double-marked possessives

[stassen]

At 13:30 16-08-99 -0500, you wrote:
>As Leon Stassen noted, Hungarian does indeed possessive
>constructions where both the possessor and the possessum are marked;
>and, furthermore, it is indeed the case that the possessor does not
>always have to be marked. However, whether the possessor is or is
>not marked for the genitive does not depend on whether it is
>definite or indefinite.
>
>Edith M.
>

This is interesting, Edith. If not on definiteness, what does the omission
of the case marking on the possessor depend on, if, indeed, it does depend
on anything structural at all (i.e., if it is not a stylistic and/or
sociolinguistic and/or dialectical thing)? Is it possible to find a rule to
it?

Leon.