Taxonomic abstraction in psychobiology

If a body of knowledge in a scientific discipline is to be extended beyond empirical observation and into the realm of laws and principles, one of the fundamental requirements is a taxonomy which supports the systematic integration of observations. Psychobiology benefits from taxonomies provided by biology and chemistry, which include not only object oriented taxonomies such as species or chemical elements, but also process oriented taxonomies, such as oxidation, metabolism, phototaxis, or predation. Psychobiology has yet to provide equivalent taxonomies for its behavioral observations, although the common use of terms such as fear, anger, arousal, stress, and memory might lead one to suppose that these are based on a well established taxonomy of behavioral measures. In this report the logical and quantitative requirements for treating behavioral measures in terms of taxonomic classes are reviewed. A sample of studies representing recent research in psychobiology was examined to assess interest in such a taxonomy and to identify elements of current practice which might contribute to its development. Recent practice displays some evidence of interest in behavioral classes, in choice of language, and in frequent use of multiple dependent measures. Multivariate methods, which might elicit from such data evidence contributing to the development of a taxonomy, are rarely used. Recommendations are given on some appropriate analytic methods for data resulting from current practice and for new exploratory paradigms which could aim directly at the establishment of taxonomic classes for behaviors.