what does 'independent' mean

[Elizabeth Bates]

It seems to me that the critical question here revolves around the
term 'independent'.  No one is denying that grammars exist.  But
how do they come about?  Once again, the vexed issue of innateness
rears its head.  Are grammars the way they are because they represent
the class of possible solutions to an extraordinarily rich communicative
problem, with some reliable and knowable properties?  Or are grammars
the way they are (at least now in our species, post-evolution) because
they are built into the mind/brain prior to birth?  The strong (and
in my view, more interesting) version of the functionalist approach
as a causal theory is the former, i.e. that the properties of grammar
*EMERGE* in response to a complex constraint-satisfaction problem
(mapping a high-dimensional meaning space with some universal properties,
determined by our social and cognitive make-up, onto a low-dimensional
channel which also has some universal properties, determined by constraints
on information processing -- memory, perception, articulation, etc.).
Now, one might argue that we have begun to build that solution into
the genome, so children don't have to start from scratch in every
generation.  That's certainly possible, and worth investigating -- but
it shouldn't be established by fiat, as an article of faith, which is
exactly how things are done in most child language work from a generativist
perspective. Most of the generative folks cannot envision what it would
mean for emergence to work as a developmental principle.  They don't trust
it, don't believe it, and hence they cite vague (invariably unspecified)
notions of unlearnability, poverty of the stimulus, etc.  This is precisely
where the neural network simulations of bits of language learning have
played an important role in the last few years: establishing by an
existence proof that certain phenomena that were previously believed to
be unlearnable are learnable after all.  These are toy simulations, who
knows if they will ever scale up to the level of a complex nervous system
in a social world, but they *DO* help the functionalist enterprise by
showing that emergence (as opposed to innateness, or its polar opposite,
learning-by-copying) is a viable approach to language development, and
that many of the conjectures (because that's all they ever were, there
are no serious proofs) of the nativist learnability program need not
concern us after all.

Brian MacWhinney has a very interesting edited book coming out shortly
(as in, in the next few days, I think) on "Emergence" in language
development.  Funknetters might find the volume useful in the framework
of the general formalist/functionalist debate.  -liz bates