No final statement can be made regarding the relations between science and policy-making. Knowledge, values, and techniques are interrelated, cumulative, and constantly changing. They are derived from man's responses to the complicated interactions between physical, biological, and cultural phenomena. Final answers are impossible because the answers themselves are part of the world and therefore are factors in changing it. We see through a glass darkly, whether it be the giant glass of Palomar or the eye-piece of the electron microscope.

For the case of statistical theories, the criteria of explanation, prediction, and testability can all be viewed as particular instances of a more general evaluation scheme. Using the ideas of a gain matrix and expected gain from statistical decision theory, these three criteria can be compared in terms of the elements in their associated gain matrices. This analysis leads to (1) further understanding of the interrelationship between the current criteria, (2) the proposal of an ordering for the criteria, and (3) the suggestion of a new criterion.

The naive misunderstanding among economists of the relationship between economics and philosophy is as profound as it is prevalent. To be sure, economists have concerned themselves with the philosophy of economics only so much as to familiarize themselves with methodology and its application to economic inquiry. Since economics is the oldest and perhaps the most “respectable” of the social sciences its scholars have developed a phobia for anything they might suspect to be even remotely unscientific. Consequently, they have erroneously narrowed their understanding of the nature of economics. This narrowness prevails largely because most economists possess little or no training in general philosophy and especially in the philosophy of science. As a result of this apparent deficiency there exists in the minds of these economists confusion and inability to distinguish between teleology, axiology, and ethics as they do or do not apply to economics.

My purpose here is to elaborate the reasons I maintain that Salmon has not been completely successful in reporting the history of work on explanation. The most important limitation of his account is that it does not emphasize the critical necessity to embrace a suitable conception of probability in the development of the theory of probabilistic explanation.

Popper and Miller (1983) have presented an argument purporting to establish the impossibility of inductive probability. Here I discuss critically their characterization of a deductive part of nondeductive support, a point that has not figured centrally in previous responses.

To overcome sociopsychologism and historical relativism, the growth of science is deduced from the combined effect of postulated invariant controls, in the form of enduring ideals of science, in their interaction with nature. The thus constituted “cybernetics-of-science” concept permits extrapolation from present to future states of science. The ideal scientific theory is the goal or target toward which the scientific process is oriented, by virtue of its invariant controls. The form of the ideal theory can thus be extrapolated, and some speculative hypotheses advanced as to its contents, taking those of the recently emerged constructs of science as basis which best accord with the predicted form of the theory.