wh-clusters

[Martin E HULD]

----------------------------Original message----------------------------
I am a native speaker of Kurath's Dialect 10, and although a number of
school teachers tried to enforce (in those days educators assumed they knew
right from wrong) the why/Y, which/witch distinction, it was entirely gone
from actual speech by all natives in the small coal mining towns of Western
Pennsylvania where I grew up, both in my generation and in the speech of
those from the earlier years of the twentieth century.  Nevertheless, I do
believe that I am sensitive to its presence in the speech of others.  I was
therefore amazed to hear a colleague a few years older than I, now in her
early fifties and raised in Red Creek NY, distinguish whether/weather but
fail to distinguish whale/wail; I then listened more closely and heard
[hwith] not [with].  It seems in her speech that the [hw]/[w] contrast was
generalized to function words but lost in nouns and verbs.  Has anyone else
noticed this or a similar pattern?

A second point, a number of texts, eg O'Grady et all. Contemporary
Linguistics symbolize [hw] as a unit phoneme with IPA inverse-w.
Historically, this is a poor choice since by the same w-deletion rule that
governs sword and two, we have an h-pronunciation for who and unetymological
whore; moreover, the Ayenbit of Inwit consistently represents initial PG [f]
by <u> (uerste = first) and PG [s] by <z> (zalt = salt), but [hw] or [xw] is
<hu> (huer = where; huich = which).  Therefore, despite Campbell (Old
English Grammar p. 20) who takes hw (and hl, hr, and hn) as digraphs of
voiceless segments, it seems best to regard them as clusters at least in Old
and Middle English.  I think the same is true of New English where whine is
parallel to swine and twine [hwajn, swajn twajn] in which there is an
initial voiceless segment which partially devoices the following glide.
Additionally, treating <why>, presuming you have it in your dialect, as
[hwaj] in contrast to [waj] (<Y>) is parallel to treating hue [hjuw] versus
yew [juw] as a case of an initial cluster.

I was wondering if anyone felt as I do that the inverse-w is an
inappropriate strategy for analyzing the remaining cases of phonemically
distinct <wh> in American dialects.

Martin E HULD