Tree structures, the work of listening, and schizophrenic discourse: a reply to Beveridge and Brown

A. W. Beveridge and K. Brown (Brain and Language, 24, 174-181) have criticized a study reported by R. Hoffman, L. Kirstein, S. Stopek, and D. Cicchetti (Brain and Language, 15, 207-233), which examined processing tasks imposed upon listeners by schizophrenic speech. The authors argue that our discourse analysis was not sufficiently objective since different discourse structures can at times be generated on the basis of a particular deviant speech segment. However, the analysis which we described does not purport to yield an "objective" structure of schizophrenic texts but rather attempts to reflect representations that individual listeners generate in processing speech; our model only requires that different interpretations of a particular segment generate different discourse structures. Deficiencies in hierarchical structure associated with schizophrenic speech increase processing requirements for listeners and may cause interpretive variability. Incoherence scores generated on the basis of our analysis were used to estimate the amount of extra "work of listening" that particular speech segments require. Because these incoherence scores were blindly and reliably determined, they can be used to make statistical inferences about the prevalence of structural deficiencies of schizophrenic speech compared to the speech of controls.