Skip to navigation – Site map

HomeNuméros8Pastoralist appropriation of land...

Pastoralist appropriation of landscape by means of rock art in Ennedi Highlands, Chad

L’art rupestre comme moyen d’appropriation du paysage par les pasteurs du massif de l’Ennedi au Tchad
Tilman Lenssen-Erz
p. 27-43

Abstracts

Research into the Saharan rock art has from the very first day been linked to the question of how the people of the past managed to live in this environment. It was about the enigma of climatic conditions and cattle herders who obviously once inhabited regions that meanwhile have become uninhabitable desert. The first rock picture discovered by a European researcher was the “Apollon Garamante” showing two masked persons on the side of cattle (Barth 1857: 210). This made the discoverer Heinrich Barth wonder about the possibilities of keeping cattle in this region, the Libyan Messak Settafet, and the necessarily more advantageous former climate (ibid.: 215-218). At this juncture, it is worth noting that there is ubiquity of cattle in Saharan rock art on the one hand, but, on the other hand, that the “green Sahara” never was anything like the deep green meadows of Europe. Accordingly this land never sustained permanent habitation that could settle in a certain place for centuries (except for the oases), even if the economic system would not have necessitated mobility such as in a hunter-gatherer economy. Therefore this paper seeks to elucidate the hypothesis that the appropriation of the land and the advertisement of identities by prehistoric Saharan pastoralists of the Ennedi Highlands were not effectuated by settlements and built structures but through rock art in various expressive forms.

Top of page

Full text

Introduction

1The Ennedi Highlands are located in the east of the Republic of Chad on the southern margin of the Sahara Desert (map, fig. 1). Its remote location as well as political insecurity lasting over decades precluded it from intense research activities. Habitation of the region until the present day is in the long tradition of pastoralists herding various domesticates, i.e. camels, cattle, goats and sheep with some horses on the side. While the sandstone highland provides relatively rich vegetation and natural water reservoirs which regionally enable semi-permanent settlement, vast regions in the north and east of the highlands can only be used by people in years of advantageous seasonal conditions. The picturesque precipitous western flank is even more favourable compared to the less spectacular central and eastern parts, due to two permanent water holes.

Figure 1 – Satellite picture of the Ennedi Highlands

Figure 1 – Satellite picture of the Ennedi Highlands

Scale bare: 100 km

© Heinrich Barth Institut MRAC

2A salient feature of the entire highlands is the ubiquitous rock art from the last five millennia. Rock art in the Ennedi Highlands belongs to the least studied art traditions in the Sahara and beyond. Only around the northwestern tip was intensive research carried out by G. Bailloud in the late 1950’s (Bailloud 1997). After the first known exploration of this region by French colonial troops in 1914 there has been no mention of rock art (Tilho 1920) although it is hardly conceivable that these first European visitors might have overlooked the pictures. The same pertains to Hassanein Bey (1924) who passed through the mountains on a route coming from northeast. Franz Kollmannsperger (1957) found very few engravings and paintings in the western interior but did not report any archaeological find while Peter Fuchs, who also negotiated the highlands, reported rock art mainly from the better accessible outskirts of the highlands and only three painted sites from the central highland parts (Fuchs 1957: 133). While Fuchs only noted painted sites in the central highlands, Théodore Monod (1968), who crossed the highlands on an ENE course from Fada, only described engraving sites. More frequent reports on rock art of the highlands surfaced after the country had become politically somewhat more stable in the 1990’s (e.g. Negro et al. 1996; Jacquet 2000; Gauthier & Gauthier 2006).

3Apart from this, before 2000 there were only short reports on rock art sites such as Niola Doa in the north of the highlands, where magnificent larger than life engravings of women can be found (Breuil 1956; Simonis et al. 1994; Fuchs 1957; Faleschini et al. 1996). Other scattered sites along the outer rim of the highlands were reported by travellers.

4Due to so little previous research it was an open question whether the central parts of the highlands had a density of rock art comparable to other Saharan highlands. Therefore a research project of the University of Cologne within the research scheme ACACIA (Lenssen-Erz 2007) focussed on the study of the art of representative sections of the landscape, combining it with archaeological excavations. These regions were intensively surveyed in order to systematically record all archaeological finds. As a result some “windows” were opened allowing a good overview of the character and distribution of rock art across the highlands.

5Significant detailed investigations relating to the climatic development in the Holocene were also part of the ACACIA research, based on coring of lake sediments in the Ounianga Lakes 250 km northwest of the highlands. Preliminary results can be summarised as follows (Kröpelin et al. 2008) while bearing in mind that due to the Ennedi’s location somewhat further south the timing and absolute figures of the climatic conditions may vary systematically from the core results: at ca. 6000 calBP the region was a savannah with annual precipitation of ~250 mm and representing a Sahelian climatic zone (Andigue et al. 2006: 13); this was a period when the climatic optimum was already past and aridification prevailed (Kuper & Kröpelin 2006). By 4300 calBP precipitation went down to ~150 mm per year, nourishing an open grass savannah with a modest population of trees which today are found 300 km further south. Roughly around 2700 calBP the present-day hyperarid status with ~50 mm annual precipitation was reached. This implies that the Ennedi Highlands would since that time be part of the Saharo-Sahelian climate zone (Andigue et al. 2006), receiving between 100 and 200 mm of annual precipitation.

6Today this means that people herd camels throughout the area and to a lesser extent also goats and sheep, paying occasional visits to the permanent gueltas (waterholes) at Archeï and Bechikelé in the west, or Omena in the southeast. Further gueltas only in advantageous years provide water, such as Koinamena or Oura in the northeast. The only absolutely reliable waterholes are the two in the west, evinced by the fact that in Archeï a small but ancient population of crocodiles has survived since the huge Megatchad-Lake has shrunk after 6000 BP (Sepulchure et al. 2008) making it impossible for the crocodiles to later move in from another habitat.

7In the Archeï region still today horses can be found occasionally while cattle may be encountered at Bechikelé some 80 km further south (being the northernmost place to encounter cattle in Chad). Judging from the distribution of rock art motifs one may conclude that at times cattle were kept all over the highlands which later was not possible at any time with horses whose pictures are extremely rare beyond the western foot area. Questioning dwellers of the Ennedi today about the possibilities of pasturing cattle in the central and eastern regions (where masses of rock pictures of cattle can be seen), they deny this categorically even for years with good rainfall. And they assert that this was impossible also for the generations before them.

8The archaeological background of rock art production needs to take account of the cultural development during the Holocene. Despite the fact that besides some rock art spotting only very few systematic archaeological studies were carried out (Bailloud 1969, 1997; Lenssen-Erz & von Czerniewicz 2005; Lenssen-Erz 2007; Keding et al. 2007) there is a rough understanding of the settlement history of the highlands. Hunter-gatherer’s presence can be determined only in few scattered locations, identified by the Dotted-Wavy-Line pottery (e.g. Keding 1997), indicating sparse hunter-gatherer populations from early to mid Holocene a fact that makes one wonder, since this was the period of best climatic conditions and should have facilitated comparatively carefree foraging.

9First pastoralists become archaeologically visible around 3000 BCE when the decrease of precipitation had already begun. Based on the data of S. Kröpelin et al. (2008) and hypothesizing that the southward increase of annual precipitation roughly parallels today’s isohyets (Andigue et al. 2006: 13) then the Ennedi Highlands should have received around 250 mm of precipitation in this phase, which today prevails south of the Ennedi where cattle herding is current. This first occupation period seems to have established or may have been enabled through cultural links eastwards to the Wadi Howar region (Keding et al. 2007; Jesse et al. in press), but also with the NNE to Jebel Ouenat. These links become visible through parallels and similarities in pottery and, concerning the Jebel Ouenat, in rock art.

10Among the pottery finds there is a notable presence of Hohou and geometric decorated pottery related to the Leiterband/Halbmondleiterband and Handessi A pottery respectively, found in the Wadi Howar in the east (Keding 1997: 179-183; Keding et al. 2007: 36-37). At roughly 1500 BCE these eastward links vanished due to the fact that the Wadi Howar region became uninhabitable (e.g. Jesse 2003: 53-57; Jesse et al. in press). The connection to the Jebel Ouenat may have been kept up much longer judging by the fact that Hassanein Bey (1924) found a guide who could pilot him from there to and partly through the Ennedi Highlands. But this connection may have had not much intensity since the typical rock art motifs of the Iron Age abounding in the Ennedi are lacking at Jebel Ouenat.

11Before the next cultural innovations reached the Ennedi there is a decrease of archaeological signals in the region (between 1200 BCE and 200 CE) so that a hiatus in occupation may be considered for the first millennium BCE (Keding et al. 2007: 42; Jesse et al. in press), correlating with a discontinuous occupation of the West African Sahel (Breunig & Neumann 2002). It is during this period that in other parts of northern Africa horses and later camels were introduced but currently it is assumed that they reached the Ennedi Highlands only in the 1st century CE perhaps synchronically (Bailloud 1997: 15) and thus differ from other Saharan regions where the introduction of horses antedates that of camels (Clutton-Brock 1997: 421; Le Quellec 1998: 119-121).

12This innovation of new domestic animals roughly coincided with the introduction of iron technology to the region which seems to have started at ca. 200 CE, the time to which first iron artefacts are dated. The full productive iron technology was clearly established at 400 CE, being archaeologically visible in furnaces and tuyeres (Keding et al. 2007: 37). At this time cultural links are directed to inhabited regions in the west (Borkou/Djourab) and northwest (Tibesti), indicated by similarities and parallels in pottery (after Treinen-Claustre 1982) and in rock art.

Data base established through Acacia research

13From 2003 to 2005 two field research campaigns of several weeks and another one of approximately ten days were carried out (Lenssen-Erz 2007). Thus three landscape units (mainly distinct valley systems) were explored rather completely and another one partially, with seven excavations, long survey walks, systematic documentation of rock art and some short interviews with local dwellers.

14The research for conventional archaeological sites sums up to 262 new finds, seven of which were excavated while the rest was filed according to the established ACACIA surveying system. Judging from the find material, 51% of the sites were habitation sites while 32% were unstructured artefact scatters. 10% of the observed phenomena obviously were tombs (cairns), none of which has been excavated however. Pottery as the most reliable and best understood archaeological proxy for the wider region through the works of B. Keding (e.g. 1997) and F. Jesse (2003) in the Wadi Howar, allows a rough assessment of the chronological status of the sites discovered (159 archaeological sites, largely without rock art, were classified based on the most frequent pottery, data from Keding pers. comm.):

  • 1% yielded Dotted Wavy Line pointing at the early Neolithic of 5000-4000 BCE ;

  • 21% had pottery of Hohou type (equals Leiterband in Wadi Howar) and indicate the middle to late Neolithic, c. 4000-1500 BCE ;

  • at 24% of the sites pottery with geometric decoration was found (Handessi A in Wadi Howar), which represents the phase of transition from Neolithic to Iron Age at about 2200-1200 BCE ;

  • 26% showed Djoki type ceramics that belongs to the Iron Age 200-500 CE.

15The following part of this paper will concentrate on the rock art findings and investigate the relation of the art to the landscape. Rock art at all of the listed sites was recorded by systematic photography of pictures and sites, by specific forms for the motifs, forms for specifications of engravings and forms for the landscape context. In some sites a detailed analysis of every figure was carried out following the paradigm of the textual rock art archaeology (Lenssen-Erz 2001).

Rock art and the Longue durée

16The Ennedi Highlands are an area where rock art production seems to have a legacy going back 5000 years from today. Of course it would be inadequate to speak of an ongoing tradition without interruptions the possible hiatus of occupation between 1200 BCE and 200 CE has already been mentioned and also in the more recent past there may be more interruptions that still need to be identified. But any one period in history since the Neolithic could without cogent contradiction be represented by some rock art, up to the present as can be seen in the engraving of a man with heels and an automatic gun near the waterhole Oura (fig. 2).

Figure 2 – Engraving next to the waterhole Oura in the northeast of the Ennedi shows a man with heeled shoes and an automatic rifle on his side, the animal is probably a dog so that this engraving may refer to hunting activities. It is proof that even in our times people produce engravings although context and meaning have changed since prehistory

Figure 2 – Engraving next to the waterhole Oura in the northeast of the Ennedi shows a man with heeled shoes and an automatic rifle on his side, the animal is probably a dog so that this engraving may refer to hunting activities. It is proof that even in our times people produce engravings although context and meaning have changed since prehistory

Height of person ca. 15 cm

© A. Pastoors

17During research in the southern central area of the Highlands a young nomad was encountered near a rock art site at which household gear was deposited. This young man asserted that some of the simple geometric signs found at this site besides engravings and paintings of camels and people on camels he made himself with his brother (fig. 3). The sign he made was the branding mark for the camels of his family and he insisted that the signs at the site he made for no purpose at all not even to mark this place as the one where his family deposited their household gear. This meets with interviews Russell has led with young Turkana pastoralists in northern Kenya, who asserted that they pecked branding marks into rock surfaces as a pastime (Russell 2012: 8-9). This statement that there is no deeper sense behind these engravings is corroborated by the opinion of older Turkana who maintain that the marks make no sense at all on inanimate things. On the camels of the Ennedi of today such signs serve the purpose of marking one’s property (Meike Meerpohl pers. comm.; see also Russell 2012: 18-19 for ownership marking among the Rendille of northern Kenya), but as Russell points out for the Turkana (ibid.: 4-6), also other meanings can be communicated by branding marks, such as warding off illness but especially to advertise clan affiliation, not ownership. This kind of sign in the Ennedi engravings, usually looking barely patinated, is found very loosely scattered over the regions where engravings occur. The engraved signs do not have an equally rich spectrum of varieties as do the marks on camels but this could mean that the area, where the engraved signs can be found, was temporarily used be people from restricted numbers of clans (or similar social communities).

Figure 3 – This young nomad by the name of Mahamat, living in central Ennedi Highlands, asserted that the engraving marked here with a circle was done by himself, being the sign with which his family marks their camels

Figure 3 – This young nomad by the name of Mahamat, living in central Ennedi Highlands, asserted that the engraving marked here with a circle was done by himself, being the sign with which his family marks their camels

18Despite the claim of the local nomad of having produced a geometric rock art sign only very recently this is no proof that all signs apparently derived from potential camel branding marks are of modern origin. They may instead be of unknown antiquity since they seem to be in use for very long. During her research in eastern Chad, Meike Meerpohl (2007) was told by local Zaghawa that they understand their provenance as originating from Syria since there the same camel branding marks are in use as the Zaghawa use today (Meerpohl in litt. 2007). The insecurity in view of the antiquity of motifs extends to pictures of camels, too, since there are only few and unreliable features that are time sensitive: styles are not helpful because engravings may appear in any stage of stylization obviously in any period and patination differs widely dependent on micro-spatial conditions. It appears, however, that fine line paintings of camels were only made during the Iron Age but not in later, more recent periods. Obviously the process of producing rock art did not require the signal of high aesthetic elaboration (Lenssen-Erz 2012) in the younger periods so that a relatively complex technique like painting was skipped while engravings which can be made by incision with any hard and pointed tool or still more simply pecked with any hard rock could be realised any time without much preparation. This finds a parallel in observations made by Russell among northern Kenyan pastoralists where camels are being treated and symbolised with much less “infatuation” than cattle (Russell 2012: 19).

19Another kind of motif by which the “longue durée” (after Braudel 1958) of rock art, i.e. its ongoing practice into recent times, can be determined are lybico-berber inscriptions (Tifinagh). But while these are current in more westerly regions of the Sahara, only one such typical incised inscription was found in the Ennedi, i.e. in the southern central Highlands obviously left by a traveller. During his research in the Tassili n’Ajjer, Issak Oukafi Cheikh (2012) not surprisingly found many such inscriptions but he also encountered an old Tuareg lady who pointed out a site where she had incised Tifinagh inscriptions when she was young. As is typical for other such inscriptions, her message only contained her name and greetings to passers by (Oukafi Cheikh 2012: 63-64). From the recount of the lady it transpires that she has an admiration for the old painted art (obviously an aesthetic appreciation) but she would herself only produce inscriptions, not being picky about the material with which to execute them. She also points out the correlation of rock art sites in shelters and the beneficial setting for habitation, configurations which the people of the region still seek today. A parallel to this was observed in the Ennedi Highlands where storages of household gear were only found in rock art sites probably chiefly because they occupy the best shelters (as is asserted by the Tuareg lady). This anecdotal observation can be supported empirically (see below).

20These examples show that there are modern repercussions of the old art practice even if they are rather rare and meanings will probably have changed. Nevertheless this specific type of communication retains a cogent link between signification and landscape simply through the immobility and durability of the signs which are the same irrespective of the epoch in which signs are produced. Therefore the cognitive and subconscious associations of the artists/producers, concerning the art’s immobility and durability, may have remained rather similar through time.

Spatial connotations of rock art

21Surveys of the rock art in the Ennedi Highlands (Lenssen-Erz & von Czerniewicz 2005; Lenssen-Erz 2007) have revealed a prima vista contradictory distributional pattern: generally, negotiating the Highlands is not fraught with special difficulties; wide stretches can be explored with motor vehicles which normally find a way to proceed even though there are no routes or tracks at all in the entire highlands. Only the western precipitous cliffs constitute an obstacle for motor cars, for which there is only one feasible descent (but which is impossible to ascent), but on foot or with animals there are many options to leave or enter the Highlands on its western side, north, east and south are unproblematic anyway. The general ease of mobility and the consequential ease of exchange between areas are contrasted by a certain compartmentalization of the landscape becoming discernible in the distribution of certain rock art motifs. These motifs are very distinct and they are to be found only in narrowly restricted sections of the landscape (table 1).

Table 1 – Regional particularities

Motif

Description

Exclusive area

© Lenssen-Erz

horses in «flying gallop»

western foot region

© Lenssen-Erz

village scene

western foot region

© Lenssen-Erz

riders on fantasy «horses»

western foot region

© Lenssen-Erz

camels in full gallop

western foot region

© Lenssen-Erz

emphasis of the navel of cattle (shoulder height 100 cm)

northeast

© Lenssen-Erz

life size cattle engravings

small area in the northeast

© Lenssen-Erz

figures en miniature (animals and people)

one valley in the northeast

© Lenssen-Erz

engravings with an «aura» (shoulder height 100 cm)

few valleys in the northeast

© Lenssen-Erz

life size decorated women (© G. Jacquet)

one valley on northern fringe

22Some of the motifs listed in table 1 are well known from other parts of the Sahara, such as the horses in “flying gallop” or the “village scenes”, both of which are current motifs in regions west of the Ennedi. But their distribution here stops at the foot of the high cliffs that border the Highlands in the west. This is the region of the rather intense research by G. Bailloud (1997) and even though the area of his research measures only some 60x80 km (ibid.: 20) he established a complex differentiation of stylistic groups (ibid.: 13-19) beginning with the Archaic period (before the introduction of cattle) in which he distinguishes five styles. In the following Bovine Period he determines a succession of four chronological sub-periods which again are subdivided in six styles while the concluding Dromedary Period it is divided in three styles. Even if conceding the rather long chronological sequence and an intuitive identification of styles by G. Bailloud, one may accept that his distinctions grasp morphological entities with a spatial restriction. Whether they can indeed be accepted as proper styles (whatever this is) or not, the differentiation of 14 rather distinct phenomenal groups in a very small section of the landscape may be a strong indication of rather clear manifestations of social units marking and fencing off their identity with rock art.

23While G. Bailloud focussed only on styles, the distribution of motifs as listed in table 1 corroborates his interpretation of a rather fine-grained distribution of different social entities. One may hypothesizes that behind the distribution of some motifs ecological reasons may be at work, such as that horses and larger settlements (“villages”) were not possible to sustain in the inner Highlands due to the lack of water. However, while the horses can indeed only be attributed to the 1st millennium CE when the full aridity of today prevailed in the area, the “village scenes” can, in confirmation of Bailloud’s analyses, be attributed to the (final) Bovine Period which, according to the archaeological record, seems to have lasted until 1200 BCE. This is a time when precipitation had declined but was not as low as today so that the lack of “village scenes” in the inner Highlands may not necessarily be ecology-driven.

24Other motifs, however, have no ecological bias since their idiosyncratic character is based on stylistic or aesthetic features (Lenssen-Erz 2012), such as specific sizes (either particularly big as in some engravings, or particularly small as in some paintings) or elaborations on the navel or body decoration. In a small area northeast of Hajjer Mornu a specific engraving technique was applied which gives the figures some kind of glooming aura. After the motif (mainly cattle) had been incised or pecked carefully into the dark crust of the sandstone, on both sides of the clear line the surface was rubbed (probably with a stone) to light up the dark patinated surface and make the sharp line blurred. This technique is not particularly sophisticated or laborious nevertheless it was not used anywhere beyond this small region.

25Similarly, in a small valley on the northern fringe of the highlands, in and around a place called Niola Doa, there are several instances of almost life size engraved women with extensive interior linear ornamentation (e.g. Fuchs 1957; Faleschini et al. 1996). They also display very specific objects and clothing details, all displayed in a sophisticated and elaborate manner that one might expect to find in a well established, far distributed style, but here this complex signature is confined to few square kilometres in a single wadi (Simonis et al. 1998).

Lifeworld as a spatial concept and rock art

26The restriction of the distribution of such idiosyncrasies to small compartments of the landscape can be linked to certain patterns of behaviour in the use of a landscape. It appears that in the Ennedi the spatial entities harbouring an idiosyncratic rock art motif preferably vary around a certain size: in the inner parts of the highlands single valleys or valley systems provide clear spatial entities. In the western foot region only the cliff side is well defined while all other sides are open land so that the extent is not naturally confined except that at a point not very far from the cliffs there are no more rocky outcrops. At any rate, all these areas measure largely only so few kilometres across that they can be easily crossed on foot within a day or less and they seem to be the arenas for playing out one’s identity manifestations.

27These spatial units seem to fit to that part of a landscape that may be used by a (even temporarily) settled group during one day, i.e. it is their daytime lifeworld (table 2 for a very brief outline of this lifeworld concept, based on Schütz & Luckmann 1975). As opposed to the seasonal (or annual) and the night-time lifeworlds, it is a part of the landscape that contains all basic resources which are needed during a normal day. This largely relates to the basic needs of food and water but also to fuel and raw materials that are needed and used up in high frequency. Sharing of the daytime lifeworld with others is likely but it would most probably require approval and would underlie quantitative limitations due to restriction of resources.

Table 2 – A brief outline of the concept of three lifeworlds as different spatial entities of use (for an elaboration of the concept cf. Lenssen-Erz in press)

LIFEWORLD

Annual/saisonal lifeworld

> Everyday lifeworld

> Night time lifeworld

Time

A whole year or season (e.g.rainy season, dry season)

The bright day from dusk to dawn

Night time with the lowest personally motivated mobility

Space

That which is being utilised within a year or season

That which is being utilised within one day (to and fro)

That which is needed for the nocturnal basic needs (consummation, hygiene, shelter)

Most frequent activity

Exploitation of the full range of resources, even those of sporadic use

Procurement and handling of all things necessary for everyday life

Processing and consummation of everyday provisions, satisfaction of basic needs

Ritual/religious use

Possible

Likely

Very likely

Archaeological correlation

Macro-region, ‘habitat’

Meso-region (various definitions and practices)

Micro-region, site

28The seasonal/annual lifeworld harbours not only several daytime lifeworlds but also further resources, such as rare or seldom used raw materials or only seasonally available faunal or floral assets. It may be shared with strangers who may have access to resources even without previous coordination or consent.

29The night-time lifeworld (camp, house, village etc.) is not resource oriented because it is the place of the treatment and consummation of the resources. As a “home” it is the safest place people have. It is a clearly defined space and access to it is only with the consent or upon invitation of the principal “owner”. If necessary this place is most decidedly defended.

30A comparison with other regions worldwide has shown (Sauvet et al. 2006) that the restricted region of the Ennedi is as divert in its motif-frequency correlated with distribution as much larger areas like southwestern France and the Pyrenees in the Late Pleistocene or northern Australia in the Holocene. One may therefore conclude that under the arid climatic conditions, and the consequential relatively sparse population, the Ennedi saw rather high social dynamics that stirred pictorial manifestations of various identities. This implies that rock art is understood here as a means to express identity in order to make it operational in social interaction (Domingo Sanz 2008: 125; Gallardo & De Souza 2008: 82; Lenssen-Erz 2008).

31As a consequence from the distribution of idiosyncratic rock pictures in daytime lifeworld landscape entities, and the therein implied manifestations of localised identities, it emerges that these groups would have had to cooperate in order to get access to resources not available in their daytime lifeworld. At first sight this assumption would postulate that a number of idiosyncratic motifs were contemporaneous. But this is not necessarily so because the spatial restriction of a specific motif, lets say “cow with aura”, in itself points out that these people did not opt for using (or rather marking) wider stretches of landscape most likely because other parts were occupied by other people who may have or have not produced rock art. Thus cooperation would have been the most liable option of interaction since the terrain of the Highlands is not conducive for an aggressive or conflicting strategy with its wide open plains and low ridges that can hardly hide anything.

The painting-engraving dichotomy

32As has been pointed out above, the relatively fine-grained separation of identity groups is based on a number of stylistic elements and idiosyncratic motifs. These are specifications within a foundation based on a far reaching accord about motifs where the main domestic animals dominate; this indicates a common cultural consensus that temporarily obviously channelled into a cattle centred ideology which in turn was informed by the economic strategy. This common base may also be indicated in the uniformity of pottery that can be discerned throughout the highlands and even extends further east (Keding et al. 2007; Jesse et al. in press). Besides style and idiosyncrasies constituting two levels of a fine “granularity” distinguishing small landscape compartments on the one hand, and the overarching economic strategy of pastoralism prevailing countrywide on the other hand, there seems to have existed an intermediary level of cultural distinction which is expressed by the implementation of the two basic art techniques, namely painting and engraving. Only in part are they restricted to certain areas, like engravings being almost absent from the western forelands at the foot of the falaise. This apart, both techniques can be found mixed in many areas of the Highlands. But looking at the details, a number of discriminating characteristics become evident that point at different context and meaning of both techniques, notwithstanding the use of the same motifs.

33Basic data on painting:

  • 5795 figures in 104 sites => 55,7 figures per site

  • Clear local concentration besides a wide distribution of sites throughout the Highlands

  • 35% open presentation of pictures

  • 34% in isolated locations

  • 91% of the paintings are in shelters which provide options for habitation

  • 24% at the foot of elevations, 1% waymarks

  • No dominant pattern in how painted sites are linked to elements of the landscape

  • Paintings show animals 52%, humans 31% and signs etc. 17%

  • Painting was the preferred technique to depict goats (100%), humans (96%, i.e. only 4% of the human figures are engraved), horses (95%), cattle (92%) and giraffe (54% listed here are only motifs with a frequency >0.5% in the whole corpus)

34Basic data on engravings:

  • 773 figures in 54 sites => 14,3 figures per site (a few sites have paintings and engravings)

  • Loose scattering over restricted parts of the landscape, almost without concentrations

  • 70% open presentation of the pictures

  • 54% in isolated locations

  • 31% in roofed shelters => restricted options for habitation

  • 43% at the foot of elevations, 11% waymarks

  • => a clear patterning in the linkage to elements of the landscape Engravings show animals 66%, signs 27 % and humans 7%

  • Engraving was the preferred technique to depict ostrich (79%) and camels (58% listed here are only motifs with a frequency >0.5% in the whole corpus)

35It emerges from differences in these data that painting and engraving are not technical options to be chosen randomly according to availability of means (such as postulated e.g. by Lewis-Williams & Loubser 1986: 259) but quite obviously each technique had its own complex practice and they served different ends. Based on these differences some hypotheses concerning the appropriation of the land by means of rock art can be forwarded, first in view of the paintings:

36Paintings are often narrative, containing many scenes, the recognition of which being facilitated by the relatively large number of human figures. The communication that was practiced seems to have drawn upon metaphoric contents which directly linked to everyday experiences. This way of communicating evoked the lifeworld as a world of shared, quasi empirical experiences.

37Places and with them the landscape were marked in the context of secular use which is clearly determinable since rock painting sites are usually habitation sites. Accordingly the marking by means of paintings took place in the everyday lifeworld, painting activities were not separated out into some inaccessible or tabooed sacred spaces or ethereal contexts. Therefore, when choosing a site, the painters had no special symbolism or sacredness of the rock formation in mind, nor did they look for particularly suited canvasses, but most important was the suitability of the place for habitation.

38The appropriation of a site then was mainly by using it for habitation and related activities, only subsequently was the appropriation completed by the production of rock paintings, but rock art was not the main token of the appropriation. Painting may have been part of particular ritual behaviour, but it was embedded in daily doings. The entire space for the rock art sites was the daytime lifeworld while each site had the capacity and attributes of a night time lifeworld.

39This entire process was an “active” appropriation of the landscape because people imposed their patterns of use on the land and used the sites not only to deposit household gear but also to “deposit”, as it were, the insignia of their livelihood and wealth in pictures of their cattle.

40The characteristics of engravings and their distribution instigate other conclusions since engravings are non-narrative, with symbolic meaning but hardly any scenes (largely due to the lack of humans). Engravings do not directly relate to the concrete lifeworld, rather they seem to express a kind of abstract religiousness like a “worshipping” of the land. This land primarily seems to be defined by sacred places which become signified with rock art motifs. The motifs in themselves are largely derived from the everyday but they are exempted from the mundane context. Engravings seem to be the results of special ritual activities having been prepared only for these peculiar places. Apparently the whole landscape was seen as symbolically charged and specific configurations of the landscape were meaningful or sacred independent of human use. Such places of symbolic significance existed independent of rock art and the prevailing significance was only made visible or emphasised through engravings. Obviously few pictures and often just a single one sufficed to achieve this emphasis. They marked the sacred landscape beyond the everyday and even if it took place in a daytime lifeworld, the places would rest outside the secular routine so that they hardly ever would be suited to become the night time lifeworld, i.e. ordinary habitation sites.

41This entire process constitutes a “passive” appropriation of the landscape since the engravers entered and used a landscape containing “ready-made” sacred places that they would endow with their marks due to the places’ pre-existing significance. Their art production was not an active way of Gestaltung but a passive way of “subjugation” under a given sacredness to which the engravings meant a kind of service.

42The fact that in ten sites of the corpus paintings and engravings turn up together is not in contradiction with the hypothesis of two ways of appropriation. In view of the long period of time in which these principles seemingly were operational a diachronic sequence would be no surprise with territorial bounds shifting over time. But even a synchronic use of parts of the landscape could have taken place, thus rather corroborating the hypothesised cooperative behaviour than refuting it.

43The hypotheses about an active versus a passive appropriation of the landscape can be extended into a picture of how the specific practice of art production links to the identities of the different groups and which function art fulfilled for them.

44Apparently for the “painters” the landscape was the stage on which they advertised their predilection for cattle (63% of the depictions of animals are cattle). In as much as the whole livelihood was revolving around the main stock also the whole thinking was dominated by it. The symbolic capital which cattle represented was “deposited” in the landscape at the most common places of their lifeworld. In this process the pictures of cattle were self-referential (not meaning and symbolizing anything else than cattle in its complex meaning), inscribing the significance and value of cattle into the landscape and thus making them to meta-symbolic capital: if cattle, beside their role for livelihood, were also symbolic capital, then a picture of a cow would be a symbol for symbolic capital, i.e. two symbolic transcriptions away from the real capital (which would have been made real if, e.g., a cow was bartered for some other goods). This use of meta-symbolism of cattle constituted a feedback loop enhancing and reassuring the significance of cattle in everyday situations since this is where the pictures were encountered. It served the end to articulate the identity of the prosperous cattle herder and by this also playing with or trying to manipulate secular assets such as wealth, status and power (comparable to what T. Russell [2012: 21] describes for the Rendille camel keepers).

45For the “engravers”, by contrast, cattle are the primary means to express “the extraordinary” that is set apart, locally and semantically, from the mundane everyday. Engravings of cattle are like donations to the sacred landscape, being detached from secular use of the land and its resources. Therefore engraved cattle were probably rather linked to religiousness and metaphysic symbolism than to status and wealth. That such views were likely embedded in religious ideology may be corroborated by the fact that certain practices concerning the expressive repertoire and styles of engravings remained stable over millennia (the raw data and characteristics listed above have no potential for chronological distinction). Concomitant with the postulated “passive” appropriation of the landscape, the “engravers” seem to have expressed with their art their identity as humble beneficiaries of a sacred landscape. Engravings (at least those of cattle) were rarely used to express secular power relations and other social values, even though the “engravers” used the landscape in no other way than the “painters”, with whom they shared the identical economic base under identical ecological conditions. Nevertheless, the co-existence of the two techniques may be interpreted in terms of two clearly distinct groups.

46T. Russell (2012) has distinguished two different principles reigning the marking of stock among Nilotic and Cushitic pastoralists in northern Kenya, where the Cushite Rendille brand their camels for ownership, while the Nilote Turkana express belonging (to a clan) with branding marks. The cutting into the skin of an animal “allows humans to speak to the spirit world” (ibid.: 9) which distances the animal from its economic function and conceives it as a means of religious communication, analogous to the role assigned here to the engraved cattle. However, a plain transfer of recent customs to prehistoric rock art in the Ennedi would be highly speculative, also because among the Turkana marks make only sense on living being and they loose this sense if they are applied to inanimate matter (ibid.). Moreover, T. Russell asserts that the differences of marking customs may be as diagnostic on the level of which type of animal is being herded as on the level of linguistic affiliation (ibid.: 21), i.e., the differences may be an economic or an ethnic marker.

47A competing hypothesis about the authorship of paintings and engravings could be set up by conceiving two distinct groups within an otherwise closed society. Without further knowledge about this society one can be sure that there were young and older people as well as men and women. Both contrast pairs constitute meaningful differentiations in most societies and entail rules of different behaviour as would be required to explain the differences between paintings and engravings. This however would mean that across the entire highlands one specific language or ethnic group would have occupied all areas, keeping certain ageor gender-specific behaviours constant over long periods, even perpetuating them into times with technological innovations (iron production). This scenario I consider rather unlikely since it is based on an assumption of little or no societal dynamics through time because the behavioural patterns of the groups would have remained stable over long periods. By contrast, the hypothesis of even linguistically different groups would more plausibly locate the distinct behaviours on the broader societal level of religiousness and identity as a group with a specific ideology that is not easily changed.

Conclusion

48Paintings and engravings epitomize different idioms of pictorial communication not only based on the difference that one is an additive technique and the other a subtractive, but also based on the differences expounded above. They also entail different conceptualizations of the human-landscape relationship, expressed in an “active” versus a “passive” appropriation of the land by means of rock art. The different kinds of landscape appropriation are based on rather far reaching cognitive conceptualisations that may be too wide apart to be compatible within a compact society.

49In view of such critical divergences I dare hypothesize that “Painters” and “Engravers” were different people to an extent that they may have spoken different dialects or even different languages. Nevertheless it is obvious that they co-existed largely peacefully since they not only shared the scarce resources of the Highlands, but they also lived so very close to each other that larger stretches of the landscape, those of the seasonal or annual lifeworld, had to be shared too.

50The shared pastoral economy was a uniting as well as a dividing factor: on the one hand occasional large meetings between the different pastoral groups may have taken place in order to barter various goods and to get into contact with exogamous marriage partners. This could explain the far-flung spread of certain traits such as homogenous pottery or stylistic commonalities in rock art. On the other hand more or less constant neighbourhood despite (semi-)nomadic or transhumant way of life as well as cooperation or competition for scarce resources may have advanced efforts for the advertisement of distinct identities which defined specific “home areas”. This however would not only have been functional at the level of groups as distinct as Painters and Engraves, but also within these groups where the idiosyncrasies mentioned above define very restricted “home areas” of few square kilometres that under the prevailing ecological conditions may have sustained nor more than an extended core family.

51While the ecological conditions seem to have imposed a narrow choice of economic strategies it was perhaps the growing scarcity of resources which favoured a parallel development of cooperative strategies and differentiating symbolism channelling into cultural multi-vocality.

I thank Sylvie Amblard-Pison for smooth cooperation in editing this paper and three anonymous reviewers for helpful comments on it. Research presented here was enabled by funding of the Deutsche Forschungsgemeinschaft in the SFB 389 ACACIA program, and the cooperation facilitated by the CNAR, Ndjamena, particularly by the director Dr. Baba Mallah and research officer Mahamat Hamdo; the entire project benefited from Dr. Stefan Kröpelin’s engagement and many colleagues of the African Archaeology, University of Cologne.

Top of page

Bibliography

Andigue P. J., Some B. & Alkhalil A. (2006) – Climat. In: D. B. Yahmed (dir.), Atlas du Tchad. Les Éditions J.A., Paris, p. 12-13.

Bailloud G. (1969) – L’évolution des styles céramiques en Ennedi (République du Tchad). In : Mémoires I, Actes du Premier Colloque International d’Archéologie Africaine, Fort Lamy (Rép. Du Tchad) 1966. Institut National Tchadien pour les Sciences Humaines, Fort Lamy, p. 31-45.

Bailloud G. (1997) – Art rupestre en Ennedi. Éditions Sépia, Saint-Maur.

Barth H. (1857) – Reisen und Entdeckungen in Nordund Central-Afrika in den Jahren 1849 bis 1855. Justus Perthes, Gotha.

Braudel F. 1958. – Histoire et Sciences sociales : La longue durée. Annales. Économies, Sociétés, Civilisations, 13 (4), p. 725-753.

Breunig P. & Neumann K. (2002) Continuity or Discontinuity? The 1st Millennium BC-Crisis in West African Prehistory. In: Jennerstrasse 8 (eds), 2002: Tides of the Desert Gezeiten der Wüste. Contributions to the Archaeology and Environmental History of Africa in Honour of Rudolph Kuper, Africa Praehistorica 14. Heinrich-Barth-Institut, Köln, p. 277-293.

Breuil H. (1956) – Le massif montagneux de l’Ennedi (territoire du Tchad) et les danseuses préhistoriques de Niala Doa, découvertes par les géologues André Bonnet, Jean-Michel Freulon, Albert F. de Lapparent et Pierre Vincent (description de la roche gravée). Cahiers d’art, Juin, p. 7-16.

Clutton-Brock J. (1997) – Animal domestication in Africa. In: J.O. Vogel (ed.), Encyclopedia of Precolonial Africa. Altamira Press, Walnut Creek, p. 418-424.

Domingo Sanz I. (2008) – From the Form to the Artists: Changing Identities in Lavantine Rock Art (Spain). In: I. Domingo Sanz, D. Fiore & S.K. May (eds), Archaeologies of Art: Time, Place, Identity. Left Coast Press, Walnut Creek, p. 99-129.

Faleschini G., Negro G. & Simonis R. (1996) – Niola Doa. In: G. Negro, A. Ravenna & R. Simonis (eds), Arte rupestre nel Ciad Borku, Ennedi, Tibesti. Sahara Dossier 1, p. 90-94.

Fuchs P. (1957) – Felsmalereien und Felsgravuren in Tibesti, Borku und Ennedi. Archiv für Völkerkunde, vol. XII, p. 110-135.

Gallardo F. & De Souza P. (2008) – Rock Art, Modes of Production, and Social Identities during the early Formative Period in the Atacama Desert (Northern Chile). In: I. Domingo Sanz, D. Fiore & S.K. May (eds), Archaeologies of Art: Time, Place, Identity. Left Coast Press, Walnut Creek, p. 79-97.

Gauthier Y. & Gauthier C. (2006) – Nouveaux abris peints de l’Ennedi (Chad). Sahara, vol. 17. p. 165-172.

Hassanein Bey A.M. (1924) – Through Kufra to Darfur. The Geographical Journal, vol. LXIV, n° 5, p. 353-366.

Jacquet G. (2000) – Piste oubliée en Haut-Ennedi. Sahara, vol. 12 : p. 141-149.

Jesse F. (2003) – Rahib 80/87. Ein Wavy-Line-Fundplatz im Wadi Howar und die früheste Keramik in Nordafrika. Africa Praehistorica 16. Heinrich-Barth-Institut, Köln.

Jesse F., Keding B., Lenssen-Erz T. & Pöllath N. (in press) – “I hope your cattle are well” archaeological evidence for early cattle-centred behaviour in the Eastern Sahara of Sudan and Chad. In: M. Bollig, M. Schnegg & H.P. Wotzka (eds), Pastoralism in Africa: Past, Present and Future. Berghahn, New York & Oxford, p. 66-103.

Keding B. (1997) – Djabarona 84/13. Untersuchungen zur Besiedlungsgeschichte des Wadi Howar anhand der Keramik des 3. und 2.Jahrtausends v. Chr. Africa Praehistorica 9. Heinrich-Barth-Institut, Köln.

Keding B., Lenssen-Erz T. & Pastoors A. (2007) – Pictures and Pots from Pastoralists Investigations into the Prehistory of the Ennedi Highlands, NE Tchad. Sahara, vol18, p. 23-45.

Kollmannsperger F. (1957) – Drohende Wüste. Brockhaus, Wiesbaden.

Kröpelin, S., Verschuren D., Lézine A.-M., Eggermont H., Cocquyt C., Francus P., Cazet J.P., Fagot M., Rumes B., Russell J.M., Darius F., Conley D.J., Schuster M., von Suchodoletz H., Engstrom D.R. (2008) – Climate-driven ecosystem succession in the Sahara: the past 6000 years. Science, 320, p. 765-768.

Kuper, R. & Kröpelin S. (2006) – Climate-controlled Holocene occupation in the Sahara: Motor of Africa’s evolution. Science, 313, p. 803-807.

Lenssen-Erz T. (2001) – Gemeinschaft Gleichheit Mobilität. Felsbilder im Brandberg, Namibia, und ihre Bedeutung. Grundlagen einer textuellen Felsbildarchäologie. Africa Praehistorica 13. Heinrich-Barth-Institut, Köln.

Lenssen-Erz T. (2007) – Rock art in African Highlands: Ennedi Highlands, Chad Artists and Herders in a Lifeworld on the Margins. In: O. Bubenzer, A. Bolten & F. Darius (eds), Atlas of Environmental Change and Human Adaptation in Arid Africa. Africa Praehistorica 21. Heinrich-Barth-Institut, Köln, p. 48-51.

Lenssen-Erz T. (2008) – Space and Discourse as Constituents of Past Identities the Case of Namibian Rock Art. In: I. Domingo Sanz, D. Fiore & S.K. May (eds), Archaeologies of Art: Time, Place, Identity. Left Coast Press, Walnut Creek, p. 29-50.

Lenssen-Erz T. (2012) – Adaptation or Aesthetic Alleviation Which Kind of Evolution Do We See in Saharan Herder Rock Art of Northeast Chad? Cambridge Archaeological Journal, vol. 22, n° 1, p. 89-114.

Lenssen-Erz T. (in press) – Cooperation or conflict? Identity and scarce resources of prehistoric Saharan pastoralists. In: Michaela Pelican & Junko Maruyama (eds), VAD Proceedings.

Lenssen-Erz T. & von Czerniewicz M. (2005) – Résultats préliminaires des recherches archéologiques dans l’Ennedi. Revue Scientifique du Tchad , vol. 7 n° 2, p. 5-18.

LeQuellec J.L. (1998) – Art rupestre et préhistoire du Sahara. Éditions Payot & Rivages, Paris.

Lewis-Williams J.D. & Loubser J.H.N. (1986) Deceptive Appearances: A Critique of Southern African Rock Art Studies. Advances in World Archaeology, vol. 5, p. 253-289.

Meerpohl M. (2007) – Mobile oases in the Libyan Desert. In: O. Bubenzer, A. Bolten & F. Darius (eds), Atlas of Environmental Change and Human Adaptation in Arid Africa. Africa Praehistorica 21. Köln, Heinrich-Barth-Institut, p. 138-141.

Monod T. (1968) – Rapport sur une mission exécutée dans le nord-est du Chad en Décembre 1966 et Janvier 1967. Fort Lamy : Institut National Chadien.

Negro G., Ravenna A. & Simonis R. (1996) – Arte rupestre nel Ciad Borku, Ennedi, Tibesti. Sahara, Dossier 1.

Oukafi Cheikh I. (2012) – L’histoire d’Eharir (Tassili n Azjer, Sahara) dans la perception locale de l’art rupestre (Analyse et interprétation dans la perspective ethnoarchéologique). Unpubl. Master Thesis, Köln.

Russell T. (2012) – Through the skin: Exploring pastoralist marks and their meanings to understand parts of East African rock art. Journal of Social Anthropology, vol. 0, n° 0, p. 1-28. Published online 7 Dec. 2012 http://jsa.sagepub.com/content/early/2012/12/05/1469605312455772

Sauvet G., Layton R., Lenssen-Erz T., Taçon P. & Wlodarczyk A. (2006) – La structure iconographique d’un art rupestre est-elle une clef pour son interprétation? Zephyrus, vol. 59, p. 97-110.

Schütz A. & Luckmann T. (1975) – Strukturen der Lebenswelt. Luchterhand, Neuwied und Darmstadt.

Sepulchre P., Schuster M., Ramstein G., Krinner G., Girard J.F., Vignaud P. & Brunet M. (2008) – Evolution of Lake Chad basin hydrology during the mid-Holocene; a preliminary approach from lake to climate modelling. Global and Planetary Change (2008), vol. 61, n° 1-2, p. 41-48.

Simonis R., Faleschini G. & Negro G. (1994) – Niola Doa, “il luogo delle fanciulle” (Ennedi, Ciad). Sahara, vol. 6, p. 51-62.

Simonis R., Campbell A.C. & Coulson D. (1998) – A Niola Doa “lost site” revisited (Ennedi, Chad). Sahara, vol. 10, p. 126-129.

Tilho J. (1920) – The exploration of Tibesti, Erdi, Borkou, and Ennedi in 1912-1917. The Geographical Journal, vol. LVI, n° 3, p. 161-183.

Treinen-Claustre F. (1982) – Sahara et Sahel à l’âge du fer. Borkou, Tchad. Mémoires de la Société des Africanistes. Musée de l’Homme, Paris.

Top of page

List of illustrations

URL http://journals.openedition.org/aaa/docannexe/image/414/img-1.jpg
File image/jpeg, 536k
Title Figure 1 – Satellite picture of the Ennedi Highlands
Caption Scale bare: 100 km
Credits © Heinrich Barth Institut MRAC
URL http://journals.openedition.org/aaa/docannexe/image/414/img-2.jpg
File image/jpeg, 1.9M
Title Figure 2 – Engraving next to the waterhole Oura in the northeast of the Ennedi shows a man with heeled shoes and an automatic rifle on his side, the animal is probably a dog so that this engraving may refer to hunting activities. It is proof that even in our times people produce engravings although context and meaning have changed since prehistory
Caption Height of person ca. 15 cm
Credits © A. Pastoors
URL http://journals.openedition.org/aaa/docannexe/image/414/img-3.jpg
File image/jpeg, 2.2M
Title Figure 3 – This young nomad by the name of Mahamat, living in central Ennedi Highlands, asserted that the engraving marked here with a circle was done by himself, being the sign with which his family marks their camels
URL http://journals.openedition.org/aaa/docannexe/image/414/img-4.jpg
File image/jpeg, 1.8M
Credits © Lenssen-Erz
URL http://journals.openedition.org/aaa/docannexe/image/414/img-5.jpg
File image/jpeg, 2.1M
Credits © Lenssen-Erz
URL http://journals.openedition.org/aaa/docannexe/image/414/img-6.jpg
File image/jpeg, 1.6M
Credits © Lenssen-Erz
URL http://journals.openedition.org/aaa/docannexe/image/414/img-7.jpg
File image/jpeg, 1.7M
Credits © Lenssen-Erz
URL http://journals.openedition.org/aaa/docannexe/image/414/img-8.jpg
File image/jpeg, 1.4M
Credits © Lenssen-Erz
URL http://journals.openedition.org/aaa/docannexe/image/414/img-9.jpg
File image/jpeg, 1.8M
Credits © Lenssen-Erz
URL http://journals.openedition.org/aaa/docannexe/image/414/img-10.jpg
File image/jpeg, 1.5M
Credits © Lenssen-Erz
URL http://journals.openedition.org/aaa/docannexe/image/414/img-11.jpg
File image/jpeg, 1.6M
Credits © Lenssen-Erz
URL http://journals.openedition.org/aaa/docannexe/image/414/img-12.jpg
File image/jpeg, 1.8M
Credits © Lenssen-Erz
URL http://journals.openedition.org/aaa/docannexe/image/414/img-13.jpg
File image/jpeg, 1.3M
Top of page

References

Bibliographical reference

Tilman Lenssen-Erz, Pastoralist appropriation of landscape by means of rock art in Ennedi Highlands, ChadAfrique : Archéologie & Arts, 8 | 2012, 27-43.

Electronic reference

Tilman Lenssen-Erz, Pastoralist appropriation of landscape by means of rock art in Ennedi Highlands, ChadAfrique : Archéologie & Arts [Online], 8 | 2012, Online since 13 October 2015, connection on 28 March 2024. URL: http://journals.openedition.org/aaa/414; DOI: https://doi.org/10.4000/aaa.414

Top of page

About the author

Tilman Lenssen-Erz

Universität zu Köln / Université de Cologne, Forschungsstelle Afrika, Institut für Urund Frühgeschichte, AAArC African Archaeology Archive Cologne, Jennerstr. 8 50823 Köln, Allemagne, lenssen.erz@uni-koeln.de

Top of page

Copyright

The text and other elements (illustrations, imported files) are “All rights reserved”, unless otherwise stated.

Top of page
Search OpenEdition Search

You will be redirected to OpenEdition Search