Chapter 2

ON THE STRATIFICATION OF LANGUAGE.

There are few sensations more pleasant than that of wondering. We have all experienced it in childhood, in youth, and in our manhood, and we may hope that even in our old age this affection of the mind will not entirely pass away. If we analyze this feeling of wonder carefully, we shall find that it consists of two elements. What we mean by wondering is not only that we are startled or stunned,--that I should call the merely passive element of wonder. When we say “I wonder,” we confess that we are taken aback, but there is a secret satisfaction mixed up with our feeling of surprise, a kind of hope, nay, almost of certainty, that sooner or later the wonder will cease, that our senses or our mind will recover, will grapple with these novel impressions or experiences, grasp them, it may be, throw them, and finally triumph over them. In fact we wonder at the riddles of nature, whether animate or inanimate, with a firm conviction that there is a solution to them all, even though we ourselves may not be able to find it.

Wonder, no doubt, arises from ignorance, but from a peculiar kind of ignorance; from what might be called a fertile ignorance: an ignorance which, if we look back at the history of most of our sciences, will be found to have been the mother of all human knowledge. For thousands of years men have looked at the earth with its stratifications, in some places so clearly mapped out; for thousands of years they must have seen in their quarries and mines, as well as we ourselves, the imbedded petrifications of organic creatures: yet they looked and passed on without thinking more about it--they did not wonder. Not even an Aristotle had eyes to see; and the conception of a science of the earth, of Geology, was reserved for the eighteenth century.

Still more extraordinary is the listlessness with which during all the centuries that have elapsed since the first names were given to all cattle, and to the fowl of the air, and to every beast of the field, men have passed by what was much nearer to them than even the gravel on which they trod, namely, the words of their own language. Here, too, the clearly marked lines of different strata seemed almost to challenge attention, and the pulses of former life were still throbbing in the petrified forms imbedded in grammars and dictionaries. Yet not even a Plato had eyes to see, or ears to hear, and the conception of a science of language, of Glottology, was reserved for the nineteenth century.

I am far from saying that Plato and Aristotle knew nothing of the nature, the origin, and the purpose of language, or that we have nothing to learn from their works. They, and their successors, and their predecessors too, beginning with Herakleitos and Demokritos, were startled and almost fascinated by the mysteries of human speech as much as by the mysteries of human thought; and what we call grammar and the laws of language, nay, all the technical terms which are still current in our schools, such as _noun_ and _verb_, _case_ and _number_, _infinitive_ and _participle_, all this was first discovered and named by the philosophers and grammarians of Greece, to whom, in spite of all our new discoveries, I believe we are still beholden, whether consciously or unconsciously, for more than half of our intellectual life.

But the interest which those ancient Greek philosophers took in language was purely philosophical. It was the form, far more than the matter of speech which seemed to them a subject worthy of philosophical speculation. The idea that there was, even in their days, an immense mass of accumulated speech to be sifted, to be analyzed, and to be accounted for somehow, before any theories on the nature of language could be safely started, hardly ever entered their minds; or when it did, as we see here and there in Plato’s “Kratylos,” it soon vanished, without leaving any permanent impression. Each people and each generation has its own problems to solve. The problem that occupied Plato in his “Kratylos” was, if I understand him rightly, the possibility of a perfect language, a correct, true, or ideal language, a language founded on his own philosophy, his own system of types or ideas. He was too wise a man to attempt, like Bishop Wilkins, the actual construction of a philosophical language. But, like Leibniz, he just lets us see that a perfect language is conceivable, and that the chief reason of the imperfections of real language must be found in the fact that its original framers were ignorant of the true nature of things, ignorant of dialectic philosophy, and therefore incapable of naming rightly what they had failed to apprehend correctly. Plato’s view of actual language, as far as it can be made out from the critical and negative rather than didactic and positive dialogue of “Kratylos,” seems to have been very much the same as his view of actual government. Both fall short of the ideal, and both are to be tolerated only in so far as they participate in the perfections of an ideal state and an ideal language.[2] Plato’s “Kratylos” is full of suggestive wisdom. It is one of those books which, as we read them again from time to time, seem every time like new books: so little do we perceive at first all that is pre-supposed in them,--the accumulated mould of thought, if I may say so, in which alone a philosophy like that of Plato could strike its roots and draw its support.

But while Plato shows a deeper insight into the mysteries of language than almost any philosopher that has come after him, he has no eyes for that marvelous harvest of words garnered up in our dictionaries, and in the dictionaries of all the races of the earth. With him language is almost synonymous with Greek, and though in one passage of the “Kratylos” he suggests that certain Greek words might have been borrowed from the Barbarians, and, more particularly from the Phrygians, yet that remark, as coming from Plato, seems to be purely ironical, and though it contains, as we know, a germ of truth that has proved most fruitful in our modern science of language, it struck no roots in the minds of Greek philosophers. How much our new science of language differs from the linguistic studies of the Greeks; how entirely the interest which Plato took in language is now supplanted by new interests, is strikingly brought home to us when we see how the _Société de Linguistique_, lately founded at Paris, and including the names of the most distinguished scholars of France, declares in one of its first statutes that “it will receive no communication concerning the origin of language or the formation of a universal language,” the very subjects which, in the time of Herakleitos and Plato, rendered linguistic studies worthy of the consideration of a philosopher.

It may be that the world was too young in the days of Plato, and that the means of communication were wanting to enable the ancient philosopher to see very far beyond the narrow horizon of Greece. With us it is different. The world has grown older, and has left to us in the annals of its various literatures the monuments of growing and decaying speech. The world has grown larger, and we have before us, not only the relics of ancient civilization in Asia, Africa, and America, but living languages in such number and variety that we draw back almost aghast at the mere list of their names. The world has grown wiser too, and where Plato could only see imperfections, the failures of the founders of human speech, we see, as everywhere else in human life, a natural progress from the imperfect towards the perfect, unceasing attempts at realizing the ideal, and the frequent triumphs of the human mind over the inevitable difficulties of this earthly condition,--difficulties, not of man’s own making, but, as I firmly believe, prepared for him, and not without a purpose, as toils and tasks, by a higher Power and by the highest Wisdom.

Let us look then abroad and behold the materials which the student of language has now to face. Beginning with the language of the Western Isles, we have at the present day, at least 100,000 words, arranged as on the shelves of a Museum, in the pages of Johnson and Webster. But these 100,000 words represent only the best grains that have remained in the sieve, while clouds of chaff have been winnowed off, and while many a valuable grain too has been lost by mere carelessness. If we counted the wealth of English dialects, and if we added the treasures of the ancient language from Alfred to Wycliffe, we should easily double the herbarium of the linguistic flora of England. And what are these Western Isles as compared to Europe; and what is Europe, a mere promontory, as compared to the vast continent of Asia; and what again is Asia, as compared to the whole inhabitable world? But there is no corner of that world that is not full of language: the very desert and the isles of the sea teem with dialects, and the more we recede from the centres of civilization, the larger the number of independent languages, springing up in every valley, and overshadowing the smallest island.

Ἴδαν ἐς πολύδενδρον ἀνὴρ ὑλατόμος ἐνθὼν Παπταίνει, παρέοντος ἄδην, πόθεν ἄρξεται ἔργω.

We are bewildered by the variety of plants, of birds, and fishes, and insects, scattered with lavish prodigality over land and sea;--but what is the living wealth of that Fauna as compared to the winged words which fill the air with unceasing music! What are the scanty relics of fossil plants and animals, compared to the storehouse of what we call the dead languages! How then can we explain it that for centuries and centuries, while collecting beasts, and birds, and fishes, and insects, while studying their forms, from the largest down to the smallest and almost invisible creatures, man has passed by this forest of speech, without seeing the forest, as we say in German, for the very number of its trees (_Man sah den Wald vor lauter Bäumen nicht_), without once asking how this vast currency could have been coined, what inexhaustible mines could have supplied the metal, what cunning hands could have devised the image and superscription,--without once wondering at the countless treasure inherited by him from the fathers of the human race?

Let us now turn our attention in a different direction. After it had been discovered that there was this great mass of material to be collected, to be classified, to be explained, what has the Science of Language, as yet, really accomplished? It has achieved much, considering that real work only began about fifty years ago; it has achieved little, if we look at what still remains to be done.

The first discovery was that languages admit of classification. Now this was a very great discovery, and it at once changed and raised the whole character of linguistic studies. Languages might have been, for all we know, the result of individual fancy or poetry; words might have been created here and there at random, or been fixed by a convention, more or less arbitrary. In that case a scientific classification would have been as impossible as it is if applied to the changing fashions of the day. Nothing can be classified, nothing can be scientifically ruled and ordered, except what has grown up in natural order and according to rational rule.

Out of the great mass of speech that is now accessible to the student of language, a number of so-called families have been separated, such as the _Aryan_, the _Semitic_, the _Ural-Altaic_, the _Indo-Chinese_, the _Dravidian_, the _Malayo-Polynesian_, the _Kafir_ or _Bâ-ntu_ in Africa, and the _Polysynthetic_ dialects of America. The only classes, however, which have been carefully examined, and which alone have hitherto supplied the materials for what we might call the Philosophy of Language, are the Aryan and the Semitic, the former comprising the languages of India, Persia, Armenia, Greece and Italy, and of the Celtic, Teutonic, and Slavonic races; the latter consisting of the languages of the Babylonians, the Syrians, the Jews, the Ethiopians, the Arabs.

These two classes include, no doubt, the most important languages of the world, if we measure the importance of languages by the amount of influence exercised on the political and literary history of the world by those who speak them. But considered by themselves, and placed in their proper place in the vast realm of human speech, they describe but a very small segment of the entire circle. The completeness of the evidence which they place before us in the long series of their literary treasures, points them out in an eminent degree as the most useful subjects on which to study the anatomy of speech, and nearly all the discoveries that have been made as to the laws of language, the process of composition, derivation, and inflexion, have been gained by Aryan and Semitic scholars.

Far be it from me, therefore, to underrate the value of Aryan and Semitic scholarship for a successful prosecution of the Science of Language. But while doing full justice to the method adopted by Semitic and Aryan scholars in the discovery of the laws that regulate the growth and decay of language, we must not shut our eyes to the fact that our field of observation has been thus far extremely limited, and that we should act in defiance of the simplest rules of sound induction, were we to generalize on such scanty evidence. Let us but clearly see what place these two so-called families, the Aryan and Semitic, occupy in the great kingdom of speech. They are in reality but two centres, two small settlements of speech, and all we know of them is their period of decay, not their period of growth, their descending, not their ascending career, their Being, as we say in German, not their Becoming (_Ihr Gewordensein, nicht ihr Werden_). Even in the earliest literary documents both the Aryan and Semitic speech appear before us as fixed and petrified. They had left forever that stage during which language grows and expands, before it is arrested in its exuberant fertility by means of religious or political concentration, by means of oral tradition, or finally by means of a written literature. In the natural history of speech, writing, or, what in early times takes the place of writing, oral tradition, is something merely accidental. It represents a foreign influence which, in natural history, can only be compared to the influence exercised by domestication on plants and animals. Language would be language still, nay, would be more truly language, if the idea of a literature, whether oral or written, had never entered men’s minds; and however important the effects produced by this artificial domestication of language may be, it is clear that our ideas of what language is in a natural state, and therefore what Sanskrit and Hebrew, too, must have been before they were tamed and fixed by literary cultivation, ought not to be formed from an exclusive study of Aryan and Semitic speech. I maintain that all that we call Aryan and Semitic speech, wonderful as its literary representatives may be, consists of neither more or less than so many varieties which all owe their origin to only two historical concentrations of wild unbounded speech; nay, however perfect, however powerful, however glorious in the history of the world,--in the eyes of the student of language, Sanskrit, Greek, and Latin, Hebrew, Arabic, and Syriac, are what a student of natural history would not hesitate to call “_monstra_,” unnatural, exceptional formations which can never disclose to us the real character of language left to itself to follow out its own laws without let or hindrance.

For that purpose a study of Chinese and the Turanian dialects, a study even of the jargons of the savages of Africa, Polynesia, and Melanesia is far more instructive than the most minute analysis of Sanskrit and Hebrew. The impression which a study of Greek and Latin and Sanskrit leaves on our minds is, that language is a work of art, most complicated, most wonderful, most perfect. We have given so many names to its outward features, its genders and cases, its tenses and moods, its participles, gerunds, and supines, that at last we are frightened at our own devices. Who can read through all the so-called irregular verbs, or look at the thousands and thousands of words in a Greek Dictionary without feeling that he moves about in a perfect labyrinth? How then, we ask, was this labyrinth erected? How did all this come to be? We ourselves, speaking the language which we speak, move about, as it were, in the innermost chambers, in the darkest recesses of that primeval palace, but we cannot tell by what steps and through what passages we arrived there, and we look in vain for the thread of Ariadne which in leading us out of the enchanted castle of our language, would disclose to us the way by which we ourselves, or our fathers and forefathers before us, entered into it.

The question how language came to be what it is has been asked again and again. Even a school-boy, if he possesses but a grain of the gift of wondering must ask himself why _mensa_ means one table, and _mensæ_ many tables; why I love should be _amo_, I am loved _amor_, I shall love _amabo_, I have loved _amavi_, I should have loved _amavissem_. Until very lately two answers only could have been given to such questions. Both sound to us almost absurd, yet in their time they were supported by the highest authorities. Either, it was said, language, and particularly the grammatical framework of language was made by _convention_, by agreeing to call one table _mensa_, and many tables _mensæ_; or, and this was Schlegel’s view, language was declared to possess an organic life, and its terminations, prefixes, and suffixes were supposed to have sprouted forth from the radicals and stems and branches of language, like so many buds and flowers. To us it seems almost incredible that such theories should have been seriously maintained, and maintained by men of learning and genius. But what better answer could they have given? What better answer has been given even now? We have learnt something, chiefly from a study of the modern dialects, which often repeat the processes of ancient speech, and thus betray the secrets of the family. We have learnt that in some of the dialects of modern Sanskrit, in Bengali for instance,[4] the plural is formed, as it is in Chinese, Mongolian, Turkish, Finnish, Burmese, and Siamese, also in the Dravidian and Malayo-Polynesian dialects, by adding a word expressive of plurality, and then appending again the terminations of the singular. We have learnt from French how a future, _je parlerai_, can be formed by an auxiliary verb: “I to speak have” coming to mean, I shall speak. We have learnt from our own language, whether English or German, that suffixes, such as _head_ in _godhead_, _ship_ in _ladyship_, _dom_ in _kingdom_ were originally substantives, having the meaning of quality, shape, and state. But I doubt whether even thus we should have arrived at a thorough understanding of the real antecedents of language, unless, what happened in the study of the stratification of the earth, had happened in the study of language. If the formation of the crust of the earth had been throughout regular and uniform, and if none of the lower strata had been tilted up, so that even those who run might read, no shaft from the surface could have been sunk deep enough to bring the geologist from the tertiary strata down to the Silurian rocks. The same in language. Unless some languages had been arrested in their growth during their earlier stages, and had remained on the surface in this primitive state exposed only to the decomposing influence of atmospheric action, and to the ill-treatment of literary cultivation, I doubt whether any scholar would have had the courage to say that at one time Sanskrit was like unto Chinese, and Hebrew no better than Malay. In the successive strata of language thus exposed to our view, we have in fact, as in Geology, the very thread of Ariadne, which, if we will but trust to it, will lead us out of the dark labyrinth of language in which we live, by the same road by which we and those who came before us, first entered into it. The more we retrace our steps, the more we advance from stratum to stratum, from story to story, the more shall we feel almost dazzled by the daylight that breaks in upon us; the more shall we be struck, no longer by the intricacy of Greek or Sanskrit grammar, but by the marvelous simplicity of the original warp of human speech, as preserved, for instance, in Chinese; by the child-like contrivances, that are at the bottom of Paulo-post Futures and Conditional Moods.

Let no one be frightened at the idea of studying a Chinese grammar. Those who can take an interest in the secret springs of the mind, in the elements of pure reason, in the laws of thought, will find a Chinese grammar most instructive, most fascinating. It is the faithful photograph of man in his leading-strings, trying the muscles of his mind, groping his way, and so delighted with his first successful grasps that he repeats them again and again. It is child’s play, if you like, but it displays, like all child’s play, that wisdom and strength which are perfect in the mouth of babes and sucklings. Every shade of thought that finds expression in the highly finished and nicely balanced system of Greek tenses, moods, and particles can be expressed, and has been expressed, in that infant language by words that have neither prefix nor suffix, no terminations to indicate number, case, tense, mood, or person. Every word in Chinese is monosyllabic, and the same word, without any change of form, may be used as a noun, a verb, an adjective, an adverb, or a particle. Thus _ta_, according to its position in a sentence, may mean great, greatness, to grow, very much, very.[5]

And here a very important observation has been made by Chinese grammarians, an observation which, after a very slight modification and expansion, contains indeed the secret of the whole growth of language from Chinese to English. If a word in Chinese is used with the _bonâ fide_ signification of a noun or a verb, it is called a _full word_ (_shi-tsé_); if it is used as a particle or with a merely determinative or formal character, it is called an _empty word_ (_hiu-tsé_[6]). There is as yet no outward difference between full and empty words in Chinese, and this renders it all the more creditable to the grammarians of China that they should have perceived the inward distinction, even in the absence of any outward signs.

Let us learn then from Chinese grammarians this great lesson, that words may become empty, and without restricting the meaning of empty words as they do, let us use that term in the most general sense, as expressive of the fact that words may lose something of their full original meaning.

Let us add to this another observation, which the Chinese could not well have made, but which we shall see confirmed again and again in the history of language, viz.: that empty words, or, as we may also call them, dead words, are most exposed to phonetic decay.

It is clear then that, with these two preliminary observations, we can imagine three conditions of language:--

1. There may be languages in which all words, both empty and full, retain their independent form. Even words which are used when we should use mere suffixes or terminations, retain their outward integrity in Chinese. Thus, in Chinese, _jin_ means man, _tu_ means crowd, _jin-tu_, man-crowd. In this compound both _jin_ and _tu_ continue to be felt as independent words, more so than in our own compound _man-kind_; but nevertheless _tu_ has become empty, it only serves to determine the preceding word _jin_, man, and tells us the quantity or number in which _jin_ shall be taken. The compound answers in intention to our plural, but in form it is wide apart from _men_, the plural of _man_.

2. Empty words may lose their independence, may suffer phonetic decay, and dwindle down to mere suffixes and terminations. Thus in Burmese the plural is formed by _to_, in Finnish, Mordvinian, and Ostiakian by _t_. As soon as _to_ ceases to be used as an independent word in the sense of number, it becomes an empty, or if you like, an obsolete word, that has no meaning except as the exponent of plurality; nay, at last, it may dwindle down to a mere letter, which is then called by grammarians the termination of the plural. In this second stage phonetic decay may well-nigh destroy the whole body of an empty word, but--and this is important--no full words, no radicals are as yet attacked by that disintegrating process.

3. Phonetic decay may advance, and does advance still further. Full words also may lose their independence, and be attacked by the same disease that had destroyed the original features of suffixes and prefixes. In this state it is frequently impossible to distinguish any longer between the radical and formative elements of words.

If we wished to represent these three stages of language algebraically, we might represent the first by RR, using R as the symbol of a root which has suffered no phonetic decay; the second, by R + ρ or ρ + R, or ρ + R + ρ, representing by ρ an empty word that has suffered phonetic change; the third, by rρ, or ρr, or ρrρ, when both full and empty words have been changed, and have become welded together into one indistinguishable mass through the intense heat of thought, and by the constant hammering of the tongue.

Those who are acquainted with the works of Humboldt will easily recognize, in these three stages or strata, a classification of language first suggested by that eminent philosopher. According to him languages can be classified as _isolating_, _agglutinative_,[7] and _inflectional_, and his definition of these three classes agrees in the main with the description just given of the three strata or stages of language.

But what is curious is that this threefold classification, and the consequences to which it leads, should not at once have been fully reasoned out, nay, that a system most palpably erroneous should have been founded upon it. We find it repeated again and again in most works on Comparative Philology, that Chinese belongs to the _isolating_ class, the Turanian languages to the _combinatory_, the Aryan and Semitic to the _inflectional_; nay, Professor Pott[8] and his school seem convinced that no evolution can ever take place from _isolating_ to _combinatory_ and from _combinatory_ to _inflectional_ speech. We should thus be forced to believe that by some inexplicable grammatical instinct, or by some kind of inherent necessity, languages were from the beginning created as _isolating_ or _combinatory_, or _inflectional_, and must remain so to the end.

It is strange that those scholars who hold that no transition is possible from one form of language to another, should not have seen that there is really no language that can be strictly called either isolating, or combinatory, or inflectional, and that the transition from one stage to another is in fact constantly taking place under our very noses. Even Chinese is not free from combinatory forms, and the more highly developed among the combinatory languages show the clearest traces of incipient inflection. The difficulty is not to show the transition of one stratum of speech into another, but rather to draw a sharp line between the different strata. The same difficulty was felt in Geology, and led Sir Charles Lyell to invent such pliant names as _Eocene_, _Meiocene_, and _Pleiocene_, names which indicate a mere dawn, a minority, or a majority of new formations, but do not draw a fast and hard line, cutting off one stratum from the other. Natural growth, and even merely mechanical accumulation and accretion, here as elsewhere, are so minute and almost imperceptible that they defy all strict scientific terminology, and force upon us the lesson that we must be satisfied with an approximate accuracy. For practical purposes Humboldt’s classification of languages may be quite sufficient, and we have no difficulty in classing any given language, according to the prevailing character of its formation, as either isolating, or combinatory, or inflectional. But when we analyze each language more carefully we find there is not one exclusively isolating, or exclusively combinatory, or exclusively inflectional. The power of composition, which is retained unimpaired through every stratum, can at any moment place an inflectional on a level with an isolating and a combinatory language. A compound such as the Sanskrit +go-duh+, cow-milking, differs little, if at all, from the Chinese _nieou-jou_, _vaccæ lac_, or in the patois of Canton, _ngau ü_, cow-milk, before it takes the terminations of the nominative, which is, of course, impossible in Chinese.

So again in English _New-town_, in Greek _Nea-polis_, would be simply combinatory compounds. Even _Newton_ would still belong to the combinatory stratum; but _Naples_ would have to be classed as belonging to the inflectional stage.

Finnish, Hungarian, Turkish, and the Dravidian languages belong in the main to the combinatory stratum; but having received a considerable amount of literary cultivation, they all alike exhibit forms which in every sense of the word are inflectional. If in Finnish, for instance, we find _käsi_, in the singular, hand, and _kädet_, in the plural, hands, we see that phonetic corruption has clearly reached the very core of the noun, and given rise to a plural more decidedly inflectional than the Greek χεῖρ-ες, or the English _hand-s_. In Tamil, where the suffix of the plural is +gaḷ+, we have indeed a regular combinatory form in +kei-gaḷ+, hands; but if the same plural suffix +gaḷ+ is added to +kal+, stone, the euphonic rules of Tamil require not only a change in the suffix, which becomes +kaḷ+, but likewise a modification in the body of the word, +kal+ being changed to +kar+. We thus get the plural +karkaḷ+ which in every sense of the word is an inflectional form. In this plural suffix +gaḷ+, Dr. Caldwell has recognized the Dravidian +taḷa+ or +daḷa+, a host, a crowd; and though, as he admits himself in the second edition (p. 143), the evidence in support of this etymology may not be entirely satisfactory, the steps by which the learned author of the Grammar of the Dravidian languages has traced the plural termination +lu+ in Telugu back to the same original suffix +kaḷ+ admit of little doubt.

Evidence of a similar kind may easily be found in any grammar, whether of an isolating, combinatory, or inflectional language, wherever there is evidence as to the ascending or descending progress of any particular form of speech. Everywhere amalgamation points back to combination, and combination back to juxtaposition, everywhere isolating speech tends towards terminational forms, and terminational forms become inflectional.

I may best be able to explain the view commonly held with regard to the strata of language by a reference to the strata of the earth. Here, too, where different strata have been tilted up, it might seem at first sight as if they were arranged perpendicularly and side by side, none underlying the other, none presupposing the other. But as the geologist, on the strength of more general evidence, has to reverse this perpendicular position, and to re-arrange his strata in their natural order, and as they followed each other horizontally, the student of language too is irresistibly driven to the same conclusion. No language can by any possibility be inflectional without having passed through the combinatory and isolating stratum; no language can by any possibility be combinatory without clinging with its roots to the underlying stratum of isolation. Unless Sanskrit and Greek and Hebrew had passed through the combinatory stratum, nay, unless, at some time or other, they had been no better than Chinese, their present form would be as great a miracle as the existence of chalk (and the strata associated with it) without an underlying stratum of oolite (and the strata associated with it;) or a stratum of oolite unsupported by the trias or system of new red sandstone. Bunsen’s dictum, that “the question whether a language can begin with inflections, implies an absurdity,” may have seemed too strongly worded: but if he took inflections in the commonly received meaning, in the sense of something that may be added or removed from a base in order to define or to modify its meaning, then surely the simple argument _ex nihilo nihil fit_ is sufficient to prove that the inflections must have been something by themselves, before they became inflections relatively to the base, and that the base too must have existed by itself, before it could be defined and modified by the addition of such inflections.

But we need not depend on purely logical arguments, when we have historical evidence to appeal to. As far as we know the history of language, we see it everywhere confined within those three great strata or zones which we have just described. There are inflectional changes, no doubt, which cannot as yet be explained, such as the _m_ in the accusative singular of masculine, feminine, and in the nominative and accusative of neuter nouns; or the change of vowels between the Hebrew _Piel_ and _Pual_, _Hiphil_ and _Hophal_, where we might feel tempted to admit formative agencies different from juxtaposition and combination. But if we consider how in Sanskrit the Vedic instrumental plural, +aśvebhis+ (Lat. _equobus_), becomes before our very eyes +aśvais+ (Lat. _equis_), and how such changes as _Bruder_, brother, and _Brüder_, brethren, _Ich weiss_, I know, A.S. _wât_, and _Wir wissen_, we know, A.S. _wit-on_, have been explained as the results of purely mechanical, _i.e._, combinatory proceedings, we need not despair of further progress in the same direction. One thing is certain, that, wherever inflection has yielded to a rational analysis, it has invariably been recognized as the result of a previous combination, and wherever combination has been traced back to an earlier stage, that earlier stage has been simple juxtaposition. The primitive blocks of Chinese and the most perplexing agglomerates of Greek can be explained as the result of one continuous formative process, whatever the material elements may be on which it was exercised; nor is it possible even to imagine in the formation of language more than these three strata through which hitherto all human speech has passed.

All we can do is to subdivide each stratum, and thus, for instance, distinguish in the second stratum the suffixing (R + ρ) from the prefixing (ρ + R), and from the affixing (ρ + R + ρ) languages.

A fourth class, the infixing or incapsulating languages, are but a variety of the affixing class, for what in Bask or in the polysynthetic dialects of America has the appearance of actual insertion of formative elements into the body of a base can be explained more rationally by the former existence of simpler bases to which modifying suffixes or prefixes have once been added, but not so firmly as to exclude the addition of new suffixes at the end of the base, instead of, as with us, at the end of the compound. If we could say in Greek δείκ-μι-νυ, instead of δείκ-νυ-μι, or in Sanskrit +yu-mi-na-j+, instead of +yu-na-j-mi+, we should have a real beginning of so-called incapsulating formations.[9]

A few instances will place the normal progress of language from stratum to stratum more clearly before our eyes. We have seen that in Chinese every word is monosyllabic, every word tells, and there are, as yet, no suffixes by which one word is derived from another, no case-terminations by which the relation of one word to another could be indicated. How, then, does Chinese distinguish between the son of the father, and the father of the son? Simply by position. _Fú_ is father, _tzé_, son; therefore _fú tzé_ is son of the father, _tzé fú_, father of the son. This rule admits of no exception but one. If a Chinese wants to say _a wine-glass_, he puts _wine_ first and _glass_ last, as in English. If he wants to say _a glass of wine_, he puts _glass_ first and _wine_ last. Thus _i-pei thsieou_, a cup of wine; _thsieou pei_, a wine-cup. If, however, it seems desirable to mark the word which is in the genitive more distinctly, the word _tchi_ may be placed after it, and we may say, _fú tchi tzé_, the son of the father. In the Mandarin dialect this _tchi_ has become _ti_, and is added so constantly to the governed word, that, to all intents and purposes, it may be treated as what we call the termination of the genitive. Originally this _tchi_ was a relative, or rather a demonstrative, pronoun, and it continues to be used as such in the ancient Chinese.[10]

It is perfectly true that Chinese possesses no derivative suffixes; that it cannot derive, for instance, _kingly_ from a noun, such as _king_, or adjectives like _visible_ and _invisible_ from a verb _videre_, to see. Yet the same idea which we express by invisible, is expressed without difficulty in Chinese, only in a different way. They say _khan-pu-kien_, “I-behold-and-do-not-see,” and this to them conveys the same idea as the English _invisible_, though more exactly _invisible_ might be rendered by _kien_, to see, _pou-te_, one cannot, _tí_, which.

We cannot in Chinese derive from _ferrum_, iron, a new substantive _ferrarius_, a man who works in iron, a blacksmith; _ferraria_, an iron mine, and again _ferrariarius_, a man who works in an iron mine. All this is possible in an inflectional language only. But it is not to be supposed that in Chinese there is an independent expression for every single conception, even for those which are clearly secondary and derivative. If an arrow in Chinese is _shi_, then a maker of arrows (in old French _fléchier_, in English _fletcher_) is called an arrow-man, _shi-jin_. _Shui_ means water, _fu_, man; hence _shui-fu_, a water man, a water carrier. The same word _shui_, water, if followed by _sheu_, hand, stands for steersman, literally, water-hand. _Kin_ means gold, _tsiang_, maker; hence _kin-tsiang_, a goldsmith. _Shou_ means writing, _sheu_, hand; hence _shou-sheu_, a writer, a copyist, literally, a writing-hand.

A transition from such compounds to really combinatory speech is extremely easy. Let _sheu_, in the sense of hand, become obsolete, and be replaced in the ordinary language by another word for hand; and let such names as _shu-sheu_, author, _shui-sheu_, boatsman, be retained, and the people who speak this language will soon accustom themselves to look upon _sheu_ as a mere derivative, and use it by a kind of false analogy, even where the original meaning of _sheu_, hand, would not have been applicable.[11]

We can watch the same process even in comparatively modern languages. In Anglo-Saxon, for instance, _hâd_ means state, order. It is used as an independent word, and continued to be so used as late as Spenser, who wrote:--

“Cuddie, I wote thou kenst little good, So vainly t’ advaunce thy headlesse hood.”

After a time, however, _hâd_, as an independent word, was lost, and its place taken by more classical expressions, such as _habit_, _nature_, or _disposition_. But there remained such compounds as _man-hâd_, the state of man, _God-hâd_, the nature of God; and in these words the last element, being an empty word and no longer understood, was soon looked upon as a mere suffix. Having lost its vitality, it was all the more exposed to phonetic decay, and became both _hood_ and _head_.

Or, let us take another instance, The name given to the fox in ancient German poetry was _Regin-hart_. _Regin_ in Old High German means thought or cunning, _hart_, the Gothic _hardu_, means strong. This _hart_[12] corresponds to the Greek κράτος, which, in its adjectival form of κρατης, forms as many proper names in Greek as _hart_ in German. In Sanskrit the same word exists as _kratu_, meaning intellectual rather than bodily strength, a shade of meaning which is still perceivable even in the German _hart_, and in the English _hard_ and _hardy_. _Reginhart_, therefore, was originally a compound, meaning “thought-strong,” strong in cunning. Other words formed in the same or a very similar manner are: _Peranhart_ and _Bernhart_, literally, bear-minded, or bold like a bear; _Eburhart_, boar-minded; _Engil-hart_, angel-minded; _Gothart_, god-minded; _Egin-hart_, fierce-minded; _Hugihart_, wise-minded or strong in thought, the English _Hogarth_. In Low German the second element, _hart_, lost its _h_ and became _ard_. This _ard_ ceased to convey any definite meaning, and though in some words which are formed by _ard_ we may still discover its original power, it soon became a mere derivative, and was added promiscuously to form new words. In the Low German name for the fox, _Reinaert_, neither the first nor the second word tells us any longer anything, and the two words together have become a mere proper name. In other words the first portion retains its meaning, but the second, _ard_, is nothing but a suffix. Thus we find the Low German _dronk-ard_, a drunkard; _dick-ard_, a thick fellow; _rik-ard_, a rich fellow; _gêrard_, a miser. In English _sweet-ard_, originally a very sweet person, has been changed and resuscitated as _sweet-heart_,[13] by the same process which changed _shamefast_ into _shamefaced_. But, still more curious, this suffix _ard_, which had lost all life and meaning in Low German, was taken over as a convenient derivative by the Romance languages. After having borrowed a number of words such as _renard_, fox, and proper names like _Bernard_, _Richard_, _Gerard_, the framers of the new Romance dialects used the same termination even at the end of Latin words. Thus they formed not only many proper names, like _Abeillard_, _Bayard_, _Brossard_, but appellatives like _leccardo_, a gourmand, _linguardo_, a talker, _criard_, a crier, _codardo_, Prov. _coart_, Fr. _couard_, a coward.[14] That a German word _hart_, meaning strong, and originally strength, should become a Roman suffix may seem strange; yet we no longer hesitate to use even Hindustani words as English suffixes. In Hindustani +válá+ is used to form many substantives. If +Dilli+ is Delhi, then +Dill-vállá+ is a man of Delhi. +Go+ is cow, +go-válá+ a cow-herd, contracted into +gválá+. Innumerable words can thus be formed, and as the derivative seemed handy and useful, it was at last added even to English words, for instance in “Competition wallah.”

These may seem isolated cases, but the principles on which they rest pervade the whole structure of language. It is surprising to see how much may be achieved by an application of those principles, how large results may be obtained by the smallest and simplest means. By means of the single radical î or +yâ+ (originally +ya+), which in the Aryan languages means to go or to send, the almost unconscious framers of Aryan grammar formed not only their neuter, denominative, and causative verbs, but their passives, their optatives, their futures, and a considerable number of substantives and adjectives. Every one of these formations, in Sanskrit as well as in Greek, can be explained, and has been explained, as the result of a combination between any given verbal root and the radical _î_ or +yâ+.

There is, for instance, a root +nak+, expressive of perishing or destruction. We have it in +nak+, night; Latin _nox_, Greek νύξ, meaning originally the waning, the disappearing, the death of day. We have the same root in composition, as, for instance, +jîva-nak+, life-destroying; and by means of suffixes Greek has formed from it νεκ-ρός, a dead body, νέκ-υς, dead, and νέκ-υ-ες in the plural, the departed. In Sanskrit this root is turned into a simple verb, +naś-a-ti+, he perishes. But in order to give to it a more distinctly neuter meaning, a new verbal base is formed by composition with +ya+, +naś-ya-ti+, he goes to destruction, he perishes.

By the same or a very similar process denominative verbs are formed in Sanskrit to a very large extent. From +râjan+, king, we form +râjâ-ya-te+, he behaves like a king, literally, he goes the king, he acts the king, _il a l’allure d’un roi_. From +kumârî+, girl, +kûmârâ-ya-te+, he behaves like a girl, etc.[15]

After raising +naś+ to +nâśa+, and adding the same radical +ya+, Sanskrit produces a causative verb, +nâśa-ya-ti+, he sends to destruction, the Latin _nêcare_.

In close analogy to the neuter verb +naśyati+, the regular passive is formed in Sanskrit by composition with +ya+, but by adding, at the same time, a different set of personal terminations. Thus +náś-yá-ti+ means he perishes, while +naś-yá-te+ means he is destroyed.

The usual terminations of the Optative in Sanskrit are:--

yâm, yâs, yât, yâma, yâta, yus,

or, after bases ending in vowels:--

iyam, is, it, ima, ita, iyus.

In Greek:--

ιην, ιης, ιη, ιημεν, ιητε, ιεν,

or, after bases ending in o:--

ιμι, ις, ι, ιμεν, ιτε, ιεν.

In Latin:--

iêm iês iet ---- ---- ient, îm, îs, it, îmus, îtis, int.

If we add these terminations to the root +AS+, to be, we get the Sanskrit +s-yâm+ for +as-yâm+:--

syâm, syâs, syât, syâma, syâta, syus.

Greek ἐσ-ίην, contracted to εἴην:--

εἴην, εἴης, εἴη, εἴημεν, εἴητε, εἶεν

Latin _es-iem_, changed to _siêm_, _sîm_, and _erîm_:--

siêm, siês, siet,[16] ---- ---- sient. sim, sîs, sit,[17] sîmus, sitis, sint. erîm, erîs, erit, erîmus, erîtis, erint.

If we add the other termination to a verbal base ending in certain vowels, we get the Sanskrit +bhara-iyam+, contracted to +bháreyam+:--

bharêyam, bharês, bharêt, bharêma, bharêta, bharêyus.

in Greek φέρο-ιμι:--

φέρο-ιμι, φέρο-ις, φέρο-ι, φέρο-ιμεν, φέρο-ιτε, φέρο-ιεν

in Latin _fere-im_, changed to _ferem_, used in the sense of a future, but replaced[18] in the first person by _feram_, the subjunctive of the present:--

feram, ferês, feret, ferêmus, ferêtis, ferent.

Perfect Subjunctive:--

tul-erîm, tul-erîs, tul-erit, tul-erimus, tul-eritis,[19] tul-erint.

Here we have clearly the same auxiliary verb, i or +ya+, again, and we are driven to admit that what we now call an optative or potential mood, was originally a kind of future, formed by +ya+, to go, very much like the French _je vais dire_, I am going to say, I shall say, or like the Zulu 1 2 3 4 1 2 3 4 +ngi-ya-ku-tanda+, I go to love, I shall love.[20] The future would afterwards assume the character of a civil command, as “thou wilt go” may be used even by us in the sense of “go;” and the imperative would dwindle away into a potential, as we may say: “Go and you will see,” in the same sense as, If you go, you will see.

The terminations of the future are:--

Sanskrit:--

syâmi, syasi, syati, syâmas, syâtha, syanti.

Greek:--

σω, σεις, σει, σομεν, σετε, σοντι.

Latin:--

ero, erĭs, erĭt, erĭmus, erĭtis, erunt.

In these terminations we have really two auxiliary verbs, the verb +as+, to be, and +ya+, to go, and by adding them to any given root, as, for instance, +DA+, to give, we have the Sanskrit (+dâ-as-yâ-mi+):--

dâ-s-yâ-mi, dâ-s-ya-si, dâ-s-ya-ti, dâ-s-yâ-mas, dâ-s-ya-tha, dâ-s-ya-nti,

Greek (δω-εσ-ιω):--

δώ-σ-ω,[21] δώ-σ-εις, δώ-σ-ει, δώ-σ-ομεν, δώ-σ-ετε, δώ-σ-ουσι

Latin:--

pot-ero, pot-erĭs, pot-erit, pot-erĭmus, pot-erĭtis, pot-erunt.

A verbal form of very frequent occurrence in Sanskrit is the so-called gerundive participle which signifies that a thing is necessary or proper to be done. Thus from +budh+, to know, is formed +bodh-ya-s+, one who is to be known, _cognoscendus_; from +guh+, to hide, +gúh-ya-s+, or +goh-ya-s+, one who is to be hidden, literally, one who goes to a state of hiding or being hidden; from +yaj+, to sacrifice, +yâj-ya-s+, one who is or ought to be worshipped. Here, again, what is going to be becomes gradually what will be, and lastly, what shall be. In Greek we find but few analogous forms, such as ἅγιος, holy, στύγ-ι-ος, to be hated; in Latin _ex-im-i-us_, to be taken out; in Gothic _anda-nêm-ja_, to be taken on, to be accepted, agreeable, German _angenehm_.[22]

While the gerundive participles in +ya+ are formed on the same principle as the verbal bases in +ya+ of the passive, a number of substantives in +ya+ seem to have been formed in close analogy to the bases of denominative verbs, or the bases of neuter verbs, in all of which the derivative +ya+ expresses originally the act of going, behaving, and at last of simple being. Thus from +vid+, to know, we find in Sanskrit +vid-yâ+, knowing, knowledge; from +śi+, to lie down, +śayyâ+; resting. Analogous forms in Latin are _gaud-i-um_, _stud-i-um_, or with feminine terminations, _in-ed-i-a_, _in-vid-i-a_, _per-nic-i-es_, _scab-i-es_; in Greek, μαν-ί-α, ἁμαρτ-ί-α, or ἁμάρτ-ι-ον; in German, numerous abstract nouns in _i_ and _e_.[23]

This shows how much can be achieved, and has been achieved, in language with the simplest materials. Neuter, denominative, causative, passive verbs, optatives and futures, gerundives, adjectives, and substantives, all are formed by one and the same process, by means of one and the same root. It is no inconsiderable portion of grammar which has thus been explained by this one root +ya+, to go, and we learn again and again how simple and yet how wonderful are the ways of language, if we follow them up from stratum to stratum to their original starting-point.

Now what has happened in these cases, has happened over and over again in the history of language. Everything that is now formal, not only derivative suffixes, but everything that constitutes the grammatical framework and articulation of language, was originally material. What we now call the terminations of cases were mostly local adverbs; what we call the personal endings of verbs were personal pronouns. Suffixes and affixes were mostly independent words, nominal, verbal, or pronominal; there is, in fact, nothing in language that is now empty, or dead, or formal, that was not originally full, and alive, and material. It is the object of Comparative Grammar to trace every formal or dead element back to its life-like form; and though this resuscitating process is by no means complete, nay, though in several cases it seems hopeless to try to discover the living type from which proceeded the petrified fragments which we call terminations or suffixes, enough evidence has been brought together to establish on the firmest basis this general maxim, that _Nothing is dead in any language that was not originally alive_; that nothing exists in a tertiary stratum that does not find its antecedents and its explanation in the secondary or primary stratum of human speech.

After having explained, as far as it was possible in so short a time, what I consider to be the right view of the stratification of human speech, I should have wished to be able to show to you how the aspect of some of the most difficult and most interesting problems of our science is changed, if we look at them again with the new light which we have gained regarding the necessary antecedents of all language. Let me only call your attention to one of the most contested points in the Science of Language. The question whether we may assign a common origin to the Aryan and Semitic languages has been discussed over and over again. No one thinks now of deriving Sanskrit from Hebrew, or Hebrew from Sanskrit; the only question is whether at some time or other the two languages could ever have formed part of one and the same body of speech. There are scholars, and very eminent scholars, who deny all similarity between the two, while others have collected materials that would seem to make it difficult to assign such numerous coincidences to mere chance. Nowhere, in fact, has Bacon’s observation on this radical distinction between different men’s dispositions for philosophy and the sciences been more fully verified than among the students of the Science of Language:--_Maximum et velut radicale discrimen ingeniorum, quoad philosophiam et scientias, illud est, quod alia ingenia sint fortiora et aptiora ad notandas rerum differentias; alia ad notandas rerum similitudines. . . . . . Utrumque autem ingenium facile labitur in excessum, prensando aut gradus rerum, aut umbras._[24] Before, however, we enter upon an examination of the evidence brought forward by different scholars in support of their conflicting theories, it is our first duty to ask a preliminary question, viz.: What kind of evidence have we any right to expect, considering that both Sanskrit and Hebrew belong, in the state in which we know them, to the inflectional stratum of speech?

Now it is quite true that Sanskrit and Hebrew had a separate existence long before they reached the tertiary stratum, before they became thoroughly inflectional; and that consequently they can share nothing in common that is peculiar to the inflectional stratum in each, nothing that is the result of phonetic decay, which sets in after combinatory formations have become unintelligible and traditional. I mean, supposing that the pronoun of the first person had been originally the same in the Semitic and Aryan languages, supposing that in the Hebrew _an-oki_ (Assyrian _an-aku_, Phen. _anak_) the last portion, _oki_, was originally identical with the Sanskrit +ah+ in +aham+, the Greek ἐγ in ἐγ-ώ, it would still be useless to attempt to derive the termination of the first person singular, whether in _kâtal-ti_ or in _ektôl_, from the same type which in Sanskrit appears as +mi+ or +am+ or +a+, in +tudâ-mi+, +atud-am+, +tutod-a+. There cannot be between Hebrew and Sanskrit the same relationship as between Sanskrit and Greek, if indeed the term of relationship is applicable even to Sanskrit and Greek, which are really mere dialectic varieties of one and the same type of speech.

The question then arises, Could the Semitic and Aryan languages have been identical during the second or _combinatory_ period? Here, as before, the answer must be, I believe, decidedly negative, for not only are the empty words which are used for derivative purposes different in each, but, what is far more characteristic, the manner in which they are added to the stems is different too. In the Aryan languages formative elements are attached to the ends of words only; in the Semitic languages they are found both at the end and at the beginning. In the Aryan languages grammatical compounds are all according to the formula rρ; in the Semitic we have formations after the formulas rρ, ρr, and ρrρ.

There remains, therefore, the first or isolating stage only in which Semitic and Aryan speech might have been identical. But even here we must make a distinction. All Aryan roots are monosyllabic, all Semitic roots have been raised to triliteral form. Therefore it is only previous to the time when the Semitic roots assumed this secondary triliteral form that any community could possibly be admitted between these two streams of language. Supposing we knew as an historical fact that at this early period--a period which transcends the limits of everything we are accustomed to call historical--Semitic and Aryan speech had been identical, what evidence of this union could we expect to find in the actual Semitic and Aryan languages such as we know them in their inflectional period? Let us recollect that the 100,000 words of English, nay, the many hundred thousand words in all the dictionaries of the other Aryan languages, have been reduced to about 500 roots, and that this small number of roots admits of still further reduction. Let us, then, bear in mind that the same holds good with regard to the Semitic languages, particularly if we accept the reduction of all triliteral to biliteral roots. What, then, could we expect in our comparison of Hebrew and Sanskrit but a small number of radical coincidences, a similarity in the form and meaning of about 500 radical syllables, everything else in Hebrew and Sanskrit being an after-growth, which could not begin before the two branches of speech were severed once and forever.

But more, if we look at these roots we shall find that their predicative power is throughout very general, and therefore liable to an infinite amount of specification. A root that means to fall (Sk. +pat+, πί-πτ-ω) comes to mean to fly (Sk. +ut-pat+, πέτομαι). The root +dâ+, which means to give, assumes, after the preposition â, the sense of taking. The root +yu+, which means to join, means to separate if preceded by the preposition +vi+. The root +ghar+, which expresses brightness, may supply, and does supply in different Aryan languages, derivations expressive of brightness (gleam), warmth (Sk. +gharma+, heat), joy (χαίρειν), love (χάρις), of the colors of green (Sk. +hari+), yellow (_gilvus_, _flavus_), and red (Sk. +harit+, _fulvus_), and of the conception of growing (_ger-men_). In the Semitic languages this vagueness of meaning in the radical elements forms one of the principal difficulties of the student, for according as a root is used in its different conjugations, it may convey the most startling variety of conception. It is also to be taken into account that out of the very limited number of roots which at that early time were used in common by the ancestors of the Aryan and Semitic races, a certain portion may have been lost by each, so that the fact that there are roots in Hebrew of which no trace exists in Sanskrit, and _vice versâ_, would again be perfectly natural and intelligible.

It is right and most essential that we should see all this clearly, that we should understand how little evidence we are justified in expecting in support of a common origin of the Semitic and Aryan languages, before we commit ourselves to any opinion on this important subject. I have by no means exhausted all the influences that would naturally, nay necessarily, have contributed towards producing the differences between the radical elements of Aryan and Semitic speech, always supposing that the two sprang originally from the same source. Even if we excluded the ravages of phonetic decay from that early period of speech, we should have to make ample allowances for the influence of dialectic variety. We know in the Aryan languages the constant play between gutturals, dentals, and labials (_quinque_, Sk. +panca+, πέντε, Æol. πέμπε, Goth. _fimf_). We know the dialectic interchange of Aspirate, Media, and Tenuis, which, from the very beginning, has imparted to the principal channels of Aryan speech their individual character (τρεῖς, Goth. _threis_, High German _drei_).[25] If this and much more could happen within the dialectic limits of one more or less settled body of speech, what must have been the chances beyond those limits? Considering how fatal to the identity of a word the change of a single consonant would be in monosyllabic languages, we might expect that monosyllabic roots, if their meaning was so general, vague, and changeable, would all the more carefully have preserved their consonantal outline. But this is by no means the case. Monosyllabic languages have their dialects no less than polysyllabic ones; and from the rapid and decisive divergence of such dialects, we may learn how rapid and decisive the divergence of language must have been during the isolating period. Mr. Edkins, who has paid particular attention to the dialects of Chinese, states that in the northern provinces the greatest changes have taken place, eight initial and one final consonant having been exchanged for others, and three finals lost. Along the southern bank of the Yang-tsï-kiang, and a little to the north of it, the old initials are all preserved, as also through Chekiang to Fuh-kien. But among the finals, _m_ is exchanged for _n_; _t_ and _p_ are lost, and also _k_, except in some country districts. Some words have two forms, one used colloquially, and one appropriated to reading. The former is the older pronunciation, and the latter more near to Mandarin. The cities of Su-cheu, Hang-cheu, Ningpo, and When-cheu, with the surrounding country, may be considered as having one dialect, spoken probably by thirty millions of people, _i.e._, by more than the whole population of Great Britain and Ireland. The city of Hwei-cheu has a dialect of its own, in which the soft initial consonants are exchanged for hard and aspirated ones, a process analogous to what we call _Lautverschiebung_ in the Aryan languages. At Fu-cheu-fu, in the eastern part of the province of Kiang-si, the soft initials have likewise been replaced by aspirates. In many parts of the province of Hunan the soft initials still linger on; but in the city of Chang-sha the spoken dialect has the five tones of Mandarin, and the aspirated and other initials distributed in the same manner. In the island of Hai-nan there is a distinct approach to the form which Chinese words assume in the language of Annam. Many of the hard consonants are softened, instead of the reverse taking place as in many other parts of China. Thus _ti_, _di_, both _ti_ in Mandarin, are both pronounced _di_ in Hai-nan. _B_ and _p_ are both used for many words whose initials are _w_ and _f_ in Mandarin. In the dialects of the province of Fuhkien the following changes take place in initial consonants: _k_ is used for _h_; _p_ for _f_; _m_, _b_, for _w_; _j_ for _y_; _t_ for _ch_; _ch_ for _s_; _ng_ for _i_, _y_, _w_; _n_ for _j_.[26] When we have clearly realized to ourselves what such changes mean in words consisting of one consonant and one vowel, we shall be more competent to act as judges, and to determine what right we have to call for more ample and more definite evidence in support of the common origin of languages which became separated during their monosyllabic or isolating stages, and which are not known to us before they are well advanced in the inflectional stage.

It might be said,--Why, if we make allowance for all this, the evidence really comes to nothing, and is hardly deserving of the attention of the scholar. I do not deny that this is, and always has been my own opinion. All I wish to put clearly before other scholars is, that this is not our fault. We see why there can be no evidence, and we find there is no evidence, or very little support of a common origin of Semitic and Aryan speech. But that is very different from dogmatic assertions, so often and so confidently repeated, that there can be no kind of relationship between Sanskrit and Hebrew, that they must have had different beginnings, that they represent, in fact, two independent species of human speech. All this is pure dogmatism, and no true scholar will be satisfied with it, or turn away contemptuously from the tentative researches of scholars like Ewald, Raumer, and Ascoli. These scholars, particularly Raumer and Ascoli, have given us, as far as I can judge, far more evidence in support of a radical relationship between Hebrew and Sanskrit than, from my point of view, we are entitled to expect. I mean this as a caution in both directions. If, on one side, we ought not to demand more than we have a right to demand, we ought, on the other, not to look for, nor attempt to bring forward, more evidence than the nature of the case admits of. We know that words which have identically the same sound and meaning in Sanskrit, Greek, Latin, and German, cannot be the same words, because they would contravene those phonetic laws that made these languages to differ from each other. _To doom_ cannot have any connection with the Latin _damnare_; _to call_ cannot be the Greek καλεῖν, the Latin _calare_; nor Greek φαῦλος the German _faul_; the English _care_ cannot be identified with Latin _cura_, nor the German _Auge_ with the Greek αὐγή. The same applies, only with a hundred-fold greater force, to words in Hebrew and Sanskrit. If any triliteral root in Hebrew were to agree with a triliteral word in Sanskrit, we should feel certain, at once, that they are not the same, or that their similarity is purely accidental. Pronouns, numerals, and a few imitative rather than predicative names for father and mother, etc., may have been preserved from the earliest stage by the Aryan and Semitic speakers; but if scholars go beyond, and compare such words as Hebrew _barak_, to bless, and Latin _precari_; Hebrew _lab_, heart, and the English _liver_; Hebrew _melech_, king, and the Latin _mulcere_, to smoothe, to quiet, to subdue, they are in great danger, I believe, of proving too much.

Attempts have lately been made to point out a number of roots which Chinese shares in common with Sanskrit. Far be it from me to stigmatize even such researches as unscientific, though it requires an effort for one brought up in the very straitest school of Bopp, to approach such inquiries without prejudice. Yet, if conducted with care and sobriety, and particularly with a clear perception of the limits within which such inquiries must be confined, they are perfectly legitimate; far more so than the learned dogmatism with which some of our most eminent scholars have declared a common origin of Sanskrit and Chinese as out of the question. I cannot bring myself to say that the method which Mr. Chalmers adopts in his interesting work on the “Origin of Chinese” is likely to carry conviction to the mind of the _bonà fide_ skeptic. I believe, before we compare the words of Chinese with those of any other language, every effort should be made to trace Chinese words back to their most primitive form. Here Mr. Edkins has pointed out the road that ought to be followed, and has clearly shown the great advantage to be derived from an accurate study of Chinese dialects. The same scholar has done still more by pointing out how Chinese should at first be compared with its nearest relatives, the Mongolian of the North-Turanian, and the Tibetan of the South-Turanian class, before any comparisons are attempted with more distant colonies that started during the monosyllabic period of speech. “I am now seeking to compare,” he writes, “the Mongolian and Tibetan with the Chinese, and have already obtained some interesting results:--

“1. A large proportion of Mongol words are Chinese. Perhaps a fifth are so. The identity is in the first syllable of the Mongol words, that being the root. The correspondence is most striking in the adjectives, of which perhaps one half of the most common are the same radically as in Chinese; e.g., _sain_, good; _begen_, low; _ic‘hi_, right; _sologai_, left; _c‘hihe_, straight; _gadan_, outside; _c’hohon_, few; _logon_, green; _hung-gun_, light (not heavy). But the identity is also extensive in other parts of speech, and this identity of common roots seems to extend into the Turkish, Tatar, etc.; e.g., _su_, water; _tenri_, heaven.

“2. To compare Mongol with Chinese it is necessary to go back at least six centuries in the development of the Chinese language. For we find in common roots final letters peculiar to the old Chinese, _e.g._, final _m_. The initial letters also need to be considered from another standpoint than the Mandarin pronunciation. If a large number of words are common to Chinese, Mongol, and Tatar, we must go back at least twelve centuries to obtain a convenient epoch of comparison.

“3. While the Mongol has no traces of tones, they are very distinctly developed in Tibetan. Csoma de Körös and Schmidt do not mention the existence of tones, but they plainly occur in the pronunciation of native Tibetans resident in Peking.

“4. As in the case of the comparison with Mongol, it is necessary in examining the connection of Tibetan with Chinese to adopt the old form of the Chinese with its more numerous final consonants, and its full system of soft, hard, and aspirated initials. The Tibetan numerals exemplify this with sufficient clearness.

“5. While the Mongol is near the Chinese in the extensive prevalence of words common to the two languages, the Tibetan is near in phonal structure, as being tonic and monosyllabic. This being so, it is less remarkable that there are many words common to Chinese and Tibetan, for it might have been expected; but that there should be perhaps as many in the Mongol with its long untoned polysyllables, is a curious circumstance.”[27]

This is no doubt the right spirit in which researches into the early history of language should be conducted, and I hope that Mr. Edkins, Mr. Chalmers, and others, will not allow themselves to be discouraged by the ordinary objections that are brought against all tentative studies. Even if their researches should only lead to negative results, they would be of the highest importance. The criterion by which we test the relationship of inflectional languages, such as Sanskrit and Greek, Hebrew and Arabic, cannot, from the nature of the case, be applied to languages which are still in the combinatory or isolating stratum, nor would they answer any purpose, if we tried by them to determine whether certain languages, separated during their inflectional growth, had been united during their combinatory stage, or whether languages, separated during their combinatory progress, had started from a common centre in their monosyllabic age. Bopp’s attempt to work with his Aryan tools on the Malayo-Polynesian languages, and to discover in them traces of Aryan forms, ought to serve as a warning example.

However, there are dangers also, and even greater dangers, on the opposite shore, and if Mr. Chalmers in his interesting work on “the Origin of Chinese,” compares, for instance, the Chinese _tzé_, child, with the Bohemian _tsi_, daughter, I know that the indignation of the Aryan scholars will be roused to a very high pitch, considering how they have proved most minutely that _tsi_ or _dci_ in Bohemian is the regular modification of _dugte_, and that _dugte_ is the Sanskrit +duhitar+, the Greek θυγάτηρ, daughter, originally a pet-name, meaning a milk-maid, and given by the Aryan shepherds, and by them only, to the daughters of their house. Such accidents[28] will happen in so comprehensive a subject as the Science of Language. They have happened to scholars like Bopp, Grimm, and Burnouf, and they will happen again. I do not defend haste or inaccuracy, I only say, we must venture on, and not imagine that all is done, and that nothing remains to conquer in our science. Our watchword, here as elsewhere, should be Festina lente! but, by all means, Festina! Festina! Festina!