Sunteți pe pagina 1din 92

Introduction to

Discourse Studies
Jan Renkema
University ofTilburg

, ... '"
... c .
:)"t. :t
e'>.
'l
!...-1-
lohn Benjamins Publishing Company
Amsterdam I Philadelphia
N
00
paper used in thii publication meet$ the minimum nquiumenu
of American National Standard for InformatIon Sciences _ f'crmaneet
of PapH for Ubnry MatttWs. AtUI z.39.4S1984.
Ubrvr of Coqrut .. b1ical loll 0 ...
.......... ,
Introducloon lodiscoulV lludoa I Im Rnlktma.
p. C1t\.
Indudn bibbopplucal rdffmca .,.d ,nda.
I. DlKourx anatysas. I. Tilk
PlO2.R4S8 2004
<f01'4i-dc12 lOO405)766
I"" 90 212 2610 5 (Eur.) 1 1 saal I 529 I (05) (Hb: olk. papa)
IS." 90 212)121 O(Eur.) , I saal I SXl5 (05) (Pb; alk. papa)
C 2004 - ,ohn BtnJamins B.V.
No pari oflhis book may bt; nproducrd in any (orm, by pnnl, phol:oprml. microfilm, or
;1IIy otMr 1IIt;lIl$, W1thool wnllen permwlon from the publWter.
Jolm Bent-lTuns Publistunl Co. . P,O, Box )6224 . 1020 ME AllUltrdam . Tht Nttherbnds
lohn Benjamins Norlh Amtrica . P,O. Box 21519 . Phlladdphia P ... 19118 0519 . IJS'"
Table ofContents
Acknowledgments
Introductton
1.1 A rough ddinition of discourse studies
1.2 Aim and Siruciure ofl his book 1-
1.3 The presenl3tion oflhe material 5
QU6tions and assignments 6
Bibtiographicalmformation 7
PART I
General orientation
2 Communication 35 action
2.1 The Organon model 11
1.1 Speech act theory 11
1.) IUocutions 10 discourse 16
1.4 The (ooperative principle 18
l.S Relevance theory 10
1..6 Pohteness Iheory 24
Qucstions and assignments 19
Bibliographical information ]2
3 Discoul"SC! in communication
).1 lhe pragmalic persptive )5
).1 Rules for symbolic interaction 37
).) Messages betwn sender and receiver )9
).4 The discourse situation 4)
).5 The socio-semiotic approach 45
).6 What makes discourse discourse? 48
Queslions and assignments 51
Bibliographical information 55
ox
9
"
3S
Tabte of contenu
PART 11
Backpacking for a scientificjoumey
4 Discourse types
4.1 The variely of functions and forms 59
4.1 Written languagc and verbal inleraction 65
4.) Everyday and lilerary language 67
4. 4 Electronic discoutse 69
4.5 Conventionaliud forms fot conventionalized oceasions 7)
4.6 Mullimodalily 76
Queslions and assignments 81
Bibliographical information 85
5 Structured content
p Propositions 87
5-1 Topics 90
5.) Macrostruclures 94
54 Superstructures 97
Questions and assignments 100
Blbliographical information 101
6 Discourse connections
6.1 Cohesion 10)
6 . 1 Referential elements 106
6 .) Coherence 108
6.4 Rhetorical Struclure Theory 111
6.5 Discourse relation research 11)
Queslions and assignments 116
Bibliographieal information 110
7 Contextual phenomena
7. 1 Deixis 111
7. 1 Staging i2)
7.) Perspectiviz.ation 116
7.4 Given- new management 1)0
7. 5 Presuppositions IJ1
7.6 Inferenees 1)6
Queslions and assignments 1)9
Bibliographical information 14)
S7
'9
8,
'0'
'"
8 Style
8., Form, content and situation
'"
8., Views on style
'47
8., Stylistic analysis "0
8., Examples of 5tylistic research
'53
Queslions and assignments
'S7
Bibliographieal information
,,.
PART 111
Special modH of communieation
9
Conversation analysis
9'
Transcription systems
,6,
9'
The turn-taking model '6,
9'
Sequentialotganiz.alion
,"
9'
Discourse markers ,68
Questions and assignml"nts 171
Bibliographical information 174
10 Informative discourse
10.1 Readability in a formula 175
10.1 The ml"asurement of understanding 178
10.) Judging discourse quality IBo
10A Thl" improvemenl of documents 184
Questions and assignments 188
Bibliographiea! information 190
11 Narratives
tl.1 The slruetutl" of fairy tales 191
II.l The sociolinguistic approach 19)
11.) The psycholinguislic approach 195
11.4 The organtz.ational approach 198
Questions and assignments 101
Bibliographieal informat ion 101
11 Argumentation aod persuasion
I:U The structure of argumentation 10]
11.1 The pragma-dialmical approach 105
12.] The social psyehological approach 107
Table of contents vii
'"
"9
,6,
'75
'9'
'0'
ii Tableofconttnts
11. 4 The quahtyof argumentation 1 11
Questions and assignments 116
Bibliographieal information 1 18
PART IV
Speci al interests
'3
Oi sc:ourse and cagnition
1). 1 Modeling discourse production
'"
1).1 Product and process analysis
".
1).) Processing and prior knowledge
".
1).4 Aspecu of processing
'" 1).5 Modeling discourse processing ,,8
1).6 The metaphor in cognitive research
'4'
Questions and assignments
'47
Bibliographieal information
's-
'4 Discourse and institution
14.1 The agent - dient approach
'53
14.1 Polities
'" 14.) Law
's8
14.4 Bureaucracy
,.,
14.5 Media , ..
14.6 Health care ,,-
Questions and assignments
'7S
Bibliographieal information ",
's
Discourse and culture
151 The Sapir.Whorfhypothesis
'" Ip Critieal Discourse Analysis ,8,
15) Gender ,8,
154 Racism
'88
155
Intercultural communication
'.-
Questions and assignments
'95
Bibliographical informalion
'97
Key t o the questions
References
Index
".
'"
'53
'"
'"
".
'57
Acknowledgments
This Introduclion 10 Discourse Studies is the result of almost twenly-five years of aca-
demie and nonacademie in teaching discourse sludies. Guided by the
quest ions and answers of hundreds of studenls and communication professionals, I
have selected and (re)formulated what 1 eonsider the key coneepls and major issues
in discourse sludies.
In 19921 published myfirsl introduelion, DiS{ouNeStudies: An Introductory Text-
book. which was reprinted scveral times and published in four languages. I1 has been
the basis for this book but the enormous number of discourse studies publications and
new insighls of Ihe past deeade thai have been incorporaled here make this a very dif-
ferent book with over 300 sauree reCerenees and about 500 basie eoncepts explained.
This book louches on various disciphnes that prevlously did nOl consider dlscourse,
and the broader, dynamic. contexts of discourse in communlcation scienee.
At the TUburg University campus a foot bridge connects the ltbrary annex e-Iearn
ing center with the area of the lecture halls. In lecturing I have always eonsidered
myself a human bridge between aU the books, journals and WebSltes, and the quest ion
marks in the students' eyes. And when, in rare moments, these students left my lcelores
with more insight and more quest ions, I was gratcful to have been a medIum for this
lransfer. I hope that by studying this book and trying to deal with a11 the quest IOns and
assignmenu SludenlS will ealch a glimpse ofthis joy oflectunng. In my opmion this is
the best guarani for a frul tful future in activities eoncerning discourse.
I would like to thank my eolleagues in the Discourse Studies Group al Ti lburg
Universi ty, especialty for Ihe opportunity to do research in the USA as a Fulbrighl
fellow. Furthermore, they gave me Ihe chance to be confronled with many student
questions abroad while Ituring in different academic cuhures, such as at Japanese,
South Ameriean, South African and German universities. J guess that in past deeades
J have discussed dozens ofthe problems brought up in this book with over a hundred
colleagues from all over the world. Most of them are mentioned in the bibliography.
There ue two people who 1 would liu to Ihank personally. First, Bertie Kaal at
Benjamins Publishers for stimulating me to found a journal on Document Design,
which provides me wlth many new contaets, and for supporting me in -just writing
down the of my lectures. Second, Erie Daamen who was of invaJuable support
in assistmg me in and index work and all the time-consuming aspecl$ oftext
production. Ouring this work he became more than only an aSSlslant.
Acknowledgmt'nt5
I hopc that this course book will help teachers in discourse studies and relate<!. or
overlapping domains such as communication science, pragmatics, rhetoric, stylistics,
conversational analysis and design studies. Ir they consider this introduction a text -
book for their basic cour5e$ thaI can lake half of the work OUI of their hands, then I
have done a useful job. Most of alI, however, I hope that the bridge I built between
the insights In the phenomenon discourse and the students' interest will motivate
students to cross it. Perhaps after graduating they will even be stimulated to build a
bridge themselves. somewhere between Ihe academie island and the various profes-
sional areu where they end up. In any case, I would lilee to encourage both tuchers
and students to send me an)' comments or suggestions Ihat might arise from using this
textbook. All contribulions are welcome. Please see www.janrenkema.nl for contact
information.
Fall 2003, Tilburg University Jan Renkema
r
...
1 Introduction
A rough definition of discourse studies
Discourse studies is the discipline devottd to the investigation of the relationship
form and funclion In verbal communication. This short but rough definition is
the point of departure for Ihis book. The definition prompts the following qUe5tions:
I. What is meant by the relationship between form and function?
2. 15 it really necessary to have a separate discipline for the Ivestigation of this rela
tionship?
Answers 10 these two quest ions are given in this section. lhe aim and structure of this
book are discuSSfii in the next two sections.
Whal is meant by the rdationship between form and Consider the fol
lowing example of a fragment of verbal communication.
(I) A: Sly, Ihere's a good movie playing lomgb!.
8: Actually, I have to sludy.
A: Toobad.
8: Yes, I'm sorry.
A: WeH, I gum I don't necd to ask you if you want me to p1ck you up.
In this example, A's first utterance is in the form of a statement thai there is a good
movie playing that night. "fhe function of this statement, however, is that of an invi
talion to . B knows that A's stalement is meant 10 be an invitation. B could have
respondtd by simply saying. "Thats niee" or"'1 didn'l know Ihal." But B responds with
a statement in lurn expressing a need 10 slUdy that evening. B's response counts as a
refusal of Ihe invitation. 1\$ Slalement of regret shows that this interpretation is nol
mere conjecture.
In this fragment the form statemeHt has the function of an iHvitatiOH (first
ance of A) and a rtfusal thereof(first utterana of8). 8elow is another example: a pas-
sage from a concerning a newly built office complex and the same passage
in a slightly different form.
1. Introduttion
(2) a. 1he is situaled in old city The archlttural
firm of Wilkinson and Sons designed it.
b. is siluale<! in old city cenler. lt was designed by
architectural firm ofWilkinson and Sons.
lhe aClive voice is uscd in the $eCond sentence in (2a): firm designed the new
office.
M
Whereas in (2b) a passive variant is uscd: "lhe new offi ce was designed by
the firm." What is the difference in function these two In the
aetive form the aecent is on the firm that provided the design. In the passive form
the office complcx is elaborated on. \'{hen different forms are uscd for getting across
approximately the same contenI, they often lead 10 differences in function. The aim
of discourse: studies is to provide an explanalory description of the intricate relations
forms of discourse elements and their functions in communication.
The second queslion is more difficult 10 answer. Why should there be aseparate
discipline studio? To rnany researchers the best answer is thalthe investiga-
tion of the relation belween form and funclion rcquires contributions from different
disciplines such as hnguistics, literalure, rhetoric, stylislics and pragmalics as weil as
other fields concerned with verbal communication such as communication seience,
psychology, sociology and philosophy. Discourse phenomena cannol he Sludied
adequately from jusl one of Ihese perspt"tives. Becausc the concepts dealing with
these phenomena are taken from many disciplines. a common ground is nessary.
Discourse studies is Ihis common ground. lt serves as an inler- or multidiscipline that
enables different research schools to have the nessary interaction so that specific
contributions can be made to research into Ihe relationship between form and func-
tion in verbal communicalion.
1_2 Aim and structun of this book
The alm of this book is 10 familiarize the prospecti\'e sludent with the mosl important
concepts and the major issues in the field of discourse studies. Knowlcdge ofthe basic
concepts will serve as a scientific Mtoolltit" that Ihe student can use in advanced courses
in dlscourse sludies. Tbis inlroduction is also as a to
readmg in handbooks on different discourse topics and 10 studying research results in
scientific journals on discourse.
Thls book consiSl5 of chapters spread over four parts. The is
inspired by Ihe metaphor ofthe student who is supposed to be a seientific
journey. After this inlroductory chapter, Part I provides information about the basic
characteristics of Ihis ,ourney through the diverse landscapes of discourse studies.
Part I I invite5 sludents to backpacks with some travehng material.
1
1.l Alm and struttureofthis book
In Part II [ the differenl ways of mak.ing a scientific journey are prescnted In Part IV
some specific domains of can be chosen.
In ParI I, Chapters 2 and 3 provide a general orientation towards the field. The
focus is on the last part of the rough definition given earlier: discourse sludies is the
study of communication. Chapter 2, Communication as aClion, is Ihe most
philosophical chapler in this book. [I tries 10 answer queslions likt: What 15 (ver-
bal) communication? What are the principlcs governing Ihe usc of the instrument
What are the strategies Ihat are brought to bear when we communicate?
Verbal communication is prescnlcd as performance of acts which must have same
for partners involvtd in il. This chapter stimulatcs the student to think of
what verbal communication i5 aboul. Chapter 3, Discourse in cornmunication, focuses
on discourse as part of the situation in which pcople communicate. Discourse is no
discourse at all without asender and a rdver. Therefore, discourse: cannot be stud-
icd adequately without Ihe discourse situatIOn being taken inlo account. Within Ihis
framework two basic discourse queslions are an5wered: How can discourse besl be
studied in a perspective in which forms are re\ated to funetions in a discourse silua-
lion? What makes astring of sentences or a cauple of unerances
In Part 11, Chapters 4, 5, 6, 7 and 8 presenl the basic conceplS for studying dis-
course:. Chapter 4, Discourse Iypes, gives an overview of the variety of forms of dis-
course: such as wrinen (a tax form), oral (3 conversallon), electronic
(an e-mail), eie. It also discusses the allempts thai have heen made 50 far in putting the
main varieties into sorne sort of model that reflects the eSSoential differences betwccn
them. dassic inlriguing quesllon behind this modeling is: Whal exactly is a dis-
course Iype? The discussion on construcling a model is followed by the question of
how to sludy new electronic communicalion situations in which combinalions of
modes (oral, wrilten and visual) are Chapter 5, Structured conlent, prescnls the
approaches 10 discourse wilh rcspect 10 the structuring of the message content. How
can this structure be described for the different levels or Three levels are
distinguishcd the global struclure (the discourse as a whole),lhe mesostruclure
(the study oftopics and themes) and the local slructure (the smallest meamng units).
Ihe central question in Chapler 6 is: What are Ihe formal ties Ihal keep Ihe different
content elements together? Dcscending from a global an overview of dif
ferent knols and links 10 connecl content dements is given. Special allention is paid
10 techniques of referring back- and forward in discourse and 10 so-callcd discourse
relations and their markers. Chapter 7, Contextual phenomena, deal s wilh discourse
elements from bolh the production and the perception side, which are directly linked
to the conteXI. Examples are the fact thaI the rneaning of "l" depends on the person
using It or Ihe possibilily of putting some information more into the back- or fore-
ground in order 10 pmduce a special cffecl on readers or liSleners. However, link

I. Inlroduction
not only from discourse to context, it goes the olher way round as weU. We always
deal with discourse on the basis of knowledge and attitudes thai we already have. We
are no black boxes. Otherwise we would all give exactly the same rendering after hav-
ing seen the same movie: neverlheless Ihese renderings do dfer in conlent.
In Chapter 8, Style, the last chapter of Ihis first part, an overview of stylistic vari-
ation in discoursc is given. After abrief discussion of dassieal rhetoric,lhe concepts
of style and register are darified. Special attention is paid to stylistie phenomena that
can euHy be studied by students and 10 stylistie research that highlights the different
manifestalions of seemingly the same messages. Consider, for eumple, the different
renderings of one movie again, now focusing on the differcnces in formulation.
Part I11 deals wilh fourcenlral modes of communicalion. In everyday life we can
"jusllalk- or make conversation inluilively wilhoul knowing Ihe outcome. but we an
also use language intentionally 10 give information, to tell a story or to try to convince
someone. Chapter 9, Conversation analysis, gives insight into a more sociological
way of discourse studies: conversation as a kind of glue betv.-een the members of a
community or a society. Chapler 10, Informative discourse, focum on Ihe readability
of information and the improvemenl of documents. In this chapter old melhods to
meuure readability are dealt wilh, as weU as Ihe notion of discourse quality in a more
contemporary view. Chapter 11, Narratives, starts with a more literary approach to
discourse and illuslrates how narratives are studied from Ihree different perspectives:
sociolinguistic, psycholinguislic and organizalional. Chapler 12, Argumentation and
persuasion, starts with approaches to analrzing Ihe validlly of reasoning in everyday
language and presents a social-psychological framework for studying the way in which
discourse can be persuasive. Here the emphasis is on stylistic elemenlS.
In Part IV, Special interests,lhe lhree most important domains of discourse studies
are presented. In Chapler 13, Discourse and cognition, the focus is on what goes on in
our brain during Ihe production and Ihe perception of discourse. Cognitive psycholo-
gists have done extensive research into modeling the way we speak and listen or wrile
and read. Several oflhese models and some highlights of current cognitive approaches
are presented here. Chapter 14, Discourse and institution, focuses on the institutional
aspects of discourse withm the sociological approach. Some key publications are pre-
sented, dealing with institutions such aslaw, health and media. Chapter 15, Discourse
and culture, presents thc major lopics in the study of discoursc from a socielal point
of view. The main quest ion is: Can discourse tell w; something about the way in which
the producer views the world? ThlS is made more concrete by addressing quest ions
such as: Can discourse analysis reveal something about power relations in sociely or,
for example, the place of women in masculine cultures? These lypeS of que'tions are
of special importancc Ifl the study of intercultural communicalion.
1.3 The ofthe material
1.3 The prese:nlaon ofthe material
The material in this book has been organiztd to serve as a first introduction to dis-
course studies at university level. Inherent in the interdisciplinary nalure of the field
of discourst studies is the fact that each phenomenon can bc looked at from different
viewpoints. Moreover, the danger ensts oflrivializing theoretical concepls, as thcy are
taken out of their disciplinary context. Special attenlion wiU therefore be paid to the
origins of key concepts in discourse studies.
Inter- or multidisciplinary discourse studies arose during the t980s. However, it is
rooted in dassical rhelorie and language philosophy and in dassic psychological and
sociological studies from both the Anglo-American and Ihe European traditions. 1t
is for that rea50n that relatively much attention is given 10 dassic or impresSive land-
marks in the field of discourse studies.
When dealing with Ihe conccplual arsenal, examples of scientific applicalions are
given whenever possible. The research exampleschosen are not always Ihe most ret:ent
ones. In Ihis book attention is also paid 10 approaches upon which conlemporary
developments are based In the bibliography almost half of the references date from
bcfore 1990 as in Ihls introductory lextbook Ihe focus 15 on concepts and approaches
that have proved 10 be soundly based and not just trendy. After sludying this book the
student will have most of Ihe fundamental apparatus 10 do his or her job. The index
al Ihe end of this book., containing about 500 entries, is a good basis for studymg the
most important concepts in Ihe field. The index entries only refer 10 the page on which
the concept in quest ion is most elaborately explained. This makes it a concise
to the major concepts in discourse studies.
Obviously, an introductory work cannot delve into discussions about defi-
nitions of key concepts or elaborate on issues. For sludents who wish 10 study more
spedfic topics. each chapler or se<:tion is accompanied by a list of suggested readings.
The main aim of Ihis bibliographical information 15 10 indte the reader 10 study Ihe
dassic or key publications in the fjeld. Books comprise about 80% ofthese refercnces.
The other 20% consist of what are considered seminal articlcs. All Ihe references in
each chapter's bibliographical information are lisled at the end ofthe hook, wilh refer-
ence to the specific stetion they refer 10.
Each chapler ends with queslions and assignments. These are meant to Slimulate
reeclion upon and discussion of seemingly unproblematic topics. which it is hoped
may encourage students 10 initiate reasoned arliculallon of their own aSlomshment
aboul (mis)communication. The key at the end of thls hook provides answers to these
queslions. Thc assignmenls can be seen as proposals for studenlS 10 work on in the
Iibrary at Iheir own campus. The menl 10 whieh the assignments are to be worked
oul depends on conditions of time and curriculum and on special wishes the leclurer
5

I. Introduction
may have. The number of questions and assignmenls ffiay vary pU section. depending.
for example. on the length of the section or the nature or the subject thai the scction
deals wilh.
Questions and assignments
Questions
1.1.1 upl.ain in yourown wonls wlUI discourw nudin is.
1.1.2 uplain in terms oHOf'lTI and funclion wha! i$ going on in the following fragment of dinnercon
wn..lItion.
A: Could)'Otl pass IM
B: Of COUfW. (8 continues udng withool the salt.)
1.1.] Dncribf: lne dift'errlKe5 in (orm aod funClIOn between the following two
a. A general praclitionerat our hralth his practice yesterday after loul demon-
stntions.. Hewas susptfll of moInting patients.
b. A general practitioner at our l'Iealth (cnler, who was suSpecltd of molesting patients, closed
his practi )'t1terday .nef" Iotal demonstrations.
Assignments
\.1.1 One joum.ll issUC' on disoourw analysis can conum _ral p.apel"li that, aa;ording 10 lheir
authOfS, all have todo with discoun;e. uplaln how the following subjects could fit the definition
of discourw studies that is glven in th.e introductory chaptuorthis book. This usignment is
by a p.ass,age in Johnslone (2002).
1. Descriptive terms uU'd ohhe accLl5e<! in the media cOYerage or 01 murder trial.
2. A discussion ofdiffereOCfl English and 'apanese.
]. An analysis of expressions oridenllty in Athabaskan (Nalive Ameriun) studenl writing.
4. A diSCU5sion orwnnets by Shakespeare,
5. A paperabout tlle epitaph oflhe spiritual masterof 01 sec! orMuslims.
6, A discLission on wnether the pronoun 1 should appearin f'omlal writing.
7. Astud)' ofpolitica.1 debate.
1.I.a The bibliographieal infOmlatlOO ofthts chaplercontains a lin oflhe most known JdUmals in
lhe field or discourse sludies. HoweYer, Ih.e notion dlSCOIIl'Sffrequenllyoccul"li in allsorts ofjournals
not speaficalty linked to discouIW studies. Same eQmples are an articlt on the mmmunitalion
I
8ibliographkal information
sltills of peopIto wilh dementia (in lhe JOIII1IOI 0( CommunKOlIOI1 0is0I'dtr), one on tut mmprehension
in reb.tion to children's narratiYes (in the BntIr Journal of Drudopmtnttll Ps!/dIoIo9yJ and research Into
Of'ganizalional changes as discourw (in the AcDdfmy 0( ,/01111101). Select any issue or one of
the journals menlioOf'd in the bibliographiea' information of Ihis chapterOf' find a JOUrnal mmpa
rableto theones mentlOOed alleM. LooIc at the summaryofone oflhe artiele-$ and uy 10 determine
whether it falls wilhin Iht definition ofdiscourse studies. Support )'OlIropinion.
Bibliographical information
1hc artlCle DiKOUru onal)'S1J by Zdhg Harns (1952) is V1ewed by many u (he Ilanmg pomt of
discourse IILldleS. Harris WB the fim 10 LISt tm term dlS(oliru anolysis In Irllele, He
discussed M ad\'eT!1SaIltll1 IUI by anaIyr.i118 w wa)' in which smlmca Ire hnknl and IM way In
which Ihe tnl correlates Wlth .ocinyand cullure. A sahenl ddall 15 the faCllh"'l III the firsl fOOlnOle
Harru thanlu hl5 research usislanl, who rrukd. Llp developing lhe m051 IIlflLlrnliallheory on eie
menUi wIlllU! a smlmce: Noam Chomsky.
In Ihe decades folloWlng Ihe year in which Harns wrole hislrticlc, a VlSI qLlanlll)'ofbooks and
artieies on dlScoune 51udlts has becn pub/Ished. So it is obviow thai Ihis tUlbook i$ not the onl)'
mlroduchon 10 IhlS fie\d of research. On the conlrary, one of lhe IIIllS orlhll book IS to scrvc u a
ufrrence gUlde to a qualilative scleclion of other handbooks .nd rraden.ln Ih,s firsl bibliographieal
OVttView morr general works Ire menlioned wheuu in the followuig chapeers lhe referrnces are
lr"I!ln gured 10 5peC16c upeclS of di.5cour5t lIudies.
Scvenl,mportantlnlrodll(lions 10 d'KOUrse sludies were publlshed m Ihe nrly 19805. Prom,
nenl GeTman.langllagt pub/KatIons Ire Kalverkimper ( 1981 ), Co5enu (198 1), Sowmskl (1983) and
SchemeT (1984). 1M n\O$l Wlddy used Enghsh-linguagc pLiblia.lions of thai time: are !Je BeaLi
grande and Dresslcr (1981), Brown and YLile (1983), Slubbs (1983) and Coole (1989). In thc plSI
dccadr nnow ncw mtroductionl werr publishcd, which IInderiina W V11a1lly of IM bro..d field
o( discourw nudies. 1hc n\O$l Important publicallON o( recrnt dalr indude the foIlowmg: Salk,e
(1995) 11. workbook, wilh UCfCI$e5 10 dctecl MOUS discoLlrsc pheoomena: jawonki & Coupland
( 1999) glve a colJechon of somc Ihirt)' pauages from kcy pubhcallON on discourse stLidles; Goad)'
(2000) provKles a dear mtroduclloo Wllh practical exercises ror dcvdopIng cnhcal awareness oflhc
rdationship Ixlwttn Iut produCllon Md CO/lSumption: John5l0ne (2002) Ipproaches d,SCOLirse not
as I doopbne 001 u I rescarch mclhod. coonlIng II Wllh II.l lopIUlhal $hapr 11, IlKh &S world,
ILidiences, mrdlum and purpoae.
lhcre au also lIllroductionito $pecI1k p;lIt$ ofthe research fidd and boob; Ihll indirectl)' have
ttus funclloo. 1hc most ligm6canl ones dlJllIIl from 1980s IOd nrly 19901 dnJ Wllh the an.al)'$lS
of converJlllOn (Edmonson, 1981: Hrnne aod Rehbock. 1982: MClaughhn, 1984, NoUingcr, 1991)
and Wlth 51ylistiQ (McMenamin, 1993). Reeeni publlClltions emrha.wu the 5OCJOcullura] aspect.s of
ducoune 1tudJes. for enmple, GH (1999), IOd IWTlIllYU, for aampk. '" Gouoos
7
I. Introdllction
(1997) and Toolan (2001). For good pubbcallonun discOllrK anaIyIII an gtfItI"*I CouIthard (1985),
Nunan (1987), McCarthy (1991), Mann and Thompson (1992) and Hocr (200 I).
In to booki, numefO\l.$ journab (mm various nsearch traditions havc bcen foundC'd in
d,5COUrsc stud'e$, Bclow an: tIM: most wilkly known tlUes. consldcrN the cort: journils
Irt: marted Wlth In asterisk.
Applicd Lingulslics
Cogrullon
LlOjljuistiCI
CoUege Composillon Communication
Computational Linjljuistics
DUcoune and Socicty"
DiscouIR PmcrMCS"
lMcOUrK SlU<lles'
Information and Document Design Journal
Human Commumcatlon Research
IEEE tranSictlOns on profasional communicllion
Infomulion and Manajljt'ment
Journal of 8usmess CommunlClltiOl\
Journal o( Commumcation
journal of
journal o( I.anjljuage and Social Psychology
JournalofPragmiitlCs"
Journal o(SemanliQ
journal of SociollOgulstics
Langw.ge and CosnJllve ProcesKS
Langw.ge and CommuniOltion
PragmiltlQ
Tu<
Wrillen CommuniCiltion"
Trun van Dijk 11 by many u tht founding fathcr of contemporary dllCOUr5e studies. 11e
founded a number of indudmg the above-mmuoned penodx:als Tul, DucOUf'Se /IM Soady
and Dtscoutw S/ud,a Morcover, he IS known fw ed'lil18 tWIl handboob m chscoune
Iwne/y 1M lUmdbook ofDuanme AnuIym (I98S) and DUcourw Slud.a; A Mullld,sapilNJr)' I",rodwe-
hOIl (1997). Two other notrworthy hartdboob dall", (mm thc p.st dcode an: Schlffnn (1994) and the
waustlvc work by Schiffrin, Tannen & Hilmilton (200 1). A lH'W library on di$COUf'SC studies $hould
undoubtcdJy mcJudc Ihcsc four worb.
PART I
General orientation
2 Communication as action
2.1 The Organon model
In thc last decades. the slogan is has CQmt ioto fashion, bul
in fact Ihis view cf communication i5 mort than two thousand rean old. Oe of the
earliest works on language. Plato's Cralylus (a dialogue on the origin oflanguage wril -
len in about 390 B.C.), describes speech as a form of action and words as instruments
Wlth which actions can be performed.
The German philosopher and psychologisI Karl BUhler was refc-rring 10 this work
when he described languagt as a 1001, which propie use in order 10 commu
nieate with oe anolher. Bhler's Orpnon modd (19J4/ 1990) has had a major impact
on the way languagt is dealt with in discoul'St studies. BhieT stale<! that asound can
onlr qualify as a linguistic sign ir a thrtc-fold relationship exists connecting the sound
10 a a recciver, and an ob;1 thai is being uferred 10. ParaUd 10 Ihis Ihlft-
pronged relalionship, each linguislic sign (5) has threc funClions simullaneously:
I. A sign funclions as a symptom as it sars something about asender, for example,
whether the sender is femate or male or what the intention of the ulterance iso
2. A sign is a symbol beause it refers to objecls and states of alfairs.
3. A sign &erves as a signal because a receiver must interpret it or uact to what has
been said.
This three-part division can be ilIustrated with any utterance. Below is an example.
(I ) Have you hurd thai strange Itoryaboutlhe drunk who decided to play barber
and cut olfhi5 friend's
symptom
I i
Figure I. Buhler', Organon model
symbol
IIIAnal
rc,vcr
"
1. CommUnicalion lS actIOn
By aslting Ihis quest ion, tht speaker indicalcs that he want! information {rom Ihe
person who is being addrtsSN. By using thc word Ihe speilir is also express-
ing an opinion. This is Ihe symptom aspect. In Ihe utterancc a referenee is made 10 a
story. areal even!. lhat is thc symbol aspect. The quest ion is an appeal 10 a listener. A
lislener is nOI expectw 10 JUS! answer "yes" or "no and change Ihe lopie. Somclhing
along Ihe Hnc o("No, tell mt about it" or "Yes" followed by the listener's own Teaclion
is expccted. This is the signal aspccl.
In this chapter Ihe {ocus is on thc basic assumptions of Ihe Organon model,
namely, thaI languagt is an instrument with whieh objectives can bc achieved and
thaI Ihis instrument canno! considerfii 10 bc separate (rom speakers and lislentrs, or
writcrs and in ptrforming eommunieative aets.
Languagt, and themort is a two way instrumtnl, an instrumtnt for a
speaktr and a listener or a wriltr and areader. Or as Ihe Danish linguislic philosopher
0110 Jespersen wrole in the introduetion to his Philosophy ofGrammar (1924):
(2) The essence oflanguage is human aetiVity - aetivity on the part of one
individual 10 make hirnself underslood by another, and activity on the part of
that other 10 undersland whal was in the mind ohhe first.
Iflwo parties use an instrument for an then such an aetivity can only suc.
cessful ifboth partics adhere to general rulrs or prindples and thtreby utilize certain
slrategies. This ean bc ilIustrattd with a non linguistie example. If two people want to
hang a painling (activity), they use a hammtr, nails. and a ladder (instruments), and
ther have to coordinate their actions. Therr will have 10 some form of cooperation;
while one is standing on the ladder, the other ean hand the tools to the first, ete. Rules
coneeming politeness will also have to follo .....ed; while one person is on the lad der,
the other should nOI try to push the first off. One general prindple of colleetive aeliv-
ity i$ and an olten-used strategy to achieve this is poIiteHw. This is also
ITUe in Ihe ea.se of verbal comrnunieation. On the basis of Ihis cooperation principle
and guidtd by so-ealJrd politeness strategies the comrnunicators have to perform thrir
eommunicative aets. Sul what precisdy are Ihose eommunicali\'e aels? Tht theory,
catltd oet theory, provides an answer 10 Ihis quest ion.
:1.1. Speech aet theory
In speech aet theory, language is seen as a form of acting. This theory sterns from the
school of philosophy that is ealled ordinary language philosophy. The proponents of
this $(hool, whieh Hourishtd tn England in Ihe middle of the last century, wanttd to
analyze philosophieal problems by looking at ordinary language and trying to ascer-
1.1 Sp<h acttheory
tain what insights it could offer into reality. For example, the ethieal question of why
human aetivity is judged to be good or bad, demands that Ihe way individuals apolo-
giu for bad also be studitd. An apology is an aet in which a justifieation is
given. Sy sludying how people perform speech aelS such as apologizing, promising,
ordering. Cle., "philosophers of ordinary language wIshtd to contribule to the
solution of philosophieal problems.
Speech act Iheory has had a slrong influence on Ihe field of studies as
Ihis theory focuses on Ihe quest ion of what people are doing whcn they language.
Consider Ihe next example. There is a strik.ing difftrence Ihe following Iwo
sentrncrs.
(3) It's rainiog.
(4) I that I will give you one hundred dollan tomorrow.
In (3) a statement is made that may or may not true. As for (4), ho .....ever, it is not
possible to say that il islrue or that it is nOI true. With verbs sueh as 10 promlK (in the
first person), nOI only is something being saidi importanlly, something is being
done. In (4) an act is being performed through an Ullerance. By saying promise ...
a promise is made. But saying "'1'5 raining" does not make il rain.
The English philosopher John Auslin (1976) used the terms COrlstative and peifor-
mative to Ihis In constativcs, such as senlenee (3), somelhing is
stated about realitYi in performatives, such as (4), an aet i5 performed by Ihe ulteranee
itself. Auslin was not suceessful, however, in cstablishing eriteria for describing the
differenee these Iwo concepts. 11 can, alter all, argued that an ael is being
performed in the case of eonstalive utterances 3S wtll; a warning given or a statemenl
made as in the ca.se of (3).
This lcd Austin to the conclusion that a11 expressions oflanguage must be viewed
as aets. He distinguishtd Ihrer kinds of aelion within each utteranee. Fint, Ihere i5 the
loculion, Ihe physieal aet of producing an utteranee. Sccond, is the iIlocutiol1, Ihe
aet that is committed by producing an utteranee: by uttering a promise, a is
made; by unering a threal, a threat is made. Thlrd, is the Ihe pro-
duclion of an effect through locution and iIIocution, for example, the exeeution of an
order by Ihe addressee.
Consider anolher example. In Ihe Stalement "There is a draft Ifl the locutlon
is the produclion ofthe utteranee. Depending on the situation, the ilIoculion could bc
a request. an order, a complaint, Clc. The perlocution could be Ihat a door or window is
dosed or that the addressee replies that he is nol a servan!. It is important 10 emphasiu
that the 10 an illoculion, the so-calJtd uploke thatleads to a perloculion, ean
differ dependmg on Ihe situalion. Below is an example offour different uptakes ofthe
same utterance.
'.
'.
1. Communication u action
T.ble I. V.rious upt.kcl of the "me utleranet
!.ocution orthc .pubr
lhcn:'I. good movie tomght
lhen:'1 a good movic lomght
Therc; a good movie lonighl
'Iht-rc:'I. good ltIOV1e 10m,11I
IIloculiOIl
[nvitahon
Advict
IUCUK
O/f ..
Uplalr.c by thc liSlcn(r
O.K.ld.go
O.K. I will SO thcre
N.,,,,rund
n.ru.",,'
In spttch aet theory the illoculion is the focus of attention. Languagc philosophen
havc tried 10 givc an ovcrvicw of all possiblc ilIocutions. from assertivH 10 ftqU6tS.
(rom promises 10 udarn.lions. This, howevcr, proved 10 bc a >,efY dlfficult lask,
b.uu: il is by no means cleat wh.t ex,clly thc characteristic differences bctween thc
proposcd illocutions Fot aamplc. a could bc a Ihreat in the locution ",
promise.l'll gct you!" First. thc phenomenon illocution itse!fhas 10 bc studied.
Among thc intriguing problems with iIIocutions, Ihue is ont thai has drawn
cial attention, n.mely, the issue of successful iIIocutions. It is easy to Stt that certain
minimum mUSI be mel for an iIIoculion 10 be succmful. If anyone olher
Ihan a church leader excommunicales someone. then the act of excommunication has
nOI bet'n execuled. If in a casino someone al the roulette lable suddenly calls - Rien ne
va plus" rNo more bets!M or -Game overn. this cannol be construed as being the
illocution refusing if Ihis person is not the dealer.
The philosopher lohn Searle (1969) formulated four felicity conditions that iIIocu-
lions must meet. These four conditions are illuslraled below using Ihe iIIocution to
promise.
(5) Felicity conditions for 10 (speech act)
a. the propositlonal contenl
In the case of-promising': the act thai the sptaker commlts himselfto (Ihe
proposition) must be a future aCl 10 be carried out by Ihe sptaker himsel(
One cannOI malte a promist for SOffi">ne eise or promJst 10 do 50melhing
thai has already been done.
b. the preparalory condition
lhis condllion concerru lhast circumstances thaI are essential for Ihe
uplake of an illoculion as IM inlended iIIoculion. In the cue of promising.
these ClrcumSlancn wouId require thalthe content of Ihe promise i.s not a
matter of course. Another preparatory condition 1$ that the prornlse must be
advantageous 10 Ihe addressee: one cannot promise something tha} i550lely
dlsadvantagcous.
l .l ael theory
c. the $incerity conditlon
The speakeT must honcstly be willinglo fulfill the promise. Even ifhe is nOI
willing, he can be held to his promise.
d. the essenlial condition
This i5 the condition Ihat separates the illocution in queslion from other
illoculions. In the case of-promising-, this means. among other Ihings, thai
the sptaker takes upon hlmself the rnponsibility of carrying oul the aet
staled in Ihe contenl of Ihe promise.
Searle uscd thest felicityconditions to show that the suceessful exchange ofillocutions
is also bound by certain rules. In terms of form and function, this means that a form
can only acquire a valid function given ccrtain conditions..
Anolher approach is provided by the German sociologist Jrgen Habermas
(1981). According 10 Habermas, sptakers claim that Iheir illOCUlions are valid. In the
cast oflhe illocution predictlrlg, for example, the speaker claims that Ihe statement will
come ttue in the fulure. In the case of congrtllUlting, the claim to validity is based on
an expression of emotion on the part of Ihe speaker, namely, that the congratulations
are In the cast of ordering. the sptaker bases the claim to validity on assumed
authority to issue the order.
Habermas based these validity claims on Bhler's Organon model and the thr
aspects that can be distinguished in language signs: symbol. symptom and signal (s
Section 2.1). Through the symbol aspcct of an ullerance, a claim is made as to the lrulh
of the statement as in Ihe prediction example above. Through the symptom a5pect, a
claim is made regarding sinccrity, see Ihe congratulation example. Through the signal
aspect, a claim is made regarding legitimacy as in Ihe order example. In Habermas's
view, an illOCUlion is only successful when the addressee acknowledges the claim to
validity. Take the example of a teacher asking a student Ihe following queslion:
(6) Could you bring me a glassofwater?
The student can refusc this requesl as invalid on the basis of allihree aspects.
(7) Dispute of the validity of (6)
a. symbol aspect: truth. (The contenl ofthe statement does not correspond to
reality.) -How can you request soffiething lilte that? lhe nearest faueet is so
far away thatl would never be able to make it before the end of dass.-
b. symptom aspect: slcerity. -No. )'Ou don't really want any water. You' re JUS!
Irymg 10 make me 100II bad I front ofthe other students.-
c. signal aspecl: leglllmacy. You can', ask me to do something Iike that. I'm
not to feteh and carry for the tcacher!"'
'5

13
2. Communicatioo as action
Table 2. Basic ilIocutions according 10 Habermas
ASpt of thr uttuance
symbol
symptom
signal
Claim 10 validity
trum
sincerily
legitimacy
of illocution
constativl"
cxprcloSivt
regulative
Using Bhler's three-way division, Habermas defines threc main types of ilIocution:
constatives (with a symbol aspecl), expressives (with a symptom aspect), and regula-
tives (with a signal aspect).
The iIIocutions claiming and dacr/bing are examples of constatives; promising and
congratulating of expressives; inviting and requesting of regulatives.
lUocutions in discourse
How does the more philosophical speech act theory in the previous seclion contribute
10 the study of discourse? First, it can provide insights inlo Ihe requirements thaI a
form (the loculion) must meet 10 ensure that the illocution and the intended uptake
take plaee. This illocution serves as aprerequisite for the aehievement of the perIocu-
tion the speaker or writer has in mind. Second, this theory ean serve as a framework
for indieating what is required in order to determine the relationship between form
and funetion, between locution, on the one hand, and ilIocution and perlocut ion, on
the other hand.
There are a numher of eases in which the utterance itself, the loeution, provides
an indieation of the intended illoeutions. lohn Searle (1969) ealls t hese indicat ions
IFIDS, illocutionary force indicating devices. 'I'IOS include performative verbs, word
order, intonat ion, aeeent, certain adverbs, and the mode of the verb. If an ' FID is
present, the utteranee is said to have an explicit illoeution; in all other eases the
utteranee is said to have an implicit or indirect illoeution. Below are a few examples
of explicit itlocutions.
(8) I request that you put out your cigarette.
(9) He is putting out his cigareue.
(10) Is he pulting out his cigarelte?
(11) An: you going to put thaI cigareue out or nOI?
(12) Would you please pul out your cigareUe?
In (8) the performative verb request" makes the illocutionary intent explicit. The
differenee in word order between (9) and (10) is indicath'e oflhe iIIoculionary intent,
1.3 III()(Ulions in discourse
in this case "statement" and "question", respectively. Ascending intonation and an
aeeent on the word can also eonvey an expression of surprise. In ( 11) the tag
"or not" is indicat ive of the imperative charaeler of the iIIocution. In (12) the mode of
the verb indieates that this is arequest; the adverb depending on the intona-
tion, ean make this request ei ther caut ious or insistenl. It is also possible to eonvert
(12) into an order by placing a special aeeent on "please" and
It should be noted that I"IDS do not always provide a definitive answer regarding
iIIocutionary intent. Tbe IFIO if ... the" in the following two examples would suggest a
conditional promise, but in fact only (13) contains a eonditional promise.
(13) If you take the garbage out, I will give you a heer.
(14) If you keep this up, you will have a nervous break down.
In (14) the IFID is not the only relevant factor; more background information is
needed, specifieally that a nervous breakdown is dangerous. Otherwise, it is impossible
to deduce why (14) is generall y seen as a warning. If ro much additional information
is needed to determine the function of explidt language utteranees, then it should be
dear that this is even more diffieult in the ease of implicit or indirect utterances. See
the example in Table 1 of Section 2.2 again, is a good movie which
eould funetion as an indirect invi tation or excuse.
Mueh knowledge is needed to link the right illocution to a locution. Consider the
next example in the form of an interrogative.
(15) Can you stop by in aminute?
Why is this interrogative generally interpreted as arequest? A request can be identified
by the foUowi ng felicity conditions:
(16) Felicity conditions for requests
a. the propositional content
The content must refer 10 a future act, X, which is to he carried out by the
addressee.
b. the pn:paratory condition
I. The addressee is capable of executing X and the speaker helieves thaI
the addressee is capable of doing it .
2. Jt is obvious 10 both conversational participants thai the addressee will
not perform the act withoul being asked.
c. Ihe sincerity condition
The speaker aClually wants the addressee to do what has been requested.
d. the essential condition
The ulterance serves as an attemptlo persuade Ihe to excCule X.
.,
,.
1. Commurllc;ltion u anion
(22) A: I am out of petroL
B: Ihere is agarage round Ihe corner.
A can deduce from B's reaction thaI B means that Ihere is agarage around Ihe corner
that is open and seils gasoline. S, however, has not mentioned thcst (acts. A can ooly
make these assumplions ifhe assumes thaI B is acting in accordance with Ihecoopera-
live prindple aod is adhering 10 Ihe maxim of relevance.
In discourse studies Ihe principle and its muims are olten rtferrt"<! 10
as they provide a lucid description ofhow listeners (and readers) can distill informa-
tion from an utterance eYcn though thaI information has not menlioned oUlrighl.
This is ofimponance 10 resurch on thc relationship bctwcen form and function.
Grice did. however, havt a number of additional comments concerning Ihe coop-
entive princ:iple. First. the muinu are only valid for language use that is meantto be
informative. This exdudes. for example. such calegories as debating and 5mall talk.
Sccond. there are. from the esthetie or sodal point of view. other possible muims.
Grice 5Uggesl$ the maxim "Be polite": Third. another prineiple is at work here. Con-
sider the quantity muim. An ovenbundanee of information does not necessarily
mean that it is Ihis muim thai is being violated, since it an al.so be seen as a waste of
time and energy and thus as a violation of same effieieney prineiple.
In addition the Grieean muims have beton eritieized for several reasons. Some
muims are rather vague. For example, how ean it be delermined whieh information
IS required (first muim of quantity)? The four maxims have beton as being
of equal importanee. but there are situations in whieh the muim of quality is more
important than the muim of manner, and vice versa. Nevertheless, Griee's cooperative
principle has had a great impact on discussions in the field of discourse studies. This is
probably because Griee showed with everyday examples that communieation, which
seems to enroll withoul rules, is organized by basic rules.
1.5 Relevance t heory
In the IiterOlture on Griee's muims special altention is given to the maxim of relevance.
One reasons for Ihis iSlhal it is uodear how it ean be determined whether a eontribu-
tion to a eonversation is relevant or not. A number of suggestions have beton made in
the direction of a dear descnptlon of relevance. It has, however. proved to be exeeed
ingly difficult 10 determine exaet1y when the maxim of relevance has been violated.
Regard the following example of a question and a number of possible answers:
(23) A: Where's my box of ehocolates?
8: a. Where are the snows of yesleryear?
b. I was feeling hungry.
e. I've gol a train to eateh.
d. Where's your diet meeH
e. The children were in your room Ihis morning.
15 Relevanee thCOf)'
Speaker A could react wilh surprise and ask why speaker 8 is suddenly quoting a line
of poelry, in the case of answer (a), or with was lalking about chocolates and now
you're lalking about the in the ease of answer (1'). At firsl sight. it sms
that B is not acting wilhin the constraints of the maxim of relevance.. However, if A
assurnes that B is adhering 10 the maxim of relevance,lhen any reaction B gives could
be construed as being relevant .
(24) a. B is not just quoting poetry; B is not really asking a question. B, by reacting
the way he does, is 5imply mahng dear that the ehocolates, llu the snows
of the palt, have gradually disappeared and that there is 00 good answer to
& quest Ion.
b. 8 is making dear that he has eaten & chocolates.
c. 8 does not want to answer the questIOn because he is 10 a hurry. Or, B is
evading the quest ion with an exeU5e; he knows more than he is letling on.
d. B is po5tponing giving an answer; first he wants to know whether or not A
should be eating chocolate.
e. B is suggesting that the children ate the chocolates. Or, B is suggesling thai
the ehildren know where the chocolates are.
Obviously, numerous other possible reactions for 8 are coneeivable. The main point is
that every reaction can be eonstrued as being relevant. lt is, of course, possible to imag-
ine eontributions 10 conversations thai would, at first sight, appear to be irrelevant. hut
these usually end up sounding like excerpts from a comedy rouline.
(25) A: Wouldyoucaretodanee?
B: I\:1love to. Do you know anyone else who ....-ould like to?
(26) A: (leacher) You should have been here at nme oelock.
B; (student) Why? Did something happen?
However, even in these examples, 8's reaetlon could be inlerpreted as being relevant
ir in (25) A is a walter or jf in (26) school docs not Slart until9:30. The problem now
is that it can be fairly objectively eSlablished when or whether the maxims of quality,
quantity and manner are violated, but 11 seems qUlte impossible to delermine when
"
"
1. Comrt1unication u action
an 00 longer counts as relevant . This makes it undear wha! thc nluc of thc
muim of relevance iso
Thc omnipresencc of relevance in communication has Icd 10 the Relevance Theory
of Dan Sperber aod Deirdre Wilson (1995). Ther look a different starting point, aod
made the relevante concepl the cornerstone of thdr view of communication while
abandoning the other Ihree maxims. Thai i5 10 ur, they tumed the disadvantage ofthe
vagueness of relevance 10 the benefit of a dearer thoory of wha! we mean by uoder-
standing each
A good start 10 getting their point is 10 realizc thai language in use is characterized
by whal is ealle<! indefuminacy or ul1derspiJication. We have a1ready seen that thc
example in Section 2.2. "Thefe is a good movie lonight': can be vague Of amblguous ir
oe does nOllake into account the discourse situation. This can be seen as a form of
undenpecilicalion. Bul even when the discoune silualion is known and the locution
is dear, the locution is often underspecified. A good example is Ihis notice often found
on the door of a lecture hall:
(27) Doors must be locked and windows dosed when leaving this room.
lhis locution can bear the meaning that a student must lock up whenever he leaves
dass for, a toilet break, but obviously no student will interpret the utterance in thaI
way. Almost anyone will understand that the notice only applies or is relevant when
at the end of the day the last lecture has ended and people leave lhe room without
returning. None ofthis extra information is incJuded in the thus underspecified notice,
and slillthis missing information is lilled in and as a resuh the utteranee is understood
correctly.
Sperber and Wilson argue that such ambiguities are dissolved in the right eontext
on lhe basis of the relevance coneept. While an addressee interprets the meaning of
an utteranee such as in (21), he assumes that it makes sense. that il is relevant and
that it forms a coherent whole. The addressu only seleets the relevant features of the
context and recogni7.es whalever the addresser communieates as relevant. But how
docs that work. attaching relevance 10 contributions such as the movie example or
example (27)?
First. It is importanl 10 recognize an utlerance as an aCI of "ostensive communi-
cation': i.e . an act of making something manifest to Ihe addressee and hdping Ihe
addressec to understand the meaning of Ihis. Second, communication must nol be
seen as jusl getting the thoughts oflhe speaker into the mind ofthe addressee but as a
means of enlarging mutual "cognitive environments": The cognilive environment is a
sei offacts perceptible In reality or inferable from knowledge aboul reality. Hearers and
readers makI' sense of (a piece of) discourse, they interprel the connections between
utterances as meaningful, drawing conclusions based on their background knowledge
2.5 Relevllnce 2]
of the world: they use their cognilive environment and, as a result of interpretlOg the
ulterance, their eognilive environment is enlarged.
Ir an addresser knows or at least can assess the eognilive environment of an
addressee by an aet of ostensive communicat ion, lhen he knows what kind of assump-
tions the addressee will aetulilly makI' in reaelion to an utterance. In the case o( (27)
the addresscr knows that the cognitive environment of an addressee in a lecture hall
will somelhing Iike: when Ileave dass for a toilet break, other people willstay lhere,
I will return in a few minutes and eontinue lislening 10 the leclurer. The addresser will
assess the addressee's reaetion 10 the utteranee as something llke: there is no need to
lock up lind dose the windows because of what I know about the situation. And thus
the addresser knows that a notiee as in (27) will suffiee in the given situation and will
only be judged as relevant in the right eontext: at the end of the day when the lecture
hlill is definitively lef!.
lhe theory of Sperber and Wilson added two important nOlions 10 studying the
way people understand each other: explicature and degree of relevance. These not ions
can be darified in discussing lhe following passage from a Mticket bUying
whieh is laken from CuUing (2002).
(28) A:
. ,
A,
.,
"
Weil there's a shuttle service sixty eUlOS oneway. When do you want to go?
Atlhe weekend .
What .... eekend?
Next weekend. How does that work? You just turn up for the shuttle service?
ThaI might be chnper. Then that's fifty.
The parlicipanlS have to interpret lhe verblli aets of the other as allempts to change
their mutual eognitive environment. This ostensive eommunieation is the input for
Ihe explicature, which is Ihe enricnment of the underdelermmed locution or the for-
mulation of the intended explicit conten!. It is like lilling in missing ....'Ords. Sperber
and Wilson eonsider Ihis explicalure a neeessary stage before making a
impliealure.
This explicature. this specification of underspecified ulterances, is ruled by the
principle of relevanee. Only that information is filled in thai IS relevant to the com-
munication situation. In Ihis conversation, B assumes Ihal A will under5land "At the
10 mean MNexl B assumes thatlhis underspeci fieation is relevant
enough. However, sinee A is going to seil a ticket, he nuds to verify if this is true. !U
last answer, MThat might be cheaper. Then thafs is nol a complete ans .... er. If A
had wanted to be more explicil, he could have said: "If you purchase a ticket now. you
have hooked aseat, which costs 60 euros. If you buy the ti cket when you turn up, I1
costs 50 euros." A, however, presumes B 10 be able to infer all ofthis and filJ in the
missing words.
..
l. Communic31ion u actIOn
Through this it can be made reasonable thai not a11 utlerance$ are
equally relevant aod thaI not all utterances are equally successful. Utterances cannot
be divided inlo relevant or irrelevant ulttrances. There is a deg!': of relevance. This
degr of relevance of an utterance i$ determined by two (aclors: contexlual elfeets aod
processing efforts. Contcx-tual elfecIS concern the war new information can ioterae!
with wha! is already known, i.e., everything that contrihutes to the addresscc's repre-
stntation of Ihe world. Processing effOT! pt:rtains 10 Ihe dfort of decoding lingulslic
information aod the elfort of accessing information in the contut 10 link the new
information 10. The degrce of relevance can then be described as folIows: the greater
Ihe contcxlual e/feet and the lcss effort it takes to create that elfect, the greater the rel -
evance is. So, in this example Ihe utterance by 8 At the has low relevance,
hause the information caonot be linked to wha! is already known. as can be sn
by I\s reacHon. The processing elfort can only be successful when the enct weekend
is known. Aod in this contexl Ihe last utterance by A has a high degree of relevance.
lhe new information cheaper- and "fifty- can easily be linked to information that is
already given in A's fi rst utterance.
So much for a more philosophical -inspired theoryabout the foundations of com-
munication. lhislheory has influcnced the analysis of mainly through the
concepl of underspedfication and the focus on the relation bnween discourse and
the situation.
2.6 Politeness theory
Notions such as cooperation and relevance are mainly valid for informative language
use. Language are nOI, however, always interesied in the elftive transfer of
information or relevance of an utterance. In the foUowing examples the speaker wants
the addressee to dose the door.
(29) a. Clem the dOOf.
b. Theres a draft.
c. Would 10u dem the door?
d. Would 10u so kind as to dem the
According 10 the maxims of the cooperatlve prindple, (29a) is sulfiden!. language is,
however, often used more indirectly, as in (29b). Somellmes certain politeness forms
such as in (29c) and (29d) are appli ed as weil.
An imporlanl source of inspiration in the study of politeness phenol1lena is thc
work done by Erving Goffman (1956). This sodal psychologisl introduced the concept
off ace. By this he meant the image that a person projecls in his sodal contacts wllh
1.6 PoIiteneu 15
others. Face has the meaning as in the saying "to lose face': In Goffman's opinion, every
partidpant in the sodal process has Ihe need to be appriated by others and the nred
10 be free and not interfered with. Goffman calls the need to be appredated "posi tive
aod the nd to not be disturbed "negative face':
Goffman wanted social inleraction, which indudes verbal communication, to be
studied from the perspective that participants are striving for stability in their rela-
tionships with others. Partkipanls in conversations should, therefore, not violate one
anolher's face. Refusing a request or reproaching someone is an action thaI can form a
threat to the other's positive or negative face. In the case ofthese threatening aets"
(FTAS), something is needed which will reduee the violation offace to a minimum aod,
therefore. preserve stability as much as possible. This can be achie\ed by using
work thniques': Examples are broad circumspect formulations of refusal5, which
make it dear thaI the request made is impossible 10 grant.
How does polileness fit into this approach? Politeness prevents or repairs the
damage caused by FTAS. The greater the threat 10 stabi lity, the more politeness, face
work thnique, is necessary. Just as there are two types of face, there are IWO types
of polileness. Face werk that is aimed at positive face is called -solidarity politeness;
this kind of politeness i5, for example, achieved by giving compliments. Face work thaI
deals with negative face is known as -respect politeness", and cao be achieved by not
infringing anolher's "domain- in the communication. Below are a fewexamples. When
a personnel manager has to turn down a}ob applicant who should nol have applied in
the first place owing to lack of education, this is an FTA that threatens the positive face
of the applieant, and Ihat of the manager. Fot this reason the pcrsonnel manager will
be more apt to write (JOb) than (30a).
(30) a. We do not understand why 10u bothered 10 apply.
b. We have some doubt$ eoncerning 10ur prior education.
In the following interaction betwn an inslructor and a sludent at the end of a tutor-
ing session, the second variant is more polite as it is less damaging to the insttuctor's
face and that of the studenl .
(31) A: J've tried to explain this as clearly as possible. Now I have to leave as I have
another appointment. I hopt' that the homework will be easier next time.
8: a. I tlill don't understand Ihe material.
b. If problems should arise, is it aU righl ifl stop by tomorrow?
Inspired by Goffman's werk, Penelope Smwn and Stephen Levinson (1978) dcveloped
a theory on the relationship between Ihe intensity of the threat to face and lingui st i-
cally realiz.ed poli teness. The intensilyoflhe threat 10 face is upressed by a weight (W)
that is linked 10 an FTA. This weighl is the sum of Ihree social parameters: (a) thc rate

I. Communicllion IS action
ofimposition, whjch is thc Mabsolutc weight" of a parlicular aet in a specific cuhure; (b)
Ihr sodal distance thr speaker aod Ihr person addressi; (c) thr power that
Ihe person being spoken 10 has over the speaker. The term absolute weight refers 10 the
faCllhat, for example. the request -Mar I borrow YOUT car?" is in a category other than
-May I borrow YOUT pen?" lbe request 10 borrow a ear 15 of COUrJC not quitt such a
greal demand ifthe person reque5ting the ear is Ihr ear owner's brot her. This illustrates
that the ractors distanu aod powu influence thc ultimate weight.
lhc ultimate weight of an PT" an be by a value according 10 thc
mula:
(32) Intensity of threat 10 face
W(FT") = R + D + P
Weighl of Face Threatening Act = Rate of imposllion + social DiSlance -+ Power
Brown aod Levinson did not indicate how values are 10 be assigned 10 R (rate of impo-
sition), D (sodal distanee), and P (power). But it should be dear that the value for P
is different in the following examples.
(33) a. Excuse sir, would it bc aJl right ifI dose window?
b. Mind me c10sing window?
Vtterance (33a) is more likely to be said by an employee to his OOss, while in the same
situation, (33b) might be said by boss to the employtt. In these examples param-
eters R and D the same values.
In their research on linguistically realized politeness, Brown and Levinson inves-
tigated a numher of languages. Their analyses indieate that !here are many ways of
committing an fTA with a given weight. All ofthese variants ean, according to Brown
and Levlnson, be redueed to five strategies:
/ ,. wHho" "'''ui" "'''"_ "''''r
/ on rtcord \ . / 2. POSItlW poIiten""
/ Do lhe FrA \ with rtdresslYe actIon \
\ 4. off rtCQrd 3. negallve pohteneu
5. Don'tdo the FTA
Figure 2. Pouible stntegies for doing FTAS
The fifth strategy is implemented when the risk of speaking is too great. when, for
Lnstance, an individual does not risk answering an impertinent and face.threatening
1.6 Pohtentss theory 17
quest ion and simply remains silent. "Off means that the FTA is not recogniz.able
:u such. An ofthis is Ihe (3k) variant ofthe request Ihat is in (33).
(33) c. I'm just so cold.
When the addressee replies "Then dose the speaker can still maintain
that arequest has not been made. "RedresslVI' action'" refers 10 an action that IS meant
to improve the stability betwttn conversational partners and is, therefore, an action
that minimius or prevcnts a loss of face.
Selow are examples of different strategies for asking a person fer a hundted dollars.
(34) a. Hey, lend Oll' a hundred dollars. (baldly)
b. Hey, friend. could you lend me a hundred budu?
polite)
c. I'm sorry I have 10 ask, but could you lend me a hundred dollars?
(ntgati\'e polile)
d. Oh no. I'm out of cash! I forgot to go 10 Ihe bank loday. (olf record)
The strategies are numbered according to their degreI' of politent'ss. (Slrategy 5 is,
from Ihis point on. lrJt out of consideralion.) If the W of an fTA is high, Ihe speaker
will choose a slrategy wi!h a higher number. This uptains why grave aeeusallons or
inconvenient requests are orten formulaled indirKt1y (strategy 4).
The Goffman approach and Brown and Levinson's theory provide an adequate
research framework for delermining gradallOns of potiteness and for anal)'lmg indi-
reet language. The foUowing queslion is an example of an indirec:t request:
(35) Are you doing anything special lonight?
The form of Ihis utterance makes it dear Ihat this is an inquiry about an Individual's
planned activilies. lbis queslion ean, however, also he an invitalion on the part of the
speaker 10 the address 10 SO out togelher.
How can a quest Ion in Ihis form have an enlirely different funelion? According 10
Levinson (1983), Ihe answer iSlhat in some eases speakers first make a pre-request in
order 10 find out whether they will gel a positive response to their request. Levinson
describes this in an underlying structure eonsisling of fout poSitions. Below are an
example and the underlying struclure.
(36) A: (I) Are you doing anything special tomght ?
B: (2) No. not really. Why?
A: (3) Weil, Iwanted to ask ifyou would like to go out 10 dinner with me.
8: (4) 1'cI love 10.
28 1. Communicahon as action
(37) The underlying structure of (36):
(I) Pre-request
(2) ahead" reaclion
(3) Request
(4) Consenl
Goffman's work on face afTers an explanation for the pre-requesl phenomenon. If B
had given an evasive answer 10 the pre-requesl, then Ihat would have eliminated the
necessity of making the main request, preventing the loss of face ofboth parlicipants.
A does not have 10 deal with a refusal and B does not have 10 refU$(: the request in
a direct manner; after the pre-request, B can claim 10 be extrcmely busy which will
soften the blow of Ihe refusal.
Indirect requesls have certain similarities with pre-requests in thai both are
allempts 10 ascertain whcthcr or not there are grounds for refusing a direct request.
Consider Ihe following example. A customer walks inlo a shoe store and I5ks:
(38) Do you seU jogging shoes?
This queslion is actually a preliminary check 10 sec if the sales clerk will be able 10 give
an affirmative response 10 a reque$t to see an assortment of jogging shoes. In Levin -
son's (1983) opinion, indi rect requests can be viewed as pre-requests in an underlying
slruclure consisling offour positions.
(39) A: (I) 00 you seil jogging shoes?
8, (2) Yes.
A: (3) Would )'ou show me seme, pIeaSt?
8, (4) l' Ugogetlhemforyou.
In many cases the reaction to a pre-request is Ihe same as 10 Ihe direct requesl.
(40) A: ( I) Doyouselljoggingshoes?
B: (4) Yes, 1'11 show you seme.
A: Thank you.
This reduction an be explained with the politeness strateg)'. Jt ensures that the eustomer
does not lose face; Ihe customer is no longer obliged to fonnulate a dlrect request.
QuesllOf1$ aod nsignmenlS 19
Questions and assignments
Questions
l.I.1 Use Ihe Organon modet 10 dininguish lhe functions in lhe foltowing ulterance:
Thll is qUlltlm i"ltmllng modtU
1.1.1 Formutale a possible iltoculion aod aponible penocution forlhe fOllowing ullennces:
I . 11" ratfling.
b. Here comes a dog.
1.1.1 Using the foltowiflg sentence, indicate ""hat is wrong with Ihe proposltional felieit)' eondition for
promist as staled in 5tion 1.1:
I pron!l-K you tht SO!ntoM IIttIl 'OlM lOII\OmIIII
1.1.) Using lhe illocutions flalteraod lie, $how Ih1t problems Inse If In itlocution is onlyconsidered
suceessful ""hen the Iddressee uflderslaods which illocution is meant.
1.1.4 Ctassify the foltowlng illocutions usi"g Habermas's basic t)'pes:
1.).1
mtHlt, pmIIlM, dtfy, offtr condoltllCd, mjllOl, lluc:nbc, l/ul!, SUII/ankt, onItf.
Whll kind o(knowledge is reqUlred to Mduce from lhe following Slalemefll th1t lnd not
a promlSC: is implied?
I proIIIl-K you thaI yow will Stt a whlppins if yOll do Iha! IIglun!
1.) .Z Give an example of an ullerancewhich, depending on the situation, ean have Ihe illocutiooary
force Ofdtf", reqoest, warn, Ind complain.
13_) lodicltewhich elements ofthe (oIlowing utterances are the reaSOll forthe asslg"ment cf In ilto-
eutiooary folU.
a. I Im _mins)'CU, there's a bull cornlng tOWlrdS)'CU.
b. a bull coming towards)'CU.
1-4.1 Using Ihe term cOIIUtfSllt1Oflll1 imptKlltun:, elpl1in wh)'A un deduce rrom 8's remarkwhat hme it iso
A: What time is it?
8: Weil , lhe mail's amved.
2 .... 2 Pnwide arguments that woold support IM stalemenllhat not alt ofGrice's ma"ims are eqtJally
important.
'0
1. Communication as action
1 .... ] "rgut fororagaiOlSl l/le following 'ine of rusoning from Lnch (1983:15,16).
..... ofGrice's two Muims ofQuality (which I call subm.:u:im5), dle sood 1ft1'T1S 10 be a
predictable extension orille first:
Muim "Do not sa)' what you belieYe 10 be (alse.
Muim 1: Do not u)' that forwhich you IKk evidenu.
Ir we ur somethillg forwhich we lack evidence, we da not knowwhetherwhat we u)' is trot
orfalse. Th1!'refol"t Muim 1 simply S<1ys '00 not pul yourwlfin a position you risk
brtaking MUlm .'; and both can be summarizN in the prKCpt 'A'iOid ttmng untroths'."
2 .... 4 Which muims orlht cooperative principle Ire belng in 1M following dialogutlllndicatr
whil;:h implicaturn this to.
a. A: Artwegoingtoeatsoon?l'mhungry.
B: In iI minute. ' jusl h.ave 10 frythe l;...er.
A: Suddenly, I've lost my appl'tltt.
b. A: Mn;. Johnson is an old .....itch.
B: It'5 wonderful weat,*,"for thts .ilm' of year, don't you thinkl
J. .... 5 plain howdeliberate violJ.tions, or as Grice ealls them, ofthe eooperative prindple
;u in (J.) and (b) ean nill beaf meaniog. Also uplain when such deliberate vioIalions do lead,
fore.ample, 10 IYlng, by di5Cussing some conditlOns Ihat must be met fofflouting to renderthe
appropriale or intendtd effect . eThis question is inspired by an e.ample In Cook '989.)
a. 11ow,1 whrn!/Oll wns 0IIt ofUy oll tN: t,nK.
b. My ctll phoM's runs d(od (IIfry fill( m,nUln.
2.5.1 Is B's reaetion a counter-ex.ample 10 the proposition thaI every uneranee can be relevant in a
eomoersation7
A: (waiter) Can I get)'Oll something to drink7
S: (customer) Nliturally, everybody drinks.
J..5.a Recall eumple (17) from Section 1.5. Nowtry toe.plain in )'Ourown words how lhe relc-vance
prineiple worb by applying ,110 lhe fol!owlngeumple: a nolice often found in lne london
underground (subway).
- Oqjs mUSI b1! rClrrite! (NI tsCClICllor.-
J..6.1 Use the terms fllc( and focr worlr Irchnlqun 10 explain Ihe misunderstanding In Ihe followlng dia
Iogue.
A: Are)'Oll golng 10 da anylhmg Wlth those oId chairsl
B: No,)'Ou can have Ihem.
A: Oh, no, Ihat's not whal I meant.
QuestionS aod usignments
2.6.1 In the following dialogue, is B being posiliwlyoroegatiwly poIite7 (B Ihinks Iht dress is ugly. )
1.: So, whal do)'OU think of my
B: Wtll, II's Ihat's for sure.
J..6.) RlIInk tht foIlowing staltmenlS fTOm -utl'fmtly 10 Mlns pelltt
M
using Brown and ll!Vin-
san's Ihtory. Indiealt which slraltgy has bftn
a. Da)'Ou IIgrrt 10 pay half oftht bill thirtydays before dtliwry7
b. Thirtydays will rectivt a bill for half oflne order.
c. You haV't to pay half oflM bltl before dtlivery.
d. Though we do not like to make Ihis dtmllnd, it is Ihis company's policy lhat half ofthe bill be
paid thlrty days
1.6... uplain why S does not answerwllh but immtdiattly makts an offff in llit foIlowlng dia-
IogUt.
A: Da)'Ou haV't ice cream?
B: Da you ehocolatt loppiogl
Assignments
15' In Seclion 1.5 it is argutd thai is uodtrspecifitd. Consider the example
MThaok)'Ou forobserving no smoking-. Try loexplain why language is often underspecifitd.
Think, foreumple, oftllt tconomy prineiple, which applies in mosl human behlVlOf.
1.5.1 Mikhail Sakh,in was a famous Runian philosopher and language tlleoriSI al Ihe beginning of
tne last century (Sft also Secuon ,p). In Tht Sokhlln lWI.ur by Morns ('99<\ :16- 37) IM foIlowiog
quole from Volosinov (a frieod's name ullderwhich Bakht;n wrole);s IOcluded:
-Ta undentand anotherptrson's ulteranct mtans 10 orienl ooeselfwith respect 10 It , 10 find
Ihe properpiace fofit in IM conteKt. Foreaeh ward oflht ulteranee Ihal_
in proctsS of underolanding, we, as il were, laydown a sei of ourown answeringwords.
T1It greatertheir number aod weighl , Iht deeper and substanlial ourundenlandlOg
will
Try 10 point out Ihe differences andlor similari l its Volosmov's reuoning and the view 00
rell!Vaoct ofSperber lind Wilson, as prtStnttd in Section 1.5 ofth,s chapter.
,.
.'
u
1. CommunicabOn as Ktion
Bibliographical information
For P]atO$ ",ew on l;mguagt' iU a form of actIon. the reader;5 .der<! 10 the Fowltr rd,hon (]\ln),
Part IV, pp. 19- 23. The View thaI b.ngulit is an "!lv,lr can also bc found in Olht. worb, rOT
tnmple. IM Gnman bngut and philOlOphf:r Wesrna- (1885) and tIM: EngiiSh Imgulst Gardu\er
(1969, 2nd edmon), who dedicated hIs work 10 Wrgentr. Gardme. also slrn.$N tht coopuauW'
aspect m
Buhlu dC"idoped hu viSIOn In publk"hOl1$ al IM btgInning of lhe last ct"!lIry. A more dabo-
nllt aplanatiOn ofhis ",ews is givm in hlS opus SpFlKhtMont (1934). A traru.lihon oflhlS work
aJ'IX'ared in 1990.
The work done by Jespoerscn has aI50 bttn influenllill. H,s major work. dalmg from 1924, has
bcm rqmnted lTWly IIiMS. the moIt reccnt being In 1992.
1.1 Austin first h,s ideu in I SC:Tla ofWiIl,am lames l.turcs 1\ Harvud Univtrsity in 1955.
The leciuml wur publuhtd posthumously by Urm$(ln and smu. an.:! tided How to Do lhmgJ Wirh
Words. S AlISlin (1976),
Tht mOSllllfluential phtlosophtr in thearea of $peech aet thtory was lohn Searle. An oft ' ql,loted
J'lIbhcatlOfl of hl$ 15 Spuch AC/s (I969).
Habamu presenlN lus ldeu m hIS major work 7Mont da Hllnde/1I$ (1981).
In thls work he also enticltts Searle. The cJusdicallon of Illocullons is not dealt WI!h furlher in Ihis
inlroduction, u this probkm IS noI in Ihe mainstream of conlemporary discol,IlK studie$.
1.J The aample of fclKity condmons for recjl,lOIS was taken from Searle (1969). Fm a good panorama
of whal is being donc in (Onlemporary speech ICl lhtory, stC' Vandervektn and Kubo (2001). lhis
coUcctlOll of iJ more DneolN towards illocuuonary logIc than 10 "daalCal- speech lC1
Ihcory. lt should bc menllOnN thaI Ihis " oot a book.
1.<4 Grice was mamly interwed m alurallangl,l . He wanlec! to provc thai a nalurallanguage was as
precllt u Ilogocal language, pTOVldtd In extra 5t1 of rults SOVttmng natural Ianguage iJ laken mto
KCounl. Sy slalmg Ihls, he look ilSl,le wllh IhOK philosophen who claimcd thlt nllllTlllanguage
was 100 imprecL5C' for Kic'ntLlic pl,lrpoioeli. Gnu, hke AUShn (5 2.2), presentcd hIS proJlO'als in I
Wilh:ml lames I.ectures senes al Harvard Uni>"efS.lly in 1%7, twdve ynn Il\er Austin. A summary
olthcse Icctuml was publishcd in 1975.
1.5 Thc standard work on relevance LI Sperber ilJ\d Wilson ( 1995). Canlon (2002) rrovida lihorough
and comprchmslve panorwna olIM rdev.ance debite up unlLl now. Example (26) is takm from
Smllh and Wilson (1979); (28) and (29) from Luch (1983). Exlmple (31) is laken from Grundy
(2000). Blakemore (1992) gtVH an inlrodLlCUOO 10 pragmaliQ thaI 1$ based 00 Rdevance lheory.
,.,
Bibhographic.al informnion
Goffmln bucd his concept of fou Oll hLS research Inlo ritual demenlS in ,ocul mIeneIlon, thaI
15, lhose slandardizcd &cu Wlth whoch IndIViduali uprc:ss rc:specl or defn-rnce for each othrr or
objects. Goffman lim prcsentec! his ideu In Tht Presentlwon o/$elf in Ewryday U/t (1956). A
good way of gctlmg acquamtcd wilh hu Kleu LI 10 lud Goffman (1967). OlM' olhis lut pl,lbl,ca-
t'OI\$ dales from J 98 J
Luch IpproaCha poIL1eness phenomcna in an entlfely different mlnner. In Lch (1983) the
politeness principle is proposed 10 be from the (ooperatlVe princlple, compkte wLth accom-
panying mulms such u Ihe mulms oflacl and mOOesty. Since, however, tM numbcr of mulms is
greatly upandcd and Viobtion oflllt; maxims clon noI kad 10 ImplicalUmi such as OCCI,II"$ ... 'th Ihe
cooperalive principle, Ihe thCQf)' hiS not gamed . Iarge following.
For mon on Brown and I..evJnion'l thcory, tIM: reader il rd"erred 10 Ihnr 1990 publicalLon that
reporu on rcseudt dOOf: 00 poIllmns phmomcna in I numba oflanguasn. Thc proposaI on Ihe
analysis ofindircct requeslS LS from l.evmson (1983).
"
3 Discourse in communication
J.l Thc pragmatic perspeClive
In the previous chapter communication has 1>0 a5 action, precisely
as the combint'd action of speaken and lisleners. or wrilers and readers. in wokh they
exchange illocutions. following the principle of cooperation aod the relevance muim
and obeying polilentsS strattgies. Ir anything has bccome dear, then it must thaI
the verbal part of commumcation, cannot be studkd without taking into
accounl the context in whkh the communicative aets take plaet. Only in this way can
lhe relation belween the form and the function of discourse be clarified.
A good framework for studytng in the form (unctlOn approach is
pragmatics. Pragmatics. hterally -the study of is itself part of a philosophieal
approach 10 the phenomenon sign, specilically the queslion o(how signs, and Iherc(ore
also linguislie signs, (unction. This is known as semioties. Two names associated wilh
semiOlics are those o( Ihe Ameriean philosophers CharIes Peiree and Charles MOfTis.
Peirce's ideas. firsl published allhe beginning ofthe previous century, were elaboraled
by Morris and gained promincnce in lhe 1960s. For a good insighl inlo lhe stat us of
the pragmalic perspective in discourse sludies some information about semiotics in
general is needed.
The cenlral conceptlO semiolics is the notion or sign. A sign, according 10 Peirce,
cannot be seen independenlly of its object and il$ interprelan\. A sign, in the form of
a rcpresenlamen, lel5 say the ",urd cast//!, stands for ils objeel. A sign ean only be a
sign if it is addressed 10 somebody, and ereates an in the mind of the addressee;
this is called the interpretan!. In other words, a sign 15 "nothing" without its funetion,
referring 10 an objt and ereating an idca. This Ihree-part approach caUs to mind a
similar approach in the Organon modd (see 5tion 2.1), but there is more. Peiree
dislinguished lhrec typet of signs. A sign an be an ieon, whieh means that the sign
rescmbles some object. e.g . a picture of a castlI' on a billboard, the picture o( a man
on a toHel door or an emoticon in e mail. A sign can be an mdex. which means that
it directs attention to the objcel. For example. a weathereock is an index of the di rcc-
lion or the wind. the phone ringing is an index of someone who wants to talk to you
and an arrow on a CfOssroads can be an index to a castle. The third sigo is the symboL
A symbol i5 associated with an obJ! by "rule". For I'xample. we have learned Ihat a
"
]. Discoul'K in communicahon
building with battlements, a drawbridge aod towers i$ called a castle. A symbol rep-
resents its ob,! aod determines its inlC'rpretant on the basis of conventions. Most
words art symbols.
It is important to note thaI manr signs art mixed signs. PeiTee himself gives the
examplc of a man walking wilh a child aod poinling into the sky. sayiog. Ihert
is a halloon.
R
Thc pointing arm is an index aod is as such essential for understanding
the symbol. lf the child then asb, "Wha! is a the father an answer by using
an ico: -0, that is something Hke a big soap bubble.
R
In comparison with the Organon
modd. the sign phi losophy in semiotics is mueh rieher, slimulating us 10 Ihink about
wha! Ihe symbols that we use in discoursc precisdy are. Figure land Tabi!' I summa-
riu Ihe aspects of a sign and its thru categories.
Elaborating on Pei rcc, the philo$Opher Morris (1938) three areas in
the field of semiotics: I. syntax, the relationship betwun signs within a sign system, 2.
semantics, the relationship belween signs and the objeclS they refer 10, 3. pragmatics,
the relalionship between signs and the people who use them. Pragmatics is
with such quest ions as why an individual uses a specific sign, which circumstances call
for the use of a specific sign and how we interpret signs. Pragmatics, in other words,
deals with questions about how signs function. Applitd to discourse, the pragmatie
approach deals with the question of how discourse is produeed and interpreted in
eontext, in specific situations. With the study of production and perception we find
ourselves in the area of eognitive studies (see Chapter 13). When the focus is on dis-
course in conten, in specific siluations we Ire in the research area of langulge and
llStltutions (see Chapter 14) and languIge and culture (see Chapter 15).
interpretanl
(the KnK made ofthc sign)

reprcsenlamcn
(the sign vchic1e)
objecl
(whltlhe sign slands for)
Figure I. Peirce'J Kmlotic triangle
Table I. Thr c.ategories of 'lgDS
Icon
I hkeno.s 10 5OfIl(' obJl
(natural or by convcrUlon)
Index
I Cilu.saJ cOltn.tet>on 10 thc
obI'"
Symbol
a 5t1pUblcd.
relation (most wams)
3.: Rliles forsymbolic Interaclion
1t is important to note that in discourse studies pragmatics is mostly not eomid
ercd a separate branch oflinguistks in the ....... y that phonetics, syntax and semanlics are.
Should this be the ease, however, pragmalles is often used as a kind of wastebasket
term for Ihe study oflinguislic phenomena that cannot be described at Mlower
in ImgUistics. On the contrary, pragmalics in discourse studies means a pragmalic
view of all levels of eommunication through the use of symbols, thus also syntaetic and
semantic aspecls. Compatt Ihe senttnces in the following IWO uamples.
(I) Marywenl to Ann because ,ht wantcd coff.
(2) Mary gol angry with Ann because ,ht was not al home.
Howcan il beexplained thai Mshe- in (I) can rtfer 10 Maryand Ann, but in (2) usually
only to Ann? To explain this difftrtnet in grammatical refertnce Wt nud a rule about
use lilre Mlf there art two possible antecedents then the pronoun is normally
used to refer to lhe subject, unless the stale of affairs represented gives an ovtrruling
cue for the othtr Now consider a semantic example. The word ''Just'' can
have the meaning "a short timt But there i5 a differtnee belwun the following
sentences.
(3) I Just eame from the toilet.
(4) 1 Just had my holidays in India.
80th sentenecs refer to the past, but the action described in (3) is usually neatet 10 Ihe
presentlhan lhe action in (4). From a pragmatic perspectivt it is inttresling 10 study
the manner in which rather fixed meanings can vary in language use.
In discourse Sludies Iht pragmalie approach has become very important, since
discourse is nol just an abstract combination of syntax and .scmanlics, possibly com
bined with phonctics in spoken discourse. In the pragmatic approach the focus is on
languige in use in specific situations. What are the regularilies, what are the rules that
pragmatics wants to det1 in language use?
3,2 Rules for symbolic inl eraction
Pragmatics is about the sodal rules for the interchange of symbols. These rules differ
from rules in physics or logic, for example. An example of I physical rule is the law of
gravilY; this rule exists indepcndenllyofhumans. An example of a logical tule is that if
A is bigger than B, and B is bigger Ihan C. we can draw the eondusion lhat A is bigger
Ihan C. By conlraS!, a social rult is nOllogical and exisls only be<:IUse propIe exhibit
SQme regularity in behavior, in Ihis cast verbal bchavior.
31
"
3- DiSCOUrR ,n communic,ltion
Sodal ruks have six charactcristics. which are important for studying lan-
guage in USt, These" characlerislics can be illustrated by the following fragment of
conversalion.
(5) A: Howareyoudoing?
B: Fine. And you?
Thc rules by which A and B aCI are ( I) acquired and (2) usually not applitd conse/aus/f.
A language user has to learn how to react 10 a question about his health. The fact that
rules are acqULred implies thai they are leamahle and that il is possible to aet accord-
ing 10 these rutes. Nevertheless. mosl of the time a [anguage user will nol bc aware of
rules.
Furthermore, social rules are (3) commrma/. They are nol private mies, hut are
acted upon by groups of people. This mean$ that spttific eXptations exist about
verbal behavior. On the basis of a communal rule for reacting to an enquiry about
someones health, A is allowed to expect B to react in a certain manner.
The fourth characteristie is based on the third. Rules are (4) /.l fr/.lmework for
ImderSl/.lnding and Judging an iIIoculion in terms of(in)consequence. impl icalions and
appropriatencss. In certain circumstances B can interpret As queslion as an attempt to
start a conversalion. In this case A may interprei the lenor of Bs as
a notification that B wants 10 have a conversation as weil. If B were 10 start a conversa-
lion with C immediately after his utterance to A, A would judge this as inconsequent
or inappropriate.
The last two characteristics are important a5 weil. Rules (5) can be viola/ed and
rules are (6) /iable to change. The faci that rules can bt violated is proved by B's pos-
sible reaclion of starting a conversation with C. The rule "answer Bs counler questlon
only by saying 'fine'" can change in the sense that the person addressed is expected 10
provide something more Ihan just a formal answer, for example: uWell, actually quite
fine these days" or "Weil, not SO good, to be
In relation 10 Ihe above-mentioned characterislies of sodal rules, an often used
pair of concepls must be broached, namely descriptive and pre5Crrplive. These concepts
are of help in explaining the nalure of rules, and in c1arifying Ihe di"erence belween
a rule and a norm.
By using rules it is possible 10 describe how language is used, but the same roles
can also indicale how language musl be used, for example. thaI a greeling has 10 be
answered wllh a counter-greeting. long as one purely describes, 11 is a maller of
descriptlVe rules. If one prescribes or dictales, it is a maller of prescriplive rules. From
a discourse-analytical point of view, rules are descript ive. However, rules are prescrip-
live for a language user thai wanlS 10 know how 10 use language. One rule can thus be
viewed from both a descriptive and aprescripllve perspeclive.
H HessagleS bel\Oftfl sender and receiver 39
What sets rules aparl from norms? Norms always concern the question ofhow one
should bthave. They are guidelines wilh an elhie aspect. Norms are Ihe values Ihal one
uses when answering queslions about what is (morally) r ight or wrong: for example,
to eal meal or nOl, or 10 use foreign words in your own language or not. A language
user can mlerpret norms for language use as prescriptive. The concepl of norm has 10
be clearly discerned from the concept of rule, since the laller docs nol always imply
a guideline.
The prtvious paragraph discussed the dlstinction between rules and norms. Bul
whal is Ihe difference between rules and maxims (discussed in Seclion 2.4)? Rules can
be represenled in Ihe form MlfX then An example is the rule Ihat if one is greeled,
one responds wilh a counter-greeting. The if-senlence indicales Ihe situalion in which
the rule applies. Such a situalion serves as the demarcalion of a rule's reach. Othcr
restraints can be an enumerallon of terms or condilions under which a role is valid.
For example, the illocution 10 promise implies Ihe rule thai the speaker commits him-
selfto do whal he hal promised.
There art also general rules wilhout a condition in an if-senlence, for example,
this rule in language use: "Avoid Such a general rule is called a maxim or a
ground rule. Another example would be "Avoid unnecessary When scveral
of these maxims can be allached 10 one underlying concepl, one sj>taks of a principle,
in this case for example, the efficiency princlple. On the basis of muims of ambiguily
and protixilY, which an be allached 10 this principle,language use in cerlain situations
can be described by means of roles.
The difference belween rules on Ihe one hand and maxims on Ihe olher can be
summarited as folIows. Muims are assumed 10 always be valid, whereas rules only
apply in specific situations or are valid for specific iIIocutions.
Now Ihal the aspects of signs, thedi"erenl types of signs and mies for the exchange
of symbols have been deall wilh, .....e take a doser look in Seclion 3 al how signs are
combined mlo messages between sender and receiver.
3.3 Messages between sender and receiver
In layman's lerms, a discourse is I. a sei of connecled meaningful sentences or utter-
ances(lhe form) bywhich a sender2. communicatesa message 10 a receiver (Ihe func-
lion). This popular approach to discourse as a coherenl series of symbols 15, however,
not useful for scholarly purposes. Two problems arise. First, il is not dear whelher or
nOI a lexl qualifies as a seI of coherenl meaningful sentences. Second, Ihe terms sender
and reuiwr mayobstrucl one's view of whal senders and receivers are actually doing as

3. OiKOOrw in communicalion
participants in communicating tht meaning of series of symbols. With regard 10 these
two problems, consider the following example, an experimental poem.
(6) DIa
DIa ota ota
800
Ola OIa
DIa OIa OIa boo
0000
00 00 ota OIa ola
Does Ihis (anglicized) poem by DUlch poet Jan Hanlo given in (6) qualify as a mean-
IIlgful Jet of coheren! utlerances? With 50me imagination, il i5 possible for areader
10 argue that the poem 15 50mehow coherent, for example, as a composition of aural
effects or as a satire on traditional pOdic conventions. The attribution of cohrrence
is, therefore. partial1y on the disposition of the receiver. So. perhaps il is
better 10 define discourse as a seI of utlerances thaI can be conceivcd of as meaningfu1
or rdevant.
The use of the terms sender and rclver presupposes that there is an information
packagC' Ihat has 10 bC" transmitted from onC' person to anothC'r. In fact. Ihis viC'w of
thC' participants in discourse sterns from a viC'w of communicatlon that is inspirC'd and
influenced by the invention of thC' telephone at the beginning of the lasl cenlury. It is
known as Ihe general communication model. which originatC's (rom ClaudC' Shannon
and Warren Weaver (1949). FOT a better understanding of this concC'pl of discoursc. a
short C'xplanation is required.
Ihis diagram is to be read as fotlows. In thC' information SQurce a message is
selC'Cted. Ihis can consis! of wrinen or spoken words. imagC's. music, dc. IhC' transmit -
1I\forrnal'Otl

InlnSmllltr

codmg
l ignal

-
Lr'
rt'vt<!
mtssagt -
51gnal
--
.. . doding

Figurt 2. Shannon and Wnvtr's gtneral communicalion model
].] HeSY8n !ltnder and rec::eilll:r
Ur encodes the message in a signal that is sent through the communication channel
to Ihe receiver. The re<:eiver translaIes Ihe signal into a message - decodes it - and
transmits this mC'ssage to the destination. In communication, the message can undergo
changes caused by noist'; this is called
The general communication model has oftC'n btC'n appliC'd 10 discourse. ThC'
speaker/writer thC'n isthe information source. (What a speaker/writer has in mind can
of course bt baSC'd on other 5Ourcesofinformation.) The message is coded into worUs.
This is then transported through a communication channd, voice, paper or screC'n. 10
Ihe receiver. The recelvC'r dC'Codes thC' message, aftC'r which the messagt' arrivt'$ at the
final destination, thC' reader's mind.
This communication modC'1 is appealingly uncomplicated, bul the process of
communication 15, in facl, much more compla. There are two major objC'Ctions that
can be put forward. First, nothing can be said about iIIocutionary force in Ihe sender-
message-rC'Ceiver approach. lhe message "1'11 come tomorrow' can be a promise. a
statement. or a thrC'at. If the communication is to be successful, the recC'iver must not
only understand that the in thC' scntC'nce will bt prescnt on the day foliowlg the
message, but also understand what kind of sptcch aCl is being committed. Ihe same
holds tTUC' for indirect language usc. An utterancC' likc -Are you doing anythingspecial
can. in ccrtam situations, bc seen as an invilalion. It is lC'ft to thC' rC'CC'ivC'r 10
deducC' this from the mC'ssage.
Second.the modd does not take into accountthC' discoursc situation in which the
communication originaled. Tbe situation does, however, playa role in the interpreta-
tion of a message. Consider the following sentencC':
(7) Pete told lohn that he was sick.
Tbe scntence can mean that Pele is siek. but il can also mean thai lohn is sick. In
practiee Ihis amblgulty rarely leads to difficultiC'S bccauSC' thC' situation usually makC's
it fairly obvious which meaning is intended. In thC' SC'nder-message-recei\'er approach
no dC'SC.Tiplion can be given ofhow thC' process of dC'Coding determines the applicable
interpretation. becau$C' thC' situation is nottaken into account.
In short. discoursc is more than a messagt' from sender to receiver. In fact, SC'nder
and receiver are metaphors that obfuscate what is reaUy going on in communicalion.
Specific illocutions haye to be linked to the message depending on the situation in
which discou(sc takes place. The illocutions have been dealt with in Section 2.4. Ilow
the sendC'r-receiver metaphor obfuscates what is reaUy going on in communication is
very insightfuUy demonstraled in a sludy by thC' psycholinguist Herbert Clark ( 19%)
Using Pursuing old ideas of, for example. JesperSC'n (see SC'clion 2. 1). that
language is a human activity on the part of al least two individuals, Clark compares
languagC' in USC' with a business transaction. paddhng together In a canoe, playing
.'
"
1- in communication
cards or performing music in an orcheslra. Cornmunication is a form of social action,
of joint action.
A (entral notion in Clark's sludy is common ground. The joint ICllvllY i5 under-
taken to accumulate Ihe common ground ofthe participants. With common ground is
meant Ihe sumoft hejoint aod mutual kno .....ledge. beliefs aod suppositions ofthc par-
licipanls. This view is pcrhaps raSier 10 understand by considering Ihe Latin roots of
lhe verb to (ommunkate: Ihr meaning "to make In Clark's view this means
thai communication as a joini activity adds mformation 10 thr common ground ofthe
participants (and not "senders" and -reivers"). In dealing with all kinds of
conversations Clark showed thai dl$COursc cannot be studied without twog inlo
account this common ground. According to him common ground divides into three
types, as shown in Table 2.
Applied 10 the joint action of a cllent and a clerk in a drugstore, these three types of
common ground can be filled in as folIows. The initial common ground pertains 10 the
fact that both the dient and the clerk presupposc the standard procedures in their cul-
ture for exchanging money for goods. They also presupposc that they both speak the
same language and thaI the elient has enough money, etc. The current state of affaif5
1$ different from the initial state of affalrs, and visible 10 the dient and the clerk. After
the client has. for example, placed two items on the counter and said like these two
il is clear 10 both hirn and the clerk that one is the the other the current
customer, that the dient refers to the hems on the counter, that the clerk has to ring
Ihem up on the cash register, C1C. When the clerk then, for uample, says "Here's your
both he and the client presuppose a sequence of public events so far: the elient
has caught the clerk's attention, he has specified the items to be bought, the clerk has
rung them up, the dient has given money and the clerk has returned the change. With
his focu.s on comman ground Clark stresses the importance of studying communica-
tion on the basis of what participants verbally exchange in a real situation.
Table 2. RepresentaliOIl5 cl ,ammon ground
of common pound
Imtial common ground
Explanation
The $el ofbackground W:11i.assumptlons.nd beliris
the parILcipanl$ presuppoud when ther entered the
joinl acll\lily.
2. Curnnl Sille of Iht: joinl acu\llIy Whallht: participanis prcsuppok' 10 be Ihe state o(lhc
3. Public C'o'enll; so hr
;&<:1'\I1IY I1 Iht: momenl.
The evcnlS thc partiClpanlS prc$uJ'PO't" hlIve OCCUTfcd
Ln publK up 10 Iht; CUlTCnlstalC.
3.4 The discount: Sltualion 41
).4 lbe discourse situation
In the discussion on illocutions, il was noled that these could be seen as functions of
certain forms. The form -announcement" can funClion as order, request, C1c. Obvi-
ously, this does not adequately desc.ribe the term jutlctiotl. Function here means the
objective and dfect in a given situation. For instan:
(8) A: Oe you smoke?
B: Weil, if you'VC' gOI a cigarette.
The funetion of A's utteranee coold be that A wants to malte B fcel at ease by usmg the
question form for the iUocution /0 olfer. IU objecllve has a specific effect: B makes lt
dear Ihat the iIIocution is understood, and counters wlth. as a perlocution, a sugges-
tion which makes it dear that A's objective has bcen achieved. lbe interpretation of
possible objeclives and effects, however, can be strongly influenced by Ihc situation
in which the utterance takes place. If, for example. the quest ion "Do you smoke?- is
asked by a physidan, it does not function as a means of starting a conversalion, bUI as
a medical quest ion.
The situation in which discourse is produced and processed an be analyzed and
defined using a large number offaclOl"S thaI can have an influence on possible objectlVes
IOd dfects of dlSCOUtse. Such a description is avaLlabie for tlv: speaking situation. It was
developed by the anthropologist DeU Hymes (1972) who, on the basis of ethnographie
research, summed up the componentsofthe Hymesdistinguished sixleen
components. which he grouped using the word SPE"KING as an acronym. In the following
outline the components are ilalicized.
..
). in communicatioll
(9) Hymes's SPEAIONG model
S
p
E
A
K
N
G
Set/mg pIa. an.:! other physicJJ cond!tions JUrmundin& tht
speech acl.

Tht psychoIogJaJ COlIntcrput 10 sctI.l!- What ;1 ma.nt httc
11 that I Ktting an chanp, for fTOm formal 10
informal. by IM parlicipants.
Partk'panl$ Tht or Smdcr, tht NIdnJsn, tht HttlrYr. Rt:UiVtt or
Audimce, and IM Addrm.
En<h The Purpm - an<! - goah.
Ac! Tht Form and Ihc Cont'"l ofthc mcssagr.
"'"
Thc 10m 01 the COflV'mIIIOII. ror aample, smous GI"
mock.mg_
1hc ClII:mntU; wnttm, rtc..., and the Fomu af
Spudt; d,akct, standard \.anguagt, de.
Norm.

1hc Norms of Inltracll()fl, e.g., Interruption and Nomu of
IntnprffallOl<. for aamplt, how a hSlmnJ suddtnIy IooIunl!
aw;lY muS{ M interpmed.
faLry taIe. advrnllCffiellt, dc..
This model became popuIar larse1y becal.lR of the handy grouping using the letters
SPEAKING. It 15 undear, hOwn'er, what the in"urnte of the components is.
Moreover, the outHne is nOl complete. Background knowlcdge share<! by the speaker aod
the listener, and possible in background knowlcdge, can inlluencc discoursc.
Tbc same holds lrue for or attitude.
Though the model is fairly general. by using it, the factors eomprising the discourse
situation ean be cJarifitd. Morwver,the differences between two comparable situations
can be highlighted. For example, take on the one hand the discoune situation in a
cJassroom where a student and a teacher have a conversation and on the other hand
the situation In which the student talks with his roommate in the dormitory. 8y
using a few letters (rom the model, possible differenees can be explained.
The discourse situations ean be defintd using elements from the components 5,
P and N: from setti"g the plaee of occurrence, and the scene, the psychologieal oecur
rence, (rom the pnrtlt:iptmts their SOCletal role, and from Ilorms those thai are bound
by plaee. It 15 obvious that this sludent finds hirnself in two different settings: in a
dassroom and in a student dorm. The scene is different as .... under the influenee of
the different roles that the participants lake: the SludenHeaeher conversation wiU be
more formal than the student student eonversation in the dorm. lhis student has to
].S
].5 The socio-semiotlC approach
ael aeeording 10 varying norms as weil, when talking with a teacher or with a fellow
student.
The remaining fanon eomprismg the SPEAKING model darify the discourse situ
ation as weil. lhese (actors deal with the relationship between (un"ion and form as
mentiontd earlier. For example, the student-teacher convenation w,U be of a different
genre than theeonversation hetween the student and his roommate: al school he asks
a quest ion, at horne he teils a joke. And pcrhaps the student uses dial1 when he is
at horne. whereas he will probably use standard language m a eonversation with his
teacher. The student's eonversations will serve different ends 100. In dass the purpose
is to have a quest ion answered; in the dorm the goal is to amuse his roommate.
A veryimportant point 10 remember is that discourse is not only apart ofthe situ-
ation, but can change the situalion or even ereate a context as weil. The way in whieh
discourse can change the situation can also be observtd in the dassroom example.
Suppose that one ofthe students knows the teacher as a Ileighbor and wants to let him
know thatlast night's party, which the leacher organized, v.-as too nolS)'. This changes
the situation in a dassroom from a studenlto-teacher conversatlon into a neighbor-
toneighbor conversation.
An example o( discourse ereating a context is the opcmng sentence of a conversa
tion. This aetivates amental mapping for both speaker and listener. Iftwo strangers
are silting on a park beneh, a eontexl is creattd when one asks the other "Could you
perhaps tell me the time?" or "I think It'l1 be Both persons have an idca
of the situation that enrolls and about the way it mayevolve.
A last.termlnologieal, question concerns the following. lhe term coll/a/ is often
use<! instead of situation or discourse situation. This term can, howevcr, cause some
confusion as the word eOIl/allS also used to denote the piece of discoursc surrounding
an element in discourse, for example, the context of a word, a senten ce, or a paragraph.
For this reason the word contul is often qualificd: the "verbal context" or
environment" as opposed 10 the "soeial contut" or "pragmatie The soeial
context is often divided into the context of situation and the contcxt of eu\ture. The
word cotut ean also be used to denote \'erbal context, thereby dlstmguishing it from
COlltext in the sense of d,scourse situation.
The soeio-semiotic approach
So far in thisehapter we have seen that diseourse cannot be studied adequately without
taking into aecount the addresser and addressee who use discourse for all the purposes
eommunication can be used for (Section 3.1). We have also establishtd that the main
aun o{ studying discourse can be formulatcd as to detect the rules that underlie this
"
,6 3- OiscQUBe in commUl1icatio01
symbolic (Scction 3.2), Within Ihis approach Iwo rdated aspe<ts are
vtry important. First, an addressee i5 not jusl a receiytr of Ihe message; in fact receiv-
ers are aclive. cooperating participants in Ihe communication ($celion 3.3). Second,
discoursc i5 a1ways situated in a social contcxt and in a spific situation ($cction 3.4).
The most important question in di$CQursc studies as an (inter}discipline has not yet
been Which theory or approach offen a good general framework for analyz-
ing allthe different aspeCIS of discourse? The bnl candidat, secms 10 be Ihe so-callt'<!
socio-semiotic approach.
The founding father and mother of Ihe socio-semiotic approach are Ihe Iinguist5
Michael Halliday (1978) and Ruqaiya Hasan. Thcir approach is generally tabete<! as
!Imctio"al grammar, becausc il focu$CS on tht funetion of discourst in eonltxt. 11 is
also ealled syswmc funclionallillguistics. becausc it studies Iht funcllons in a system-
alie wholt. Howtver, Iht ttrm socio-umiotic Sms more prteisc as il uplains tht
rools in semiolics (S Slion 3.1) and tht sociologieal focus. With tht prefu. soeio
tht social context 15 meant, in whieh the context of eulture and the conlext of silualion
can distinguishoo. With semiolic the aet of eonveying meaning with symbols. in
our eue In discourst, is mtant. lht etntral claim in this approach is: Evcry (pitet of)
diseourse has to be studied in ils social conttXI, in tht culture and situation in whieh
il
Hallidayand Hasan (1985) describe thru aspects of sodal contat: fidd, tenor and
mode. lhese concepls make il possible to interpret the social contat of a diseourst ,
the environment in whieh meanings are exchanged.
I. Field
lhe fidd of discourst refers to what is happening and to the nature ofthe sodaJ action
that is taking place. More specifieally, it answers quest ions about what the participants
are engaged in, in which Iht languagt figures as some essential component. In genual
It is the gist of what the discourst is about. Tht fidd rtftrs 10 all diITtrent kinds of social
actions, from doing tht dishes to a parliamenlary debate. Fidd is moslly rtstricted to
institutional settings Iike Itures, visits to a doctor, etc.
2. Tenor
The tenor of discourse refers to who is twng part, to the nature of the partieipants, 10
their statuses and roles. It 53YS somtthing about the kinds of rolt relalionshlp that exist
between Ihe participanls. It explicates both the types of speteh role that participants
assurnt in dialogue and the whole clusler of socially signi!leanl rdationships in which
they are involved_
Tenor can analyred using categories such as power and social distance. which
have already bun addressed in Stion 2.3 on politeness strategie$. Tenor also refers to
].5 The socio'Kmlotlc approach 47
aITtet. the of emotional charge in the relationship between the participants, the
attitudes and emotions thaI playa role in communication. Compare a more objectlve
discussion between judges and a discussion at a protest meeting.
3. Mode
The mode of discoursc refers to what part the language plays and the participants'
txpectations about what language ean do for them in that situation. It is about the
symbolic organization of the text, its status, and its funcllon In the conlext. including
the ehannel (spoken, written or a eombination of the two) and the rhetorical mode.
Finally. it says something about what the text achieves in terms persuasive. expository,
and didaetic categories, and the like. In discourst studles the mode aspecllS central.
The most important quest ions are: "Whal is the organiulion ofthe "What
is its function in context?R and "What is achievoo bydiscourst?"
The three aspects of the sodal context outlined above eorrelale with thr aspecls
of the discourst: ideational, interpersonal and textual meaning. Meaning conneets
discourse to the contul. Simply stated. this is tht reference using symbols to elements
in the context. Hence the connection betwn discourst and context be situated on
the semantic level of semiotics.
Table 3 gives an overview oflhe correlation between the three aspects ofthe rocial
eontext and the three aspects of the discourst.
Table l. Key conceplS in functional grammar
Situaon
compontnU of ,ontut
Fidd
Tenor
Mod.
DiJCoUrk
IIp<<t.of muning
ldeational
Ilerpersonal mtanmg
Tatull mumng
From the discourst" point of view of discourst. Ihe ideatlonal meaning of a discourse
eorresponds to the !leid. 11 is the content of a dl$COUrst" as it refers 10 what i5 going on
in a particular situation or a specifie topie. The ideational meaning must be given as
an answer to the question: What is this discourse about?
The meamng of a dlSCoul'$e corresponds with the tenor of context.
and can be detected by analyzing how partidpants in the discourse are related 10 the
contenl or ideational meaning and how they use language to aet. For example.lhe use
of a command can reveal thai the has the power to gwe a command. Or an
adjecti\'e such as horrible reveals an affeetual aspect of a person who uses Ihis word.
"
3. Oiscourse in rommuni<:.lI!ion
Finally, tbe lextual meaning corresponds 10 thc mode of the contexI. The textual
meaning is the organization of the conlen! elements in a larger structurt, t.g., the
speclive in which a lopie is dealt with, and the techniquesofputtingsome information
in a prominent plaee or Ihe combining of sentences.
A last point 10 be menlioned i5 the war discourst is embedded in thc differtnt
levels of sodal (ontexi and thc correspondence with discourse characteristics. Hal -
lidayand Hasan distinguished two levels in sodal (Ontexl: thc contexi of culture and
thc contexi of situation. Up lilI now neilher thc different types of discourse nor thc
different slytes or registers in which discourst occurs have been discUSSfii Bul with
Ihis division of conlnt ioto two 1eo.'els, the phenomena of different tYPfi and styles
have a well-defined place within this sodo-semiotic approach. The notion of cultural
context is linked to discourse type and the not ion of situational context to style. In this
view cultural aspects have given discourse Iypes such as a news story or an instruction
their form. And the way in which language can differ in style - informal, bureaueratic,
persuasive, etc. - is dictated by the siluation in which a discourse takes place. In the
next part ofthis book, the first chapter (Chapter 4) is about types, the cultural context,
whiJe Ihe last chapter (Chapter 8) is about style and regisler, the situalional context.
The intermediate chapters deal with the mode aspects of discourse.
3.6 What makts discourse discourse?
Discourse has many different manifestatlons in many different situations, from chat to
decd of purchase, from sermon to shopping list. These manifestations are so different
thaI one may wondC'r ifthe lC'rm discount: is too vaguC' 10 span aU differences.
What is actually meant by discourse?
What all differC'nl kinds of discourse have in common can be indicated with ref
C'rence to the etymology of discourse. The word sterns from the Medieval Latin word
which means drculate. Literally, il means "10 run to and fro" or "10
run like a pC'rson who gives a speech and runs on aboul a topic. A discourse is
something that runs from one person to another. By the way, the shopping lisl ean
only do thai by "running back" 10 the pC'rson who made thislis!. lbe erymology ofthe
rdated word tut ean also be useful forclarifying the nature of discourse. The 'word tut
has the same rool5 aSltxtilt: the Lalin verb which means weave": Wovcn
in a text, in a discourse, are the different meaning units 10 a bigger whole. The textIle
metaphor wc know from such C'xpressions as lose the thread of communication"
or "10 tie up loose ends" in finishing a plea or argumC'nt , and "to spin a yarn" (dishmg
up an incredible story) exemphfies this.
] .6 What makes discourse discoursel <49
Of course, this etymological approach is not enough for the studyofthe
of discourse. What makes a sequC'ncC' of sentences or uttC'rancC'S a discourse? lbC' fol -
lowing fragment is, in any case, not anormal discourse.
(10) lohn wanl5 to visit his girlfriend. Mr. Smith Jives in a small village ntarby. The
vacuum deanC'r didn'l work. The barber down the stlftt couldn't help. The last
paJ'C'r had been sold. It is gOlng to M a Iong dull talk.
This fragment sterns to have come inlo existence by a number of unrelatcd sentC'nces
Ming plaeed in random order. But if some words are changerl, a piecC' of discourR is
the resulI.
(11) lohn wanl5 to visil his girlfriend. Mary Jives in a ,mall viIlage nearby. lhe car
wouldn't start. The garage down Ihe SIreei couldn't help. The last bus had already
!eft. 1I 1$ going to a long hol W<1lk.
The example iUustrates that the existenee of connections betwecn sentences is an
important characteristic of discourse. The term comrection i5, however, somewhat
vague. Robert de Beaugrande (1981 l, one of the grand old men 111 discourse studiC's,
has formulated enteria for tC'xtuality, that is, eriteria thai a sequence of sentences
must meet in ordC'T to qualify as a discourse.
a. Cohr.sion is the connection that results when the interpretation of a textual element
is dependent on anothC'r demenl in the lext. Consider the following example.
(12) lbe store no longer sold porcelain figurines. 11 used to, the man behind the
counler said, bul they didn't seil very weil. Sinee the business had switched to
plastic. sales were doing a lot beller.
Tbe interpretation is dependent on that of "store" just as "they- is dependent
on that of"porcelain The meaning of"used to" is dependent on "sold por-
cdain figurines": Tbc word "plastic" can only be completely inlerpreted in relation to
"(porcelain) figurines": Cohesion refers 10 Ihe conneclion Ihal exists betwttn elements
in the tC'xt .
b. Coheretrce is the connection that is brought about by something outside the tC'xt.
This "somelhing" is usually knowledge which a listenC'r or readC'r is assumerl to have.
The following example is not problematic in terms of cohC'sion even though the scn-
tenees hardly seem 10 be conne<:ted.
(13) lhe procedure is aClually quite simple. First you arrange things into different
groups. Of COUTSC', one pile may M sufficient depending on how much Ihere is
5
I
i. Discourw in communic.lion
10 da. Ir you have 10 go SOffiewhere else dur 10 lack of fadlities, thai is Ihe nal
s!ep. otherwise you pretty weil set.
[t is importan! not 10 overdo things. Thai is, it is better 10 da [00 few things
at on,e than 100 olaoy. In Ihe ShOT! run Ihis may nOI stern important, but
complicalions can (illiHr arise. A mistakc can be expensive as weil.
At fiTtiChe whole procedure will stern complicated. Soon, however, 1I will
bome JUSI another (actl arlife. 1I i$ difficult 10 {ores any end 10 the necessilY
for Ihis task in the immediate futuu, but then 001' never can 11'11.
After the proccdure is completcd. one arranges the materials into different
groups again. Then (hey am be pul inlo thrir appropriatc Evenlually
they will be used onee more and the whole eyde wiIIthcn have 10 be rtptalM.
Howcver, thai is pan aflire.
This seemingly disJoinled paS53.gr becomrs cohrrenl whrn crrtain Imowlrdgr of Ihr
world, Lr" knowledgr of washing clolhrs, is applied 10 Ihr lai. Thr Irxlthrn bomrs
rasy 10 mtrrpret.
c. IlIltnltOnalily mcans that writrrs and spcakers must have thr conscious intention
of achieving specific goals wlth their mrssagt, for instance, convrying mformation or
argumg an optnion. According to this cnterion, Ihe sequence of words in the experi-
mental poem DIa Boo in Section 3.3 can only be called a discourse after an authorial
intention has been assigned to it. When no intention is assigned, the word sequence
becomes the equivalent of a page of random words not unlike Ihe penmanship praclice
of elementary school pupil s.
d. Acceptabilily requires thaI a .sequence of senlences be acceptable to the intended
audience in order to qualify as a text. Consider Ihe claim book is mine. Don't
10u see my name is in This example has a somewhat skewed internallogic and is
Iherefore unacceptable to many people.
e. InformntiYenes5 is necrssary in discourse. A discourse most contain new informa-
tion. If areader knows everything conlained in a discourse, Ihen It does not qualify.
Likewise, if areader does not understand what is in a discoursc, it also does not qualify
as a discourse.
f. is essential to tatuality. So, it is importantto consider Ihe situalion m
which the discourse has bn produced and dealt with (see Section 3.4).
g. Inttrtulualily means that a sequence of sentences is rdated by form or mramng 10
olher .sequences of sentences. Ihis chapter is a discourse because it is relak!d 10 the
other chapters of this book. And this book is a discourse because it is a member oflhe
group oftexlbooks. An examplr of mtertextuality where the two sequences are related
3.6 What make-s discoursel
by meaning is a nrws bullelin on a topic that has previously heen dealt with in a news
program.
Criteria c, d, and t are somewhat subjective. Recognition of intentionalilY, acceptability
and informaliveness arr observer.deprndent. lt is eonceivable that within Ihe bounda-
ries of the situationality criterion, Ihe following example can be seen as an acceptablr
and informative fragment or discourst:
(14) Shawpeare wrote more Ihan 20 plays. Will rau havr dinnrr with me lonight?
This non-discourse might at first Sttm 10 be Ihr lastline of a newspaper atlide fol-
lowed by Ihr first line of an entlrely unrelated artide. Yet il is possiblr to think of a
situation in which these two srntellCeS could form part of a discourse, for rumple.
the situation in whjch lhe speaker has wagf'red a dmnrr a5 to Ihr number of plays Ihat
Shakespeare wrolf'.
Not all crileria areconsidered equally Important m discourse studies. Intertextuality
is mainly dealt wilh in Ihe field of discourse typology. Situattonnlily and Ihr subjeclivt
characteristics iPltePltionalily and II1formattvenm an of secondary importancf'. They do
playa rolr in research into textual fuoclions where function is defined as Ihr goal (i"len-
liollalily) and the e!feet (primarily the transfer of information) in a specific situation.
The crilrrion acceptabilityonly occun in nonnalive approaches 10 discourse studies, for
example, in Ihe investigation into the quest ion: Whal is a good discourse? In disco urst
studies most attention has been paid 10 the first criteria of rohesiml and coherenet, some-
limes laken logelher as conPletlivily. Cohesion is usually defined as the connectivity that
is lilerally deleclabJe in discourse, e.g., by synonyms and pronominal words such as slle, il,
ete. Coherence is Iheconne<:tivity thaI can be inferred from the discourse by Ihe readeror
lislener, e.g., we can place the word Ihmforeor Ihereafterbetwten Ihe followingsentences
in order 10 explicale lhr relation we havr inferred: '"She had a child. She married':
5'
"fR..l.Cj,;
r;Pp@'
'''' c:. 't
'-;t. . ...
-,i
... c ......... J
.'
]. in communicalion
Questions .. nd assignments
Questions
).1.1 In the preface 10 his Pnnc:ipln ofProgllHltlC.! (1gB]:.:) leh Ippro.lIched pngm.1ltics in the
following manntt
-8ut my appfOIth 10 pragmatks is by WlI)' ohlle thesis that communkatioll is problem-sohr
ing. A speaker, qua communicatOf, has 10 solW! Iht' problem, 'Given that I Wint 10 bring about
such-aod-such a in lhot consciousnns, wha! is the best war 10 accomplish this
aim by
Indicate the similantin and difTC'n!nces between Ihe pragmatic apPfOIch 10 discourse in Ste-
tion ].1 al'ld LeK"'s approach.
].:l.1 Is I. gramm.ticaJ rufe dncriptive Cf Support)'Our .n,wer.
].].1 Commen! on the proposition ",. tut is a (oliKtion ofconnttd ulterances- on lhe basis orille
following 00 they, in )'OUropinion, count u t6ts1 Enm!* {al is a graffiti tut aod turn-
pie (bl is a oottce on a shop door.
I. In the springlimt this building blossoms.
b.
33J ThHtymology oftht words ttxt and ttxtik gots baek to tIM: samt utin vtrb, "IWIt", which
muns "to wtavt- or "to join t.thtr". Defint Itxl usingworos Ihat also dtscribt eharaeltriSlics
of ttxtiles.
3.3.3 Ost txamples 10 illustralt thaI thecauses ofnon-compreIM:nsibilityofltllls can bt in
the compontnl5 oftht gtntral communication modtl.
3 .. 1 A susptCt apptars in court and reael' 10 a qutSllan postd by tht judge in IM foIlowingway:
A: You alt' John Smlth1
8: l'vt that thlft tlmes already during tht invtstig.tions. Yoo sllould know by now.
hplain in ttrms of objtctivts ;lIld siluation wh.1 is going wrong in Ihis eommunication.
3-4.J 8y using IM Itlltrs S, [, N and G from tM SPL\I(JI'IG modtI, poinl OUI Iht difTtrencts bttwttn
the discourst si tuations in which a doclor and. nurst find thtmstlvts whtn thty alt' working
logtthtral Iht hospilal and whtn tht)' SO lOrdrinks .1 a locll baraft.trwork.
3,5.1 Givt tumplts oftlM: thlft types o( mtaning (idtallonal, inltrptrSOflal an<! ttlllual) in txamplt (I)
in Stetion 1.1.
Qut'$tlOf1S and assignmtnu
3.6.1 In Iht foIlowing tu.mplts thtwonl"run" has difTtrenlmt.llIngs. Dots IM OtItf1Tlin.tJOn ohhe
COll'tct mt.ning havt to do with cohnion orcoMrenct1
a. I'm going 10 wind up thtst old docks I fOtJnd in tht attic, but I den'l know i(thtywill run or
not.
b. A numbtrofltsstr-known andidatnwtre promistd governmtnt funding, butt don'l know
ifthtywilt run or nol.
3.6.J Explain whtthtror not tht cohtrenct crittrion is in t ...amplt (a) in qutsl!on 3.P.
3.6.3 Ost tIM: eooptrativt prlll(iplt and tM mUlm of reml'lCt to iIIuslratt thai IM informatlvtntss
crittrion (st'C Stetion 3.6) is also applicablt in situations in which osttnsibly no ntw information
is btiflg givtn. An txamplt is tht situation in which A and 8 both know thatJohn is aslt'Cp and
thty also both knowthat tht othtr knows Uns, but A 51111 SIrs 10 B: "jolm is slttping."
Assignments
p.l Compare IM Ihlt't (uncliOns thaI Biihlerassigns 10 a linguistic S'gfl (see StetiOn 1.1) wuh the
three-part approach to a Sigfl ofPtirct. tM similaritltS u wtll as on ponible diffeT-
tncts MtWt'Cn thest two ptrsptctiYl!S.
3.4.1 Explain howdlscourst ean ereatt or mocIify a situation by usiflg an tumplt)'Oll l\avt
)'OUrstlf'.
3.4.1 Johnstont (1001) mentions si ... asptCts ohht shaping ofdlSCourst. Elch ofthtcategorits cor-
responds 10 ont way in which COflttlllS shapt discourst and discOtJrst shapesconttJls. Tht list is
givtn btlow. Glvt eumplts ofhowdlscourst shapes and is shaped (o(ueh ofthtse si ... wl)'S.
1. Discourst is Sh3ped b)' the world, an<! discOtJrst shapes Iht world.
1. Discout'St is Sh3ped b)' 13ngu3ge, and discourst shapts language.
3. Oiscourst is Sh3ptd by participants, an<! Sh3pes particlp.lnts.
... DiscOtJrR is shaptd by priordiscourst, 3nd discourst shapes tht possibilitylOt'futuredis-
courst.
S. Discourst is shaptd by ;ts and discOtJrst shapes tht ponibihlyofits
6. DiscOtJrst is shaptd by purpost, and discourst shapes possible purposn.
3.5.1 In Stetion 3.5. tht e.planatlon o(tht flot,on modt con"IIlS' stnttnct tfldlllgwith tht words "and
tht Try to fill ifl. in )'Our own words, wh,t "afld tht tikt- can mt.n in this cantul.
3.5.l Takt a printadvtrtisemtnt (rom a ntWSp.lptrOl" magazlnt and analyzt vtrbal and
"';sual codts b)' USiflg thc triad fitld, ttnorand modt.
Sl
54
3- Di5COUrR in commumcatJon
).6.1 In SecbOfl ) .6 it was oplaioed whll., is that mam d.scourse d.scourse. Now COI'Isidrrthe follow-
In' game" thai is InspIrre! by a rnean::h melhod dlscuss! in johnson-uird ('98)). Wil" l group
offourto five people, have the first person write down a on a piece ofp,i.per.ltl thf' ned
person read this $enlener and Ihen add a srntence. Then fold the plperso Ihal onlr Ihe Sf'Cond
person's is visible 10 the third person. Now the mird perwn adels a sentencr. Repeat this
procedure untileveryone hu rud the predKtssor'S an<! COI'Itribvled a senten. Discuss
10 wh ..., ellent Ihr producrd discourse qualines u discour5e, liven IM definition in Chaptrr,
and Ihr critffia in Se<:tion ]_6. 15 Ihr previous contribution always uSfd 01' is il ... tionally
Repeat lne usignmeflt, 001 inltrad of ooe sentence per person let everyone I'\OW write
down tillft' or senttnCts. Don this yourjudgmentoftl'le qualifiulion ohhe dis-
COIlrse as dis(Oursel
).6.a In this il WlU wh;lt It is 111111 mllkr!;
almosl as manydrfimtions as lherr arr trJltbooks. klow arr a fewuampln (rom
p1Jblk;ltKnu byotllrr ;Iuthon;. Try to point out thr dil'krrm:H be'tlllftn vllriOU$ on
discoul!.
FflH!I Sl:ubln
MOjSCOUl! is language the sentrnce Of the dause."
FflH!I Iro"," lind rult
MThr ;lnalysis of discoul! is, nreessarily, thr IInalysis ofJanguagr in usr."
From TanMn (1988:.;):
MThat moving thaI Cl'ffpi in Ihr words and the
spar\o;ing idrlll, images lind emotions thlll arr not contaiMd in an)'words one at 11 time - the
face that makes words intodiSCOUl!."
From FCllrdoIIQh (IHU&):
'''DiSCOtll!' is formt' rnorr th,n just uw; it is IIW, whether speech Of
writing, sei!n as a type ofsocial pnctice."
,.'
,..
,.,
BlbllOIraphical infotm,tion
Bibliographical infonnation
lhrrr Irr I number of dllfrrenlipproaches and de6m1lons m pragmalla. Though nations of pug-
malics alterN during the pasl decades. Ihr publicalion by Wat71awick, Heavin and Jrlwn
(1961) may AlU bc rcgmled u an oulSl,ndmg conlribullOfi 10 fidd.. I1 u wntten ll\ I magntfi
cent and dur SI)'''', dnllg Wlth I grell numba 01 nnp"icaiAudlO!S, tlllUtral\'d Wllh aamplcs
from life as weil u from the aca&mic domam.
Good introduclionl to pragmalta ale (19g9) and Ihe olllslandUlg OVCTVleW
by lLvinson ( 198l). Of more rreent date is the work by Mey (2001). lhoogh more conci$e Ihan thc
lallu pubbcallOn, Cnmdy's (2000) work usriul.offullI splendid UCTCUcsand qucstloru.
For a suceinci and accesslblc mtroduclion, Wlth commenls., IIvely examples and exer-
cises., s Peccei (1999). Se-r abo Cutllng (2002).
A Iandffi;J.rk pubhutIOn .1 thc accwmed Handboolc afPrflgmlltiCf by Vencllueren, 5tman
and BJomma.ert (1995), 'pannlllg Ihe multiple and ;lpproaches 10 pngmallCS and aplonng
tn dep\h .mportant and central iuues l pragmalics. An onhne venion Oflhc Hllndboolr. O! Prog
",/llia became available in 2003 (sec www.bcnjamins.,comJonhne). also Verschucrcn (1999) for
a comprehtllllVl: and at$5lblt llliroduction to pragmallCL
Astili vnr good inlroduclJon 10 stmlOtlCJ 1$ Uu.1 byTC}IOl (19811).
A good overview ofthe ulensive liltralUrt on rules'5 1ven by Vannesle (1980). SliII one oflhe
bm pubhcallOfl$ on lhe $ludy 01 rules and norms is thc man: philosoplucal approach l Bartsch
(1981).
The general communiUllon model was dtveJoptd by Iht Amer;can malhemllidan Cl.udt Shan
non, who publuhed an Ithc lt .n 1948 enlitJed TM Mllrhtmlll.caI of CommllnWlfIOrl This
art>dt _ reprinled Wlth the additJOn of an mtrodUCIIOn by the scienllfic advisor W;mcn
Se-r Shannon and (1949).
In dlscourse sludies dlfferenllerms art ust<! for wr;\ten and spoken communkallon Ind the
Klon;. 1hf ",us.agt WlIh I and I rrenlYT an be a lul or a rIocumtnl Wlth I wTlfn and a
rtIldnor a con>'Cnlllwn or an mleroetum wtth pattlOJNlnU (a JPmkn and ./Ultnnor an
or a rtC.pltn,). In this booIt Ihe lum discolI"" is uscd u the gentrallerm
)..4 For thc SPlAKll'IG modd, _ Hymes ( 1912). Tbc IttractlOfl mncmomcalldler dasstliatlion is
Arikingly ,UU51nlro by thc Frmch MbpcahOll ofthis propoAl Altu rrgroupmg. Hymcs arnves at
"AII.LANT, which stanch for, parlIOpants, /Kla, rruson (rm.JtIlt), Ioclllt, 4Ktn's, nonrra, lOri , rypes. See
also Vannestc (1980) whut the slluallon is descnbtd wlth the SITUI modeJ: - sdllf', ,nstrumtnwllth,
tliOM, tfftt:
The (QlKepu s.lllllllon and conlm an drlincd in Ihc hlcraturr 111. number of dlffert'nl W<I)'5
rangtng frorn Iht "cry bro.ad 10 lhe more preclSC 5, amons othrn. Van D'Jk (19n ), In Stau
55
s6 3. OiKOUrw in communM;;IItion
gnndr (1981) and Brown and Yuk (1983). whlCh d,stlnguidld bdWttn disannweo",tx' and
eo"ttxf of SltulUro"
For furthrr study the artkId colltttal by Aurr ud LuzlOn (1992), and DuTlnli and
Goodwm (1992). which providr inSlshl inlo thr crntn.l issUd conarning the .d.hon brlwn
langllage ;lind conleXI.
).5 Thr 50o-semlotk approKh ongmatd (rom Halhday and HUhn lhr Ihr kty p"bllOilions on
this maUer HaUiday (1978) and (1994). and Halliday and H;II5$ln (198S). Tlblr 3 is I.sed on
HaJliday and Huun (l98S). Many other di$COUt$C' anaJysu. mainty from the Ult and AUSirala,
work on Ih,s frarmwork. 5, for rumple, KrrsJ (1989) and Marlln (1992). A pod onlroducuon to
lhe systrmic funetional apprwch on di$CourK analysis 1$ Manm and ROK (W(2). wh,eh is basal
on M;IIrtin (1992),
).6 Eaampld {lO);IInd (ll) WeK t;llktn {rom Dm Uyi (1983), aamplr (IJ) (rom Braruford ud Jolm
son(l97J).
lhr KVcn critrria for Itxtllahty wrrr formulalal by Dr 8raugrandr (1981) wilh Ihr cla,m Ihat
tUI iJ non-<DmmUniOlUvr ifthe cnteria do noI apply 10 iI. 1M COrKept oe IrlftTlutUIl/,ty is aJso
Importanl m dlscourK lheory. For an Introductory publicauon in th,S Arr;l, Jtt Mac-
dondl (I986). In the many defimhDns are: provided for Ihr cDncepU roI!a,o"
arnI cohcrc:ncc. A dlstincllOn brtwn Synlu and $Cmantics (for cohesion) on Ihr onr hand and
pragrnaliCi (forcohcrc:ncr) on Ihe other hand iS oftrn madc. Sometimes thr term eomuctll'ltyor
co"tICe'cdMJS iJ uscd for both lerms.
PART 11
Backpacking far a scientific jaurney
4 Discourse types
4.. The varietyoffuncons aod forms
Discourse has manr different fUelions. Therefore, il is not surprising thai discourse
has lIlany different forms as weil. Following Ihe thret function! of languagt given in
Ihe Organon model (S Section 2.1) wt can distinguish Ihr main discoursc types.
Ir lhe symbol a5p1 of languagt. Ihe reference 10 reality. is predominant, Ihen Ihe
fUelion is Ihe transmission of infonnation. Ir Ihe aceen! is on Ihe symptom aspect,
then Ihe function is txpression. r.g., in a story or in poctry. When Ihe signal aspecl is
accentuatw, Ihen Ihe fUction is for example, in an argumentative text.
In schemat ic form this becorn:
Table I . Tbe Organon model as. 5tarting point
Organon modd
.ymbol
symptom
"gnal
Fnctions
InforlJ\.;ltion
fipl'n$lOII
pcrsUUlon
Typa
mfomt.lllVC discool'K
narnuw d,scoorK
d,seount
However, this threefold division is much too simple to serve as a basic scheme for
eovering all the varieties of discourse. Moreover, the functions seldom oceur in their
pure forms. A writer can tell a story in order to persuade people about a certain issue.
This three-part division says $Omelhmg about aspects of language that can playa
role simultaneously. A mOTe cnllcal objection, however, ls that rnany more fune-
t ions are possible. For instance. language ean be uS! to coneeal information. to give
instruelions or 10 instLiI a feeling of camaraderie. One of Ihe most mfluentialscholars
in li nguistics and literary sciences, Roman Jakob$On. who was educated at Russian
universities. distinguishcd in a famous Mdosing statement
H
al a conference on "Style
in (1960) six funclions that can also oceur in combmation. He base<! his
distincllons on an extended version of the communication model discusS! in Sec-
tion 3.3. Below are Ihe functions presented within his communication model .
So .. _ types
2. addrnsc:r
lmoliw
I. oontext
rrfounlll>l

pot/ic
S. mmnd, (Onl"'l
"'",.
6. cO(k
mdlllmgamJ
Flgure I. FunClion$ according 10 JakobSOn
3. addresseoo
((mIlI,,..
A is unI from 10 the This message refers 10 some-
thing in the world. the (ontat, and 15 transporte<! using a code, t.g., symbols. via a
ehannel, the participants in the communication. Tbc ehannel coosins of a
physical and a psychologieal connection, hence, the two words channd aod (ontaet.
Thc most importan! function iSlhe Ot: at the top, i.c., the refcrence 10 something in
the world. the referential fUelion. Its pendant in the Organon model is the symbol
aspccL Two other fUchons are linked with the participants. Tbc emotive or cxprfiSive
function expresses the attitude of the addresser. Thc conalivc function is the orienta-
tlon towards addresstt. giving a command or an instruction. These two func-
lions related to participants or as functions relaled to
symptom and signal in Organon
fourth function in Ihis is language in focus on the message
itsclf. It is called poetie function the most pure form of Ihis function is
poetry. poetie function is also apparent in not only
in puns in ads bUI also when argues that "'John and Margeritha'" sounds
nieer than and lohn': is not only ustd for giving information.
Language is also used for checking the channe! or making contacl. This is
phatic communion, fifth function. Examples of Ihis are "HeUo, you are still
there?" and - in a context both partieipanis know thalthey will have to wail
for a long time - "I gUe5S will havea long wait?" tht languagt focusts
on the codt itself, for "I don't follow you" or "Am 1 ThlS is
or glossing function.
So much for this functional approach to varieliC'S in discourse. that
functions can distinguishtd, for example giving explanations in a course book
Ihis one or making shudder by telling a horror story. Neverthe!eiS, Jakob
son's six-parl division is generally a good framework for discussing the
variety of (mixed) functions of discourse.
4.1 The aod farms
AnOlhtr way to map the variety of discourse is starting with allthe differtnt types
and forms Ihemsclves, from ttltphone cal1s to telephont bills, from hypertext 10 chat,
frorn inttrrogations 10 lectures. lhe names of these various kinds of discourse suggest
that there is a difference between a scientific article and an between a sermon
and a political spch, ete. Especially intriguing is the fact that languagr users can
distinguish different kinds of discourse: "This is not a business Irlter, it's 30
personal one." Ptople can recogniu miStakes in classificatlon: "This is ulled a fairy
lak, but in faCI it is a saga." Changes in the character of discourse an be
as weil: ""At this point the news bulletin took on the character of an editorial." People
also have opinions about the suitability of given kinds of discoune for specific types
of messages: "lhat's not the kind of remark you want to put into the And,
finally, discourse types can be parodied: 30 story can be molded into the form of an
explorer's diary or a civil service letter. It would appear thai people have intui -
tions regarding discourse types. In order to ascertain what intuitions are based
on, it is necessary 10 a system of discourse classification wlthin which discourse
characteristics can be relaled to kinds of discoune.
Many attempts have been made 10 design a classification system. Three appr03ches
can be distinguished in the classification of the variety of discourse. In the first
approach a discourse typology is based on the relation betwecn the discourse situation
and general discourse characteristics. In the second approach abstract forms are the
basis for distinguishing general discourse types to which different kinds of discourse
can be assigned. In the third approach specific laieal and syntact ic characteristics are
re!ated to communicative functions. Below are reprC'Senlative and more or less classic
examples of these three approaches.
In the classification of oral discourse by Hugo Steger el al. (1974). Ihr point of
departure is the discourse situation. On the basis of sociological analysis, six discourse
situations distinguiShed. each with a distinclive discourse type: I. presenlatlon; 2.
message; 3. 4. public debate; 5. conversation; 6. interview. The double quotation
marks denote the fact thai these are not everyday designations but abstraci discourse
types. The discourse situations can be dislinguished on the basis of a large number of
characteristics. Some examples are given in Figure 2.
FOT oral discourse the main division is inlo monologue as opposed to dialogue.
This is the first faClor in Stegers approach. In this cast the enterion is the or
absence of interaction. Dialogue or inleractive discourse ls divided according 10 the
division symmetricn!- asymmetrical. In this system this characteristic is called rank.
Unequal Oleans that the conversational participants do not have Ihe same righls, for
instance, in Iht case of an interrogation or a class discussion where Ihe judge and the
teacher, respeclively, determine who gets to say what. An example of an Iual dis
course is a conversation betwecn two friends.
..
6, Discourse types
Num'-oi OntSptaur
Sptilirt
Multiple Spuurs
Rank Equal
Unrqual
lMmt predetumined
Theme noc prtdl'ttmtined
Mtthod uf Thtme
Treatment
Argument.live
Auociallve



[
.... -. __ ...



Figurt 2. Classificalion by SIeger et al. (1974)






C
5

:g
E
,


1


I


e
.,









In Ihe ease of a concrele discoursc it 15 necessary to first determine the main type,
and Ihr determine the faetors in the discourse situation on the basis of whjch this
kind of discourse distinguishes ilself from Telate<! kinds of discoursc. FOT example. a
doctor- patient talk ean be categorized as an interview. Bul an interrogation is also a
form ofinlerview. Using this model as a point of departure, it is possible to inve5Iiga!!'
which faetor in the discourse situation can bcsl be used 10 describe the ditrerence. for
example, Ihe role of the participants and the goals of the interview.
Egon Werlieh's discourse typology (1982) refers to the second approach: from
abstract forms 10 discourse Iypes. Werlich distinguishes five basie or ideal forms Ihat
are fundamental to discourse types. He argues for Ihe choiee of Ihese five basic forms
by referring to studies on innate categorization possibilities in human thinking. The
basie forms are given in the first column ofTable 2 below.
Werlich rdates these basie forms to specific sentence structures. The character-
istic type of sentence for the rnstrucfive, for example, is the imperative. The mlrrative
requires a certain type of informative sentence whieh has a verb in the past tense and
indicalions oftime and place, as in the senlence: "The passengers arrived in New York
in the middle of the
The live basie forms are each divided into !WO melhods of presentallOA: subjec-
live (Ihe wnler's perception) and objective (which can be verilied by readers). Here
100, discourse characteriSlics are named. The passive voiee is, in Werlieh's opinion, a
4.1 Thevanetyoffunctionsad (orms 6]
Table 2. Werlich', dUcourse typology (1982)
ukForms
(I) Descriptive
(2) Narrative
(3) Uplanalory
(4) ArgumentalLve
(S) InstructlVe
Subjective
Impreuioni$lic deKription
"J'Ort
-,
Imment
Instructions
Objtht
te<:hnical dcscription
ncW$ $tory
expli, atlOn
argumentation
dirtC1ions, rutes, regulations
and statutes
charaeteristic of objective presentation, while the active vaice is typical of subjective
discoursc type$. Thc discourse types determmed in this way must then be further
subdivided in lerms of the channd. for example, the oral channel as opposed to the
writlen channeI. After this subdivision, a specification can be given ofkinds of dis-
course. Then, as Werlieh observes, il will become dear that a specilic discourse can
contain a number of different basic forms. for example, a story thaI opens with an
impressioOlslie description. An important pornt of discussion in Werlieh's approach
is the slatus of the live basic forms. The existenee of innate categorization possibil it ies
is difficult to prove. For Ihis rearon, altempts have 1>n made to make divisions on
Ihe basis of other criteria.
In Ihe lirst approach the line went from discourse situations to general discourse
eharaclenSlics. In Ihe second approach the hnc went from genel"2l forms 10 discounc
types. The Ihird approach relates Ihe co-oceurrence of linguistic features to com-
munieative funetions. In Biber's typology (1989) a restrieted seI of text prototypes
is distinguished on the basis of five sets of lexical and grammallCal featuTC$. Biber
analyz.ed aooul 500 lexts by looking at Ihe way 10 whieh sevenly linguistic features
co-oceurred. The Iinguistie featuI"CS had a wide range, from the tense of verbs 10 lexi-
cal classes (nouns, prepositions. modals, eIe.), from passivizalion to the use of specilie
words. After statistical analysis he found Ihat there wert' live clusters offealures. These
live dimensions .....ere labeled as follows.
(I) The live dimen.!ilons of Biber (1989)
I. Involved versus informalional producllon
2. Narrative versus non+narrative concerns
3. Elaborated versus situatlon-dependent reference
4 Oven expression of persU8SlOn
5. Abstract ,-crsus non-abstract style
In dimension 1 are, on the one hand interactive and affective discourse types, like con-
versations and personalleiters. and, on the olher hand highly informative lexlS. tike
.. 4- DiKoue Iype-s
editorials aod academic prosc. The typH characterized by the presence or absence
of a set o(features. In dimension 2, narrative text s - with, amons other features, many
past-teoSt' verbs aod third-person pronouns - are distinguished (rom non-narrative
texts. Dimension 3 sets Ihe highly explicit contcxt-independent texts, like offidal
documenls, apart (rom all other discoursc Iypes. Dimension 4 characterizes all text
with persuasive elements, such as ads aod politicians' spuches. Dimension 5. with
Ceaturt'Slike push-es, characterizes Ihe abstract aod formal style.
On the basis of these live dimensions Biber distinguished eight lext prototypes
such as -inlimate personal intemction': -imaginative narrativeM aod "situated reporl -
With his stat istical analysis of Ihe co-occurrence of linguistic features and the
linking to communicative functions Bibtr showed that general concepts like narrative
form. upumatory form, upository form and interactivt discouru in other models au
much too vague. According 10 theu is no single cxpoSltory form. Theu are, for
example, blg diffeunces in linguistic features between (abstract) scientifu: cxpository
texts and the socalled "learned expository" in "litente Pr()U" with a more aclive style.
Moreover, in inleractive discourse there proved to bt a big difference between the
linguistic fealures of "intimate interaction" (lhe more phatic communion) wilh a 10w
informationalload and the characteristic.s of the more informalional inlenclion$. as
in telephone convemtions betwcen business associates.
The three typologies mentioned here have not been fullydeveloped into a typology
of aU possible discourse types. In contemporary discourse studies. few altempis at an
all.encompassing dassification can found. In the last few deeades much attention
has been paid to the differences between spoken and wrilten communicalion and
the dialogic aspects of wrilten discourse; see Seetion 4.2. NOlwithstanding the fact
Ihat discourse studies is mainly restricted to everyday discourse, the poetic function
of language manifests itself frequently in "prosaic" discoursc. So. more information
about the poetic funclion is needed to understand the practice of discoursc analysiS;
sec Stclion 4.3. lhe most recent mode in discourse is the computer scrcen. Some char
acteristic.s of electronic discourse are dealt with in a special secllon; see $cction 4.4.
After Ihese Ihree seclions, the problem of type and genre are dealt with again, but now
from a new perspeclive on ordering the varietyof discourse; sec Section 4.5. Chapler 4
condudes with a section about a topic that has been a focus in discourse studies in
reeent yeaes and that will probably shed new light on the facIor mode in variety. lt is
the phenomenon of mu/timoda/lty, which means that communicalion is increasingly
being realized in different modes 5imultaneously; sec Section 4.6.
4.2 Wntten tanglUge and I'fiilal inte1'lllcllOn
4.1 Wrillen language and verbal interacHon
The term discourse is used far all forms of onl and wriuen communication. There
are, however. important differences bnween oral and writlen discourse. According 10
WaUace Chafe (1982), two ractors uplain the differences belween written discoursc
and verbal intenction: 1. Writing takes Ion ger than speaking; 2. Writers do not have
contacl with readers. The first factor is responsiblt for whal Chafe caUs "integration"
in weitten language as opposed to the fragmentation" that supposedly takes place in
verbal inleraction. This integration is achieved through, among other things, the use
of subordinale conjunclions. These subordinate conjunctions occur more often in
written language than they do in verbal inleraction. The second factor is responsible
for the delachment from the reading public in wnlten language as opposed to the
involvement that is present with verbal interaction. Speakers and listeners are more
involved in communication than writers and readers. Thls expresses itself, according
to Chafe, in references 10 the participants in the convemlion and comments on Ihe
topic of conversation. That the involvement in written language is nOI as great is made
clear, among other things, by the more frequenl ose of Ihe passive voice in which Ihe
person who is acting remains in the background.
The diffeunce can also be described in lerms of situation. Verbal interaction is
part of a shared situation that indudes both speakers and listeners. In such a situation,
information is also passtd along through means other than language, such a5 posture,
intonation, hand gestures, etc. Moreover, speakers can quickly react 10 nonverbal reac
tions on the part of listeners. A written discoursc, on the other hand, is not parI of a
shared situation uisling writers and readers.
This difference obviously has farreaching consequences. Yet , there are a large
number of discourse studies issues in which Ihis difference hardly plays a roll' at all. In
both forms of verbal communication. phenomena can be studied that are related to the
cooperative principle, politeness strategies. cohesion and coheunce, stylistic variation,
etc. h is for Ihis reason that Ilddressee or receiver can used 10 denote both readers
and liSleners, and producer can be used for both speakers and writers.
One similarity belween Im and dialogue that is often overlooked is thai. although
wrilers cannot process an addressee's reactions, they can anlidpale probable ructions
and write the text accordingly. lhe following illustrates this phenomenon.
"
66 4. Oiscourse
(2) 1. Discourse studle5 is not a separate science. 2. 11 ean be seen from Ihe
discoursc studies pubticalions thaI have appeare<! up until Ihe presto! Ihal
Ihere are no cammon targets or goals that can be formula!!. from Ihe variouS
research loria, 3. Thi5 is Ihe least thaI ean bt expected from researchers wishing
10 do work in a new fitld of research. 4. Thete art researchers who see in Ihe
cancept Mbreakdown in communication
M
a binding element, bUI cven in this
approach the theoretical underpinnings are al best rudimenlary.
lhis passage can sren as a dialogue in which Ihe contributions made by Ihe coo-
vCr$3.lional partner have becn omine<!. The relationship betv.'cen Ihe sentences can be
made apparen! by inlerjming qutslions.
(3) 1. Di.scourst studlel is not I separate
Ja. How does the author reach Ws conclusion?
2. It can bt sn from the discoufSl." studil."S publications that haVl." appl."ued
up untillhe prl."Sellllhatlhere ue no common targets or pis that can be
formulatl."d from the rl."Sl."arch topics.
2a. 15 this really an argument?
3. This is the least thai ean be upted from rl."Searchers wishmg 10 do work in
a new field of rl."Sl."arch.
3a. There are other eriteria for aseparate science, aren't there?
4. There are rl."Searchers who see in Ihe concept "breakdown in
communicalion" a bindmg element, but even in this approach the
theorelical underpinnings are at besl rudimenlary.
This example of dialogue aspecls in discourse shows that wrilten communication ean
also be studied from the perspective of a situation in whieh verbal interaction takes
place. The scholar who gave a big impulse 10 Ihis view of (wrillen) discourse as dia-
logic interaction was the Russian philosopher and liter.ny theorist Mikhail Sakhlin.
He developed his ideas in the 1920s, but only after the translation ofhis work ioto
English, more Ihan half a cenlury later, did his 'deas become influentia!. A central idea
in his work is that language in use cannot be considerl."d a set of words with abstract
meanings as described in dictionaries, but thai the meaning of words is actualized in
discoufSl." OWlg to the Iteraetion ofthe participants.ln other for all uller
anees only same potential meanings ean be given with referenee to dictionaries. but
it is the parlicular SlluatlOn that determines which meaning is actualized. UUerances
Hke "lhat is a ni ce can ha\e different meanings. but the situation aClualizes
the sprcific meaning of admiration. irony or desire to buy il. etc. So. if you eonsider
language to be an abstract system of struelures and meaning. il will nol bring you far
in discourse studies.
4.] Evoeryday and 111.11"}' language
The right way of studying discourse. aceording to Sakhtin (1981) and many other
researehers, is to view discourse as inherently dialogie. Also in wrillen discourse utter
anees are responses to other ulteranees. for uample. quest ions provoked or expected
by readers. And for discoune eonlaining references to different persons. for example.
a nove! with quotations from different pcrsons, Sakhtin eoined the music metaphor
polyphon(. This "multivoicedness" is an important faetor in many types of discourse.
for example. a news story in whieh the journalist reports the viewpoints of different
aetors, or in a judicial report describing statements made by diffe.rent witnesses.
It is prrciscly this dialogic and sometimes polyphonie aspect of discoursc that
provides an argument to study written language and verbal interaction in the single
discipline of discourse studies. The issues Ihat specifically apply to verbal interaction
are deall with in Chapter 9. In the other ehapters. both text and talk are under eonsid-
eralion unless otherwise slatcd.
4.3 Everyday and Iiterary language
In principle the term diSt:ourK is not used 10 dislingulsh belween everyday and liter-
ary language. Of course. there are important distinClions. Literary language does. after
all. scrve a very different purpose than. for example. informative language. Literary
elements, however, can also be found in everyday language. But what makes a kind of
language use literary?
An important difference between everyday and literary language ean be demon
stratcd by way of the following statement by Jakobson, who was introdueed in Sec
tion 4.1 on funetions of eommunication. He was one of the first researehers to pay
attention to Ihe poetic funClion in communicalion. This is his famous, bul diffieult
statement about the poetie funelion .
(4) The poetlC funetion proJI."'C1S the principle of equivalenee from the axis of
sdrction into the axis of combm3tion.
For a beller understanding of Ihe terms seleclion and combination. two important
asprcts of language, the syntagmatie and the paradigmatlC aspecl , need to be dis
eussed. These terms can be explamed u$mg the concepts horizontal and vutical.
The syntagmatic or horizontal aspect has to do with syntax, the combination of words
in a sentence. The way these eombinations are made is go\"emed by tue<! rules. The
eombinations possibl e in everyday language can be described using rules of grammar.
for example. the rule that a verb like to go cannol be followed by a direct object_ Sen-
tenee (53) is not English. sentenee (Sb) iso
"
68 4- OiscourR typ"
(5) a. lohn went the schooL
b. John went 10 school.
Thr paradigmatic aspeet is the vertical as in in Ihe sense of a list
of verb {orms: I wall<, you walk, he walks, eie. Instead of John in (Sb), a word can be
substituted from a whole list of ather words such as Ihe man or Ihe girl. Thc same holds
lrue for Ihe wordSlhat follow "'ahn" in the sentence. In this way il i5 possible 10 gener-
ale a sentence hke: Pete drove 10 the beach."
In everyday languagt the paradigmalic selection procw is simply a malter of
choosing words thaI are calegorically equivalent. "Equivalent" in Ihis case means that
the elements mus! have something in comman. lohn can be replaced by Pete hut not
by is or two. Thc commonality can consist ofboth words being Ihe same kind of word,
or of both words possessing the same meaning element, in this case a person doing
something.
The syntagmatic element involves the horizontal combination axis, while the para-
digmatic element involves the vertical selution axis. Jakobson's remark implies that in
poetic language the syntagmatic axis is somehow special and that this special quality
has to do with the choice based on equivalence along the paradigmalic axis. It might be
$lid that in poetic language. the syntagmatic axis is oflesser imporlance than the para-
digmalic axis because Ihe syntagmatic axis is influenced by the paradigmatic one.
Consider the following example. When an individual wishet to mue it dear thai
he would rather take Ihe car instead of the train, he has a number of possibililles,
including those in (6):
(6) a. Give me the car any day.
b. Driving i5 nice.
e. It's great to bt behind the wheel.
d. Alive when I drive.
Example (6<1) is by far the mosl poelic. The influence of equivalence from the para-
digmatic selcction axis on Ihe horizontal combinalion axis is obvious in Ihe first and
Ihe last word. In everyday language the equivalence is limite<! to one position on the
combination axis; in (6a) and (6b), another word may be chosen for or
In poetic language, equivalence manifests ilself in multiple positions. In Ihe eumple
above, the words -alive" and -drive- are equivalent because they rhyme. This type of
equivalence is called pro)Ktion. Equivalence is obviously nol Jusl a quest ion of rhyme.
Jakobson al.so mentions the repcaled use of the same grammatical construction and
makes special note of the parallelism phenomenon, for example, the repelition of Ihe
same patterns in different lines of poelry. This kind of repetition does not nave to be
containe<! in a single text. A sentence can also have a poelic funclion on the basis of
4.4 ElKtronic diKOUrse 6g
intertextuahly, forexample. because Ihe structure of Ihat sentence IS reminiscent ofthe
structure of a sentence from another kmd of discourse. This is the case in the following
example taken from an advertisement.
(7) Quiet type seeks acquaintance with provocative sweatshirt.
The slructure of this sentence bears a distinct resemblance to Ihe type of phrase
often seen in personal advertisement5. This form of parallelism is. in Jakobson's view,
responsible for the poetic character of such a sentence.
Jakobson's statement explains clearly the difference between language with a focus
on the encoded message and olher typcs of language USt. lnsight into this poctic
funclion is important for discourse studies because this function often occurs in
nonliterary discourse, for example, in advertising texls, graffiti, flyers and newspaper
headlines. Also stylistic techniques such as plconasm (e.g., round metaphor
or personiflcation (e.g., Mthe wind howls") cannot bc studied adequately without refer
ence to this poetic function.
4.4 Eleclronic discourse
Since the 19705 a new mode of communication has emergnl. lt is called electronic
discourse, web communication, computermediated communication or
e.language. The impact ofthis communication mode is considered as big as the impact
ofthe telephone and television was. Hence, many researchers and professionals expect
that this new communication channeI, and especially the Internet, will change our
communication pauerns.
Just as spoken and written language are cover terms for different discourse
types, computer-mediate<! discourse also refers 10 different discourse types. A dis-
tinction currently used is that betwun synchronous communication and asynchro-
nous communication ofwhich the most common discourse types are presented in
Table 3.
Table 3. DiKoutse typn in computer. mcdiatcd communkallon
Enmple
5ynchronolls
asynchronoul
chat groups, instant MUDs (multi user dim(niIOns,
for rKreation .nd (ducation)
email,discussion hsts, W(\)silcs
"
4. DiKourse types
acronymslike "bbfn" ("bye-bye for now") often in combination with littral"chal11cter
pronunciatioo" as in "ur ("you 100") and "icq" n Sttk you").
A more essential dlfference with oral or written communication would be a sty-
listic difference in wording or syntax. A representative example of research in Ihis
arta is the sludy of oral aod writlen aspecls in computer conferencing by Yales (1996).
He compared large corpora of spoken, written and computer communication, and
analyzcd languagt U ~ according 10 two of thc thrtt aspts of the socio-semiotic
approach ~ Section 3.5): field aod tenor. In field, Le., what it is in discourse thai
participants are engage<! in, he counted the use of modal verbs like musl, mar and
should, indicating how the participants !"elate 10 the aclions described in the discourse.
In tenor, the relationship bet .....een the partlcipants. he counted the use of pronouns in
the first and second persons. Table 4 shows the resulls.
Electronic discourse differs significantly from speech and writing in the use of
modal verbs. This could indieate that the field of computer. mediated communica-
tion differs from the fields of speech and writing, presumably because it contains
more informal discussions m whieh the use of modals is more appropriate. Al; for
pronouns, electronic discourse is more Iike speech than writing. An explanation
could be that participants In writing are not as salient as in the other two modes of
communiealion.
A very restrietive condition of computermediated discourse is that only one
screen at a time is available to produce and understand a message. Tbis phenomenon
has led to an extensive use of a device we previously knew only from sc:ientific books,
Le., referring 10 other information with a number or olher link to footnotes or end
notes. Moreover, the demand for extracting information from large databases on the
basis of individual needs and preferences, such as in reference books, alphabetic Iists
or tabtes and schemas, hu prompted furlher developmenl of non-linear struclure,
as, for example, in menu-based telephone systems (- If you would Iike information
about X, press 3 ..... ).
The number referring 10 a note or a special question answer pattern in pro-
grammed phone conversalion is called a link or hyperlink in electronic discourse.
Tbls Iinking is the main structural characteristic ofhypertexl and is the focus of much
research into the design of computer-mediated discourse. HyperteJ;t is electronic dis-
course in whieh pieces of information are I1I1ked through references, mostly leading
Table 4. Some dilferences belween spuking, writing Ind eleclronic dlscourse
modal verb. (permillage)
w l ~ and . you' (pcrcentage of rronouns)
.p<h
14.5
58
wrillnll
1).7
27
dectronic difCoulK
18.3
..
.s
4.5 Conventionalized forms forconYtnlionalizfd occasions
towards the information referred to. This link - usuallya highlightt'd word or phrase,
or an icon - may just bc an annotation like a pop-up link eJ;pllming I difficult term.
Sut hypertinks Cln also refer 10 sections and subsections (like headings in a book)
or to other information on topies and sublopics related to the message outside the
webpage. Research into the struclure ofhypertexts ineludes queslions like: How does
Ihe -departure and arrival information" on the Hnhng labels aff1 the Usablhty of a
.....ebsite? Where on the screen are hyperlinks preferred? How to give information about
a tinking path 10 prevent the reader from gelling lost? From a discourse perspeclive
this research is very important because the non linear structure of many etectronic
discourse types seems 10 bc the characteristic in whieh eomputer. mediated commu
nication difrers mostly from spok.en or wrillen communieation.
Conventionalized forms for conventionalized occasions
Tbe attemplS 10 describe the different functions of language use, discussed in Sec-
tion 4.1, proved not 10 be comprehensive enough to eatch the parameters offunctional
variety, Moreover, the development of elassification models prompted quest ions about
what precisely the characleristics of the differem discourse types are. For some main
types the differences seemed to bc elear, see the sections 4.2 and 4.3, but an accurate
description of the different genres proved to be an unfeasible goa!.
Tbe difficulties in describing discourse types led to another approach to genres,
which was again inspired by the Russian philosopher ofliterature, Bakhlin, who was
already introduced in the discussion about the dialogieal aspect of written discourse in
Section 4.2. Important in Bakhtin's work. is the focus on discourse use in spccific situa
tions. In situations that are more or less the same, the discourse will have more or less
the same characteristics. In siluations such as crossexamining suspects, presentlOg
acquired knowledge in an edueahonal situation or commenting on news facls, spedal-
iud forms of discourse will emerge which can fulfitl the demands of the participants.
In the aforementioned situations these forms would be interrogations, doctorallheses
and edilorials. And precisely knowledge of how 10 pracliee these conventionalized
forms of discourse gi\'es the opportunity to fully participate in a spedfic comnlUnica
tive situation.
Tbc most important approach to genres as conventionalized forms was developed
by Swales (1990). His characterizations of genres ean bc summarized as folIows:
"
"
4. Iyp"
(8) GtnrC', as ddined by Swales (summary)
A genre is a dass of communicative tvl"nts with share<! rccogniz.ablC'
communicative purpost!l. These purposes give rise 10 C'xploitable constraints
concerning conten! and form.
A communiCitive event is an evtnl in which discoursc plays an indispensable' toll',
fot examplt. a Iture as opposed 10 driving a ear (which can be done almost without
discourse). Genres are not ddint'd on thc basis of similarity in ltxical and gram
maticaJ ftaturts or lotende<! audiencc or channeI. Genres are primarily conceived of
as communicative vehicles for achieving purposes. These purposes are shared byand
recogniz.able 10 the participants in the communicativc event. Were Ihey nOl, then ooly
a partidpant's private purposes would be at issuC', and private a10nt art: not
strong enough 10 shape lhe discourse in slruclure and slyle.
Gwen Ihe shared purposes, lhere are re5triclions on conlenl and form. Swales
givt'S the cxample of the posilive lelter and the rejection lener after a job interview.
The purposes of these two genres dictate the content, the SlruClure and lhe slyle of the
discourse. A positive leiter will have an enthusiastic and invlting style; the negative
leiter will ha\'e a more formal style and be focused on ending the conlact. These con-
straints, however, are exploilable. Expert genre writers can be very creative in dealing
with these constraints. Hcmce, there i5 a difference a standard rejeclion Ittter
and one wllh more quality in which, for example, same personal statements are made
aboul the opportunities of the rejected applicant . On the other hand, these
constraints have to be learned before they can be exploited. Swales gives a niee exam
pie: an opening sentence by a Ph.D. student in aquaculture who has trouble writing a
research artiele.
(9) In aquaculture. the relations among nument$, slockmg rate, water quality and
weather arc compla.
Why does this sentence nol fit into the genre of a scientific artiele? This sentence bet
ter suits a textbook, as the specialized readers of a journal about aquaculture already
know that the relations are complex. In adding that knowledge to the formulation the
sentence is made more suitable as an opening sentence.
(10) In aquaculturc, Ihe relations among nutrients. slodang rate, waler quality and
weather are known to be complex.
This reformulation can only be made if one knows the impact of the shared purposes
of the genre "research artiele" on the style of discourse. How can this genre theory he
made applieable to the analysis ofthe variely of discourse? In Shatia (1993) ihis Iheory
is demonslrated in an analyt,ing techmque in whieh the shared purposes of a genre are
4.5 forms oc:caSIOflS
reformulated Ifl moves that must be recogni:r.able in the dlscourse. Selow is an example
ofthe analysis of a research artiele abstract. The purpose of an abslract is 10 indicate the
conlent of a paper in advance. Why does the following passage count as an abstract?
(11) This paper stU out to examine two findings reported in the literature: one, that
during the one-word stage a mild', word productions are highly phonetically
vanable, .nd two, !hat the one-ward stage is qualltatl"dy distinct from subsequent
phonologIcaI devdopmenl The complete se1 of word fonns produced by a mild I1
the one word stage were: colltcted and analyzed both cross-sectionally (month by
month) and longitudinally (Iooking for changes over time). It was found that the
data showed very liltle variability, and that phonologieal development during the
period studled was qualitatlvtly eontinuous wilh .wbsequent devdopment. It Is
suggested that phonologicaUy pnnelpled de\"C!.opment oflhis chils fust words is
rtlated 10 hiSlate onset of spe-ech.
An abstract is characterized by the purpose of providing eoncise information on the
four aspects of research that it describes: J. What the author did; 2. How the author
did it; 3. What the aulhor found; 4. What the author conduded. Answers to these four
qucstions are given using the following four moves.
(12) FOUf movts in. research artide abstract
1. Introducing the purpose: This move gi\"eS a precise indlCation of the
author's mtention, thtsis or hypothtsiS that formed the basis ofthe research
bring reported. It mayaIso mdude the goals or obJccliVCS of research or the
problem that the author tackled.
2. Describing the methodology: In this move the author givts a good
indiealion of the aperimmtal design, induding mformation on the data,
procedurts or method used and, ir nccessary, tM scope of the research
brmg reported.
3. Summarilmg the resullS: Th_s i5 an important aspect of abstracU wherc the
author mention! his observations and findings and also suggests solutions
to the problem, if any. pose<! in the fiT5t move.
4. Prestnting the conelusions: This move is meant 10 interpret results and
draw inferences. Jt typically includts some indlealion ofthe implications
.nd appliealioru of the presenl findmgs.
When the four moves are applied to the abstract given I example (11), it can be secn
that the first scntenee introduces the purposc, thaI the second sentence describes the
methodology. and that the third sentence summarizes tht" rcsults. Tbe last sentence
prescnts the conclusions. Figure 3 provides a schematic overv;ew.
75
,.
4. Oiscourw typn
Thi.$ K'U 0\.1\10 uamint two findinp rql'Orlrd in the lIteralure: one, Ihal dunng the
one-woni '''ge I child's wort! produclionlare highly phonctically variable. and two.lhat
the oneword Ilagt 11 qualitalively disl!"'1 {rom ,ubsequent phonologle.l dtve]opment.
Ihr compltlc of word forms produad by. child at Ihe oe-word stage Wert collccud
and analy-ft'd both cl'OsHtionaily (month by month) and longiludilUllly (looking for
dwlgn over timt) ,
h WH found thaI Ihc dill showrd very hlde Yam.bdlty, and thaI pbonologK.al devdop-
ment dunng thc period studit(\ wu qUlhllhvcly continuous with lubsequmt
mcnl.
It is suggcsted that phonolO(!;iully principltd dcvclopmcnl of Ihis child', first worcb b
rdatt<! 10 his1.lc onse! 0( 5pCh.
Thcfourmovnin(ll)
2
J

In the genre thcory dC'alt with in this section, the focus is on conventions in form in
recurring situations. In Ihis view, discourse types can !>est be characterized by consid-
ering them as a scries of moves to reach communicativc goals. In the last scction of
this chapter the focus is on the mOOes that are uscd in rcaching a goal.
4.6 Multimodality
lhis chapter started with the various functions of language usc: giving information,
persuading an addre.ss, making contact, etc. These functions, howcvcr, seldom occur
in a pure form. Mostly they are merged, for example, in a TV program that mixes
information with entertainment, or when a company tries 10 seil
more prOOucts by giving information about something new, the
The channeIs, which are used are also mixed. We watch television and simultaneously
we read thc subtitles and hear lhe sounds. Thi$ mix of mOOes, whkh is nearly always
present in communication is alted multimOOality. In discourse studies the simulta-
neous use of mOOes was neg1ected a long time. The study ofthe discourse itself, b<- il
spoken or written, uncovered more than enough interesting problems. But the last few
decades saw so much mixturc of modes, especially the visuali:r.ation of communica-
tion, that multlmOOality has become an Important factor in discourse studies.
One upect of multimOOality which has rcceived special attention in discourse
studies is the co-deployment of the visual element in written discourse: a text with a
diagram, a plcture, etc. There is old folklorc wisdom in the saying pkture is worth
more than a thousands words." Below is a well-knowo example from Bransford and
4.6 Multlll'looality
!ohoson (1973). Many people who rcad the following passage have trouble under-
standing what it is about. They cannot a pieture" out ofthis.
(13) If the balloons popped the sound wouldn't be able to carry since everything
would be 100 far away from the coneet f1oor. A closed window would also
prcvenl the sound from c.arrying, 5ince most buildings tend to bc weil insulated.
Since Ihe wholc operation depend5 on a 5teady f10w of electricity, a break in
the middle of the wirt would also cause problems. Of course, the fellow could
shout. butthe human voke is not loud cnough 10 c.arry that far. An additional
problem is that astring could break on the instrument. Thcn thcre could be no
accompaniment 10 the message. It is dear Ihat Ihe best situation would invol ...e
1tS5 dislance. Then there would b<- fewcr potential problems. With face-Io-face
contacl, the Insl number of things could go wrang.
1t is hetter for Ihis passage, with signs allhe symbolic level, 10 be vizualized, thus prc
sented at an konie level. Then, words are even supcrfluous (see Figure 4).
\J .
. . (,
f
'" I 0
, 1-::':
o "
o I \I

't
Figurc 4. Pieluraliution of (13)
77
"
4- OiKOUrw I)'pn
Sometrmcs it is clear how a piclurc. a schema or a diagrarn can replace a text. hut it is
far (rom dear how visual elements can add 10 verhai information. In studies
same analysis schemes have been presented for research into illustrated discourse. A
scherne thai is suitable 10 start with is the GeM model of Judy Delin and John Balernan
(2002). GeM stands for genre and nwltrmodality. Thc aim of Ihis model is 10 provide
dala on different genres using combmations oftext, layout, graphics. pictures and dia-
grams. In Ihis model. multimodal documents are analyud on five levels of structurt
thai ddlne the possibili ties fot e1tronic or paper documents.
(14) Thc GeM modd
I. GontImt structure
Thc raw dala out of whrch the document is constructtd.
2. Rhetorical structure
lhe way the conlent is "argued': the r<,lations b<'tween
content elements.
3. Layout structun
The nature, app<"aranc<, and poslhon of communicatlV<' elements on pap<"T
4. NaYigation strue/un
The ways the int<,ndl mode{s) of consumphon of the document iJ/afe
supported.
5. nguistk sf1"\lefure
lhe struclure ofthe language used to TeaIlze Ihe layout el<'ments.
lhe main advantage of this model LS thai il can b<' used for analyzing both textual and
visual meaning. Subsequ<'ntly it can motivate and question some of th<' chokes that
have been made in the presentation of information on a page. How does this seherne
fOT analysis work? The authors ofth<' GeM model present, among others, the following
uample: a page from a blrd book (Figure 5).
4.6 Multimodaloty
Figure 5. Ganne!, 7he Observer Book 0/ Bird5 (1972:22)
The content structure, !he -gannet can bt organized into topie segments as
illustrated in Figure 6. lhe hierarehical representation shows the relations between
the content parts, wlthout distinguishing linguistlCally and graphically presented
information. The hierarchical structure does not account for why which information
i5 placed wher<' in the page shown in Figure 5. Moreover, the content hierarchy does
not explain why some ofthe informalion is presentN in one way and some in another,
but it queslions whether Ihis was the right deeision.
79
So Oiscourw typ"
gannel
I
I I I
name appearancc bthavior IilZC geographiea1
r
----"----, dlstnbution
I I I I
mature (<<ding brKding habitat u ll
rl h
nest YOU"g usual breeding
Figurc 6. Hicrarchical Tqlrnc'ntatiOn oC content of Ihe gannet pagt
The rhelorical slruclure pertains 10 the way in whien the content is argued and the
various segments are interrelated textually, Figure 7. At the levd, the con-
lenl is divided ioto the pkture and thc texl. The major part oflhe text is organized ioto
text segments thai provide other facts aboul Ihe gannet: wings and wingspans in the
first 5cnlence. fishing behavior in the second, followed by a segment on plumage, elc.
Spific lexl segments at lower levels further specify thc concept of gannet. The lilie
information scrves as a 10 the main [ontent.
The layout structure is characterized by a division in three zones: the title infor-
mation, the gannet illustration and Ihe text below. and spacing in the lilie

""",,"Ni

family
\ JOInt
Sulicbe pnndS


"

""",,"Ni

latin ganlWt
illustration

Highl, lishing
.u
-L
CO-S/,ijialt/Ofl


1"'"'
...
L
L ktc. 1

joint

sighling
fi'"
dropping adult immatu",
Figurt 7. Rhttorical SlrUClure o( the ganne! page
4.6 Mullllnocb.hl)'
information vafY. The bird's name in English al the top left in hold print signilies that
it is the most important part of the tille information and serves a5 a starting point for
reading. However, il remains undear why, for example, "Sulidae
M
is in small caps aod
the English and Latin family names are illogically placed. lhe piclure ofthe gannet has
a central place. The two textual components that follow both have a different alignment
IOd font size. The split of comparable mformation across these two different modes of
presentation seems arbitrary.
At the navigation level it can be seen that the bird entry number 27" has to
do with the interna! struclure ofthe bird book and has nothing to do with gannets.
It is nol necessary to place this information within the block that is meant for gan-
net -related lilie information. Moreover, because Ihe entry number is nol sufficiently
distinguished (typographically) from other unrelated information, il fans 10 stand out
as a usable navigation element.
Intereslingly, the hnguislic structure is the last oflhe live structure levels. In the
main textual 5ection complele daU5eS are used, while the labeled lisl is in telegraphic
style. lhese two types of linguistic slructure evoke different expeclalions related 10
these typographie conventions. The main bodytext is $Omewhat involved, as Ihe author
evaluates the gannel ralher than presenting il factually, e.g., the bird is described as
"this greal having "magnificenl wings- and it will like an arrow': Language
use and typography dearly go together, as the author refrains from lively descriplions
and complele sentences in the labcled list.
According to Delin IOd Baleman the GeM model makes 1I possible to detect Ihe
motivation for some oE Ihe choices thaI have been made in rendering information on
a page through analysis al differenllevels of struclure. Using the GeM model makes
I1 possible 10 uncover slippage between various levels of a multi modal documeni, and
enables an adaplation of typography and imagery 10 conlent and rhetorical structure.
Finally it may give insighl into why, in some cases, a designer has made a de<:ision that
is cJearly a compromise.
s.

4. Oiscourse typn
Questions and assignments
QUt'$lions
4.1.1 Which ofWtrlich'$ boisic forms un occurin <In Informative discourse1
4.1.1 Using tlte model, in which category would you platt Ihr foIlowing kinds ofdiKourwl
J.f:rmon,llIdio nlWS bulk!I". orol uom
".l.1 Oncribe some oflne important betwe-tn verbal intef'Kt;on aod written discourse
using IM (oIlowing 9.lmpl4:s lakm from a .-sp.aperarticlt on Ihr atttmpted bribery of a drug
Thc newspaper .. rtielt is given in (a) and Ihe oflhe COIM!.1.tion IMnllOflcd
in (a) is given in (b).
Indicatc !owha! CJ.lrn! Ihr diffcreocn n;une<! byChafr apply in this ease.
a. Oru .. ( ... ) gave uttnsive tntirnooy orhis cootaets with Revenue SeMct
M,V. S.E. h,lld In incomc' tu. usnsmenl noticr whidl had Ied 10 a lien pul on
his house. Wheon No_nt on 10 talk .. boot Ihf: 80,000 turDS V. Wliling 10
lowerthc amourlt 10 30,000 rums. In a latcrconversation V. oITere-!! 10 lower S.E'$ IU dtbt by
.. noiller '3,000 rUrM ifS.E. WOtJld pnwide incrimil1alil1g Intimoll)' agail1st Ihe Happy Famil)',
a )'OI.Ilh arlter. S.E. had a l.lItet"discussiol1 il1 his carwith V. eoncernil1g Ihis issur. Thr dealet"
laped Ihis COOYrrsaIlOI1.
S.E. tntined that V. lold him that his statemenl5 had to torrrspood wilh two depositioos
that the Revenue Service had provided Oll the Happy Farni l)'. S.E. stated that his impmsiol1
was t hat lhe Revel1ue Service wu a\temptil1g to -l1ail- lhe Happy FamiI)' il1 Ihis wa)'. V. did
1101 sa)'exactlywnal he wallted from S.E., but thr tonversatlOIl Ied S.E. to believe he was 10
ronfirm givel1 tntimony. Tbis tntimony deah wilh thr of drug saJn
takll1g place al the Happy Farnity, al1 issue Ihat had led t lle Revel1ue Service to start proce-
dum agail1st the )'Outh teIlter.
b. S.E.: listeIl, I want 10 malle sure thai ifl pa)' the '7,000 eums 1'11 be rid ofit afterl tntif)'.
V.: That tal1 be arral1ged, belleYt me. And I wallt, wilh)'Ol.l, wilh rtgard to Ihat, I will
dec:lare il1 Ille prtSellCe of ... (IUlme ul1il1telhgible, ed.l who is thefoe, I hopr il's oot someorM'
else thel1)'01.1 cal1 always sllrt civil proc:edures agail1sllhe RtveIlue Service, Ihrn)'Oll can
sa)' ifl dedare it in fronl ofboth of)'Oll, then I will give)'Ou same legalltverage which WOtJld
make me prett)' uh ... al1d besidn thai 1,)'Ou eould verif)' it )'OUrsclf, jusl call Mr. B, 110 prob-
lem, aod lhen )'01.1'11 hearuactl)'w11at the agl'ftment iso There WOI1'1 be all)' mal1,ptJlatil18,
that's oot In)' style. I wal1l us 10 be able to Iookexh otherstl1light il1 Ihr eyr allCl ifit's worth
it to us, we'lI pa)' for it. You shouldn't have to deal with il afterthe (act. Wilh rtg,ud to Ihal,
Ihere is one eonditiol1 and Ihat condition is that the statement,s eoncrete, that 'Nt ein do
QueSlions al1d assignmenls
somethmg \!fuh itallCl there are, juS! between)'Ol.l allCl me, two others wnich have 10 do w,th
Ihe Happy Famity and we know a Iot about Ihal.
4]-1 UsingJakobsol1's statemenl, explail1 the portie (uncliOl1 of-Ufe is a disaster in spite o()'OIlr mas-
ters 01'1 a b.a11M1'duri"lt a students' protest).
4." " Whal does mean il1 a COOYrrsation and il1 I hypertext orenal group1 uplain tnrdifTer-
eIleH .
.. .... 1 15 lhere an)'dlfl"rrellCe betwe-e11 lurk,l1g,l1 a ehal group and taYHdroppmg m a (a'-IO-fa' coo-
vrrsatiol17
.. .... ) In astud)' by Oavis al1d Brewer(I997) il wu showl1 thaI il1 ehal group cOl1lributiol1s participal1u
relied on privale verbs (e.g., !hink, (<<I, kllDW). Private verbs are lhose where thr activit)'cal1not be
publiclyobscrvrd; thtyCOfltrast w,lh thr ptJblic verbs such U SIl!l allCllr!l. uplail1 this mull and
why it is in line Wllh mults Inftllioned il1 Section 4.4 Oll thr use of peOl1al prooouns.
4-5" Which characteristics of !iOdal rules that were mentioned il1 Section 3.1 are important in Ihr ootion
of nJles il1 5wales's defil1ition ofgel1re7
... 6.. Considerrumple (I) and Ihr illustration il1 4. Name thlft' elements Ihal Ihr texl il1 (I)
adds to the Ylsual prescntatjon ", F'gure 4.
Assignments
4-1.1 A large l1urnbero( attempu 10 deslgl1 a discoursc dasslficauon system are ofliterary
scirntific rrstan:h done il1 thr area of genre Iheory. 111 Ihis theory Ihr fourgemn -fairy tale-,
"saga, and "legend- dislil1guished accordiIlI to Ihe two faetors -religious and
"hislOrical-.look up in Ihe literature whal is said about these (our gel1rrs al1d Iry 10 dtlil1e them.
GiYe an rumple o(eaeh (booIc er slory lilie). Then poil1t OtIllhr d,ffe1oencn through a das-
sificatiol1 by scormg thr grnm (+ or-) 0/1 lhe futon; religious" and "historieal".
4.1.1 See Bakhtin's viewo( writtel1 discolll'Se in 1.4 (dialogk il1leractiol1): the meal1;ng ofthe
words js aelualiled il1 discoursc ow"'l to the interaction o(the partlcipal1u. In which wa)' does
this appro;Jeh difl"erfrom the lat", eXPlYssion "lItrbo uoltnt !/Su-, wnlCh IMal1S "words get Iheir
meanil1g il1 their lIsc"7
... )-1 Take an advert,scment lext and try to al1al)'Ze whieh SI)'listk or li terary techniQun (such as pleo-
num, metaphorand prrsonilication) were used in wril,ng this te.l.
"
.. 4- Discourse types
4-4.1 Try 10 why ar\ t-mai! message ofttn lacks an opening builimosl alwars has an endlng .
..... . J. Vcrify in handbookl oron websitcs on clearwrlting wht san of advice is given on dur langu.ge
use in computcr-mtdiate-d commLlnintion that dooe5 not appl)' 10 writtcn ororallanguage use.
4.5.1 rake Ille summaryOt"abstract mentiooed in uSlgomenl,.l.l ofChaptcr I. and detcnninewheth.er
or not it complies with IM fourlT\OYH INiI are mentioned in 5lIon 4-5-
4.6.1 Takt l tut of)'Ourchoict' and anatyzc it aecording 10 Ihe levels orlhe GeM model.
,.'
Bibliogr.llphical information
Bibliographical information
jakobscm (1960) borrowffi Ihe term pMt,( "'rtHlllmion fmm IM lrllhropologl5l MalinowsJU (1930),
WM omcrvfd Iht phenOllltnOn iUllOng a "J'r!mitlW' pc"Qplc in tht 1920$.
lakobson !IJIlrW Iht disclWlon .boot funchons. PubhcatlOfll ofhlstoncal imporUn<t In this
Iidd an 1",* by (1974) and Werlich (1982). For _ Lnrormahon on dusifia.hon thaI
mcludn tht Slt\UlI/Oll;U a cntmon, 1ft Ountrr ( 1981). Nowadaf$, Ihr woB of BJbrr ( 1989) JS Ihr
mosl ciled wori:..
.4-) Bakhtin developcd his Kkas in Ihr period aftrr thr Commulst Revoluhon. BrcaUK of
con/hels wilh official Sovid htrnrydogma, B.khlL is $upposcd 10 havr also publuhed undrr Ihr
rrallWm1 of othrr peopk. See, for aampIt>. Morris ( 1994) for I coI.lhon of ilrtKks by Bakhtm.
Mcdwdev and Voloshmov. 1M lattlff" twoarc' fnmds undtt whose IWM Bakhtl wrotr. Srr funhrr
Wales ( 1988) ror an Itmduction to Bakhtm.
The suggtstlon th.;J,t a dl$COIII"K can br &$ "half a an also br foond in Roulct
(1984). Set; furthrr Nystrand (1986), whlch glve5 much attentIon to an!LClpallg rradrrs' upta-
lions and to thrit necds. A good start;ng pomt fot further l"r$Cuch is Van Kuppcvclt
(1995).
4-) 1"ho!' ddUuUon of pocbc Ianguage [a bc found in Jakob$on (I960). 11us ddiru\)()lllw
much dlsowlOO m thf, 6dd ofhtcnry $C1mCc. For a cntiquc, $C\" Wmh (1976).
4.4 Dm ofthe firsl publlCa\lOn$on computrr-mcdJated di.socoursr iJ /On" (1995), who focUSCI on ncws
groops. Roud, l.evoncn and Dillon ( 1996) present emplrieal research ;nlO IM proccsscs
involved m IUmg hypcrlut. Srr abo Hnnng (1996) for a colItlOO of lChoIarIydiscunJonl from
cross-culturaJ t%anuC dtroruc: dlSCOUI"K. A good mtrodu<;llOO to thc anaI)'$I$ of
MlJD5I!Chcmy (1999).
Consult Crystal (2001) whm slarIll\g Wlth the ITlIIn issues and problems concmung rlcctronlC
d,SOOUtSC. Pembcrton and Shun'llle (2000);5 I vaned collcclion of Irtldcs on Janguagt' use in (ern
putcr.bun! media. lt providts a good ovcrvicw ofthe vanous reseat(:h qutstions on, for uample.
languagc $lru(ture, hypcrtut links and Intrl1lction.
For Ihc 5ludy of Nid$ton (1999) 1$ an essrnlial publicallon Fot a mon: pm:tK:ally
orimlated book, $C\" Pncc and Pric:c (2002) On thc Inlrrntt. abo $C\" thc Jou,,"" o/CompultrMcd'
/lud Commullic4l114l1l. which Iw bttn onlinr .mu 1995.
4.J Besides Blhcr (t Seclion 4. 1), Sw.tlts (1990) 1$ now influentlal. Bhalla (1993) builds 41n thls
o(thought. Example ( 11) and (12) .nd Figure 3 .re taken from his work.
"
... Oiscourse typH
4.6 A good mtroducuon 10 multunodahty is KlUI and V:an Lewwm mo5t reanl
of ;ltl.oo on multtmodahty is provKkd by Charles:and KaJtenbacher (2004). For funheT
Sludy also (2002). Figure 4 is Imn from Bransford a.nd Johnson (1973: 394). FiguffS
S. 6 and 7 are taken from Ddm a.nd Batema.n (2002). In addition 10 the five Jevels menllOned in
lhlS Stion, the GeM model also mcludes thrff constramts. Whlch I. not dellt W1th bauw
of their spificuy.
5 Structured content
50' Propositions
The building blocks for discourse are (written) or (spoken) utterances.
Within these smallesl units of discourse il is possible to convey about the same con-
tent in a numlxr of different ways. A well-known example is the similarity between a
sentence in the active voicc and oe in the passive voice. The following sentences have
some content elements in common.
(I) This buteher seils only steak.
(2) Only steak is sold by Ihis butcher.
For certain typn of discourse analysis. it is convenient to disn:gard differenees In
formulation with approximately the "nle meaning. Diffen:neC'S in formulation are of
less importance when the focus is on the information itself and not on the discourse
situation. It is likewise convenienl to other aspects, e.g., the wriler's attitude
conccming the senlence. This aspect plays a role in the following examples.
(3) If only this butcher sold steak!
(4) This butcher only seils sleak?
$entence (3) expressn a wish while in (4) Icredulity or surprise is expressed.
The four sentences show imponanl diffen:nces. but they are also similar in a
number of ways. They all refer 10 a buteher and selling steak. This common element is
referred to as aproposition. The proposition can Ix described aS the meaning of a sim-
ple usertive sentence. The addition of the word simple makes it clear that a sentenee
ean conlain more than one proposition. Assertive signifies that it is irrelevant whether
the sentence is a question, a wish, an exclamation, eie. There are four propositions in
the following aclamation.
(5) Whal a pity that the poor boy can't cape wilh the hornble: truth!
I. his a pltythat x.
2. The boycan't cope with tht truth.
3. The boy is poor.
4. The truth is horrible.
"
5- COfltent
In a propositional analysis the situation in which the uttern! and Ihe
writer's or attitude as weil as Ihe forms in which occur are disregarded.
Thr concept proposition is taken from Ihe fields of philosophy and logic. There it
has a well -defined meaning. In discour$e studies proposition i5 use<! in a more general
sense, 10 denote Ihe minimal unil of meaning. What dOC5 such a uni! of meaning look
Ilke? Aproposition has averb, the' predicate, as its eore and oe or more arguments
thaI felate 10 it. Iklow is an example of a proposition. (63), whieh is the basis of a
scnlenee, (6).
(6) lohn finaUr bought a prescnt for mother.
a. 10 bur ((John)oubjKI(present)objt(mother)md.fKtobjtc.)
The predicate is the verb to buy. 1I is accampanie<! by thrtt arguments in a relation-
ship which is in (in this past
of verb to buy) and modal aspect are not taken into aeeounl.
ofthis method of notation is that it immediatdy becomes thai
following propositional
(7) For lohn boughl nothing.
(8) Could lohn bought anything for
A proposition eonsists of a and or more Bdow is an
of a propositional analysis of a
(9) Paper
I( you to begin 10 enumerate the various uses o( paper, you would find
list almos! without end. Yet, IheTe was a time when Ihis familiar iltm was
a pmious rarity, when the shttt of paper you now tass into lht wasttbasket
without thinking would betn purchased at a great priet and cartfully
p""'-
Thc propositional analysis is givcn below. The in propositions are not to
be equated with the words in the as ean be phrased in ways. It is
for this reason that propositions printed in small capitals. lhe numbers in the
propositions to the propositions which or follow. first proposition
given eontams condition for second; this eondltion funetions as a prroi-
Nott that third proposition is, through second proposition,
in first,

5.1 PTopositions 8g
(10) Proposltional analysis of(9)
J. (CONOITION,2)
2. (ENUMERATE, 3)
,.
(USES, PAPER)

(vulous.3)
5. (FIND, 6)
6. (WITHOUT END, 3)
7. (ALMOST, 6)
8. (6, CONTRAST. 9)
,.
(RARITY, PAPER)
10. (9, TIMB: PAST)
1J. (FAMIUAR. PAPER)
12. (PREClOUS, 11)
13. (PURCKASE, PAPU)
14. (13, PiUCE)
15. (14,OREAT)
16. (PRI!.SERVE, 15)
17. (16. MANNER: CARBFUUY)
18. (17. CONTRAST, 21)
19. ("ross. PAPER)
20. (19, PLACE: WASTUASk.f:T)
21. (19, TIME.! NOW)
22. (19, MANNER: WITIIOOT THINKING)
Al> can be setn in Ihis analysis, propositions do not eonsist of predieates and
with Usually,there art two elements. One
can be setn as subjcct, for example, paper in proposition 13J. other can be a
prerucative element. enumerate in 12J and purcha.sein 113J, but also a noun that fune-
tions as an object as in [31 uses in eonjunetion with paper. Such dtsignations as plaee
and time in 1201 and 1211 are also predieatts. Sometimes, aproposition has a logical
slructure as in (6J. In (181, a division can be
made. In Ihis analysis. compound propositions also oceur. for example. 1161 in which
115J - and thtough 1151, (14J and I13J - is also incorporated.
A propositional analysis can best be dcscribed as a hst of mmimal units
showing whleh ontl are directly relatro. relation of these units with propositlons
as a subject-predicate or a predicate-argument strueture somewhat It
should also be mcntioned that are hardly any eriteria which could be given to test
aecuraey of the analysis. In the analysis of the first sentenet, for example, proposi -
tions 131 and 161 eould be mentioned fil'lit because hold a position.
.. S. Structun!'d tOllten!
others 10 thtse two propositions.12] aod 13J could also combintd using Ihe
cnumuatc and Ihe argument usa 0/ papa. This uample is meant 10 illustrate
howa proposilional analysis works in practicc.
In discourse studies, the (OCllS is mainly on the relations between proposition,.
Take. for examplc. Ihe first part ofthe SKond sentence ((81-1121): Yet, there was a
time when Ihis familiar item was a rarity: Thc relation bclwn these propo-
sitions can bc illu$lrated in a diagram as shown in Figure I below, whrre 112) is Ihe
most embedded proposition.
This of analysis i$ important for measuring Ihe difficuJty of discQursc or for
gaining in5ighl into the proctsS of discoursc underslandmg. Examplcs of hypothesn
for which Ihis analysis 15 nu6Sary are: Thc more embeddcd propositions Ihere are,
the more diRicult the discourse will be. Aproposition al the lower level 4 is not as easy
to remember as one at level 3. For a given group of readers, there ean be no more than
thrtt propositions at levd 2.
So much for propositions, whkh can be linked endlessly 10 build a discourse. In
discourse analysis il is useful 10 eonsider a level between the microlevel of proposilions
and the macrolevel ofthe discourse as a whole: a mesolevel ofiopics, whieh encompass
series of propositions that are linked together.
p. Topies
A topie or a theme i5 what a discourse, a discourse fragment or a sentenee is about. It
is the shorlest summary of a discourse,lhe main proposition of a paragraph or whal is
eommented on in a sentence. The term topic is usually defined as the of a
unil of discourse. The vagueness of the definition maus awareness ofthe distinelions
important.
A distinetion has 10 be made bctwttn a discourse lopie and a senlence topk, i.e., the
topk deall with in a discourse or a senlenee, respectivdy. Beloware two examples.
9 level I

8
"
J
I
12 bd4
Figure I. Thf rflalion bel ... n proposition,
5.1 Topies
(11) The NfW York Yankees won.
(12) A: Did you see the Yankf'eS'Sox gamf yesterday?
B: Yuh, who would haVf thought thai thf YanUfS would win!
When normal inlonation is used, sentenee (11) is about the New York Yankees; there-
fore, the New York Yankees is the senlenee lopie. What is staled about the topie is
ealled the eomment. In fragment (12) the Yankees-Sox game is the topie of eonversa-
tion, and therefore the discourse lopie, wilh 8'5 remark serving as the eommen!.
The concepts topic and eamment often lead 10 confusion as the distinclion belween
these and relaled eonctpts remains undear. Firsl, among thosc other concepts is the
concept set Iheme-rheme. A theme i5 what is "undtr discussion
M
in a given situation;
often il isthe subjeel of a sentenct. The rheme is whal is said aboutlhe themt; usually
Ihis is the predicate of a sentence. In (11) the themt-rheme dislinction runs parallel
to the topiecomment division, bul this is not nessarily always true. As the Iheme-
rheme coneepts are more or less synonymous with the subject and the predieate. these
terms are used le55 frequenlly. A sentenee topic is nol necessarily the subjcct of the
sentenee; s A's in (12) whtre tht topic IS the objt.
Second, there is the coneepl set (5 also Seclion 7.2). As tht topic is
wha! is dealt with in Ihe and is, therefore, usually known, lopir and giYen are
often used inlerchangeably. Yel, there is a distinction. Below is an example; pay careful
attention 10 B's utterance.
(13) (A. B, and C are parlicipanlS in a mtfling)
A: Shall wt dlscuss the minutts
B: I didn'l receive a copy.
e: Mine is unreadable.
In 8 '5 ulterance, "I" i5 topie and the comment is that a copy has nol been received.
In the commenl there is, however, a word which owing 10 the question about IV;
is already given: the word lhe new element in Ihe comment is that
It has not been received.
Third, there is the concept set foregraund-background mformation (see also 5.
tion 7.2). Since the topic is whalthe senlence i5 aboul, it usually does nol contain the
mosl important informalion in a sentence. Often the topie i5 more in the background
But this is nol always Ihe case, as can b<! seen in the foUowing example. In B's ulteranee
the tlemtnt about tht neighbor can be sn as the lopi c, e\'en though this information
is in the foreground.
(14) A: I had enff I1 Mary's yesttrday.
B: Say. did you hear that her nelghbor wanlS 10 get a divorce?
.'
"
5. Stf\ICtun!'d conlen!
Thc conCepllopi( lhus deals with 50mnhing which is discusscd in a $entence or dis.
course. And thai an be defined as background, foreground,
given, new,eIC. Below is oe more example 10 clarify Ihis somewhat elusive dislinclion.
(15) A:
B: YH, Ihere bas bttn another flood in Bangladesh.
lhe discoursc topic here is "news": In 8's utlerance the sentence topic is that Ihere is
news, and the commen! is that there has bun a flood in Bangladesh. 111e theme is
"floOO" aod Ihe rherne thaI it look plaee in Bangladesh. IMpending on Ihe intonation,
"f100d" or "Bangladesh" ean bc given or ('W. Thc word "another" is also more fore.
grounded than "flood" or "Bangladesh":
Although Ihere are 00 unequivocal criteria for delermining thc topic of a sentence,
some tendencits ean be given. A topic is: I. more likely 10 be definite !han indefinite;
2. sooner pronoun !han noun, 3. sooner subJtet than object. In the following example
all thrte tendendes ean steno
(16) lhe blonde wornan saw a man eross Ihe street. She lrnrnediately slarted walking
futer.
Btcaust woman" is ddi.nite and in the subjec! position, it ean $OOner claim lhe
topie status. The topicality is strengthened by the pronoun in the stntenee whieh
folIows. That Ihtsc are only lendendes is proven by the following uamplc.
(17) The blondt woman uw a man eross the street. The man looktd scared.
In this case. appears likely to have topk status. iso however. indefinite
in the first stntenee and does not return as a pronoun. Ihere also appears to bt a
tendeney in the order: firsl the topie and then the eomment. But this is only a slighl
tendeney.
For the analysis of the stnlenee topie, eertain lendendes ean be indicated. For the
analysis ofthe discourse topk, only intuitions apply. 11 is usually possible to come tO a
consensus a5 to what the tOpie of a given discourse fragment is. It is more diffieuit 10
deterrnine where a subtopie begins or if there is, in facl. lopie eontinuity. topk shift or
topk digression. In example (14), for instanee, topk shift lakes plaee. This shift resuits
in a subtopie jfthe eonversalion eventually returns to drinking coffte al Marys. It has
proven quite diflicult to generale adequate erileria for topie shifts.
Wilhin a (sub)topie, topie digression ean take place if a sidetraek is taken. An
example ofthis would be ifthe eonversation in (14) were to turn to the special way in
which the eoffee was made al Mary's. However, as in lhe ease of topie shifts. it is dif.
fieuit 10 formulate eriteria for digressions. The same is also Irue for lopie eontinuity.
Look I1 the following examples.
(18) I. The Prime Mmisler sleppe<! offthc plane. Joumalists immediately
surroundtd him.
Sl Topics 93
b. The Prime Minister steppe<! off the plane. He was immediately surrounded
by journalisis.
In (18b) there is topieeonlinuity. "The Prime Minister" remains Ihesubject in the
lowing sentenee in the form of a pronoun. In (l8a) there would appear to be a lopie
shift. as Ihe following sentenee starts with another subjtct. But il would depend on how
the discourse wenl after these two sentenecs.
Intuitions about sublopics and topk shifts have provtn to be quile intersubjeetive.
This was shown in an experimtnt by the Brilish linguist, Eugene Winter (1976). He
rearranged the sentences in a text and had students atlempt tO pul Ihem baek in the
correct order. that iso arrange thcm so that they linked up topkally. Winter also asked
the sludenlS 10 mm the tnds of paragraphs, thai is, mark where subtopics com
meneW. This experiment is repeatw in !ht last assignmenl of this ehapter.
Aside from this consensus of intuitions, another pattern can be pointed out in
discourse. Jt has to do with the relationship between the degree of topicality and the
amount of"language material": Talmy GivOn (1989) has caJlw Ihis reJationship the
code qUQl1tity pril1ciple. Aceording to this prindple. a topie is defined as that whieh is
predktable or accessible.
(19) GivOn's code quantity prineiple
The 1CS5 predielable or aeeessible a referent iso the mort phonological material
will bc used to code il.
Compare the following uamplcs.
(20) a. He walchtd how the gas sial ion attendant hooked up Ihe hose.
b. lhe man watched how the gas station attendant hooked up the hose.
e. The man behind Ihe wheel watched how the gas station atltndant hook.ed
up the hose.
If a "man behind the wheeJ" and a "gas station attendant" are bolh eharacters in a story.
and the first man possesses topie status through then the topie status is lowcrcd
as more phonological material is used. The code quantity prindple would appear to
provide a good basis for topie analysis.
..
S. Struclured content
5.3 Macroslructures
Readers are generatly ahle to give a summary of the topics they have juS! fead. Belo ....
is an example of a short story and Iwo possible summaries.
(2 1) Pete decided \0 80 on a skiing vlcalion thai rear. Up unti] thcn he had only
gone hiking in the mounlains in the summerlirne, but he had declded thaI he
wanted to ltarn bow 10 ski and lhe winter mounlain alT mighl be bendic.ial 10
his health. He wenl to a lravd agency 10 get information SO that he could choosc
a dtstination. Utah Smed the most attraclivc. Once he had milde his choiee he
wen! Inck 10 the Irave! agtncy 10 book lhe flight and reserve a room at a hotd
thaI he had found in one of the folders. Naturally, he also nme<! .skis, poles.
and boots, but slnee he did nOI have the money 10 OOy them, he declded to rent
tMm when he got there. In order 10 avoid the SCOlSOnai rush, he decided to go
after the New YeaT. \Alhen tht big day 6naUy arrived, he was taktn to Ihe a.Lrport
by his fath so that he would not have to deal with his luggage on his own. He
took the night flighL He was aetually abl.e to sJeep on the plane. The foUowlO8
morning Pele arrived, weU rested. at his destination. It was snowmg. lhe hotd
was right nut to the ski resort. The view ofthe rnountains was btaulifuLl1e
immediatdy feit right at horne.
a. Pete wanted 10 go skiins in Utah Ihat wiler. He made the nects.sary
arrangements. He went by plane. He liked Ihe hOld in Ihe rnountains.
b. Pete went skii ng. He really liked it.
How do readers manage to arrive at these types of summaries? This can only be
explained by assuming that a discourse has a structure of meaning that make5 dear
what does and what does not belong to the core of the content, or the gist of the
discourse. Teun van Dijk (1980) introdueed the term maerostrueture 10 denote this
structure of meaning. This term is the opposite of microstrueture. The term micro-
struetul't denotes the relations betwn sentences and sentenee segments. these ean
be represented with the help of propositions. see Section 5. 1.
A macroSlructurt is the global meaning of discourse. Thus, the maerostructurt
(2Ia) or (2Ib) can be attributed to text (21). 8elow i5 an explanation ofhow mae-
rostructum are formed using thr macrorules.
51 Macrostructures 95
a. Dtll'tiol1 rule
This rule eliminates those propositions that are not relevant for the interpretation of
other propositions in di scourse. Take the following example, which contains three
propositions.
(22) A Sirl in a yellow dress passed by.
I. A girl passed by.
2. She was wearing a dress.
3. The dress was yellow.
By using the deletion rule, propositions 121 and 131 can be eliminated, leaving only 111
as aproposition. lhe deletion rul e is a n ~ t i v e formulation: eliminate irrelevant prop-
ositions. \Alhen formulated POSltlvely, it is a selection rule: select those propositions
that are necessary for the interpretation of other propositions. The deletion role ean
be split into a weak and a strong vanant. lhe weak deletion rule eliminates irrelevant
propositions; the strong deletion rule only eliminates propositions that are relevant at
thr mierolevel but not al the macrolevel. Bdow is an example.
(23) lohn is siek. Hr wdl not be going 10 the meeting.
At the microlevel the proposition "lohn is siek" is relevant for the interpretation of the
senlence which folIows. lf, however, the text does not continue with the therne ofJohn's
iIlness, then this proposition is irrelevant at the maeroleveL
b. Gel1eraliztlliol1 rule
Using this rule aseries of specific propositions art converted into a more general
proposition. Here is an example.
(24) Mary was drawing a picture. Sally was skipping rope and Daniel was building
something with Ugo blocks.
I. The children were playing.
This rule does not just eliminate irrelevant details. Rathet, specific predicatC'S and
argumrnts in aseries of propositlons are replaeed by more general terms so that one
proposition may suffiee.
c. Col1slruellol1 rule
By means ofthis rule OM proposition an be constructed from a numberof propositions.
See the foUowing example and the macroproposlliol1 that was eonslrueted from iL
..
5- ronltnt
(25) lohn wen! 10 tht' station. He boughl a ticket. starttd runmng when he saw whl!
IIme it was, and was forccd 10 conclude that his walen was wrong when he
reached Ihe platform.
I. John missed the train.
The difference this role and the generalization rule is thai the propositions
on the basis of which a general proposition can bt: conSlructed do nOI all have 10 be
contained in discourse. In (25) neither nor are mentioned. Yet, on the
basis of general knowledge, it is possible 10 construct aproposition from Ihis incorn-
plete description.
How do these macrorules 'NOrk in determining the global meaning structure of
discourse? Below are a lexl fragment and a simplified version of a short example of
macroanalysis,
(26) I.
2.
3.

5.
6.
7.
8.
9.
A lall stirn blondt in a white summer frock wallongjust ahnd ofhim caught
Ken Holland's cyc.
He studie<! her, watching her gende undulations as she walked.
He quiddy shifted his q'ts.
He hadn't looktd at a woman like this sill he had tirst met Ann .
What's the maUer with Me? he asked himself.
I'm getting as bad as Parker,
He looked again at the blonde.
An evening out Wlth her, he thought, wou!d bt: sensational
What the eye doesn't see, Parker was always sayiog. the heart doesn't grieve
abouL
10. That was true.
11. Annwouldneverknow.
12. After aU. other married men did it.
13. Why shouldn', he?
14. But when the girl crossed the road and he lost light ofhtt. he jerked his mine!
back with an dron 10 the Irtttt he had received that moming (rom Amt
15. She had been away now ror live weeks, and she wrote to say that her mother
was no beuer. and she had no idea when she was coming back.
The deletion rule aod the generaliution rule app!y to 111. The mrormation about
dothmg can be eliminaled. The description or the blonde can be generalized to an
altractive woman': The message about iso al the microlevel. relevant ror Ihe
interpretation or 121. and can, thererore. nOI be eliminaled according '0 the .....eak vari-
ant or the deletion rule. It iso hOlO."Cver. possible 10 eliminate waIk" and undulations"
by applying the strong deletion rule and the generalization "an atlraCllve woman': The
S" Superstructurn 97
way in which the woman walks is or secondary importance. The rollowing discourse
elemenls can be generalized to "Iooking at": "caught Ken Holland's eye" 111, "studied"
and "watehing" (21, "shifted his eyes" (3). "Iooked" r 41 . and "Iooked again" [71. On the
basis or 13J. (41, 15J and 161. it can bt deduced that Ken Holland feels guilty because
he, a married man. wanLS to dale another woman. Scntences 191through 1131. which
provide. as it were, the argumentation. can bt generalized into "There is no reason not
to go out with another woman," The generalizalion can possibly be e\iminated when
it has become dear that the argumentation is irrelevant foe the rest or the story. Sen-
tences 1141 through 115) are linked to (1) and 12) and provide inrormation about the
main character. On the basis ofthis information, it can be construed that Ken Holland
is unhappy. This information is nol in the texl, but can bt deduced using presupposed
knowledg(: ofthe married man's psyche. After this analysis, van Dijk arrives al the fol-
lowing macroslructure.
(27) 1. Ken Holland is looking al a btautiful gIrl In th(: Sll"t (from )1], 12]. )7] and
181 by generali7.ation).
2. He has a gUllty consdenet aboot thai because he is mamed (from 131.141.
151 and 161 by eonstruction).
3. He is fruslrated hecause his wife is absent (from 1141.( 151 by construCllOn).
Clearly. Ihis is not the only possible macrostructure. The texl about Ken Holland can
also be summarized as follows.
(28) A man shonchanged two warnen.
Macrorules are not rules thai can be used in order 10 lrace the meaning structure of
discourse. The rules only describe the procedures with which a meaning slructure can
be assigned.
S.4 Superstructures
In many cases discouT$e conlains nOI only a meamng strUClure, but also a kmd of
prefab structure 10 present a structured contenI: a supentrue/ure. A good example
is a letter of application. This type or letter usually has a specilie form: an mlroduc-
Hon to Ihe application. which is rollowed by an argumentative segment or sales pitch
and, in conclusion, ptrhaps a reference 10 the curriculum vitae or references. Within
such a discourse schema, the content ean vary. For this formal Slructure, Van Dijk
introduced the term supentrueture. Superstructures arc convenlionalized schemas
that provide the global form for the macrostructural content of a dlscourse. In other
words. macrostructures deal with the contenl and superstructures with the form. The
.. 5. Structured contenl
OOsnvation (I)
I
theory(2)
I
hYJlOlh:tsu (3)

laun (4) condusiont(l)
apmmcnls{S) oulC(Jm!:(l2)

/Z cnculions(9)
ruuIu(IO) dlJClWions (11)
subjKls (7) ronditiotu (8)
Figurc' 2. Supentructure o( a SCientific artkk reporting on uperimenta.l ftsC'arch
term supentrue/ure also illustrates Ihe fact thai the discoursc form stands above Ihe
content in some sense. When a Inter of application is lKing wriuen. an txisting dis-
course form can bc used with a specific contenl. Thc of Ihe leiter can Ihen
easily determine where 10 find specific information.
Superstruclures are also used for other types of discourse. Thc superstruclure
of a scientific artide in whien experimental is reported on could look like
Figure 2 abovc.
As an illustration, a fake investigation 15 described thai has Ihe concept super-
sfructure as ils lopie. Newspaper arlicles often have a slruClure that an be represenled
approximately like this: First. there is a kadlil1e (sometimes accompanied by a subtitle),
which seITes as a short summary. Following this is a lead: bold print containing the
basic gist of the news item. lhen comes the news article itself, the f/at text, which is
the detailed report of the news item. This iso of course, not a complete description.
Furthermore, these rules do not apply 10 editorials. For simplicity's sake, however, it
is assumed that the following observation (I) is eorreet; news bulletins have a super
structure with a heading, a lead, and a f/at text.
An interesting question is to what degree the superstructure influences the
assimilation of the lext Or, to phrase the question in a more limited fashion: Is more
or less information acquired from the flal text when there ;5 no lead? On the basis of
a theory (2) on text eomprehensibility, il ean be assumed that a reader will gain nlOre
information from a texl if il is den beforehand what the text is about. 1he following
hypothesis (3) can be derived from this theory: More information can be deduced
from news articles that have a lead !han (rom those that do not. Thls hypothesis can
be tested (4) experimentally (5). For example. readers are given news articles wilh or
54 Superstruclufn 99
withoutleads and are asked queslions about the text after they have read it in order to
determine the of mformation assimilation. The set-up oflhe experimenl (6),
information on the subJI.'Cls (7), and the eonditions (8) will have 10 be induded in the
report. If, for exampll.'. il were 10 become dear Ihal some subjects possessed a good
deal of prior knowledge concerning the topic of the news artic1e, the results would be
less reJiable. It must also be reported how the experimenl was performed (9), whether
the subjecls were given equal amounts oflime, etc. The results (10), in this case the
differences belween the answers 10 the quest ions, are discu.ssed (11) with the outcome
(J 2) leading 10 a condusion (13). in which it is staled whether or not the hypothesis
has been confinned. The components mentioned in the above schema do not have to
occur in exactly the order described here. This superslructure or a varialion on il can,
however, be found in many research reports.
One queslion thaI has frequently been invesligated is whether a sludytal is easier
10 learn if the text itself provides c1ues about the macrostrueture or superstructure.
These dues are called aavance organizers. An organizer can be a lilie or a subtille
that indicales the content, bul it can also be an introductory paragraph in which the
siructure ofthe text is t:xplained. Ac1ually, every text fragment that describes the texl
thaI follows i$ an advanct organi:ter. Numerous experimt:nts have made it cltar thaI
in certain circumslances advance organizers can aid the learning proass, for instance,
when a sludenl has very littlt: prior knowledgt: of the topic. For this reason, many
texlbooks contain introduetory sections that t:xplam the contt:nt (macrostrueture) and
construction (superstrucIUrt) ofthe text.
~ :r

g-
H
H

...
n
n
,n
H
-~
if ~
H
s
15" :;
, =
&. ~
n [

, .
,
u
.' ,
"
2 ~
U
"
7


1


..
' 01 5. Structured content
,.,
Bibliographical informati on
The aamples ofthe propositional analysis of discourse fragments wen tmn from the dissertation
by Baten {l981). For additional uamples, sn Kintsch (1998).
Po The term top;, is. accordI8 to Brown and YuIc (198J), "the frfijuently uscd uncxplalncd term
in the of disoourse." Consult this soun:e for furthcr information. Example (28) on topic
continuity W3$ also takt:n from this votumc.
"
,.,
The 'hemc-rheme concept set originated in the Prague school. This Khool, which f10unshcd
primarily in the 19JOs, indudcd a number oflinguists whosc maln interesI was the communica-
funelion of word order differenees. Str, for uample, Firbas (1992). Mueh ofthe lilerature on
thematlGS is scalleTl'd GYer dIfferent disdplines (hterature, psychology, linguistics and cognilive
stienee). Louwerse and Van der Pur (2002) have cdilcd a dClailcd eotltclion 10 provide a eoheTtnt
ovcrvicw.
For further study, sn Van Dijk (1980). Thc examples hen: wcre for the most part tmn from
this publication. lhe Kf,n Holland passage is an uccrpt from TIgu Iry ,he T.1I1 by lama Hadley
Ch1SC (1966).
For funhcr study, litt Van Dijk (J980). Thc term udV<lmc 0'i"niur W3$ introduccd by Ausubd
(1960). An OVl'rvicw ofthc rescarch is givm in Marer (1979). For funhcr 5ludy, Stt lonasscn (1982)
in which sU: fundions of adVi\Il(C orgalli=-s arc defiJ>Cd S also PoIanyi and Van den Berg (1996),
who dcmonstratc a framcwork for hdping to idcntify and resolvc compla issucs in thc 51ruclurc
and interpretation of discounc.
6 .
6 Discourse connections
Cohesion
The most salient phenomenon of discourse is the fact that sentences or utterances
are linked logether. For this this two concepts are used:
cohesion, referring to the connections which have their manifestation in the discourse
itself, and coherellce, referring 10 the connections which ean be made by the reader or
listener based on knowledgc outside the discourse. In a sentence like "Mary goI preg-
nant and she the fact that she refers to Mary is an example of cohesion, and
the interpretation that her prcgnancy was the reason for her to marry is an example
of coherence (see Section 6.3).
This first seetion deals with the connections evident in the discourse, with cohe-
sion. Michael Halliday and Ruquaiya Hassan ( 1976), who were already introdueed in
Section 3.6, were the first to analyze this kind of discourse conntion. They distin-
guished five types of cohesion.
a. Substitution
Substitution is the replacement of a word(group) or sentence segment by a -dummy"
word. The reader or listener can fill in the correct element based on the preceding.
Threc frequently occurring types of subst itution are that of a noun (1), of a verb (2)
and of a clause (3).
(I) These biscuits are stale. Get some fresh Olles.
(2) A, Have you called the
B, I haven't done il )'ct, but I will do it.
A: Though actually, I think you should do it.
(3) A: Are they still arguing in
B, No, il just seems so.
b. El/ipsis
Ellipsis is the omission of a word or part of a sentenCe. Ellipsis is closely related to sub-
stitution, and can be described as "substitution by zero". The division that is normally
used is nominal, verbal and clausal ellipsis.
'04
6. Oiscoorse connKtions
(4) bL$CUJIS are siale. are fush.
(5) He partieipate<! in the dd>alt, but you didn'l.
(6) Who want' 10 go shoppingr You?
c. Rtf erence
Reference concerns thc relation a discourse element and a preceding or (01
lowing element. Reference deals with a semanlic relationshIp whereas substitution and
ellipsis deal with the relat ionship between grammalical uni ls: words, sentenee parts
and clauSC$.ln thc ca$(' of referencc, thc meaning of a dummy word can be determined
hy what is imparted bcforc or after the occurrcnce ofthe dummy word. In general. Ihe
dummy word is a pronoun.
(7) I see lohn is here. He hasn'! change<! 1I bit.
(8) She certainly has changed. No. bebind lohn. I mtan Karin.
Bul reference can also be achieved by other meaos, for inSlance, by the use of adefinite
or an as in the followi ng examples:
(9) A man eromd the strttt Nobody saw whal happcned. Suddenly the man was
l)'ing Ihere and ",lling for help.
( 10) We grew up in Ihe 1960s. We were idealislie fhen.
d. Conjunction
ConJunction is Ihe rdationship whieh indicates how the sentenee or c1ause
should be linked 10 the preceding or the followmg (parts of the) senlence. This is usu-
ally aehieve<! by the usc of conjunclions (also known as eonneclives). The following are
examples of thrtt frequently oceurring relalionshlp5; addition, temporality, eausalilY.
The rclationship can be hypotaclic (as in the a-examples, which combioe a main c1ausc
wi th a subord.oate c1ause or phrase) or paratactie (as in the b-examples, whieh have
two mai c1auses).
addition
( 11) a. Besldes being mun, he is also hateful.
b. He no Ion ger goes 10 school and is planning to look for a job.
lemporality
( 12) a. After the car had been repaired, .... e were able 10 conlmue our JOurney.
b. The ear was repalred. Afterward$ we were able 10 conlinue our journey.
causalily
(13) a. He is not gomg 10 school tOlhy bwwu he is sidt.
b. Ann gol a beauliful Job last yeararrd now SM is rieh.
6.1 Conesion lOS
e. Laica/ cohnion
Lexical cohesion refers to Ihe links betwttn Ihe content words (nouns. verbs. adjec: -
tives, adverbs) which are use<! in subsequent segmentso( discourse. Two type5of lexi-
eal eohesion can be distinguished: reiteration and colloca' ion.
Reiteration includes not only repetition but also synonymy. Reiteration can also
occur Ihrough Ihe uscof a word that is systematica1ly linke<! to a previous one, for exam-
pie. young and old. In rciteration is divided into Ihe five following Iypes.
J. (often involvmg refercnce)
(14) A eon/eunce will be held on national environmenlal policy. At this confrrence
the !ssue of sa!inat ion will play an imporlant roll'.
2. synonymy (often involving referencel
( I S) A eon/erence will bt held on natK>nal environmenlal policy. This envlrDnmtntal
sympa$ium will be primarily a canference dtahng with water.
3. hyponymy/hyperonymy (e.g., the rdation of flower 10 tulip and vice versa. subordi -
nalion and
(16) We were in town tooay shopping for furmture. We saw a lovely table.
( 17) Did you see the wooden igloos in this new lown? Oh, they build cven stranger
hou$l'S heu.
4. meronymy (part V$. whole)
(18) At its six-month eheekup, the "rakes had to bt repaired. In general, however, the
car was in good condition.
S. antonymy (e.g., white V$. black)
(19) The old movit'S jusl don', do it any more. The new ones are more appcahng.
Collocation, the Sond type of lexieaJ cohesion, deals with the rdattonship belween
words on the basis of Ihe fael that often occur 10 the same surroundings. Some
examples are sheep and woo/, congrt!$S and palilicilln or college and study.
(20) Red Cros.s helicopters .... ere in the air eontinuously. The blood bank will soon be
in ne! of donON.
(21) The hedgehog scumed accossthe road. Its speed surprised me.
In the five main types of cohesion (substitution, ellipsis. refercnee, eonjunetion and
lexical cohesion), the interpretation of a discourse demenl IS on a"other
element Ihat can be polOted out in discourse. In (2 1), for instance, the eorreet
....
6. DiKOOrse conllKtions
interpretation of the word is oolr possible by feading thc preccling senlen
within which the word is of primary importance.
6.:1 ReferentiaJ elements
A special type of rtfetenlial cohesion ruults from the use of proous.
(22) John said thaI he was not going 10 school.
(23) When he camt' in John tripped aver the blocks.
Back-referential proous, such as the proou in (22), are called anaphora. The term
is derived from a Creek word which means -10 lift up or -'0 bring back": Forward-
referential pronouns, such as the one in (23), are called calaphora: cata- is the opposite
of una-.ln the examples mentione<! here. Mht
M
can also refer 10 anolher person. Ihen
it is caJled an uophor or a deictle element.
Anaphoric relations are nol only found when personal pronouns are used. See the
proverb in the following example.
(24) Iflohn is nOI going loschool, Ihen I won't do it eilher.
lhe research into anaphol'2 is focused on the following quest ion: How are anaphol'2
and whieh (actors playa role in the interpretation process? Compare tne
foUowing fl'2gments.
(25) Mary said nOlhing to Sally. She would not understand the finl thing aboul it.
(26) Mary told SaUy everything. She could not keep her mouth shut.
In (25) ean only refer to "'SaUy": In (26) both referenees are grammatieally pos-
sible. While in (27). "lhe" can only refer to "Sally":
(27) Mary told Sally everything. She eould nOI kttp her moulh shut and Mary really
told her off for doing it.
An interesting phenomenon can be observed in the following sentences.
(28) 'uhus leh. He was siek.
(29) He was siek. ,ulius It"n.
(30) He was siek. lhat's why 'uhus len.
In (28) ean refer 10 Juhus. In (29) il i5 mueh more plausiblt thai "nt" re(ers 10
someone olher Ihan JuHus while, in (30) "he" ean be interpretcd as referring forward
to "Julius": These differences ean be explained by assuming an interpretation principle
suggtsled by Peler Boseh (1983).
6.2 eltmenu 107
(31) Prineiple of natural sequential aboulness
Unless Ihere IS $Omc reason to assurne the eontrary, eaeh following sentt"nce is
assumed 10 say something about objt"Cls imroduet"d in previous sentt"nct"S.
On Ihe basis of this principlt", aeeording 10 Boseh (1983), the in (29) cannot
be interpreted as JuHus. The (act of !uHus leaving says nothing about the prt"eeding
sentenee: "Ht was In (30), on the other hand, the word "that" indieatts that
something is going to be said whieh is linkt"d to the preuding sentence. Tbis indlca
tion is reinforced by the reader's knowledge that oe consequenee of "being siek" is
(ound in the words whieh follow, thai is, that sickness ean be a realon for leavi ng. lt
is for this reason that Ihe sentenee about 'ulius ean be linked to the preet"ding sen-
tenee. Tbis interpretation is, therefore, vt"ry mueh dependent on the reader'$ general
knowltdgt. Tbis ean also be seen in the following example, 10 which the rdatlon is
tht same as in (30).
(32) Ht $Creamed. Thai is why Julius Itft.
Al> someone's sertaming is not usually a reason for thai same person's leaving, il ean
be assumed on tht basis of the interpretation principle Ihatthe suond does
nol say anything about the person In tht first sentence. ThU$, the "he" in (32) cannot
be interpreted as referring to
Exptrimenlal has dl'termined which (aClors playa role in the interpreta-
tion o( anaphol'2. In an experiment eondueted by Susan Ehrlich (1980), subjuts wt"re
given senlenees o( tht following type.
(33) Steve blamed Frank because ht spille<! the eoff.
(34) Jane blamed Bill bause he spille<! the coff.
Tbt timt ittook for tht subjectsto dettrmint" whien name was tht anttcedent for Ihe
anaphor "he" was measured. Most o(the subJecls dettrmint"d that "he" in senlenee (33)
referred 10 Frank.. Tbis duision did not requirt grammatical knowledge but gtneral
knowledgt. Spilling coffee is dumsy and inconvenienl and is, therefore, a reason for
blame. I( Steve is blaming Frank. !htn it is most likely the lalter who spilltd the eof-
f. Tbt US(' of general knowlt"dge is a pragmalic faetor. ln (34) this is not
ntce5Sary for the interpretation of-he": Knowtt"dge of grammar makt"S It dear that
being a male-gender pronoun, can only refer to Bill.
If pragmalic faetors always playa roll' in the interprttation o( anaphora, then the
subjtcts would have spent equal amounts oftime in delermining the for
both sentenet (33) and (34). 1(, however, readers firsl apply Iheir grammatieal knowl -
edge and only then their gentral knowledgt, if nccwary, Ihen the inltrpretalion of
(33) will take less timt than that of (34). After all, in the case of (34) grammatical
.oB
6.,
6. conJlKlions
is sufficicnt. Thc experiment did indecd provc thaI thc interpretation of
(34) look less time than thaI of (33). This 1110 the conclusion thaI pragmatic ractors
oolr playa role when grammatical ducs are lacking.
Coherener
Ir proposilions are thc building blocks of discourse, then discoursc relations ace the
eement between thc blocks. Below are 50me examples from thc many different kinds of
discourse ulations thaI exist. We are looking 31 thc I"C'lalion betWN'n thr two sentences
in each discourse fragment
(35) a. The government has laken emergency measures. will becomc effeclivc
next reaf.
(36) a. Thc president Will probably run for reelection neXI reM. This was
announccd yeslcrday by thc \Vhlte House press sccretary.
(J7) a. The pmidenl wu not available for commenl. At that particular moment hc
was recciving his Chinese counterpart.
In (35a) thc follow-up sentence elaborates on onc constituent, In (36a) the
second sentence encapsulates thc first sentence. In (37b) the situation is diffeunl: thc
followup provides an explanation for the content of thc firsl senlen.
ThC' relations in (35a) and (36a) add very littlC' 10 the muning oflhe sentences.
In (37a), howcvcr, a meaning dement ls added. Ihis can bc scen ifthc scntcncC's are
rewriHen as one single scntcncc. Cnly in (37b) will a meaning-Iaden conjunction
be nccessary.
(35) b. The govcrnment ha5 taken measures which will become dftetivc
nat year.
(36) b. The While Hause press secrelary announccd yesterday thai the president
will run for reeltelion nexi year.
(37) b. Thc presidC'nl was not available for commC'nt 1115 hc was TtcC'iving hiS
Chinese' counlC'rpart at that particular momcnt.
Research into discourse relations has conccnlrattd on Ihose links betweC'n sC'ntC'nces
which hur meaning. This i5 not lhC' case in examples (35a) and (36a). This discoursc
does not contain a mcaningful link between the main sentC'nces and the adjtetival
subordinalC' clause (35a) and the objtet complC'mC'nt (36a). Thc link in example (37a),
howevcr, does have its own mC'aning: reason.
In the research done into (meaning bearing) discourse relations, two basic types
are distinguishtd: Ihe additivC' relalion and the causal relation. The additivC' relation
6.3 Coherence log
can be tractd back to a conJunction and 1115 such is relattd to variou$ typcs of coordma-
lion. Among the coordinating relations are Ihose which can be rC'presenlcd by words
such as a,.d (conjunclion or addition), hut (contrasi), or (disjunction), or an equivalenl
of these words. Below is an example of a contrast relation.
(38) John bought a presenl for his mother. (But) he forgot 10 takt it W1th hirn.
A causal relation can be traccd back to an implication, and is as such relaled 10 sub-
ordination. The most importanl causal relalions are Ihe sevcn types distinguished in
traditional grammar:
(39) cause
John did not go to schaoL He was sick.
(40) reason
John did not come with us. He haIes parties.
(41) muns
Would )'Ou mind opening lhc Here is Ihe kty.
(42) consequence
lohn is siek. He is not going to schooL
(43) purpose
The inSlructions should be printcd in capitallC'lIers. 11 is hoped thai in Ihis way,
difficulties in reading them wiU be avoidcd.
(44) condition
You can get a job thlS summer. Bul first rou have 10 pus )'Our exams.
(45)
HC' was rich. Yet he nevC'r gavC' lOything to charilY.
These discoursC' relations can be distinguished as folIows. A cause indicates a con-
sequence that is outside the domain of volition. A reason always indicates that a
volitional aspect is present. A mcans is a deliberatc utilization of a cause in order 10
achieve a \'olitional consequence. A purpose is a volitional consequence. A condition
is a necessary or possible cause or reason for a possible conscquence. A concession is
a cause or reason for which the txpected consequence fails 10 QCcur, or the yieiding
of a point.
Discoutse relations can IM> groupe<! or classified according 10 specific character-
islics which lhey share. ane of these charactcristlcs is the k'manllcpragmalic dimen-
sion. The hterature includes various definitions ofthese terms. The following are fairly
common. Scmantic relations conntet segments on the basis of Iheir propositional
".
6. Discourw conntions
content. the loculions of the segment, linking the situations that are referred 10 in the
propositions. Pragmatic rdations connec:! .segments on the basis oflheiT iIIocutions.
A good example of a semanti c relation is (40). A hearer can interpret lohns haliog
parties 3S a reason, without having 10 deal with the illocutions of Ihe segments. It is
the two situations in the consecutivt sentencH that are related: the situation
in the last scnlenee is a reason for the situation "not (oming along
R
In Ihe first
scnlenee. An uample of a pragmatic relation can bc seen in the following sequence.
(46) J'1l getlhe grocenes. l have 10 go shopping anrway.
In this example, the rdalion does not pertain 10 Ihe two situations in both $Cnlences,
but 10 the illocutions. After all, "going in the last scntence is not necessar-
ilya reason for "getting the groceries" in the first sentence as far as its propositional
content is concerned. If this were the casc. then anyont: who was to go shopping
would also get the groceries.
It is, however, sometimes difficult to draw a precisc boundary between the seman-
tie and pragmatie relations. For examplt:. is the relalion in (45) scmantic or pragmatic?
The relation is a scmantie one in a wotld wheTe it is \Try unconventional for somwne
who is rich not to make donationsto charity. But the relation is pragmatic when the
speaker has the apparent intention of making an accusation.
A special subset of pragmatic relations is rhetorieal relations. These are the rela
tions with whieh speaken or writers apparently have the intention ofbringing about
.11 change m opinion. position or bchavior of readers or Iisteners. Usually the five fol
lowing rhetorieal relations are distinguished.
(47) evidence
No single measure has had an effect. The traffic j.1lms are still.1l$ bad as ever.
(48) conclusion
The wmdow is open. Thete must have bt:en.1l burglat.
(49) justification
Now I am throwing in the towtl. rvt- tried It tt:n limes.
(SO) solut ion
No single measurt has had an effect. Wilh Ihis proposal our goals will be
achleved.
(5 1) motivation
Da you w.1lnt to know Send us a stamped sclfaddresstd enveloJ.X'
Other types of pragmatic relations are distinguished in the Hterature as weil, for exam-
pie the following by Eve (1990): eplstemie, spCh act .lind metahnguistlc
1
6.4 Rhelorinl SlruclUIl! Theory 111
relations. Epistemie relations are pragmatie relations, expressing a writer or speaker's
condusion based on a causal relation in reality. An example is (52), whieh shows that
the wrilt:r or speaker's knowledge that the has drunk .11 lot produces the condusion
about the headache. lhe connection then does not He in the external realilY, but in the
mental domain of the speaker.
(52) lle must have a headache. He has drunk too much.
In a speech act relation the speech act is motivated by reference 10 I situation conSlitut -
ing the reason for it, forexample: "What are you doi ng tonight , because there is a good
movie on': Metalinguislie relations refer to discourse itself, for example: "In condusion
I would like to remark .....
In !his seetion various sorts of discoune relations have been presented. Section 6.4
introduces a notabit theory on how 10 analyu these relations, while in 5e'Ction 6.5
somt problems and restarch topic.s concerning discourse relations Ire dealt with.
6.4 Rhetorical Structure Theory
In the past two decades stveral attempls have been made to creatt a method for the
analysis of discourse and discourse relat ions betwecn text stgmtnts. One of tht best
known proposals is tht: Rhetorieal Structute Theory (RST) by William Mann and
Sandra Thompson. This thwry, developed in the 1980s, considers .11 discourse to be a
hierarchical organization of text segments.
An RST analysis starts by dividing a text into minimal units, such as indtptndent
dausts. Then the conneclion bt:tween these units is labeled by choosing a relation
name. Mann and Thompson propose a set of over 20 relations. They distinguish
subject matter relations and prestntational relations, .11 division that roughly cor
responds to the semantic-pragmatic dichotomy. A schematic overview ofthis das-
sification i5 given below.
(53) QassifiCiltion of RST relations
Sub)tct matUr relations
Elaboration, Circumslanct, Solutionhood, Volitional cause, Volitional result,
Nonvohtlonal C.1IUSC, Non vohtional 'Sult, Purpose, Condi\lon, Otherwisc,
Interpretation, Evaluation, Restatement , Summary, Sequence, Contrast
Prmntatiomll
MO\lvation, AntithesIs. Background, Enabltment, Evidtnet , JU5tificalion,
Concession
'"
6. Discourse COI'!nectioos
Mostly. Iht units in a relation art dther nudem or satdlilc. This means that ont'
member of thc pair, thc nudeu5. is mon: 10 the writer's purposc, whik the
supporting element is the satellite. A pair consisting of a nudeus aod a satellite uni! is
(alled a span. Spans can bt linke<! 10 other units or spans. 50 thai the lext as a whole is
connecttd together ioto a hierarchie slruclure. Thc largesl span erealt'<! in this manneT
encompasscs thc whole lut.
Below is an example of a lext aod thc corresponding RST diagram. lhe text is
divided ioto six units, beginning with thc lilie, indicated by numbers which have
beton added.
(54) I.
2.
3.
4.
5.
6.
l.eading indicators
Steep dechne:s in capllal spmding rommitmenu and building permits, along
with a drop in thc moner stock pushcd the lrading composite down for thc
fifth time in thc pasl 11 monlhs to a Itvel 0.5% below ils high a ytar ago.
Sueh a diint is highly unusual allhis slagt in an expansion;
for exa.mple, in the thrt't most Tent expansions, the leadtrs Wtrt rismg. on
average, at about a 7% clip al eomparable phasts in tht eyde.
While nOI signaling an oUlright rteession,
tht eurrtnl protracttd sluggishnw of tht Inding indiealors appears
consisttnt with our prognosis of sluggish rtal GNP growth ovtr Ihe next
ftw quarters.
Unit 4, which cefers to previous expansions, forms tvidenee for unit 3, which states
that tht prtStnt dedines are unusual. Units 3 and 4 together are a span that elaborates
the dedines" thaI are mentioned. in unil 2. Units 5 and 6 interpret the span that
is formtd by units 2-4. Unit 5 provides limils on the degree of interpretation, while
I'
intt: :alion
'6
1
,..

2 3-4

,..

s ,
3 4
Figurt I. llST anal}'5l' of tat (54)
6.5 Oiscourw relation me.reh 11)
uni! 6 provides the interpretation. Finally, unit 1 is a tille that prtpares the reader for
whal is to eome. Note, though, thai the preparation relation was not included In the
original dassification.
6.5 Discou.rse relation research
Up unlil now, IlST has remained one of the besl-known and widtly applied meth-
ods for discourse analysis. However. being the center of attention. it has also been
the subj1 of debate and erilique in Ihe pasl ftw deeades. In the discussion seven
topics can be distinguished thaI are dealt wilh in Ihis section. One major eriticism
is the fact that the sei of relations in RST is purely dtSCriptive. not doingjusliee to
the differenees between specific relations, but throwing them all on two heaps. To
Ihis date, no agreement has been reaehed on exactly which relations and ealegories
should be distinguished. As a result. there is no gene rally aeknowledged standard
set of rtlalions.
As often as not, a set of relations ls presented without further slrueturing. There
is no order in the dassifiealion. Sanders el aL (I992) have argued thai a seI of dis-
course relalions must not only be descriptively adequate. but also be psychologically
plausible. In thdr view, the latter does not apply to, for eumple, RST, as the relations
Ihal Mann and Thompsan distinguish are all on the same level. However. same rela-
tions (e.g., volitional cause and evidenee) are more similar than olhers (e.g., vohtional
caust and contrast). Sanders et al. propose a dassification based on the assumption
that discoune relations are ordere<! in the human mind by four fundamental ordering
principles whieh ther eall "primitives":
(55) Sanders et al:S four primitives
I . Basic operation: Each relation has a cau$lIl or an eompontnt.
2. Souree of coherenee: Each relalion iJ roherenion JrnIantlcor progmatie
grounds.
J. Order of segments: lhis distinetion only appties 10 causal relations. These
have a bask ordu when the anttcedent (t.g.. cause) is to the left ofthe
eonsequence (e.g., rrsult) and a non-basic ordu when lht anttcedenl is to
the righl of the consequenct.
4. Potarity: A relation is positive when Ihe basic operation conntclS the contenl
of!wo text segmenls as given and when it eonntets the contenl of
ont of the segments with tht negation of tht content ofthe other.
The foue primitives can be eombined in order to obtain twelve classes of discourse
relations. The set of rdations can then be organiz.ed In terms of 115 own -meaning
".
6. Oiscourse connedions
when a relation contains causality. il belongt 10 a different group than
a relation thai is nol causa!. The causc-conscquencc relation, for example. 15 defined
as Basic Operation = Causa!, Souret of Coherence = Semantic. Order", Non-Basic,
PolarilY '" Positive. An example ofthis i5 (56).
(56) Last week il raint<! I 101 in Scotland, bause there was low pressure
Ireland.
A lhird point of attention is the definition of discoursc relations, i.e., thc assignment
of thc corrl" relation label. Owing 10 thc vagueness ofthe description. it cannol be
untquivocally determined which relations are applicable in an analysis. Consider the
foUowing example.
(57) SUt i5 corporate presldent. YOll should take mis 10 her.
Is this a reason relation or a conclusion? Ifthe aeeent is pul on thc first sentence, are
the relations of motivation or justification also possible? If so, a third problem ariscs:
relations are applicable that in, for example, RST belong to two separate categories of
relations, in this case subJect matter (e.g., reason) and presemalional (e.g., motiva-
tion). The view that perhaps two (or more) relations can hold at the same time is
known as the Multi -Level Hypolhesis (MLH). In essence;1 argues Ihat discourse rela
tions can exist on more than one levcI simultaneously, thus on both a semanlic and a
pragmatic level.
1\.'0 fervent supporters of MLH are Moore and Pollack ( 1992). They claim that
the interpretation of discourse requires the co-existenu of so-calle<! informational
and intentional relations. Informational relations, corresponding 10 semanlic rela-
tions, pertain to the relation between information in two consecutive text segments.
Intentional relations corrcspond to pragmallc relations. They conurn the notion that
texts are meant to reaHle changes in a reader's mental Slate. Moore and Pollack use
the following example 10 prove thai two relations exist at the same time. According 10
Ihem, a volitional cause as weil as an evidence relation applies between the two text
segments: the first sentence can be both a volitional cause and evidence for the action
in the second sentence.
(58) The president supports big business. He's sure 10 veto House Bill 1711.
A fifth important issue in discourse relation research is the nucleussatellile division. If
there i5 an asymmetry betwun the pUlS of a relation, e.g., if one member ofthe pair is
more essential to the wnter's purpose,then Ihe most imporlant element is the nudeus.
In (58), for example, the first sentence would be the sateUite in case of a volitional cause
or evidenee relation. In additive relations such a5 (59). there ean be two nudei, hut not
necessarily. The conlext decides which one is the nudeus. Ifin the case of(59) the topie
1
6.5 Discouf$e relation research 115
in the conttxt is forgetfulness. Ihen the sccond sentenee is the nudeu5. If the tOplC is
the fact that lohn loves cooking, then the first sentenu is the nudeus.
(59) lohn prepared a pie for his parents. (But) he forgot to pul it in the oven.
Sixth, there is the order of the pans. In (60), for example, the condition comes after
the stalemenl, bul the reverse order is also possihle, as in (61). where the concesslon
precedes the statement. This raises some intriguing quest ions. 1$ there a marked and
unmarked order of parts? And if so, under what conditions will an unmarked order
appear?
(60) You can go to thai party this Saturday. But first you have 10 dean up )'Our room.
(61) He hked laking care ofhis sister's kids. Yet he and his wife never had children
themselves.
A last issue worlh mentioning is the division inlo explicit and implicit relations. This
depends on the presence or absence of a conjunction. In the examples given above,
such as (59). it is dear that the use of conjunClions is optional. Moreo\'er, a conjunc+
lion can indicate more than one relation. Look again at example (60), in which the
word bul, indicating a coordmating contrast relation. marks a conditional relation.
The queslion i$ under which condit ions the use of conjunctions can enhance com-
prehensibility.
.. , 6. Discourse connretions
Questtons and assignments
Questions
6.1.1 Using YOUfown aamplH, showthat substitution and ellipsis deal wilh gr.amlniltical relationships
while deals wilh semanlic rtlationshlps.
6.1.1 Usi"g yourown v:ampln, show that Ihe conjunclion aMl can relationships
addition.
6.1.3 Wbich cohnion are cootained in lhe flOm in Wooderbod1 For
an analysis, see Halliday aod Hasan (1976: MO).
(I) lasl word in a 10ng 50 like a Ihal "li slarted. (1) Iooked
allhe who Sttmed 10 ha...e suddenlywrapped herselfup in wooL (J) "Ike rubbed her
q'n, aod Iookrd again. W SM couldn't wh1.t had 01.1 alL (5) Was she in
a shop1 (6) And was that really - w;u it reaJIy a shetp that WlIS sitting on IM OIt!er side of the
(7) Rub as SM would, she COIJld nothing lTIOf1! ofil ...
6.a.l "re lhewords in italIes in the following discourse (laken from '980) ana-
phora. orcataphor.al
As for your SKond questlOll : to thal (,) I ooly wi$h to say IM mIlCh (1): thi'S party fiods lhe
proposaJs and salisfactOf)'. Tltmln (J) lies a possibility forallewiating tl!e situation.
Thai C4) is why I Olm sure will ralify Ihi5 bilL I would 10 add D!N' IlIlns (5): WI! are
ofthe opinion thaI dlls (6) is only a fir11 step. ThiS (7) has always ourstandpoint; Il!at (8)
will not And finally, to set back layour Ihird question, th( foIlounns (g): discussions
are still soingoo. Thot (10) ts why I cam\Ol 5.1y at this time. In my.new it{n)
is possible: that the will deal wilh Ihis issue when il nexi IJIC'tts.ln anycaM', IMI
(11);5 whal I will propose; ift am correct, IIIhol (13) will happen is Ihal your suggestions will
be taken inlo account.'" IhlS IIICIJI ('4) il will be bill 10 be presented to Parlia
menl the summer ress.
6.1.1 Tryto a:plain why Ust OCCUr1 mucll more olten Ihan

6.1.) In which is the all,llpllor"he" lTIOf1! difficult 10 inleqlret aod whyl
a. lobn told Petc thaI ht would to quit.
i
Questions and ;usignmcnts "7
6.).1 The foIlowlO' (taken from Michael HOqt, l!J8j)un be placed in 14
seqlRnCH:
a. I was on stntf)' duty.
b. I saw Ihe fncm)' approaching.
c. I opened fire.
d. I beat offlne allack.
8elow thlft ",mpln using subordinuing, cooo:linating aod other link'ng constituents.
a. I beilt offtht a!tack IHJ openl"g firt. Thtn I sawlhe enem)' whilt I was on senlryduly.
b. I SI"'" Ihe ellcrn)' approaching bauSf I was 011 senlryduty. BUlt Art and I beat ofTtl1e
.nack.
c, SHing the ellern)' appmach wu for UlUW10 ope1I fire. In th.s IIIIIY I beat offthe allxk. At
thol 'HM I was on senuyduly.
Providc additional rumplcs of yourown.
6.3.a Oesignale the relalion(s) betwttn two M'ntenus In rach ofthe eumple:s below.
a. In the memo, possibilitits forretrainlllg and a:tra training_re dealt Wlth. ofils
COOlenlS will impmvt yourchaoces ofemployment el-n.ere in Ihis company.
b. in ourdep;lrtment are IIOt motivated to go to night $Chool. "5
employees' representati...e 00 the boan:ll would like to stress importance of continuing
tducation.
c. NoiM' poIlullon has not decreascd. This is btaust IIIsulallOll factorofthe souodproofing
was Iower than was a:pecttd.
d. Study Ihe plans 1I your Then it will become elrar tllat all the specificalions lIave bten
followed to tlle
e. Tht govemment lias a numberof new meawres. They will becomedfl!'Cti...e mext
year.
r. The ntw plans can bt t<:uted. Tht bo.rd of direclor1l1as allocated a sum of money
for juS! tllis purpose.
6)3 Otlermint relations in )8-4S and 56 are implicit aod which
paru are nucltl and which are satt!hlts. Ottnmint which words are netded or netd 10 bt Itft out
in order to change Iht order of tht parts.
6.4.1 an RST analysis oftlle following lext taktn from Mann and Thompson ('988:l53).
, . farmington police lIad 10 htlp conlro! traffic
2. when hundrtds lined up to be among II!e fir1t applymg fOr job-I; at lhe yet-to-open
b. Jeton toId should stopcomplalning. Marriott holt!.
3 hotel's - for 300 openings - was a rare opportunily for many
unemployed.
:'
-
120 6. OiscourK ConnKlion$
6 .
6 .
6.,
6.,
6.,
Bibliographical information
The discussion of coht:Sion is based on tht: standlrd werk by HIUiday and Hl$;ln (1976) in whlCh
tm, dllferentlypt:l of cohesion a/l' ddint:d Tht: definition in Scction 3.6, a5 Wt:ll a5
I numMo{ aalples, was aha tmn (rom HaUiday and Hasan. For lakai cohe:sion, 5t:t:
Kn:ss (l989).
Tht: dlSlinctlO bdwn anaphora and cataphon i$ takm from Bhkr (5t:t: tn.: 1990 t:dlhOn). See
Hallldayand Hl$;ln (1976) fOl" tn.: term uopllor. Easy t:nlry to on anlphon i$ offertd
m Bosch (1983) and Poil: (1987). An QVn"V)ewof apcnlTKIltai IJ !'n"n by Sanford Ind
Wrrod (19811. For approach to mnmtial clenKnts and cOOttt-nct, 5t:t:
Chapter 18 in Juraf"5ky and Martln (2000).
A Iandnwt: pubhcauon on diJcount rdaoons that 1$ sull wonh rnding is Grimt:S (1975). Though
rdatlvdy oId, quallty coUtcuon of paptr$ on d!scoulK rdatiom IS Haunan and Thomp$Ofl (1988).
For I pubbcauon on lht POSIUon 0{ advt"l"b!a1 dluses in $pokm languagt from I dJSCOUIK ptf5pt(-
tln",5t:t: Ford (1993). (1996) I' I must far a good basts m cohes!on phc:nollM'na. Eumples
of rt:KVCh Lnto connectiVU are IIVt:Tl In Couptr-Kuhltn and Kortmann (2000).
Tht: hy publicluon on Rhtton1 lheory is Mann and Thompson (1988).
F..xamplc (l9) iI taken from Mann and Thompson (200I}.
For urlou.s vicwpoints on thc issuc of thc dassifition of discoursc relahOns. thc readcr is
to Hovy and Scotl (1993) and Knoll (1993). rumple (55) was laken from Knott and Sandtrs
(1998). lhe concept of initntionalily in dlSC()llrsc was inlroduccd by Grosz and Sldner (1986). Set:
Rambow (1993) and MOSCT and Moore (1996) for furthcr publications on thJS ,"uc. The laller also
louches "pon the nudtus-Jatdlue di.sunctlO Ln rdatlO 10 inttnllonality. For a good contribu
lIon 10 thc dlscussion on oyeriappins discoulK rdahOlU, 5t:t: Ford (1986). For a cntlCal analysis of
and PoUad:'s .pproach, 5t:t: Sandcrs and Spoort:n (l999). Many good pap<"'"' on dlKOUI"SC
research an: f()llnd In Sanders. Schdp<"roord and Spooan (2(01).
1
,..
7 Contextual phenomena
Deixis
In the study of the relation between discourse and contell:t siI eoneepts are used fre-
quently. The first is deixis, the phenomenon in which the dependency of discourse on
tht situation is most striking. The ward deixis, which is derivcd (rom the Greek word
meaning "to show" or "to is U$ed to denote tbose elements in a language
whieb refer directly to the discourse situation. Dtietie warm are words with a refer-
tnce point that is speaker- or writer-dependent and is determincd by the speaker's or
writer's position in spaee and time. See the following example.
(I) I am now slandingon the roof.
lbe word "," reftrs to the person uttering tbe sentenee. lbe time whicb "now" denotes
is dependent on the moment the statemtnt is uttercd. This situation dependeney does
not oceur with words such as "roof"; the meaning of this word remains more or less
eonstant in different sit uations.
The research into deixis was ins pi red by Karl Bhler ( 1934/1990), who also dcvel -
oped the Organon model (see Section 2.1). Bhler was one ofthe first to map out dcic-
tic phenomena. He distinguished two fields in language: the deictie field (das Zeigfeld)
and the symbolic field (das Symbol/eid). Words sueh as roof. run, niee, ete., belong to
the symbolie field. These words - eallcd Nemlworter by Bhltr - have a more or less
eonstant meaning, independent of the situation.
Bhler compared the wards in the deictic field to signs on a footpath that dircct
walkers to thtir destination. The ward I points out the speaktr and the word you the
listener. Likewise, therr points 10 a spe'Cific place and yesterday to a spific time. Bh-
ler distinguishes person, plaee, and time deIXis in contrast to mental or phantasmatic
deixis. This latter form refers to amental or fantasy field. The phanlasmatic form of
deixis (Deixis (.Im Phantasm(.l) can be seen in noveIs in whieh Ihe first-person narrator
does not necessarily have to refer to the authoT. It can also OCCUT in quotes.
(2) Pett said: "1'11 do somtthing abaut It lomorrow!"
'"
1. Contextual phCf1OfI1ena
Thc deietic fjeld of Ihe quote- is different (rom that ofPete lhrec time fields play
a role here: a. the time at which lhe speaker ullem:! thc senlence; b. the time 31 which
Pele said wha! is being quote<!; c. Ihe moment -," refers 10.
AI the center ofthe deiclic neid. which Bhler ealls the Origo. are the words I,
hert, and IIOW. Deictic words are genen.lIy focused (rom Ihe spcaker's pcrspective.
In other words. deixis is egocentric, wilh an I-hert:-now Origo in person, plaet and
time deixis.
a. Person deixu
Deixis 10 person is rcalized using personal pronouns. Thc speaker 35 first person, I,
directs the utterance 10 the listener as SKond person, you, and (ould be talking about
a third person. he or she. In many languages person deixis can also conlain ather
meaning elements. for example, the gender of tht third ptrson. The manner in which
the second ptrson is addressed an, in some languages, also provide ;uJ in5ight Into
tht relationship bt'tween the first and the second person. Thi5 phenomenon i5 olten
ealled sodal deixis. The best-Imown example ofthis is Japantse, which has an elaborate
system of polittness forms called The ehoict of a speeifie form of address
i5 determined by, among other things, the gender and social status of the addressee.
An phenomenon in this regard takes place with the deixis of the first
person plural, w('. This word ean mean the group as a whole:
(3) Do wt have time for that? (when Ihe utteranee is being dlcwtd al the group in
general)
This is the indusive MIt'. The word Wt' ean also be used to denote a segment of a group
exduding the other members of the group: the exdusive Wt'.
(4) Do we have time for that? (when yoo ask someone dSf: for adviee)
Oddly enough, the exdusive w(' an also be used to denote precisely that ucludtd
group.
(5) Do we have time for that? (asked bya mother who 55 her children taking oul a
new toy two minutes befOTe bedtime)
b. Pfau deixil
The following i5 an example of plaee deixiS.
(6) Lelt afMr. A Sits Mrs. B.
This statement is initially interprettd from the speaker's viewpoint and not from Mr.
/\S. lf a statement like this is made to an audienee from astage, the referenee point will
be mentioned in order to avoid eonfuslon, e.g., me left, but for you right of Mr.
1
7.1 Staging 11)
A.- In plaee deixis a speaker ean refer to somtthing thai i5 in the vidnity or further
away: this, these as oPpoSf:d to that, thQst'. Plaee deixis ean be not only by the
uSt' of demonstrative pronouns, but also by the uSt' of adverbs of plaet: here and there.
In other languagts there are more subtle distinetions. Latin possesses, in addition to
the words hic, whieh means "that whieh is dose to the speaker" and iste, which meilns
"that which IS dose to the lislener': the word ilIe which meiln5 -that which is nelther
dose to the speaker nor the lislener".
An interesting phenomenon in plaee deixis is the ilmbiguity that aTiSe5 benuse
ean take plaee from different spatial positions. The following sentence ean
have at least two meanings.
(7) Mary is standing in front of the car.
a. Milry is standing betwn the car IOd the speaktr.
b. Mary i5 standmg in fro! of the car's front end.
If (7) has the meaning of (7a), the plaee-hound deixis is to the sptaker; if it
has the meaning of (7b), it is rclated to the ear. This i5 the differenee belween speaktr-
oriented and objectoriented deixis.
e. TIme deixis
IMixis to time would seem 10 be a simple form of deixis. The language resource:s are
the adjectives oftime m the sequence ... yeslerday ... now ... tomorrow .... IOd the verb
tenses. The verbs, however, sometimes also have another funetion besides referring to
a spedfie time. See the following examples.
(8) I had been walking thm. (past perftet progressive)
(9) I have bn walking there. (prestnt perfect progressive)
The past perfect and the present perfecl (whether progressive or not) both refer to
evenlli or aetions that starttd in the past. One of the main differenees is
that the perfect always indicates that either the event or the time frame I
which it takts plaee is still going on, which cannot be said ofthe past perfect. Time
deixis i5 often aecompanied by other meaning elements and is, therefore, diffieult to
isolaie.
Staging
Words in discourse follow eaeh other in a linear fashion. This dOts not mean, however,
that the information in discourse is presented linearly. The information i5 presented
in line with the Importanee it is supposed to have in a given eontut. Compare the
following examples.
... 7- Conte.tu .. 1 ..
(10) a. lohn issick.
b. 16hn is siek.
In both sentenees, information is provided about lohn. Yet Ihere is a difference. In
(IOa) the aspt of sickness is in lhe foreground, while in (lOb) the fact that it is lohn
who is siek is in the foreground. Wha! is foreground information in (IOa) i5 hack-
ground information in (lOh). Below is another examp]t,
(11) a. Every y('ar I go on vacation 10 Aruba for two weeks.
b. Every YUf I SO 10 Arubil on vacation for '''"0 wetks.
C. foT two wedts cvcry year I go on vacation 10 Aruba.
d. I SO on vacalion 10 Aruba for two wecks every year.
Using normal intonation, will be slightly a(eente<! in sentence (1Ia). In (llh),
the importan! demen! is that the activity in question is a vacation; the information
given on the destmation and the time spenl there is background mformation. In
(11 cl, two is foreground information and in (l i d) every year
M
is in the fore
ground.
The phenomenon of foaground and background information is calli, using
a lhtater melaphor, slaging. Speakers and wnters can present thtir information 10
such a way that some elemenls wi ll be in the foreground while others remain in the
background. The theater metaphor can, however, be misleading on one point. The
rdalionship between foreground and background in discourse can be much mort
eomplex than thaI on astage.
The head- tai! principle is a good starting point for analyzing the presentation of
IOformation. The more 10 the left (head) or right (Iail) the information is presented,
tht more important, promintnt and in tht foreground it becomes. This is ilIustrattd
by the following two sentences.
(J I) e.
r.
Every year I go to Aruba for Iwo wecks on vaeation.
Aruln is where I go on ncation for two wecks every yen.
In (lle), "on is more in the forcground than it is in (Ilb), where it is in tht
middle position. In (11 f), "Aruba" is more in the foreground. The front positioning can
be aecentualed bya so-caUtd eleft conslruclion in which a sentence oftht form "x does
isgivtn thtstructurt"z is whtrex does T .... odifftrtnt mtthodsofslagingcan
be observed in Iht following examples. FollowlOg the neutral order in (12a), Ihere is a
topicalizalion in (12b) and another eiert eonstruclion in (12c):
(12) a. I asked her 10 marry me in the middle of an autumn slorm.
b. In Ihe middle of.n .ulumn storm I asktd her 10 marry me.
e. lt w ...s in the middle of an aulumn storm thai I asked her to marry me.
r
7.1 Staging 115
The head- tai! principle is also al work in paragraphs and longer pas.sages oflex!. Com
parI' the following passages.
(13) a. I am against an expensive overseas vaeatlon. \VI' have already spent $0
much money on special things Ihis year And after all, there are so many fun
things we can do in our own country.
b. We ha\'e already spent $0 much moneyon speciallhmgs Ihis year. And after
all, there are so many fun Ihings we can do in our own CO\lntry. That is why
I am against an expensive overseas vaeallon.
c. \VI' have a1ready spent 50 mueh money on speciallhings Ihis year. That is
why I am ag ...insl an expensive overseas vacahon. And after all, Ihere are 50
many Fun things wecan do in our own country.
The most important message in this passage is the opposition to an expensive overseas
vacation. This message must, therefore, hold a promment position; in (13a) It is al the
beginning, the head, and in ( 13b) al theend, the tai!. The middle position is the least
conspicuous; Ihus (I3e) seems to be leu eohesive Ihan (13a) or (13b).
On the basis oflhe head- tai! principle, il can also be dtduetd why sentences some
times appear not 10 link up very well. Compare the following passages.
(14) a. lhe htalth departmenl, in a repon on canle neglect, states that more and
more farmers Ihese days are confronled wlth financial and infrastructural
problems, whde in Ihe past caille maltreatm('nt was usuaUy caused by lack
of food and expertise. The problems cannot, howe\'er, be blamed solelyon
the failing EU policy of recent years.
b. The health deparlment, in areport on eaule negiect, stateslhal while in the
past cattle maltre ...tmenl wu usually caused by lack of food and expertise,
more and mot(' farmers these days are eonfronled wlth financial and
infraSlructural problems. The problems cannot, however, be blamed solely
on Ihe fililing EU policy of recent yeal$.
The sentences in (14b) are linked logelher better Ihan those in (14a).ln (l4a) the ele-
menls "in the and -lack offood and are lowards the end ofthe sentence
and will, thcrefore, anraet more attention. lhe reader may thus expecl thai the fotJow-
ing sentence will eontinue dealing with these elements. This is. however, nOllhe case.
In ( 14b) the element at the end of Ih(' 5Cntenee atttacts more attentIOn than
in (14a). lhe following $entence continues dealing with this. This taU head linking is
why (14b) ean be judged as more coherent Ihan ( 14a).
An intriguing problem is Ihe question 10 wha! degree the malO-subordinat(' dause
distinction relltets the foreground- baekground relalionship, and to what exlen! Ihe
head or lail posllion oflhe dause is of inlluence. Compare Ihe following examples.
.. 6
,-,
7- Contedual phenomena
(J 5) a. 1I was already dark when OUT hcra awoke.
b. OUT hero awoke after it had already become dark.
1I appears thaI the information in Ihe main dause is more foregrounded Ihan thaI in
the subordinate Assuming that discourse proceeds with the foregrounded de-
ment, it ean be expecled that (ISa) will procccd 10 deal with the element -dark': while
(ISh) would center on Ihe element -awoke-, Consider thesc" final examples.
(16) a. Thc 34-year-old soldier, who had been arrested Ihr limes bcfore for
exlortion and fnud, has been sentenced 10 ten rean' imprisonmcm for
laking an afmy colond hostage. Thc court bc:lieved that Ihe soIdier WlII5 not
n'Sponsible for the dtath ofthe (olond. Thc colond was actually
killed by a marksman's shot.
b. The 34-year-old soldier, who was senlenccd 10 ttn ynrs' imprisonment for
takiog an army colonel had been arm;ted thrce times fot
610tllOn and fr.lud. 1he court that soldier was not personaUy
respomible for the dealh of the colonel. The colonel was aClually killed bya
marksman's shot.
Passage (16b) 10 be inferior to (16a). One possible explanation is thaI the con
tinualion in (16b) deals with the backgrounded information from the relative sub
ordinate clause in the first sentence rather than the foregrounded information abaut
"exlqrtion and fraud':
Perspectivization
Information ean be presented from a number of different perspeclives. Comparc the
following examples.
(17) a. was a man at the bar. 1hc door opened A woman and achlid came in.
b. 1hert was a man at bar. The door optntd. A "'"Oman and a ehild waIked
inside.
c. was a man at bar. looked up when the door opened. A woman
in, followed bya child.
d. A waman opened the door for ehild He walked in and saw a man
sitllng atthe bar.
In (17.) narrator is present inside the bar. In (I7b) the narralor is appe.rently not
lhe bar, otherwise the sentence would not have read "walked The nar
rator could, for example, be looking through a window inlo the bar in a position from
T
73 Perspecti'liullon 117
which he can the man at the bar but not the people outside the door. In (17c) the
story is told from the man's perspecllve and in (17d) (rom that of Ihe ehild.
The term perspective is used 10 describe these different points of view. The com.
parison to einemalic arl is often made by defining perspective as the eamera position.
In discourse studies three approaehes are of importance: firslly, the sociologi -
eally-inspired research into the ideologieal persptctive or vision; secondly, the more
literary-oriented research into the narrator's persptctive or focaliUllion; thirdly, the
syntactically-oriented research into the speaker's attitude, which is called empalhy.
a. Vision
Information can be presented from an ideological perspective: a system of norms and
values pertaining to social relations. This explains why lwo newspapers reporting on
lhe same tvent can produce different reports.. The following examples are the opening
sentences from a conservative righl -wing daily and a leftist daily dealing with a large
peace demonstration in the Nelherlands. Try to delermine which is which.
(t8) With 400,000 demonstrators puticifnting. double that of the organizers' highesl
estimatt$, the peace demonstration in Amsterdam has a1ready been labded an
important political evenl.
(19) The (ears on the fnrl o{ thousands of DUlchmen that peace demonstration
in Amsterdam would culminate in an aggressive anti -America orgy werc not
fu1filled.
Most readers will inslantly recognize Ihe progressive (18) and the conservative (19)
ideological perspeetives. The central quest ion in the research on vision is how an ideol -
ogy affects language use. Below is an example of experimental research that has been
done in this framework..
As part of a rcfresher course, a group of journalists were asked to write a news
story based on a fictilious event: a schoolteaeher who was on the verge of being fired
from her job at a Christlan school for becoming pregnant oul of wedlock. Afterwards,
the journaHsts, who were not aware of the research goals., wert a questionna..ire
that asked their personal views concerning the issue. One qucstion, for example,
iffiring the teacher was justifitd By selling the investigation up in this manner, it was
possible to ascertain that these personal views delermined the way in which a given
event was reported. Below is an example of differences in reporting. The malerial thai
the journalists were given included the ttanserip! of a telephone conversation with t he
teacher. At a certain point in Ihe conversation the teacher answered the question MO
O
they want to get rid o( as folIows:
u. 7. Contextual phenomena
(20) weil, I find il difficult 10 comment on this. yeah, weil, I don', think il i5
wise, wilh lhe dlsmis.sal and Ihe atmosphere at school wherc C'vcryone is lurned
againsl rot.
This anSwer was worke<!. inlO Ihe artides in different ways. Compare Ihe following
accounts.
(21) Thc leacher has dtcided 10 Will! and sec: wha! happtns: -\ find il ddficult 10
comment on Inis. Wilh Ihe dismissal and Ihe almospherc at school where
evcryone i5 turne<! agamst roe,-
(22) Thc ccntral figure in Ihis controwrsy has no idea why sM is being dismiSKd.
Thc first accounl was given by areporter who. according 10 the questionnaire. was on
the tuchers side. Thc hesitation in (20) is intcrpreted in a positive manneT as bring
a wait-and-see auilUde. Thc sond account was given by a journalist whose position
was neutral. The leacher is, nevertheless, portrayed in a negOlotive fashion. In (21)
the teacher is hesitant whereOlos in (22), it is stated that she reaUy does not know why
she is being fired. On the basis of the analytical model developed in this investigation,
it was possible to show that even journalists who say that their position concerning a
given issue is neutral also report in a subjective manner.
b. Focaliwtion
An entirely different approach is provided by perspective analysis which incorporates
narrative theory. The central idea is that the narrator could be someone other than
the individual who has witnessed or LS wilnessing an event. This is dear in (17c) and
(l7d). The person who is telling the story is not the man looking at the door or the
child who SttS the man. Following the French Iiterary theorelician Gerard Genette, the
term Jocalizalion is used to describe this. This relationship can be signaled in discoune
through verbs of observalion (to sn, 10 hear, 10 nOliee, etc.). In focalization, there is a
subject and an objcct, an observer and something that is being observed. The subjecl
of the focalization is called the focalizcr. This could be a narrator who is observing
everything from an external viewpoint as in (l7a) and (17b): in this case the subJcct
is called an external focalizer. It could, however, also be a charaeter in the story ilself,
as in (I7e) and (I7d); these are ealled eharacter-bound focalizcrs. Below are further
examples. The verbs of observalion havc been ilalicized.
(23) a. Pele gavc astart when he hlllrd thc man commg up the stairs.
b. Mary fell thaI Pele was stulled when he hellrd Ihe man commg up Ihe stairs.
In (23a), lhere is a charaCler-bound focalizcr and in (23b) Pele is embedded in Mary's
obJcct of focalization as a focallzer. Focalization analysis helps determine from which
observation poin! a st ory is being lold and if, for example, a change of pcT5pcctive has
1
7.3 Perspeclivi:Ulion 119
taken place. It also helps to determine how tension is buHt up in the story. Below is a
more elaborate version of the first story.
(24) There was a man at the bar. He looked despondenl. He was mumbling
something about murdenng his great love .nd his only futurc.
R
The door opened. A .....oman and a linie boyentercd. The boy gazed'l the
CU$lomer at the bar. 5uddenly he feIt the woman's hand in front ofhis eyes.
Through her fingers he could s ...
From Ihe verbs of observation used, il can be deduced that al first there is .n external
focalizcr. A change lakes plaee when the child appears on the scene, at which point the
siory conlinues from the childs perspcctive. The tension in this story is established by
the fact that the reader, thanks to the external focalizcr. knows more than the woman
and the child, namc1y, the Ihreat of murder.
c. EmfHlthy
Empathy in discouT$C studies is used to describe the degree to which a speaker identi
fies with a person or objeet which is part of an event or condition that is described in a
senlence. Theterm was introdueed by Susumu Kuno (1987). He mowed that empathy
is expressed in the syntaetic Slrueture of a sentence. Some examples:
(25) . lohn hit Mary.
b. lohn hit his wire.
c. Mary's husband hil her.
In (25a), the empathy is almost equally divided. In (25b), however, more empathy is
directed towards lohn than Mary. One indication oflhis is that "lohn
R
is in the subjt
position: another is that Mary islabeled as lohn's wife. In (2Se), the speaker identifies
more with Mary than with her husband. Kuno concluded that if a possessive noun
phrase, such as "Mary's is used, the empathy will be eloser to the referent of
the possessive (Mary). He also slated that two conflicting empathies cannot oceur in
one sentence. This .....ould explain why the following sentence sounds odd.
(25) d. Maryli husband hit his wife.
In the subJe'CI position, the speaker expTCS$C$ empathy with Mary aecording to thc
rule of the possessive noun phrase. In the objeet position, empathy is expressed for
lohn as Mary is referred to as "his Kuno also showed that Ihere are rest riet ions
on changing empathy. Comparc the following examples.
(26) a. Mary had quite an expcrience last night. 5he insulttd an important guest.
b. Mary had quite an expcrience lasl nigh\. An important guest was insulted
by her.
')0
]. Contexlual phenomena
Example (26h) is not as good as (26a), a fact whieh ean bc cxplaine<! as folIows. lbe
empalhr in Ihe first i5 with Mary. In thc second senlenee of(26a), the empa-
Ihy remains with Mary duc 10 Ihe she- in subject position. In (26b), on Ihe othrr hand,
I new eharleter is introduced in subj! position in the second sentenee and bccomes
Ihe focus of empalhr as a muh. Wilh Ihis example and SCOrtS of others. KURO showed
thai Ihe empalhr of the spe.lker i5 evident in Ihe scntential siructurc.
7-4 Given-new management
So rar we havc seen thaI three faetors play an important mle in presenting information.
First, Ihere art dir! links with Ihe situation via deixis 10 person, plact and time (7.1).
In Ihe discourse situation some dements ean COffiC 10 the foreground while olhers
remain in Ihe background (7.2). The information can also be described from certain
perspeclives (7.3). In order to determine which perspectives, the following quest ions
have to be answered. From which ideologicaJ viewpoint is the issue presented (vision)?
Whose viewpoint is being communicated (focalizalion)? Which characler does the
speaker identify with the most (empathy)?
There is a fourth important variable in the prtsentation of information: the
knowledge on the part of readers or Iisteners that is assumed by the speaker or writer.
Research on the use of definite and indefinite artides provides a good introduction to
the research being done into given and new information.
The psycholinguist CharIes Osgood (1971) did the followmg experiment with
his students. He aske<! them to describe simple observations. He used the following
objects: an orange ring, a black ball, a red cup and a green cup. With these objI$, he
gave demonstrations which can be described as folIows.
(27) Osgood's demonstrations
I. There is an orange ring on the table.
2. Someone is holding a black baU in his hand.
3. There is a black b;lll on the table.
4. Someone is holding a red cup in his hand.
S. There is a green cup on the lable.
sgood asked his students to describe, as simply as possible in one sentence, what
they observed. The demonstrations [11, [3J and [5J are equal in the sense Ihat Ihey all
deal with an objecl on the table. Yet, the descriptions proved 10 be different. BeJow are
a few eumples.
1
(28) There i5 an orange ringlying on the lable.
(29) The black ball is on the table.
7.4 Given-newmanagement 13 1
In descriptions of demonstration 1I J, as in (28), the word -ring- has no definite arti-
cle. Demonstration 131, which followed a demonstration in which a black ball had
been shown, Ie<! to a goOO number of descriptions contaimng definite articles, as in
(29). The same held true for the final demonstralion in which students wert shown a
green cup after having heen shown a red cup. With this relatively simple experiment,
05goOO showed that adefinite artide is used when of the objects involved
is presumed.
However, the distinction given-new isless dear-cut than it sterns. It has been sug-
geste<! that new information is information whjch cannot be ascertained from the pre-
ceding discourse. The following eumple shows how problematic this definition iso
(30) Your f;l,tner did it.
The new element hert is "father': Butthi5 element can in part Sttn as given, as the
speaker can assurne thaI the concepl fa/h" is in the listener's consciousness. For this
reason. it has heen suggested that the listener's consciousness be incorporaled into Ihe
definition of givtn and new. Wallace Chafe (1976) provided the following definition:
"Given (or old) information is thaI knowledge which the speaker assurnes to be in the
consciousness oflhe addressee at the time ofthe utterance. Socalled new informalion
is what the speaker assumes he is introducing inlo the addressee's consciousness by
what he $315'-
Within Ihis approach, even finer distinctions are necessary; Ihere are a
of gradations givtn and new. Ellen Prince (1981) suggesied the following
distinctions.
(31) Prince's given- new taxonomy
Ntw Infemblt
brand new
""'"'"
E'I1Oked
sltuational
texlual
An example of brand new is "an orange ring- in (28). An example of ummd is (30);
the concept of father is known but not yet aClivated. In the following examples the
inftrable elements have been ilaliciud.
(32) My whole suitcase ....7.5 searched by Inem. Luckily, it didn', occur 10 Ihern to
open the tube oftoothprutt.
(33) I was approachlg the Ilersection at high speed. lhe traffie light was grn.
.,>
,.,
1. pllenomena
In (32) Ihe demenl of loothpastc" is somewherc bttween new and given. 1I
has not bern mentioned berore, but LI ean be infcrre<! from prior knowledge pcople
have concerning Ihe conlenlS of travelers' luggage. Approximalely Ihe same is true
for (33). MThe traffie light
H
is neither given nOf new but ean be inferred on the basis
of Minterslion':
lhe evokw elements are given. This is possible becaust of the situ-
ation. Ir in (33) thr first.person narrator is already tdling Ihe story. thC'n Ihe - ,- is
situationally evoked. Texlually evoked refers 10 those elements that have already been
mentioned in thc discourse. Below are IWO furlher examplcs.
(34) It has been S<1id thilt a good deal ofhashish is used Ihere. But while I was Ihere
nobody smoke<! a )Ol1It.
(l5) There was a young couple walking in front of me. While walking, he put hIS arm
around her.
For those students who find Prince's five part division too findy.mtshed, Ihere is
always Chafe's division: active, semiactive and inactive. A concept, accord
ing to Chafe, can be active (given or evoked) or inaclive (brand new or unused) in Ihe
Iistener's consciousness. A concept is semiactive (inferable) when it is quickJy acti
vated on the basis of al1 avail able knowledge. It is dear that, regardless of which analy-
sis is applied, a simple binary distinction given and new will not suffice.
Presuppositi ons
Examples were given in 7.4 that indicate that listeners and readers can somelimes
infer information from the In fact, this is a mort characteristic of
more can be derh'ed from discourse than is explicitJy slated. Considtr Ihe
example below.
(36) It look lohn seven yean; 10 complele his sludies.
followmg information can he derived from Ihis sentence.
(36) a. There is a person named John.
b. John was a student.
c. John was not a brillianl student.
information that there is an individual named John is notstaled explicltly in (36),
but an he derived from the factlhat aperson is mentioned who 15 called by.that name.
The fact that lohn was a Sludent is likewise not stated explici tl y, bul Ihis can be derived
from the statement thai he took seven yeats 10 finish his studies. Depending on the
75 PreslIppolhions ' 33
concrele situation, more information could he derivtd. (36) could contain
(36c) as implicit information if it had jusl heen slaled that the program John was in
usually takts four years 10 complete.
A speciallypc ofimplicil information is called prl!$upposllion, meaning - 10 aume
a term which originaled in the philosophy of logic. Information which
is explicitly stated is referred to 15 I claim or an lUSertion. The example above makts
il dear that alt kinds ofinformation can be derived from a sentence. The term presup
position is reserved for the implicit information which must he true for the sentence
in queslion to be ilselftrueor false. A sentence such as have stopped smoking" can
only be true or false if the person saying it in fact used 10 smoke. The presupposition
of this sentence is thus "I used 10 smoke." Put another 'A-"Y: a prtsupposition 15 the
only type of information that is unaffccted by denial of the origmal sentence. Look at
the following examples.
(l7) John is (nol) opening the window.
a. 'Ihe window i5 dostd.
(Ja) Democracy mUSI (not) be rtstOl'ed in Sunnam.
a. Sunnam was once a democracy.
The asentencts given here are presupposi lions because they are also true if (37) and
(38) 3rt denied. Of course, the whole sentence has to be denied, and not jusl one or
more constituents, for its presuppositions 10 be maintained. Note that a negative sen
tence can be denied; the result is then a positive sentence. In a more formal notalion,
the presupposition is written out as folIows.
(39) Bis a presupposition of A if and only if(A -+ B) and (.,A -t B)
The symbol -+ is the implication sign for and the symbol., is the symbol for
negation. 1be definition given in (39) is known as negation test.
Apresupposition is thus tht implicit information Ihat must be true for the sen
lence to be either t rue or false and which is not alfected by a ntgation. The implicit
information can be derived from different elements in a sentence. In (37) and (38) il
is derivtd from tht meaning of tht words. In (37), use of tht vtrb open" suggests
the window is now dosed, and in (38) Iht word "restored" can lead to condusion
that at one point or another thert was a democracy in Surinam.
Presuppositions can be prompted by the words themselves or by the sente nce
structure:
(40) Car! has thr flu again.
a. Carl has had Ihe flu before.
.,. 7. Conttl<tual pl!ellOfTlorna
(4 1) Garl is a better linguist Ihan Ptlt.
a. Pelt is a linguist.
Presupposition (40a) can be derived from thc word Magain': In (41) the comparison
implies that Pete has the same profession as Carl.
Emphasis also plays an important role in deriving presupposilions. Somelimes thc
emphasis is al ready dear owing 10 the syntactical slruclure as in deft COllstructlonS, for
uample. Ot in which MX is doing y" is given the structure is x who is doing (see
also Section 7.2). This pUlS extra emphasis on x, as in the following example.
(42) [\ was Pete who pointed out thc problem 10 mt.
a. Somebody pointe<! out the problem 10 me.
In the following sentence are four possibililies. dependmg on which word
re(cives extra stress.
(43) Pele seils p;untings 10 museums.
a. (htt) Pele, and 00 at dse.
b. (ulls) don nol away.
c. (pmntings) don nOC sdl sculptum.
d. (museums) Ptte does not seil painlings to individuals.
Similarly. a ccrtain presupposition can be prompte<! bya specific emphasis in (41) and
(42). If in (41) -linguist
M
is strnsed. then this implies (4Ib) below. If in (42) pointed
out" is heavily stressed. then (42b) is a more obvious presupposition than (42a). Pre-
suppositions can. therefore. be prompte<! not only by lakaI and syntactical elements
but also by intonation phenomena.
(41) b. Gar! is in other areas Inferior to Ptte.
(42) b. the problem myself.
One of the sentences in presupposition rfit'arch was originally uS!. in
an artide published at the beginning of the last century by the phllosopher Bertrand
RusseU. 'Ole reads as fo11ows.
(44) The king ofFrancc is bald.
lbis sentence has the following existential that iso apresupposition
whkh can be derived from a proper name or a nominal constituenl containing a defi
mte artide (5 also example 363).
(44) a. lhereisoneandonlyonekingofFranee.
,
75 PrestJppositions '35
Following Ihe definition of presupposilion. Ihere is an 0pposLte claim with Ihe same
presupposition.
(45) The king ofFnnce is not bald.
In the of (45). the same presupposition. (44a). LS presumed to be true. These
sentences pose a diffieult problem for philosophers and logicians. If I1 is assumtd thaI
either a claim or its negalion is true. soeilhcr (44) or (45) must be true, and ifil is also
assumed thaI (44a) can be derivcd from (44) or (45). then a presupposillon ean bc
deduccd which is logical but unlrue: France is, after all, a republic.
Although Russe11 suggested a way of getting around this problem, the solution
remained unsatisfactory. A half-ntury later the issue bccame a topie in presupposi -
lion research. In 1950 Ihe philosopher Peter Strawson provided a pragmatic analysis,
Ihe gist of which is lhe following: sentences can only he true or false if their presup-
positions are met (i.e., are true). Only in the situation before the French Revolution
was (44a) true; thereafter il was no longer lrue. So. only before the French Revolution
were the presuppositions of either (44) or (45) met and could they have a lruth value
(he either Irue or false).
lhe debale between Russell and Strawson played an important role in launch-
ing the research into presupposltions. If Ihe situation in whieh an utterance takes
plaee is laken into account, then Ihe research becomes far more complex. And yet.
this extension is a natural one. A stciel approach using the results ofthe negation
lest is only a partial mapping,oul ofthe information implicit in an utterance. From
sentence (46). for example. mueh more ean be dedueed Ihan Just the existential
presupposilion (46a).
(46) Gotothestudenladvisor.
a. There is a advisor.
Since (46) is an order. it can be deduced that the speaker is in a position to give orders
to the addresse-e. The problem iso however. thaI mueh unspoken information can be
derived from language in use. ApresuppositIon can even be instantly denied. ham-
ple (45) has (44a) as apresupposition. Language in use. on the other hand. is not
hampered by the conclusion that there i5 a king of Franee. The foUowing utterance is
aeceptable, at least for some language users.
(47) The king of Fnnce is not bald; then IS no king of Fran.
\Vhen discourse is looked at in a specific situatIon. ill$ not jusl the implicil infonna
tion derived from the negation test.the presupposition. that is available; other implicit
information is also derivable from a given senlence. lhe term for thlS is inference.
,,.
,.6
7. ConltJltual phtnomena
Infcrcnces
Inference is the colle<tive term for all possiblc implicil information which can be
dcrivcd {rom a discourse. The term inference (from thc Lalln "inferTe" meaning "to
carry in) is used 10 denoie thc phenomenon thaI discoursc summon! up knowledge
Of information which can be used 10 understand the informailon. The most significant
cases of Ihis, besicles presupposilion, are tnlai/mem, convtnlionai implicature. conwr-
sational implicature, aod connalolion. Bclow is an cxarnple of uch.
Entailmctll is a term taken (rom logic. Ir A is grealer Ihan B aod B is greater Ihan
C, thcn it can be conclucled thaI A is grealer Ihan Cln discourse studies the term can
be used more broadly. Look allhe following example.
(48) Pete bought oranges.
a. bought fmlt.
Example (4&1.) is an entailment of(48). The difference betwecn an entailment and a
presupposition is clear here. The entailment does not have to be true if the claim 1$
denied. Understanding entaUments is often necessary in order 10 make connections.
The inference that oranges are in the category of fruil is necessary iE discourse con
tinues as folIows.
(48) b. Pete boUghl oranges. Unfortunately, he completely forgot that he had bought
them. After St"'eral weelu his cupboard started 10 smell of spoiled fruil.
The term convenlional Lmplicature was coined by Grice (sec Seelion 2.4). Griee gav('
the following example.
(49) He is an Enghshman; he is therefore brave.
From the word "therefore'" one can, through the fiAed meaning or by convention,
derive the conclusion that Englishmen are brave. Grice calls Ihis type of implicalure
"conventional" in order 10 dislinguish it from conversational Lmplicature (see also S
tion 2.4). An example of the laller is given in the following conversalion.
(SO) A: Did you already buy fruit?
B: The are already in the refrigerator.
On the basis of Grice's second maxim of quantity (00 nol man your contribution
more informative than is required.), it would have sufficed for B to aoswer A
can assurne that B is complying with the cooperative principle and can, therefore,
also assurne thai B has a reason for providing what at first appears to be extraneous
information. Depending on the sit uation, B can implidtly communicate one of the
following.
1
7.6 Inferencn ']7
(SO) a. J'1I dlde whal kind of frulI is bought.
b. You know that 1 buy oranges every weck.
c. 1 have done even more than you requested; I have already put the fruil in
the refrigeralor.
An example of connotation is provided by the following story, which causes problems
for many readers.
(SI) A father and a son are silling In a car. They are m a St'rious aceiden!. The father
15 kille<! on Impact and the son is laken to the hospital in cntical condilion. As
the victim is whecled into the operating room, the surgeon exclairns; Oh no, I
can't operate. That's my son!-
The profession of surgeon, at least in some western cultures, evokes the imageof a man.
This association makes the: story puz:r.ling. The same is true of the: story about the: two
Indians, one: of whom is the son of the other while the other is nol his father. When
readers de:rive from the word surgeon" or that the: individual is a male, this
is an inference on the basis of a culturally-determined association or connotation.
The term inference covers quite a broad area of meaning. A number of attempts
have bttn made in the lilerature to develop a classification system. The two main dis-
tinctions made are those between nessary- and "possible:- and between - forward
R
and -backward". Compare the following e:xamples.
(52) No longer able 10 conlrol his anger. the husband threw the dehcate porcelain
vase against the wall. lt cost hirn over one hundred dollars to replace the vase.
(53) No longer able 10 control his anger, the husband threw the delicate porcelain
vase against the wall. He had been feeling angry for wecks. but had refused 10
.seele help.
To undersland (52) properly, it is ne:cessary to make the inference that Ihe vase has
been broken. In (53) this inference is not necessary. If the inference that Ihe vase has
been broken takes place in the St'cond sentence of (52), it is called a backward infer
ence or bridging inference. If in (53) lhe inference concerning the broken vase is
already drawn in (he firsl sentence, then it is called forward infe:rence or elaborative
inference.
To what exten( are inferences reaUy made during the reading process? One of
the best-known txperiments carried out (0 answer this quest ion was done by Susan
Haviland and Herbert Clark (1974). lhey had subjecls read fragments of discourse
such as the following:
(54) a. Herb took the picmc supplies from the car. The bttr was warm.
b. Herb look the bttr from the car. The bttr was warm
'0'
7. Contatual
Thc senlcnces werc shown on a and Ine subjKl$ werc 10 press a button when
they had finishcd reading the sentence. Thls made il posslble 10 measure rcading time..
If an inferenc!! is made, it is logical 10 assurne that il will take time. In (54a) an infer-
ence is necessary. namety, thai beer was included in Ihe picoi, supplies. The results
show! that Ihe time taun 10 rcad Ihe second sentence in (54a) was langer Ihan Ihe
time needed 10 rcad thc second sentence in (54b). It could bc argucd thai Ihis has
more 10 da with word repetition, which occurs in (54h) and not in (54a), Ihan with
inferenc!!. lbis explanation was, howtvt r, ruled. out in another experiment in which
fragments such as the following. in Ihe (ontext of a picnic, were used:
(54) c. Iierb was espially fond ofbeer. The beer was warm.
In lhis fragment an inferencc is also made, namcly, that Herb gOI some bccr. As in the
eanier experiment, the rcsult was that the sccond scntence in (54c) took longer to read
than the $CCond scntence in (54b). Inferences are indccd made duri ng the process of
lislening and reading.
Quntions aod usi,nments ' 39
and assignments
Questions
7.1.1 PI1Mdeeumpin which 5how mal thedeictical I lnd gou Cln ,Iso bt UsN lna-
phoricl1t)' and Cltlphoricll1)'.
7.1.1 Point out the lkictical in the fojlowing sentenccs.
a. phone ofJohn is ringing.
b. I've got
c. is no sIKh thing as ghosts.
d. Pete is sltting in thc glrden.
7.1.} According to ln estimalien bythc language philosopherYehoshua ('954),
our in nine out ohen Ilnguage Utlffill'lCfl.. Blr-Hillel cited thc f'oIlowing thouatll
6pcriment.
A logkian. Tom Brown, the NcwYur's oot 10 usc which contain
deictical He ootifiH his wife ofthe lnd dccidcs upon awakeningon
Ncw's Year Da)'that hrwould like bcd.
Whal senlence could he usc 10 malte his wish durto his wifewlthout u5ing deictical elements?
7. :1.1 Give a countercumplc for thc idca thaI mam-subordinatr riausc mirrors thc
ship forqround and backiround informallOfl.
7. ) .1 Is an externat focalileror a ch.al'3clrr-bound focaliurin thr sccood sentence in thc f'oIlow-
ing exarnplc1 Pmvidr arguments.
A man was si tting al thr bar looking dcspondenl. Thr dooropcned with 1 ud groan. A
warnln and a child Ullll!' in.
7.) .1 USII' Kuno's rmpath)' apPf"Oach 10 6plain wh)' the following Sll'nlrl1Ce is flawrd.
HIS brothrr WClJ hIt byJohn.
7-4-' DHc:ribr thr bctwren t he following senlenccs.
a. I was dnving down thr highway. Suddenly thr car brgan 10 wcaVl!'.
b. I was drMna down thc highway. Suddmly a car Mgan towcave.
What afC ur" and, car'" called in Princr's fivc-part division?
7.4.1 Ortrnnine which typcs of COMsion ($CI!' Scction 6.') quaJif)' as rIements and which as

' ..
7. ConttKtual plleoomena
7-5.1 Wl'oich inferencn in example (fo-C) c.tn be as presuppositionsl
1.S.1 15 B answering the claim orlhe presuppositioll in the following t'lth.nges?
A: hlld enough orthat nicket from Ilexl doof'.
B: Weil, then do somethlng about it-
A: 1"-': had tflOUgh ofthat racket from "ext door.
8: An! lhe neighbars making that much noise?
]'5-3 Howdoes B rne! 10 the pre'5upposition in the following exchanges?
A: Democracy muSI be rntOl'!'d in Surinam.
B: BUIlhere hn an)' democracy in Surinam.
A: Do)'OLI nill drink a lotl
BI: NO, I'W! cut down.
B1: I've MYer bc-en a hc .. yY drinker.
7.6.1 Indicate ir an is OnWlry or possible in IM fotlowing sentencn aod .flhe inlfl-enct;5
fol'Wlrds orbackwlllrds. Also determine ..... hat kind in ordertodl'llw the

a. His Yiews Ire quitt conservatiw, oot all in all he's a niet guy.
b. The Ileighbof's are having a partytonight. 11'5 png 10 be a short "ight's
c. A wudooe by S. Md!lam. wu tIM! foUowing.
d. He opeoed thedoor and immediately $tneCllnal someone el$t hOld bttn in thc 100m. Sud-
denly he feit cold 011 his neck. He stood still, p,aralyzed with Fora short while was
deathly silent lt WOIS 0I11y .ftera Ii!w minutK Ihat he noticed thai he was sunding in. draft.
].6.J. 01 abool thedlfferena in rnding time fthe last $tntente in the followlng frag
ments, assllming that when the name "John";s firstt"IKountere<t, IM forward-directed inference
is made John i5 01 schoolboy. {The $tntences were taken from Sanford and Garrod, 19BI).
a. john was on his way 10 school. The bus trundled slowty along the road. Last IOoftk he hOld
trouble controlling etus.
b. The 011 his way to school. The!xls trundled 5I0\Il1)' llong the U5t he
hOld trouble controllingetass.
Assignments
] .. 1 ComIdeI" lhe "To IM Ieft. ofMr. A is MIS. B". In terms of dN.is, this descriptlOll ein be of
two t)'peS (LeYell, .gB21. lt is i( Ais to IM leftofB (rom the speakers pointo(_. The
description is p,attern-oriented when 8 is to the left.o(A from A's point of view. This distinction
also be conSlrued 15 objt oriented. Now consider the rotlowing diagnm.
QutSlions alld assignmenu '41
""
Now 11")" 10 describe the diagram aoo.oe in weh a WOIY that a person who nas not seen
the figure would be able toduplicate 11. Compare yourdescription .....th tIM! two 1)'peS glven in
and (bI. Exptain whetheryoo have given an ego-oriented (speaker) or p,atlern-oriented {ob.J11
description.
a. I. In the middle, to begin, grey nodal poIni .
,. From lhere upw2rd, 01 nodat point.
,. Then 10 the left, 01 pink point
,. Then from pink again to IM left 01 blue nodaJ point.
,. Then ru.ck again

Then from red to the right a yellow nodal poinl .
,. And from yellow again 10 the fight 01 glftn nodal point.
b. ,. Istart at crossing point gra)'.
,. Go ur.aight on to red.
, Go left to pink.
,. Go str.aight on to
,. Turn around to go 10 pmk.

Go slnighl on to red .
,. Straight on to ye!low.

Straight on 10 green .
].1.1 Take an Ksay Ihat NlS been judged as inco/M:'rent aod 1I$t tne head- tail principle to analyu the
o(,n(OI"TIQtion.
].].1 Appl)'the three nations ofvision, (ocalizalion and empathy to the discourse ofa tennis match.
>
1

.,
I
[
5
'
"
g .

8 Style
8.1 Form, content and situation
It is possible 10 say approximaldy the same thing in any number of different ways. The
word style is 10 denote "different WiiYS": This word is derivtd (rom the Lalin
ward stilus which means pen. The form aflettcn is influenced by the war in which a
peo (fealher quill) is cut. ret it is possible 10 wrile the same letters with different pens;
the letters only differ in their style. WhC'n we rxamine thr u.sc of thr ward "pen"' in thc
expression "His pen is dippe<! in we ean see that "how 10 writc"' also means
"how 10 forrnulate':
The concept of style sterns from old. dassical rhetoric, thc theory and pnactice of
effective language u.sc thaI can persuade a pub]ic in a special situation, t.g., 10 reduct a
suspect's sentence or 10 gain support for going 10 war. Stylistics has a history of about
2,500 The Grk philosopher Aristotle (4
1h
century B.C.) says in hl.5 famou$
Rhetori(a that an orator has to use a style, a form for his material, which is appropriate
and also "unaccustomtd" by using metaphors, for example. And the Latin rhetorician
Quintilianus (1
11
century A.D.) presenttd in his Ik Inslltutione Ora/oria four "virtues
of style": correctness, perspicuity, appropriatcness and ornamentation. Nowadays,
questions concernlg correctness are mainly dealt with in writing tducation. The
of a text is now studied in the cognillve approach to discourse, how people
understand information. The other two characteristics of style still play an important
role in stylistic research. In fact, the tension between the appropriateness (to the situ-
ation) and the elegant deviation or ornamentation (to draw attention) still lies at the
hurt of discussions about good style.
The starting pomt for styhstic research is that the same conten! (a be expressed in
different forms. A well -known example is the foJlowing. When an individual wishes to
tell about a man he had seen acting strangely on the bus that morning and who he had
met again, coincidentaUy, that aftemoon, there is an infimte number of ways 10 do so.
Bdow are Just two ofthe ways it may hedone (see Raymond Queneau, 1947/ 1981):
( I) How ILghtly packtd in we were on that bus platform! And how stupid and
ridiculous that young man looktd! And what was he dOlng? Weil, ifhe wasn't
actually trying to pick a quarrel with a chap who - so he dairned! Ihe young (op!
,.'
8. Stylt'
_ kept on pushing hirn! And then he didn't find anything heUtT 10 da than 10
rush offand grab a seat which had herome freel1nstead ofleaving it (or a lady!
T ....,o houn; after, gutSS whom I mt"! in front ofthe gare Saint-Lawe! ThC' same
fancypants! Being given somC' sartorial advice! By a friend! You(i never believe it!
(2) I was not displeased with my attire Ihis day. I was inaugur3ting a new, falher
spnghtly hai, and an ovtrcoat of which I thought most highly. Met X in front of
thc gare Samt -Laure who tried 10 spoil my pleasurt by trying 10 prove thai Ihis
overcoat is cut 100 low allhe lapds and that I ought 10 have an txll1ll button on
il. At least he didn't dare anack my headgeilT.
A bIt tuher I had roundly lold off a vulgar type who was purposcly iII-lrtillmg
me every IIme anyone wenl by getting off or on. This happentd In 001' ofthast
foul omnibi which fiU up with hoi polloi prccisdyat thast timts
when I have to conscnt to usc mem.
Many diffeunccs can be POlflted out between the tv.'o slorics above. In (I) the accent
is on the storyteller's at meeting the same person again, while lfl (2) the focus
is on the protagonist's experiences. Furthermou, in (2) adeparture is made (rom
chronological order. -Ibere are also differences in scntence structure and ward choice.
Still, the two stories have the same basic content. However, it proves to be very difficult
to describe the differences in vocabulary, grammar and discourse structure. And, even
more important 1$ the question to what extent these differences can be seen as stylistic
variation, i.c., variation that can be linked to appropriateness and ornamentation.
In discourse studies the concept of style is mainty studied with a focus on appro-
priateness, and not with a focus on the more literary ornamentation. Stylistics is more
or less restricted 10 prose stylistics, to everyday use of language. Within this domain
approprialeness is mainly related to the situational contut. lbis narrow use ofthe con-
cept llyle is covered by the term register, in the scnse of a church organ's registenlhat
can make the same piece of music (the content) have various diffeunt sound Mcolon"
(the forms). Register is the stylistic variation that can be uplained with the situation,
cr more pucisely, with the sixteen faetors of the SPEAKtNG model, introduced in Sc<:
lion 3.4. The IOOuence of these situational faeton is also evident 10 the terms used to
denote different styhstic variations: telegram style (the ehal1nel factor), city hall style
(the sct/ing faetor), and court ruling style (thc EtnfC' factor), etc. In many studlcs the
terms style and register are used without any distinction, but in eases where nol only
the formulatlon but also the relation between language and situation is meant, the
term register is to be preferred.
In the study of style and register the starting point is not only that 5OI11e content
ean have different formulations but also that these different formulations highlight
8.1 '47
different aspects of the possible meaning that has 10 be conveyed. So in the example
below there must be (slight) diffeunces in meaning as a rcsult of Ihe formulalion.
(3) a. John does not go to school, becausc Mary is ill.
b. Bausc Mary is iII, John docs not go 10 schooL
c. Mary is ill. So, John dOC5 not go to schoot
d John docs not go 10 school. Mary is ill.
Following the theme-rheme distinction Section 5.2), Ln formulalion (3a) the
rheme is the ilIncss of Mary, and in (3b) Ihe fact thaI lohn does not go to school.
The themerheme distinction prtsupposes what is called the len-right principle. This
means thaI new information (the rheme) is mostly placed allhe end of an utterance
ifthC're are no other uuons 10 put some information In focus, for example, by into-
nation in -Not Mary, hut lohn docs not go to school." So formulations (3a) and (3b)
differ in focU5 on what the wants to convey as kerncl information. Thc first
two fonnulations contain a and a subclalL5C, a subordination, and the last two
formulations have two full scntences, a coordination. Connecting or not connccling
full with connectives likt and '"so" can highlighllhe salience ofthe
relation. In (Je) the condusion is more salient than in (3b) of the usc of
and in (3d) the explanation is a more independent proposition.
From a stylistic viewpoint there is no free variation in formulation. Each formula-
tion has ils own stylislic" meaning. The assumption that there is a one-onone rela-
tion belween form and meaning is known as Humboldt's principJe. This inHuential
nineteenth-century German philosopher coined this principle in his research into
linguistic change. To give a simple example, if a language adopts the word bar as .....ell
as the word pub (with more or Jess the same meaning), a difference in meaning will
develop. Another example would be that if the same propositional content can be
expressed in different ways, say in an active and a paSSIve form, there must be a dif-
ference in meaning. Following Humboldt's principle every language has a tendency
againsl homonymy, one form wilh diffeunt meanings, like bank (of a river) and bank
(money) and synonymy, different forms wlth one meaning, like liberty and frudom.
Humboldt's principle forms the basis of much stylistic analysis: at first sight different
formulations may equivalent, but application of the principle will enable us to
recognizc subtle $hades of meaning aod shifts lfl focus.
8.2 Views on style
One of the most challenging problems in stylistks is the problem of defining style.
A description of allthe differences between the two storics at the beginning of this
,.'
8.
chapter d06 not guaranI Ihat the style has bccn mappcd out. In the hterature on
stylistics a great deal of attention has bttn given 10 tht' ddinition of style. The numer-
OU5 views on style can be divided ioto thrtt categories, corresponding 10 the Organon
mode!'s division iolo symbol, symptom aod signal (see Seclion 2.1).
First, when the symbol aspecl oflanguage (the reference 10 reality) i5 central, style
can IK seen as a possible form for a spetific conlenl. Second, (rom the angle of the
symptom aspect of expression (from the perspective of the writer or spcaker), style
can be a5 the choice of spedfic forms. Third, from the angle of the signal aspecl
of persuasion (the perspective of the reader or lislener), languagt can be seen as a
dmalion (rom a givtn expectation. Every interpretation. when viewed more dosely.
poses problems.
a. Style IU a poss,ble form for a specific COlllent
When dealing with style as a possible form for a spedfic content, the question arises
whether it is possible to alter the form of language without changing the content . At
the word level. the main queslion is: 00 true synonyms reaUy uist? Do the words dad.
father, and my old man have the same meanmg? In part ther da. narnely, the procrea-
tor of a child. The ans wer to Ihe quest ion whclher or not synonyms have the same
meaning depends on the definition assigned to the word mean'ng. If lhe definition
of meamng also indudes a reference 10 lhe dass of propie who use certain words, for
example. lhe fact thai predominantly YOllng children use the word dnddy 10 denoIe
lheir fathers. Ihen there are no lrue synonyms and Ihe condusion must be drawn that
the form does change Ihe content at least partially.
A similar argument holds true for differences in senlence composition. (See the
remarks on the difference between the aclive and the passive voice in Section 1.1.) The
research inlo stylistic variation presupposes that lhe lexts to be compared have some-
thing in common. This somelhing- may be called basic conteni, as in the IWO
in Ihe preceding seclion. In thai case, however, the basic content in question is no more
Inan "meeting on a bus': lhe propositional content of these stories is rather different.
In discourse studies "tnat which remlined unchanged" can be defined in a number
of ways. It is gene rally used to refer to propositional conteni. The focus when texts
are compared is then on differences in formulation. Tbe phrase un. however, also be
use<! to denote basic content, for example, sequences of evenlS such as in the stories by
Queneau given abovI'. Sometimes, from a more pragmatic viewpoint, a nlore abstract
definition is given. In this case it is assumed that only iIIocutionary force rernams the
same. Wilhln Ihis defimlion different Insic contents can be used. for uample. to utter
the same Ihreal. The interprelation of style as a posslble form for a specific contet
proves. therefore. 10 be too vague.
8.l Views 01'1 style '49
b. Style as a choiee of spljie palterns
To see style as choice patterns, one takes the point of view ofthe wriler or speaker. who
has a number of different possibilities in phrasing what he would like 10 say. Here is
a weU-known example. If A and Bare logether in a room and A wants B to dose an
open window. then A can makI' this dear in a number of ways.
(4) I. Could)'Ou perhlps dose the window?
b. Hey, cln't that window be dosed?
Co The window's open.
d. Oose Ihe window!
e. YOlI should be clreful about drafts, the way you're feehng.
i I'm not paying to hell the outdoon.
When dealing with style as choice, il is necessary to know what choices writcrs or
speakers have or had al their disposaL Some are partially determined by the
situalion. Perhaps A IOd 8 Ln the exarnple above do not have a relalionship in which
orders would be appropriate. The style-at-choice approach 11. however, suitable for
determining in which way the Situation limits choice, for example, which facton in Ihe
situation might eontribute to (4a) being preferable to (4b) in some eISes.
c. Style as a deviation from exptations
In a third view, style is tn as a deviation from expeclations. Owmg 10 long-term
exposure 10 certain routine patterns, readers and listeners develop expectations about
the way in which a specific content can be given form, and about the choice of certain
forms. When readers make a stytistic judgment aboUI a given form of language use, it
is apparently because the form deviates from what ther are used to.
Takt the genre "State of the Union which is reasonably well-known.
Because the genre is familiar, people have developed expectations concerning the
language in such a speech. Here are possible introductory sentences.
(S) a. In reeent rears our counlry has Iruly been put to the test. Many na\'e lost
lheir job$. Thousands of young people "''('re unable to find employmenl.
b. Our counlry has nOI had it easy 10 reeent years. Amons adult1. bul
especially among younS people. unemployment IS high.
c. These last rean have been touSh. reaUr tough. So manr have lost their jobs!
So manr roung people were never able to get one!
It is very Iikely Ihat the style used in (5a) will be characteriud as somewhat formal and
Ihal reAders will evaluate Ihe wording in (5b) as too informal for a Slate oflhe Union
address. Words like lively will probably be found applicable to the style used in (5c).
Formulation (5.1) will in alllikclihood be judged as the most sui lable.
'so
8. SlyIe
One problem with the style-as-deviation approach is thai readers' expcctations
can bc quitt divergent. It has thaI a norm for languagt use bc $t!t and
thai every deviation from that norm bc viewed as a stylislic characteristic. In order 10
do Ihis, however, it is obvious that normallanguage must be characterized first. FOT
Ihis rtason 11 has bccn suggested thai mate<! texts bc used as a point of comparison,
a5 a norm. In the stylistic analysis of, for instan. dvil service style. the differences
betwecn civil service documenls and murdeT mystery novds are not 31 issue, but the
differences betwecn these documents and editorials or informationalleaflets are.
lbese thr views of style can bc T:ogniud in currenl research, in which a great
deal of attention is paid 10 the problem of describing differences betwecn texts and
measuring the effect of stylislie on the attitude of Ihe reader.
8.] Stylistic analysis
The concepl styk In ilsdf is an evaluatlve concept, as we can s in exprcssions like
"Thai guy has Slyle normaUy means "good This can explain why stylistics
has a long trllldition in a more qualitative and normative approach deahng with
Hke: "How to prod.uce a good texl
M
or "How to avoid a 'complex', a 'rigid', a 'harsh' or
an 'informal' style': Apart from this approach. attempts have been made to connect
Ihe rllIther vague stylistic qualificatlons, like the ones given in the previous sentence, 10
patterns in formulation. These two approaches art dealt with in this section.
a. 7he normative approach
In many languages style have published in an educational setting in which
the style virtue referre<! 10 in Section 8.1, plays a central role. These
handbooks on dear or effective writing provide many hints and tips on how 10 prod.uce
a good text. One of the famous examples for the English language is the booklet The
Elements of Style by Strunk and White (2000), originally published in 1918, and still
influential al the end of Ihe last century. This guide to writing in a satisfactory style
concludes with 21 suggestions, a few of whieh are given below.
(6) Some caulionary hinls concerning good style
1 Place yourself in the background.
2. Write in a way that comes nalurllllly.
] . Work from a suitablt design.
4. Write nouns and verbs.
( ... )
19. Do not lake shoriculS at the cost of cluit)'.
I
8.] StytisticanaJ)'Sis '5'
20. Avoid fOl'C'ign languagn.
21. Prtfer the slandard 10 the offbeat.
Suggestions like thest cause different problems for a discourse studies perspective.
Let us look at the thrtt most important ontl. First, it is seldom cltar how these sug-
gestions can be implemented in producing discourse. What kind of stylistic deviees
must a writer avoid or use in order 10 place hirnself in Ihe Whal is
naturllli wriling? Second, must suggtltions always be foJlowed? In somt cases,
the use ofa mort unconventional slyle (see suggestion 21) or using foreign werds
(number 20) can be mort effective in reaching your goals. And Ihird, il is by no
means dear thaI acting upon a piece of advice results in a better texl. Is it really true
that, whalever tht circumstances. nouns and verbs should be prderred 10 adjectives
and adverbs (suggestion 4)?
Anyway, hints on using a good style still play an importanl role in educatienal
settings and organizations' corporlllte style guides. Thelr merits art thai writers are
focused on elements in discour,," that before were neglected. Moreover, they provide
a good starting point for criticizlg lhe way in which content is molded into a form.
Just one examplc ofthislanguage crilicism will suffice in Ihis section: the deceptive
metaphor.like the foUowing remark from Stion 3.3.
(7) Discourse is more than a message from sender to receiver. In facl. sendtr and
receiver al'C' metilphors Ihat obfuscate what is rtaUy gamg on in communicatlon.
First some information aboutlhis figure of speech. whieh was alrndy mtntioned as
a stylistie device by Arislotle. The metaphor is a form of figurllltive language in wh ich
an objecl or concept is denoted using another object or concept. This assignment
of one obJect or cancept 10 anolheT takes plact on the basis of cerlain similarities
between the two. Since the of a mountain resembles the "base" of a human in
some way, we can speak ofthe "foot" of a mountain. When human behavior begins
to resemble that of a certain animal. the namt of the animal is given 10 that indi -
vidual: for example, a sly fox". In everyday language many metaphors occur that
are no longer considered to be figurts of speech. These are called petrified or dead
metaphors. When the metaphor suggests a similarity that is non-existent. it is called
a deceptive metaphor.
A generllli suggestion in normali"e stylislies is to use figures of speech -hke meta-
phor, personification, irony, etc. - with a certain restraint, and thai ifthey are applied,
they mllSt enlighten some aspects ofthe conteni. Example (7) is a formulation Ihat can
be criticized with regard 10 perspicuity. When a discourse is characterized as a message
from sender to receiver, thcn a similarity is suggested betwn a receiver and a listener,
rtader or addresS. Ln Ihis telephone metaphor the suggestion i5 made thai a listener
'5'
8. Slyle
picks up signals in a passi\-c way. Nothing could be further from the Iruln. A listener
or reader i5 obliged 10 take an aclivc part in the communication proccss. For more on
Ihis. set' the on coherence in Scction 3.6.
b. The approach
In stylistic analysis Ihere is a lang tradition of describing differenct5 in formulation
patterns. In thest descriptions mostly impressionistic and C'valualivC' adjmives arc
used, whien prompt crities 10 draw comparisons with lasting wines or advertising cars.
Stylistic tvaluations of someone's writing stylt. such as "Iucid- or "objec-
may secm 10 have a fixe<! meaning, hut it is orten unclear what is meant by these
characterizations. A sinking example is cited in Grammllr os Style by Tufte (1971).
High-school teachers of English were asked 10 characterize the slyle of the novell/l
Cold Blood by Capote. The adJeclives range<! the whole alphabet and many of them
contradicted others. Here are some characteristics.
(8) Style. labels for Capote', tn Cold Blood (by Tulte, 1971)
alhterative. blunt, chisei!. discerning. elaborate. forceful, general, harsh,
impressionislic, hterary, meditative, natural, omniscient, prosaic. rambtmg,
stylized, thoughtful. unconventlonal, vivid, wordy
Even if people agree about the characteristics of, for exampte, the language ofbureauc-
racy, so-called officialese, it is undear 10 whal elements of style they must be linked.
In different cultures governmenllanguage is said 10 be difficult, impersonal and tradi -
lional. But II is unclear whether Ihe difficulty is caused by the conlenl, sentence length,
sentence complexity or Jargon, and whether, for example. a passive form always has
to be labeled as impersonal. AClually, even more imporlant is Ihe argument that an
official document has to be somewhat traditionaI, in accordance with the authority of
the governmenlal institution thaI produced it.
This more qualitative approach in style analysis has inspired much stylistic research
in dISCOUf5C Sludles in attempl$ to detect in what stylislic elements documents may
differ. lf an objective description can be produced ofthe stylislic phenomena of cerlain
documents, then there is a base for research into the link with stylistic labels gi\-en by
readers and the elfeci ofthe message on the reader. In Table 1 bclow some elements
are lisled which have been featured in stylislic research.
lt is remarkable that the concept of style is not confined to composition and formu -
lation. choices in content and peculiarilies in usage are somelimes considered stylist ic
phenomena as weil. The analysis of spelling errors, for example, is important in what
is called forensie stylist ics. a br2nch of stylistics that is used to find additional proof
in detecting the authorship of documents that playa role in legal cases (blackmail,
elc.). In analyzing composition and formulalion, the focus is on the elements that are
8.4 uamples ofstylistic mean:h '53
Table l. SQme elements in stylislic resean;:h
,.".1 mauet:
Content number .nd sort or topia. propolltional dtnsity, of elaboralions
Discoul"SC! "ruc!un pan.graph length, introduclion. conch15ion. rhetorical rt:lahons,
l'1ument slruclure
Syntu:
LuM:on
"",.
$Crlttnce length. variallon in senten<:e knsth, $Cnten.ce-inillalstr1.l(Iure.
number of relative clauses, sub- and coordination, passive voict,
deA coostructiom, nominalizationJ
W(lrd Icnsth, lexicaJ dlVCf$IIY ratio), frequencin and rallOS of
vuious parts oe speteh (advt"rbs. qualifiers. functton wordt., d(..),
hapu: legomml (W'Of"ds occurring only once), impenonal construclions,
liguus oe spttcb
frequmdes of characten, punclualion marks, spelling erTOrs
countable, for example, the type-loken ratio (TTR) as an indicator of lexical richness.
This ratio is the number oe different ....-ords (the types) divided by the numberof words
in a text (the tokens). The TTR ofthe previous sentence is 0.71 because words like the
aod wards were used more than once (in tOlalthere are IS typet and 21 tokens). This
TTR can be an indicator ofthe difficultyof a text. more types can make a text more dif-
ficult. AnOlher r2tio is the noun-verb ratio. If 11 could be prm'ed thaI some documents,
e.g., bureaucratic documents, have a more nominal style, fOT example, than newspaper
articles or noveIs (wilh a more verbal style), then this could bc an explanation of style
labels for officialese like "abstract" or
Of course, an analysis on the basis of these quantitative elements cannot produce
a full-f1edged description of the style. And many style elements shirk quantitative
analysis. How 10 measure Ihe st re;ngth of a good metaphor, for example? The style of a
document is as difficuh to describe as the character of a person. An old French saying
expresses this perfectly; -U slyle est I'homme mime- ("Style is Ihe man hlmself-).
8.4 Examples of stylist ic resear ch
In thlS section four examples of stylistic research are provided. The first example, Car
roll (1960), is important as it is the first altempt to describe all possible characteristics
that could be ofimportance stylistically. The nexl, Sandell (1977), is the first to embed
style analysis in the main purpose of rhetoric: persuasion. From research ofthe last few
decadcs, concerning many stylistic phenomena, two topi cs werc selected: Ihe rhetori -
cal quest ion, Howard ( 1990), aod intensifiers, Harnlhon et al. (1990).
.S< 8. Style
Carroll (1960) provided a description of linguistic characteristia thai may bc
responsible for the war prost styles are judged. Ht baS! his work on ISO passages,
each containing over 300 words, culled from many different types of discourse: news-
articles, letters. essays, scienlific articles, ete. Theu passages wert investigated
with attention 10 39 characteristics including sentence leogth, number of subordi-
nale dausc5, percentag(' of verbs aod the number of personal pronouns. In addition,
experts were asked 10 give their opinion of the texts. Thc exputs were told 10 mark
on an answer Shtel whal they thought orlhe texts: personal -impcrsonal, hard-easy,
seriou$-humorous, de. In totalthere wert 29 opposites. This yidded 68 x ISO pieces
ofinformation: 39 characteristics and 29 expert answers for tach of the ISO passages.
of thlS large amount of information broughtlo light a of
betv.n charaeteristics of discoursc and the way that discourse is judged.
There appeared to among other things, a rcJationship between the judgments
and and the discourse characteristics numerals-, and
A link was also discovered between the judgment and
the discourse charaeteristic "lots of adjectives". Carroll dKi, however, reffiark that this
did not prove that the charaeteristics described were responsible for the differences in
style. In this study relatively 5uperficial eharaeteristics were dealt with as the first step
towards a stylistie analysis.
Is there a specifie style for discoursc that is intended 10 persuade? One of the first
researchers who attempted to answer this question was SandeU (1977). He investigated
different Iypes oflexts; a type oftext that is definitively persuasive, advertisements. and
a texl that ahogether lacks persuasive elements, the short foreignnews item. Passages
were selected from news papers and analyud with attention to, among other things,
word length, ellipsis, and alliteration. The advertisements contained signifieantly
higher numbers of adjectives and intensifiers (for example. a superlative or words
like newr and always). The average length of the words was also shorter and ellipsis
oceurred more frequently. It was concluded on the basis of a statistical analysis thai
the primary charaeteristic of advertisements is thaI Ihey contain a large number of
adjectives. SandelI explalOed this by pointing out that for the description of a produci
it is these adjectives that have an eva.!uatlVe value and. influence consumer
attitudes. SandelJ poinled out that his research was based on a small random selection
of advertisements and news reports and thai the persuasive effect of other factors can-
not be ignored. Among other things. SandeIl mentioned the faetor domail1. Advertise-
ments can deal with topics that are enllrely different from those in newspaper articles
and may for this reason alone be persuasive.
Many different stylistlc elements have been invesllgated in the re!learch into
persuasive style. Here are two representative examples. Sometimes a claim is more
8.4 Eumpln of 51ylistic '55
convincing when it is formulated as a rhetorical quest ion. If a lawyer wantl a e hent
aequitted. the summation is more likely to end with (9b) than wlth (9a).
(9) a. The defendant did not intend 10 hurt his neighbor. He was always a very
peaceable man.
b. The defendant did not inlend 10 hurt his neighbor. Was he not always a very
peaceable man?
Research on the use of rhetorical questions such as in (9b) has shown thatlhis type of
question invites Ihe answer intended by the questioner. On Ihe queslion of why this
is true. opinions vary. It eould be that rhetorical questions are onen used when strong
arguments are presented and that !hey therefore become associated with forcef"ul
arguments. From the use of rhetorieal quest ions receivers infer that the arguments are
strang. 1fthis is true, rhetorical questions will inerease the persuasiveness of a message
regardless of argument strength. Another opinion is that a rhetorical question elicits
a judgment from the receivers. In order to baek their judgment. receivers will pay
more attention to the arguments presented If the argumentl are strong. they will pass
a more positive judgment than will receivers who heard the statement rather than the
rhetorical question (in the example, (9a and were not stimulated to pay attention 10
the arguments. If the arguments are weak, the opposite will hold. Receivers who hear
Ihe rhetorical quest ion will then pass a more negative judgmenl than will receivers
who heard the statement.
Research by Howard (1990) supports the latter explanation. Howard investigated
the constquences of asking rhetorical questions before presenting the arguments.
Again, the rhetorical queslions evoked a judgment. but this judgment was not founded
on the arguments. After hearing the rhetorical questions. receivers were no longer
influeneed by arguments. They had passed their Judgment and kepl 10 it This would
seem to prove thai rhetorical quest ions followlOg the arguments ean increase their
persuasiveness. whereas rhetorical questions preceding the argument ean anmhilate
the persuasive power of a message.
In the following example. It is madedear how language intensity can be ofinfluenee.
Which ofthe following styks is more eonvincing. the one used in (10a) or thaI In (lOb)?
Each of the two following paragraphs is the eonduding segment of an address ealling
for the legaliution of the sale of heroin.
,.'
8. Stylt
(10) il. the sale ofherom provides socicty with sevcral advantages.
1I would discouragc cTime by making heroin relatively inexpensive and
availahle 10 addicu. 1I would help in the fight agalnsl organiud crime by
taking away an important source oflhe underworld's incorne. Finally, il
would ntarly tliminate police corruption related 10 heroin trafficking by
moving the saJe ofheroln outside their jurisdiclion. Legalizing heroin
would also bc advantageous 10 thc USCT. It would gradually rtduce thc
number ofheroin-related inJurics due 10 diseasc and ove:rdose. In addition,
UStrs would bc able 10 better aKord other health-rdated producu.
b. Legalizing the Ale ofheroin provtdes society with several dear advantages.
It would deter crime by making heroin relatively inexpe:nsivt and available
\0 addicts. It would help In the fight against organiz.ed erime by taking away
an important source of the underworld's income. Finally, it would virtually
eliminale police corruption relaled 10 heroin trafficking by moving the
sale of heroin outside their jurisdiction. Legalizing heroin would also be
advanlageous 10 the user. lt would sharply reduce the number ofheroin-
related deaths due to disease and overdose. In addition, users would be able
10 beuer afford other heallh-rdated produeu.
These texts wert in uperimtnts done by Hamilton et aL ( 1990). Three faeton
were investigated: language intenslty, $Ouree eredibility and gtnder. To testthe effeet
of language intensity, the language was intensified in the b-version eompared 10 Ihe
a-version. The word "elear" was placcd in front of "advantages': The word "diseour-
was replaced by "deter", "nearly" by "virlually': eIe. To test the effeet of souree
credibi lity,lhe aulhor in one ease was said 10 be an Assistant Direelor ofthe Drug
Enforeemenl Ageney with degrees from Berkeley and Slanford, and in the other
ease Ihe author ....'a.s said to be a former addicl . To test Ihe effeel of gender, the name
of the author was varied simply by using eilher Ihe name lohn or the name !oan.
Before the subjects were shown the IUIS, Ihey were asked Iheir views concerning
Ihe legalization ofhemin.
Tht txperiments showed Ihat intensifiers had a positive mfluence on attitude
change. There did not, however, appear to be a direct link between language intensity
and attitude change. The lext was percei\'ed as being dearer due to the intensive of
language, and this elarity facilitated a change in attitude. In this investigation, the large
extent to which other faelors outside language are influential also beeame elear.
Intensifiers had a positive elfeel iflhe souree was seen as reliable; the same language
was eompletely uneonvindng if Ihe souree was feIt 10 be unreliable. Dther examples
of research on convincing people using stylistie devices are de;tlt w\th in Chapler 12
on ugumentation ;tnd persuaslon.
aod assignrTlffllS '51
Questions and assignments
Questions
8.1.1 Which o(Grice's muims (5ft Stelion 1.<4) <l1l' importance 10 sly1islic rHean::hl What
othe1- muims could be formulaled from a stylinic ptrspet"tivd
8.1.1 Which (acton (rom lne SPU,KI,..G modtl (see StlOf1 3. 4) all' ofimportance in tne foIlowlng types
oflangu;tgel
tmotioruJl formal stylt, Ifllaltst, of!'KiIIlnf, pttSUIIS/W JII11SulI,gr, SllCI!II wotkm' JIIr,gClft, Illd fJt1:hClrwt
Ial19uagt.
8.1.1 Describt, using p;trameters P and 0 from politeJlt5S thtOry (see Stetion 1.6), when which o(the
utterans "'al and (4b) would be appropriate.
8. 1.1 If style can be 5ftn as a ...ariatlon in form of appf"Olumaltly tht same cClfltent, what in ",el and (4f)
in Section S.lIS that $ImecClnttntl
1.3.
1
Mtt1phoric languagt' IS usu<llly IlOl suit.1.bIe lOr providing pIl'Cise dncriptioos. Camiderthe term
IOtllrclmmarwhich isoften used in discoul"Ststudin. ExpIain why this ttrm is <I metlphor, and try
to point out why it is deptM.>e in this tase.
8.3.1 What is the typt-toktn rallCl oftht first thm:' stntenct5 o(thls chapttrl
8.4.1 Shootd a sptaktr ust rhetorical qut5tions ifne wanlS his auditoct to think about his arguments?
g.<4.1 Should a sptaktf"ust intensifiers ifne w;tnlJi his auditnct to thlnk about his argumentsl
Assignmtnts
8.1.1 In Stetion 8.1 it is argued that (f"I'I'varialion in language has a function (Humboldt's principlt).
Campall' this with (Il'e variation in n1lull', which has 00 functlCln (tht form ud coIorofthe
human eyt or the form of ont typt ortll"t). Try to (ormulatt arguments against Humboldt's
principle, one outside linguistics {e.g., (Il't variation in and ont w;thin linguistics (t.g., by
Iooking at homonyms and synonyms).
8.1.1 CoIltct a o(tne concepl ofstyle, dncribe the difTt ll'ncn and tr)' to calegorize
tnem according to tht thrH appn>aehn to style mentlOf\l'd in this 51101\.
'58 8. Style
8.3.1 Form a group wi th 50me fellow students. All members a!l' 10 provide a le.lllle)' have written
Provide normative as...-ell as objtive stylistk criticism of each other's texts.
8. 4.1 Write an advertising te.t williout using adjtives.
Bibl iographical information
8. 1 StylistiOi lIas a rich history, particularly in the fields of rhetork and literary throry. Aristotlc men
tions 1M two style requiremmts appropnalClos and unalLSlomtdnos in Book 3. Chapter I. Quin
tilimus mcntions the style vutucs in books 8 and 11 olms wo", Ihr OnIlors Edu(<lllOn. A good and
acccssibic publicahon on these and othcr das.sical rhetork 'oY'()fk$ is thaI by Kennedy (1994), Lanh:ml
(1991) is handy for I lit$t introduclion to rhetoric.
Reid (1956) inlroduced the term "Kiste . Ccntral figures in oonlemporary register ll:5elrch
arr Biber and Conrad (2001). lhe literature mentioned in thtir article provides a very good over
view. A good starting point for furthtr study of rtgiSler is Ghadessy (1993). Urszula Clark (1996)
provides a practkal introduction 10 stylistic analysis wilh acrei5e material Ind good aplanations
requiring no prior knowledge ortht nessary grammaticalterminology.
II;s only in laltr inlerpretations ofVon Humboldt's (1836) work that he i$ credited with hav
ing introduced a principle. Von Humboldt hlmsdf, however, never spoke or a prindple.
The eumples of Ihe Queneau (1947/1981) slory variations are translations by Barbara Wright.
8.l A good Englishlanguage survey conctrning definition problems is provided l Enkvist (1973).
8.3 Table I is insplred by OveTVICWS in the publication by McMenamin (1993), which is I good
introduction 10 stylistic analysis. Anothtr good starting point for stylt analysis is Crystal and
Davy (1969).
A fine overvicw or research on mttaphon is providtd by Ortony (1993). See Seclion 13.6 for
Ihe eogmlive approach 10 metaphon.
8.04 Weber (1996) gives a good overview of the fidd or stylistlCS in tht last four dades of th( last century,
also tow::hing upon rt(tnl in critical and cognitive 5ty1istics. Semina and Culpt"per
(2002) providc a slarting pOInt for I study of mOf( littrary styli5tics. This coLlectioll of artldes shows
how liteTaT)' analysis ean be in a mQll: oogmllvdy orienttd study of language producuon
and compll:henSlOn.
PART 111
Special modes of communication
9 Conversation analysis
9.1 Transcription systems
Thc study of verbal interaction r[uim a method of wrilten reprtstntahon, a lran-
5cription system, as the regular spelting conventions are not sufficicnt for transcrip-
lion. Intonation, for instancc, can oolr be partially reproducN using punctuation aod
stress marks. Furthermore, it is important 10 know exactly who said whal when. 11 is
also important 10 be ahle 10 register silenee. FOT Ihis reason, various transcription sys-
tems have been dtvclopcd. Thc Iwo most prominent systems are score notation aod
dramaturgiea! notation.
Score notation was insplred by thr wnlten representatlOn of music. A convenation
is morde<! like tones on rnusic staves, with a liDe rtstn'ed for each participant in the
conversation, as in (I) below.
(I) I. R:
[
Pet, weil he almost never rats anything nevu hes hungry
(0,2) . .
J: That Lssurpnsing
2, R: [ Strange, isn'tllr
I: You wouldn't say that by Iooking at hirn. Yes. Not
3, R: [ No, but he/who don't eat very linie
J: that he's fat but Those rnuscles mus! be eoming
((Iaughs
4. R: [ Very, very, strange,
J: (rom somewhtrt,
With score notation it is possible 10 indicate when eaeh partieipant is speaking, where
overlap takts plact, and where silent periods are locate<!. A silence can be indicated by
seconds denoted between brackets. Comments that are necessary for undtrstanding
the conversation can be provided between double braekets. The numbers placed in
front of the Koring bars are meantlo simplify reference. Usually. a length of ILrne per
number o(bars is defined in order 10 denote Ihe length of the entire conversation.
Score notation was deve10ped by the German discourse Konrad Ehlich
and lochern (see. for aample, Ehlich's 1993 publicatlon). lhey called thciT
system a interprelatieve (abbreviated 10 IIIAT). lhe first
,6. 9. ConYeru.tion analysis
parI of the name indicates thaI transcription also entails interpretation. When, for
example. a parlicipant in a conversation stops speaking for a moment aod then later
continues, his words can be regislered as oe turn or as two different turns with a pause
in between Ihem. Sy using Ihe term "Arbeitstranskription", Ehlich and Rehbein made
il elear that it is nol possible 10 UR thc HIAT system 10 produu adefinite transcriplion
in Oe altempt. Ooly by listening 10 a recording of a conversalion again aod again is it
possible 10 gn a precist Iranscription.
Dramaturgical notation was by Gail Jdferson for aample. her
1978 publicalion), ont of the pioneers of conversalion analysis. This form of notation
is ba,sed on the written of stagt: disoourse. Uttl!rances are ordl!red onl!
under another aecordmg to the ordl!r of participation. Whenever possible individual
aelS arl! repTesf!nted on singll! linf!S. Below is thl! same discourse used in fragml!nt (I)
in dramaturgieal notation:
(2) I. IC Pcter, weil he almost nr.ver eals a:nything
2. (0.2)
3. ne:::ver (.) [ He's ntver ]

"
Th.t .:5 surpnsmg
5. You wouldn't say that by loo:king al hirn'"
6. IC :Strange: ISn't II?'"
7.
"
",Yes:: Nol that hl!'S fa:t (.) bu:1
8. (.)
9. R,
No but hel [WhO don', eat (.) vE::ry little ]
10. Iaughs))
1 1. J: Those museles must coming
from 5O::mewhere
12. R: Very/ve: ry stra:nge
Overlap is denoted by a separate symbol, square brackets. The meanings of the other
.symbols are presented bdow.
(3) Symbols in dramaturgic.] notation
no interruption
(al the
and the end of a line)
word correclion
"
extension
..
XX
(.)
(0.2) -
stress point
ullered 10udly
tone rises
ShOft pause
0.2 second pause
In both transcription systems only verbal elements are recorded. Posture al'ld mimiery
of the conversalional partieipants aft not dealt with although ther can influence the
course of an inleraction. For thls reason. video recordmgs aft also used. Of the !WO
9.2 The tum-takina modd 16]
sY5tl!mS, dramaturgieal notation i5 thl! most commonly used. It i5 important 10 Teallu
that any mcthod of transcriplion i5 more or less 5ubjective, not only in the encoding
of pauses and prosody, but also in the interpretation of non verbal aspts and in the
presentation of, for example, half an utleranee as a full turn in a conversation.
9.2 Thl! turn-taki ng model
At first g1ancl!, most conversational activities seern rathl!r chaoti,- One phenoml!non,
however, sel!ms to be eonstant: verbal interaction i5 rl!aliud by turn-ta king. But even
this turn taking can be quite varied In conversation5, there is no limit to the length
of a turn, A turn can vary in length from a single word 10 a complete story. There are
also no rules for determining the order of turns among conversational participants.
Likewise. there are no rules coneerning the number of turns a participant can take or
the possible content of a turn.
Despite the enormous number of variations possible. it i5 rare for silences to
result from participants not knowing whose lurn 11 is. A doser look al conversations
shows Ihat exactly Slmultaneous turn'laking also seldom occurs. In cO\'ersations
there i5 a dear lendency to speak in orderly turns Wlth only one speaker speaking at
any given moment. This lendeney is described in the turn-taking modd developed
by Harvey Sacks, Emanud Schegloff and Gail Jeffel"$On (1974). The model consists of
two components: the turn -construction eomponent and the turn -taking component.
In the first component a turn i5 constructed, built up of syntactical unilS: senlenees,
sentenee fragments, or words. The first point al whieh an as5ignment of turns can
take plaee is at the end ofthe firsl unit. Thi5 point is caIJed the transilion -relevanee
plaee, a possible point of lurn transferal. As soon as such a point is reached, Le., at Ihe
end of every syntactieal unil. the turn-taking component becomes applicable. This
component consists of four rules.
(4) The rules for turn-Iaking
I . For any turn, al the initial transition-relevan place of an milla] turn-
construetional uni!:
a. Iflhe turn-so-far i5 so constructed as to involve Ihe use of a 'curr<,nl
speaker Stlet:ts nexl' technique, then the participant thus selected has
the fight and i5 obliged 10 lake Ihe next turn to speak; no othen have
such rights or obligations, and transfer oceurs atthat place.
b. If the turn-so-far is so construeled a5 not to involve the USt of a
'cunent speaker selects next' lechnique, Ihen selfseleetion for the
neXI speakership may. but need not . be instlluted. The person who
' ..
9. Conwrsalion analysis
first starts al thai moment aequira thc fight 10 a turn, and transfer
occun altha! place.
c. If the turn-so-far is so oonstructed as not 10 involve the use of a ',urren!
sdects next' lechnique. then the ClIrftn! speaker mar, but nted
not continue, unlest anOlher self-sdecl$.
2. Ir, at thr imllallransillon-relevance plaee of an initial turn-construclional
unil, 1. nor Ib has opt'rlltd. and, followm8 the provision of le, the
eurrent speaker has continued, then the rule-set (a) 10 (cl re-applies al the
next transition-relevance place. and mursive:ly at each ensuing transition-
rtlevance plaet. unlillransfer is effected.
The following example illustrales how Ihis turn-taking model works.
(5)
WI'U (eh) pmty actually UUSC 1(1 really Icarned
I. A, [
B how du;! tht rxam go yesttrday!
"
2. : ,' [IM 'tuff you know but whm I was J,ltlln& "' thatlectun: haU (eh) 1 JU5t cou.ldn't 'OrM
" 3. A; (1oI)ynh,lwas
B: you BOt a blKkout
" 4. A:. [Ir)'lo, 10 oon,cntrale. bul , ould only Ihink of not ,omin, up wilh proper lswers
.,
"
5. A, [
B: weil. you $houldn't worry abouL it 100 mudl now
C; yeah no. indced
In (5.l) Speaker B chooscs Aas Ihc subscquent speaker according 10 rule la. Speaker
A contlOues untiI, after a moment of silenee, B takes a turn in (5.3) following role Ib
or role Ic. After the silenee Ihal then folIows. rule Je becomes applicable.
A number of objeclions have been raised againsl this model. Firsl of aII, in the
analysis of eonversations it is often impossible 10 say which rule applies. Take the fol-
lowing example:
(6) A: Andy just paid me back my fifty bucks.
B: Greal. You guys wanla go out ror pizza?
C: Hey A. Can you !oan me ten dollars?
9.2 The tum-taking model .65
In this phase of,he conversatlon, I1 is not dear whether C is reacling to Xs words or
is ignoring Band is jusl taking a turn. In other words, it is not possible to ascertain if
C is getting a turn according 10 rule la (eurrent speaker chooses subsequenl speaker)
or rule Ib (a conversational parlicipant tak6 a turn when no subsequenl speaker is
chosen). Determining which role is applicable has turned out to be more difficult than
the model suggnted
Secondly, it is assumed in Ihe model thai conversational participants can recog-
nize a construction unit. This may be true for queslions and answers, but m many
utterances it is undear where the possible poinls of turn assignrnent are. Moreover, it
is possible for a speaker to neutralize th6e pomts by beginning a turn with aremark
such as: are two pomts that I would Hke to make dear ..... Formally spcaking,
the roles for lurn assignment become effeclive at the end of the first point. The content.
however, indicates that this point docs not demarcate the end of the turn.
Thirdly, conversalions do not eonsist solelyoflurns but include remarks irrelevanl
10 Ihe flow of Ihe conversation such a5: - Would you like something to drink?", "Can I
have a handkerchieP.", eie. Morcover, eonversational participants who do nOI "have Ihe
floor" often volee their involvemenl with such utterane6 as hm, rraUy?, weil, weil, eie.
This type of utterance i5 clas.sified as backchannel behavior or col/a/era1 commu"ica
lioIJ. The lurn-taking model does nOI make dear how the diSlinetion is made between
turns, on the one hand, and ancillary remarks or back-channel behavior that docs not
trigger the rol6 of assignmenl. on the other.
The lurn laking model prompls thoughl on the qU6tion of what exaclly a turn
iso Is baek channel behavior, the hm made by speaken or Ihe um by which lisleners
indicate a wish 10 speak, also a turn? Ifit is appropriate to speakof a turn when a par-
ticipant takes the floor. then these minimal reaClions will nol qualify as lurns. Instead,
another's turn is supported by hstener reaetions such as "How about "You ean
say thai again" or similar utterances. The speaker can, after such areaction, sLmply
eonti nue with his turn. If, however, the turn applieation um is sn as a complete lurn,
then the rules of the turn taking model do not always apply. InleresIingly, a silenee
ean also be sn a5 a turn. Participants can, by remaining silent, answer a qU61ion or
agr 10 a requesl.
Obviously, it is tOO simplistic to speak only of a turn when participanls become
the main speaker, but it is equally wrong to view cvery ulteranee, no malter how
minimal, as a turn. One solution IstO view backchannel behavior as a pre turn wilh
whLch partLcipants make it c1ear thai they want a turn, jusl as thc so-called tnbreathe
indicat6 thai participants want 10 say something. Thai a silence can sometimes also
constitute a turn can be explained by the assumption that positions can be filled by a
verbal reaction. or a - nulr verbal reaelion.
,66 9- CooversatiOfl analysis
9.) organiutio
A conversational sequence is a systcOlatic succession of turns. In Ihe analysis of
sequences the fOCU5 l1as primarily on the adjaceney pair. This term refers 10 thc
phenomenon thai , In a conversalIOn, Ot uneranct has a role in determining the sub-
sequent uUerance Of at least in raismg upectallOns concerning its oonlenl5. Bdow is
an example of the adjacency pair -question. answer':
(7) A: How da you likc
B: (0.3) Weil, what can T say?
Schegloff (1977) points out thaI in an adjacency pair, the second utlerance is condi-
tionally relevant': This means that ifthere is an adjacency pair consisting of parts A and
8, and part A has 1>n uttered, then part B is expted And, when 8 has 1>n uttered,
then il is viewed by the participants as being relevant to A. 8 is therefore relevant on
the condition that A has been uttercd. If 8 dors not occur, then this is not random but
a significant or absence and conclusions can be drawn from this. 80th
possibilities can be in the following example.
(8) A: Would you likr to SO and .. . uh ... get same
B: (2.0)
A: Or aren't you in thr mood?
8: ( 1.5) Whatdoyoumtan?
1\5 first utterance creates eXptations of areaction. Questions are, after alt, usually
followed by an answer. It is for th is reason that 8's silence is not viewed by A as being
random. J\s second ulterance is areaction to an observably absent answer. 8'5 second
utterance is conditionally relevant to this reaction by A, A's question makes 8's utter-
ance relevant, that is, interpretable as an ans wer in the form of arequest for more
precise information.
In fact, the designation adJautlcy pair is nottotally correct. The parts of a pair are
often not adjacent. In the following example.the opening question and the answer to
Ihis question are separated by another question-and-answer pair.
(9) A: Can you tell me how 10 gel to the mall?
B: Do you sec thal big neon sign?
A: Yes.
B: You have to make a left turn there.
The adjacency pair is an important buildmg block in conversation. Ihe adja-
cency pair, a thrre-part sequence also occurs. Below are same examplrs.
9] Sequenllala.ganlution ,67
(10) A: Wtll, Paul, can YOli corne up and find AUSll1lha on thc
B: That's here, I guess.
A: Indeed, 10ll are right.
(li ) A: How about having a drink
B: Yt'S, good ideal
A: O.K., Ihere is a taxi.
In in-dass interaction, such as in ( 10), the teacher often asks a quest ion and commenlS
on the answer given by the pupil This thrre-part sequence is called a question. answer.
evaluation chain. And if sameone proposessomethingas in (I I), the positive react lon
is usually foJlowed bya suggestion for further action. This sequence is called the offer-
agreement-affirmation chain.
An answer to a quC:S(ion is oftm foUowed by a comment as in thr following example.
(12) A: Can you tell mr how 10 get 10 the mall?
B: Thrn right at Ihe third light.
A: Terrilic, thank )'ou.
Research on the SYSlematic sequencing of turns concenlrated on Ihe beginnings and
Ihe ends of conversalions. This approach started with a study by Scheglotf and Sacks
(1973) in which Ihe techniques used by participants 10 reach a point at which Ihe
conversation can be dosed were inventoried In every conversation there is a point at
whieh Ihe condusion of one turn no longer leads to a subsequent turn and the silence
that follows cannot be interpreted as the silence of one of the participants. Scheglotf
and Sacks analyzed a large number of telephone conversations and found that many
of the conversalions ended with the following dosing pair.
(13) A: Oka)'?
B: Alright.
Should 8 not want 10 end the conversation, then the possibility WSIS for B to continue
after A's ulterance. However, if B fiUs in the second part of the dOSLg pai r with an
affirmation of the first parI, then Ihe conversation is rssenliall )' over (except possibly
for mutual grtetings). What is is that a pair Iike the one abcl\'e can also
occur in the middle of a conversation. Speaker A can, following 8'5 reaction, contmue
with a new topic. Apparently, changing phrases such as only serve 10 end a
conversalion if there is nothing left to discuss.
How do conversalional participants know that Ihere is nothing left 10 discuss?
When reviewing their material. Schrgioff and Sacks found that topics were usually
ended wi lh words such asgood, okay, and wtll, pronounced with a declining intonation
,68 g. Corwerution 3nalysis
whieh the started a oew lopie. Their analysis showed thaI these types of
lopie dosing are also used as a way of suggesting the end of a conversation. Below is
an example taken (rom Schegloff aod Sacks' material in .....hieh thc word occurs
Ihree limes. The first Moka( is a lopie dosing aod therehy a possible pre-announce-
ment of a convtrsalion closing. Thc .second Mokay" serves as an announcemenl or
drclanuion of inten! 10 end the convcrsalion. The third "okay" serves as a sign of
agreement with Ihis closing.
( 14) 1'.:
S,
A,
S,
A,
S,
A,
S,
... and uh, uh, we're gonna see if "'"C can', uh tie in our plans a linie better.
Okay 11 fine
Alright?
Righl.
O"'yboy.
O"'y
Byellbye
G'mghl.
Obviously, multiple fun(lions ean bt combined in one In the following excerpl
it ean be seen that "okay" serves as bolh a topie dosing aod as a dedaralion of inlenl
10 dose the conversalion.
(15) (A has eall! 10 invile B. bul has bccn lold B is going OUllo dinner.)
A: Yeah. Wcll gel on YOUT dothd and get out and colleet some of Ihat frcc food
and wc'lI make il some other lime Judy Ihen.
B: Okay then lack.
1'.: Bye bye.
B: Byc byc.
Schegloff and Sacks'$ analysis shows that analyzing a number of lums containing
is insufficienllo make il dear why a double exchange can be followcd by
a dosing in Ihe form of a greeling and relum ofthe gruting. foT Ihe analysis of a turn
or a pair of turns, it is necessary to look at Ihe conlexl wilhin which it occurs.
9.4 Discourse markers
In the sludy of wrillen discourse Ihe focus of many researchers is on conneclh'es,
which link the various segments; see Section 6.2 and 6.3. In the study of spoken
discourse many analysts focus on discourse markers or pragmalic parlides, such as
okay in the last example of Ihe previous section. Discourse nlarkers have as their main
1
9.4 Oiscourse markerti ,6g
funclions marking something in Ihe struelure and indicaling some aspects of anitude.
Bcloware same examples (tbe italiciud words).
(16) A: I Ihink I will stay horne. I fecl /rke I ran half a maralhon.
B: ....nd yeslerday you said you would (ome!
(17) A: But Ilold you nOI to open the door, not for anybody!
S, Weil, I do ha\'e my own WIll, y'know.
(18) A: So, in Ihe end you have decid! to join os then.
S, After all, I had to bc here anyway.
Discourse markers Ihal are used: connectives (like antI). adverbials (like anyway),
inlerjcctions (Hke wtlf), prepositional phrases (Hke after alf) and lexical phra.ses (like
y'/alOw). As wilh mosl important concepts, definitions of discoursc markers given in
tbe lilerature vary, depending on tbe theorelical approach. However, most describe
discoursc markers as signaling devices outside the propositlonal conlenl. indicaling
the expressive funetion of a piece of discourse. The expressive function denotes the
attilude of the speaker toward the locution (su Section 4.1). The piece of discourse
can be, for example, a turn in a conversa,tlOn or a topie. Because discourse markers are
not apart of the propositional conteni, Ihey are mostly found at the beginmng or Ihe
end of an ulteran. DiscouTSe markers are usually distinguishcd from connectives in
that connectives wign all kinds of semantic and pragmallc funcllOns 10 paragraphs.
dauses and subclauses, while discourse markers only indicate the allitude of a speaker
(or possibly a writer), mostly marking a turn or a topie.
The Iwo markers in example (16) illustrate the attitude approach combined with
the function of a turn. The discourse marker like indicates looseness. The speaker
expresses that the of exhaustion docs nol have to be taken precisely as it is
formulated. This can be seen as an altempt by the addresser 10 enhance the of
relevance of the turn's contenl for the addressee, who is now alJowed or obligcd to also
take inlo consideralion a more kind oftirednes5. The .... nd at the beginning
OfS'slurn does not indicale a connection such as in "apples and expresses
the indignation of the whole turn, which with Ihis marker becomes more subjective
than a formal reproach.
DiSCOUTSe markers have been studled in various ways 10 detect their usage and
functions in different communication silualions. In cO\ersat ions. partieipanIs use
discourse markers nOI only 10 express attitudes, but also to delect or to confirm which
information is given and which is new (see Scction 7. 4 about given- new manage
ment). Discourse markers. in other words, also mark the presence or absence of
common ground (see Section 3.3). This has been nicely demonstratcd in a study by
Jucker and Smith (1998). They asked their sludents 10 have conversations aboul lopics
'7"
g. Converution Jrl31ysis
likt sports, travtl, opera eie., and dlvidcd the group ioto pairs of students who ,",'cre
friends and pairs who were strangen. In total they transcribed thrcc and a halfhours
of conversation by 15 participants. They counted aod liste<! the discourse marken and
found almost 3,000, which means one about every live seeonds. Hefe is one passage
from their transcriptions with the discourse markers ilalicized:
(19) A: I ptay basketball
B: Oh ytah what position ( .. )
A: yeah
B: that's cool
A; JUs: (,.) juS! playing with thc friends you know
With thc reaction yeah
ft
B does not conllrm the information that A givts. aslhe
yeah- in A's second turn does. 8 thai the information has bttn rcceived
aod can with other information available at thai point in thc conversallon.
The discourse marker"you know" does nol remind B ofknowledge a1ready given, hul
presents more or less an invilation 10 make Ihe fight inferences for Ihe assignmenl of
relevance to an ullerance, fot example that person A is not a real top sportsman.
lucker and Smith divided the discourse markers into markers thai serve as indica
lors ofinformatlon reception, like )'tah. and markers that are used to present informa-
lion, like you know. They counled these markers in Ihe conversations betwun friends
and strangers, and found some remarkable differences in Iheir material (Table I).
The most frequenl reception marker is ytah. However, betwun friends it i5 sig-
nificantly leu frequent (every 18 seconds) than between strangers (every 13 seconds).
This can be explained by the nature of Ihe conversation. Between strangers it is more
necessary 10 indicate that Ihe information has been received than between friends,
who have more common ground (see also Section 3.3) based on shared experiences.
TIMc I . Frequcncy of rcccpllon marUl"$ and prescntation markcn (Ivrragc.' number of
toktns per mlOutc.') in COnvtl'SallOnS bttw:n pairs of strangc.'rs and pairs offric.'nds
Marktr Strangfl't Friendt
RKeption marktn
Y''"'
'.5 3.'
.. 1. 1.1
rtGU, 0.3 0.'
Prescntation markers
/," 2.' '.5
)'Oll krrow 1.0 1.'
,
9.4 Discourw markers ' 7'
The analysis ofthe conversations also indicated thai there is a difference in information
reception between )'tah. oll and really. Compare Ihe following examples.
(20) A: J likr playing basketball
8, Yeah, ...
(21) A: I hke playing basketball
B: Oh, ...
(22) A: llike playing basketball
8,
These thru markers indieate a difference in the ease of integration of new informa-
lion. Yeah indicates that Ihe integratIon ptocess is "ety easy, oh marks that some extra
processing effort is needed, and reallY5uggesis that more information is nmed befote
integralion an be successfuL
The ptesentation markers also showed a siriking difference. Ftiends use Ihf'm more
OMen than strangers. This can be explained hy thf' same factor of common ground. If
there is more shared experience (bf'tween friends), there is a belter basis for provid.
ing indicators about how to process information (for uample, 10u know can be taken
literally betwun friends). Hencf', presenlation markers like those mf'ntioned above,
and others Iike wdl and I mean are more frequf'nt in cQnvf'rsalions betwun friends
than hetwun strangers. \Vith this study Jucker and Smith nicely show that dlscourse
markf'fS are not only indicators of attitude hut also signposts fot the exchange of infor-
mation based on an important characteristic of communication: common ground.
",
9. Comotru.lion analysis
Questions and assignments
Questions
9_',1 In this IWO thalll'e igJ\Of'ed in transcriplion systrrm bul In!
important for lhe CO\IIW of a Name one olMr phenomenon.
9.1.1 Demonstrate that, besides Ihe turntaking model, Ihere is 11 separate ruh: flecessary fortllt
dosing of a COIMBllion.
9.1.::1 Give an rumple ora COOYersational txeerpt in which a silenceon lhe part afa p.articip.ant must be
interpreltd u a turn, lIking ;nlO ltooUfll thc fifth OrITM (se-t stion 2.6).
9.1.3 Indinle wby beh<lVlornnoot be SfI as I pore-turn.
9.].1 Give e.amples of other type'S of adjac:ency pairs besidn que5lion-answer.
9.3. 1 uplain wfIy telephone CorMrutions an! often Ihe focus of analysis of votrb.tl interaction.
9_4.1 Provide uamples OfOIMrdiS(ou!"W markers besidH those given ill Section 9.4.
9 ... 1 Tryto point out 1M d,fkrence In muning in eumpln ('7)'00 (8) when the discoue is
not uSf:d.
9.4.) In a conYerwtion betVOftn str.angers it is l1tSury to Ust indic:OiltOl'l in onleno enhance the
pro<:eSsingofinfonnOiltlOll. Ollt would then tlIptCt pttStntation mar1(ers betwttn str1lmgen
tooccur mort frtqutntly, bpiain why this doei not show in, fOf"uample, thC' rncan:h dollt by
Jucktl' and Smith.
Ass ignments
9. ' ,' Make a tapt rt'COf"ding of an t"o'trydayoon...el"ulicm betVOftn Ihlft people, a eon-
Yersation in a living room 01' a vi511 by fritnd5. Try 10 transcribe two mmutts ofthis conYersation
u5mg dn.mOilturgic:al notation.
9 l .' A siitnce Cln be difTertntially usignt<! on t hC' basis o(thC' rules ror tumlaking. Silence can be
Sn.u ellhC'r. I. a glp btfott I substquent Ippiicalion of rulf: .bO( 'C; J. a lapse ofthe non
ap-pllcatlOll ofrult II"b or 1(; 01'). I signineant siltnce ofthC' MXt selecttd 'ptaktr .Ifltf"lhC'
applintion ofrult la.
1
QutstlOllS Ind usignments ' n
Try 10 find eumpln of eac:h of thC'se thlft typn of siltnce In ordtf" to Yerify whtthtr thC'
IKm worts. Stlrt, fortumpIe, by lookmgat lhC'eumpln used in (1m).
9.).1 In the pan allempIS haYe bttl"l made 10 desigl"l computer programs thai simulate human I,n
luage behaviOl". 11"1 tx!l1!mt cun, software "boiS wert crtattd 10 bel'laYe 11 ifthty wert human
inlerlocutOl'lln inten.ctive wrillen discourse. [lilI, ereilt<! in .g66 by JOStph WtizenbllUm, is
such a eomputtf" program.1t rtformulatts optlOltor answers in orderto kttp a conYersll1Oll going
and.u such simulales psycholnalytic:OilI IMrapy.
Find an Elill challerbot on Ihe Inlerntl and start a COl1'o'tfS.llion. SaYe the l ranscripl of your
convtrSation.lnd tryto point oul the successfuladjac:ency pairs and Iht mort eurious turns.. Try
10 txpllin how [ lill worlu out how to rtSpond 10 your lums. Which turns Art Iht mosl diffi<:u ll for
fliza and why would thai bt Ihe easel
9.4.' Find a corpus (500 10 ',000 words) of formaland inlOrmal corwersation in yourOWfl Ianguagt and
COIJl"lt Iht l"Iumbtt-of disooul"St markers il"l il. Tryto find difftrtnces il"l frtqutllC)' betwttl"l formal and
il"lformal cOlMrulion.
174 9. Conversation analysis
, ..
, ..
,.,
, ..
Bibliographical information
For more on tM HLAT wt EhLch (1993). in whkh a more drtailed aUed HLAT LI IS
suggested. See (1978) for aamp]ts or dl'illfTlilltursiu]llOtalJOn. For an introdu( -
uon to C()IlVff$;Ition ilfIalysiJ wt Ten (1999) . Set: also Hutchby and WoofIiu (1998). llus pub
lieallon demonstnLttJ the practical relcvance of conversalLon analYSI$, for by diliClSLng
LU mc:rllS in the anatysis
Edwank (2001 ) .Iso prov!dts I good ovtrv>cw of the VUIOW tntnKnp!1OII methods and
IIOns thc,r (dls)iIId .... ntagts. After In introductory WQrk to analysIS {sec Chapter I}
Egins and Sladc (1991) KI'Vts U a good further acquamtance. The founding father or (onver-
anaI)"LS IS Ihf IOCloIosLit Sacks, aod hLi lcctura wen pubhslltd by Itlferson; 5
Ind (1992).
tUTO-tumg modd was pubhshed m.n .mc!e A. S4mplnt SystCIMflOjor flK T'"1U1-
Ilon ojTurn-lakingjor ConvtrSation; Set Sacks. Schegloff and (1914). Set: McLaughlin
(1934) for cnheism ofthlS; aampk (6) wu tmn frorn tJus publication.
lnslead of lurn. the 1$ oftcn used. ThLi is. mtlllg term U 1II;"'flMo fIoor LI yotIl"$."
IS iU yel no cku definition. For the dlshnction be!Wetn turn and position, sec
(1983). For furtMr iludy.sec also kark (1992), who cntKlus thc turn-talong modd.
Jaworslo (1991) is an ac:dlmt colleclion of papen on silence in, among othcT things. verbal
intcraction and narratives.
For notIOn adjaWICY fHlIT, see Schegloff (19n). For on thrtt -part see
Edmoodson (1981) . Set: Ochs, Schcgl.off and Thompson (1996) for a coUechon of papus on mter -
actJOllai ULU and turn orgamUtlOn.
anal)"ls h:u very specific roots 111 tthnomethodology. For I good intmduction,
see NofsLger (1991).
Schlffrin's dIssertatIon f)UCOl,lru Manrn 1$ il bndmark publiution in thlS field of sWdy. Ind was
pubhshe<! u a book In 1987. Sch,ffnn taltes dbcourse markers 10 be rrwkert offour different
typcs or knowle<!V apresslvt, $(Kial and tatu.aI. ln tJus book. only
aprcs,s,ye and tUtuaJ types are dlscusstd. BOOd monograph u Dnnton (1996). An
ucdlent startLg POLOt for ItiHrch L$ tne coIlKtion of arlKa by Jucker:and ZJv (199lI ).
study by Jucker aod Smlth (1998) that 1$ to In this sectlon can also be fOllnd in tJus
publica\lon The table LO thlS sechon was from thelr artlde. Sm'th and lucker ,Iso contrib-
ute<! an aTllde to Fdur ilnd (2002) In whloCh!hey scruhmu discourse muur weil.
For more on IM discourse marker oUy, 5 Heuler (19%).
For mort informatIon on the cognltive aspcct of common ground that Jucker and Sm,th
sec Clark', (1996) lmportilflt publia tlOn on thIS IUb/tCt, which waI.1so
111 thll book.
,
10 Informative discourse
10.1 Readability in a formula
lt is often said that discoursc serves main funetions. In discourse the
wants 10 inform the address, or his own opinions and and by doing
50 he wants to persuade the addressee to adopt a change in attitude or action, or both
(sec $echon 4.1 abaut this tripartition in information, exprtsSion and argumentation).
Chapter 11 focuses on narratives, the most suitable discoursc mode for exprts.sion.
Chapter J 2 deals with argumentation and persuasion. This chapter is about informa-
tive discoursc, which is mainly referrcd to as expository or explanatory discourse.
One of the oldest problems in research concerning discourse thai tS meant to
inform Ihe is how to delermine whelher or nOI a discourse is comprehen-
sible 10 a specific target group. In the last century researchers tried 10 50Ive this
problem by de"e1oping readability formulas. Some of these formulas are still popular
in the field of educalion, where there is a great need for an emdenl method of deter-
mining whether a textbook is suitable for a given scholutic level. The procedure for
developing a readability fonnula consisis of four steps.
I. The first step is to colleel a number oftexts thai are Irnown 10 have different levels
of difficuhy. This colleetion can consist oftexts that have bet'n used in schooltests,
and which have known score levels for each educalionallevel. A lexi can then be
aMumed to be suitable for a certain level if pupils at that level get a given avt'rage
score, for eumple, seven out of ten correet answers.
2. The lexts can Ihen be analyzed for all the possible characteristics which may have
an influence on readabililY: the length of words, the percentage of abstract words,
Ihe number of subordinate dauses per sentence, Ihe number of preposttions ]X'r
hund red words, etc. for example, Ihe style analysis done by John Carro))
discussed in Seclion 8.4.)
3. StatlSltCal proce.ssing can then aid in deternllning to what exlent the dlfferences in
difficulty (see step 1) can be ascribed 10 characteristics in the texts (see step 2).
4. On the basis of step 3, it can be delermined which text characlerislics contribute
the mosl to the outcome of a reading comprehension test . Further statislical analy-
sis can be done 10 determine how often these characteristics should or may occur
in order for a lexl 10 be reiidable for a certain le,el.

S-ar putea să vă placă și