THE WESTERN PROVINCES

Philippe leveau

The western provinces can be divided into two zones according to their relationship to Rome, the center of power: a Mediterranean zone in which contacts via the sea prevailed, and a continental and oceanic zone separated by the Alps from Italy. They comprise the Libyan, Iberian, Celtic, and Germanic linguistic zones where Latin, the language of administration imposed by the conqueror, also became the idiom of culture. Since they had never been part of the great Hellenistic empires, their population had no experience of state organizations. With a few notable exceptions, administrative practices characteristic of ancient cities were recent and Roman in origin.

These territories were thus new zones open to Roman initiatives. This marks a first difference in character compared with the east where exploitation had started earlier. Moreover, our written sources for understanding the economy are less abundant and particularly so for the continental sector. They favor the urbanized coastal areas of the Mediterranean, the sectors of the economy where the state intervenes — the emperor’s laws and administration — and areas of military interest.

Our knowledge of regions that were less urbanized, less subject to administrative control, and not affected by military operations is mainly dependent on archaeological sources which are difficult to organize and use as evidence for economic history. Nevertheless, these sources are primarily responsible for the striking increase in our knowledge of the provinces of continental Europe over the last twenty years. Our understanding of rural settlement patterns has been profoundly altered by the introduction and increasingly general practice of landscape archaeology and rural field surveys. In towns and their immediate surroundings, the suburbium, the remarkable development of rescue archaeology has led to a proliferation of excavations. At the same time, the application of modern methods in the earth sciences and environmental studies has allowed us to discover traces of craft activities which had previously been undetected.

I more and larger cities

Population figures remain modest: around 25 to 30 million altogether according to most recent relevant estimates for the provinces of north Africa, Spain, Gaul, Germany, and Britain.1956 The remarkable urbanization of the west which goes hand in hand with a numerical and qualitative increase of rural sites bears witness to the dynamic growth of the early empire.

Rome’s foundation of citizen colonies and later of veteran settlements strongly accelerated the process ofurbanization ofthese provinces. In a few provincial areas, the density of cities comes close to those of the regions of Italy and of the east that had long since been urbanized. This is the case in the eastern part of Africa, in Baetica, and southern Lusitania, in Gallia Narbonensis along the Mediterranean coast and extending north up the valley of the Rhone.

But outside these privileged sectors, the west cannot compare with the east, which was unquestionably more densely populated, more urban, and in general richer and more developed. In this part of the empire, the main concentrations of people are agglomerations of barracks (canabae) capable of uniting 40,000—50,000 around the army camps of the limes.2 Carthage was the only city which, with more than 100,000 inhabitants — estimates range from 60,000 to 300,000 inhabitants — might qualify as a “megapolis.” For regional centers, including provincial capitals, our estimates barely exceed 10,000—20,000 inhabitants. In Gaul and the north-west, some of these settlements do not meet the criteria of “urbanitas’” and hence qualify only as vici. In the Iberian peninsula under the Julio-Claudians, a quarter of the 399 autonomous cities were exempt from the stipendium; beyond the few large agglomerations, cities were but modest centers, the population of most of which would vary between 1,000 to 2,000 inhabitants.3

Ii the countryside

The increasing number of rural sites, revealed more distinctly from the beginning of the Christian era by the inventories and archaeological surveys, is too pervasive to be explained as an overrepresentation ofthose sites which, because of the ceramic evidence, can be “read” more clearly than in previous and later periods.

The quintessential rural Roman settlement and center of agrarian production, the villa, witnessed a boom paralleling that of urbanization. In recent years, systematic surveys and excavations everywhere lead to a dramatic increase of finds, so that the figure of about 1,000 villas for the Gauls,

Germanics, Spain, and Britain given by Smith in 1997 has turned out to be much too low.

One and a half centuries of archaeological research have revealed the diversity of models to which this structure corresponds. Italian archaeologists have defined a type of villa which corresponds with the texts of the Latin agronomists, the “Settefinestre model,” divided into three parts: residential building, villa rustica/production facilities, and accommodation for the slave labor force.1957 The model of the slave villa which produces for the market corresponds with a situation found in Italy towards the end of the Republic and in the early empire. But the Roman villa must equally allow for otium and thereby be a true country mansion.

This ambivalence explains why archaeologists find it difficult to establish a typology of structures emerging from two architectures: an architecture of amenity, which combines the basic forms of decorative Roman architecture in two broad classes of layouts, the peristyle villa and the porticus villa; and a functional architecture that varies regionally and according to mode of production and period. Like the city (or unlike it as the case may be?), the villa can be seen as a factor of economic development or as a parasitic structure expressing the elites’ domination of the countryside. Chronology, purpose, or use are not the only factors dictating the typology of the villa. Location matters too. The agrarian economy of Italy, as opposed to the provinces, is marked by the influx of slaves. The interplay of these factors explains why it is difficult to observe the “Settefinestre” model outside Italy. In the western provinces, the producer-villa did not become established before the middle of the first century ad. But this type arrived earlier on Mediterranean coasts than in central Gaul. In some regions we find what are essentially villas with considerably more modest residential sections. This is the case in Africa, around Caesarea Mauretaniae, beyond the villae maritimae of the coast.5 In Gallia Narbonensis, the number of villas with developed residential quarters is relatively small taking into account (given) the wealth of the provincial aristocracy. In Belgica, Germania, and Britannia, excavations have produced many examples of a succession of estates, one replacing the other: a “small” villa of the first century ad, a “medium-sized” villa of the Antonine epoch, and a large palatial villa at the end of antiquity. This sequence has also been observed in the Iberian Peninsula and in Aquitania. Numerous attempts at evaluating the extent of domains have foundered on a fundamental problem: the relationship between the sizes of the buildings of an estate and that of the land which belonged to it. There is no direct relation between the centuriation, the grid of land assignation, and the villa, the building. Excavations record the successive developmental stages of a building whose owners are known only in exceptional cases. The alternative resulting from this artifact of our source evidence lies at the heart of the problem of the economic development of the west. Is the construction of the villas, a century after the conquest, the consequence of general prosperity brought by Rome, or does this new prosperity rather derive from the investment of wealth previously accumulated by indigenous aristocrats or from plundering? The urban elites and the owners of villas belonged to the same class. But the same sites have permitted two types of interpretations: first, an interpretation in terms of discontinuity — a Roman has taken the place of a local aristocrat; families have concentrated the properties, and alternatively in terms of continuity: the increasing wealth of an individual family finds expression in the construction of more and more luxurious buildings. Our interpretations are intimately linked to paradigms which have been shaped by written Italian sources or inspired by better documented modern history. This difficulty makes it necessary to look at studies of individual regions.

Iii the vicus

The place of the vicus in the rural economy of the west can only be understood in terms of some elaborate historiographical constructs. In a simple model of role division of city and country, the political functions are carried out by the city; the productive functions are divided between villa — for agricultural production — and the vicus for most craft production. In the provinces of the west and of Gaul and Germania in particular, the vici would have gathered a middle class of farmers, owners, or coloni and of artisans who accepted the standardization of the pottery production proposed by the traders. This free population generated regional economic development. There, innovation would have found a more favorable framework than in Italy where the massive influx of slaves and the persistence of slave labor as the mode of production would have prevented it from flourish-ing.1958 This theory rests on research showing that the Roman occupation of the countryside was not limited to the import of the Roman villa model. Resembling the oppidum, as a pre-Roman type of group settlement, the vicus has attracted the attention of those archaeologists who are particularly interested in identifying continuity. On the other hand, the identification of production, in particular the activities of artisans, which are not well represented, is a task for a discipline capable of studying the social strata that are less well represented by written sources and neglected by classicists.

Inventories of excavations have allowed us to identify many settlements of all sizes characterized by their functions as trading post vici, artisan vici,

Peasant vici, mining vici, etc. But excavations in the vicinity of villas compel us to reject the idea that these sites were isolated. Some vici are directly attached to villas, while others are placed several hundred meters from a great villa. In Saarland, the case ofBliesbruck-Reinheim offers an exemplary illustration of the degree of proximity between a great villa and a village.1959 A dynamic approach to the evolution of the rural habitat in the west takes account of such typological diversity. Because of the conquest, the habitat begins to decline. This marks the beginning of new groupings: the rural habitat tends to agglomerate into a vicus developed around a center belonging to a private estate, the villa, or in its immediate proximity. The same process of aggregation explains why in Britain and Germany, and along the limes in Africa, townships formed themselves around forts and sometimes survived the latters’ disappearance. Therefore, the relationship between city, villa, and vicus deserves to be described in terms of complementarity and succession rather then opposition and mutual exclusion.

It is not certain whether the three Gallic provinces were the zone in which vici first appeared. In the context of still modest urbanization and importance of monumental centers, the relatively expansive extension of the suburbium contributes considerably to the impression that northern Gaul was the most developed region in terms of material craftsmanship.

Iv mastering the land and the development of the ager

In archaeological research, two complementary procedures explain the progress achieved in our knowledge of the extension of land cultivated by Rome in the west: archaeo-morphology and archaeological surveys. The former, aerial archaeology and landscape archaeology, have recorded spectacular successes. Vertical aerial photographs and cartographic analysis have demonstrated the crucial role the Roman period played in the west in dividing the land into plots. Centuriation went hand in hand with the foundation of colonies from the end of the civil war onwards and ensured the takeover of the ager publicus. A military framework and forced indigenous labor made it possible to commence the cultivation of large expanses of land, from the north African steppe to the forest of oceanic Europe, and everywhere to make inroads into the marshland. Centuriated, divided, distributed or rented out, these new lands considerably increased the size of the ager (the cultivated territory within the empire). Meanwhile, since the 1990s, attempts to describe the appropriation of provincial land by Rome by studying fragmented fossils have become increasingly contested. It seems now excessive to compare the centuriation to a vast net thrown by the

Roman conquerors over the west and to call it “domestication” of the landscape, analogous to that of the American west in the nineteenth century. In its actual state, the map of centuriations in Europe contrasts in a way which is too schematic to be plausible a Europe of Celtic fields with a centuri-ated Europe, where France and Spain are particularly prominent as the two countries where the “cadastre paradigm” has been most successful. The elimination of many pseudo-centuriations in Gallia Narbonensis and the extension ofarchaeo-morphological research in the regions ofthe north and in Britain leads to the formulation of some plausible propositions.1960 This evolution fits with the assertions of archaeological surveys, which from the 1970s onwards have become a scientific activity. This has multiplied the number of known sites, contributing unarguably to the picture of demographic expansion at the beginning of the Roman period in the west.

More recently, thanks to palaeo-environmentalist studies, already used convincingly in northern Europe,9 a third means of evaluating the development of the ager at the expense of the saltus has emerged. Thus, some environmentalists attribute all perceptible changes in the vegetation from the beginning of the Bronze Age to the establishment of the Mediterranean climate, in particular the growing importance of forms of vegetation which have adapted to the dry summers that characterize this climate. A change in annual rainfall is said to have started in the south of Spain and to have moved up to the Gulf of Lions. But, observing that agricultural and pastoral practices have similar effects to a natural drying-up of the climate, others have attributed these changes in vegetation to human activity. In fact, micro-regional studies in Languedoc show a link between the dynamics of vegetation and those of economic development. The early start of the deforestation in western Languedoc (Aude basin) is linked to its vicinity to Narbo (Narbonne), the capital of the province, to the regional urbanization which is much stronger than along the eastern coastal part of the Languedoc, and to agricultural growth identified by anthracological studies. In the same way, in the coastal zone ofCatalonia, the difference between the territories of Barcino and Tarragona has less to do with natural conditions than with the development of the Roman occupation in the Ebro delta. The south-north gradient observed in the evolution of vegetation certainly does not reflect an evolution of the climate, but the conquest of the ager starting from the southern shores of the Mediterranean.10

V agricultural production

Cereal production was surely ubiquitous, and ought to be the best known form of agriculture. This is particularly so in the area of trade of which the state kept control for the annona of Rome and the frontier army. Even though grain was part of a market which was, in principle, free, this market sector was watched over by local authorities keen to avoid shortages. As a result, there are written sources which — and this is unusual — do not only concern the city of Rome. The areas of production are less well attested. Outside Africa, for which the data of the annona can be used, the available evidence does not allow us either to evaluate the quantities produced or characterize the zones of production with any degree of precision. The geography of grain production is too often extrapolated by economic historians from a matching up of the map of the units of production — the villas — the centers of consumption — the cities and the limes — and of the land suitable for grain production. Partial exceptions are Gallia Belgica and the north-west of Europe, where regional cereal cultivation is by now well documented thanks to palaeo-carpology and research on technical innovation.1961 The same holds for animal production. It was of fundamental importance for the food supply, but archaeological data are still not explicit enough to identify specific zones of husbandry. Meanwhile, archaeo-zoology contributes qualitative data which fuel debates over the increase in size of domestic animals; as this happened from the beginning of the conquest, it seems it can be attributed to improvement in animal husbandry.

Since the essence of what can be observed about arable techniques and animal husbandry relates to the consumption of their products, archaeology allows us to “read” the data better and to identify zones of olive and wine production and to study their evolution.12 The techniques of studying amphoras as economic indicators developed in Italy by international teams have enabled us to write a history of the trade of foodstuffs and to identify the routes which brought them to the centers of consumption, in the first place Rome, then the camps of the limes, finally to the provincial sites.13 This history for the provinces of the west shows a reversal of the situation analogous to that of pottery production: first Gaul, but also the Iberian Peninsula, evolved from importers of Italian wine in the Dressel amphoras to exporters to Italy in amphoras of a specific type.

Olive oil is an exclusively Mediterranean product whose importance is linked to a cultural phenomenon: it plays an important role outside its zone of production, namely in the diet of the army on the limes and of the Roman and Romanized elites. This brought about a growing demand which is one of the important factors of the agricultural conquest in the Maghreb, well established by archaeological surveys and epigraphic data. The inscriptions of the saltus of the Mejerda occupy an essential place in this dossier. In the field, recognizing production zones is helped by the stone components of oil presses. At the end of the seventies, Carandini, basing his claims on

Studies of pottery, proposed to explain the urban density in the north of modern Tunisia as a function of the export of African red slip ware and of agricultural products: grain — though the income from production on imperial domains did not benefit the cities — and above all oil, which was going to compete mainly with that of Baetica.1962 Some twenty years later, drawing together archaeological facts available for three regions, the south of Spain, the Tripolitana, and Tunisia, Mattingly showed that there was a considerable increase in the amount of land devoted to olive growing.1963 In the centuriated regions of Africa vetus, surveys conducted on foot and from the air have provided evidence for a rapid expansion in the numbers of olive presses on geometric plots of land or for traces of plantation holes. These observations have also been used by Hitchner as evidence for the capacity of the Roman economy to generate growth.16

Since then, the microregional study conducted by a Tunisian Danish team of a space of 40,000 hectares around Segermes, one of the many little towns of Africa Proconsularis, has led 0rsted to insist on the diversity of soils and production methods. He also reminds us that it is necessary not only to see the agrarian economy of north Africa with regards to the Roman market, but also to take into account the regional dimension.17 The material found by surveys clusters chronologically in a period of at least fifty years during which the finds are rarely strictly contemporaneous; therefore one has to refrain from drawing too much of a parallel between antiquity and the Tunisian olive production in the nineteenth century. It is here that observations on the presence of Olea in pollen diagrams based on the series of marine sediments from the Gulf of Gabes become relevant:18 there, the Roman period shows a share of about 10 percent, i. e., the same level as observed in the sediments from the harbor of Carthage during the same period.19 These percentages, double what palynologists would accept as evidence for olive cultivation, confirm the importance of olive trees within this landscape. But they do not allow us to assume a monoculture comparable to that of the ninteenth century. In the same sediment sample, the percentage increases to nearly 20 percent in the Middle Ages and up to 40 percent during the French mandate (late nineteenth to mid-twentieth century). Such percentages were reached in antiquity too, but only elsewhere, namely in Baetica, another major region of olive production. In the pollen profile of Laguna Medina, near Cadiz, a peak of Olea reaches nearly 40 percent for a period datable to the second/third century ad. These measurements contrast with previous and later periods during which the curve settles between 5 and 10 percent.20 The combination of these facts thus justifies the hypothesis ofcommercial agricultural production, but also documents relative differences in importance for different regions.

From the 1990s onwards, studies of viticulture, whose importance for the economy of Roman Italy had already been known from the precise information offered by ancient treatises on agronomy, have once again become a major preoccupation of archaeological research. Outside Italy, its importance is tied to the inclusion of wine in the ration of legionaries in the camps of the limes and to consumption by Italian settlers and the provincial elites as a departure from a pattern of consumption which had reserved wine for special occasions only. In the same way as for olive oil, the identification of the provenance of the amphoras in which wine was distributed has advanced our knowledge of the zones of production.1964 In Provence and Languedoc, where both plants are cultivated, the excavation ofwine storage facilities led to a reevaluation ofthe importance ofviticulture relative to olive cultivation. It is difficult to distinguish the two just through careful study of the possible or likely stone bases and supports of oil and wine presses which are rarely distinguished by casual field survey. Brun and Laubenheimer have admirably synthesized some twenty years of research on viticulture in Gaul.22 They have shown its essential place in the agricultural economy of southern Gaul. But there is a crucial difference between the two plants: the farmers of the ancient world were able to move production nearer to the places of consumption by making the Mediterranean plant adapt to more northern climatic conditions. Its importance in Aquitania has been confirmed. Hence “with the exception of isolated regions and it seems the Alsace, the Mosel and Rhine areas,” the vine had conquered “from the first century ad all the territories it was to occupy in the Middle Ages.”23 Archaeological data prove that it reached Britain as well. Viticulture on the Mosel and Rhine offers remarkable examples of integration into the local economy. The archaeological evidence for its rise clearly postdates the construction of the legionary camps, which means that in this sphere their presence can be linked to political developments. The oldest evidence dates from the third century. The civic and military elite, and later the imperial court which was established in fourth-century ad Trier, stimulated “the demand for good wine and provide[d] the financial means to establish both wine production plants and the vineyards themselves.”24 But one should not imagine the Roman west as one huge vineyard as in the past, based on Arab writers, one had pictured Africa as one huge olive grove.

Vi mines and quarries

Urban and rural building works, road construction, and hydraulic works necessitated the opening of quarries and the construction of ovens to burn bricks and lime. The market was local or at most regional. Where the environment lent itselfto it, quarries were opened near the building site and

Maintained for as long as building went on: the quarry of L’Estel near the Pont du Gard provides an example of this practice. In other cases, building programs generated local supply: the lime of the ovens of Iversheim sur L’Erft supplied the sites along the Rhine; the stone from the Midi moved up the Rhone towards Lyon; brick cargos circulated on the Guadalquivir. The precious marbles from Africa, Italy, and the east are found on sites along the Rhine.

The pattern is different for the products ofmines: lead; which is generally used in particular for urban water supply; silver, which is gained from galena, its ore; and gold, whose exploitation was controlled by the state. Mining activities were stimulated by urban development, the luxury of the aristocracy, and the Roman state’s demand for metal for coinage. While the mines of the eastern provinces had already been exploited in earlier periods, the mining areas of the west were put to use successively and in a coordinated manner to respond to the imperial needs and to ensure the monetarization of its economy. The activity of mines in Spain, brought to light again by the research of Domergue, is at present the best known.25 The “industrial” scale of the exploitation is demonstrated by the importance of the workforce, reaching (according to Polybius) 40,000 workers in the mines of Carthago Nova (Cartagena) (Strabo 3.2.10). In principle, one cannot count as enrichment of the region the considerable means invested by the Roman state in the exploitation of the resources of their provincial subjects. However, the Iberian example shows that during the empire the system was able to evolve. In the south-east of the peninsula, the mining sectors of the eastern Sierra Morena and the Sierra of Cartagena and of Mazaron were progressively integrated into the regional economy. The activity of the mines of Carthago Nova largely came to an end towards the Augustan period. But the families of Italian origin which had been involved in their exploitation for several generations ensured the durability of the economic development of Carthago Nova, which became a colony under Caesar or Augustus. This situation distinguished this zone from other mountain districts which remained marginal.26

Vii manufacture (a) Metallurgy

The working of precious metals with their specific uses has to be contrasted with the working of iron. The use of this metal is crucial for the efficiency of tools used in both agriculture and manufacturing. In principle, its primary production did not differ from that of other metals: after a period

Orejas 2001—3.

When mining concessions were given to rich Italian entrepreneurs, production came to be controlled by procuratores ferrariarum attested in Gallia Narbonensis and Lugdunensis. But on the level of consumption there are fundamental differences. The example of the Swiss plateau is typical. In an area systematically studied by surveys, the Roman period is characterized by a real abundance of iron in everyday life.27 The mediocre conservation of this metal and the recycling of objects had led to an underestimation of its use. Systematic studies of countless known piles of slag allowed only identification of slags from the forge rather than from smelting, which would be proof of primary production. The metal was imported in the form of ingots: from zones nearby, located in France? Noricum? Burgundy? The question remains open. But the last one of these regions is among those where primary iron production of major importance has been recognized. Systematic survey has in fact allowed us to identify mining works and piles of slag in central Gaul, among the Senones, the Haedui and the Bituriges and in Gallia Narbonensis, in the Montagne Noire, where production greatly increased from the middle of the first century bc onwards.

Containing hundreds if not thousands of workshops, these zones can be distinguished from districts of middling importance, where there are some hundreds of sites. Systematic use of Carbon 14 together with traditional archaeological techniques has enabled us to distinguish them from regions which had been thought to be important in antiquity but only had become so during the Middle Ages. The region between the Sambre and Moselle rivers is a case in point. On the other hand, in Morvan and Berry, the clever use of field survey enabled researchers to propose the existence of areas that had been specifically used for the reduction of iron ore.28 This example illustrates the unevenness of regional development. Agetencum, the main settlement of the Senones, where we know mainly about primary production, is described as a vicus. By contrast, Autun and Avaricum, the main settlements ofthe Haedui and the Bituriges Cubi, were endowed with all the public monuments commonly associated with Roman urbanism. A class of metal-working craftsmen is attested there. This continued into the late empire with the installation offabricae for arms in these two cities, in Autun itself and in Argentomagus. Since the 1980s, when Cleere reviewed the organization of this production in the western provinces of the empire, research spurred on by Mangin has allowed us to draw a map of several zones of production in Gaul.29 Recent fieldwork and the systematic use of geo-chemistry to differentiate slags from the smelting of the mineral from forge slags of smithies have provided more information than textual sources would have offered.30 The same is true of the metallurgy of lead. Isotopic geo-chemistry shows the diversity of its provenience in western Switzerland:

' Leroy 2001: 90.  29 Mangin et al. 1995. 30 Serneels 1998.

Local (from the Valais), regional (from the Vosges) and material from further away (Britain, Spanish peninsula).1965

(b) Pottery

Craftsmen worked in all kinds of settlements. There is not a single case of an urban agglomeration without workshops, located in the suburbium rather than the civic center. Detailed excavations have shown that they were to be found just as much in villas as in village-type agglomerations. By “detailed excavations” we mean those excavations which not only investigate buildings and systematically collect all visible remains including slags, crucibles, and fragments of moulds, but also thoroughly sieve excavated soil (or subject it to flotation, thereby recovering all plant remains, bone, and debris from craft production). This procedure has brought to light, for example, tiny droplets of bronze which allowed us to determine the location of manufacture of objects and metal vessels typically found elsewhere, exported from the Rhineland to the Orient. They have corrected the focus of discussion of the artisans, which once put more emphasis on their social status and the distribution of their products than on production itself.32

It is from this angle that one would expect insights on pottery, which survives in large quantities. We have underlined the importance of widespread use of terracotta for the production of building materials and containers for agricultural products. The real importance of this product within the wider economy is of course much more limited. The value of amphoras is mainly that of their content. To a lesser degree the same is true of pottery dishes, the main chronological indicator in stratigraphy of Roman sites. This limitation does not prevent us from paying proper attention to the massive production of Gallic terra sigillata. Its presence in Pompeii in ad 79 signals the reversal of the relationship between the Gauls and Italy and the rise of the provinces. A map of sigillata workshops lets us follow an evolution which began with the settlement of Italian potters in the suburbium of Lyon. It continues with their migration towards the northwestern margins of Gallia Narbonensis, at Montans and La Graufesenque, a suburbium of Condatomagus, the capital of the Ruteni, then from there successively towards Lezoux and finally towards the north-west, Argonne and Lorraine (Gallia Belgica) and along the Rhine.33 In fact, production of fine pottery becomes more and more regional. In the Iberian Peninsula, from Claudius onwards, regional products replaced pottery from Gaul and were distributed throughout Mauretania Tingitana from two main centers in the Ebro valley around Tritium Magallum and Andujar in Baetica. They differ in importance: Tritium Magallum is as important as La Graufe-senque, while Andujar wares were disseminated less widely. But in both cases we are dealing with highly urbanized areas. The third provincial area which reveals comparable productivity, Africa, exported its sigillata. The works of Peacock have shown that one of these, African Red Slip C, was produced in central Tunisia.1966 But the origin of African Red Slip A, widely distributed around the western Mediterranean, is still unknown. It would therefore be premature to pronounce on its role in a regional economy. These workshops are also a sign of progress: current research on the workshops of central Gaul — the most important are around the site of Lezoux in the Allier valley — between the first and the fourth centuries ad challenges overly linear views of the evolution of centers and patterns of production. An archaeology of furnaces shows how the potters responded to the need “to adapt their working tools to the constraints of highly competitive markets.”35

Viii the geography of trade and regional development

In the west, where a common currency had developed over the course of the first century ad through the disappearance of municipal coinages and increased monetarization of the economy, trade depended primarily upon the development of infrastructure, most importantly the systematically organized road network radiating from Rome outwards. Of course, the Roman state limited its efforts to the viae militares which were essential for the cursuspublicus. But trade followed swiftly in the wake of conquest and administration; traders were to contribute for their maintenance with their taxes. We must not exaggerate the scope of the road system. Modern maps of the Roman road network do not take time into account: they represent the sum total of all road works in antiquity. “Some routes are bound to have shrunk and disappeared while others, appearing later, replaced them.”36 Roads were not a completely new development either: before Rome, during the Barcid period, an organized network already existed in the Spanish peninsula; in southern Gaul, the movements of the Cimbri and the Teutones and those of Roman armies show the importance of movement on land routes between northern Europe and the Mediterranean. Strabo drew attention to “the harmonious agreement (homologia) which characterizes” the Gauls “in relation to the water routes and the two seas which form its borders” (4.1.14). The Roman investment in infrastructure to make these river systems more navigable, which was considerable, also contributed to the development of this network. It was chiefly through the Rhone valley since the Hallstatt period that objects produced in Etruria, Greece, and the east penetrated the country. It remained the main commercial route from the Mediterranean to the Rhineland.

Punctuated by a string of cities, whose construction marks a break with the geography of the tribal territories of the pre-Roman period, there existed a land route that ran parallel with the river. Its development had begun with the access canal to the Rhone which Marius had dug in 102 bc. It combined more or less developed river sections. Beyond Lyon, the canal from the Saone to the Mosel planned by the legate for Germania Superior (Tac. Ann. 13.53.2) was never realized. This axis was competing with the routes along the Atlantic and the Danube. From the second century ad onward it was possible to travel from the east directly to the regions of the north-west by following first the Danube and then the Rhine. As elsewhere, trade benefited from the same network of roads or waterways designed to enable the movement of troops along the limes. The Danube route did not completely replace the central axis. The same is true of the route along the Atlantic, which linked Mauretania Tingitana and the Iberian provinces, in particular Baetica, to Britannia. In theory, this doubled the old route linking the Mediterranean to the Atlantic via the Aude and the Garonne. In reality, however, there were obstacles to its use in the Gulf of Gascogne which even advances in maritime navigation did not fully overcome.

Until Rome’s assumption of control of the Rhine—Danube regions, most trade followed the routes along the Rhone or over the Brenner and widely skirted a mountain range destined by its altitude to remain of marginal importance for trade. Once Rome had extended its empire northwards, geography demanded that the state exercised total control of the Alpine passes. This explains why during the early empire the administrative geography of these provinces was defined by the routes and itineraries which ensured travel from Italy: the alpine provinces commanded the passes.1967 Here too the merchant followed the army, which led to the type of settlement which was to become the alpine city. On the Italian side these grew into “real” cities. On the provincial side, we find vici located along the routes leading down from the passes. These were closely linked to transalpine traffic but also increased the value of local resources. In Gallia Narbonensis, the civitas of the Allobroges is attested in a pre-alpine form as in Vienne and in two alpine civitates, Cularo/Grenoble and Geneva. On the other hand, there does not seem to be any decisive change in the exploitation of the other two resources specific to the mountains, namely vegetation (forage from the pastures and rangelands and wood from the forests) and the mines. The very different evolution of the economy of the

Map 24.1 Communications and development in the western provinces Graphics: M. Sintes-Aioutz

Mountain areas demonstrates the impact of the new centrality of Rome in defining an administratively fragmented area. When describing the borders of the Gallic provinces Strabo (4.1.i) stresses the opportunistic character of the division of the provinces which takes no account of the old borders. This was particularly true in the Alps, where the new organization was necessitated by a completely new circulation network.

Ix the weight of the regions

The example of the Alps leads to the central question which now can be addressed: what was the impact of “Romanization” on the regions? This concept is open to criticism on account of its vagueness, but its very flexibility allows it to be used as a footbridge providing a tenuous passage between different domains of provincial life (religion, language, art, economy. . .) where one can recognize the influence of Rome. I have used it above in connection with an activity specific to mountains, mining, which in the south-west of the Iberian peninsula allowed the integration of certain mountain regions — the Sierras of eastern Morena, Cartagena and Mazaron — while others retained their marginality. The Roman empire made use of the technical means at its disposal to master an environment in very specific contexts which have yet to be categorized. This leads us to the question of the conquest of the cultivated land which also had been touched upon earlier on with reference to southern Gaul. In the first century ad, the Elder Pliny compared Gallia Narbonensis to Italy (HN3.31). But this complimentary generalization conceals extremely diverse conditions. The Narbonensis featured vineyards, olive groves, fields ofgrain on the silt ofthe Rhone, transhumant flocks in the Crau, towns aligned along the coast and along the Rhone and Aude-Garonne valleys. But it also included remote alpine valleys.

Similarly, the contradictions in Pliny’s observations on the Iberian provinces might be explained both chronologically, in terms of different stages, and geographically, in terms of regional differences. Their urbanization came to a halt under the Julio-Claudians when a quarter of the autonomous cities escaped the stipendium. “Beyond some sizable agglomerations, the large majority of cities were centers of modest size which could vary from i, ooo to 2,000 inhabitants.”1968 Meanwhile, pace Pliny, who calls the award of Latin rights by Vespasian “rash in a period of turmoil for the state” (HN 3.4.30), the municipalization of the peninsula cannot be understood without the context of a strong economic development. Baetica, where Pliny counted 175 cities, and neighboring Lusitania saw a level of development which recalls that of Gallia Narbonensis. In Baetica since the Flavian epoch, Gades, Hispalis, Corduba, and Italica produced a remarkable series of aristocratic dwellings. Birthplace of two emperors, Italica spread a model pattern of the aristocratic house in the peninsula.39 Emerita, capital of Lusitania, received successive waves of settlers. Its cen-turiations and the construction of dams associated with large villas and with irrigation zones, fit with the picture ofa true “pioneer frontier.” In the north, the vast area ofTarraconensis was divided into three conventus whose centers formed nuclei of development: Tarraco and the coastal cities of its conventus, Caesaraugusta in the Ebro valley, and in the south Carthago Nova with its mining zone. In North Africa, where the way for Romanization had been paved by the encounter with Punic civilization and, from Masinissa onwards, the wish of Numidian princes and Moorish kings to embrace classical civilization, is not fundamentally different. Africa Proconsularis and Numidia stand out by virtue of the sheer number of towns, the quality of their urban culture and the position of their elite within the empire: under Marcus Aurelius, 15 percent of the senators whose local origins are known were Africans. This development affected above all northern Tunisia and its extension into the north-east of Algeria. But the idea of a radical separation of these zones from that of a mountainous Mauritanian west left to primitive Berber tribes is not acceptable. This contestable conception1969 revives the old contrast between a “Roman Africa” and a “forgotten Africa”; the three capitals of the Mauretaniae — Sitifis on its plateau, Caesarea and Tingi on the coast — constitute just as much poles of urbanization and development. In this the western Maghreb, the sea routes play a principal part while a geo-ecological factor, the rise of the desert to the Algero-Moroccan border, accentuates the peripheral character of the region.

In the early empire, the preponderance of the Mediterranean regions is indisputable. Did this change later when the emperors established themselves in Trier? Did they simply respond to a military necessity or was this the consequence of a reorganized imperial economic system? If so, after Rome lost its central position, peripheral regions would have acceded to the status of “new core regions.” Haselgrove maintains that the historic rise of the regions of northern Europe clearly predates the break-up of the unity of the Mediterranean brought about by Islam.41 He argues that between Augustus and the crisis of the third century ad, the countries between the Mosel and the Rhine had seen remarkable development in which their support for Rome in the crisis of ad 69 constituted an important stage. The civitates of the Treviri and the Mediomatrices involved in providing supplies for the frontiers would thus have become poles open to innovation.42 The military camps and the cities of the limes would also have attracted long-distance trade. These incontestable facts, which justify attributing to the frontier regions this role in economic development, should be contrasted with others, which put this overestimation into perspective and attribute the shift of the center of power to a (geographical) periphery to geopolitical and military considerations rather than to the economic weight the periphery had acquired. Under the Severans only a small number of senators came from the three Gallic provinces. The contrast between a less urbanized western Gaul and an east traversed by the commercial axis from the Rhine to the Saone remains noticeable. The cities of the north-west were both fewer in number and less prosperous. In Britain, where economic development has been re-evaluated, urban activity remained modest compared to that of the Mediterranean regions, despite the blossoming of building activities after Hadrian’s visit.

X a typochronology of regional development in the roman west

An analysis of the Roman west in terms of regional development shows a diversity which one could categorize according to the fourfold modern classification of J. Friedmann.1970 Without a doubt, only the region of Carthage could perhaps be classed as a “core region.” In light of what has been said about the development of the provinces of the north-west, they cannot be counted among this category, but among the “upward-transition regions,” a status they reached by the end of the early empire. As discussed above, the regions of the Iberian peninsula had been in this category for a long time already. Its south-west constituted a unique key economic unit which asserted itself despite the provincial divisions. But none of its large cities, Emerita, Corduba, and Hispalis, dominated the region clearly enough to allow classing the region as a “core region.” The same is true of the valleys of the Rhone and Saone: stretching from Arles to Autun between Alps and the Massif Central, they acquired this status as a result of an evolution which recent research has retraced.44 They formed a corridor of strongly urbanized development, but there was no single pole (neither Lyon nor Vienne).

All these regions previously went through the stage of “resources frontier regions,” the third of the categories defined by Friedmann. Among these one would count also the large cities of western Gaul, in particular Pictons and Bituriges. The fourth and final category, “downward-transition region,” might comprise regions which were already urbanized or on the way to urbanization at the time of conquest, and saw a decline during the early empire. I have dismissed this hypothesis for the Mediterranean regions as a whole. But this proposition does not assume that all would need to belong to the same category and some might be considered to be “downward-transition regions.”

Regions which experienced their first phase of development at the end of the Republic and in the early empire are also those where the first withdrawals, which took place from the middle of the second century ad onward, signalled the crisis of the third. Gallia Narbonensis is a good example. While the alpine zone became more urbanized, in the south-west

’ Woolf 1998:193-202.

Of the province the disappearance of Ruscino shows a contrary trend. In the countryside, surveys show a decreasing number of sites from the reign of Marcus Aurelius onwards. Was there a crisis in the province comparable to that attributed to Italy, with which Pliny had compared it? This is the subject of a debate connected with the interpretation of the archaeological data. The “abandonment” of the countryside can lead to a pessimistic interpretation of the change in settlement pattern towards larger agglomerations. Hence the impression of a crisis could simply be an effect of the scale of observation.

In his classification, Friedmann thinks of societies of the industrial age. For completeness, a fifth category is needed which covers regions where forms of economic life dominant in prehistoric times had persisted. These regions owed their marginality to geo-ecological factors: the Mediterranean hill regions, the steppes of the Iberian peninsula and Africa, the high Alps and Pyrenees, the great forests of the Gauls and of the north-west and the marshes, both inland and coastal as well as in deltas.

Xi conclusion

The data presented here can be read in different ways. It is possible that the urbanization of these provinces and the related efforts to provide roads or sea or river routes served simply to transfer the wealth of the provinces for the benefit of Rome and its elites. Taking into account the rigidity of the system in the absence of major technological innovations, the facts cited to argue for economic development would in reality be behind a major crisis of which there were signs already from the Antonine epoch onwards. In the fourth and fifth centuries ad, in the context of practically no (economic) growth, the concentration of wealth in the hands of a shrinking minority and the military collapse of Rome would have liberated its peripheries, whereupon Africa and the north-west of Europe became new “core regions.” Such a paradigm fits quite well with models which underline the subordination of the provinces in the service of Rome and the provinces and in particular with the model of the “consumer city.” The latter insists on the dissemination of a common political and cultural model, the urban model, derived from the system of the city state, the spread of Latin and of negotiatores who did business with the whole population, not merely with a small minority of aristocrats. Rome becomes the common fatherland of a united Occident. This reading favors an approach which sees all development as a spread of the economic model of Rome. Yet as critics of the concept of Romanization emphasize, the history of the Roman west cannot be reduced to the spread of a cultural model. Regional specificities play an essential role as well. In an evolutionist perspective, the twin concepts of “center and periphery”, developed in the 1960s by Marxist theorists1971 are particularly relevant in accounting for the disruption in economic circulation caused by the integration of the defeated into the imperial system. On the scale of the empire and on a geopolitical level, the subordination of the conquered territories to the service of Rome and the army of the limes characterizes all the provinces of the west as dependent periphery. A regional analysis, however, documents differences between them which cannot be explained by either their original ethnic diversity or by the deceptive heterogeneous nature of the data at our disposal. Since prehistoric times, a small number of regions, in contact with the trading economies of the Mediterranean, distinguished themselves from the rest of Africa and Europe. Often the trader preceded the soldier. In the west, the restructuring of regions which promoted the integration into the Roman system was gradual and uneven. Between Rome and the limes, regional diversification was the result of the interplay of internal factors leading to the emergence of the regional poles presented above. This does not contrast cities with main local capitals which related to their countryside in a manner analogous to the relationship between Rome and its empire, but generates differentiated regional units. From an evolutionist perspective, regional differentiation caused by this process led to a positive dynamic of economic development. New regional differences did not result in a simple recreation of the previous situation.

Thus, in an evolutionist perspective, regional diversity would be the motor of innovation. Economic development would be real. Demographic development would correspond to the increase in the number ofcities built and of rural sites and, contrary to Malthusian theory and in accordance with what Boserup proposed, would justify belief in the possibility of sustainable development in the west. Demographic issues therefore seem to be essential, and we may wonder whether the suggested population total of 25 or 30 million for the western provinces is compatible with the current portrayal of the development conveyed by archaeological research.