Friedrich Hayek

As we advance we find more and more frequently that we can in fact ascertain only some but not all the particular circumstances which determine the outcome of a given process; and in consequence we are able to predict only some but not all the properties of the result we have to expect. Often all that we shall be able to predict will be some abstract characteristic of the pattern that will appear — relations between kinds of elements about which individually we know very little. Yet, as I am anxious to repeat, we will still achieve predictions which can be falsified and which therefore are of empirical significance.