Let’s begin with a striking and much-debated case of outright rural revolt in late antique Gaul: that of the Bagaudae or Bacaudae. It was once argued by E. A. Thompson that they were peasant revolutionaries who played a role in the fall of the Roman empireFootnote ¹ – the ultimate case of historical change coming bottom up as well as top down. However, subsequent studies have substantially undermined both the significance of the Bagaudae and even scholarly confidence in what we can actually say about them.Footnote ² We first hear about the Bagaudae in northern Gaul in the 280s, and then again, after a gap of more than a century, in the first half of the fifth century, at this time also in the Ebro valley in Spain. They are mentioned – by name and otherwise – by a substantial cast of late antique authors, albeit mostly briefly.Footnote ³ It is unclear, however, if these ancient accounts actually refer to a single, identifiable phenomenon. So who were the Bagaudae? Our ancient sources refer to them variously as agrestes, agricolae and rustici/rusticani (rustics and farmers), as well as using other related rural terms, including aratores and pastores (ploughmen and shepherds) but also latrones (bandits). It seems most likely that the Bagaudae were peasants taking advantage of political and military instability, as well as landlord absenteeism, in a context of social stress. The Bagaudae were reputedly active in areas that were both geographically and politically marginal – very far from and indeed unlike Provence. Ultimately, they remained an isolated phenomenon in the west, and had no real successors.Footnote ⁴ And yet we might still wonder whether the case of the Bagaudae could hold some relevance for our understanding of peasants in southern Gaul. However vexed the case of the Bagaudae turns out to be, it might remind us that peasants could exercise agency, and that the potential for violent resistance, as well as flight, existed – even if only very rarely exercised and under very particular, highly contingent circumstances.

But let’s turn now to a much more local case study: that of the shepherds of the Crau plateau. The pastoralist was a category of rural person traditionally treated with ambivalence – at the very least – in the ancient world, often associated with nomadism, uncivilized behaviour and not infrequently unlawful activity, including brigandage – and indeed in association with the Bagaudae in 280, as we just saw.Footnote ⁵ Bishop Hilary of Arles, a remarkably robust individual, as we have already seen, was not daunted by this reputation and made a visit to the plain area of the Crau one Eastertime to baptise and teach the shepherds there.Footnote ⁶ Although large-scale sheep-rearing on the Crau had dwindled by late antiquity, there were clearly enough shepherds to merit not just the spiritual but also the economic attention of the church of Arles: in the following century Caesarius left pasturage on the Crau to his monastery in his will.Footnote ⁷

The Crau shepherds constitute an intriguing case study for several reasons. The countryside of southern Gaul was a highly variegated region, home to a range of diverse landscapes as well as types of inhabitants. Pastoralists constitute a class of non-elite rural dweller who we might consider at one end of a spectrum in terms of relative levels of autonomy. Earlier in the Roman period the Crau had hosted multiple clusters of shepherds, living in ‘bergeries’ together with their families and (according to the season) their sheep.Footnote ⁸ The fact that the clusters were built to identical plans, around a well and an oven, suggests a structure whereby individual landowners owned their own flocks of sheep and hence groups of bergeries.Footnote ⁹ From the late antique period, however, this large-scale rearing seems to have come to an end and only a few scattered ‘cabanes’ (huts or cabins) have been found,Footnote ¹⁰ although it has been suggested that shepherds also lived seasonally in caves.Footnote ¹¹ This very tangible change indicates the decline of large-scale sheep-rearing, doubtless related to changes in the wool trade. It thereby represents one of a number of notable social and economic shifts in the late antique countryside. This particular shift is clearly compounded by the role of ecclesiastical interests, with the intervention of the church of Arles visible in terms not just of spiritual but also of economic and seigneurial activity and oversight, aspects we shall pursue elsewhere in this chapter. In the case of the Crau we are able to combine both archaeological and literary testimony in a way that is not always possible, making it a neat example with which to begin this chapter.

As in the previous chapter, I am aiming to elucidate late antique popular culture as fully embedded in its local environment. I shall again make use of recent archaeological work to produce as granular as possible a picture of rural life.Footnote ¹² Of course, we cannot understand the late antique countryside in isolation: the mutual interdependence of town and country is a commonplace but nonetheless bears repetition. Cities only functioned within a broader economic context, not least as the wealth of the urban social and political elite was based on their landholdings in the countryside round about.Footnote ¹³ The interrelationship between town and country was, moreover, social and cultural, as well as economic.Footnote ¹⁴

In this chapter we shall consider some of the distinctive landscapes that made up the territorium of Arles (with some excursions outside this boundary, where they are of particular interest and relevance) (see Map 2). These will include the lagoons south of Arles – such as the Étang de Berre and the Étang de Vaccarés – the nearby Camargue and Crau (which we have just visited) and the hills of the Alpilles, as well as the coastal part of Languedoc: a strikingly diverse landscape for a relatively small geographical area. We shall examine a range of different types of settlement, examining the diverse trajectories of earlier villa sites, as well as the development of other types of settlement clusters, such as the widespread phenomenon of ‘perchement’ (the re-occupation of previous Iron Age hilltop oppida), as well as the emergence of rural Christian sites and networks. We will see both the development of brand-new enterprises and the reestablishment of previously abandoned industries.

A number of questions emerge from this developing picture that are relevant for our understanding of popular culture. How was rural society changing in this period? And how far can we understand any of these changes as emerging ‘bottom up’? How far do the substantial changes visible at villa sites represent an actual change of social organization on the ground? Can we trace a weakening of aristocratic control and rise in peasant autonomy? Can we trace a move towards more ‘horizontal’ communities at all? Can we see struggles over popular culture as an ongoing response to these social and economic changes? And how do clerical discourses, in particular, represent a response to these shifts? What kinds of Christian structures were available to rural societies and how did they impact upon communities and, specifically, the non-elite? How far can these questions be answered outside a micro-regional context? These issues will help shape the discussion in this chapter and those that follow.

Several scholars have explored related questions in ways that offer stimulating and suggestive models to think with. Reference DosseyLeslie Dossey’s 2010 book, Peasant and Empire in Christian North Africa, raises interesting questions and comparisons for my study. Dossey’s study of the North African countryside seeks to restore agency to the non-elite and does so to a large extent by examining their material culture. Dossey looks at the previously underestimated prosperity of the countryside of the region and writes: ‘The archaeological record of late Roman North Africa represents thousands of small acts of entrepreneurship and aspiration, undertaken by people unable to communicate to us in any other way.’Footnote ¹⁵ Despite the famous case of the Donatist schism, and indeed the much-discussed violence of the so-called Circumcellions, often understood in socio-economic terms,Footnote ¹⁶ Dossey insists that ‘what really disrupted power structures in the countryside was not opposition movements against the Roman government or reactionary efforts to return to traditional cultural and economic norms, but rather a desire by rural populations to participate in the same material culture, community structures, and intellectual currents as were affecting urban society’.Footnote ¹⁷ She sees non-elite communities as aspiring to and sometimes achieving a level of self-governance, and benefiting from the fragmentation of the elite,Footnote ¹⁸ going so far as to argue that the ‘end result was a weakening of the social and political controls that had previously kept peasants in their places’.Footnote ¹⁹ This does seem to offer some new insights into and suggestions as to how we might interpret some of the processes visible in the southern Gallic countryside.

Equally of interest, Reference GreyCam Grey’s 2011 book, Constructing Communities in the Late Roman Countryside, likewise works to attribute agency to rural non-elites in a context of changing economic, social and political conditions. Grey refuses to see peasants as an undifferentiated and downtrodden mass and provides a fascinating examination of rural communities and their role in the broader social economic structures. We see that peasants and their communities developed their own strategies for dealing with the problems and situations they found themselves in, and indeed with the power structures above them. Through a series of close readings, Grey argues, ‘they emerge from our sources as conscious actors, both individually and collectively, whose decisions and behaviour could impact upon the more powerful individuals with whom they came into regular contact’.Footnote ²⁰ In this way, Grey’s work provides another helpful framework for my own project. I shall discuss ‘The Roman Peasant Project’, directed by Kim Bowes together with Cam Grey, in more detail later, but it provides an exceptional example of how archaeology can illuminate the lives of ancient peasants beyond the still prevalent focus on the villa.Footnote ²¹

My aim is to build up a more rounded picture of rural society and culture in southern Gaul in order to make sense of aspects of popular culture which will be examined in more detail later on, such as the ‘lived religion’ explored in Chapter 5 and the festival of the Kalends in Chapter 6. In this chapter I shall explore the fundamental elements of non-elite rural life, including evidence for social organization and changing patterns thereof, as well as the various types of economic activity that determined the rural livelihood. These included not just crop cultivation and livestock-rearing but also a whole range of activities including viticulture, oil and salt production, fishing and various types of artisanal activity. I shall also consider the material culture of the countryside, especially as evidenced through pottery finds. Finally, I will look briefly at the key features of the religious landscape (to be returned to in more detail in Chapter 5). In this way, a picture of rural life in southern Gaul will be built up that provides a pair to Chapter 2 and a foundation for what is to come.

The Rural Environment: Landscape and Micro-Regions

It is both difficult and inadvisable to generalize too far when discussing the late antique countryside. My aim is to investigate late antique settlements in their local, micro-regional contexts, as well as in the context of broader late antique trends. This micro-regional focus has several methodological advantages. First, it favours an interdisciplinary approach, enabling us properly to see the interaction of nature, culture and society in a given environment.Footnote ²² Next, for Peregrine Horden and Nicholas Purcell, as expressed in their hugely influential 2000 book The Corrupting Sea, micro-regionality is an essential characteristic of the Mediterranean.Footnote ²³ However, for Chris Wickham, it is a particularly post-Roman characteristic.Footnote ²⁴ The methodological challenge for the historian, as others have noted, is how to link the micro-picture to the bigger one,Footnote ²⁵ and I shall return to this challenge later on.

We are dealing with a highly varied landscape (see Map 2). The territorium of Arles includes some diverse and distinctive zones. Closest to Arles itself, and still very much part of the cultural identity of the city even today, lies the delta region of the Camargue where the large and disconcertingly flat vista seems to stretch interminably.Footnote ²⁶ Today’s visitors find an area that is carefully managed and preserved, including several distinctive features (such as the presence of flamingos) and activities, some of which postdate the period at hand (including rice-growing and the rearing of horses and bulls). The north of the Camargue, in particular, is highly fertile and most ancient settlements were grouped around the banks of the various branches of the Rhône. To the south is an area of saltwater lagoons and marshes; salt production indeed made up a crucial feature of the region’s economy in antiquity and beyond.

The rocky plain of the Crau, a roughly triangular-shaped (though Strabo thought it was circular, as we shall see) steppe of some 55,000 ha formed out of one of the beds of the river Durance, constitutes another distinctive, and indeed not entirely prepossessing, landscape to the east and south of Arles. The area is referred to as ‘Stony Plain/Field’ throughout antiquity;Footnote ²⁷ Strabo describes it as follows:

The Crau was rocky, arid and windy territory and so it is not surprising that there are no traces of villas in the landscape; as noted by Pliny, the area favoured sheep-rearing beyond all other types of agriculture, and we have indeed already ‘met’ the local shepherds.Footnote ²⁹

The northern Crau borders the hill range of the Alpilles, which provides a highly picturesque backdrop to the site of Glanum (itself abandoned after the third century ce). This area was traversed by the important Via Domitia, which connected Spain to Italy, and unsurprisingly therefore contained many villas, some of which saw continued occupation into the fifth century.Footnote ³⁰ The Alpilles also provide a literary backdrop to several anecdotes in the Life of Caesarius, which, as we shall see elsewhere, likes to depict the bishop paying frequent visits to the various rural areas of his territorium. One of Caesarius’ miracles took place in a large aristocratic domaine in the Alpilles, the productivity of which had been seriously damaged by terrible hailstorms until the episcopal baculus of Caesarius miraculously saved the day.Footnote ³¹

In the south of the territorium lies a different kind of habitat again, characterized by inlets and lagoons (known as étangs), as well as wooded, hilly areas. The micro-region of the Étang de Berre and its environs during antiquity have been brought to life by the archaeologist Frédéric Trément (see Map 6).Footnote ³² The area has provided material testifying to a wide range of types of occupation: villas, hamlets, artisanal and other productive sites; it is also the location of two significant examples of late antique reoccupations of earlier oppida at Saint-Blaise and Constantine (to be discussed later). The archaeological evidence testifies to a high population level in the region during the fifth and sixth centuries. As with other areas, this is clearly a micro-region which benefited from the prosperity of Arles and Marseilles, although at other times it also seemed to suffer from tensions between the two.

The final area of especial interest is that of south-eastern Languedoc, parts of which at times belonged to the territory of Arles. This area has benefited from the archaeological work associated with the TGV extension and thus provides unparalleled material with which to build up a micro-level picture.Footnote ³³ It contains a number of important villa sites (including the villa of Près-Bas at Loupian) and shows signs of new economic developments in our period, often in connection with the important coastal economic sector. Having sketched out the different landscapes of the region, I shall now seek to describe their occupants.

Who Lived in the Countryside? Defining the Inhabitants of the Late Antique Countryside

How close can we hope to get to our rural dwellers? It is famously hard to access the rural non-elite of antiquity (and indeed beyond). Cam Grey has aptly noted the tendency in the historical literary sources for ‘the focus of attention to slip upwards from rustici to the lowest echelons of our aristocratic authors’ own social stratum’ – as well as a similar tendency among recent historians.Footnote ³⁴ The texts through which we have sought to understand the conditions of the non-elite are of course far from transparent, a striking case in point being the work of Salvian of Marseille, who wrote around the middle of the fifth century. Salvian gives a much-discussed picture of the impoverishment and exploitation of poor tenants, largely due to their oppression by taxation, claiming that free men were being forced into becoming coloni.Footnote ³⁵ But how poor are Salvian’s ‘poor’? The reliability of Salvian as a witness has come under increasing attack in recent scholarship: his picture of the late antique countryside is rhetorically and ideologically shaped by the larger demands of his argument – his need to explain the misfortunes of the era – which leads him to find villains in Roman power and ‘the rich’.Footnote ³⁶ Salvian’s pauperes, as Cam Grey has written, ‘seem to function as a sliding point of comparison with the rich’,Footnote ³⁷ with Salvian able to describe himself too as a pauper when it suits him.Footnote ³⁸ Moreover, it is clear from the examples he gives that the oppressed pauperes the writer has in mind are often actually landowners, albeit ‘lesser’ ones. Caesarius also discusses the plight of such smaller landowners, suffering at the hands of their richer and more influential neighboursFootnote ³⁹ – undoubtedly a widespread phenomenon but one that takes us beyond the truly subaltern of the late antique countryside.

We need to use the work of archaeologists – supplemented at times by comparative and anthropological scholarship – alongside the familiar textual sources if we are to have a hope of accessing the rural non-elite. Recently, ‘The Roman Peasant Project’ has represented an important step forward towards our understanding of peasant sites and lifestyles. The project’s detailed analysis of sites in central Italy during the Roman period seeks to elucidate the lives and material culture of the rural poor as never before.Footnote ⁴⁰ While Bowes and Grey stress the importance of giving primacy to the historical particularity of the peasant experience, their study nonetheless highlights relevant points of comparison and insight. For instance, they stress the evidence for rural exchange and mobility, the lack of a distinctive ‘peasant’ material culture, and the significance of temporality rather than timelessness in our understanding of the ancient peasant.Footnote ⁴¹ While the nature and extent of the ‘Peasant Project’ is unique, the excavation and publication of non-elite sites in our own region, including lagoon-side hamlets, shepherds’ huts and potteries, provide a striking body of material for this chapter.

Many, but by no means all, of our rural dwellers can be defined as ‘peasants’. In line with the approach taken by both Kim Bowes and Chris Wickham, it is most helpful to apply a broad, ‘maximalist’ definition. For Bowes, as well as smallholders, peasants were ‘individuals who could, in other contexts, be described as tenants, wage labourers, seasonal labourers, slaves, and rural craftspeople’. Wickham observes further that peasants themselves could also hire tenants and labourers.Footnote ⁴² Furthermore, the non-elites we are looking at actually comprise more than agriculturalists; they include fishermen, food producers and artisans, and those engaged in a whole range of other economic activities. Next, when looking at popular culture in the countryside, our ‘non-elite’ by necessity include not just those clearly at the bottom of the pile but also those in the upper echelons of peasant society. There is, nonetheless, clearly a point when such figures cease to be ‘peasants’ or subaltern – such as, as Wickham suggests, when they own so much land that they no longer need to farm it themselves and now become, rather, ‘medium’ landowners.Footnote ⁴³ All peasant societies, he reminds us, contained ‘a kaleidoscope of differentiation’.Footnote ⁴⁴

What about the legal status of our rural dwellers? The complexities of Roman agricultural production were such that it was entirely typical that ‘tenancy, slavery, wage labour, and thus coloni, free peasants and slaves coexisted in the same countryside’, as Kyle Harper has put it.Footnote ⁴⁵ That is, our rural non-elite individuals could be slaves, coloni or free peasants, without our generally being any the wiser – and these categories certainly do not tell the whole story about the people so defined in any case. The much-cited Testament of Remigius, the will of Remigius, bishop of Reims (in northern Gaul), who died in 533, is a useful case in point.Footnote ⁴⁶ Remigius bequeathed a number of individuals (about a hundred in total), together with his estates; some are described as slaves (servi, ancillae, pueri), others are called coloni, and others nothing at all: but all are disposed of in his will, just the same.Footnote ⁴⁷

That many slaves worked in the countryside of southern Gaul during our period is clear.Footnote ⁴⁸ For instance, as we saw in Chapter 2, the corpus of Caesarius throws up a number of (often negative) references to slaves.Footnote ⁴⁹ The sermons do not give us much information about what work slaves actually did in the countryside, though slaves do appear in one sermon as vineyard workers.Footnote ⁵⁰ Preaching against abortion, Caesarius says that no free woman would allow her slave girl to abort her child lest the landowner lose the profit that could be made through the offspring.Footnote ⁵¹ More generally, as we saw in Chapter 2, slaves appear in our sources as members of the household, belonging to members of his congregation, who could subject them to corporal punishment.Footnote ⁵² The authors of his Vita think it worth noting that Caesarius took care that his slaves should receive no more than the canonical thirty-nine lashes by way of corporal punishment – unless their crime had been especially serious, in which case several more blows could be bestowed, after an appropriate interval of a few days.Footnote ⁵³ Of particular interest here is the fact that the passage refers not only to slaves but also to dependent free men (ingenuis obsequentibus sibi). This reminds us once more that legal status was far from absolute in defining the status and condition of those at the bottom of the social hierarchy and that the slave/free binary was far from absolute in late antiquity.

This inevitably brings us to the issue of the colonate – the subject of a perhaps disproportionate amount of scholarly attention in the past. The existence of coloni – still most easily defined as ‘tied labourers’ – is not in doubt, but the nature and significance of the status of colonus continues to inspire debate.Footnote ⁵⁴ In practice, quite apart from doubts about the actual de facto differences between legal statuses, it is often hard to deduce from our evidence which type of status we are dealing with (as we saw in the case of the will of Remigius). What is most notable is the susceptibility in late antiquity not just of slaves and coloni but also of supposedly free peasants to coercion, just as the example from the Life of Caesarius showed. Caesarius’ own words also make clear the physical realities of such coercion: late Roman law prescribed the use of manacles, shackles and fetters – such as have been found in Roman villas – on fleeing coloni, and the bishop of Arles urged his audience to use them on recalcitrant, idol-worshipping tenants.Footnote ⁵⁵

Accessing peasants as historical individuals is – sometimes – just about possible. In late antique and early medieval wills there are some named persons of securely non-elite status. The will of Remigius lists no fewer than eighty named individuals (giving the reader something of a small historical thrill), comprising individuals and families of the lowest social status, who were passed on as part of the bishop’s various estates.Footnote ⁵⁶ For instance, there is the vinitor Enias, his wife Muta, their daughter Nifastes and son Monulfus.Footnote ⁵⁷ The overall picture is one of small, nuclear families engaged in a mixture of broadly agricultural activities.Footnote ⁵⁸ This discussion of families raises the question of rural women specifically: accessing women in the ancient countryside is of course doubly difficult, and late antiquity is no different. It is probable that women worked in the countryside of southern Gaul as both slave and wage labour, as well as being part of peasant households.Footnote ⁵⁹ Literary texts are unsurprisingly uninformative when it comes to women’s labour, although female work at the loom is well known.Footnote ⁶⁰ The late antique agronomist Palladius provides grudging testimony as to the suitability of women for rearing hens as well as for collecting and storing acorns (this together with children).Footnote ⁶¹ The presence of women at rural sites can be archaeologically attested through the identification of particular artefacts: for instance, women seem to have been present in the earlier ‘bergeries’ on the Crau.Footnote ⁶² Bowes and Grey cite the presence of pins, loom weights and other small finds among the pottery excavated at Marzuolo, southern Tuscany as evidence of a broader pattern whereby women worked in potteries when seasonal pressures affected labour supply, ‘slotting in ceramic tasks round child rearing’.Footnote ⁶³ Scholars in the past all too often envisaged a productive landscape devoid of women but this was a false picture; more generally, we need to be conscious of how changing social, economic patterns could have impacted on women, and on their role in popular culture more specifically.

Finally, it is worth considering, as we considered for urban contexts in Chapter 2, whether we can identify any cultural or indeed ethnic diversity among our rural non-elites. Once more, the question of language arises. A good deal of work has been done tracing the presence of Celtic language (largely ‘Gaulish’), in addition to Greek, in the region, as well as the interaction of both these languages with each other and with Latin.Footnote ⁶⁴ While Greek was at issue in the cities as we saw, in the countryside the question is rather one of Gaulish: was it still spoken in the countryside of southern Gaul in late antiquity?Footnote ⁶⁵ The ‘general consensus’ agrees that Gaulish was moribund ‘by the fifth, or perhaps sixth century ce, though it may well have met its end earlier in the South’.Footnote ⁶⁶ Certainly any testimonies that could possibly point to the continuance of spoken Gaulish in our period are ambiguous at best.Footnote ⁶⁷ While any positive evidence for the continuance of Celtic languages among the non-elite is missing, perhaps most importantly we can tellingly contrast Augustine’s concern to find Punic-speaking clergy in North Africa with the total lack of comparable evidence from Gaul of pastoral concern regarding non-Latin speakers: there was clearly no linguistic barrier in operation.Footnote ⁶⁸

What about other elements of linguistic or cultural diversity? While scholars have understandably sought to find the early history of the Jews in France in urban settings, as we saw in the previous chapter, chance findings of examples of Jewish material culture in the countryside suggest a more complex, indeed intriguing picture. Both examples take the form of clay lamps, cheap and portable items, displaying menorah designs. The first was found in a ravine close to Orgon in the Alpilles and is dated to as early as the late first century ce.Footnote ⁶⁹ The second find hails from a securely late antique context, the late occupation of the oppidum of Lombren in the Hérault (discussed later), most likely dating from the late fourth/early fifth century.Footnote ⁷⁰ The case of this second lamp is particularly intriguing in terms of what we might call object biography. The lamp was found in a potters’ workshop but was itself a North African import: was it being used (in part?) as a model for local production?Footnote ⁷¹ It must have been seen as valuable: Although the disk of the lamp was broken, this did not stop its owner from using it. It would be overly and unnecessarily reductive to assume that the presence of ‘Jewish’ lamps equals the presence of Jewish rural inhabitants: an unproblematic identification in this way cannot and should not be made. Nevertheless, the presence of the two lamps remains of interest.

Having introduced the inhabitants of the region, we shall now look in more detail at the habitats and forms of social organization in which they lived, which provided the granular contexts for the development of forms of popular culture.

Habitats and Social Organization

We shall now consider how these different forms developed and changed in late antiquity. These changing forms provide the context for developments in popular culture: we are looking for forms of social and economic organization that might represent a less hierarchical form; specifically, a move away from the classical type of villa organization. As summarized by Simon Esmonde Cleary, the ‘exceptional suite of data’ for south-eastern Gaul enables us to see ‘profound changes’ by the turn of the fifth/sixth century. These included the abandonment of the villa as an aristocratic residence and its replacement with other types of settlement ‘quite probably reflecting unrecoverable changes in tenurial and productive relations, along with a substantial flattening of the social hierarchy’.Footnote ⁷² Although, as we shall see, this represents something of an oversimplification of a highly diverse pattern, this arc still provides a useful working hypothesis to explore further.

The villa, both as aristocratic residence and centre of production, was of course a key feature of the Roman countryside. As Chris Wickham has remarked, while villas varied greatly in scale, their broad homogeneity was itself ‘a marker of the homogeneity of Roman culture’.Footnote ⁷³ The villa stood at the centre, and the summit, of rural settlement, and rural settlement hierarchies.Footnote ⁷⁴ It also functioned as a crucial location for the display of aristocratic status and prestige and this is the aspect of the villa that undergoes the most drastic change during our period. Christophe Pellecuer’s study of seventy-one villa sites from across Narbonensis, built between the republican period and the beginning of the second century ce, establishes that while 50 per cent of sites were still occupied during the fifth century, only 20 per cent had occupation in the sixth century, with as few as five examples surviving after 600.Footnote ⁷⁵ However, Pellecuer and Pomarèdes’ study of western Narbonensis (modern Languedoc), making use of recent archaeological data, gives a rather different picture. Almost three-quarters of republican/early imperial villas were still occupied during the fifth century; even though a decline followed in the sixth, this was less extreme: 42 per cent were still occupied, and nearly a quarter in the seventh century.Footnote ⁷⁶ In the area of the Alpilles, however, a well-connected region with many villas, their survival rate into our period is much lower, with only about one-quarter surviving into the fifth century.Footnote ⁷⁷ The notable chronological differences here are of course in part down to the availability of relevant excavation publication, no doubt, but are clearly also down to regional and micro-regional diversity.

Regional differences are indeed clearly marked: neighbouring Aquitaine, as revealed by Catherine Balmelle’s study, provides a striking point of comparison with Provence. Here the prestige domestic buildings of the aristocratic villa show continuing architectural development and embellishment during the fifth century.Footnote ⁷⁸ It is indeed not a coincidence that the picture of the vitality of traditional aristocratic villa-based life in late antiquity found in the letters of Sidonius Apollinaris is largely (if not uniquely) located in Aquitaine. On several occasions we find Sidonius nagging his friends for spending too much time in their villas, while visits to friends’ villas still seem to make up a substantial part of the aristocratic lifestyle. For instance, there is a nice description of two neighbouring villas in the Gard region, whose owners (one of whom was Sidonius’ own uncle) are both (re?)constructing baths.Footnote ⁷⁹ Some caution is required: we have already seen the difficulties of using Sidonius as a ‘source’ in Chapter 2. Sidonius’ use of classical topoi (and in particular here his debt to Pliny’s villa letters) is well known, but it is nevertheless hard to imagine his accounts as entirely fictional.Footnote ⁸⁰ Nonetheless, while these few examples are striking, they cannot be taken as representative, certainly not outside Aquitaine. The archaeological evidence in Provence indeed represents a different picture: most excavated villas demonstrate a rather different type of occupation at this time.Footnote ⁸¹ Overall, we can see what we might call a broad downturn in ‘traditional’ aristocratic villa occupation in our region in late antiquity.

Some villas were still being renovated and indeed embellished up to the early fifth century, the most striking example being the villa at Près-Bas, Loupian, near the Bassin de Thau, in the Hérault in Languedoc (see Map 2), which has been extensively excavated and is very well presented for both scholars and the general public.Footnote ⁸² This villa, on a site of 3 ha, underwent a series of developments from the middle of the first century bce onwards, but the most striking modifications were made in late antiquity (in the mode of the Aquitanian examples). During the middle of the fourth century new residential buildings were developed on a grand scale, involving the building of monumentally large and high-ceilinged apsed rooms – including a reception room of 85 square metres. These buildings were then decorated with a series of mosaics in the early fifth century. Thereafter we can see occupation of a more functional nature, as seen elsewhere in our region: developments include the covering over of the mosaics and the installation of a hearth in the baths. At the same time, the coastal port area of Bourbou associated with the villa was redeveloped and a church was built at some distance from the main villa buildings, both to be discussed further later. There are a few (scant) comparable examples: baths at the villa site of Saint-Pierre in Eyguières in the Alpilles were refitted at the end of the fourth/early fifth century, exceptionally for this micro-region.Footnote ⁸³ Moving later, the villa at Saint-Pierre-de-Pabiran in Languedoc was still occupied throughout most of the sixth century and only saw a real change towards its end, although the villa was still ‘frequented’ even in the seventh and eighth centuries.Footnote ⁸⁴ In general, we can conclude that where occupation of villa sites continued throughout the fifth and sixth centuries, it was an occupation less focused on high-prestige lifestyle activities: areas of occupation were often reducedFootnote ⁸⁵ and in most cases villa baths ceased to function.Footnote ⁸⁶ Nonetheless, and this is a point which should not be ignored, despite often substantial changes, we tend to see the continued presence of imported goods and coins, indicating continuing participation in networks of exchange.Footnote ⁸⁷

What do we mean when we talk about new forms of occupation in late antiquity? ‘Re-use’ of residential buildings for artisanal purposes is commonly found during the period, but the nature and significance of this activity is often hard to interpret. For instance, the villa site at Barrou at Sète in Languedoc was occupied intensively, without interruption, up to the sixth century. The late antique occupation was of a markedly different nature from that of the villa’s earlier period, however. Tanks for salting and making garum were placed on the mosaic pavements of the residential area as early as the third century, while other rooms were simply backfilled.Footnote ⁸⁸ Who made these alterations and who was doing the salting? Are we to imagine that an absentee landlord had had his villa remodelled for commercial and productive activity? Or had the villa been abandoned and taken over by local farmers and fishermen? One type of occupation, frequently (but unhelpfully) described as ‘squatting’ (as in the cities), is visible at a number of sites, such as the villa at Pataran in the Gard. This villa had enjoyed a luxurious redevelopment in the third century but was abandoned early in the fifth century and was then partially re-occupied by hut-style dwellings during the fifth and sixth centuries.Footnote ⁸⁹ At both Barrou and Pataran we can clearly identify a change of use, and can with all probability posit the end of the villa as a site for aristocratic residence and prestige. However, this transformation does not necessarily mean that the land associated with the villa had changed hands – only that the ways in which the estate was used by its owner might have changed, as we shall discuss further.

Finally, the presence of chapels/churches and cemeteries at, or at least in clear relation to, villa sites in late antiquity forms another distinctive phenomenon, albeit one that is sometimes difficult to interpret effectively. More precisely, it is often unclear whether the presence of a late antique church or chapel represents a new use of the villa, one related to its abandonment as an aristocratic residence, or whether it represents an initiative of the (continuing) owners, so that the church in question should be understood as a villa chapel, such as I shall discuss in more detail in the later section on Religious Structures and Landscapes. Saint-Laurent de Carbardel at Pélissane in Bouches-du-Rhône does present a clear case where the villa had been abandoned as an aristocratic residence for some considerable time before the construction of a chapel at the start of the sixth century.Footnote ⁹⁰ The site of Près-Bas provides an interesting example once more: we can see new forms of economic activity at the port site of Bourbou, as well as the building of a church (Ste. Cécile) at Loupian.Footnote ⁹¹ Church building on villa sites was thus part of a broader trend, whereby new patterns of activity appear on the margins of villa settlement as the residential occupation of the villa ends – at least sometimes marking the role of the church as landowner.

I have discussed the developments in villa occupation at some length here because the changing nature of the villa is so important when it comes to understanding larger patterns of settlement and settlement hierarchy, which in turn has significant implications for the economic, social and cultural life of the rural non-elite. Whether or not villas continued to act as traditional rural centres, with continued local aristocratic dominance, is a central question when it comes to understanding the social and economic transformations of late antiquity. And yet, as we saw, it remains incredibly difficult to establish this from the archaeological evidence and the nature of landowning and land-use is much debated. Clearly, not all aristocrats left the countryside. For one thing, as we shall see, Caesarius’ sermons certainly tend to presuppose an audience of more or less traditional domini, as do many of our other literary and indeed legal sources. So how should we interpret the archaeological evidence that shows so many of our villas being used in such a different way? Tamara Lewit has argued that rather than assume abandonment as the reason for the change of nature of occupation, we should instead consider that what we can observe is actually landowners ‘living in a different, non-classical style’.Footnote ⁹² Just as elite lifestyle was changing in the cities, Lewit argues, it was also changing in the countryside: that is, both the traditionally maintained urban centre and the private villa were becoming ‘socially irrelevant’.Footnote ⁹³ In response, Kim Bowes and Adam Gutteridge have argued that such an analysis is flawed: the changes going on were more wide-ranging and more complex. They suggest that we should go beyond assuming the presence of the elite as an ‘a priori category’. They also suggest that we should instead understand villa sites as occupied by ‘agglomerations of élites and non-élites living together in multi-family groups’; this is fascinating, but detailed evidence remains rather lacking, making such an interpretation somewhat speculative.Footnote ⁹⁴

In any case, the villa was certainly not the only form of social organization or indeed of settlement in the late antique countryside, despite its dominance in the scholarly picture. We can and should consider a whole continuum of types of small rural settlements, under a range of definitions, including scattered (‘dispersed’), individual settlements, settlements which French scholars call ‘polynuclear’, and ‘secondary’ settlements, but also those described, perhaps in a manner more familiar to many, as ‘hamlets’ and vici. The surviving textual sources from late antiquity do not, however, specify quite such a large range. The Council of Riez in 439 talks about parish churches in castellis ac vicis,Footnote ⁹⁵ two terms that frequently appear together. We should try to avoid imposing precise and likely anachronistic definitions of these types of settlement: the term castellum certainly implies an elevated settlement, but whether it also implies an extensive degree of fortification is less clear.Footnote ⁹⁶ The problems in identifying and classifying settlement types are of course due in large part to the difficulty of interpreting the archaeological remains, but are also down to broader problems of terminology as well as of methodology. Laurent Schneider commented that whether we call these small settlements vici, villages, ‘protovillages’ or hamlets in large part depends on whether we are ancient or medieval historians, archaeologists or indeed geographers – also perhaps on what particular line of historical trajectory we favour.Footnote ⁹⁷

Perhaps the most distinctive and most discussed aspect of changing occupation patterns in late antique southern Gaul is the re-occupation of earlier, Gallic hilltop oppida sites, so-called perchement. This striking development, once understood simply as a reaction to crisis and uncertainty, is now understood as part of a broader shift in patterns and networks of occupation as well as of changes in social and economic organization.Footnote ⁹⁸ I shall briefly consider two of the most striking cases – Lombren in the Gard (Languedoc) and Saint-Blaise/Ugium in the Berre region of Provence – which together demonstrate the different ends of the scale: the site of the former covers about one-quarter of a hectare, while the latter covers nearly six hectares.Footnote ⁹⁹

Lombren at first glance at least seems to meet the ‘response to crisis’ interpretation quite neatly: the re-occupation of this limestone-topped oppidum began towards the end of the fourth century, and the site was abandoned before the middle of the sixth century. Charmasson, the excavator, stressed the difficulty of access to the site and saw it as a place of refuge from the barbarians, imagining a population of about 200 inhabitants. Today, however, the typicality of such upland occupations is accepted without need for a ‘crisis’ explanation. The occupation at Lombren was sheltered behind a dry-stone rampart, with streets of sixty or so modest dwellings within.Footnote ¹⁰⁰ There is evidence of artisanal activity, including pottery, metal work and possibly glass making.Footnote ¹⁰¹ The material culture of the site was described by Charmasson as indicative of a modest lifestyle and standard of living but he himself demonstrated that this material culture included imported decorated North African ceramics and some other exceptional items, which seems rather to contradict his own interpretation.Footnote ¹⁰²

The site of Saint-Blaise/UgiumFootnote ¹⁰³ at Saint Mitre-les-Remparts in the Berre region, still highly impressive to visit today, represents a much more extensive, developed and socially diverse re-occupation.Footnote ¹⁰⁴ It is a fortified oppidum, originating as early as the seventh century bce, occupied up to 120 bce and then re-occupied, after a considerable break (probably not total), from the fifth to seventh centuries ce. This was a substantial re-occupation: while the Hellenistic walls can still be seen at the site today, a new rampart, with ten towers, was built during late antiquity. Successive campaigns of excavation have identified various phases in the development of the site, showing periods of intensive activity, followed by apparent breaks and crises, with final abandonment occurring at a time hard to pinpoint, perhaps around 600. Excavation of the late antique occupation shows substantial artisanal activity, while the finds also include a fair quantity of imported ceramics. Apart from the walls, the most impressive aspect of the late antique settlement is a church (‘Saint-Pierre I’) built in a prominent site near the main entrance to the Hellenistic fortifications. The church’s features included a portico, an apse and a presbyterium, with evidence of polychrome mosaics and a raised solea.Footnote ¹⁰⁵ A second church (‘Saint-Pierre II’) was built outside the walls, but is less well understood, thanks to many later phases of construction.Footnote ¹⁰⁶ The existence of two churches and one of the largest known cemeteries in Provence from this period suggests the importance of this settlement during the fifth and sixth centuries.Footnote ¹⁰⁷ Indeed, we should probably see Saint-Blaise as a small town at this point. The presence of the churches here, paralleled with chapels/churches at other oppidum sites, including nearby ConstantineFootnote ¹⁰⁸ and Saint-Estève, in the Var,Footnote ¹⁰⁹ is again indicative of how broader social and cultural forces were at work in forging this new settlement pattern. Clearly, the redevelopment of Saint-Blaise, and other related sites, cannot simply be linked to a brief crisis; rather, this development must be understood in the context of broader social and economic change. An economic impetus for the redevelopment, connected, for instance, to the nearby saltworks (to be discussed further later), is quite possible but need not constitute a sole cause.Footnote ¹¹⁰ Links to political and administrative changes, in terms of both secular and ecclesiastical organization, are also likely to be significant and we shall return to these later on.Footnote ¹¹¹

Finally, the area of the Étang de Berre and its environs, brilliantly elucidated by Frédéric Trément, constitutes a fascinating micro-regional case study (see Map 6). We cannot, of course, fully understand this area in isolation but need rather to look in terms of the relationship between such rural areas such as these and nearby urban centres. This micro-region had smaller centres, in Saint-Blaise, and the less-well-known port town of Fos: the changing fortunes of Arles and tensions between Arles and Marseille as well as their relative prosperity would have impacted upon the Berre area, including Saint-Blaise.Footnote ¹¹² Shifts in the relationship between city and countryside, in the case of western lower Provence, do seem to indicate what Trément has described as ‘un mouvement de reconquête agraire’.Footnote ¹¹³ As many as thirty-one new sites appeared here during the fifth century, mostly after 450. Meanwhile, Sebastian Heath points to the number of imported goods in the area and suggests that what we see happening in the middle of the fifth century is new consumers being brought into the countryside.Footnote ¹¹⁴ We can certainly posit a population increase in this micro-region in late antiquity: the archaeological evidence suggests pressure on resources, leading to attempts at both diversification and intensification.Footnote ¹¹⁵ Trément identified a number of small establishments clustering around the shores of the lagoons and convincingly argues that small communities needed to come together like this in order to make a living in a milieu he describes as ‘difficult’ and even ‘repulsive’.Footnote ¹¹⁶ We are not talking here about a settlement hierarchy centred on a villa but rather an interesting possible model of a less hierarchical and more horizontal form of social organization, which would pose a suggestive context for late antique popular culture. Thinking about the presence of imported goods, and these communities as made up of consumers, rather than thinking in terms of pure subsistence, also provides an interesting angle for considering the development of popular culture, suggesting affinities with the model presented by Leslie Dossey for North Africa discussed earlier. Indeed, Sebastian Heath’s cataloguing and analysis of ceramic finds across a wide range of sites in southern Gaul enable us to see not just survival but even new growth in consumption, of imported as well as local products, including wine and oil. With this archaeological evidence in mind, Heath permits us to read Caesarius’ preaching, which attacks rustici for having lengthy bibulous parties, less sceptically than we might otherwise – a theme to which we will return in the following chapter.Footnote ¹¹⁷

Next we shall turn to look in more detail at the diverse built environments and living conditions of our rural dwellers. Pottery is obviously the most abundant and best-studied element of this material culture. Pottery finds range from coarse ware, through the local ‘dérivées des sigillées paléochrétiennes’ (DSP), to imported fine ware. Most communities represented in the finds used a combination of these types, as in the case of the inhabitants of the agricultural site at Soires studied by Trément. The buildings at Soires (in the Berre micro-region) were made of earth and dry-stone; their inhabitants used predominantly local crockery but also had access to imported ceramics and amphorae, demonstrating their participation in commercial exchange.Footnote ¹¹⁸ A similarly diverse material culture is to be found at the late antique oppida sites: while some have sanctuaries and fortifications and ‘soigné’ building constructions, they also feature dry-stone buildings, including the ubiquitous ‘cabanes’.Footnote ¹¹⁹ Finds from Saint-Blaise include plenty of imported artefacts, including North African imports, which continued throughout the entire sixth century.Footnote ¹²⁰ The finds at Lombren, meanwhile, as we saw, include items of jewellery (pendants, a hair pin, a bronze bracelet and rings) as well as North African red slip ware.Footnote ¹²¹

Pottery finds are indicative of (changes in) local eating habits and dining customs. For instance, at the high-level occupation at Saint-Estève in the Var, J. P. Brun notes the presence of many more bowls than plates or other crockery, seemingly representing a change in dining practices.Footnote ¹²² The relative use of local versus imported pottery can likewise be linked to more than just the accessibility of each type: DSP and ARS were often used together, with distinct repertoires, in terms of both morphology and decoration. Local makers and local consumers chose distinctive vessel types and decorative motifs.Footnote ¹²³ These finds also of course tell us about changing patterns of exchange: most of the crockery at Saint-Estève is very local, consisting of DSP from the lower Rhône valley.Footnote ¹²⁴ However, not all the ceramics were so local: experts have managed to trace the works of the local workshops used across several elevated sites and we can see how the same DSP products were used at Saint-Propice as at Saint-Blaise, and indeed Marseille.Footnote ¹²⁵ Nonetheless, a substantial transformation in exchange takes place in the Rhône valley in the course of the sixth century, with a progressive withdrawal of long-range exchange.Footnote ¹²⁶ In Chapter 6 I shall consider some of this material culture – the pottery – in more detail, in order to consider its role in the construction of popular culture. For now, we shall turn to the material production of our non-elite rural dwellers themselves.

Making a Living

My discussion of settlement in Provence and Languedoc has already demonstrated that our non-elites were involved in a wide range of activities, extending well beyond the cultivation of the Mediterranean triad, including fishing and salt production, animal-rearing and artisanal work of various kinds. We can also find pockets of dense polycultural activity, with all or most of these activities taking place in close proximity.

The production of oil and wine on a large commercial scale and for export had more or less disappeared even before our period began. Small-scale production can be hard to pinpoint archaeologically, while it is also difficult to distinguish the manufacture of wine from that of olive oil, but there are some interesting instances identified. There is a late antique oil press (or, at least, one that was adapted during the late antique period) at Collet de Carbonnière in the region of the Étang de Berre.Footnote ¹²⁷ A medium-sized oilery with two tanks and (probably) two presses has also been identified nearby at La Pousaraque, which seems to have continued production into the fifth century – which is notable as the general picture is one of decline in oil production after the mid-fourth century.Footnote ¹²⁸ Mentions of viticulture, by contrast, occur in many of our literary and documentary sources. Although large-scale commercial viticulture had long been abandoned in the region, vineyards, and wine production on a smaller scale, were still central parts of the rural economy.Footnote ¹²⁹ Archaeological work in the 1990s in Languedoc has again been particularly valuable, with excavation showing, for example, intensive new work on the vines during the fifth to sixth centuries at Dassargues, near Lunel-Viel (Hérault),Footnote ¹³⁰ while at the port site of Bourbou, near the villa of Près-Bas, there is evidence of wine production in small units.

Who can we imagine working such units? J. P. Brun suggests that they would have been worked by coloni, and that this was part of a broader pattern whereby vineyards were handed over to coloni in place of direct exploitation by landowners.Footnote ¹³¹ While Brun has no specific evidence for his arguments, a number of texts do indicate that slaves and coloni frequently fulfilled the role of vinitores.Footnote ¹³² The ubiquity of ownership of small parcels of vineyard, often known as vineolae, is clearly demonstrated by early medieval estate records from other areas of early medieval France and also by conciliar acts, where these small portions are seen as suitable to give to newly manumitted slaves.Footnote ¹³³ Even if members of the non-elite did not own their own small vineyards, they rented them from their landlords.Footnote ¹³⁴ Furthermore, references to the cultivation of grapes are plentiful in Caesarius’ sermons, showing the continuing popularity of vineyard cultivation as a source for metaphor and analogy in rural contexts.Footnote ¹³⁵

This being a Mediterranean region, there was also plenty of economic activity related to the sea, including at coastal lagoons. The Étang or Bassin de Thau in eastern Languedoc, a coastal lagoon of over 7,000 ha, was a significant fishing area. To get a sense of scale, during the late nineteenth century 140,000 kilos of fish and 720,000 hectolitres of shellfish were caught there each year.Footnote ¹³⁶ The exploitation of seafood beds in antiquity, especially for oysters, is indeed well known. Returning to Provence, looking at the lagoons of the Berre region, we can see that large amounts of this shellfish were eaten across the region in late antiquity, probably more than in earlier periods.Footnote ¹³⁷ In fact, the conchological evidence demonstrates overexploitation during late antiquity, as evidenced both by the wide range of species eaten and the small size of many of the finds.Footnote ¹³⁸ However, fishing was likely just one activity in a polycultural society, particularly as it was often seasonal. Again, comparative evidence is suggestive: in the nineteenth century the fishermen who fished the Thau basin also owned plots of vineyard.Footnote ¹³⁹ Furthermore, coastal lagoons were highly productive not just for fishing but also for salt, with a connection between the two activities – in the form of the salting of fish – being very common.Footnote ¹⁴⁰

Hilary of Arles once more makes an appearance. In a chapter of his Vita designed to show his industry, very much in line with the ascetic model, the author makes the intriguing claim that Hilary fastened nets while dictating.Footnote ¹⁴¹ In this same chapter a much-discussed passage recounts that the bishop walked 30,000 steps (c. 44 km) – a distance that corresponds very nicely with the 45 km between Arles and Saint-Blaise – in order to inspect and help out at some saltworks before undertaking his liturgical and pastoral duties:

The very fact that Hilary made such a journey to visit the saltworks demonstrates their economic importance. This passage does not specify the involvement of the church in this enterprise but the bishop’s arduous excursion surely makes it likely, his ascetic enthusiasm for manual labour notwithstanding. There is unfortunately no surviving archaeological evidence which would enable us to locate the saltworks precisely at Saint-Blaise. The location is made all the more probable, however, by the low water level of the lagoon in this period, suggestive of a high level of salinity. There is also evidence of saltworks at the nearby Étang de Lavalduc, and early medieval documents refer to saltworks in the area.Footnote ¹⁴³ Coastal lagoons are ideal locations for salt production and the region contains a number of still active ‘salines’, including those in the Camargue.Footnote ¹⁴⁴ Villa sites in the region certainly contain evidence of salt-related activities, including the salting tanks in the villa at Barrou, discussed earlier, and the well-preserved basins at the villa near Notre-Dame d’Amour in the Camargue.Footnote ¹⁴⁵ It seems likely, then, that salt production was significant not just in the development of Saint-Blaise but also in the smaller hamlets of the region.Footnote ¹⁴⁶

To return to another interest of Hilary’s – pastoralism – we have already seen how this seems to have undergone a transformation by the late antique period, with the decline of large-scale sheep-rearing on the Crau and the switch to smaller-scale pastoralism. In some cases, goat-rearing and goat-meat consumption might have overtaken sheep-rearing and consumption; indeed, Columeau suggests widespread goat-raising in the Berre region based on zoo-archaeological finds.Footnote ¹⁴⁷ We saw from the Life of Hilary that shepherds seem to have retained their rough and ready reputation.Footnote ¹⁴⁸ As with the visit to the saltworks, we can see the church of Arles exercising a combination of pastoral and economic interest. The visit to the shepherds, however, does have the notable feature of constituting intervention by a bishop among one of the most distinctively marginal groups in the ancient countryside.

Finally, we come to rural artisanship. While the production of materials for use on site (such as the manufacture of roof tiles and (in earlier periods) of amphorae) is to be expected, I shall focus here on more substantial artisanal production, destined for the market, in particular looking at rural ceramic production which we can see taking place on different scales. First, a rural ceramics workshop, producing the local DSP ceramics at a former villa site at La Quintarié in Clermont-Hérault, was uncovered by excavations for the new TGV extension, providing by far the best evidence of ‘secondary’ rural production in our area.Footnote ¹⁴⁹ This was a site that had been abandoned for more than two centuries and was (re)developed in late antiquity, and was active from the end of the fourth century up until perhaps the beginning of the sixth. Three units, each with a distinctive production, have been tentatively identified, each focusing mainly on tableware made for local distribution.Footnote ¹⁵⁰ The question of the social and economic organization of the potters and their workshops is intriguing: there is no villa nearby that can be related to this production site. A cluster of buildings of various types at the site of La Madeleine II, a short distance away, could represent the home of the potters.Footnote ¹⁵¹ But who owned the workshops? The existence of shared facilities for storage and for clay treatment does suggest, as the excavators note, the likelihood of centralized organization – but, as is usual, conclusive evidence is lacking.Footnote ¹⁵² We might well conclude that here is a case of rural artisans working with what we might term relative autonomy. Pottery was not, of course, the only form of artisanal work to be found in the countryside, including at hilltop sites: iron and glass production were also carried out in the re-occupied oppidum of Lombren in the Gard.Footnote ¹⁵³ The ceramic production there is of particular interest, however: pottery workshops have been located with clay, potters’ tools and an oven among the finds, as well as three stamped oil lamps. Here we clearly have a case of small-scale production, taking place in the same very basic ‘cabanes’ that provided both homes and working spaces in this community.Footnote ¹⁵⁴

As we have seen, these forms of production were not generally practised in isolation; polycultural activity is a notable feature in this period and its growth can be clearly linked to changes in settlement and social organization. Returning once more to Bourbou, the port site previously associated with the villa at Près-Bas: amphorae had been made here but the site saw new development during the fifth century. The activities carried out are familiar by now: pottery, fishing, agriculture and viticulture. Crucially, this activity seems to have been ‘dispersed’ rather than centrally organized, as was the case in the earlier period.Footnote ¹⁵⁵ Finally, it is the micro-region of the Étang de Berre that provides the fullest snapshot of polycultural activity. As we would expect, there is evidence of a wide range of cultivation on various scales, including that of cereals, vines, fruit, olivesFootnote ¹⁵⁶ and even cattle-rearing.Footnote ¹⁵⁷ Products of the lagoons were also of great importance to the economy of the region: both fish and shellfish but also salt, as we have seen. Artisanal production included ceramics (with two ateliers for the production of amphorae identified at Velaux)Footnote ¹⁵⁸ and metalwork (with related finds at Saint-Blaise).Footnote ¹⁵⁹ The evidence for the diversification of activities, even within one small farm, must surely again be placed in a context of rising levels of population, clustering in new settlements, and putting concomitant pressure on resources. According to Trément, we can see, alongside production clearly intended for exchange, evidence of ‘un système vivrier ou semi-vivrier’ (i.e. a subsistence or semi-subsistence system) as part of a wide spectrum of economic activity.Footnote ¹⁶⁰

What even these brief surveys show is that the rural inhabitants engaged in popular culture are far from homogeneous in terms of social and economic structure and working conditions, as well as in their levels of relative autonomy. Just as in the case of urban production, discussed in Chapter 2, it is hard to come to any firm conclusions about the organization of labour on the evidence we have. I have suggested at various points that we might be able to see relatively autonomous working conditions, both at former villa sites and in hilltop villages, but this cannot be straightforwardly proven. Might we also want to look for ecclesiastically sponsored artisanal production at this time? Giulio Volpe has identified what he calls ‘un vero e proprio “artigianato ecclesisastico”’ developing in our period in southern Italy, going so far as to brand the bishop of Canosa ‘un vescovo (bishop) manager’.Footnote ¹⁶¹ There is a lack of comparable evidence (e.g. of stamps) for our region but it is not improbable that such interest was present in some cases, given the proximity of new ecclesiastical buildings in developing settlements. I have already suggested an example where we can probably see new oversight from ecclesiastical ownership, in the case of the saltworks. Next, we will look at the impact of the church on the rural landscape and society in more detail.

Religious Structures and Landscapes

The final element in this sketch of the countryside of south-eastern Gaul in late antiquity is the religious landscape. To consider the religious landscape of the late antique countryside purely in terms of ecclesiastical buildings would seem like a bishop’s fantasy come true: of course, late antique Christian cult sites were built in a rural topography that was already suffused with a variety of religious meanings, across many locations. Penny Goodman has nicely summarized the range of ritual sites we need to imagine in late antique Gaul: ‘altars, sacred enclosures, sculpted columns, stone basins, hearths, pits, trees, rocks, water-features, road-ways and hill-tops’.Footnote ¹⁶² I shall consider how religion was embedded in the society and culture of the countryside in much more detail in Chapters 4 and 5, so here I shall just provide an outline of the Christian religious topography and infrastructure as revealed in the archaeological and literary records.

Nonetheless, I still want to avoid a top-down survey which presents ‘the Church’ as ‘Christianizing’ the countryside through a steady and incremental spread of centrally controlled infrastructure, both architectural and clerical: the reality was undoubtedly more complex. As we have already seen, villas increasingly included Christian cult buildings, which thus constituted a substantive element of the religious landscape, alongside parish churches. This distinction is indeed an important one: the parish church, under the direct supervision of the city-based bishop, was not the only or even most numerous form of Christian establishment in the countryside. Moreover, while even in the countryside bishops dominate the late antique sources and modern scholarship alike, they were in reality a rare presence.Footnote ¹⁶³ This is not to say that we cannot detect the ambition of the church to extend its reach over the countryside in terms of preaching, liturgy and pastoral oversight – this will indeed be a key theme of Chapter 4 – however, we need to place this ambition within the constraints and contingencies of the rural landscape.

In terms of actual structures, this was of course a significant period for the building of churches in the countryside, even if we know of few in comparison, for instance, with the Touraine at the same time.Footnote ¹⁶⁴ Nonetheless, Jean Guyon writes that in the fifth and sixth centuries the Provençal countryside had begun to be covered with ‘une “blanche robe” d’églises’.Footnote ¹⁶⁵ Distribution, however, was uneven, just as with villas: the Rhône and Durance valleys and coast were favoured, as well as locations close to major axes of communicationFootnote ¹⁶⁶ (see Map 2). Our textual sources give a picture of two types of ‘church’: on the one hand, the parrocchiae/dioceses, associated with local communities, and on the other, churches/chapels associated with villas, and hence their proprietors – although this distinction, as we shall discuss, is probably too crude.Footnote ¹⁶⁷ As well as the churches and chapels named in literary sources, there are also a number that have been identified archaeologically.Footnote ¹⁶⁸ As we have already seen, these include several sites that can be associated with villas, but ascertaining the nature and chronology of this association is tricky. While in some cases we can clearly link the building of the chapel to the period of aristocratic occupation of the villa, in other cases we cannot.Footnote ¹⁶⁹ What the frequent and clearly significant relationship between the villa and the church does suggest, however, is the continuing importance of villas in rural settlement and social and cultural organization.

The Council of Agde in 506 distinguished between rural parrociaeFootnote ¹⁷⁰ and oratoria in agris: it decreed that worshippers at the latter should attend either the cathedral or parish churches for major feasts, in a clear attempt to retain liturgical control.Footnote ¹⁷¹ Evidently there was considerable ecclesiastical concern over the dangers posed by the autonomy of villa churches.Footnote ¹⁷² Kim Bowes has shown how it is wrong to see the emerging religious structure of the countryside in the fourth and fifth centuries as entirely clerically led, stressing instead the importance of the initiative, influence and sponsorship of the elite.Footnote ¹⁷³ Indeed, that the clergy of the villa churches were effectively part of the villa workforce, rather than coming under the direct supervision of the bishop, is shown by Roman imperial legislation.Footnote ¹⁷⁴ As Bowes points out, we should not expect the interests of the landowner and the bishop to be identical,Footnote ¹⁷⁵ and indeed later in the book we can see points where they clearly diverged.

As for parish churches, during the episcopacy of Caesarius we see a concerted effort to support and shore up their financial position. For instance, the Council of Carpentras in 527 allowed parishes in the province of Arles to use their own revenues for the upkeep of clergy and building, instead of having to hand these over to the bishop.Footnote ¹⁷⁶ This of course required there to be substantial resources available, raising the question of ecclesiastical income and the extent of ecclesiastical landholding. It is in fact very difficult to estimate the extent of land owned by the church – that is, held by various churches and monasteries, rather than assuming a single monolithic ‘Church’ – in our period, given the lack of evidence from charters available for the later period.Footnote ¹⁷⁷ It is nonetheless clear that this period was one of major growth in ecclesiastical landowning. We do not have evidence of major gifts in our region at this time (to compare, for instance, with the donation of more than ninety estates to various institutions made by Desiderius of Cahors (c. 655), according to his Vita, or the 120 units of land left in the will of Bertram, bishop of Le Mans in 616).Footnote ¹⁷⁸ We do get glimpses of the property portfolio of the church of Arles: the dispute between Arles and Marseille over ecclesiastical territory in 417 was caused by Arles building churches on land they owned outside their own jurisdiction, at Citharista (generally identified as La Ciotat, east of Marseille) and Gargarius (Saint-Jean-Garguier in modern-day Gémenos, long associated with Arles).Footnote ¹⁷⁹ The Life of Caesarius also refers to an agricultural property belonging to the church of Arles at Succentriones.Footnote ¹⁸⁰ It is notable that the Council of Agde of 506 was preoccupied with questions of church property, demonstrating its growth and consolidation at this time.Footnote ¹⁸¹ Ultimately, the expansion and sustaining of churches and monasteries required substantial resources. What we cannot know is how much of the land in southern Gaul at this time was in ecclesiastical ownership.

Conciliar legislation during this period shows a real concern for the pastoral as well as the financial dimension of parish churches. The Council of Vaison in 529 supported parish presbyters to foster a local clergy by allowing them to bring up young lectors in their own homes.Footnote ¹⁸² The same council explicitly confirmed the right of presbyters to preach, with even deacons able to recite pre-written homilies in their absence.Footnote ¹⁸³ (By contrast, as Klingshirn points out, this legislation did not affect the priests of ecclesiastical foundations based at villa sites.)Footnote ¹⁸⁴ The number of clergy, many of whom were managed by their bishop, would have been quite substantial, taking into account all the junior clergy; that is, deacons, sub-deacons, lectors. (Beck concludes that at least half of the clergy in south-eastern Gaul would have been rural clergy.)Footnote ¹⁸⁵ Caesarius might well have been exceptional, but the amount of work that he put into the provision and training of rural clergy is striking, including, of course, the licensing of the reading of ready-made sermons.Footnote ¹⁸⁶ His Vita presents the bishop as hard at work on the christianization of his territorium and mentions several rural parishes visited by or in connection with the bishop, to which we shall return later.Footnote ¹⁸⁷ This text shows the programme by the church of Arles to extend its reach, its discipline and indeed its control into the countryside. In the chapters that follow I shall seek to explore the reality on the ground.

We shall now consider what the surviving archaeological evidence for church buildings in our region might (or might not) tell us. Let’s go back to one of our best-understood villa sites, Près-Bas, at Loupian, in the Hérault.Footnote ¹⁸⁸ This is in fact just one example from a number of seemingly villa-associated churches in our region, although it is perhaps the most striking due to its level of preservation.Footnote ¹⁸⁹ Here we have a very substantial villa which underwent significant developments at the start of the fifth century. As we saw, alongside extensive remodelling of the residential part of the villa, there were also seemingly bottom-up economic developments around the same time at the port site to the south of the villa, at Le Bourbou. In the 1980s a late antique church with attached baptistery was excavated, this time around 800 metres north-east of the villa. Dating this church (‘Sainte-Cécile’) precisely is difficult due to the lack of finds and the paucity of the existing remains, but a fifth-century date is generally accepted. According to Kim Bowes, it is clear that the church is part of the villa’s ‘prestige apparatus’.Footnote ¹⁹⁰ However, the large size of the church is striking: the remains suggest a length of 35 metres, with a baptistery attached. Do these factors suggest that it was meant for a wider community than that of the villa alone; indeed, that it could be identified as a ‘parish’ church? Or does it rather demonstrate that it is too crude to make a clear distinction between parish churches and villa chapels?Footnote ¹⁹¹ Alongside chapels/churches at villas, churches have also been identified at several re-occupied oppida. The site of Saint-Blaise is remarkable for having not one but two churches (one intra- and one extra-mural) that can be dated to the fifth/sixth century, the first being particularly prominent and well appointed. Saint-Blaise is not unique: churches/chapels have also been identified at nearby Constantine and at Sainte-Propice.Footnote ¹⁹² As ever, matching the archaeological evidence with the testimony of texts is difficult. However, we have already looked at the visit of Hilary to the saltworks putatively located at Saint-Blaise, suggesting wider interests of the church at the site; that is, encompassing economic as well as pastoral concerns.

The evidence of the church councils suggests a process, albeit slow and halting, towards controlling and standardizing the structure of rural church organization, but we cannot expect the material remains to match any clearly ordained pattern and, as ever, we have to be careful how we read normative texts. Combining the surviving material and textual evidence builds a picture of a rural landscape dotted with a number of churches, and supplied with clergy at various levels. Overall, we can imagine a context where churches and clergy of diverse kinds and statuses were probably a visible part of rural life but not necessarily via the parish structure. Meanwhile, it is clear that churches played a significant role in the evolution of late antique settlement types. The implications and details of these developments will be considered in more depth in Chapter 5, which considers the practice of religion in the late antique countryside.

Conclusions

In this chapter I have sought to provide a useful synthesis of key features of the late antique countryside in order to understand as fully as possible the lived culture of the rural non-elite and how significant transformations in this period shaped the development of popular culture. Variety rather than uniformity has been key, with our non-elites living in a range of habitats, across an assortment of different settlement types and in different forms of social organization. Nonetheless, certain trends are discernible. One is in terms of settlement type, including the re-occupation of oppidum sites and the growth of polynuclear hamlet clusters. In terms of the villa, while the debate goes on, it looks likely that aristocrats were differently present in the countryside than in earlier centuries, and this would have had an effect on the social and economic situation of (and opportunities for) rural non-elites. We will have cause to consider this further as we look at the struggles taking place in the sphere of popular culture.

In some cases, we can see opportunities for members of the non-elite to develop new (or previously abandoned) economic occupations, apparently working without direct input from aristocratic landlords. While the notion of purely peasant (or rather non-elite) self-sufficient communities might remain something of a long shot, there are signs of developments in this direction. However, the growth of the ecclesiastical infrastructure represents a less ambiguous development. Church buildings played a key role in the restructuring of landscapes, most notably at both villas and re-occupied oppida, and also in the local economy (as with the case of Hilary and the saltworks) – a development which tends against autarchy. Finally, the presence of imported goods in the countryside suggests the presence of new, emerging consumers where they did not previously exist.

The question to ask is this: how did these new developments both reflect and indeed provoke new developments in the area of culture and religion? How did elites and non-elites, as well as the church, respond to this changing economic and social landscape? How were these changes reflected in the textual sources of our period, especially those of the church? These are some of the key questions to be addressed in what follows.