For me the question of interest is the complexity of descriptions of corpora. Is a complete grammar possible, or is the corpus of a language is the most compact description of itself?

If Chomsky's work is relevant to that, why not talk about it?

Personally I could do without the "he was right"/"he was wrong" stuff too. It buries the interesting issues and is meaningless in the abstract.

So let's keep it tight to the science/engineering issue.

But let's talk about it.

If the completeness of grammatical descriptions of corpora is not an appropriate topic for the Corpora list, what is?

-Rob

Hello Professor Wilks,

Thanks for that interview. Nice to hear a really pithy theoretical discussion.

I have a great respect for historical perspectives. They force you to see where other people have come from.

It is fascinating to hear the key conclusion of a man who was there at the beginning of the field, now after 50 years, is that we must go back to scratch and re-examine all our assumptions.

Among comments which stuck in my mind:

He abandoned his "depth hypothesis" because he was forced to conclude that "there are different ways of drawing tree structures."

That our models of syntax, semantics, pragmatics etc. all failed because we need to "get rid of our assumptions of grammar and syntax."

"Look at your assumptions and criticize them."

Personally I think we can clear up a lot of the mess, and get a very predictive model, by abandoning just one assumption. I believe much of machine learning to be quite sound for instance. We can use it (right from the level of sound waves.) We can even keep grammar, in a sense.

The assumption I believe we need to abandon is the one that there is only one grammar to be found.

Why do we insist on the assumption of global generalizations?

Instead we need to accept there are many grammatical perspectives possible, and many of them contradict. (By this view the grammars we learn now, "statistical language models" and the like, only appear random because they are many contradictory grammars summed together.) Instead of attempting global generalizations we should keep the corpus, and let context select the generalizations we need at the time we need them.

I think Chomsky saw this too. That is the "value" I see in his rejection of the phoneme. But he preferred to keep his assumption of one grammar, and abandon the idea grammar could be learned instead.

-Rob