34. Linguistic Importance of a Correct Solution of the Problem of Aspiration and Bilabialization

Still greater difficulty in solving this problem is met with in Turk where no h or x are found. However, some words of the supposedly common origin are found with the initial b. These occurrences are explained as due to the preservation of the original initial bilabial of the Altaic pra-language. However, what the actual origin of this consonant in these words is is not yet established. The process of bilabialization in Turk might have the same place as it is in Mongol or as it is in Tungus. And lastly, the group of languages known under the name of Uralian has been brought by A. Sauvageot to join the Mongol, Turk, and Tungus under the same idea of the hypothetic loss of the initial labial tenuis in the «Altaic» languages. In the work of this author, the history of the occlusive tenuis is represented in an absolutely definite manner, as he says that «l'his-toire de 1'occlusive sourde en position initiate a ete la suivante» (op. cit., p. 62), as shown in the table below. I quote it in full as the most completely and definitely formulated statement based upon the work of several predecessors and his own, and also as a case of a consonant which in the «Ural-Altaic» languages, according to this hypothesis (promoted to the grade of «law») had a very complex history

The analysis of the phenomenon of aspiration and bilabialization in Tungus has already brought us to the conclusion that the phenomenon of occurrence of the initial f, p, v, w, h, x, and zero in the words of the same meaning does not require the hypothesis that there was such a complex history of the loss of the initial bilabial tenuis. On the other hand, we have also seen that the similar phenomenon of aspiration in Mongol is quite possible. Perhaps the case of bilabiate in Turk may also be connected with the process of bilabialization, if it is shown that the words are really of common origin. Indeed, my generalization is not spread over Turk, for I am not familiar with this language in the same degree as with the Tungus language, and the material that may be compared is rather limited. This may be due to two conditions; namely, the remoteness of similar processes in this language, or perhaps an insufficient development of any of them in the Turk. Here it must be pointed out that the remoteness must be understood, not only absolutely, in terms of centuries, but also in the relative sense; namely, how many phonetic changes had occurred since these phenomena, if any, had taken place.

Besides the fact that the phenomena of aspiration and bilabiali-zation may nowadays be observed in Tungus, we are in an extremely good position, owing to the existence of Goldi-Olcha, where, as shown, owing to the transplantation of bilabialized elements to an alien soil, the character of bilabialization has received a very strong and sharp expression through the modification of even a slight bilabialization into a stable element like p. It is curious to note that this language has been greatly responsible for the hypothesis of the loss of this consonant. As a matter of fact, without the accidental preservation of Goldi, it would be difficult if not impossible to build up this hypothesis. In Mongol and Turk it is applied by the simple method of analogy [97]. Yet with the loss of Manchu the hypothesis would be absolutely impossible. However, the preservation of Manchu is also a mere historic accident. The Goldi and Manchu languages are fortunately preserved, but how many other languages are lost for ever? How many phonetic processes have escaped the attention of now living observers? But the facts which are lost might perhaps bring the observers to entirely different conceptions. The hypothesis of the loss of the initial p is based, not upon the facts, but chiefly upon certain theoretical reasoning for illustration of which the facts are picked up here and there. One of the presumptions which is usually left without discussion is that the languages under discussion were formed from a common ancestor in rather recent time, and this moment is referred to the historic period. The fact that there are little changes in Turk as seen in Uigur, in Mongol as seen in the mediaeval language, and in Southern Tungus as seen in Nuichen produces a certain disappointment, for one has to remove the moment of hypothetical separation to the periods when there were no documents. This is a reasoning by analogy with the present Indo-European languages, most of which were formed in historic time witnessed by numerous written documents. Another important presumption is that the languages lose their phonetic elements implied by the transition from tenues to media lenes and zero. Indeed, these phenomena do exist in some particular languages which supplied facts for this conclusion, but the generalization of this conclusion cannot be supported by the contrary facts; namely, the alteration of media lenes into tenues, which also exists in some other languages. However, the former alteration is in good agreement with the ethnographical complex of the gradual «weakening and deterioration of human species.» Such terms as «fortis,» «stark,» and «strong» as opposed to «weak» support the idea of «weakening.» This conception remains effective even in the case of a change of terminology. The third important presumption is that the changes in languages always take definite direction conditioned by the factors appearing in the language as an organic entity. This is also a reasoning by analogy with the organisms conceived in a simplified manner.

If we now summarize what has been previously stated, we may see that the formation of this hypothesis is in some way independent on the facts and the latter originally were used not as facts needed for conclusions but as illustration, just as the artist in mosaic work uses differently coloured materials for combining them in such a way as to express his own idea. Such is the work of G. Ramstedt in which this idea was illustrated with a series of incidentally picked-up facts. The choice of facts was not always fortunate, so the mosaic work appeared but as a simple essay. The same methodology is seen in other works. G. Ramstedt has started his work with forty-two parallels from «Altaic» languages, P. P. Schmidt has added some more, P. Pelliot has also added some, and A. Sauvageot, who has included the «Uralian» languages, has brought forth eighty-five parallels (cases) to illustrate G. Ramstedt's hypothesis. In this form the hypothesis seems to come out of facts and its origin is unintentionally camouflaged from the eyes of laymen. In its present form it has a new function; namely, to prove in the most daring form the existence of an ancestor language — Ural-Altaic. So that volens nolens, we have now to proceed together with these authors to the analysis of evidences.

97. My approach to this problem in Mongol is also made by the analogy with Tungus, so I do not insist upon my theory in Mongol as a «law» but as a simple working hypothesis. In reference to Turk I cannot classify it even as a suggestion. The processes of loss are not rare, indeed. What is good for a particular case of Tungus may happen to be invalid for Turk.