> At 2:03 pm +0100 23/2/00, BP Jonsson wrote:
> >At 07:04 +0100 21.2.2000, Raymond Brown wrote:
> [....]
> >that it may be different in some areas with a Welsh substrate, where
> >gemination was probably introduced by first-generation English-speakers as
> >a spelling pronunciation.
> Not at all - it is the Welsh substrate.
> In Welsh voiceless, aspirated plosives are regularly gemminate after a
> stressed syllable, and the aspiration is retained, tho the spelling has
> only a single letter.  The Welsh for 'happy' is 'hapus' /'happ_h1s/.  Welsh
> 'criced' /'k_hrIkk_hEd/ (the game, not the insect) sounds quite different
> from its English origin.
> >Or is the /p/ of that pronunciation you are hesitant about aspirated?
> >Perhaps [h&p_hi] has been re-interpreted as /h&p.hI/!
> Not likely - Few, if any, English speakers who are phonetically illiterate
> are aware of aspiration of voiceless plosives.

That reanalysis needn't occur consciously.

But anyway, this gemination applies to sonorants and fricatives too, after
a short vowel. E.g. "sunny", "silly", "fussy". And not just before /i/.
Also "runner", "willow" etc. Strictly speaking, I shouldn't call it
gemination, because I don't know what the best phonological analysis
would be; a safer term would be "consonant lengthening". Even at a
phonological level gemination might not be accurate, because e.g. in
"thunder" other accents have thu[n]der to S. Walian thu[n:]der, fi[l]ter
to fi[l:]ter etc. So it strikes me as a rule of phonetic realization that
is sensitive to syll structure: "lengthen realization of the coda" (assuming
short vowels always precede a coda).

On the happY vowel:
> At 2:35 pm -0500 23/2/00, John Cowan wrote:
> >For me, a GA speaker, the last phoneme is unequivocally /i/, so there
> >may be other dialects for which this is also true.
> Certainly, it's alway been true of me also, a speaker of south east England
> origin.  AFAIK it's the most common sound for final /i/.   The pronuciation
> /I/ is heard but is IME regarded as rather 'upper class' or affected.

/I/ or [I] for happY has a weird distribution. RP and most of the north
of England, though not coastal cities (Liverpool, Newcastle, ?Hull), but
also parts of the USA (i.e. whoever's accent Mick Jagger is imitating
when he sings happ[E]), but outside of RP the realization tends to be
even opener, towards [E]. Also a problem for phonemic analysis, like the
windOW vowel.