Re: THEORY: phonemes and Optimality Theory tutorial

Dirk,

I have trouble seeing how your exposition of Optimality Theory
is an argument against the phoneme. In your analysis, it seems
that there is this phoneme, '/B/', with a number of conditioned
allophones, and the OT constraints concern the conditioning of
the allophony.

When I silently concurred with John's "death of the phoneme"
statement (while strongly agreeing that like Newtonian physics it
is a useful & convenient approximation of the 'truth'), I had in
mind things like the following:
* Phonemics baselessly privileges the segment over subsegmental
  features and over larger elements, such as syllables.
* The structural/prosodic/syntagmatic properties of segments are
  not treated as defining features of phonemes, despite the fact
  that different structural/prosodic/syntagmatic positions support
  different sets of contrasts. In other words, if you wanted to
  insist on defining contrastive *segments*, there should be
  separate inventories of such contrastive segments for each
  separate structural/prosodic/syntagmatic position.

--And.