A few remarks about the latter must be given here without any attempt at completeness. That preliminary criticism is necessary seems hardly to need justification; to look at the irreducible inconsistencies and contradictions of a series of statements concerning any given point is enough. But such criticism must not be arbitrary; it must conform to strict rules.[27] [27] The ethnographic sources do not differ essentially from the historical ones. The purpose of a descriptive ethnography (such as is adopted in this monograph) is also nearly identical with the historian's task so far as positive statement of an actual state of things is the aim of both. The rules of criticism of sources apply therefore to our subject as well as to any other historic science. This will become obvious to every ethnographer who reads the excellent manual of MM. Langlois et Seignobos. The proof of the indispensability of such a criticism is to be found there, too. The method of criticism of sources adopted in this paper was accepted after a careful study of this useful book. The first point to which attention must be paid, is to ascertain the exact meaning of a given statement. As many of our informants do not use exact terminology but write in a colloquial language, often spoilt by literary pretensions, we occasionally run the risk of being misled by a word or by a turn of expression.[28] In other words, it never seems advisable to cling blindly to the verbal meaning of a statement before having put it to the test. So, for instance, in the problem whether the natives live in families or tribes--the _family_ and _tribe_ having been exactly defined, a phrase like "the aborigines live in families" may not be accepted as argument, for by the word "family" the author may possibly have understood what we have designated by the word tribe.[29] I shall, as a rule, quote each statement _in extenso_, and give, if necessary, an interpretation or correction.[30] The sense in which a word is used may be, in the majority of cases, easily settled from the context, examples given by the author, and other instances where he uses the same word. When a phrase is hopelessly ambiguous, it is wrong to make any use of it. [28] We obviously do not speak here of such writers as Howitt, W. E. Roth, Spencer and Gillen, authors who had a theoretical training. But, as may be pointed out here, in the present work no writer is excluded, provided that he has had opportunities of first-hand observation, or opportunities of private information from first-hand observers. [29] See below the discussion of this topic, Chap. V., I.