Military Demography

For military history, the problem of assessing the size and make up of armies is of critical importance, and the question “how big was an army?” echoes in the mind of any historian interested in the history of war. This “million dollar” question has stirred up such a heated debate within the academic community in the last century that historians should tread with caution when referring to issues of military demography. Therefore, it is my intention here to assess the various factors that would have affected the army estimates of the primary sources for the Battle of Manzikert, and to explain to what extent the figures they provide for army sizes are reliable, both in absolute numbers and in the ratios given between mounted and dismounted troops.

In terms of how contemporary historians obtained their numbers, we may consider two broad categories: first, those who had no access to official reports and data because of their position in society and/or because of any secrecy act/policy by the state; and second, those who did have access to official reports and data, and therefore provided numbers based either on military reviews or on regulations or custom. To begin with, the accounts of historians belonging to the first category would have been affected by three factors, depending on whether they were eyewitness to the events they describe. The first factor relates to the time when their work was written, and the reliance on oral testimonies by the historians. Such accounts always bore the risk of inflation and/or miscalculation. We should also consider that memories inevitably focus on the outcome and significant incidents. In fact, Thucydides observes that battles are difficult enough to reconstruct because each witnesses only knows what is happening in their immediate vicinity. This becomes more acute the lower an eyewitnesses was in the military hierarchy and, as a result, would have had a limited opportunity both to observe and to grasp the reality on the battlefield, which could provide us with a distorted picture of the day's events (although not always).

Chaos in the Camp

Because he was present in the fortified Byzantine camp, Michael Attaleiates 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.1 There is a mixture of feelings in Attaleiates's account; panic and confusion about how the events unfolded, and uncertainty over who to ask for information: “Meanwhile, the flight of the others had led them to cluster outside the camp palisade and all were shout-ing incoherently and running about in disorder; nobody could say what exactly was going on.” Critically for Attaleiates, the fate of the emperor was of utmost importance, hence we read the contradicting reports that he received throughout Friday afternoon, “some claimed that the emperor was firmly resisting with what was left of his army and that he had routed the barbarians. Others said that he had been killed or captured. Everyone had something different to report, claiming victory for each side and then alternately denying it.”

Tthere was the growing awareness of the outcome of the battle, coupled with Atta-leiates's desperate attempts to avert the unfolding disaster by trying to convince some of the retreating soldiers to return to their posts, like in the aftermath of the desertion of some of the Oghuz on Thursday (August 25): “Finally, many of the Kappadokians who were with the emperor, one group after another, began to desert. Whether I myself was trying to stop many of the soldiers from running away and getting them to return to their posts to save us from defeat, I leave it to others to report.”

It was when Attaleiates had finally accepted that everything was lost, and a feeling of sadness and desperation set in that we read his harrowing words about the fate of the emperor and his army. There are reports of the Turks seizing and looting the imperial camp, but the loss of prestige was far harder to bear:

THE MAIN AIM of this chapter is to examine the primary sources for the Battle of Manzikert strictly from a military perspective and attempt to reach some conclusions regarding their value for the history of eleventh-century warfare in the region of Asia Minor. The major questions that will be raised in this chapter are: how accurate and detailed are their descriptions of pitched battles and clashes between enemy units during the Byzantine campaigns of 1068 to 1071 in the East? How far do their narratives permit an accurate reconstruction of chains of events, especially regarding the battlefield manoeuvres of armies in action?

The “Battle-Piece”

Why do historians write “battle-pieces”? And how should a modern reader attempt to understand the narrative of such a dramatic and complex event? Examining what a battle narrative can tell us and how to read it should be a central part of our approach to military history. But first, what is a “battle-piece”? We could say that it is the battle account that forms the essential element in the warfare narrative; a literary piece that was incorporated into hagiographies, letters, rhetorical works, epic poetry, romances, and, of course, historical texts. Yet, what remains unclear in the aforementioned definition is the perspective from which the battle narrative is written. Because the vast majority of the battle narratives from ancient to early modern times focus on events from the perspective of the commander(s) who inevitably had a disproportionate influence on the course of events, modern scholars have followed the same well-trodden course of writing about battles, adopting a traditional “command-centred” approach to the strategic, operational, and tactical movements of armies, where the thoughts, decisions, and actions of the commander(s) are far more visible than those of the common soldier(s).

Accordingly, one of the toughest challenges involved in writing about medieval warfare is describing what battle in that period “was really like”: how it might have felt to be there, including not just the physical but also the emotional sensations. What would have been running through the combatants’ minds? What were the motivations used to suppress their primal instinct for running away from danger?

Diogenes's Campaigns of 1068 and 1069

Emperor Romanos's choice of Hierapolis and northern Syria as the main area of opera-tions for the 1068 campaigning season, rather than, for example, the northeastern (“anti Taurus”) region in Asia Minor, should be attributed to the military pressure brought upon the ducate of Antioch by the Seljuk raiding parties led by Afşin Bey, Sunduq al Turki, Ahmetşah, Türkman, Demleçoğlu Mehmet, Duduoğlu, Serhenkoğlu, and others. These particular groups, but especially the one led by Afşin, were launching increasingly destructive raids both in the frontier regions and in the interior of Asia Minor, as we saw in the previous chapter. Afşin's intention may well have been to disrupt the communica-tions between Edessa and Antioch, considering his victory over the doux of Edessa in 1066, and his wintering at the foot of the Black Mountains, only 90 kilometres north of Antioch, the following year.

Attaleiates makes a reference to Alp Arslan “detaching a fairly large contingent that he divided into two parts and encamped in upper Asia, placing the one more to the north and the other to the south,” which should be understood as the sultan's strategic response to the news of Diogenes's campaign preparations in the theme of Anatolikon, assigning large detachments in both the northern (Caucasus and Armenia), and the southern (Edessa and Syria) operational areas. A few lines later Attaleiates also notices that “the Persian army to the north…gave the impression that it was withdrawing before the emperor's advance,” hinting at the apparent departure of the main Seljuk army from the Caucasus region. Accordingly, we see Romanos's reaction in his advance southeast to the small theme of Lykandos, to the west of Melitene, instead of marching north to Sebasteia and Koloneia, evidently in preparation for an invasion of Syria.

The strategic and tactical inability of the cumbersome imperial forces to block or intercept the smaller and more mobile Turkish raiding parties became apparent when the emperor had to march back north to Sebasteia to deal with the aftermath of an “utterly unexpected” Turkish raid and pillage of Neokaisarea on the Lykos River, around 160 km to the northwest of Koloneia.

HAVING A WIDE variety of primary sources at a historian's disposal can be a blessing or a curse, depending on how ready one is to peel away the multiple layers of “literary goals” and “‘political agendas” to reach “true history” (as opposed to Anthony Kaldellis's “unhistory”). A historian writing about the Battle of Manzikert, therefore, must tackle a dizzying array of sources not only from both the Chalcedonian and the non-Chalcedonian historiographic traditions but also from the medieval Arabic and Persian ones. Some of them were contemporaries to the events they describe, which means they had access to stories and evidence from the recent past, while others repeated well-established historical episodes for didactic and/or propagandistic purposes several centuries after the battle; some of them were (high-ranking) civil servants, while others were religious scholars and famous preachers.

To answer the question of how we are to write the history of the Battle of Manzikert in light of what we know about our sources, we should appreciate their accounts not as collections of “facts” that can be understood as being either true or false, but rather as “literary artefacts” with their own aims, skills and aesthetic.” Or, as Stamatina McGrath aptly put it, “the act of composing history should be understood as a deliberate, political act with the intention of influencing how the past would inform the present and future concerns of the authors, their patrons, and their worlds.” Therefore, to be able to read and better evaluate the battle-narrative of another author, we need to be able to understand the frame of reference with which the author constructed narratives. Accordingly, this chapter will refer to some basic points in our authors’ social and educational backgrounds, the date and place of the compilation of their works, their own sources and the way they collected their information. I will try to make sense of their biases and sympathies, and how far their “socio-political agendas” shaped accounts, paying particular attention to their treatment of the personalities and reigns of Emperor Romanos and his successors, Michael VII and Nikephoros III.

The Turkoman and Seljuk Incursions into Eastern Asia Minor, 1017–1067

The late 1030s witnessed a struggle for supremacy in Khorāsān between the Ghaznavids and the Seljuks who, by taking advantage of a growing Ghaznavid weakness that followed the death of Sultan Mahmud of Ghāzni (r. 998–1030), gradually seized territories formerly administered by the Ghaznavids. Merv went over to the Seljuks in 1037, followed by the cities of Herāt and Nīshāpūr less than a year later. On 8 Ramażān 431/23 May 1040, outside Dandānqān near the city of Merv (in modern Turkmenistan), a force of some sixteen thousand Turkmen led by the Seljuk brothers Ṭuḡril Beg and Čaḡrī Beg, defeated a heavily armed Ghaznavid army under Sultan Masʿūd of Ghāzni (r. 1030–1041), whose combat effectiveness had been badly impaired by famine and drought conditions in the region. Not only did the victory abruptly end Ghaznavid dominion there, as Ṭuḡril was proclaimed emir of Khorāsān on the battlefield, but it also opened up Khorāsān to the Seljuks. Rayy and Hamadhan fell within the next three years, thus becoming strategic jumping-off points for further expansion westwards into Azerbaijan and Upper Mesopotamia.

Attacks in eastern Asia Minor by bands of the so-called ‘Iraqiya Turkmen began many years before the Battle of Dandānqān, with the Armenian sources dating them to either the year 465 ae (1016–1017 ad)2 or 467 ae (1018–1019 ad), about half a decade before Vaspourakan was officially surrendered to the Byzantine Empire by Senek’erim-Yovanēs. There is a consensus among historians that these were nothing more than raids conducted by unruly Turkmen to acquire pasturelands and loot, which were “side effects” of the Seljukid infiltration of Azerbaijan and Upper Mesopotamia. These ‘Iraqiya Turkmen were followers of one of Seljuk's sons, Isra’il/Arslan, who had been taken prisoner by the Ghaznavids and imprisoned in India in 1025, where he died (in 1032), after which his followers migrated to Persian Iraq and Azerbaijan, areas that would serve as an ideal location for their flocks. Ibn al-Athīr notes some two thousand Iraqiya migrated towards Mount Balkan (modern Uly Balkan) in the west of modern Turkmenistan, whilst those reported raiding in Rayy were five thousand strong.

THE I PERIAL ARMY under Diogenes was based in Theodosiopolis probably until the high summer of 1071, before advancing into “unfriendly territory” and having to rely on their two-month supplies that the emperor had ordered his troops to carry for the forthcoming march to Lake Van. Yet, it would seem inconceivable that Romanos was unaware that Alp Arslan was assembling an army. As we saw in the previous chapter, we must assume that, in June or July, the Byzantine leader was receiving some very bad or even misleading intelligence reports that would have a disastrous effect on his strategic decisions for the following stage of the campaign.

Strategic Miscalculations

Sometime in late July or early August, Diogenes's overconfidence led him to make a critical strategic error. After leaving Theodosiopolis to march southeast to Manzikert via Xnis/Xinus (modern Hinis, Erzurum province), a distance of around 230 km, the emperor decided to divide his forces because, according to the sources, he was under the impression that the Turks and Daylami defenders of Manzikert would be easily over-come. Matthew of Edessa reports that Diogenes also ordered the dispatch of some ten or twelve thousand troops to the “Abkhazes,” which is a rather dubious report in my view, although we should not exclude the possibility of sending smaller detachments to Georgia to procure supplies. Other sources also account that the emperor dispatched separate detachments of troops to march to the strategic fortress-city of Khliat that con-trolled the western approaches to the region. These detachments comprised of “Uze mercenaries and the Franks under Rouselios…[sent] towards Khliat to forage for pro-visions,” and because the emperor “considered them [Turkish defenders] of no great importance since they were few in number, he detached another not insignificant por-tion 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 magistros Trachaneiotes “consisted of select troops, difficult to withstand, taking the initiative in close combat and other types of battle and ready to face dan-ger, and far more numerous than the soldiers retained by the emperor.”

IT HAS BEEN described as a dreadful day, a most shattering and shameful defeat, a debacle, an unmitigated catastrophe, the breaking of Byzantium, the beginning of the end of medieval Hellenism, and the day that opened the gates of Anatolia to the Turks, while being pondered over whether or not it was “un désastre militaire” for the Byzantine army of the late 1060s. Whatever the case, the battle that was fought on August 26, 1071, between the forces of the Byzantine emperor Romanos IV Diogenes and the Seljuk sultan Alp Arslan is—rightly so—regarded as one of the most significant turning points in medieval history. In recent years, an increasing number of scholars specializing in the period have lent their voices in support of the view that the Battle of Manzikert was not the real military disaster as it was once portrayed, framing the upcoming collapse of the Byzantine rule in Anatolia that followed as the direct outcome of two factors: Byzantine aristocratic infighting and civil war, and the increased involvement of the Seljuks in these civil wars that “eased” their penetration into Anatolia.

Indeed, it is also my firm conviction that Romanos Diogenes's “Manzikert campaign” of 1071 may have been a strategic failure for the emperor, but it was not a tactical disaster. It should have only confirmed Turkish domination of the Armenian highlands in eastern Asia Minor, rather than almost all of Anatolia, which itself was overrun by the Turks within eight years. Therefore, even though no great military figure is said to have died in the battle, and it seems clear that the actual losses incurred were limited to the emperor's immediate retinue, the decisive outcome of the Battle of Manzikert ushered in an element of chaos in the geopolitical history of the Byzantine Empire in the 1070s: the civil wars of the early 1070s that followed the battle marked the political and military collapse of the empire in Anatolia, the Balkans, and Italy as well as central and eastern Asia Minor.

In the introductory part of this book found in the first three chapters, I examine the Christian and Muslim narrative sources for the battle strictly from a military perspective, and I reach several conclusions regarding the value of these sources for the his-tory of warfare in Anatolia and the Eastern Mediterranean in the eleventh century.