But, as ages pass on, the root and the addition get so closely combined that neither of them alone has, as a rule, a distinct meaning, and the language arrives at its third stage of grammar-formation. It is not difficult to find examples of a language in this condition, for such is the case with all the languages by which we are surrounded. All the tongues which the majority of us are likely to study, almost all those which have any literature at all, have arrived at this last stage, which is called the inflexional. For instance, though we might divide _actionis_ into two parts _actio_ and _nis_, and say that the former contains the essential idea, and the addition the idea implied by the genitive case, there are only a few Latin words with which such a process is possible, and even in the case of _actio_ the separation is somewhat misleading. In _homo_ the real root is _hom_, and the genitive is not homo-nis but _hominis_. So, again, though we were able to separate ‘asmi’ into two parts--‘as’ and ‘mi’--one expressing the idea of being, the other the person ‘I,’ this distinction is the refinement of the grammarian, and would never have been recognized by an ordinary speaker of Sanskrit, for whom ‘asmi’ simply meant ‘I am,’ without distinction of parts. In our ‘am’ the grammarian recognizes that the ‘a’ expresses existence, and the ‘m’ expresses I; but so completely have we lost sight of this, that we repeat the ‘I’ before the verb. Just the same in Latin. No Roman could have recognized in the ‘s’ of _sum_ ‘am’ and in the ‘m’ ‘I;’ for him _sum_ meant simply and purely ‘I am.’ It was no more separable in his eyes than the French _êtes_ (Latin _estis_) in _vous êtes_ is separable into a root ‘es,’ contracted in the French into ‘ê,’ meaning _are_, and an addition ‘tes’ signifying _you_. This, then, is the last stage upon which language enters. It is called the inflexional or inflected stage, because the different grammatical changes are not now denoted by a mere addition to an intelligible word, but by a change in the word itself. The root may in many cases remain and be recognizable in its purity, but very frequently it is unrecognizable, so that the different case-or tense-endings can no longer be looked upon as additions, but as changes. Take almost any Latin substantive, and we see this: _homo_, a man, the genitive is formed by changing _homo_ into _hominis_, or, if we please, adding something to the root _hom_--which has in itself no meaning; _musa_ changes into _musæ_; and so forth. * * * * * [Sidenote: The five stages in the formation of language.]