pronoun trace

[Arnold Zwicky]

dInIs sez:
 >I recommend that we label deletion of the sort first recommended as
 >grammatical (which at least now Ben, Alice, Arnold and I find
 >ungrammatical) "Presumptive Resumptive Pronoun Deletion."

 >Remember, it's hard to talk about stuff without labels

granted.  but this is not a good label, since the pronoun in
question isn't resumptive.  (nice rhyme, though.)  i don't
have a better proposal at the moment, though.  (in the provacy
of my own mind, i've started calling the construction Gwendolyn,
after a friend.)

 >(and I'm intrigued by Arnold's notion that there is a a sort of
 >superficial analogy going on here which causes speakers to find
 >deletion appropriate here).

i have a (slowly) developing interest in these amalgam/blend
constructions, having spent some time on examples like
  I don't know how many people that there were at the party.
(where the subordinate clause has *both* an initial WH-containing
phrase *and* the subordinator "that", so it has properties of
subordinate interrogative WH clauses and also properties of relative
clauses).

by the way, i'm not discounting the judgments of people who find
the missing-pronoun examples fine, though i think the matter
deserves a lot more investigation.

arnold (zwicky at csli.stanford.edu)