Wesley C. Salmon

Causality without Counterfactuals

This paper presents a drastically revised version of the theory of causality, based on analyses of causal processes and causal interactions, advocated in Salmon (1984). Relying heavily on modified versions of proposals by P. Dowe, this article answers penetrating objections by Dowe and P. Kitcher to the earlier theory. It shows how the new theory circumvents a host of difficulties that have been raised in the literature. The result is, I hope, a more satisfactory analysis of physical causality.

Causality and Explanation: A Reply to Two Critiques

This paper discusses several distinct process theories of causality offered in recent years by Phil Dowe and me. It addresses problems concerning the explication of causal process, causal interaction, and causal transmission, whether given in terms of transmission of marks, transmission of invariant or conserved quantities, or mere possession of conserved quantities. Renouncing the mark-transmission and invariant quantity criteria, I accept a conserved quantity theory similar to Dowes--differing basically with respect to causal transmission. This paper also responds to several fundamental constructive criticisms contained in Christopher Hitchcocks discussion of both the mark-transmission and the conserved quantity theories.

Why Ask, ‘Why?’? An Inquiry Concerning Scientific Explanation

[In ‘The Philosophy of Hans Reichenbach’ (above), I remarked upon the fertility of his posthumous work, The Direction of Time, especially its fourth chapter, which deals with the causal and statistical relations among macrophenomena. I emphasized the value of his ideas in connection with the explication of scientific explanation — in particular, their potential utility for purposes of developing a full-blown theory of causal explanation. Since that article was written, I have been devoting considerable attention to just such issues, and I feel that my hopes have been —at least to some extent — vindicated.

Causality: Production and Propagation

A theory of causality based upon physical processes is developed. Causal processes are distinguished from pseudo-processes by means of a criterion of mark transmission. Causal interactions are characterized as those intersections of processes in which the intersecting processes are mutually modified in ways which persist beyond the point of intersection. Causal forks of three kinds (conjunctive, interactive, and perfect) are introduced to explicate the principle of the common cause. Causal forks account for the production of order and modifications of order; causal processes account for the propagation of causal influence.

On Vindicating Induction

This paper deals with the problem of vindicating a particular type of inductive rule, a rule to govern inferences from observed frequencies to limits of relative frequencies. Reichenbachs rule of induction is defended. By application of two conditions, normalizing conditions and a criterion of linguistic invariance, it is argued that alternative rules lead to contradiction. It is then argued that the rule of induction does not lead to contradiction when suitable restrictions are placed upon the predicates admitted. Goodmans grue-bleen paradox is considered, and an attempt to resolve it is offered. Finally, Reichenbachs pragmatic argument, hinging on convergence properties, is applied.

Scientific Explanation: Three Basic Conceptions

By contrasting three general conceptions of scientific explanation, this paper seeks to clarify the explanandum and to exhibit the fundamental philosophical issues involved in the project of explicating scientific explanation. The three conceptions--epistemic, modal, and ontic--have both historical and contemporary importance. In the context of Laplacian determinism, they do not seem importantly distinct, but in the context of irreducibly statistical explanations, the three are seen to diverge sharply. The paper argues for a causal/mechanical version of the ontic conception, and concludes by exhibiting some striking consequences of this approach.

The Justification of Inductive Rules of Inference

Publisher Summary This chapter discusses the justification of inductive rules of inference. If the logic of science can be explicated without reference to inductive inference, the problem of justification of induction can be dismissed, cheerfully dispensed with inductive logic. The only mode of inference available for the acceptance or rejection of hypotheses is modus tollens, and it is suitable only for rejection. Hypothetico-deductive theorists and inductive probabilists maintain that positive instances tend to enhance the probability of the hypothesis or give it inductive support. When a hypothesis has been falsified, it is discarded and replaced by another hypothesis that has not yet experienced falsification. Hypotheses differ from one another with respect to the ease with which they can be falsified. A highly falsifiable hypothesis which is severely tested and survives the tests without actually being falsified becomes highly corroborated. However, hypotheses must not be regarded as true because they are highly corroborated.

A Third Dogma of Empiricism

In 1951, W. V. Quine published his provocative and justly famous article, ‘Two Dogmas of Empiricism’. At about the time Quine was mounting this attack, a number of ‘empiricists’ were busily establishing what has subsequently become, in my opinion, a third dogma. The thesis can be stated quite succinctly: scientific explanations are arguments. This view was elaborated at considerable length by a variety of prominent philosophers, including R. B. Braithwaite (1953), Ernest Nagel (1961), Karl Popper (1959) and most especially Carl G. Hempel.1 Until the early 1960s, although passing mention was sometimes made of the need for inductive explanation, attention was confined almost exclusively to deductive explanation. In 1962, however, Hempel (1962a) made the first serious attempt to provide a detailed analysis of inductive (or statistical) explanation. In that same year, in a statement referring explicitly to both deductive and inductive explanations, he characterized the ‘explanatory account’ of a particular event as ‘an argument to the effect that the event to be explained ⋯ was to be expected by reason of certain explanatory facts’ (Hempel, 1962b, my italics). Shortly thereafter he published an improved and more detailed version of his treatment of inductive-statistical (I-S) explanation (Hempel, 1965, pp. 381–412). In this newer discussion, as well as in many other places, Hempel has often reiterated the thesis that explanations, both deductive and inductive, are arguments.2 The purpose of this present paper is to raise doubts about the tenability of that general thesis by posing three questions — ones which will, I hope, prove embarrassing to those who hold it.3

Hans Reichenbach’s Vindication of Induction

Hans Reichenbach believed that he had solved Hume’s problem of the justification of induction, but his arguments have not proved persuasive to most other philosophers. The majority of those who addressed the problem held, for one reason or another, that it is a pseudo-problem. In a number of articles during the 1950s and 1960s I tried to refute this position.1 I still believe it is incorrect — that the problem of justification of induction is a genuine and profoundly important problem — but I shall not rehearse that issue here. In this article I shall first discuss Reichenbach’s justification and the problems confronting it. I shall then consider other attempts at vindication that in one way or another pursue a similar goal. In the end, I shall maintain, Reichenbach’s program can succeed, given certain additional considerations, a crucial one of which is found in his own writings.

Should we attempt to justify induction

IN THE broadest sense, an inductive inference is any non-demonstrative inference to a matter of fact. An inductive rule, then, would be any nondeductive rule of inference for drawing matter of fact conclusions, provided that such a rule does not sanction drawing self-contradictory conclusions from any consistent set of premises (including the null set). I regard the problem of justifying induction as the problem of justifying a choice from among the wide variety of possible inductive rules. The question whether past experience is to be a guide to the future is included in the problem thus formulated, for among the possible rules are some which render evidence about the past irrelevant to predictions of the future.* In recent years a rather large number of philosophers have argued that the attempt to justify induction ought to be abandoned. They have supported this claim by arguments designed to show that a justification of induction is either impossible or unnecessary or both. Within this paper