Version classiqueVersion mobile

Rodumna (Roanne, Loire), le village gaulois

 | 
Marie-Odile Lavendhomme
, 
Vincent Guichard

Abstract

Texte intégral

1 The site

1.1 Geography and history

1The only ancient references to Roanne are those in the Peutinger Table (Roidomna), and in Ptolomy (Rodumna), which indicate that it is a town of the Segusiavi, together with their capital, Forum Segusiavorum (Feurs). At the time of the conquest, the Segusiavi constituted an autonomous political entity within the Ædui sphere of influence. The heart of their territory is the Forez plain (fig. 1). Roanne is at the center of its own distinct area, the Roannais plain, situated further to the north.

2Roanne is at the point where the Loire is no longer navigable and at the junction of several cross-country routes whose existence is certain for the Roman period and assumed for the preceding. Outside the town itself, the prehistoric period is scarcely documented. There have been sporadic discoveries for the early periods of the Iron Age (before the 2nd century BC). On the other hand, there is another important late Iron Age (LIA) site, the oppidum of Jœuvre, which occupies the top of an escarped hill in a vast meander of the Loire (70 ha), 10 km to the south of Roanne (fig. 2). Its fortifications have been identified, as a vallum covering an older murus gallicus, but the précisé extent of its occupation remains generally unknown in the absence of archaeological exploration.

1.2 Topography of the site

3The Iron Age seulement of Roanne is situated at the heart of the plain, on the left bank of the Loire and on top of a sandy terrace, surrounded by the main riverbed to the east and a marshy zone to the west (fig. 4). Its main cemetery also lay alongside the river, 800 meters to the south. The site probably developed at a road junction whose existence is certain in the Roman period. It is known that there was organised continuous occupation from the 2nd century BC, taking the form of a modest nucleated settlement whose extent does not appear to exceed four hectares in the middle of the 1st century BC. Its size then increases quickly to reaches its maximum area (30 ha) from the Augustan period to the middle of the 1st century AD, between which there is no fundamental reorganization of its plan, which was dictated to the routes passing through it.

1.3 Excavations

4The importance of the early occupation of Roanne has been well known since the last century. Yet excavation of the LIA settlement was not started until the end of the 1950s. The publication of the first excavation (Saint-Joseph) by M. Bessou in 1976 marked a turning point in our knowledge of the LIA in central France, because of the tight chronology he proposed for the 1st century BC – it actually appears, twenty years later, that the span of time covered by the artifacts is much longer and that it starts a century earlier. This monograph describes another excavation (Saint-Paul) carried out in 1987 and immediately adjacent to Bessou’s and next to another opened in the 1970s (Rue Gilbertès). The total surface area uncovered was small (less than 1500 m2) but it has produced important information, through the abundance of the artefacts excavated and the quality of the stratigraphy. This volume presents a general synthesis of the results of the excavations of the last thirty years.

1.4 Limits of the period studied

5The upper limit of this study is determined by the artefacts, which do not predate the turn of the 2nd and 3rd centuries BC, with some minor exceptions (cf. infra). On the other hand, the lower limit is arbitrary. In the absence of a stratigraphic break, it has been fixed at the moment when several new categories of ceramics appear which are typical for the 1st century AD.

1.5 The earliest Iron Age Settlement

6The structures of the 2nd and 1st centuries BC are preceded by earlier occupation found at the base of the stratigraphy. It is very diffuse and totally lacking in features. It only produces very dispersed and fragmentary finds which seem to cover a long period including the 4th and 3rd centuries BC. Worth of note are several coins (silver minims originating from Marseilles) and fragments of painted flagons originating in the Rhône Valley.

2 Settlements structures and stratigraphy

2.1 Stages of development of the settlement

7The excavation at Saint-Paul shows a particularly clear stratigraphy. We can identify three phases of occupation dating back to the 2nd and 1st centuries BC. Phase 1 corresponds to a slow rise in the surface, due to several reconstructions of houses on the same spot; phase 2 to the establishment of general dumping and terracing ; and phase 3 to the resiting of the settlement on a different orientation which is then preserved throughout the Roman period. The study of the finds also allows us to distinguish six chronological horizons which can be closely linked with the stratigraphy (table i).

2.2 Organisation of the settlement and stratigraphy

8For phase 1, the excavations have produced evidence of a row of rectangular buildings, probably backing on to an alley and opening on to an open space for domestic activities (fig. 12). Three or four phases can be distinguished, which represent many rebuildings with almost identical plans. After the brief episode of dumping of phase 2, other houses were constructed on a different orientation with an east-west axis initially marked by a palisade, and later by a stone wall (fig. 21). The architecture is without doubt more variable in phase 3, but it is not possible to produce a single house plan due to the number of disturbances from the Roman period.

9The continuity of the stratigraphy shows that the LIA settlement, already in existence at the beginning of the 2nd century BC, was not greatly affected by the appearance (at the start of the 1st century BC) of an oppidum nearby – contrary to what we observe in other cases in central and eastern Gaul (e.g. Levroux, Basel and in the région of Clermont-Ferrand) where the foundation of the oppidum was accompanied by the disappearance of pre-existing sites in the plain. This unusual situation (shared by the regional sites of Feurs and Goincet, which also form part of the territory of the Segusiavi) allows us to follow clearly the evolution of the nucleated settlement over a long period of time: systematic organisation of the settlement as early as the first half of the 2nd century BC and reorganization in the first half of the following century, which remained unaffected by Romanisation.

2.3 Settlement structures

10The ephemeral buildings of phase 1 are particularly well preserved, when one considers the nature of the building material. They are one-roomed rectangular cells (4 x6 m). Their limits are identified by a thin wall of daub anchored directly into the soil. The nature of the superstructure is uncertain. The floor, of beaten earth, is sometimes laid on a foundation of potsherds or gravel. The doors generally open at the gable end. A rectangular hearth at floor-level occupies the center of the house. Small domed ovens, movable fired clay braziers and large external hearths complete the range of heat sources. The interior organisation includes ceramic pots set in the ground. It is also characterised by the presence of numerous burials of foetuses and neonatal infants. The structures around the houses comprise a well, a granary (?) and a number of pits whose original function can not be exactly identified.

11In the construction of phase 3, sherds of Italie wine amphora are intensively used, either as a base for floors or as a foundation for walls. Construction techniques of Mediterranean origin (roof tiles made of baked clay, lime mortar) make their appearance in the last quarter of the 1st century BC. The variation in buildings and other structures suggests a greater range of functions than that of their predecessors of phase 1, but these can not be precisely defined. We can note in particular two buildings standing on heavy posts, a storeroom paved with roof tiles and several pits with highly variable profiles and fills.

3 The finds

3.1 Relative chronology

12The first objective of the study of the finds was to establish a chronology. For this, twenty-six particularly rich assemblages (pit fills) were selected and classified by seriation. In order to facilitate the reading of this section of the text, the assemblages have been re-ordered according to the results (F1-F26) and the finds displayed in the same order on plates 1-95, at the end of this report.

13Only the ceramics were selected for the statistical study. The other categories of finds were too variable to allow integration into the same calculations and are only used as a check on the validity of the classification afterwards. The numerous ceramics are divided into twenty categories corresponding to well-defined production groups (Campanian, samian, etc.) or to groups defined by their technical characteristics (type of firing, methods of fabrication, preparation of clay etc.) (table vii, no 2). The algorithm used is correspondence analysis. The resulting graphs (fig. 63) show that a good seriation is effectively obtained with, at first glance, three chronological stages which correspond to the three periods defined by Bessou. Each period is characterised by the presence of distinctive ceramics from the principal categories (fig. 65). A doser analysis shows two gaps in our ceramic sequence: one between periods 1 and 2, and another between periods 2 and 3. The first hiatus may, however, be filled by the finds from the site of Feurs. A subdivision of each main period is proposed, to which we will assign the term « horizon ». The stratigraphy (fig. 66) is fully compatible with the results.

3.2 The pottery

14At first glance, we can distinguish well-characterized groups which generally correspond to Mediterranean imports and to vessels inspired by Mediterranean styles. The remaining pottery, of (supposedly) regional manufacture, clearly the commonest category, is given special consideration. It is submitted to the detailed typological classification similar to that already developed for the study of the finds from the LIA site of Feurs (Vaginay, Guichard 1988). At its first level, this classification takes into account technical parameters (quality of the clay, method of assembly, presence of slip, and method of firing), and secondly uses hierarchical morphological criteria (general form of the container, aperture opening index, form of the rim, etc).

15The evolution of the regional ceramic style is rapid and matches perfectly with that witnessed at Feurs (fig. 121). In effect, each horizon presents specific characteristics (fig. 65 and fig. 123128). The most striking feature is the slow disappearance of handmade pottery, replaced by wheel-turned pottery. The latter is mainly fired in a reducing-oxydizing atmosphere (« A ») (at the end of the 2nd century and the beginning of the 1st century BC), then in a reducing atmosphère (« B »).

16The imported pottery forms an increasing percentage of the assemblage and witnesses a diversification of the kilns providing it. Up to horizon 3, there is only the regular importation of Campanian wares, Lower Rhône Valley wares (with a typical light, fine fabric) and (more rarely) grey Catalan jugs. Later thin-walled beakers appear (from horizon 4) and Italie samian (from horizon 5) then. The regional products imitating Mediterranean forms become more and more frequent, until there is a complete change of the assemblage. This starts during horizon 2 with imitation Campanian, in tum replaced, in horizon 5, by imitation samian. At the same time, beakers and Jugs also appear. More generally, a radical change is observed in the local pottery assemblage between horizons 4 and 5 – i.e an improvement in the manufacturing techniques, the abandonment of the decorative techniques of the 2nd century BC (painting, burnishing and stamping) in favour of other more mechanical techniques (rouletting, comb-decoration). This transformation, which starts a half-century before the conquest, shows evidence for the reorganization of the modes of production, which are becoming more centralised and more concerned with profit. Particular mention must be made of the so-called Besançon wares, the only category of indigenous pottery which witnesses a large-scale commercial organisation in the 2nd century BC. These are handmade containers of a fabric containing large feldspar inclusions with a distinctive finish (lip moulded with a characteristic groove, coated in mica). They are considered as imports from the Morvan and are very common in all periods. The reasons for their popularity are uncertain: they seem to have been used mainly as cooking pots, and perhaps also as transport vessels.

17There are large numbers of amphorae, the largest group being wine amphorae of Roman Republican type (Dressel 1). They are present mainly from horizon 3 onwards. Becoming rare by horizon 6, they are not replaced by other types. Despite the difficulties posed by their typology, the evolution of the shape of the rims tends to confirm the seriation carried out for the other ceramics. Equally, the results provide us with a synchronisation of the local chronology with other LIA sites.

3.3 The instrumentum

18This term groups together a highly varied selection of objects: Personal ornaments, metal vessels, weapons, tools, hardware and the evidence of craft activities. The assemblage of ornament is characterised throughout by the scarcity of glass objects (beads and bracelets) in relation to those in metal (essentially brooches). The large collection of brooches (86 ex.) shows the common use made of iron and the existence of regional types (especially one with the bow covering the spring). Filiform brooches of La Tène II construction, always in iron, characterise horizons 1 and 2. Those with a frame foot of La Tène III construction (filiform, covered-spring and Nauheim types) appear in horizon 3. The later horizons are less well documented (fig. 129),

19Weapons and, more surprisingly, tools as well, are very scarce. The only clearly identifiable craft production is that of weaving, with many loom-weights. There is also slight evidence of metal working (lost-wax and shaped moulds, and occasionaly slag). The manufacture of pottery is highly probable, based on distinctive ceramic styles which contrast with those from the site of Feurs and on the importance of this craft industry at Roanne in the 1st century AD.

3.4 The coins

20The collection of Gallic coins found in the site is large (250). it includes, exceptionally, a large number of examples which were found in datable archaeological contexts. The distribution ties in unambiguously with the Aedui-Sequani domain, characterised by high percentages of potin coins « à la grosse tête » (57 % of the identified coins). The chronological data (fig. 133) show the existence of an initial phase of circulation (3rd century) characterised by Marseilles minims. In a later phase (horizon 1, first half of 2nd century), imitations (regional?) are added to the minims. Cast bronze coins (« potin ») appear in horizon 2 and dominate the circulation until horizon 6, when Roman coinage starts competing with them.

21The supply source of the coins does not change recognisably until the massive introduction of Roman coins (fig. 131). The regional division which characterises the usage of coins in LIA Gaul is maintained throughout the duration of the 1st century BC.

3.5 The animal bones

22Thirty bone assemblages were studied, which totaled more than 14,000 individual fragments. Overall, the distribution of the large domestic animals presents the usual patterns encountered in temperate Gaul on large LIA settlements, despite the over-representation of cattle and the underrepresentation of sheep (fig. 155). It is also possible to observe a regular regression on the part of pigs. The cattle and the pigs show, in the same period, a larger size than that reached in the north of France, which does not seem to be due to the introduction of foreign animals of better breed. An example of large-sized imported horse is nevertheless noted, in horizon 4. The mode of selection of the animals for butchering (importance of beef, frequency of old individuals) gives an image of poor meat diet.

4 Synthesis

4.2 The chronology

23This section first considers the difficulty of linking distant regional chronologies – due to the variability of regional assemblages, especially ceramics. The finds used primarily for making synchronisation are the Mediterranean imports, the coins and the Personal ornaments, with the assumption that an object is used simultaneously in all the regions where it appears, unless there is explicit proof to the contrary.

24The chronological System proposed, divided into six horizons, is more precise than that traditionally used, which distinguishes only four stages during the period 2nd-1st centuries BC (La Tene C2, D1, D2 and Augustan). Within this general framework, more precise chronologies have been proposed within diverse regions (e.g. the north-east of the Parisian Basin, the central Rhine Valley, the Swiss plateau), but it is difficult to establish a precise correspondence with the System set up for the Segusiavi area.

25Using varied fixed points (the dating of imported Mediterranean pottery and of the coins deposited on sites providing with a terminus ante quem such as Carthage, Alesia or the camps of the Germanic limes), an approximate date for all horizons can be proposed. These dates are slightly older than those previously proposed for the nearby site of Feurs (Vaginay, Guichard 1988). They are more consistent with tree ring dating now available on several sites of eastern Gaul (table i).

4.2 The economy of the settlement

26The data permit us to document the rise of a monetary economy, which is achieved by the end of the 2nd century BC. From this point of view, romanisation seems to bring in some decline in the use of coins, with the abandonment of the lightest base-metal types (with a subsequent drop in coin losses). The intensification of the relations with the Mediterranean world is equally well documented. It becomes really important at the end of the 2nd century with the start of massive importations of Italian wine. The « transverse » exchanges between the regions of central Gaul are much more difficult to identify. Nevertheless, there is evidence of large-scale exchange for two particularly significate items: querns, probably imported from the Morvan and the Auvergne from the beginning of the 2nd century; Besançon wares, equally originating in the Morvan. We can perhaps add glass ornaments, which are too scarce to suggest a local production.

27The subsistence economy is indicated only by the evidence of archaeozoology. This shows a type of consumption characteristic of the larger LIA settlements. The industrial production attested by excavation is limited (weaving is the main one) but this is probably due to the small area of excavation. The detailed study of the pottery, which is presumed to have been mainly made locally, shows indirect evidence of some aspects of the evolution of industrial activities. Evidence of mass production for the market only appears at the beginning of the 1st century BC, as documented by improvements in the techniques of manufacture and standardisation in styles (accompanied by a decline in the range of decorative techniques). It may even be assumed that the same period witnesses the setting up of some craft workshops such as are known in great number in central Gaul in the following century.

4.3 The material culture of the Segusiavi

28The composition of the coin losses attaches Roanne (and more generally the territory of the Segusiavi) to the sphere of the Ædui. This confirms the assertion of Caesar that the Segusiavi were dependent on their strong neighbours. The local pottery also shows evidence of the cultural community of the sites lying within the presumed territory of this people, which includes the Roannais plain and its marches, the Forez plain and also Lyon and a part of its surroundings. This community is discernable from the 2nd century BC. At this time, the pottery style is clearly distinguishable from that of the adjacent areas of the north of the Massif Central, in particular the Auvergne. This does not mean that these various styles did not evolve in parallel: they share several cultural characteristics, like the major development of painted ceramics in the 2nd century BC.

29The impact of the foundation of an oppidum in the territory of the Roanne, probably right at the beginning of the 1st century BC, is difficult to define. The archaeological data suggest a decrease in activity on the site during the 1st century, but this is not precisely measurable. Indeed, it is striking that the site is not abandoned, as happens elsewhere in central and eastern Gaul. On the contrary, it develops rapidly from the beginning of the Augustan period. Meanwhile, the oppidum appears to be abandoned shortly after the conquest. This is a general scenario, common for the whole territory of the Segusiavi, characterised by a brief episode of oppida construction and a much more significant importance of the sites founded in the 3rd/2nd centuries BC on the subsequent Gallo-Roman settlement pattern. More precisely, the structure and the economy of the Gallo-Roman site of Roanne seems to be already fixed to a large extent at the time of Romanisation.

Le texte et les autres éléments (illustrations, fichiers annexes importés) sont sous Licence OpenEdition Books, sauf mention contraire.

Cette publication numérique est issue d’un traitement automatique par reconnaissance optique de caractères.

Lire

Open access

Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search