It has been held by many scholars, whose opinions deserve respect, that an extensive literature of _Cantilenae_[12], or short historical ballads, preceded the lengthy epics which we now possess, and was to a certain extent worked up in these compositions. It is hardly necessary to say that this depends in part upon a much larger question--the question, namely, of the general origins of epic poetry. There are indeed certain references[13] to these Cantilenae upon which the theories alluded to have been built. But the Cantilenae themselves have, as one of the best of French literary historians, the late M. Paulin Paris, remarks of another debated product, the Provencal epic, only one defect, 'le defaut d'etre perdu,' and investigation on the subject is therefore more curious than profitable. No remnant of them survives save the already-mentioned Latin prose canticle of St. Faron, in which vestiges of a French and versified original are thought to be visible, and the ballad of Saucourt, a rough song in a Teutonic dialect[14]. In default of direct evidence an argument has been sought to be founded on the constant transitions, repetitions, and other peculiarities of the Chansons, some of which (and especially _Roland_, the most famous of all) present traces of repeated handlings of the same subject, such as might be expected in work which was merely that of a _diaskeuast_[15] of existing lays. [Sidenote: Trouveres and Jongleurs.] It is however probable that the explanation of this phenomenon need not be sought further than in the circumstances of the composition and publication of these poems, circumstances which also had a very considerable influence on the whole course and character of early French literature. We know nothing of the rise or origin of the two classes of _Trouveurs_ and _Jongleurs_. The former (which it is needless to say is the same word as _Troubadour_, and _Trobador_, and _Trovatore_) is the term for the composing class, the latter for the performing one. But the separation was not sharp or absolute, and there are abundant instances of Trouveres[16] who performed their own works, and of Jongleurs who aspired to the glories if not of original authorship, at any rate of alteration and revision of the legends they sang or recited. The natural consequence of this irregular form of publication was a good deal of repetition in the works published. Different versions of the legends easily enough got mixed together by the copyist, who it must be remembered was frequently a mere mechanical reproducer, and neither Trouvere nor Jongleur; nor should it be forgotten that, so long as recitation was general, repetitions of this kind were almost inevitable as a rest to the reciter's memory, and were scarcely likely to attract unfavourable remark or criticism from the audience. We may therefore conclude, without entering further into the details of a debate unsuitable to the plan of this history, that, while but scanty evidence