You are here

Philip E. L. Smith
Problems and possibilities of the prehistoric Rock Art of Northern Africa

African Historical Studies, I, 1 (1968): 1-39

The Age and Origins of the Rock Art

These two subjects cannot easily be considered separately, for the question of the antiquity of the art as a whole and of that of each period has a very direct bearing on the interpretations which are likely or possible concerning origins.

In the question of the antiquity and chronology of North African art we have not yet reached the extreme point attained in the controversy over the age of the art of the Spanish Levant, which in Pericot's recent words “has assumed the proportion of an international polemic.” 9 But nearly identical difficulties are met in establishing absolute, and to a lesser degree relative, chronologies in the two regions, even though the disputes concerning the age of the older forms of art in North Africa are somewhat more muted. If such uncertainty can exist in an area such as eastern Spain, which has been methodically explored and studied for over half a century and where the purely archaeological sequence is fairly well known, we need hardly wonder that in North Africa we are still very far from certainty, particularly in the Sahara, where intensive study of the art is a recent feature and where archaeological investigations oriented to rock art problems have only recently advanced beyond the surface collecting stage.

The Age and Origins of the Rock Art

This is not the place to present a history of the various points of view expressed in the past century about the age of the earlier forms of rock art in North Africa 10. After the initial interpretations of the art as Phoenecian or as the work of “idolaters” of recent times from the south, a feeling slowly developed that they were in part prehistoric. This was apparently first suggested by Bonnet in 1889, while in the 1890's Pomel, because of the fauna represented, suggested that some of the art was Palaeolithic 11. At any rate, from the end of the last century, but especially from the 1920's on, there was a continuing controversy over the age of the oldest group of rock drawings, a controversy which is still not settled. In brief there are three main viewpoints. One group favors a beginning for the art rupestre in the Palaeolithic and includes, or has included, Dalloni, Frobenius, Pomel, Bosch-Gimpera, Solignac, Kuhn, and A. J. Arkell. Obermaier originally shared this opinion but in 1931 abandoned it in favor of a post-Pleistocene age, a position which was also adopted by Vaufrey, Graziosi, Huzayyin, Almagro, Lhote, Reygasse, Balout, Monod, and a number of other writers 12. Finally, there is a third group of prehistorians who are either undecided or who decline to commit themselves in the present state of knowledge. These include Alimen, Breuil, McBurney, and Mori.

Until the present time there is only a limited number of ways of assigning absolute or relative ages to prehistoric rock art. In rare but fortunate cases scenes depicted on walls may be covered by later archaeological deposits, fragments fallen from roofs or walls may be found in the datable archaeological levels, or, in the case of caves, the entrance may be blocked by demonstrably later deposits. All these situations provide ante quem datings for the wall art. On the other hand post quem dating may be obtained if a fallen slab with designs is found buried over an archaeological level which can itself be dated by radio-carbon or other methods. If the fallen piece is sandwiched between two datable levels, then the age may be narrowed down even further. Occasionally a more precise dating can be inferred if some object clearly related in style to the wall art (e.g., a sketch or rough draft of a feature shown on the wall) is found in good association with a datable archaeological level. Unfortunately, such associations of wall art with buried or stratified archaeological deposits are very rare in northern Africa, and in most cases other means have had to be used to provide datings. Superpositions of engravings or paintings give an idea of the relative ages and sequences of different styles or motifs. Degrees of patination can sometimes yield similar information. Objects whose age is known from other sources (e.g., weapons, certain animals, clothing, and ornaments) may give useful clues when depicted in the art. Artifacts whose age is known within broad limits may be found as occupation debris in the neighborhood of the sites, and, although the possible errors in this reasoning are obvious, it can be useful if the associations are consistent in a large number of cases 13. The geographical distribution of archaeological cultures may be compared with the distribution of rock art, and relationships of one with the other may be inferred as has been done in southern Africa, where the distribution of engravings on boulders seems to coincide closely with the extent of the Smithfield A industries in the Later Stone Age. Technical criteria may also be used, for example, the use of a pecking, grooving, or polishing technique which is thought to be time-restricted. Finally, the dating may be inferred on purely stylistic grounds, based either on parallels with styles whose age is known in other egions or on some stylistic sequence or scheme which assumes a development from, say, naturalism to stylization or schematism. Needless to say, the latter method is sometimes based on a priori reasoning which does not resist close examination and can often lead to circular arguments.

The case for the post-Pleistocene age of the earliest North African rock art was presented by Obermaier in a classic paper 14. His argument was that, since (a) no extinct Pleistocene fauna is shown, (b) domesticated animals are represented, (c) most or all of the animals were known in the area in classical times, and (d) the wild animals shown are executed in the same style as the domesticated ones, a Pleistocene age cannot be supported, and he suggested a “Neolithic” age instead for the earliest forms. Writers such as Reygasse proposed the same arguments, and in 1939 Vaufrey, as a result of a long and methodical examination of the art and prehistoric archaeology in the Southern Oran region of Algeria, proposed in his comprehensive memoir that in this region and probably in the whole Sahara “aucune gravure naturaliste de cette grande region n'est plus ancienne que le Neolithique de tradition capsienne,” i.e., no older than the fifth millennium B.C. 15. However, other writers have suggested a somewhat earlier beginning in the Upper of even Typical Capsian 16 or perhaps between an evolved Capsian and the Neolithic (Flamand, Lhote), that is, in Holocene times since about 9000 B.C. but in a pre-Neolithic context or at the most in a very early non-ceramic Neolithic. In particular Lhote has criticized Vaufrey's use of such elements as horses and rams to prove a recent dating. He has argued 17 that true wild horses (Equus caballus) were present in North Africa in the Neolithic long before their introduction into Egypt by the Hyksos about 1600 B.C., and, therefore, their presence in at least six drawings in the Maghreb contemporary with Bubalus engravings cannot be used as evidence of a very recent age for these engravings 18. Another element in the art of the Maghreb to which Lhote has apparently succeeded in assigning an earlier age than that allowed by earlier writers 19 is the engraved ram with the disc or spheroid which has usually been regarded as domesticated and a reflection of influences from the Ammon cult of the Egyptian New Empire. Indeed Lhote 20 now claims that these engravings are the same age as the Bubalus group and that the spheroid is no evidence of domestication (though he has suggested taming of individual animals for ritual purposes in a religious cult). It must be admitted that, since similar discs are now known between the horns of the extinct Bubalus in Southern Oran 21, and since Huard 22 shows that collars, head ornaments, and neck ornaments are also found on such large and unquestionably wild animals as giraffes, rhinoceros, elephants, and hippopotamus in at least thirty instances in the Sahara, we can no longer rely on these criteria alone to establish the status of such potential domesticates as cattle, goats, or sheep. Whether these ornaments and paraphernalia reflect a special interest in wild animals which might be interpreted as a stage of manipulation or incipient domestication is something we cannot answer at present, though I suspect this argument would be advanced if similar evidence were present in the early Holocene or final Pleistocene horizons of southwestern Asia.

A single example of stratigraphic archaeological evidence is available to support an age at least as old as the Neolithic of Capsian tradition for some of the engravings in the Maghreb. At the cave of El Arouia in Algeria a deeply engraved equid (a true horse, according to Lhote), said to be done in the same style as the earliest or Bubalus engravings, was found by Vaufrey covered by deposits of Neolithic of Capsian tradition 23. A group of traits capsiens (simple grooves) and pit-marks carved at the entrance of a nearby cave might give a further clue. The presence in 32 out of 36 cases of microlithic industries (but no pottery) said to belong to the Neolithic of Capsian tradition in the immediate environs of the engravings studied by Vaufrey is generally taken to suggest a probability that they date the art. But it is no more than a probability, as Lhote emphasizes in pointing out that in other cases in North Africa Aterian and even Acheulian-type artifacts are found near the engravings although this does not indicate a Palaeolithic age for the art 24.

So far there is no direct archaeological evidence to provide more precise dates for these earliest engravings, either in the Maghreb or in the Sahara. Indirect evidence from the Acacus region of the Fezzan in Libya suggests that they may have been under way well before 6000 B.C. since they seem to precede the Round-Head style paintings, which in turn precede the first Bovidian Pastoral style paintings believed to be dated by radiocarbon at ca. 5500 B.C. at one of Mori's Acacus sites. The earliest engraved rock art thus still hangs in mid-air. Until some lucky find allows us to pin down the age and duration of this group more securely, we shall have to be content with cautious speculation concerning its beginnings and origins. Several writers have commented on the fact that it seems to appear fully formed and without obvious local antecedents. Monod remarks:

It is astonishing that the oldest rock art chronologically is also the most “beautiful.” Such consistent workmanship obviously implies attempts, gropings, an apprenticeship. Why have these left no traces on the rocks? Were there not other media even more favorable to “scrawls” (ostrich eggs, wood, plates of schist, skin, bone, etc.)? 25

This problem might be answered in several ways. In the first place we are perhaps not justified theoretically in assuming that, if a purely local origin is demanded, there need have been such an “apprenticeship” stage; in archaeology it is notoriously difficult to establish the beginning stages of traditions or industries, and Spaulding 26 has suggested some very cogent reasons based on the principle of quantum advance why there is usually a clustering in time of events at the beginning of a period rather than a spreading out to reveal archaeologically discernible developmental stages. Second, there is evidence of quite early art in North Africa, in the Epipalaeolithic or even in the Palaeolithic. It is rare, and it cannot yet be linked directly with the rock art we are discussing here, but nevertheless it must be kept in mind. In the last few years Roche has described his discoveries at the stratified occupation site of Taforalt in Morocco of a small quartzite nodule worked to resemble simultaneously male and female sexual organs in a level of the Iberomaurusian (= Oranian) culture dated by radiocarbon to 10, 120 B.C., i.e., presumably final Pleistocene times. In succeeding levels at the same site dated to about 8800 B.C. were found a pebble with a rough engraving of an elephant, an ostrich egg shell disc with rough incised lines, and a grinding stone with engravings which are difficult to interpret but which may represent mouflon horns or an anthropomorphic figure 27. In an Upper Capsian site in Algeria, Khanguet el-Mouhaad, there is a plaque with an engraving of a horn recalling those of Bubalus in Maghrebian wall art 28. At El Mekta site in Tunisia, in the Typical Capsian, sculptures as well as engravings on portable objects and on walls are reported 29; the absolute age is uncertain but is probably at least as early as the sixth millennium B.C. Finally, we might mention the geometric engravings at Abka in the Nile Valley of the northern Sudan, some of which are claimed to date between 7500 and 7000 B.C. 30. The evidence is meager, but at least it indicates that art as a phenomenon was already present in this part of the African continent by the beginning of Holocene times or even in the final stage of the Pleistocene. There are, in other words, seeds from which the later art might have grown.

The problem of diffusion of art from southwestern Europe during the Upper Palaeolithic to North Africa is a time-honored one. But a European origin for the earliest North African art does not depend on a Pleistocene dating for the latter, for the diffusion may have taken place during the European Mesolithic. Boule, Frobenius, Breuil, and a number of others have in the past expressed belief in a basic unity of European Palaeolithic art and the engravings of North Africa, and this viewpoint is still apparently shared by Bosch-Gimpera 31,  who advocates transmission from north to south. In more recent years this viewpoint has had fewer defenders, and although several authors, such as Pericot and Bandi, still believe in at least some stylistic relations between North Africa and the Spanish Levant art in Holocene times, this opinion does not seem to be generally shared today 32. Even Breuil had apparently abandoned this viewpoint shortly before his death, although as recently as 1957 he expressed a belief in contacts between North Africa on the one hand and the Spanish Levant and Sicilian art on the other. Today opinion concerning these linkages appear to be diluted to the force of assumed “unities of mentality” between the two traditions 33 or to the existence of a general “Mediterranean province” of post-Palaeolithic art which shared a number of stylistic elements such as the shapes of animal horns or hooves, especially as shown at Addaura and Levanzo in Sicily, at Ebbou in southern France, and at the Fezzan in northern Africa 34. These shared stylistic elements described by Graziosi are difficult to interpret just now; they are certainly suggestive and merit further examination, but it is hard to say whether they are not too generalized to be significant in establishing an artistic tradition over so wide an area. In the last analysis, however, an answer to this question must await settlement of the old problem of whether there were direct trans- Mediterranean cultural contacts sufficiently early to have given rise to the similarities.

Another problem still to be satisfactorily resolved is the relation of the North African art to prehistoric African artingeneral. Seen in a broad context, the art of northern Africa presents many of the same problems and opportunities as does the art of the rest of the continent. Much of it is undoubtedly recent and offers at least the possibility of being linked with the direct ancestors of modern peoples. The emphasis at times on human figuration north and south of the Sahara is striking. And we are faced with very similar problems in establishing the ages, the sequences, and the relationships between different period and spatial styles.

To what extent can we regard prehistoric African art as a whole sui generis, at least in the earlier periods? This is a difficult problem which the author of this paper is not qualified to answer one way or the other. It can only be decided, perhaps, by investigators with a profound knowledge of the art of northern and sub-Saharan Africa and after the research methods of recent years have been carried considerably farther. Breuil, who was probably more familiar with the whole of ancient African art than anyone of his generation, claimed that Saharan rock art could not be separated from or treated apart from the rest of prehistoric African art 35. Like a number of others (Boule, Graziosi, Gautier, Joleaud, Kohl-Larsen) Breuil was inclined to favor connections between the prehistoric art of northern and southern Africa, though he admitted that the area he considered a “contact area” (Tanganyika) offered no links with such zones as Tassili and that it was possible that both groups of art had developed, with few contacts, from a very distant common base 36. Other authorities today are dubious about such linkages, and Lhote insists that art was developed independently in situ in each region and that each region developed its own character in art 37. In any case, it is unlikely that this problem can be answered with a simple yes or no. It must be investigated separately on each time horizon, with the environmental and cultural features encouraging or inhibiting diffusion in mind, just as is the case with more orthodox aspects of archaeology. Considering what is now known of the pastoral groups in the Sahara and immediately south in prehistoric times, for instance, it would be quite normal to expect at least some diffusion of Saharan art into the sub-Saharan zone during the postulated centrifugal movements as environmental conditions in the Sahara became less favorable for cattle herders about the third millennium B.C. Perhaps Monod's concept of the Sahara as a device for sorting and filtering elements can be usefully applied to this problem of art diffusion 38.

Certainly, before seriously tackling this problem, we must have more reliable chronological frameworks than now exist for the art sequences in both northern and sub-Saharan Africa. The dating of the earliest form of art in sub-Saharan Africa is notoriously difficult to establish, and estimates have ranged from the Magosian of the Second Intermediate, perhaps between 14, 000 and 8000 B.C. 39, to considerably later. This question is too detailed to go into here and it is enough to say that at the present moment there seems to be no art in sub-Saharan Africa known to be as old as the earliest dated examples from northern Africa. The radiocarbon date of 4360+ 250 B.C. for the painted engravings at Chifubwe Cave in Zambia seems to be no longer acceptable, and, even if the new dates of ca. 3400 B.C. and >5000 B.C. from the Matjes River shelter in South Africa 40 are accepted as reliable for several examples of art found there, they are younger than some dates now known from the Sahara and Mediterranean hinterland and possibly from Nubia as well. Certainly none of the naturalistic paintings which have survived in southern Africa can be considered as ancient as the oldest ones from the Sahara on the basis of existing knowledge. The fact that in both areas engravings seem to appear earlier than paintings in the rock art may not demonstrate a general chronological synchronism between them. At the moment, therefore, it does not seem likely that the roots of the earliest African art are to be found south of the Sahara. Perhaps in the realm of art sub-Saharan Africa occupied much the same kind of role it apparently did in the case of animal and plant domestication: it was an area of secondary rather than primary discovery and development.

When we speak of the diffusion of art into northern Africa from other continents, there is perhaps an unspoken assumption that art could not have developed indigenously in Africa but required outside stimulation, or even that all art springs from a single source in western Europe about 30,000 B.C. when it is observable archaeologically for the first time. I should say now, before proceeding farther, that I am by no means convinced that this is the case. Whatever may be the justification for employing concepts involving psychic or spiritual unity in evaluating prehistoric art, it seems fair to say that, from what we now know of the abilities of hominids after 30,000 B.C. or so, there are no good reasons to deny any of them the intellectual qualities necessary to produce expressions we can classify as art; particularly when the idea of art is one of those expressions of culture which are probably ecologically free, though the form of expression may at times be ecologically bound.

In the past much —probably too much — has been written about the possible European origins of African art, especially via Spain and Italy. I feel that, if diffusion is to be called in, southwestern Asia has been neglected in this respect in spite of the fact that the purely archaeological evidence for diffusion from that region to northern Africa during the late Pleistocene and Holocene is incomparably better than the evidence for movements between Europe and North Africa.

True, there is still no incontrovertible evidence for the existence of rock art in southwestern Asia during Palaeolithic times, though there have been some claims for a Palaeolithic age for some engravings in Turkey 41. But art is certainly present in the Lower Natufian of the Palestinian region in the form of sculptures and, just possible, rock engravings on outcrops around 9000 B.C., and in the Pre-Pottery Neolithic B of Jericho in the seventh millennium B.C. Furthermore, the recent surprising findings of very elaborate paintings on house walls in the early Neolithic levels of Chatal Huyuk, in the late seventh or sixth millennia, indicate how rapidly new discoveries can change our viewpoints. However, it must be admitted that we can still see no stylistic resemblances between the art of southwestern Asia and northern Africa on these early horizons. Such hypotheses as Rhotert's 42 that naturalistic art crossed from southern Arabia to Eritrea and then moved up and down the Nile Valley and into the Sahara are still unsubstantiated.

9. L. Pericot Garcia, “Sobre algunos problemas del arte rupestre del Levante espanol,” Prehistoric Art of the Western Mediterranean and the Sahara, L. Pericot Garcia and E. Ripoll Perelló, eds. (New York, 1964), 157.
10. A review is given in J. Forde-Johnson, Neolithic Cultures of North Africa (Liverpool, 1959).
11. The fact that similar claims were being made about the same time for the age of cave art in France and Spain is probably relevant to this subject, but the influence of European events in prehistoric research on North African archaeology is not a suitable topic for this paper.
12. This neat division is complicated by the fact that some writers believed the earliest drawings were Capsian at a time when this culture was believed to be late Pleistocene in age, while some other writers such as Flamand, though calling the art “Neolithic,” actually favored a considerable antiquity since they thought the North African Neolithic developed in the late Pleistocene.
13. This method has been used in eastern Spain to suggest post-Pleistocene ages for the controversial Levant rock-shelter art, and Vaufrey was able to demonstrate in Algeria that in 32 out of 36 cases archaeological remains which he classified as Neolithic of Capsian tradition were found in close proximity to the engravings of the Bubalus type. R. Vaufrey, L'art rupestre nord-africain (Paris, 1939).
14. H. Obermaier, “L'âge de l'art rupestre nord-africain,” L'Anthropologie, 41 (1931), 65-74.
15. Vaufrey, L'art rupestre.
16. Lionel Balout, “La préhistoire,” Revue Africaine, 100 (1956), 446-449, 74.
17. Henri Lhote, “Faits nouveaux concernant la chronologie relative et absolue des gravures et peintures parietales du Sud Oranais et du Sahara,” Prehistoric Art of the Western Mediterranean and the Sahara, L. Pericot Garcia and E. Ripoll Perelló, eds. (New York, 1964), 191-214.
18. The palaeontologist Arambourg agrees Equus is represented but suggests E. conga rather than E. caballus.
19. Allegedly “for psychological rather than archaeological reasons ”on the part of Vaufrey, Breuil, and Obermaier. Lhote, “Faits nouveaux,” 196.
20. Ibid., 196.
21. See Vaufrey, L'art rupestre, fig. 23.
22. Huard and Massip, “Gravures rupestres,” 192-197.
23. Vaufrey, L'art rupestre, pl. 11.
24. Lhote, “L'évolution de la faune,” 90.
25. Th. Monod, “The Late Tertiary and Pleistocene in the Sahara,” African Ecology and Human Evolution, F. C. Howell and F. Bourliere, eds. (London, 1963), 208.
26. A. C. Spaulding, “The Dimensions of Archaeology,” Essays in the Science of Culture in Honor of Leslie A. White, G. E. Dole and R. L. Carneiro, eds. (New York, 1960), 437-456.
27. J. Roche, L'Epipaleolithique marocain, 2 vols. (Lisbon, 1963); J. Roche, “Représentation humanie bisexuée trouvée à la Grotte de Taforalt (Maroc),” Miscelanea en Homenaje al abate Henri Breuil, 1877-1961, E. Ripoll Perelló, ed., 2 (Barcelona, 1965), 307-308.
28. Lionel Balout, Préhistoire de l'Afrique du Nord, essai de chronologie (Paris, 1955), 442, 478.
29. E. G. Gobert, “El Mekta, station princeps du Capsien,” Karthago, 3 (1952), 3-79.
30. O.H. Myers, “Abka Again,” Kush, 8 (1960), 174-181.
31. P. Bosch-Gimpera, “Le problème de la chronologie de l'art rupestre de l'Est de l'Espagne et de l'Afrique,” Actes du Congrès Panafricain de Préhistoire, II session, Alger 1952 (Paris, 1955), 695-699; P. Bosch-Gimpera, “The Chronology of the Rock Paintings of the Spanish Levant,” Prehistoric Art of the Western Mediterranean and the Sahara, L. Pericot Garcia and E. Ripoll Perelló, eds. (New York, 1964), 125-129.
32. E. Ripoll Perelló, “Proceedings of the Wartenstein Symposium on Rock Art of Western Mediterranean and Sahara,” Prehistoric Art of the Western Mediterranean and the Sahara, L. Pericot Garcia and E. Ripoll Perelló, eds. (New York, 1964), x.
33. H. Breuil, quoted by Ripoll Perelló, “Proceedings,” x.
34. P. Graziosi, Palaeolithic Art (London, 1960); P. Graziosi, “L'art paleolithique de la 'province mediterranéenne' et ses influences dans les temps postpaléolithiques,” Prehistoric Art of the Western Mediterranean and the Sahara, L. Pericot Garcia and E. Ripoll Perelló, eds. (New York, 1964), 35-46.
35. Ripoll Perelló, “Proceedings,” x.
36. Breuil, “Les roches peintes,” 150.
37. Henri Lhote, “Sur les rapports entre les centres d'art préhistorique d'Europe (province franco-canabrique et Levant espagnol et celui du Sahara),” Prehistoric Art of the Western Mediterranean and the Sahara, L. Pericot and E. Ripoll Perelló, eds. (New York, 1964), 215-224.
38. Monod, “The Late Tertiary,” 122.
39. J. Desmond Clark, The Prehistory of Southern Africa (Penguin, 1959), 265.
40. C. Gabel, “African Prehistory,” Biennial Review of Anthropology, B. J. Siegel, ed. (Stanford, 1965), 64.
41. E. Y. Bostanci, “Researches on the Mediterranean Coast of Anatolia. A New Palaeolithic Site at Beldibi near Antalya,” Anatolia, 4 (1959), 129-178.