botany; that it grows and changes entirely independently of the will of man, being governed by laws which are not under his control; that "although a poet may knowingly and intentionally invent a new word, its acceptance depends on circumstances which defy individual interference." He appears to us here to mingle things which should be kept distinct. In speaking of the growth and changes of a language, he is speaking of the inflexional system; this undoubtedly, having been originally not consciously invented, but blindly evolved, is changed from time to time by an instinct rather than with a purpose, and therefore is removed away from the mental sciences. The invention of new words, on the other hand, stands on a different footing, and is a conscious act, and the acceptance is no less a conscious act; and they consequently belong to human, not to natural history. We have seen the word " telegram,"-lately invented by a newspaper, because it was found convenient, and its analogy to diagram, anagram, program, seemed to justify it,-meet with universal acceptation. And it is also essential to notice that the two processes here mentioned belong to entirely different stages of society, and that the one excludes the other; the change of the inflexional system belonging to a period destitute of literature, and when men spoke with as little consciousness of the principles of the process as they breathed; but the invention and acceptance of new words, and the establishment of old words in peculiar significations, belonging to an age of literature, when the practice of individuals does influence the language very considerably witness Homer, Luther, Shakspere, Johnson, Carlyle. But about the former Müller is quite in the right; and it is important to understand this clearly, since only on this hypothesis can we expect to find language, in the changes of its inflexional system, conforming to laws as clear and as binding as those of chemistry.

It will be observed that the scientific comparison of languages originated, and has at present been successfully prosecuted, on one field only-that of the Aryan languages; and that there the feature which gives character and expression to the language, and which yields the most reliable data for comparison among various idioms, is the inflexional system. When we attempt to push the experience thus gained into other fields, however, and to determine the mutual relationship of the outlying languages, we are met by the difficulty that most of these have no

* Lest we should be supposed to intend to justify this formation, we will remind the reader that "telegraph" is from the adjective form 77λeypápos, "writing afar" (like τηλεμάχος, τηλεσκόπος, τηλεβόλος), and that " a thing written from afar, telegraphed,” must be either τηλέγραφον (proparox.) or τηλεγράφημα (cf. τηλεβολέω): whereas τηλεγράφω, and therefore τηλέγραμμα, are quite impossible forms.

inflexional system at all, at least not in the same sense. The morphological classification of languages as isolating, agglutinative, and inflexional, established by William Humboldt, is thus explained by Müller.

"As all languages, so far as we can judge at present, can be reduced in the end to roots, predicative and demonstrative, it is clear that, according to the manner in which roots are put together, we may expect to find three kinds of languages, or three stages in the gradual formation of speech.

1. Roots may be used as words, each root preserving its full independence.

2. Two roots may be joined together to form words, and in these compounds one root may lose its independence.

3. Two roots may be joined together to form words, and in these compounds both roots may lose their independence.

66

The first stage, in which each root preserves its independence, and in which there is no formal distinction between a root and a word, I call the Radical Stage. This stage is best represented by ancient Chinese. Languages belonging to this first or radical stage have sometimes been called Monosyllabic or Isolating. The second stage, in which two or more roots coalesce to form a word, the one retaining its radical independence, the other sinking down to a mere termination, I call the Terminational Stage. This stage is best represented by the Turanian family of speech; and the languages belonging to it have generally been called agglutinative, from gluten, glue. The third stage, in which roots coalesce so that neither the one nor the other retains its substantive independence, I call the Inflexional Stage. This stage is best represented by the Aryan and Semitic families, and the languages belonging to it have sometimes been distinguished by the name of organic or amalgamating."

In his earlier work on the Turanian languages Dr. Müller had thus happily indicated the difference between the two last stages:

"The difference is the same as between a compositor and a reader. The compositor puts the s to the end of a word and looks on the type s in his hand as producing the change of pound into pounds. To the reader the s has no separate existence (except on scientific reflection); the whole word expresses to him the modified idea, and in his perception the same change is produced by penny and pence as by pound and pounds."

This classification being morphological, it may or may not tally with the genealogical or historical classification,—that is, there may be many families of languages historically entirely unconnected, which nevertheless all conform to the same type of formation, say the agglutinating type or the inflexional type. And, on the other hand, it is conceivable that a language of the second stage should be developed out of one of the first,

and that one of the third should be developed out of one of the second; and so the genealogical families would overstep the limits of the morphological. This latter hypothesis is, indeed, in the highest degree probable. For the distinct word, which in the isolating languages indicates relation (as when by juxtaposition of úo, house, and li, interior, in Chinese, uo-li is employed to denote "at home"), easily loses its accent, and sinks down into a mere termination; and thus the language becomes an agglutinative one: thus uõli, if li be treated as a mere affix, is precisely similar to Becsben, "at Vienna," in Hungarian (an agglutinating language), where ben appended denotes in. And as in the agglutinating languages the root, which theoretically ought always to maintain itself perfectly sharp and distinct, unaffected by the addition of any affixes, tends with time to accommodate itself to these, and thereby to give up its separate existence, the language tends to become inflexional. These are not merely theoretical possibilities; they are, in some instances, ascertained facts.

"Though each language, as soon as it once becomes settled, retains that morphological character which it had when it first assumed its individual or national existence, it does not lose altogether the power of producing grammatical forms that belong to a higher stage. In Chinese, and particularly in Chinese dialects, we find rudimentary traces of agglutination. The li which I mentioned before as the sign of the locative has dwindled down to a mere postposition; and a modern Chinese is no more aware that li meant originally 'interior' than the Turanian is of the origin of his case-terminations. In the spoken dialects of Chinese agglutinative forms are of more frequent occurrence. Thus in the Shanghai dialect wo is, 'to speak,' as a verb; woda, 'a word.' Of woda a genitive is formed, woda-ka; a dative, pela woda; an accusative, tang woda. In agglutinative languages, again, we meet with rudimentary traces of inflexion. Thus in Tamil the root túngu, to sleep, has not retained its full integrity in the derivative túkkam, sleep."

We may advance a step farther, and affirm that this growth of an isolating language into an agglutinative, and of an agglutinative into an inflexional, affords the only satisfactory explanation of the rise of the two latter classes. One of the first achievements of Aryan comparative grammar was the establishment of a connexion between the inflexions of person and the corresponding pronouns, and between those of tense and mood, and certain auxiliary verbs. The inflexions of ἐσ-μί, ἐσ-σί, ἐσ-τί, are obviously connected with the pronouns μe, σE, TO. The σ that forms the future in Soow is the root eo, to be; and the older Attic forms φευξούμαι, and still more the Doric πραξίομες, preserve the original Sanskrit syámi, syámas, in all its purity, in which, moreover, the syllable ya (which is also used in the formation of the potential dadyam, and becomes in in Greek,

CC

Sidoinv) is the root ya, to go (like je vais dire, I am going to say, amatum iri). The d that forms the aorist, I loved, is the root of the verb to do, as appears clearly from the inflexions of the Anglo-Saxon ner-e-de, ner-ë-dest, ner-e-de, ner-e-don, nere-don, ner-e-don (I nourished), compared with dide, didest, dide, didon, didon, didon (I did), but still more from the plural of the Gothic verb to nourish,-nas-i-dédum, nas-i-dédup, nas-i-dedun (where the inflexion retains the reduplication of 'I did'). The b of ama-bo, ama-bam, is the root of the verb bhú, to be, in Sanskrit, which gives us fui in Latin. The augment denoting past time (Sanskrit a, Greek e) is probably the Sanskrit demonstrative root a, transferring the action yonder, away from the present. The syllable ya, which forms the passive in Sanskrit (with which may be compared the e of pendeo, veneo), is from the root ya, to go. These instances might be multiplied so as to cover nearly the whole field of inflexion. Now this similarity between the inflexions and separate pronouns or verbs can only be adequately accounted for by supposing a primitive composi tion of words precisely after the manner of the agglutinative languages, and that subsequently that more perfect fusion of the two words into one took place which is the characteristic of the inflecting languages. If this be the history of the elaborate inflexional system which we find at the earliest period to which we can trace back the Aryan languages,-in the Vedic Sanskrit at 1500 B. C.,-how many ages must we assume during which the earlier agglutinative language was developing this inflexional language? and how many earlier still till we can reach the original monosyllabic or isolating stage? Truly the study of comparative grammar tends to modify our conceptions of the antiquity of the history of man, in the same way as geology deals with that of the earth's crust.

To return to the point we were discussing, viz. the classification of the agglutinative languages. Dr. Müller, in another work, has devoted much learning to this subject. He admits that on many whole groups of dialects we have as yet scarcely any information; yet he thinks that enough is known to warrant us in assigning a common origin to all languages bearing this general character; and he consequently calls them by the common name "Turanian." He shows how little similarity could a priori be looked for among languages which, although spring

The only other hypothesis would be that the inflexional stage is the older, and that the separate pronouns, &c. had been formed by lopping off the inflexions, and treating them as words. We should not have thought that this idea, which is confuted by the invariable greater fullness of the separate pronouns, and by the distincter notion of action embodied in the separate verbs "to go," &c., could be entertained by any sane man, had it not been propounded by the late accomplished scholar M. Fauriel, and by Rapp.

ing from the same original centre, have diverged so far as to the shores of the Arctic Sea on the one hand, and to Malacca, South Africa, and the islands of the Pacific on the other, and for whose dissemination, therefore, a vastly longer period must be assumed than for the Aryan. And he shows how the peculiar genius of these languages causes them to diverge much more rapidly than is the case in the inflexional languages. For where the root and the relational affix are kept apart so carefully as in the languages in question, the consciousness of each as a separate word never dies, and of two synonymous affixes, one dialect may choose the one, and another the other; and where this extends over the whole relation between the languages, there may be in a few generations no similarity in the relational words to attest the common origin. And inasmuch as these languages are very copious, and one tribe adopts one of a number of synonymous roots, and another adopts another to more familiar use, and the neglected words gradually die out, a comparatively short time of separation may efface most of the radical affinities also. When it is further remembered that these languages are chiefly spoken by tribes such as the Mongols, Tatars, and Huns, which rose and fell rapidly, generally led a nomad life, had little intercourse with the world beyond their tribe, and no history or literature to fix the language at one point and prevent further changes, sufficient reasons are presented to account for a vastly wider divergency than we find in the Aryan languages, which have generally been spoken by settled nations, and whose earliest accessible language is a literary one. If, then, we admit neither the root nor the relational affixes as means of comparison, but conceive it possible for languages to have a common origin in which neither of these display any common features, what is left upon which we can base our argument for community of origin? Dr. Müller relies upon the agglutinative character itself. And so, when he has satisfied himself that this is the character of the languages of Japan, of Malacca, or of Polynesia, he unhesitatingly includes these in his great Turanian family. Yet if we accept the idea previously explained, that the three classes of isolating, agglutinative, and inflexional languages are not generically distinct, but are rather three stages, which may be run through in succession by the descendants of one race, it is manifest that the common agglutinative character creates not even a presumption of community of origin, but is equivalent only to a declaration of the languages standing on the same level of civilization. And that Dr. Müller still holds to the idea of those three classes being not generically distinct, is evident from his calling them the radical stage, terminational stage, &c.

The fact seems to us to be, that the larger the number of