Re: [tied] Re: Eggs from birds and swift horses

24-02-2004 13:42, elmeras2000 wrote:

[Piotr:]

>> Not _very_ different. *O > *o when it turns syllabic (i.e.
>> interconsonantally) but it's possible that *O > w before a vowel.
>
>
> How do you define very? [...]

I mean the only real difference is the direction of the suggested
metathetic movement: *CROhV- > either *COrhV- (> *CorhV-) or *CrhOV- (>
*Crhwo-). Is that _very_ much of a difference?

> But not *very* hard? /H3/ was rounded, so if assimilation rounds the
> whole group become a long voiced rounded fricative, it will not be
> unlikely to produce a segmental /w/ if a vowel follows.

That's certainly a possibility.

> The result looks like it ended in a segment that was not in the
> input. That is not well-described as metathesis, though there may be
> a thousand ways it can have involved one on its way.

Phonemes are bundles of features. If features migrate (through
assimilatory spread) to a new location a couple of segmental slots away
where a new linking point is created for them while their original host
segment disappears, the result can legitimately be described as
metathesis. I agree, however, that we can only speculate about the
phonetic process involved and the only thing we really see is the
outcome, which admits of several different explanations, all of them
reasonably plausible. What we certainly agree on is the phonetic, not
morphological origin of the *-w- in *pr.h3wo- etc. A *-wo- morpheme has
no business to be there.

Piotr