On Fri, 18 Jul 2003 09:33:48 +0100, Richard Wordingham
<
richard@...> wrote:
>What keeps *sW out of the standard reconstructions?
Probably the fact that they haven't been written by me. The standard
reconstructions also fail to mention Jens' nom.sg. *-z. As far as they're
concerned, an *s is an *s.
There is of course a small number of words (most notably the numeral 6 and
certain forms of the 2pl. personal pronoun) where the reflexes (Av. xs^v-,
Bryth. chw-, Slavic s^-) show something funny going on, but there is
probably too little of it to be totally convincing.
Some further points:
As in the case of *kW vs. *k^w, we have to distinguish between *sW and *sw.
Both become <k`> in Armenian (*swesoresW > k`ork`), but Slavic has
<sestra>, not *<s^estra>. I believe *sW gives /x/ ~ /s^/ in Slavic, but
*sw gives /s/. Unfortunately, there is little overlap between Slavic and
Armenian:
Arm Slav
*sW- k`- x-, s^-
*-sW- -0- -x-, -s^-
*-sW -k` -0
We can only compare initial *sW (it's unfortunate that neither Slavic nor
Armenian have maintained the distinction between nominative *so (which I
believe must be *sWo < **tu-) and oblique *to < **ta-).
After a consonant, Armenian *sW gives /s/ (e.g. acc/loc.pl. -s < *-ns(W)
and *-C-sWu). I do believe the *sW is responsible for the colouring of -u
in loc.pl. *-su, from **-sW-i. Slavic has -xU everywhere, not only in the
i- and u-stems, which is usually explained as analogy. The Slavic sigmatic
aorist also has -x-/-s^-, which I explain as *sW (originally from 2/3sg.
aorist *-sW, taken as an aorist marker instead of as person markers, cf.
the Old Irish t-preterite from 3sg. past tense *-t). The Slavic thematic
2sg. -es^I (OCS -es^i) also reflects *-sW- (*-e-sW-i), as expected.
=======================
Miguel Carrasquer Vidal
mcv@...