February 11, 2004

Word counts without lexical facts

I doubt that I'm alone in finding Mark Liberman's ruminations on camel spit genuinely fascinating (vastly more so than I would have predicted had someone asked me yesterday to say whether I wanted to hear something about this topic; one really must try not to prejudge).

And I hope that no one will have missed the key difference between what he does and what I was grumbling about in the previous post. It's the difference between the qualitative and the quantitative. Mark cites specific qualitative facts about the meanings and etymologies of particular Somali words, and speculates on what they mean for the view of the world you get through Somali lexicon and metaphoric imagery. And he has studied this language for a semester, and he has a dictionary of it, and unlike some people, he has learned to use it. What I grouse about is people who reduce the wonder to bald quantitative assertions concerning ethnic groups they know nothing about (tribes with 50 words for this or 92 words for that), having no actual quantitative data to back it up, and having not even asked if there are any such data. Comparative lexical census-taking without actually counting; statistics without the numbers. That's what gets my goat about the people who prattle on about the "abundance of words" this or that tribe has for shoes or ships or sealing wax or camel spit or kings.

Posted by Geoffrey K. Pullum at February 11, 2004 01:30 PM