Studies spanning several decades have revealed how the complex forces of antigen processing distinguish those epitopes of a protein that dominate the immune response from those that remain cryptic. Since foreign antigens and self-proteins are subjected to the same proteolytic pathways before presentation to the T-cell repertoire, it has long been assumed that they comply equally with the established rules of immunodominance. Nevertheless, the pathological determinants of some autoantigens appear ill-equipped for the dominant role they adopt, displaying features more befitting subdominant or cryptic epitopes, such as low affinity for their MHC restriction element. These findings may be reconciled by suggesting that, far from remaining sequestered during ontogeny, many classical autoantigens participate in the establishment of self-tolerance, the efficiency with which individual epitopes purge the T-cell repertoire being determined by the conventional rules of immunodominance: while those epitopes that are truly dominant induce profound non-responsiveness, those that are poorly presented may leave residual reactivity, manifest in the periphery as responses to epitopes that appear inappropriately dominant. Here we review recent evidence showing the process of self-tolerance to be uniquely responsible for the reversal of immunodominance which promotes such epitopes to an undeserved position of importance within the determinant hierarchy.