Why reread? Evidence from garden-path and local coherence structures

ABSTRACT

Two eye-tracking experiments were conducted to compare the online reading and offline comprehension of main verb/reduced relative garden-path sentences and local coherence sentences. Rereading of early material in garden-path reduced relatives should be revisionary, aimed at reanalysing an earlier misparse; however, rereading of early material in a local coherence reduced relative need only be confirmatory, as the original parse of the earlier portion of these sentences is ultimately correct. Results of online and offline measures showed that local coherence structures elicited signals of reading disruption that arose earlier and lasted longer, and local coherence comprehension was also better than garden path comprehension. Few rereading measures in either sentence type were predicted by structural features of these sentences, nor was rereading related to comprehension accuracy, which was extremely low overall. Results are discussed with respect to selective reanalysis and good-enough processing.

KEYWORDS:

• Garden-path sentence
• Good-enough processing
• Local coherence
• Rereading
• Language comprehension
• Eye movements

Acknowledgements

Portions of this work were presented at AMLaP 2011, Paris. We thank the audience of that presentation for helpful questions and discussion. We also are grateful for the excellent work of the members of the EdPsych Psycholinguistics Lab, Beckman Institute, especially Cassie Palmer.

ORCID

Erika K. Hussey http://orcid.org/0000-0002-9782-5943

Notes

1 The disambiguated version of this sentence was faulty, so the entire item in both LC and GP conditions was removed from analysis.

Additional information

Funding

This work was supported by National Science Foundation (NSF) CAREER Award [grant number BCS-0847533] to K. Christianson and National Institutes of Health (NIH) Training [grant number T32-HD055272] (PI: Sarah Brown-Schmidt) to E. Hussey.