In lieu of an abstract, here is a brief excerpt of the content:

Reviewed by:
  • Postverbal behavior by Thomas Wasow
  • Georgia M. Green
Postverbal behavior. By Thomas Wasow. (CSLI lecture notes 145.) Stanford, CA: CSLI Publications, 2002. Pp. 185. ISBN 1575864029. $25.

Postverbal behavior is a gem of a book about variation in the order of phrases that follow the verb in English sentences, in particular, the phenomena discussed in the transformational literature as heavy NP shift, dative alternation, and particle movement. If this review has a conversational flavor, it simply is reflecting the relaxed and accessible expository style of the monograph.

Ch. 1 (‘Introduction’) poses the question that the studies described in Chs. 2–4 investigate: Can we define ‘weight’ in such a way as to make testable the Principle of end weight whereby language users prefer alternatives where phrases are ordered by increasing ‘weight’ (Quirk et al. 1972:14.8, Hawkins 1994) so that pronouns come earlier and long phrases later? Chs. 2–4 describe corpus studies and controlled judgment and elicitation experiments which indicate that the tendency for heavier phrases (however defined) to follow lighter ones in the three constructions in question reflects clear—but for the most part, not categorical—preferences. That is, despite confirming the general tendency, Wasow found substantial numbers of instances of heavier phrases preceding lighter ones, both unshifted heavy NPs, for example, and shifted nonheavy NPs. The last two chapters of the book reflect on theoretical, metatheoretical, and methodological issues raised by the investigations discussed in the core of the book. As dry as that might sound, and as interesting and thought-provoking as the first four chapters are, I found them the most impressive and exciting parts of the book. [End Page 327]

Ch. 5 (115–48) takes on three assumptions that have achieved paradigmatic status in generative grammar and argues that the results of the investigations discussed in Chs. 2–4 provide reasons to reject them all. The first is that rules (principles, laws) of grammar are categorical—they are about what is or is not allowed and not about what is more or less likely or probable. The second is that while the study of linguistic competence informs the study of performance, models of linguistic behavior cannot contribute to theories of linguistic competence. The third assumption W rejects is that the study of syntax is exclusively the study of I-language, to the exclusion of language as it is manifest outside the individual. The positions being rejected here might be overstated. While practicing generative grammarians have indeed rejected performance-based accounts of grammatical knowledge and inductive generalization from particulars, they have always depended on eliciting judgments, clearly a kind of linguistic performance, to test the empirical validity of hypotheses.

W’s interest, naturally, is in the mechanisms generative grammars provide for describing/explaining/correctly predicting1 word order possibilities of the three types he investigates. It is a sad commentary on the ignorance and small-mindedness of recent theoretical commentary that W found it necessary to comment on his use of the term ‘generative’ to include all varieties of formal syntax currently being pursued with an explanatory footnote:

It has become common to restrict the use of the term ‘generative’ to describe only work within some version of transformational grammar. I continue to use it in its broader sense to include nontransformational theories that share the original generative goal of providing fully explicit mechanisms (rules, constraints, principles, parameters, etc.) sufficient to characterize all and only the well-formed sentences of a language. In particular, I include such theories as Generalized Phrase Structure Grammar, Lexical-Functional Grammar, and Head-driven Phrase Structure Grammar (inter alia) under the heading of ‘generative grammar’.

(115)

W concludes that the principle of end weight ‘has categorical manifestations’ (such as the prohibition against a nonpronominal dative object preceding a pronominal theme object in declarative clauses seen in *give the men it), so that ‘trying to segregate facts about frequency from grammar evidently leads to missing a generalization here’ (128). This may be an overstatement; since language change happens piecemeal, with pieces not always defined as originally, it often leaves categorical islands in a stochastic sea. For example, in a number of Germanic languages, grammars that were strictly...

pdf

Share