Wayne Cowart

Experimental syntax : applying objective methods to sentence judgments

Introduction Are Judgements Stable? Error Variance in Sentence Judgements Designing Experiments on Acceptability The Sentence Judgement Task Presenting Sentence Materials to Informants Response Methods and Scaling Issues Sampling Settings for Experiments The Organization and Construction of Questionnaires Coding and Decoding the Data Summarizing the Data Statistical Issues

Experimental evidence for a minimalist account of English resumptive pronouns

In this article we provide evidence for a Minimalist account of English-type resumptive pronouns. Our findings provide empirical support for syntactic theories that, like Minimalist accounts, allow for competition among derivations. According to our account, resumptive pronouns are spell-outs of traces. For reasons of economy, the resumptive pronoun surfaces only when the derivation with the trace is precluded by syntactic principles. This account predicts that resumptive pronouns should only improve violations of constraints on representation, and not violations of constraints on movement. We tested this prediction by conducting an acceptability judgment task with 36 native speakers of English. The results bore out our prediction; subjects preferred the resumptive pronoun over the trace in cases where the trace itself was illicit, but not in cases where only the movement operation was illicit.

Anchoring and Grammar Effects in Judgments of Sentence Acceptability

This study examined the relation between anchoring effects, as demonstrated in 1992 by Nagata, and grammar-based effects in judgments of sentence acceptability. 35 subjects judged the acceptability of target sentences representing six different syntactic types. There were highly robust differences among these sentence types arising from differences in sentence structure. For one group of subjects the target sentences were mingled with a long list of highly acceptable sentences (High Anchor Set). A second group saw the same target sentences with an Anchor Set in which one-third of the sentences were of very low acceptability (Mixed Anchor Set). Target sentences seen in the context of the Mixed Anchor Set were judged more acceptable (an anchoring effect); however, the effect of Anchor Set did not interact with other factors. The relative acceptability of the six target types was unchanged in the two anchor conditions. Implications for the psychological theory of sentence judgments are considered. In particular, it is argued that anchoring effects do not arise in the cognitive mechanisms that evaluate sentence structure.

Equivocal Evidence on Field-Dependence Effects in Sentence Judgments

The study reported here examined field-dependence effects on judgments of grammaticality. Several sentence types involving coordinate structures were used. The types we tested were based on examples such as, The patients pretended that Elinor injured themselves and the dietician. Earlier work shows that the presence of the coordinate structures (themselves and the dietician) greatly improves acceptability of apparently ungrammatical sentences such as this. The present results show no reliable differences in the over-all pattern of judgments with these and related coordinate structures for subjects classified as field-dependent and field-independent; however, some intriguing differences in pattern were nonetheless apparent across subject types, although they were not reliable in all relevant tests. In addition to reliable differences in gross response level, field-dependent people showed an attenuation of certain differences noted for field-independent people. Thus, our results constitute equivocal evidence of differences in the grammatical preferences of field-dependent and field-independent people. We argue, however, that these results reflect differences in the processing styles of the two types of individuals, not differences in the principles or logic implemented by the grammatical systems of the two types.