Brittany is rich in archaeological remains from prehistory.Footnote ¹ At many junctures in the remote past, the peninsula has been a centre of cultural innovation or a corridor by which innovations have passed between the Mediterranean, inland Europe and the Atlantic. At the dawn of the Neolithic, it was the home of some of Europe’s earliest and most spectacular megalithic funerary monuments, in particular the sequence of long mounds, passage graves and tumuli around Carnac built between 4700 and 3500 BC. At about the same period, thousands of dolerite axes made from the local stone at Plussulien, Côtes-d’Armor, were transported all over western France. In the third millennium BC, rich grave goods and votive deposits show the region benefiting from its central position along the riverine and ocean trade routes from the Mediterranean to Britain. In the later Bronze Age Brittany may have been relatively isolated, a possible sign of this being the manufacture of thousands of non-functional bronze axes with a high lead content, purely for ritual burial. But from ca 500 BC Brittany’s external contacts revived, with signs that it developed a ‘middleman’ role in channelling materials such as tin and copper from southern Ireland and south-west Britain to the power centres of west-central Europe, and later to the Mediterranean. It may have been by way of Brittany that La Tène art spread to Britain in the fifth century BC – the region’s elegantly decorated pottery making use of the sinuous motifs of central European metalwork in a new medium.Footnote ² A dense settlement pattern reveals an elite able to assert itself with defended, banked and ditched enclosures and, on the coast, ‘cliff castles’. Another distinctive artefact that survives in thousands from Iron Age Brittany is the stela, or shaped stone column. On the eve of Julius Caesar’s conquest of Gaul, we glimpse the political make-up of the peninsula. It comprised five territorial units, civitates, that each issued its own sophisticated coinage and had become rich from trading in Mediterranean wine with southern Britain; it was part of a larger maritime region called in Gaulish Aremorica, the land facing the sea.Footnote ³ ‘Armorica’ (French Armorique) is often used by modern writers as a synonym for Brittany, or as a convenient term for the peninsula in the prehistoric and Roman periods before it was settled by Britons. It must be borne in mind, however, that historically the name ‘Armorica’ referred to different extents of land at different times, usually including more territory than what later became ‘Brittany’.Footnote ⁴

There has long been an appreciation of the complexity, and sometimes intensity, of the peninsula’s prehistoric contacts with Britain.Footnote ⁵ The exports and imports changed over time, and the preferred routes varied, yet the network of seaways had some constant features. These reflect the influence of coastlines, the prevailing winds, tides and seamanship – factors that remained significant in the early medieval period. In the early prehistoric period and at the end of the first millennium BC, seafarers used the mid-Channel crossings linking Christchurch and Poole Harbour with the bay of Saint-Brieuc (via the Channel Islands).Footnote ⁶ These routes were relatively reliable in terms of visibility and the chance of difficult conditions, although they could be more dangerous than the shorter Channel crossings to the east.Footnote ⁷ Even now, the Portsmouth to Saint-Malo crossing takes eight to eleven hours. A longer sea route was developed from Iberia to western Brittany, passing through the daunting waters around Ushant to Cornwall. This crossing involved at least ten to twenty hours out of sight of land, thus requiring deep-sea navigation skills.Footnote ⁸ Vessels will have varied in size and capability; by the Iron Age some had sails, such as the boats that Caesar admired amongst the Veneti of southern Brittany. To cross the Channel, these vessels would have had to perform with the prevailing south-westerly and westerly winds almost at a right angle.Footnote ⁹ Prior to the Roman interlude, then, seafarers had the aptitude and equipment to connect various parts of the Breton coast with southern Britain and the Irish Sea region. Prehistoric routes continued to be attractive to medieval seafarers, but political and economic conditions changed significantly under the Roman Empire.

During the early Roman period, Brittany’s archaeological visibility was maintained: the trappings of Roman culture were introduced here as elsewhere in the empire.Footnote ¹⁰ Yet the fact that Europe’s economic life was now entirely, and increasingly, harnessed to the political needs of a centralised empire was fatal to the peninsula’s distinctive role as a zone of communication.Footnote ¹¹ The main artery from the Mediterranean to Britain was now the Rhône–Rhine corridor, with the Atlantic coast reduced to marginality; under the later Empire the military demands of the Rhine frontier increasingly dominated the movement of goods between Britain and the Continent. Even at its most prosperous, Roman Armorica was relatively a backwater. When the expansion of the empire gave way to retrenchment after its political crisis in the 260s AD, the region’s prosperity rapidly evaporated. From this point until the central Middle Ages, the material remains of life in Brittany become all but invisible.

With Patrick Galliou, one can only lament this lack of evidence as ‘all the more unfortunate since in the midst of those dark years the old Roman order faded into oblivion and a new world was born’.Footnote ¹² These were the centuries that saw the redefinition of the peninsula as ‘Brittany’, the introduction of a British Celtic language and the formation of a distinctive society and Christian culture. Similar transformations elsewhere in early medieval Europe have been reinterpreted in recent decades, thanks to the greater availability and better understanding of archaeological evidence. But with such a shortage of information, how is similar progress to be made in understanding Brittany?

Brittany within Gaul

Understanding the antique/medieval transformation of Brittany is embedded in a larger problem, that of understanding the end of Roman imperial power in Gaul as a whole. The reaction of a student who is reasonably familiar with Roman Britain, and approaches the study of Roman Gaul for the first time, is likely to be astonishment at how little has been written about it, compared to the intense scrutiny that Roman Britain has received from Anglophone scholars.Footnote ¹³ Gaul, a much larger and more varied area than Britain and with a longer Roman history, tends to be treated either in broad-brush terms as part of the Western Empire as a whole or on a regional level; certain themes and regions are much more thoroughly covered than others.Footnote ¹⁴ A largely descriptive approach to the very visible culture of the ‘high’ imperial period gives way, in the later Roman period, to two major themes: the rise of Christianity, with associated developments in social forms and mentalités; and the history of the frontier, with the supposed development, in its hinterland, of a ‘barbarised’ or ‘militarised’ society.Footnote ¹⁵ In fact, Gibbon’s famous duality, ‘the triumph of barbarism and religion’, is alive and well as a lens through which to see late antique Gaul.Footnote ¹⁶ It is hard to shake off the approach that spends the entire long fourth century ‘waiting for the barbarians’, examining only phenomena that explain, or tend towards, the dramatic collapse of the fifth century. This approach also entails a regional bias. The source material is concentrated, on the one hand, on the Rhine frontier and its hinterland in north-east Gaul and, on the other, on Gaul south of the Loire, which remained connected to the Mediterranean, where villa and town life survived longer and we possess written sources, such as the letters of Sidonius Apollinaris, that illuminate the interactions of Roman and barbarian. This leads to analyses in which a broad contrast between northern and southern Gaul is outlined, but in which the north-west, including Brittany but also the Pays de la Loire and western Normandy, is rarely discussed. As will be seen, this brings difficulties both in characterising the transformation of Brittany and in assessing to what extent it shared in broader regional developments.Footnote ¹⁷

The current understanding of the history of the Armorican peninsula under Rome may be summed up as follows (Map 1.1). From Caesar’s conquest in 58–50 BC until the late fourth century AD, the peninsula is mentioned in no written sources, but judging by its archaeology it seems to have differed little from the rest of Gaul in its absorption of Roman culture and its participation in the imperial economy. Some decades after the conquest, once the empire had been stabilised under Augustus, the five civitates in what would later become Brittany – the Riedones, Coriosolites, Osismi, Veneti and Namnetes – like the rest of conquered Gaul, were incorporated into the provincial administration and provided with capital cities and a road network. A classic Roman built environment was created in the civitas-capitals and in a number of smaller towns or vici, and Roman religion and burial customs were taken up there, even if they did not penetrate deeply into rural areas. In the countryside a large number of villas attested to the intensification of agriculture and growing prosperity. There was some economic specialisation seen, for instance, in the fish-processing facilities at Douarnenez and elsewhere, and mining operations including extraction of tin at Abbaretz-Nozay.Footnote ¹⁸ However, the imperial instability of the third century and the first substantial Germanic invasions of Gaul in the 260s dealt the region a blow from which it never recovered. After ca 280, town life was severely retrenched, with defensive walls being built around the civitas-capitals of Rennes, Nantes and Vannes, while those further west – Corseul and Carhaix (Vorgium) – may have had their administrative functions transferred to the newly built coastal forts of Alet and Brest.Footnote ¹⁹ Occupation continued on some villa sites, but it was no longer the lifestyle of the cultured landowner: open hearths, craft activities and rubbish tips intruded into the living quarters, while aqueducts and baths fell into disuse.Footnote ²⁰ A limited urban revival took place in the fourth century, when coins and imported pottery show that Brittany remained economically connected to the rest of the north-western provinces to a modest degree, but the latest Roman coins found on occupation sites anywhere in Brittany are from the first decade of the fifth century.Footnote ²¹ While the Western Roman Empire disintegrated during the fifth century, archaeological evidence in Brittany is reduced to a few burial sites, mainly in the east, and a handful of settlement sites at an extremely basic level.

Map 1.1 Roman Brittany. The Roman road network is from M. McCormick et al. 2013, ‘Roman Road Network (version 2008)’, DARMC Scholarly Data Series, Data Contribution Series #2013–5. DARMC, Center for Geographic Analysis, Harvard University, Cambridge MA 02138 URL: https://darmc.harvard.edu, accessed 19 February 2020. DOI: https://doi.org/10.7910/DVN/NIAWTY

The dramatic decline of Roman material culture in the Armorican peninsula – and in most of northern Gaul – has tended to be attributed to a range of causes: the political turmoil of the middle years of the third century, with their rapid succession of military coups; the resulting ‘barbarian’ incursions into Gaul across the Rhine frontier; the ongoing menace of ‘Saxon’ and ‘Frankish’ pirates on the Channel and Atlantic coasts; social unrest in Gaul linked to high taxation and rural poverty; and the implantation of ‘barbarian’ populations on imperial territory – in the case of Brittany, it has been suggested that these were the ‘non-Romanised Britons’ who gave the region its Brittonic language.Footnote ²² However, the real impact of all these factors, as well as sometimes their very existence, is difficult to assess.Footnote ²³ Why should villas have been abandoned, and cities in the far west of Gaul have shrunk and been forced to put up massive defences because of incursions across the Rhine hundreds of miles away? The peak in coin-hoards, that used to be adduced in support of a generalised invasion panic in Gaul in the 260s–270s, has been reinterpreted as being linked to the debasement of the coinage and the demonetisation of the coins of the ‘Gallic Emperors’ who ruled from 260 to 274; the coins, mainly of low value, were ‘dumped’ rather than hidden.Footnote ²⁴ The written sources for piracy and seaborne raids in the fourth century are scattered and imprecise, including a high proportion of panegyrics. They cannot necessarily be connected with archaeological evidence for the destruction and/or abandonment of Roman buildings.Footnote ²⁵ The locations, motives and impact of recorded social revolts are likewise well-nigh irrecoverable.Footnote ²⁶ The reoccupation of Roman buildings in ways that negated or ignored their original functions – what used to be called ‘squatter’ occupation – is a very widespread phenomenon in northern Gaul in the fourth century, and elsewhere in the Western Empire in the fifth and sixth centuries; it is no longer assumed to be associated with ‘barbarian’ settlement or even with the flight or drastic impoverishment of landowners, but rather with a switch in the favoured modes of display by the elite, from a leisured secular civilian lifestyle to either military trappings or the Christian Church.Footnote ²⁷ Recently, it has been suggested that climatic factors (a change to cooler, drier growing seasons, reducing crop yields across Europe) may have contributed to all these political and economic changes.Footnote ²⁸ However, it is problematic to claim, and still more so to explain, a ‘catastrophic’ breakdown in the Roman order in northern Gaul from the late third century onwards; and within this already enigmatic picture, it still needs to be shown whether, and to what extent, the future Brittany was exceptional.

As stated earlier, northern Gaul tends to be treated as a single region within the late Empire, but it contained great disparities and is unevenly studied. Scholarly attention has been focused on Gallia Belgica and particularly on the frontier regions. When the evidence for the better-studied parts of northern Gaul (basically, an arc running east from Normandy to the Rhineland) is generalised, it can be claimed that the salient feature was the ‘militarisation’ of society – that the fourth-century economy was geared to the support of a higher density of defence works and intensified military production. The regional decline in villas from the later third century onwards has been shown to have been partly offset by the appearance of new rural settlements, less substantial and less formally organised, with wooden instead of stone buildings, but clearly still agriculturally productive and in touch with the outside world: indeed, an increased number of grain storage and drying facilities suggests intensified production.Footnote ²⁹ It is possible to talk of a ‘medium-term cycle with increasing numbers of rural settlements from the late Iron Age into the earlier Roman period and then a settling back through the later Roman period, rather than a precipitous decline in the third century’.Footnote ³⁰ In much of northern Gaul, moreover, the reduction in settlement remains is to some extent balanced by the greater visibility of burials: the furnished burial rite, with weapons, brooches, buckles and other grave goods, provides evidence of social organisation and a degree of wealth, although the interpretation of these burials is controversial. They were long supposed to be those of ‘Germanic’ federate troops established within the empire for internal defence, but several scholars have pointed out the Roman associations both of the burial rite and of the items buried, and have reinterpreted the burials as symptoms of social insecurity and competition on the part of local elites or as representing founder-status in a new (but not necessarily exogenous) kind of landholding.Footnote ³¹

A large intermediate area of Gaul, including Brittany but also the modern region of the Pays de la Loire, is even harder to characterise. Here, villa buildings did not continue in use as they did south of the Loire, nor is there the evidence of a militarily mobilised economy that occurs further north and east. Furnished burials of the fourth and fifth centuries are much rarer. There seems to be little available study of late Roman burial customs in the Loire region where furnished burials did not come into vogue: one example is the excavated mid-third- to early fourth-century cemetery at Angers where an unassuming range of Roman burial styles, without grave goods, was continued – straight-to-ground burial, burial in wooden coffins, cists and tegulae (tile surrounds).Footnote ³² However, some evidence suggests that Brittany should be associated with north-eastern Gaul rather than the Pays de la Loire. Important evidence comes from the Notitia Dignitatum, a list of military commanders and their units throughout the Roman Empire, the main western section of which can be dated between 399 and 408.Footnote ³³ Here we find that the coastal towns of the Armorican peninsula were elements in a larger defended coastal zone, the Tractus Armoricanus et Nervicanus, reaching from the mouth of the Somme to the mouth of the Loire (or, according to one of the two descriptions, as far south as the Gironde).Footnote ³⁴ Perhaps in keeping with this militarised status, the furnished burial makes a tentative appearance in Brittany in the late fourth and fifth centuries. At Guer, Morbihan, two burials which may have been part of a larger graveyard yielded an axe, spearheads and a knife, glassware, a ‘chip-carved’ Roman official buckle and a Fowler C2 type penannular brooch, metalwork typical of the north-eastern Gaulish cemeteries of the second half of the fourth century.Footnote ³⁵ At Ty Korn, Gouesnach, in the far west of Finistère, thirteen graves (also possibly belonging to a larger cemetery) offer a similar late fourth-century assemblage including glassware, local and Argonne pottery, a belt-buckle and a C2 brooch.Footnote ³⁶ Three graves excavated at Carhaix in 2012 appear similar.Footnote ³⁷ At Saint-Marcel le Bourg, near Vannes, a slightly later cemetery of forty-five burials was excavated in 2006: some of the graves contained pottery and glass, and one in particular, dating from the second quarter or the middle of the fifth century, included an axe-blade, a spear-point, a knife and official or military regalia with belt-buckles in the ‘Quoit Brooch Style’ associated with early fifth-century south-east Britain, and an early fifth-century bow brooch (fibule en arbalète) similar to examples found in ‘Germanic’ women’s graves in Normandy – or occasionally in male graves in imitation of official Roman ‘crossbow’ brooches: an interesting mixture of status markers.Footnote ³⁸ A Quoit Brooch Style buckle was also found with glassware in a single burial at Goas-an-Eyec, near Pont-de-Buis, Finistère, in 1911.Footnote ³⁹ The accidental discoveries of the Ty Korn and Saint-Marcel cemeteries – the first late antique cemeteries to be excavated to modern scientific standards in Brittany – place the region’s late Roman history in a new light. More such finds are likely, and might change the picture entirely.Footnote ⁴⁰ Such burials seem to constitute, at the least, a ‘fringe’ participation in whatever social development gave rise to the much higher numbers of furnished graves in Normandy, Picardy and further east, although the scale as seen so far is modest.Footnote ⁴¹ The excavators continue to view the burials as ‘Germanic’ and ‘military’ in character, associating them with the presence of Germanic federates (such as the Frankish laeti placed at Rennes in the Notitia Dignitatum) or of other incoming soldiers stationed in the countryside. Alternatively, they may simply represent leading local families asserting their position by means of whatever symbols were most available and effective. Yet on present evidence it seems that most burials in Brittany – including those in the larger early medieval cemeteries in the west, those of Saint-Urnel-en-Plomeur and Lostmarc’h en Crozon, Finistère – seem to have continued in Roman traditions, bodies being buried very simply with few or no grave goods or sarcophagi.Footnote ⁴² In this, Brittany is quite typical of late Roman Gaul outside the north-east.

The northern Gaulish, informal type of rural settlement with rectilinear, wooden, post-supported buildings, extends as far as the borders of Brittany but no further, with examples dating from the fifth century found in the Rennes area, at Montours, Janzé and Visseiche.Footnote ⁴³ In a more southerly, Loire Valley distribution, the elaborately built and decorative fourth-century town walls of Rennes, Nantes and Vannes have points in common with those of Angers and Le Mans: all these cities were to have a role as Christian centres, taking part in conciliar activity, in the fifth century and beyond.Footnote ⁴⁴ The Loire Valley as a whole played a notable part in late antique socio-religious developments in Gaul. The careers of St Martin, bishop of Tours (371–97) and his friends and disciples Liborius of Le Mans and Maurilius of Angers, began a well-studied movement of urban church-building, monastic enthusiasm and episcopal leadership in the region.Footnote ⁴⁵ Here was a new, officially recognised channel for aristocratic co-operation and local leadership, but to what extent it was able to compensate for the region’s apparent economic stagnation and social unrest is far from clear. To all appearances, the specifically Christian late antique urban renaissance penetrated no further west than Vannes.

The reduction in archaeological evidence becomes more extreme as one moves further west. No examples of new, non-villa rural sites of the fourth or fifth centuries have been found in Brittany west of Rennes, and many pre-existing rural sites seem to have been abandoned: Loïc Langouet calculated that in the civitas of the Coriosolites in central northern Brittany, 75 per cent of the rural sites active in the earlier Roman period show no activity datable after the end of the third century.Footnote ⁴⁶ Given the reduced range of datable artefacts in the fourth century and after, a 75 per cent figure for abandonment is certainly artificially high.Footnote ⁴⁷ Recent work has queried the generalised picture of villa abandonment and ‘squatter’ occupation in the fourth century: some luxurious villas (mainly on the coast) were given new and ambitious architectural features, or at least continued in high-status use for the first half of the fourth century.Footnote ⁴⁸ In the westernmost civitas-capital, Carhaix (Vorgium), too, recent rescue excavation shows an important house continuing to be occupied in ‘high Roman’ style until the mid-fourth century.Footnote ⁴⁹ At Carhaix and in the burials at Ty Korn, a considerable amount of locally manufactured, wheel-thrown pottery of the fourth century, of reasonable quality although plain in style, has been found. If it were possible to identify and catalogue these pottery types, this might provide a more sensitive dating indicator that could be used to illuminate fourth-century settlement and activity more generally.Footnote ⁵⁰ In general, work in progress is tending to undermine the idea of a definitive late third-century crisis in western Gaul and to point to a gentler decline some decades later. However, the decline did eventually come.

Patrick Galliou has argued that fourth-century Brittany was still part of the ‘normal’ late Roman world to the extent of being part of its established trading networks: this is demonstrated by the fact that a range of urban and rural sites have yielded fourth-century imported pottery including Argonne ware, céramique à l’éponge from southern Gaul, and Black Burnished ware from Dorset.Footnote ⁵¹ (But while fourth-century southern Gaulish ceramics à l’éponge are found throughout Brittany, the southern Gaulish sigillées paléochrétiennes that took over from this type in the fifth century are not found further west than Morbihan.Footnote ⁵²) Galliou attributes the fortifications of the ‘Saxon Shore’ and its continental counterpart the Tractus Armoricanus et Nervicanus to the need to safeguard such coastal trade. The evidence for military investment in fourth-century Brittany – the remodelling of the road network, the construction of city walls, the building or rebuilding of coastal forts and the provision of garrisons recorded in the Notitia Dignitatum – reveals a level of central imperial commitment that seems to argue the peninsula was still of value to the Roman government.Footnote ⁵³ Yet the main value may have been political and broadly strategic, rather than economic. Was the trade there to supply the forts, as much as the forts to safeguard the trade?

It might be suggested that the history of Brittany from the Roman conquest onwards was that of a ‘sidelined’ region, cut off from its ancestral sources of prosperity by the political priorities of the empire to which it now belonged. During the early imperial period, this was not apparent, because the general growth in the economic output of the Western Empire floated the peninsula upwards as well. However, under the late Empire, prosperity largely became dependent on closeness to the imperial court or to its major areas of expenditure, the land frontiers, and Brittany was a long way from both.Footnote ⁵⁴ Moreover, its elite under Rome had priorities that might easily override their regional ones. If native to Brittany, their greatest ambition will have been to make a mark on the all-Gaulish or imperial stage; if from outside, then Brittany will have been only one item in their portfolios. It seems likely that much of the investment and spending that went on there – the industrial enterprises like mining and fish-processing, and the most lavish villas – was the work of incomers or people with widely spread interests. (The late second- or early third-century inscription discovered at the Douarnenez fish factory, stating that the statue to the god Neptune Hippius was raised by C. Varenius Varus – a name that seems to hail from the far south of Gaul – and the local ‘club’ (conventus) of Roman citizens, makes the point nicely.)Footnote ⁵⁵ If political factors or other changes made it marginally more attractive to operate elsewhere in the empire, then they would do so, depriving the area of the leadership on which it had come to rely. Wendy Davies and Greville Astill have suggested that the disuse of villas after the third century may indicate that villa-owners relocated, perhaps outside the province.Footnote ⁵⁶ Subsequently, under the late Empire as reorganised by Diocletian and his successors, the state became the main buyer of goods and initiator of exchange. Brittany’s specialist products were apparently no longer required, and its people could not independently reactivate their former exchange networks, especially with Britain, since British trade and production was being directed elsewhere, as will be seen in the next section. If those local leaders who remained attempted to set their own priorities, they risked being seen as rebels.

The building of forts and city walls and the stationing of garrisons in Brittany may have had a two-fold purpose on the part of the imperial government: to quell incipient rebellion if necessary, and also to head it off by the assurance of a minimum of imperial favour and expenditure, in that a measure of tax revenue was being injected into the region. John Drinkwater suggested that the fourth-century ‘barbarian threat’ beyond the Rhine was a Roman artefact, stage-managed to justify spending on the imperial army, which was primarily needed to deal with internal threats; might the late-antique building programme in Brittany have had equally political, as opposed to strategic purposes?Footnote ⁵⁷ The remaining local bureaucrats and the army were the only reliable market left for primary producers who, under the early Empire, had catered to entrepreneurs and major regional landowners. When first Britain and then northern Gaul as a whole passed beyond imperial control in the course of the fifth century, even this market was put into jeopardy.

In the late fourth century the region ‘Armorica’ (including, although not confined to, the future Brittany) makes its first appearance in written sources since Caesar’s time.Footnote ⁵⁸ Repeated revolts in Armorica are mentioned from the early to the mid-fifth century. Whether these revolts represented, at different times, social risings by overtaxed peasants, attempted coups by sections of the army, or assertions of autonomy by local leaders who filled a power vacuum when Roman imperial authority could not make itself felt in their region, they fit into a background of increasing fragmentation of power in northern Gaul: groups of Saxons, Franks, renegade Roman generals and (in the 460s) a British army tussled for control and aligned themselves variously with rival imperial candidates, their army commanders and their barbarian allies.Footnote ⁵⁹ Fifth-century sources mesh awkwardly with those of the mid-sixth century and later and a coherent narrative cannot be created without a great deal of guesswork. However, by ca 550 it becomes evident that the Frankish kingdom of Clovis and his successors, the Merovingians, had been left as the last and only power standing in most of Gaul. The Loire valley and the eastern civitates of the Armorican peninsula, the Namnetes and Riedones and at first the Veneti, fell within this power bloc and their archaeology reflects this fact. Nantes, Rennes and Vannes housed Merovingian mints.Footnote ⁶⁰ Elaborate burial took a new lease of life: the distinctive sixth- to seventh-century practice of burial in limestone sarcophagi flourished in the Loire Valley region, and on the fringes of Brittany, the excavated cemeteries of Bais, Visseiche, Bréal-sous-Vitré, Rétiers and La Mézière, with their sarcophagus and cist burials and (in the case of the first three) their early association of cemetery with church, show the same sort of development as sites deeper into the Merovingian kingdoms.Footnote ⁶¹ Western Brittany, however, was by this time settled by Britons: according to Gregory of Tours’s account, British leaders (Brittani) captured Vannes in 578 and their power reached as far east as the River Vilaine.Footnote ⁶² What, if anything, can archaeological evidence suggest about the nature of this British takeover?

Late Roman Gaul and Britain

That Brittany received migrants from Britain we know – but there is no clear answer to the question of which parts of Britain provided the majority of the migrants, or exactly when. It might be hoped that archaeology would illuminate such questions. But as E. G. Bowen bluntly pointed out half a century ago, ‘of this extensive movement there remains no archaeological evidence whatsoever’, and he entered an early caveat against the assumption that ethnic and political identity are necessarily associated with material culture.Footnote ⁶³ Since he wrote, the absence of archaeological evidence for Breton origins has come to stand out even more starkly against the greatly increased (though uneven) amount of archaeological evidence available for the period in general. Why might this be?

In late third- and fourth-century northern Gaul, the archaeological picture is one of stagnation partly offset by militarisation. The Roman provinces of Britain present a strong contrast.Footnote ⁶⁴ It was in the fourth century that Britain’s ‘villa economy’ reached its peak of wealth and showiness in what is known as the civil zone, especially in the west of that area.Footnote ⁶⁵ The pottery industry expanded and diversified. One of the production centres was on the coast of Dorset: its distinctive ‘Black Burnished Ware’ was transported to the north and south coasts of Brittany as well as to Normandy, much of south-west Britain, and the northern British frontier.Footnote ⁶⁶ Towns changed in character, with an emphasis on private dwellings and the gathering and production of state supplies replacing the public buildings of earlier centuries, but if civitas-capitals declined somewhat, small towns grew.Footnote ⁶⁷ Britain was well connected with the new fourth-century imperial capital at Trier, and some of its agricultural produce went to feed the Rhine frontier army.Footnote ⁶⁸ In the military zone, that is, most of the north and west of Roman Britain, there was also the economic stimulus of catering to a substantial military force stationed on the island. Military infrastructure was focused on the long-term bases of Chester and Caerleon and on Hadrian’s Wall, an actively defended frontier if on a smaller scale than the Rhine limes. In Britain, the army several times enabled generals to make bids for imperial power – some successful, some not. Brittany, and the west of Gaul in general, was less important strategically, economically and politically, and can never have known even a fraction of the military spending that Roman Britain enjoyed.

It is interesting to consider to what extent the post-Roman trajectories of western Britain and Brittany were determined by their different fortunes under the late Empire. If their archaeology fails to show common characteristics, and their politics follow different courses, perhaps this is to be expected given that Britain’s ‘end of empire experience’ was arguably a rapid collapse, followed by a long period of political faction and conflict on multiple fronts; while that of western Gaul was a long, slow, obscure decline, during which it remained on the fringes of the main arenas of warfare.Footnote ⁶⁹

The theory that the British identity of Brittany was based on the establishment there of military forces of British origin before AD 400 has become, as Bernard Merdrignac remarked, ‘the standard view of the region’s history’ (la vulgate de l’historiographie régionale).Footnote ⁷⁰ The archaeological evidence adduced in support of early British military immigration included the presence of weapons and non-local metalwork and pottery. None of the material, however, is both unequivocally British and of the sort to have been brought by migrants rather than to have arrived by trade. The discovery of Dorset Black Burnished pottery together with military equipment in the excavation of the coastal fort at Le Yaudet (Finistère) is significant for the ‘military migration’ theory, yet the fact that Le Yaudet was a military base and received supplies from southern Britain in the fourth century is far from proving that it was manned by significant numbers of British, let alone British-speaking, troops.Footnote ⁷¹

The fact that there is no clear material evidence certainly does not of itself disprove the presence of soldiers from Britain. Both in Britain and in northern Gaul, military personnel during the late Empire are difficult to trace archaeologically, although they must have been omnipresent. In the absence of fortifications and purpose-built barracks, even large bodies of troops such as must have been present at the capital, Trier, are invisible.Footnote ⁷² In Brittany, likewise, the garrisons listed by the Notitia Dignitatum at Nantes, Rennes, Vannes, Alet and probably Brest, although their existence is scarcely in doubt, have left no archaeological evidence that has been discovered to date. How much more invisible would be irregulars from Britain billeted among the general population, particularly if they did not usually use weapon burial?

The other side of the question is the probability of an occupying force in Brittany being recruited entirely or mainly from the less Romanised, British-speaking population of Britain – from Wales and the far south-west, as has been argued – during the fourth century.Footnote ⁷³ What would be the context for such recruitment? Northern Wales had remained under military control throughout the Roman period, without civil local-government institutions, and neither it nor the south-west show many traces of classically Roman material culture.Footnote ⁷⁴ It has sometimes been argued that these remote regions remained effectively under native British rule.Footnote ⁷⁵ However, there seems little reason to believe that they served as a source of military manpower to be deployed to other parts of the empire. In frontier regions which did thus serve as recruiting-grounds, there is archaeological evidence for soldiers returning as well as leaving, bringing prestigious Roman artefacts with them. Where local leaders retained independent power, the imperial authorities ensured their good will with gifts and subsidies, as well as controlling them with strategic fortifications.Footnote ⁷⁶ No evidence of this kind has been found in Wales and south-west Britain: Roman artefacts at ‘native’ Welsh sites consist of modest quantities of civilian goods like domestic pottery.Footnote ⁷⁷ The Britons had no martial reputation in the wider empire; the idea of western Britons as turbulent frontier tribesmen (tribus frontalières, in Soazick Kerneis’s words) comes from a later period.Footnote ⁷⁸ The Roman military presence in Wales had been much reduced since the conquest period and by the fourth century the remaining garrisons were used to protect the region from seaborne raiders from Ireland, rather than to police the local inhabitants.Footnote ⁷⁹ Finally, north Wales and the south-west were sparsely inhabited. It may be doubted whether even active recruitment in these areas would have supplied enough manpower to form the majority of an effective defence force for Brittany: Brittany’s three westernmost civitates alone have about the same area as the whole of modern Wales and are larger than the four south-western counties of England combined. No recent specialised research on the archaeology of Roman western Britain takes any account of a possible drain of manpower to late Roman Brittany, or what social or material conditions might be associated with it. A re-examination of the question from this angle would be desirable.

The Material Culture of Post-Roman Britain and Brittany

If the archaeology of fourth-century Britain and Brittany provides no clear-cut evidence of population movement from one to the other, what of the period after Roman rule in Britain ended, the fifth and sixth centuries? Discussions of the Insular links of Brittany during this formative period have generally avoided making much use of archaeological evidence. This was understandable until the 1970s, when post-Roman Insular Britons were mostly archaeologically unidentifiable, and it was thought that they had suffered massacre, expulsion or assimilation to incoming Anglian and Saxon groups. However, now that there is a growing and comparatively well understood body of evidence for the material culture of some post-Roman Britons, particularly in the western highland zone, it becomes necessary to compare it with what we know of Brittany. Part of the accepted view of sixth-century Brittany as expounded by Francophone scholars is one of high-level political co-operation with Britain, within identifiable dynasties with Welsh, Cornish and Breton branches. It is legitimate to ask whether archaeology can shed any light on such putative contacts.Footnote ⁸⁰

In Britain, there is an apparent hiatus between the early fifth century and the 470s when developments in material culture cannot be detected among the Britons. This is the ‘twilight’ period when Roman coins and pottery no longer provide reliable dating for archaeological sites, when Roman material culture may have been used residually, but was no longer being ‘refreshed’, and innovations (such as sunken-featured buildings [Grubenhäuser] and furnished burials containing pottery, weapons and jewellery) have been read as the work of Germanic immigrants. Exceptional as a new and dateable artefact type during this period is metalwork decorated in the ‘Quoit Brooch Style’. Brooches, belt plaques and similar objects in this style, which combines late Roman and Germanic decorative motifs, have been found, mainly in graves, in south-east England, western Normandy and Brittany: several new finds have come from the early to mid-fifth-century cemetery at Saint-Marcel, Morbihan.Footnote ⁸¹ Attempts have been made to link the style to political groupings in Britain, or to Germanic mercenaries, but these are inconclusive. The style is probably continental in origin, and shows communication between north-western Gaul and southern Britain persisting through the political chaos of the fifth century. This can be connected with the movement of individuals and small groups of migrants that occurred throughout the Roman period and after but has no particular implications for the formation of Brittany.Footnote ⁸²

From about 450, archaeologically distinctive phenomena began to emerge in still-British areas. Iron Age hill forts were being reoccupied and refortified in the south-west and in South Wales, showing signs of metalworking and other craft activities: South Cadbury (Dorset), Cadbury-Congresbury (Somerset) and Dinas Powys (Morgannwg) are the best known examples. The Somerset and Dorset examples may hint at a political rupture between the Romano-British civitas of the Durotriges and the (now Saxon-ruled?) territories to the north.Footnote ⁸³ From about 475 to 550, a number of sites in the south-west and in Wales (as well as in the Irish Sea zone more generally) show quantities of imported East Mediterranean and North African pottery (Phocaean Red Slipware and African Red Slipware, and amphorae – food storage containers – from the East Mediterranean).Footnote ⁸⁴ This reveals renewed (or perhaps continued) contact between western Britain and the Eastern Roman Empire, in the form of directed trade in which Cornish tin may have been the main commodity exchanged.Footnote ⁸⁵ It overlaps with a longer-lasting sequence of imports of continental pottery (‘D’ and ‘E’ ware) and glassware, harder to date because the styles are less distinctive, continuing approximately until the end of the seventh century.Footnote ⁸⁶ Concurrently, the British-controlled parts of Britain were developing local artefacts in which Iron-Age Celtic stylistic motifs were rediscovered, in contrast to the ‘Germanic’ Salin I style being developed in the south-east. The most distinctive artefact types are the ‘hanging bowls’ of ca 550–650, and the penannular brooches, Fowler Type G, which are thought to have originated around the Severn estuary in the late fourth or fifth century and spread, with local variations, to many parts of Celtic Britain and Ireland.Footnote ⁸⁷

The fifth and early sixth centuries, despite or perhaps because of political turmoil, seem to have seen the progress of Christianity to become the only visibly practised religion in western Britain. However, funerary practices, which are the most widespread material evidence of religious or ritual behaviour, did not change very much in western Britain at this time. A large-scale shift from cremation to oriented inhumation burial had already taken place in Britain, in common with the other western provinces of the empire, in the third and fourth centuries. In many places in western Britain this involved a revival of an Iron Age custom that may never have completely died out in its heartland from north Dorset to South Wales: that of burial in graves sided with stone slabs, otherwise known as cists.Footnote ⁸⁸ Many excavated cemeteries overlap the Roman and post-Roman periods, and contain a mixture of cist burials and burials dug straight into the ground, with a sprinkling of other types, over the entire period of a cemetery’s use. For instance, the cemetery at Cannington, Somerset, was in use between the fourth century and the eighth or ninth, and contained east-to-west oriented, plain dug, rock-lined and cist burials; at Caerwent, where two cemeteries were in use between the late Roman period and the eighth century, one contained cist burials, the other did not.Footnote ⁸⁹ In Cornwall, however, cist burial was revived only in the post-Roman period after what seems to have been a long period of disuse; the same may have been true of Devon, although the county has a dearth of burial evidence of the Roman and post-Roman periods.Footnote ⁹⁰ Burial rites did change dramatically in most of the lowland zone of Britain from Hampshire to Yorkshire; however, from the fifth to the seventh centuries, this zone contains highly visible furnished burials, in which the dead were buried fully clothed (or cremated) with suites of jewellery, weapons and sometimes other grave goods. These burials have usually been connected to Anglian and Saxon ethnic identity and used to identify the areas settled by these groups, but they also correlate closely to the distribution of Roman villas in Britain, and Guy Halsall has argued that they represent a response to social crisis on the part of ‘Anglo-Saxons’ and ‘Britons’ alike throughout what had been the prosperous grain-producing parts of Britain, analogous to the appearance of furnished ‘weapon burials’ in fourth-century northern Gaul.Footnote ⁹¹

In the far west and south-west, particularly in north-west and south-west Wales and in Cornwall, there was a new fashion in the late fifth and sixth centuries for erecting monuments – usually undressed, roughly pillar-shaped stones – bearing the name of a commemorated person or people in the Latin alphabet and sometimes in the Irish Ogam script, and occasionally some additional information such as a Christian phrase or epitaph, or the deceased’s age, status or occupation.Footnote ⁹² This development may indicate the rise of a new elite and a reorganisation of landownership, perhaps connected to the endowment of the Church: this may also perhaps be indicated, in Cornwall, by gradual changes in the settlement pattern in the sixth and seventh centuries, as the most visible settlement type of the Roman period – the fortified farmstead, or ‘round’ – was replaced by the unfortified tref.Footnote ⁹³

This summary of the main developments in post-Roman southern British material culture has been brief and perhaps excessively simplified: but even a simple presentation is enough to show a considerable contrast with the archaeological information available in Brittany at the same time. In Brittany we find no distinctively Celtic, post-Roman decorative metalwork: no hanging bowls, no penannular brooches, nor, indeed, any precious metalwork objects from between ca 400 and 800, except for a handful of minimally decorated silver and bronze bracelets and rings.Footnote ⁹⁴ There are a few signs of post-Roman occupation of earlier fortified sites, such as Kastel Kerandroat, Le Yaudet and Comblessac.Footnote ⁹⁵ However, in no case is there evidence of any elite activity – luxury imports, metalworking, glassworking – that would point to the places in question being foci of secular leadership of the sort that are found in most parts of Celtic Britain in the ‘Dark Ages’. It is perhaps less surprising that there are no precious finds, when there seem to be no central places in which to find them.

From this lack of high-status settlement evidence, one might simply conclude that Brittany remained more similar to the rest of Merovingian Gaul than to Britain. In the rest of northern Gaul, too, there is a lack of archaeological evidence for elite settlement during this period. However, unlike in Brittany, there is alternative evidence of aristocratic wealth: notably grave-goods, but also the continued widespread distribution of high-quality pottery.Footnote ⁹⁶ As regards pottery, post-Roman Brittany did not apparently participate either in continental or in insular distribution networks. The Mediterranean and continental pottery imports that are a diagnostic feature of elite centres in Cornwall, Devon and Wales are extremely rare in Brittany. Two sherds of a Bii (LR1) amphora have been found at a putative monastic site on the Île Lavret and one at Le Yaudet.Footnote ⁹⁷ To all appearances the Mediterranean trade that targeted western Britain passed Brittany by, but for these two northern coastal sites that may have obtained their examples through personal contacts or by accident. It is perhaps more astonishing that ‘E ware’, which has been thought to originate in the Charente region of southern France, some 200 km south of Brittany, is so rare there when it occurs on many sites in western Britain and Ireland. Charles Thomas catalogued three examples from Brittany: one from Les Cléons, Loire-Atlantique; one from Plaudren, Morbihan; and one from a sand-dune site at Guissény, Finistère, to which Ewan Campbell adds a sherd from Le Yaudet.Footnote ⁹⁸ In no case is there evidence for more than a single pot, making it impossible to assume that the pottery was being imported in any quantity. No other pottery except rough local wares can be seen in Brittany between the fifth and the ninth centuries. If pottery that was being made for export in the coastal region practically next door to Brittany failed to penetrate there, it seems to suggest that Brittany was not seen as a market for goods or as a target for gifts or diplomacy. There are hints that tin- and zinc-producing areas were active in sixth-century Brittany, as in Cornwall, but there are no archaeologically visible effects on the wider economy of Brittany.Footnote ⁹⁹ Similarly, sixth- and seventh-century East Roman and Merovingian coins have been discovered as single finds in coastal and riverine locations in the west of Brittany, but in insufficient numbers to imply any specific trade, diplomatic or ritual activity.Footnote ¹⁰⁰

An unknown quantity affecting this whole analysis is the availability of seaborne transport. While the longue durée of Atlantic contact hints at similarities between prehistoric and early medieval travel, the political and economic background was ever-changing, and this will have affected the availability and desirability of transport.Footnote ¹⁰¹ It is probable that at first the Roman military and merchant transport system was used, as it often was by fifth-century ‘migrating peoples’.Footnote ¹⁰² Breton rulers of the sixth century had some sea power at their disposal, and historians have suggested that they had taken over whatever was left of the Roman military fleet of Britain – the classis Britannica of the Notitia Dignitatum.Footnote ¹⁰³ Humbler British migrants might have travelled on vessels taking Dorset Black Burnished pottery to north-east Brittany, a trade which may have continued through the fifth century; in the sixth, Ben Guy has speculated that ships bringing wares to western Britain from the Mediterranean may have taken paying passengers to Brittany on the return journey, but this is a hypothesis that cannot easily be tested.Footnote ¹⁰⁴ There is no evidence that Breton rulers after the sixth century used ships for military purposes. Marine archaeology has yielded no vessels from early medieval Brittany, although inferences can be drawn from earlier shipwrecks and those found in surrounding waters. The combined evidence of excavated wrecks, textual allusions and surviving artworks suggests that there would have been two shipbuilding traditions available. One was the carvel-built ‘Romano-Celtic’ craft made of wooden planks nailed together, perhaps with leather sails, descendants of the kind of ships Caesar described as belonging to the Veneti in the first century BC, an example of which may be seen in the second-century wreck at St Peter Port, Guernsey; the other was the older, light Atlantic ‘curragh’ made from animal skins stretched over a wicker frame, of the sort described in the Voyage of St Brendan and Adomnán’s Life of St Columba and still used in the Aran Islands in the early twentieth century.Footnote ¹⁰⁵ Turning to hagiography (which provides nearly all the source-material), most references to ships and voyages are vague and stylised.Footnote ¹⁰⁶ Jean-Christophe Cassard sees an empaysannement (‘peasantification’) of the Bretons, a turning away from the sea and the coastline, particularly from the tenth century to the twelfth.Footnote ¹⁰⁷ Even in the twelfth century, when the Angevin empire of Henry II (1154–89), stretching from Ireland to Aquitaine, was geographically centred on Brittany, there is little evidence that Henry exploited the peninsula’s maritime potential: strategically, he was mainly concerned with its long land borders with his other Continental possessions.Footnote ¹⁰⁸ Brittany’s paucity of goods for long-distance trade must have inhibited communication by sea, and made the time and risks involved in any crossing less worthwhile. The Loire estuary stands out in this context because of salt production in the vicinity, which attracted Irish, English and Frankish traders.Footnote ¹⁰⁹

On the other hand, Cassard’s evidence is mainly for clerical attitudes as expressed in hagiography. More indirect evidence – for instance that of texts and manuscripts – shows that Brittany never fell entirely out of contact with the Insular world. The paucity of visible trade and military sea power does not necessarily mean that the Bretons turned away from the sea: it may simply mean that they travelled chiefly in pursuit of archaeologically invisible goods, such as social advancement, education, employment and religious benefits.Footnote ¹¹⁰

Early Medieval Christian Archaeology

A substantial and growing subset of the archaeological record in early medieval Celtic Britain relates to the growth of Christianity. In this field as in others, the differences between Brittany and the other Celtic-speaking regions are greater than the similarities, and have become more noticeable as the Insular record has been more thoroughly studied.

In the 1970s, almost as little archaeological material relating to early medieval Christianity was available for the British Celtic regions as for Brittany. Gildas Bernier’s 1982 monograph Les chrétientés bretonnes continentales barely adduced any archaeological evidence, but neither did Siân Victory’s The Celtic Church in Wales, published in 1977. Since then the Insular Celtic churches have been the subject of intensive study.Footnote ¹¹¹ Philippe Guigon has proposed abandoning the designation of the early medieval Breton Church as ‘Celtic’, since the material evidence it left was much more directly comparable to the Church elsewhere in Merovingian and Carolingian Gaul than to the Insular churches.Footnote ¹¹² However, this conclusion does not seem to apply positively to any parts of Brittany except the eastern borderlands. With regard to Brittany west of the Vilaine, the most noticeable feature of the evidence deployed in the article is its scarcity, compared both to that available for the Gaulish Church in general, and that examined for the Insular churches in other papers in the same volume. Unless more archaeological evidence is discovered, Brittonic-speaking Brittany stands out more for the unique elusiveness of its early Christian material culture than for either ‘Insular’ or ‘Continental’ characteristics. Nevertheless, some comparisons and contrasts can be attempted between Brittany and neighbouring regions.

An important strand in early Christian archaeology is the study of burials, and the way that burial places and places of Christian worship coalesced in the course of the early medieval centuries. In France as in the Breton borderlands, this association was present from the Merovingian period onwards; in Wales it gradually developed from the seventh to ninth centuries onwards – although in both regions, as also in England and Ireland, there was a great variety of burial-places and churchyard burial does not seem to have become obligatory until the central Middle Ages.Footnote ¹¹³ These conclusions are based on abundant data in France, England, Ireland, and in Wales, where ‘there are now over a hundred locations where cemeteries of extended inhumations can be identified during the period c AD 400–1200’, although some of these may need to be re-dated to the late Roman period.Footnote ¹¹⁴ By contrast, in Brittany west of the Vilaine, only four cemeteries dating from a comparable period (between the end of Roman power and the beginning of the eleventh-century Gregorian reform) have undergone modern excavation: the dune sites at Saint-Urnel-en-Plomeur (F), apparently in use from the fifth to the tenth centuries, and Pléherel, or Fréhel, near Saint-Malo (CA); and the monastic burial grounds of Landévennec (F) and Île Lavret (CA).Footnote ¹¹⁵ For a number of others, including the apparently large dune cemetery at Lostmarc’h, Crozon, the only available information is reports of discoveries made in the late nineteenth or early twentieth centuries.Footnote ¹¹⁶ Part of the problem is the difficulty of discovering sites given the poor survival of skeletal remains in the acid soils of Brittany combined with the prevalence there of burial in plain-dug graves, lacking both the grave goods of early English and Frankish burials and the stone cists that make some burials in Celtic Britain conspicuous. Even so, many potentially promising sites, such as medieval parish graveyards with possibly early above-ground features, are uninvestigated.Footnote ¹¹⁷ More data are gradually becoming available: recent excavation at Leslouc’h, Plouédern (F) has revealed two small quadrangular and three circular ditched enclosures of the seventh to eighth centuries, which the excavators suggest may have had a funerary use, with comparanda both in the Merovingian cemeteries of western Gaul and in Anglo-Saxon England.Footnote ¹¹⁸ As yet, however, it is impossible to generalise about early medieval burial in Brittany, for instance to say whether burials of the early Middle Ages often re-used prehistoric monuments (as in the Insular world in general), or to identify features such as ‘special graves’ and mortuary chapels or eglwysi-y-bedd, which marked the progress of the Christianisation of burial in Wales.Footnote ¹¹⁹ It does not seem that burial practice in Brittany was influenced by Insular western British practice. Cist burials can be found in post-Roman Brittany but only in the extreme east of the peninsula, under the influence of a separate cist-using tradition in northern Gaul; they were not used in western Brittany, which might be thought the most likely to have been influenced by the cist-users of south-west Britain.

Some comments can be made about the dune cemetery at Saint-Urnel, with the proviso that they cannot necessarily be generalised. This cemetery has been excavated several times, most recently in a thorough programme led by Pierre-Roland Giot in 1973–5.Footnote ¹²⁰ The cemetery, containing some thousands of burials at successive levels in a wind-blown dune, was in use for several centuries. The sand provided an unusually favourable environment for the preservation of human remains, but the absence of grave goods, or even traces of clothing, coffins or shrouds, made dating almost impossible until radiocarbon dating became available. This revealed, to archaeologists’ surprise, that the cemetery was early medieval rather than prehistoric. The first burials may have dated from as early as AD 320 and the latest from around the year 1000. The only signs of ceremony were the careful disposition and alignment of the bodies and the presence of ‘surrounds’ of stones, or occasionally whalebone, on the surface of the graves. The conservatism of burial practice at this site throughout late antiquity and the early Middle Ages is remarkable. Giot held in 1977 that the characteristics of the skeletons were more closely comparable to those in contemporary burials from south-west Britain than those from Iron Age or Roman Brittany, and that ‘the only serious interpretation’ was that this was a population of British immigrants.Footnote ¹²¹ In subsequent publications he was more cautious.Footnote ¹²² A reinterpretation of the evidence using newer techniques such as carbon isotope analysis might be of interest. But whatever the origins of the Bretons of Saint-Urnel, their social profile fits well with what has been noted about the material culture of early medieval Brittany in general. The burials do not indicate any kind of social hierarchy in death, yet the skeletal data show that the people were healthy and well-fed by medieval standards with few signs of medical or violent trauma.

The field study of parish churches in Celtic Britain has yielded a considerable amount of information about their early medieval origins. In Wales and Cornwall, numerous churchyard enclosures can be identified as early, with the corollary that existing churches, built much later, are still on their early medieval sites. Some churchyards are reused Iron Age enclosures, mainly circular earthworks but occasionally promontory forts; others, circular, rectangular or concentric, were apparently created in the early medieval period to mark the limits of burial grounds or sanctuary space.Footnote ¹²³ The possibility that similar features existed in Brittany remains to be followed up. Ninth-century hagiography and charters give the approximate locations of many early churches, yet little research seems to be available on the question of whether the existing churches remain on their earliest sites and whether any early features might be identifiable through fieldwork or excavation.Footnote ¹²⁴

Within Brittany (west of the Vilaine) as a whole there are doubtless many local variations, yet some general features make comparison with Celtic Britain difficult and suggest some fundamental differences. The late medieval and modern development of the ecclesiastical landscape in Brittany is very different from that in either Wales or Cornwall. In these latter regions, substantial parish churches and a number of chapels were built in the later Middle Ages, after which there was little church-building until the nineteenth century. In Brittany, most existing church buildings are later in date, having been rebuilt in the sixteenth century and after with probable substantial remodelling of the surrounding features: this is particularly apparent in the case of the enclos parroissiaux of northern Finistère, where the creation of walled churchyards with monumental entrances in the sixteenth and seventeenth centuries may have destroyed earlier enclosures and other features. There are also large numbers of chapels, most of them comparatively recent foundations but some probably occupying early sites of worship. The potential data set is therefore enormous, but its historical exploitation is difficult.Footnote ¹²⁵ At first glance, the average parish church in Brittany has a very different aspect from that of a typical Welsh or Cornish parish church. Even in small villages, the church tends to be set in a built environment that is determinedly civic, often sharing a central square with the mairie and/or market hall or attached to other ranges of buildings. Some churches can be entered directly from streets or squares rather than being set back within churchyard enclosures as is the case with most rural churches all over Britain – not just in the Celtic regions.Footnote ¹²⁶ Churchyards are typically rectilinear, walled, levelled and paved rather than grassed, suggesting the obliteration and modernisation of earlier boundaries. In Cornwall and Wales many parish churches have remote settings: in Cornwall, in particular, many are in waterside (riverine or maritime) locations.Footnote ¹²⁷ In Brittany, isolated waterside churches exist but they are chapels rather than parish churches.Footnote ¹²⁸ The principal churches in most parishes, and even the majority of chapels, appear more deliberately located to serve a lay community than is typical in Wales or Cornwall.Footnote ¹²⁹ This may be partly a result of Brittany’s early modern prosperity and devoutness, attracting settlement clusters around churches, or perhaps of the even later spread of the civic ideals of metropolitan France; however, charter-evidence suggests that local churches were spatially central and community-oriented as early as the ninth century. ‘Houses clustered around them and roads radiated from them … transactions were often performed in or in front of the local church’.Footnote ¹³⁰ We may be dealing with a fundamental difference dating as far back as the late antique period. However, in any case, the corollary is that medieval and modern re-workings of the built environment are much more likely to have removed the earliest medieval evidence in Brittany than in Britain. By the same token, archaeological investigation there would be more disruptive than similar work on British churches. For the foreseeable future, it appears that students of the early Breton Church will have to do without much archaeological data relating to local churches; however, it is possible that some promising sites may be selected and investigated. At the least, the observation made for some sites which also have inscribed stones (Plourin, F; Louannec, CA; Sainte-Tréphine, CA) that the churchyard area stands up to 2 m above the surrounding ground level, gives reason to think that the sites have been in use for many centuries – or possibly that early churches might be built on artificially raised areas of ground, as the place-name-element podum implies in Wales.Footnote ¹³¹ Site-by-site observation of how church buildings fit into surrounding settlement patterns, roadways and field systems is a promising avenue of study.

In Wales and Cornwall, as stated earlier, early medieval inscribed and carved stones survive in large numbers (around 550 in Wales, over fifty in Cornwall and west Devon) and their presence can suggest an early origin for known ecclesiastical sites.Footnote ¹³² Moreover, they can assist in prospecting the landscape for as yet undiscovered burial places and religious enclosures.Footnote ¹³³ Monuments can illuminate individual churches’ history and contacts: for instance, the inscriptions revealing royal patronage and/or commemoration at Llanilltud Fawr (Llantwit Major), Morgannwg, and Llangadwaladr, Môn, and the pair of pillar crosses found at Llandochau and Llandaf.Footnote ¹³⁴ In Brittany, there are far fewer of these monuments: twenty-six inscriptions dating from the eleventh century and before (plus two from the Channel Islands), of which seven are inscribed slates found inside burials in the sixth- to seventh-century cemeteries of Ille-et-Vilaine. Their rarity – as in northern Britain and eastern Wales – may reflect the absence of the influence of Irish epigraphy.Footnote ¹³⁵ Few though they are, the inscribed stones of Brittany are a precious source of information on the early Church in general, and on their find-sites in particular. The range of scripts used reinforces the reality and longevity of Brittany’s insular connections.Footnote ¹³⁶ One example is the pillar with a cross in the parish churchyard at Lanrivoaré in Finistère. The inscription features the personal name Gallmau in Insular decorative capitals reminiscent of those in the Lindisfarne Gospels.Footnote ¹³⁷

As several inscriptions date from the seventh or eighth centuries, they help to illuminate a period of Breton history that has produced very little other evidence. They allow, or reinforce, the identification of some important ecclesiastical sites. In some cases, as at Lanrivoaré and Plourin in Finistère, toponymy and charter-evidence work together with the inscriptions to suggest the existence of early religious foci. In others an inscription is the initial clue to the early origins of a religious site, as in the case of the famous fifth- or sixth-century inscribed granite sarcophagus at Lomarec (M), where excavation revealed that the chapel built around it stands on a foundation of Gallo-Roman brick and tile.Footnote ¹³⁸ However, when all due allowance is made for losses, it seems clear that, in contrast to Cornwall and the west of Wales, inscriptions and stone sculpture were never a defining feature of the growth of Christianity and of a post-Roman elite in Brittany.

Similarly, fewer early medieval portable items connected with Christian cult survive from Brittany than from anywhere else in the Celtic world. Almost the only metalwork to survive, where Ireland, Wales and Scotland boast chalices and patens, croziers and reliquaries, consists of six bronze handbells (out of a total of 101 early medieval bells from the Celtic-speaking regions) – and even these have been thought to be of Welsh or Cornish manufacture.Footnote ¹³⁹ There is also a group of seven anthropomorphic bone figurines used as pendants, which are as yet undatable for want of comparanda.Footnote ¹⁴⁰ The only early monastery sites to have been excavated are Landévennec (definitely monastic), Île Lavret (probably) and Le Yaudet (possibly).Footnote ¹⁴¹ The findings at Landévennec prove the antiquity of a monastic foundation for which the earliest written evidence dates from the ninth century. It parallels Llandochau, Llanilltud and Bassaleg in south-east Wales in that it was sited adjacent to a large Roman villa and may possibly have made use of still-standing villa buildings in its earliest phase. In this respect it appears unique in western Brittany, although in eastern Brittany as elsewhere in Gaul there are examples of Roman buildings being reused as ecclesiastical buildings.Footnote ¹⁴² A stone oratory and two other rectangular buildings were built at Landévennec in the seventh or eighth century, before the introduction of the Benedictine Rule in 818 led to the building of an enlarged church and cloister. There is also a cemetery (containing plain-dug and wooden coffin burials) in which one of the earliest graves gave a radiocarbon date of AD 470–635, implying that the monastery was indeed founded in the generations immediately following the end of Roman Britain, as the Vita of its founder St Winwaloe implies.Footnote ¹⁴³ However, all three ‘monastery’ excavations are notable for the modesty of their remains. There is no trace of luxury objects: no evidence of craft-working, as there is at Clonmacnoise and Portmahomack, or imported pottery or glassware such as that found in the cemetery at Llandochau.

In brief, the most general contrast that can be drawn between early Christian archaeology in Brittany and those in most other parts of western Europe, apart from sheer lack of evidence, is the absence of an obvious hierarchy of sites. This is apparent in toponymy (to be more fully discussed in Chapter 6), as well as in archaeology. Almost everywhere in Insular and Continental Europe in the post-Roman centuries, it is obvious from both written sources and archaeology that ecclesiastical sites were a vital part of the topography of power. By the seventh century, various ranks of religious establishment could be distinguished physically, from major urban cathedrals and basilicas (on the Continent) and rural monasteries endowed with large estates, down to hermitages and chapels. In many parts of the Insular world, even where written source material is lacking, archaeology has allowed the identification of ‘pairings’ between high-status secular and ecclesiastical sites, revealing an early symbiosis (though not identity) between lay and ecclesiastical leadership. Examples include Burghead and Portmahomack in northern Scotland; Bamburgh and Lindisfarne in Northumbria; Dinas Powys and Llandochau in Morgannwg; and Tintagel and the church-site of St Matheriana in Cornwall.Footnote ¹⁴⁴ At one level down from episcopal and favoured royal establishments were important rural churches staffed by several priests: the baptismal churches of France, the minsters of Anglo-Saxon England and the ‘mother churches’ of Wales were all of this general type.Footnote ¹⁴⁵ They seem to correlate with secular administrative units such as pagi, hundreds and cantrefs. Below these again were estate churches and chapels. It was not until the central Middle Ages that these different levels of religious provision were fully articulated into a diocesan and parish network.Footnote ¹⁴⁶ Nevertheless, variations in wealth and political importance appear to have been present from the beginning. In Brittany, such variations can scarcely be seen. West of the river Vilaine, there was a surviving Gallo-Roman episcopal centre at Vannes and there may have been others at Dol and Alet, although Dol is archaeologically unidentified.Footnote ¹⁴⁷ For other episcopal sees, there is no evidence pre-dating the Carolingian conquest. The only church centre west of Vannes with unequivocally pre-Carolingian material evidence for its existence is Landévennec. Toponymy and later evidence – in particular, the three hundred or so ninth-century charters preserved in the cartulary of Redon, a monastery that was founded in eastern Brittany in 832 on the Carolingian Benedictine modelFootnote ¹⁴⁸ – suggest an explanation for the apparent shortage of high-status Breton churches. From an early period (before the Carolingian conquest) there was an unusually dense provision of churches serving the village communities known as plebes in Latin, ploue in Breton: territorial units smaller than English hundreds, Welsh cantrefs or Frankish pagi or conditae.Footnote ¹⁴⁹ The close association of community and church within the plou may account for much of the difference between Brittany and Celtic Britain in the physical remains of early medieval church-sites, a question which will be discussed further in Chapter 6. There were also many small monasteries (monasteria): seventeen are referred to in the charters of Redon.Footnote ¹⁵⁰ Episcopal authority over all these establishments seems to have been distant, nor were most monasteries large or well-endowed enough to gain dominance over other churches, Landévennec again being the only notable exception. The plebes of Brittany may, as Largillière put it, have been ‘little autonomous republics which, for a long time … recognised no superior authority’.Footnote ¹⁵¹

Unless further archaeological research broadens the evidence base, the conclusion must be that the shortage of material evidence for the early medieval Breton Church points in the same direction as the lack of secular evidence, and both suggest a state of society noticeably different from that of Brittany’s neighbours in either Gaul or Britain, more decentralised, with weaker elites.

Some Conclusions

The absence of archaeological evidence for Breton migration does not in itself cast doubt on the migration having taken place. The situation is quite different from an ostensibly comparable case, the Gaelic-speaking region of Dál Riata in Argyll, Scotland. Bede, along with annalists and genealogists based in the Gaelic-speaking world, wrote that the kingdom of Dál Riata had been founded through Irish immigration in the early historic period; but Argyll and north-western Ireland at this time had distinctive settlement-forms and portable artefacts which are sufficiently different as to argue (on one view) against such a population movement, while the historic evidence is far from contemporary and the language-evidence not closely datable.Footnote ¹⁵² In Brittany, the contemporary testimony of Gregory of Tours and the strong evidence for language-change puts the fact of migration beyond reasonable doubt. Why, then, would Britons moving to Brittany take with them, and perpetuate, their language and sense of identity but not the visible cultural traits that they developed in Britain itself?Footnote ¹⁵³

The difference between the archaeological profiles of western Britain and Brittany in the fifth to seventh centuries is perhaps one of degree rather than kind. If Bretons at this time are almost entirely invisible in the archaeological record, many western Britons are equally so. John Blair has recently concluded that large areas of early Anglo-Saxon England are bare of the material culture that supposedly defines Angles and Saxons; the same thing is even truer of Celtic Britons.Footnote ¹⁵⁴ Devon, for instance, lacks re-occupied hill forts and ‘late Roman’ and ‘Celtic’ metalwork, and also cist burials, while inscriptions and evidence for Mediterranean trade are confined to the fringes of the county.Footnote ¹⁵⁵ The absence of such features in Brittany might be due to the regional origin of the migrants or it might be socially determined: a version of Ben Guy’s suggestion that it was the poorer members of society who migrated to Brittany, rather than the elite. But the class argument need not be a rigid one: ‘peasants left, landlords stayed’. Perhaps, rather, surviving as a member of the post-Roman British elite or moving to Brittany were mutually exclusive strategies. You either stayed, and learned to live in a hill fort, wear a penannular brooch, fight your neighbours and trade your prisoners for wine and tableware or perish – or you went to Brittany in the hope of avoiding these necessities. Perhaps indeed the hope of maintaining a more ‘Roman’, less ‘barbarian’ identity was a factor in at least the early stages of migration.

D. H. Miller has written of the prevalence of a ‘frontier’ dynamic in what we think of as typical early medieval societies.Footnote ¹⁵⁶ In the economic slump that followed the removal of the compulsion of the Roman imperial state taxation and supply system, operating a frontier was perhaps the only (not very) reliable way of accumulating enough of an economic surplus for any kind of cultural display. In Britain, multiple frontiers rubbed up against each other, providing opportunities for the conflict and plunder that alone would generate a surplus (however ephemeral) for the winners. The old, increasingly permeable Roman frontiers of Hadrian’s Wall and the Irish Sea; tribal and factional frontiers within the former province that might or might not coincide with an ethnic demarcation, Briton against Saxon; religious frontiers, Christian against pagan and, later, ‘Roman’ against ‘Celtic’; and the old divide between the grain and cattle lands, the villa zone and the rest, along which there formed the most ruthless early Anglo-Saxon kingdom, the Mierce – men of the March, Mercians. The most successful kingdoms grew through having vulnerable ‘plunder zones’ to exploit, and thus were able to afford cultural achievements that have echoed down the ages: as Peter Brown put it, ‘the Golden Age of Northumbria … rested on much real gold’.Footnote ¹⁵⁷

Brittany, by contrast, was rarely in the front line either to plunder or to be plundered. If, to quote Peter Brown again, the Mediterranean itself after Rome’s fall was a ‘numbed extremity’ of the Eurasian land mass, then Brittany was an appendix.Footnote ¹⁵⁸ Bypassed by such trade routes as remained, it developed a low-pressure, self-sufficient economy.Footnote ¹⁵⁹ Its imports were things that cost nothing: language, script types and saints. Compared to most parts of early medieval Europe, its lack of cultural display may indicate that it was a relatively safe place to live.

But if high-status trade and high politics passed Brittany by, that does not mean it was entirely isolated from the outside world. Recently, rescue archaeology has permitted extensive excavations that have at last begun to discover early medieval settlement evidence in western Brittany. On the whole, this evidence fits with the conclusions presented here, in that it does not suggest high levels of wealth or power, the trading or manufacture of luxury objects, or objects or practices introduced directly from south-western Britain. It does, however, reveal inhabitants sharing at a modest level in the rural lifestyle and technology of neighbouring regions, insular and continental. Several sites datable to the seventh to ninth centuries in western Brittany apparently consist of individual family farms comprising dwellings and various craft activities within curvilinear ditched enclosures (contrasting with the rectilinear enclosures more commonly found in Merovingian northern Gaul).Footnote ¹⁶⁰ The presence of sunken-featured buildings is typical both of northern Gaul and early Anglo-Saxon England; corn-drying ovens, on the other hand, are well known from insular sites but within Gaul are specific to Brittany, being an adaptation to damp climates. Some higher-status sites have also been investigated, marked out by their elevated and otherwise strategic locations, large size and earthworks rather than by any more spectacular material: for instance Bresselien (F) and Leslouc’h at Plouédern (F). At Leslouc’h there is a sequence of seventh- and eighth-century pottery, made from local materials, but in forms that hark back to late Roman Dorset Black Burnished Ware and Crambeck ware, while others resemble late Anglo-Saxon ‘shelly’ ware.Footnote ¹⁶¹ Investigation of sites like these, which so far fall into the dating range of the seventh to ninth centuries, with a new departure in the tenth, turns attention away from the ‘migration period’ to the basic conditions of rural life which Brittany, to a great extent, shared with the rest of north-western Europe.