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 Interpretations of Art

At each level of interpretation of art (technical, typological, chronological, ecological, ethnographic, esthetic, meaning) there arise certain methodological problems. Some of these have been touched or skirted in the discussion already presented in this paper. Others, as indicated in the introduction, will not be handled at all. Nevertheless, there is a group of problems very intimately related to any anthropological interpretations of prehistoric art whose methodological treatment must at least be mentioned.


In attempting to understand any prehistoric art which has a considerable time range and is not static, we are faced with the same necessity as in ordinary archaeology: very little can be accomplished beyond a superficial level without an accurate and detailed chronology. It goes without saying that some of the attempts at correlating the art of different sites and different regions of northern Africa have been made by using methods which are somewhat intuitive and erratic. This is not to say that most of the chronological reconstructions have been of this kind or that the majority of writers have been uncritical of their own assumptions and theories. Nevertheless, there is always a temptation in evaluating this kind of data to choose individual elements or themes and to trace their distribution through time and space in order to demonstrate that the flow was in a certain direction. A recent example is provided by Bosh-Gimpera when, to demonstrate the Palaeolithic age of certain African art, he used stylistic parallels or analogies which occur in Spain and Tanganyika to document his argument for relationships 82. But Lhote has given a very useful illustration of how deceptive simple parallels of form in art can be when divorced from historical context by comparing skirts worn by women in Saharan art with those worn in the art of the Spanish Levant (at Cogul) 83: in spite of the stylistic similarities they are not contemporary since the former belong to the Horse Period about 1200 B.C. while the latter are very much earlier.
The index-fossil concept is a particularly popular one in analyzing prehistoric art on a chronological basis and is often used even when not explicitly recognized as such. An example in northern Africa is the use of the extinct buffalo, Bubalus antiquus, which Lhote explicitly describes as the fossiledirecteur for the earlier engravings, at least in Southern Oran, Tassili, and Fezzan 84. The method, taken alone, leaves a good deal to be desired. If an animal or some other single element is to be used as a time marker, then just as in geology or archaeology certain conditions must be established including a good knowledge of its typological variations, its geographical distribution, and its temporal range. If these can be established by purely archaeological means, so much the better; but comparative studies can also be effective in confirming or rejecting the usefulness of such markers as Lhote himself has attempted in rejecting the presence of discs or other head-elements on animals as indicators of domestication. It may be that Bubalus should be reexamined now that archaeological excavations have revealed that this animal lasted into the fourth millennium B.C. at Hassi Meniet in the Ahaggar massif, where its bones are found in a Neolithic site; further evidence from the southern Sahara suggests that it is found in engravings which can be no older than the second millennium B.C., and there is even some possibility, judging from a drawing on a Tunisian mausoleum, of its survival into the Roman period 85. A similar problem, this time using a stylistic device rather than extinct fauna, was created in the Upper Palaeolithic art of western Europe by Breuil's use of perspective tordue — the horns of bovids facing the viewer rather than in profile — as a time marker indicating an earlier, less accomplished stage of naturalistic art. This concept bedeviled the chronology of Palaeolithic art studies for decades, until the realization that this artistic convention existed in the later stages of Franco-Cantabrian art removed a source of error. In ordinary archaeological comparisons of excavated materials most archaeologists have long recognized the weaknesses of the indexfossil method, that is the identification of archaeological cultures solely by a few specific types of artifacts and of drawing wide conclusions concerning origins, diffusion, and other problems by spotting individual artifacts or attributes through space. This is not to deny that individual artifacts or elements can diffuse, or that it is unnecessary to trace the similarities in widely dispersed elements. But, just as archaeologists today have come to think of assemblages of artifacts as functioning wholes designed to accomplish certain purposes and capable of traversing space or time while sustaining greater or lesser changes, so we ought whenever possible to consider art groupings as wholes which are to be considered as particular kinds of assemblages. It is true, of course, that the problem of deciding just what the group is in prehistoric art, or even whether it exists, is a particularly troubling one, particularly in the European Palaeolithic; but this problem is often less difficult in northern Africa. I shall come back to this topic later.
Fortunately there is today considerable awareness among the foremost investigators of the art of northern Africa of the necessity of establishing both period styles (similarities among groups of elements from the same time period) and local styles (similarities or resemblances among elements from adjacent regions). Huard and Massip have suggested that in the case of the “Hunter Period” art, regional classifications are the necessary conditions for new progress in understanding 86. We are, after all, dealing with a very large region, and even in the central part there is not merely one Sahara but many Saharas, as Monod has emphasized 87. It seems a principle as relevant to prehistoric art as to prehistoric archaeology that the spatial dimension is one of the essential variables to consider in evaluating the variations in artifact form. With this in mind we can appreciate that differences between two regions need not be solely a factor of time but that motifs, styles, and techniques might be retained in some regions after they had been dropped or modified in others. There seems to be an instance of this in Tanganyika, where the naturalistic tradition of art lasted longer than it did in Zambia and Mozambique, where schematic and geometric art took over 88; and the same thing may well have occurred at times in North Africa.


From the nineteenth century on prehistorians and others have made use of information in prehistoric art to draw conclusions concerning past climate and the fauna and flora of the area. The faunal remains excavated from Palaeolithic sites had demonstrated this in the Pleistocene range, but the “Neolithic” art, especially the great herds of domesticated cattle painted in regions of the Sahara where no cattle can survive today, was a most dramatic proof of the very different environment and very drastic changes even within Holocene times. As the archaeology became better known and the art styles arranged in something approximating their proper sequences, attempts were made to use the art as a more precise index of ecology and environmental changes and even to reconstruct specific vegetation patterns, precipitation isohyets, and wild faunal ranges through time as evidence for schemes of Holocene climatic shifts involving sub-pluvial and arid phases 89. This assumes, of course, that the art really does reflect the presence in the past of those particular animals in the same zones where the engravings and paintings are found today. There have been many claims in the past that this was not the case but that the pictures, especially those of Ethiopian-type fauna, were done from memory of those seen in far-off regions to the south or in circuses in Mediterranean cities, or were copied as animals were brought across the Sahara en route to the Roman games. Today few such claims are seriously advanced, and most prehistorians are prepared to share the opinion voiced by Lhote 90 and others that the art does reflect fairly faithfully the evolution of the past fauna through the vicissitudes of Saharan climatic changes from prehistoric to modern times.
Nevertheless, there are too many cases known in ethnographic or prehistoric art in other parts of the world of fauna from distant regions being shown in the local art for us to ignore this possibility when we are dealing with small numbers of pictures. The cases of fish depicted on the local Mimbres pottery in prehistoric New Mexico, although no fish occur naturally in the region, and of wall paintings of marine fish in the interior of southern Africa come to mind. Presumably these may have been actual imports through trade, and one would not expect pachyderms to be exchanged in this way. Nevertheless the memory factor must be kept in mind, and this is particularly true when no osteological remains of a particular animal are found in palaeontological or archaeological sites in the region. For instance, there seems good stylistic and palaeontological reason to believe that the only rhinoceros so far found depicted in the rock art of the Nile Valley of Upper Egypt is such a caricature, done from memory 91.

Two other sources of error are possible. Schulz, a zoologist, has pointed out the dangers in too easy conclusions about climate and environment based on prehistoric fauna shown in prehistoric art since most large mammals can live under very varied climatic conditions and, provided they can obtain suitable food at all seasons, can often adapt to widely different sustenance 92. Finally we must recall, especially when dealing with animals believed to be domesticated, that the artists may have been careless or inconsistent, and precise details of iconography on which the specialist depends may not be altogether reliable. W.S. Smith points this out for Dynastic Egyptian art, where too exacting conclusions should not be drawn from the fauna shown in the wall paintings and reliefs, because the artists did not have a scientific interest in the modern sense 93. This should be kept in mind for prehistoric art as well.

Meaning and Purpose

 The problem of meaning in the interpretation of prehistoric art is probably the most formidable, for, in discussing the function or purpose, we are in many cases touching on the ideology of the people responsible for the art. An idea of the discrepancy possible in interpreting even the art of a modern group of primitives who had been intensively investigated by ethnographers is given by the dispute between Ascher 94 and Seligman 95 over the alleged functional relationships between subsistence patterns and motifs in Vedda paintings from Ceylon. On the prehistoric levels we can hardly expect less uncertainty or ambiguity. Some attempts have been made, nevertheless, to reconstruct something of the purpose and ideology of Saharan and other North African art. Mori suggests that the Round-Head style reveals a complex world of rites and beliefs that are more religious than magical and that revolve around a semi-divine or divine anthropomorphic being 96. Lhote considers differences in religious beliefs in the earliest (Bubalus) engravings of Southern Oran and the Tassili because of the differences in frequency of presentation of certain symbols, animals, or associations: in the Round-Head art he sees an “essentially African and Negro” style marked by distinctive symbolism (rounded heads) and animism (horned and masked figures); while the Bovidian Pastoral art, he thinks, was no longer exclusively inspired by magical and religious feeling but was also concerned with “art for art's sake,” for pleasure, and for purposes of narration and description 97.

It is extremely difficult to criticize statements of this kind, for they are presented with a good deal of tentativeness by students well aware of the risks involved and with a very intimate first-hand knowledge of, and empathy for, the art in question. The possible sources of error are so manifest that one is underlining the obvious in pointing them out. Argument based on the rarity or absence of a motif or theme, for instance, must be tempered by the reminder that it is possible for major cultural interests to be expressed only rarely in the art of a group — the absence of the human figure in official Islamic art comes to mind immediately, and we also have Suggs's documentation of the fact that sexuality, though an explicit and integral part of aboriginal Marquesan religion and a major cultural interest, was not often expressed in their graphic art 98.

Perhaps the greatest difficulty in getting at the meaning of prehistoric art is that we do not know the symbolic conceptions which were involved even in naturalistic representations. Are these to be taken literally, that is as signs? Or are they loaded symbols, part of a code to be broken? According to the Soviet anthropologist Okladnikov modern Siberian primitives consider the moose a symbol for a female deity of fertility and abundance 99; without knowing this “code” we would have no obvious clue to its meaning to the artists, and would be forced in most cases to proceed on the superficial level of analysis. The matter of “breaking” such “codes” where and if they are suspected to exist is a very troublesome one today especially to students of European Upper Palaeolithic art, and here I can only make a brief mention of one such attempt — which has by no means escaped sharp criticism — by Leroi-Gourhan in France. This writer has proposed a solution by attempting to establish the combinations or associations of one motif with others, and of their positions inside the caves, and by postulating a binary division based on sex whereby two principles, maleness and femaleness, account for certain species of animals and certain kinds of signs (e.g., horses and arrows are male, bovids and triangles are female) 100. Whatever the merits of this scheme, there has been to my knowledge no effort as yet to interpret the art of northern Africa in this way, but something along these lines may well be attempted in the future.

Stylistic analysis

One fairly common assumption in studies of North African art is that a style can emerge as the result of fusion between two or more other styles or cultures. Thus, it has been suggested by Rhotert 101, followed by Butzer 102, and Huard 103, that the engravings of the Bovidian Pastoral period wall art in the Sahara were born as a consequence of contacts between the indigenous hunters-engravers and immigrant “Hamitic” pastoralists from the east. Without attempting to deny that in theory this kind of intercourse might produce new styles, it remains true nevertheless that a safe distance between speculation and statements of what happened in history is not always preserved in writings of this kind. Although there may have been some overlap in time between the early engravers and the Pastoral ones in some places in the Sahara — indeed it would be curious if there were not — Lhote seems justified in denying the existence of any good evidence to support this hypothesis of contacthybridization in spite of certain resemblances between the two groups of engravings 104. In any case it is the Round-Head style, not the Bovidian Pastoral, that follows immediately the Bubalus art in the central Sahara. Nor are there any reliable data at this time that pastoralism was introduced from the east. The same criticisms can be made of suggestions that the paintings of the Pastoral period derive their inspiration from the Spanish Levant; there are certain analogies, certainly, but they are not convincing enough at the moment to support a belief in contacts between the two regions.

Another fairly common assumption in judging the dating and subsistence background of prehistoric art is that the cultural status of the groups concerned can be deduced from the technique and style of the art itself. Thus McBurney 105 proposes that certain naturalistic figurations can be ascribed to hunters (and so dated early), because hunters' art is naturalistic and involves a keen observation of the details of the fauna on which they depend, in contrast to the art of peasant communities 106. On the other hand Vaufrey seems in part to have based his belief in a Neolithic age for Southern Oran art on the alleged absence of a style naturaliste which would be expected in a hunting people 107. I suspect that, though in general this may often be true — and Australian art shows many exceptions — there are a good many other variables to be taken into account before we can use technique or style alone to denote cultural status. Breuil has made an interesting attempt to link naturalistic art with subsistence patterns and environment to show that hunters of large animals tend to emphasize naturalistic and large figures, but adds that in forested regions, where easy observation during hunting is not possible, this will not hold true, and great naturalistic art will be limited to open or steppe-type country 108. This is a reminder of the necessity of taking the former geographical and environmental factors into account in drawing conclusions; undoubtedly there are many more factors as well.

Several other essential points might be made in discussing the analysis of styles. One is that, whenever possible, like should be compared with like, that is engravings with engravings and paintings with paintings. It appears that the two techniques may not necessarily deal with identical themes and subjects even during the same cultural period in North Africa, e.g., in the Bovidian Pastoral phase. This is simply a confirmation of the principle recognized in art history that there may be cultures with two or more collective styles of art at the same moment 109. There is no need here to go into the possible reasons for this — male vs. female art, religious vs. profane art, and so on — but a recognition of this fact may help avoid a number of errors in interpretation. The second point is that the interpretation of North African art cannot be done en bloc by treating it as a whole through time. Each period style, like each local style, must be interpreted as far as possible in terms of its own economic and subsistence background. Unlike, say, the Upper Palaeolithic art of western Europe which represents a very long and essentially homogeneous tradition based on hunting and gathering alone, the art of northern Africa, regardless of the amount of cultural, physical, and stylistic continuity carried from one phase to another, is not based on a single exploitative tradition but on several. Throughout the long period represented by the strictly prehistoric art (perhaps more than six millennia) several quite different forms of social and economic life prevailed, ranging from large-game hunters at the beginning through various grades of good-producing based on animal, and perhaps plant, domestication, culminating in historic times in most areas in camel nomadism. Attempts at interpretation of the art have to be tackled in the light of these different cultural statuses, and the pertinent analogies must be drawn from the art and behavior of peoples pursuing similar kinds of subsistence patterns today. The relationship between the art and the economy is not necessarily a direct one: “between the economic relationships and the styles of art intervenes the process of ideological construction, a complex imaginative transposition of class roles and needs, which affects the special field — religion, mythology, or civil life — that provides the chief themes of art” 110; but it must always be borne in mind.

82. Bosch-Gimpera, “Chronology of the Rock Paintings,” 128.
83. Lhote, “Rapports,” 219-220.
84. Lhote, “L'évolution de la faune,” 93.
85. Huard and Massip, “Gravures rupestres,” 192-197.
86. Huard and Massip, “Gravures rupestres,” 192-197.
87. Monod, “The Late Tertiary,” 117-229.
88. J. Desmond Clark, “The Rock Paintings of Northern Rhodesia and Nyasaland,” Prehistoric Rock Art of the Federation of Rhodesia and Nyasaland, R. Summers, ed. (Salisbury, 1959), 216.
89. K. W. Butzer, Studien zum vor- und fruhgeschichtlichen Landschaft-wandel der Sahara und Levante seit dem klassischen Altertum. II. Das okologische Problem der Neolithischen Felsbilder der ostlichen Sahara, 1 (Mainz, 1958), 20-49.
90. Lhote, “L'évolution de la faune,” 83-118.
91. W. Gowers, “The Classical Rhinoceros,” Antiquity, 24 (1950), 61-71.
92. H. Schulz, “On the Zoomorphic Representations,” Prehistoric Art of the Western Mediterranean and the Sahara, L. Pericot Garcia and E. Ripoll Perelló, eds. (New York, 1964), 253.
93. W. S. Smith, The Art and Architecture of Ancient Egypt (Penguin, 1961), 138.
94. R. Ascher, “Function and Prehistoric Art,” Man, 61 (1961), 84.
95. B. Z. Seligman, “Function and Prehistoric Art,” Man, 61 (1961), 136.
96. Mori, “Contributions,” 173.
97. Lhote, “Rapports,” 215-224; Lhote, “L'évolution de la faune,” 83-118.
98. R. C. Suggs, Marquesan Sexual Behavior (New York, 1966).
99. A. P. Okladnikov, “The Temperate Zone of Continental Asia,” Courses toward Urban Life, R. J. Braidwood and G. R. Willey, eds. (New York, 1962), p. 277.
100. André Leroi-Gourhan, Préhistoire de l'art occidental (Paris, 1965).
101. H. Rhotert, Libysche Felsbilder (Darmstadt, 1952).
102. Butzer, Studien, 20-49.
103. Huard, “Art rupestre,” 122-147.
104. Lhote, “L'évolution de la faune,” 102.
105. C. B. M. McBurney, The Stone Age of Northern Africa (Penguin, 1960), 267 -268.
106. A rather similar judgment is expressed in McBurney's appraisal of rock drawings from Egypt, which are “suggestive of an impoverished marginal tradition far removed from the main centre of development.” Ibid., 272. Here the spatial dimension alone is relied upon for interpretation and the possibility of their representing an early, archaic stage is dismissed.
107. Vaufrey, L'art rupestre.
108. Henri Breuil, “L'Occident, patrie du grand art rupestre,” Mélanges Pittard (Brive, Correze, 1957), 101-113.
109. M. Schapiro, “Style,” Anthropology Today, A. L. Kroeber, ed. (Chicago, 1953), 294.
110. Ibid., 311.