IV.— The Reconstruction of the Original Chanson de Roland. 



It should be stated, at the very outset, that throughout this ar- 

 ticle the terms Chanson de Roland, and Oxford manuscript (for 

 which we shall use the abbreviation O), signify respectively the 

 Roland, and O, exclusive of the so-called ' Baligant Episode.' Practic- 

 ally all Roland students admit that this episode was an interpolation 

 in that lost intermediary manuscript through which all the extant 

 redactions^ derive from the orimnal.^ 



Although the Chanson de Roland has been studied for three- 

 quarters of a century, many of its problems, including several of 

 the most important ones, are as yet unsolved. In the opinion of 

 the present writer, however, a great number of these problems are 

 solvable if the following thesis be proved — that the original Chanson 

 de Roland was a poem of marked and consistent technical excel- 



^ There is, of course, a difference of opinion as to whether the Carmen 

 de Prodicione Guenonis and the so-called Chronicle of Turpin derive through 

 the same lost intermediary as the other redactions (cf. Gaston Paris, in 

 Romania^ xi, 465—518 ; and Stengel, in Zeitschrift f. roman. Philol.^ viii, 

 499—521). The solution of this problem, however, does not affect the 

 contentions of tliis article. 



^ Cf. especially Scholle, in Zeitschrift f. reman. Philol.., i, 26—40 ; Donges. 

 Die Baligantepisode im Rolandsliede (Marburg, 1879) ; Pakscher, Zur Kritik 

 U7id Geschichte dcs franzosischen Rolandsliedes (Berlin, 1885), pp. 42—59 ; 

 Lindner, in Romanische Forsch-imgen.^ vii, 568—569 ; Gaston Paris. Extraits 

 de la Chanson de Roland^ pp. xxi— xxii. There are slight differences of 

 opinion on the delimitation of the episode ; these, however, do not affect 

 the contentions of our article. We shall consider the episode to consist 

 of the following lines (throughout this article we use the numeration of 

 Stengel's Genatter Abdruck der Oxforder Hs. Digby 23^ Heilbronn, 1878) : 

 2496-2844, 2974-3681, 3975-3987, 3990.— Donges, 1. c, p. 47, says that 

 arguments based on the technic of the episode would confirm his con- 

 clusion as to its spru-iousness. That neither he nor any one else has 

 made a detailed and complete presentation of such arguments is doubt- 

 less due to the Heedlessness of confirmation. Perhaps, however, such 

 confii'mation would have a certain value, for now and then there appears 

 ^n obstinate defender of the episode : for instance, Tavernier, in liis Z7ir 

 Vorgeschichte des altfranzosischen Rolatidsliedes (Berlin, 1903), pp. 155—173. 



