Wednesday, April 08, 2009

When goals create blind spots

You're watching a ball game attentively. A person in a gorilla suit walks right through the middle, remaining visible for 5 seconds. Can you imagine not noticing the gorilla guy? Well, it turns out that nearly half of all people undergraduate volunteers don't, if they're busy trying to count passes - and the authors of that study cite 7 other experiments confirming the same principle.

It strikes me that there's a lesson there for linguists. Often linguists study a language for a specific theoretical goal - looking at Malagasy primarily to see what VOS syntax is like, or Oneida primarily to learn how polysynthesis works, or Songhay primarily to see whether it's related to Nilo-Saharan or not. That's fair enough; no one can focus on everything at once. But we can miss some really interesting stuff by focusing on one aspect of the language to the exclusion of others. For example, when Laoust studied Siwi, he was interested almost exclusively in its Berber origins - and as a result, his generally excellent study somehow ignored the vowels e and o (which are found even in Berber words, but are not phonemic in the Moroccan Berber varieties he was more familiar with), and mistakenly attributed the Arabic elements of Siwi to the adjacent Bedouin dialects, when in fact they show some very distinctive non-Bedouin characteristics. This is something we all need to watch out for.


Mattitiahu said...

I'd believe it. There's this one guy here whose primary research interest is the psycholinguistics involved in how people read whom I've had some interesting lectures/dialogues with.

The physical mechanics of it is interesting, like for instance how we tend to skip over function words for the most part (for proficient adult readers anyway), yet mentally reconstructing the original target phrase in our minds without actually physically looking at each and every written word. I really don't know much about it myself, but it seems analogous to the same process of how we mentally process normally spoken speech with all the regularly occurring reductions and elisions... but only through a visual medium rather than a sonic one. Heck, I'd like to read more on it sometime when I have more time. It's an incredibly fascinating subject.

As for second point, yeah, I agree. In-depth specialization can be a double-edged sword, especially if one gets too myopic.

Unknown said...

This has been my own experience when studying Cicipu.

I don't think the problem is helped by the nature of linguistics as a academic discipline. Coming from a background in IT where the idea of a single person designing, developing, testing, and releasing their own software fills people with horror, it's been hard to get to grips with the lack of accountability in linguistics.

How many other 'sciences' would tolerate an infrastructure that encourages 'lone wolf' scientists whose work is, in practice, untestable by their peers?

Lameen Souag الأمين سواق said...

Stuart: that comment would be well worth expanding on. The one linguist per language model does make our work less reliable; as they say, with enough eyes, all bugs are shallow. That's a good argument for publicly available glossed audio corpora, like yours - but let's face it, not many linguists are actually prepared to wade into a corpus in an unfamiliar language. Perhaps we need to change that ethos too - but how?

Certainly historical linguistics could be done much more productively with a more collaborative approach - no one person can possibly have detailed enough knowledge of every branch of, say, Nilo-Saharan to weed out all the mistakes - but that's another story.

Anonymous said...

An additional difficulty is when all linguists working on a given language family share a common bias: for example, in Romance studies there is an unfortunate tendency, when describing some dialect or other, to focus upon retentions or losses from the vantage point of either Latin or some well-attested intermediate ancestor (Old French, Old Provencal, for example), and conversely to downplay or outright ignore features alien to these prestigious ancestors. I strongly suspect the same is true in the case of language families such as Sinitic, Mongolic, Tibetan, Hellenic, where knowledge of a prestigious older written language typically leads scholars examining modern vernaculars to examine them solely through the lens of said older written language.

While I'm no Arabist, I strongly suspect the same is true in Arabic studies: Bedouin dialects of Arabic, for instance, seem to be considered of interest solely because of their having preserved certain Classical Arabic features.