Date: Sun, 20 Oct 96 17:48:18 EDT
Message-Id: <9610202148.AA10150@JESPERSEN.CS.NYU.EDU>
From: Ruth Reeves <reevesr@jespersen.cs.nyu.edu>
To: linconf@tamvm1.tamu.edu
Subject: Comment on Zlatic's comment
To Larisa Zlatic;
In response to Seely's question about the status of indexing labeled
"?*" in (4) and "?", "?*" in (9), you proposed a "pragmatic/processing
explanation."
(4) Jovan (i) je procitao [Marijin (j) clanak o sebi (i, ?*j)].
John AUX read Mary's-ADJ article about self
'John read Mary's article about herself.'
(9) Jovan (i) je procitao [Marijin (j) clanak o njemu (?*i) / njoj (?j)].
John AUX read Mary's-ADJ article about him / her
'John read Mary's article about him/herself.'
>From your response this afternoon to that question, i take it that Seely
miunderstood the idea. Here's how I'm taking it: Its not the ambiguity
between argument vs non-argument of Marijin itself which would cause a
processing or pragmatic failure, but rather that if the NP(j) is interpreted
as an arguement in (4), say, then the indexing given becomes somewhat
acceptable -- or perhaps the rather different claim -- it becomes acceptable
to a small set of speakers on that interpretation
this being also what would allow njemu to be co-indexed to Jovan in (9):
an intervening subject allows it.
I'm guessing that the ambiguity you are talking about turns on something
like whether Marijin is the possessor of the article or its author;
the first being interpreted as a non-argument and the second as an argument.
two questions:
If my interpretation of your idea is right,
In (9) the "?" on njoj --co-indexed to Marijin-- indicate that its ok when M.
is interpreted as the posssessor (non-argument) but crucially NOT when M. is
interpreted as the author (argument). True?
My second comment is a general worry about how to cash out this kind of idea.
The claim seems to be that people waver in their judgements where its not
clear which meaning to give the nounphrase in SPEC-NP
I wonder about the possibility that this is a pragmatic effect. after all,
the speaker can't have meant to utter an ambiguity and if the hearer knows
which sense of possession is meant, then it won't count as ambiguous in any
pragmatic sense.
So suppose the effects are processing ambiguity effects, where >1
possible assignments to the position cause momentary parse failure
upon reaching the argument co-indexed where both possibilities are still
"active".
This seems good until we we return to the explanation which motivates it. It
was selecting a PARTICULAR interpretation in the first place which allows the
given indexation. Perhaps, in my attempt to understand your position, i have
misrepresented it. i would be happy to learn that this is so because i am
continually bothered by data involving more than one set of intuitions
in a single grammar which seem to push each other back and forth.
Ruth Reeves