Editor for this issue: <>
cantva.canterbury.ac.nz>Some time ago I posted a query about gender systems. Specifically I was interested in finding out about gender systems which are relatively *large* (4 or more genders) and also *covert* rather than overt. I offered no precise criteria for the covert/overt distinction but suggested that one should err on the inclusive side, so that e.g. Latin would count as covert. The reason for the inquiry is that I suspect that covert gender systems (appropriately defined) are subject to some sort of 'economy' constraint, reminiscent of though not identical to 'paradigm economy' as propounded in my _Allomorphy in Inflexion_ (1987), whereas covert systems are not. In the terms of Corbett _Gender_ (1991), this would mean that, given a certain number of target genders, the number of controller genders which they can be organized into is subject to certain constraints if the gender of controllers is covert, but escapes these constraints if the gender of controllers is overt. I got replies from Bob Beard, Ellen Contini-Morava, Larry Hutchinson, John Koontz and Bob Port. Bob Beard predicted that my findings would support his distinction between natural gender (a lexical category), grammatical gender (an inflectional category) and 'agreement' (another inflectional category, distinct from grammatical gender). On that I cannot say yes or no as yet. Ellen and Bob Port pointed out that gender in Swahili is not so uniformly alliterative as introductory textbooks sometimes claim. This is important in the sense that, if (part of) the Swahili gender is system is (or is becoming) covert in the relevant sense, I must predict that the freedom with which individual nouns can 'choose' different singular-plural pairings of target genders (Bantuists' 'classes') should become restricted. This has yet to be investigated in detail. John mentioned the Dhegiha group of Mississippi Valley Siouan languages. These have animate and inanimate genders, each subdivided into four covert 'genders' relating to movement and posture. But these do not seem to be genders in the required sense, since they are not lexically fixed; for example, the verb accompanying a noun meaning 'tent' can be conjugated according to the 'erect' or the 'supine' pattern, depending on whether the tent is erected or not. Larry mentioned that the Sierra Leone language Temne has 28 genders, but I have no more details. All in all, the inquiry produced no clear counterevidence to my tentative hypothesis, so I will press on. A progress report will be presented at the Typology and Parameters Workshop at the international Conference on Historical Linguistics at UCLA in August. Andrew Andrew Carstairs-McCarthy Department of Linguistics, University of Canterbury, Private Bag 4800, Christchurch, New Zealand Phone +64-3-364 2211; home phone +64-3-355 5108 Fax +64-3-364 2065Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue