The result of the foregoing observations is to show that the formal varieties of language present no obsta- cle to the theory of a common origin. Amid these va- rieties there may be discerned manifest tokens of unity in the original material out of which language was formed, in the stages of formation through which it has passed, in the general principle of grammatical expres- sion, and, lastly, in the spirit and power displayed in
the development of these various formations. Such a result, though it does not prove the unity of language in respect to its radical elements, nevertheless tends to establish the a priori probability of this unity; for if all connected with the forms of language may be re- ferred to certain general laws, if nothing in that depart- ment owes its origin to chance or arbitrary appoint- ment, it surely favors the presumption that the same principle would extend to the formation of the roots, which are the very core and kernel of language. Here, too, we might expect to find the operation of fixed laws of some kind or other, producing results of a uniform character; here, too, actual variety may not be incon- sistent with original unity.
(II.) Before entering on the subject of the radical identity of languages, we must express our conviction that the time has not yet arrived for a decisive opinion as to the possibility of establishing it by proof. Let us briefly review the difficulties that beset the question. Every word as it appears in an organic language, wheth- er written or spoken, is resolvable into two distinct ele- ments, which we have termed predicaljle and formal, the first being what is commonly called the root, the second the grammatical termination. In point of fact, both of these elements consist of independent roots ; and in order to prove the radical identity of two languages, it must be shown that they agree in both respects, that is, in regard both to the predicable and the formal roots. As a matter of experience, it is found that the formal elements (consisting, for the most part, of pronominal bases) exhibit a greater tenacity of life than the others; and hence agreement of inflectional forms is justly re- garded as furnishing a strong presumption of general radical identity. P^ven foreign elements are forced into the formal mould of the language into which they are adopted, and thus bear testimony to the original char- acter of that language. But though such a formal agreement supplies the philologist with a most valuable instrument of investigation, it cannot be accepted as a substitute for complete radical agreement: this would still remain to be proved by an independent examina- tion of the predicable elements. Tlie difficulties con- nected with these latter are many and varied. As- suming that two languages or language-families are under comparison, the phonological laws of each must be investigated in order to arrive, in the first place, at the primary forms of words in the language in which they occur, and, in the second place, at the corresponding forms in the language which constitutes the other mem- ber of comparison, as has been done by Grimm for the Teutonic as compared with the Sanscrit and the classical languages. The genealogy of sound, as we may term it, must be followed up bj' a genealogy of signification, a mere outward accordance of sound and sense in two terms being of no value whatever, uidess a radical affin- ity be proved by an independent examination of the cognate words in each case. It still remains to be in- quired how far the ultimate accordance of sense and sound may be the result of onomatopojia, of mere bor- rowing, or of a possible mixture of languages on equal terms. The final stage in etymological inquiry is to decide the limit to which comparison may be carried in the primitive strata of language — in other words, how far roots, as ascertained from groups of words, may be compared with roots, and reduced to yet simpler ele- mentary forms. Any flaw in the processes above de- scribed will, of course, invalidate the whole result. Even where the philologist is provided with ample materials for inquirj' in stores of literature ranging over long pe- riods of time, much difficulty is experienced in making good each link in the chain of agreement; and yet in such cases the dialectic varieties have been kept with- in some degree of restraint by the existence of a liter- ary language, which, by impressing its authoritative stamp on certain terms, has secured both their general use and their external integrity. Where no literature exists, as is the case with the general mass of languages
TONGUES, CONFUSION OF 4 TO TONGUES, CONFUSION OF
ill the world, the (liffioultics are infinitely increased by the combined effects of a prolitic j^rowth of dialectic forms, and an absence of all means of tracing out their progress. Whether, under these circumstances, we may reasonably expect to establish a radical unity of laii- guai'e is a question which each person must decide for himself. Much may yet be done by a larger induction and a scientific analysis of languages that are yet com- paratively unknown. The tendency hitherto has been to enlarge the limits of a •' family" according as the ele- ments of atlinity have been recognised in outlying mem- bers. These limits may perchance be still more en- larged by the discovery of connecting-links between the language-families, whereby the criteria of relationship will be modified, and new elements of internal unity be discovered amid the mini fold appearances of external diversity.