But would the dog ever accomplish much more than to attach a kind of concrete significance to the figures of the letters; than to associate his necessities or his natural or artificial wants with them? It is very doubtful, and that is what is indicated by Sir John Lubbock's experiments:
Sir John painted six cards, two blue, two red, and two yellow. Three of these were put before the dog, who was to bring his master the card of the color that was shown him. Although he was rewarded every time he succeeded, he never fairly understood what was wanted of him. This was because the action of bringing the card of the right color did not appeal directly enough to his senses. Sir John obtained no better result with six cards marked I, II, III, etc. Van never exactly grasped the conformity of the figures.
What was it prevented my dog, upon whom I tried experiments in numbers, grasping the difference between three and four pieces of meat? He failed because he had to abstract the ideas of the numbers 3 and 4 from the variety of the figures which were presented to him. I have no doubt he might in time have learned to distinguish the triangles from the squares which I formed on the plate with the three and the four pieces of meat. The thing that baffled the beast—we must not forget that the dog carries the faculty of observation to a considerable length—was the incessantly variable diversity of the figures. Under these circumstances, the problem was made too complicated for his head, those means only being given which I had prepared for entering into communication with his intelligence.
If any of the readers of these pages is tempted to teach a dog arithmetic, he would do well, I think, to begin by making him distinguish one from two, permitting him to touch only a single piece at the word one, and two pieces at the word two. Then he could pass on to three, and, if he went so far, to four. After that, he might essay addition: one and two, one and three, and two and three. The experiment would be very interesting and instructive, whatever the result might be. For, as Sir John Lubbock says, we ought not to aim for any one result rather than another, but for the truth.
Is the dog, after all, a suitable subject to experiment upon, in regard to the distance that separates man from animals? Would it not be better to select the monkey, intractable as he is, but formed like us, and not only able to imitate our gestures but fond of doing so? We might by this means attempt a verification of M. Noiré's seductive hypothesis respecting the origin of language: that it is the product of a social state already considerably advanced, and that the sounds, being at first simply utterances accompanying the movements of the whole, finally become the signs of those movements. But suppose, for a moment, that the dog acquires some notion of number, what are we to conclude from it? Is the advance of such a kind that it can be communicated to the whole species or to a particular breed? That would
