Predicative possession (review)

All languages that have become known in some detail appear to dispose of conventionalized grammatical expressions for possessive notions. Possession has been described as a universal characteristic of human languages and its typology has been the subject of a number of book-size publications (in particular, Creissels 1979, Seiler 1983, Chappell & McGregor 1996, Heine 1997, Baron et al. 2001). This book is another attempt to provide a comparative analysis of such expressions by using a typological framework of morphosyntactic analysis.

The book is divided into three parts. Part 1, consisting of seven chapters, deals with the typological diversity of predicative possession. Ch. 1 provides the background of the study. It introduces relevant distinctions and delimits and defines the domain of the study. Stassen decides to restrict discussion to 'indefinite predicative possession' (as in Ann has a computer), which means that neither 'definite possession' (The computer belongs to Ann) nor attributive possession (Ann's computer) are within the scope of the book. There are further restrictions: it is only alienable possession that is considered, and only topical possessor noun phrases and possessee phrases that are not modified or quantified (35). The typological analysis rests on a sample of 420 languages of world-wide distribution that can be assumed to be representative of the world's genetic and typological diversity of languages.

The typology that S proposes is formal in nature, based on the encoding of the grammatical functions of the possessor and possessee. Four basic types of predicative possession are distinguished, and S is satisfied that these are the only types that he needs for his typology (48). These types, discussed in Ch. 2, are the locative possessive, the 'with'-possessive, the topic possessive, and the 'have'-possessive. Deviations from the expected structures of types are discussed in Ch. 3, while the processes of diachronic change that possessive constructions may undergo are the subject of Chs. 4, 5, and 6.

Part 2 is the main part of the book, where the factors determining the structure of predicative possession are discussed. Ch. 8 deals most of all with temporal sequencing and deranking and their relevance to the analysis of possession, and for each of the types distinguished, a universal of predicative possession encoding is proposed (274). The framework of analysis is illustrated in the subsequent Chs. 9-12. Each of these chapters is devoted to one of the types, and treatment is divided up according to geographical regions or genetic groupings.

Part 3 contains only one chapter, in which a model for encoding predicative possession is proposed. Two appendices conclude the volume. One of them provides an alphabetical list of all languages of S's sample with bibliographical and geographical information (Appendix A), while the other groups the languages according to their genetic affiliation and the possessive types found in them (Appendix B).

The main concern of the book is with the syntax of possessive and related constructions. S decides to ignore semantic criteria in the construction of his typology, maintaining that 'it will not be appropriate to employ anything else than formal criteria in the construction of the typology' (39-40). The practice adopted is 'to define the domain first in terms of a range of semantic phenomena, and then to use formal criteria to limit the domain to a set of constructions that is crosslinguistically manageable' (4).

S does not aim centrally at establishing typological classes based on necessary and sufficient criteria. For example, his reasons for diverging from some previous authors in not distinguishing a separate class of 'adnominal possessives' (i.e. genitive possessives in other terminologies) include observations such as that 'at least a significant number' of alleged adnominal possessives are instances of his locative or topic possessives with reference to criteria of constituency, or 'at least a number of them' derive their formal features historically from locational possessives (135). Typologies are, as S puts it, 'fuel' for explanatory theories. That is, he is primarily concerned with explanation: 'typological distinctions are not an end in...