all kinds, and that even talking birds are able to distinguish between names as severally names of objects, qualities, states, or actions. Thus we find that to almost any order of intelligence which is already surrounded by the medium of spoken language, the understanding—and, in the presence of any power of imitative utterance, the acquisition—of denotative names as signs or marks of corresponding objects, qualities, &c., is, if anything, a more primitive act than that of using a sentence-word; but that in the absence of such an already-existing medium, sentence-words are more primitive than denotative names. Nevertheless, it is of importance to note how low an order of receptual ideation is capable of learning a denotative name by special association, because this fact proves that as soon as mankind advanced to the stage where they first began to coin their sentence-words, they must already have been far above the psychological level required for the acquisition of denotative words, _if only such words had previously been in existence_. Consequently, we can well understand how such words would soon have begun to come into existence through the habitual employment of sentence-words in relation to particular objects, qualities, states, actions, &c.; by such special associations, sentence-words would readily degenerate into merely semiotic marks. How long or how short a time this genesis of relatively “empty words” out of the primordially “full words” may have occupied, it is now impossible to say; but the important thing for us to notice is, that during the whole of this time—whatever it may have been—the mind of primitive man was already far above the psychological level which is required for the apprehension of a denotative name.[264] So much, then, for the first class of considerations which has been opened up by throwing upon the results of our psychological analysis the independent light of philological research. I will now pass on to a second class, which is even of more importance.