sense does not know what an error is, and needs more light upon the subject. Let common sense not disturb us, then, in our further search, by the constant and indignant protest that error must somehow exist, and that doubt on that subject is nonsense. Nobody has any doubts on that subject. We ask only how error exists and how it can exist.
For the rest, what follows is not any effort to demonstrate in fair and orderly array, from any one principle or axiom, what must be the nature of error, but to use every and any device that may offer itself, general analysis, special example, comparison and contrast of cases, — anything that shall lead us to the insight into what an error is and implies. For at last, immediate insight must decide.
We shall study our problem thus. We shall take either some accepted definition of error, or some special class of cases, and we shall ask: How is error in that case, or in accordance with that definition, possible? Since error plainly is possible in some way, we shall have only to inquire: What are the logical conditions that make it possible? We shall take up the ordinary suppositions that common sense seems to make about what here determines the possibility of error. We shall show that these suppositions are inadequate. Then the result will be that, on the ordinary suppositions, error would be impossible. But that result would be absurd, if these were the only possible suppositions. Hence the ordinary suppositions must somehow be supplemented. When, therefore, we seem to say in the following that error is impossible, we shall mean
