Language structures are unique but comparative grammar is nevertheless useful

That it is useful to compare language structures may sound trivial, but it has sometimes been regarded as problematic or dangerous, because in describing one language, we may be unduly influenced by what we know from some other language(s). In the worst case, we overlook our prejudices and impose a procrustean bed onto another language. Especially the structuralist movement emphasized the uniqueness of each language and made a point of describing each language in its own terms, with its own categories.

In this talk, I begin with the observation that the uniqueness perspective is actually well-motivated, and I note that a rigorous approach to language structures can hardly get around it: Describing language structures with general terms (rather than with language-particular terms) may lead to strange or ethnocentric analyses. But on the other hand, it is also clear that languages exhibit many similarities that are worth highlighting, and many linguists find comparative and universalist frameworks satisfying even when they focus on a single language.

Perhaps surprisingly, I will propose that the relation between language-particular analyses and the study of cross-linguistic contrasts is best seen not as mutual dependency (let alone as unilateral dependency of analytic proposals on comparative work), but as relative independence: Unique language structures can be analyzed without much concern with universals, and the study of universals does not depend on the correctness of specific analyses. I will provide examples from a range of (mostly European) languages, from a range of different domains of grammar.