such words as ĝin, man, kiai, many, to form the plural ğin kiai, to the perfect grammar of Sanskrit and Greek, everything is intelligible as the result of the two principles of growth which we considered in our second Lecture. What is antecedent to the production of roots is the work of nature; what follows after is the work of man, not in his individual and free, but in his collective and moderating, capacity.

I do not say that every form in Greek or Sanskrit has as yet been analyzed and explained. There are formations in Greek and Latin and English which have hitherto baffled all tests; and there are certain contrivances, such as the augment in Greek, the change of vowels in Hebrew, the Umlaut and Ablaut in the Teutonic dialects, where we might feel inclined to suppose that language admitted distinctions purely musical or phonetic, corresponding to very palpable and material distinctions of thought. Such a supposition, however, is not founded on any safe induction. It may seem inexplicable to us why bruder in German should form its plural as brüder; or brother, brethren. But what is inexplicable and apparently artificial in our modern languages becomes intelligible in their more ancient phases. The change of u into ü, as in bruder, brüder, was not intentional; least of all was it introduced to expressed plurality. The change is phonetic, and due to the influence of an i or j1 which existed originally in the last syllable and which reacted regularly on the vowel of the preceding syllable; nay, which leaves its effect behind, even after it has itself disappeared. By a false analogy such a change, perfectly justifiable in a

1 See Schleicher, Deutsche Sprache, p. 144.

certain class of words, may be applied to other words where no such change was called for; and it may then appear as if an arbitrary change of vowels was intended to convey a grammatical change. But even into these recesses the comparative philologist can follow language, thus discovering a reason even for what in reality was irrational and wrong. It seems difficult to believe that the augment in Greek should originally have had an independent substantial existence, yet all analogy is in favor of such a view. Suppose English had never been written down before Wycliffe's time, we should then find that in some instances the perfect was formed by the mere addition of a short a. Wycliffe spoke and wrote: I knowlech to a felid and seid pus; i. e. I acknowledge to have felt and said thus. In a similar way we read: it should a fallen; instead of it should have fallen;" and in some parts of England common people still say very much the same: I should a done it. Now in some old English books this a actually coalesces with the verb, at least they are printed together; so that a grammar founded on them would give us "to fall" as the infinitive of the present, to afallen as the infinitive of the past. I do not wish for a moment to be understood as if there was any connection between this a, a contraction of have in English, and the Greek augment which is placed before past tenses. All I mean is, that, if the origin of the augment has not yet been satisfactorily explained, we are not therefore to despair, or to admit an arbitrary addition of a consonant or vowel, used as it were algebraically or by mutual agreement, to distinguish a past from a present tense.

1 Marsh, p. 388.

If inductive reasoning is worth anything, we are justified in believing that what has been proved to be true on so large a scale, and in cases where it was least expected, is true with regard to language in general. We require no supernatural interference, nor any conclave of ancient sages, to explain the realities of human speech. All that is formal in language is the result of rational combination; all that is material, the result of a mental instinct. The first natural and instinctive utterances, if sifted differently by different clans, would fully account both for the first origin and for the first divergence of human speech. We can understand not only the origin of language, but likewise the necessary breaking up of one language into many; and we perceive that no amount of variety in the material or the formal elements of speech is incompatible with the admission of one common source.

The Science of Language thus leads us up to that highest summit from whence we see into the very dawn of man's life on earth; and where the words which we have heard so often from the days of our childhood"And the whole earth was of one language and of one speech" assume a meaning more natural, more intelligible, more convincing, than they ever had before.

And now in concluding this course of Lectures, I have only to express my regret that the sketch of the Science of Language which I endeavored to place before you, was necessarily so very slight and imperfect. There are many points which I could not touch at all, many which I could only allude to: there is hardly one to which I could do full justice. Still I feel grateful to the President and the Council of this Institution

for having given me an opportunity of claiming some share of public sympathy for a science which I believe has a great future in store; and I shall be pleased, if, among those who have done me the honor of attending these Lectures, I have excited, though I could not have satisfied, some curiosity as to the strata which underlie the language on which we stand and walk; and as to the elements which enter into the composition of the very granite of our thoughts.

APPENDIX.