argued to be associated with feminine gender. Une baie, for example, opens to surrounding waters, whereas un lac, which is landlocked, is “closed” and masculine . Other attributes are analyzed as being associated with words ending in a pronounced vowel (“rough skin,” like cornichon) versus a pronounced consonant (“smooth skin,” like banane). Of course, most words have multiple traits of this kind. Therefore, these attributes are said to be ranked in terms of saliency. For example , although une autruche is “flightless,” which is related to masculine gender, it is more importantly the largest of all birds and thus “unique,” a feminine property. Variable rankings of such attributes account for words that take both genders, words that have historically changed genders, and cases in which grammatical gender does not match with semantic gender (for example, une victime for a masculine victim). In this analysis, these properties frequently take on the form of oppositions between semantic features asserted to extend to other lexical fields and, possibly, to the organization of nominal classifications in other languages. Beckett makes a caveat for certain nouns denoting women that take masculine grammatical gender due to sociohistorical forces often considered antiquated today. This systematization is somewhat successful in explaining native speaker intuitions regarding grammatical gender. However, Beckett’s ambitious undertaking also generates questions, the most obvious of which is whether native speakers are aware of these two semantic systems. Francophones undoubtedly know that grammatical gender can mirror semantic gender and that masculine forms, unlike their feminine counterparts, are often vowel-final (for example, chien versus chienne). Whether native speakers know that “dark” is masculine, “aerial” is feminine or “edible/moist” is associated with consonant-final words, however, is less obvious. Some evidence of speakers’ awareness of such rules is found in words that have historically changed genders. Beckett suggests that grammarians may have changed the original feminine gender of un mélèze (“larch”) and un lierre (“ivy”) to conform to the notion that certain woody plants carry masculine gender. Another such question touches on the concept of saliency of traits. For instance, why is it more important that une autruche is “unique” than “flightless” in determining its gender? Beckett mentions this matter in passing , but the question is never fully resolved. While Beckett’s conclusions may be of interest to teachers and language learners, the exhaustive, dissertation-style analysis is more likely to appeal to those interested in a detailed semantic study on this topic. Nonetheless, the book’s conclusions do uncover potential rules constraining gender and word-final phonology in French, challenging the less than compelling results of previous models. These conclusions also indicate how nominal classification in French may be similar to typologically dissimilar languages, such as Australian Aboriginal languages. In this respect, Beckett’s book makes an important contribution to the study of the structure of the French language and to the field of linguistics in general. University of Texas, Arlington Christopher M. Stewart BEECHING, KATE, NIGEL ARMSTRONG, and FRANÇOISE GADET, eds. Sociolinguistic Variation in Contemporary French. Philadelphia: John Benjamins, 2009. ISBN 97890 -272-1865-0. Pp. xi + 257. $149. A North American reader approaching this volume would probably not 604 FRENCH REVIEW 85.3 need a great deal of convincing to concur with the position that a “Labovian” quantitative approach can still shed valuable light on variation and change in a language such as French, despite criticisms that some researchers have directed at the paradigm in recent years. That said, the defense Beeching articulates for the variationist perspective in the introduction to this collection of papers is a cogent one that effectively addresses concerns of critics who often tend to miss the point. At the same time, it provides those of us who do not need to be convinced with a reminder of the perennial value of corpus-based empirical studies. The editors offer us a collection of thirteen well-presented papers, whose heterogeneity is given a good deal of orderliness thanks not only to Beeching’s general introductory chapter, but also to the organization into three thematic sections, each preceeded by an introductory chapter that brings together unifying strands: phonological variation and leveling (seven papers); syntactic and stylistic variation (four papers...