In this article, I address a number of important issues regarding the nature of repression in bilingual analyses. Repressed memories are assumed to hold the key to pathological formations and to the development of transferential manifestations that arise in attempts to work through the conflicts associated with these memories. The nature and quality of the individual's dreams, as well as the whole associative process, are also assumed to be influenced in this way. Analysis of patients who know more than one language to communicate their experience, however, is beginning to raise some questions as to the best way to understand the work of repression in these individuals. I contend that because certain memories are more accessible in one language than the other, repression for bilingual individuals may be partly a function of linguistic accessibility. The concept of repression will therefore need to be reassessed in the context of these observations.