xxfifififlv.‘m1m

m‘

.i

.
_
.

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 
 
 
 
 

.V . t . ‘:.. _ a, _ I ‘ V ‘a:.;._;...:.€t..r; .. 1.1.“: .. . ,1,

 

lllllllllllllllllllllllllllllllllllllllllllllllllllll

31293 00898 7517

This is to certify that the

thesis entitled

HUNTER GATHERERS AND NON-HUNTER—GATHERERS:
A LITHIC ANALYSIS FROM N!OMA, TSODlLO HILLS, BOTSWANA

presented by

Salalenna Greek Gaotshabege Phaladi

has been accepted towards fulfillment
of the requirements for

M ' A ' degree in Wogy
yaw 94/ .MW

Major professor

 

Ms M, W
/ /

O-7639 MS U is an Affirmative Action/Equal Opportunity Institution

LEBMRY
Michigan State
University

‘ J

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or botoro date duo.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU In An Affirmative Action/Equal Opportunity Institution
campus-m

 

HUNTER GATHERERS AND NON-HUNTER-GATHERERS:
A LITHIO ANALYSIS FROM N20HA, TSODILO HILLS, BOTSWANA

BY

Salalenna Greek Gaotshabege Phaladi

A THESIS

Submitted to
Hichigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF ARTS

Department of Anthropology
1991

ABSTRACT

HUNTER GATHERERS AND NON-HUNTER-GATHERERS:
A LITHIC ANALYSIS FROM N!OHA, TSODILO HILLS, BOTSWANA
BY

Salalenna Greek Gaotshabege Phaladi

This thesis describes the lithic variability of the
site of Nioma, Tsodilo Hills, Botswana in the light of
interactions between hunter-gatherers and their neighbors.

Nioma is an agropastoralist site that dates to AD 700 -
A0 1000 and has lithic artifacts and iron, copper, wild and
domestic plant and animal remains. There are also shells
which indicate regional trade.

An analysis of the lithic artifacts shows that tools
were made at the site. Also there are activity areas in
several levels of the site.

Most of the artifacts are made of quartz, which is
locally available.

This thesis provides lithic data on an agropastoralist
site, and that data is often lacking in the literature on
southern African Archaeology. But the data is insufficient
to conclusively answer the question of whether the lithic
artifacts were made by pure hunter-gatherers,

agropastoralists or both.

 

Copyright by
SALALENNA GREEK GAOTSHABEGE PHALADI

1991

To Keitirile MmaLebeya Marope and

Phaladi RaMoloko Kootshotse

iv

ACKNOWLEDGEMENTS

Thanks go to all members of my extended family for the
encouragement they gave me. I would like to single out
Thatayaone, Virginia and Galetshetse Gaotshabege, Segametse
Blundell and Neo Kolanyane.

Again, I thank my Maunatlala Primary School teachers,
particularly the following: Miss Keromeng Morebodi, Mr.
ngetse Bogopa, Mr. King Thomas Mogwai, Mr. Seduke Maifala,
Mr. Philip Sekhute and Mr. Seoforeng.

From Maruapula Secondary School, I thank the following:
Mrs. Moncho, Mr. Olebile Gaborone, Mr. Zachariah Matumo, Mr.
Derek James and Mrs. Margaret Ballinger.

I also thank the teachers who taught me at the London
School of Oriental and African Studies.

Also, I thank Mrs. Kelebone Sekelemane, Mrs. Hallgerd
Dyrsson, Moshe Mookodi, Nobantu Kalake, Mothomahura Selato,
Alfred Dube, A.W. Kgarebe, David Rendoh, Bashi Ikitsing,
Miss Motsumi, and Anita Johnson.

My gratitude also goes to the workers of the Merafe
Road Branch of Barclays Bank of Botswana Limited in
Gaborone. Also, the workers of the Main Office of Bank One
in East Lansing were helpful. Also, I thank my housemates at

the Howland House Co-operative. Again, the community at the

Ingham County Medical Care Facility has been helpful.

I also thank Mrs. Onuma Ezera, Joseph Lauer, Norman
Sauer, Bernard Gallin and David Dwyer, as well as my
Guidance Committee--Lawrence Robbins, Robert McKinley,
Charles Cleland, and Edwin Wilmsen.

Again, thanks go to Alex Matseka, Gocdiwang Motswedi,
Leonard Ramatokwane and Hildi Alhild Hendrickson, as well as
Elda Keaton, Joseph Cousins, Kenneth Ebert, Penny
Butterfield, Glynda Moorer, Helene Jones, Barbara Knuth,
Tammra Wright, Tammy Brakora, William Lovis, Stephen Godek,

Caroline Ang and Pens Ang.

vi

TABLE of CONTENTS

Page
LIST of FIGURES ........................................... xi
LIST of TABLES ............................................. x
INTRODUCTION ............................................... 1

CHAPTER 1

Interactions Between Hunter-Gatherers and Non-Hunter

Gatherers in Southern Africa: A Case Study from the Tsodilo

Hills, Botswana.... ........................... . .......... ..2
Introduction ............................................ 2

The Debate: Rousseauians versus

The Pragmatic Opposition ................... 10
The Isolationist View ................................ ..15
The Pragmatic Viewpoint ............................... 34
The View of the Botswana Scholars ...................... 56
Conclusion... .................................... . ..... 59
CHAPTER 2
The Tsodilo Hills--Setting of the study ................... 63
CHAPTER 3
A Summary of Previous Archaeological Excavations in the
Tsodilo Hills Region .................................... ..70
CHAPTER 4
The Site of Ntoma ....... ............ ............ ..........76
CHAPTER 5

Methodology and Definitions of the Categories Used in
Classifying the Lithics from Ntoma ........ . ..... .... ...... 81

vii

CHAPTER 6

Comments on the Tables of Artifacts by Levels,

for the Eight Different Areas of the Site of Nioma ........ 87
CHAPTER 7

Conclusion............ ..... .... ....... . ........... . ...... 103
APPENDICES

Appendix 1
Codes for N!oma Artifacts ............................. 114

Appendix 2
Nioma Artifacts-List of Cateogories that
I grouped into smaller divisions for ease of

classification.. .................... ....... ........... 119
FIGURES.. ................................................ 121
TABLES... ..... . .......................................... 135
REFERENCES CITED ......................................... 151

viii

LIST of FIGURES

Page
Figure 1
Some of the countries in the area of discussion ....... 121
Figure 2
Some of the prehistoric sites in Zambia ............... 122
Figure 3
Some archaeological sites in Zambia and Malawi ........ 123
Figure 4
Some of the archaelogical sites in Zimbabwe ........... 124
Figure 5
Some of the dated archaelogical sites in Botswana ..... 125
Figure 6
Tsodilo and part of its environs ...................... 126
Figure 7
The Tsodilo Hills ..................................... 127
Figure 8
Nioma-areas. .......................................... 128
Figure 9
Blades ................................................ 129
Figure 10
Cores ................................................. 130
Figure 11
Edge damaged .......................................... 131
Figure 12
Retouched and scraper ................................. 132
Figure 13
Crescents; backed and retouched ....................... 133
Figure 14
Informal knives .................................... ...134

ix

LIST of TABLES

Table of artifacts by material, showing variability
within areas

Page

Table 1 .......................... . ....................... 135
Table of artifacts (rows) by level (columns) - Area 3
Table 2 .................................................. 137
Table of artifacts (rows) by level (columns) - Area 4
Table 3 .................................................. 138
Table of artifacts (rows) by level (columns) - Area 5
Table 4... ........... . .. ..... .................... . 139
Table of artifacts (rows) by level (columns) - Area 6

' Table 5 ........................................ . ......... 141
Table of artifacts (rows) by level (columns) - Area 7
Table 6........... .................... ...... ............. 143
Table of artifacts (rows) by level (columns) - Area 8
Table 7 .................................................. 144
Table of artifacts (rows) by level (columns) - Area 9
Table 8 .......................................... . ....... 146
Table of artifacts (rows) by level (columns) - Area 10
Table 9 .................................................. 148
Table of artifacts by material
Table 10 ................................................. 149
Table of artifacts by material, showing variability across
site according to artifact category
Table 11 ................................................. 150

HUNTER-GATHERERS AND NON-HUNTER-GATHERERS: A LITHIC ANALYSIS

FROH N!OHA, TSODILO HILLS, BOTSNANA

INTRODUCTION

Human beings have lived by hunting and gathering for a
long time, but this adaptation has been disappearing. In
order to ask some questions about how human groups organize
themselves, it helps to study groups that hunt and gather.
This subject causes debate, so this thesis will deal with
the debate about the nature of hunting and gathering
societies and their relation with their neighbors. The focus
of the study is an archaeological site in the Tsodilo Hills
of northwestern Botswana. The Hills will be described, as
well as the archaeology of the region. A description of the
site of Nioma will also follow, after which I will mention
how I analyzed the lithics from Ntoma. Next, I will comment

on the analysis of the lithics, and draw conclusions.

CHAPTER 1

INTERACTIONS BETNEEN HUNTER-GATHERERS AND NON-HUNTER-
GATHERERS IN SOUTHERN AFRICA:

A Case-Study from the Tsodilo Hills

Introduction

The aim of this thesis is to explore the relationships
between hunting and gathering peoples and their non-hunting
and gathering neighbors by looking at lithic artifacts from
an important early Iron Age village located at the Tsodilo
Hills in northwestern Botswana. Radiocarbon dates show that
the site of Nioma dates to between ADTOO and AD1000. This
site contains evidence for agropastoralism, use of iron
tools and participation in long-distance trade. The site of
Ntoma also contains thousands of Late Stone Age artifacts.
Part of the thrust of this thesis is to describe the lithic
variability of the site of Ntoma. The purpose of this
description of the lithic artifacts from Nioma is to help
fill a gap in our knowledge of lithic data on Early Iron Age
sites, for detailed classification and presentation of such
data on Early Iron Age is often lacking. At the same time,

the analysis permits of some basic exploration of the

3

overall problem of the nature of the interaction between
hunter-gatherers and non-hunter-gatherers. Below follows an
exploration of the general theoretical background of the
question of hunter-gatherers and their relations with their
neighbors. This theoretical background will be followed by a
description of the Tsodilo Hills of northwestern Botswana,
where the site of Nzoma is located. This archaeological site
serves as a case-study for looking at the question of
hunter-gatherers' relation with their neighbors.

The question of the nature of contacts between hunting
and gathering peoples and non-hunters and gatherers is a
fascinating one. According to Lee and DeVore (1968:3),
during almost 99% of the time that human beings have existed
on planet Earth, most of their livelihood has depended on
hunting and gathering. Thus, if we view success in terms of
longevity, subsequent societal processes pale into
insignificance when we view them in terms of their success
as humankind's modes of production. These include the
processes of sedentary agriculture and the concomitant
developments of settled dwellings and more complex ways of
culture. Additionally, there is the more recent (in terms of
prehistory) process of industrialization (Leacock and Lee
1982:5). Therefore, the question of how hunter-gatherers'
ways of life articulate with those of their non-hunting and
gathering neighbors bears more than academic significance.
For example, among present hunters and gatherers, we as

anthropologists may find avenues of understanding how

4

humankind lived in the past (Leacock and Lee 1982:1-2,13;
Lee and DeVore 1968:vii,3-4,12; Schrire 1984:xiii-xiv).Let
us note this change in the life of hunters and gatherers.
Pure hunters and gatherers living in a world of hunters and
gatherers no longer exist, and the remaining hunter-
gatherers are fast disappearing (Lee and DeVore 1968:4).
Today, most hunters and gatherers have been involved in
processes of assimilation into the wider world of which they
form a part (Schrire 1984:xv). In various parts of the
world, this process of assimilation has proceeded at varying
speeds, but the direction seems to be one of incorporation
(e.g. Yellen 1985). It would appear that the basic processes
governing interaction between the two societies would be the
same in the various parts of the world, with slight
differences pertaining to local circumstances. Today, there
is a web of complex interrelationships between hunters and
gatherers and their neighbors (e.g. Bahuchet and Guillaume
1982; Vierich 1982; Chang 1982 and Lee and Hurlich 1982).
These neighbors may herd animals, practice agriculture, live
in mission stations, villages, towns or cities, or on farms.
This diversity and complexity has contributed to a lively
debate among anthropologists about the nature of contacts
that existed in the past. The question, "How did Prehistoric
Humans live in the past?" is a baffling one and raises a lot
of disagreement.

To narrow the geographical focus, this present

discussion of contacts between hunter-gatherers and their

5
neighbors will deal primarily with the contact situation in
Southern Africa. References to other parts of the world will
be incidental and not detailed (see Figure 1).

In the question of contacts between hunter-gatherers
and non-hunter-gatherers, many anthropologists have been
critical about the way hunter-gatherer studies have been
used (e.g. Barnard 1983, Chang 1982, Denbow and Hilmsen
1986, Parkington and Hall 1987). Wobst (1918) raises a
number of important and pertinent issues which I will
summarize below. I will stick closely to his text.

Nobst maintains that the spatial, temporal and
behavioral constructs that order archaeologists'
understanding of past human "behavior and its precedents and
products" contain serious limitations (Nobst 1978:303)
Basically, the shortcomings derive from ethnographic
fieldwork's incapacity to deal sensitively with the wider
ramifications of hunter-gatherers' behavior. Therefore,
Nobst argues against an uncritical acceptance of
archaeological research that is founded on "expectations,
implications, and measurements derived from ethnography"
(Nobst 1978:303).

A quote from Hobst will set the tone of his argument.
He argues that all hunter-gatherers that ethnographers have

studied

...were intimately tied into continent-wide cultural

matrices, be it through the world market or through

6
other direct and indirect contacts with more complex
societies. Thus, the ethnographic record should be a
veritable gold mine of information on regional and
interregional process, among hunter-gatherers and
between hunter-gatherers and other populations. Yet,
the literature is remarkably silent on the behavior of
hunter-gatherers as it articulates within larger social

and spatial entities [Wobst 1978:303-304].

Apart from being dissatisfied with this curious state
of affairs, Wobst continues to analyze it. He argues that
preoccupation with salvage ethnography and the "ethnographic
present" (Wobst 1978:304) has stimulated the blooming of the
parochial model of hunter-gatherers.

In the first case, salvage ethnography dealt with
observing how hunter-gatherers differed from the
agriculturalists who gradually intruded upon the lands and
rights of hunter-gatherers. Wobst maintains that the
rationale of salvage ethnography disclaimed those behaviors
which united hunter-gatherers with their neighbors (Wobst
1978:304). By focussing on locally distinct process, salvage
ethnographers minimize inter- and intra-regional process;
hence the high profile of the parochial model of hunter-
gatherers.

Another factor that contributes to homeostasis in
studies of hunter-gatherers is ethnographers' concern with

the "ethnographic present." Here, Wobst indicates that the

7
ethnographic present refers to the unreal moment in time
when hunter-gatherers were affected by culture contact to a
smaller degree (Wobst 1978:304). By filtering out the
results of cattle ranches, forts, trade and missions, the
effect is to limit processes to really narrow "units of
social, economic, and geographic space" (Wobst 1978:304). By
not situating these processes in their wider regional and
interregional context, "ethnographers are forced to
attribute what remains to the most local and internalized
stresses" (Wobst 1978:304).

What Wobst says is that ethnographers' temporal and
spatial constraints of their fieldwork heighten the
resilience of the narrow model of hunter-gatherers. He
argues that it is easy for Lee (1969) to talk of the "two-
hour-walk territory" or Jarman (1972) to put forward the
”catchment area" because they conceptualize hunter-gatherers
in terms of limited periods of time. These constraints limit
ethnographers' length and breadth of sensitivity towards
understanding how hunter-gatherers respond to major
stresses, hazards, and catastrophes. Failure to gain this
perception leads to ethnographers' attribution of behavior
to what is within their narrow range of vision: "small group
dynamics, small units of space, and temporal and spatial
variability of low amplitude" (Wobst 1978: 304). Thus
ethnographers pass on information which archaeologists,
without using strong inference, then base their research on

(Platt 1964). In this way, archaeological research does not

8
challenge the "overwhelming ethnographic stereotype that
hunter-gatherers articulate exclusively with local
variability, and that regional and interregional process
among hunter-gatherers is a symptom of degeneration and
culture contact" (Wobst 1978:304).

Conversely, Wobst maintains that it is among hunter-
gatherers that we find conditions which are conducive to
broad processes covering a wide geographical area, rather
than the narrow area that the parochial model puts forth.
Wobst says hunter-gatherers have more flexibility of
movement than agriculturalists. The former's ease of
movement arises from hunter-gatherers' exploitation of a
broad range of food sources. Also, they have weakly
developed facilities, storage, ownership, and claims to
land. In addition, the game and plant foods that hunter-
gatherers exploit are not limited to a restricted
geographical area; they are spatially continuous. Moreover,
other hunter-gatherers and predators from other areas
compete for these resources. Therefore, Wobst concludes that
we should find a great deal of interdigitation among hunter-
gatherers and their neighbors (Wobst 1978:304-305). He
asserts that the hunter-gatherers' resources and personnel
are not "confined to the small unit of space that some
general models want to allot to hunter-gatherers" (Wobst
1978:305).

Wobst also says that another dimension to the parochial

model of hunter-gatherers is the reliance of ethnographers

9
on informants. The latter depend on observation and hearsay
in making sense of behavior. It then follows that in dealing
with the behavior of other groups, much invention goes on.
That arises out of informants' inability to know what other
groups' behavior is really like, apart from the patterned
and homogeneous conduct that is displayed in public (Wobst
1978:305-306).

Wobst suggests that archaeologists must predict
"variability of behavior in all of its spatial and temporal
dimensions" (Wobst 1978:307). He lays the responsibility on
Archaeology's lap because it alone among the various sub-

disciplines of Anthropology deals with

information about behavioral variance in all of its
dimensions: in personnel from single individuals in
private to the largest structural poses; in space from
the smallest catchment area to the largest continent-
wide population matrix, and in time from single events

to millennia [Wobst 1978:3071.

Thus, archaeologists have to create "theory more
permissive of behavioral variability and more sensitive to
culture process in all of its spatial dimensions“ (Wobst
1978:307).

Having outlined some of these theoretical
considerations, I would like to discuss the views of some of

the scholars involved in the debate of how past hunter-

10
gatherers and their neighbors related to each other in their
day-to-day existence. While we are dealing with the sub-
discipline of Archaeology within Anthropology, the evidence
for such prehistoric contacts is by nature diverse. I say
"by nature diverse" because a lot of the evidence relates to
many disciplines. We have an array of information and
knowledge gleaned from oral history, linguistics,
geomorphology, physical anthropology, ethnography,

archaeology and archival sources.

The Debate: Rousseauians versus The Pragmatic Opposition

We may compare and contrast the views that researchers
have about the relative degree of isolation of hunter-
gatherers in several ways. One of my prime considerations is
the focus on Southern Africa, so some noteworthy scholars of
hunter-gatherers do not meet this criterion - e.g. Colin
Turnbull and James Woodburn, who have worked primarily in
Central and East Africa (e.g. Turnbull 1961, 1983; Woodburn
1970). Another criterion I used in discussing the question
of contacts between hunter-gatherers and their neighbors is
the attitude that I feel researchers take towards human
nature. For the purposes of this discussion, I have chosen
to group researchers into two categories - those who take a
Rousseauesque attitude towards hunter-gatherers and those
whose attitudes I will call pragmatic. In using the terms

"Rousseauesque camp" and the "Pragmatic camp," I am using

11
them for convenience. My aim is not to denigrate the
contributions of the individuals in this debate on hunter-
gatherers. Moreover, I do not mean that the people in each
camp always agree with each other. Again, it must be borne
in mind that there may be overlap between these two camps,
for they are engaged in the pursuit and furtherance of
knowledge, and the nature of the academic enterprise
consists of debate and confrontation, as well as
collaboration.

In general, the Rousseauesque "school of thought" sees
modern hunters and gatherers as exemplifying man's original
state of nature. These researchers see that state of nature
as having been benign and full of affluence. This is seen as
an idyllic state of nature, so that we may say, without
meaning to be pejorative, that hunters and gatherers were
noble savages. The Rousseauians tend to treat modern hunters
and gatherers as though they have not been subject to the
winds of change that have been blowing across our planet. It
is as if hunter-gatherers have lived an unchanging existence
for millennia. They are seen as representatives of aspects
of human beings' existence that are fixed and only change
now when hunters and gatherers come face to face with the
dictates of the industrializing, mechanizing and uprooting
modern world.

In the Rousseauesque camp, we have Richard Lee, Sheryl
Miller, David Phillipson, George Silberbauer, Jiro Tanaka

and John Yellen. To illustrate some of the points of view of

12

this "school, here is a quote from Richard Lee:
...life in the state of nature is not necessarily
nasty, brutish, and short. The Dobe-area Bushman (of
Botswana) live well.today on wild plants and meat, in
spite of the fact that they are confined to the least
productive portion of the range in which Bushman
peoples were formerly found. It is likely that an even
more substantial subsistence base would have been
characteristic of these hunters and gatherers in the
past, when they had the pick of African habits to

choose from [Lee 1968:43].

Now, Jiro Tanaka commented on the G/wi of the Central

Kgalagadi that:

...even nowadays most of them still live a life of
hunting and gathering away from the influence of modern
civilization, relying on primitive "stone age"
techniques... The fact that a group of people with a
population of several thousand is still living in the
same fashion as human societies of almost 10,000 years
ago is a miracle [Tanaka 1974:iii].
At this juncture, allow me to clarify some points
regarding my use of some terms relating to Botswana - the
land and its peoples. First, I would like to use the term

"San" or "Basarwa" in referring to the people that are

13
commonly referred to in the literature as "Bushmen." The
term Bushman has now fallen out of favor and is regarded as
discriminatory and pejorative. As the people of Botswana
largely use the name Basarwa to refer to these hunter-
gatherers, I have chosen to use that name. Also note that
the prefix "Mo" refers to one person, thus a "Mosarwa,”
while the plural prefix is "Ba", thus Basarwa. The prefix
denoting the languages is "Se" - therefore we have Sesarwa.
The different groups that I call Basarwa also have their own
names for themselves, so it is essential to realize that the
names I am using are those that are used by the Setswana-
speaking people of Botswana.

Let me delineate the arguments of the opposing
epistemological camp. This is the "school of thought" that
for the purpose of this discussion, I have called the
Pragmatic Opposition. It is represented by James Denbow,
Robert Hitchcock and Edwin Wilmsen.

One major strand of this camp's argument is the
emphasis they lay on a flexible and open response to changes
in the hunter-gatherers' environment (Denbow 1984; Hitchcock
1979, 1982a, 1982b; Wilmsen 1979, 1982, 1989). Thus, they
stress the hunter-gatherers' readiness to take advantage of
opportunities available to them, in order to maximize their
security. Part of the motivation also lies in enhancing
hunter-gatherers' existence through a wide range of survival

strategies. As Denbow puts it:

14
...people are decision-makers and, as such are capable
of using the same environment in a number of different
ways and of inventing new strategies and technologies
to adapt to changes in ecological and social conditions

through time [Denbow 1984:189].

So, we have a far-reaching and wide-ranging repertoire
that encompasses complex interaction with non-hunting and
gathering peoples, who, given the dictates of circumstances,
adopted the behaviors of hunter-gatherers (Lee 1968:40).
Therefore, according to the second group of researchers, the
line between hunters and gatherers and their neighbors is
not clearcut. It is not exactly clear who is a hunter-
gatherer and who is not, and as Bahuchet and Guillaume say,

even though some hunter-gatherers

...enable us to observe ancient forms of social and
economic organization, they are by no means fossils of
a prehistoric state, magically preserved. It is highly
likely that centuries of proximity and contacts have
had profound influences on both sides [Bahuchet and

Guillaume 1982:1891.

Moreover the "second school of thought" also differs
from the Rousseauians in that Denbow, Hitchcock and Wilmsen
state that there has been environmental change. They say

that the environment in which we find today's hunter-

15
gatherers has not always been the undesired, parched and
inhospitable wilderness that some present-day scholars deem
it to be - so that the Kgalagadi may have been occupied by
non-hunter-gatherers for a long time (Denbow and Wilmsen
1979, 1982, 1989; Hitchcock 1982a:47; cf Phillipson
1969:35). Part of their argument is to cite evidence for

past environmental change in the Kgalagadi Desert.

The Isolationist View_

Sheryl Miller is one archaeologist whose views on
contacts in the sub-continent of Southern Africa I will
follow at length. I have chosen her views to illustrate the
isolationist position that the Rousseauians espouse. Again,
she uses evidence which at the time of writing was new, and
which predates some significant radiocarbon dates (Miller
1969). I am using Miller (1969) as a historical point of
departure and her view was an appropriate view at that time
given the state of research and knowledge. I am not
asserting that her views are immutable, but am using her
research to provide a framework for the debate on contacts
between hunter-gatherers and their neighbors in the past.

In her interpretation of the nature of contacts between
hunter-gatherers and their neighbors, Miller deals with two
periods in the prehistory of Africa, the Later Stone Age and
the Early Iron Age. Some scholars have suggested that terms

like the Later Stone Age and Early Iron Age, as well as

16
Later Iron Age, are ambiguous, and should best not be used
or should be used with care (e.g. Huffman 1982; Parkington
and Hall 1987; Wilmsen, personal communication 1988). I use
terms like the Later Stone Age and Early Iron Age for
convenience. My aim is to illustrate some of the points
raised by discussants of the question of contacts between
hunters and gatherers and their neighbors. Later, I will
discuss some of the problems involved in using
archaeological remains and ethnography to designate the
ethnicity of archaeological sites.

In dealing with Southern Central Africa (the
territories of Zambia, Malawi and the then Southern Rhodesia
or present-day Zimbabwe), Miller associates the Later Stone
Age with autochthonous hunting and gathering groups (e.g.
Miller 1969:89). Therefore, particularly with reference to
Miller's work, I will occasionally use "Later Stone Age
people" to refer to hunters and gatherers. On the other
hand, Miller argues that during the early contact period,
the Early Iron Age people had a more varied material culture
than that of the hunter-gatherers. Components of this Early
Iron Age culture probably included such technological and
economic variables as agriculture, stock-rearing, techniques
of manufacturing iron and pottery, as well as a sedentary
lifestyle based on villages (Miller 1969:81).

This broad repertoire differs from the simpler hunting
and gathering one, which she does not largely describe. That

is probably because she maintains that its cornerstones were

17
hunting and gathering with simple weapons of stone, and
perhaps iron. Also, it may reflect her interest in the Early
Iron Age, as well as more evidence on that period at the
time of publishing her work in 1969. But Miller, a
pioneering researcher, does provide a caveat that certainly
applied in 1969. She says that the dynamics of the Early
Iron Age are not exactly clear, a problem which
archaeologists have yet to resolve (Miller 1969:84). In her
time, it was even more baffling because studies of the
African Iron Age had not reached the sophistication and
intensity of research that now exists in the 19805 (e.g.
Denbow 1979, 1980, 1981, 1982, 1983, 1986: Denbow and
Wilmsen 1983, 1986; Huffman 1970, 1978, 1982; Mgomezulu
1981; Phillipson 1974, 1975a, 1977, 1985).

Essentially, Miller's conclusions regarding the long
contact situation in Southern Central Africa form two
strands. The first component of this structure states that
the Later Stone Age way of life existed for nearly two
thousand years after the Early Iron Age first appeared in
Southern Central Africa (Miller 1969:81). That is to say
that various pockets of the Later Stone Age way of life
persisted in fairly remote areas into fairly recent times.
The second part of Miller's analysis postulates a near-
technological segregation between the two cultures (Miller
1969:89). Under this scenario of cultural isolation,
peaceful barter probably facilitated exchange of a few

material goods. According to Miller

18

. . the two peoples remained culturally distinct
until the present, the hunter-gatherers adopting only a
few innovations from the agriculturalists and
maintaining their basic subsistence pattern intact

until driven to retreat or extinction [Miller 1969:89].

To lend credence to this isolationist analysis, Miller
draws on ethnographic examples from various parts of sub-
Saharan Africa. Those cases show several kinds of contacts
between hunter-gatherers and their non-hunting and gathering
neighbors. From these recent examples, Miller draws some
possible contact scenarios for the past.

Let us now turn closely to the archaeological evidence
as discussed by Miller. Primarily, I will adhere to Miller's
1969 work, and will only add more recent work to clarify a
few of her points, especially those dealing with physical
anthropology. My reasons for dealing at length with Miller's
work include a need to familiarize myself with the sites and
dates she mentions in South-Central Africa. Again, as a
representative of the Rousseauian "school," she
counterbalances the "school" of Denbow, Hitchcock and
Wilmsen. Thus, I have chosen to give a lot of attention to
her work in order to avoid bias towards Denbow, Hitchcock
and Wilmsen.

There follows an examination of the archaeological

evidence from Zambia, as presented by Miller. I will look at

19

the contact situation in Zambia first because of the
relative abundance of sites and dates from this territory.
Moreover, the Zambian contact situation serves as a basis
for comparison with the contact situation in the adjoining
territories of Malawi and present-day Zimbabwe. Therefore, I
will look at several key sites in Zambia, and discuss their
significance.

In terms of contacts between Late Stone Age and Early
Iron Age peoples, a survey of the Zambian evidence shows
variation by region. For example, the southern Zambian
plateau is typified by a quick replacement of the Late Stone
Age by the Early Iron Age, so argues Miller. The reason
could be ecological: good soils and the presence of light
savannah woodland are conducive to agriculture. Therefore,
the argument follows that hunter-gatherers were forced into
less desirable areas (Miller 1969: 82). At this juncture,
one may add that Miller does not provide evidence to show
that the ecological conditions on the southern Zambian
plateau have been existent for so long that they cover the
period of early contacts between the hunter-gatherers and
their neighbors. Moreover, her assertion "that hunter-
gatherers there were soon absorbed into the new economy or
forced into less desirable territory" runs counter to her
premise of the "lack of cultural exchange" (Miller 1969:
82). My argument is that if the hunter-gatherers merged into
the economy of the agriculturalists, then the process cannot

be said to have been one of splendid isolation. The very use

20
of the term "absorbed" suggests a process which had wider
ramifications than Miller's stance of cultural isolation
would lead us to believe.

Regarding the evidence from archaeological sites,
Miller cites several sites and dates (see Figures 2 and 3).
Her basic argument postulates a short co-existence between
the Early Iron Age and the Late Stone Age on the southern
plateau. She cites evidence which suggests that contact
first occurred in Zambia during the early part of the first
millennium A.D.. For example, the Late Stone Age horizon at
Lusu is dated to 400 1_100 B.C. and 75 1_230 B.C. (C-830).
Clark and Fagan (1965) say the Early Iron Age layer above
this Late Stone Age horizon contains shards of Situmpa Ware.
And at Leopard's Hill Cave, the Late Stone Age layer lies
right below an Early Iron Age layer dated to 535 :_125 A.D.
(SR-126) (Miller 1969a).

But as the millennium wore on, Miller argues that the
agriculturalists who replaced Late Stone Age hunter-
gatherers on the southern Plateau (and also had contacts at
the northern border of the country) caused a transformation
in their relationship with the hunter-gatherers. That is to
say, they expanded their territory closer to Late Stone Age
retreat areas (Miller 1969:84).

The valley of the Middle Zambezi River could have
served as such retreat territory. Here, the archaeological
evidence cones from an aggregate of lithic materials and

potsherds, which were apparently in place together. Also,

21
there is the Muchinga Escarpment, which was an "ideal" and
"long-term retreat area for hunter-gatherers," unlike the
short-term retreat area of the valley of the Middle Zambezi
River (Miller 1969:82-83). On the Muchinga Escarpment,
several sites yielded evidence that Miller interpreted as
showing the persistence of hunting as a "mainstay in the
economy, still depending on lithithechnology even after the
introduction of iron" (Miller 1969:83). For example,
Nachikufu Shelter yielded pottery and iron, in addition to
stone artifacts. To substantiate her conclusion that hunter-
gatherers maintained their tradition of lithic technology
1,000 years after meeting Iron Age agriculturalists in the
Muchinga Escarpment, Miller cites the site of Nakapapula.
Here, lithics and pottery from the Late Stone Age level were
dated to 770 1 100 A.D.. Also, Nachikufu Cave provided
similar, supporting evidence of this kind of interaction
between agriculturalists and hunter-gatherers. Again, Nsalu
Cave yielded more evidence of Early Iron Age and Late Stone
Age contact: Late Stone Age material and pottery; stem of an
iron arrowhead; coiled copper wire and several glass beads
(Miller 1969). According to Miller, these findings are
attributable to the fact that hunter-gatherers did not
change their basic subsistence pattern. Then, she argues
that contact went on till late, at least in limited areas.
That is because evidence from Nachikufu Shelter shows a Late
Stone Age level going back to 890 1_95 A.D., and a top

horizon of lithics that are associated with a date of 1750 1

22
100 A.D. (Miller 1969:83).

On the northern plateau of Zambia, Miller's evidence
leads her to suggest that Late stone Age peoples stayed
there for some time following contact with Early Iron Age
peoples. She cites the evidence from Mwela Rocks Shelter,
near Kasama. The battery of evidence is as follows: at 45 cm
below the surface, a Late Stone Age horizon contained
shards. Moreover, there is no major change in the relative
frequencies of stone tool types following the onset of
pottery. Again, continuity of the lithic tradition reaches
the surface of the deposit. In addition, remains of three
iron arrowheads came from the top part of the deposit.
Furthermore, the shards from this site bear resemblance to
Early Iron Age wares from Kalambo Falls, which dates to 345
1_40 A.D. (GrN-4646), and was excavated by J. Desmond Clark
in the extreme north of Zambia (Miller 1969:84). She
speculates that the "poor local soils and dense hgaghxstggia
woodland of the region" may have "discouraged the rapid
expansion of the Early Iron Age agriculturalists there"
(Miller 1969:84).

Similarly, contact occurred in other parts of Zambia.
Thus, Chondwe farm on the Copperbelt produced some "Later
Stone Age type lithic material, Early Iron Age pottery, and
iron slag" (Miller 1969: 84). And in eastern Zambia, the
site of Makwe had shards in its upper levels (Phillipson
1968).

Likewise, we find evidence of contact in Malawi. Miller

23
says that the patterns of interaction in Malawi are
reminiscent of the Zambian contact situation. But she does
assert that because of scanty evidence, it is hard to draw a
clear picture of the Malawi contact situation. However, her

optimism leads her to express the likelihood that

. when more sites and dates are known, the Malawi
contact situation may well be found to approximate that
of Zambia, with Early Iron Age agriculturalists soon
established in favorable regions and Later Stone Age
groups preserving their traditional way of life in

remote areas [Miller 1969:85].

But on the Nyika Plateau, Miller's evidence shows that
contact began at about the same time as it did in Zambia.
Here, at the sites of Chowo and Phopo Hill, Robinson and
Sandelowsky (1968) report the presence of lithics and
Mwavarambo (Mwabulambo) Ware in Early Iron Age levels, which
may overlay Late Stone Age levels. Miller reports evidence
for a similar contact situation between the mountainous
northern region of Malawi and the Muchinga Escarpment of
Zambia. She argues that the northern region of Malawi may
have functioned as a retreat area. Her evidence stems from
J. Desmond Clark's work at Mphunzi Shelter, where several
potsherds and some Late Stone Age lithics were found. More
Mt. Cave also gave supporting evidence of early contact

(Miller 1969:85).

24

In the other territory of the then Southern Rhodesia,
now Zimbabwe, contact also dates to at least the turn of the
first millennium A.D. For example, Calder's Cave, in the
north-western part of the country, yielded pottery dated to
20 1_80 B.C. (UCLA-929). Evidence for this contact comes
from various parts of the country (see Figure 4). For
example, in the Fort Victoria area, Early Iron Age sites
show occupation by agriculturalists. Citing Robinson (1961),
Miller says that the site of Mabveni produced a wide range
of Early Iron Age artifacts, in addition to Late Stone Age
components: pottery from the eponymous site of Gokomere; hut
and granary remains; iron and copper; and ostrich egg-shell
and snail-shell beads of Late Stone Age type (Miller
1969:85). Mabveni is dated to 180 1_120 A.D. (SR-43), while
Robinson suggested an anomalous date of 570 1_110 A.D. (SR-
79) which may have been due to intrusion or contamination.
Again, another Gokomere Ware site had elements similar to
Mabveni, except for the fact that the date of earliest
occupation is much later: 540 i 120 A.D. (SR-26).

Another difference relates to the absence of typically
Late Stone Age artifacts (Robinson 1963). Miller concludes
that maybe "by this date the later Stone Age peoples no
longer inhabited the area" (Miller 1969:85).

In north-east Zimbabwe, the Inyanga region has some
Early Iron Age Ziwa Ware sites that date to the beginning of
the fourth century A.D.. And Miller cites Summers (1967) as

saying the artifacts found included pottery, iron, gold and

25
Late Stone Age elements like bone points and ostrich egg-
shell beads (Miller 1969:85).

Miller's interpretation of the above sites is that
Early Iron Age people were settled in Zimbabwe by the middle
of the first millennium A.D.. As in most cases, Late Stone
Age material culture elements are found in the earliest
agricultural horizons; Miller argues that this is more
evidence for contacts. As these Late Stone Age material
culture elements later disappear altogether, Miller says
that means these items of material culture were made by Late
Stone Age people who later disappeared. She suggests that
the Early Iron Age people did not make these items (Miller
1969:86).

There is also evidence to show that as in Malawi and
Zambia, Late Stone Age people retreated to remote areas;
this followed contact with immigrant agriculturalists. Sites
in the Matopo Hills of southwestern Zimbabwe, e.g. Tshangula
(Cooke 1963) and Bambata (Schofield 1940) contain Late Stone
Age horizons, some of which contain pottery. And Robinson
(1966) tentatively suggests that Bambata Ware may represent
contact with Early Iron Age people. In addition, some
topmost horizons of Late Stone Age deposits in the Matopo
region have pottery styles that "indicate that hunter-
gatherers remained there for at least several centuries
after their initial contact with" Early Iron Age groups
(Miller 1969:86).

Let us examine the types of contact that existed

26
between Late Stone Age and Early Iron Age peoples during the
centuries of contact. First, we need to note archaeological
data alone is insufficient to give a clear and broad picture
of the nature of the interrelationships that existed between
Late Stone Age and Early Iron Age peoples. Miller argues
that there was peaceful co-existence, as evidenced by the
presence of potsherds and rare metal objects in Late Stone
Age horizons. She also says that another indicator of
peaceful trade is the presence of Late Stone Age shell beads
and bone points sometimes found in Early Iron Age deposits.
Modern analogies for peaceful trade come from the Ituri
forest, and here Miller cites Turnbull (1961), who says the
Mbuti exchange their products for those of their
agricultural neighbors. And according to Fagan (1967), the
presence of wild animals in Early Iron Age deposits
indicates that hunting had a supplementary role to the
domestic economy. Perhaps agriculturalists traded their
goods for game when they encountered Late Stone Age peoples.
It is also conceivable that Early Iron Age people may have
been involved in hunting. In times of stress, when domestic
animals and domesticated crops failed, the Early Iron Age
people may have hunted. I would say Miller does not
seriously consider this possibility, and in parts of
Botswana where there are no hunter-gatherers, the
agropastoralists sometimes supplement their economy by
hunting. I

Still on the question of peaceful trade, Miller

 

I‘ll

DI

tr

DC

6'»
Ir
tr
In
DP
ma
thl

Ict

$0,

27
suggests that honey may have been an item of trade between
the Early Iron Age and Later Iron Age peeples. Here, she
postulates that the Late Stone Age peoples probably provided
the honey, while they received the agriculturalists' produce
in return. I would say that one may add that in this
exchange scenario, we need to note that the agriculturalists
may have collected the honey themselves. I do not think that
this item of trade was necessarily the preserve of the Late
Stone Age people, for the agriculturalists may conceivably
have learned how to hunt for honey.

Miller also seems to contend that the Late Stone Age
people did not know how to make pottery. Thus, they probably
traded it from the agriculturalists. She further states that
pottery found in Late Stone Age horizons tends to be
typologically similar to pottery from Early Iron Age sites
(Miller 1969:87).

Concerning iron-smelting, Miller argues that the
evidence from Nachikufu shows that Late Stone Age people
knew how to smelt this metal. Here, Miller goes against the
traditional view that iron-smelting, like pottery, was not
known to the hunter-gatherers. She suggests that the
presence of a few iron objects in a Later Stone Age horizon
may lead to the inference that trade occurred. Conversely,
the presence of slag and even furnace remains indicates
actual smelting on the site, carried out by itinerant Iron
Age craftsmen or by the Late Stone Age occupants themselves,

so Miller argues. She argues that Late Stone Age lithic

 

 

Iecl
Sto:
COm
sme
VIII
5%!
pr

III'

28
technology coexisted with iron-smelting done by the Later
Stone Age people themselves (Miller 1969). The evidence
comes from Nachikufu. But Miller's assertion that iron
smelting was known to the Late Stone Age hunter-gatherers,
while pot-making was not, seems to be contradictory in the
sense that both creative methods, iron-smelting and pottery,
probably went together. A people that could smelt iron,
which is a complex technological process that entails a high
degree of cultural innovation, was probably able to make
pots as well. Therefore, for Miller to argue that Late Stone
Age people did not make pottery themselves, but knew how to
smelt iron, does not seem to follow a logical process.

Turning to the extra-commercial relationships of the
Late Stone Age and Early Iron Age peoples, Miller faces a
dearth of knowledge. It is unclear exactly how displacement
of the hunter-gatherers by the agriculturalists occurred,
she argues. While the hypothesis of major skirmishes is
unsupported, rock paintings in Zimbabwe depict small-scale
fighting. Therefore, Miller argues that it is possible that
the inchoate Early Iron Age involved minor skirmishes
between the Late Stone Age hunter-gatherers and the Early
Iron Age agriculturalists.

Regarding the evidence from Physical Anthropology,
Miller states that there is not much evidence from this sub-
discipline as far as the characteristics of each group are
concerned. She attributes this lack of knowledge to scanty

skeletal evidence. But we also need to realize that Physical

29
Anthropologists do not agree regarding the degree to which
"races" can be determined from skeletal material.

Aside from the debate among Physical Anthropologists,
Miller also states that some researchers disagree regarding
the "stock" of the hunter-gatherers and the
agriculturalists. She cites Gabel (1965) as saying the
hunter-gatherers could have been of Khoi-San stock, while
the agriculturalists had the physical characteristics of
Bantu-speaking peoples. But, according to Fagan (1967), the
evidence from very early agricultural sites is insufficient
to confirm the preceding interpretation.

In a later context, Rightmire (1970, 1975) delineates
some of the problems dealing with the identification of sub-
fossil skulls and other skeletal remains from prehistoric
sites. There are difficulties pertaining to dating,
preservation and provenance. Then, again, Rightmire states
that the origin of the Negro is uncertain, and more work
needs to be done to shed light on these matters. His work
disputes some of the claims that have been made regarding
the ascription of some skulls disinterred from Iron Age
sites to a mixed stock with Khoisan, Negro and "Boskop"
features. He disclaims the presence of a Boskop race as a
myth. Likewise, Schepartz (1988) disputes the assertion that
the later Pleistocene Eastern Africans were of Khoisan
stock. Instead, Schepartz's skeletal work suggests they were
tall and linear peoples. Therefore, we need to realize that

there is a great deal of debate regarding the interpretation

30
of skeletal remains from the prehistoric sites in Africa and
the physical characteristics of the groups that inhabited
those sites. It is likely that as more work proceeds, other
interpretations will also arise to challenge Miller's
interpretation.

Another area of uncertainty pertains to the degree of
mixing of people from the different cultures. According to
Rightmire, in this case of "long-term contact between
divergent human populations, exchange of both genes and
cultural traits must have played some role" (Rightmire
1970:148). Miller also shares this view, insofar as she
states the probability that at the level of the individual,
physical mixing occurred. At the group level, mixing may
have occurred when one group became acculturated without
losing its physical identity. However, as indicated earlier,
Miller subscribes to an isolationist view of interactions
and contact between these two population groups. Thus, her
interaction scenario postulates a situation in which
separate identities remained intact. She argues that each
culture retained its identity and traditions, adopting only
a few new material elements from the other culture. So,
after learning to smelt iron, the Late Stone Age people
still maintained their lithic technology until they had to
retreat into remote areas, were forced into extinction or
became incorporated into the agriculturalists' social system
as time wore on. Only a few hunter-gatherers were not

incorporated, hence the persistence and existence of San

31
hunter-gatherers (Miller 1969:89) in South Africa, Botswana,
Zambia and Zimbabwe, and by extension, Angola and Namibia.

Another scholar who subscribes to this isolationist
view of contacts and interactions between Late Stone Age
hunter-gatherers and Early Iron Age agriculturalists'is
David Phillipson. Based on his extensive research into the
prehistory of Eastern and Southern Africa (e.g. Phillipson
1968, 1974, 1975a, 1977, 1985), Phillipson argues that the
Early Iron Age Industrial Complex spread into the central
and southern parts of Africa in two distinct streams. The
Eastern Stream, which like the Western Stream is based on
pottery typology and the possession of cattle, is
represented by regional groups in northern Zambia, eastern
Zambia, Malawi and Zimbabwe. The above stream covers a wide
geographical region (including Kenya, Tanzania, and the
Republic of South Africa and Botswana and Mozambique).

The Eastern Stream is the better-known one, in terms of
chronology and geographical distribution. Using a battery of
radio-carbon dates, Phillipson (1977) argues that the Early
Iron Age progressed in a north-to-south movement from East
Africa. Its representative sites in the regions above had
been settled by Eastern Stream peoples around the fourth
century A.D.. But the Western Stream sites show a settlement
by people at a later date, after the turn of the fifth
century A.D..

Regarding contacts and interactions between the

indigenous hunter-gatherers and the immigrant metal-using

32
farmers, Phillipson argues that relations between these two
population groups were characterized by aloofness or mutual
avoidance, albeit not in all areas (Phillipson 1985:180). He
further says that at the Zambian sites of Makwe, Thandwe and
Nakapapula, the lithic material buttresses this isolationist
interpretation. That is, there were no big typological
changes signifying significant contacts between the two
population groups. Only minimal contact is suggested, he
maintains (Phillipson 1985:180-181).

Prior to the final displacement of the hunter-gatherers
or their conquest or absorption, Phillipson postulates a
scenario that included client relationships between the
hunter-gatherers and the cattle-keeping farmers. He borrows
this idea from Silberbauer (1965). Ultimately, Phillipson
sees the expanding population of the second millennium A.D.
bringing about greater competition for land between the two
populations. Consequently, the hunter-gatherers were pushed
into "the few areas which were unsuitable for farming"
(Phillipson 1985:181).

Now I will delineate some of the features of hunter-
gatherers' contacts with pastoralists and other neighbors,
as seen by other members of the Rousseauian school such as
Richard Lee, John Yellen and George Silberbauer. They all
have worked among San communities in the Kgalagadi Desert of
Botswana. This is the area that is normally referred to as
the "Kalahari" Desert. However, the Setswana term for this

area is "Kgalagadi" which means "the land of thirst." This

WC

9:
he

he

an

93

re:

3“

33
noun is derived from the Setswana verb "go kgalega," which
means "to be thirsty." (Morulaganyi Kgasa, personal
communication December 1988). That is why I use the term
"Kgalagadi" rather than the misnomer "Kalahari," as the
Setswana term is more appropriate.

Apart from having a common geographical area of
research interest, the above three scholars also see the
Kgalagadi as a place that the San retreated to, where,
according to Lee, "life in the state of nature is not
necessarily nasty, brutish, and short" (Lee 1968:43). As
indicated, he further states that the San people of his
research area, Dobe, in north-western Botswana, subsist
sufficiently on wild flora and fauna. Thus, he continues,
"it is likely that an even more substantial subsistence base
would have been characteristic of these hunters and
gatherers in the past, when they had the pick of African
habitats to choose from" (Lee 1968:43). Likewise, Yellen
(1985) and Silberbauer (1981) see the Kgalagadi Desert as
having been a retreat area that the San went to when other,
more aggressive groups avoided it because of its
inhospitable climate.

It is not entirely true to say that all the above
anthropologists categorically see the modern San hunter-
gatherers as replicas of an ancient and unchanging way of
life. What can be said is that there is a belief that they
represent a Rousseauesque type of "noble savage.” They are

seen as representatives of Sahlins's “original affluent

34
society" (Lee and DeVore 1968). Hence, the point of
departure in the above scholars' work is to study the San of
Botswana because they have not changed much in their way of
life. So, at the time of the "ethnographic present," the San
are taken as not having had much contact with their

neighbors.

The Pragmatic Viewpoint

This traditional picture of San hunter-gatherers as
exemplars of an unchanging existence in the Kgalagadi comes
under attack from James Denbow (1982, 1983, 1986a, 1986b),
Robert Hitchcock (1982a) and Edwin Wilmsen (1979, 1982). For
the purposes of this discussion, I will refer to this group
of researchers as the Pragmatists. The thrust of their
studies postulates a large degree of symbiosis between the
hunter-gatherers and their neighbors. Rather than seeing the
hunter-gatherers as isolated people who did not mix with
their neighbors, the Pragmatist "school of thought" argues
for interdigitation. Denbow, Hitchcock and Wilmsen argue
that the history of interaction in the Kgalagadi shows an
intermeshing of different lifestyles involving the hunter-
gatherers and their herding and agricultural neighbors
(Denbow 1984; Denbow and Wilmsen 1986; Hitchcock 1982). In
other words, flexibility characterized the contacts and
interactions between the peoples in this area. Denbow,

Hitchcock and Wilmsen argue that as pragmatic dwellers of

35
the Kgalagadi, the San, just like the other groups they
interacted with, made decisions based on maximizing the
benefits that would accrue from the choices they made. Thus,
we find that in the Kgalagadi, the various groups involved
lived interdependent lives. They did so, the Pragmatists
argue, in order to take advantage of opportunities that were
available in that environment. As one protagonist puts it:
"Hunter-gatherers are opportunists who take advantage of
circumstances as they arise" (Hitchcock 1982:61).

Before proceeding with an examination of the data
pertaining to the evidence for at least 1,500 years of
interaction in the Kgalagadi, I will consider the concept
"Mosarwa" or "Bushman." Here, I will deal with Yellen's
insights on this question, and finally, I will add some
points regarding his critique of this term and its
importance insofar as contacts and interactions are
concerned (Yellen 1985).

As indicated earlier, "Mosarwa" is the singular form
for a San person, while the plural term is "Basarwa." These
terms are used in Botswana, and because the term "Bushman"
is pejorative and discriminatory nowadays, the terms
"Basarwa" and "Mosarwa" are preferred for the purposes of
this discussion. Therefore "Mosarwa" or "Basarwa" can be

used interchangeably with "San,' which is used in the
anthropological literature to designate the hunting and
gathering people that were referred to as "Bushmen."

Now, "Basarwa," as used in Botswana is a term that

36

agrees with Schapera's (1930) term "Khoisan. Basarwa" are
speakers of "Bushman" and "Hottentot" languages, and
"Sesarwa" is the term used in Botswana to refer to the
languages of the Basarwa.

According to Yellen (1985), the concept "Mosarwa" needs
some elaboration. It is hard to define, even though it may
be seen as a clearcut category in theory. A fourfold
definition based on criteria of language, physical
characteristics, lifestyle and self-identity is often
untenable in reality. He states that a Mosarwa is supposed
to speak a "click" language that belongs to the Khoisan
language family of African languages. Although genetically,
the Khoisan languages are not related to Bantu languages,
the Bantu languages of Zulu and Xhosa do contain clicks in
their phonological make-up (Yellen 1985:15). (But note that
Westphal (1963:242) says that other Bantu languages with
clicks are Swati, Ndebele, Suthu (Sotho), Yei, Kwangari,
Gciriku and Mbukushu.)

Second, Yellen states that the definition of a Mosarwa
in Botswana means having a livelihood that encompasses
nutritional dependence on hunting and gathering wild foods.
Therefore, a certain amount of "backwardness" or eking out a
living based on a "rural" or "traditional" lifestyle is
associated with Basarwa (Yellen 1985:15). But Yellen further
points out that this definition runs counter to much of the

acculturation that has characterized relations between

Basarwa and non-Basarwa. His argument is that today's

37
Basarwa and their neighbors have interrelated closely since
they began living together in the same environment.

Third comes the genetic argument. According to Nurse
and Jenkins (1977), there are physical characteristics which
are peculiar to Basarwa peoples. These include such
descriptions as

Basarwa are generally short, have yellowish rather than

blackish-brown skin, tightly coiled "peppercorn" hair,

a malar bone conformation which produces a flat face

which is small in relation to the cranium, and a skull

shape which is pentagonal when viewed from above

[Yellen 1985:15).

The fourth criterion relates to self-identity. Yellen
states that an individual will indicate his belonging to the
general Basarwa group or will say he belongs to one of the
Basarwa "tribes" (Yellen 1985:15).

But Yellen disputes this four-fold definition as being
out of touch with the concrete reality of Basarwa
integration into the Botswana social and economic system
(Yellen 1985). In brief, he argues that there has been a
long process of interaction between "pure” Basarwa hunter-
gatherers and Tswana, Herero, English and Afrikaner groups
in the past. So he argues that Silberbauer's (1981) and
Tanaka's (1980) works on those Kgalagadi interactions
minimize the degree of interaction (Yellen 1985:1b).
Similarly, genetic markers as distinguishing features of

Basarwa are also tenuous as shown by the case of the "Black

38
Bushman" of the Okavango Swamps. While they consider

themselves Basarwa, and speak a Khoisan language,

physically, they are "indistinguishable from non-Basarwa

Bantu speakers" (Yellen 1985:16). Also, in the Nata area,

Chasko et al. (1979) found that the Denesana, a Basarwa
group, are genetically closer to black African groups than
to San groups. The implication is that related groups in
Botswana, Zambia and Zimbabwe may be "biologically Negro
rather than San" (Chasko et al. 1979:18).

Yellen further questions the neat and clearcut textbook
case of ideal Basarwa. He substantiates his argument by
SCY‘ing that many people in the Kgalagadi area of Botswana
are multi-lingual. Thus, Basarwa may speak Herero and
Setswana in addition to a San language. Also, Herero and
Tswana people may speak a San language. This is esppcially
true in areas where there is a significant number of Basarwa
in the area. Moreover, in cases of intermarriage, Yellen
a"‘St-Ies that the question of ethnic identity may not be an
°V°Priding issue. That is, the concerned person navigates
his Worlds in a manner which makes him feel at home in a
given situation (Yellen 1985:16).

The above points made by Yellen are pertinent in the
3°“se that the concept of a "Mosarwa" is fluid in reality.
“t as an analytic unit, it may appear to be more cut-and-
drin. In fact, the protagonists and antagonists of this
qu°stion of contacts between farmers and hunters and

98therers agree on this question. This is evident insofar as

Ba

Ba
Fe
an;

Ulla

39
they recognize the vicissitudes of terms like "Bantu,"
"Basarwa" and "Bushman" (Hitchcock 1982b).

Actually, Yellen tends to romanticize the relations
between Bantu-speaking people and Basarwa in Botswana. Many
Bantu-speaking people in Botswana do not regard Basarwa as
fully human. They are regarded as if they are sub-human or
childlike. This is a point which I find deficient in
Yellen's analysis. To illustrate this position of servility
and degradation, there is a Setswana proverb that says
"Mosarwa ke yo motonanyana" - "the male Mosarwa is the
Mosarwa." This saying is interesting for some of the
I“Sative attitudes it expresses towards Basarwa. First, the
reference to gender, "yo motonanyana," depersonalizes
Basarwa. That term is used to refer to animals, and is
hardly ever used to refer to human groups. Here, Basarwa are
being equated with animals. Again, this saying expresses the
dPI‘OSatory sentiment which Basarwa have to contend with for
It Says that Basarwa women are valuable insofar as they have
"PPr‘oductive powers. The implication is that with regard to
8”(Hal activity, Basarwa women may be regarded as non-
Basfirwa. But the same consideration cannot be extended to
en‘s-we males, who are considered the lowest of the low.

So, I think we have to consider Yellen's notions of
B“saw-ma assimilation in this light. Basarwa institutions are
rel Gsated to the bottom of the ladder of social categories,
”‘d the modern state of Botswana does not take pains to

understand them and alleviate the pressures accompanying

40
moves towards incorporation. As Basarwa are numerically weak
and do not have economic and social power, they are being
treated very badly in the name of bringing them into the
mainstream of Botswana society. The process of "integration"
involves severe insensitivity towards Basarwa and does not
really create substitute institutions to take the place of
Basarwa's lost privileges. Those privileges include access
to land and a kin network that supports Basarwa in times of
stress, as well as a sense of belonging coming from a
feeling of community. The process of "assimilation" destroys
these mechanisms, but does not attempt to fill the vacuum,
for Basarwa are powerless and lack a strong voice.

Now, let us consider the "Pragmatic School's"
conceptualization of the contacts between Basarwa and non-
Basarwa in the Kgalagadi area of Botswana. In this section,
I will refer primarily to Denbow. The understanding here is
that his work is a collaboration with Wilmsen. The choice is
partly one of convenience: it is easier to refer to one
person only. Again, Denbow has undertaken a lot of
archaeological excavations in Botswana, so, that is partly
why I choose his name over Wilmsen's. To round off this
discussion, I will add Hitchcock's work in the Sandveld area
of Botswana. My intention here is to provide a modern-day
example of how Denbow's scenario of interaction could have
worked.

Next, in this discussion of contacts will follow the

work of Mabunga lehwa Gadibolae (1985) and Gaontatlhe

41
Mautle (1986). Here, my aim is to bring in a measure of
comparison between Western and indigenous scholars of
Botswana. As people who have their roots in Botswana, and
have lived there for a long time, I consider the perspective
of Mautle and Gadibolae important. The views and insights
they have are likely to be different from those of Western
and Western-trained anthropologists. Again, they have an
advantage in having stayed in the country of Botswana longer
than the other scholars. Most of the discussants of this
question of contacts visit Botswana for short periods of
time, and then return overseas. These brief journeys are
likely to have serious implications on their analysis of
contacts and interaction between Basarwa and their
neighbors.

In dealing with the Kgalagadi contact situation, Denbow
believes that a "reassessment of the antiquity and presumed
evolutionary status of contemporary foragers in the
Kalahari" is called for (Denbow 1984:175). His
reconsideration of recent Botswana prehistory challenges
conventional anthropological views regarding Botswana and
its contribution to understanding "what life may have been
like at some earlier stage in humankind's cultural
development" (Denbow 1984:176). According to him, San
hunter-gatherers have not lived unchanged lives since 10,000
years ago. His research leads him to disagree with a number
of researchers regarding their view of a "static picture of

pre-nineteenth-century foragers, a caricature set in a

42
timeless dimension with no social or cultural change until
recent times" (Denbow 1984:176).

The first person to come under attack is David
Phillipson. He argues that many parts of Southern Africa
were not climatically conducive to settlement by
pastoralists. Hence hunter-foragers found it opportune to
settle in those inhospitable areas. Again, Phillipson (1969)
argues that over the past 2,000 years, rainfall did not fall
so much as to have made the Kgalagadi a haven for
agriculturalists. Therefore, he says that "most of Botswana
was occupied largely, if not exclusively, by hunting-
gathering peoples of Late Stone Age stock until very recent
times" (Phillipson 1969:39).

Other researchers with whom Denbow disagrees include
members of the Rousseauian school. He assails George
Silberbauer (1965, 1981), Richard Lee (1979) as well as Jiro
Tanaka (1974, 1980) for taking the view that Khoisan peoples
in the Kgalagadi are nearly "pure survivals from another
age, people who for millennia have lived out of touch with
the outside world" (Denbow 1984:184). Consequently, Denbow
argues for a reconsideration of anthropological models based
on aspects of Khoisan territorial organization, family and
Broup structure, and demography. He says so because of the
evidence he uses to reconstruct change and continuity that
Went on for about 1,500 years in the recent prehistory of
the Kgalagadi (see Figure 5).

Primarily, Denbow relies on archaeological data to

43
buttress his reassessment of the prehistory of the
Kgalagadi. Additionally, the evidence comes from
geomorphology, ethnography, oral traditions, archival
sources and linguistics.

Denbow uses data based on over 400 surveyed sites and
detailed excavations at 16 selected localities (Denbow
1984:176). These excavations and reconnaissances help
buttress Denbow's interpretation of contacts and interaction
between foragers and agropastoralists. He argues that
patterns of exchange and interaction reach back to about
1,500 years ago. This is a longer time-span than the
conventional anthropological view argues. According to
Denbow, these relations were not easy to understand,
contrary to the rather simplistic view that "the standard
anthropological categories of hunter-gatherer or herder-
farmer would suggest" (Denbow 1984:179). His basic premise

is that

the association of particular types of subsistence
strategies with specific ethnic or linguistic groups is
not straightforward and the relations among the Khoi,
San, and Bantu seem to have been both of longer
duration and greater complexity than was formerly

thought to be the case [Denbow 1984:1791.

Concretely, Denbow argues that in the past, the

Kgalagadi apparently witnessed an oscillation between

44
foraging and food production by some hunters and gatherers.
Likewise, some hunters and gatherers earlier on learned to
make pottery, and herd animals, so that they are "now almost
indistinguishable, both genetically and culturally, from
their Bantu-speaking neighbors" (Denbow 1984:179).

Regarding the prehistoric past, there is evidence from
several sites in eastern Botswana to show forager-
pastoralist interaction. From the sites of Taukome and
Maiphetwane come such tell-tale characteristics of Southern
African Late Stone Age sites (Sampson 1974) as small
scrapers, crescents and other finds (Denbow 1984:180). The
presence of these artifacts on some Early Iron Age sites and
not on others leads Denbow to say that these artifacts
indicate exchange and interaction between Early Iron Age
peoples and Late Stone Age groups that happened in the
western Botswana Sandveld. Moreover, he argues that
supporting evidence comes from the scrapers' and crescents'
identity with lithics typical of Late Stone Age deposits in
other parts of Southern Africa. Additionally, Denbow asserts
that these artifacts were not necessarily manufactured and
used by Iron Age peoples.

Conversely, he states that the tools indicate
interaction between Iron Age groups and Stone Age peoples
(Denbow 1984:180). As for the time range of this
interaction, radiocarbon dates from seven of the over 300
Early Iron Age sites on the eastern fringes range from the

seventh to the fourteenth centuries A.D.. These sites

 

45

include Bisoli, Taukome, Thatswane, Maiphetwane, Toutswe,

Moritsane and Broadhurst. At Toutswe, a date from the upper

level brings the range up to the sixteenth century (Denbow
1984 : 180).

Denbow's research also shows that there was dense
agropastoral settlement on the eastern Kgalagadi, and this
sett 'l ement dates to the prehistoric past. This is indicated
by deep and extensive cattle dung deposits on several sites.
In fact, over 75% of the 320 sites located on the eastern
Sandveld of Botswana contain these deposits of vitrified
cmutt—clung. These deposits range from 30 to 100 m in diameter
and have a depth range of 25 to 150 cm (Denbow 1984:180).

Below follows some evidence from some parts of the
OBSth-n Sandveld of Botswana to indicate past interactions
‘5 We ll as suggest past environmental changes. According to
°°hbow, the first millennium sites of Toutswe and Taukome
prov ided a yield of bones. About 8096 of the meat eaten there
can-.9 from domesticates - cattle, sheep and goats (e.g.

We 1 bourne 1975). Also, the presence of springbok in the
f‘uhal assemblages suggests either past radical
"‘V 1' ronmental change or trading and hunting links with the
K9‘ 1 agadi to the west. Denbow further says that springbok
‘nd impala tend to exploit mutually exclusive areas, for
thgy subsist on the same resources. Springbok occupy the
dbiQr territories while impala are found in wetter areas.
“‘5‘" . the better-watered Toutswe area supports impala, and

hiStorical records show that was the case in the past - but

46
there are no springbok now and they are not reported in the

historical records (Denbow 1984:180, note 1).

We will now look at the evidence for prehistoric
interaction in the northern Kgalagadi. Again, I will follow
Denbow closely. He states that eight sites that are linked
to the Early Iron Age and Late Stone Age have been dated to
the f’ irst millennium A.D.. The dates range from 680 1 80
A.D. (I-12,801) at Matlapaneng to 980 j; 50 A.D. (Wits-836)
at Outpost I (Denbow 1980; Denbow and Wilmsen 1983; Denbow
1984 : 177-178, Tables 7.1 and 7.2.). On the drainage systems
01’ the Chobe and Okavango rivers are the sites of
"Ct '3 apaneng, Serondella, Chobe and Hippo Tooth (on the
3011 letli River). The sites of Society and Outpost I are
I':’<=&‘ted on the Tsodilo Hills, which lie 60 km west of the
°k§VQngo Delta. The site of Nxai Nxai is near the Namibian
border, 160 km in the Kgalagadi Sandveld. Finally, there is
sfind- lowsky's (1979) site of Kapako along the Okavango River
in horthern Namibia.

Denbow further catalogs the remains from these sites.
with the exception of Kapako, all these sites contain Early
”on Age and Late Stone Age artifacts. Also deviating from
t“- conmonalities of the other sites are Nxai Nxai and Hippo
1'0ch which do not have charcoal-tempered Early IP00 A90
c°"§anics. Other artifacts include iron and cOpper tools -
such as cutting implements, projectile points, pendants,

b°‘ds and linked-chain segments (Denbow 1984:81). Denbow

1"“‘ther reports the presence of smelting slag from

47

Matlapaneng, Society, Serondella and Chobe. In addition,

Matlapaneng had a grinding stone, while Serondella and

Outpost I had pole-impressed daga fragments characteristic

of Iron Age structures. More importantly, from the

perspective of interaction in this area, is the recovery of
cover i e shells from Society and Outpost I, as well as cane
glass beads from Society and Matlapaneng. Denbow deduces
that such trade items "provide further evidence that
trans continental trade networks stretched as far as western
Botswana by the seventh century A.D." (Denbow 1984:182).

As for the use of food resources, several patterns
merge. The further away one moves from the Okavango Delta
and enters the Sandveld, the greater is the prevalence of
amesticated fauna over wild animal remains. Denbow asserts

thgt there is a predominance of domesticated species near

the Okavango swamps (Denbow 1984:182).

Thus, Serondella, Society, Outpost I and Matlapaneng

have sheep and goats. Then, at Society and Outpost I, there
erg mostly cattle, which outnumber sheep by 8 PPOPOf‘tIOD 0f
2 to 1. Only Kapako has no cattle remains (Denbow 1984:182).
Moreover, the use of aquatic resources exhibits a
p‘t'tf-ern somewhat similar to the above. Fish bones are
“999",“; from Matlapaneng, Society (from the Okavango river
dPQ-‘i naga system), Serondella, Chobe and Hippo Tooth. Denbow
further states that hippo bones come from Kapako and Hippo
">ch (Denbow 1984:182). The further away one moves from the

0“Qvango, the faunal remains become mostly of hunted

 

48
species. However, even at Nxai Nxai, Wilmsen found evidence
of cattle - he dug out cattle teeth in association with a
hearth at 60-70 cm below surface. The radiocarbon date is
800 L 60 A.D. (Beta-397), (Denbow 1984:182).

But some controversy arose over the above date. Yellen
disputed the association, but Denbow supports Wilmsen,
because of the fact that the teeth and charcoal-tempered
pottery accompany each other in a manner that lends credence
to the considerable age suggested by the radiocarbon date
(Denbow 1984:182, note 2). Also, charcoal-tempered pottery
'58 reported from Nxai Nxai (Denbow and Wilmsen 1986;

"i IUnseen, personal comunication 1988).

The lithic assemblages also display a similar kind of
p‘ttern: away from the Okavango Delta, there is a change in
the nature of the prehistoric deposits. At Nxai Nxai and
"1990 Tooth, Late Stone Age lithics increase in proportion,
"hi 1 e the number of ceramics and metal tools decrease.
Ac"’t'hlelly, at these sites, stone tools predominate in the
rishth and ninth centuries A.D. levels. These sites are not
tr“ 1y Early Iron Age sites in Huffman's (1970, 1982) sense
Of What constitutes Iron Age sites, so Denbow sees them as

"be 'i ng probably best classified as Late Stone Age sites with
°Vi dance for contact and exchange with Iron Age Con'munities"
(°°nbow 1984:182).
Let us note these points which bear upon the nature of
°°htacts between hunter-gatherers and their non-hunting and

9“filtering neighbors. First, the evidence supplied suggests

49
that it can no longer be said that contact between foragers
and food producers in the northwest Kgalagadi is a recent
phenomenon. That is because the site of Nxai Nxai lies only
45 km from Dobe; Dobe is in the area where Yellen and other
members of his "school of thought" argue for little contact
between the two worlds we are discussing (Denbow 1984:182).
Moreover the Nxai Nxai teeth lead Denbow to suggest that in
the f‘ irst millennium A.D., some foragers had an association
With domesticated cattle. But the precise nature of that
assoc iation is unclear - were the cattle herded or hunted?
Again, the view that the Kgalagadi has been "eternally
unpty , its peoples long segregated and isolated from each
°th°P " comes under more attack (Denbow and Wilmsen
1986 = 1514). Denbow's view is substantiated by the presence
of ceramics and iron tools in the trade networks of the
"9‘ 1 agadi in the first millennium A.D.. There is also the
°V '3 deuce from the rock paintings at the Tsodilo Hills (see
Fig‘-lr‘es 6 and 7). These hills are 110 km north of Dobe.
outpost I and Society are located on the Tsodilo Hills, as
"h- the sites of Nqoma (Nioma) and Divuyu (Denbow 1984;
0°“bow and Wilmsen 1986:1511). These sites date the
°gt ‘blishment of the Iron Age to the eighth and ninth
c°nturies A.D. (Denbow 1984: Denbow and Wilmsen 1986: Table
1) -
These hills abound with rock paintings. There are over
2'000 prehistoric rock paintings at the Tsodilo Hills
(embell et al. 1980; Rudner 1965). Research indicates

50
these paintings had an extra-decorative purpose - they were
instrumental in the religious and ritual symbolism of the
prehistoric painters (e.g. Vinnicombe 1976). As the
paintings bear stylistic resemblances to thousands of others
attr ibuted to Khoisan artists in Southern Africa, it may be
said that the Tsodilo paintings may have attracted foragers
from a wide area. These foragers would then have come into
contact with Iron Age herders and new ways of life over a
thousand years ago (Denbow 1984:183).

‘l’hat there was contact and interaction is shown by some
Paint ings portraying cattle. Denbow cites the painting that
dep 'icts men raiding or tending cattle. While this painting
is uhclear insofar as it does not reveal whether the herders
"‘9 ethnically different from the painter, I presume the
p‘ihter was a Khoisan-speaking person. Also, this painting
b°§hs close stylistic resemblance to numerous typical
out, 1 ‘ines depicting wild animals and human figures found on
“‘9 Tsodilo Hills. Denbow thus views the paintings as proof
Of “direct evidence of contact and interaction between
for ‘sing and agropastoral peoples in the region and suggest
th‘t concomitant changes in economic organization may have

°°°ur~rod at the same time" (Denbow 1984:183).
Denbow hypothesizes that the introduction of the Iron
"9‘ in the Tsodilo Hills area led to the demise of rock art.
‘f Qo, he further reasons that twentieth century San ritual
"‘d religious behavior differs from that of 1,000 years ago.

This argument is part of the Pragmatists' argument of

51
openness to opportunity. Thus, Denbow produces more evidence

to show that

rather than being static, uniform relics of an ancient
way of life, San societies and cultures have undergone
transformations in the past 2,000 years that have
varied in place and time in association with local
economic and political alterations involving a variety

of peoples [Denbow and Wilmsen 1986:1514].

‘There is more evidence of forager-herder interaction in
the second millennium A.D.. This evidence comes from the
°th°l~ Tsodilo Hills site of Depression Cave. Denbow says "no
d.‘==<>l"‘ated shards were found, so it is unclear what type of
she"‘tis the undecorated shards from the 50 cm level below
surface were (Denbow 1984:184). However, Lawrence Robbins
(9°Fsonal conlnunication 1988) says that there are decorated
E‘r‘ ‘ y Iron Age shards in the levels above 50 cm. But
uncertainty regarding the identity of the undecorated shards
may still remain - whether they are Khoisan ceramics or
E‘r‘ 1 y Iron Age ones is a point that has not been resolved.

Regarding Khoisan manufacture of pottery, Rudner (1979)

8t"‘tes that the Khoikhoi and the San have a history of
mak ‘ing ceramics. The Khoi ware tended to have these features
- t"tin walls; a pointed base; reinforced lugs: pounded

qu‘rtz temper; a rare decoration of simple horizontal lines

rculled the neck: and the vessels would be red or black

52
(Rudner 1979:10). The San ceramics, on the other hand,
tended to be thick-walled bowls with a grass temper; they
would also be generally badly fired and black in color
(Rudner 1979:10).

Turning to Botswana again, Denbow cautiously suggests
that the Depression Cave ceramics may be of San manufacture
- he does not "rule out the possibility that the pottery
ment i oned was of San, not Bantu, manufacture" (Denbow
1984 : 184). Also Denbow points out associations of the Khoi
with pottery during historic times. One site is at the
KQWebe Hills, 40 km south of Lake Ngamii. Here, Denbow found
both Tswana and Khoi ceramics at a mid-eighteenth century
"'1' dden. This midden is associated with the settlement of the
”‘9‘ by the Tawana, a Tswana group, in the eighteenth
°°ntUry (Schapera 1952; Tlou 1972).

Although Khoi ceramics from over 50 Late Stone Age
sites in the Lakes Ngami and Dow area suffer from lack of
"‘t 1° '19 of the sites, Denbow surmises that the sites'

“di fferential distribution suggests that indigenous Khoi pot
m‘nufacture is of some antiquity in the region" (Denbow
‘984 : 184).

There is more data pertaining to Khoi manufacture of
put’-‘:ery. Present-day Dete Khoi informants in the area
furnish information that indicates their forebears made
potitery before the intrusion of Bantu-speaking peoples in
“‘9 eighteenth century (Denbow 1984:185). Their descriptions

“‘dicate that the pottery resembles pottery from

 

Ird

ninet
altho
their

of tI

cont:
Botlc
andq
firs
othe
from
that
area
Botl

(Den

53
archaeological sites and surface assemblages from a
nineteenth-century Dete village (Denbow 1984:185). Also,
although modern-day Dete Khoi groups do not make pottery,
their wares resemble descriptions of Namibian Khoi ceramics

of the early twentieth century (Schwartz 1923).

Linguistic data also throws light on the question of

contacts and interaction in the area concerned, i.e. the

Botletli River and the margins of the Okavango Delta. Date
and other Khoi groups are shown to have lived here in the
first millennium A.D.. Also, Westphal (1963, 1979), among
Others, argues that other Khoi groups may have split off
from the proto-Khoi nuclear area in Botswana. Denbow argues
that; Khoi pot-making may have originated in the Botletli
are; ; or, pot-making and herding may have come to the

3°"; 1 etli area from Namibia in the West at a later date
(Denbow 1984:185).

Oral traditions also shed light on this theme of
'"th‘action in the Kgalagadi area of Botswana. Dete oral
trgd itions say the Date lived for a long time in the
Bot; ‘ etli area of northern Botswana. They lived as herders,
hunters and fishers before the Bantu-speaking peoples

°nt§red the area (Denbow 1984:185). Denbow further mentions

orig cross-cutting folk tale that describes how the Date
do""‘Qsticated cattle: it is conmon to many Khoisan ethnic and
‘i"sauistic groups. Also, the explorer, David Livingstone
CK”'"IIiented on the Date herding lifestyle during his travels

I“ the area in the nineteenth century (Livingstone 1858).

54
Also supportive of Denbow's picture of flexible

interaction are the serogenetic studies I have already

referred to (Chasko et al. 1919). These studies support the

idea. of long contact and interaction among various groups in
the Kgalagadi. Khoi speakers who hunt and forage today are
shown to be genetically closer to Bantu than San speakers.
Thi s , Denbow argues, shows that people are able "to
transcend the expectations of conventional anthropological
categories" which present a "static picture of Kalahari
hunter-gatherers" (Denbow 1984:186).

From the recorded past, Denbow (1984) again produces
evidence to show that there was a frontier situation in the
Ksaj agadi area of Botswana. He argues that this frontier
3 ituation goes back at least 1,000 years into the past; it
"‘3 characterized by different political and economic
abh‘hgements intermeshing with each other. He argues against
‘ 9eegraphical separation of these different social and
°°°hom1c systems (Denbow 1984:186-190). We find that in the
"inQ‘teenth century, Khoisan people became involved in trade
ian lving products such as skins, ostrich feathers and ivory
- th is linked the economies of foragers and agropastoralists
(Gohdon 1984, Wilmsen 1989). Denbow (1984) also states that
has thoring Bantu groups competed for Khoisan labor. These
thQesses served to cause close links between the two

QQQhomies.
Denbow (1984) argues that parallels exist today. For

exQmple, Lee (1979) says Dobe lKung San are involved in the

55
modern-day Bantu groups' (Tswana and Herero) pastoral
economies because the Bantu speakers want labor for their
animals. This is especially so since the large increases in
the number of cattle, sheep and goats that followed upon the
outbreak of rinderpest in 1896 (Lee 1979). We may say that
the net of Khoisan-Bantu relationships encompasses a broad
spectrum of activities ranging from economic activities like
provision of labor and food to social activities like
visiting, dancing and settlement of disputes (Denbow
1 984 : 186).

Now, Denbow agrees with Hitchcock's (1982b) version of
versatile lifestyles among Kua San in Botswana - a shift
From foraging to cattle-posts and back to foraging,
d953°nding inter alia on social and climatic circumstances
C‘Derilzoow 1984:186). So, in Denbow's view, San people have
MVQd to cattle-posts as part of the wider process of
eye 1 ical interaction and change in the Kgalagadi. He does
not subscribe to the view that the shift to cattle-posts is
thg San's way of entering the "Iron Age" (Denbow 1984:187).
Ac§§rding to him, anthropologists who say these compound
3e'Dt'iomies are recent phenomena miss the broader picture. The
f‘ ‘ 1acy lies in the fact that twentieth-century interviews
with independent foragers occurred after the collapse of the
E""‘cpean market in game products (Denbow 1984:187, Wilmsen

‘SBSL Again, such interviews occurred after the 1896
g"‘idemic of rinderpest, which left few cattle in the region.

8°. Denbow sees the independent foragers of the 19605 and

 

 

56
1970s as belonging in his scheme of broad interactions and
contacts.

Hitchcock's work tends to fit into Denbow's dynamic
picture of interaction between the Khoisan and their Bantu-
speaking neighbors (Hitchcock 1982a, 1982b). Like Denbow,
Hitchcock argues that the Khoisan flexibly decide to take

advantage of prevailing opportunities in order to maximize

their chances of survival. Therefore, they opt for various

 

methods of interaction with the Bantu-speaking peoples;
these include working in cattle-posts and owning some goats
”“1 cattle, as well as doing some hunting and foraging and
pl‘nting crops (Hitchcock 1982a, 1982b). Referring to groups
”at are conventionally referred to as "hunter-gatherers,"
Hitchcock (1986:94) observes that "even mobile groups will
plant crops of melons, maize, beans, or sorghum, then will
cont ‘i nue to move, returning to the fields when the crops are
ripe - " In this way, Hitchcock's work exemplifies the

p
F‘smatic School's contention that relations between

f
c"‘Qsers and agropastoralists are not clearcut, and are

aubt 1e and complex.
h Vi of h otswa a lars

Before dealing with the views of Gaontatlhe Mautle
(‘986) and Mabunga lehwa Gadibolae (1985), I need to
‘ddbess several points. First, the other scholars have been

ch“ ing with the prehistoric past. Second, their use of data

57
from the ethnographic record as well as other records of the
nineteenth and twentieth centuries has been a way of
illuminating contacts and interactions in the past. Third,
Gadibolae and Mautle deal primarily with the nineteenth and
twentieth centuries, but do not specifically aim at
elucidating prehistoric contacts and interactions. Fourth,
let me say that the nineteenth- and twentieth-century
Processes occurred mostly under the aegis of colonialism and
1°WIF>er“ialism. Therefore, all those researchers who use
latter —day analogies are bound to have their conclusions
affected by the nature of colonialism and imperialism. In
this sense, perhaps the Botswana scholars' work is more
‘ff9cted by these processes, as they do not compare their

work with the archaeological record.

But I chose to use their work in spite of the above
cg"QQ‘ts because it illuminates the nature of client-patron
re'gt ionships. Again, as people from inside Botswana, 1 find
‘3th their insights and perspectives differ from those of
reSQQrchers who come from outside the country. My contention
is tZitat they may have some insights which foreigners will
probably never have because the situations of the Botswana

schQ 'lars differ from those of the foreigners, e.g. the neo-

co"Qt'iial experience.

It is in this light that I find Mautle and Gadibolae's
contribution to the debate on contacts and interactions in
”‘9 Kgalagadi interesting. They present a radically

d"H’i’erent picture of the relationships involved. By

 

58

"radical ," I mean that the other scholars have not
emphasized violence and dependency in the Bantu-Khoisan
relationships. But these two Botswana scholars argue
basical 1y that the Khoisan and one group that is regarded as
low on the ladder of social stratification, the Bakgalagadi,
were treated as serfs or slaves (Gadibolae 1985; Mautle
1986). They emphasize the unequal nature of the
relationships, and here, I need to point out that implicit
in the ir discussion is the theme of the expropriation of the
servi 1 e group's will. In addition, we may say that the
39"‘V1' 1 e group came to regard the controlling group as

super icr, and this reinforced the subordinate status of the
"colonized" group. Also, the groups of Botswana had guns,
"hi ‘9 the Basarwa and Bakgalagadi did not, and lacked
efFee‘tive regimental organization. These factors facilitated
the subjugation of Bakgalagadi and Basarwa by Botswana
groubs in Botswana.

Noteworthy is Gadibolae's and Mautle's emphasis on the

”‘9 I oitative nature of the above relationship. According to
"gut 1e (1986:19), "this relationship...was more of bolata
(3‘ every) than botlhanka (servitude)." He is referring to
the Bakgalagadi's relationship with the Bakwena, a Tswana
°thhic group. Gadibolae (1985:25) talks of the "ill-
treatment of Basarwa by Bangwato and other dominant groups."
On the other hand, the other scholars seem to view the
r°‘§tionship between the Khoisan and their Bantu-speaking

n°"9hborsias almost advantageous to all parties. They see it

 

59

as being primarily a response to structural realities.

C n ion

. when are we looking at transitional economies,

when are we seeing the results of exchange
relationships, when may we suspect class distinctions

or clientship and when are we merely dealing with a

"mm—"‘1

poorly resolved archaeological sequence? [Parkington

and Hall 1987:2]

The last two millennia in Southern Africa bore witness
t° "Sreat technological and demographic dynamism"
(P‘Pkington and Hall 1987:15). Perhaps it should not be
supp, ising then that the various discussants of the theme of
corflificts and interactions in Southern Africa disagree
spa“I‘.ly on several points.

There is disagreement relating to the degree of mixing
‘nd interaction between hunter-gatherers and
aghopastoralists. The Rousseauians take the view that
re] Qtions were characterized by mutual aloofness. But the
pr~‘9l'l'oatic "school of thought" disagrees with this
is° 1 etionist stance. Researchers like Denbow and Wilmsen
a"‘SLle for a lot of fine interrelationships between the two
9r°ups. They argue for a symbiotic relationship between
f°hfisers and agropastoralists, and also say that these

r°‘§tions have gone on for millennia. That is in contrast to

60
the Rousseauians' contention that recent changes are only a
new introduction to the hunter-gatherers' splendid isolation
from their neighbors. The Pragmatists argue that these
relations are not new, but have continued for millennia. The
Pragmatists argue that as decision-makers, people respond to
change, and are not static.

The Botswana school, while dealing with the colonial
situation, emphasizes the use of force in bringing about
UNOQual relations between Basarwa and Bantu-speaking
Peoples. The pertinent point here is that the Botswana
scholars seem to probe deeper into the values underlying the
"“39" actions between Basarwa and Bantu-speaking peoples.
This situation may reflect the Botswana scholars' inside
vi°w which the other non-Botswana scholars do not have.

But there are other issues as well in this question of
contacts and interactions in Southern Africa. One of them
Pel‘tgs to the use of "ages" in Southern African prehistory.
We heve found that in Late Stone Age sites, there were Early
lb°n Age artifacts, and vice versa. So, we have to realize
th‘t there are "problems in applying the traditional system
of tWO ‘ages' in Southern African prehistory" (Parkington
‘nd Hall 1987z2). The problems are due to the terminology's
fa‘ ‘ ure to capture "the blended traditions and interactive
comp laxity" of the "ages" (Parkington and Hall 1987:15).
Thi‘ problem of typology is also raised by Mason (1976). The

p°iht is that the typological classification used fails to

P‘sbond sensitively to "complex intergroup relationships

61
where boundaries of various kinds (linguistic, economic,
technolcegical) do not necessarily coincide" (Parkington and
Hall 1987:15).

What we have to bear in mind is that it is extremely
difficmslt to determine a wide range of past behavior from
dates and artifacts. A radiocarbon date does not tell us who
made the accompanying artifacts, where the material came
from, whether it was bartered or bought or stolen; and other
related issues are masked (Parkington and Hall 1987:14). I
am a"Suing that facts do not speak for themselves, and this
brings into question the use of artifacts to designate
ethrri<3ity. We are dealing with a subtle issue, an issue
Which defies mechanistic distinctions and definitions. What
further makes the determination of "ethnic affiliations of
"‘cthological sites" difficult is "the failure of the
irflfilltgitants to behave in ways compatible with ease of
c"‘€3ssification" (Mason 1976:357).

So we may note that relations between hunter-gatherers
and agropastoralists are complex and variable. This is a
point which the isolationist position implicitly recognizes.
‘wh£“t» I mean is that even though Lee and members of his
school may emphasize lack of co-operation and highlight
st“E3'ility through time, they still realize that hunter-
S‘therers interact with their neighbors. This is a point
V"high the other discussants specifically point out through
their emphasis on widespread contacts and interactions.

The following section provides a description of the

62
setting of the Tsodilo Hills. These hills have sites that
provide the archaeological case-study area that is my avenue
into examining the basic problem of hunter-gatherers'

interaction with non-hunter-gatherers .

 

CHAPTER 2
THE TSODILO HILLS - SETTING OF THE STUDY

The Tsodilo Hills, a National Monument, lie in the
northwestern part of Botswana in the Northwest District.
These hills are located at approximately 18 deg 45 South
latitude, 21 deg 48 East longitude. To the north lies the
Caprivi Strip, while Namibia is found to the west. The
Tsodilo Hills are approximately 31 miles from the Okavango
River, while the "annual rainfall is about 20 inches"
(Rudner 1965:11). The area in which these hills lie "has a

Mean annual rainfall of about 400-500 m" (Smithers
1971:21). And this area is characterized by a wet season
"'1 ich begins in October, and ends in April. The rainfall
tends to be localized, and sometimes occurs as
thunderstorms. The remainder of the year consists of the dry
sQason, when barely no rain falls (Andringa 1984:117). But
"‘Ost of the rain occurs during the months of November to
Mlurch (Wilson et a1. 1976:34) in this area, for the Tsodilo
Hi lls are not far from the Okavango Delta, with which Wilson
9t al's work deals.

The coming of the rains is unpredictable, and they may

be delayed for several months. Also, there is a lot of

63

64

variation regarding the amount of rainfall from year to
year. Temperatures are highest during early afternoon in the
summer, but they markedly decrease during the winter months,
when ground frost may occur at night (Andringa 1984:117).
Temperatures often exceed 40 deg C in the summer (Andringa
1984:118). Other parameters associated with the rainy season
are increased wind speed, evaporation and relative humidity
(Andringa 1984:118).

Rainfall data from the weather station at Shakawe,

about 25 miles to the north of Tsodilo, may give some

measure of the amount of rainfall in the Tsodilo Hills area.

lTie measurements covered the period from 1932 to 1980, and
titere were 51 raindays of 0.1 mm of rainfall or more in the

Period March to November; annual normal amount of rainfall

Wes 538 m (Andringa 1984:119).

The Tsodilo Hills are unusual in that they lie in an

a"‘ea of Botswana that is undulating and lacks hills. They

“i e in the Kgalagadi Sands, the Great Thirstland of

BOtswana. In fact, the Kgalagadi Sands include parts of

NQmibia in Southern Africa, and reach as far north as Zaire.

I will call the area in which the Tsodilo Hills lie the
£3llndveld, as opposed to the hardveld in the eastern part of
E3(stewana. These two broad physiographic zones have been used
by others (eg Denbow and Wilmsen 1983:405): "the Kalahari

sandveld, which covers the northern and western two-thirds

<>f the country; and (sic) the eastern hardveld, which is an

65

extension of the adjacent Zimbabwean and Transvaal

highlands." The Sandveld is characterized by the Okavango -

Chobe - Zambezi river systems in the north as well as the
Okavango - Botletli drainage, which cuts the Sandveld into
two parts (Denbow and Wilmsen 1983:405). Regarding the
hardveld further east, erosion of the Kalahari beds has
produced a more rugged topography; river systems in this
area flow eastwards into the Limpopo (Denbow and Wilmsen
1983:405). The boundary between the two broad physiographic
zones is sometimes clearly marked by an upwarped escarpment,
but sometimes the divide is imperceptible.

Turning to the Tsodilo Hills again, they consist of
f<>ur separate hills (see Figure 7). There is the Male Hill,
as well as the Female and the Child. These names are used by
‘tfie local lKung San people who live in the Tsodilo Hills
af‘ea. The fourth hill is unnamed. The Male Hill is the
*Ii sheet. "Its western face, which is solid rock, has a broad
base and rises sheer for over 300 m" (Main 1987:31). The
hfihtight of the hills descends, for the Female Hill is less
hi 9h, and the Child Hill is smaller still. The unnamed hill
‘i53 ‘the least high - 80 m. These hills are unconnected to
eQch other. The basic rock that the hills are made of is
cluartzite (Rudner 1965:51). Campbell (1980:290) says the

*‘i lls are made of "micaceous quartzite schist."

Regarding the flora of the Tsodilo, Rudner cites Banks

(1963 in lit.) to say that

66
There is a distinct difference between the heavy grass
and bush vegetation (mopani, "wild syringe" and acacia)
of the surrounding sandy plains (about 3,000 feet above
sea level) and that on the Hills and their immediate
environs with an abundance of wild fruits such as
baobab, Combretum spp., camel thorn, mispel, mangetti

(Ricinodendron rautenii), tsamma melons and strychnos

or monkey orange [Rudner 1965:51].

Regarding the fauna, the evidence comes from Smithers
(‘1971). He recorded remains of numerous mammals that were
netted or captured. These included animals like the vervet
"Konkey, wild cat, bateared fox, rusty spotted genet, banded
Inc>ngoose, common duiker, steenbok, kudu, scrub hare,
l3<>rcupine, spring hare and bush squirrel (Smithers
15371:99,124, 130-131, 166-167, 182-183, 215, 219,223, 261,
26 9, 277-280). Other mamals were only sighted, eg chacma
bElboon, cheetah, wild dog, antbear, elephant, Burchell's

zQbra, warthog, gemsbok, roan, blue wildebeest and porcupine

(Smithers 1971:101, 113-114, 140, 188, 192, 204, 206, 238,

241, 246, 268).
There are two groups of people who inhabit the Tsodilo

”l‘ills area. These are the San or Basarwa, who belong to the
2Kung group of Basarwa. They have a small village near the
"file Hill. Traditionally, they hunted and gathered wild
f'l‘uits, but they now supplement their subsistence through

|Seeping some domestic animals, such as goats and sheep.

67
Moreover, they have settled down, instead of being
residentially mobile. These San people are acknowledged to
be the original inhabitants of the Tsodilo Hills. By
"original" I mean that the "owners of the land" (bang ba
lefatsbg) are said to be the San villagers. This term is
applied to those inhabitants of an area that have been
occupying the area longer than other groups.

The lKung villagers are in a kind of client
redationship with the Hambukushu villagers, whose settlement
is further from the Male Hill than the San's. The Hambukushu
keep cattle, and in good rains they plant cereals. They also
hunt game.

The Hambukushu originated from Central Africa, and
IDIeing river people, they moved into the swamps of the
Okavango. They were fleeing from the dynastic troubles of
'ti1e Lozi kingdom in Western Zambia (Main 1987; Tlou 1972).
They were also able to work iron and keep cattle.

The Hambukushu, like many Bantu-speaking groups and
c":hers, look down on the San. So, the relationship is one of
ihequality. However, interaction does flow both ways. Some
Of the Hambukushu men marry Basarwa (San) women, or father
‘3iiildren out of wedlock. Also, visiting between the San and
the Hambukushu takes place frequently.

As noted previously there are rock paintings on the
Tsodilo Hills. These paintings are under study by Alec
Campbell. The number of the paintings exceeds 2,750
(Campbell et al. 1980). Cooke (1969) says they are found on

68
many open shelters as well as on individual rocks. He
further states that these paintings have more differences
than similarities "with paintings from elsewhere in Southern
and South-Central Africa" (Cooke 1969:27). Moreover, Cooke
states that the "naturalistic paintings of Tsodilo are more
in the form of caricatures than true representations" (Cooke
1969:27). These paintings are mostly taken to have been
executed by San people, even though local San today do not
paint rocks. Furthermore, surviving San people in the
'Tsodilo Hills area do not have oral traditions which
attribute the paintings to them - except for a few found at
tiie White Shelter on the Male Hill (Campbell and Coulson
1988). But a number of authorities suggest the paintings are
‘tfie work of San people, and Cooke (1969) says categorically
'tliat they were made by San people.
As for the age of the paintings, uncertainty abounds.
.rfiat is because of the extreme difficulties entailed in
dllting rock paintings (Main 1987:35). Stylistic changes and
8'--aperimposed paintings, as well as using wear and tear
(degree of fading) and the paintings' subject matter all
r‘filve drawbacks (Main 1987:35). Archaeological associations
‘irid radiocarbon dating are the more reliable methods, but
tiFiey still do not tell us the exact age of Tsodilo Hills

"teck paintings (Main 1987:35). However, he states that

The small handful depicting what are clearly cattle are

unlikely to be older than the oldest traces of those

69
animals found in the hills. This suggests that they

were not painted before A.D. 700-900 [Main 1987:35].

The next section gives a review of the general
archaeological background of the area or region in which the

Tsodilo Hills lie.

CHAPTER 3

A SWARY OF PREVIOUS ARCHAEOLOGIOAL EXCAVATIOHS IN THE
TSODILO HILLS AREA

In order to understand the importance of archaeological

work in the Tsodilo Hills area of Botswana, it is necessary

to remember that the Kgalagadi region was conventionally
seen as a peripheral area that owed its marginality to its
inhospitable climate. Therefore, the excavations at Tsodilo
are significant in the sense that they enlarge our knowledge

01’ past cultural processes in this area that was hitherto

cOnsidered to be a backwater. As stated previously, the

fif‘ sument stated that the Kgalagadi was a retreat area where
San conmunities fled when they came in contact with food-

pr‘cducing agropastoralists who also brought new technologies

8Llch as metallurgy. It follows that the incursion of Bantu-

8Deaking peoples into the Kgalagadi overwhelmed the San, and

the argument also runs that the San led an almost-static

9x istence that continued for millennia.

The excavations in the Tsodilo Hills are important in

yet another respect. The dates from these sites indicate

that there has been a long period of interaction between the

sen-speaking peoples and their Bantu-speaking neighbors. In

70

71
this sense, our knowledge of the time depth of interaction
between the two communities is also enhanced. Previously, it
was thought that the process of interaction is a more recent
phenomenon. But the trajectory involved is much longer in
terms of time.

Moreover, the significance of the work at Tsodilo lies
in the fact that there have been few excavations in the
Tsodilo region. Therefore, work at Tsodilo helps to fill the
lacunae in the archaeology of the Kgalagadi. Thus, work in
the Tsodilo area is instructive because the Tsodilo sites
yield data that enable us to make comparisons with research
from better-researched regions. Apart from the comparative
value of the Tsodilo sites, we also have to note that they
aid knowledge insofar as they are building blocks on which
we can base our inquiry into past cultural processes in the
Kgalagadi. We can ask a different order of questions that we
could not ask without the knowledge based on excavations at
Tsodilo. Admittedly, more archaeological work needs to be
carried out in the Kgalagadi, for the work that has been
done is insufficient to flesh out the bare skeleton that has
been exposed.

The Tsodilo sites discussed in this work lie on the
Female Hill at Tsodilo. These sites were excavated by
several American archaeologists and their Botswana
assistants.On the sites of Outpost 1 and Society, James
Denbow and Edwin Wilmsen worked with Alex Matseka, Alec

Campbell, Calvin Sebole, Salalenna Phaladi and Samochao

72
Mokate among others in 1979-1980. Later, Society and Outpost
1 were incorporated into the more extensive work of Nloma
that took place in 1985. Another site on the Female Hill is
Divuyu, which was also excavated in 1985. Also, there is the
site of Depression Rockshelter, which Lawrence Robbins
excavated with Alec Campbell and their assistants
extensively in 1987. The Depression Rockshelter is also on
the Female Hill.

The sites of Society and Outpost 1 have yielded
evidence for agriculture during the period from 700 to 1000
A.D. (Denbow 1987). Especially at Society, direct evidence
for agriculture is attested by the presence of shallow
grindstones, carbonized seeds of millet, sorghum, cowpeas
and cucurbits.

Still at Outpost 1 and Society, faunal remains are
mostly those of cattle and ovicaprids (Denbow and Wilmsen
1986). Society largely has cattle remains, while sheep and
goats preponderate at Outpost 1. There are also fish bones
at Society.

Lithic remains from the two sites belong to the Late
Stone Age. There are small scrapers, crescents, backed
bladelets and steeply retouched segments. As for ceramics,
Society and Outpost 1 are characterized mainly by charcoal-
tempered Early Iron Age ceramics.

There are also indications of coastal trade. This is
represented by cowrie shells at both Society and Outpost 1.

In addition, caneglass beads come from Society.

 

73

Other artifacts include copper and iron tools. The
range includes pendants, segments, beads, projectile points
and cutting implements.

Additionally, smelting slag came from Society, while
Outpost 1 had pole-impressed daga fragments, indicating
house structures.

At the site of Nloma, we also have a time range of AD
700 to AD 1000, except that NIoma flowered in the ninth and
tenth centuries. Its economy was centered on cattle, but
sheep and goats were also present. Also, other components of
the economy included crops like sorghum, millet, and perhaps
melons. But the importance of foraging in this
agropastoralist economy is shown by the present of Mongongo
nuts and Grewia berries and wild faunal remains.

Other artifacts from Nloma include daga houses, ivory,
iron and copper ornaments. Also, there are iron tools, cane-
glass beads and marine mollusk shells (including cowrie
shell). Also, freshwater mussels and fish came from Okavango
river (Wilmsen 1989:71-72).

The other main site of Divuyu dates to the sixth and
seventh centuries AD. It was a fully developed Early Iron
Age site. It had many ceramics, iron and copper tools and
ornaments. Ivory was also present. Again, iron slag shows
that iron smelting took place at Divuyu (Wilmsen 1989:70).

Still at Divuyu, sheep and goats were the mainstays of
the economy. Cattle were rare. In terms of trade over large

areas, two Atlantic coast marine shells and two iron

74
pendants show that there was trade with the interior and the
coast (Wilmsen 1989:70).

At Depression Rockshelter, charcoal-tempered pottery
was found (Denbow 1984:184). The pottery was mainly
undecorated (Campbell and Robbins 1989; Robbins 1990:334),
but some of it had incision and comb-stamping (Robbins
1990:335). Other finds included a variety of lithics: small
scrapers, crescents, casually retouched pieces, sharp-edged
flakes, backed microliths, points, blades, bladelets, and
backed pieces. The fauna was not well-preserved, and is
entirely of wild animals. Again, Dr. Robbins found that
mongongo nuts (Ricinodendron rautaneii, Schinz) were used
during the Holocene. These nuts are a very nutritious staple
of San diet as well as other peoples. Again, the dates for
Depression Rockshelter extended to 19,000 years ago (Robbins
1989:1-3).

These sites shed light on a number of the theoretical
issues discussed above. The presence of lithics and pottery
of the kind made by Bantu-speaking agropastoralists shows
that interaction between the San-speaking peoples and the
Bantu-speaking agropastoralists is of greater antiquity than
has often been thought. Also, we can see that foraging and
agropastoralism complemented each other. Also, the Tsodilo
area was involved in trade over an extensive geographical
area.

The following section deals with the site of Nloma,

which is the source of the lithic material I looked at in

75
this study. I will outline the different areas we excavated,
discuss the nature of the levels, as well as look at the
stone artifact classification. As I have already stated, the
site of Nloma dates to between A0700 and AD1000; it also
shows remains of agro-pastoralism, iron technology and

involvement in long distance trade.

CHAPTER 4
THE SITE of HIOHA

The site of Nloma is on the Female Hill at Tsodilo.
This site covers a large area measuring 100 m x 200 m. As
mentioned previously, this site was excavated under the
direction of James Denbow and Edwin Wilmsen. It was part of
their project investigating the coming of pastoralism into
Botswana. The site was excavated in May-June 1985, with
three previous excavations at the test sites of Society and
Outposts I and II. The latter excavations were undertaken in
1979, 1980 and 1982.

The excavations divided the site into 10 "areas" and
these areas are used in relation to my lithic analysis
(Figure 7). Each area, apart from Area 10, encompasses a set
of squares or units that lay close together in a
geographical area that was distinct from other areas. Only
Area 10 did not meet this criterion, for Area 10 includes
all the test squares that did not belong to Areas 1 through
9.

Before presenting a list of the 10 areas of N!oma, I
will mention some details pertaining to the "levels" we used

to excavate the site. The levels are arbitrary, and in cases

76

77

where there were actual, distinct occupation layers, the
levels cut across those layers. (I have no record of any
cultural stratigraphy). In addition to mentioning the levels
we used to excavate the site of Nloma, I will also mention
the distances between the test squares.

A "level" means that we excavated the squares at NIoma
in intervals of 10 cm. Thus, Level 1 will range from 0 to 10
cm and Level 2 will include the next 10 cm, and so on down
to the deepest level. The exceptions include those squares
where we reached bedrock before the full 10 cm level could
be completed. Another instance would be with burials, which
did not fully extend into the depth of the square. One more
exception includes floor scrapings, wherein in order to
clear up the square, the bottom of the square was scraped
and the resulting artifacts made up their own level.

Regarding the spacing of the test squares (or units),
we excavated by dividing the site into quadrants measuring
10 m by 10 m (see Figure 8).

Below follows a list of the 10 different areas and
their geographical range.

Area 1: Squares between 00.20N and 10W and 30N to 06E

30N.

Area 2: Squares 0.0, 02W 00, 03W 00, 02W 018, 03W 018.
Area 3: Squares contiguous to 40W 255 and 40W 208.
Area 4: Squares between 35W 608 and 50W 60S, 40W 698.

5: Squares with numbers in 70s and 80s West and

APO.

20s and 30s South.

78

Area 6: Squares 80W 608, 81W 608, 80W 618 and 81W 618.

Area 7: Squares 60W 60N, 61W 60N, 60W 61N, 61W 61N.

Area 8: Squares 170sW, 180sW, 190sW.

Area 9: Squares 223W 127N to 239W 132N.

Area 10: All remaining squares.

A point regarding Areas 1 and 2 needs discussion. When
I began analyzing the artifacts from NSoma, I began with the
lithics from Areas 1 and 2. Because I was beginning to
undertake this analysis, I omitted the artifacts from these
two areas from further examination. Therefore, I looked more
closely at the artifacts from Areas 3 through 10. These are
areas I looked at more extensively when I had more
experience, which I gained from the initial work at Areas 1
and 2.

Another point regarding the areas I examined concerns
the sampling of some areas. Time became a constraint, and I
decided to sample some areas, instead of looking at all the
squares. The affected areas are Areas 4, 6, 8 and 10.

Here are the details of the areas I looked at more
extensively.

Area 3 had 15 units, with Level 8 (70-80 cm) as the
deepest level that yielded lithics. Area 4 had four units.
Again, Level 8 was the deepest level. Area 5 had 13 units,
all but Unit 79W 348 with one level (0-10 cm). The
exceptional unit had a deepest level of 30-40 cm. Area 6 had
three units. Its deepest level to yield lithic artifacts was

100-110 cm. Area 7 had five units, with 130-140 cm as the

79
deepest level. Area 8 had seven units, with 90-100 cm as the
deepest level. Area 9 had 11 units, with 80-90 cm as the
deepest level. Area 10 had 19 units (scattered across the
site), with 120-130 cm as the deepest level.

The total number of units I examined for the final
analysis is 67, encompassing 263 levels. Regarding the above
statements about the depths of the various areas, I have to
point out that several factors account for the variable
depths of those areas. One reason is that we reached
bedrock, and could not dig any deeper. Another factor is
that when we excavated, we reached a depth where we found no
more cultural remains. This level differed from one area to
another, hence the various areas' variable depths. In
addition, some areas did not yield lithic artifacts beyond
certain levels, even though those levels might produce bones
and other artifacts.

Turning to the computer analysis, I collapsed the 91
artifact categories into 20 categories in order to make
analysis easier and more statistically meaningful. Without
breaking the artifact categories into broader groups based
on the presence of retouch, lack of retouch, presence of
edge damage and a miscellaneous category of two artifacts
that did not fit neatly into the top three groups, the
statistical analysis would have been void. That is, by being
sensitive to the differences in the designation of the
lithic artifacts from Nloma, I produced a long list of

artifacts, but the various designations often had few (such

80
as less than five) items belonging to them. Therefore,
running statistical tests would be meaningless for many
individual artifact categories, as the number of
observations would be too small. So, I decided to use the
twenty categories to group individual artifacts and generate
groups more amenable to statistical analysis using the
computer.

The total number of cases I entered into the computer
was 5,757. These artifacts were made of quartz and chert.
These two raw materials presented different challenges when
artifacts had to be classified according to the type of
their constituent raw material. In the case of quartz, the
problem was easier because the quartz had a distinctive
shininess that was uniformly present in artifacts that were
made of this material. Conversely, the chert artifacts
showed a lot of variation. This variation pertained mainly
to color. The range in color variation included dark brown,
light brown, gray, dark red and yellow. To simplify
classification and analysis, I grouped all these shades of
chert under the single name of chert.

I will list the definitions of the NIoma lithic
categories in the next section, where I will also deal with
the methodology I used to analyze the lithic artifacts from

Nloma.

CHAPTER 5

HETHODOLOGIES AND DEFINITIONS OF THE CATEGORIES USED IN

OLASSIFYING THE LITHIOS FROM N!OHA

Various classifications for microlithic stone
artifacts are used in Southern Africa, such as Deacon
(1984). My classification uses terms such as flake,
scraper, and crescent, among others; but it also brings in
terms that are more specific to the material in this
collection, eg informal knife.

The main purpose of my classification is descriptive.
However, I do use some terms that may be functional. For
example, a use wear study recently carried out by R. Donahue
and L. Robbins on similar artifacts from the nearby
Depression site indicates the drills were actually used for
drilling a hard substance such as ostrich eggshell and the
scrapers were in fact used to scrape hide and hard materials
(L. Robbins personal communication 1989).

Below are the definitions I used to classify the
lithics from Nloma (see Figures 9-14 for some
illustrations).

Angular waste - A stone artifact that is not a tool and also
lacks a striking platform and a bulb of percussion.

81

82

Angular waste with metallic material - Angular waste that
contains a luminous substance and is an iron oxide
hematite.

Angular waste with metallic material? - Angular waste that
contains a luminous substance that may be an iron oxide
hematite.

Backed - A lithic with a series of small flake scars on the
edge or edges. These scars are at an approximate right
angle with the main surface of the tool. In many cases,
these marks have been purposely made on the part of the
artifact opposite the sharp edge. Also, the marks tend
to be well-made.

Atypically backed - An artifact that has backing that is
unusual and differs from the backing on most of the
artifacts in the sense that the backing is very steep
and covers various parts of the artifact.

Backed and edge-damaged - An artifact that has backing and
edge damage.

Backed and retouched - An artifact that has backing and
retouch.

Backed, round-shaped tool? - An artifact that has backing
and is circular in form.

Backed, start of a crescent? - A backed artifact that may be
the beginning of a crescent.

Double-backed - An artifact that has backing on two sides.
Heavily-backed - An artifact that has extraordinary backing.

Partly-backed - An artifact that does not have full backing
along the entire edge.

Traces of backing - Slightly backed.

Blade - a blade is an artifact with long sides that have a
short width and may have functioned as part of a
cutting tool.

Core - An artifact that functioned to provide the basis for
other pieces that were made into tools or became flakes
and angular waste.

Blade-core - A core that served to produce blades.

Blade-core? - A core that may have produced blades.

Broken-core - A core that broke into several pieces.

83

Flake-core - A core that produced flakes.

Piece of core? - An artifact that may be a fragment of a
core, but is less complete than a broken core.

Split-core - A core that shows evidence of having been
broken apart along a horizontal plane that is above the
base and sides.

Retouched split core - A split core that has retouch.
Split core with edge damage - An edge-damaged core.

Split core? (Broken core?) - A core that may have broken
apart along a horizontal axis.

Start of core - Beginning of a core.

Worn-out core? - A core that appears to have been completely
reduced.

Crescent - A small, almost semi-circular tool that is
usually backed. These artifacts are also known as
lunates or segments.

Broken crescent - A crescent that was forcibly separated
into different parts.

Crescent? - An artifact that appears to be a crescent.

Atypical crescent - An unusual crescent. It has a shape that
differs from other crescents'. Its retouched, curved
side is much wider.

Crude crescent - An unrefined crescent.

Large crescent/piercing tool? - A long crescent with a big
width, and it could also have been used to make holes
in artifacts such as hides.

Retouched but not backed crescent - A crescent that has no
backing but is retouched.

Drill - A lithic with sides that converge on a small pointed
end.

Drill? - An artifact that appears to be a drill.

Backed drill - An artifact that is backed and appears to be
a drill.

Retouched drill/awl? - A retouched artifact that could be a
drill. On the other hand, it could also be an awl, for

 

84

it is slender and tapered, with a sharp end. Possibly,
either as a drill or, as an awl, this artifact was used
to pierce holes, such as in leather.

Edge-damaged - A lithic that has indentations or sharp cuts
on the edge or edges. These cuts or indentations do
not clearly show signs of having been purposely made.

Flake - A stone artifact that has a striking platform and a
bulb of percussion. The piece resulted from the making
of tools or the reduction of a core.

Core refresher flake/core trimming flake - A flake that
resulted from making a core more amenable to producing
more artifacts.

Flake with possible edge-damage - A flake that may be edge-
damaged.

Flake with shiny metallic material - A flake that contains a
bright, luminous and reflective material that is an
iron oxide probably hematite.

Flake with some clear, transparent quartz - A chert flake
that contains a large amount of chert material, but has
a slight amount of quartz material in one place.

Freshly broken flake - A flake with a newly-occurred break.

Split flake - A flake that has been separated along a
horizontal axis.

Notched piece - An artifact that has a series of V-shaped
marks or indentations on the edges. This notching is
due to retouch in some cases.

Point - A retouched artifact that has an edge that appears
to have been used as a projectile for hunting or
perhaps to pierce other objects.

Point? - An artifact that may be a point.

Backed point? - A point that may have backing.

Unretouched point? - A (point?) that does not have retouch.

Retouched - A lithic that has been improved or refined by
removing some tiny pieces off the edge of the stone

artifact.

Broken retouched piece - A retouched piece that is broken.

Possible retouch? Retouched - An artifact that may have
retouch.

85

Possibly retouched or heavily edge-damaged - An artifact
that may have retouch or has heavy edge damage.

Retouched (with double backing and broken tip) - An artifact

that has retouch and is backed on two sides; it also
has a broken end.

Slightly retouched, minimally retouched; partly retouched;
minimal/minimum retouch - An artifact that has a little
amount of retouch.

Steeply retouched; steep retouch? - An artifact that either

has a high angle of retouch or may have a high angle of
retouch.

Scraper - A stone artifact with a relatively continuous
retouched edge that may have been used to rub or add

pressure to another surface such as an animal skin or
wood.

Scraper with minimal retouch - A scraper that is slightly
retouched.

Utilized? - An artifact that may have been used for a
purpose that is not certain.

Worked - Same as retouched; I use these terms
interchangeably.

Informal knife/unretouched knife/Informal, unretouched knife

- A stone artifact with a fine, sharp cutting edge that
lacks retouch.

Large bifacial tool - A long and wide artifact with two
opposite surfaces that converge. Both have retouch and
may have been used.

In the next section, I deal with the results of the
artifacts that I investigated. I classified each artifact
according to these traits:

(a) raw material (quartz or chert or undetermined ie
not sure whether it is chert or quartz)

(b) area (the specific site area)

(c) square

(d) level

 

86

(e) category (type of lithic artifact)

(f) length, breadth and depth in millimeters.

Each artifact also had a specific number, and I entered
all these data on the computer. The people who helped me
regarding computer programs and statistics are Dr. Stephen
Godek of James Madison College and Dr. William Lovis of the
Anthropology Department at Michigan State University. The
results of the lithic classification are presented in the
accompanying tables (Tables 1-11). In the following section,
I comment on the tables of artifacts from N!oma. The tables
are for Areas 3 to 10, and exclude Areas 1 and 2. As
mentioned previously, I excluded these first two areas from
the analysis because I was beginning to learn how to analyze
the lithics, so the results of my analysis were suspect.
When I reached Areas 3 to 10, I then felt more certain of
what I was looking for and how to look for it. The comments
are descriptive. They also include lists of the codes I used
to classify the artifacts from the site of Nloma. Again, I
include the final 20 categories that incorporate all the
lithic artifacts from Nsoma. These 20 categories make it

less cumbersome to look at the data from the whole site.

CHAPTER 6

COHHENTS ON THE TABLES OF ARTIFACTS BY LEVELS FOR THE EIGHT
DIFFERENT AREAS OF THE SITE OF N!OHA

For Area 3, a few patterns are evident (see Table 1).
First, Level 1 does not have many tools, except for those
that are backed, retouched, and edge damaged. As for Level
2, it is also interesting because all the informal knives
lie in this level. Also, there is one scraper for the whole
area, and this is the level in which it was found. This
observation is similar to that for the backed artifacts, for
both of them lie in the first level of excavation. Another
observation concerns the blades, which are distributed
almost evenly over Levels 3 through 8. As for the cores, the
concentration is on Levels 2 to 5. But a comparison of the
blades and cores shows that there may be stratigraphic
significance as five of the seven blades in Area 3, and
eight of the 10 cores in the same area, lie in Levels 3
through 5. Then, the edge damaged pieces concentrate on
Levels 1 to 3 (12 out of 22), and Levels 4 to 6 (nine out of
22). The mixed pieces do not show a particular observation
of being concentrated in a certain area. The only drill is

in Level 5. To round out the observation of distribution of

87

88
artifacts in Area 3, the flakes lie mainly in Levels 2 up to
5, while the angular waste lies mainly in Levels 3 to 6.

There is also a pattern of the flake and angular waste

levels as activity areas. By "activity areas,‘ I mean that
the flake and angular waste levels are also the levels
wherein fall the tools--such as the blades, cores, informal
knives, scrapers, the combination lithics, drills and edge
damaged pieces. But it is difficult to say what kinds of
activities were carried out, as I cannot tell what the
various tools, eg the blades, were used for. Also, there is
only one scraper in Area 3, and that number is too small to
speculate that woodworking or hide-scraping went on.

As for Area 4, an interesting feature emerges: Level 5
has the only core of the area (see Table 2). It also has
three of the area's four retouched pieces. Again, it has one
of the six backed pieces. Turning to other levels, we find
that Level 7 contains two backed tools, a retouched piece
and one combination lithic. Level 8 rounds out the
observation of the deeper levels' having several tools:
three out of the six backed tools lie in this level. Again,
it is noteworthy that in the first four levels, there are no
cores, backed, retouched and combination lithics. Also, this
«observation would hold for all the tools in Area 4, except
tfliat there is one edge damaged lithic in Level 3. This
OtDservation contrasts with the presence of angular waste in

3] l eight levels, and there are flakes in all but the first

1eD‘vel.

89

I sampled four squares out of the total of twelve

squares in this area. I looked at all the bags in those

selected squares. I wanted squares from each of these

points: 40Ws, 41Ws, 43 Ws, and 44Ws--there were no 42Ws.

Also, I selected squares that had at least five levels or

more, and omitted the ones that had fewer levels.
For Area 5, the striking pattern relates to the first
level (see Table 3). Most of the artifacts lie in this

level, and all the artifact categories have at least one

lithic represented here. In fact, all the blades (five),

combination lithics (four), backed pieces (one) and informal
knives (one) are in this level. For the remaining tool
types, six of the seven cores, eight of the 11 retouched
lithics and 21 of the 25 edge damaged lithics lie in the
first level. Also, while the flakes and angular waste are
represented in every level, most of them lie here: flakes--

179 out of a total of 224; angular waste--32 out of a total

of 45. Clearly, the first level witnessed tool making. Also,

since Area 5 has only four levels, I argue that occupation

occurred in the topmost level, and did not extend much into
the remaining 30 cm of the area. This area can be seen as a
homogeneous entity that was recently occupied, hence the
«concentration of lithic artifacts in the topmost level.
Since the occupants of Area 5 lived there recently, their
.r‘tifacts are found mostly in level one, and not in the
0earlier levels. This interpretation may need backing from

radiocarbon dates for Area 5, as well as data pertaining to

90
the other types of artifacts found in Area 5--such as metal
objects and beads. This issue will receive further attention
in the following chapter as part of the conclusion of the
investigation of the lithics from the site of N!oma.

For Area 6, there is a high density of artifacts in the
last four levels: eight, nine, 10 and 11 (see Table 4). For
the cores, 16 of 19 specimens lie in these levels. The
pattern holds true for the backed, retouched and edge
damaged artifacts: the last four levels contain most of the
lithic categories. The numbers are as follows: 10 out of 11
for the backed lithics; 15 out of 21 for the retouched
lithics and 35 out of 36 for the edge damaged artifacts.
This pattern also applies to the angular waste and the
flakes: 802 out of 887, and 66 out of a total of 70--all in
the last four levels.

Again, the interpretation that can be made about the
concentration of artifacts in these deeper levels of Area 6
is that these levels were activity areas. The presence of
flakes and angular waste would lend credence to this
interpretation. Also, this view of the deeper levels having
been activity areas derives support from the presence of
other tool types in these levels, albeit in a less striking
way. For the informal knives, four of the six are in levels
eight and 11. For the combination pieces, both lie in levels
10 and 11. For the blades, all nine are found in levels
eight and nine. For the drills, both are found in level

eight.

 

91

Toolmaking did not occur much in the upper levels of
Area 6, for these levels contain few artifacts. Level 2 has
only 10 flakes, and the rest of the artifacts are not
present. For Level 1, this observation is continued, for'
there would be no artifacts at all were it not for the eight
flakes, two cores, one informal knife and one crude
crescent. Level 4 also has some artifacts but they are not
present in significant numbers: one core; one informal
knife; one backed piece; one edge damaged lithic; one
possibly retouched lithic and four retouched pieces.

I also sampled squares from this area by selecting
three of the four squares, and looking at all the bags in
those squares.

For Area 7, the patterns differ from those of the
preceding areas (see Table 5). Here I will concentrate on
the levels, and not on the artifact categories. I will use
the levels as the organizing features for describing the
patterns I see and other considerations will be subordinate
to this process of ordering. First, several levels stand out
in terms of their having many different kinds of artifacts,
while others have fewer kinds of artifacts. The levels with
many artifacts are 5, 8, 12 and 13. The number of tool types
varies from five to seven. The level with the highest number
of artifacts is level 12, which has seven tool types. Level
5 has the fewest number of artifacts, five.

Furthermore, other observationss are evident in Area 7.

Levels 1, 6 and 11 hardly have any tools at all. Level 1 has

 

92
three tools (one core, one combination piece and one edge
damaged lithic); Level 6 has two tools (one retouched and
one edge damaged) and Level 11 has two tools (one blade and
one core). Another observation that emerges is that the two
blades lie in Levels 10 and 11. Moreover, the edge damaged
lithics are mostly found in the first three levels (8 out of
40). For the cores, half lie in levels 10 to 14 (nine out of
a total of 18). For the retouched pieces, the concentration
is in Levels 2 to 10, with 18 out of the total 24 lying in
these levels.

I think the main reason Area 7 has a diffuse pattern of
tool concentration is that this area consists of only four
squares. Because Area 7 is limited in area (number of
squares), especially in comparison with other areas that
contain more than four squares, this fact helps to explain
the small numbers of tools in the area. I would argue that
this area was not a major activity area, and the tools and
flakes, as well as the angular waste, reflect a wider
settlement in N!oma. However, the presence of these
artifacts does not represent the carrying-out of specific
tool-making activities and occupations in the area. Also, a
look at a map of the site (see Figure 7) shows that Area 7
is not near the other areas. It is isolated, whereas the
other areas tend to be nearer each other. It may be possible
that this area is on the edge of the occupation.

As for Area 8, I would argue that activities of a tool-

making nature occurred here (see Table 6). Suggesting that

93

Area 8 witnessed occupation by people who made stone tools
is the fact that Level 5, in addition to the deeper levels
of Area 8, contains a large number of tools and debitage
(flakes and angular waste). Levels 7, 8, 9 and 10 contain
112 of the 183 tools present in Area 8. As for Level 5, it
contains 31 tools out of the remaining six levels' 71 tools.

Among the tool types in Area 8, the edge damaged
lithics are the most numerous. There are 75 of them. They
are found mainly in Levels 8, 9 and 10: 39 of the 75 were
found in these levels. Also, Levels 4 and 5 have a
significant number of tools, as 19 were recovered in these
two levels.

The retouched pieces are also numerous. Their total is
52. This total is higher than the total for the angular
waste: 45. It is also higher than the total for the backed
artifacts and cores, which number 31 and 25 respectively.
The explanation for this area's unusual number of retouched
pieces may lie in the fact that I sampled material from
squares in this area, as I had a large number of artifacts
to analyze. Moreover, time was becoming a constraining
factor. It is likely that my sampling strategy leaned
towards squares that had an abundance of retouched tools. It
is uncommon to find retouched tools outnumbering the backed
land core tools. Additionally, it may be that in this area,
'the occupants were able to make retouched tools more
intensively, for they form just over 25% of the total number

«of tools present here (26.94%). (This number is the second

94
highest, as the edge damaged lithics form 38.86% of the
tools in Area 8). Area 8 may have so many retouched pieces
because there was a big need for those tools. Additionally,
it is possible that there were other presently unknown
factors that governed the presence of so many retouched
tools in Area 8.

Again, the retouched pieces are concentrated in the
last levels, with Levels 7, 8, 9 and 10 containing 32 of the
52 retouched artifacts. The same pattern applies to the
backed pieces as well. They are also concentrated in the
last levels. In this case, Levels 8, 9 and 10 contain 15 of
the 31 backed pieces present in Area 8.

For the cores, we find a similar pattern. Levels 7, 8,
9 and 10 contain 14 of the area's 25 cores. We also find
that Level 5 has many cores, eight, which is the highest
number in any level. I speculate that Level 5 may coincide
with long, heavy occupation of the area, which may mean that
we can then find these tool types being highly frequent, as
was the case with the retouched pieces. Also, as with the
retouched pieces, it may be that when I sampled the squares
from this area, I chose bags that had a preponderance of
cores; these bags then may have had an effect on the
distribution of cores in the area.

For the remaining tool types, the pattern still holds;
the last levels still have a large number of the tools
present. However, the actual numbers involved are relatively

small, when compared to the numbers for the retouched

95
lithics and cores. There are three informal knives; one
«drill and six combination lithics. Continuing the trend of
'tools being concentrated in the last levels, the tool that
is probably a drill lies in Level 9. With the informal
lsnives, one of the three lies in Level 7. For the
«sombination tools, three of the six tools lie in Level 9.

This pattern of tools being concentrated in the deeper
levels of Area 8 is also shown by the concentration of the
‘Flakes and angular waste in these levels. Also Level 5 has a

large concentration of flakes and angular waste. For the
'flakes, 563 out of the total of 1120 found in Area 8 were
Icecovered from Levels 7, 8, 9 and 10. With Level 5'8 254
flakes the total reaches 817. As for the angular waste, 20
‘of the 45 pieces lie in the last four levels. Level 5 does
have the largest number of angular waste, 16: this makes the
total for Levels 5, 7, 8, 9 and 10 reach 36. Therefore, I
argue that the later levels, as well as Level 5, were
<occupation levels, and the upper levels show an absence of
prolonged occupation; that is because they do not have a
:sizable amount of artifacts, especially tools.

Area 9 shows several patterns pertaining to the
(distribution of artifacts over the area's nine levels see
lrable 7). First, a significant point is that Area 9 has 16
Clifferent types of tools, in addition to the flakes and
ahgular waste. Also, the area has a large amount of tools:
1 68 of the 477 (35.22%) artifacts in Area 9 are tools.

Generally, this area has many tools in the upper levels, and

96
I.evels 8 and 9 are underrepresented. Area 9 has one tool in
l_evel 8, while Level 9 has six tools. The rest of the tools
in Area 9 lie in Levels 1 to 7. Actually, the first four
levels from the surface contain 106 of the area's 168 tools.
Idowever, it is noteworthy that Level 3 has the smallest
Inumber of tools, apart from Levels 8 and 9.

A breakdown of the tool types by levels shows that most
<3f the tool types in Area 9 are edge damaged (38, or
122.62%), blades (31, or 18.45%) and cores (29, or 17.26%).
“These three types of tools constitute 98 of the 168 tools
present, or 58.33%.

All the blades lie in Levels 1 to 6. Levels 7, 8 and 9
have no blades at all. Level 2 (eight blades) and Level 4
(eleven blades) have 19 of the 31 blades present in Area 9.
As for the cores, Levels 8 and 9 have none, as does Level 6.
The cores are distributed almost evenly in the remaining
levels. With the edge damaged lithics, 20 of these 38
artifacts lie in Levels 1 and 2 (52.63%), while Levels 4
through 7 also have many artifacts that make up almost the
total number of edge damaged lithics (39.47%) that are
Ioutside Levels 1 and 2. Levels 3 and 8 have no edge damaged
artifacts, and Level 9 has only three pieces.

The pattern of heavy tool concentration in the upper
levels of Area 9 also continues with other tool types. For

Clxample, with the retouched lithics, only three pieces lie
'ili Levels 6, 7 and 9 (Level 8 has none). All the remaining

levels, except for Level 3 (which has none), contain 11 of

97
‘the total number of retouched lithics (14, or 78.57%).
Similarly, the backed pieces are concentrated mainly in the
'top levels (1 through 5). Only three out of a total of 11
lie in Levels 6 and 7, with no backed artifacts in Levels 8
and 9.

Likewise, all the scrapers were found in the first six
levels. There are no scrapers in Levels 7, 8 and 9. With the
scrapers, there is no heavy concentration in a particular
level, as the number in a given level is either one or two.
As for the crude crescents, the pattern of the upper levels
predominating by having many artifacts becomes readily
apparent--Levels 1 and 2 have five of the six crude
crescents present. This pattern is also borne out by the
unretouched points--there are only two, and one lies in
Level 1, while the other lies in Level 3. In the case of the
utilized? tools, we find a similar pattern: one is in Level
2, and the other is in Level 3.

This pattern does not hold in the case of the possible
burin and the drills?--they are located in the lower levels,
where there have not been many other tool types. The
possible burin is in Level 7, while the two drills? are in
Levels 7 and 8.

In the case of the tool that is most likely an informal
tool, the pattern is hard to discern, since there is only

<3ne tool. Also, this lone tool lies in Level 5, which is a
fnjddle level. As for the flakes and angular waste, the

ceattern that applies to the majority of the tools in Area 9

98

applies: Levels 7, 8 and 9 have few artifacts. For the
angular waste, six of these 59 artifacts lie in the last
'three levels. For the flakes, Levels 7, 8 and 9 contain only
39 of the 250 flakes present in Area 9.

The interpretation of the distribution of the data in
,Area 9 shows that the upper levels of the area were activity
areas. In this case, by "activity area" I am referring again
(as I did with reference to Area 3) to the presence of tools
and debitage, which shows that tool-making went on.
Similarly, I cannot clearly say what kinds of activities
took place, beyond tool-making. I can, however, speculate
that the presence of scrapers may suggest that hide or wood
was worked (or both). So, the heavy presence of tools and
debitage in the upper levels suggests that upper levels of
Area 9 witnessed a significant occupation. Conversely, the
low presence of artifacts in the deeper levels means that
these levels did not witness a significant occupation. As
the lower levels were used earlier than the upper levels, I
argue that in Area 9, the occupation by human groups was
restricted to later times. If the area had little or no
artifacts in the upper levels but had a preponderance of
tools, flakes and angular waste in the deeper levels, then I
‘would argue that in terms of time, the deeper levels were
Occupied; thus, the area would have been without much
Settlement as time progressed.

Area 10 differs from the other areas in one crucial

Blement: it is a combination of all the test squares that

99
belong outside Areas 1 to 9. Hence Area 10's squares do not
completely lie adjacent to each other, but are scattered all

over the site of NIoma. This lack of a coherent geographical

lorigin, unlike in the other areas, is bound to affect the

patterning of artifact types through time, in terms of the

levels (see Table 8). Also, I sampled squares in this area.

I looked at 17 squares only.

For Area 10, the retouched, edge damaged and backed

lithics predominate in the tool assemblage. These tool types

form 143 of the 196 tools present. Including the 34 cores,

the total reaches 177; the rest of the tools include five

informal knives, five blades, seven combination tools, one
crude crescent and blade with edge damage or retouch?

For the area as a whole, the lithics lie mainly in
Levels 2 to 10. Levels 1 and 13 contain only 11 of the 196
‘tools present in Area 10. Also, Levels 11 and 12 have no
artifacts at all--no tools, no flakes and no angular waste.

A breakdown of the area's tool types shows that the
edge damaged artifacts are spread all over the area. They
are present in Levels 1 to 10, as well as in Level 13. There
‘is a fair distribution throughout the levels, except that
leavels 4 and 13 have two pieces and one piece respectively.
111e cores also show an even spread, with only Level 13
having no cores at all. For the retouched pieces, only
Levels 1, 2, 6 and 13 have one or two pieces each. The other
levels have a larger number of tools; the highest number is

1'1 in Level 5. For the backed artifacts we find that they

100
are spread across the area evenly. Only one level has no
backed pieces. Level 10 also breaks the pattern because six
of the 19 pieces that are backed were found in this level.
The rest of the levels have one or two pieces each.

The remaining artifacts are few in number. These
remaining artifacts comprise combination tools, informal
knives and blades. The combination artifacts lie in the
deeper levels, for six of them are in Levels 7, 8, 10 and
13. The total for the combination tools is seven. Similarly,
the five informal knives lie in the deeper levels, as four
Iof them lie in Levels 7, 8 and 9. But the blades, totalling
five, are divided almost equally between the shallower
levels (2 and 3), and the deeper levels (7 and 8). There are
three in the first levels, and two in the latter.

The crude crescent and the blade with edge damage or
light retouch? are also few in number--there is only one of
each. The crude crescent is in Level 3 and the other tool
type is in Level 9.

I selected seventeen squares in this area, which
consisted of squares that did not belong to Areas 1 to 9.

For the site of Nloma as a whole, we find that the chi-
square test shows that the differences between the different
areas are statistically significant except Area 5 (see

Tables 1-8).

101

Here is the Pearson chi-square data:

Area value df prob.
3 94.262 70 .028
4 65.191 42 .012
5 19.621 24 .718
6 188.600 72 .000
7 217.350 156 .001
8 90.753 72 .067
9 192.785 136 .001
10 162.630 110 .001

It is unlikely that the observed differences between
the areas occurred by chance alone. However, we have to
remember that the results for the various areas are tempered
by the small frequencies of artifacts in certain artifact
categories. In cases where there are less than five
artifacts, the results are suspect. Therefore, I have to
list those artifacts, for each area.

The affected artifact categories are:

1. Area 3 - informal knives, crude crescents, scrapers,
combination tools and drills?

2. Area 4 - crude crescents, combination tools,
retouched pieces and edge damaged lithics.

3. Area 5 - informal knives, backed pieces and
combination tools.

4. Area 6 - crude crescents, combination tools, drills?
and utilized? lithics.

5. Area 7 - blades, informal knives, scrapers, points,
notched piece and drills.

6. Area 8 - informal knives and drills?

JOE

bl:

dr.
cou

ne

102
7. Area 9 - points, large bifacial tool, drills?
possible burin, utilized? unretouched point and informal
tool.
8. Area 10 - crude crescents, combination tools and

blade with edge damage or light retouch?

The following section will consist of the conclusions
drawn from looking at the lithics from Nloma in the broader
context of hunter-gatherers' interactions with their

neighbors.

CHAPTER 7

CONCLUSION

At the beginning of this study, one question that I
posed dealt with anthropologists' rationale for studying
living groups of hunter-gatherers and how they interact with
their neighbors. I indicated that there were several reasons
for this undertaking. An obvious explanation is that as
human beings, we are curious to know about the past, so we
have a scientific desire to discover our roots. Related to
this scientific desire to know about our past is the fact
that, as Lee and DeVore (1968) stated, hunting and gathering
has been the main adaptation of human beings for most of the
time that human beings have been in existence as human
beings. Therefore, studying extant societies of hunter's and
gatherers can enable us to inquire into the hunting and
gathering way of life. We can find out about what human
beings do--social organization, labor organization,
cooperation, regional organizations, trade, exchange,
relations between neighboring human groups (Binford 1990:
508). Also, there are not many living groups of hunters and
gatherers, so the few remaining ones offer us opportunities

to ask those questions that enable us to advance theoretical

103

104
knowledge and help these societies become less marginalized
in today's nation-states. It is true that groups of hunting-
and-gathering peoples often have low socio-economic standing
as opposed to their neighbors, be they farmers or traders or
Ininers, etc. Also, archaeological studies provide a longer-
term perspective than ethnographic studies, which take a
shorter temporal view. Again, ethnographic studies have also
produced some confusion in hunter-gatherer studies.

In the case-study I chose, that of the Tsodilo Hills,
the San hunter-gatherers help highlight some of the concerns
that anthropologists have been dealing with. For instance,
the question of what to call these hunter-gatherers has been
a vexing one. The term "Bushman" has been used, but is now
considered to be pejorative. The term "San" or "San-speaking
peoples" is also used. Then, in Botswana, the term "Basarwa"
is also used. Richard Lee also used the term lKung. Also,
the term "Zhu" is often used. This question of what to call
these hunter-gatherers points to the disagreement that
characterizes studies of hunter-gatherers, not only in
Southern Africa, but in many parts of the world.

I also have to point out that the hunter-gatherers of
the Kgalagadi in Botswana have a high profile in the
anthropological literature and the anthropological world and
the popular imagination. These people take a pre-eminent
position among hunter-gatherers partly because of the work
of Richard Lee and the Harvard Kalahari (Kgalagadi) Project.

Also, several films have contributed to the popularization

105
of the Basarwa, eg "The Hunters" by John Marshall, and the
commercially successful, "The Gods Must be Crazy" by Jamie
Uys, made in 1958 and 1980 respectively. They have
introduced the Basarwa to a worldwide audience. According to
David Wiley, The Hunters is the most widely shown film on
American campuses (David Wiley, personal communication
1991). Coupled with this cinematographic presentation is an
appealing portrayal of the Basarwa as child-like, innocent
and peaceful people who have not been contaminated by the
disastrous dehumanizing effects of the industrialized and
modern 20th century. It is partly because of this portrayal
of the Basarwa as people without history, caught in a static
capsule of time that we have a debate on the nature of
hunter-gatherers and their interactions with their
neighbors.

For convenience, I divided the main protagonists into
two camps. I have to reiterate that this division was for
ease of writing and my intention was not to belittle any of
these scholars who have contributed significantly to our
understanding of social and environmental processes in the
Kgalagadi and, by extension, in the world at large. Again, I
have to indicate that by grouping these scholars in the
manner I did, I am in no way saying that their positions are
fixed and unchanging, and that members of each camp
uncritically agree with each other, and always oppose
members of the other group. Again, I need to point out that

the views of these scholars, given the lapse of timnpthat

106
has passed since I undertook my study have of necessity been
changing; for instance, Sheryl Miller was a pioneering
scholar in the 1960s, and John Yellen is still working on
his Kgalagadi research. Furthermore, Edwin Wilmsen and James
Denbow continue to carry out archaeological research in
Botswana.

I also have to point out that the African voice in this
study is represented by two Botswana scholars. In studies on
the Basarwa, it would be ideal to quote Basarwa scholars,
but at the moment, that is not possible as there are no
secondary school and university scholars of Basarwa stock.
The closest I came to this undertaking was to deal with the
work of some Botswana scholars who have also touched on the
question of socio-economic relations between Basarwa groups
and their neighbors, including members of the main Tswana
groups in Botswana and other subordinate ethnic groups.

Therefore, the study I undertook primarily concerns the
views of the group that I termed the "Rousseauesque camp."
This is a term that I chose in a narrow sense to illustrate
the lack of a diachronic perspective in their approach
towards hunter-gatherers and their contacts with their
neighbors. By contrast, the opposing group of scholars tend
to emphasize responsiveness to opportunity and change, so,
again, in a narrow sense, I labeled this camp the "Pragmatic
camp." As the third group has a lesser contribution in terms

<>f published output, this group does not figure largely in

'hy study. However, I called it the "Botswana scholars." So

107
in the Rousseauesque camp, we have Sheryl Miller, John
Yellen, Richard Lee, George Silberbauer, David Phillipson
and Jiro Tanaka. Among the Pragmatists, we have James
Denbow, Robert Hitchcock and Edwin Wilmsen. Lastly, the
Botswana scholars consist of Mabunga lehwa Gadibolae and
Gaontatlhe Mautle.

Essentially, the Rousseauians emphasize little contacts
between hunter-gatherers and their neighbors; they argue for
a distinct lifestyle of each group. They argue that lithics
show that there was isolation between hunter-gatherers and
their Bantu-speaking neighbors. Also, while trade may have
taken place, it did not lead to intermingling that led to a
loss of cultural identity. Also, hunting and gathering as a
lifestyle is seen as having persisted for millennia, so that
today's hunter-gatherers live as they have lived for
thousands of years. Thus, studying them, especially in their
retreat areas, places of inhospitality, offers us a window
into how Early Man lived. But the Pragmatic camp challenges
this view; they argue that as times change, and the
environment changes, men with the times also change. Their
evidence comes from geomorphology, archival sources,
ethnohistory, linguistics and archaeology. They argue that
far from being isolated, the Tsodilo Hills area of
northwestern Botswana has had at least 1,500 years of
interaction with its wider region. Long-distance trade
occurred as evidenced by trade items like cowrie shells.

Also, people in this area domesticated livestock. Also, they

108
used domesticated plants such as sorghum. In addition, they
used iron, and lived in mud houses. At the same time,
hunting and gathering went on, and stone tool production
also occurred. More recently, historical sources show that
the Basarwa used to be involved in trade, and kept domestic
stock. Again, today, Basarwa do have livestock in some parts
of Botswana, and have adapted to their environment by
seeking employment on farms and mines, as well as growing
food crops. Again, using biological studies, the Pragmatists
show that the distinction between a "Mosarwa" and a "non-
Mosarwa" in Botswana is a cultural, and not biological,
designation as some "Basarwa" are not genetically different
from "non-Basarwa" Botswana and other Negroid groups.

The Botswana scholars‘ perspective is largely
historical and more limited in time. Again, they highlight
the use of force in the subjugation of less powerful and
less closely knit ethnic groups by more powerful ethnic
groups. Again, they provide an insider's view of the
interaction process, and through their knowledge of some of
the local languages, customs, politics, economics, and
moral, philosophical and religious systems, they bring out
perspectives that are different from those portrayed by the

Rousseauians and the Pragmatists.

Ingvy1gg from $0: §jte of Nigma, Isggilg Hills, Botswana

I chose to tackle the question of interaction between

 

hunter-gatherers and their neighbors by taking Tsodilo Hills

109
as a case-study. The Tsodilo Hills lie in the area near
Dobe, where Richard Lee and his co-workers have done a lot
of their research. Therefore, making comparisons is easy
because both Tsodilo and Dobe lie in the same area.

This part of Botswana is relatively unexplored in terms
of archaeological reconnaissance, so the site of Nloma helps
to advance our understanding of the prehistory of the area.
Again, the site of Nloma has evidence of settlement by
Bantu-speaking peoples and their technology, while there is
also evidence of stone-tool using, which is often associated
with San-speaking hunter-gatherers. In following the
terminology that is currently used, N!oma is an Early Iron
Age site with Late Stone Age components.

As has been stated previously, this site on the Female
Hill of Tsodilo is an agropastoral site. It dates to between
AD 700 and AD 1000. Its principal components date to between
the ninth and tenth centuries. Its finds include cattle,
which were the mainstay of the economy. Also, there were
sheep and goats, as well as sheep and millet. Sorghum and
millet were also present. Mongongo nuts and Grewia berries
attest to foraging. Wild faunal remains are also present.
Trade items include ivory, iron and copper ornaments as well
as iron tools. Long-distance trade is shown by the presence
of cane-glass beads and cowrie shells. Contact with the
Okavango Delta, 70 km away, is shown by the presence of
freshwater mussels and fish. Also, there were pole and daga

(daub) houses. Also, the site has stone tools. Also, in one

 

 

110
area of the site, stone tools, tuyeres, millet and sorghum
remains, as well as cane-glass beads and cowrie shells were
found on burned hut-clay floor (Wilmsen and Denbow 1990).

Certainly, the presence of over 5,700 lithics at the
site of NIoma shows that the makers of the stone artifacts
were settled at the site for some time. I argue that they
must have lived there in a non-itinerant way. If, indeed,
this scenario held true, then it could support the
interpretation of Wilmsen and Denbow that local hunter-
gatherers were part and parcel of the local economy of
Tsodilo. At the time of writing, this fact cannot yet be
documented, but a way to test this hypothesis could be
through burial evidence. A way to demonstrate this
connection between stone artifacts and human groups could be
to see what kinds of human remains were found buried with
stone tools, and if there were morphological differences,
then this demonstration might hold.

In addition, the large numbers of flakes at the site,
as well as the large numbers of the angular waste, show that
the tools were made at the site, and did not come from
elsewhere. Moreover, the quartz raw material is available on
the Female Tsodilo Hill; so I argue that tool production
occurred locally.

Furthermore, the presence of activity areas in Area 3
to 10 shows that when people settled at Tsodilo, they
carried out tool production at the site. Additionally, a

wide range of tools were made, including crescents, backed,

111
retouched and edge damaged pieces, as well as informal
knives, scrapers and blades.

However, several questions remain unanswered at this
particular stage of the research. For instance, I am not
able to tell what the scrapers were used for. This calls or
follow-up research on micro-wear. Also calling for further
research is the question of the source of the chert raw
material. Right now, I cannot tell where the chert used to
make the Nloma lithics came from. So, future research must
locate the sources of the chert before this question can be
addressed. Currently, Wilmsen and some of his students are
working on this problem. Another question awaiting
investigation relates to how long the different activity
areas were settled for. Yet another question that needs
addressing is the point of who made the lithic artifacts--
were they hunter-gatherers or agropastoralists? Were these
people the same as those from the nearby sites of Divuyu
(Denbow and Wilmsen 1990:499) and Depression Rock Shelter
(Robbins et al 1989)? There is also the question of
comparison of the lithic variability of this site with other
sites that also have Early Iron Age and Late Stone Age
components together--eg sites in Zambia, Zimbabwe, South
Africa, and perhaps Namibia. It would also be fruitful to
look at the faunal and trade items, as well as the pottery,
and see how they correlate with the concentration or lack of
concentration of the lithics in certain areas and levels of

the site of NIoma.

112

Nloma is a thought-provoking site with exciting
possibilities for future research that will help us reassess
the "dynamics underlying social and economic forms in the
Kalahari [Kgalagadi]" (Denbow 1986:35). In raising questions
for future research, the site of NIoma and its stone
artifacts help us advance our understanding of the past of
the Kgalagadi. In undertaking this task, I submit that a
process of integrating the methods, theories and experiences
of a wide range of fields of inquiry is called for. For
instance, the realms of archaeology, ethnoarchaeology,
ethnohistory, history and linguistics can be drawn upon.

Again, I contend that this process of further inquiry
into the remote past needs to include more indigenous
scholars. There are several advantages to the greater
incorporation of scholars from their native lands. I submit
that it will be a boon to further archaeological research to
have native scholars carry that research as they grew up in
the local environment. They are more likely to be in tune
with the local culture, languages, customs, peoples, laws,
politics, economics, and moral and religious-philosophical
system. As products of the local environment, it makes sense
that they should be more involved in interpreting the past
of their native lands. Again, their being involved in this
research will bear personal, professional and national
rewards (Dr. Yacob Fisseha, personal communication 1991).
Research by local scholars builds their own self-esteem, and

increases the self-confidence of their communities. They

 

113
became role-models to younger children, who also are able to
identify with them. Also, it becomes easier for discerning
informants to pass information on to these indigenes as
there is the commonality of the cultural factors mentioned
above. Again, it makes sense to have native scholars create
and interpret their own past, as the past is "created" and
"reconstructed" in view of the prevailing socio-political

and intellectual times (Parkington and Smith 1986, Shaw

 

1989). Again, local researchers may help lessen the
dependence on foreign scholars; in some cases foreign
scholars become an economic drain on the meager financial
resources of developing nations as they have to be given
financial inducements to attract them to the host country.
Again, since foreign scholars often come from western
countries, this process delays the emergence of productive
debate, collaboration and fruitful confrontation, for the
West is currently in a dominant political and economic
position vis-a-vis non-western countries. When foreign
researchers work in their host countries, they are thus
working as westerners (Hodder, 1986). This can have a
debilitating effect of stifling initiative and blossoming
ideas, partly because knowledge is power. Yet, for research
to go forward, an attitude of cultural humility is called

for.

 

APPENDICES

 

01010.50)“

.3

10
11
12
12
13
14
15

APPENDIX 1

CODES FOR NIOHA ARTIFACTS

slightly edge damaged

edge damaged? .

edge damaged

edge damage on one side

blade

blade?

broken blade?

partly backed and retouched blade
blade with edge damage

broken blade

blade with edge damaged tip

core trimming blade

blade with heavy edge damage
blade with shallow-angled retouch
blade with edge damage or light retouch
backed

broken backed piece

backed and edge damaged

partly backed

backed and retouched

114

 

16
17
18
19
20
21
22
23
22
24
22
25
26
27
29
30
28
31
32
33
34
35
28
36
37
37
38

115
heavily backed
atypically backed
double-backed
backed, manufactured, round shaped tool?
backed, start of a crescent?
traces of backing
core
small core
broken core
worn-out core?
piece of core
piece of core?
blade core
blade core?
retouched split core
crude core
split core
bladelet core?
split core?
flake-core
start of core
backed bladelet
split core with edge damage
backed bladelet
crescent
broken crescent

retouched but not backed crescent

 

39
40
41
42
43
44
45
46
47
48
49
75
50
50
50
50
50
50
50
50
51
51
51
52
53
54
55

116
backed and retouched crescent
crude crescent
large crescent/piercing tool?
minor crescent
crude crescent?
atypical crescent
start of a crescent
crescent?
possible burin
drill
drill?
backed drill?
flake
split flake
freshly broken flake
flake with possible edge damage
flake with some clear, transparent quartz
core refresher flake
flake with shiny metallic stuff
core trimming flake
angular waste
angular waste with metallic stuff
angular waste with metallic stuff?
point
unretouched point
point?

backed point?

 

 

56
57
58
59
59
60
69
61

62
63
64
65
66
67
68
70
71
72
73
74
75
76
77
78
79
80

117
notched piece
unretouched knife (informal knife)
unretouched knife?
retouched
broken retouched piece
retouched, edge damaged
retouched drill/awl?
slightly retouched, minimally retouched,
minimal retouch
retouched (with double backing and broken tip)
possible retouch? retouched?
retouched and backed
possibly retouched or heavily edge damaged
retouched and edge damaged
steeply retouched; steep retouch?
broken retouched piece
scraper
scraper with minimal retouch
large bifacial tool?
utilized?
utilized
backed drill?
utilized? light retouch
bladelet
slightly retouched?
core?

piece off blade core

 

81
82

57
83
84
85
86
87
88
89
20
14

118
backed?
small scraper with backing
edge damaged or retouched flake
knife/edge damaged
backed blade
small scraper
small scraper?
scraper?
small blade core
retouched/edge damaged
informal tool?
backed, start of a crescent?

partly backed?

 

APPENDIX 2

N20MA ARTIFACTS--LIST OF CATEGORIES THAT I GROUPED INTO
SHALLER DIVISIONS FOR EASE OF CLASSIFICATION
A. Pieces without retouch
1. Flakes - code 50.
2. Blades - codes 4, 5, 8, 77.
3. Cores - codes 22 to 28, 30 to 34, 79, 80, 87.
4. Angular waste - code 51.
5. Informal knives - codes 57 and 58.
6. Crude crescents - codes 40 to 46.
8. Pieces with retouch
7. Scrapers - codes 70, 71, 84 to 86.
8. Backed - codes 12 to 21, 35, 36, 48, 55, 75, 81 to
83 and 88.
9. Retouched - codes 59, 61 to 63, 67 to 69, 76, 78.
10. Mixed/combination - codes 6, 10, 29, 37, 38, 39, 60,
64, 65, 66, 69, 71, 88.
11. Points - codes 52 and 54.
12. Large bifacial tool? - code 72.
13. Notched piece - code 56.
14. Drill? - code 49.

15. Possible burin - code 47.

119

120
C. Pieces with edge damage
16. Edge damaged - codes 1 to 3, 7, 9.
17. Blade with edge damage or light retouch? - code 11.
18. Utilized - codes 73 and 74.
D. Others
19. Unretouched point - code 53.

20. Informal tool? - code 89.

F IQIRES

121

 

 

 

 

 

Figure 1. Sane of the countries in the area of discussion
(modified after Phillipson 1977: Figure 1)

 

 

 

122

 

 

,- ‘Qfifilumbo 15a “5
\
.ch1}\
/

   

 

 

Figure 2. Some of the prehistoric sites in Zambia
(modified after Phillipson 1975b)

123

 

 

 

.'§S}§“;§°°
Mwela Rack“ *3: Key:
"“6“ k “ flu 2f“ " ”Wm
Nakn P‘PW. x : ArcmeOlogical
site in Malawi
Zambsi Raw 1

 

Figure 3. Sane of the prehistoric sites in Zambia and
Malawi (modified after Phillipson 1977: figure 33)

 

 

 

 

 

 

 

 

 

Figure 4. Some of the archaeological sites in Zimbabwe
(modified after Phillipson 1977: Figure 37)

 

§ “0“) ’L‘
.. L «WWW .
5Q “W ' (f ‘1
.5 'N‘lM' “New, _ $.59“ ..
E ‘ “\“OVW‘M ~ ktbflafl 1.33%“
...a‘ Mm ? o 0“”

I

i

I
”"7“”’fl
\-
I

V

 

 

 

Figure 5. Sane of the dated archaeological sites in Botswana
(modified after Denbow 1984: Figure 7.1)

 

   

 

 

 

Figure 6. Tsodilo and part of its environs (modified after
Wilmsen 1989: Figure 3.1)

127

 

1!.
Wu

Figure 7. The Tsodilo Hills (modified after Rudner 1965)

 

128

N!OMA-AREAS

 

6
1'.-
a

30 a
. 2. .

’0. .0. .0. 40. .0. 39. )0. 00 1;)! 29' OOH 40H CON .0” IO. .00! DON 100N11ON 120N130“

Figure 8. Nloma-Areas

129

)( ,
m

 

Figure 9. Blades

 

130

Figure 10 . Cores

131

Figure 11. Edge damaged

132

 

Figure 12. Retouched and scraper

133

 

Figure 13. Crescents; backed and retouched

134

 
 

Figure 14. Informal knives

TABLES

135

WuN

NN
‘v-Q
F!

V
N m

hem—Nmm—N

“ouch

~~ooo~oo~ V vm
H

0

@D m

"00000000 H

h.

WN m

00000000“ fl

W

fifiooooflfih

m

~m>mA

Nm

nN

mmooomoo~ m

V

m
m

NW
V

“VDDDDHDW :

m

N
m

Ow
N

Demuoouom :

N

m
p.

ODDDNNDDN m mfi

pa

gmu0h acute

wflmmz Lmuamcc
mfixfluk

muoou Lou "ouch

flwmflEQU 006w
bun—«LO
COmHQCmflEOU
U¢LOJOahfl
taxonm
MLQQQLom

m0>mcx uflELOkCu
MOLOU

nmtw—m

mmmocmaoflLu

nooumutc

m HOLE l aflCEDnOOV ~0>0~ in AWBOLV “HOkafiLfl $0 Ounflh

n Onnflh

1 (cont’d)

Tabla

Level

HrtiFact

888888888

888888888

fidflfiflflfiflfl

888888888

VDDDDRDO?
u-u-n

888888888

300069000

888888888

OODDDDOOV

888888888

zeoooogga

888888888

88000900?

888888888

:SOODOgOa

888888888

°288°°8°8
888888888

136

100.00

11.11 20.37 20.37 16.67 20.37 1.85 1.85 7.41

Total For tools

100.00

7.51 15.02 27.70 15.02 16.% 3.76 4.23 9.%

Grand Total

Total

Level
8

Table of arti?acts (rows) by level (columns) - 8rea 4

Frequencies

flrtiFact

Table 2

137

“WVNN

14
230
27
271
00
00
00
00
00

CHDDCDD (fl r~m

23
00
00
00
00
00

DN~~D v DN
m

56
00
33
00
00
00

DCDDCDfi d Rd!

32
00
00
00
00
00

DCDDCDB
~~m°° w 8“ 8 88888
gag-Edd
OCDDCDD O 800 g; §E38E38
ddddo’

GOOD“ fl RN

28
00
00
00
00
00

DCDDCDD 0 NC“

1

19
00
00
00
00
00

00000 0 ON N 88888

5 E 8 ~ 3‘
.8)?- .:§ § .85
1) "‘ "‘ °"‘ 9 3"
333383 33} g 5338“

.43 100.00
7.41 11.11 100.00

10.00 21.74 7.39 100-00

29.63

.96
18-52
0.74 7.01 10.33 14.39 26.57 11.81 20.66 8.49 100.00

0.00 0.00 7.14 0-00 35.71 7.14 28.57 21

0.00 7.39 10.87 15.65
7.41 7.41 7.41 11.11

Total For tools
angular Haste
Grand Total

Flakes

138

 

909:

209:
Q02:

909:

909:
909:
909:
gag:
909:
909:
Ran:

mun—uva Y tm m
e m a? g

~0uoe

F4

0~00N00 m hm

'

8.8.8.3888. 2. 8.8. 8.

N
p.

000000N N :d

m

AQ‘J

w
08 0

3.3.8.8883. 8.
0000000 m

.-. 8

0000fl0N m *0

N

hhdfi

nu.nh
«065

0000

H040
909:
NENN
R19:
R19:
uhfia
R39:

uldOb flCle

Ofi‘ZLSSQE
882E

m—ooa Lou -uOh

“‘833ugiw
Exfizfltflu
Elgafé
Riga
n$:&_ifa%;
uyka

8‘82u

ungfiaki

fizprgukw

.58i.'1&5u
881:—

flash

.8933u860
53238130
Ezeafi

2:25
88‘s.:ak¢£_

“950
ӣ25

80m0CUJUILL

finafitfi

m IOLC I AICID—OOv ~I)In in A830Lv MfiOflbmfiLfl $0 Unfilh

Niiih

‘ul‘lI‘u

k0 i~nuih

139

hmD— mun mam omN mVV 0m Du Nu uflfiOh UCILQ
N n D u D D O D fiOCMILO¢OflCD
on Nu n mm —m v o a tuna: Lauamcc
ham mm no" NVN Ohm hm On Q floxlnu
mom mu mu MN VV m D V IuOOJ Lam uluOh
u D D O D u D O hUONmnmuD
mm v w m mu ~ 0 a town-Iv omum
N O O O N O O D Flu—«LO
N n u D D D O D COmuchDIOU
"N n u nu V V O D UO£OJO£IN
an N u u m n O D uoxoam
— D D O D D O — IuCOOMOLU OflJLU
w n D O m u D a M0>mcx mlILOKCH
m— V V n h a D N uOLOU
m D O m m D O O UOflflnm
ImmacODUOLu

wouOh an Du m m V N u
~O)OJ aOImmflLm
m 'OLC I a3OLv ~I>In ma nClJuOOv “HOIEMuLO b0 Onfllh

V Onnlh

140

00.00u

00.00u
00.00n

00.00u

00.00n
00.00n
00.00n
00.00”
00.00n
00.00n
00.00n
00.00n
00.00"

~auoh

0k.0n

MV.~
mm."—

V0.N—

h0.0«
00.0
00.00
0N.V
00.0
00.0
00.0
00.uN
00.0

0n

0n.hN

NV.~m
0N.NN

0n.mN

nk.~V

0N.VV
uh.nV

VN.0V

VV.VV
00.00"
00.0
00.0"
00.V0
00.0
00.00
V0.00
50.00

00.5
nh.0
00.h
mm.0

V0.0
00.0
m~.~
00.0

00.0

00.0

N—.u

00.0
00.0

0h.m

00.0
00.0
00.0
00.0
00.0
00.00n
h0.0—
mm.0n
00.0

"sank ucoLm

Gall: LanUCC
fiixflnu

UnOOJ Low "00°F

thNM—mfi0
UIUIIIU 0000
fil—umLD
COmaICMAIOU
005030000
UOXOID
h£COOIOL0 OvJLO
lt>mcx nlILOhC~
IOLOU

non-um

utml¢¢00Lln

noummuLc

afi-0C00v V OuflOh

hues

fihhofrfifidsGudnymwdwhmnw)-Wufl'

lMl

numu:

M'hhl

Q 13

u

m

h:

M1

Ngmmoag~~mg

0V0000H000W

0NH000NH000

o~~~~o~oo~~
~~000000000
“~000Nm000m
ooooamwoooo
0~~~~m0000N
0000fiN00000
00000~0000~
0~000d00fi~0
0N00~~00000
0000~m000~m
0H0d0N0000Q

0~0000fl000fi

Ibbhtmn
lbhb

fihhdfibx

hfimflJNMB
DiMfi
Edg-

mlhl
mm;

10 l 13

E

m

m

3

'hhflfirhnk

a...
“N
,3.
m
a...
Q.

N0

N0

00

&n
#0

m

43 25 EB 2%

EB 15

a

Q'dflflfl

142

8.93 8.0 0.00 8.0 0O.N

8-8O 0N.VO 0.0 8 0 0.0
8&0.— NV.0 05.0 0 0O 84
8 8O 8.0 NN.0O 8.N k O
8.8O 8N.— BJ... 8.0 8.0
8.8O 8.0 8.0 8.00 8.0
8.8O 8.0 8-0 8.0 8.0
8.9.: 8.0 8.8.. 8.0 8.0
8.93 8.00 8.8 8.00 8.0
8.8O 8.0 BR 8.0 8.0
8.8O 8-0 8.0 n0.NO 8.0
8.9”: 8.0 8.0 00.8 8.0
BAH-O 8.0 8.8 00.00 8.0
8.9.: N.“ O0.00 0.0 0.0
8.9”: 8.0 8.0 8.0 8.8

('0
F.
N
O.
F!
O.

0O.h 0V.OO N0.0
0N.VO 0B.V 8.0
8.0 0N-VO N0.0
8.0 NV.V 0K
n0.h 8.0 8.0
8.0 8.0 8.0
8.0 8.0 8.0
8.0 8.0 8.0
8.8 BAN 8.0
8.0 n0.NO B.NO
8.0 8.0 n0.NO
8.0 8.0 8.8
8.0 8.0 8.8
0.0 8.0 0.0

8.0

0N.0 0.0 8.0

0O.m 8.0 0N.V
00.0 O0.0 6V0

N0.00 k; 0O .0

$.N 0.h O0.00 V0.V
8.0 0h.V 8.0 8.0
0V.N NS 0N.VO NVo0
0.0 8.0 0O.VO a...“
8.0 n0.NO 8.8 80
8.0 8.8 8.0 8.0
8.0 8.0 8.0 8.0
8.0 8.0 8.0 8.0
8.0 8.0 8.0 8.00
NO.V n0.NO 8.0 8.0
H0.NO n0.NO 8.0 8.0

8.0 8.0 8.8 8.0
8.0 8.0 8.0 8.0
O0.00 8.0 0.0 0.0
8.0 8.0 8.0 8.0

V

0 N

O

3». 3.0

i ting
Iv‘un—

uOoBbm O30»

tum-Iv imam.
52.:

83a unfit:
gnu
Smite-8

..On'hmmufc

gain—Ow 0 Omn‘h.

ID Tdal

9

Table 0? artifacts (rout) by level (coluuns) - Bro. B
Lmnl

[Wupunhs
urn:

fliflhd:

Tdflofi

8”83”"P

YO0N000
O O

00*h0ut
F0

N00000N
O

N~~0~00

000000N

0000000
F.

0NH0~00

N0~f000

00~0~0~

~0fl000?

hfiammlhfiwn
Enhw
Ombfiwfiaw

Drills?

fiwmnhd

um

24

18

18 31

12

9

TdalFa'unh:

5613-de

“3

SN
a0

8O

Indus

133

161

74 113 301 69 119 106

31

Patumqps
Gr":
hfiammlhfiwn

Hmphrihflw
Sndedal

8888888

8888888

fiFIMMv-‘flfi

8888888

000000h
OON O

8889883

édddddd
N OOmO—n
O

8828888

0000000
O O

8&3 383

0000000
MOO

88$k88$

000000N

8838888

”0&0000

£33333.

0800302

8888885

0000000

BSRSQBB

000000“
.4

883k888

$000000

lbdfimfihm
Drills?

lhumdnd

Bu*«i

Tdalfa'unhltfié IJEELZEELEIflLaitJS 9£BIZJM 1&431EJ31GLUJ

Eirchmgu!

CI!!!

152.
0h

FIFO

.3.

339!
RN
.53
*5.
WV
3.3.
N8
3‘5".
WV
83

I 0
N0

8%

9.6

Ind“:
fimphrihfln
Gndedal

 

5.52 2.29 5.45 9.32 22.19 5.99 9-79 7.91 22.75 11.99 199-99

M4

fim N
W h
V

~vmm—w~8~~— 8
CI.

FIFO

Nmm
a

3am

93

III

oooooo~~~aoomooa W RN

m

mo 0

ODDODDOODDflDDDDD d

m h?
g

—

omN—o~~coo~~mooo

h

VOdOdeddDODWDOO

W

WYDDNdeOODDVDDfi

m

ut‘d

8

km
Vu

N
N

fiVNONmDfiDfiODmDOO

Y

~m~o~~~ooocoo~~o

m

§

NO
mu

1%

DNYN~~N~NODDW~DO

N

samba

II.

NwflNNfimfiaOOOYDfiO

unuOh uanm

Risaizarc
“3‘3.

a_oou Lou _o»0h

buOOfi ulILOQCu
deOQ t6£030¢0Lc3
Ptiflmnmfifi
flimliflfl Umfiw
CmLJD Ounmanl
bunnmLD

gooa game-«ma omLIJ
IACmOQ
com¢lCmniou
ultoaouoa

tOJOIm

ILOQILOW
RJCOouiLO ODJLU
It>mcx alCLOKCu
IOLOQ

Ilfllnm

fl0m0C0300Lu

“fiszye

m IOLC I ah€I3~00v ~I>On In AllOLv U¢OIkm¥LI $0 O—Dlh

hlfiwh

Tabla 7 (cont’ :0

Love 1

Mi?act

otal

T

145

8888888888888888

8888888888888888

MI'IHHFIF‘HF‘FIHF‘F‘dF‘I-IF.

8888883388888888

OOOOOONEROOOBOOO

8888888888888888

OOOOOOOOOO8000OO

883883888888888

onmwomhoooognooo
«mu m

8888223588883888

NOOOZONO OOONOOO

SNOONZ8000OOOOO8

$888NR 888888888

:203NO:;OOOO8OBO

'é ..
£3 ~ .5 is
sé §§ g} g

gaifiigg-Eaizé BE:

aszsmaésmggz

HILOO

3.57

8.93 0.60

19.64 20.83 6.55 16.07 14.29 9.52

Total For tools

16.95 1.69 2

IDJIJ 24.40 9.60 1
.%

ll

Flakes
thulr Haste

NILE!)

5.24

7.55 1.26

13.63 22.22 7.55 18.87 “16913.!!!

Grand Tot. l

146

an

Wam~em~m~

figiuh

E

O $OO

OOOOflfiOv-‘O

OmN 9O
n O

5.. m—
an RN
mu «N
O n

O u u

u O

V O

O N

O O

O N

N O

O O
Ow O

O
mun
Odd

uNnONw-tho

a?“

O

Van

um—IOHID'OO

mun

BOO

OYOOF‘NOOO y")

\D

. . .33 99a

5
H
p.
3
O
23

3mg.—
Iall! *3
£03.".

a 8“"
am~
O BOO m

38.. Lo... .10..

End»! in: ...o 8......
an... 5...... 8.3m

nun-in bum
5.8.88

“India

nan-m

38.5 13.6

82...». .85...

p.00

nan-E

N
V
d

m.—

om—no—czogo
OtOONIDONO
nlDOv-‘NNOOO
NNOOv-tv-H-‘h-O
OOOOO—IOVO

wimp-lg“.

:33 aunt;

9 our: .. al.-:33 12: mn 3.5.0 nan-5...... no 3......»
m 03!.

147

BAH:
8.8a

8.8"
8-8—
BOA
8.8g
8.8g
8.8g
8.8a
8.8.—
HOOn

130...

5.?

On .O
NO

B..—

8O
mm.—
8O
84
wow
H0-0
8O
8.0
n0.0

 

Kw.bn O?.—N

9.x. «NAN
O?.mn 0?.“

$.m Bfiu

8O 8.8—
BO fi.m~
ON .?n 8.0
OOK. um .hn
BJO B.Ou
8.0 8.0
8O 8.9
mom OWN"
8.0 8.0

On m

WOO 8O

8.”. B.O
N— .O 8O

nNOu nkO—

8O 8.0
NKO 8O
mN.?u ?u.hm
O2: N0.0
BOa mgm
8O 8-0
BAN 8.8
a.“ 8.0
8.8 8.8

B.N 8.h

Rd
Rd

RN.—
8......

Dw.h 8.?—

BO 8O
3.: Odd
8.0 8O
8d On.“
ad ”O
8.0 8O
8O 8.8
Oth 8O
8O 8O

~24

8O

NOR...
8“

OOO

H0.0
Oh.N
8O
N0.0

BOa

8O
8O

ONE:

8O

um.h O?.m
8O BO
nOO 8.?
N“— ?n .h
8O SO
and" m.m
8O m.?«
8.0— 8."
BOA Rd
8.8— BO
H0.0 8O
nh..?n 8O
8.8 8.9

6.?

8.0
O?.?

8.?

8O
84m
8O
84
8.0
8O
8.0
8O
8O

fi$£

Guanihun-v O .11...

 

 

Table 9

Table 0F artifacts by material

Hrea 3 Hrea 4 fires 5 Hrea 6 Rrea 7 Hrea B Brea 9 firea ID
a c T u c T a c T u c u T u c T a c T a c u T a c T Total‘
Frequencies
Blades 2 5 7 o 5 o 4 1 5 1 5 0 9 1 1 5 o o 23 7 1 31 3 2 5 59
Cores 5 2 1o 1 o 1 2 5 7 14 5 o 19 7 11 1 10 15 25 25 4 0 29 23 11 34 143
InFormal knives 1 2 3 o o o 1 o 1 3 3 o 5 o 3 3 1 2 3 12 u o 12 2 3 5 33
Crude crescents o o u 0 0 u o 0 o 1 o 0 1 o o 5 5. o 5 5 1 0 5 0 1 1 5
Scrapers 1 o 1 o o o o o o 0 o 0 u o 3 3 5 o o 7 2 a 9 u 0 0 13
Backed o 2 2 o 5 5 5 1 1 2 9 0 11 5 5 5 4 2? -1 1o 1 o 11 4 15 19 59
Retouched 2 3 5 1 3 4 5 5 11 5 15 o 21 1 23 24 5 47 52 12 2 0 14 22' 30 52 153
Cambination 1 2 3 o 1 1 o 4 4 1 1 o 2 3 7 15 1 5 5 5 1 a 5 3 4 7 39
points a o o o o u o a o o 0 0 0 1 o 1 5. o 0 3 o n 3 o o o 4
Large biracial tool 0 o u 0 U o o 0 o o o n o 5 5 o 0 5 G 1 o o 1 u o o 1
Notched pieco o o o o 0 o o 0 o a 0 o o 1 o 1 5 o o 0 o o 5 o o o 1
Drills 1 o 1 o o o o o u 1 1 o 2 1 2 3 5 1 1 1 1 o 2 a o o 9
Possible burin o o u o o o o o o u o 0 o o o o 0 0 0 a 1 o 1 o o o 1
Edge damaged 13 9 22 o 2 2 15 9 25 21 15 o 35 17 23 45 35 4o 75 25 9 1 35 45 27 72 310
Blade with edge a 5 o o 5 o o u a o o o 5 c o 5 0 a 0 o o a o 1 0 1 1
damage or light retouch?
Utilized? n o o o 5 o o o o 1 0 a 1 2 o 5 a o a 2 0 0 2 o a n 3
Unretouched point 0 u o 0 o n a o o o o o o o o 0 a o n 2 o o 2 o o o 2
Informal 5551 o c o o 0 u 0 5 o o o o o o 5 5 o a 0 1 o a 1 o o o 1
Total Far tools 29 25 54 2 12 14 23 25 54 51 57 o 105 32 51 113 55 137 193 137 29 2 155 103 93 195 900
Flakes 125 15 142 215 14 230 192 32.224 557 220 o 557 23? 92 329 554 315 1125 205 42 2 250 995 294 1292 4474
Hngular waste 15 2 17 27 o 27 37 5 45 53 7 u 75 19 2 21 4c 5 45 55 2 1 59 93 1 94 375
Indeterminate o 0 u o o o o 5 n o a 2 2 o o a n 0 o a o a o 1 2 3 5
2 1057 255 175 453 950 455 1355 399 73 5 477 1195 390 1555 5757

Total 170 43 213 245 26 271 257 66 323 781 284

 

 

 

RroaS

fire. 4 Rroa

fire. 3

149

Shauing varxabilitg across sit: according to ariFact cat390fig

Table 0? artiFacts b9 naterxal.

T401. 10
Percentages

...V‘w V. m E “1.1..“ x r T. ‘4‘ x. .1. ..PW I m. .«IW ma X.W.H_\U....f-.A .i.

L 1... 3. r .. ..1...

‘
\.
\

s.-

4axwax¢snnwga a amm.m aux fl
SSEZO.RM.S.1.O»0%.CLG 3.10 R mum”, ...w
mmommmmmo 0mmw mm m mmm m
on. u a. D .0.
nunnnrnuuutmuuc non u now 1
$wmmmu ms m mumum mmm a flaw
0.. .0. Q
33900082000u040 000 1 410 4
mnmmmamxoomumn am. y mmm a
L1111111111L110 300 S 460 3
'5 I. 3 I51 1
7%mmoxmxmmmmmmm mmw m mmm m
14111L111111111 110 1 SUD s
1

mammmzzzmow mmm

00
00
30
2.89
£
00
1.15

L1111L1m110 C1111 000 020
no mmmm ommmmmzm mmm u 37m m
1L1111111111111 111 1 491 4
mm mmmwmfimmmwwaw mmm % Hum n
111116111111.11 111 L 110 L
mmmmm7afimmmmmgm mm a mmm 6
111111L11111111 111 L 110 1
0m0000500010000 000 Q 340 6
0 0000500000000 000 a 81 2
111111111111111 111 1 «70 L
mgmmmfi wmm “mm mmm m ”mm m
L1111111111m111 111 1 1L0 1
nammmsyammmmmmm mmm n xgm n
1L1111L11111111 111 1 110 1
69309m960001090 000 a 70 S
35006 050001010 000 90 9
111111L1111m1a1 111 1 111 1

7.
l W
W e
r t
n
35 0.5.1 3
w: 1 n 9n m 1
gm _. i mm 1 o 0"
0m m xm ”1 1 MW 0 ,m
d, :9 amp? r a;
10‘ mu m unmowm; o .
n wd cm s “.3b4u $2 on F s r”
w ”w oo.em 1t ow‘nx 9 ngtr 1 C‘st 1
do n9 .Lsodmleo D Rue a
2.511312030wm0,3,:
B‘LT.C Q P.LNAW ERUHUU I T :.F.1 T

some; :9 (2‘1; 1‘

.,‘ 5 '0. 7
g ' E: C
:9 : .7 2.9.51.69
_ ...0
“‘fi 5 VI ~::E3.oa
...nn '-
..; :1: 11:12.57
5 3nd ""
..r .a t$‘&;.17'95
:51v iv " ._
'. ~~331wfiloo
.30 :13 ”5343.00
-3 .1. T» ;°.00
. 4 mm
A: 13‘.” , 2.22
-' h A.
-. w“; .1100)
[“14 v- ‘17-‘2
B 855'. 10-00
3: 3-33 '*
,. -:°-00
3:; 2.3983100
3 ;£;f'.1co
7- -3 ;L”
’._faooo
(3; "
g ofifl‘
:2.06
’1 3:1.
.~ 5::‘1 g0.53
2:5 '1} .100
.0 " _ all ,l
‘3 2'1, "
5 .3.04
a?‘

H
OWN”

wfi.
$888$3

..-
w
I

I"
p-

.3393“

'0

9 811 1.888 agamaaar

O

(a 0.01"

6-22
17.9?
10-58

0.00

15.63

fire: 8

0.00
10.49
6.06
0.00
0.00
30.34
23 68
12.82
0.00
0.00
0.00
11.11
0-00
12-90
0.00

0-00
0.00
0-00
15-22
7.06
1-32
0-00

7.96

0.00
17.49
9.09
0.00
0.00
34.83
28.42
15.38
0.00
0.00
0.00
11-11
0.00
24-19
0-00

0.00
0-00
0.00
21-44
25-03
11.90
0-00

23.59

38.98
17.48
36.36
62.50
53.85
11.24
6.56
12.82
75-00
100.00
11.11
0.0“
9-03
0-00

66-67
100.00
100.30

15-22
4.60

14-81
0.00

3‘0; 9
C U
11.86 1.69
2.00 0.00
0.00 0.00
12.50 0.00
15.38 0.00
1.12 0.00
1.09 0.00
2.56 0.00
0.00 0.00
0.00 0.00
0.00 0.00
11.11 0.00
100-00 0.00
2-90 0.32
0.00 0.00
0.00 0.00
0.00 0.00
0.00 0.00
3.22 0-22
0.94 0.04
0.53 0.26
0-00 0.00
1.27 0.09

H
o I 0 o

O

888 Bussééésaaasaga

p
Lf)O(DC)O‘Qb)¥CJO¢FU\M

H
QC)? 8

RFQJ 10

3.39
7-69
9.09
12.50
0.00
16.85
16.39
10.26
0.00
0.00
0.00
0.00
0-00
8.71
0.00 1

0-00
0.00
0.00
10-33
6.57
0.26
40.00

6.77

T

8.1?
23.78
15-15
12.50
0-00
21.35
28.42
17.95
0.00

. 0.00

0.00
0.00
0-00
23-23
00.00

0.00
0.00
0.00
21-78
28.88
24.87
60.00

27.53

TuLal

100.00
100-00
100.00
100.00
100.00
100.00
100 00
100.00
100.00
100.00
100.00
100.00
100-00
100.00
100.00

100.00
100-00
100.00
100-00
100.00
100-00
100.00

100.00

 

 

W939

W934 W265 FreaE- FY93? Frees

T331? OF a‘tiFas‘ts 1:13 material, mirg va'iabilitg uifliin areas
8'63 3

T3319 11

C

p

Black-s
Cone-s

2?2Dno.~a3&0n00%u

wammmaxmmmmwmmamww
00

lmaL06§4ooumoaso

asymmaxmmmwmwewm

2210031200000600000

mmu afisawmmmgmemwm?

.

8.7356&310010203110

mmmmmmmmmmmmmmmmmmm

Ex0101000n...0&0.umwnwnfuluo

«mu5m4m mmmfifimmmm .mm

30363630003310.0000

_-L

mamau mm$9nmnmumm%%n

b883573320000m00110

msfimmm wu mmem$mmmmm

021006510000 $00000

%:mwn 1.5mmmnmmmmwmm

51095a300000200100

wga mmagmmmmmmmmmmmm

0710078100000m00000

wagmamzwwmmammmmwmm

1120271800020£00000

awmmmamm mmmflmwwmmmm
LBana-imanwo .azuaamnau

flimmmmaaamfiamflammmm

3&00003930310QDO L000

332%mg.$mmm$m1mm%mm

r7.UULq,100010-301000

:m wwwoammmmmmux mm

1070(0U000000040100

07amnnysnwman.13mwm

406005£100010500000

$5 m%m9m%mmm%mgmw%mm

L35103n100010100100

z%$mm$y4mmmmmmwmmmm

9Q1001m700000600000

gammmgmfimmmmmammwmm
amuaaaasuuaauuauuua

349mmm$mmmmmmn mmmmw

4?300L7000000900000

m4mmm&W4mmmmmamwmmm

07000Q$700000400000

mmmmmmmammmmmsmmmmm

Doooomzaoooooeo £000

mwmmmmmmmmmmmmmmmmw

0$00u0$000000000000

%a%m$m%$mmm$m7mmmmm
28501395000110m00000

mmmmmmmmmmmmmmmmmmm
£880082&00000%03000

m$5ms mmfimmmamammmmm

6230306300010400000

UUU

mfiws mw avmmmmm7mmmm

m
on
mm

Qfimdescents

SG‘QCI‘EFS

Backed

Ir‘chrmal krlive-S
Curbinaticn

Pom

Lzrgp bi?acial tcol

8-62 ZELEE 12-37

I)

36-

14.21 34-34 39-73 40-01)

6- 2 29-91

46-29 24.41

11.11

[,0 1CD.CU 0.03 ICXlCD ImII] 103.03 100-00 1CD_UJ ICU-U) IUlCD 103.03 103-CI] 103-CD ICELGJ MILO? 103.113 100-00
0.03 10.12

16-72 6-53 23.07

5.1? 10.89 39-39

ICU-CD ICELOCI ICCLCU 103-CD 1113.03 HILCU 103.03 ICC-CD 103-CD 103.
0-82 46. 15

17.125 58. 14 5.35

For tools

\

Total;

Total Fcr tools
Rig-113’ masts
IffiE‘ber‘mirde

 

m%m
man
i...
500
mam
£10
“mm

20
mmm
man
gum
920
Q...

51

paw
mag
mam

$10

30m
amu
mam

14.&

54m
$10.
5......

Q60

B$B
wsa

0mm

.2...

nus

9%m

gmm

$70
flaw

 

 

[I] 1111-1-13 I'll-LCD 18.110 103-03 IOU-[I] 103.03 103.07! 1113.01 100.03 103-CI] ICU-CD 1113-130 ICU-CD 1113-03 ICU-CI] 103-DI.)

 

100-00 HILCD 103-[13 1(D-CD ICU-CU ICU-(,0 103 CD 101.03 1113.03 103.

Gwdmml

REFERENCES CITED

REFERENCES CITED

Andringa, Jouke
1984 The Climate of Botswana in histograms. Botswana
Notes and Rocordg 16:117-125.
Bahuchet, Serge, and Henri Guillaume
1982 Aka-farmer relations in the northwest Congo Basin.
Translated by Sheila H. Van Wyck. In Politics ond
fliotory in Bood Societies, edited by Eleanor Leacock
and Richard Lee, pp. 189-211. Cambridge University
Press, Cambridge and Editions De La Maison Des
Sciences De L'Homme, Paris.
Barnard, Alan
1983 Contemporary hunter-gatherers: current theoretical
issues in ecology and social organization. Aoogol

Review of Anthcooology 12:193-214.
Binford, Lewis B.

1990 Comment on Paradigmatic History of San-Speaking
Peoples and Current Attempts at Revision by Edwin
Wilmsen and James Denbow. Curront Anthropologx_31:507-
508.

Campbell, Alec

1980 Ibo goioo to fiotowana. Winchester Press,

Johannesburg.

151

 

152
Campbell, Alec and David Coulson
1988 Cultural confrontation at Tsodilo. Qotimo 36:12-19.
Campbell, Alec, Robert K. Hitchcock, and Michael Bryan

1980 Rock Art at Tsodilo, Botswana. South Africon Jourool

of §oienco 76:476-478.

Chang, Cynthia

1982 Nomads without cattle: East African foragers in

historical perspective. In Politics and Hjotory in
Band Societies, edited by Eleanor Leacock and Richard
Lee, pp. 269-282. Cambridge University Press,
Cambridge and Editions De La Maison Des Sciences De
L'Homme, Paris.

Chasko, N., Henry Harpending, Trefor Jenkins, and George

Nurse

1979 Sero-genetic studies on the "Masarwa" of north-

eastern Botswana. Botowono Notog goo Roooroo 11:15-32.
Clark, Desmond J., and Brian Fagan
1965 Charcoals, Sands, and Channel-decorated Pottery from
Northern Rhodesia. Americon Anthrooologiot 354-371.
Cooke, C. K.

1963 Report on Excavations at Pomongwe and Tshangula
Caves, Matopo Hills, Southern Rhodesia. South Africoo
Archaeological Bullotin 75-170.

Cooke, Cranmer K.

1969 Rock Paintings - Botswana and elsewhere.

Botowona Notos god Recoroo 2:24-28.

 

153
Deacon, H.J.

1984 The Loter Stone Agooof Southernmost Africa. Oxford,
The British Archaelogical Reports. International
Series S213.

Denbow, James R.

1979 Cenchrus Ciliaris: an ecological indicator of Iron
Age middens using aerial photography in eastern
Botswana. South African Journal of Science 75:405-408.

1980 Early Iron Age remains from the Tsodilo Hills,
northwestern Botswana. South African Jogrnal of
Soioooo 76:474-475.

1981 Broadhurst: a fourteenth century A.D. expression of
the Early Iron Age in south-eastern Botswana. South
Afrioon Archooologiool Bullotin 36:66-74.

1982 The Toutswe tradition: a study in socio-economic

change. In Settlement in Botowaoa, edited by Renee
Hitchcock and Mary Smith, pp. 73-86. Heinemann,

Johannesburg.
1983 on A e economics: herdi wea th nd li ' s
f i f th i

oorly Iron Agoo_Unpublished Ph.D. dissertation,
Department of Anthropology, Indiana University.
1984 Prehistoric Herders and Foragers of the Kalahari:
The evidence for 1500 years of interaction. In Eoot_
Pr n ' Hun r G rer tudi , edited by
Carmel Schrire. Academic Press, Orlando, Florida.

1986a A new look at the later prehistory of the Kalahari.

154
Journal of Afrioon Hjotory 27:3-28.
1986b Comment on Janette Deacon's Review of Egomufluotono_
to Farmers, edited by Clark and Brandt. South African
Acohoeologjoaofiullotin 143:35.

Denbow, James and Edwin Wilmsen

 

1983 Iron Age pastoral settlements in Botswana. South

Africao Journal of Science 79:405-408.

 

1986 Advent and course of pastoralism in the Kalahari.
§oience 234:1509-1515.
Eibl-Eibesfeldt, 1., and Polly Weissner
1984 Foreword. In Past and Preseot in Hunter-Gatherec
Studios, edited by Carmel Schrire, pp. xiii-xiv.
Academic Press, Orlando, Florida.
Fagan, Brian M.
1967 The Iron Age peoples of Zambia and Malawi. In

Bookground to Eyolution in Afrioo, edited by Walter w.
Bishop and J. Desmond Clark, pp. 659-686.

 

Gabel, Creighton
1965 Stone Ago Huntero of the Kafue. Boston University
Press, Boston.
Gadibolae, Mabunga N.
1985 Serfdom (Bolata) in the Nata Area 1926-1960.
WW 17:25-32.
Hitchcock, Robert K.
1982a Prehistoric Hunter-Gatherer Adaptations. In
t n in swa , edited by Renee Hitchcock and

Mary Smith, pp. 47-65. Heinemann, Johannesburg.

 

155
1982b Patterns of sedentism among the Basarwa of eastern
Botswana. In Politics ond History in Bond Sociooioou
edited by Eleanor Leacock and Richard Lee, pp. 223-
267. Cambridge University Press, Cambridge and
Editions De La Maison Des Sciences De L'Homme, Paris.
1986 Sandveld Agriculture in Botswana. fiotouooo_flo§oo
ond Bocords 18:91-105.
Hodder, Ian
1986 A Reply to the Editor. Africon Archoologico! Reviou
4:5-6.
Huffman, Thomas N.
1970 The Early Iron Age and the spread of the Bantu.

South Afrioon Arohooologjoal Bullotjo 25:3-21.

1978 The origins of Leopard's Kopje: an eleventh century

Difaqane. Arnolojo (Rhogooja) 8:1-23.
1982 Archaeology and ethnohistory of the African Iron

Age. Annuol Boyjou of Anthrooology 11:133-150.
Jarman M.

1972 A territorial model for archaeology. In Mooolo io
Aoohooologyd edited by D. L. Clarke, pp. 705-734.
Methuen, London.

Leacock, Eleanor, and Richard Lee (editors)

1982 Eoljtjoo 59g fljotory jg aono Sociotieo. Cambridge
University Press, Cambridge and Editions De La Maison
Des Sciences De L'Homme, Paris.

Lee, Richard

1968 What hunters do for a living, or how to make out on

156
scarce resources. In Man the hunt r, edited by R. B.
Lee and I.DeVore, pp. 30-48. Aldine, Chicago.

1969 IKung Bushmen subsistence. In Environment and
Culturol Bohovjoc, edited by A. P. Vayda, pp. 47-79.

Natural History Press, Garden City.

1979 Eolaharj huhtor-gatherors: mehI uomen ond uorh in o

foraging society. Harvard University Press, Cambridge. 4

 

Lee, Richard, and Susan Hurlich

 

1982 From foragers to fighters: South Africa's
militarization of the Namibian San. In Politioo ono
hiotory in Bond Sociotioo, edited by Eleanor Leacock
and Richard Lee, pp. 327-345. Cambridge University
Press, Cambridge and Editions De La Maison Des
Sciences De L'Homme, Paris.

Lee, Richard, and Irven DeVore

1968 Moo the hunter. Aldine, Chicago.

Livingstone, David

1858 misoionary travolo and resoorchoo in South Africa.

Harper and Brothers, New York.

 

Main, Michael

1987 K ar': L'fe's variet 'n dun and delta. Southern
Book Publishers, Johannesburg.
Mason, Ronald J.

1976 Ethnicity and Archaeology in the Upper Great Lakes.

In Qultucol Qhohgo ohg Continuity. Eooays in Honor of

am s e i f' , edited by Charles E. Cleland,

 

pp. 349-361. Academic Press, New York.

157
Mautle, Gaontatlhe
1986 Bakgalagadi-Bakwena relationship: A case of
slavery, c.1840-c.1920. Botswana Notoo ono Recoroo
18:19-31.
Mgomezulu, Gadi G. Y.

1981 Recent Archaeological Research and Radiocarbon Dates
from Eastern Africa. Journal of Africon Hjotory
22:435-456.

Miller, Sheryl F.
1969 Contacts between the Later Stone Age and the Early
Iron Age in Southern Central Africa. Azanjo 4:81-90.
Nurse, G. T., and T. Jenkins
1977 Hoolth ago the huhtor-Qothoroh. S. Karger AG, Basel.
Parkington, John, and Andrew B. Smith
1986 Guest Editorial. A r' Ar 0 i l
Bullotih 144:43-44.
Parkington, John, and Martin Hall
1987 Patterning in Recent Radiocarbon Dates from Southern
Africa as a Reflection of Prehistoric Settlement and
Interaction. uouhhol of Africon Hiotory 28:1-25.
Phillipson, David W.

1968 The Early Iron Age in Zambia - regional variants and
some tentative conclusions. uouhhol of Afriooh_fliotoh1
9:191-211.

1969 Early iron-using peoples of Southern Africa. In

r e i i , edited by

Leonard Thompson, pp. 24-49. Heinemann, London.

158
1974 Iron Age history and archaeology in Zambia. uouhool
of AfrioohoHiotory 15:1-25.
1975a The Chronology of the Iron Age in Bantu Africa.
Joucool of Africon Hiohory 16:321-342.
1975b Iho Iron Ago in Zambia. Neczam, Lusaka, Zambia.

1977 Tho Loter Prohistory of Eootorh oho §outhorn Afrioo.

Africana, New York.

 

 

1985 Afrioon Archooology, Cambridge University Press,
Cambridge.
Platt, J. R.
1964 Strong Inference. Science 146:347-353.
Rightmire, G. P.

1970 Iron Age skulls from Southern Africa Re-assessed-by
Multiple Discriminant Analysis. ric r
Physical Anthropologists 23:147-168.

1975 Problems in the Study of Later Pleistocene Man in
Africa. Amoricon Anthropologjo£_77:28-52.

Robbins, Lawrence H., and Alec Campbell.

1989 The Depression Rock Shelter Site, Tsodilo Hills.

Botsuana Notes oho Recoroo 20:1-3
Robinson, K. R.

1961 An Early Iron Age site from the Chibi District,
Southern Rhodesia. th f i n A h i
Bullotjh 16:75-102.

1963 Further excavations in the Iron Age deposits at the

Tunnel Site, Gokomere Hill, Southern Rhodesia. South

WM 72:155-171.

 

159
1966 Bambata Ware: Its Position in the Rhodesian Iron Age
in the Light of Recent Evidence. South Afrjoan
WM 82:81-85.
Robinson, K. R., and B. H. Sandelowsky
1968 Further work in the Iron Age of Northern Malawi.
Azanja 3:107-145.
Rudner, lone
1965 Archaeological Report on the Tsodilo Hills,
Bechuanaland. Sough Africon Archooological Sulletih
20:51-70.
Rudner, Jalmar
1979 The use of Stone Artifacts and Pottery among the
Khoisan Peoples in Historic and Proto-Historic Times.

South Africon Archooologjool Bullotih 29:3-17.
Sampson, Clavil Garth

1974 The Stono Ago Archaoology of Southorn Afrioo.

Academic Press, New York.

 

Sandelowsky, Beatrice
1979 Kapako and Vunga Vunga: Iron Age Sites on the
Kavango River South African Archaeological Society.
gooouin Serieo 3:52-61.
Schapera, Isaac

1930 Tho Khoison Peooles of South Afrioo:_Bushmen oho
hottontoto, Routledge and Sons, London.

 

Schepartz, L. A.

1988 Who were the later Pleistocene eastern Africans?

Afcjooh Arohooologiool Roview 6:57-72.

 

160
Schofield, J. F.
1940 A Report on the Pottery from Bambata Cave. South

Afhjooh Journal of Soiehco 37:361-372.

Schrire, Carmel (editor)

1984 Eoot_oho_flhoooht in Huntor Gathoher Stuojoo.

Academic Press, Orlando, Florida.

 

Schwartz, E. H. L.
1923 Iho Kalahari ond its Native Races. H., F., and G.

Witherby, London.

 

Silberbauer, George B.
1965 ort to he Governm t f Be h e
Bushmon Survey. Government Printer, Gaborone.
1981 b' in e l i rt.
Cambridge University Press, New York.
Shaw, Thurstan C.
1989 African Archaeology: looking back and looking
forward. Ath1ooh_Ahohooo1ogjool_fioyiou 7:3-31.
Smithers, Reay
1971 The mammals of Botswana. fluooum_homoih_No. 4. The
Trustees of the National Museums of Rhodesia,
Salisbury.
Summers, R.
1967 Iron Age industries of Southern Africa, with notes
on their chronology, terminology, and economic status.

In Soohground to Eyolution in Afrioo, edited by Walter
W. Bishop and J. D. Clark, pp. 687-700.

 

161
Tanaka, Jiro
1974 The Sushmen. Translated by D. Hughes and G. Barnes.

Shisakusha, Tokyo.

1980 The Son hunter-gathecors of tho Kalahari: o Stuoy 1h
Ecologioo! Ahthrooology. University of Tokyo Press,

Tokyo.
Tlou, Thomas

1972 The taming of the Okavango Swamps - the utilization
of a riverine environment 1J750 - 11800. EQLEESDB.
Noteo and Roconoo 4:147-160.

Turnbull, Colin
1961 Iho:Eoro§t Pooolo, Simon and Schuster, New York.
1983 The Mbuti Pygmios:ohohge ohg odoototioh. Holt,
Rinehart and Winston. New York.
Vierich, Helga I. D.

1982 Adaptive flexibility in a multi-ethnic setting: the
Basarwa of the Southern Kalahari. In Politios oho
fljotory in Bong Societies, edited by Eleanor Leacock
and Richard Lee, pp. 213-222. Cambridge University
Press, Cambridge and Editions De La Maison Des
Sciences De L'Homme, Paris.

Vinnicombe, Patricia
1976 fiooo1o_o£_tho_§loho, Natal University Press,
Pietermaritzburg.
Welbourne, Robert

1975 Toutswe Iron Age site: its yield of bones. fiotouoho,
W 7:1-16.

162
Westphal, Ernst O. J.

1963 The Linguistic Prehistory of Southern Africa: Bush,
Kwadi, Hottentot, and Bantu linguistic relationships.
Africo 33:237-264.

1979 Languages of Southern Africa. In Porsooctivoooonotho
Southern Afri an a t, edited by the Centre for
African Studies, pp. 37-68. University of Cape Town
Press, Cape Town.

Wilmsen, Edwin N.

1979 Prehistoric and Historic Antecedents of a

Contemporary Ngamiland Community. fiotouoho_flotoo_oho
Rocoros 10:5-18.

1982 Exchange, Interaction and Settlement in Northwestern
Botswana: Past and Present Perspectives. In Sottlomoht,
jh Botswano, edited by Renee Hitchcock and Mary Smith,
pp. 98-109. Heinemann, Johannesburg.

Wilmsen, Edwin N.

1989 tond Filloo:with Flioo: A Politjool Economy of tho
Koloharj. University of Chicago Press, Chicago,

 

 

 

London.
Wilmsen, Edwin H., and James R. Denbow
1990 Paradigmatic History of San-speaking PeOples and
Current Attempts at Revision with CA coment. m
Anthrooology 31:489-524.
Wilson 8. H. and T. Dincer

1976 An introduction to the hydrology hydrography of the
Okavango Delta. In WM

 

163
Wm. pp . 33-47 .
Botswana Society, Gaborone, Botswana.

Wobst, Martin H.

1978 The Archaeo-ethnology of Hunter-gatherers, or the
Tyranny of the Ethnographic Record in Archaeology.
Amoriooh Ahtiguity 43:303-309.

Woodburn, James
1972 Ecology, nomadic movement and the composition of the

local group among hunters and gatherers: an East

 

African example and its implications. In goo,
§35511m2nt_oho_uhhohiom, edited by Peter J. Ucko, Ruth
Tringham and G. W. Dimbleby, pp. 193-206 Gerald
Duckworth, Landon and Schenkman, Cambridge,
Massachusetts.

Yellen, John

1985 The Process of Basarwa Assimilation in Botswana.

WM}. 17:15-23.

 

"IIIIIIIIIIIIIIIIIIIIT