- Describing and modeling variation in grammar
Morphosyntactic variation has been an object of detailed study for quite a while both in formal linguistics and in sociolinguistics and dialectology, but with different points of view. The book under review here is a rare attempt to bring research on morphosyntactic variation from various camps together. The editors are convinced that each of these camps will benefit from the others’ perspectives. I first discuss the notion of variation in grammar as well as its value for linguistic theory, and give a broad evaluation of how this book supports our understanding of variation. I then give an overview of the papers of the collection.
The understanding of variation that is used throughout this volume comes close to how allomorphy is understood in morphology. Two expressions are variants if they share the same meaning or grammatical function, but differ in more shallow aspects. One author (Bert Capelle) even introduces the term ‘allostruction’ for the kind of ‘intimate variation’ (a term cited by the editors in their introduction from Weinreich, Labov, & Herzog 1968:167) in (morpho)syntax that is under debate here. There are several ways of determining syntactic variants. Two structures can be seen as variants if they are alike, for instance, with respect to dependency structure, but have different constituent structures or linear orders. Most authors deal with variants that are not semantically different, at least not with respect to propositional content. A number of papers in the collection use construction grammar, a framework that has some affinity to such a conception of variation since it treats syntactic structures, roughly, as morphemes.
Formal syntax has, from the very beginning, developed many tools for describing intimate relations between variants in this sense. For instance, the relation between active and passive clauses was an early application of transformations in generative syntax. The phenomena dealt with in this book do not differ in principle from such a straightforward case of syntactic variation. One might wonder whether it is necessary to dedicate a book particularly to variation if it is such a central feature of languages and present in most syntactic work even when it is not stated explicitly. The editors’ agenda is a bit more ambitious, as they state in their introduction: ‘In any event, however, a theory of grammar that simply chooses to dismiss variation, for technical reasons or some notion of elegance, is of little help in describing and modeling linguistic reality’ (14). As noted above, theories of grammar in general are well equipped to deal with variation, contrary to what is stated here. I understand this passage as a plea for a more realistic idea of what a language is. It is, in fact, the sociolinguistic idea of language that takes the speech community much more seriously, including the distinction among standard, dialect, and sociolect, formal and informal communicative situations, and different registers, as well as language change. The formalist tradition has a psychological view of language and differs from the sociolinguistic view already with regard to whether the German or English language even exists. The common interest of both approaches, to my mind, lies in the dynamics of language and language change, because here the interaction of linguistic and extralinguistic factors becomes crucial. And here is the place for theoretically interesting research on linguistic variation. When two expressions A and B are (‘allo-’)variants, it will not contribute very much to my understanding of the linguistic phenomenon itself if I know that Ais preferably used by men and B by women. But it will certainly deepen my understanding of the phenomenon if I know how this came about. It is therefore a pity that only a few papers in the volume deal with linguistic change.
The papers are collected from two workshops on variation that were held in Bielefeld, Germany, and Berne, Switzerland. This explains the strong overrepresentation of contributions dealing with German and Germanic (fourteen out of sixteen papers, among which seven are on...