concord, agreement and suppletion

[Rolf Noyer]

In response to Ora's question -- this may be a silly answer -- but, as far
as I understand the term "declension class" the phenomena mentioned in (2)
and (3) below are impossible *by definition.*  The difference between
declension class and gender -- as I understand these terms -- is simply
that declension class features do not propagate syntactically, while gender
features do (or may). In other words, if a declension class property showed
the kind of agreement/concord relations mentioned in (2) and (3) we would
then call it a "gender" property. (Aronoff has made this definitional
difference explicit in various places.)

Did you have in mind some other definition of declension class?

Rolf


>Dear everyone,
>
>Would anyone know if the following three phenomena are possible:
>(1)	Definiteness agreement: the main verb/predicate reflects the
>definiteness of the subject. Things like the Hebrew accusative "et" on
>definite objects and Spanish "a" don't count :)
>
>(2)	Declension class agreement: the declension class of the noun is
>reflected on the main predicate/verb.
>
>(3)	Declension class concord: the declension class of the head noun is
>reflected on the modifying adjectives, the article, the possessor, the
>demonstrative...
>
>NB: In order not to confuse declension class (grouping that determines how
>Case markings sound) with gender (noun class, I would guess it's the
>grouping that determines the pronoun, or?), I would rather ask for data
>from languages that have both (e.g. Latin, Russian - which don't have any
>of the above)
>
>Any other proposals for features that do not/may not trigger agreement or
>concord?
>
>And finally, are Case and definiteness ever involved in suppletion (as far
>as I can recall from Carson's email, possessives do)?
>
>Many thanks in advance,
>
>O
>Ora Matushansky
>
>CNRS - UMR 7023 (Paris 8)
>email: matushan at noos.fr
>page web: http://mapage.noos.fr/matushan/