A basic assumption ofthe behavioral sciences is that there exists a direct, self-evident relationship between the behavioral categories, or types, used to describe behavior and how the organism actually behaves. This assumption is challenged. It appears that behavioral types are typically created apriori, either with an eye to convenience of measurement or on the basis of anthropomorphic or intuitive inferences. Furthermore, it appears that human observers do not record behavioral types accurately or consistently. The elementary units, or tokens, ofbehavior are not recorded and, therefore, play no explicit role in the creation of the types used to describe behavior. It is suggested that behavior be described explicitly through the use offormal notation systems. This suggestion is supported by a review of studies that have used such systems to reveal striking new insights into the organization ofbehavior. Ofthese systems, Eshkol-Wachman Movement Notation appears exceptionally powerful for extracting invariant features ofbehavior. We suggest that the adoption of such methods will lead to behavioral taxonomies that are both ecologically and neurologically valid.Until recently, psychology was a philosophical discipline concemed with the nature of mind. During the latter part of the 19th century, it emerged as a scientific discipline. The spirit of the times dictated that mind, like all phenomena, should be regulated by identifiable, regular laws similar to those of Newtonian physics. Mind, however, was not amenable to experimentation because it was not directly observable. In order to found a "science" of psychology, it was necessary to shift the focus of psychology from the mind to overt behavior. As J. B. Watson, the most articulate advocate of a science ofbehavior, proclaimed, the key to unlocking general psychologicallaws was the systematic observation and measurement of behavior. lnitially, this approach seemed to be a simple one. Behavior appeared easy to observe and measure, and the task of creating a set of general psychologicallaws did not look insurmountable. In addition, given reasonably generallaws ofbehavior, it seemed that anchoring them to the physiological substrate would be straightforward. An understanding of the organization of the physiology could, in turn, provide predictive explanations of behavior and useful tools for the clinic. The