Prolegomenon to a Theory of Argument Structure (review)

For those who are familiar with Ken Hale and Samuel Jay Keyser’s previous collaborative work, it was probably just a matter of time before the extremely influential ideas espoused in Hale & Keyser 1993, among others, were made more explicit, refined a little, and presented collectively in a monograph. But this book goes well beyond that. In addition to laying out the general framework at length and revising parts of previous analyses, this monograph contains new, follow-up work—the culmination of eighteen years of collaborative research, as the reader learns in the ‘Preface’ (ix–x), from the inauguration of the MIT Lexicon Project in 1983 to H’s untimely death.

Ch. 1 presents ‘The basic elements of argument structure’ (1–27) and constitutes a considerable update of the research program reported in Hale & Keyser 1993. The guiding intuition H&K pursue in their search for a theory of argument structure is to look closely at ‘the syntactic configurations projected by a lexical item’ (1), their definition of argument structure, yielding a syntactic process of derivational morphology (and beyond). Motivation for doing so comes from the desire to find an answer to the question of why there are so few thematic roles, such as ‘agent’, ‘theme’, or ‘goal’. Theta theory, as conceived in the principles-and-parameters framework (and government-binding theory in particular), may have something to say about how argument structure is licensed qua thematic roles—but it has rather little to say about why there are only two or three thematic roles and not, say, twenty-three.

‘Bound features, Merge, and transitivity alternations’ (29–45) are the concerns of Ch. 2, which explores further the implications of the limited types of argument structure in language, namely that argument structure needs more than just complement and specifier relations (see below for a brief exposé on H&K’s structural assumptions). By integrating recent views of the operation Merge (Chomsky 1995, 2000), H&K update their previous account of verbs like put or smear (cf. *Spurs put on Leecil, *Saddle soap smeared on my chaps) in that the complement of this verbal head is not some full predicative projection of P with a specifier, but the result of simple Merge between this P and its complement (in previous X-bar-theoretic terms, P′). This type of argument structure yields the well-formed We put spurs on Leecil and Leecil smeared saddle soap on my chaps.

A theory of ‘Conflation’ (47–103)—the ‘fusion of syntactic nuclei’ (47)—is presented in Ch. 3. Similar in spirit to incorporation, conflation differs in at least one important aspect: ‘a verb cannot “conflate” with the specifier of its complement’ (103), because conflation, due to its selectional nature, is a strictly local relation between a verb and the head of its complement. (Incorporation, by contrast, does allow the configuration of a specifier incorporating into the next higher head.)

The empirically rich Ch. 4 deals with ‘A Native American perspective’ (105–57) on argument structure. As H&K humbly put it, the ‘purpose [of this chapter] is merely to lend a certain degree of cross-linguistic perspective to this work’ (105). H&K discuss mainly the transitivity alternation in four ‘Native American linguistic traditions’ (105), the genetically and/or a really quite different languages Navajo (an Athabaskan language), Ulwa (Misumalpan), Tohono O’odham (aka Papago), and Hopi (both of which are Uto-Aztecan).

The remainder of the book deals with interesting extensions of the system. For example, H& K’s approach also has something to say ‘On the double object construction’ (159–88), going beyond the initial remarks in Hale & Keyser 1993, which they discuss in Ch. 5. The complex structure argued for is essentially one that takes a recursive, fully selecting, predicative head (i.e. a head that selects for both a specifier and a complement) to be the complement of a head that only selects for...