402. Another supposition, akin to the last discussed, is that in the words which originally had initial ϝ the ordinary effects of an initial consonant remained after the sound itself was no longer heard. Such a phenomenon would be by no means without parallel in language. In French, for instance, elision is not allowed before certain words beginning with h, as le héros, la hauteur, though the h is no longer pronounced. Similarly, then, it may be held that the facts of Homeric meter only prove the habit or rule of treating certain words as if they began with ϝ.
On the other side it may be urged that the h of héros, hauteur, etc., is only traced in one way—by hiatus, and that only in a small number of combinations; whereas the ϝ not only protects hiatus, but also makes position. Moreover the retention of a traditional usage of this kind is very much easier in an age of education. Anomalies which would naturally disappear in a few years are kept alive by being taught to successive generations of children. It seems difficult to believe that the ϝ would have kept its present place in the memory of the poets unless it were familiar, either to the ear as a present sound, or to the eye as a letter in the written text.