In the current literature of linguistics and psycholinguistics, one can frequently find the dictum that "linguistic competence" is the proper object of study for linguistics while performance is the proper object of study for psychology or psycholinguistics. Such a division of labor has its attractive aspects but it assumes that one can draw an absolute and principled distinction between competence and performance. I shall try to show that no such distinction is possible with the present state of knowledge which does not have the consequence of restricting in an unacceptable manner the study of meaning.

Chomsky (1965, p. 4) l-a.5 noted the .similarity between the competence-performance distinction and that of the de Saussurian langue-parole; but whereas langue is merely a "systematic inventory of items," competence refers to the earlier Humboldtian conception of 'a system of generative processes." The motivation for the distinction in both cases stems from the observations of fluctuations in grammaticality of the speech of individuals and the ascription of a proper theoretical significance to this observation, namely, that the speech of individuals does not directly reflect their grammatical knowledge. If one assumes that a mature speaker's knowledge of his language does not fluctuate from moment to moment as does grammaticality of his utterances), the linguist's task in building a grammar of English (say) becomes in effect, one of describing the speaker's "permanent knowledge" of his language, or, his linguistic competence. It is ten left for the psychologist to describe how the interfering effects that manifest themselves during speaking interact with the speaker's linguistic-competence to produce the grammatically impaired utterances that are typical in everyday situations.

It is important to realize that the view that has just been expressed applies equally to both speaking and understanding, since the psychologically limiting (or interfering) conditions that impair the grammaticality of spoken (or written) sentences (sometimes called "encoding") also produce similar effects on the understanding ("decoding") of well-formed sentences (viz., we sometimes fail to understand a perfectly grammatical sentence, or, misunderstand, it). This is clear enough, although it is worth emphasizing'. because it has sometimes been misunderstood (students, especially, are prone to this mistake). But there is a more subtle mistake, committed more often, which is to consider "the intuitions of the native speaker" as synonymous with ''the ideal speaker-hearer's intrinsic competence." Chomsky (1965, Chapter 1, and elsewhere) is himself quite clear on this point. Introspective data or linguistic intuitions provide no more of a direct indication of competence than overt performance (speech).

The intuitions of a native speaker about his language (e.g., Is X a well-formed, sentence or not') represent performance data under a special set of conditions which are, to be sure, different from ordinary speech performance, but are nevertheless subject to error. As Chomsky is clearly aware, and explicitly states, this fact lends a definite uncertainty to the linguistic enterprise as currently conceived: since there is no known (i.e., mechanical) direct procedure for inferring the under lying competence from the performance data (overt speech and intuitions), the attempt at specifying a grammar may fail altogether. Furthermore, even if the inferential process were to be successful, ,In additional difficulty is that there will remain uncertainty as to the body of facts to be explained or, in other words, the assessment of the adequacy. of the inferences thus made (i.e., the proposed grammar) problematic. At best, the evaluation of the adequacy of proposed grammar is limited to those sentences it generates over which native speakers' intuitions agree (the recalled "clear eases").³

Faced with this intrinsic difficulty with linguistics as it is currently conceived, Chomsky is candid enough to raise the issue as to whether linguistics is to be considered a ''science" since its data are not all "objective." He dismisses the issue as unproductive (agreed!) and points out that, for the moment, the number of clear cases" far exceeds our ability for account for them. It is probably true, as he points out, that all increase objectivity of linguistic data would not lead to easier and insights.⁴ The competence-performance distinction has been misconstrued at another level, especially by Psychologists of the behaviorist tradition. This relates to the nature of the explanation to be provided for what constitutes knowledge a language (i.e., linguistic competence). Generative linguists of both modern and earlier times, have adopted "rationalist" views and attribute to the individual inborn mechanisms of a highly specific character (viz.., they insist on going beyond the obvious and empty concession that human language is species .specific). This, of course, a psychological (or psycho-neurological) theory, and psychologists of the behaviorist tradition (whether of the Lockian associationistic bend or of the strict Watsonian tradition or of the ''neo-Behavioristic'' sort have rejected such speculations as "animistic," "non-materialistic non-objective," and "non-empirical." As Chomsky points out (see especially note 1 in Chapter 1, 1965), there are three issues involved in this criticism. One is the issue of, "dualism", namely, whether the underlying competence is to be found in a non-material "mind" or a material "brain." This is an idle controversy for both fields and belongs to metaphysics, not science. The second issue is that of "behaviorism," viz., whether one should be concerned with observable data only or whether there is an interest in the substrata of observable behavior.

Although strict behaviorism has never been seriously practiced, not even by Skinnerians (see James, 1966), the issue is whether the concept. introduced in the explanatory system are to have behaviorally defined meaning or not (i.e. mentalistic in nature) It is possible to give behavioral definition of mental concepts such as "knowledge," "decision," and "intention." but such a reductionism strips the mentalistic concepts from their explanatory value. Chomsky has consistently and persuasively argued that this leads to a totally sterile approach that is incapable of handling the complexities to be explained.⁵ The third issue is that of "introspectionism" and relates to the problem of objectivity raised earlier. Chomsky's reasonable position is that if intuitions lead to linguistic insights about grammars, there is no more justification to exclude them as data than overt speech performance.

The specific ways in which linguistics or the study of competence is non-objective and mentalistic should now be clear. Charges that the rationalism of current work on linguistic competence is non-empirical is an obvious and blatant fallacy in view of the empirical criteria required for the evaluation of grammars as to their adequacy (both descriptive and explanatory). Charges of mentalism are certainly true, but the use of mentalistic concepts is an advantage, indeed a necessity for freedom from theoretical sterility. Charges of non-objectivity are also true, but only in the ultimate sense as discussed earlier, not in the sense of l9th century psychological introspectionism where the accessibility and reliability of the primary data were unacceptably low. Neither of these caveats applies to a large and significant body of linguistic data.

3. Anyone who has attended linguistics colloquia in recent times has no doubt experienced he humorous situation in which the speaker offers a sentence as an illustration for a particular rule he proposes, only to be faced by vigorous head shakes in the audience on the part of the people who reject the speaker's sentence as "non-grammatical," whereupon the speaker makes another attempt with a different sentence, sometimes with no greater success.

4. Descriptive adequacy is no the sole issue since, theoretically, several different proposed grammars can be equally adequate in accounting for the available data. On "explanatory adequacy" see Chomsky, 1965, Chapter 1.

5. This claim is by no means generally accepted and it appears that we are faced here with a Kuhn-type clash of paradigms. The only resolution possible, Kuhn (1962) suggests, is the loss of power of the "old guard" and the firm entrenchment of the "young Turks" on independent levels. It is less well known that Lashley (1951) before Chomsky, has come to similar conclusions from an entirely different perspective. Osgood's (in press) recent work on "perceptual presuppositions" referred to earlier must be considered, despite his claims to the contrary, as falling within the mentalistic frame of reference. His analysis of sentential presuppositions makes use of cognitive processes that do not receive an adequate behavioral definition: e.g., on p. 26 he states: " Use of Definite the involves an implicit kind of redundancy -- the entity in question is assumed by the speaker to be 'familiar' to the hearer." Such statements are indistinguishable from those found in mentalist theories such as Chomsky, 1965, Katz, 1966 Vendler, 1967, and others.