Originally developed about recent lithic productions (Upper Paleolithic blade production, Neolithic specialized productions: e.g. Pelegrin, 1985 and 2007; Pigeot, 1987; Karlin, 1991; Appel et Knutson, 2006) the identification of skill levels is critical for the study of Lower Paleolithic industries. However, the apparent variability visible in these industries as well as the diversity of the raw materials used, such as different tenacious volcanic rocks within, among others, the East African context, form as many constraints for an evaluation of the levels of technicality represented in the assemblages (e.g. Jones, 1979; Toth, 1982; Edwards, 2001). In this article we present the methodology and results of the Ph.D. of S. Sánchez-Dehesa Galán, on the Acheulean assemblage of Garba I (Ethiopia) dated to about 600,000 years ago (Sánchez-Dehesa et al., 2022) and thus attributed, indirectly, to Homo heidelbergensis (Profico et al., 2016). We use our results to infer possible mechanisms of learning and knowledge transmission of Acheulean groups. Garba I is part of the Melka Kunture complex (Ethiopia) and is known to be one of the richest and most extensive handaxe accumulation of the African Acheulean: excavated over 240 m2 and a maximal thickness of 20 cm, it gave more than 900 shaped pieces and numerous smaller tools and fragments of eroded bone (Bailloud, 1965; Chavaillon, 1967, 1970 and 1978; Chavaillon et al., 1978; Chavaillon and Berthelet, 2004). Excavated half a century ago by the French mission at Melka Kunture directed by J. Chavaillon, the chronology and formation process of the accumulation had remained so far scanty, and the lithic assemblage unpublished. New fieldwork and precise dating conducted during the PhD of S. Sánchez-Dehesa Galán (2020) confirmed the anthropic in situ origin of the accumulation and provided a approximative datation of 600 ky of the occupation (Sánchez-Dehesa et al., 2022). For this study we focus on the Large Cutting Tools (“LCT”, n=828), handaxes and cleavers, as they are the tools with the highest technical investment allowing for a more precise appreciation of technical skill. Among these, the well knapped pieces (n=151) are easy to distinguish; they show a good control of the knapping and bear no -or almost- accidents and errors. Considered as accidents are a deep hinge and an edge crush. An irregular profile and/or outline, an avoidable prominence on a face, are considered as errors. These descriptive characters are objective. These well knapped pieces can be attributed to the best expression of the mature knappers of Garba I and allow for the identification of three main objectives: - typical U-shaped cleavers (sensu Tixier, 1956), made out of a large flake from which an adequate edge is selected and preserved as the distal cutting edge of the cleaver (n=42; fig. 3b); - bifaces (sometimes U-shaped; n=71; fig. 3c), that are bifacially shaped, yet whose distal end displays a transversal sharp edge created by an ultimate “tranchet blow”, which consists in the removal of a special sharpening transversal edge-flake. - typical almond-shaped bifaces (n=38; fig. 3a). The observations made on the Garba I assemblage highlights several important aspects related to the diversity of Acheulean industries, and most specifically to the stability of the shape of bifaces and cleavers: Garba I LCT’s show a small range of shapes that go from cordiforme to limande shapes (Bordes, 1961), whose general dimensions can vary, but their proportions (length/width ratio) are remarkably constant, indicative of an homothetic outline. The reproduction of distinctive tools forms irrespective of raw material type and blank size shows a remarkable capacity of adaptation of the knapping tools and gestures. Moreover, this successful adaptation of bifacial knapping techniques (through gesture and force, as well as the preparation of the knapping point before the removal) to raw materials as different as basalts (rather tenaceous) and obsidian (a very fragile volcanic glass), is impressive, testifying of a deliberate intention. Therefore, we support that the production of these pieces proceed from deliberate models (mental templates), successfully reproduced by the mature knappers of Garba I. The overall outline, homothetic (similar proportions), appears much more specific than that of a purely functional tool, which would simply associate a cutting edge, mass and adequate hand manipulation grip. Concerning teaching behaviors, the detailed analysis of the large bifacial pieces has enabled an appreciation of different skill levels in the assemblage that correspond to the gradual acquisition of knapping control and efficiency for the production of large bifacial tools. However, the presence of children during prehistoric times is not a great discovery (e.g. Coutouly et al. 2020). As we can detect from the many “apprentice pieces” they left, they were motivated to learn and reproduce the stone tools made by the elders. More important, the elders were possibly interested in their success. At Garba I we have documented: - The apprentices had access to blanks for exercise, which may be seen as a “facilitation”. - The tranchet blow, the production of very large flakes, the delicate preparation of obsidian edges for thinning are each a complicated and exigent technical procedure that can be suspected of requiring specific demonstration. - Bifacial pieces of reduced size showing immature skill proceed most probably from beginners/early apprentices. These would have used their own knapping tools or tools facilitated by the elders. In the latter case, the fabrication of these smaller “copies” could have been suggested by the elders as exercise. - Finally, we suggest that well knapped miniaturized bifacial tools can be interpreted as tools fabricated by adults for their children, and could also be models produced during demonstrations so as to show in detail the actions and gesture required. To conclude, distinguishing skill levels in an archaeological collection has two main components. On the one hand, recognizing the pieces produced by experienced knappers bears a methodological interest as it results in a “purification” of the assemblage and allows for a better understanding of the intentions underneath the knapping activity. On the other hand, identifying the products of less experienced knappers (from beginners to apprentices) may possibly reveal different clues of intervention by the experienced, as a complement to the natural acquisition of the young (whose progression has been finely studied by N. Pigeot (1987) in Étiolles and later enriched by various authors as D. B. Bamforth and M. Finlay (2008), and the contributors in L. Klaric (2018). These possible clues of teaching do not provide new information to the understanding of modern Human, but are important elements of discussion on the history of learning, from a cognitive and psycho-social point of view.