Re: Advice required: phonologic or phonetic?

Francois:
> I asked something similar some months ago, but did not get any
> satisfactory feedback.
> How would you call a (part of) consonant (e.g. /k/, phonemic) which is
> still considered the same consonant when it becomes /g/ (allophone) and
> the /k/--/g/ opposition has grammatical meaning? Knowing, of course, that
> the native script uses the same symbol for both?
> (This last point is very important: due to intensive use of infixes, the
> very fact that [k]aned [g] are considered the "same" consonant is the only
> way to distinguish, e.g., /LOT/ from /LOK/, while phonetically they may
> become [luG] or [lowD] respectively)
> It is just a matter of technical vocabulary and notation, I think. But, I
> would like to avoid any misunderstanding.
It's much more than a matter of technical vocabulary and notation.
/k/ and /g/ are different *phonemes*, according to your description.
If [lok] and [log] contrast, even if the contrast is a grammatical
one, then [g] is not an allophone of /k/.

Here are two alternative possible analytical strategies:

A. /k/ and /g/ are different phonemes, but lexical forms (stems?)
are composed not of phonemes but of archiphonemes, such as 'L',
'O', 'K'. (Morpho)phonological rules specify which phoneme
realizes a given archiphoneme in a given environment. The native
script's symbols represent archiphonemes.

B. Phonemes are not important; morphemes are composed instead
of (strings of) partially specified segments -- such as 'velar
obstruent' and 'voicing' -- and affixation can make two segments
simultaneous (e.g. an affix consisting of 'voicing' can be
simultaneous with the last segment of a stem, e.g. 'velar
obstruent').

Strategy A sounds like it might be the more appropriate one in
the present instance.

--And.