Abstract
This paper addresses the problem of intermodal perception, that of how warranted perception arises of objects having characteristics in multiple sense modalities. It first shows the inadequacy of the currently popular explanations of such perception in terms of special, innate mechanisms. It proposes instead a phenomenological account in terms of an infant's general capacities for observation and thought. To this end it prepares the terrain with brief investigations into four topics: spontaneity, non-symbolic thinking, the role of spontaneity in perception, and the lived body. It then traces how an infant may, through spontaneous observations of invariable co-presence and sequence, integrate tactile objects, then tastes, smells, and sounds, and finally, with the help from topological and dynamic congruence, visual objects into a single world with the lived body as centre of orientation.