Reviewed by: Dimensions of possession ed. by Irène Baron, Michael Herslund, and Finn Sørensen Lise M. Dobrin Dimensions of possession. Ed. by Irène Baron, Michael Herslund, and Finn Sørensen. (Typological studies in language 47.) Amsterdam: John Benjamins, 2001. Pp. vi, 337. ISBN 1588110621. $95 (Hb). Based on a 1998 workshop, this most welcome volume is a collection of papers that look at possession primarily from typological, functional, and cognitive perspectives. As Michael Herslund and Irène Baron note in their introduction (1–25), possession is a difficult notion to unify, since it is not a type of linguistic construction, but rather a family of conceptual relations between entities (e.g. such that one is part of the other, is located near it, or controls it) that may be variously embedded in the grammar. Framing the remainder of the volume are chapters by two authors with established research programs on the theory of possession. The first, by Hansjakob Seiler (27–40), offers a schema of abstract possession relationships, a typology of morphosyntactic techniques that languages use for representing possession, and some reflections, derived from Cahuilla, on inalienability. After some summarizing metatheoretical remarks on possession research, the final chapter by Bernd Heine (311–28) presents a grammaticalization account of locative possessives in Khoe. Four of the chapters explore possession in Danish. Ole Togeby (41–55) surveys the range of formal expressions used for Danish possession and concludes that the concept is not, strictly speaking, grammaticized in the language. Finn Sørensen (57–65) sketches a theory of possessive semantics based on the creation of relative spaces, and Lars Heltoft (115–46) distinguishes the Danish indirect object from cosubject constructions in terms of part-whole relations. Irène Baron and Michael Herslund (85–98) use Danish examples to illustrate the semantics of the verb have, which they explain using a hierarchy of inclusion relations that is reflected in turn in grammatical organization. Another four chapters deal with modern Romance languages. Inge Bartning (147–67) looks at French NP de NP and elaborates a typology of interpretations critically distinguishing between the discourse level and the ‘micro-level’ of the complex NP. Henrik Høeg Müller (169–86) looks at the analogous Spanish structure from a cognitive perspective, rejecting possession as a useful semantic framework of analysis and focusing instead on the more fundamental relational meaning of de and the restrictive/nonrestrictive nature of the first NP. Martin Riegel (187–200) describes possession in French as comprising four types of semantic ‘participation’ relations, with the part-whole relation prototypical in its high degree of integration among the participants. Anne-Marie Spanoghe (227–42) shows that Portuguese differs from French and Spanish in tending not [End Page 829] to express a possessor for body part terms; her corpus-based study finds that where an indefinite article is used to introduce a Portuguese body part term, the possessor is likely to be indeterminate or nonexistent, the referent nonprototypical, and the term situated in a rhematic sentence position. Two further chapters by A. Machtelt Bolkestein (269–83) on Classical Latin and Per Durst-Andersen (99–113) on Russian compare the semantics of alternative cases used in each language to express possession. Although the papers are heavily skewed toward European languages, there are a few particularly sweet exceptions. William McGregor (67–84) explores predicative possession with the verb ‘have’ in the Nyulnyulan languages of Australia; Nyulnyulan ‘have’ covers a range of senses including ‘hold’, ‘keep’, and ‘grasp’ and appears to be more active than its Indo-European counterpart. Kari Fraurud (243–67) uses data from Komi, Udmurt, Turkish, and Yucatec Mayan to test (and critique) the hypothesis that possessives grammaticalize into definite articles. Her results are useful for clarifying two types of possession based on the discourse constraints on their referents. Marianne Mithun (285–310) considers Lakhota, Kathlamet, and Mohawk verbs that appear to express attributive possession with a bound affix; while these constructions might seem to be associated with...