Abstract :
[fr] En France, le Gravettien moyen est caractérisé par deux unités archéologiques lithiques dont les relations chronologiques restent difficiles à appréhender : (1) le Noaillien, défini par la présence de burins de Noailles au sein des séries archéologiques qui lui sont rattachées, et (2) le Rayssien, défini par une méthode de débitage lamellaire dévolue à la production l’éléments d’armatures de chasse : les lamelles de la Picardie. La répartition géographique de ces deux entités se recoupe dans le nord du Bassin aquitain, et depuis leurs découvertes respectives, les hypothèses se sont succédées pour expliquer leur co-occurrence au sein des mêmes ensembles archéologiques dans cette région. Toutefois, peu de sites se prêtent à une analyse diachronique fiable du Gravettien moyen en raison de l’ancienneté des fouilles ou de l’action de processus post-dépositionnels ayant pu affecter l’intégrité des ensembles archéologiques. La séquence la plus fiable de la région reste celle de l’abri Pataud, qui permet de poser l’hypothèse d’une succession – tout du moins partielle – du Noaillien puis du Rayssien.
Dans cet article, nous revenons sur un cas de figure classique d’inversion stratigraphique par rapport à la séquence de l’abri Pataud : il s’agit du gisement archéologique des Jambes à Périgueux, fouillé dans les années 1960 par Guy Célérier, dans lequel ce dernier décrit la présence de burins de Noailles dans la couche 2 (supérieure) du site, tandis que les burins du Raysse sont présents à la fois dans la couche 2 (supérieure) et la couche 3 (inférieure). Cette proposition repose néanmoins sur une publication partielle de cette fouille, et plusieurs auteurs ont également soulevé la possible action de processus de versant ayant pu compromettre la validité de cette inversion stratigraphique. Afin de clarifier ce cas de figure, nous avons entrepris la révision archéo-stratigraphique et typo-technologique des couches 2 et 3 du gisement des Jambes. Cette analyse a démontré qu’il est difficile de distinguer différents ensembles archéo-stratigraphiques au sein du gisement, ce qui nous conduit à proposer la présence d’une unique nappe de vestiges probablement affectée de processus caractéristiques des versants. Elle a en outre permis de préciser la place de ce site au sein du Gravettien moyen en apportant une description plus précise des comportements typo-technologiques associés à ses industries lithiques. Ainsi, la présence du Rayssien est clairement attestée, tandis que l’expression du Noaillien y est plus discrète étant donné la quasi-absence de burins de Noailles. Toutefois, l’analyse des modalités de débitages lamino-lamellaires permet de tracer des liens avec d’autres séries à burins de Noailles dans la
moitié ouest de la France. L’apport de ces nouvelles données constitue un nouveau pas vers
une meilleure connaissance des industries lithiques du Noaillien et du Rayssien en Périgord, et permet de rediscuter des hypothèses de leur possible contemporanéité partielle. Il s’agit par ailleurs d’un excellent exemple de reprise de données anciennes, dont la description bibliographique initiale était incomplète et ne permettait que difficilement d’évaluer la qualité des conclusions archéologiques proposées, compromettant ainsi leur réutilisation dans le cadre d’analyses de plus grande ampleur.
[en] In France, the Middle Gravettian (31.2-28.5 ka cal. BP) is characterized by two archaeological units, which chronological relationships are difficult to understand: (1) the Noaillian, defined by the presence of Noailles burins in archaeological assemblages, and (2) the Rayssian, defined by a reduction method devoted to the production of microlithic hunting armature elements, called Picardie bladelets. The geographic distributions of these two entities overlap in the north of the Aquitaine basin, and since their respective discoveries, many hypotheses have been proposed to explain their co-occurrence in the same archaeological assemblages in this area: some authors consider that they reflect a functional or cultural complementarity of the Noaillian and the Rayssian, while others interpret these association as the result of post-depositional processes or unprecise excavation methods that mixed initially distinct occupations. However, few sites are fit for a reliable and fine-grained diachronic study of the Middle Gravettian, precisely because of the antiquity of excavations, the lack of knowledge of discovery/stratigraphic contexts, or because of the post-depositional processes that might have affected the integrity of archaeological assem-
blages. The most reliable sequence in the region is the long abri Pataud Upper Palaeolithic sequence, which allows one to hypothesize a succession (at least partial) of the Noaillian (couche 4: lower) and the Rayssian (couche 4: upper).
In this article, we study the archaeo-stratigraphic sequence of Les Jambes (Dordogne), which provides a classic case of stratigraphic inversion when compared to the abri Pataud reference sequence. It was excavated by Guy Célérier between 1964 and 1968. The first and only publication (1968) presents the first two years of the excavation, in which G. Célérier describes the presence of Noailles burins in layer 2 (upper layer), while Raysse burins are present in both layer 2 (upper layer) and layer 3 (lower layer). However, this proposition relies on the partial publication of the excavation and associated archaeological assemblages, and many authors have also underlined the possibility that the action of slope processes on the site may compromise the validity of this stratigraphic inversion. To clarify this case, we revised the archaeo-stratigraphy of Les Jambes site’s layers 2 and 3 and produced a typo-technological description of the associated lithic industries. We examined the excavation archives, composed of field notebooks, photographs, correspondence and plans, as well as the archaeological flint collection, composed of 6 761 objects (layer 2: 2’761; layer 3: 3’936) and 13’422.1 g of fragments and splinters of size smaller than 3 cm (layer 2: 5’657.7 g; layer 3: 7’764.4 g). Examining archives allows us to better understand how the artefacts were collected at Les Jambes: excavation methodology consisted of coordinating the most notable objects and dry sieving the sediment. Notes in field notebooks and their comparison with Célérier’s own working documents indicate that the archaeo-stratigraphy was difficult to read on the field. We then proceeded to digitize all object’s coordinates concealed in field notebooks, which corresponds to ca. 20% of the total collection. Coordinated objects were then projected and visually analyzed using the archeoViz web application. Spatial visual analysis in both horizontal and vertical dimensions shows that it is difficult to distinguish different archaeo-stratigraphic units at the site, which leads us to hypothesize the presence of a unique patch of artefacts. To test the internal coherence of the archaeological layers defined by G. Célérier, we also performed intra- and inter-layers systematic refitting of broken blades, bladelets and backed microliths. The results of this process are mostly inconclusive, as only 8 refits were found despite the exhaustivity of methodology (204 308 tries). However, the only refit between coordinated artefacts shows an important slope corresponding to the general slope of coordinated artefacts, suggesting that the site was probably affected by slope-characteristic processes. To document typo-technological aspects of layers 2 and 3 lithicindustries, the collection was analyzed using the “French technological approach”. It consisted of typo-technological classification of artefacts, diacritical schemes and drawings, and the collection of different qualitative and quantitative information, such as the dimensions of blades or the presence/absence of diagnostic traits. These data were then used to propose general hypotheses about the
chaîne.s opératoire.s employed to produce blades and bladelets in each layer.
The results of this study bring new data on the typo-technological behaviors associated with the Middle Gravettian in northern Aquitaine, thus allowing to better replace this site in its chronocultural regional context. Results show that the Rayssian is clearly present in the assemblages: the Raysse bladelet reduction method presents very characteristic expressions when compared with other well-studied Raysse burin-cores assemblages, such as La Picardie, Grotte du Renne, Bassaler-nord, Le Raysse or Le Flageolet I. In addition, blades and blade cores presenting typical Raysse method elements, such as lateralized implementation of débitage or oblique lateralized faceting of the platform prior blade extraction, were identified in significant number. These observations thus support the technical transfer of the Raysse method from bladelet cores towards blade cores described at La Picardie and Grotte du Renne by L. Klaric. In contrast, the Noaillian is very poorly represented, with almost no Noailles burins identified in the collection. However, the analysis of the blades-bladelets débitage modalities suggests some similarities with other assemblages with Noailles burins in the West of France. These similarities are complemented by the presence of backed bla-
delets in the assemblages. Finally, a few potentially Recent Aurignacian artefacts were described, and cast doubts on the stratigraphic integrity of the sequence. Unfortunately, layers 2 and 3 are extremely similar – if not identical – in terms of typo-technology, and spatial visual analysis did not allow to spatially discriminate artefacts associated with the Rayssian, the Noaillian and the Recent Aurignacian.
These new data are an important step towards better understanding of Noaillian and Rayssian lithic industries in the Périgord region. They are also valuable for rediscussing their potential contemporaneity in the north of the Aquitaine basin. The case of Les Jambes is also an excellent example of an old excavation data reconsideration, data for which the initial bibliographic description was incomplete and did not allow a full and reliable evaluation of the proposed archaeological conclusions. This aspect necessarily compromised their reusability in the framework of bigger scale analyses; in conclusion, we advise researchers to consider bibliographic data carefully and to restudy old excavations when necessary and possible, to enhance the quality of bigger-scale archaeological approaches.
