gender

[Larry Trask]

On Mon, 8 Mar 1999, roslyn frank wrote:

> And finally, has anyone contemplated the possibility that there
> might have been an even earlier stage that needs to be reconstructed
> (e.g.  perhaps in the case of Euskera) that eventually gave rise to
> a animate/inanimate dichotomy (e.g., as it is found today in
> Euskera)? Any ideas on that, Larry? This would imply that
> cognitively speaking, there could have been an earlier structure
> that was not based on an "animate/inanimate" contrast but on another
> ontological type or definition of "being."

Basque generally lacks gender and noun-classes, but there is a wrinkle:
animate NPs typically construct their local case-forms differently from
inanimate NPs.  Inanimate NPs just take the case-endings.  But animate
NPs use an extra morph which is, or probably once was, a postposition.
The facts vary according to region.

In the west, the additional morph is <-gan>, which appears to be a
fossilized postposition.  This is usually attached to the genitive case
of the animate NP, categorically so in some circumstances, but the rules
differ from variety to variety.  In the east, the equivalent item is
usually <baita>, which is variously attached to the genitive or to the
absolutive of the animate NP.  (But <-gan> is also used to varying
extents in eastern varieties.)  In both cases, the local case-suffix is
added to this extra morph.  A further difference is that the use of
<-gan> is generally categorical with animate NPs in the west, while the
use of <baita> is not categorical with animate NPs in the east: in the
east, we sometimes find the local case-suffixes attached directly to
animate NPs, though the rules governing this usage are obscure.

The origin of <-gan> is unknown, though it may possibly be a reduced
variant of <gain> `top', which itself serves to form postpositions in
the language.  The origin of <baita> is likewise unknown, though this in
its locative form <baitan> is frequently used in the older literature to
mean specifically `at the house of', as opposed to merely `in, on, at'
-- something which is not true of <-gan>, so far as I know.  This has
induced speculation that <baita> may once have been a word for `house',
but there exists no evidence to support such an interpretation.

The majority view is that both elements are native in Basque and of some
antiquity.  We may therefore surmise that the special treatment of
animate NPs in the local cases is also of some antiquity, but we can't
say how much, for lack of data.

It is perhaps curious that two different formations exist in the same
function, but maybe we are just seeing the remnants of an earlier, and
unrecorded, state of affairs in which the language offered several
resources for the purpose, and some degree of selection has taken place
-- a common kind of historical development in languages.

As to what happened in Basque before the 16th century, it's impossible
to say.  But I have two observations.  First, animate NPs are not
distinguished from inanimate NPs in any other way: in particular, all
the non-local cases are formed identically for all NPs.  Second, it
seems clear to me that the local cases of the modern language -- with
the likely exception of locative <-n> -- are of recent formation.  We
find substantial variation in the formation of the local cases, both in
time and in space.  So, if the modern local case-endings are in general
not very old, it hardly seems likely that the distinctive formations
used with animate NPs can be very old, either.  That leaves open the
question of how long Basque has been distinguishing animate and
inanimate NPs at all.

Larry Trask
COGS
University of Sussex
Brighton BN1 9QH
UK

larryt at cogs.susx.ac.uk