The three-way "grammatical gender" of nAIE
is just one of many possible noun-classification schemes, and it would be
Euro-parochial to regard its development as something unique. A language may
have no noun classes at all (i.e., no semantic markers determining the
grammatical behaviour of the words that carry them, especially with respect to
agreement rules), but if it has only a binary opposition, it will usually be
"animate/inanimate", "feminine/masculine", "human/non-human", etc. (but
more exotic classifications, e.g. "small/large", or "solid/liquid" are also
possible). The "masculine/feminine/neuter" system is hardly a rarity
cross-linguistically, while not being the only possible three-way
classification.
Needless to say, once the classification
has been grammaticalised, gender assignment becomes to a large extent
conventional -- that is, purely formal rather than semantically based; in other
words, the meaning of a word does not necessarily suffice to determine its
gender. In many modern IE languages (e.g. Russian or German), the masculine
gender includes "men plus a (large) residue", the feminine includes "women plus
a (large) residue", and the neuter is one large residue -- a dustbin for
whatever remains. A word meaning, say, "table", "star" or "tree" can
be placed in any of the three residues. Djirbal (Australia) has the following
genders:
(1) "masculine": men, most birds, fishes,
reptiles and insects, moon, storm, boomerang, ...
(2) "feminine": women, dogs, fireflies,
scorpions, crickets, sun, stars, fire, water, most weapons, ...
(3) "vegetable": fruit and vegetables,
honey, miscellaneous plants, ...
(4) other: body parts, meat, bees, most
trees and vines, grass, sounds, language, ...
The classification has many exceptions, and
it's hard to predict where, say, "rainbow" or "stone" should be classified ((1)
and (4) respectively, as it is). Even the "countable/uncountable" distinction in
English (which is gender-like in that it has grammatical consequences) is partly
arbitrary -- other languages may use the same semantic categories for a slightly
different grammatical classification (as any student of English as a foreign
language knows only too well)
Very large systems of noun classes (15-20+)
are found in Bantu; middle-sized systems are common in some parts of Australia
and the New World. Noun-class systems are often simplified diachronically, but
new ones arise and undergo complication at the same time. Erosion is often
gradual, whereas complication tends to happen abruptly. The nAIE three-way
system has been reduced to "fenimine/masculine" or "animate/inanimate" (a.k.a.
"common/neuter") in many languages. English gave it up altogether a thousand
years ago. On the other hand, some Slavic languages (especially Polish, Slovak
and Upper Sorbian) have recently complicated the inherited system by
introducing subtle hierarchical subclassifications ([masculine, animate,
human] < [masculine, animate] < [masculine]).
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. 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.
Piotr
----- Original Message -----
Sent: Sunday, November 04, 2001 6:29 AM
Subject: Re: [tied] Re: Dating PIE
*****GK: Is the gender reality distinction as between AIE and
NAIE unique in the world's linguistic systems, i.e. are there other language
families which exhibit a similar innovation as to gender awareness in some
members of the family only as does IE?*****