Background: The ability to generate words that follow certain constraints, or verbal fluency, is a sensitive indicator of neurocognitive impairment, and is impacted by a variety of variables. Aims: To investigate the effect of post-stroke aphasia, elicitation category and linguistic variables on verbal fluency performance. Methods & Procedures: Twenty-eight persons with aphasia (PWA) with a single left-hemisphere lesion and 40 age-matched neurotypical community-dwelling adults were administered three verbal fluency tasks--two semantic (animals and actions) and one phonemic (the letters F, A and S). Data analysis included comparison of total scores, clusters and perseverations. Individual responses were coded for frequency of occurrence, age of acquisition and syllable length to investigate qualitative differences in word generation. Outcomes & Results: PWA performed worse than neurotypical participants across all verbal fluency tasks, and animal fluency scores were farthest from neurotypical performance. PWAs' animal and action fluency were correlated with other language measures, while phonemic fluency was uncorrelated with language measures. While some PWAs showed dissociations between verbal fluency tasks, the dissociations did not pattern along with aphasia fluency. PWAs produced fewer clusters and responses with higher word frequency across all three verbal fluency tasks. Responses had earlier age of acquisition and shorter word length for animal and phonemic fluency, but not action fluency. Conclusions & Implications: Verbal fluency, particularly animal fluency, is sensitive to even mild aphasia. PWA produced lexically simpler responses than their neurotypical peers. This study identifies the relevance of qualitative analysis of verbal fluency responses.