Uralic and IE

[Miguel Carrasquer Vidal]

"Glen Gordon" <glengordon01 at hotmail.com> wrote:

>MIGUEL:
>  In the secondary verbal endings -(e)t and -(e)nt, -t has probably
>  been restored analogically from -(e)ti, -(e)nti, but we also have
>  lautgesetzlich -e:r < **-ent.

>Yes, thus justifying **-nD becoming *-r, not **-n > *-r. If I have this
>right, **-ent would first become **-e:n before becoming *-e:r because
>according to you **-n > *-r. So, the change of **-VCs to **V:C (such as
>**-Vns to *-V:n) must occur later:

>           1.  *-n > *-r
>           2.  *-ns > *-:n   (: indicates lengthening of prec. vwl)
>               *-rs > *-:r

>Wait a minute, how do *-ter endings react in Hittite nominative then??

I don't think there are any. Non-neuter r-stems go: Nom.
sakuwassar-as, Acc. sakkuwassar-an.  The word kessar "hand" is
neuter in Old Hittite, or at least doesn't distinguish nom. from
acc. (kessar), later it has Nom. kessaras, Acc. kessaran.

>Why would **-nC (in this case, **-nt) be simplified to *-:n FIRST before
>the *-n>*-r change? This would mean that the simplification of the cons.
>plus neuter *-d/*-t occured well before the cons. plus animate *-s
>changes instead of concurrently! Thus:

>           1.  **-nD > *-:n    (inanimate simplification)

>           2.  **-n  > *-r     (heteroclitic)

>           3.  **-ns > *-:n    (animate simplification)
>               **-rs > *-:r

>Saying that **-nD > *-r and **-n > *-n is much simpler because you have
>this scenario instead:

>           1.  **-Cs > *-:C    (inanimate/animate simplification)
>               **-nD > *-:r

>See? Animate, inanimate AND heteroclitic can be explained in one big
>swoop.

But you seem to imply that inanimate nouns (always?) had an
ending -d/-t, which I cannot agree with.  The *-d is pronominal
only.  We do have cases like Skt. yakrt "liver" (as well as asrk
"blood", with unexplained -k), and generalized Greek -at- < *-nt-
(onoma, onomatos etc.).  But except for the n and n/r stems, all
other neuters have a zero ending [or *-m in the o-stems].  Surely
the neuter nom.acc. was unmarked.  What happened was that there
were neuter stems in -nt (and -nk?), as well as plain -n (and
-r?).  Just like we have m/f stems in -n(s), -nt(s) and -r(s).
In absolute auslaut (because of -0 ending), these developed to -r
(-n, -r) and -:r (-nt, -nk) and tended to merge into a single r/n
heteroclitic paradigm.  One might object that there *are* a
number of neuter n-stems, but most of them end in -m(e)n, where
the preceding nasal consonant may have prevented the regular
development of -n > -r.

>MIGUEL:
>  While some traces of *-t have remained, we have no trace at all
>  of *-k and *-p in PIE.  It is tempting to reason by analogy and
>  hypothesize that if **-t > *-H1, then **-p > *-H3 and **-k >
>  *-H2.  In the case of *k ~ *H2 we have just a few interesting
>  clues, such as Grk. gune:, pl. gunaikes "women".

>I don't find it so tempting. I think there's a very good reason why *-k
>and *-p don't exist in IE. Simply put, words either end with pronominal
>endings of some kind or with a declensional suffix - none of these
>possible suffixes have *-p or *-k and exposed roots are non-existant as
>well.

The neuter nom/acc (and sometimes the loc. of all nouns) had an
exposed root.  Furthermore, isn't it rather odd [not that that
proves anything, but still...] that neither *-k nor *-p (nor *-t,
verbal 3rd.p.sg. secondary and Hittite instr. excepted) occur at
all in grammatical endings, while *-H1 and *-H2 abound?

>If you're saying that IE *-t > CS *-H1 then you have to say that IE *-k
>> CS *-H2 and IE *-p > CS *-H3. This means that we should see Anatolian
>languages with a cornucopia of *-k's and *-p's. Is this what we find?

No.  We must simply assume that the loss of **-p was earlier and
more complete than the loss of **-k, and the loss of **-k was
earlier and more complete than the loss of **-t.

=======================
Miguel Carrasquer Vidal
mcv at wxs.nl
Amsterdam