schwa-raising - formants

[Mark J. Jones]

----------------------------Original message----------------------------
Bob Rankin comments on the pronunciation of 'communication' by the BBC's
Lise Doucett (see appended text).

He states that the pronunciation involved a lax high front vowel /I/, with
a fronted realisation of /k/ preceding it. He concludes that regardless of
phonetics, this realisation is a phonological /I/.

Firstly, I would like to point out that some speakers may indeed have
phonologised this pattern, though it remains a moot point whether /I/ or
/@/ would be the target of closest auditory similarity. There is also the
issue of a fronted /u/, normal in many varieties of British English. I find
this an easier alternative than /I/ for 'communication' on auditory
grounds, but there may be variation here too.

Further, in 'communication' (which in my speech has a schwa-like unstressed
vowel), the nasalisation may play a perceptual role: nasalisation will make
the bandwidth of F1 broader, and also affect the higher formants. The
untressed vowels in question are very short, and if in addition formant
transitions are moving throughout (as in the case of F2 i reported for
schwa in 'Copernicus'), quality variation in perception is likely.

So it seems that a great deal of variation is heard. The same comment is
made for schwa realisations of the STRUT vowel in Cardiff (Mees and
Collins, p. 189, in Urban Voices, eds. Paul Foulkes and Gerard Docherty,
Arnold, 1999) and for schwa in Californian English by Peter Ladefoged in
his description for the IPA Handbook (Cambridge University Press, 1999).

I'm not sure on what grounds Bob Rankin makes the claim that these vowels
are phonologically /I/. If it is on the basis of fronted /k/, the
articulation of /k/ before /@/ in 'Copernicus' or the /^/ vowel in 'cud',
or the reverse epsilon/long schwa in 'curd' is definitely more similar in
auditory terms to /k/ in 'kid' than /k/ in 'card', as it will have a
relatively front articulation. In other words, /k/ sounds fronted for many
central vowels. I would venture to say that the degree of fronting of /k/
proves nothing in this case, and that the phonological status of IPA
reverse E as an allophone of a particular vowel remains to be demonstrated.
>From the phonetic point of view, the two are non-identical in my own
speech. Clearly more instrumental work is needed.

Mark

Mark Jones
Department of Linguistics
University of Cambridge

Bob Rankin's original comments:


> Sensitized to this question by the recent correspondence I was listening
> to the BBC newshour this noon and heard Lise Doucette (sorry if I
> misspell her name) say very clearly [kImyunikey$In] 'communication' with
> a very clear phonemic /I/ (small cap I) as in the proper name Kim. Her
> male colleague said the same word a moment later with a phonemic schwa.
> Whatever their etic makeup, phonologically these things are lax high
> front vowels. It was very striking. It also seemed to me that the fronted
> variant of /k/ was used with the /I/.
>
> Bob Rankin
>