[In the following essay Kay investigates the interaction of character and plot in various chansons de geste, particularly in Raoul de Cambrai, and argues that neither aspect holds a simple priority over the other.]

The relationship between character and plot confronts the literary critic with a chicken and egg problem of ostrich proportions. How far can we dissociate what a literary character is from what he does, and supposing such dissociation to be operable, which is the proper, primary level of description in any text? Should we say, for instance, that Othello is a character liable to violent jealousy, and so can be set up to kill his ever-loving wife, or do we say that the killing of an unjustly suspected wife leads us to perceive her murderous husband as violently jealous? Do texts differ in this regard? As with Shakespeare plays, some chansons de geste are assigned titles featuring a principal character (Chanson de Roland, Aiol, Gaydon, Chanson de Willame), others an aspect of the action (Charroi de Nîmes, Prise d’Orange, Siège de Barbastre, Aliscans). The prologues of some, like the Couronnement de Louis, announce the matter that is to follow as centered primarily on character:

De Looÿs ne lerai ne vos chant
Et de Guillelme au cort nes le vaillant,
Qui tant soffri sor Sarrazine gent;
De meillor home ne quit que nus vos chant.

[7-10]¹

Others, of which Girart de Roussillon is an example, summarize part of the plot:

Ceste muet de Folcon e de Folchui,
Et de Girart le conte la vos revui,
Quant prestrent guere a Charle el e li sui,
Per quant sunt espandut de sanc mil mui.

[12-15]

Critics of medieval literature writing in a traditional idiom have explored the chansons de geste both with an assumption of the primacy of plot, and from the point of view of character; their choice reflects their generation. Among the more senior, this judgement on the Roland by Edmond Faral has been much (if incorrectly) cited: “Les idées […] constituent le principe vital de [la Chanson], […] mais ce que ces idées ont,à l’application, de force poétique et de beauté morale, c’est dans l’âme des personnages qu’il faut le chercher.”² This passage is quoted with approval by Le Gentil in 1955 and 1967,³ and by Misrahi and Hendrickson in 1980.⁴ Le Gentil continues, “Rien de plus exact: dans notre poème, les caractères commandent l’action et les problèmes posés le sont dans et par des ‘études d’âmes’” (p. 123). Tony Hunt is representative of the younger school when he asserts, of the same poem, “[the poet] is not interested in motivation but in repercussion […] [His] technique is entirely the result of his interest in tragedy as the product of action, not character”; and, later in the same article, “Too much emphasis has been placed on the Roland as a psychological drama.”⁵

There are good reasons why we should not simply turn our backs on this issue.⁶ The gross categories of “actor” and “plot” are those to which the reader reacts first, and beyond which the naïve reader rarely progresses. Through them interest, sympathy, and excitement are elicited in the reader, and ethical and especially political themes conveyed.⁷ They are universals of narrative; to quote the words of Barthes, “the characters […] constitute a necessary plane of the description, outside of which the commonplace‘actions’ that are reported cease to be intelligible, so that it may safely be assumed that there is not a single narrative in the world without ‘characters’ or at least without ‘agents’.”⁸ They are not far from being universals of criticism, too, to judge by the frequency with which titles on “The hero” continue to appear in bibliographies.⁹ Finally, the relationship between dramatis persona and plot in a fiction will relate to (though not necessarily correspond with) contemporary prejudice about the relation between character and action in the real world. This should certainly qualify it as a legitimate concern of literary history; the more so in a genre such as the chanson de geste which belongs to a para-historical or even historical mode of discourse.¹⁰

Important though this area of investigation seems to me, it is clear that most literary theorists have their sights trained towards other horizons. Receptional aesthetics with its revalorisation of literary history and concern to explicate reader reaction was an obvious place to look; and indeed an article on “Levels of identification of hero and audience” by Hans Robert Jauss¹¹ testifies to the recognition accorded to the notion of hero in this critical school. His aim in this article is to revise Northrop Frye’s famous modes (mythic, marvellous, high and low mimetic, and ironic)¹² from a receptional standpoint, suggesting modes of reader identification with character ranging from the admiring to the ironic. Comparing the two studies, however, it seemed to me that Jauss was even less interested than Frye to examine how the hero is constituted (through “characterization,” authorial intervention, emplotment, or whatever), and while he suggests ways in which the reader can react to character, he says nothing about how he or she perceives it in the first place.

Hard-core structuralist theory also proved a disappointing hunting ground. In La Sémantique structurale, Greimas formulates admirably the question which preoccupies me when he acknowledges the possibility of opting between a character-based and a plot-based description of a given corpus of mythological material.¹³ The researcher could either (1) identify a series of plots in which the god is active, and so attribute a set of functions to him (“description fonctionnelle”), or else (2) identify a range of descriptions of him (“description qualificative”). The results of these approaches would be, he says, (“dans certaines conditions”) complementary and “comme convertibles de l’un à l’autre modèle: le dieu pouvait agir conformément à sa propre morale; ses comportements itératifs, jugés typiques, pouvaient lui être intégrés comme autant de qualités” (pp. 172-3). Sadly the discussion goes no further. Greimas offers no argument for his decision in favour of a functional analysis (one presumes he is merely bowing to the force of the formalist tradition transmitted by Propp),¹⁴ nor does he investigate under what conditions the two descriptive models might cease to be equivalent.

Greimas’ approach, then, assumes the primacy of plot and the subordination to it of “character.” “Actantial categories” (the term is Greimas’), which are the primary unit of analysis, do not correspond directly with individual surface structure characters but with the roles that underly them. They are held to be reducible to six: Subject (the agent), Object, Adjuvant (the role of any person or element assisting), Obstruant (any element that hinders the activity of the agent), Destinataire (the indirect object of beneficiary) and Destinateur (the role under whose auspices, or with regard to which, the Subject is activated). These actantial categories are identified by reducing successive stages of the plot to minimal narrative units of “deep structure” (that is, to sentences summarizing the surface text), and by reading off the grammatical functions within those sentences as the corresponding category, i.e. grammatical subject = Subject, grammatical Object = object, etc. But as Greimas’ conception of deep structure—like that of another discipline one can think of—has got deeper with the years, it is not always easy to determine how the actantial categories are realised by the actual characters in a particular text.¹⁵ Application of his theory to the chansons de geste has produced some startling results. According to one self-confessedly experimental reading of the Roland, for example, it appears that God, who is not a “character” at all, is Destinateur and Destinataire, and at times also Adjuvant, namely three out of the total of six categories, whereas Roland is Subject under the modality of will (vouloir) only (though with some savoir)¹⁶, and as such inferior to Charlemagne who combines vouloir, pouvoir, faire, savoir, and devoir. Such an analysis, one might pardonably observe, if it lacks an excessive regard for character, is not over-preoccupied with plot either.¹⁷

More nouvelle vague criticism (one is tempted to call it nouveau vague), with its diligent glossing of arbitrarily chosen fragments of surface structure text (referred to, since Barthes, as “lexies” or “reading units”), has no interest in crude mimetic categories such as actor and plot which are simply so much noise on its semiotic codes. In an ingenious analysis of a fragment of Chrétien’s Perceval by Jean-Michel Adam,¹⁸ for instance, the action—Perceval’s meeting with the first knights he has ever seen—and his character of naïve persistence, simply disappear down the cracks between the semiotic building-bricks that realise the “isotopies” of nature and religion versus chivalry.

This paper—mercifully—does not set out to offer a broad theory of how character and plot should best be defined and integrated into literary criticism. More modestly I investigate some of the ways in which they interact in a body of literature whose exciting narrative stance has always attracted both theoretical and practical criticism—the chansons de geste of the twelfth and thirteenth centuries—with particular reference to the text among them best known to me, that of Raoul de Cambrai.

.....

It is manifest that one way an author can present his actors as primary, the action as flowing from them, is by serving up slabs of circumstancial and atemporal description of a character, and then offering morsels of incident as being merely illustrative of his or her nature. The example of Balzac leaps to mind. It was to defer consideration of this style of authorial “characterization” that I began with a reference to Othello rather than a narrative hero, since in this respect the chansons de geste have always been compared with the theatre. This commonplace of epic criticism is clearly formulated by Misrahi and Hendrickson: “We learn to know and judge epic figures as we do those of medieval theatre, by what we see them do” (p. 360).

The element of portraiture, which is in any case slight, is so traditional in content that its aim can only be to categorize, not individualize. I have pointed out elsewhere that the descriptions of hero and villain in the Roland are substantially alike.¹⁹ Attempting a character sketch of Raoul in Raoul de Cambrai based on observations ex persona poetae and by other characters, I found that nothing was said about Raoul that was not also said of at least one other character. If he is handsome, so are Bernier and Gautier.²⁰ Did he inherit his clear face from his mother, and his fierceness from Guerri?²¹ But then fierceness is found among the Vermandois too.²² Other people can be wise on occasion, and then carried away on a tide of uncontrollable emotion.²³ Raoul may be a marvellous knight, but so are Bernier, Rocoul, and others;²⁴ as for the qualities of nobility, valour, rank, and breeding, they are two a penny.²⁵ He is not alone in his bad qualities either. His tendency to perfidy and immoderateness is shared by the king and even by virtuous Bernier.²⁶ The poet, in short, makes no attempt to particularize his central character through description. Indeed, a recent series of articles by Moroldo has shown that the descriptive elements which make up the portraits in the Chanson de Roland are taken from a common stock which does not differ materially through six other chansons de geste of varied tone and date.²⁷

This is not to say that epic characters necessarily lack “psychological depth” (though some do), but that perceiving it involves a moral and imaginative effort on the part of the reader/hearer (with perhaps an actor’s contribution from the jongleur).²⁸ While literary characters should not be confused with real people, the fact remains that epic heroes are knowable in a way similar to that in which we know other people in real life; whereas the heroes of a novel whose author is constantly imparting insights about them are known to us more as we know ourselves.²⁹ This gives epic characters a directness and an otherness which challenge the reader / hearer’s imagination along lines which are suggested and guided chiefly by the plot. The descriptive element in the chansons de geste is therefore not such as would lead to a distinction between characterization and emplotment; rather one would see them as indissolubly bound up together.

Another feature of the composition of these poems that militates against a dissociation of character from action is repetition. The emplotment of chansons de geste is intensive rather than extensive; character and plot are at once reduced in scope, though often with a corresponding gain in intensity: the hero is of such a type that he unrelenting repeats the same action; the plot is so organized that certain aspects of the hero are insistently brought into play.

In the Loheren cycle, the families of Lorraine and Bordeaux do battle unstintingly through thousands of lines of siege and countersiege, sortie, ambush, and the occasional pitched battle (this clearly perceived as more glamorous). Not surprisingly, then, the character of the Loheren heroes emerges as grim and embattled, as austerely resolute as their unvarying text.³⁰ The Couronnement de Louis reveals a comparable sobriety of invention. William starts by protecting young King Louis at home; then he frees Rome from agressors; then he protects Louis at home; then he frees Rome from aggressors (incidentally protecting Louis); and finally he protects Louis at home. In the simplest imaginable way the plot analyses the character of its hero into two principal facets which it then proceeds to dramatize, viz. feudal competence at home and abroad.³¹

The Chanson de Guillaume repeats a single though more extended cycle: feasting, resting, setting out to fight, fighting, returning, feasting, resting, setting out to fight, etc.³² Only one character, William himself, survives this punishing routine repeatedly; his true grit marks him as the hero. Other characters achieve only part of it. Thiebaut and Esturmi feast and rest to perfection, but when it comes to setting out to fight they turn tail and run away: clearly they have no grit at all. Vivien sets out to fight and fights, but does not return: he is a figure of doomed youth, inspiring but unsuccessful. Girart possesses more of the strength of his uncle, for he sets out to fight, fights, returns, feasts, rests, sets out to fight, and fights, but then fails to return. His score is midway between Vivien and William. Rainouart, finally, completes the cycle successfully but only once, and so lacks the solidity of William’s achievements which, in any case, he burlesques. The Frankish warriors are thus all carved of one material. William is the complete man whose nephews are partial copies of himself, and in comparison with whom the cowards are shown to be totally inadequate.

In Raoul de Cambrai there are four types of scene in which the hero appears: court scenes (with the king), family scenes (in which I include his meeting with Bernier’s mother), council scenes (with his vassals, including Bernier), and fighting scenes. We should not, however, be misled by this apparent diversity. Nearly all the characters whom Raoul meets are subsumed to the role of enemies, and he behaves throughout in a spirit of violent antagonism, threatening the king, insulting his uncle and his own and Bernier’s mother, and hitting his vassal with a stick.

This belligerence is invested with ambiguity. As heroic bravery, it is endorsed in Raoul’s encounter with the forbidding Vermandois knight, John of Ponthieu, who is almost a giant. Raoul is on the verge of panic when he remembers his father’s courage before him and gains the victory through sheer force of character and pride of lineage (ll. 2744-8). In a macabre dissection scene following the death of both knights (l. 3239 ff.), Raoul’s heart is found to be as big as that of an ox, while John’s is the size of a child’s. These scenes present Raoul’s warlike temperament in bono. For the most part, however, his violence is perceived as excessive and misdirected, an effect largely achieved through the patterning of scenes. Starting at ca. l. 900 the order of events is as follows: (1) Raoul decides to go to war to claim the Vermandois, (2) quarrels with his mother, (3) sets off on the campaign threatening destruction, (4) quarrels with Bernier’s mother, (5) destroys Origny, (6) quarrels with Bernier, and finally (7) fights the Vermandois army. This interleaving of scenes connected with actual fighting (1, 3, 5, 7) and scenes of domestic or feudal conflict (2, 4, 6) shows Raoul’s aggressivity spilling out into relationships which should be peaceable and supportative.

For much of the text, Bernier is presented as the exact opposite of Raoul. He also appears in the same types of scene but manifests an aversion to violence. He avoids quarrelling with Raoul’s mother and his own; he makes only a peaceful protest to the king; he is reluctant to fight; and he endures extraordinary ill-treatment from his overlord before obtaining his release—which he does by proposing himself as victim to Raoul’s assault.³³ For both characters, emplotment and characterization are handled with radical simplicity and are in practice impossible to dissociate.

.....

The concatenation of similar scenes is the most obvious compositional principle in the chansons de geste, most of which are structurally open-ended: one could devise additional scenes at little cost to the imagination, or else abridge them with no sense of irreparable loss. The episode is their basic unit of composition, just as the laisse is of the episode. It is probable that the episode was also the commonest unit of performance: one can imagine an audience clamouring for “the bit where…” and certainly the jongleurs’ catalogue in the romance of Flamenca lists many items which are in fact episodes from larger works.³⁴

This fragmentation in practical performance no doubt partly accounts for the tendency to invest individual scenes with drama at the expense of overall coherence, which is well attested in the chansons de geste. Consistency in the presentation of individual characters may well be sacrificed to heighten the clash between them. When Ganelon is brought before Marsilie in the Roland, for instance, he infuriates the Saracen king by demands purported to come from Charlemagne but which formed no part of the Frankish deliberations as reported in the preceding scene, and therefore seem to be Ganelon’s own invention. Only a few lines before, Ganelon had seemed fearful even to undertake the embassy, but now he is emboldened to the point of greatly increasing the dangers inherent in it. The payoff of an exciting scene, with Ganelon and Marsilie almost coming to blows, quite eclipses the incoherence of characterization that produced it;³⁵ as T. Hunt observes, the Roland poet is more interested in interaction than in identity.³⁶

In her monumental study of Huon de Bordeaux, M. Rossi traces a similar lack of consistency in the central character. Initially Huon demonstrates a lucidity and good sense unusual in an epic hero. Once he is under the magical protection of Auberon, however, the plot risks stagnating: excitement and danger can only be revived when Auberon’s friendship is withdrawn, and this has to be engineered by Huon’s provoking him with hitherto uncharacteristic folie.³⁷ By adopting a laxer standard of consistency, the poet achieves the double drama of conflict with Auberon and physical risk to his hero. Concern with action as primary has made characterization take second place.

A third example of a text where the personality of a major character is bent to meet the author’s conception of the story-line is the Charroi de Nîmes.³⁸ The essentially reduplicative structure whereby in the first half William argues with King Louis and in the second has a battle of wits with the Saracen kings Otran and Harpin sustains the hero in a dramatic conflict of personalities without, however, masking the fact that his own personality undergoes a considerable shift in the process. In the first scene he is all righteous anger as he defends the ideal of an enlightened feudalism that would reward the meritorious while protecting the weak; in the second he emerges as a wily play-actor, a knight disguised as a merchant pretending to be a thief, sardonic manipulator rather than innocent victim, his high-mindedness replaced by bravado and deceit.³⁹

In the laisses similaires of the better epics, an act is presented from different standpoints which, taken together, are less contradictory than complementary. In the same way, differences in the presentation of a character from one episode to another may be reducible in a creative reading to a coherent presentation of a moral or political theme. In such a case, character is subordinated to a plot conceived not just as a series of acts but as the articulation of an ethical point of view. In the Charroi de Nîmes, for instance, the inconsistencies of William’s behaviour may be partly neutralised by subsuming them to an apologia for the petite chevalerie.

In Raoul de Cambrai, the pacific Bernier occasionally rises to Raoul’s implacable belligerence, while Raoul declines into moderation and goodwill. After their rupture and Bernier’s return to his family with news of the invasion, his prudent father and uncles adopt a policy of appeasement. An ambassador is sent to Raoul’s camp who behaves with exemplary decorum (not in the least like an Irishman, l. 2154) and pledges the Vermandois to assist in the reconquest of the Cambrésis from Giboïn. Even Guerri is struck by the generosity of their offer (ll. 2170-73), but Raoul is unbending and accuses him of cowardice. In a second embassy scene, however, the roles are reversed. Bernier is insulting and provocative, calling down a curse on Raoul and enumerating his grievances against him. At first incensed, Raoul is moved by Bernier’s message of peace. From fix a putain (l. 2252) he switches to addressing him as ami (l. 2287), promises to accept the terms and looks forward to renewed friendship (ll. 2286-90, 2296-7). It is because of Guerri’s obstinacy and wounded pride that the offer is rejected. Learning this, Bernier is filled with vindictive delight: he thanks God for the war, defies Raoul and at once rushes to attack him, thereby killing an unarmed man who had run to Raoul’s defence.

A second diptych, composed of duels, demonstrates a similar changeover of roles between Bernier and Raoul’s successor and nephew, Gautier, who is a watered-down version of his uncle. First Gautier, with uncharacteristic moderation, challenges Bernier to single combat in order to avoid further bloodshed. He prepares for the fight by scrupulous attendance at religious services (ll. 4290-5); courteous in word and deed, he ministers at Aliaume’s death bed and then actually saves Bernier’s life (l. 4715 ff., 4745 ff.). Bernier’s pride or folly in insisting that their two seconds, Guerri and Aliaume, should fight as well, is condemned by both of them (ll. 4710, 4731-5). In the events surrounding the second duel, however, Gautier with ruthless intransigeance rejects repeated appeals for peace from the king (l. 5139), Bernier (ll. 5181 ff. and 5267 ff.), the abbé of St. Germain (l. 5295 ff.) and the whole Vermandois family (l. 5346 ff.). His piety shows a marked decline, for he is reluctant to yield to the abbé’s offer of absolution, and does so only with a bad grace: “con je le fas dolent” (l. 5360). Bernier, by contrast, has been conciliatory from the start, excusing Gautier to the king on the grounds of his youth (l. 5163), offering him his sword (l. 5256 f.), and prostrating himself on the ground (l. 5365, etc.).

In such scenes the author’s conception of character is manifestly subordinate to his desire to expose the psychology of vendetta and reprobate the incessant renewal of strife. This principle of applauding the doves and blackening the hawks—unless, of course, they are engaged in a worthwhile warlike crusade—is constant in the so-called epic of revolt, and applies without regard to the integrity of individual characters. Subrenat comments on the dénouement of Gaýdon that the hero

“maintient ses exigences certes et, en cela, ne change pas, mais il se rend compte aussi, de lui-même cette fois-ci, que la solution du conflit n’est pas dans les armes. […]

“Et curieusement, à ce moment-là, c’est Riol le sage qui devient plus violent. Ce renversement d’attitude entre les deux personnages souligne le caractère plus conciliant de Gaydon.”

⁴⁰

In Girart de Roussillon the judgments passed on the hero by both author and characters reflect his swings of mood between conciliatoriness and bitter antagonism towards Charles Martel, and it is no coincidence that the extraordinarily long and laudatory description of Foulques occurs shortly after he has condemned Girart’s implacability (l. 4832 ff.) and immediately before this line: “E sapçaz d’esta gera mout li desplaz” (l. 5010).⁴¹

Bertrand de Bar-sur-Aube retained something of this spirit in his often fatuous recasting of the same legendary material in the chanson de geste Girart de Vienne. Burlesque though the grounds of Girart’s resentment against his sovereign are (he has been tricked into kissing the queen’s toe instead of Charlemagne’s, and the emperor has refused to punish his wife for this shameful outrage!), he persists in belligerent opposition until he successfully takes Charlemagne’s godson, Lambert, prisoner. At once he changes character: hoping to win back Charles’s friendship, he offers to waive a ransom demand; and now it is Charlemagne’s turn to appear in a bad light. When Oliver, Girart’s nephew, comes to his court to sue for peace, the emperor is intent on revenge (ll. 4020-1). He wants to humiliate his opponent, even though legally he himself is in the wrong (ll. 4046-8). Even when accused of treachery by Roland, there is no longer any question of rancour on Girart’s part: his moral stock has risen as ineluctably as Charlemagne’s has declined.⁴²

One means whereby an author can spell out the moral thrust of his tale so that even the dullest wit shall perceive it is by writing in a divine intervention. Since this device deprives the human actors at least temporarily of that agency which is the principal form of expression of character, it again tends to diminish the status of the actors vis-à-vis the plot. This is the more true when, as sometimes happens, God is transparently an alias for the author’s undistinguished moral prejudices, and the characters dangle puppet-like from unsubtle strings.

An example of such a scene occurs in Girart de Vienne. Roland, Charles’s champion, and Oliver, Girart’s champion, are to fight a duel to determine whether or not Girart has been guilty of irregularity in his feudal obligations. This episode, the most protracted of the text and apparently perceived by the author as its high point, lasts for some thousand lines (from the arming of the heroes, l. 4880 ff., to a conclusion ca. l. 5957), during which the poet spares us nothing of how they vie with each other in prowess and gentility. So many blows, or such nice manners, should, one feels, bring a conclusion to the point at issue: but not a bit of it. God literally brings the curtains down on their efforts by enveloping both combatants in a dense cloud (l. 5891 ff.) while an angel announces that they would be better employed fighting side by side against the Saracens in Spain (ll. 5908-5920).⁴³ The heroes at once fall in with this way of thinking. The duel is abandoned, and Charlemagne and Girart reach a reconciliation by quite another means. While this divine intervention does facilitate a realisation of the many promptings of mutual affection and esteem between the two young knights, it also deprives their great show of courage and endurance, their eager championship of their respective uncles, and their bitter political arguments of much of their point. The characters, in other words, are sacrificed to the author’s concern to express his somewhat banal view of the proper purpose of chivalry (and also, incidentally, to bring them into line for the Chanson de Roland, to which his text is a retrospective introduction).

The two redactions of the Moniage Guillaume exemplify different approaches to the use of angelic intervention. The shorter first redaction, which proceeds at a fabliau-like pace, dispenses with a “psychological” consideration of William’s sanctity by assigning an angel to summon him first to the monastery of Genevois-sur-mer (ll. 58-60), then to his hermitage at St. Guillaume le Désert (ll. 820-37), leaving William to a relatively uninhibited exercise of his former, worldly characteristics of pugilism and voracity, with splendid comic effect. The second redaction takes William’s holiness much more earnestly. The angel vanishes, to be replaced by much explicit comparison between monasticism and knighthood, and rather too much pious meditation. The result is a far more serious view of character which is not without interest, despite its dampening effect on the cartoon-strip style of comedy of redaction 1.

Even when God is not a direct participant, the ambiance of the Old French epic tends to the religious and moral. In so far as they are “structured” at all, the stories display patterns of reward or punishment measured according to the preoccupations of the author, rather than of success or failure in terms of the desires or intentions of the characters. Heroes and traitors are alike exemplary, and a chief requirement of “realism” is to confer sufficient density on a character for the audience to engage with the type (not an individual) and thereby concur with the author’s judgment on it. How characters act will show what they are made of, and elicit audience response; the ultimate test, however, is not what they do, but what happens to them. Thus E. B. Vitz writes of the Roland that “religious causation (God’s will, revealed or unrevealed) undermines human agency as a narrative force in plot structure,” and “men are responsible for their actions, but their actions are not responsible for what occurs.”⁴⁴ She uses her observations of this and other texts to argue against the validity of a Greimasian notion of “Subject” for medieval literature; certainly I think they confirm that a sense of divine purpose is one way in which characters can be said to be subordinated to plot.

The classic reward for good conduct is a fief and a wife, always beautiful and preferably rich. W. C. Calin has argued that in Aymeri de Narbonne, the themes of the conquest of a city and the winning of a bride, which together make up most of the action, are interrelated: with land comes power and position, and with marriage a lineage to maintain it.⁴⁵ Aymeri’s career in this chanson de geste is merely an amplification of that of nearly all his relatives, for his grandfather won the fief of Monglane and the lovely Mabille (Garin de Monglane), his son William wins Orange and “la belle Orable” (Prise d’Orange), and his grandson becomes king of Andrenas and marries Gaiete (Guibert d’Andrenas). Those whose careers involve more danger or excess are rewarded with martyrdom (Renaut de Montauban, Roland (?)), or sainthood (Garin Le Loheren, Girart de Roussillon), or at least with glory. Whoever wrote the second death scene of Vivien in the Chanson de Guillaume clearly felt it was less incongruous for the hero to appear to die twice over in very different circumstances than for his efforts to go unacknowledged, and hence rewarded him (l. 1988 ff.) with a touching end, in an odour of sancity, attended with the last rites by his dear uncle, and in a Christian-symbolic setting of fountain and olive tree, even though he had been cut to pieces in a desert landscape some ten to fourteen days before.⁴⁶ Conversely punishment usually takes the form of death without honour. Characters like Raoul in Raoul de Cambrai and Isembard in Gormont et Isembard are “des maudits qui ne sont pas tout à fait des damnés, puisqu’ils se repentent et que Dieu leur laisse une chance de le fléchir, mais des maudits qui n’ont pas droit à un autre sort que la calamité, de la mort subite.”⁴⁷ Outright traitors are condemned in court (Ganelon in the Chanson de Roland), summarily executed (Acelin in the Couronnement de Louis, l. 1884 ff., or Makaire in Aiol, l. 10900 ff.), or insouciantly killed with a blow from William’s tremendous fist (as Aymon le viel in the Charroi de Nîmes, l. 678 ff.),⁴⁸ all lending their weight to the conviction expressed in Fierabras that “Tous jours vont traïtours à male destinée; U en pres un en loing, jà n’i aront durée.”⁴⁹

Happily the hand of Providence is sometimes more discreetly gloved. In the events leading up to the hero’s death in Raoul de Cambrai there is a conviction of fatality combined with confusion as to its operation. Cursed by his mother, sacrilegious and blasphemous, Raoul seems an obvious target for divine retribution, and yet his actual death is presented in almost wholly secular terms.⁵⁰ Is Bernier an agent of the Almighty? The poet spares us this conviction; but then he excels in conveying a sense of the tangibility of moral questions together with the elusiveness of moral answers. A scene which shows a similar, masterly reluctance to dilute the complexity of human problems with an infusion of divine justice is the death scene of Garin in Garin le Loheren (ll. 16511-93). It is a martyrdom of a sort: Garin is murdered by the treacherous Bordelais as he kneels at an altar in penitence for the sins of a violent life. Yet before he is dead, a follower with commercial flair and a quick eye to a future saint piously hacks his right arm off for use as a relic: which assault Garin pardons before he dies. Sanctity or burlesque? An air of ambiguity pervades the episode, proof against the easy answer or pat judgment. Two final, and more obvious examples of a sense of the numinous adding to the significance of a human drama without detracting from the alterity and integrity of the characters are the deaths of Roland and Vivien.

.....

I will conclude this paper by considering two features of the composition of the chansons de geste which seem to me positively to privilege character vis-à-vis the individual events of the plot, and to confer on certain heroes an almost mythic status.

In these poems the distinction between “subject” and “fable” drawn by the Russian formalists and taken over by more recent critics with the terms “récit” and “histoire”⁵¹ is more than a theoretical construct, it is a vital part of the mechanics of composition and reception. Whereas in modern fiction there tends to be a systematic blurring of the two, the Old French epic poets are quite clear that their narrative (or “récit”) is only one possible rendering of a tale (“histoire”) from which it is conceptually distinct. Without wanting to go into the subtleties of this distinction here, the devices by which it is most straightforwardly indicated include allusions to sources or other, rival versions; summaries of the histoire in prologues and elsewhere; recapitulations and prospections or prophecies which temporarily suspend the flow of the récit to reveal at least a partial overview of the histoire; and tense usage which opposes an unfolding, present tense récit against a complete and known past tense histoire.

The “normal” situation is for author and audience to share in this knowledge of the existence of the histoire while the characters slog along the line of the récit, experiencing and acting as though it were the ultimate reality. At times, however, the characters are received into the complicity of author and audience; they gain insight into the totality of the story in which they figure; and though they continue to enact the roles apportioned to them and to feel the emotions that are their lot, they do so with a curiously enhanced stature which can border on myth.

The commonest device whereby a character learns more about the plot of his text is the prophetic dream. From the structured diversity of dream-lore inherited from Classical Antiquity the authors of chansons de geste fixed on one form only: the dream which comes true.⁵² It usually forewarns of trouble, conflict, death, or some other disaster. The dreamer is temporarily removed from the flow of the récit to perceive some part of the underlying story.⁵³ In Raoul de Cambrai, for instance, Aalais learns of Raoul’s death (ll. 3511-20), Beatrice of her husband’s (ll. 8467-76). In the Chevalerie Ogier, the hero dreams that Charlemagne’s army will pursue him (ll. 8213-21). On waking he learns that he has just been betrayed in Castel Fort and that the enemy are upon him. This stock device is amusingly parodied in the Roman de Renart, Branche II, ll. 131-60.

Although belonging to a recognizably rhetorical convention, as the Renart version shows, these dreams do not utterly conflict with mimetic realism, and do, indeed, reflect contemporary folk belief. More uncanny is when the characters intuit the plot without the need for such mediation. In the Chanson de Roland, for instance, Charlemagne receives veiled premonitions of future strife from his angel, but Roland himself appears to have a hotline to foreknowledge.⁵⁴ In the opening council scene he is the only one of Charlemagne’s barons to discern Marsilie’s treachery behind his lavish offers of peace (l. 196 ff.). He seems to divine Ganelon’s guilt at the time of his appointment to the rearguard, when he breaks out in violent recrimination of his stepfather and alludes to the earlier scene of Ganelon’s being chosen for the embassy—that scene being, of course, the reason for Ganelon’s betrayal (ll. 763-65). He expresses no surprise at Oliver’s announcement that the Saracens will attack, and quashes his denunciation of Ganelon (ll. 1008, 1026-7), only admitting his guilt later (l. 1457). This admission is accompanied by a prediction of Charlemagne’s vengeance (l. 1459). His forecasts of the battle are proved true, though not exactly as he anticipated: fierce blows are struck, great honour won, the Saracens are vanquished or slain, Charles hears the horn and returns. Only as the battle draws towards its close does Roland reckon the full cost of these achievements to himself and his men in the so-called lament:

Barons franceis, pur mei vos vei murir,
Jo ne vos pois tenser ne guarantir.
Aït vos Deus ki unkes ne mentit!
Oliver frere, vos ne dei jo faillir,
De doel murra[i], se altre ne m’i ocit.

[1863-68]

Embracing martyrdom, Roland anticipates Charles’ arrival on the field to honour their remains (ll. 1922-31), and also his grief at the irreparable loss to his imperial army (ll. 1985-7). The death scene with the mediating angel sets the seal on Roland’s participation in a mythic or quasi-divine level of awareness. As a character, he enjoys an exalted status mysteriously situated above the blow-by-blow unfolding of the action as expounded by the récit, a status aligned within the text with God and the angels, and outside it with the foreknowledge granted to author and audience.

Vivien in the Chanson de Guillaume presents an interesting mixture of true and false prediction of his future. On the eve of the battle his sober realism is contrasted with the Dutch courage of the inebriates Thibaut and Esturmi: they imagine they can win without reinforcements, but Vivien advises sending for William because “Od poi conpaignie ne veintrun pas Arabiz.” (l. 71) The next morning he amplifies this advice: Thibaut should assess the enemy’s strength, and then either engage battle or conceal his troops until William can be sent for: with God’s help, the victory will be theirs (ll. 171-82). In the course of reconnoitring, Thibaut rashly exposes himself to enemy view and Vivien, feeling the Franks will be dishonoured if they do not at once attack, embarks on a policy of desperation. Though he assures Thibaut they will win, he does so quoting the drunken arrogance of Esturmi (cf. ll. 68 ff. and 207 ff.). A more realistic appraisal of the prospective encounter in ll. 245-51 contains predictions of the deaths of cowards, lesser men, and youngsters, and recognition that only God’s power can overcome the heathen. As the battle takes its tragic toll, Vivien is divided between desperate hope that William or Louis will come to bring them victory and realisation that his followers are going to receive a martyr’s death (ll. 485-6, 561-2, 573; 502, 538, 545-7). For a moment of terrible misapprehension, he takes Girart’s cry of Monjoie for the arrival of William or Louis (ll. 453-54); this lack of true perception culminates in two despairing and false predictions, that he himself alone will win the fight, even though all his companions flee (l. 589 “Jo les veinterai ben solunc la merci Deu”), and that William and Louis will arrive in time, on that very day (ll. 751-2). As his last knights are cut down around him, however, Vivien returns to his former lucidity which now, with tragic prescience, encompasses his defeat and death:

Forz sui jo mult, e hardi sui assez,
De vasselage puis ben estre sun per; [ie. William’s]
Mais de plus loinz ad sun pris aquité,
Car s’il fust en l’Archamp sur mer,
Vencu eust la bataille champel.
Allas, peccable, n’en puis home gent
Lunsdi al vespre.

[830-6]

Through this mixture of erroneous belief and terrible lucidity, the author of the Chanson de Guillaume gives his hero a stature comparable with Roland’s while stressing his greater humanity.

The device of authorial prophecy is used with particular insistence in Raoul de Cambrai, making it the more remarkable when certain of the characters suddenly join with the author in announcing what will happen next.⁵⁵ From the outset, the audience has been alerted to grief that will follow from Louis’s gift of a fief to Raoul (cf. vv. 516-17, 536-37, 639-40, 697-98, 745-46, 778-780, etc.). Quarrelling with her son about his intended campaign, Aalais too predicts that the longed for investiture will in fact bring death to its recipient:

Qui te donna Peronne et Origni,
Et S. Quentin, Neele et Falevis,
[Et] Ham et Roie et la tor de Clari,
De mort novele, biax fix, te ravesti.

[987-90; cf. -1004-06]

Asking how he expects to win such a war, she taunts him with an inquiry after Bernier: what will become of him (l. 1077)? Raoul’s reply is perhaps unwittingly prophetic: Bernier will help his family if the need arises (l. 1085; contrast the rather different declaration of intent by Bernier in ll. 944-5).⁵⁶ In a flash of clairvoyance, Aalais puts her finger on the nub of the whole plot:

Bien le savoie, a celer nel vos qier,
Ce est li hom dont avras destorbier,
C’il en a aise, de la teste trenchier.

[1089-90]

Again this prophecy echoes, but how much more strikingly, predictions already made by the author (see ll. 6-11, 391, and 741). Raoul therefore proceeds to the war against the Vermandois as the plot and his own logic require, but he does so both as a character enmeshed in the récit, and as a participant in a dialogue regarding the histoire which exactly parallels a similar dialogue engaged between author and audience. This is a highly effective device for privileging the heroic characters over the action.

My last point concerns discrepancies between histoire and récit which operate to undermine the credibility of the presentation of events and focus instead on the intensity of character reaction as the only reliable reality. In his highly influential study of L’art épique des jongleurs,⁵⁷ Jean Rychner established a terminology for classifying the various ways in which laisses could be grouped together. Laisses parallèles were those in which similar events were presented in similar terms, so that what might be conceived as successive parcels of narrative are in fact presented as “juxtaposed”: “à moments similaires successifs correspond une expression similaire répétée” (p. 88). It is relatively easy to find examples of such laisses; a typical instance would be the series in Raoul de Cambrai where the supporters of the Vermandois arrive one after the other and pledge themselves to kill Raoul (laisses 97 [end], 98, 99 + 100, 101, 102, 103) (and cf. Rychner, pp. 83-93). The term laisses similaires is initially used to apply to certain groups of laisses in the Roland where this lyric “juxtaposition” is particularly marked: laisses 40-42 (Ganelon treats with Marsilie), 83-85 (the first horn scene), 133-8 (Roland sounds the horn), etc. Being principally concerned with formal elaboration (though there is no clear formal distinction drawn between laisses parallèles and laisses similaires) Rychner prefers to avoid determining the narrative reference of such sequences. Commenting on laisses 40-42, for instance, in which Marsilie asks Ganelon when Charlemagne will desist from war and is told, never so long as Roland is alive, he writes:

Faut-il dire, d’ailleurs, la ou les questions de Marsile? Sous cet ensemble statique, les réponses de Ganelon marquent à chaque fois quelque progrès; ce n’est pas, à vrai dire, un progrès narratif, mais un progrès psychologique et dramatique, dévoilant peu à peu ses pensées secrètes, suggérant, insinuant la trahison. […] Il ne faut donc pas chercherà savoir, sur le plan de l’événement, si Marsile a réellement posé trois fois la même question, car les laisses similaires retiennent le récit dans une halte bien plus lyrique que narrative. (pp. 94-5)

In most of the examples which Rychner goes on to discuss, however, it is clear that the distinguishing feature of laisses similaires is the fact that they offer multiple reports of the same event. In laisses 174-6 of the Roland, for instance, it would not be reasonable to suppose that Roland delivered his glove to a descending angel any more than once; likewise in laisses 206-7, we cannot imagine that Charlemagne found Roland’s body twice. The examples adduced from other texts (pp. 100-107) confirm that whatever his formal definition, in practice for Rychner laisses parallèles are those where different events are described in similar terms and laisses similaires are those where the same event is presented more than once in slightly different ways.

This leaves a very considerable number of instances outstanding of which, as in the case of the Marsilie—Ganelon scene, it is impossible to assert whether it is the event or merely the description that is repeated. The narrative becomes unreliable on a quite simple and fundamental issue of plot but supplies instead a very insistent image of a character in the grip of an intense experience or emotion. Since the laisse patterns of the Roland have attracted so much critical attention, I will illustrate this from Raoul.

Nearly all the “lyric” patterns of laisses in this text occur in the purportedly older core (up to ca. l. 3740), and focus on the character of the hero. The scene confronting Raoul and his mother, laisses 48-54, contains a repeated pattern of maternal admonishment followed by a display of anger from Raoul. In laisses 48 and 49 Aalais’s warning that the gift of the Vermandois means death is reiterated; in laisses 50 and 51 Raoul issues dire threats against the men of Arrouaise if they refuse their support; and in laisses 52 and 53 Aalais expatiates on his vassals’ faults. Are they both so furious that they keep repeating what they have just said? or is their anger and sense of mutual injury being dwelt on repeatedly by the poet as a technique of amplification (expolitio)?

Laisses 77 and 78 create a similar ambiguity around the growing conflict between Raoul and Bernier. In laisse 77 Raoul threatens to drive the Vermandois brothers across the sea and seize every penny they own; Bernier replies that this would be outrageous treatment for such worthy knights. In laisse 78 Raoul’s threats are repeated in more emphatic terms; the penny’s worth of land, for instance, becomes:

… le montant d’un denier
De toute honnor ne de terre a baillier,
Ou vif remaigne[n]t, ou mort puise[nt] couchier.

[1633-5]

and “Tant que il soient outre la mer fui” (l. 1621) becomes “Outre la mer les en ferai naigier” (l. 1636). Bernier’s reply is likewise amplified. To the defense of Herbert’s sons as fine knights is added criticism of Raoul (l. 1639), warning that even if his campaign is successful he will not find it easy to hang on to the Vermandois (l. 1643), and the wish to be revenged on Raoul for his many crimes against him (ll. 1645-51). The descriptions contained in these two laisses are different but not incompatible, and so it is impossible to assert whether they record two exchanges, or one. What is, however, unmistakable is the escalation in both provocation and remonstrance, and the consequently sharpened focus on the conflict between the two knights. One cannot be sure exactly what they said on any precise occasion, but one forms a very strong impression of their capacity for emotional antagonism: an impression, that is, of character at the expense of a precisely informed grasp of plot.

The ensuing quarrel between Raoul and Bernier contains further patterns of repetition which create uncertainty about the narrative flow precisely as they emphasize the emotions of the participants.⁵⁸ The final phase of this episode is Raoul’s offer of reparation for the blow which he has struck Bernier. In the middle of laisse 85 he promises a public act of atonement (“Droit t’en ferai voiant maint chevalier,” l. 1745), which Bernier angrily rejects: not till the blood flowing down from his head returns there of its own accord, and not till he can obtain revenge (?), will he consider such an offer (ll. 1746-55).⁵⁹ In laisse 86 Raoul’s offer is amplified to occupy nineteen lines (1759-77) and is followed by a murmur of approval among the Franks: “Ceste amendise est bele! Qi ce refuse vos amis ne vieut estre” (1778-9). Laisse 87 then combines elements from both these preceding accounts. Raoul urges Bernier to accept reparation; Bernier refuses in almost identical terms to those which he had used before; and Guerri amplifies the reaction of the Franks, condemning Bernier for his obduracy in the face of so splendid an offer (1792-5). Once again details of who says what when are effaced before Raoul’s urgent desire to placate, and Bernier’s furious implacability.⁶⁰

This study quite explicitly forestalls any temptation to regard it as definitive by the logic-defying concern for symmetry with which it marshalls two arguments in favour of the inseparability of character and plot, and two each for the primacy of the one over the other. Perhaps the whole investigation should be reoriented away from formal relations towards an examination of reference: after all, the fact that in the chansons de geste the issues of honour and dishonour, loyalty and treachery, wealth and destitution, life and death are all superimposed and mutually reinforcing is what makes these narratives impressive, their heroes moving even today and clearly far more so to contemporary audiences in a community still strifetorn and in many ways primitive. One framework of reference for this type of study which has already yielded promising results is medieval law; further theoretical approaches are suggested by medieval personality theories and the philosophy of action. My main concern here has been to challenge the easy assumption that either character or plot has a simple, unanalysable priority of status in the economy of a literary text.

Notes

1. Editions cited are as follows:

   Aiol: chanson de geste, ed. J. Normand and G. Raynaud (Paris: Didot, Societé des anciens textes français, hereinafter referred to as S. A. T. F., 1877).

   Le charroi de Nîmes (…), ed. Duncan McMillan (Paris: Klincksieck, 1972).

   La chevalerie Ogier de Danemarche, ed. Mario Eusebi (Milan: Istituto Editoriale Cisalpino, 1963).

   Les rédactions en vers du “Couronnement de Louis,” ed. Yvan G. Lepage (Geneva: Droz, Textes li Héraires français, 1978). The text cited is the AB version.

   Garin le Loheren (…) ed. Josephine E. Vallerie (Ann Arbor: Edwards Bros. Inc., 1947).

   Girart de Roussillon, ed. W. Mary Hackett (Paris: Picard, S. A. T. F., 1953-55).

   “Girard de Vienne” de Bertrand de Bar-sur-Aube, ed. Wolfgang van Emden (Paris: Picard, S.A.T.F., 1977).

   La chanson de Guillaume, ed. Duncan McMillan (Paris: Picard, S. A. T. F., 1949-50).

   Les deux rédactions en vers du “Moniage Guillaume,” ed. Wilhelm Cloetta (Paris: Didot, S.A.T.F., 1906, 1911).

   Raoul de Cambrai, ed. P. Meyer and A. Longnon (Paris: Didot, S.A.T.F., 1882).

   Le roman de Renart, ed. Jean Dufournet (Paris: Garnier-Flammarion, 1970).

   La chanson de Roland, ed. Frederick Whitehead, 2nd ed. (Oxford: Blackwell, 1946).

   I would like to take this opportunity to thank my colleagues Dr. G. S. Burgess and Dr. O. A. C. Waite for their helpful and learned advice with this paper.

2. Edmond Faral, La Chanson de Roland, étude et analyse (Paris: Mellottée, Collection Chefs d’oeuvre de la littérature expliqués, 1933), p. 261.

3. Pierre Le Gentil, La Chanson de Roland, first printed 1955; 2nd ed. rev. (Paris: Hatier, Collection Connaissance des Lettres, 1967), p. 123.

4. Jean Misrahi and William Lee Hendrickson, “Roland and Oliver: Prowess and Wisdom, the Ideal of the Epic Hero,” Romance Philology 33 (1979-80), 357-72, p. 359.

5. Tony Hunt, “Character and Causality in the Oxford Roland,” Medioevo Romanzo 5 (1978), 3-33, pp. 12-13. See also his article “The Tragedy of Roland: an Aristotelian View,” Modern Language Review, 74 (1979), 791-805; Sarah Kay, “Ethics and Heroics in the Song of Roland,” Neophilologus, 62 (1978), 480-91; and François Suard, “Le personnage épique,” in Proceedings of the 5th Conference of the Société Rencesvals, Oxford 1970 (Salford: University of Salford Press, 1977), 167-76.

6. Cf. the defence of character study by Marguerite Rossi, Huon de Bordeaux, (Paris: Champion, Collection Nouvelle Bibliothèque du Moyen Age, 1975), 463-67.

7. See for example John Bayley, “Character and Consciousness,” New Literary History 5 (1973-4), 225-35.

8. Roland Barthes, “An Introduction to the Structural Analysis of Narrative,” New Literary History 6 (1974-5), 237-72, pp. 256-57 (translation, with additional material, of an essay originally published in Communications 8 [1966]).

9. E.g. Norman T. Burns and Christopher J. Reagan, eds., Concepts of the Hero in the Middle Ages and the Renaissance (Albany: State University of New York Press, 1975); Micheline de Combarieu du Grès, L’idéal humain et l’expérience morale chez les héros des chansons de geste, des origines à 1250 (Aix-en-Provence: Publications de l’Université de Provence, 1979).

10. See D. J. A. Ross, “Old French,” in A. T. Hatto, ed., Traditions of Heroic and Epic Poetry, 1, The Traditions (London: Modern Humanities Research Association, 1980), 79-113, pp. 90-91: “In the Middle Ages the chansons de geste seem generally to have been accepted as having a basis of valid historical fact.”

11. Hans Robert Jauss, “Levels of Identification of Hero and Audience,” New Literary History 5 (1973-4), 283-317.

12. Northrop Frye, Anatomy of Criticism: Four Essays (Princeton: Princeton University Press, first published 1957, p.b. 1966), pp. 33-4 (1966 edition).

13. A. J. Greimas, Sémantique structurale: recherche de méthode, (Paris: Seuil, 1966). It has been pointed out to me that by trying to integrate a theory of narrative within a theory of poetry, Greimas risked losing sight of the specificity of either.

14. “Dans l’étude du conte, la question de savoir ce que font les personnages est seule importance; qui fait quelque chose [est une] question qui ne se pose qu’accessoirement,” V. Propp, Morphologie du conte (Paris: Seuil, 1970), p. 29, cited by Michel Mathieu, “Les acteurs du récit,” Poétique 5 (1974), 357-67, p. 357, who endorses this bias.

15. A. J. Greimas, “Les actants, les acteurs et les figures,” ed. Claude Chabrol et al., Sémiotique narrative et textuelle (Paris: Larousse, Collection Université, 1973), 161-76.

16. A further refinement to the theory is the introduction of modality permitting the nature of the Subject’s agency to be analysed according to whether its “deep structure” action implies the presence of a modal verb.

17. Larry S. Crist, “Roland, héros du vouloir: contribution à l’analyse structurale de la Chanson de Roland,” in Mélanges de Philologie et de Littératures Romanes offerts à Jeanne Wathelet-Willem, Marche Romane No. spécial (1978), 77-101. See also Patricia Harris Stablein, “The Structure of the Hero in the Chanson de Roland: Being and Becoming,” Olifant 5 (2) (Dec. 1977), 105-19, which suggests that the role of hero is divided between Roland, Oliver, and Charles. Recent studies of other medieval genres influenced by a functionalist approach include Pierre Gallais, “L’hexagone logique et le roman médiéval,” Cahiers de Civilisation Médiévale 18 (1975), 1-14, 133-48; Donald M. Maddox, Structure and Sacring: The Systematic Kingdom in Chrétien’s “Erec and Enide,” French Forum Monographs 8 (Lexington: French Forum, 1978); and several studies by Evelyn Birge Vitz including “Narrative Analysis of Medieval Texts: La fille du comte de Pontieu,” Modern Language Notes 92 (1977), 645-75. Vitz is critical of Greimas’ concept of the Subject which, however, she seeks always to equate with a single surface structure character.

18. Jean-Michel Adam et Jean-Pierre Goldenstein, Linguistique et discours littéraire: Théorie et pratique des textes (Paris: Larousse, Collection Université, 1976), pp. 108-120.

19. Sarah Kay, “The Nature of Rhetoric in the Chanson de Geste,” Zeitschrift für Romanische Philologie 94 (1978), 305-20, p. 316.

20. Cf. Raoul de Cambrai, ll. 347, 515, 625, 1549, 1555 (Raoul) with ll. 395, 591, 5596 (Bernier) and 3824, 4326, (Gautier).

21. Cf. ll. 392, 399, 831, etc. (Raoul o le vis cler and variants) with ll. 30, 364, 1217, etc. (Aalais … o le vis cler, and variants); ll. 506 (Raoul’s fier contenant) with l. 3579 (Guerri’s fiere contenance). Guerri has a fier vis and fiere veüe in ll. 2534, 1208; Raoul’s coraige is fier in ll. 1629, 1728, etc.

22. Cf. ll. 1365 (Marsent), 2024 (Wedon), 4501 (Bernier) and 5385 (Ybert).

23. Raoul is saige, l. 781, as also Guerri, ll. 1045, 2012, and Gautier, l. 3998, but of all three is it repeatedly said that le sens quide / quida changier (e.g. ll. 3081, 3402, 4505).

24. Cf. l. 2819 (Raoul) with l. 2906 (Rocoul), and also, though the formula is a slightly different one, with ll. 146 (Giboïn), 790 (the hostage Wedon de Borbone), and 1316 (Bernier).

25. Raoul is frequently ber (e.g. l. 836), as are Giboïn (e.g. l. 116), Guerri (e.g. l. 295), Louis (e.g. l. 844), and Gautier (e.g. l. 4946); the term gentis, used of Raoul (e.g. l. 372), is also used of Wedon (e.g. l. 1980), Bernier (e.g. l. 4264), and the abbé (l. 5310); Raoul is preu in ll. 534, 5170, as are Gautier (e.g. l. 4794), Guerri (e.g. l. 3221), Bernier (e.g. l. 3436). The qualificatives au coraige vaillant and cortois are likewise shared: cf. ll. 338, 2408, 2658, 4916, 4917, and 372, 5386, 5514, 5553.

26. Cf. ll. 2010, 881, 5169 and 2883, 778. This consistency of characterisation in the poem is commented on by Italo Siciliano, Les chansons de geste et l’épopée. Mythes—Histoires—Poèmes, Biblioteca di Studi Francesi 3 (Turin: Società Editrice Internazionale, 1968), 399: “Tout divers qu’ils soient, les preux et les sages du Raoul ont sensiblement la même taille, tous sont coupés dans la coulée d’une larve ardente et mouvante. Exception faite pour la triste Marsent, ils s’assemblent et se ressemblent. Toute porportion gardée, Aalais, Raoul, Guerri le Sor, Bernier, Garnier [Siciliano means Gautier], Aliaume sont de la même race et de la même force. Ici c’est le choeur qui chante la chanson et qui en fait l’unité, et c’est un choeur inquiet et inquiétant qui s’agite, court d’un bout à l’autre de la pièce, tourne en rond, s’épuise en interminables disputes, vit et périt de sa violence.”

27. Arnaldo Moroldo, “Le portrait dans la chanson de geste,” Le Moyen Age 86 (1980), 387-419 and 87 (1981), 5-44.

28. Cf. Barthes’ assertion, in “The Structural Analysis of Narrative,” p. 263, that the creation of “psychology” in a fiction presupposes mobility of stance from 3rd to 1st person and back again, which is transparently based on post-medieval narrative.

29. Cf. Jean Frappier, Les chansons de geste du cycle de Guillaume d’Orange, 1 (Paris: Société d’édition de l’enseignement supéricur, 1955), pp. 173-4: “les personnages existent de l’intérieur, ils ne révèlent pourtant leur nature morale que de l’extérieur, par leurs gestes, leurs paroles et leurs actes […] Sans chercher à rendre nos héros plus complexes qu’ils ne sont, sachons goûter une simplicité expressive et franche qui parvient à montrer ou à suggérer la qualité des âmes.”

30. The treacherous Bordelais, with whom for the most part the initiative lies, are more obviously imbued with “personality.” In a sensitive study aimed at differentiating Garin from his brother Begon, Anne Iker Gittleman notes that often it is the absence of a descriptive element or motif that characterises Garin, to whom she attributes “une grandeur dépouillée”: see Le style épique dans “Garin le Loheren,” (Geneva: Droz, 1967), p. 215.

31. Cf. W. W. Ryding, The Structure of Medieval Narrative (The Hague: Mouton, 1971), pp. 45-6.

32. Cf. Beate Schmolke-Hasselmann, Boire et mengier: essai d’une analyse structurale de la “Chanson de Guillaume,” paper read at the 17th International Congress on Medieval Studies, Kalamazoo, May 1982. I would like to thank Dr. Schmolke-Hasselmann for kindly sending me a copy of this as yet unpublished paper.

33. See Pauline Matarasso, Recherches historiques et littéraires sur “Raoul de Cambrai” (Paris: Nizet, 1962), p. 146.

34. Le Roman de Flamenca, ed. Ulrich Gschwind (Berne: Francke, 1976), vol. 1, 621 ff., and vol. 2, notes to ll. 630-1, 641-2, 677-8.

35. Cf. Suard, “Le personnage épique” (cited in fn. 5), p. 168, for another example of possible inconsistency in Ganelon.

36. Hunt, “Character and Causality in the Oxford Roland” (cited in fn. 5), p. 11.

37. Rossi, Huon de Bordeaux, p. 484.

38. ———Pace D. D. R. Owen, “Structural Artistry in the Charroi de Nîmes,” Forum for Modern Language Studies 14 (1978), 47-60.

39. In a fascinating article “Type et individu dans l’autobiographie’ médiévale,” Poétique, 6 (1975), 426-45, Evelyn Birge Vitz attempts a study of the medieval concept of personality based on Abelard’s Historia Calamitatum. A part of her exposition deals with the question of consistency: concerned to present his experiences more from the point of view of their intensity (and of his own superiority, or extreme sensibility) than of their fundamental unity, Abelard allowed various parts of his career to appear completely compartmentalised one from another.

40. Jean Subrenat, Etude sur “Gaydon”, chanson de geste du XIII^e siècle, Etudes Littéraires. (Aix-en-Provence: Editions de l’Université de Provence, 1974), pp. 210-11.

41. For a study of the pacific ethos of Girart de Roussillon, see René Louis, De l’histoire à la légende: 2, Girart, comte de Vienne, dans les chansons de geste (Auxerre: Imprimerie Moderne, 1947), 1: 405-16. Cf. also the remarks under the heading Condamnation de l’attitude belliqueuse in vol. I of Micheline de Combarieu’s L’idéal humain, p. 147 ff.

42. For further inconsequences in Girart de Vienne, see Ryding, Structure of medieval narrative, pp. 51-2.

43. There is a similar scene in La Chevalerie d’Ogier de Danemarche, ll. 10454-70. A more successful analogue is the supernatural fire which consumes the standards of Girart and Charles Martel in Girart de Roussillon, thus forcing an end to their battle at Vaubeton (l. 2874 ff.), since it operates as an effective metaphor of the violence and irrascibility of the two antagonists, as also of the cauterising process of self-examination in the two camps following this disturbing experience. Charles Martel is pronounced to be accursed of God (ll. 2906-26); despite a rehearsal of his grievances, Girart is condemned by his uncle for not seeking peace (ll. 3018-30). God’s intervention is in harmony with, but does not cause, the signal act of personal abnegation which actually finalises the peace negotiations, namely Tierri’s voluntary exile (ll. 3122-35, especially ll. 3123-4: “Ne place a Damlideu, au manne rei, Que ja mais per mon cors nus on gerrei!”). In this case, the credibility and integrity of the characters seem to me to have been maintained alongside the expression of God’s will.

44. Evelyn Birge Vitz, “Desire and Causality in Medieval Narrative,” Romanic Review, 71 (1980), 213-43, pp. 223-24.

45. W. C. Calin, The Epic Quest: Studies in Four Old French Chansons de Geste (Baltimore: John Hopkins University Press, 1966), especially pp. 5-31. Calin stresses the hero’s representative character, p. 23: “All of society participates in Aymeri’s quest.”

46. See Minette Grunman, “Temporal Patterns in the Chanson de Guillaume,” Olifant, 4 (1) (Oct. 1976), 49-62, p. 56.

47. Jean-Charles Payen, Le motif du repentir dans la littérature française médiévale (des origines à 1230) (Geneva: Droz, 1967), p. 180.

48. For remarks on William’s celebrated rabbit punch, consult Frappier, Guillaume d’Orange, 1:94-96. On the post mortem prospects of traitors, see Payen, Le motif du repentir, p. 158 and notes.

49. Quoted by William Wistar Comfort, “The Character Types in the Old French Chansons de Geste,” Publications of the Modern Language Society of America 21 (1906), 279-461, p. 356. Cf. also Garin le Loheren, l. 6357, “Hom desloiax ne puet longues garir,” cited by Micheline de Combarieu, L’idéal humain, 1:98.

50. See my article La composition de “Raoul de Cambrai,” forthcoming in the Revue Belge de Philogie et d’Histoire.

51. See Sarah Kay, “The Contrasting Use of Time in the Romances of Jaufre and Flamenca,” Medioevo Romanzo 6 (1979), 37-62, especially pp. 39-40 and notes.

52. See Herman Braet, Le songe dans la chanson de geste au XII^e siècle, Romanica Gardensia, 15 (Gent: Romanica Gandensia, 1975), 198.

53. On the integration of the dream to other forms of annonce, see Braet, Le songe, p. 103 ff.

54. Thus E. Vance writes in “Roland et la poétique de la mémoire,” Cahiers d’Etudes Médiévales 1 (1975), 103-115, p. 106: “Les héros du Roland non seulement parlent-ils dans les mêmes formules métriques que le poète, mais ils ont recours aux mêmes épithètes, aux mêmes listes, voire à la même pré-science que le poète.” This article has appeared in English as “Roland and the Poetics of Memory,” in Textual Strategies: Perspectives in Post-Structural Criticism, ed. Josué Harari (Ithaca: Cornell University Press, 1979), 374-403.

55. For some other, less convincing examples of the characters of Raoul predicting their own future, see Carlo Pica, “Raoul de Cambrai: crisi di un sistema,” Cultura Neolatina 40 (1980), 67-77, pp. 71-72.

56. The Brussels fragment of Raoul de Cambrai adds after the passage corresponding to ll. 944-945 “Mais tant vous di que ie lor aiderai; Se pooir ai, sachiez ie vous nuirai,” which would, of course, deprive Raoul’s words of their prophetic force. See A. Bayot, “Raoul de Cambrai (Brüsseler Fragmente),” Revue des Bibliothèques et Archives de Belgique, 4 (1906), 412-29.

57. Jean Rychner, La chanson de geste: essai sur l’art épique des jongleurs (Geneva: Droz, and Lille: Giard, 1955).

58. Laisses 82-83 repeat in very similar terms Bernier’s account of his mother’s forcible abduction by Ybert; the mention of his mother’s nobility in the very short laisse 80 is amplified in laisse 81.

59. The text is not very clear; Meyer suggests the possibility of a lacuna between ll. 1754 and 1755. I would punctuate and translate as follows:

    Ja enver vos ne me verres paier
    Jusqe li sans qe ci voi rougoier
    Puist de son gre en mon chief repairier.
    Qant gel verai, lor porrai esclairier
    La grant vengance qe vers ton cors reqier—
    Je nel laroie por l’or de Monpeslier.

    [1750-55]

    “You will not see me reconciled with you until the blood that I see running red returns to my head of its own accord. When I see that, then I shall give vent to the terrible vengeance I desire against you—not for all the gold of Montpellier would I forego it!”

60. A further example of ambiguity is presented by laisses 143, 144, 147, and 151 describing Ernaut’s flight from Raoul on the battlefield.