|A very simple theriomorph -- a horned human figure -- from African rock art.|
In this post, I provide a translation for an article by A. Muzzolini on rock art of the central Sahara. To see the figures he refers to, please look up the French original, which appeared in the journal, Archeo-Nil, May 1991, on pages 17 through 42. Bold numbers in square brackets are the page numbers. Muzzolini's notes and references appear in the original article as well.
MASQUES ET THEROMORPHES DANS L’ART RUPESTRE DU SAHARA CENTRAL
(MASKS AND THERIOMORPHIC FIGURES IN CENTRAL SAHARAN ROCK ART)
by Alfred Muzzolini
Archéo-Nil May 1991, pp. 17-42 (plus figures 1-13 on subsequent unnumbered pages)
Abstract (in English as presented in the original article)
MASKS AND THEROMORPHIC FIGURES IN CENTRAL SAHARIAN ROCK ART. After recalling the rock art chronology, Saharian masks and theromorphic figures are described. They are found for the most in the “Naturalistic Bubaline” school (mainly at Oued Djerat in Tassili and Mathendous in Fezzan) and in the Tassilian Round Heads. Their positon with regard to peripheral areas is specified. Contacts with the pharaonic Egypt or the Predynastic are very unlikely. But a very old, common archaic substratum, peculiar to the Afro-asiatic linguistic group, is perhaps reflected by a common “africanness” in artistic themes. [as in original]
It is wholeheartedly that the author of these lines adheres to the project recorded in the Editorial of the inaugural issue of this magazine: "Provide a meeting ground for some (prehistorians) and others (Egyptologists)." For the vast plains and mountains stretching to the west of the last oases of the Western Desert, these desolate lands where the evil Seth reigned, were less empty of human occupation than the Egyptians believed. And during the long night that preceded the unification of the Kingdom of the Two Crowns, structured social groups, perhaps even peoples, in certain epochs traveled a desert much less arid than it is today. Hunters or shepherds, their fate could sometimes cross that of the groups that gradually took root in the alluvial plain of the Nile. Do traces of such contacts, or snatches of perhaps formerly common beliefs, still exist? Almost hidden in a hollow rock on the bank of a tributary of the Mathendous, south-west of Libya, a rock carving suddenly arises, a disquieting image of a man with the head of a jackal: it inevitably evokes the god Anubis. Could it be?
Our curiosity is legitimate. Masks and theriomorphs in Saharan rock art, which we propose to study here, are required immediately as a possible rapprochement. Masks and humans with animal heads  are known forms in many ideologies. Do not those of the Sahara and those of Egypt, however, have a mutual foundation – an original "Africanness"?
We must first define the rules of such a methodological study, a necessary precaution so as not to cause an "archeological fiction" to be written on the basis of some formal similarities that could only have arisen due to accidental convergences. After a brief chronology of Saharan rock groups - a dimension that our study cannot provide - we will describe the actual groups wherein masks and theriomorphs are found, and those where they are not found, an absence that may be significant. This will provide a cultural mosaic that we will try to interpret, more broadly, in its African context.
I. Method of study
The only masks in Saharan rock art that we apprehend as such are masks of animal heads. A difficulty arises when we undertake an inventory of the instances. How does one distinguish, in the rock art images, a mask – a material object borne, for symbolic purposes, by a real man to give the appearance, only the appearance, of a different head, for example of an animal - and a theriomorph with an animal head, a mythical being, half man, half animal? Theoretically, the mask has a lower limit, which should be marked by an almost horizontal line at the height of the bust, or encircling the neck, while the theriomorph shows no trace of such a demarcation. Some authors demand this criterion and decide between the two entities according to it. But this requirement appears very questionable.
On the one hand, masks can in fact reach the ankles: for example, even today, dancers, in some African tribes, cover themselves completely with an entire skin of an ox, and the appearance that results is that of a theriomorph with a horned head. On the other hand, a line under the neck can also sometimes mark a necklace, or the top of a garment, as much as the bottom of a mask (e.g., fig. 3 h, 4 e). More generally, one cannot expect of an engraver of rock representation a respect for detail, which only appears rational in our designs of the twentieth century. In many arts called "primitive," the important thing is not formal similarity, but the mere presence of the object represented: here, the head of an animal. And the engraver could disregard the detail of the dividing line at the base, like so many other details considered unnecessary or even inappropriate. This is true both in rough or schematic or slightly naturalistic styles, and in the so-called naturalistic styles, because even in these, it is a matter of naturalism in overall expression which is only required by rendering forms deemed essential for the function of the image. In other words, we cannot accept the overly simple criterion of the marking of the lower limit of the mask: whether it exists or not, there is usually indecision over the interpretation.
This consideration substantially affects the basis of our inventory. In any event, lines of demarcation at the base are very few - maybe half a dozen cases in all central Saharan rock art – and such indecision prohibits the possibility of a study distinguishing masks and theriomorphs. It seems more sensible to establish one set from these two categories. Not only because a study, thus globalized, then becomes possible, but because the two categories certainly belonged, in the collective imagination, to a single conceptual archetype: one that admitted a certain  identity (of whatever nature, function, value, or place in a symbolic language) between man and animal. The mask is only an artifice creating the image of a referent conceived as real, the mythical theriomorphic being. It will be objected that we allow a speculative component to intervene already in the inventory records, an inventory that should confine itself to the "objectivity" of the collection. We recognize that. But our hypothesis which is to replace the two categories, "mask" and "theriomorphs," with a single category, "representations using the hybrid image of a human-animal for symbolic purposes," seems to us acceptable at least as a working hypothesis.
It should be noted already that, in Saharan rock art, only the heads of these hybrid representations are of animals; the body is still human. The opposite - the sphinx or centaur type for example - is not found.
Whether it is a matter of true masks or of theriomorphs, we encounter an additional difficulty whether in groups of rough style, or in naturalistic groups where a taboo prevails, or conventions, for the representation of the face. Sometimes we cannot recognize the animal species represented, nor probably even distinguish between the face of an animal and the face of a human summarily carried out or deformed. Some groups frequently practice caricature drawing, including that of giving faces "a muzzle." We will discuss this point in different contexts. Already we note that many hasty sketches found in the literature and labels such as "masks" seem suspect to us, being equally interpretable as a caricature of a human or as awkwardness of execution. It will be better to preclude from the debate those representations in which symbolic intent is not obvious.
We will divide representations of masks and theriomorphs of the central Sahara into the various artistic, historical, and possibly ethnic units. For each ethnic group has its own collective imagination. Some authors use a global structuralist approach for the whole of Saharan rock art: this postulates a homogeneous Saharan block, a coherent ideological system, that is never justified a priori, and no one sees why, in this case, they do not extend this block, for example, to Egyptian art or Levantine art, which is roughly contemporary. Mixing it all together, even limiting it to central Saharan rock art, considering the ideologies of the various specific groups recognized by classical archeology or rock art studies, yields a single body of myths and beliefs, invariable through the millennia. This oversimplification has exposed monumental mistakes. If we were to admit this hypothesis in the study of the symbolic European universe, for example, we would strongly risk - in the absence of written texts - putting on the same footing the Cretan goddesses, Athena or Venus, and the medieval worship of the Virgin. We would seek common elements of their legends – there are some - ending up with erroneous results, with false convergences, for example, ignoring the crucial fact of the emergence of Christianity. We need to avoid these mistakes, treating separately as much as possible the various discrete blocks of collective imagination. And that, before any airing, projecting our representations onto the appropriate chronological units (fig. 1).
II. Chronology of the rock art representations
Various teams are apparently on the verge of success in possibly directly dating rock engravings (Nobbs-Dorn, 1988) and some Australian paintings will soon be dated by C14 (Loy et al., 1990). But for now, no Saharan rock carving can yet be dated directly: the C14 dates  that are found in the Saharan literature are only dates of carbon that was found at the foot of decorated walls, dates extended – a bold speculation – to the representations on these walls. As a result, many archaeologists believe, even if they do not express it bluntly, that Saharan rock art researchers have no serious, "scientific" basis for sequencing their representations chronologically, and therefore their classifications are based on "intuition"…. This is just a preconceived notion and a misconception. It is correct that, in previous generations, many personal prejudices were integrated into the classificatory systems. But for the last 20 years, the time frames of Saharan prehistory have been largely confirmed. If some researchers have not adopted the necessary updates, others have not failed to take the new view of rock art ensembles suggested by the new data.
The researcher in Saharan rock art primarily provides a classification into artistic groups, distinguished by various criteria (techniques, patinas, figurative anthropological types, and especially styles). He then orders these groups in a relative chronology thanks to various chronological gradations (patinas, superpositioning, sequence of fauna, certain themes). He finally anchors this “floating” sequence thanks to absolute dates provided by other disciplines. Remnants of habitats dated by C14, zooarchaeology, and climatology, in particular, have provided precise dates unknown to previous generations, and that help locate populations and faunal contexts recorded in the representations. For example, to trace the beginning of rock art to the final Pleistocene, as asserted by Mori, now proves unthinkable, because we now know that the central Sahara was then extremely arid and empty of wildlife and people; similarly, one can imagine vigorous pastoral groups and a fauna of large, wild mammals during the “Arid post-Neolithic,” a climatic episode that can be quite precisely dated to 2500-1000 BC (+/- 500). For recent phases, milestones that are clear enough – approaching a few centuries – are provided by the dates of introduction, from peripheral countries that had already entered into history, of some domestic animals (the horse, the camel), of certain objects of material culture (carts, various weapons), of writing, etc.
Ultimately, we come to a chronological timeline that is fairly well assured and accurate but imprecise – the two concepts are not to be confused – in absolute chronology: this means that the sequence is certain but the dates of various stages cannot be fixed except within a notable range of uncertainty, on the order of +/- 500 years for the oldest periods of rock art. This is embarrassing, of course, especially for comparisons with well dated peripheral ensembles, for example that of Egypt. But it is a handicap that must be accepted for rock art. It is not fatal. Many “scientific” disciplines only deliver results stated either with margins of uncertainty or in terms of probability, and such data can be dealt with.
The present author, after reconsideration of the inventories and discussion (1986), which cannot be repeated here, has proposed the timeline summarized in fig. 2 for the major blocs of rock art in the central Sahara and the Saharan Atlas. For readers unfamiliar with these problems of chronology, we only emphasize that our synthesis deviates from the traditional chronologies on two essential points:
1. It objects to the notion of a “bubaline period” or a “phase of great wildlife” described by Lhote, Monod, Camps,  Mori, etc., as “prepastoral” or “prebovid.” For us, this purported “bubaline period” is a “style,” a “school,” of an age that was already “bovid” or “pastoral”: clearly domesticated cattle are related, for example, to the “decorated rams” of the Atlas.
2. It objects to the very high dates attributed by Lhote or Mori to the “bubaline period” (and thus to the “decorated rams” of the Atlas and to the large buffaloes that are incontestably related to them): 5000 BC as the “minimum age” for Lhote, the “upper Pleistocene” for Mori, dates that are incompatible with zooarchaeological data from excavations.
This correct scenario of time frames is necessary because the representations of masks and theriomorphs prove to be confined to certain groups. In fact, they appear with some regularity in two groups that are among the oldest: those of the engravings of the “Naturalistic Bubaline” – and again, only on a portion of this group’s area of extent – and those of the Tassili paintings of the Round Heads. They then become scarce and disappear in more recent periods.
III. Masks and theriomorphs in engravings of the “Naturalistic Bubaline” group
This well known group of engravings – the most naturalistic, the most “beautiful” according to our Western canons of the twentieth century – comprise the oldest engravings of the Saharan rock art sequence. This group has a wide distribution, from the Rio de Oro to the Fezzan. Masks or theriomorphs are neither rare nor frequent (1). But they are found almost exclusively in two regions only: the Oued Djerat (Tassili North) and in the rock art ensemble of the Messak Settafet (Fezzan). The latter is more commonly called the ensemble of the Mathendous.
At Oued Djerat, masks and theriomorphs mainly occur in the upper part of the valley, that is to say on a stretch of about thirty kilometers (mainly the Rock of Ahana). These compositions are often called “erotic” – “obscene” or “pornographic” would be more appropriate in terms of our modern codes of value – where the masks are worn by ithyphallic men. These, in situations without mystery, are often adorned with a huge caricatural phallus (fig. 3). This association of mask-phallus is almost constant only in the Djerat. We note that scenes of coupling and of phallic persons, but without masks, are not uncommon as well in the Djerat.
What is the animal represented? One perceives a subtle game of referents, as Michel Foucault would have seen in “Words and things.” For example, one type of fairly stereotyped mask – also found at Mathendous – seen from the front, vaguely triangular, quite symmetrical, is that of an animal that may resemble a cat (2) somewhat, with short ears above the head (e.g., Lhote, 1976, fig. 372, 682) (fig. 3 a, 3 b). But this “cat” is endowed with eyes, a mouth, and sometimes a nose, which confer on it an almost human expression. So one wants to give the phallic person a mask symbolic of an animal, but one also wants that animal mask to “signify” some concrete human iconically. To complicate matters, in one case, two non-realistic horns, not proper to the species represented by the mask, surmount heads: it is sometimes a matter of simplified horns in a circular arc – perhaps cattle horns, perhaps also simplified horns not interpretable as to species – but when the drawing is careful, they are twisted horns evoking an antelope – hartebeest or addax. On the other hand, the person very usually wears a false tail (a long, feline one, when it is well drawn, and not that of an antelope). In one third of cases, however, the mask  takes on an elongated look, with a muzzle: it is that of a canid – a motif, we shall see, that is classic to the Mathendous – and the thing is sometimes quite clear (e.g., fig. 3 d, 3 f, 3 h). But here again, the expression sometimes seems humanized. Thus, the hunter in fig. 3 h, who holds a bow, with hair reduced to some above the head (he recalls exactly the jackal-headed man of Ti-n-Lalan, Acacus, fig. 7 c). Other heads, quite human, simply have two horns (fig. 3 g) (Lhote, 1976, fig. 377) or indeed long ears (of a hare? of an ass?) (fig. 3 e( (ibid., fig. 571) sometimes analogous to the two lobes that we note on the heads of the Round Heads, and even known exceptionally on women (e.g., ibid., fig. 2082). This mixture of humanity and animality appears confusing to us. That a man should take on, just as the shamans of the San do, the appearance of an animal – as Lewis-Williams explains to us – we still understand, but that the face of this animal should be given a human expression (that of the shaman, or that of another man?) requires a still more complex interpretative thesis for which we have no key.
Many of the masks of the Djerat are, however, of very rough draftsmanship, unusually coarse in this group which knows how to use, on occasion, very naturalistic renderings. Thus, at Ouan-Rechla (Soleil-havoup, 1988, p. 66), the man with a giraffe’s head walking at the head of a row of giraffes and elephants, probably of the “Naturalistic Bubaline” (3). Often, a reading such as “bestialized human face,” without precision concerning the suggested “beast,” would also be sustainable (e.g., Allard-Huard-Huard, 1981, p. 16, 46 – Lhote, 1976, fig. 372, 1695) (fig. 3 e). We note, finally, an exceptional representation (fig. 3 c): on a rhinoceros of good “bubaline” workmanship, a man with a huge phallus and using a dummy tail, has a badly defined head but from which a rhinoceros horn seems to have come loose. A squiggly line goes from the phallus to the eye of the rhinoceros. It probably depicts a jet of sperm, following a convention known elsewhere in Saharan rock art, for which various glosses have been proposed: transmission of life, or of vital energy, etc. – but none can be clearly assured.
Some masks, of various shapes, are found exceptionally, in the Djerat, on women (Allard-Huard-Huard, 1986, p. II). But the style of these representations does not allow attribution to the “Naturalistic Bubaline”; it is perhaps a matter of works of a little later date.
In the Mathendous, masks or theriomorphs are a little more common than at Djerat (cf. n 1). A conventional type (e.g., fig. 4 c, 4 g, 5) is that of a man with a head generally described as that of a jackal. In fact, dog or jackal can hardly be distinguished (both species are documented in lists of fauna of the Sahara). The representation of this animal head indeed observes conventional canons (while the naturalism of the animals in the Mathendous is often perfect). The muzzle is often square, not very realistic. Long ears or ears with large lobes as among the Round Heads emphasize animality. Yet the general expression of this animal face, rendered by the eyes, the mouth in a faint smile, appears humanoid here also. We note that everything that has to do with the representation of the face, whether of an animal in mask position, or of a man, appears in the Bubaline, and especially in the Mathendous, tainted by taboos: human faces without masks are often little dealt with, sometimes left without any internal detail, sometimes even deliberately replaced by a smooth, shapeless expanse. Sometimes we hesitate between mask (of a cat?) or caricature, or simply conventional sketch (e.g., Castiglioni-Negro, 1986, fig. 484). Sometimes the faces are endowed with non-realistic details expressing a graphic design or tattoos (tribal?), considered more important than real organs, and which  make the facial features appear strange to us (e.g., ibid., fig. 82) (fig. 4 f).
This man with the “jackal” mask is sometimes, but more rarely than in the Djerat, represented as ithyphallic, and even more rarely with a huge phallus. We note that, more generally in the “Bubaline” of the Mathendous, a few scenes of coupling are known, without masks, but they are rare, and the atmosphere of these scenes (with hieratic women holding mysterious ovals, for example) have nothing of the priapic character of the Djerat scenes. We even note here a few cases of coupling involving masked men: a bestial coupling of a jackal-man with an elephant (Jelinek, 1985, I, fig. 49, and perhaps also fig. 50), and a probable coupling of masked men that are sometimes horned (ibid., fig. 51 and 1984, I, fig. 59).
The jackal-man is found either in isolation – for example (fig. 5), engraved in a fracture in the rock, he looks toward the dark interior of a severe fracture – or in many cases (about half) in the heart of compositions apparently interpretable as related to hunting activities. Thus, in the well known scene of men dragging a rhinoceros shot by an arrow (fig. 4 g), or in that of men clutching an antelope’s neck (fig. 4 e). Moreover, one of these jackal-men carries an ox on his shoulders (Castiglioni-Negro, 1986, fig. 122). This ox is probably a wild ox killed in the hunt (its “pincer” horns, unusual in the domestic cattle of the Mathendous, suggest this). In a dozen cases, the masked man carries a bow, sometimes strung (ibid., fig. 406), or various other weapons (ibid., fig. 480). But it is clear that these varied scenes are symbolic and are not intended simply to describe hunting exploits.
Theriomorphs or cat masks are also known from the Mathendous (e.g., Graziosi, 1970, fig. 176). We also note the famous “gatti-mammoni” (“cat-monkeys,” “Meerkätze” of Frobenius, Castiglioni-Negro, 1986, fig. 61, 125) because their attitudes more resemble a man than a macaque. The long tail and the head with short ears are reminiscent of the heads of the Djerat, and are those of cats, here in good “Naturalistic Bubaline” style (fig. 4 d). There are also in the Mathendous some bestial features, just as at Djerat (e.g., fig. 4 b) (ibid., fig. 266, 267, 334), on identical persons, seen from the front, with short ears, legs spread and huge phalli. The facial expression is, again, vaguely humanized. Their patinas are a little light, however. Their rough style and light patinas do not allow them to be included in the “Naturalistic Bubaline” group. They belong to a more recent group, which we have not succeeded in identifying.
Some masks have an equid muzzle (e.g., ibid., fig. 121 – Jelinek, 1984, p. 130 – 1986, II, p. 247-48, with horns, however), others, long ears (e.g., Jelinek, 1985, II, p. 228), but the species often remains undetermined (e.g., Castiglioni-Negro, 1986, fig. 192 and 406, again with archers related to hunting activities). The masks frequently bear horns (e.g., ibid., fig. 70, 121 – Jelinek, 1985, I, fig. 51-1985, II, p. 247-48) but these are often not realistic, or unrelated to the species of animal represented by the mask (fig. 4 a). They must represent some independent, conventional value. The false tail is quite often found, instead, on persons, masked or not, or is a classic loincloth hanging down, with two or three divisions, apparently ceremonial, typical as among the Round Heads. Also as among the Round Heads, there are big ears (or coiffures?) with two lobes, that are unrealistic and conventional (e.g., the “gatto-mammone” or cat-monkey, Castiglioni-Negro, 1986, fig. 119, the man-jackal) (fig. 3 e). Some muzzles and horns of the hartebeest antelope are noted here also. 
There is a single case of a mask with sheep horns and a single one with the horns of a chevaline antelope, for their consonance with the “sacred” animals of the Round Heads. Finally, some unmasked men simply wear horns (fig. 6).
Other masks from the Mathendous leave one hesitating between cat and canine (e.g., ibid., fig. 82) and the species of the animal is often indeterminable (ibid., fig. 70). Some strange details: one jackal-man has a rhinoceros horn (ibid., fig. 80), a man with a very long muzzle, probably of a hippopotamus, pursuing two hippopotami (fig. 4 c), an equid mask (or that of a bovid?) with horns, surmounting a quite human head, with detailed hair and profile (ibid., fig. 121). That the mask may be worn atop the head is not necessarily an error of the engraver: some masks of the Dogon are worn this way (Dieterlin, 1990, p. 8). One engraving, the “Garamantian Apollo” (Graziosi, 1981) (fig. 4 a), famous because it was the first rock depiction reported in the literature, also shows a muzzle that could be that of an equid or a hartebeest (or an ox as Graziosi suggests), surmounted by fanciful horns. In one case, the muzzle is of exaggerated length, like an elephant trunk (Castiglioni-Negro, fig. 414).
These masks and theriomorphs of the “Naturalistic Bubaline” group are almost non-existent outside the Tassili and Fezzan. The Hoggar, where one still notes representations of this style, contains almost none. One does not find them either in the vast and important part of the area of distribution of this linguistic group constituting the Saharan Atlas, southern Morocco, and the Rio de Oro (4). The only hybrid figure that we may note there is that of a well known engraving from Mouchgeug (Mountains of Ksour), which is also clearly symbolic. Before a decorated ram, two symmetrical people – or animals? – are enthroned in a hieratic posture, joined back to back. Their heads are small and bizarre; one believes one sees there an ovid [or sheep]. Their common tail forms a spiral around them. One of them is ithyphallic; the sex of the other is disputed (Lhote, 1970, p. 67). All this highlights a spirit quite different from that of the masks of the Djerat and Mathendous.
These differences in themes within the same “bubaline” stylistic group well illustrate that “styles” and “artistic groups,” when they are found over areas as large as that of the “bubaline” group, may not, for reasons that escape us, follow exactly the same style (see discussion in: Muzzolini, 1986, p. 110-18).
IV. Masks and theriomorphs of the Round Heads
The Round Heads comprise a group of paintings confined almost exclusively to the Tassili-Acacus and more or less contemporary with the “Naturalistic Bubaline”; that is to say, it too dates to a very ancient period, essentially. Some paintings show, however, at least the survival of its techniques, of some themes and of some of its styles, highly degenerate, into later periods contemporary with the “Final Bovidian” (Muzzolini, 1989).
Masks and theriomorphs are known (5) in several of the sub-groups that have been defined by the Round Heads, both in the oldest (the “primitive Martian”) and in the most recent (“figures in solid ochre”) (Muzzolini, 1986). This group follows a schematic, expressionist style both for human faces, which respect a taboo or convention banning, as in Islam today, the representation of the sense organs, and for the faces of animals. Thus, the determination of the species of animal represented by the masks  proves difficult. One finds (6):
1. Figures decked out with a vaguely humanoid head but surmounted by two large lobes. These lobes may be the large ears of a donkey or hare, two species known both from excavation and from representations. They may instead simply represent conventional hairstyles. In any case, highly symbolic values are evident in the context of these “bi-lobes.” They are primarily found in the subgroup of the “Martians.”
On a wall in Sefar, next to one of these “bi-lobes,” there are two other persons whose heads are replaced by a design significantly larger than the head and representing a “jellyfish,” a strange animal or graphic unique to this group of Round Heads and sometimes represented independently (fig. 7 f). These “jellyfish,” so called because wavy filaments hang down, sometimes resemble a turtle because it comes out with a very short head or limbs, with hands or claws. Obviously we do not know how to explain the meaning of these intertwined symbols. We maintain for our purposes only that human heads are replaced here by “jellyfish” as “equivalents.” A confirmation of this “equivalence” is suggested at Tan-Zoumaitak, where there is a human head in the style of the heads of “Martian” Round Heads – a simple circle, decorated inside with typical chevrons –represented alone, but the usual filaments hang down from the “jellyfish.”
2. Paintings, at Sefar and Aouanrhet (central Tassili), represent the mask alone, seen from the front, perfectly symmetrical, without even the hint of a neck. In one case, we face a horizontal series of three of these masks, joined together, and identical. They are indeed masks because they are also found in mask position on a human body (Fig. 8) (Lajoux, 1977, p. 55-56). They are depicted in a vigorous, expressionist style, very modern looking, close to pure graphic and exceptional among the Round Heads. They outline, but very schematically, the sensory organs, unlike the usual taboo of the group. What referent did they represent, if they were intended to “denote” anything? For Lhote (1973, p. 87), one of these masks, that of the “sorcerer” of Aouanrhet, evoked an antelope. But the high forehead is rather humanoid and some of these masks present a detail of head adornment, the “compartmentalized band,” that is often found on humans of the latest Round Heads. That the mask of the “sorcerer” is surmounted by horns proves nothing among the Round Heads, one is trying to say: the group even endows real un-masked humans with horns, like the White Lady (Lajoux, 1977, p. 60) or like the humans of Uan Tamauat (Acacus) (Fig. 7 a). That such drawings of masks constitute a symbolic motif as much as an image and only suggest, in broad strokes, the basic structure of man or animal’s face, is ultimately all that we can advance. These expressive masks belong to subgroups of recent Round Heads.
3. Some “cat” masks: they resemble ogres in children’s books, but an interpretation as “cats” remains possible (e.g., Lajoux, 1977, p. 63).
4. Genuine masks represent a very elongated snout, short ears, and sometimes horns, an animal very similar to that of the “Garamantian Apollo,” and quite difficult to interpret. Thus, at Asadjan-oua-Mellen (central Tassili) (Fig. 7 e), a masked hunter bends his bow (an unusual trait among the Round Heads); another is seated at his feet. Both muzzles are elongated, though not identical, but one bears horns – a bit fanciful, but capable of being seen as twisted – the other not. This prompts one to consider the horns as an additional value , independent of the facial mask. One can see an equid in these masks as easily as a hartebeest [or bubaline antelope]. We note on this subject that the animals of the Round Heads obviously invested with “sacred” value are the sheep and the chevaline antelope, not the bubaline antelope. But sheep and chevaline antelopes are not represented among the masks. Such a disjunction is probably significant. The bubaline antelope, itself, hardly appears in the Round Head frescoes: despite the stylization of the drawings, it is other species of antelope that one perceives.
Except in a remarkable case from Sefar, which involves precisely a mask. Between two worshipers and a large ox, all in “Martian” style and white in color, apparently in connection with a “great god” further to the right, a large hartebeest antelope occurs, identifiable by the shape of its horns and by the bearing of its head. Its “sheath” indicates the male (Fig. 9). Precisely from this sheath comes a descending row of three small antelopes, a row that crosses a mysterious circular sign in ochre, traced beneath the belly of the antelope, and visibly tied to this belly, where it stops. The last small antelope is held at the neck by a masked man with long ears, who appears to be in a position for coupling with the antelope. Our reading is not certain, because such a scene is unique in the Tassili. We are obviously tempted to project onto the composition a meaning related to the theme of procreation, but this may only be a matter of subjective interpretation. We maintain only that the masks may be linked to themes of this nature.
5. Men who are only horned are frequent in the group, especially among the schematic little devils. Often it is unclear whether the extensions above the heads should be interpreted as feathers or as horns, but in some cases horns are certain. Their shape evokes cattle or antelopes (e.g., Uan Tamauat, Acacus, Fig. 7 a). These are not masks, strictly speaking, but the relationship to an animal probably comes from the same category in the collective imagination of the group. The symbolic value of these horns is particularly manifested in the very neat horns of the White Lady, previously mentioned. They recall exactly those of the great bubaline realization of the Mathendous, the engraving of a “goddess,” legs apart. In both cases “objects” appear between the horns that are scarcely interpretable and obviously symbolic.
In the most recent groups, the faces are sometimes crossed by a reserved area in the form of an unpainted, rectangular surface. Here, it is not a real mask, either, though this reserved area was sometimes interpreted as a wolf or veil. We would simply see this as evidence, again, of the taboo against facial features, so common in the group of the Round Heads (e.g., White Lady, Tamauat, Fig. 7 a). We also note that faces, in all the groups of Round Heads, frequently bear paint.
Note that false tails are rare among the masked depictions of the Round Heads. The tails only hang down from cloth, often bifurcated or trifurcated. An enigmatic object, in the shape of a fork or magic wand, is carried by certain masked figures (e.g., Lajoux, 1977, p. 47) and we also note it, identically, among some “bubaline” masks (e.g., at Mathendous, Fig. 4 g). This confirms a certain contemporaneity and probably some known symbolic value.
We note, finally, much less frequently than in the “bubaline” group,” some masks or theriomorphs with the muzzle of a canid. We also recall that the taboo on sex is typical among the Round Heads, so one scarcely sees in this group the magnified phallus of the “Bubalin” (8). 
V. Paintings of the bovidian period
The absence of masks or theriomorphs is total in the paintings of the “Old Bovidian,” of Negroid anthropological type (“group from Sefar-Ozanéaré).
In contrast, the mixed Abaniora group (“mixed” because it represents both some of Negroid type, some Europoid types, and especially a type of non-Negroid Blacks similar in appearance to modern Fulani) represents some there, very rarely. In the final Bovidian, the group from Iheren-Tahilehi, exclusively of Europoid types, are represented similarly: they are few but clear. The symbolic character, in these groups of the “Bovidian,” is still evident.
Thus, at Tirehart Wadi (near Tamadjert, Tassili N.W.) (Kunz, 1988, Pl. I), a pitched battle seems to oppose two rival bands. On the left, the Good Guys: a troop of archers, painted flat as in the Abaniora style. They are all decked out with masks of canids perfectly reminiscent of those of the “bubaline” figures of the Mathendous, and with short tails. They draw their bows against the Bad Guys, represented on the right by a troop of baboons, done in a very different style, that of the late Round Heads. These baboons, although riddled with arrows (many already have two or three stuck in their bodies) are facing their foes. Their aggressive stance, standing first, some with feet that are too humanoid, suggest they could be human enemies that are bestialized. The baboon is rarely represented in rock art but a few others are known. Thus, a site of Round Heads in the Tassili, Tizzeine, presents a good example, painted in “Martian” style.
A battle with the same aspect as that of Tirehart, with masked men, but more confused, occurs at the Wadi Imirhou (central Tassili) (fig. 10). At Baidikoré (N.W. Tassili), three men with canid masks (fig. 11), whose style evokes both that of Iheren-Tahilahi and the late Round Heads, brandish bows and throwing sticks simultaneously, and rush toward prey that is not visible.
At Tazerouk (near Dider, central Tassili), a shelter with paintings in the style of Iheren-Tahilahi, a little elf bears horns (fig. 12). Nearby, there is a composition that is quite effaced; nevertheless, it leaves visible a figure who pulls off the mask of another (fig. 13). At Sefar, before a classic mounted ox, a man is walking who wears an elephant mask with big ears and trunk. At Ti-n-Tarleften A (N.W. Tassili), a shelter in the style of Abaniora (Kunz, 1979, Pl. 4-4), a man wears the head of an equid or of a bubaline antelope.
We note, in the Abaniora group and especially in the group from Iheren-Tahilahi, very many simply caricatured drawings, representing faces “with dog snouts” (e.g., Boccazzi, 1990, fig. 1). All the intermediate forms exist, from the person with a very naturalistic, perfectly Europoid profile, to figures “sketched” in a hurry, with misshapen muzzles, which make us think that these are not canid masks at all. They have neither physical meaning (it is not a matter of prognathous Negroids, as has strangely been objected, in contradiction to our denomination of an “exclusively Europoid group” for the group from Iheren-Tahilahi), nor symbolic meaning, any more than the kind of oversized noses or spindly faces with which the cartoonists Sempé or Faizant depict the French today.
VI. Other pre-equid or pre-camelid schools [i.e., groups before the domestication of horses or camels]
Besides the “Naturalistic Bubaline” group, engravings of the central Sahara include:
n On one hand, the “Tazina school,”  a collection of engravings of a very original schematic style, widespread throughout the Atlas, an area as vast as that of the “Naturalistic Bubaline.” Neither masks nor theriomorphs are reported. But this negative observation is hardly significant here, because this group very rarely depicts people.
n On the other hand, a large batch of engravings, made in various styles and techniques, which we have proposed to leave (for the moment) in a pseudo-group of “unclassifiables.” This is what Lhote called “Decadent Bubaline” and “bovidian engravings” – the concepts and meanings of these two entities are too vague and subjective to be accepted – and various other local styles not reliably allocated to known artistic groups. While these are certainly prior to the camelid period, and some compositions are even prior to the period of the horse, probably, that is all that can be said as far as age (Muzzolini, 1986, p. 107-10 for discussion). Masks or theriomorphs, relatively few in total and not homogeneous in appearance, have been reported in these various groups. One notes, for example, among the engravings of the Hoggar (Trost, 1981, p. 188-1990, fig. 2, 4, 6), on one hand, a mask, apparently of a jackal, but here with an enlarged phallus and false tail, and on the other hand, “cat” masks recalling those of the Djerat. But the coarse style and overly light patina do not allow their attribution to the “Naturalistic Bubaline” period. Similarly for a masked figure with jackal head and enlarged phallus from Ti-n-Affelfelen (Hoggar) (Camps, 1974, Pl. 19), in crude style, with a dotted outline and light patina. We do not know what to relate these various masks to.
The summary character of the depictions of this “unclassifiable” group often leaves one in doubt: is it the head of an animal, or simply the head of a man drawn without concern for realism, and which presents by chance a deformation resembling, for example, a canine muzzle?
We must make a separate group for a series of engravings from the Acacus, those of Ti-n-Lalan. They include, in particular, a scene of coupling (fig. 7 c). The woman is lying down with legs apart; the details of her face, hair, and ornaments are rendered with precision. The man, himself, is drawn more summarily, but bears the head of a jackal, with a vaguely humanized expression, here again. His phallus is, as usual, enlarged. Despite a relatively naturalistic rendering, the composition cannot be attributed to the “Naturalistic Bubaline” style. As far as technique is concerned, it uses a fairly coarse pecking, not the polished or carefully pecked technique of the “Naturalistic Bubaline.” Its patina is relatively light. One can hardly attribute this composition to an already listed group. This seems quite old, perhaps even pre-horse, and one would not know how the clarify its attribution any better. However, the elements of the composition are interesting: they are close enough to the themes of the Ahana Rock in the Djerat (realistic depiction of coupling, large phallus, etc.). But while scenes of coupling are rare in the Mathendous, the mask is indeed that of a canid here, typical of the Mathendous. Perhaps this representation from Ti-n-Lalan must finally be related to the rare realistic couplings that we reported in the “Naturalistic Bubaline” period of the Mathendous, but where the symbolic aspect, instead of being expressed by masked men, is transferred to the woman.
At Ti-n-Lalan, the scene with jackal mask is repeated in a copy of mediocre style, nearby (Mori, 1965, fig. 41). But the latter replaces the jackal head with a head similar to that of the “cats” of the Djerat and the Mathendous, even adding their enormous phallus. Are the heads of jackals and of cats then to be symbolically equated? Like the cat masks of the Mathendous, this one has a light patina, thus  probably is of recent age. Some other masks of late age are also known, in the Acacus, with enormous phalli but rabbit ears. Cat heads and rabbit heads, and more generally all animal species expressed by the masks – are they therefore equivalent as far as the function of the masks? We will return to this point.
VII. Recent periods: stage of the Libyan warrior, period of the horse, period of the camel
The inventory is much simpler for all the recent periods: it is virtually impossible to note any definite representation of a masked figure or theriomorph.
Not in the vast group of engravings of the stage of the “Libyan warrior” which, in the first millennium BC then in its recent phases – which continues up to today – covers the Air and the Hoggar and touches lightly on the Djado. Nor in the group of the “Herders of Ti-n-Anneuin,” in Northern Tassili, and contemporary with the beginning of the Tassilian “period of the horse.” Not in the latter – which includes paintings and engravings – nor in the groups of chariots and horses that appear on its periphery, including as far as the Saharan Atlas. Not in the many camelid engravings and paintings of all regions (9).
The atmosphere has changed. The warrior has replaced the herder everywhere; women disappear almost completely from the compositions. The symbolic world reflects – or imposes – these profound social transformations. The ancient masked figures and theriomorphs are no longer valid. We will try to understand the meaning of this change.
VIII. Peripheral countries
To better locate the masks and theriomorphs of the depictions of the central Sahara, let us take a quick look at those represented in the surrounding areas.
Toward the north and west, the “Naturalistic Bubaline” and the Tazina school extend as far as southern Morocco and to the Rio de Oro. They are as devoid of masks and theriomorphs as the Saharan Atlas. Toward the south, the tradition of masks becomes scarce as soon as one passes beyond the Tassili. In the valley of In-Djerane, only some figures with heads “in a V” or with head ornaments that can hardly be interpreted as masks can be pointed out. In the southern Sahara and the Sahel (Air, Adrar of the Iforas, Mauritania), the rock art groups, which are numerous and very rich – with tens of thousands of engravings – only show representations from recent periods, contemporary with the “stage of the Libyan warrior,” except for some rare productions that are scarcely any older: none of these groups includes masks or theriomorphs. To find them, one must pass beyond the Bandiagara to Mali, but it is then a completely different world which opens to us, the world of Black Africa. Until the emergence of the empire of Ghana, or more or less to the network of commercial centers, such as Jennejeno (Mali), which could have been established a little earlier, the destiny of that part of Africa seems indeed to remain completely independent of the Saharan world. A possible invasion of Saharan people toward Ghana, around the “Arid post-Neolithic” (around 2500 BC), a thesis held by Davies, has not left any trace in the rock art groups of the Sahara or of the Sahel.
Europoid settlement continued in the Tassili-Hoggar-Air-Fezzan group, from ethnic groups following the bubaline style and the Round Heads, followed by those of the Iheren-Tahilahi style, to those of the “stage of the Libyan warrior,” of the periods of the horse and camel. This settlement comprises ethnic groups that are certainly varied, but all create rock art (even  the Round Heads, who are often claimed, wrongly, to be Negroid). This Europoid world ends, toward the east, at Djado, and never penetrates the western Tibesti.
For the eastern Fezzan, Le Quellec made groups of engravings known that were of quite summary design for the most part: some heads of bizarre forms may be masks, some bear long ears (or feathers?), but the identification is rarely certain. Three masks in profile that may resemble a canid or a baboon are completely certain, one of them with an enlarged phallus (1989, fig. 2, 10). The age of these engravings is imprecise, but relatively recent, post-dating the “Naturalistic Bubaline” period (fig. 8 b).
A little farther south, among the engravings of the Djebel ben Ghnema, some masks are noted: they comprise some groups that are quite unique and stereotyped with small imps in acrobatic positions, or floating, their arms bent behind their backs. The draughtmanship is very rough, the represented species poorly distinguished: equids? or canids? (fig. 7 d). Some are provided with horns of various forms (in front, behind, diverging like the horns of sheep, etc.), or large ears (Ziegert, 1967, Pl. 80, 90, 132, 137, 151). These floating figures, of small size, are also known in one of the rare paintings of the Mathendous (unpublished) and in the paintings of the Gilf Kebir (Rhotert, 1952, Pl. 30 and 31).
The Tibesti has seen other rock art sequences, of which some schools are probably as old as the “Naturalistic Bubaline” style and that of the Round heads. But very few masked figures or theriomorphs are known that are as clear as those of the central Sahara. Usually, the “masks” noted in the literature (e.g., Huard-Leclant, 1972, p. 21 – Beck-Huard, 1969, p. 165, 210, 212) correspond to very worn representations, so that indecision occurs between caricatures and true masks with symbolic intent. Some rare cases may be credited with this symbolic intent, just the same: thus the dancers in pointed facial profile (with canid heads?) from Mossei (Beck-Huard, 1969, p. 213).
We can hardly point out any representations of masks or theriomorphs in the rock art groups of the Borkou or of the Ennedi, few of which have been published. At Ouenat and Gilf Kebir, extensively published bodies of art, the absence is also just as complete (Rhotert, 1952, Pl. 32, shows one person with big ears, but we only know it as “mentioned,” and uncertain).
At Dakhla and in the representations of Nubia, examples are similarly very rare and, because of the summary nature of the engravings, uncertain.
Neither Marmarica nor the Libyan Desert include any rock art representations; Cyrenaica presents very few. Tripolitania, richer, contains some old engravings and some belonging to the “Naturalistic Bubaline” period. But it yields neither masks nor theriomorphs any more than the rest of the Maghreb.
Finally, we add that if our survey is limited to rock art representations, no Saharan archeological discovery of any era has yet confirmed the existence or use of any element traceable to masks or theriomorphs: we are absolutely reduced, in this chapter, to rock art documentation.
We can now recap, situating spatially and temporally the groups where there clearly existed the practice of representations of masked figures or theriomorphs. We disregard here those representations where we hesitate between caricature and symbolic intent, and more generally, the groups  or representations where animal heads only rarely are of that type. An isolated graffito of an animal head is, still today, a banal, individual manifestation, that may derive from multiple motivations. But science only focuses on the usual and the only groups that interest us here are those where there socially existed a symbolic tradition of this trait, obviously integrated into the official culture of the ethnic group.
Seen in this light, the overall picture ultimately appears very simple. Throughout Saharan rock art, we only see such a symbolic tradition, socially structured, with human representations bearing animal heads, in three groups and in a very specific area:
· The “Naturalistic Bubaline” group of North Tassili (Djerat) and the Fezzan (Mathendous), to which we add, though in a different style and a slightly later date, the representation of the jackal-man of Ti-n-Lalan (Acacus);
· the Round Heads (Tassili – Acacus);
· the groups from Iheren-Tahilahi and incidentally the group from Abaniora which is essentially contemporary.
The first two groups, the “Naturalistic Bubaline” and the Round Heads, are loosely synchronous. Some subgroups of Round Heads, however, were able to survive until the “final Bovidian” of Iheren-Tahilahi and transmit to the group of that name the remains of this tradition (Muzzolini, 1989). Nevertheless, the Abaniora and Iheren-Tahilahi groups are chronologically distinct from the first two and are located after the clearly defined episode of the “Arid post-Neolithic,” which certainly fragmented the Saharan populations.
All these groups are made up of non-Negroid ethnicities. The group of Negroid types of the “Old Bovidian” (the group from Sefar-Ozancaré) does not participate in the tradition of masked figures and theriomorphs, any more than the Negroid or not-clearly-Europoid groups of the Tibesti, of the Ennedi, of Nubia, or of Ouenat. The scarcity of masks and theriomorphs in the Hoggar, their absence in the Air, the Adrar of the Iforas, in Mauritania, must be related to the fact that these areas contain few or no representations from the ancient period.
In contrast, Tripolitania and the Saharan Atlas, including as far as the Rio de Oro, contain an abundance of “bubaline” representations, from the ancient period, and we are sure, from the identical nature of numerous technical, stylistic and thematic traits, of their contemporaneity with those of the Djerat and Mathendous. The discrepancy between the two symbolic worlds is, however, obvious: in the north, the cult of the ram, and total lack of “great gods,” horned goddesses, or masked figures or theriomorphs; in the south, horned goddesses, masks and theriomorphs in the “Bubalin” and in the contemporary Round Head paintings, the “great gods” of the Round Heads, the total absence of “decorated rams” – and few rams, even without ornament. Such differences must be significant.
In summary, the picture that emerges is that of a symbolic tradition of masked figures with animal heads and of theriomorphs, circumscribed to the Tassili-Acacus-Fezzan area, in existence during the “Humid Neolithic” (approx.. 4500-2500 BC), probably semi-interrupted during the “Arid Post-Neolithic” (approx. 2500-1000 BC), and only a few remnants of which are found at the beginning of the “Third Humid” period which follows, circa 1000 BC. Then the tradition is completely lost.
In this area, actually quite limited spatially and temporally, do a few constants appear in the types of masks represented?
There are very few. We first note a tendency to confer on these animal icons a few hints of their underlying humanity – the overall expression of the face, a smile, the compartmentalized headband – as well as feeling that they are there as signs of their creator.  As for the species represented, there is nothing like the rigorously structured, zoomorphic pantheon of Egypt. At most, we can recognize a certain generalization (less clear among the Round Heads) and a continuation through the millennia of the mask with canid head. The mask with cat head also obtained a certain vogue in the “bubaline” period: but it is bound to contexts termed “erotic,” with enlarged phallus, of the Djerat, and much less frequently found elsewhere. Both types sometimes occur together (e.g., in the Mathendous, Graziosi, 1970, fig. 176). No scene of worship of such masked figures occurs. The impression prevails that all our theriomorphs are indeed masked figures, perhaps reflecting some mythical being, or a belief in the transformation of a man into an animal for some magical purpose. Such a belief, well described by Lewis-Williams for the shamans of the San, is common to many cultures.
What is astonishing, however, is the variety of masked figures, humans that are simply “bestialized” with a silhouette of an indeterminable animal, or only wearing horns, etc.: for us, since we do not know how to interpret them, this is only an ideological wilderness in which no landmark emerges.
There is even a serious question for us: are we sure that these masks have meaning? Bouchet (1965) explains that, among the current masks of the Senufo of the Côte d’Ivoire, the forms of masks, whether zoomorphic or human or indeed the graphic form they represent – or rather, seem to represent, in the first sense – have no importance. These forms do not convey any meaning; they are not a privileged ideological support. Meaning is expressed in the ornamental details, in the “accessories in secondary appearance,” known only to insiders. This explains the continual evolution of the forms of these masks, and even their frequent abandonment as ritual objects. If it were the same with our bubaline and Round Head masks, any attempt, such as ours, to rationally classify based on the icons they represent, would be a futile enterprise. It could be, for example, that the meaning is expressed by the presence or shape of the horns, the frequent independence of which we noted in relation to the head of the animal depicted. Whether it is the canid mask, or that of the cat or the hare, all may be indifferently represented, under the sole condition that a certain detail of the face appears.
A more reassuring hypothesis would be that the less frequent masks or the various animal attributes reflect local beliefs that are more varied than those manifested by the “great gods” and the masked figures with a jackal head. They could, for example, simply indicate the Djenoun villagers, the equivalent of the cult of local saints in Christianity.
The reader will understand that our discussion soon stops once we enter the domain of interpretation of the meaning of masks and theriomorphs. Without a sacred text or oral tradition or some correlation with contemporary religions – such as Egyptian religion – our interpretation has only the vaguest ideological notions; or if it is precise, it contains too much subjectivity. What also appear subjective to us are attempts to “explain” the frescoes by analogies to the beliefs of modern Fulani, which would require, among other incredible things, an invariance in myth and its images for 5000 years (10). As for trying to find here the manifestations of shamans and their states of trance, as Lewis-Williams does for the Bushman frescoes, this would be practicing an ethnographic comparison with nothing to ensure its validity.
Can we at least, regardless of the meaning, glimpse the function of these masks and theriomorphs? Here again, information proves to be meager. We noted that many  compositions with jackal mask from the Mathendous seemed related to hunting. At Sefar, on a wall with paintings essentially of the Iheren-Tahilahi style, one notes several animated scenes of the hunt for various antelopes. One man wears a zebra mask, with the head of an animal that is hard to identify (11). He himself is also probably involved in a hunting scene, and runs toward the prey (fig. 14). Could the mask, an explanation often advanced, be a disguise to deceive the hunted animal? A more common method of hunters of the Iheren-Tahilahi group to lure and approach the beast is to carry animal skins on their backs, or to be followed by a sheep. The hunts we note can only be symbolic, and the atmosphere of the compositions of the Mathendous would speak rather in this sense. We can hardly overcome these many uncertainties.
Other themes perhaps related to the function of the masks are the frequent ithyphallism, enlarged phalli, and scenes of coupling. They appear even in groups, like those of the “Bubaline” of the Mathendous, who seldom use these themes aside from their appearance among the masked figures. The compositions involving cat masks have a more priapic nature than those involving jackal-men. But neither the one nor the other seem capable of being reduced to “shepherd’s entertainment” as has sometimes been claimed. Ethnography offers many avenues of explanation: the representations may attest to the sexual vigor of the authors, or have the effect of preserving it, or of obtaining it, or they have a prophylactic goal, or they intend to ward off sterility in women, etc. The scenes of coupling may, more simply, recount a “sacred marriage” included in myth. There is no thread to guide our choice among these various interpretations.
X. Relations with Egypt
Could this tradition of masks and theriomorphs, identified in the Tassili-Acacus-Fezzan group during a period roughly contemporary with the Predynastic period and Old Kingdom, with some survivals contemporary with the New Kingdom, have something to do with the many animal-headed gods of the Egyptian pantheon? These already appear in the Old Kingdom in their basic form, and certainly have their roots in the Predynastic period.
But although the dates do not prevent contact between the two traditions, Saharan and Nilotic, one must note the complete absence, in archeological excavations and also in the rock art, of any unambiguous trace of contacts or of cultural influences between Egypt and the central Sahara. Some traits are obviously common to the two regions. But without demonstration, by chronologically intermediate transmission, they can be analyzed equally well as due to accidental convergence, or as the results of parallel evolution starting from a common ancient basis. For example, the beautiful workmanship of the Tenereen flints, often close to that of the Predynastic period. Or the horned goddesses of the Round Heads, such as the White Lady, which obviously evoke the Egyptian Hathor, as well as the Asiatic Ishtar. The absence of intermediaries and the isolated character of these common traits only constitute negative findings, of course, but their general nature is striking in an area as vast as the central Sahara and the Maghreb, and in all periods as far back as the proto-historic age, and we believe it confers value to these negatives. It seems surprising, if there was contact, even indirect, that there should not be found somewhere in the Sahara, the Sahel, the Maghreb, some unambiguously Egyptian objects, attesting to these contacts: some scarabs, some  typical vases, some ushabtis, such as are found in Southwest Asia, the Sudan, Crete, Cyprus, Anatolia, and Andalusia. Objects, elements of architecture, and bas-reliefs testifying to an Egyptian presence or influence are found as far as Kharga and Dakhla. But when one plunges into the desert, at Abu Ballas, 200 kilometers to the southwest, some jars and a very awkward engraving though still respecting the Nilotic canons, are the only evidence of Egyptian influence. Another 400 km and, at Gilf Kebir or Ouenat, two mountain chains that were important centers of pastoral life in the second and first millennia BC, contemporary with pharaonic Egypt, a very rich rock art documentation is noted, but nothing, absolutely nothing, evokes an Egyptianizing or Egyptian cultural object or trait. It is the same in the Ennedi, Borkou, Tibesti, Fezzan, Cyrenaica, and Tripolitania. It was not until the middle of the first millennium BC – that is to say, after the founding of Carthage and Cyrene, with the early pharaonic expeditions in Cyrenaica, the establishment of an extensive network of relationships across the Mediterranean Sea, the early Punic navigations in the western Mediterranean and on the Atlantic coasts – that elements foreign to the central Sahara occur, deriving from the Mediterranean coasts: the horse, chariot, writing, use of the spear, shield and sword, the first metal objects, even the camel.
Nevertheless, some authors have argued for the existence of contacts with Egypt or of influences from there. Lhote, (including 1958), first advanced this thesis and published images of Tassili rock art with a strong pharaonic air: an “offering scene,” and especially his four “goddesses with bird heads,” which continue to appear episodically in the literature, and even in recent works. In the second edition of his work (1973, p. 235), Lhote remains very ambivalent: he explicitly returns to these statements, but without absolutely rejecting any influence, and publishes these scenes again with the same captions and comments, linking them to an “Egyptian influence (18th dynasty?)” : the doubt concerns the date, not the fact. He does not say – probably he was still unaware – that everyone has now realized that these “goddesses with bird heads” and the “offering scene” are fakes. The winters were long on the Tassilian plateau, for the art students from Montparnasse whom he had brought, in 1956, charged with carrying out the “recording” of the frescoes. And people amuse themselves, at times, as they can. The fun, this time, went a little too far (12).
More seriously, Huard has also, in his many writings, supported the thesis of contacts between the Nile and the central Sahara, thanks to which he explains the diffusion of many cultural traits identified in the two areas. He thus affirms “the uniqueness of the Culture of the Hunters,” the latter being formed of such cultural traits “distributed throughout the subcontinent” (Allard-Huard, 1981, p. 64). Other writings argue that there would only be an “archaic substrate” common to the two areas: their evolution eventually would have diverged. The second position seems much more acceptable and we will support it in our final attempt at the interpretation of the masks. But it seems to us inconsistent with the first. And it is hard to reconcile, especially with other writings by Huard, where these archaic traits, instead of being only cultural elements common to many specific later groups, are described as those of one group, itself specific and structured, the “Hunters.”
The latter are in fact sometimes advanced (e.g. Huard-Allard, 1977, p 659 or 1981, p. 64) as being a particular group, occupying a specific site, from a particular place, as  with real historical groups: indeed, one seems to speak in this case of an ethnic group, not a concept. We confess that we do not understand the mixture of such diverse statements, which seem incompatible to us.
In any case, the cultural traits in question only constitute characteristics that are either trivial, and therefore may only reflect accidental convergences, or very vague, sufficing to justify – or at least to illustrate – the hypothesis of a former basis that was common to particular subsequent groups, carriers of these traits, but insufficient to support the thesis of real contacts between Egypt and central Sahara in recent millennia.
It therefore appears unlikely to us, ultimately, that the tradition of masked figures and theriomorphs of the central Sahara, in the period of the “Naturalistic Bubaline” and the Round Heads, reflects some contact with predynastic Egypt and the Old Kingdom. It also appears unlikely that the masks of the Iheren-Tahilahi group, in the latest period, have anything to do with Egypt: they are explained much more normally as remnants of the tradition of the previous era, preserved only among the Europoid populations of the central Sahara (13).
XI. Attempt at interpretation
The description of Saharan masks and theriomorphs and their application in the general Saharan and African frameworks stops here: beyond, the field of speculation opens up.
It is necessary. Simply “objective” analysis, even that of the New Archaeology, which only aims to quantify the archaeological record, only reaches the obvious, measurable surface of the evolution of societies: their material culture, their economic and environmental adaptations. Now we aim to reach the heart, that which is actually the basis of the identity of the groups. But how? Conscious of the dangers of empiricism and of subjectivism, we try all the same to get somewhat beyond mere observed facts and at least integrate them into a story, the most fundamental in all human societies, that of the ideologies, beliefs, and symbolic structures that generally reflect social structures. This story is part of the level that F. Braudel called the “long term.” Unfortunately, it hardly leaves any easily quantifiable material traces.
We propose to read the meaning of the Saharan masks and theriomorphs following two scales corresponding to areas of differing magnitude, during two different durations: first, the central Sahara only (to be understood, in this section, as reduced to the Tassili-Acacus-Mathendous) during the late Holocene; then the whole of North Africa throughout the Holocene and the end of the Pleistocene.
On the scale, first, of just the central Sahara in recent millennia, the Saharan masks and theriomorphs do not confirm the existence of a religion of or belief in an “animal-god” with the head of a jackal or cat. These heads are not even “attributes” symbolizing a deity: that, if it exists as such – a “great god,” a horned “goddess,” a bull or ram – is represented independently. The difference here is essential, with the status of the zoomorphic gods of Egypt, or that of the panther-goddess of Çatal Hüyük, true deities (Cauvin, 1981, p. 26). Nevertheless, these masks and theriomorphs are symbolic elements that fit into a larger symbolic universe, specific to the same places and the same artistic groups: the “Naturalistic Bubaline” and the Round Heads. Although ethnically diverse, these groups  both correspond to Europoid ethnicities. They constitute, in the central Sahara, the only rock art groups – and, more generally, the only archeological facts – that allow us to understand a spiritual dimension, a little confusing for us, and difficult to interpret with the present evidence.
The latter term – spiritual dimension – comes, of course, from the culture of the present author, and it would be presumptuous to think that the representations had the same meaning in the “Bubaline” culture or in that of the Round Heads, or even to believe that they denoted some concept of the same nature: the distinction between mind and matter, if familiar to us, is not as clear for all cultures. By the term “spiritual,” we only mean that besides scenes of material cultural, perceived as such in all cultures at least in the primary sense (regardless of a possible secondary, symbolic meaning) – e.g., pastoral scenes, scenes of everyday, a battle, a hunt – there also appear scenes in which it is clear that the meaning is uniquely or eminently symbolic. A meaning “hidden” to us, but present, and not reducible to the preceding scenes: the “great gods” with unusual attributes, the “worshippers” – an attitude not typical in “ordinary” life – men and women with horns, “decorated” rams (the latter only in the Maghreb), persons floating in space, animals struck by a mysterious spiral or bearing a “disk” between the horns, scenes reflecting a special status for the bull, etc. This “spiritual” cultural trait is thus defined as very frequent among the Round Heads, and just frequent in the “Naturalistic Bubaline.” One barely perceives it later in the equally Europoid group of Iheren-Tahilahi, where only a few masks show it. Then it disappears completely in the later Europoid groups, contemporaries of the period of the horse. One very modern word summarizes this cultural process: the secularization of society. Insofar as these various Europoid groups are, in the same place, and despite the vicissitudes of weather, structured by related ideological systems – this is the hypothesis that we advance – we find that in their symbolic universe the “great gods” and the horned goddesses disappear, and the rock art repertoire focuses mainly, first, on pastoral scenes, then, from the Tassilian period of the horse, of the “Herders of Ti-n-Anneuin” and of the stage of the “Libyan warrior,” in the scenes glorifying the warrior, or the chieftain, or the prestigious individual with some title – the charioteer, for example. The evolution is from the collective to the individual, from the era of ideological institutions to those of chiefdoms and “entrepreneurs,” as modern sociology terms them. Our “interpretation” – a pretentious word – would not know, without knowledge of the various symbolic codes that informed the groups involved, how to go beyond these very vague notations.
Let us now take a backward glance, expanding our view through time and space. Bits of the symbolic universe of the central Sahara and Maghreb, that we perceive from the “Humid Neolithic,” around 4000 BC, strike us with a family likeness that they appear to share with similar traits from Egypt, and also, somewhat, from Southwest Asia. There are, for example, the masks and theriomorphs, the “great gods” with animal tails, the horned “goddesses,” the decorated rams preceded by an officiant, “floating” figures that appear among the Round Heads and that we have found among the masked figures, but which are also found in Egyptian astrological compositions such as the syringe of Set I (Yoyotte, 1968, p. 150), etc. Some of these features are also found in Black Africa. Conversely, they are hardly found at all in the art of the Spanish Levant, in the  Italian Bronze Age, the depictions of Val Camonica, Cretan or Mycenaean art, Geometric Greek art, etc. We term “Africanitude” [or “Africanness”] the common character that they define: the term is deliberately vague, but suffices to distinguish this character from that of the Indo-European world, on the other side of the Mediterranean. They were already past zoomorphism, a “primitive” stage of many religions, with purely anthropomorphic deities, relegating zoomorphic beings to a lesser rank, that of “demons,” griffons, sphinx, etc.
We have explained that specific contacts between Egypt and the central Sahara are very unlikely in the middle and late Holocene. This statement has no objection to one of the theses that has emerged in the works of Leclant, Huard and Allard-Huard, that of an archaic substrate, a “common, paleo-African complex,” as referred to by Leclant (1990) in his inaugural article of the first issue of this journal, where he clearly articulates this thesis. From this very ancient substrate, during the millennia of the “West Neolithic,” only traces persist – such as our masks and theriomorphs – of archaic elements, in the form of cultural traits, homologous therefore, and not just homologous, but which had time to differentiate themselves in the various cultures. So much so that we now have some difficulty recognizing the original identify of these traits. Their formal similarities are now reduced to a vague family resemblance, “Africanitude” [or “Africanness”]. The process of their differentiation is the same as for the words of an original language: the latter change, in each of the daughter languages, they diverge and, over time, no longer present anything similar but a few roots that are difficult to recognize.
This comparison with language is not a coincidence. It could be that language and artistic expression are here two symbolic systems reflecting one parallel cultural evolution from a distant original “Africanitude” [or “Africanness”]: that of the Afro-Asiatic linguistic group. This common linguistic “Africanness” is now perceived by linguists from very ancient dates. From 15,000 BC, that is to say, it is from the terminal Pleistocene, that the initial location is postulated – the “homeland” – of the Afro-asiatics, and it would not lie in the Fertile Crescent, as we once thought due to intellectual conformism as well as because we were impregnated with the Biblical tradition, but somewhere in Nubia or the eastern Sahara (Ehret, 1979, p. 163 – 1984, p. 27 – Jungraithmayr, 1989, p. 157). Glossochronology has even allowed the dating, though very approximately, of some Afro-Asiatic cultural vocabularies that would push the dating backward in time of the domestication of animals in Africa to dates at least as old as those of the Middle East, and that domestication would have begun in the eastern Sahara.
This major event could not remain without effect on the social and symbolic structures (Cauvin, 1987). Subsequently, a contraction of linguistic domains, presumably caused by the “Arid Post-Neolithic,” rejected Afro-Asiatic Saharan components toward the south and east (Jungraithmayr, 1989). The Semitic branch would only be one branch among the most recent in this bloc; it would have resulted from a centrifugal movement, which could explain some cultural similarities between Northeast Africa and Southwest Asia. If we agree to link anthropological entities, linguistic entities and cultural entities – such as the iconic expressions of ideology, for example masks and theriomorphs – and especially if we accept this reversal of traditional conceptual schemes “ex Oriente,” linguists provide us a good model to explain our artistic components. First collected into a common archaic substrate, they would  have diverged when populations were fragmented by climate change, and are found, later, in some concrete groups, reduced to a few bits that are still vaguely similar. This is speculation, of course, that the hypothesis of parallel evolution, very old, of linguistic and cultural phenomena, a process that history often belies. But let us keep it in mind, possibly to be supported by documents to come: it would make us well aware of this substrate of “Africanness” common to Egypt and the central Sahara, that we expect to find in both art and language, two fundamental symbolic systems structuring all society.
7 rue J. de Rességuier, 31000 Toulouse et
Laboratoire d’Archéozoologie, 07460
All photographs in this article are by the author.