[PS-3.37] GOOD, BAD... OR BOTH: ANIMACY AND THE PROCESSING OF COMPLEX SENTENCES

Blohm, S. 1, 2 , Kretzschmar, F. 2 & Schlesewsky, M. 3, 2

1 MPI for Empirical Aesthetics, Frankfurt, Germany
2 University of Mainz, Germany
3 University of South Australia, Adelaide, Australia

For many languages, the processing penalty for object (ORCs) vs. subject relative
clauses (SRCs) reduces/vanishes if the relativized head is inanimate and the RC-subject animate (Mak et al. 2002). Whether animacy and RC-type cues interact with information about the head's syntactic function in the main clause (MC) remains unclear. The parallel-function-hypothesis (Sheldon 1974) predicts facilitated processing (regardless of animacy), if the head's syntactic functions in MC and RC match.
We recorded ERPs (N=24) and eye movements (N=48). Participants read locally ambiguous SRCs/ORCs in German. The 2x2x2 design crossed animacy of the
head/RC-internal noun (A-I/I-A), RC-type (SRC/ORC) and syntactic function of the
head (S/O).
In experiment 1, all conditions showed a LAN-like negativity relative to SRCs
with animate heads at the disambiguating verb. In experiment 2, ORC verbs were read faster with inanimate heads. We found no preference for parallel functions. At the RC-internal noun phrase, however, processing was impaired for animate nouns (as indexed by a late positivity and increased go-past time) only if the RC modified inanimate MC-objects.
Our results corroborate the notion of animacy as a potential remedy reducing the ORC-penalty and suggest that grammatical and semantic cues interact across
clause boundaries in the processing of complex sentences.