ABSTRACT
In this chapter the reader observes a similar situation unfolding concerning contemporary theories of the process underlying form recognition. The contemporary consensus among theoreticians, if the literature is to be accepted at face value, is that recognition occurs as a result of the analysis of a form into local features. As I review the theoreticalliterature, it is shown that the problem of form recognition is nearly universally approached from a point of view that argues that the analysis of complex forms into their "parts," "features," or "components" is the initial information-processing step in classification and recognition. I believe, however, that a considerable body of evidence argues that there are global strategies at work in human form perception that have little to do with the feature-by-feature algorithms that currently are the most popular theories of form recognition. The analogy I have drawn here between synapses and learning on the one hand and features and form recognition on the other is far from complete, however. In the former case there is little inconsistency between theory and empirical data; in the latter case data clashes with theory, even if this is not generally appreciated to be the case.
