Reflections on Language

Reflections on Language is a reworking of two formulations of Noam Chomsky’s current research in linguistics, the first and larger part being an elaboration of the 1975 Whidden Lectures delivered at McMaster University in Hamilton, Ontario, and the second part being a revision of an essay printed elsewhere on the same general questions. Both parts show the broad outline and some details of what might be called the MIT school of linguistics, in the form their theories have taken from about 1973 to the present.

The book is presented as a nontechnical outline of Chomsky’s views, but if the reader understands “nontechnical” to mean something like Desmond Morris’ The Naked Ape or Julius Fast’s Body Lanuage, he will be baffled by Reflections on Language. The book demands at least an acquaintance with the work of philosophers and linguists over the last generation; it does not compromise on that demand. Chomsky’s style is severe to the point of being forbidding; he presumes the reader’s familiarity with many technical terms that the nonspecialist is unlikely to know, and his use of stipulated abbreviations cumulates through the second part until some paragraphs read like parodies of the very worst of academic styles. The conscientious reader who wishes to supplement what the book presents by reading the referenced works will frequently be stymied: roughly a quarter of the cited studies, including several crucial to Chomsky’s arguments, will be unavailable in even the most magnificently appointed university libraries. These include references to many works forthcoming at the time of the book’s publication, to more than a dozen MIT dissertations, and to a score of unpublished papers, obtainable only by writing to the author. But the reader may take some small comfort from the fact that even fulltime linguists share the difficulty of finding out exactly what is being done at Cambridge, Massachusetts.

But if Reflections on Language is not a popular book in the publisher’s meaning of the word, it is nonetheless an important one. It presents both linguistic and philosophic arguments, the linguistic ones dealing with a major change in Chomsky’s theory of language, and the more general philosophic ones discussing the importance of transformational-generative theories in solving problems of human knowledge.

The specialized linguistic arguments deal with what information is available to the hearer of a sentence (in an idealized situation) when that sentence is understood. In works published in 1955 and 1965, Chomsky outlined and elaborated a theory that postulated a three-part division of human language faculties. For convenience we may call these three the syntactic component, the semantic component, and the phonological component. It should be mentioned at the outset that all three are abstractions, parts of a theoretical formulation intended to account for human linguistic ability. Whether these components correspond to brain structure or activities of any kind is another question, a question to which neurophysiology can at present give no answers. In any case, the production of a sentence (so the early theory goes) begins in the syntactic component, which is itself divided into parts. One part is a set of rules that produces a kind of labeled parsing, or phrase-structure. These rules, which are said to generate a phrase-structure, are of the form S NP PP; the example here may be read as “a sentence is composed of a noun phrase (the subject) and a predicate phrase (the complete predicate of traditional grammar).” The phrase-structure rules incorporate, generally, the analysis of sentences of a traditional formal grammar. A second part of the syntactic component is the lexicon, a dictionary-like compilation that supplies lexical material to be attached to the phrase-structure at appropriate places. With many simplifications, an output of the syntactic component can be represented, for example, as (1):(1) [S [NP John ] [ PP [AUX past ] [VP [ V see ] [NP Sam ]]]]

The labeled bracketings provide information about the category and relationship of each item supplied by the lexicon: thus S labels the whole structure as a sentence; PP stands for “predicate phrase,” as explained above; AUX, or “auxiliary,” in this case contains only the marker for past tense, but for another sentence might contain auxiliary verbs or markers of aspect; VP, or “verb phrase,” groups the verb (V) see with any modifiers it may have; NP identifies John and Sam as noun phrases. The functions of the two noun phrases as subject and direct object are defined by their positions in the bracketing.

A third part of the syntactic component, the transformational rules, changes the phrase-structures in specified ways. In general, transformations add, delete, or rearrange material in the phrase-structure. Suppose the passive transformation is applied to example (1); the result is phrase-structure (2):(2) [S [NP Sam ] [pp [AUX Past be participle ] [VP [V see ] [NP by John ]]]]

In the standard transformational theory we are talking about, the two...

(The entire section is 2120 words.)


Books and Bookmen. XXI, June, 1976, p. 55.

Christian Century. XCIII, May 12, 1976, p. 466.

Economist. CCLVIII, March 27, 1976, p. 116.

Harvard Educational Review. XLVI, November, 1976, p. 645.

Spectator. CCXXXVI, March 6, 1976, p. 25.