Morphology and its relation to phonology and syntax Ed. by Steven G. Lapointe, Diane K. Brentari, and Patrick M. Farrell (review)

164LANGUAGE, VOLUME 76, NUMBER 1 (2000) Morphology and its relation to phonology and syntax. Ed. by Steven G. Lapointe, Diane K. Brentari; and Patrick M. Farrell. Stanford: CSLI Publications, 1998. Pp. viii, 440. Reviewed by Mike Maxwell, Summer Institute of Linguistics This book is the outcome of a workshop held at the University of California at Davis in 1995 concerning the relationship between morphology on the one hand, and syntax and phonology on the other. Included are the papers presented, a prepared response for most of the papers, and lightly edited transcripts of most of the discussion sessions. In the introduction, the editors do an excellent job of summarizing both the position papers and those of the commentators. The papers fall into two groups: those treating the interaction of morphology and syntax, particularly the realization and mapping of arguments, and those dealing with morphology and phonology, especially issues of haplology and syncretism within paradigms. The perspective throughout is that of generative linguistics. Issues which would have been more prominent a few years ago, such as the distinction between allomorphy and phonology, scarcely come up (although the issue of level ordering does, which is perhaps surprising given the predominance that nonderivational theories of phonology had achieved by the time of this conference). Rochelle Lieber ("The suffix -ize in English: Implications for morphology', 12-33) examines the semantics of the suffix -ize, abandoning her earlier analysis which attributed the properties of the derived form to the syntax of the construction. Rather, she concludes, lexical semantics calls for a different sort of representation from the hierarchical structures of syntax; derivational morphology cannot be treated as a syntactic operation. The representation which Lieber suggests for lexical semantics is based on Pinker 1989. The commentator, Patrick Farrell, argues that while Lieber's objections to the syntactic account are not fatal, the syntactic approach is nevertheless to be rejected on other grounds, namely that the structure and operations assumed under the syntactic approach are unmotivated. Hagit Borer ('Deriving passive without theta roles', 60-99) calls into question the assumption underlying a wide variety of approaches in which syntactic structure is seen as the projection of properties of lexical items, in particular of the argument structure of verbs. Instead, she argues, the syntactic structure is determined both by the meaning of the verb and by the aspectual properties of the predicate; as a result, 'there remains little motivation for projecting arguments of unaccusatives and unergatives in different syntactic positions' (62-63). The surface structure which is observed is instead the result of the movement of the verb's arguments out of unordered, nonhierarchical VPs into specifiers offunctional projections.' Differences such as those between unaccusatives and unergatives or among actives, verbal passives, and adjectival passives are directly related to the aspectual properties of the clause, properties encoded in functional categories . The connection between transitivity and aspect immediately brings to mind the situation in some split ergative languages where the split is determined aspectually (see e.g. Dixon 1979)—a point which is touched on by the commentator, Andrew Spencer. Unfortunately, the transcript of the discussion session for Borer's and Spencer's papers had to be omitted; apparently the recording failed. The next paper, by Peter Sells (The functions of voice markers in the Philippine languages', 1 1 1-37), concerns a related topic: can the typologically unusual case marking systems of Philippine languages be accommodated within a two-way split between accusative languages and ergative languages, or do they represent a third sort of system? With accusative languages, the subject is normally the highest ranking participant in the thematic hierarchy, often the agent. For a different argument to become the subject requires a passive construction in which the lower ranking argument is promoted to the subject position while the highest ranking participant in the thematic hierarchy is demoted to a chômeur status (to use the relational grammar terminol1 Borer considers, but rejects, the base generation of the arguments in the spec-of-functional category projections (see her footnote 5, pp. 66-67); but this seems to me a possibility deserving further consideration. REVIEWS165 ogy). In ergative languages, the normal choice of subject depends on transitivity, with...