Psychological verbs and the double-dependency hypothesis

The double-dependency hypothesis (DDH, Mauner et al., 1993) holds that where two dependencies of a certain kind are present, comprehension in Broca's aphasia will be random, but that where there is only one dependency, comprehension will be intact. We tested this hypothesis by examining the performance of Broca's aphasics on sentences with psychological verbs of two different classes. One class has an argument structure in which the Experiencer role is assigned to the subject. In the other class, the Experiencer role is assigned to the object. Subject-Experiencer verbs can form verbal passives which have two relevant dependencies, whereas object-Experiencer verbs can form adjectival passives and have only one relevant dependency. Thus these sentence types make contrasting predictions relevant to the DDH. Our results clearly demonstrate that patients understand the adjectival passive psychological verbs, as predicted by the DDH. On the verbal passive psychological verbs, patients perform at chance, again consistent with DDH predictions. These results firmly buttress the DDH account. They also contradict the results of an earlier study (of verbal passive psychological verbs only), a study which we argue is plagued with problems (namely, Grodzinsky, 1995b).