question about phonemes

[Suzette Haden Elgin]

April 3, 2003

My thanks to all on the list who have responded to my question about the
upper limit of (documented) phonemes in human languages. Thank you for the
help, and for the references. However, this is one of those cases in which
the cure has turned out to be worse than the disease.

After reading Spike Gildea's response -- which included the statement that
"most linguists no longer believe in the cognitive reality of the notion
'phoneme' " -- I withdraw my foolhardy remark about this question not being
a profound one, and I would appreciate a little clarification. I now have
two questions -- two question-clusters, actually.

(1) What does that statement of Spike's mean -- functionally? And how do we
define "cognitive reality," precisely? If phonemes do not function in human
speech to let speakers/listeners distinguish the meanings of words, how [I
am tempted to say "how the blazes"] does that happen? Why do we continue to
teach the concept of phoneme and to teach charts of phonemes if they're
only figments? And what should those of us who work outside the ivory tower
use as a way of helping people do useful tasks like teaching reading, if
not phonemes? I am accustomed every few years to learn that something we
linguists have written whole shelves of books about is now out of favor and
considered quaint, only to learn a few years later that the quaint little
whatever-it-is has come back around on the guitar again. It's
disconcerting, but is apparently the nature of the Linguist Beast. However,
I'd like clarification from within a functionalist framework. I'm not at
all sure that I understand this.

(2) It looks to me, from the responses, as if the following situation holds
(always remembering that the whole thing is unreal, anyway, right?):

Suppose we come across a language that has five identifiable vowels.
Suppose the language modifies vowels by nasalizing them, and there are
minimal pairs in which the meaning distinction depends on whether the vowel
is or isn't nasal. The question then is whether the phoneme inventory for
the language is to be analyzed as having five vowels or as having ten
vowels. If that is so, and if I haven't totally misunderstood your
messages, what are the criteria for making that decision? Functionally
speaking.....

Suzette