On 2006-09-06 12:20, Sergejus Tarasovas wrote:
> Interestingly, both Illic^-Svityc^ and Derksen (in "An introduction to
> the history of Lithuanian accentuation") write *kaHwlós (with a non-
> syllabic *w, like in *tenHwós, rather than syllabic *u) so that *aH is
> tautosyllabic in their interpretation. That's how I.-S. seems to
> interprete "long diphthongs". Does that, in your opinion, violate any
> phonotactic constraints and/or syllabification rules?
Phonemically, of course, it makes no difference, but syllabification is
a phonetically grounded phenomenon. In PIE, the laryngeals were less
sonorous than semivowels, so that e.g. *h2w- or *h2j- were perfectly
acceptable syllable onsets. One could imagine a syllabification like
*[kah2=wlos] ("=" symbolises a syllable boundary), but it seems to me
that *wR- clusters were only permitted word-initially. An
interconsonantal *w could only be realised as *[u], and *//keh2wlos//
would have been pronounced as *[ka=h2u=los]. Something like *[dah2jwe:r]
would not have been syllabifiable at all in PIE terms.
PIE laryngeals, like other fricatives, were also less sonorous than
nasals and liquids, but in some branches they could be lenited to the
point where their vocalisation became possible in -RhC- and *hR-
contexts. In Balto-Slavic, *hR- > R- (just like *wR- > *R-), but
apparently *-RhC- > *R&C- > *-:RC- (schwa loss with compensatory
lengthening).
Piotr