CAMB RID GE LANGUAGE TEA C HI N G LIBR ARY

A series covering cenrral issues in language teaching and learning, by a urhors who
have expert knowledge in their field.

In this series:
Conversation: Frotn
Affect in Language Learning edited by j ane Arnold
Approaches and Methods in Language Teaching second edition by Jack C. Description to Pedagogy
Richards and Theodore S. Rodgers
Beyond Training by jack C. Richards
Classroom Decisio n-Making edited by Michael Breen and Attdrew Little;ohn
Collaborative Action Research for English Language Teachers by Am1e Burns
Collaborative Language Learning and Teaching edited by David Nunan
Communicative Language Teaching by William Littlewood
Developing Reading Skills by Fram;oise Grellet
Developments in English for Specific Purposes by Tony Dudley-Evans and Maggie
jo Stjohn
Discourse Analysis for Language Teachers by Michael McCarthy
Discourse and Language Education by Evelyn Hatch
Scott Thornbury and
The Dynamics of the Language Classroom by Ian Tudor
English for Academic Purposes by R. R. Jordan
English for Specific Purposes by Tom Hutchinson cmd Alan Waters
Diana Slade
Establishing Self-Access by David Gardner and Lindsay Miller
Foreign and Second Language Learning by William Littlewood
Group D ynamics in the Language Classroom by Zolttin Dornyei and Tim
Murphey
Language Learning in Distance Education by Cynthia White
Language Learning in Intercultural Perspective edited by Michael Byram and
Michael Fleming
The Language Teaching Matrix by Jack C. Richards
Language Test Construction and Evaluation by ]. Charles Alderson, Caroline
Clapham and Dianne Wall
Learner-Centredness as Language Education by Ian Tudor
Learners' Stories: Difference and Diversity in Language Teaching edited by Phil
Benson and David N unan
Managing Curricular Innovation by Numa Markee
Materials Development in Language Teaching edited by Bria11 Tomlinson
Morivational Srcategies in the Language Classroom by Zoltan Domyei
Psychology for Language Teachers by Marion Williams and Robert L. Burden
Research Methods in Language Learning by David Nunan
Rules, Patterns and Words: Grammar and Lexis in English Language Teaching by
Dave Willis
Second Language Teacher Education edited by jack C. Richards and David Ntman
Society and the Language Classroom edited by Hywel Coleman
Task-Based Language Teaching by David N tman
Teaching Languages to Young Learners by Lynne Cameron
Teacher Learning in Language Teaching edited by Donald Freeman and jack C.
Richards
Testing fo r Language Teachers second edition by A1'thur Hughes
Understanding Research in Second Language Learning by james Dean Brown
Using Surveys i.n Language Programs by james Dean Brown
Vocabulary: Description, Acquisition and Pedagogy edited by Norbert Schmitt and ~ CAMBRIDGE
Michael McCarthy ~ UNIVERSITY PRESS
Vocabulary, Semantics and Language Education by Evelyn Hatch and Cheryl
Brown
Voices from the Language Classroom edited by Kathleen M. Bailey and David
Nunan
 Contents
~MBR I D GE UNIVERS ITY PR ESS
Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, Sao Pau lo

Cambridge University Press

The Edinburgh Building, Cambridge CB2 2RU, UK

www.cambridge.org
Information on this title: www.cambridge.org/9780521891165 Vl) l
Thanks and acknowledgements
© Cambridge University Press 2006
Introduction 1
This publication is in copyright. Subject to statutory exception
and ro the provisions of releva nt collective licensing agreements, 1 Characterizing conversation 5
no reproduction of any part may rake place without the written Introduction 5
permission o f Cambridge University Press. T he nature of conversation 5
1.1
1.2 Approaches to the analysis of conversation 27
First published 2006 37
Summary
Printed in the United Kingdom at the Universit}' Press, Ca mbridge
2 The vocabulary of conversation 40
A catalogue record for this publication is available from the B1·itish Library Introd uction 40
2.1 Lexical size 42
Library of Congress Cataloging-in-Publicatio11 Data 2.2 Lexical density and lexical va riety 43
Tho rnbury, Scott, 1950- 2.3 Lexical frequency 45
Conversarion : from description to pedagogy I Scott Thornbury and 2.4 Lex ical repe ti tion 49
Diana Slade. 2.5 Vague language 54
p. em.- (Cambridge language reaching library) 2.6 Fillers 56
Includes bibliographical references and index. 57
ISBN 0-52 1-89116-7 (pbk.: alk. paper) - ISBN 0-521-81426-X
2.7 Discourse markers and other inserts
2.8 Routines and lexical phrases 62
(hardback: alk. pa per) 65
1. Conversation ana lysis. 2. Discourse analysis. 3. Language and 2.9 Appraisal and involvement
languages-Study and teaching. 4. Communicative competence. I. Title. 2.10 Implications 69
II. Series.
3 The grammar of conversation 73
P95.45.T49 2006 73
371.102 '2-dc22 2006023282 Introduction
3.1 Complexity 75
lSBN-13 978-0-521-81426-3 hardback 3.2 Heads and tails 80
lSBN-10 0-521-81426-X hardback 3.3 Gra mmatical incompletion 83
ISBN-13 978-0-52 1-891 16-5 paperback 3.4 Ellipsis 83
ISBN-10 0-521 -89 116-7 paperback 3.5 Dei xis 85
The publisher has used its besr endeavours ro ensure tha t the URLs for external 3.6 Questions 86
websites referred to in rhis book are correct and active at th e rime of go ing to 3.7 Tense and aspect 90
press. However, the publisher ha s no responsibility for the websites and can make 3.8 Modality 94
no guarantee that a site will remain live or that the content is or will rema in 3.9 Reporting 98
a ppropriare. 3.10 What do lea rners need to know? 100
v
 Conte1'zts Contents

(j) The discourse features of conversation

Introduction
107
107
7.4
7.5
Pragmatic competence
Transfer
223
224
4 .1 Cohesion in conversation 108 7.6 Acquisition vs learning 230
4.2 Inreraction in conversation 113 7.7 Classroom talk 238
4.3 Topic management: Topic development, topic
change and topic choice 127 ,l /' Teaching conversation: A history 247
4.4 Discourse strategies 130 Introduction 24 7
Summary 137 8.1 Pre-reform and reform 247

y
..
Genres in conservation: Storytelling and gossiping
Introduction
142
142
8.2
8.3
Direct method: Learning-through-conversation
Audiolingualism: Drills, dia logues and the conversation
class
249

251
5.1 Chat a nd Chunks in conversation 142 8.4 Situational English: Conversation in context 252
5.2 Genre theory 145 8.5 Oral English: Conversation as speaking practice 254
5 .3 Storytelling genres 151 8.6 CLT: Conversation as communication 255
5.4 Lexica-gra mmatical fea tures of storytelling genres 159 8.7 Task-based learning: Conversation as a task 266
5 .5 Storytelling genres: Summary 168
5.6 Gossip 170 9 Teaching conversation: Approach, design, procedure
5.7 Lexica-grammatical features of gossip 177 and process 274
5.8 Gossip genre 180 Inrroduction 274
5.9 Classroom implications . 180 9.1 Approach 275
Summary 182 9.2 Design 281
9.3 Procedure 295
6 Acquiring L1 conversational competence 186 9.4 Process 307
Introduction ' 186 Conclusion 318
6.1 Conversational competence 186
6.2 Turn taking 188 Task key 326
6.3 Child-directed speech 190 References 342
6.4 Formulaic language 192 Author index 358
6.5 Repetition 194 Subject index 361
6.6 Scaffold ing 196
6.7 Syntax: Vertical constructions 197
6.8 Cohesion 198
6.9 Coherence 199
6.10 Functions, genres and speech acts 200
6.11 Pragmatics 203
6.12 Educated discourse: Talk at school 204
6.13 Sociocu ltural theory and 'instructional conversation' 206

7 Acquiring L2 conversational competence 214

Introduction 214
7.1 Fluency 214
7.2 Formulaic language 218
7.3 Communication strategies 219

Vl Vll
Introduction

Casual conversation is a fundamental human activity, and one in which

most of us engage many times a day. It may take the form of small talk
about the weather at the supermarket check-out, or gossip about col-
leagues around the office coffee machine, or an extended phone conver-
sation with a close friend about the meaning of life. Before getting down
to the business at hand, sales reps chat with their clients, doctors chat
with their patients, waiters with diners, and teachers with their students.
Strangers at a bus stop will start up a conversation to vent their frustra-
tion about the service. Taxi drivers famously air their opinions, seldom
solicited. Your dentist will chat away even when your responses are
reduced to grunts. Fellow passengers on a long-haul flight will exchange
pleasantries before settling in to watch the movie. Listeners will phone a
radio talk show to sound off about local crime, and teenagers will talk
for hours on their cell phones about matters of apparently enormous
consequence. Even very young children chat away with their parents,
and by the age of three are able to have fairly sustained conversations
with their playmates.
Conversational talk crosses age gro ups, gendet~ class, culture and eth-
nicity. Levelt (1989) calls it ' the canonical setting for speech in all human
societies'. Indeed, the stylistic features of conversation have extended
beyond spoken talk itself and 'crossed over' into other modes and media,
such as the popular press and advertising, a process called conversa-
tionalization by Fairclough ( 1992). And the advent and rapid expansion
of the use of email, text messaging and online chat have further blurred
the distinction between spoken and written language, while underscor-
ing the ubiquitous role of conversation in human affairs.
The centrality of conversation to human discourse owes t o the fact
that it is the primary location for the enactment of social values and rela-
tionships. Through talk we establish, maintain and modify our social
identities. The role that conversation plays in our formation as social
beings starts at an early age. Stubbs (1983: IX} asserts that ' infants learn,
as it were, to engage in conversation before they learn language', and
Hatch (1978: 404) claims that 'language learning evolves out of learning
how to carry on conversations, out of learning how to communicate'.
Even as far back as the 1930s, H arold Palmer argued that all language

1
Introduction Introduction

use is based on, and is an extension of, conversation, adding that con- • ro outline how first-language conversa tional competence develops,
versation must therefore be the start of any study of language. In and to relate this research to the development of second-language con-
Palmer's day, this meant prioritizing the teaching of pronunciation. The versational competence;
nature of spoken language itself was barely understood and for a long • to identify and analyse the kinds of difficulties that learners of English
time spoken language was taught as if it were simply a less formal version encounter when participating in conversation;
of written language. This is a view that has been rectified only recently, • to outline a range of methodological approaches, procedures and
with the advent of corpus linguistics and the consequent amassing of techniques for teaching English conversation and to illustrate these
corpora of spoken data. Findings from such data now heavily inform the approaches by reference to current materials;
content of learner dictionaries, such as the Cambridge Advanced • and, finally, to argue for a n interactive, 'integrated' model of instruc-
Learner's Dictionary (second edition 2005), and descriptive grammars, tion, informed by the descriptio n of conversation and the learning
such as the Longman Grammar of Spoken and Written English (Biber et theories outlined in the preceding chapters.
al., 1999).
Finally, sociocultural theories of learning, such as those that derive from
A note on transcription conventions
Vygotsky's research into children's cognitive development, foreground the
role of conversation as the medium for all learning, arid have contributed Wherever possible the data used as examples in this book come from
to the notion that effective teaching is, essentially, a ' long conversation' authentic sources, i.e. from spontaneous and naturally occurring conver-
(Mercer, 1995). Recent research into second-language acquisition also sations recorded in a variety of contexts. (The few instances of invented
supports the view that the learning of second languages may be success- data a re identified as s uch.) In transcribing these conversations we have
fully mediated through conversational interaction (van Lier, 1996). Such tried to capture their spontaneity and informality, but n ot at the expense
a view not only reinforces the arguments for an approach to language of their readability. This has sometimes meant ignoring the finer details
teaching that systematically deals with spoken English, but would seem ro of transcription, such as length of pauses, pitch direction and other para-
vindicate the intuitions of those legions of learners who consistently linguistic phenomena, unless these features have been expressly singled
demand inclusion of more 'conversation' in their language courses, our for discussio n. In cases where w e cite data tha t employ different tran-
For all these reasons, an account of how conversation works is' there- scription conventions from our own, we have modified these transcrip-
fore essential in the development of a p edagogy for second-language tions so as to bring them into line. Where this has not been possible, an
learning. This book aims to meet this need by providing the reader with explanation of any variant conventions wiU be found alongside the data.
first an overview of the features that characterize conversation and dis-
tinguish it from other spoken and written genres (Chapter 1 ), followed The transcription devices that we use are the fo llowing:
by a systematic description of conversational English, including its
vocabulary (Chapter 2), its grammar (Chapter 3), its discourse structure • full stops: these indicate completion, usually realized by falling into -
(Chapter 4), and its characteristic generic patterning (Chapter 5), and nation
then an informed account of irs development in both first- and second- • commas: these are used to separate phrases or clauses in order to make
language acquisition (Chapters 6 and 7). On this basis, and after a review utterances more readable
of teaching approaches to date (Chapter 8), an integrated approach to • question marks: these are used to indicate utterances that, in- their
the teaching of conversation will be outlined, along with practical class- context, function as questions, irrespective of their grammatical form
room applications (Chapter 9). or their intonation
In short, the book aims: • exclamation marks: these are used conservatively to indicate the
expression of s urprise or shock
• to introduce practising teachers to the na ture and structure of conver- • capital letters: words in capital letters are used conservatively to indi-
sation in English, drawing from a range of theoretical models; cate emphasis
• to equip readers w ith analytical techniques necessary to analyse • quotation marks: double quotation marks are used to signal that the
authentic conversation at the level of vocabulary, grammar, discourse speaker is directly quoting speech; s[ngle quotation marks are used to
and genre; · signal that the speaker is saying what they or someone else thought

2 3
Introduction

• empty parentheses: non-transcribable segments of talk are indicated 1 Characterizing conversation

by ()
• filled parentheses: words within parentheses indicate the transcriber's
best guess as to a doubtful utterance
• square brackets: information about relevant non-verbal behaviour is
given within square brackets [ )
• dots: three dots indicate a hesitation within an utterance: .. .
• dash: a dash represents a false start:
Speaker: Did you ever get that- I mean in French what is it? Introduction
• equals sign: a double equals sign is used to represent overlap p he-
nomena, such as Conversation accounts for the major proportion of most people's daily
o simult aneous utterances, i.e. where two speakers are speaking at language use but despite this (or perhaps because of it) it is not that easily
the same time: defined. Compare, for example, these three dictionary definitions:
Speaker 1: Is it still going, Studebakers? • If you have a conversation with someone, you talk w ith them, usually
Speaker 2: = =I don't know in an iniQ.r~_l situation (Collins' CO 8 Ul LD English Dictionary).
Speaker 3: = = No it's got a new name • Informal talk in which people exchange news, feelings, and thoughts
o overlapping u tteran ces: the point where the second speaker begins CLo;:;g;zan Dictionary of Contemporary English).
talking is shown by= = preceding the point in the first speaker's
• An inf~rmal talk involving a sma ll group of e~?ple or only two;
turn: the acnvity of talking in this way (Oxford A(i.vanced Leamer's
Speaker 1: Can you dance now= = Rod, can you? Dictionary).
Speaker 2: = =I can do rock'n' roll and Cha Cha and Rumbas
and Sambas and waltzes While all three definitions highlight the informal and the spoken nature
o contiguous utterances: i.e. when there is no interval between adja- of conversatio n, only o ne singles out group size as a defining feature,
1 while another focuses on topic. The distinction between a conversation
cent utterances produced by different speakers:
Speaker 1: they had to move out of the flat because the whole= = (i.e. conversation as a countable noun) and conversation (uncountable)
Speaker 2: = =roof collapsed. is either ignored or blurred in the first two definitions. Finer distinctions
between conversation and, say, chat, small talk, discussion and gossip,
are not dealt with. And, as we shall see in Chapter 8, the term conver-
sation with special reference to language-teaching methodology has
been enlisted for a wide variety of uses - ranging from speaking and
communication to dialogue and role play. In this chapter we shall
attempt to characterize conversation, first by contrasting it with other
kinds of language, and then by listing its distinguishing features. By way
of conclusion, we w ill offer a working definition of conversation that
will serve as the starting point for a more detailed description in subse-
quent chapters.

1.1 The nature of conversation

In April 1999 a freak storm devastated parts of the city of Sydney. H ere
is how the storm was reported in The Sydney Morning Herald the
following day:

4 5
Characterizing conversation The nature of conversation

(6) SS: The time mar we realised tha t it was heading for the
Text 1.1 city ...
Hail shatters c ity (Radio 2BL, Philip Clark Breakfast P resenter, 15 Apri11999)
A freak hai l sto rm swept across Sydney last night, ca using damage A couple of days later four friends were talking about how they were
worth hundreds o f millions of dollars and rriggering a massive
affected by the storm. Here is the transcript of part of that conversation:
rescue and repair effort by emerge ncy services.
Thousands of homes were damaged as roofs caved in and windows Text 1.3: Hailstorm
and skylights were smashed. Thousands more cars were wrec ked or
badly damaged in the storm, w hich struck with no official warning. (1) Odile: ... No I think 1 don't know many people w ho have
been affected except yo u and I. That much.
The ambulance service said dozens of people were treated for cuts (2) Rob: You don't know?
and lacerations after being hit by fa lling glass or hail scones, w hich (3) O dile: Well you know except for the neighbours.
witnesses described va rio usly as being as big as golf balls, lemons, (4) Rob: O h a friend of ours in Paddington, they had to move
cricket balls and rock melons. ou t of the flat = =
... At Paddington, Ms Jan Maurice said all houses on one side (5) Grace: = =tvlm.
of Prospect Street had windows smashed. M r Luc io Gallero, of (6) Ro b because the whole = =
Lucio's Restaurant at Paddingron, said: 'I had five windows in the (7) Grace: = =roof collapsed.
restaurant smashed. Water flooded in and patro ns' cars have been (8) Rob: T he tiles fell through the ceiling= =
smash ed.' (9) Grace: = =Mm
(10) Rob: into the room and they've actually had to move out
(The Sydney Morning Herald, 15 April 1999)
completely.
(11 ) O dile: Oh really?
On the day after the storm a radio talk show host interviewed a And there was the lirde old lady over the road who ...
(12) Dan:
spokesman from the Weather Bureau: (13) Rob: O h yea h. [laughs] She was sitting in her living room
and a hail scone fell through the skylight, this o ld
Text 1.2 Ira lia n woma n. She had corrugated iron but it fe ll
(1 ) PC : ... here on 2BL. Well what wenr wrong? Why didn't through the skylight. It fell through the ceil ing and
the Weather Burea u tell us w hat was happening? You landed in her la p when she was sitting= =
have heard earlier this morning reporrs that the Bureau (14) Odile: == Mm.
tho ught er saw the storm but tho ught it would go back (15) Rob: watching television.
out to sea. It didn 't. Steve Simons, a senior forecaster (16) Dan: Watching The X -(7les probably.
with t he Bureau, joi ns me o n the line this morning. (17) All: [la ugh]
Good morning Steve. ( 18) Odile: I'm so glad the kids were not there because you
(2) SS: Good morning Philip. know that hole is just above Debbie's head.
(3) PC: So what went wrong? (19) Rob: Yea h.
(4) SS: What went wrong was that the storm developed down (20) Grace: Oh yeah.
near Wollongong and we had it on the radar and we were (21) Rob: No, it is amazing more people wer en't injured.
tracking ir and the track at that stage was showing it (22) Grace: Mm.
going out to sea and rhen very suddenly it developed into (23) Rob: So erm they go back to school tomorrow?
what we call a 'supercell' which is the beginning of a (24) Odile: Not romorrow = =
severe thunderstorm and these supercells ha ve a habit of (25) Rob: == Monday.
doing some rather crazy things. It changed direction very (26) Odile: It's Sunday.
sudde nly - this was down nea r Orford Bundeena way = = (27) R ob: Monday.
(5) PC: = =Yes all right so er what was the rime interval (28) Grace: Monday.
between you first discovering this storm and then (29) Odile: Monday.
di scovering that it was in face heading for the the city? (30) Rob: Mm.

6 7
Characterizing conversation The nature of conversation

(31} Odile: Yeah. language, i.e. sentence stress, int?nation, :e~po and articulation ra t~,
(32) G race: Is the school OK? rhythm and voice quality, ~re _lost 111 transcnptt~n. In orde: to redress th1s
(3J) O dile: You mean, ge neral damage? omission, here is a transcnptton of Text 1.3 w1th prosod1c features rep-
{34 } Grace: Yeah. resented, using the system adopted by Crystal and Davy (1975), as out-
(35} Odile: I don't know. lined in the glossary below:
(36} Rob: The school's closed next to us, yea h.
{37) Grace: I was spea king to erm ... II tone-unit boundary
(38) O dile: O h my god I hadn't thought about tha t ... 1 first p rominent syllable of the tone-unit
. (Authors' dara) falling tone
rising tone
Each of these three texts deals with the same topic - the storm- but each level tone
deals with it in a very different way. These differences derive partly from rising-falling tone
the d ifferent channels oi g>tnm\J.n.li:ano~ l~~vgl~~d~ partly fm.rn. th~ dif- falling-rising tone
ferent_p.UL[LQ_se~ thatmotivared each text, ajl"'ci pagh:Jro!!!....th~J:J.if(~rent the next syllable is stressed
kinds of rol~~- and relatwnshtps existing in each oLthe C91J1.1J'l.lJI\i.~ arive i the next syllable is stressed and a lso steps up in pitch
s-itua tiqw;~ While all three texts encode instances of spoken language extra strong stress
(Text 1-1 ljoth report~~nd,..s!_i r~~!¥_J.Uotes w hat witnesses are supposed SMALL CAPITALS the word, or words, containing rhe nuclear
to have said), only Texts ~_J.,and2.J;exhibit the 'jo intJy-constr~cted-in­ syllable in a tone-unit
real-time' nature of talk, and only one of these texts - Text 1.3- IS a c;_o!'l-
v!!_!sat£on in"tfie-sense"that we will be using in this book. pauses, from brief to long
In order to arrive at a workable definition o f conversation, then, it will
be useful to look at the differences between these three texts in more
detail. By highlighting the differences, first between written and:spoken
English, a nd then between formal and info rmal spoken English; the fo l- Text 1.3 - Phonological transcription
lowing defining characteristics of conversation, and their implications,
(1} Odile: .. . \no l 'think II I don 't !know imany 'people 'who
will be discussed:
have been AFFECTED II except !you and il ll
• that (to state the obvious) it is spoken, and ITHAT 'much II---
• that this speaking takes place spontaneously, in real time, and (2} Rob: · you !do n't KNOW II
• that it takes place in a shared context; (3) Odile: !WELL you KN6W II ex'cept for the
• that it is interactive, hence jointly constructed and reciprocal; iNEIGH BOURS II
• that its function is primarily interpersonal; (4 ) Rob: oh a i friend of 'ours in PWDTNGTON ll lthey 'had
to ' move 'out of the iFLAT II
• that it is informal; and
(5} Grace: IM' MI\
• since, it is the critical site for the negotiatio n of social identities, it is
(6) Rob: belcause the WH6 LE II
expressive of our wishes, feelings, attitudes and judgements. !roof COLLAPSED II
{7) Grace:
(8) Rob: the i riles 'fell through the CEILING II
1. 1. 1 Conversation is spoken (9) Grace: IM'M ll)
(10} Rob: \into the iROOM II and they've lactually had to
Conversation is spoken (or primarily so, since computer-mediated com- ' move 'out COMPLETELY II·
munication now allows conversation to take p lace by means o f writing- {11 ) Odile: oh !R EALLY 1\
see Section 1.1.8 below). Hence the most obvious difference between (12) Dan: and !there was the little old 'lady over the R6AD
Texts 1.1, 1.2 and 1.3 lies in the choice of mode: Text 1.1 is - and was who 11 -
always- written, whereas Texts 1.2 and 1.3 are written transcriptions of {13 ) Rob: loh YEAH II [laughs]lshe was 'sitting in her
what was originally spoken. The transfer from one mode (speaking) to L1VING 'room 1\ a nd a !hail stone 'fell through the
another (writing) means that most of the prosodic features of the spoken SKYLIGHT II this \o ld ITALIAN 'woman 1\ !she had

8 9
Characterizing conversation The nature of conversation

'co~ruga ted IRON II but it 'fell through the II'm i so 'glad the iKlDS were nor TH~REII belcause you KN6WII
SKYLIGHT II it 'fell through the 'ceilin&_ and that IH6LEII is 'just above 'Debbie's HEAD<j[
'landed in her · iLAP lllwhen she was SITTING II
(14) Odile: IM-MII She uses a falling_t;_qne O!:!.'~ ids.'..to int£QQ_t,t~!'! .a ne.~ _ topjc (or to 'pro-
(15) Rob: !watching TELEvtSION 11--- claim' it, in Brazil's (1997) terminology), and a risi~!h or)~!~.rr\!~.s.:
(16) Dan: lwatching the iX-FILES II PROBABLY II rone, on 'th~re' and o~ 'h_gk' t_Q_~efer_lQ yy_b._~tiH.QnJ.JnQn gro und. The
(17) All: [laugh] other speakers have already been shown the h ole, a fact that lS sug-
(18} Odile: ll' m iso 'glad the iKIDS were not THERE II oested by the deictic expressions ' there' and ' that' which assume a
belcause you KN6W II that IH6LE II is 'just above ~hared perspective, not to mention the explicit reference to shared
'Debbie's HEAD II knowledge in the expression 'you know'. On the other ha nd, the new
(19) Rob: IYEAHII information about the proximity of the hole to Debbie's head is 'pro-
(20} Grace: loh YEAH II
Ino it its a'mazing II more lpeople weren't INJURED II claimed' using a falling tone.
(21) Rob:
(22) Grace: IMM II--- • The use of high_~~y tq ma!f!:!~i n a sp_~ak in.~ tu!n, contrasted with a fall
(23) Rob: IS6 erm llltbey go back to 'school TOM6RROW II ro low key as the speaker prepares to relinquish the turn, as in R o b's
(24} Odile: not !TOMORROW II turn 10.
(25) Rob: MONDAY II • The use of high key to signal 'high invo lvement', as in Od ile's turn 38.
(26) Odile it's \SUNDAY II
(27) Rob: IMONDAYII The e:o..-tract demonstrates wha t Dalton and Seidlhofer (1994: 89) call the
(28) Grace: IMONDAYll 'crucial and all-pervasive' role that intonation - and key in particular -
(29) Odile: IMONDAY II plays in conversation management, influencing the management of
(30) Rob: IMMII copies and of turns, the identification of information status and the sig-
(3 1) Odile: IYEAH 11-- nalling of degree of speaker involve ment. One has o nly to imagine a con-
(32) Grace: is the lschool 6K II versation between two Daleks (the robotic characters in Dr Who, wh o
(33} Odile: lyou MEAN lllgeneral DAMAGE II speak in an uninflected monotone) t o appreciate the importance of these
(34) Grace: IYEAHII
prosodic features, and how they are implicated bo th in the interactive
(35) Odile: IY don't 'know II
(3 6) Rob: the !SCHOOL'S 'closed lllnexr to Os II IYEAH II natu re of conversation, and its interpersonal function.
(37) Grace: II was SPEAKING to erm II
(38} Odile: ioh my GOD II I hadn't iTH6UGHT about 'that II
1. 1.2 Conversation happens in real time
It would be impossible to convey the full extent of the conversational ' I had .five windows in the restaurant smashed. Water fl ooded in
'work' that is ac hieved through prosody, but among the features that are and patrons' cars ha ve been smashed.'
worth noting in the above extract - and which are either completely
absem or only notionally represented in w ritten text (e.g. by the use of Notice how in the newspaper article even the quoted speech fo llows the
punctuation) - are the following: conventions of written language, in that each sentence forms a complete
entity, consisting of cla uses that combine a single subject and its predi-
• The use of intonation (i.e. changes in pitch direction), and specifically cate in wa ys tha t do not deviate from the norms of written grammar.
a rising tone to signal questions, where no other grammatical markers Moreover, there are no erms or ahs o r false starts and back-trackiogs.
of interrogation are present, as in Rob's utterances (2) and (23); Compare this to:
• T he use of high ' key' - i.e. a marked step up in pitch - to indicate
thve introduction of a new topic: (4) oh a ifriend of 'ours in (4) Rob: Oh a friend of ours in Paddington, they had ro move
PADDINGTON II; out of the flat= =
• The way intonation is used to contrast informa tion that is considered (5) Grace: = = lvlm.
to be shared by the speakers ('given') and that which is being pro- (6) Rob because the whole= =
claimed as ' new', for example, in Odi!e's utterance (18): (7} Grace: = = roof collapsed.

11
10
 Characterizing conversation The nature of conversation

Even without the addition o f prosodic features, this is clearly transcribed that talk is built up clause by clause, and phrase by phrase, rather than
speech. It conveys the sense of being locally planned in real time. sentence by sentence, as is the case w ith written text (see Chapter 3). This
Compare this to the news article, where the production process has been explains why utterance boundaries are less clearly defined in spoken lan-
elaborated through several stages of drafting, re-drafti ng, editing and guage, and why coordination is preferred to subordination (the use of
publication.f'The m.~.inJactor which disti nguishes written from spoken subordinate claus~s) . Spok.e.n_l~gl,l~ge cons1sts qfJ.Le.qJ.Lent s~R.u.ences of
l language ... is time' (Crystal and Davy, 1975: 87).1The real-time spon-
taneity of talk accounts for a number of features tlrl.t distinguish it from
short cl~uses_ j.oin.ed by and, but, then, becaus.e. For example:
what went wrong was that the storm developed down near
\ writing. The most obvious of these a re 'dysfluency' effects, which occur Wol longong and we had it on the radar and we were tracking it
{ 'when the need to keep talking . . . threatens to ru n ahead of mental and the track at that stage was showing it going out to sea and
I planning' (Biber eta!., 1999: 1048). Texts 1.2 and 1.3 include several then very suddenly it developed into what we call a 'supercell'
instances of such dysfluency: . which is the beginning of a severe thunderstorm and these
supercells have a habit of doing some rather crazy things
• hesitations: So erm they go back to school tomorrow?
• word repetition: it was in fact heading for the the city The ' layering' of phrase on phrase, and of clause on clause, allows for
• false starts: No I think I don't know many people who . .. a looser form of utterance construction than in written sentences, with
• repairs: the Bureau thought er saw the storm their canonical subject-verb- object structure. Thus, in order t o fore-
• unfinished utterances: they had to move out of the flat because the ground the theme of an utterance (i.e. the point of departure of the
whole [. .. ] message), information in the form of a noun phrase can be placed at
• ungra mmaticaliry (in terms of written norms, at least): except you the head of the utterance, in advance of the syntactic subject: a friend
and I of ours in Paddington, they had to move out. Likewise, retrospective
Other devices that 'buy' planning time, and thereby help avert the more comments can occupy a tai[slot that does not exist in written sentences:
distracting effects of dysfluency, include the use o f fillers (as in: ,Well you I don't know many people who have been affected except you and I.
know except for the neighboU?·s), and the repetition of sentence frames That much.
(but it fell through the skylight it fell through the ceiling . .. ). R\~petition Another characteristic of spoken language which is attributable to its
may also take the form of 'borrowing' chunks of the previous speaker's spontaneity is the fact that information is relatively loosely packed. One
utterance, as in Text 1.2: measure of this density is the proportion o f content words (such as nouns
and verbs) per clause. Spqken texts are not as..le.xically de.nse as written
PC: So what went wrong?
t~~ts. So, for example, in Text J.l above, of the 142 words in all, 88 are
SS: What went wrong was rhat the storm developed down near
Wollongong (... ) lexical words - that is nouns, verbs, adjectives, and - ly adverbs - giving
a lexical density (Halliday, 1985) figure of 62 per cent. In the spoken
More generally, it is now thought that a great deal of spoken language is Text 1.3, however, the lexical density is just 36.5 per cent. This lower
borrowed, in the sense that it is retrieved in 'chunk' form, not simply le_~.densi~-~is partly a consequence of production pressure, butthe
from other speakers' utterances, but from the speaker's own store of pre- more thinTY-~P~~P oc~rr,en_~<:~C!.f..PI:QIN~it.io nal co.ntenJ~ as represented
fabricated and memorized items (Nattinger and DeCarrico, 1992; Wray, in le~ical words, also helps m~oken L<!.~g~ge._e~_?.i~!.!9.. P.r9.<:_e~~.by !is-
1999). O ne class of such 'second-hand' chu nks are utterance launchers, teners·;wtrcr, like speakers, are also having to work under the constraints
examples of which include: of real-time processing.
and there was the little old lady over the road who. The lower lexical density of talk is balanced by the fact that it is often
it is amazing more people weren't injured deceptiv'ely intricate, as speakers cdhstruct 'elaborate edifices' {Halliday,
I 'm so glad the kids were not there ... 1985: 330) of loosely linked clauses and phrases {as in the extract about
you mean, general da mage? the storm, quoted above). Halliday describes this as 'the ability to
I was speaking to erm ... "choreograph" very long and intricate patterns of semantic movement
The ab ility to achieve fluency by stringing chunks together accounts for while maintaining a continuo us flow of discourse that is 'coherent
one of the basic constructional principles of spoken language, which is without being constructional' (1985: 202). It is these 'long and intricate

12 13
Characterizing conversation The nature of conversation

patterns' that can often tax the processing ability of listeners, especially • ellipsis, where what is omitted can be reco nstructed from the context,
non-native-speaker listeners. as in:
(2) Rob: You don' t know [many people who have been
1. 1.3 Conversation takes place in a shared context affected]?
(3) Odile: Well you know [I don't know many people who have
A freak hailstorm swept across Sydney last night, causing damage been affected] except for the neighbours.
worth hundreds of millions of dollars and triggering a massive
rescue and repair effort by emergency services. • non-clausal expressions that can stand alone, and whose interpreta-
tion relies on situational factors, such as
In the newspaper text, few assumptions are made about the reader's
present state of knowledge. Even the city (Sydney) is named, although (19) Rob: Yeah .
(20) Grace: Oh yeah.
most readers of the paper will be Sydney residents, and many will have No it is amazing more people weren'r injured.
(2 1) Rob:
experienced the storm themselves. The writer cannot assume, however, Mm.
(22) Grace:
that this is the case, hence most referents (that is, the people, places
and things that the content words refer to) have to be made explicit. In summary, in fac~.!9::face jn.ter.actions participants share not only-the
The only reference that a reader who is removed from the events in ph~ical context (so that explicit mention of referents is often not neces-
both space and time ma y have trouble identifying is last night. Compare sa-ry) but also th~.in~ti.t~JiQn.lll, .soc!~ ! ~Q.O. <;:JJlturalc.out.e4_ts,,as well. This
this to: sharing of contextual kno':"ledge - resu lting in, among other things, a
high frequency of pronouns, t,h e use of ellipsis and substitute pro-forms -
(1 8) Odile: I' m so glad the kids were not there because you
means that the inte_!Pretation,...of .rh~.,-co ov!!rs:g.im~. is.. dn~enge.nt., QD. the 1 ;_,. •.
know that hole is just above Debbie's head.
immediate context. By ~?~t~~st, in written CO!Dmunication, where writers ._
(19) Rob: Yea h.
(20) Grace: Oh yeah.
cannot instantly adapt tneir message according to their ongoing assess-
(21) Rob: No it is amazing mo re people weren't injured. ment of their readers' comprehension , g reater explicitness is needed to i
(22) Grace: Mm. ensure understanding.
I
(23) Rob: So erm they go back to school tomorrow?
(24) Odile: Nor tomorrow ==
1.1.4 Conversation is interactive
{25) Rob: == M o nday.
(26) Odile: lt's Sunday. Conversatio~i s~~i!-c.h but it is not a speech. It is dialogic - or, very often,
nwlti/()gic- in that it is jo intly constructed and multi-authored. It evo lves
In the conversation, w here the context is both sha red and immediate,
through the taking of successive (and sometimes overlapping) turns by
Odile can take it for granted that her listeners will be able to identify the
the two or more participants, no one participant holding the floor for
referents of the kids, there and that hole, and that they know who Debbie
more time than it is considered appropriate, for example to tell a story
is. By the same token, Rob can safely assume that they in turn 23 w ill be
(as in Rob's turn 13 in Text 1.3). Conversation is co-constructed recip-
taken to refec to the kids, and that everyone knows that tomorrow is
rocally and continge~1tly: that is to say, s£_eakers re~pond to, build upon
Monday (although in fact it is Sunday, as the others are quick to point
and refer to the_£re~!!S utter.a.n.c.es of other speakers. Thus, Rob's ques-
out). This heavy reliance on the sh ared knowledge of the participants,
tloii (in Text 3)
jncluding knowledge of the immediate temporal a nd spatial context,
acco unts for a number of features of talk that distinguish it from most (23) So erm they go back to school tomorrow?
written text. For example: while marking a shift of t opic, nevertheless makes reference back to
• the frequent use of pronouns: for example, there are 25 pronouns Odile's utterance, seve ral turns back:
(including the possessive form her) in Tex t 3, compared to only one in ( 18) I'm so gla d rhe kids were not there
Text 1;
• the frequency of deictic items (that is, words that 'point' to featur es of At the same time, Rob's question produces an a nswer (Odile: (31) Yeah),
the physical context, such as this, that, there, now, then etc); but only after a side-seq uence in which Rob's tomo rrow is corrected by

14 15
Characterizing conversation The nature of conversation

other participants (and himself) to on Monday. Meanings are jointly The fast and unpredictable nature of conversation means that it is not
constructed and negotiated to form . a complex and textured semantic always plain sailing. Occasio nally, ambiguities need to be resolved, as in
network. As van Lier puts it: this exchange:
Progression is fast, unpredictable, and turns are rightly interwoven, (32} Grace: Is the school OK?
each one firmly anchored to the preceding one and holding out (33} Odile: You mean, general damage?
expectations (creating possibilities, raising exciting options) for the Such repair work is possible because of the reciprocal nature of conver-
next one.
sation. Speakers are either physically or audibly present and can provide
{van Lier, 1996: 177)
immediate feedback on each other's utterances, by, for example, agreeing
This dual nature of utterances, w hereby they are both retrospective and ()'eah}, 'back channelling' (~m}, show in? interest (oh really?), clarifiying
prospective, is a condition that van Lier calls contingency. In order to (you mean . .. ?), or respondmg to questwns. At the sa.me ttme, speakers
anchor contingent utterances, and to signal the direction the 'fast, unpre- are having tO constantly adapt theu message accordmg to thetr mter-
locutors' reactions, both verbal and paralinguistic. Tannen observes that
dictable' talk is heading, certain words and phrases occur frequently a t
conversation is less a matter of two (or more} people alternating between
the beginning of speakers' turns, or at transition points in the flow of
the roles of speaker and listener, but that it is more a joint production in
talk, such as yes, yeah, yes all right, no, oh, well, so, etc. These are known
variously as discourse markers and interactional signals (see Chapter 2}. which 'not only is the audience a co-author, but the speaker is also a co-
listener' (1989: 12). She uses the term involvement to characteri ze this
So, in this extract from Text 1.3, such signalling devices are italicized:
quality, and identifies features such as the rhythmic and repetitive n ature
(18) Odile: I'm so glad the kids were not there because you of much conversation as being indicative of its high-involvement style. We
know tha t hole is just above Debbie's head, also saw, earlier, how the use of pitch - and high k ey in particular- con-
(19} Rob: Yeah. · tributes to a high-involvement conversational style.
(20) Grace: Oh yeah.
(21) Rob: No it is amazing more people weren't injJred.
(22) Grace: Mm. 1 1.1.5 Conversation is interpersonal
(23) Rob: So erm they go back to school tomorrow?
So far we have contrasted Text 1.1 - the newspaper story - with the
Take away the interactional signals and the conversation doesn''r seem to spoken Texts 1.2 and 1.3. But what distinguishes the two spoken texts?
hold together nor flow so easily: Why is one 'conversation' and the other not? To answer this question,
we need to identify differences in the channel, the purpose and the tenor
Odile: I'm so glad the kids were nor rhere. That hole is just of each exchange.
above Debbie's head ...
Text 1.2 is spoken but it is mediated both by telephone and by radio:
Rob: Iris amazing more people weren't injured.
[pause] this fact alone does not disqualify the talk as being conversational.
Rob: erm they go back ro school tomorrow? Telephone talk, as we shall see, very often falls within the parameters of
conversation, despite not being face-to-face communication. Radio talk-
Of course, written language employs discourse markers, too, but usually back programmes share many of the features of casual conversation,
not with anything like the frequency they are used in interactive talk. At especially when more than two interlocutors are involved. N evertheless,
the same time, there is a greater variety of discourse markers in written the purpose of Text 1.2, h owever informal the language, is essentially
language. The following, for example, would be rare in spoken language informative. The roles ~f the speake~re established froin the outse·t as
but are frequent in certain kinds of texts, such as academic prose: more- interviewer and interviewee, the interviewee having been concacted
over, therefore, however, whereas, by the same token etc. Talk has a nar- because of his expertise, and the purpose of the interviewer's questions
rower range of markers, but uses them more frequently: McCarthy (1998} being to elicit information (and perhaps with the ul terior motive of
notes that the words yes, n o, so, well, oh and right occur significantly more assigning blame: What went wrong?}. Hence, the direction of the ques-
frequently in collections of spoken data (spoken corpora} than in collec- tioning is entirely one-way: it would not be appropriate for the intervie-
tions of written data (written corpora}. wee to ask questions of the interviewer. The management of the discourse

16 17
Characterizing conversation The nature of conversation

is very muc h in the control o f the interviewer, therefore. Even when a casual conversation, howe.ver, such as in Text 1:3, s~ch rights are equ.ally
third speaker joins the talk her questions to the expert are directed l' rributed: the relationshtp between speakers JS satd to be symmetncal.
through and by the interviewer: ~~~is is not the same as saying that the relationship is one of equalit y, as
van Lier points out:
(1) PC: Erm just hang on t here for a second because Emilia
wa nts to ask a question about that. Yes Emilia, good Equal ity refers to facrors extrinsic to the talk, such as status, age,
morning. role, and other social and societal facrors that decide one person
(2) E: Good morning Philip. Look I was at the Ethnic has more power or is 'more important' (or more knowledgable,
Communities' Council meeting w ith Angela as well. wiser, richer, and so on) than another. Symmetry refers purely to
Some o f the stuff that I saw was actua lly bigger than the ralk and the interaction itself ... symmetry refers to the equal ·
than a cricket ball, I mean it was like a big huge orange, distribution of rights and duties in talk .
you know? (van Lier, 1996: 175)
(3) PC: Mm.
(4) E: But the interesting thing about it was I mean I've seen In Text 1.3 there are a number of wha t are called initiating moves, as
hail before and even big ha il and normall y it co mes opposed to responding moves. Typically, these initiating moves can take
down fa irly compact and it looks white and it's got rhe form of questions, as in
smooth edges but some of the stuff that was coming
down last night it ... you could see the crystals and it (23) Rob: So erm rhey go back ro school tomorrow?
actually had ragged edges, it wasn't even smooth, and
but they can a lso rake the form of statements:
I just wondered whether that was a particular rype of
hailstorm that had come over you know and formed ( 18} Odile: I'm so glad the kids were nor there
differently to ochers?
(5) PC: Yeah, all right. Steve? What's the answer to that? I have Even in the brief segment that has been transcribed (Text 1.3} all four
heard repo rts that that too not all the hail that fell was of the speakers (Odile, Rob, Dan and Grace} make initiating moves, sug-
inainaball. ' gesting t hat, even if their contributions are not exactly equal, t heir right
(6) SS: No hail very often isn't in a ba ll. It comes down in a ll to initiate is equally dist ributed. The equal distribution of rights in con-
sorts of jagged sha pes and l umps because very often the versation contrasts with the situation in other spoken genres such as
var ious ha ilstones aggrega te together . . . \. interviews (as we have seen ), and service encounters (such as those that
Thus, even with more than two participa nts involved, the interview take place in shops ). T he function of service encounters is primarily
structure, and the roles inherent in this, are still in place: the interviewer trmtsactional: the speakers have a practical goal to achieve, and the
(PC) manages the interaction in a way that in casual conversation success of th e exchange depends on the achievement of that goal.
bet ween friends would seem ou t of place and extremely assertive. It is Typical transactio nal exchanges include such events as buying a train
hard to imagine, for example, tha t the conversation between friends ticket, negotia ting a loan or returning a damaged item ro a store. To a
could have gone like this: certain extent it could be argued t hat the radio interview is tra nsac-
tional, too, but, rather than the transaction of goods or services, it is the
Rob: So erm they go back to school tomorrow? transaction o f information that is the objec tive. The same argument
Odile: Yeah. might a pply also co the interaction tha t characterizes classrooms
Rob: Erm just hang on there for a second because Grace (including language classrooms ), another context in which rights are not
wants to ask a question about that. equally distributed and where information is being transacted - ty p i-
Grace: Is the school OK?
cally in the form o f facts.
Odile: You mean, general damage? . ..
The storm conversation in Text 1.3 , howeve1~ does not have as its
It is clear tha t the conversation in Text 1.3 is not managed in the same objective the trading of either goods and services, nor of info r m at ion.
way as the interview in Text 1.2, where an asymmetrical relatio nship That is to say, the satisfactory achievement of t he goals of t he encounter
exists between the interactants. In other words, the r ight to initiate, to is not pro duct-o rien ted. T hese goals can be partly inferred from wbat
ask questio ns, to direct the flow of talk is not equally distributed . In participants themselves often say after a conversation: We had a nice

18 19
 Characterizing conversation The nature of conversation

chat or The conversation really flowed, or, less positively: No one had {guage - and by pronunciation features,. such .as the use of contractions.
very much to say to each other or Graham went on and on. Examples of an mformal style m the hailstorm conversation include:
What is at stake in casual conversation is the social well-being of the
participants, the aim being essentially phatic, i.e. to signal friendship and lexical: Is the school OK? (rather than undamaged, for
example) Yeah. (instead of rhe more formal yes)
to strengthen the bonds within social groups. Rather than being directed
Oh my god I hadn't thought about that ...
at the achievement of some practical goal, the talk is primarily directed (swearing)
a t the establishing and servicing of social relationships. For this reason
conversation has been labellec\.J!!l£2:~cE.,on_g]JlS-o_pp_Q~.!i:~Lt9 trm~onal. pronunciation: You don't know? (rather than do not) The
Brown and Yule further refine the distinction between these two- pur- school's closed next to us (rather than school is)
poses: AJG.<lJ!!fe_Qf CQ_Qver_sa.ti.on!s-in(Q~mal style is_theJrequcnt use..9J vernacu-
We could say that primarily interactional language is primarily lar wmmar (Biber eta!., 1999), that is, the use of stigmatized forms that ·

(.,
I listener-oriented, whereas primarily transactional language is
primarily message-oriented.
are often asso~~te~ with a pa rticular regional variety. Examples citea by
He's left now, ain't he?
( 1983: 13) Oh yeah. Whatever they wanr they gets.
Me and ]ody had a contest for the ugliest pictmes.
Because it emphasizes the persona l element, we will use the term inter-
Don't say I never gave you nothing.
personal in preference to intemctional. This is also consistent w ith
Halliday's use of the term to identify one of language's m.etafunctions: (Biber eta/. , 1999: 112-5)
'Interpersonal meaning is meaning as a form of action: the speaker or
writer doing something to the listener or reader by mea ns of )anguage' 1.1.7 Conversation is expressive of identity
(1985: 53}. The 'something' that a speaker is doing in conversation is
social 'work'- the establishing and maintaining of social ties. The use of vernacular language, mentioned above, underlines the facJ
It is important to emphasize that talk is seldom purely transactional that c~~rs~.ti?n is a critical site for the negotia.t ion ~social identities
or purely interpersonal, but that both functions are typica1ly imerwoven It 1s through mformal ta_lk that Q~PP-le es_ti!b.llsh .ffild) !Jlaintain thei1
in spoken language: even the most straightforward transactions are tem- affili~tion wicl1 a particular socia l grou2., and ' vern acular features of
pered with interpersonal language (such as greetings} and chab mongst !£~nmar can be highly prized because of their roie in establishing and
m~intaining social solidarity am~ng the speakers in selected groups' 1
friends would be ultimately unrewarding without some kind of infor-
mation exchange taking p lace (as in Text 1.3 }. Nevertheless, the primary (B1ber et al., 1999: 1121). In fact, It has been claimed (e.g. by Dunbar)
purpose of a shopping exchange is not socia l, and nor is the primary that language, tn the form of conversation, originally evolved as a kind
purpose of the storm conversa tion to exchange factual information of 'vocal grooming' in order to facilitate the bonding of large groups. ' In
about storm damage. a nutshell ... language evolved to allow us to gossip' (Dunbar, 1996:
79}. Dunbar argues that language fulfils this social (or interpersonal)
functwn more effectively than physical grooming because:
1. 1.6 Conversation is informal
I It allows us to reach more individuals at the same time; it allows
Partly be~e of i_ts spontaneous and _interactive nature, and partly us to exchange information about our social world so that we can
because of its irterperSQllal function, conversation ISCFiaracterized by an keep track of what's happening among the members of our social
informal ~tyle. A.rt. informal (or casual) style contrasts wi~h the styJeo1 network (as well as keeping track of social cheats); it allows us to
~ \. more f~a!_~<:ken genres,.s.u.c_h_a.s_~p~ech~s__~nd r~cQf_<;kd aono.u.nce.- engage in self-adverrising in a way that monkeys and apes cannot;
and, last but nor least, it allows us to produce the reinforcing
~t_s, wl;er~,;jo~1!Jal ~p~efh is defin<:;d as ~ a careful, imner$Qnal and olten
effects of grooming ... from a distance.
QU~l~ mode gf sp,p~jng used in £~t~i.£! ~tuatl.ons ..~nd wl)jfb may
influence pronunciation, chai.c_~ of words and sentence structu-te,.- · (Dunbar, 1996: 192)
(Richards and.Schmidt~ 2002: 209). Y~formality in s~cl; is chan;cterized C.o.!lY_~rg!_ioJl.i.~ .~arked, therefore, by continual expressions of likes dis-
~ lexical choices -such as th~..ll-~.e_Qf slaQg~~~.ring and colloquiallan- ~s an.d emotional state~. Interactants exp.ress rh~ir attitud~s;b9ur'ea~h
20 21
Characterizing conversation The nature of conversation

other, a bout others who are no~ present and about ~h~. y.rgrld. T_b_e~e are the laughing stock .. . etc). There are a number of other linguistic ways f
alSQ_.c.o n~t.;nt ~xpr~~_siops_.of poli.t en~s.s, such as please,_t_hank y_ou., sorry; that speakers encode attitnde, including swearing (bloody hell!); the use
woult:l.you .. . ?; do Y,.Q11Jnit'Jd etc. There is a lsS> a lot of h~m.9_ur in con- of nicknames and familiar a ddress terms (such as love, mate) a nd the use 1
versation: fugQy stg ries are told, jokes ar~ exchanged, ~-~.<:I_ paf~lcir,~ts of interjections (wow, cool ) (see Eggins a nd Slade, 1997). I
tease o ne another. Consider this tra nscript below from an a uthentic con- To summarize, the fact that the CQ.IJ.Y.~rsati oQ.i.S qoth inrerpe_r~Qnal a n9
versation between fo ur men during a coffee break at work: expressive qf perso~el .and . social. identiry i~ lin g!Ji~t\~?J.IY ~qcod~g.i.u ..~.
v:-~ riety of way~, ma11y of which. are exemplified in the storm conversation:
1. A: It'd be good p ractice
2. B: T hat's a good idea Jim- the best suggestion I've h eard • the use of supportive back-channelling (such as M m); \
you make all this year - then maybe we can understa nd • the frequency of appraisal language (that is, lan guage that expresses
yo u Jim- I don't know how Harry understands you the speaker's attitude to, or evalua tio n of, wha t he or she is saying):
3. A: Who? completely, probably, I'm so glad, oh my God, etc;
4. B: H arry • many sentences have human agents, a n d the speaker is often the
5. A: Wh o's Harry?
subject of the sentence: l 'm so glad the kids were not there; I was
6. B: Harry Krishna
speaking to erm ... ;
7 . All: (laugh)
Who's Harry: Harry Krishna • rhe telling of stories (Rob's long turn 13);
8. C:
9 . All: (la ugh ) • a preference for informal rather than formal or specialized lexis, e.g.
10. B: Didn't you say you were going there? they had to move out of the flat rather tha n they had to vacate their
11. A: (laughs) I've cold yo u- I'm breaking away from apartment;
them now • rhe use of humour;
12. D: He's changed • the use of swearing (bloody hell!);
13. A: I've changed • rhc use of nicknames a nd familiar address terms (s uch as love, mate)
14. B: You're giving it away?
15. C: He's shaved his 'mo' off (Lexical and gramma tical features encoding the interpersonal purposes
16.
17.
B:
C:
He's o nl y getting too lazy to carry his upp~ lip around
Harry Krishna
of conversation wiJJ be explored in more depth in Chapters 2 and 3.)
\
18. All: {laugh)
(Authors' data) 1.1.8 Conversation in other modes

Here the men's use of mu tual teasing serves to ensure (on the· surface) So fa r, we have been working under the assumption that conversation is
that they have a laugh and enjoy the coffee break. But the purpose of necessarily spoken. However, this assumption needs to be qua lified in the
casual talk such as this is a lso to help construct cohesive relationships light of the development of computer-media ted communication (CMC),
between a group of people who are not necessarily friends but see each such as that which occurs in internet cha t rooms, where, although com-
other on a daily basis. munication takes place in real-time (it is synchronous), it is written. That
Text 1.3 a lso had examples of humour, such as Dan's reference to The is to say, cha t participants key in utterances at their 'home' terminal that
X -files - a popular television series abo ut the paranormal: are then almost immediately available for all other participants to read and
respond to. Quite often, CMC shares many of the kinds of features of ta lk
Rob: . . . it fell through the ceil ing and landed in her la p bet~veen friends that we ha ve identified. H ere, fo r example, is an extract
when she was sitting [Odile: Mm] wa tching
of tnternet chat. The first 'speaker's' turns are indicated by the time ar
television.
which they were posted (e.g . [12:40]); the second speaker's are marked > :
Dan: Watch ing The X-files probably.
All: [laugh] Session Sta rr: Sun 26 May 1996 12:40:29
Also highly frequent in conversa tio n is the occurrence of appraisal lan- [12:40] H ow are yo u?
guage (see Martin, 2000), including evaluative voca bulary (awful, won- >good.
derful, ugly, weird, etc) a nd form ulaic expressions (What a joke; He was > fine here

22 23
Characterizing conversation
The nature of conversation
(12:41] Please cell me more about you:)
> like what
well, yes, but ... There. is humour (test ~eds for a .living). ~ven par-
(12:42) Ok, just to refresh memories, I'm 32,
. 1. guistic indicators of mvolvement are s1gnalled usmg emorrcons: ©,
.l Ill f h I k f . [ . . 1. f
divorced, English, 8 old son ... in order to compensate or t he ac o ~1sua. or mtbonat1odna mf hormda~
(12:42] rest beds for a li vi ng:) · 1. Hillier characterizes sue texts as wnnng to e rea a.s 1
(101 ear
> well, I've never really talked to you before so no (1004: 213).
refresh ing was necessary- chis is a ll new -Similar features have been identified in asynchronous electronic com-
[12:42) Ok :) 0 mnications, such as in newsgroups and email exchanges, where there is
[12:43] What about you? rime lapse between the sending of the message and its reception. Text
> I'm 40, single, American, no children ;:1essage exchanges ma y be either sync.hron~us or async~ronous, but,
> and I' m an interior designer ~irher way, they are ch aractenzed by a htghly mformal and mterpersonal
[12:43] Whereabouts in America? ... srde. Thus, the electronic medium has had the effect of dissolving many
> In Palm Beach, Florida
[12:43) I' m in southwest England
of the traditional distinctions berween written and spoken interaction:
> whereabou ts?
interacranrs are less writers than co-participanrs in an exchange that
[12:44] Sounds nice .... n:sembles live talk. It is too early to say to wha t extent these proto-genres
[12:44] :) will develop their own idiosyncratic features, both linguistic and prag-
[12:44] Wellington, Somerset. matic (but see Crystal, 2001 ). As interesting and as suggestive as these
> I've been there. It's beautiful country d~velopments are, they are outside the scope of this present study. Suffice
[12:45] Really? :) ir to say that spoken conversation remains the interactional type from
[12:45] It is lovely here which these electronically mediated interactions derive many of their
[1 2:45] I was chatti ng with Lee yesterday ... chcHacteristics.
[12:45) but she a lways seems to be busy
> yes, she's a friend of mine
[12:46) :)

I/
1.1.9 Defining conversation
[12:46] She is nice
[12:46) very sweet ... To summarize: CQn~ers~_tionjs (p-rimarily) ~poken and it is pJann~cl. a_.r:td
[12:46) but very busy produced spontaneously, i.e. in real time, which accounts for many of
> she is. We've know n each other for lots of years the ways it differs linguisticaTly from written language, .or from spoken
[12:47] You live close by? htnguage that has been previously scripted (as in news broadcasts, for
> about an hour from her example). In Halliday's formulation: '\Y(!iring_ exists,_ whe~_eas speech
[12:4 7) That's nice :) happens' (1985: xxiii). C..Qn~~rs;niqn ~-the kii.itl-nhpeech t~at ha_ppens J
© 1996-2006 Andy and Lisa H unt a nd Quantum Enterprises informally, S}'mmetrically and for the p1,1rposes of estap~_i~hing and main-
taining social ties. This distinguishes it from a number of other types of
This text shares a number of the features of conversation that we have communication, as shown in Table 1.1, although it is important to stress
isolated in our analysis of Text 1.3 (the storm conversation). Speakers rhat there is considerable variation within categories. There are sections
take turns; they respond to pcevious turns; questions are distributed of news broadcasts that are unscripted, for example; and not all emails
between participants; topics are intcoduced, developed, dropped; there ~crvc a transactional function, nor is all classroom talk dialogic.
are opening moves (how are you?), a nd presumably closing ones; there On the basis of Table 1.1, we can now offer a more comprehensive
are evaluative responses (that's nice), checking moves (really?) .and definition of conversation than those with which this chapter began:
confirming moves ( [but very busyJ she is). And, as a consequence of the Conversation is the informal, interactive talk berween two or more
constraints of real-time processing, the lang uage is syntactically rela- people, which happens in real time, is spontaneous, has a largely
tively simple, elliptical (sounds nice; about an hour from her), and often interpersonal function, and in which participants share
produced in clause- or phrase-length chunks: she is nice/very sweet/but symmetrica l rights.
:o
very busy. Discourse markers and interac tional signals are used mark (~ore that, because we have defined conversation as being informal, we
shifts in the direction of the talk, and to manage the mteractton: ok, will use the terms conversation and casual conversation interchangeably.)
24
25
 Approaches to the analysis of conversation

1.2 Approaches to the analysis of conversation

~· )( )( )( )( '>....,
Spoken language, and conversation .in particular, has only rec~ntly
started ro receive the same kmd of detailed lmgUJstJC attenuon as wntten
language. Moreover, many approaches to the analysis of conversa-
tion have been partial, focusing on particular features of conversation
through the lens of a single theoretical construct. The approach we will
be adopting in subsequent chapters is a more eclectic one, on the
grounds that a more comprehensive, and hence potentially more useful,
~nalys is should draw on a variety of theoretical models. Our starting
premise, and one of the basic assumptions shared by all the different
models to be disc ussed below, is that conversation is structurally pat-
terned, and displays an orderliness that is neither chaotic nor random
but, rather, is tightly organized and coherent. It follows that, if this
organization can be described in ways that are accessible to teachers and
learners, there are likely to be practical classroom applications. (T his
docs not mean, of course, that one such application would simply be to
'deliver' the description to learners without some form of pedagogical
mediation. )
Conversation, then, has been analysed from the perspective of a
........ . ~
.~ 0.0 number of different academic disciplines. T he most important of these
o.oo
- 0 ~

· "'
-o
....
- -::l
E
are sociology, sociolinguistics, philosophy and linguistics.
Figure 1.1 below provides a typology of these different approaches to
the analysis of conversation.

1.2. 1 Sociological approaches

Perha ps the most significant contribution to the study of conversation

has come, not from linguistics, but from sociology. A fundamental
concern of socio logists is to account for the organization of everyday
life, including the way that social activities are structured and ordered .
The sociological approach to analysing 'talk-in-interactioo' has come
to be known a.s -Coryv~_rsation Analysis_ (CA), a branch of sociology
which posits that it is in and through conversation that most of our
routine everyday activities are accomplished. CA is represented pri-
marily in the studies of Sacks, Schegloff and Jefferson (see, for example,
1974). The objective of CA is to describe and explain the orderli-
ness of conversa tion by reference to the participants' tacit reasoning
procedures and sociolinguistic competencies. To take the 'hailstorm'
conversation as an example, a researcher within the CA paradigm
would be particularly interested in showing how the speakers are ori-
ented to the rules of turntaking and how they accomplish this in an

27
Characterizing conversation Approaches to the analysis of conversation

- Sociology Conversational (23) Rob: So erm they go back tO school tomorrow?

Analysis (24) Odile: Not tomorrow = =
(25) Rob: ==Monday.

-[
Ethnography It's Sunday.
(26) Odile;
Interactional (27) Rob: Monday.
- Sociolinguistics
Sociolinguistics (28) Grace: Monda y.
(29) Odile: Monday.
Variation (30) Rob: Mm.
Theory (31) Odile: Yeah.

Speech Act The repair sequence is interpolated into another sequence, which is the
Theory simple, two-part, question-and-answer sequence of turns 23 and 31. The

-Philosophy
-L Pragmatics
Birmingham School
question and the answer would normally constitute what is ca lled an
adiacency pair, that is a two-part exchange, the second part of which is
functionally dependent on the first- as in greetings, invitations, reques ts,
and so on. In this case, a repair sequence is inserted (i.e. it forms an inser-
tion sequence) in the adjacency pair, because the first elem ent of the
Structura 1- pair - the question - cannot be answered until the question ha s been
functional 'repaired'. (Another example of an insertion sequence occurs at turn 33.)
Conversational analysts are particularly interested in w hat such
sequences demonstrate about the orderliness of conversa tion , a nd how
Systemic-functional
-Linguistics the conversational 'work' is co-operatively m anaged. (In Chapter 4 we
Linguistics :

-[
will return to the subject of conversation management.)

Social-
1.2.2 Sociolinguistic approaches
semiotic
Critical DiscQurse
Sociolinguistic approaches have emerged from the theoretical common
Ana lysis ·
ground shar ed by sociology, anthropology and linguistics. These are
especially concerned with the analysis o f language in its socia l context,
and the way that language use varies according to contextual and cul-
Figure 1.1 Different approaches to anal ysing conversation (from Eggins and
tural factors. H ymes (1972a), one of the foremost proponents of what is
Slade, 1997)
called the ethnography of speaking, proposed a ru bric fo r in vestigati ng
the contextua l factors that impact on any speech event. These factors
include, among o thers, the setting, the participants, the ends (or purpose)
of the speech event, its key (i.e. its 'tone, manner, or spirit', such as
orderly manner, as when Grace 'takes over' Rob's turn and fini shes it whether it is serious or jokey), and its genre, or type. Like an anthro-
fo r him: pologist, an ethnographer, armed with this rubric, would seek not only
(6) Ro b: because the w hole= = to describe the speech event under study, but to explain it, particularly
(7) Grace: = =roof collapsed. in relation to the culture in which it is embedded. An ethnographer
observing- or, better, participating in- the hailsto rm conversa ti on, for
Conversation analysts are also interested in how conv ersati~nal 're~air~' example, might be interested in this comment by Dan, in response to
are achieved and how these repairs also illustrate the participants on- Rob's mentioning that the woman w ho received the hailstone in her lap
entation to ;he basic rules of turntaking, as in this instance when Rob was watching television:
uses the word tomorrow to ask about an event that will in fa ct take place
the day after tomorrow: (16) Dan: Wa tching The X-files pro ba bly.

28 29
Characterizing conversation Approaches to the analysis of conversation

The fact that this th rowaway line not o nly effectively ends Rob's narra- .11 re is also a seam of evaluative language running through the story,
tion, but elicits laughter a ll round, despite the fact tha t it would seem ~u~~1 as little old lady, this old Italian woman, which se~ves to convey the
rather flip pant in the light of the serio usness of the S1tuat1.on (after all, ·akers' (a mused ) attitude to the events they are relatmg. (Spoken nar-
the woman could have been killed), suggests that th ts parttcular speech sprt:i.ves will be described in more detail in Chapter 5 .)
r.t
event has ends other than the simple relay ing of disaster stories, and that
these ends are mutually understood. The use of humour (a feature of the
speech event's key) serves to create a sense of group membership, and 1.2.3 Philosophical approaches
this has cultural implications tha t an ethnographer m1ght be keen t~ Speech Act Theory, which grew out of the philosophical study of
ex p lore. Is this light-hearted key a distinguishing feature of thts kmd ot meaning, has been influentia l in the way it has added to o ur under-
conversation in this particular culture, irrespective of t he senousness of c;tanding of how speakers' intentions are expressed in language.
the topic, for instance? And what cu ltural assumptions are shared by Ph ilosophers such as Austin ( 1962 ) and Searle ( 1969 ) re-conceptualized
the mention of a television programme that dra mat1zes supernatural speech as ' action' and attempted to describe how (a potentially infinite
events? number of) spoken utterances can be classified according to a finite- and
Sociolinguists would also be interested in the variat~on that the speak- relatively limited - set of functions. By ascribing communicative func-
ers display in, for example, their pronunciation or then· c~01ce of word~, tions ro utterances, and by attempting to describe the conditions under
and would attempt to correlate these linguistic factors wtth soctal van- which an utterance can fulfil a specific function, speech act theory helped
abies - such as class, ethnicity or gender. The fact that Odile is of. French pave rhe way for a communicative - rather than purely fo rmal - descrip-
o rigin, for example, might be reflected not only in h~r ~roD:unciatwn,.but tion of spoken language. The hailstorm conversatio n provides at least
in her interactional style- and the study of such va natwn IS the provmce rwo examples of utterances tha t might be interesting to analyse from the
o f interactional sociolinguistics (see Tannen, 1984, 1989). Proponents of perspective of speech act theory, since their function is not transparent
variation theory (see, for example, Labov, 1972) are inter~strd in track- in terms of their fo rm. That is, they are declarative sentences in terms of
ing language change and variation as evidenced in such dis~o urse umts their grammatical form, but are nevertheless interpreted as being ques-
as spoken narratives . Narratives exhibit relatively stable structures, but tions:
allo w for a great deal of linguistic variation w ithin these sth~etures, not
(2) R ob: You don 't know?
least because of the tendency of spea ke rs to adopt a ven 1<ilcular style
whe n narrating, aU of w hic h makes them an ideal site for the study of and
patterns of linguistic va riation and change. The narrative that Dan and (23) Rob: So erm they go back tO school tomorrow?
Rob co llaborate in telling (in the hailstorm conversatwn ) displays <1
number of the creneric features of n arratives that Labov and Wa letzky Note that in the transcript the two utterances have been punctuated as if
(1967) identified in their seminal stud y of the narrati ves of ur ban Afro- they were questions but, of course, in spoken language there are no such
Americans. T hese include an abstract (or brief annou ncement of the things as 'question marks'. Nor is rising intonation necessarily a reliable
topic of the story) : indicator of a question: ma ny statements (especially in Australian
English ) are uttered with a rising intonation. And many utterances that
(12) D an: And there was the little o ld lady over rhe road . ..
are both intended and interpreted as questi ons are uttered wi.t h a falling
an orientation (to the situational context and the participants): inronation. According to speech act theory, to count as a question an
utterance must fulfi l a number o f conditions, or rules. These include the
(13) Rob: Ob yea h [laughs) she was sitting in her living room .. ·
condition tha t the speaker does not know the answer, that the speaker
a complication: sincerely wa nts to know the answer, and that the answer is not likely to
... and a hailston e fell th rough the skylight ...
be forthcoming without the question being asked. But, with regard to the
first 'question ' (You don't know?), the previous speaker (Odile) has
which is in turn typically followed by some form o f resolution (although already said that she doesn't know, suggesting that Rob's question may
not in the case of the ha ilstorm story) and a coda: have some other function, such as expressing his surprise - or even his
(16) Dan: Watching The X-files probably. disbelief- at her not knowing many people who have been affected by

30 31
Characterizing conversation Approaches to the analysis of conversation

the hail st orm, i.e. the utterance has an expressive function, rather than . ·he suddenly sees the relevance of Ro b's comment a bout the school
a purely representative one. as ~r door being closed , w hich in turn relates ro Grace's question: Is the
Closely related to speech act theory, and sharing a similar philosoph- ~,~;ool OK?- a question that at first Odile couldn't quite see the rele-
ical background, is pragmatics. Like speech act theory, pragmatics is \"tnce of. (S he asks: You mean, general damage?). Only three turns later
concerned with eluc idating speake r meaning, especially where speaker d~es she see rhe point: tha t her own children's school may also have been
meaning seems to be at variance with semantic meaning, that is, the -losed because of hailstorm damage. Pragmatics, then, offers insights
literal meaning of the words and g rammatical fo rms of an utterance. ~nro how speakers and listeners co-operate in order to achieve coherence,
Moreover, pragm atics goes further and seeks to answer the question as ~laking sense of each other's utterances by searching for relevance in the
to how the speaker's meaning is retrieved by listeners, rather tha n being co-text and the context.
interpreted literall y or nonsensically. So, fo r example, Odile's utterance:
(18) Odile: I'm so glad the kids were not there because you 1.2.4 Linguistic approaches
know thar hole is just above Debbie's head ...
Originating mo re in linguistics than in any other discipline, both the
would make very little sense to anyone who had nor been privy to the Birmingham School of Discourse Ana lysis and Systemic Functional
conversation up to this point. There is no inherent logical connection [.i11g11 istics have made major contributio ns to the descrip tion and analy-
between Odile's being gla d the kids were not there, and the hole being sis of spoken language.
just above Debbie's head. And even with access to the co-text (the pre- The Birmingham School, influenced by the work of Firth (1957), was
vious utterances) the significance of the hole is not obvious. Nevertheless. c~r.tblished primarily by Cou lthard and Sincla ir, whose earlier work
the others seem to have no problem interpreting the statement as focused on the analysis of classroom discourse (see, fo r example, Sinclair
m eaning the hole where the hailstone came through is just above the and Cou lthard, 1975, and Sinclair and Brazil, 1982). They were inter-
place where Debbie's head would have been had she been thrre. They are ested in identifying the 'grammar' of interaction, and in particular the
able to make this interpretation partly thro ugh recourse to s hared con- way a speaker's discourse ch oices are pre-determined by the immediately
textual knowledge (they have just been shown the ho le). But they are also preceding utterance, analogous to the way that the choice of a word in a
interpreting the utterance through adherence to a mutually accepted set senrcnce is determined. This 'discourse grammar' was described in terms
of principles for th e condu ct of ta lk, without which coherent conversa of a hierarchy, fro m the largest units (e.g. a lesson) to the smallest, these
rion would be impossible. These principles w ere first outlined by Grier being the individual acts of which a lesson m ight be composed. These acts
(1975), who expressed them in terms o f four 'maxims': Me not to be confused witb speech acts, as mentio ned above, rather, they
an: defined in terms of their interactive function, such as eliciting, inform-
1. Maxim of quantity: Make yo ur contribution just as informative as ing and evaluating, or their turn-taking function, s uch as cueing and nom-
required. inating. Intermediate categories in the hierarchy include exchanges a nd it
2. M axim o f quality: Make your contribution one that is true. was the structure of exchanges which was the focus of particular interest.
3. Maxim of relation: Make you r contribution relevant. The identification of the t hree-part exchange structure that characterizes
4. Maxim of manner: Avoid o bscurity and ambiguity. Be brief and classroom interacrion - initiation, response, follow-up- is one of t he best-
orderly. known findings of this School (see Chapter 7). But discourse consists of
la rger units too, such a s transactions, and these are often identifiable by
Thus, in acco rdance with the maxim of rela tion, and given a ll the possi-
the di.scourse markers that frame them. In the hai lstorm conversation,
ble m ea nings that Odile's utterance could have had, her listeners were
there IS a clear division between two transactions at turns 22 and 23:
disposed to select the interpretation that was most relevant, that is the
one which (according to relevance theory, Sperbe r and Wilson, 1990) (22) Grace: Mm.
required the least processing effort in order to make sense. T his assump- [pause]
tion of relevance is fundamental to the maintenance of conversational (23) Rob: So erm they go back to school tomorrow?
coherence. It acco unts, for example, for Odile's o utburst: Ro~'s so serves to frame the introductio n of a new topic, and the trans-
(38) Odile: Oh my god 1 hadn't thought about that ... acnon that follows is composed of a number of exchanges about this

32 33
Characterizing conversation Approaches to the analysis of conversation

topic, each exchange realized in the form of question-and-answer moYes. . P onal meanings: meanings about roles a nd relationships.
2. mter ers . d
The fact that exchange structure allows considerable flexibility - more These are a reflecuon o~ tenor; an .
so than, perhaps, sentence grammar a llows - is evidenced by the way that 3. textua 1meanings· · mea mngs abou t the message. These are a reflectiOn
the exchanges are interrupted by insertion sequences, as we noted above. of mode.
Systemic Functional Linguistics (SFL) is largely derived from the work
· ·ficantly these three types of meaning occur simultaneously in
of Halliday (see Halliday, 1985; Halliday and Matthiessen 2004; Eggins, Most stg111 ' f h h ·1
even dause ortext. Take as an example, a clause rom t e a t storm con-
1994). The central concern of SFL is, in a systematic way, to relate lan-
guage to its social context a nd, in particular, to the functions ~t performs vers:~rion, Rob's comment:
in that context. Such a concern leads to a focu s on the a nalysts of actual (21) Rob: No it is amazi ng more people weren't injured.
lano-uage in use: of texts considered in relation to the social context, both
.· ltaneously expresses meanings about the world (more people
cul~ural and situational, in which they occur. Systemic Functional ~mm h I · I. .I I . .
wt•r('ll't injured) and about t ere at10nsd11p wtn t1e pafrtictpandts 10(t e
. h
Linguistics stresses the centrality of the stud y of conversation to the
. •~:rsation thro ugh the exaggerate expressiOn o atmu e zt ts
study of language, beca use conversation is the most important vehicle b~ ~un \
means of w hich social reali ty is rep resented and enacted in language.
' I f ·
r,
. h
,m:.dng). This has ~n interpersona un~ n on JO a.t'. not on y . oes It
I d .
·c ro validate Odtle's reltef that the k1ds weren t IDJured, but 1t cap-
Moreover, 'to understand the na ture of text as socia l action we a re led
mn:'> rhe general feeling of an:azement s h
:>cf\
. are d b. y t he parttcJ~ants,
· · an d
naturally to consider spontaneous conversa tion, as being the most acces- ·intorces the sense of solidanty that the1r talk msta nnares. Fmally, the
sible to interpretation' (Halliday, 1978: 140). n:
Jiscourse marker No has a textua 1 f unct1on . . t I1at tt
m . connects R o b's
Systemic Functional Linguistics is a functional approach to language utterance with the preceding ta lk, signalling an agreement to the n ega -
description. Functional description s seek to explain the internal organ- tive 1mplication of Odile's comment, i.e. that people weren't injured.
ization of language in terms of the functions that it bas evolv(!d to serve. Together the field, tenor and mode of the situation constitute the reg-
As a functional approach, SFL argues that language sh ould 1be thought ister LJariables of a situation. Texts whose contexts of sttuanon co-vary
of as real instances of meaningful la nguage in use. In turn, pecause lan- in the same way are sa id to belong to the same register. The concept of
guage - in the form of written or spok en texts - always occ'urs in social ret{isrer is a 'a theoretical explanation of the common-sense o bservation
contexts, SFL argues for the need for a descriptive framew~rk whereby rh~t we use language differently in different situations' (Eggins a nd
language and context a re systema tically and functionally related to one ~turin, 1997: 234 ). It is a useful way of explaining and predicting the
an other. relationship between features of context and features of text. Thus,
It is well know n that different contexts predict different kinds o f lan- rht: three texts about the hailstorm cited at the beginning of this chapter
guage use. SFL argues that there is a systematic correla tion between (tht: newspaper account, the radio interview and the friendly conversa-
context a nd language, and, specifically, that three dtfferent aspects of tion) al l share the same field , in that they are all a bout the hailstorm. But
context correlate with the three different kinds of meaning expressed in tht:\' differ with regard to their tenor an d mode. It is these tenor and
language. H alliday (1985; Halliday a nd Matthiessen, 2004 ) identifies mode differences which are reflected in different kinds of grammatical
the determining context factors as being: and lexical choices, and w hich account for such different wordings as
• the field of disco urse (what is being talked or written a bout); the following:
• the tenor of discourse (the relationship between the participants}; and At Paddington, Ms Jan Mourice said all houses on one side of
• the mode of discourse (whether, for example, the language is written Prospect Street had windows smashed. [newspaper repon]
or spoken ).
Steve Simons, a senior forecaster with the Bureau, joins me on the
The significance o f field, tenor and mode is that these th ree contextual line this morning. [radio interview]
dimensions are then encoded into three types of meanings represented tn Oh a friend of ours in Paddington, they had ro move out of the flat
language. The three types of meaning a re: [conversation]

1. ideational meanings: meanings about the world. These are a reflection Tht way that, within specific cultura l conrexts, register variables
o f field; influence how particular texts (whether spoken or written) are structured

34 35
 Characterizing conversation Summary

and have become institutionalized is captured in the concept of genre (see rime. CDA is particularly effective at unmasking this kind of ideological
Chapter 5). A genre is a recognizable language activity, such as a new, ·su b-rexr'.
report, or a conversational story, whose structure has become formalized · Finally, rhe analysis of conversation has been immeasurably enhanced by
over time. Speakers of a language know how to perform these language the advent of two technologica l innovations: the invention of the tape
activities in ways that are appropriate to their cultural contexts. For recorder and the computer. T he former allowed the recording and tran-
example, they know how to make stories interesting, entertaining or scription of authentic dar~, which in turn paved the way for descriptions of
worth listening to. Genre theory provides sema.o.tic.....and.g.~;-a.llllUi!!ica l spoken English that descnbe attested use, rather than basing their descrip-
j/ tools for grouping texts with similar social f~rposes into texk~q;­ tions on invented examples. The computer enables researchers to compile
exaniple~Thenadstorn1Conversati6n contains a story (in turns 12, 13 and and consult databases (ca lled corpora) of spoken language, and has given
16) that has its own internal str ucture (an orientation and a complicat- rise ro the science of corpus linguistics (see Chapter 2). Much of the au then-
ing event) an d which, in turn, is embedded within the larger conversa- ric spoken data we use in this book com es from different spoken corpora.
tion. As we will see in Chapter 5, in conversation speakers weave io and for example, the Australian data comes from a spoken-language corpus
o ut of telling stories and gossiping, and these genres are nested w ithin called OZTALK, a joint Macquarie University and University of
highly interactive talk consisting of shorter exchanges. Technology, Sydney, project (referred to hereafter as OZTALK).
Sharing with SFL a concern for the social context of language in use, Among other things, corpora permit the analysis and comparison of
proponents of Critical Discourse Analysis (CDA), such as Fairclough word frequency counts across a number of different contexts. For
(1995), view discourse as a form of social practice. They argue that dis- example, a word frequency ana lysis of the hailstorm text, using a soft-
course, including conversation, can be properly understood only in rela- ware program, shows that over 85 per cent of the words in the text are
tion to the social structures that it both shapes and is shaped by, and in in the most frequent 1000 words in English. Words that are not in this
particular the relations of power inherent in these social structpres. A job frequency band include hail, corrugated, ceiling, skylight, injured, tiles
interview and a friendly chat, for example, while sharing many super- and X-files. Moreover, 50 per cent of the words in the text are function
ficial features, manifest very different relations of power, aqa these dif- UJords, such as a, about, am, and, etc. Of these, the most frequent is the
ferences will determine the kind of language choices that the participants definite article the (12 occurrences, compared to only one of the
make, including the way turns are taken and distributed. At the same indefinite article a). A concordance of the examples of the in the text
time, the language choices they make are socially constitutive, in that (Table 1.2 overleaf) shows just how many references- to people, places,
they help to sustain and reproduce the existing social structures, as when things, TV programmes- the speakers share. This in turn underlines the
the job interviewee has to be invited to ask questions, rather than simply way that their conversation both reflects and reinforces the commonal-
being allowed to ask them unsolicited. But even conversation, which on ity of their different but interconnected worlds.
the surface would seem to be one of the most egalitarian forms of inter-
actio n, can itself be the site for significant interpersonal work as interac-
tants enact and confirm their social identities. This is particularly the case Summary
when conversation is used as a way of 'disguising' inequalities of power,
as when a boss might say to a n employee: 'Let's have a little chat about This chapter has aimed to provide a working definition of conversation.
you r future plans .. .' To do this, we have attempted to answer these questions: How is con-
The hailstorm conversation, if examined through the lens of Critical versation different from other related forms of communication? and
Discourse Analysis, would not reveal significant inequalities amongst How are these differences realized in terms of language? In the chapters
the participants, perhaps, but their choice of language to talk about - that follow we will take conversation to mean spontaneous, spoken, dia-
and to position themselves in relation to- parties who are not present logic (or multilogic) communication, taking place in real time and in a
may shed some light on unstated but shared ideological va lues. The shared context, whose function is primarily interpersonal, and in which
fact, for example, that Dan's neighbour is first characterized as a little the interactants have symmetrical rights.
old lady, rather than, say, as an elderly pensioner, and then as this old We have also looked at some of the different theoretical constructs
Italian woman, construes her not only as an object of mirth, but tends that h ave provided too.ls for analysing a nd describing conversation, and
to reinforce cultura l stereotypes of age, gender and race, all a t the same which will inform our ana lysis in the chapters that follow. In the next

36 37
 Summary

tr chapters, we will take a closer look at t he linguistic features of con-

to~sation, specifically its vocabulary a nd grammar, irs discourse structure
\·ee rhe way it is constructed sequentially and reciprocally}, and the way
II· .
·rs laruer structures ha ve become generalized into certain culrurally-
~mbedded patterns - or genres. A description of conversation a t these
hfferent levels clearly has implications for reaching, and these implica-
~ions will be explored in the second half of the book.

Task
ln the transcriptio n of spoken English below, can you identify fea -
tures that are evidence of (a) spontaneity; (b) its interactive na ture
(i .e. reciprocity); and (c) its interperso nal function?

Mary: I discovered that Adam used to be my neighbour

yeste rday.
Fran: Oh really ?
Mary: Yes. In the conversation. Over a year, oh, well over
a year ago.
Adam: Yes we used to be neighbours about, about a year
ago. About yea h well over a year ago we used to be
neighbours in Glebe.
Fran: But you never ac tua lly, did you know people in
common?
Mary: Yes.
\ Adam: Yes definitely. We knew the peop .. . person who
moved in there afterwards.
Mary: Well he moved in whi le [was moving out. He lived
wi th me for about fou r or five weeks but I was never
there. But I couldn't believe it.
Adam: [laugh] It's incredible.
Mary: We used to think there were these strange people
next door. And Adam used to think there were these
strange people next door.
All: [laugh]
Adam: And now we KNOW there were strange people next
door.
Mary: No doubts yea h.
(Autho rs' data )

39