This article reports the development and evaluation of a new intervention termed "Supported Conversation for Adults with Aphasia" (SCA). The approach is based on the idea that the inherent competence of people with aphasia can be revealed through the skill of a conversation partner. The intervention approach was developed at a community-based aphasia center where volunteers interact with individuals with chronic aphasia and their families. The experimental study was designed to test whether training improves the conversational skills of volunteers, and, if so, whether the improvements affect the communication of their conversation partners with aphasia. Twenty volunteers received SCA training, and 20 control volunteers were merely exposed to people with aphasia. Comparisons between the groups' scores on a Measure of Supported Conversation for Adults with Aphasia provide support for the efficacy of SCA. Trained volunteers scored significantly higher than untrained volunteers on ratings of acknowledging competence [F(1, 36) = 19. 1, p < .001] and revealing competence [F(1, 36) = 159.0, p < .001] of their partners with aphasia. The training also produced a positive change in ratings of social [F(1, 36) = 5.7, p < .023] and message exchange skills [F(1, 36) = 17.6, p < .001 ] of individuals with aphasia, even though these individuals did not participate in the training. Implications for the treatment of aphasia and an argument for a social model of intervention are discussed.