Was there an interchange between Cushitic pastoralists and Khoisan speakers in the prehistory of Southern Africa and how can this be detected?

Full text

Was there an interchange between
Cushitic pastoralists and Khoisan speakers
in the prehistory of Southern Africa and
how can this be detected?
Special Research Centre ACACIA project ‘Migration, Settlement and Cultural History on the Basis
of Linguistic Sources’, University of Cologne
Presented at Königswinter, March 28-30, 2007
and subsequently submitted for publication
To be published in a special volume of Sprache und Geschichte in Afrika edited by Wilhelm Möhlig & Axel
Fleisch
Roger Blench
Kay Williamson Educational Foundation
8, Guest Road
Cambridge CB1 2AL
United Kingdom
Voice/ Fax. 0044-(0)1223-560687
Mobile worldwide (00-44)-(0)7967-696804
E-mail R.Blench@odi.org.uk
http://www.rogerblench.info/RBOP.htm
Cambridge, Friday, 08 August 2008

TABLE OF CONTENTS
1. INTRODUCTION: THE ORIGIN OF THE ‘HOTTENTOTS’............................................................. 1
2. CONTACT BETWEEN CUSHITIC AND KHOISAN PEOPLES ........................................................ 2
3. LIVESTOCK...............................................................................................................................................3
4. MATERIAL CULTURE ............................................................................................................................4
4.1 Pottery.....................................................................................................................................................................4
4.2 Mat tents.................................................................................................................................................................5
4.3 Butter-making........................................................................................................................................................5
4.4 Further suggestions................................................................................................................................................5
5. LINGUISTIC EVIDENCE......................................................................................................................... 6
6. GENETIC EVIDENCE ..............................................................................................................................9
7. CHRONOLOGICAL SCENARIO AND SYNTHESIS...........................................................................9
REFERENCES..............................................................................................................................................10
TABLES
Table 1. Early dates for Southern African livestock ......................................................................................... 3
Table 2. A root for ‘cattle’ in Khoisan .............................................................................................................. 6
Table 3. The *guu root for ‘sheep’ in Khoisan and Southern Bantu................................................................. 7
Table 4. The –pene root for goat terms in Khoisan and Bantu ......................................................................... 8
Table 5. The –gumbo root for goat terms in Northern Khoisan and Bantu....................................................... 8
FIGURES
Figure 1. Distribution of fat-tailed sheep in Africa........................................................................................... 3
Figure 2. Fat-tailed sheep represented in rock-art in Mazowe district, Zimbabwe........................................... 4
Figure 3. Khoi cow depicted at the Cape, 1778 ................................................................................................ 4
Figure 4. Bisharin mat-house, Red Sea Province, Sudan.................................................................................. 5
Figure 5. Eholo butter-maker, Southern Angola ............................................................................................... 5
Figure 6. Map showing potential overlaps between Cushitic and Khoe speakers .......................................... 10

1
1. Introduction: the origin of the ‘Hottentots’
The Khoisan populations of Southern Africa are conventionally divided into the Khoekhoe and the San, the
main logic of the dichotomy being that the Khoekhoe were pastoralists at first contact, while the San
subsisted purely by foraging (Barnard 1992). This does not entirely correspond to the linguistic classification
of the Khoisan languages, but nonetheless represents a fairly significant division, especially as it later
became apparent that the pastoral groups were lactase tolerant, while the hunter-gatherers were not, arguing
for a genetic split of some antiquity (Nurse et al. 1985: 97).
In the early literature, the Khoe were usually referred to as Hottentots, a word of disputed etymology. First
recorded in Dutch in 1677, it is said to mean stutterer; or to be an ideophonic representation, hot en tot, of
the stereotyped sounds in the Khoe language. Very early, observers of the Khoe peoples noted features of
their culture that set them apart from both San and Bantu and it soon came to be assumed these were
evidence of Semitic origins. Kolb (1731) observed;
‘These customs, in which the Hottentots agree with both the Jews and the Troglodytes, being, ’tis
pretty certain, all or most of ’em as old as the time of Abraham, which was but 300 Years after the
Flood, refer their Tradition so clearly to Noah, as to put the matter almost out of doubt’.
Somewhat later, Mentzel was to account for the phenotype of the Khoe by assuming they were Asian
children shipwrecked on the South Africa coast1. However, more broadly, the tendency was to assume that if
the Khoe indeed had ‘Semitic’ affinities, they must in some way have migrated from NE Africa. This is the
model found in Stow (1905) and expanded by Jeffreys (1968), who maintained there were ‘Semitic
influences’ on Hottentot culture. Indeed Jeffreys, always an author to defend entirely improbable theories
with immense scholarly apparatus, followed Mentzel in holding that the Khoi were the result of seagoing
Semites landing on the coast and intermarrying with the local San stock. Carl Meinhof’s (1912) Die
Sprachen der Hamiten provided a spurious linguistic justification for these views, linking as it did Cushitic,
Nama and Fulfulde (a West African Niger-Congo language). All of these writers drew the conclusion that
the ancestors of the Khoe must therefore have migrated from elsewhere, most likely NE Africa. Shorn of
Semitic rhetoric, even later, more archaeologically informed writers such as Ehret (1982: Map 13) and
Elphick (1985) assumed an origin somewhere in northern Botswana, NE of the present range of the Khoi.
A secondary development of these ideas was that there must be ‘Hamitic’ elements in the culture of the
Bantu herding peoples along the Angola/Namibia border, notably the Herero, Himba and Kwanyama/Ambo.
Anthropological texts such as Irle (1906), Estermann (1950) assumed chamitique influence, a view
enshrined in Baumann’s (1940) overview of African peoples. The argument typically depends on the sacred
nature of cattle and the complex of rituals around them, although the parallels adduced are often remarkably
short on detail. The possibility that cattle pastoralists might develop elaborate rituals around the most
important element in their subsistence strategies seems not to have occurred to these authors.
Loeb’s (1962) In Feudal Africa argued for ‘Indications of early Mediterranean influence’ on the culture of
the Kwanyama people on the border of Namibia and Angola. This goes back to a Germanic tradition
describing the purported wanderings of the Hamiten (see, for example, Adametz (1920) or Lebzelter (1934)).
Loeb refers to ‘noticeably Caucasian features’ (p. 6) and a tradition that the Kwanyama ‘lived originally in
the region of the Great Lakes of Africa’ (p. 9). According to this version of prehistory, the same sort of NE
African pastoralists as originated the Khoe must have migrated to SW Africa, bringing with them herding
culture, but being largely absorbed phenotypically. This in turn was held to be responsible for ‘stratification’
in the culture of the Bantu pastoralists.
‘Hamitic’ is a conflated cultural/racial category which lumped together Cushitic/Nilotic peoples and even
Bantu peoples, such as the Tutsi, who were deemed to have Hamitic characteristics. Composite terms such
as ‘Nilo-Hamites’ were invented to cover peoples whose linguistic identity did not seem to match their
cultural traits. Greenberg (1963:24, 69) largely skewered the crypto-racial nature of the Hamitic theory,
1 By some chance, Asians (a Thai embassy to Europe) were shipwrecked on the South African coast in the seventeenth
century, unfortunately too late to be responsible for such a remarkable confluence of races (Smithies 1999).

2
particularly in relation to Fulfulde, although as Boonzaier et al. (1996:14) point out, it continues to live a
shadowy life in school textbooks.
The notion of a Hamitic culture brought together a bundle of obsessions of scholars at the period, that racial
and linguistic categories went hand in hand, that ‘tall’ people were somehow superior, that cattle-herders
trumped cultivators and that migration from an appropriate homeland could explain widely distributed
cultural features. Can anything be salvaged from this? Do any of these observations have validity or is this
just racist nonsense? This paper will argue that these earlier authors had observed something important, but
lacking an interpretative framework to make sense of it, they veered off into a wild hinterland of speculation.
Shedding the more outré assumptions of the earlier literature, a still more intricate story can be told of early
contact between different cultures and the way in which evidence can be overprinted by subsequent
population movements.
However, perhaps the ground should be cleared before developing a more positive argument by saying that;
Evidence for Semitic influence is non-existent
Evidence that the Khoe migrated from ‘elsewhere’ is entirely lacking
Evidence that the Bantu herders of Southern Angola migrated with their present culture from NE Africa
is absent
There is no proven genetic connection between Khoisan languages and Afroasiatic or any other group
Mitchell (2002: chap 9) provides a very judicious account of the archaeological context of Khoe pastoralism
but reaches no firm conclusions about its origins. Contra the Hamitic model, a more ‘indigenist’ tradition
exists in the archaeological literature. John Kinahan (1991) who has probably contributed more than any
other researcher to the archaeology of pastoralism in Namibia, states in his hypothesis 1, ‘Nomadic
pastoralism arose out of the indigenous Central Namib hunting economy when a fundamental ideological
change permitted the accumulation of property in domestic livestock’. It could be argued that this is actually
circular; once you start building up herds you inevitably undergo a ‘fundamental ideological change’.
Kinahan does not address the question of the source of specific livestock breeds nor the mechanism of their
transfer to the Khoe.
This paper will argue that the explanation for some continuities of pastoral culture between NE Africa and
the Khoe-speaking peoples is really quite simple; pastoralists speaking Cushitic languages once spread as far
as south-central Africa, where they were in contact with the ancestors of present-day Khoe-speakers. This
led to a transfer of both species of domestic animals and also some rather specific techniques of pastoral
lifestyle including dairy-processing etc. Khoe pastoral culture is known mainly from records and their
original sheep and cattle breeds have now become heavily crossbred. The explanation for related traits
among adjacent Bantu peoples is likely to be a similar, subsequent transfer from the Khoe to the Bantu,
although it is possible that there was also direct Cushitic contact with the Bantu in the same region. It is
further likely that this was connected with the expansion of the Khoe peoples, explaining why their language
subgroup is remarkably coherent within Khoisan, which is otherwise characterised by a high level of internal
diversity, reflecting its considerable antiquity. The importance of the pastoral revolution in Southern Africa
led to the borrowing of livestock terms into other branches of Khoisan.
2. Hypotheses as to the origin of Khoe livestock culture
It has long been observed that some groups of Khoe peoples of south-western Africa acquired pottery, sheep
and cattle within quite a short time window (ca. 2000 BP) before attested contact with expanding Bantu-
speakers. Links between the pastoral cultures of NE Africa and the culture of the Khoe and adjacent Bantu
herding peoples are much less certain. The mechanism for these innovations has been much debated, as the
introductory discussion shows. The pottery might be an independent invention, although given the
geographical proximity of other pottery-makers, this is unlikely, but the livestock must have been
transmitted by another group of livestock-keepers as sheep and cattle have no wild relatives in Southern
Africa. As it turns out, some very specific techniques associated with pastoral production are also shared
between SW and NE Africa, making cultural transmission the only reasonable assumption.

3
Several hypotheses might account for this: pastoralists made their way to SW Africa and were assimilated,
‘becoming’ Khoe or Bantu; Khoisan speakers were once resident in modern-day Tanzania, as the evidence
of Hadza and Sandawe appears to show, and the transfer took place there; or that both were once in contact
in an area now dominated by Bantu-speakers, such as modern-day Zambia (Blench 2006). Güldemann (in
press) takes the view that the ancestors of the Khoe were originally resident in East Africa and were not
physically Khoisanoids. Indeed, as Smith (2005:163) notes, no Khoisan-type skeletal material has ever been
found north of the Zambezi. In this model, as the ancestral Khoe migrated southwards and interacted with
the click-speaking foragers, they came to resemble them more closely, both linguistically and culturally.
Fauvelle-Aymar (2004) reviews some of the cultural connections between Khoe pastoralism and the pastoral
practice elsewhere in Africa and points to similarities he sees with Nilo-Saharan, a view similar to that
espoused by Ehret (1998). This paper2 will present some of the evidence for contact, focussing on livestock
breeds, material culture and linguistics and interpreting it in the light of modern archaeological evidence.
3. Livestock
The Khoe were in possession of cattle, sheep, goats and dogs when first encountered by European observers
(Boonzaier et al. 1996). Table 1 shows selected early dates for Southern African livestock in the
archaeological record;
Table 1. Early dates for Southern African livestock
Species Location Site Calibrated date
Sheep Namibia Falls rockshelter° 190 BC–383 AD
Sheep Botswana Toteng 190 BC–AD 20
Sheep South Africa Blombos 82 AD-215 AD
Sheep South Africa Spoegrivier 165 BC-AD 13
Ovicaprines South Africa Ma38 200–300 AD
Cattle Botswana Toteng 190 BC–AD 20
Cattle Botswana Lotshitshi >200 AD
Cattle South Africa Happy Rest >300 AD
Expanded from Sealy & Yates (1994), Henshilwood (1996),
Bousman (1998); Smith (2000), Robbins et al. (2008)
It seems likely that the ovicaprines were sheep. Goats were
herded by the Khoe when European observers first
encountered them but their exact antiquity is uncertain.
Badenhorst (2006) in a review of the evidence for goats in
Southern Africa, comments on the difficulty of distinguishing
goat and sheep bones, but notes a marked absence of early
dated goat bones. The general assumption is that the Khoe
acquired goats following contact with the expanding Bantu. A
piece of contributory evidence for this is that while southern
Bantu languages such as Xhosa have borrowed their words for
‘sheep’ and ‘cattle’ from Khoe, words for goat are not
borrowed (§5.).
The breeds of sheep and cattle found among the Khoe provide
clues to their origin. The sheep were all of the fat-tailed type,
common in NE Africa and Arabia but otherwise entirely
absent in West-Central Africa (Epstein 1971; Blench 1993).
Figure 1 shows the distribution of fat-tailed sheep in the recent
past; they have practically disappeared in recent times in the
Southern African region, but plenty of evidence for their distribution is provided by existing documentation.
Goodall (1946) first illustrated the fat-tailed sheep in Zimbabwean rock art (Figure 2) and they also occur in
paintings in the western Cape (Manhire et al. 1986; Hollman 1993). The rock art of Southern Angola is now
2 Thanks to Tom Güldemann, Maarten Mous, Karim Sadr and Bonny Sands and for helpful comments on an earlier
version. Wilhelm Möhlig kindly supplied me with further comparative data on livestock names
Figure 1. Distribution of fat-tailed sheep
in Africa
Historical occurrence;
indigenous sheep now
largely replaced by modern
Rock pa intings indicating
former presence of fat-
tailed sheep

4
quite well-known (Gutierrez 1996, 2008) and there is no trace of sheep or indeed any livestock
representations, providing reasonable evidence that the migrant early pastoralists did not colonise this area.
The cattle typical of expanding Bantu pastoralists were the Sanga,
a cross between zebu and the humpless taurines previously found
across Middle Africa. The Horn of Africa seems to have once
had both humpless longhorns and shorthorns, to judge by rock
paintings and surviving relic populations (Blench 1993). Gifford-
Gonzalez (2000) points to the disease challenges to pastoral
expansion and the taurines, as the longer-adapted race, would
have found survival easier in the high-challenge environments of
southern Africa. Although zeboids cannot always be reliably
identified in archaeozoological assemblages as identification
depends on the presence of the bifid vertebrae (or on finds of
models of humped cattle) the relatively late dates for such
indicators strongly associate their southward movement with the
expanding Bantu (Magnavita 2006 esp. Figs. 1 & 2). To judge by early representations of the Khoekhoe,
their cattle were all of the longhorn taurine kind (Figure 3 and Figs. 23, 28 in Boonzaier et al. 1996), and
similar types survived among some pastoral groups in Southern Angola until recent times (Hauenstein 1980
Figs. 1-6). Indeed these are referred to as ‘Hamitic longhorns’ in older literature. Epstein (1971, I:482)
retains the idea that the ‘Africander’ cattle are humped cattle, even though both his illustrations and indeed
quoted early literature explicitly deny this.
4. Material culture
4.1 Pottery
It has long been observed that pottery appears in the
Southern African archaeological record prior to the
arrival of the Bantu. A recent review concludes;
‘Thin-walled, fibre tempered pottery appears [in
Southern Africa] two to four centuries before the
arrival of Iron Age agro-pastoralists who were
uniformly associated with thick-walled ceramics’
(Sadr & Sampson 2006). Thin-walled pottery, is often
identified as Bambata ware in the literature since the
early 1980s although this only designates one subtype
(Robbins et al 2008). Sadr (2008) provides a comprehensive review of the differentiation between the two
pottery types. He says ‘thin ware is found in small and large, open and sheltered sites, always associated
with LSA stone tools, some combination of hunter-gatherer-forager-fisher-herder subsistence pattern and no
evidence for the cultivation of domestic crops’ (Sadr 2008:106). Surveys of the Kavango river on the
northern fringe of the Kalahari 2005-2007 have provided evidence for a Ceramic Later Stone Age (CLSA)
associated with microlithic tools, pottery and sheep dated to the 1st millennium BC (Kose & Richter 2007).
Vogelsang & Wendt (2007) map the CLSA sites for Namibia, where they are abundant along the coastal
strip but thinly scattered in the interior. Smith (2005, 2006) has argued for a connection between the thin-
walled ‘Pastoral Neolithic’ ceramics of East Africa with those of Southern Africa (described in Robertshaw
1990: 198). In particular, he compares the spouted wares of Hyrax Hill and Ngamuriak in Kenya with the
spouted pots at Bambata in Zimbabawe and those from the Western Cape (Smith 2005: 177). Indeed, he
floats the suggestion that these are actually milking pots. This comparison remains controversial but would
certainly fit with the argument of this paper. Given that the pottery is broadly contemporaneous with arrival
of pastoralism, it would not be extravagant to assume that it was part of the same wave of introductions,
although Sadr & Sampson (2006) argue for independent invention.
Figure 2. Fat-tailed sheep
represented in rock-art in Mazowe
district, Zimbabwe
Figure 3. Khoi cow depicted at the Cape, 1778

5
4.2 Mat tents
One distinctive feature of Nama culture is the matjeshuis or
mat house, a house made from a semi-circular frame
covered in layers of mats. These are illustrated in Boonzaier
et al. (1996: Figs. 25, 26, 27, 77, 78). They are
characteristic only of the Khoe and not their pastoralist
neighbours, and have persisted, at least partly, as a cultural
symbol among today’s non-pastoral descendants. The ‘mat
house’ is also typical of the pastoral peoples from Upper
Egypt to NE Kenya, particularly the Cushitic-speakers such
as the Rendille and Beja. Figure 4 shows a typical mat
house among the Bisharin Beja, but similar constructions
are found throughout NE Africa but not in intervening areas.
Lest it be thought that such a construction is typical of
pastoral peoples, it can be noted that no similar houses are
found among nomads anywhere else in Africa (Prussin
1995) or indeed in Africa in general (e.g. Oliver 1971).
4.3 Butter-making
One of the most characteristic techniques of the Bantu-speaking
pastoral cultures along the Southern Angola/Namibia borderland is
the production of butter using a leather bag (or a gourd framed in
leather) suspended from two poles and swung from side to side by a
seated producer (Figure 5). This system is found among the cattle
producers in the Horn of Africa, Ethiopia and all the way to Egypt,
but not in the region between these two areas. It also does not occur
among other African pastoralists such as the West African Fulɓe or
the Nilotic peoples of Southern Sudan. There seems to be no clear
evidence for its possible distribution among Khoe speakers.
4.4 Further suggestions
A variety of other items of material culture could be adduced, for example, the characteristic skin sandals
manufactured by the Khoe (Boonzaier et al. 1996: Fig. 31) which are virtually identical to those in use by
East African herders today. Similarly, children’s dolls, with beads and characteristic leather skirts, used by
the San and Zulu (Frobenius 1933: Taf. 107) are stylistically identical to those made by Nilotic herders in
Northern Kenya. All of these items have broadly the same distribution; a zone between southern Angola and
the Cape and NE Africa, pointing strongly to their having been brought from this region and the distribution
made subsequently discontinuous by the southwards spread of the Bantu.
Figure 4. Bisharin mat-house, Red Sea
Province, Sudan
Figure 5. Eholo butter-maker,
Southern Angola

6
5. Linguistic evidence
The classification of Khoisan has alternated between uniting all click languages in a single phylum and
regarding the different branches as independent and unrelated. Bleek (1956) by implication, Greenberg
(1963) and Ehret (1986) argued for a ‘macro-Khoisan’ including all Southern African Khoisan plus the two
East African click languages, Hadza and Sandawe. Westphal (1962) took the opposite stance, that Kwadi,
Hoa and even the three recognised branches of Khoisan were independent of one another. This position is
now considered extreme and in Güldemann & Vossen (2000) Kwadi was ‘undetermined’, Hadza and
Sandawe ‘isolates’, and Hoa an isolate within ‘Non-Khoe’. The internal classification of Southern African
Khoisan remains controversial, but it is usually agreed to form three major subgroups (Northern = Ju or Zhu,
Central =Khoe and Southern = Tuu) plus disputed isolates. Few specialists would include Hadza within
Khoisan today although evidence for a relation between Sandawe and Khoe is considered more likely (e.g.
Elderkin 1986). The extinct Kwadi language in SW Angola has previously been considered an isolate,
although Güldemann (2004) and Güldemann and Elderkin (in press) now argue it is related to Khoe. Hoa
has previously been considered an isolate but recent views relate it to Northern or Ju languages (e.g. Honken
in press).
If the linguistic case for Khoisan/Cushitic contact were obvious, it would presumably have been pointed out
by now. It is therefore likely that contact was either with groups that have now disappeared or that the nature
of contact led to relatively few lexical transfers. Meinhof (1912:231-240) proposes common lexical items as
part of his ‘Hamitic’ argument, although only a few of his comparative series include Nama. As Meinhof
wanted to prove a genetic relationship between his Hamitic peoples much of his argument hangs on
morphology, and as Greenberg (1963:69) rightly observes, his case is very weak. An alternative model has
been presented by Ehret in several places, but most recently in Ehret (1998:323). According to this view, it
was not Cushites but Eastern Sudanic-speakers (Eastern Sahelian in Ehret’s terminology) who were
responsible for the transmission of cattle culture. This seems highly unlikely, in part because the Maasai and
other Nilotic incursions into East Africa are manifestly subsequent to the Cushitic settlement of the region (a
view which Ehret has paradoxically espoused elsewhere by identifying Cushitic substrates in East African
Bantu languages; see Ehret & Nurse 1981). Ehret’s proposed cognates are with his own Eastern Sudanic
reconstructions rather than with actual forms and are generally markedly ad hoc.
Table 2 presents the most salient and widespread cattle terms in Khoisan. The common root for ‘cow’ may
possibly be cognate with widespread terms in Cushitic that have a similar form.
Table 2. Words for ‘cattle’ in Khoisan
Branch Group Language Cow Cow Bull Comparison
Northern Ju/’Hoan gúmí cf. Proto-Agaw kǝmi
!Xun gùmì
Central Khoekoe Nama koma PEC korma ‘bull’
Khoekhoegowab goma-s ||goo
Khoe //Ani góɛ̀
Khwe góɛ́kx’áò
Naro Naro gòè //òò
//Ana /Ui gúè
Shua Cara bé
Tshwa Kua dzú bé
Kwadi goe-
Southern !Xóõ gùmi
|Xam xoro ‘oxen’
Sources: Voßen (1997, 2007), Haacke & Eiseb (1999), Visser (2001), Kilian-Hatz (2003),
König & Heine (2008), Güldemann (n.d.)
Güldemann & Elderkin (in press) make the interesting alternative suggestion that ‘*goe ‘cow’ in both Kwadi
and Kalahari Khoe could be a Bantu loan which underwent similar sound changes *gombe > *gobe > *goe’.
However, it is certainly the case that cattle reach southwestern Africa prior to the Bantu incursions in the
region making the chronology of the borrowing somewhat difficult to understand. Westphal (1963) was the

7
first to observe that Southern African Bantu languages have replaced widespread inherited Bantu terms for
‘cattle’ and ‘sheep’ with loans from Central Khoisan.
Table 3 tabulates the *guu root for ‘sheep’ in Khoisan and a sample of comparative Bantu forms. This term
is apparently borrowed into the Southern Bantu languages, replacing other roots such as #-kòòkò and #-
méémé. Reflexes of #-gu occur in Bantu zones K, R and S.
Table 3. The *guu root for ‘sheep’ in Khoisan and Southern Bantu
Phylum Branch Group Language Attestation
Khoisan Northern Ju/h’oan gùú
Central Proto-Khoe *guu
Khoekhoe Khoekhoegowab guu
Khoe //Ani, /Ui gû
Khoe Khwe gùu
Naro Naro gùu
//Ana /Ui gǔ
Shua Cara gù
Kwadi Kwadi guu-
Southern !Xóõ kūu
!Ui N|uu gǂaɾu pl. ǀoaχu
Bantu Kimbundu nguli
Xhosa igusha3
Venda ǹngú
Setswana nku
Sources: Tanaka (1978), Voßen (1997, 2007), Haacke & Eiseb (1999),
Visser (2001), Kilian-Hatz (2003), Sands et al. (2007)
The occurrence of this root in !Xóõ and Ju/h’oan must be treated as loans, since the time-depth of sheep is
too shallow to account for genuine cognates. As Vossen (2007: Table 3) shows, Ju/h’oan appears to have
many of the same livestock terms as Khoe. More puzzling is the case of Kwadi. Information on this
language can almost certainly never be expanded so we depend on Westphal’s manuscript notes. Both the
‘cattle’ and ‘sheep’ terms closely resemble those in Khoe, and yet other elements of the language are quite
distinct, arguing for a long period of separation from central Khoe (Güldemann 2004). This is rather in
contradiction with the known facts about dates for livestock in the archaeological record. There are two
possibilities to explain this situation; either the livestock names are borrowed (which would explain their
near-identity with Khoe) or the idiosyncratic restructuring of Khoe has taken place within a surprisingly
short time. Maho (2000) has compiled the names for sheep in sources for now-extinct Tuu languages which
demonstrate a wide variety of unrelated roots. This suggests rather strongly that sheep were unknown to
speakers of proto-Tuu but that they developed names from a diversity of lexical sources as a result of seeing
sheep. Westphal (1963:254) provides further evidence for the extension of this root in Bantu.
The most widespread Khoisan term for ‘goat’ is shown in Table 4, together related Bantu terms. The nature
of the link with Bantu is still obscure. The evidence from Bantu (documented by Guthrie but excluded from
BLR3) is for a widespread root, *-mpene, scattered from Eastern to Southern Africa.
3 Westphal (1963:254) notes that Xhosa adopted the name with the feminine Khoe suffix –s attached, confirming the
direction of borrowing.

8
Table 4. The –mpene root for goat terms in Khoisan and Bantu
Group Language Attestation
Khoisan Ju Ju/h’oan párí
Khoekhoe Khoekhoegowab piri
Khwe míní
Naro, G//ana, G/wi piri
Tuu Western N|uu piɾi ‘ram’
Eastern N|uu miɾi
|Xam puli
Bantu Setswana pʰelau ‘ram’
Rundi im-pene
Hima em-pene
Gogo hmene
Luguru im-hene
Bena imene
Nyakyusa em-bene
Yeyi impʰèné
Sources: Bleek (1956), Guthrie (1967-1971), Tanaka (1978), Dickens (1994),
Haacke & Eiseb (1999), Visser (2001), Kilian-Hatz (2003)
The Khoisan terms are probably cognate with the Bantu root, although the vowel-raising e→i is hard to
explain. The Khwe form, míní, also suggests a distinct borrowing from a word with initial mp-. There are,
however, also #puli forms found in Tuu languages, which resemble more closely the widespread Bantu root
*-bʊdi and which probably are direct borrowings. The whole complex suggests interference and re-analysis
of two distinct roots.
Another shared root is *-kumbo, found in both !Xun and neighbouring Bantu languages (Table 5).
Table 5. The –gumbo root for goat terms in Northern Khoisan and Bantu
Group Language Attestation
Khoisan !Xun gkhúmbō
Kwadi khobo
Bantu Kimbundu hombo
Manyo shikômbo
Herero ongómbó
Ndo oshikombo
Sources: Guthrie (1967-1971), König & Heine (2008), Möhlig (p.c.)
Given its shallow extension, the Khoisan term is most likely borrowed from the neighbouring Bantu
languages.
Terms for ‘dog’ in Khoe have no clear source, although the dog presumably appeared at roughly the same
era. Existing terms point to two distinct roots,
ʔ
ari and
ʔ
aba, but no etymologies have been proposed for
these (Voßen 1997: 453). Southern African Bantu languages have replaced the usual Bantu root #bʊ
́à for
‘dog’, but their terms do not resemble Khoe. Westphal (1963:254) also points out that the term for ‘sour
milk’ (i.e. yoghurt) appears to be borrowed from Khoe languages into Southern Bantu.
Other livestock terms reconstructed to different levels of Khoe by Vossen (2007:180) include *dubi ‘to
milk’, *!hada or *kada ‘cattle-kraal’4, *n//gubu ‘to churn’, *//ãũ ‘to fence in’, *gude ‘to herd’, *ts’ao or
*/x’ao ‘to milk into container’, *tsxôm ‘to milk into mouth’. Many of these also have cognates in Ju/h’oan.
The absence of any obvious etymology for these words is quite perplexing given that they are rather specific
to herding and thus cannot be of any great antiquity. It certainly suggests that the original pastoral
communities that adopted or developed these technologies were quite distinct from any surviving in the
present. It is the case that the four surviving Southern Cushitic languages are all very closely related and
4 Though see Sandawe hado.

9
have undergone a ‘lexical revolution’. It is likely that there were other pastoral Cushitic languages, perhaps
more related to the poorly documented Asax and Qwadza, which contributed to the pastoral culture of SW
Africa.
6. Genetic evidence
Attempts to use genetics to unravel the history of populations such as the Hadza go back to the era of blood-
sampling, but results have been inconclusive or contradictory. Recent work, sampling Hadza, Sandawe and
Southern Africa Khoisan, concludes that the divergence times between Eastern African click speakers and
those in the south is very ancient (>35,000 years) and that even between Hadza and Khoisan, divergence
times could be as much as 15,000 years (Tishkoff et al. 2007). Chen et al. (2000) explored the degree of
differentiation between !Xun and Khoe and concluded both that Khoisan populations exhibit ancient
lineages consistent with present dates for the evolution of modern humans and that Khoe populations have
much more in common genetically with other African populations than !Xun. Indeed, Knight et al. (2003)
argue that clicks must be traced back to the original language of mankind. Güldemann & Stoneking (2008)
rightly argue that this is not a valid hypothesis and that contact phenomena and more recent language
evolution can just as easily explain the present observable state of affairs. More directly germane to the
hypothesis of this paper is recent work on pastoral populations of Southwest Angola by Coelho et al. (2008).
They looked at lactase persistence genes and concluded that there one distinctive mutation, -14010, was
brought directly from the East to Namibe by people speaking Afro-Asiatic or Nilo-Saharan languages. Nilo-
Saharan can almost certainly be excluded, but the connection with the Horn of Africa is intriguing, if far
from proven.
7. Synthesis and chronological scenario
Khoe-speakers of south-western Africa and their northern Bantu pastoralist neighbours share features of
their culture with the pastoral peoples of NE Africa. Archaeology has shown that pottery, sheep and cattle
appear in the archaeological record prior to the putative expansion of the Bantu into this region. Numerous
wayward explanations were advanced for this in the earlier literature, both connecting the ‘Hottentots’ with
the spurious cultural category of Hamites and proposing the Khoe were the offspring of miscegenation with
seagoing Semites. However, it cannot be an ‘indigenous’ development; the breeds of cattle and sheep are
only otherwise found in Northeast Africa.
While some of the more bizarre proposals can be summarily dismissed, the broader problem remains; how
was pastoral culture transmitted to SW Africa? The paper proposes that this was a consequence of the
interaction of Khoe speakers and Cushitic pastoralists in a location intermediate between their present area
of distribution. To simplify the model, Figure 6 depicts this in Central Zambia, a region now entirely
occupied by Bantu speakers. Cushitic pastoralists would formerly have spread down through Central Africa,
at least as far as Zambia/Northern Zimbabwe, probably intermixed with hunter-gatherers. However, given
the likely ethnolinguistic complexity of both the foragers and pastoralists, a more complex set of interactions
is probable. Certainly the pattern of innovation, borrowing and re-interpretation of livestock terms set out in
§5 points to this complexity. However, it is important to state that the pastoral communities that brought
livestock to the region would have herded fat-tailed sheep and longhorn taurine cattle and known how to
make pottery. They would not have been iron-users but would have hunted using microlithic points and be
associated with the thin-walled ware of the CLSA, dated to earlier than 2000 BP.
About 2000 years ago these two groups encountered one another and the pottery skills and livestock breeds
were passed between them along with associated material culture such as mat huts, sandals and butter-
making equipment. A diverse pastoral culture would have existed in this intermediate zone, observed by San
hunter-gatherers who both traded with the herders and painted their animals. This explains why the typical
Khoe terms for domestic animals also occur in Northern (Ju) and Southern (Tu) languages. Subsequently,
the Bantu southwards expansion and from Tanzania to Zimbabwe assimilated or incorporated the Cushitic
pastoral culture. The distinctive animal breeds became heavily crossbred and the languages disappeared or
survived only as substrates. Figure 6 presents a hypothetical map illustrating this overlap and the possible
zones of interaction;

10
Figure 6. Map showing potential overlaps between Cushitic and Khoe speakers
Current
southern limit
of Cushitic
speakers
Possible original
southern limit of
Cushitic speakers
Main modern-day nucleus
of Khwe speakers
Possible northern
limit of Khwe
speakers 2500
BP
Main
modern-d ay
nucleus of
Cush itic
speakers
?
Possible pre-Khwe
northern ex tension
The distinctive terminology of livestock and its production in Khoe has few likely Cushitic etymons despite
the very evident ‘Cushitic’ features of pastoral systems in SW Africa. This argues that there was a ‘lost’
branch of the Cushitic family whose speakers encountered the early Khoe. The diffusion of animal names to
Kwadi and the languages of both the Northern and Southern branches with little or no phonological
alteration suggests this was a relatively recent process.
As the Bantu encountered the mixed Khoe-Cushites in western Zambia/Angola, different processes of
cultural assimilation occurred for reasons as yet unclear. Language shift to Bantu took place, but much more
of the NE African pastoral culture was retained, including features lost among the Khoe, at least by the time
of the first European incursions. Hence some features of the culture of the Namibia/Angola pastoralists
(which survives relatively intact) can be more obviously identified with the Cushites in the Horn of Africa.
Significant borrowing of livestock terms into Southern African Bantu languages from Khoe suggests that the
only livestock of importance in the earliest phase of Bantu expansion in this region was the goat. The
predecessors of the Zulu and others must have encountered pastoral Khoe and borrowed intensively from
their culture including skills such as yoghurt production. Only later would the zeboid cattle, now
predominant in herds throughout the region, have replaced the longhorn taurines originally herded by the
Khoe.
References
Adametz, L. 1920. Herkunft und Wanderung der Hamiten. Wien: Forschungsinstitut für Osten und Orient.
Badenhorst, S. 2006. Goats (Capra hircus), the Khoekhoen and Pastoralism: Current Evidence from
Southern Africa. African Archaeological Review, 23:45-53.
Barnard, Alan 1992. Hunters and herders of Southern Africa: a comparative ethnography of the Khoisan
peoples. Cambridge: Cambridge University Press.
Baumann, H. 1940. Volkerkunde von Afrika. Essen: Essener Verlagsanstalt.
Bleek, Dorothea F. 1956. A Bushman dictionary. New Haven: American Oriental Society.

11
Blench, R.M. 1993. Ethnographic and linguistic evidence for the prehistory of African ruminant livestock,
horses and ponies. in: The Archaeology of Africa. Food, Metals and Towns. Shaw, T., Sinclair, P., Andah,
B. and Okpoko, A. eds. pp. 71-103. London: Routledge.
Blench, R.M. 2006. Archaeology, Language and the African past. Lanham: Altamira Press.
Boonzaier, E., C. Malherbe, P. Bernes, and A. Smith 1996. The Cape Herders. A History of the Khoikhoi of
Southern Africa. Cape Town, South Africa: David Philip.
Bousman, C.B. 1998. The chronological evidence for the introduction of domestic stock into southern Africa.
African Archaeological Review, 15(2): 133–50.
Chen, Yu-Sheng et al. 2000. mtDNA Variation in the South African Kung and Khwe—and Their Genetic
Relationships to Other African Populations. American Journal of Human Genetics, 66(4): 1362–1383.
Coelho, Margarida, Fernando Sequeira, Sandra Beleza & Jorge Rocha. 2008. On the edge of the Bantu
expansions: MtDna, Y-Chromosome and lactase genetic variation in the Namibe Desert of Angola.
Conference poster.
Dickens, P. 1994. English-Ju/’hoan / Ju/’hoan-English Dictionary (Research in Khoisan Studies 8). Köln:
Rüdiger Köppe.
Ehret, C. 1982. The first spread of food production to Southern Africa. in C. Ehret & M. Posnansky eds. The
archaeological and linguistic reconstruction of African history, pp. 158-181. Berkeley: University of
California Press.
Ehret, C. 1986. Proposals for Khoisan reconstruction. Sprache und Geschichte in Afrika, 7(2): 105-130.
Ehret, C. 1998. An African classical age: eastern and southern Africa in world history, 1000 B.C. to A.D.
400. Charlottesville, VA: University Press of Virginia.
Ehret, Christopher & Derek Nurse 1981. The Taita Cushites. Sprache und Geschichte in Afrika, 3:125-68.
Elderkin, E.D. 1986. Diachronic inferences from basic sentence and noun phrase structure in Central
Khoisan and Sandawe. In: F. Rottland & R. Voßen eds. African Hunter-Gatherers. Sprache und
Geschichte in Afrika, 7(2):131-156.
Elphick, R. 1985. Khoikhoi and the founding of white South Africa. Johannesburg: Ravan.
Epstein, H. 1971. The origin of the domestic animals of Africa. (2 vols) New York: Africana Publishing
Corporation.
Estermann, Carlos 1950. Le bétail sacré chez quelques tribus du Sud-ouest a l’Angola. Anthropos, 45:721-
732.
Fauvelle-Aymar, François-Xavier 2004. Between the first herders and the last herders: are the Khoekhoe
descendants of the Neolithic ‘hunters-with-sheep’? Before Farming, 2004/4 article 5.
Frobenius, Leo 1933. Kulturgeschichte Afrikas. Phaidon-Verlag: Zürich.
Gifford-Gonzalez, Diane 2000. Animal Disease challenges to the emergence of pastoralism in Sub-Saharan
Africa. African Archaeological Review, 17(3):95-139.
Goodall, E. 1946. Domestic animals in rock art, and some observations on Rock Paintings illustrating Burial
Rites. Transactions of the Rhodesia Scientific Society, XLI.
Greenberg, Joseph H. 1963. Languages of Africa. (Indiana University Research Centre in Anthropology,
Folklore, and Linguistics, Publication 25). The Hague: Mouton & Co.
Güldemann, Tom & Rainer Vossen 2000. Khoisan. In: African languages: an introduction. B. Heine & D.
Nurse eds. 99-122. Cambridge: Cambridge University Press.
Güldemann, Tom 2004. Reconstruction through ‘de-construction’: the marking of person, gender and
number in the Khoe family and Kwadi. Diachronica, 21(2):251-306.
Güldemann, Tom in press. A linguist’s view: Khoe-Kwadi speakers as the earliest food-producers of
southern Africa. Special Issue of Southern African Humanities, Karim Sadr and François-Xavier
Fauvelle-Aymar eds. Pietermaritzburg: Natal Museum.
Güldemann, Tom n.d. The San languages of southern Namibia: a linguistic appraisal with special reference
to Krönlein's N|uusaa manuscript. Unpublished paper downloadable from the author’s website.
Güldemann, Tom and Edward D. Elderkin in press. On external genealogical relationships of the Khoe
family. In Brenzinger, Matthias and Christa König (eds.), Khoisan languages and linguistics: the
Riezlern symposium 2003. Quellen zur Khoisan-Forschung 17. Köln: Rüdiger Köppe.
Güldemann, Tom & Mark Stoneking 2008. A Historical Appraisal of Clicks: A Linguistic and Genetic
Population Perspective. Annual Review of Anthropology, 37: 93-109.
Guthrie, M. 1967-71. Comparative Bantu. (4 vols.) Farnborough: Gregg.
Gutierrez, Manuel 1996. L’Art pariétal de l’Angola. Paris: L’Harmattan.
Gutierrez, Manuel 2008. L’art rupestre de la province Namibe, Angola: étude des sites par rapport à leur
position topographique. Journal of African Archaeology, 6 (1): 131-149.
Haacke, Wilfrid H. G. & Eliphas Eiseb 1999. A Khoekhoegowab Dictionary, with an English-
Khoekhoegowab Index. Windhoek: Gamsberg Macmillan.

12
Hauenstein, Alfred 1980. Rites et coutumes lies à l’élevage du bétail dans le sud de l’Angola. St Augustin:
Anthropos Institut.
Henshilwood, C. 1996. A revised chronology for pastoralism in southernmost Africa: new evidence of sheep
at c. 2000 b.p. from Blombos Cave, South Africa. Antiquity, 70: 945-949.
Hollmann, J. C. 1993. Preliminary report on the Koebee rock paintings, western Cape Province, South
Africa. South African Archaeological Bulletin 48: 16-25.
Honken, Henry in press. Eastern ǁHoã as a Northern Khoisan language. In: Essays in Honour of Anthony
Traill. K. Keuthmann, G. Sommer & R. Voßen eds. Köln: Rüdiger Köppe.
Irle, J. 1906. Die Herero. Gütersloh: G. Bertelsmann.
Jeffreys, M.D.W. 1968. Some Semitic influences in Hottentot culture. Raymond Dart Lectures, 4.
Johannesburg: Witwatersrand University Press.
Kilian-Hatz, Christa 2003. Khwe dictionary. Köln: Rüdiger Köppe.
Kinahan, John 1991. Pastoral nomads of the Central Namib desert. Windhoek: New Namibia Books.
Knight, A., Underhill, P.A., Mortensen, H.M., Zhivotovsky, L.A., Lin, A.A., Henn, B.M., Louis, D., Ruhlen,
M. & Mountain, J.L. 2003. African Y chromosome and mtDNA divergence provides insight into the
history of click languages. Current Biology, 13:464-473.
Kolb, P. 1731. The present state of the Cape of Good Hope. 2 vols. London: W. Innys.
König, C. & B. Heine 2008. A concise dictionary of Northwestern !Xun. Köln: Rüdiger Köppe.
Kose & Richter 2007. The prehistory of the Kavango people. Sprache und Geschichte in Afrika, 18:103-129.
Köln: Rudiger Köppe.
Lebzelter, V. 1934. Eingeborenenkulturen in Südwest-und-Süd Afrika. 2 vols. Leipzig: K.W. Hiersemann.
Loeb, Edwin M. 1962. In Feudal Africa. International Journal of American Linguistics, 28(3). [book-length
single issue]
Magnavita, Carlos 2006. Ancient Humped Cattle in Africa: A View from the Chad Basin. African
Archaeological Review, 23:55–84.
Maho, Jouni Filip 2000. Comparative wordlist of Kwi (South Khoesan) languages - part 1 (PDF) and part 2
(PDF). URL: www.african.gu.se/maho/downloads/kwilexis1.pdf
Manhire, A.H., J.E. Parkington, A.D. Mazel, T.M. Maggs 1986. Cattle, sheep and horses: a review of
domestic animals in the rock art of southern Africa. South African Archaeological Society Goodwin
Series.
Meinhof, Carl 1912. Die Sprachen der Hamiten. Abhandlungen des Hamburgischen Kolonialinstituts 9.
Hamburg: L. Friederichsen.
Mitchell, P. 2002. The archaeology of Southern Africa. Cambridge: Cambridge University Press.
Nienaber, G. S. 1963. Hottentots. Pretoria: J. L. van Schaik Beperk.
Nurse, G.T., Weiner, J.S. & T. Jenkins 1985. The peoples of Southern Africa and their affinities. Oxford:
Clarendon Press.
Oliver, P. ed. 1971. Shelter in Africa. London: Barrie & Jenkins.
Prussin, Labelle ed. 1995. African nomadic architecture. Washington & London: Smithsonian Institution
Press.
Robbins, L.H. et al. 2008. Recent archaeological research at Toteng, Botswana: early domesticated livestock
in the Kalahari. Journal of African Archaeology, 6 (1): 131-149.
Robertshaw, Peter 1990. Early pastoralists of South-Western Kenya. Nairobi: British Institute in Eastern
Africa.
Sadr, K. 2008. An ageless view of the first millennium AD Southern African ceramics. Journal of African
Archaeology, 6 (1): 235-252.
Sadr, K. & C.G. Sampson 2006. Through thick and thin: early pottery in Southern Africa. Journal of African
Archaeology, 4 (2): 235-252.
Sands, Bonny, Amanda L. Miller, and Johanna Brugman. 2007. The Lexicon in Language Attrition: The
Case of N|uu. Selected Proceedings of the 37th Annual Conference on African Linguistics, ed. Doris L.
Payne and Jaime Peña, 55-65. Somerville, MA: Cascadilla Proceedings Project.
Sealy, J. & Yates, R. 1994. The chronology of the introduction of pastoralism to the Cape, South Africa.
Antiquity, 68: 58–67.
Smith, A.B. 2000. The origins of the domesticated animals of southern Africa. In: The origins and
development of African livestock: archaeology, genetics, linguistics and ethnography. R.M. Blench & K.
C. MacDonald (eds.), 222–238. London: UCL Press.
Smith, A.B. 2005. African herders: emergence of pastoral traditions. Walnut Creek, CA: Alta Mira Press.
Smith, A.B. 2006. Excavations at Kasteelberg, and the Origins of the Khoekhoen in the Western Cape,
South Africa. Oxford: British Archaeological Reports International Series 1537.

13
Smithies, M. 1999. A Siamese embassy lost in Africa 1686: the odyssey of Ok-Khun Chamnan. Chiang Mai:
Silkworm Books.
Stow, G.W. 1905. The native races of South Africa: a history of the intrusion of the Hottentots and Bantu
into the hunting grounds of the bushmen, the Aborigines of the country. London: Swan Sonnenschein.
Tanaka, Jiro 1978. A San vocabulary of the Central Kalahari: G//ana and G/wi dialects. Tokyo: ILCAA.
Tishkoff, Sarah A. et al. 2007. History of Click-Speaking Populations of Africa Inferred from mtDNA and Y
Chromosome Genetic Variation. Molecular Biology and Evolution, 24(10):2180-2195.
Visser, Hessel. 2001. Naro dictionary. D’Kar: Naro Language Project.
Vogelsang, Ralf & Karl Peter Wendt 2007. Changing settlement patterns in Namibia during the Holocene.
In: Bubenzer, O., Bolten, A. & Darius F. eds. Atlas of cultural and environmental change in Africa. 68-
71. Köln: Heinrich-Barth Institut.
Voßen, Rainer 1997. Die Khoe-sprachen. Köln: Rudiger Köppe.
Voßen, Rainer 2007. Languages of the desert…and what they can tell us about the economic history of
southern Africa. Sprache und Geschichte in Afrika, 18:175-185. Köln: Rudiger Köppe.
Westphal, E.O.J. 1962. A reclassification of Southern African non-Bantu languages. Journal of African
languages, 1:1-8.
Westphal, E.O.J. 1963. The linguistic prehistory of Southern Africa: Bush, Kwadi, Hottentot and Bantu
linguistic relationships. Africa, 33:237-265.