Functional constraints in grammar: On the unergative-unaccusative distinction (review)

The theme of this book is the unaccusative hypothesis (UH), or the division of intransitive verbs into unergatives and unaccusatives depending on the original structural position of their subjects. The UH is widely considered, particularly in generative grammar, to play an important role in accounting for a variety of syntactic phenomena, including the English constructions there, way, cognate object, pseudo-passive, and extraction from subject NPs. The main purpose of the book is to show that ‘the acceptability status of sentences in each of [these] constructions is basically independent of the unergative-unaccusative distinction, but is derived from complex interactions of a number of nonsyntactic (as well as syntactic) factors’ (25). Thus, the book is challenging to generative grammarians who accept the UH, as well as informative for functional grammarians who are interested in how the syntactic phenomena widely discussed in generative grammar are to be handled from the perspective of functional syntax.

The book is well organized, transparently written, and easily accessible without special technical knowledge of generative and functional theories. Kuno and Takami’s argumentation concerning the aforementioned five constructions is consistent throughout all of the chapters dealing with them. They first formulate the syntactic constraints relevant to the constructions that can be deduced from the UH, then present plentiful counterexamples to the syntactic constraints, showing that the constraints based upon the UH are fallacious, and finally propose alternative accounts from their functional approach. The counterexamples the authors provide are either attested in linguistic corpora, journals, newspapers, web pages, and so on, or they have been checked by reliable native speakers of English. Their functional constraints are extremely important as descriptive generalizations concerning the constructions that no linguistic approach aiming for descriptive adequacy should disregard. In what follows, I pose some questions in passing, but these do not detract from the highly valuable contribution the authors make to detailed description of the constructions.

After a review of the short history of the UH and its basic concepts in Ch. 1, Ch. 2 focuses on the there-construction, which is generally supposed to be possible only with unaccusative verbs, whose subject position is assumed by the UH to be originally vacant so as to accommodate the pleonastic there. As counterexamples to this generally supposed constraint, K&T present ‘inside verbal’ there-constructions containing some transitive verbs (e.g. cross the mind, enter the room) and some unergative verbs (e.g. rule, race). These inside verbal there-constructions and the ‘outside verbal’ ones with unergative verbs (e.g. There walked into the courtroom two people from London) are similar in that ‘the string to the left of its logical subject is interpretable as denoting existence or appearance’ (47), which is incorporated into their functional constraint on the there-construction.

A question soon arises as to how the constraint copes with one of the most basic there-constructions, that containing the progressive, passive, or copulative be. While such ungrammatical examples as *There are happening riots in China or *There are present some nonmembers seem to meet the functional constraint, such grammatical ones as There was a man being arrested or There are some nonmembers present do not. The constraint, designed as a set of necessary conditions for acceptability, makes no reference to functional/discourse conditions on extraposed constituents (namely, logical subjects), such as those of Ward & Birner 1995 and works cited in it, which the discourse constraint on extraposition from subject NPs in Ch. 6 refers to. Another problem concerns whether the uniform derivation of all sorts of there-constructions is compatible with the oft-observed fact that inside and outside verbal there-constructions (and some other types) behave differently with regard to subject-aux inversion, embedding under ECM verbs, and so on (Coopmans 1989, Rochemont & Culicover 1990).

Ch. 3 examines the way-construction, which is extensively considered to be possible only with unergatives, whose object position...