We propose a program of research to test an explanation for the discourse impairments exhibited by stroke patients, especially those with right hemisphere brain damage. The main theme of this project is what has been termed a person's """"""""theory of mind,"""""""" i.e., his or her ability to infer and understand what beliefs, emotions, and motivations are in someone else's mind. The ability to attribute accurately mental states to other people plays a critical role in communication. For example, listeners need to infer speakers' beliefs and intentions in order to determine whether a nonveridical utterance (one that is not literally true) is an honest error, a deliberate lie, or an attempt at joking, teasing, or irony. That is, a speaker who states that the weather is great when it is raining may not know it is raining (an error), may know it's raining buy wants to deceive the listener (a lie), or may be joking about the weather (irony). Similarly, speakers need to infer what their listeners do and do not know in order to make their utterances informative and relevant: When asked what one is reading, it is not informative to reply """"""""a book,"""""""" since the speaker already knows this much. A relation between the ability to infer mental states and discourse ability has already been established in both normal and in autistic children. The proposed research has two major goals. First, we assess the abilities of both anterior and posterior groups of unilaterally right and left hemisphere damaged stroke patients to understand beliefs (which are nonobservable, mental, representational states) extends to other nonobservable states (e.g., hunger) and to nonmental representations (e.g., photographs), or reflects a selective deficit limited to understanding mental, representational states. Second, we test whether theory of mind deficits predict patients' performance in a range of discourse tasks. Some of these tasks require interpretation of nonveridical utterances as lies, mistakes, or irony, in which it is particularly critical to infer speakers' beliefs and intentions. Other tasks test subjects' abilities to adjust the form of literal utterances to reflect what a listener knows and does not know. The proposed research will be presented as a battery of tasks such that (with two exceptions explained in the text all subjects will participate in all tasks. the results of this project should provide a theoretical account of communication deficits that often characterize right hemisphere damaged patients and that can severely limit their long-term success in social and familial settings. This work may also shed more light on the surprisingly preserved discourse comprehension of left hemisphere damaged, aphasic patients.
