This paper, which analyzes the evaluative aspect of affectivity, argues that the functional internalization of organic evolution for the formation of the intellectual system of human beings required that some means of selecting behavioral variations be devised, and that the selective mechanism devised was the affective system. From this perspective, it becomes possible to relate Pugh's (1977) conception of value-driven decision systems and his account of affective mechanisms to Cellerier's (1979a, 1979b) analysis of pragmatic and epistemic transformations. Tenuous as the resulting model may be, it gives a picture more complete than other models of the role affective evaluations play in adapting action and in constructing knowledge. In a preliminary way, the model also allows the bringing together of empirical data on problem solving, on decision making, and on the evolution of scientific thought. Further study is needed to determine whether these beginnings can mature into a full-blown theory in which other aspects of affectivity will be included. (RH)