From these quotations the whole system of work may be seen quite clearly. Referring to the previous general remarks and A. Sauvageot's direct indications, it may be formulated in the following propositions:
1. The languages may be grouped into the families (filiation, genetic unity) in which the relationship is defined by the fact of their having at a certain time a common ancestor — pra-language.
As shown from this presumption, logical consequences follow, which are:
2. There must be bearer-ancestors, which is naturally presumed. In the European complex they are usually called «Mongoloids,» or something like that, although their existence has never been shown, and anthropological investigations point to a great variety of types and probably races in Asia.
3. There must be territory, which is already known. (Reference is made to the authority of K. Donner, p. xxi). Hence the value of identic structure, which is not convincing without other evidences, appropriates its full significance.
4. There is a unity of filiation in the Ural-Altaic languages and thus a genetic unity.
These propositions form the backbone of the hypothesis, which is supported by the following reasons:
5. Alternations of consonants and vowels regularly observed in some languages in the same words (stems) and particularly «la loi de Ramstedt.»
6. Vocalic harmony, which is far from being a general phenomenon.
7. The parallels which are not numerous according to him and some of which may be rejected altogether.
The method is based upon:
8. Semantic and phonetic parallels.
9. The belief that the people do not loan words for expressing primary notions.
10. The presumption that the same words (stems) found in different languages are not loan-words (emprunt). 11. The presumption that if the source of the loan is not established the stem belongs to the given «language,» either as an original word or as an old loan which becomes indigene. This is not formulated on any particular occasion, but it is a logical conclusion from the previous (9) and (10), and it is practically applied.
There are some other elements of secondary importance which form the same complex. Some of them will be occasionally mentioned in this part [69].
The comparison of the above characteristics of A. Sauvageot's complex with those discussed in the previous chapter shows that the former do not essentially differ from the latter, the chief reason of which is the recognition of language as an evolving entity. The other three elements are logical consequences of the first one, and hypothetical, too, in their nature. However, all four hypotheses are supposed to support one another, which, from the methodological point of view, cannot be accepted. In fact, the bearers of the pra-language, if it is a reality, are not found even hypothetically. These might be any of the groups, nowadays existing, or, even more than that, there might be a group now extinct. The anthropological evidences are not in favour of any of these hypotheses, for the population of Asia is mixed. The finding of bearers is hopeless, as has already shown to be the fact in the case of the pra-Indo-Europeans, — any one of existing «races»; i.e., Nordics, Mediterraneans, Alpines, or Dinarics might be the ancestors who spoke the hypothetic Indo-European pra-language.
The finding of territory is based upon lexic evidences. But every one may understand that, in the language of a people living in a certain area, one has a much greater chance to find common words for notions of local milieu and adaptation, especially if the migrations took place prior to the settlement of the people in the given geographical area. The chance of common words is still greater when different groups live for a long time side by side. Since this is so, this proposition must be supported by other evidences. The evidences are of a negative character, — recent (historic) migrations nor clear, definite historic documents pointing to this have not been found, so it is supposed that the groups lived from time immemorial in the same area, and nearly in the same distribution of groups. This is naturally mere guesswork adapted to the main hypothesis, for that there were great migrations in Asia is a fact now established. These hypotheses are supported by evidences of phonetics and again by common words which are hypothetic, and which need a detailed treatment.
69. One of these elements is, for instance, the common civilization which follows as a consequence of the fundamental postulate: » 'Langue commune' suppose civilisation commune,» as A. Meillet says («La Methode,» op. cit., p. 17), and «chacune des grandes 'langnes communes' du passe doit exprimer un type de civilisation» (id., p. 20), whence a geographical location and physical bearers of civilization and language are nothing but further logical conclusions.