dAf 19
GADY Serge
Les souterrains médiévaux du Limousin
Approche méthodologique
Souterrains have in the past been the object of both varied and fanciful interpretations. Considered at different times to be places of refuge or worship, they have been attribued to periods ranging from pre‑history to the French Revolution. The present study, first written as a doctoral thesis, examines 94 souterrains in the Limousin region and analyses their geographical and historical significance. Research has included excavation of the Bois~du‑Mont souterrain at Bessines‑sur-Gartempe in the Haute‑Vienne department. The result of the study proves that these underground chambers are related to rural settlement during IXth to XlIIth centuries. Archaeology provides firm evidence for when these structures were tunnelled, occupied and abandoned. The disciplined methodological approach of this work is exemplary in a field where scientific analysis has so far been virtually non‑existent.
Abstract
Abstract
PART ONE
Underground chambers in the Gartemp basin
The subject of this study is the Marche, a region mentioned in Carolingian texts and located in the north of the Limousin. Throughout their history the inhabitants of essentially rural area have had to adapt their farming and rearing methods to the cold thin soils wich result from the omnipresent crystalline bed-rock. Medieval rural settlement has been studied over an area of approximately 1 500 square kilometres in the western part of the Marche territory, within the hydrographic basin of the River Gartempe, itself a tributary of the River Creuse.
Ninety-four sites known to include underground chambers or storage pits are examined. These archaeological remains are often already known to local savants but have never been investigated methodically. Existing hypotheses concerning their origins are thus varied and involve every period of history. It was deemed imperative to establish a methodology based upon cartographic analysis before carrying out further research : the geographical, historical and archaeological environment of each site would be considered using numerical criteria.
The survey has demonstrated that the sites studied are located in the same geographical and hydrographical environment as medieval rural settlements, as shown by the following shared features :
- settlements of the higher ground on slopes with good southern and eastern exposures :
- settlements are not consistently close to large waterways :
- no apparent concern for occupation of strategic sites such as overlooks or the confluences of rivers :
- the relationships between underground chambers and defensive or religious structures were not premeditated, but rather the result of successive occupation of the same site;
- when these sites are compared to medieval rural habitat, on the other hand, they are seen to be functionally and chronologically identical. Archaeological evidence, written sources and toponymy reinforce this interpretation.
PART TWO
The Bois-du-Mont souterrain : applying an analytical method
From 1970 to 1977 excavations were carried out on the Bois-du-Mont souterrain (in the district of Bessines-sur-Gartempe) as part of an archaeological research project in the area. The north-west zone of the souterrain was chosen as the test site.
Architecture
The chamber's architecture comprise four basic units :
Unit I : a 5-metre-long corridor which slopes down from a narrow entrance at ground level ; its inclined wall is comprised of stone blocks, several of which jut out to form a rough set of steps. This wall retains the mass of earth used to backfill the trench originally used to dig out this part of the chamber. Two cells, of about two square metres in size, and a bench had been added. The end of the corridor opposite the entrance would have been closed off since grooves in the wall imply the presence of wooden door-frames.
Unit II : the souterrain's central chamber, 3,5 metres long and 1,5 metres wide was dug out using a side-shaft which was blocked off when excavation was completed.
Unit III : a room was later added to the central chamber ; it was only accessible through a narrow passage 50 cms in diameter. As with unit Il, unit III was dug out using a side-shaft blocked off when work was completed. Notches cut into the room's vaulting probably held a hanging shelf.
Unit IV : a 12-metre-long corridor connects unit Il and the rest of the north-west souterrain area with the as yet unstudied south-east zone. A separate entrance links this structure with the outside.
Stratigraphy
The stratigraphy of the site reveals the presence of both occupation layers and layers of primarily natural deposits.
The occupation sequences appear to be cyclical : a carefully laid floor layer is covered by an occupation layer comprised of organic material, primarily ash, and abundant pottery sherds. Four successive occupation phases can be identified in the central chamber and the two corridors (units I and IV). These same layers are not present in the side-room. The accumulation of layers slowly covered over certain architectural features and modified the way that the interior space was used.
The initially available height beneath the vaulting of 1,25 metres diminished as layers accumulated. The occupants attempted to prevent this loss of space by scraping away rock from the vaulting and spreading it on the ground to make a new floor rather than fetching additional materials from the outside. At the end of the final occupation phase the entrances to the souterrain were closed off. With time soil has filtered in and sealed the stratigraphic sequence. Two other entrances had been deliberately back-filled following the construction of the chamber.
The finds
Pottery sherds, which constitute the majority of the archaeological finds, are present in ail occupation layers. An analysis of the distribution of sherds from reconstructed vessels has enabled the identification of paths, working areas and rubbish deposits.
The pottery, ail coarse-ware, is often crudely made. Pots found in the first occupation sequence typically have a collared rim, while those vessels found in phases Il and III are either carinated with convex bases or round-shouldered. Other finds include lead and clay loom-weights, polishers, a disc and several unidentified iron objects. The geographical context and soil conditions are such that little organic material has been preserved.
Interpretation
Four successive periods of occupation have been identified, although we know neither how long they lasted nor the lapse of time between each one.
However, we observe that the new occupants recalled ,the way that the interior of chambers had been organised previously and maintained the same organisation. Changes in the intemallay-out were brought about mostly by a gradualloss of space.
The side-room (unit III) may have been a store-room or larder, while the central chamber was used as a workshop (for spinning ?).
The lack of evidence for normal domestic activities (such as cooking) tends to suggest that time was divided between the house,. outsideandthe below-groundarea.
CONCLUSION
The storage pits and underground chambers found in the Marche area are related to medieval rural settlement. Wooden buildings with associated storage pits are known from the 10th century. They are replaced in the 10th and 11th centuries by the house and underground chamber. These structures are abandoned in the 12th and 13th centuries, but in more than two cases out of three the habitat remains and forms the basis of present-day settlement.
The present study identifies major characteristics and provides a broad chronological framework, but deals neither with problems of continuity nor with local variants. A real understanding of history of rural settlement in the Limousin will only be achieved through further excavation and cartographic survey.
