Roman ‘Grand Strategy’ in Action? Claudius and the Annexation of Britain and Thrace

Abstract

A longstanding debate among ancient historians and students of Roman frontiers concerns the reality and effective reach of Roman imperial policy. Certainly when new military commitments were involved, the slowness of supply and information meant that major moves had to be planned well in advance. This paper focuses on the provincialisation of Britain and Thrace in a.d. 43 and c. 45. The dating evidence provided by tree rings, coins and milestones suggests that logistic preparation for the invasion of Britain started at least two years before the event. This pattern, of a newly installed Emperor immediately initiating a campaign on the northern frontiers, allowing two years for logistic preparation, is seen no fewer than seven times between Caligula and Caracalla.

THE ‘PROBLEM OF CLAUDIUS’

There is a very real ‘problem of Claudius’.¹ One aspect is that the literary sources for his reign are somewhat lacunose and lopsided. The manuscript tradition of Tacitus’ Annales has left us bereft of the first half of Claudius’ reign (a.d. 41–54), with his predecessor Caligula (37–41) missing altogether, while Dio's account survives partly as a discontinuous series of Byzantine excerpts and epitomes. The only complete narrative we have is Suetonius’ heavily biased biography. His Life of Claudius famously presents a marginal member of the Augustan house, an object of ridicule with otherworldly occupations, who was pulled from his hiding by the Praetorians as a convenient replacement for his just eliminated nephew. Suetonius paints a dark picture of a weak and fearful man who from his earliest childhood had suffered ‘so severely from various obstinate disorders that the vigour of both his mind and his body was dulled’ and was ‘not thought capable of any public or private business’.² When pushed into the purple, he tended to rely on those around him, notably his freedmen and wives.³ Calumny and gossip, the stock of Roman senatorial historiography, have fed modern versions and fictions, notably Robert Graves's novels and the memorable BBC series based upon them, leaving an indelible imprint on all of us.

The real ‘problem of Claudius’, meanwhile, is that contemporary documents, epigraphical, papyrological and legal, present us with a very different picture. It is as if

the shut-in scholar suddenly emerged, with energy, initiative, ideas. Ambitious projects were undertaken: Britain was to be annexed, a harbour built for Rome. Reforms were announced, opening up membership in the Senate, tightening Rome's finances. Claudius tirelessly held legal hearings, he issued floods of edicts. (…) Claudius was not a totally weak-minded man; he showed, more than Tiberius or Caligula, energy in the practical business of government.⁴

Few principates illustrate with greater acuity the contrast between the narrative source material and the legal, documentary and indeed archaeological evidence for what actually happened on the ground.

What can Roman frontier studies offer to bridge the gap? A strong asset is the shifting picture of military investment along the fringes of the Empire, the resolution of which still increases with each decade. It is important that we realise that major shifts in this area normally issued from specific imperial mandata given to governors at the start of their tenure,⁵ or as follow-up instructions, as Corbulo was to experience in a.d. 47 when summoned to place his praesidia on the south bank of the Rhine.⁶ Where it came to new financial or manpower commitments the scope for autonomous action by governors was limited.⁷ Under Claudius, interestingly, military deployment along the Rhine and Danube was subject to considerable changes. The obvious comparandum in the written sources would be those items that fall under the modern heading of ‘foreign policy’. Most of these concern the Emperor's (or, formally, the senate's) dealings with client rulers or ‘friendly kings’ who had been retained, restored and occasionally even moved around, often after having been groomed at Rome, to rule subjected or otherwise dependent tribes or polities indirectly.⁸

CLAUDIUS AND HIS CLIENT RULERS – A REVIEW OF INHERITED ARRANGEMENTS?

Given the eventual annexation of several Roman client kingdoms, the institution is easily mistaken for a purposeful antechamber, to prepare recently subjected polities for direct Roman rule. But even for the Julio-Claudian period, the heyday of client rule, the overall picture is varied. It is true that Tiberius, in preparation of Germanicus’ eastern mission, turned Cappadocia and Commagene into Roman provinces in a.d. 17, seemingly in keeping with ‘the policy begun by Augustus of hardening the frontier by taking direct control of the former client kingdoms that bordered it’.⁹ Elsewhere, however, Tiberius proved ‘reluctant to annex new territory simply to round off the Empire’ and largely continued Rome's system of indirect dominion.¹⁰

Caligula even more relied on client kings, returning Commagene to Antiochus’ son in 38 and appointing several others, including Mithridates of Armenia, his namesake of Bosporus and the three sons of Rhoemetalces of Thrace, Polemo, Cotys and Rhoemetalces, who were given Pontus, Lesser Armenia and Thrace, respectively.¹¹ From 37, the grandson of Herod the Great, Agrippa, stepwise regained parts of the old Hasmonean kingdom, which had been largely annexed in a.d. 6, with the crown jewel, Judea, returned to him by Claudius in 41. Three years later, the Alpes Cottiae, likewise, reversed from Roman province to client polity – and then back again under Nero in a.d. 63. The overall impression is that ‘there was no firm notion that the status of client kingdom was one consciously intended to prepare for full and imminent annexation’.¹² If it worked, and the balance of benefits and costs was healthy, indirect rule often proved to be the preferable option.

On his accession, Claudius confirmed, restored or repositioned most of Caligula's appointees, several of whom had lost their dominions (and some their life or liberty as well) following reports, fabricated or not, of their implication in the ‘conspiracy’ of 39.¹³ However, in the next few years, all corners of the Roman world saw significant changes to the inherited client-rule arrangements. First, Claudius decided to turn into a province the vast client kingdom of Mauretania which had been destabilised by rebellion and incursions following Caligula's execution of Ptolemy in 40.¹⁴ Then, between 43 and c. 45, a cascade of annexations happened: Britain, Judea, Lycia, Thrace; the date of Noricum remains uncertain (fig. 1).

FIG. 1. Changes in the status of former client kingdoms and provinces around the start of the reign of Claudius.

It is not clear in all cases what exactly triggered (or was staged to justify) the transition. All we can say is that Agrippa conveniently died late in 43, before he could finish his suspicion-arousing city wall for Jerusalem,¹⁵ having been suddenly struck, while attending games at Caesarea, by a severe pain in his belly which ‘began in a most violent manner’.¹⁶ About two years later, Rhoemetalces III of Thrace was reportedly eliminated by his wife/cousin/co-ruler Pythodoris II, producing unrest – and a time-honoured excuse for Rome to interfere.¹⁷ The question may legitimately be asked how far the imperial government was overtaken by such events and whether Claudius’ splitting-up of the great Balkan command in a.d. 44 was a consequence or the cause of the events in Thrace.¹⁸ Recent studies underline the foresight and strategic sense of Claudius’ new provincial commitments.¹⁹ There is more than a suggestion here that a systematic review of existing arrangements may have taken place in the first years of his reign – and that Rome was a less passive player in all of this than our literary sources may suggest.

In this paper, I will focus on two cases, Britain and Thrace, which were turned into Roman provinces in a.d. 43 and around 45, respectively, having been ruled by client kings for several generations. For Thrace, the prior client status is well known,²⁰ as it was amply expressed in acts of military assistance²¹ and, most visibly, in the kingdom's coinage, pairing Roemetalces I and III with Augustus and Caligula, respectively.²² After the bloody suppression of the Thracian revolt in a.d. 26, stirred up by rumours that the native levies were to be turned into regular auxilia,²³ Roemetalces, who had supported the Roman cause, was continued.

Incomparable as Late Iron Age Britain may seem to Hellenised Thrace, the Roman perspective was that the island had been formally subjected by Caesar and the relevant part of it ruled by dependent kings ever since. In a series of stimulating studies, John Creighton has shown how the coinage of the southern and eastern British dynasties broadly adopted the political imagery of the Augustan peace, sharing some of its types with Mauretania and the Bosporan kingdom,²⁴ and also how their capitals, funerals and other trappings of power reflected the conscious appropriation of Roman models, possibly informed by youthful experience at Rome as obsides – the usual grooming of client rulers.²⁵ Augustus received two British refugee kings, Dubnobellaunus and Tin[comarus],²⁶ and appears to have considered, or just threatened, a change to direct rule several times, ‘since the people there would not come to terms’.²⁷ Shortly after his adoption in a.d. 4, Tiberius happened (?) to be in Boulogne on the Channel coast where he received embassies to congratulate him, not just from Aizanoi in Asia Minor²⁸ but surely from the British client kings as well. After his accession, ten years later, he continued the inherited situation whereby Britain was ‘virtually Roman property’ with ‘heavy duties’ imposed on both imports and exports, so that there really was ‘no need of garrisoning the island’.²⁹

Tiberius’ principate was enveloped by the reign of Cunobelinus, king of the Catuvellauni (c. a.d. 10–40/43). The overtly Romanising style and appropriation of imperial portraiture of his coinage are difficult to see as anything other than political expressions of allegiance, if not evidence for the use of professional Roman dies or die-cutters.³⁰ It is not clear why, or even whether, the British system broke down in the late a.d. 30s. The shifting ‘tribal’ coin distributions of the preceding decades have long inspired narratives of aggressive expansionism, notably by the Catuvellauni, Cunobelinus' brother Epaticcus seemingly carving out a big swath of the southern kingdom of the Atrebates around a.d. 25, but such shifts need not reflect more than intermarriage or territorial rearrangements within or between the dynasties of client rulers.³¹ All we know is that one of Cunobelinus’ sons, Adminius, had fallen out with his father in 39/40 and sought refuge with Caligula who was then just engaged in his nebulous northern campaign.³² The ensuing anecdote of the Emperor ordering his troops to collect seashells on the beach in a mock victory over Oceanus sums up the problem of our literary sources.³³

PASSIVE PLAYERS?

A running theme of Greek and Roman historiography is the variability of human nature, culminating in the freak show of ‘bad Emperors’ presented to us in the hostile senatorial tradition. The ‘psychological’ interest tended to place centre stage the protagonists’ actions, spontaneous or in response to events, so that history could become a mirror of virtues and vices. Suetonius’ Caligula was capricious and cruel, Claudius a weak-minded marionette under the influence of the people around him. Unsurprisingly, Claudius in a.d. 43 is presented as reacting to extraneous events. After Cunobelinus had died, ‘Aulus Plautius […] made a campaign against Britain; for a certain Bericus [obviously Verica of the southern dynasty], who had been driven out of the island as a result of an uprising, had persuaded Claudius to send a force thither’.³⁴

There is a wider dimension to Roman passiveness in ‘foreign affairs’. A time-honoured theme from the days of the Republic was the ideology of bellum iustum: Rome never actively sought war and only reluctantly took over direct rule, preferably on a formal request by local ‘friends’ and allies.³⁵ Modern observers are often inclined to see such attitudes continued under the principate, as in the case of Thrace whose annexation c. a.d. 45 is believed to have been triggered by a palace revolt – not by expansionism on the part of Claudius³⁶, although there really is little evidence either way. A specific, and legitimate, cause for action, certainly in the imperial context, was the violation (iniuria) of Roman honour (decus) which, in the context of client rulers, could be caused by any sign of arrogance (superbia) or failure to acknowledge Rome's maiestas.³⁷ Official communication would have framed the interventions in Britain and Thrace as provoked by internal events that had offended or violated Rome's honour, rights or interests. Their survival in imperial historiography suggests that the supplications of Adminius and Verica in a.d. 39/40 and 43, respectively, were eagerly received, and publicised, as requests for help by ‘friends’.

There is also the modern debate about the mirage of Roman imperial power – far-reaching and mighty in its rhetoric and imagery, limited and ‘thin’ in its effective means and reach. This is not the place to trace developments since Fergus Millar's landmark study of The Emperor in the Roman World which presented the princeps’ role as essentially passive, manifesting itself mostly in response to petitions and embassies.³⁸ The conditions of the Empire – vast spaces, thin bureaucracy, slow communication, costly transportation – meant that there was ‘little scope for the active dissemination of “policy” or “propaganda”’.³⁹ However, following Zanker's seminal work on the Augustan age,⁴⁰ art historians have explored the conscious creation and projection of imperial imagery,⁴¹ while other areas of imperial initiative and agency have been noted, not least by Millar himself.⁴² A notable shift has occurred in the last two decades. Ancient historians and classical archaeologists are now more aware that the literary, legal and documentary evidence ‘privileges the reactive and adjudicatory roles of Emperors and governors and underplays the regulatory and extractive functions of the imperial state’.⁴³ Signal contributions in this vein include Ando's exposition of Roman provincial government and Mattingly's analysis of the imperial machinery, tellingly titled ‘Ruling regions, exploiting resources’.⁴⁴

A special branch of this debate, concerning the elusive ‘Grand Strategy of the Roman Empire’, was sparked off by the homonymous study by Edward Luttwak, equally influential to Millar's Emperor and published almost simultaneously – but maximalist this time.⁴⁵ Millar himself replied authoritatively,⁴⁶, pointing out the practical and institutional obstacles that stood in the way to the promulgation and pursuit of informed policies, let alone strategies. Isaac has questioned whether Rome had the required grasp of geography, the intelligence position and indeed the intellectual vocabulary to formulate a conceptual strategy.⁴⁷ Despite Wheeler's revindication,⁴⁸ the present consensus among ancient historians is that the Principate did not develop a ‘grand strategy’ in any meaningful modern sense, and that the term is simply misplaced.

Meanwhile, an attempt has been made to revive the notion in a more modest form.⁴⁹ While it is true that Roman imperial government, if it was to work at all, had to radically decentralise and delegate much of its acting authority, especially to provincial governors, this did not obviate the need to take many crucial decisions at the centre, notably including the choice of governors and other appointees, the mandata given to them and the disposition of troops and frontier installations. ‘Emperors made decisions about how to allocate resources to meet objectives empire-wide, and thus definitely thought about grand-strategic issues’.⁵⁰ Precisely because the resources at the Emperor's disposal were limited, there would have been an inbuilt incentive at least to think strategically, set priorities, perhaps even formulate policies,⁵¹ although any such statement would usually remain secret.⁵² All of this is certainly true where it comes to the arrangements with client rulers, as any alterations would entail new military and administrative commitments. We can expand this paradox: precisely because communication was slow and transportation cumbersome, the Emperor, if he wanted to get something done, would have had to plan things well in advance. Unsurprisingly, preparations for Hadrian's first inspection journey, along with the works to be inspected themselves, appear to have been set in motion some two years in advance.⁵³

In the shady world of Roman imperial government, then, logistic preparation, if traceable, can become a hidden pointer to premeditated policies. Wilkinson has exposed a glaring contradiction in the tradition of Caligula's northern campaign: ‘the expedition was on a whim and speedily done and yet the sheer logistics of such an operation would have required massive preparations and some sort of planning’.⁵⁴ There is clearly something wrong in Suetonius’ account: ‘on a sudden impulse’ the Emperor ‘was seized with the idea of an expedition to Germany […]. He assembled legions and auxiliaries from all quarters, holding levies everywhere with the utmost strictness, and collecting provisions of every kind on an unheard of scale’.⁵⁵ In the case of Britain and Thrace, recent archaeological advances are now starting to shed light on the sheer scale, and runup, of the logistic effort that accompanied the annexations of a.d. 43 and c. 45. If anywhere, it is here that we may find the fingerprints of Roman premeditation.

RHINE AND DANUBE: LIFELINES OF MILITARY SUPPLY

Claudius’ new commitments in Britain and on the Balkans entailed major redeployments, bringing about a cascade of legionary displacements. More consequentially, the new provincial garrisons needed to be reliably supplied. In Britain, no less than four legions with their associated auxiliary units would have to be paid, fed, clothed, housed, armed and part of them mounted. The reorganisation of the Balkan provinces in a.d. 44,⁵⁶ whether cause or consequence of the provincialisation of Thrace, meant that a new legionary fortress was planted at Novae on the Lower Danube to shore up the now detached Moesian command. Both new provinces had their natural lifelines. The main feeders of Britain were the Rhine and the Atlantic ports and river outlets of Gaul. The eastern Balkans faced towards the Black Sea for part of their supplies, but the Moesian garrison also relied on the Danube for tapping specific resources from the western parts of the Empire, certainly until the early second century.

The Rhine no doubt played a vital role in supplying the British campaign, conquest and consolidation. The annexation broadly encompassed with the south-eastern triangle of Britain that had seen all the major developments of the Augusto-Tiberian period: ‘territorial’ oppida, royal burials, classicising coinage, continental imports.⁵⁷ Some of the traditional continental outlets, including a handful of Channel ports in Brittany, are well known through archaeology.⁵⁸ Strabo, writing in the early years of Tiberius, says that the four most used passages to Britain are from the Rhine, Seine, Loire and Garonne, with the first route following the Flemish coast to the narrowest crossing-point.⁵⁹ A notable shift towards the Rhine had occurred in the mid-Augustan period, to judge from the concentration of Dressel 1B amphorae in the eastern kingdom and especially Essex.⁶⁰ The Samian and Gallo-Belgic wares from sites with high numbers of continental imports like Braughing–Skeleton Green and Colchester–Sheepen are broadly similar to late Augustan-Tiberian assemblages from the Rhineland.⁶¹

The Rhenish link continued after the Conquest, as can be seen through the distribution of e.g. Eifel mortaria, dating to c. a.d. 40, or Lyon ware.⁶² An archaeological classic is the trail of Dressel 20 olive oil amphorae, like those of C. Antonius Quietus and LQS, all along the Rhône–Saône–Moselle–Rhine corridor and on to Britain.⁶³ Such was the volume of this supply traffic that plans were made, in a.d. 58, for the construction of a canal between the Saône and Moselle. ‘The design of this’, Tacitus comments, ‘was that the supplies conveyed over the Mediterranean and, next, the Rhône and Saône might follow that canal, continuing via the river Moselle to the Rhine and further downstream into Oceanus, so that by removing the difficulties of overland traffic the shores of the western Mediterranean and the North Sea would be interconnected for shipping’.⁶⁴ The project eventually foundered in bickering between governors about their competences, but the point is that the canal was designed to ease the transport of military supplies (copiae) all the way to the northern Oceanus – obviously meaning Britain, where Nero had just decided to resume a forward policy.⁶⁵

A similar story applies to the Danube. By its sheer length (and through tributaries like the Sava and Drava) this river opened up vast swaths of land for the provisioning of the Pannonian, Moesian and Dacian garrisons. Around a.d. 105, the Cohors I Hispanorum veterana, tucked away at Stobi in Macedonia, had an unknown number of men on a mission ‘in Gaul, to collect clothing’.⁶⁶ The flow of Spanish olive oil, which an increasing body of expert opinion now sees as part of the annona militaris, can be followed through Pannonia and Moesia into Dacia⁶⁷, while the close correspondence of Italic amphorae and Samian ware on the Lower Danube has recently led Arioli to conclude that the supply network based on the Po valley/North Adriatic was ‘at least partly controlled or supported by the state: its main purpose was supplying the military units’.⁶⁸ The Danubian supply system had several feeders. One of them started at Aquileia. It had developed stepwise since Octavian's Illyrian campaigns (35–33 b.c.), with the hubs of Nauportus and Poetovio servicing the military corridors along the Sava, Drava and Amber Route and Dressel 6B olive oil amphorae from northern Italy and Istria serving as an archaeological proxy (fig. 2).⁶⁹ By the late Tiberian period, the Upper Danube had been sufficiently secured to join the system, which can be seen in full operation when we consider the predominantly western provenance of the amphorae and fine wares supplied to Novae, the base of legio VIII Augusta, in the pre-Flavian period⁷⁰ – certainly if it is realised that fine wares often travelled on the back of bulkier cargos, if only to profit from tax immunity for military supplies.⁷¹ According to Dyczek, this west-facing dependency indicates that logistic organisation of pre-Flavian Moesia was lagging behind, with legio VIII Augusta relying on old supply networks inherited from its former base at Poetovio.⁷²

FIG. 2. Early supply routes from the northern Adriatic basin to the Middle and Lower Danube regions as indicated by the distribution of Dressel 6B amphorae in Dalmatia, Pannonia and Upper Moesia. (Red dots; after Egri 2008, fig. 1)

FORTIFIED CORRIDORS

It is no wonder that rivers, or rather the supplies that moved along them, soon became the object of specific security concerns themselves. This is especially true for remote river tracts that lay outside the remit or practical reach of the troop concentrations that were supplied by them. Continuous escorts, whether by cavalry or Oberstimm-like patrol ships, might not always be feasible everywhere. The Danube, for example, passed through several heavily forested areas, mountain ranges and gorges, notably the Weltenburger Enge, Danube Knee and Iron Gate, with stages becoming less freely navigable upstream of Ulm. Before the age of river normalisation, shipping conditions and modes of propulsion varied considerably from place to place. As a consequence, the pre- and early-modern European riverscape boasted a great variety of ship types, each with their own areas of deployment and points of transhipment.⁷³ The resulting mosaic of river-navigation provinces required all kinds of mooring, revictualling, handling, storage and ship-maintenance facilities. This was no less true in Roman than in pre-modern times. In the context of Roman military supply, especially, exposed areas with challenging shipping conditions would be well served with static riverbank infrastructure to monitor, secure and support the river traffic.⁷⁴

The textbook example of such a ‘fortified corridor’ is the Iron Gate system in present-day Serbia, with its dense chain of forts, fortlets, towpath and harbour facilities.⁷⁵ Trajan, before launching his first Dacian war in a.d. 101, took two years for improving the towpath and other infrastructure of this crucial corridor, which included a harbour basin at Prahovo/Aquae and a 3.2 km long canal to bypass the rapids at the Iron Gate proper (fig. 3).⁷⁶ Earlier works on the towpath, in a.d. 92 and 46, had apparently been prompted by the resumption of Domitian's Dacian wars and Claudius’ new commitments on the Lower Danube, respectively.⁷⁷ Interestingly, the earliest dated works on the towpath belong to a.d. 33/4, i.e. the late Tiberian period.⁷⁸ In those years, a similar disposition was starting to take shape along the Upper Danube between Hüfingen and Oberstimm.⁷⁹ The logistic vocation of this system is highlighted by the fort at Rißtissen, with its unusual collection of storage buildings. ‘In all phases of the military use of the base at Rißtissen between c. a.d. 45/50 and 110, the fort […] fulfilled an important function as a central entrepot between Rhine and Danube, prior to the transit of commodities per ship down the Danube’.⁸⁰

FIG. 3. Example of a ‘fortified transport corridor’: the system of forts, fortlets, towers and supporting infrastructure that accompanied the Danube's passage through the southern Carpathians. Reconstruction of the situation on the eve of Trajan's Dacian wars. (Largely following Gudea 2001, 23 and 56–91, with a few modifications based on Jęczmienowski 2015 and the locations of the rock-cut towpath taken from Petrović 1990, fig. 1)

A third member of this early group of ‘fortified corridors’ is the chain of installations between Vechten and Valkenburg in the Rhine delta (fig. 4).⁸¹ In this secluded landscape, supply traffic would have run into all kinds of issues to do with security, propulsion and transhipment. The corridor broadly coincided with the zone where the tidal cycle would have governed towed river transport, so that a string of halting places with protected mooring facilities was required here. West of Vechten, the Rhine had several side branches and occasional parallel channels, potentially exposing supply traffic to the type of waterborne raiding Pliny had heard about, or perhaps witnessed, when stationed on the Lower Rhine around a.d. 47.⁸² Van Dinter has noted that most forts in the Rhine delta were situated precisely at such nodal points in the riverscape.⁸³ The few known early watchtowers were placed such that they would visually cover long reaches or complete meanders of the river. In one case, a tower was moved 30 m downstream, apparently to make up for a corresponding migration of the river bend.⁸⁴ It would appear that the river itself, or rather the traffic that travelled along it, was the prime object of security concerns.

FIG. 4. Palaeogeographical map of the Netherlands c. a.d. 100. (Vos et al. 2020, cropped to the Rhine delta, with reduced legend and military installations of the early a.d. 40s added; © Rijksdienst voor het Cultureel Erfgoed, TNO en Deltares)

THE BUILD-UP FOR BRITAIN

It is now time to look at the finer chronology. Scholarly opinion had long suspected a close connection between the British invasion of a.d. 43 and the massive investment in the Rhine delta in the a.d. 40s. ‘In all probability’, Schönberger wrote in 1985, ‘this policy arose from the fact that Claudius and his advisers had no plans for large-scale military activity in western Europe after the conquest of southern Britain, but were concerned rather with maximum security’.⁸⁵ But was this a response to the new commitments in Britain? Recent advances in Dutch archaeology now suggest that investment in the Rhine delta largely preceded the annexation of Britain.

It is now clear that the Rhine delta was a hotspot of military activity around a.d. 40. Caligula himself appears to have been there in the course of his nebulous northern campaign of 39–40, following a hibernation in Lyons and limited operations on the middle Rhine, possibly early in a.d. 40.⁸⁶ The Emperor's famous show of force against Oceanus⁸⁷, whatever its objective, may actually have taken place on the adjacent North Sea coast. The father of Brinno, a leader of the local Cananefates, appears to have been an eyewitness (and made a mockery) of Caligula's beach battle.⁸⁸ More materially, staves of wine barrels inscribed with the Emperor's name, C CAE AUG GER, have been found at both Vechten and Valkenburg.⁸⁹ It had long been intuited that Valkenburg may well have derived its Latin name, Praetorium Agrippinae, from Caligula's mother, although its special vignette on the Tabula Peutingeriana seemed unexplained by the status of a normal auxiliary fort. In the autumn of 2020, a full-blown legionary fortress was discovered, partly new, partly pieced together from the 1985–8 excavations at the Marktveld site to the south-east of Van Giffen's famous castellum at Valkenburg village.⁹⁰ Structural timbers which can now be re-contextualised as part of the fortress's defences had earlier been dated to winter/spring 39/40 and 39±6.⁹¹ An association with Germanicus’ stay in the Rhine delta is an attractive option, but whether the fortress was built with a view to a planned British invasion, and whether it was ever completed, is a different matter.⁹²

A marginally later foundation date is in evidence for Alphen aan den Rijn, two stations upstream. Major excavations in 2001–2 produced a surprising series of dendrochronological dates for the first three construction seasons at this site, starting in a.d. 41, and therewith for the building order of a fort generally (table 1).⁹³ A second striking feature is the sharp spike of Caligulan coins: almost 66 per cent of the identified coins (n = 486). Two-thirds of those never received a countermark, which ought to mean that most of the local coin pool was supplied before the damnatio memoriae of Caligula. Of the countermarked Caligulan coins (n = 110), all but one of the legible ones read TICLAVIM, obviously the initials of Caligula's successor. This particular countermark appears to belong to Claudius’ first years in power.⁹⁴ The clear implication is that, whatever had happened in the Rhine estuary in early a.d. 40, this had soon been followed up by the order, still under Caligula, to start work on a series of smaller installations in the Rhine delta.

Table 1 dendrochronological datings of various elements of the early roman fort at alphen aan den rijn in the western netherlands*

* Based on Polak et al. 2004, 277–80.

Alphen was not alone. Apart from Valkenburg (45 per cent Caligula, n = 51)) we now also have the published coins of De Meern (n = 745). More than a quarter of the coin finds from the fort site proper are those of Caligula, 78 per cent of which have no countermark. This places the foundation of De Meern also in c. a.d. 41, more or less synchronous with Alphen and Valkenburg.⁹⁵ The base at Vechten was four decades older and had a rather different coin pool, but it still has a notable Caligulan spike at 20.7 per cent (n = 696).⁹⁶ The start date of Woerden need not be much later.⁹⁷ Interestingly, the TICLAVIM countermark almost exclusively occurs in Lower Germany, with a near monopoly at Alphen and a strong presence at Valkenburg, i.e. the auxiliary fort at the village, and De Meern. This distribution suggests the local presence, in the Rhine delta and at the start of Claudius reign, of a person with the required ‘authority and standing’.⁹⁸ Whatever the purpose and fate of the fortress at Valkenburg, which may have been a short-lived affair, the evidence for the smaller installations seems to point to a sustained build-up in the years leading up to the British invasion, with the first trees at Alphen felled in autumn/winter a.d. 40/1 and work continuing into 42.

It is important to note that, when this unusual string of installations was built in the Rhine delta in the early a.d. 40s, there was also continued investment in the area to the north. At Velsen, 30 km north of the Rhine mouth, a new major installation was constructed in Claudius’ first years in power. Recently, a fortress in the vexillation size class (at least 11 ha) has been teased out of earlier finds and watching briefs, one of which had yielded a felling date of winter 42/3 for a post pad.⁹⁹ One possible scenario is that this new installation functionally replaced the fortress at Valkenburg.¹⁰⁰ Like Valkenburg, Velsen occupied a turntable position at an intersection of major overland and water routes. This was clearly a nodal point, situated in the heartland of the ‘minor’ Frisians and close to an open-air sanctuary.¹⁰¹ At that stage, their territory, roughly coinciding with the present-day province of North-Holland, was apparently still considered part of the Roman orbit. Pomponius Mela, writing in 43/4, could be taken to imply that Germania libera started on the right bank of Lake Flevo, the smaller precursor of the IJsselmeer.¹⁰² It may well be that this ‘minor’ Frisia was the area where, a few years later, Corbulo was about to introduce Roman-style administrative institutions, indicative of formal administrative integration.¹⁰³ Velsen would have been ideally placed to monitor that delicate process.¹⁰⁴

It all serves to underline that the Rhine delta forts of the early a.d. 40s were not built to mark and monitor the edge of Roman-held territory, but primarily to offer security and logistic support along a vital transport corridor through a difficul-to-monitor landscape. The corridor's southbound continuation, the fossa Corbulonis, which connected the Rhine and Meuse estuaries, is now believed to have had a precursor.¹⁰⁵ If so, this had largely silted up by a.d. 50, when a new watercourse was dug and lined with revetments, which may push the first canal to the early 40s. The British connection of this corridor is further hinted at by the altar Claudius had put up at the Brittenburg, the lost fort at the Rhine debouchment.¹⁰⁶

But the Rhine delta obviously was not the only continental outlet that mattered in the context of a.d. 43. The Atlantic ports of Gaul and Strabo's river outlets were also involved, as the distributions of amphora types like Haltern 70 and Richborough 527 illustrate.¹⁰⁷ The provenance of the fine wares imported to south Britain broadly confirms the continued importance of the Channel and Atlantic ports of Gaul for trade links with the new province.¹⁰⁸ Apart from Barzan on the Gironde, a number of traditional ‘nodal points’ on the north coast of Normandy and Brittany would have been involved.¹⁰⁹ It is now time to look how all those continental outlets were linked with the hinterland and the wider trade and supply networks of the Empire.

THE MESSAGE OF THE MILESTONES

One archaeological category has been curiously neglected in this context: milestones.¹¹⁰ Of course, milliaria, as a proxy of road-building, are not unproblematic. As monumental distance signs along the imperial highways they were highly charged with publicity potential. During the third-century anarchy, milestones duly came to serve as bill-boards expressing the loyalty of local administrative bodies.¹¹¹ An increased role for local authorities in road maintenance is manifest from the turn of the second century, with milestone series now often limited to particular civitates and the first expressions of civic pride occurring under Trajan.¹¹² However, for the first century patterns and practices are different. Early series with identical formulas tend to transcend civitas and even provincial boundaries.¹¹³ In the first century, the dedicatory dative case is still rare, the nominative or ablative cases usually highlighting the Emperor's authorship.¹¹⁴ Moreover, the geographical distribution is very uneven: in many cases, there appears to be a link with the Emperor's family background, his personal experience or impending visits, and this pattern continues into the Severan period (fig. 5). Michael Rathmann, who has studied the material in detail, has concluded that, in the early Principate, ‘the impulses to road-building’, insofar as accompanied by milestones, should be qualified, ‘in large part seem to have issued from the emperor’.¹¹⁵

FIG. 5. The distribution of milestones over the western provinces under the successive Roman Emperors up till Severus Alexander. (Based on the inventory of Rathmann 2003)

To focus on the western provinces, under the first three Emperors, the milestones are confined to Spain, Africa and Narbonensis, with only one early Tiberian exemplar occurring in coastal Dalmatia. This is followed by a spike of early Claudian milestones in Gaul, Upper Germany, Raetia, Noricum and Dalmatia, with the bulk in Gaul and Germany dated to a.d. 43–45 (fig. 6).¹¹⁶ This curious surge has led Rathmann to suggest that the implied road-building campaign issued from Claudius’ stay in Gaul in the months following the British invasion.¹¹⁷ However, even Roman roads did not just appear overnight. One would guess that the imperial highways, as nowadays, would often take several years to construct, as we happen to know of the via nova Traiana.¹¹⁸ One would also assume that milestones were normally put in place upon, or towards, completion of road works.¹¹⁹ Now, the problem is that several of Claudius’ Gaulish milestones were placed in a.d. 43 already, while those in the Rhône valley actually predate the British invasion. In the Rhône delta, at what appears to be the stem of the ‘tree’ of road branches in Gaul, work was already taking place in a.d. 41. The ‘Claudian’ infrastructural project appears to have been initiated well before a.d. 43.

FIG. 6. The distribution of milestones over the north-western provinces under the Julian and Flavian Emperors. (Based on the inventory of Rathmann 2003)

If we plot the stones on a map, an intriguing pattern emerges (table 2, with figure 7). The earliest milliaria, datable in 41, belong to the Via Domitia between Beaucaire and Montpellier (nos 4–14, later extended to Narbonne: 52), with a branch road from Nîmes to Arles (1–3). Interestingly, the building programme also involved the Via Iulia Augusta between Aix and the fleet base Fréjus (32), creating a firm Mediterranean baseline. Next there are two stones, dated to the first half of a.d. 43, from the road on the east bank of the Rhône, leading up to Lyon (25–26). To the same year belongs (the completion of) work on the roads from Lyon to Geneva (27) and up north along the Saône (28) – two main routes to the Rhine indicated by amphora distributions. Another crucial south–north axis was the road from Besançon to Langres (29–30), which was also delivered in a.d. 43, possibly as part of Strabo's route from Italy to the Channel coast, as suggested by the milestone from Lannes on the road from Langres to Naix (31).¹²⁰ Its more challenging Alpine extension may have followed a couple of years later, as the milestones on the Martigny–Lausanne road suggest (50–51). The next stage, a.d. 44, involved the road from Reims to Trier (33), possibly geared to the Moselle, and the Rhine valley between Bingen and Koblenz (34–36). What we have here is a coherent system of roads geared towards the north, starting in 41 in the Rhône delta, then moving up the river and proceeding with the main routes to the crucial Rhine corridor via the Swiss lakes and the Saône valley, with the Besançon–Langres–Naix axis pointing to the Channel coast. With most of the road works delivered between 41 and 44, the combined thrust towards Britain seems unmistakable.

FIG. 7. The successive road-building initiatives in Gaul and Upper Germany as evidenced by the early Claudian milestones, based on table 2. The colours of the triangles correspond with those in the left margin of table 2.

Table 2 claudian milestones from the north-western provinces in chronological order*

* The colours in the left margin correspond with the symbols in figs 7–8. The second column refers to the volumes of the Corpus Inscriptionum Latinarum (volumes in Roman numerals), Année Epigraphique (AE) and the Inscriptiones Latinae selectae (ILS). The final three columns give the number of tribunicial powers (TP), imperial salutations (IMP) and consulates (COS, Des referring to consul designatus), respectively.

The next wave of milestones is equally interesting. These are consistently dated to a.d. 45. Again, a specific pattern emerges. Intriguingly, several remote Channel ports were serviced, including Bayeux (44) and Roscoff (45), as well as the main Atlantic hub on the Gironde, Saintes (43). A road leading to Poitiers from the east (38) may also be geared towards Saintes. The cluster of stones in the region around Clermont-Ferrand might, at first sight, be understood as part of a local road-building programme for the civitas of the Arverni, were it not for the fact that the neighbouring civitas of the Vellavi was also involved (37). Two milestones (38–39) indicate that the old Agrippan axis from Lyon to Saintes was part of the programme. It looks as though the capital of the Arverni was serviced as a regional node, with two milestones (40–41) signalling work on feeder roads from Montluçon and Autun, the capital of the Haedui. No. 37, from Beaulieu, is less easily understood, not least because of the difficult passage of the Cevennes, although in this case, too, a west-bound destination has been proposed.¹²¹ The a.d. 45 series, then, not only transcended civitas boundaries but involved two provinces. What appears to unite this disparate group is the wish to improve overland connections of some of the major civitas capitals of central and western Gaul with key ports on the Atlantic coast. It may be noted that Bayeux was also among the main ports prepared for the great British campaign of Septimius Severus.¹²² The shift of emphasis to the west, corresponding with the first occupation of Britain, is obvious.¹²³ Claudius himself may have initiated this second wave while residing in Lyon for several months in late 43–early 44.¹²⁴

A SHIFT TO THE EAST

With Thrace, Claudius's other new province, the hidden history is much the same. As we saw earlier, the traditional sources will not tell us whether Claudius’ reorganisation of the Balkan provinces was set in motion by the murder of Rhoemetalces III in c. 45 or, inversely, whether the palace revolution in Thrace was simply a concomitant regime change. The date of the coup is somewhat uncertain as items in the spatium historicum of Eusebius’ Chronicle, the sole source of our knowledge,¹²⁵ are known regularly to have migrated one or two years in the course of the complex transmission and re-formatting history of the work.¹²⁶ Is there a way to circumnavigate these uncertainties? The answer may again be in the roads that were to support the new commitments in the Lower Danube region. We have already seen that logistic preparation included works on the Iron Gate towpath, the completion of which is dated to 46.¹²⁷

The milestones, again, are suggestive of a coordinated campaign, this time opening up a series of crucial Alpine and Dalmatian passes linking the Adriatic with the upper Danube (47–48, 62), the river Sava (53–59) and the central road node of the Balkans, Naissus (60: fig. 8). To judge by the milestones, the two Dalmatian roads were only delivered in a.d. 47. However, these appear to be new routes which had to negotiate some difficult terrain, so that a start of work several years earlier seems a reasonable assumption. Support for this is provided by a rather overlooked, almost narrative milestone referring to the building of a road between Trieste and Rijeka across the Istrian peninsula, extending the baseline around the head of the Adriatic, in a striking parallel to the earliest Claudian works in the Rhône delta.¹²⁸ The road was built ‘on the orders’ (iussu) of Claudius by military personnel acting under the authority of the governor of an adjacent province, almost certainly Dalmatia, seeing that the legate in question was none other than Aulus Plautius, who went on to lead the British invasion of a.d. 43, taking one of the Pannonian legions, IX Hispana, with him.¹²⁹ This results in a narrow time window, with the date of delivery probably closer to 43 than 41.

FIG. 8. The successive road-building initiatives in the Alpine and Dalmatian regions as evidenced by the early Claudian milestones, based on table 2. The colours of the triangles correspond with those in the left margin of table 2.

Crucial evidence comes from the famous via Claudia. Its two surviving milliaria (47–48) explicitly name the Danube as the road's destination (ad flumen Danuvium), with the Cesiomaggiore stone identifying Altinum, the port at the head of the Adriatic, as the caput viae – a strong pointer to the road's purpose.¹³⁰ The stones were inscribed in late 46 and January 47, respectively. But, again, these would be the dates for the completion of work. Our good fortune is that we also have dendrochronological dates which ought to be a close indicator of the start of work. These pertain to two log roads that were part of the parallel Reschenpaß–Fernpaß and Brennerpaß routes. Of the former, which crossed the Lermoos bog, most timbers were harvested in autumn/winter of a.d. 45/6¹³¹; for the latter, across the Eschenloher Moos, all trees were felled between May and August 43.¹³² Work on this dual Alpine corridor was apparently going on between c. 43 and 46. The implication is that, like the British project, preparations for Claudius’ reorganisation of the Balkans started about two years before the event.

PREPARING FOR PROVINCES (AND PERSONAL PRESTIGE)

The evidence of the road-building campaigns strongly suggests that the annexation of Britain, and probably that of Thrace too, rather than being triggered by extraneous events, was premeditated and carefully prepared, at least two years in advance.¹³³ Pretexts for occupation could be found, or fabricated, later, like the call for help by Verica or the coup against Rhoemetalces. The situation in Britain may have been deteriorating since Cunobelinus had died a couple of years earlier. However, with the end of his reign foreseeable and a pool of potential successors available, such unrest might have been neutralised with the usual, and extended, methods of Roman ‘diplomacy’ – if that was the preferred outcome. But in this case a Roman agenda appears to have been the hidden driver. A coherent road-building programme geared towards, and including, the Rhine corridor had been stagewise unfolding in Gaul and Upper Germany from Claudius’ very first year in power. The fortified transport corridor in the Rhine delta, one of Britannia's early lifelines, had also been under construction from the very start of Claudius’ reign, as the continuous series of dendrodates at Alphen indicates.

The first road works in the Rhône delta, advertised by the milestones of a.d. 41, must have been commissioned early in the same year at the latest. It is conceivable that this was just a quick refurbishment of existing infrastructure (the milestones’ formula consistently speaks of refecit), commissioned right after Claudius’ accession. But the stronger likelihood, perhaps, is that Claudius inherited, and simply continued, a project in progress, as with Mauretania.¹³⁴ The evidence from the Rhine delta, with its massive supply of Caligulan coinage and the start of work on its fortified corridor now firmly dated to Caligula's last year in power, unequivocally points to Claudius’ predecessor as the instigator of what was to unfold in a.d. 41–43. The felling date of the trees for the rampart base of Alphen (autumn/winter 40/41), in particular, seem to confirm Barrett's carefully argued scenario that Caligula, whatever the initial plans for his great northern expedition, had to conclude, early in a.d. 40, that a successful invasion of Britain needed better logistic preparation.¹³⁵ This would almost certainly have included an upgrade of the port infrastructure at Boulogne, which makes sense of Caligula's order to build a lighthouse, probably there, in the wake of the a.d. 40 campaign.¹³⁶

Suetonius says that Caligula launched his campaign ‘on a sudden impulse’.¹³⁷ This cannot be true. Whatever the rationale of the German expedition and its relation with Britain, all the evidence points to an enormous operation. The campaign, even if not nearly fielding Dio's 200,000 or even 250,000 men,¹³⁸ would have required massive logistic support and careful preparation, as Suetonius well knew.¹³⁹ Caligula almost certainly was responsible for the raising of two new legions,¹⁴⁰ and these, again, would have to be fed, clothed, equipped, trained, paid and housed. Three short-lived and roughly contemporary installations of legionary size now are in the picture at Mainz-Weisenau, Ludwigshafen-Rheingönheim and Groß-Gerau-Wallerstädten.¹⁴¹ The scale of preparatory troop movements and encampments, with XV and XXII Primigeniae apparently raised as replacement for the legions destined for Britain, is obvious. Barrett makes a strong case that Caligula's objective with his ‘German campaign’ always was to prepare and secure the launch-pad for a full-blown invasion of Britain, thereby emulating his father and achieving what Caesar had abandoned.¹⁴² This great undertaking must have been planned long before, probably soon after Caligula's accession, two years before the event.¹⁴³

CONCLUSION

The detailed chronologies elaborated in this paper have a transcending meaning for the debate about the scope of Roman imperial policies. The Emperor obviously had far-reaching interests, if only in the sphere of personal prestige and public image, aspects studied in detail for Claudius by Osgood (fig. 9).¹⁴⁴ He also certainly had a long and strong arm – the military apparatus commanded by his legates. It was the Emperor's key interest to keep this apparatus well paid, fed and otherwise supplied.¹⁴⁵ This is why frontiers and ‘foreign policy’ matter. New commitments often meant military investment which, in turn, required logistic preparation. It is here that we can best see imperial policy and foresight in action. The central concern always was to prepare and secure the stream of supplies. This could entail serious and structural investments, witness Trajan's works along the Iron Gate. A recent study argues that his investments reflect wartime conditions and that the Iron Gate would normally remain a bottleneck.¹⁴⁶ However, the volume of military supply in peacetime was not essentially different. After the dust had settled on the Balkans around a.d. 45/6, Novae continued to receive the latest fashions in Italic and Gaulish Samian in the following decades,¹⁴⁷ meaning that the Danubian supply line remained important. In the far west, likewise, connections with Britain continued to be upgraded after a.d. 43, witness the roads to the Atlantic ports of Gaul.

FIG. 9. Claudius overcoming Britannia. Relief from the Sebasteion at Aphrodisias in the Aphrodisias Museum. (Photo by Dick Osseman, licensed under the Creative Commons Attribution-Share Alike 4.0 International licence)

The nature and scale of logistic investment were such that, more often than not, at least two years of preparation were necessary. The roads to Saintes, Bayeux and Roscoff delivered in 45 would have been commissioned in 43, after the British invasion had proved successful, perhaps with Claudius himself overseeing things from Lyon. The preceding work on the Rhône–Rhine corridor had started early in a.d. 41 at the latest. The reorganisation of the Balkans appears to have been prepared in 43–45/6. The planning of Caligula's northern expedition would have started soon after his accession, two years before the event. We have seen that Trajan took at least two work seasons to pave the way to Dacia, while the frontier works that Hadrian was to inspect in 121/2 had been ordered in 119.¹⁴⁸ The occupation of south Scotland by Antoninus Pius, a relative walk-over effectuated by August 142,¹⁴⁹ was preceded by work on the horrea at Corbridge, the logistic turntable on Dere Street, in 139 and 140.¹⁵⁰ Severus’ Scottish campaign of 208–211, likewise, appears to have been carefully prepared, with the roads to Boulogne and Bayeux delivered in time, in 207/8.¹⁵¹ For the start of preparations for the expeditio felicissima Britannica we would, again, have to look for relevant dendro dates. The Rhine delta, gateway to Britain,¹⁵² now has on offer the harbour basin of Forum Hadriani on the Corbulo canal as well as one of the Zwammerdam barges (generally a good proxy for peaks of military activity), for both of which trees were felled in a.d. 205.¹⁵³ Finally, the German campaigns of Augustus come to mind, with the spring-board and supply streams prepared, possibly over several years, while the Emperor was managing affairs from Lyon (16–13 b.c.).¹⁵⁴

This 2/3-year rule of thumb brings us to a somewhat uncomfortable observation. If we limit ourselves to the first and second centuries, we find that those rulers who lacked a military track record of their own typically launched a campaign on the northern frontiers in the second or third year after their succession: Caligula, Claudius, Domitian, Trajan, Antoninus Pius, Commodus¹⁵⁵ and Caracalla (table 3). In most cases, we see logistic build-up taking place in the preceding two or three years. For Domitian's Chattan war (83) and Caracalla's expeditio Germanica (213) we have, again, evidence for roads completed in time in the respective hinterlands.¹⁵⁶ With the exception of Antoninus Pius, the Emperor normally participated in person, albeit briefly in most cases. While the lack of military prestige has been recognised for most candidates, the inclusion of Trajan may surprise some. However, Michael Speidel has rightly underlined that the optimus princeps, on his accession, was not yet the great military man and would have felt a dire need to bolster up his battlefield credibility.¹⁵⁷ In Trajan's case, logistic preparations appear to have been set in motion shortly after his sudden promotion to the purple, while our 2/3-year rule of thumb suggests much the same for the others. It all serves to identify the northern ‘barbarian’ frontiers as reserved hunting-ground for newly enthroned Emperors with thin martial credentials – and remind us of the potentially predatory nature of the Roman state.

Table 3 the incidence of imperial campaigns on the northern frontiers in relation to the (effective) start of their reign

The protagonist of this paper, Claudius, has challenged us to probe beyond the anecdotal nature of imperial biography and the pretexts of Roman foreign policy. Although the modern usage and apparatus of ‘policy’ and ‘strategy’ almost render the terms useless in the ancient world, a restricted use advocated by Kagan and Osgood seems viable and helpful.¹⁵⁸ Most Emperors would have had clear objectives, even if some might have cared mainly about a reliable revenue and retinue. Paradoxically, it was precisely the limited means at their disposal and slowness of communication that forced them to set clear priorities and plan things well in advance. In this paper we have followed the trail of evidence left by two major foreign policy projects that marked the beginning of Claudius’ reign. Far from a passive player who could do little more than respond to appeals, embassies and extraneous events, like the flight of Verica or the elimination of Rhoemetalces, we see an Emperor with clear goals who planned his projects years in advance – a reality almost irreparably distorted by the passive routine of the documentary sources and, above all, hostile senatorial historiography.