This is the second edition of the first volume in a four-volume set under the general editorship of Daniel N. Osherson; the others are titled Visual cognition, Thinking, and Conceptual foundations. It is a second edition in name only. Its fourteen chapters are entirely new, and were chosen 'to reveal why language holds a special place in cognitive science' (xix). The book is suitable as a text in an undergraduate or graduate course or seminar on language and cognitive science for students who have already been introduced to the basic concepts of phonology, syntax, and semantics. Each chapter includes suggestions for further reading, all but two of which are annotated, and problems (but no answers). Several chapters also contain questions for further thought designed for students with the requisite background knowledge (e.g. Prolog programming) or for individual and group projects.
Half of the chapters are devoted to topics in grammatical analysis, and the other half to psycholinguistics. I discuss the grammar chapters first, each of which describes a problem and works toward a solution which attempts to engage the reader in the process. The three most successful of these are ch. 3 'The sounds of Mawu words' by Mark Liberman, which leads the reader step by step through the intricacies of word formation in Mawu, a Manding language of Ivory Coast, to a very satisfying conclusion; ch. 11 'Lexical semantics and compositionality' by Barbara H. Partee, which explores the semantic properties of adjective-noun combinations in English; and ch. 14 'Some philosophy of language' by James Higginbotham, a play in two acts which analyzes the syntax and semantics of the English bare infinitive construction exemplified by Mary saw John leave.
Two of the grammar chapters argue for the virtues of particular grammatical theories: ch. 9 'Computational aspects of the theory of grammar' by Mark Steedman, which leads to a particular view of constituent structure, using combinatory categorial grammar; and ch. 10 'The forms of sentences' by Howard Lasnik, which leads to a quite different view, using transformational grammar of the extended standard theory variety. No effort is made in the book to compare, much less reconcile, these or any other grammatical theories, though the issue is raised in the suggestions for further reading section of ch. 8.
Ch. 12 'Semantics' by Richard Larson is narrower than its title suggests, focusing on the entailment relations induced by the English quantifiers every, some, and no, and ends just as the story is getting interesting. Ch. 2 'The case of the missing copula' by William Labov suffers from the opposite difficulty. Labov not only delivers what his title promises in the first four sections of his chapter (25-48), but tantalizingly opens up an entirely new problem in the final section (48-52) which is potentially of greater significance than the problem of the missing copula in such Afro-American Vernacular English (AAVE) sentences as He fast in everything he do. Labov shows that the only forms of the copula that can be omitted in AAVE are the contracted forms of is and are, just as in the Standard English (SE) casual-speech pronunciation of Where're you going?. In other words, the grammatical systems of AAVE and SE do not differ significantly with regard to such forms of the copula.
In the final section, Labov goes on to discuss stressed been, which is 'one of many grammatical elements of AAVE that are not found in other dialects and are generally not recognized by speakers of other dialects' (51). Labov then raises the question why speakers of other dialects know nothing about these '"deeper" or "more different" aspects of AAVE' (apart from habitual be, which is occasionally recognized by white speakers), whereas they are generally cognizant of the fact that the copula may be omitted in AAVE. His answer is that in interactions between AAVE speakers and others, and in the portrayal of AAVE in the media, these aspects are not represented, whereas media presentations 'reflect more or less faithfully the variation in the use of the copula that we have studied here' (52). That is, the 'African-American English of the media is a familiar idiom, a part of the joint world shared by black and white citizens of the United States'. Consequently, the declaration that AAVE (or, more recently, Ebonics) is a separate language from SE is met with incredulity in both white and black communities, a perception that is not helped if the only feature that proponents of AAVE as a separate language cite as a distinguishing property is the omission of the copula. But if AAVE is possessed of numerous distinctive grammatical elements, then its status as a linguistically distinct entity deserves serious public consideration, especially as it relates to the teaching of SE to native speakers of AAVE.
I turn now to the seven chapters on psycholinguistic topics. These are, with one exception, survey articles. Ch. 1 'The invention of language by children' by Lila Gleitman and Elissa Newport, and ch. 6 'Language acquisition' by Steven Pinker together thoroughly discuss the general problem of language acquisition. Ch. 4 'Exploring developmental changes in cross-language speech perception' by Janet Werker manages to be comprehensive in fewer than twenty pages, while ch. 7 'Speaking and misspeaking' by Gary Dell does the same for speech-error research on English in just over twenty pages. Ch. 8 'Comprehending sentence structure' by Janet Dean Fodor, and ch. 13 'Brain regions of relevance to syntactic processing' by Edgar B. Zurif focus on the recognition of empty categories in English, specifically those resulting from wh-movement, in the first case by normal adults, and in the second by those suffering from Broca's aphasia.
Ch. 5 'Why the child holded the baby rabbits' by Steven Pinker is the one problem-oriented psycholinguistics chapter; it focuses on the problem of overregularization in young children's speech, e.g. why English-speaking children on occasion say holded instead of held. Central to the solution is the 'Blocking principle', which Pinker does not define, but which he first describes as 'a piece of adult psychology that causes the experience of hearing an irregular form like held to block the subsequent application of the regular "add -ed" rule to that item' (111) and subsequently as 'part of the child's innate universal grammar' (114), together with the evidently psychological principle that memory retrieval is fallible, but improves with increased exposure to exemplars.
Pinker contrasts this solution with one which uses analogical pattern matching as implemented by a parallel distributed processor (PDP), and argues that the solution involving the Blocking principle and the fallibility of memory is superior, except for cases in which past-tenses are formed on analogy with other irregular past-tense patterns, e.g. forming brang as the past-tense form of bring on analogy with sang as the past-tense form of sing (124-7). He suggests a resolution of the 'ongoing debate in cognitive science over whether rule systems or connectionist networks are better models of language processes. Both approaches might be right, but for different parts of the mind: rules for the combinatorial system underlying grammar, networks for the memory underlying word roots' (128).
This solution is unlikely to please partisans on either side. In any event, Pinker's solution is marred by his elevation of the Blocking principle to the status of a linguistic universal. If it is a psychological mechanism, as he first suggests, it does not rule out the possibility that a given verb such as light can have distinct past-tense forms such as lit and lighted. At most, it imposes a strong tendency for verbs to have unique past-tense forms. However, if it is a linguistic universal, as he later suggests, then that possibility is incorrectly ruled out. Hence it cannot be a linguistic universal, at least not an inviolable one.