Re: Evidence for Nostratic? (was Re: Proto-Uralic?)

On Mon, 7 Jul 2003 11:24:22 +0200, =?iso-8859-1?Q?J=F6rg=20Rhiemeier?=
<joerg_rhiemeier@...> wrote:

>I don't believe in a monovocalic system at any point of PIE's prehistory.
>While there are instances of /i/ and /u/ that arose from syllabic
semivowels,
>there must have been /i/ and /u/ in pre-ablaut PIE.  I also think that
these
>took part in ablaut, being diphthongized to /ai/ > /ei/ and /au/ > /eu/
>in full grade.
While /ai/ > /ei/ seems realistic to me, /au/ > /eu/ does not.  The problem
is solved if you have /aya/ > /ey/ and /awa/ > /eu/.

>One moment.  Only neuters had original endingless nominatives.
>Masculine and feminine nouns had a nominative in *-s, which was
>lost in those forms that don't have it at a rather late stage,
>and I think that *-s is from a suffixed animante demonstrative *-sa.
PIE *kerd and *genu are inanimate nouns.  The reconstructed form for 'foot'
is *pots, which I think may be erroneous.  Latin has pes, pedis; Greek has
pous, podos.  These point to a problem with current reconstructed PIE: what
was the original genitive suffix, *-es or *-os?  Perhaps we can find the
answer to this question.

One thing that I did find out is that there might be a way to reconcile
Latin and Greek disparate genitive forms.  First, we must assume that PIE
originally had a stress-accent system that fell regularly on the penultima,
and that that system was subsequently abandoned.  We must then assume that
either Proto-Italic or Proto-Hellenic separated from the rest of PIE before
a new accent scheme was finalized.

Also, why do you think that animate nom. sg. *-s comes from a suffixed
animate demonstrative *-sa?  The sigmatic nominative seems to be intimately
connected with the sigmatic genitive.

>This is the first time I see the PIE word for "name" reconstructed
>with an initial *g.  So far I thought the initial consonant was *h3
>(/x^w/).
That's an idea of my own.  It seems logical that a root *gno- 'know' (more
likely 'be acquainted with'), when coupled with an extension *-men, which
was some sort of abstract suffix, that a word with a meaning akin to 'name'
would result.  But perhaps I am wrong (it won't be the first time!).

>The genitive singular ending must have been longer.  Nom. sg.
>was just *-sa, while gen. sg. was *-a-sa.  What I suspect here
>is a case of suffixaufnahme, i.e. the genitive ending was *-a,
>but took the case suffix of the possessum in addition to that,
>as if it was an adjective.  Thus, *-a-sa was the nominative of the
genitive.
>The thematic noun/adjective class might also come from such genitival
forms.

I'm gonna throw something out here, maybe it's a crazy idea but I think it
might be correct.

According to Sihler's New Comparative Grammar of Greek and Latin, Proto-
Latin (or something like that) often shortened the nominative singular
forms of i-stems and o-stems, i.e. frons, frondis 'leaf,' ager,
agri: 'field.'  Perhaps PIE did such a thing with its root nouns.  After
all, Latin preserved a distinction between nom. sg. frons (earlier
*frondis) and gen. sg. frondis (earlier *frond(e)yes, which would've
quickly become *frondis).

>> Agreed.  One question though: why did (Pre-)PIE /a/ remain /a/
before /x/?
>
>Because /x/, being velar, prevented palatalization of adjacent vowels.
>I think this is quite natural.  (And /x^w/ prevented palatalization and
>added rounding to the vowel, thus leading to /o/.)
Where does this supposed palatalization come from?

>Which Greek forms have /x/ > /h/?  What I know from Greek is
>*h1 > e, *h2 > a, *h3 > o in initial position before a consonant,
>and in forms with syllabic laryngeals.
Sorry, that must be my mistake, then.

>These endings are used when the object is indefinite, right?
>They look more like a Hungarian innovation to me than anything else.
>But the possibility of it coming from a Uralic stative conjugation
>cannot be ruled out: 1sg. -k might be cognate to IE stative 1sg. -h2
>(PIE seems to have changed final stops into fricatives, cf. PIE pl. *-s:
>PU pl. *-t).
I think the PIE stative conjugation is more likely 1st **-xa, 2nd **-dha,
3rd **-ha or **-ya.

>So do I.  What do the Selkup forms look like?  Are they similar to the
>Hungarian ones, and what is their meaning?
Unfortunately, I don't know very much about Selkup, but I will look into it.

>It seems as if Proto-Uralic verbs agreed to their objects only in number,
>not in person, and to their subjects in person and number.
I think it's more likely that Proto-Uralic verbs did not agree with their
objects at all.

>> PIE reconstruction, as it stands today, is very muddled.  I think the
>> reason behind this is that PIE went through at least two changes in its
>> accent system during its 'lifetime.'  There's also a very real
possibility
>> that some of the accepted reconstructed forms are incorrect.
>
>Yes.  There are so many open questions and controversial issues
>that one has to assume that much of what we reconstruct is wrong.
I certainly agree.  Perhaps other, more learned, people will take our cue
and start rehashing commonly accepted PIE reconstructions.

>The form *wl.kWos is indeed problematic as it requires a reduction
>of an accented vowel, which seems unlikely.  I think the accent was
>on the *o, and later retracted by analogy with other thematic nouns.
>This might indicate that it is of different origin than most thematic forms
>which might be adjectival in origin.
If you are correct, and the accent was originally on the *o, then the
entire tonal accent theory can be thrown out the window, unless it can be
proven that labialized velars rounded their following vowels.  The problem
can be best illustrated graphically:

wálkWa, walkWása > wálkW, wl.kWás > *welkW, wl.kWes

This is assuming that tonal accent had a high tone wherever the earlier
stress-accent had a stressed syllable.  However, if the tonal accent
instead had either high tone falling regularly either on the initial
syllable or the penultimate, we get this:

wálkWa, walkWása > wálkW, wl.kWás > walkW, wl.kWas > *welkW, wl.kWos

The latter reconstruction also assumes that syllabic resonants were not
vocalized regardless of tonal accent.

- Rob

Replies