In my book, The Quark and the Jaguar, I mention the possibility that the predictions are probabilistic, that our universe may have one out of a number of possible solutions, and that the probability of such a solution may be calculable. Perhaps it's e to the minus 2 script S, where script S is some kind of loop corrected and Euclideanized action for the situation under discussion. In that case, the lowest S would be… state… the lowest S solution would be the one with the highest probability, or if they degenerate, the lowest S solutions would be the ones with the highest probability. That would greatly reduce the predictive power of the theory, but not necessarily to the point where it would be unsatisfactory as a theory. It might reduce the predictive power. I don't regard it as such a horrible thing if that were to happen because it… we know that the history of the universe is co-determined by the fundamental laws and a very long sequence of chance events or accidents, this would just add an initial accident. It would adjoin an initial accident to the list, which is perhaps unfortunate from the point of view of predictive power, but it wouldn't invalidate the whole idea. I don't know whether it will turn out to be that way or not. Perhaps there's one phase that's selected in some… some clear manner.