Illicit Acceptability in picture NPs

Four experiments examine the interaction between extraction and specificity in picture NPs. The results indicate that the acceptability judgements of naive speakers show highly robust patterns that do not conform well to widely held assumptions about the relative acceptability of several theoretically important kinds of sentence. There is also evidence that the difference between argument and non-argument extractions has a marked impact on acceptability (though no such acceptability difference has figured in linguistic theory). Further, the paper argues that there are circumstances in which ungrammatical sentences may be rendered acceptable via the intrusion of extragrammatical mechanisms in comprehension. Thus, the acceptability of these sentences is 'illicit'.The work reported in this paper was supported in part by a grant from the National Institutes of Health (1 R01 NS22606-01)