[Corpora-List] Corpus vs Intuition

[Gemma Boleda]

Dear Mai,

in my experience, it's fruitless (and very frustrating) to try and 
convince "armchair linguists" to change their methodology, they just 
don't see the point. I remember once I tried to convince a PhD student 
that the data-gathering methodology in linguistics is less than 
satisfactory, and argued that the community should start using 
Psycholinguistic methods to gather reliable linguistic data. He said 
that's not an adequate methodology, because the goal of linguistics is 
to model the internal grammar of a single speaker (an argument that 
somehow sounded familiar to me...). I asked him how one could falsify 
the hypotheses made, or even reproduce the results, and he looked kind 
of puzzled. But I didn't manage to move his position even half a 
milimeter. (Maybe I should have asked what happened when that single 
speaker died...)

So I think the way to go is to "predicate with the example" (is that an 
expression in English too?). That is, instead of convincing people that 
using corpora (or analysing linguistic phenomena with statistical 
methods, or gathering data with Psycholinguistic methods) is useful, do 
it yourself and try to publish in linguistic fori (which is sometimes 
difficult, because, you know, editors and reviewers are usually armchair 
linguists; but if you have good linguistic arguments it should be 
feasible; see, e.g., Sorace & Keller 2005). My hope --and I am a born 
optimistic, I must say-- is that little by little the message will 
spread and theoretical linguists will start using corpora, statistical 
analysis, or maybe even computational techniques to carry out research 
on theoretical linguistics. Thus quantitative linguistics, corpus 
linguistics, glottometrics, etc. will no longer be separate areas of 
research (with separate communities, journals, etc.), and we'll all be 
doing simply linguistics. But it's us who have to provide the proof of 
concept for the change to take place.

REF: Sorace, A. and Keller, F. 2005. Gradience in linguistic data . 
Lingua 115: 1497-1524.

Best,

Gemma Boleda
Universitat Politècnica de Catalunya


_______________________________________________
Corpora mailing list
Corpora at uib.no
http://mailman.uib.no/listinfo/corpora