agreement

[Matthew S Dryer]

Edith's response to my query does not address what I see as the central
issue.  Consider the case of a language where the antipassive verb form is
used if and only if the "object" is indefinite and the basic
non-antipassive form is used if and only if the "object" is definite.  One
can infer from the antipassive form of the verb that the "object" is
indefinite and from the basic form that the "object" is definite.  Assume
further than in basic clauses, the verb agrees with the definite objects
for various features like person and number.  From Edith's
characterization in her earlier message, it would follow that the basic
vs. antipassive morphology is coding agreement in definiteness, which is
counterintuitive.  My question is whether there is some argument that
Hungarian is different from this.

Matthew