Neolithisation in the Northern French Alps: First Results of the Lithic Study of the Industries of La Grande Rivoire Rockshelter (Isère, France)

1 Introduction

The moment of the “transition” between the end of the Second Mesolithic and the beginning of the Early Neolithic is a singular phase of the Prehistory of Western Europe. Following a diffusion movement initiated from the Near East between 11000 and 7000 cal. BC, the spread of the “Neolithic package” to the west was relatively fast, reaching the Atlantic margin around 5000 cal. BC. Until recently, it was still considered that its set up across western Europe would have followed two fronts: on the one side, a colonisation course by the Rubane, going up the Danube River to the northern regions and, on the other side, an acculturation movement of the Mediterranean zone, originating in the Italian peninsula, with a first incursion of the Impressa facies on the coast, followed by the Cardial, which will go back up through the Rhone Basin (Ammerman & Cavalli-Sforza, 1971; Guilaine, 2001).

This dichotomous reading of the neolithisation of the west of the European continent, opposing Danubian colonisation and Mediterranean acculturation, is debated today (Guilaine, 2001; Mazurié de Keroualin, 2003a,b; Rasse, 2008; Zvelebil & Dolukhanov, 1991). Thus, the process would have followed an otherwise more complex and arrhythmic movement, marked by a succession of progressions and pauses in some environments, which would have led to transformations or stabilisation times necessary for the agro-pastoral populations (Guilaine, 2001). Several factors would have influenced these changes. These include a readjustment of agricultural practices in the approach of a territory that might be unsuitable for a food production economy, or even in front of encounters with indigenous hunter-gatherer groups, still present in the area, and who might be recalcitrant to the appropriation of this new technological background, or simply to interact with these newcomers (Guilaine, 2001; Mazurié de Keroualin, 2003a,b; Perrin, 2013; Rasse, 2008).

This article focuses on the first results of the study of archaeological assemblages from recent excavations at La Grande Rivoire rockshelter (Isère, France). Emphasising the potential recognition of behavioural changes that can be observed in the reduction sequences of lithic industries from the recent phase of the Second Mesolithic to the Early Neolithic pre-shepherdship, this work was undertaken to define the chronocultural framework surrounding the neolithisation process more clearly in the Northern French Alps. The site, excavated annually between 2000 and 2017 (Angelin, 2017; Angelin, Perrin, & Nicod, 2018, 2021; Nicod & Picavet, 2003; Nicod et al., 2012), has indeed revealed a complete stratigraphy for the period and new data allowing us to update our knowledge of the modalities of the settlement of the Early Neolithic in the northern Alpine region. A complete study of the shaping methods of the diagnostic pieces could highlight possible technical transmissions between Mesolithic and Neolithic spheres.

2 The Northern French Alps: A Particular Context of Neolithisation

The first signs of neolithisation in the Northern French Alps appear in the second half of the sixth millennium BC, in a range between 5500 and 5350 cal. BC (Figures 1 and 2). Even today, the very question of the arrival of the Neolithic in the northern Alpine region is still difficult to understand, especially when we are concerned about the rate and modalities of its implantation. These flaws are related to the presence of numerous imprecision in the local chronocultural sequence (Bintz, 1991; Nicod et al., 2012). These gaps are, among other things, the result of a small number of sites with a sufficiently developed stratigraphy for the period or by the existence of sometimes significant gaps between the phases of occupation. The materials of the referenced sites tend to be tenuous, due to occupations that may be of relatively short duration, if not coming only from pedestrian prospections or reduced surveys. Moreover, we are often faced with old excavations, which have their share of biases, among other things because of insufficient sieving conditions for microlithic industries (Nicod et al., 2012).

Figure 1

Chronocultural evolution of the Second Mesolithic to Early Neolithic in the Northern French Alps, the Rhone basin and Provence and the Jura and Swiss Plateau (compilation of data from Bintz, Bocquet, Borel, & Olive, 1989; Bintz, 1991; Beeching, 1995; Bintz, Picavet, & Evin, 1995; Voruz, Nicod, & de Ceuninck, 1995; Perrin, 2001; Angelin, 2017).

Figure 2

Hypothesis of chronocultural evolution in the Rhone Basin between 5500 and 5200 cal. BC (modified from Perrin, 2003b).

The presence of numerous taphonomic phenomena also adds to the confusion of the understanding of the neolithization process in the Northern French Alps. These are, for instance, the result of a sedimentary accumulation sometimes erratic on some deposits, frequently accentuated by the natural disturbances of the freeze/thaw cycles that are frequent in high-altitude context and leads to asynchronous mixing of occupations (Bintz, Picavet, & Evin, 1995; Picavet, Angelin, & Moulin, 2014). These phenomena can also be aggravated by a slowing down of the sedimentary input experienced locally during the “transition” period from the Second Mesolithic to the Early Neolithic, having the effect of artificially bringing together occupations that are far apart in time (Bintz et al., 1995; Nicod et al., 2012; Picavet et al., 2014).

In addition, the chronocultural sequence is partly disturbed by a significant reoccupation rate of Mesolithic sites at medium and low altitude by the first agro-pastoral groups to have penetrated the northern Alps. Some sites thus yield archaeological assemblages known as “mixed sets,” revealing characteristic elements of Second Mesolithic and Early Neolithic lithic industries within the same level, which are only differentiated by the technical systems specific to each tradition (Figure 3). Generally referred to in the literature as “Epicastelnovian,” “Meso-Neolithic” or “Proto-Neolithic” (Bintz, 1991; Nicod et al., 2012; Picavet, 1999), these assemblages are considered to be unreliable, as such associations are interpreted as the result of asynchronous occupation shifts (Nicod et al., 2012; Perrin, 2002, 2008, 2013).

Figure 3

Arrowheads from layers d15 to d30 of the SU16-22 sector at La Grande Rivoire rock shelter. Drawings: Régis Picavet (modified from Nicod et al., 2012).

However, certain behavioural changes may have taken place on some sites, following probable borrowings and interactions between Mesolithic and Early Neolithic groups. Indeed, the context of neolithisation in the northern Alps is all the more complex in that there is a certain temporal and geographical proximity between the last indigenous hunter-gatherer-collectors and the first agropastoralists in the upper Rhone Valley around 5200 cal. BC, which could have led to possible interactions between the two human groups (Perrin, 2003a,b,c, 2008, 2013). In this case, such patterns could be observed within perceptible changes in the technical systems of either or both industries.

3 Problematisation

To better recognise the chronocultural framework and the mechanisms inherent to the neolithisation process of the Northern French Alps, the primary objective of this work is to characterise the reduction sequence specific to each of the lithic industries of La Grande Rivoire occupations, from the recent phase of the Second Mesolithic to the pre-shepherdship levels of the Early Neolithic. Through the study and analysis of the lithic productions, we seek to identify the criteria that may be common, or inversely distinct, to recognise the intentions and characteristic modalities of the “mixed” assemblages that allow us to distinguish their cultural attributions. On a diachronic level, it would be possible to identify a continuity of traditions across the occupations of the rockshelter, indicators of exchanges and/or interactions between them or even, on the contrary, to recognise a more or less clear break between the technical systems of the end of the Mesolithic and the beginning of the Early Neolithic.

The study of lithic industries and the technical systems associated with them favours the appreciation of behavioural changes during neolithisation. Indeed, stone knapping was practised by both the Mesolithic and Neolithic peoples, each with their own intentions and traditions. Moreover, arrowheads, as elements with an important cultural charge, are suitable for a typo-technological study on coherent archaeological assemblages. In other words, such a study of lithic technical systems makes it possible not only to identify the intentions of the prehistoric knappers but also to observe the succession of cultural entities according to the phases of occupation of the site.

In a second time, we also questioned the possible role played by the last hunter-gatherer-collector during the neolithisation of the North Alpine region. Although the presence of assemblages with “mixed” industries is widely considered to be the result of blend in asynchronous occupations, it is possible that such assemblages may reflect, on the one hand, the persistence of certain Second Mesolithic traditions within the material productions of the early local Neolithic, or on the other hand, the appropriation of allochthonous technical traits by indigenous people. Such evidence could be observed in the heart of the knapping process, in the shaping of tools and arrowheads, or even in the management of raw materials.

A final underlying question is the origin of the transversal arrowheads. These inverse bitruncations with direct low-angle retouches – more commonly called “Montclus arrowheads” – are generally attributed to the Neolithic. However, this reality is now questioned by the results brought by recent excavations of some deposits, notably those of La Grande Rivoire (Isère, France) or Cuzoul-de-Gramat (Lot, France) rockshelters, where these arrowheads are also found in Second Mesolithic contexts (Nicod et al., 2012; Perrin, Marchand, Valdeyron & Sam, 2019; Valdeyron et al., 2011). We can therefore wonder about the real context of appearance of these arrowheads, which could be presented in different scenarios: (1) the first hypothesis would suggest that this is a truly Neolithic concept, as these arrowheads constitute the dominant type of the Cardial and Epicardial assemblages and their presence in the industries of the last hunter-gatherers is the result of a late acculturation of these; (2) a second hypothesis is that they have a Mesolithic origin, through the recognition of the practice of low-angle retouches already in this period, as well as the absence of these arrowheads in the material productions of the Neolithic Impressa facies, and a later adoption by the Mediterranean Early Neolithic (Perrin et al., 2019); (3) or – although this hypothesis is more complex and still difficult to apprehend with the state of research – would come back to a possible simultaneous development in the two cultural spheres, with techno-functional schemes, intentions and, possibly, hafting modes characteristic to each one (Philibert, 2016; Perrin et al., 2019). The latter scenario would, in other words, the emergence of similar but distinct arrowheads concepts. In this case, a better understanding of the “mixed sets” and their technical systems is therefore necessary.

4 La Grande Rivoire Rockshelter

Lithic assemblages from recent excavations at La Grande Rivoire rockshelter were considered for this study because of the great potential represented by the site (Figure 4). Discovered in 1986 by an archaeology amateur from Grenoble (France), the site is located in the north of the Vercors massif, more precisely on the western slope of the Furon Valley, the latter constituting the main access route from the Cluse de l’Isère to the highlands. The shelter opens at the foot of a senonian flint limestone massif at an altitude of 580 m a.s.l., covering a surface of about 80 m².

Figure 4

Geographical location of sites in the Vercors massif that has yielded material with components from the Second Mesolithic to the Early Neolithic (modified from Nicod et al., 2012). 1: Grande Rivoire; 2: Pas de l’Échelle; 3: Balme Rousse; 4: Coufin 1; 5: Machiret; 6: Pas de la Charmate; 7: Bouvante; 8: Ferme d’Ambel; 9: Vassieux-en-Vercors; 10: Fontaine de la Baume; 11: Gerland; 12: Pré-Peyret; and 13: Pas de l’Aiguille.

Between 1986 and 1994, the site was the subject of five rescue campaigns led by the archaeologist (Picavet, 1987, 1991, 1999). These first operations allowed not only to identify the high potential of the rockshelter but also to intervene on the levels threatened with collapse due to the presence of a quarry front, remnant of the widening of the departmental road below in anticipation of the Grenoble Olympic Games of 1968 (Picavet, 1987, 1991, 1999). Considering the importance of the chronocultural sequence uncovered, and the ever-present danger of destruction despite the support structures erected, a more exhaustive annual scheduled excavation was finally launched in 2000 – and lasted for 18 years – at the request of the Ministry of Culture and the Conseil Général de l’Isère, under the direction of Nicod (Angelin et al., 2018, 2021; Angelin, 2017; Nicod & Picavet, 2003; Nicod et al., 2012).

The archaeological operations carried out at La Grande Rivoire have led to the discovery of numerous elements that facilitate the study of the neolithisation processes in the northern Alpine region. The site has revealed one of the few reliable reference stratigraphies, which shows, at a height of nearly 6 m, an almost uninterrupted occupation from the First Mesolithic (8301–8223 cal. BC) to the Gallo-Roman period (897–58 cal. BC). The “transition” phase from the end of the Mesolithic to the beginning of the Early Neolithic is observed in five sedimentary horizons (levels B3a to B1b of Picavet operations) characterised by a complex stratification where ash and organic levels intersect (Figure 5). Among these horizons, we note the presence of an occupation phase attributed to an Aceramic Early Neolithic period. The excavations have yielded a substantial amount of material, counting several thousand lithic pieces, faunal remains, bone industry as well as palaeo-environmental elements (charcoal, seeds and so on) (Nicod & Picavet, 2003; Nicod et al., 2012).

Figure 5

Stratigraphic section S40 showing the archaeological levels from the First Mesolithic to the Early Neolithic pre-shepherdships in band R17-21 (sector NR16-21). The coloured lines correspond to the limits of each decapages (identified in the rectangles). CAD: Cyril Bernard (AVDPA) and Marc-André Dallaire.

5 Methodology Applied to the Lithic Study

5.1 Stratigraphy and Planimetric Observations

The levels attributable to the late phase of the Second Mesolithic to the Early Neolithic pre-shepherding were identified and excavated in planimetry in four sectors covering an area of approximately 55 m² (Figures 5 and 6): NR12–15 (d35–d43), NR16–21 (d127–d147), SU12–15 (d5–d18) and SU16–22 (d19–d34) (Nicod & Picavet, 2009, 2011, 2017; Nicod et al., 2003, 2006). These assemblages are marked by a great wealth of lithic, faunal and – for the upper “decapages” – ceramic furniture. The totality of the lithic elements, which amount to 38,872 pieces, was recorded in a database (FileMaker Pro) during a preliminary study of the material productions. Each piece was thus detailed in an individualised sheet considering the spatial, technological and typological data. The typo-technological study of the tools and arrowheads was carried out following the typologies proposed in the thesis of Binder (1987) and echoed by the work of Perrin (2001) (Perrin, Angelin, & Defranould, 2017^[1]), which take into account the nature of the blanks used as well as the shaping techniques employed, and the inclination and orientation of the retouches.

Figure 6

Localisation of the main excavation areas at La Grande Rivoire rock shelter and of our study area (in orange). CAD: Cyril Bernard (AVDPA) (modified from Nicod & Picavet, 2014).

Subsequently, an analysis of the spatial distribution of the remains was carried out to distinguish the area(s) revealing the best criteria for a complete study of the neolithisation process at the rockshelter. To achieve this, we followed the methodology previously proposed by Angelin in his thesis (2017, pp. 130–139; Angelin et al., 2018), which focused on the typo-technological study of the first to the early Second Mesolithic at La Grande Rivoire. This particular approach was chosen essentially because it allows us to observe in a more practical way the complex succession of occupations of this deposit. First of all, we proceeded to the realisation of the profiles of each square metre excavated in the four above-mentioned sectors, virtualised from the base elevations of the “decapages” noted on the excavation surveys, following the principles of the verticalisation of planimetric data (VPD; Sabatier, 1995).

These profiles, carried out with the Open Source software QGIS 3.10, were then used to project the density of lithic materials (Figure 7). Calculated by relating the number of lithic pieces to the quantity of sediment removed during “decapage” in each square metre excavated (Angelin, 2017, pp. 132–134), these projections allowed us to identify the areas of highest concentrations. One area, located in sector NR16-21, was then revealed to be the most relevant for further study. In addition to presenting higher density of materials, this sector is entirely protected by the rocky overhang of the shelter and therefore less affected by erosion and soil leaching, which favour the mixing of substrates (Figure 6). Our study of the lithic productions was subsequently limited to the elements from “decapages” d127 to d147 from square metres PR17–21 and P16 (sector NR16–21), attributable to the recent phase of the Second Mesolithic to the Early Neolithic pre-shepherding. However, the sample studied remains consequent, with a total of 26,881 pieces (Table 1), constituting a little more than 69% of this corpus. The rest of the recorded pieces will nevertheless be considered for a second time during a global study of the site to ensure the conformity of our results.

Figure 7

VPD on the NR18 axis displaying the density of lithic furniture for every square metre of each decapages (from decapage d127 to d147 in the sector NR16-21 of La Grande Rivoire rockshelter).

Table 1

Count of lithic pieces for each set in our study area

Sets	Pieces (total)	Pieces (without undetermined)	Tools (with arrowheads)	Arrowheads
	Nbr.	Nbr.	Nbr.	Nbr.
A–B	3,497	2,341	154	22
C	12,081	6,482	320	164
D	2,281	1,483	72	41
E	2,934	2,086	97	59
F	2,908	2,307	235	140
G	3,181	2,124	146	90
Total ( N )	26,882	16,823	1,024	516

Finally, additional projections were made to discern the presence of coherent archaeological assemblages allowing us to observe probable behavioural changes through our analyses. All the material – lithic, faunal and ceramic – recorded in three dimensions during the excavation was first projected onto the different density profiles. Due to the presence of an SW-NE dip in our study area, these projections were limited to a depth of 25 cm to reduce certain irregularities that could possibly disturb our observations. To complete our reading, the types of arrowheads and domestic tools identified in the recording were added to the projections, which finally allowed us to individualise six coherent sets (Figure 8):

1. a first set (set A–B), corresponding globally to “decapages” d144 to d147, has a medium density of lithic material, characterised by the presence of bitruncations with direct abrupt retouches (BG2) using the microburin technique. The number of faunal remains is significant, particularly in the lower “decapage” (d147);

2. a set with a very high density of lithic material (set C), corresponding to “decapages” d141 to d143, to which is added one facies – LGM(C) – from “decapage” d140, is characterised by a great diversification of the types of arrowheads and by the appearance of inverse bitruncations with direct low-angle retouches (BG32);

3. a set (D), comprising the rest of the “decapage” d140 and distinguished from the lower and upper sets by a very low density of material, dominated by inverse bitruncations and direct low-angle retouches (BG32);

4. the set E, corresponding to “decapages” d136 to d138, again with a high density of material, still dominated by inverse bitruncations (BG3), but with an important number of pieces with bifacial retouches (BG4 and PB31);

5. a set F, corresponding to “decapages” d131 to d134, rich in lithic furniture, and dominated by inverse bitruncations with direct low-angle retouches (BG32); and

6. and an upper set (G), corresponding to “decapages” d127 to d129, also rich in lithic furniture dominated by the “Montclus arrowheads” (BG32) and characterised among other things from the underlying set by the presence of a number of ceramic fragments.

Figure 8

VPD on the PR18 axis displaying the density of lithic furniture and the vertical projection of lithic, faunal and ceramic remains (from decapage d127 to d147 in the sector NR16-21 of La Grande Rivoire rockshelter).

6 From the End of the Mesolithic to the Beginning of the Early Neolithic at La Grande Rivoire Rockshelter: First Results

A complete study of the technical systems was undertaken for each of the six assemblages identified. These analyses let us recognise several changes in technical behaviour throughout the neolithisation process.

At the bottom of the sequence, represented by set A–B, we note the presence of two reduction sequences. The first one is oriented toward the production of short bladelets, or even narrow blades, with a rectilinear profile and a tendency to have distal curve. While most of them are made of regional flint (72%; n = 470), the use of good quality local materials is also noted. The shaping of the cores is rudimentary, limited to the opening of a plain striking platform, from which the knapping is immediately begun on a wide surface, sometimes semi-turned. On average, the first-line blanks have a width of 9.7 ± 3.2 mm and a thickness of 2.4 ± 1 mm. Their length is relatively short, with an average dimension of 34.8 ± 12.8 mm, due to the use of modules as cores not exceeding 40 mm. The butts are relatively wide and thick, with an average width of 6.3 ± 2.1 mm and a thickness of 2.7 ± 1.1 mm. With regular edges and ridges, the bladelets are equally divided between the two- and three-sided elements. However, there does not seem to be any particular reduction sequence or intention between the two types of bladelets, the number of sides resulting from the variability of the production. Indirect percussion is most often used for the extraction of bladelets, carried out mainly on plain butts, occasionally faceted, with frequent abrasion of the edges, although not systematically. The soft direct percussion, revealed by the presence of some dihedral butts, is also used jointly, exerted with a tangential movement, and appearing generally with the reduction of the volume following the sequences of knapping and marked by the production of appointed bladelets.

Although it represents little more than 32% of the lithic assemblage (n = 650), the production of bladelets dominates within the retouched elements (94%; n = 144). The most regular bladelets are preferably selected for the shaping of geometrical bitruncations (n = 22) made by means of the microburin technique. These are largely dominated by direct trapezoidal bitruncations (BG2; n = 20; Table 2), most often asymmetric (BG21; n = 13). The rest of the domestic tools, fashioned on slightly irregular bladelets or from other knapping phases, are made up of Montbani-type notched elements (BA25; n = 26), truncations (TR2; n = 18) and simple semicircular scrapers (GR2; n = 15). The second reduction sequence is oriented on the production of flakes, mostly extracted from small cores made of local raw materials. Although the number of flakes is important, constituting almost 70% of the assembly (n = 1,573), they are rarely retouched (n = 10). The products from all the knapping phases are used to make a few truncations (TR1; n = 3), backed edges (BA3; n = 3) or pieces with irregular removals (IR1; n = 2).

In the middle of the sequence can be subdivided into two distinct parts. The first one brings together sets C and D, which show a continuity of the technical schemes, despite the presence of disparities within the retouched elements. First, we observe a supply that favours more and more regional flints, which then amount to a respective rate of 70% (n = 3,777) and 80% (n = 1,064) of the two assemblages. In both sets, the production is mainly oriented on thin and short blades, of good workmanship, generally with a triangular section. Regional flints are preferentially used (84% (n = 932) in set C and 87% (n = 201) in set D), exploiting mostly blond/light brown flints with a very fine grain displaying a good knapping quality. The shaping of the cores, whether on a pebble or a small-volume block, remains relatively reduced, always limited to the opening of a plain striking platform, from which the knapping begins on a large unipolar surface. When the morphology of the core allows it, a bidirectional knapping on opposite faces is sometimes made according to the same parameters. Indirect percussion is manifested by predominantly plain [62% (n = 259) and 64% (n = 61)] or punctiform [14% (n = 57) and 12% (n = 11)] butts, diffuse bulbs [92% (n = 384) and 88% (n = 84)] and careful abrasion of the edges. The length of the first-line blades is reduced (27.1 ± 12.1 mm in set C and 27 ± 10.7 mm in set D) and again determined by the small size of the cores, which rarely exceeds 40 mm long. The blades have an average width of 9.6 ± 3.9 mm for 2.3 ± 1.2 mm thickness in set C and 9.6 ± 3.5 mm and 2.1 ± 0.9 mm in set D. The butts are small, averaging 4.9 ± 2.3 mm in width for 1.8 ± 0.9 mm in thickness in set C, and 5.3 ± 1.9 mm for 1.9 ± 1 mm in set D. The presence, more rarely, of dihedral or linear butts, with a high and short bulb and a marked lip, testifies, however, of the occasional use of a soft direct percussion, appearing with the reduction of the volume of the cores, resulting in the extraction of less regular blades with a curved profile.

First-line products, most regular and straight, are preferentially selected for shaping arrowheads, which constitute nearly 83% (n = 265) of the retouched elements in set C and 75% (n = 54) in set D. These arrowheads show a great diversification of types – notably in set C with 13 different types (Table 2), marked by the appearance of significant numbers of inverse bitruncations (BG3; n = 103 in set C and n = 24 in set D). These are mainly represented by inverse bitruncations with direct low-angle retouches (BG32; n = 68 and n = 12), which will then dominate over all the other transversal arrowheads until the top of the sequence. These “Montclus arrowheads” are, however, combined with numerous direct bitruncations (BG2; n = 47 and n = 9). Some of them, which had never been seen before, are also adorned with direct low-angle retouches realised before the retouching of the truncations (BG24; n = 24 and n = 3). Nevertheless, their intentions are comparable to the BG32. The microburin technique disappears in set C, the blanks being now segmented by means of the flexing break. Few microburins (MB; n = 12), identified in set C at the interface with the underlying set, would be inherent of sedimentary mixes. For both sets, the domestic tools are mainly shaped on blades from the other phases of knapping, mostly for the realisation of notched blades (BA25; n = 24 in set C and n = 2 in set D), truncations (TR2; n = 21 and n = 2), semicircular scrapers (GR2; n = 12 and n = 2) or various pieces with irregular removals (IR22; n = 32 and n = 9). Flake production by direct percussion remains important on these levels, although its representation rate will gradually decrease up through the stratigraphy. These are mainly used for the realisation of pieces with irregular removals (IR; n = 34 in set C and n = 9 in set D), some truncations (TR; n = 9 and n = 2) and simple scrapers (GR; n = 6 and n = 2). There is also a decline in interest in local raw materials, with production increasingly favouring the use of regional flints.

The second part of the middle of the sequence is composed of the lithic productions of set E, which show some pronounced changes with the trends observed so far. Although regional flints are still predominantly exploited (64%; n = 1,329), there has been a change in the strategies for their supply. These are now dominated by flints with yellow/mustard brown hues with orange zonation, which represent more than 50% of the regional flints (n = 585), whereas they were only anecdotal in the underlying levels. Nevertheless, the blond/light brown flints, much finer and more homogeneous, are again preferentially selected for the shaping of the tools and the arrowheads. The “Montclus arrowheads” (BG32), always numerous (n = 13), are accompanied by many pieces with bifacial retouches: BG41 (n = 10), BG42 (n = 5) and PB31 (n = 11) (Table 2).

The blade production follows the same tendencies observed in the first part, with an indirect percussion practiced on small modules, following a unipolar knapping on a large surface. The first-line blades have regular edges and a straight profile. They have a predominantly plane butt (69%; n = 104) and an almost always diffuse bulb (96%; n = 148). The butt is small with a width of 5.5 ± 2.5 mm and a thickness of 2 ± 1.1 mm on average. The abrasion of the edges is carefully but not systematically done, being observed in only 38% of the butts (n = 56). On average, first-line blades show measurements of 9.8 ± 2.8 mm in width and 2.1 ± 0.9 mm in thickness. Again, the reduced dimensions of the cores, which do not exceed 40 mm, determine the small size of the blanks, which has an average length of only 22.7 ± 8.3 mm. The first-line products, with a preference for the widest elements (12 mm in average), are selected for the shaping of the domestic tools [66% (n = 63) of the retouched pieces], and more particularly of the various types of arrowheads (i.e. 75% of the tools on blades; n = 54).

At the top of the sequence, a resurgence in the supply of regional raw materials is again observed in sets F and G. The proportions increase to 73% (n = 1,303) in the former and 85% (n = 1,628) in the latter, thus recovering rates comparable to those observed before in sets C and D. The types of flint will then diversify, several rocks already present since the bottom of the sequence showing higher rates now. Nevertheless, the flints with a better quality are favoured over the others – especially the blond/light brown type – for the shaping of domestic tools and arrowheads. To these are added grey to light brown flints, fine-grained, which will take a more important place in set G. The top of the sequence is also distinguished by a significant shift away from local raw materials, constituting less than 15% of the total of the two assemblages (n = 499 and n = 265).

For both material productions, the knapping is mainly oriented for blades, carried out by indirect percussion on small pebbles. The shaping of the cores is once again reduced to the opening of a striking platform, from which the knapping is directly engaged in a narrow unipolar frontal surface allowing the extraction of three or four blanks per sequence. The production is focused on obtaining blades with a high degree of straightness of the edges, ridges and profile. The proximal part has butts predominantly plain [68% (n = 137) in set F and 70% (n = 74) in set G], if not punctiform, and an almost always diffuse bulb [89% (n = 180) in set F and 93% (n = 98) in set G]. In both assemblages, the edges commonly display a neat abrasion. The butts are relatively narrow and thin, with measurements 5.4 ± 2.3 mm wide by 1.9 ± 0.9 mm thick in set F, and 5 ± 2.2 mm by 1.7 ± 0.8 mm in set G. The average dimensions of the first-line blades are on average wider than in the rest of the sequence, with a width of 11.5 ± 3.4 mm and a thickness of 2.5 ± 1 mm in set F, and 10.6 ± 3.7 mm for 2.5 ± 1.2 mm in set G. However, the length of the products remains small – 28 ± 10.4 mm in set F and 27.5 ± 13.4 mm in set G – due to the reduced dimensions of the cores, which rarely exceed 50 mm. The presence of very well-made laminar blanks, usually with an “S” profile and parallel edges, could indicate occasional use of pressure. Nevertheless, as the technical diagnostic criteria are often similar to those of indirect percussion in these conditions (François Briois, pers. comm.), its identification remains uncertain.

The first-line blanks, showing the greatest straightness, are exclusively selected for the shaping of inverse bitruncations (BG3; n = 110 in set F and n = 71 in set G), mostly represented by the “Montclus arrowheads” (BG32; respectively n = 80 and n = 47) (Table 2). The domestic tools – mainly made up of truncations (TR3; n = 12 and n = 7), end-scrapers (GR2; n = 20 and n = 9) and few notched blades (BA25; n = 10 and n = 2) – are carried out on knapping waste or products from other phases of the reduction sequence. The flakes, although still strongly present in these assemblages (n = 1,573 and n = 1,692), are rarely retouched (n = 35 in both sets). The same is true of local flints, which are in the minority in the domestic tools of set F (n = 8) and totally absent in set G.

7 Discussion

7.1 The Occupation of La Grande Rivoire: Continuity vs Break

On a diachronic level, these first results already allow us to observe the presence of several common elements throughout the different occupation phases at La Grande Rivoire, which could indicate a certain continuity of traditions during the neolithisation. First of all, we notice the persistence of raw material supply networks across the sequence. In all the lithic assemblages, we note the presence of the same local and regional flint types, despite certain variations in their proportions and the appearance or disappearance of some anecdotal types. We also observe a comparable management of siliceous materials for all productions. The regional flints showing the best quality at knapping – generally flints with blond/light brown shades – are selected preferentially for the shaping of domestic tools and more particularly for the arrowheads. Due to their relatively heterogeneous quality, the local flints are used for the production of flakes, which are recovered as blanks for domestic tools or as expedient tools. The local flints, often of inferior quality, are preferred for the production of flakes, those being used for domestic tools or simply in their raw state. Even the set E, which will be distinguished by a sudden dominance of regional flints with yellow/mustard brown hues and orange zonation, shows a preferential selection of the blond/light brown flints for the shaping of arrowheads, following the same trend observed in the rest of the sequence.

The modules used for the cores – mostly pebbles or small blocks – have relatively comparable dimensions throughout the assemblages. On the basis of the examination of the intact cortical elements, the tested blocks and the few cores having known a short exploitation, it was determined that the selected volumes measures on average between 40 and 50 mm of sides, only rarely exceeding 60 mm. The dimensions of the blades are therefore largely dictated by these limits with first-line products displaying an average length around 30 mm. Only a growth of the measurement of the widths of the blades will be really perceptible, passing on average from 10 mm at the beginning of the sequence to a progressive normalisation at 12 mm in the top of the stratigraphy (Figure 9a).

Figure 9

Evolution of the average widths of the first-line laminar blanks (a) and arrowheads (b) according to the raw materials used.

On the contrary, other criteria will rather evoke ruptures in the technical systems, whether they are clear or progressive. First, there is a redefinition of the management of local raw materials. While these flints are exploited almost equally with regional resources in the lower part of the sequence, their rate gradually decreased toward the top of the stratigraphy to reach only 15%. This devaluation of local materials can also be seen in the volume of flint used: while the mass represents 78% in set A–B – dominating regional flints – this rate drops to 36% in set G. Their use for the shaping of domestic tools and arrowheads also follows a similar trend. While local materials are used for almost 10% of the retouched pieces of sets A–B (n = 12) and C (n = 24), then rising to almost 20% in sets D (n = 11) and E (n = 15), they are only used for the realisation of 3% (n = 8) of the tools of set F and are completely absent from the retouched elements of set G. They gradually give up their place in the blade production throughout the neolithisation period, with a selection that favours more and more regional siliceous materials. The latter, showing a better overall quality of the extraction of regular blades, are better suited to the extraction of wider and standardised blanks in the upper levels, what local flints, with a coarse grain often desilicified or with and diaclasts, cannot offer.

We also note the sudden disappearance of the microburin technique as the method of splitting for the blanks, which is commonly used for the manufacture of direct bitruncations (BG2) in the Second Mesolithic. Thus, from the set C onward, this method is replaced by flexing break. This technical change can be explained, in part, by a production of thinner blades, which can be simply fractured by bending, and would have led to the rejection of the technique adopted by the indigenous (F. Briois, pers. comm.). Despite this change, the number of geometric bitruncations attributable to the Mesolithic remains significant within the arrowheads of this assemblage. Although fragmented by flexing, these pieces are still shaped following the same technical patterns as in the lower part of the stratigraphy, suggesting the preservation of some Mesolithic traditions in the Early Neolithic period, at least in the early stages, and not a complete change in technical behaviour. Although a dozen microburins are counted within the assemblage of set C, they are present only at the interface with the underlying level and seem to be relegated to taphonomic phenomena and not to a truly anthropogenic fact.

The direct bitruncations (BG2) globally follow a movement similar to that of the abandonment of the microburin technique. Dominant in the lithic assemblages at the beginning of the sequence (i.e. in set A–B), the direct bitruncations are quickly surpassed, then replaced, by inverse bitruncations (BG3), several of which are thinned by direct low angle retouches (“Montclus arrowheads”, or BG32), which are in the majority of the rest of the stratigraphy. This type of arrowheads, used exclusively as transversal pieces and no longer as apical elements or barbs, thus becomes the dominant type for the occupations of the middle and upper levels of La Grande Rivoire. In parallel, there is an important dimensional differentiation between the bitruncations found at the bottom of the sequence and at its top (Figure 9b). The arrowheads of set A–B show an average width of 10.1 mm for 2.3 mm thickness. These dimensions are significantly smaller than those of set G, with respective measurements of 12.9 mm by 2.4 mm. The difference is even more marked with the pieces of set E, where the bitruncations have an average width of 13.5 mm for 2.6 mm in thickness.

Nevertheless, we note the presence of several direct bitruncations with direct low-angle retouches (here identified as BG24) in the middle of the sequence (sets C–E), also employed as transversal arrowheads, as seems to be indicated by the observation of burin-like spin off removals perpendicular to the slicing edge, suggesting a violent impact characteristic of such use (Albarello, 1986; Nuzhnyi, 1989; Philibert, 2016). However, these pieces are distinguished from the typical “Montclus arrowheads” on a technical level. In this case, after the segmentation of the blank, the direct low-angle retouches are carried out in the first instance. The two truncations are then shaped in a second step by means of direct abrupt retouches. Although these particular direct bitruncations show comparable intentions to the BG32, their low-angle retouches, however, are limited only to the edge of the truncations, whereas it has strictly a thinning objective for the “Montclus arrowheads.”

7.2 Evolutionary Sequence of Lithic Assemblages

The six assemblages identified in “decapages” d127 to d147 of square metres PR17–21 and P16 (sector NR16–21) at La Grande Rivoire yielded 1,023 retouched pieces, mostly made of blades (82%; n = 835). Regional raw materials are preferentially selected for their shaping, with a representation rate of 80% (n = 669). The tools are mainly made on first-line blades, i.e. a little more than 70% (n = 587) of all the retouched lithic elements.

7.2.1 General Evolution of Lithic Tools

The entire lithic assemblage – except for indeterminate backed edges (BA42; n = 5), a flake with bifacial retouches (PB1) and a flake with clactonian retouches (CC12) – was processed using factorial correspondence analysis (FCA, Figure 10). Microburins (MB1) – as waste of the breaking of the blanks used for the shaping of Mesolithic bitruncations – were added as supplementary elements. Arrowheads with bifacial retouches (BG4 and PB31) were also displayed in this way.

Figure 10

FCA based on the percentages of the main lithic tool classes by archaeological sets. The dots correspond to the archaeological sets and the triangles to the different types of tools. The dark triangles correspond to the supplementary elements.

Initially, we observe the formation of a parabola, following the Guttman effect, between the coherent archaeological assemblages and the different types of lithic tools. Axis 1 (59% inertia) shows an opposition between backed edges blades (BA2), truncations (TR2 and TR3) and direct bitruncations (BG2) – subsequently with microburins (MB1) – associated with sets A–B and C, and inverse bitruncations (BG3) or related types (BG4 and PB31), these latter being related to sets D–G. Second, axis 2 (29.1% inertia) provides a coherent diachronic dimension, opposing the “mixed” industries or those in the neolithisation process (sets C–E), from assemblages with a stable or stabilised composition, which may be related to the Second Mesolithic (set A–B) or to the beginning of the Early Neolithic (sets F and G).

The parabola effect obtained by the multivariate analysis of the typological compositions of the lithic industries in the different assemblages suggests a certain evolution. The sequence begins with the first group, composed of the furniture of the set A–B, with the presence of blades with backed edges blades (BA2) and side-scrapers on blades (RA2). This assemblage is second related to the microburin technique, identified by its characteristic waste, which is used for the splitting of the narrow bladelets. The first group thus stands out from the other lithic assemblages (sets C–G), which favour the use of flexing break as splitting method for the blanks. A second group is subsequently identified by set C, which shows an affiliation to direct bitruncations (BG2) and asymmetrical bitruncations with oblique truncations (BG1). A third group distinguishes itself from the previous one, uniting sets D and E, with the presence of pointed blades or piercers (BA3) and pieces with irregular removals (IR2). This group is further characterised by a similarity to the geometrics with truncation and bifacial retouches (BG4). Lastly, we find a fourth group, made up of the materials from sets F and G, with a significant presence of inverse bitruncations (BG3), scrapers on flakes (GR1) and – in a second time – pieces with short bifacial retouches (PB31).

7.2.2 General Evolution of the Arrowheads

The totality of the lithic arrowheads, which amount to 515 pieces, was also subjected to an FCA (Figure 11). This analysis of the proportions of the different types of arrowheads allows us to make observations comparable to the conclusions already present with the retouched lithic industries, although with more precision. Thus, there is a clear contrast between the material from the lower “decapages” (d145 to d147) and the rest of the chronological sequence, separated by axis 1 (80.5% inertia). In this case, the first group, formed by the set A–B, is characterised by a strong connection with the direct bitruncations with symmetrical points with oblique truncations (BG17) and geometrics with asymmetrical (BG21) and symmetrical (BG23) bitruncations.

Figure 11

FCA based on the percentages of the main lithic arrowhead classes by archaeological sets. The dots correspond to the archeological sets and the triangles to the different types of tools. The dark triangles correspond to the supplementary elements. Drawings: R. Picavet et J.-B. Lajoux.

On the other end, axis 2 (10.9% inertia) traces a diachronic subdivision in the succession of the sets. Above the horizontal, we find the bitruncations with direct low-angle retouches and direct truncations (BG24), altern bitruncations (BG33) and the altern bitruncations and direct low-angle retouches (BG34), which are associated with sets C and D. To these, we can add in second time the geometrics with direct and bifacial truncation (BG41). In direct opposition are the inverse bitruncations (BG31) and the inverse bitruncations with direct low-angle retouches (BG32), associated with sets E–G. These arrowheads are linked to the geometrics with inverse and bifacial truncation (BG42) and pieces with short bifacial retouches (PB31).

Again, this multivariate analysis of the arrowheads typology reveals a parabola that evoked a chronology through the material productions of the site (Figures 11 and 12). The first group, which appears at the bottom of the stratigraphy with industries of set A–B, is globally linked to the traditions of the end of the Second Mesolithic. It is characterised by the presence of symmetrical points with oblique truncations (BG17) and asymmetrical (BG21) and symmetrical direct bitruncations (BG23), whose blanks are fractured by the microburin technique.

Figure 12

Proposed interpretation of the lithic arrowhead typologies of the different archaeological sets. Drawings: R. Picavet and J.-B. Lajoux. CAD: M.-A. Dallaire.

The middle of the sequence (sets C and D) is characterised by a significant diversification of arrowheads types. The presence of altern bitruncations (BG33), altern bitruncations and direct low-angle retouches (BG34), as well as bitruncations with direct low-angle retouches and direct truncations (BG24), can be recognised. The latter, being original arrowheads, show intentions comparable to those of the “Montclus arrowheads” (BG32), without being a faithful copy. Less well standardised, the direct low-angle retouches of the direct bitruncations have never a thinning function, unlike the “Montclus.”

The upper assemblages, comprising the material of sets E–G, are characterised by a dominant presence of inverse bitruncations (BG31), but more particularly of inverse bitruncations with direct low-angle retouches (BG32). These transversal arrowheads, generally well normed, are associated in a second time with other types wearing bifacial retouches (BG42 and PB31), which display intentions close to the “Montclus arrowheads,” without revealing such a sustained morpho-typological standardisation.

8 Conclusion and Perspectives

The results presented here of our typo-technological study of the lithic industries of La Grande Rivoire already make it possible to establish a first sketch of the chronocultural evolution of the occupations stretching from the recent phase of the Second Mesolithic to the beginning of the Early Neolithic. We also have a better understanding of the character of the “mixed” lithic industries, a key element of our studies, and thus of the Mesolithic–Early Neolithic “transition.” Moreover, although behavioural changes are discernible between the lower and upper “decapages,” attributable, respectively, to the last hunter-gatherer-collector and the first agro-pastoral groups in the Vercors massif, there is a significant variation over time.

In this case, on a diachronic level, the gap between the technical patterns of the two “pure” traditions is not as straightforward as we might have thought in the past. Several behavioural changes are observed, suggesting the possibility of exchanges or interactions between the indigenous societies and the newcomers. Thus, although new concepts are appearing in the industries – such as the inverse bitruncations or the use of flexing break, which replaced the microburin technique – certain practices of the indigenous traditions remain perceptible until the establishment of the Early Neolithic. Given the complexity of the “mixed sets” and the richness of the lithic material, new analyses of the spatial distribution of the artefacts have yet to be carried out to have a more complete understanding of the occupations at the heart of this “transition.” New statistical analyses will also allow us to refine our knowledge of the rates of the neolithisation in this particular pre-Alpine context.

Of course, these results cannot claim to answer the whole question of the neolithisation in the Northern French Alps, focusing only on lithic technical systems. They will thus be enhanced by the contribution of complementary studies covering the other industries of the last hunter-gatherer-collectors and agropastoralists, particularly the bone industry, ceramics – especially for its appearance in this particular context – as well as the exploitation of animal or plant species, whether domesticated or not. Our study nevertheless allows us to further refine the representation of the chronocultural sequence of lithic industries from the recent phase of the Second Mesolithic to the Early Neolithic at La Grande Rivoire rockshelter. It also enables us to synthesise a certain reference for the neolithisation in the Vercors massif, which will allow, in future perspectives, to clarify our perception of this particular process in the northern French Alps through comparative studies of the industries of other shelters and full sites in the region, and even more distant spheres.