CHAPTER 8

8
Absolutives, ta-Participles and Infinitives; Introduction
to Internal Sandhi

There is a crucial difference between how Sanskrit and English use verbs. In a sentence that de-
scribes more than one action, English typically uses a chain of finite verbs:

She leaves the house, goes into the forest and sits down.
After she left the house and went into the forest, she sat down.

Sanskrit, on the other hand, frequently has no more than one finite verb in a sentence. Other verbs
appear in non-finite form, either as absolutives (also sometimes called gerunds) or as participles.
Literally translated, Sanskrit would say:

Having left(absolutive/participle) the house and having gone(absolutive/participle) into the forest,
she sat(finite verb) down.

This section will introduce you to the most frequent of these non-finite forms, the absolutive and
the ta-participle.

FORMATION AND MEANING OF THE ABSOLUTIVE AND THE TA-PARTICIPLE

Both of these are formed by adding a suffix to the zero-grade root of a verb. The suffix added in the
case of the absolutive of simple verbs is -tvā, that of the ta-participle is -ta-.

The absolutive of a verb ‘x’ means ‘having x-ed’. For example:

√द्रु ‘to run’ –› द्रुत्वा ‘having run’

√कृ ‘to do’ –› कृ त्वा ‘having done’
√गम् ‘to go’ –› गत्वा ‘having gone’ (on the zero grade of √गम् –› Chapter 7)

80
 (Absolutives of compound verbs, i.e. verbs that have a preverb (to be introduced
in Chapter 10), add -tya if they end in a short vowel, and -ya otherwise. For 8
example, अप-√द्रु ‘to run away’ ( अप- = ‘away’) –› अपद्त्य रु ‘having run away’; प्रति-
√लिख् ‘to write back’ (प्रति- = ‘back’) –› प्रतिलिख्य ‘having written back’.)

The ta-participle of a verb ‘x’ has a past passive meaning ‘having been x-ed’. For example:

√भृ ‘to carry’ –› भृत- ‘having been carried’

√कृ ‘to do’ –› कृ त- ‘having been done’

If a verb cannot sensibly form a passive, the ta-participle has active meaning:

√भू ‘to become/be’ –› भूत- ‘having become’

√गम् ‘to go’ –› गत- ‘having gone’

(Note how a passive expression such as +‘having been become’ would not make sense.)

ta-participles are a-stems and thus employ the endings introduced in Chapter 6 (Nom Sg -taḥ,
Acc Sg -tam etc.). They are adjectives in the sense that they agree in case, number and gender with
whatever noun they refer to (see the examples under ‘Use of participles and especially the ta-par-
ticiple’ below), and can be used to refer to any noun in a sentence. The absolutive, on the other
hand, is indeclinable, i.e. it does not add any ending after the -tvā. It can refer only to the agent of
a sentence, i.e. the subject of an active verb or the instrumental-case agent of a passive verb (again,
see ‘Use of participles and especially the ta-participle’ below on this use of the instrumental). Thus,
in e.g. गृहं गत्वा नरः विशति ‘having gone to the house, the man enters’, it is understood that the person
going to the house is the same as the person entering the house, the subject of the main verb.

Some Formal Irregularities

There are a number of ways in which the shape of absolutives and ta-participles can diverge from
the regular pattern just described. It is necessary to be aware of them so as to still recognise which
verb a particular form comes from.

-i-
Some verbal roots that end in a consonant add the vowel -i- before the suffixes -tvā, -tya or -ta-
(i.e. before suffixes beginning with a t-): √पत् ‘to fall, to fly’ –› पतित- and पतित्वा ‘having fallen’;

81
 √लिख् ‘to write’: लिखित- ‘having been written’, लिखित्वा ‘having written’. This -i- does not change
8 the meaning of these forms in any way, and one cannot predict from looking at a root whether the
-i- will appear or not. Across different times and literary genres, there are verbs that sometimes dis-
play the -i- and sometimes do not. It would be ideal to memorise whether and when an -i- does or
does not appear in any given verb; it is simplest just to be able to recognise the -i- whenever it does
appear (and thus to recognise that a form such as लिखित्वा does not come from a hitherto unknown
root +√लिखि, but simply from √लिख्).

The only predictability is found in causative stems and in roots that form Class X presents: these
regularly form their ta-participles in -ita, their absolutives in -ayitvā and their infinitives in -ayitum:
from √पत् ‘to fall,’ (I पतति) we find the causative पातयति ‘he causes (someone) to fall’, the ta-participle
पातित- ‘having been caused to fall’, the absolutive पातयित्वा ‘having caused to fall’ and the infinitive
पातयितुम् ‘to cause to fall’. Note that causatives keep their strengthened root vowels in these forms.

See also the Linguistic Note on seṭ/aniṭ in Chapter 21.

Unexpected Appearance of Guṇa

Some roots have guṇa (–› Chapter 7 on vowel grades) in both their absolutives and their ta-parti-
ciples (such as √रक्ष् ‘to protect’ –› रक्षित-; √पत् ‘to fall; fly’ –› पतित-). This cannot be predicted; be
prepared to recognise these forms (by looking at their suffixes and endings) in spite of this formal
exception.

BRIEF INTRODUCTION TO INTERNAL SANDHI

The unexpected presence of the -i- does not cause any noteworthy difficulties in identifying the verbal
root that a participle or absolutive is derived from. Yet whenever a root-final consonant comes to stand
directly next to another consonant (here, the initial t- of the suffixes), it may change, and this may make
the resulting form more difficult to recognise. This process of sounds changing according to their envi-
ronment is known as sandhi (Sanskrit सं धि-, literally ‘putting together’). We have already encountered it
in the change of -n- to -ṇ- in words that contain r, ṛ, ṝ or ṣ (–› Chapter 6).

This phenomenon may seem odd at first sight, yet consider how it occurs in English: the ending
-s (whether to mark the plural of a noun or the 3rd Sg of a verb) is pronounced either as voiceless
[s] if a voiceless sound precedes (he sinks, cats) or as a voiced [z] when a voiced sounds precedes
(he sings, dogs). The sound environment decides which exact form the ending takes, in English
just as in Sanskrit; the main difference is that English usually does not mark this in writing, while
Sanskrit always does. (Sometimes, even English reflects such pronunciation changes in its spelling:
the prefix ‘in-’ that negates words (inevitable, inadequate etc.) is changed both in pronunciation
82
and spelling to im- in words such as impossible or improbable. Technically put, the dental nasal n
turns into the labial nasal m in front of a labial stop (here, p or b).) 8
Sandhi is found both within words (‘internal sandhi’) and between words (‘external sandhi’). The
various forms of internal sandhi will be explained throughout this book whenever they become
relevant. External sandhi will be introduced in Chapters 11, 13 and 16.

Internal Sandhi before t

Much of internal consonant sandhi involves sounds changing and becoming easier to pronounce
together, often by becoming more similar to each other. t is a voiceless, unaspirated stop. If a ver-
bal root ends in a stop, that stop will also become voiceless or unaspirated when it is followed by
(voiceless, unaspirated) t: d changes into t, b changes into p etc. Palatal stops, furthermore, turn
into a velar – in this case, specifically, the voiceless unaspirated velar, k. Some examples:

√छिद् ‘cut’ –› छित्त्वा ‘having cut’ (with d > t)

√मुच् ‘to release’ –› मुक्त्वा ‘having released’ (with c > k)
√युज् ‘to join, link’ –› युक्त- ‘having been joined’ (with j > c > k)

Buddha Sandhi
There is one exception to this. When a root ends in a stop that is both voiced and aspirated, it does
not lose these qualities in front of a following t; instead, this voiced aspirate passes them on to the
t and thus turns the t into a dh. In that process, the original voiced aspirate loses its aspiration, and
so a combination of:

dh + t results in ddh
bh + t results in bdh
h + t results in gdh

What appears in Sanskrit as h was often originally a gh. Between vowels it appears as an h; yet before
a consonant it still behaves like the gh it once was; thus it also appears in this list. Some examples:

√बुध् ‘to awake’ –› बुद्

्ध वा ‘having woken (someone) up’
√लभ् ‘to take’ –› लब्ध्वा ‘having taken’
√दह् ‘to burn’ –› दग्ध- ‘having been burnt’

बुद्ध- is ‘having been awakened’ and thus ‘enlightened’; the Buddha literally is ‘the
awakened one’. The virāma in बुद्
्ध वा is solely there to keep the combination of -ddhv-
legible.

83
 ś plus t
8 There is one more situation in which both the stem-final consonant and the -t- of the -ta- change:
before t, root-final ś changes into ṣ. After this retroflex ṣ, t turns into the equivalent retroflex ṭ:

√दृश् ‘to see’ –› दृष्ट- ‘having been seen’

Nasal before t and Other Consonants

Before the dental stop t, nasals turn into the dental nasal n. A parallel phenomenon is found
word-internally before other stops: before a velar stop, the velar nasal ṅ appears; before a palatal
stop, the palatal nasal ñ, and so on. Before consonants that are not stops (y r l v, ś ṣ s, h), nasals
appear as anusvāra ṃ.

USE OF PARTICIPLES AND ESPECIALLY THE TA-PARTICIPLE

Participles are adjectives formed from verbal roots, and thus known as verbal adjectives (–› Chap-
ter 9 for details on adjectives in Sanskrit). They have three kinds of usage: 1) attributive (like an
adjective), 2) substantivised (like a noun), 3) instead of a finite verb.

1) Attributive
As was outlined in the Road Map on nominals (Chapter 5), adjectives, including participles, have
forms for all cases and all genders, and need to agree with whatever noun or pronoun they de-
scribe. Yet, as they are derived from verbs, participles such as e.g. दृष्ट- ‘having been seen’ often have
a more strongly verbal meaning than regular adjectives such as प्रिय- ‘dear’. This often means that a
Sanskrit participle is best translated into English as a relative clause:

दू तं मुक्तं पश्यामि not + ‘I see the having-been-freed messenger’ but ‘I see the
messenger who has been freed’
नरः मुक्तः ग्रामं गच्छति ‘the man who was freed (not +‘the having-been-freed man’)
goes to his village’

2) Substantivised
Participles, like adjectives in general, can be used on their own. In its masculine form, an adjective
like प्रिय- ‘dear’ would then mean ‘the/a dear man’, in its feminine form ‘the/a dear woman’ and in
the neuter ‘the/a dear thing’. If we change the above example to मुक्तां पश्यति it literally translates
as +‘he sees the having-been-freed feminine one’ –› ‘he sees the woman who has been freed’.

84
3) Instead of a Finite Verb
The ta-participle may also be used as the main verb of a clause. One can recognise that this is the 8
case when there is no finite verb present. Depending on whether the specific ta-­participle has
active or passive meaning, it needs to be translated as a past active or passive verb, respectively:

नराः पुरं गताः । ‘The men went to the city.’

पुरं दृष्टम् । ‘The city was seen.’

The agent of a passive verb form (whether finite (–› Chapter 21) or not) is expressed in the instru-
mental (or sometimes the genitive) case:

नरेण/नरस्य अश्वः दृष्टः । ‘The horse was seen by the man.’

नरैः/नराणां पुरं दृष्टम् । ‘The city was seen by the men.’
नरेण/नरस्य श्रुतः अश्वः द्रवति । ‘The horse that was heard by the man is running.’

All examples in this section used just one participle – the ta-participle – as that is the
only one that has already been introduced. Especially usages 1 and 2 are very common
also with all other participles, which are introduced in the following chapters: present
and future active (25), present middle and passive and future middle (22), perfect active
and middle/passive (35).

THE INFINITIVE
The Sanskrit infinitive, expressing ‘to do’, ‘to run’, ‘to eat’ etc., is formed by adding the suffix -tum
to the verbal root in guṇa: e.g. √नी –› नेतम ु ् ‘to lead’; √श्रु –› श्रोतुम् ‘to hear’. Infinitives display the
same formal variations as ta-participles and absolutives do: a number of roots insert an -i- before
the -tum (such as √पत् –› पतितुम् ‘to fall’). In verbs that do not insert the -i-, there is sandhi of the
root-final consonant (identical to that before -ta- or -tvā; –› √बुध् –› बोद्म धु ् ‘to understand’). As de-
scribed above, m changes into n before t-, giving e.g. गन्तुम् ‘to go’ from √गम्. There is some overlap
between verbs that add this -i- in their ta-participle and absolutive and verbs that add the -i- in
their infinitive; yet the two groups are not identical (–› e.g. √भू –› भवितुम् ‘to be’, but भूत- and भूत्वा
‘having been’). And again, a number of verbs have different infinitives in different times or literary
genres (e.g. √शुच् ‘to mourn’ has both शोचितुम् and शोक्तुम्).

85
 Infinitives of Class X verbs and causatives are formed from their present stem minus the final -a-:
8 चिन्तयति ‘he thinks’ –› चिन्तयितुम् ‘to think’; भारयति ‘he causes to carry’ –› भारयितुम् ‘to cause to carry’.

Infinitives are used after verbs expressing ‘to want/to hope/to intend/to go’ (etc.) to do something:

नगरं गन्तुम् इच्छामि ‘I want to go to the city’

THE NUTSHELL
ta-Participle Absolutive Infinitive
Form zero-grade root + -ta- + ending zero-grade root + -tvā guṇa root + -tum
Meaning ‘having been x-ed’ or ‘having x-ed’ ‘having x-ed’ ‘to x’
Examples भृत- ‘having been carried’ पतित्वा ‘having fallen’ भवितुम् ‘to be’
गत- ‘having gone’
(Absolutives of compound verbs: -tya after short vowels, -ya elsewhere:
अपद्त्य
रु ‘having run away’ –› Chapter 10.)
-ta-, -tvā, -tya, -tum: -i- may appear before these without effecting any change in meaning.
Internal sandhi: adjacent sounds change and become easier to pronounce together.
Stops lose either voice or Buddha sandhi
­aspiration. of voiced aspirates:
Before t: Palatal stops become velars: -ś + t- –› -ṣṭ- (g)h/dh/bh + t –›
त्यक्त-. -gdh-/-ddh-/-bdh-
m > n: गन्तुम्.

VOCABULARY

Note: The ta-participles and absolutives of the verbs√कृ , √श्रु and √वच् are introduced here as they
are used very frequently. These verbs belong to classes not yet introduced; thus, ignore the question
of how to form any of their other forms for the time being.

√दह् (I दहति) ‘to burn’

√जीव् (I जीवति) ‘to live, be alive’
√वस् (I वसति) ‘to live (in a place)’

86