On Sun, 13 Mar 2005 16:42:40 +0100, Miguel Carrasquer
<
mcv@...> wrote:
>On Sun, 13 Mar 2005 13:26:12 +0000, elmeras2000
><jer@...> wrote:
>>> I see nothing wrong with Illich-Svitych's tu~kU.
>>
>>I think there is plenty wrong with it. It has a zero-grade alternant
>>with short -u- in tùkti (prs. tuñka) 'grow fat'; the Latv. form is
>>tàuks 'fat'. LIV posits *tewk- besides *tewH2-. It looks like a case
>>of laryngeal hardening, the -k- being then a variant of the -H2-,
>>not an addition to it.
>>
>>If Balto-Slavic had a more reduced propvowel with laryngeals after u-
>>diphthongs than after sonants proper, sequences like *-ewHt- might
>>work just as *-eyHt- and *-eHt- by attracting the ictus from the
>>following syllable.
>
>I don't have the time now to study the material. I'll be
>back on this later.
I've had a preliminary look, but it's not an easy task.
The question is: which PIE sequences *VRH are affected by
Hirt's law and which not?
The sequences in question are:
eiH > í ei > i~
euH > (j)ú eu > (j)u~
erH > ér er > e~r
elH > él el > e~l
emH > éN em > e~N
enH > éN en > e~N
oiH > ê' oi > ê~
ouH > ú ou > u~
orH > ór or > o~r
olH > ól ol > o~l
omH > óN om > o~N
onH > óN on > o~N
A complication is that í can also come from *iH, éN from
*n.H/*m.H, e~N from *n./m., ê' from *e:, *eh1.
There's not much point in looking at verbal forms, where we
would expect for roots containing *VRH:
no Hirt's Hirt's
C-stem bar. a.p. a a.p. a
C-stem mob. a.p. c a.p. a
V-stem bar. a.p. c a.p. c
V-stem ox. a.p. a a.p. a
Thematic barytone verbs are not affected by Hirt's law, and
oxytone (theme-stressed) verbs are affected by the
jábloko-rule. The difference can only be detected in PIE
athematic mobile verbs.
In nouns, we have more chance of success:
no Hirt's Hirt's
C-stem bar. a.p. a a.p. a
C-stem mob. a.p. c a.p. a
V-stem bar. a.p. a a.p. a
V-stem ox. a.p. c a.p. a
Here, vowel-stem masculines and feminines can show the
difference: if they were oxytone in PIE, but a.p. a in
Slavic, Hirt's law has worked. If they are mobile, Hirt's
law hasn't worked.
The problem is finding out which words were oxytonic in PIE
(and, of course, whether a laryngeal was involved at all).
It seems pretty clear to start with that *eiH (= *ei&) was
_not_ affected by Hirt's law (*leiH- > li"ti, lila`), but
*oiH _was_ (*poiH- > pê"ti, pê"la; poiHmé:n > OLith.
píemuo). An reasonable initial hypothesis would then be
that this can be generalized to all *VRH-sequences: *eRH
gives *er& and does not trigger Hirt's law, *oRH does not
vocalize the laryngeal and triggers Hirt's law. The
infinitives bo"rti, po"rti, po"lti etc. and their Latvian
cognates seem to confirm that. So does me"lti if it
acquired the accentuation of *mo"lti (= Lith. málti). The
"swamp/mud" word *bol&tóm > bo"lto is another case in point
(Slavic bo"lto could reflect the jábloko rule, but Latv.
ba~lts "white" unequivocally points to Hirt's law).
That would mean that mobile su~rovU "raw" would have to be
from *seuHr- (vrddhi from *suHr- = sy"rU ?), mobile dus^a`
and du~xU from *dheuh2- (du"lo might be from *dhouh2-ló- but
as a neuter is undecisive, as is the jé-verb du"joN, du"ti:
retraction here is the result of the jábloko-rule).
More later, time permitting, when I've looked at the data in
more detail.
=======================
Miguel Carrasquer Vidal
mcv@...