Effective Categorization of Objects , Scenes , and Faces Through Time

We first describe SLIP (Strategy Length & Internal Practicability) a formal model for thinking about categorization, in particular about categorization through time. We then discuss an early application of this model to basic-levelness. We continue with novel evidence for discrete processing cycles through time. We then turn to aspects of categorization through time that have been neglected in the categorization literature: our limited processing capacities; the necessity of having a flexible categorization apparatus; and the paradox that this inexorably brings about. Finally, we spend several pages on a twofold resolution of this paradox. Throughout, but especially toward the end, we attempt to bridge work done in categorization, vision, neuropsychology, and physiology. 3 1. A model of categorization Figure 1 shows four artificial scenes synthesized by combining two different luminance patterns (that we call flat and hilly) with two different chromatic patterns (labeled grassy and sandy). This toy-example captures some of the essential characteristics of real-world categorization. These stimuli can be categorized as either " field " (the combination of is_flat and is_grassy), " desert " (is_flat and is_sandy), " mountain " (is_hilly and is_grassy), or " dune " (is_hilly and is_sandy) at the most specific level of categorization. At the most abstract level of categorization there are more than one possibilities – we will come back to this point latter – but, for now, let us only consider the two categories " flat " and " hilly "