the case of my opponents, and thus clearly to mark out the field of discussion, that I have instituted the distinction between names as receptual and conceptual, or denotative and denominative. This distinction having been clearly understood, the next point was that both kinds of names admit of connotative extension—denotative names within the receptual sphere, and denominative within the conceptual. That is to say, when a name has been applied to one thing, its use may be extended to another thing, which is seen to belong to the same class or kind. The degree to which such connotative extension of a name may take place depends, of course, on the degree in which the mind is able to take cognizance of resemblances or analogies. Hence the process can go much further in the conceptual sphere than it does in the receptual. But the important point is that it unquestionably takes place in the latter within certain limits. Nor is this anything more than we should antecedently expect. For in the lengthy account and from the numerous facts which I gave of the receptual intelligence of brutes, it was abundantly proved that long before the differential engine of conception has come to the assistance of mind, mind is able to reach a high level in the distinguishing of resemblances or analogies by means of receptual discrimination alone. Consequently, it is inevitable that non-conceptual or denotative names should undergo a connotative extension, within whatever limits these powers of merely receptual discrimination impose. And, as a matter of fact, we found that such is the case. A talking bird will extend its denotative name from one dog in particular to any other dog which it may happen to see; and a young child, after having done this, will extend the denotative name still further, so as to include images, and eventually pictures, of dogs. Hence, if the receptual intelligence of a parrot were somewhat more advanced than it happens to be, we can have no doubt that it would do the same: the only reason why in this matter it parts company with a child so soon as it does, is because its receptual intelligence is not sufficiently developed to perceive the resemblance of images and pictures to the objects which they are intended to represent. But the receptual intelligence of a dog is higher than that of a parrot, and some dogs are able to perceive resemblances of this kind. Therefore if dogs, like parrots, had happened to be able to articulate, and so to learn the use of denotative names, there can be no doubt that they would have accompanied the growing child through a somewhat further reach of connotative utterance than is the case with the only animals which present the anatomical conditions required for the imitation of articulate sounds. Both dogs and monkeys are able, in an extraordinary degree, to _understand_ these sounds: that is to say, they can learn the meanings of an astonishing number of denotative names, and also