Navigation – Plan du site

African Elephants, Roan Antelope, Language and Rock Art

Éléphants d’Afrique, antilopes, linguistique et art rupestre
J. Francis Thackeray
p. 17-22


Les données linguistiques et éthologiques concernant les éléphants sont examinées en lien avec l’art rupestre d’Afrique du Sud, avec une attention particulière pour le comportement de ces grands animaux à défense qui deviennent dangereux et agressifs quand ils sont blessés. C’est le cas également de l’antilope rouanne (hippotrague) qui possède de longues cornes courbes. Dans cet article, je tente de démontrer que le comportement des éléphants et des antilopes blessés a pu constituer le point de départ de l’expression de quelques concepts identifiables dans l’art rupestre. Les mots peuvent être considérés comme des « fossiles » dans le contexte de ce que je nommerais une paléontologie linguistique.

Haut de page

Texte intégral


1Elephants (Loxodonta africana) and southern African rock art have recently been the subject of discussion by Andrew Paterson (2007, 2018) and David Lewis-Williams (2019). Of special interest are therianthropes (part animal and part human). A. Paterson (2018) cleverly coins the term “elephanthrope” for the Ebusingata (Drakensberg) and Cederberg (Western Cape) therianthropes, where “ant” is the common form at the end of the word elephant and the beginning of “anthrope”. Its derivation can be traced to Greek forms: ἐλέφᾱς for elephas, meaning elephant; and Greek ἄνθρωπος, anthropos, meaning human. (Latin nominative singular is elephas, whereas the nominative plural is elephantes). It should be noted that elephanthropes are very rare in southern African rock art. They are represented at only two sites in the Cederberg, and another two in the Drakensberg (Paterson 2018).

2A. Paterson (2007, 2018) gives a cogent argument that certain Cederberg elephant scenes can be interpreted in the context of social behaviour and communication, where sounds may be represented by lines surrounding the animals portrayed in the art (Parkington & Paterson 2017). Some of the elephant scenes in the Cederberg may also relate to trance (Maggs & Sealy 1983; Deacon 1994). The elongation of an elephant in a painting in the Tsodilo Hills (Botswana) may be trance-related (fig. 1). It is pertinent to supplement these interpretations with an exploratory study of linguistics and ethology (the study of animal behaviour).

Figure 1 – The figure of an unusual elongated elephant in the rock art of the Tsodillo Hills, Botswana, painted with white pigment

Figure 1 – The figure of an unusual elongated elephant in the rock art of the Tsodillo Hills, Botswana, painted with white pigment

© J. F. Thackeray

San words for elephants and other animals

3A San term for an elephant cited by A. Paterson (2007, 2018) is !xo. Other San terms for the animal, as reported by Dorothea F. Bleek (1956) in A Bushman Dictionary, are ‡xoa and |xwa: (southern San SII). They may be considered cognate with koa (central San CIIb) and ‡khoab (central San CII and northern San NIIa).

4These words are of great interest in the sense that //khoab is a Khoi term for Hippotragus equinus, the roan antelope (Thackeray 2005). Furthermore, koa is a central San term (CIIb) for sable (Hippotragus niger). Kwa is a central San (CI) term for roan (CI) and kwar is the southern San (SI-B) word for the same animal, as recorded by D. F. Bleek (1956). Both roan and sable are very large antelope with long curved horns.

5These Khoisan terms for elephant and roan/sable are summarised as follows, linking words of similar if not identical form:


Roan/sable (hippotragine antelope)









kwa, kwar

6The similarities between certain terms for elephant and hippotragine antelope give reason to question whether they are coincidental, or whether they relate to the way in which these animals were classified (linguistic taxonomy) and perceived in prehistory, in terms of behaviour or other factors. One response to this question (and not necessarily the only one) is that these particular animals become highly aggressive, especially when wounded either as a consequence of ritual fights between males for territory and/or access to females, or when wounded as a consequence of hunting.

7The concept of “wounded roan” appears to have been important in African prehistory (Thackeray 2005), perhaps with considerable time depth such that a full understanding of the symbolism may be difficult to appreciate, solely on the basis of ethnographic data that have been recorded in recent centuries. Linguistic data may facilitate the identification of ancient conceptual associations not fully expressed ethnographically. For example, J. Francis Thackeray (2005) drew attention to the term for a Khoi “supreme being”, Tsuni//Goam (Hahn 1881), otherwise identified as Tsui//Khoab in the contect of tsui (wounded) and //khoab (roan). The question arises as to whether Tsui//Khoab might at one time have been identified also with the behaviour of wounded elephant, recognising that ‡khoab is a San term for this large tusked animal.

Other terms for elephants and potential associations

8Among Nguni we have the form ndhlovu (isiZulu and Siswati) and in-dlovu (isiXhosa) as terms for elephants (Walker 2015; McLaren 1984). In Xhosa the form –dlovu is found in the word -dlovula meaning “to make havoc”; furthermore, -dlovu is cognate with dlova, as in isidlova, a noun referring to a violent or destructive person (McLaren 1984). Associations between elephants and these concepts of havoc, violence or destruction can potentially be appreciated by understanding aspects of the behaviour of elephants. Relevant here is the description of the destructive behaviour or havoc caused by “violent” dangerous elephants, as reported by J. D. Skinner & R. H. N. Smithers (1990: 550):

“Elephants are normally peaceful, but individuals, especially if sick or injured, can be aggressive and extremely dangerous. Females in groups with young are particularly unpredictable”.

9Adult male elephants can become very aggressive when in “musth” due to increased level of androgen. In the context of destruction, J. D. Skinner & R. H. N. Smithers (Ibid.) state “With their tusks elephants can break down substantial tree branches. Furthermore, they go on to say that fights between males are extremely rare… when they become enraged, however, they do great damage to each other with their tusks, and aggression can persist until one or the other is seriously wounded or killed”.

10Analogous aggressive behaviour is displayed when roan or sable antelope fight each other in ritual combat (as in the case of males), or when wounded (Thackeray 2005), the very conditions in which they are dangerous.

11Kwa is a word not only for elephants and roan antelope (see above). Its form is also found in the San term ¯kwa ¯kwa which refers to being dangerous or aggressive. These are precisely the concepts to which I have alluded in the context of both elephant and roan.

Elephants and rock art

12The observations reported here may be relevant to certain examples of rock art. Notably, in figure 2, images of an elephant and part of an antelope with long curved horns (which I cautiously regard here as a hippotragine bovid, possibly Hippotragus leucophaeus) are superimposed in the darkness of a passage deep in the Cango Caves in the southern Cape near the town of Oudshoorn (Thackeray & Gerstner 2008). The cave was visited briefly by Henri Breuil in 1929 (as documented on September 20 of that year in the Oudshoorn Courant), but he did not copy the elephant and antelope which were recorded carefully by J. F. Thackeray and Hein Gerstner (fig. 2). The details were extremely difficult to record, possibly due to their antiquity and lack of preservation despite being in a protected and secluded area of the cave. In this instance, the superimposition of an elephant and a hippotragine antelope may reflect a conceptual association of the kind that developed in prehistory and which may be expressed linguistically on the basis of the kinds of words examined in this exploratory study. Notably, small subparallel incisions can be observed on the middle of the body of the elephant. I propose that these incisions relate to symbolic wounds.

Figure 2 – Two components of an elephant, and part of an antelope (here identified as a hippotragine bovid) in the Cango Caves, Southern Cape, South Africa

Figure 2 – Two components of an elephant, and part of an antelope (here identified as a hippotragine bovid) in the Cango Caves, Southern Cape, South Africa

© J. F. Thackeray & H. Gerstner

13Secondly, a therianthrope at Ebusingata in the Drakensberg (Pager 1975) incorporates a human body with the head and trunk of an elephant (fig. 3). It is associated with a swarm of bees. This juxtaposition may relate not only to an association between sounds (humming or buzzing) and perceptions in trance states, but also to the aggressive behaviour of bees when swarming, if not also to the aggressive behaviour of elephants under certain conditions. Notably, the human body of the elephanthrope has dots or linear features which may represent scarifications of the kind which relate to symbolic wounds and the principle of so-called “sympathetic magic” (Thackeray 2019). The latter expression may more appropriately be called “empathy”. Scarifications or stripes as symbolic wounds may be identified for example in the case of a hunter painted in the called “White Lady” panel, copied by Henri Breuil in 1947 in the Brandberg in Namibia (Thackeray 2019).

Figure 3 – A therianthrope (elephanthrope) at Ebusingata in the Drakensberg

Figure 3 – A therianthrope (elephanthrope) at Ebusingata in the Drakensberg

It incorporates a human body with the head and trunk of an elephant, associated with a swarm of bees; scale: 10 cm

© H. L. Pager 1975

Discussion and Conclusions

14Linguistically, similarities in form of certain words may relate to conceptual associations of the kind associated with semantic shift and which may be expressed in prehistoric art (Thackeray 1994). Similarities in words for elephants, roan antelope and other concepts (as indicated in this exploratory study), may potentially relate to perceptions of animal behaviour as understood by Khoisan or Nguni Bantu-speaking peoples (and their ancestral populations) in southern Africa. Aggressive behaviour of elephants and roan, especially when wounded, may have been the basis for certain beliefs in African prehistory.

15I have previously emphasised that stripes painted or engraved on animals might represent symbolic wounds (eg. Thackeray 2005, 2019). One example is the Melikane painting in Lesotho where a therianthrope has painted stripes, corresponding closely to a photograph of a person beneath the skin of a roan antelope with painted stripes, which J. Thackeray (2005) associates with a “wounded roan”. In this study, I direct specific attention to a set of small subparallel incisions on the middle of the elephant in the Cango Cave panel (fig. 2), incorporating the hippotragine antelope with long curved horns (possibly the extinct Hippotragus leucophaeus, commonly known as the “blue-buck”, a close relative of roan antelope). It is remarkable that in this particular instance we have a superimposition of the very kinds of animals which are the subject of interest in this linguistic study, whereby it has been suggested that hippotragine antelope and elephants were conceptually associated in the context of their aggression when wounded.

16Of course many animals can become aggressive when wounded, but the roan is one of the largest African antelopes (apart from eland), and elephants are the largest extant terrestrial mammals. Thus one may venture to say that both large size and aggression, in combination, were factors that contributed to belief systems in prehistory. Further, recognising that southern African rock art is known to have been related to a large extent to shamanism, one might speculate that the aggressive behaviour of certain animals such as roan and elephants was associated with shamanistic concepts of the kind known by the San terms /num, /nom or !gi. Perhaps not coincidentally, the word !gi relates to a cut (cf. scarification), and such scarifications (cf. wounds) were believed by San hunters to be essential for success in a hunt.

17In Africa we have the advantage of having access to linguistic, ethnographic as well as ethological data which can used together with rock art in exploratory studies to identify potential associations which developed in prehistory, perhaps outside the range of visibility of modern ethnographic records. In this paper I suggest that observations of the behaviour of wounded elephants or roan may have contributed to the development of at least some concepts that may be identified in prehistoric art. Words can be regarded as fossils in the context of what may be called linguistic palaeontology.

I thank Emmanuelle Honoré for the invitation to submit this article for publication. I express my deep gratitude to Sylvie Amblard-Pison for her assistance with the manuscript. Hein Gerstner provided information regarding the Cango Caves. I extend thanks to José Braga for facilitating my visits to Rouffignac Cave where I was fascinated by Upper Palaeolithic images of mammoths, which will be the subject of research in a separate study. The late Harald Pager and David Lewis-Williams stimulated my interest in southern African rock art. In my retirement I have had the opportunity to explore a diversity of topics, including palaeoanthropology, linguistics, ethnography and rock art as aspects of prehistory. Part of this study relates to my research while employed and supported by the National Research Foundation (South Africa).

Haut de page


Bleek D. F. (1956) – A Bushman Dictionary. New Haven, American Oriental Study.

Deacon J. (1994) – Some views on rock paintings in the Cederberg. Cape Town, National Monuments Council.

Hahn T. (1881) – Tsuni-//Goam: The Supreme Being of the Khoi-khoi. London, Trübner.

Lewis-Williams J. D. (2019) – More about elephants and related matters in San rock art research. The Digging Stick, 36(1): 10-13.

Maggs T. M. & Sealy J. (1983) – Elephants in boxes. In: J. D. Lewis-Williams (ed.), New approaches to southern African rock art, Cape Town, South African Archaeological Society Goodwin Series, 4: 44-48.

Mclaren J. (1984) – A new concise Xhosa-English dictionary. Cape Town, Maskew Miller.

Pager H. L. (1975) – Stone age myth and magic as documented in the rock paintings of South Africa. Graz, Akademische Druck-u. Verlagsanstalt.

Paterson A. (2007) – Elephants (!xo) of the Cederberg Wilderness Area: a re-evaluation of the San paintings previously referred to as “Elephants in boxes”. The Digging Stick, 24(3): 1-4.

Paterson A. (2018) – Elephanthropes of the Cederberg. The Digging Stick, 35(3): 1-6.

Parkington J. & Paterson A. (2017) – Somatogenesis: Vibrations, undulations and the possible depiction of sound in San rock paintings of elephants in the Western Cape. The South African Archaeological Bulletin, 72(206): 134-141.

Skinner J. D. & Smithers R. H. N. (1990) – The Mammals of the Southern African Subregion. Pretoria, University of Pretoria.

Thackeray J. F. (1994) – Animals, conceptual associations and southern African rock art: a multidisciplinary, exploratory approach. In: T. Dowson  & J. D. Lewis-Williams (eds), Contested images: diversity in southern African rock art research, Johannesburg, Witwatersrand University Press: 223-235.

Thackeray J. F. (2005) – The wounded roan: a contribution to the relation of hunting and trance in southern African rock art. Antiquity, 79(303): 5-18.

Thackeray J. F. (2019) L’apport de l’ethnographie au concept de « chasse magique par empathie » (ou compatissante) et de chamanisme dans l’art préhistorique d’Afrique austral. L’Anthropologie, 123 (1) : 180-188.

Thackeray J. F. & Gerstner H. (2008) – Rock art in the Cango Caves. The Digging Stick, 25(3): 17.

Walker C. (2015) – Signs of the wild. A field guide to the spoor & signs of the mammals of southern Africa. Cape Town, Penguin Random House Struik.

Haut de page

Table des illustrations

Fichier image/jpeg, 236k
Titre Figure 1 – The figure of an unusual elongated elephant in the rock art of the Tsodillo Hills, Botswana, painted with white pigment
Crédits © J. F. Thackeray
Fichier image/jpeg, 140k
Titre Figure 2 – Two components of an elephant, and part of an antelope (here identified as a hippotragine bovid) in the Cango Caves, Southern Cape, South Africa
Crédits © J. F. Thackeray & H. Gerstner
Fichier image/jpeg, 96k
Titre Figure 3 – A therianthrope (elephanthrope) at Ebusingata in the Drakensberg
Légende It incorporates a human body with the head and trunk of an elephant, associated with a swarm of bees; scale: 10 cm
Crédits © H. L. Pager 1975
Fichier image/jpeg, 128k
Haut de page

Pour citer cet article

Référence papier

J. Francis Thackeray, « African Elephants, Roan Antelope, Language and Rock Art », Afrique : Archéologie & Arts, 15 | 2019, 17-22.

Référence électronique

J. Francis Thackeray, « African Elephants, Roan Antelope, Language and Rock Art », Afrique : Archéologie & Arts [En ligne], 15 | 2019, mis en ligne le 15 décembre 2019, consulté le 18 février 2020. URL : ; DOI :

Haut de page


J. Francis Thackeray – Evolutionary Studies Institute, University of the Witwatersrand, PO WITS, Johannesburg 2050, South Africa

Haut de page

Droits d’auteur

CNRS - ArScAn. Cartographie d’après

Haut de page
  • Logo ArScAn - Archéologies et Sciences de l’Antiquité (UMR7041)
  • Logo Ethnologie Préhistorique
  • Logo CNRS
  • OpenEdition Journals