- Complex predicates in Oceanic languages: Studies in the dynamics of binding and boundedness
The diverse collection of papers in Complex predicates in Oceanic languages constitutes both a geographical and typological complement to Terry Crowley's 2002 Serial verbs in Oceanic. Like Crowley's work, these studies are bottom-up, data-driven descriptive typologies within the tradition of Foley and Olson 1985, Crowley 1987, and Durie 1997. One should not be misled by the subtitle; the analytical frameworks employed owe very little to what used to be called "Government and Binding" theory. Most of the papers explore the muddy boundaries of verb serialization and compounding.
The collection originated in a workshop organized by the editors to focus on "serial and compound verbs in Oceanic languages" in March 2001—a bit too late for any contributors to react directly to Crowley 2002. Nevertheless, three papers on New Caledonian languages (by Isabelle Bril, Françoise Ozanne-Rivierre, and Jean-Claude Rivierre) help fill one major gap in Crowley's coverage. Four other papers on Melanesian languages enrich our understanding of the variety of related phenomena on Oceanic serialization's home turf: a look at core-layer junctures in Saliba, an SOV language of Milne Bay (by Anna Margetts); a glance at nuclear-layer junctures in Teop, an SVO language of North Bougainville (by Jessika Reinig); and two skeptical studies from northern and southern Vanuatu (by Alexandre François and John Lynch, respectively). François finds that (nuclear) serialized verbs in Mwotlap fill a syntactic slot within the "macro-verb" that is reserved for general adjuncts, "open to adjectives and nouns, plus many lexemes ('pure adjuncts') exclusive to this function" (115). Lynch finds that serialization is the least frequent conjoining strategy in , where echo-subject marking prevails, as it does in Crowley's 2002 exemplar from southern Vanuatu, Erromangan. It should be noted, however, that echo-subject constructions in southern Vanuatu languages perform many of the same functions that serial verb constructions do in northern Vanuatu.
It is not too surprising that serial skepticism also pervades the four papers on Polynesian languages: East Uvean (by Claire Moyse-Faurie); Pileni (by Åshild Næss); Tahitian (by Mirose Paia and Jacques Vernaudon); and Samoan (by Ulrike Mosel). In languages where most lexical roots are multifunctional, it is often hard to tell whether the noninitial element in a verbal complex is another verb, an incorporated noun, or an adverbial adjunct. Compare Tahitian inu huna 'drink and hide' or 'drink secretly' vs. huna inu 'hide and drink' or 'hide [one's] drinks' (256). This problem, also noted with respect to Mwotlap "macro-verb" slots, proved so intractable in the Polynesian languages that the editors elected to omit the word verbs from the title of the book, even [End Page 247] though it figured prominently in the workshop title (x). For both geographical and typological reasons, the Polynesian section could well be subtitled, "At the far edge of serialization." Unfortunately, the far northern edge—in Micronesia—remains entirely unrepresented in this collection.
Some of these boundary phenomena are intriguing. The closest thing to core-layer serialization in Polynesian languages seems to be found in Pileni, an outlier in the Solomons: Lha-ko toa lha-ko mot-ia te pakola la na (3DU-TA take 3DU-TA cut-TR ART giant DEM DEM) 'They cut the giant (to pieces)' (238). (It is not clear whether the same pattern is found in non-Polynesian languages of the Solomons.) In purely Polynesian territory, East Uvean allows a degree of nuclear serialization along the lines of 'He eats makes-noise his meat', but the "deserialized" alternative is preferred: 'Is noisy his eating of the meat' (211).
The most challenging of the Polynesian papers, however, is Mosel's study of "Juxtapositional Constructions" that are underdetermined for syntactic category. She finds (263) that "while noun phrases and verb complexes are well defined by the presence of articles and...