Recent discourse in stigma literature has put the spotlight on mental illness, drawing attention to the dehumanizing effects of biological explanations of illness acquisition. Research has demonstrated dehumanization as multidimensional, in that targets may be denied certain capacities (e.g., emotion) while still attributed others (e.g., memory). Dehumanization, in turn, has been associated with decreased helpful and increased harmful behavior, this link theoretically consequent of the absence of empathy. Research into dehumanization's capacity-specific effect on empathy is lacking, such as how denying a person the capacity to feel emotion impacts the extent to which one empathizes with their emotions, irrespective of attributed cognitive capacity. The current research will examine if empathy in fact depends on the type of dehumanization associated with particular mental illnesses. In a pre-test (n = 95), mental illness dehumanization was found to vary on two dimensions, namely, moral and executive function capacities. For the main study (n = 85), participants were assigned to one of five conditions, varying the target group label on these dimensions: Morality (high vs. low) x Executive function (high vs. low), and a mentally healthy control. Participants observed video footage of the target experiencing embarrassment. We measured participant embarrassment mimicry whilst watching the video, and explicit reports of embarrassment following the video. We predicted participants to experience stronger empathic embarrassment towards targets high in morality, as increased morality corresponded to attributed embarrassment capacity in the pre-test. A between-participants ANOVA with orthogonal contrast codes was used to test the hypothesis. There was no evidence for a difference in empathy towards mentally healthy and mental illness targets. Furthermore, no evidence was found for an effect of morality. Interestingly, marginal effects revealed targets low in executive function to elicit the highest explicit empathic response, while those high in executive function elicited the least empathy. Results from the pre-test and main study demonstrated executive function to correlate significantly with competence and responsibility for behavior, pointing to potential explanatory correlates of the effect. While state inference of target embarrassment did have a marginal positive relationship with participant explicit embarrassment, there was no evidence that state inference explained the executive function-empathy relationship. Overall, the results neither support a general nor capacity-specific dehumanization-empathy link. The effect of executive function on increased explicit empathy corroborates emerging evidence in the literature that targets low in competence and lacking responsibility for their plight attract stronger empathic responses, regardless of their valence (i.e., warmth, group affiliation). The present findings and considerations imply we also shine a spotlight upon an unexpected source of stigma in mind attribution — the consequences of attributing capacities to individuals, especially to those deficient in them. In specific contexts, attributing the capacity for executive function to targets may in fact allow for lesser empathy. Future research should investigate the relationship between executive function, competence, and attributions, as well as their impact on empathic responding and stigma.