Sunday, January 11, 2009

Pleasant Surprise at LSA

Being a cynic by nature, I am typically underwhelmed by linguistics conferences (I suspect that I'd be equally underwhelmed by other academic conferences too, but I'm a linguist). The Linguistics Society of America's huge annual meeting was held in my backyard this last week in San Fransisco so I attended a few sessions. Unfortunately, I largely had the same experience I typically have: squirming in the audience while smart, accomplished professionals drone on about their topic of choice. The problem is the format: 20 minute presentations with 10 minutes for Q&A. That's a tough set to play. Only academics and professional comedians are ever asked to perform under those kinds of conditions, and professional comedians get hundreds if not thousands of times more experience before they get good at it. Professional academics get maybe one or two chances a year to perfect the art of presentation.

Nonetheless, occasionally there is a person who has a talent for presenting complex information in a helpful and productive way, and I was lucky enough to see one presentation at the LSA by just such a linguist: Chris Golston of CSU, Fresno. He and co-author Tomas Riad presented an OT account of metrical phenomenon. By all rights, I should have been half asleep. I am neither a phonologist nor an OT adherent. But Chris, who was the primary presenter, was engaging, funny, and damned good and getting me to understand what the issues were and what their solution was. He is a natural teacher. His students at Frenso probably have no clue how lucky they are to have such a good teacher.

Their Presentation: Chris Golston (California State University, Fresno), Tomas Riad (Stockholm University): A constraint-based view of English meter.

Sunday, January 4, 2009

What Is a Word?

(picture definition of the "word" zazen)

The nature of a word's meaning has been an Achilles Heel, stumbling linguistics for hundreds of years. For example, Pāṇini's 4th Century Sanskrit grammar Astadhyayi (Aṣṭādhyāyī - अष्टाध्यायी) appears to have accepted that "the authority of the popular usage of words … must supersede the authority of the meaning dependent on derivation. The meanings of words (the relations between word and meaning) are also established by popular usage" (more here).

The Urban Dictionary is a great example of this kind of approach to dictionary making and now The Photographic Dictionary is trying to use pictures to define words (HT: Daily Dish). It's an interesting project, linguistically as well as artistically. I doubt these pics have been normed for their "meaning" (to be fair, it's more of an art project than linguistic research), but it's a good move towards functional definitions of words. I'd prefer to see multiple pictures (and videos??) for each word that have been normed to some extent for the meanings they are supposed to represent. For example, when I looked at the picture definition for the "word" zazen (above), a word I had never seen before, I did not feel that one picture helped me understand the meaning of the word. If anything, it confused me because I could imagine any number of conflicting meanings associated with that one pitcures. No one meaning was salient. This is classic function/structuralist linguistics. Cognitive semantics grew out of exactly this kind of problem.

And, for the record, my answer to the question in this post's title is this: I have no idea. See
Princeton's Construction Site for more on my confusion.

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...