We revisit, within Harsanyi’s impartial observer setting, the question of foundations underlying procedural fairness concerns in welfare judgments. In our setup—that of allocating an indivisible good using a lottery—such concerns, presumably, matter. We draw from the social preferences literature and relax a typical assumption made while addressing this question, namely, that individuals in society do not care about procedural fairness and such concerns arise exclusively at a societal level, which are captured by non-linear social welfare functions (SWFs). In our model, individual attitudes towards procedural fairness are identified and factored into welfare judgments. Specifically, we provide an axiomatic basis within Harsanyi’s (1955) framework to represent procedural fairness sensitive individual preferences by the representation in Karni and Safra (2002). We then show, in terms of underlying axioms, how such individual assessments incorporating both risk and procedural fairness attitudes can be aggregated by means of utilitarian and generalized utilitarian SWFs.