VAN DIJK-KINTSCH - Strategies of Discourse Comprehension

7/26/2019 VAN DIJK-KINTSCH - Strategies of Discourse Comprehension

1/195

Strategies of Discourse Comprehension

Teun A. van Dijk

Walter Kintsch

1983

New York: Academic Press

GUATEMALA: NO CHOICES*

Compared with the relative shades of gray in El Salvador, Guatemala is a,study in black and white. On the left is a collection of extreme Marxist-Leninist

groups led by what one diplomat calls a pretty faceless bunch of people. Onthe right is an entrenched elite that has dominated Central Americas mostpopulous country since a CIA-backed coup deposed the reformist governmentof Col. Jacobo Arbenz Guzmn in 1954. Moderates of the political center.embattled but alive in E1 Salvador, have virtually disappeared in Guatemala-

joining more than 30.000 victims of terror over the last tifteen vears. Thesituation in Guatemala is much more serious than in EI Salvador, declares oneLatin American diplomat. The oligarchy is that much more reactionary. and thechoices are far fewer.

Zero: The Guatemalan oligarchs hated Jimmy Carter for cutting off U.S.military aid in 1977 to protest human-rights abuses-and the right-wingers hired

marimba bands and set off firecrackers on the night Ronald Reagan was elected.They considered Reagan an ideological kinsman and believed they had a specialfriend in White House aide Michael Deaver, whose former PR firm hadrepresented a Guatemalan businessmens group, Los Amigos del Pas (Friendsof the Country). But after a year of Reagan, the Guatemalans have beendisappointed. If Reagans team has proved friendlier than Carters, the persistentU.S. demands for political moderation continue to grate on the Guatemalans. Asone diplomat in the Guatemalan capital puts it. U.S. leverage on the regime iszero.

Cold Shoulder: The Guatemalans have snubbed visiting U.S. congressmen.calling some of them communists. They have even given the cold shoulder to

Reagans special emissary, Gen. Vernon Walters, who visited Guatemala twicelast year. On the first trip Guatemalas President Romeo Lucas Garcia finallyagreed to receive Walters at the last minute. On the second visit the Presidents

plane was preparing for takeoff just as Walterss landed. They feel they arewaging our war in Central America, and were not helping them. says oneAmerican in Guatemala. They say, Wed rather do it with you-but with you orwithout you-well do it.

The government is certainly making every effort to do it. Guatemala hasacquired Brazilian armored vehicles and an array of arms and equipment fromFrance, Yugoslavia, South Korea and Romania. But the principal ,ourre ofweapons for the Lucas Garcia government appears to be Israel. The Guatemalan

Army is equipped with everything from Israeli-made cartridge belts and helmetsto Galil assault rifles, Uzi submachine guns and araw transport planes. TheIsraelis are also reported to be training radio

*Copyright 1982 by Newsweek, Inc. All rights reserved, Reprinted by permissionfromNewsweekMarch 1, 1982, p. 16.

1/195


2/195

techniciansatarecentlyopenedmilitary-communicationsschoolinGuatemala.

AndnowthattheReaganAdministrationhasloosenedJimmyCartersmilitary

embargo,theGuatemalanArmypatrolsthecountrysideinAmerican-made

trucks,jeepsandhelicopters.

Still,ReagancanhardlymendrelationswithaGuatemalanGovernmentthat

seemstohavemademurderanofficialpolicy.Morethan13,000peoplehave

beenkilledsince1978,andtheStateDepartmentestimatesthat300moreare

murderedeachmonth.Themorningpapersarefullofreportedkillingsby

desconocidos(unknowns).Mostmurdersseemtobetheworkofright-wing

deathsquadsliketheSecretAnti-CommunistArmy,widelyassumedtobe

covertagentsofthegovernment.AU.S.missionary,JohnArnoldMiller,was

killedtwoweeksago,theseventhclericmurderedinfourteenmonths.AndinthevillageofUspantanlastweek53peasantswereroundedupandbeheaded.

Thegovernmentblamedguerrillas,buttheevidencewasskimpy.Thereareno

politicalprisonersinGuatemala,formerVicePresidentFranciscoVillagran

Krameronceremarked.Onlypoliticalmurders.

Hydra:Thewholesalekillinghasnotyetdauntedtheregimesguerrilla

opponents.EarlierthismonththefourprincipalguerrillagroupsannouncedtheyhadjoinedinaGuatemalanNationalRevolutionaryUnity.Withinhoursof

thedeclarationGuatemalaCitywasalmostblackedoutbyaseriesofbomb

blasts-evidencethatunificationmayworktotheadvantageoftheestimated

3,000to5,000guerrillas.Theymayalsobebolsteredbyassistanceandtraining

fromCommunist-bloccountries.Lastsummerthegovernmentdiscoveredmorethan25guerrillasafehousesstockedwithChineseandSoviet-blocweapons,

includingacacheofU.S.M-16rifleswhoseserialnumbersmatchedthoseleft

behindinVietnam.In1981therewere383reportedclasheswithguerrillas-

almostfourtimesasmanyastheyearbefore.Wehavebeenabletopacify

someregions,saysCol.JaimeRabanalesReyes.Buttheguerrillasarelikea

hydra-theirheadsalwaysshowupinsomeotherplace.

ThechoicesfacingtheUnitedStatesinGuatemalaarefewandunpleasant.

Washingtoncanhardlyhopeforacentristpoliticalsolution:thesameChristian

DemocratstheUnitedStatessupportsinElSalvadorcontinuetobedecimated

inGuatemala.Inthelasteighteenmonths238ChristianDemocraticleaders

havedisappeared.Somediplomatsthinktheextremerightisdestroyingthemoderatecenterpreciselytoprecludethesortofreformistsolutionthatthe

UnitedStatesadvocatesinElSalvador.Ultimatelythepolarizationservesthe

causeoftheleftistrevolutionaries.Thegrowingstrengthoftheleftconceivably

couldpersuadetheUnitedStatestobackyetanotherCentralAmerican

dictatorship-thisone,inGuatemala,themostbrutalofthemall.

LARRYROHTERinGuatemalaCity

StrategiesofDiscourseComprehesion

2/195


3/195

Strategies ofDiscourse Comprehension

Teun A. van DijkDepartment of General Literary StudiesSection of Discourse StudiesUniversity of AmsterdamAmsterdam, The Netherlands

Walter KihtschDepartment of PsychologyUnivcrsity of ColoradoBoulder, Colorado

1983

ACADEMIC PRESSA Subsidiary of Harcourt Brace Jovanovich, Publishers

New York LondonParis San Diego San Francisco Sao Paulo Svdnev Tokyo Toronto

COPYRIGHT 1983, BY ACADEMIC PRESS, INC. ALL RIGHTS RESERVED.NO PART OF THIS PUBLICATION MAY BE REPRODUCED OR TRANSMITTEDIN ANY FORM OR BY ANY MEANS, ELECTRONIC OR MECHANICAL,INCLUDING PHOTOCOPY, RECORDING, OR ANY INFORMATION STORAGE

AND RETRIEVAL SYSTEM, WITHOUT PERMISSION IN WRITING FROM THEPUBLISHER.

ACADEMIC PRESS, INC.111 Fifth Avenue, New York, New York 10003

United Kingdom Edition published by

ACADEMIC PRESS, INC. (LONDON) LTD.24/28 Oval Road, London NWl 7DX

Library of Congress Cataloging in Publication Data

Dijk. Teun Adrianus van, DateStrategies of discourse comprehension.

Bibliography: p.

Includes index,1 . Discourse analysis--Psychological aspects.

2. Comprehension. 1. Kintsch, Walter, 1932-II. Title.P302.D472 1983 401.41 82-22671ISBN 0--12--712050 --5

PRINTED IN THE UNITED STATES OF AMERICA

83 84 85 86 9 8 7 6 5 4 3 2 1

3/195


4/195

Contents

Prefaceix

Chapter1TowardaModelofStrategicDiscourseProcessing

1.1.THESTUDYOFDISCOURSE11.2.BASICASSUMPTIONS41.3.ANOVERVIEWOFTHEMODEL101.4.CONCLUSIONS19

Chapter2

ObservationsontheStatusofExperimental

ResearchonDiscourseComprehension

2.1.PERCEPTUALPROCESSES:LETTERANDWORDIDENTIFICATION212.2.SYNTACTIC-SEMANTICPARSING272.3.AMBIGUITY332.4.SEMANTICUNITS:PROPOSITIONS372.5.COHERENCE432.6.KNOWLEDGESTRUCTURES462.7.INFERENCES492.8.SCHEMATICSUPERSTRUCTURES522.9.OUTLOOK59

Chapter3

TheNotionofStrategyinLanguageandDiscourseUnderstanding

3.1INTRODUCTION61

3.2THENOTIONOFSTRATEGY623.3COGNITIVESTRATEGIES683.4LANGUAJESTRATEGIES703.5GRAMMATICALSTRATEGIES73

3.6DISCOURSESTRATEGIES783.7THEREPRESENTATIONOFSTRATEGIES953.8ASAMPLEANALYSIS:THENESWSWEEKTEXT98

Chapter4

PropositionalStrategies

4.1.PROPOSITIONS1094.2.COMPLEXPROPOSITIONS1194.3.PROPOSITIONSASCOGNITIVEUNITSANDASSTRATEGIES1244.4.ASAMPLEANALYSISOFTHENEWSWEEKTEXT1344.5.EXPERIMENT1:PROPOSITIONFUSION144

Chapter5

LocalCoherenceStrategies

5.1.LOCALCOHERENCE149

5.2.UNDERSTANDINGSENTENCES1515.3.LOCALCOHERENCESTRATEGIES1545.4.TOPICALITYANDSTRATEGICPRONOUNUNDERSTANDING1605.5.EXPERIMENT2:PRONOUNIDENTIFICATION1725.6.ANALYSISOFTHENEWSWEEKTEXT182

Chapter6

Macrostrategies

6.1.THENOTIONOFMACROSTRUCTURE1896.2.MACROSTRUCTURESANDDISCOURSEUNDERSTANDING191 6.3.CONTEXTUALMACROSTRATEGIES196

6.4.TEXTUALMACROSTRATEGIES2016.5.MACROSTRATEGIESINACTION:SOMEEXAMPLES2066.6.EXPERIMENT3:INTERESTANDLEVELOFDESCRIPTION

ASMACROSTRUCTURECUES222

6.7.EXPERIMENT4:PRIMINGMACROPROPOSITIONS226

Chapter7

SchematicStrategies

7.1.SUPERSTRUCTURES235

7.2.SCHEMATICSTRATEGIES2377.3.SCHEMATICANALYSISOFTHENEWSWEEKTEXT242

4/195


5/195

7.4. PREDICTIONS AND IMPLICATIOS OF THE THEORY OFSCHEMATIC SUPERSTRUCTURES 251

7.5. EXPERIMENT 5: THE ROLE OF RHETORICAL STRUCTURE INDESCRIPTIVE TEXT 253

Chapter 8

Production Strategies

8.1. PROBLEMS OF A PRODUCTION MODEL POR DISCOURSE 2618.2. INTERACTIVE AND PROGRAMATIC PRODUCTION STRATEGIES 2648.3. SEMANTIC PRODUCTIONS STRATEGIES 2728.4. STRATEGIES FOR ESTABLISHING LOCAL COHERENCE 2788.5. PROPOSITIONAL PRODUCTION STRATEGIES 2808.6. SOME CONSEQUENCES FOR SENTENCE PRODUCTION STRATEGIES 2838.7. SOME PRODUCTION STRATEGIES FOR THE NEWSWEEK TEXT 2868.8. TOWARD A PROCESS MODEL OF DISCOURSE PRODUCTION 293

Chapter 9

Strategies for the Use of Knowledge

9.1. KNOWLEDGE SYSTEMS 3049. 2. KNOWLEDGE USE IN METAP HOR COMP REHENSION 3129.3. REDINTEGRATION, REMINDING, AND PARTIAL MATCHES 3159. 4. STRATE GIES FOR KNOWLEDGE USE IN DIS COURSE 3179.5. KNOWLEDGE USE IN THE NEWSWEEKTEXT 3199. 6. EXP ERIMENT 6: GLOB AL BIASES IN KNOWLEDGE 324

APPENDIX: TEXTS USED IN EXPERIMENT 6A 331

Chapter 10The Cognitive Model

10.1. FROM THE TEXT REPRESENTATION TO THE SITUATION MODEL 33610.2. A FRAMEWORK FOR A PROCESS MODEL 34610.3. SHORT-TERM MEMORY USE IN DISCOURSE PROCESSING 35210.4. RETRIEVAL FROM EPISODIC TEXT MEMORY 35610.5. A PROCESSING MODEL FOR UNDERSTANDING AND SOLVING WORD

ARITHMETIC PROBLEMS 36410.6. EPILOGUE 383

References 387

Author Index 405Subject Index 413

Contents

Preface ix

Chapter 1

Toward a Model of Strategic Discourse Processing

1.1.1.2.1.3.1.4.

THE STUDY OF DISCOURSEBASIC ASSUMPTIONSAN OVERVIEW OF THE MODELCONCLUSIONS

14

1019

Chapter 2

Observations on the Status of Experimental Research

on Discourse Comprehension

2.1. PERCEPTUAL PROCESSES: LETTER AND WORD IDENTIFICATION 212.2. SYNTACTIC-SEMANTIC PARSING 27

2.3. AMBIGUITY 33

2.4. SEMANTIC UNITS: PROPOSITIONS 37

2.5. COHERENCE 43

2.6. KNOWLEDGE STRUCTURES 46

2.7. INFERENCES 49

2.8. MACROSTRUCTURES 52

2.9. SCHEMATIC SUPERSTRUCTURES 54

2.10. OUTLOOK 59

5/195


6/195


7/195


8/195


9/195


10/195


11/195


12/195


13/195


14/195


15/195


16/195


17/195


18/195


19/195


20/195


21/195


22/195


23/195


24/195


25/195


26/195


27/195


28/195


29/195


30/195


31/195


32/195


33/195


34/195


35/195


36/195


37/195


38/195


39/195


40/195


41/195


42/195


43/195


44/195


45/195


46/195


47/195


48/195


49/195


50/195


51/195


52/195


53/195


54/195


55/195


56/195


57/195


58/195


59/195


60/195


61/195


62/195


63/195


64/195


65/195


66/195


67/195


68/195


69/195


70/195


71/195


72/195


73/195

denotedbypictorialexpressions,denoting(maybe)anindistinctpolitical

situationandasituationofmarkedcontrastoropposition.Strictlyspeaking

S1isungrammatical:theSubordinateclauseshouldspecifyacomparison

forthesubjectinthemainclause,namelyGuatemala,andthatwould

74/195


74/195

137138

Figure4.4.PropositionalstructureofS1ofNewsweektext.

withanimplicitagent,namely,we,I,orNewsweek,denotingtheauthorof

thetext.Wehaveidentifiedthisimplicitagentwiththeconstantx0.Inother

words,thisfirstclausecanbeinterpretedasaconditional,forexample,Ifwecompare....

Atheoreticalproblemalsoexistsforexpressionswithagenitive,suchshadesofgray,whichwetakeasacomplexobject,describedbyaconjunctionofatomicpropositions,inwhichshadesofisabinarypredicate.Wemightalso

takethislatteratomicpropositionasamodifierofgray.

Notealsothatthelocativeparticipant-namely,ElSalvador-isoneoftheparticipantsdependentonthepredicatecompare:ItisaboutgrayinEl

Salvador,andnotthecomparingwhichtakesplaceinElSalvador,whichwould

placethelocationinthehigherlevelcircumstancecategory,alongwithtime.We

seethatforallcomplexpropositionsthiscircumstancecategoryfeaturesatime

categorywhichhasbeenfilledwiththepragmaticallydefinedconcept(now),

denotingthetimeofspeaking/writing,expressedbythepresenttensesinthe

sentence.IfweinspectTable4.1andFigure4.4,weseethattheon-linestrategic

analysisofthefirstsentenceisratherunsuccessful,mainlybecauseofthe

unresolvedinterpretationotthemetaphor:Areaderdoesnotknowyetwhattheauthoristryingtoconvey.Whatthereaderdoesunderstandisthattwocountries

arecompared,orratherthe(political)situationinthesetwocountries,

denotedbypictorialexpressionsdenoting(maybe)anindistinctpolitical


75/195

whitein(S1)asmarkedcontrast.Theexpectationgeneratedfrom

knowledgeonthebasisofontheleftissatisfiedbyextremeMarxist-

Leninistwhichindeedisapoliticalpropertyofsomegroupsontheleft.

76/195


76/195

141142

*Prettymodifiesapredicate,notanargument.

Figure4.5.PropositionalstructurcofS1ofNewsweektext.

mentofmeaningandreferencemayhavetowaituntiltheendofthesentenceor

evenuntiltheinterpretationofthenextsentence.Itisinthisreferencecolumn

thatweindicatetheindividualobjectsorpropertiesreferredto,whichmayin

latersentencesbecoreferredwith,eitherimplicitly(bypresupposition)or

explicitly.Thenextsentence,S2,islessvague(Table4.2,Figure4.5):

(S2)OntheleftisacollectionofextremeMarxist-Leninistgroupsledbywhatonediplomatcallsaprettyfacelessbunchofpeople.

Theprevioussentencehadintroducedtwocountriesandcomparedthem,

presumablywithrespecttotheirpoliticalsituation.Thus,apoliticsframewas

activatedwhichturnsouttobecrucialfortheinterpretationofthesecond

sentence.Thus,firstwehavetoknowthatontheleftisnotjustaconcretelocationidentification,butapoliticalqualificationofapositioninthepolitical

spectrum.Theassociatedknowledgeissomethinglikeprogressive,liberal,

orsocialist.Andtheexpectationisthattherightwillalsobementioned.

Thisisparticularlythecaseafterthepoliticalinterpretationofbluckand

hiti(S1)kdttThttitdf


77/195


78/195


79/195


80/195


81/195


82/195


83/195


84/195


85/195


86/195


87/195


88/195


89/195


90/195

211

Table 6.2

Macrostrategies in the Comprehension of the NewsweekText

Propositional inputas

Provisionalmacro-operation

A. Provisionalmacroproposition(s)

Specific mucrocu

Title: ZERO (SELECTION) There are no (political) Title/Headline

212

Table 6.2(Continued)

Prupositional inputas

Provisionalmacro-operation

A. Provisionalmacroproposition(s)

Spccific macroc

S7 DELETION G. regime first wclcomed History and

irrelelection of Reagan (6) Referent: Rcaga


91/195

ZERO (SELECTION) There are no (political) Title/Headline

NO CHOICES choices in G. (I) First positionThere is no solution to the Bold, large type

problems in G.

(Overall macroproposition)

S, GENERALIZATION (Metaphorical interpr. )

or ZERO The political situation in G. First sentenceis more extreme than in Mention of impoES (2) referents (G.,ES)

DELETION There are unorganized com Opinion diplomGENERALIZATION munists on the left (3) irrelevant

S, DELETION On the right is an elite who History and spcc

has the power and who names are irr.

was helped by US (4)S; DELETION Political center has been Specifics irr. (30

GENERALIZATION murdered by regime in victims)G.. but still exists in ES

(5)

SS ZERO (2) ()pinion diploma

DELETION mcleNarit

s6 ZERO w and (1) REPETITION

I .ml o( paratr,t

election of Reagan (6) Referent: RcagaNew paragraph (

SH DELETION Gs regime had counted on Names and detail

CONSTRUCTION Rs administrations help(7)

SM

ZERO Gs regime now disappointed

about R.s adm. (8)

General predicate(con-

clusion about actual

17.

135: Ufavors

more moGENERALIZATIO Gs regime does not like dc- 18 =B17

mands from US for politi

cal moderation (8)

DELETION US has no influence in G. Opinion diplomat irr. 19 BI: if a

(10) country,

of millGENERALIZATIO Gs regime acts negatively New paragraph 20. B15: Cong

C ONSTRUC TION towards US represent. Head(11)

91/195

92/195


92/195

213

SISCONSTRUCTION=(II)DELETIONGsregimethinkstheyOpinionirr.

aUSwaralone(13)

S17DELETION/ZEROGsregimewillfight

withoutUShelp(l2)SIXDELETION=(12)REPETITION

NewparagrapSlvCONSTRUCTIONGsreimeacuiredAfterSls(=S1-

frommanycountriestInitialsentenararah

CONSTRUCTION,or=(13)SELECTIONAlsoIsraelhels14bulrincial

CONSlRUCTION_(S14)Detailsofw

Isr.

SzCONSTRUCTION(S14Otherdetailaid

S,CONSTRUCTIONG.usesUSmilitaryResultof

DELETIONmcnt(14b)

Sza1.[:RO(,ccA9)

IitficultrclationStill(v,reimeandRadmNcuararahIIS~

214

Table6.2(continued)

PrupositionalinputasProvisionalA.ProvisionalB.Knowledge,

beliefs

expressedbysentencesmacro-operationmacropropo;ition(s)Specific

macrocucsopinions.attitudes

(secn5):

Gsregimecommitsmassmurders(16)

Reintrod.RcaFirstsentencenewparagraph

SGENERALIZATION(16)Specification

CONSTRUCTION(16)Detailofmurd

(conscqucncc)media

S-~CONSTRUCTION(16)Agcntsofmur

(componentacts)

S,HCONSTRUCTION(16)Victimsalso

S,yCONSTRUCTION(16)Innocentvicti

CONSTRUCTION(Ib)Murderersbla

opponcntsSCONSTRUCTION(16)Detailsirr.

UF?LF:TI()N

S,,Z.I:RO(16)Repetition

SZER0IDFI_ETION(16)doesnotdauntNewparagrap

opponents(17)Head:HvdruChanecofLIumrilla

215

Sm OI.NLKAI I/.-\I1( N (uriill,i inm ~wiiicalor CONSTRUCTION (18)

S;s CONSTRUCTION (17) Dctails of fi,-,ht

(component acts)S ~6 CONSTRUCTION Guerrilla backed by

or GENERALIZATION nist countries (19

216

Table 6.3

Higher Level Macrostrategies for the Newsweek TextIis,t levelr~ iu:

Provisionalmacro-operations

M, ZERO n9i

M: ZERO M;


93/195

CONSTRUCTION (19) Detalla of aid

CONSTRUCTION (19) rr,ults

S) CONSTRUCTION (19) I [,[Its results

S4 ZERO (1) Ncw pal. TITLE

Rc(crent: l S5ai GENERALIZATION (2) No ccntrist Solution. Rcf,cicnt: centri

in FS 20

S4, GF.NERAI,IZATION Also political center i, Dctail,, uf r or CONSTRICTION sacred (?I)

S41 CONSTRUCTION (21) Details ofinurde

Sai CONSTRUCTION t211 Reaction to (2 1

Sa; CONSTRUCTION 1I ) help,, left (22) Conclusion

Sa6 GfNER.4l_IZAhIUN Will Us help Gs le-- Final Scntcl l,C1? winch i hrutal f1TI l?:

;

llEI.ET]ON/ZGKO M;

zl:xo M2

M5 DELETION MS

DF[_ETIUN G

r)l l l HON cj

i~i i i io)^.

93/195


94/195


95/195


96/195


97/195


98/195


99/195


100/195


101/195


102/195


103/195


104/195


105/195


106/195


107/195


108/195


109/195


110/195


111/195


112/195


113/195


114/195

260

herence, parsing, etc., as discussed in these chapters. A test for one level mayhave no bearing on another, as we have just shown in these experiments.Comprehension is just a convenient term for the aggregate of these processes; itis not to be reified. not to be tested for.

Results like the present ones (Kintsch & Yarbrough, 1982) and those ofMeyeret al. (1980) are less ambiguous with respect to the psychological role of

schematic superstructures than are the studies in the narrative domain. As wehave seen, the difficulty in the latter studies is to unconfound textual structurefrom the structure of the content itself. Stories are about actions and readersknow a lot about the structure of actions, inside or outside of stories. The contentin descriptive texts is much less constrained Temporal and causal relations play

Chapter 8

Production Strategies


115/195

in descriptive texts is much less constrained. Temporal and causal relations playa decisive role in stories (see the discussion of Johnson & Mandler, 1980) but areabsent in some types of descriptive texts. This is not to say that content does notconstrain the form of an essay, but the constraints are certainly much weakerthan in the case of stories. Hence demonstrations that rhetorical superstructuresaffect comprehension and memory in these cases can be accepted with someconfidence.

8.1. PROBLEMS OF A PRODUCTION MODEL

FOR DISCOURSE

It is well known that most work done in psycholinguistics is about comprehen-sion. In this respect the psychology of discourse is no exception. Our earliermodel of discourse is also mainly about comprehension. Production was an issuelimited to some aspects of reproduction, mainly in the framework of recallexperiments. How language users go about producing a new discourse, eithermonological or as part of a dialogue, is a problem we know little about (but seeClark & Clark, 1977: Chapters 6 and 7; Butterworth, 1980; Gregg & Steinberg,1980). In this book we will not even attempt to supply an adequate answer to thisimportant question of language use, but will merely formulate a number ofspecific problems and some suggestions about the strategic aspects of discourse

production as they seem to emerge from the previous chapters oncomprehension.

A first observation in this respect is the restatement of an insight, which grewover at least 10 years of psycholinguistic research, that processes of production arenot simply the reverse of processes of comprehension (Hirsch, 1977; Olson, Mack,

& Duffy, 1981; Rubin, 1980). The application of early ideas from generativetransformational grammar suggested that the grammar was a model of linguisticcompetence, and that such a model was neutral regarding the production or com-

prehension processes of language use. Although this assumption may have somevalidity for the account of such abstract objects as sentential or textual structures assuch, it certainly does not hold for the cognitive model itself in which analysis and

115/195


116/195


117/195


118/195


119/195


120/195


121/195


122/195


123/195


124/195


125/195


126/195


127/195


128/195


129/195


130/195


131/195


132/195


133/195


134/195


135/195


136/195


137/195


138/195


139/195


140/195


141/195


142/195


143/195


144/195


145/195


146/195


147/195


148/195


149/195


150/195


151/195


152/195


153/195


154/195


155/195


156/195


157/195


158/195


159/195


160/195


161/195


162/195


163/195


164/195


165/195


166/195


167/195


168/195


169/195


170/195


171/195

PERSET.Withitcometworequestsforsubsets,oneforeachofthejoint

owners.ThepreviouslyestablishedsetsS1andS2satisfytheserequestsandthe

roleofsubsetisassignedtothemintheproblemrepresentation,whichsetsthe

conditionforthearithmeticoperatorADD.

InCombine(2)asupersetisestablishedfirst,carryingwithitarequestfor

appropriatesubsets.SetsS2andS3canbeidentifiedassuchmerelyonthebasis

ofthespecificationslot.However,aslightrewordingofthesecondandthird

sentenceinCombine(2)couldprovideamoredirectlinguisticcuetotheirsubset

roles:Joehasthreeofthesewouldbetheconditionforaspecialsubsetoperator.

whichwoulddirectlyestablishthesubsetrolesofS2(andS3)andthereby

obviatethenecessityforaninference.Indeed,Rileyetal.reportthatsuchslightrewordingmakestheproblemsomewhateasiertosolve-presumablybecauseit

suppliedthechildrenwithanextralinguisticcuethatwasmissingintheoriginal

version.

Compareproblems.InCompare(1)(Table10.14),tworolelesssetsareestab-lishedfirst.Thesearethenassignedrolesaswholeandmatchsetsonthebasisof

P,j(x,)S1Object:P3M

PztX2Role:Sec:PIPzP4-J&TP3MIuantit:PS--~8P4HAVE(xl&xz~superP58(REUEST:S,sub

St,sub

P6M(Y2)-S2Object:P6-MS,P,HAVE(x,yz)Spec:P7~J&,,

Pa3vzuantit:Ps-3STsubRole:sub

P9M(Y3)(_S3Object:P9~M~S21&T,superPtoHOWMANYSec:PII~TSZlsub

P1HAVEx3uantitPlo~*REUEST:Srn//LRole:subCALCULATE:Subtract

SIArs.iS21-nS3*

T.sub

172/195


172/195

375

Table10.11

Change(5)

ACTIONSHORT-FERM

MEMORl

P,j(-r,)F_-S]Object:P,-NtFMc1P14

FHAVE.rSec:+.IFSOME(c1-~P16

uantit:P4PiRule:;P,-,start

PTHEN(F,Ph)Object:P,MS,

P(f(c-Sec:P,,PKJPMc-uantit:P-iP,GIVE(.r-.r,c=Role:Ps-R5(NzREUEST:S;,

S,-It PiNOW(P,-)_S,Object:P11M

PMcSec:P--J5PHAVI?rc3uantit:P?KREIEST:S.Pi8()-z)Role:P,-result

P,M(r4)MATCHS,onP16P17

PHOWMANYh,4FSiFHAVExzP,PAST(P16CALCULATE:Concert-subtract

376

Table10.12

Combine(1)

ACTION

P,Ax1)S,Object:Pz-+M

PzM(,Sec:P,P3-JP3HAVE(xluantit:P4-3P43(yl)Role:/sub/

P5t(X2)SzObject:P6-M

P6Mv2Sec:PSP~-~TPHAVExzzruantit:P8-5Ps5(yz)Role:/sub/

P9M(y3)~S3Object:Py-M

P,oHOWMANY(y3)Spec:P

P11HAVE(x,&xzy3)Quantity:P,o

SHORT-TERMMEMORYRole:~super

REQUEST:Sj,sb//ReinstateS,Sr,sub//

CALCULATE:Add

S~i.s.bS2,-n

Table10.13

Combine(2)

ACTIONSHORT-TERMMEMORY

P,j(x,)S1Object:P3-M

377

Table 10.14

Compare (1)

ACTION SHORT-TERM MENIORYPiP,P,P4

#X,) rObjcct: P, -M I-S,f o ) I Spec: P, P,--

MJ8/whole/

PS r(,v_)~ yS_, ~ Object: P, ---

M

P1, M(y,) i Spec: P; P,-- I

P, HAVE(.c, Quantity: PK--r 5

P;~ 5(v~)---~ ~Role: hnatch/

R) Mlcz) ~ S, -Object: l,,---y M

PM HOWMANY i S ec: P11- --- JP H.AVE(-c,yz) Quantity: P,-- I -

P11 MORE(yzr_) ~-Rolc: P,: rcnu,indcr

~Rh;QUES f: S,

S1v Rcinstute Si

378

Table 10.15

Compare (3)

ACTION

Pi 0 i) -Si

P= M6 1 S ec: P, -P,-+ JP, HAVE(.r, i uantit : P-t ~ 3P, 3(ry -Role: /match/

P; t(.r=) ~S- Object: P,, -- NI

W S ec: P P7-i f P~ HAVE (s= s-) Quantity: PK~ 5PH 5(~v Role: Pa rernainder

P) MORE(y, N*,) REQUEST. S,,,,,,

Piv M6 1) S~

P HOWMANY Ob ect: Pir --~ M I S cc: P,- ~ T uantit :Ill HAVE r- l

-Role: whole


173/195

S1v -. Rcinstute SiCALCULATE: Subtract I F S,

their ownership when the remainder set is constructed. Note that what isinferered is entirely a conceptual structure belonging to the problemrepresentation-not a set of propositions in the textbase. The final two problems(Tables 10.15-l0.16) introduce no new devices, though the by now familiaroperators are combined here in novel ways. In all of the compare problems, thefinal outcome is a MORE-THAN pattern, with a whole set, a match set, and aremainder set, and a calculation goal assigned to one of these. We can eitherassume that these patterns are by themselves conditions for the appropriate

arithmetic operations, or that the whole- match-remainder pattern gets convertedinto a super-sub-sub pattern, the usual condition for addition and subtraction.

What these examples demonstrate is, first of all, that the model works as weclaimed it did. One can indeed specify comprehension strategies at various levelsthat result in the construction of a text representation and a situation model that aresuitable for the problem-solving task. As Kintsch and Greeno (1982) pointed out intheir original report on this work, such a demonstration has some interesting im-

plications. It tells us, for instance, why some problems are harder to solve thanothers. Different problem types require different knowledge structures and differentstrategies (or the application of that knowledge. Change problems, for instance,require that the child know about the TRANSFER schema, which is simple andconcrete and, apparently, is quite within the grasp of even the kindergarten chil-dren. The SUPERSET schema and especially the MORE-THAN schema are muchless available and might require some special training, both at the level of concrete

Table 10.16

Compare (5)

ACTION SHORT-TERM

Pi /O-1) Object: P, = M

P, M(y, ) Spec: P, P,~ ,lPz HAVE(-, , uantit : Pt ~ 8

P, 8(y1) - _Role: ;whole/PS M(c=) ~S_, Object: P;-~ M

Pr, HAVE(.r, c- S ec: Pf, ~ J

P, MORE ( c, 3 Role: PHPa r(.r-) PyPmPi i

Pio M(r, llP11

HAVE(s= y a) SPi- M(c4) S, Object: P - M S,P,zP,,, HAVE Spec: P,a -~ T S=, Quantity:

-Role: match REQUEST: S,-MATCH P,u=Pi-. P,,-P,,

CALCULATE: Subtract Si i.,n..,~

Szl

S=

173/195


174/195


175/195


176/195


177/195


178/195


179/195


180/195


181/195


182/195


183/195


184/195


185/195


186/195


187/195


188/195


189/195

Chapter1

TowardaModelofStrategicDiscourseProcessing

1.1.

1.2.

1.3.1.4.

THESTUDYOFDISCOURSE

BASICASSUMPTIONS

ANOVERVIEWOFTHEMODELCONCLUSIONS

1

4

1019

Chapter2

ObservationsontheStatusofExperimentalResearch

onDiscourseComprehension

2.1.PERCEPTUALPROCESSES:LETTERANDWORDIDENTIFICATION212.2.SYNTACTIC-SEMANTICPARSING27

2.3.AMBIGUITY33

2.4.SEMANTICUNITS:PROPOSITIONS37

2.5.COHERENCE43

2.6.KNOWLEDGESTRUCTURES46

2.7.INFERENCES49

2.8.MACROSTRUCTURES52

190/195


190/195

413

Yngve,V.H.Thedepthhypothesis.ProceedingsqJtheX/!SmpositimillApplied

Mathematics.Providence,R.L:AmericanMathematicalSociety,1961.

Yuille.J.C..&Paivio,A.Abstractnessandtherecallofconnecteddiscourse.

JournalofE_sporinuntctlPscchologc.1969,82,467-472.

7_ammuner.V.L.Speechproduction:Strategiesilldiscourseplanning:A

tlreorWirulandempiricalcnguir_c.Hamburg:BuskeVerlag.1981.

Cont

ents

Prefaceix

2.9. SCHEMATIC SUPERSTRUCTURES 54

2.10. OUTLOOK 59V

The Notion of Strategy in Language and Discourse Understanding

3.1. INTRODUCTION 61

3.2. THE NOTION OF STRATEGY 62

3.3. COGNITIVE STRATEGIES 68

3.4. LANGUAGE STRATEGIES 70

3.5. GRAMMATICAL STRATEGIES 73

3.6. DISCOURSE STRATEGIES 78

3.7. THE REPRESENTATION OF STRATEGIES 95

3.8. A SAMPLE ANALYSIS: THE NEWSWEEKTEXT 98

C

ha

pt

er4

Propositional Strategies

4.1. PROPOSITIONS 109

4.2. COMPLEX PROPOSITIONS 119

4.3. PROPOSITIONS AS COGNITIVE UNITS AND AS STRATEGIES 124


191/195

4.4. A SAMPLE ANALYSIS OF THE NEWSWEEKTEXT 134

4.5. EXPERIMENT l: PROPOSITION FUSION 144

C

ha

pt

e

r5

Local Coherence Strategies

5.1. LOCAL COHERENCE 14

5.2. UNDERSTANDING SENTENCES 151

5.3. LOCAL COHERENCE STRATEGIES 15

5.6. ANALYSIS OF THENEWSWEEKTEXT 18

Cha

pt

er6

191/195

9

StrategiesfortheUseof

Knowledge

9.1.KNOWLEDGESYSTEMS304

9.2.KNOWLEDGEUSEINMETAPHORCOMPREHENSION312

9.3.REDINTEGRATION,REMINDING,ANDPARTIALMATCHES315

9.4.STRATEGIESFORKNOWLEDGEUSEINDISCOURSE317

9.5.KNOWLEDGEUSEINTHENEWSWEEKTEXT319

9.6.EXPERIMENT6:GLOBALBIASESINKNOWLEDGEUTILIZATION324

APPENDIX:TEXTSUSEDINEXPERIMENT6A33i

C

hapt

er1

0

TheCognitiveModel

10.1.FROMTHETEXTREPRESENTATIONTOTHESITUATIONMODEL336

10.2.AFRAMEWORKFORAPROCESSMODEL346

10.3.SHORT-TERMMEMORYUSEINDISCOURSEPROCESSING352

192/195


192/195

Macrostrategies

6.1.THENOTIONOFMACROSTRUCTURE1896.2.MACROSTRUCTURESANDDISCOURSEUNDERSTANDING1916.3.CONTEXTUALMACROSTRATEGIES196

6.4.TEXTUALMACROSTRATEGIES201

6.5.MACROSTRATEGIESINACTION:SOMEEXAMPLES206

6.6.EXPERIMENT3:INTERESTANDLEVELOFDESCRIPTION

ASMACROSTRUCTURECUES222

6.7.EXPERIMENT4:PRIMINGMACROPROPOSITIONS226

C

ha

pt

er7

SchematicStrategies

7.1.SUPERSTRUCTURES235

7.2.SCHEMATICSTRATEGIES237

7.3.SCHEMATICANALYSISOFTHENEWSWEEKTEXT242

7.4.PREDICTIONSANDIMPLICATIONSOFTHETHEORYOF

SCHEMATICSUPERSTRUCTURES251

7.5.EXPERIMENT5:THEROLEOFRHETORICALSTRUCTUREIN

DESCRIPTIVETEXTS253

Chapter8

ProductionStrategies

8.1.PROBLEMSOFAPRODUCTIONMODELFORDISCOURSE261

8.2.INTERACTIVEANDPRAGMATICPRODUCTIONSTRATEGIES264

8.3.SEMANTICPRODUCTIONSTRATEGIES272

8.4.STRATEGIESFORESTABLISHINGLOCALCOHERENCE278

8.5.PROPOSITIONALPRODUCTIONSTRATEGIES280

8.6.SOMECONSEQUENCESFORSENTENCEPRODUCTIONSTRATEGIES283

8.7.SOMEPRODUCTIONSTRATEGIESFORTHENEWSWEEKTEXT2868.8.TOWARDAPROCESSMODELOFDISCOURSEPRODUCTION293

Ch

apt

er

9

10.4. RETRIEVAL FROM EPISODIC TEXT MEMORY 356

10.5. A PROCESSING MODEL FOR UNDERSTANDING AND SOLVING WORD

ARITHMETIC PROBLEMS 364

10.6. EPILOGUE 383

Author Index 405Subject Index 413

References

387

,7

Figure 2.1. A few of the neighbors of the node tier the letter T in the first position in a word, and theirinterconnections. (From J. L. McClelland & D. F. Rumelhart, An interactive activation model of contexteffects in letter perception.Psychological Review,1981, ,48. 375-407. Copyright 1981 by the AmericanPsychological Association. Reprinted by permission of the author.)


193/195

193/195


194/195


195/195

Documents

VAN DIJK-KINTSCH - Strategies of Discourse Comprehension