Someone might object that this has gone a bit fast. No matter how plausible C2 may be as a category of skills, no matter that we can imagine language's being understood without appeal to rules, there is a strong reason in favor of placing language in C3 rather than C2. One of the anonymous reviewers raised the following objection.
11...[w]ork in the Chomskian tradition, following the `poverty of the stimulus' argument, has shown that the amount of negative feedback from the environment is not enough to account for language acquisition. Language is much too complex, and evidence of negative feedback is much too sparse to account for it.
According to this Chomskian objection, even if we suppose that
it's possible to imagine someone's acquiring language without
appealing to propositionally known rules, the fact of the matter is
that such an account does not accord with what actually happens.
Children acquire language much too quickly, and, worse yet, on the
basis of far too little evidence for an account such as mine to
capture the facts of language acquisition. Language belongs in C3
because it is only by attributing knowledge-that (innate
knowledge-that) of the rules of language to the child that we are able
to explain how acquisition actually occurs.
As compelling as it might seem, this objection misses the point
of my arguments. I am not concerned with the mechanics of learning,
but with the conceptual requirements for explaining someone's
possessing a skill. Goddard's account of learning to count shows
that, even if one agrees that an appeal to theoretical knowledge of
rules could account for the acquisition of some intentional and
purposive skill, attribution of knowledge-that of the rules of the
activity is not the only option available for acquiring the
skill.
The confusion in the objection traces in part to a mistaken
assumption that I am advancing a behaviorist account of language
acquisition. Goddard and I are not committed to such an account; nor
does my account depend upon any such commitment. It is open to me to
appeal to all manner of `innate' endowments or mechanisms to account
for our ability to `catch on' to counting or language.
That humans do acquire language in a relatively short period of
time, and on the basis of exposure to very little `good' data, is not
at issue. But the same is true of walking. Children `learn' to walk
in a very short time, with virtually no `input.' There is no
reason to demand knowledge-that of the physiology of walking, the
rules of maintaining balance, and so on in explaining this feat. My
argument is designed to show that there is no reason to do so in the
case of language. More significantly, my arguments show that
abandoning claims of knowledge-that of the rules of language does not
force us to abandon claims about language's standing as a rational,
purposive and intentional activity. For this, all that I am
required to establish is that it is not necessary that all of these
endowments consist in or rely upon knowledge-that of propositionally
available rules. Behavior which ``seems like rule-following'' is, as
Ryle tells us, far more varied than the intellectualist legend would
have us believe.