THEORY: universalism in phonology

On Fri, 25 Feb 2000, And Rosta wrote:

> I'm not all that keen on the universalism that is even more dominant in
> phonology than it is in, say, syntax. I would prefer that theories of
> individual languages take priority over theories of language in general,
> but am a pretty lone voice. Anyway, it means that most good cutting-edge
> work in phonology (and syntax) is addressed to issues that to me are of
> secondary importance.
Recent work by Paul Boersma adresses this concern. His idea is
that phonological contrasts are of certain limited but universal
types, but their instantiation in particular languages is
language specific. (This may be close to tautological; bear with
me.) To take an example, we might claim that all languages
distinguish vowels along a dimension we might call Height. Some
languages distinguish only two degrees of Height, some three,
and some four. For each language, learners will establish the
relevant categories based on the available data; this
categorization can be made without recourse to prior primitives
such as distinctive features. If a language has only two degrees
of Height, we may well call them [High] and [Low], but that is
not to say that [High] and [Low] will be defined in precisely
this way for the language with three or four degrees of Height.
So while the contrast type is universal, its instantiation will
vary from language to language and is not dependent on
procrustean constructs such as Distinctive Features.

Similar arguments hold for any type of contrast available to
speakers of languages. You may want to find his "Elements of
Functional Phonology" at the Rutgers Optimality Archive and take
a look; there's lots of techy phonetic stuff that I generally
just skim through, but the overall picture is clear enough from
this work. An interesting feature of his theory is that he uses
an explicitly formal model (Optimality Theory) to make an even
more explicitly functional statement: phonology is the way it is
because people need to communicate efficiently.

Dirk

--
Dirk Elzinga
dirk.elzinga@m.cc.utah.edu