references on ineffability as conflicting forms

[Carson Schutze]

Hi everyone,

This is a citation question, not intended to open an issue for debate
(though of course if people feel like debating it that could always be fun).

Some DMers have claimed in talks that one way (perhaps the only way?) in
which a generally non-filtering morphological component can induce
ineffability is by an unresolvable conflict between output forms. A classic
case of what is meant by this can be seen in German free relatives, as
discussed by Taraldsen a couple of decades ago, and analyzed by Uli
Sauerland in an MIT ms. in the mid-90s.

Ich zerstoere *wer/*wen/was mich aergert.
I   destroy   *who-nom/*who-acc/what-nom~acc  me upsets

That is, the matrix clause wants the wh-word to be ACC while the embedded
clause wants it to be NOM, a conflict that is intolerable unless the
relevant NOM and ACC forms happen to be homophonous.

So, my literature questions are

1) is there any DM work that has claimed in print that this is the only kind
of ineffability there is?

2) are there any known cases where this kind of pattern has been claimed to
arise as the result of a verb having to agree with two different NPs?

3) is there any literature that ought to be cited offering plausible
arguments *against* the spirit of this analysis for cases of the German type
above or something similar, i.e. that a conflict between two surface forms
is not at the heart of the problem? (Uli's actual account was not quite as I
described it above, rather he had two separate NP positions and a rule that
could delete one of the wh-phrases just in case it was phonologically
identical to the other; I take that as in the same spirit as the above.)

Thanks!

    Carson