|
Abstract:
Across languages it is a general fact that the likelihood of
observing conceptual agreement, as opposed to purely syntactic
agreement, increases along the following hierarchy (Comrie, 1978;
Corbett, 1985): modifier-noun agreement, subject-verb agreement;
subject-predicative adjective agreement; subject-pronoun agreement.
In moving along this hierarchy, two factors covary: (a) the
syntactic relationship between the two agreement elements; (b) the
processing distance between the two elements. In a series of two
experiments (one in French and one in Italian) we assess the
contribution of processing distance to an increase of conceptual
agreement. In the experiments, the agreement relationship is kept
constant (agreement in gender between the subject and a predicative
adjective) but the distance between the two agreeing elements is
manipulated, the question is whether the manipulation affects the
likelihood of observing conceptual instead of syntactic agreement.
In both languages, we selected nouns (to be used as sentential
subjects) that have a fixed grammatical gender (masculine or
feminine) despite referring to both male and female entities (e.g.,
in Italian: "vittima" [victim, fem]; "soggetto" [subject-masc]).
Predicative adjectives have to agree with the grammatical gender of
the subject noun; conceptual agreement with the sex of the referent
is considered an error by native speakers. However,
subject-anaphoric pronoun agreement may be either syntactic or
conceptual.
The experiments use the sentence completion paradigm introduced
by Bock and Miller (1991), in which speakers are asked to complete
sentence fragments composed of a subject NP and a modifying NP. The
crucial manipulations are: (a) the sex of a discourse referent,
introduced by a context sentence, which matched or mismatched with
the gender of the subject noun; (b) the length of the modifying PP,
which was either short (det+N) or long (det+Adj+N, in Italian; and
det+2 Adjs+ N) in French. Adjectives were either pre or
post-nominal. First, participants were presented with a sentential
context, which they were asked to read silently (examples are in
Italian):
Un camion ha investito Fabio/Fabiola che correva in bicicletta
ascoltando musica
[A truck hit Fabio/Fabiola who was riding the bike while listening
to music]
Then they were presented with a sentence fragment to
complete:
La vittima dello scontro (terribile)
[The-fem victim-fem of-the-masc (terrible) crash-masc]
In the example, the noun vittima, refers to a man (Fabio) or to
a woman (Fabiola). Preliminary data are reported in Table 1;
agreement errors were more common when the sex of the referent
mismatched the gender of the noun, then when the two matched, and
errors were more common for long preambles than for short ones.
However, crucially, the interaction between the two factors was not
significant, suggesting that the likelihood of finding conceptual
agreement instead of syntactic agreement does not depend on
processing distance.
Table 1
A. French (long preambles are generated by adding two adjectives
to the modifying PP)
Short
Long
Context Congruent 6.5 8.1
Context Incongruent 21.7 36.1
B. Italian (long preambles are generated by adding one adjective
to the modifying PP)
Context Congruent 2.9 6.1
Context Incongruent 12.5 23.9
Bock, J. K., & Miller, C. A. (1991). Broken agreement.
Cognitive Psychology, 23, 35-43.
Comrie, B. (1975). Polite plurals and predicate agreement.
Language 51: 406-418.
Corbett, G. G. (1983). Hierarchies, targets and controllers:
Agreement patterns in Slavic. University Park, PA: The Pennsylvania
State University Press.
|
