Transactions of the Asiatic Society of Japan/Series 1/Volume 2/Has Japanese an affinity with Aryan languages
HAS JAPANESE AN AFFINITY WITH
ARYAN LANGUAGES.
BY
W. G. ASTON, Esq.
Read before the Asiatic Society of Japan,
on the 17th June, 1874.
———o———
It has long been fully admitted that the languages constituting the Indo-European or Aryan family are sprung from a common ancestor, and it is even possible by comparing them with each other to arrive at a tolerably definite idea of the principal characteristics of this primitive Aryan speech, and of the state of civilization of the people who used it.
The nation or tribes by whom it was spoken now more than 3000 years ago probably inhabited the table-land which forms part of the modern Persia. They had already acquired not a few of the arts of civilization. They had settled habitations, and even towns and castles, and they possessed the principal domestic animals. They were also workers in several of the metals, and were acquainted with the arts of weaving and of brewing a fermented liquor.
Their language had passed the agglutinative stage, and reached that of inflexion. It had declensions of nouns and adjectives, conjugations of verbs, and the grammatical distinction of gender. The distinctions between the parts of speech had become comparatively well defined, and they had a system of numerals extending as far as a hundred.
The question naturally suggests itself whether this primitive Aryan tongue stood alone, wholly isolated from any languages which may at that period have been spoken around it, or whether it was merely one of a group, the members of which were related to each other in a somewhat similar way to that in which the languages of the Aryan family are now connected. The latter alternative seems in itself the more probable one, and the facts adduced by Mr. Edkins in his suggestive and valuable treatise “China’s place in Philology” appear to point unmistakeably to the conclusion that the ancient Chinese was one of these languages. Have we any grounds for supposing that Japanese, or rather that ancient tongue from which the modern Japanese is descended, occupied a similar position?
Before adverting to those points of resemblance between Japanese and Aryan languages which lend countenance to this supposition, it may be well to clear the ground a little by indicating in what respects it is evident that little similarity exists. Broadly speaking, the grammatical systems are entirely different. Japanese has no declensions, no conjugations (properly speaking) no grammatical distinctions of gender, number, or person such as all Aryan languages possess or have possessed at some period of their history. The construction and syntax proceed upon wholly different principles. The numerals are different, so are the names of the metals, and altogether the differences in the most important and fundamental respects are so numerous that it is quite plain that no near affinity need be looked for. In other words the point of divergence must be thrown back to a very remote period, and the common ancestor which this theory would give to Japanese and Aryan tongues must have been a language rude and undeveloped to a degree of which it is difficult for us to form an adequate conception. It must have been spoken in a remote antiquity by a tribe of ignorant savages belonging probably to the stone period of the history of our race.
Granting, however, that an affinity does exist such as would be implied by a distant common origin of this kind, where are we to look for it? Evidently not in the grammatical inflexions and terminations, nor in the construction or syntax, for all these belong to the comparatively later stages of a language’s development. Nor is it likely that after a separation of many thousands of years, any considerable part of the vocabulary should be the same. All that we can reasonably expect to find is that a number of the more essential roots in Japanese and in Aryan languages should upon examination appear to be identical, and if this were shown to be the case, enough would have been proved to substantiate the hypothesis above suggested.
The method and conditions of such an inquiry deserve a few words of remark. Some of the principles which should guide it may he stated as follows:—
The oldest words and forms of words attainable should be compared.
All grammatical additions should be carefully eliminated.
No letter-changes should be considered probable which are not supported by well established analogous cases.
The great desideratum, however, in comparing Japanese with Aryan languages is the discovery of the law or laws which govern the letter-changes which take place, in short of such a law as that discovered by Grimm for the principal branches of the Aryan family. It can hardly be doubted that if any real affinity exists, there will also be some such law by which their relations are governed.
The peculiarities of the languages concerned should not be forgotten. Those of the European languages it is impossible to enumerate here, but it may not be out of place to note briefly a few characteristics of the Japanese language which have a bearing on this question.
R is never found at the beginning of a word.
L is not a Japanese sound. It is represented by r.
N is not found at the end of verbal roots.
M and b are often interchangeable.
H and f are in Japanese identical.
Every Japanese syllable ends with a vowel.
Double consonants are unknown—at least in the older forms of the language.
T becomes ch before i, and ts before u.
D becomes j before i, and dz before u.
Grimm’s law divides the languages of the Aryan family into three classes: 1st, Latin, Greek, Sanskrit, etc.: 2nd, Low German, including English: and 3rd, High German. It is a statement of the different forms assumed in these classes of languages by what is substantially the same sound, little or no distinction having probably been recognized in the common original Aryan language. For example, the sound which is aspirated in the first class is the corresponding flat mute in the second, and the corresponding sharp mute in the third. The letters with which Grimm’s law is concerned, are the sharp mutes p, t, k, the flat mutes b, d, g, and the aspirates ph or f, th and gh or ch. It will be useful to examine how these letters are represented in the Japanese alphabet. Of the series p, b, ph or f, p is entirely wanting in the old language and may safely be passed over, and of the other two series the aspirates are absent, so that in each case we have two consonants instead of three viz: b and f, d and t, and k and g. It is unnecessary to say that in Japanese these pairs of sounds are represented each by one letter only. The difference in sound was not marked in writing until a comparatively late period when the diacritic mark known as the nigori came into use for this purpose. Even at present the distinction between k and g, t and d, etc. is somewhat hazy in the mouths of Japanese, and it may be assumed that in ancient times it was entirely disregarded. In other words we may take it that in Japanese k and g, b and f (or h) and t and d are the same letters.
It follows therefore that where in European languages we find p, b, ph or f, Japanese may have either b or f, (practically it is often the corresponding nasal m which as above stated is in Japanese interchangeable with b); where we find k, g, gh or ch, Japanese may have k or g, and where we have t, d or th, Japanese will have t or d or the letters which replace them before i and u. Upon the whole, however, Japanese would appear to approach more closely to the sounds of the principal branch of the Aryan family, viz., that represented by Greek, Latin, Sanskrit, etc.
Grimm’s law is subject to numerous exceptions. For example in English an h is often found, especially at the beginning of a word, where the rule requires an aspirated k or g i. e. ch or gh. Such differences must be allowed for, in comparing with Japanese, languages in which these exceptions are found.
The principles, some of the more important of which have now been briefly indicated, have been to some extent applied in preparing the following list of examples of apparent identity between Japanese and Aryan roots. It is not maintained, however, that these cases of similarity establish conclusively the hypothesis in question. They are rather to be looked upon as so much raw material, the real value of which it is difficult in the present state of our knowledge to determine, and they are given merely for what they are worth. The most that can be said of them is that they present a plausible primâ facie case in favour of the proposition that Japanese is remotely related to Aryan languages.
The following are a few examples out of a much larger number which might have been given:—
In the following cases a Japanese k is represented by h in English.
The Japanese m is usually b or p in European languages as:
N is the negative sound in Japanese as well as in European languages. It appears in ina, no, inamu, to refuse naki, not, the na of the neg. imperative etc. The use of n for negatives in European languages is too well known to require illustration.
There is no more essential part of a language than its system of pronouns, and it is therefore important to see if any resemblances can be traced in these parts of speech in Japanese and the Aryan languages. The following considerations seem to indicate a possible connection. Taking the oldest forms of the Japanese personal pronouns and stripping them of terminations, we have for the first personal pronoun a, for the second na, and for the third ka. These is no difficulty in supposing a to be identical with the Sanskrit aham, Greek egō, Lat. ego, Ger. ich and the Engl. I. The final guttural which is found in all these languages except English may easily have been lost in Japanese as it has been in English. There are numerous other cases in which gutturals are lost in Japanese, as for instance the k of adjectival terminations and g in such forms as sawaide for sawagide, and if we accept Bopp’s opinion that the root of the first personal pronoun in the Aryan languages is a, the guttural being an addition of the nature of a demonstrative, it is not even necessary to have recourse to this supposition.
Na, the pronoun of the second person, presents greater difficulties. N is however the nasal of t and as seen above the Japanese language often prefers the nasal in the case of m which is found instead of b. It is also well known that in the Yedo dialect at least, a nasal g has taken the place of the ordinary g hard. The same tendency would substitute n for t, and it is perhaps to this that such forms as nameru for taberu, to eat, are due. It also seems likely that nani, what, and tare, who, contain the same root. If this view is correct, na is identical with the Sanskrit tuam (stem tua) the Lat. tu, the Greek tu or su and the English thou.
The third personal pronoun ka is perhaps the Greek ekeinos, the initial e of which is a mere prefix, as may be seen by comparing its dialectical forms.
The Japanese reflexive pronoun shi may be compared with the Latin se, Engl. self, Ger. sich.
The root so or sa which appears in sore, sono, soko, saru (for sa aru) may be identical with the English definite article, which in its old form was a demonstrative pronoun with the same meaning as sono and had separate forms for the three genders viz. masc. se, fem. seo, neut. thaet, our modern article being a modified form of the last. What strengthens this supposition is the circumstance that just as from so as a stem are formed a number of adverbs as soko, sochi, sasuga, the English the serves as a stem from which are derived the adverbs there, thither, thence, thus.
By a similar analogy the root ko ‘this’ may be the same as the English pronoun he (as seen above a k in Japanese at the beginning of a word is commonly h in English) and its derivatives koko, kochi, will correspond to the English words here, hither.
It is evident that this paper covers but a very small portion of the field of inquiry. Some of the branches of the Aryan family of which no mention has been made, as for instance Russian, may present much closer analogies than any which have been adduced, and no reference has been made to resemblances in points of grammar, which are not entirely wanting and deserve to be investigated. Nor has the subject of the Turanian languages cognate to Japanese been taken up. Such languages might be expected to aid the inquiry by furnishing intermediate forms which would assist in identifying roots apparently different and in many other ways. The difficulty, however, is to find a cognate language with a sufficient resemblance to Japanese for this purpose. In so far as I am aware, Japanese has no near relation among the other languages of Asia, and it has certainly none whose relationship has been clearly established. There seems to be no such close affinity between the branches of the Turanian family as there is between the members of the Aryan family of languages, and it is doubtful whether any other language of Asia resembles Japanese as closely as the most widely separated branches of the Aryan family resemble each other. There are several which present striking analogies with it in point of construction and grammar, but when we come to the vocabulary, but little resemblance can be traced. The Aino language is perhaps the most promising in this respect, but hardly enough is known of it to enable one to speak with confidence.