Oxford Handbook of Linguistic Typology (review)

Abstract

Reviewed by: Oxford Handbook of Linguistic Typology Jeffrey Heath The Oxford Handbook of Linguistic Typology. Jae Jung Song, ed. Oxford: Oxford University Press. 2011. Pp. xxi + 704. $150.00 (cloth). There are eleven other linguistic volumes in the Oxford Handbook series. This one contains thirty chapters, from ten to forty-five pages each. It is especially useful for readers who are not themselves engaged in mainstream typology but who wish to follow or catch up on it. For this journal, instead of a technical review aimed at typologists, or complete coverage of the chapters, a selective appreciation with a linguistic anthropological flavor is appropriate. If by mainstream typology we mean the inheritors of Joseph Greenberg’s first-order worldwide codings of grammatical data points (e.g., constituent order, morphological agreement), and second-order correlations among them, the main recent news has been the intrusion of geography and history. This “epiphany” (as the editor calls it on p. 3) began around 1990 with work by, e.g., Matthew Dryer and Johanna Nichols. It had once seemed that worldwide typology would flourish as conclusions drawn from Greenberg’s thirty languages were refined by the addition of hundreds of new languages; Chomsky’s absolute universals would be supplemented or replaced by powerful probabilistic correlations. As it happens, worldwide typology has run into snags. Careful studies have found systematic macroregional biases, making even the ideal of a global typology with bias-purified results deeply problematic. This issue is covered in Dik Bakker’s chapter on “Language Sampling” (pp. 100–127), but it crops up in several others. And geography leads to history. If each macroregion has its own trait distributions and correlation patterns, this can only be explained by an unexpectedly profound conservatism of grammatical patterns, going back far beyond the brick wall of ten thousand years that has plagued comparative historical linguistics. Edward Sapir lives. And a number of seemingly robust correlations from the Greenberg era have imploded. For example, the correlation between noun-adjective and verb-object order has not held up (Song on “Word Order Typology,” pp. 253–79). In this instance, the problem can be solved by refining the theoretical analysis (an object is phrasal, while an adjective is not). But typology faces weekly crises as new reference grammars from the Amazon and New Guinea stream out, often revealing categories and constructions that resist labeling and coding (Patience Epps on “Documentary Linguistics,” pp. 634–49). Greenberg-style typology is bureaucratic; it needs cross-linguistically codable traits based on reliable etic grids, fed into large databases that output powerful correlations. But what if the codings are forced, or if the correlations are unimpressive (sixty to forty instead of eighty-five to fifteen)? One consequence of the bureaucratization of the subfield is a striking lack of interest in extreme languages. Pidgins, creoles, and sign languages are not listed in seventeen pages of language and subject indexes at the back, and I noticed only one or two passing references in the text. Perhaps there are good (bureaucratic) reasons for these omissions. The recent controversies about Pirahã played out in the pages of Language and The New Yorker are too recent to have been covered in a volume several years in the making. However, the neglect of configurationality is hard to excuse—it gets one brief footnote. Yet this is the most crucial dividing point among human languages, since the absence of multiword phrasing has consequences for the entire grammar and lexicon (think of quantification and negation, for example). It deserves to be prominently featured in a [End Page 287] subfield whose major aim is to document the range of human language structures. A chapter on configurationality would have made a nice complement to Walter Bisang’s chapter on “Word Classes” (pp. 280–302), which describes languages that lack a well-defined noun-verb distinction, essentially subordinating lexicon to syntax and discourse. But there are not enough fully nonconfigurational languages to make interesting statistics, and it is hard to code their features. John Haiman (“Competing Motivations,” pp. 148–65) does deal directly with another kind of extreme language, namely, Cambodian, which he provocatively calls an “everything-drop” language. The extensive ellipsis affecting all stem...