I do not think it is the original existence of unit voiceless aspirates
that is at stake here. For typological admissability that is not required.
If only there were phonetic segments of the type /ph, th, kh/ in PIE,
there were four series as far as phonological typology is concerned. I
also think there are cases where laryngeal-induced aspiration is retained
even when the laryngeal is lost. My best oone is Gk. terthron 'summit',
analyzed as *terH2w- 'win, conquer' + suffix *-tro-m (allomorph of *-tlo-m
used after roots containing aliquid) 'means or place of -ing'. Before a
consonant cluster a sequence -RHU- (sonant + laryngeal + semivowel) loses
the entire sequence -HU- (here *-H2w-); thus also here, but in this case
the aspirating effect of /H2/ on preceding /t/ had already worked. If
correct, this is a clear case of phonemicization of an aspirate caused by
"preaspiration".
The Greek aspirated perfect was explained by Sapir a long time ago on the
basis of the IE 1sg perfect: pepompha would then reflect *-p-H2a,
generalized from there to the rest of the paradigm. The explanation has
lost its flavour because the forms are limited to Attic-Ionic and are not
Homeric. But surely an event of analogical spread, as this must be, can
show different results in different dialects.
Jens
On Sat, 27 Sep 2003, P&G wrote:
> > >Paul Elbourne has formulated the
> > >following fine law for Greek: th > t /[+cons, +cont]
>
> On the other hand, Greek inexplicably aspirates voiceless consonants in
> quite a few of its perfects, and elsewhere. I still have great
> difficulty
> taking the "evidence" for aspirated voiceless consonants seriously,
> especially when we know that in I-I many of them are from unaspirated
> stops
> + laryngeal, and in Greek there was a tendency to aspirate for causes we
> do
> not yet understand.
>
> Peter
>
>
> Yahoo! Groups Sponsor
>
> Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.
>