SALMON, WESLEY
(1925–2001)

The American philosopher of science Wesley Charles Salmon was born August 9 in Detroit, Michigan, and died April 22 near Madison, Ohio. He pursued undergraduate studies at Wayne University and the University of Chicago Divinity School, received an MA in philosophy from the University of Chicago in 1947, and a PhD in philosophy from the University of California at Los Angeles in 1950. His principal academic appointments were at Brown University, Indiana University, the University of Arizona, and the University of Pittsburgh; he retired from this last institution in 1999. At UCLA his dissertation advisor was the philosopher of science Hans Reichenbach and much of Salmon's subsequent work was influenced by Reichenbach's philosophy. A lifelong defender of empiricism, Salmon made significant contributions to a wide range of topics, primarily in explanation, causation, inductive inference, and the philosophy of probability.

Work

Beginning in 1971, Salmon developed a widely discussed alternative to Carl Hempel's covering law model of scientific explanation. The key element of Salmon's statistical relevance model was its insistence that explanatory factors must be statistically relevant to the occurrence of the event to be explained. This undermined in two ways Hempel's view that an explanation must lead people to expect the explanandum to occur. It showed that this condition was not necessary because events with low probability, such as the occurrence of lung cancer, can be explained in terms of statistically relevant factors such as cigarette smoking. It also showed that Hempel's model did not provide sufficient conditions for an explanation because irrelevant factors such as a man's taking birth control pills, when included in a Hempelian explanation, undermine the effectiveness of an explanation of his not getting pregnant.

In the course of developing the statistical relevance model, Salmon began to stress the importance of the causal relevance, rather than the statistical relevance, of explanatory factors; his 1984 book Scientific Explanation and the Causal Structure of the World contains an account of probabilistic causality grounded in an "at-at" theory of causation within which spatiotemporally continuous markable processes connecting cause and effect play a central role. The aim was to provide an account of causation free of appeals to counterfactuals and thus acceptable to an empiricist, yet different from Hume's in stressing the importance of connecting processes. In the light of criticisms that the markability criterion required tacit appeals to counterfactuals, Salmon abandoned it in the early 1990s and adopted a position where the transmission of conserved quantities was what distinguished causal from non-causal processes.

This appeal to causal processes and the distinction between genuine processes and pseudo-processes meshed with Salmon's interests in space-time theories. In that area he defended a causal theory of space and time within which the direction of time was to be grounded in causal asymmetries. He also maintained a long-term interest in conceptions of synchrony and in defending a conventionalist approach to simultaneity relations.

In his 1984 book, Salmon argued for a form of scientific realism based on the principle of the common cause. This principle states that if an association is observed between two types of event then, in the absence of a direct causal connection between instances of the events, there exists a common cause responsible for generating the association. This principle is general and can be used to argue for the existence of unobserved entities. It lies behind the reasoning used by Bertrand Russell in inferring the continued existence of a cat from its occasional observed appearances and was employed, Salmon claimed, by Jean Perrin in using the similarity of values obtained from different experimental techniques to determine Avogadro's number to argue for the reality of atoms. Despite its appeal, the principle does have its limitations. It is inapplicable in certain quantum mechanical situations where there are no hidden variables. It is also easy to find cases where two properties each increase over time but there is no common cause underlying the two. Nevertheless, Salmon's emphasis on this principle has led to an important new way of thinking about scientific realism.

Much of Salmon's early work concerned issues in probability and induction. For many years he defended Reichenbach's pragmatic vindication of induction, which argues that inductive inferences, more specifically the "straight rule" that projects the existing relative frequency of an event's occurrence into the future, is at least as likely to be successful as any other rule. In the light of criticisms by Ian Hacking, Salmon tempered his advocacy of this approach while continuing to insist on the importance of linguistic invariance for inductive rules.

Throughout his career Salmon defended a relative frequency interpretation of probability, including its use in his accounts of causation. For Salmon, the correct frequency to attribute to an event was the frequency within the broadest homogenous reference class to which the event belongs—that is, that class of events for which no further statistically relevant factors exist. Although he occasionally displayed sympathy for a propensity approach to probabilities and appreciated the role played by logical probabilities in Carnap's inductive logic, he developed important criticisms of both. In many of his writings, Salmon argued that an objective form of Bayesian inference could illuminate a number of issues in the philosophy of science. Most notable, Salmon's insistence on preserving the context of discovery/context of justification distinction and on using Bayesian methods in the latter area led him to argue that Kuhn's account of theory choice could be made more objective by employing Bayesian techniques of theory justification.

In addition to his philosophical contributions, Salmon was an outstanding teacher and was much admired for his personal qualities. His introductory logic book was widely used as an undergraduate text, went into three editions, and was translated into five foreign languages. The exemplary clarity of his writing is evident in all of his publications.

Bibliography

primary works

The Foundations of Scientific Inference. Pittsburgh: University of Pittsburgh Press, 1967.

Space, Time, and Motion. 2nd ed. Minneapolis: University of Minnesota Press, 1981.

Scientific Explanation and the Causal Structure of the World. Princeton, NJ: Princeton University Press, 1984.

Four Decades of Scientific Explanation. Minneapolis: University of Minnesota Press, 1990.

Causality and Explanation. New York: Oxford University Press, 1998.

Reality and Rationality, edited by M. Salmon and P. Dowe. New York: Oxford University Press, 2005.

secondary works

Fetzer, J., ed. Probability and Causality: Essays in Honor of Wesley C. Salmon. Dordrecht, Netherlands: D. Reidel, 1988. Contains a comprehensive annotated list of his publications through 1988.

McLaughlin, R., ed. What? Where? When? Why?: Essays on Induction, Space and Time, Explanation. Dordrecht, Netherlands: D. Reidel, 1982. Contains a brief autobiographical essay.

Paul Humphreys (2005)