|
Abstract:
The aim of this poster is to discuss Frazier and Clifton's
distinction between primary and non-primary relationships, and more
precisely their claim that "an underspecified analysis is assigned
only to nonprimary relations." An eye tracking experiment,
conducted in French, is reported, whose results suggest that in
case of conflict between a primary and a non primary relation, the
decision concerning the primary relation can be delayed. The
sentences were of the following type (simplified English examples)
:
(1) Tired (masculine) of calling the man he left the room.
(2) Tired (feminine) of calling the man she left the room.
The first NP, "the man," can be interpreted either as the direct
object of "calling" or the subject of "left" (competition between
two primary relationships). However, "the man" is also the
potential subject of the adjunct predicate "tired" (depending on
gender agreement). That is, if "tired" and "the man" have the same
gender (first example), the reader will have a tendency to consider
that "the man" is the one who is "tired." As a consequence "the
man" also becomes the likely subject of "left." By contrast, in the
case of gender disagreement, "the man" cannot be the subject of the
adjunct predicate and thus keeps being a potential object for the
preceding verb, in accordance with the usually preferred
interpretation for a noun phrase in such a position.
The analyis of regressions suggested that gender agreement was
checked at the moment when the ambiguous NP was being processed,
and that a decision concerning the adjunct predication was made at
that moment. Moreover, the analysis of first fixation duration at
the diambiguating point suggested that the ambiguous NP was indeed
given the status of potential subject of the following verb in the
gender agreement condition. The decision concerning the non-primary
relationship did not trigger a full commitement in favor of the
subject interpretation however, as indicated the data recorded on
the subject-disambiguated sentences, that the participants were
also presented with. Two types of subject-disambiguated sentences
were compared:
(3) Tired(masc) of calling(optionally trans) the man left the
room.
(4) Tired(masc) of chatting(intransitive) the man left the
room.
First fixation durations and first-pass gazes at the
disambiguating point were shorter in the intransitive condition
(4), suggesting that the status of the ambiguous NP was left
undecided until the disambiguating region was reached in the
optionally transitive condition.
|
