Thinking is important because, as we have seen, it is that function in which given or ascertained facts stand for or indicate others which are not directly ascertained. But the process of reaching the absent from the present is peculiarly exposed to error; it is liable to be influenced by almost any number of unseen and unconsidered causes,--past experience, received dogmas, the stirring of self-interest, the arousing of passion, sheer mental laziness, a social environment steeped in biased traditions or animated by false expectations, and so on. The exercise of thought is, in the literal sense of that word, _inference_; by it one thing _carries us over_ to the idea of, and belief in, another thing. It involves a jump, a leap, a going beyond what is surely known to something else accepted on its warrant. Unless one is an idiot, one simply cannot help having all things and events suggest other things not actually present, nor can one help a tendency to believe in the latter on the basis of the former. The very inevitableness of the jump, the leap, to something unknown, only emphasizes the necessity of attention to the conditions under which it occurs so that the danger of a false step may be lessened and the probability of a right landing increased. [Sidenote: Hence, the need of regulation which, when adequate, makes proof] Such attention consists in regulation (1) of the conditions under which the function of suggestion takes place, and (2) of the conditions under which credence is yielded to the suggestions that occur. Inference controlled in these two ways (the study of which in detail constitutes one of the chief objects of this book) forms _proof_. To prove a thing means primarily to try, to test it. The guest bidden to the wedding feast excused himself because he had to _prove_ his oxen. Exceptions are said to prove a rule; _i.e._ they furnish instances so extreme that they try in the severest fashion its applicability; if the rule will stand such a test, there is no good reason for further doubting it. Not until a thing has been tried--"tried out," in colloquial language--do we know its true worth. Till then it may be pretense, a bluff. But the thing that has come out victorious in a test or trial of strength carries its credentials with it; it is approved, because it has been proved. Its value is clearly evinced, shown, _i.e._ demonstrated. So it is with inferences. The mere fact that inference in general is an invaluable function does not guarantee, nor does it even help out the correctness of any particular inference. Any inference may go astray; and as we have seen, there are standing influences ever ready to assist its going wrong. _What is important, is that every inference shall be a tested inference_; _or_ (since often this is not possible) _that we shall discriminate between beliefs that rest upon tested evidence and those that do not, and shall be accordingly on our guard as to the kind and degree of assent yielded_.