On Thu, 29 May 2003 04:26:42 +0200, Miguel Carrasquer <
mcv@...>
wrote:
>On Thu, 29 May 2003 03:52:51 +0200, Miguel Carrasquer <mcv@...>
>wrote:
>
>>It seems to me that verbal -mmek and -ttek can be explained in a
>>similar manner: there were separate forms in Proto-Baltic-Finnic (PBF)
>>for transitive verbs with singular, dual and plural object, something
>>like:
>>
>> sg du/pl
>>1. -m > -n -k/t-m > -nn
>>2. -t -k/t-t > -tt
>>1. -mek -k/t-mek > -mmek
>>2. -tek -k/t-tek > -ttek
>>
>>This is the way Samoyed and Ob-Ugric work, so it would make sense if
>>PBF had the same thing.
>>
>>When the distinction between sg/du/pl. objects in the verb was given
>>up (as well as that between transitive and intransitive forms) the two
>>(three, if the intransitive was different) series merged as -n, -t/-d;
>>-mmek, -ttek.
>
>OK, I see the problem. If the Uralic verb agrees with the _object_ in
>number (as it certainly does in Samoyed and Ob-Ugric, and in my
>analysis of PBF), the same as in Eskimoan, then the given analysis of
>the accusative *-m can't be right: "the man is the dog's killer" ~
>"the men are the dog's killers" would imply number agreement with the
>_subject_.
>
>We can get it to work if the analyze : "the killing of the dog by the
>man" ~ "the killings of the dogs by the man".
Given that the subject is expressed as a possessive in Samoyed and
Ugric, that should rather be: "the man's killing of the dog" ~ "the
man's killings of the dogs"
>This requires a marking
>on the subject as well as on the object, which is OK in IE (*-m, *-s),
>but not in Uralic, where the subject is not marked (of course the
>marking, being redundant, may have been lost there).
And this does not exhaust the logical possibilities.
There is the cautious linguist's possibility that we have a PIE acc.
*-m, a Uralic acc. *-m, an Eskimoan gen./erg. *-m and a Kartvelian
ergative *-m(a), but that they need not be related at all. After all,
grammatical markers are short, and they tend to use the most unmarked
phonemes around, such as /m/.
If we do accept the relatedness of PIE, Uralic, Eskimoan and
Kartvelian, there is still the possibility of several cycles of
reanalysis having taken place. Languages can go back and forth
between ergative and accusative systems using passive and antipassive
constructions, stative verbs can reintroduce possessive and genitival
constructions as subjects and objects, the (partitive) genitive can
always replace the accusative, etc. For instance, the Eskimoan,
Uralic and Indo-European systems may all have emerged from a common
pattern (e.g. a stative construction like "the man's killing is the
dog"), with an unknown genitive marker on what became the subject,
which was subsequently eroded away in all the languages. Eskimoan
then may have simply renewed the construction using its productive
genitive marker *-m, while IE and Uralic switched to a partitive
model, with the object now marked by genitive *-m (cf. the Slavic and
Hungarian accusatives where this has happened [again]).
Another interesting possibilty is that *-m was not the genitive marker
at all. The only proof we have of that is Eskimo, where *-m/*-p is
both genitive and ergative. But Eskimo lacks the genitive in *-n, and
I can't think of any other grammatical morphemes in Eskimo that must
be derived with final *-n. If so, it's possible that *-n in the
absolute Auslaut became *-m in Eskimoan, and that the genitive in *-m
is etymologically identical to the Uralic genitive in *-n. The IE and
Uralic acc. *-m (as well as the Kartvelian erg. *-m(a) and perhaps
even the Eskimoan ergative -m) must then have a different origin. Now
in constructions such as "the man's killing (is) the dog" or "the man
(is) the dog's killer", the copula has been hidden away behind
parentheses, but what if it was overtly expressed? What if *-m(a) is
in origin an enclitic copula? In Japanese, -wa is used as a topic
marker. In Akkadian, the topic/predicate marker is -ma. In Sumerian
the verb "to be" is <me>, and the enclitic copula is -(a)m. Such a
copula could easily have been reinterpreted as a subject marker if it
directly followed the subject (S-Aux-O), or as an object marker if it
followed the object (S-O-Aux).
=======================
Miguel Carrasquer Vidal
mcv@...