wh

[Sam Martin]

In my (now I realize "old-fashioned") midwestern version of /hw/ I
perceive a definite [h] segment, followed after a quick voweless [w] by a
voiced [w], so I have no difficulty with the phonemic analysis as a
cluster. I do the same sort of thing with the "hu-" of "huge", "human" and
stress-downgraded "humanity": [h] followed by [yuw] with a voiceless y
(ich-laut) between that is only brief, at best. For the record I am
familiar with many speakers in the northeast (and probably elsewhere) who
fail to distinguish "Hugh" or "hew" from "you", just as they cannot hear
the difference between "which" and "witch".

I should think this would lead to spelling problems, but then who bothers
to spell any more?