Was the Byzantine defeat at Manzikert the result of treachery on the battlefield—or a series of strategic miscalculations? A closer reading of the medieval chroniclers suggests that the search for betrayal began soon after the battle itself.
By George Theotokis
Few figures in Byzantine military history have been so harshly condemned in the aftermath of a single defeat as Josef Trachaneiotes, Roussel of Bailleul, and Andronikos Doukas. According to the dominant narrative, the first two disobeyed their emperor by failing to return and support him at a critical moment, while the third allegedly abandoned Romanos on the battlefield, ordering a withdrawal precisely when his intervention was needed to prevent the encirclement and collapse of the imperial line. Yet the credibility of these accusations is far from certain. Much depends on which sources one chooses to trust.
The Prelude to the Battle
The Battle of Manzikert marked the geopolitical high point of a prolonged contest between the Byzantine Empire and the Seljuk Turks over dominance in eastern Asia Minor and northern Syria. Despite its later significance in Byzantine–Seljuk relations, many scholars argue that Sultan Alp Arslan’s chief strategic concern was not Byzantium itself, but rather the Shiʿa Fatimid Caliphate. From this perspective, the Turkish operations of 1070–1071 that culminated at Manzikert were intended primarily to prepare the ground for a future campaign against Egypt.
To secure his position in Syria and protect the flanks of this anticipated advance, Alp Arslan moved to seize key fortresses, including Artchesh and Manzikert. At the same time, sustained Turkish incursions over preceding years had rendered the territory between Theodosiopolis and Manzikert largely depopulated and devastated. These developments ensured that the Seljuk sultan could operate in Syria without immediate threats to his rear.
The Byzantine Empire and its provinces (themes) and neighboring countries at the death of Basil II in 1025 AD. Map by Cplakidas / Wikimedia Commons
Alp Arslan’s winter campaign of 1070–1071 in Syria can therefore be regarded as a strategic success, as it considerably expanded his authority across northern Syria and the Upper Euphrates region. Nevertheless, his focus soon shifted away from Syria due to developments unfolding further north.
Insight into the circumstances that led Emperor Romanos IV to undertake his ill-fated expedition toward Lake Van is provided by the historian Nikephoros Bryennios the Younger. He recounts that during a military council convened while the imperial army was marching between Kaisarea and Sebasteia, both the Georgian magistros Joseph Trachaniotes and Nikephoros Bryennios the Elder advised against advancing into Turkish-held territory. Instead, they urged the emperor to remain at Theodosiopolis, thereby compelling Alp Arslan to engage the Byzantines on ground more favorable to Roman forces. Their counsel, however, was overridden by more aggressive voices within the council. As a result, the advance toward Khliat and Manzikert was approved—a decision that Bryennios later described with evident dismay and confusion.
Dividing his Forces: Overconfidence or Lack of Intelligence?
Diogenes’ imperial army remained stationed at Theodosiopolis until approximately mid-summer 1071. In late July or early August, however, the emperor’s increasing confidence prompted a decisive strategic miscalculation. Upon departing Theodosiopolis and advancing southeast toward Manzikert via Xnis/Xinus (modern Hınıs in the province of Erzurum), a march of roughly 230 kilometers, Romanos chose to divide his army. According to Michael Attaleiates, this decision was based on the emperor’s belief that the Turkish and Iranian Daylamite garrison defending Manzikert posed little serious resistance.
Additional sources indicate that Romanos detached contingents of Frankish and other mercenary forces, placing them under the command of Roussel de Bailleul and the magistros Joseph Trachaneiotes, and ordered them to advance toward the strategically vital fortress-city of Khliat. Control of Khliat was essential, as it dominated the western access routes to the Lake Van basin.
Coin of Romanos IV Diogenes. Photo: International Numismatic Club / Wikimedia Commons
Two points are especially significant. First, the force led by Roussel and Trachaneiotes was composed of elite troops and was, in fact, substantially larger than the contingent the emperor retained under his direct command. Second, Romanos intended to capture Manzikert with his own forces and did not plan to reunite with the detached army unless an extreme emergency arose. Nor did he envisage advancing on Khliat until Manzikert had first been brought back under imperial authority.
In my view, Romanos anticipated a rapid and straightforward operation that would result in the swift recovery of both Manzikert and Khliat. Events, however, would soon demonstrate how gravely mistaken this assumption was.
In order to contextualize Romanos’ strategic decisions, it is essential to examine the quality of Byzantine intelligence and, in particular, the role played by Leon Dabatenos. Dabatenos belonged to a prominent Armenian family, and a figure of this name is attested as vestarches and doux (or katepano) of Edessa between roughly 1071 and 1077. He appears to have assumed this position shortly after the Turkish siege of Edessa during the winter of 1070–1071.
Nikephoros Bryennios the Younger is our sole source for an official intelligence report said to have been delivered to Romanos by Dabatenos while the emperor was likely still stationed at Theodosiopolis with the main body of his army. According to this account, the report claimed that the Seljuk sultan, having learned of Romanos’ campaign against Khliat and Manzikert and fearing the strength of the imperial forces, had withdrawn from Persia toward Baghdad. Bryennios emphasizes the consequences of this information: trusting its accuracy, the emperor divided his army, retaining one portion near Manzikert while dispatching the other toward Khliat.
This testimony should be considered alongside a striking remark by the Muslim preacher and historian Sibṭ Ibn al-Jawzī (1185–1256), who notes that a Byzantine envoy returned to the imperial camp in high spirits and that this report reinforced Romanos’ determination to pursue Alp Arslan and engage him in battle. The parallel accounts raise an important question: are Bryennios and Ibn al-Jawzī referring to the same individual responsible for supplying intelligence to the emperor?
It would have been entirely reasonable for Romanos to request detailed information from the doux of Edessa concerning the size, composition, and morale of the Seljuk army. The crucial issue, however, lies in the reliability of that intelligence and the extent to which Leon Dabatenos could be trusted. A revealing passage from Matthew of Edessa’s Chronicle casts a long shadow over Dabatenos’ conduct. Matthew records that the doux of Edessa supplied Alp Arslan with horses, pack animals, and provisions, allowing the sultan to pass safely through the territory of Edessa before continuing eastward toward Mount Lesun.
Such cooperation strongly suggests that the governor of Edessa had, at least tacitly, recognized Seljuk overlordship. This naturally raises doubts about the credibility of the intelligence reportedly sent to Romanos regarding the Seljuk army’s intentions and disposition. Was Dabatenos deliberately providing misleading information, encouraging the belief that Alp Arslan would remain on the defensive in Iraq rather than confront the emperor in Armenia? Absolute certainty is impossible, but the circumstantial evidence is troubling—particularly when one considers that Dabatenos retained his position throughout the reign of Michael VII, remaining in office until around 1077.
Turning to the emperor’s decision to summon back the forces previously detached to Khliat under the command of Roussel and Trachaneiotes, this development occurred on Thursday, 25 August, the day before the Battle of Manzikert. Both Attaleiates and Skylitzes record that Romanos initially postponed engagement but ultimately abandoned any expectation of reinforcement from the Khliat detachment, having concluded that some serious impediment was preventing their return.
At the same time, the emperor appears to have believed that these elite units were merely delayed rather than having withdrawn altogether from the area. This assumption explains his resolve to offer battle on the following day, Friday, 26 August, relying on the troops immediately at his disposal, while still entertaining some hope that the absent forces might arrive shortly thereafter.
Attaleiates further notes that Romanos was unaware that Trachaneiotes, upon learning of the sultan’s advance against the imperial army, had led his entire force in retreat through Mesopotamia until reaching Byzantine territory. The historian characterizes this withdrawal in harsh terms, accusing Trachaneiotes of disgracefully abandoning both his sovereign and his obligations. Yet this raises an important question: how justified is such a severe judgment of the conduct of Roussel and Trachaneiotes?
Given the ideological divisions within the Greek narrative tradition—often categorized along pro- and anti-Romanos lines—Attaleiates and Skylitzes tend to mitigate the emperor’s responsibility for having detached nearly half of his army to Khliat. They argue that the division of forces was neither irrational nor devoid of strategic reasoning, attributing the disastrous outcome instead to fate or to divine will beyond human comprehension.
By contrast, Nikephoros Bryennios the Younger stands alone in preserving details of the military council held at Theodosiopolis, where the decision to advance toward Manzikert was confirmed over a more cautious alternative strategy. That alternative—remaining on the defensive and compelling the Seljuks to approach—was advocated by Trachaneiotes and Nikephoros Bryennios the Elder.
A map showing Byzantine territory (purple), Byzantine offensives (red) and Turkish offensives (green). Image: Wikimedia Commons
At this point, a fundamental issue must be examined more closely. Once Romanos understood that he was facing not dispersed Turkmen bands but the main Seljuk field army at Manzikert, he reportedly sent messengers to Roussel and Trachaneiotes with instructions to return. The decisive question, however, is whether those instructions ever reached their destination. There is no concrete evidence that they did.
On the contrary, the territory between Manzikert and Khliat was almost certainly dominated by hostile mounted forces, rendering the transmission of messages extremely dangerous, if not altogether unfeasible. In such conditions, the failure of the Khliat detachment to rejoin the emperor’s army may be better explained by the operational realities of warfare in enemy-controlled terrain than by accusations of cowardice or disloyalty.
A passage in Attaleiates has often been treated as decisive proof that Trachaneiotes did, in fact, receive Romanos’ recall order and chose to disregard it. Attaleiates states that the emperor was unaware that Trachaneiotes, having learned of the sultan’s attack against the imperial army, had withdrawn with all his troops through Mesopotamia until reaching Byzantine territory. Yet this information need not have derived from a formal recall order. It is equally plausible that Trachaneiotes’ own scouts, attempting to re-establish contact with the emperor—or even to warn him—had independently discovered the Seljuk advance and reported it to their commander.
Likewise, Attaleiates’ observation that Romanos still hoped the Khliat detachment would not be delayed beyond the following day has led some scholars to conclude that the emperor’s orders must have been successfully delivered. Such an inference, however, is unwarranted. This expectation reflects nothing more than Romanos’ wishful thinking and a dangerously optimistic reading of the tactical situation, rather than firm evidence of functioning communications.
Indeed, the assumption that imperial messengers could evade Turkmen cavalry not once but repeatedly strains credulity. The journey from Manzikert to Khliat—approximately sixty kilometres—would already have been hazardous; to then return from Khliat to Manzikert, a distance of roughly 120 kilometres, through terrain saturated with enemy cavalry, would have required extraordinary good fortune. Romanos’ belief that such exchanges could occur uninterrupted suggests a serious misjudgement of the operational environment.
It is far more likely that the Byzantine commanders near Khliat quickly grasped that a substantial Seljuk force had arrived from the south and that Turkish cavalry had effectively flooded the region between Khliat and Manzikert, severing communications with the emperor. Under these circumstances, the most practical course of action for the Byzantine units around Khliat would have been a withdrawal westward along the Arsanias River valley into Upper Mesopotamia, as reported by the Greek sources. Such a retreat would logically have proceeded first to the fortified city of Muş, approximately 120 kilometers to the west, and then onward to the relative security of Melitene.
An additional perspective is provided by Ibn al-ʿAdīm, the Aleppine jurist and historian (1192–1262), who records that, following the Byzantine defeat, only the contingent besieging Akhlat remained intact and that, upon learning of the disaster, they abandoned the siege and withdrew in fear. This testimony may suggest that Trachaneiotes’ detachment remained in the vicinity of Khliat and continued operations there, likely in accordance with its original orders.
Taken together, this evidence casts serious doubt on the traditional portrayal of Roussel and Trachaneiotes as traitors or convenient scapegoats for the catastrophe of 26 August 1071. Such judgements overlook the constraints imposed by unreliable intelligence, severed communications, and an overwhelmingly hostile tactical environment. In light of these factors, the decisions made by the two commanders appear far more defensible than later narratives have allowed.
A Failed Manoeuvre?
Finally, attention must be directed to the crucial failure of the Byzantine support line to reinforce Romanos’ main formation as it struggled against the Seljuk tactic of feigned retreat. On the afternoon of Friday, 26 August—the day of the battle—Alp Arslan launched his decisive counterattack shortly after the conclusion of the afternoon prayer. This moment may coincide with the point at which the sultan was informed by his scouts of a sudden reversal affecting the Byzantine army, prompting urgent advice that he seize the opportunity and strike.
What, then, constituted this “sudden misfortune”? Earlier that afternoon, Romanos had issued a pivotal command: the imperial standard was to be turned, signalling the army to withdraw toward the camp. The emperor had likely become aware that his forces had advanced too far from their fortified base, which was sparsely defended and exposed. At the same time, the cohesion of his battle line appears to have deteriorated, with widening gaps opening between the centre and the wings. Fatigue, thirst, and dwindling supplies would have further undermined the effectiveness of the troops.
A 15th-century depiction of the Battle of Manzikert – Bibliothèque nationale de France, MS Français 226, Fol. 265
The signal to disengage, however, proved disastrous. Some units on the flanks evidently misread the movement of the banner as evidence that the centre had collapsed or that the emperor had been killed. Leaving aside accusations of treachery, the inability of these formations to execute a controlled withdrawal strongly suggests deficiencies in morale, training, and discipline within the imperial army. This momentary breakdown was precisely the opening Alp Arslan had anticipated, and his reconnaissance units were quick to detect and relay the opportunity.
Our sources do not clearly indicate which flank collapsed first. Nikephoros Bryennios the Younger alone records that the Seljuks, having turned back with force and raised their battle cries, struck the Byzantine line and drove the right wing into flight. This wing included Cappadocian and other Anatolian contingents. Bryennios does not, however, connect this rout explicitly with the misunderstood withdrawal signal.
Under such circumstances, the role of the rear guard became critical. As the right wing disintegrated and streamed back toward the camp, the rear guard—rather than moving forward to stabilize the situation—accelerated its own withdrawal. Pro-Romanos historians, most notably Attaleiates and those who followed him such as Skylitzes and Zonaras, interpreted this action as a deliberate betrayal by the rearguard commander, Andronikos Doukas. Attaleiates’ formulation is revealingly vague, claiming that “many report” that someone hostile to the emperor circulated false news of his death among the troops.
It is important to note, however, that Attaleiates, though present on the campaign in his capacity as military judge (kritēs tou stratopedou), remained in the camp during the battle itself. He was neither a combatant nor an eyewitness to events on the field, and there is no indication that he possessed a clear vantage point from which to observe the unfolding engagement. His testimony must therefore be treated with caution.
By contrast, Nikephoros Bryennios the Younger—often sympathetic to the Doukai—attributes the collapse not to treason but to the failure of the rear guard to support the emperor’s division, which allowed the Seljuks to encircle Romanos and attack him from all sides. Notably, Bryennios’ narrative contains no accusation of deliberate betrayal by Andronikos Doukas.
A more plausible explanation is that, over the course of the afternoon, a considerable distance had opened between the rear guard and the main battle line, rendering timely intervention impossible. A useful parallel may be drawn with the Battle of Dyrrachium in 1081, where the Varangian Guard advanced far ahead of the main Byzantine line and was subsequently annihilated when Alexios I failed to provide adequate support. As Anna Komnene carefully obscures her father’s misjudgment at Dyrrachium, so too might we expect pro-Romanos sources to seek scapegoats for the catastrophe at Manzikert.
From this perspective, the branding of Trachaneiotes, Roussel, and Andronikos Doukas as traitors appears less a reflection of their actions than a rhetorical strategy designed to deflect responsibility from the emperor himself. Romanos misread Seljuk intentions following Alp Arslan’s withdrawal from Aleppo; he underestimated the threat posed by the garrisons of Manzikert and Khliat, leading him to divide his army at a critical juncture; and he overestimated the ability of his forces to carry out a complex and hazardous manoeuvre such as a coordinated withdrawal under pressure. Ultimately, the interpretation of Manzikert depends largely on perspective—and on whom later narratives chose to blame for one of Byzantium’s most consequential defeats.
George Theotokis received his PhD in History from the University of Glasgow (2010), specializing in medieval military history. He has published numerous articles and monographs on the history of warfare in Europe and the Mediterranean in the medieval and early modern periods, including The Art of War in Byzantium (Arc Humanities Press, 2024). He works as an Assistant Professor of European history at Ibn Haldun University in Istanbul.
Was the Byzantine defeat at Manzikert the result of treachery on the battlefield—or a series of strategic miscalculations? A closer reading of the medieval chroniclers suggests that the search for betrayal began soon after the battle itself.
By George Theotokis
Few figures in Byzantine military history have been so harshly condemned in the aftermath of a single defeat as Josef Trachaneiotes, Roussel of Bailleul, and Andronikos Doukas. According to the dominant narrative, the first two disobeyed their emperor by failing to return and support him at a critical moment, while the third allegedly abandoned Romanos on the battlefield, ordering a withdrawal precisely when his intervention was needed to prevent the encirclement and collapse of the imperial line. Yet the credibility of these accusations is far from certain. Much depends on which sources one chooses to trust.
The Prelude to the Battle
The Battle of Manzikert marked the geopolitical high point of a prolonged contest between the Byzantine Empire and the Seljuk Turks over dominance in eastern Asia Minor and northern Syria. Despite its later significance in Byzantine–Seljuk relations, many scholars argue that Sultan Alp Arslan’s chief strategic concern was not Byzantium itself, but rather the Shiʿa Fatimid Caliphate. From this perspective, the Turkish operations of 1070–1071 that culminated at Manzikert were intended primarily to prepare the ground for a future campaign against Egypt.
To secure his position in Syria and protect the flanks of this anticipated advance, Alp Arslan moved to seize key fortresses, including Artchesh and Manzikert. At the same time, sustained Turkish incursions over preceding years had rendered the territory between Theodosiopolis and Manzikert largely depopulated and devastated. These developments ensured that the Seljuk sultan could operate in Syria without immediate threats to his rear.
Alp Arslan’s winter campaign of 1070–1071 in Syria can therefore be regarded as a strategic success, as it considerably expanded his authority across northern Syria and the Upper Euphrates region. Nevertheless, his focus soon shifted away from Syria due to developments unfolding further north.
Insight into the circumstances that led Emperor Romanos IV to undertake his ill-fated expedition toward Lake Van is provided by the historian Nikephoros Bryennios the Younger. He recounts that during a military council convened while the imperial army was marching between Kaisarea and Sebasteia, both the Georgian magistros Joseph Trachaniotes and Nikephoros Bryennios the Elder advised against advancing into Turkish-held territory. Instead, they urged the emperor to remain at Theodosiopolis, thereby compelling Alp Arslan to engage the Byzantines on ground more favorable to Roman forces. Their counsel, however, was overridden by more aggressive voices within the council. As a result, the advance toward Khliat and Manzikert was approved—a decision that Bryennios later described with evident dismay and confusion.
Dividing his Forces: Overconfidence or Lack of Intelligence?
Diogenes’ imperial army remained stationed at Theodosiopolis until approximately mid-summer 1071. In late July or early August, however, the emperor’s increasing confidence prompted a decisive strategic miscalculation. Upon departing Theodosiopolis and advancing southeast toward Manzikert via Xnis/Xinus (modern Hınıs in the province of Erzurum), a march of roughly 230 kilometers, Romanos chose to divide his army. According to Michael Attaleiates, this decision was based on the emperor’s belief that the Turkish and Iranian Daylamite garrison defending Manzikert posed little serious resistance.
Additional sources indicate that Romanos detached contingents of Frankish and other mercenary forces, placing them under the command of Roussel de Bailleul and the magistros Joseph Trachaneiotes, and ordered them to advance toward the strategically vital fortress-city of Khliat. Control of Khliat was essential, as it dominated the western access routes to the Lake Van basin.
Two points are especially significant. First, the force led by Roussel and Trachaneiotes was composed of elite troops and was, in fact, substantially larger than the contingent the emperor retained under his direct command. Second, Romanos intended to capture Manzikert with his own forces and did not plan to reunite with the detached army unless an extreme emergency arose. Nor did he envisage advancing on Khliat until Manzikert had first been brought back under imperial authority.
In my view, Romanos anticipated a rapid and straightforward operation that would result in the swift recovery of both Manzikert and Khliat. Events, however, would soon demonstrate how gravely mistaken this assumption was.
In order to contextualize Romanos’ strategic decisions, it is essential to examine the quality of Byzantine intelligence and, in particular, the role played by Leon Dabatenos. Dabatenos belonged to a prominent Armenian family, and a figure of this name is attested as vestarches and doux (or katepano) of Edessa between roughly 1071 and 1077. He appears to have assumed this position shortly after the Turkish siege of Edessa during the winter of 1070–1071.
Nikephoros Bryennios the Younger is our sole source for an official intelligence report said to have been delivered to Romanos by Dabatenos while the emperor was likely still stationed at Theodosiopolis with the main body of his army. According to this account, the report claimed that the Seljuk sultan, having learned of Romanos’ campaign against Khliat and Manzikert and fearing the strength of the imperial forces, had withdrawn from Persia toward Baghdad. Bryennios emphasizes the consequences of this information: trusting its accuracy, the emperor divided his army, retaining one portion near Manzikert while dispatching the other toward Khliat.
This testimony should be considered alongside a striking remark by the Muslim preacher and historian Sibṭ Ibn al-Jawzī (1185–1256), who notes that a Byzantine envoy returned to the imperial camp in high spirits and that this report reinforced Romanos’ determination to pursue Alp Arslan and engage him in battle. The parallel accounts raise an important question: are Bryennios and Ibn al-Jawzī referring to the same individual responsible for supplying intelligence to the emperor?
It would have been entirely reasonable for Romanos to request detailed information from the doux of Edessa concerning the size, composition, and morale of the Seljuk army. The crucial issue, however, lies in the reliability of that intelligence and the extent to which Leon Dabatenos could be trusted. A revealing passage from Matthew of Edessa’s Chronicle casts a long shadow over Dabatenos’ conduct. Matthew records that the doux of Edessa supplied Alp Arslan with horses, pack animals, and provisions, allowing the sultan to pass safely through the territory of Edessa before continuing eastward toward Mount Lesun.
Such cooperation strongly suggests that the governor of Edessa had, at least tacitly, recognized Seljuk overlordship. This naturally raises doubts about the credibility of the intelligence reportedly sent to Romanos regarding the Seljuk army’s intentions and disposition. Was Dabatenos deliberately providing misleading information, encouraging the belief that Alp Arslan would remain on the defensive in Iraq rather than confront the emperor in Armenia? Absolute certainty is impossible, but the circumstantial evidence is troubling—particularly when one considers that Dabatenos retained his position throughout the reign of Michael VII, remaining in office until around 1077.
Turning to the emperor’s decision to summon back the forces previously detached to Khliat under the command of Roussel and Trachaneiotes, this development occurred on Thursday, 25 August, the day before the Battle of Manzikert. Both Attaleiates and Skylitzes record that Romanos initially postponed engagement but ultimately abandoned any expectation of reinforcement from the Khliat detachment, having concluded that some serious impediment was preventing their return.
At the same time, the emperor appears to have believed that these elite units were merely delayed rather than having withdrawn altogether from the area. This assumption explains his resolve to offer battle on the following day, Friday, 26 August, relying on the troops immediately at his disposal, while still entertaining some hope that the absent forces might arrive shortly thereafter.
Attaleiates further notes that Romanos was unaware that Trachaneiotes, upon learning of the sultan’s advance against the imperial army, had led his entire force in retreat through Mesopotamia until reaching Byzantine territory. The historian characterizes this withdrawal in harsh terms, accusing Trachaneiotes of disgracefully abandoning both his sovereign and his obligations. Yet this raises an important question: how justified is such a severe judgment of the conduct of Roussel and Trachaneiotes?
Given the ideological divisions within the Greek narrative tradition—often categorized along pro- and anti-Romanos lines—Attaleiates and Skylitzes tend to mitigate the emperor’s responsibility for having detached nearly half of his army to Khliat. They argue that the division of forces was neither irrational nor devoid of strategic reasoning, attributing the disastrous outcome instead to fate or to divine will beyond human comprehension.
By contrast, Nikephoros Bryennios the Younger stands alone in preserving details of the military council held at Theodosiopolis, where the decision to advance toward Manzikert was confirmed over a more cautious alternative strategy. That alternative—remaining on the defensive and compelling the Seljuks to approach—was advocated by Trachaneiotes and Nikephoros Bryennios the Elder.
At this point, a fundamental issue must be examined more closely. Once Romanos understood that he was facing not dispersed Turkmen bands but the main Seljuk field army at Manzikert, he reportedly sent messengers to Roussel and Trachaneiotes with instructions to return. The decisive question, however, is whether those instructions ever reached their destination. There is no concrete evidence that they did.
On the contrary, the territory between Manzikert and Khliat was almost certainly dominated by hostile mounted forces, rendering the transmission of messages extremely dangerous, if not altogether unfeasible. In such conditions, the failure of the Khliat detachment to rejoin the emperor’s army may be better explained by the operational realities of warfare in enemy-controlled terrain than by accusations of cowardice or disloyalty.
A passage in Attaleiates has often been treated as decisive proof that Trachaneiotes did, in fact, receive Romanos’ recall order and chose to disregard it. Attaleiates states that the emperor was unaware that Trachaneiotes, having learned of the sultan’s attack against the imperial army, had withdrawn with all his troops through Mesopotamia until reaching Byzantine territory. Yet this information need not have derived from a formal recall order. It is equally plausible that Trachaneiotes’ own scouts, attempting to re-establish contact with the emperor—or even to warn him—had independently discovered the Seljuk advance and reported it to their commander.
Likewise, Attaleiates’ observation that Romanos still hoped the Khliat detachment would not be delayed beyond the following day has led some scholars to conclude that the emperor’s orders must have been successfully delivered. Such an inference, however, is unwarranted. This expectation reflects nothing more than Romanos’ wishful thinking and a dangerously optimistic reading of the tactical situation, rather than firm evidence of functioning communications.
Indeed, the assumption that imperial messengers could evade Turkmen cavalry not once but repeatedly strains credulity. The journey from Manzikert to Khliat—approximately sixty kilometres—would already have been hazardous; to then return from Khliat to Manzikert, a distance of roughly 120 kilometres, through terrain saturated with enemy cavalry, would have required extraordinary good fortune. Romanos’ belief that such exchanges could occur uninterrupted suggests a serious misjudgement of the operational environment.
It is far more likely that the Byzantine commanders near Khliat quickly grasped that a substantial Seljuk force had arrived from the south and that Turkish cavalry had effectively flooded the region between Khliat and Manzikert, severing communications with the emperor. Under these circumstances, the most practical course of action for the Byzantine units around Khliat would have been a withdrawal westward along the Arsanias River valley into Upper Mesopotamia, as reported by the Greek sources. Such a retreat would logically have proceeded first to the fortified city of Muş, approximately 120 kilometers to the west, and then onward to the relative security of Melitene.
An additional perspective is provided by Ibn al-ʿAdīm, the Aleppine jurist and historian (1192–1262), who records that, following the Byzantine defeat, only the contingent besieging Akhlat remained intact and that, upon learning of the disaster, they abandoned the siege and withdrew in fear. This testimony may suggest that Trachaneiotes’ detachment remained in the vicinity of Khliat and continued operations there, likely in accordance with its original orders.
Taken together, this evidence casts serious doubt on the traditional portrayal of Roussel and Trachaneiotes as traitors or convenient scapegoats for the catastrophe of 26 August 1071. Such judgements overlook the constraints imposed by unreliable intelligence, severed communications, and an overwhelmingly hostile tactical environment. In light of these factors, the decisions made by the two commanders appear far more defensible than later narratives have allowed.
A Failed Manoeuvre?
Finally, attention must be directed to the crucial failure of the Byzantine support line to reinforce Romanos’ main formation as it struggled against the Seljuk tactic of feigned retreat. On the afternoon of Friday, 26 August—the day of the battle—Alp Arslan launched his decisive counterattack shortly after the conclusion of the afternoon prayer. This moment may coincide with the point at which the sultan was informed by his scouts of a sudden reversal affecting the Byzantine army, prompting urgent advice that he seize the opportunity and strike.
What, then, constituted this “sudden misfortune”? Earlier that afternoon, Romanos had issued a pivotal command: the imperial standard was to be turned, signalling the army to withdraw toward the camp. The emperor had likely become aware that his forces had advanced too far from their fortified base, which was sparsely defended and exposed. At the same time, the cohesion of his battle line appears to have deteriorated, with widening gaps opening between the centre and the wings. Fatigue, thirst, and dwindling supplies would have further undermined the effectiveness of the troops.
The signal to disengage, however, proved disastrous. Some units on the flanks evidently misread the movement of the banner as evidence that the centre had collapsed or that the emperor had been killed. Leaving aside accusations of treachery, the inability of these formations to execute a controlled withdrawal strongly suggests deficiencies in morale, training, and discipline within the imperial army. This momentary breakdown was precisely the opening Alp Arslan had anticipated, and his reconnaissance units were quick to detect and relay the opportunity.
Our sources do not clearly indicate which flank collapsed first. Nikephoros Bryennios the Younger alone records that the Seljuks, having turned back with force and raised their battle cries, struck the Byzantine line and drove the right wing into flight. This wing included Cappadocian and other Anatolian contingents. Bryennios does not, however, connect this rout explicitly with the misunderstood withdrawal signal.
Under such circumstances, the role of the rear guard became critical. As the right wing disintegrated and streamed back toward the camp, the rear guard—rather than moving forward to stabilize the situation—accelerated its own withdrawal. Pro-Romanos historians, most notably Attaleiates and those who followed him such as Skylitzes and Zonaras, interpreted this action as a deliberate betrayal by the rearguard commander, Andronikos Doukas. Attaleiates’ formulation is revealingly vague, claiming that “many report” that someone hostile to the emperor circulated false news of his death among the troops.
It is important to note, however, that Attaleiates, though present on the campaign in his capacity as military judge (kritēs tou stratopedou), remained in the camp during the battle itself. He was neither a combatant nor an eyewitness to events on the field, and there is no indication that he possessed a clear vantage point from which to observe the unfolding engagement. His testimony must therefore be treated with caution.
By contrast, Nikephoros Bryennios the Younger—often sympathetic to the Doukai—attributes the collapse not to treason but to the failure of the rear guard to support the emperor’s division, which allowed the Seljuks to encircle Romanos and attack him from all sides. Notably, Bryennios’ narrative contains no accusation of deliberate betrayal by Andronikos Doukas.
A more plausible explanation is that, over the course of the afternoon, a considerable distance had opened between the rear guard and the main battle line, rendering timely intervention impossible. A useful parallel may be drawn with the Battle of Dyrrachium in 1081, where the Varangian Guard advanced far ahead of the main Byzantine line and was subsequently annihilated when Alexios I failed to provide adequate support. As Anna Komnene carefully obscures her father’s misjudgment at Dyrrachium, so too might we expect pro-Romanos sources to seek scapegoats for the catastrophe at Manzikert.
From this perspective, the branding of Trachaneiotes, Roussel, and Andronikos Doukas as traitors appears less a reflection of their actions than a rhetorical strategy designed to deflect responsibility from the emperor himself. Romanos misread Seljuk intentions following Alp Arslan’s withdrawal from Aleppo; he underestimated the threat posed by the garrisons of Manzikert and Khliat, leading him to divide his army at a critical juncture; and he overestimated the ability of his forces to carry out a complex and hazardous manoeuvre such as a coordinated withdrawal under pressure. Ultimately, the interpretation of Manzikert depends largely on perspective—and on whom later narratives chose to blame for one of Byzantium’s most consequential defeats.
George Theotokis received his PhD in History from the University of Glasgow (2010), specializing in medieval military history. He has published numerous articles and monographs on the history of warfare in Europe and the Mediterranean in the medieval and early modern periods, including The Art of War in Byzantium (Arc Humanities Press, 2024). He works as an Assistant Professor of European history at Ibn Haldun University in Istanbul.
Click here to read more from Georgios Theotokis
Subscribe to Medievalverse
Related Posts