term “grammatical device” in a similar way

in 8.3 with reference to the function of the resulting state
infix -(ȩ)l- in deriving nouns from intransitive verbs.

3. In the sentence below, the reduplicated form of beot ‘easy’

occurs with a possessed form of reng ‘heart, spirit’ (see
17.4) to yield an expression meaning ‘be undecided (about
something), not take (something) seriously’:

Ng ko ȩr a bebeot a rȩngul a ‘Droteo is rather

undecided
Droteo ȩr a omȩrael ȩl mo ȩr a about travelling to
Hawaii. Hawaii.’

4. The special expression mȩnga ȩr a chull ‘(lit-) eat the rain’

corresponds to English ‘get caught in the rain’. Its use is il-
lustrated in the following additional sentences:

a. Lak monga ȩr a chull e ‘Don’t get caught in the rain

kȩ mo smechȩr. or you’ll get sick.’
b. Ak killii a chull mȩ ak ‘I got caught in the rain, so
mlo smechȩr. I got sick.’

600
 Notes

5. Some speakers also use reduplicated forms of the possessed

nouns kirek/kirel ‘my/his obligation’, etc. and sȩbȩchek/
sȩbȩchel ‘my/his ability’, etc. These reduplicated forms are
illustrated in sentences like the following:

a. Ng kekirek ȩl mong. ‘I sort of have to go.’

b. Ng sesȩbȩchek ȩl mȩlȩkoi a ‘I can sort of speak
tȩkoi ȩr a Siabal. Japanese.’

6. For a discussion of mochu ‘about to go’, see 11.12.5 below.

7. The related imperfective transitive forms for these verbs

are mȩlatȩch ‘clean’, mȩlemd ‘mop’, and mȩlamȩl ‘cut
(grass)’.

8. Some of these verbs have additional reduplicated forms

lacking the first syllable of C 1 eC 1 V(C 2). For example,
the forms mȩsisaik, mȩrȩngȩringȩl, and mȩkȩrkar are all ac-
ceptable, though mȩkȩrkar has a very different meaning, as
we shall see in 11.6 below.

9. All of these verb stems can occur independently as nouns.

10. The o in the C 1 V(C 2) syllable of mȩrerorael also occurs

in the possessed forms of rael ‘road’—e.g. rolek ‘my road’,
rolel ‘his road’, etc. Note also omȩrael ‘trip’—omȩrolek ‘my
trip’, etc.

11. The reduplicated form mȩsesilil is unpredictable: possibly

the s was part of the verb stem at some earlier stage of the
language. Compare mȩiusȩch ‘calm (sea)’ and its redupli-
cated form mȩsesiusȩch ‘rather calm’.

12. Evidence from this form and from the reciprocal verb re-
lated to rȩborb—namely, kadȩrȩborb ‘sit with each other’
(cf. chap. 10, note 14) seems to indicate that the verb stem
in question is dȩrȩborb and that, for unknown reasons, the
initial syllable dȩ- drops in the simple intransitive form.

601
 Notes

13. Some speakers omit C 2 when producing this form, giving

mȩlelȩlu chȩs.

14. Some speakers reduplicate this ergative verb form ac-

cording to the C 1 eC1 V(C 2) pattern, giving obebȩlȩbalȩch.

15. Notice the appearance of l as C 2 in the reduplicated syl-

lable. This is further evidence that the basic stem for ‘eat’ is
indeed kal (cf. 6.5.b).

16. The pattern of reduplication is slightly different here, since

the a of mȩlatȩch ‘clean’ has been deleted in mȩleltȩlatȩch
‘clean compulsively’. The a in question is deleted in other
forms of this verb as well—note, for example, the derived
state verb ngȩltachȩl ‘cleaned’ (cf. 7.8.2).

17. In ombibtar ‘keep swinging’, the reduplicated syllable has

the form C 1 i instead of C 1 e. This represents a much less
common pattern.

18. The technical term inchoative is ultimately derived from a

Latin verb meaning ‘begin’.

19. The -ng- inserted before the predictive and inchoative suf-
fixes also turns up in tȩkingek ‘my words’, tȩkingel ‘his
words’, etc., which are the possessed forms of the related
verb stem tȩkoi ‘word, language’ (cf. 3.4.4).

20. For further discussion of this construction, see 15.7.7.

21. The extra -ng- in mȩkȩrȩnga ‘waking up’ is unpredictable.

22. The -ng- occurring before the inchoative suffix in klunga

‘getting big’ also appears in klungek ‘my size’/klungel ‘his
size’, etc., which are the possessed forms of kllou ‘size’, a
noun derived from klou ‘big’ (cf. chap.8, note 13).

23. Note the unusual loss of the l of ungil ‘good’ before the in-
choative suffix.

602
 Notes

24. A small number of action verbs also take -o or -e for the

inchoative instead of -a. Among them are (mo) mȩrek
‘finish’—mȩrko ‘is just finishing’, omȩngur ‘eat (a
meal)’—omȩngro ‘starting to eat’, and omes ‘see’—mȩsȩnge
‘has just seen’.

*25. It is possible that the -ch- is inserted between mo and the

following suffixes -u and -a to prevent vowel blending or
vowel cluster reduction, which would obscure the identity of
the suffix. This possibility was suggested to me by Donald
Topping.

12 IMPERFECTIVE VS. PERFECTIVE VERBS

1. Focusing on this aspect of the meaning of imperfective

verbs, Wilson 1972:120–128 uses the term progressive in-
stead of imperfective. Similarly, she uses the term pro-
gressive affix for what we have been calling the imper-
fective marker.

2. Any possible English equivalent for 21b would likewise be

contradictory and unacceptable—e.g., *‘I’ve read the book
completely, but I haven’t finished it yet.’

3. This time span expression is a special type of temporal

phrase. See 14.6 for further details.

4. For a more complete list, see 4.9.4, ex. 53.

13 DIRECTIONAL VERBS

*1. Notice the English equivalents for the sentences in 1 and

2. In 1B, the speaker uses come because he expects A (the
person addressed) to be awaiting his arrival: B seems to be
looking at the situation from A’s viewpoint. In dialog 2, the
use of either come or go appears possible to describe a sit-
uation in which A does not intend to be at home upon B’s

603
 Notes

arrival, but note the following difference surrounding the

circumstances of the dialog itself. If A is now at home (say,
talking to B on the telephone) but does not intend to be at
home tomorrow, he can say the following to B:
Can you come to my house tomorrow (to feed the dogs)?
But if A is not at home now (say, talking to B on the street)
and furthermore does not intend to be at home tomorrow, he
can say the following to B:
Can you go to my house tomorrow (to feed the dogs)?

2. For clarity of presentation, we regard this mle (the past

tense form of me ‘come’) as a different word from
homonymous mle ‘was, were’, which is used as an auxiliary
word to indicate the past tense with state verbs. Ultimately,
however, these two words are probably one and the same el-
ement. Cf. chap. 5, note 1.

3. The basic form of me is actually mȩ (verb marker) +

be (verb stem), and that of mo is mȩ (verb marker) + bo
(verb stem). The phonetic rules which apply to these basic
forms to give the actually-pronounced forms are explained
in 6.2.1, ex.12.

4. This mle is the auxiliary word mle used with state verbs.
Cf. 5.1.3 and note 2 above.

5. This directional phrase, which literally means ‘to the sea’,

is used together with the directional verb mo to mean ‘go
fishing’.

6. Note that the change of state expression in 13d consists of

mlo and the reciprocal verb kaodȩnge ‘know each other’.
The fact that a verb like kaodȩnge can appear following mo
in a change of state expression is further evidence that rec-
iprocal verbs are really a subtype of state verbs (for addi-
tional discussion of this point, cf. 10.1, ex. 4). In 13g, the

604
 Notes

change of state expression consists of mla mo and the neg-

ative verb diak ‘isn’t, doesn’t exist’ (see 18.3); here, too, we
have evidence that diak is actually a state verb.

7. For further discussion of the perfective forms of mȩrkui

‘finish’ and of how they are used in specifying clauses, see
15.7.4.

14 RELATIONAL PHRASES

1. When mla ‘existed, was (located)’ is followed by a locational

phrase containing a specific place name, the result is an
expression indicating what country or place the subject is
from, or what place the subject has come or gone from. Note
the following:

a. Ak mla ȩr a Ngchesar. ‘I’m from Ngchesar.’

b. Ak mla ȩr a Merikel ȩl mei. ‘I’ve come from
America.’
c. A Toki a mla ȩr a New York ‘Toki went from New
ȩl mo ȩr a Paris. York to Paris.’

In b and c the sequences ȩl mei and ȩ l mo er a Paris are

types of specifying clauses (see 15.7).

2. The question word ker ‘where?’ is never preceded (or intro-

duced) by a (cf. 2.6).

3. The nouns bab and eou can also be used to indicate rank,
as in the following example:

A Droteo a ngar ȩr{bab/eou} ȩl chad.

‘Droteo is {higher/lower} in rank.’

They also occur in the expressions bab ȩl daob ‘(lit.) upper

sea’ and eou ȩl daob ‘(lit.) lower sea’. The former is the
name of Babeldaob (often spelled Babelthuap on maps),

605
 Notes

the large island in the northern part of the Palau group,

while the latter refers to the islands south of Babeldaob,
including Koror.

4. Observe the following sentence, in which the best English

equivalent for bebul is ‘surface’:

Ng ngar ȩr ngii a bȩtok ȩl ‘There’s a lot of trash on

komi ȩr a bebul a daob. the surface of the ocean.’

5. Mȩdal can also mean ‘his eye’ or ‘his face’. In addition, it

can refer to the point or edge of something, as in mȩdal a
oluchȩs ‘pencil point’ and mȩdal a oles ‘cutting edge of a
knife (including the point)’.

6. Unlike bab(bebuk, bebul, etc.), mad (mȩdak, mȩdal, etc.),

and other nouns discussed above, ullel must always occur
with a possessor suffix (ullek, ullel, etc.). In other words,
ullel is an obligatorily pos sessed noun (cf. 3.5)

7. A more literal English translation for this example might be

something like

‘Droteo is bathing in the {upper/lower} part of the

river.’

8. In some cases, it is very difficult to tell whether uche is

being used to refer exclusively to space or to time. In the fol-
lowing sentence, for example,

A bilsȩngel a Droteo a mlo ȩr a uche ȩr a klaidȩsachȩl.

‘Droteo’s boat {came in first in won} the race.’

the expression mlo ȩr a uche ‘won’ (lit. ‘went to the front of’)
(pronounced as if spelled mloruche) seems to blend the
concepts of being first in space and being first in time.

606
 Notes

9. Chȩlsel ‘inside’ is used in cases like this because taoch

‘channel (in a mangrove swamp)’ and omoachȩl ‘river’ are
viewed as enclosed spaces into which the boat has entered
by means of some opening or passageway.

10. The noun iikr ‘area/space outside’ is used in a locational

phrase in the sentence below:

A rȩngalȩk a milil ȩr a iikr.

‘The children are playing outside.’

11. The unpossessible noun bita expresses any relationship of

proxi mity (i.e., nearness) between two objects—hence, the
two possible English equivalents. This word also refers to
one half of a coconut (after being split).

12. Dȩlongȩlel is a possessed form of dȩleongȩl ‘area/space be-

tween’, which can also refer to relationships between two in-
dividuals or groups of individuals. This meaning is observed
in the sentences below:

a. A dȩlongȩlel a Droteo mȩ a Toki a kmal ungil.

‘Droteo and Toki are on very good terms.’

b. A dȩlongȩlir a rȩsensei mȩ a rȩngalȩk ȩr a skuul a
mȩkngit.
‘The teachers and students don’t get along well with
each other.’
c. A dȩlongȩlel a John mȩ a Mary a ko ȩr a klsakl.
‘Something is wrong between John and Mary.’
d. A Korea a mlo dokurits ȩr a uriul ȩr a mȩkȩmad ȩr a
dȩlongȩlel a Merikel mȩ a Siabal.
‘Korea became independent after the war between
America and Japan.’

607
 Notes

13. The other points of the compass are dimȩs ‘south’, chongos
‘east’, and ngȩbard ‘west’. The four nouns designating
points of the compass are all unpossessible.

14. The word bȩlngel ‘middle’ is an obligatorily possessed noun.

15. The obligatorily possessed noun tkul refers to the (outside)

edge or corner of anything, as in tkul a blai ‘corner of the
house’, tkul a tebȩl ‘edge of the table’, tkul a chȩldukl ‘edge
of the dock’, tkul a mȩkȩsokȩs ‘corner of the yard’, etc.

16. The obligatorily possessed noun mobȩdul ‘general location/

direction’ is related to bȩdul ‘direction, area/space facing
onto’. In the sentences below, expressions like bȩdul a skuul,
etc. are used as specifying clauses (see 15.7) following
other verbs:

a. A Droteo a millutk ȩl bȩdul a kbokb.

‘Droteo pointed at the wall.’
b. A Toki a rirurt ȩl bȩdul a skuul.
‘Toki ran in the direction of the school.’
c. A skuul a dȩchor ȩl bȩdul a M-dock.
‘The school faces in the direction of M-dock.’
d. A Toki a mle dȩngchokl ȩl bȩdul a chȩlȩchol.
‘Toki was sitting facing the beach.’
Bȩdul can also mean ‘upstream’, as in
e. A Droteo a mȩlȩcholb ȩr a bȩdul.
‘Droteo is bathing upstream.’
With e compare example 16 above.

17. The grammatical structure of sequences such as mȩrael ȩl

mo… ‘walk to’ and rȩmurt ȩ l mo… ‘run to’ will be clarified
in 15.7.1.

18. Most speakers feel there is a subtle difference in meaning

between tiluu (cf. present tense tmuu) and silisȩb (cf.
present tense soisȩb), both of which we have translated as

608
 Notes

‘went into’. While silisȩb implies that there was some diffi-
culty in entering (perhaps the hole was too narrow for the
mouse), tiluu has a more neutral connotation. Cf. chap. 9,
note 8.

19. The perfective forms of these verbs, which are rather ir-
regular, are listed below according to the usual format:

a. mȩlȩcha ‘put’: b. omȩche ‘leave’:

loia lilia mȩchire milȩchire
locha lilȩcha mȩche milȩche

The verb omȩche ‘leave’ is used in the very common ex-

pression Bȩchire mȩ a uriul ‘Leave it till later.’ The im-
perative perfective forms of omȩche ‘leave’ (see 19.5),
when suffixed with the appropriate object pronoun,
can be followed by the word mȩ‘(lit.) so that’ and a
hypothetical verb form to result in expressions cor-
responding to English ‘let (someone) (do something)’.
Observe the examples below, which will be explained
further in 22.1.1:

a. Bȩchire a ngalȩk mȩ lȩbo ‘Let the child go play.’

loilil.
b. Bȩchikak mȩ kbo kusuub. ‘Let me study.’
c. Bȩchititȩrir mȩ lȩbo ‘Let them go
longȩdub. swimming.’

20. Expressions of this kind will be analyzed in 15.7.1.

21. The words ultungii and ulsisȩbii are the 3rd pers. sg.
object past perfective forms of the causative verbs oltuu
‘put/push in’ and olsisȩb ‘put/push in’ (cf. 9.4, ex.24). The
meaning difference between these two causative verbs par-
allels that found between tmuu and soisȩb, the intransitive
action verbs to which they are related (cf. chap. 9, note 8,
and note 18 above).

609
 Notes

22. In this sentence, ȩr a kȩdȩrang ‘from the beach’ is inter-

preted as a source phrase out of contrast with (ȩl mo) ȩ r
a stoang ‘(going) to the store’, which expresses the termi-
nation point of the movement. However, if ȩl mo ȩ r a stoang
is omitted, the resulting sentence
A John a rirurt ȩr a kȩdȩrang. ‘John ran at the beach.’
only makes sense if ȩr a kȩdȩrang is interpreted as a lo-
cational phrase meaning ‘at the beach’ (cf. 14.2 above).
Further analysis of the depend ent clause ȩ l mo ȩ r a
stoang ‘going to the store’ will be given in 15.7.1.

23. The word ultȩbȩdii ‘pulled him out’ is the 3rd pers. sg.
object past perfective form of the causative verb oltobȩd
‘pull out’, which is related to the intransitive action verb
tuobȩd ‘come out’ (cf. 9.4, ex. 24).

24. While cheisȩch ‘stained’ of 29e refers to a permanent un-

washable stain due to such substances as betel nut juice,
etc., kikiongȩl ‘dirty, soiled’ refers to a temporary state of
dirtiness due to such washable substances as mud, blood,
soot, etc.

25. Idȩlsȩbel is the possessed form of idȩlisȩb ‘the day before

yesterday’.

26. This expression, containing uche ‘before’, means, literally,

‘before/ previous to three days ago’.

27. Since these expressions contain tia ‘this’ and mlo mȩrek
‘finished’, they mean, literally, ‘this finished week’, etc. Note
that the word for week—sande—is borrowed from English
Sunday.

28. Since these expressions contain se ‘that’ and mlo mȩrek

‘finished’, they mean, literally, ‘that finished week’, etc.

29. See 22.2 for a discussion of temporal clauses introduced

by ȩr se ȩ r a ‘when’, as in the following:

610
 Notes

Ak milsa a John ȩr se ȩr a kngar ȩr a Hawaii.

‘I saw John when I was in Hawaii.’

30. For this and the previous expression, some speakers re-
verse the order of the parts, giving ȩr a klukuk ȩr a tutau ‘to-
morrow morning’ and ȩr a klukuk ȩr a kȩbȩsȩnge ‘tomorrow
evening’. This reversed order, however, is less preferred.

31. Ngiosel is the possessed form of ngiaos ‘the day after to-
morrow’. Note the similar pattern for past time—cf. note 25
above.

32. This expression, containing ikrel (the possessed form of iikr

‘area/ space outside’—cf. example 18g and note 10 above),
means, literally, ‘outside of/beyond three days from now’.

33. Since these expressions contain tia ‘this’ and me ‘come’,

they mean, literally, ‘this coming week’, etc., which also rep-
resents acceptable English.

34. Since these expressions contain se ‘that’ and me ‘come’,

their literal meaning is ‘that coming week’, etc.

35. These expressions, containing the ordinal numbers kot

‘first’, ongeru ‘second’, etc. (see 24.4.4), mean, literally, ‘the
first (day of) work’, etc.

36. Since kȩbȩsȩngil is the possessed form of kȩbȩsȩnge

‘evening’, the literal translation of this expression is some-
thing like ‘on the seventh evening (of July)’.

37. There are other ways of expressing frequency of occurrence

in Palauan. Note, for example, the italicized portions of the
sentences below:

a. A Toki a me mȩngȩtmokl ȩr a blik a lȩ tutau.

‘Toki comes and cleans my house in the morning.’
b. A rȩsȩchal a mo mȩlasȩch a mlai a lȩ suelȩb.
‘The men go making canoes in the afternoon.’

611
 Notes

The italicized expressions in a-b are actually conditional

clauses meaning something like ‘if it is morning’, etc.;
further details on these and other conditional clauses
will be given in 19.4. If we compare example a above
with 35a, we note the following difference of meaning.
Since 35a contains bek ‘each, every’, it implies that Toki
comes and cleans morning after morning, on a very
regular basis. Sentence a above, however, merely asserts
that Toki usually comes and cleans in the morning (as
opposed to the afternoon, for example); furthermore, the
routine may not be as regular as that implied in 35a.

38. The question word oingara ‘when?’ is never preceded (or

introduced) by a (cf. 2.6)

39. Contrast the following sentences:

a. Kȩ milȩngȩsbrebȩr ȩr a blai ȩl oba a ngarang?

‘What (instrument) did you paint the house with?’
b. Kȩ milȩngȩsbrebȩr ȩr a blai ȩr a ngarang?
‘What color did you paint the house?’

While sentence a contains the instrument clause ȩl oba

a ngarang ‘(lit.) using what?’ and therefore asks the
person addressed to name the instrument (e.g. brush,
roller, etc.) used in painting the house (see. 15.3), b is
similar to 43c and asks a question about the medium (or
material) used in painting the house.

40. See. 17.8.1 for further discussion of the constructions con-

taining omȩlmil and omȩngȩlir of 43g and 43h. These con-
structions contain the possessed forms of the derived
action nouns omȩlim ‘drinking’ and omȩnga ‘eating’ (cf.
8.6).

612
 Notes

15 DEPENDENT CLAUSES

1. Note, in addition, that the verb form directly following ȩl

is not preceded (or introduced) by the word a. As we saw
in 2.6, every Palauan verb form is introduced by a unless
the preceding subject is a pronoun. In 3a-b, the absence of
a before the verb forms following ȩ l is probably due to the
fact that these verb forms have no overtly-expressed sub-
jects.

*2. In other words, there is a deep structure constraint which

requires identity between the subject of the main clause and
the subject of the purpose clause introduced by ȩl. The latter
is deleted by an Equi-NP Deletion transformation during the
process of derivation.

3. The noun techall can also refer to an opening in physical

space, as the following sentences indicate:

a. Ng diak a techȩllek ȩl mo ȩr a bitang.

‘I don’t have any space/room to get to the other

side.’
b. Ng diak a techȩllek ȩl soisȩb ȩr a bas.
‘I don’t have any room to get into the bus.’
c. A rȩkangkodang a mlo diak a techȩllir ȩl tuobȩd ȩr a
bas ȩl milsesȩb.
‘The tourists didn’t have any room/way to get out of
the bus which was burning.’

The concrete meaning of techall found in a-c above was

probably extended to include the more abstract idea of
an “opening” or “space” in time, thus resulting in the
meaning ‘opportunity, chance’ observed for techall in 7c-
e above.

613
 Notes

4. The Palauan verb oba is unusual in that it has only per-

fective, but no imperfective forms. The present perfective
forms of oba are listed in 4.9.4, ex. 53; the past perfective
forms are derived simply by replacing word-initial o- by
ul(ȩ)- (e.g. oba ‘use it’—ulȩba ‘used it’, olab ‘use
them’—ullab ‘used them’, etc.) The verb oba also means
‘carry’ or ‘take’, as in the sentences below:

a. Ak oba a ngalȩk. ‘I’m carrying the

child.’
b. Ak ullab a ilumȩl ȩl mo ȩr a ‘I took drinks to the
party. party.’

5. Recall that we used a similar “contextual” approach in 12.5

to confirm our analysis of the basic difference in meaning
between imperfective vs. perfective verb forms.

6. The k of this word is not pronounced because it is preceded

and followed by another consonant. Hence, the phonetic
transcription for this word is [uləbəntərir]. Can you explain
why the ng is pronounced as [n] rather than as [ŋ]?

7. In the following sentences, the italicized specifying clauses

contain past tense verb forms; therefore, the sentence de-
scribes a totally completed past event. The independent
clauses contain change of state expressions consisting
of the directional verb mo ‘go’ followed by the state verbs
ngȩltȩngat ‘lucky’ and mȩkȩrior ‘unlucky’ (cf. 13.5).

a. Ak mlo ngȩltȩngat ȩl miltik a dart ȩl kluk ȩr a rael.

b. ‘I was lucky to find $100 in the road.’
c. Ak mlo mȩkȩrior ȩl ririid a dart ȩ l kluk ȩ l ududek.
‘I had the misfortune of losing $100 of my money.’

8. The perfective forms of this verb—nguu ‘bring/ take it’,

ngmai ‘bring/ take them (non-hum.)’, etc.—are listed in
4.9.4, ex. 53.

614
 Notes

9. If we omit the second specifying clause of 33b, we get the

following sentence, which, although grammatical, differs in
meaning from 33b: Ak mlo ȩr a kȩdȩra ȩl di ngak. ‘I went to
the beach on my own.’ The sentence above has the special
implication that the subject was able to get to the beach or
find his way to the beach without any assistance.

10. Because the expression klde ȩl hong ‘three books’ neces-

sarily designates a plural object, it is impossible to have a
sentence like the following:
*Ak mla rokir a klde ȩl hong.

This sentence is ungrammatical because the perfective form

rokir implies a singular object, and therefore occurrence
before klde ȩl hong is contradictory.

11. For this reason, a more appropriate English equivalent for

blȩchoel might be ‘do habitually’ or ‘invariably do’.

16 OBJECT CLAUSES

1. Notice that the temporal phrase ȩr tia ȩl mlo mȩrek ȩl

rak ‘last year’ happens to contain an occurrence of the past
tense form of mo mȩrek. Cf. chap. 14, note 27.

2. In 8d, the object clause following mla mo mȩrek contains

the state verb bȩchiil ‘married’ (cf. buch ‘spouse’).
Therefore, a word-for-word translation of this sentence
would be something like ‘Droteo has finished being
married.’

3. The transitive verb mȩngoit ‘quit, throw away’ has the fol-
lowing perfective forms:

Present Past
3rd pers. sg. object choitii chilitii
3rd pers. pl. (non-hum) object chȩmoit chiloit

615
 Notes

Some simple sentences containing the imperfective and per-

fective forms of mȩngoit followed by a concrete object
include the following:

a. Ak mo mȩngoit ȩr a ‘I’m going to throw away the

komi. trash.’
b. A Toki a chilitii a ‘Toki divorced Droteo.’
Droteo.
c. Ak chilitii a skuul er ‘I skipped school yesterday.’
a elii.

4. This is the 3rd pers. sg. object present perfective form of

mȩlasȩm ‘try’. The appearance of the full vowel e ([ε]) in the
first syllable is quite unusual (cf. 6.3.1–2).

5. Some of the present and past perfective forms of olȩngȩseu

‘help’ (past: ullȩngȩseu) are listed below:

Person and Number of Present Past

Object
1st pers.sg. ngosukak ngilsukak
2nd pers.sg. ngilsukau ngosukau
3rd pers.sg. ngilsuir ngosuir
3rd pers. (hum) pl. ngosutȩrirngilsutȩrir

The set of object pronoun suffixes observed in the above

perfective forms is identical to that illustrated in 4.9.4,
ex.53, except that the 3rd pers. sg. obj. suffix is -ir (cf.
4.9.4, ex. 54).

6. Some of the present and past perfective forms of oldurȩch

‘tell, ask’ (past: ulȩldurȩch) are given below:

Person and Number of Object Present Past

1st pers.sg. odȩrchak uldȩrchak
2nd pers.sg. odȩrchau uldȩrchau
3rd pers.sg. odȩrchii uldȩrchii

616
 Notes

3rd pers. (hum) pl. odȩrchȩtȩriruldȩrchȩtȩrir

Since oldurȩch is a causative verb in torm, its perfective

forms follow the patterns for causative verbs described
in 9.4. 3rd pers. (hum) pl.

17 PROCESSES OF SENTENCE FORMATION:

SUBJECT SHIFTING AND PREPOSING OF
POSSESSOR

1. The noun following words like soak and chȩtil does not nec-
essarily have to be inanimate (i.e., non-living) as it is in ex-
amples 1a-b. Thus in the sentences below, soak and chȩtil
are followed by human nouns:

a. Ng soak a sensei. ‘I like the teacher.’

b. Ng chȩtil a Satsko. ‘He dislikes Satsko.’

2. This is a noun phrase of characterization—cf. 3.7 and

3.9, ex. 31.

3. Observe the sentences below, which are equivalent to each

other in meaning:

a. Ngak mȩ a Helen a ‘Helen and I are

kausȩchȩlei. friends.’
b. Aki kausȩchȩlei ngak mȩ a
Helen.

Here, the shifted plural subject contains the emphatic

pronoun ngak ‘I’, which refers to the speaker, and the
noun Helen. Since the speaker is part of the shifted
subject, the pronominal trace in b appears as aki, the 1st
pers. pl. excl. non-emphatic pronoun.

4. Cf. the discussion in 16.5, where other sentences con-

taining two objects are analyzed.

617
 Notes

5. The 3rd pers. sg. object pronoun suffix -a is extremely rare.

Cf. 4.9.4.c and chap. 4, note 10.

6. As the English equivalents for 30a and 31 show, the

grammar of English also contains a rule which can inter-
change the two objects of give.

7. Another verb which takes two objects is olisȩchakl ‘teach’.

The order of these objects can be reversed, as the following
sentences show:

a. A Toki a olisȩchakl ȩr a rȩngalȩk a tȩkoi ȩr a

Merikel.
‘Toki is teaching the pupils English.’
b. A Toki a olisȩchakl a tȩkoi ȩr a Merikel ȩr a
rȩngalȩk.
‘Toki is teaching English to the pupils.’

8. For ease of understanding, this discussion has been

somewhat oversimplified. See 17.8.c below.

9. In all of the examples presented in this section, the de-

pendent clause following the possessed forms of soal, chȩtil,
sȩbȩchel, and kirel contains an action verb. Occasionally,
such clauses contain state verbs, as in the example below:

A Droteo a chȩtil ȩl ‘Droteo dislikes being

smechȩr. sick.’

10. Any adequate explanation of the occurrence of ȩl before

the object of omȩruul, omȩnguiu, and omȩlamȩch in 41a-c
would require a complex, lengthy analysis which would be
beyond the scope of the current discussion. Though far over-
simplified, it is sufficient to say that the object of a transitive
verb must be “marked” with ȩl when this transitive verb is
changed into an action noun by the prefixing of o-.

618
 Notes

11. This sentence can also be interpreted as ‘My father dislikes

the things I do at parties.’

12. After the negative verb diak, all nouns must be prefixed
with the 3rd pers. sg. hypothetical pronoun—hence, lsoak.
Further discussion of this phenomenon will be provided in
chap. 18.

18 NEGATION

1. In the most natural dialog, this occurrence of oles would be

prono minalized to ng ‘it’ (cf. 4.8).

2. Though mla is used as the past tense of ngar, it is not really

clear whether these words are related to each other.

3. Indeed, the widespread practice of spelling these sequences

as single words—ngarngii and mlarngii—shows that most
Palauans “feel” them to be indivisible units. The spelling
used in this text, which was also approved by the 1972 Palau
Orthography Committee, is based on the assumption that
ngar ȩr ngii and mla ȩr ngii each consist of three words
which are otherwise always spelled separately. The same
discussion applies to mo ȩr ngii ‘there will be’ and mochu
ȩr ngii ‘there is about to be’, which have been spelled as
morngii and mo churngii.

4. The only situation in which ngar need not be followed by a

locational phrase is illustrated in the sentence below, where
this word occurs in the related meaning ‘alive’:

A rubak a dirk ngar. ‘The old man is still alive.’

619
 Notes

5. Recall that subject shifting must also be applied obligatorily

in certain other types of sentences. Thus, in 17.2, we pro-
posed that obligatory subject shifting is the only plausible
way to account for sentences containing the special pos-
sessed nouns soal ‘his liking’ and chȩtil ‘his disliking’—e.g.,

Ng soal a biang. ‘He likes beer.’

Ng chȩtik a rrom. ‘I dislike liquor.’

6. This sentence can be further transformed by the preposing

of possessor rule to give
A rubak a mlo diak a urerel. ‘The old man lost his job.’

7. As we saw in 6.2.1, hypothetical verb forms normally lack

the verb marker. Thus, if we compare hypothetical
lsȩngȩrengȩr with its non-hypothetical counterpart
songȩrengȩr ‘hungry’, we conclude that the -o- in the latter
form is due to presence of the verb marker. We therefore
assume that songȩrengȩr is derived in the following way:

mȩ + (basic form=verb marker+verb stem)→

sȩngȩréngȩr
s + mȩ + (by metathesis of verb marker) →
ȩngȩréngȩr
s+m+ (by deletion of ȩ)→
ȩngȩréngȩr
s+u+ (by change of verb marker to u in
ȩngȩréngȩr unstressed syllable) →
songȩréngȩr (by vowel blending)

The step-by-step derivation above exactly parallels that for

certain 3rd pers.sg.obj.present perfective forms (cf.
6.3.2, especially ex. 23).

8. Though the verb marker is absent in most hypothetical verb

forms (cf. 6.2.1), a notable exception is found among intran-
sitive state verbs which have the prefixed verb marker mȩ-

620
 Notes

in their pronounced forms. Thus, in the hypothetical forms

of mȩkngit ‘bad’, mȩched ‘shallow’, etc., mȩ- is not deleted,
and we get lȩmȩkngit and lȩmȩched.

9. For a complete listing and discussion of the hypothetical

forms of the directional verbs mo ‘go’ and me ‘come’, see
4.10.

10. Recall (cf. 6.7) that the basic structure of hypothetical per-
fective forms is hypothetical pronoun + verb stem + object
pronoun.

11. Since ergative verbs (cf. 5.4) simply consist of the sequence
verb marker + verb stem, their hypothetical forms have the
structure hypothetical pronoun + verb marker + verb stem.
Note that while most hypothetical verb forms lack the verb
marker, the hypothetical forms of ergative verbs retain it (cf.
6.2.1).

12. As we saw in 4.10.6, complex verb phrases include se-

quences like mo mȩruul ‘will make’, me mȩngȩtmokl ‘come
and clean’, mo mȩrek ‘finish’, mo/me rȩme ‘come/go back’,
mo ungil ‘get better’, etc. In the hypothetical forms of such
expressions a hypothetical pronoun is often prefixed to each
of the parts.

13. We can also speculate that di kea might contain a con-

tracted form of diak followed by e ‘and (then)’. Under this
approach, suggested to me by Robert Gibson (personal com-
munication), we can at least explain why di kea has a neg-
ative meaning.

621
 Notes

19 USES OF HYPOTHETICAL VERB FORMS

1. It is not clear whether the a which introduces Palauan con-

ditional clauses is really a different word from the a which
introduces all noun phrases and verb phrases (cf. 2.6). For
purposes of simplicity, however, we will refer to the a of con-
ditional clauses as a separate word meaning ‘if’.

2. The sequence ngar ȩr ngii ‘there is/are’, which is used in af-

firmative expressions of existence, is explained in detail
in 18.2.

3. With 4c compare the following sentence, which is somewhat

different in structure:

A lak a ududem, e lak chobo ȩr a mubi.

‘If you don’t have any money, (then) don’t go to the

movies.’ In this conditional sentence, the consequential
clause is expressed as an order or command (see 19.5
below).

4. With the general questions of 6, compare the following

general statements:

a. A dȩbo ȩr a che, e ng kired ȩl ousbech a chȩlais.

‘To go fishing, we need a basket.’
b. A dolasȩch a mlai, e ng mȩringȩl.
‘Carving canoes is difficult.’

Though interpreted as general statements, a and b above

are actually conditional sentences which mean some-
thing like ‘If we go fishing, (then) we need a basket’ and
‘If we carve a canoe, (then) it’s difficult.’ The conditional
and consequential clauses of these examples can be per-
muted, resulting in the following sentences:

a’. Ng kired ȩl ousbech a chȩlais a dȩbo ȩr a chei.

‘We need a basket to go fishing with.’

622
 Notes

b’. Ng mȩringȩl a dolasȩch a mlai.

‘It’s difficult carving canoes.’

5. It is not clear how to analyze the word kmu. There is some

possibility that it is related to the kmo of ȩl kmo, which is
used to introduce quotations (see 21.1), or that it is an un-
usual form of the verb dmu ‘say’.

6. Following kmu, the a of ak ‘I’ is deleted. As a result, a kmu

ak ‘if I…’ is pronounced [akmuk].

7. The following conditional sentence, which exhibits the

pattern of 16a-e, contains lak in the conditional clause:

A dȩmal a Droteo a soal a Droteo a lak lolim a biang.

‘Droteo’s father wants him not to drink beer.’

Since the italicized conditional clause requires a hypo-

thetical verb form, and since this clause is negative, we
would expect to find lȩdiak. This form is not acceptable,
however, and we therefore conclude that lak is derived
from lȩ + diak by a phonetic rule of contraction (cf. 18.3
and exs. 4–5 above).

8. The sequence lȩbo lak is the hypothetical form of mo diak

‘will not be, will become non-existent’. Here, too, we have
evidence that the hypothetical form lak is derived from lȩ +
diak by contraction. Cf. note 7 above and 18.3.

9. While ȩr tiang ‘here, at this place’ functions as a locational

phrase in this sentence, ȩ r tiang ‘here, from this place’
functions as a source phrase in 24h. Cf. 14.2 and 14.4.

*10. In 6.2.1, exs. 11–12, we observed just the opposite phe-

nomenon—namely, an initial b is deleted before a following
m. This is found in the derivation of mo ‘go’ from basic mȩ
+ bo (i.e., verb marker + verb stem), where metathesis of
the verb marker and deletion of ȩ result in the sequence

623
 Notes

b + m + o, which becomes mo ‘go’ when the initial b is

deleted before m. The phonetic processes under discussion
suggest the following principle: when two bilabial conso-
nants appear (or come to appear) in word-initial position,
the first one is always deleted.

11. For a discussion of how Palauan passive sentences differ

from ergative sentences, see the concluding remarks in 5.6.

20 QUESTIONS

1. The state verb mȩduch, translated here as ‘strong’, is also

used as a (transitive) state verb meaning ‘know how (to), be
skilled at’ (cf. 16.3).

2. The examples of 20 can also be expressed with subjects

containing chad ‘man, person’ followed by a relative clause
(see chap. 23) which describes or modifies chad. Thus, with
20a-b, for example, compare the following equational sen-
tences, which are identical in meaning:

a. A chad ȩl mlad a Droteo.

‘The one who died is Droteo.’
b. A chad ȩl chillȩbȩdii a Toki a John.
‘The person who hit Toki is John.’

Some linguists would propose that the italicized subjects

of 20a–b are derived from those of a and b above by
deleting the noun chad (which is redundant given the
context) and the relative clause in troducer ȩl.

*3. In 4.6, ex.22a-b, we listed negative sentences like the fol-

lowing, which emphatically deny that some person or
persons were connected with a particular event:

a. Ng dimlak lȩngak a silsȩbii a blai.

‘It wasn’t me who burned down the house.’

624
 Notes

b. Ng dimlak ltir a milkodir a bilis.

‘It wasn’t them who killed the dog.’

The derivation of a–b, though complicated, should now be

clear. Thus the source sentence of a, for example, is the
following:

c. [A silsȩbii a blai a ngak] a dimlak.

As we saw in chap. 18, the subject of the negative verb diak

(past: dimlak) can be a whole (bracketed) sentence, as
in c. In c this bracketed sentence is of the equational
type, and furthermore its subject (italicized) contains a
verb phrase being used as a noun phrase (cf. 20
and 24 above). The source sentence c is transformed as
follows: First, the entire bracketed sentence is moved to
the right of dimlak by the subject shifting rule, leaving
the pronominal trace ng:

d. Ng dimlak [a silsȩbii a blai a ngak].

Next, since the subject of a shifted equational sentence

cannot remain in initial position (cf. 18.6), d must be
changed into the following:

e. Ng dimlak [a ngak a silsȩbii a blai].

At the same time, the hypothetical pronoun lȩ- must be pre-

fixed to the noun immediately following dimlak, thus de-
riving sentence a.

4. When preceded by the specifying word ȩr or the relational

word ȩr, the question word tȩcha ‘who?’ cannot be intro-
duced by a. The reason for this restriction is unclear.

5. A more complicated way of expressing this question is the

following, which uses a passive verb:

625
 Notes

Ng tȩcha a chomoba a hong ȩr ngii?

‘Whose book do you have?’

6. When used as a possessor following a possessed noun,

tȩcha is normally not introduced by a. Cf. note 4 above and
chap. 4, note 3.

7. As this example indicates, the common way of asking

someone’s name in Palauan is to use tȩcha ‘who?’ together
with the appropriate possessed form of ngakl ‘name’.
Therefore, the literal translation for 32e would be ‘Who is
your friend’s name?’ (which is of course unacceptable in
English).

8. In this and the following sentence, the objects tia ȩl babier

‘this letter’ and blai ‘house’ of the transitive verbs milluchȩs
‘was writing’ and silsȩbii ‘burned it down’ have been pre-
posed. Questions of this type seem to be used when the
things referred to by the objects represent old information
for the speaker and hearer—i.e., when they have already
been introduced into the conversation as a topic of dis-
cussion.

9. It is interesting to note that when ngara is preposed in this

way, it is not introduced by the word a.

10. The sequence introduced by ȩl functions as a purpose

clause (cf.15. 2).

11. By applying the rule of subject shifting to this sentence, we

get Ng soam a ngarang? ‘What do you want?’/‘What would
you like?’

12. This and the above example are passive sentences in

which the subject noun phrase (ngara ȩl tȩkoi ‘what lan-
guage?’ or ngara ȩ l mubi ‘what kind of movie?’) corre-
sponds to the object of the related active sentence. As we
saw in 19.7.3, the subject noun phrase of a passive sentence

626
 Notes

sometimes corresponds to a noun phrase which would occur

in a relational phrase in the associated active sentence.
This is true for the examples below:

a. Ngara ȩl kȩdȩra a dȩbo dongȩdub ȩr ngii?

‘What beach are we going swimming at?’
b. Ngara ȩl delmȩrab a losuub ȩr ngii a Droteo?
‘What room is Droteo studying in?’
c. Ngara ȩl blsibs a lȩtilobȩd er ngii a beab?
‘What hole did the mouse emerge from?

The italicized subject noun phrases of a-b designate the lo-

cation of an action, while that of c refers to the source.

13. For an explanation of the meaning and use of the pre-

dictive word ku, cf. 11.12.6.

14. For a discussion of this type of sentence, which contains a

condi tional clause, cf. 19.3.

21 DIRECT AND INDIRECT QUOTATION

1. It is possible that kmo has developed from a form of the

verb dmu ‘say’ (past: dilu). Even though we spell various
combinations of non- emphatic pronoun + dmu in a con-
sistent manner—namely, ak dmu ‘I say’, ng dmu ‘he/she
says’, aki dmu ‘we (excl.) say’, tȩ dmu ‘they say’, etc.—dmu
has a special pronunciation in several cases. After ak ‘I’ and
ng ‘he/she’, the d of dmu changes to [k]; this [k] results from
assimilation (cf. 1.3.3) of d to the preceding velar conso-
nants k and ng. Because of this assimilation, we have the
normal pronunciations [akkmu] and [ŋkmu]. The pronunci-
ation of dmu as [kmu] has spread to other forms where there
is no phonetic reason for the d to change to k: thus, many
speakers pronounce tȩ dmu, for example, as [təkmu]. Al-
though its final vowel is different, the kmo of ȩl kmo might

627
 Notes

be related to the “favored” variant [kmu] of dmu. This same

[kmu] might also appear in the sequence a kmu ‘if’, which
introduces a type of conditional clause (cf. 19.3 and chap.
19, note 5).

2. The double quotation marks used in this and similar sen-

tences are merely a convenient (though artificial) device
for identifying directly-quoted statements. In actual speech,
however, there is no phonetic difference between state-
ments which are directly quoted vs. those which are in-
directly quoted. For this reason, 2a is ambiguous when
spoken—that is, it can also be interpreted as an indirect quo-
tation with the meaning ‘I told Droteo that I wanted to go to
your house.’ Here, of course, blim refers to the house of the
person to whom this utterance is addressed.

3. The verb mȩlȩbȩdebȩk can also mean ‘think of/about’, as in

the sentences below:

a. Ak mȩlȩbȩdebȩk ȩr a {party/beluak}.
‘I’m thinking about {the party/my home}.’
b. Ak millȩbȩdebȩk ȩr a sȩchȩlik ȩl mlo ȩr a Merikel er
se ȩr a lȩme a babier ȩr ngii.
‘I was thinking of my friend who went to America
just when a letter from him arrived.’

In these sentences, mȩlȩbȩdebȩk seems to be followed by a

special subtype of directional phrase which designates
the goal of some activity or state (cf. 14.3.1).

4. As we saw in 7.4, mȩdȩnge is a transitive state verb which

has perfective forms and which can be followed by concrete
objects, as in the following:

a. Ak mle mȩdȩngȩlii a Toki er se ȩr a lȩngalȩk.

‘I knew Toki when she was a child.’
b. Ak mȩdȩngȩlii a kotai ȩr tia ȩl ochur.

628
 Notes

‘I know the answer to this math problem.’

c. A John a mȩdȩnge a tȩkoi ȩr a Sina.
‘John knows Chinese.’

5. Our use of double quotation marks helps to identify this

sentence as containing a directly-quoted question. As we
will see below, the same sentence without double quotation
marks involves an indirectly-quoted question and differs in
meaning:

A Droteo a ulȩker ȩr ngak ȩl kmo ng soam a biang.

‘Droteo asked me if you wanted a beer.’

In this sentence, soam ‘your liking’ of course refers to the

person to whom the utterance is addressed. Whereas
there is no phonetic difference between directly-quoted
vs. indirectly-quoted statements (cf. note 2 above) there
is a significant difference in the pronunciation of
directly-quoted vs. indirectly-quoted questions. While a
directly-quoted question (e.g. 11b) ends with a rise in
intonation (cf. 20.1), an indirectly-quoted question (e.g.,
the example above) does not. A similar discussion can be
applied to 11c-d.

6. While omdasu corresponds to ‘wonder (if…)’ in this ex-

ample, it usually means ‘think, believe’, as in 5a-c above.

22 REASON CLAUSES, RESULT CLAUSES, AND

TIME CLAUSES

1. The temporal phrases ȩr a uche ȩr a chȩldȩchȩduch ‘before

the meeting’ and ȩr a uriul ȩr a chȩldȩchȩduch ‘after the
meeting’ contrast in meaning with the temporal phrases ȩr
a uchȩlel a chȩldȩchȩduch ‘at the beginning of the meeting’
and ȩr a rsel a chȩldȩchȩduch ‘at the end of the meeting’ (cf.
3.5).

629
 Notes

23 RELATIVE CLAUSES

1. The concept of “distribution” is introduced in 2.1 and 2.3.

2. Recall that bracketed sentences are found only in struc-

tures which serve as the (abstract) source for sentences
that are actually spoken. Bracketed sentences must be
shifted, or must have some of their elements deleted, when
the source sentences of which they are a part are trans-
formed into actually-spoken sentences. These concepts
are discussed in detail in 17.2, 17.7, and 18.2.1.

3. As we saw in 15.1 and in chap. 16, the same word ȩl

also functions to introduce dependent clauses and object
clauses. A further use of ȩl will be observed in chap. 24.

4. This sentence can also be translated as ‘I’m sad about my

friend’s having died.’ A similar example is the following:

A Toki a ungil a rȩngul ȩr a Droteo ȩl mȩsisiich.

‘Toki is happy that Droteo is well.’

5. This sentence, which translates literally as ‘There’s some

news of Toki which is known by me’, has a rather unfa-
vorable connotation—that is, it implies that the news is
about something bad, unusual, etc. that Toki did.

24 MODIFIERS

1. For a review of how the concept of bracketed sentences is

used in the derivation of relative clauses, cf. 23.2.

2. Since this difference was described in detail in 23.5, it will

not be repeated here.

3. The sequence ngodȩch ȩl chad can also mean ‘someone/

anyone else’, as in the sentence below:

630
 Notes

a. Ng dimlak a ngodȩch ȩl chad ȩl mle ȩr a party?

‘Wasn’t there anyone else at the party?’

4. The word kid is added to an equational sentence when

the speaker wishes to emphatically point out the location of
someone or something.

5. Through sheer oversight, the author failed to include yet

another set of Palauan demonstrative words representing a
fourth category of distance. These omitted items designate
persons or things near the speaker but relatively far from
the hearer. Thus, the demonstrative words tie ‘this’ and aile
‘these’ should be added to the chart in 15, while the demon-
strative words ngile ‘this person’ and tirile ‘these people’
need to be added to the chart in 16. The over-all analysis
presented in 24.3 must also be modified accordingly.

6. The i of ngi- is deleted optionally before the k-initial mor-

phemes -ka ‘located near speaker and hearer’ and -ke ‘lo-
cated far from speaker and hearer’. When this i is dropped,
the word-initial ng becomes syllabic (cf. 1.3.5)—i.e., we
have ngke [ŋkey] and ngka [ŋkaŋ].

7. The i of tiri- is obligatorily deleted before the k-initial mor-

phemes -ka and -ke. Cf. note 6 above.

8. Note that a human noun like sensei ‘teacher’ does not take
the plural prefix rȩ- (cf. 2.5) in constructions of the form
demonstrative + ȩl

+ noun. In such cases, the plural prefix rȩ- is unnecessary (or

redundant) because the demonstrative word itself automati-
cally designates singular vs. plural. Note, therefore, the con-
trast between the following:

ngilȩcha ȩl sensei ‘that teacher’

tirilȩcha ȩl sensei ‘those teachers’

631
 Notes

Needless to say, it is also the demonstrative word alone

which distinguishes between singular and plural in a pair
like the following,

tia ȩl hong ‘this book’

aika ȩl hong ‘these books’

where the head noun is non-human and therefore could

never be preceded by rȩ- anyway.

9. As we will see in 24.4.3 below, Palauan number words are

always preceded by a (cf. 2.6) when they are used in sen-
tences or as part of expressions such as tȩruich mȩ a ta
‘eleven’ or tȩruich mȩ a eru ‘twelve’, etc. In tȩruich mȩ a ta
‘eleven’ the ȩ of mȩ is not pronounced, resulting in [təruyɂə-
mataŋ] (cf. 1.5.e, ex. 54). And in tȩruich mȩ a eru ‘twelve’,
etc. both the ȩ of mȩ and the a before vowel-initial eru
‘two’ are omitted, giving the pronunciation [təruyɂəmεruŋ]
(cf. 1.5.a, ex. 40).

10. Further expressions of this type are given in 14.6, ex. 34g.

11. Further expressions indicating hours of the day are found

in 14.6, ex. 34d.

12. Recall that the plural prefix rȩ- can be optionally attached
to the number word in expressions like these. Thus, in ad-
dition to teim ȩl sensei ‘five teachers’ we can have rȩteim ȩ l
sensei, with no change in meaning. Cf. the discussion in 2.5,
especially ex. 25.

13. It is possible that teliud ‘one bunch’ is a contracted (or

shortened) form of what was originally the sequence ta ȩl
iud. This very same structure is actually found in tȩruich
ȩl iud ‘ten bunches’. It also seems plausible that teluo ‘one
(long object)’ of Set V has developed by contraction from a
sequence ta ȩl uo.

632
 Notes

14. The word ma (not to be confused with mȩ a… [ma] ‘and…’)

also means ‘first’. This word does not have to be used to-
gether with kot, as the following example (equivalent to 42b
in meaning) shows:

Kȩ ma mo ȩr a skuul, e ngak ekong.

‘You go on ahead to school, and then I’ll follow.’

15. In the sentence below, the ordinal number ongeru ‘second’

is used in exactly the same way:

Ak ongeru ȩl ‘I’ll run second (in the race,

rȩmurt. etc.).’

16. Temporal phrases containing this and the preceding ex-

pression are given in 14.6, exs. 34b-c.

17. The qualifying word di ‘only, just’ can also precede nouns,
as in the following sentences:

a. Aki di ngalȩk ȩr a skuul. ‘We’re just students.’

b. Ak mȩnga a di iasai. ‘I eat vegetables only.’

18. The qualifying word kilo ‘almost, nearly’ may be the past
tense form of the special verb ko ‘just’ (cf. 15.7.7, ex. 56).

25 THE CONNECTING WORDS MȨ AND E

1. The Palauan sentences in 3b-c indeed contain result clauses

introduced by mȩ, even though this is not directly reflected
in the English translations. Detailed discussion of such cases
is given in 22.1.1.

2. The slang expression mo bad ‘go to sleep’ means, literally,

‘become (like) a rock’.

633
 Notes

3. It is possible that a lȩchub is some kind of conditional

expression (cf. 19.1) in which a ‘if’ is followed by a “fos-
silized” hypothetical verb form containing the 3rd pers. sg.
hypothetical pronoun lȩ-. The fact that the clause following
mȩ a l ȩchub is introduced by e lends support to this spec-
ulation, since e introduces consequential clauses which
follow conditional clauses (cf. 19.1 and 19.3).

4. In the clause ng mirrael a Droteo ‘Droteo left’, the subject

Droteo has been shifted to clause-final position, leaving the
pronominal trace ng (cf. 17.2).

634
 Appendix

Appendix
Guide to Phonetic Symbols Used in This Textbook

Phonetic Symbol Corresponding Sample Palauan Word

Letter
in Palauan
Spelling

[b] b blai [blay] ‘house’

[p] b brer [prεr] ‘raft’
[t] t tkul [tkul] ‘its edge’
d dmak [tmakh]
‘together’
[th] t chat [ɂath] ‘smoke’
[d] d dub [dup] ‘dynamite’
[θ] d dngod [θŋoð] ‘tattoo
needle’
[ð] d mad [mað] ‘eye’
[k] k ker [kεr] ‘question’
[kh] k brak [prakh] ‘taro’
[g] k rȩkas [rəgás]
‘mosquito’
[x] k chȩrmek [ɂərmέx] ‘my
animal’1
[ɂ] ch charm [ɂarm] ‘animal’
[s] s sers [sεrs] ‘garden’
[m] m mlim [mlim] ‘your
canoe’
[ṃ] m mdak [ṃðákh] ‘be
together!’
[ŋ] ng ngor [ŋor] ‘mouth’
[ŋ] ng ngklem [ŋklέm] ‘your
name’
[n] ng iungs [yuns] ‘island’
n nas [nas] ‘eggplant’
[ṇ] ng ngduul [ṇdúwl] ‘clam’
[l] l rael [ráεl] ‘road’

635
 Appendix

[ḷ] l lmuut [lmuwth]

‘return’
[l:] ll kall [kál:ə] ‘food’
[r] r reng [rεŋ] ‘heart,
spirit’
[ṛ] r rsel [ṛsέl] ‘its end’
[r̄] rr rrom [r̄om] ‘liquor’
[h] h hong [hoŋ] ‘book’
[ǰ] j daijiob [dayǰyób] ‘all
right’
[z] z keizai [kéyzay]
‘economics’
[i] i sils [sils] ‘sun’
[u] u bung [buŋ] ‘flower’
[ε] e oles [olέs] ‘knife’
[ə] ȩ chȩlat [ɂəláth]
‘smoked (fish)’
(sometimes not ralm [rálmə] ‘water’
spelled)
[o] o ngor [ŋor] ‘mouth’
[a] a bad [bað] ‘stone’
[y] i blai [blay] ‘house’
e eolt [yóltə] ‘wind’
[w] u uel [wεl] ‘turtle’
o oach [waɂ] ‘leg’
[í], [ú], etc. i, u, etc. chȩtil [ɂətíl] ‘doesn’t
(stressed vowels) like’

636
 Glossary of Terms

GlossaryofTerms
abstract noun—a noun which identifies a concept, idea,
quality, or emotion—i.e., something which lacks any dis-
cernible physical attributes and therefore cannot be per-
ceived by any of the five senses.
accompaniment clause—a type of dependent clause which
identifies the person(s) with whom the action or activity of
the preceding independent clause is being performed.
action noun—a noun which designates an action or activity as
an abstract concept. Action nouns are derived by prefixing
o- to transitive or intransitive action verbs.
action verb—a verb which describes an action, activity, or
event.
active sentence—a sentence type in which the agent appears
as the subject of a transitive verb and the person, animal,
or thing affected by the action of the transitive verb appears
as its object. In active sentences, attention is focused on the
agent and what the agent does.
affirmative command—an imperative sentence in which the
speaker orders the hearer to do something.
affirmative expression of existence—a sentence which as-
serts the existence of something or introduces something
into a conversation as new information.
affirmative sentence—a sentence which asserts or affirms the
occurrence of some action, event, state, condition, etc.
affix—a morpheme which cannot occur as an independent word
but which must be attached to, or inserted into, another
word. The three types of affixes are prefixes, suffixes, and
infixes.
affix pronoun—a pronoun which does not occur as a separate
word but which must be attached to another word as a prefix
or suffix.
agent—the doer of an action; the person who performs, carries
out, or causes an action.

637
 Glossary of Terms

allophones—the non-contrastive predictable phonetic varieties

of a particular phoneme. The allophones of a given phoneme
never provide us with minimal pairs.
alveolar—a consonantal sound involving the alveolar ridge.
alveolar ridge—the hard ridge behind the upper teeth which
serves as a point of articulation for the Palauan consonants
s, r, and l.
ambiguous—having two (or more) different meanings; inter-
pretable in more than one way.
animate noun—a noun referring to something which is alive
and can move by itself—i.e., to a human being or animal.
anticipating state suffix—a suffix of the form -(ȩ)l, -all, or long
vowel + l which combines with verb stems to form antici-
pating state verbs.
anticipating state verb—a state verb formed with the suffixes
-(ȩ)l, -all, or long vowel + l which describes the state or
condition that someone or something is expected to be in.
Many anticipating state verbs function as nouns with special
meanings.
appositional structure—a structure of the form noun phrase
of possession + ȩl + noun in which the two elements linked
by ȩl are equated with each other.
archaic—term used to describe a form or usage which has
nearly disappeared from a language.
articulator—a speech organ (or part of a speech organ) which
moves and touches some other speech organ during the pro-
duction of a speech sound.
aspiration—an audible puff of air accompanying the pronunci-
ation of certain Palauan stop consonants.
assimilation—a process by which the pronunciation of a given
sound is changed so as to become similar or identical to that
of an adjacent or neighboring sound.
auxiliary word—a helping word such as mo ‘go’ or mle ‘was,
were’ which accompanies a verb and shows the tense of the
verb or some other information.
back vowel—a vowel pronounced with the tongue retracted to-
wards the back of the mouth.

638
 Glossary of Terms

basic stem—an abstract representation of a noun or verb stem

which does not occur in actual pronunciation or writing but
which allows us to predict the correct pronunciation of the
various forms in which it occurs.
basic structure—a formula which shows all of the elements
(morphemes) which compose a particular noun or verb form.
The elements in the basic structure of a form must often be
modified considerably in order to obtain the actually-spoken
form.
bilabial—a consonantal sound produced with the two lips.
bound form—a morpheme which can never occur alone as an
independent word but must always be connected (or bound)
to some other morpheme(s). All affixes and some verb stems
are bound forms.
bracketed sentence—a whole sentence which functions as a
subject noun phrase in certain source sentences. Bracketed
sentences must be moved by the process of subject shifting.
causative marker bȩk—a morpheme with causative meaning
which is part of the causative prefix omȩk-.
causative prefix—a prefix of the form omȩ(k)- or ol(ȩ)- which is
added to a verb stem to derive a causative verb.
causative verb—a special type of transitive action verb formed
with the causative prefix. Causative verbs describe actions
in which the subject causes or forces someone or something
to perform a particular action or be in a particular state.
cause phrase—a relational phrase which indicates the cause
of, or reason for, some state, condition, or action.
central vowel—a vowel pronounced in the center of the mouth
with the tongue in “neutral” position—i.e., neither advanced
(as for front vowels) nor retracted (as for back vowels).
change of state expression—a sequence consisting of the
auxiliary word mo ‘go, become’ and a state verb. Such ex-
pressions imply a change from an earlier state to a later, re-
sulting state.
characterizational phrase—a type of relational phrase which
characterizes the immediately preceding noun by describing
its function, purpose, content, origin, etc.

639
 Glossary of Terms

complex noun—a noun which contains at least two morphemes

and is formed by adding certain types of affixes to verb
stems or to other nouns.
complex verb form—a verb form which is composed of two or
more morphemes.
complex verb phrase—a verb phrase consisting of two or
three separate words.
concrete noun—a noun which identifies something that has
discernible physical attributes and can therefore be per-
ceived by one or more of the five senses—i.e., something we
can see, hear, touch, taste, or smell.
condition—an event, action, state, etc. whose occurrence is
suggested or put forth as a possibility; occurs together with
the consequent to form a conditional sentence.
conditional clause—the clause of a conditional sentence
which expresses the condition. Conditional clauses contain
hypothetical verb forms and are introduced by the condi-
tional clause marker a ‘if’.
conditional clause marker a—the word, equivalent to English
‘if’, which introduces Palauan conditional clauses.
conditional sentence—a sentence type consisting basically of
two parts—a condition and a consequent—and expressing
the idea that if some event, action, state, etc. occurs, then
something else will happen.
connecting words mȩ and e—words which connect two
simple sentences into one and relate the ideas which they
represent.
consequent—an event, action, state, etc. which takes place
as a result or consequence of some other event, action, or
state; occurs together with the condition to form a condi-
tional sentence.
consequential clause—the clause of a conditional sentence
which expresses the consequent. Consequential clauses are
introduced by e ‘(and) then’ if they follow the conditional
clause.
consonant—a sound produced by partial or complete inter-
ruption of the outward air flow.

640
 Glossary of Terms

consonant cluster—a series of two adjacent consonants in the

same syllable.
consonant-final (word or stem)—a word or stem ending in a
consonant.
consonant-initial (word, stem, or suffix)—a word, stem, or
suffix beginning with a consonant.
context—the broad environment of a word, phrase, or sen-
tence; the words or sentences which surround a particular
word or sentence and have bearing on its meaning.
contextual restrictions—the restrictions imposed by the
context or environment on the correct or acceptable occur-
rence of particular forms.
contrastive sounds—sounds which contrast with each other
(= phonemes) in otherwise identical environments. The
sounds represented by Palauan t and d are contrastive be-
cause they occur in the environment -ub and therefore dif-
ferentiate tub ‘spit’ from dub ‘dynamite’. See also
phonemes, significant sounds.
coordinate noun phrase—a noun phrase consisting of two or
more nouns (or noun phrases) joined by the connecting word
mȩ.
demonstrative—a word which is used to point out or draw at-
tention to a particular person, animal, or thing.
dental—a consonantal sound involving the teeth.
dependent clause—a clause without an overtly-expressed
subject whose full interpretation depends on information
found in the preceding (independent) clause.
dependent clause introducer ȩl—a word which appears at
the beginning of, and therefore introduces, dependent
clauses.
derived action noun—see action noun.
derived plural noun—a plural noun formed (or derived) by
prefixing the plural prefix rȩ- to a state verb or action verb.
derived state verb—a state verb which is derived by adding
one or more affixes to a verb stem or a noun stem.

641
 Glossary of Terms

directional phrase—a relational phrase which indicates the

goal or termination point of some action involving movement
or the characteristic goal of some action or state.
directional verb—a special type of intransitive action verb
which indicates movement in a particular direction.
direct quotation—word for word repetition by the speaker of
someone else’s statement or assertion.
dissimilation—a process in which the pronunciation of a given
sound is changed so as to become different from that of an
adjacent or neighboring sound.
distribution (of a sound or word)—the positions (or environ-
ments) in which a given sound or word occurs relative to
other sounds or words.
duration—the period of time over which an action or state
takes place.
emphatic pronoun—a type of independent pronoun which,
when used as sentence subject, makes exclusive reference
to a person or persons and implies a contrast with other
persons who might be (but are not) associated with the
action or state in question. Emphatic pronouns also occur
after the relational word ȩr or the specifying word ȩr, but
without any sense of exclusive reference or contrast.
environment—the sound or sounds adjacent or close to some
other sound; the word or words adjacent or close to some
other word.
equational sentence—a sentence type in which two noun
phrases are equated with each other.
ergative sentence—a sentence type containing an ergative
verb. In ergative sentences, the subject of the ergative verb
is viewed as having undergone the effect of some action, and
the agent or cause of this action is de-emphasized.
ergative verb—a verb form consisting of the verb marker and
a verb stem which takes as its subject what would be the
object of the related transitive verb.
existential (state) verb—a verb which asserts the present or
past existence of the subject in a particular location—i.e.,
ngar ‘be (located)’ and mla ‘was (located)’.

642
 Glossary of Terms

expanded noun phrase—a noun phrase consisting of a head

noun followed by a relative clause or of a head noun pre-
ceded by a modifier.
explicit—shown by an actual marker or form; overt.
first person plural exclusive—a pronoun or possessor suffix
which refers solely to the speaker and at least one other
person associated with him, but which excludes reference to
the person(s) spoken to.
first person plural inclusive—a pronoun or possessor suffix
which not only makes reference to the speaker (and anyone
associated with him) but also includes reference to the
person(s) spoken to.
first person pronoun—a pronoun (‘I’ or ‘we’) which refers to
the speaker of the sentence or the group to which he be-
longs.
formal—referring to the actual form or structure of a word,
grammatical construction, or sentence.
fricative—a consonantal sound involving partial closure or con-
striction between the articulator and point of articulation,
resulting in audible friction.
front vowel—a vowel pronounced with the tongue advanced to-
wards the front of the mouth.
full vowels—the non-neutral vowels i, u, e, o, and a, which are
pronounced with full vowel quality or “color”.
future tense—a tense used to describe an action or state which
will occur in the future (i.e., at some point in time following
that of the utterance). It is represented by the auxiliary mo
‘go’ followed by an action or state verb.
general question—a question about a habit or general truth
rather than a specific event.
general statement—a sentence which expresses a broad gen-
eralization, habitual occurrence, or general truth rather
than a specific event.
glide sounds—sounds characterized by a gliding movement of
the tongue towards a high front position (y-glide) or high
back position (w- glide).

643
 Glossary of Terms

glottal—a consonantal sound (usually a stop) involving the

glottis (or vocal cords).
glottis—the space between the vocal cords.
grammatical system of a language—the contrasting units of
meaning (morphemes or words) and their distribution.
habitual statement—a sentence which describes a habit or re-
peated action which the subject engages in.
head noun—the noun which precedes a relative clause or
follows a modifier construction.
high vowel—a vowel pronounced with the tongue raised high
in the mouth and close to the palate.
homonyms—words which are pronounced the same but have
distinct meanings. Some nouns are homonyms in their inde-
pendent forms but have distinct possessed forms.
human noun—a noun which refers to a human being.
hypothetical pronoun—a type of affix pronoun which is pre-
fixed to verbs in a large variety of complex grammatical con-
structions which involve hypothetical (i.e., unreal) events or
situations.
hypothetical verb form—a verb form containing a prefixed
hypothetical pronoun and usually lacking the verb marker.
Hypothetical verb forms designate unreal, supposed, or
imagined events or states rather than real ones.
imminent (action or event)—an action or event which is just
about to occur or is likely to occur in the very near future.
imperative verb form—a verb form used to express an order
or command. Palauan imperative verb forms are actually hy-
pothetical verb forms prefixed with the second person hypo-
thetical pronoun.
imperfective marker—a morpheme with various forms (-l-, -
ng-, and -m-) which functions to mark or identify a particular
transitive verb as imperfective. The imperfective marker is
placed between the verb marker and the verb stem and
causes the initial consonant of the following verb stem to
delete.

644
 Glossary of Terms

imperfective verb—a transitive verb form which consists of

the verb marker, the imperfective marker, and a verb stem
and which designates an action which is in progress and has
not been brought to completion or perfection.
implicit—not overtly expressed by some marker or form; im-
plied.
inanimate noun—a noun referring to something which cannot
move by itself—i.e., to something other than a human being
or an animal.
inchoative suffix -a—a suffix added to verb stems which de-
notes an action or state that has just come about and is
therefore new and unexpected.
independent clause—a clause which precedes a dependent
clause or object clause in certain types of complex sentences
and which can occur alone as a separate (or independent)
sentence because it contains a fully specified subject and
verb.
independent form of a noun—the unpossessed form of a
noun—i.e., the noun stem without any prefix or suffix.
independent pronoun—a pronoun which occurs as a separate
word rather than as a prefix or suffix.
indirect quotation—a summary or rephrasing by the speaker
(i.e., from the speaker’s point of view) of someone else’s
statement or assertion.
infix—any morpheme which is inserted into a word.
initial state—a state which existed before some process of
change began.
instrument clause—a type of dependent clause which iden-
tifies the instrument used in performing the action or ac-
tivity mentioned in the preceding independent clause.
instrument noun—a concrete noun which designates the tool,
implement, or utensil used in performing a particular action.
intentional causation—a type of causation in which the
subject of a causative verb means or intends to bring about
the action or state in question.

645
 Glossary of Terms

internal structure (of a word)—the structure of a word in

terms of the morphemes (stems and affixes) from which it is
formed.
intervocalic—appearing between two vowels.
intonation—the rising or falling pitch of the voice, sometimes
used to distinguish Palauan questions from statements.
intransitive action verb—a verb which names an action or ac-
tivity that involves only a doer, but no receiver.
intransitive sentence—a sentence type consisting of a subject
noun phrase and an intransitive verb (followed optionally by
a relational phrase).
intransitive state verb—a state verb which involves only a
sentence subject, but no sentence object.
liquid—a consonant pronounced by making a partial closure in
the mouth, but without any audible friction.
locational phrase—a relational phrase which indicates the lo-
cation or place of some action or state.
long l—a pronunciation of l which is held about twice as long as
that of the single consonant; spelled ll and transcribed pho-
netically as [l:].
long vowel—a vowel pronounced with greater length than the
corresponding short (or regular) vowel. Palauan long vowels
are also characterized by a gliding articulation.
low vowel—a vowel pronounced with the tongue low in the
mouth, relatively distant from the palate.
main clause—another term for independent clause.
manner of articulation—the way in which a speech sound is
produced; involves complete vs. partial stoppage of air flow,
nasal vs. non-nasal articulation, etc.
marked forms—forms which are identified or marked by the
overt presence of a particular morpheme (usually an affix).
For example, Palauan past tense verb forms are easily iden-
tified because they always contain the past tense marker -il-
or -l-.

646
 Glossary of Terms

means of transportation clause—a type of dependent clause

which specifies the means of transportation used to move
from one location to another.
metathesis—a phonetic process in which sounds or groups of
sounds exchange positions. As a result of this process, the
verb marker mȩ- appears to have been infixed into verb
stems as -(ȩ)m-, -u-, or -o-.
mid vowel—a vowel pronounced with a tongue position some-
where between high (close to the palate) and low (relatively
far from the palate).
minimal pair—a pair of words which differ from each other
with respect to a single phoneme—e.g. blai ‘house’ vs. mlai
‘canoe’, chad ‘person’ vs. chat ‘smoke’, etc.
modifier—any grammatical construction in which a word or
group of words is linked by ȩ l to a following (head) noun.
morphemes—meaning-bearing units which combine with each
other in the formation of words; minimal units of meaning.
morphology—the study of how different morphemes combine
with each other in the formation of words.
nasal sound—a consonant pronounced with air flowing freely
through the nose but not the mouth.
negative command—an imperative sentence in which the
speaker orders the hearer not to do something.
negative expression of existence—a sentence which intro-
duces something into a conversation as new information
but at the same time denies the existence of this particular
thing.
negative sentence—a sentence which denies the occurrence
of some action, event, state, condition, etc. Nearly all
Palauan negative sentences contain some form of the neg-
ative verb diak.
negative verb diak—an intransitive state verb, equivalent to
English ‘isn’t, doesn’t exist,’ which is found in nearly all
Palauan negative sentences.

647
 Glossary of Terms

neutral vowel—the mid central vowel schwa (phonetically

transcribed [ə]), so called because it is articulated in a po-
sition which is least extreme or deviant in terms of tongue
height and tongue advancement.
non-contrastive sounds—slightly different pronunciations of a
phoneme (= allophones) which are usually predictable and
automatic and which do not result in minimal pairs.
non-emphatic pronoun—a type of independent pronoun
which, when used as sentence subject, refers to a person
or persons without any particular sense of emphasis or con-
trast.
non-human noun—a noun which refers to something which is
not human—i.e. to an animal or thing.
non-intentional causation—a type of causation in which the
subject of a causative verb does not consciously mean or
intend to bring about the action or state in question.
non-specific noun—a noun which does not refer to any par-
ticular person, animal, or thing but instead refers to some-
thing in a general way.
noun—a word (or part of speech) which names or makes ref-
erence to various types of things and living beings.
noun phrase—any single word or group of words which can
occur as sentence subject or object, or following the re-
lational word ȩr. All Palauan noun phrases must at least
contain a noun.
noun phrase of characterization—a noun phrase consisting
of either of the following:

(i) a possessed noun with a 3rd pers. sg. or pl. possessor

suffix followed by another noun (phrase) which
characterizes the possessed noun by describing its
function, purpose, content, origin, etc.
(ii) the independent form of an optionally possessed noun
followed by a characterizational phrase.

noun phrase of possession—a noun phrase consisting of any

of the following:

(i) a single possessed noun.

648
 Glossary of Terms

(ii) a possessed noun with a 3rd pers. sg. or pl. possessor

suffix followed by another noun (phrase) which specifies
or identifies the possessor.
(iii) an unpossessible noun followed by a possessor phrase.

noun stem—way of referring to a noun when it appears in com-

bination with a prefix or suffix.
number morphemes—morphemes designating numbers that
combine with various prefixes to form number words for
counting different categories of things.
number word—any word which designates a number. Most
Palauan number words consist of a prefix identifying the cat-
egory of thing being counted followed by a number mor-
pheme.
object clause—a type of dependent clause which is used in sen-
tence object position following certain transitive verbs.
object pronoun—a type of affix pronoun which is suffixed to
perfective verb forms and which identifies the person or
persons affected by the action of the verb.
obligatorily possessed noun—a noun which has no inde-
pendent form and must therefore always occur with a pos-
sessor suffix.
obligatory—must occur, apply, etc.
optional—may or may not occur, apply, etc.
optionally possessed noun—a noun which may (but does not
have to) occur in a possessed form—i.e., a noun which can
appear independently as well as with possessor suffixes.
ordinal number—a number used to indicate the order or rank
of something—i.e., first, second, etc.
overt—shown by an actual marker or form; explicit.
paradigm—a set or group of related forms.
parts of speech—classes or groups of words in a language
whose members have common distributional character-
istics.
passive sentence—a sentence type containing the hypothetical
form of a transitive verb in which the person, animal, or
thing affected by the action of the transitive verb appears

649
 Glossary of Terms

in sentence subject position and the agent appears in

sentence-final position. In passive sentences, attention is fo-
cused on the object and what happens to it.
past tense—a tense used to describe actions or states which
occurred or were in progress in the past (i.e. at some point
in time preceding the time of utterance).
past tense marker -il- or -l-—a morpheme indicating the past
tense which is infixed after the initial consonant of a verb.
perfective verb—a transitive verb form which contains a suf-
fixed object pronoun and which designates an action which
is brought to completion or perfection.
permutation—exchange of position (as in the permutation of
conditional and consequential clauses).
phonemes of a language—units of sound which contrast with
each other and serve to differentiate words from each other.
See also con trastive sounds, significant sounds.
phonetic symbol—a symbol written in square brackets and
used to identify the actual pronunciation of a sound—e.g.
[b], [ð], [a], etc. The phonetic symbol for a sound may not
necessarily correspond to the letters used in Palauan
spelling—e.g. [ɂ] for ch. See also phonetic transcription.
phonetic transcription—the system (or practice) of writing
down the words of a language in phonetic symbols. Phonetic
transcriptions of words are enclosed in square
brackets—e.g. [mað] for mad ‘eye’. See also phonetic
symbol.
phrase—a group of associated words. Palauan has three major
types of phrases—noun phrases, verb phrases, and rela-
tional phrases.
plural—referring to two or more persons, animals, or things.
plural prefix—a prefix (rȩ- or r-) which can be added only to
human nouns to form plurals.
point of articulation—the place where contact is made during
the production of certain speech sounds.
possessed noun—any noun consisting of a noun stem followed
by a possessor suffix.

650
 Glossary of Terms

possession—relationship between a noun stem and a possessor

suffix such that the person or thing to which the suffix refers
owns or possesses whatever is designated by the noun stem.
possessor phrase—a type of relational phrase which identifies
the possessor of the immediately preceding (unpossessible)
noun.
possessor suffix—a suffix which is added to a noun stem to des-
ignate the possessor.
predictive suffix -u—a suffix added to verb stems which ex-
presses the speaker’s prediction that the action of the verb
is about to happen.
prefix—any bound morpheme which is attached to the be-
ginning of a word or morpheme.
preposing of possessor—a grammatical process in which a
noun phrase designating a possessor is preposed (or moved
forward) to sentence-initial position, where it substitutes for
a pronominal trace. The possessor to be preposed must al-
ready have been moved to the right of the verb phrase as
part of a shifted subject.
preposition—term used in English grammar to refer to relating
words such as by, with, on, for, etc.
present tense—a tense used to describe actions or states
which are in progress at the present time (i.e. at the time
when the sentence is uttered). The present tense is also
used in general statements and habitual statements.
productive—term used to describe a pattern of word formation
which is widely used by the speakers of a language in
coining new words.
progressive—in progress; in the process of taking place.
pronominalization—a grammatical process by which a noun
(or noun phrase) is replaced by a non-emphatic pronoun
under certain circumstances.
pronominal trace—a 3rd pers. sg. or pl. non-emphatic
pronoun which occurs in sentence-initial position as a result
of the grammatical process of subject shifting.

651
 Glossary of Terms

pronoun—a short word referring to persons such as ‘I’, ‘you’,

‘he’, etc.; a word which substitutes for a noun or noun
phrase in the proper context.
prepositive verb form—a verb form used when the speaker
wishes to propose or suggest that he and the hearer(s)
perform some action or activity together. Palauan propos-
itive verb forms are actually hypothetical verb forms pre-
fixed with the first person plural inclusive hypothetical
pronoun.
purpose clause—a type of dependent clause which explains
the purpose of the action or activity mentioned in the pre-
ceding independent clause.
qualifying word—a word which qualifies or limits the meaning
of a directly following verb.
question—any sentence which asks for or demands information
of one kind or another.
question word—a word which occurs in question sentences
and is used to ask about the specific identity of a person,
place, thing, etc.
reason clause—a clause introduced by e le ‘because’ which ex-
plains the cause or reason for the event or state described
in the preceding clause.
recent past (tense)—a tense which refers to an event that hap-
pened in the recent past (i.e., at a past time point not too
distant from that of the utterance itself) or to past expe-
rience. It is represented by the auxiliary mla followed by an
action verb.
reciprocal—affecting each other; having a mutual effect.
reciprocal prefix—a prefix with many variant forms (kai-, kau-,
kaiu ȩ-, kȩ-, ka-, and cha-) which is added to verb stems or
nouns to derive reciprocal verbs.
reciprocal verb—a verb formed with the reciprocal prefix
which describes a situation in which two or more persons or
things are affecting each other in some way.
reduced vowel—the neutral vowel ȩ (schwa), often resulting
from the reduction or weakening of the full vowels i, u, e, o,
and a in unstressed syllables.

652
 Glossary of Terms

redundant—containing superfluous or unnecessary infor-

mation.
reduplicated syllable—a syllable which has been formed by a
process of reduplication.
reduplication—a phonetic process which involves repeating
(or reduplicating) part or all of a verb stem or noun, usually
resulting in a significant change of meaning. Palauan redu-
plication patterns are very complex and have many irregu-
larities.
relational phrase—any sequence consisting of the relational
word ȩr followed by a noun (phrase) or a pronoun. Relational
phrases provide such information as the place or time of an
action or state.
relational word ȩr—a widely-used Palauan word which func-
tions to express various types of relationships such as lo-
cation, time, possession, and the like. This word corresponds
to English ‘on’, ‘at’, ‘in’, ‘from’, ‘of’, etc.
relative clause—a clause introduced by ȩl which relates a par-
ticular piece of information to the preceding (head) noun.
relative clause introducer ȩl—a word which appears at the
beginning of, and therefore introduces, relative clauses.
result clause—a clause introduced by mȩ‘(and) so’ which ex-
plains the result or consequence of the event or state de-
scribed in the preceding clause.
resulting sentence—a sentence derived as a result of applying
certain grammatical processes to a source sentence.
resulting state—a state which has come about as the result of
some process of change.
resulting state infix -(ȩ)l—an infix which is inserted after the
initial consonant of a verb stem to form a resulting state
verb.
resulting state verb—a state verb formed with the infix -(ȩ)l-
which is used to describe the state or condition someone or
something is in as a result of a particular action. Many re-
sulting state verbs function as nouns with special meanings.
rounded vowel—a vowel pronounced with simultaneous
rounding of the lips.

653
 Glossary of Terms

schwa—a mid central vowel occurring only in Palauan un-

stressed syllables; transcribed phonetically as [ə] and
spelled as ȩ in this text.
schwa release—a predictable pronunciation of schwa following
a consonant cluster in word-final position.
second person pronoun—a pronoun (‘you’) which refers to the
person or persons addressed.
sentence object—the noun or noun phrase which follows a
transitive verb and designates the person, animal, or thing
affected by the action of the verb.
sentence subject—the noun or noun phrase which appears
at the beginning of a sentence preceding the verb or verb
phrase.
sentence type—a variety of sentence characterized by a par-
ticular sequence of elements. Palauan has sentence types
such as transitive sentence, intransitive sentence, ergative
sentence, etc.
sequential time relationship—relationship between two
clauses joined by the connecting word e such that the clause
introduced by e designates an event, state, etc. which takes
place after the event, state, etc. of the preceding clause.
shortening—a phonetic process in which a sequence of two
vowels (i.e., a long vowel or a vowel cluster) shortens to a
single vowel.
significant sounds of a language—units of sound which con-
trast with each other and serve to differentiate words from
each other. See also contrastive sounds, phonemes.
simple noun—a noun which consists of a single morpheme.
simple noun phrase—a noun phrase consisting of the inde-
pendent (or unpossessed) form of a noun.
simple state verb—a state verb which consists of a single mor-
pheme.
simultaneous time relationship—relationship between two
clauses joined by the connecting word e such that the
events, states, etc. of both clauses are happening at the
same time.
singular—referring to just one person, animal, or thing.

654
 Glossary of Terms

sound system of a language—the contrasting units of sound

(= phonemes) and their distribution.
source phrase—a relational phrase which indicates the point
of origin of some action involving movement or the person
from whom something is received, learned, etc.
source sentence—the sentence from which another sentence
is derived by applying grammatical processes such as
subject shifting, preposing of possessor, and the like. Some
source sentences are themselves grammatical, while other
source sentences must obligatorily undergo certain gram-
matical processes.
spatial relationship—a relationship between the locations of
two or more persons or things. Spatial relationships involve
such concepts as on top of vs. under, in front of vs. in back
of, between, etc.
specific noun—a noun which refers to a particular person,
animal, or thing; a noun which is associated with a par-
ticular occasion.
specifying clause—a possible cover term for purpose, in-
strument, means of transportation, and accompaniment
clauses, all of which have the common function of spec-
ifying, qualifying, or giving further information about the
action or activity of the preceding independent clause. Use
of this term, however, is restricted to any dependent clause
which fulfills the rather general function of specifying or
narrowing down the scope of the action or state of the
preceding independent clause, but which cannot be con-
veniently assigned to any of the four types of dependent
clauses mentioned above.
specifying word ȩr—a word which precedes certain types of
sentence objects that refer to specific persons or things.
statement—any sentence which provides or supplies infor-
mation by describing events, actions, states, etc.
state verb—a verb which describes a state, condition, or
quality which temporarily or permanently characterizes a
person, animal, or thing.
stop—a consonantal sound in which the outward air flow is
completely stopped or interrupted at some point in the
mouth or throat.

655
 Glossary of Terms

stressed syllable—the loudest or most prominent syllable in a

word; sometimes identified in this text with a stress mark (′).
stress mark—a mark (′) used to identify the stressed (i.e.,
loudest and strongest) syllable of a word, placed over the
(prominent) vowel of the stressed syllable.
subject shifting—a grammatical process in which the subject
noun phrase of a sentence is shifted to the right of the verb
phrase (or the second noun phrase, if the sentence is equa-
tional). When subject shifting occurs, a 3rd pers. sg. or pl.
non-emphatic pronoun is left in the original subject position.
suffix—any bound morpheme which is attached to the end of a
word or morpheme.
syllabic consonant—(nasal or liquid) consonant which is pro-
nounced as a separate syllable.
syllables—the pulses of air with which a given word is pro-
nounced.
tapped r—a pronunciation of r made with a quick tapping
movement of the tongue tip against the alveolar ridge.
temporal phrase—a relational phrase which names the time of
an action or state.
tense—the time of the action or state designated by a verb.
third person pronoun—a pronoun (‘he’, ‘she’, ‘it’, or ‘they’)
which refers to someone or something the speaker is inter-
ested in talking about.
time clause—a clause which is used to express a temporal
relationship between two events, actions, states, etc. Time
clauses are introduced by such expressions as er se ȩr a
‘when’, ȩr a uche ȩr a ‘before’, and ȩr a uriul ȩr a ‘after’.
time word (or expression)—any word (or expression) which
identifies a point or period of time in the present, past, or
future.
tongue advancement—the relative degree of advancement or
retraction (front, central, or back) of the tongue during the
pronunciation of a given vowel.
tongue height—the relative height (high, mid, or low) of the
tongue in the mouth during the pronunciation of a given
vowel.

656
 Glossary of Terms

transitive action verb—a verb which names an action or ac-

tivity that involves both a doer (or agent) and a receiver (or
object); a verb identifying an action which is done to or di-
rected at some person, animal, or thing.
transitive sentence—a sentence type consisting of a subject
noun phrase, a transitive verb, and an object noun phrase
(followed optionally by a relational phrase).
transitive state verb—a state verb (usually referring to a
mental state or ability) which involves both a subject and an
object.
trilled r—a series of two or three tapped r’s pronounced in
rapid succession; spelled rr and transcribed phonetically as
[r̄].
unaspirated—not involving aspiration—i.e., not accompanied
by an audible puff of air.
unmarked forms—forms which convey a particular meaning
in spite of the fact that they do not contain any overt mor-
pheme (or marker) for that meaning. For example, Palauan
present tense verb forms (e.g. mȩsuub ‘study’, mȩnguiu
‘read’, etc.) are identified by the absence of any overt tense
marker.
unpossessible noun—a noun which cannot take possessor suf-
fixes.
unrounded vowel—a vowel pronounced without any simulta-
neous rounding of the lips.
unstressed syllable—any syllable in a word other than the
loudest and strongest.
utterance—the act of saying (or uttering) a sentence.
velar—a consonantal sound involving the velum.
velum—the membrane behind the soft palate which serves as a
point of articulation for the Palauan consonants k and ng.
verb—a word (or part of speech) which describes an action or
state.
verb marker—a morpheme with various forms (mȩ-, -m-, -u-, -o-
, etc.) which functions to mark or identify a particular word
as a verb.

657
 Glossary of Terms

verb of communication—a verb which denotes the trans-

mission or reception of information—i.e., a verb involving
saying, telling, hearing, etc.
verb of mental activity—a verb which denotes such mental
processes as thinking, believing, knowing, and the like.
verb phrase—any single word or group of words which follows
the sentence subject and describes an action or state in
which the sentence subject is involved.
verb stem—a morpheme (sometimes occurring as an inde-
pendent noun) which combines with various affixes to form
different types of verbs.
vocal cords—a set of elastic membranes in the larynx which
can vibrate, producing voiced sounds.
voiced sound—a sound pronounced with a simultaneous vi-
bration of the vocal cords.
voiceless sound—a sound pronounced without any simulta-
neous vibration of the vocal cords.
voicing—the “buzzing” sound produced when air passes be-
tween the vocal cords and causes them to vibrate.
vowel—a sound produced by holding the tongue in various po-
sitions in the mouth without any contact or friction.
vowel blending—a phonetic process in which two vowels in an
unstressed syllable change or blend into a single vowel. The
resulting vowel preserves phonetic features from each of the
original vowels.
vowel cluster—a series of two adjacent vowels in the same syl-
lable, as in oách ‘leg’ and ngáu ‘fire’. Usually, one of the
vowels in a vowel cluster is stressed.
vowel cluster reduction—a phonetic process in which a vowel
cluster is reduced to a single vowel (or sometimes ȩ) in an
unstressed syllable.
vowel deletion—the loss of a vowel under certain circum-
stances (usually in unstressed syllables).
vowel-final (word or stem)—a word or stem ending in a vowel.
vowel-initial (word, stem, or suffix)—a word, stem, or suffix
beginning with a vowel.

658
 Glossary of Terms

vowel qualities—the different vowel sounds which are pro-

duced by changing the position of the tongue.
vowel reduction—a phonetic process, found in Palauan and
many other languages, in which a full vowel (i, u, e, o, or
a) reduces to the weaker, more neutral mid central vowel
schwa under certain conditions.
vowel triangle—a triangular configuration which schematizes
the relative positions of articulation of the vowels.
word-final(ly)—occurring at the end of a word.
word-initial(ly)—occurring at the beginning of a word.
word-internal(ly)—occurring within or inside a word.
yes-no question—a question which can be answered by ‘yes’
or ‘no’ and which asks whether or not such-and-such is the
case.
zero object pronoun—an object pronoun which has no actual
phonetic realization but which nevertheless is of structural
significance because it contrasts with the other object pro-
nouns. The zero object pronoun (symbolized as Ø) refers to
non-human plural objects.

659
 Bibliography

Bibliography
Carlson, Clayton H. 1967. “Lessons in Palauan” (revised edition).
Mimeographed. Honolulu: University of Hawaii Peace Corps
Training Center (PALI).

——1968. “Palauan phonology.” Unpublished M.A. thesis, University of

Hawaii.

Dyen, Isidore. 1971. Review of Die Palau-sprache und ihre Stellung zu

anderen indonesischen Sprachen, by Klaus Pätzold. Journal of the
Polynesian Society 80:2, 247–58.

Flora, Sister Marie Jo-Ann. 1969. “Analysis of the segmental phonemes

of Palauan.” Languages and Linguistics—A Student Review
(Georgetown University) 4:1, part 2, 1–30.

Greenberg, Stephen, Larry Hall, and Taddese Beyene. 1969. “The

Palauan verb.” Languages and Linguistics—A Student Review
(Georgetown University) 4:1, part 2, 68–90.

Mancill, Grace S. and Frances S. Woods. 1969. “Morphology and syntax

of Palauan nominals.” Languages and Linguistics—A Student
Review (Georgetown University) 4:1, part 2, 31–67.

Wilson, Helen I. 1972. “The Phonology and syntax of Palauan verb af-
fixes.” Working Papers in Linguistics (University of Hawaii) 4:5.

660