On the emergence of probability

Abstract

The modern theory of probability is usually dated from the second half of the 17th century. The famous Pascal-Fermat correspondence of 1654 began a rapid advance in the subject, and by the completion of Jacob Bernoulli's Ars Conjectandi (published posthumously in 1713, but written and discussed long before) one can say that the subject has more or less fully emerged. This of course raises an important historical question: what factors are responsible for the sudden growth of the theory of probability? Why did it happen when it did? In this paper we will examine an answer to this question recently put forward by Ian Hacking. In his book The Emergence of Probability,1 Hacking proposes that the sudden development of the theory of probability is to be explained by an important conceptual change in the way people thought about chance and evidence. The claim, in brief, is that the modern theory of probability emerged when it did because it was not until the middle of the 17th century that we possessed the modern concept of probability. We believe that Hacking is wrong. After presenting an outline of his thesis and the principal arguments that he offers for it, we will show that Hacking's explanation for the sudden activity in the theory of probability cannot be correct, since many of the concepts that Hacking believes constitute the core of our modern notion of probability were present long before the mid- 17th century. We will argue instead for a different explanation, one that accounts for the history of the theory of probability without appeal to radical conceptual revolution.