Few people have been vilified in such a way after the disastrous outcome of a decisive battle other than Trachaniotes, Roussel of Bailleul, and Andronikos Doukas. The first two failed to follow their lord’s orders and come to his aid, while the latter deserted the emperor and ordered a withdrawal from the battlefield when he should have supported the Byzantine front line to avert its envelopment and subsequent collapse. But how credible are these accounts? It depends on who we read.
Because Michael Attaleiates was left behind in the fortified Byzantine camp, he was able to offer a highly emotional personal account of the unfolding disaster, as he experienced it through what he saw and what he heard from the soldiers who fled from the battlefield. But how objective is Attaleiates’ account in terms of his personal bias and sympathies compared to Nikephoros Bryennios (‘the Younger’)?
Advertisement
The Prelude to the Battle
The Battle of Manzikert was the geo-political climax of the struggle for control of key cities and towns in eastern Asia Minor and northern Syria between two major powers, the Byzantine Empire and the Seljuk Turks. However, it is widely believed that the Seljuk sultan Alp Arslan’s primary strategic target was not the Byzantine Empire but the Shi’a Fatimid Caliphate. For that reason, the Turkish campaign of 1070/71 that preceded the fateful clash at Manzikert in 1071, aimed at laying the foundation for a major future invasion of Egypt, and that meant the targeting of Aleppo to force the emir of the city to renounce his allegiance to the Fatimid Caliph al-Mustansir.
Therefore, Alp Arslan wished to have his flanks covered for the critical campaign in Syria, and for that reason he conquered the fortress towns of Artchesh and Manzikert, on the north shore of Lake Van, sometime in the early autumn of 1070. Attaleiates reports that Manzikert was, subsequently, garrisoned by a strong contingent of Daylami fighters from northern Iran, while the incessant Turkish raids of the previous years had transformed the area stretching between Theodosiopolis and Manzikert into ἀοίκητον και ἠρημωμένη.
Advertisement
Alp Arslan then marched towards the Upper Tigris Valley between Amida (Diyār Bakr) and Mayyafariqin, the heartland of the Marwānid emirate, on his way to his first objective of Byzantine-ruled Edessa, a strategic city that had to be taken before he could march on to Aleppo. The Turkish army besieged Edessa in mid-winter (of 1070/71), failing to capture it due to the stubborn resistance of the formidable Byzantine garrison, and the competent command of the proedros and doux of Edessa, Basil Alousianos.
Alp Arslan decided to withdraw from Edessa on 19 January 1071, marching his army across the Euphrates and against the much larger prize of Aleppo. Eventually, just a few weeks after the start of the siege (perhaps on 4 May 1071), the emir of Aleppo, Mahmūd bin Naşr, submitted to Alp Arslan and, handing over to him the keys of the city, was confirmed as governor of the city by Alp Arslan.
Alp Arslan’s campaign in Syria in the winter of 1070/71 should be considered a strategic success because it significantly boosted his influence over the regions of northern Syria and the Upper Euphrates by imposing Seljuk suzerainty over the Mirdasid emir of Aleppo while putting further pressure on the remaining imperial outposts of Edessa and Antioch. The sultan also performed a necessary show of force in the regions of his other vassals, the Marwānids and the Uqaylids, and before all of these had taken place, he had inflicted yet another significant blow to the defences of the Byzantine empire in Vaasparakania by capturing Manzikert and Artchesh. However, Alp Arslan’s attention would be drawn away from Syria by events taking place farther north.
The historian Nikephoros Bryennios (‘the Younger’) provides an important view of the background behind Romanos’ fateful decision to launch a major strike towards Lake Van. He reports that in the military council that took place sometime during the march of the imperial army between Kaisarea and Sebasteia both the respected Georgian magistros Joseph Tarchaniotes and Nikephoros Bryennios (‘the Elder’) sought to convince the emperor not to take to the field against the Turks but, rather, to wait for them in Theodosiopolis so that Alp-Arslan would “be forced to clash with the Romans where it is in their [Romans] interest to make war. But they [Trachaneiotes and Bryennios] were under the impression that they were talking to a deaf man … more attention was paid [by Romanos] to the flatterers and not to those who advised the right thing.” Unsurprisingly, the more bellicose officers in the council prevailed and the march towards Khliat and Manzikert was confirmed, much to Bryennios’ bewilderment and consternation.
Advertisement
15th-century French miniature depicting the combatants in contemporary Western European armour – Bibliothèque Nationale MS Fr. 226
Dividing his forces: overconfidence or lack of intelligence?
The imperial army under Diogenes was based in Theodosiopolis until, probably, the high summer of 1071. Sometime in late July or early August, his growing overconfidence led the emperor to make a critical strategic error. After leaving Theodosiopolis to march south-east to Manzikert via Xnis/Xinus (modern Hınıs, Erzurum province), a distance of around 230kms, the emperor decided to divide his forces because, according to Attaleiates, he was under the impression that the Turks and Iranian Daylami defenders of Manzikert would be easily overcome.
Other sources also account that the emperor dispatched separate detachments of troops to march to the strategic fortress-city of Khliat that controlled the western approaches to the region of Lake Van. These detachments comprised of “Uze mercenaries and the Franks under Rouselios … [sent] towards Khliat to forage for provisions,” and because the emperor “considered them [Turkish defenders] of no great importance since they were few in number, he detached another not insignificant portion of his army [to march to Khliat] and placed it under the command of the magistros Joseph Trachaneiotes.”
There are two crucial points we need to emphasize here; first, that the detachment under the Norman Roussel de Bailleul and the magistros Trachaneiotes “consisted of select troops, difficult to withstand, taking the initiative in close combat and other types of battle and ready to face danger, and far more numerous than the soldiers retained by the emperor.” Second, the emperor planned to take Manzikert with the units he kept under his command, but he did not plan to join forces with Roussel’s and Trachaneiotes’ detachments unless there was a dire emergency and, certainly, not to march against Khliat before “Manzikert was restored to Roman rule.” I believe that our sources make it clear that the emperor was planning for a swift campaign to capture Manzikert and Khliat, and, perhaps, even subdue the remaining fortress-towns on the northern shore of Lake Van, replace their garrisons with imperial troops and head back home before there was any reaction from the Seljuk leadership. Romanos was in for a shock!
Advertisement
To understand the background behind Romanos’ strategic decision, we need to refer to the decisive issue of Byzantine intelligence. When it comes to the supply of intelligence to the emperor in this critical juncture of the campaign, with the imperial army stationed at Theodosiopolis in the summer of 1071 and in need of accurate information about the Turkish intentions before marching against the main target of Khliat and Manzikert, the role of Leon Dabatenos must be studied more closely. Leon Dabatenos (Armenian: Lewon Dawatanos) came from a well-known aristocratic family of Armenian origin that found itself on the rise around the middle of the eleventh century but was already in decline in the twelfth century.
A certain Dabatenos was katepano or doux of Edessa and died during the siege of Amida in 1062, while another Leon Dabatenos is identified as vestarches and doux/katepano of Edessa, who held the office ca. 1071-1077. It is interesting to note that the latter Dabatenos had replaced the doux of Edessa Basil Alousianos who was responsible for the defence of the city from the Seljuks in the winter of 1070/71, with Alousianos disappearing from the record shortly after the Turkish withdrawal from Edessa. This change of leadership at such a decisive moment is intriguing, but I was unable to identify the reason for Alousianos’ dismissal.
Image by Vartolu3566 and Bakayna / Wikimedia Commons
Moreover, Nikephoros Bryennios (‘the Younger’) is the only source to account for an official report that reached Diogenes by Dabatenos while the emperor was probably still in Theodosiopolis with the bulk of his army, according to which “the sultan, it read, learning about the campaign [of Romanos to Khliat and Manzikert] and afraid of the power [of Romanos’ army], left Persia for Babylon [Baghdad];” but what is more important about the outcome of the campaign is what immediately follows, “the emperor, putting his trust on these words, divided his army in two: he kept one in place [Manzikert], and he dispatched the other to Khliat.”
The aforementioned report should be combined with another intriguing comment by Sibṭ Ibn al-Jawzī (b. 581/1185 – d. 654/1256), a renowned Muslim preacher and historian whose account of the campaign and battle of Manzikert offers significant details about Alp Arslan’s activities leading up to the battle itself, who writes that “the envoy of Byzantium returned [to the imperial camp] rejoicing to his master [Romanos] and that strengthened the resolve of the king of Byzantium to pursue him [Alp Arslan] and to fight him.”
Advertisement
Do both Bryennios and al-Jawzī basically describe the same person compiling an intelligence report for the emperor that was delivered either in letter (Bryennios) or in person (al-Jawzī)? Was Leon Dabatenos the head of the embassy dispatched by the emperor to Alp Arslan while the latter was in the vicinity of Aleppo in the early spring of 1071? There is some food for thought here, but the sources raise more questions than provide answers. Nonetheless, it is very reasonable for the emperor to have demanded intelligence from the doux of Edessa about the size, units and morale of Alp Arslan’s army; the critical point, here, is how far could Romanos trust Leon Dabatenos.
When it comes to the matter of allegiance, technically, Dabatenos’ was with the Byzantine emperor, but an extract from Matthew of Edessa’s Chronicle may suggest otherwise:
The doux [of Edessa], who resided in the city [of Edessa], provided him [Alp Arslan] with horses, mules, and victuals. Taking these, the sultan passed through the confines of Edessa, unharmed, and went in an easterly direction towards the mountain called Lesun.
Although he is not mentioned by name, what Matthew suggests here is that the doux of Edessa may have acknowledged the suzerainty of the Seljuk sultan, thus having to provide him with the necessary supplies for his march north from Syria back to Iran. But what about the report dispatched to Emperor Romanos on the strength and size of the Seljuk army? Was Dabatenos deliberately attempting to mislead the emperor with false intelligence? We will never know with certainty, but the evidence is damning, especially if we consider that he remained in office throughout Michael VII’s reign until ca. 1077.
Finally, we have Attaleiates’ comment that “spies had already reported to him [Romanos] that the sultan was in a hurry to return to Persia.” If we couple this with the report dispatched by Leon Dabatenos, both reports would, certainly, have boosted the emperor’s confidence and conviction that, following his strategic withdrawal from Syria, Alp Arslan would adopt a defensive strategy in Iraq and would not attempt to march against him in Armenia. The failure of the intelligence reports is critical here!
Moving on to the emperor’s decision to recall the units that had been dispatched to Khliat under Roussel and Trachaneiotes. It is only Bryennios (‘the Younger’) who reports of a military council summoned by the emperor on the eve of the battle, Thursday, August 25, to assess the situation and plan accordingly, while, again, he highlights the polarization within the council between those who wished to go out immediately and face the main Seljuk army in a pitched battle, and those who urged caution and to await the arrival of those troops recalled from Khliat who would, of course, never arrive. Attaleiates and Skylitzes fail to mention any military council summoned on Thursday morning, adding that “he [Diogenes] put off the battle. But then he gave up on receiving any help from them [units dispatched to Khliat], for he suspected that some powerful obstacle was preventing their arrival.”
However, Romanos was still under the impression that the elite units in Khliat were, simply, delayed rather than withdrawn from the region, hence the following statement of the emperor’s intention that “[the emperor] decided to do battle against the enemy on the following day [Friday, 26 August] as bravely as he could with the soldiers he had at hand. He had, moreover, some hope that the others [units dispatched to Khliat] would not be delayed beyond the next day,” with Attaleiates adding that “he [emperor] was unaware that their general [Trachaneiotes], upon learning of the sultan’s attack against the emperor, had taken all his soldiers and shamefully retreated through Mesopotamia until he came into Roman territory. The coward took no account of his lord or even his duty.” But how fair is this assessment of the actions of Roussel and Trachaneiotes?
With the Greek sources being categorized as pro- and anti- Romanos, Attaleiates and Skylitzes are inclined to excuse the emperor’s decision to dispatch almost half of his troops to Khliat. They emphasize that “dividing his [Romanos] forces was not irrational or lacking in strategic calculation. It was fate, or rather the wrath or design of God beyond our understanding.” Unsurprisingly, Bryennios provides a different view of the background behind this fateful decision, being the only one to report the decisions of the war council in Theodosiopolis that confirmed the plan to advance to Manzikert instead of the -more cautious- strategy of having to wait for the Seljuks to come to them, a strategy that was supported by Trachaneiotes and Nikephoros Bryennios (‘the Elder’). But there is a surprising inference by Skylitzes in that Trachaneiotes was, indeed, privy to the plan already formulated by Andronikos Doukas to betray the emperor at the first opportune moment, and he even managed to persuade Roussel to join forces with him and retreat to Roman territory “through Mesopotamia” instead of rushing back to Manzikert to reinforce Diogenes, as he intended to do.
There is a critical question we need to ask at this point; if the emperor dispatched messengers to Trachaneiotes and Roussel after he realised that his army was facing the main Seljuk army instead of units of irregular Turkmen, did they manage to deliver the emperor’s orders for a recall back to Manzikert? We have no evidence to suggest they did, as the area between Manzikert and Khliat would have been swarming with enemy (mounted) troops.
Earlier, I quoted a key statement by Attaleiates that has been considered sufficient evidence to assume that Trachaneiotes did, indeed, receive the order for a recall by Romanos and, thus, chose to ignore it; let me quote the sentence again: “for he [Romanos] was unaware that their general [Trachaneiotes], upon learning of the sultan’s attack against the emperor, had taken all of his soldiers and shamefully retreated through Mesopotamia until he came into Roman territory.”
In my view, however, this comment does not necessarily mean that Trachaneiotes received the news of the unfolding crisis at Manzikert from Romanos’ messengers. Rather, the intelligence could very well have come from his own scouts who were trying to re-establish contact with the emperor and, possibly, warn him!
Furthermore, Attaleiates’ comment that “He [Romanos] had, moreover, some hope that the others [units dispatched to Khliat] would not be delayed beyond the next day,” has led scholars to believe that “the focus on Diogenes’s anxious expectation for news from Khliat means that his orders had reached Trachaneiotes safely.” No, it does not! It is, simply, wishful thinking on behalf of the emperor, and a catastrophic assessment of the tactical situation.
Additionally, Romanos’ expectation that the messengers would have evaded the roaming Turkmen once (from Manzikert to Khliat, distance: 60 kilometres) and delivered the recall order to Trachaneiotes and, then, evade them twice (Khliat to Manzikert, again, distance: 120 kilometres) by returning to the imperial camp to report to the emperor, all the while the area between Manzikert and Khliat was swarming with enemy cavalry was, to say the least, over-optimistic. We should remind ourselves that these comments come from the pen of a pro-Romanos source, Attaleiates, who is looking for scapegoats about seven or eight years after the battle, and the same source who is ready to accuse Andronikos Doukas for treachery although, as I will show further down, there are more credible causes for the collapse of the imperial army on Friday, August 26.
The broken and rocky terrain at the battlefield of Manzikert. Photo by Georgios Theotokis
Another critical question is how to make sense of the strategic reaction of Trachaneiotes and Roussel to retreat to “Mesopotamia” in the face of the rapidly deteriorating situation in the vicinity of Khliat that pitted them against strong(er) Turkish reinforcements dispatched by Alp Arslan exactly to prevent these two armies from marching back to reinforce their emperor? Following the reinforcements dispatched by Alp Arslan to the garrison of Khliat, Trachaneiotes’ army would have launched a siege of the city according to the orders of the emperor, as it is reported by our Muslim sources, yet we are at a loss as to what happened after that, because the focus shifts to the siege of Manzikert. We can only speculate that the Byzantine commanders would have realised that they were up against a strong Seljuk army that had arrived from the south, with Turkish cavalry effectively ‘flooding’ the region between Khliat and Manzikert and, consequently, cutting communications between Trachaneiotes’ army and the emperor.
Assuming that this interpretation of events is correct, then the most realistic option for the beleaguered Byzantine units around Khliat would have been to retreat west along the Arsanias River valley that headed to (Upper) Mesopotamia (as reported by the Greek sources), first to the fortified city of Muş/Mûş, some 120 kilometres to the west, and then to the relative safety of Melitene. Moreover, this explanation could also explain the reason why no warning got through to Romanos, as any messengers would have been easily intercepted by the Turks. Otherwise, even if Trachaneiotes did receive the recall order from Romanos (which I believe he did not, but for the sake of the argument, let’s assume that he did), and before rushing to brand him as a traitor, we should consider that his decision to withdraw west would have been based on his tactical intelligence and assessment of the situation on the ground; perhaps he considered that any attempt to reach the emperor would have been fraud with dangers bordering on the suicidal. We will never know!
Finally, there is another intriguing suggestion in the history of Ibn al-ʿAdīm, a jurist and historian born in Aleppo (b. 588/1192 – d. in Cairo in 660/1262), about the aftermath of the Battle of Manzikert. He wrote that “Of the Byzantine army no one survived except the contingent which was besieging Akhlat. When the news of the defeat reached them, they left the town in terror.” This could very well indicate that Trachaneiotes’ detachment remained in the region around Khliat and carried on with the siege of the city, as he had, probably, been ordered to when he was dispatched there in the first place; again, does this make him a traitor for not rushing to the emperor’s help? We do not know the background to his decision!
Interestingly, the fact that Trachaneiotes was appointed as protoproedros and doux of Antioch to replace the pro-Romanos Chatatourios almost immediately after the emperor’s blinding on June 29, 1072, may point to the Armenian’s anti-Romanos or, in my view, neutral, stance during the campaign – he was considered to be a highly experienced officer yet trustworthy enough by the new Doukas regime for this critical post that was far away from the centre of power to cause any trouble. Therefore, I believe that branding Roussel and Trachaneiotes as traitors and designating them as scapegoats for the disaster that befell the emperor and his army on Friday, August 26, 1071, is an unfair assessment of the decisions made by the two men based on the intelligence in hand and the tactical situation on the ground.
Betrayal, or a failed manoeuvre?
On Friday, August 26, the Turks made full use of their usual nomadic tactics of ‘feigned retreat’ that sought to exploit their extraordinary skill with the bow and their mobility as light horse-archers, usually by staying out of reach of their opponents’ weapons and using hit and run tactics in waves while ‘showering’ the enemy with arrows. Then, when the enemy’s forces were drawn out in pursuit and had become exhausted, the nomads would wheel around and deal a decisive blow. The critical question is for how long would the Seljuk horse archers have been able to sustain such a concerted barrage of fire and, to the same effect, for how long would the advancing imperial units have been able to withstand the onslaught?
Timing is of the essence here, and the Muslim sources agree that Alp Arslan’s counterstroke was delivered at the end of the afternoon prayer, when Muslim preachers would be delivering their sermons from the minbars or pulpits. This could very well have been the time when the sultan was informed by his scouts, who were concealed behind the rocky outcrops in higher ground observing enemy movements, about the “sudden misfortune that hit the Romans, [and] told the sultan what was going on, and insisted that he turn around.” What, exactly, was this sudden misfortune?
Sometime earlier that afternoon, Romanos issued a critical order for “the imperial banner to be turned around as a signal for the troops to return to the camp.” The emperor had probably realised that he had marched his lines far from his fortified camp, which was itself almost empty and vulnerable to attack, while it is likely that the gaps between his centre and wings were growing longer than they should have been. On top of that, his troops would have undoubtedly gotten tired, thirsty, and perhaps short of supplies as well, although we have no indication as to what happen between the mid-morning and the afternoon order by Romanos to withdraw to camp.
Unfortunately for the Byzantines, the signal to retreat to the camp was misinterpreted by some of the divisions in the flanks, who apparently thought that the centre, led by Romanos, had been defeated and/or that the emperor had fallen. It should be noted that every Byzantine military manual cautions the general against using too many ‘voiced’ signals during the battle: “We do not consider it desirable that many trumpets be sounded or blown during battle, for this causes disturbance and confusion, and commands cannot be properly understood.”
Setting aside any accusation of treachery, the fact that some of the Byzantine units proved unable to perform an orderly withdrawal back to the relative safety of the camp is a strong indication of the low morale and the lack of training and discipline in the imperial army. According to the Taktika of Leo VI, “Untrained men turn out to be totally ignorant and blind in the face of sudden and unexpected actions,” adding a little later that “it is necessary for you [general] to exercise the army and to make the soldiers skilled in war by getting them used to it,” arguing that “Even if they do acquire a little experience, they swiftly retreat, unable to bear the labour and the dangers.”
It is this kind of advice that could help us understand what happened in the afternoon of Friday, August 26th – an order to withdraw that disintegrated into a rout because of the collapse in communications and order. It was the moment that Alp Arslan had been awaiting, and his scouts were quick to spot and relay the intelligence back to him!
Another reason for Alp Arslan having to hold his attack against the Byzantines until late in the afternoon was the wind. Al-Ḥusaynī (d. after 622/1225), a late twelfth-century Iranian author whose account of Manzikert contains details not found in earlier accounts, notes that “a dusty wind blew which blinded the eyes of the Muslims, and the sultan’s army almost took flight … The sultan kept on repeating this entreaty and weeping until the directions of the winds were reversed and the eyes of the infidels were blinded.” This factor may have also played a role in the Turkish barrage being less effective than anticipated. But, as the sources assure us, there was a ‘miracle’ in the late hours of the afternoon and the wind changed direction, a development that was interpreted as divine approval of Alp Arslan’s counterattack.
We are not told which of the two flanks of the imperial army was the first to buckle, but Bryennios is the only source to report that “then they [Turks] turned back, and when they attacked with great violence and shouting against the Romans, they put the right horn to flight,” which included the Kappadokian and other Anatolian units, although our source does not mention the miscommunication with the banner signal to withdraw back to the camp. The role of the rear guard was essential under these circumstances, in that “if the first line happens to retreat, it may find refuge in the empty spaces between the divisions [of the second battle line], and with them it will seem to form one battle line.” Therefore, at this critical moment in the battle, as the troops of the right flank fled towards the safety of the camp, the rear guard, instead of rushing to assist, reportedly speeded up its own withdrawal.
The ‘pro-Romanos’ historian Attaleiates, followed by Skylitzes and Zonaras, interpreted this move by the commander of the rearguard, Andronikos Doukas, as a betrayal of the emperor. Attaleiates’ dubious comments read as follows:
Many relate (ὡς δ’ οι πολλοί πληροφοροῦσιν) that one of those who was waiting for a chance to get at him, a cousin of the emperor’s stepson Michael who had previously plotted against him, spread this report [of the emperor’s death] among the soldiers. He [Andronikos] quickly got his men together – for the emperor, with his good heart, had entrusted a large contingent to this man’s command – and fled back to the camp.
We should bear in mind that Attaleiates accompanied the emperor in his 1071 campaign in the capacity of a military judge (krités tou stratopédou), and although was present at the infamous defeat at Manzikert, he was not an eyewitness of the events but, rather, had remained in the relative safety of the camp. This is a critical point, because, his role was that of a judge and advisor, not a soldier, and there is nothing to suggest that Attaleiates ever enjoyed an appreciable view of any battle. Moreover, Attaleiates depended on oral and written sources about the events he did not personally witness, including the Battle of Manzikert, and he consistently refers to reports and messages as a source of information for distant events; a typical example of this is his compiling of conflicting reports on the fate of the emperor by eye-witnesses and informants returning to the imperial camp after the battle.
On the other hand, the ‘pro-Doukas’ Nikephoros Bryennios (‘the Younger’) asserts that the failure of the Byzantine rear guard to support the main division under the emperor enabled the Seljuks to “surround the emperor and assail him on all sides;” but there is no hint of treason by Andronikos in the Ύλη Ιστορίας. Additionally, Michael Psellos also conveniently ignores the alleged treachery of Michael’s cousin Andronikos Doukas in his Χρονογραφία, instead, attributing the defeat to Romanos as the key reason for the collapse. Although the ‘pro-Doukas’ Psellos bemoans the futility of Romanos’ campaigns, Attaleiates shows that they were moderately successful in reconstituting the army and making some valuable defensive gains.
Therefore, I believe that the failure of Andronikos’ division to come to the aid of the collapsing front should be attributed to the great distance that would have developed between the rear guard and the main line during the hours of the afternoon.
We can draw some parallels here with the annihilation of the unit of Varangians at the Battle of Dyrrachium, on October 18, 1081, when the emperor Alexios Komnenos, seeing the right wing of the Normans in disorderly retreat, signalled a general advance. But with the centre and right divisions of the Byzantines already involved in a moderate skirmish with their opposite Norman divisions, and the Varangians advancing well beyond the Byzantine line of march, the tactical mistake was Alexios’ failure to catch up and support the Varangians’ advance, thus leaving them cut off to be annihilated by the Normans. As Anna Komnene is very careful to hush her father’s blunder at Dyrrachium, arguing that “the axe-bearing barbarians [Varangians] and their leader Nabites had in their ignorance and in their ardour of battle advanced too quickly and were now a long way from the Roman lines, burning to engage battle with the equally brave Franks [Normans].” We should expect the same hunt for ‘scapegoats’ from the pro-Romanos sources.
It is my conviction, therefore, that the brandishing of Trachaneiotes, Roussel, and Andronikos Doukas as traitors only served to mask/excuse the strategic and tactical errors made by Diogenes during the later stages of the campaign of 1071. It was the emperor who misinterpreted the Turkish strategy after Alp Arslan’s retreat from Aleppo; it was he who miscalculated the threat of the local garrisons at Manzikert and Khliat that led him to divide his troops at a critical stage of the campaign; and it was he who overestimated the ability of his troops to perform a difficult and complex military manoeuvre like the withdrawal to the camp. It is all a matter of perspective!
Georgios Theotokis: Ph.D History (2010, University of Glasgow), specializes in the military history of the Eastern Mediterranean in Late Antiquity and the Middle Ages. He has published numerous articles and books on the history of conflict and warfare in Europe and the Mediterranean in the Medieval and Early Modern periods. His latest book is Twenty Battles That Shaped Medieval Europe. He has taught in Turkish and Greek Universities; he is currently a postdoctoral researcher at the Byzantine Studies Research Centre, Bosphorus University, Istanbul.
By Georgios Theotokis
Few people have been vilified in such a way after the disastrous outcome of a decisive battle other than Trachaniotes, Roussel of Bailleul, and Andronikos Doukas. The first two failed to follow their lord’s orders and come to his aid, while the latter deserted the emperor and ordered a withdrawal from the battlefield when he should have supported the Byzantine front line to avert its envelopment and subsequent collapse. But how credible are these accounts? It depends on who we read.
Because Michael Attaleiates was left behind in the fortified Byzantine camp, he was able to offer a highly emotional personal account of the unfolding disaster, as he experienced it through what he saw and what he heard from the soldiers who fled from the battlefield. But how objective is Attaleiates’ account in terms of his personal bias and sympathies compared to Nikephoros Bryennios (‘the Younger’)?
The Prelude to the Battle
The Battle of Manzikert was the geo-political climax of the struggle for control of key cities and towns in eastern Asia Minor and northern Syria between two major powers, the Byzantine Empire and the Seljuk Turks. However, it is widely believed that the Seljuk sultan Alp Arslan’s primary strategic target was not the Byzantine Empire but the Shi’a Fatimid Caliphate. For that reason, the Turkish campaign of 1070/71 that preceded the fateful clash at Manzikert in 1071, aimed at laying the foundation for a major future invasion of Egypt, and that meant the targeting of Aleppo to force the emir of the city to renounce his allegiance to the Fatimid Caliph al-Mustansir.
Therefore, Alp Arslan wished to have his flanks covered for the critical campaign in Syria, and for that reason he conquered the fortress towns of Artchesh and Manzikert, on the north shore of Lake Van, sometime in the early autumn of 1070. Attaleiates reports that Manzikert was, subsequently, garrisoned by a strong contingent of Daylami fighters from northern Iran, while the incessant Turkish raids of the previous years had transformed the area stretching between Theodosiopolis and Manzikert into ἀοίκητον και ἠρημωμένη.
Alp Arslan then marched towards the Upper Tigris Valley between Amida (Diyār Bakr) and Mayyafariqin, the heartland of the Marwānid emirate, on his way to his first objective of Byzantine-ruled Edessa, a strategic city that had to be taken before he could march on to Aleppo. The Turkish army besieged Edessa in mid-winter (of 1070/71), failing to capture it due to the stubborn resistance of the formidable Byzantine garrison, and the competent command of the proedros and doux of Edessa, Basil Alousianos.
Alp Arslan decided to withdraw from Edessa on 19 January 1071, marching his army across the Euphrates and against the much larger prize of Aleppo. Eventually, just a few weeks after the start of the siege (perhaps on 4 May 1071), the emir of Aleppo, Mahmūd bin Naşr, submitted to Alp Arslan and, handing over to him the keys of the city, was confirmed as governor of the city by Alp Arslan.
Alp Arslan’s campaign in Syria in the winter of 1070/71 should be considered a strategic success because it significantly boosted his influence over the regions of northern Syria and the Upper Euphrates by imposing Seljuk suzerainty over the Mirdasid emir of Aleppo while putting further pressure on the remaining imperial outposts of Edessa and Antioch. The sultan also performed a necessary show of force in the regions of his other vassals, the Marwānids and the Uqaylids, and before all of these had taken place, he had inflicted yet another significant blow to the defences of the Byzantine empire in Vaasparakania by capturing Manzikert and Artchesh. However, Alp Arslan’s attention would be drawn away from Syria by events taking place farther north.
The historian Nikephoros Bryennios (‘the Younger’) provides an important view of the background behind Romanos’ fateful decision to launch a major strike towards Lake Van. He reports that in the military council that took place sometime during the march of the imperial army between Kaisarea and Sebasteia both the respected Georgian magistros Joseph Tarchaniotes and Nikephoros Bryennios (‘the Elder’) sought to convince the emperor not to take to the field against the Turks but, rather, to wait for them in Theodosiopolis so that Alp-Arslan would “be forced to clash with the Romans where it is in their [Romans] interest to make war. But they [Trachaneiotes and Bryennios] were under the impression that they were talking to a deaf man … more attention was paid [by Romanos] to the flatterers and not to those who advised the right thing.” Unsurprisingly, the more bellicose officers in the council prevailed and the march towards Khliat and Manzikert was confirmed, much to Bryennios’ bewilderment and consternation.
Dividing his forces: overconfidence or lack of intelligence?
The imperial army under Diogenes was based in Theodosiopolis until, probably, the high summer of 1071. Sometime in late July or early August, his growing overconfidence led the emperor to make a critical strategic error. After leaving Theodosiopolis to march south-east to Manzikert via Xnis/Xinus (modern Hınıs, Erzurum province), a distance of around 230kms, the emperor decided to divide his forces because, according to Attaleiates, he was under the impression that the Turks and Iranian Daylami defenders of Manzikert would be easily overcome.
Other sources also account that the emperor dispatched separate detachments of troops to march to the strategic fortress-city of Khliat that controlled the western approaches to the region of Lake Van. These detachments comprised of “Uze mercenaries and the Franks under Rouselios … [sent] towards Khliat to forage for provisions,” and because the emperor “considered them [Turkish defenders] of no great importance since they were few in number, he detached another not insignificant portion of his army [to march to Khliat] and placed it under the command of the magistros Joseph Trachaneiotes.”
There are two crucial points we need to emphasize here; first, that the detachment under the Norman Roussel de Bailleul and the magistros Trachaneiotes “consisted of select troops, difficult to withstand, taking the initiative in close combat and other types of battle and ready to face danger, and far more numerous than the soldiers retained by the emperor.” Second, the emperor planned to take Manzikert with the units he kept under his command, but he did not plan to join forces with Roussel’s and Trachaneiotes’ detachments unless there was a dire emergency and, certainly, not to march against Khliat before “Manzikert was restored to Roman rule.” I believe that our sources make it clear that the emperor was planning for a swift campaign to capture Manzikert and Khliat, and, perhaps, even subdue the remaining fortress-towns on the northern shore of Lake Van, replace their garrisons with imperial troops and head back home before there was any reaction from the Seljuk leadership. Romanos was in for a shock!
To understand the background behind Romanos’ strategic decision, we need to refer to the decisive issue of Byzantine intelligence. When it comes to the supply of intelligence to the emperor in this critical juncture of the campaign, with the imperial army stationed at Theodosiopolis in the summer of 1071 and in need of accurate information about the Turkish intentions before marching against the main target of Khliat and Manzikert, the role of Leon Dabatenos must be studied more closely. Leon Dabatenos (Armenian: Lewon Dawatanos) came from a well-known aristocratic family of Armenian origin that found itself on the rise around the middle of the eleventh century but was already in decline in the twelfth century.
A certain Dabatenos was katepano or doux of Edessa and died during the siege of Amida in 1062, while another Leon Dabatenos is identified as vestarches and doux/katepano of Edessa, who held the office ca. 1071-1077. It is interesting to note that the latter Dabatenos had replaced the doux of Edessa Basil Alousianos who was responsible for the defence of the city from the Seljuks in the winter of 1070/71, with Alousianos disappearing from the record shortly after the Turkish withdrawal from Edessa. This change of leadership at such a decisive moment is intriguing, but I was unable to identify the reason for Alousianos’ dismissal.
Moreover, Nikephoros Bryennios (‘the Younger’) is the only source to account for an official report that reached Diogenes by Dabatenos while the emperor was probably still in Theodosiopolis with the bulk of his army, according to which “the sultan, it read, learning about the campaign [of Romanos to Khliat and Manzikert] and afraid of the power [of Romanos’ army], left Persia for Babylon [Baghdad];” but what is more important about the outcome of the campaign is what immediately follows, “the emperor, putting his trust on these words, divided his army in two: he kept one in place [Manzikert], and he dispatched the other to Khliat.”
The aforementioned report should be combined with another intriguing comment by Sibṭ Ibn al-Jawzī (b. 581/1185 – d. 654/1256), a renowned Muslim preacher and historian whose account of the campaign and battle of Manzikert offers significant details about Alp Arslan’s activities leading up to the battle itself, who writes that “the envoy of Byzantium returned [to the imperial camp] rejoicing to his master [Romanos] and that strengthened the resolve of the king of Byzantium to pursue him [Alp Arslan] and to fight him.”
Do both Bryennios and al-Jawzī basically describe the same person compiling an intelligence report for the emperor that was delivered either in letter (Bryennios) or in person (al-Jawzī)? Was Leon Dabatenos the head of the embassy dispatched by the emperor to Alp Arslan while the latter was in the vicinity of Aleppo in the early spring of 1071? There is some food for thought here, but the sources raise more questions than provide answers. Nonetheless, it is very reasonable for the emperor to have demanded intelligence from the doux of Edessa about the size, units and morale of Alp Arslan’s army; the critical point, here, is how far could Romanos trust Leon Dabatenos.
When it comes to the matter of allegiance, technically, Dabatenos’ was with the Byzantine emperor, but an extract from Matthew of Edessa’s Chronicle may suggest otherwise:
The doux [of Edessa], who resided in the city [of Edessa], provided him [Alp Arslan] with horses, mules, and victuals. Taking these, the sultan passed through the confines of Edessa, unharmed, and went in an easterly direction towards the mountain called Lesun.
Although he is not mentioned by name, what Matthew suggests here is that the doux of Edessa may have acknowledged the suzerainty of the Seljuk sultan, thus having to provide him with the necessary supplies for his march north from Syria back to Iran. But what about the report dispatched to Emperor Romanos on the strength and size of the Seljuk army? Was Dabatenos deliberately attempting to mislead the emperor with false intelligence? We will never know with certainty, but the evidence is damning, especially if we consider that he remained in office throughout Michael VII’s reign until ca. 1077.
Finally, we have Attaleiates’ comment that “spies had already reported to him [Romanos] that the sultan was in a hurry to return to Persia.” If we couple this with the report dispatched by Leon Dabatenos, both reports would, certainly, have boosted the emperor’s confidence and conviction that, following his strategic withdrawal from Syria, Alp Arslan would adopt a defensive strategy in Iraq and would not attempt to march against him in Armenia. The failure of the intelligence reports is critical here!
Moving on to the emperor’s decision to recall the units that had been dispatched to Khliat under Roussel and Trachaneiotes. It is only Bryennios (‘the Younger’) who reports of a military council summoned by the emperor on the eve of the battle, Thursday, August 25, to assess the situation and plan accordingly, while, again, he highlights the polarization within the council between those who wished to go out immediately and face the main Seljuk army in a pitched battle, and those who urged caution and to await the arrival of those troops recalled from Khliat who would, of course, never arrive. Attaleiates and Skylitzes fail to mention any military council summoned on Thursday morning, adding that “he [Diogenes] put off the battle. But then he gave up on receiving any help from them [units dispatched to Khliat], for he suspected that some powerful obstacle was preventing their arrival.”
However, Romanos was still under the impression that the elite units in Khliat were, simply, delayed rather than withdrawn from the region, hence the following statement of the emperor’s intention that “[the emperor] decided to do battle against the enemy on the following day [Friday, 26 August] as bravely as he could with the soldiers he had at hand. He had, moreover, some hope that the others [units dispatched to Khliat] would not be delayed beyond the next day,” with Attaleiates adding that “he [emperor] was unaware that their general [Trachaneiotes], upon learning of the sultan’s attack against the emperor, had taken all his soldiers and shamefully retreated through Mesopotamia until he came into Roman territory. The coward took no account of his lord or even his duty.” But how fair is this assessment of the actions of Roussel and Trachaneiotes?
With the Greek sources being categorized as pro- and anti- Romanos, Attaleiates and Skylitzes are inclined to excuse the emperor’s decision to dispatch almost half of his troops to Khliat. They emphasize that “dividing his [Romanos] forces was not irrational or lacking in strategic calculation. It was fate, or rather the wrath or design of God beyond our understanding.” Unsurprisingly, Bryennios provides a different view of the background behind this fateful decision, being the only one to report the decisions of the war council in Theodosiopolis that confirmed the plan to advance to Manzikert instead of the -more cautious- strategy of having to wait for the Seljuks to come to them, a strategy that was supported by Trachaneiotes and Nikephoros Bryennios (‘the Elder’). But there is a surprising inference by Skylitzes in that Trachaneiotes was, indeed, privy to the plan already formulated by Andronikos Doukas to betray the emperor at the first opportune moment, and he even managed to persuade Roussel to join forces with him and retreat to Roman territory “through Mesopotamia” instead of rushing back to Manzikert to reinforce Diogenes, as he intended to do.
There is a critical question we need to ask at this point; if the emperor dispatched messengers to Trachaneiotes and Roussel after he realised that his army was facing the main Seljuk army instead of units of irregular Turkmen, did they manage to deliver the emperor’s orders for a recall back to Manzikert? We have no evidence to suggest they did, as the area between Manzikert and Khliat would have been swarming with enemy (mounted) troops.
Earlier, I quoted a key statement by Attaleiates that has been considered sufficient evidence to assume that Trachaneiotes did, indeed, receive the order for a recall by Romanos and, thus, chose to ignore it; let me quote the sentence again: “for he [Romanos] was unaware that their general [Trachaneiotes], upon learning of the sultan’s attack against the emperor, had taken all of his soldiers and shamefully retreated through Mesopotamia until he came into Roman territory.”
In my view, however, this comment does not necessarily mean that Trachaneiotes received the news of the unfolding crisis at Manzikert from Romanos’ messengers. Rather, the intelligence could very well have come from his own scouts who were trying to re-establish contact with the emperor and, possibly, warn him!
Furthermore, Attaleiates’ comment that “He [Romanos] had, moreover, some hope that the others [units dispatched to Khliat] would not be delayed beyond the next day,” has led scholars to believe that “the focus on Diogenes’s anxious expectation for news from Khliat means that his orders had reached Trachaneiotes safely.” No, it does not! It is, simply, wishful thinking on behalf of the emperor, and a catastrophic assessment of the tactical situation.
Additionally, Romanos’ expectation that the messengers would have evaded the roaming Turkmen once (from Manzikert to Khliat, distance: 60 kilometres) and delivered the recall order to Trachaneiotes and, then, evade them twice (Khliat to Manzikert, again, distance: 120 kilometres) by returning to the imperial camp to report to the emperor, all the while the area between Manzikert and Khliat was swarming with enemy cavalry was, to say the least, over-optimistic. We should remind ourselves that these comments come from the pen of a pro-Romanos source, Attaleiates, who is looking for scapegoats about seven or eight years after the battle, and the same source who is ready to accuse Andronikos Doukas for treachery although, as I will show further down, there are more credible causes for the collapse of the imperial army on Friday, August 26.
Another critical question is how to make sense of the strategic reaction of Trachaneiotes and Roussel to retreat to “Mesopotamia” in the face of the rapidly deteriorating situation in the vicinity of Khliat that pitted them against strong(er) Turkish reinforcements dispatched by Alp Arslan exactly to prevent these two armies from marching back to reinforce their emperor? Following the reinforcements dispatched by Alp Arslan to the garrison of Khliat, Trachaneiotes’ army would have launched a siege of the city according to the orders of the emperor, as it is reported by our Muslim sources, yet we are at a loss as to what happened after that, because the focus shifts to the siege of Manzikert. We can only speculate that the Byzantine commanders would have realised that they were up against a strong Seljuk army that had arrived from the south, with Turkish cavalry effectively ‘flooding’ the region between Khliat and Manzikert and, consequently, cutting communications between Trachaneiotes’ army and the emperor.
Assuming that this interpretation of events is correct, then the most realistic option for the beleaguered Byzantine units around Khliat would have been to retreat west along the Arsanias River valley that headed to (Upper) Mesopotamia (as reported by the Greek sources), first to the fortified city of Muş/Mûş, some 120 kilometres to the west, and then to the relative safety of Melitene. Moreover, this explanation could also explain the reason why no warning got through to Romanos, as any messengers would have been easily intercepted by the Turks. Otherwise, even if Trachaneiotes did receive the recall order from Romanos (which I believe he did not, but for the sake of the argument, let’s assume that he did), and before rushing to brand him as a traitor, we should consider that his decision to withdraw west would have been based on his tactical intelligence and assessment of the situation on the ground; perhaps he considered that any attempt to reach the emperor would have been fraud with dangers bordering on the suicidal. We will never know!
Finally, there is another intriguing suggestion in the history of Ibn al-ʿAdīm, a jurist and historian born in Aleppo (b. 588/1192 – d. in Cairo in 660/1262), about the aftermath of the Battle of Manzikert. He wrote that “Of the Byzantine army no one survived except the contingent which was besieging Akhlat. When the news of the defeat reached them, they left the town in terror.” This could very well indicate that Trachaneiotes’ detachment remained in the region around Khliat and carried on with the siege of the city, as he had, probably, been ordered to when he was dispatched there in the first place; again, does this make him a traitor for not rushing to the emperor’s help? We do not know the background to his decision!
Interestingly, the fact that Trachaneiotes was appointed as protoproedros and doux of Antioch to replace the pro-Romanos Chatatourios almost immediately after the emperor’s blinding on June 29, 1072, may point to the Armenian’s anti-Romanos or, in my view, neutral, stance during the campaign – he was considered to be a highly experienced officer yet trustworthy enough by the new Doukas regime for this critical post that was far away from the centre of power to cause any trouble. Therefore, I believe that branding Roussel and Trachaneiotes as traitors and designating them as scapegoats for the disaster that befell the emperor and his army on Friday, August 26, 1071, is an unfair assessment of the decisions made by the two men based on the intelligence in hand and the tactical situation on the ground.
Betrayal, or a failed manoeuvre?
On Friday, August 26, the Turks made full use of their usual nomadic tactics of ‘feigned retreat’ that sought to exploit their extraordinary skill with the bow and their mobility as light horse-archers, usually by staying out of reach of their opponents’ weapons and using hit and run tactics in waves while ‘showering’ the enemy with arrows. Then, when the enemy’s forces were drawn out in pursuit and had become exhausted, the nomads would wheel around and deal a decisive blow. The critical question is for how long would the Seljuk horse archers have been able to sustain such a concerted barrage of fire and, to the same effect, for how long would the advancing imperial units have been able to withstand the onslaught?
Timing is of the essence here, and the Muslim sources agree that Alp Arslan’s counterstroke was delivered at the end of the afternoon prayer, when Muslim preachers would be delivering their sermons from the minbars or pulpits. This could very well have been the time when the sultan was informed by his scouts, who were concealed behind the rocky outcrops in higher ground observing enemy movements, about the “sudden misfortune that hit the Romans, [and] told the sultan what was going on, and insisted that he turn around.” What, exactly, was this sudden misfortune?
Sometime earlier that afternoon, Romanos issued a critical order for “the imperial banner to be turned around as a signal for the troops to return to the camp.” The emperor had probably realised that he had marched his lines far from his fortified camp, which was itself almost empty and vulnerable to attack, while it is likely that the gaps between his centre and wings were growing longer than they should have been. On top of that, his troops would have undoubtedly gotten tired, thirsty, and perhaps short of supplies as well, although we have no indication as to what happen between the mid-morning and the afternoon order by Romanos to withdraw to camp.
Unfortunately for the Byzantines, the signal to retreat to the camp was misinterpreted by some of the divisions in the flanks, who apparently thought that the centre, led by Romanos, had been defeated and/or that the emperor had fallen. It should be noted that every Byzantine military manual cautions the general against using too many ‘voiced’ signals during the battle: “We do not consider it desirable that many trumpets be sounded or blown during battle, for this causes disturbance and confusion, and commands cannot be properly understood.”
Setting aside any accusation of treachery, the fact that some of the Byzantine units proved unable to perform an orderly withdrawal back to the relative safety of the camp is a strong indication of the low morale and the lack of training and discipline in the imperial army. According to the Taktika of Leo VI, “Untrained men turn out to be totally ignorant and blind in the face of sudden and unexpected actions,” adding a little later that “it is necessary for you [general] to exercise the army and to make the soldiers skilled in war by getting them used to it,” arguing that “Even if they do acquire a little experience, they swiftly retreat, unable to bear the labour and the dangers.”
It is this kind of advice that could help us understand what happened in the afternoon of Friday, August 26th – an order to withdraw that disintegrated into a rout because of the collapse in communications and order. It was the moment that Alp Arslan had been awaiting, and his scouts were quick to spot and relay the intelligence back to him!
Another reason for Alp Arslan having to hold his attack against the Byzantines until late in the afternoon was the wind. Al-Ḥusaynī (d. after 622/1225), a late twelfth-century Iranian author whose account of Manzikert contains details not found in earlier accounts, notes that “a dusty wind blew which blinded the eyes of the Muslims, and the sultan’s army almost took flight … The sultan kept on repeating this entreaty and weeping until the directions of the winds were reversed and the eyes of the infidels were blinded.” This factor may have also played a role in the Turkish barrage being less effective than anticipated. But, as the sources assure us, there was a ‘miracle’ in the late hours of the afternoon and the wind changed direction, a development that was interpreted as divine approval of Alp Arslan’s counterattack.
We are not told which of the two flanks of the imperial army was the first to buckle, but Bryennios is the only source to report that “then they [Turks] turned back, and when they attacked with great violence and shouting against the Romans, they put the right horn to flight,” which included the Kappadokian and other Anatolian units, although our source does not mention the miscommunication with the banner signal to withdraw back to the camp. The role of the rear guard was essential under these circumstances, in that “if the first line happens to retreat, it may find refuge in the empty spaces between the divisions [of the second battle line], and with them it will seem to form one battle line.” Therefore, at this critical moment in the battle, as the troops of the right flank fled towards the safety of the camp, the rear guard, instead of rushing to assist, reportedly speeded up its own withdrawal.
The ‘pro-Romanos’ historian Attaleiates, followed by Skylitzes and Zonaras, interpreted this move by the commander of the rearguard, Andronikos Doukas, as a betrayal of the emperor. Attaleiates’ dubious comments read as follows:
Many relate (ὡς δ’ οι πολλοί πληροφοροῦσιν) that one of those who was waiting for a chance to get at him, a cousin of the emperor’s stepson Michael who had previously plotted against him, spread this report [of the emperor’s death] among the soldiers. He [Andronikos] quickly got his men together – for the emperor, with his good heart, had entrusted a large contingent to this man’s command – and fled back to the camp.
We should bear in mind that Attaleiates accompanied the emperor in his 1071 campaign in the capacity of a military judge (krités tou stratopédou), and although was present at the infamous defeat at Manzikert, he was not an eyewitness of the events but, rather, had remained in the relative safety of the camp. This is a critical point, because, his role was that of a judge and advisor, not a soldier, and there is nothing to suggest that Attaleiates ever enjoyed an appreciable view of any battle. Moreover, Attaleiates depended on oral and written sources about the events he did not personally witness, including the Battle of Manzikert, and he consistently refers to reports and messages as a source of information for distant events; a typical example of this is his compiling of conflicting reports on the fate of the emperor by eye-witnesses and informants returning to the imperial camp after the battle.
On the other hand, the ‘pro-Doukas’ Nikephoros Bryennios (‘the Younger’) asserts that the failure of the Byzantine rear guard to support the main division under the emperor enabled the Seljuks to “surround the emperor and assail him on all sides;” but there is no hint of treason by Andronikos in the Ύλη Ιστορίας. Additionally, Michael Psellos also conveniently ignores the alleged treachery of Michael’s cousin Andronikos Doukas in his Χρονογραφία, instead, attributing the defeat to Romanos as the key reason for the collapse. Although the ‘pro-Doukas’ Psellos bemoans the futility of Romanos’ campaigns, Attaleiates shows that they were moderately successful in reconstituting the army and making some valuable defensive gains.
Therefore, I believe that the failure of Andronikos’ division to come to the aid of the collapsing front should be attributed to the great distance that would have developed between the rear guard and the main line during the hours of the afternoon.
We can draw some parallels here with the annihilation of the unit of Varangians at the Battle of Dyrrachium, on October 18, 1081, when the emperor Alexios Komnenos, seeing the right wing of the Normans in disorderly retreat, signalled a general advance. But with the centre and right divisions of the Byzantines already involved in a moderate skirmish with their opposite Norman divisions, and the Varangians advancing well beyond the Byzantine line of march, the tactical mistake was Alexios’ failure to catch up and support the Varangians’ advance, thus leaving them cut off to be annihilated by the Normans. As Anna Komnene is very careful to hush her father’s blunder at Dyrrachium, arguing that “the axe-bearing barbarians [Varangians] and their leader Nabites had in their ignorance and in their ardour of battle advanced too quickly and were now a long way from the Roman lines, burning to engage battle with the equally brave Franks [Normans].” We should expect the same hunt for ‘scapegoats’ from the pro-Romanos sources.
It is my conviction, therefore, that the brandishing of Trachaneiotes, Roussel, and Andronikos Doukas as traitors only served to mask/excuse the strategic and tactical errors made by Diogenes during the later stages of the campaign of 1071. It was the emperor who misinterpreted the Turkish strategy after Alp Arslan’s retreat from Aleppo; it was he who miscalculated the threat of the local garrisons at Manzikert and Khliat that led him to divide his troops at a critical stage of the campaign; and it was he who overestimated the ability of his troops to perform a difficult and complex military manoeuvre like the withdrawal to the camp. It is all a matter of perspective!
Georgios Theotokis: Ph.D History (2010, University of Glasgow), specializes in the military history of the Eastern Mediterranean in Late Antiquity and the Middle Ages. He has published numerous articles and books on the history of conflict and warfare in Europe and the Mediterranean in the Medieval and Early Modern periods. His latest book is Twenty Battles That Shaped Medieval Europe. He has taught in Turkish and Greek Universities; he is currently a postdoctoral researcher at the Byzantine Studies Research Centre, Bosphorus University, Istanbul.
Click here to read more from Georgios Theotokis
Click here to read this article with footnotes as a PDF file
Related Posts
Subscribe to Medievalverse