Chomsky's reply!

From:chomsky@MIT.EDU
Subject:Re: lsa242: problems and prospect of minimalist program
Date: August 8, 2005 10:23:33 PM CDT
To:hsiangyun@gmail.com

Deluged with mail, including course comments, so have to be brief, below.

Noam

At 11:36 AM 8/8/2005, you wrote:
Dear Prof. Chomsky,

I take the course for credit and here is my question/comment.

It seems to me that the overall picture of the minimalist program is really a "philosophical theory" of what human language is and how it works; it has some theoretical merits and some linguistic evidence, but it is to be explored and verified by further scientific findings.

The minimalist program is standard biology. It has always been understood that apart from genetic endowment and external environment, growth and development depend on general principles of structural architecture, laws of form, and far-reaching constraints that hold far beyond the organism or organ under investigation. These questions have been at the horizon of inquiry for the most part, not because there is any question about the significance of such factors, but because they reach too far beyond current understanding, though there have always been exceptions -- Alan Turing is perhaps the most famous in the modern period -- and in recent decades these have become major topics in biology. The Minimalist Program has always been on the horizon for the same reasons, but also became feasible in more recent years. This is perfectly normal science. There is nothing philosophical about it, any more than investigation of the various factors that determine other aspects of the nature of organisms. Since it's a program, not a theory, it cannot be verified or disconfirmed, as a matter of logic, though one can certainly ask whether the program is premature.


This "philosophical" flavor of this 'program' reminds me of the "innateness hypothesis." The innateness hypothesis has renewed the debate between empiricism and rationalism, and if it is correct, rationalism is to be favored. I'm wondering how far this preference can go and how deep it is.

The debate about the "innateness hypothesis" has been quite a curious one. There is a substantial literature of criticism of the hypothesis, but no one defends it. The reason is that there is no such hypothesis, and cannot be. There is, of course, a truism: it is not controversial that my granddaughter has some property that enables her to develop language in the normal way, but not her pet kitten (chimp, songbird, bee, etc.), sharing exactly the same environment. And she cannot attain their capacities, cognitive and other. Apart from mysticism, that yields the conclusion that they have different innate properties. That's no more controversial than the fact that animals develop different visual systems even if they share the same environment. These are biological truisms, considered controversial only in the study of higher mental faculties, an interesting illustration of the irrationality that has seriously hampered these pursuits. Why this should be so is an interesting question, but the fact seems pretty clear.

The only serious questions about innate endowment have to do with what it is, not whether it exists. The same in every other domain of biology.



The question is how we should understand what "innate" mean here, or exactly "what" is innate in the proposed hypothesis. One possibility is to interpret the innateness as 'written in our gene' and it's a form of biological growth. If so, I'm a little doubtful if this may be called 'innate knowledge.'

No one uses the term "innate knowledge" except informally. In fact, even the concept of mature knowledge (say, of English) is in part an idiosyncrasy of English usage. The technical term "competence" was introduced 40 years ago to try to bypass confusions about these issues, based on misinterpretation of the significance of informal English usage, and misleading philosophical theories of knowledge. Innateness means genetic endowment.


Regarding the notion of biological growth, we know that infants cannot eat solid food and they can't even drink certain liquid food until certain months. This is so because their digest system and stomach are not ready yet; the biological growth is not ripe. If what's special about human language is just like our stomach's ability to digest food, then what is innate is at most a kind of 'ability,' rather than 'knowledge;' do we want to say that human have 'innate knowledge' of how to digest?

As noted, the term "innate" (or "mature") knowledge is informal usage, and should not be misunderstood. However, it is clearly false that mature competence ("knowledge of language") is a kind of ability. There is ample literature on that, going back many decades.


Perhaps we can apply the distinction between 'knowing how' and 'knowing what' and clarify that the innateness hypothesis does claim that human language is innate, but it's an innate 'knowing how.' The task of the linguists, then, is to spell out the 'knowing-how' in terms of 'knowing-what.' But it may be a question whether all 'knowing-how' and be translated into some 'knowing-what.'

Again, there is ample literature on this going back many years, showing, in particular, that even the concept "knowing how" involves a cognitive element and cannot be reduced to clusters of abilities, propensities, etc. These are just philosophical errors, which have to be unravelled.


Another way to think of the innateness hypothesis is really to think of it as knowledge and 'knowledge of what', and this knowledge is written in the gene. Then I wonder how it can correctly be labeled as knowledge, for knowledge must have justification.

That is an invention of modern philosophy, which has devised a technical notion of knowledge divorced from common usage. There's nothing wrong with devising technical terms, but it makes no sense to insist that the invented technical usage replace ordinary usage, or different technical usages.


Depending on one's choice of what counts as justification, whether our knowledge of universal grammar is 'innate knowledge' can have very different answers. I think if we adopt the internalist notion of justification, 'innate knowledge' of language can't really be knowledge, simply because we really don't know, and are not aware of what justifies it, even though he mechanism that produces it is innate. However, if we take the reliablist approach, as long as the relevant production mechanism is reliable, its product counts as knowledge, then innate knowledge of language makes sense. And indeed, what's written in our gene should be pretty reliable, since it's the result of evolution. In this sense, whether linguistic knowledge is innate knowledge turns on which theory of justification we endorse.

Again, there is ample literature decades back exploring these issues, and showing, I believe, that the philosophical inventions are extremely misleading and should be abandoned.


Going back to how much weight the innateness hypothesis can put on the empiricism/rationalism debate, we've seen that there are many factors to be considered. We need a clear definition of innateness, a notion of knowledge and justification. To say that the success of 'innateness hypothesis' definitely favors rationalism just oversimplifies the issue.


Since there is no "innateness hypothesis" (apart from truisms), one cannot say that it favors rationalism. That would be true even if there were some clear notion of rationalism. There isn't. That's why in discussing these issues over 30 years ago I suggested that we introduce technical terms (like R and E) to try to avoid the huge mass of philosophical confusion that has developed over these issues.