Part 22
At the end of our morphological classification a problem presents itself, which we might have declined to enter upon if we had confined ourselves to a genealogical classification. At the end of our genealogical classification we had to confess that only a certain number of languages had as yet been arranged genealogically, and that therefore the time for approaching the problem of the common origin of all languages had not yet come. Now, however, although we have not specified all languages which belong to the radical, the terminational, and inflectional classes, we have clearly laid it down as a principle, that all languages must fall under one or the other of these three categories of human speech. It would not be consistent, therefore, to shrink from the consideration of a problem, which, though beset with many difficulties, cannot be excluded from the science of language.
Let us first see our problem clearly and distinctly. The problem of the common origin of languages has no necessary connection with the problem of the common origin of mankind. If it could be proved that languages had had different beginnings, this would in nowise necessitate the admission of different beginnings of the human race. For if we look upon language as natural to man, it might have broken out at different times and in different countries among the scattered descendants of one original pair; if, on the contrary, language is to be treated as an artificial invention, there is still less reason why each succeeding generation should not have invented its own idiom.
Nor would it follow, if it could be proved that all the dialects of mankind point to one common source, that therefore the human race must descend from one pair. For language might have been the property of one favored race, and have been communicated to the other races in the progress of history.
The science of language and the science of ethnology have both suffered most seriously from being mixed up together. The classification of races and languages should be quite independent of each other. Races may change their languages, and history supplies us with several instances where one race adopted the language of another. Different languages, therefore, may be spoken by one race, or the same language may be spoken by different races; so that any attempt at squaring the classification of races and tongues must necessarily fail.
Secondly, the problem of the common origin of languages has no connection with the statements contained in the Old Testament regarding the creation of man, and the genealogies of the patriarchs. If our researches led us to the admission of different beginnings for the languages of mankind, there is nothing in the Old Testament opposed to this view. For although the Jews believed that for a time the whole earth was of one language and of one speech, it has long been pointed out by eminent divines, with
## particular reference to the dialects of America, that new languages might
have arisen at later times. If, on the contrary, we arrive at the conviction that all languages can be traced back to one common source, we could never think of transferring the genealogies of the Old Testament to the genealogical classification of language. The genealogies of the Old Testament refer to blood, not to language, and as we know that people, without changing their name, did frequently change their language, it is clearly impossible that the genealogies of the Old Testament should coincide with the genealogical classification of languages. In order to avoid a confusion of ideas, it would be preferable to abstain altogether from using the same names to express relationship of language which in the Bible are used to express relationship of blood. It was usual formerly to speak of _Japhetic_, _Hamitic_ and _Semitic_ languages. The first name has now been replaced by _Aryan_, the second by _African_; and though the third is still retained, it has received a scientific definition quite different from the meaning which it would have in the Bible. It is well to bear this in mind, in order to prevent not only those who are forever attacking the Bible with arrows that cannot reach it, but likewise those who defend it with weapons they know not how to wield, from disturbing in any way the quiet progress of the science of language.
Let us now look dispassionately at our problem. The problem of the possibility of a common origin of all languages naturally divides itself into two parts, the _formal_ and the _material_. We are to-day concerned with the formal part only. We have examined all possible forms which language can assume, and we have now to ask, can we reconcile with these three distinct forms, the radical, the terminational, and the inflectional, the admission of one common origin of human speech? I answer decidedly, Yes.
The chief argument that has been brought forward against the common origin of language is this, that no monosyllabic or radical language has ever entered into an agglutinative or terminational stage, and that no agglutinative or terminational language has ever risen to the inflectional stage. Chinese, it is said, is still what it has been from the beginning; it has never produced agglutinative or inflectional forms; nor has any Turanian language ever given up the distinctive feature of the terminational stage, namely, the integrity of its roots.
In answer to this it should be pointed out that 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.(308) 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_.(309) In agglutinative languages again, we meet with rudimentary traces of inflection. Thus in Tamil the root _tûngu_, to sleep, has not retained its full integrity in the derivative _tûkkam_, sleep.
I mention these instances, which might be greatly multiplied, in order to show that there is nothing mysterious in the tenacity with which each language clings in general to that stage of grammar which it had attained at the time of its first settlement. If a family, or a tribe, or a nation, has once accustomed itself to express its ideas according to one system of grammar, that first mould remains and becomes stronger with each generation. But, while Chinese was arrested and became traditional in this very early stage the radical, other dialects passed on through that stage, retaining their pliancy. They were not arrested, and did not become traditional or national, before those who spoke them had learnt to appreciate the advantage of agglutination. That advantage being once perceived, a few single forms in which agglutination first showed itself would soon, by that sense of analogy which is inherent in language, extend their influence irresistibly. Languages arrested in that stage would cling with equal tenacity to the system of agglutination. A Chinese can hardly understand how language is possible, unless every syllable is significative; a Turanian despises every idiom in which each word does not display distinctly its radical and significative element; whereas, we who are accustomed to the use of inflectional languages, are proud of the very grammar which a Chinese and Turanian would treat with contempt.
The fact, therefore, that languages, if once settled, do not change their grammatical constitution, is no argument against our theory, that every inflectional language was once agglutinative, and every agglutinative language was once monosyllabic. I call it a theory, but it is more than a theory, for it is the only possible way in which the realities of Sanskrit or any other inflectional language can be explained. As far as the formal part of language is concerned, we cannot resist the conclusion that what is now _inflectional_ was formerly _agglutinative_, and what is now _agglutinative_ was at first _radical_. The great stream of language rolled on in numberless dialects, and changed its grammatical coloring as it passed from time to time through new deposits of thought. The different channels which left the main current and became stationary and stagnant, or, if you like, literary and traditional, retained forever that coloring which the main current displayed at the stage of their separation. If we call the radical stage _white_, the agglutinative _red_, and the inflectional _blue_, then we may well understand why the white channels should show hardly a drop of red or blue, or why the red channels should hardly betray a shadow of blue; and we shall be prepared to find what we do find, namely, white tints in the red, and white and red tints in the blue channels of speech.
You will have perceived that in what I have said I only argue for the possibility, not for the necessity, of a common origin of language.
I look upon the problem of the common origin of language, which I have shown to be quite independent of the problem of the common origin of mankind, as a question which ought to be kept open as long as possible. It is not, I believe, a problem quite as hopeless as that of the plurality of worlds, on which so much has been written of late, but it should be treated very much in the same manner. As it is impossible to demonstrate by the evidence of the senses that the planets are inhabited, the only way to prove that they are, is to prove that it is impossible that they should not be. Thus on the other hand, in order to prove that the planets are not inhabited, you must prove that it is impossible that they should be. As soon as the one or the other has been proved, the question will be set at rest: till then it must remain an open question, whatever our own predilections on the subject may be.
I do not take quite as desponding a view of the problem of the common origin of language, but I insist on this, that we ought not to allow this problem to be in any way prejudged. Now it has been the tendency of the most distinguished writers on comparative philology to take it almost for granted, that after the discovery of the two families of language, the Aryan and Semitic, and after the establishment of the close ties of relationship which unite the members of each, it would be impossible to admit any longer a common origin of language. It was natural, after the criteria by which the unity of the Aryan as well as the Semitic dialects can be proved had been so successfully defined, that the absence of similar coincidences between any Semitic and Aryan language, or between these and any other branch of speech, should have led to a belief that no connection was admissible between them. A Linnæan botanist, who has his definite marks by which to recognize an Anemone, would reject with equal confidence any connection between the species Anemone and other flowers which have since been classed under the same head though deficient in the Linnæan marks of the Anemone.
But there are surely different degrees of affinity in languages as well as in all other productions of nature, and the different families of speech, though they cannot show the same signs of relationship by which their members are held together, need not of necessity have been perfect strangers to each other from the beginning.
Now I confess that when I found the argument used over and over again, that it is impossible any longer to speak of a common origin of language, because comparative philology had proved that there existed various families of language, I felt that this was not true, that at all events it was an exaggeration.
The problem, if properly viewed, bears the following aspect:—“_If you wish to assert that language had various beginnings, you must prove it impossible that language could have had a common origin._”
No such impossibility has ever been established with regard to a common origin of the Aryan and Semitic dialects; while on the contrary the analysis of the grammatical forms in either family has removed many difficulties, and made it at least intelligible how, with materials identical or very similar, two individuals, or two families, or two nations, could in the course of time have produced languages so different in form as Hebrew and Sanskrit.
But still greater light was thrown on the formative and metamorphic process of language by the study of other dialects unconnected with Sanskrit or Hebrew, and exhibiting before our eyes the growth of those grammatical forms (grammatical in the widest sense of the word) which in the Aryan and Semitic families we know only as formed, not as forming; as decaying, not as living; as traditional, not as understood and intentional: I mean the Turanian languages. The traces by which these languages attest their original relationship are much fainter than in the Semitic and Aryan families, but they are so of necessity. In the Aryan and Semitic families, the agglutinative process, by which alone grammatical forms can be obtained, has been arrested at some time, and this could only have been through religious or political influences. By the same power through which an advancing civilization absorbs the manifold dialects in which every spoken idiom naturally represents itself, the first political or religious centralization must necessarily have put a check on the exuberance of an agglutinative speech. Out of many possible forms one became popular, fixed, and technical for each word, for each grammatical category; and by means of poetry, law, and religion, a literary or political language was produced to which thenceforth nothing had to be added; which in a short time, after becoming unintelligible in its formal elements, was liable to phonetic corruption only, but incapable of internal resuscitation. It is necessary to admit a primitive concentration of this kind for the Aryan and Semitic families, for it is thus only that we can account for coincidences between Sanskrit and Greek terminations, which were formed neither from Greek nor from Sanskrit materials, but which are still identically the same in both. It is in this sense that I call these languages political or state languages, and it has been truly said that languages belonging to these families must be able to prove their relationship by sharing in common not only what is regular and intelligible, but what is anomalous, unintelligible, and dead.
If no such concentration takes place, languages, though formed of the same materials and originally identical, must necessarily diverge in what we may call dialects, but in a very different sense from the dialects such as we find in the later periods of political languages. The process of agglutination will continue in each clan, and forms becoming unintelligible will be easily replaced by new and more intelligible compounds. If the cases are formed by postpositions, new postpositions can be used as soon as the old ones become obsolete. If the conjugation is formed by pronouns, new pronouns can be used if the old ones are no longer sufficiently distinct.
Let us ask then, what coincidences we are likely to find in agglutinative dialects which have become separated, and which gradually approach to a more settled state? It seems to me that we can only expect to find in them such coincidences as Castrén and Schott have succeeded in discovering in the Finnic, Turkic, Mongolic, Tungusic, and Samoyedic languages; and such as Hodgson, Caldwell, Logan, and myself have pointed out in the Tamulic, Gangetic, Lohitic, Taïc, and Malaïc languages. They must refer chiefly to the radical materials of language, or to those parts of speech which it is most difficult to reproduce, I mean pronouns, numerals, and prepositions. These languages will hardly ever agree in what is anomalous or inorganic, because their organism repels continually what begins to be formal and unintelligible. It is astonishing rather, that any words of a conventional meaning should have been discovered as the common property of the Turanian languages, than that most of their words and forms should be peculiar to each. These coincidences must, however, be accounted for by those who deny the common origin of the Turanian languages; they must be accounted for, either as the result of accident, or of an imitative instinct which led the human mind everywhere to the same onomatopoëtic formations. This has never been done, and it will require great efforts to achieve it.
To myself the study of the Turanian family was interesting particularly because it offered an opportunity of learning how far languages, supposed to be of a common origin, might diverge and become dissimilar by the unrestrained operation of dialectic regeneration.
In a letter which I addressed to my friend, the late Baron Bunsen, and which was published by him in his “Outlines of the Philosophy of Universal History”(310) (vol. i. pp. 263-521), it had been my object to trace, as far as I was able, the principles which guided the formation of agglutinative languages, and to show how far languages may become dissimilar in their grammar and dictionary, and yet allow us to treat them as cognate dialects. In answer to the assertion that it was impossible, I tried, in the fourth, fifth, and sixth sections of that Essay, to show _how_ it was possible, that, starting from a common ground, languages as different as Mandshu and Finnish, Malay and Siamese, should have arrived at their present state, and might still be treated as cognate tongues. And as I look upon this process of agglutination as the only intelligible means by which language can acquire a grammatical organization, and clear the barrier which has arrested the growth of the Chinese idiom, I felt justified in applying the principles derived from the formation of the Turanian languages to the Aryan and Semitic families. They also must have passed through an agglutinative stage, and it is during that period alone that we can account for the gradual divergence and individualization of what we afterwards call the Aryan and Semitic forms of speech. If we can account for the different appearance of Mandshu and Finnish, we can also account for the distance between Hebrew and Sanskrit. It is true that we do not know the Aryan speech during its agglutinative period, but we can infer what it was when we see languages like Finnish and Turkish approaching more and more to an Aryan type. Such has been the advance which Turkish has made towards inflectional forms, that Professor Ewald claims for it the title of a synthetic language, a title which he gives to the Aryan and Semitic dialects after they have left the agglutinative stage, and entered into a process of phonetic corruption and dissolution. “Many of its component parts,” he says, “though they were no doubt originally, as in every language, independent words, have been reduced to mere vowels, or have been lost altogether, so that we must infer their former presence by the changes which they have wrought in the body of the word. _Göz_ means eye, and _gör_, to see; _ish_, deed, and _ir_, to do; _îtsh_, the interior, _gîr_, to enter.”(311) Nay, he goes so far as to admit some formal elements which Turkish shares in common with the Aryan family, and which therefore could only date from a period when both were still in their agglutinative infancy. For instance, _di_, as exponent of a past action; _ta_, as the sign of the past participle of the passive; _lu_, as a suffix to form adjectives, &c.(312) This is more than I should venture to assert.
Taking this view of the gradual formation of language by agglutination, as opposed to intussusception, it is hardly necessary to say that, if I speak of a Turanian family of speech, I use the word family in a different sense from that which it has with regard to the Aryan and Semitic languages. In my Letter on the Turanian languages, which has been the subject of such fierce attacks from those who believe in different beginnings of language and mankind, I had explained this repeatedly, and I had preferred the term of _group_ for the Turanian languages, in order to express as clearly as possible that the relation between Turkish and Mandshu, between Tamil and Finnish, was a different one, not in degree only, but in kind, from that between Sanskrit and Greek. “These Turanian languages,” I said (p. 216), “cannot be considered as standing to each other in the same relation as Hebrew and Arabic, Sanskrit and Greek.” “They are radii diverging from a common centre, not children of a common parent.” And still they are not so widely distant as Hebrew and Sanskrit, because none of them has entered into that new phase of growth or decay (p. 218) through which the Semitic and Aryan languages passed after they had been settled, individualized, and nationalized.
The real object of my Essay was therefore a defensive one. It was to show how rash it was to speak of different independent beginnings in the history of human speech, before a single argument had been brought forward to establish the necessity of such an admission. The impossibility of a common origin of language has never been proved, but, in order to remove what were considered difficulties affecting the theory of a common origin, I felt it my duty to show practically, and by the very history of the Turanian languages, how such a theory was possible, or as I say in one instance only, probable. I endeavored to show how even the most distant members of the Turanian family, the one spoken in the north, the other in the south of Asia, the _Finnic_ and the _Tamulic_, have preserved in their grammatical organization traces of a former unity; and, if my opponents admit that I have proved the ante-Brahmanic or Tamulic inhabitants of India to belong to the Turanian family, they can hardly have been aware that if this, the most extreme point of my argument be conceded, everything else is involved, and must follow by necessity.
Yet I did not call the last chapter of my Essay, “On the Necessity of a common origin of Language,” but “On the Possibility;” and, in answer to the opinions advanced by the opposite party, I summed up my defence in these two paragraphs:—
I.
“Nothing necessitates the admission of different independent beginnings for the _material_ elements of the Turanian, Semitic, and Aryan branches of speech;—nay, it is possible even now to point out radicals which, under various changes and disguises, have been current in these three branches ever since their first separation.”
II.
“Nothing necessitates the admission of different beginnings for the formal elements of the Turanian, Semitic, and Aryan branches of speech;—and though it is impossible to derive the Aryan system of grammar from the Semitic, or the Semitic from the Aryan, we can perfectly understand how, either through individual influences, or by the wear and tear of speech in its own continuous working, the different systems of grammar of Asia and Europe may have been produced.”