| |
Abstract:
Shapiro and Hestvik (1995) performed a cross-modal lexical
decision study with VP-ellipsis sentences such as 'The policeman
defended himself and the fireman did too, according to someone
who was there.' Recognition of a probe word that was semantically
related to the first clause subject (e.g. 'robber') turned out to
be facilitated at the location of the ellipsis ('did too') and a
subsequent position. Shapiro and Hestvik interpret this result as
evidence for the on-line computation of the strict reading of the
anaphor 'himself', as part of the reconstruction of the VP. This
is a surprising finding, in light of the observation that the
strict reading, which supposedly invokes vehicle change
(transformation of an anaphor into a pronominal) is strongly
dispreferred in intuitive judgments and other off-line tasks. We
surmise that Shapiro and Hestvik's result is not related to the
reflexive pronoun at all. We accept the suggestion that the first
clause VP is reconstructed at the location of the ellipsis.
However, we hypothesize that in order to reconstruct the VP,
readers/listeners have to search a memory representation of the
preceding clause for a valid 'filler'. As an (unintended) result
of this search, the first clause subject is re-activated. This
account yields a straightforward prediction: Re-activation of the
first clause subject in the reconstruction of ellipsis will also
take place if the antecedent VP contains a non-reflexive
predicate. We ran an end-of-sentence probe recognition experiment
with English stimuli similar to those used by Shapiro and
Hestvik, e.g. "The renowned judge severely criticized himself,
and the prosecutor did too," probe: "renowned." The subjects were
native speakers of English. The results confirm that first clause
subjects are re-activated in VP-ellipsis with a reflexive
predicate, just like Shapiro and Hestvik report. Moreover, the
probe recognition times also provide evidence for re-activation
of the first clause subject in VP-ellipsis constructions with
transitive (non-reflexive) predicates. These results cast doubt
on the 'processing reality' of vehicle change (strict
interpretation), and support the idea that the comprehension of
ellipses depends on a memory scan of immediately preceding
linguistic input.
Shapiro, L.P. Hestvik, A. (1995). "On-line comprehension of
VP-ellipsis: Syntactic reconstruction and semantic influence."
Journal of Psycholinguistic Research,
24, 517-532.
|
