by
Leon James
Department of Psychology, University of
Hawaii
Originally published in Studies in Second Language Acquisition, 1982, 4(2),
205-10 (under the name Leon A. Jakobovits)
Carol Kates, Pragmatics and Semantics: An Empiricist Theory. Ithaca: Cornell University Press, 1980. Pp.253. Hardbound. (In SSLA 1982)
This relatively short book is yet a considerable challenge requiring persistence and patient hours for many of its concise pages. Every paragraph requires deep reflection and cumulative processing to maintain one's grip over the building argument throughout its twelve brief chapters. The author draws her affiliation from both philosophy and linguistics, and incorporates the literature in psycholingusitics of the 1960's and 1970's. Thus, the reader must be able to handle technical details from these three fields. Nevertheless, the experience of getting through the book was an exciting intellectual adventure, and I am left with the agreeable feeling of well invested professional time. I feel somewhat better educated now as compared to before. I will describe what is exciting about this book and why it is worthwhile reading for those who attempt to figure it out.
HISTORICAL CONTEXT
The author presents her two main sources of ideas as coming from, on the one hand, the BloomfieldÐHockett axis of behaviorism in linguistics, and on the other hand, the positivistic phenomenology of Husserl. Within this philosophical oddÐcouple Kates intersperses the psychological literature on language development as a means to shore up many empirical deductions that result from such a marriage. The book's chief and important claim is to have effected a theoretically productive integration between a type of radical empiricism and genuine phenomenology. As a result, the nature of communicative competence is illumined, and we have for the first time
a proposal in the language sciences that appears to be successful in repairing the great rift wrought by the Chomskyan revolution.
PHILOSOPHICAL CONTEXT
Lest I be thought of exaggerating Carol Kates' theoretical claims or my claims as to the difficulty level of her book, let me quote just one paragraph of average difficulty level. .It is the opening paragraph of Chapter 10 which is called "The Authority of Logical Truth."
"In this section I shall further develop the empiricist theory of utterance meaning I have been presenting by showing how this theory can account for the necessity of logical truth and the evident distinction between such necessary truths and contingent empirical statements. I shall contrast phenomenological empiricism with a metaphysical form of rationalism and with classical empiricism and show that the first theory can account for intuitions of logical necessity without positing an a priori knowledge of metaphysically objective ideas and truths, and without falling into a stultifying form of psychologism" (188)."
To debunk the technical language a bit, let me clarify the terms. By "empiricist theory" is meant a theory based on observables (rather than innate or other unobservables). "Necessary truths" vs. "contingent truths" captures the difference between universal and invariant facts vs. relative and inconsistent facts ("psychologisms" and "psychologistic reasonings"). This distinction is crucial for Kates' proposal since it is the place where the marriage is superior to the divorce; that is, currently known empirical accounts in the language sciences are all based on contingent truths which a speaker acquires through cultural experience and socialization, and thus, are not universal or culture free. As will be shown below, Kates' proposal is the first serious attempt to work out an argument for empiricism that allows it to deal with the notion of observable ideal objects, i.e., objects
that are not in space and time (ideal) yet are observable. QuiteÐa claim! "Phenomenological empiricism" is the title of the proposed marriage, where "phenomenology" is to be identified with the ideas of Husserl, as modified by Kates, and where "empiricism" is to be identified with Hockett and with psycholingusitics. "Metaphysical rationalism" stands for Chomsky, and GÐT grammars in general. "Classical empiricism" means J.S. Mills, chiefly Husserl's critique of him. To paraphrase the quoted paragraph, Kates is arguing in favor of an empiricism that transcends itself or a phenomenology that constricts itself, meeting in the middle where they can together engender scientific accounts of abstract knowledge ("linguistic intuitions" and "universal semantic and grammatical categories").
PARADIGMATIC STRUCTURES vs. GENERATIVE RULES
Kates freely acknowledges Chomsky's good influence in linguistics, which was that he successfully redefined linguistics science assigning a central focus to competence. In effect, Chomsky evolved techniques that allowed researchers to investigate the relation between linguistics and cognitive processes (operations functionally equivalent to knowledge, intuition, or competence). The big rift came when Chomsky turned to Cartesian rationalism as the theoretical vehicle for dealing with ideal objects or forms. Because Cartesian and Kantian rationalism posit a scientifically empty duality, which is to say that the second, underlying world is unobservable, therefore it is that Kates rejects GÐT oriented research today. To her, unobservables are unscientific. She finds in Husserl a more agreeable turn, permitting her to invent the notion of observable ideal objects. Note however that she retains the rationalistic concept of "ideal" as real, but not as in place and time. This is the great leap for the empiricist who is still stuck at the monistic level of crude behaviorism; that is, an empiricism that relies entirely on analogy or generalization to account for novel language use (creativity). Kates rightly rejects an oversimplistic objectivism that
may satisfy psychologists but leaves the linguist totally without instruments. The facts of linguistics and of linguistic knowledge cannot be encompassed by analogic reasoning. This much Chomsky has accomplished by shifting focus on competence. There is no turning back to preÐChomskyan behaviorism when semantics was not a formal focus for distributional grammars, though acknowledged and used in practice by the grammarian...Kates finds great comfort in Hockett's 1958 restatement of Bloomfield's behavioristic model of semantics, and again in 1978 (through personal communication with him). According to Kates, Hockett's approach is to equate the distributional structure of language use with the notion of an individual speaker's stock or repertoire; then, this corpus "of the most productive patterns in the community" is equated with "communicative competence."
But (as shown by Chomsky) distributional grammars with fixed repertoires cannot explain the actual as well as the theoretical novelty in language use. Hence, if one rejects GÐT grammars, one needs another method for relating utterances to cognitive structures (reasonings) and communicative intentions (topicÐcomment predications). Kates accomplishes this astonishingly difficult theoretical task through a Husserlian definition of experience, particularly the idea that the experience of individual objects in space and time can be "bracketed" or transfigured into "intentional structures within experience"; these are outside place, time, and individual variation, forming a real but ideal existence "within experience" and equated with linguistic universals, logical (necessary) truths, and essential or generic reference (category form classes). Let us take a look at this tour de force.
GENERIC REFERENCE AS IDEAL INTENTIONAL STRUCTURES
"Individual reference" is a specific, context situated communicative act. This is obvious when the object is given to perception (e.g., that red pencil over there). But it is also true in thought, imagination, or
utterance (sentence) as when one thinks of a red pencil, or even when one reflects upon the meaning of the sentence that contains "the red pencil" as its topic. This is because objects, both perceptually and cognitively, remain particular and individual, i.e., have particular characteristics, and exist in space and time. In contrast, "generic reference" involves ideal objects (structures) only. That is, particulars of individuality and spaceÐtime location have been "bracketed out" (Husserl) leaving behind in experience an "idealization" (Kates) which has real existence, but outside placeÐtime. Yet it is observable when it functions as intentional structures (ideal meaning).
Carol Kates' important contribution is apparent here. This notion that generic reference is transformed into observable intentional structures gets empiricists in the language sciences off the hook. It gives them a new and powerful tool to investigate language use, linguistic intuition, communicative competence, and speech acts. Her proposal may be outlined in three steps as follows:
a. the phenomenological premise: There is "a link between logic and ontology" such that only that which is logical or necessarily true can organically exist in experience. Thus, a contradiction cannot exist in experience either as individual or ideal; instead, a contradiction has a hypothetical existence, not a real one. Logical truths exist as real, but ideal structures, and are available for observation in experience. Contingent truths, which are variable and psychological (e.g., beliefs) exist as real, but individual structures, and are also available in experience (e.g., memories).
b. the empiricist premise: All utterances or individual acts of reference (speech acts) have meaning (are interpretable) not by analogy (or generalization) but by pattern matching and recognition between the individual and ideal meanings of the utterance. In other words, communicative competence is a matching skill between individual and generic, both being given in experience.
c. the pragmatic premise: The communicative function (meaning) of an utterance is nothing else than "the intentional structure" of the speaker. Utterances in context are produced in three simultaneous phases as follows:
First, there must occur an intentional structure in experience, such as when the context or the topic focus is perceived and a comment or reaction is elicited, either in speech or in thought and imagination. These intentional structures are built up in experience and acquired through bracketing operations, as pointed out before. Intentional structures are idealized, generic, universal objects that are shared by the members of a speech community ("langue") and correspond to our logical intuitions. Second, there must be available a stock or repertoire of paradigmatic structures (linguistic categories) which may be lexical, semantic, or syntactic). These stock patterns are also built up and acquired in experience (perception and cognition). Third, there must be individual acts of performance or presentation which make use of these stock patterns.
In summary, Kates' proposal analyzes communicative competence into a threefold unit of behavior. The performance of the act or, the presentation of the utterance in context, is the ultimate effect produced by two prior causes, arranged in series. The immediate cause of the production of an utterance is the cognitive activity of selecting from among the stock patterns from the speaker's repertoire. The original cause of the utterance (its essential meaning) is the speaker's intentional structure, built on individual experience but idealized and existing invariantly (logical intuition). Thus, synthetically viewed through a performance model, Kates' proposal is that a speaker's ideal intentions cause pattern selection and matching to produce interpretable utterances.
IMPLICATIONS FOR SECOND LANGUAGE TEACHING
Carol Kates' proposal of a marriage between phenomenology and empiricism in the language sciences seems to me to have revolutionary implications for second language teaching. By way of demonstrating this claim, I'll present two or three basic issues that concerns us collectively and indicate how this new approach provides very interesting solutions. 1.Where do utterances come from? What i~ communication? The pragmatic account stimulated by Kates shows me that human communication has three constituents: the origin of an utterance; the intermediate of an utterance; and the outermost of an utterance. These three correspond to goals, means, and uses, respectively.1 Goals are the innermost of an utterance. They are to be observed in "intentional structures." Psychologically, these correspond to stimuli, impulses, drives, needs, wants ÐÐÐ i.e., striving or motivational issues. Raising a question, asserting something, arriving at a judgment, are "predicative acts" which are responses to the dynamic cognitive implications of the context (environment). This is where the utterance originates from and has its dynamic, spontaneous character. So the very origin of an utterance (its essential meaning) is the speaker's intention.
Consider the following scenario. A question is raised or some contention is perceived; this contention has some relation to an "ideal structure" from prior experience; under the striving issue, this is transformed into an observable "intentional structure." Next this intentional structure finds a corresponding match in the next or intermediate constituent from the available stock of grammatical or cognitive patterns and acquired before; this is the means and is called ''grammatical knowledge" or "linguistic intuition." Finally, the outermost constituent of the utterance is reached, and this has corresponding relation to presentational features (mappings); these are the ultimate effects or uses of the utterance.
2.What is the relation between competence and performance? In the view I've outlined, performance models deal with the ultimate effects or uses in human communication. These are called "PÐmodels" and involve the outermost or surface features of utterances in context. The data in PÐmodels are those features of utterances that are explicitly marked, on the surface. We can call this the form of the utterance. Next;, competence models dip below the surface into underlying facts and implications. This is a cognitive sphere and "CÐmodels" deal with cognitive structures. Here we find grammatical patterns and stock phrases which are shared in common by a speech community. These facts belong to the intermediate constituent of an utterance (means) and may be called the structure of an utterance. Finally we dip still deeper into the innermost constituent of an utterance, what we can call the function of an utterance. These are to be dealt with by what we might call from now on "IÐmodels" which are "intentional structures" and as was explained before, are observable "ideal" objects built up in experience through phenomenological "bracketing."
It is clear therefore that an utterance in context is constituted of three simultaneous but distinct spheres of operation requiring three distinct data models, as outlined just above. These three data models are in correspondence relation to each other, i.e., are arranged in a (vertical) ontological series. The function of the utterance is its origin and its deepest, consists of observable intentional structures having relation to generic universals (essential or ideal meanings). The structure of the utterance is given by the speaker's or the speech community's available ,repertoire ("corpus") and relates to competence. The form of the utterance is given by the speaker's or situation's presentational matches or mappings. These are the surface constituents and relate to performance.
3.What is the predicative (topic/comment) function of speech? In the view I've outlined, the deepest level of an utterance is its function as given by "intentional structures." This belongs to the affective sphere of speech and language. It's what we usually call 'background knowledge" or "contextual information" and is the source for significatives, i.e., it is the essential meaning of signs. These function as pointers, indicators, deictics. They are what distinguishes between figure and ground. This deepest level of an utterance can be equated with the notion of topic in predicative acts (speech acts), namely that which is 'given' ('old information'). Kates mentions the term "nomination" for this deepest level of a linguistic function.
Next we have the intermediate level of depth of an utterance, what I've equated above with structure. It deals with competence issue. We may call this activity "association" and is evidently the comment constituent of predicative acts. It is organized at the level of cognitive structures (categories and classes). Finally, we have the surface level of an utterance, which is its forms and represents the conjoining operation whereby "nomination" and "association" appear simultaneously in the surface of the utterance.
Thus, predication is evidently a conjoining operation between topic functions at its deepest level and utterance forms at its surface level through comment structures at its intermediate level.
FOOTNOTE 1. My descriptions would be more easily grasped
with a system of graphic illustrations I use to depict the three depth levels of speech,
but there is not enough space for this here. Readers who are interested in this may write
for further particulars. The idea that language and human communication have three levels
of depth comes to me from Emanuel Swedenborg, the 18th century scientist/theologian whose
works in English are published by the Swedenborg Foundation, New York.
Back to Index of Articles
To a related article on Scientific Dualism