The ability to categorize is imperative in the emergence of language (Medin & Schaffer, 1978). The present study examines the effects of typicality, a component of categorization which involves having semantic features in common within a category (Stanczack & Waters, 2006). Galizio, Stewart, and Pilgrim (2004) used a trial and error procedure to successfully train typically developing college-aged students to categorize 24 nonsense figures into 3 classes based upon typicality. The present study reports efforts to systematically replicate the findings of Galizio et al.'s (2004) trial and error procedure on 3 individuals with aphasia. A second study investigates the effectiveness of an error-reduction learning procedure on another group of 3 individuals with aphasia using the same stimuli and classes as Galizio, et al. (2004). Typicality effects did not remain consistent between the present study and the systematically replicated study. Unlike the present study, typicality effects were observed in college students in that stimuli with more class-defining features were learned with fewer errors, responded to more promptly, and interpreted as better prototypes of the class than stimuli with less class-defining features (Galizio, et al., 2004). There was a statistically significant reduction of errors in the error reduction procedure compared to the trial and error procedure. Only 1 out of 3 participants, however, demonstrated successful learning of the nonsense word to novel picture relationships. The results provide a framework for future methods to develop error-reduction procedures to address categorization in the aphasia population.