Principle adopted by English economist John Maynard Keynes (1883-1946) to underpin his Bayesian approach to induction by finding a justification for assigning the relevant probabilities.
The principle says that, for at least that sphere we are investigating, the number of objects and qualities it contains may be infinite, but the number of independent groups into which they fall is finite - or at least that there is a non-zero probability of this.
One of the objections to the principle is that there is no adequate reason to think it true, and that even if true it will not help us to assign actual figures to the probabilities unless we know how many independent groups there are; otherwise we cannot know that any progress we make by the Bayesian procedure is more than infinitessimal.
J M Keynes, A Treatise on Probability (1921), ch. 22,9