| |
Abstract:
n this study we will compare Fodor's (1998) "same size sister
principle" to the "attachment-binding dualism" as proposed by
Hemforth et al. (in press a). Essentially, Fodor's (1998) principle
states that ambiguous modifiers such as prepositional phrases (PP)
or relative clauses (RCs) are preferentially attached to a head of
similar size. For example, in two-site ambiguities of the type
NP1-NP2-modifier, a long RC or PP is preferentially attached to NP1
while a short one is preferentially attached to NP2. The
"attachment-binding dualism", on the other hand, states that the
attachment of RCs is determined by a combination of anaphoric and
syntactic principles: while anaphoric processes try to bind the
relative pronoun to the most salient host (e.g., a main discourse
referent, which is NP1 in two site contexts), syntactic processes
preferentially attach the RC to the most recent host, i.e., NP2 in
two-site contexts. In languages that use relative pronouns
consistently such as German, anaphoric processes override the
recency-based component and RCs are preferentially attached to NP1
(cf. Hemforth et al., in press a, b). In the case of PPs, on the
other hand, the anaphoric binding of a pronoun is not involved;
consequently, PPs are preferentially attached to NP2 according to
syntactic recency.
Since the modifiers we used in our previous experiments (e.g.,
Hemforth et al., in press a, b; Walter & Hemforth, 1998) varied
slightly in length, Fodor's principle might be an alternative
explanation for the results that we have proposed to support the
attachment-binding approach.
In order to directly compare the two approaches, we are
presently conducting a magnitude estimation study to be followed by
an eyetracking study using sentences like (1)-(8) in which we vary
the factors pragmatically forced attachment (high: 1-4 vs. low
5-8), the type of modifier (PP: 2, 4, 6, 8) vs. RC: 1, 3, 5, 7) and
the length of the modifier (long: 3, 4, 7, 8 vs. short: 1, 2, 5,
6).
The predictions of Fodor's approach depend on the length of the
ambiguous modifier: "same size sister" predicts an NP1 attachment
preference for long modifiers, independent of whether the modifier
is a PP or an RC. The predictions of attachment-binding, on the
other hand, do not depend on the length but on the type of the
modifier: for RCs, attachment-binding predicts an NP1 preference
since the RC is preferentially bound to the most salient host. In
the case of PPs, anaphoric binding is not involved such that PPs
are preferentially attached to NP2 according to syntactic
recency.
References
Fodor, J. D. (1998). Learning to Parse.
Journal of Psycholinguistic Research,
21 (5), 1303-1321.
Hemforth, B., Konieczny, L., & Scheepers, C. (in press a).
Syntactic attachment and anaphor resolution: Two sides of relative
clause attachment. In M. Crocker, M. Pickering, & C. Clifton,
Jr. (Eds.),
Architectures and Mechanisms for Language Processing.
Cambridge, UK: Cambridge University Press.
Hemforth, B., Konieczny, L., & Scheepers, C. (in press b).
Modifier attachment: relative clauses and coordinations. In B.
Hemforth & L. Konieczny (Eds.),
German Sentence Processing.
Dordrecht: Kluwer Academic Press.
Walter, M., & Hemforth, B. (1998). Syntactic and thematic
influences on relative clause attachment in three-site
constructions. Talk presented at the 4th Annual Conference on
Architectures and Mechanisms for Language Processing (AMLaP),
Freiburg, Germany, September 24-26, 1998.
|
