British Dialects Book

[Mark A Mandel]

        [dInIs, round 1 (note that these brackets don't indicate
         phonetics!)]
#>#>I don't understand what a phonemic (since it is a mental
#>#>representation) pronunciation is. Nobody ever pronounced a phoneme.

        [Mark's example omitted]

        [dInIs 2]
#Nope. I still don't grasp the concept of a phonemic pronunciation.
#Your example shows us that some people have two phonemes - /O/ and
#/a/ - represented by the phones [O] and [a]. Others have one phoneme
#/a/ represented by the phone [a]. So far it looks like people are
#"pronouncing" phonemes and that phones are gratuitous.

Well, insofar as "pronunciation" refers to the articulatory and acoustic
production of speech, the expression "phonemic pronunciation" is subject
to your criticism. But I, and I think others in this list, have been
using "pronunciation" here to refer to the graphical representation of
the way(s) a word is pronounced. At Dragon Systems we called such
representations "prons".

Pronunciations in the first sense, which you seem to be using
exclusively, *exist* in articulation and acoustics, and the only way to
*present* them is to perform them, play a recording, display a sonogram
or articulatory diagram, etc. In writing, they can only be
*REpresented*.

Pronunciations in the second sense, or "prons", are symbolic entities
whose existence is in ink on paper or phosphors on screens. They can be
phonemic or phonetic. Being symbols, they are related to the things they
represent only by way of the meaning they have in people's minds. The
pron
        [sOt]
 is phonetic because we have agreed that the square brackets mean "this
is a phonetic representation". The pron
        /sOt/
 is phonemic because we have agreed that slashes mean *that*.

You say that nobody ever pronounced a phoneme. I could reply that nobody
ever wrote a phone.

-- Mark A. Mandel