Many with psychosis experience substantial difficulties forming and maintaining social bonds leading to persistent social alienation and a lack of a sense of membership in a larger community. While it is clear that social impairments in psychosis cannot be fully explained by symptoms or other traditional features of psychosis, the antecedents of disturbances in social function remain poorly understood. One recent model has proposed that deficits in social cognition may be a root cause of social dysfunction. In this model social relationships become untenable among persons diagnosed with psychosis when deficits in social cognition result in inaccurate ideas of what others feel, think or desire. While there is evidence to support the influence of social cognition upon social function, there are substantial limitations to this point of view. Many with psychosis have social impairments but not significant deficits in social cognition. First person and clinical accounts of the phenomenology of psychosis also do not suggest that persons with psychosis commonly experience making mistakes when trying to understand others. They report instead that intersubjectivity, or the formation of an intimate shared understanding of thoughts and emotions with others, has become extraordinarily difficult. In this paper we explore how research in metacognition in psychosis can transcend these limitations and address some of the ways in which intersubjectivity and more broadly social function is compromised in psychosis. Specifically, research will be reviewed on the relationship between social cognitive abilities and social function in psychosis, including measurement strategies and limits to its explanatory power, in particular with regard to challenges to intersubjectivity. Next, we present research on the integrated model of metacognition in psychosis and its relation to social function. We then discuss how this model might go beyond social cognitive models of social dysfunction in psychosis by describing how compromises in intersubjectivity occur as metacognitive deficits leave persons without an integrated sense of others' purposes, relative positions in the world, possibilities and personal complexities. We suggest that while social cognitive deficits may leave persons with inaccurate ideas about others, metacognitive deficits leave persons ill equipped to make broader sense of the situations in which people interact and this is what leaves them without a holistic sense of the other and what makes it difficult to know others, share experiences, and sustain relationships. The potential of developing clinical interventions focused on metacognition for promoting social recovery will finally be explored.