Background: Conversation is the archetypal mode of communication. As a process it draws on numerous skills, and predispositions, adapting to dynamic contexts and coordinated in highly sophisticated ways for successful interaction. The combination of these abilities with contextual variables coalesces uniquely to represent what this research proposes to be conversational intelligence. It is argued here that high levels of conversational intelligence rest to a large degree on executive functions (EF) which are steadily becoming more widely acknowledged and researched within the communication domain. The impairment of EF in neurologically injured individuals has significant, though as yet undisclosed, repercussions for recovery, response to therapy and ability to integrate communication skills in every day interactions to support conversational success. This study incorporates some new approaches to the study of communication disorders following stroke, including conversation, executive functions and the possibility of pharmacological intervention. Aims: The aims of this study were to describe in detail the language, executive function and conversational characteristics of ten individuals who had experienced strokes and to examine the relationships among these three areas of functioning. In addition, this research investigated the response of these ten participants to pharmacological therapy on a one month trial of Leviteracetam (LEV). Methods and Procedures: Ten individuals who had suffered single incident strokes were recruited from local community and rehabilitation facility referrals. All ten participated in a four stage randomised, double blind investigation including baseline, active, placebo and withdrawal phases. At each stage participants underwent testing on an EF battery and were recorded having conversations with familiar interlocutors. In addition, significant others completed a rating scale assessing affective features and behaviours and language testing was conducted at the baseline phase using the WAB. The language and EF data were scored and the conversations subjected to Conversation Analysis. For each participant, profiles were created and assessed for interrelationships between the executive characteristics and conversational features representative of each executive construct. Repeated measures analysis of variance was conducted on EF data for the four phases of the study to determine significant pharmacological effects. Outcomes and Results: The majority of the sample presented with significant EF deficits across most areas assessed. Two participants presented with essentially intact profiles which were not explained in terms of types of aphasia or site of lesion. Language results proved to be inconsistently associated with EF deficits, but conversational features reflected underlying executive strengths or deficits with greater consistency. Two participants experienced amelioration (though not statistically significant) of interference control during the active phase, with observable improvements in conversational skill. One participant demonstrated improved conversation without a change in EF scores. The response to LEV is evaluated with reference to potential alternatives. Conclusions: The existence of EF deficits in individuals with stroke is demonstrated. Furthermore, the impact of these impairments is considerable and observable during naturally occurring conversations, suggesting the centrality of the EF contribution to conversational intelligence. The lack of association between formal language test scores and EF impairments argues for the inclusion of more authentic assessment approaches for stroke patients. The data is explored in terms of specific consequences of different lesion sites on EF and communication and briefly addresses bilingualism as a potential variable in explaining some of the variations in the data. Pharmacotherapy is addressed as an important focus of future research protocols. Implications for assessment and treatment are discussed as well as proposals for future study.