The battle of Roncesvalles was an ambush by a group of Vascon soldiers on August 15, 778 at the pass of Roncesvalles in the Pyrenees, during which the rearguard of Charlemagne”s army, returning from Saragossa, was destroyed. Several personalities of the Frankish kingdom were killed in this battle, including the knight Roland, prefect of the march of Brittany, who commanded the rear guard.
This battle in French history is recounted by the monk Éginhard in the Vita Karoli Magni (it is most famous for the epic, not historical, account of the Chanson de Roland, a chanson de geste composed in the eleventh century whose main character is the knight Roland and which attributes the attack to the Saracens. The exact location of the battle is uncertain, but a memorial recalling the legend of Roland stands in the present-day village of Roncesvalles.
At the death of Pepin in 768, Charles was elevated to royalty. Liuba II (traditionally called Wolf II), Duke of Gascony, swore an oath to him. The following year, he entrusted Charles” court with the education of his son Sanz, and asked him to protect his goods and lands. Gascony extended from the Garonne River to the southern Pyrenees, including cities such as Pamplona.
In 777, at the plea of Paderborn, Charlemagne received the ambassador of the Muslim governor of Barcelona, Suleiman al-Arabi (also spelled Sulayman) – who had revolted against Abd al-Rahman I, the emir of Cordoba – who asked for the help of the Franks to hold the city of Saragossa.
Perhaps he offers himself as a vassal to Charles, seeking his protection against the emir whom he has betrayed twice. Perhaps he offers Charles to repel the Emir by taking territory from him, constituting an allied buffer state, avoiding raids.
Zaragoza was a major strategic military and economic issue, allowing it to control the Ebro. The city is also a high place of Christianity in the Iberian Peninsula, and an enclave of Christian confession in a territory then under Muslim domination. Prudence in the fourth century in the Peristephanon sings the city and is the image of greatness attached at the time to Zaragoza. The cathedral contains the tombs of many Christian martyrs, including the relics of St. Vincent. It is not impossible that the so-called miraculous pillar of the Virgen del Pilar was already made at the end of the 8th century.
But Charles was surely less attracted to Zaragoza than he was worried about the activities of the ambitious Banu Qasi clan, an ancient Visigothic Islamic lineage, led by Abu Tawr, whose father had already joined the alliance with the emir. From their strongholds in Olite and Tudela, they sought to take control of Pamplona, which was under Frankish rule, as well as Huesca and Girona, which were dependent on the Emirate.
Charles, if he leaves – supported by Pope Adrian I who wishes him “happy victory” – to defend the oppressed Christians, it is the Franci homines of Pamplona, those that the Muwallads (recent Muslims) have just subdued, and this on the territory of the Frankish kingdom. The Banu Qasi had subdued the city that Liuba II had placed under royal protection nine years earlier. Charles was thus part of the old fight against these sons of the Goths, who were considered capable of all heresies (since homoeism).
Charles crossed the Pyrenees with two armies: one in the east, composed of Bavarians, Burgundians, Austrasians, Provençals, Septimanians and Lombards, crossed the Perthus Pass. The western army, led by Charles, was composed of Neustrians, Bretons, Aquitanians (newly organized territory between the Loire and the Garonne) and Gascons (from the southern Garonne).
The gates of Pamplona open at the sight of Charles. Abu Tawr tells him the submission of his cities and gives his son and brother Abu Talama as hostages as promised, as a guarantee. Suleiman led Charles to Zaragoza where the junction was made with the eastern army that had just subdued Girona, Barcelona and Huesca.
But in Saragossa, El Hussayn, who ran the city with Suleiman, refused to open the city gates to the Franks. Charles was not in a position to conduct a siege and did not want to linger over this plot at the risk of weakening his army and risking a trap being sprung on him. He takes Souleiman as a hostage. The heat, the risk of running out of food and leaving the kingdom too poorly defended, commanded him to send the army back east.
Charles learns that the Banu Qasi are taking over Pamplona and agitating the population. Charles, before crossing the Pyrenees, returned to Pamplona and found the doors closed. But the Banu Qasi probably expected the destruction – or at least the weakening – of the Frankish army at the siege of Zaragoza; their surprise finally forced them to abandon their ambitious acquisition. Charles convinced the Navarii – defenders of Pamplona – to stop obeying the Banu Qasi. These Navarii took an oath to him. In order to prevent Pamplona from being targeted again by ambitious people because of the strategic nature of its defenses, Charlemagne had the walls of the city razed to the ground – certainly while waiting to be able to install a substantial defensive force.
In 1867, Léon Gautier wrote in the introduction to his analysis of the Chanson de Roland, which closes the second volume of his monumental French Epics: “Roncesvalles is at the center, it is at the heart of the whole cycle of Charlemagne. Roncesvalles is the capital fact of all the Geste du Roi, it is the core of all the Carlovingian poems”. The emotion caused by this literary battle of Roncesvalles prompted medieval historians and medieval literature to take an early interest in the historical reality that served as its backdrop. However, no archaeological research has been able to shed light on it, and it is therefore known only through historiographical sources. In 1850, François Génin knew of only two contemporary texts of the facts: the Annales Royales, up to 829, and Eginhard”s Life of the Emperor Charlemagne.
From the first half of the nineteenth century, medievalists, including Gaston Paris, who published his Histoire poétique de Charlemagne in 1865, based largely on the ancient manuscripts collected in Monumenta Germaniæ Historica . Scriptores edited by Georg Heinrich Pertz from 1826, have been constantly searching for the historical foundations of this battle to determine how the reality could have served as inspiration for several major texts of medieval literature. When Francisque Michel published in France in 1837 the first edition of Chanson de Roland in its version of the Oxford manuscript, it seemed to him to be a given that the famous battle of the song referred to a real ambush suffered in 778 in the Pyrenees by the rearguard of Charlemagne”s army on its return from the Spanish campaign.
Already in 1817 and the first study referring to the Oxford manuscript by Louis de Musset, the historicity of Roland is defended on the basis of Eginhard”s Life of the Emperor Charlemagne, which is considered a historiographical source of reference. The identification of the battle of Roncesvalles in the chansons de geste with the defeat of the Pyrenees is much older. This is what Jean Papire Masson does, for example, defending in 1577, also based on Eginhard but also on ecclesiastical chronicles such as that of Flodoard, the idea that the Chronicle of Turpin, which also describes the battle of Roncesvalles, is largely legendary.
In 1959, the scholar Ramón Menéndez Pidal attempted to synthesize the research that had been carried out over the last century and a half in the sum he devoted to the Song of Roland. He includes extracts from the main medieval texts that allow us to get a closer look at the reality, and he organizes them into two categories: the Carolingian annals, composed of sixteen Latin texts written approximately between 791 and 906, and three extracts from late Arabic chronicles. To this corpus, some historians, including Ramón Menéndez Pidal himself, add various secondary sources.
The passages from the Carolingian annals of 778 relating to the Spanish expedition are copied and reworked to such an extent that it is possible to determine their “family tree” and thus follow the evolution of their modifications over time. In order to allow an easy comparison, Ramón Menéndez Pidal separates them into four groups composed of texts clearly inspired by each other: the first, the Annals of Metz, which do not know the battle of Roncesvalles; the second, the Royal Annals, which see the ambush appear in 829 and are later rewritten in a literary style; the third, the brief annals, which do not know the battle but seem to have a little more information about the Saracens, and finally the group of very brief annals with a lapidary text:
All of these sources echo Charlemagne”s expedition to Spain in mid-778, but only four of them mention an ambush in which the Frankish army fell as it crossed the Pyrenees to return north to confront the revolting Saxons.
The very brief annals do little more than confirm the reality of Charlemagne”s Spanish expedition in 778. Thus the second continuation of the Annals of Saint-Amand, one of the oldest texts since it was written before 791, is satisfied with a single sentence: “778 (779) Carlus rex fuit in Hispania ad Caesaraugusta”, which translates as “778 (779) King Charles was in Spain at Saragossa”.
As for the more extensive chronicles, they give crucial details about this campaign. The Royal Annals, up to 801 for example, written probably in 788, specify the constitution of the Frankish army and name the subjected peoples:
A few years later, perhaps in 805, the Annals of Metz also describe the whole expedition, but in a much more hagiographic and religious style:
These short passages indicate that Charlemagne came to Spain with two army corps that met in Saragossa. He received Muslim hostages there, then destroyed Pamplona, and finally returned to Frankish country to deal with the rebellious Saxons. But beyond the general context of the campaign, they present variations. For example, some chronicles, such as the Annals of Lorsch or the Chronicle of Moissac, which are influenced by the clerical point of view, replace the Basques with Saracens, and for them, Charlemagne took Pamplona from the Muslims. These chronicles are also incomplete and imprecise. They do not explain, for example, why hostages were handed over in front of Zaragoza or why Pamplona was destroyed. They are even difficult to understand at times. Who are the Hispani Wascones, translated as “Hispanic Basques” in the Royal Annals, until 801? What is the difference between the Basques
There is no question of ambush or defeat: the king”s victory is total. If the annals do not mention it, perhaps it is because these courtier texts were to present, during Charlemagne”s lifetime, the Spanish expedition as a success by concealing what could look like a failure. The Swiss philologist Paul Aebischer goes further, speaking of “imperial censorship with the aim of hiding the disaster of the Pyrenees, minimizing its consequences, and preserving the king”s reputation as an invincible leader. For his part, the historian Robert Fawtier considers that the Carolingian annals are similar to official communiqués published in times of war, emphasizing, as everywhere and at all times, the victories to the detriment of the defeats. But perhaps this defeat was insignificant, as Joseph Bédier maintained. The annalists would then simply not have considered it relevant to report it.
However, around 814, the year of Charlemagne”s death and the beginning of the reign of his son Louis the Pious, the Annals of St. Gallen published by Baluze summarize the year 778 with an obscure sentence heavy with meaning: “DCCXXVIII. Hoc anno domnus rex Carolus perrexit in Spania et ibi dispendium habuit grande” translated as “778. In that year, the Lord King Charles went to Spain, where it cost him dearly.
While the previous annals say nothing about an ambush, the Royal Annals, until 829, provide details, unknown until then, about the battle of the Pyrenees:
Once Pamplona was razed and the army returned to the north, the Gascons, in Latin Wascones, attacked and decimated the rearguard of the Frankish army in the Pyrenees. The date of this admission of an important setback is debated. It is between 801 and 829, that is, between the end of the reign of Charlemagne and the beginning of that of Louis the Pious. The late unveiling of the sad reality, at least twenty years after the facts, is often explained by the fact that the truth being known by all, it was no longer possible for the annalists to continue to hide it. Jules Horrent, who believes that the reworking of the Annales royales took place after Charlemagne”s death, considers that it was no longer necessary to hide a disaster that had so “darkened the king”s heart”. Going against the consensus of historians, Bernard Gicquel considers that the new version of the Annals is later than 824, the date of the defeat of Roncesvalles against the Vascons during the reign of Louis the Pious, and that they invent a defeat of the father in 778 at the same place to serve the imperial ideology for the benefit of the son.
The Annales reworked designate the attackers by the Latin word Wascones, which historians interpret with great difficulty in the context of the end of the eighth century. Some, such as Évariste Lévi-Provençal or Pierre Narbaitz, translate it as “Vascons,” others, such as Gaston Paris or Joseph Bédier, as “Basques,” and still others as “Gascons,” which is the choice of François Guizot in his 1824 translation. But some also alternate “Basques” and “Gascons” in their studies, thus showing the difficulty they have in identifying the mountain people who attack the rearguard. The Wasconia is one of the most problematic regions for historians of the High Middle Ages, and it is not known whether the Frankish annalists were aware of a dichotomy between the Vascons of the North, often called “Gascons”, and those of the South, traditionally called “Basques”. This separation is all the more delicate since the Basque language was then spoken in Aquitaine as far as Toulouse.
Eginhard wrote his Life of the Emperor Charlemagne, in Latin Vita Karoli Magni imperatoris, probably between 826 and 829 in the palace of Aix. This book, of which 134 complete manuscripts have been preserved, is a fundamental source for historians for the knowledge of Charlemagne”s reign and person. Chapter 9, entitled by Strabo “What he did in Hispania and the blow that the Basques inflicted on his army” describes the ambush in which Charlemagne”s army fell:
The king”s friend and the master of his palatine school recounts the battle half a century after the fact: the army advancing in single file through the Pyrenees on the return from the Spanish campaign, the ambush in which Charlemagne”s army was defeated in a single day, and the prestigious deaths that could not be avenged. This short text is clearly inspired by the reworked Royal Annals, but it adds details that they ignore. Joseph Bédier believes that Eginhard, who was admitted to the court in the early 790s and lived in the emperor”s immediate circle, may have frequented those who had taken part in the Spanish campaign. He would have reported their memories in his Vita Karoli.
Ramón Menéndez Pidal is the first to point out the singularity of this chapter. He notes, for example, that the short Spanish campaign of 778 occupies more lines than any of the other nine wars fought by Charlemagne. For each of them, Eginhard makes an effort to synthesize and omit events of significant historical importance. Conversely, he offers an unparalleled wealth of detail in describing the disastrous ambush in the Pyrenees. Finally, contrary to his habit, he mentions by name three Palatine soldiers killed in the attack, although their names are not mentioned in the Annals. Ramón Menéndez Pidal then suggests that Eginhard was inspired, in addition to the Annals, by a sung history contemporary with the writing of the Vita Karoli, which he calls a “topical song”, and which would give rise, among other things, to the Chanson de Roland nearly three centuries later. The historian Michel Rouche goes a little further in asserting that popular history eventually supplanted the official history conveyed by the clerics. Eginhard, but also the annalists of the Royal Annals, would have recorded while censoring the orality “singing the true sufferings and the true hero”, that is to say Roland.
The mention of the prefect of the march of Brittany next to two other well-known personalities, however, has been the subject of controversy since the first quarter of the 19th century when it was discovered that not all the manuscripts of the Vita Karoli contain it. These were classified into several categories called A, B and later C, according to minor editorial details such as the dedication to Louis the Pious, the chaptering or even the mention of Roland in chapter 9. The Swiss medievalist André de Mandach went so far as to propose in 1961 that the name of Roland, absent from the type B manuscripts which were then supposed to be the oldest, had been added to the text four centuries after its initial writing. Later epigraphic studies suggest, however, that the three types of manuscripts date from the same year 820, suggesting that Eginhard produced several versions of his work, for example for a first reading or for corrections.
The battle of the Pyrenees is also mentioned in the Vita Hludovici pii translated as “Life of Louis the Pious”, also known as Vita Hludovici imperatoris i.e. “Life of the Emperor Louis”, written in 840 or 841 by an anonymous person known as the Astronomer. Louis was born during the Spanish expedition of his father Charlemagne, whom the Astronomer describes in the following pompous terms:
As for the chronicle of L”Astronome, in his Life of Louis, if it designates Saracens as general enemies of the expedition, it does not mention any Gascons concerning the battle itself
The main Arabic sources relating to the Spanish expedition are few: a short passage from the Akhbar Madjmu”a, a collection of chronicles compiled in the eleventh century, and two extracts from Ibn al-Athîr”s Kâmil dating from the thirteenth century. These three texts provide valuable information about the belligerents, but only Ibn al-Athîr”s annal for the year 157 of the hegira, i.e. from November 21, 773 to November 10, 774 in the Gregorian calendar, suggests that Muslims attacked the Frankish army on the way back:
Ibn al-Athîr uses the lost history of Ahmed al-Rasi, who died in 955, who himself had much earlier annals. Therefore, even if he makes a mistake regarding the date of the expedition, some medievalists such as Ramón Menéndez Pidal or Gaston Paris accept his late chronicle as reflecting a part of historical truth that can shed light on the designation of the protagonists of the battle. Others, however, such as René Basset, Robert Fawtier and Joseph Bédier, reject these sources completely as inconsistent and containing anachronisms. The historian Louis Barrau-Dihigo even considers that they are strongly influenced by Latin sources, which makes them worthless. In an intermediate attitude, some medievalists like Jules Horrent exclude them while accepting their authenticity. They consider them to be of little relevance to the battle itself because they do not refer directly to it. Finally, others, such as the professor of medieval literature Michel Zink or Michel Rouche, make the hypothesis that the chronicle of Ibn al-Athîr is closer to historical reality than the Latin sources.
August 15 is the day of Aggiard”s death, as it appeared in his epitaph, the text of which in elegiac distichs has been preserved by ms 4841, a Latin manuscript kept at the Bibliothèque nationale de France :
This manuscript, published for the first time by the German historian Ernst Dümmler (de) in 1873, attracted the attention of Gaston Paris who established the correspondence with the text of the Vita Karoli of Eginhard. He deduced that the person to whom the text refers is the seneschal Eggihard, who died during the battle, which took place on August 15, 778, if we are to believe the date inscribed on the epitaph: “the eighteenth day of the Kalends of September.
The historian René Louis suggests that the church of Saint-Vincent to which the epitaph refers and where Eggihard was buried would be in Metz. This implies that the body of the seneschal must have been transported for most of the return journey from Spain. It seems that the journey was relatively short because Charlemagne arrived in Herstal on September 24, 778, i.e., a little more than a month after crossing the Pyrenees. But this journey of about 1,000 km in the middle of summer, with the coffin perhaps installed on an oxcart, seems hardly credible to Professor Bernard Gicquel, who comes to doubt the authenticity of the manuscript.
Robert-Henri Bautier does not believe in the transport of the body over such a great distance when the army was in a hurry to reach the Rhine. But he doubts René Louis” hypothesis and assumes that, as had been envisaged for a long time, the shrine of Saint-Vincent would be that of Dax. He therefore admits the authenticity of the epitaph and with the community of historians, recognizes that this date is the most probable. This date has excited the imagination, causing the medievalist Robert Lafont to write, for example, “Chance prepared the myth: August 15 is the Marian feast day, the day of the Dormition of the Virgin or of her Assumption.
As no archaeological evidence has ever been found, the location of the battle remains unknown. Various hypotheses have been put forward and the battle was not only located near the pass of Roncesvalles but all along the Pyrenean chain, from the Basque Country to Catalonia. For most historians, the route used would have followed the line of ancient Roman roads. It is the route and the place where it crosses the Pyrenees that differs according to the authors.
For most authors, the action took place on the ab Asturica Burdigalam road (from Astorga in Leon through Pamplona to Bordeaux) which crosses the Pyrenees at Roncesvalles. The expression “porz de Sizer” in the Song of Roland refers to the passes of the Cize region. Contrary to what popular tradition and some authors, such as Ramon d”Abadal i de Vinyals, affirm, the ancient route does not cross the Pyrenees at the pass of Roncesvalles itself (or the pass of Ibañeta, after the name of the nearby mountain): in fact, the present road was not opened until 1881; as for the name Roncesvalles (Orria or Orreaga in Basque), it only appears in the 12th century and does not exist in any document of the time.
Several authors (including Ramón Menéndez Pidal and Pierre Narbaitz) think that the route used passes a few kilometers to the east. The passes of Bentarte and Lepoeder, close to the Astobizkar, would be among the most probable.
In 1933, Robert Fawtier, taking up a hypothesis of Joseph Bédier, thought that the Roman road ab Asturica Burdigalam passed through the pass of Belate, north of Pamplona and 25 km west of Orreaga: the route envisaged from Pamplona would pass through the pass of Velate, the valley of Baztan, the Rio Maya, the pass of Otxondo, and would follow the valley of the Nive to Bayonne: he placed Roncesvalles there. “Bédier wondered whether Charles” defeat took place in the pass of Roncesvalles or in that of Velate.
Another location, proposed by Antonio Ubieto Arteta and retained by Robert Lafont, uses the Roman road Cæsar Augusta linking Zaragoza to Bearn. Passing through the valley of the Rio Gallego, the forest of Oza (Valle de Echo, province of Huesca), the pass of Pau (puerto del Palo) close to the Somport to come back down through the valley of Aspe, it was still maintained in the 9th century. From this point of view, the burt Sizaru of the Arab geographers and the porz de Sizer of the Song of Roland would be Siresa, where a monastery is reported from the ninth century, and the “Tere Certeine” of the Song would be the mountains Gibal-el-Sirtaniyyin mentioned by an Arab geographer as the place of the source of the Gallego river.
Other hypotheses are based on the absence of a place called Roncesvalles in the documents of the time, on the mentions in the Song of Roland of a return of Charlemagne by Narbonne and Carcassonne and of the Saracens” ride through the Cerdanya (the “Tere Certaine”) to support a passage through Catalonia: the possibilities include the Cerdanya. (Llívia valley) according to Adolphe d”Avril in 1865, the Perthus pass according to Rita Lejeune for whom the “Pyrenei saltus” mentioned by Eginhard (“Pyrenei saltum ingressus est”) designates the eastern Pyrenees, or even the high ports of Andorra for Marcel Baïche who notes that the toponymy of the Chanson is not Basque but Catalan: the porz de Sizer would be the port of Siguer. These hypotheses do not hold that Charlemagne took a Roman road, nor that he was returning from Pamplona, and they sometimes consider that his rearguard was not confronted with the Vascons but with the Saracens.
According to Jean Claret, a self-published author, the battle of Roncesvalles would not have taken place there, but rather in France, at La Unarde, a desolate place in the mountains in the present-day commune of Aston in Ariège mentioned in the IGN map (42° 41′ 30″ N, 1° 35′ 49″ E): “For 1,200 years, Éginhard led us to believe that the expedition was circumscribed to the Basque Country and that Roland died during an ambush led by Vascons. Fortunately, some weaknesses remained in his reasoning and by confronting them with those of the Arab chroniclers and others, we have been able to re-establish what seems to be the reality of the facts.”
In La baronnie de Miglos : étude historique sur une seigneurie du haut comté de Foix, published in Toulouse in 1894, Casimir Barrière-Flavy, devotes a chapter to an exploration at the Unarde site, presenting sketches of a scramasaxe and a knife found there.
: document used as a source for the writing of this article.