The conflation of m. and f. is of course
possible, but as opposed to Scandinavian, Anatolian preserved the PIE
word-structure quite well. It did not reduce inflectional endings, for example.
The only word classes in which the three non-Anatolian genders are
systematically contrasted are the "thematic" nouns and adjectives
(represented here, for the sake of clarity, by the familiar nom.sg.
endings *-os, *-a:, and *-om). What Anatolian shows is _not_ a merger of the
historical *-os and *-a: classes (this could have happened most naturally
through the assignment of Hittite endings of the -as [< *-os]
declension to old *-a: feminines) but the complete absence of anything that
might be connected to the *-a: type.
Danish and Swedish have not struck out
Germanic feminines from their vocabularies; they have only eliminated their
specific gender markers, which were residual anyway. Many historical *-a:
feminines have Danish and Swedish cognates; it's only their distinctive gender
that has been lost. By contrast, there is no systematic correspondence between
non-Anatolian *-ah2 stems and Hittite -as stems. Only two possible examples have
ever been proposed, as far as I know: Hitt. hissa- 'shaft' : Skt. i:s.a:- (but
<hiss(a)-> is not attested in the nominative, so it isn't really certain
what declension it belongs to, and even if it were the noun could have been
independently thematised in both groups, since other cognates suggest that the
original stem was consonantal), and hassa- 'hearth' : Lat. a:ra 'altar' (but
there is evidence that Hittite <hassa-> was actually a nasal stem). Very
few (maybe half a dozen) non-Anatolian *-a: stems are in any way related to a
Hittite substantive, whether thematic or not.
In other words, there is no
evidence of a putative feminine declension having been absorbed
into any of the Anatolian declensional types. The most reasonable
interpretation of these (and a number of other similar) facts is that the *-ah2
formation arose after the separation of Anatolian.
Piotr
----- Original Message -----
Sent: Monday, November 05, 2001 5:25 PM
Subject: Re: [tied] Gender (Was: Dating PIE)
The probability of Anatolian having reduced a three-gender
system
into a animate_inanimate system:
When assessing the
probability for this having taken place, in may be
useful to see if there
exist examples of this in living languages.
Swedish and Danish are
examples of modern languages with an animate-
inanimate classification
system. At the same time Norwegian that is
so close to Swedish and Danish
that it is nearly the same language,
still has the old three-gender
system.
My native language is Norwegian. When waching what has occured
nearly
in my own backyard in the last hundred years, I am sure the same
could have taken place when Anatolian broke away from the common
indoeuropean stock.
Of course that does not proove that this
has taken place in
Anatolian, it only shows that such a transformation
actually could
have taken place.