All images: Matthew Symonds, unless otherwise stated

Hadrian’s Wall: creating division

Could the relationship between the Roman army and local communities living in the shadow of Hadrian’s Wall hold the key to understanding this remarkable monument? Matthew Symonds investigates why it was built.

Start

Hadrian’s Wall is rarely seen as much of a mystery. Even its modern name reveals who ordered this frontier system into being, while a wall is something we can all relate to. As for purpose, the sole Roman statement on the matter has a simple clarity. The Historia Augusta, which was probably compiled more than 200 years after Hadrian’s death in AD 138, records that he ‘was the first to build a wall, 80 miles in length, to separate the barbarians from the Romans’. With the who, what, and why taken care of, there may seem little scope for any real mystery. But one key consideration remains thoroughly enigmatic: what did Hadrian’s Wall do?

All images: Matthew Symonds, unless otherwise stated
Milecastle 39 on Hadrian’s Wall. The ruins of this Roman frontier system remain impressive, but what was the army attempting to achieve by building it? All images: Matthew Symonds, unless otherwise stated.

As the Wall’s course cut through stable farming communities that were centuries in the making, its likely consequences must also have excited keen interest back when construction started. Even at a conservative estimate, the precise nature of the division that the Roman authorities desired would have determined the future of thousands of people. Much depended on how easy it was to cross the Wall. Frontier gateways within small military posts set at sufficiently regular intervals to coin the name ‘milecastles’ are central to this debate. Although their presence raises the possibility that passageways were provided to minimise inconvenience for local groups, an increasing number of scholars now see these gateways as primarily intended to aid military manoeuvring. If so, major disruption for local people seems certain, but were such groups merely passive recipients of Roman policy? My new book takes a fresh look at the frontier (see ‘Further reading’ box on p.42), arguing that responses by some local communities played an important role in determining the nature of Hadrian’s Wall.

Drawing a line

Clues about what the Wall did can be found in both its design and its interaction with the wider landscape. The Wall’s location on the Tyne–Solway isthmus was presumably largely a consequence of Britain narrowing abruptly there, making a shorter barrier viable. But this region also presents the finest natural east–west corridor north of Stainmore, about 50km away, and south of Tweeddale, roughly 60km distant. The Tyne–Solway passage is created by the valleys of two river systems, which are linked by a 2.55km- wide bottleneck of land known as the Tipalt–Irthing gap. This key junction not only eases east–west movement, but also provides the only place on the isthmus where north–south passage is possible without making a river crossing. As the course of Hadrian’s Wall seals off the lateral east–west valley corridor from the north, it would fit with the Roman authorities paying close attention to the local landscape. The design of the Wall, though, has long been taken to imply that the nuances of the terrain were largely irrelevant to its smooth running.

So far as we can tell, the original Wall concept featured a cordon of small posts strung out along the curtain wall. Milecastles were placed approximately every 1,479m (or one Roman mile), with two turrets in between, resulting in a manned military installation roughly every 495m. The connecting curtain was originally built of stone to the east of the River Irthing and turf to its west. A ditch was usually cut to the north of the curtain, while timber obstacles likened to Roman barbed wire could also be present. The result was a formidable barrier, with a relatively modest garrison. Milecastles are unlikely to have held more than roughly 32 men, with perhaps as few as two soldiers on duty in each turret at any given time.

left Milecastle 39 on Hadrian’s Wall. The ruins of this Roman frontier system remain impressive, but what was the army attempting to achieve by building it? above Hadrian’s Wall, as completed at the end of Hadrian’s reign.
Hadrian’s Wall, as completed at the end of Hadrian’s reign. Map: courtesy of David J Breeze.

One of the more remarkable features of this uniform frontier concept is the way it departed from the established Roman military practice of placing posts so that they drew maximum advantage from their surroundings. Instead, imposing a regular cordon on irregular terrain threw up absurdities. Although some spacing flexibility was permitted, milecastles could still be built on steep slopes, while gateways sometimes opened on to awkward drops, even when passes on level ground lay close at hand. David Woolliscroft has suggested that signalling was one reason why posts were not always shifted to more suitable plots. Between Carlisle and Corbridge, forts already existed to the south of the Wall along the line of a Roman road known as the Stanegate. As these garrisons offered the only ready source of reinforcements for the milecastles and turrets, it is easy to see why a signalling link might have been desirable. If so, though, a change of plan while construction was under way emphasises this arrangement did not prove satisfactory.

Specialists refer to this overhaul of the original concept as the ‘fort decision’, because the most eye-catching alteration was the addition of forts to the Wall line. Once again, regular spacing seems to be in play, with the forts usually described as lying at intervals of between 7¹⁄³-7²⁄³ Roman miles. A notable quirk is that many forts were built astride the curtain, so that three of their four main gateways projected to the north. To the south – that is, within Roman territory – an extraordinary earthwork known as the Vallum was created. This usually took the form of a massive ditch flanked by earth banks, creating an obstacle about 36m wide. Another change at around this time was a reduction in the width of the Stone Wall curtain, from a Broad Wall of roughly 2.9m to a Narrow Wall of 2.3m. This is helpful for providing a sense of where early work was carried out, while the uniform ‘blueprint’ adopted for Hadrian’s Wall can also offer a glimpse of military priorities. Because the basic concept is predictable, we can see when the army – for want of a better phrase – broke its own rules. Examining where special measures were taken reveals a clear pattern.

above Looking west over the Tipalt Irthing gap, from near the position of milecastle 46. The gap connects the valleys of the main east and west river systems, creating a natural corridor for cross-country movement.
Looking west over the Tipalt Irthing gap, from near the position of milecastle 46. The gap connects the valleys of the main east and west river systems, creating a natural corridor for cross-country movement.

Conquering a landscape

image: after Breeze and Dobson (2000)
A schematic diagram of Hadrian’s Wall: (A) shows the Wall as it was seemingly originally conceived, with a cordon of milecastles and turrets along the curtain; (B) illustrates how it was reworked during construction, to add a series of forts and the Vallum. Image: after Breeze and Dobson (2000).

Thanks to the Broad Wall, we know that early construction work on the curtain was focused in the east, so it would also be natural for Broad Wall milecastles to cluster in that region. Instead, milecastles that were certainly or probably entirely Broad Wall in plan were widely distributed along the Stone Wall. Significantly, all of the known examples can be associated with potential north–south routeways. Among them, milecastles 47 and 48 controlled the Tipalt–Irthing gap, milecastle 27 blocked the natural passage created by the North Tyne valley, milecastle 23 lay near Dere Street – a major Roman road that in places followed a prehistoric predecessor – and, as Al McCluskey has pointed out, milecastle 10 was built north of key Tyne fords exploited by invading Scottish armies in 1346 and 1640.

PHOTO: Stephen Greep
The linear barrier elements of the original Wall concept created an impressive obstacle. Although most of the Wall curtain was built of stone, a considerable length running west of the River Irthing was made of turf, earth, and timber. PHOTO: Stephen Greep.
The linear barrier elements of the original Wall concept created an impressive obstacle. Although most of the Wall curtain was built of stone, a considerable length running west of the River Irthing was made of turf, earth, and timber (above). A ditch was normally cut to the north of the curtain (right).
A ditch was normally cut to the north of the curtain. PHOTO: Stephen Greep.

Assessing the Wall forts corroborates this impression that important natural and artificial landscape features received special treatment. The distance between Chesters Fort and its neighbour, for instance, was reduced to six Roman miles, allowing it to dominate the North Tyne valley. An additional fort was later founded only 3²⁄³ Roman miles away at Carrawburgh, creating a concentration of force in the vicinity of the North Tyne valley. The same is true of the Tipalt–Irthing gap, where the forts on its flanks at Birdoswald and Carvoran lay only 3¹⁄³ Roman miles distant, while a third fort was established just three Roman miles away. Indeed, when various unusual Wall elements are superimposed on a single map, they reveal a concentration of anomalies in and around the Tipalt–Irthing gap. That this natural junction in the landscape occasioned so many departures from what might be considered normal suggests that an ability to clamp down on movement is central to what the army was attempting to achieve with Hadrian’s Wall.

left The regular spacing system resulted in several posts occupying imperfect plots. Here, milecastle 48 lies on a one-in-five slope.
The regular spacing system resulted in several posts occupying imperfect plots. Here, milecastle 48 lies on a one-in-five slope.

There can be little doubt that military measures to tighten control of key routeways, and the Wall concept more generally, risked aggravating any tensions between the army and pre-existing local communities. Archaeological work to the north of the Wall on the Northumberland coastal plain, for instance, has revealed settlements that surprised by virtue of their size, quantity, and longevity (see CA 277). This farming community had an estimated pre-Roman population of 10,000-15,000 people. There are hints that elite status was at least partially conferred by ancestral claims to land, while some landscape features were probably of ritual significance. Sufficient prehistoric metalwork has been found in the Tyne to suggest that it was a focus for some kind of sacred rites, while there are other indications of interest in watery places, such as rivers, springs, and bogs. Many people would have depended on movement for a range of activities, be they farmers attending markets, cattle drovers, pilgrims, envoys, seasonal labourers, or those journeying for marriage. Seen this way, Hadrian’s Wall would sever a complex political, working, and spiritual landscape.

UnRoman warfare

It is easy to assume there was little that local groups could do to resist this. Accounts of warfare in Roman Britain often focus on the great set-piece battles of the Boudican revolt or Mons Graupius, when Roman armies could slaughter their enemies in vast numbers. At Mons Graupius, for instance, the Roman historian Tacitus claims 10,000 enemy dead, for the loss of just 360 soldiers. But ancient authors also report that another form of fighting was common in Roman Britain, especially Wales and Scotland. It typically involved ambushing Roman soldiers while they were vulnerable or otherwise disadvantaged, an approach described by Tacitus as ‘in the manner of banditry’. Today this style of combat is often popularly referred to as guerrilla warfare or an insurgency. Such fighting has long been a hazard for colonial powers, and an influential treatise on the subject was written by Colonel Callwell back in the 19th century. He noted that such campaigns ‘are always most trying to the [regular] troops’.

left The North Tyne river, as seen from the east gate of Chesters fort. This commanding location did not come about by chance, as the fort lay closer than usual to its eastern neighbour.
The North Tyne river, as seen from the east gate of Chesters fort. This commanding location did not come about by chance, as the fort lay closer than usual to its eastern neighbour.

Guerrilla fighters typically attempt to overcome the advantages that regular troops enjoy in numbers, equipment, and training by launching surprise attacks and – crucially – escaping before the enemy’s superior force can be brought to bear. There are numerous accounts of such attacks in Roman Britain, including one on legionary cohorts undertaking construction duties in Wales, but when their comrades responded ‘the enemy escaped with trivial losses’. Regular forces often respond by developing tactics to diminish the advantage of surprise by accelerating their response time. There can be a tendency to see such struggles as trivial affairs of little significance, but plenty of 20th-century insurgencies illustrate how misleading that can be. The Northern Ireland Troubles, for instance, became the longest conflict in the history of the British Army, and featured its largest infantry deployment since the Second World War. Victory in a counterinsurgency also requires a different approach to winning conventional wars. Because guerrillas rely on communities that support or harbour them, local populations are key to understanding and resolving such conflict. Given guerrilla warfare was seemingly widespread in Roman Britain, could this help shed light on Hadrian’s Wall?

below Plotting unusual elements of the Wall on a single map produces a cluster of anomalies in and around the Tipalt Irthing gap. These unusual features (some of which have been moved a little to the north or south to aid legibility) include Broad Wall milecastles, large milecastles, Wall miles with additional towers, unusually close forts, unusually large forts, unusually small forts, milecastles where causeways over the Wall ditch were retained, and milecastles and turrets that appear to have major timber components in the late period.
Plotting unusual elements of the Wall on a single map produces a cluster of anomalies in and around the Tipalt Irthing gap. These unusual features (some of which have been moved a little to the north or south to aid legibility) include Broad Wall milecastles, large milecastles, Wall miles with additional towers, unusually close forts, unusually large forts, unusually small forts, milecastles where causeways over the Wall ditch were retained, and milecastles and turrets that appear to have major timber components in the late period. Image: courtesy of David J Breeze, with additions.

Ingredients for unrest were certainly present in the decades leading up to the Wall’s construction. Roman forces had abandoned southern Scotland under Hadrian’s predecessor, Trajan, violating Callwell’s ‘one great fundamental rule’ when fighting guerrillas: ‘to seize and to keep the upper hand, to advance constantly’. This withdrawal would certainly have encouraged any belief that Rome could be compelled to retreat further. We also know that there was trouble in Britain at the beginning of Hadrian’s reign, when the Historia Augusta records that ‘the Britons could not be kept under Roman sway’. A military taskforce was dispatched to the island, most likely in 122, at around the time work on the Wall started. Its origins are also placed in the aftermath of fighting by the reconstructed text of an inscription found at Jarrow, which presumably refers to Hadrian: ‘after the barbarians had been dispersed and the province of Britain recovered, he added a frontier line’. Describing Rome’s foes as ‘dispersed’ is interesting, as it suggests enemies that were hard to bring to battle and defeat decisively, which would fit with guerrilla warfare.

above The Gask Ridge system in Scotland, where forts and fortlets were combined with a tower cordon along a key stretch of highway.
The Gask Ridge system in Scotland, where forts and fortlets were combined with a tower cordon along a key stretch of highway. Image: courtesy of David J Breeze.

Waging counterinsurgency?

below The Vallum is an extraordinary earthwork that generally consists of a 6m-wide ditch flanked by earth mounds to the north and south. As the decision to add it seems to have been taken after construction was under way, does it reflect a shift in the security situation?
The Vallum is an extraordinary earthwork that generally consists of a 6m-wide ditch flanked by earth mounds to the north and south. As the decision to add it seems to have been taken after construction was under way, does it reflect a shift in the security situation?

It is potentially illuminating to compare the Wall to military control strategies in Wales and Scotland, where the ancient histories indicate protracted insurgencies. A network of forts and fortlets was established in Flavian Wales, creating a web of posts reminiscent of measures that Callwell advocates as the best bet for victory in guerrilla warfare. An enhanced version of this approach can arguably be found along the Gask Ridge in Scotland, where a chain of towers was added to the model. Broad similarities between the Gask Ridge and subsequent frontier systems have long been noted, making it possible that the design of Hadrian’s Wall drew on and adapted preceding counterinsurgency techniques. If so, the relatively small milecastle and turret garrisons suggest the Wall was initially expected to face relatively low-level resistance. A sense of scale may come from documents found in Egypt, which indicate that Roman fortlet garrisons probably roughly comparable to those in the milecastles were tackling hostile bands of 18-61 ‘barbarians’.

Perhaps early work on posts near key routeways signalled Roman intentions to seal off the region, exacerbating tensions with local groups, and hit-and-run attacks were mounted from the north.

The sudden need for Wall forts, and the approximately 9,090 soldiers they held, is most easily explained by a rapid deterioration in the security situation. One possibility is that early work on posts near key routeways signalled Roman intentions to seal off the region, exacerbating tensions with local groups. Hit-and-run attacks being mounted from the north would explain why forts like Chesters were placed astride the Wall. Maximising the number of gateways opening north would fit the counterinsurgency cliché of a regular army seeking to reduce its response time and engage elusive enemies before they vanish. That Roman cavalry squads were engaging mobile groups is indicated by an altar that probably came from Chesters. It was dedicated by a prefect of cavalry, and commemorates the ‘slaughtering of a band of Corionototae’. A similar example is known from Carlisle. The Vallum would hinder ambushes from the south: even if attackers snuck across it under cover of darkness, the clean escape essential for guerrilla warfare was compromised.

below Turret 41a on Hadrian’s Wall. Could this frontier system have been designed as a component of a counterinsurgency strategy?
Turret 41a on Hadrian’s Wall. Could this frontier system have been designed as a component of a counterinsurgency strategy?

Both the initial and revised plans for the Wall, then, appear explicable as a response to security concerns that included guerrilla warfare. The milecastle and turret cordon along the curtain would help isolate southern resistance from northern support or safe havens by clamping down on north–south movement, while the Wall forts aided the interception of nimble assailants, and the Vallum guarded against ambushes. That the new measures had an impact on local groups is indicated by at least some settlements on the Northumberland coastal plain being abandoned around the time the Wall became operational. Modern counterinsurgency methods, though, advocate seeking a political solution that addresses the original grievance. On that score, it may be telling that there are signs southern resistance slackened in the later 2nd or early 3rd century. This is roughly the time a town was founded at Carlisle and perhaps also Corbridge, restoring a measure of autonomy to local groups.

Further reading
Matthew Symonds (2021) Hadrian’s Wall: creating division (Bloomsbury Academic, ISBN 978-1350105348), £19.99 (paperback).