“Having been all my life since school-days engaged in the practice of architecture and civil engineering, I can assure Prof. Max Müller that designing and invention are done entirely by mental pictures. I find that words are only an encumbrance. In fact, words are in many cases so cumbersome that other methods _have_ been devised for imparting knowledge. In mechanics the graphic method, for instance.”[20] 2. Its aim is always practical. It should never be forgotten that at the outset, the faculty of cognition is essentially utilitarian, and cannot be otherwise, because it is employed solely for the preservation of the individual (in finding food, distinguishing enemies from prey, and so on). Animals exhibit only _applied_ reasoning, tested by experience; they feel about and choose between several means,—their selection being justified or disproved by the final issue. Correctly speaking, the logic of images is neither true nor false; these epithets are but half appropriate. It succeeds or fails; its gauge is success or defeat; and as we maintained above that it is the secret spring of æsthetic invention, let it be noticed that here again there is no question of truth or error, but of creating a successful or abortive work. Accordingly, it is only by an unjustifiable restriction that the higher animals can be denied all functions beyond that of association, all capacity for inference by similarity. W. James (after stating that, _as a rule_, the best examples of animal sagacity “may be perfectly accounted for by mere contiguous association, based on experience”), arrives virtually at a conclusion no other than our own. After recalling the well-known instance of arctic dogs harnessed to a sledge and scattering when the ice cracked to distribute their weight, he thus explains it: “We need only suppose that they have individually experienced wet skins after cracking, that they have often noticed cracking to begin when they were huddled together and that they have observed it to cease when they scattered.” Granting this assumption, it is none the less true that associations by contiguity are no more than the _material_ which serves as a substratum for inference by similarity, and for the act which follows. Again, a friend of James, accompanied by his dog, went to his boat and found it filled with dirt and water. He remembered that the sponge was up at the house, and not caring to tramp a third of a mile to get it, he enacted before his terrier (as a forlorn hope) the necessary pantomime of cleaning the boat, saying: “Sponge, sponge, go fetch the sponge.” The dog trotted off and returned with it in his mouth, to the great surprise of his master. Is this, properly speaking, an act of reasoning? It would only be so, says James, if the terrier, not finding the sponge, had brought a rag, or a cloth. By such substitution he would have shown that, notwithstanding their different appearance,