Page:Mind (New Series) Volume 15.djvu/317

 THE EXPERIMENTAL THEORY OF KNOWLEDGE. 303- which the rose formerly occupied. By reason of disappoint- ment, the person may turn epistemologist. He may then take the discrepancy, the failure of the smell to execute its own intended meaning, as a wholesale, rather than as a specific fact : as evidence of a contrast in general between things meaning and things meant, instead of as evidence of the need of a more cautious and thorough inspection of odours and execution of operations indicated by them. One may then say : Woe is me ; smells are only my smells, sub- jective states existing in an order of being made out of con- sciousness, while roses exist in another order made out of a radically different sort of stuff ; or odours are made out of ' finite ' consciousness as their stuff, while the real things, the objects which fulfil them, are made out of an ' infinite ' consciousness as their material. Hence some purely meta- physical tie has to be called in to bring them into connexion with each other. And yet this tie does not concern knowledge ;. it does not make the meaning of one odour any more correct than that of another, nor enable us to discriminate relative degrees of correctness. As a principle of control, this tran- scendental connexion is related to all alike, and hence con- demns and justifies all alike. 1 It is interesting to note that the transcendentalist almost invariably first falls into the psychological fallacy ; and then having himself taken the psychologist's attitude accuses the empiricist whom he criticises of having confused mere psychological existence with logical validity. That is, he begins by supposing that the smell of our illustration (and all the cognitional objects for which this is used as a symbol) is a purely mental or psychical state, so that the question of logical reference or intention is the problem of how the 1 The belief in the metaphysical transcendence of the object of know- ledge seerns to have its real origin in an empirical transcendence of a very specific and describable sort. The thing meaning is one thing ; the thing meant is another thing, and is (as already pointed out) a thing presented as not given in the same way as is the thing which means. It- is something to be so given. No amount of careful and thorough inspection of the indicating and signifying things can remove or annihilate this gap. The probability of correct meaning may be increased in varying degrees and this is what we mean by control. But absolute certitude can never be reached except experimentally except by performing the operations indicated and discovering whether or no the intended meaning is fulfilled in propria persona. In this experimental sense, truth or the object of any given meaning is always beyond or outside of the cogni- tional thing which means it. Error as well as truth is a function of knowing. But the non-empirical account of this transcendent (or be- yond) relationship puts all the error in one place (our knowledge), and all the truth in another (absolute consciousness or else a thing-in-itself).