| |
Abstract:
Abstract: Current models of sentence processing and
reanalysis make different predictions concerning the effect of
increasing distance between syntactically dependent elements, and
the nature of the intervening material on processing load. This
study investigates whether the processing of subject-verb agreement
is influenced by (i) the distance between the subject and the verb,
keeping syntactic complexity constant; and (ii) by the presence of
a noun phrase with interfering number features. Event Related
Potentials were recorded while participants were reading Dutch
subject-object-verb adjunct clauses and performing an
end-of-sentence acceptability judgment task. The subject and the
verb were separated by either five words (2 constituents, 2 new
discourse referents) or two words (one constituent, 1 new discourse
referent); the verb either agreed or did not agree with the
subject. ERPs to the ungrammatical vs. grammatical verbs showed a
P600, preceded by an early negativity, and followed by a negativity
(N400) that was limited to the clause containing the violation. The
distance between the subject and the verb did not have any effect.
This is somewhat problematic for theories such as Gibson (1998),
but supports Fodor & Inoue (1998). In addition, the P600
latency for violations was shortest when both subject and object
were singular, compared to other number combinations. This
partially confirms results from production (Bock & Cutting,
1982) and reading studies (Pearlmutter et al, 1999).
|
