The Song of Roland

Spread the love

Although some people—usually those with a religio-socio-political axe to grind—claim great accuracy for “oral history,” the earliest examples, epics, don’t offer much support for such claims.  In their creation, the memorable events of a culture are told and retold, growing longer, more dramatic, and larger-than-life as time goes on.  There is probably never an intention to falsify history; in fact, the notion of “history”—as a record of the actual events of a period—insofar as they can be determined by comparing and evaluating source materials and known facts—doesn’t exist at all for them.  Instead, there is the intention of telling the story of certain events in a way that illuminates their significance by making the narrative more rich and meaningful, since that is often what is important from a society’s perspective.  In short, it is myth-making.  (And, of course, there is also the intention of staying in business as a bard and increasing one’s prestige by making it a really good story.)


We see this in the Iliad.  To the extent that there is an historical basis for this epic, it was probably no more than a hit-and-run, marauding expedition in some remote area:  but that area inevitably became rich and fabled Asia Minor; the town attacked became a famous and powerful city, Troy; the cause became justice and revenge instead of loot; and the motive became adultery and the violation of the guest-host relationship rather than lust for power.  There isn’t a shred of evidence for any of this except the poem itself.  (It is a very great work of literature—but for reasons that have little to do with the notion of the “epic.”)  The Odyssey has even less going for it.  It is really no more than a frame narrative—Odysseus comes home and reclaims his wife after years of wandering—which is packed with every intriguing bit of mythology that the poet can find, becoming a sort of Greek Arabian Nights in the process.


Literary epics are those written at a later stage of culture (or in an urban center) as opposed to those told at an earlier stage of culture (or in rural areas where few people can read or write).  Oral epics differ from their literary imitations in important ways.  To begin with, there is a complicated set of poetic devices that must be learned, and passed down to apprentices, to permit one to tell the same long story over and over.  Repetition of various kinds is obviously an important way of accomplishing this.  (The oral technique has been studied for less than a century, beginning with the work of Milman Parry in the 1930’s.  Its discovery has proven to be one of the most important advances in understanding the epic form.)  But literary epics don’t actually need the oral devices—they can be written by authors sitting at a desk—and their authors probably didn’t even know that that an “oral technique” existed.  What they did know was that the epics which had come down to them from the past had a recognizable “style” and that is what they tried to imitate.


Another important characteristic of the literary epic is likely to be the presence of a deliberate social or political motive for writing it.  This isn’t true for the author(s) of the the Iliad, an oral epic, , who are telling a story which glorifies the Achaean past but does little more.  In contrast, Virgil, in writing the literary epic, the Aeneid, chooses the myth of Aeneas, the mythical survivor of the royal family of Troy—and who was, even more-mythically, the ancestor of Julius Caesar.  Virgil does this because it sets Rome apart from the remnants of the Hellenistic kingdoms, and in opposition to them, further glorifying the Romans.  Aneneas’ legendary offspring supposedly founded Rome, and what’s more, Augustus (the first emperor and the beneficiary of Virgil’s propaganda) was Julius Caesar’s adopted son, so you see how nicely this fiction works for him.  In writing this, Virgil is serving as one of those loathsome creatures, a poet laureate, glorifying Augustus as the appropriate heir to Rome, and justifying Roman domination of the known world.


Of course, all of the oral epics we have were written down at some time, and that raises the possibility that they have undergone an additional period of creation as a literary epic after having originated as an oral epic.  This is probably not the case with the Iliad, for we are told that the oral poets who knew the story were gathered together in the sixth century BCE to work out a “best” version, which was then written down.  But it does appear to be the case with Beowulf.


Although the Anglo-Saxon pagans were converted to Christianity in about the seventh century, many pagan beliefs survived for a long time.  (Just look, at the adoption of the Christmas tree, and at the dates of Christmas and Easter which stem from those of pagan celebrations.)  A poem like Beowulf might well have been preserved by monks who loved the story, just as the Greek romances (rather trivial bodice-rippers) were preserved by monks, who justified doing so by “interpreting” them in Christian terms.  In the case of Beowulf, such monks might have added a Christian gloss to the story, leaving a poem that is partly oral and pagan, and partly literary and Christian.  There is some controversy about this, but I think it is probably true.


La Chanson de Roland has some parallels with Beowulf.  It is probable that it was first composed in the eighth century CE and was written down later, reaching its final form in an eleventh-century manuscript that reveals the work of two scribes.  The principal focus of the poem, the destruction of Charles’ rear-guard, was an actual event that occurred on August 15, 778 CE.  The hero of the poem is Roland, and a ninth century history, the Vita Karoli (“The Life of Charles”), tells us that one of the dead was Hruolandus, Prefect of the Breton Marches.  (The “h” is a linguistic feature similar to the “h” which appears in many names in Beowulf, such as “Hrothgar.”)  The oldest existing manuscript (The Oxford manuscript, Digby 23) was probably written between 1225 and 1250, and is in Anglo-Norman French.  These things (the Breton Marches and Anglo-Norman French) suggest that the poem may well have been first told and written on the west coast of France.  Let’s look at the background of the story.


The Roman Empire had fallen in 410 and there was a general breakdown of the Roman institutions, which was greatly accelerated by invasions of Germanic tribes including the Goths and Vandals.  (German scholars like to call these “folk wanderings.”)  This Early Medieval Period has sometimes been called the “Dark Ages” because of the loss of literacy and art.  In many areas, the people would have been only a few generations away from a Germanic, warband structure in which they owed loyalty to a tribal leader.  Charles’ grandfather, Charles Martel, had largely unified Western Europe, and feudalism had become an important institution.  He also defeated the Islamic forces at Tours in 732 CE, and in other battles throughout southern France.  (Some Islamic power remained in Spain until 1492.)  Charles (who only became known as Charlemagne after being crowned Holy Roman Emperor in 800 CE) began what has been called the Carolingian Renaissance, bringing in scholars from throughout Europe, but this would have had little immediate influence throughout much of his empire.  Among other things, such scholarship was largely confined to clerics.  This roughly describes the world in which the historical events that inspired the poem take place.


Sulaiman ibn Yaqzan ibn Al-Arabi, who was the governor of Barcelona and Gerona, had entered into a revolt against the Emir Abd al Rahman of Cordoba.  In 777 CE, he went to Charles with a proposal:  he would submit to his authority if Charles would send an army to assist him.  (Charles was 37 years old and had not yet taken the title of Emperor.)  In 778 CE, he entered Spain, captured Pamplona, and advanced on Saragossa.  The governor of the region was supposed to open the city to Charles, but he reneged on his agreement, and Charles began a siege which lasted for a month-and-a-half.  At this point, Charles decided to return to France.  Basque bandits attacked the baggage train, took what they wanted, and scattered into the surrounding hills.  Hruolandus was one of the dead.  There is very little glory in the bare bones of this account.


Now, suppose that you are a semi-barbaric teller of heroic tales living in Brittany a few generations after these events.  Some of Hruolandus’ men had returned with confused and contradictory accounts of the slaughter, but none of them would have wanted it to seem an absurd, almost pointless affair.  Many tales of heroism were told.  Hruolandus was said to have been betrayed, to have fallen into a trap prepared for him:  how else could such a great warrior have been defeated!  Stories that focus on different parts of the event would have circulated widely.  They have the character of folk-songs, long on feeling and short on coherent detail:  Roland is a great hero, loyal to the king; returning from Spain he is given a task; his step-father betrays him and he is killed in an ambush.  Al-Arabi and the governor of Saragossa are collapsed into one person:  Marsilion.  Roland is not killed by Basque brigands but by the flower of the Pagan (Islamic) cavalry.  Since Marsilion is held up as the representative of the Pagan enemy, he is said to be trying to defeat Charles instead of  revolting against the Emir.  (And, of course, the actual facts aren’t known to the people who tell the stories and sing the songs.)


You, O barbaric teller of tales, collect these and begin to fit the details together into an ur-Roland, a simplified but dramatic account of Roland’s death, an epic,  something long and dignified, something worthy of the theme:


I      Charles has retaken all of Spain from the Pagans except Saragossa, and Marsilion offers hostages to persuade him to go back to France.  Ganelon, Roland’s stepfather, plots with Marsilion to have Roland delegated to the rear-guard.

II     Roland’s men are killed by Marsilion’s cavalry.  Roland has burst a blood vessel in his temple from the force with which he has blown his horn, the Oliphant.  He alone survives against the remaining Pagans, and finally, they flee.  Then Roland dies of the wound in his head.  (He can’t be said to have been killed by Pagans—he’s too heroic for that.)

III    Charles returns and vanquishes the Marsillion and his Pagans.

IV    He punishes Ganelon for his treason.


Now, a hundred-and-fifty or two-hundred years have passed, and you are now a rustic teller of tales intent on recasting the epic to reveal its significance as a turning point in the conflict between Christianity and Islam.  This is the kind of motive that might underlie the creation of a literary epic, even though you are not deeply learned.  You will need to make a number of changes:


  1.  You will create a parallel between Ganelon and Judas.
  2.  You will create a parallel between Roland and Jesus.
  3.  You will bring from the East the ultimate representative of Islam to engage Charles in an apocalyptic battle:  Christianity against Islam.
  4.  Charles will frequently appear to be like Jehovah (and sometimes like Jesus).



This is the version that, with further editing of the existing version, becomes the Chanson de Roland that we know.


Let’s look first at some of this “further editing,” and then we will take up these changes in structure.  After listening to Marsilion’s emissary, Charles asks his barons for advice in laisse (verse) 13.  Roland recommends that he refuse Marsilion’s offer; after all, Marsilion had given his word on a previous occasion and then killed the envoys that Charles sent to him.  Ganelon immediately calls Roland a fool and says that it is in Charles’ best interest to accept the offer.  After some discussion, Charles agrees with Ganelon and asks his barons who should be sent to Marsilion as an envoy.  Roland volunteers, but Charles says that he won’t send his most important warriors for the task.  After more volunteering by his comrades, Roland suggests that Ganelon would be a good choice, and Ganelon is outraged insisting that it is a death-sentence for him—that any envoy will be killed.


There is a deep contradiction between Ganelon’s saying that Marsilion can be trusted, and his later claim that any envoy will be killed.  If this isn’t carelessness on the part of the scribes—and we will see scribal carelessness in a moment—it must mean that Ganelon’s advice was hypocritical:  he had taken a position opposed to Roland’s out of his hatred for him, even though he didn’t believe what he what he was saying.  (Regardless of how it occurs, it is obviously inconsistent.)  When Ganelon is given the emperor’s glove as a symbol of his authority, he accidently drops it, and when the Franks wonder what this may portend, he says:  “Lords, you’ll be hearing news.”  (laisse 25, line 336)  This is best explained as a foreshadowing of his intention to do evil.


After Ganelon returns from plotting treason with Marsilion, he proposes that Roland command the rear-guard.  (The death of Roland would weaken Charles and satisfy Ganelon’s wounded vanity:  Roland had laughed at him.  And in laisse 60 Roland says:  “did you think the glove would fall from my hands/as the staff fell from yours.” (Lines 764-5)  And in the very next laisse he is offered a bow and he says:  “And no man here, I think, will say in reproach/I let it drop, as Ganelon let the staff drop.”  (Lines 768-9)  You will remember that Ganelon had dropped the glove in laisse 25, but he did not drop the staff which is given to him in the following laisse.  Some critics have said that the reference to the staff being dropped shows that laisse 60 is a late addition made by an inattentive scribe.  They also argue that a reference to the staff is added to laisse 61 to make the previous error more consistent.  I don’t find this very convincing.  The best reason for thinking that all three laisses (60, 61, and 62) are inappropriate late additions is their content:  in them, Roland is represented as being fearful for his life and furious that he is chosen as Charles’ envoy:  This contradicts laisse 59 where he responds to his nomination in the way a brave and loyal vassal would:  “Lord Stepfather, I have to cherish you!/You have had the rear-guard assigned to me.”  As a warrior hero, he is glad to have trust placed in him, glad of the chance to win glory; moreover, he had already volunteered.  It is clear what has happened:  an insensitive scribe added the three suspect laisses to create a parallel to those in which Ganelon is outraged because he has been chosen as the envoy to Marsilion.  Such parallels are a frequent occurrence in the poem—but this is not the place for one.


Now, let’s look at the structural changes.  First, the basic parallel between Ganelon and Judas.  In the twelfth laisse, as we have seen, Charles calls for his barons, wanting their advice as to whether he should accept Marsilion’s offer.  There are twelve of these trusted followers (in this, Charles is like Jesus), and they are named for us:  Ogier, Turpin, Richard . . . Oliver, and then, “Ganelon came, who committed the treason.”  (Line 178)  (In Luke (6: 16) Jesus chooses twelve apostles from among the disciples, and they are named for us:  Peter, Andrew, James . . . Judas the son of James, and then “Judas Iscariot, who became a traitor.”)  Ganelon’s treason is emphasized throughout the rest of the poem.  For example, in laisse 38 when Marsilion agrees to plot Roland’s death, he takes Ganelon by the hand and leads him into an orchard where, “they plotted that criminal treason.” (Line 511)  In laisse 46 Ganelon swears on the holy relics in his sword that he will betray Roland.  And in laisse 56 Charles dreams that Ganelon grabs a great lance from him and shakes it until it flies into splinters.  The meaning of the dream becomes clear to him when Ganelon proposes that Roland command the rear guard in laisse 58.  Charles bursts out, “you are the living devil.” (Line 746)


And now we have the basic parallel between Roland and Jesus.  At the height of the battle against Marsilion’s forces we have this:


it is, in truth, a trembling of the earth.

From Saint Michael-in-Peril to the Saints,

from Besançon to the port of Wissant,

there is no house whose veil of walls does not crumble.

A great darkness at noon falls on the land,

there is no light but when the heavens crack.

No man sees this who is not terrified,

and many say:  “The Last Day!  Judgment Day!

The end!  The end of the world is upon us!”

They do not know, they do not speak the truth:

it is the worldwide grief for the death of Roland.


The death of Jesus is described in similar terms in Matthew (27:45-55):


Now from the sixth hour there was darkness over all the land until the ninth hour . . . And Jesus cried again with a loud voice, and yielded up his spirit.  And behold, the curtain of the temple was torn in two, from top to bottom; and the earth shook, and the rocks were split; the tombs also were opened, and many bodies of the saints who had fallen asleep were raised . . . .


We are then forewarned of the desperate struggle between Christianity and Islam that is yet to come by a dream filled with apocalyptic imagery which comes to Charles in laisse 185:


All through the night the angel stands at his head;

and in a vision he brought the King dread tidings

of a great battle soon to come against him:

revealed to him its grave signification:

Charles raised his eyes and looked up to the sky,

he sees the thunder, the winds, the blasts of ice,

the hurricanes, the dreadful tempests,

the fire and flames made ready in the sky.

And suddenly all things fall on his men.

Their lances burn, the wood of ash and apple,

and their shields burn down to their golden bosses,

the shafts of their sharp spears burst into pieces,

then the grating of hauberks, helmets of steel.

He sees his warriors in great distress—

leopards and bears furious to devour them,

serpents, vipers, dragons, demons of hell,

swarms of griffins thirty thousand and more,

and all come swooping down upon the French


Charles wants to help, but his way is blocked by a gigantic lion.  It leaps upon him and it is not clear who is winning.  The apocalyptic imagery tells us that this is no mere battle between armies.  The lion is Charles’ counterpart on the Pagan side–Baligant–and now we need to bring him upon the scene.


After Charles returns to help Roland (who is dead by the time he arrives) and defeats Marsilion’s army, there is almost a new beginning to the rest of the poem.  To see that this is so, let’s look back at the first lines of laisse 1 at the beginning of the epic:


Charles the King, our Emperor, the Great,

has been in Spain for seven full years,

has conquered the high land down to the sea.


Now look at this “second beginning” in laisse 189:


The Emperor with all his mighty host

has warred seven full years in the land of Spain,

Takes many castles, takes city upon city.


And we are told, again, of Marsilion’s efforts to find a way to make Charles withdraw.  And then we are told that In the first year, Marsilion had sent letters to the Amiral, Baligant, in Babylonia pleading for help.  In response, Baligant has called for troops from forty realms, and at this moment his army is nearing land.  It is as if we are beginning a sequel to the ur-Roland in which there is no treachery, no Ganelon to punish.  This is the “Balignant episode.”  It runs from laisse 189 through laisse 263, and it takes only the slightest of “bridges” to connect it at its beginning and end to the ur-Roland.  In fact, I find it far easier to think of the poem as being two stories fused together than one rewritten, but either is possible.  (Charles’ dream in laisse 185 is possibly one such “bridge.”)


When Baligant hears of Marsilion’s defeat, he goes to him, and Marsilion gives him all of Spain.  Meanwhile, Charles has begun to search the battlefield at Rencesvals, looking for Roland’s body.  When he finds it, he has the hearts of Roland, Archbishop Turpin, and Oliver cut from their bodies, wrapped in silk brocade, and placed in a marble coffin.  The bodies are washed and wrapped in shrouds.  At this moment, Baligant’s vanguard appears.  Battle ensues, and Baligant and Charles meet and begin to fight.  Baligant offers Spain to Charles if he will serve him.  Charles offers Baligant friendship if he will become a Christian.  Neither accepts the other’s offer, and after exchanging terrible blows, Charles kills Baligant, and the Pagans flee.  Marsilion dies of grief and shame in a laisse that is left as part of the concluding “bridge,” and his soul is taken by the “living devils.”


Now Charles can get back to his business.  He leaves the remains of Roland and his two friends at Blaye, and at his home in Aix he arranges the trial of Ganelon.  Pinabel, Ganelon’s champion, intimidates the remaining barons, threatening them with death if they find him guilty.  But Tierri, one of the weaker knights, breaks the impasse by condemning Ganelon, and a trial by combat is decided upon.  God will decide who is right by favoring the winner, and as it happens, Tierri defeats Pinabel, and Ganelon is pulled apart by war-horses as punishment for betraying Charles.


What has been accomplished by adding the Balignant story?  We have moved from an epic of mainly local significance (but rooted, to a degree, in fact) to one of importance for all of Christendom:  to a complete victory over Islam instead of a local victory (and this has no factual basis at all).  In the eyes of the writer(s), it is also a victory of good over evil.  The poem becomes a sort of passion play, representing the death of the favored son (nephew, actually) and the revenge taken by the father (uncle) upon the sources of evil.  Throughout there are symbolic representations of virtue, doubt, death, etc.  (It is important to note that in the warband society, the children of a widowed sister become particularly important to their uncle!)


I have no proof for this.  All we have is the poem as written, and there is no evidence to suggest that there was an ur-Roland (and there is no evidence for a pre-existing sequel like the Balignant episode)–but think about it.  Isn’t this the way that it must have happened?  Would the Baligant episode have been present at the beginning in a poem composed by poet(s) celebrating a local hero?  And as pure invention, doesn’t it bear the marks of a literary rather than an oral epic—after all, it is entirely fictitious?  Alfred Lord, a colleague of Millman Parry’s believes that the remains of the oral technique are too prominent for The Song of Roland to be entirely a literary epic, although some will dispute that.  I think that it is clearly the case.


The fact that the poem (including the Balignant episode) is the product of rustic rather than learned poet(s) is supported by the improbability of the account.  We can readily accept that the heroes are capable of amazing deeds.  That is expected from such a poem.  But you must remember that the Oxford Roland is thought to date from the first half of the twelfth century, and Charles was crowned Emperor on Christmas Day in 800 CE.  In other words, the completion of the Oxford Roland follows the events by only about 400 years, though Charles is said to be 200 years old at the time of the battle.  He is said to have retaken all of Spain from the Pagans—who will still be there (to some degree) for another 700 years.  Roland’s 20,000 Franks are said to have defeated 200,000 Pagans; and when only 60 Franks are left, 100,000 of the pagans flee from them; at one point, Roland, Oliver, and Turpin kill 4,000 by themselves; at the end, only Roland, Gautier and Turpin are still alive, but they stand off 41,000 who are too terrified to close with them; Turpin, himself, kills 400.  Etc., Etc.  (Remember what I said at the beginning about the accuracy of “oral history.”) This doesn’t sound at all like a poem produced at a literate court which has relations with other states; it sounds like a poem produced in the hinterlands where the audience is far more credulous and there is a personal interest in the topic . . . perhaps somewhere in Brittany?



The quotations from Roland are from The Song of Roland, translated by Frederick Goldin.  W. W. Norton & Company:  New York, 1978.  I like the translation, but I find the introduction far-fetched.  We are supposed to believe that this poem marks the turn from the old feudal political structure to a more modern nation-state kind of structure:  Ganelon’s betrayal of Charles is moot, and God’s championship of Tierry shows that the King must be supreme.

My students have always enjoyed reading the poem, even though it is long and complicated, sometimes tedious, and sometimes moving.  When I first taught it in an undergraduate class on medieval world literature (1987?), I recognized that the Balignant episode was much better treated as a later addition to the story, and began to work out the interpretation I have given here.  As it turned out, I was not entirely alone.

Joseph J. Duggan’s The Song of Roland:  Formulaic Style and Poetic Craft (University of California Press:  Berkeley, 1973) is a splendid book.  I discovered it when I began to write this, and I wish that I had found it much earlier.  His discussion of the oral technique and how it applies to Medieval French, and this poem in particular, is excellent.