Autoantibodies have the potential of pathogenicity in several diseases. In rheumatoid arthritis (RA), however, this has not been ultimately proven. RA is characterized by a variety of autoantibodies. Newer insights into characteristics of rheumatoid factors indirectly suggest their pathogenetic involvement. In contrast, antibodies to collagen, despite the availability of an experimental model, do not appear to be pathogenetic in man. Anti-hnRNP antibodies, particularly anti-A2/RA33, are present in RA and experimental models of RA, and therefore, aside from their diagnostic value in established and early RA, could also be involved in the disease process. The nature of Sa, another target antigen in RA, has not yet been elucidated. Filaggrin is the antigen recognized by antikeratin antibodies and antiperinuclear factor; however, citrullin is the target amino acid in filaggrin, and anticitrullin antibodies have a high predictive value. Among a series of cartilage proteins, most have not yet been characterized sufficiently; one, gp39, appears to be of particular interest. Whether or not these antibodies are involved in RA pathogenesis is not yet known. It can be speculated that autoimmunity to some, if not all, of these autoantigens mirrors events important in the development of RA, but further studies on T-cell reactivities and in experimental models are needed to fully understand the involvement.