These modifying constants, moreover, are not restricted to suffixes, infixes, and affixes attached to roots, so as to constitute single (or compound) words: they also occur as themselves separate words, which admit of being built into the structure of sentences as pronouns, adverbs, prepositions, &c. And they may occur likewise as so-called “auxiliary verbs,” in the case of some languages, while in the case of others their functions are served by grammatical “inflection” of the words themselves. Thus, according to the “genius” of a language, its roots are made to lend themselves to significant treatment in different ways, or according to different methods. But in all cases the roots are present, and serve as what may be termed the back-bone of a language: the demonstrative elements, in whatever form they appear, are merely what I have termed modifying constants. From this general fact we may be prepared to expect, on the theory of evolution, that in all languages the roots should be the oldest elements; those elements which serve only the function of “demonstrating” the particular meaning which is to be assigned to the roots on particular occasions, we should expect to have been of later growth. For they serve only the function of giving specific meanings to the general meanings already present in the roots; and, therefore, in the absence of the roots would themselves present no meaning at all. Consequently, as I have said, we should antecedently expect to find that the roots are the earliest discoverable (though not on this account necessarily the most primitive) elements of all languages. And this, as a general rule, is what we do find. In tracing back the family tree of any group of languages, different demonstrative elements are found on different branches, though all these branches proceed from (_i.e._ are found to contain) the same roots. Of course these roots may be variously modified, both as to sound and the groups of words to which in the different branches they have given origin; but such divergent evolution merely tends to corroborate the proof of a common descent among all the branches concerned.[147] I have said that all philologists now agree in accepting the doctrine of evolution as applied to languages in general; while there is no such universal agreement touching the precise method or history of evolution in the case of particular languages. I will, therefore, first give a brief statement of the main facts of language-structure, and afterwards render an equally short account of the different views which are entertained upon the question of language-development. Or, to borrow terms from another science, I will first deal with the morphology of the main divisions of the language-kingdom, and then proceed to consider the question of their phylogeny. * * * * *