| |
Abstract:
Doubly quantified (DQ) sentences (sentences such as "Some
circle hits every square") are widely accepted to be ambiguous (May
1993, Kurtzman & MacDonald 1993, Greene 1992, Johnson-Laird et
al 1989). Whether logically naive subjects can perceive the
alternate readings of DQ sentences presented in isolation (without
context) remains unclear. Preceding context has a demonstrated
effect (Johnson-Laird 1989, Greene 1992). Not clearly demonstrated
in the literature is what effect succeeding sentences or queries
will have on the interpretation of a particular sentence, although
an effect is likely: the subject's choice of interpretation could
be biased toward one interpretation over another, especially if the
various interpretations are available in parallel. Also not clear
are whether there are subject variables that influence the choice
of one interprestation over another.
We devised a paradigm which allows subjects to respond
intuitively to the meanings of DQ sentences without compelling bias
in favor of one interpretation over another. In this paradigm, we
requested subjects to graphically draw their interpretations of
various DQ sentences. Since subjects were given no context from
which to judge each sentence, and were allowed to draw only one
interpretation, they could only respond with the first
interpretation that came to mind.
Our data show that multiple interpretations for each sentence
are available, with, not surprisingly, the most widely chosen
interpretation for any given sentence being the subject wide-scope
reading. However, it was by no means universal. The cross-subject
frequency of this choice was affected by the syntactic structure of
each sentence (e.g., where passive structures tend to defocus one
quantifier), and by the specific pairs of quantifiers chosen. More
importantly, certain subject variables tended to influence the
choice of one interpretation over another. In interpreting active
sentences, males and females tended to choose object-wide readings
with equal frequency. However, with passive sentences, males chose
more frequently object-wide interpretations. This was most
pronounced in left-handed males, with a less pronounced, although
still significant, effect with familially lefthanded (FS+) males.
Similar interactions between handedness, FS and gender are
documented in the literature (Bever et al 1990), with the greatest
effects noted at the word level, not at the structural level, as
our data show.
Our data point to possible genetic factors influencing
structural interpretations of sentences. If one subscribes to a
multi-pass strategy for parsing, as described in (Bever et al
1998), it is apparent that FS- subjects are more likely to accept
the surface interpretation available from the first-pass, with FS+
subjects more likely to engage in more extensive structural
manipulations available in the second and subsequent
syntacic/semantic passes (possibly in parallel). Although the
genetic locus of FS is still indeterminate, its effects on language
processing areas are strongly suggested by our data.
Bever, T.G., Carrithers, C. Cowart, W., Townsend, D.J. (1990).
Language processing and familial handedness. In Galaburda, A.M.
(ed), From reading to neurons. Cambridge, MA: The MIT Press.
Bever, T.G., Sanz, M., Townsend, D.J. (1998). The emperor's
psycholinguistics. Journal of Psycholinguistic Research, v. 27 n.
2., pp. 261-284.
Greene, S. B. (1992). Multiple Explanations for Multiply
Quantified Sentences: Are Multiple Models Necessary? Psychological
Review, v. 99, n. 1, 184-187.
Johnson-Laird, P. N., Byrne, R. M. J., & Tabossi, P. (1989).
Reasoning by Model: The Case of Multiple Quantification.
Psychological Review, v. 96, n. 4, 658-673.
Kurtzman, H. S. & MacDonald, M. C. Resolution of Quantifier
Scope Ambiguities. Cognition, 48, 243-279. May, R (1993). Logical
Form. Cambridge, MA: MIT Press
|
