Massive modularity theory has replaced classic, Fodorian modularity as a major focus of research within cognitive science. The massive modularity thesis posits that there are a large number of knowledge and action stories, designed in a piecemeal fashion over evolutionary time to solve specific, adaptive problems. We criticize massive modularity as a general theory of human cognition, with particular attention to the issue of context-sensitive perception, thought, and language. First, the experimental methods used to uncover individual modules are notoriously inadequate and fail to meet reasonable standards by which modules may be dissociated from one another. Second, input criteria, by which modules are presumably defined, may be impossible to discover given the context-embedded nature of human performance. Third, catalogues of experimental effects that are assumed to demonstrate the presence of modules do not constitute a comprehensive theory of the acknowledged interaction of brain, body, and world in ordinary cognition. An alternative conception of cognitive performance, based on principles of self-organization, better explains the embedded, context-sensitive mechanisms of adaptive cognition.