Identity research largely emerged from clinical observations. Decades of empirical work advanced the field in refining existing approaches and adding new approaches. Furthermore, the existence of linkages of identity with psychopathology is now well established. Unfortunately, both the directionality of effects between identity aspects and psychopathology symptoms, and the mechanisms underlying associations are unclear. In the present paper, we present a new framework to inspire hypothesis-driven empirical research to overcome this limitation. The framework has a basic resemblance to theoretical models for the study of personality and psychopathology, so we provide examples of how these might apply to the study of identity. Next, we explain that unique features of identity may come into play in individuals suffering from psychopathology that are mostly related to the content of one's identity. These include pros and cons of identifying with one's diagnostic label. Finally, inspired by Hermans' dialogical self theory and principles derived from Piaget's, Swann's and Kelly's work, we delineate a framework with identity at the core of an individual multidimensional space. In this space, psychopathology symptoms have a known distance (representing relevance) to one's identity, and individual multidimensional spaces are connected to those of other individuals in one's social network. We discuss methodological (quantitative and qualitative, idiographic and nomothetic) and statistical procedures (multilevel models and network models) to test the framework. Resulting evidence can boost the field of identity research in demonstrating its high practical relevance for the emergence and conservation of psychopathology.