>> You did not respond. If I have misconstrued the argument, please say how. If
>> not, refutation is hardly required.
>
> Okay. Children are aware of what sentences are grammatical and agrammatical
> from a very early age, as young as 3 in some cases. Yet, roughly 95% of the
> sentences they hear are agrammatical or incomplete. The majority of the
> utterances they are exposed to will remain incomplete from a grammatical point
> of view all their lives. Yet, they know what kinds of sentences are right and
> which are wrong. [Most of this is cribbed from what I remember of Pinker's
> PhD thesis.]
>
If I may paraphrase: look how fast children learn language.
> On the basis of this data, it is difficult to see how a purely behaviourist,
> data-driven account of language is adequate to explain language acquistion.
>
That it may be difficult proves nothing. Paraphrasing again: therefore
behaviorism is false.
> Comparitive studies of children acquiring a first lanaguge and adults and
> children acquiring a second one point out that the type of errors made by
> adults trying to acquire a language in a rule-driven fashion are very
> different from the types of errors acquired by children learning a first
> language or young children acquiring a second. [This is what I remember from
> Guy Lefrancois's work in modeling language acquisition in the 70's.]
>
This is irrelevant. It may also be a badly taken point, in that the training
involved in the two cases is different.
> Furthermore, the sharp difference in language acquisition ability before and
> after puberty (as little as 6 months makes a huge difference in ability) is
> strong evidence for a biological mechanism in language acquisition.
> Regardless of learning environment, reinforcement or stimulus, all but a small
> minority of adults will never master a second language.
>
I believe that given the appropriate learning environment and reinforcement
virtually all adults can aquire a second language. Berlitz agrees.
> The magnitude of
> difference between the language acquisition ability of children and adults is
> unparalleled by differences in the ability to acquire other skills. [This
> undermines Geoffrey Sampson's argument that children acquire all skills more
> quickly than adults.]
>
This magnitude documented where? What other skills are we talking about?
Walking? Grasping? Visual tracking? What units are abilities to aquire
skills measured in? I note here, and below, the reference to age barriers to
language aquisition. These are generally hard to study and such claims rely
on a very few cases. I would caution against broad generalizations based on
such scantily studied phenomena.
> Now, Chomsky requires a universal grammar in order to explain this. I suspect
> a more universal semantic system is far more likely, and some universal
> limitations on how meaning is restructured into linear utterances.
>
> However, children acquire language because they are programmed to do so. The
> consistency in language acquistion among children, regardless of their
> environment; the rigidity of age barriers to language acquistion; the
> acquistion of language even by children who can neither hear nor speak; the
> existence of language among **all** humans raised in contact with as little as
> one other person, even when that other person is also an infant with no
> language; the inability of children raised in isolation from all contact (like
> that poor girl in Ohio who was raised in a closet) to learn language at the
> same level later in life - all these things point straight to biology as the
> motive for language acquistion.
>
Again paraphrasing: therefore language is innate.
This is more or less the argument as I originally outlined it (though I omitted the hand-waving).
Let me again recommend "Rethinking Innateness".
-- bill