Sunday, September 30, 2007

Crystal Clear ...

Thanks to cko (who lent/loaned/leaned me her copy of Crystal's Language Death) I was finally able to read Chapter 2 Why Should We Care and Chapter 3 Why Do Languages Die. I have to say, I found the general writing style disappointing. It’s a lightweight volume that reads like it was pasted together from notes and speeches (which it may very well have been). He tends to make the same points over and over, in no systematic order.

I read nothing in the first three chapters of this book which caused me to re-evaluate my gut feeling that there may be some favorable outcomes to language death (and we linguists ought to study that possibility more closely).

Only 3 main points relate to why language death is bad:

I. Languages are like an ecosystem = ecosystems have mutually reinforcing relationships between members/elements (i.e., hurt one, hurt the system)

II. Languages are repositories of data (i.e., we can learn stuff from them: history, culture, linguistic feature space)

III. Language = identity

There’s no proof of (I) and Crystal is quick to caution against taking the analogy too far (he claims that humans are in complete control of language death factors; I suspect he is wrong about that, but ...); nonetheless, I suspect that it’s somewhat analogous. However, by the same ecosystem analogy, it may be the case that some language death may have favorable outcomes (which has been my guess all along). As I noted on cko’s blog recently, “I suspect that recent work in language learning and evolution by Partha Niyogi and folks like him will bear greatly on this topic in the coming decade.”

Argument (III) is garbled at best. Crystal claims that “Identity makes members of a community recognizably the same” (p39). Hmmmmmm. I thought it was the opposite -- identity makes members of a community recognizably different. In any case, this argument is vague at best, and does not relate directly to language death. There are various cultural factors that go in to “identity”, whatever that is.

It is argument (II) which I find most compelling, and the one I agree with most readily and without debate. Yes, I agree that all languages have unique linguistic properties that are well worth studying in themselves. But just because we find valuable data in every language does NOT mean we should stop language death per se. we need a broader understanding of the system of language interaction and language evolution, otherwise zealously stopping language death may be as irresponsible as zealously causing language death. Like a protected species over-grazing or over-hunting a locale, language over-population may serve some ecosystem harm. We just don't know.

No comments:

A linguist asks some questions about word vectors

I have at best a passing familiarity with word vectors, strictly from a 30,000 foot view. I've never directly used them outside a handfu...