Temporal Processing and Syntactic Complexity

John Hoeks and Laurie Stowe
Behavioral and Cognitive Neurosciences
University of Groningen

In studies on temporal connectives a distinct processing difference has been observed between the connectives 'before' and 'after'. For instance, an ERP study by Münte et al. (1998) showed that sentences starting with 'before'- clauses (e.g., Before <event1>, <event2>.) elicited a larger negative slow wave than sentences starting with 'after'-clauses, suggesting that extra processing is required when events are presented out of chronological order. A later ERP study (Hoeks & Stowe, 2000) replicated these findings and also showed that temporal processing is affected by semantic complexity (i.e., establishing co-reference), suggesting that temporal and other semantic processes share a common resource.

In the present study, two self-paced reading experiments will be discussed, aimed at investigating whether temporal processing is also influenced by syntactic complexity. The first experiment examined the effect of temporal connectives in subject-relative (SR) and object-relative (OR) sentences; the second experiment looked at temporal processing in less complex active versus passive sentences.

The results of the first experiment showed that 'before'-sentences cause (slight) processing difficulty in OR sentences only; in SR sentences they were read significantly faster than 'after'-sentences. The second experiment showed no interaction of active/passive with temporal processing, but a small overall processing advantage for 'before'-sentences as compared to
sentences starting with 'after'. This was especially apparent in the second clause of the experimental sentences; in the first clause 'before'-sentences were read somewhat more slowly, though this difference was not significant.

On the basis of these findings we would like to propose that in the first clause of 'before'- sentences a kind of 'temporal ordering frame' is set up, which under normal circumstances may require only very modest processing effort. With this time-frame in place, the rest of the sentence can be read quite easily. In 'after'- sentences on the other hand, this frame might not be constructed a priori, but applied and evaluated during integration with the second clause. This could explain why 'before'-sentences were read faster than 'after'-sentences in Experiment 2 and in the SR sentences of Experiment 1. If, however, the construction of a time-frame is complicated by increased memory demands (as in the OR sentences, or due to word-by-word presentation used in ERP experiments), the temporal ordering of the events may have to be constructed during or at the end of the sentence, which will require more processing resources for 'before' than for 'after'-sentences.

References

Hoeks, J.C.J., & Stowe, L.A. (2000). Slow-wave ERP-patterns reflecting semantic processing during sentence comprehension. Journal of Cognitive Neuroscience, S11, 122.

Münte, Th.F., Schiltz, K., & Kutas, M. (1998). When temporal terms belie conceptual order. Nature, 395, 71-73.