The role of argument structure in grammar: Evidence from Romance By Alex Alsina (review)

BOOK NOTICES 459 dressing aspects ofMayan culture and history as well as the calendar, numbers, and writing system itself. Since the phonetic decipherment of the glyphs was a fairly recent event, only a few general books on writing systems treat them adequately, as this one does. The final and longest section of the book deals with writing systems in use today and their precursors . R discusses the Chinese and Japanese scripts in most detail ('Chinese writing' 182-97, 'Japanese writing' 198-209), and the alphabet, which he traces from its Phoenician beginnings to its descendants in various comers of the world ("The first alphabet' 158-67, 'New alphabets from old' 168-81). Onedisappointment is that R neglects the scripts of India, remarkable for the varied forms they take, and the Korean writing system (Hangul), arguably the most sophisticated in the world. Each is given a mere page, and one cannot help but think that they deserve more discussion in a book of this sort. The concluding chapter is entitled 'From hieroglyphs to alphabets—and back?" (210-17). R dismisses those scholars who argue that 'with enough imagination and ingenuity, a system of signs could be expanded into a "universal" writing system . . . capable ofexpressing the entire range ofthought that can be expressed in speech' (21 1). He is equaUy suspicious of the view that the alphabet is superior to logographic or syllabic scripts. This colorful and well-written book deserves a spot on the bookshelf of any true aficionado of writing systems. [Kirsten Fudeman, Cornell University.] The role ofargument structure in grammar : Evidence from Romance. By Alex Alsina. (CSLI lecture notes, 62.) Stanford: CSLI Publications, 1996. Pp xi, 306. $69.95. This monograph is a detailed and very readable study on argument structure and its relation to other levels of grammatical representation. Although the focus lies mainly on Catalan, most of the analyses are intended to hold for other Romance languages as well. The book consists of nine chapters including an introduction and a summary of the main conclusions. In Ch. 2, 'The theoretical framework' (15-79), Alsina presents the theoretical foundations, a version of lexical functional grammar (LFG), which differs considerably from other LFG versions. For example, the correspondence between c(onstituent)-structure and f(unctional)-structure is not governed by functional equations but merely constrained by mapping principles. Furthermore, a(rgument)-structure maps to f-structure in the syntax rather than in the lexicon. A-structure is thus conceived of as an autonomous level of syntactic structure connecting the lexicalsemantic representation of a predicate and its syntactic subcategorization. Building on ideas of David Dowty, A assumes that arguments are merely specified as either Proto-Agents or Proto-Patients or neither or both. Ch. 3 and Ch. 4 deal with reflexive clitics in Romance . In Ch. 3, "The Romance reflexive clitic' (81-1 14), A presents evidence showing that the subject of a reflexive construction like Catalan La directora es defensa "The director defends herself corresponds both to the internal and the external argument of the predicate. In Ch. 4, 'Solving the paradox ' (115-47), it is argued that these facts do not create a paradox unless the relation between arguments and grammatical functions is required to be one-to-one. According to A's analysis the Romance reflexive clitic functions as a morpheme which triggers the coindexation of two arguments in the astructure of the predicate it combines with ('a-structure binding'). Both arguments are then linked to the same grammatical function and, as A points out, to the same 'semantic participant.' In Ch. 5, Objects and case marking' (149-84), A distinguishes two types ofobjects in Romance claiming that the syntactic differences between them can be attributed to different case features, namely dative vs. nondative. However, the concept of morphological case A employs appears to be somewhat peculiar as he regards the preposition a in Catalan (and other Romance languages) as a case marker. In Ch. 6, 'Causatives' (185-235), A proposes to analyze causative constructions in Romance as involving predicate composition, which means that the a-structures ofthe causative and the embedded predicate are merged. Consequently, all causative constructions are analyzed as monoclausal structures. The a-structure...