Rational Belief and the Common Cause Principle

Abstract

During the past decade, Wesley Salmon has successively refined and elaborated Reichenbach’s principle of the common cause, as part of a wide-ranging inquiry into statistical inference and explanation. Being as convinced as he of the central importance of the probabilistic concept of common cause, but skeptical of its universal applicability, I have corresponded with Salmon on this subject, and some of that correspondence has made an appearance in both our published writings.² In this paper I shall try to state exactly what I think is the significance (especially the epistemological significance) of the principle and provide a new (I hope, improved) version of one line of argument in our correspondence. I should emphasize that this has been a very cooperative enterprise: my arguments always began in rather feeble, intuitive form and in answering them, Salmon would gently correct my mistakes, or restate the arguments in a stronger and more precise form than I had managed, and often add striking, concrete illustrations.

Research for this paper was partly supported by a National Science Foundation grant, and it was written while I was released from teaching duties by the University of Toronto Connaught Senior Fellowship. Besides my main, and very large debt to Wesley Salmon, I benefitted greatly from correspondence and conversations with Nancy Cartwright, Ellis Crasnow, Joe Hellige, Jonathan Katz, Edwin Levy, Ben Rogers, and Ron Webster.