Chapter 3

Long-term Problems, the Nature of Roman Instability in Transmitting Power, and What Could Easily have Happened

Earlier ‘What Ifs’: the crises that led to the creation of the ‘Later Empire’

In a similar manner to the disaster at Adrianople, a victory against the Goths in their earlier incursion of 251 – when Emperor Decius was killed and the Balkans opened to ravaging – would have avoided sparking off a round of ‘copy-cat’ invasions by emboldened Germans. An experienced commander who his troops had hailed as Emperor after he defeated earlier attacks and forced him to turn on his master Emperor Philip (or so he claimed), Decius had had problems in bringing the Goths to battle in Thrace in 250–1. Logically the invaders had spread out over too wide an area to be confronted quickly in one grouping and had to be confined in a manageable area. This was an inevitable result of the time lag between them crossing the Danube and the Emperor arriving from Italy. There had been trouble on the Danube from the Carpi through the 240s, so a major attack was not that surprising, but placing a local commander in the area with a large enough army was politically dangerous, as he might use his troops against Decius as Decius had done to Philip. The delay in tackling the Goths probably owed much to Decius’ caution after recent revolts. A stable and unchallenged ruler could have risked having a senior commander or more troops in the region or both. This is with the caveat that manpower was available and had not been diverted to face Persia, we simply do not know. In any case, both Decius and his elder son were killed in battle. A bout of internal Roman revolts followed Decius’ death, with his surviving son Hostilianus, young and inexperienced, being superseded by one general after another.

The wave of German attacks, at sea across the Black Sea as well as on land, was followed by another invasion by the opportunistic Sassanids and the diversion of Emperor Valerian and many of his troops to the East to fight them in 260. The Emperor’s disastrous military failure there and capture at a parley by Great King Shapur’s troops broke the Empire up into chaos and opened many more provinces to ravaging. The humiliation was played up by Shapur, whose ambitions seem to have been to restore the ancient Achaemenid Empire of Darius and Xerxes that had reached to the Aegean. Reviving the ancient tradition of rock-cut glorification of the Persian Great King for all to see on the main road from Iraq up onto the Iranian plateau, he created a carving of Valerian kneeling before him as a suppliant, and was rumoured to have stuffed his body as a trophy when he died. In practical terms, the Persian invasion of Syria had to be driven back by Odenathus, the Roman client-ruler of Palmyra, as the shattered Roman army in the East retreated into Asia Minor and became split up in claims of Imperial power by its commanders. A wave of rebellions left Valerian’s son Gallienus with control of the central Roman lands but the West (Gaul, Spain, Britain, the Rhine) lost to a breakaway regime under Postumus and the Levant controlled by the autonomous city-state of Palmyra under Odenathus and later his widow Zenobia. Both states were reconquered by Aurelian after 270, but the multiple crises led to the emergence of a much more centralised, bureaucratic, and tax-heavy Roman state.¹

Would this form of state have ever emerged had the Empire not faced disaster in the 250s? Or was the loss of Roman manpower in the plague of the 250s sufficient to embolden attackers and fatally weaken the Roman army anyway? Indeed, it can be argued that the nature of copycat revolts, one successful rebel general’s example encouraging another to challenge him later, means that the chronic and disastrous instability of the mid-third century Empire owed a lot to recent political failure. The Empire had been invaded by large-scale German tribal forces at a time of plague in the 160s, but Marcus Aurelius had fought them off and the Empire did not collapse. Was this because the Empire of the 160s did not have to cope with the Sassanid state too, or was it due to mid-second century Roman political stability? It is noticeable that the Emperors whose death and capture sparked disaster in 251 and 260 (Decius and Valerian) were both usurpers, not secure long-term rulers from an established dynasty.

Would the Empire have fared better had it had a stable line of unchallenged rulers after the extinction of the Severan dynasty in 235?

There were no long-lived rulers or secure successions from this event until the creation of Diocletian’s new system of government after 284. It is possible that the loss of both Emperor Septimius Severus’ adult sons in 211–17 (Geta murdered by his brother Caracalla, the latter a tyrant murdered by his most senior commander) was what ushered in this dangerous instability, given the inadequacies of their distant cousins and successors Elagabalus and Alexander Severus. The tyrannical Caracalla was murdered for self-preservation by his competent Praetorian Praefect Macrinus in 217, but the latter was seen as dynastically illegitimate by his mutinous troops; some of the latter duly acclaimed Caracalla’s cousin (and reputed son) Elagabalus, the teenage High Priest of the Syrian sun-god at Emesa. Macrinus was defeated in battle and killed, and the victor and his entourage moved to Rome, taking the sacred stone (was this a meteorite?) of the god with them. A transvestite bisexual exhibitionist, Elagabalus was murdered in 222; Alexander, his cousin and successor, was seen as dominated by his mother and was murdered too on a Rhine campaign (235). A round of coups commenced, with the troops’ capable but lowborn choice of sovereign, the Thracian ex-ranker Maximin, facing revolt in Rome and Africa.

After 235 no Emperor could secure stability, even the militarily competent Maximin and Philip the Arab. Arguably, the fault for all this lay with Septimius Severus for not killing his violent, and fratricidal elder son Caracalla, who had possibly already plotted his murder, and ensuring that the less dangerous Geta succeeded him; the latter and his capable adviser Papinian could have secured stability for a vital period of the early-mid third century. Severus’ stated plea to his sons on his deathbed to live at peace with each other, keep the troops happy, and not bother about anyone else was wishful thinking. If he did not want to kill his son he could have despatched him to a remote island, as Augustus did with his allegedly violent and politically dangerous grandson Agrippa Postumus.²

The Empire’s problems in the third and fourth centuries. How might they have been reduced by earlier military successes?

In the longer term, it is also arguable that the nature of the Germanthreatened Empire’s outer defences in the later fourth century, easily crossable rivers, the Rhine and Danube, weakened its defences. In the first decade of the millennium Augustus’ generals had attempted to annex the lands between Rhine and Elbe, bringing many of the local tribes into the Empire, only to meet with disaster in the Teutoberg Forest in AD9. The conquest of lower Germany had then been abandoned. Even if it had succeeded, a combination of Roman parsimony about garrisons and the ever-likely civil wars could easily have occasioned a successful revolt before the fourth century. In the 100s Trajan had responded to repeated Dacian attacks on the middle Danube by advancing his frontier to the eastern Carpathians, and in the 170s Marcus Aurelius had temporarily overrun the Czech lands.

Maintenance of all three occupations would have brought many of the tribes who threatened the Empire under its rule, with their warriors serving in the Roman army, like the previously hostile Gauls from the 50s BC, instead of raiding Roman lands. The remaining Germanic territory South of the Carpathians, that of the Iazyges between middle Danube and Theiss, could have been occupied or left as an allied kingdom under pro-Roman chieftains. The military occupation of a slice of territory was in any event less important than its neutralisation as a threat. Rome had long operated through a cheap system of allied kingdoms that did not entail direct rule, as with the Germanic tribal realm of Maroboduus on the Danube and the multiplicity of Levantine Greco-Aramaic states. Provided that a territory was not immanently hostile to the Empire, occupation usually occurred when a Roman ruler needed to prove his military credentials by an impressive conquest, as with rising politician Caesar in Gaul in 58BC and ageing new Emperor Claudius in AD43. Indeed, it is worth remarking that despite the disastrous defeat of a large Roman army in the German forests East of the Rhine by Arminius in AD9 the situation had been partly rectified by Tiberius’ nephew Germanicus before he died in 19.³ The latter was ambitious and arrogant, and made much of his physical resemblance to Alexander the Great.

Had Germanicus succeeded Tiberius as planned, could he have decided to add to his reputation by invading and annexing territory beyond the Rhine or Danube? Doing this would have entailed either moving part of the garrison of Lower Germany (four legions on the Rhine) into the area, risking revolt to their rear, or raising more legions. Augustus had had trouble in finding new troops after losing three legions in the German disaster of AD9, having to arm slaves, so would it have been difficult to create even two new legions to help hold down the lands to the Elbe?⁴ There was a political advantage in creating this new force for a new province (Transrhenus?), as it would serve as a check on the ambitions of the military commanders in Lower Germany. In AD69 their commander Vitellius rose in revolt against new Emperor Galba and fought his way to Rome; could he have done so had he faced a local rival who was still loyal?

The precedent of Roman occupation of a similar agriculturally based tribal society in Gaul shows that local rebellion was still a problem (occasionally) in Tiberius’ time, seventy years after Caesar’s conquest, and there was to be a major eruption under Julius Civilis in northeastern Gaul once Rome fell into civil war. The latter revolt in 69 was aided from beyond the Roman frontier on the Rhine, and had the Empire advanced to the Elbe some time after AD9 or had Varus defeated Arminius then and saved three legions for use in an occupation, revolt was still probable later. Crucially, the Empire kept existing tribal units as administrative sub-provinces in Britain and Gaul, and would probably have done the same in Germany. This aided a sense of identity among the locals, as shown in the tribal-based nature of revolt in Gaul in the early 20s and 69. Breaking up the existing tribal landowners’ landed power-base by atomising the social structure would have been more effective in preventing revolt, which would have entailed massdeportations as in Dacia under Trajan after 100. Was this the successful policy that the Romans adopted in keeping the Iceni quiet after Boudicca’s revolt in 60–1? The name of the Iceni never re-emerged in the fifth century, unlike other British tribal kingdoms. And could a similar break-up of Germanic tribal polities between Rhine and Danube in the first century AD have produced an invaluable long-term German boost to the Roman army that aided it in its third and fourth century wars?

Such territory may not have been fully occupied as a province (like Germany west of the Rhine and the Gallic Belgica) but just dominated by legionary outposts, leaving open the chance of revolt at times of weakness, as with the Batavian region of the Rhine-mouth in AD69–70. The rule of this region had been left to its own local chieftains, provided that they supplied troops to the Empire; the same strategy was followed for allied British kingdoms beyond the frontier in the 40s and 50s, e.g. Cartimandua’s Brigantes and Prasutagas’ Iceni. The main aim of the Antonine occupation of Bohemia in the 170s seems to have been to prevent more invasions of Italy, and this aim may have been achieved without creation of a formal province. In the event Marcus’ death in 180 and Commodus’ withdrawal meant that Marcus’ war-aims of around 177 are unclear.⁵

The effect of military domination and partial occupation would still have been the same: to prevent the emergence of the new super-tribes, the coalitions of disparate German peoples under single dynamic leaders which invaded the Empire in the mid-third century. The use of new names to identify them in place of the terminology of the first century AD suggests new tribal groupings emerging, probably under active warlords who forged coalitions. The regular recruitment of their menfolk to the Roman army would have kept the latter as allies to, not preying on, the Empire. The Germans would have been sent to serve well away from their home territory to minimise the chances of revolt, as with the Empire’s Sarmatian nomad allies from the lower Danube.

Holding down such an extended dominion would have had its problems, not least revolt. Romanization of tribal peoples was a slow process and some of the Gauls revolted in the early 20s AD , followed by the major Rhine revolt of Civilis in 69–70. But a mountain frontier eastwards from the Elbe Gap near Leipzig to the Iron Gates on the lower Danube, broken only by a few passes and occasional low-lying regions like the Ostrava Gap/Beskids, would have been easier to defend from penetration than the river frontier, and would certainly have required no more troops. The main danger would have been of a still-restless tribal population hankering after its freedom and ready to revolt at times of crisis with help from beyond the frontier, as with parts of northeast Gaul as late as AD69, over a century after Caesar’s conquest. That problem, however, had not been insurmountable once Rome regained its military cohesion and Vespasian could send troops to suppress the Gallic revolt under Civilis; using the same argument, the Danube- Carpathian lands were no less controllable for the period after 180. There would however have been longer distances involved, hampering quick reaction. Also, the poorer soil of the north German plains and the forests would not provide useful produce for the Roman economy; costs would have been high.

The numbers of invasions that the Empire faced from the 170s onwards would thus have been reduced, though attacks from Wallachia across the lower Danube (e.g. the Gothic invasion of 251 and Gothic refugeemovement of 376) would have been unaffected. The Persian threat from the 230s would have been equally serious, though the disaster of 260 (when Emperor Valerian was captured and the East dissolved into chaos) owed much to the distractions of the Rhine and Danube invasions that resulted in an under-manned Eastern army. The third century Empire would still have been subject to major losses from the plague of 252, though not necessarily as many civil wars given the absence of certain domestic crises such as Commodus’ reign, the civil war of 193–7, and the successive coups and revolts from 235 to 260. Internal stability would have decreased the distractions of civil war, which aided the invaders, enabling an unchallenged Emperor to meet the main attacks head-on with his army as Marcus Aurelius did after 169.

State structures and the succession: from ‘First Citizen’ to hereditary autocrat

It can be argued that each successful coup or revolt from 192 had a cumulative effect on the Empire’s stability, and thus on its long-term chances of survival. No arrangement for the political succession is of course infallible, though a rigorous selection process or a definitive genetic right of heirship can present distinct advantages. The most stable political systems have relied on a collective leadership rather than one individual, as with the oligarchy of medieval Venice with its figurehead Doge, but this was unrealistic for an age of monarchy and military leadership, particularly in the Greek-Middle Eastern areas of the Empire.

The whole tradition of government in Rome had evolved from the concentration of political and military power in two annually appointed senior magistrates, with a temporary dictatorship for emergencies but a taboo on any notion of kingship; it was Augustus’ genius to introduce a hidden monarchy while avoiding the open single rule which Caesar had not troubled to disguise. The vital role of military power in a far-flung state needing large armies that the government had to control to avoid the turbulence of the late Republic made one man’s ascendancy inevitable. There was a need for a visible individual as the centre of control (and of cultic worship), as Caesar, Antony, and Augustus duly recognised. The last attempt to re-create a stable collective aristocratic leadership, by Sulla in 82–80BC, had foundered on patrician feuds and powerful provincial army commanders. The same would have been likely to happen had Pompey and the Senate defeated Caesar in 48BC or Brutus and Cassius defeated the Caesareans in 42BC.

There was not yet the necessary structure of government, or acceptability for it, to create a rigid, top-down bureaucracy in the early Empire. In this case a cabal of ministers heading permanent departments would rule the state and the Emperor be its nominal front-man, whose personal failings and possible removal did not affect the viability of the state. This sort of regime emerged in centralised dynastic China from the middle Han period, under a succession of weak Emperors. Ultimately, in China this did not preventfailure of leadership; inter-ministerial feuds erupted, civil war and revolt followed, and repeated break-ups into rival provincial polities followed that. But China had by then evolved a tradition of centralised bureaucracy, as created in the state structure of the kingdom of Jin in the ‘Era of Warring States’ and imposed on the entire country by the unifying ‘First Emperor’ Jin Shih Huangdi. The nearest bureaucratic equivalent in the Mediterranean world was the complex governmental structure of Ptolemaic Egypt, which carried on functioning irrespective of the incompetent sovereigns and bloody feuds within the ruling dynasty. In traditionalist Rome, government had been a simpler and more ad hoc affair with the small number of senior officials assisted by their private households and their groups of personal ‘amici’.

It had only slowly adapted to the massive political and economic demands of empire in the last two centuries BC, with its resistance to change aiding the political chaos of the Late Republic. Even with power and Mediterraneanwide official business concentrated on the Emperor, Augustus and his successors governed through the traditional means of a senior magistrate’s personal household. There was contemporary criticism of the emergence of low-class freedmen household officials wielding immense power under the later Julio-Claudians. There was always something ad hoc about the early Empire’s Imperial government, as analysed by Fergus Millar, and much was left to local self-rule by governors and city councils who sought Imperial advice and instructions when necessary, as shown by the famous correspondence between Trajan and Pliny the Younger.⁶

The creation of a large-scale and intrusive bureaucracy had to wait until the later third and early fourth centuries, and is plausibly ascribed to a specific political strategy formulated by the administratively minded Diocletian. This bore some resemblance to the hierarchic governmental systems of Sassanid Persia and China, and notably had a Persian-style formal court based at a ‘Sacred Palace’ instead of the more democratic courts of most early Emperors. The Emperor was cut off from direct contact with his subjects, surrounded by a hierarchy of court officials and Persian-style eunuchs, and in an increasingly religious age an effort was made to have the court reflect the order and ceremonial of Heaven (firstly in its Olympian guise, but soon in a Christian context).

It may be significant that Diocletian, unlike his militarily successful predecessors Claudius II, Aurelian, Probus, and Carus was a civilian (Greek) bureaucrat not a general. He may have decided to trust in a foolproof state system, not personal charisma as the Illyrian soldier-emperors from 268 to 284 had done. The system of a godlike Emperor isolated at court, which he created, was then kept on by the next long-ruling Emperor, Constantine, who could have reverted to a less formal mode of ruling. It is not sufficient to claim that a huge court was the inevitable accompaniment of autocracy, as the Illyrian Emperors, and even Septimius Severus much earlier, had ruled by naked force rather than in tune with early Empire deference to the Senate.

Megalomaniac Emperor Caius ‘Caligula’ had seen himself as a god and used the Temple of Castor in the Forum as the entrance to his palace. Nero had built a colossal statue of himself as the sun-god Helios and ruled as a sun-king from his ‘Domus Aurea’ in the mid-60s, and Domitian in the 80s had ruled as ‘dominus et deus’ (‘lord and god’) and built a large throne-room in the Imperial residence on the Palatine.⁷This fitted in with Eastern notions of the ruler as semi-divine but not with Roman practice; it was a different matter from so-called ‘Emperor-worship’, a universal practice in the Early Empire which was more a matter of a cult of respectful loyalty to the status of the ruler than treating him as divine. On all three occasions this innovation had been reversed by their successors, so the same could have been done to Diocletian’s ‘Oriental’ Court.

But instead Constantine continued the process, and made a further break with tradition by creating a new Christian capital, inaugurated in 330: ‘New Rome’ (later called ‘Constantinople’) on the Bosphorus. The idea of a permanent Imperial capital in the East for a locally resident Emperor had been Diocletian’s, but he had chosen the established Bithynian city of Nicomedia. Constantine set up a much more radical project, with a new and carefully planned city (based on a previous, smaller Greek town) with the facilities that were to become appropriate for medieval Christian cities (e.g. a cathedral). He also set up a Senate and encouraged aristocratic families to move to the city, going further than his predecessors had done; the new capital was clearly meant to supersede the old one.

It was also linked to a huge and dominant court and a regular round of Imperial ritual that added mystique to the Imperial office, though the unique and non-Roman nature of this may owe more to hindsight, as Constantinople was the only Late Roman capital to survive for centuries. Constantine still kept the senior role of Western Emperor for his eldest sons, Crispus (killed 326) and then Constantine II, who were based at Trier. The literary evidence certainly hints at him having a form of megalomania in his later years, which may have influenced the boldness and grandiosity of his plans. His family life ended up replaying ancient Greek myth, as he killed Crispus at the behest of the latter’s stepmother Fausta and then killed her for false accusations, an echo of Theseus, Hippolytus, and Phaedra.⁸

Augustus had laid claim to the loyalty of the troops of his late greatuncle’s armies from 44 BC as his genetic heir, the new Caesar, and duly outmanoeuvred the more experienced senior general Antony. The new system focussed loyalty on the ‘Princeps’ in his role as ‘Caesar’, with a family surname turning into an administrative title, and the hereditary basis of power was established. Significantly, the ailing Augustus’ designated heir in 23BC was his young and untried genetic heir, nephew and son-in-law Marcellus, not an experienced political and military lieutenant like Agrippa; in later years the choice fell on his equally inexperienced grandsons. The unique position he had created for himself became a hereditary monarchy, though with the technical caveat of confirmation of powers by the Senate. When the latter attempted to choose their own (or no) candidate for ruler after Caligula’s murder in January AD41 they were swiftly reined in by the Praetorian Guard.⁹

Augustus eventually attempted to lay down a form of succession by the most experienced member of the Imperial family rather than by father-son descent, arranging for his stepson Tiberius to be followed by Tiberius’ older, militarily senior nephew Germanicus, who was married to Augustus’ own grand-daughter, not by Tiberius’ son Drusus. Drusus (maybe two years younger than Germanicus) would then be followed by Germanicus’ sons. Tiberius kept to this faithfully, and put Germanicus’ sons ahead of Drusus’ (under-age) son Tiberius Gemellus in the queue to be Emperor. Arguably, Claudius did the same in putting his stepson Nero ahead of his own, younger son Britannicus. This system then evolved to father-son descent in the second century, when an Emperor had a son or brother, which was not the case from 81 to 161.

But a wholly hereditary system of rule, whereby the next eligible adult male (son, brother, or cousin) inherits irrespective of capability, introduces the risk of incompetence or insanity. This is followed by the overthrow of the incumbent by a more competent but illegitimate successor, who can then be challenged by ambitious relatives or military commanders. One coup leads to another, a minority is the inevitable opportunity for adults to overthrow the under-age sovereign, and prolonged instability is only ended by a strong ruler. This is what happened to the English monarchy in 1399, 1461, 1470–1, and 1483–5, though other kingdoms (such as France in 987–1328) had a luckier run of unchallenged capable heirs (usually adult) in direct succession.

If a state dominated by its officer-corps or provincial generals turns into a hereditary monarchy, any succession of an inexperienced minor can lead to coups by the military, as seen by the repeated fate of Sultans’ under-age heirs in Mameluke Egypt after 1260. This fate happened to Gordianus III of Rome, teenage ruler in 238–44, at the hands of his Praetorian Praefect Philip. If there is an heir entrusted with a governorship and army outside the capital, they can then overthrow their sovereign, as Julian did to Constantius II in 360–1, and as occasional Ottoman Sultans did, e.g. Selim I to Bayezid II in 1512. If a state is split among a multiplicity of eligible adult heirs as corulers, mutual assistance is less likely than endless struggles for supreme power, of which invaders then take advantage, as with Carolingian Francia among Louis the Pious’ sons in the 840s. If the state is lucky, one heir can quickly destroy his rivals and reassert central authority, as Bayezid II and Selim I did in the Ottoman civil wars of 1481 and 1512. If it is not, a standoff and permanent division ensues as in Francia post-843.

In Rome’s case, the split of power among multiple heirs in the fourth century often led to civil wars among the rivals which only ended with one candidate’s victory, as after Diocletian’s abdication in 305 and Constantine’s death in 337. The division between Valentinian and Valens in 364 was better managed. It was not fatal to a militarily strong Empire, despite the loss of manpower in internecine warfare; but after the split of 395 the mutual mistrust of Arcadius’ East and Honorius’ West was to give Alaric’s Gothic armies a crucial opportunity to play one Emperor against the other and militarily overshadow both.

In the case of Rome’s principal contemporary equal, Han China and its heirs, a succession of weak rulers in thrall to a feuding court bureaucracy and the repeated accession of minors led to instability and decline at the centre of power and ultimately to successful military challenge from the provinces. Once a regular succession of competent adults to the throne failed, central power weakened and the Emperors became puppets of their ministers, as with several initially successful Chinese dynasties after the Han, most notably the T’ang in the ninth century. In Japan, the powerful court dynasties surrounding the throne in the ninth century and afterwards even kept the throne restricted to under-age rulers in order to secure a succession of lucrative regencies. In Sassanid Persia from 226, a more centralised state with a more unified army than that of its Parthian predecessor, a long-lasting dynasty survived under rulers of varying merits with only a few internal non-dynastic coups, e.g. in 590. This was probably due to the overwhelming power of the central as opposed to provincial armies, which coup-prone Parthia had lacked.

No provincial warlord would think it worth challenging a State with overwhelming military superiority. That factor probably kept most ambitious Roman provincial commanders from challenging the Emperor in Rome until the rule of Nero deteriorated and made the central authorities vulnerable in 68. Once there was a hiatus in Rome or one commander dared to rebel, it was open season for would-be rebels to join in, as in AD69, 193, and 260.

The reliance of the Roman political system on the merits of one man, as introduced by Augustus, had ended earlier instability. The repeated politicomilitary struggles and usurpations of the period from the Gracchi to Actium had been caused primarily by the feuding over power of the rival senior noble families and ambitious new men of the late Republican patriciate, and centred on the senior provincial governors’ possession of armies. Ultimately, with Julius Caesar and then Octavian-Augustus, only one political leader was left with full control of all the Republic’s armies, and the latter, as ‘Princeps’, ensured that the armies remained loyal to him and his family and that the political system was unobtrusively turned into a monarchy while remaining technically a republic. Unlike Caesar, he did not flout the cherished ‘mos maiorum’ and he enabled the Senatorial aristocracy to live within the fiction that the traditional constitution was being maintained, once he had slaughtered all real or potential serious challengers.

The succession problems of his dynasty are well known, and were survived by luck as much as by good judgement in the case of two mentally unstable rulers (Caius ‘Caligula’ and Nero) who were murdered and one civil war (AD69). Luck could have been better and there have been no tyranny, for example if Tiberius had been succeeded by his original heirs, but his nephew Germanicus, his son the younger Drusus, and the former’s son Nero Caesar all predeceased him, possibly violently. Caligula was only Germanicus’ third son and was an unlikely successor until the destruction of his elder brothers, Nero by the jealous Sejanus and Drusus (II) by Tiberius for betraying his family to Sejanus. Nor did Caligula’s worst traits become apparent until after a serious illness months after he succeeded Tiberius in 37; did this illness emotionally unbalance him?

Britannicus was only in his early teens when Claudius died, and the latter had political reasons for advancing his older stepbrother Nero as senior heir. Nero was more closely descended from Augustus and his mother’s family were popular with the troops.¹⁰Possibly Claudius even suspected that Britannicus’ nymphomaniac mother Messalina had used one of her lovers to father the boy, or that his enemies, led by Nero’s mother Agrippina, would say that. But an older and more viable Britannicus could have succeeded his father as an adult around 60–62. Claudius might not have adopted his stepson Nero as senior heir, or Claudius discovered Agrippina’s poison plot in time in October 54. The Julio-Claudian succession system would have worked then and not have seemed inferior to that of the Antonines in retrospect. The Empire would not have had to wait to 96–180 for a run of good Emperors.

Thereafter the Roman Empire had been lucky in its transmission of the succession from 96 to 180, no ruler except Marcus Aurelius having an adult close relative to succeed him. But the succession was not always smooth, as with the mysterious political executions of four consuls early in Hadrian’s reign and the early death of his chosen heir Aelius Verus. It was unlucky thereafter. There was not a conscious system of choosing the best man as Edward Gibbon and other historians believed; if an Emperor had a son of whatever age or quality, as Marcus with Commodus, the choice of heir was clear, and the army would probably have baulked at accepting any substitute.¹¹ Marcus would have had problems had he proposed to set his incompetent son Commodus aside, assuming that the latter’s faults were already visible by the time of Marcus’ death; the most blatant acts of Commodus’ misrule occurred after some years in power. The new Emperor (aged nineteen) was taken advantage of by flattering Court favourites such as Saoterus and later Cleander, a perennial problem for a vain and impressionable young autocrat (as shown by the initially good Nero).¹² For that matter, it should be remembered that Commodus had a twin and a younger brother, both of whom died young; had either of them replaced him on his murder in December 192 there would not have been a succession crisis and civil war then either. Instead the chosen new ruler, the competent but brusque and disciplinarian general Pertinax, alienated the over-indulged Praetorian Guard and was soon murdered too.

The run of poor or easily challenged rulers in the third century was not inevitable, and some Emperors ducked their responsibility to provide an adequate and unchallenged heir, most crucially Septimius Severus with Caracalla and Geta in 211. It should also be remarked that the Empire had been lucky in that a new civil war did not erupt in 97–8 after the unexpected extinction of the Flavian dynasty. The shaky regime of the elderly, obscure, and heirless Nerva, defied by its own guardsmen, swiftly adopted a powerful and charismatic military commander (Trajan) as its heir to ward off another civil war. Nerva was fortunate to avoid the fate of the similarly placed Galba thirty years earlier. There was no certainty in 193 that Pertinax would not be able to control the Guard, or that he would be killed rather than just defied as Nerva had been in 97. Had he been more careful or tactful, this veteran commander (aged sixty-six in 192/3) could have averted murder and civil war and passed on the throne to his chosen heir, either Septimius Severus, an earlier protégé, or the latter’s rival Clodius Albinus. Instead, the son-less second and third century Emperors sought to bolster legitimacy by adopting their heirs, which was bizarre at times, as when Elagabalus adopted his cousin Alexander Severus, four years younger than him. Septimius Severus retrospectively had himself adopted posthumously by Pertinax.¹³

The habit of provincial military challenges to the centre of power had been a threat to the Roman polity ever since the emergence of powerful provincial armies. Sulla had used his armies in Greece to overthrow the regime of Marius’ heirs in 83–2BC, Caesar had marched on Rome from Gaul in 49BC, and in 43 the Senate was helpless before the triple alliance of Antony, Octavian, and Lepidus. It recurred throughout the early Empire at times of crisis and uncertain leadership in Rome, potentially (though aborted) with Gaetulicus in AD39 and Scribonianus in 42 and fully with the civil war of 69. Augustus, lucky in the possession of a large family, had where possible kept the armies in his loyal and competent male relatives’ hands. It resumed with the instability following Pertinax’s murder in 193 (though there was an abortive revolt as early as the rumours of Marcus’ death in 175), and became endemic after 235. But once the chances of a revolt succeeding had become greater, with the lack of an recognisably dominating ruler and stable dynasty in Rome, the temptation to challenge the Emperor became greater. One successful revolt bred another, and in turn the preoccupation of the current incumbent with survival meant that external enemies were emboldened.

The safest military solution to this was that adopted by the late Roman state, probably through a deliberate plan by Diocletian which Constantine reinforced; the central army at the Emperor’s disposal, the ‘comitatus’, outnumbered any provincial army and the old Augustan provinces were split into many smaller ones so that no provincial commander had enough men to risk challenging the government. Each province also now had a separate civil and military governor. This did not stop some bold commanders, as with Maximus in Britain in 383, a desperate man like Silvanus in Trier in 355, or a junior Imperial prince commanding a ‘comitatus’ on a threatened frontier, as with Julian in Gaul in 360; but it made revolt more risky and so halted the epidemic of risings in the 250s and 260s.¹⁴

In the 250s the instability at the centre, coherent military challenge from Persia, and opportunistic attacks from Germanic tribal coalitions across the Rhine and Danube – all feeding off each other – came together at a time when the Empire’s manpower was being undermined by plague. The results were catastrophic. But what if there had been greater political stability within the Empire at this juncture?

The 250s and after: dynastic mischance and its exploitation

Apart from the military advantages of a firm and continuing response to outside threats, greater political stability from a secure succession process would have enabled the Empire to call upon its full revenues for and troops from all unaffected provinces to aid the government and military. The plague of 252 would still have diminished both, making the chances of defeat and a civil war higher, particularly with the opportunistic Persian Great King Shapur I ready to invade. The defeat of the Eastern armies, sack of Antioch, local power-vacuum, and seizure of the politico-military initiative by Odenathus and Zenobia of Palmyra in the 260s were probable if the Rhine or Danube wars had tied down the Western armies. If the Emperor marched East and had no capable colleague to guard the threatened Rhine, a revolt on the latter was probable. In real life the local commander Postumus deposed Gallienus’ young son and set up a breakaway Gallic-Spanish-British realm in 260 and it was not reconquered until Aurelian had dealt with Palmyra in 272–3.

Recovery should have been quicker without the multiplicity of revolts in 259–67, and the strain on resources to pay for the enlarged Diocletianic army and civil service less without the economic dislocation caused by ravaging across many provinces. The efforts usually attributed to a bureaucratically minded Diocletian to secure adequate manpower for vital professions (the military and agriculture in particular) centred on the solution of making them hereditary, while economic problems were tackled by similar legislation. Inflation was solved by being banned, the reaction of a despot like a modern Third World dictator. Diocletian’s crucial lack of a male heir led to an ingenious attempt to solve the endemic problem of the succession by another administrative solution. The two new Emperors, of East and West, would each adopt a competent adult heir who would serve as deputy ruler or ‘Caesar’ under him before succeeding to the throne. Was this only suggested because Diocletian had no son? This idea was unworkable given human nature and the desire of the men involved to pass on their power within their families, and it was duly wrecked by a complicated power struggle among Diocletian’s heirs after he retired in 305.

The outcome was the personal ascendancy of Constantine as sole ruler in 324, followed by his own attempt to divide up the Empire among his sons and nephews, which also collapsed in bloodshed in 337. The successive bouts of political instability and civil wars which followed saw no dynasty surviving with stable adult male rule for more than a few decades, although the loss of manpower in civil war did not immediately affect the Empire’s survival when it lacked external challengers. But indirectly the effects of dynastic strife commenced the process of political disintegration; Valentinian I’s unwise choice of his brother Valens as his co-ruler in 364, criticised at the time, presented the East with the man who mishandled the Gothic crisis of 376–8 and was killed in Rome’s first serious military defeat at German hands since the 250s. After the death of Valens and destruction of the Eastern army at Adrianople in 378 the Goths, initially a flood of refugees from Hunnic incursions into their steppe homeland not a hostile invading army, were able to maintain their own polity within the borders of the Empire. Nominal military vassals at first, after their accommodation with Theodosius in 381, their Gothic-commanded forces were outside the Roman chain of command, and were able to exploit the vacuum in leadership that followed Theodosius’ death in 395.

At this point, the Empire’s physical loss of control of the provinces, and their manpower and revenues, commenced, and the Germanic warlords within its borders began to be a serious military challenge. The spiral of Decline and Fall, began, more specifically meaning a growing loss of resources and military power in the West at a time of rising challenges from unchecked Germanic warlords who could not be intimidated or bought off indefinitely. There has been much argument over the size of the barbarian hordes, the amount of damage and economic dislocation, and the possible exaggeration of their depredations. But the loss of Imperial political control of the outlying provinces, leaving a rump state at the mercy of its Germanled armies in the 460s, speaks for itself.

The survival of the Western Empire: feasible with better luck?

A larger state: a match for the Germans and the East?

The survival of Theodosius I, only forty-seven at his death in 395, for another decade or two, putting him in the position to combat the Germanic crossing of the Rhine into Gaul in 406, should have made that attack containable like the previous invasions of the mid-270s and 350s. Probably Alaric the Goth would not have risked his attacks on Italy from 402. Even if the Empire had faced at least one of these challenges, an experienced adult Emperor would have been in a stronger position to meet them than the regent Stilicho. The Germans would not have had the opportunity to spread unchecked over Gaul, Spain, and later Africa from 406, detaching rich provinces from the Empire and so reducing its ability to pay for a militarily superior central army that could defeat tribal based armies led by opportunistic German warlords. Barbarian kingdoms would not have coalesced around successful warlords in Roman territory and become more powerful than the Imperial armies, or the central Imperial army been reduced to over-reliance on king-making German generals with their own loyal entourages, e.g. Ricimer and Gundobad. Crucially, the Vandals would not have been able to set up their kingdom in Africa from 429 and deliver the major blow of the sack of Rome in 455, ravaging the Italian (and other) coasts for decades thereafter and undermining trade. The Goths would not have been able to operate freely in Italy against an undefended city of Rome and a militarily weak government in Ravenna after Stilicho’s assassination in 408, or set up a kingdom in Southern Gaul in 418.

Thereafter, a more powerful Western Roman army, though probably still with a major autonomous allied German contingent, would have been available under Aetius to meet the attacks of the Huns. The latter might still have been undefeated until 451–2, due to earlier concentration on the weaker East, but German refugees from Attila’s empire fleeing to the West would have reinforced the Roman army. Even in the circumstances of a weakened Empire that had lost control of its African corn-supplies and much of Gaul, Aetius, who had usefully lived in exile among the Huns earlier and knew their tactics, was able to muster a Romano-German coalition to defeat Attila’s incursion into Gaul. How much better would he have fared had the Empire not already lost much of its revenues and power?

The assassination of Aetius by his jealous sovereign Valentinian III in 454 led to a vacuum in Roman military leadership, a new power-struggle at court, and Gaiseric’s physically and psychologically damaging sack of the capital in 455. It also led to Aetius’ successor Petronius Maximus, allegedly implicated in his murder, seeking alliance with the Goths, which Rome had avoided since 408, and the latter securing a free hand to operate in Spain and extend their power there. What if Aetius had escaped the attack and overthrown his monarch? (He was already rumoured to be intending to install his son Gaudentius as the next ruler, married to Valentinian’s daughter.)¹⁵ As praised by his contemporary and panegyricist Merobaudes in his laudatory poems of around 439 and 443, the indefatiguable Aetius had restored order to Gaul in the 430s and defeated Germans and ‘bacaudae’ alike – and even held up Gaiseric’s advance in Africa temporarily by bringing in Aspar and Eastern troops in 435. The relative stability he brought after three decades of chaos speaks for itself, as does his successful leadership of the Romano-German coalition to defeat Attila somewhere near Chalons in 451. The collapse of Roman power in Gaul and Gaiseric’s attack on the capital only followed his death, so what if his rule had continued?

The defeat, containment, and death of Attila in 451–3, followed by Aetius’ contining ascendancy at court, would have enabled Aetius to recruit many of the subject tribes who revolted against the Huns in 454 to be allies of Rome. The Western Empire would have continued as a major military power into the later fifth century with most of its provinces intact, and Aetius had competent officers to succeed him in power such as Aegidius and Majorian. In real life Aegidius ruled parts of northern Gaul as a Romano-Gallic warlord after 455, and Majorian became Emperor in 457. Such competent and energetic rulers, with appropriate armies, would probably have dissuaded or defeated further provincial revolts, and if North Africa and its corn supplies had still been in Vandal hands reconquest would have been a priority. The Empire would have been manageable without outlying regions of Gaul (lost to Goths, Burgundians, and Franks) and Spain (lost to the Suevi).

Either a strong military leader or a legitimate Theodosian would have served as a focus for stability in the 460s and after. With or without the continuation of the dynasty of Theodosius, the Empire’s military leadership and control of resources would have been adequate for survival on the politico-socio-economic basis of the state of the fourth century. It would have been the military equal of Justinian’s Eastern Empire provided that it was not undermined by further civil wars. Therafter it would quite possibly have been an unviable target to conquer. Indeed, the notion of an Eastern attack on the West owed much to the aggressive (and ultra-Christian) ambitions of Justinian himself, a conqueror, builder, and would-be theological arbiter on the scale of Constantine the Great.

But what if this humbly born Balkan peasant-boy had never become Eastern Emperor? He did not take the throne by obvious military talent as had the humbly born Balkan Emperors of the later third and early fourth century like Claudius II, Aurelian, Maximian, Galerius, and Constantius I.

His introduction to the capital, Court, and politics was due to his (childless) uncle Justin becoming an ex-ranker Guards officer and inviting him to the capital around 500 as his protégé. The disputed succession to eighty-eight year old Emperor Anastasius in July 518 then saw Justin, commander of the ‘Excubitors’ guards regiment, selected as a compromise candidate instead of the ambitious civilian minister Celer or Anastasius’ nephew Hypatius.¹⁶ The latter was the ‘legitimist’ choice of the Nika rioters to replace Justinian in 532.

Justin, then aged around sixty, may well have been seen as a stopgap by his selectors, less dangerous and strong-willed than Celer and a respectable Catholic in place of the unpopularly pro-Monophysite Anastasius. The latter had recently had to abandon his pro-Monophysite policies due to a Catholic military revolt led by the charismatic general Vitalian, who still had an army to hand in 518 and so was a potential Emperor. Instead, Justinian took control of the administration as his semi-literate uncle’s civilian strongman, had Vitalian bought off with a consulship and quietly murdered, and was duly made co-Emperor. In 527 he succeeded his uncle, despite tension over his marrying the ex-actress and alleged prostitute Theodora. But would the West have been invaded in 533–7 had Celer, Hypatius, or Vitalian secured the throne in 518?

Had the Vandals been kept or driven out of Africa, the only significant military challenge the West would have faced after Attila’s death would have been the armies of Theodoric the Ostrogoth (encouraged to leave the East after defying its rulers through the 480s) around 490–3. In real life Theodoric was able to overthrow the post-Imperial regime of Odovacer in Italy, but a strong Western central army could have held him at bay as Stilicho had done to Alaric in 402. Indeed, there is a possibility that a militarily strong West could have sent troops to assist the Catholic military revolt of Vitalian against Anastasius around 513 and helped to save the cause of orthodox religion in the latter’s capital. The West had intervened successfully in the East before, as Constantine had defeated the suspiciously lukewarm pro-Christian Licinius in the name of orthodoxy in 324 and Julian had attacked his uncle Constantius II in 361. A Western Empire controlling most of its provinces (and maybe with Gothic allies from Aquitaine) was a formidable foe for an East whose armies were in disarray, and in real life Anastasius had to treat with Vitalian as his armies could not defeat him. A West that had held or restored the Rhine frontier could have militarily outmatched Anastasius’ divided armies in 513–18, especially if its Catholic Emperor allied to Vitalian in the cause of orthodoxy.

A smaller state

Alternatively, if the Western Empire had been weakened by invasions and poor leadership in the fifth century, a core of Mediterranean provinces, probably minus Gaul, the Rhineland, and Britain, could have survived as a small state into the 530s. This scenario was possible from the recovery of the Empire in the 410s under Constantius III, which left a Gothic ‘federate’ state ruling south-west Gaul, Germans roaming at large in Spain, and the Rhineland and Britain permanently lost. Had the Vandals still managed to conquer North Africa, in which a power-struggle between Aetius and Boniface to control the regency in Rome around 428–33 aided their advance, the West would have faced a new pirate kingdom with a fleet raiding Italy, worryingly based at the ancient foe Carthage, and the loss of African corn and revenues. That need not have led to the devastating sack of Rome in 455, the result of the murder of the military leader Aetius by Valentinian III and the chaos that followed, which gave Gaiseric the excuse to intervene. But the surviving Empire would have been weakened further by regular raids and the defeat of a retaliatory expedition (e.g. that of Majorian in 462) would have been a signal for more political instability.

A massive expedition from East and West like that of 468 could have evicted the Vandals if competently led. The expedition apparently consisted of a huge armada of 1,100 ships according to Byzantine sources, and was only defeated due to skilful use of fireships by Gaiseric as the Eastern navy reached the Tunisian coast – against which the wind helped in trapping the Roman ships so they could not sail or row to safety. The Eastern commander Basiliscus was accused later by the sixth century historian Procopius of accepting a bribe to delay the attack, which enabled the Vandals to prepare and use their fireships (though nobody could have known that the wind would change and trap the Eastern ships against a lee shore). Once an army was ashore it had a reasonable chance of defeating the incumbent power, then blockading Carthage into surrender and driving resistance out into the deserts, as carried out by Scipio Africanus in 203–2BC and Belisarius in AD533. Had the alleged overwhelming size of the combined Eastern and Western force secured victory, the corn and tax revenues of prosperous North Africa would have been restored to the Western Empire and it could have raised the men to tackle the smaller kingdoms of German-held Spain (e.g. the Suevi).

A success by Majorian in Africa in 462–3 (his expedition was defeated en route and he was overthrown by Ricimer) would have been more useful in regaining Spain than a success by Anthemius and Basiliscus in 468, as by 468 the Goths were more entrenched in Spain. But either expedition could have regained Africa and tipped the balance of resources in the Western Empire’s favour, at least securing it control of nearer parts of Spain. This would have helped the Empire to hold onto central and southern Gaul too, with or without a war against the Goths who only gained control of the Auvergne region around 470. The North was partly ruled by survivors of Aetius’ Gallic army, under Aegidius and his son Syagrius, until 486 and this force (based at Soissons) was likely to rally to the Empire if the latter was in the ascendant. In that case, a revived Mediterranean-based Western Empire that Majorian or Anthemius had secured North Africa would have been viable for decades but for an external threat or civil war.

This smaller Empire could still have avoided ruinous civil war if there had been a stable succession within one dynasty, which could have been that of Valentinian III, his son-in-law Olybrius, and grandson Areobindus from 425 to around 510 or Anthemius’ family, or the role could have been taken by a series of strong military leaders such as Aetius, Majorian, Marcellinus, and Julius Nepos. It was actual or potential Roman civil wars that gave the Germanic tribal leaders their opportunities, from the time when Stilicho neglected tackling the invasion of Gaul in 406–8 to concentrate on the Eastern succession to the power-struggle in Rome in 455.

Holding onto its remaining Mediterranean territory with a viable army led by Roman generals, the Empire should have avoided becoming the puppet of German officers such as Ricimer. The latter’s emergence as commander-in-chief and Emperor-maker in the later 450s would have been inconceivable had Aetius been alive, and in any case he could have been overthrown by his Emperor (for instance the capable Majorian) as in the East the Germanic commander-in-chief Aspar was killed in 467 by his puppet Leo I. Majorian’s defeat of the Vandals, rather than his own defeat, in 461–2 would have strengthened both him and his rump Empire, and in 467–8 the Eastern fleet could have defeated Gaiseric and reconquered North Africa for the new Emperor Anthemius.

If the Western Empire had defeated the invasion of Theodoric in circa. 492–3 the state should have had no more military challenges into the sixth century, when it would have posed a tempting challenge to the aggressive Eastern ruler Justinian. (Being orthodox instead of Arian like the Gothic kingdom of Italy would not have saved it; the East had attacked the West already in 351–2, 388, 394, 425 and 467). The likelihood is that Justinian would have sought to reabsorb the West even had it been ruled by a Roman Emperor rather than several disunited heretic German kings, but he might have regarded this as a lower priority and taken on the Sassanids for a longterm war first. The attempt at reunification would have been risky, as in real life Justinian could only spare small armies to carry it out (due to the Persian threat to the Eastern frontier) and if the West had survived he would have been facing a Roman ‘comitatus’. But he could still have attempted it, particularly in order to remove a theologically unorthodox, possibly Arian, Emperor or to exploit a civil war. It appears from the contemporary accounts of the great plague that commenced in 542 that this cost millions of lives, and thus would have reduced military capability to send an adequate army West thereafter.¹⁷ But an attack before that date was still feasible, with the caveat that independently of internal Roman politics a new wave of Germanic, Asiatic (Hun and Avar), and Slavic attacks was disrupting the Balkans by the 540s.

Assuming Justinian and his general Belisarius had succeeded in invading the West, there would have been one Empire as last seen under Theodosius I in 394–5. The Imperial writ would have run from the upper Euphrates and upper Nile valleys to Southern Gaul and the Straits of Gibraltar, with or without a surviving Rhine frontier (which even if it had not been breached in 406 could have fallen to a migration of refugees from Attila’s empire in the 430s). The West could then have been given to a separate ruler in a new division of power in the later sixth century, as appears to have been considered as an option in real life by Tiberius II in 582. At the time, the possible division of power between Tiberius’ son-in-law Maurice (East) and Germanus Postumus (West) would have left a weakened West struggling to hold back the Lombards in an Italy ravaged by Gothic wars.¹⁸ But there could have been been no Gothic occupation of Italy in 493 (or any grant of land to Germans in 476), or a more decisive victory for the East by 540 without Totila’s subsequent fight-back. In these scenarios, Italy would have been free from the presence of German settlers with homes to defend against the Eastern troops and a ruinous war would not have occurred.

Even in the circumstances of real-life 540, the East had driven the Goths back into the Po valley and their demoralised remnants were reduced to inviting Belisarius (evidently admired as a chivalrous foe) to become their ruler. (This did not do his reputation any good with his suspicious Emperor.) The Goths’ initially successful recovery in the early 540s was partly due to the capable Totila taking on the command; partly due to Justinian’s recall of Belisarius and many of his soldiers, and partly due to Roman manpower losses in the devastating plague of 542–3. The Persian attack on Syria and sack of Antioch made the recalls probable, and the plague meant that Justinian could not send an adequate army West until his nephew Germanus’ mission in 550 (aborted by the general’s death).¹⁹ Justinian, like George Bush in Iraq in 2003, seems to have been too eager to proclaim ‘mission accomplished’ and not alert to the possibility of revolt. But what if Totila had been won over or killed quickly, or Belisarius had not been recalled? Italy would have been in a far better condition to meet any new invasions in the later sixth century, though still denuded of manpower if the plague of 542 had occurred. As it was, the land had already been ruined by decades of Romans and Goths fighting over it before the Lombards moved in, making conquest easier, and Justinian’s extortionate tax demands did not help agricultural recovery either.

The survival of the Western Empire during 476–535 would have provided a ‘comitatus’ to be incorporated into Justinian’s army at the reconquest, and a stronger force available to hold back the Lombards. The Western Empire, possibly incorporating southern Gaul and Spain as well as Berber-raided North Africa, should have survived as a viable political entity into the seventh century. Crucially, providing there were good relations with the current Eastern Emperor it would have been able to send him troops to fight the Persian incursions in the 610s, and later to fight the Arabs. But if the plague of 542 had carried off up to half the population, as estimated by Procopius, military manpower (and tax revenues to hire troops from outside the Empire) would have been smaller in the later sixth century than in the fourth and fifth. The military challenges from the new nomad threat, the Avar empire (centred in the Hungarian basin and Wallachia like Attila’s), and the mass-movements of its fleeing enemies (e.g. the Lombards) would have prevented a peaceful and prosperous future for the united or divided Empire of the period 570–600.