----- Original Message -----
From: Mitko Sabev
To: phoNet@egroups.com
Sent: Wednesday, April 12, 2000 11:13 PM
Subject: Re: [phoNet] Webster/IPA
(1) My transcription was intended to be (not very narrowly) phonetic rather than phonemic, hence some contrasts which MAY be regarded as superfluous (depending on your theoretical standpoint) in phonology, such as the one between the two vowels of abut. One could argue that whereas the vowel of cut is a FULL vowel with a relatively stable quality and a well-defined articulatory target, the schwa proper is a cover symbol for a variety of weak vowels used according to the phonetic context (e.g. rather open word-finally, rather high before a velar or a palatoalveolar) and not having such a target.
(2) The same concerns vowel length. I agree that quantitative contrasts in GenAm are not salient enough to be really relevant, but I decided to use a legth mark for [i:] and [u:] as a rough equivalent of transcriptions which show them as diphthongs (complex vowels), such as non-IPA [iy], [uw], still used by some American linguists.
(3) Final vowels, even if unstressed and weak, are typically quite long, so I don't think [i:] is quite unsuitable. The weak vowel has the same diphthongoid character as its stressed counterpart.


\&\ as a and u in abut [ə], [ʌ] (unstressed/stressed respectively)
Both the stressed and the unstressed vowel in 'abut' can belong to one and the same phoneme only for those American speakers who do not have a phonetic distinction between them.
For those who have different phonetic values for the two vowels, the distinction is, as elsewhere, not only phonetic but phonemic, since /ʌ/ may occur in unstressed positions as well:
'unending' /ʌnˈɛndɪŋ/ vs 'an ending' /ənˈɛndɪŋ/.
WWWebster's transcription is representative of a widespread American accent where /ʌ/ and /ə/ have merged under a phonetic value more close to schwa [ə]. So I believe WWWebster's "&" is more likely to correspond to IPA [ə] only.
\&r\ as ur and er in further [ɝː], [ɚ] (stressed/unstressed r-coloured vowels; worry = [ˈwɝːiː])
(i) As above, for those Americans who do not make any distinction between the two, what is customarily transcribed as (ː)/ can occur in both stressed ('curve' /$ kɝv, £ kɜːv/) and unstressed ('excerpt' /$ˈɛksɝpt, £ˈɛksɜːpt/) positions. Again, I think IPA [ɚ] would be an adequate match for W's "&r".

(ii) I don't consider vowel length relevant to (phonemic) transcription of GenAm. In the standard accents of MnE quantitative phonemic opposition of vowels no longer exists -- all contrasts being qualitative. In transcribing RP, however, vowel length is, more often then not, marked. There are two reasons (or rather "excuses") for this: (a) in this way the "traditional" division of short vs long is preserved; (b) although not distinctive, the length of the "long vowels" is phonetically salient in RP.
In GenAm vowel length is by no means as prominent as in RP -- it is thoroughly conditioned by environment:
$/ɑ/ (/α/) : 'pa' [pʰɑː] ~ 'pod' [pʰɑˑd̥] ~ 'pot' [pʰɑt]; 'see' ~ 'seed' ~ 'seat'; etc. (Although in RP the same environmental conditions are operative, there are quantitative differences inherent to different phonemes -- /ɔː/ is inherently longer than /ɒ/.)
Furthermore, weak vowels are usually of rather short duration so I can't account for the length mark in the second syllable of your [ˈwɝːiː].

In the lines above I have been mostly concerned with the relevance of quantity to phonemic transcription of GenAm, and although your transcription is not explicitly phonemic, I find it unnecessary to state length in GenAm unless a very narrow transcription is intended.

\A\ as a in ace [eɪ] (a diphthong!)
Like GOAT, for many Americans it is a pure vowel -- [e(ː)].
\ä\ as o in mop [ɑ(ː)] (lengthened when stem-final, as in pa [ˈpɑː], in this position it may be replaced by [ɒː]) (underlining is mine)
$[ɑ(ː)] and [ɒ(ː)] (or up to [ɔ(ː)]) are in free variation in various other environments too: 'strong', 'dog', 'broad', 'brought', etc.
\E\ as ea in easy [iː] (usually slightly diphthongal [ɪi]; may be unstressed, as finally in very [ˈvɛɹiː])
WWWebster's "E" can stand both for a weak (and therefore inherently phonetically short) and a strong vowel — /i/; a length mark, if needed, should be added only to the strong vowel.
\hw\ as wh in what [ʍ] (= the voiceless counterpart of [w], also transcribed [hw]; most speakers use [w] instead)
Phonetically, [hw] is not the same as [ʍ], although together with [w] these are in free variation whenever the graphic realisation is < wh >. [hw] is a sequence of two sounds: "voiceless glottal fricative" + "voiced labio-velar approximant", while [ʍ] is a single sound — "voiceless labio-velar fricative".
All this variety is a reflection of the gradual lenition that took place since Old English: OE cluster < hw > [xw] developed into a single (voiceless) fricative — LOE, EME < hw, wh > [ʍ] (which is preserved in some accents), and then into an approximant — MnE < wh > [w]. The cluster that exists nowadays, [hw], is most probably an artificial formation, an attempt to imitate [ʍ] in order to keep < w > and < wh > apart. Indeed, many people (with no special ear training) will find [ʍ] and [hw] identical.
\l\ as l in lot [l] (pronounced as ‘dark’ [ɫ], except when followed by [j]; syllabic in bottle, whistle) (underlining is mine)
For many Americans /l/ is "dark" in all positions. I will never forget my first American teacher pronouncing 'volume' [ˈvɑɫjum] with a 100% valorized /l/.
\O\ as o in go [oʊ] (a diphthong; often unstressed, as in follow)
(i) As with [e(ː) ~ eɪ], this is often a pure vowel — [o(ː)].

(ii) Indeed, unstressed /o(ʊ)/ is (in J. Wells's theory of weak vowels) a weak vowel in GenAm. This is why it is much more frequent than RP (strong) /əʊ/: 'obituary' is usually /oʊˈbɪʧuɛri/ in GenAm, while the normal RP form, as regards the initial vowel, is /əˈbɪʧuəri/ (/əʊˈbɪʧuəri/ being typical of more careful speech).