Linguistic universals and language variation (review)

This volume is a collection of thirteen papers by twenty-one scholars on topics broadly related to linguistic universals and language variation, plus an easily accessible synopsis by the editor Peter Siemund. Some of the papers were first presented at the workshop of the same title held in 2007 at the University of Hamburg's Research Centre on Multilingualism. The thirteen contributions are organized thematically into four sections: Part 1, 'Varieties and cross-linguistic variation'; Part 2, 'Contact-induced variation'; Part 3, 'Methodological issues of variation research'; and Part 4, 'Variation and linguistic theory'. As the sectional titles suggest, the book is heavy on language variation and light on linguistic universals.

Part 1 consists of four papers that investigate the variability of selected morphosyntactic features. HOLGER DIESSEL and KATJA HETTERLE ('Causal clauses: A cross-linguistic investigation of their structure, meaning, and use') examine the positioning of the causal clause in some sixty genetically unrelated languages, and find that it follows the main clause (exclusively) in 45% of the languages. The temporal and conditional clauses, by contrast, score a mere 1.7%. So the causal clause stands out among the adverbial clauses. MICHELE LOPORCARO ('Two euroversals in a global perspective: Auxiliation and alignment') studies the variation of perfective auxiliation and accusative alignment in Romance languages. Adducing both synchronic and diachronic data, Loporcaro shows that the selection of 'have' or 'be' to mark the perfective is not neatly in line with accusative marking. TANJA KUPISCH and ESTHER RINKE ('The diachronic development of article-possessor complementarity in the history of Italian and Portuguese') report that modern Italian and Portuguese, unlike English, require that the possessive pronouns be preceded by articles (Italian il mio libro vs. English *the my book). They show that the two languages exhibit variation in the diachronic development of the possessive noun phrase since the thirteenth century. While Old Italian is not much different from modern Italian, Old Portuguese allows 'bare' possessives, suggesting variable paces in the spread of the article-possessive structure. The contribution by SALI A. TAGLIAMONTE ('Variation as a window on universals') is an extensive and detailed study of default agreement in English (they/he was) in terms of the feature's geographic spread in Britain and Canada, as well as the morphosyntactic or communicative contexts in which it is used. The variation is enormous. For example, the incidence of use in existential contexts (there was ... ) varies from 30.2% in Toronto to as high as 95.7% in a small English town. For Tagliamonte, default agreement is a universal in the sense that it is widely attested in English dialects, despite the variability in usage pattern. Its status as a universal is an effect of the interaction of universal constraints.

Variation, of course, can be caused by language contact, which is the theme of the two papers in Part 2. HANS-OLAV ENGER ('Gender and contact: A natural morphology perspective on Scandinavian examples') studies gender marking in Norwegian. The modern Bergen dialect has a two-way system of marking grammatical gender, in contrast to the three-way marking of Old Norse and almost all other modern regional dialects of Norwegian. Enger attributes the reduction in gender marking to language contact in Bergen, which was a major commercial center in the Hanseatic League (ca. thirteenth century). The reduction of the Old Norse gender system is interpreted as evidence in favor of contact-induced simplification. No contact-specific evidence, however, is provided. YARON MATRAS ('Universals of structural borrowing') surveys the literature on structural borrowing and proposes various hierarchies that govern the likelihood of grammatical structures being borrowed in a contact situation: the higher a form is on the hierarchy, the more likely it is to be borrowed. So, derivational affixes outrank inflectional affixes, and the possessive construction outranks the attributive construction, and so on. For Matras, the hierarchies encode universals that govern...