It is out of the question to set forth in this place in detail the contrast between transcendental epistemology and an experimental theory of knowledge. It may assist the understanding of the latter, however, if I point out, baldly and briefly, how, _out of the distinctively empirical situation_, there arise those assumptions which make knowledge a mystery, and hence a topic for a peculiar branch of philosophizing. As just pointed out, epistemology makes the possibility of knowledge a problem, because it assumes back of knowledge conditions incompatible with the obvious traits of knowledge as it empirically exists. These assumptions are that the organ or instrument of knowledge is not a natural object, but some ready-made state of mind or consciousness, something purely “subjective,” a peculiar kind of existence which lives, moves, and has its being in a realm different from things to be known; and that the ultimate goal and content of knowledge is a fixed, ready-made thing which has no organic connections with the origin, purpose, and growth of the attempt to know it, some kind of _Ding-an-sich_ or absolute, extra-empirical “Reality.” (1) It is not difficult to see at what point in the development of natural knowledge, or the signifying of one thing by another, there arises the notion of the knowing medium as something radically different in the order of existence from the thing to be known. It arises subsequent to the repeated experience of non-fulfilment, of frustration and disappointment. The odor did not after all mean the rose; it meant something quite different; and yet its indicative function was exercised so forcibly that we could not help--or at least _did_ not help--believing in the existence of the rose. This is a familiar and typical kind of experience, one which very early leads to the recognition that “things are not what they seem.” There are two contrasted methods of dealing with this recognition: one is the method indicated above (p. 93). We go more thoroughly, patiently, and carefully into the facts of the case. We employ all sorts of methods, invented for the purpose, of examining the things that are signs and the things that are signified, and we experimentally produce various situations, in order that we may tell _what_ smells mean roses _when_ roses are meant, what it is about the smell and the rose that led us into error; and that we may be able to discriminate those cases in which a suspended conclusion is all that circumstances admit. We simply do the best we can to regulate our system of signs so that they become as instructive as possible, utilizing for this purpose (as indicated above) all possible experiences of success and of failure, and deliberately instituting cases which will throw light on the specific empirical causes of success and failure.