This article was downloaded by: [University of Sydney]On: 14 March 2013, At: 02:40Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,37-41 Mortimer Street, London W1T 3JH, UK

Inquiry: An Interdisciplinary Journal of PhilosophyPublication details, including instructions for authors and subscription information:http://www.tandfonline.com/loi/sinq20

Towards a theory of communicative competenceJürgen Habermas aa J. W. Goethe University, Frankfurt a. M.Version of record first published: 29 Aug 2008.

To cite this article: Jürgen Habermas (1970): Towards a theory of communicative competence, Inquiry: An InterdisciplinaryJournal of Philosophy, 13:1-4, 360-375

To link to this article: http://dx.doi.org/10.1080/00201747008601597

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions

This article may be used for research, teaching, and private study purposes. Any substantial or systematicreproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form toanyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contentswill be complete or accurate or up to date. The accuracy of any instructions, formulae, and drug doses shouldbe independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims,proceedings, demand, or costs or damages whatsoever or howsoever caused arising directly or indirectly inconnection with or arising out of the use of this material.

Inquiry, 13, 360-75

TOWARDS A THEORY OFCOMMUNICATIVE COMPETENCE

Jürgen Habermas

J. W. Goethe University, Frankfurt a. M.

In this, the second of two articles outlining a theory of communicative competence,the author questions the ability of Chomsky's account of linguistic competence tofulfil the requirements of such a theory. 'Linguistic competence' for Chomsky meansthe mastery of an abstract system of rules, based on an innate language apparatus.The model by which communication is understood on this account contains threeimplicit assumptions, here called 'monologism', 'a priorism', and 'elementarism'.The author offers an outline of a theory of communicative competence that is basedon the negations of these assumptions. In opposing the first two assumptions he in-troduces distinctions, respectively, between semantic universals which process ex-periences and those that make such processing possible, and between semanticuniversals which precede all socialization and those that are linked to the conditionsof potential socialization. Against elementarism, he argues that the semantic contentof all possible natural languages does not consist of combinations of a finite numberof meaning components. Differences in systems of classification preclude this, andsuch differences can be seen to infect all respects of intercultural comparison. Usingthe notion of 'performative utterance', the author elucidates the role of dialogue-constitutive universals as part of the formal apparatus required of a" speaker'scapacity to communicate. He then notes what would be required of a generalsemantics based on a theory of communicative competence; and finally points outhow this theory might be used for social analysis.

Chomsky1 starts from two experiences: the creativity of the speakerand the grammaticalness of language. Considering the finite capacityof human consciousness, every natural language must consist of a finitenumber of elements. Irrespective of this fact, everyone who masters alanguage can, with the aid of these elements, understand and producean infinite number of sentences, some of them unpredictably new.Moreover, every competent speaker can decide ad hoc whether asequence of linguistic expressions is formed correctly or violates thesystem of linguistic rules; he can differentiate 'intuitively' between

Dow

nloa

ded

by [

Uni

vers

ity o

f Sy

dney

] at

02:

40 1

4 M

arch

201

3

A THEORY OF COMMUNICATIVE COMPETENCE 361

correct and deviating formulations. And with the aid of the samecapability he can also partially understand semantically senseless orgrammatically garbled sentences and classify them according to degreeof grammaticalness. For these two particular achievements thecompetent speaker must possess a knowledge grossly disproportionateto his empirical information; the competent speaker must know morethan he can have learned in his previous contacts with his linguisticenvironment. Chomsky explains this asymmetry between knowledgeand experience by postulating (1) an abstract linguistic system whichconsists of 'generative' rules. I shall not comment on this, but go ondirectly to introduce three further assumptions that Chomsky makes.

The asymmetry evident when an adult speaker 'knows' more thanhe can have learned empirically is especially conspicuous in the caseof language acquisition in infants.2 Chomsky therefore assumes (2) thatthe development of the abstract system of linguistic rules is based uponthe interaction of phase-specific stimulus conveyance and organicmaturation processes. In other words, the system of linguistic rules isinnate. Chomsky further assumes (3) that this innate languageapparatus consists of linguistic universals which predetermine the formof all potential natural languages. The difficulties he encountered inhis attempts to ascertain this system of rules by means of the usualinductive methods of segmentation and classification led him, finally,to the assumption (4) that the given linguistic sequences are surfacestructures which result from the transformation of deep structures. Thebasic assumption of a transformational grammar proves useful,moreover, in explaining grammatical ambiguities in phrase structure.3

'Linguistic competence' is Chomsky's name for the mastery of anabstract system of rules, based on an innate language apparatus,regardless of how the latter is in fact used in actual speech. Thiscompetence is a monological capability; it is founded in the species-specific equipment of the solitary human organism. For such acapability to be a sufficient linguistic basis for speech, one would haveto be able to reconstruct the communication process itself as a 'mono-logical' one. The information model of communication is suitable forthis purpose. I consider this model to be monological because itconsistently attributes the intersubjectivity of meaning — that is, themutual sharing of identical meanings — to the fact that sender andreceiver — each an entity for itself — are previously equipped withthe same programme. It is this pre-established code that is supposedto make communication possible. Speech, the actual language be-

Dow

nloa

ded

by [

Uni

vers

ity o

f Sy

dney

] at

02:

40 1

4 M

arch

201

3

362 JURGEN HABERMAS

haviour, would then have to be explained as the result of an interactionbetween linguistic competence and certain psychological, as well associological, peripheral conditions which restrict the application of thecompetence. While the system of linguistic rules determines, on theone hand, whether a sequence of expressions is correct or deviant,these restrictive conditions determine, on the other, whether a correctlyformed expression is unusual or acceptable in a given situation.4

To begin with, I would like to show some of the difficulties inherentin this model. If general linguistics restricts itself to giving a rationalreconstruction of the abstract system of linguistic rules which the idealspeaker has in mind, as it were, prior to all communication, and if thetheory of language performance analyses solely restrictive extra-linguistic conditions for applying linguistic competence, then not onlygrammar and phonetics but semantics, too, would have to be developedindependently of the pragmatic dimension of language performance asan element of a monological ability. In this framework generalsemantics has two problems in particular to solve. First, it must clarifythe apparatus of rules which help us to project lexical units intogrammatical deep structures and to maintain constancy of meaning intransformational processes; secondly, it must reduce the lexica ofnatural languages to a finite number of meaning components, out ofwhich the basically solitaiy speaker can construct all possible semanticcontents. The elementaristic research strategy,5 which is supposed toreconstruct any lexical unit with the aid of general semantic markersand reduce them in the form of deductively classified marker hier-archies to some few universals, results conclusively from the mono-logical language model. At the same time it is assumed that languagepossesses an a priori meaning-structure. Bterwisch6 introduces thisstructure, matching it with the a priori phonetic structure:

Phonological markers represent acoustic properties of objects just aslittle as semantic markers directly represent environmental character-istics. The universal phonological inventory is a hypothesis aboutdistinctions which the human being can exploit linguistically accordingto the structure of his organism, i.e. the ear, the speech organs, and thenervous systems controlling them. A universal inventory of semanticmarkers would be, similarly, a hypothesis about the differentiations inregard to his environment, which the human being can achieve withthe help of his sense organs, his nervous system, or in general, hisapperceptive constitution. In other words, a complete inventory ofsemantic markers, from which each individual language makes aspecific selection, would in the end be an intrinsically exact character-ization of the apperceptive apparatus by which the human beinganalyses his surroundings practically and intellectually.

Dow

nloa

ded

by [

Uni

vers

ity o

f Sy

dney

] at

02:

40 1

4 M

arch

201

3

A THEORY OF COMMUNICATIVE COMPETENCE 3 6 3

Bierwisch's commentary elucidates the implicit assumptions for aprogramme of general semantics. We can summarize these in thesis-form as follows. The thesis of monologism assumes that the universalmeaning components belong to the basic equipment of the solitaryorganism of the speaking subject. The thesis is incompatible with theproposition that semantic universals could also be parts of an inter-subjectively produced cultural system. The thesis of a priorism assumesthat the inventory of ultimate meaning elements — as the conditionwhich makes semantic differentiation possible — precedes all ex-perience. The thesis is incompatible with the proposition that universalsemantic fields can also reflect the universality of specific scopes ofexperience. The thesis of elementarism assumes, finally, that the semanticcontent of all possible natural languages consists of combinations of afinite number of meaning components. This thesis is incompatible withthe proposition that semantic fields can be formed and shifted in struc-tural association with global views of nature and society (Weltbilder).

I believe that the propositions which are incompatible with thetheses can be more credibly argued than the theses themselves.Regarding theses (1) and (2): Universal meanings, which arise in allnatural languages, neither automatically precede all experience, norare they necessarily rooted in the cognitive equipment of the humanorganism prior to all socialization. The universal distribution ofmeanings, and even of meaning components, is not a sufficient criterionfor the a priorism and monologism of general semantics urged by theChomsky school of linguistics. Some meanings are a priori universal inas much as they establish the conditions of potential communicationand general schemes of interpretation; others are a posteriori universal,in the sense that they represent invariant features of contingent scopesof experience which, however, are common to all cultures. For thatreason we differentiate between semantic universals which processexperiences and semantic universals which make this processingpossible in the first place (i.e. a posteriori / a priori). Furthermore, somemeanings are intersubjectively universal in the sense that they are fixedon structures which first develop with the cultural level of linguisticcommunication itself; other meanings are monologically universal inas much as they refer to structures of the solitary human organismprior to all communication. Therefore we differentiate betweensemantic universals which precede all socialization and semanticuniversals which are linked to the condition of potential socialization(monological/intersubjective). The combination of these points of viewcomprise four classes of semantic universals:

Dow

nloa

ded

by [

Uni

vers

ity o

f Sy

dney

] at

02:

40 1

4 M

arch

201

3

3 6 4 JURGEN HABERMAS

Semantic universal*

a priori a posterioriintersubjective dialogue-constitutive cultural universal

universal

monological universal cognitive universals of perceptiveschemes of interpretation and motivational

constitution

A few examples can suffice as illustration.7 The dialogue-constitutiveuniversals include personal pronouns, interrogative, imperative, andassertive formators, modal formators and the like. In the cognitiveschemes of interpretation which have been absorbed in the linguisticsystem of rules we include causality and substance, space and time —in fact the deictic formators in general which establish the system ofpossible denotations. The best analysed example of cultural universalsis the system of kinship words and, correspondingly, of universals oforganic constitution of the system of colour words. Whether we can alsoassume a general vocabulary of basic drives and emotional tendenciesfor the motivational equipment is a more difficult question.

As for thesis (3), the a posteriori universal meanings surely make quiteevident the limits within which an elementaristic meaning-analysis cantake place. The institution of family, for instance, which is based onthe sexual privilege of the parents, i.e. on the incest barrier betweenthe generations, as well as between brothers and sisters, is indeeduniversally distributed;8 a system of kinship relations which is formedaround this institutional core is to be found in all cultures. But aninter cultural comparison of the kinship vocabulary shows clearly thatthis same semantic field is differently classified depending on theprevailing status system, i.e. on the specific definition of the age-, sex-and descent-linked primary roles. It is true, as the ethnological classi-fications themselves indicate, that we can find descriptive systemswhich allow us to reconstruct any given kinship vocabulary in termsof a small number of semantic markers. But there is no generalcriterion for the selection between several conventional descriptivesystems. If we follow the standpoint of cognitive validity and choosethat descriptive system which corresponds to the relevant evaluationcriteria of a culture itself, then practically every culture demands itsown system. Even in the case of coincident vocabularies tor twodifferent cultures, the kinship relations can be interpreted differently;

Dow

nloa

ded

by [

Uni

vers

ity o

f Sy

dney

] at

02:

40 1

4 M

arch

201

3

'A THEORY OF COMMUNICATIVE COMPETENCE 3 6 5

and in that case, which of the alternative interpretations is cognitively'valid' depends solely on the acknowledged definition of social roles.9

The system of colour expressions is a case similar to that of kinshipvocabulary. The neuro-physiological sense organization establishes aculturally invariable scope of experience for colour perception. Thesemantic field 'colour-words' is as universal as that of kin-terms, butagain the classification of this field varies; notwithstanding therelatively high degree of congruence, here too we find no universalcognitively valid descriptive system, for the common patterns ofinterpretation classify the colour spectrum in different intervals andnot only within the physically determinable dimensions of hue,luminosity, and saturation.10

On the basis of ethno-linguistic evidence, we must conclude that thefields of meaning which depend upon culturally universal institutionsor upon organically universal equipment probably do representsemantically equivalent scopes (family, colour), but that, in addition,the interpretation of these scopes already depends upon the socio-cultural background.

The dependence of semantic analysis upon the non-exceedablecommon context of the society11 to which the speakers belong isconnected with a fundamental difficulty of semantic elementarism.The examples of semantic analysis given by Chomsky, Fodor, Katz,and others reduce complex meanings to simple semantic markers.These markers are usually specifications of the following type: physicalitems versus non-physical, animated versus inanimate, male versusfemale, old versus young. Such marker-pairs are introduced asdisjunctive specifications, which can then be classified hierarchicallyor divided into new disjunctions; for example, living creatures aredivided into human beings versus non-humans (animals, plants);human beings into male and female; male into men (adults) andchildren (non-adults); men into married and unmarried (bachelor),etc. Thus it is possible to create conceptual hierarchies with meaningcomponents of increasing complexity inserted in their intersections.This procedure depends on the fundamental relations of hyponymy,and incompatibility can be differentiated as antonymy, comple-mentarity, and converseness.12 The vocabulary of all natural languagesis structured along these fundamental relations. That is undisputed.An elementaristic meaning-analysis must assume, however, that thereis, in addition, a universally valid list of ultimate, i.e. independent andirreducible, meaning components. It is the fundamental semantic

Dow

nloa

ded

by [

Uni

vers

ity o

f Sy

dney

] at

02:

40 1

4 M

arch

201

3

366 JURGEN HABERMAS

relations holding between these components which, according to suchan analysis, determine the fundamental relations between compoundmeanings. I wish to dispute this assumption.

The empirical evidence of comparative ethno-linguistics indicatesthat the structures of culturally and historically changing world views(Weltbilder) determine both (a) whether a finite number of inde-pendent and irreducible meaning components is assumed at all, orwhether, instead, a system of basic, mutually interpreting meaningcomponents is to be presupposed; and (b) which meaning componentsare recognized to be the ultimate ones in a given case or which systemof meaning components is to be recognized as basic. The conceptualhierarchies which the semantic analysis of a given common vocabularydiscloses change in accordance with the world view, i.e. the globalinterpretation of nature and society, which is valid in a social systemat a particular stage of development. It is apparent that the examplespresented by Chomsky and his colleagues are likewise guided by aglobal pre-understanding, though admittedly one that possesses acertain plausibility for us as sharers of the ontology governing theeveryday understanding of enlightened members of our civilizationafter three hundred years of modern science and the criticism ofreligion, a hundred years since Darwin, and fifty years since Freud,i.e. after a subjectivization and privatization of belief systems, to-gether with the rationalization of social life which Max Weber hasanalysed.

Only the briefest consideration suffices to show whether dichotomieslike those between physical and non-physical objects, organic and inor-ganic nature, and human and non-human being, which are basic for us,could have just the same position in the conceptual hierarchy of ani-mistic, mythical, religious, philosophical or scientifically oriented viewsof the world. But then, if we are right in saying that they could not,the programme of general semantics could only hope to succeed if thecategorial frame of reference for a structuralist analysis of alt possibleglobal interpretations of nature and society could be elicited systemat-ically from the general theory of language itself.

It follows from these considerations that general semantics cannotbe adequately developed on the narrow basis of the monologicallinguistic competence proposed by Chomsky. The general competenceof a native speaker does not extend merely to the mastery of anabstract system of linguistic rules, which — pre-programmed by hisorganic equipment and the processes of stimulated maturation — he

Dow

nloa

ded

by [

Uni

vers

ity o

f Sy

dney

] at

02:

40 1

4 M

arch

201

3

A THEORY OF COMMUNICATIVE COMPETENCE 3 6 7

introduces into a communication in order to function as sender orreceiver during the transfer of information. That is, it is not enough tounderstand language communication as an application — limited byempirical conditions — of linguistic competence. On the contrary,producing a situation of potential ordinary-language communicationis itself part of the general competence of the ideal speaker. In otherwords, a situation in which speech, i.e. the application of linguisticcompetence, becomes in principle possible, depends on a structure ofintersubjectivity which is in turn linguistic. This structure is generatedneither by the monologically mastered system of linguistic rules, norby the extra-linguistic conditions of its performance. On the contrary,in order to participate in normal discourse the speaker must have athis disposal, in addition to his linguistic competence, basic qualifica-tions of speech and symbolic interaction (role-behaviour), which wemay call communicative competence. Thus communicative competencemeans the mastery of an ideal speech situation.13

We can elucidate this, in the first instance, by studying the exampleof a category of verbs to which J . L. Austin ascribed a performatoryuse.14 As is well known, verbs like 'promise', 'announce', 'warn','report', 'desire', 'determine', etc. can be used to perform the acts theyrespectively designate rather than refer to or describe them. The mean-ing of a 'performative utterance' includes a reference to (a) an act ofutterance in a particular and appropriate interaction relationship ('Ihereby promise . . . '), (b) the definition of a (suitable) situation whichis explicitly determined by the performance of the speech act itself, and(c) the propositional content of the dependent clause. Austin differen-tiates between the levels of 'saying something' and of 'doing something'(locutionary level v. illocutionary level). Being composed of speechacts and dependent clauses of propositional content, utterances ingeneral have, in addition to the meaning of their propositional content,a meaning which is linked to the speech situation as such. This,following Austin, we can call their 'illocutionary force'. When they useperformative expressions, the speech acts are linguistic representationsof that illocutionary force, i.e. the universal pragmatic power ofutterances. Expressions of this kind retain no given pragmatic featureof contingent speech situations; they explain the meaning of certainidealized features of speech situations in general, which the speakermust master if his competence is to be adequate for participating at allin situations of potential speech. A theory of communicative com-petence can thus be developed in terms of universal pragmatics.16

Dow

nloa

ded

by [

Uni

vers

ity o

f Sy

dney

] at

02:

40 1

4 M

arch

201

3

368 JURGEN HABERMAS

The performatives have been called discourse operators, i.e.formators of the speech situation which belong to a meta-language forthe placing of linguistic expressions in speech situations. Perhaps weshould say that ordinary language contains elements which enable itto be used as the meta-language of the speech situation as well. Theperformatives are not the only elements to be included here. Deicticelements represent the other class of universal pragmatics. In amanuscript on 'Pragmatics, Speech Situation and Deixis', DieterWunderlich has analysed six elements which serve to verbalize featuresof the ideal speech situation.16 I shall mention two examples of deicticelements. Firstly, personal pronouns. The linguistic description canonly explain why the sentence

'I apparently am hungry'

deviates from

'He apparently is hungry',

if ' I ' is understood not only as one nominal pronoun among many, butas a reflexive specification of a speaker in a particular situation. Ananalogous consideration pertains, secondly, to deictic expressions ofplace and time. The deviant status of a sentence such as

'I fear that it is raining here (now)'

would not be linguistically apprehensible if the pragmatic sense of'here' and 'now' were to be ignored and the words merely understoodas some adverbial specifications among others. The choice betweendefinite and indefinite articles, or between different forms of pro-nominalization, does not express a characteristic of the nouns but,rather, certain pragmatic relationships.

What can we learn from these considerations ? If communicativecompetence meant no more than that the speaker can relate himselfreflexively to speech situations and copy speech situations meta-linguistically, then the speech situation itself would have to be pre-supposed in turn and understood as a simple empirical state — at leasta non-linguistically determined state. This is a mistaken view. The'verbalization' of pragmatic features inherent in speech situationssimply lifts onto the level of linguistic communication that illocutionaryforce which is already generated with the structure of speech situationsitself. Utterances are never simply sentences. Even if they do notexpressly make pragmatic relations their subject, they are, due to their

Dow

nloa

ded

by [

Uni

vers

ity o

f Sy

dney

] at

02:

40 1

4 M

arch

201

3

A THEORY OF COMMUNICATIVE COMPETENCE 369

illocutionary force, integrated from the beginning into a form ofintersubjectivity of mutual understanding. This structure of potentialspeech is, in a manner of speaking, itself of a linguistic nature, for thespeech situation already contains all of the reflexive relations. Theseuniversal pragmatic features can be expressed in explicit discourse withthe aid of the linguistic elements mentioned. If this is the case, however,we should assume that these elements do not serve as a subsequentverbalization of a previously coordinated speech situation; on thecontrary, they must be the very factors which enable us to generatethe structure of potential speech.17 It is the dialogue-constitutiveuniversals, as we now prefer to say, that establish in the first place theform of intersubjectivity between any competent speakers capable ofmutual understanding.18 It may be added that the structure ofpotential speech is present in the speech process even if pragmaticrelations are not contained in the explicit content of the linguisticcommunication, i.e. when we omit the dialogue-constitutive universalin explicit speech. In this case understanding is incomplete on thelocutionary level. Only when the dialogue-constitutive universalswhich give a sentence the status of an utterance are substituted bynon-linguistic means, e.g. by gestures and context, can an implicitunderstanding on the illocutionary level be added compensatorily.

Above all, communicative competence relates to an ideal speechsituation in the same way that linguistic competence relates to theabstract system of linguistic rules. The dialogue-constitutive universalsat the same time generate and describe the form of intersubjectivitywhich makes mutuality of understanding possible. Communicativecompetence is defined by the ideal speaker's mastery of the dialogue-constitutive universals, irrespective of actual restrictions underempirical conditions. We shall disregard the question of how far themotivation of actions involved in language-games is linguistically opento public communication. We shall also disregard the question ofwhether and to what extent a systematically distorted communicationactually takes place. Thus the idealization exists in the fact that wesuppose an exclusively linguistic organization of speech and inter-action. The ideal speech situation can then be analysed according tothe functions of pure dialogue-constitutive universals.

(1) The personal pronouns (and their derivatives) form a referencesystem between potential speakers. The identity of meanings, thefoundation of every communication, is based on intersubjectivelyvalid rules. Their validity is intersubjective, in the strict meaning of

Dow

nloa

ded

by [

Uni

vers

ity o

f Sy

dney

] at

02:

40 1

4 M

arch

201

3

370 JURGEN HABERMAS

the word, if at least two speakers understand the meaning of a symbolon the basis of reciprocal recognition. For only in that case is it possiblefor both speakers to comprehend and identify the meaning from theirown position and from that of the other at the same time. Only thisinterlacing of perspectives makes an intersubjectively valid meaning,and thus identity of meaning, possible. The system of personal pro-nouns enables every participant to assume incompatible roles simul-taneously, namely that of the I and that of the You. Every being, whosays T to himself asserts himself towards the Other as absolutelydifferent. And yet at the same time he recognizes himself in the latteras another I, and is conscious of the reciprocity of this relationship;every being is potentially his own Other. These dialogue roles of I andYou are reproduced on the level of We and You, while He, She, andThey describe roles of virtual or potential participation in the dialogue.

(2) The deictic expressions of space and time, as well as articles anddemonstrative pronouns, form the reference system of possible de-notations. They link the levels of intersubjectivity on which thesubjects converse and interact reciprocally with the levels of objectsabout which the subjects make propositions. This linking of levels ispossible because the cognitive content of the deictic expressions isambiguous; they contain, respectively, two different schematisms ofeach underlying category (space and time, substance and causality).We relate articles and demonstrative pronouns to persons just as tothings; space and time deixes refer to 'experienced' as well as tomeasured spaces and times.

(3) Forms of address (vocative), forms of social contact (greeting),of speech introduction and speech conclusion, indirect discourse,questions and answers, are performatory in that they are directed atthe act of speaking as such. They determine the structure of potentialspeech in as much as they explain the pragmatic meaning of speakingitself. Speech is the medium of communication which already pre-supposes a tacit consensus about what it means to communicate andan awareness of the possibility of misunderstanding, as well as of errorand deception.

(4) The further performatory speech acts form a system whichfinally enables us to mark the basic differentiations which are funda-mental for any speech situation.19

(a) Being and appearance. Expressions which refer to the truthvalue of utterances (not of statements) according to the prototype of'to claim' and 'to dispute' form the dimension of being and appearance

Dow

nloa

ded

by [

Uni

vers

ity o

f Sy

dney

] at

02:

40 1

4 M

arch

201

3

A THEORY OF COMMUNICATIVE COMPETENCE 371

(assure, confirm, deny, certify, testify, doubt, question). All speechacts imply an intended consensus on that which really is, as distinctfrom that which subjectively only appears to be (the propositionalcontent). This presupposes a differentiation between a public world ofintersubjectively acknowledged interpretations and a private world ofpersonal feelings and impressions.

(b) Being and essence. Expressions which refer to the self-represen-tation of persons according to the prototype of 'to reveal' and 'to hide'form the dimension of being and essence (expose, present, allude,express, conceal, obscure, betray, symbolize, virtualize, take at one'sword, deceive, etc.). In all speech acts the subjects in their speech-actperformances unavoidably express their own selves at the same timeas they converse with one another on some propositional topic. Thispresupposes a differentiation between a communication'on objects anda meta-communication on the level of intersubjectivity.

(c) What is and what ought to be. Expressions which refer to thenormative status of rules according to the prototype of 'to prescribe'and 'to follow' form the dimension of the 'ought' (order, obey, allow,demand, refuse, resist, recommend, advise, warn, oblige, violate, callto account, etc.). All speech exists in a context of actions and in-tentions. The mutual recognition of the subjects who communicatewith one another includes the certainty that they can conduct them-selves reciprocally towards one another's expectations, i.e. act accordingto valid norms. This presupposes the differentiation between validrules, which are intentionally followed, and regularities of observableevents, which can be stated empirically.

If one thus analyses the structure which we generate and describeby means of pure dialogue-constitutive universals, one arrives at anumber of symmetrical relations for the ideal speech situation. Pureintersubjectivity is determined by a symmetrical relation between Iand You (We and You), I and He (We and They). An unlimitedinterchangeability of dialogue roles demands that no side be privilegedin the performance of these roles: pure intersubjectivity exists onlywhen there is complete symmetry in the distribution of assertion anddispute, revelation and concealment, prescription and conformity,among the partners of communication. As long as these symmetriesexist, communication will not be hindered by constraints arising fromits own structure.

(1) In the case of unrestrained discussion (in which no prejudicedopinion cannot be taken up or criticized) it is possible to develop

Dow

nloa

ded

by [

Uni

vers

ity o

f Sy

dney

] at

02:

40 1

4 M

arch

201

3

372 JURGEN HABERMAS

strategies for reaching unconstrained consensus; (2) on the basis ofmutuality of unimpaired self-representation (which includes theacknowledgment of the self-representation of the Other as well) it ispossible to achieve a significant rapport despite the inviolable distancebetween the partners, and that means communication under con-ditions of individuation; (3) in the case of full complementarity ofexpectations (which excludes unilaterally constraining norms) theclaim of universal understanding exists, as well as the necessity ofuniversalized norms. These three symmetries represent, incidentally, alinguistic conceptualization of what are traditionally known as theideas of truth, freedom, and justice.

A speech situation determined by pure intersubjectivity is anidealization. The mastery of dialogue-constitutive universals does notitself amount to a capacity actually to establish the ideal speechsituation. But communicative competence does mean the mastery ofthe means of construction necessary for the establishment of an idealspeech situation. No matter how the intersubjectivity of mutualunderstanding may be deformed, the design of an ideal speech situationis necessarily implied in the structure of potential speech, since allspeech, even of intentional deception, is oriented towards the idea oftruth. This idea can only be analysed with regard to a consensusachieved in unrestrained and universal discourse. Insofar as we masterthe means for the construction of an ideal speech situation, we canconceive the ideas of truth, freedom, and justice, which interpret eachother — although of course only as ideas. On the strength of com-municative competence alone, however, and independent of the empir-ical structures of the social system to which we belong, we are quiteunable to realize the ideal speech situation; we can only anticipate it.

It should be possible to demonstrate the deformations of pureintersubjectivity, induced by the social structure, • on the basis ofasymmetries in the performance of dialogue rules. The unevendistribution of dialogue-constitutive universals in standard com-munication between individuals and social groups indicates theparticular form and deformation of the intersubjectivity of mutualunderstanding which is built into the social structure. But here I amafraid I must leave the matter at that. The considerations I havespoken of can only claim to be a first attempt to grasp communicativecompetence in terms of linguistic theory. Let me, however, draw twoconclusions, the first of which will lead back to the problem of generalsemantics.

Dow

nloa

ded

by [

Uni

vers

ity o

f Sy

dney

] at

02:

40 1

4 M

arch

201

3

A THEORY OF COMMUNICATIVE COMPETENCE 3 7 3

I imagine that the particular form of intersubjectivity of mutualunderstanding — that is, the particular structure of potential speech —is the basic linguistic framework which also determines the scope andstructure of corresponding world views. Then, the classification ofsemantic fields is predetermined by the question of how far the net ofintersubjectivity must be spread in order to stabilize the identity of theindividuals, as well as that of the social group in a given culture orsubculture at a given time. The structural differences between theanimistic, the mythical, the religious, the philosophical, and thescientistic views of life lie clearly in this dimension. The range of thoseglobal interpretations of nature and society extend from the case oftotal identification of the individual and his group with all non-humanphenomena, within an all-embracing association of motivated actions,to the case of total reification of all intersubjective relationships withinthe framework of objectifying sciences. At this time I cannot pursuethis topic further. However, one consequence seems to me to beimportant in our context. If we could succeed in describing defor-mations of pure intersubjectivity in the dimension in which dialogue-constitutive universals are applied; and if it were possible, moreover,to distinguish also the categorial frameworks of potential views of lifein terms of distributions of dialogue-constitutive universals, thengeneral semantics could be developed on the basis of a theory ofcommunicative competence.

The second and final point is how that theory of communicativecompetence might be employed for social analysis. As already men-tioned, the 'idealization' of the concept of the ideal speech situationdoes not consist simply in the fact that we disregard contingentempirical limitations. It consists rather of the supposition that themotivational base of all actions is organized linguistically, i.e. withinthe structure of potential speech. By this idealization we imagine theactual motivations of the actor being identical with the linguisticallyapprehensible intentions of the speakers. This model of pure com-municative action is included in the design of pure intersubjectivity.Now we have reason enough to assume, however, that social action isnot only — and perhaps not even primarily — controlled by motiveswhich coincide with the intentions of the actor-speaker, but rather bymotives excluded from public communication and fixed to a pre-linguistic symbol organization. The greater the share of pre-linguis-tically fixed motivations which cannot be freely converted in publiccommunication, the greater the deviance from the model of pure

Dow

nloa

ded

by [

Uni

vers

ity o

f Sy

dney

] at

02:

40 1

4 M

arch

201

3

374 JURGEN HABERMAS

communicative action. I would propose to make the empiricalassumptions, first, that these deviations increase in proportion to thedegree of repression which characterizes theinstitutional system withina given society; and secondly that the degree of repression depends inturn on the developmental stage of the productive forces and on theorganization of authority, that is of the institutionalization of politicaland economic power.

N O T E S

1. N. Chomsky, Aspects of the Theory of Syntax, M.I.T. Press, Cambridge, Mass.1965.

2. D. McNeill, 'Developmental Psycholinguistics', in F. Smith and G. A. Miller(Eds.), The Genesis of Language, M.I.T. Press, Cambridge, Mass. 1966, pp. 15-84.

3. N. Chomsky, Language and Mind, Harcourt, Brace & World, New York1968.

4. J . Fodor and M. Garret, 'Some Reflections on Competence and Performance',in J. Lyons and R. J. Wales (Eds.), Psycho-linguistic Papers, Edinburgh UniversityPress, Edinburgh 1966, pp. 135-63; R. J . Wales and J. C. Marshall, 'TheOrganization of Linguistic Performance', ibid., pp. 29-80; C. B. Cazden, 'OnIndividual Differences in Language Competence and Performance', in Journalof Special Education, Vol. I (1967) No. 2.

5. J . J . Katz and P. M. Postal, An Integrated Theory of Linguistic Description, M.I.T.Press, Cambridge, Mass. 1964.

6. M. Bierwisch, 'Strukturalismus', Kursbuch, Vol. 5, Frankfurt a.M. 1966, pp. 97 f.7. J . H. Greenberg (Ed.), Universals of Language, M.I.T. Press, Cambridge, Mass.

1963, 1966.8. C. Lévi-Strauss, Les Structures élémentaires de la Parenté, Mouton & Co., Paris

1967.9. A. Romney, 'Cognitive Aspects of English Kin-terms', in American Anthro-

pologist (1946), pp. 36-170.10. H. C. Conklin, 'Hanunvo Color Categories', in D. Hymes (Ed.), Language in

Culture and Society, Harper & Row, New York 1964, pp. 189-92.11. J . Lyons, Introduction to Theoretical Linguistics, Cambridge University Press,

London 1969, pp. 419 f. and pp. 470 ff.12. Lyons, Introduction, op. cit., pp. 446 ff.13. I propose to use this term in a way similar to that in which Chomsky uses

'linguistic competence'. Communicative competence should be related to asystem of rules generating an ideal speech situation, not regarding linguisticcodes which link language and universal pragmatics with actual role systems.Dell Hymes, among others, makes use of the term 'communicative competence'in a socio-linguistically limited sense. I don't want to follow this convention.

14. J. L. Austin, How to do Things with Words, Clarendon Press, Oxford 1962.15. J . R. Searle pursues a similar approach with his theory of speech acts: Speech

Acts, Cambridge University Press, London 1969.16. Manuscript T.U. Berlin, Sept. 1969.

Dow

nloa

ded

by [

Uni

vers

ity o

f Sy

dney

] at

02:

40 1

4 M

arch

201

3

A THEORY OF COMMUNICATIVE COMPETENCE 375

17. This is why Searle conceives the linguistic rules which govern speech acts aswhat he calls 'constitutive rules'. 'Constitutive rules do not merely regulate, theycreate or define new forms of behavior' (op. cit., p. 33). 'The hypothesis ofthis book is that speaking a language is a matter of performing speech actsaccording to systems of constitutive rules' (ibid., p. 38).

18. Searle puts the same argument in the following way: 'If I am trying to tellsomeone something, then. . . as soon as he recognizes that I am trying to tell himsomething and exactly what it is I am trying to tell him, I have succeeded intelling it to him. Furthermore, unless he recognizes that I am trying to tell himsomething and what I am trying to tell him, I do not fully succeed in telling itto him . . . In the case of illocutionary acts we succeed in doing what we aretrying to do by getting our audience to recognize what we are trying to do.But the "effect" on the hearer is not a belief or response, it consists simply inthe hearer understanding the utterance of the speaker. It is this effect that Ihave been calling the illocutionary effect. The way the reflexive intentionworks then . . . is: the speaker S intends to produce an illocutionary effect IEin the hearer H by means of getting H to recognize S's intention to produce IE'(ibid., p. 47).

19. Austin claims that there are about a thousand performatives in English. Theclassification proposed by Austin himself is not convincing. Searle, who presentsthe most penetrating analysis of the structure of the speech act (cf. op. cit.,Ch. 3, pp. 22-71) does not give a systematic account of the classification ofspeech acts. My proposal is intended to have the role of such an account, butthe three criteria offered still lack a reasonable explication.

Dow

nloa

ded

by [

Uni

vers

ity o

f Sy

dney

] at

02:

40 1

4 M

arch

201

3