Date: Thu, 06 Feb 2003 12:27:44 -0500 From: Phaedra Royle Subject: Review: The Language Organ: Linguistics as Cognitive Physiology
Stephen R. Anderson, and David W. Lightfoot (2002) The Language Organ: Linguistics as Cognitive Physiology. Cambridge University Press, Paperback ISBN 0-521-00783-6, xix+263pp, $24.00.
This book was announced on LinguistList Issue 132382 Web link http://linguistlist.org/issues/13/13-2382.html
Phaedra Royle, School of Communication Sciences and Disorders, McGill University.
In The Language Organ (hereafter TLO), Anderson and Lightfoot aim to position Linguistics squarely into the realm of biological sciences, where the authors believe it should be. The authors want to convince the reader of the biological basis of the cognitive faculty we call "language", and show that the organization of the language "organ" is complex. The authors address in turn, the history of language study, basic issues in linguistic research, syntax, phonetics, phonology, morphology, language change, language acquisition and the biology of language in order to provide support for the notion that the study of language "...is every bit as "physiological" as the study of processes that take place in the kidney, even though the "organ" involved is defined...in functional rather than anatomical terms."(xiii)
To linguists, the innateness hypothesis does not come as a surprise, and is often implicitly assumed. However, we don't have to search as far a field as the general public to find a significant body of people who do not believe that there necessarily is an innate language faculty. A case in point is the domain of cognitive psychology, where language representation and processing theories run the gamut between connectionist (where patterns arise as a consequence of the input, but are in no sense innate rules) and rule-based models, with dual-route models (incorporating rule- and association-based processes) positioned somewhere in the middle ground between the two extremes.
The reader's interest will be piqued by the comment, stated boldly in the book's forward, that "we need to understand the "processes, activities and phenomena" characteristic of our mental life, the organization, development, and interaction of cognition. ... to do that we have to look at the right things; and in the case of language, that means studying language as a highly structured kind of knowledge [or I-language] rather than as a collection of external events [or E-language]." (xiii) Anderson and Lightfoot proceed to analyze research in different sub-domains of Linguistics with the stated goals of convincing the non-specialist of the biological origins of language, and of reminding the specialist of the importance of the innate and internal nature of language, not to be confused with it's external manifestation(s).
The premise of TLO is that Linguistics does not belong in the Humanities, nor in the Social Sciences but rather in the "hard sciences" on par with Biology. This is based on the notion that Linguistics is not the study of E-language (or external manifestations of language) but rather I-language (the internal "knowledge" of language). The book is intended to show (or remind) the reader what the difference is between E- and I-language and how to avoid the E-language trap. For example, in Chapter 5 "Describing linguistic knowledge" the authors present arguments against the use of phonological rules in favor of Optimality Theory (hereafter OT), while presenting loanword data from Fula. Fula is a language spoken in West Africa and contains many borrowings from French. As those who have studied borrowing know, borrowing languages modify the phonological structure of loanwords in order for them to obey the phonology of the borrowing language. Anderson and Lightfoot argue that rule-based accounts of loanword adaptations are insufficient as descriptions of I-language. If a language (like Fula) does not independently have a representation for a specific phonetic segment (say, /v/), there is no reason for the existence of a rule in that language converting /v/ to something else. Anderson and Lightfoot conclude that "[t]he regularities of sound structure seem to have a reality whether or not the language can be argued to contain rules to modify its own forms so as to make them conform."(97) They conclude that the constraint-based approach of OT manages to capture the universal, cross-linguistic, and innate nature of speaker's linguistic representations.
It is interesting to note the parallels between OT and Connectionist approaches to language representation and processing. Both integrate a notion of evaluation of multiple outputs and the idea of the best fit being the output of the system. However, major differences do exist. Most importantly, OT specifies innate and universal conditions on well-formedness (NoCoda and so on), while Connectionist approaches believe in the emergent nature of linguistic "rules". That is, patterns appear based on the input rather than on innate properties of the language organ. Despite these differences, it appears to me that OT theorists could (and should) use the computational tools developed by Connectionist modelers in order to be able to model their theoretical assumptions, and to see which account will provide a better match for linguistic data in domains such as acquisition, loan-word borrowing (and probably others). But I digress.
Anderson and Lightfoots's book arrives at a time where a number of linguists are reconsidering issues in the field (see for example, Hale and Reiss, 2000, and Jackendoff, 2002). In particular they question what Linguistics is, how we study language, what type of data is relevant to the study of Linguistics, and where Linguistics is heading as a science. The authors review a number of different areas of linguistic investigation (syntax, phonology, morphology, acquisition, language change, etc.) while attempting to provide a unified, theoretical, approach to the study of Linguistics. The only area they do not explicitly address is semantics. In fact, this book could be considered complimentary to Jackendoff's Foundation of Language, as the authors have decided to leave aside a thorough discussion of semantics, referring the readers to previous Jackendoff articles (his more recent book had not been published at the time of publication of the present book). Anderson and Lighfoot's goal is to present an account for a large number of linguistic data with a restricted number of innate constraints (on syntax, phonology and so on).
Considering the goals of the authors, the book is quite succinct - a bit less than 300 pages for ten chapters. However, the theoretical arguments presented are thorough and to the point. No content is lost in the name of brevity. The discussion is thorough and elegant. For example, in the syntax section, they manage to account for a number of seemingly very different phenomena (constraints on 'that' deletion, traces, incorporation, 'wanna' contraction, and so on) with a few simple principles. Over and over again, they show that approaching the study of language from the point of view of E-language is misleading because of the isomorphy problem-linguistic knowledge often is not isomorphic with external manifestations.
On a number of occasions, Anderson and Lightfoot make cavalier statements about language without seriously backing up their claims. For example, they state that children come to language with the predetermined knowledge on "the nature of clitics, and the fact that they cannot by themselves satisfy the requirement that phrases be represented by at least one phonological word [and that this] is contributed by the linguistic genotype and is part of what the child brings to language acquisition." (31) It might be the case that children have the ability to recognize clitics based on innate knowledge, but this is a bit of a leap of faith. It is highly improbable that children have a representation for "clitic". I know this is not what the authors are trying to say but their phrasing is unfortunate, and the inexperienced reader will find this type of assertion daunting. Another example can be found in chapter 6 (Phonetics and the I-linguistics of speech) where the authors endorse a Motor Theory of Speech Perception without discussing why, but rather stating that they are convinced by the arguments presented by proponents of Motor Theory (Liberman & Mattingly 1985, Mattingly & Studdart-Kennedy, 1991). One wonders what is the point of presenting a theory that is not discussed. It does not add anything to the book because it is simply a statement. No implications can be drawn from it.
TLO is intended for a number of different audiences. In spite of this, the author's statement that it is available to the uninitiated reader may be a bit misleading. The discussion of disadvantages of different approaches might seem a bit out of reach of the non-linguist, or even the neophyte. It would probably be more appropriate for an advanced Theories of Language class at the undergraduate level, or even a graduate seminar in Linguistics. TLO could be a useful tool for researchers with an understanding of linguistic issues (for example, neurologists, cognitive scientists, and so on) and that are interested putting their work into perspective, or in bringing their knowledge up to date in domains that are not their specialty. The chapters could also be used as quick and ready introductions to more recent areas of theoretical linguistic inquiry such as Optimality Theory and Minimalism.
A final caveat is that a number of typos were found in the text. Fortunately this has little consequence for general understanding. However, one typo was found in the French examples in the discussion of borrowing. This makes one wary of the other unknown language examples, since the non-native speaker cannot evaluate whether the examples and transcriptions are accurate, or not.
Hale, Mark & Reiss, Charles (2000) Phonology as Cognition, in Burton-Roberts, N., Carr, P., & Docherty, G. (Eds.) Phonological Knowledge: Conceptual and Empirical Issues. Oxford University Press, pp.161-184. Jackendoff, Ray (2002) Foundations of Language: Brain, Meaning, Grammar, Evolution. Oxford University Press. Liberman, A. M. & Mattingly, I. G. (1985) The motor theory of speech perception revised. Cognition, 21, 1-36. Mattingly, I. G. & Studdart-Kennedy, M., eds. (1991) Modularity and the motor theory of speech perception. Hillsdale, NJ: Lawrence Erlbaum.
ABOUT THE REVIEWER:
Phaedra Royle holds a Ph.D. in Linguistics from the Université de Montréal. Her interests lie in psycholinguistics, language disorders (Specific Language Impairment), language acquisition and morphology. Her thesis investigated lexical access in language-impaired French-speaking adolescents and adults. She is presently carrying out postdoctoral research on early language acquisition in French-speaking children with and without language delay, at McGill University in the School of Communication Sciences and Disorders.