>Yes, one can argue that having the relative in the case of the antecedent
>makes it easier to locate the antecedent. However, when case attraction
takes
>place where the relative immediately follows the antecedent then this
>argument seems to be undermined, because there is no difficulty in
>identifying the antecedent.
>
Touche/. Excellent point.
>Also the observed inconsistency of this phenomena undermines this argument.
If
>the reader can not count on the case of the relative being the same as the
>antecedent, then this explanation of "attraction" is also cast into
question.
>What good is the attracted case of the relative for identifying the
antecedent
>if it only occurs sporadically?
Well, my contention was that this was probably an unconscious occurrence, not
intentional, which would account for its sporadicity (?). But as a
hypothesis, it is probably pretty weak. Your hypothesis is quite interesting.
I don't have the tools for isolating occurences of "attraction" so I'm afraid
I can't really test it. I hope somebody who can, will.
God Bless,
Rev. Craig R. Harmon.