ABSTRACT This article discusses current perspectives on foreign language loss and offers a reinterpretation of regressionist theories of attrition in which last-presented FL material is predicted to deteriorate first. Before reporting the results of a small-scale study of the retention and loss of Spanish among college undergraduates, the author describes a processing model of FL maintenance which holds that FL material having cognitive associations is most likely to be retained. Results of the investigation, which provide evidence for a processing view, are then discussed with respect to developing a more widely-based research paradigm.