syncretism w/o paradigms

[Jonathan David Bobaljik]

On Martha's remakrs re: acquisition, see Noyer's diss/book, extending
ideas of Andrea Calabrase's - learning as the suppression of
feature-cooccurrence restrictions (filters) = positive evidence for
the acquisition of (contrastive) features.

A clarification:

Meta-syncretisms (as I used the term) do not mean that a contrast is
systematically absent from the language. I (and I think Williams,
implicitly) use the idea only for cases of syncretism in which the
relevant contrasts are indeendently attested in the language.

For example, Russian  Class I nouns (and adjectives and pronouns)
["masculine" and "neuter", except those that decline as Class II -a]
never show a distinct accusative form; it is always syncretic with
nominative (inanimate) or genitive (animate). But the contrast
certainly exists in the language: can't understand the syntax without
the acc-nom contrast, and elsewhere in the morphology (class II
singular nouns). The vocab-insertion based impoverishment appears to
be missing a generalization (though the set of inflections is finite,
hence this is a tricky notion).

Thus, I'm not sure I understand Martha's prediction. If
impoverishment applies before "agreement", the prediction arises, but
if impoverishment arises after "agreement" then the prediction does
not arise. Imagine a language like Russian, but where the
meta-syncretism does not extend to adjectives: ACC=NOM/GEN for some
class of nouns, but the distinction is still marked on the agreeing
adjectives (this arises, in principle, in Russian for nouns that
decline as "masculines" but may take a feminine adjective when
refering to a woman: e.g., % xoros-aja vrac 'good-fem doctor',
corresponding ACC should be: xoros-uju vrac-a: the adjective is
non-syncretic accusative - because it's feminine - but the noun,
being masculine animate, is syncretic with the genitive).

Is there any reason to think that impoverishment (or one's favourite
corresponding device) necessarily applies before agreement (or even
that it can)?

-Jonathan


--
_______________________
Jonathan David Bobaljik
University of Connecticut
Department of Linguistics, Unit 1145
337 Mansfield Road
Storrs, CT 06269-1145
USA

tel: (860) 486-0153
fax: (860) 486-0197

http://bobaljik.uconn.edu/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/dm-list/attachments/20040301/7225cf85/attachment.htm>