One of the inevitable consequences of the theory of evolution applied to languages is the recognition of the existence of a relationship between the languages as there is between the progeny and ancestors of living organisms. Although the absurdity of such an absolute analogy with the organisms is evident for every one who is familiar with the languages, yet it is silently admitted that the relationship of the same type does exist. The reason is that no Other theory exists sufficiently well developed, and especially there is no terminology generally adopted, to take the place of the old conception; while the latter is, so to say, unconsciously adopted by the youth in the schools, together with the present cultural complex of the so-called «civilized nations.» It is already so deeply rooted in this complex that in public opinion those who do not agree with the «theory of evolution» are merely backward and uneducated persons [50]. The phenomena disturbing the scheme of linguistic relationship are looked upon as accidental conditions. The list of these disturbing factors, as, e.g., borrowing, phonetic «degeneration,» the loss and mixture of language, etc., together with the increase of knowledge of facts, is becoming longer and longer. At last they come into conflict with the theory of relationship between the languages. The specialists have then either to abstain from undertaking any attempts at a further investigation of the problem of relationship — the attitude characteristic of several groups of linguists — and confine themselves to the minor problems, or to revise the fundamental problem of the nature of language, and thus they come to the same problem —the revision of the theory of evolution in its application to language.
The idea of the organic evolution of languages- results in another important element; namely, the recognition of the existence of a certain pra-language which was the ancestor of a group of languages. Although the hypothesis as to the existence of such a language in some cases may practically be absolutely useless from the point of view of establishing a relationship between the languages and impossible to ever be shown, yet the search for a pra-language constitutes one of the important items of the linguists' work. If the facts disagree with this presumption, then they must be either unconsciously overlooked or discredited under any handy pretext which does not come into conflict with the given complex. The attempts at the construction of hypothetic pra-languages have been made for all linguistic groups, or «families.» Indeed, the controversy between the reconstructors is inevitable, for the existence of such a pra-language itself is only a hypothesis. The connexion between a certain language and culture is one of fundamental postulates adopted by all who practise [51] the application of the organic evolutive principle in linguistics. Owing to this, the search for such a culture (civilization) constitutes one of the important items of reconstruction as made from the languages themselves [52]. Methodological fallacy is indeed evident.
Since the existence of pra-languages is presumed, there must also be the existence of some people who spoke it. This forms another inevitable consequence of the original conception. The search for the bearers of language naturally constitutes an important problem of the origin of language. The conflict between the creators of theories is also inevitable, for the existence of a pra-language itself is only a hypothesis. Here the reconstructor and the creator of theory as to the bearers of a pra-language bring forth a new series of facts, namely, other cultural elements and physical characteristics of hypothetical people. Both the ethnographical elements and the. anthropological characters of the bearers of a pra-language are a mere hypothesis, for, as shown, the ethnographical elements and the anthropological features are subject to changes in the ethnical units owing to the complex processes of variation and substitution. The hypothesis of a pra-language is thus supported by the hypothesis of the bearers of the language and the hypothesis of the bearers is supported by the hypothesis of the continuity of the ethnographical complex and a presumption (again a hypothesis) of physical characters.
But this is not all. The hypothesis of the bearers of the language results in another consequence; namely, the location of the bearers. Here the vicious circle is closed, for the geographical area is again reconstructed from the reconstruction of a pra-language [53].
50. A. Meillet gives («Linguistique,» op. cit., p. 102) an ethnographically interesting reason: «L''expression [parente] est trop etablie pour qu'on y renonce; il suffit de la definir pour n'en etre pas dupe.» His definition is that one must not understand parente as a relationship between mother and daughter, but as that of a «transformation» of one language. This definition is only a paraphrase of an old idea chiefly in its psychological and historical aspects. So A. Meillet again says (id., p. 81), «Ainsi la parente de langues resulte uniquement de la continuite du sentiment de l'unite linguistique.»
51. I say «practise,» for in some instances the theory denying the connexion of language as organical entities is proved but not applied.
52. Cf. supra, Chap. II, Sec. 16. A. Meillet in this respect goes very far, but he bases himself on combined data, — linguistical and historico-cultural.
53. From the above remarks it must not be inferred that my idea is that no people, no culture, no territory, existed with which a certain language in its elements continuing in other languages was connected, bat I want to point ont that it ought to be first shown that such a language, people, culture, and territory did exist. The same holds good for the reconstruction of cultural complexes, which may not always be connected with definite languages. In the case of the Latin language, which is well known, as well as the history of the Roman nation, the anthropological characters of populations, and cultural variations, the problem of continuity in «Romanic culture» is relatively simple. It is more difficult with the Chinese written language (characters), for here we have a limited language complex confined to the written symbols and the historic anthropological data are not yet complete enough. However, such easy cases as to their documental material are rare, and they are not sufficient for building up a general theory regarding laws of correlation, which may have a reverse force.