dAf 76
CHARDRON-PICAULT Pascale et PERNOT Michel
Un quartier antique d'artisanat métallurgique à Autun.
Le site du Lycée militaire
Vestiges of Gallo-Roman workshops that were in activity between the 1st and 3rd centuries AD in the city of Augustodunum were found in the course of work carried out at the Lycée militaire in Autun. The workshop buildings are described in detail and situated in their urban setting. The different crafts activities are studied closely: metalworking with copper and iron alloys, ceramics, crafts of bone. This work has enhanced our knowledge of the chain of production in metallurgical crafts in particular. Fresh data on the manufacture of common objects such as brass fibulae and bells of bronze coated iron sheet is obtained from the analysis of mould fragments, of furnace remains and of forge slag and waste.
Crafts emerge as a major feature of the economy in the city of the Ædui, according to processes already identified at Bibracte.
The rich documentation of this work constitutes a significant contribution to research on crafts in ancient times.
Abstract
Abstract
The archaeological excavation of the site of the Lycée militaire covered an area of 9500 m2, located in the south-eastern section of the city, inside the walls, near the theatre and bordering the ramparts. The remains exhumed date exclusively from the Gallo-Roman period. The space is structured by the rampart wall, by a decumanus and two cardines that terminate with two of the towers on the enclosure wall. Five lots have been distinguished. The extent of the three southern lots was limited by the rampart wall. Of the five lots, the middle one (lot C) was entirely excavated, the four others (lots A, B, D and E) only partially. The relatively steep slope of the terrain led the occupants to make terraces; as a result the quality of preservation of the vestiges is highly variable : the uppermost zones are greatly eroded, while at the foot of some of the terrace walls the subsisting stratigraphic layer attains 1,5 meters.
The streets are remarkably wide: 11 to 12 meters. The roadway structure can be described thanks to cross-sectional cuts, and it can be seen that the roads were not superimposed on an earlier structure. The entire section that lies alongside the rampart wall could not be excavated; the two wide trenches made close to the towers did not yield any decisive information for establishing a relative chronology for the construction of the rampart wall and the roadways. ln contrast, the traces of the implantation of ail these features clearly predate ail buildings.
The stratigraphic data on the occupation of the lots are most noteworthy for their illustration of actual conditions at the site: there are many series of numerous thin layers that are the result of changes in crafts activities. Three major states of occupation can be defined using relative chronology diagrams, schematically outlined; the boundary layers, marked by architectural changes, are generally accompanied by changes in activities. Alterations in the configuration of a room can be used to divide a period into phases. The initial periods of occupation (phase 0) and the time of abandonment of the site were given particular attention.
In each lot the benchmarks of absolute chronology are provided by a study of fine ceramics. Given the small size of the sample, associated with a significant proportion of residual material, the conclusions drawn must be interpreted with prudence. The principal conclusions are supported by the analysis of 83 coins found in exploitable contexts.
In the second part of the study, the general floor plans of the buildings are outlined for the different chronological periods. Ali of the structures excavated are described in detail- walls, recovery ditches, remains of ovens or pits, to give just a few examples. Generally speaking the findings are given in succession for the various chronological states of occupation for each lot. Lot C is the exception, and the three distinct buildings are presented individually. Whenever possible an initial interpretation of the functional use is given for each room (an enclosed and covered unitary space) during each period, on the basis of structures and objects brought to light. The crafts activities discovered are, in order of importance: metalworking of copper base alloys, ironwork and pottery. In most instances the activity of a given place changed with the chronological period.
In general there are sidewalks, spaces for circulation between the buildings and the roadways. Because of the lack of chronological links, they are studied separately from the built environment. Little data remain concerning the finishing of the building facades, due to the eroded state of the remains. Porticoes over the sidewalks must have existed along some of the roads.
Questions persist concerning the location of the spaces used by the craftsmen for their domestic pursuits, which appear to have occupied only a small part of the ground-level surface area. The presence of stair wells and vestiges of fallen wall paintings are evidence that an upper storey was erected over some of the ground-level rooms. In lot C it is certain that there was a correspondence between the erection of upper storeys and the lever of terracing.
The chapter devoted to crafts activities opens with the study of the remains of the activity that was predominant throughout the occupation of the site: the manufacture of small objects made of copper base alloys. The main categories of objects are directly related to this production, and are analysed following a short review of the techniques used during this period.
The whole set of recovered fragments of lost-wax moulds weighs a total of 280 kg. Among those pattern cavities that can be interpreted, the vast majority correspond to the main part of two-part spring-clasp fibulae. One mould, of which two-thirds remain, was found abandoned before casting. Five different models of fibulae have been recorded; for several models there are two classes of modules, in addition to variants. Large fibulae with a 50 mm wide spring-plate were produced in large quantities at the site. The pattern cavities of small bells, with circular or quadrangular bases, have also been identified in large numbers; the mouth opening measures between 40 and 100 mm. No vestige related to statuary was identified, nor for hinged fibulae. Several dozen chunks of limestone (over 30 kg in all) have been interpreted as fragments of permanent moulds made up of several pieces. The legible pattern cavities (over 50) indicate that Some of the items made were dishware; the maximum diameter recorded is 300 mm. Crucibles, generally highly fragmented, were also recovered in great quantity (650 kg), along with lids. Different shapes are present, and at least three sizes, containing between 0.15 and 1.5 litres, corresponding to weights between 1 and 12 kg. Among the metallic objects made of non-ferrous alloys, only the rough castings, scraps and waste which are evidence of the work done locally were studied. Forty compositional analyses were carried out. The copper alloys fall into five groups; out of 11 fibulae abandoned in the course of manufacture, the six pieces analysed are made of brass. The white metals analysed are unalloyed lead or a lead-tin alloy. Small stone tools (polishing tools and stakes) have been assigned an initial classification on the basis of a short study. The manufacture of bells made of bronze coated iron sheet is attested by the presence of fragments of earthen shells (over 60 kg) used for this production. A quick study of the fired clays yields some information on the recipes used by the bronzesmiths.
Forty furnaces used for working the copper alloys were examined, and grouped into five categories. The other types of structures related to this craft (floors, pits, buried pottery, postholes, etc.) were systematically examined. The correlation of structural data and information on materials drawn from 38 cases supports and justifies the notion of a workshop. Two types are distinguished; casting and post-casting activities are found in the first type (FPF), while only post-casting activities occur in the second type (PF).
Most of the remains of ironwork found at the site can be interpreted as the results of work coming after reduction of the ore. Forge work is clearly attested by kiln and anvil structures and by typical wastes (slag and scale); the forge was active during periods 2 and 3. The disposition of the workshop space is described for seven cases; the differences observed lead to the assumption that there was a certain degree of specialisation among the crafts workers, even if the type of production is not known. The iron items, only superficially studied, show evidence of rough forms and scraps; these findings may in the future be used to specify the nature of some of the work produced.
Two potters' workshops were in operation during period 2; kilns and rubbish heaps have been found. A preliminary assessment of the production carried out at the site can be based on the pieces spoiled in firing found in the rubble which filled in these structures. This production was primarily dishware for domestic use, common in 2nd and 3rd century Gaul; ordinary light-coloured ceramics of non-limestone clay, fired in mode A, and ordinary grey ceramics, fired in mode B. No fewer than 544 vases were examined, and over 30 shapes are represented; oscillum fragments in the form of masks were also discovered. This pottery appears to have been produced in a relatively modest context, without industrial characteristics.
Boneworking is evidenced by waste and rejects, but no workshop was found within the extent of the excavations. The types of objects manufactured could not be identified either. Use of goat and sheep horn is attested by the dumping of a hundred or 50 bone pegs found in the vicinity of an oven in a metalworking shop.
Anthracological study of charcoal samples taken for the most part in areas of crafts activity reveals that changes occurred between period 1 and the following periods: the use of oak drops off considerably, giving way to beechwood.
The detailed excavation of this metalworking site has yielded exceptionally important data, by the quantity of items preserved and the structures recorded, but above all by the quality of the documentation on the associations between these two types of vestiges. The "technical" results lead to "social" findings. It is clear, for example, that the artisans worked in small groups, in the context of a competitive economy. In addition, despite changes in the materials used (passing from copper alloys to iron to pottery), the crafts-based nature of the district persisted throughout the period of occupation, over two centuries. The crafts workers of the Lycée militaire site can thus be situated in relation to the legacy of Bibracte and to what is known of the city of Autun. It is now clearly established that crafts activity was a major component of the economy of Autun, capital of a Roman civitas.
