Reviewed by: Deixis and alignment: Inverse systems in indigenous languages of the Americas Denis Creissels Deixis and alignment: Inverse systems in indigenous languages of the Americas. By Fernando Zúñiga. Amsterdam: John Benjamins, 2006. Pp. 309. ISBN 9027229821. $180 (Hb). The treatment of inverse systems constitutes a particularly challenging issue in morphosyntactic typology. They are crucial for the discussion of typological hierarchies, which have been a major concern for typologists in the last decades, but their status in alignment typology is controversial. This book includes detailed descriptions of the relevant subsystems of five Amerindian languages or language groups, and a thorough discussion of the theoretical implications of the data presented for a general typology of direction marking and its interactions with other morphosyntactic phenomena. The book includes: an introduction; two chapters addressing general issues in the study of alignment, hierarchically based morphosyntactic mechanisms, and direction (Ch. 1, 'Alignment and direction'; Ch. 2, 'A theory of direction'); five chapters devoted to case studies (Ch. 3, 'Algonquian languages'; Ch. 4, 'Kutenai'; Ch. 5, 'Sahaptian languages'; Ch. 6, 'Kiowa-Tanoan languages'; Ch. 7, 'Mapudungun'); a conclusion (Ch. 8); three appendices on Algonquian paradigms, analysis of Kiowa personal prefixes, and optimality-theoretic syntax of inverses respectively; references; and language, author, and subject indices. The central part of the book provides precise and detailed descriptions of a variety of systems involving direction marking. The morphological intricacies that often make it difficult to choose [End Page 492] between two alternative analyses of the same data are presented as clearly as possible, but without any oversimplification. The accounts given by other linguists are thoroughly analyzed, and when the data allows for alternative interpretations (which is often the case), the competing analyses are carefully compared. After summarizing the basics of alignment typology as exposed in works such as Comrie 1981, Dixon 1994, Harris & Campbell 1995, and Dixon & Aikhenvald 2000, Ch. 1 discusses the way the classic model of alignment can accommodate various types of splits, and in particular the proposals to recognize a 'hierarchical type' (Nichols 1992) or an 'inverse type' (Klaiman 1992) as distinct from the basic types commonly recognized. The indexability hierarchy (Bickel & Nichols 2007)—also known as Silverstein's hierarchy, the empathy hierarchy, the animacy hierarchy, and so on, and omnipresent throughout the book—is described as involving an invariant core 'SAP [speech-act participant] > 3rd person pronouns > [+human] > [+animate] > [-animate]', but allowing for considerable crosslinguistic variation in further details, for example, the possibility to rank SAPs relative to each other as 1 > 2 or 2 > 1, or to leave them unranked. Most importantly, a single language may exhibit distinct variants of an indexability hierarchy with respect to different construction types, as discussed in detail for Algonquian: contrary to the commonly received opinion, the 2 > 1 ranking accounts for one particular aspect of Algonquian indexation systems only (access to prefix slot), but is contradicted by some aspects of the behavior of suffixed person marks. Direction is defined as reflecting 'the alignment between the indexability hierarchy and a relational hierarchy where A's outrank O's' (28), and Zuñiga insists that direction marking (direct vs. inverse) is a possible manifestation of the indexability hierarchy, logically independent from other possible manifestations such as access to marking slots or access to syntactic function. The emphasis on the necessity to clearly distinguish these notions in analyzing the morphosyntax of individual languages, and to consider the question of their interactions not as a theoretical but as an empirical and typological issue, is certainly one of the strongest points of this book. Some confusion may arise, however, from the fact that 'direction' is not always used in conformity with the clear and precise definition formulated on p. 28—which is somewhat surprising, given that, in general, Z is very careful with terminology. On p. 31, DIRECTION is defined as 'dynamic actional deixis', that is, the deictic status of couples of entities encoded as the A and P arguments of transitive verbs: 'he-me', 'you-us', and so on. The concept of DIRECTION2 (formulated on p. 31) involves no reference to any hierarchy, and the relationship between direction2 (defined on p. 31) and direction1 (defined on p. 28...