The constraints obeyed by a measure of a theory's logical content are exactly the same as the probability calculus, so a probability assignment can always be consistently reinterpreted as a statement of a theory's degree of logical weakness—saying one theory is more probable than another can mean it says less.