Abstract
Psychologists sometimes discuss the need to refine clear designations of the observable units comprising their subject matter. This paper links such discussions to (a) Dewey and Bentley's (1949) account of specification as relatively accurate unit-designation, and (b) the logical base of scientific classifications and abstractions in observable particulars. The paper then reviews, clarifies, evaluates, and contrasts the psychological units proposed by Kantor (behavior segment), Skinner (operant), and Lee (deed). Overall, Lee's deed is found to be the sharpest, least ambiguous designation, and the only specification. Deeds, fields of contributors, and contingencies are then used to selectively integrate aspects of all three units. The resulting integration is consistent with field-based approaches to causal relations within and among units, where the noun cause is synonymous with one of many contributors. It is also applicable to the analysis of feedback loops, which are designated as circular networks of dependency among subclasses of deeds.