Rob wrote:
> Are you sure that that English sound change was incomplete? If so,
> please explain.
It left a residue of unshifted forms which are not accounted for by any
simple structural condition. It's true that they _usually_ have a labial
(other than /m/ or /v/, for reasons chimerical) before the vowel, and
_often_ a lateral or palatoalveolar after it:
wolf, wool (with an etymological short vowel), bull, bulrush, bullet,
bulletin, pull, pulley, pullet, full, fulmar;
push, bush, butcher.
But why <cushion> with an initial velar, if we have <cull>, <gush>,
etc.? And why was <put> not affected if <butter> was? Why <bulb>,
<bulk>, <budge>, <budget>, <Bulgar>, <bulge>, <pulse>, <pulp>, and
vacillation in <pulmonary>, <fulminate> and <pulpit>?
It seems that certain segmental configurations hindered the change
probabilistically without blocking it absolutely, so the lowering was
distributed unevenly across its potential lexical domain and was never
carried through to completion.
Piotr