Page:The Song of Roland.djvu/26

 frankness which befits the scholar whether old or young. It seems to me that this is not merely in detail but in general effect, the most faithful version I have ever seen of the great Song that Turoldus did something absolutely uncertain with.

The obstacles to assonance in English, and its probable disagreeables, are many and various. In the first place (and no wise person will minimise or misunderstand this) we “have not proved it;” it has never been an accepted and familiar form with us. In the second, we know it best as a failure of something else—a slovenly or careless substitute for rhyme. Thirdly, there are certain stumbling-blocks hard to get over or avoid in the sound-habits (I never use the word phonetic if I can possibly help it) of English as a language. We are so fond of throwing back the accent that we have comparatively few words sounding fully on the ultimate. The habit of slurring vowel-sound, though not so usual with well educated and well-bred people as phoneticians seem to think, does to too great an extent deprive us of the sharp, full ringing effect that assonance requires, and that Old French, and Spanish of all times, afford. Lastly, there is the multiplicity,—valuable in itself and not to be sacrificed to any simplifying simpleton,—of our sound-values for the same vowel. All these are dangerous lions in the path (to vary the comparison), and some of them are disagreeable beasts as well as dangerous ones. Captain Scott Moncrieff has, I think, managed the stumbling blocks, and met the beasts, with a most creditable amount of skill and courage and with a very considerable success. He has had, of course, to avail himself of some licenses, none of them, however, unjustified by good precedent.