| |
Abstract:
According to centering theory (Grosz, Joshi & Weinstein,
1983, 1995), a pronoun is read faster than a repeated name when the
anaphor refers to the focused entity in the preceding sentence,
that is, to the subject and/or first mentioned entity (Gordon,
Grosz & Gilliom, 1993). According to grammatical parallelism
(Stevenson, Nelson & Stenning, 1995), pronoun resolution is
facilitated when the pronoun occupies the same grammatical role as
its antecedent. Recently, Chambers and Smyth (1998) showed that the
repeated name penalty also occurs for non-subject pronouns, as long
as the anaphor and antecedent are grammatically parallel and
contained in structurally congruent sentences. This result poses
problems for centering theory, which does not predict a repeated
name penalty for nonsubject anaphors. However, research on both
centering theory and parallelism has focused exclusively on animate
entities. The present study redressed this imbalance by testing for
the presence of a repeated noun phrase penalty with subject and
nonsubject anaphors that referred to inanimate entities.
Participants were presented with short texts in a self-paced
reading time task. The first sentence introduced the two
antecedents, the second contained a subject and a nonsubject
anaphor. The materials varied according to four factors:
Parallelism (the anaphors and their antecedents filled either
parallel or non-parallel grammatical roles); Type of Subject
Anaphor (pronoun or repeated noun phrase); Type of Nonsubject
Anaphor (pronoun or repeated noun phrase); and Syntactic Structure
of the sentence introducing the antecedents (active or passive).
For example, The diamond is normally stored in the vault. The vault
is double locked to protect the diamond. is nonparallel, has a both
subject and nonsubject noun phrase anaphors, and uses a passive
sentence to introduce the antecedents.
When the antecedents were introduced in the active voice, there
was a clear effect of parallelism. Both subject and non-subject
anaphors showing repeated noun phrase penalties in the parallel but
not the non-parallel condition. When the antecedents were
introduced in the passive voice, there was no effect of
parallelism. Instead there was a repeated noun phrase penalty in
both parallel and non-parallel conditions whenever the anaphor
referred to the subject antecedent. The parallelism in the active
condition replicates Chambers & Smyth's result and suggests
that both subject and nonsubject contribute to coherence in active
sentences, a view more consistent with Sidner's (1979) model of
focusing than with centering theory. The repeated noun phrase
penalty for subject antecedents in the passive condition is most
likely because passive sentences topicalise the surface subject,
making it more accessible to both subject and non-subject pronouns.
We conclude from these results that parallelism may be over-ridden
by focusing when there is a highly topicalised antecedent.
Chambers, C.G. & Smyth, R. (1998). Structural parallelism
and discourse coherence: A test of centering theory. Journal of
Memory and Language, 39, 593-608.
Gordon, P.C., Grosz, B.J. & Gilliom, L.A. (1993). Pronouns,
names and the centering of attention in discourse. Cognitive
Science, 17, 311-348.
Grosz, B.J., Joshi, A.K., & Weinstein, S. (1983). Providing a
unified account of definite noun phrases in discourse. Proceedings
of the 21st Annual Meeting of the Association for Computational
Linguistics. Association for Computational Linguistics.
Grosz, B.J., Joshi, A.K., & Weinstein, S. (1995). Centering: A
framework for modeling the local coherence of discourse.
Computational Linguistics, 21, 203-225.
Sidner, C.L. (1979). Towards a Computational Theory of Definite
Anaphora in English Discourse. PhD. Thesis, MIT. Stevenson, R.J.,
Nelson, A.W.R. & Stenning, K. (1995) The role of parallelism in
strategies of pronoun comprehension. Language and Speech, 38,
393-418.
|
