650 ROMANCE true notion of the richness and copiousness of Frankish arly rench allads. romance. We know from Eginhard that the Frankish heroic ballads were reduced to writing by Charlemagne's order, and thus the first step was taken which led to the creation of similar ballads about himself and his principal warriors. His own large and catholic spirit seems to have embraced all the people within his dominions, and thus indirectly brought about the official employment of the French lan- guage in the famous compact between his grandsons in 841. He was then only twenty-eight years dead, yet his influence was still so mighty that even the Gauls and Aquitanians are declared in a 9th-century chronicle to have gloried in bearing the name of Franks. This both implies an amal- gamation of the two races more complete than is usually believed, and accounts for the creation of French as well as of Frankish ballads on the life and exploits of Charle- magne in the second half of the 9th century. The number of persons who could speak only Theotisc (Teutonic lan- guage) and of those who could speak the two languages was of course constantly diminishing, and the chanson de geste soon displaced the Heldengedicht within the limits of modern France. Of all the French ballads current in the 9th and 10th centuries some have perished utterly, others survive only in later refusions ; the most ancient now extant is the Chanson de Roland, in the modified form which was given to it soon after 1066 by a Norman called Turold. This poem contains so many references to others on Charlemagne and his dome pairs or paladins as to make it certain that such a ballad-literature existed in the 9th and 10th centuries. None of the existing chansons de geste represents those older forms; all are rifaci- menti of the 12th and 13th centuries, and bear evidence of additions, interpolations, and arbitrary changes. In the 13th century we find the older episodical ballads re- arranged in the form of " cyclic poems," and falling into three groups, which take each its name from the central personage or subject. One is the Geste of the king (Charlemagne, his father, and grandfather); the next is the Geste of Provence or of Garin de Montglane ; and the third, the most heterogeneous, is the Geste of Doon of istorico- Mayence. Each of these is composed of many separate ythical parts, but the first may be generally described as the entire r ' J8 ' mythical history of Charles the Great, his family, and his faithful peers; the second a separate and independent set of narratives concerning his conquest of Narbonne; the third a history of his wars with rebellious vassals and with traitors, including Ganelon, through whom the peers were defeated and slain at Roncesvalles. The number and names of the peers are variously given in nearly all the poems, but Roland and Oliver are included in all the lists, united as in a proverbial English phrase. Roland is the daring warrior, Oliver the wise one; the one is the Achilles, the other the Ulysses of the Carolingian epopee. Many of the early chansons give the name of Turpin, arch- bishop of Rheims (an actual historical contemporary of Charlemagne), as one of the pairs, warrior and priest com- bined ; and there is a chronicle bearing his name which has furnished the later romancers with a goodly propor- tion of their matter. This Pseudo-Chronicle of Turpin was written in Latin, by various hands and in various places between 1000 and 1150, being apparently constructed from the chansons for the purpose of forging history to suit monastic ends. It took its final and existing form between 1160 and 1180, when edited by Geoffroy de Brueil, and was for many centuries regarded as actual his- tory. This work was not the first so-called history which embodied monastic fiction in the narrative of Charlemagne's career. A monk of St Gall wrote about 890 a chronicle De Gestis Karoli Magni, based partly upon oral tradition, in which certain fabulous incidents appeared for the first time, such as Pippin's fight with the lion, and the conversa- tion about the Iron Emperor between Ottokar the Frank (better known as Ogier the Dane) and the Lombard king Desiderius, on the walls of Pavia, when Charlemagne was advancing to besiege it. Another fabulous incident of great moment in the romances is Charlemagne's sup- posititious journey to the Holy Land, which was related for the first time by Benedict, monk of St Andre 1 , about 968, in his Descriptio qualiter Carolus M. Clavum et Coronam Domini a Constantinopli Aquugrani attulerit. A great deal of historical truth underlies the absurdities of the Turpin Chronicle and the rhapsodies of the chansons de geste. Compare ROLAND, LEGEND OF. In fact all the older poetic literature of this cycle is based upon purely historical events and real personages ; it is only at a later date that the events are multiplied or variously misapplied, and that the personages also are arbitrarily distorted and augmented, according to the fancy and local sentiments of the various writers. As for the language in which the older poems were written, the idea that they were chiefly the work of troubadours in the langue d'oc is now aban- doned. Gaston Paris holds the curious theory that the French language (langue d'oil) was popularly current over the north of Italy, instancing the works of Rusticien of Pisa as an illustration, besides certain works in Italianized French which belong to this class. Such a notion cannot be accepted readily, as we know that the langue d'oc was the general language of southern France, western Spain, and north-west Italy. But it is possible that the French language (langue d'oil) may have been used as a general literary vehicle for the Charlemagne cycle of poetical fiction, and that Rusticien or Rustighello may have com- piled an abridgment of the Frankish stories. Such a work, if it ever existed, has perished ; and it is in the Italian language of Tuscany that we find the first prose compila- tion of Carolingian romance. The Reali di Francia (Princes of France), if it had been completed, would have occupied a corresponding position to the Morte Arthur of the British cycle ; for, while no such popular compilation appears to have ever been made in France itself and in the French tongue (unless the late 15th-century Fiera- bras may be considered to take that rank), the Reali in verse and in prose was current in Italy early in the 14th century. From some peculiarities in the language it is conjectured that the author, although writing in Tuscan, was a Venetian. In France at the same period we find only the separate fictions, mostly in verse, but a few in prose. The first French prose compilation of the whole cycle was made by David Aubert in 1458 for Philip of Bur- gundy ; but it was dead-born and has never been printed. The second, in three books, was made a few years later by Jean Bagnyon, for Henri Bolomier, canon of Lausanne ; it was first printed in 1478, and is entitled in some editions La Conqueste que fist Charlemagne es Espaignes, and in others Fierabras ; both titles are insufficient, having appa- rently been merely created to supply the lack of a general heading. The first section is a summary chronicle of the history of the Franks from Clovis to Charlemagne, the second an abridgment of the old poem of Fierabras, and the third an account of the Spanish expedition, taken from the Pseudo-Turpin. This work became very popular in and out of France, and most readers during the 16th and 17th centuries derived their entire knowledge of the Charlemagne romance from it. Although the French prose works of the cycle were for the most part very late in their construction, there were three French prose transla- tions or adaptations of the Turpin Chronicle executed soon after 1200. The names as well as the offices of the douze pairs varied
