Ridiculous as this notion sounds, it may serve to mark a downward limit from which the rudest types of human speech are not so very far removed.

Their well-known tendency to alter their whole character in twenty years or less is due largely to the fluid nature of primitive utterance; it being found hard to detach portions, capable of repeated use in an unchanged form, from the composite vocables wherein they register their highly concrete experiences.
Thus in the old Huron-Iroquois language _eschoirhon_ means "I-have-been-to-the-water," _setsanha_ "Go-to-the-water," _ondequoha_ "There-is-water-in-the-bucket," _daustantewacharet_ "There-is-water-in-the-pot." In this case there is said to have been a common word for "water," _awen_, which, moreover, is somehow suggested to an aboriginal ear as an element contained in each of these longer forms.

In many other cases the difficulty of isolating the common meaning, and fixing it by a common term, has proved too much altogether for a primitive language.

You can express twenty different kinds of cutting; but you simply cannot say "cut" at all.

No wonder that a large vocabulary is found necessary, when, as in Zulu, "my father," "thy father," "his-or-her-father," are separate polysyllables without any element in common.
The evolution of language, then, on this view, may be regarded as a movement out of, and away from, the holophrastic in the direction of the analytic.