Re: the sound [a]

Joe said:
> Mark P. Line wrote:
>
>>So, when we're studying the basic color terms of a language, we believe
>>that we can ascertain what a speaker's perceptual prototypes are.
>>
>>But in the case of phonemes, we can't...?
>>
>>Maybe we could just ask native speakers which sound they hear.
>>
>>
>
> But they'd have to be phonetically trained natives.  Which, when
> recording endangered and/or unknown languages, this is probably rare.
No, just phonemically trained. We train our informants to give us the
kinds of answers we're looking for when we're eliciting grammaticality
judgements, so I see no reason not to train them to give us the kinds of
answers we're looking for when we're eliciting phoneme categorization
judgements.

Nothing in the procedure involves any knowledge of phonetics -- just the
ability to assign a sound to one of the subject's native phonemes (or to
reject the sound as non-linguistic or foreign or whatever).


> An oddity, it seems to me, is the transcribing of Spanish [D]~[d],
> [B]~[b] and [G]~[g] as /d/, /b/, and /g/, respectively, despite the fact
> that [D], [B], and [G] are used in more situations.
Again, I think you would have to determine which sound ([D] or [d]) is
prototypical for /d/ among a population of speakers, just as the aspirated
stops (not the unaspirated stops) are the prototypical allophones for
speakers of English. The folks who design writing systems for previously
unwritten languages have been all over this subject. Google makes the wise
man happy.


-- Mark


Polymathix
San Antonio, TX