My old buddy the archiphoneme

[Dennis R. Preston]

Definitely not an archiphoneme (though it warms my heart to see the
word early on Saturday morning). Not all neutralization is
archiphonemic. In this case the merger of /O/ and /a/ (as in most
western US dialects, Eastern New England and a growing band across
the Midland) does not result in an archiphoneme; it simply results in
phoneme loss. I have one more phoneme than such speakers (and clearly
a great deal of moral superiority by possessing it). An archiphoneme
arises from neutralization of phonemic oppositions within certain
contexts and has to do with the economy of representations. For
example, English has the word "spin" /spIn/ but does not (and CANNOT)
have the word "sbin' /sbIn/. Therefore, the voiced-voiceless
opposition for /p/ and /b/ is lost in clusters with /s/, and the
phonemic (not phonetic) representation /spIn/ is overspecific since
it includes the information -voice on the /p/ segment. The
archiphonemic proposal (Trubetskoy's by the way) was to supply a
"larger" category which did not specify the predictable voicing
characteristic (often done with a cap, e.g., /sPin/).

So much modern phonology stems from this that I dare not write on.
Glad to see "archiphoneme" at dawn. ( .. the snail's on the thorn ...
all's right with the world ... whatever).

dInIs

>  > On 9/20/2002 21:16, Mark A Mandel wrote the following:
>>
>>  >Let's try an example, made up a-purpose:
>>  >
>>  >         sought. vb: past tense and perfect participle of "seek".
>>  >         /s O t/
>>  >
>>  >where the phonemic symbol /O/ is defined as the vowel of "caught",
>>  >*whatever that is phonetically for you*. If you merge it with /a/ as
>>  >[a], you'll say [sat]. If you merge it with /a/ as [O], or don't merge
>>  >it, you'll say [sOt]. In any case, your pronunciation will be the
>>  >correct one for whatever dialect you speak.
>>
>>  Wouldn't that just be an archiphonemic representation?
>>
>>  Cheers,
>>  Scott
>>
>
>
>How can anyone reading all this purport to wonder why other people find
>linguists abnormal?
>   ;-)   (Yes, I am winking) -- Millie

--
Dennis R. Preston
Professor of Linguistics
Department of Linguistics and Languages
740 Wells Hall A
Michigan State University
East Lansing, MI 48824-1027 USA
Office - (517) 353-0740
Fax - (517) 432-2736