Page:Language and the Study of Language.djvu/412

390 on. This is very absurd; and yet, save that the absurdity of it is made more palpable to us by being put in terms of our own language and another with which we are somewhat familiar, it is hardly worse than what has been done, and is done, in all soberness, by men claiming the name of linguistic scholars. It is even now possible for such a man to take an African vocabulary, and sit deliberately down to see what words of the various other languages known to him he can explain out of it, producing a batch of correspondences like these: abetele, 'a begging beforehand' (which he himself defines as composed of a, formative prefix, be, 'beg,' and tele, 'previously'), and German betteln, 'beg' (from the simpler root bit, bet, our bid); idaro, 'that which becomes collected into a mass,' and English dross; basile, 'landlord' (ba for oba, 'master,' si, 'of,' and ile, 'land'), and Greek basileus, 'king:' and the comparer, who is specially versed in the mathematical doctrine of chances, gravely informs us that the chances against the merely accidental character of the last coincidence are "at least a hundred million to one." More than one unsound linguist has misled himself and others by calculating, in the strictest accordance with mathematical rules, how many thousand or million of chances to one there are against the same word meaning the same thing in two different and unconnected languages. The calculation is futile, and its result a fallacy. The relations of language are not to be so simply reduced to precise mathematical expression. If words were wholly independent entities, instead of belonging to families of connected derivatives; if they were of such precise constitution and application as so many chemical formulas; if the things they designated were as distinct and separate individualities as are fixed stars, or mineral species, or geographical localities—then the calculations of chances would be in place respecting them. But none of these things are true. The evidences on which linguistic science relies to prove genetical connection are not identities of form combined with identities of meaning: forms may differ as much as hijo and filius; meanings may differ as much as German bekommen, 'get,' and English become, 'come to be,' and become, 'suit;' form