• -10 000
  • -9 000
  • -8 000
  • -7 000
  • -6 000
  • -5 000
  • -4 000
  • -3 000
  • -2 000
  • -1 000
  • 0
  • 1 000
  • 2 000

Docteur Jean-Georges Rozoy

Résumé des abréviations utilisées dans les articles : consulter la liste.





The chronological limits of the Epipalaeolithic ('Mesolithic') -like those of all periods in prehistory - must be based on human achievements, such as industries and methods of production, and not on extrinsic geological phenomena. If the beginning of the period is defined by the end of glaciation, then circular reasoning is being employed (climatic change would have caused industrial change) and it becomes impossible to make comparisons between different natural regions.

Each of the major periods of prehistory has its own technical and social unity. This unity derives from the techniques used for the most important activity - in this context, the procurement of food. The basis of subsistence in the Epipalaeolithic was hunting with a bow (the technique of the bow and arrow is discussed). This technique is demonstrated by the microlithization of armatures. The large-scale and persistent use of sharp-pointed armatures weighing less than 5 g must serve to define and delimit this epoch. The geometric character of these pieces is of no significance.

Three major chronological subdivisions can be recognized based on diversification in the classes of armatures and their techniques. These factors suggest that a greater variety of hunting methods came to be used. The technical changes (which were gradual, inter-related and developed independently) occurred in a geographical mosaic, and this provides evidence for the multiplicity of centres of invention and the easy circulation of ideas.

From an economic viewpoint the Epipalaeolithic was definitely not evolving towards food production. Rather the reverse - it received this technique from outside contacts. On the other hand, in demographic, cultural and intellectual spheres, it achieved a slow maturation - substantial increases in population, detailed familiarity with the land, improved capacity for abstract thought, and the understanding of biological mechanisms. Indeed, these developments provided the starting point from which the spread of the Neolithic became possible (but not necessary).

The successive changes which marked the Epipalaeolithic occurred, on three occasions, each shortly before a change of climate; if cause and effect is to be perceived here, then it might seem that it was the inventions that caused the climatic change. It can also be suggested that these innovations were a response to constant pressure from the environment - and not just from its modifications - by a species that is particularly adaptable.


In contrast to other divisions between prehistoric periods, which are well defined - sometimes too well defined, thus concealing continuity between them -the distinction between the Palaeolithic in its final stages and the 'Mesolithic' remains particularly obscure to many people. This is true not only for specialists in other periods, which is perhaps understandable, but also for most Mesolithic scholars. The contrived term 'Epipalaeolithic' is often used to indicate the short interval of post-Magdalenian time before the advent of 'true cultures with geometric microliths', and care is taken not to specify whether this 'Epipalaeolithic' should be assigned to one side of the dividing line or the other. In the end, since students must be given guidelines, the date of 10,000 BP is usually adopted; this has the advantage of being a 'round number' (!) and also refers to the bipartition of the Scandinavian ice sheet, the official end - such things have to be fixed - of the 'Würm' or 'Weichsel' Glaciation (neither its name nor its subdivisions are important, since only its closing stages are relevant in the present context). Thus one finds authors, some of them being most authoritative in this respect, stating that the Mesolithic is 'the period which follows the Ice Age and precedes the introduction of food production'.


As I had occasion to recall in a recent issue of Mesolithic Miscellany (Rozoy 1984 a), this definition of the Epipalaeolithic is fundamentally unsound, since it begins with a geological event, extrinsic to man (with very different consequences in different regions), and ends with a purely human advance. A prehistoric epoch must be defined by human activities. To draw the boundary between Palaeolithic and Mesolithic as corresponding to the end of glaciation immediately precludes all possibility of recording the same events in other climatic zones; what does 'the end of the Ice Age' mean in the Near East - a most sensitive area in this respect - or in the heart of Africa? It is by no means certain that the stages in the evolution of human society witnessed in Europe also occurred in the same way in India or Central America; what was done in three stages in Europe could have been done in four, five, or only two stages elsewhere. If we are to make comparisons, however, we must discount local circumstances in order to concentrate on the criteria of general applicability.

The most striking quality of general applicability possessed by mankind is his remarkable independence, of widely varying environmental conditions.

What other animal or vegetable species could we refer to which finds a livelihood from the Poles to the Equator, as well in desert as amid swamp, in mountain or plain, fishing as much in fresh water as in the sea? - we could go on ad infinitum. Imagine trying to transport ants from Europe to the Amazon (where different species of ants live), reindeer to the Sahara, or wheat to the Tropics; wheat, reindeer and ants are all strictly conditioned by their environment and could not tolerate major variations.

Man lives in every part of this planet (and looks forward to conquering space), which is why C.J. Thomsen with his Three Age System (Stone, Bronze, Iron Ages) as early as 1836 (Thomsen 1836, 1837) and then G. de Mortillet (1869) based the prehistoric classification - which is still broadly in use - on products of human activity, as also did Sir John Lubbock (1865) in making the distinction between Palaeolithic and Neolithic. All our classifications are based on human activity, whether concerned with tool shapes, economy, or even social organization.

How, then, does it come about that intelligent people still accept ('for lack of something better') for the Palaeolithic/'Mesolithic' boundary alone a principle which would take us back to the 'Ages' of the cave bear, the mammoth and the reindeer in the ephemeral geological classification used by Lartet (Lartet and Christy 1865) ? The reason is simple; it is not possible for them, as it was for Lubbock when considering the Palaeolithic and the Neolithic, to contrast the two epochs point by point, since they have had no success to date in relating the tools described (burins, scrapers, piercers and javelins, on the one hand, and Tardenois points, trapezes and notched blades, on the other) to the basic technical actions which produced them. A perfect example is the paper by Gilead (1984).

Some of the tools of the Upper Palaeolithic show similarities over the time-range from the Aurignacian to the Magdalenian (and the Pavlovian), to the extent that the same type-list can be used to describe assemblages, with only a few groups of tools being specific to particular periods (e.g. Solutrean tools -shouldered points, laurel leaves, etc.). Our purpose here is not to argue for or against the utility or efficiency of such lists of tool types, but merely to record that the Upper Palaeolithic toolkit can be described using just a single type-list. The same applies to the toolkits of the Middle Palaeolithic, on the one hand, and the 'Mesolithic' (or Epipalaeolithic - the word is unimportant) on the other. This provides evidence for the existence of a technical community and, beyond, a socio-cultural one. For the Upper Palaeolithic this unity was based on hunting in large groups with the spear and javelin.

Whether we have elucidated more or less completely the why and the wherefore of these facts for each period, or not, should not prevent us from retaining them; each of our great periods is based on technical and social unity. This is particularly clear for the Epipalaeolithic ('Mesolithic'), for it is a state of society whose basic technique is at least as well established as that of the Neolithic. The latter is based on arable cultivation and animal husbandry -although some hunting, with its own individual character, was still practised occasionally - and the very diverse forms of Neolithic industries do not prevent us from distinguishing the common basis. The common basis for the Epipalaeolithic ('Mesolithic') is also well established - hunting with a bow.


G. Schwantes (1923, 1952) sought to use the axe as the distinctive trait separating the Palaeolithic from the 'Mesolithic' - or, to be exact, the Palaeolithic from the Neolithic, since in 1923 the idea of the 'Mesolithic' was not generally accepted and cultures with microliths were assumed by many people to be early Neolithic. The use of the axe as a distinctive trait only appears to be an advance over the use of the phrase 'Postglacial', because the axe, although it is of course a tool, is strictly dependent on climate and the growth of forests. In any case, many cultures with microliths lack flint axes. This does not mean that the hunters in question did not know how to cut wood, for they had blades of bone which may have been used as axes (Rozoy 1978: 988-996). For a discussion of the criteria used to define the Palaeolithic/Epipalaeolithic boundary, see Rozoy (1978: 894-895).

We must focus our attention on the most important activity - namely, the procurement of food. There had been no appreciable deforestation which could have contributed to subsistence activities in the Epipalaeolithic. The essential item of equipment for the acquisition of food was the bow and arrow, involving the use of sharp-pointed microliths. The widespread distribution of microlithic armatures is much better established in Europe than that of the axe - with the exception of the great Russian loess plain where there is no stone, and where armatures, tiny ones, were made of bone. The proportion of armatures in Epipalaeolithic assemblages varies from c. 20% (Ardennian) to c. 60% (Tardenoisian), and is sometimes as high as 85% (the Causses group). This proportion is seldom less than 10% of the tools present, while the rare Magdalenian sites which have produced armatures (e.g. Gare de Couze) have only 3-5% at most. And so we find once more the high proportions of armatures (10-50%) that were known in the Solutrean, but the Solutrean pieces were much larger (Fig. 3).

Some odd myths concerning microlithic armatures must be dispelled once and for all. Firstly, there is their so-called 'geometric' character, something which has no chronological unity nor, more importantly, any technical significance. Non-geometric Tardenois points were made by the same methods as the scalene triangles or crescents which accompany them, and they were used for the same purposes. This applies also to the 'Sauveterre points' when compared to the little short scalenes from the Midi, and to the points with oblique truncations found everywhere - the simplest, even the basic, type. It is not the geometric character which is important, but the microlithization of the sharp-pointed armatures which began long before the appearance of the geometrics (Azilian points, Valorguian points, etc.).

Another well-worn expression is the term 'microlith' itself. Many tiny piercers or retouched bladelets are, in the strictest sense, 'micro-liths' - being less than 5 cm long and less than 4 mm thick. So, of course, are the backed bladelets that the Magdalenians used so extensively. What is significant - and most scholars realize it, however confusedly - is the concept of a pointed armature of low weight (less than 10 g and, in practice, usually less than 2 g).

The technique of oblique truncation on bladelets is the basis of nearly all European microlith production. Very sharp points can, of course, be made in other ways - by shaping the end of a bladelet (see, e.g., Rozoy 1978: pl. 5, nos 4, 5), for example, or by adapting the technique of backing. Other methods were subsequently developed in the Neolithic, with transverse arrowheads and points shaped by surface retouch. Essentially, however, the very sharp points produced in the Mesolithic were obtained by the conjunction of an oblique truncation with an un-retouched cutting edge (Fig. 1). These elements both have their uses, as we shall see in the following sections.

It is worth recalling that the first appearance of this technique, as a systematic and generally adopted phenomenon, occurred in north Germany during Dryas III - in the Ahrensburgian. Rahir (1920) and Rust (1937, 1943) understood its importance well and rightly assigned the Ahrensburgian of Remouchamps and of Ahrensburg to the 'Mesolithic' - initially called 'Tardenoisian' throughout Europe. It is important to remember that over one hundred arrows were found at Ahrensburg (Fig. 1), some with the flint tang inserted into the wooden arrow-shaft (for a list of finds of hafted armatures, see Rozoy 1978: 956-968). Later Mesolithic arrows (Fig. 2) are no different.


In terms of technique, Epipalaeolithic toolkits are distinguished (from those used throughout the 25,000-year tradition of the Upper Palaeolithic) by the use of numerous microlithic points made of flint, replacing earlier types of points which were characteristically much larger, scarce, and made of antler (and for this reason necessitated burins for their preparation). There was, of course, the Solutrean phase (Smith 1966) when carefully shaped stone points were used, but these were the same size as Aurignacian and Magdalenian antler points and were much heavier on account of the greater density of the material. Nevertheless, the essential difference in size remains. Palaeolithic points were much heavier, weighing more than 10 g and, in the majority of cases, more than 20 g (Fig. 3); exceptions are rare. It can also be observed that microlithic armatures could be produced much more quickly (in 10-20 seconds), which made it possible to produce them in large quantities. Spears are retained, but some arrows are lost.


The arrows recovered measure about 90 cm in length, with a diameter of one centimetre. They were fitted with points weighing 0.5-2.0 g, so that, together with a little mastic and fletching, the total weight would be 20-30 g; a heavier tip would have unbalanced them. An arrow travels at 100 km per hour, this initial high speed being the reason for its effectiveness (1/2 mv2). It is possible to shoot over distances of 100 or even 200 m, but in practice arrows are generally used over 20-50 m; beyond this, the force of penetration diminishes too much. An arrow is designed to shoot medium- and large-sized game, which it would pass through completely. The fundamental advantage of the bow is its precision. The role of the flint point is obviously to pierce the fur and hide of the animal; the armature is positioned so that its cutting edge hits the target obliquely and cuts the skin and organs it meets. Flint points are more efficient in this respect than steel points (Pope 1962). A second armature placed along the shaft, as on the Loshult arrow (Fig. 2), makes the cut wider and enlarges the wound not only in the skin, but in all the tissues pierced. They are thus not barbs, like those on Magdalenian harpoons, but lateral cutting edges. Epipalaeolithic groups whose assemblages contain a very high percentage of armatures probably used several of these cutting tools on each shaft. It is not impossible that some of the backed bladelets of the Upper Magdalenian were mounted in much the same way, but these would be parallel to the shaft; the shaft would also have been much larger, judging from the dimensions of the points mounted at the head, and the analogy is therefore of secondary importance.

Behind each armature (or each group of two or three armatures) we should imagine such an arrow, propelled by a bow 1.60 m long and able to penetrate and pass right through a bear at fifty metres. The Epipalaeolithic was essentially the era of hunting with a bow. Modern attempts to study the social evolution of the Epipalaeolithic (e.g. Constandse-Westermann, Newell and Meiklejohn 1984; Newell 1984) would be even more interesting if greater account were taken of its socio-technical basis (for more details, see Rozoy 1978: 956-968, 1008-1020).


In the very early period (twelfth and eleventh millennia BP) there was no uniformity; each culture adapted its traditional technique, in its own way, to include the new invention. This was a transitional period - in one area armatures were points with curved backs, in another points with straight backs, elsewhere points with truncated bases, or with a tang. Only the Ahrensburgian culture invented truncation on bladelets, while other cultures solved the problem in other ways. There was, however, one common characteristic: each regional group made only one class of armature at a time - Istres points in the Valorguian (Fig. 4), Tjonger points in the north, the pre-Tardenoisian of Les Blanchères, etc. Only the Ahrensburgian, dated to Dryas III, used tanged points at the same time as simple points made by truncation, together with some triangles and atypical trapezes (Fig. 1).

In the early and middle periods (tenth and ninth millennia BP) oblique truncations on bladelets spread everywhere, although treated differently in different regions to produce either geometric or non-geometric armatures. More importantly, in each regional group two or more classes of armatures were in use at the same time (Fig. 5), and one class was only abandoned (gradually) when a third had become established. This diversity strongly suggests diversification of types of bows and arrows, although this cannot currently be proven.

In the recent and final periods (eighth and seventh millennia BP) trapezes made on more regular blades and bladelets were used, with diverse types occurring in different regions, and still with several classes in use at the same time (Fig. 6). Nothing is known of the nature of the intervening technical development - it must have been very important since it spread rapidly by 7850 BP - and it is difficult to pinpoint its origin geographically. In the southern two-thirds of France trapezes were made with retouch almost invariably on the left side (point upwards), thus following on the unvarying tradition of the middle period. In Belgium the point is retouched on the right side; in the Tardenoisian 75-80% of trapezes are retouched on the right side. This practice must have been acquired from an external source, since the earlier tradition used retouch on the left side, in accordance with the rest of France. In Belgium, however, from c. 8200 BP a regional variety of armatures was being made (the 'feuilles de gui' group) with retouch on the right (Fig. 5). It is probable, therefore, that this was the geographical origin of trapezes. The time-range over which trapezes existed alongside armatures of the middle period varies from region to region (Fig. 7).

Trapezes did not penetrate (or penetrated only slightly) some regions (e.g. Ardenne, Beaugency), where people adapted their traditional armatures but using ancillary techniques similar to those of their neighbours (e.g. style of retouch, flat inverse retouch, etc.).

The final period (seventh millennium BP) is characterized mainly by the use of flat inverse retouch, probably indicative of a new hafting method, and this was applied to quite different types of armature. Small transverse arrowheads (Montclus and Le Châtelet types - Rozoy 1978) appeared in the southern half of France.


The changes in armatures, as in the industries generally, occurred in a mosaic-like way (Chavaillon et al. 1978). These changes were gradual; for instance, the armatures of the middle period took more than a millennium to disappear after the invention of the trapeze. On the one hand these changes were inter-related within a single culture; on the other, they were autonomous from one culture to another, and their diffusion around each centre of invention varied. Where diffusion was widespread, as in the case of oblique truncations in the early and middle periods and of trapezes in the recent and final periods, it led to the development of many variations, reflecting the personality of the groups using them.

The diffusion of inventions, which was sometimes very rapid, is evidence of the fundamental homogeneity of technical and cultural levels, and of the existence of a network of communications between regional groups.


It has often been postulated that this period represents a 'transition' (this is the etymological meaning of the term 'Mesolithic') towards the Neolithic. In a recently published work, Grahame Clark even speaks of a 'prelude' (Clark 1980). This point of view can be considered from at least three aspects, whether related or not: (i) technical progress towards food production, for example by specialization on a single animal species, systematic harvesting of wild cereals, and so forth; (ii) the objective preparation (obviously unplanned and unconscious) of a human setting favourable to the introduction of food production, in terms of population density and familiarity with the land; and (iii) subjective preparation (no less unplanned and unconscious) as a consequence of intellectual evolution. It is appropriate to examine these elements separately, without losing sight of their constant overlapping and interactions.

(a) Economic aspects

In Europe the Epipalaeolithic (“Mesolithic”) reveals no signs whatsoever of movement towards food production. There was no specialization on a single species, rather the contrary. The Magdalenians, who relied on reindeer for 80% of their subsistence (at least at certain times and places, perhaps for ecological reasons) were far more specialized than Epipalaeolithic peoples. In the latter case, conversely, a clear and systematic diversification is perceived (Fig. 8), reflected in the variety of classes of armatures mounted at the tips of arrow-shafts.

Most of the traces of harvesting or of the domestication of animals which some people thought they had perceived in the Epipalaeolithic have, in the long term, been shown to derive from old or very insecure data; for example, the sheep at Téviec -represented by a single tooth which has since vanished (Péquart et al. 1937; Rozoy 1978: 787). Those which can be substantiated from recent research are all in the Mediterranean region. The sheep (Ovis) at Gazel and at Dourgne (Guilaine 1976a, 1976b, 1980) are contemporary with Neolithic cultures, from which the animals could have been stolen or perhaps borrowed. In any case, this concerns only the very end of the period. Some harvested plants (Lens, Vicia, Pisum, Cicer) are known from ninth millennium BP contexts at Fontbregoua and L'Abeurador (Guilaine 1980); but gathering is not cultivating, and in this instance did not continue.

No major deforestation, even on a local scale, took place before the middle of the Atlantic period -that is, before the intrusion of the Neolithic (this event took place earlier in the Mediterranean zone, and precisely where the Cardial culture was in existence as early as 8000 BP). Sedentism has often been claimed to exist in the Epipalaeolithic, but this is very difficult to prove; it seems unlikely in most regions where there are numerous sites with only slight occupation debris. Traces of solid buildings are not found until the Neolithic, when they appear as early as its earliest stage (LBK) in the region under discussion. An exception could be claimed for coastal sites, but the coasts of this period are only known in a part of Denmark and in the north of the British Isles (Bonsall 1988), the rest now being submerged. For example, at the time of their occupation, Téviec (Péquart et al. 1937) and Hoédic (Péquart and Péquart 1954) were several kilometres inland (Rozoy 1978: 779-796, pl. 198).

There is no serious evidence of a trend towards food production during the twelfth, eleventh, tenth or ninth millennia BP. As for the cultures with typical trapezes and notched blades, the evidence which does exist is confined to periods and regions where borrowing - whether of objects or practices -from Neolithic neighbours was possible and, therefore, likely.

Naturally, such conclusions cannot be extended to apply to the Earth as a whole. The transition certainly occurred somewhere and probably at many points, especially in the Near East. But it cannot be inferred from this that analogous processes occurred in Europe, even if the technical starting point (bow, arrows and microlithic armatures) was the same. Nor can the possibility be excluded that in other regions, in America for example, a direct transition from the Palaeolithic to the Neolithic took place, without the intervening period of the bowmen. Such problems must be investigated without any preconceptions.

(b) Demographic aspects

The situation is rather different. It is certainly an extremely delicate matter to attempt calculations of past populations from archaeological evidence. Whatever basis is considered - skeletal remains (Constandse Westermann et al. 1984), faunal remains, or site distributions - it would be necessary, as Bordes et al. (1972) have stated, to revert to Biblical chronology - to compress the whole of the development of mankind into the period since 4004 BC -and this period would still be too long! Too little is preserved, still less brought to light, and of that only about one per cent is excavated. Thus at Fère-en-Tardenois (Allée Tortue), which is considered to be an extremely rich deposit, with some ten concentrations of flint artifacts, we found some five thousand tools 'relating to occupation during the seventh millennium BP - that is, about five tools a year (or 50 for each decade). This does not imply a very large population, and certainly not even a tenth (or even a hundredth) of the population that actually existed.

We are thus reduced to estimating populations from the nutritional capacity of the territory (Rozoy 1978: 1055-1060, 1105-1109), and to making comparisons between periods in terms of the number and importance (based on numbers of tools) of sites known. Nonetheless, these, albeit approximate, comparisons give the following indications for the Epipalaeolithic:

a) The population had expanded to occupy the whole of the available land area, even areas which might be the size of a modern French 'canton' or even 'commune', in contrast to the Upper Magdalenian when the population was scattered in isolated islets (Rozoy 1985: map), with very large territories completely empty (Brittany, Champagne).

b) Sites were much smaller (both in area and in the number of tools), but occur in much greater numbers than in the Upper Magdalenian. This leads to the conclusion that social groups were more restricted in size (but with a larger number of primary groups) or that there was more mobility; probably both of these were true.

c) Cultural units were more restricted in size -about 100 km across compared to 500 km in the Magdalenian. Thus the composition of Upper Magdalenian toolkits is the same in Belgium and in the area to the south east of Paris (Rozoy 1984b, 1985, in press), the qualitative difference being at best the presence of 'zinken' in Belgium (even then not found at every site). In the Epipalaeolithic in the same region a number of cultural units can be distinguished: the Limburgian, the Ardennian group (which ought perhaps to be subdivided), the northern Tardenoisian, the southern Tardenoisian, the group from the Loing valley (Les Richoux, Les Champs Bertin - Pigeot 1973), the Beaugencian and, on the fringes, the groups from Belloy and from Luxemburg (Spier 1982), without counting those which have not yet been identified. There is a corresponding number of different assemblage types, each with its own style and special tools.

Population estimates suggest the strong probability that each cultural group contained roughly the same number of people (1000-3000 persons, including children) in both the Upper Palaeolithic and the Epipalaeolithic, but that the overall population showed a marked increase by a factor of between five and ten. On the other hand, the primary groups appear to have been smaller (15-20 people instead of 50-80 people), with greater individualism. Thus each of the regional Epipalaeolithic groups maintained itself (with minor variations in boundaries) from millennium to millennium in the same territory, as can be seen in the persistence of certain technical details ('bordage', for example - see Rozoy 1967: 212, 1968: 336) peculiar to each region.

From this one can deduce no sedentism proper but well-defined movements within a limited territory of which the topographical detail would be well known and appreciated (Turnbull 1961, 1968). In other words, people were acquiring an exhaustive and intimate knowledge of the land, to a degree that was not possible in the final Palaeolithic. Moreover, the evident cultural unity indicates the existence of many close relationships between primary groups in an awareness of their regional unity. The rapid diffusion of inventions also points to the existence of good inter-cultural relations. The 'frontiers' of cultures were certainly easily penetrated, and socio-technical levels were absolutely equivalent.

All of these elements produced conditions that were, by their very nature, favourable to the subsequent diffusion of the Neolithic. Nevertheless, it is unlikely that any single element was an indispensable preliminary, and these conditions certainly did not in any way make the transition to food production necessary - a hunting way of life could have continued for a long time. In any case, this traditional mode of subsistence continued for several millennia alongside cultures in which food production was already well developed. Within France itself, nearly two thousand years passed between the Cardial neolithization along the Mediterranean littoral and the general spread of food production to the rest of the country. The evidence from Montclus (Rozoy 1978: 298-299, 1190) seems to show that, far from evolving inexorably towards cultivation and stock-breeding, Epipalaeolithic peoples preferred to sever contact with their neolithized cousins.

(c) Cultural and psychological aspects

The cultural and intellectual changes that occurred during the Epipalaeolithic, although very difficult to demonstrate, were perhaps the most important of all. Some of them derived from the sociological changes brought about by the use of the bow. Restricted group size encouraged the existence and autonomy of the nuclear family and probably made possible the greater role of individualism, initiative and responsibility. A more numerous overall population would have permitted greater social interaction. Life was more fulfilling and more easily assured.

Art, and non-utilitarian evidence in general, gives only a very broad indication of the intellectual levels attained by the whole population. First, there is the almost total disappearance of the figurative animal art of the Upper Palaeolithic. Modern research, and that of A. Leroi-Gourhan (1965) in particular, has clearly established that the naturalistic figures constituted symbols, and were not a basis for magic. It can be argued that these signs, or abstractions, which still took concrete form, had to be figurative in order to be understood. Towards the end of the Magdalenian phase abstract symbols multiplied, sometimes associated systematically with particular naturalistic figures (e.g. Rozoy 1984b, 1985) and in the Epipalaeolithic only the abstract symbols survived. It could be argued, therefore, that sufficient numbers of Epipalaeolithic people were now able to grasp these abstractions. This would indicate an increase in the capacity for abstract thought amongst the general population. The grouping of the figures in scenes is an analogous indication, and the centring of these scenes on humans instead of animals is yet another.

Abstract thought is also seen in the ability to count. In the Upper Palaeolithic there is evidence of this in the form of sets of signs engraved on bone ('marques de chasse' or 'hunting tallies'), but very few of these convey a system which we can comprehend or even perceive. In this context, reference can be made to the 'churinga' from the grotte de la Roche at Lalinde, dated to the very end of the Palaeolithic (Fig. 9), on which there are alternating series of five horizontal and five vertical marks. There are five groups of each, but the last vertical group contains only four marks, and the aesthetic character of the whole, surrounded by other marks, casts doubt on the view that its primary purpose was for calculation. The engraver at least would have known how to count to five on his fingers. The common use of abstract numbers up to five is thus assured.

In the Ahrensburgian assemblage from Remouchamps (microlithic and 'Mesolithic', it must be stressed) there is a bone which is very similar to the hunting tallies, but the subtle organization of this piece is beyond our understanding, in spite of the bold hypothesis put forward by its finder (Dewez 1974a, 19746). It was accompanied by another example with more systematic markings - groups of five cupmarks, arranged like the number 'five' on dominoes (Fig. 9). This piece seems therefore, and even more clearly than at Lalinde, to involve first-order notation, which was in itself a considerable advance, occurring around 11,000 BP. Nevertheless, there are no groups of 5x5 or 10x5. Therefore, second-order notation (i.e. groups of 5 or 10, counted as second-order units) is not attested, but it is possible that it was reached later. This cannot be taken further for lack of other evidence, but it does show continuity from the Magdalenian and the extent of progress in the capacity for abstract thought. There is little need to emphasize the importance of a system of numeration for the transition to food production.

At the same time we see the disappearance of female figurines centred upon maternity, or rather potential maternity - for Magdalenian outline drawings and statuettes, and even Gravettian statuettes, do not portray pregnancy, merely those parts of the female anatomy connected with reproduction and childbirth. The Epipalaeolithic statuette from Gaban, for instance, is already realistic and complete. It is possible that this corresponds to an understanding of the process of procreation. Such an advance is clearly important in facilitating the transition to food production – to sow to obtain a harvest, to put the cow to the bull to obtain calves.

Relations to the Neolithic - summary

As we have seen, the 5 000 years of the Epipalaeolithic, considered as a whole, constituted as a matter of objective fact a basis without which neolithization could not have taken place. This basis, sociological as much as intellectual, made the transition to food production possible. It did not make this necessary and, in temperate Europe, it did not include any technical element tending towards food production. In short, neolithization needed this social and intellectual basis which had developed slowly over more than 100,000 years during the Palaeolithic, not really reaching maturity until the Epipalaeolithic ('Mesolithic'). From this basis, food production made its appearance in those places with a favourable climate (the Near East, southern China, Central America), while less favourable regions followed the traditional way of life based on hunting. These regions then received the new techniques and attitudes from external sources and were able to adopt these easily since the necessary social and intellectual level had already been widely achieved.


It is often stated that the technical changes that occurred in prehistoric societies, and in particular that which marked the beginning of the Epipalaeolithic, were the result of environmental change. This thesis appears so self-evident to many workers, that they confine themselves to formulating it without attempting any demonstration. For them, it is in a sense a postulate, a presupposition, an act of faith, even a ritual invocation. Underlying it, whether consciously or not, is for some people undoubtedly an unwillingness to accept any explanation for change other than a material one, with any doubts on this point being automatically denounced as spiritualism, even deism or creationism. Praehistoric cultures are thus described as being so many 'adaptations' to the environment or, above all, to changes in the environment. It is in fact a mis-use of the scholastic principle 'post hoc, ergo propter hoc'.

I have already pointed out the splendid circular reasoning which begins by defining the 'Mesolithic' as being Postglacial (disregarding the Ahrensburgian people and certain other groups who manufactured armatures of 1 g in weight), and then concludes that it was climatic change that caused changes in industries.

It goes without saying that we must applaud the desire of scientists to study nature (including man) exactly as it is, and - I would add - on the basis of the facts, all the facts, and nothing but the facts; and consequently without any presupposition - most particularly about causes and mechanisms. The facts are these ...

The technical basis of the Epipalaeolithic is characterized by certain essential changes. Firstly, there was the microlithization of armatures, which was gradual and lasted for more than a millennium. Its starting point was Palaeolithic armatures and techniques even in the Ahrensburgian region, since microlithization began in the Tjongerian with curved backed points. This microlithization began as early as the end of Dryas II; it was already evident in some Magdalenian assemblages when various experiments were tried out (e.g. Gare de Couze - Bordes and Fitte 1964; Rozoy 1978: pl. 85, 85b) and in which the Azilian point appears before the end of Dryas II. The process of microlithization continued through the Alleröd and was only accomplished fully in Dryas III (La Borie del Rey - Coulonges 1963; Remouchamps, etc.). In the Valorguian, too, Istres points ('pointes fusiformes' - cf. Rozoy 1978: 275) which are roughly shaped in layer 12 at Cornille (Fig. 4), dating to the Lateglacial (Dryas II), are abundant in layer 10 which is dated to the Alleröd (layer 11 is sterile). The glaciation had by no means ended, and in Dryas III it was still extremely cold; yet microlithization continued and became increasingly significant. It is evident, therefore, that this first technical change preceded climatic change (Dryas II to Alleröd) and intensified with returning cold.

The second change marked the beginning of the early and middle periods. It consisted of the widespread adoption of the technique of oblique truncation on bladelets (with or without the microburin as a waste product, according to which group is considered) and, more importantly, the diversification of classes of armatures made easier by this technique, which allows for multiple combinations. Thus one truncation makes a simple point, two could make a scalene or a Tardenois point, and so forth (Fig. 5). It is this change which other authors see as marking the beginning of the 'Mesolithic proper'. On the whole these innovations, which were also introduced very gradually, correspond to the Pre-Boreal. However, as mentioned above, their invention goes back to the Ahrensburgian in Dryas III, where tanged points co-existed with points with oblique truncation, points with unilateral retouch, and a few triangles (Fig. 1; Rozoy 1978: 115-155). Once it had appeared here, the technique spread very widely. Again, technical innovation preceded climatic change.

The third major technical change was the introduction of typical trapezes and debitage in the 'Montbani' style. This is securely dated everywhere to c. 7800-7850 BP (Fig. 6) - yet the Atlantic period did not begin until c. 7500 BP.

So we can see that, three times, the changes in industries occurred before the changes in climate. Once may be chance, twice is a coincidence, but three times becomes a proof. Since it is out of the question to contest the principle of cause and effect ('post hoc, ergo propter hoc'), we are led to the following revolutionary conclusion: changes in industries caused climatic changes (three times) !

Such a conclusion may seem somewhat surprising if looked at only superficially; it has, nonetheless, some elements of merit - it is just as solidly founded as the previous hypothesis, according to which climatic changes caused industrial change.

Moreover, this conclusion has the subordinate advantage of actually conforming to facts. Its socio-logical and philosophical implications are considerable - mind over matter. But these considerations lie outside our present concerns, and the task of developing them must be left to others. Let us simply note in passing that popular wisdom has decided that atomic experiments have unsettled the weather (Fig. 10).


If we really wish (as we should) to study nature purely as it is, 'without any external addition', the very first consideration must be the constant pressure that the environment exerts on all its constituent parts, including the human species. It is not only environmental change which affects man, but rather the ever-present environment itself: there is the speed of running animals; there is the danger of wolves; there is cold, snow, floods which come with the thaw, permafrost, the shortage of game if spring comes too late, the inevitable 'plague' of mosquitoes when spring comes at last; there are females and young animals that are more wary than the males, the charge of bison when attacked, and so on. These are the stimuli for continual adaptation. Of course, increasing warmth was far from being a catastrophe. The reindeer departed, but red deer multiplied (to such an extent that roe deer were not hunted, or only occasionally). Wild boar also proliferated. Mammoth had disappeared, but there were more aurochs and more bison than ever (so long as the forest did not become too dense). Moreover, fuel for warmth was abundant, although there was less need of it.

Moreover, if a milder climate was the cause of technical change, it would have occurred tens of millennia earlier in the warm regions of the Earth; yet this did not happen. The same would be true of Europe at the beginning of each interglacial.

The question of whether modifications in animal species (morphological changes or behavioural changes, if such there were) were a response to climatic changes, must be left to the zoologists. It is well known, however, that the main characteristic of the human species is non-specialization, which gives rise to adaptability; and it is this trait, quite natural though it is, which sets our species apart.

The course of praehistory is marked out by inventions, some of considerable significance - stone throwing, the club, manufactured tools, fire, clothes, artificial shelter, bifacially flaked axes, the Levallois technique, bone-working, blade debitage, figurative art, the bow and arrow, the domestication of plants and animals, and so forth. These inventions were responses to permanent pressure from the environment, and their general spread depended on the ability of the whole population to exploit them according to the level of conceptual advance at the time. The bow and arrow were probably invented several times in different periods and regions (e.g. in the Spanish Solutrean of Parpallo). They could not spread and become the standard of that time until mental evolution (probably conditioned by its physical basis, i.e. the biology of the brain) was sufficiently advanced throughout the population as a whole. The driving force behind this evolution has not been provided just by changes in the environment (from which one could escape by moving, as animals did), but rather by the constant pressure exercised by the environment on a species that has been far more capable than all others of continual adaptation.


1. Elsewhere in this volume, the spelling 'Epi-Palaeolithic' has been adopted. At the author's request, the spelling used in this paper follows that (Epipaléolithique) in his previous publications (e.g. Rozoy 1978). (Ed.)

Acknowledgements :

Both the author and the Editor would like to thank Jane Murray for producing the basic translation of the original French text. Brian Barron, Department of French, University of Edinburgh, gave assistance with the more difficult pieces of translation. The author would also like to express his gratitude to Clive Bonsall for his kindness and patience.


BONSALL, C. (1988) Morton and Lussa Wood: the case for early Flandrian settlement in Scotland. Scottish Archaeological Review, 5: 30-33.

BORDES, F. and FITTE, P. (1964) Microlithes du Magdalénien supérieur de la Gare de Couze, Dordogne. Miscelanea en Homenaje al Abate Henri Breuil. Barcelona, Instituto de Prehistoria y Arqueologia: 259-267.

BORDES, F., RIGAUD, J.P. and SONNEVILLE-BORDES, D. DE (1972) Des buts, problèmes et limites de l'archéologie paléolithique. Quaternaria, 16: 15-34.

CHAVAILLON, J., CHAVAILLON, N., HOURS, F. and PIPERNO, M. (1978) Le début et la fin de l'Acheuléen à Melka-Kunture: méthodologie pour l'étude des changements de civilisation. Bulletin de la Société Préhistorique Française: 105-115.

CLARK, J.G.D. (1980) Mesolithic Prelude. The Palaeolithic-Neolithic Transition in Old World Prehistory. Edinburgh, University Press.

CONSTANDSE-WESTERMANN, T.S., NEWELL, R.R. and MEIKLE-JOHN, C. (1984) Human biological background of population dynamics in the western European Mesolithic. Proceedings of the Koninklijke Nedelandse Akademie van Wetenschappen, B87: 139-223.

COULONGES, L. (1963) Magdalénien et Périgordien post-glaciaires. La grotte de la Borie del Rey (Lot et Garonne). Gallia Préhistoire, 6: 1-29

DEWEZ, M. et al. (1974a) Nouvelles recherches à la grotte de Remouchamps. Bulletin de la Société Royale Belge d'Anthropologie et Préhistoire, 85: 5-160.

DEWEZ, M. (1974b) New hypotheses concerning two engraved bones from la grotte de Remouchamps, Belgium. World Archaeology, 5(3): 337-345.

FAIZANT, J. (1982) Cartoon from 'Jours de France'.

GILEAD, I. (1984) Is the term 'Epipalaeolithic' relevant to Levantine prehistory? Current Anthropology, 25(2): 227-229.

GUILAINE, J. (1976a) Grotte Gazel. Le Mésolithique et le Néolithique. Abri de Dourgne II. IXème Congrès U.I.S.P.P. (Nice). Livret-Guide de l’Excursion C2 Provence et Languedoc Méditerranéen, Sites Paléolithiques et Néolithiques: 279-284, 312-315.

GUILAINE, J. (1976b) Les civilisations néolithiques dans les Pyrénées. In J. Guilaine (ed.), La Préhistoire Française, tome 2. Paris, Centre National de la Recherche Scientifique: 326-327

GUILAINE, J. (1980) Dossiers de l’Archéologie, 44 (June 1980).

LARTET, E. and CHRISTY, H. (1865-1875) Reliquiae Aquitanicae. London, Williams and Norgate. Paris, Baillière.

LEROI-GOURHAN, A. (1965) Préhistoire de I'Art Occidental. Paris, Mazenod.

LUBBOCK, J. (1865) Praehistoric Times as Illustrated by Ancient Remains and Customs of Modern Savages. London, Williams and Norgate.

MALMER, M.P. (1968) Die Mikrolithen in dem Pfeilfund von Loshult. Meddelanden frân Lunds universitets historiska museum, 1966-1968: 249-255.

MATHIASSEN, T. (1948) Danske Oldsager. I. /Eldre Stenalder. Copenhagen, Nordisk Forlag.

MORTILLET, G. DE (1869) Essai d'une classification des cavernes et des stations sous abri, fondée sur le produit de l'industrie humaine. Comptes Rendus de I'Académie des Sciences de Paris, 1.3.68; and Bulletin de la Société Géologique de France, série 2, 26 (1868-1869): 583-587; and Matériaux pour servir à l'Histoire Primitive et Naturelle de l'Homme, serie 2,_n° 3: 172-179.

NEWELL, R.R. (1984) On the Mesolithic contribution to the social evolution of western European society. In J. Bintliff (ed.), European Social Evolution: Archaeological Perspectives. Bradford, University of Bradford: 69-82.

PEQUART, M., PEQUART, ST-J., BOULE, M. and VALLOIS, H. (1937) Téviec, Station-Nécropole Mésolithique du Morbihan (Archives de l'lnstitut de Paléontologie Humaine, 18). Paris, Masson.

PEQUART, M. and PEQUART, ST-J. (1954) Hoédic. Deuxième Station-Nécropole du Mésolithique Côtier Armoricain. Antwerp, De Sikkel.

PETERSSON, M. (1951) Mikrolithen als Pfeilspitzen. Ein Fund aus dem Lilla Loshult-Moor, Ksp. Loshult, Skane. Meddelanden frân Lunds universitets historiska museum, 1950-1951: 123-137

PIGEOT, N. (1973) Analyse typologique d'une série de 163 pointes du Tardenois. Cahiers du Centre de Recherches Préhistoriques de I'Université de Paris 1 (U.E.R. Art et Archéologie), 1: 19-29.

POPE, S.T. (1962) A Study of Bows and Arrows. Berkeley, University of California Press.

RAHIR, E. (1920) L'habitat tardenoisien des grottes de Remouchamps, Chaleux et Montaigle. L'industrie tardenoisienne et son évolution en Belgique. Bulletin de la Société Royale d'Anthropologie de Bruxelles, 35: 31-91

ROZOY, J-G. (1967) Essai d'adaptation des méthodes statistiques à l'Epipaléolithique ('Mésolithique'). Liste-type provisoire et premiers résultats. Bulletin de la Société Préhistorique Française, 64: 209-226.

ROZOY, J-G. (1968) Typologie de l'Epipaléolithique ('Mésolithique') franco-beige, introduction. Bulletin de la Société Préhistorique Française, 65: 335—342.

ROZOY, J-G. (1978) Les Derniers Chasseurs. L'Epipaléolithique en France et en Belgique. Essai de Synthèse (Société archéologique champenoise, numéro spécial). Charleville, J-G. Rozoy.

ROZOY, J-G. (1984a) The age of red deer or of bowmen? Mesolithic Miscellany, 5(2): 14-16

ROZOY, J-G. (1984b) Le Magdalénien supérieur à gravures de Roc-La-Tour I. Congrès de Nivelles (2ème Congrès de I'Association des Cercles Francophones d'Histoire et d'Archéologie de Belgique, 1984), vol. 2: 122-132.

ROZOY, J-G. (1985) Deux modes de chasse sur le plateau ardennais. Jagen und Sammeln. Festschrift fur H.G. Bandi (Jahrbuch des Bernischen Historischen Museums, 63/64 1983-84). Berne, Stämpfli: 245-252.

ROZOY, J-G. (in press) Le Magdalénien supérieur de Roc-Tour I. Helinium.

RUST, A. (1937) Das altsteinzeitliche Rentierjagerlager Meiendorf Neumünster, Karl Wachholtz.

RUST, A. (1943) Die alt- und mittelsteinzeitlichen Funde Stellmoor. Neumünster. Karl Wachholtz.

SCHWANTES, G. (1923) Das Beil als Scheide zwischen Paläolithikum und Neolithikum. Archiv für Anthropologie, N.F. 20: 13-41

SCHWANTES, G. (1952) Wann endete das Paläolithikum? Zeit schrift für Morphologie und Anthropologie, 44(1-2): 237-244.

SMITH, P. (1966) Le Solutréen en France. Bordeaux, Delmas

SONNEVILLE-BORDES, D. DE (1960) Le Paléolithique Supérieur en Périgord. Bordeaux, Delmas.

SONNEVILLE-BORDES, D. DE (1967) La Préhistoire Moderne Périgueux, Fanlac.

SPIER, F. (1982) Les stations épipaléolithiques-mésolithiques de la commune de Hespérange (Grand-Duché de Luxembourg In A. Gob and F. Spier (eds), Le Mésolithique entre Rhin et Meuse. Luxembourg, Société Préhistorique Luxembourgeoise: 229-256).

THOMSEN, C.J. (1836) Ledetraad til Nordisk Oldkyndighed Copenhagen, Det Kongelige nordiske oldskriftselskab.

THOMSEN, C.J. (1837) Leitfaden zur nordische Althertumskunde Copenhagen, Det Kongelige nordiske oldskriftselskab.

TROELS-SMITH, J. (1961) Et Pileskraft fra Tidlig Maglemosetid Aarb0ger for Nordisk Oldkyndighed og Historie, 1961: 122-146.

TURNBULL, CM. (1961) Le Peuple de la Forêt. Paris, Stock.

TURNBULL, CM. (1968) The importance of flux in two hunting societies. In R.B. Lee and I. De Vore (eds), Man the Hunter Chicago, Aldine: 132-137.


Fig. 1. - Ahrensburgian: armatures from Geldrop III—1.

Biologisch-Archeologisch Instituut, Groningen ; arrows from Stellmoor (after Rust 1943). Microlithization is achieved during full glacial conditions.

Fig. 2. - Arrows from the early and middle periods.

Loshult (after Petersson 1951; Malmer 1968), Vinkelmose (after Troels-Smith 1961), Holmegaard (after Mathiassen 1948). There is no technical difference from the Ahrensburgian examples.

Fig. 3. - Upper Palaeolithic armatures.

1-4, Solutrean (after Smith 1966); 5-9, Magdalenian (after Sonneville-Bordes 1960). The dimensions are quite out of proportion with those of the Epipalaeolithic, and hence inconsistent with use as arrowheads.

Fig. 4. - Valorguian armatures.

7-11, Cornille layer 12 (Late-glacial, proto-Valorguian); 1 6, Cornille layer 10a (Alleröd). These form a single class of armatures that is already clearly microlithic.

Fig. 5. - Armatures from the early and middle periods.

1-17, Roc-La-Tour II; 18-26, Oirschot V (B.A.I., Groningen). There are several classes of armatures at each site. Note the armatures with invasive retouch (24-26) which have the truncation on the right side as early as the middle period.

Fig. 6. - Armatures from the recent and final periods.

1-21, Montclus layer 16; 22-30, Allée Tortue X, Fère-en-Tardenois. Stippling indicates marks (of hafting?). In each group there are three main classes.

Fig. 7. - The co-existence of trapezes and armatures inherited from the middle period.

This varies in different cultures, but the trapezes themselves can include several types (e.g. Fig. 6).

Fig. 8. - Proportion of animals killed (in terms of meat weight) at Epipalaeolithic sites.

There are always at least two staple sources; these differ in different areas (even adjacent areas) and periods.

Fig. 9. - Rare examples of evidence of numeration.

1 - 'churinga' from Lalinde, final Magdalenian (slightly reduced, after Sonneville Bordes 1967); 2, 3 - engraved bones from Remouchamps (after Dewez 1974). First-order numeracy only is proven.

Fig. 10. - They are upsetting the weather with their flint-chipping experiments (after Faizant 1982).

This opinion has as much scientific value as saying that changes in environment are the causes of changes in industries.

Résumé des abréviations utilisées dans les articles : consulter la liste.

© Jean-Georges Rozoy - Tous droits réservés 2016-2019