The current consensus among psychoanalysts holds that direct infant observations are one means for testing the developmental propositions of psychoanalytic theory; that the observations have already falsified some of the theory's basic propositions; and that they hold the key to a qualitatively different developmental theory of psychoanalysis. The consensus, although not universal, has motivated a wide range of research programs on early infancy, whose findings are commonly interpreted as disclosing psychoanalytic metapsychology and clinical theory in an entirely new light. This essay examines some of the assumptions that have motivated such investigations, as well as the research strategies by which the new versions of theory are promulgated. On the basis of these explorations it is concluded that psychoanalytically informed infant observations may be the source for new theories of social-emotional development, but that they are essentially irrelevant for psychoanalysis as a psychology of meanings, unconscious ideas, and hidden motives.