From: Andrew Jarrette
Message: 63843
Date: 2009-04-16
>Perhaps because /p/ is sufficiently different from /s/ that it did not become assimilated to the /s/ and then elided, and its retention therefore prevented compensatory lengthening. Perhaps Doric <po:s> originated from *pots > *poss > *po:s. Only /ts/ could undergo this shift perhaps.
> --- In cybalist@yahoogroups.com, "Andrew Jarrette" <anjarrette@> wrote:
> >
> > --- In cybalist@yahoogroups.com, "dubbelax" <dubbelax@> wrote:
> > >
> > > --- In cybalist@yahoogroups.com, "Andrew Jarrette" <anjarrette@> wrote:
> > > >
> > > > Well, wouldn't one expect Greek *po:s in the nominative singular from IE *po:ts? Odoús has <ou> from *ons (< *onts), which became nasalized *õs which became a long vowel higher than Greek <o:> which was therefore written <ou> (hence <-ous>. At least that's what I've always assumed. Where would <ou> come from in <poús>? Normally *ts becomes <s> in Greek, does it not, without influencing a preceding vowel, long or short? Or am I wrong here? What is the 2nd sg. indic. pres. of *h1ed- in Greek? What other monosyllabic consonant stems end in *-ts < *-ds in PIE (and in Greek)?
> > > >
> > > > Andrew
> > > >
> > >
> > > You are right, Andrew - my odoús is an irrelavant example.
> > > I do not recall other cases of a final dental + s. I am not an expert in Greek, though. Anyway, methinks that the word in question sounded [pots] in some period of time, with a dental possibly restored after óps etc. What do you think of this?
> > >
> >
> > And you mean that the *o: in the nominative singular arose from compensatory lengthening due to the loss of *t, which was later reintroduced on the analogy of <óps> etc.? Sounds good enough for me, but then what about *wo:kWs? Why does it have a long vowel?
> >
> > Andrew
> >
>
> I have a counterquestion: Why is <óps> without a long vowel? :-)
> The only thing I can say is that wo:kWs is a hypothetical word.Well of course, all PIE words are hypothetical words. *wo:kWs is based on Latin <vo:x> and Sanskrit <va:k, va:c->, so it has a solid basis.
> If there were really no "original" long vowels in PIE (no vrddhi, no contractions, no nothing)Except vrddhi and contractions are certainly not "original" long vowels, they are later developments.
>, we should reconstruct the protoform with a short vowel. The real forms in various languages, resulting from some loss/compensation, could then have been wo:kW, wo:s or even wo: (if we stick to o-vocalism in Nominative). Everything resembling wo:k(W)s would be secondary and so could be wokWs (although some IE languages might have preserved the original, unlengthened form).Oh, I see what you mean, that the *-s on *wo:kWs might be secondarily added on by analogy after it had been lost and produced *wo:kW by compensatory lengthening. The only thing is, neuters without *-s also could be lengthened, e.g. *k^e:r(d) "heart" and *do:m "house" (of course, both in these neuters and in masculine and feminine root nouns analogical levelling of the vowels of inflected forms often occurred in various languages). It seems to me that this suggests that the primary lengthening factor for these root nouns is that their nom. sg. (and voc. sg.) was a stressed monosyllable. This is what Beekes has also said, casually, in his "Comparative Indo-European Linguistics: An Introduction", and I'm sure many if not almost all IE linguists have said the same. I forget easily, though.