classifiers and plural marking

[David Gil]

Edith's summary of the discussion so far is really edifying.  In her
conclusion she proposes two conjectures; here are some comments on each of
them:

>       a/ There is no exceptionless implication between the occurrence of
> classifiers in a language and the occurrence of plural marking: languages
> can have either, both, or neither. Nonetheless, most languages do
> not have both; there are some (a few?) that have neither; and most have
> one or the other only.

I think Johanna Nichols was on the right track with her comment on this
issue.  If I may attempt to paraphrase and elaborate on it here:  Both
classifiers and number marking exhibit areal patterning at the global level.
This is an observable empirical fact, which, though this is irrelevant for
the discussion here, suggests that these two features are relatively
"conservative", ie. they tend to persevere and not to change their values
quickly over time.  Now in general, if any two features exhibit global
patterning, there's a pretty good chance that they will end up producing what
appears to be a typologically significant generalization, even though the
generalization may be spurious.

Imagine the following.  Here's the map of the world.  Along comes the
number-marking genie, who gets to fling one big blob of number-marking at the
world.  She hits Africa, Europe, and much of Asia.  Next comes the classifier
genie, who gets to fling one medium sized blob of classifiers at the world --
she hits East Asia.  Is it that surprising (ie statistically significant)
that the two blobs happened not to overlap?  Not at all.  If what we're
counting is blobs, rather than points on the globe (or individual languages),
then what looks like a typological generalization turns out to be the result
of an accident.

Of course this is an overgeneralization: both genies actually did a rather
messy job, with a certain amount of splattering.  But the global areal
patterning is still very much in evidence.  So what's crucial, when examining
proposed typological implications such as these, is not *just* to see that
one's sample is geographically and genetically balanced, but to ascertain
that the features themselves *don't* exhibit the type of areal patterning
which, effectively, reduces a large number of facts to a much smaller number.

>       b/ If a language has (non-generic) nouns with plural
> referents that are not marked for number in any context other
> than numerals, it also has such nouns unmarked for number when they
> occur with numerals. (I.e., there is no language where plural-referent
> nouns are plural-marked when occurring with numerals but not
> plural-marked in other contexts.)

A possible counterexample to this generalization is provided by Singlish
(Colloquial Singapore English) -- though the facts are somewhat difficult to
ascertain (and, unfortunately, my notes are still in Southeast Asia, so I'm
reconstructing from memory here).  In Singlish, in general, number marking is
optional:  _John eat apple_ can mean "John ate the apple" or "John ate the
apples".  But, as I recollect, when a numeral greater than one is present,
there is either a strong tendency, or in fact an obligation, to mark the noun
with the plural -s.

Be that as it may, Edith's generalization b/ is almost certainly valid at
least statistically; and of course it makes obvious functional sense -- to
the extent that languages try to avoid redundancy.

David


--
David Gil

Department of Linguistics
Max Planck Institute for Evolutionary Anthropology
Inselstrasse 22, D-04103 Leipzig, Germany

Telephone: 44-341-9952310
Fax: 44-341-9952119
Email: gil at eva.mpg.de