>Some words seem to have an original *i or *u (e.g. *mus). At the very
>least we'd need a system of /i u a/ probably with length contrasts.
>One could then try to derive *e and *o from them. Perhaps *a, which
>we'd expect to be the most common and the one partaking in ablaut,
>was raised to *e or *o in certain environments (e.g. apophonic *o
>appears to occur primarily before nasals).
I've always been curious about those etyma. They are clearly unusual
since they contain an otherwise "reduced" vowel in an accented syllable.
I've considered the possibility that in some instances an early Late IE
*Vi/Vu may have become PIE *i/*u when it otherwise shouldn't have.
In pronouns like *tu:, I'm sure that the vowel has been simply reduced
from earlier *teu. The reduction here can be explained as a special
development of pronouns and demonstratives (such as *i- < *ei,
*kWi- < *kWei, etc). However, *mu:(h)s "mouse" is strange.
On the other hand, one idea I have to explain it involves possible
early Late IE rules of syllable shape vis-a-vis laryngeals. If we
reconstruct *mu:hs "mouse" in the nominative, *u is long as all
vowels are in the nominative form, while the presence of *h ensures
that *u always appears long in other case forms. The realization
of an accented *u here then can be explained as the product of
an early Late IE nominative *mouh-s& becoming in this special
case *mu:h-s, rather than **mo:uh-s due to the presence of the
laryngeal which would otherwise make the one syllable unusual
in its spoken length and complexity. Thus this etymon might be
the result of syllable restructuring and simplification.
I suspect that this same syllable restructuring might be to blame
for other otherwise unsolvable oddities in IE morphology.
The new MSN 8: smart spam protection and 2 months FREE*