Introduction

A recent article by Bacon et al. (2023) makes the argument that the main objective of Palaeolithic art (parietal and portable) was to provide a numeric system serving as a time marker or calendar and therefore to be at the origin of “proto-writing.” This hypothesis had already been proposed by several authors (e.g., Aujoulat, 2005; D'Errico, 1989; de Smedt & de Cruz, 2011; Hayden & Villeneuve, 2011; Jègues-Wolkiewiez, 2005, 2014; Marshack, 1991 and recently by Taylor, 2021). These events are thought to have been very important for hunter-gatherer groups because their survival depended on knowing the life cycle of the fauna around them (Bacon et al., 2023: pp. 375–376). This proposal is consistent with Mithen’s idea that Palaeolithic art serves only for practical purposes linked to the procurement of food (Mithen, 1988).

According to the hypothesis of Bacon et al. (2023), Palaeolithic artists associated certain figurative representations with a type of sign to determine various important moments in the life of the animals, such as migration, mating, and birth. The signs used for this purpose are dots and lines (treated as equivalent as already suggested by Marshack (1991)) and the Y-shapes (two lines converging into one), which is considered to symbolize “giving birth.” The fact that they generally appear in a horizontal arrangement and with a standardized size and spacing suggests to Bacon et al. (2023) that they are numerical elements. The number of signs and the place of the “Y” (when present) are supposed to provide ethological data about the time at which an animal migrates, mates, or gives birth. Thus, this theory proposes the existence of a notational system acting as an extension of artificial memory, as other authors have already tried to demonstrate (e.g., d'Errico, 1995; d'Errico et al., 2018; Overmann, 2013). The theory focuses on this explanation without taking into consideration other possibilities.

According to the authors, “as none of the sequences in our database [...] contains more than 13 marks, they are consistent with the 13 lunar months of a year” (Bacon et al., 2023: p. 376). They envisage a calendar of thirteen months, although in their database cases appear such as Candamo, Montespan, Lascaux, La Madeleine, and Chauvet, with more than 13 signs. For instance, a horse from Lascaux bears 29 red dots or a reindeer from Chauvet Cave which displays 20 dots. Without any explanation, the authors start the transition from one year to the next during the so-called bonne saison when “rivers unfreeze, the snow melts, and the landscape begins to green” (Bacon et al., 2023: p. 377).

To demonstrate their hypothesis, the authors applied a series of statistical analyses that indicate two things. First, the calendar was better adjusted to mating and birth events (not so much with migration) in animals such as aurochs, bison, horses, mammoths, and fish. In the cases of cervids and especially caprids, the correlation between signs and ethological event appears less robust. Secondly, groupings without the Y-sign would indicate mating times and those including a “Y” the birth season. No argument is provided for this hypothesis. A bias, evident in the statistical analysis (see their Fig. 3), is that numbers higher than 13 have been discarded. In this way, migration disappears from the graphs and only mating and birth remain.

In spite of the commendable efforts made by the authors to objectify their hypothesis, we believe that the analysis suffers from certain errors and that certain liberties were taken that together make the results unreliable. Some of the archaeological data on which the hypothesis is based are doubtful or outdated and lead to a biased interpretation, as will be discussed in the following section.

Discussion

The Bacon et al. (2023) study takes liberties with the analysis that will be discussed in this section. We should not forget that the meaning of Palaeolithic art will remain unresolved forever. Erwin Panofsky (1939) taught us in “Studies of Iconology” that it is impossible to interpret works of art of an unknown civilization without having an intimate knowledge of the symbols, beliefs, and ways of life of those people, so that any interpretation of it must always be approached with a sound methodological basis avoiding presentism.

Creation of the Corpus

A striking point in the study discussed here is that the authors rely mainly on reproductions and photographs to construct the study corpus. However, many of the tracings consulted are now out of date. Paradigmatic examples are those made by Breuil at the beginning of the twentieth century. Many of his works are still a compulsory reference for caves such as Altamira, La Pasiega, El Castillo, La Pileta, or Les Trois-Frères, but researchers in the last century did not have the same means at their disposal as we do today. It is important to try to use the most recent bibliography with which we can better analyze the execution of the artistic motifs. It is the only way to avoid a bias in the information.

Below, we discuss some of the examples that we found to be erroneous in the work of Bacon et al. (2023). These types of inaccuracies invalidate the interpretation of Palaeolithic art as an ethological calendar. The errors may be classified in three ways: signs unrelated to animal figures, tracings or photographs with non-existent signs, and tracings or photographs with misinterpreted signs.

  1. A)

    Signs Not Related to Animal Figures

    The Asturian site of El Pindal Cave is an illustrative example of signs that are not fully related to animal figures. One of the most iconic depiction in this cave corresponds to one of the few mammoths that exist in the Cantabrian region. It is located on a panel in which only the proboscidean can be seen, with no association other than a large filled circle inside it. Despite the existence of more recent references, the authors have used the work of Alcalde del Río et al. (1911) to show the relationship of a series of lines with the mammoth. However, it is well known that Breuil did not pay attention to the spatial location and sometimes combined pictures from different places in the cave. This is the case of the panel with the mammoth. The team currently researching the cave, led by Olivia Rivero and María González-Pumariega, can corroborate that these sets of lines are not located in close contact to the mammoth (Fig. 1). Moreover, the horse shown in Fig. 1A is located far away (about 150 m) on a roof.

    Fig. 1
    figure 1

    Panel of the mammoth in the El Pindal Cave. The sets of lines appear further from the mammoth (168 cm away for the most visible and 75 cm for the least visible). A According to Alcalde del Río et al. (1911). B Photograph by Eric Robert

    Full size image
  2. B)

    Non-existent Signs

    In the Pair-non-Pair Cave, a Y-sign is assumed to be associated with the foreleg of a horse, as can be seen in the very schematic sketch from Daleau’s field notebook (1898) used by the authors (Fig. 2A). However, more recent work (Delluc & Delluc, 1991) and more modern photographs show that this Y-shape graphic unit does not exist and could even be mistaken for cracks in the wall (Fig. 2B, C).

    Fig. 2
    figure 2

    Horse from Pair-non-Pair. A Depiction used by Bacon et al. (2023) in their interpretation based on a tracing from F. Daleau's field notebook. B The same depiction according to Delluc and Delluc (1992) Redrawn by M. García-Bustos. C ©The Wendel Collection, Neanderthal Museum.

    Full size image

    The panel with a horse in Sotarriza Cave is another example of a Y-sign associated with a figurative image used by Bacon and colleagues. The authors use the reference of García Guinea (1975) to demonstrate this association. They have chosen this tracing among others that do not show any Y-sign under the horse’s belly (Fig. 3), for instance, the tracing by Breuil (Alcalde del Río et al., 1911) and the most recent by González Sainz and San Miguel Llamosas (2001).

    Fig. 3
    figure 3

    Horse from Sotarriza Cave. A Depiction by García Guinea (1975) used by Bacon et al. (2023) in their interpretation. Tracing by P. García Bustos. B According to Alcalde del Río et al. (1911). C According to González Sainz and San Miguel Llamosas (2001). D © César González-Sainz

    Full size image

    A last example concerns Atxurra Cave, which houses a Magdalenian iconographic ensemble with more than 100 engravings and paintings. In this case, the authors seem to have discovered "4 white dots in front of the breast on the horse’s left" (Supplementary 1, row 530 of Bacon et al. (2023)) from a low-quality photograph in an English digital newspaper (The Guardian Newspaper 27 May 2016). As can be seen in Fig. 4, these dots do not exist.

    Fig. 4
    figure 4

    Engraved horse in Atxurra (photograph by O. Rivero). As can be seen, there are no dots in front of the horse

    Full size image
  3. C)

    Misinterpreted Signs

    In Altxerri A Cave, there is a mistake in the morphology of the signs drawn by Barandiarán (1964) who interprets the group of signs superimposed on a bison as made up of lines, angles, and a Y-sign. However, in the most recent study of this figure carried out by Ruiz-Redondo (2014), these signs appear to be formed only by lines and angles in the form of a “V” (Fig. 5). Once again, the chosen tracing is the one which fits most closely what the authors are seeking.

    Fig. 5
    figure 5

    Bison with a group of engraved signs from Altxerri A. A Depiction by Barandiarán (1964) used by Bacon et al. (2023). Tracing by P. García Bustos. B More recent interpretation according to Ruiz-Redondo (2014)

    Full size image

    Another example comes from the main panel in Covaciella Cave where there is a composition with four large bison among other figures. In the monograph of the cave (García-Diez et al., 2015), it can be seen that there is a group of red signs in the lower part of the composition. The authors have used a photograph from a Web source with little detail to count six dots. However, it can be seen that there is in fact a V-shaped sign and that the number of the dots is larger, which causes a bias in the subsequent calculations (Fig. 6).

    Fig. 6
    figure 6

    Main panel in Covaciella Cave. A Photograph by M. González-Pumariega. B Tracing by P. García Bustos

    Full size image

    In the case of Tito Bustillo Cave, some reindeers have a dewlap of flesh and hairs below the neck which are often conventionally represented by a series of short lines. This convention is common to many caves, such as Font-de-Gaume, Altxerri, and Les Combarelles (Fig. 7). Bacon et al. (2023) incorrectly consider these conventions to be abstract signs in order to prove their ethological calendar. Another over-interpretation of signs associated with reindeer can be found in dots distributed along the flank of the animal, as can be seen in examples in such caves as La Forêt, La Mouthe, Font-de-Gaume, or Les Trois-Frères (Fig. 8). Sometimes they can also be represented as zigzags or lines. We think that these are fur patterns rather than signs per se, as they appear continuously in the same arrangement in numerous examples in different caves and are not placed differently as in the rest of the samples in their database.

    Fig. 7
    figure 7

    Main panel in the Tito Bustillo Cave. The lines are actually the delimitation of the reindeer’s chest fur. Photograph from Balbín Behrmann et al. (2022) and tracings from Balbín Behrmann and Moure Romanillo (1982)

    Full size image
    Fig. 8
    figure 8

    Reindeer with dots or lines on the body that may represent conventionally the colour gradient of the fur. A La Forêt (Leroi-Gourhan, 1984). B-C Altxerri A (Ruiz-Redondo, 2014). D-E La Mouthe (Breuil, 1952). F Font-de-Gaume (Capitan et al., 1910). G-I Les Trois-Frères (Breuil, 1952). J) La Madeleine (Tosello, 2003). K-L La Madeleine (Lartet and Christy, 1875). M Isturitz (tracing by O. Rivero).

    Full size image

Association Between Signs and Animals

The argument made by Bacon et al. (2023) rests on the notion of “association” of signs with animals. The use of this term is one of the weak points in their discussion because it remains implicit and is never defined. As shown in the figures in their article, Bacon et al. (2023) indistinctly consider various different cases: the signs are contained inside the animal, superimposed on its outline, or even located far away. The places they occupy within the figure, their size, or the combination of techniques are not given any importance in this study. All these drawbacks may lead to a bias in the results obtained.

Generally, an association is defined by the presence of various individual motifs (figurative or not) in a single space called a “panel.” It is often established that the maximum distance between motifs should not exceed the length of the biggest one (e.g., Barrière et al., 1986). For this reason, strictly speaking, the mammoth and the lines in Fig. 1 do not constitute a panel of associated figures.

Apriorisms

We believe that the authors make a series of apriorisms and assumptions that may lead to inaccurate results or conclusions. These ideas should have been discussed in their study in order to better understand and contextualize the analyses they subsequently carry out.

The first and most obvious case concerns the signs chosen for the study. The authors only consider a total of three abstract motifs: dots, lines, and Y-signs. Moreover, they do not differentiate between dots and lines, using them indiscriminately in their statistical analysis. Why are other geometric motifs or more complex signs such as claviforms or quadrangular ones not taken into account? The authors do not discuss the possibility that signs may have other meanings than their assumed numerical character, as for example representations of wounds, conventions for fur, changes of color, or simply as a technical licence of the artist (Fig. 9). A good example is given by the sculpted bison in La Madeleine, for which we share the interpretation of Paillet (1999). Most probably, the parallel lines represent the animal’s hairiness and not a numerical sequence (Fig. 10). Evidently, signs may fulfil other functions that we are unable to understand.

Fig. 9
figure 9

Different possible interpretations of dots and lines in some cases. A Les Trois-Frères (Breuil, 1952). B Chauvet (tracing by P. García Bustos). C Marsoulas (Breuil, 1952). D Candamo (Corchón et al., 2017). E, F Altxerri (Ruiz-Redondo, 2014)

Full size image
Fig. 10
figure 10

Sculpture from La Madeleine (Paillet, 1999). Bacon et al. (2023), in the supplementary materials S3, are not sure that “the sequence should be read left-right or right-left (IIIIIIIIIIIIY or YIIIIIIIIIIII).” As in many other pieces of Magdalenian art, the artist is representing the hairiness of the animal in the greatest detail with a series of parallel strokes

Full size image

The authors assume that dots and lines represent a notation system and therefore act as numbers. But a notation system can have many everyday or symbolic uses. They have chosen arbitrarily a lunar calendar, taking for granted the studies by Marshack (1991) and Jègues-Wolkiewiez (2014). They accept that the total number of dots/lines forming a group is equivalent to the month in which an important event in the life cycle of the “associated” animal occurs. Furthermore, they accept that each sign represents a month in the lunar year, although this is a presentist perspective, which cannot be demonstrated. On the other hand, the combination of these signs is given a great importance in their discourse, particularly the place of the Y-sign in a group formed by dots/lines. The authors assume that the Y-shape would symbolize birth by the convergence of two lines into one or representing two parted legs (Bacon et al., 2023: p. 377) and therefore would signify “to give birth.” The location of the sign in the group would indicate the month of parturition of the associated species. This hypothesis is very subjective and methodologically questionable.

Moreover, the interpretation of a sequence supposes that the reading direction is known. For instance, if the sequence “•Y••” is read from left to right, the Y is in second place, but if read in the other direction it is in third place. The authors assume that hunter-gatherer groups read these signs as numbers from left to right, again projecting Western culture onto a prehistoric society. The only piece of artwork for which they ask themselves the question is the sculpted bison of La Madeleine (Fig. 10), but in this case, the engravings are depicting the pilosity of the animal. It should also be noted that, consulting in the corpus presented by the authors in the supplementary materials, they mix indiscriminately dots and lines with other signs such as “V,” “Ʌ,” or “¦.” It is not said whether they have been removed from the statistical analysis or if they have been finally taken into account.

Interpretation

Finally, the authors propose an interpretation of Palaeolithic art as an ethological calendar, since hunting is essential for the survival of the group. However, we know that the consumed fauna does not match the fauna represented in rock art (e.g., Altuna, 1994; Moure Romanillo, 1990; Lorblanchet, 1995). For example, reindeer, which is one of the most consumed faunas in French Magdalenian, do not appear in their statistics, probably because they are unduly confused in the category “cervids”. As well we can mention the case of ibex and red deer, species hunted during the Palaeolithic period in Western Europe.  Both cervids and caprids are the species that are not statistically significant in their analyses. Also, small mammals frequently consumed such as rabbits, are lacking (another bias). All this suggests that a direct interpretation linking Palaeolithic art to hunting is inaccurate, as it is an interpretative framework that has actually been discarded (Rivero et al. in press).

Conclusion

In the paper discussed here, Palaeolithic people are assumed to have used a notational numerical system in rock art and portable art in order to establish an ethological calendar aimed at memorizing the key periods for fauna in their environment (mating, birth, migration). The authors claim that this kind of external memory could constitute the emergence of a “proto-writing.” However, all the reasoning is based on the accumulation of inaccurate images taken from old publications without verification. Thus, the lack of pertinence of the examples selected for the demonstration and some methodological weaknesses largely invalidate the authors’ interpretation.

Palaeolithic art is a particularly difficult field of study, requiring specific training in observation and first-hand knowledge of the art. It is also an art that is very distant from our culture, which is why we must be very cautious when presenting hypotheses in accordance with today’s mentality. Moreover, it is an extraordinarily widespread artistic production in time and space, of enormous complexity. It involved actors and generations over thousands of years and undoubtedly responded to multiple motivations far beyond those linked to the mere observation and subsistence of hunter-gatherer groups.