Analyse des artefacts lithiques du site de Longgupo

Excavation has enabled recovery of 854 artifacts within 30 archaeological levels in the south sector and 11 in the north (chapter 3). These levels are quite probably contemporaneous, or even the same. The quantitative disparity in the number of strata between the two sectors is simply due to the fact that only the lower half of the northern zone was completely investigated. Similarly, the number of artifacts recovered by level varies according to the surface area excavated, although is some cases the density of material is significant despite the small area excavated, for example stratum C IV 5 which contains 174 lithic artifacts in 2 m². Before undertaking the technological analysis of the artifacts, given the preceding polemics provoked by the great age of this site and its implications for the spread of the first populations out of Africa, it was deemed important to carry out a plurifactorial analysis combining all of the data related to: the stratigraphy, taphonomic processes (post-depositional disturbance), analysis of natural processes that may have produced eoliths, experimentation and techno-functional analysis of the material. The stratigraphy shows clear interstratifications of fine and coarse fluviatile levels with often very clear particle size sorting of the coarse fraction. The archaeological material is typically found at the interface of these strata, either at the base of a clayey matrix, overlying a preceding coarse level, or in the superficial part of a sandy-clay deposit underlying coarse deposits. Post-depositional disturbance revealed during the new excavations in 2003–2006 cannot alone be the cause of eoliths. Excavation of a 6 m² zone in the modern river bed, located below the site, has demonstrated that the technological traits of eoliths recovered cannot in any way be confused with the technological traits of the artifacts recovered at the site itself. Similarly, viewed quantitatively, the 6 m² zone excavated in the river bed yielded around 20 eoliths while the 30 m² zone excavated at the site yielded 854 artifacts, one stratum alone yielding 184 artifacts in a 3 m² zone. During the experimental phase, adopting the same conditions of procuring raw material, from the same river bed, we very quickly realized the rarity of types of adequate volume that had been generally used at the site and the need to use certain operational processes to create such a form. In addition, the hardness and presence of several natural fracture planes in the Triassic limestone explain the choice of different operational processes and the very high number of knapping accidents, including those occurring during bipolar percussion. Although 90% of the raw material used was cobbles or broken blocks of local Triassic limestone, 10% of the tools were made on exogenous raw materials – siliceous or gravelly limestone, quartzitic sandstone, chert, volcanic rock – that are absent from the immediate environment of the site. These raw materials were brought to the site in the form of tools: worked cobbles, large retouched flakes, backed double-truncated flakes, a plaquette with a lateral bifacial edge, etc. The 854 artifacts have been classified into six object classes: worked cobbles with transversal edge (39%), worked cobbles with lateral edge (2%), unipolar flakes (27%), bipolar objects (half-blocks, half-cobbles including some flat “split” cobbles, “orange slices”, flakes and diverse fragments) (17%) and fragments resulting from knapping of blocks or cobbles (13%), hammerstone (2%). When the frequencies of these classes are calculated for each of the sectors, percentages are similar, indicating a high degree of homogeneity in the archaeological assemblages at the site. The situation is somewhat different when assemblages are compared within a single sector. Slight differences appear in the percentages of bipolar pieces and unipolar flakes. These differences seem to be random, like the frequency rate of knapping accidents in bipolar reduction, or economic, such as the choice of operational schemes to create worked cobbles based on the availability of suitable raw materials. The technological affinity between each of the archaeological assemblages tends to demonstrate great stability in technological knowledge through time. The class of worked cobbles is by far the most important and, apart from a few flakes produced intentionally, it appears to group all of the tools. To avoid placing these tools in a restrictive, semantically meaningless, class, we prefer the concept of matrix to the term worked cobble. A matrix is a structured arrangement of a series of technological traits, in a form as close as possible to that of the future tool. The matrix phase leads to the tool production phase, which may be unnecessary if the matrix phase includes fictionalization. In other words, the concept of matrix enables separating the phase of preparing a predetermined volume, such as a blade, Levallois flake or bifacial piece, from the tool production phase, consisting in creating the type of transformative edge intended, if necessary. The tool is thus an artifact of a specific form with an integrated edge and an operational scheme both specific to the function attributed to it and means of use associated with the form. Observed variability relates to: the size of the volume ranging from to 20, morphology, the form of the line formed by the edge which can in frontal view be curved, linear, sinuous or denticulated, and in transversal view curved, linear, sinuous or saw-toothed, and the length of the edge ranging from 1 to 10. Matrices with a simple bevel are distinguished from double bevels. In the framework of the technological analysis of production schemes to produce matrices with a simple bevel, a broad range of variability in production schemes can be observed, divided arbitrarily into two stages. The first stage consists in creating as closely as possibly the technological traits of the future matrix due to five general schemes. The first scheme (A) consists in selecting a cobble or block naturally possessing at least some of the technological traits needed. The missing traits are added by various preparations, including bipolar percussion 3 times out of 5. The second scheme (B) consists in knapping a flake from the block with some of the technological traits required for the matrix present on one of its surfaces. The third scheme (C) consists in the choice of a plaquette from which a bipolar shock creates the main traits of the matrix. The fourth scheme (D) consists in choosing a volume very similar to the intended matrix. The fifth and final scheme (E) consists in knapping a flake with technological traits very different from those intended. Depending on the distance between intention and realization, a second stage may be necessary. In general, this second stage perfects or creates the intended active edge, which is rarely obtained in the first stage. To produce a matrix with a double bevel, it is sometimes necessary to add an intermediate stage in order to prepare the second bevel. The first stage remains the same, with the use of the five operational schemes. By contrast, a clear difference exists in the percentages for the use of these schemes. For a matrix with a simple bevel, scheme A is dominant, followed by scheme D, while the situation is reversed for a matrix with a double bevel, where scheme D is dominant. Unipolar flakes, representing 27% of the assemblage, are produced in three different ways. The most important is flakes resulting from matrix production. The two others are flakes produced during different knapping schemes, some flakes in relation to the few cores present, other flakes in exogenous raw materials produced elsewhere and generally much larger. The other classes are dominated by bipolar products resulting essentially from knapping accidents. To summarize, these assemblages are characterized by: the search for tools differentiated by form and active edges; more than 90% of the tools made on two kinds of supports: a matrix with a simple bevel or a double bevel; matrices obtained using different operational schemes successively associating if necessary a knapping stage and a shaping stage. While the Triassic limestone is hard and thus imposes a strong constraint on knapping, the range of operational schemes appears to have been a “cultural” response diversified to this constraint and the presence of tools on exogenous raw materials. At the scale of China, comparison of this industry is impossible since it is the only site of this age and to contain so much material. The site of Majuangou, the only site of similar caliber, is younger by several hundred thousand years and is located several thousand kilometers to the north, making comparisons meaningless. We note only that most of the tool supports at Majuangou are knapped flakes. On an inter-continental scale, the comparison of sites of equal age is more promising. But lithic analyses are based on different methods, preventing comparison of similar data. However, if we make a simple summary of the data available, we can first say that in Africa, during these periods, different development technological stages were present and stages that are considered more evolved are manifestly less common using our approach. While these stages are more or less contemporaneous, which counters the idea of uniqueness, they would more surely be evidence of populations that were not in direct contact and had separate lines of development. In Asia, the Longgupo industry evidences a different technological option than that of contemporaneous populations in Africa. By contrast, when we take into account its developmental stage, we realize that this is an “evolved” stage in which the form of the support of the future tool is predominant. If we compare Africa and Asia in terms of stages, we are a priori at the same stage with different options being selected.