Reviewed by: The universal structure of categories: Towards a formal typology by Martina Wiltschko Karen Zagona The universal structure of categories: Towards a formal typology. By Martina Wiltschko. Cambridge: Cambridge University Press, 2014. Pp. 356. ISBN 9781107038516. $99 (Hb). The nature of functional categories has been a topic of long-standing interest for descriptive, typological, and theoretical linguistics. Essentially three approaches have been assumed, which differ with respect to whether the language faculty (universal grammar, or UG) provides a universal set of functional categories, and if so, whether every grammar has all of them or only a subset. The ‘variationist’ approach posits that languages may vary arbitrarily in their grammatical categories (Comrie 1976, Jespersen 1932); this is assumed to be possible without limiting the expressive power of the language because there are alternative ways of communicating the information that is carried by functional elements. The (strong) ‘universalist’ position assumes as a working hypothesis that grammars are essentially invariant, which implies that all functional categories that are available in UG are present in every grammar. This is the approach adopted in Cinque 1999 and in an extensive body of research in the ‘cartographic’ research program.1 A weaker universalist position claims that only a subset of the functional categories that are available in UG are present in any individual grammar. There has been lively debate in the literature over the past decade or more, both about the universality of individual functional categories and about the nature of crosslinguistic variation. Martina Wiltschko’s monograph presents a new approach to these questions, combining elements of the variationist approach (there is no universal [End Page 744] inventory of functional categories) with a new proposal for a different type of universal: a class of functional elements that are themselves devoid of substantive content (such as definiteness, number, tense, or aspect) but that provide the basis for constructing functional categories in individual languages. The claims are thoughtfully developed and are illustrated with data from a variety of languages. The centerpiece of the discussion is in-depth analyses of Halkomelem and Blackfoot, languages that have been studied extensively by W and collaborators (Ritter 2014, Ritter & Rosen 2010, Ritter & Wiltschko 2009, 2014, among others), and supporting evidence is drawn from comparisons with a number of other languages. W begins by stating the principal issues for a theory of functional categories: whether or not they are universal, and if they are not, in what ways they differ across languages. W also raises a methodological issue: since syntactic categories need to be motivated on distributional grounds, the fact that most category diagnostics are language-specific poses a methodological problem: that is, how is it possible to ensure that crosslinguistic comparisons are in fact comparisons of the same category? W’s proposals are introduced in Chs. 1–3. Ch. 1 argues that there are two language-independent tests for grammatical categories. One test is the existence of ‘multi-functionality’, such as the occurrence of perfect have in English alongside possessive have. W notes that, although these phenomena are usually analyzed either as accidental homophony or are attributed to grammaticalization paths, their similarity across languages implies that they reflect a universal process of recategorization. W incorporates this generalization into her approach, claiming that all functional categories are formed in particular grammars by processes of (re)categorization. The outcome of this process is that a category label mediates between the item (have, for example) and an interpretation. The nature of this process is discussed in Chs. 2 and 3, to which I return below. The second language-independent diagnostic is based on patterns of contrast in the value of features. In English, the morphological alternation between plural -s and its absence corresponds to a semantic alternation. In Halkomelem, the plural/∅ morphological alternation does not correspond to a semantic alternation. A zero-marked nominal may be either singular or plural in interpretation. W argues that the difference is based on whether plural marking is realized as a functional category. If it is, then the category will have semantic values regardless of the absence of morphological marking. If the language does not have the category Number, then the absence of overt marking is...
Read full abstract