Don't touch my phonemes (was: minimal pairs ex: PIE e/o Ablaut)

[Dr. John E. McLaughlin]

David L. White wrote:

> I hesitate (perhaps not enough) to get into this, but people DO know what the
> phonemes of their language are, they just don't know what the term means....

> I believe Sapir has a comment somewhere to the effect that if you want to
> know whether something is a phoneme in an Indian's language, ask him.  You
> just have to ask the right, i.e. indirect, question, not "Is this a
> phoneme?", but "Is this a possible different word?"  I think Sapir probably
> had enough experience that he should know.

You're probably referring to the famous paper "On the Psychological
Reality of Phonemes" (reprinted in David Mandelbaum's collection of
Sapir's most influential papers).  To briefly summarize and simplify
for clarity what Sapir experienced with Tony Tillohash (his Kaibab
Southern Paiute consultant), and simplifying the phonetics for the
ASCII character set, the citation form for 'water' in Colorado River
Numic (the single language of which Southern Paiute, Ute, and
Chemehuevi are dialects) is [pa], but when it is the second element of
a compound, it is [-va], as in 'big water' [piava].  When Sapir asked
Tony to say the word with compounded pa ([-va]) slowly, he
consistently said, [pi-a-pa] rather than [pi-a-va].  He then used this
to demonstrate how surface phonetics do not interfere with native
speakers' intuition about what the sound "really is".  My experience
in working with Numic languages over the past 25 years, however, has
shown that this is not quite so simple.  When untrained native
speakers have developed their own spelling systems for these seven
languages, they have not written the equivalent of pa and piapa, but
fairly consistently write the equivalent of pa and piava.  The
"linguistically accurate" orthography that Wick Miller and Beverly
Crum (a native speaker) developed for Gosiute and Western Shoshoni
over 30 years ago (which would spell pa and piapa) has not spread
within the Shoshoni community beyond the areas where Crum was
influential.  In contrast, more recent linguists, such as Chris
Loether and Drusilla Gould (a native speaker), working on Northern
Shoshoni; Pam Bunte and Rob Franklin, working on San Juan Southern
Paiute; Arnie Poldevaart, working on Nevada Northern Paiute; Kay
Fowler and Harold Abel (a native speaker), working on Nevada Northern
Paiute; Talmy Givon, working on Southern Ute; Maurice Zigmond, Chris
Booth, and Pam Munro, working on Kawaiisu; Chris Loether and Rosalie
Bethel (a native speaker), working on Western Mono; and Lucille
McClung and Albert Nahquaddy (both native speakers) along with Alice
Anderton, working on Comanche, have all used a more phonetically-based
writing system that native speakers seem to prefer much more than that
based on "native intuition".  While writing systems are not always the
best evidence of "native intuition", it's something we can actually
put our hands on rather than anecdotal evidence.  The Numic consonant
gradation processes run the gamut from language to language from being
totally frozen relics, to being highly productive modern processes,
but throughout the family, the evidence is clear that native speakers
prefer piava over piapa.  Probably the best indication of the
direction which native speakers are actually leaning is the Idaho
State University Shoshoni system (the one that I promote with Shoshoni
bands that haven't adopted an official writing system yet).  There
they use b instead of v and spell ba and biaba.  That's a lot closer
to the psychological reality which Sapir observed, but working from
the "inside out", so to speak.

John E. McLaughlin, Ph.D.
Associate Professor, Linguistics

Program Director
USU On-Line Linguistics
http://english.usu.edu/lingnet

English Department
3200 Old Main Hill
Utah State University
Logan, UT  84322-3200

(435) 797-2738 (voice)
(435) 797-3797 (FAX)