Introduction: Chauvet, the most dated cave in the world?

Discovered in 1994, figurative and non-figurative images in the Chauvet-Pont-d’Arc cave (Ardèche, France) number in excess of 450 and constitute some of the most impressive examples of Upper Palaeolithic cave art known to archaeology. The images were initially identified through their style, content and technique as Mid and Late Upper Palaeolithic (Gravettian–Early Magdalenian) in age, but the results of a preliminary radiocarbon dating programme led to the reassignment of many of the cave's impressive ‘black series’ charcoal drawings to the Early Upper Palaeolithic (Aurignacian). If this assessment is valid, our whole understanding of the development of more than 25000 years of Upper Palaeolithic art would be shown to be incorrect. Debate began almost immediately, however, and continues to this day, although the scholarly community seems largely to have accepted these preliminary results.

We believe that the current chronology of human and animal activity in Chauvet is unrepresentative and problematic, and we urge that the radiocarbon dating of its art, as it stands, should be wholly rejected. No comprehensive statement exists detailing the number of radiocarbon dates that have been produced on materials from the cave: we estimate that about 99 have been published. Ostensibly, this is a large sample—Valladas et al. (Reference Valladas, Clottes and Geneste2004: 82) referred to the site as “le mieux daté du monde [the best dated in the world]”—but that is actually misleading. It probably contains the most comprehensively dated context in the world (29 measurements on a single hearth), but this does not make its art the best dated; instead, it is the most problematic. About 47 of the published measurements are on charcoal fragments, over 40 of which derive from its floor. The entire case for an early age for its art rests solely on eight dubious measurements on charcoal from six images in three chambers (e.g. Valladas et al. Reference Valladas, Tisnerat, Arnold, Evin, Oberlin and Clottes2001, Reference Valladas, Tisnérat-Laborde, Cachier, Kaltnecker, Arnold, Oberlin and Évin2005; Petrognani Reference Petrognani2013: 31). Comprehensive information detailing pretreatment and measurement has never been published, and for only one image was the charcoal split into humic and humin fractions. Results were taken at face value to indicate an Aurignacian age for the art, although they were clearly problematic and have been seen by the Chauvet team themselves as being in need of further verification (see below). As this simple protocol of splitting into humic and humin fractions was not followed for the remainder of the assays, we have no confidence in them.

We feel that it is now imperative to conduct a measured, objective, informed and detailed debate about the age of Chauvet's art. As the Chauvet team have been unable to address the numerous contradictions in their model point by point, we suggest that their ‘early chronology’ should now be rejected. In its place, we propose a later chronology for the cave. We suggest that this is a more appropriate default chronological model for Chauvet, at least until it can be modified or eliminated on scientific grounds. It will be shown that all evidence from the cave indicates artistic activity over several phases from the Gravettian to Early Magdalenian. More speculatively, some of the cave's red images may belong to the Aurignacian, although these are not among the images that the Chauvet team have suggested belong to this period.

The debate so far

We do not wish to rehearse the history of the debate over the Chauvet images. Suffice it to say that there are a number of problematic issues. A fuller explanation of this debate is included in supplementary online material. Our main conclusions are:

• • The few dates on the art itself cannot be trusted and do not reflect the age of the art.

• • The wider radiocarbon dating programme for the cave indicates only that a small and unspecified number of hearths were lit in the cave 30–32 ka BP, and other human activity occurred after this. It is irrelevant to the age of the art.

• • The attribution of some of the cave's art to the Aurignacian is highly problematic and requires a number of assumptions that are not justified.

• • The number of entrances used by the cave's artists has not been established, and the closure of the current entrance has been inaccurately dated: whatever the case, data are consistent with access to the cave until at least 18 ka BP, and possibly much later.

• • The archaeology and palaeontology on the cave's floor are chrono-culturally undiagnostic and are irrelevant to the age of Chauvet's art.

• • In all cases, the art of Chauvet can be attributed to the Gravettian, Solutrean and Magdalenian. Wide artistic parallels with securely dated art from elsewhere make this the most parsimonious—and most likely—interpretation.

In light of this, we can now propose a new understanding of the Chauvet art in its chronological context.

Chauvet's art: stratigraphic, stylistic and thematic perspectives

The starting point for understanding Chauvet's art has been its division into two series: red and black. Clottes (Reference Clottes, Chauvet, Brunel Deschamps and Hillaire1996) drew attention to superimpositions from which it can be inferred that many (but not all) of the cave's ‘red series’ of images and signs were created earlier than most (but not all) of its ‘black series’. This superimposing occurs, for example, on the Daguet pendant in the Chamber of the Skull (Feruglio & Baffier Reference Feruglio and Baffier2005: 153; see also Alcolea & de Balbín Reference Alcolea González and de Balbín Behrmann2007). While the red series is relatively homogeneous and genuinely seems to pre-date much of the black series, the latter, by contrast, is heterogeneous and represents at least two distinct phases, and thus at least two periods. Relatively simple animal outlines were created in both red and black, although only the black series contains the cave's impressive charcoal drawings of horses, bison and lions. On technical grounds, therefore, it is over-simplistic to speak of an earlier ‘red’ and later ‘black’ series; in reality, there is an earlier series of relatively simple animal outlines and signs, generally in red but with black examples, followed by several phases of exclusively black depictions. As Züchner (e.g. Reference Züchner1999a), Alcolea and de Balbín (Reference Alcolea González and de Balbín Behrmann2007) and Combier and Jouve (Reference Combier and Jouve2012, Reference Combier and Jouve2014) have argued, the red animals can be assigned to the Gravettian and perhaps early Solutrean on the basis of stylistic comparison with examples clearly shown to relate to these periods (see Table 1); nobody would argue with this. It therefore follows that most (but not all) of the cave's black images must be later than the red, i.e. they must be late or post-Gravettian.

Table 1.

Summary of thematic, stylistic and technical aspects of Chauvet's art, and the chronology of parallels elsewhere.

Note that in almost all cases these parallels are consistent only with an attribution to the Gravettian-Magdalenian, and, with the exception of aspects of the red series, are never consistent with an attribution to the Aurignacian. It should be noted that all other examples of cave art in the broad region (i.e. Ardèche) are Solutrean and Magdalenian; it is conceivable that some are Gravettian, but there have been no serious arguments for any of an earlier age. Thus, our alternative chronology for the cave's art fits with this broader, regional picture.

The earliest phase of the black series is characterised by animal outlines drawn with technically simple (although highly naturalistic) methods, whereas the later phase includes several elements of preparation of the cave walls, shading, considerable attention to detail, group composition and perspective and movement, as well as a number of stylistic and thematic traits not seen anywhere else to date before the Solutrean (Feruglio & Baffier Reference Feruglio and Baffier2005: 152; see also Table 1). Some thematic differences also seem apparent between the two; Feruglio and Baffier (Reference Feruglio and Baffier2005: 154) note that the first phase includes depictions of cave bears, reindeer and mammoths, the former of which does not appear later. Thus, thematic, stylistic and technical differences distinguish all three recognised phases. A spatial distinction between most of the red and black images is also evident.

Most red signs and animal outlines are located close to the cave's current entrance, whereas the black drawings and engravings are located in the cave's current central and rearmost areas (e.g. Chauvet et al. Reference Chauvet, Brunel Deschamps and Hillaire1996: 111–13). Clear stylistic and spatial distinctions occur between the cave's depictions of cave bears (which cluster close to the current entrance) and lions (deeper in the cave). Robert-Lamblin (Reference Robert-Lamblin2005) assumed that the depictions of the two were contemporaneous and inferred a symbolic distinction between the two, but this is hard to justify. The cave bears are drawn as simple outlines, mostly in red; the lions, most in black, vary technically but are often drawn with shading, considerable attention to detail and are usually incorporated into scenes that depict perspective and movement (Clottes & Azéma Reference Clottes and Azéma2005a & Reference Clottes and Azémab). It is therefore highly likely that they belong to separate periods; the black and red outline drawings of bears dating to the earlier period, and the black charcoal drawings of lions to the later period. It is impossible to quantify the time separating these: as has been noted, “les deux phases de dessins noirs sont séparées par un temps difficilement quantifiable [the two phases of black drawings are separated by a period of time that is difficult to quantify]” (Feruglio & Baffier Reference Feruglio and Baffier2005: 153).

We can use the activities of another of the cave's denizens to improve this picture further. Cave bears left numerous claw marks on its walls, many of which ‘deface’ red and black depictions. Where claw marks clearly overlie lines in both black and red in the end chamber (Clottes Reference Clottes2003a: fig. 125), these lines plainly form part of relatively simple outline paintings of lions in both red and black (Clottes Reference Clottes2003a: fig. 126), assigned to the first black phase of the art by Feruglio and Baffier (Reference Feruglio and Baffier2005: 158). As one of these is depicted in red—‘sandwiched’ between the two black examples—it constitutes clear evidence of the contemporaneity of simple red and black images during the cave's earlier phase. All other images defaced by claw marks are similarly simple, e.g. on the panel of the rhinoceros in the Megaloceros gallery (Feruglio & Baffier Reference Feruglio and Baffier2005: 150–52). As we hope we can safely assume that the bears were not making decisions about which art to deface and which to ignore, the stratigraphic information of bear defacing provides an important chronological marker within the art phases. Numerous examples of claw marks deface the images of the first black phase, but no such marks deface images of the second (Feruglio & Baffier Reference Feruglio and Baffier2005). Thus, if we can assume that the bears had not ‘religiously respected’ the latter (Alcolea & de Balbín Reference Alcolea González and de Balbín Behrmann2007: 447), we can infer that the cave bears were present in the cave during or after the creation of the red series and the earliest black series, but that they were not present during or after the creation of the second black series. The Chauvet team conclude that cave bears did not use Chauvet after 23 ka BP (Bon et al. Reference Bon, Caudy, de Deiuleveult, Fosse, Philippe, Maksud, Beraud-Colomb, Bouzaid, Kefi, Laugier, Rousseau, Casane, van der Plicht and Elalouf2008, Reference Bon, Berthonaud, Fosse, Gély, Maksud, Vitalis, Philippe, van der Plicht and Elalouf2011; although taking errors into account, this should be 22 ka BP), and thus we can infer that the art of the second black series must be more recent than 22 ka BP. Further support for this relative chronology is the fact that the depiction of cave bears in Chauvet's art is always technically simple—fitting the earlier outline phase—and in no case is this species depicted in the second black phase (Feruglio & Baffier Reference Feruglio and Baffier2005: 157).

We can also use examples where art is clearly superimposed over claw marks or other art to develop the relative chronological scheme. Examples of this fall into two groups: the second series of black drawings, and engraved animals. A shaded charcoal drawing of a bison is superimposed upon claw marks on a wall perpendicular to the lion panel (Chauvet et al. Reference Chauvet, Brunel Deschamps and Hillaire1996: fig. 92). Another example is the image of three lions on the left panel in the Hillaire chamber, noted above; the lions are covered by claw marks—which are overlain by a simple outline engraving of a mammoth, clearly visible in fig. 79 of Chauvet et al. Reference Chauvet, Brunel Deschamps and Hillaire1996. Fine engraved lines—some of which seem to be part of animal engravings similar to a mammoth engraved in the Morel chamber—are superimposed upon bear claw marks (Le Guillou Reference Le Guillou and Clottes2003: 62). Therefore, shaded charcoal drawings and engraved animal outlines must post-date the simple red and black series and bear denning, and must be more recent than 23–22 ka BP.

Such a relative scheme fits with what we know of the wider artistic context, against which we must surely evaluate Chauvet's art. In all cases, wider thematic, stylistic and technical parallels support the notion that Chauvet's function as an artistic sanctuary spans the Gravettian, Solutrean and Early Magdalenian, and may even have persisted until the Middle Magdalenian (e.g. Züchner Reference Züchner1999a & Reference Züchnerb; see Table 1). The complex, four-phase sequence of production employed in the depiction of horses, aurochs and lions of the second black phase (Tosello & Fritz Reference Tosello and Fritz2005), and the attention to anatomical detail, perspective and movement have no parallels elsewhere before the Solutreo-Magdalenian (Züchner Reference Züchner1995, Reference Züchner1996, Reference Züchner1999a & Reference Züchnerb; Alcolea & de Balbín Reference Alcolea González and de Balbín Behrmann2007: 446; Pettitt et al. Reference Pettitt, Bahn, Züchner and Bahn2009). In Azéma's (Reference Azéma2004, Reference Azéma2010) erudite analysis of the depiction of movement in Upper Palaeolithic art, the overwhelming number of examples are post-Gravettian. The Chauvet team assign the cave's mostly naturalistic depictions of reindeer to the Aurignacian, yet no depictions of reindeer occur before the Magdalenian in any other Palaeolithic art, parietal or portable (Züchner Reference Züchner1999a; Djindjian Reference Djindjian and Otte2004). Feruglio and Baffier (Reference Feruglio and Baffier2005: 154) assign reindeer to the first of the black series on stylistic grounds; this would imply that in Chauvet this species was depicted in the Gravettian. But what are the grounds on which this assignation is made? Stylistically, one can discern an earlier phase in which reindeer are depicted “statiques et proportionnés, modelés et détaillés [static and well proportioned, modelled and detailed]” and a later in which they are “plus souvent en movement, aux extrémités non-terminées, aux membres démesurés et aux formes quasi expressionistes [more often in movement, with unfinished extremities, huge limbs and almost expressionist forms]” (Feruglio & Baffier Reference Feruglio and Baffier2005: 154). We see no reason why reindeer need not belong to two post-Gravettian (i.e. Solutrean or Magdalenian) phases; there are several similarities between these two broadly differing stylistic conventions (e.g. shoulder pelage is always depicted as a sinuous line) and no stratigraphic distinction between the two is apparent; we therefore suggest that they are best viewed as part of a stylistic continuum, and are thus probably close in time.

In a recent stylistic study of pre-Magdalenian cave art, an ‘Analyse Factorielle des Correspondences [Correspondence Factor Analysis]’ was carried out on multiple anatomical features in 148 complete horse figures from 26 sites (Petrognani Reference Petrognani2013: 138). To Petrognani's great surprise (Reference Petrognani2013: 148), the Chauvet horses were found to occupy a medial position between sites reliably dated to 25000 BP and sites dating to 18000 BP. His chosen explanation, as he fully accepts the early chronology for the cave's imagery, was that either the Chauvet horses were millennia ahead of their time, or (more likely, in his view) the result comes from the great variety of anatomical criteria used in the analysis, and means that Chauvet's horses are made up of both early and late features. Needless to say, we consider the result to be a clear endorsement of our own view—namely that the Chauvet horses are not early at all, but can indeed be attributed to periods around 25000 and 18000 BP and even later.

Finally, the ‘confronted rhinos’ provide a useful example of how best to interpret what seem initially to be contradictory sets of data. These animated images are clearly an integral part of the panel of horses in the cave's Hillaire chamber (Fritz & Tosello Reference Fritz, Tosello and Clottes2003). The composition, style and techniques demonstrated on this panel clearly show that it belongs to—and indeed could serve to define—the cave's second black phase. As discussed above, this must post-date 23 ka BP. Despite this stylistic evidence, the results of only three questionable radiocarbon measurements on samples of charcoal from the two rhinos have been taken to indicate that these images, and thus the whole panel of horses, date to 30–32 ka BP, that is up to ten millennia before the cave bears became locally extinct. Notwithstanding the chrono-stylistic problems these dates raise, cave bears must have religiously avoided the most technically achieved of Chauvet's art in this case. The alternatives, of course, are: either there is something wrong with the radiocarbon measurements that exist, or, because they pertain to the creation of the charcoal in the cave, they bear no relation to the date that the art was produced. Neither of these scenarios has been adequately addressed by the Chauvet team, despite the fact that Valladas et al. (Reference Valladas, Cachier, Maurice, Bernaldo de Quiros, Clottes, Cabrera Valdés, Uzquiano and Arnold1992) acknowledge that it is a serious possibility. Furthermore, in order to eliminate the possibility of contamination, charcoal samples can be split into two fractions—humic and humin—which should really be a matter of course with such minuscule samples. This has, however, only been undertaken and published for one of the images, a horse head. The results on each fraction were statistically distinct (around 20 and 29 ka BP respectively), and in the light of our criticism of such poor results these have been withdrawn pending ‘future verification’ (Valladas et al. Reference Valladas, Tisnérat-Laborde, Cachier, Kaltnecker, Arnold, Oberlin and Évin2005: 111). As this methodology apparently has not been used on the samples taken from the rhinos, how can we know if the results are reliable? The reputation of the early chronology for Chauvet may stand or fall on the confronted rhinos.

We also note in passing that the remarkably early dates (c. 32–33000 BP) obtained by the Gif laboratory for some black dots in Candamo Cave (Asturias), on which we cast doubt at the time (Pettitt & Bahn Reference Pettitt and Bahn2003), have now been declared unreliable by the laboratory, which has redated the same dots to c. 18–22000 BP (Corchón et al. Reference Corchón, Garate, Valladas, Rivero, Pons-Branchu, Ortega and Hernando2014).

An alternative chronology for Chauvet cave

A handful of highly problematic radiocarbon measurements were used 20 years ago to suggest that against all other indications the earliest art at Chauvet is Aurignacian. No further dates on the art have been added, nor has information pertinent to the complex chemistry of the samples been published. If we ignore these results—and we can justifiably do so for the reasons stated above—all lines of evidence point towards a parsimonious set of phases for activity in Chauvet cave, which we summarise here. We can, of course, only consider those phases of activity for which there is tangible evidence; we do not mean to imply that these are the only phases of activity in the cave. We regard the issue of whether there are any ‘Aurignacian age’ examples of art in the cave as genuinely open; although, if these do exist, we suggest a different set of examples drawn from sound parallels elsewhere, which the Chauvet team have missed. Our proposed alternative chronology for Chauvet cave can be summarised as follows:

• • Cave bears denned in the cave's central galleries. This was repeated frequently from at least 37 ka BP until their regional extinction by or after 23 ka BP.

• • Humans—either Homo neanderthalensis or Homo sapiens—entered the cave at least once around 30–32000 (¹⁴C) BP, either through the current entrance chamber or via the Morel chamber. They lit at least one hearth. They may have left a small amount of art; if this were the case we suggest it would be elements of the red series; some combination of hand stencils, dots and lines (Pike et al. Reference Pike, Hoffman, García-Diez, Pettitt, Alcolea, González-Sainz, de las Heras, Lasheras, Montez and Zilhão2012).

• • Humans (culturally Gravettian Homo sapiens) entered the cave around 28 ka BP, leaving torch wipes and at least one hearth, and art in the form of simple, naturalistic animal outlines in both black and red and a tracé chinois. Among the art, they depicted cave bears, having observed this animal in the cave's locale, possibly as part of wider activities that included moving the bones/carcasses of the bears. If examples of the red series do not belong to the earlier phase, they will belong to this phase. Artistic activity was restricted for the most part to the cave's outermost chambers, probably relating to access (through the current entrance or Morel chamber).

• • A later period of culturally Gravettian activity is possible. Some of the red series or early phase black series could belong to this period, although this is unclear. Cave bears could still be extant at this time.

• • The cliff face exterior to the current entrance collapsed somewhere between 22 and 15 ka BP, partially or wholly sealing this entrance.

• • The entrance to the cave was now elsewhere as a result of the collapse, perhaps the northern wall of the gallery of cross-hatchings. Solutrean or Early Magdalenian activities included the movement of cave bear bones and stalactite blocks, and the creation of most of the second phase of the black series. Unsurprisingly, this has numerous thematic, stylistic and technological similarities with Solutreo-Magdalenian art from elsewhere, from which it derives. Activity occurred much more frequently in the cave's ‘depths’ than in previous phases due to the redefined entrance(s).

Until the unlikely ‘official’ chronology has been proven beyond reasonable doubt, we suggest that our alternative chronology best fits the partially represented, partially studied, and partially published palaeontology, archaeology and art of the cave. As a result, it should stand as the default hypothesis for activity in Chauvet until it can be eliminated scientifically.

Supplementary material

To view supplementary material for this article, please visit http://dx.doi.org/10.15184/aqy.2015.21.