On Sat, 24 Apr 2004 23:19:38 +0000, elmeras2000
<
jer@...> wrote:
>--- In cybalist@yahoogroups.com, Miguel Carrasquer <mcv@...> wrote:
>
>> >Is the analysis involving the numeral 'one' really wrong? Is *te-
>sm-
>> >o:y not the dative of a compound made of *te- + *sm-o-,
>originally
>> >meaning 'this one'?
>>
>> I think so. Hittite does not have this form, but it has
>> pronominal oblique cases with interposed -ed(an)-, which
>> might be linked to Slavic *ed-i:nU "1", Arm. ez "one, only"
>> (PIE *h1edh- "one").
>
>But is it not the second part of jedinU that means 'one'?
Yes (*oihnos), but there is a first part, and that must have
meant something too. That the second part means "1" does not
exclude in the least that the first part had a similar
meaning (cf. English "only one" = Pol. jedyny = Arm. ez).
>> >And has the feminine *te-sy-aH2-ay not in that
>> >case lost an /-m-/ in the clustering (Johannes Schmidt again)?
>>
>> I don't think so. A simpler solution, I think, is that
>> *tosyah2 is in fact the feminine form corresponding to masc.
>> *tosyos, n. *tosyod (the source of gen. *tosyo). Gen.
>> *tyosyah2 acquired an *-s, and *tosyah2i etc. were
>> backformed on that.
>
>But the enthusiasm for the 'one' solution for the Hittite forms
>favours an analysis of the very same kind.
I don't follow.
>And the stem of dat.-
>loc.sg. ke:dani (ka:s 'this') is ke:-, cf. gen. ke:l, abl. ke:z and
>instr. ke:t. The -d- even appears in the plural, dat.-loc.pl.
>ke:das. Is that a good sign?
This is Hittite. If neuter sg. *-od > -at "it" (enclitic)
can come to mean "they" (neuter and masc./fem.), I don't
think there's a big problem with those plurals in -ed-as.
The stem of ke:dani etc. is ke:- (< *k^e-), and it doesn't
make a difference whether we add *-(h1)edh- or *-dh-
(*k^e-edh-an-i also gives ke:dani). That *e- is really part
of the "infix" is shown by the personal pronouns amm-ed-az,
tu-ed-az.
>And does it really invite
>identification with the segment /ed(h)-/ of the other languages?
I think it does. Hittite -ed- appears only in pronouns, and
only in the dative/locative and the ablative/instrumental
cases. In other words, *exactly* in the same places where
we find *-sm- in PIE.
>So, if there is "one" in the dative *tesmo:y, the ablative *tesmV:t
>and the loc. *tesmi, why is it so bad for the feminine?
It's not bad. It's just less well attested: Baltic and
Slavic have *-sm- in the dat/loc/ins masculine/neuter, but
not a trace of it in the feminine. And less obvious
phonetically: it requires an ad-hoc reduction of *-smy- >
*-sy-. And would that be exclusively athematic, when the
masculine "equivalent" [with the exception of loc. *esmi-n,
*tosmin, etc.] is (largely) thematic? And why is there -sy-
in the genitive, where we don't have *-sm- in the masculine?
>In fact we
>expect the "one" part also in the gen. masc.-ntr.. That should make
>us derive *tesyo by reduction from older *tesmesyo. Unfortunately I
>see no way of checking this.
I don't see why we should expect it in the genitive. The
PIE distribution of *-sm-/*-edh- in the "local" cases
[assuming dat. is secondary to loc.] only, not in the
grammatical cases, is very similar to what we see elsewhere,
e.g. in the Basque (indefinite, inanimate) declension:
mendi "mountain"
abs. mendi-0
erg. mendi-k
dat. mendi-ri
gen. mendi-ren
ins. mendi-z or mendi-ta-z
loc. mendi-ta-n
abl. mendi-ta-tik
all. mendi-ta-ra
KO mendi-ta-ko
I'm of course not suggesting Basque -ta- has the same origin
as Hitt. -ed(an)-. Not even the meanings match: -sm-/edh-
looks like a singulative infix, while Basque -ta- is an
indefinite marker (the unmarked local cases Irun-en[*],
mendi-tik, mendi-ra, mendi-ko have gone over to the definite
paradigm).
There is no reason to assume a priori that the genitive
should behave like the local cases, only because it does so
in its Ablaut structure (weak case). That's a historical
accident. In matters of a more semantic nature (like
perhaps the infixation of *-sm-), we can expect the genitive
to pattern with the grammatical cases (nom., acc.).
[*] plain loc. -n is only used for (place-)names. The def.
loc. of plain nouns is -an ( < *-ga-n), with another infix
-ga-, which properly belongs in the local cases of animate
nouns: gizon-aren-ga-n "in the *ga of the man",
gizonaren-gan-dik etc.
=======================
Miguel Carrasquer Vidal
mcv@...