Here one could see a dreadful sight: those celebrated Roman regiments who had brought both East and West under their sway, consisting now of only a handful of men - and men, moreover, bowed down with poverty and ill-health, no longer even fully armed. Instead of swords and other weapons they held, as the Bible has it, only pikes and scythes. And this was not even in time of peace. Yet because it was so long since any Emperor had fought here they lacked war horses and equipment of every kind. And since they were considered weak and cowardly and of no serious use they had received no subsistence money, nor their customary allowance to buy grain. Their very standards rang dully when struck, and looked dirty and as if blackened by smoke; and there were few to care for them. All this caused great sadness in the hearts of those who saw them, when they thought upon the condition from which the Roman armies had come, and that to which they had fallen.
John Scylitzes
Within weeks of Isaac's death it had become clear to all with eyes to see that his brief reign had constituted only a momentary pause in the imperial decline. This had begun immediately on the death of Basil II in 1025, with the accession of his hopeless, hedonist brother; it had continued all through the long, unedifying reigns of Zoe, her husbands, her sister and her adoptive son; and now, under Constantine X Ducas -arguably the most disastrous ruler ever to don the purple buskins - it reached its nadir. Not that there was anything evil or malevolent about Constantine. He was, as we have seen, the close friend, former pupil and to a certain extent the creature of Michael Psellus, on whose advice Isaac had named him his successor; he was a scholar and an intellectual, and -by Byzantine standards, which would certainly not be ours - a superb orator. Finally, he was a scion of one of the oldest and richest families of the military aristocracy. Had he but remained true to his background, had he continued Isaac's work for the eight years that he was to reign, building up the army in preparation for the challenge that so obviously lay ahead, the situation might even at this late stage have been saved. But Constantine X was not one of nature's soldiers. He preferred the ease and comforts of Constantinople, spending his time in learned discussions and the drafting of intertninable dissertations on the finer points of law. And the price that the Empire paid for him was heavy indeed.
Once again the bureaucracy was all-powerful, operating on a scale unmatched anywhere else (with the possible exception of China) for several centuries; for it has to be remembered that the Byzantine Empire, absolute monarchy though it might be, ran its economy on distinctly socialist lines. Capitalism was allowed, but rigidly controlled at every stage; production, labour, consumption, foreign trade, public welfare and even the movement of population were all firmly in the hands of the State. The consequence was a vast army of civil servants, taking its orders theoretically from the Emperor - though effectively, more often than not, from Psellus and his friends - and inspired, so far as one can see, by one overriding principle: to curb - if not actually to destroy - the power of the army. In the past seventeen years, they might have argued, the Empire had experienced three military insurrections: two had been quelled more by luck than anything else, the third had succeeded. It followed that the army must be humbled, and reduced to a proper state of subordination. It must be starved of funds, the authority of the generals must be limited, the former peasant-soldiers - many of whom had followed government advice and bought their exemption from military service - must be progressively replaced by foreign mercenaries.
What Constantine X and his government of intellectuals could never apparently understand was, first, that these were the very measures most likely to provoke further coups; second, that mercenaries were by their very nature unreliable, being loyal to their paymasters only for as long as they received their pay, or until someone else offered them more; third, and most important of all, that the enemy — the most formidable enemy that Byzantium had seen since the appearance of the Saracens 400 years before - was at the gates.
That enemy was a people who, being relatively new arrivals on the scene, have so far received only a passing mention in this book. The Seljuk Turks first appear as a distinctive tribe in the latter half of the tenth century in Transoxania, that region of Central Asia which lies between the rivers Oxus and Jaxartes south-east of the Aral Sea, where they quickly adopted the prevailing faith of Islam. At this time they were still entirely nomadic and leading a life of brigandage: fighting with neighbouring tribes, pillaging and plundering wherever the opportunity arose, finding amid the constant warfare of the local princes plenty of employment for their tough little war ponies, their swords and above all their bows, which they could string in the saddle; shooting as easily to the rear as to the front, they seldom wasted an arrow. By 104 j, under their leader Tughrul Bey, they had spread across Persia; ten years later they were to make themselves masters of Baghdad, establishing a protectorate over the moribund Abbasid Caliphate and proclaiming Tughrul 'Sultan and King of East and West'.
The Caliphate, however, had never been their ultimate objective. Still less had the Byzantine Empire, whose existence had always been accepted by the rulers of Islam. This history has already recorded raids and incursions aplenty, in both directions across the frontier; but the idea of annihilating Byzantium would have struck the Seljuk Sultans as completely unrealistic, even ridiculous. The final goal on which their attention was fixed was Fatimid Egypt, whose Empire now extended across Palestine and Syria as far as Aleppo. As orthodox Sunni Muslims, blazing with all the fervour of recent converts, they detested these Shi'ite upstarts, who represented in their eyes not only unspeakable heresy but - since they had actually dared to set up a rival Caliphate in Cairo - a rupture in the fundamental unity of Islam. They knew that the Fatimids would not rest until they had taken Baghdad; and they were determined to destroy them before they had a chance to do so. First, however, there were certain matters to be settled nearer home; and the most important of these was Armenia.
From 1045 onwards, according to the agreement that John Smbat, King of Ani, had made with Basil II nearly twenty years before,1 the larger part of Armenia was in Byzantine hands. Its annexation after John's death had been virtually the only diplomatic success of Constantine Monomachus; at the time he had made much of it, but it would have been better if he had left it alone - particularly since his subsequent policy, and that of Constantine Ducas, towards Armenia could hardly have been more short-sighted. The Empire's reason for
1 See p. 264.
acquiring this great mile-high mountain barrier on its north-eastern frontier can only have been strategic; yet one of Constantine's first acts after its acquisition had been to institute a fierce religious persecution of the staunchly monophysite Armenians, the surest possible way of turning them against him. His second successor Constantine X, while maintaining the persecution, was to go still further in idiocy. Armenia maintained a local militia of some 50,000 men, for the maintenance of which certain imperial taxes were remitted; in his constant search for new sources of money the Emperor ordered the taxes reimposed and the militia disbanded.
Thus it was that Byzantium lost an invaluable buffer state and gained instead, not an Armenian bulwark as it had hoped, but what might a century ago have been called an Armenian Question — a disaffected and discordant minority within the Empire which created more problems than it solved. The Armenian princes, left to themselves, would have put up as stiff a resistance to the Muslim invaders as they always had; now, demoralized and resentful, they found themselves wondering whether even conquest by the Turks would prove appreciably worse than their present subjection to the Greeks.
Tughrul Bey was not slow to turn this vastly improved situation to his advantage. His first attack on Vaspurakan as early as 1046 was a failure: the Byzantine Governor deliberately left a bait in the form of an undefended camp and ambushed the Turks as they were plundering it.-Two years later, however, his unruly half-brother Ibrahim Inal took advantage of the temporary removal of Byzantine troops at the time of Leo Tornices's revolt and overran the city of Ardzen. The Armenian historian Matthew of Edessa speaks of 150,000 massacred, and goes on to describe 'the sons taken into slavery, the infants smashed without mercy against the rocks, the venerable old men abased in public squares, the gentle-born virgins dishonoured and carried off'; Matthew doubtless exaggerates, but the Seljuk sack of a wealthy city cannot have been a pleasant sight.1 Thenceforth the raids continued almost annually. At one point Constantine Monomachus, obliged to withdraw troops from the East in order to deal with a more immediate threat from the Balkan Pechenegs, concluded a truce with Tughrul, but it did not last long: in 1054 the Seljuk Sultan personally led an expedition which ravaged
1 The survivors are said to have escaped to the neighbouring city of Theodosiopolis, which they renamed after their old home, Ardzen er-Rum (Ardzen of the Romans) - a name which over the years was corrupted into the modern Erzurum.
northern and central Armenia and the plain of Erzurum, pressing on to within some fifty miles of Trebizond itself. Although his entry into Baghdad in 1055 provided a brief interval of relief, the old pattern was resumed all too soon; now, too, the sufferings of the local populations were increased by the activities of the Turkomans - Turks who, though outwardly Islamicized, had never abandoned their nomadic habits, refused to accept the authority of the Sultan and cheerfully continued the brigand life of their ancestors.
Tughrul died in 1063, and was succeeded - after a good deal of family strife - by his nephew Alp Arslan, the son of his brother and co-ruler Chagri. Alp Arslan's moustaches were said to have been so long that they had to be tied behind his back when he went hunting; apart from that, the chroniclers tell us little about his personal appearance. As to his character, their accounts differ. Matthew of Edessa predictably calls him a drinker of blood, and Aristakes considers him one of the forces of Antichrist; in the opinion of Michael the Syrian, on the other hand, he ruled justly and well. The Arab historian Ibn al-Adim tells a story which suggests that he did not invariably observe the Prophet's strictures against the drinking of wine, but this failing was not unusual among Muslim princes of his day; all that we know of him for certain is that he was a superb commander in the field. At the time of his accession he was about thirty-three. Early the following year, 1064, he led a huge expedition against Armenia and besieged its capital, Ani.
No traveller visiting what remains of Ani today can fail to catch his breath in sheer astonishment at the splendour of the site: the towering walls, still partially standing, the rolling plain beyond it from which rise the ruins of some of the most magnificent churches of their time (Matthew of Edessa claims that there were a thousand and one of them) and - invisible until one is at the very brink - the sudden chasm formed by the river nowadays known as the Arpa Cay and one of its tributaries, thanks to which the city enjoyed one of the strongest defensive positions of any in the region. This was, however, of little use against the Seljuks. Unlike many of its neighbours Ani put up a show of resistance, holding out for twenty-five days before it surrendered. At the last moment its inhabitants are said to have sent out all their loveliest maidens and most handsome young men in an attempt to avoid a sack; but Alp Arslan was, as usual, merciless. The Arab historian Sibt ibn al-Gawzi quotes a purported eye-witness of what took place:
The army entered the city, massacred its inhabitants, pillaged and burned it; leaving it in ruins and taking prisoner all those who remained alive... The dead bodies were so many that they blocked all the streets; one could not go anywhere without stepping over them. And the number of prisoners was not less than 50,000 souls. I was determined to enter the city and see the destruction with my own eyes. I tried to find a street in which I would not have to walk over the corpses; but that was impossible.
Even the Armenian historians admit that the fall of Ani was largely due to the low morale of its inhabitants; there seems, too, to have been little love lost between the local population and the Byzantine officials. Be that as it may, the city constituted the only serious interruption to the Seljuks' progress. From there they were able to advance deep into the centre of Anatolia, where the state of the Byzantine defences by that time can be judged from the fact that in 1067 they penetrated as far as the Cappadocian Caesarea, which was subjected to another merciless sack, and to within a hundred miles of Ancyra (Ankara) before withdrawing. More shameful still, scarcely a sword was lifted against them.
That same year saw the death of Constantine X. Even on his deathbed he had done his best to perpetuate his catastrophic policies, obliging his young wife Eudocia to swear that she would not remarry and demanding from all those around him written commitments that they would recognize only a member of his own family as his successor. In this the dying Emperor was almost certainly abetted by his brother, the Caesar John Ducas, and by Psellus himself, who must have known that he would get short shrift if a member of the military aristocracy should come to power: he had already once - to his deep disgust - been exiled to a monastery and was determined not to repeat the experience. But by this time the fate of Caesarea was known and alarm was widespread throughout Constantinople. Even among the civil bureaucracy there were many who saw that the Empire could be saved only by a radical change of policy. The problem was that, short of a coup d’etat, the only way to ensure the kind of Emperor that was required was for Eudocia to marry him - the one thing that she had sworn not to do.
The Empress herself was perfectly willing to take another husband if she could be freed from her oath, but for this she needed a dispensation from both the Patriarch and the Senate. Unfortunately, since the former was Psellus's old friend John Xiphilinus - with-him virtually a founder member of the new bureaucratic party — while the latter was almost entirely composed of Constantine's appointees, her chances appeared slim. Eudocia, however, was nothing if not resourceful. Aided by one of the Palace eunuchs, she put it about that she was seriously considering marriage to the Patriarch's brother, a rather dashing man-about-town; and Xiphilinus, who knew that his brother had always enjoyed considerable success with women, believed the story. He therefore summoned the Senators one by one, explaining to each in turn the iniquity of the oath that Constantine had forced on his widow. It was, he told them, unlawful and unjust, an attempt by one man to gratify his own personal vanity without thought for the good of the State. This latter, he continued, could be achieved only by the Empress's remarriage to some nobleman of distinction who would then, through her, inherit the imperial crown. Certain of the Senators agreed wholeheartedly, others needed persuasion of one kind or another; but eventually all gave their consent. Now at last Eudocia could announce her true intention: she would marry, not the Patriarch's brother, but a man who seemed, more than anyone else, to epitomize the Anatolian military aristocracy. His name was Romanus Diogenes.
The Emperor Romanus IV, who was probably married and certainly crowned on i January 1068, came of an old and distinguished military family with vast estates, in Cappadocia. Still only in his early middle age, he had already served as Governor of Sardica in which capacity he had won several victories over the Pecheneg invaders; while in Bulgaria, however, he had been accused of conspiracy against the throne. The sentence of death pronounced on him after his return to the capital was subsequently commuted to one of exile, but this too was lifted on the death of Constantine X; Romanus was released and brought before the Empress, who is said to have wept when she saw him. The significance of her emotion is uncertain; we can hardly attribute it to love at first sight, although Romanus - according to Michael Attaleiates, who served with him in the field and knew him well - was quite exceptionally good-looking, with broad shoulders and bright, flashing eyes.1 In any event, tears or no tears, the primary reason for the marriage was unquestionably one of policy: to raise a soldier to the throne and thus to save the State. Eudocia's earlier inclination had been towards the other leading general of the day, a certain Nicephorus Botaneiates, of whom we shall be
1 Attalciates also implies that Eudocia anyway disliked sex; but since her only experience of it had been with her first husband this - from what we know of him - is not altogether surprising. Somehow, in any case, she had contrived to bear him three sons.
hearing more in due course; as a result of this single audience, however, she changed her mind. Romanus, who had left for Cappadocia immediately afterwards, had not even reached his home before he was summoned to return; the marriage and coronation followed a few days later.
It is impossible not to feel sorry for Romanus Diogenes. Although an arrogant man with a strong sense of his own importance, he was also an able and hard-working administrator and a brave soldier; and he fully recognized the gravity of the Seljuk menace. Already once he had risked his life to overthrow an Emperor whom he believed to be leading the Empire to its destruction; now that he found himself to be that Emperor's successor, he set to with spirit and determination to restore imperial fortunes - and it was not his fault that he failed. Wherever he looked, whatever he tried, the dice were loaded against him. In Constantinople, he had to contend with Psellus - who loathed him - and the Ducas family, who bitterly resented his rise to power and were resolved, sooner or later, to bring about his destruction. In the field, he found a hopelessly demoralized army composed largely of vacillating mercenaries, ill-fed, ill-equipped and frequendy on the point of mutiny. Both 1068 and 1069 saw him leading military expeditions to the East, and it is a remarkable tribute to his powers of leadership that he achieved some degree of success - particularly in Syria, where he captured Hieropolis (the modern Mambij) and appreciably strengthened the Byzantine position; but the account of those expeditions by Attaleiates, who accompanied him as a member of his military tribunal, makes almost unbearably depressing reading. The Emperor's personal courage, his determination not to be defeated by conditions and circumstances that would have driven most generals to despair, shines out like a beacon in the darkness; for the rest, we have a seldom-interrupted saga of frustration and disorganization, of cowardice and chaos.
Fighting in the East continued through the first half of 1070, and was followed by a truce. This campaign was not, however, under Romanus's personal command. Every time that he left the capital there was a risk that his enemies - Psellus, the Ducas and their followers - might attempt a coup,and internal tensions during that year were such as to rule out any prolonged absence from Constantinople. In fact, his enforced stay had one major advantage: it enabled him to devote his energies to improving the lot of the army, settling arrears of pay, procuring new equipment, instituting training programmes and generally making good at least some of the damage wrought by Constantine IX and X and their predecessors. It also gave him time to recruit new forces. The experience of the past two years had shown him that the existing army of the East, even when in top fighting form, would never be sufficient to inflict on the Seljuk hordes a defeat decisive enough to ensure the safety of Armenia and, through Armenia, of Anatolia. The truce he had recently concluded with Alp Arslan was being continually broken by the raiding Turkomans and was, so far as he was concerned, a dead letter; he was therefore already planning a campaign for 1071 in which he would be able to throw some 60,000 to 70,000 men1into the fray.
That expedition crossed the Bosphorus in the second week of March 1071, and immediately headed eastward. Once again, Michael Attaleiates was present; and his version of the events of that summer, while it leaves a tantalizing number of points unexplained and questions unanswered, remains by far the most detailed and trustworthy account that has come down to us. As always, he makes no secret of his admiration for the Emperor; he notes, however, that after the army had gone some zoo miles Romanus's demeanour seemed to change:
He became a stranger to his own army, setting up his own separate camp and arranging for more ostentatious accommodation. When the army crossed the river Halys, for example, he did not cross over at the same time but remained behind and spent some days at a fortress which had recently been built at his command. Soon afterwards he issued an order to separate his private possessions from those of the army.
Scylitzes, whose narrative for this period clearly derives from Attaleiates, suggests that Romanus had been perturbed by various supposed portents, among them the sudden breakage of the central pole of his tent and an unexplained fire in which he lost much of his personal equipment, together with several of his best horses and mules. It may well be so — this was a superstitious age and the Byzantines were even more credulous than most; there seems little doubt, at any rate, that to his natural arrogance the Emperor now added several other faults which threatened to become liabilities in the days to come: remoteness, ill
1 Such figures as these are, as always, highly conjectural. Byzantine sources give no information, Muslim ones range from 200,000 to 600,000 and Matthew of Edessa claims, grotesquely, 1 million. Most modem historians seem to agree that the estimate given above is the most probable, though the true number could have been as high as100,000.
temper, impatience with advice and a streak of cruelty which we have not seen before.1
Strangely enough, while the immense Byzantine army was marching across Anatolia, Alp Arslan was heading in a completely different direction. Being totally unable to control the Turkoman raiders and having frequently disclaimed all responsibility for their activities, he believed the truce of the previous year to be still in force; he had therefore decided that the moment had come to fulfil his long-cherished ambition of destroying the Fatimid Caliphate. Late in 1070 he left his headquarters at Khurasan, captured the Armenian fortresses of Manzikert and Archesh, then marched south-west to Amida and on to Edessa where, towards the end of March, he drew up his army beneath the walls. Scarcely had he begun the siege when he received a message from the Emperor proposing a renewal of the truce and the exchange of Manzikert and Archesh for Hieropolis in Syria, which he had captured three years before. He accepted and, leaving Edessa untaken, continued his march. Six weeks later he was besieging Aleppo when there arrived another envoy from Romanus, now in Armenia, repeating the offer but this time presenting it in distinctly threatening terms.
If Romanus had received Alp Arslan's previous acceptance of his offer, he had played the Sultan false. Even if he had not, he was now in a position of strength and putting forward what amounted to an ultimatum. Seeing at once that he had no choice but to abandon his Fatimid expedition, Alp Arslan turned about and made straight back for his homeland. Such was his haste that he failed to take adequate precautions in recrossing the Euphrates: many of his horses and mules were carried away by the current and drowned. But this hardly mattered to him. He knew that he would anyway have to raise a very much larger army before he could meet the Emperor on his own terms. Sending his vizier Nizam al-Mulk ahead to raise troops in Azerbaijan, he himself headed for Khoi (now Khvoi), between Lake Van and Lake Urmia, picking up some ten thousand Kurdish cavalry en route; there he was joined by his new army and set off in search of his foe.
1 At one point of the campaign a Byzantine soldier accused of stealing a donkey from a local inhabitant and brought before the Emperor was sentenced to have his nose severed - a punishment that had fortunately gone out of fashion in the eighth century. It was when Romanus upheld this sentence even after the poor man had invoked the intercession of the victory-bringing icon of the Holy Virgin of Blachernae (which was always carried by the basileus into battle) that Attaleiates reports his first presentiment of the divine vengeance that would follow.
Romanus had meanwhile encamped near Erzurum, where he had split his army into two. The greater part he dispatched under the command of his general Joseph Tarchaniotes to Khelat (now Ahlat), a strongly defended Seljuk-held fortress a few miles from the northern shore of Lake Van, while he himself - together with his other principal commander, Nicephorus Bryennius - set off with the remainder for the litde fortress-town of Manzikert, which he believed would offer no serious resistance. He was, as it turned out, perfectly right: the garrison gave in without a struggle. Tarchaniotes, on the other hand, was less fortunate. We do not know precisely what happened. Later Muslim historians refer to a pitched battle in which Alp Arslan scored a decisive victory, but there is no mention of such an engagement in any Byzantine source — the most trustworthy, Attaleiates, simply reporting that the news of the Sultan's arrival with his new army was in itself sufficient to send 'the scoundrel' Tarchaniotes into headlong flight, his men after him. They never stopped until they reached Melitene on the Euphrates, and did not reappear again during the whole campaign.
But it cannot, surely, have been as simple as that. Joseph Tarchaniotes was a highly respected general, in command of a force of 30,000 to 40,000 - larger, very probably, than the entire Seljuk army. If we reject the Muslim version - that he was soundly beaten in the field - we are left with various other possibilities, all in varying degrees improbable: that he was angry with Romanus, whom he had strongly advised not to split the army, and determined to prove him wrong whatever the cost; that Alp Arslan had taken his army by surprise, and that since it had no opportunity to regroup a general sauve-qui-peut was the only answer; or -most intriguing of all - that Tarchaniotes was a traitor, a tool of the Ducas, who had actually set out from Constantinople with the deliberate intention of abandoning his Emperor when the moment came. If such a theory seems at present far-fetched, it may seem less so by the end of this chapter; and it would also go a long way to explain another mystery - why it was that no word of what had happened was sent to Romanus, only thirty miles away at Manzikert. But however much we may speculate about the cause of the rout, its consequence is all too clear: that by the time the Emperor finally met the Seljuks on the field of battle, more than half his army had deserted him.
Romanus Diogenes had captured the fortress of Manzikert; he did not have long, however, to savour his triumph. On the very next day some of his men, out on a foraging expedition, were set upon by mounted Seljuk bowmen and suffered heavy casualties. The Emperor, assuming that he had to deal with nothing more serious than a small band of marauders, sent out a small detachment of troops under Bryennius — and flew into a fury when he received, an hour or two later, an appeal for reinforcements. After some hesitation he dispatched a rather larger force under an impulsive Armenian named Basilacius; they tried to pursue the bowmen, but were trapped by them and quickly surrounded. Basilacius himself was captured, but few of those who followed him escaped with their lives. Bryennius, riding out once again - this time with the entire right wing of the army behind him - in an attempt to rescue his rescuers, found himself confronted with what must have been a considerable proportion of the entire Seljuk army. He and his men retreated in good order to the camp, but not before he had received no less than three wounds, two from arrows in the back and one, in the chest, from a lance. Fortunately all three proved to be superficial, and he was able to continue the campaign.
That night there was no moon - and little sleep for the Byzantine army. The Seljuks kept up an unremitting pressure, loosing hail after hail of arrows and generally causing so much tumult and confusion in the darkness that again and again they were thought to have broken down the defences and overrun the camp. It was a pleasant surprise to everyone the next morning to find that the palisades had held - but a most unpleasant shock to learn that a large contingent of Uz mercenaries had defected to the Seljuks; there were several other Turkic units in the army, any or all of which might at any moment follow their example. In such circumstances, and with half his army — including one of his best generals - vanished without trace, one might have expected the Emperor to welcome the delegation that arrived a day or two later. It came, officially, from the Caliph in Baghdad (though in fact it was obviously sent by Alp Arslan in the hopes that it might fare better than one from himself) and it proposed a truce.
Why, we may ask, did the Sultan want one? Almost certainly because he was far from sure of victory. We know that shortly before going into battle he spoke of possible martyrdom in the field; he dressed himself in white in order, as he explained, that his garment might also serve as his shroud; and he enforced oaths from those around him that his son Malik-Shah should succeed him in the event of his death. In the past the Seljuks had always relied upon their skill in irregular warfare - the warfare of raids, ambushes and surprise attacks. They disliked pitched battles, which they avoided whenever possible; and recent Byzantine humiliations notwithstanding, they still retained a healthy respect for the imperial army.1 Besides, was there, from the Sultan's point of view, any real reason to fight? The one serious difference of political opinion concerned Armenia, which had considerable strategic value for both him and Romanus. If only the two of them could agree on a mutually acceptable division of Armenian territory, both armies could remain intact - and Alp Arslan could turn his attention back to what really interested him: the Fatimids.
But the Emperor's determination remained firm. This was, he knew, his only chance of freeing his Empire once and for all from the Turkish menace. Alp Arslan was only a few miles away with his full army; he himself, despite the disaster of Khelat, still commanded a force which, even if not quite its equal in size, was certainly larger than he was ever likely to muster again. Finally, if he were to return to Constantinople without having so much as engaged the Seljuks on the field, what chance would he have of keeping his throne - or indeed his life - against the intrigues of the Ducas? He dismissed the embassy with the minimum of courtesy and prepared for battle.
It is a curious and somehow frustrating fact that neither the date nor the location of one of the most decisive battles of the world can be universally agreed. Muslim historians are unanimous that it took place on a Friday, and the month was indubitably August; but scholars still argue whether it was fought on the 5 th, 12th, 19th or 26th. Most European historians seem to have opted for the 19th, but in doing so they have ignored an important clue: the fact that according to Michael Attaleiates - who was there - the second or third night before the battle was moonless (aselenos). Now in August 1071 the full moon was on the 13th (Julian calendar), which means that the nights would still have been bright on the 16th and 17th; they would have been far darker on the 23 rd and 24th, when the moon would have been visible only as a thin crescent, an hour or two before dawn. Leaving aside the possibility that Attaleiates simply meant that the sky was overcast - something extremely improbable in that place and season — we cannot but conclude that it was
1 But this raises another question: if the Scljuks had indeed defeated the Byzantines in pitched battle at Khelat, would the Sultan not have been rather more optimistic? The more one thinks about it, the more one comes to suspect that that battle never took place.
on Friday, 26 August that the fate of the Byzantine Empire was settled.1
As to the location, we know that the battle was fought on fairly level steppe, within a mile or two of the fortress of Manzikert (now marked by the modern Turkish town of Malazgirt). The chronicler Nicephorus Bryennius, grandson and namesake of Romanus's general and another valuable source, adds that in the closing stages the Byzantines ran into ambushes, which suggests - since steppes by their very nature afford few hiding-places - that there must also have been rougher, hillier country close by. Now Armenia is a mountainous country, but there is indeed just such a steppe, some three or four miles across, extending for perhaps ten miles on a south-west-north-east axis immediately to the south and east of the town. Beyond it a line of foothills, cut through with gullies and ravines — ideal ambush country - soon gives way once more to the mountains. Somewhere in those forty-odd square miles the two armies drew up opposite each other, early that Friday afternoon, and battle was joined.
Or was it? The fact of the matter is that Manzikert, for all its historical significance, was until its very final stages hardly a battle at all. Romanus had formed up his army according to the traditional army manuals, in one long line several ranks deep, with the cavalry on the flanks. He himself took the centre, with Bryennius on the left and, on the right, a Cappadocian general named Alyattes. Behind was a substantial rearguard composed, we are told, of the 'levies of the nobility' - in fact the private armies of the great landowners - under the command, somewhat surprisingly, of Andronicus Ducas, son of Caesar John Ducas and nephew of the late Emperor. This young man seems to have made little secret of his contempt for Romanus, and the wonder is that the latter should have ever allowed him to participate in the campaign at all. Presumably he thought it safer to have him under his eye as a potential hostage rather than to leave him to stir up trouble in Constantinople; if so, it was the gravest mistake of his life.
All through the afternoon the imperial army advanced across the steppe, but instead of coming forward to meet them the Seljuks steadily withdrew in a wide crescent, leaving the initiative to their mounted archers who galloped up and down on the Byzantine flanks, showering
1 For this information, as for much else in this chapter, I am indebted to the late Alfred Friendly, whose book The Dreadful Day: The Battle of Manzikert, has proved invaluable.
them with arrows. The infuriated cavalry would then break line and pursue them into the foothills - straight, it need hardly be said, into the carefully prepared ambushes; but for the increasingly frustrated Emperor in the centre there remained, where the enemy should have been, an empty void. On and on he rode, apparendy convinced that if he continued as far as the mountains he could somehow force his antagonists to turn and fight; then, suddenly, he noticed that the sun was fast declining and that he had left his camp virtually undefended. There was no point in further pursuit - if, indeed, he was actually pursuing anyone. He ordered the imperial standards to be reversed - the recognized signal for turning back - and wheeled his horse.
This was the moment for which Alp Arslan had been waiting. From his observation post in the hills above he had watched Romanus's every move; now and now only, he gave the order for the attack. As his men poured down on to the steppe the Byzantine line broke in confusion; some of the mercenary units, seeing the reversed standards and not understanding their true significance, assumed that the Emperor had been killed and took to their heels. Meanwhile the Seljuks cut across immediately behind the broken line, separating it from the rearguard. At this point that rearguard should have justified its existence by moving forward, squeezing the enemy between it and the forward line and preventing its escape. Instead, Andronicus Ducas deliberately spread the word among his troops that the Emperor was defeated and the battle lost. He and they thereupon fled from the field and, as the panic spread, more and more contingents followed them. Only the troops on the left wing, seeing the Emperor in difficulties, rode to his rescue; but the Seljuks bore down upon them swiftly from the rear and they too had to flee.
Romanus stood his ground, his personal guard around him, calling in vain on his troops to rally. But the chaos and confusion were too great. As Attaleiates describes it:
Outside the camp all were in flight, shouting incoherently and riding about in disorder; no one could say what was happening. Some maintained that the Emperor was still fighting with what was left of his army, and that the barbarians had been put to flight. Others claimed that he had been killed or captured. Everyone had something different to report...
It was like an earthquake: the shouting, the sweat, the swift rushes of fear, the clouds of dust, and not least the hordes of Turks riding all around us. Depending on his speed, resolution and strength, each man sought safety in flight. The enemy followed in pursuit, killing some, capturing others and trampling yet others under their horses' hooves. It was a tragic sight, beyond any mourning or lamenting. What indeed could be more pitiable than to see the entire imperial army in flight, defeated and pursued by cruel and inhuman barbarians; the Emperor defenceless and surrounded by more of the same; the imperial tents, symbols of military might and sovereignty, taken over by men of such a kind; the whole Roman state overturned - and knowing that the Empire itself was on the verge of collapse?
Who survived? Effectively, those who took flight in time. We cannot blame the Armenians for doing so; they had little love at the best of times for the Greeks, who had conquered their country and were even now persecuting their families for the faith that they had always upheld. To the mercenaries we need feel less sympathetic; admittedly they had no emotional loyalty to the Empire, and they understandably resented the Emperor's discrimination against them, together with his unconcealed preference for his native troops. But they were under contract, their wages had been paid, and - though their conduct was probably no worse than that of other soldiers of fortune the world over - they might have shown a little more spirit than in fact they did. The only real villains were the 'levies of the nobility' who formed the rearguard, and their commander Andronicus Ducas. Their shameful flight was probably due to treachery rather than cowardice, and was not a jot the more excusable for that.
There was another survivor too: Romanus Diogenes himself. Left almost alone, he had refused to flee; he fought like a lion to the end. Only when his horse was killed under him and a wound in his hand made it no longer possible for him to hold his sword did he allow himself to be taken prisoner. His captors must have known who he was, but they gave him no special treatment. All night he lay among the wounded and dying. Only the following morning was he brought before the Sultan, dressed as a common soldier, and in chains.
Few chroniclers of the period, Greek or Islamic, have been able to resist giving an account of the interview that took place between the victorious Sultan and the defeated Emperor on the morning after the battle. What is surprising, however, is not the number of these accounts but their similarity. At first, we are told, Alp Arslan refused to believe that the exhausted captive who was flung at his feet was indeed the Emperor of the Romans; only when he had been formally identified by former Turkish envoys and by his fellow-prisoner Basilacius did the Sultan rise from his throne and, ordering Romanus to kiss the ground before him, plant his foot on his victim's neck.
It was a symbolic gesture, nothing more. He then immediately helped Romanus to his feet, bade him sit down at his side and assured him that he would be treated with all the respect due to his rank. For the next week the Emperor remained a guest in the Turkish camp, eating at the Sultan's table; never once did Alp Arslan show him anything but friendliness and courtesy - although he frequently condemned the faithlessness of those who had deserted him in his hour of need and, we are told, occasionally permitted himself a few gende criticisms of Byzantine generalship. All this was, of course, in the highest tradition of Islamic chivalry; we are inevitably reminded of the similar stories told of Saladin a century later. But it was also sound policy on the part of the Sultan: how much better, after all, that Romanus should return safely to Constantinople and resume the throne than that it should be taken over by some inexperienced and headstrong young man of whom he would know nothing and who would think only of revenge.
Considering the circumstances, the peace terms were both merciful and moderate. The Sultan demanded no extensive territories - not even Armenia, to which he believed that he possessed at least as much right as did the Empire. All he asked was the surrender of Manzikert, Antioch, Edessa and Hieropolis, together with one of the Emperor's daughters as a bride for one of his own sons. There remained the question of a ransom. Alp Arslan first suggested that 10 million gold pieces might be appropriate; but when Romanus objected that since the fitting out of his great expedition the imperial treasury simply did not possess such a sum, the Sultan willingly reduced his demand to a million and a half, with a further 360,000 in annual tribute. He also took the point that the Emperor should return to Constantinople at the earliest possible moment; there was otherwise a very real danger that he might be dethroned in his absence, in which case his successor would be highly unlikely to recognize the validity of the agreement that had just been made. It was thus only a week after the battle that Romanus set out on his homeward journey; Alp Arslan rode with him on the first stage, and for the remainder granted him an escort of two Emirs and a hundred Mamelukes. He had left his capital as an Emperor; as an Emperor he would return.
Or so, at least, he hoped. His feelings were not however shared in Constantinople, where the news of the defeat had come as the second shattering blow in this most catastrophic of years. The previous April -just a month after Romanus had left for the East - the Normans under Robert Guiscard had captured Bari. Since the days of Justinian Bari had been capital of Byzantine Apulia and headquarters of the imperial army: the largest, wealthiest and best defended of all the Greek dues of the peninsula. By the time the siege had begun, it had also been the only one still remaining under imperial control. The Bariots had resisted valiantly, for no less than thirty-two months; but surrounded as they were by an impenetrable blockade by land and sea, they finally had no choice but to surrender. It was the end, after more than five centuries, of Byzantine Italy.
The reports from Bari had, however, at least been clear; from Manzikert, they were hopelessly confused, creating at the court of Constantinople an atmosphere of uncertainty and indecision. On one point only was everyone — with the possible exception of the Empress herself - in unanimous agreement: that even if Romanus were alive and at liberty he was now defeated and disgraced; there could be no question of his being allowed to continue as basileus. But who was to take his place? Some called for Eudocia to resume the supreme power that she had wielded before her marriage; others favoured Michael, her son by Constantine X - perhaps as a co-ruler with his younger brothers Andronicus and Constantine; yet others saw in the Caesar John Ducas (who now hurriedly returned from Bithynia, whither Romanus had exiled him before his departure) the best hope for the Empire in this moment of crisis. In the event, it was John who acted - though not, ostensibly, on his own behalf. There can be no doubt that he coveted the throne; on the other hand his own faction was not large enough to give a direct attempt upon it any real hope of success. Fortunately his nephew Michael was a weak-willed youth, with whom - once his mother were out of the way - John could do as he liked. Fortunately too he had the Varangian guard behind him. While the rest were still debating what was best to be done, he divided it into two groups. One, under the command of his recently-returned son Andronicus, charged through the Palace proclaiming Michael Emperor; with the other he marched straight to the Empress's apartments and arrested her.
It was all over quite quickly. The terrified Eudocia was exiled to a church she had founded at the mouth of the Hellespont, where she was shortly afterwards tonsured and compelled to take the veil. A similar sentence was passed on Anna Dalassena, sister-in-law of the late Emperor Isaac Comnenus, as a warning to the only other family in the capital from whom trouble might be expected.1 Michael VII Ducas was crowned with due ceremony by the Patriarch in St Sophia. It remained only to deal with Romanus Diogenes.
Romanus's movements after his departure from the Seljuk camp are hard indeed to trace: our authorities are few, and tend more often than not to contradict each other. All that we can deduce with any certainty is that he somehow managed to gather together what remained of his once-great army, with the intention of marching on the capital. John Ducas was, however, ready for him. There seem to have been two battles: one near Dokeia (Tokat) against a force commanded by the Caesar's youngest son Constantine, and one near Adana in Cilicia in which Romanus found himself confronted with the general who had betrayed him at Manzikert, Andronicus Ducas. In both he was defeated. After the second he finally gave himself up to Andronicus, agreeing to renounce all claims to the throne and to retire to a monastery, and receiving in return a guarantee from the new Emperor - endorsed by the Archbishops of Chalcedon, Heraclea and Colonea — that no harm would come to him.
It could perhaps be argued that the decision of Andronicus to mount the ex-Emperor on a mule and to parade him in his degradation the 500-odd miles from Adana to Cotiaeum - the modern Kiitahya - was not actually harmful, though it seems a curious interpretation of the terms of his undertaking. In the light of what happened afterwards, however, the question seems academic. Referring to the archbishops, Scylitzes writes:
Although they wished to help him, they were weak and impotent when cruel and harsh men took him and pitilessly, mercilessly, put out his eyes. Carried forth on a cheap beast of burden like a decaying corpse, his eyes gouged out and his face and head alive with worms, he lived on a few days in pain with a foul stench all about him until he gave up the ghost, being buried on the island of Proti where he had built a monastery.
He was given a rich burial by his wife, the Empress Eudocia, leaving memories of trials and misfortunes too terrible to be told. But in all his
1 Nicephorus Bryennius records that she was made the victim of a show trial, in which forged letters were produced to implicate her in a conspiracy to restore Romanus. Her action in suddenly producing from the folds of her gown an icon of Christ, and bidding those present behold the true judge of the proceedings, was applauded as a splendid theatrical gesture - but cut, we are told, no ice.
misfortunes he uttered no curse or blasphemy, continuing always to give thanks to God and bearing courageously whatever befell him.
Michael Psellus, who always detested Romanus and never loses an opportunity to denigrate him, predictably attempts to justify the blinding:
I am reluctant to describe a deed that should never have taken place; and yet, if I may alter my words slightly [sic], it was a deed that should unquestionably have taken place. On the one hand the scruples of religion, allied to a natural unwillingness to inflict pain, would forbid it; on the other, the political situation and the sudden changes of both parties made it absolutely necessary . .. since the more enthusiastic element in the imperial council1 feared that Diogenes might succeed in his conspiracies and once again cause the Emperor embarrassment.
Nor was that his last insult. A few days before his death in the summer of 1072, Romanus received a letter from his old enemy. It was couched in the friendliest terms, and congratulated him on his good fortune in losing his eyes - a sure sign that the Almighty had found him worthy of a higher light. As he lay in his final agony, the thought must have given him profound comfort.
The battle of Manzikert was the greatest disaster suffered by the Empire of Byzantium in the seven and a half centuries of its existence. The humiliation was bad enough, the performance of the imperial army having been characterized by a combination of treachery, panic and ignominious flight; the fate of the Emperor, too, was unparalleled since the capture of Valerian by the Persian King Shapur I in AD 260, before Constantinople was even thought of. The real tragedy, however, lay not in the battle itself but in its appalling epilogue. Had Romanus Diogenes been permitted to retain his throne, all would have been well: he would have observed the terms of the treaty he had made with his captors, and Alp Arslan - who (let it never be forgotten) had no intention of making outright war on the Empire, far less of conquering it - would have resumed his expedition against Fatimid Egypt. Even had Romanus been succeeded by an Emperor worthy of the name, the damage could easily have been contained: a Nicephorus Phocas or a John Tzimisces - let alone a Basil II — would have restored the status quo in a matter of months, and the Seljuks did not begin any systematic move into Anatolia
1 I.e., Caesar John Ducas. Psellus goes on to say that he gave the order for the blinding without consulting the Emperor Michael.
until the summer of 1073 - two years after the battle. By then they can hardly be blamed for doing so. Michael VII's refusal to accept the obligations of the treaty signed with Romanus gave them a legitimate motive for their action, while the chaos that reigned within the Empire and the collapse of the old defensive system based on military holdings ensured that they met with no resistance.
Thus it came about that tens of thousands of Turkoman tribesmen swarmed into Anatolia from the north-east, and that by 1080 or thereabouts the Seljuk Sultan Malik-Shah1 controlled a broad tongue of territory covering perhaps 30,000 square miles and extending deep into the heartland: an area which, in recognition of its former history as part of the Roman Empire, he named the Sultanate of Rum. The Empire still retained western Asia Minor and its former Mediterranean and Black Sea coasts; but it had lost, at a single stroke, the source of a considerable portion of its grain and more than half its manpower. And it had done so not because of the superior fighting power of the Seljuk Turks, but as the result of its own inefficiency and short-sightedness. The battle through which it had suffered that loss had been directed against an unwilling enemy; it need never have been fought and could easily have been won. Even after defeat, its long-term consequences might have been avoided by judicious diplomacy. But those in power in Constantinople, led by Caesar John Ducas and inspired by the odious Michael Psellus, systematically refused to take the steps that were so obviously necessary. Blinkered by their own smug intellectualism and obsessive personal ambition, they made every mistake, threw away every opportunity offered to them. In doing so they martyred a courageous and upright man who, though no genius, was worth more than all of them put together and could, with their loyalty and support, have saved the situation; and they dealt the Byzantine Empire a blow from which it would never recover.
The reign of Michael VII continued as disastrously as it had begun. The year after Manzikert saw a serious uprising in Bulgaria as a result of which a certain Constantine Bodin, son of Prince Michael of Zeta,2 had
1. The son of Alp Arslan who had been physically attacked by one of his junior commanders and had died of his wounds on 24 November 1071 at the age of forty-one.
2. Formerly known as Dioclea and theoretically a semi-independent principality within the Empire, Zeta. had broken away in about 1055 and had refused to recognize Byzantine overlordship - the first Slav state in the Balkans to do so after the death of Basil II.
been crowned Tsar in the city of Prizren. Thanks largely to the efforts of Nicephorus Bryennius the Empire eventually managed to reimpose control, but at a heavy cost; it was clear that further outbreaks could not be long delayed. Meanwhile Rome was steadily extending its influence beyond the Adriatic, in the lands over which Basil II had formerly claimed suzerainty. In 1075 the legates of Pope Gregory VII crowned a vassal named Demetrius Zvonimir King of Croatia, and in 1077 there came a further blow for Byzantium when Michael of Zeta also received a papal coronation. As the imperial hold weakened, the Pechenegs and Hungarians became increasingly troublesome. Thus, within half a century of Basil's death, his whole magnificent achievement in the Balkans was already crumbling away.
At home the situation was very little better. Inflation was rising, to the point where a gold nomisma would no longer buy a whole measure of wheat, but only three-quarters. Before long the Emperor became known as Michael Parapinaces, or 'Minus-a-quarter', a nickname which stuck with him until his death. Weak and feckless as always, he also allowed himself to fall under the influence of a sinister new arrival on the political scene, the eunuch Nicephoritzes, who took over the effective government of the Empire, cast aside Psellus and the Caesar John and rapidly became the same sort of power in the capital that John the Orphanotrophus had been forty years before. Determined still further to strengthen the centralized bureaucracy of the State, he went further than any of his predecessors by turning the corn trade into a government monopoly, building a vast official granary at the port of Rhaedestum on the Marmara at which all corn shipped to the capital was to be stored until resale. The attempt, predictably enough, proved yet another disaster. The landed proprietors in those areas of Anatolia still in imperial control sustained heavy losses, while the urban consumers found that Nicephoritzes was less interested in ensuring adequate supplies than in increasing state revenues by screwing up the price of bread. This in turn led to a general increase in prices and a further twist of the inflationary spiral.
Then there were the military insurrections. The first of these was inspired by the leader of the Norman mercenaries, a soldier of fortune named Roussel of Bailleul. His fighting record was not unblemished, since he had been involved with Joseph Tarchaniotes in the mysterious events at Khelat; somehow, however, he had charmed his way back into imperial favour - it sometimes seemed that anyone who had betrayed Romanus Diogenes was by definition a friend of his successor - and had later been sent with a mixed force of Norman and Frankish cavalry against Seljuk marauders in Anatolia. Once deep in Turkish-controlled territory he had yet again betrayed his trust and, with 300 loyal followers, had set up a self-declared independent Norman state on the south Italian pattern. Had Michael VII and his advisers thought for a moment, they would surely have realized that, compared with the Turkish tide that threatened to engulf them, Roussel was little more than a mild irritant; instead, so determined were they to liquidate him that they turned to the Seljuks for aid, offering them in return the formal cession of territories that they already held and thus immeasurably strengthening their hold on Asia Minor. Even then Roussel managed to escape; only when an army was sent out from Constantinople under the command of the ablest of the Empire's younger generals, Alexius Comnenus, was he hunted down and brought back in chains to the capital.
But Alexius Comnenus could not be everywhere at once and, owing to the neglect of the army during the previous half-century, experienced generals were in short supply; thus it was that, a year or two later, when the Empire was faced with two new and far more serious insurrections, one in the East and one in the West, Roussel was suddenly released from his captivity and found himself fighting at Alexius's side against the two new pretenders to the imperial throne. The first of these was Nicephorus Bryennius who, having fought with distinction at Manzikert, had been installed as Governor {dux) of Dyrrachium, where he had been principally responsible for putting down the Slav revolt of 1072. No longer prepared to accept the incompetence of Michael Parapinaces and his government - and having learnt that the eunuch Nicephoritzes had listed him for assassination - he raised the standard of revolt in November 1077 and marched to his native city of Adrianople, where he was acclaimed basileus. A week later he and his army were beneath the walls of Constantinople.
His insurrection might well have succeeded had it not been for another, almost simultaneous, rising in the East. Its leader was the strategos of the Anatolikon Theme, Nicephorus Botaneiates. He has made one previous, passing appearance in this story as a prospective husband for the Empress Eudocia, before the arrival of Romanus Diogenes banished the idea from her mind. Presumably because he had doubts about his loyalty, Romanus had deliberately excluded Botaneiates from the Manzikert expedition; the general had returned to his extensive estates in Anatolia, where he had received his present appointment soon after Michael's accession; but now he too - probably, like Bryennius, for the highest motives - took up arms against the Emperor.
Of the two rival claimants, Bryennius was first in the field; Botaneiates was, however, of far nobler birth, being a kinsman of the Phocas and thus a member of the old military aristocracy; he was also the stronger, particularly since he had managed to suborn the Seljuk forces hired by Michael to oppose him. Neither made a direct attack on Constantinople, knowing full well from secret contacts within the capital that popular discontent over rising prices would soon bring matters to a head -which, in March 1078, it did. Riots broke out in every corner of the city. Many government buildings were burnt to the ground, among them Nicephoritzes's new granary. The eunuch himself was seized by the mob and tortured to death. The miserable Michael, lucky to escape with his life, quickly abdicated and withdrew to the Studium, and on 24 March Nicephorus Botaneiates entered Constantinople in triumph. His rival Bryennius was captured and blinded.
It was an inauspicious start to the new reign. Botaneiates had been a competent general, but he knew nothing of politics or statesmanship; besides, he was getting old — he was already well into his seventies — and his bid for the throne, successful as it had been, had used up much of his remaining strength. Utterly incapable of coping with the crisis he had inherited, he could do little but preside, helpless, over the further disintegration of the Empire, during which one revolt followed another and the State drifted further and further into anarchy. The old party of the civil bureaucracy had collapsed with the murder of Nicephoritzes, and with it the authority of the Senate; all that was left to the Byzantines was to pray that of the several military commanders now struggling for power, one might establish himself above the rest; and that that man might prove himself a leader capable of putting an end to the chaos.
Three years later - in the nick of time - their prayer was answered; and answered more completely than anyone could have dared to hope. The pathetic old Botaneiates abdicated in his turn in favour of an aristocratic young general who, coming to the throne on Easter Day 1081, was to reign for the next thirty-seven years, giving the Empire the stability that it so desperately needed and governing it with a firm and steady hand. That general was Alexius Comnenus, nephew of Isaac I and father of the celebrated Anna Comnena, who was to make him the subject of one of the most wholly enjoyable of all medieval biographies. Not even Alexius could undo the damage done by the battle of Manzikert; that, alas, was past repair. But he could, and did, restore to Byzantium its reputation and its good name among nations, thus preparing it to play its part in the great drama that was to begin to unfold even before the end of that turbulent century: the Crusades.