classifiers and ligatures

[David Gil]

I am in complete agreement with Alan that classifiers in Chinese and
other languages can be viewed as a particular case of ligatures /
linkers.  But I think it is important to point out that not all
ligatures are classifiers.  This is true, trivially so, for the
"prototypical" examples of ligatures, from Austronesian languages such
as Tagalog.  But it is also true for Chinese languages.

There is a very easy semantic diagnostic to distinguish classifiers from
non-classifier ligatures in Chinese languages (and elsewhere).
Classifiers are *inherently singular*; in the absence of an overt plural
quantifier (such as a numeral), the interpretation of an NP containiong
a classifier is invariably singular.  Hence facts such as the following:

CANTONESE

(a) ping11-guo35
    apple
    [SG or PL]

(b) go33 ping11-guo35
    CLF apple
    [SG]

(c) a33-faay55 go33 ping11-guo35
    Ah Fai CLF apple
    [SG]

(d) a33-faay55 ge33 ping11-guo35
    Ah Fai LIG apple
    [SG or PL]

HOKKIEN

(a) pheng22-ko53
    apple
    [SG or PL]

(b)*liap21 pheng22-ko53
    CLF apple


(c) a44-beng24 liap21 pheng22-ko53
    Ah Beng CLF apple
    [SG]

(d) a44-beng24 e22 pheng22-ko53
    Ah Beng LIG apple
    [SG or PL]

(All the data are from speakers of the Singapore dialects, but I think
they hold generally.)  In the above, the NPs in (a) show that in
Cantonese and Hokkien, bare nouns have general number, ie. are unmarked
for singular / plural.  The NPs in (b) and in (c) show that classifiers
force singular intepretations.  And the NPs in (d) show that these
languages also have a ligature that doesn't force singular
interpretations.  (It is the possible plural interpretation of the NP in
(d) in Hokkien which is what made me take issue with the claim that
[e24] is a classifier in Hokkien.)

Thus, classifiers have inherently singular semantics: their meaning is
"one unit of a particular type".  This is true of all the languages that
I've had occasion to examine.

With one very puzzling exception.  In Shanghainese, constructions such
as those in (b) and (c) above can, for some but not all speakers, be
understood as either singular or plural.  A few years ago, there was
some discussion of this on the linguist list -- but I was never able to
figure it all out.  Maybe in Shanghainese all the classifiers are
getting reanalyzed as ligatures.  Any ideas, anyone?


--
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