Page:The Song of Roland.djvu/20

 I have added a fifth, which I number 1777 a, from the Venice and Paris MSS. This line is quoted in a note by M. de Julleville. I have also followed Muller’s arrangement of the lines 1466–1670, which are displaced in the Oxford, but not in other MSS. The comparative result is as follows:

With these precautions, my translation may, I hope, be used as a “Companion to the Study” of the Oxford MS.

I do not propose to discuss the operation of the Law of Assonance on our language, beyond suggesting that it is an operation under local anæsthetics, which some degree of painfulness may accompany. For variety, there are twenty-two different vowel-endings in the original poem, of which half are feminine or double endings. This number I have not attempted to match. For consonance, I know that in the old language the predominance of vowel over consonant sounds makes it almost always rhyme; and in this belief I have indulged in sequences of rhyme to which the professors of assonance may easily take exception. I claim only that my translation is literal: if it cannot be read with enjoyment, there is no more to be said.

Proper names I have spelt mostly as in the original, anglicising such words as England and Spain—as also Rhone (1583), Toledan (1568), and some others; some I have further varied to improve my assonances. I claim also the privilege