Some three years ago I published a paper [1] devoted to a complex linguistic and ethnographie problem — the terms of orientation amongst the Tungus. This paper has illustrated my idea that the problem of common words in the so-called Altaic languages cannot be settled before we hâve a complète analysis of the complexes compared, both from the linguistical and the ethnographical point of view. My attitude was not in agreement with the current opinions as to this question—the paper was actually directed against the practice of the uncritical using of Tungus parallels with the intention of including the Tungus languages into the larger unit of the Altaic languages supposed to hâve originated from an ancestor-language. My publication has been noticed by P. Pelliot [2], who was particularly concerned in the treatment of the Tungus terms of orientation, for he had previously tried to bring forth some Tungus parallels with a quite definite intention of supporting the hypothesis of the Altaic pra-language [3]; so, referring to my paper, he says: «De bonnes remarques, et d'autres contestables; p. 184, n. 45, M. Sh. veut que M. Rudnev n'ait pas commis une inadvertance que j'ai signalée incidemment; toute l'anthropologie ne peut prévaloir contre M. Rudnev lui-même qui m'a écrit autrefois qu'il avait bien commis cette inadvertance.» This note formally contains nothing offensive (the discovery of the spécial System of orientation amongst the Tungus as well as the analysis of terms is a minor thing, the mentioning of which can be omitted), but in the above quotation one may see that the emphasis is put on my suggestion as to the origin of a minor half-mistake made by A. D. Rudnev. Indeed, the latter could not formally deny it. P. Pelliot has made of it a case which has at the same time screened the fundamental problem discussed in my paper. Yet, in his remark, he has mentioned anthropologie, which is not so simple as that. I do not need to explain that anthropologie has nothing to do with the case, for I have used an ethnographical explanation of A. D. Rudnev's half-mistake and not an anthropological one, so that there is some other purpose in mentioning anthropologie. This is a veiled form of protest against using neighbouring sciences for the analysis of phenomena which are supposed to be reserved for linguistics [4]. As an attitude it is rather characteristic, so I shall dwell on it a little longer. Who may say that any knowledge, be it anthropology or any other science, may not be useful for linguistical work? For instance, L. de Saussure's work on Chinese astronomy («Les Origines de l'astronomie Chinoise») was possible only on condition of this author's familiarity with astronomy. Then may not anthropology be helpful too? I think that any knowledge, even without a direct bearing on the subject, is useful. It is particularly true of anthropology, familiarity with which greatly helps in avoiding a heated discussion about the origin of the «Chinese,» the «Altaics,» the «Aryans,» and other linguistical groups directly concerned and even sometimes discussed by linguists without their being familiar with anthropology. Indeed, there is a serious objection as to including anthropology into the group of sciences with which the linguist must be familiar; namely, the difficulty of mastering two seemingly distinct fields. But this question is technical, and shall I say personal, so it has no theoretical interest.
I have dwelt on this topic, for in the present work I will occasionally refer to anthropology, and in the general setting of the problems I shall base myself on their ethnological aspects. I cannot shut my eyes to the fact that linguistical phenomena have also their ethnographical features, and the Ural-Altaic hypothesis in its nature is an ethnological problem. I think, therefore, that anthropology, ethnography, and especially ethnology, may be very helpful, and for this reason the chief subject of the present paper—«The Ural-Altaic Hypothesis»—is preceded by a short discussion of the general problem of ethnos and language. As a matter of fact, it has been necessary, for the problem of ethnos and language has not yet been treated, so references to previous publications could not be made. However, without such a preliminary introduction, the treatment of the Ural-Altaic hypothesis would even be impossible. On the other hand, I did not want to publish the general and special parts separately, for the former had been written ad hoc and could not be extensive enough to become an exhaustive treatment of the problem of ethnos and language — to this problem I shall return later in my other publications.
Since the problem here treated is approached from various points of view, I shall often refer to the facts and conclusions which have already become commonplace, almost truisms, for the specialists. However, they are not so for those who confine their work and interests to the restricted fields of scientific inquiry, and therefore they need to be pointed out. Yet I shall need them as a ground for approaching my chief goal. Besides the well-known facts and conclusions, I shall refer to various theories which have not yet been generally accepted; yet I shall offer some new facts and theories which have never been published. In this respect, the present work does not pretend to be uniform.
My point of view as to the methods and possibilities of discussing the problem of the common origin of the so-called Ural-Altaic languages in many a respect is different from that of other authors. The difference does not consist in the discussion of evidences proving or disproving a common origin, but the difference is due to my idea; namely, that the «origin» and «relationship» between the languages in this way cannot be discussed at all. This point of view ought to be supported by theoretical reasons, which have occupied a large portion of Part I. Another point to make clear is that of finding how such a difference in attitude exists, for without showing the process of formation of the opposite attitude my point of view might appear to be a mere «point of view,» as good as any other, while the difference is much deeper than this. This has necessitated other sections in which the theory or evolution and its consequences are shortly discussed.
The difference in method and the original point of procedure for the investigation requires a revision of terms used. L. K. Ogden and I. A. Richards have shown in their work («The Meaning of Meaning. A Study of Influence of Language upon Thought and of the Science of Symbolism,» London, 1923) how important symbolism is if one wants to be understood. However, the task of revision of terms requires more than a mere substitution of the old meanings by new ones, but a «symbolization» of old symbols, for some of the «symbols» cannot be used any more. Naturally, I cannot undertake this task in reference to a foreign language, such as English is to me. I have, however, permitted myself to introduce a few new terms: such as «ethnos,» which has already been used; and «starter,» which is here explained: and some other old terms to which I give a slightly different «meaning.» I deem it necessary to make these remarks in order to avoid a useless and tiresome «discussion about the words.»
In the treatment of the problem of ethnos and
language, I shall follow the shortest way of expounding some salient points of
my theory. Their complete treatment would require a much greater space than may
be allowed in discussing such a particular problem as the Ural-Altaic
hypothesis. So in many a case I shall confine myself only to the conclusions,
omitting altogether the method of my investigation and the enumeration of
facts. I must here point out that I, as well as many other authors, notice that
some critics make an elementary error; namely, in their mind every work which
is not supplied with «facts» is classed as a «speculative»
one. It is a good word, indeed, but one must use it with a certain amount of caution
and with correct reference. They refer, as a matter of fact, not to the
essentials of the method, but to the technique of writing. For a genuine,
speculative work, facts are not much needed, and they may be lacking altogether
in the written form. However, when the current fashion requires
«facts,» a speculative work may be supplied with facts in any
desirable quantity,—in fact, volumes of facts, — but the genuine speculative
character as to method will remain as it is [5]. On the other hand, the bringing
up of well-known facts from other publications may sometimes be absolutely
unnecessary for the expounding of a new theory or hypothesis built upon these
facts and based upon a purely inductive method. Such a work will not be
overburdened by the repetition of facts already reproduced hundreds of times by
those who want to avoid the reproach of speculation. In the present work (Part
I), I do not cite facts in all the cases discussed, but this does not at all
mean that my conclusions are not based upon facts, and that they are merely
«speculative.» All of them, with the exception of specially
pointed-out cases, are based upon facts, and at any moment the technique of
investigation may be shown and the conclusions illustrated. However, in most cases
this seems to be useless, for the facts are already known. It is different with
Part II, where I am dealing with the analysis of facts and bring up new facts.
Here one has a case in which the difference of method is evident. The work of
A. Sauvageot discussed here is based upon and inspired by a series of
hypotheses, and only seemingly based upon facts, while the facts are actually
used for supporting the hypotheses. My attitude is to have reliable facts
first, so, for the time being, I am not worrying about the problem whether the
pra-Ural-Altaic language existed or not.
1. «Northern Tungus Terms of Orientation,» in Rocznik Orjentalistyczny, Vol. IV, Lwow, 1928.
2. T'oung Pao, Vol. XXVI, 1929, pp. 411, 412.
3. «Les Mots à h initiale aujourd'hui amuie dans le mongol de» XIIIe et XIVe siècles,» in Journal Asiatique, avril-juin, 1925.
4. The mentioning of anthropology may also be interpreted as an intention of discrediting my linguistical evidences, for the anthropological investigations constitute one of the items of my work; Indeed, it would not be a fair way of discrediting them, so for this reason I reject this supposition.
5. In the field of ethnography and linguistics, there are hundreds of publications in which facts occupy more space than the simple fundamental speculative proposition, but the critics do not notice the mosaic work of the authors in using facts as a shield to protect themselves from the critics.