Gender and Noun Class

William Croft wcroft at UNM.EDU
Fri May 10 15:43:29 UTC 2013


In the first nine responses to Don's request, some stated that "noun class" is the generic term, and others that "gender" is the generic term; and some stated that using (but explicitly defining) existing terms is the best route to take, and others that the best route is to create new terms that don't carry the baggage of older terms. So consensus is unlikely. 

The reason for this, of course, is the vexed question of defining grammatical categories both within and across languages. The real problem here is the assumption of the primacy of categories in language description and comparison (the "building block model", as Langacker [1987] calls it). Instead, language description and comparison should be on the basis of the grammatical phenomena that are used to define those categories, in particular constructions, as I (and others) have argued. 

This perspective implies that the categories so defined are language-specific, since the constructions or other grammatical phenomena which define the categories are language-specific. The proposal that Mark refers to, to capitalize language-specific categories, was first made by Bernard Comrie in "Aspect" (1979) as far as I know, and was followed by Joan Bybee in "Morphology" (1985), and several others including myself since then.

This perspective also implies that the terms used for categories should make explicit their basis for definition in a particular grammatical construction or other grammatical criterion, as I suggested in "Radical Construction Grammar" (2001, chapter 1). In this particular situation, Grev's proposal seems most appropriate: "Inflection(al) Class" if the Uduk phenomenon is one of formal morphological declension, or "Agreement Class" if the Uduk phenomenon is one of syntactic agreement/indexation.

Of course, Grev's proposal implicitly assumes "class" rather than "gender" is the better term for this phenomenon (one could easily have "Inflectional Gender" or "Agreement Gender", for example). Or rather than a completely new term. There is a rather ironic reason why I incline towards reusing existing grammatical terms. Many grammatical terms, "gender" for example, are used to refer both to a language-specific grammatical category and to a semantic category (biological sex, for biological referents). This is confusing and problematic in both language description and language comparison. But as typologists since Greenberg have argued, the primary basis for crosslinguistic comparison is semantic/functional. So ironically that confusion in the use of grammatical category terms means that the use of a term such as "gender" or "tense" in a language description is going to allow me to connect the language-specific category to semantically related categories in other languages I am familiar with, even if the language-specific Gender categories deviate from biological sex, and the language-specific Tense categories deviate from time reference. (And of course, typologists know that such deviations often follow regular patterns as well.) So the fact that these grammatical category terms have some semantic connotations, even though I know that they are language-specific (which can be flagged by capitalizing them), makes the language description more accessible, as Matthew notes. Of course, there are cases where there are no traditional grammatical category terms, or there is such a proliferation of terms, that it is advisable to coin a new one (e.g. evidentials in the first case, converbs in the second), as Martin proposes.

In this situation, I am inclined towards (Inflectional, Agreement, etc.) Class rather than Gender because there are some systems whose semantic core appears to originate in animacy (human, animal, inanimate) without sex playing a role. But they appear to be in the minority crosslinguistically, so I also understand why Gender is advocated (also no doubt in part due to European grammatical tradition, thanks to the systems in European languages). But whatever choice is made, the most important thing is to explicitly describe the grammatical constructions/phenomena that led you to posit the grammatical category in the first place.

Bill


On May 10, 2013, at 3:19 AM, Don Killian <donald.killian at HELSINKI.FI> wrote:

> Dear all,
> 
> I have read quite a number of books and articles by this point on this subject, but despite everything I haven't been able to come to a conclusion on something, so I thought to ask the list for suggestions (particularly since some of the authors on the subject are on this list!).
> 
> The difference between gender and noun classes seems to be mostly tradition rather than actual linguistic differentiations (perhaps noun classes are generally viewed to have more categories, but even that isn't absolute), and I've run into a terminology problem with a current grammatical description I'm working on... mainly on what might be a more neutral term incorporating both of these ideas.
> 
> Uduk differentiates all nouns into two categories which are for the most part arbitrary, both phonologically and semantically (in contrast to Corbett's comment: "When we analyse assignment systems of languages from different families we find that genders always have a semantic core.")
> 
> As Uduk is NOT using semantics as the main criteria for differentiation (at least not synchronically), I would like to use a more neutral term than gender or noun class to refer to these categories. Each time I have used gender or noun class, a number of readers have associated biological gender/animacy with the first or Bantu-style noun class systems with the second, and it can often end up detracting from my focus.  I'd rather avoid any sort of general debate on what a noun class/gender system actually is, and instead focus on the actual grammatical system of Uduk.
> 
> Hence my question to the list.. IS there a more neutral term than noun class or gender to refer to grammatical categories of nouns in a language?  Agreement class isn't quite adequate because it also doesn't necessarily refer to this being a nominal property (and noun agreement class is too cumbersome of a term). Nominal category is awkward, although possible.
> 
> I'm open to further suggestions people have.
> 
> Best,
> 
> Don
> 
> 
> -- 
> Don Killian
> Researcher in African Linguistics
> Department of Modern Languages
> PL 24 (Unioninkatu 40)
> FI-00014 University of Helsinki
> +358 (0)44 5016437
> 



More information about the Lingtyp mailing list