[Avery Andrews 930915.1040]
(LHL@fs.law.wlu.edu, 14 Sep 93 08:49:55 EDT)
>One can have an empirical theory of an unobservable; physics is full
>of them. The reason for choosing I-language over E-language is that
>I-language is the biological uniformity; ...
Not quite. You're identifying `I-language' with the `initial state
of the language faculty', which is a biological uniformity (to some
degree, at any rate), but what linguists investigate are derived states,
which are prima facie arenot. Chomsky has recently speculated that
`grammar' is fixed and universal, with variation confined to the
lexicon, but then we need to explain how the lexicon is acquired, so
Plato's problem, etc., is simply popping up somewhere else.
The derived states of the language faculty are not uniform, but are
shaped by interaction with the environment (regardless of where the
variation is located); these are still `I-language' (and subject to
variation). But to get any actual information about these states,
you have to examine what people say, and how they react to what
people say, and this requires assuming quite a lot about performance,
just like you need to assume things about friction to investigate
Newton's laws in an actual laboratory.
You don't need a good theory of performance (just like you don't need
a real theory of friction - coefficients of sliding and static friction
will do), but you do need some assumptions, and of course you'd
naturally be happier if you had a real theory to back up the
assumptions.
So something sort of like `E-language' (glossed E-language), is an
essential methodologial concept, even though it is probably not a
good object of study.
I think Chomsky may also be momentarily ignoring the fact that certain
`near-subsets' of E-language (e.g. the stuff that people actually say)
has a rather pivotal role, being for example, the transmission channel
for the variable aspect of I-language. There is also quite a lot of
evidence that there are control systems which tend to cause people's
individual grammars to converge in their extensional effects
(`accomodation' in sociolinguistics - I can observe this when we visit
my wife's family and her speech becomes more Iowegian). So in spite of
the alleged incoherence of the concept of `language' as conventionally
understood (dialect with an international airline), there may be some
real concept out there after all, along the lines of `a group of
speakers who control (as individuals, in this case) for a perception
of being able to talk to each other.'
Avery.Andrews@anu.edu.au