Robert M. W. Travers's 1980 essay is the focal point of this paper. He argues that evaluation research requires a scientific taxonomy of human behavior and learning. Such a taxonomy must be experimentally based, mathematically expressed, theoretically sound, and predictive. He bases such criteria on the development of scientifically successful taxonomies, specifically, the Periodic Table of the Elements. Unscientific taxonomies are either metaphysical (Aristotle) or standardized practices (Bloom). But, Travers has misread his own historical example; he cannot eschew both metaphysics or human practices. Moreover, his example of a potential source for a scientific taxonomy, Piaget's cognitive science, is rife with problems. The paper concluded with reasons for the unrealizability of any attempt to scientifically describe human actions outside of historical situations. (Author)