Re: [tied] Gender (Was: Dating PIE)

From: Piotr Gasiorowski
Message: 10980
Date: 2001-11-04

 
----- Original Message -----
From: george knysh
To: cybalist@yahoogroups.com
Sent: Sunday, November 04, 2001 4:58 PM
Subject: Re: [tied] Gender (Was: Dating PIE)


> [GK:] ... I am not asking whether the "3-way" classification is unique, I am asking whether in other language families which have it there is a similar distinction between some member languages which do and some which don't as exists in IE between AIE and NAIE.
On the basis of what Piotr had written so far I would have guessed that the answer is "No". If I have guessed wrong let me know.
 
The typology of noun classes is extremely rich and for that reason it may be difficult to come up with an example that exactly parallels the IE situation. If I find such a typological precedent, I'll let you know. But it is not unusual at all to find a variety of noun-class systems in one and the same family.
 
>> [PG:] What militates against reduction in Proto-Anatolian as opposed to complication in Proto-non-Anatolian-IE is the fact (already mentioned by Miguel) that an older binary class system is implied by internal reconstruction within nAIE.

> [GK:] Hmm. Nothing more obvious?
 
This is obvious enough -- precisely what should be expected if the ternary gender classification is a nAIE innovation. Comparative reconstruction points to the existence of three genders in Proto-nAIE, while internal reconstruction reveals relicts of an earlier two-class system (see below). Note also that, by contrast, the Anatolian languages display no "fossil" evidence of ever having had a m./f. gender distinction in their prehistory. It seems, in particular, that the familiar *-ah2- (or *-ih2-) femininity marker, so productive in all nAIE branches, doesn't correspond to _anything_ in Anatolian.

>> [PG:] Linguists of a century ago were already well aware of that -- fossil traces of the old system are quite clearly visible even in Greek and Latin. What we see in Hittite corresponds very closely to what had been independently reconstructed before Hittite was discovered.

> [GK:] If I understand this correctly, the claim is that these "fossil traces" indicate that NAIE, without the M-F "innovation" would have developed like AIE? Viz., an "animate/inanimate" binary opposition.
 
Well, without the m./f. split the system would have _continued_ like Anatolian, with an archaic binary opposition.
 
> [GK:] What are some examples of these traces? When you say "even in Greek and Latin" do you mean that this fossilized evidence exists in all NAIE languages?
 
Generally, the deeper you dig, the more traces can be unearthed. Not surprisingly, late developments (like the loss of noun gender in English or the independent loss of the neuter in several groups) have often obliterated the archaisms in question. However, some traces survive in virtually all branches. Latin, Old English, etc. show relict traits lost in the modern Romance languages or in modern English. The most important facts are the following:
 
The masculine and feminine genders are not formally distinguished e.g. by having different inflectional endings (cf. *bHrah2te:r : *mah2te:r) except in the _least_ archaic declensional classes. At the same time they are jointly differentiated from the neuter. This means that, in general, the m. or f. gender of consonantal, *-i- and *-u-stem animate nouns cannot be deduced from the way they are declined (if you don't happen to remember whether Latin nouns like <pe:s>, <ops>, <li:mes> or <piscis> are m. or f., you have to look them up in a dictionary).
 
Even some "thematic" adjectives may lack the m./f. distinction (*-o-s vs. *-ah2 -- thus often in classical Greek), and athematic ones (in *-i-s or *-u-s) lack it as a rule (cf. Lat. gravis [m., f.] : grave [n.]).
 
There is some ambiguity as regards the correlation between m./f. assignment and morphological markers in nouns: there are quite a mumber masculine *-ah2-stems (plenty of them in Slavic, for example), and there are some feminines in *-o-s. There is no such crossover involving thematic neuters. It seems that the formal separation of the two non-neuter genders has never become complete even within the major declensions.
 
There is more evidence, all pointing in the same direction. Hittite declensions look just like what must be reconstructed as the most archaic stratum of nAIE. It is virtually unthinkable that a language should have lost its most productive paradigms and preserve only minor and irregular ones -- and that's what would have to be proposed for Anatolian under the "simplification" scenario. Apart from Holger Pedersen, who rejected everything that might point to the special position of Anatolian and so _had to_ insist on the antiquity of the three-way system, most linguists (at least since Meillet) have accepted the secondary character of the m./f. distinction.
 
> [GK:] A follow up: would this evidence show that this is unique to PIE or would it have shared this with other Nostratic languages?
 
Ask the Nostraticists; I am not one :). There are several different noun-class systems in languages regarded as Nostratic (including total absence of noun classes). Even if Proto-Nostratic really existed, there has been enough time for its noun-class system (if any) to have changed beyond recognition.
 
Piotr