Words and Rules: The Ingredients of Language by Steven Pinker

Abstract

Words and Rules: The Ingredients of Language by Steven Pinker New York: Basic Books: New York, 1999, 348 pp. Reviewed by Nancy Jones University of California, Los Angeles Words and Rules, the offer latest popular offering from the prolific MIT lin- guist Steven Pinker, although characteristically clever and well-written, does not first any surprises to those familiar with his other works. Pinker introduced the premise underlying Words and Rules then, is in his 1994 The Language Instinct: The way language works, rules that tal that each person's brain contains a lexicon (a of words and the concepts they stand for (a mental dictionary) and a set of combine (p. the words to convey relationships among concepts men- grammar, Words and Rules expands upon this premise and provides some new twists on an old theme, as Pinker focuses on regular and irregular verbs as a means to show that words and rules are the ingredients of language. Over the course of his argument, Pinker also attempts to rule out two com- peting language theories: Chomsky and Halle's (1991) generative phonology model and Rumelhart and McClelland's (1986) connectionist model. For those interested in the ins-and-outs of the past tense or an overview of the connectionist model as it applies to verbs, this book is worth a read. However, Pinker fails to convince the reader that connectionist models can immediately be dismissed — nor does he ad- grammar equately demonstrate cognition. how his own words and rules model gives us insight to Pinker's theory states that language consists primarily of a for pro- duction of novel utterances (rules) and a lexicon of memorized sound combina- tions (words). Pinker asserts that the grammatical behavior of regular verbs dem- onstrates how we use rules to generate words while the behavior of irregular verbs illustrate the role that memory plays in word generation. Pinker gives a later, historical explanation for the fact that certain irregular verbs have similar forms. that the irregular verbs originally He states were generated by rules and to through lan- guage change, became memorized forms. Comparing tive theory this words and rules theory Chomsky and Halle's genera- and to the connectionist/neural net theory of Rumelhart and McClelland, Pinker concludes that neither Chomsky and Halle's model nor Rumelhart and and irregular forms McClelland's model adequately explain the behavior of regular and irregular verbs. Chomsky and Halle's model asserts that both the regular Issues in Applied Linguistics ISSN 1050-4273 Vol. 10 1999, Regents of the Umversity of California No. 2,173-178