>> I have important
>> disagreements with all of the schools of linguistics
>> that I know of, even the ones I'm promoting here,
>> but I wouldn't be a good linguist if I didn't. :^)
> As a linguist, why do you suppose the disputes in that
> field seem to be so hostile? I'm just an observer of
> the academic scene, but it seems to me that the
> linguistics field's infighting is worthy of MTV's
> "Celebrity Deathmatch". Is it because it's not
> enough of a "hard" science yet to allow for civilized
> differences, or are there political dimensions to
> these disputes that I'm unaware of?
Yes, there are political dimensions to these disputes. Linguistics is, in a sense, a branch of psychology, and how we think about people's minds is fraught with political repercussions. But I don't think that's the most important source of dispute. Linguists argue over which type of analysis is best without very much knowledge of how language actually works in the brain. Since the early days of structuralism, it has become possible to make theories that, if you fiddle with them enough, will explain all the data. Unfortunately, many different kinds of theories, with fundamentally different kinds of analyses, can all be made to fit the data more or less, and it's very hard to justify any of them on empirical grounds.
The biggest single problem that I have with generative grammar is that it invokes one kind of analysis without ever explaining why that one analysis is better than any other. We can write a generative grammar for any language, if we fiddle with it enough and define language in a very limited way. I just don't see any good reason to do so unless we have some evidence that the brain actually works that way, which strikes me as unlikely. This complaint basically holds true for other grammatical theories too, although most of them do try to make some sort of argument for why their analysis is better than the others.
There is a culture in linguistics that refuses experimentation and makes a lot of analysis based on very small amounts of data. I remember one of my profs justifying his theory of language on the grounds of "impecable logic," but he had no data that supported his theory over any other. He had never made any attempt to falsify his theory, he just criticised the logic of the authors of other theories. And this is one of the people whose theories I mostly agree with!
This tendency is as bad or worse everywhere else in the linguistics world too, as far as I can tell. When I studied government and binding theory, the prof never once, in all the time I was there, introduced actual data to the class. We were presented with G&B as a fait accompli which needed no justification to a mere group of undergrads.
I don't think the problem is that linguistics isn't "hard" enough, it that it isn't done in a very scientific manner at all. I came into linguistics after washing out of graduate school in high energy physics, so I knew what the scientific method was and I knew that people weren't using in linguistics.
They could use it, but it takes more than just the scientific method to make that happen, it takes a culture of science. People need to know that the very first thing that will happen when they publish a paper is that someone will criticise it on the grounds that it can't be falsified, and revise what they write accordingly. They need to understand that results have to be repeated, and that studies have to have controls in order to be meaningful, and that they have to report data that is contrary to their hypotheses and try to analyse the amount of error in their work. In physics, this is something you get enculturated to - it's part of a sort of scientific boot camp training - but linguistics programmes generally don't teach these kinds of things. No matter how much math you have or how much formalism you use, without these falsification you just aren't doing science.
A physicist with a theory has to show that some kind of experiment could falsify it. A physicist must, therefore, consider the possibility that he or she is wrong. When you realise that your ideas may be completely discredited, you tend to treat your opponents with much more respect. In linguistics, no one ever has to concede that they were wrong, because they don't really try to falsify their hypotheses and they can fiddle them endlessly in order to evade falsification, and so they have no reason to fear their opponents.
This is especially true because a lot of linguistics is done in journals with a particular slant. There is relatively little impartial peer review. People working in G&B publish in G&B oriented journals, reviewed by G&B people and are only read by G&B people. Other schools of linguistics are unfortunately similar in behaviour.
Until there is a scientific culture in linguistics - something that is growing slowly, but is growing - all that will distinguish one theory from another will be the largely subjective judgement of how "intuitive" and "neat" it is, and as long as results are judged on such a subjective basis, the disputes will never get quieter.
I have some hope that cognitive scientists, who are enculturated into scientific culture and do publish in common journals rather than setting up a journal for each sect, will help change some of this in linguistics.
Scott Martens
---
Envoyé par Moncourrier.com Vous aussi obtenez gratuitement votre adresse électronique sur Moncourrier.com