Ladd's posting on the 'limits of scientific investigation' seems right
on to me. The problem here is not unique to the IPA and phonetics
research, but is found in morphosyntactic theory and typology as well.
Discussions similar to the 'dark-L' issue take place in syntactic
typology all the time. For example, "is this construction in language x
a 'passive' or not?" It turns out that constructions across languages
share some properties but not others. This is why I believe efforts to
come up with a universally applicable syntactic vocabulary have to be
evaluated very cautiously. In fact, in recent discussions on Linguistic
Typology, I used the IPA as an example. IPA symbols are useful tools
for representing a general area of articulation, for example, but must
always be supplemented in any grammar of a language or narrower study
by copious annotations and measurements (and measurements are equally
applicable in syntax in terms of statistical studies of relative
acceptability, say, of a given construction) which say something like
'the symbol [L] means "x" in this language.'
Unfortunately, the potential usefulness of the IPA is too often
overriden as linguists just use an IPA symbol, a standard definition of
the sound, and let it go at that (this is extremely common in most
descriptive grammars, for example). This is a mistake. Unfortunately it
permeates most linguistics research.
Each grammar, each study is to some degree relative. It is unlikely to
be the case that any component of any grammar is exactly like another
component of another grammar (or, as Ladd mentions, perhaps even
exactly like any other utterance by the same speaker).
William James recognized this more than a century ago. In his
pioneering work in psychology, James argued that "no two sensations,
however, much they resemble each other, can be correctly judged as the
same" (paraphrased in Myers's (1986) biography of James.
Much of what we have to do empirically is to show how much languages
differ from one another. I suspect that linguists of the future will be
bemused by our current all too common desire to make languages look
like one another, from their phonetics to their morphosyntax, etc.
Understanding the nature of similarities is important, of course, but
it is arguably equally important to understand differences and
particulars.
Phoneticians in my experience tend to be much less guilty of this than
syntacticians or phonologists, but the problems is there nonetheless.
Dan
------------------------------------------
Daniel L. Everett
Department of Linguistics
The University of Manchester
Oxford Road
Manchester, UK M13 9PL
http://ling.man.ac.uk/info/staff/de
Fax: 44-161-275-3187
Office: 44-161-275-3158
|