Print

Print


STAN MULAIK wrote:
[quotation cut]
 
>
> There are rumblings at sci.lang that Chomsky may be changing his positions
> in a radical way at the present time.
>
 
Well, that wouldn't surprise me at all - he does that regularly.  It's
Chomskyans, not Chomsky, who tend to get stuck in rigid theoretical models.
 
>
> I think perhaps the thing to object about with Chomsky is not so much
> the hard-wired aspect of his theory, but the particular conception he
> had of how the connections functioned.  I too am familiar with and
> sympathetic with categorization and metaphor theory (via George Lakoff's
> "Women, Fire and Dangerous Things" and his writings with Mark Johnson).
> But they in turn ground their metaphor theory in embodied schemas of
> perception and action in the world, which brings in their own hard-wired
> side of the story.  Both are modern variants of Kantian a priori category
> theory. But metaphor theory shows there is much greater flexibility in
> how the embodied perceptual schemas are extended in conceptualization.
> Perhaps they would also concede to cognitive psychologists that there is
> also some freedom in cognitive development for how these schemas will
> form.
 
Quite.  I believe some aspects of language are indeed hard-wired, and have
argued for this point of view on the cogling list.  Your characterisation of
this view as Kantian is very apt.  What people (inevitably) disagree on is the
degree and type of this hardwiring, early versions of Universal Grammar being at
one extreme of the spectrum, the more radical cognitivists at the other, and
people like Jackendoff and Langacker somewhere in the middle, perhaps.  On the
innateness question, what I would propose is the following:
 
1. All animals store categories in the brain (even if it is only
"edible/inedible").  Some of these categories may be hard-wired.
2. Some animals can associate categories with vocal/visual signs e.g. alarm
calls, aggression displays.
3.  A few species can group such signs into strings (as in captive dolphins
memorising sequences of instructions).
4.  A few species can organise categories hierarchically (e.g. Sarah the chimp's
ability to classify "banana" and "apple" as "fruit").
5.  As far as we know, only humans can apply hierarchical categorisation to
strings so as to create syntax (X-bar theory).
 
Getting back to the IAL question, I think the important thing in IAL design is
not the hard-wired constraints on language, but the functional ones.  IMO the
hard-wired constraints are so much a part of our thinking that it would take a
real effort of imagination to construct a language that didn't conform to them.
Most of the choices that can be made in language design involve functional
considerations - e.g. whether the advantages of pro-drop or SOV order outweigh
the disadvantages.
 
Robin Turner