tgpedersen wrote:
> Would an irregular shortening followed by irregular lengthening have
> done the trick?
Jespersen regarded /gO:n/ as a compromise between /gOn/ and /goUn/ (the
latter regular but no longer used in the standard accents), combining
the quality of one of the variants with the quantity of the other. He
likewise claimed that /fa:D&r/ had originated as a compromise between
the free variants /feID&r/ (with open syllable lentgthening) and /faD&r/
(without it). There may be something in this explanation, although it
isn't widely accepted.
Piotr