(3) But the greater number of primary words do not admit of derivation from foreign languages; they must be resolved into the letters out of which they are composed, and therefore the letters must have a meaning. The framers of language were aware of this; they observed that alpha was adapted to express size; eta length; omicron roundness; nu inwardness; rho accent rush or roar; lambda liquidity; gamma lambda the detention of the liquid or slippery element; delta and tau binding; phi, psi, sigma, xi, wind and cold, and so on. Plato’s analysis of the letters of the alphabet shows a wonderful insight into the nature of language. He does not expressively distinguish between mere imitation and the symbolical use of sound to express thought, but he recognises in the examples which he gives both modes of imitation. Gesture is the mode which a deaf and dumb person would take of indicating his meaning. And language is the gesture of the tongue; in the use of the letter rho accent, to express a rushing or roaring, or of omicron to express roundness, there is a direct imitation; while in the use of the letter alpha to express size, or of eta to express length, the imitation is symbolical. The use of analogous or similar sounds, in order to express similar analogous ideas, seems to have escaped him. In passing from the gesture of the body to the movement of the tongue, Plato makes a great step in the physiology of language. He was probably the first who said that “language is imitative sound,” which is the greatest and deepest truth of philology; although he is not aware of the laws of euphony and association by which imitation must be regulated. He was probably also the first who made a distinction between simple and compound words, a truth second only in importance to that which has just been mentioned. His great insight in one direction curiously contrasts with his blindness in another; for he appears to be wholly unaware (compare his derivation of agathos from agastos and thoos) of the difference between the root and termination. But we must recollect that he was necessarily more ignorant than any schoolboy of Greek grammar, and had no table of the inflexions of verbs and nouns before his eyes, which might have suggested to him the distinction.