dAf 71
COLIN Anne
Chronologie des oppida de la Gaule non méditerranéenne
In this work stemming from her doctoral thesis, Anne Colin establishes a chronological framework for fortified and open settlements at the end of the Iron Age, in order to further our understanding of the pre-urbanisation phenomena that affected the Celtic world of this period. Examining ancient and recent documentation on some thirty sites and using a statistical series method applied to Gallic and imported furnishings, Anne Colin defines five phases which she then compares to existing regional chronologies. She analyses the diversity observed in regional trends, diversity due to differing political and economic circumstances, and also to the complex role played by contacts with the Mediterranean sphere, throughout a large part of Gaul.
The study volume is completed by a site catalogue that includes unpublished data and bibliographical references.
Abstract
Abstract
Introduction
The discrepancy in date between late Iron Age fortified sites with urban functions (oppida) in Gaul, and their earlier counterparts in central and eastern Europe, is the starting point of this study. Its principal aim is to establish a firm chronological framework for late Iron Age settlements by examining the available documentation, both older and more recent, for a selected group of sites, and, by so doing, to clarify and help understand the phenomenon of proto-urbanisation as it affected the Celtic world. The study encompasses not only fortified sites but also open settlements which are integral to this process.
1. Background and definitions
The history of research on oppida has conditioned the nature of the data available on the problems which concern us here. It was the desire to locate on the ground the places described in Caesar's Commentaries on the Gallic War -nearly all oppida- that first led 19th century archaeologists to excavate this type of settlement. The mass of material recovered by the beginning of the 20th century allowed Déchelette to construct a periodization of the Second or La Tène Iron Age. His third and final phase corresponded with the occupation of the oppida and spanned the whole of the first century BC. He also established the concept of an "oppida civilisation" covering the whole of temperate Europe.
Déchelette's death in 1914 brought this period of intensive research in France to a close. Between then and the late 1960s only a few individuals continued this tradition of research, whilst the academic initiative in later prehistoric archaeology shifted towards Great Britain, Germany and central Europe, where large numbers of new cemetery and settlement excavations led to the abandonment of the chronological scheme constructed by Déchelette. In France, the existence of classes of settlement a1ready well attested in Germany and Switzerland (lowland oppida and large open villages) was eventually acknowledged at the end of the 1960s, but the lack of a reliable chronological framework for the late La Tène period added to uncertainty surrounding their origins and development, questions which form the basis of the present work.
What is meant by the term oppidum in Caesar's text and its usage by archaeologists is discussed in depth. Oppidum is a term frequently used by Caesar, without any precise definition. The sites thus designated -with their often elaborate ramparts, their streets and their squares, where commerce took place and important political decisions were taken- are placed at the top of the settlement hierarchy. This type of oppidum, which unites various urban functions, is essentially limited to east-central Gaul. Caesar's evidence, however, does not permit us to determine to what extent this model can be applied to other parts of the country; and in places he gives the impression that the situation was quite different from one people (civitas) to another. From a strictly chronological standpoint, his Commentaries provide no other information beyond the existence, at the time of the Roman Conquest, of the sites mentioned in the text.
The current archaeological definition of oppidum is based largely on the work of German scholars: a large fortified site, Completely enclosed by a continuous rampart broken only by inturned entrances, characteristic of Celtic Europe at the end of the Iron Age. To this basic . definition, various functional considerations have been added: the presence of craft activities located in specialised zones, and evidence for long distance commercial exchange. Without being able to talk of a town or city in the Classical sense, these attributes give the oppidum a proto-urban character. While there is a fair degree of consensus over the essentials of this definition, opinion is divided on details which are nonetheless important, for example surface area. Equally, many fortified sites do not possess all the requisite physical attributes, although they do house thriving craft and commercial activity. In short, the oppidum is not an homogenous class of settlement. Accordingly, sites were selected for study on the basis of their chronological potential, rather than on strict interpretation of the above criteria which are in any case not especially satisfactory.
2. The chronological framework of the Late Iron Age
This section starts by setting out the principal chronological schemes in use for the Late Iron Age, ultimately deriving from the work of Déchelette in France and Reinecke in Germany. The absolute dating evidence for the last two or three centuries BC is then considered; this derives from a variety of sources: historical events referred to in ancient texts, Roman coins present in archaeological contexts, and dendrochronology. Utilisation of this data is not straightforward: the direct impact of a given historical event on the occupation of a settlement can never be demonstrated with absolute certainty; Roman coins yield only a terminus post quem and could have been in circulation for a long time; dendrochronological dates - the number of which has fortunately increased in recent years are not always of great value for want of associated archaeological material, and can be difficult to interpret. The absolute dates given in publications are in practice almost always the product of a process of interpolation, whereby the phases of relative chronology are replaced by spans of 20-30 years according to individual author, within a chronological framework calibrated by just a few fixed points.
Next, the material culture of most relevance for chronological purposes -imports, personal ornaments, pottery and coins- is examined.
The chronology of imported pottery is based on material from wrecks or from dry-land sites which have been dated by reference to historical events. These goods originate from Italy, or more rarely from Spain, southern Gaul or Greece, and until the last third of the first century BC mainly comprise Dressel 1 amphorae and Campanian A and B pottery. Their different variants are broadly contemporary, but the relative proportions change through time.
Brooch chronology rests on dating derived from cemeteries (in Germany, Switzerland, and occasionally France), from the stratified sequences of Mediterranean settlements, and from the Roman military camps on the Rhine frontier. Some brooch types can be given a terminus ante or post quem (e.g. from dendrochronological dates, or presence in the ditches at Alesia); the Nauheim is one of the types for which new dating evidence has had most consequences for late Iron Age chronology.
Gaulish pottery tends to be fairly standardised as a result of technological developments such as widespread use of the potter's wheel or the perfection of new firing techniques, as well as due to typological borrowing from imported ceramics. There is however perceptible regional variation: significant local differences cloud the interpretation of statistics based on fabric, manufacture and firing. But certain trends in the evolution of form and decoration go beyond local or regional bounds and allow a typologically-based chronological classification to be put forward.
French numismatic studies have come a long way since the time of J.B. Colbert de Beaulieu, but using Iron Age coins as a means of dating remains fraught with difficulties. While the chronological succession of alloys and of certain individual coin series is well enough established in general terms, the absolute chronology is still open to debate, especially for the final phase of La Tène which is characterised by a particular abundance and variety of coin types. One recent advance is acceptance of the production of low-value coinage (potin) well before the Gallic War, which for a long time was denied by numismatists because it would have involved abandoning the link between the relative chronology based on alloys and the chronology of the Gallic War (one of the basic tenets of Colbert de Beaulieu's schema), as weil as implying the existence of an economic system sufficiently complex for coin to have been used in commercial transactions.
3. Chronological analysis
The character of Late Iron Age settlement remains in temperate Europe -a lack of surviving occupation levels and thus of stratified sequences, and an abundance of material culture necessitates a more complex approach to chronology than simply recording the presence or absence of a given type fossil. Statistical methods have therefore been used in settlement analysis since the late 1960s; such methods are based on the principal that the varying proportions of different types of object associated in archaeological assemblages carry more information collectively than individual items. The overall facies of material obtained permit the characterisation of successive stages in the history of a site, or its placing in time relative to others.
This approach can come up against a range of obstacles: statistical problems caused by shortages of certain categories of material; and problems of comparability due to the heterogeneity of the contexts studied (stratified and unstratified sites; stratigraphic units or site assemblages) or due to dealing with sites which are often very far apart. The present study attempts to minimise these obstacles by selecting for analysis only data from recently excavated contexts which are also sufficiently rich in finds. Accordingly, some 30 sites were selected, including half-a-dozen open settlements.
The method adopted consisted of classifying and grouping the settlements in chronological order based on the percentages of each category of material considered to be chronologically significant -specifically, amphorae, imported pottery, brooches, ceramics and coins. Groups of sites were defined, characterised by identical populations of material, both by type and by proportion. To counter the difficulties of identifying very fragmented Italian amphorae from settlements, a method of classification based on rim height was devised; statistically it can be postulated that the relative proportions of the different height classes can be interpreted in chronological terms. By comparing all these different classifications, a series of chronological phases can be established, each defined by contexts yielding comparable assemblages of material.
4. Periodization
Five chronological phases are
characterised by five groups of
assemblages in which some or all of the
following material may be present:
- Phase 1 : Imports (present only in small numbers): Campanian A, more rarely Campanian B, Graeco-Italic and Dressel1 amphorae (the latter less common), pale-coloured coarse wares, Ampurian pottery; brooches of La Tène II construction, accompanied in some instances by filiform and / or Nauheim types; lignite and glass personal ornaments (notably Haevernick type 7 and 8 bracelets), absence of beads of Haevernick types 23, 24 and 25; reduced pottery with post-firing oxidisation (mode A) or irregularly-fired (mode B primitif), predominantly coarse -or semi-fine tempered, generally handmade; coins of gold, silver and potin, but only in small numbers, some of the potin series copy the small Massiliote bronzes with a butting bull; in central and east-central France, Massiliote obols or their imitations also occur.
- Phase 2 : Campanian A and B (absence of Campanian A dating to the first half of the second century BC) ; an overwhelming majority of Dressel 1 amphorae with short rims (less than 4.5 cm in height), a few Dressel 1B; among the brooches (all of which have uncovered springs), the Nauheim type predominates; among glass personal ornaments, plain glass bracelets (particularly Haevernick type 2) predominate; appearance of copies of Campanian A pottery; introduction of struck bronze coinage and of silver coinages based on the weight standard of the Roman denarius.
- Phase 3 : predominance of brooches of La Tène III construction with uncovered springs, introduction of brooches with moulded bows (Almgren type 65 and its variants), spoonbow brooches, etc; rare examples of brooches with hooks or wings (eg collared brooches); amphorae principally Dressel 1 (A, Band C), with occasional examples of less common types; Campanian Band its derivatives (Campanian A becomes rare); disappearance of Ampurian pottery; appearance of lamps, occurrence of coin types certainly issued around the mid first century BC, as well as of Roman coins which were previously very rare.
- Phase 4 : increased presence of amphora types other than Dressel 1, Dressel 1 with short rim (less than 35 mm) much less common; appearance of vases à paroi fine, platters with internal red slip, Campanian pottery and its derivatives; development of brooches with hooks, introduction of hinged brooches (Alésia type); development of Campanian B imitations; decline of carinated bowls, except in Belgic Gaul and Armorica; reduced pottery (grey wares) accounts for an increasingly significant proportion of the fine ware assemblage.
- Phase 5 : Terra sigillata and its imitations (terra nigra and other products), vases à paroi fine, including ACO beakers; increasing variety of amphora types; decline or disappearance of simple bow brooches in favour of brooches with hooks, especially types with a covered spring or a hinge; incorporation into the ceramic repertoire of various vessel forms of Mediterranean origin; occurrence of coin types issued after 30 BC and more especially after 15 BC.
Phase 5 comprises two successive stages which are distinguished principally by the types of imported or mediterranean style pottery present: Italian terra sigillata and its relatives and earlier varieties of vases à paroi fine (beakers with studded decoration, with concave rims and ACO beakers) in the earlier stage; southern Gaulish sigillata and local imitations of vases à paroi fine in the latter. It is during the second of these stages that brooches with covered springs or hinges become the commonest types, while the forms which first appeared at the beginning of La Tène D - Nauheims and filiforms - finally disappear.
The increased number of reference sites (notably the Roman forts on the Rhine ' frontier) allow the absolute dating of this last phase to be established with greater precision: Phase Sa begins' around 20 BC at the latest, Phase 5b dates to the first century AD, the transition occurring shortly before the turn of the millennium. Phases 3 and 4, with Dressel 1 B amphorae and later varieties,- centre on the first century BC, the transition from Phase 3 to Phase 4 occurring in the middle of the century. Several dendrochronological dates place the start of Phase 2 before the end of the second century BC and suggest that its attributes were already well established by the last quarter of that century; Phase 2 certainly ends well before the mid first century Be. Phase 1 lies entirely within the second century BC; its starting date is difficult to establish, but should be before the middle of the century, as shown by the presence of Campanian A with palmette decoration and the large number of Graeco-Italic amphorae from Levroux.
Comparison of this periodization with other regional chronologies (those established by A. Haffner and A. Miron for the Trier region, by J. Metzler for the Titelberg area, by P. Pion for the Aisne Valley, and by V. Guichard for Forez) fully supports the chronological divisions proposed for Phases 2 to 5 :
Phase 2 equates to Miron's La Tène D1b ; Pion's Etape 3
Phase 3 equates to Miron and Metzler's La Tène D2a ; Pion's Etape 4
Phase 4 equates to Miron and Metzler's La Tène D2b ; Pion's Etape 5
Phase 5a equates to Pion's Etape 6 ; Haffner's early Gallo-Roman horizon and Metzler's early Gallo-Roman (GR1).
Correlations for Phase 1 are, however, much less easy to establish; it evidently overlaps La Tène C2 and Pion's Etape 1, but it also includes elements of Miron's La Tène D1a and Pion's Etape 2. Among the chosen sites, there is no context which enables these different periods to be distinguished (as is now being attempted for sites like Levroux).
The validity of this periodization is then , tested by examining other sites not previously considered. This exercise demonstrates that the periodization works well in central and east-central France, in eastern Picardy, in the Aisne Valley, in Champagne and in the southwest - those regions where the majority of sites selected for chronological analysis are located. On the other hand, western France fits much less well, as its material culture includes little or no metalwork, coins, or imported pottery and its local imitations. The difficulty of integrating this latter region into the proposed chronological framework presumably reflects genuine cultural, socio-economic and political differences.
5. Historical overview
Mapping the settlements belonging to the 5 phases (all the selected sites had evidence of at least one specialised craft activity and / or a commercial role) highlights not only how widely oppida were distributed but also reveals contrasts between certain regions.
The sites of Phase 1 were almost entirely open settlements, confirming that the characteristics of La Tène Finale first gained hold in unfortified settlements; such sites are markedly more numerous south of the Loire and in east-central France than in the north. In Phase 2, the first oppida appear, as well as new undefended villages. Phases 3 and 4 represent the peak of the phenomenon, although some open villages disappear. From Phase 5 onward, the oppida dominated system starts to break up. Gaul is thus distinguished somewhat from the rest of the Celtic world by the late appearance of its oppida (apart from a few examples which may date back to Phase 1, the vast majority of Gaulish oppida do not appear before Phase 2: the end of the second / start of the first century BC) and by their longevity, as well as by the importance of open villages in the settlement pattern of the most «urbanised» zones.
All five phases are represented in central and east-central France as well as to some extent in the south-west; there, as many large open villages specialising in artisanal and commercial activities are found as fortified sites. Monetary developments in this enormous zone are characterised by a progressive alignment of its silver coinages on the weight standard of the Roman denarius, starting well before the Conquest. This zone was in receipt of numerous and varied imports, at least from the beginning of Phase 2. This is also the zone where Caesar describes the most highly developed administrative organisation and political institutions. In contrast to this image of a region which looked towards the Mediterranean from an early date, western Belgic Gaul and above all western Gaul are characterised by a near total absence of large open villages until at least Phase 3, by the short-lived existence of its oppida, by the rarity of imported pottery -which moreover only appears at a late date and by the continued existence of a monetary system based on the stater standard until at least the Conquest. Settlement evolution is thus neither synchronous nor homogenous from one part of Gaul to another, and the variability of the evidence testifies to differing degrees of economic and political development from place to place.
Although the oppida -like the open settlements- yield evidence for varied craft activities together with large numbers of imports, their creation cannot be considered a simple progression from the large open villages. Nor can their defensive function any longer be considered the prime mover. In reality, the oppida were the product of a centralised political and economic system, the origins of which are interwoven with the emergence of the civitas, in which they played the role of central places, exercising a range of urban functions. This role was embodied in a traditional form, generally that of the hilltop fortification, the distinctive means of expressing the power of the social group among protohistoric peoples.
The origin of the socio-economic transformations which culminated in the creation of oppida remains obscure. It pre-supposes in any case a climate of economic growth in which increased agricultural production permitted the development of craft specialisation and commerce. Contacts with the Mediterranean world played a complex role in this process of change. The penetration into Gaul, during Phase 1, of both Mediterranean goods and coinage (Massiliote obols) does not coincide with a general move towards the establishment of oppida -a process already well under way in the rest of Celtic world"':' yet the early opening up of the Mediterranean world may well have influenced the process of urbanisation by favouring the development of unfortified settlements in certain regions (central, east-central and south-western France). Analysis by region shows in any case that this process takes diverse paths and progresses at different rates in different places.
The Roman conquest of Gaul did not thwart the development of oppida; on the contrary, most of them enjoyed considerable prosperity during the second half of the first century Be. Only in the last quarter of the century, coinciding with the administrative reforms of Augustus, the creation of the first towns and the construction of the first Roman roads, was this equilibrium upset: whilst many oppida disappear at the end of phase 5, no doubt partly due to their distance from routes of communication, almost all of the lowland open settlements continue into the Gallo-Roman period, sometimes with a slight shift in the focus of occupation.
