Tuesday, December 22, 2009

How Many Linguists Are There?

The Independent recently published an article about the language documentation efforts of Mark Turin and his colleagues at The World Oral Literature Project. In the article, Turin was lamenting the large number of undocumented languages (a fair lament) and was quoted as saying this:

There are more linguists in universities around the world than there are spoken languages – but most of them aren't working on this issue. To me it's amazing that in this day and age, we still have an entirely incomplete image of the world's linguistic diversity. People do PhDs on the apostrophe in French, yet we still don't know how many languages are spoken.

I found this passage remarkably agitating. I appreciate Turin's passion for language documentation and I support language documentation efforts, but there are two claims in this passage (one explicit, one implicit) that I object to:

First, I'm not sure there really are more linguists than languages. Linguistics is a small field (this fact is relevant to both of my objections). The article uses the fairly common number of 6500 languages. This is a guesstimation at best. We don't have a good definition of a language (vs. a dialect), so it's not clear what counts. This is a non-trivial point. Figuring out what exactly language is, is the core of linguistics (imho). I take the problem seriously. The answer will likely be disappointing to non-linguists. The answer will likely be something like: there no such thing as a language as traditional conceived. In any case, I'm fine with using the 6500 number publicly because people like numbers. They want a number? Okay, we'll give them 6500. But within the field, there is no number. Next, I'm not sure what a linguists is. This is also non-trivial. We could say anyone with a PhD in linguistics from an accredited institution is a linguist. But even that definition requires refinement. Do cognitive scientists count? Professors of French? We could define this as anyone with the skill set required to go into the field and document a language. Wow, that would actually be a highly restricted set. No computational linguists. No psycholinguists. Not most syntacticians. Not most phonologists. Even if we were fairly generous in our definition, I think we'd be hard pressed to count up 7000 linguists.

Second, I object to the notion, implicit in Turin's quote, that language documentation is so critical a goal of linguistics that most linguists should devote their careers to it. Again, I'm pro-documentation, but there are lots of important tasks to be completed within linguistics. I believe that understanding how language works in the human brain is the absolute center of linguistics. All efforts follow from that. Language is first and foremost a cognitive product of individual human brains. Yes, there are very interesting sociolinguistic processes that are well worth studying; important cultural interactions that language takes part in to be sure. But understanding how the individual human brain produces and comprehends language is the key to understanding those sociocultural process.

Look, this was a bit of hyperbole on Turin's part and he doesn't deserves to be beaten about it, but it just got under my skin.  It all needs to be studied. I get that. I want to quadruple the number of linguists in the world and set an army of linguists into every town and hamlet, every village and urban center, documenting and analyzing every linguistic feature they can get their greedy hands on. I also want an army of theoretical linguists, psycholinguists, computational linguists, neurolinguists, and a host of other *linguists. But there must be method to the madness.  There must be something that coordinates those efforts around a shared goal. I see that shared goal as understanding how language works in the brain.


KEN said...

If we don't document a language that dies out and it has some feature that didn't exist in any other language, then we will have an incomplete picture of how language works in the brain. Just a thought.

Chris said...

Ken, it's a fair point. We certainly want to document a wide range of features, and I support documentation efforts.

But it's like chasing infinity: we'll never document them all because so many languages have come and gone that have zero documentation and always will (thousands? tens of thousands?), there will always be a gap in our documentation.

NLPers: How would you characterize your linguistics background?

That was the poll question my hero Professor Emily Bender posed on Twitter March 30th. 573 tweets later, a truly epic thread had been cre...