Re: [tied] Vowel Lengthening from V + Voiced Stop (was Vowel Lengt

On 2006-10-02 19:47, Patrick Ryan wrote:

> http://wwwgeocities.com/proto-language/WiikStudy.jpg
> <http://www.geocities.com/proto-language/WiikStudy.jpg>
>
> All the above remarks only have relevance to the discussion if
> the "Final" heading caption of the table is interpreted as
> designating a monosyllable, i.e. CV#.
>
> I interpreted "Final" to mean simply that the V was final, thus
> designating an open syllable, CV, in whatever position.
>
> Do you have information that would substantiate either
> interpretation?

Yes. "Final" means "word-final" (not necessarily in a monosyllable, but
also in words like <degree> or <guffaw>. Intervocalic consonants affect
the preceding vowels just like final ones, so the whole thing is not
dependent on syllabification. There are even some accents of American
English in which medial /t/ and /d/ fall together as a voiced alveolar
tap, but the vowel still shows the historical difference between them,
being shorter in <latter> than in <ladder>. There is no consensus among
phonologists on how to interpret this phenomenon. It seems that the very
intention of realising a voiceless/voiced consonant shortens/lengthens
the vowel.

> Many phoneticians have abandoned the terms fortis/lenis; and I
> am not sure what they offer in benefit to voiceless/voiced in
> the context of English.

OK, call them voiceless/voiced instead, but just bear it in mind that
the factor that affects the duration of the vowel is not the phonation
alone. Consonants also differ in length and in English voiceless
obstruents are normally longer than voiced ones, so there is a trade-off
relationship between adjacent segments in VC combinations.

> But this is NOT the same as saying V + Voiced Stop > V:, is it?

Not phonemically, at any rate, though the difference could be
phonemicised if the conditioning factor disappeared. For example, we may
be dealing with phonemicisation in progress in those accent that
distinguish <ladder, rider> from <latter, writer> using different VOWELS
rather than different CONSONANTS.

> For an inherently short vowel, the measure of [I] in an English
> word like <bilabial>
>
> ***CORRECTION: <bitumen> ***
>
> would be more likely to give a benchmark duration of the vowel,
> in my opinion, i. e. not stressed, not word-, phrase- or
> sentence-final.

All right, but the same phoneme will be realised a little longer in
<ligament>. Which of the two durations is the benchmark one?

Piotr