. 1
( 11)


African Archaeology

Research in Africa is now accepted as an integral part of global archaeological
studies. As well as providing archaeologists with their oldest material, Africa is
also widely recognised as the birthplace of modern humans and their charac-
teristic cultural patterns. Archaeological study of later periods provides unique
and valuable evidence for the development of African culture and society, while
ongoing research in Africa provides insights relevant to the interpretation of
the archaeological record in other parts of the world. In this fully revised
and expanded edition of his seminal archaeological survey, David Phillipson
presents a lucid and fully illustrated account of African archaeology from pre-
history and the origins of humanity to the age of European colonisation. The
work spans the entire continent from the Mediterranean to the Cape of Good
Hope and demonstrates the relevance of archaeological research to the under-
standing of Africa today.

DAV I D W. P H I L L I P S O N F B A is Professor of African Archaeology and
Director of the Museum of Archaeology & Anthropology at the University of
Cambridge. He is a Fellow of Gonville & Caius College, Cambridge.
Frontispiece The principal physical features of Africa
African Archaeology
Third edition

University of Cambridge
©¤§ µ®©© °
Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo

Cambridge University Press
The Edinburgh Building, Cambridge  µ, UK
Published in the United States of America by Cambridge University Press, New York
Information on this title: www.cambridge.org/9780521832366

© Cambridge University Press 1985, 1993, 2005

This book is in copyright. Subject to statutory exception and to the provision of
relevant collective licensing agreements, no reproduction of any part may take place
without the written permission of Cambridge University Press.

First published in print format 2005

©®-± ·-°-µ±±-±±·- eBook (NetLibrary)
©®-±° °-µ±±-±±·- eBook (NetLibrary)

©®-± ·-°-µ±-- hardback
©®-±° °-µ±--µ hardback

©®-± ·-°-µ±-µ°°-µ paperback
©®-±° °-µ±-µ°°- paperback

Cambridge University Press has no responsibility for the persistence or accuracy of
µ¬s for external or third-party internet websites referred to in this book, and does not
guarantee that any content on such websites is, or will remain, accurate or appropriate.
To Laurel

List of illustrations ix
Sources of illustrations xiii

Preface xv

1 1
Elucidating the African past 1
Archaeology in Africa 3
Linguistics 6
Oral traditions 9
Ethnoarchaeology 10
Africa in world prehistory 10

The emergence of humankind in Africa
2 15
De¬nition and process 15
World-wide precursors of the hominids 20
The earliest hominids 22
The oldest discoveries in eastern Africa 32
The Lake Turkana Basin and Olduvai Gorge 34
Central and south-central Africa 42
South Africa 42
The earliest tool-makers 47

The consolidation of basic human culture
3 52
Acheulean and Sangoan in Africa 52
Acheulean in eastern Africa 60
Acheulean in south-central Africa 68
Acheulean in southern Africa 73
Acheulean in West Africa and the Sahara 75
Acheulean in North Africa 77
Sangoan assemblages 81
Acheulean/Sangoan artefacts and their makers 84

Regional diversi¬cation and specialisation
4 91
The ˜Middle Stone Age™ and the ˜Late Stone Age™ 91
Southern Africa 96
South-central Africa 108
Rock art in southern and south-central Africa 111
Central Africa 116
Eastern Africa 122

viii Contents

West Africa 128
North Africa and the Sahara 131
Changing life-styles and technology 141

The beginnings of permanent settlement
5 147
North Africa, the Sahara and the Nile Valley 147
East Africa 156
Overview 159
African peoples 10,000 years ago 160

Early farmers
6 165
Cultivation and herding 165
The Sahara and North Africa 172
The Nile Valley 181
West and Central Africa 195
Ethiopia and the Horn 203
East Africa 206

Iron-using peoples before ad 1000
7 214
Iron 214
North Africa 216
Egypt and the Arab invasion 221
The Sudan 224
Ethiopia and adjacent regions 228
West Africa 234
Central Africa 245
Eastern and southern Africa 249
The contribution of Bantu linguistic studies 261
Mode of dispersal 265
Madagascar and the Comoro Islands 269
Stone-tool-using herders of southwestern Africa 269

The second millennium ad in sub-Saharan Africa
8 274
The last 1000 years 274
West Africa 275
Ethiopia, the southern Sudan and adjacent regions 284
The east coast of Africa 288
Bantu-speakers north of the Zambezi 291
Southeastern Africa 297
Southwestern Africa 307
Epilogue 308

Bibliographic guide 310

Bibliographic references 311
Index 369

Frontispiece The principal physical features of Africa ii
Fig. 1 The classi¬cation of recent African languages 7
Fig. 2 Africa™s major language families in recent times 7
Fig. 3 Modes of lithic technology 12
Fig. 4 Geological periods of the last 24 million years 21
Fig. 5 Classi¬cation of the order Primates 21
Fig. 6 Tentative ˜family tree™ of African hominids 23
Fig. 7 Geological, magnetic and oxygen isotope subdivisions 25
Fig. 8 Climate and phases of accelerated evolution 26
Fig. 9 Skulls from Sterkfontein, Olduvai and Koobi Fora 30
Fig. 10 Skeletons of gorilla, Australopithecus and Homo sapiens 31
Fig. 11 Sites of Australopithecus, Homo habilis and Oldowan artefacts 33
Fig. 12 Hominid footprints at Laetoli dated to about 3.8 million years ago 35
Fig. 13 Mode-1 artefacts from Hadar, Koobi Fora and Omo Valley 36
Fig. 14 Stratigraphy of the Koobi Fora Formation 37
Fig. 15 Olduvai Gorge 40
Fig. 16 Stone circle at site DK in Bed I, Olduvai Gorge 41
Fig. 17 Breccia and travertine at Taung 43
Fig. 18 Ages of the South African australopithecine sites 45
Fig. 19 Horn and bone tools from Swartkrans 46
Fig. 20 Oldowan artefacts from site DK, Olduvai Gorge 48
Fig. 21 Acheulean-type artefacts from Montagu Cave 54
Fig. 22 Principal Acheulean and related sites in Africa 56
Fig. 23 Core preparation 57
Fig. 24 Air view of the Victoria Falls 59
Fig. 25 Artefacts of the Karari industry from Koobi Fora 61
Fig. 26 Homo ergaster from the Lake Turkana Basin 61
Fig. 27 Stone artefacts from Bed II at Olduvai Gorge 63
Fig. 28 The archaeological sequence at Olduvai Gorge 64
Fig. 29 The Bodo skull 65
Fig. 30 Acheulean-type artefacts preserved at Olorgesailie 66
Fig. 31 The 220-metre-high Kalambo Falls 69
Fig. 32 Wooden objects from Acheulean levels at Kalambo Falls 71
Fig. 33 Skull from Broken Hill Mine, Kabwe 72
Fig. 34 Wonderwerk Cave 74
Fig. 35 The ˜Victoria West technique™ 75
Fig. 36 West African Acheulean-type artefacts 76
Fig. 37 Artefacts from Ain Hanech and Sidi Zin 79
Fig. 38 Mandible of Homo ergaster from Terni¬ne 80
Fig. 39 Deposits exposed at Sidi Abderrahman quarry 80
Fig. 40 The distribution of principal Sangoan sites 82
Fig. 41 Sangoan artefacts from Luangwa Valley and Kalambo Falls 83

x List of illustrations

Fig. 42 Early traditions of lithic technology 85
Fig. 43 Mounted microliths 93
Fig. 44 The Klasies River Mouth sites before excavation 98
Fig. 45 Mode-3 and mode-5 artefacts from southern Africa 99
Fig. 46 Mode-3 and related industries in southernmost Africa 101
Fig. 47 Ochre with carved lines, from Blombos Cave 102
Fig. 48 Sibudu Cave 103
Fig. 49 Kalemba rockshelter during excavation 106
Fig. 50 Early mode-5 artefacts from Kalemba 109
Fig. 51 Artefacts from Gwisho 110
Fig. 52 Early paintings from ˜Apollo 11 Cave™, Namibia 112
Fig. 53 Naturalistic rock paintings at Makwe, Zimbabwe 113
Fig. 54 Rock painting at Mpongweni, KwaZulu-Natal 114
Fig. 55 Line-engraving from Doornkloof, South Africa 115
Fig. 56 Pecked engravings, Klipfontein, Northern Cape 115
Fig. 57 Rock paintings showing people in trance 116
Fig. 58 South African schematic rock paintings 117
Fig. 59 Occurrences of post-Acheulean/Sangoan stone industries 119
Fig. 60 Lupemban artefacts 123
Fig. 61 Artefacts from Gobedra and Gamble™s Cave 126
Fig. 62 Mode-3 artefacts from Nigeria 129
Fig. 63 West African mode-5 artefacts 130
Fig. 64 Aterian artefacts from Bir el Ater and Adrar Bous 132
Fig. 65 Artefacts from Haua Fteah 135
Fig. 66 Oranian and Capsian artefacts 137
Fig. 67 Capsian carvings on stone and ostrich-eggshell 138
Fig. 68 Khormusan and Halfan artefacts from Nubia 140
Fig. 69 Location of settlement sites discussed in chapter 5 148
Fig. 70 Rock engraving of long-horned Bubalus antiquus 151
Fig. 71 Stone structures at Ti-n-Torha rockshelter, Acacus 152
Fig. 72 Pottery from Amekni 153
Fig. 73 Khartoum-related and Shamarkian artefacts 154
Fig. 74 Artefacts from Early Khartoum 155
Fig. 75 Old beach deposits at Lowasera 157
Fig. 76 Artefacts from Lowasera 158
Fig. 77 Location of rock art, early cultivation and herding 166
Fig. 78 Areas of domestication of indigenous African crops 168
Fig. 79 Rock engravings at Jebel Uweinat 174
Fig. 80 Artefacts from Adrar Bous 177
Fig. 81 Round-headed painted ¬gure, Tassili 178
Fig. 82 Rock painting of a pastoral scene, Tassili 179
Fig. 83 Dhar Tichitt stone enclosures 180
Fig. 84 Artefacts from Esh Shaheinab 182
Fig. 85 A-Group artefacts from the Wadi Halfa area 184
Fig. 86 Brick substructure of royal burial mound at Kerma 186
Fig. 87 The temples at Abu Simbel 187
Fig. 88 Pre-Dynastic Egyptian artefacts 188
Fig. 89 The chronology of ancient Egypt 190
Fig. 90 Ancient Egyptian hieroglyphic writing 191
List of illustrations xi

Fig. 91 Egyptian royal tombs 192
Fig. 92 Relief carving at Kalabsha, Nubia 193
Fig. 93 The Eighteenth-Dynasty pharaoh, Tuthmosis III 194
Fig. 94 Sorghum 196
Fig. 95 Traditional yam-storage in eastern Nigeria 197
Fig. 96 Shell middens on the coast of Senegal 199
Fig. 97 Artefacts from Kintampo sites 200
Fig. 98 Excavation through deep deposits at Daima 201
Fig. 99 Rock paintings of domestic cattle 205
Fig. 100 Representations of Punt and its inhabitants 206
Fig. 101 Stone bowl from North Horr 207
Fig. 102 Artefacts from Hyrax Hill, Kenya 209
Fig. 103 Nderit ware pottery 210
Fig. 104 Traditional African iron-smelting, Ghana 215
Fig. 105 Rock painting of a horse-drawn chariot, Acacus 217
Fig. 106 Greek and Phoenician colonies in North Africa 218
Fig. 107 Ruins of the Roman city of Timgad, Algeria 220
Fig. 108 Septimius Severus 220
Fig. 109 Map of Africa by Ptolemy of Alexandria 222
Fig. 110 The mosque of Sidi Okba at Qayrawan 223
Fig. 111 Meroitic temple at Naqa 225
Fig. 112 Meroitic cursive script 226
Fig. 113 Early metal-working and modern vegetation 227
Fig. 114 Fresco from Faras cathedral 228
Fig. 115 Pre-Aksumite altar from Addi Gelemo, Ethiopia 229
Fig. 116 Aksumite gold coin of King Endybis 230
Fig. 117 The third-largest Aksumite stela 231
Fig. 118 Artefacts from Jebel Moya 233
Fig. 119 Nok terracotta heads 235
Fig. 120 Reconstruction of a Taruga iron-smelting furnace 236
Fig. 121 Lost-wax bronze castings from Igbo Ukwu 239
Fig. 122 Early metal in West and central Africa 240
Fig. 123 Location and plan of Jenne-Jeno 242
Fig. 124 Megalithic stone circle at Sine Saloum, Senegal 243
Fig. 125 Pottery vessel and stone axes from Batalimo 246
Fig. 126 Pits at Oveng, Gabon 247
Fig. 127 Urewe ware from sites in southwestern Kenya 250
Fig. 128 Sites of the Chifumbaze complex 252
Fig. 129 Chifumbaze-complex pottery from Malawi 255
Fig. 130 Schroda, viewed from an overlooking cave 257
Fig. 131 Terracotta head from Lydenburg 259
Fig. 132 The distribution of Bantu languages 263
Fig. 133 Rock painting at Silozwane Cave, Zimbabwe 267
Fig. 134 Schematic rock painting at Sakwe, Zambia 268
Fig. 135 Sites in southwestern Africa with early sheep remains 270
Fig. 136 Cape coastal pottery 271
Fig. 137 Kasteelberg 272
Fig. 138 Plan of stone-built houses at Tegdaoust 277
Fig. 139 West African sites and kingdoms 278
xii List of illustrations

Fig. 140 Trumpet and trumpet-blower from Begho 280
Fig. 141 Terracotta heads from Ife 281
Fig. 142 Linear earthworks in the region of Benin 282
Fig. 143 Benin ˜bronzes™ 283
Fig. 144 Second-millennium sites in eastern and southern Africa 286
Fig. 145 The rock-cut church of Abba Libanos, Lalibela 287
Fig. 146 Principal sites on the East African coast 289
Fig. 147 Ruins of a mosque at Gedi 291
Fig. 148 Salt crystallising at Kibiro 293
Fig. 149 Luangwa-tradition pottery 294
Fig. 150 A Kisalian grave at Sanga 296
Fig. 151 Gold artefacts from Mapungubwe 298
Fig. 152 Clay-walled house excavated at Great Zimbabwe 299
Fig. 153 Inside the great enclosure at Great Zimbabwe 300
Fig. 154 Reconstruction of the main building at Nhunguza 302
Fig. 155 Copper cross-ingot from Ingombe Ilede 303
Fig. 156 Elaborate stone walling at Naletale 303
Fig. 157 Terracing and enclosure at Nyanga 304
Fig. 158 Stone-walled structures at Makgwareng, Free State 306
Fig. 159 Rock painting of an ox-drawn wagon 308
Sources of illustrations

I am most grateful to all who have assisted with the provision of illustrations.
The line drawings in Figs. 13, 20, 21, 25, 27, 36, 37, 45, 50, 51, 61--6, 68, 72--4,
76, 80, 84, 85, 88, 97 and 105 are owed to Dr Laurel Phillipson, while Figs.
55, 99, 101, 102, 115, 127, 129, 136, 149 and 154 are the work of the late
John Ochieng™. I am responsible for the drawings in Figs. 3, 4--8, 10, 14, 18,
23, 32, 35, 39, 41--3, 60, 90--2, 100, 120, 123, 138 and 142.
Acknowledgements are due to the following for permission to reproduce
photographs: Fig. 12 -- John Reader; Fig. 15 -- Cambridge University Press
(M. D. Leakey 1971); Figs. 17, 19 -- Dr Francis Thackeray, Transvaal Museum;
Fig. 24 -- Zambia Survey Department; Fig. 26 -- National Museums of Kenya;
Fig. 29 -- Leverhulme Centre for Human Evolutionary Studies, University
of Cambridge; Fig. 33 -- Trustees of the Natural History Museum, London;
Fig. 38 -- Mus©e de l™Homme (Archives de l™Institut de Pal©ontologie Humaine, 32);
Fig. 44 -- Dr John Wymer and the University of Chicago; Fig. 47 -- Image cour-
tesy of Christopher Henshilwood; Fig. 48 -- Professor Lyn Wadley; Fig. 52 --
Heinrich-Barth-Institut; Fig. 54 -- A. R. Willcox; Figs. 58, 159 -- Dr Benjamin
Smith, Rock Art Research Institute, University of the Witwatersrand;
Fig. 70 -- Dr Rudolf Kuper (Stehli 1978); Fig. 71 -- Professor Barbara Barich
(1987); Figs. 81, 82 -- Dr Karl-Heinz Striedter (Stehli 1978); Fig. 83 -- Dr Augustin
Holl (1989); Fig. 86 -- Museum of Fine Arts, Boston; Figs. 93, 112 -- Trustees
of the British Museum (ea986 and ea901 respectively); Figs. 98, 148 --
Professor Graham Connah (1976, 1996); Figs. 103, 118 -- Cambridge University
Museum of Archaeology and Anthropology (1931.755 and 1953.211, 303, 409
respectively); Fig. 104 -- Dr Len Pole; Fig. 107 -- Thomas Nelson and Sons Ltd
(A. A. M. van der Heyelen and H. H. Scullard, eds., Atlas of the Classical World,
1959); Fig. 108 -- Fitzwilliam Museum, University of Cambridge; Fig. 111 --
Dr Julie Anderson; Fig. 114 -- National Museum, Warsaw; Fig. 121 -- Profes-
sor Thurstan Shaw; Fig. 124 -- Professor Merrick Posnansky; Fig. 131 -- South
African Museum; Fig. 137 -- Dr Peter Mitchell; Figs. 141, 143 (left) -- Professor
Frank Willett; Fig. 143 (right) -- Glasgow Museums and Art Galleries; Fig. 150 --
Dr Francis Van Noten (1982); Fig. 151 -- Professor Andrie Meyer, University of
Pretoria; Fig. 157 -- Dr Robert Soper (2002); Fig. 158 -- Dr Tim Maggs (1976).
All other photographs are my own.

D. W. P.


The aim of this book, as of previous editions, is to provide an overview and
guide to an increasingly complex subject. As information accumulates, the
need for such an overview becomes even greater, not only for historians,
general Africanists and specialists in allied disciplines, or for archaeologists
of other continents, but also for Africanist archaeologists themselves, for
the days are fast disappearing when any one individual can be expected to
be familiar with all periods in all regions of that vast continent.
The book is intended not only to summarise data and their interpretation,
but also to serve as a guide to the literature. Unlike many broad surveys, the
book therefore contains a comprehensive bibliography. For each major topic
I have attempted to cite primary sources, even when their interpretation is
now outdated, together with the most recent contributions or evaluations.
When a full account of a piece of research has been published, earlier pre-
liminary reports are not cited. Where the full account is still awaited, or
where research is ongoing, important data may be scattered among many
preliminary reports. Where available, I have preferred to cite publications
in English as most readily accessible and comprehensible to the majority
of readers. I have not cited unpublished dissertations, conference papers or
contributions to informal newsletters. Despite the length of the bibliogra-
phy, much selectivity has been exercised in its compilation. It is preceded
by a brief bibliographic guide, which I hope readers will ¬nd useful.
As will rapidly become apparent to the reader, the quantity and quality of
research varies enormously between regions, countries, periods and topics.
I have tried to even things out. The result is that some parts of the narrative
(such as those on Angola) are based on a few minor discoveries, while others
(on ancient Egypt, for example) attempt brie¬‚y to summarise many decades
of intensive specialist research. This imbalance is inevitably re¬‚ected in the
In this book, use of geographical names has followed, wherever possible,
current African usage while seeking to maintain comprehensibility for the
non-African reader. Names of countries are given in their current form, with
the former Za¨re designated ˜D. R. Congo™ and its namesake across the river
simply as ˜Congo™. Care has been taken to obviate confusion between ancient
Ghana and the modern republic of the same name which has a different
geographical location, and between Benin City in Nigeria and the Republic
of B©nin. The name ˜Sudan™ with a capital ˜S™ refers to the modern Republic
xvi Preface

of the Sudan; spelled with a small ˜s™, ˜sudan™ refers to the open savanna
country which extends across Africa south of the Sahara and north of the
equatorial forest.
Names of provinces and geographical features have sometimes been
subject to repeated changes and foreigners can ¬nd this confusing. In
this book, old and new names are sometimes used together, as for the
uKhahlamba/Drakensberg mountains or Lake Edward/Rutanzige. Where a
new name has become accepted in common international usage, as in the
case of Lake Turkana, it is used alone. When referring to archaeological sites,
the form used at the time of investigation and ¬rst substantive publication
has been retained, so that ˜Terni¬ne™ is used rather than the new name
˜Tighenif™, and ˜Broken Hill™ rather than ˜Kabwe™. Similarly, although the
South African town formerly known as Pietersburg is now called Polokwane,
the eponymous stone-tool industry retains its old name.
This book is the result of four decades™ study and involvement in African
archaeology. The staff of Cambridge University Press have been unfailingly
helpful; an anonymous reader whom they engaged has made many sugges-
tions for the improvement of the text. Numerous friends and colleagues have
contributed to my knowledge and understanding although, needless to say,
I have not always followed their advice and all errors and omissions are
my sole responsibility. By far the greatest contribution has been made by
my wife, Dr Laurel Phillipson, whose support -- in addition to her own sub-
stantial contributions to archaeological research in Africa -- has permitted
my involvement, whose scholarship has contributed to almost every para-
graph of the book, and whose incisive but tactful criticism has immeasurably
improved it. With love and gratitude, the book is dedicated to her.

D. W. P.
1 October 2003
Elucidating the African past
International understanding of the African past has increased and improved
enormously during the past one hundred, more particularly the past ¬fty,
years. This understanding extends both throughout the continent itself and
far beyond, so that Africa™s contribution to the whole story of human devel-
opment and achievement is coming into focus. This is a matter for celebra-
tion, not for ignorance ( J. G. D. Clark 1961) or doubt and backward-looking
criticism (M. Hall 2002). It is a story to which the contribution of archaeology
is paramount and which needs to be told and understood both in African
and in world-wide contexts (D. W. Phillipson 2003b).
This book attempts to provide an up-to-date summary and interpretation
of the archaeological evidence for the past of humans in Africa from their
¬rst appearance up to the time when written history becomes the primary
source of information. In chronological terms, this period covers all but
a tiny fraction of the time that the earth has had human inhabitants. It
now seems very probable that it was in Africa that humankind ¬rst evolved,
during the period between 5.0 and 2.5 million years ago. At the other end
of the time-scale, the earliest written records relating to Africa, those of the
ancient Egyptians, began about 5000 years ago; for many other parts of the
continent, notably the interior regions south of the equator, no such records
are more than one or two centuries old.
The period of time before written history is conventionally known as
prehistory. The term is not entirely appropriate in Africa, for a number
of reasons. First, there were long periods, especially in the northern part
of the continent, about which written records, although available, are not
generally informative on many aspects of contemporary life. There are also
numerous instances where the only available written records were produced
by outsiders and frequently give an incomplete account of matters which
the writers did not properly understand. These are situations which con-
front prehistorians in many parts of the world, but they give rise to par-
ticular problems in some parts of Africa because of the generally shallow
time-depth of indigenous literacy. A different approach is required to those
aspects of African culture which, to a very large extent, take the place of
written literature in other regions. These include the developed oral tradi-
tions which, in many societies, preserve the accumulated wisdom of the
2 afric an archaeolog y

people, including details of their past history. Again, language itself plays
a large part in determining a people™s or an individual™s sense of identity.
Where written examples of ancient languages do not exist, much can be
learned through the study of present-day linguistic forms and distributions
concerning the nature and interactions of past populations. The methods of
interpreting these sources of information about the African past will be dis-
cussed below, and an attempt made to link their evidence with that derived
from archaeology.
This book offers a particularly African perspective on the continent™s past.
Developments elsewhere are discussed only where they are directly relevant
to an understanding of African archaeology. For this reason, the most recent
periods in North Africa are excluded from consideration, since at that time
the regions north of the Sahara were ¬rmly part of the Mediterranean world.
Likewise, the archaeology of European colonialism in sub-Saharan Africa is
omitted from the present narrative.
Africa is a vast land-mass, representing approximately one ¬fth of the
habitable area of the globe. It is home to over 500 million people who,
despite ever-increasing urbanisation, practise a great variety of economies
and life-styles almost all of which are rooted in the continent™s diverse envi-
ronments and resources. The principal physical features of Africa are shown
in the frontispiece to this volume. It is useful to visualise African envi-
ronments as forming a series of roughly concentric zones, centred on the
low-lying equatorial rain-forest of the Congo Basin and the West African
coastlands. Around this, on the north, east and south, are belts of progres-
sively drier country, generally at increasing altitudes, with forest merging to
woodland savanna, to open grassland, to semi-desert or sahel. In the eastern
part of the continent, this zoning pattern is interrupted by highlands, the
Rift Valley system and its associated lakes, extending from Eritrea southward
to Lesotho and South Africa. It is in this eastern region that the continent™s
highest mountains lie; some retain year-round snow and glaciers. In both
northern and southwestern Africa are areas of true desert, where almost no
rain currently falls; both are low-lying but interrupted by mountain massifs.
Finally, at Africa™s northern and southern extremities, along the coast of the
Maghreb and around the Cape of Good Hope, are areas of Mediterranean
vegetation, while parts of the east coast are fringed by mangroves.
This pattern has not, of course, remained static. Throughout the time with
which this book is concerned, and continuing at the present, ¬‚uctuating
temperatures and patterns of rainfall have resulted in large-scale changes in
the extent and location of the environmental zones. At times, the equatorial
forest has shrunk to isolated enclaves, and the deserts have expanded and
become even more arid. On other occasions, the forests have been far more
extensive than they are today and the deserts have almost disappeared. All
Introduction 3

these processes have been accompanied by corresponding ¬‚uctuations in the
intervening savannas. During the earlier periods of humankind™s presence
in Africa, earth movements in the Rift Valley continued. These changes are
noted, albeit brie¬‚y, at appropriate places in the narrative which follows.
In view of the enormous time-span of African prehistory and the great vari-
ety of the human societies that have inhabited the continent, it is clear that
very varied methods have to be employed in elucidating the past. Studies
of linguistics and oral traditions are, of course, only applicable to relatively
recent periods. For the vast majority of the period of time with which this
book is concerned archaeology is our main and often our only source of
primary information about human activities.

Archaeology in Africa
Archaeological data provide a picture of the past which is essentially differ-
ent from, and in many ways complementary to, that which may be recon-
structed from written or oral sources. The archaeologist studying the mate-
rial remains of a pre-literate people will hardly ever be able to learn the
names or characters of individuals. He or she will often ¬nd it dif¬cult to
make more than very general inferences about social systems or political
situations. On the other hand, the archaeologist™s interpretation of techno-
logical skills and economic practices, such as hunting, agriculture or the
herding of domestic animals, will generally be far more complete and reli-
able than that which can be obtained by other types of research. For this
reason it is not just to the study of prehistory that archaeology can make an
important contribution; it is also an approach that greatly aids our under-
standing of more recent societies, even those for which abundant written
records are available.
To the student of Africa, the ¬ndings of archaeological research represent
a major source of information about the continent™s past; it is the principal
source for most of our understanding of prehistory, and it makes a signif-
icant contribution to knowledge about more recent periods. In Africa, the
shallow time-depths to which archaeological investigations may often be use-
fully applied greatly increase our historical perspective of recent trends and
events; and it may justly be claimed that our understanding of such press-
ing contemporary problems as deserti¬cation and tribalism is enhanced
through the results of archaeological research, with obvious implications
for economic and political development (di Lernia and Palombini 2002;
D. W. Phillipson 2003b). In addition, African archaeology is relevant far
beyond its own continent. The archaeologists and prehistorians of other
regions have much to learn from the African record, not only from its unpar-
alleled evidence for the earliest periods of human development, but also
4 afric an archaeolog y

methodologically. Because much of Africa has undergone environmental
change on a scale which is relatively minor when compared with the for-
merly glaciated regions of Eurasia and North America, abundant data are
fairly readily available to aid the interpretation of prehistoric subsistence
practices. In many parts of Africa the scale of modern development has like-
wise been comparatively slight. Africa also provides excellent opportunities
for contrasting the testimony of archaeology with that of linguistic and oral
historical studies, and for interpreting the meaning of rock art in the light
of the belief systems of recent peoples, as will be discussed below.
The modern study of African archaeology has developed in two princi-
pal directions. The literate civilisations of ancient Egypt and North Africa
were of interest to Graeco-Roman historians and, more recently, have been
investigated through more than 200 years of changing approaches, while
the prehistory of more southerly regions ¬rst received serious attention in
South Africa at the end of the nineteenth century (Robertshaw 1990a). The
two studies have long remained separate, and their methodologies are only
now beginning to converge as is illustrated, for example, by recent investi-
gations in Ethiopia. In the Saharan and sub-Saharan latitudes the emphasis
and geographical coverage of research has always been irregular, and the
archaeology of large areas still remains virtually unexplored. Of necessity,
much effort has gone into demonstrating basic sequences and setting up
a terminological framework for prehistory. Only fairly recently has it been
practicable to present a comprehensive overview such as that attempted
here, and to propose plausible explanations for the trends and developments
which are detected in the archaeological record.
The classi¬cation of archaeological materials into successive phases, indus-
tries and complexes is now seen as an arti¬cial compartmentalisation of
what was usually a continuous variation, both in time and space. It must,
however, be recognised that closely de¬ned boundaries between culture
areas could and did exist in certain circumstances, and that stylistic and
technological change proceeded more rapidly at some times than at others.
This problem is considered in greater detail below (pp. 84--6).
Contrary to assumptions frequently made in the past, it is now recognised
that the parameters of material culture distributions do not necessarily coin-
cide with those of human societies as recognised on socio-political, linguistic
or other bases. Uncertainty about the signi¬cance of material culture group-
ings is greatest in the case of the earlier prehistoric industries, because it
is only very rarely that we can understand the purposes to which particular
artefacts were put. It is frequently dif¬cult to distinguish between varia-
tion due to different functions, to stylistic traditions and preferences, or to
other factors such as the availability of particular raw materials. Further-
more, a single society may engage in multiple distinct life-styles, sometimes
Introduction 5

on a seasonal basis, perhaps in separate areas and environments: contrast-
ing archaeological assemblages may thus represent different activities of a
single community or of sub-groups within that community. On the other
hand, it has also been shown that certain items or styles of material culture
may ful¬l a symbolic function through their association with a particular
society or section of a society, as is the case, for example, with some iron-
smelting furnaces (Childs 1991). Although material culture distinctions do
not necessarily coincide with socio-political ones, it is equally incorrect to
assume that such correlations may not, in certain circumstances, exist.
For reasons such as these, African prehistory is here presented with
emphasis on economic development and general life-style (including, where
practicable, socio-political systems and ideology), correspondingly less atten-
tion being paid to the de¬nition, succession and inter-relationship of named
cultures and industries. However, in the present state of African archaeolog-
ical studies, the old framework needs to be partially retained. In many parts
of the continent, as will be made apparent in the following chapters, con-
certed programmes of archaeological research have never been undertaken.
There are several regions, even whole countries, where chance discoveries
or isolated excavations, often poorly documented, provide the only data on
which a synthesis may be based. Here, the archaeologist may only be able
to propose an outline succession of industrial stages, such being an essen-
tial pre-requisite for the detailed study of ancient life-styles and resource-
exploitation patterns.
An ever-increasing contribution to our understanding of the past is now
being made by genetic studies, not only of past and present human popula-
tions, but also of the plants and animals on which people™s livelihood has
depended. The potential of such studies, and the methodologies involved,
are described by M. K. Jones (2001).
In writing a concise overview of African archaeology such as that con-
tained in this book, it has been necessary to select and, on occasion, to
simplify. While some geographic areas have yielded a wide range of archae-
ological data, in others very little is yet available. Thus it is that in certain
sections of the book almost every site that has been investigated receives
mention, while elsewhere a more general picture emerges from a series
of comparable investigations. As a result, major changes of emphasis and
interpretation may be expected to occur as research and discovery progress.
Topics which have been the subject of recent in-depth research receive com-
paratively detailed treatment.
In the building up of an overview of African prehistory, particular atten-
tion must be paid to erecting a sound chronological framework. Several
methods are available; for details of the methodologies the reader is referred
to the comprehensive survey presented by Klein (1999). Age estimates based
6 afric an archaeolog y

on radiocarbon analyses are particularly problematic, since it is only for the
more recent periods that the relationship between radiocarbon and true
ages is known. In this book ages are cited in the following manners in order
to minimise confusion and to aid comparison between dates derived from
different sources and methods.
(a) In chapters 2--5 ages are given in the form ˜about . . . years ago™. These
ages apply to periods beyond the last 7000 years and should all be regarded
as approximations. They are derived from a variety of sources, mostly -- for
the last 50,000 years or so -- radiocarbon, and no attempt has been made to
calibrate or correct them unless otherwise stated.
(b) In chapters 6--8 dates since about 5000 bc are given in years bc or ad,
these conventional designations being retained as those most widely and
most readily understood. Here, radiocarbon dates have been calibrated and
are expressed in calendar years according to the calculations presented by
Stuiver and Kra (1986; Stuiver et al. 1998). At certain periods this calibration
permits only approximate ages to be proposed because of variation in the
radiocarbon content of the atmosphere. Precise dates such as 146 bc are
derived from historical sources. All ages noted in these three chapters are
thus intended to be comparable with one another, but they are not neces-
sarily compatible with those cited in chapters 2--5.
Since more plentiful data relating to absolute chronology are now avail-
able than could be employed by the writers of previous syntheses, and in
view of the evidence for disparate rates of development in different parts
of the continent, this book does not employ the conventional terminology
based upon broad chrono-technological subdivisions such as ˜Late Stone Age™,
˜Neolithic™ or ˜Iron Age™. It has long been recognised that such terms can-
not be precisely de¬ned, but their informal use has continued, often at the
expense of clarity; they are avoided in this book.

Language provides an important means of classi¬cation for African pop-
ulations. It has a major bearing on an individual™s sense of identity and
membership of a group. It also has historical validity, since people usually
learn their ¬rst language from the other members of that group to which
they belong by birth and/or upbringing.
There is good but by no means unanimous agreement among linguists
concerning the major language families of Africa (Greenberg 1963; Heine
and Nurse 2000; Fig. 1), whose present distribution is shown in outline
form in Fig. 2. In the northern and northeastern regions of the continent,
the languages which are spoken today belong to the super-family generally
known as Afroasiatic. This includes the Berber languages of North Africa
Introduction 7

Family Main divisions Examples

Afroasiatic Semitic Arabic, Amharic, Gurage,
Berber Berber, Tuareg
Cushitic Somali, Oromo, Afar, Sidamo,
Chadic Fali, Hausa
Nilo-Saharan Sudanic Acholi, Shilluk, Mangbetu, Jie
Saharan Kanuri, Teda, Zaghawa
Songhai Songhai
Niger--Congo West Atlantic Dyola, Fulani, Temne
Mande Mwa, Mende
Voltaic Dogon, Mossi, Talensi
Kwa Akan, Bini, Ibo, Igala, Yoruba
Bantu Gikuyu, Bemba, Shona, Xhosa,
Adamawa-Eastern Mbaka, Zande
Fig. 1: The
!Kung, ‡Khomani, Nama
KhoiSan South African KhoiSan
classi¬cation of
?Sandawe Sandawe
recent African
?Hadza Hadza
languages (after
Greenberg 1963)

Fig. 2: The
distribution of
Africa™s major
language families
in recent times
(simpli¬ed from KhoiSan
Greenberg 1963)
8 afric an archaeolog y

and the Cushitic tongues centred on Ethiopia and Somalia, as well as the
widespread Semitic family, the modern members of which include Arabic,
Amharic and Hebrew.
To the south is a very irregularly shaped area covering much of the cen-
tral and southern Sahara, the southern Sudan and parts of the adjacent
savanna with an extension into parts of East Africa, where most of the mod-
ern languages are classed as Nilo-Saharan, with Nilotic and Sudanic as the
principal subdivisions. Songhai, spoken around the Niger bend, may also
have Nilo-Saharan af¬nity. It may be that the present fragmented distribu-
tion of the Nilo-Saharan languages indicates that they were formerly spoken
over a more extensive area.
Most of the modern languages of West Africa belong to the Niger-Congo
family, which may be extended to include Kordofanian, spoken in the west-
ern Sudan. Within West Africa these languages have developed considerable
diversity. On the other hand, the distribution of one sub-group of Niger-
Congo extends over the greater part of central and southern Africa, exclud-
ing the extreme southwest. These are the Bantu languages which, despite
the enormous area of their distribution, show a relatively strong degree
of similarity with one another. The northern limit of the Bantu languages
approximates to the northern edge of the equatorial forest. In the savanna
woodland to the north, the Adamawa and Ubangian languages also belong to
the Niger-Congo family. Niger-Congo is sometimes linked with Nilo-Saharan
to form a Niger-Saharan macrophylum (Blench 1999).
As will be shown in chapter 7, there is good evidence that the Bantu-
speaking peoples have expanded from a northwestern area into sub-
equatorial latitudes during the course of the last few thousand years. In
signi¬cant parts of this region, these new populations replaced or absorbed
people who spoke languages of the KhoiSan family, such as still survive
in the southwesternmost parts of the continent. These are the languages
of the Khoi (formerly sometimes called by the derogatory term Hottentots)
and San (or Bushmen), who have retained into recent times their traditional
herding or hunting life-styles beyond the country of the Bantu-speakers.
There are indications that in earlier times KhoiSan-related languages may
have been spoken as far to the north as the modern Kenya/Tanzania bor-
der area, but in regions further to the west their northerly extent is less
In the absence of writing it is on modern languages that historical lin-
guists must, of necessity, base their conclusions (Nurse 1997; D. W. Phillip-
son 2003a). Through studying the distribution of recent linguistic forms it
is often possible to reconstruct certain features of the past languages from
which the modern ones are derived, and to suggest the areas in which these
ancestral languages may have been spoken. The vocabulary that is attested
Introduction 9

for these ancestral languages can tell us something about the life-styles of
the people who spoke them, and about the things with which they were
familiar. As different peoples came into contact with one another words
were borrowed from one language into neighbouring ones; these ˜loanwords™
too can often be traced. It is through studies such as these that the linguistic
prehistory of Africa may tentatively be reconstructed.
We have of course no precise information about the varying speeds at
which particular linguistic changes proceeded. It is only in the case of lan-
guages which have a long written history that such speeds can be calcu-
lated at all precisely. Linguistics alone can provide only a relative ordering
of processes and events, together with a rough estimate of the lengths of
time that may have been involved. Use of glottochronological formulae to
calculate the dates at which linguistic developments took place should be
regarded with great suspicion, particularly when applied in non-literate con-
texts, since these formulae assume that language change occurs at a uni-
form rate. However, when links can be demonstrated between independent
sequences, based respectively on archaeology and on linguistics, the chronol-
ogy of the latter is placed on ¬rmer ground. Historical reconstructions based
on linguistic studies may be of particular value in supplementing the tes-
timony of archaeology in areas where little excavation can be undertaken,
and for those aspects of inter-group relationships illustrated by linguistic
studies but which are dif¬cult to demonstrate on the basis of archaeologi-
cal evidence. Such linguistic reconstructions mostly relate to the past 5000
or 6000 years, although tentative attempts have been made to apply these
methods to still earlier periods.

Oral traditions
In order properly to interpret the oral historical traditions which are pre-
served in many African societies, we must understand the rˆle that they
play in those societies and the reasons for their recollection (Henige 1974;
Vansina 1985). It is generally found that oral traditions are most carefully
preserved and re-told among peoples who have a strong centralised political
system. In such cases the function of the historical traditions is often to
support the established authority, for example by explaining the origin of
the ruling clan or family and the manner by which its members claim their
right to rule. Several societies recognise this aspect of oral tradition and
have of¬cial historians, whose task it is to preserve and transmit orthodox
versions of their state histories.
Traditions which purport to relate to events of more than four or ¬ve
centuries ago must generally be interpreted with particular caution. Abso-
lute chronology (in the western sense) is not often a major interest of the
10 afric an archaeolog y

custodians of oral tradition; events may sometimes have taken place at sig-
ni¬cantly earlier periods than a literal interpretation of the traditions would
suggest (e.g. J. C. Miller 1972). As with written histories, oral historical tra-
ditions tend to concentrate their attention on political events and on the
activities of important individuals. They are not, therefore, a substitute for
archaeology as a source of historical information, and a comprehensive pic-
ture of the African past can only be built up through the use of all available

Increased attention has been given in recent years to the application of
archaeological perspectives to the study of recent societies and their mate-
rial culture (David and Kramer 2001). To some practitioners, the principal
aim of ethnoarchaeology is to gain insights to aid the interpretation of
archaeological data relating to earlier periods. This is an exercise which must
be approached with great caution. It cannot be emphasised too strongly that
observations of recent societies are of value only in suggesting possible inter-
pretations of archaeological data; they can never themselves provide conclu-
sive proof. Secondly, use of ethnoarchaeologically based interpretations may
carry the hidden implication that the recent peoples studied are in some
way backward or primitive. Those who practise or make use of ethnoar-
chaeology must beware of thus unintentionally insulting those who have
provided their inspiration. This is not to belittle the value of such studies,
provided that the models which they help to generate are applied with care
and sensitivity. Two examples may be cited: particularly valuable insights
have been obtained into the uses and signi¬cance of stone tools (e.g. Brandt
1996) and into non-western views of the past and of time (e.g. Schmidt 1995;
Stahl 2001).

Africa in world prehistory
The archaeological picture of the African past now discernible is one of
paramount importance for the study of human prehistory. As will be shown
below, evidence for the life-style and physical characteristics of the earliest
hominids comes at present only from African sites. While it cannot be con-
clusively demonstrated that human beings ¬rst evolved in Africa, there is a
very strong probability that this was indeed the case. Virtually every major
subsequent stage in humankind™s development may be illustrated from the
African record.
The succession of African hunter-gatherer societies is the longest and one
of the most varied known. It extends from the origins of humanity to the
Introduction 11

present day and, potentially, provides an evolutionary link between the stud-
ies of the social anthropologist and those of the specialist in the behaviour
of non-human primates. This does not, of course, imply that modern hunter-
gatherers are ˜primitive™ survivors from earlier times, somehow less devel-
oped than their farming neighbours. Hunter-gatherers have adapted to envi-
ronments of great diversity, ranging from deserts and high-altitude glacial
margins on the one hand to rain-forests and coastal swamps on the other.
Major environmental changes have, of course, taken place during the period
of some 2 to 3 million years with which we are here concerned, but these
have not generally been so drastic as those in more northerly latitudes
and, although their distribution has undergone great shifts, the range of
situations exploited by prehistoric hunter-gatherers has been preserved in
Africa to an extent not paralleled in other continents that were settled in
Middle Pleistocene or earlier times. This continuity, both in environment
and in the hunter-gatherer life-style itself, offers in Africa an unrivalled
series of opportunities for interpreting major trends and processes in human
It has been noted by J. G. D. Clark (1969) that developments in stone-
tool technology followed broadly similar stages over much of the world
(Fig. 3 on p. 12). Backed-blade industries of the type represented in the Euro-
pean and West Asian ˜Upper Palaeolithic™ were thought prerequisite for the
development of food-production and, ultimately, of literate civilisation. This
premise, at least so far as Africa is concerned, has been seriously questioned
(Shaw 1971), and the point made that, although blade industries of this type
are uncommon in Africa south of the Sahara, microlithic industries such
as were made by the earliest European and Levantine food-producers were
produced in Africa at a remarkably early date. More recent research, sum-
marised below, has con¬rmed this observation, and microlithic technology
is now attested in sub-equatorial Africa at dates far earlier than in any other
part of the world; not only is it signi¬cantly older than the European or
Asian blade industries, but it is apparently associated with the oldest fossils
so far known which may con¬dently be attributed to fully modern people,
Homo sapiens or H. s. sapiens. These discoveries become even more important
as a result of recent genetic research which suggests that all modern human
populations may be descended from an African ancestor (Stringer and McKie
1996). Far from being a backwater, as has sometimes been suggested, Upper-
Pleistocene Africa may have been a world leader both in the evolution of
our species and in its development of technology.
Current research has also involved a major re-evaluation of the evidence
for early African settled life, herding and cultivation. At least south of the
Sahara, these developments are now seen to have taken place very gradually
and essentially independently of comparable processes in other parts of the
12 afric an archaeolog y

Fig. 3: The ¬ve
modes of lithic
proposed by J. G.
D. Clark in 1969.
The illustrations
are of African
reproduced to a
common scale.
Introduction 13

world. Settled life appears to have come about in what is now the southern
Sahara and sahel in the context of the rich and well-watered environments
which prevailed there some 10,000 years ago. This same general area is now
recognised as the homeland of many plant species which were subsequently
brought under cultivation and dispersed to become important food crops
there and in other parts of sub-Saharan Africa. Additional important crops
were developed in Ethiopia, and others on the fringe of the equatorial forest.
Cattle were probably locally domesticated in the eastern Sahara. Although
the domestic animals herded south of the Sahara appear all to have been
introduced from elsewhere, it is now clear that Africa was a major area for
the initial cultivation of vegetable foods as diverse as yams, enset, rice and
other cereals.
With the possible exception of the Egyptian Nile Valley, no part of Africa
saw the rise of a wholly indigenous literate civilisation, for strong exter-
nal in¬‚uences made signi¬cant contributions to such developments along
the Mediterranean littoral, in the Sudanese Nile Valley, in the Ethiopian
highlands and on the East African coast. What were the main factors con-
tributing to the rise of literate civilisation in Egypt and elsewhere which
were absent in other parts of Africa? The Egyptian Nile Valley was given
great fertility by the annual inundations of the Nile. Its surrounding desert
constrained the physical expansion of the dense population which the fer-
tile valley could support. Comparable situations prevailed in other centres
of early civilisation: the Indus Valley, Mesopotamia, the Hwang Ho Valley
and Mexico (Trigger 2003). The social stress engendered by a dense popu-
lation required for its control the development of elaborate socio-political
systems, an established and sanctioned order of a complexity which would
have been out of keeping with the smaller, less constrained societies of most
other parts of the continent. It may be argued, at least in part, that it was
the richness of the African environment and its lack of physical barriers
which permitted many African societies to develop their own forms and
orders without the constraints imposed by literate civilisation.
Research in Africa can provide archaeologists of other regions with impor-
tant insights which aid their interpretations (D. W. Phillipson 2003b). Par-
ticularly important examples relate to socio-political evolution and people™s
views of history (S. K. McIntosh 1999a; Stahl 2001).
It is far more relevant to consider what Africa did achieve, than what
it did not (cf. Shaw 1971). It is also pertinent to view the development of
African societies, as revealed by archaeology and other disciplines, from an
essentially internal viewpoint before comparison is made with their coun-
terparts in other parts of the world. Thus we are able to evaluate African
achievements in terms of their African context and to appreciate the range
of economic practices, technologies, socio-political systems and beliefs which
14 afric an archaeolog y

was developed in the context of varied population densities, physical bound-
aries, communications and available resources. Then we can see the com-
prehensive manner in which varied environments were exploited. We can
begin to understand why some aspects of indigenous technology reached
the high level of expertise evidenced, for example, in West African casting
of copper alloys, while other aspects -- such as methods of transport -- saw
little change over prolonged periods. We can appreciate oral tradition as a
counterpart of written literature, and wonder at the possible antiquity of
art-forms in wood and other perishable materials which are at present only
known from the most recent periods. African archaeology provides a unique
view of cultural development leading to recent societies that are now appre-
ciated, not as failures that have fallen by the wayside in the rise of industrial
civilisation, but as examples of different -- perhaps (as in several other parts
of the world) more viable in the long term -- expressions of human cultures
fully adapted to their practitioners™ circumstances.
The emergence of humankind in Africa
De¬nition and process
Before attempting an account of the African evidence for human origins, it is
¬tting to make some general observations about the nature of that evidence
and of the research on which it is based; it is necessary also to offer some
de¬nitions of the terms that are employed both here and in the writings
of others. The ¬rst point that needs to be emphasised is that the evidence,
although still very incomplete, is accumulating rapidly. New discoveries or
analyses can frequently necessitate radical revision of our interpretations.
However, there is often a long lapse of time between the initial announce-
ment of a discovery (whether it be made in the ¬eld or in the laboratory) and
its de¬nitive publication. There are often strong political and ¬nancial pres-
sures on researchers to make prompt -- even premature -- announcements.
Indeed, a disconcertingly high proportion of the evidence on which the
present synthesis is based comes from such preliminary accounts. Further-
more, very rarely is a truly continuous record available for study: we have
a number of brief and incomplete glimpses separated by long intervals for
which no information is available. Any attempt to reconstruct a continuous
sequence must take account of this limitation.
Much of the ongoing research here described is being undertaken by
collaborative teams of specialists. This chapter attempts to synthesise four
main interlocking strands of evidence which are often the work of distinct
experts: geology and faunal studies can yield information about habitat and
chronology, examination of the hominid fossils has the potential to illus-
trate human physical evolution, genetic studies of both fossil and modern
populations are beginning to make major contributions to knowledge of
their inter-relationships, while associated artefacts and their associations
elucidate some aspects of cultural development.
Clearly, location of ¬eld research is determined both by the anticipated
availability of evidence and by whether practical and political conditions
are conducive to large-scale investigations. For example, the Rift Valley in
eastern Africa offers unparalleled conditions for the preservation and subse-
quent discovery of materials relevant to this research, together with volcanic
associations which can be dated fairly precisely; yet even here ¬eldwork was
impracticable in Ethiopia during the 1980s because of that country™s politi-
cal and economic conditions. Relatively little ¬eld research relating to early
16 afric an archaeolog y

hominid evolution has been undertaken in more westerly parts of Africa
because the relevant remains are there more rarely preserved and less read-
ily discovered. It follows that apparently limited geographical distributions
must be interpreted carefully, and that ancient phenomena may have been
far more widespread than their present physical attestation might suggest.
Evidence for environmental change must be sought and its impact on past
developments assessed.
Next, it is necessary to offer some observations about the terminology that
is employed to classify fossils, whether of human ancestors or other crea-
tures. This classi¬catory system is still based on the Systema Naturae devised
by Carolus Linnaeus in 1735, which attributes each animal or plant to a
genus (e.g. Homo) and, within the genus, to a species (e.g. sapiens). Linnaeus
(who wrote more than 120 years before the publication of Charles Darwin™s
On the Origin of Species in 1859) was concerned with the classi¬cation of
contemporary living things and, so far as animals were concerned, his prin-
cipal criterion for de¬ning a species was that its members should be capa-
ble of interbreeding to produce fertile offspring. This, clearly, is something
which can never be demonstrated for individuals known solely from their
fossil remains. Linnaeus recognised that differences in outward appearance
could be misleading: for example, some birds showing only minor differ-
ences are assigned to different species, while recent breeding has produced
very diverse dogs which all belong to the same species. The amount of phys-
ical variation that can be accommodated within a single species thus varies
greatly. In some instances, much of this variation may be accounted for by
sexual dimorphism: males and females of certain birds, for example, are
indistinguishable to ornithologists without close anatomical examination,
whereas female gorillas are half the size of males from which they may
also be differentiated on the basis of dentition, musculature and osteology.
Adoption of the Linnaean system by those who study the fossil evidence for
human (and non-human) evolution thus gives rise to great dif¬culties, both
practical and philosophical. The de¬ning characteristic of a Linnaean species
cannot be established from a fossil. Many fossils cannot be attributed with
certainty to a male or a female individual; we may thus be unable to deter-
mine the degree of sexual dimorphism in a fossil population. In fact, we
can obtain little if any idea from the fossil record of the amount of physical
variation that can be accommodated within a fossil species. Linnaeus had no
concept of evolution, so it is hardly surprising that his classi¬catory system
cannot logically be applied to a diachronic situation of species-change. If we
do make such an attempt, we must accept either the former existence of
individuals intermediate between recognised species or that at some stage
in the evolutionary process a female of one species gave birth to offspring
of another species or even genus. Both are such manifest absurdities that
The emergence of humankind in Africa 17

they call into question our whole system of classifying fossils. The problem
has often been recognised (e.g. H. J. and J. Deacon 1999), but no adequate
remedy has been proposed.
As explained, we have only very hazy ideas about the amount of physical
variation that should be accommodated within a Linnaean species recog-
nised in the fossil record. Some attempts have been made (e.g. Foley 1991)
to consider the extent to which an ideal classi¬cation of this material should
re¬‚ect perceived similarities rather than differences. It is, of course, recog-
nised that geographical isolation can accelerate differentiation, and this has
led to the questionable practice of de¬ning species as much by the geography
of their occurrence as by their morphology. Examples (considered further
below) are the division of the robust australopithecine Paranthropus into East
African and South African species respectively designated boisei and robustus,
and the division of the species formerly recognised as Homo erectus between
African H. ergaster and H. erectus in Asia (cf. Klein 1999).
It is not only in the case of fossils that conventional classi¬catory terminol-
ogy actually hinders our understanding of diachronic processes. Compart-
mentalisation is a very useful preliminary stage in developing an under-
standing of many variable phenomena; later, however, it can mask the
nature of variation. This will be a recurring theme in many chapters of
this book (cf. Panchen 1992). Colonial administrators (and some others) have
sought to group their subjects into tribes which have then been regarded
as watertight compartments whose validity may even extend through time
(M. Hall 1983). Classi¬cation of languages has encountered similar problems.
This book is concerned with the past lives of people, especially but not exclu-
sively as illustrated by their artefacts: when archaeological study of a par-
ticular period or region is beginning, classi¬cation of artefacts into named
industries is a useful approach. Such industries need to be described in
terms of the technology employed and the morphology of the end-products
as well, ideally, as the uses to which individual items were put and thus,
in conjunction with other evidence, the activities that they could support.
Eventually, reference can also be made to the geographical area where they
occur and to the span of time during which they were made and/or used.
Degrees of similarity may then be assessed, leading to the proposal of a clas-
si¬cation. Here we run into dif¬culties which are remarkably analogous to
those encountered by biologists: how can processes of change be understood
and accommodated within a classi¬catory system which implies the exis-
tence of rigidly de¬ned boundaries? There may, of course, have been times
when a sudden change did take place, or situations where the practitioners
of one industry were separated by a sharply de¬ned frontier from those of
another. Far more often, however, the transition will have been more grad-
ual, sometimes extending through hundreds of generations or over great
18 afric an archaeolog y

distances. The past, notably but not exclusively the human past, has seen a
multi-dimensional reticulation of variation and change which in some cir-
cumstances and at some times has been more marked and/or more rapid
than otherwise. There seem to have been, and continue to be, situations of
virtual stasis and others of pronounced change -- cultural and/or biological.
Conventional labelling of categories cannot adequately re¬‚ect this situation.
It may be possible to recognise certain overarching trends which cut across
the conventional terminology. I do not argue for abandonment of classi¬ca-
tion, but for the adoption of methodologies and terminologies which re¬‚ect
processes of variation as these are now understood.
A modest step in this direction, which the present writer has long advo-
cated, is the application to studies of stone-tool industries of the categorisa-
tion into ¬ve successive modes, as proposed by J. G. D. Clark in 1969. This
taxis, summarised in Figure 3, avoids the tendency to correlate stages of tech-
nological development with ¬nite periods of time; and it also helps to min-
imise the arti¬cial compartmentalisation of what were frequently continu-
ous processes of development (D. W. Phillipson 1977a: 23; Foley 2002). The
modes appear to form a homotaxial sequence of world-wide applicability;
they do not represent watertight compartments, it being recognised that ele-
ments of the technologies of earlier times often continued alongside more
recent innovations. This last point has frequently been overlooked (e.g. J. D.
Clark and Schick in de Heinzelin et al. 2000: 51--181; J. D. Clark 2001b), which
may account for prehistorians™ limited adoption of J. G. D. Clark™s categori-
sation. Although its value is restricted to discussion of very broad trends, it
provides a useful means of technological comparison while avoiding poten-
tially misleading terms such as ˜Early Stone Age™ which can too easily be
assumed to correlate with ¬nite periods of time.
The synthesis offered in this book takes account of these factors. It gen-
erally retains at least some aspects of conventional terminology, if only to
facilitate use of other literature, but it seeks to note circumstances where,
in the author™s opinion, such terminology has a tendency to distort or to
hinder understanding. In particular, since more plentiful data relating to
absolute chronology are now available than could be employed by the writ-
ers of previous syntheses, and in view of the evidence for disparate rates of
development in different areas, this book does not employ the conventional
terminology based upon broad chrono-technological subdivisions such as
˜Late Stone Age™, ˜Neolithic™ or ˜Iron Age™. It has long been recognised that
such terms cannot be precisely de¬ned, but their informal use has contin-
ued, often at the expense of clarity; they are avoided here.
A further source of misunderstanding is the conventional application to
types of stone artefact of names such as ˜scraper™ or ˜arrowhead™ which carry
The emergence of humankind in Africa 19

an implication as to function, when there is in fact little if any de¬nite
information as to how the tools were actually used. In this book, the attempt
has been made to restrict use of such functional terms to cases where there is
good evidence as to the use to which artefacts were originally put, although
it is recognised that some element of subjectivity necessarily remains.
Next, it is necessary to de¬ne some parameters. This is particularly impor-
tant in the present chapter which attempts to discuss the initial processes of
both physical and cultural evolution by which humans came to be differenti-
ated from other animals. It has long been conventional to apply the informal
terms ˜hominoid™ to humans and the closely related great or anthropoid
apes, and ˜hominid™ to humans and their ancestors after the separation
from the apes. We shall turn below to the problem of de¬ning the term
˜human™; ¬rst it is necessary to record that recent genetic studies demon-
strate a remarkably close af¬nity between apes and humans (Goodman et al.
1990; Gagneux et al. 1999), to the extent that it has been suggested that a
distinction at this level cannot be substantiated and that apes should be
regarded as hominids, the term ˜hominin™ being proposed in its place to
exclude the apes. Certain recent writers (e.g. Mitchell 2002) have adopted
this change but the present book does not follow them: the proposed ter-
minology does not improve understanding and is not accompanied by more
precise de¬nition of the categories to which it relates.
The once-conventional wisdom that ˜man is a tool-making animal™
(Oakley 1952) has been called into question by studies of chimpanzees. For
years the question was fudged by invoking contrasts between tool-using, tool-
modifying and tool-making, but the old de¬nition has ¬nally been rendered
untenable by the demonstration that a Bonobo chimpanzee can learn to make
stone tools, admittedly under laboratory conditions (C. and H. Boesch 1990;
Toth et al. 1993). It must also be stressed that the earliest tools are not easily
recognisable in the archaeological record: only stone artefacts are generally
preserved and, in most circumstances, readily recognised. Modern apes use,
by preference, a range of more perishable but more easily worked materials,
and there is no reason to believe that early hominids were different in this
regard. (The distinction must be made between an artefact, which is simply
any arti¬cially produced thing, and a tool or implement which is an artefact
made, modi¬ed or selected to facilitate the accomplishment of a task.) There
is no reason to suppose that early tools were made from stone before or in
preference to those of other materials. Thus it is likely that the ¬rst tools,
by whatsoever creature they were made, will not have been preserved.
For reasons that will be set out later in this chapter, the view has been
expressed that Homo ergaster, ¬rst recognised in the African fossil record
around 1.75 million years ago, represents both physically and culturally a
20 afric an archaeolog y

major advance from the preceding hominids, whether australopithecine or
Homo habilis. In recognition of this, use of the terms ˜person™ and ˜people™ is
here restricted to H. ergaster and later humans.
A recent development has been the adoption of the phrase ˜anatomically
modern people™ in place of the Linnaean designation Homo sapiens or H. s.
sapiens. While, in view of the problems outlined above, this abandonment of
Linnaean classi¬cation might be welcomed, it must be questioned whether
the new term has actually improved understanding. It has the welcome
advantage of emphasising the virtual identity of all modern people (but
see A. W. F. Edwards 2003), avoiding implications that some ˜races™ might
be more or less ˜primitive™ than others. On the other hand, it has had two
particular effects which give rise to concern. Because it has not been accom-
panied by a corresponding revision of terminology relating to earlier repre-
sentatives of the genus Homo, it has imposed a conceptual barrier between
modern people and their ancestors. Secondly, and arising from this barrier,
it has strengthened the view that modern people may also be distinguished
on cultural and behavioural grounds, leading (as will be argued in chapter 4)
to unjusti¬ed emphasis on certain aspects of the archaeological record
and to a tendency to underestimate the cultural achievements of earlier
hominids. It has even given rise to the suggestion (Klein 2000a) that these
behavioural developments may have been brought about by genetic muta-
tion as little as 60,000 or 50,000 years ago -- a Eurocentric argument which
approaches circularity.
The term ˜human™ is retained in this book as a purely informal noun or
adjective referring to any australopithecine or member of the genus Homo,
whatever his or her tool-making capabilities may have been, in contexts
where the more precise designation ˜hominid™ seems inappropriate. The
growing tendency to apply the word ˜human™ exclusively to anatomically
modern Homo sapiens is to be deplored.

World-wide precursors of the hominids
The story of the emergence of humankind extends far back into geological
time (Fig. 4). The modern species of Old World and New World monkeys, apes
and hominids are all classed as members of the Anthropoidea sub-order of
the order Primates (Fig. 5). Other members of this order, with which we are
not here concerned, include such animals as tarsiers and tree-shrews. Fossil
remains of early primates (Klein 1999) have been recovered at many sites in
the Americas, Europe and Asia as well as in Africa, extending back in time
as far as the end of the Cretaceous period about 70 million years ago. The
modern Old World Anthropoidea are believed to be descended from small
but ape-like primates, notably that named Catopithecus, whose remains are
The emergence of humankind in Africa 21

Fig. 4: The
periods of the
last 24 million
years, showing
the approximate
ages of the
hominoid genera
attested in the
African fossil

Fig. 5: The
classi¬cation of
the hominids
within the order

best known from deposits in the Fayum Depression of Egypt, dating from
the late Eocene/early Oligocene period about 36 million years ago (Simons
1990; Simons and Rasmussen 1994).
By the beginning of the subsequent Miocene around 24 million years ago,
it appears that primate evolution had proceeded suf¬ciently far to permit
22 afric an archaeolog y

the differentiation of lines of descent that have led, on the one hand, to the
modern Old World monkeys, and on the other to the great apes and modern
people. The best-known fossil primate of this time is the forest-dwelling
Proconsul from Rusinga Island in Lake Victoria (A. C. Walker et al. 1993).
Subsequently, the hominoid line is represented around 15 million years ago
in western Kenya by Kenyapithecus, which shows important developments in
skull, teeth and wrist. It probably lived in the open woodland that became
widespread in East Africa in mid-Miocene times, before the completion of
the great earth movements which resulted in the formation of the Rift
Valley (Pickford 1983, 1986).
In later Miocene times, between about 12 and 5 million years ago, further
evolutionary development must have taken place which eventually led to
the emergence of the family to which all human types, past and present,
belong. With the colonisation of new environments, resulting in greater
geographical spread and subsequent isolation, several distinct varieties of
early hominoid developed. Fossil-bearing deposits of this time are relatively
rare in Africa, and in those that have been investigated primate remains
are extremely uncommon. At present the fossils of greatest relevance to
the study of human origins are those from Europe and southern Asia
attributed to the genera Dryopithecus and Sivapithecus respectively (Bene¬t
and McCrossin 1995). The faces of these creatures were less snout-like than
those of their ancestors, their jaws were more massive and their teeth were
further adapted to use as grinders. Associated fossils of other species from
the same sites indicate that these creatures favoured open savanna wood-
land environments. This continued adaptation to a less restrictive habitat,
doubtless linked (as the teeth indicate) with the adoption of a more omniv-
orous diet, may have been a major step in the evolutionary processes which
led to the emergence of humankind.
When the African fossil hominoid record resumes around 5 million
years ago, it is almost exclusively in eastern and southern Africa that true
hominids are represented; they occur with an abundance that contrasts
markedly with the earlier periods. Despite the wide Old World distribution
of the mid-Miocene hominoids, the evidence currently available suggests
that it was probably in Africa that hominids ¬rst evolved.

The earliest hominids
This section attempts an overview of the principal fossil evidence from
Africa, individual sites and their archaeological materials being discussed
subsequently. The picture now offered is radically different from that pro-
posed in the second edition of this book (D. W. Phillipson 1993b), which
serves to emphasise how rapidly knowledge is developing and how new
The emergence of humankind in Africa 23

Fig. 6: A tentative ˜family tree™ of African hominids during the last 5 million years. The most
widely accepted evolutionary relationships are indicated by hatched lines; discontinuities by
continuous lines. (Australopithecus robustus and A. boisei are regarded as regional races in
southern and eastern Africa respectively.)

discoveries may necessitate substantial revisions of current interpretations.
Figure 6 attempts a summary representation of the system now proposed.
It is important to realise that in both southern and eastern Africa the
majority of the fossil discoveries have been made within recent years. Study
of many is as yet at a preliminary stage, and there is often considerable
24 afric an archaeolog y

controversy about their attribution to named species and, on occasion, their
dating. Likewise, new ¬nds are steadily being announced; these may require
the modi¬cation or abandonment of existing theories. Any account of the
current state of research must, therefore, be both tentative and provisional.
A number of different systems are used to order the chronology of this time;
these are summarised and explained in Figure 7.
It is clear from the results of recent research that it was during the period
between about 8 million and 5 million years ago that the ¬rst individuals
generally acknowledged as hominids must have developed. The very sparse
fossil evidence is supported by genetic studies which independently suggest
that the most recent common ancestor of humans and chimpanzees lived at
this time (Miyamoto and Goodman 1990). So far, however, ¬eldwork in Africa
has revealed only two relevant fossils which date from this crucial interval.
These are the fragments attributed to Orrorin tugenensis from Lukeino in the
Tugen Hills of Kenya (Senut et al. 2001; see also Kingston et al. 2002) and,
less certainly, a skull from the Djourab Desert in Chad (Brunet et al. 2002),
concerning which elaborate claims but little factual detail have been pub-
lished. Estimates for the age of the Lukeino deposits range between 6.1 and
5.7 million years. Slightly more recent are fossils from the Middle Awash
area of the Ethiopian Rift Valley which have been attributed to Ardipithecus
ramidus kadabba and dated 5.8--5.2 million years ago (Yohannes Haile-Selassie
et al. 2001). None of these recent discoveries, from Kenya, Chad and Ethiopia,
has yet been published in detail or fully evaluated. Controversy (not to men-
tion rivalry) continues, particularly with regard to the possible bipedalism
of these creatures and their relationships to later hominids. Likewise, it has
not yet been ascertained how these new discoveries relate to earlier reports
of very fragmentary hominid fossils from late Miocene and early Pliocene
sites west and south of Lake Turkana in northern Kenya, such as a piece of
a jaw from Lothagam, dated to some 5.5 million years ago, and a skull frag-
ment from a somewhat more recent context in the Chemeron Formation
near Lake Baringo (Howell 1982).
It is noteworthy that the period just prior to 5.5 million years ago, which
saw the acceleration of evolution resulting in the appearance of the ¬rst
hominids, was one of global cooling and aridity (Williams et al. 1993). Fur-
ther instances will be cited below when major evolutionary developments
occurred under relatively arid conditions. It is only from the period begin-
ning about 4.5 million years ago that a rather more comprehensive picture
becomes available (Fig. 8). Virtually all the important fossils which illus-
trate this process have been recovered from sites in eastern Africa and,
subsequently, from southern Africa also. It is now convenient to describe
the different types of hominid that have been recognised in this crucial
time-span, and the theories that have been put forward concerning their
The emergence of humankind in Africa 25

Fig. 7: The principal
subdivisions of the last
5 million years, as de¬ned by
geology, magnetic polarity
and oxygen isotope studies
(for details, see Klein 1999).
Periods of normal (i.e. north)
magnetic polarity are shown
in black, those of reversed
(i.e. south) polarity in white;
there were also many briefer
periods of change which are
not shown here. Oxygen
isotope stages alternate
between comparatively warm
and damp periods (with odd
numbers, shown black) and
cooler arid phases (even
numbers, shown white).
26 afric an archaeolog y

Fig. 8: A chart of climatic ¬‚uctuations during
the last 6 million years (after Foley 2002),
showing how ¬ve principal African phases of
accelerated anatomical and/or cultural
hominid evolution appear to have coincided
with periods of comparative aridity

inter-relationships. Details of individual sites with their archaeological com-
ponents and interpretations will follow in later sections.
A site beside the Aramis tributary of the Middle Awash River in Ethiopia
has yielded a partial hominid skeleton plus fragments of several other
individuals, all now attributed to Ardipithecus ramidus and dated to about
4.4 million years ago (White et al. 1994). Originally attributed to the genus
Australopithecus, this creature had teeth rather like those of a chimpanzee,
although it was probably largely bipedal; it evidently inhabited a moist
The emergence of humankind in Africa 27

woodland environment. Somewhat later, around 4.2--3.9 million years old,
are remains of Australopithecus anamensis from two sites in the Lake Turkana
basin of northern Kenya (M. G. Leakey et al. 1995). Both in dentition and in
gait this creature shows development away from the features of Ardipithecus
(than which it seems to have been signi¬cantly larger) towards those of later
and better-known australopithecines.
A later species, Australopithecus afarensis, is much more comprehensively
represented by fossils from at least ¬ve localities in Ethiopia, Kenya and
Tanzania, all dating between 3.9 and 3.0 million years ago. Details of the
principal occurrences are given below. Here, it is appropriate to note that
A. afarensis was a relatively lightly built or gracile species showing several
features, principally in the teeth and face, which distinguish it from later
australopithecines (Johanson and White 1979; Day et al. 1980). Its bipedal
gait is indicated both by its anatomical features and by two series of well-
preserved hominid footprints at Laetoli in Tanzania (Fig. 12, below; M. D.
Leakey and Harris 1987) which are demonstrably contemporary with fossils
attributed to A. afarensis. Its forelimbs, however, suggest a continued ability
to climb trees. Several features of the skull, both braincase and dentition,
continue to display ape-like characteristics. Cranial capacity was in the order
of 400 cubic centimetres (as compared with at least 1350 cubic centimetres
for a modern person) and adult individuals were generally 1.3--1.5 metres tall.
The more comprehensive fossil record means that A. afarensis is the earliest
hominid for which the extent of sexual dimorphism may be estimated: it
was signi¬cantly greater than in later hominids and closer to that seen
among apes (Richmond and Jungers 1995). It has been suggested that this
may indicate ongoing male competition for mates, rather than long-term
The general trend between 6 and 3 million years ago emphasised bipedal
development, while retaining powerful arms which suggest frequent climb-
ing of trees. Indeed, there are indications that all hominids known from
this time inhabited well-wooded environments, effectively disproving the
former belief that development of bipedalism accompanied a shift in pre-
ferred habitat from forest to savanna. Dentition only gradually evolved away
from the small, thin-enamelled teeth still seen in chimpanzees towards the
heavier forms characteristic of many later hominids. It is noteworthy that
both these developments are attested before there is evidence for any sig-
ni¬cant increase in brain-size. In no deposits of this period has any trace of
artefacts been found.
There is broad agreement that Australopithecus afarensis occupied an evo-
lutionary position that was ancestral to all later hominids. By 3.0 million
years ago australopithecines were present also in South Africa, as A. africanus,
while their East African counterparts are generally designated A. aethiopicus
28 afric an archaeolog y

or Paranthropus aethiopicus (see Klein 1999). Both these South and East African
australopithecines, especially the latter, had more massive teeth than
A. afarensis; some individuals (perhaps the males) had particularly power-
ful jaw muscles extending up the sides of the face to a ˜sagittal crest™ of
bone running along the top of the skull from front to back. Even more

. 1
( 11)