Sie sind auf Seite 1von 231

Studies in Contemporary German Social Thought Thoma s McCarthy, general editor

(partial listing)

James Bohman, Public Deliberation: Pluralism, Complexity, and Democracy James Bohman and Matthias Lutz-Bachmann, editors, Perpetual Peace: Essays on Kant's Cosmopolitan Ideal Craig Calhoun, editor, Habermas and the Public Sphere

Maeve Cooke, Language and Reason: A Study of Habermas's


Jürgen Habermas, Between Facts and Norms: Contributions to a Discourse Theory of Law and Democracy Jürgen Habermas, Justification and Application: Remarks on Discourse Ethics

Jürgen Habermas,

Jürgen Habermas,

Jürgen Habermas, Moral Consciousness and Communicative Action

Jürgen Habermas,

Jürgen Habermas,

Jürgen Habermas, Philosophical-Political Profiles

Jürgen Habermas, Postmetaphysical Thinking: Philosophical Essays

Jürgen Habermas,

Jürgen Habermas, The Structural Transformation of the Public Sphere: An Inquiry into a Category of Bourgeois Society

Jürgen Habermas, editor, Observations on "The Spiritual Situation of the Age"

On the Logic of the Sotial Sciences The Inclusion of the Other: Studies in Political Theory

The New Conservatism: Cultural Criticism and the Historians' Debate The Philosophical Discourse of Modernity: Twelve Lectures

On the Pragmatics of Communication (edited by Maeve Cooke)



The Critique of Power: Reflective Stages in a Critical Sotial Theory



The Struggle for Recognition: The Moral Grammar of Sotial Conflicts


Honnet h an d

Hans Joas, editors, Communicative Action: Essays on Jürgen Haber-

mas 's Th e Theory of Communicative Action Axel Honneth, Thomas McCarthy, Claus Offe, an d Albrecht Wellmer, editors, Cul- tural-Political Interventions in the Unfinished Project of Enlightenment Axel Honneth, Thomas McCarthy, Claus Offe, an d Albrecht Wellmer, editors, Philo- sophical Interventions in the Unfinished Project of Enlightenment

Maurizio Passerin d'Entrèves and Seyla Benhabib, editors, Habermas and the Unfinished Project of Modernity: Critical Essays on Th e Philosophical Discourse of Modernity

Ernst Tugendhat,

Self-Consciousness and Self-Determination

Albrecht Wellmer,

Endgames: Essays

and Lectures on the Irreconcilable Nature of Modernity

Albrecht Wellmer, The Persistence of Modernity: Essays on Aesthetics, Ethics, and Postmod- ernism

On the Pragmatics of Communication

Jürgen Habermas

edited by Maeve



MI T Press, Cambridge, Massachusetts

©1998 Massachusetts Institute of Technology

This volume is published by arrangement with Suhrkamp Verlag, Frankfurt a m Main, Germany. Th e sources on which the translations are based are listed in the acknowl- edgments.

All rights reserved. No part of this book may be reproduced i n any form or by any

electronic or mechanical means, (including photocopying,

storage an d retrieval) without permission in writing from the publisher.

recording, or information

Thi s boo k was set i n Baskeryille by Wellington Graphics and in the United States of America

was printed an d bound

Library of Congress Cataloging-in-Publication


Habermas, Jürgen. O n the pragmatics of communication / Jürgen Habermas :

edited by Maeve Cooke.


cm . —

(Studies in contemporary

German social

Essays translated from

Includes bibliographical references an d index.

ISBN 0-262-08265-9 (alk. paper)


1. Pragmatics. I . Cooke, Maeve.

II. Title.

III . Series









Acknowledgment s


vi i

Introductio n





1 What I s Universal Pragmatics?




2 Social Action,

Purposive Activity, an d Communication






Rationality and the Theories



an d Action






Speech Acts, Linguistically Mediated

Interactions, and the Lifeworld (1988)




Comment s on Joh n Searle's "Meaning,


and Representation"




6 Toward a Critique of

the Theor y of




7 Som e Further Clarifications of

the Concept



Rationality (1996)



8 Richar d Rort/ s Pragmatic Tur n (1996)



9 O n

the Distinction between Poetic and









10 Questions and Counterquestions (1985)

Selected Bibliography and Further

Inde x





Editor's Acknowledgments

I woul d

ing, for

and for considerable help over and above his editorial duties with

regard to translation difficulties. M y thanks are also due to Jürgen Habermas for encouraging the project, for replying promptl y to my many queries, and for his unfailing cordiality. Ronald Böhme helped

Depart -

men t o f German, University College Dublin, and he also compiled the index . Here , too , I a m ver y grateful . Thanks , finally, t o Marti n Sauter, no t just for his painstaking help i n checking translations bu t for his love and support over the many years i t took to pu t this boo k together.



like to thank Thomas McCarthy for initiating this undertak- fulfilling the role of series editor i n an exemplary manner,


th e

proofreading ,

wit h

assistance fro m

th e


T h e

translations of








o n

earlier published translations, as



translation o f


1 by Thoma s

McCarth y was


i n


and the Evolution

of Society, ©197 9

Jürge n Habermas , by Beacon Press,




translation o f


2 by Thoma s

The Theory of Communicative Boston.

McCarth y was


Jürge n Habermas , by Beacon Press,


1 ,

i n

©198 4

vol .

A translation o f chapter

first publishe d i n Communicative Action: Essays on Jürgen

T h e Theor y o f Communicative Action, edited by Axe l Honnet h

Hans Joas, ©1991

by Polity Press, Cambridge,

by Jeremy



an d Doris L . Jones






Editor's Acknowledgments

A partia l translatio n o f chapte r 4 was publishe d i n Philosophical Prob- lems Today, edite d b y Guttor m Fl0istad , ©199 4 b y Kluwe r Academi c Publishers, Dordrecht.

A translatio n o f chapte r 5 was publishe d i n John Searle and His Critics,



Blackwell, Oxford , U.K.

A translation o f chapter 6 by Willia m M . Hohengarte n was publishe d

i n Jurge n Habermas , Postmetaphysical Thinking, ©199 2 b y Massachu -







Gulick, ©1991


setts Institute o f Technology,

A translation o f chapter 9 by Frederick G. Lawrence

i n Jurge n

Cambridge ,




©198 5

Habermas ,

The Philosophical Discourse

of Modernity,

by Massachusetts


o f Technology,



A translation o f



by James

Bohma n


publishe d

i n

Habermas and Modernity, edite d b y Richar d Bernstein , ©198 5 b y Polit y


Chapters 7 and 8 and the final part o f chapter 4 have been translated





Cooke and appear i n English for the first time.


Maeve Cooke


translation, te n essays that present the mai n concern s o f Habermas's progra m i n forma l pragmatics. Its ai m is to convey a sense o f the overall purpose o f his linguistic investigations, while introducing the

reader to their specific details. Habermas's forma l pragmatics fulfills two mai n functions. First, i t serves as the theoretical underpinnin g for his theory o f communicative action, whic h is a crucial element i n his theor y o f society. Second, i t contributes to ongoing philo - sophical discussion o f problems concerning truth, rationality, action,

This anthology

brings together for

the first time, i n revised or

and meaning. Correspondingly, the ai m o f the present


is twofold . First, i n providin g better access to essays by

Haberma s

that focus explicitly o n language, i t may help those interested i n

social theor y to assess critically the linguisti c basis fo r his account s

communicative rationality. Second, i t mor e traditional philosophical prob-

lems to understand and to appreciate Habermas's treatment o f


may help those interested i n

o f communicative

action an d


"universal pragmatics." Th e adjective "universal" was meant to indi -


matic analyses o f language. Whereas earlier pragmatic approaches to




original ter m for his linguistic research


his linguistic project


progra m



language ha d tended to analyze particular contexts o f language




ou t



universal features

o f



guage .

Thi s

explain s

th e


o f

hi s

programmati c


"Wha t







Is Universa l Pragmatics?, " first publishe d i n 1976. However , i n a footnote to the 1979 English translation, Habermas expresses dis-

satisfaction with the label

"formal pragmatics." On e advantage o f the latter terminology, i n his

view, is that i t reminds

semantics. As we shall

larly crucial i n Habermas's accounts o f meanin g and truth.

Wha t is mean t by universa l or, as we shoul d no w say, forma l prag - matics? Habermas's starting poin t is that forma l analysis o f language

should no t be restricted to semantic analysis, for forma l investigation

o f the pragmatic dimensions o f language

important. By the "pragmatic" dimensions o f language, Habermas

mean s those pertainin g specifically t o th e employment o f sentence s i n

utterances. H e

tolerant sense to refer to the rational reconstruction o f general

intuition s o r competencies . Forma l pragmatics , then , aims at a sys- tematic reconstruction of the intuitive linguistic knowledge o f com - petent subjects, the intuitive "rul e consciousness" that a competent

makes clear that "formal " is to be understood i n a

"universal" and a preference for the ter m

us that forma l pragmatics is related to forma l see, the natur e o f this relationshi p is particu-

is equally possible and

speaker has o f her own language.

I t aims to explicate pre theoretical

knowledge o f a general sort, as

opposed to the competencies

o f

particular individuals and groups. Forma l pragmatics thus calls


min d the unavoidable presuppositions that guide linguistic ex- changes between speakers and hearers i n everyday processes o f com-

munication i n any language. I t makes us aware that, as speakers and hearers, there are certain things we must—as a matter o f necessity—

to be success-

always already have presupposed i f communicatio n is

ful. I n focusing o n the forma l properties o f speech situations i n


general, Habermas's progra m may thus be distinguished fro m

pirical pragmatics—for example, sociolinguistics—which looks pri- marily at particular situations of use.


formal-pragmatic investigations into everyday


tic practices i n moder n societies are attempts to reconstruct the universal competencies that are involved whe n social actors interact wit h th e ai m o f achievin g mutua l understandin g {Verständigung)} Communicative competence is crucial for Habermas's social theory, whic h is based o n the thesis that action oriented toward reaching understanding is the fundamental type o f social action. His name for

his analysis o f i t turns o n


validity. Mor e precisely, linguistic utterances as they are used i n

as claims t o

validity. Fro m his perspective, everyday linguistic interaction is pri- marily a matter o f raising and responding to validity claims. Haber¬


strategic, figurative, o r symboli c interaction , bu t h e contend s tha t


I n its simplest terms, communicative action is action whose success

claim raise d b y identifies three

basic types o f validity claims that are raised by a speaker with her speech act: a claim to the trut h o f wha t is said o r presupposed, a


action o f this sort is "communicative," and











processe s o f communicatio n ca n be

construe d






o f

linguistic interaction, such


are parasitic o n


depend s o n th e hearer' s respondin g t o th e validity

the speaker with a "yes" or a

"no." Here , Habermas

claim t o th e normative Tightness o f th e

speech act i n th e given

context or of

o f the speaker.

speaker raises all

communicative exchange, however, jus t one o f the claims is raised

explicitly; the other two remai n implicit presuppositions o f under-

standing the utterance. Th e three validity "universal " b y Habermas , i n th e sense o f communicatively used speech act.

communicatively, th e

claim to the truthfulness

the underlying norm , and a

I n

usin g

a linguisti c expression



three o f

simultaneously. I n a typical

claims are described as bein g raise d wit h every

T h e three universal validity claims—to truth, normative lightness,

a n d truthfulness—provide a basis fo r classifying speech acts. Thus ,

communicative utterances can be divided into three broad catego- ries according to the explicit claims they raise: constative speech acts are connecte d i n th e first instanc e wit h trut h claims , regulativ e speech acts with claims to normative lightness, an d expressive speech acts with claims to truthfulness.

T h e

thesis o f

three universal validity claims




bot h language





O n




i t


mean t to provide a mor e

convincing basis fo r classifying speech


than, for example,

the proposals

o f Austin and his followers or


mor e theoretically motivated typologies o f Searle and his followers.

O n



hand ,










validity claims,

thereby giving rise


a particular




conception o f social order as reproduced throug h communicative


I n showing that everyday linguistic interaction depends o n raising a n d recognizing validity claims, Haberma s presents a picture o f so- cial order as a network of relationships o f mutual recognition that have tw o significan t characteristics. The y are , first, cooperativ e rela - tionships o f commitmen t an d responsibility: participants i n commu - nicativ e interactio n undertake t o behav e i n certai n ways, an d th e success o f th e interactio n depend s o n th e cooperation o f bot h partie s

mutual recognition charac- inheren t rational dimen -

sion: the communicative actor undertakes an obligation to provide


while his counterpart i n action may either accept the proffered

the basis o f better reasons. I n this action involves a rudimentar y prac-

involved. Second, the relationships o f teristic fo r communicativ e actio n have

a n

reasons for the validity o f the

claims he

reasons or challenge the m o n sense, everyday communicative

raises with his

tice o f "argumentation." Furthermore , these everyday practices o f

giving reasons for and against controversial validity claims—some- times referred to by Habermas as naive communicative action — poin t toward the possibility o f other, mor e demanding forms o f

argumentation, whic h he calls "discourse." Everyday communicative

the reasons sup- Whe n this back-

groun d consensus is shaken—as will happen mor e frequently i n

posttraditional societies—communicative action cannot continue routinely. Participants then have three options: they can switch to strategic action; they can break off communicatio n altogether; or

action normall y operates o n the assumption that porting the validity claims raised are goo d ones.

they can recommenc e their communicative activity at a


mor e reflective level—namely, argumentative speech. I n

the proc-

esses o f argumentatio n know n as discourses, certain idealizing sup-

everyday communicative action are

formalized. These presuppositions are unavoidable i n the sense that they belong to the very meaning o f what i t is to take part i n argu- mentation; they are idealizing i n the sense that they are typically counterfactual and will not as a rul e be satisfied more than approxi - mately. Thus, Habermas claims, participants i n argumentation nec- essarily suppose, amon g other things, that they share the commo n

positions already operative i n


aim o f reaching agreement with regard to the validity o f the disputed validity claim, that no force except that o f the better argument is exerted, that n o competent parties have been excluded fro m the discussion, that no relevant argument has knowingly been sup- pressed, that participants are using the same linguistic expressions

i n


same way, an d so on . These idealizing suppositions refer bot h



practice o f argumentation and to its outcome. For Habermas,

the various idealizing suppositions unavoidably guiding argumen-

tation are what give meaning to the

ideas that transcend all local the validity claims raised i n


everyday processes o f argumentation


ideas o f truth and justice as

o f validity. To

the extent

have a

connection i n principle with possible vindication i n dis-


they have an inheren t context-transcendent power. This

power is the rational potential built into everyday processes o f communication.

Habermas's picture o f everyday communicative action thus has

important implications for critical social theory. For one thing, i n

network o f cooperation involving com -

i t opposes

take interactions between strategically acting subjects as fundamen-

tal, for example, models grounde d i n decision or game theory. For another, i n the context-transcendent potential o f the validity claims raised i n everyday communicative processes, i t locates a basis for a "postmetaphysical" conception o f communicative rationality and , ac- cordingly, a standard for critique. As that conception refers to a potential already built into everyday communicative action, it situ-


whic h i t points


presenting social order as a mitmen t and responsibility,


o f




i n



the ideas o f truth an d justice toward i n idealizing suppositions that are part

are grounde d


f everyday huma n activity. Moreover, communicative rationality is


o t





o f


prevailing i n any local

context o f communicative activity. Rather,

the idealizing supposi-

tions o n whic h i t rests provid e standard s fo r criticizin g loca l practice s

o f justification, bot h with regard to the outcomes o f the agreements

reached and with regard to practices o f justification themselves. Thus the idea o f communicative rationality is meant to provide a

truth an d

postmetaphysical alternative to traditional conceptions of

justice that nonetheless

avoids value-relativism.







Fro m a mor e strictly linguistic-philosophical poin t o f view, Haber- mas's forma l pragmatics offers a n approach to questions o f meanin g

a n d truth that radicalizes th e linguistic tur n

phy. I n his view, traditional formal-semantic approaches to meanin g have been guilty o f three kinds o f abstractive fallacies: a semanticist abstraction, a cognitivist abstraction, an d a n objectivist one . Th e semanticist abstraction is th e view that th e analysis o f linguistic

meanin g ca n confin e itself t o th e analysis o f sentences, abstractin g

fro m th e pragmatic contexts

that all meaning can be traced

back t o th e propositional content o f utterances, thus

ducin g meanin g t o th e meanin g o f assertoric sentences . Th e objec - tivist abstractio n is th e view tha t meanin g is t o b e define d i n term s

o f objectively ascertainable truth conditions, as opposed t o th e knowledge o f th e trut h condition s tha t ca n b e impute d t o speakers o r hearers. Fo r Habermas, pragmatic theories o f meaning have th e advantage that they focus no t o n sentences bu t o n utterances (he is thinking here primarily o f th e use-oriented theories o f meanin g suggested b y th e later wor k o f Wittgenstein, o n th e on e hand , an d the wor k o f Austin an d Searle, o n th e other) . Furthermore , prag- matic theories o f meaning d o no t emphasize only th e assertoric o r descriptive modes o f language use; they draw attention to th e mul - tiplicity o f meaningfu l ways o f usin g language . Finally, suc h theorie s

T h e cognitivist abstraction is th e view

withi n moder n philoso-

o f th e use o f sentences i n utterances.

indirectly re -

stress th e connection practices; they draw

o f th e forms o f life i n whic h communicative activity is always em -


between th e meanin g o f utterances an d social attention t o th e institutions an d conventions

t o

t o those o f forma l seman-

tics. Th e great strength o f forma l semantics has been its attempt t o retain a connection between th e meanin g o f linguistic expressions

a n d some notio n o f context-transcendent validity. I n th e mai n prag-

matic approaches,



th e later wor k o f


meanin g have weaknesses complementary

I n


view, however,

existing pragmatic



this connection

either slips fro m

o r is interpreted to o narrowly i n a cognitivist way. Fo r

use theories

o f meaning

derived fro m



i n effect


a context-transcendent

no -


o f

validit y

b y reducin g

i t

t o

th e prevailin g validity

o f

loca l

language games an d particular forms o f life. O n th e other hand , pragmatic approaches that have attempted t o avoid such a reduc- tion—Habermas mention s Searle's speech-ac t theory—typicall y have succumbed t o th e cognitivist abstraction, interpreting validity to o narrowly as propositional truth. Habermas sees his ow n pragmatic

theory o f meanin g as a n attempt t o combine th e productive insights

o f existing formal-semantic an d pragmatic approaches to meaning

while avoiding their respective weaknesses. H e regards speech-act

theory as a fruitful starting point , bu t insufficient as i t stands, an d

attempts t o

assertibility conditions. I n a sense, then, Habermas's pragmatic the-

o r y o f meanin g ca n be

Austin an d Searle with Frege

regarded as th e proposed happy marriage o f

buil d into i t th e formal-semantic emphasis o n truth o r

an d Dummett .

Fro m th e speech-act theory o f Austin an d Searle (who m h e praises for renderin g Austin's theory mor e precise), Habermas takes over

the emphasis o n utterances rather than sentences as th e central uni t

o f

an d descriptive modes o f lan-

a n equal footing—othe r ways

th e

traditional narro w focus o n assertoric

guage use t o include—potentiall y o n

o f using language, such as acts o f promising , requesting, warning , o r

confessing. I n addition , h e finds fruitfu l speech-ac t theory' s empha - sis o n th e illocutionar y forc e o f utterances , that is, o n th e fact that a

speaker i n saying somethin g also does something . However , i t ma y b e

o f illocu-

tionary force, whic h goes beyond Austin's i n a numbe r o f significant

respects. Austi n used th e notio n o f illocutio n t o refer t o th e act o f

helpful here t o notice Habermas's distinctive conception

analysis. H e also associates himself with




uttering sentences with propositional content. Fo r him , th e force

o f

an utterance consists i n th e illocutionary act—i n th e attempt

t o

reach a n uptake; h e contrasted th e force o f a n utterance with its

meaning, conceived as a property o f th e sentence uttered. Haber-

mas's objectio n t o this is threefold : first, Austin' s distinctio n betwee n force an d meanin g overlooks th e fact that utterances have a mean- i ng distinct fro m th e meaning o f the sentences they employ; second,

it is connecte d

constatives an d performatives, whereby initially, fo r Austin, only con - statives are connecte d with validity claims; third, i t neglects th e ra - tional foundatio n o f illocutionary force. B y contrast, Habermas

with a problematic classification o f speech acts into



propose s a n accoun t o f utteranc e meanin g tha t brings together th e

categories o f meanin g an d force ; h e

ary force to all utterances that are used communicatively; and he

emphasize s th e rational foundatio n see, Habermas's pragmatic theor y

o f utterances as inseparable fro m the act o f uttering them, an d

defines utterances as acts o f raising validity claims. Hi s definitio n o f illocutionary force follows fro m this: illocutionary force consists i n a

speech act's capacity to motivate a hearer to act o n the premise that the commitmen t signalled by the speaker is seriously meant. O n this

conception , illocutionar y

sumption o f a

o f the validity o f the claims she raises. So understood, illocutionary

force is a rational force, for i n performin g a speech act, the speaker

undertake s t o suppor t wha t

although Habermas acknowledges speech-act theory as the most fruitful poin t o f departure for his progra m o f forma l pragmatics, he engages with i t critically, makin g use o f some o f its central categories

i n distinctive ways.

forc e is boun d u p wit h th e speaker's as-

warranty, i f challenged, to provide reasons i n support

extends th e nodo n o f illocution -

o f illocutionar y force . As we shal l gives a n account o f the meanin g

she says wit h reasons, i f necessary. Thus ,

Fro m the poin t of view o f Habermas's progra m o f forma l pragmat-

ics, the mai n weakness of speech-act all communicatively used utterances

principle context-transcendent. H e attempts to make goo d this deficiency by drawing o n Michael Dummett' s account o f under- standing meanin g i n terms o f knowin g assertibility conditions. I n analogy with Dummett' s formulatio n o f what i t is to understand the meanin g o f an assertoric expression, Habermas proposes that we understand an utterance whe n we kno w what makes i t acceptable. Truth-conditional semantics runs into difficulties whe n i t explains the meanin g o f sentences i n terms o f their truth conditions without mediatio n throug h th e knowledge th e speaker o r heare r ma y have o f such conditions. Thus Habermas adopts Dummett' s "epistemic turn " and criticizes Donal d Davidson for offering an objectivist reading o f Frege's and Wittgenstein's thesis that to understand an utterance is to kno w what is the case i f i t is true. H e rejects this objectivist reading as tacitly assuming that for every sentence, or at least for every assertoric sentence, procedures are available for effectively deciding

theory is its failure to connect wit h validity claims that are i n



whe n the truth conditions are satisfied. Such an assumption, he

knowledge that

regards the simple predicative sentences o f an observational lan- guage as fundamental. Habermas the n follows Dummett , wh o suggests replacing the emphasis o n truth conditions with a consid- eratio n o f wha t i t is fo r a speaker t o know when th e trut h condition s woul d be satisfied. This is what he refers to as Dummett' s epistemic turn; he, however, wants to tur n even further. As Habermas reads it , Dummett' s theory o f meaning has two mai n shortcomings that pre- vent his developing fully the inheren t potentials o f the epistemic turn. Th e first is a prioritization o f trut h claims over other kinds o f validity claims: Dummett' s notio n o f assertibility conditions accords priority to assertoric utterances. I n order to make roo m o n an equal footing for nonassertoric utterances such as promises, imperatives, o r avowals, Haberma s prefer s t o spea k o f acceptability conditions . Th e second is that Dummett' s notio n o f assertibility conditions is in - sufficiently pragmatic: i t remains o n the semantic level of analysis inasmuch as i t relies o n an ideal o f validity that is conceptually independent o f discursive practices o f redeeming validity claims. This last objection takes us to the heart o f Habermas's pragmatic theory o f meaning.

argues, implicitly relies o n an empiricist theory o f

Before considering it , however, i t may be helpful to clarify the status o f the theory. Broadly speaking, i t seems possible to distin- guis h betwee n tw o account s o f its status. Accordin g t o th e first, a pragmatic theory o f meaning is merely an extension o f truth-condi-

tional semantics i n the sense that i t broadens its focus.

Habermas's theory leaves the basic assumption o f the formal-seman- tic account o f the meaning o f sentences intact, while expanding its range , first, t o includ e nonassertoric linguisti c expressions and , sec- ond , t o embrac e utterances as wel l as sentences . Hi s earlie r essay "What Is Universal Pragmatics?" suggests this account o f the tasks o f

a pragmatic theory o f meaning. However, i n most o f his later writ-

ings, he seems to offer a mor e radical account. According to this, a pragmati c theor y o f meanin g undercuts th e formal-semanti c ap- proach to meaning. This view is suggested, for example, i n chapters 2 and 3 i n the present volume, where Dummett' s assertibility-condi- tional theor y o f meanin g is criticize d fo r failin g t o carr y throug h

O n this view,



completely the move fro m the semantic to the pragmatic level o f analysis. I n a recent response to objections raised by Herber t Schnadelbach (see chapter 7) , Haberma s clarifies the status o f his pragmatic theory of meaning i n a way that suggests that bot h o f these interpretations are correct. Starting fro m a distinction between the communicative and noncommunicative use o f language, he ac-

knowledges that epistemically used propositional sentences and teleologically used intentional sentences have a meaning content that is i n some sense independent o f the illocutionary acts i n whic h they can be embedded. I n order to understand propositional sen-

represent states of affairs or facts, i t is

tences that serve purely kno w their trut h conditions. I n order to understand intentional sentences that serve to calculate action consequences monologically—without reference to a second person—i t is suffi-

cient to kno w their success conditions.

Such sentences, whic h are

used noncommunicatively, can be analyzed exhaustively wit h the tools o f forma l semantics. However, they are special cases of lan- guage use, due to a feat of abstraction that suspends their pragmatic

possible communicative situations i n whic h a speaker

dimension: the

woul d assert th e propositio n "p," o r declar e th e intentio n "p," wit h the ai m o f findin g agreement wit h an addressee are abstracted from.

As a rule, however, propositional sentences and intentional sen-

tences are embedde d i n illocutionary acts i n the for m o f assertions a n d announcements. The meanin g o f assertions and announce- ments, whic h are par t of the communicative use o f language, can be explicate d onl y pragmatically. Fro m this we can see that Haberma s

the forma l semantic approach to meaning, for he

acknowledges its ability to account for the meaning of noncommu -

nicatively used propositional an d intentional sentences. A t the same

challeng e th e claims o f formal-semanti c theorie s t o

explai n th e meanin g o f utterances suc h as assertions an d announce -

ments, or mor e generally, o f communicatively used linguistic expres- sions. Moreover, i f formal-semantic theories of meaning can account only for the noncommunicative use o f language, then their re- stricted scope suggests that this approach to meaning is itself limited.


does no t reject

time, h e doe s

We have ascertained that a pragmatic theory is required to expli- cate the meaning of communicatively used linguistic expressions. I t



remains unclear, however, i n what sense such a theory is pragmatic.

As indicated, i n his earlier essay o n universal

had justified his preference for the category of acceptability condi- tions, as opposed to truth or assertibility conditions, o n the grounds that i t avoids the prioritization of the assertoric mod e o f language

pragmatics, Habermas

use implicit i n the latter categories. I n these later writings, however,

assertibility conditions seems to go beyond

this. They are said to rest o n faulty pictures o f trut h and justification tha t fail t o recogniz e internal , conceptua l link s wit h pragmatic contexts of justification an d thu s remai n trappe d i n abstractiv e fallacies o f a

cognitivist and semanticist kind. I n Habermas's view, validity an d justification—and hence utterance meaning—are inescapably prag- matic notions. They cannot be explicated independendy of discur-

sive processes o f redeemin g differen t kind s o f validity claims. Whil e Dummett' s notio n o f assertibility conditions pushes i n the direction of a pragmatic account of justification an d validity, i t does not quite

arrive there; i t remains a

semantic theory to the extent that i t fails

t o explicat e these notion s as conceptually linke d t o discursive proc - esses o f redeemin g disputed—assertoric an d nonassertoric—validity claims.

his objection to trut h or

Habermas proposes that we understand the meaning of a speech

act whe n we kno w what makes

i t acceptable. We kno w what makes


speec h act acceptabl e whe n we kno w th e kinds o f


tha t


speaker can offer, i f challenged, i n order to reach understanding

with a hearer concerning the validity o f the disputed claim. I n every- day processes o f communication, the kinds of reasons that a hearer

must kno w i n order to understand

scribed contextually. Let us imagine a request to a passenger by an

airline steward to stop smoking. I n order to understand this request,

reasons that

the airline steward could provide i n order to justify his request, i f necessary. These reasons migh t include the argument that smoking is unpleasant for other passengers or that i t is against the regulations of the airline or against an international code o f airline practice. These reasons are of certain kinds. I f other kinds o f responses were offered as reasons—for instance, that i t is raining outside, or that Finnegans Wake is Jame s Joyce' s best book , o r tha t ther e are n o snakes

the passenger has to be able to reconstruct the kinds o f

a given utterance are circum-



i n Ireland—th e context i n question woul d render the m irrelevant

and, indeed, unintelligible. Thus, althoug h the set of reasons consti-

open-ended , i n

everyday contexts of communication contextual considerations act as a constraint o n the kinds o f reasons that are relevant to justification.

tutin g a give n kind o f reasons is always i n principl e

T h e hearer no t only has to kno w the kinds o f reasons the


coul d adduce i n

a given instance,



to kno w how



migh t use the m i n order to engage i n argumentatio n with a hearer concerning the validity of a disputed claim. This focus o n knowin g

reason s t o suppor t a dispute d validit y

claim clearly recalls Dummett' s epistemic turn. Like Dummett ,

Habermas also stresses that the principle be decided once an d


construed fallibilistically, that is, as always i n principl e subject to

revision i n light of new arguments based o n new evidence and

insights. Thi s is on e sense i n whic h th e questio n o f validity is tied t o pragmatic contexts o f justification, an d i t constitutes a furthe r rea- son for describing Habermas's theory of meaning (and, indeed,

Dummett's) as pragmatic. However, there is a second,

possibly mor e

h o w th e speake r migh t use


o f these reasons can never i n

for all. Rather, their validity must

contentious sense, i n whic h Habermas ties validity to pragmatic contexts of justification. I n this second sense, validity is no t only

always subjec t i n principl e t o discursiv e réévaluation , i t is in itself

pragmatic. The pragmatic dimension is no t something

the idea o f validity externally, as i t were; rather, i t is internal to the very concep t o f validity. A theory o f meanin g that sees itself as

pragmatic i n this stronger sense must

account of validity itself. To this extent, Habermas's pragmatic theo- ries of trut h (empirical and theoretical validity) and justice (moral validity)—and, indeed, his accounts o f ethical and aesthetic valid-

therefore offer a pragmatic

attached to

ity—are crucial ingredients

o f his pragmatic theory of meaning.

of mora l validity has bee n the subject of exten- criticism. Fro m the poin t o f view o f the theor y

sive commentar y an d

o f meaning, our question is the following: how is the conception o f mora l validity it proposes internally connected with processes of discursively redeeming validity claims? A nor m or principle is mor - ally valid (right or just), for Habermas, i f i t is the possible object of





a discursively achieved consensus to the effect that i t is equally i n the interest o f all affected. Therefore, agreement reached i n discourse— idealized rational acceptability—contributes constructively to the va- lidity o f mora l norms. I t is clear fro m this that Habermas conceives moral validity as internally linked to the idea of discursively achieved consensus and hence to pragmatic contexts o f justification.

Habermas also proposes a pragmatic theory of truth. Discussion of this is complicated by the fact that he significantly amended the

accoun t he originall y presented i n the 1973 essay, "Wahrheitstheo¬

fully revised version. How-

ever, a recent essay o n Richar d Rorty's neopragmatism (included

here as chapter 8) can be seen as an attempt to rectify this deficiency. For our present purposes, what is most interesting about these re- cent remarks is their continued insistence o n the pragmatic nature

of truth. Habermas associates himself with Rorty's ai m of radicalizing

the linguistic tur n withi n moder n philosophy by moving to a prag- matic level of analysis. H e criticizes him , however, for drawing the wron g conclusions fro m his critique of the philosophy of language. Rorty reduces truth to practices of justification, thus losing sight of the potential power of validity claims to explode actual contexts of justification. Habermas, by contrast, wants to hol d onto the momen t of unconditionality that is part o f the idea of truth, while retaining

an internal relation between truth an d justifiability. His aim, i n other words, is to wor k out a theory o f truth that is inherently pragmatic yet retains the idea o f an unconditional claim that reaches beyond all the evidence available to us at any given time. What woul d such

a theory look like? I n the 1980s, Habermas defended a view no t

unlike Hilary Putnam's conception of trut h as idealized rational

rien, " withou t subsequently presenting a


unde r conditions o f an ideal speech situation. Truth, on this ac- count, is a regulative idea, the anticipation o f an infinite rational consensus. I n th e recent essay, however, Haberma s acknowledge s

objections to this earlier conception. On e set of objec-


tions is directed against some conceptual difficulties with the very

acceptability: a proposition was said to be tru e i f i t coul d be

an ideal speech situation, i n particular, the paradox in - aimin g for "complete" or "conclusive knowledge." Th e

objection has been raised, for instance, that i t woul d be paradoxical

notio n o f volved i n



for huma n beings to strive to realize an ideal, the attainment of

whic h woul d be the en d o f huma n history. Anothe r set o f objections

draws attention to the difficulties involved i n conceptualizing

connection between truth and justified acceptability. O n the one hand, i f there is an unbridgeable gap between de facto an d ideal acceptability, the idea of an idealized rational consensus seems so far

removed fro m actual huma n practices of justification as to

min e the regulative role ascribed to it. O n the other hand, such a gap seems to be necessary i n order to preserve the intuition that trut h has a momen t of context-transcendence.

I n the face of these and other difficulties, Habermas n o longer conceives trut h as idealized rational consensus. H e now focuses o n

t h e idealizin g supposition s guidin g th e process o f rationa l argumen -

tatio n rathe r tha n o n th e idealizin g supposition s markin g its outcome.

T h e


They include the idealizing suppositions that participants are moti - vated only by the force of the better argument, that all competent

parties are entitled to participate o n equal terms i n discussion, that

n o relevant argument is suppressed or excluded, and so on. I t is

fro m such idealizations, whic h guide the process of argumentation, that the idea of truth draws its power as a regulative idea. This power is expressed i n the idea that a claim, i f true, could withstand all attempts to refute i t under ideal discursive conditions. The idea of trut h has a "decentering" functio n that serves to remin d us that what is currendy regarded as rationally acceptable may conceivably be called into question i n the future, as the limitations of our curren t understanding of argumentation becom e apparent.



forme r idealizations

pertain to the conduct of discourse rather

to the agreement to whic h participants i n discourse aspire.

It is importan t here to beware o f confusing Habermas's explica- tion of the idea of truth with an explanation of what makes a proposition true. The thesis that a proposition, i f true, can stand u p to attempts to refute it unde r the demandin g conditions of rational argumentation explicates the pragmatic meaning of truth. I t is not, however, an explanation of what makes the proposition true. As to the latter, Habermas's position is the standard one that a proposition is true i f and only i f its trut h conditions are satisfied. Althoug h we

c a n establish whethe r th e trut h condition s o f a give n propositio n are





satisfied only i n argumentation, their satisfaction or is no t itself an epistemic fact. Whereas, as we have

seen, idealized

rationa l acceptabilit y

o f

indicates th e trut h

constitutes th e validit y o f mora l norms , i t merel y propositions . Nonetheless , i t is clear fro m th e

foregoin g

that, o n

Habermas' s

account ,

th e concept

o f trut h mus t b e


pragmatically; we have n o access to truth except

by way

o f

a concept of validity explicated i n terms o f how we talk about truth,

that is, i n terms o f an idealized practice

A further concer n of Habermas's progra m of forma l pragmatics

use o f linguistic expressions is

the basic mod e o f language use o n whic h other modes, for example,

strategic o r fictional ones , are parasitic. Otherwise , i n ignorin g thes e other modes, the demonstration that everyday communicative ac- tion has an in-built connection with context-transcendent validity claims woul d be seriously limited. I n arguing for the derivative status of the strategic use of language, Habermas initially drew on Austin's distinction between illocutions an d perlocutions (see chapter 2) . I n response to criticisms of his interpretation o f this distinction, how- ever, Habermas subsequendy modifie d an d clarified his under- standing o f Austin' s categories (see chapters 3, 4, an d 7) whil e

continuin g

is to argue that the communicative

o f argumentation.

to insist that the strategic use o f language is parasitic o n

the use o f language with an orientation toward reaching under- standing. His argument for the parasitic status of the symbolic, the

figurative, an d th e fictional mode s o f languag e use is tha t th e every- day communicative use o f language fulfills indispensable problem- solving functions that require idealizing suppositions not demanded by the world-creating and world-disclosing use o f language charac- teristic for the aesthetic realm. Th e idealizing suppositions of, for example, consistency of meaning or a shared orientation toward

o f lan -

mutua l understandin g are suspende d i n th e fictional use

guage, and with these, the illocutionary bindin g and bondin g power o f everyday speech acts (see chapters 9 an d 10).




theory o f meaning

attempts to





relations between







ance is always to understand i t as an utterance i n a given situation,

whic h


i n whic h they are

tur n



embedded. For to understand an





i n

part of







Habermas draws attention to various kinds o f background knowl -

edge: for instance, knowledge o f the speaker's personal history or familiarity wit h the (culturally specific) contexts i n whic h a given topic is normall y discussed. These kinds o f knowledge, although usually only implicit i n acts o f understanding, are relatively close to the foregroun d and can be rendered explicit without difficulty. Thu s they can be contrasted with the deep-seated, prereflective, taken-for- granted background knowledge o f the lifeworld that, as a horizo n o f shared, unproblematic convictions, cannot be summone d to con- sciousness at will or i n its entirety. This background knowledge o f the lifeworld forms the indispensable context for the communicative use o f language; indeed without it , meanin g o f any kind woul d be

impossible. I t also functions tion tha t arises whe n a social

mechanisms o f communicative action. I t is thus a necessary comple- men t to Habermas's theories o f meanin g and communicative action

to absorb the risk o f social disintegra- orde r is reproduce d primaril y throug h


i n particular, chapters 2, 4, an d

8) .

T h e essays collected i n this antholog y were selected wit h the ai m o f providing general access to Habermas's treatment o f forma l prag- matics, fro m his earliest programmati c essay (chapter 1) to his most recent attempts to resolve some perceived problems with his ac- counts o f meaning and trut h (chapters 7 an d 8). Whereas, i n the

process o f

translating, revising existing translations, and retranslat-

ing, every effort has been made to ensure terminological consis- tency, n o attempt has been made to impose consistency o n the

argument s as they are presented i n the various essays. We have seen, for instance, that Habermas's earliest proposal for a pragmatic the- o r y o f meanin g differs i n some respects fro m his subsequent propos- als, an d that he himself has modifie d his distinction between

as initially drawn . I n later writings (see distinctio n withi n th e categor y o f Ver-

stdndigung betwee n a wea k an d stron g orientatio n towar d consensus ,

chapter 8) he takes o n boar d objections to the conceptio n

illocutions an d perlocutions chapte r 7) h e introduce s a

a n d (see

o f trut h hinte d at i n chapter 3 o f the present volume. Wit h the exception o f the last two pieces, whic h are no t directly concerne d wit h the question o f meaning , the anthology presents the essays i n

roug h chronology i n order to show developments and revisions; the

reader is encouraged

In chapter 1 we are introduced to forma l pragmatics as a research progra m aime d at reconstructing the universal validity basis o f speech. Th e procedure o f rational reconstruction is elucidated

empirical-analytic approaches and to

throug h reference bot h to

Kantian transcendental analysis. This is followed by a sketch o f a theor y o f speech acts, whic h diverges fro m Austin's an d Searle's theories i n several importan t respects, an d i n whic h speech acts are

characterized i n terms o f claims to validity.

to look out for them.

Chapters 2, 3, 4, and 6, thoug h situating forma l pragmatics i n

relation to Habermas's theory o f communicative action, focus o n the theor y o f meaning . Th e coordinatin g power o f speech acts is ex- plained throug h an account o f understanding utterance meaning i n

terms o f knowin g acceptability

o f

conditions. Thi s pragmatic theory

meanin g

is presented as an attempt to overcome the limitations o f


theories throug h drawing o n Karl Buhler's schema o f lan-

guage functions and o n speech-act theory. I n addition, a typology o f speech acts based o n their connection with one o f three universal validity claims is set u p i n chapter 2, formin g the background for

Habermas's discussion i n subsequent chapters.

worl d as a kin d o f deep-seated, implicit, background knowledge is

also introduce d i n chapter 2 and developed, i n particular, i n chapter

Th e concept o f life-

4. Habermas stresses the importance o f this concept, o n the one

hand, as a presupposition for understanding utterance meaning and, o n the other, as a risk-absorbing counterpoise to the potentially disintegrative effects o f action oriented toward reaching under- standing. Further, Austin's distinction between illocutions an d per-

locutions is a

Habermas to support his thesis that the strategic mod e o f language use is parasitic o n the communicative use. Thi s involves hi m i n discussion about the status o f simple imperatives (for example, threats), whic h as a type o f utterance no t apparently connected wit h validity claims, seem to undermin e his claim that strategic utterances have a derivative status.

thread runnin g throug h these chapters, and is used by

Chapter 5 is a critical discussion o f Searle's theory o f meaning as



fro m



1970s onwards.





problems attached to Searle's view, whic h he reads as a modified intentionalist one, arguing that his own pragmatic theory is better able to account for the meaning of, i n particular, imperatives and promises.

Chapter 7 responds to Herber t Schnadelbach's criticisms of Habermas's concept of communicative rationality. Accepting Schnadelbach's criticism that he has hithert o accorde d it a privi- leged position, Habermas now identifies three core structures of rationality; this leads hi m to make some new distinctions between different modalities of language use. One noteworthy modification here is his introduction of a distinction between action oriented toward reaching understanding i n a weaker sense and action ori- ented toward agreement i n the strict sense and, correspondingly, between weak and strong communicative action. Some implications of these distinctions for the theory of meaning are also discussed.

Chapter 8 examines Richard Rorty's neopragmatism, interpreted by Habermas as an attempt to carry the linguistic tur n through to its conclusion, and criticizes i t for its assimilation of truth claims to



Chapte r 9 focuse s o n the relatio n betwee n the fictional o r poeti c use of language and language as i t is used i n everyday communica - tive action; i t criticizes Derrideans for faulty accounts of everyday and poetic language, for a consequent problematic leveling of the distinction between literature and communicative action, and for a failure to appreciate the distinctive mediating roles of philosophy and literary criticism.


theory of communicative action. Against Rorty, he defends his view of philosophy as guardian of reason, while acknowledging that this rol e mus t be define d i n a new way. H e the n clarifies his positio n wit h respect to moder n art and the validity claims connected with it, reaffirms his position that interpretive understanding inescapably involves evaluation, clarifies his idea of the unity of reason as an interplay of validity dimensions, and concludes with a discussion of the objection that his theory concentrates on justice at the expense of happiness.

I n








o f




1. Verständigung (n.): "reaching understanding," "mutual understanding," or "com-

munication. The corresponding verb is sich verständigen. As Habermas acknowl- edges, this term is ambiguous even in German. Although it embraces linguistic comprehension (Verstehen), it goes beyond this to refer to the process of reaching understanding, in the sense of reaching an agreement with another person or per- sons. However, despite having previously used the two terms interchangeably, Haber- mas now distinguishes between Verständigung and Einverständnis, agreement or consensus in the strict sense (see chapter 7). Finally, Verständigung can also be used as a synonym for "communication"; thus, for example, communicative rationality is occasionally rendered by Habermas as Verständigungsrationaätät.


What Is Universal Pragmatics? (1976)


T h e task o f universal pragmatics is to identify and reconstruct uni - versal condition s o f possibl e mutua l understandin g (Verständigung). 1 I n other contexts, one also speaks o f "general presuppositions of

communication, " bu t I prefer to speak of general presuppositions

action aimed at


assumption (without undertaking to demonstrate it here) that other forms of social action—for example, conflict, competition, strategic action i n general—are derivatives of action oriented toward reaching understandin g (Verständigung). Furthermore , sinc e languag e is th e specific mediu m of reaching understanding at the sociocultural stage of evolution, I want to go a step furthe r and single out explicit speech actions fro m other forms of communicative action. I shall ignore nonverbal actions and bodily expressions. 2


communicative action because I take the type o f reaching understanding to be fundamental. Thus


start fro m

T h e Validity Basis o f


Karl-Otto Ape l proposes the following formulatio n i n regard to the general presuppositions o f consensual speech acts: to identify such presuppositions we must, he thinks, leave the perspective o f the observer o f behavioral facts and call to min d "what we must neces- sarily always already presuppose i n regard to ourselves an d others as


Chapter 1


a n d i n this sense, what we mus t necessarily always already

accepted." 3 Ape l her e uses th e aprioristi c perfec t (immer schon: always


transcendental constraint to whic h we, as speakers, are subject as soon as we perfor m or understand or respond to a speech act. I n or after the performance of this act, we can becom e aware that we have involuntarily made certain asssumptions, whic h Apel calls "norma-

tive condition s o f th e possibilit y o f reachin g understanding. " Th e adjective "normative" may give rise to misunderstanding. On e can say, however, tha t th e genera l an d unavoidable — i n this sense tran - scendental—conditions of possible mutua l understanding have a

dimen -

normative content when one thinks not only of the validity

sion of norms of action or evaluation, or even of the validity dimen - sion of rules i n general, bu t also of the validity basis of speech across its entire spectrum. As a preliminary, I want to indicate briefly what



the possibility of










mod e


necessity i n order

I mean


must, i n performin g any speech act, raise universal validity claims

a n d



wants to participate i n a process of reaching understanding,

cannot avoid raising the following —a n d indeed precisely the follow-

by the "validity basis of


tha t



the y ca n

b e



I shall



suppos e

vindicate d

(einlösen). Insofa r


ing—validity claims. She

claims to



utterin g somethin g




givin g (th e hearer )

something t o

understand ,

c. makin g herself thereb y understandable ,

d .

an d

another person.

comin g

t o a n understandin g wit h

T h e

speake r mus t choos e a n intelligibl e (verständlich)

expressio n

speake r

mus t have th e intentio n o f communicatin g a tru e (wahr)

prepositional content, the existential presuppositions o f whic h

are satisfied) so tha t th e heare r ca n share the knowledge o f th e speaker.

T h e haftig)



speake r



propositio n

tha t speake r an d heare r


speake r

mus t wan t

t o


tha t

th e

heare r


ca n

comprehend one another. Th e


expres s he r

ca n


th e

intention s truthfully

utteranc e

o f

th e


trust her) . Finally, the speaker must choose


What Is Universal Pragmatics?

anc e tha t is righ t (richtig) wit h respec t t o prevailin g norm s an d values

speaker an d another wit h

respect to a recognized normative background. Moreover, commu - nicative action can continue undisturbed only as long as all partici-

pants suppose that the validity claims they reciprocally raise are raised justifiably.

ai m o f reachin g understandin g (Verständigung) is to brin g

abou t a n agreemen t (Einverständnis) tha t terminate s i n th e intersub -

jective mutuality of reciprocal comprehension, shared knowledge, mutual trust, and accord with one another. Agreement is based o n recognition o f the four corresponding validity claims: comprehensi- bility, truth , truthfulness , an d Tightness. We can see that the word "Verständigung" is ambiguous . I n its narrowes t meanin g i t indicates

so that the hearer can accept the utterance, an d bot h heare r can , i n th e utterance , thereb y agree with one

T h e

that two subjects understand

i n its broadest meaning it indicates that an accord exists between two

subjects concerning th e Tightness o f a n utterance i n relation t o a

a linguistic expression i n the same way;

I n addition, the partici-

pants i n communicatio n can reach understanding about something i n the world , and they can make their intentions understandable to one another.

I f full agreement, embracing all four o f these components, were

a norma l state of linguistic communication, i t woul d no t be neces-

sary to analyze the process of reaching understanding fro m the dynami c perspective o f bringing about a n agreement . Th e typica l

states are i n the gray areas between, o n the one hand , lack o f understanding an d misunderstanding, intentional and involuntary untruthfulness, concealed and ope n discord, and, o n the other hand, preexisting or achieved consensus. Reaching understanding is the process o f bringin g about an agreement o n the presupposed

basis of validity claims that are mutually recognized. I n everyday life,

we start fro m

tations taken for granted among participants. As soon as this con - sensus is shaken, an d as soon as the presupposition that the validity claims are satisfied (or could be vindicated) is suspended i n the case of at least one of the four claims, communicative action cannot be continued.

mutually recognized normative background.

a background consensus pertaining to those interpre-


Chapter 1

T h e task of mutua l interpretation, then, is to achieve a new defini-

tion o f the situation that all participants can share. I f this attempt fails, one is basically confronted with the alternatives of switching to strategic action, breaking off communicatio n altogether, or recom- mencing action oriented toward reaching understanding at a differ- ent level, the level of argumentative speech (for purposes of discursively examining the problematic validity claims, whic h are now regarded as hypothetical). I n what follows, I shall take int o

aside bot h dis-

course an d strategic action. I n communicative action, participants presuppose that they kno w what mutua l recognition of reciprocally raised validity claims means. I f i n addition they can rely o n a shared definition o f the situation a n d thereupon act consensually, the background consensus includes the following:

consideration only consensual speech acts, leaving


Speaker and hearer know

implicitly that each o f the m has to raise




claims i f there is to be

communicatio n



(i n









standing) .

b. Bot h reciprocally suppose that they actually do satisfy these pre-

suppositions o f communication, that is, that they justifiably raise

their validity claims.

c. This means that there is a commo n

claims raised either are already vindicated, as i n the case o f the comprehensibility of the sentences uttered, or, as i n the case o f


that any


truth , truthfulness , an d


coul d be vindicate d becaus e th e

sentences, propositions,


intentions, an d utterances satisfy

the corresponding adequacy conditions.

Thu s I distinguis h (i ) th e conditions fo r th e validity o f a grammati - cal sentence, true proposition, truthful intentional expression, or normatively correct utterance appropriate to its context fro m (ii) the claims wit h whic h speakers deman d intersubjectiv e recognitio n for the well-formedness o f a sentence, trut h of a proposition, truth- fulness of an intentional expression, and Tightness of a speech act, as wel l as fro m (iii ) th e vindication o f justifiabl y raise d validit y claims . Vindication means that the proponent , whether throug h appeal to


What Is Universal Pragmatics?

intuitions and experiences or throug h arguments and action conse-

quences, justifies the claim's worthiness to be recognized and brings abou t a suprasubjective recognitio n o f its validity. I n accepting a valid - ity claim raised by the speaker, the hearer recognizes the validity of the symbolic structures; that is, he recognizes that a sentence is

grammatical, a statement true, an intentional expression

an utterance correct. The validity o f these symbolic structures is

truthful, or


satisfy certai n adequac y condi -

tions ; bu t th e meanin g o f th e validit y consists i n thei r worthiness t o be recognized that is, i n the guarantee that intersubjective recogni-


proposed the name "universal pragmatics" 5 for the research

validity basis of

progra m aimed at reconstructing the universal

speech. 6 I woul d now like to delimi t the theme

progra m i n a preliminar y way. Thu s befor e to the theor y o f speech acts, I shall prefix

b y virtu e o f

th e fact tha t the y

ca n

b e

brough t abou t unde r suitabl e

conditions. 4

I have

o f



passing o n (i n par t II )

a few guidin g remarks

dealing with (i) an initial delimitation o f the object domai n of the proposed progra m of universal pragmatics; (ii) an elucidation o f the procedure o f rational reconstruction, as opposed to an empirical- analytic procedure i n the narrower sense; (iii) a few methodological

the status of

a reconstructiv e science ; an d finally (iv) th e questio n o f whethe r th e proposed universal pragmatics assumes the status o f a transcenden- tal theory o f reflection or that o f an empirically substantive recon- structive science. I shall restrict myself to guidin g remarks because, while these questions are fundamental and deserve to be examined independently, they for m only the context o f the topic I shall treat

a n d must thus remai n i n the

difficulties resulting fro m the fact that linguistics claims


Preliminary Delimitation of the Object

Domai n

I n several o f his works, Ape l has pointe d to the abstractive fallacy

that underlies the approach to the logic of science favored by con- temporary analytic philosophy. 7 Th e logical analysis of language that originated with Carnap focuses primarily o n syntactic and semantic properties of linguistic formations. Like structuralist linguistics, i t delimit s its objec t domai n by first abstractin g fro m th e pragmati c


Chapter 1

properties o f language, and subsequentiy introducin g the pragmatic dimension i n such a way that the constitutive connection between

the generative accomplishments o f subjects capable o f speaking an d acting, o n the one hand, an d the general structures o f speech, o n the other, cannot come into view. I t is certainly legitimate to draw an abstractive distinction between language as structure and speak- i ng as process. A language will the n be understood as a system o f rules for generating expressions, such that all well-formed expres-


the other hand , subjects capable o f speaking can employ such ex-

pressions as participants i n a process o f communication; for in -

they can utter sentences as well as understand the m an d

respon d t o them . Thi s abstractio n o f language fro m th e use o f lan -

(langue versus parole), whic h is mad e i n bot h th e structuralist analysis o f language, is meaningful.

Nonetheless, this methodological step is no t sufficient reason for the view that the pragmatic dimension o f language fro m whic h one abstracts is beyond forma l (or linguistic) analysis. A n abstractive fallacy arises i n that the successful, or at least promising, reconstruc-

linguistic rul e systems is seen as justification for restricting

formal analysis to this object domain. Th e separation o f the two

analyti c levels, language an d speech, shoul d no t be mad e i n suc h a way

left to exclusively em-

pirical analysis—that is, to empirical sciences such as psycholinguis- tics an d sociolinguistics.

(e.g., sentences)

may count

as elements

o f this language. O n


guag e

i n


logical an d the

tion o f

that the pragmatic dimension o f language is

I woul d like to defend the thesis that no t only language bu t speech

too —tha t

ble to forma l analysis. Like the elementary units o f language (sen-


fro m the methodological stance o f a reconstructive

Approaches to a general theory o f communicatio n have been developed fro m the semiotics o f Charles Morris. 8 I n their framework o f fundamental concepts they integrate the mode l o f linguistic be-


is, the employment o f sentences i n utterances—is

the elementary units o f speech (utterances)

can be



haviorism (the symbolically mediated stimulated individual organism) with

transmission (encoding and decoding signals between sender and receiver for a given channel and an at least partially commo n store

behavioral reaction o f the the model o f informatio n


What Is Universal Pragmatics?

o f signs). I f the speaking process is conceptualized i n this way, the

fundamental question o f universal pragmatics concerning the gen-

era l condition s o f possible mutua l understandin g (Verständigung) cannot be pose d i n an appropriate way. For example, the intersub-

meanings that are identical for at least two speakers does

n o t even becom e a proble m

to extensionally equivalent

don e i n linguistic behaviorism, 9 or (ii) i f i t is preestablished at the analytic level that there exists a commo n code an d store o f signs between sender an d receiver, as is don e i n informatio n theory.

I n additio n to empiricist approaches that issue, i n one way or another, fro m the semiotics o f Morris, there are interesting ap- proaches to the logical analysis o f general structures o f speech an d action. Th e following analyses can be understood as contributions along the way to a universal pragmatics. Bar-Hillel pointe d out quite early the necessity for a pragmatic extension o f logical semantics. 1 0

(i ) i f th e identit y o f meaning s is reduced classes o f behavioral properties, as is

jectivity o f

Als o o f not e are th e proposal s fo r a deontic logic (Hare , H . vo n Wright ,

N. Rescher) 1 1 and corresponding attempts at a formalization o f

speech acts such as assertions an d questions (Apostel); 1 2 approaches

to a logic o f nondeductive argumentation (Toulmin , Botha) belong

the side o f linguistics, the investigation o f pre- Petöfi), 1 4 conversational postulates (Grice, Lak-

off), 1 5 speech acts (Ross, McCawley, Wunderlich), 1 6 and dialogues

a n d texts (Fillmore, Posner) 1 7 lead to a consideration of the prag-

matic dimension o f language fro m a reconstructionist point o f view. T h e difficulties i n semantic theory (Lyons, Katz) poin t i n the same

direction. 1 8 Fro m th e side o f formal semantics, i n particula r th e dis- cussion—going back to Frege and Russell—of the structure o f propositions, o f referential terms an d predicates (Strawson) 1 9 is sig-

nifican t fo r a universa l pragmatics . Th e sam e

theory (Danto , Hampshire , Schwayder) 2 0 and for the discussion that

o f inten- of mean-

i ng introduce d by Wittgenstein has universal-pragmatic aspects (Alston), 2 2 as does the attempt by Grice to trace the meaning o f sentences back to the intentions o f the speakers (Bennett, Schif- fer). 2 3 As the most promising poin t o f departure for a universal

hold s fo r analytic action

here as well. 1 3 Fro m suppositions (Kiefer,

has arisen i n connection with the logic o f the explanation tional actio n (Winch , Taylor, vo n Wright). 2 1 Th e use theory


Chapter 1

pragmatics, I shall draw primarily o n the theory of speech acts



lytic philosophy o f language have the commo n goal of clarifying processes o f language use fro m the viewpoint o f forma l analysis.

with regard to the contributio n they


ent . I n man y cases I see a dange r tha t th e analysis o f condition s o f possible mutua l understanding is foreshortened, either

However, i f one evaluates the m make to a universal pragmatics,

by Austin (Searle, Wunderlich) , 2 4 approaches developed fro m logic, linguistics, an d the


their weaknesses also becom e

a. because these approaches do no t generalize radically enough and

do no t push throug h the level of fortuitous contexts to general and unavoidable presuppositions—as is the case, for instance, with most of the linguistic investigations of semantic and pragmatic presuppo- sitions; or

b. because they restrict themselves




the instruments developed i n are inadequate for capturing



an d

grammar, even whe n




i n syntactic


the performative character

c. because they mislead one

o f speech acts; 2 5


int o a formalization of basic



What Is Universal Pragmatics?

a m no t usin g forma l analysis i n a sense tha t refers, say, t o th e stan- dard predicate logic or to any specific logic. Th e tolerant sense i n whic h I understand forma l analysis can best be characterized

rational recon-

struction o f

an d schemata. Thus we speak

throug h the methodological

attitude we adopt i n the

concepts, criteria, rules,

o f the explication o f meanings an d concepts, o f the analysis o f presuppositions an d rul e systems, an d so forth. O f course, recon-

structive procedures are also importan t for empirical-analytic re-

search, for example, for explicating frameworks o f basic concepts, for formalizing assumptions initially formulate d i n ordinary lan-

particular hypothe- so on . Nonetheless , o f sciences that de-



knowledge of competent


velop nomological


guage, for clarifying deductive relations amon g ses, fo r interpretin g results o f measurement , an d

procedures are no t characteristic

hypotheses about


of observable



rather, these procedures are



ence s tha t systematically reconstruct the intuitive subjects.

I n clarifying the distinction between empirical-analytic and recon- structive sciences, I woul d like to begin with the distinction between

sensor y experienc e o r observation an d communicativ e experienc e o r

that have no t been satisfactorily analyzed—as can, i n my view, be

understanding ( Verstehen). Observatio n is directe d towar d

perceptibl e

shown i n the case o f the logics of norms whic h trace norms of action

things an d events (or states); understanding is directed

toward the

back to commands;



meaning of utterances. 2 7 I n experiencing, the observer is i n


d. becaus e they start from th e mode l o f th e isolated, purposive -

ple alone, even i f the categorial net i n whic h experiences are


rational actor and thereby fail—as do, for instance, Grice an d Lewis 2 6 —t o reconstruct i n an appropriate way the specific momen t o f mutuality i n the understanding of identical meanings or i n the

recognition of intersubjective validity claims.

It is my impression

that the theory of speech acts is largely free

these and similar weaknesses.

Some Remarks o n the Procedure

of Rational



ized as experiences laying claim to objectivity is always already shared

by several (or even all) individuals. I n contrast, the interpreter wh o understands meaning undergoes her experiences fundamentally as

a participant i n communication, o n the basis o f an intersubjective

relation established throug h symbols with other individuals, even i f she is i n fact alone with a book, a document, or a wor k of art. I shall

relationship between observation and

understanding any further; I woul d like to direct attention to just one aspect o f this: the difference i n level between perceptible reality a n d the understandable meaning o f a symbolic formation. Sensory

n o t here analyze the complex


have been employing the expression "formal analysis" i n opposi-


is related to segments o f reality withou t mediation, com-

tion to empirical-analytic procedures (in the narrower sense) with-


experience only mediately, as illustrated i n the diagram

o ut providing a detailed explanation. This is, at least, misleading. I




Chapter 1

What Is Universal




Level 2

Level 3

Observabl e






I - Observation sentence










This diagram represents

three different relationships:

a. Epistemic relations between


sion (here o f the observation sentence), i n a way similar to how the

act o f observation relates to the objects an d events

b. Relations of representing an aspect o f reality i n a propositional

sentence. I n this sense, the interpretation represents the semantic

content (here o f the observation sentence), i n a way similar to how

the observation sentence represents certain objects and events.

c. Relations o f expressing




this sense, the act of understanding relates to the symbolic

experiential acts an d their objects. I n


intentional acts. I n this sense, the





o f



i n

T h e two pairs o f concepts—"perceptible reality" versus "symboli-

cally prestructured reality" and "observation" versus "under-

standing"—can be correlated with another pair: "description" versus

"explication." Wit h the aid o f a sentence that represents an observa-


sentence that represents an interpretation o f the

boli c formation , I ca n explicate th e meanin g o f suc h a n utterance .

formatio n is un -

clear does the explication need to be set of f as an independent

that we use to describe objects

and events, there can be a lack o f clarity at various levels. Depending

o n the level, we deman d explications o f different kinds. I f the phe-

nomeno n described is i n need o f explanation, we deman d an expli-

cation that makes clear how reality operates and how the

phenomeno n i n question comes about. If, by contrast, the descrip-

tion itsel f is

makes clear what the observer meant by his utterance and how the

symboli c expressio n i n nee d o f elucidatio n come s about . I n th e first

case, a satisfactory

explication will have the for m o f an explanation

Naturally, only whe n the meaning o f the symbolic



ca n

describe th e observe d aspec t o f reality. Wit h

th e

ai d

o f

meaning o f a sym-

analytic step. I n regard to sentences

incomprehensible , we deman d a n explicatio n tha t

propositional content

o f the interpretation, just as the


we undertake with the aid o f a causal hypothesis. I n the second case,



i n




o f



we speak o f explication o f meaning. (O f course, explications o f


Apar t fro m the fact that all three types o f relation simply poin t to

fundamental problems, there is an additional difficulty i n specifying

the precise differences between the epistemic relations o f the ob-

server an d the interpreter to their respective objects and between

the representational relations o f the observation sentence to reality,

o n the one hand, and that o f the interpretation sentence to (sym-

bolically prestructured) reality, o n the other. This specification

woul d require a comparison between observation and interpreta-

tion, between description and explication. For the time being, the

diagram is intended merely to illustrate the two levels o f reality to

whic h sensory and communicative experience respectively relate.

T h e difference i n level between perceptible and symbolically pre-

structure d realit y is reflecte d i n th e ga p betwee n direct access


throug h observatio n of realit y an d communicativel y mediate d

throug h understandin g a n utteranc e concerning reality.

meanin g need no t be limited to descriptive sentences; any meaning-

fully structured formatio n can be subjected to the operation o f

meaning explication.)

Descriptions and explications have different ranges; they can be-

gi n

o n the surface and push throug h to underlying structures. We


familiar with this fact fro m the explanation o f natural phenom -

ena—the mor e general the theories are with whic h we explain natu-

ral phenomena, the more penetrating the corresponding theoretical

descriptions. Th e same is true o f explications o f meaning. O f course,

i n the case o f meaning explications, the range o f explication does

n o t depend o n the level

the structures o f an external reality accessible to observation bu t

o n knowledge

understanding— a reality o f symbolic formation s produce d ac-

cording to rules. Th e explication o f natural phenomena pushes

i n a different direction fro m the explication o f the meaning o f


o f the deep structures o f a reality accessible to

o f generality o f theoretical knowledge about


Chapter 1


I want


distinguish two





meaning. I f the meaning of a written sentence, action, gesture,

wor k

of art, tool, theory, commodity, transmitted document, and so

o n is

o f meanin g is directe d first t o th e semanti c

content of the symbolic

tent, we take u p the same position as the "author" adopted

wrote the sentence, performe d the gesture, used the tool, applied


formation . I n trying to understand its con-

unclear , th e explicatio n

whe n

the theory,

an d



Often ,



must go


wha t was



intended by




take int o consideration


contex t o f whic h he was

derstanding of content pursue s connection s tha t lin k th e surfac e struc -

no t conscious. 2 8 Typically, however, the un-

tures of the incomprehensible formatio n with the surface structures

o f other, familiar formations. Thus, linguistic expressions can be

throug h


i n

translation int o expressions of another language; i n bot h cases, com -

throug h







speakers draw o n


know n




obtain withi n the lexicon

of one


or between

the léxica




I f


canno t

attai n he r en d

i n this way, th e interprete r ma y


it necessary to alter her attitude. She the n exchanges the attitude of


whic h she, as i t were, looks throug h symbolic formations to the worl d


focuses o n the generative structures of the

T h e


author must have produced

tion, th e interprete r draws o n semanti c meanin g relation s


manner, so to speak, i n that she

wit h competent speakers o f that language. I n this sense, the role of interpreter can (under suitable conditions) be attributed to the

soon as the inter-

prete r tries no t onl y t o apply thi s intuitiv e knowledg e o f speakers bu t

o f

t h e symboli c formation ; she n o longe r look s throug h i t intentione

symboli c











whic h

uttered—fo r

attitude i n whic h


interpreter then attempts to

formatio n with



o f


the meaning of a sym-

to whic h


the rules according

it . I n norma l paraphrase an d transla-




i n


a knowledge

(fo r in -






different words

simply applies

author himself. Th e attitude changes, however, as

t o reconstruct it . She

the n turn s away fro m

recta t o

th e world . She

attempt s instea d

t o

th e surface

structur e

pee r int o th e


What Is Universal Pragmatics?

formation—penetrating throug h its surface, as i t were—i n order to discover the rules according to whic h this symbolic formatio n was

produced (i n our example, the rules according to whic h the lexicon of a language is constructed). The object o f understanding is n o

longe r th e

content o f a symboli c expressio n o r wha t specific author s

mean t b y i t i n specifi c situations bu t rathe r th e intuitiv e rule conscious-

ness tha t a competen t speake r has o f hi s ow n

language .

distinguish between understand s ho w t o

produc e o r accomplis h something , an d know-that, th e explici t knowl - edge o f how i t is that he is able to do so. I n ou r case, what the autho r

means by an utterance and what an interpreter understands o f its

conten t are a first-level know-that . T o th e exten t tha t his utteranc e is correctly forme d and thus comprehensible, the author produced it i n accordance with certain rules or o n the basis of certain struc- tures. H e knows how to use the system o f rules o f his language and understands their context-specific application; he has a pretheoreti- cal knowledge o f this rul e system, whic h is at least sufficient to

rul e

consciousness is a know-how. The interpreter, i n turn, who no t only shares bu t wants to understand this implicit knowledge of the com- petent speaker, must transform this know-how int o explicit knowl - edge, that is, int o a second-level know-that. This is the task o f reconstructive understanding, that is, o f meaning explication i n the sense o f rational reconstruction of generative structures underlying the productio n o f symbolic formations. Since the rule consciousness to be reconstructed is a categorial knowledge, the reconstruction depend s first o f al l o n th e operatio n o f conceptua l explication .

enable hi m to produce the utterance i n question. This implicit

Following a suggestion

made by Ryle, 2 9 we can

know-how, th e abilit y o f a competen t subjec t wh o


pu t forwar d four



concept must fulfill i n order to be


the explication o f


i. Th e explican s shoul d be similar t o th e explicandum , tha t is, fro m now o n the explicans should be able to be used i n place o f the explicandum i n all relevant cases.










o f



(i n

connectio n

wit h othe r

scientifi c concepts )

i n

a n


manner .


Th e

explican s

shoul d

prov e

t o



wit h

respec t

t o

th e

formulation o f general



Chapter 1


shoul d b e


that requirement s i-ii i can be

possible. 3 0

as simple as

met )



Wunderlic h sums u p his reflections o n the status of concept expli-

cation as follows:

Explication always proceeds (i n conformity with Carnap's requirements i-iv) -with regard to theories; either such centra l concepts (as "meaning") are explicated that entire theories correspond to them as explieans, or different concepts are explicated interconnectedly.

We explicate always with regard to clear cases, so as to be able (i n connectio n

with these) to replace our intuitions with exact arguments. However, the theory can then also provide answers to borderline cases; or we explicate separately what a clear borderline case is.

the same level as the explicandu m lan- standardized version derived from it) .

Accordingly, it is not a question here of a descriptive language or a metalan- guage relative to the language of the explicandum (the explieans does not describe the explicandum) . 3 1

Th e language o f explication is at guage (e.g., ordinary language or a

I n these reflections o n the explication o f concepts, one poin t

strikes m e as insufficientl y worke d out—th e evaluative accomplishments

of rule consciousness. Reconstructiv e proposal s ar e directe d towar d

domain s o f pretheoretica l knowledge, tha t is, no t t o jus t any implici t

opinion , bu t to a proven intuitive preknowledge. Th e

rul e conscious-

ness o f competent speakers functions as a court o f

evaluation, for

instance with regard to the grammaticality o f sentences. Whereas

understanding of content is directed toward any utterance whatever,

reconstructive understanding refers only to symbolic objects charac-

terized as "well formed " by competent subjects themselves. Thus, for


example, syntactic theory, propositional logic, the theory of science,

a n d ethics start with syntactically well forme d sentences, correctly

fashioned propositions, well-corroborated theories, and morally un -

objectionable resolutions o f nor m conflicts, i n order to reconstruct


rules according to whic h these formations can be produced. To


extent that, as i n the following examples, universal validity claims

(the grammaticality of sentences, the consistency o f propositions,

the trut h o f hypotheses, the Tightness o f norms o f action) underli e

intuitive evaluations, reconstructions relate to pretheoretical knowl-

edg e o f a genera l sort, t o universal capabilities, an d no t merel y t o


What Is Universal Pragmatics?

particular competencies o f individual groups (e.g., the ability to

utter sentences i n a Low-German dialect or to solve problems i n

quantu m physics) or, indeed, to the ability o f particular individuals

o f

the twentieth century). Whe n the pretheoretical knowledge to be

reconstructed expresses a universal capability, a general cognitive,

linguistic, or interactive competence (or subcompetence), then what

begins as an explication o f meaning aims at the reconstruction o f

species competencies. I n scope and status, these reconstructions can

be compared with general theories. 3 2

(e.g., t o writ e a n exemplar y Entwicklungsroman

even i n th e middl e

have developed this idea i n the

case o f grammatica l theor y (fo r th e first time i n Syntactic Structures,

1957). Roughly speaking, i t is the task o f grammatical theory to

reconstruct the intuitive rul e consciousness commo n to all compe-

tent speakers i n such a way that the proposals for reconstruction

represent the system o f rules that permits potential speakers, i n


least one language L , to acquire the competence to produce an d to

understand any sentences that count as grammatical i n L , as well


to distinguish these sentences well-formed i n L fro m ungrammatical

It is the great merit o f Chomsky


sentences. 3 3


versus Empiricist Linguistics


hop e I have sufficiently characterized the reconstructive procedure


f sciences that transform a practically mastered pretheoretical

knowledge (know-how) o f competent subjects into an objective an d

explicit knowledge (know-that), so that i t is clear i n what sense I a m

using the expression "formal analysis." Before mentionin g some

methodological difficulties with reconstructive linguistics, I woul d

like to contrast, i n broad strokes, two versions o f linguistics, one

empirical-analytic an d one reconstructive. (Wunderlic h speaks o f an

empirical-descriptive and an empirical-explicative linguistics. 3 4 ) I

compare bot h approaches unde r four headings.




throug h observation alone, the data o f linguistics consist o f meas-


the extent that the experiential basis is supposed to be


Chapter 1

ure d variables of linguistic behavior. By contrast, insofar as recon-

structive understanding is permitted , the data are provided by the

speakers, maieutically ascertained

(i.e., throug h suitable questioning with the aid o f systematically or-

dered examples). Thus the data are distinguished, i f you will,

i ng to their ontological level: actual linguistic behavior is

perceptible reality, and rule consciousness points to the productio n


consciousness of





of symbolic formations i n whic h something is uttered about reality. 3 5

Furthermore , observations always mean a knowledge o f something

particular, whereas rule consciousness contains categorical knowl -

edge. Finally, observational

other case, competent

possible data fro m the

data are selected only fro m the analytic

viewpoints o f the linguist, whereas, i n the

speakers themselves evaluate an d preselect

poin t o f view of their grammatical well-formedness.

Theory and Object Domain As lon g as natural languages count as the object of linguistic

tion an d no t as th e for m o f representatio n o f a reconstructibl e


pretheoretical knowledge, linguistic theory relates to its object do-

mai n as a causal-analytic theory that explains linguistic descriptions

of linguistic reality with the aid of nomological hypotheses. If, o n the

contrary, linguistic theory is supposed to serve to reconstruct pre-

theoretical knowledge, theory relates to its object domai n as an

of meaning to its explicandum. Whereas i n the empiri-

cist version the relation of linguistic theory to the language to be

explained is basically indistinguishable fro m that between theory

and reality i n other nomological sciences, i n the

the linguistic character of the object necessitates a relation that can

hol d only between different linguistic expressions: the relation be-

tween explication and explicandum, whereby the language o f expli-

explicative version


cation (that

is, the construct language of linguistics, whic h is a


version of ordinar y language)

belongs i n principle


the same level as the natural language to be explicated. (Neither i n

the empiricist no r i n the explicative case of theory formatio n can

the relation of linguistic theory to its object domai n be conceived as

that of metalanguage to object language. 3 6 )


What Is Universal Pragmatics?

Theory and Everyday


There is yet another peculiarity arising fro m these differendy ori-

conceptualizations. A n empirical-analytic theory i n the narrow

sense can (and as a rule will) refute the everyday knowledge o f an

object domai n that we initially possess prio r to science and replace

it with a correct theoretical knowledge regarded provisionally as

true. A proposal for reconstruction, by contrast, can represent pre-


theoretical knowledge mor e or less explicitly an d adequately, bu t i t

can never falsify it. A t most, the representation of a speaker's intui-

tion ca n prov e t o be false, bu t no t th e intuitio n itself. 3 7 The latter

belongs to the data, and data can be explained bu t no t criticized. A t

most, data can be criticized as being unsuitable, that is, either erro-

neously gathered or wrongly selected for a specific theoretical


T o a certai n extent , reconstruction s mak e a n essentialist claim . On e

c a n say, o f course , tha t theoretica l description s "correspond " (i f

true) to certain structures of reality i n the same sense as recon-

structions "bear a likeness" (if correct) to the deep structures expli-

cated. O n

descriptive theory and its object admits many epistemological inter-

pretations apart fro m the realistic (e.g., instrumentalist or conven-

tionalist) ones . Rationa l reconstructions , b y contrast , can reproduc e

the pretheoretical knowledge that they explicate only i n an essential-

ist sense; i f they are true, they have to correspond precisely to the

rules that are operatively effective i n the object domain—tha t is, to

the rules that actually determine

tures. 3 8 Thus Chomsky's correlation assumption, according to whic h

linguistic grammar is represented o n the part o f the speaker by a

menta l gramma r tha t correspond s exactl y t o it , is, at least i n th e first

instance, consistent.

the other hand, the asserted correspondence between a

the productio n o f surface struc-


To be sure, serious methodological difficulties have arisen i n con-