But having thus resolved words into their elements, we reach forms which obstinately resist all further analysis, and hence are treated as radical. Words, then, are regarded as made up of two roots: Predicative roots, which form the body of the word, and Demonstrative or pronominal roots, which specialize or limit the others by adjuncts of place, number, person, time, &c. These roots Prof. Müller divides into three classes: Primary, consisting of a vowel, or of a vowel preceding or following a consonant: Secondary, of two consonants inclosing a vowel; and Tertiary, of two consonants and a vowel, of a vowel and two consonants, of two consonants a vowel and a consonant, or of two consonants, a vowel, and two consonants. The primary roots are the most important; but their general import is slightly modified by the consonants of the secondary and tertiary roots. From these roots, flexible and prolific, expressing always general ideas which admit of indefinite analogous applications, a vast and all-sufficient frame-work of language is easily constructed. We have but to take the Latin root spec, and its Greek cognate axeл, to see the vast flexibility and pliancy of one of these predicative roots. Run it through its innumerable Greek derivations, whence come sceptic, scope, episcopal; through the vast range of Latin forms, as species, special, specific, respectable, speculation, circumspect, expect, auspicious; see it turning up, by accidental association, in the French espiègle, waggish, and épicier, a grocer; in the Italian spezieria, an apothecary's shop, and the English spices; and we get a little idea of the thriftiness of Mother Nature in turning to account the linguistic materials entrusted to her ménage. Sanscrit grammarians have reduced the whole growth of their language to about 1700 roots, which Prof Müller thinks may be still reduced to about 500. The Hebrew has about as many, and the Chinese is satisfied with about 450. This, says Prof. Müller, shows a wise spirit of economy in primitive language, for the possibility of forming new roots was practically unlimited, since with only twenty-four letters, the possible number of biliteral and triliteral roots is about 15,000. Now from 500 roots, assuming that each root gave rise to fifty words, we should get a vocabulary of 25,000. Yet a well educated Englishman

rarely uses more than 3,000 or 4,000 words, in actual conversation, although eloquent speakers, and acute dialecticians may rise to about 10,000. The Old Testament utters its revelations with less than 6,000; Milton contents himself with 8,000; and the limitless range of Shakspeare's genius employs but 15,000. But along with these predicative elements, we also need the demonstrative elements, for the formation of words. We must have not only the root luc, shine, but this, with the pronominal elements, luc-s, shining there, light; with the second person of the pronoun, luc-e-s, thou shinest. These two elements are essential to make up language; and it is absolutely certain that every inflected language was once in a state in which both these elements were clearly visible.

Armed with this doctrine of roots, we may now resume our work of classification, and make our onset upon those languages which were proof against any verbal affiliation. Nothing, we have seen, is more changeable than the words of a language when bound down by no political organization, and by no written literature. It creates and throws away forms with reckless prodigality, and hence it is in no way surprising that the dialects of barbarous and nomad tribes should display slight traces of verbal affinity. But they cannot emancipate themselves from the general law of language. They cannot form words on radically different principles from those which have been already examined. All speech must consist originally of these two classes of roots—the general, fundamental or predicative, and the demonstrative or specializing. These roots may co-exist in three different ways: 1. Both the roots may remain entirely unmodified and distinct. 2. One of the roots may retain its integrity, and the other be worn away or modified. 3. Both the roots may have been so modified, worn away, or blended with each other, that their distinct form is lost. The first class, in which both roots are distinct, is called monosyllabic; the second, in which the changeable ending is as it were glued to the root, is called agglutinated; the third, in which both the parts are liable to be fused entirely into a new compound, is called inflected. We have here a principle of classification absolutely exhaustive. In some

one of these stages all languages must be found; those which are in the second have reached it by passing through the first, and those in the third by passing through the first and second. Of the first, or monosyllabic stage, the chief specimen is the Chinese: to the third, or inflexional stage, belong all the Indo-European languages: while the so-called Turanian, comprising the numberless dialects spread over Central and Northern Asia, and a few sporadic European tongues, islanded in the encompassing ocean of Arian speech, belong to the agglutinated class.

If it is doubted whether, as these seem to have the fixedness of a permanent type, they are not in reality three originally distinct classes of speech, it is answered that it can be demonstrably shown that the inflected languages must have once existed as agglutinated and monosyllabic; that the Chinese exhibits symptoms of that abrasion which approximates it in some points to the agglutinated; while some of the Turanian dialects, as the Turkic and Finnic, having considerable literary culture, have also so far taken on the inflected form as to stand nearly midway between the inflected and the ruder Turanian dialects.

Thus the languages which escape the genealogical classification, submit to the morphological. Where they do not display affinity of words, they display affinity of origin and original construction; and when we add that the same principle applies to languages which, like many of the African and American, are polysynthetic, or consist of many roots, we find we have now a principle of classification which applies absolutely to all languages.

Having reached an ultimate analysis, by virtue of which we may unite all languages under this common law of classification, we are prepared to see what light our science sheds on the question of the common or diverse origin of human speech. Prof. Müller insists on keeping the problem distinct. He will confound it with no theological, and with no other scientific questions. It is not an inquiry into ethnical affinities, nor into the original unity of the human race. Men may have had, he thinks, diverse origins, and yet the language

of one people have gained an ascendency over and supplanted all others. Or all nations may have sprung from one original source, and yet language have broken out at different times, and with a complete diversity of verbal forms. Thus the ́original unity of language is really a question by itself alone.

Both of the above positions are undoubtedly conceivable. But they are so violently improbable, that, like the infinitesimal elements in fluxions, they may with entire safety be neglected in our calculations. Men will always regard the questions of the original unity of language and the original unity of the race, as de facto identical. If, indeed, we accept a theory which Prof. Müller unconditionally and contemptuously rejects, viz., that language is a matter of arbitrary agreement, the product of a body of speechless sages, agreeing together to originate a vehicle of inter-communication, then we see not why there might not be as many independent tongues as there were bodies of mute bipeds who should happen to combine for this high purpose. But if such a theory is as absurd as, with Prof. Müller, we suppose it to be, and if language is really a natural and necessary product of the unique human organism, then the two questions can never be so disengaged from each other that the solution of one will not involve the virtual solution of both. Meantime we concur entirely with Prof. Müller in the propriety of examining the question wholly apart from any extrinsic considerations, and especially from any theological consequences supposed to be involved in it. Science must reign supreme in its own department. It must be allowed to tell its own story to render in its own unforced and unbiassed testimony. The moment that we bring either fear or favor to bear on the witness whom we have placed on the stand, his testimony is worthless in our favor, while we raise a violent suspicion against the soundness of the cause that needs such supports. Prof. Müller, though a firm believer in the common origin both of language and of the race, yet concedes that, scientifically, the question is yet unresolved. He contends, however, that from the diversity of languages, no impossibility of their common origin has been or can be shown, the nature and degree of their diversity be

ing perfectly otherwise accounted for. On the kindred subject of the common origin of the race, he says, incidentally, and, as we think, with great justice:

"If I am told that no quiet observer would ever have conceived the idea of deriving all mankind from one pair, unless the Mosaic records had taught it; I must be allowed to say in reply, that this idea, on the contrary, is so natural, so consistent with all human laws of reasoning, that, so far as I know, there has been no nation on earth which, if it preserved any traditions on the origin of mankind, did not derive the human race from one pair, if not from one person. The author of the Mosaic records, therefore, though stripped before the tribunal of Physical Science of his claims as an inspired writer, may at least claim the modest title of a quiet observer; and if his conception of the physical unity of the human race can be proved to be an error, it is an error which he shares in common with other quiet observers, such as Humboldt, Bunsen, Prichard, and Owen."

But passing from the inductive, we reach the last and theoretical stage of our science, which deals with the ultimate nature and origin of the radical elements into which our complex speech has been resolved. What are these roots, so wondrous in their character, and how and whence were they originated? In other words, we now take in hand the difficult problem of the origin of human speech.

1. Let us state the different suppositions. Is language the product of conventional agreement-men going to work to construct it as they would to build a city-the deliberate product of the felt need of an instrument of social intercourse?

2. Is language a direct gift from God, superinduced upon man's primary faculties, as a sort of dowry with which to set up his rational and social house-keeping?

3. Is language the immediate, spontaneous, natural product of his blended material and rational organism, and does he speak by the same rational instinct by which he thinks? Is it a part of the very idea of man to speak, as of a singing bird to sing?

The doctrines involved in the two first questions, held as they have been by many very able men, may be regarded as effectually exploded. Language is not a matter of subsequent convention, for men must have spoken before they could have agreed to speak. And certainly they must have been created