Gender

[Miguel Carrasquer Vidal]

Jens Elmegaard Rasmussen <jer at cphling.dk> wrote:

>On Thu, 11 Mar 1999, Miguel Carrasquer Vidal wrote:
>> [...]
>> The lack of feminine gender in Hittite (Anatolian) suggests that
>> the PIE three-gender system (masculine, feminine, neuter) is
>> datable to the time between the split-off of Anatolian and the
>> break-up of the rest of IE (beginning with Tocharian).
>> [...]

>I don't think this can be true, because:

>(1) How could the feminine marker *-yeH2-/*-iH2- have been hit by the
>working of ablaut if it only arose after the split-off when the process
>must have been over?

>(2) Adjectives like dankuis 'dark' appear to have the same background as
>Lat. svavis, viz. the feminine in *-w-iH2- of u-stem adjectives (Skt.
>sva:du'-s, fem. sva:d-v-i:').

>(3) The allative of 'one' is sa-ni-ya in the Anitta text, rather obviously
>based on *s(V)m-iH2- (Gk. m!a fem. 'one'), conflated with some form where
>the /m/ was word-final and so changed to /n/ (cf. Gk. ntr. he'n from IE
>*se'm).

>All in all, it looks like a two-bit reduction of the system of three
>genders to two, whereby masc. and fem. formed a "common gender"
>opposed to the surviving neuter just as in Dutch, Danish and Swedish.

Supposing the two examples above are correctly analyzed as
containing *-iH2, it only follows that Anatolian had nouns and
adjectives ending in *-iH2.  As it probably (had) had *-eH2 and
*-uH2.  The question is whether it can be shown that these
suffixes (if suffixes they are) at one stage served to denote
feminine gender in (Pre-)Proto-Anatolian.  Now, if the neuter of
dankuis were danku, you would have a case, but I believe it's
dankui.  We also have *-(e)H2 in Anatolian to mark the neuter
plural, but no trace of it as a feminine marker.

By the same token, I cannot prove that such forms are *not*
relics of what once were feminine markers in Anatolian, instead
of what I believe they are: merely root endings or extensions
which were grammaticalized into feminine markers in non-Anatolian
IE.

There are indications even outside of Anatolian that the feminine
gender is of relatively recent origin.  The lack of formal
marking in many common nouns (Beekes mentions dhugH2te:r,
snusos), the adjectival classes that do not distinguish a
separate feminine form.

The general impression one gets of Hittite is that of an "active
language" in Klimov's terms (see Lehmann, G & I), with a central
role played by the opposition animate ~ inanimate.  The use of
the "ergative" suffix -ant(s) (inanimate <watar> --> animate
<wetenants> when the subject of transitive sentence), the mi-
(active) and hi-/-ha (stative/middle) conjugations.  There is no
sign or trace of feminine gender, no 3rd. person feminine
pronominal forms [tell-tale sign in Dutch, Swedish and Danish
that these languages once had a feminine gender; but of course
also absent in e.g. Armenian], the lack of any formal marker for
feminine nouns (except suffixed -sara) and the lack of feminine
agreement in the adjective.

In view of this, I prefer the more parsimonious explanation that
Hittite maintains the ancient state of affairs (active/inactive
nouns and verbs) and the other languages have developed a
3-gender system out of an earlier animate ~ inanimate one, than
to suppose that Hittite once had a feminine gender, then lost it,
and reverted exactly to the state PIE must have had in the first
place before feminine gender developed.

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