Extant research on dirty work—occupations involving physical, social or moral taint, which affect worker identities—has been read primarily through the lens of social identity theory (SIT). There are two notable shortcomings that emerge as a consequent of dirty work being too heavily reliant upon the precepts of SIT, which we seek to remedy in this article: (1) the overemphasis on the symbolic to the detriment of the material, which has led to false optimism regarding the ability for subjects doing dirty work to exercise agency in constructing their own sense of selves and, (2) the failure to substantively account for the role of identity differences, which suggests that empirical research on the phenomenon is devoid of proper historical and cultural contextualization. Drawing on a qualitive study on low-caste toilet cleaners in Pakistan, our findings were largely incongruous with the scholarly conceptualization of dirty work that has been propagated to date. We explicate the embedded role of power and context in dirty work, which are not adequately considered using SIT alone. Repudiating the overly romanticized version of the concept, we argue that SIT’s account of the concept ought to be complemented by social construction theory going forward.