This paper provides an integrative review of two schools of psychological research on memory. In one, the developers of associative network models (e.g., Anderson, 1983; Bower, 1981; Rumelhart and McClelland, 1986) base their work on experimentation and computer simulations. In the other, clinicians studying dissociative disorders (e.g., Bliss and Jeppsen, 1985; Braun, 1984a; Kluft, 1985a; Putnam et al., 1986) develop theories through the observation of patients and the collection of clinical statistics. Both schools of research are attempting to understand how items in memory are variably accessible and inaccessible in order to conceptualize an overall structure of the way memory is connected and organized as well as to develop a vocabulary of rules about how memory works. Connections found between the two areas result in three topics of particular interest. First, the central role of affect in dissociative disorders suggests a reevaluation of the minor role that it has been given in most network models. Second, the extreme levels of affect present in clinical observation, not ethically reproducible in mood and memory experimentation, call attention to the necessity of examining clinical data in establishing experimental theories. Third, the role of inhibition in “neural” network memory models, taken for granted in clinical research, must be accounted for in network models of memory.