Paperback
-
PICK UP IN STORECheck Availability at Nearby Stores
Available within 2 business hours
Related collections and offers
Overview
Originally published in 1983.
The Princeton Legacy Library uses the latest print-on-demand technology to again make available previously out-of-print books from the distinguished backlist of Princeton University Press. These editions preserve the original texts of these important books while presenting them in durable paperback and hardcover editions. The goal of the Princeton Legacy Library is to vastly increase access to the rich scholarly heritage found in the thousands of books published by Princeton University Press since its founding in 1905.
Product Details
| ISBN-13: | 9780691613550 |
|---|---|
| Publisher: | Princeton University Press |
| Publication date: | 07/14/2014 |
| Series: | Princeton Legacy Library , #439 |
| Pages: | 246 |
| Product dimensions: | 7.00(w) x 9.80(h) x 0.40(d) |
Read an Excerpt
Explanatory Models in Linguistics
A Behavioral Perspective
By Pere Julià
PRINCETON UNIVERSITY PRESS
Copyright © 1983 Princeton University PressAll rights reserved.
ISBN: 978-0-691-06524-3
CHAPTER 1
Psycholinguistic Context
1.1 Scientific progress requires the realistic demarcation of fields of inquiry. Yet there is often a certain degree of arbitrariness associated with the delimitation of related topics, as the proliferation of "hyphenated" disciplines — say, biochemistry, astrophysics, neurophysiology, etc. — in recent decades clearly attests. Sciences with a long-standing tradition or a reasonably well-established set of procedures, techniques, and vocabulary, combine their independent knowledge in order to solve borderline problems more effectively. Sometimes a new and relatively autonomous field emerges out of this interdisciplinary contact.
There is a fundamental difference, however, between this sort of cooperation, where the contributing sciences have already proved capable of generating a reliable body of knowledge, and what amounts to a juxtaposition of fields lacking that minimum of methodological stability necessary for cumulative progress. In the latter case, the result may well be an even greater confusion about the status of the problems involved and the means available for coping with them; more generally, greater conceptual chaos is likely to ensue. Writing about physics and biology more than eighty years ago, Mach (1897) discussed the possible outcomes of the first type of cooperative enterprise in the following general terms:
It often happens that the development of two different fields of science goes on side by side for long periods, without either of them exercising an influence on the other. On occasion, again, they may come into closer contact, when it is noticed that unexpected light is thrown on the doctrines of the one by the doctrines of the other. In that case, a natural tendency may even be manifested to allow the first field to be completely absorbed in the second. But the period of buoyant hope, the period of overestimation of this relation which is supposed to explain everything, is quickly followed by a period of disillusionment, when the two fields in question are once more separated, and each pursues its own aims, putting its own special questions and applying its own methods. But on both of them the temporary contact leaves abiding traces behind. Apart from the positive addition to knowledge, which is not to be despised, the temporary relation between them brings about a transformation of our conceptions, clarifying them and permitting their application over a wider field than that for which they were originally formed.
It is probably too soon to tell the extent to which psycholinguistics will follow the pattern described by Mach. After roughly thirty years of virtually complete dissociation, psychology and linguistics were formally brought together in the fifties. Since then, linguistics has undoubtedly played the role of "absorbing science": modern linguistic metatheory includes explanatory power as an adequacy criterion for particular theories of languages or grammars; it also specifies the properties to be met by those grammars, with a previously unknown degree of specificity and formal rigor. This has lent syntactic analysis an unprecedented prestige and the a priori formulation of related "explanatory" hypotheses an unwarranted degree of credibility.
Although there are undeniable signs of growing skepticism, the period of buoyant hope and overestimation that has characterized the psycholinguistic community for the past two decades cannot be said to have given way to disillusionment yet. It is quite possible that both fields will again recoil into their own areas of expertise in their treatment of verbal phenomena. It may be worth asking then what the abiding traces of this temporary contact and the ensuing positive addition to knowledge are likely to be, and in what way it can bring about "a transformation of our conceptions, clarifying them, and permitting their application over a wider field than that for which they were originally formed." Perhaps what we must ask first, however, is whether psycholinguistics is bona fide symbiosis of the kind alluded to by Mach or merely a complex case of juxtaposition. The best way to appraise the situation is to review, if in broad outline, the nature of this temporary contact and to weigh carefully the kinds of claims made by the engulfing field.
1.2 Osgood and Sebeok (1954) characterized psycholinguistics as the convergence of linguistics, information measurement, and learning theory. Not surprisingly, Saporta (1961) could still write: "it is clear that psycholinguistics is still an amorphous field, and the topics chosen as well as their arrangement represent only one arbitrary attempt at shaping a large body of available information about language." Adopting the above threefold demarcation two years later, Gough and Jenkins (1963) further pointed out the three aspects of "learning theory" relevant to psycholinguistics: (1) Osgood's theory of meaning, formulated in mediational terms; (2) Skinner's analysis of verbal behavior; (3) verbal associations, "a set of informal theoretical assumptions about verbal habits, elaborated and developed by individual psycholinguists into distinct theoretical conceptions." Presumably in an attempt to bring some order into this mélange, Osgood (1963) expressed the opinion that the important issues crystallize around the question of whether contemporary psychological theories can shed light on man's ability to create and understand new sentences. (Notice the uninhibited reference to the plurality of theories and the appeal to "sentences" as a priori analytical units. Osgood later became much concerned over the problem of "sentencehood.") A later characterization (Cofer, 1968) just about sums up the direction of subsequent trends: "information measurement is perhaps less influential than it once was; and of course, linguistics has recently become more influential, especially on the question of what knowing a language means for the psychology of the user of the language." This is a peculiar way of referring to psychology, one which in turn raises rather fundamental questions about practically all of the adjacent terms. It becomes perhaps clearer in the light of the following statement: "there has been relatively little effort on the part of learning theorists to deal with language as a system." Cofer then refers to various writers who, leaning heavily on linguistics, have concentrated on language acquisition and who "see their findings as making difficulties for current learning theory."
These developments can be attributed in large measure to the impact of Chomsky's speculations about the structure of language and the explanatory value claimed for his proposed model of transformational-generative grammar (TGG). Some underlying assumptions are important. Language is made up of an infinite set of sentences and its "use" is a matter of linguistic rule. The job of a grammar is to specify the structural features of this infinite set by finite means: a generative system of rules will meet the case. The evaluation of competing grammars is made in accordance with an overall metatheory that says what any natural language must be like; this general theory is identified with a model of the innate abilities responsible for language learning. The metatheory thus becomes an explanatory model. Moreover, speakers of natural languages constantly produce and understand sentences they have never heard before, and this unlimited use of language on the basis of finite experience suggests a relation between the finitary descriptive means proposed and mental processes of some sort. (The details, not always equally explicit, of Chomsky's proposal will be taken up below.)
These are strong claims, elegantly expressed as a general rule. To account for the "use" (that is, production and understanding alike) of novel utterances has become something like the raison d'être of psycholinguistics, with developmental studies conforming to specification as well. This is not to say that all psychologists concerned with language have embraced Chomsky's proposal uncritically or to the same degree. But it is fair to say that research on verbal behavior has been weakened (sometimes seriously) by this latter-day linguistics-dominated outlook: it is not unusual to see its findings contrasted with, if not discussed as a function of, the latter's demands and expectations as a matter of course (see, for example, Glanzer, 1967; Clifton, 1967; and more generally, Dixon and Horton, 1968). Mainstream psycholinguistics has found itself progressively committed to the "justification" of the formal outlook on language (see, for example, Miller, 1962,1965; and, despite Fodor and Garrett, 1966, Blumenthal, 1970; Fillenbaum, 1971; Johnson-Laird, 1974; Danks and Glucksberg, 1980). The new psycholinguists, as they were soon called, reversed the prevailing empirical outlook in rather drastic ways: the well-rooted tendency to keep close to the data and thus to concentrate on the emerging functional relations (for what they might have been worth) gave way to a frank priority to theoretical speculation; all too often, experimental work has become ancillary. Indeed, many linguists and a number of philosophers have automatically become experts on the psychology of language and, emboldened by their immediate impact, on psychology at large. Some aspects of psychology have necessarily been emphasized at the expense of others.
Whatever its intrinsic interest, the forcefulness with which the new standpoint has been put forth is probably not sufficient in itself to explain its overwhelming impact. Traditional psychological study of language was vulnerable; it was based on poorly understood behavioral processes and half-hearted formulations. Gough's and Jenkins' characterization of the study of verbal associations above must be reckoned as accurate. Similar remarks apply to the venerable field of verbal learning. The two traditions share a commitment to conditioning principles (within a narrow Stimulus-Response conception) and a moderate indulgence in theoretical speculation. Whatever measures of theoretical analysis are available have merely a local interest; even so, the gap between data and theory seems to be growing wider. These "miniature" theories are usually anchored in largely inconclusive results, and they are generally disconnected from one another. Moreover, it does not seem reasonable to expect any significant degree of integration in the future. The major outcome of a large-scale conference convened with this purpose in mind was a more direct scrutiny and ultimate rejection of the S-R frame of reference as well as a general consensus as to the need for "greater theoretical complexity." It should perhaps be added that this position was strongly reinforced by the presence of mainstream psycholinguists, who argued in the same sense on grounds of formal principle. Some participants rather unnecessarily appealed to perceptual variance, ethology, and computer simulation of cognitive functions to add strength to what should have been nothing more than a reasonable and straightforward recognition of empirical insufficiency.
Cofer (1968) points out "that students of verbal learning have concentrated their efforts on retention and transfer, that the psychological psycholinguists have emphasized meaning and associative processes (and, from information theory, sequential aspects of language), and that the overall system, which a language is, has not been given much consideration. It may also be said that human learning and retention, in some overall or broadly gauged perspective, has not been given attention either." The basic problem is less likely to be a neglect of language as an "overall system" (whatever this may ultimately prove to mean) than the obvious lack of procedural unity and theoretical direction in the study of learning among students of verbal learning and the psychological psycholinguists. (These are by no means the only areas of psychology similarly affected.)
It is partly understandable that against this background of experimental and theoretical chaos, the linguist's formally impressive, all-encompassing descriptions of "what is learned" should appear particularly enticing and (at least to some) his corresponding broader claims persuasive. Anyone looking at the proceedings of the aforementioned conference (Dixon and Horton, 1968) is bound to conclude that psychology at large, and the behavioral study of language in particular, are in serious trouble. A more informed point of view, however, puts the SR paradigm and the broad variety of traditional "problem areas" there represented in proper historical perspective and clears the way to a more realistic view of "language" as an object of behavioral science.
1.3 A generally recognized feature of the expository style of the new psycholinguists is the ease with which rival proposals have been habitually dismissed as "uninteresting" and "trivial," almost since the very inception of the TGG movement (see Verhave, 1972, for an exposition). Information theory, for example, was originally introduced as a framework for the regularization and possible quantification of language and other complex behaviors; as Glanzer (1967) also points out, it eventually played an important part in bringing into focus the formulation of theories of linguistic structure incorporating the assumptions listed above. Chomsky appropriately discussed the inadequacies of related finite-state models in his early work, and similarly, those of any form of assignment of grammatical categories to consecutive phrases roughly on the phrase-structure pattern (see chapter 3). Reference to these shortcomings became part and parcel of any argument in favor of the TGG model for years; typically, these references tended to become indiscriminate (and sometimes equivocal) as the movement gathered momentum.
Outside of linguistics, the (psychological) study of meaning should not count as an alternative position in itself; but it is rendered so by the extreme formalist's insistence that syntactic analysis must be performed first. The following early statement illustrates the subsidiary role accorded semantics in general and (to go back to Gough's and Jenkins' appraisal) to Osgood's work in particular: "To introduce the process of comprehension, however, raises many difficult issues. Recently there have been interesting proposals for studying abstractly certain denotative (Wallace, 1961) and connotative (Osgood, Suci, and Tannenbaum, 1957) aspects of natural lexicons. Important as this subject is for any general theory of psycholinguistics, we shall have little to say about it in these pages. Nevertheless, our hope is that by clearing away some syntactic problems we shall have helped to clarify the semantic issue if only by indicating some of the things that meaning is not" (Chomsky and Miller, 1963). Meaning is unquestionably a difficult subject. Any clarification of what it is not would be an important contribution in itself. But this is not what the reader is presented with, probably because the implied dependence of semantics on syntactic analysis cannot be upheld. Neither is prior to the other. Indeed, where natural languages are concerned, the entire notion of "abstract study" of lexicons may well prove meaningless altogether.
Criticisms have been particularly hostile with respect to behavioral science in general and what Chomsky understands to be Skinner's position in particular. They originate principally with Chomsky's widely reprinted review (1959) of Skinner's Verbal Behavior (1957), which went largely unchallenged for over ten years. MacCorquodale (1970) provided a comprehensive reply, in which he accurately characterized the review as "an amalgam of some rather outdated behavioristic lore including reinforcement by drive reduction, the extinction criterion for response strength, a pseudo-incompatibility of genetic and reinforcement processes, and other notions which have nothing to do with Skinner's account."
Chomsky missed the point of Skinner's experimental framework in rather fundamental ways; having misunderstood the "premises," against which his review is mainly directed, his conclusions concerning a strictly empirical extrapolation of laboratory findings to the verbal field could hardly be expected to be correct. For all its superficial brilliance, Chomsky's review gave a thoroughly distorted, if not specious, picture of Skinner's analysis. (MacCorquodale fittingly calls its tone "ungenerous to a fault, condescending, unforgiving, obtuse, and ill-humored.") It also provided a series of statements that have become, over the years, standard (rather repetitive and unimaginative) invectives against any unqualified view of language as behavior. Postal (1964a) offers a typical early example: "it is clear that enough is already known about the nature of language to show that views of language learning which restrict attention to the gross phonetic properties of utterances, either by adherence to psychological theories which do not countenance concepts more abstract and specific than 'stimulus,' 'chaining,' 'response,' etc., or linguistic structure representable by final derived phrase markers, cannot teach us very much about the fantastic feat by which a child with almost no direct instruction learns that enormously extensive and complicated system which is a natural language." The operant formulation of behavior, verbal and nonverbal, is still usually referred to as an example of S-R psychology. Here too psychologists are partly to blame.
In fairness to Chomsky, it must be pointed out that he does acknowledge the existence of "certain nontrivial applications of operant conditioning to the control of human behavior":
A wide variety of experiments have shown that the number of plural nouns (for example) produced by a subject will increase if the experimenter says "right" or "good" when one is produced (similarly, positive attitudes on a certain issue, stories with particular content, etc.). ... It is of some interest that the subject is usually unaware of the process. Just what light this gives into normal verbal behavior is not obvious. Nevertheless, it is an example of positive and not totally expected results using the Skinnerian paradigm. (1959, fn. 7; his one reference is Krasner, 1958)
(Continues...)
Excerpted from Explanatory Models in Linguistics by Pere Julià. Copyright © 1983 Princeton University Press. Excerpted by permission of PRINCETON UNIVERSITY PRESS.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.
Table of Contents
- FrontMatter, pg. i
- CONTENTS, pg. vii
- ACKNOWLEDGMENTS, pg. ix
- INTRODUCTION, pg. xi
- CHAPTER 1. Psycholinguistic Context, pg. 1
- CHAPTER 2. Structuralist Background, pg. 19
- CHAPTER 3. The Transformational-Generative Proposal, pg. 42
- CHAPTER 4. Explanatory Models, pg. 65
- CHAPTER 5. Subsequent Refinements, pg. 93
- CHAPTER 6. Performance and Competence, pg. 110
- CHAPTER 7. Mentalism in Linguistics, pg. 126
- NOTES, pg. 143
- REFERENCES, pg. 204
- INDEX, pg. 219