This study applies the computational theory of the ‘discriminative lexicon’ (Baayen et al., 2019) to the modeling of the production of regular and irregular English verbs in aphasic speech. Under impairment, speakers with memory loss have been reported to have greater difficulties with irregular verbs, whereas speakers with phonological impairment are described as having greater problems with regulars. Joanisse and Seidenberg (1999) were able to model this dissociation, but only by selectively adding noise to the semantic units of their model. We report two simulation studies in which topographically coherent regions of phonological and semantic networks were selectively damaged. Our model replicated the main findings, including the high variability in the consequences of brain lesions for speech production. Importantly, our model generated these results without having to lesion the semantic system more than the phonological system. The model’s success turns out to hinge on the use of a corpus-based distributional vector space for representing verbs’ meanings. Joanisse and Seidenberg (1999) used one-hot encoding for their semantic representation, under the assumption that semantically regular and irregular verbs do not differ in ways relevant to impairment in aphasia. However, irregular verbs have denser semantic neighborhoods than do regular verbs (Baayen and Moscoso del Prado Martín, 2005), and we show that in our model this greater density renders irregular verbs more fragile under semantic impairment. These results provide further support for the central idea underlying the discriminative lexicon: that behavioral patterns can, to a considerable extent, be understood as emerging from the distributional properties of a language and basic principles of human learning.