BackgroundParanoia is associated with a multitude of social cognitive deficits, observed in both clinical and subclinical populations. Empathy is significantly and broadly impaired in schizophrenia, yet its relationship with subclinical paranoia is poorly understood. Furthermore, deficits in emotion recognition – a very early component of empathic processing – are present in both clinical and subclinical paranoia. Deficits in emotion recognition may therefore underlie relationships between paranoia and empathic processing. The current investigation aims to add to the literature on social cognition and paranoia by: (1) characterizing the relationship between paranoia and empathy, and (2) testing whether there is an indirect effect of emotion recognition on the relationship between empathy and paranoia.MethodsParanoia, empathy, and emotion recognition were assessed in a non-clinical sample of adults (n = 226) from the Nathan Kline Institute-Rockland (NKI-Rockland) dataset. Paranoia was measured using the Peters Delusions Inventory-21 (PDI-21). Empathy was measured using the Interpersonal Reactivity Index (IRI), a self-report instrument designed to assess empathy using four subscales: Personal Distress, Empathic Concern, Perspective Taking, and Fantasy. Emotion recognition was assessed using the Penn Emotion Recognition Test (ER-40). Structural equation modeling (SEM) was used to estimate relationships between paranoia, the four measures of empathy and emotion recognition.ResultsParanoia was associated with the Fantasy subscale of the IRI, such that higher Fantasy was associated with more severe paranoia (p < 0.001). No other empathy subscales were associated with paranoia. Fantasy was also associated with the emotion recognition of fear, such that higher Fantasy was correlated with better recognition of fear (p = 0.008). Paranoia and emotion recognition were not significantly associated. The Empathic Concern subscale was negatively associated with emotion recognition, with higher empathic concern related to worse overall emotion recognition (p = 0.002). All indirect paths through emotion recognition were non-significant.DiscussionThese results suggest that imaginative perspective-taking contributes to paranoia in the general population. These data do not, however, point to robust global relationships between empathy and paranoia or to emotion recognition as an underlying mechanism. Deficits in empathy and emotion recognition observed in schizophrenia may be associated with the broader pathology of schizophrenia, and therefore not detectable with subclinical populations.