PART THREE
15
`Falling on the barbarians without warning [Stilicho] utterly destroyed their whole force. Scarcely anyone escaped, except a few whom he accepted as auxiliaries.' - Zosimus, fifth century.'
All Gaul was filled with the smoke from a single funeral pyre.' - Orientus, describing the impact of the barbarian invasion in 406.2
Aremarkable document from the imperial bureaucracy survives from .the end of the fourth century. Its formal title is `The list of all offices, both civil and military' - in Latin, the Notitia Dignitatum omnium, tam civilium quam militarium, although scholars usually just refer to it as the Notitia Dignitatum - and it was produced originally for the senior notary or clerk (primicerius notariorum) of the emperor Honorius in 395. The recent division of the empire into eastern and western halves is clear throughout the text, the two being throughout shown as utterly distinct. It is highly detailed, setting down each post, saying something about its responsibilities, and in the case of army commanders naming the regiments in their charge. What makes it all the more fascinating is that it is colourfully illustrated. The insignia of each rank is shown, along with symbols of their work - for instance, weapons for those in charge of the factories making equipment for the army or loaves of bread for those responsible for collecting levies of food. Provinces are shown both as personifications and as highly stylised pictures, with miniature walled cities labelled to show the principal communities of the area. Field army regiments have circular images of the devices that were supposed to be painted on their shields.3
We do not, of course, have the original document written and illustrated by the staff of the senior notary in 395. Substantial works, whether literary, legal or administrative, survive only because copies were made over the centuries, and the earliest surviving texts are usually medieval. In the case of the Notitia Dignitatum we have several sixteenth-century copies of a version made for a Carolingian king in the early middle ages. The illustrations present an odd mixture of Roman and medieval styles. Anything the artist recognised tended to be painted as it looked in his world - hence the little walled towns have a decidedly medieval look. Things he did not understand were more likely to be copied exactly.
Even in the Roman period the Notitia Dignitatum seems to have been modified on several occasions. It was clearly used by some parts of the imperial government in the Western Empire, for the sections on the Eastern Empire do not seem to show any changes after 395. The western sections, particularly some of those dealing with military organisation, were altered in the fifth century, perhaps as late as the 420s. The updates were patchy and not always consistently made throughout all of the relevant sections. There are also clearly errors in the pages showing the shield patterns of army units. Quite a few of these are blank, and others look like invented variations on a theme. They are also all shown as neatly circular, when in fact the army used oval shields. Judging from the intricate patterns on the painted shields found at Dura Europos, the miniature versions are also likely to be greatly simplified.4
For all its errors and confusions, the Notitia Dignitatum is unique in providing an official survey of the imperial bureaucracy and army. It contains a vast amount of information about the structure of the divided empire, including many details not recorded elsewhere. Yet its very uniqueness is also highly frustrating, for if we had one or two comparable documents from earlier or later decades we could trace the developments in the imperial structure far more clearly. There are hints of changes in the text we have, while other sources confirm the existence of much of the administrative and military structure described at other periods. Ammianus, for instance, mentions many of the ranks and posts listed in the Notitia, as well as a number of the regiments, and seems to confirm other aspects of organisation. Much of the structure at provincial level, as well as the distribution of limitanei, seems unlikely to have changed as a result of the division of the empire in 395.5
All this is most encouraging for the wider usefulness of the Notitia Dignitatum, but scholars have sometimes been more than a little reckless in its use. Those studying the army have been especially inclined to stretch its lists of units both forward and back for more than a century. It is conventional to assume that a regiment named after an emperor was also formed by him, ignoring the real possibility that already existing units were renamed as a reward or to encourage loyalty. Armies are not always the most logical of structures, especially when it comes to names and titles - so that, for instance, in the modern British army a soldier in different infantry regiments may be ranked as rifleman, fusilier, kingsman or guardsman instead of the more prosaic private. It is also unwise to calculate the losses at Adrianople or other disasters on the basis of deducing `missing' units from the lists in the Notitia Dignitatum. As already noted this ignores the probability that some units were lost in other campaigns, while others may simply have been renamed, merged or disbanded for any number of reasons. Field army units at the largest estimate were just a fraction of the size of the old legions and hence more vulnerable to losses and short-term decisions to alter the structure of the army. The frequency of civil war made it all the more likely that the army list would be confused rather than neat and logical.
It is equally important to remember that the objectives of the men who first drew up the Notitia Dignitatum were limited. The chief concern of the notaries was in the issuing of commissions for the various military and civil posts listed - a writing desk and a bundle of commissions accompany the insignia of the primicerius notariorum. It was the appointments themselves, along with their seniority, that mattered. This was not a work intended primarily to explain how the army and administration functioned. Imperial posts brought their holders power while they were in office and, more permanently, rank, pay and privileges. There were always plenty of men seeking each post and the senior officials with power to bestow or influence appointments expected future favours or immediate bribes to secure success. Inevitably, this encouraged the proliferation of offices, and in some cases these were mere sinecures, for the individual never had any intention of actively performing his supposed duties. As usual, the frequency of imperial legislation intended to curb such abuses suggests that it had only limited success. An archive of letters left by an officer commanding a unit of limitanei in Egypt in the fourth century describes how on arrival at the garrison he discovered that several other men had been granted the same commission. Only after considerable difficulty and appeal to higher authority was his own claim acknowledged.6
The Notitia Dignitatum listed the officially recognised positions - there would be a commander for every army unit even if he is not named specifically - just as its title claimed. There is no mention of the many forces of allies that formed a major part of armies at this period or of the officers who commanded them. Presumably such posts were not the concern of the notaries. Similarly, those who subsequently adapted the text seem to have been interested only in certain sections. Changes in civil posts are rarely noted, and even in the military sphere the priority seems to have been certain units, presumably those of concern to the officer whose staff kept the list.'
The overall impression remains of a vast, highly organised and powerful empire, recently divided into two distinct hierarchies in east and west, but still in spirit part of the same entity. Without doubt the Notitia Dignitatum has reinforced the views of those scholars who depict the fourth- and to a lesser extent early fifth-century empire as still inherently strong and generally efficient. Its list of army units is the principal basis for the claims that the army was massive, perhaps well over 6oo,ooo strong and so almost twice the size of the forces at the disposal of emperors like Marcus Aurelius. Most will note that such figures would represent `paper strength' and that the actual number of effectives was likely to be lower, but do not then seem to absorb the full implications of this. The picture still remains of a very large army.'
Yet it is very hard to reconcile this with the course of events in the late fourth and fifth centuries. The army - and to some extent the imperial state itself - at times seems invisible, with regions supposed to have been strongly garrisoned apparently undefended. Time and again the question arises of where this supposedly massive army actually was. This raises the fundamental issue of how far the Notitia Dignitatum reflected day-today reality, especially where the army was concerned. Clearly, the listing of a regiment along with its shield device meant that it actually existed as far as officialdom was concerned. In some cases this may genuinely have meant that the unit had a substantial part of its full complement of soldiers, trained, equipped and ready for service. Alternatively, it could have been massively under strength, although still able to take the field. Another option would be a small cadre, with some key staff and the documentation preserved, all waiting to be turned into a proper unit if ever it was allocated sufficient recruits and other resources. Finally, the regiment may only have existed on paper, its existence reflecting the status of the general in command and perhaps showing what forces he would control in an ideal world. This in itself did not necessarily mean that someone was not enjoying the salary and privileges of being its commander.
All of these options are possible and it is more than likely that there were examples of each of them at various times amongst the regiments listed in the Notitia. We have already seen in the conflict with the Goths in 376-382 that the Roman army found it very difficult to deal with a comparatively small number of enemies. At times, when reading descriptions by modern historians of the warfare in this period, it is difficult to avoid the image of Hitler in his last days, planning grand offensives on a map with divisions that had long since ceased to exist. The situation was not so desperate in the years following 395, nor was the enemy so powerful and organised, but the reigns of Honorius and Arcadius were desperate enough.
Stilicho
Even when emperors were strong, their senior officials and commanders routinely and ruthlessly struggled for power, promotion and influence. When emperors were weak or young there was even less restraint in this never-ending contest for dominance. In 395 Theodosius' sons were both young and later events would prove their characters to be extremely weak. Incapable of restraining their subordinates, when old enough they settled for simply playing them off against each other. Anyone able to dominate the emperor effectively gained supreme power. In the east this was first achieved by Arcadius' praetorian prefect Rufinus, who was to be followed by a succession of court officials who virtually ran the Eastern Empire. These men held a range of formal posts and this was in most respects far less important than the hold they were able to develop over the young emperor. This control was never secure, and all eventually fell from power and died violently.
In the Western Empire the real power tended to rest with the man who controlled the bulk of the army, rather than with civilian officials. For the first thirteen years of Honorius' reign this was Stilicho, whose rank - apparently created for and by him - was `Count and Master of all Soldiers'. This made him formally the supreme military officer in the western armies. There was no equivalent to this post in the east, where several Masters of Soldiers held equal power. In 395 Stilicho had the added advantage that many field army units from the east were still in Italy following the defeat of Eugenius in the previous year. These, combined with western regiments, including many which had previously fought for the usurper, gave Stilicho a military force that none of his potential rivals could match.9
Stilicho's father was a Vandal who had commanded a cavalry regiment under Valens, and his `barbarian' ancestry would be thrown at him by his critics. However, there is no reason to see him as anything other than fully Roman. He had begun his service in the protectores and risen rapidly after winning the favour of Theodosius. He married Serena, daughter of the emperor's brother and raised in the imperial family after her father's death. Claiming that the dying Theodosius had entrusted the care and protection of his sons to him, Stilicho swiftly secured control of the tenyear-old Honorius, effectively ruling the Western Empire as regent. He does not ever appear to have sought imperial status for himself, but in due course he would arrange the marriage of his daughter to Honorius.
The new emperors - or perhaps better, the men who controlled them - were faced with problems almost immediately. In 395 predatory bands of Huns raided both Sassanid Persia and Rome's eastern provinces, plundering widely in Armenia, Mesopotamia, Syria and even into Asia Minor. The presence of many eastern army units with Stilicho may well have reduced the capacity of local commanders to deal with this attack. As usual with such raids only a small minority of communities were actually struck, but fear spread much more widely. Away from the borders with Persia, most of this region had been peaceful for over a century and this was clearly a traumatic episode, especially when the Huns returned two or three years later."
Another, closer threat also erupted in 395 when some of the Goths settled within the empire by Theodosius rebelled. They were led by Alaric, an officer commanding troops serving as allies with the Roman army. He was probably from an aristocratic family in one of the tribes, and in due course would be called king, but the source of his authority is not clear. Some of his men may have been bound to him by ties of kinship, others simply by serving under him in the army. He was clearly a man of considerable personality, who would keep his followers loyal for fifteen years. Like the precise root of his power, his long-term objectives are unclear. There had been resentment and some minor outbreaks of rebellion amongst the Gothic communities when they were called upon by Theodosius to provide troops to fight for him against Maximus and Eugenius. However, Alaric and his men had not been part of this and had served with some distinction at the River Frigidus. Perhaps the belief that the Goths had been cynically sacrificed at that battle fuelled resentment. More probably, the succession of two new emperors simply offered an opportunity for profit.
Quite a few of the Goths appear to have resented the settlement imposed on them in 382, even though it was very favourable by Roman standards. We do not know how many men Alaric led, or whether they consisted primarily of warriors who had not settled down in their new lands or also included whole communities. Certainly, they were strong enough to be seen as a dangerous force. From the beginning Alaric sought a senior military appointment from the Romans. Such a commission would have brought with it the right to food supplies for his men from the state. It is possible that all he and his followers hoped for was an improvement in their status and livelihoods well within the Roman system."
The Goths began plundering the provinces in Thrace and Macedonia. Once again, the absence of many regiments in Italy prevented Arcadius and Rufinus from dealing with them. Yet, as usual, they were more concerned with the threat posed by Roman rivals. Stilicho proclaimed that Theodosius had entrusted him with the guardianship of both his sons and led his army eastwards, ostensibly to deal with the rebellious Alaric. However, Arcadius rejected his assistance and instructed him to send the regiments from the eastern army back to Constantinople. There seems to have been friction between the western and eastern troops and it may well be that Stilicho could not control all of his men. In any case, he obeyed the order. He withdrew with all the regiments from the western army and did not engage Alaric.'2
The eastern regiments were led by an officer named Gainas, who was himself of Gothic extraction. When they arrived outside Constantinople they murdered Rufinus as he rode out to meet them. The latter's influence over the emperor was taken over by his chamberlain, the eunuch Eutropius, who for the moment was more concerned with consolidating his own power at court. For two years Alaric was left to plunder the provinces. In 397 Stilicho returned with an army bolstered by large contingents of barbarian allies. Eutropius and Arcadius proved no more enthusiastic about accepting his aid and thus acknowledging his right to intervene in the east. Alaric was blockaded and forced to retreat into Epirus, but then Stilicho himself withdrew without achieving any permanent victory. Acting on the advice of his favourite, Arcadius had declared Stilicho a public enemy. Around the same time the eastern government began talking to Alaric and eventually agreed to make him Master of Soldiers in Illyricum. The former army officer turned rebel had now become a Roman general. Clearly, Eugenius and Arcadius found this preferable to accepting the dominance of Stilicho."
Late in 397 Gildo, the man left in command of the North African provinces since Theodosius' reign, decided to defect with his province to the Eastern Empire. There was an immediate crisis since Italy and Rome relied so heavily on grain and other food from the region. Stilicho sent Gildo's brother and bitter enemy Mascezel in command of a small expeditionary force, which sailed to Africa from Italy. He was quickly successful. Stilicho was generous with his praise, but shed no tears when Mascezel fell into a river and drowned soon after his return. There were rumours that he had been thrown in by Stilicho's bodyguards.14
In the east Eutropius seems to have taken personal command in a successful campaign against the Huns, and this success prompted him to arrange to become consul in 399. It was customary at this time for the eastern and western emperors each to name one of the pair of consuls who took up office on i January and gave their names to the year. The post itself was one of prestige rather than real power, but it was ancient - there had been consuls now for nine hundred years - and aristocratic opinion was outraged by the idea of this hallowed office being held by a eunuch. As his unpopularity grew, so other senior men began to see Eutropius as vulnerable. In the same year some Gothic troops operating against bandits in Asia Minor chose instead to rebel, as their commander had a personal score against Eutropius. Gainas led a force against them, but then allied with the rebel troops and urged the emperor to grant their demand to dismiss his favourite. Arcadius' wife Eudoxia joined the chorus of condemnation of the chamberlain and eventually the emperor gave in. Eutropius at first sought sanctuary in a church, but gave himself up when he received the promise that his life would be spared. He went into exile in Cyprus, only to be executed a little later on the false pretext of plotting against the emperor.
For the moment the empress and her favourites dominated the court, but Gainas marched on Constantinople and the implicit threat of this force granted him a short-lived supremacy. He was named as consul for 400. However, his Gothic soldiers were unpopular when they were stationed in Constantinople and eventually he decided to send them to Thrace. However, as the columns formed up to leave, the rearmost parties were attacked by mobs. Large numbers were killed, including many of the soldiers' wives and children. One large group sought sanctuary in a church, but died when the building was set on fire with them still inside. Shortly afterwards Gainas was defeated by an army led by a general called Fravitta, who was also a Goth. Fleeing across the Danube, Gainas eventually died at the hands of a Hunnic king. Fravitta was executed by the Roman authorities just months after his victory on allegations of disloyalty. Generals who were too successful or too popular were seen as dangerous by the powers at court. Eudoxia and her allies had regained control of the court and thus the Eastern Empire."
In 401 Alaric left his haunts of the last few years and headed for Italy. The situation in the east had changed and the new regime was unlikely to be well disposed towards him when it had risen to power by criticising the prominence of `barbarian' and specifically Gothic generals such as Gainas. Free for the moment from internal disputes, there was a real chance that the imperial government might decide to withdraw his commission and use force against him. For the moment, Alaric judged that he could hope for better terms from Stilicho. We do not know how many men followed him. Those Goths who remained on the land they had been granted in 382 and contentedly farmed inevitably do not get mentioned in our sources. It is probably wrong to think at this stage of a whole people once again on the move, a delayed resumption of the migration that had brought the Goths across the Danube in 376. In the main his followers are likely to have been the young and restless. Perhaps they had failed as farmers or were the children born inside the empire for whom there was not enough land to share. There was also a long tradition in tribal societies of youths seeking glory and wealth as warriors or soldiers, something on which the Roman army had long relied to supply it with recruits. There cannot have been many who had fought at Adrianople more than twenty years before. Some of the warriors were doubtless accompanied by their wives and families, just as camp followers often trailed behind Roman army units. The behaviour of Alaric and his men in subsequent years was not that of a migrating people, but rather of an army.
Alaric's hope was to win negotiated concessions from the Western Empire, most likely including senior military rank and the use of state resources to feed and supply his followers. He was rebuffed. Therefore, while Stilicho was north of the Alps dealing with barbarian raids into Raetia (very roughly equivalent to modern Austria), Alaric invaded Italy in 402, brushing aside one small Roman force to besiege Milan. The city was a frequent imperial residence, but during these disturbed years the court spent more time and eventually settled in Ravenna, which was surrounded by marshland and very hard to attack. It was also more isolated. Stilicho returned to Italy and fought two (perhaps three) battles against the Gothic army. He captured Alaric's wife and children along with other distinguished prisoners, but also suffered losses himself and failed to win a decisive victory. There was a truce, before fighting broke out again and another battle was fought outside Verona, which again left no clear winner. Eventually Alaric withdrew, most probably through lack of food, and moved back into the Balkans for the next few years. There, on the border between the Eastern and Western Empires where neither side could exert much control, he waited, plundering or extorting the supplies he needed from Illyricum. Around 405 Stilicho was willing to negotiate and granted the Gothic leader the rank of Master of Soldiers. The court in Constantinople refused to acknowledge this, especially since it again implied the right of Stilicho to dictate to both east and west. 16
Stilicho soon faced more immediate problems. Near the end of the year a large force of Goths led by King Radagaisus launched a deep raid and once again reached northern Italy. These warriors were from the tribes beyond the Danube and had no connection with Alaric's men, save that they were all broadly Goths and spoke versions of the same language. Zosimus claims that there were 400,000 of them, but such a figure is clearly absurd for any tribal army, let alone a raiding band. He also tells us that Stilicho concentrated thirty units along with allied contingents in the force that met and utterly defeated the raiders. The Notitia Dignitatum lists 181 regiments for the field armies of the Western Empire, forty-six of them in Italy and forty-eight in Gaul. Stilicho seems to have summoned substantial forces from the northern frontiers to form this army. Precise numbers and distribution may well have been different in 405, but this still suggests that it was next to impossible to concentrate more than a small minority of the supposed mobile field army in one place, even assuming that all of Stilicho's units on this occasion were comitatenses. Once again, the lack of knowledge about the sizes of units makes it impossible to calculate the size of the army. Nevertheless, it proved sufficient to win a clear victory - the price of slaves is supposed to have fallen sharply when the market was flooded with captive Goths .17
During this campaign the first of a series of usurpers was proclaimed emperor by the army in Britain. He was murdered by the soldiers within a matter of weeks and his successor suffered the same fate after just a few months. The third in the line was Constantine, allegedly chosen because of his famous and imperial name, and he proved far more capable as a politician. Distant Britain often seems to have felt neglected by the imperial government and so inclined to make its own emperor, but rarely did these men prove content with just the rule of the island. Like others before him, Constantine crossed the Channel, probably in 407, and soon controlled most of Gaul as well as large parts of Spain. Stilicho sent a Gothic officer named Sarus, a bitter personal enemy of Alaric, to fight against the usurper. He enjoyed some success, but was then in turn forced to retreat.'8
On New Year's Eve, traditionally in 406, but a good case has been made for 405, raiding bands of two separate Vandal groups, the Silings and Asdings, as well as Suevi and Alans crossed the Rhine near Mainz. Once again, these seem to have been predominantly groups of warriors and not entire tribes migrating in search of new homes. There is no direct evidence for the often repeated story that the river had frozen, although this is certainly possible. Similarly, the claim that this movement was prompted ultimately by pressure from the Huns is unsubstantiated and generally unlikely. It is far more probable that the apparent weakness of the Roman frontier just seemed to offer an opportunity. If the earlier date is correct, then the withdrawal of troops from the Rhine to face Radagaisus could well have created this impression. In this case the usurpations in Britain and Constantine's occupation of Gaul may have been fostered by the failure of Honorius' representatives to stop the invaders. If the attack was only launched at the end of 406, then the warbands themselves may have been taking advantage of the confusion that inevitably followed the outbreak of civil war within the empire. No leader is named for this attack, but the co-operation between several distinct groups suggests the presence of one, or perhaps a few chieftains of considerable charisma. Very quickly the warrior bands overran and plundered the communities near the Rhine, before pushing on into the interior of the provinces. Constantine won some minor victories over them, greatly bolstering his support in Gaul, and seems to have kept them bottled up in the northern regions of Gaul. Yet he did not break them, and for the next few years these bands ranged individually or together through this region, plundering or extorting at will.'I
In 407 Alaric decided to take advantage of the situation and led his army back towards Italy in the hope of wringing an even better deal from the beleaguered Stilicho. In the next year he demanded 4,000 lb of gold as the price for not launching a new invasion. Stilicho, acting in a way now very familiar for Roman leaders, decided that Constantine was the greater threat and that Alaric could be hired to fight against him. He agreed to pay the gold, and went to the Senate, since such a vast amount was very difficult to secure at short notice and the wealthy senators were one of the most logical sources. In addition, temples and artworks in Rome were stripped to provide the necessary sum, but the senators were bitterly resentful of this bribing of an enemy. One described it as `not a treaty, but a pact of slavery."'
The priorities of the leading Romans shifted abruptly when news arrived that Arcadius had died on i May 408 at the age of just thirtyone. He was succeeded by his seven-year-old son Theodosius II - the infant had been named as Augustus when he was barely a year old. Stilicho and Honorius both announced their intention to go in person and supervise the accession of the new emperor, in the process no doubt making clear the primacy of the western court. Rivals ambitious to supplant Stilicho had long been encouraging Honorius - who was himself still only twenty-three - to mistrust him. They claimed that the general planned to make his own son emperor, perhaps instead of Theodosius. He certainly was determined to maintain a close link with the imperial family. When Honorius' wife and Stilicho's daughter died, Stilicho promptly replaced her as imperial consort with her sister. The young emperor resented his marginal role in running the empire, just as Valentinian II had before him.
Just who engineered the final confrontation is unclear. The payment to Alaric made Stilicho deeply unpopular and his enemies quickly scented an opportunity to attack him. There was a mutiny of the troops concentrated at Ticinum ready to be sent into Gaul. Several officers and senior civil servants - probably all men appointed by and loyal to Stilicho - were murdered. Honorius was there, but survived the bloodletting. Stilicho was some distance away, attended only by some barbarian troops who were staunch in their loyalty. By the time he reached Ravenna the emperor had ordered his arrest. He refused to fight, in spite of the fact that the soldiers with him were willing. Instead he sought sanctuary in a church, but gave himself up on the promise that his life would be spared. He was promptly executed, once again ordering his men not to protect him. His end was dignified, especially since it was rare for a senior Roman commander to accept death rather than take the chance of fighting a civil war. Perhaps he realised that he had been utterly outmanoeuvred and that his own position was now too weak for him to have any prospect of winning a struggle with Honorius. However, it is hard not to want to believe that he put the good of the empire before his own fate. It may even be true.2'