It is assumed that confidentiality is not one singular ethical entity but a conglomerate of quite different issues depending upon clinical context and the sector of information sharing at stake. The focus here is on how to think psychoanalytically about requests for information from third parties (payers, courts, public security). Defining confidentiality as a promise to 'never tell anything' outside of the relationship omits evaluation of the impact of the third's listening on the combined freedom of thought and freedom of speech in analyst and analysand. Circulation of information outside the dyad need not be toxic, need not disrupt the analytic couple's openness to new meaning. Key to contamination and inhibition of analytic work is whether or not disclosure serves an analytic end. Current defense of confidentiality relies heavily on the models of protection of privacy and professional secrecy, which, though useful and relevant, fail to encompass the transitional, intersubjective space engendered by the analytic process. Suggestions are made for alternate sources of paradigms better suited to represent the latter. Offered for discussion is a draft of a confidentiality policy with respect to third parties that is informed by psychoanalytic theory and clinical practice rather than by local legal jurisdiction or original disciplines' ethics codes.