Peter,
Fair question. Personally, I have found Chomsky's ideas
sometimes right on and sometimes wildly implausible. I think that,
in the final analysis, the field would have benefitted more if the
ideas had been presented as just that -- as ideas, rather than as
established scientific fact. Transformations, the cycle, rule-
ordering, feeding, bleeding, the grammar gene, the sudden evolution
of language, speech is special, the centrality of recursion,
minimalism, mapping interfaces, principles and parameters, poverty of
stimulus, modularity, early full competence, and the like. These are
all fascinating ideas. But so are competition, cue validity,
constructions, entrenchment, transfer, thematic structure, dependency
relations, coevolution, sensorimotor bases, gesture-speech linkages,
linguistic relativity, thinking for speaking, dynamic systems,
learning to learn, resonance, and neural plasticity. What troubles
me is that the first set of ideas was presented as a single take-it-
or-leave-it package and contrasted with the second package, which was
often dismissed as something that had "already been shown to be
implausible."
My personal assessment of the developments of the last 50 years
is that, the single-package approach largely outlived its usefulness
by about 1980. By then, it was clear that the pieces of the overall
package were not logically or empirically co-dependent. If we had
been able to "get over" this barrier, I think we would now have made
quicker progress on (1) linking L1 and L2 theory, (2) grounding
acquisitional theory on data, including video recordings, of real
interactions, (3) made more headway on understanding real-time
processing of language by normal children and children with
disabilities, (4) developed a more solid quantitative methodological
base for ongoing work, and (5) been able to field full working
simulations of the acquisition of language. In this regard, science
is really often just a trade-off of time and manpower. If we, as a
field, had devoted less energy to attempts to characterize systems
for parameter setting, I think we would have made more progress on
these five fundamental fronts, building a firmer basis for ongoing work.
In my mind, Chomsky's most important contribution is his initial
one -- the emphasis on generativity. If our models of language
learning are good at all, they should be able to model and depict the
actual course of language development. In 1980, it was basically
impossible to do this, because we did not have the relevant corpora.
My own focus has been on solving this problem through the
construction of CHILDES. Now, I believe, we are in a position to get
back to the central task. We should be able to show how, using data
available in real interactions with real children with real video and
audio, children succeed in learning language.
I don't think that the issue here is about Chomsky's presence or
absence. Some of his questions are good and cannot be forgotten.
But it is time to move on.
-- Brian MacWhinney
On Oct 12, 2007, at 2:47 AM, Gordon, Peter wrote:
> I think it would be an interesting exercise if Robin (and other
> Anti-Chomskians) could give us a sense of how child language might
> have developed without Chomsky and why it would have benefited from
> his absence.
>
> Peter Gordon
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/info-childes/attachments/20071012/64703217/attachment.htm>