but slightly acquainted. And, of course, the smaller the vocabulary, the greater is such necessity; so that it is greatest of all when an infant is only just emerging from its infancy. Therefore, as just remarked, on grounds of psychological analysis alone, I do not think we should be justified in concluding that the first-speaking child has no appreciation of what we understand by parts of speech; and it is on account of the uncertainty which here obtains as between necessity and incapacity, that I reserved my consideration of “sentence-words” for the independent light which has been thrown upon them by the science of comparative philology. Now, when investigated by this light, it appears, as already observed, that the protoplasmic condition of language prior to its differentiation into parts of speech was of much longer duration in the race than, relatively speaking, it is in the individual. Moreover, it appears to have been of relatively much greater importance to the subsequent development of language. How, then, is this difference to be explained? I think the explanation is sufficiently simple. An infant of to-day is born into the medium of already-spoken language; and long before it is itself able to imitate the words which it hears, it is well able to understand a large number of them. Consequently, while still literally an _infant_, the use of grammatical forms is being constantly borne in upon its mind; and, therefore, it is not at all surprising that, when it first begins to use articulate signs, it should already be in possession of some amount of knowledge of their distinctive meanings as names of objects, qualities, actions, states, or relations. Indeed, it is only as such that the infant has acquired its knowledge of these signs at all; and hence, if there is any wonder in the matter, it is that the first-speaking child should exhibit so much vagueness as it does in the matter of grammatical distinction.