Demonstration effects only a conditioned, mediated certainty; by virtue of it, something is certain if another thing is certain. If any doubt arises as to the certainty of this other, then this certainty must again be appended to the certainty of a third, and so on. Now, is this retrogression carried on _ad infinitum_, or is there anywhere a final link? I know very well that some are of the former opinion; but these men have never considered that if it were so, they would not even be capable of entertaining the idea of certainty—no, not even of hunting after certainty. For what this may mean: to be certain; they only know by being themselves certain of something; but if everything is certain only on condition, then nothing is certain, and there is even no conditioned certainty. But if there is a final link, regarding which no question can be raised, why it is certain, then, there is an undemonstrable at the base of all demonstration. They do not appear to have considered what it means: to have proven something to _somebody_. It means: we have demonstrated to him that a certain other certainty is contained, by virtue of the laws of thinking, which he admits, in a certain first certainty which he assumes or admits, and that he must necessarily assume the first if he assumes the second, as he says he does. Hence all communication of a conviction by proof, presupposes that both parts are at least agreed on something. Now, how could the Science of Knowledge communicate itself to the dogmatist, since they are positively _not agreed in a single point_, so far as the _material_ of knowledge is concerned, and since thus the common point is wanting from which they might jointly start.[10]