In reviewing the conditions of the existence of linguistical and other cultural phenomena, we have seen that the elements are grouped into complexes. The linguistical complexes, just as the economical, technical, and social ones, in the different ethnical groups may have a certain similarity. In a mind which is not armed with a knowledge of the nature of the complexes, the similarity is often mistaken for an indication of a common origin of complexes and even bearers. Referring to the language, we find that the evidences of «common origin» are found in «common words»; so that this problem is quite important, especially in view of the second part of this work where we shall deal with «common words.»
I might confine myself in referring to the recent works on language, quoted here, also to many others which have not been quoted, and where the problem of common words and limitations in using them as evidences are perfectly well shown; but since, in the second part of this work, I shall have to refer to the various aspects of the problem of common words, I shall now point out some facts and conclusions. It is rather surprising (later on, we shall see that it is not so) that the analysis of the nature of common words and possibilities in using them as evidences do not reach many investigators who practically ignore what is known as to the nature of «common words.» These general linguistical works remain a «pure theory,» somewhat impractical. The common words met with in two and several languages may be grouped into two sets; namely, common words recognized as «common» and common words which are classed sometimes as loanwords, convergence, etc. So there is a very definite distinction between two groups, for «common» is very often referred to as «genetically common, of the same origin,» an idea of which will be better understood when we analyse the complex of evolution and the idea of the organic character of language. However, since a «common» word may happen to be a «loan-word» for the given language, and since the «ancestor-language» is very often absolutely unknown, the labelling of «common» and «not-common» words becomes arbitrary. In a mind which postulates the common origin of languages as organical, evolving entities, the distinction of «common» and «loan-words» is of primary importance, so I shall start from the last type, that of common loan-words. However, since our chief goal is a discussion of the Ural-Altaic hypothesis, we must point out that we have no history of these languages (vide infra, Chapter IV) and we have no isoglosses, nor even sufficiently extensive dialectological works for all existing dialects.
It is not easy to give a short definition of a loan-word, because all words are to a certain moment «loaned» by the coming generations. But if we take it in a narrow sense, how may a «loan-word» be recognized actually as one? A loan-word is one which has not been created by the given ethnos speaking a certain language, but which has been borrowed from another ethnos speaking a different language. Naturally one always meets with the difficulty of definition, whether the word is a «loan-word» or not, for the starters are borrowed in great number, but the language which loans them disappears, so that if we do not know the history of the language in all its details we cannot say whether the given word is «loaned» or «created» or «received» from ancestors during the existence of the given language. Since historic documentation is lacking, yet the idea of distinction of loan-words from genuine ones is essential, other methods are brought forth for proving this distinction. I shall now point out some of these corroborative evidences. It is said that «the loan-word is one which in all details resembles that found in the other language.» But here one needs a series of corroborative evidences, including the presence of a series of words «loaned» from the same language and historic data of that loan. These evidences are often lacking altogether. It has been suggested that we must consider as genuine (not-loan) words all those which are found for a long time in the language. A. Meillet («Linguistique,» op. cit., p. 103) gives the definition thus: «Les emprunts sont les elements, pris a des parlers quelconques, et qui ne reposent pas sur une tradition continue.» But how may it be recognized when we know no history of the language? This definition is supported by the following: «The loan-word is not one which is assimilated by the language in such a manner that it gives a series of derivatives.» But we know foreign words which, during a short period,— less than one generation, — produce a series of derivatives and some other words which in this sense remain «loan-words» for ever, so that this evidence may become entirely misleading. And lastly, «the loan-words cannot be those which designate primary, elementary phenomena.» This negative characteristic cannot hold good, for we need an exact definition of «primary» and «elementary,» which is often impossible, and we know that such starters are sometimes loaned. Another negative characteristic, rarely used, is that «a word met with in several dialects cannot be a loan-word» [46.] But in this case it must be shown that it was not loaned, for the lack of evidence is not indicative that the word was not borrowed by a group of languages from a common source, and in different times, and one language from another. A. Meillet, in his discussion with Schuchardt, points out that the psychological condition is essential in the problem. So he says: «Les sujets qui ont transmis les elements indigenes ont eu constamment, d'une maniere plus on moins [italics are mine] nette, le sentiment et la volonte de parler leur langue traditionnelle» (id., p. 104). He concludes that without this consideration «la doctrine classique me parait… theoriquement insoutenable» (loc. cit.). This pessimistic conclusion is the last protection of the «doctrine classique.» Further investigations into the psychological complexes have shown that they cannot be used as evidences — they are mere functional mechanisms. The processes and elements may be percepted and reactions may be produced by ethnoses (made of individuals), but they are not always correctly referred to, and the reactions are not always effective enough to be observed.
It is thus evident that the operation with «common» and «loanwords» is extremely dangerous, for «loan-words» may easily happen to be included into the misleading series of «common» words. The idea of «loan-word» is simple, however, if one does not give to it particular contents implied by the idea of organic evolution. The words are of different origin and those whose origin may be connected (not arbitrarily, of course) with corresponding words in other languages, yet those whose migration cannot be surely established, must not be included in the series of «common» words; and till the establishing of their history, they may be conventionally designated as words of so-and-so origin.
Such a treatment of «loan-words» certainly greatly affects the length of the list of «common» words, and still the actual loan-words may happen to have no traces of having been borrowed, and thus these series of common words may become misleading.
The common words may be different in origin; namely, they may originate from (1) a statistical phenomenon of convergence; (2) spreading in the territory of the certain phenomenon together with the term; (3) terms referred to the local phenomena, i.e., confined to certain territory; (4) words resulting from the imitation of various sounds (onomatopoetic); (5) words transmitted through the mechanism of imitation from one to another ethnical group; (6) words transmitted from one to another generation by the tradition and spreading over the territory together with the bearers. The quantitative side of the problem is also of importance, especially from the point of view of the origin of the commonness of words.
46. In fact, it may not be so, for the loan might take place prior to the differentiation of languages in which it is found.