Rob:
> Something to think about: Perhaps there were two stages of accent in
> PIE. The first stage was a stress accent -- this is where the
> syncope occurred. The second stage was a tonic accent, which gave
> rise to most of the Ablaut distinctions.
I've already assimilated that into my theory a few years ago. Tonal
accent developped during Late IE, necessarily so if stress accent
caused Syncope as it no doubt did. If the accent was stressed during
Syncope which marks the start of early Late IE, then tonal accent
must occur after it. The stress accent must have lingered even till
Acrostatic Regularization since stems ending in *-รก- became *-&-
without the accent (suggesting vowel erosion). After that is when
I'd place the change of stress accent to tonal accent.
As for before the event of Syncope, such as in OIE or IndoTyrrhenian,
I feel that the stress accent was much lighter than it came to be
in late MIE since there is relatively little to show in way of
vowel erosion before that. The only piece of evidence I see might be
the reduction of final *-e and *-a in ITyr to OIE *-a. Also, the
accent shifted in OIE from a fixed initial accent like in Germanic
to a quasipenultimate one by the time MIE was in contact with a
northern dialect of Semitic/ish from NW Turkey. IndoTyrrhenian
must have had an initial accent (like that of Uralic and Altaic)
since this also explains Tyrrhenian's accent which was "quasiinitial",
that is, found either on the initial syllable or sometimes shifted to
the second syllable to avoid accented *a when possible.
> Another possibility: Much is made from the contrast between o-grade
> root-accented thematic nouns and ending-accented thematic nouns. I
> suspect that one or the other is original and the other is a later
> innovation. Logic says that, since "original" thematic nouns have
> initial accent, the ending-accented ones must be later.
Yes.
> A better riddle is to try to figure out where Latin canis fits into
> all of this. Furthermore, how the Latin root noun genitive singular
> -is (presumably from *-es) can be reconciled with everything else.
I don't think it's a major issue. Rather, Latin is a clear exception
to the norm. While it's interesting and I have no immediate
explanation for it myself, it can have no possible pertinence to
the issue of *kwon- itself which is nonetheless well substantiated
in plenty of other languages without difficulty, including all its
precise case forms.
= gLeN