Here then, let it be noted, we are in the presence of exactly the same distinction with regard to the origin of language, as we were at the beginning of this treatise with regard to the origin of man. For we there saw that while we have the most cogent historical evidence in proof of the principles of evolution having governed the progress of civilization, we have no such direct evidence of the descent of man from a brutal ancestry. And here also we find that, so long as the light of history is able to guide us, there can be no doubt that the principles of evolution have determined the gradual development of languages, in a manner strictly analogous to that in which they have determined the ever-increasing refinement and complexity of social organization. Now, in the latter case we saw that such direct evidence of evolution from lower to higher levels of culture renders it well-nigh certain that the method must have extended backwards beyond the historical period; and hence, that such direct evidence of evolution uniformly pervading the historical period, in itself furnishes a strong _primâ facie_ presumption that this period was itself reached by means of a similarly gradual development of human faculty. And thus, also, it is in the case of language. If philology is able to prove the fact of evolution in all known languages as far back as the primitive roots out of which they have severally grown, the presumption becomes exceedingly strong that these earliest and simplest elements, like their later and more complex products, were the result of a natural growth. Nevertheless, as I have said, it is important to distinguish between demonstrated fact and speculative inference, however strong; and, therefore, I will begin by briefly stating the stages of evolution through which languages are now generally recognized by philologists to have passed, without at present considering the more difficult question as to the origin of roots. Supposing we take such a word as “uncostliness.” Obviously here the “un” the “li” and the “ness” are derivative appendages, demonstrative elements, suffixes and affixes, or whatever else we care to call _modifying constants_ which the speakers of a language are in the habit of adding to their root-words, for the sake of ringing upon those words whatever changes of meaning occasion may require. These modifying constants, of course, have all had a history, which often admits of being traced. Thus, for instance, in the above illustration, we know that the “li” is an abbreviation of what used to be pronounced as “like;” the “ness,” however, being older than the English language; while the “un” dates back still further. The word “cost,” then, is here the root, as far as English is concerned—though it can be followed (through the Latin _con-sta_) to an Aryan root, signifying “stand.”