To some, it might seem obvious that prehistory is no respecter of modern national borders. For example, even the most cursory examination of hillforts across the northern part of Britain readily reveals that the dense concentration of such sites apparent in south-eastern Scotland easily spills into what today is north-eastern England. The same pattern can be seen in other forms of Iron Age settlement, such as rectilinear enclosures and roundhouse architecture – but this is only part of the picture. Some kinds of constructions, namely brochs, duns, crannogs, and souterrains, are found widely across Scotland, but this architectural array is not replicated in northern England.
Why were Iron Age peoples in Scotland building and occupying these structures when their contemporaries to the south were not? There is no environmental reason behind this difference. Drystone architecture was common enough amongst Iron Age communities in northern England, but brochs and duns were evidently not to their taste. Nor is there any inherent environmental condition that largely prevented souterrains being constructed south of Scotland (until we reach Cornwall). It is also surely not credible that not one lake in England was conducive to crannog construction – even the Lake District, actually visible from much of crannog-rich Galloway, contains no known examples. There is a single crannog in Wales, but this is early medieval in date and is the result of Irish influence. Indeed, Ireland possesses numerous crannogs but the vast majority of these are much later than most Scottish examples, demonstrating that, while equivalent, they are not part of the same contemporary settlement pattern. Souterrains are also found in Ireland, but again these are predominantly much later in date (see CA 263), suggesting that they too are an independent cultural response to analogous circumstances. Perhaps, then, this differentiation can be explained through cultural reasons – the collective choices one community made that were distinct from those of other groups and which were not dictated by environmental factors.
Surprisingly, the fact that various types of Iron Age settlement do not breach the modern Anglo-Scottish border is something that has not been previously examined in detail. The reasons for this may lie in the way Iron Age archaeology is studied in Scotland, much like in the rest of Britain – that is, largely by region. This approach is understandable given the sheer volume of evidence, and there undoubtedly is considerable regional variation apparent in the settlement and material record of later prehistoric Scotland – we are not talking about one homogenous group. Take, for example, the dominance of different settlement types, with brochs and duns common in the north-west, and hillforts and souterrains more common in the east. Material culture, meanwhile, is particularly diverse, with a pottery-rich Atlantic zone and a pot-poor remainder, while differential patterns of quern types and hoarding can be seen between northern and southern Scotland. We can also see regional idiosyncrasies in the personal ornamentation worn by people in different areas, such as the ring-headed and spiral finger rings that were common to the Atlantic zone but largely absent in Galloway. What can we understand from these trends? My research, recently published in the Proceedings of the Society of Antiquaries of Scotland (see ‘Further Reading’ on p. 45), set out to make sense of these shifting patterns and perceptions.
Iron Age culture in Scotland is, to some degree, defined by the relative dominance in Atlantic Scotland of brochs, duns, and crannogs, in contrast with the prevalence of hillforts and large enclosed settlements in south-east Scotland; between a society composed of single-household settlements in the west and multiple-household settlements in the east; a pattern solely comprising individual farmsteads to that comprising villages, hamlets, and farmsteads.
The situation is not quite that simple, however. While enclosed settlements in lowland Scotland might often be bigger than their highland and island equivalents, many are also single-household settlements, but ones that were enclosed with large surrounding farmyards. Nor is it the case that Atlantic Scotland had no large hillforts: it is just that so few of them have been excavated that it creates an impression of absence. Moreover, although some western regions like Galloway have no large hilltop enclosures of the kind evident elsewhere in southern Scotland – such as Burnswark in Dumfriesshire (CA 316), Traprain Law in East Lothian (CA 203 and 283), and Eildon Hill North in the Scottish Borders – more significant than their specific size is the fact that these enclosed large regional landmarks. In Galloway, the most dominant natural feature – recognised at least as early as Ptolemy’s Geographia of the early 2nd century AD – is not a hill, but a mull on the southernmost tip of Scotland. The ramparts cutting across the Mull of Galloway enclose 40ha, which would make this the largest of all later prehistoric enclosed sites in Scotland. The same pattern is evident across the regions of Scotland: the bulk of Iron Age settlements are below 0.8ha in size, with only a handful surpassing 1ha and usually just one regional site being significantly larger than all the rest.
It is not the comparative size of settlements that is the most crucial aspect to consider, though. A number of distinctive attributes – exotica (eg Roman objects), ornamental metalwork, ironworking, shale-working and non-ferrous metalworking – have been identified amongst Iron Age settlements in Scotland as being related to social networks and specialised skills. Most settlements possess one or two of these, but only a very few large settlements preserve evidence of three or more. Across southern Scotland, excavations of multiple-household settlements like Traprain Law and Broxmouth in East Lothian (CA 307) and Dunragit in Galloway (CA 378) have revealed a wider range of these attributes than is found in contemporary Iron Age single-household settlements. This is evident in the large multiple-household settlement at Culduthel in the Highlands (CA 383) and further west too, where Balloch Hill in Argyll and Dun Deardail in Lochaber – the only two hillforts on the Atlantic seaboard to have been extensively excavated in modern times – have both yielded evidence of exotica, ironworking, and non-ferrous metalworking.
The earlier excavations of Dunagoil on Bute yielded perhaps the richest assemblage per cubic metre examined of any hillfort in Scotland. This suggests not only that the economic and organisational basis of multiple-household settlements differed from single-household settlements, but that both Highland and Lowland Scotland had both types of settlements. It is true that larger settlements are more numerous in the south and east, but this reflects a greater population density than in the north and west (where single-household settlements predominate), rather than a completely different nature and structure to Iron Age society. In short, archaeological evidence suggests that Iron Age Scotland was not so much a patchwork but a spectrum.
Turning to those types of settlement peculiar to Iron Age Scotland – crannogs, souterrains, duns, and brochs – although their distribution across the country is patchy, to ascribe this purely to regional or topographical diversity does not do justice to the evidence. Despite a preponderance of crannogs in the west, for instance, recent research shows that where suitable loch conditions exist, crannog sites are found right across Scotland.
Souterrains, meanwhile, are more common to the north of the Forth, but are also present in southern Scotland and on the Atlantic seaboard too. It is also notable that the souterrain at Cults Loch in Galloway originated in the last two centuries BC, demonstrating that southern Scottish souterrains do not derive from a late spread emanating from the north-east, but belong within the same chronological context and are part of a far wider distribution across Scotland than has been previously understood. They form part of a cultural and economic pattern of accruing a food surplus stemming from the intensification of farming, a process that pollen analyses indicate began in the last few centuries BC.
Though the origins of brochs grew out of the development of complex Atlantic roundhouses in northern Scotland during the later centuries BC, the brochs of southern Scotland, of the early centuries AD, are not a homogeneous group distinct from their northern neighbours. A similar and contemporary pattern of materially wealthy broch households is apparent in central and southern Scotland – for example, at Torwoodlee, Leckie, and Buchlyvie – as in northern Scotland, at sites such as Scalloway and Dun Vulan. Though barely scratched in comparison, Galloway brochs such as Teroy and Crammag Head have also yielded some of the same social network and specialised skill attributes, such as access to Roman exotica and ironworking, that we discussed above.
There is also as much variety among the architecture of southern brochs as there is among northern ones; idiosyncratic features within some of the Galloway examples, for instance, such as double entrances and staircases and diminutive internal floor areas, are architecturally analogous with brochs in Atlantic Scotland. Meanwhile, some lowland brochs, such as Edin’s Hall and Torwoodlee, may have formed the nucleus of multiple-household settlements, comparable with northern broch villages such as Gurness in Orkney or Old Scatness in Shetland, but others are discrete single-household settlements, like the majority of Atlantic brochs. Even then, brochs like Teroy and Bow Castle that occupy elevated spurs seem hardly the same as Doon Castle and Stairhaven that cling to the Galloway coast on precariously overlooked locations. Wherever they were built, however, it is likely that many of these constructions were intended as physical statements by locally prominent households.
The brochs of lowland Scotland reflect cultural choices that are consistent with settlement patterns elsewhere in the country, but these same choices were not made elsewhere in contemporary Britain or Ireland. The same is apparent for crannogs – dating from around the middle of the 1st millennium BC across both northern and southern Scotland but predominantly occurring much later in Ireland – a similar pattern to that observed between the souterrains of both countries. An equivalence can also be drawn between the scattered distribution of small stone-walled settlements across south-west Scotland, the homesteads of Perthshire, the ringforts of north-east Scotland, and the duns of north-west Scotland – again, not recorded south of the Tweed or Solway. That it is often difficult to draw a line between brochs and duns, between forts and duns, and even some crannogs and duns too, accentuates the clarity of divergence between opposite sides of the Scottish/English border.
A state of anarchy
Given these divergences, how different was the nature of Iron Age society in Scotland to that elsewhere in Britain? Archaeological evidence from the lowland brochs at Torwoodlee, Leckie, Buchlyvie, and Edin’s Hall suggests the presence of wealthy households during the early centuries AD, who chose to define their homes with monumental architecture. It is doubtful, though, that every broch, whether in the south or north, was of equally high status. The fact that occupation at many of these sites seems to have come to an end in catastrophic episodes of destruction, such as fires, combined with excellent preservation conditions, might bias our view as to their social prominence, but comparable levels of material wealth may have been enjoyed by other nearby but less well-preserved duns and crannogs, as well as the communities inhabiting larger settlements such as Traprain Law, Broxmouth, Dunragit, Burnswark, and Dunagoil.
What is difficult to demonstrate at any of the excavated multiple-household settlements, though, is the hierarchical nature of these settlements, either between different households within, or in contrast to households without. There is no evidence that the inhabitants themselves were of a higher status than those of other contemporary households. Ornamental personal jewellery is apparent in a range of contemporary Iron Age settlements but production of this does not appear to have been controlled by small elites. Given that enclosing ramparts and ditches are often interpreted as bolstering the status of Iron Age settlements, the absence of such features at Dunragit in Galloway or Culduthel in the Highlands hardly suggests an elite status for these complex settlements. Within the Rhins of Galloway, where Dunragit lies, excavations of a palisaded settlement, crannog, and promontory fort in and around Cults Loch revealed a dynamic and sequential settlement pattern between the mid-6th and 1st centuries BC rather than a hierarchical one.
This lack of clear hierarchy amongst a multitude of settlements is apparent elsewhere in Scotland. The prevalence of brochs across Atlantic Scotland suggests that these were less prehistoric castles than stout farmsteads in the main, though the development of broch villages such as Gurness suggests that some households climbed to pre-eminence. Across the Forth Valley, Perthshire, and Angus, a range of unenclosed and enclosed settlements of varying sizes occurs, as well as clusters of brochs and duns. Within what might be perceived as a competitive non-hierarchical society, then, access to Roman imports during the early centuries AD could have provided temporary impetus to some of the substantial single households. The full scale of architectural aggrandisement and variety originated prior to this, however. What are we to make of evidence for differential access to Roman goods? They could represent the result of Roman patronage, local rejection, or archaeological preservation. And even for a household with access to Roman imports, this process was not plain sailing – or long-lasting, given the fiery demise of so many of the settlements in the Forth Valley, some perhaps the result of Roman aggression, as the heat-cracked catapult stone and ballista bolt found at Leckie suggest. There is little evidence, too, of hierarchy amongst the variety of enclosed and unenclosed farmsteads and settlements in East Lothian until the 1st century AD, when, instead of substantial stone roundhouses, Traprain Law returned to prominence.
Drawing a line
This lack of hierarchy may have played a crucial part in explaining why the Romans failed to absorb Scotland into their empire, despite three major military campaigns that appear, at least in Roman accounts, to have been overwhelmingly successful. This failure is often attributed to the changing political and military priorities of Rome, but it may owe more to the nature of Iron Age society in Scotland, which archaeologists are beginning to recognise was anarchic in nature – not chaotic, but composed of autonomous households and communities that lacked institutionalised leadership. Unlike the tribal kingdoms the Romans encountered to the south, which were more easily absorbed into the Empire’s way of doing things by creating client kings from local leaders, the Iron Age peoples of Scotland were seemingly immune to Roman charms.
Although the territories seized by the construction of the Antonine Wall were officially part of the Roman Empire for almost a generation during the 2nd century AD, no towns or villas were founded in this region. Indeed, only a handful of extramural settlements appear in southern Scotland at all, and their inhabitants seem to have packed up and headed south as soon as their adjoining garrisons were redeployed to what became the permanent northern frontier at Hadrian’s Wall.
The distinctive differences in the archaeological record that we have discussed are especially significant because they demonstrate that this divergence occurred long before the Roman frontier zone of Hadrian’s Wall severed the island of Britain in two. Crossing where the British mainland narrows, Hadrian’s Wall followed a line on the best strategic course – and this line ran through a broader zone of cultural divergence that already existed.
The nuclear option
Clear evidence for the adoption of Roman culture is not apparent in Scotland until the 5th century AD, after the official end of the imperial project in Britain. This takes the form of ecclesiastical Latin-inscribed stones, bearing the Latinised names of indigenous inhabitants, accompanied by Christian terminology and symbols, which were erected across southern Scotland at this time. It appears that communities in Scotland only truly adopted aspects of Roman culture when the region’s Iron Age society had become distinctly more hierarchical. This more structured social hierarchy, indicated most clearly by evidence of direct international trade, gold and silver metalwork, the production of ornamental metalwork, weapons, complex fortifications, and royal inauguration features at a limited number of small settlements, is not apparent until around the 5th and 6th centuries AD. Here, further comparison with brochs is relevant.
Brochs in lowland Scotland are distributed in large but distinct clusters: within a 700km² area around the Forth Valley, a 600km² area around the Firth of Tay, and a 400km² area in the Rhins of Galloway. Those at Torwoodlee and Bow Castle in the Scottish Borders, meanwhile, are separated by 40km from Edin’s Hall and may constitute the remnants of another group. These concentrations of prominent households within 400-700km² areas during the first two centuries AD may represent an Iron Age precursor to the 400-700km² clusters of pre-eminent households that emerged in the 6th and 7th centuries AD: settlements that often come in the form of nucleated forts, a type of early medieval hillfort which is also absent south of the border. Nucleated forts often occur in discrete clusters of elite settlements – in Galloway, Argyll, the Scottish Borders, Fife, Tayside, and Aberdeenshire – and excavations have revealed several of these to be royal strongholds.
In these clusters of early medieval elite settlements, is there an echo of much earlier social organisation and households accruing power and status that had been arrested in development during the early centuries AD? This is not to suggest that any broch is directly related to or equivalent to a nucleated fort – investigations at these latter sites have never revealed the underlying remains of a broch – but expressions of power and prestige distinctive to early medieval Scotland suggest that profound cultural divergence continued in the centuries that followed the demise of Roman Britain.
Pictish symbols, whether carved on stone or inscribed upon artefacts such as massive silver chains and silver ornaments, were only used in Scotland. While these are overwhelmingly concentrated north of the Forth, they are also encountered within non-Pictish contexts to the south and west in the Lothians, Lanarkshire, Galloway, and Argyll. The direction of influence was not one-way. Massive silver chains, which are also unique to Scotland, are concentrated in the south-east, reflecting their cultural origin there in the appropriation of Roman silver for a new way of expressing status and power. That silver chains are also found north of the Forth but not south of the Tweed and Solway is another sign of mutual cultural values in the expression of power and prestige among the Britons of southern Scotland and the Picts of northern Scotland, but not, apparently, among the Britons, Angles, and Saxons of England and Wales.
Culture should not be conflated with identity. The peoples of early medieval Scotland may have identified separately as Britons, Picts, and Scots but they nevertheless shared some distinct cultural traits. So too, it appears, did their Iron Age predecessors. Archaeological evidence suggests that Hadrian’s Wall was not a cause but instead an effect of an existing cultural divergence between the peoples of what later became Scotland and England: a divergence that was enough to stop the Roman army in its tracks and continued into the medieval period.
Further reading R Toolis, ‘Shifting perspectives on 1st-millennia Scotland’, Proceedings of the Society of Antiquaries of Scotland, Volume 150, https://doi.org/10.9750/PSAS.150.
Dr Ronan Toolis is a Director of GUARD Archaeology Limited.