Вы находитесь на странице: 1из 367

Daggers

Knowledge
&
Power
The Social Aspects of Flint-Dagger Technology
in Scandinavia 2350-1500 cal BC

Jan Apel

Uppsala
200l

ISSN 1404-1251
ISBN 91-973674-2-7
Jan Apel 2001
PhD dissertation 2001
Coast to coast-book 3
Series editor: Helena Knutsson
Design and layout: martin hgvall
Cover design: Lotta Sundberg and martin hgvall
Cover drawing: Errett Callahan
Revision of the English text: Neil Tomkinson
Published and distributed by: Coast to coast project
Department of Archaeology and Ancient History
Uppsala University
St Eriks torg 5, SE-753 10 Uppsala, Sweden
Department of Archaeology
University of Gothenburg
Published with aid from The Bank of Sweden Tercentenary Foundation
Printed by Wikstrms, Uppsala 2001
341 051

Contents

Preface .............................................................................................. 5
1 Introduction ............................................................................... 7
The theoretical stance ....................................................................... 7
Scandinavia, 2350-1500 cal BC .................................................... 10
Transitions and material culture .................................................... 13
Outline of the dissertation ............................................................. 14

2 Technology and the Division of Labour .............................. 17


Technology and a definition of skill ............................................... 17
The inside/outside of body dichotomy .......................................... 18
Technological knowledge ................................................................ 20
The operational chain ..................................................................... 22
A description of the operational chain ........................................... 24
Knowledge and know-how ............................................................. 27
Skill and flint daggers ..................................................................... 31
Raw material ................................................................................... 31
The production process .................................................................. 32
To estimate craftsmanship and skill ............................................... 34

3 Apprenticeship as a Social Institution ................................ 45


Social facts and collective representations ...................................... 46
Religion, the first institution ......................................................... 49
Durkheimian approaches to cultural reproduction ....................... 52
Post-processual critique of holistic approaches ............................... 56
Collective versus individual intentionality ..................................... 60
Realism or relativism ...................................................................... 66
Social structure ............................................................................... 69
A revision of the Durkheimian view of social reproduction ........... 73
The past in the present and the present .................................. 75
Conclusion ...................................................................................... 83

Jan Apel

4 The Emergence of Social Inequality .................................... 85


Some definitions ............................................................................. 85
The division of labour in society .................................................... 88
Mechanical and organic solidarity ................................................. 89
Population pressure and marginalization ....................................... 96
Pathways to power ........................................................................ 101
Technological traditions as vehicles for the reproduction
of symbolic capital ........................................................................ 104

5 Previous Research on the Determination


of Debitage from Bifacial-reduction Sequences ............. 127
Lithic-production stages ............................................................... 128
Earlier attempts ............................................................................ 132
Recent attempts ............................................................................ 134
Concluding discussion ................................................................. 151

6 Flint technology in Southern Scandinavia ....................... 157


The Limfjord area ......................................................................... 157
Late Neolithic sites in the province of Thy .................................. 178
Eastern and southern Jutland and northern Germany ................ 183
The Danish isles and Skne .......................................................... 193
Late Neolithic flint mines ............................................................ 197
Conclusion .................................................................................... 199

7 Flint Technology in Eastern Central Sweden ................. 203


The sites ........................................................................................ 205
The flint material ......................................................................... 211
An interpretation of the flint technology ..................................... 217
The value of disc-shaped flakes .................................................... 224
Conclusions .................................................................................. 228

8 The Classification System of the Scandinavian


Flint Daggers ......................................................................... 231
The flint-dagger typology ............................................................ 231
Some chronological trends in the Scandinavian
flint-dagger material ..................................................................... 248
The flint-dagger typology and its relevance ................................. 252
A revision of the chronology ......................................................... 264
Conclusion .................................................................................... 272

Contents

9 The Scandinavian Flint Daggers ....................................... 277


Representativity .......................................................................... 279
Swedish flint daggers .................................................................. 279
Scandinavia ................................................................................. 292
Schleswig-Holstein, Germany .................................................... 294
Mecklenburg, Germany ............................................................. 295
Poland ......................................................................................... 298
Thringen, Saxony and Sachsen-Anhalt, Germany ................... 301
North Rhine-Westphalia, Germany ........................................... 301
The Netherlands ........................................................................ 302
Finland and the Baltic countries ................................................ 303
Austria ........................................................................................ 303
The British Isles ......................................................................... 304
General distribution of the Scandinavian flint daggers ............. 304
Down-the-line or direct exchange ........................................... 305
The character of the distribution ............................................... 319

10 The Formation, Reproduction and Reformation ......... 323


Rsum ....................................................................................... 323
The social structure of the Late Neolithic ................................. 329
The formation ............................................................................ 331
The reproduction ....................................................................... 336
The reformation ......................................................................... 340

References ........................................................................... 343

Preface

I do not consider my research to have been a lonely endeavour. I have


always been able to hand out rough drafts and discuss ideas with friends
and colleagues and, even if consensus was never reached, I have greatly
valued the ongoing discussion.
I would like to express my sincere thanks to the following persons
who, in different ways, have helped me to finish this book.
My supervisor Kjel Knutsson has given me theoretical and methodological insights as well as contributing numerous comments and
ideas on drafts throughout the years. Kjel encouraged me to study
technology and social theory and thereby changed my way of looking
on archaeology. Errett Callahan shared freely his vast knowledge of flintknapping and archaeology with me during two summers in Denmark
and Sweden and a two-week stay at his home in 1995, and in many
letters over the years. Bo Grslund provided valuable comments on
drafts, as well as encouragement throughout my years in the Department of Archaeology here in Uppsala. During the last stage of this work,
Ola Kyhlberg gave me moral support and made valuable comments on
the final draft.
The participants and the scientific advisory board of the project From
coast to coast have supported me throughout the time I spent on the
project. I would especially like to mention Helena Knutsson for the
interesting discussions I had with her and the effort she put into the
practical details of making this book, and Helle Vandkilde who served
as my external supervisor.
My fellow PhD students at the Department of Archaeology provided
encouragement and friendship. My room-mate Fredrik Hallgren, Per
Lekberg and Michel Notelid all helped me in different ways. Lars
Sundstrm took valuable time to read the final draft and we also share
5

Jan Apel
a peculiar interest in Danish flint something that a customs officer at
the port of Gothenburg was once made aware of.
Donald Broady and Thomas Hkansson commented on drafts and
contributed to the content of the book. I have had interesting discussions
with Debbie Olausson and Anders Hgberg throughout the years.
Berndt Andersson generously placed his collection of Neolithic artefacts
from sterunda at my disposal. Wladyslaw Duczko helped me to
translate Polish texts and Neil Tomkinson revised my English. Kalle
Lindholm helped me to make distribution maps, Martin Hgvall put
the book together and Lotta Sundberg and Alicja Grenberger made
drawings. High Voltage fuelled my spirit.
I am especially grateful to Carola and Ellinor for being there, for all
their support and for putting up with me, especially during the last
couple of months. We shall spend this summer together all four of us.
A significant part of my time as a full-time PhD-student has been
spent on museum collections, studying flint daggers and flint flakes.
The staffs at the following museums have helped me out with practical
details: the Archaeological Museum in Stavanger, the Archaeological
part of the Gothenburg City Museum, the Gvleborg County Museum,
the Halland County Museum, the Historical Museum in Bergen, the
Historical Museum in Stockholm, the Historical Research Centre at
Lejre, the Jmtland County Museum, the Kalmar County Museum,
the Lund University Historical Museum, the Moesgrd Museum in
rhus, the Gustavianum Museum at Uppsala, the Museum of Natural
History and Archaeology in Trondheim, the National Museum in
Copenhagen, the Oldsagsamling of Oslo University, the Uppland
County Museum, the Vsterbotten County Museum and the Vstmanland County Museum.
During my years as a PhD student, I received funding from the
following foundations: the Berit Wallenberg Foundation, the Birger
Nerman Foundation, the Knut Stjerna Foundation, the Mrten
Stenberger Foundation and the Vallsgrde Foundation.
Jan Apel
Berthga, 26 Mars 2001

Chapter 1

Introduction

This book concerns two main subjects. On the one hand, it deals with
the ways in which the division of labour and the use of prestige technologies are related to social complexity in traditional societies. This
discussion concerns the theoretical aspects, as well as the questions and
observations raised by my involvement in technological experiments in
flint. On the other hand, it also deals with a cultural-historical problem
the nature of the social organisation of the Late Neolithic agrarian
communities in Scandinavia. The social interpretation presented in Chapter 10, springs from a dialectical research strategy in which the theoretical issues of Chapters 2-4 are confronted with the results of the studies
of the flint industries in Chapters 5-9 here seen as a prestige technology
in a stratified society.
This introduction is divided into three sections. It begins with a
presentation of the chosen theoretical stance. This section is followed
by a brief presentation of the Late Neolithic and the Early Bronze Age
in Scandinavia. A third section, in which the subject of each chapter in
the book is presented, follows.

The theoretical stance


Archaeologists cannot base their understanding of prehistory on first
hand observations of how living people act and interact with others and
with the environment. This means that archaeologists concerned with
the social aspects of prehistory often import theories of human action
from subjects in which humans are studied at first hand. During the
last 20 years, theories of practice by sociologists such as Pierre Bourdieu
and Anthony Giddens have been used to interpret how prehistoric social
practices were formed, reproduced and changed. I argue that the dialectic relationship between, or the merging of, structure and individual
7

Jan Apel
agency, maintained in modern social theories in general, cannot be applied
uncritically to questions regarding traditional societies. Due to current
politics, individual agency is considered to be a far more important
factor in social change than it ever was in traditional societies, and
probably more important than it is in modern societies as well. This
does not mean that we should abandon modern social theory for more
dogmatic, classical, social theories. However, we have to consider the
specific characteristics of traditional societies and relate them to the
modern actor/structure-based theories.
In this context, traditional societies is a general term referring to all
prehistoric and historically known societies that did not adhere to
modern economic principles. In Chapter 3, the holistic perspective on
culture reproduction of the French sociologist Emile Durkheim is used
to argue for the necessity to adjust modern social theory in discussing
traditional societies. Since important aspects of Durkheims sociology
have been incorporated into Pierre Bourdieus theory of practice,
Bourdieus ideas will play an important role in this book as well. I do
not believe that the recognition of the fact that individual agents, through
strategic actions, can improve the quality of life for themselves, as well
as ruining it for others, is the greatest achievement of the social sciences.
We all know that this is true and we all feel that we have a considerable
scope in which to act according to our wishes on many occasions in our
social life. Thus, as I see it, the scholars that today, in their polemics
with the proponents of holism and structuralism, have brought these
ideas into the scientific debate are only stating the obvious. No, the
greatest intellectual achievement in the history of the social sciences is
the recognition that important aspects of the ways in which we act are
governed by conditions that lie outside ourselves, in the form of norms,
institutions and social facts that we are seldom aware of. In fact, quite
often these cultural dispositions are difficult to recognise and appear to
be natural. I feel that these achievements are neglected in the current
archaeological debate and that they therefore need to be re-introduced
once again. Chapters 2 to 4 will be devoted almost exclusively to an
attempt to understand this relationship, and certain aspects of this
problem will be returned to in the other chapters as well.
In this dissertation, I reject the idea of the existence of a fundamental cultural relativism that is suggested in some quarters in the
archaeological debate of today. I also argue that archaeology, like any
8

Introduction
other contemporary social phenomenon, is based on the experience of
the object that is given by the senses. However, since this realism is one
of the two pillars of positivism (von Wright 1993 [1957]:145), I must
add that I do not think that archaeologists must adhere to the second
pillar of positivism, namely that the experience given by the senses should
be used only to formulate general, cross-cultural laws. Even though
such laws can be useful tools in trying to understand certain aspects of
prehistory, an archaeology that concerns itself only with such work is
sterile. An outside, modernistic perspective must incorporate an emic,
inside perspective if the social aspects of prehistory are to be discussed.
Thus, I strive in this book to establish a dialectic relationship between
experience and a more formal approach. In Chapters 2 to 4, a theoretical
preconception of Late Neolithic society is established by references to
social theory and anthropological parallels. The conclusions of these
chapters are confronted with different types of flint materials in Chapters
5 to 9.
I also propose that an external reality is able to communicate itself
through scientific principles. This is my personal view of how the world
works and it is based on the ideas of scholars like John Searle, Steven
Lukes, Maurice Bloch and others (see Chapter 3). They maintain that
there are cognitive faculties of perception that are common to all humans.
However, the important discussion of social archaeology should not
revolve around whether or not there is an external reality. A
constructivistic approach to this matter, for instance, that presented by
Pierre Bourdieu, also takes the social aspects seriously, even though it is
based on entirely different principles. The important thing that should
be acknowledged is that sociality, present in the external world or constructed can be objectively studied through concepts like symbolic capital,
habitus and production fraternity. In fact, I suspect that the material
culture of any age may be studied within the framework of Bourdieus
economy of symbolic assets. It has been appreciated that, while studies
of material culture diminished during the 20th century within academic
disciplines such as ethnology and ethnography, we have become more
and more dependent on artefacts in our everyday life (Miller 1987:3).
This state of affairs actually justifies archaeological-artefact studies
because, in the long run, they may contribute to a better understanding
of our own society.

Jan Apel

Scandinavia, 2350-1500 cal BC


In the older archaeological literature, the Late Neolithic period was
defined as the part of the Late Stone Age when flint daggers were made
and deposited in graves. The transition between the Middle and the
Late Neolithic was set in the period when flint daggers replaced battleaxes as the male prestige gifts in the upper graves of the Single-Grave
Culture in Jutland (Lomborg 1973:256 ff.).
The transition between the Late Neolithic and the Early Bronze Age
was more difficult to pinpoint. While the Late Neolithic was defined
by grave types and grave goods, Oscar Montelius (1885) defined Period
I of the Bronze Age mainly by the flanged axes in bronze and these axes
were never found in graves. Consequently, there was a considerable
chronological overlap between the periods. In fact, in many regions of
Scandinavia, gallery graves, with flint daggers as grave goods, were built
at least until the transition between periods II and III of the Early Bronze
Age. Conversely, the local production of flanged, bronze axes in southern
Scandinavia began already around 1950 cal BC, that is, in the middle
of the Late Neolithic period. A more exact definition of the period in
calendar years was therefore awaited.On the basis of a thorough
chronological investigation of the early metalwork in Denmark, the
Late Neolithic period in southern Scandinavia has been divided into
two phases: LN I (2350-1950 cal BC), and LN II (1950-1700 cal BC)
(Vandkilde 1996:139 ff.). LN I is thus contemporary with the middle
and late Beaker phases in the British Isles, the Veluwe and Epimaritime
Beaker phases in the Rhine delta, and the Early Bronze Age in central
Europe. The second phase roughly coincided with the late Beaker phase/
Early Bronze Age, Wessex I phase in the British Isles, the barbed-wire
Beaker phase in the Rhine delta and the classical netice phase in central
Europe (Vandkilde 1996:140, Fig. 134). Since the distinction between
the Late Neolithic and the Early Bronze Age obviously does not reflect
any significant cultural changes, the chronological scope of this dissertation includes period I of the Scandinavian Early Bronze Age and can
thus be set to 2350-1500 cal BC.
During the early part of LN I, in the Limfjord area of northern
Jutland, a Bell-Beaker phase manifested itself in single burials and
settlements with Bell-Beaker pottery, slate wristguards and heart-shaped,
bifacial arrowheads in flint. In these contexts, the earliest flint daggers
were made under the influence of western European forms (see Chapter
10

Introduction
8). All through LN I, northern Jutland remained a core area for the
production of lancet-shaped, flint daggers (Rasmussen 1990:37,
Vandkilde 1996:13, Apel 2000a & b:141, see also Chapter 8). Since
these were the first flint daggers produced in Scandinavia and since my
thesis uses daggers to discuss the development of social complexity as
shown by related prestige technologies it was necessary discuss extensively
the conditions in northern Jutland during LN I in this book.
The battle-axe and pitted-ware traditions of Denmark, the southern
and central parts of Sweden and the south-eastern parts of Norway
merged into a unified material expression in the first half of LN I. In
the second half, a Late Neolithic tradition with several common traits
embraced the central and southern parts of Scandinavia. It included the
production and consumption of characteristic flint artefacts, such as
bifacial sickles and arrowheads, daggers and spoon-shaped scrapers. In
addition, objects such as shaft-hole axes in stone, slate pendants, bone
needles and bronze axes also characterise this period. The netice area
in central Europe influenced this tradition during LN II. The singleburial tradition and the production of fishtail-flint daggers and elaborately
shaped, bone needles, obvious copies of metal artefacts, are cases in
point.
A collective burial tradition was practised in the southern and central
parts of Scandinavia during LN I. In northern Jutland, a type of gallery
grave with a single chamber was common; it was initially intended for
a single burial, but was later used as a collective grave (see Fabech 1988).
A type of gallery grave with several chambers, sometimes with round
portholes made of stone, was at the same time erected mainly in the
central parts of Sweden and in the south-eastern parts of Norway. This
tradition apparently had its roots in northern France (Johansson 1961).
While these collective graves were used, a proportion of the population
was buried in inhumation graves in flat ground in Skne (Strmberg
1975) and certainly elsewhere too, although they are not common in
the archaeological record.
The Late Neolithic is characterised by evidence of a considerable
landnam that is interpreted as the result of an increased population.
Thus, at the transition between MN A and B, c. 2800 BP cal, a landnam
took place in areas that earlier had few indications of Neolithic
populations, such as the central part of Jutland (Kristiansen 1982:263).
Eventually, a landnam also affected the southern and central parts of
11

Jan Apel
Scandinavia during the Late Neolithic (Kristiansen 1982:263). There is
today consensus among archaeologists that the population of Late
Neolithic Scandinavia practised an economy that included crop growth
and animal husbandry (Hedeager & Kristiansen 1988, Welinder 1998,
Ethelberg 2000). This was also true of large areas of Scandinavia with
no previous indications of Neolithic technologies, such as the Norwegian
west coast (Prescott 1995:175), the remote, forested areas of Swedish
provinces such as Smland (Lagers 2000) and Vrmland (Heiman 2000)
and along the Swedish north-eastern coast (Knigsson 1989, Welinder
1998). However, in the inland areas of Sweden, north of the Mlaren
valley in the east and the province of Vrmland in the west, huntergatherer groups were still present and would be so for a long period to
come.
The population of southern and central Scandinavia lived in long
houses during the Late Neolithic. Two-aisled, long houses have been
recovered from the Norwegian west coast and eastern central Sweden
in the north (Bakka 1976, Johnson & Prescott 1993, Andersson &
Hjrthner Holdar 1989) to Bornholm and the southern parts of Jutland in the south (Nielsen & Nielsen 1985, Ethelberg 2000). At Myrhj
in the province of Himmerland in Jutland (Jensen 1973), Fosie in the
south-western part of Skne (Bjrhem & Sfvestad 1989) and Stngby
in western Skne (Artursson 2000), several, Late Neolithic, long houses
that may have been in use at the same time have been excavated. Twelve
house structures of a slightly different kind, dated to the Late Neolithic
and Early Bronze Age, have been recovered at the Hagestad 44 site in
south-eastern Skne (Strmberg 1992:61). Although the excavator was
reluctant to consider these houses as forming a small hamlet (Strmberg
1992:88), such an interpretation must at least be considered, especially
since none of the houses overlapped one another. Consequently, it may
be assumed that several domestic units lived together in permanent long
houses forming small hamlets during the Scandinavian Late Neolithic
and Early Bronze Age. During the latter part of the Late Neolithic, the
size of the houses increased, particularly in central and eastern Denmark, and continued to do so in the Early Bronze Age as well (Vandkilde
1996:285). Helle Vandkilde (1996:285) has suggested that this trend
should be connected with the concentration of metal objects that this
region displays.
It has been pointed out that few social interpretations of the Late
12

Introduction
Neolithic have been made, compared with the Early Bronze Age. Most
of the recently published descriptions of Early Bronze Age society in
Scandinavia acknowledge the institutionalised, hereditary, inequality.
As regards the Late Neolithic, however, consensus has not been reached.
Some scholars regard the period as socially undifferentiated as while
others regard it as socially differentiated (see Vandkilde 1996:259, Fig.
278). Vandkilde suggests that Late Neolithic society was group-oriented
and that the emerging production and consumption of bronze objects
were collectively controlled. During the first period of the Bronze Age,
larger quantities of metal became available and this contributed to the
disintegration of the collective structure (Vandkilde 1996:306).

Transitions and material culture


Some of the artefact categories used on Late Neolithic sites are remnants from the Middle Neolithic period. These include blades with
gloss, apparently used as harvesting knives, and transverse arrowheads
found in the settlements, for instance, on the Myrhj site dated to the
beginning of LN I (Jensen 1973:83, 89). Blades also appear at some of
the Late Neolithic sites investigated by the Thy project (Steinberg 1997).
The production of blades detached from conical-shaped flint cores and
used as harvesting knives and as blanks for transverse arrowheads
continued into the early part of LN I. This production ceased when
bifacial sickles were introduced in the later part of LN I a technology
that was certainly inspired by the dagger technology of the Limfjord
area. During the early phase of LN I, harvesting knives of the Middle
Neolithic kind, made of blades, were still in use on Late Neolithic sites
such as Myrhj in Himmerland (Jensen 1973) and possibly also Stngby
in Skne (Knarrstrm 2000b:40).
The production of square-shaped, thick-butted, flint axes also
continued into LN I. Of 16 southern-Scandinavian sites dated to the
LN I phase, 11 contained flint axes, while only one of 17 sites dated to
LN II did so (Vandkilde 1996:269, Fig. 283, 271, Fig. 284). Thus,
Vandkilde has suggested that the production of flint axes rapidly
decreased at the transition between LN I and II coinciding conveniently
with the earliest evidence for a local production of bronze axes (Vandkilde
1996:271 f.). While lancet-shaped, bifacial arrowheads were made during
the Middle Neolithic period (Larsson 2000:98 f.), there is no evidence
of heart-shaped, bifacial arrowheads earlier than the LN I Bell-Beaker
13

Jan Apel
phase in Jutland. There are no signs of bifacial sickles and spoon-shaped
scrapers on sites dated to the early part of LN I and consequently they
should probably be dated to the later part of LN I. These sickles were
produced well into the Early Bronze Age. Daggers with lancet-shaped
blades date to LN I (types I-III, according to Lomborg 1973), while
daggers with fishtail handles date to LN II (types IV and V according to
Lomborg 1973). Daggers of Lomborgs type VI, as well as bifacial strikea-lights (Lomborg 1959), belong to the Early Bronze Age. Thus, the
Late Neolithic archaeological material stems from differences of local,
regional and chronological origin.

Outline of the dissertation


In Chapter 2, the theoretical principles for an operational-chain analysis
are unfolded and applied to the results of flint-dagger-production
experiments carried out by Errett Callahan, assisted by myself which
will be fully published in a joint book (Callahan & Apel, ms). The
conclusion of Chapter 2 is that the production of prestigious flint daggers
must have required an apprenticeship system in which the knowledge
and know-how of dagger making was transmitted from one generation
to the other.
This conclusion is, further explored in Chapter 3, from the theoretical point of view. Here, a theoretical discussion starts with Emile
Durkheims definition of social facts in relation to socially formed,
technological gestures and ends by describing a terminology concerning
ritual, originally introduced by Maurice Bloch, for the flint-dagger
technology. This chapter also initiates a discussion of the importance of
the relationship between the division of labour and social complexity
that will be continued in Chapter 4. The conclusion of Chapter 3 is
that every practical action will contain repetitive, past-in-the-present,
aspects, as well as generative the-present ones. In other words, all
technologies include a social message that may be used in strategies for
creating and maintaining power.
The ways in which technological specialisation relates to social
complexity in traditional communities are further explored in Chapter
4. Here, it is argued that it may be fruitful to reconsider antiintellectualistic ideas in trying to understand the relationship between
social change and specialisation. Brian Haydens ongoing investigation

14

Introduction
of the pathways to power in transegalitarian societies and Pierre
Bourdieus discussion of the economy of symbolic assets are also
introduced. The chapter also includes an ethno-archaeological example,
based on Jean Browns close study of traditional iron-working in Kenya,
in which the theoretical concepts are brought to bear. The chapter ends
with a description of the social role that the flint-dagger technology
might have played during the Late Neolithic and the Early Bronze Age
in Scandinavia.
In Chapters 5 to 9, the ideas discussed in Chapters 2 to 4 are
confronted with archaeological materials. Thus, Chapter 5 consists of a
research history where previous attempts to define bifacial-reduction
sequences in the flint production debitage are presented. The main goal
of this chapter is to create analytical tools that can be used in the
following investigations of flint debitage from an area close to natural
flint resources (Chapter 6) and flint debitage from an area without
natural flint sources (Chapter 7). In each of these two chapters, the flint
technologies of each area are related to the previous theoretical discussion.
Thus, these chapters contain descriptions of several Late Neolithic/Early
Bronze Age sites with flint artefacts that have been analysed by myself.
In Chapter 8, the flint-dagger typology is discussed in relation to the
previous results concerning the organisation of the production. I hereby
deal almost exclusively with previous research on the Late Neolithic
flint daggers. The main idea here is that the typology does not only
reflect a chronological sequence but also certain recipes of action related
to controlled production in workshops. Here, the system used to classify
the flint daggers referred to in the following chapter is presented. This
chapter also necessarily investigates the chronological significance of the
flint-dagger typology, proposed by Lomborg (1973) to be able to define
contemporaneous dagger types and thus workshops.
In Chapter 9, the Swedish flint-dagger material is presented and
related to materials from other parts of Scandinavia, as well as northern
Europe. In this I investigate whether the geographical distribution of
flint daggers, understood as expressing exchange networks, can be used
in order to understand the character of the political organisation of the
Late Neolithic and Early Bronze Age. I also discuss the possible presence
of workshops as shown by the spatial organisation of dagger types.
Finally, in Chapter 10, the results from the other chapters are

15

Jan Apel
summarised and a detailed interpretation of the social structure of the
Late Neolithic/Early Bronze Age, agrarian societies in Scandinavia is
presented.

16

Chapter 2

Technology and the Division


of Labour

The aim of this chapter and the two following chapters is to present a
theoretical framework that can be used to understand the way in which
the Late Neolithic societies in southern Scandinavia organised and used
their flint technology, especially how it was used as a vehicle for social
reproduction. It will initially proceed from an evaluation of the flintdagger technology based on experiments (Callahan & Apel, ms). Important issues to be discussed are the degree of skill needed to make
daggers and, more importantly, how this craftsmanship may have been
reproduced through the generations. The purpose of this investigation
is to delimit the production context of the flint daggers, both geographically and socially. This initial discussion ends with an interpretation of the reproduction strategies involved in the production of flint
daggers, which is fairly tentative in its nature. In order to formalise this
interpretation, it will be related to a school of thought concerning the
question of social reproduction established by Emile Durkheim and
further developed and modified by Marcel Mauss, A. R. RadcliffeBrown, Raymond Firth, Maurice Bloch and others. This theoretical
discussion will be presented in Chapter 3, and in Chapter 4 it will be
related to the complexity debate within archaeology and anthropology
and to an explicit theory of practice suggested by Pierre Bourdieu. Subsequently, terms such as symbolic capital and habitus will be used to give
social meaning to the reproduction of the technology and its role,
actively and as a metaphor, in the reproduction of the Late Neolithic
society.

Technology and a definition of skill


In order to understand the social mechanisms behind the production of
flint daggers, there is need for a definition of craftsmanship and skill. In
17

Jan Apel
classical philosophy, the term techne embraced what we today call craftsmanship and artistic know-how (Lbcke 1988:540, Bernstein
1987:205). A definition of techne, in relation to Late Neolithic flintknapping, is anticipated and discussed in the following section of this
chapter. It will in turn be used to delimit areas of dagger production
within a larger consumption area, but, more importantly, to try to delimit
those segments of society that made and used flint daggers. Yet, in order
to get further understanding of the Late Neolithic societies in
Scandinavia, it is not sufficient to define techne. The social aspects of
dagger technology from an inside perspective, its fronesis, i.e. the ethical
knowledge involved in the process (Bernstein 1987:204 f.), or the good
and valuable aspects of the technology in a culture-specific sense, have
also to be considered. Since a proper understanding of technology can
be found only in the dialectic (or hermeneutic) relation between techne
and fronesis, Chapter 3 is aimed at gaining an understanding of the social
aspects of technology.

The inside/outside of body dichotomy


In this context, technology is regarded as a coherent system of artefacts,
behaviours and knowledge that can be handed down from one generation
to another (Schiffer & Skibo 1987:595). This definition is ultimately
concentrated on the reproduction of technology and thereby recognises
technology as integrated with other aspects of society, such as the social
institutions that serve to reproduce technological knowledge, i.e. systems
of apprenticeship, and how such institutions function in society. A
holistic perspective is regarded as especially important in this context,
since flint daggers were produced over a period of 800 years. However
this definition does not account for individual expressions and the role
which such expressions may play in the production of artefacts.
Technological studies are often based on an analytical distinction
between knowledge required from a source outside the body and knowledge acquired within the body. Ian Hodder has argued that this line of
reasoning follows a western tradition rooted in classical philosophy
(Hodder 1992:205). Accordingly, people in other cultures do not
necessarily share such a view, and Hodder is most certainly right in this.
However, it is important to remember, regarding non-industrial
societies, what Pigeot says:

18

Technology and the Division of Labour


Unlike in our modern societies, no sharp distinction is made, in
the acquisition of skills, between theoretical knowledge and actual
practice. On the other hand, it would seem that long held views
regarding the purely imitative and non-institutional educational
processes in these societies need be moderated. The pertinent
even if discrete intervention by adults, and the undoubtedly
educational character of many rites and ceremonies are cases in
point. The organisation of youth education among traditional
societies is probably more complex than previously realised (Pigeot
1990:136).

The distinction between the modern western view and other views
on this matter can be traced to the dichotomy between formal and
informal ways of learning. It has been suggested that formal, institutionalised learning, for instance, in the form of the western educational
system, differs in a radical way from learning by doing, which is the
way in which traditional apprenticeship systems supposedly worked.
To support the relevance of this dichotomy, differences, such as the use
of language versus activity as the major vehicle of instruction and the
creative versus conservative outcomes of formal versus informal teaching,
have been put forward (Pelissier 1991:87 f.). However, recent studies
indicate that so-called informal apprenticeship systems include a fair
proportion of theoretical learning. This has led Catherine Pelissier to
suggest that the dichotomy between formal and informal learning should
be avoided (Pelissier 1991:88 ff.). This conclusion may be part of a
tendency in the social sciences of today to avoid the terms hinting at
cultural evolution that pervaded earlier writings on the subject. Earlier
informal ways of learning based on apprenticeship were, implicitly or
explicitly, regarded as a primitive form of education that had been
replaced by formal means of learning in developed countries. It is
interesting to note here that Pierre Bourdieu, at least in his earlier writings,
stresses the differences between institutionalised education versus learning by doing in traditional societies (Bourdieu 1977:87), especially as
regards the form of asset which is referred to as cultural capital. However, recently Bourdieu seems to have moderated his opinion on this
subject in a way that makes his view more compatible with Pigeots.
I shall now proceed to unfold technology and try to pinpoint its
ingredients in relation to the character of learning discussed above. Two
19

Jan Apel
ideas on how technology can be studied in archaeological contexts will
be referred. These two concepts, (1) technological knowledge and (2)
the operational chain, differ to some extent from each other, but I find
that the following discussion will profit from including both these
approaches and relating them to each other.

Technological knowledge
The term technological knowledge (Schiffer & Skibo 1987) embraces
the theoretical and practical knowledge that defines a technology and
has three essential components: (1) recipes for action, (2) teaching frameworks, and (3) techno-science. A recipe for action consists of the theoretical knowledge and all the formal rules needed to make an artefact
from the acquisition of raw material through the different stages of
production (Schiffer & Skibo 1987:597). A recipe for action includes a
list of the raw materials used, the tools and facilities employed, and a
description of the sequence of specific actions carried out in the technological process. The term also embraces the rules used to solve any
problem that may arise. Accordingly, a recipe for action will guarantee
that the theoretical knowledge that is involved in a technology can be
transmitted from generation to generation. A teaching framework will
allow the practical knowledge involved in a technology to be reproduced.
This is accomplished through imitation, verbal instructions, practical
demonstrations and self-teaching by trial and error. In a way, it can be
said that the understanding of a technology is related to the apprenticeship that guarantees its survival (Pigeot 1990:136). The teaching frameworks are not solely involved in the transmission of practical knowledge. They are also concerned with the legitimacy behind a technology.
This form of ideology can be based on anything from purely rational
arguments to arguments like we have always manufactured arrowheads
in this way. These types of explanations can be regarded as rationales
that give the master or teacher authority during the class. It is also likely
that this kind of knowledge, especially when it takes on an esoteric
appearance, can be used in order to control the practical know-how
that is the foundation of a technology (Hodder 1990:156). In other
words, as we shall see more clearly in the following chapters, those who
formulate the rationales of a technology also control it.
The third component of a technology is, according to Schiffer and
Skibo, techno-science. In this part of technology, the scientific founda20

Technology and the Division of Labour


tions for the facts that a recipe for action and a teaching framework will
lead to the production of the desired artefact and that this artefact will
function in the way it is supposed to are presented (Bunge 1974:30,
Schiffer & Skibo 1987:597). It is obvious that a technology does not
have to originate from scientific principles, even though it can be
partially or completely explained by such principles. On the contrary,
history demonstrates that many of the scientific principles that pervade
modern science originated from technological traditions (Schiffer &
Skibo 1987:597). It is therefore not necessary to control the technoscience in order to execute a specific technology. In traditional societies,
the techno-science is often unexpressed. There is simply no need to
understand the scientific principles behind the manufacture and use of
a certain kind of arrowhead in flint, for example; it is sufficient to know,
through experience, that it will work. Techno-science should probably
be regarded as a form of rationale that dominates the modern, western
view of technology. Obviously, this does not mean that archaeologicalresearch strategies should avoid methods based on techno-science. On
the contrary, such methods can produce results that enable us to explain
archaeological patterns with references to the scientific laws of nature.
Lewis Binford has argued that a successful archaeological-research strategy should aim at separating the known from the unknown (Binford
1983a:56). Since an archaeological site consists of deposits created by a
variety of events, the archaeologist should strive to sort out the remains
of events that can be explained with references to our knowledge of
how the world works. What is left when the knowns have been
identified is what the archaeologist should try to understand by engaging
in experimental, historical or ethnographical studies. Such a procedure
will require some constant in the archaeological material that we today
have knowledge of. Thus, the anatomy of the modern reindeer has
been used in order to understand Palaeolithic hunting and butchering
strategies (Binford 1983a:33 ff.). The law-like way in which quartz
nodules break when subjected to mechanical stress has been used to
understand prehistoric, culture-specific, reduction methods (Knutsson
1998:209, Knutsson & Lindgren in press). Petrographic examinations
of different raw materials have been used to formalise the raw-material
choices of prehistoric axe-makers (Sundstrm & Apel 1998). If our
study of the past is regarded as a form of communication, in which the
archaeological material must be able to participate and talk back, the
21

Jan Apel
main problem is the lack of living informants. In archaeological, technoscientific studies, living informants are simply replaced by, for instance,
a petrographic study and it is anticipated that the result of this study
will reveal patterns that will explain certain aspects of the selection of
raw materials in prehistoric times.

The operational chain


The chane opratoire (operational chain) concept, which in this chapter
will be applied to the flint-dagger technology, was developed within
the Durkheimian tradition of sociology. Emile Durkheim defined a
social fact as a human act, thought or feeling that exists outside the
individual consciousness and is imposed on the individual by society.
Accordingly, cognition was regarded as something that was handed down
to the individual from the surrounding society and not something that
could be reduced to the biological heritage or personal psychology of
individuals. Since human agency was seen largely as a result of social, as
opposed to individual, factors, there was very little that the individual
could actually bring to society. Social scientists no longer needed to rely
on the personal ideas of society that the free, individual agents had, but
could instead concern themselves with objective studies of social
institutions and other stable patterns within society. Consequently, the
sociologists should regard social facts as objects, accept them as such
and explain their existence in the external world (Campbell 1986:173).
As we shall see, these ideas inspired not only the development of the
operational-chain concept in technological studies, but also Bourdieus
work on the development and refinement of a general theory of practice
which will be presented in Chapter 4. However, it should be pointed
out that Durkheims realism, i.e. his idea of the existence of an external
world, has been modified by Bourdieu, as well as by another of
Durkheims followers, Mauss (Broady 1991:253).
In an important article, Techniques de corps, published in 1936, Mauss
argued that a technical act, a gesture1, emerges from individually and
collectively constituted, practical reason (Mauss 1979:97 ff.). Consequently, gestures, even those that seemed to reflect biological aspects of
humanity, were regarded as a kind of social phenomena that was learned
1

22

In the following text the term gesture refers to the kind of socially constituted gestures that
Mauss referred to as technical acts.

Technology and the Division of Labour


within social contexts (Broady 1991:253 f.). In other words, human
gestures were seen as a form of social facts. The child, the adult, imitates
actions which have succeeded and which he has seen successfully
performed by people in whom he has confidence and who have authority
over him. The action is imposed from without, from above, even if it
is an exclusively biological action, involving his body (Mauss 1979:101
f.). Mauss exemplified this by pointing out that certain gestures involved
in swimming techniques, running techniques and even sleeping
techniques varied from culture to culture. For instance he recognised
that the girls of Paris changed their way of walking when the cinemas
started to show American movies. Mauss maintained that the new way
of walking appeared as a result of the fact that French girls were impressed
with the American movie stars and accordingly changed fundamental
gestures involved in their walking techniques (Mauss 1979:100). He
also pointed out that during the First World War, when a division of
English soldiers was replacing a French division at the front, 8000 spades
had to be replaced in the trenches. The English soldiers simply did not
know how to master the French type of spade (Mauss 1979:99). From
these and other examples, it was apparent that not only cognition but
also practical gestures were learned within the frames of socially defined
contexts. This was a ground-breaking observation, since the implication
was that the study of gestures concerned sociology. Because the social
perspective on techniques was essential to Mauss, he suggested that
techniques should be classified according to the nature of the education
and training in which the technique was transmitted (Mauss 1979:108).
He maintained that the reasons for making a certain gesture, and not
another one, could not be fully explained by physiological factors but
to an important degree also by knowledge of the tradition which
imposed it (Mauss 1979:109).
Mausss opinion has been summarised in the following way: The
technical act should be apprehended throughout its unfolding, for in
each of its moments, in each of its forms, in each of its gestures, the
social nature of the techniques find their expression (Schlanger 1990:23).
Interestingly, Mauss used the word habitus to describe these socially
constituted gestures. As we shall see, this definition has a lot in common
with Bourdieus discussion of the type of embodied symbolic capital
that he also refers to as habitus. As a consequence of his findings, Mauss
urged ethnographers and technologists to pay close attention to the
23

Jan Apel
different stages and events of a technology and to comprehend their
operational chain (Schlanger 1994:144).
The concept was introduced into archaeology and further developed
by the French archaeologist Andr Leroi-Gourhan (Leroi-Gourhan
1964, 1965, Schlanger 1994:144 f.). In his work, important aspects of
Mausss ideas were incorporated in the development of a functionalistic
view of techniques. It was pointed out that archaeologists had to be
aware of the structural and functional aspects of techniques and thereby,
much in line with Durkheims functionalistic interpretation of social
institutions and practices, techniques were analogous to biological
organisms. In a similar way to that in which the limb or organ is
structured, serves a purpose and operates in a certain manner, LeroiGourhan regarded the technical element, gesture or procedure. An important conclusion from this idea was that archaeologists were urged to
document the elementary means of matter of a technology, because
they were considered to be the building stones that, put together in a
sequence, would reveal the technical aspects of a technology. In the case
of lithic technologies, the elementary means of matter was regarded
as the recognition of different techniques, such as percussion, abrasion,
etc. (Schlanger 1994:145). It also deserves to be mentioned in this
context, since it shows Leroi-Gourhans close attachment to the
Durkheimian tradition, that he was the first archaeologist to use
structuralistic principles explicitly to explain an archaeological material
the Palaeolithic cave art of western Europe (Leroi-Gourhan 1968).

A description of the operational chain


Within this field of research, the gesture is regarded as the lowest
common denominator recognisable in archaeological materials. Different gestures involved in the production of an artefact will produce
different material remains that are diagnostic for that gesture. It is not a
coincidence that the operational chain has been most successfully applied
by archaeologists working on lithic materials. First of all, these materials
are usually preserved in archaeological contexts. Secondly, since tools in
stone are the result of an irrevocable process of reduction, lithic debitage,
as well as the finished object, will reveal important aspects of the production process, providing that they are studied with this in mind, using
technical methods and relate its specific materiality to gestures. In this
context, a gesture is a body movement that, together with tools, raw
24

Technology and the Division of Labour


materials and other gestures produces, for example, a flake: Each
percussion act is expressed into a flake and its negative, and each debitage
sequence leaves on the ground a series of products and by products.
These elements retain, to a various degree, some evidence of the succession of gestures carried out prior to their own detachment. On this
basis, it becomes possible to decipher and reconstruct, with greater precision, the coherence of the knapping process, the techniques employed,
and the aims of the actor (Pigeot 1990:127 f.).
Technological studies of artefacts aimed at an understanding of their
life cycles have a strong tradition in Scandinavian archaeology. Bo Madsen
has demonstrated that important technological knowledge can be
generated by thorough analyses of the life cycles of artefacts, including
experiments (Madsen 1986, 1994, Hansen & Madsen 1983). During
the 1970s, Madsen worked in France together with Francoise Bordes
and Jacques Pelegrin and subsequently brought the idea of the operational
chain to Scandinavia. Kjel Knutsson explicitly applied Robson
Bonnichsens the decision set model (1977) in his analysis of the
production and consumption of quartz and flint artefacts on Neolithic
sites in Vsterbotten, northern Sweden (Knutsson 1988a). The operational-chain approach that will be used in this context differs in important aspects from other methods aimed at an understanding of the life
cycles of artefacts, such as the behavioural chain (Schiffer 1976), or the
examples mentioned above. These differences or aspects include the use
of a neuro-psychological terminology in discussing the degree of skill
involved in the production process of an artefact (see below). In that
respect, the operational chain does not stop interpretation at the technical
level but is well suited to be used in social interpretations of technology.
An operational-chain analysis begins with an in-depth reading of
the archaeological data (Pelegrin 1990:116), which includes the
following steps: identification of the raw material, techniques and
methods, and a technological definition of the production stages that
are represented at a site or within the specific operational chain in question.
This work is carried out with the help of information obtained by
experiments or, when it concerns flaked assemblages, by refitting
(Pelegrin 1990:116). In order to make a chronological division of the
actions and mental processes in use during the making of a stone tool,
definitions that can be used on archaeological materials are needed. Thus,
during the replication of a prehistoric artefact type or during the refit25

Jan Apel
ting of the waste material from the production of a prehistoric flaked
artefact, the different production stages must be defined. Each of these
definitions must correspond to an end product or a series of complete
or fragmented flakes that can be distinguished from other flake types
(Geneste 1989:443, Sellet 1993:108). An important consequence of
this is that, even if the experimenting flint-knapper intuitively feels that
he or she has moved from one production stage to the next, because of
a changed mental template, this stage definition cannot be used until it
has been empirically defined in the lithic debitage. The lack of prehistoric
informants ensures that the credibility of the experimenting flint-knapper
is based on his or her ability to replicate all the conceivable aspects of
the prehistoric artefacts in question. This means that a true replication
of a prehistoric artefact demands an initial analysis in which the production has been defined in detail in archaeological materials. Thus, a
true replica, as opposed to a simulation, is defined as a copy that is alike
in all attributes, while a simulation is alike in only some attributes
(Callahan, personal communication). Because of the fact that there are
few circumstances in which the entire production process can be defined in archaeological contexts by refitting, true replicas are seldom
made by flint-knappers today. Instead, unknown stages, for example,
in the making of flint daggers, can be simulated by the knapper and the
debitage can be defined. These definitions can then in turn be confronted with a prehistoric debitage to see where the newly defined characteristic traits occur. In this way, the work of the archaeologist/lithic
analysts and the flint-knapper proceed in a hermeneutic fashion.
A classic example of how such a emic research strategy works is
Kjel Knutssons investigation of knapped quartz assemblages at the
Bjurselet site in Vsterbotten, northern Sweden (Knutsson 1988b). The
main goal of this investigation was to understand the categorisation of
lithic materials that was used by the prehistoric populations of northern
Sweden, i.e. to identify emic categories in a prehistoric material
(Knutsson 1988b:11). First of all, the quartz material from the Bjurselet
site was classified into basic formal categories, such as cores, scrapers,
debitage, etc (Knutsson 1990:18). At random, a sample of the number
of quartz flakes recovered during the excavations of the Bjurselet site
were subjected to a use-wear analysis. A certain proportion of the flakes
displayed use-wear. These were then tentatively divided into several emic
categories based essentially on the character of the edge. Knutsson then
26

Technology and the Division of Labour


returned to the quartz material from the site and picked out flakes
according to this emic-classification system. After subjecting these flakes
to a use-wear analysis, it was concluded that a relatively larger part of
these pieces displayed traces of use-wear compared to those in the
random sample. It was thereafter concluded that the definitions of emic
categories were reasonably correct (Knutsson 1990:18). Knutsson then
took this dialectic-interpretation strategy further, but there is no need
to recapitulate more in this context, since the point has been made. By
engaging in a communication with the archaeological material, Knutsson
increased his understanding of which pieces of quartz were preferred in
prehistory. In my view, this is a very strong research strategy that justifies artefact studies, since it actually broadens the frame of reference
that the archaeologist has.

Knowledge and know-how


The psychological factors that are involved in the making of each gesture
can be analysed after the initial, in-depth reading is completed and the
gestures have been defined in the archaeological material. Pelegrin has
introduced the terms connaissance (knowledge) and savoir-faire (knowhow) into archaeology. These terms define two fundamental elements
of a distinctly neuro-psychological nature involved in the execution of
a gesture (Pelegrin 1990:118). These two forms of memory, which
have also been referred to as declarative and procedural (Squire 1986),
are involved in almost every practical action performed by humans, but
the proportion of each may vary considerably as between different
activities. In relation to the concepts presented by Schiffer and Skibo,
knowledge is an integral part of a recipe for action, while know-how is
an important part of the teaching framework, especially self-teaching
by trial and error (Schiffer & Skibo 1987:597). Pelegrins terms have
the advantage that they make a sharp distinction between information
acquired from a source outside the body and the type of know-how
that can only be achieved by co-ordinating the muscles involved in a
gesture (Table 2:1).
Knowledge is explicit and declarative (Pelegrin 1990:118); it can be
spread between actors (teacher and pupil) by word of mouth, signs or
written language or by simple observation. In other words, knowledge
is communicative. Know-how, on the other hand, is an unconscious
form of memory that springs from practical experience (Pelegrin
27

Jan Apel
Knowledge

Know-how

Explaining

Acting

Explicit memory

Unconscious memory

Communicative

Intuitive

Theoretic memory

Muscle memory

Lost in case of lost memory

Not lost in case of lost memory

Words (two dimensional)

Mental pictures (three dimensional)


Experience

Table 2:1. Key concepts used to unfold knowledge and know-how.

1990:118). In 1972, when Don E. Crabtree wrote that the experiences


and habits involved in flint-knapping will eventually cause the workers
muscles to respond subconsciously to induced forces (Crabtree 1972:3),
he was implicitly referring to know-how as defined in this context. In
other words, know-how is intuitive, connected with body movements
and is learned foremost by repetition. Where flint-knapping is concerned, this type of muscle memory corresponds to: operations on
the assessed adequacy of the knapping parameters invoked in the current
operation. The mass and quality of the striking tool, as well as the mass
and morphological characteristics of the object to be knapped, are
appreciated through vision and tactile sensibility (Pelegrin 1990:118).
During flint-knapping, abstract knowledge, mental templates (theoretical knowledge) and experience (know-how) are confronted with raw
materials and tools. In this situation, culture and nature are conjoined
and negotiated (Hodder 1990:155). The quality of the raw material
can never be fully anticipated, the gestures made will not always
correspond to the mental templates and consequently there will always
be a certain amount of insecurity involved in the process, which will
continuously force the flint-knapper to re-assess certain situations. In
order to face this challenge, he or she will need to acquire a deep level of
practical know-how that will enable him or her to adapt to the actual
circumstances. In this way, intuitive judgements, i.e. on the morphology of the platform remnant and the presence or absence of ridges and
how these will affect the next flake removal, are vital ingredients of
know-how. This means that the technology will have to be adjusted to
the raw material and vice versa and, as a result of this, a dialogue will
emerge. Hodder is very enthusiastic about this and concludes: This
28

Technology and the Division of Labour


creative linking of the general and particular is what I would call
interpretation, organised by hermeneutic principles (Hodder 1990:155).
Practical know-how cannot be communicated and can be acquired
only through practice. This conclusion will have implications for the
understanding of the technology and distribution of flint daggers,
especially if the technology is concluded to require a deeper level of
practical know-how. A technology with a low degree of know-how can
be spread over large areas during short spaces of time simply through
communication or imitation. On the other hand, a technology which
demands a deeper level of know-how will be restricted to areas and
circumstances in which the raw material and the time needed to apply
the practical know-how involved in the production and to maintain
and develop the technology are available.
To teach someone how to strike a flake from a core is an activity in
which the gestures involved require a larger share of knowledge than of
know-how. It is quite possible for the pupil to make his or her own
flakes after receiving information on the tools to be used, angles, striking directions etc. This large proportion of knowledge, compared with
know-how, thus makes it possible for this activity to be spread over
large areas in a short time. If the method allows for different raw
materials, it can be spread even further. Pelegrin has referred to lithic
productions such as these as ordinary production (Pelegrin 1990:123).
Lithic industries based on simple bipolar or platform methods belong
to this category.
On the other hand, complex technologies as, for instance, the
production of long blades at Grand-Pressigny in France, Spiennes in
Belgium or Varna in Bulgaria, the production of Maya eccentrics in
Meso-America or the square-axe and bifacial production of Scandinavian
daggers during the Neolithic, require a great deal of know-how. The
analysis of such productions with the assistance of experimental
reconstructions indicates that a suitably larger flint nodule, and a
little more patience, are not sufficient for the successful production of
blades, axes and daggers 30 to 40 cm long. These large products require
a much deeper level of know-how (one which is motor and particularly
ideational for the construction of critical sequences) (Pelegrin 1990:123).
This means that the acquisition of knowledge is not sufficient to
produce these artefacts. They require a deeper level of know-how and,
because of the fact that this can only be acquired by practical experience.
29

Jan Apel
No artefacts will be produced until the flint-knapper has learned to
control the repertoire of gestures involved in the production. In relation
to Schiffer and Skibos terms previously discussed, one may say that, in
order to make remarkable objects, it is not sufficient to have access to
the recipe for action; one must also take part in the teaching framework
(in which practical know-how is transmitted from generation to generation). A modern example can be used to illustrate how knowledge
and know-how interact in a systemic context. When someone is learning how to ride a bicycle, theoretical knowledge in the form of the
positions to be adopted, where to put hands and feet, and so on, can be
provided. However, information of this kind is not sufficient. Practical
training will be a vital part of learning how to ride a bicycle, since this
act involves the co-ordination of different gestures and muscles (Hodder
1992:206).
The geographical distribution of a technology will ultimately be
dependent on whether or not different groups of people choose to incorporate it or not. But, independently of free choice, the spatial distribution of a technology will be influenced by at least three other factors
that in turn are related to each other: knowledge, know-how and rawmaterial availability. Individuals or groups can control each of these
factors. The choice of whether or not to embrace a new technology is
open only to those who accept the requirements of these factors. Theoretical knowledge can be controlled by keeping secret the recipes for
action that lie behind the technology. The level of know-how that is
needed to be able to perform a technique can be controlled by introducing exclusive apprenticeship systems open only to members of certain
segments of society. Raw material can be controlled in different ways
by monopolising the sources of raw material.
This conclusion leads to the idea that an advanced technology, i.e. a
technology which demands specialists, will occur only in societies in
which time and labour can be invested for its development and
reproduction. In such a scenario, specialists will have to be subsistentially
supported by others and this will in turn demand a redistribution system in which economic and administrative power is organised beyond
the family unit (Sahlins 1972). On the other hand, one may also argue
that advanced technologies, such as the Scandinavian flint-dagger technology, may appear in communities with a surplus of food that in turn
enables certain persons to become full- or part-time specialists. Such a
30

Technology and the Division of Labour


situation may, for instance, have appeared in France between 20 000
and 17 000 years BP, when the famous, Solutrean, laurel-leaf like, flint
points were produced (Fig. 2:1). The difficulties involved in producing
these artefacts are equivalent to those involved in the production of a
percussion-finished, type I, Scandinavian flint dagger and, as we shall
see, a great deal of knowledge and know-how is involved in the making
of such an artefact. Perhaps it would be possible to argue that a surplus
of food, for instance, in the form of wild horses, made it possible to
enforce an increased division of labour during the late part of the
Pleistocene. In more general terms, this is a scenario that has been put
forward by, among others, Brian Hayden (Hayden 1990). However, it
has also been argued that traditional societies do not encourage technological inventions, since developments of this kind inevitably tend to
provide room for maverick and entrepreneurial behaviour that in turn
threatens the collective values of the community (Binford 1983a:146).
I shall return to these questions in Chapter 4.

Skill and flint daggers


In the previous section of this chapter, some terms that may be used to
define techne were presented. In order to use these terms in a discussion
on the production of flint daggers, the raw materials and the repertoire
of gestures involved in this production must be identified. These data
must then be organised in a way that will enable me to grade the gestures
involved in each production stage according to their relative degree of
theoretical knowledge and practical know-how. The results of this
investigation will be used to discuss the degree of craftsmanship involved
in the production and thus make way for a discussion of how the flint
dagger technology was involved, both directly and metaphorically, in
the reproduction of the Late Neolithic social organisation.

Raw material
The majority of the Scandinavian flint daggers were made of highquality Senonian flint. This type of flint occurs naturally in chalk cliffs
on the Danish islands (south-eastern Zealand, Mn, Lolland and Falster)
and around the Limfjord in northern Jutland (Fig. 2:2, Becker 1988:46).
In south-western Skne, ice-transported chalk with Senonian flint has
been deposited at Sallerup and Kvarnby outside Malm and on the
western part of Jravallen, an ancient shoreline outside Malm (Becker
31

Jan Apel

Fig. 2:1. A Scandinavian, type I, flint dagger (left) dated to LN I (2350-1950 cal
BC)(from Lomborg 1973) and Solutrean laurel-leaf-shaped point in GrandPressigny flint (right), dated to the period between 20 000 and 17 000 BP
(Drawing made by Errett Callahan).

1988:47). Even though there are rare examples of simpler daggers produced from other raw materials, such as Danien flint, Kristianstad flint
and even quartzite, it is obvious that the overwhelming majority were
made of Senonian flint. It is therefore likely that big pieces of Senonian
flint were a prerequisite for the flint-dagger technology. Accordingly,
this raw material was used in our experiments on order to understand
the production process.

The production process


In this study of the production process of the Late Neolithic flint daggers, the definition of each production stage of a generic biface (Fig.
2:3) has been used as a constant (Callahan 1996). The transition from
one production stage to another is defined by the completion of a mental
32

Technology and the Division of Labour

Fig. 2:2. Primary flint sources in southern Scandinavia according to Becker (1993). (A)
Kristianstad flint. (B) Senon flint. (C) Danien flint. (D) Senon flint from Sjlland, primary
and large, secondary sources. (E) Senon flint of the Hov variant. (F) Senon flint from
Jutland with characteristic white spots. (G) Plain senon flint from Jutland.

template, i.e. the completion of a preform with specific characteristics,


according to an idea that in its turn is a prerequisite for continued
reduction. Each of the production stages included several different
techniques and methods and, as a consequence, several different gestures
will be performed. In my analysis, only diagnostic flakes were used to
define each production stage. Diagnostic flakes are created by gestures
or techniques that are unique to specific stages and therefore gestures
producing flakes that are used and created within several stages are not
considered. This means that, within the defined, cognitive framework,
only a certain proportion of the flakes from a production sequence can
be assigned to a certain production stage. Types of flakes that are the
result of gestures that are used during several stages will not be included
in the definition of a stage. Such flakes usually constitute the majority
33

Jan Apel
of the flakes found on a production site. For instance, all flakes created
during platform preparation will belong to this category, since such
flakes are produced during every stage of the production of a bifacial
tool. A consequence of this is that only a portion of the flakes from the
production of a lithic artefact will be assignable to the different production stages.
In this case, stages relevant to the production of flint daggers have
been formulated after production experiments conducted by Callahan
and documented by myself. In order to be able to use this information
on prehistoric materials, a classification system that will tie the
production debitage to the production stages has to be drawn up. In
this context, it is not possible to give an account of the classification
system used on the experimental material. However, several gestures
that in turn generate different types of diagnostic flakes have been
identified during the experiments. These results are based on a dialogue
between the experimenters and an analysis of the experimental debitage
and original flint daggers and preforms. If one considers these results as
valid, they can also be used in interpretations of the prehistoric making
of flint daggers. The following judgements regarding the degree of craftsmanship are based mainly on Callahans vast experience of making flint
daggers, but also to a certain extent on my own limited experience in
making bifacial flint tools.

To estimate craftsmanship and skill


For two summers and during a three-week, flint-knapping course, one
could say that I was acting as an apprentice to Callahan. During these
periods, Callahan produced several flint daggers and in the meantime
explained the ideas, goals, techniques and methods involved in the
production, i.e. the recipes for the actions required to complete each
production stage and in the end a complete dagger. Before I describe
the skill involved in each production stage, it is important to know that
the skill needed to make bifaces will depend on the size of the biface. A
large dagger will demand a much higher level of skill than a small or an
average dagger will, all other aspects being kept constant (Callahan,
personal communication). In the following text, the stages defined by
Callahan (1996) will form the contextual basis for the discussion (Fig.
2:3).

34

Technology and the Division of Labour


Stage 1
BLANK

Stage 2
ROUGH OUT

Stage 3
PRIMARY
PREFORM

Stage 4
SECONDARY
PREFORM

Stage 5
FINAL
PREFORM

Stage 6
FLAKED
IMPLEMENT

Stage 7
RETOUCHED
IMPLEMENT

Obtain a blank (unmodified) piece of raw material. A blank may be a spall,


nodule, irregular chunk, cobble, or any other form suitable for the end
product. Action may vary from simply picking up a suitable piece to
systematic flaking of a suitable spall from a core. Edges may vary from thin
and sharp to thick and squared. Shape is irrelevant.
Create a circumferential, roughly centered edge which is neither too sharp
nor too blunt (ideally between about 55 - 75). Work sh ould focus on the
outer zone with little or no attention being paid to the central zone, crosssection, or shape. Shape and width-thickness ratios may vary in the
extreme. The edge should end up being roughly centered and bi-convex,
without such concavities, convexities, steps, squared edges, or other
irregularities as would hinder successful execution in the next stage.
Create a symmetrical, handaxe-like outline with generous, lenticular lateral
cross-sections and a centered bi-convex edge. Width-thickness ratios
should fall between roughly 3.00 to 4.00 while edge-angles should fall
between about 40 - 60. Focus on the mi ddle zone without loosing control
of the outer zone. Principal flakes should generally just contact or overlap
in the middle zone, except on thin pieces, and be without such
concavities, convexities, steps, or other irregularities as would hinder
successful execution in the next stage.
Create a symmetrical outline with flattened, lenticular cross-sections and a
straight and centered, bi-convex edge. Thickness should gradually diminish
during reduction, so that width-thickness ratios end up falling between
roughly 4.00 to 5.00 or more. Edge angles should fall between about 25 45. Focus should be on the middle zo ne without loosing control of the
outer zone. Principal flakes should generally overlap, often considerably, in
the middle zone. Generalization of the final shape may start now and
patterned flake removals may be implemented. The resultant piece should
be without significant concavities, convexities, steps, or irregularities as
would hinder successful execution in the next stage.
Create a symmetrical, more-or-less parallel-sided outline (if final shape is
to be parallel-sided) of specific shape, with appropriately flattened,
lenticular cross-sections, and a straight and centered, bi-convex edge. The
outline and the thickness should be within one set of principal flake
removals from the final product (i.e., within about 2 - 4 mm at either edge).
Patterned flake removals may be employed, with flake terminations being
feathered. Principal flake scars in the middle zone may or may not overlap
those of the previous stage. Width-thickness rations and edge-angles should
be about the same as on the final product, which may (or may not) be
greater than the secondary preform. Focus on the middle zone while giving
special attention to outer zone regularity. The resultant piece should be
without such concavities, convexities, steps, or irregularities as would
hinder successful execution in the next stage.
Create an implement of specific, symmetrical shape, cross-sections, widththickness ratios, thickness, and contours with a particular flake removal
sequence and flake scar appearance, as appropriate to the type or
anticipated function. The edge should be more or less straight but without
final retouch and alignment, if needed. The focus should be upon the outer
zone, with the flake scars penetrating into the middle zone as appropriate to
the type or function. Fluting, if applicable, is done at this time.
Create a finished implement, with edges and hafting elements being
retouched as appropriate to the type or anticipated function. Focus on the
outer zone only so as to create a sharp, very straight and centered edge, not
prepared in anticipation of another set of flake removals but form function.
Execute basal hafting or finishing elements such as notching, shouldering,
stitching, etc. Lateral notching sequence, if applicable, are applied at this
time. Basal abrasion may also be done now, as appropriate.

Fig. 2:3. Definitions of the stages involved in the production of a generic


biface, according to Callahan (1996:vii).

35

Jan Apel
The purpose of the first production stage (blank) is to acquire a
suitable piece of raw material. This stage involves a certain degree of
theoretical knowledge and a low degree of practical know-how. It is
important that the nodule does not have internal cracks and that it has
the correct proportions: it must be big enough and have the right
thickness in relation to its width and length. After receiving verbal
instructions and participating in collecting potential flint-dagger nodules,
it was possible to collect suitable nodules from shores, fields and chalk
mines.
The purpose of the second stage (rough-out) is to create a centered
edge around the preform. It is possible to create an edge by soft and
hard, direct techniques. I was able to learn this stage after a few days of
practice and it is obvious that the average archaeology student can grasp
this stage quickly, at least on small preforms. It may be concluded that
this stage does not require any great degree of practical know-how.
During the third stage (primary preform), the most important aim
is to see that the preform is given a generous, lenticular cross-section.
This will be achieved by applying a soft, direct technique, for instance,
using an antler billet. If the lenticular cross-section is not achieved, it
will be impossible to make the preform smaller without losing width.
This has to be avoided, since the later stages will be concentrated upon
obtaining a fairly large, width-thickness ratio. The lenticular cross-section
is therefore needed to allow flakes during later stages to reach over the
middle zone of the biface, thereby thinning it down in the process.
During the fourth stage (secondary preform), a dagger preform with
symmetrical contours and with a blade that has a width-thickness ratio
of between 4:1 and 5:1 is demanded. During a considerable part of my
apprenticeship period, I was stuck in the transition between the third
and the fourth stage. I could not thin the preform down without making
it smaller at the same time. In spite of the fact that, through Callahan,
I had access to all the theoretical knowledge needed, in the shape of
what tools to use, striking angles, positions, etc., I was not able to
complete the stage. It became apparent that somehow I lacked a great
deal of the practical know-how needed to cope with the process. This is
a problem that all flint-knappers will encounter to some degree in their
learning processes. This threshold regarding lithic ability and skill can
also be recognised in the evolutionary perspective. It took several hundred
thousand years for Homo to develop the first bifacial handaxes
36

Technology and the Division of Labour


(Abbevillian types with an edge angel between 55 degrees and 75 degrees;
see Fig. 2:4) into Acheulean type handaxes (with edge angles between
40 degrees and 60 degrees; see Fig 2:4). These two types of handaxes
actually correspond to stage-2 and stage-3 bifaces (Callahan 1979:35).
Only Homo sapiens has ever made bifaces with a width-thickness ratio
that is consistent with Callahans fourth production stage. It took
Callahan ten years of practice before he achieved this level of knowhow. But then one has to consider the fact that he was forced to
rediscover the recipes for action only by confronting himself with original
artefacts. In a living technological tradition, these recipes will be transmitted verbally or simply by observation, and all those who are suppose
to be learning the technology will have access to this information. In
the same way I do not need to rediscover the recipes for action involved
in the making of a type-IV flint dagger, since this has already been done
and I have access to this information. In this way, the difference between
theoretical knowledge and practical know-how is enhanced. This is due
to the fact that it is frustrating to know exactly what to do, but not be
able to do it. This may need a further explanation. In Fig. 2:4, the
relationship between the width and thickness of the blade of a generic
biface in each stage of the production is visualised.
To produce stage-4, bifacial, thinning flakes successfully, the striking tool has to hit the preform on the edge (Bradley 1975). Since the
edge on biface preforms has a low angle, the platform (i.e. the edge)
must be prepared, in order that it may not break during the reduction.
If the preparation is too extensive, it will be more difficult to release big
thinning flakes, since the greater part of the energy will be spent on
breaking the platform. On the other hand, if it is not strengthened
enough, the platform will be crushed during impact and, as a result, the
energy will be lost and the flake will be released by a hinge fracture. In
order to make correct judgements regarding the degree of platform
preparation needed, the flint-knapper has to build on experience created through practical know-how. During flint-knapping, fractures will
follow convex surfaces and if a flake reaches into a concave area, the
fracture will most likely stop in a hinge fracture and create a deep scar in
the preform. When a preform has been taken through the first production stages and displays a lenticular cross-section, it is critical to know
where on the preform to reduce the mass in the next removal. Wrong
judgements at this stage cannot easily be repaired. The problems that
37

Jan Apel
$ ! %&&'& &






 
   



 #

  
 
 !" #,' 
# ,'
 !"   
  &   ()
& ( * )
   ! +&

 #.

  

 #

 
  
 !" #' 
#'
 !" '  ' 


 

 
   



 #

  

 #

 
  
 #'  #
 !" '   


 

 
   





 #.

  

 #

 

 #

 

  
 !" #  -#,
 !"   -, 
  &   )
& (  #)
   

Fig. 2:4. Width-thickness ratios of stages 2-5 of a generic biface, according to


Callahan 1996.
38

Technology and the Division of Labour


arise during the fourth stage will affect all those who try to learn bifacial
work. However, the time needed in order to achieve this level of skill
will vary from person to person (Olausson 1997). Eventually, during a
two-week stay at Errett Callahans house in Virginia in 1995, I managed
to complete the fourth production stage on a small, dagger-like biface
made of obsidian with a percussion technique. The skill needed to produce consistently flakes that overlapped in the middle zone of the dagger
blade was firmly associated with the acts that preceded and followed
the actual impact. When these acts were established in my body, they
did not even appear to be difficult any more. This is a result of the fact
that the procedure involved a great deal of practical know-how that can
never be fully grasped without knapping your way through large amounts
of flint. However it was also apparent that this stage also involved a
fairly large degree of theoretical knowledge concerning impact angles
and so on. This means that, even if the practical know-how involved in
these acts will always be engraved in the muscle memory of my body,
the theoretical knowledge involved will tend to be forgotten if not
repeated on a regular basis.
This is as far as my own experience with bifacial flint work will take
me. The following description of the difficulties inherent in different
stages of flint-dagger production stems from information provided by
Callahan. He has been working with flint and flint-like materials since
the mid-1950s. During the last 40 years, he has produced thousands of
bifaces of different kinds and has written numerous articles and a book
on this subject (Callahan 1979 & 1996). Since 1980, he has made over
200 Scandinavian-type daggers in flint and obsidian. I asked him to
give a verbal description of the relative difficulties involved in the different
production stages that he has defined for type-IV daggers (Callahan &
Apel, ms). These stages are in many respects similar to the stages defined
in Fig. 2:3. The most important differences concern the making of the
thicker handle, the grinding and the parallel retouch present on some
dagger types. For instance, during the production of type III, IV and V
daggers, the base will not be worked by the bifacial method after the
second production stage, only to be shaped by square technique during
the latter stages of production (for a description of the square-axe method,
see Hansen & Madsen 1983). In some cases, the dagger blades were
ground in order to facilitate parallel flaking. In such cases, the grinding
took place between the fifth and the sixth stages. It is important to note
39

Jan Apel
that both prestigious and simpler daggers of each defined type were
produced during the Late Neolithic. Callahans description is concerned
with the prestigious type IV daggers:the degree of skill varies considerably from type to type and up and down the staging. Type IV is tops,
type I C (with parallel retouched blade) next... Type I D is also very
skillful...but is more attainful as its percussion only (2 less stages)
(Callahan, unpublished letter, 17 November 1997).
Callahan is of the opinion that the first three stages contain a relatively low degree of practical know-how. Thus, they can be carried out
by relatively inexperienced flint-knappers after they have acquired the
proper theoretical knowledge. It is tempting to consider the idea that
the early stages and the later grinding stage were undergone by youths
in an apprenticeship system on their way to becoming artisans. As we
have seen earlier, the fourth production stage demands a greater deal of
practical know-how. According to Callahan, this stage is 8-10 times
more difficult to complete than the previous stage. This means that
beginners were not likely to be able to succeed. It is likely that the
fourth stage was unxdergone by journeymen with a great deal of theoretical knowledge and practical know-how, but not yet masters. The
gap in required skill between the fourth and the fifth stages is as great as
that between the third and the fourth stages. Consequently, the fifth
stage (final preform), according to Callahan, is 8-10 times more difficult
than the previous stage. Any mistakes that the flint-knapper makes
during this stage will be difficult to repair: its not just a matter of
skill, as I said, but of perception, observation, and intelligence. Rectify
by looking at the originals (I keep the Hindsgavl cast at hand as I make
my daggers. It's my greatest aid). One must duplicate the angles, flake
scar types, et., as said above. Locking a 3-dimensional template in your
brain isnt easy. It comes with know-how (Callahan, unpublished letter,
17 November 1997).
The fifth stage, according to Callahan, is the most difficult stage to
master. During the production of certain dagger types, for example,
subtype I D with a percussion, flake-scar finish this is the last
stage before the final retouch. If the fifth stage is conducted without
flaws, the following stages, i.e. grinding, parallel retouch and final retouch, will be completed fairly easily in comparison. The following
stages, with the grinding stages as an exception, demand a certain degree
of practical know-how, but they are not as demanding as the fifth stage.
40

Technology and the Division of Labour


In order to visualise and formalise Callahans information, the degree of practical know-how in relation to the production stages has
been plotted along an axis (Fig. 2:5). This axis can be interpreted as
representing a division of labour within an apprenticeship system, based
on selected individuals, for instance on certain age-based groupings. It
is not entirely unlikely that the flint-dagger technology was
institutionalised in this way, at least if the Late Neolithic communities
were sedentarily organised with more or less permanent houses forming
small hamlets, cattle, farming and access to large networks for the
exchange of prestigious objects. These kinds of institutionalised organisations would be based on the control of the recipes for action, the
practical know-how and the raw material. It is likely that the technology,
in itself, and not only the finished object, will mirror the order of society
and thereby the structure of power within the society. In such a scenario,
the organisation of a complex technology is metaphorically fundamental
for the reproduction of the society in general.
In Fig. 2:6, the degree of theoretical knowledge involved in each
production stage is visualised. Owing to the obvious difficulties involved in judgements of this character, it should be stressed that the
interpretation presented in the figure stems from my own experience
and it is likely that it differs in one way or another from the prehistoric
reality. However, it is possible to tie some typological elements in the
flint daggers to different recipes for action that are based mainly on
theoretical knowledge.
An example of this would be the pronounced handles on subtype-I
D and type-II daggers from the Danish islands in relation to the nonexistence of pronounced handles on the western, type-I subtypes. The
different handle constructions cannot be explained by the fact that some
handle variations would be significantly more difficult to produce. It
must instead be understood as a form of stylistic variation that probably
stemmed from a need to form the handles in different ways. In this
case, it is likely that the dagger typology has captured two real dagger
types that were created within the frameworks of regional and normative,
technological traditions.
Another example of how schemes involved in the production of
flint daggers were kept secret is certain differences involved in making
the pressure-flaked seams that are present on the handles of type-IV
daggers. The seams on the handles of the most prestigious, type-IV
41

Jan Apel
1 2 G

Apprentice/youth

Journeyman/young adult

7 6

Master/Old adult

Fig. 2:5. The seven production stages defined by Callahan (and a grinding stage = G),
graded according to the degree of practical know-how.

1 G

Learning by observation,
simple recipe

4 5

Moderate recipe

7 6

Complex recipe

Fig. 2:6. The seven production stages defined by Callahan (and a grinding stage = G),
graded according to the degree of theoretical knowledge.

daggers, such as the one from Hindsgavl on Fyn (National Museum in


Copenhagen) or the one from Skatelv in Smland (National Museum
in Stockholm), are meticulously formed according to specific norms
and with specific tools. The seams on ordinary type-IV daggers are made
with a rougher technique and method, much like those used to make
the seams on type-III daggers. These differences cannot be explained by
references to skill, since the seams on prestigious daggers do not entail a
significantly greater degree of skill than the seams on the ordinary daggers.
They do, however, demand that the knapper has access to a certain
schemes of action that includes information on which tools to use,
which angles to work from, which bodily positions to take, which
gestures to use, etc. It is tempting to consider the possibility that this
scheme of action was actually kept secret by a workshop, a family or a
community, in order to maintain exclusive rights to the daggers, as well
as the technology.
It can be concluded that the production of prestigious flint daggers
demanded great portions of both knowledge and know-how. It is
therefore likely that the reproduction of the flint-dagger technology
demanded some form of institutionalised apprenticeship system. Because
of the fact that such apprenticeship is based on the acquisition knowhow that can be achieved only through much practical training, it follows
that the dagger technology can only be reproduced over generations in
areas where there is an abundance of high-quality raw material. This
42

Technology and the Division of Labour


leads to the conclusion that the production took place in the close vicinity
of the raw material, i.e. the Limfjord area in Jutland, in the southern
parts of the Danish islands and in the south-western part of Skne.
Consequently, these areas are regarded as the core production areas of
Late Neolithic and Early Bronze Age flint daggers (Apel 2000a, 2000b).

43

Chapter 3

Apprenticeship as a Social
Institution

We speak a language we did not make; we use instruments we did


not invent; we invoke rights that we did not found; a treasury of
knowledge is transmitted to each generation that it did not gather
itself (Emile Durkheim).

In the previous chapter, my tentative preconception of the social


complexity in the Late Neolithic communities of southern Scandinavia
was presented. The use of certain words, for instance, the term
apprenticeship, implies a fairly stratified, chiefdom-like community,
in which craft specialisation was institutionalised (for definitions, see
Chapter 4). It was further suggested that the main function of the
apprenticeship system was to reproduce the knowledge and know-how
of dagger-making in order to guarantee that those who were supposed
to learn the craft were able to do so. It was argued, with references to
experiments, that the production process involved in making prestigious
flint daggers was very complex. The concentration of suitable raw
material in small regions in southern Scandinavia suggested that the
production areas were severely limited. From these conclusions, it was
inferred that the craft was most likely inherited and that the daggermakers were born into fixed social positions. It was suggested that a
hereditary principle was instrumental in the selection of suitable
apprentices and thereby served to keep the technology exclusive to certain
segments of society.
In the following chapters, I shall try to formalise this suggestion
into an ongoing discussion of the idea of the growth of social complexity in traditional societies. It will be related to theories using data based
on first-hand observations of human agency. Thus, the discussion will
rely on views of humans and human societies described in anthropo45

Jan Apel
logical and sociological texts. It will involve a discussion of social and
cultural reproduction that is essential to an archaeology that aims at
considering the meaningful aspects of prehistoric technologies in relation
to social structure. Some key terms used to shed light on the social
structure of the Late Neolithic societies in Scandinavia will be defined.
Certain issues that have been discussed in the complexity debate in
archaeology and anthropology will be recapitulated and related to the
tentative interpretations of the Late Neolithic societies in Scandinavia
presented above. Last, but not least, concepts such as technology, specialisation and division of labour will be discussed in an attempt to
understand their roles in the strategies of gaining and maintaining power
in a stratified society, strategies that, according to the interpretative framework used, led to the emergence and growth of institutionalised
inequality in Scandinavia during the Neolithic and Early Bronze Ages.
The theoretical stance that is proposed in this section originates from
a Durkheimian tradition seen chiefly in sociology and anthropology.
As we have seen earlier, the theories of Durkheim concerning the
reproduction of social facts inspired the development of the operativechain method in archaeology. This method is explicitly used in this
book, as well as in previous articles, to discover the complexity and
geographical distribution of the flint-dagger technology (Apel 2000a,
2000b, Callahan & Apel, ms). Durkheims view of social reproduction
was also instrumental in the development of Bourdieus general theory
of practice. This theory will be used below to elucidate how a technology
in itself, and not only its products, form a symbolic capital (see the
discussion on Bourdieu and his relation to the Durkheimian tradition
of sociology in Broady 1991:100 ff.). Thus, it seems appropriate to
start off by presenting the main ideas of this school of thought and to
follow its development towards the theories of human practice and
individual versus collective agency presented by Bourdieu (1977, 1999a).

Social facts and collective representations


Durkheim defined a social fact as a human act, thought or feeling that
exists outside the individual consciousness and is imposed on the individual by society (Durkheim 1978 [1895]:22). He maintained that
such collective phenomena were impossible to reduce to psychological
or physiological factors and had to be ascribed to society (Broady
1991:70). Durkheim thus criticised the intellectualism proposed by
46

Apprenticeship as a Social Institution


social evolutionists, such as Herbert Spencer and E. B. Tylor, who
regarded cultural reproduction as a process in which the individuals of a
social group made calculated and conscious choices in order to improve
the quality of life. In this sense, Durkheim was one of several social
scientists during the late 19th century who denied simple, uni-linear,
evolutionary explanations in favour of an anti-intellectualistic approach
emphasising the role of social factors in the reproduction of social values
and ideas.
The analogy that Spencer used in order to explain social evolution
was the growth and decline of the human body. He thereby came to
regard social evolution as an inevitable and unilinear process (Persson
1999:669). Spencers analogous treatment of organic and super-organic
(social) evolution can, according to Radcliffe-Brown (1952:8), be reduced
to the following propositions. (1) That both in the development of
forms of organic life and in the development of forms of human social
life there has been a process of diversification by which many different
forms of organic and social life have been developed out of a smaller
number of original forms. (2) That there has been a general trend of
development by which more complex forms of structure and
organisation (organic or social) have arisen from simpler forms. In
contrast to these ideas, Durkheim argued that social complexity would
occur in areas where for some reason there was a high moral density
varying with the process of the division of labour. The main factor that
promoted moral density and thereby also the degree of the division of
labour in a society was population density (Giddens 1978:27). In this
context, the degree of social complexity is defined in relation to the
number of formal institutions present within a defined society. Thus,
Complexity refers to that which is composed of many interrelated
parts (Price & Brown 1985:7). This definition actually echoes that of
Herbert Spencer, who regarded an increasing complexity as something
that is characterised by more parts and more relations between parts.
In 1910, inspired by Durkheims work, Lucien Lvi-Bruhl defined
collective representations as sentiments common to all members of a
social group, transmitted from generation to generation and imposed
on individuals from an external source, both preceding them and surviving them (Horton 1973:251). Such collective representations, in
the form of normative role castings, were seen as shaping the individual
during childhood and upbringing, because they were embedded in
47

Jan Apel
peoples habits, traditions and common ways of living and took the
forms of institutions, laws, moral values and ideologies (Durkheim 1978
[1895]:21 f.). A socialisation process resulted, accordingly, in feelings
of respect and a sense of duty towards the values of social norms. These
values and norms, that to the child appear as compelling, are through
the socialisation process, including upbringing, internalised into the
individual, i.e. they are integrated into the individual personality and
thereby become part of a persons self-identity (Guneriussen 1997:311).
The individual actor will, generally speaking, try to establish a balance
between himself and the community to which he belongs. Attitudes,
values and feelings among individual actors of a tightly knit group of
people will thereby, through time, tend to merge into a common
consciousness (Durkheim 1993 [1897]:86). Durkheim referred to such
embodied moral styles as ethos and, as we shall see, this term was later
used in the early works of Bourdieu as almost synonymous with habitus
(Broady & Palme 1986:12) . Bourdieu later confirmed that ethos, as
well as other classical concepts, such as eidos and hexis, all formed parts
in his habitus concept. The term habitus was initially used by medieval
scholars as synonymous with Aristotles term hexis, which means
posture or gesture (Bourdieu 1991:147). Apparently both Durkheim
and, and as we have seen, Mauss, also used the habitus concept in their
work. Durkheim used it to explain how the early Christian schools
were forced to use a pagan habitus in order to solve the problems that a
change of religion actually meant for the population (Bourdieu
1991:148).
Accordingly, the reproduction of society stemmed from the fact that
cognition was determined by social factors (Bloch 1989:1). In this
respect, the Durkheimian view of social reproduction has been seriously
challenged only by proponents of the idea that culture is the extrasomatic means of human adaptations to nature within the subject of
archaeology (Preucel & Hodder 1996:205). As we shall see, the fact
that Durkheim believed that social and cultural variability should be
assigned to social factors, and not be regarded as adjustments to natural
constraints, did not prevent him from making analogies to Darwins
evolutionary biology. However, in contrast to the evolutionary biologists, who at that time regarded change as a result of natural selection
on the individual level (Persson 1999:672), he was convinced that the
mechanisms of cultural change operated mainly on the collective level.
48

Apprenticeship as a Social Institution


The functionalistic aspects of Durkheims work receive only a peripheral treatment in this context. However, it deserves to be mentioned
that he argued that social institutions and society, or social integration,
were comparable with the different organs in a biological organism. It
was proposed that the functions of social institutions were to maintain
or enhance social integration, just as the function of an organ was to
maintain the organism. As a result of this conviction, Durkheim was
able to argue that the social sciences no longer needed to rely on the
subjective views of individual social actors in order to understand and
explain social phenomena, including the questions regarding social
production and reproduction which concern us here. Yet, as regards
these matters, I share the opinion, that the function of any social
phenomenon does not exists a priori but has to be revealed by empirical
studies (Merton 1967, Guneriussen 1997:83).

Religion, the first institution


Durkheim argued that fundamental categories of thought, such as space,
time, class and causality, had social origins. They derived from culture
and were regarded as relative categories, in the sense that they lacked
fixed, cross-cultural meaning (Morris 1987:132). It was maintained
that the rhythm of the social life forms the conception of time, while
the extension and appearance of the inhabited territory is the basis of
the categorisation of space. Through studies of traditional societies (the
Australian Aborigines, the Zuni and Sioux Indians and traditional
Chinese communities), it was argued that not only conceptions of time
and space, but also the systems used to classify the surrounding world
were cultural creations handed down from generation to generation
(Durkheim & Mauss 1963 [1903]). In The Division of Labour in Society
(1984 [1893]), Durkheim suggested that mechanical solidarity, i.e. the
form of social solidarity present in traditional, pre-literate societies (see
the definition of transegalitarian societies in Chapter 4), was based on a
common value system forming a collective consciousness. Collective
norms were maintained by repressive sanctions and the individuals shared
identical beliefs and sentiments. Such collective representations, or belief
systems, were always of a religious nature. Religion was therefore regarded
as a system of ideas by means of which individuals imagine the society
of which they are members (Durkheim 1995 [1912]:227). In short,
religion was seen as the way in which traditional societies reproduced an
49

Jan Apel
ideal social order; it was the worship of values created by society, or
rather society worshipping itself. Durkheim believed that he had
demonstrated that religion was a primitive form of social organisation
simply because totemism, as such, was inseparable from the clan-based
social organisation of the Australian Aborigines (Durkheim 1995
[1912]:169). The members of a single clan were united first and foremost
because they had the same name and shared the same emblem. Since
clan members were not necessarily consanguineous and were often
scattered throughout the tribal territory, this seemed to be a logical
conclusion. According to Durkheim, clan unity sprang from a collective
belief that all members had the same relation with the same categories
of things, and from practising the same rites in other words, from
the fact that they commune in the same totemic cult (Durkheim
1995:169). Since Durkheim argued that the role of religion in the
reproduction of the social order was gradually replaced by other
institutions, he had an evolutionary perspective on cultural change, in
the Darwinian sense of the word, that is (see the further discussion on
this subject in Chapter 4).
Thus, religious rituals were regarded as one way of transmitting
knowledge of the social order to the individuals and that orderly social
life among humans presupposed certain sentiments, shared by all
members, that controlled the behaviour of the individual in relation to
the collective. One of the important functions of rituals was therefore
to regulate, maintain and hand down such sentiments from one
generation to another. This functionalistic approach to religion and ritual
may appear blatant and one-dimensional to modern scholars but, as
Morris (1987:111) has pointed out, throughout history there are
examples of how the social role of religion has been acknowledged by
at least certain segments of the population. The following quotation
may indicate that the ruling aristocracy in ancient China was aware that
rituals served social functions: Ceremonies are the bond that hold the
multitudes together, and if the bond be removed, those multitudes fall
into confusion (Morris 1987:111; translation of a quotation from the
ancient Chinese Book of Rites from the third century B.C.). As society
became more and more complex and thereby gradually moved from
mechanical to organic solidarity, religion lost its importance as a vehicle
of social and cultural reproduction in favour of other kinds of institutions
(see Chapter 4). It is important to remember that Durkheim, as op50

Apprenticeship as a Social Institution


posed to almost all other scholars of his time, saw no fundamental
difference between primitive and modern ways of thought. Consequently he thought that the moment at which religion is born is the
moment at which the possibility of all higher forms of thought, including
science, is also born (Horton 1973:260). One of the deficiencies in
Durkheims thinking was that he never considered the possibility that
religion could be seen as a form of ideology by which groups or individuals dominated others (Giddens 1978:103 f.). In my opinion,
Durkheims ideas have to be complemented in some ways and such a
complementary perspective will indeed be discussed below. However,
the importance of collective, as opposed to individual, intentionality in
social reproduction deserves to be stressed, since such aspects tend to be
neglected in favour of individual approaches in the archaeological
discussions of today. As I see it, the complicated production process
involved in the making of flint daggers seems to suggest that it is
especially important to consider the collective aspects, as they are reflected in the apprenticeship that secured the reproduction of the
technology.
The idea that cognitive systems are socially determined had a tremendous influence on the sociology and anthropology of the 20th century. One could perhaps draw a parallel to the impact that the idea of
the subconscious mind had on psychology during the same period.
Although Sigmund Freud has been severely criticised, few have argued
that the idea of the subconscious is irrelevant to psychology. In a similar
way, Durkheims idea of collective representations, and the view of society as something that encompasses the individual, both in time and in
space, have permeated the greater part of the anthropological research
carried out since the turn of the century (see Bloch 1989:3). However,
while Durkheim, like most social scientists of his time, was a firm
believer in a reality outside the human senses, few of those who followed
in his footsteps shared his realism. This state of affairs did not offer a
great problem since the new theoretical stance did not challenge the
fundamental subject that concerns us here, namely that cognition has a
social origin. In this context, it is maintained that the usefulness of
social theories is not necessarily dependent on highly generalised questions of epistemology (Giddens 1984:xviii). No science can await a
clear-cut solution to epistemological problems. However, I shall discuss a minor part of these problems in relation to the social theories put
51

Jan Apel
forward, because the idea of social facts is intimately connected with
such questions.

Durkheimian approaches to cultural


reproduction
Durkheims ideas on the reproduction of society have been criticised
from two perspectives in the social sciences (Guneriussen 1997:65).
Proponents of the utilitarian-individualistic tradition have criticised the
holistic approach, because it implies that the functions of social institutions in society are analogous to functions of the different organs in the
human body. According to this tradition, a social analysis, on the
contrary, always has to be based on the knowledge, preferences and acts
of individuals.
From a hermeneutic perspective Durkheims objectivism manifested
foremost by his tendency to regard social realities as an objective and
causal phenomenon that determine the behaviour of individuals, has
been criticised (Guneriussen 1997:65). According to this tradition, the
proposed outside perspective, which in this context should be seen as
the idea that the functions of social phenomena cannot be recognised
on an individual level, is regarded as a fallacy. As we shall see, these two
critical perspectives have been embraced by the post-processual
archaeologists and are often combined in the critique of traditional and
processual archaeology in general. In the following pages traditional
and processual archaeology are merged as conventional archaeology
(Hegardt 1997) . Yet this does not mean that I share the conviction that
there is a fundamental difference between conventional archaeology and
post-processual archaeology. The term conventional archaeology is used
only because it is more convenient.
Later scholars tended to regard social facts not only as relative, but
also as individual creations in themselves, simply because of the fact
that an understanding of perception in itself is theory-dependent. The
realism and positivism of the late 19th-century social sciences were by a
linguistic turn replaced by hermeneutic phenomenology and language
philosophy (Giddens 1984:xxii). This new line of reasoning was put
forward by, among others, Martin Heidegger in arguing that there is a
unity between being and thinking and that the materialisation of the
world is an individual matter and is not connected with an external
world. Consequently, concepts such as time and space were seen as
52

Apprenticeship as a Social Institution


constituted by human perception and not as universal phenomena
existing outside the senses (Heidegger 1972 [1927]). In line with this
reasoning, the reality, as perceived by an individual in a cultural system
of thought, is a social construction and this conclusion is, as we have
seen, consistent with Durkheims view, although it is based on entirely
different premises. In the former scenario, the individual does not receive
social facts from an outside source, as Durkheim proposed, but creates
his or her world-view through the individual senses.
In an important article on the social determination of truth, Lukes
states that the urge to see the rules specifying what counts as true and/
or what counts as valid reasoning as themselves relative to particular
groups, cultures, or communities is a temptation that has attracted
thinkers of several disciplines. It has found its way into the sociology of
knowledge (Whinch), linguistics (Whorf ), social anthropology (LvyBruhl), the history and philosophy of science (Kuhn) and Marxist
research (Lukes 1973a:231). Levy-Bruhl adhered to the relativist position
when he stated that individuals in traditional societies live, think, feel,
move and act in a world which at a number of points does not coincide
with ours (Levy-Bruhl 1922:47). Franz Boas pointed out that a simple
symbol, such as a triangle with the tip pointing upwards, is regarded by
different traditional communities as a symbol of a tent, a mountain, a
cloud, a bears paw and a fish-tail (Boas 1955:120 ff.). Boas used
empirical examples such as this one to demonstrated his deeply rooted
cultural relativism. Benjamin Lee Whorf maintained the relativistic idea
when he stated that all observers are not led by the same physical evidence to the same picture of the universe, unless their linguistic backgrounds are similar, or can in some way be calibrated (Whorf 1956).
According to Whorf, it is language that determines the human perception of the world. Consequently, a Hopi Indian will experience and
sense the world in a completely different way than, for instance, an
Englishman, because their languages are structured in different ways
(Asplund 1970:77). Johan Asplund elaborate further on this by referring to the biologist Jacob von Uexkll. According to Uexkll, the
ways in which any organism experiences the world are dependent on its
psycho-physiological equipment. For instance, certain one-celled organisms will experience nothing else of its surroundings than the stimuli
which trigger the only type of reaction that they are capable of, namely
flight (Asplund 1970:77). Personally, I feel that this is a weak analogy,
53

Jan Apel
since all living humans share the same biological faculties of perception.
The relativistic, theoretical stance can be further illustrated by the
following statement, accompanied by a clarification by Bloch: If lions
could speak we could not understand them (Wittgenstein). In other
words, that communications with creatures with a fundamentally
different system of ideas and life is not possible, and surely people with
a different concept of time would in this respect be like lions, since
everyone agrees about the particularly fundamental nature of this proposition. (Bloch 1989:8). Richard Rorty has gone very far in this respect
and sees everything as the result of linguistic principles (Rorty 1989).
He maintains that not only the reality around us but also our personalities and our roles in society are constituted by linguistic praxis. According to this view, humans are shaped and constituted by the same linguistic praxis that creates and reproduces the external reality
(Guneriussen 1997:206).
As regards anthropology, the linguistic turn eventually divided the
field into two groups that both adhered to cultural relativism but based
their assumptions on two fundamentally different principles. On the
one hand the social anthropologists, functionalists and structuralists such
as Spencer, Durkheim, Radcliffe-Brown, Levi-Strauss and Edmund
Leach, emphasised how long-lived, external structures formed human
cognition. On the other hand, cultural anthropologists, from Boas to
Clifford Geertz, maintained that culture is something that is always
negotiable and is created on an individual level. In short, social aspects
are constructed and studied from an outside perspective, while cultural
aspects require an inside perspective: Culture is the fabric of meaning
in terms of which human beings interpret their experience and guide
their action; social structure is the form that action takes (Geertz
1957:33 f.). Perhaps it is not entirely unfair to claim that cultural
anthropologists today are empirically orientated and will have a hard
time seeing the forest for all the trees, while social anthropologists
may be seen as having a mechanical view of society that neglects
individuality. Personally, I prefer the outside view, although I accept
that it was constructed by an anthropologist and is not readily available
for study. As I see it, in line with Durkheim & Mauss (1963) the world
has to be classified in order to be understood and it is not something
that will reveal itself through empirical studies. However, as we shall
see, the cognitive faculties involved in this process cannot be solely
54

Apprenticeship as a Social Institution


assigned to social factors in the way that Durkheim imagined.
In the prevailing strands of the post-processual archaeology, the idea
that each cultural system of thought is unique is alive and prospering.
Several authors have maintained that the past is a foreign country
which cannot be understood, since the social facts that gave meaning to
the world as its people knew it have long since been forgotten (Lowenthal
1985, Eriksen 1993, Solli 1996, Olsen 1997, Burstrm 1999).
Accordingly, it has been suggested that the acceptance of the culturally
relativistic idea in archaeology can make us realise that the culturally
foreign once shared our own territory and thus make it easier for us to
accept cultural differences: It stirs the imagination to think that in the
same well-known place which we call home and where we live our
daily lives, other people at another time were living in a world that was
essentially different: They filled their lives and their surroundings with
other meanings than ours. We share the same place yet we live in different
worlds (Burstrm 1999:25). Consequently the notion of the uniqueness of prehistoric cultures may be used as a political tool to gain
acceptance for other cultures today, provided, of course, that the correct
ethics are followed. Mats Burstrm hopes that the will to regard prehistoric cultures as relative in this sense will lead to a greater interest in
and understanding of cultural pluralism in the past, as well as in the
present (Burstrm 1999:25).
It has also been suggested that archaeologists should face up to the
consequences of Heideggers idea of how perception is constituted, i.e.
that there is a unity between being and thinking which forms the worldview of an individual (Karlsson 1998). The implication of this is, at
least to me, that not only archaeologists, and not only all scientists and
researchers, but also all humans would have to face up to the consequences of this world-view. The cornerstone of all empirical sciences,
and also of human conduct in day-to-day life, is the acceptance of our
perception of the world, as we receive it through our senses. If we want
to go beyond that, we have to remember that all understanding of the
mystery of perception and thinking is in itself theory-dependent and
that the answer to this question presupposes the belief in a primary
philosophy, such as Heideggers. If, on the other hand, we choose to
maintain that archaeology is an empirical science, this critique is, as I see
it, irrelevant. Basically, then, my view on this subject is that, even if
humans do profit from a contemplative approach to the surrounding
55

Jan Apel
world, archaeologists can safely assume that it is enough to define and
construct the prehistoric object through our senses.

Post-processual critique of holistic approaches


From these examples, it may seem as if the culturally relativistic idea
has been embraced by post-processual archaeologists in general. However, an emphasis on the individual as the prime mover of social change
has in turn led to the rejection of the holistic proposition on social
reproduction made by Durkheim and his followers. Subsequently, the
ideas of social facts as something that encompass the individual, and of
social structures in which individuality is trapped and from which the
individual receives important cognitive abilities, have been severely
criticised (Tilley 1982, Hegardt 1997:225 ff.). This may appear strange,
since the theory of social facts, as we have seen earlier, was the cornerstone in the development of the culturally relativistic approach in sociology and anthropology (and subsequently also in archaeology).
Before continuing, it must be stressed that there are internal discrepancies between Durkheims theoretical framework, based on holism and
positivism and presented in his methodological writings, on the one
hand, and the views put forward in his more empirical and theoretical
writings, on the other. It has been pointed out that, especially in his
work after 1900, he positioned himself somewhere between the
positivistic search for causality and a more contextual way of working
(Guneriussen 1997:310). Anyone who reads, for instance The Elementary Forms of Religious Life, which was written in 1912, will no doubt
realise this. It is obvious that this work is very much based on the
personal experiences of the world that Durkheim had acquired rather
than on the causal relationships between objective facts. Robert A. Nisbet
has suggested that Durkheim did not receive his controlling idea of the
social aspects of suicide from a preliminary examination of the vital
registers of Europe, any more than Darwin got the idea of natural
selection from his observations during the voyage of the Beagle. Instead,
Nisbet argued that the idea of the social significance of the suicide was
in his mind well before he collected his data on the subject. In fact, it is
even suggested that Durkheim might have formulated his main idea
from an intuition born of personal loneliness and marginality (Nisbet
1963, Asplund 1970:57). In her introduction to the new English
translation, Karen Fields has indeed pointed out that The Elementary
56

Apprenticeship as a Social Institution


Forms of Religious Life can be read through many theoretical spectacles:
American postmodernist theorizers of discursive practices and
representations will recognize through Formes the Durkheimian pedigree
of Michel Foucault. Psychologists will repeatedly glimpse old and notso-old ways of thinking about phenomena that the scientific study of
memory, identity, language, and intelligence must be able to account
for. Philosophers will find old problems interestingly tackled, if not
necessarily solved (Fields 1995: xxiii). There are even sociologists who
regard Durkheim as a forerunner of the post-modern project in general
(Mestrovic 1992:7 ff.). These remarks go to show that to regard
Durkheim simply as a holistic positivist is to make things much too
easy. However, since post-processual critique of Durkheim has taken
this form, these are the sides of Durkheim that will be discussed in the
following pages.
The post-processual critique of Durkheim obviously stems from
the fact that he had a holistic view of society, in which the cornerstone
was his belief that it was possible to explain human agency with references
to groups of people or to society as a whole. This way of doing science
necessitates definitions of objects, such as social facts, that are not
dependent on any individual subjects own interpretation of the external
world. In other words, it requires an outside perspective considered by
post-modernists to be a fallacy. Post-processual archaeologists argue that
this objectivism tends to deny the active human subject the
individual. They tend to regard the individual as an autonomous and
complete human being who is not shaped by the social relations in
which he or she participates. In the social sciences, the conflict between
these two ideas represents one of the most fundamental differences in
the field and goes a long way back. Liberal-minded, social scientists,
from Adam Smith to Max Weber, had an individualistic view of society,
while, for instance, Durkheim and Marx, adhered to holistic approaches.
In this context, the latter term refers to the view that social phenomena
are to be understood from studies of the collective rather than from
studies of the individuals within the collective. It has been argued that
conventional archaeologists (as opposed to post-processual archaeologists) have over-emphasised the importance of structure in relation to
agency in the reproduction of societies and have thereby neglected the
possibility that the free individual can change his or her social situation
(Hegardt 1997:235). Curiously, Christopher Tilley (1982), and Johan
57

Jan Apel
Hegardt (1997), have argued that Durkheims functional and positivistic
approach made him a forerunner of the processual school of thought in
archaeology. However, as I see it, Durkheims ideas of social facts and
the way in which they function as reproducers of society have very little
in common with, for instance, Leslie Whites or Lewis Binfords view
of culture as mans extra-somatic means of adaptation to nature. On
the contrary, it has been pointed out that these two ways of regarding
culture as determined by nature or by society are two traditions of
thought which exclude each other (Preucel & Hodder 1996). However,
as we shall see below, Binfords view on the emergency of inequality
does have certain aspects in common with Durkheims theory explaining
the principles behind the division of labour in society.
While Tilley claimed that Durkheims holistic view of society influenced processual archaeology, he was also of the opinion that Weber
greatly influenced post-processual archaeologists. Hodder has also
maintained that Webers ideas were neglected by conventional archaeologists and that the latters emphasis on subjective meaning and historical approaches in interpreting social and cultural reproduction
deserved to be introduced into the archaeological-theory debate (Hodder
1986:81 ff.). One of the main ingredients in Webers theory of practice
was that it is not possible to understand the complex set of relations
that make up society without a thorough knowledge of the subjective
aspects of the members of society in their internal affairs (Campbell
1986:203). For Weber, it was vitally important that the positivist search
for causal explanations should be combined with an understanding of
the meaningful individual act, i.e. the meaning that the social act had to
the individual actor (Campbell 1986:203). As we shall see below, Webers
hostility towards holistic, over-individualistic ideas concerning society
can perhaps be traced to his upbringing and political views. Weber observed the world from the perspective of an ontological individualist,
i.e. he regarded everything that exists as phenomena of an individual
nature and he regarded collective phenomena, such as society, group,
class as constructions in the minds of individuals (Andersson & Andersson
1991:7). Thus, Weber maintained that social facts had to be preceded
by an idea that had been formed by an individual agent. Or, as Edmund
Leach has put it, the idea of, for instance, a city must precede the actual
forming of the earliest cities in the Near East (Leach 1973).
This conflict between processual and post-processual archaeologists
58

Apprenticeship as a Social Institution


that Tilley traced to a conflict between holistic and individualistic
approaches in the social sciences in general echoes an earlier debate on a
similar subject that took place in France during the 1930s. This debate
deserves to be mentioned in this context because the political implication
of the post-processual embracement of Webers ontological
individualism, ironically, seems to have been neglected. In 1935, the
Durkheimian tradition was challenged by sociologists with an individualistic view of social change (Broady 1991: 89). Thus, Raymond Aron,
using references to Weber, accused Durkheim and his followers of being
dogmatic positivists, in the sense that they strove to imitate the methods
of the natural sciences. Since Aron had right-wing political sympathies,
while the Durkheimians tended to harbour left-wing sympathies, one
of the consequences of the conflict was that Webers sociology became
connected with the right-wing political movement in France during
this time (Broady 1991:90).
In a similar way, it is perhaps possible to regard the newly found
interest in individualistic approaches in post-processual archaeology as a
result of the current political situation, in which the practical use of old
collectivist ideologies, such as Marxism, has diminished. Personally, I
am convinced, on the basis of the discussion above and on personal
reflections, that ontological individualism has to be moderated in
discussing the reproduction of societies. The idea that free agents can
negotiate their positions in society through material culture greatly needs
to be complemented with a theory that emphasises the importance of
structure as a framework for individual agency. Accordingly, to my mind
the makers of the Late Neolithic flint daggers of Scandinavia did not
choose to become flint-knappers they were destined to be flintknappers.
If we do not realise how much tradition and structure determine us
as human social beings but believe in references to undefined and vague
terms such as natural talent, we will not be fully prepared to face up to
and understand the modern society around us. I believe that many
important aspects of individual human behaviour can be understood as
results of the ways in which the social structure shapes the individual
actor. As I see it, the importance of structure, as a force that encompasses
the individual in the shaping of the members of society, is apparent in
Sweden today and not least in my own surroundings. It is impossible
for me to disregard the fact that dispositions to different forms of social
59

Jan Apel
abilities are inherited within different social segments and that such
abilities are used in social strategies. This conclusion is even more striking if we consider the fact that these problematic tendencies have been
actively confronted and opposed within the Swedish educational system
during the last 40 years. It is difficult for any individual to actively
change traditional patterns and this may be due to the fact that important aspects of our social abilities are reproduced at an early age within
the family unit (Bourdieu 1999a:120).
Consequently, one of the important theoretical aims of this thesis is
to demonstrate the importance of the social structure in the reproduction of society. My field of inquiry is the flint-dagger technology of the
Late Neolithic period in Scandinavia and I shall try to uncover its social
facts in a Durkheimian sense. I want to discuss the principles which
determined who should be involved in the production of these daggers
and who should be excluded. In my opinion, it is not sufficient to
regard technological talent (an individualistic view) as something that
individuals are born with, i.e. a form of natural talent. On the contrary,
technological talent, just like other social facts, is, according to this
view, mainly reproduced within social contexts. If not, we shall inevitably
maintain a charismatic ideology in which the importance of individual
agency is allowed to completely overshadow the importance of structure,
as regards the social positions of individuals (Featherstone 1994:75). I
shall return to and elaborate on this crucial subject in discussing
Bourdieus theory of practice in Chapter 4.

Collective versus individual intentionality


The political history of Europe during the 20th century, with the rise
and fall of totalitarian regimes seemingly without any respect for humanity, seems to have resulted in a reaction in which totality and structure tend to be feared, neglected and subsequently avoided or dismissed.
However, I maintain that the role of the individual actor as the prime
mover of social reproduction tends to be exaggerated in present-day
archaeological discussions and that the importance of structure as a vital
ingredient in this process needs to be returned to and investigated. John
R. Searle (1997:38) have pointed out that many animal species, and
especially humans, are able to share a form of collective intentionality.
In this context, I would like to use parts of his discussion to clarify
what I understand as the relationship between Webers ontological in60

Apprenticeship as a Social Institution


dividualism and Durkheims holism. According to Searle, collective intentions are not only constituted by the ability to perform co-operative
tasks but should also be regarded as beliefs, wishes and intentions which
are shared by a collective. There are obvious similarities between this
definition and Durkheims definition of social facts. An important matter
in this context is to investigate what kind of relationship exists between
collective intentionality, on the one hand, and individual intentionality, on the other. Most attempts to answer this question has reduced
collective intentionality to individual intentionality (Searle 1997:38),
i.e. different forms of the kind of ontological individualism proposed
by Weber for instance. Yet Searle maintains that there are reasons for
questioning whether the reduction of collective intentionality to the
individual actor is the only possible solution. In studying humans in
action, individual intentions seem to appear only within frameworks
of collective intentions or as Durkheim would have put it, social facts.
The individual intentions observed in war, sport, politics and religion
always tend to be rooted in structures formed by collective intentions.
As we shall see, this does not mean that individual actors, through their
strategic actions, are unable to change structures or to take advantage of
situations within the structures and thereby gain prestige, power, etc.
On the contrary, such aggrandizer strategies have been studied extensively
in the anthropological literature (see Chapter 4) and one of the reasons
for acknowledging structures is that knowledge of them makes it easier
to recognise and understand this kind of agency. Searle proceeds to
explain why the notion that collective intentionality can be reduced to
individual intentionality has become so popular. The argument that,
implicitly or explicitly, has been put forward is that, because all
intentionality, as such, exists in the brains of individual actors, it has
been erroneously concluded that intentionality can only be addressed
to the individuals in whose brains it exists (Searle 1997:40). In
opposition to this kind of argument, Searle simply suggests that collective
intentionality guides many of the joint efforts in everyday life. The
decision that two people make regarding taking a walk in the park
together can from these points of view be formulated in two
fundamentally different ways (Table 3:1).
In sociology, the theoretical discussion from the 1930s up to the
post-modern debate has been focused on the problematic relationship
between structure and agency and between collective and individual
61

Jan Apel
Type of intentionality

Person 1

Person 2

Individual intentionality

I intend, and I think that you intend, to

I intend, and I think that you intend, to

take a walk in the park

take a walk in the park

We intend to take a walk in the park

We intend to take a walk in the park

Collective intentionality

Table 3.1. An example of how the ideas of individual and collective intentionality may
differ in a concrete situation.

agency. In Talcott Parsons The Structure of Social Action Durkheims


collectivism and Webers individualism are combined (Campbell
1986:226). Parsons acknowledged Weber by giving the individual act
important meaning, while on the other hand, he also recognised
Durkheim in that he regarded each act as taking place in a situation in
which the possibilities of individual control are severely limited
(Campbell 1986:227). In recent years, the theory of structuration
presented by Giddens (1994), the theoretical contribution of Fredrik
Barth (1981) and the theory of human practice presented by Bourdieu
(1977, 1999a) have tried to reach beyond a simple relationship between
individual actors and structure in the process of maintaining and
reproducing human societies. It also deserves to be mentioned that, in
philosophy, a similar development has arisen with the urge to try to
reach beyond simple dichotomies, such as objectivism and relativism
(Bernstein 1987; for a discussion in relation to archaeological problems, see Sundstrm, in preparation).
Giddens (1984) has presented an eclectic theory in which the importance of the duality of structure is stressed. Structuration theory
proposes that orthodox theories regarding social structure have failed to
recognise the way in which the daily affairs of individuals also incorporate and reproduce the overall institutions in societies (Giddens 1984:19).
Giddens suggests that most daily practices on the individual level are
not directly motivated but should be seen as routinized practices:
Routine, psychologically linked to the minimizing of unconscious
sources of anxiety, is the predominant form of day-to-day social
activities (Giddens 1984:282). Accordingly, on the one hand, routinized
practices serve as a means of securing social continuity and, on the other,
the enactment of day-to-day activities in the form of routines gives the
individual a sense of ontological security. In this way, structure and
individual agency are seen as closely linked concepts in the reproduction
of a social order (Giddens 1984: 60 ff ). Another important aspect of
62

Apprenticeship as a Social Institution


the structuration theory is that it recognises the simple fact that any
mechanism of social organisation or social reproduction acknowledged
by social analysts can also be recognised by lay actors and subsequently
be used in individual social strategies (Giddens 1984:284).
According to Barth (1981), social institutions will provide role castings that limit the possibilities of each agents action, but within these
given frames, individuals will try to maximise their capital through
different manoeuvres. However, sometimes these actions will trigger a
backlash, in the form of unintended consequences of action, since many
individual actions made in order to maximise the outcome of each
individual, will devalue the profit and, as a consequence, a change in
structure will appear. Barth thereby stresses, in contrast to the
Durkheimians, the importance of individual, entrepreneur, behaviour
as a means of changing structure. An entrepreneur is, in this context,
regarded as an individual that through innovation seeks to maximise
profit. The entrepreneur will, by being involved in multiple transactions, be able to make commensurable some forms of value which
were previously not directly connected. This may be so because the big
potentialities for profit appear where there is a disparity in value between
two kinds of goods (Barth 1981:56). Barth is obviously taken by the
idea of the thick description proposed by empirically orientated,
cultural anthropologists, such as Clifford Geertz (see below). However,
he does accept structuralistic ideas in that he recognises how empirical
social patterns and cultural premises (equivalent to social structure)
limit individual agency (Barth 1981:147).
Through his definitions of terms such as capital, field and habitus,
Bourdieu has tried to create categories that enable him to study objective relations within defined contexts, instead of studying substantial
realities such as individual actors and the collective (Bourdieu 1999a:7).
Consequently, a field is defined as a system of relations between different positions and a certain habitus receives meaning and effects only in
relation to the shared habitus of people in other groups (Broady 1991:
462). In many ways, Bourdieus sociology can be seen as a continuation
of the Durkheimian tradition. Yet, according to Donald Broady
(1991:419 f.), Bourdieus sociology differs in three important aspects
from orthodox Durkheimian sociology: (1) Bourdieu does not adhere
to the evolutionist tendencies in Durkheimian sociology. (2) With the
habitus concept, in a similar way to Giddens, Bourdieu has reintro63

Jan Apel
duced the importance of the active individual that Durkheim denied in
regarding collective representations as springing from society. (3)
Through the field concept, Bourdieu can delimit smaller social units.
This allows him to regard the social world as more differentiated than
Durkheim, through his concepts, was able to do. I shall return to
Bourdieus theory of practice in the next chapter.
An alternative view of the relationship between structure and agency
which I find appealing has been put forward by Maurice Bloch (1989). It
will be recapitulated below. However, before proceeding any further, I
would like to suggest, in line with the general aim of this chapter, that the
differences in tradition and upbringing, i.e. structure, might have guided
Weber and Durkheim to opposed views on cultural reproduction.
Webers father was a member of a family that had prospered in the
German linen industry. He was a lawyer and a successful, liberal-minded,
nationalistic politician effective in official circumstances and as a
father very authoritative. His mother was interested in arts and culture,
was strictly religious and concerned herself with charity, something that
was despised by her husband. On the emotional level, Weber took his
mothers side, even though he did not support her religious views. He
came to following his fathers footsteps in that he studied the law and
showed an interest in public matters, but he also followed his mother
in that he cherished cultural refinement and spirituality (Campbell
1986:205). As a political writer, he was a defender of bourgeois ideals
and much of his research can be regarded as an indirect critique of
Marxism (Andersson & Andersson 1991:7) and other holistic traditions
in the social sciences.
Durkheim was the son of a rabbi and grew up in an orthodox Jewish
family in a long-established Jewish community in Alsace-Lorraine. He
was himself intended to become a rabbi and studied for some time in a
rabbinical school until, while still a school boy, he decided not to follow
the family tradition (Lukes 1973b:39). His background marked him
with several lasting traits: Scorn for the inclination to conceal effort,
disdain for success unachieved by effort, horror for everything that is
not positively grounded: the life of the individual within the framework
of the group, truths through their rationally established implications,
conduct by its moral regulation (Davy 1960:15, Lukes 1973b:40).
Lukes further writes: From the time of his childhood, he retained an
exacting sense of duty and a serious, indeed austere, view of life; he
64

Apprenticeship as a Social Institution


could never experience pleasure without a sense of remorse. During his
time as a student at the cole Normale Suprieure in Paris, he was never
concerned with philosophies which did not have a political and social
application and, on the other hand he never engaged in politics without
a sound philosophical basis (Lukes 1973b:46).
This may be dismissed as quasi-psychology, but I cannot escape the
feeling that the social backgrounds of Weber and Durkheim in some
parts shaped their views on society and the causes of social reproduction
and change. The liberal-minded Weber, who adhered to ontological
individualism, certainly received some of his opinions from his family.
It is also likely that Durkheim, being brought up in an orthodox Jewish
milieu, must have felt the influence of professed belief (structure) upon
day-to-day conduct very strongly (Giddens 1978:80). In a sense then,
the lives of Weber and Durkheim can serve as a demonstration of the
importance of role castings in the way in which we perceive the world.
Yet it should also be recognised that Durkheims decision to abandon
the rabbinical school was an individual act that must to some degree
have been in conflict with a social structure.
This subject obviously touches upon the core of Bourdieus sociological work. Upbringing will form the habitus of individuals in ways
that will make them develop dispositions especially suitable for certain
practices. Donald Broady has pointed out that the power struggle between the so-called epistemological fraction and the existentialists tradition in French academic circles during the 20th century may partly be
explained by differences in the backgrounds of the individuals involved
(Broady 1991:47 ff.). In an article published in English in 1966, Sociology
and Philosophy in France since 1954: Death and Resurrection of a
Philosophy without a Subject, Bourdieu and Jean-Claude Passeron asked
themselves whether a sociological explanation could be put forward
that would clarify why the scientific works of scholars of the epistemological fraction, such as Bachelard, Piaget, Guroult, Canguilhem and
Vuillemin, differed from existentialists such as Sartre and Camus when
they all had virtually the same education. According to Bourdieu and
Passeron there was no other explanation of this fact than their social
and geographical background (Broady 1991:47, note 77). The
proponents of the epistemological fraction generally had their roots in
the working class or the lower middle class and came from the provinces,
while the existentialists tended to be born into the intellectual world of
65

Jan Apel
Paris. Apparently there is a connection between a form of socially created ethos and a form of intellectual moral that is, in the case of the
former group, demonstrated in the urge to re-instate science and scientific
research (Broady 1991:47, note 77).
A social situation of a similar kind appeared in archaeology during
the 1960s. It has been pointed out that the emerging processual school
in American archaeology during the early 1960s was developed in a,
middle class, mid-western context (University of Chicago) far away from
the prestigious universities of the east coast (Trigger 1993). It is also a
fact that its foremost proponent, Lewis Binford, with his claims to a
scientific, anti-authoritative, positivistic archaeology, never received a
chair in any of the prestigious universities. In fact, whatever deficiencies
the early processual archaeology may have had (and they have all been
severely scrutinised during the archaeological debate over the last 20
years), its aims were to develop objective methods that would work
independently of the individual researchers personal positions, views
and history. It is clear that the claims of such a research strategy include
the ideas that research should be valued according to the accuracy of the
results and not by the reputation of the individual scholar.

Realism or relativism
Lukes has pointed out that the urge to regard what counts as truth or
valid reasoning as relative to cultural systems of thought has its roots in
assumptions as to how reality is constituted (Lukes 1973a:231, 235
f.). For example, it has been argued that since human perception, and
therefore also the human understanding of the world, is theorydependent, there is no such thing as an theory-independent, external
world outside the senses (Lukes 1973a:235). But just because ones
understanding of how human perception is constituted must be theorydependent, since humans cannot be positive on anything outside their
senses, it does not automatically follow that there are no theory-independent objects of perception and understanding (Lukes 1973a:236).
In order to establish this realism, Lukes is obliged to present social facts
that can be regarded as universal. Thus, he continues to illustrate his
critique with a linguistic example in which the main argument is that
the existence of a common reality is a necessary precondition in order
to learn and understand foreign languages something that we know
is possible. The argument is concluded with the proposition that if
66

Apprenticeship as a Social Institution


G has a language in which it expresses its beliefs, it must, minimally,
possess criteria of truth (as corresponds to a common and independent
reality) and logic which are not and cannot be context-dependent
(Lukes 1973a:239). Lukes also contributes more specifically to the
discussion on social reproduction and social change. According to him,
it is only by assuming that there are non-relative means of identifying
differences between collective representations, on the one hand, and
social realities, on the other, that it is possible to see whether belief
systems prevent or promote social change (Lukes 1973a:243).
The point that Lukes is trying to get at has been developed further
in language philosophy. It has been suggested that all representations
present in human languages refer to two forms of facts: brute and
institutionalised facts. The distinction between brute and institutionalised (or social) facts is a distinction between facts that are independent
of our existence, as opposed to facts that are dependent upon our
existence (Searle 1997:163). In any given language, representations in a
sentence that refer to brute facts refer to something that is not contextdependent. Any human being with a language will be able to make a
representation of his or her perception of a brute fact and it will be
understood by other human beings, provided that they make an effort
to learn the language in question. According to this view, for instance,
during a game of football, an observer from any culture will be able to
see that people are running around on a flat surface chasing and kicking
a round, rolling object and will in turn make representations of these
brute facts in his or her own language. In this respect, brute facts are
non-relative in the sense that Lukes claimed above. Institutionalised facts,
on the other hand, are context-dependent facts. In order to understand
their meaning, the image given by perception is not enough. To be able
to classify and use them properly, the individual in a culture has in one
way or another to receive knowledge from an outside source. When it
comes to understanding the rules in football, for instance that a goal is
scored when the ball ends up behind the white line that runs between
the two posts, the information required has to be learned in an explicit
manner. In most circumstances, however, the knowledge of the meaning
behind institutionalised facts is not obtained in a conscious way. We
receive them, as Mauss (1979) demonstrated, from an early age, by
observing and imitating people that we trust and admire. Searle refers
to such unconscious knowledge as the background. The background is
67

Jan Apel
simply the frame of references that humans receive from an outside
source. As we shall see, Bourdieus use of the term habitus is to an
important degree equivalent to Searles use of the word background. It
is important to note that Searles discussion does not verify the actual
presence of an external reality outside our senses. It does, however, demonstrate that all languages, and thereby also the representations that are
made in them, presuppose the presence of an external reality.
A critique of relativism and anti-realism is often dismissed by proponents of these schools of thought simply because it comes from the
outside. The urge to view and study in a distant perspective, the
cornerstone of all positivist/modernist research, is regarded as useless,
since it cannot account for the personal opinions of individual actors. If
cognition is a relative affair, every culture and field of enquiry will have
its own standards, as regards truth and valid reasoning. Critique from
an outside source will appear strange and of lesser value, because it is
not operating within the proper discourse. Post-modern proponents
will therefore tend to regard all systems based on the principles of the
Enlightenment as oppressive (Lyotard 1984). In line with this, a project
aiming at creating universal standards, such as positivism, is considered
to be dogmatic, because it stipulate rules for what is true and what is
not (Hegardt 1997:193). However, post-modernist claims to the discovery of real historical discontinuity are not unproblematic (Featherstone 1997:17 f.). The post-modern critique of universality and totality, the post-modern feeling of being beyond traditional, normative
and modernistic class analysis and the emphasis on discontinuity, written
texts, paradoxes, ironies and the critique of the Enlightenment and the
modernists way of thought may just not be the end of history after all.
On the contrary, some suggests that the post-modern project may be a
necessary break with the dominance of 19th-century, orthodox social
theories but that the relevance of large-scale studies of social change will
prevail (Giddens 1987:41). It is therefore maintained that the necessary
critique of orthodox social theories would be more aptly regarded as
high modernity rather than as post-modernism (Giddens 1990). In archaeology, however, this debate is of less concern, since post-processual
archaeologists in general work in a conventional way but with a new
jargon.
The Durkheimian view of the nature of social reproduction, however tempting and convincing it may seem, is problematic for those
68

Apprenticeship as a Social Institution


who wish to make general assumptions regarding the ways in which
social and cultural ideas are reproduced and changed. Since the field of
archaeology is obviously a subject in which such assumptions are of
great value, these problems are pronounced in this context. First of all,
it is evident that, if humans receive all their cognitive abilities from an
external society, it is difficult, if not impossible, to explain internal,
socio-cultural change and it is consequently not possible to explain the
cultural diversity that is in evidence in the world around us. Secondly, if
cognition is solely transmitted from society to the individual, different
cultures will have different cognitive systems and it will thereby be
impossible to establish and maintain social relationships over cultural
borders, whatever they may be. Since we know that the values and
ideas transmitted through the generations, i.e. the social facts or the
collective representations, can change as a result of internal conflicts of
interest or as a result of contacts over cultural borders, we know
instinctively that something is missing in this analysis. Personally, I also
find it problematic to accept the relativistic notion that the individuals
of a social group share all cognitive functions and that these are unique
to them. I shall try to discuss these problems in relation to some of the
followers of Durkheims idea of social reproduction.

Social structure
In the British functionalistic tradition of anthropology, Durkheims ideas
became especially important. A. R. Radcliffe-Brown was heavily
influenced by the French functionalism and his definition of social structure shares important similarities with the meaning of social facts and
collective representations. According to Radcliffe-Brown, structure
should be regarded as an ordered arrangement of components (RadcliffeBrown 1952:9). The components of social structures were individual
human beings. Social structures were created as a result of the fact that
observable social phenomena within a society were not the immediate
result of the nature of the human beings (Radcliffe-Brown 1952:190).
Thus, social structures appeared as soon as a group of individuals structured their internal relationships in some other way than those based on
flock behaviour: There is continuity of the structure, just as a human
body, of which the components are molecules, preserves a continuity
of structure though the actual molecules, of which the body consists,
are continually changing. In the political structure of the United States
69

Jan Apel
there must always be a president; at one time it is Herbert Hoover, at
another time Franklin Roosevelt, but the structure as an arrangement
remains continuous (Radcliffe-Brown 1952:10).
It is important to recognise that the term social structure was created
under the influence of Durkheims idea of society as something that
encompasses the individuals within a social group, especially since the
term is frequently used in the archaeological literature. Radcliffe-Brown
regarded society as a homogeneous, organised and self-reproducing entity
and thereby shared Durkheims holistic view of cognition as originating
from society (Bloch 1989:3). According to Radcliffe-Brown, persons
could not be studied except in terms of their social structure and, conversely, social structure could not be studied except in terms of the
persons of which it was composed (Radcliffe-Brown 1952:194). However, the social structure was part of a social system that also included
certain functional aspects (Hatch 1988:224). Thus, a social system was
seen as consisting of three aspects: (a) the social structure, (b) the totality
of all social usages, and (c) the special modes of thinking and feeling
which we can infer or assume (from behaviour and speech) to be related
to the social usage and the social relations that make up the structure
(Radcliffe-Brown 1937:152). This implies that there may be room for
cultural change and generalisations regarding different societies in his
theory. The main problem with Radcliffe-Browns definition is that it
includes all social relations between persons. Or, as Firth has put it,
When I knock into someone on the street and apologise, we have a
social relation. But is this to be reckoned as part of the social structure?
(Firth 1964:37). In this context, it must be stressed that the (b) and (c)
aspects are always dependent on the shape of (a), i.e. the social structure.
Radcliffe-Brown thereby maintained Durkheims idea that cognition
in general has a social origin. This also means that he shared a relativistic
view, in that he regarded each cultural system as a unique, closed system
and rejected generalisations. In order to argue this point, he used an
example from the Australian kinship systems that were known from
anthropological field studies. The following quotation illustrates his
theory: It was impossible for a man to have any social relations with
anyone who is not his relative because there is no standard by which
two persons in this position can regulate their conduct towards one
another(Radcliffe-Brown 1913:157). The fact that Radcliffe-Brown
used the Australian Aborigines, the people who at the time were seen as
70

Apprenticeship as a Social Institution


having the lowest known degree of social structure, is intriguing and
may be related to his rejection of the idea of a single, evolutionary
development of societies. It is also worth mentioning that, when
Durkheim wrote The Elementary Forms of Religious Life, he also chose
to exemplify his views with examples from central Australia, because he
regarded totemism as an example of the most primitive form of religion (Giddens 1978:84).
In order to come to terms with the inherent problems in Durkheims
view of cultural reproduction mentioned above, Raymond Firth criticised the all-inclusive definition of social structure that Radcliffe-Brown
had maintained, because it tended to merge together permanent structures with temporary or ephemeral action (Firth 1964:37). Firth presented a solution in which there are two levels of social facts: social
structure and social organisation. According to Firth, the social structure was, just as Radcliffe-Brown had argued, the rules that society
imposed on the individual. The second level, social organisation, was
the social patterns produced by individuals who chose to follow, or not
to follow, the social structure (Firth 1964:35). Accordingly, the structures
were seen as providing the frameworks for action, while Firth maintained
that circumstances always seemed to give rise to new combinations of
factors: Fresh choices open, fresh decisions have to be made, and the
result affect the social action of other people in a ripple movement
which may go far before it is spent. Usually this takes place within the
structural framework, but it may carry action right outside it. If such
departure from the structure tends to be permanent, we have one form
of social change (Firth 1964:35). Systematic disobedience of rules
reproduced within the social structure would lead to changes in the
social organisation, to a point at which it became necessary to change
the social structure. Initially, it seems as if Firths approach offers a
solution to the problem of how to account for cultural change. The
dialectic process, between the social structure and the social order, that
in turn results in a new social structure, accounts for cultural change.
However, Bloch has pointed out that it is an illusion to think that
Firths modification of Radcliffe-Browns theory of social structure actually provides a satisfactory explanation. In other words, it is difficult
to understand how people can contribute to fundamental cultural
changes if all their concepts and categories are derived from society (Block
1989:4). It was because of this paradox, which can be traced to the
71

Jan Apel
Durkheimian view of cultural reproduction, that Firth failed to produce
a sufficient explanation of the question of cultural change. Even though
Radcliffe-Brown and Firth tried to overcome the problem by introducing
a practical level of cognition, operating in a dialectic process at a
theoretical and ideal level, they failed in their attempts to produce a
satisfactory explanation. According to Bloch, this failure can be traced
to the fact that the second level descends from the first. That is, the
different cognitive levels belong to the same cognitive system, which in
turn is unique to a certain cultural system. Consequently, one of the
main problems with the Durkheimian view of cultural reproduction is
that the organic model of society also leads to the assumption that
societies can be regarded as closed systems in equilibrium (Leach
1973:165). In the systems theory laid out by, for instance, biologist
Ludwig von Bertalanffy (1969) or within archaeology by Lewis Binford
(1972) a distinction is made between open and closed social systems. In
order to be fully effective, a pre-condition of the Durkheimian view of
cultural reproduction seems to have been the idea of a closed social
system in equilibrium. It was suggested that social systems in balance
(equilibrium) should be anticipated in social studies for analytical reasons
(Parsons 1951, Guneriussen 1997:75). However, even if closed social
systems exist, they are, just as biological systems, never in equilibrium
but quite frequently exhibit the characteristics of equifinality (Leach
1973:164). In general systems theory, the term equilibrium refers to
the condition that every closed system strives towards, namely maximum
entropy, i.e. the dissolution of all differentiation. In order to be consistent
with general systems theory, Parsons and, for that matter Binford, should
have used the term homeostatic open systems, i.e. systems that strive to
maintain a high degree of organisation (Guneriussen 1997:76). It does
not take a genius to figure out that social systems tend to be open in the
respect that they are in contact with other social systems. This means
that the members of a social system will always be aware of other ways
of living, and inspiration to change the social structure may very well
come from the outside. What should be considered in the study of
human societies is the degree to which they are open. For instance, the
archaeological remains from the Late Neolithic period in Scandinavia
reveal evidence of contacts over large areas, so in this context we obviously
need to operate with a fairly open social system in mind.

72

Apprenticeship as a Social Institution

A revision of the Durkheimian view of


social reproduction
During the 1960s and 1970s, new ethnographic evidence appeared that
changed the view of the relation between social life and concepts such
as social facts, collective representations and social structure. These
findings were linked to a newly found interest among anthropologists
in using evolutionary theories to account for cultural change. Field works
among hunter-gatherer communities in central Africa, such as the Hadza
in Tanzania (Woodburn 1962), the BaMbuti pygmies in Zaire (Turnbull
1961), and the !Kung Bushmen of Botswana (Lee 1972), revealed that
these peoples lacked permanent institutions, except those based on sex
or age groups. They seemed to be devoid of any superstructures in the
form of rituals or religious institutions (H. Knutsson 1995:34 ff.), i.e.
institutions that were regarded, by Durkheim for instance, as the first
signs of social facts. They appeared to be organised in ways in which no,
or very few, social ideals were passed through the generation in the form
of occupations or permanent institutions. They seemed to be original
affluent societies in the respect that they enjoy an unparalleled material
plenty with a low standard of living, i.e. all conceivable material
wants were easily provided for (Sahlins 1972:2). Situation-based inequality seemed to be the predominant form of social organisation, i.e.
the best hunter became the leader during hunting trips etc. At first, it
may seem as if the neo-evolutionary trend was responsible for a newly
found interest in functional aspects and day-to-day activities among
hunter-gatherers as opposed to the interest that Durkheim and RadcliffeBrown had shown in Aboriginal rituals and religion. However, at least
in the cases of the Mbuti and the Bushmen, older ethnographic descriptions are available that confirm these societies lack of interest in
the reproduction of an ideal social structure (H. Knutsson 1995). These
studies had great implications for the Durkheimian view of social
reproduction, since they implied that traditional societies existed in
which religion did not seem to function as a social glue but which were
organised and lived a natural life. As we shall see below, Bloch takes the
consequences of these new findings. It may be added though, that this
view has, during the 1980s and 1990s, been criticised by a revisionist
school that basically points out that it is impossible to regard these
societies as closed systems. This was demonstrated by pointing out that
these egalitarian hunter-gatherers lived in a symbiotic relationship with
73

Jan Apel
pastoralists and farmers and that their economy was integrated with
these societies. The revisionists urged anthropologists and archaeologists to regard the traditional societies in Africa as formed by historical
and social factors and not as the last remnants of past evolutionary stages
(Wilmsen & Denbow 1990).
Regardless of these views, it may even be suggested that, if we return
to the anthropological descriptions of these societies today, they display
signs that indicate that they have what Durkheim called a strong mechanical solidarity. Obligations, such as sharing, can be regarded as
social facts maintaining a rigid social structure, albeit without any
obvious religious undertones. If this view is accepted, the egalitarianism
of hunter-gatherers is ironically a social construction that may very well
have been created and maintained in order to prevent social stratification.
Bloch seems to be aware of this when he states that the difference between
hunter-gatherers in Australia and the Hazda lies not in their economy
but in their degree of social inequality (Bloch 1989:17). This view is
also consistent with Marxs opinion that ideology appears only in class
societies (Campbell 1986:152 f.).
It was not only the new ethnographic evidence touched upon above
that inspired Bloch to modify the Durkheimian/anthropological view
of social reproduction. Bloch aimed to reintroduce a non-ideological
level of cognition, an idea that was originally suggested by Marx and
Engels in their distinction between ideology and non-ideology. These
concepts were accordingly used to separate social categories and ethical
notions, such as the bourgeois notion of the law of property , that are
imposed on the individual by society, from a kind of cognition that
grows out of experience of the surrounding world. As we have seen, a
non-ideological aspect of cognition has been avoided by the anthropologists who followed in Durkheims footsteps. In order to find further
evidence for a deficiency in the Durkheimian view that cognition must
always preced action and therefore be created by society, Bloch refers to
the important progress made by psychologists studying the development
of children (Bloch 1989:113). Jean Piaget, for instance, suggested that
the child forms concepts as a result of its interaction with the environment and that consequently categories are constructed from the childs
experience of the world. The world from which this categorisation stems
can, according to Piaget, be reduced to the following three components:
(1) the childs own body, (2) the non-human environment and (3) other
74

Apprenticeship as a Social Institution


human beings. This conclusion was based on an extensive body of
experimental data (Maltn 1981:65) and it clashed with the
Durkheimian view that seemed to imply that the child learns or absorbs
modes of reasoning from outside sources consisting of parents, society,
etc.

The past in the present and the present


Inspired by these ideas, Blochs solution to the inherent problems in
Durkheims theory is to argue that there are not only different levels of
cognition present in human societies, but also two fundamentally
different types of cognition. He refers to these two types of cognition
as the past in the present and the present. This terminology displays a
contrast between ritual communication and other strange ways of thinking, on the one hand, and non-ritual communication and universal
concepts, on the other (Bloch 1989:11). In this context, strange should
be regarded as a cultural phenomenon that an outsider cannot make
immediate sense of. It is important to note that these two concepts
should not be conceived of as two levels of the same cognitive system,
as seems to be the case with Firths social structure/social organisation
pair. On the contrary, they represent two fundamentally different kinds
of cognitive system present in the reproduction of human societies, one
that is unique within each cultural system of thought (the past in the
present) and the other which is universal in all human societies (the
present). In this respect, Bloch seems to share Lukes view that there are
universally known categories of thought, among those social facts that
have been defined and studied by anthropologists and sociologists, that
enable people from different cultural systems to communicate and to
understand each other. The degree to which Bloch hereby suggests that,
while religion and rituals are examples of how an ideal past is transmitted
to the present, there is also another cognitive system among humans in
which categories such as time and space are common to all humans.
The problem, according to Bloch, is that most anthropologists and
sociologists have been interested in the unique ways by which the past
in the present is expressed in different cultural systems and this interest
has led them to overemphasise the importance of unique social facts.
Bloch chooses to illustrate this with an example from Bali. The people
of the Indonesian island of Bali have a ritual calendar in which the dates
of the rituals do not seem to follow any logical structure from a western
75

Jan Apel
point of view. This has led cultural anthropologists, such as Geertz, to
suggest that they have a unique perception of time which cannot be
understood by outsiders and that their social life takes place in a
motionless present(Geertz 1973:404). Geertz article is a well formulated
justification for the cultural-anthropological view of social reproduction.
Accordingly, human thought is consummately social: social in its origin,
social in its functions, social in its forms, social in its applications (Geertz
1973:361). This is a huge claim, since it suggests that anthropology is
indeed the first science, and it is not surprising that this view has caught
on among anthropologists. Geertz is empirically orientated, in that he
denies any underlying structures of human thought; there is no need to
speculate on what goes on inside peoples heads; the meaning will be
apparent through everyday life. In this respect, the article should be
regarded as a critique of structuralism, as it was presented by Claude
Lvi-Strauss, in which the importance of a universal, underlying, social
structure in the reproduction of societies was emphasised. So, even if
Geertz follows Durkheim and Weber in regarding cognition as socially
determined, he parts from Durkheim when he adopts an extremely
individualistic and relativistic approach to cultural studies. It should
be kept clearly in mind that, both here and in the discussion to follow,
distinctions are formulated from the actors point of view, not from an
outside, third-person observer (Geertz 1973:366 f., note. 5). Here he
explicitly refers to actor-oriented (sometimes miscalled subjective) constructs in the social sciences, as they appear in Webers methodological
writings.
The theoretical stance that I take in this book follows Durkheim,
Lvi-Strauss and Bourdieu, in that an outside perspective, in which
theoretical concepts are constructed, is preferred. Bloch, in sharp contrast
to Geertz, claims that, if one chooses to study everyday life on Bali, for
instance, the agricultural work, one will find that the farmers share a
perception of time, formed by the constraints of nature, such as the
weather, the seasons, climate, geology, etc. This is an understanding
shared by all humans, since it is based on brute facts. It seems as if
cultural phenomena related to the past in the present have been preferred
by sociologists and anthropologists simply because it has been in their
interests to study particular aspects of culture. Even today, when there
is general agreement that there is a biological foundation for human
behaviour, there seems to be a consensus in the social sciences that humans
76

Apprenticeship as a Social Institution


are able to adjust behaviour according to circumstances in infinite ways.
Accordingly, the social sciences should concentrate on cultural variation
in agency and leave universal human behaviour to social biology and
human ethology (Elster 1990:43). This is, of course, not surprising,
since the acceptance of the importance of the biological foundations of
human behaviour would make anthropology superfluous. It is important to note that Bloch, in line with Piaget (1968), does not mean that
the everyday concept of time is given by nature, only that it is created as
a result of the constraints of nature (Bloch 1989:214). Durkheim, like
others after him, rejected the notion that cognition was constrained by
nature, by pointing to the variability of concepts, especially of concepts
of time; but if he is wrong, his objection cannot hold (Bloch 1989: 11
f ).
It may very well be that Durkheims idea has been embraced in an
unreflective manner by anthropologists in a way that demonstrates the
power that tradition has in scientific cultures (for an archaeological
example, see Knutsson 1998). While Bloch maintains, as Durkheim
did, that religion and rituals are the main vehicles by which the past in
the present is transmitted from generation to generation in traditional
societies, I would like to expand this idea further. In this context, I
would remind the reader of Mauss article Technique de corps, which
was mentioned earlier in this chapter. Mauss was able to demonstrate
that body techniques (gestures) involved in even the most natural acts
of human agency are social facts in the sense that they are learned, i.e.
the knowledge and know-how of the gestures have to be passed down from
generation to generation. Anthropologists and sociologists have only
recently acknowledged the implications of this article. However, it is
now commonly known that technologies that are transmitted from
generation to generation should not be regarded only as means of solving
a practical problem but also as serving as vehicles in a process of passing
ideal knowledge of the past in the present to younger generations. In
other words, technologies are as vital parts in the reproduction of society,
in the Durkheimian sense, as are rituals, myths and religion (see various
articles in MacKenzie & Majcman 1985 and in Dobres & Hoffman
1999). In fact, I believe that these two kinds of cognitive abilities
function together in intricate ways.
We all know that every object that we own, no matter how insignificant it may seem, not only answers practical needs but also carries a
77

Jan Apel
social message (Olausson 2000, Hgberg 2000). Artefacts such as pencils, cars, gloves or trousers all contain these two aspects. It is therefore
important to recognise that the division between the reproduction of
the past in the present and the universal, everyday, practical behaviour in
the present is an academic reduction of reality. This does not mean that
the demarcation of these categories is useless. On the contrary, scientific
reductions are necessary in that they tend to reveal patterns that are not
obvious and therefore are difficult to appreciate by common sense. Bloch
has exemplified how these two modes of cognition may work within a
continuous process rather than as two opposites. In his article Symbols,
song and dance: Is religion an extreme form of traditional authority? he
argues that there is a continuum in which communication from
repetitive (formalised) communication to generative (everyday)
communication represents the two ends (Bloch 1989:22 ff.). In this
example, ritual communication is formalised and repetitive and belongs
near the Past-in-the-present end of the continuum, while everyday language stems from the present cognition. In Table 3:2, a few variables in
speech acts that separate the past in the present from the present are
listed.
It is important to remember that Bloch does not consider the contrasts displayed in Table 3:2 as a dichotomy of two types of speech but
as a continuum between two extreme forms of speech (Bloch 1989:25).
I have found it very interesting to note the circumstances under which
these two extremes are displayed in speech by myself and by people in
my own surroundings. There are many situations in which individuals
refer to the past in the present, using a formalised language in order to
gain audience confidence and support. For instance, at the beginning of
the 20th century, Max Weber pointed out that there was a low degree of
correlation between good scientists and popular academic speakers at
the German universities. He concluded that the popularity of certain
speakers was not necessarily based on the knowledge that they had of
their subject but rather on what were, from an academic point of view,
pretty shallow factors, such as temperament and the chosen tone of
voice (Weber 1991:14). I think that this observation can be seen as
recognising the importance of formalised speech acts in public contexts.
In his article, Bloch convincingly demonstrates how the repetitive and
formalised, ritual language is used in traditional societies to uphold a
form of traditional authority that is used as a means of creating and
78

Apprenticeship as a Social Institution


maintaining power and political control (Bloch 1989:25). The significance of this is that we should perhaps see the origin of religion in
this special strategy of leadership, the use of form for power, which we
have found in a lesser form in our study of the communication of
traditional authority, we would then see the performance of religion as
serving a special form of authority (Bloch 1989:45).
Using the operational-chain approach it is easy to see how these two
extreme forms of speech acts are analogous to the performance of a
technology. While expedient technologies have a certain degree of
flexibility as regards the sequencing of the gestures, the choices of raw
materials, etc., it is obvious that specialised technologies require a larger
body of normative rules, manifested in strict apprentice systems and
recipes for action. As we have seen in Chapter 2, a technology can be
dismantled into defined acts and stages. The quality of the dissolution
achieved by this process is dependent on the quality of the archaeological material. A gesture (regarded in this case as analogous to a speech
act) in a technology will be dependent on knowledge and know-how
that have been transmitted from previous generations, i.e. the repetitive
aspects of technology, and from the natural constraints of the human
body, i.e. generative aspects of technology. The degree of repetitive
gestures will inevitably function as a form of vehicle for the reproduction
of the past in the present in that they presuppose a certain order of
things that will give history, the teacher and his or her position, authority.
In the same way, the choice of raw material will also be dependent on
these two types of cognition. The intended artefact will in itself demand
a certain kind of raw material, but with functional equivalents (or
reasonable equivalents), the choice will be dependent on aspects of a
social nature. The craftsman will, through his apprenticeship, have learnt
what material to use and where to collect it and it is important to remember that, because of the social aspect, these choices may not seem
rational today. The same goes for the choice of tools or the place of
work etc. The relationship between the generative and the repetitive
aspects of each gesture naturally varies from gesture to gesture, but there
is no way of escaping from the social aspects of any technology, however
simple it may be to perform. Even with the knowledge and know-how
of extremely expedient technologies, there is at least a small portion of
social facts involved in the reproduction. Binford has appreciated this,
and his solution to the problem was to suggest that archaeologists should
79

Jan Apel
Everyday speech acts

Formalised speech acts

Choice of loudness

Fixed loudness patterns

Choice of intonation

Extremely limited choice of intonation

All syntactic forms available

Some syntactic forms excluded

Complete vocabulary

Partial vocabulary

Flexibility of sequencing of speech acts

Fixity of sequencing of speech acts

Few illustrations from a fixed body of accepted

Illustrations only from certain limited sources, for

parallels

example scriptures and proverbs

No stylistic rules consciously held to operate

Stylistic rules consciously applied at all levels

Table 3:2. Aspects of speech that contrast formalised speech (the past in the
present) from everyday speech (the present), according to Bloch 1989:25).

try to formulate covering laws, i.e. cross-cultural laws. Such laws can be
established only by working on the generative aspects of, for instance,
technology and in this respect Binford is one of the few who have
accepted the consequences of his theoretical conviction.
This means that archaeological artefacts are always to some degree
material expressions of an ideal social structure. Some artefacts, for
instance, blades from conically-shaped flint cores in the Scandinavian
battle-axe context or Late Neolithic flint daggers, contain a large portion
of the past in the present. Other artefact categories, for instance, bipolar
flakes in quartz that has been used in an expedient manner, may contain
a smaller degree of the past in the present. However, since all archaeological
objects are social facts, in the sense that they are the result of what Searle
(1997) has defined as institutionalised facts, since meaning has to be
ascribed to them from an outside source (society), their original meanings
will inevitably be lost when the social structure collapses or undergoes
radical changes. Since the meanings of institutionalised facts, as opposed
to raw facts, are relative, the story behind prehistoric technologies with
a large degree of repetitive cognition (the past in the present) will appear
to be more difficult for us to understand than technologies with a large
degree of generative cognition (the present).
Even if we accept that every technology, as defined in this context,
includes both repetitive and generative aspects, it has proved to be fruitful
to sort technologies into two categories, defined according to their
inherent degree of these aspects. Binfords (1962) division of technology
into technomic and socionomic types or Haydens (1998) distinction
between practical and prestige technologies reflect the need to separate
80

Apprenticeship as a Social Institution


technologies with strong social messages (of the past in the present)
from technologies that seem mainly to answer to practical needs (H.
Knutsson 2000:27). In fact, Pelegrins ordinary and complex
technologies, referred to in Chapter 2 also seem to reflect a will to
identify the generative and repetitive aspects in the production contexts
of lithic tools. However, just as in the case of speech acts, even though
these terms seem to describe a dichotomy, they should be regarded as
two extreme forms in a continuum. Personally, I believe that the
reduction of this continuum to separate categories is necessary in order
to reach some kind of understanding. Since the flint daggers are extremely
formalised objects that stem from a normative tradition in which the
repetitive aspects are pronounced, they must be regarded as artefacts
produced in a technology that carried a strong, past-in-the-present
message to the people of the Late Neolithic communities of Scandinavia.
A striking example of how such a distinction between practical and
prestige technologies can be used to give meaning to an archaeological
material is the study of Middle Neolithic blade technologies in
Scandinavia presented by Helena Knutsson (1995, 1999, 2000). In the
Middle Neolithic, Swedish-Norwegian, Battle-Axe Culture, a certain
type of blade, made from conical flint cores, was frequently deposited
in graves. Use-wear analysis reveals that, when the blades displayed any
sign of use, as they do, particularly in graves far from natural flint sources,
they seem to have been used as harvesting knives (Knutsson 1999:41).
No production place of these blades has ever been found and only five
blade cores of this type are known in Scandinavia, all deposited in ritual
contexts, such as graves and wetlands (Knutsson 2000:18). None of
the known battle-axe settlements in Scandinavia have produced evidence
of the production of these blades and thus it seems likely that the
production context was very specifically secluded.
During the same period, similar blades, although made with a slightly
different, methodological approach, were deposited in Pitted-WareCulture contexts. These were struck from cylindrical flint cores and
were predominantly retouched and used as arrowheads. The production took place on the settlements where large numbers of cylindrical
cores are found, especially in areas with natural flint sources.
In this case, it seems as if two similar technologies were used and
apprehended in the two extreme ways described above. On the one
hand, the blades from the Boat-Axe-Culture contexts seem to have been
81

Jan Apel
produced in an exclusive way near the flint source and distributed to be
used as harvest tools and then to be deposited as grave gifts. It is,
according to Knutsson, not a coincidence that this ritualised form of
technology was used in a highly formalised, grave ritual and by a people
that accepted and adhered to the Neolithic package (Knutsson 2000:30).
On the other hand, the production of the Pitted-Ware blades was
apparently not restricted, since evidence of production is found on the
settlements. These blades were also used in a seemingly profane, hunting
activity. Knutsson suggests that these differences corresponded to two
types of mentalities. One ritual version was formed by a sedentary,
Neolithic lifestyle, in which there was a strong need to reproduce what
Bloch would refer to as the past in the present. The other day-to-day
version was formed by a mobile, hunter-gatherer way of life, in which
there was little need for reproducing the past in the present through a
prestige technology or through other ritual activities.
The prestigious blades in evidence in Scandinavian Boat-Axe contexts are the result of an old tradition that can be traced over large areas
in northern, central and southern Europe, as well as the western part of
Asia (H. Knutsson 2000:18 ff.). Knutsson suggests that the grave ritual
in which the technology is embedded should be related to a ritual motif
common to a large region covering most parts of Europe and western
Asia, which was probably rooted in the emergent Neolithic way of life
in western Asia (H. Knutsson 2000:29). These blades, and the
technological process by which they were produced, obviously served
as vehicles for the reproduction of the past in the present in the Neolithic
societies of western Asia and Europe between 8000 B.C. and 2000
B.C. These objects served as a form of insignia for a certain segment of
these societies and their symbolic meaning, i.e. the story that they told,
was widely recognised in the early farming communities. It is interesting
to note that in this case it is not the technological complexity or the
normative style that separates these two blade technologies. Rather it is
the way in which the objects were produced and deposited that gives
the interpretation credibility. Thus, in this case, the interpretation springs
from the archaeological contexts. A parallel situation to these circumstances, concerning the production and consumption of bifacial arrowheads in eastern central Sweden during the Late Neolithic and Early
Bronze Age, has been published (Apel 2000c) and is presented in Chapter
5.
82

Apprenticeship as a Social Institution

Conclusion
In this chapter, it has been suggested that the Durkheimian view of the
reproduction of social facts and collective representations can be used
to understand the reproduction of prehistoric technologies, such as that
of the Late Neolithic, flint-dagger technology over an 800-year period.
However, while Durkheim tended to regard all categories of thought as
social phenomena, I have followed Maurice Blochs suggestion that
cognition basically takes two fundamentally different forms. The past
in the present refers to a type of cognition that is of a social origin. It can
be regarded as analogous to Durkheims social facts and may serve to
reproduce the ideal structure of a society from generation to generation.
It is therefore imposed on the individual, through imitation, upbringing and education by society. The present, according to Bloch, refers to
a cognition that springs from individual experience of the surrounding
world. This form of cognition is not social in its nature, in the sense
that it is handed down through the generations, but is grounded in
human biology. This type of cognition makes it possible to reach an
understanding over cultural borders and over time. The present-abilities
make themselves known in an everyday perception of time, as well as in
the universal way in which the basic structure is identical in all known
human languages.
The past in the present in its purest form is displayed in rites, myths
and religion. Bloch has demonstrated how the formalised language used
in rituals serves as an extreme form of traditional authority by which a
idealistic notion of society is imposed on its members. It has been
suggested, in this chapter, that the past in the present is also a vital
ingredient in aspects of social life that were earlier regarded as purely
functional in their nature, such as technologies. Since the purpose of
the-past-in-the-present strategies is to reproduce an idealistic notion of
the society itself, it follows that technologies that are based on normative
rules and formalised sequences of gestures also serve as vehicles, perhaps
through metaphors, by which ideal notions of the social structure are
reproduced.
It has been demonstrated, by experiments, that the reproduction of
the flint-dagger technology required theoretical knowledge, in the form
of a complicated recipe for action, and practical know-how, involving a
fixed sequencing of gestures. Thus, it follows that the organisation of

83

Jan Apel
the flint-dagger technology, in addition to the finished daggers, should
be regarded as a vital ingredient in the reproduction of the social structures in the agricultural Scandinavia of the Late Neolithic. Since social
facts of this kind are always used in competitions for prestige and power,
the following chapter is directed at an understanding of how technologies function in society.
I hope that the discussion above has shown that the people who
used and made flint daggers were not only the creators of tradition but
that their values and expressions were also formed by tradition. As I see
it, the main point about the dagger tradition was that it should be
reproducible. It was only by being transmittable from generation to
generation that the technology was able to serve its purpose, i.e. to be a
vehicle of the dominant social values of the community.

84

Chapter 4

The Emergence of Social Inequality

In the previous chapters it was argued that technologies not only meet
functional needs but also play important roles in the reproduction of
the social structure in a community. In the present chapter this knowledge will be related to certain ideas, on how institutionalised inequality
emerges and grows, that have been put forward in the complexity debate. This will lead over to a discussion of the relation between dagger
technology and social structure during the Late Neolithic and the Early
Bronze Age that will be continued in Chapter 10. Thus, in this chapter
it will be argued that the production of large quantities of high-quality
flint daggers distributed over several generations required a hereditary
system that could be upheld only in a fairly stratified society with
institutionalised heredity.

Some definitions
All human societies display certain aspects of inequality (Price 1995:130).
It is therefore not sufficient, in this context, to define egalitarian societies as societies in which there is no inequality between their members.
An acceptable definition might be that an egalitarian society has no
division of labour based on other principles than sex or age groupings.
In other words, these societies do not have permanent social institutions that can be achieved, or inherited, by persons with ambitions and
a cultural capital that allows them to do so. Since status differentiation
is present among primates and in other animal societies, Price has pointed
out that it is not an issue that concerns anthropologists and archaeologists.
A discussion of the development of hierarchical societies should aim at
an understanding of how ascribed and institutionalised status is introduced, maintained and developed (Price 1995:130).
Since the complexity debate has been focused on the growth of hier85

Jan Apel
archies from egalitarian to complex, a definition of the latter term is
also called for in this discussion. In this context, complex societies are
defined as possessing social and labour relationships in which leaders
have sustained or on-demand control over non-kin labour and social
differentiation is hereditary (Arnold 1996:78). The most complex form
of society that concerns us in this book is the chiefdoms that, according
to Service (1962) bridge the gap between tribal societies and archaic
states. Chiefdoms are politically stratified societies with settlement hierarchies, institutionalised inequality based on hereditary principles and
lite control over domestic labour (Earle 1991:1, Arnold 1993:77).
After Marshall Sahlins elaboration of a theory developed by Karl Polanyi
(1944), chiefdoms have been connected with concepts such as redistribution and pooling (Sahlins 1958, 1972). In such situations, surplus
commodities are collected by a political lite (as taxes or tributes) and
then redistributed downward in order to provide for warriors, specialists, etc. In recent years, mainly as the result of Timothy Earles work on
Hawaian chiefdoms (1977, 1978), the importance of redistribution
and pooling in the formation of complex chiefdoms has diminished.
In most cases chiefs will indeed receive tributes, but these tributes are
not the driving force in the development of complexity, even in cases
when they are actually used to provide specialists and warriors. Instead,
the incitement behind surplus production seems to be the need to create
and acquire prestige goods that are used to establish exchange networks
and thereby form the basis of a prestige economy (Earle 1991). It has
also been suggested that the development of segmentary tribes into
chiefdoms is not the result of a linear process but may take on many
different forms (Kristiansen 1991). In chiefdoms, external contacts with
lite fractions in other communities, often situated quite far away, make
up the foundation for the emergence of the organised production of
prestige items that can be exchanged. In economic terms, the prestige
items will not be as valuable where they are made as they are further
away from the production area. A thorough, cross-cultural, ethnographic
analysis has demonstrated this principle in traditional societies, i.e. statusor wealth-indicating ornamental substances are more often non-local
than local (Taffinder 1998:31 and Fig. 2:5). As regards the Scandinavian
Neolithic period, Jaqueline Taffinder concludes that there is evidence
of prestige exchange, according to her model, in Early and Late Neolithic
contexts but not to Middle Neolithic ones. Especially during the Late
86

The Emergence of Social Inequality


Neolithic, this is apparent, since during the Late Neolithic the
quantitative evidence for social differentiation is much greater (Taffinder
1998:144). Even though this conclusion is somewhat offset by Helena
Knutssons results as regards blade production and distribution in the
Swedish, Middle Neolithic, Battle-Axe contexts (see Chapter 3), it is in
accordance with archaeological evidence from the Middle and Late
Neolithic on Jutland (see Chapter 10).
Most traditional societies that are used in analogies with prehistoric
situations by archaeologists (who receive their preconceptions of social
structure from subjects such as anthropology and sociology) appear to
have a level of complexity that is neither egalitarian nor chiefdom-like
but falls somewhere in between. These societies are defined chiefly by
the practice of a segmental principle whereby smaller groups are combined into larger ones through several levels of incorporation (Sahlins
1968:15). Families tend to be parts of local lineages and lineages parts
of village communities. Villages are in turn parts of regional confederacies
that make up tribes (Sahlins 1968:15). Usually, the members of the
tribes speak a common language and have common practices. As we
shall see, Durkheims The Division of Labour in Society initiated the
anthropological treatment of these societies (Sahlins 1968:114). They
have been referred to as segmental (Durkheim 1984 [1893]), ranked,
tribal (Service 1962, Sahlins 1968) or transegalitarian (Clark & Blake
1989, Hayden 1995). Most of the historically known, tribal societies
consisted of farmers or herders, but there are several examples in which
hunter-gatherer communities had a social organisation of tribal complexity. Instead of focusing on economic issues, it might therefore be
fruitful to consider the degree of mobility as making the big difference
between egalitarian and tribal societies (H. Knutsson 1995).
All knowledge of tribal societies is based on ethnographic evidence,
mostly collected during colonial times in North America (south of
Canada and north of the Valley of Mexico), in the Caribbean and
Amazonia, in portions of sub-Saharan Africa, in the inner parts of Asia
and Siberia, in the hinterland of South-east Asia, and on the islands of
the Pacific Basin (Fig. 4:1). The interest in generalised schemes of the
development of cultural complexity coincided with the neo-evolutionary trend that permeated anthropology during the 1950s and 1960s.
When these ideas went out of fashion, so did the generalised classification systems. However, they persisted in certain strands of processual
87

Jan Apel

29
46
27
35

38

9
36

24

25

12 39
8
3 11
21
34
49

33
48

10
6
28

16

19

41

37
5 47

42

32

18

23

30
14

17

43

26
2

40

44

7
20

45
13

22
31

Fig. 4:1. The geographical distribution of historically known or present-day societies with
a social structure of tribal complexity (after Sahlins 1968).

archaeology, in which it was considered to be convenient to use the


anthropological generalisations (see, for instance Earle 1991). It is also
interesting to see that an awareness of the relevance of these generalisations is re-emerging in economic anthropology as well (Hkansson
1996:263). In this context, the degree of social complexity is measured
according to the number of formal institutions present in a society. An
increasing number of formal institutions, exclusive to certain population segments, are equivalent to a greater social stratification.

The division of labour in society


Since the late 18th century, it has been understood that a high degree of
division of labour corresponds to a high degree of social complexity. It
follows that, if a division of labour based on anything other than sex or
age groupings can be recognised in an archaeological context, it is also
possible to discuss institutionalised social inequality in prehistory
(Olausson 1993). Adam Smith, who coined the term division of
labour, regarded it from a materialistic point of view as the result of
conscious choices made by individuals in order to increase skills, reduce
manufacturing time and increase output (Clark & Parry 1990:292).
Smith exemplified his ideas by describing a pin factory where, as a result
88

15

The Emergence of Social Inequality


of a division of labour, the individual production rose from 1-20 to
4800 pins for each factory worker (Smith 1970). This conclusion suggested that the division of labour, and consequently also craft specialisation, was the result of economizing behaviour (Clark & Parry 1990:293).
In other words, Smith proposed an intellectualistic explanation of the
division of labour in regarding it as the result of conscious choices by
individuals in order to make life easier and improve its quality. This
view was further enhanced by proponents of utilitarian philosophy during the 19th century, such as Spencer, who argued that the problem of
solidarity in the division of labour is solved by the fact that each individual acts in his or her own interest in economic exchanges with others.
An increased division of labour was seen as occurring because it increased
human happiness by maximising the productivity of labour (Giddens
1978:27). It has been common to regard the division of labour mainly
as a result of rational and utilitarian behaviour in the archaeological
debate on craft specialisation and important ideas and results have been
produced (Earle 1981, Brumfield & Earle 1987, Clark ms, Olausson
1993, 1997, Hayden 1995). However, in accordance with the discussion in previous chapters, I feel that it might be fruitful to re-introduce
some anti-intellectualistic explanations of the emergence of the division
of labour into the debate. Later on, I shall try to expand on this question in an archaeological setting (see Chapters 6, 7 and 10).

Mechanical and organic solidarity


In 1893, Durkheim presented an alternative view of the division of
labour in his doctoral thesis The Division of Labour in Society (1984
[1893]). He was opposed to the utilitarian view for two reasons. First
of all, he argued that the contractual exchange, regarded by Spencer as
sufficient in itself to regulate human interaction in industrialised nations, presupposed a moral foundation on which obligations to share,
give and receive were based (Durkheim 1984 [1893]:162). The moral
foundation was seen as being unconsciously assumed by men and women
during their upbringing. This view is consistent with Durkheims strongly
held belief that collective intentionality preceded individual
intentionality. Secondly, the rising suicide rates in industrialised nations
suggested that the progression of the division of labour had not necessarily
increased human contentment (Giddens 1978:27). In adhering to a
holistic approach as regarded the organisation of society, he did not
89

Jan Apel
regard the division of labour as something that concerned only
economic life it affected all types of institutions (Durkheim 1984
[1893]:2).
During the late 19th century, there was general agreement among
social scientists that one of the most important social aspects of industrialisation was that it seemed to free the individual from the social
restraints of traditional societies. According to the utilitarians, the older
type of social solidarity was simply replaced by the contractual agreement, which, as it was based on intellectual decisions, would regulate
the behaviour of humans towards one another. Durkheim, on the other
hand, was convinced that an unfettered individualism would inevitably
lead to social decay and that the modern cult of the individual had to
be based on a more fundamental principle to secure social solidarity.
Durkheim was a passionate believer in the autonomy of the individual
and the main question that he strove to answer was how this autonomy
could be reconciled with the necessary regulations and discipline needed
in order to maintain and reinforce the bonds that held society together
(Coser 1984:xiii). In his doctoral thesis, he aimed to understand how
social solidarity was maintained as the degree of the division of labour
increased. In this respect, he worked in opposition to the leading German
social scientists at the time, such as Gustav Tnnies, who argued that
the only true solidarity was to be found in the village communities of
the past and that modernity served to undermine this Gemeinschaft
(Coser 1984:xiv). Durkheim agreed that an all-embracing, moral
consensus characterised traditional societies with a low degree of the
division of labour. However, he objected that, even if an increased
division of labour does not result in an increased, all-embracing, moral
consensus, it does not follow that there is a lack of social solidarity in
modern societies with a high degree of the division of labour. Since
laws and sanctions are expressions of the character of the moral codes
within a certain society, Durkheim believed that a classification of the
sanctions practised in different types of societies would shed light on
the relationship between social solidarity and the degree of division of
labour. In comparing societies with a high degree of division of labour
with societies with a low degree of division of labour, Durkheim noticed
a pattern. In traditional societies, with a low degree of division of labour,
a penal law with repressive sanctions served to protect collective and
strongly held, moral beliefs (Giddens 1978:25). In a repressive legal
90

The Emergence of Social Inequality


system, crimes never have to be explained and defined, since they are
obvious to all members of society because of the shared value system.
Characteristic of societies with a high degree of division of labour, on
the other hand, was the existence of restitutive sanctions aimed at restoring
the state of affairs as they were prior to the transgression of the law. In
restitutive law, the crime has to be verbally defined, because the law
does not stem from a shared value system. Durkheim concluded that
traditional societies practised penal law, because the individual in these
societies had to subordinate himself to a collective consciousness shared
by all members of society. Any transgression of the law had to be
punished, since it threatened the social cohesion of the collective
(Durkheim 1984:45). Durkheim saw no fundamental difference
between the repressive law of traditional societies and restitutive law of
civilised states; both types of laws were seen as created in order to protect
society. But if there was no fundamental difference in the functions of
primitive versus modern law, the reasons for their inherent differences
had to be found elsewhere. Durkheim suggested that the collective consciousness the totality of beliefs and sentiments common to the
average member was very strong in traditional societies. Accordingly,
the individual had to subordinate himself to the collective through the
principle of mechanical solidarity.
This view is in many ways similar to Bourdieus claim that a single
habitus is common to all members of the Kabylian communities a
Berber population in North Africa (Bourdieu 1977:17). In an article
from 1959, Le sense de lhonneur, Bourdieu describes how the sense of
honour practically makes up an economy of symbolic assets that is based
on a collective consciousness common to all members of the community (Bourdieu 1986). Each Kabylian village had its own Quanum, a
listed collection of rules and the proposed punishment of anyone who
transgressed these rules. By studying the Quanum in the village of
Agouni-n-Tesellent of the Ath Akbil tribe, Bourdieu noticed that, of
249 articles, 219 were of a repressive character, while only 25 were of a
restitutive character. The remaining five were connected with the foundations of the political system (Bourdieu 1986:61). It is obvious that
this predominance of repressive over restitutive sanctions is related to
the high degree of mechanical solidarity present in the Kabylian
communities. Bourdieu points out that, in communities such as these,
where everyone knows everyone else, the social control of the individual
91

Jan Apel
is always strong. It is interesting to see that Bourdieu, in discussing
these questions, refers to a saying from his own background amongst
the farmers of Barn in south-western France (Bourdieu 1986:43, note
17). It is apparent that Bourdieu, like Durkheim, has first-hand
knowledge of how mechanical solidarity tends to shape individual agency
in accordance with the collective consciousness.
An increasing division of labour seemed to promote the autonomy
of the individual. Since groups with different forms of collective consciousness had to function together, the collective consciousness of any
of these groups could not serve as the sole foundation of law. In order
to maintain social solidarity and thereby to unite people, a new form of
social solidarity emerged: organic solidarity. In such a system, each
specialised group of people (field) functioned together like organs in a
body, since they were dependent on each other. As a result, repressive
sanctions, which, in order to protect society have to rely on a shared
value system, were gradually replaced by restitutive, formalised law.
Hence, a greater complexity in the division of labour correlates with a
progressive transition from mechanical to organic solidarity (Giddens
1978:25).
Durkheim did not believe in unilinear evolution. On the contrary, a
growing complexity, as measured by the degree of the division of labour
in society (or, as in this context, by the relative number of institutions)
was not something that inevitably had to occur in all societies. In order
to understand why some societies developed complex social systems,
Durkheim employed a perspective based on the theory of evolutionary
biology. As we have seen earlier, the term division of labour was introduced by Adam Smith in order to describe a social phenomenon
that became apparent during the early phase of the Industrial Revolution. It was later adopted by evolutionary biologists to explain the
principles of diversity in nature. It is therefore, perhaps, not surprising
that Durkheim regarded the division of labour, not as springing from
the intelligence and will of humans, as the utilitarians would have put
it, but as a general biological phenomenon (Durkheim 1984:3).
For Durkheim, the crucial point to be explained in order to understand the mechanisms behind the growth of complex societies was the
moment at which the segmentary type of society was dissolved as a
result of an increasing division of labour. According to Durkheim, an
advanced division of labour was not possible in segmental societies and,
92

The Emergence of Social Inequality


in order to understand the origins of the division of labour, it had to be
decided why clan-based societies in some contexts had abandoned the
segmental principle. It appeared that, once this principle was broken,
the disappearance of segmental organisation would be rapid and the
preconditions for the growth of complexity would be created. A condition for the dissolution of the segmental principle was that, at some
point, separate clan groups with no earlier contact had to establish social
relations. In other words, the moral gaps between the segments, which
were the result of different collective consciousnesses, had to be filled
or, as Durkheim put it, the society had to reach a higher degree of social
or moral density. Accordingly, the division of labour varied with the
degree of moral density. In most cases, an increased moral density was
promoted by population growth since it would force different clans to
share space (Durkheim 1984:210).
Thus, the emergence of the division of labour was analogous to
what Darwin had suggested would happen if a similar situation occurred among animals. According to the theory of biological evolution, the more a single species varies in genetic properties, the more
suitable it will be to explore the different niches in nature. This will in
turn guarantee the survival of the species, as well as allowing the number
of individuals in it to increase. Therefore, an increasing population
density among animals will tend to lead to specialisation of function,
in order to allow different species to co-exist, especially in areas where
the natural resources are limited in relation to the size of the population
(Darwin 1976:88). Archaeologists may marvel at this analogy, since it
will reminds them of Oscar Montelius use of an analogous treatment
of evolutionary biology and artefact typology. However, while it has
been firmly established that this analogy is false (Grslund 1974), the
Darwinian view of culture change still prevails among some participants in the complexity debate (see, for instance, Binford 1972, 1983)
and has recently gained popularity among archaeologists, especially in
America (Persson 1999).
In this context, it may be appropriate to mention that Darwin
presented his own theory of the origin of agriculture. He proposed that
knowledge was the determining factor in the process of developing the
Neolithic techniques. When humans understood that seeds would grow
into plants, they would inevitably have chosen to become farmers,
because it would, in Darwins eyes, give a reproductive advantage to
93

Jan Apel
those who chose this strategy (Darwin 1868). It is ironic that Durkheim
and Binford drew analogies with Darwinian biology in order to avoid
intellectualistic explanations of the emergence of inequality, while
Darwin himself proposed an intellectualistic explanation for the
emergence of Neolithic technologies.
It is important to remember that Durkheims explanation of the
emergence of the division of labour is fundamentally anti-intellectualistic, since it excluded the free will of individuals as an important part
of the process. Durkheim regarded segmental societies, and probably
all types of social formations, as conservative in their nature. He therefore
concluded that any fundamental change in the organisation of society,
such as the abandonment of the segmental principle, most probably
had to be provoked from a source outside society. In this respect,
Durkheims view of social change radically differs from the Marxist
view. The view that Durkheim proposed accordingly provoked a critique
from Marxist scholars, since they regarded it as conservative (Giddens
1978:105 f.). The segmental principle would not be abandoned if population pressure occurred in a society in which surplus of resources and
space was available the problem would simply be solved by letting
new domestic units move to new territories. So, as long as there was no
external phenomenon creating a moral density that forced the society
to abandon the segmental principle, it would, owing to the conservatism of any social organisation, maintain the same principle.
There are ethnographic examples that support the idea that groups
of people, when forced to move to new territories because of economic
stress, will strive to keep their traditional social structure in the new
environment. In the northern Tanganyika region (present day Tanzania), groups of people from two different traditions gradually inhabited the slopes of Mount Meru during historical times. Both the Meru
(mountain farmers) and the Arusha (pastoral Maasai) settled in this
environment during the 17th-19th century as a result of population
pressure within both groups (Spear 1997). Initially the population grew
faster among the Arusha because they were successful in stealing their
neighbours cattle. Thus, successful age-sets rapidly colonized the slopes
as each generation of warriors retired and settled down to clear new
land to farm. The Meru population, however, expanded slowly because the lineages slowly increased in numbers, filled the land and forced
younger members to move. Spear concludes his findings in the follow94

The Emergence of Social Inequality


ing way: Both peoples thus expanded across the mountain slopes and
intensified their agricultural production in parallel but different ways
to bring about their own distinctive agricultural revolutions (Spear
1997:237). This investigation demonstrates that the same ecological
niche can be inhabited by the peoples of two quite different traditions
without any fundamental changes in their social structures. Therefore,
the social structure, at least in this case, must be regarded as a normative
system that is based on tradition and transmitted vertically through the
generations. There are also many historical examples from our own
cultural sphere in which people were forced to move but took with
them, and sometimes even reinforced, traditional values in the new
country.
In a sense, following the ideas of Durkheim, one could argue that
cultural norms work strongly in traditional societies in a mechanical,
normative way. On the contrary, in industrialised societies people from
different social groups with different forms of collective consciousness
will be forced to act together in an organic fashion. The development
of the division of labour had earlier always been considered to be an
evolutionary process proceeding from an organic (natural) structure to
a mechanical structure. It is important to recognise that Durkheims
view of the origins of the division of labour among humans implies
that the social structure of traditional societies is very much governed
by cultural norms and does not reflect a natural way of life. The division
of labour varies in direct proportion to the volume and density of
societies and, if it progresses in a continuous manner over the course of
social development, it is because societies become regularly more dense
and generally more voluminous (Durkheim 1984:205).
Durkheims anti-intellectualistic approach to the emergence of the
division of labour in human societies did not have an impact on the
archaeological theories regarding the emergence of complex societies in
prehistory until the 1960s. However, in an attempt to explain the human adaptations to the post-Pleistocene environment which resulted
in the Neolithic way of life, Lewis Binford presented an interpretation
that was in many ways similar to Durkheims theory presented above
(Binford 1968, 1972, 1983). As we shall see, the arguments presented
are relevant to a discussion on the division of labour and its relation to
the emergence and growth of social complexity and inequality.

95

Jan Apel

Population pressure and marginalization


The new ethnographic data available to Binford, especially concerning
egalitarian hunter-gatherers (see Chapter 3), meant that the interest shifted
from the transition of segmental tribes into chiefdoms and archaic states
to under which circumstances egalitarian, mobile hunter-gathers are
transformed into sedentary tribes. Binford adheres to the antiintellectualistic approach to the emergence of complex social systems in
general: It is far from clear, in any case, why anyone should want a
complex system to the extent of investing effort in over-production;
there must exist some pressures for change in the Darwinian sense.
However, we yet have to identify what they are and how they operate
(Binford 1983:231). He points out that one of the most misleading
explanations of social evolution, in this respect, are the profit-seeking,
economic arguments that have been put forward by utilitarians (in
economic theory) and formalists (in anthropology). According to
Binford, the dynamics involved in ecological processes are non-rational
and not guided by intellect.
Binford rejected earlier explanations of the post-Pleistocene adaptation of humans in different parts of the world, resulting in the
introduction of Neolithic economies (Binford 1972). Gordon Childe
argued that the Neolithic revolution was the result of a direct adaptation by humans to the rapidly changing environment that appeared at
the end of the last ice age. These changes included the proposition that
the Sahara and other deserts were becoming increasingly desiccated.
Animals and humans were therefore forced together in oasis-like places
in river valleys and other delimited areas with rich natural resources. In
these new environments, humans, plants and animals were able to live
in a symbiotic relationship to each other and humans could eventually
settle down and practice husbandry and agriculture (Childe 1928,
Binford 1983:196). It was in fact this theory that Robert Braidwood
set out to test when he started his fieldwork in the Near East in the late
1940s. However, after applying a variety of techniques, including pollen
analysis and sedimentology, it was concluded that there was no evidence
of environmental changes occurring prior to the Neolithic revolution
(Braidwood & Howe 1960, Binford 1983:197). Having to abandon
Childes oasis-theory, Braidwood instead came up with an intellectualistic explanation of the emergence of food-producing techniques.

96

The Emergence of Social Inequality


According to this theory, the development of the human mind eventually allowed humans to invent the Neolithic techniques. Thus, according
to Braidwood, the origin of agriculture was a cultural achievement.
In opposition to these theories, Binford (1972) argued that major
demographic changes were responsible for the emergence of the
Neolithic technologies and, consequently, also for agriculture. This
explanation shares an important aspect of Durkheims view of the
emergence of the division of labour presented above. Binford was of
the opinion that population pressure in itself, and under special circumstances, would lead to a division of labour and social inequality. In
an article published in 1968 they took over the scenario, presented by
Childe, in which small, oasis-like areas, with an abundance of especially
maritime resources, were created all around the temperate zone. These
changes were mainly seen as a result of the isostatic changes in the general
sea-level that occurred as a global phenomenon due to the smelting of
the ice during the early, post-Pleistocene period. However, according to
Binford, it was not in these areas that the Neolithic technologies were
developed. A consequence of their Darwinian approach was the belief
that specialisation had to be provoked and that it would not occur in
areas where there was a satisfying plenty. Thus, it was argued that the
more sedentary life the result of the newly available resources
would lead to an increased population density. This situation would
not be a problem in areas where resources were widespread and new
social units could move to new territories and support themselves in
the traditional way. However, a density problem would occur in areas
with limited resources; under such circumstances the new social units
would, sooner or later, have to move to areas that were not as rich as
those that they came from. According to Binford, in order to maintain
the traditional social order, such situations would lead to an increased
specialisation in the anti-intellectual sense of the word. The new
Neolithic techniques, including husbandry and crop-growing, would
thus tend to be developed in marginal areas and be practised by a
marginalized population. It was further claimed that the earliest evidence
of food-producing techniques in the Near East had indeed been located
in areas adjacent to the areas where the earliest evidence of sedentary life
had been found (Binford 1972). Thus the development of agriculture
was preceded by a sedentary way of life that in turn was the result of
population pressure. While it is obvious that the spread of a technology,
97

Jan Apel
after it has been invented, cannot solely be explained by this theory, it is
interesting to bear it in mind in discussing the evidences for technological
specialisation that occur in the Late Neolithic communities of southern
Scandinavia. In Chapter 10, I discuss whether the archaeological evidence
from Jutland, as regards the emerging flint-dagger technology at the
transition between the Middle and Late Neolithic periods, can be interpreted as the result of a marginalization process.
In later writings, Binford modified his opinions but maintained the
anti-intellectualistic proposition that the origins of agriculture, and
complexity in general, should be regarded as the result of population
pressure. He now argued that the reason why population pressure occurred all over the world almost simultaneously was that humans are
most reproductive in the temperate zone at the mean earth bio-temperature of 14.4 degrees Celsius. From a biological point of view, this
makes sense, since the main regulatory mechanism seems to be mortality in the tropics and fertility in the Arctic, the two coming together
in the temperate zone (Binford 1983:209). This means that when
Homo sapiens reached the temperate zone, it was only a matter of time
before density problem would occur that in turn would provoke new
economic strategies, such as husbandry and agriculture. This process
evolved especially quickly on the American continent, since the population there had gone through a disease filter, and were no longer
threatened by native disease organisms, which allowed a population
explosion (Binford 1983:209 f.).
Binford maintained that the role of craft specialisation in the development towards greater complexity in traditional societies should not
be overemphasised (Binford 1983:221). The implications of the antiintellectualistic ideas presented by Durkheim and Binford suggest that
it is problematic to explain specialisation in traditional societies by references to modern economic theories. It is, in fact, rather common in
traditional societies for specialists to be low-class or outcast individuals
or families with no real political influence. This implies that the occurrence of craft specialisation is a secondary effect of hierarchical differentiation. In such a scenario, specialists are people or families who, as a
result of, for example, population pressure, are forced to move into
low-production areas and to earn their livings as part- or full-time
specialists. Dean Arnold followed this line of reasoning in explaining
individuals or families who have turned to ceramic craft specialisation
98

The Emergence of Social Inequality


as people who have been separated from land tenancy or ownership
(Arnold 1985). The main idea being that, if the population density of
a region increases beyond the carrying capacity of the natural environment, it makes sense to manipulate the natural environment in order to
be able to maintain the traditional social structure. However, this intention is not always fulfilled. As regards Neolithic techniques, for
example, Esther Boserup argued that argriculture in itself tend to promote technological development, since an increased population demands
it (Boserup 1966).
Most archaeologists who have proposed that population pressure
may have triggered complexity have, however, not leaned on Darwinistic
theories, such as those presented by Durkheim and Binford. Instead,
Boserups ideas on the emergence and development of agriculture were
the main source of inspiration for the conventional archaeologists of
the 1970s and 1980s. Recently, Stephen Shennan (2000) has presented
a theory aimed at interpreting culture change in prehistory as a result of
population fluctuations. This is, as I see it, a new attempt to discuss
culture change as a result of demographic circumstances in anti-intellectualistic terms, as opposed to most processual archaeologists who, during the 1970s and 1980s, followed Boserups intellectualistic idea.
Shennan points out that there is a strong element of vertical
transmission of the techniques that produce artefacts in traditional
societies. In other words, crafts, in such contexts, are often handed on
from parent to child within lineage or clan groups. This would indicate
that the distribution of sites displaying a certain technology would, in
fact, just as the old Kulturkreis theories suggested, be markers for a
social territory. Shennan suggests that there were severe fluctuations in
the prehistoric populations and that these fluctuations, combined with
the idea of evolution were as influential on cultural patterns (the archaeological record) as they are on genetic patterns (Shennan 2000). In
other words, Shennan, just like other archaeologists who discuss culture
change in Darwinistic terms, regards artefacts as parts of the human
phenotype (Rudebeck 2000:209).
As we have seen in Chapter 3, within hunter-gatherer communities,
the egalitarian social structure can be regarded as a culturally created
protection from hierarchical tendencies. In such systems, individuality
and entrepreneurial behaviour are under strict collective control. In such
homeostatic, open systems a steady, low population density is
99

Jan Apel
maintained by limiting the number of children, for instance by abortion, contraception, abstinence or infanticide (Binford 1972:437). This
may indicate that the social problems following from increased population density are known in these societies. In more open social systems,
such as transegalitarian societies, a steady population density is maintained by letting surplus families, or individuals, move to new areas. In
such societies, Binford suggests that during a period of population
growth, the increase in the size of the local family or extended family
will tend to split into equal units, as long as there is space enough to
allow the new family to act as a unit of production. When this is no
longer an option, the segmentary principle will break down and groups
of people will become marginalized, while others will keep or develop
their privileges (Binford 1983:2).
The main problem with the population pressure theory is that several studies of mobile hunter-gatherers published during the 1960s demonstrated that these societies never seemed to allow the population to
grow above the carrying capacity of the land. It has also been suggested
that larger, more complex societies, such as the Iroquois, in some cases
strive towards maintaining egalitarian structure in villages with several
thousands of inhabitants (Trigger 1990). So the main question one
should ask, as regards specialisation and the emergence of complexity is
perhaps why the Neolithic way of life was ever considered, or allowed,
in the first place. The crucial point that needs to be explained is why
hunter-gatherers are forced to become sedentary (Binford 1983:208).
Since sedentism is regarded as the ideal way of living in western societies, this question has seldom been asked before. It has simply been
taken for granted that humans will prefer a sedentary way of life before
mobility, if they can find ways to support themselves in one place.
Binford, however, presents several examples that indicate that mobile
hunter-gatherers never consider a sedentary way of life because all their
knowledge of the world and how to use aspects of it in order to live a
good life is based on mobility. In other words, they regard mobility as
the ideal way of living. Binford does not accept Childes and Braidwoods
intellectualistic explanations of the emergence of the Neolithic technologies which are based on the assumption that humans will become
sedentary when there is an abundance of available resources in one place.
On the contrary, he maintains that, when present day hunter-gatherers
(he explicitly refers to the Nunamiut Eskimos in Alaska and to Aus100

The Emergence of Social Inequality


tralian aborigines, among whom he conducted fieldwork) find a place
where there is an abundance of resources, they tend to move. The security
that surplus resources in a certain area give is used as travel insurance.
Hence, when such situations arises, the group will use this security to
explore areas in their territory that they have not visited for a long time;
if resources are scarce here, they can always go back.
Now we have to ask ourselves, did the abandonment of the segmental principle really unleash a maverick or aggrandizer personality that in
egalitarian and transegalitarian societies was subject to firm social control.
The frameworks presented by Durkheim and Binford seem to suggest
that this may have been the case. In this context, the works of Brian
Hayden will be introduced into the discussion.

Pathways to power
From an empirical point of view, Hayden has tried to answer the questions about the emergence of inequality. He has conducted a crosscultural analysis of anthropologically documented, transegalitarian societies and has divided them into sub-groups according to their relative
degrees of complexity. These sub-groups are defined by the different
strategies that they use to gain and maintain power. Hayden discusses
only communities with hunter-gathering and/or agriculture in their
economies and does not consider pastoral groups. In Scandinavia, pollen evidence speak in favour of an economy that included cultivation
during the Late Neolithic (Hedeager & Kristiansen 1988, Welinder
1998, Heiman 2000, Ethelberg 2000, Lagers 2000). An increasing
number of excavated sites with houses and large structures implying
sedentary hamlets in both the southern and central parts of Scandinavia
are strong arguments for this case (Jensen 1973, Bjrhem & Sfvestad
1989, Strmberg 1992, Artursson 2000). According to Hayden, the
main reason for the growth of social inequality was the continuous and
increasing production of surpluses and a social order in which aggrandizers were encouraged to use these surpluses in strategies to achieve
power. Hayden maintains that the aggrandizer concept covers a human
type that has an aggressive and acquisitive nature and is present in all
human populations.
Initially, Hayden distinguishes between transegalitarian communities in which the social structure is reproduced within independent
households and those in which it is reproduced within a larger corpo101

Jan Apel
rate group consisting of several households that share the exclusive control of an important subsistence resource (Hayden 1995:18 f.). Within
these two types of communities, the strategies used to accumulate power
are different in their natures. In the case of the Late Neolithic societies
in southern Scandinavia, we shall have to consider both of these possibilities. It is plausible that a corporate community had exclusive control of the raw-material sources for dagger production and that the
apprenticeship was open to all children belonging to this group. A
situation like that would indeed have certain advantages for the reproduction of the technology. It would, for instance, mean that a larger
base was available to choose apprentices from. On the other hand, it is
also plausible that the flint-dagger technology was organised by families in control of the knowledge and the know-how inherited in the
technology rather than of the flint source. Such a scenario would also
guarantee the exclusiveness of the technology to a certain segment of
the community and would in fact also make it possible for communities with interesting recipes for action to use them in social strategies. I
shall return to these questions below.
In both communities based on independent households and those
based on corporate groups, three different levels of complexity have
been recognised and described: Despot, Reciprocator and Entrepreneur
societies (Hayden 1995:19, Fig. 2). The developments of these forms
relates to how an increasing amount of surplus was used by aggrandizers in strategies to maintain and increase their power (Hayden 1995:25).
Haydens approach to an explanation of the occurrence of institutionalised inequality is explicitly materialistic in its nature. This is revealing
in many ways, especially since it makes the argumentation logical within
its own realm. However, Haydens ideas are, as I see it, compatible with
the general theory of human practice and symbolic capital that has been
developed by Bourdieu. The incorporation of some key concepts from
Bourdieus theory of praxis into the discussion will demonstrate that a
purely materialistic approach to the problem can be moderated (see
below).
In Despot communities, the strategies used by aggrandizers to maintain and increase control include provoked warfare and reciprocal exchange (Tab. 4:1). In Reciprocator communities, provoked warfare and
reciprocal exchange are still used, but bride-wealth and child-growth
payments, and to a certain degree also different rituals, are also intro102

The Emergence of Social Inequality


duced as vehicles for social advancement. In Entrepreneur societies, the
use of provoked warfare and reciprocal exchange diminishes, while bridewealth and child-growth payments become more important. Competitive feasts and investments are also strategies that are introduced in Entrepreneur communities (Tab. 4:1). According to the preconceptions
of the organisation of the dagger technology that was presented in
Chapter 2, institutionalised heredity was proposed as an important aspect of the reproduction of the technology. In independent household
communities, strong heredity is reserved to entrepreneur societies.
However, strong hereditary systems are not equivalent to institutionalised heredity. For instance, among Big Men, a type of aggrandizer that
Hayden assigns to Entrepreneur communities, there is no evidence of
institutionalised heredity. But the fact that every Big Man has to create
his own position does not mean that hereditary principles are not in
action. In 85 % of the cases, the child of a Big Man does in fact become
one himself. I would not be surprised if similar figures turned out to be
relevant in describing other specialist traditions in transegalitarian communities as well. One of the archaeological consequences of Investment
exchange, which is an aggrandizer strategy exclusive to Entrepreneur
communities, is craft specialisation, which is a form of investment
strategy (Fig. 4:1). As we shall see, child-growth strategies does also
include aspects of craft specialisation. Thus it can be concluded that the
presence of institutionalised heredity and craft specialisation in the late
Neolithic societies in southern Scandinavia suggests a high degree of
social complexity. On Haydens scale, such communities would be classified as Entrepreneur or Chiefdom communities.
If the interpretation of the Late Neolithic flint technology, as reproduced in an apprenticeship system based on hereditary principles, is
correct, Haydens Entrepreneur level is the appropriate place to start a
comparative analysis. As we have seen in Fig. 4:1, child-growth strategies are, according to Hayden, one path which can be taken by aggrandizers in Entrepeneur societies to create and maintain inequality in
Entrepreneur communities. Three different kinds of child-growth strategies are used within Entrepeneur communities (Owens & Hayden
1998):
(1)Ceremonies in connection with puberty.

103

Jan Apel
(2)Initiation rites (into secret societies).
(3)Special formal training (in a certain technology, often involving esoteric knowledge of some kind).
If a family is able to let children participate in such rituals or training, their symbolic capital will increase and the symbolic capital can
then be used in marriage negotiations. Those wanting to marry a child
who has undergone an expensive ceremony, or initiation rite or a special
form of apprenticeship have to be able to afford it, i.e. they will have to
have an equal amount of symbolic capital.
It may now be suggested that one of the main functions of the flintdagger apprenticeship system was to maintain and enhance the symbolic capital that this craft gave, through the generations. It is not necessary to believe that the flint smiths belonged to the lite of their
communities in order to suggest this. The ontological security that the
membership of a production fraternity of this kind gave its members
is, in itself, a sufficient explanation of its reproduction. I shall now
proceed to relate Haydens important ideas to the theory of symbolic
economics (in the widest meaning of the word) presented by Bourdieu.

Technological traditions as vehicles for the


reproduction of symbolic capital
Bourdieus attempt to present a general theory of human practice has
resulted in some concepts that may be useful in this context (Broady
1991, Bourdieu 1999a). The strength of Bourdieus theory-building is
that it has been used in various empirical studies and that it is based on
first-hand observations of how individuals and groups of people act
and interact. Regarding the problematic relationship between cognition as being socially transmitted or picked up on an individual level as
the result of the childs interaction with its surroundings, Bourdieu
suggests that this reasoning is tautological in character. On the one hand,
he acknowledges the psychological view of cognition, as presented by
Piaget, that the child picks up his or her cognitive abilities by interacting
with the surrounding environment. However, according to Bourdieu,
environments are never neutral in character, since they are coloured by
ideology, in the form of physical structures, artefacts, etc. Or as Bloch
puts it: A child brought up in a Berber house by Berber parents picks
up Berber notions, just because the material nature of the house, as well
104

The Emergence of Social Inequality


Aggrandizer Strategies
Territorial Conquest
Provoked War
Bride-wealth
Obligatory Initiations

Egalitarian

Despots

Reciprocators

Entrepeneurs

Chiefdoms

..xxxxxxxxxxxx
xXXXXXXXXXXXXXXXXXXXxxxxxxxxxxx..
..xxxxxxxxxxxxxxxxXXXXXXXXXXXXX lite/Nonelite
xxxxxxxxXXXXXXXXXXxxxxx.

Child growth

.xxxxxxxxxxXXXXXXXXXXXXXXlite/Nonelite

Investment

.xxxxXXXXXXXXXxx

Feasts
- Solidarity
- Reciprocal Exchange
- Competitive
Cults: Ancestors and Others

XXXXXXXXXXXxxxxxxxx.
xxxxxxxXXXXXXXXXXXXXXXxxxxxxxxxxxxxxx
.XXXXXXxxxxxx(lite only)
..xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxXXXXXXX

Table 4:1. Strategies used by different types of transegalitarian aggrandizers to increase


power and control (after Hayden 1995:31, Fig. 4).

as the behaviour of the people with whom he interacts, contains in


itself the specific history of the Berber house (Bloch 1989:118). In
this way ideology is transmitted from generation to generation in society, partly as the unintended result of the cultural influence of the environment.
Before continuing any further, I would like to devote a few paragraphs to Bourdieus epistemology, since I find it interesting and believe it to have special significance for the subject of archaeology. Donald
Broady has argued convincingly that Bourdieus sociology is firmly based
on the idea that the scientific object has to be constructed (Broady
1991:397). This construction cannot be based on the kind of ideas
based on common sense appearing through intuition born out of human
experience of the empirical reality. On the contrary, it must be based on
the conviction that the scientific way of thinking is, in fact, able to
produce concepts and mental tools that can never be induced from
empirical observations alone. To be a useful tool for sociologists, a hypothesis evolving from such concepts must be applied to the experienced world in a systematic, scientific way, in order to allow the empirical data to talk back (Broady 1991:397). In other words, an idea
based on scientific principles has to be confronted with the experienced
world, for instance, by rigorously applied, statistical methods, in order
to be rejected or confirmed. The dialectic process between theory and
105

Jan Apel
empirical experience, and the rejection of intuition as a useful method
in sociology, are things that Bourdieu shares with the Durkheimians
and makes him a proponent of a kind of applied rationalism. Broady
claims that the classical opposition between explanation and understanding is transgressed by this dialectic relationship between the hypothesis
and the statistical verification (Broady 1991:397). The fact that Bourdieu
claims that sociology should construct its own subject in this way means
that he promotes a continuous dialogue between closed rationalism, on
the one hand, and naive empiricism, on the other. This applied
rationalism is thus not deductive or inductive but is best described as
constructivistic in its character (Broady 1991). In archaeology, a similar, theoretical, approach has been presented by Kjel Knutsson and
successfully applied during archaeological excavations in eastern central
Sweden (see Apel et al. 1994, 1996 & 1997, Hallgren et al. 1996 &
1997, Knutsson & Lekberg 1996). As we shall see in Chapter 5, an
archaeology that concerns itself only with rational arguments that can
be verified in formal ways will tend to be very sterile. On the other
hand, an archaeology that is not firmly embedded in a scientific way of
thinking but concerns itself only with learned speculations is also, at
least to my mind, not worth the trouble.
One of the cornerstones in Bourdieus theory of practice is actually
the idea that social complexity is the result of a certain kind of diversification process. The main idea being that all institutions derive from a
common ancestor institution. As we have seen, Durkheim believed
that religion was an institution of this kind. Bourdieu considers the
domestic unit, the family, as the fundamental human institution
(Bourdieu 1999a). The domestic unit is one of the few institutions
that actually practice a positive reciprocity (Sahlins 1972). It is also in
the families that important aspects of habitus are created. According to
Bourdieu, the strategies for social reproduction proceed from the fact
that families are corporate bodies that aspire to maintain their social
being and accordingly preserve all their possibilities and privileges. This
is achieved through strategies for childbirth, marriage and rules of inheritance, economic strategies and strategies of education (Bourdieu
1999a:31). Among transegalitarian societies, strategies for concentrating power include warfare, bride-wealth, child-growth payments, membership of corporate lineage groups, and claims to supernatural power
(Hayden 1995:25). In this respect, Bourdieu and Hayden share funda106

The Emergence of Social Inequality


mental theoretical principles in the form of a anthropological view of
humans that stresses that individuals will strive to gain and maintain
power and influence over others if they are given the opportunity to do
so (as regards Bourdieu, see Broady 2000:23). The most fundamental
social institutions, such as the family unit with positive reciprocity or
an egalitarian social system, can thereby be interpreted as restricting
individual agency on favour of the collective good.
If all institutions stem from one original institution, such as the
family unit, it is perfectly clear why the rules that govern such seemingly different institutions as the modern monetary system, the legal
systems and the intellectual field are actually organised according to the
same fundamental, underlying principles. It is these fundamental
principles that Bourdieu has investigated in different contexts using tools
such as symbolic capital, habitus and social field. I shall discuss the possible
relevance of these concepts to interpretations of prehistoric conditions
later on. But first I would like to define them. As Broady has pointed
out, Bourdieu himself seldom defines the terms that he uses. On the
contrary, he seems to avoid presenting strict definitions, presumably in
order to maintain the opportunity to improve and refine his tools
through his research. Fortunately, Broady has tried to define Bourdieus
most important concepts as distinctly as possible and the following
definitions rely on his work (Broady 1991).

Symbolic capital
According to Bourdieu, different assets, in the form of different kinds
of capital, i.e. physical, economic, cultural or social capital, are transformed into symbolic capital when they are perceived by social agents
with categories of perception that allow them to recognise them, acknowledge them and appreciate their value (Bourdieu 1999a:97). Broady
(1991:169) has defined Bourdieus term capital as values, assets and
resources that are symbolic or economic in their nature. Bourdieu has
been especially interested in symbolic capital, and the shortest definition of symbolic capital is what is recognised as valuable. In this sense,
symbolic capital is a very general term, but it is important to recognise
that it does not concerns single individuals; it solidly rests on the
common consciousness of groups of people (Broady 1991:169). It is
therefore not enough to have a certain possession or skill; it is also required
that this asset shall be accepted as valuable by a group of people. In
107

Jan Apel
relation to the subject of this book, one could argue that exceptional
flint-working skills cannot, in themselves, be regarded as a form of
symbolic capital until they are recognised by the collective as such. One
important consequence of Broadys definition of symbolic capital is
that economics, which in certain schools of economic thought has been
regarded as an independent field of research following its own logic,
can in this context be regarded only as a variant of symbolic capital.

Cultural capital
This form of symbolic capital is, according to Bourdieu, a very important form of capital in western societies, such as France. It is connected
with the kind of prestige that knowledge of the fine arts, literature and
classical music gives persons that possess such knowledge. It can also be
the kind of prestige that a certain dialect gives its speaker. Bourdieu
(1991:138) has pointed out that those born in Arrondissement 7 in
Paris will gain linguistic prestige as soon as they speak. In order to make
his point, he concludes that the overwhelming majority of those who
led France at this particular point in time, 1978, spoke this dialect.
Degrees from prestigious schools are also to be counted as cultural
capital. It is important to recognise that, as opposed to symbolic capital,
which may be specific to certain segments of society, cultural capital is
a form of capital that is acknowledged by society as a whole. Symbolic
capital, on the other hand, also incorporates assets that are only acknowledged within certain fields of society (Broady 1991).
It has been suggested that the acknowledgement of cultural capital
emerged as the result of specific historical circumstances that led to the
possibility of more or less permanently stored, symbolic capital in the
form of degrees, rules and regulations, written documents, etc. (Broady
1991:171). According to this view, cultural capital could not have been
important as a dominant form of capital before the introduction of
written language, printing techniques and the development of formal
institutions for education that can provide printed degree certificates.
However, as we have seen earlier in this chapter, Haydens generalisations regarding the more complex transegalitarian societies indicate that
different kinds of cultural capital are recognised within traditional societies as well. Families that can afford to let their children go through
initiation rites or let them participate in time-consuming education,
such as certain apprenticeship systems, can use the value that these
108

The Emergence of Social Inequality


strategies create in order to increase bride-prices etc. The value of these
rites and kinds of education is recognised by society and this gives the
child and his or her family a form of cultural capital that can be used in
marriage negotiations. One may suspect that the formal aspects of a
Late Neolithic, apprenticeship system, such as is was presented in Chapters
2 and 3, are a precondition that allows cultural capital to be used in
traditional societies.

Social capital
This form of capital, according to Bourdieu, can be regarded as assets
that take the form of personal relations based on kinship, on ties between old school friends or on personal contacts that have been established during work. Personal relations of this kind can obviously be
used as a form of capital in different forms of social strategy (Broady
1991:169). As regards traditional societies, Bourdieu has noted that
certain Arabic-speaking tribes in Algeria have recourse to a certain capital
initial that is attached to the kinship with certain mythological ancestors, a capital which can be converted into power and prestige (Bourdieu
1958:85, Broady 1991:183). We may therefore assume that certain
kinship ties were used in prehistory as well. As we have seen in Chapter
3, this was actually suggested by Helena Knutsson when she connected
the use of a certain blade type as grave goods with the communication
of mythical, ancestral relations. In a similar way, it is tempting to regard
the Late Neolithic flint dagger as an insignia that referred to a kinship
with mythological ancestors.
It is only the complexity of the social structure of the society in
question that sets the limits to the number of types of capital that can
be defined. Thus, in the modern western world an infinite number of
such types can be defined, while, in traditional societies, the number of
types of capital that are acknowledged as such within the communities
may be considerably fewer.

Habitus
Habitus is a system of arrangements that will enable people to act, think
and orient themselves in the social world and, as such, it can be regarded as a form of embodied symbolic capital (Broady 1991:17). In
the habitus concept, Bourdieu has demonstrated how individual and
collective intentionality merges into the individual actor. The habitus
of a certain individual is formed by the life that he or she has lived and
109

Jan Apel
that will guide his or her imagination and practice and thereby, in most
cases, contribute to the recreation of the social world. However, in
situations in which the habitus in fundamental ways does not coincide
with the structure, it may help to change the social world (Broady
1991:225).
Habitus can under certain circumstances be regarded as form of symbolic capital, although it differs from, for instance, cultural capital, in
that it does not take the form of degrees or titles, etc. Habitus is embodied symbolic capital; it takes the form of a certain dialect, the capacity
to distinguish good taste in literature, theatre, film, etc. I did not put
quotation marks around the word good in the preceding sentence
since I referred to the objective form of good taste in society the
taste of the dominant segments of society and consequently the kind of
taste that is rewarded in social strategies. Someone who speaks a country
dialect and has simple tastes as regards, for instance, literature but is a
good motor mechanic will most probably only be able to use his or her
habitus as capital to promote his or her status and influence in a limited
circle of people. This means that peoples habitus can be related to the
constitution of their inherited or acquired capital (Broady 1991:228).
Any given habitus will facilitate the composition of a certain register of
strategies that will define and limit the freedom to act of the person in
question. It is very difficult to change ones habitus. Accordingly,
Bourdieu has pointed out that different dispositions that are contained
within the habitus concept will remain long after the social situations in
which they were created have ceased to be. A persons habitus is primarily
formed in his or her family (Bourdieu 1999a). A person in a familiar
environment will intuitively know what to do and what to say if the
habitus is in accordance with the surrounding milieu. Such a situation
will make the person feel at ease with the situation and the surroundings will appear to be natural. Durkheim referred to these types of
feelings as springing from a form of social instinct (Durkheim
1993:118).
In traditional societies, the habitus can sometimes take on a very
tangible appearance, for instance the way in which the degree of fatness
corresponds to prestige on certain islands in the Pacific Ocean. Certain
initiation rites that will result in characteristic, physical traits in the
individual who undergoes them are also encountered in the ethnographic
literature. In some communities, girls are subjected to long stays in110

The Emergence of Social Inequality


doors without being able to work, as a form of initiation rite. Their
weak physical appearance, i.e. a body that has not had to perform manual
labour, will also be recognised by all in the society and can thereby be
used as a form of symbolic capital, especially in marriage negotiations.
Bourdieu has pointed out that, among Kabylian farmers, the physical
representation of an honourable farmer is connected with a certain
posture. An honourable farmer will thus walk in a certain way, not too
fast and not too slow, and keep his head high. The members of the
society will intuitively know the social significance of these gestures,
since they all share a common consciousness. It is not unlikely that the
Scandinavian flint-smiths shared an almost identical habitus, especially
if the craft was transmitted vertically through the generations in an apprenticeship system. It is also possible that the working conditions,
including being exposed to large quantities of silica dust, would give
the flint-smiths certain exclusive physical traits, such as a certain kind of
cough.

Social field
A social field is a system of relations between positions that are held by
people or institutions that argue and negotiate on something that they
have in common (Broady 1991:9 f.). Usually, what unites the participants in a field is certain common factors of their habitus. The existence
of a field presupposes specialists, institutions and value hierarchies that
are recognised by the participants in the field. A field is also characterised by different agents having different positions. This results in a
continuous power struggle between the dominant and the dominated
fraction of the field. The members of a field will share certain characteristics in the habitus with each other. This means that they will have a
deeper understanding of each other and each others actions than is usually the case among individuals in society in general. Most archaeology
students will, for instance, realise this when they think back to the feelings
of mutual interest that occurred when they started to socialise with
other archaeology students. However, there are also differences in the
dispositions that make up the habitus of the members of a field. These
differences stem from the fact that the members have different social
backgrounds. Some members of the field will have backgrounds that
have given them dispositions that are more in accordance with the
objective structure of the field than other members have. For such
111

Jan Apel
persons, the objective order of the field will simply appear to be in a
natural state. In intellectual fields, such as archaeology, it is not unusual
for persons who have a high degree of their habitus in accordance with
the ideal structure of the field to regard themselves as being smart and
being talented. Members with a lesser part of their habitus in accordance with the structure cannot avoid reflecting on the fact that parts of
the objective structures of the field make them feel uncomfortable.
They will use their influence to change the structure of the field if they
are given the opportunity. It is the dynamics, formed by the objective
structure of the field and the variations of the constitutions of the members habitus, which promote continuous change in modern production
fields.
As regards prehistoric institutions that produce symbolic capital, for
instance, in the form of dagger production, the field concept may not
be an appropriate term to use. In traditional communities, in which
crafts are transmitted vertically through the generations within lineage
groups, clans or castes, the members of the guilds will, owing to their
similar social backgrounds, have certain almost identical traits in their
habitus. Thus, the dynamics that exist between the objective structure
and the variations in the different habitus of the members that are present
in a modern production field will be present to a much smaller extent
in a craft guild in a traditional society. In other words, in traditional
societies, in which almost everybody acknowledges the same type of
symbolic capital there is less chances of internal change since there is less
variation in the habitus of the members of any congregation. Instead,
social structures will tend to be very conservative. Therefore, it may not
be appropriate to use the term production field in discussing prehistoric guilds. In this context, the term production fraternities will be
used to describe the traditional form of Bourdieus term production
field, i.e. it is a group of people that in some way produces symbolic
capital that is distributed unequally between its members. The similar
traits in the habitus of the memebers of a production fraternity will
make it far more conservative in its nature than the modern production
fields. A historical example of a production fraternity in which such a
conservative nature eventually led to a internal collapse as the result of
outside pressure is the production of Swedish iron works in Bergslagen
from medieval times up to the 19th century. In this almost secluded
community, the different crafts were passed down from father to son
112

The Emergence of Social Inequality


for generations (Per Lekberg, personal communication). These iron
works were, on account of their inherent conservatism, outdone by the
modern iron works of western Europe during the 20th century.

Apprenticeship systems
The most commonly known type of apprentice organisation is probably the medieval guild. The complexity of the medieval societies, as
regarded production and consumption, meant that artisans became specialised in producing different types of items that were needed for everyday life. Hence, bakers, butchers, tailors, blacksmiths, etc. were in
evidence in every little town (Pirenne 1936:179). In order to protect
both consumers and producers, it became necessary to create craft guilds
that guaranteed quality and controlled prices. It has been suggested that
this new type of organisation was a development of the fraternities that
appeared among associated craftsmen of different trades during the 11th
century (Pirenne 1936:181). Even though these fraternities had a charitable and pious aspect, they must also have functioned in a protectionistic
way (Pirenne 1936:181). It is important to note that these fraternities
were based on free association and that the introduction of the guilds
meant a strict legislation for the crafts. The guilds, as opposed to the
fraternities, had legal authority to force any artisan either to join the
guild or to close down his workshop. I shall not dwell on the subject of
the medieval guilds any further. However, they are good examples of
traditions becoming institutions as a result of an increasing moral
density. So even though the medieval guilds were political creations in
a fashion unlikely to appear in prehistorical contexts, they stemmed
from a tradition with roots far back in prehistory and well into preRoman times.
There are at least three different ways in which apprenticeship systems are known to have been organised in traditional societies: as guilds,
castes or as kin association groups (Clark & Parry 1990:334). The anthropological use of the term guild means considerably less than it
did in the history of the medieval guilds but seems to have more in
common with their predecessors, the craft fraternities. In anthropology
a guild is defined as an association of craftsmen that is not necessarily
based on kinship ties or on a causal association based on residential
groups (Clark & Parry 1990:334). Castes and kin association groups
differ from guilds in that they base recruitment on kinship ties. The
113

Jan Apel
degree of kinship differs for various groups it can be lineage- or clanbased, but caste organisations are distinguished from kin association
organisations by the fact that they are endogamous (Clark & Parry
1990:334). In some cases, when the craft is based on a limited rawmaterial source, the corporate forms of kinship-based, apprenticeship
systems may appear. These may consist of the people who are living
near the raw material and can control it.
Now, how may a prehistoric craft fraternity have been organised? It
has been suggested that the segment or caste, of the society that controlled the flint resources and the dagger technology formed a kind of
corporate group based on close kinship. The reproduction of the
technology took the form of an apprenticeship system and the selection of apprentices was based on an institutionalised, hereditary system.
In order to illustrate how an apprenticeship system may be organised in a traditional context, I have chosen the example of traditional
metal-working in Kenya. This tradition has been studied comprehensively and the following descriptions will focus on questions regarding
the importance of kinship in the reproduction of the technology.

Traditional metal-working in Kenya


This description is based on Jean Browns important work on metalworking in traditional societies in Kenya (Brown 1995). This book
contains material assembled from almost all of the Kenyan tribes and
field data were collected during a ten-year period, 1964-1974 (Brown
1995:ix). According to Brown, the iron-working outside the major
cities in Kenya has remained virtually unaffected by the modern laws of
monetary economy. However, the frequent use of scrap metal from
industrial production has made the smelting of iron obsolete and the
increasing amount of industrially produced, iron objects has reduced
the number of smiths (Brown 1995:x). The belief in traditional taboos
and curses has declined as a result of modern education and Christianity
and this has in turn made youths more inclined not to follow the craft
of their forefathers, since they no longer fear divine retribution (Brown
1995:x). The following description is basically a summary of Chapter
IV of Browns book, which, among other things, is concerned with the
training of smith apprentices and the hereditary aspects of the craft
(Brown 1995:119 ff.).
In Kenya, iron-working is a specialist craft that is hereditary and
114

The Emergence of Social Inequality


conducted by men. The other specialist craft in this region is potterymaking, which is generally carried on by women. The families involved
in the craft are never restricted to one clan, even in tribes where smiths
have low status. Smith families are present in most clans within a tribe
and this apprenticeship system is based on close kinship, i.e. it is a kinassociated craft (see above).
The craft is handed down from father to son, but not all sons are
thought to be suitable. Generally, only the most interested sons are
considered as apprentices, since they generally show most aptitude for
the work. In order to avoid over-establishment, only one son, usually
the eldest, is allowed to enter into an apprenticeship. Brown is a little
unclear about what would happen in a situation in which the most
interested and suitable son was not the eldest. It would be interesting to
know whether structure or practice would be dominant in such a case.
Generally the youngest son is considered if the oldest son does not learn
the craft and, curiously, youngest sons are considered to make the best
smiths. Brown claims that there is a tendency to choose the eldest or
the youngest sons in the family rather than in-betweens (Brown
1995:119). Usually, there is no place for more than three sons in a
smithy and if any other son insists on learning the craft he has to promise
to leave the area afterwards. In order to secure successors, some smiths
will put a curse on their sons, in order to get them to follow in their
fathers footsteps.
If his own sons are missing, a smith will take a son of his brother or
sister or a grandson as his apprentice. In the case of a brothers son, who
is not in the direct line of succession, a ceremony will have to take place
whereby the smithy is divided in two. The apprentice, and in turn his
descendants, will thereby be allowed to run the ancestral smithy after
the smith is dead. If no apprentices are available within the family, they
have to be taken from outside it. However, they must come from
families with smith forebears on either parents side. Usually, grandsons
or great-grandsons of smiths, chosen from among the sons of the smiths
best friends and age-mates, will be considered. If, for some reason, a
youth without smith ancestry is chosen, he can never have his own
smithy and can work in the craft only as long as the smith is alive. This
situation rarely occurs. However, sometimes a youth from a tribe without
smiths might become an apprentice in order to return to practice
amongst his own people.
115

Jan Apel
A smith whose master is still alive will often obtain his blessing
before taking on apprentices from outside of the family. Yet this practice varies from tribe to tribe. Sometimes it is prophesised that a woman
will bear a son destined to become a smith or that a young man with
distant smith ancestry has been chosen to be a smith. These apprentices
are more expensive than ordinary ones, since they have to be purified in
a ceremony in which they are obliged to provide animals, often a goat,
to be sacrificed. Their relatives are obliged to build a smithy, to which
objects from the ancestral smithy are brought, where they can be taught
in privacy. They will also have to pay a higher fee than an apprentice in
the close family does.
Usually, a smiths son starts to appear in the smithy when he stops
sleeping with his mother around the age of seven or eight years. However, serious training does not begin until the voice starts to break. If
the apprenticeship starts at this age, it will usually last eight to ten years.
When an apprentice is brought in from the outside, he is generally much
older and will receive intense training, lasting between three and six
months, provided that he is quick to learn and can pay the necessary
fees. If he is not good with his hands he may have to stay as an apprentice
for one to three years in order to acquire the necessary skills. It is
interesting to note that the apprenticeship system is flexible enough to
teach the skills to people with very different qualifications. In view of
the theoretical considerations presented in this chapter, one would suspect
that a person growing up around the smithy, other things being equal,
would appear to have a larger degree of natural ability. Tutoring fees are
usually demanded from all apprentices. Sons will pay smaller fees than
apprentices from outside the immediate family. The fees are paid in the
form of sheep, cows, hens, baskets of grain and honey beer. An apprentice
from outside the family may start his apprenticeship by giving the smith
a goat or sheep and liquor. He will then pay a hen during the first stage
of tuition, a goat or sheep during each of the second and third stages
and finally one or two animals and liquor during the ceremony of
initiation to the craft.

Stages of apprenticeship
In order to become a smith, the apprentice must go through three stages
of apprenticeship. Each of these stages includes certain taboos that must
not be broken. Usually, the apprentice deliberately overstep a taboo in
116

The Emergence of Social Inequality


order to notify and persuade his master that he believes himself to be
prepared to enter the next stage. After it has been ascertained why the
taboo was broken, he is fined and thereafter he is allowed to proceed
with his apprenticeship.
Stage 1. The first task of a new apprentice is to learn how to prepare
and look after the fire. An important aspect of this is to be able to blow
the bellows properly. He learns which wood will make the best charcoal
and how to prepare it. How to find ore and how to construct, load and
tend a smelting furnace He will also be taught all the taboos connected
with the craft. If related to the terminology introduced in Chapter 2,
this stage requires some recipes for action but a low degree of know-how
Stage 2. The task which requires the highest degree of practical knowhow in iron-working is to learn how to forge the iron. Consequently,
most of the apprenticeship is taken up in learning this. The second
apprenticeship stage starts when the apprentice is given a large stone to
be used as an anvil. Gradually he learns how to heat the iron to the right
temperature, beating on the anvil, forming simple tools like hoes and
axes and roughing-out preforms for the master to finish. He also acquires
skills in making handles and hafting the tools correctly. He also learns
phrases and repetitive songs that ease the monotony of the work and
provide a good working rhythm. This stage is characterised by
demanding a large share of know-how.
Stage 3. During this final stage, the apprentice learns most of the
secrets of the craft, for example, how to finish the more complicated
tools. In many cases these secrets take the form of what were earlier
(Chapter 2) referred to as recipes for action and it may thereby be
concluded that they often exhibit a large degree of theoretical knowledge. The apprentice is trusted to work on his own during the masters
absence. This stage demands a certain degree of both knowledge and
know-how.

Initiation into the craft


When the apprentice is fully trained and has paid his fees he is subjected
to a test in which he has to make good examples of the different products
whose manufacture he is supposed to have mastered. If successful, the
smith contacts his fellow smiths and an initiation rite is organised. In
order to be introduced into the craft, the apprentice has to have been
initiated into adulthood and in many cases he has also to get married
117

Jan Apel
and have children. Since smiths are often classed as elders and are expected
to function as counsellors, it is believed that only married men with
children can command the necessary respect.
The details of the initiation rite will vary from tribe to tribe but
usually follow the same general pattern. All the smiths in the neighbourhood are supposed to attend the initiation rite. In some cases smiths
will travel up to thirty miles in order to participate in the ceremony.
Usually, the rite starts with a practical demonstration of smelting and
forging by the master smith. He will then describe the progress of his
pupil. The apprentice will then demonstrate his skills in smelting and
forging, in order to show that he is worthy to join the assembly. The
highpoint of the ceremony is when the apprentice is presented with his
tools and particularly his hammer, which is the symbol of the craft.
The apprentice is then taught to curse and the remaining innermost
secrets of the trade, which he swears never to disclose. He is informed
that he will die if he does and it is understood that he must remain a
smith for the rest of his life. The oldest smith in the congregation blesses
the new smith. The sacrifice is made and must be eaten during the
ceremony. When a man is initiated as a smith the mystical power of
his smith ancestors, derived ultimately from god, is passes on to him to
enable him to smelt and forge iron successfully (Brown 1995:122).
The most powerful curses in the traditional societies of Kenya are
those uttered by a smith. Regardless of their status, which may vary
from tribe to tribe, smiths are always feared and respected. They are
regarded as being different from everybody else in society because of
the mystical powers they have inherited from their ancestors and from
god. Close contact with smiths or their products is believed to cause
ritual pollution as a result of innate and automatic curses.

The status of smiths


In Kenya, smiths generally have a higher status among farmers than
they have among nomadic pastoralists. In agricultural societies, the smith
plays an important role for the livelihood of the whole group. Iron
tools are essential for intensive agricultural work, weapons of iron are
necessary for hunting and defence, and iron ornaments keep people
healthy and prevent misfortune.
War further increases the importance of smiths, since they make the
necessary weapons.
118

The Emergence of Social Inequality


During times of agricultural expansion, smiths generally manage to
build up their wealth and prestige to a point where it becomes necessary
to maintain their positions by making the craft exclusive (Brown
1995:122). This is achieved by surrounding it with a veil of secrecy and
mystery, which in turn is preserved by a host of strict taboos designed
to keep people away, and by making use of deliberate and most potent
curses which smiths among pastoralists rarely need to use (Brown
1995:122). In agricultural communities, the smiths may use curses in
order to protect the property of the community, as well as that of the
smith. This would never happen in pastoralist communities, where curses
are only used in self-defence.
There are fewer smiths working among pastoralist peoples. In these
societies, there is a constant demand for iron products and there are not
as many opportunities to accumulate wealth as a result of an expansive
economy as there are in agricultural societies. Since manual labour is
regarded with disdain and contempt among most pastoralists, the dirty
craft of iron-working is despised and scorned. Smiths are also regarded
with contempt because they do not fight. The warrior herdsmen, who
gain prestige by prowess in fighting and raiding, tend to regard men
who do not fight as effeminate. However, it is important to remember
that even smiths with a very low status are potentially some of the most
powerful people in society. When indispensable to their society, smiths
can manipulate their mystical powers to their own advantage, maintain
the exclusiveness of their craft and attain the highest positions in society.
In all Kenyan communities, the families of the smiths form a separate
occupational group. They live in special areas on the edges of the villages.

Discussion and conclusions


Before embarking on the concluding discussion, I would like to explain why ethnographic data are useful to archaeologists and how such
data should be used in archaeological interpretations. An explicit use of
ethnographic analogies, especially analogies concerning activities carrying a large portion of the past in the present, may seem to contradict the
Durkheimian view on the reproduction of societies discussed above.
The relativistic view of culture that resulted from Durkheims view of
cognition and which is also maintained in Blochs view of the reproduction of the past in the present (see Chapter 3) seems to suggest that
comparative analyses are useless in this respect. Yet it is important to
119

Jan Apel
recognise the close connection between Durkheims view and the structuralism proposed by Claude Lvi-Strauss.
According to Lvy-Strauss, many cultural phenomena, such as rituals and myths, rules for kinship and descent, and settlement patterns,
are external expressions of the way in which the human mind structures
data. Accordingly, if this unconscious, general structuring of data can
be recognised beneath the obvious structure of a specific institution,
such as iron-working, it can be used to explain other institutions and
customs (Levy-Strauss 1958:28). In the following discussion, I shall
avoid picking out specific aspects of traditional iron-working in order
to use them as direct analogies with the Late Neolithic flint technologies of Scandinavia. Instead, I shall try to pin-point some principles of
technology which can be assumed to be more general in their character.
However, if Durkheim and Lvi-Strauss were correct in their assumption that there is no fundamental difference between primitive and
modern ways of thinking, the obvious implication is that ethnographic
analogy does not have any more explanatory value than a modern,
western analogy. So why not use modern, western examples instead of
traditional metal-working in Africa? As I see it, there are two reasons
for doing this. First of all, the scientific principle that guides all
conclusions based on analogy must be put forward. This principle simply states that the more common features that can be established between the object of analogy and the object of study increase the possibility that there are further similarities between them. Therefore it makes
sense to compare societies that share similarities in, for instance economy
and climate in order to understand the social structure. As we have seen
earlier in this chapter, Hayden has successfully used this strategy. Secondly,
I believe that the outside perspective that the ethnographic examples
provide gives huge advantages just because the archaeologists and
anthropologists are not born in the system and thereby stand a better
chance of understanding the underlying principles of the activity studied.
The detour over an unfamiliar tradition is necessary in order to break
up the deceivingly well-known relationship that we have to our own
tradition (Bourdieu 1999b:13). In confronting a foreign tradition, the
anthropologist cannot take anything for granted and the judgements
will to a lesser extent be based on what Searle refers to as the background
of this particular society. Bourdieu uses analogies to Berber communities to understand aspects of the French societies because he feels that
120

The Emergence of Social Inequality


these traditional societies display a concentration of the historical Mediterranean social structure, which, as regards the dominance of men over
women, still is present in the societies of the western world. Consequently, in this case, there is a historical continuity between the object
of analogy and the object of study (Bourdieu 1996).

Heredity
As we have seen rigid hereditary principles are followed concerning those
who are allowed to serve apprenticeships, as opposed to those who are
not. Some remarks indicate that these hereditary rules may be transgressed
if the person who is supposed to learn the craft does not show an aptitude
to do so. The fact that apprentices from outside the immediate family
have to pay higher fees also underlines the importance of heredity. In
iron-working, the question of talent is not that important, probably
since almost anyone will be able to master the practical know-how and
gain the theoretical knowledge involved in the technology. I hereby
suggest that a technology that includes a great deal of practical knowhow will be much more dependent on the motor abilities of the
apprentices. In such situations, the apprenticeship will last longer, since
motor skills are established in the body through repetition As I see, it
we shall have to accept that any natural ability, in the form of biological
heritage, will be distributed evenly between the members of any social
group. It has been suggested that natural abilities may have been decisive,
as regarded who would be considered as a flint-knapper and who would
not (Olausson 1998). This reasoning is appealing, since it is impossible
to avoid seeing that certain individuals find flint knapping easier than
others, especially while teaching students the basics of flint knapping.
However, I would like to add that, in most traditional crafts, hereditary
principles determine who is to follow the family tradition and who is
not. It may be difficult to judge whether, what we consider to be natural
ability, is dependent more on biological than on the social aspects and
vice versa. In most traditional cases, the combination of biological
heritage and social encouragement within family units is decisive, as
regards becoming talented artisans. This conclusion is actually in line
with ethnographic investigations, such as Barbro Santillo Frizells (2000)
study of the craftsmen that build the traditional, Apulian, dome-shaped
houses in southern Italy. This architectural tradition, based mainly on
the silent science of the hands (i.e. on know-how, as it was defined in
121

Jan Apel
Chapter 2), is transmitted vertically through the generations according
to hereditary principles. However, in cases in which the eldest son does
not show the appropriate interest or aptitude, another son will be chosen
to become an apprentice (Santillo Frizell 2000:75 f.). In a broader
context, it is tempting to regard prehistoric apprenticeship systems,
organised according to hereditary principles, as emerging because of a
strong need to keep natural ability within families, who in turn could
use this symbolic capital in order to gain and maintain social power and
prestige. Apprenticeship systems were thus created in order to guarantee
that as many as possible of those who were supposed to learn the craft
were able to do so, regardless of their intellectual and motor abilities. It
is important to remember that, in referring to natural ability one should
pay heed to the present social situation.

The importance of esoteric knowledge, ritual taboos and


prohibition
An important aspect of Browns description of traditional iron-working in Kenya is that the smelting process is the one that is surrounded
by the most taboos. The blacksmiths consider a strict observation of
taboos as essential, in order to succeed with the smelting process. The
ancestral smiths set up these prohibitions in order to secure the process
by keeping it pure and free from pollution. In short, this means that
everyone but the smith and his apprentices is excluded from witnessing
the smelting process. In other parts of central Africa, the smelting process
also seems to be surrounded with taboos and prohibitions. It almost
seems as if there is a general rule to severely restrict attendance at the
furnace during smelting. Among the Toro people of present-day Uganda
during pre-colonial and early colonial times, the smelting process was
also strictly regulated. In order to control those who were not allowed
to be present during the smelting, a stockade was often built around
the area where the furnace was situated, in order to warn other people
of the activity (Childs 1999:32). In the 19th-century kingdom of
Karagwe, now the north-western part of Tanzania, an iron-smelting
centre was located at Igurwa. It is obvious from archaeological finds
that this activity was also incorporated within a wider framework of
symbolism and social-power games. Here, the ritual aspects of the
smelting were very pronounced. Indeed, this became a present-day
problem when archaeologists tried to reconstruct a furnace and re-enact
122

The Emergence of Social Inequality


a smelting, because local participants and informants may not have
disclosed or elucidated all the secrets involved in the process (Herbert
1993:16 f.). Only men were allowed to help out with the smelting.
Women were excluded and, in order to avoid female influence, none of
the men involved in the smelting were supposed to have any sexual
relations during the smelting.
The taboos that surround the smelting should be contrasted with
the relative openness that is connected with forging. In the Toro communities, the forge is regarded as a social place, similar to a traditional
market and all that passed by were encouraged to pump the bellows
(Childs 1999:36). It is also perfectly clear from the description of ironworking in Kenya that forging, even though also surrounded by taboos
and prohibitions, is a more public affair than smelting (Brown 1995:71).
It is tempting to regard the taboos and restrictions as means of
controlling and manipulating knowledge in order to gain or maintain
power (Reid & MacLean 1995:145). I would like to remind the reader
that learning to forge is by far the most time-consuming aspect of the
Kenyan apprenticeship described above. As I see it, the reason for this is
that forging contain the greatest degree of practical know-how involved
in the making of iron objects in these societies. To learn to forge requires
long hours of practice, so that the muscles and senses involved in the
actions are co-ordinated. In contrast, the smelting seems to be an activity
that indeed requires a great deal of theoretical knowledge in the form of
complicated recipes for action. In line with the theoretical framework
presented earlier, I would like to propose that taboos should to a large
extent be seen as a way of keeping the production stages secret, in order
to maintain exclusive control over the technology. In some parts of
Africa, the inherent differences between the know-how needed to smelt
iron and the know-how needed to forge have indeed been appreciated.
In these areas, there is a general agreement that, while anyone can smelt
iron, only the smiths can forge it properly (Englund 2000:66). It is
important to remember that these conclusions do not imply that the
rituals that surround certain stages of the production appear to have
this meaning to those that practice the technology. On the contrary,
when born into a social system of this kind it will appear as natural and
be in perfect accordance with the habitus of the participants of the system.
It will become a form of embodied symbolic capital. In this context it
is necessary, so that there may be no misunderstandings, to remind the
123

Jan Apel
reader that there is no fundamental difference between traditional Ironworking in Africa and modern technologies today.

The production fraternity of the dagger-makers


The production fraternity that produced flint daggers functioned as a
religious congregation that served the well-being of its members, as
well as adjusting the group to other parts of society. It is suggested that
the fraternity was of what we would call a religious nature, including
common initiation rites, ceremonies and feasts and most likely a common burial practice. They would also have moral and ethical rules,
perhaps even formalised laws exclusive to the group. In this way, the
production fraternity of dagger-makers was organised very much in the
same way as the medieval guild. This also implies that the social mobility
within such groups would be severely restricted. In most cases, children
that were born into these groups would inherit an almost identical
habitus (Bourdieu 1996:60 f.). This contrasts to some extent with the
production fields of modern western societies where the individuals
share important aspects of the habitus, but not all, and were the
differences provide the dynamics for change. Heredity principles might,
as in the case of the traditional ironsmiths in Kenya, have been put
aside, in cases in which no suitable apprentices were available. If, for
instance, a flint smith did not have children of his own, we suspect that
apprentices were brought in from outside the lineage. In such cases, just
as in the Kenyan example, they would have to pay a higher fee. This
would be an opportunity for an individual from outside the families
that controlled the dagger-making to gain symbolic capital that could
be used in, for instance, future marriage negotiations. In a prehistoric
production fraternity, common interest would always come before individual interests. This description of the field corresponds in large parts
to Durkheims suggested solution to the anomity problem in modern
secularised societies (Durkheim 1984:xxxii ff.). According to Durkheim,
the modern, secularised, organic society has made people insecure, since
convincing guidelines no longer exist, in the form of religious ideas that
can be relied on for moral guidance. Durkheim believed that corporate
companies should be organised in a way that would give their employees
ontological security. Of course, such projects were seldom initiated,
although some Japanese corporations seem to have a similar kind of

124

The Emergence of Social Inequality


organisation. In other words, they should form an organ within the
social body in which all the needs of its members are taken care of. It
must therefore follow that the production fraternities, in traditional
societies, still with a certain degree of mechanical solidarity but with an
incipient social complexity, such as complex, segmental, or chiefdom
societies, gave their members ontological security in a way that most
modern production fields can never do. This may explain why people
originally accepted social hierarchies. In Chapter 10, a more detailed
interpretation of the Late Neolithic social structure will be presented.

125

Chapter 5

Previous Research on
the Determination of Debitage
from Bifacial-reduction Sequences
In Chapters 6 and 7, Late Neolithic sites in southern Scandinavia and
eastern central Sweden with flint will be studied and discussed in relation to my main concern, i.e. to see whether it is possible to recognise
different production stages between and within sites. Such findings
would shed some light on the ways in which daggers and other artefacts
in southern Scandinavian flint were distributed, these finding could
also be used to understand the character of the Late Neolithic political
landscape and the organisation and complexity of the social structure.
The Late Neolithic, flint technologies of southern Scandinavia are mainly
based on bifacial-reduction strategies. The presence of bifacial-reduction debitage on sites in southern Scandinavia habitually used to date
them to the Late Neolithic period. Hence, it is appropriate, as a starting-point, in this context, to describe the formal ways in which production debitage from bifacial-reduction sequences can be distinguished
from other flake types in archaeological contexts. These kinds of flakes
are actually not difficult to recognise; any experienced archaeologist or
lithic analyst will immediately recognise flakes from bifacial-reduction
sequences when such flakes are recovered in the field. In spite of this,
the formal ways in which these flakes can be distinguished from other
types of flakes, such as flakes from unifacial platform cores or the production of square-shaped axes needs to be evaluated. In my specific
discussion of the bifacial technology, in relation to the idea of craft
specialisation that will follow later on, there is a need to be able to
distinguish flakes from different stages in the production of flint daggers. If this can be done, it will be possible to understand and designate
the production debitage from different production stages between and
within sites. Consequently, I am not interested solely in identifying and
separating bifacial flakes from other types of flakes but also in being
127

Jan Apel
able to classify them to different stages of the production. Thus, in this
chapter, in preparation for the analysis of original production debitage
in Chapters 6 and 7, earlier attempts to classify debitage from bifacialproduction sequences are referred to and discussed.
The following paragraph summarises some elementary concepts in
the art of flint-knapping that the reader may find useful to know in
reading the rest of this chapter. During flint-knapping, flakes are struck
from larger flint pieces (cores) with the help of hammerstones or antler,
wooden or metal billets. This technique is in this context referred to as
percussion technique. In some cases, a punch, in the form of an antler
or copper billet, is placed on the platform of the original core and is
positioned between the hammerstone and the core. This technique is
referred to as indirect percussion. If the percussion hammer is of a harder
material than flint, for instance, a harder variant of quartzite or metals
such as copper or steel, the technique is referred to as hard percussion
technique. If the percussion hammer is made of a material that is softer
than flint, such as sand and limestone, wood or antler, the technique is
referred to as soft percussion technique. Flakes can also be removed by
pressing a bone, antler or copper-tipped, pressure flaker on the edge of
the platform. This technique is referred to as pressure technique (Pelegrin
1991, Madsen 1996, K. Knutsson 1988a).

Lithic-production stages
In 1972, Don E. Crabtree defined the term method in flint-knapping
as ... the logic manner of systematic and orderly flaking process, or the
preconceived plan of chipping action based on rules, mechanics, order
and procedure (Crabtree 1972:2). Among flint workers and lithic analysts, method is the phenomenon that verifies the idea that the production of knapped artefacts is not haphazard but carefully planned and
answers to specific needs formulated before the actual flint-knapping
begins. In other words, The shape, length, width, thickness, form and
technique of applied force to fashion the tool was predetermined by
the toolmaker before the initial fracture of the raw material (Crabtree
1972:2). The preconception of the goal (a flake, core or tool with characteristics) that the flint worker has in mind before the actual knapping
begins can be referred to as a form of mental template.
Technique is the way in which the force is applied to the core during
the removal of each flake. It may be direct or indirect, done with a hard
128

Previous Research on the Determination of Debitage


or soft hammer, etc. (Crabtree 1972:2). Manner is the mode of
preparation and technique used to form the artefact according to a
method, i.e. all those decisions, as regards, for instance, platform preparation, determined impact angle and application of force, that will affect the result of a chosen technique (Crabtree 1972:2). A rather curious consequence of these labels is that the term manner always includes technique, while technique, on the other hand, can be described
independently of manner and method.
In later works, the concepts of manner and method tend to be
combined under the term method. Accordingly, the production
debitage of any technological strategy can be understood as being a result
of technique and method (Pelegrin 1981, Madsen 1986, K. Knutsson
1988a). These two concepts, together with the natural constraints of
the flint, will determine the appearance of a flake. Technique refers to
the way in which the power is transmitted from the knapper to the raw
material. Method, on the other hand, refers to the choices of techniques, impact angles and working positions, as well as the intentions
of the knapper during the production process. The relative importance
of these concepts, as regards the appearance of the production debitage,
has been extensively discussed in the literature (Bradley 1978, Patterson
& Sollberger 1978, Pelegrin 1981, Patterson 1982, Hayden & Hutchings
1989, Dibble & Pelcin 1995, Pelcin 1997 and Bradbury & Carr 1999).
In recent years, it has become apparent that the choice of method plays
a greater role in determining the formal shape of the flakes than was
previously realised. It has, for instance, been acknowledged that experienced flint workers can produce flakes with similar attributes by different techniques (Callahan, personal communication). The obvious
conclusion of these findings is that the intentions of the flint worker,
which are a vital ingredient in the method concept, will determine not
only the morphology of the flakes but also their technique related attributes (Patterson & Sollberger 1978, Patterson 1982, Pelcin 1997,
Bradbury & Carr 1999).
The term stage is important for the process of classification and interpretation of technology in this thesis because it is used to relate the
production process to different levels of knowledge and know-how.
However, it is defined in different ways in the papers referred to in this
chapter. Usually it is thought of as a major change in technique or manner
(Whittaker 1994:201). In order to be useful to the lithic analyst, this
129

Jan Apel
change has to be great enough to affect the appearance of the debitage
in a recognisable way. That is, a number of variables that define and
isolate flake types from different production stages have to be present
in the debitage. Some authors argue that a natural stage occurs when a
stage defined in this way coincides with the completion of the mental
template that is needed to secure a continued reduction (Callahan 1979).
Other authors prefer to work with definitions that are independent of
any technological changes and instead are based on an artificial division
of the reduction sequence, i.e. experimental materials are divided into
four, equally large sections. The former way of working is demonstrated
in the following examples.

Identifying Stages, Example 1


During the reduction of a biface, several changes in manner will be
made in order to secure the success of the continued reduction. One of
the more significant of these changes appears at the transition between
the rough-out stage and the primary preform stage (for definitions of
the production stages, see Fig. 2:3). During the rough-out stage, the
objective is to create an edge around the preform. This requires concentration on the outer zone of the preform, and the appearance of the
flakes, as well as the preform, will be affected by this decision. The
manner of this operation will to a large extent involve behind-theedge flaking (Fig. 5:1), i.e. the impact point of each removal will ideally be located behind the margin of the preform. Flakes removed in
this manner are referred to as non-marginal flakes (Bradley 1978:8 ff.).
They tend to be wide at the expense of length and will display a relatively thick, platform remnant. During this stage, platform preparation
is redundant, since the point of impact is located away from the margin
of the preform and the platform will survive the impact owing to its
relative thickness. These flakes will not display any significant platform
preparation.

Identifying Stages, Example 2


At the transition between the rough-out stage and the primary preform
stage, the mental template, and consequently the manner, will have to
change. During this new stage, it is important to create a biface with a
lenticular cross-section and without convexities, major humps or ridges
(Callahan 1979 & 1996). To achieve this, flakes have to reach into the
centre of the preform and make contact with flake scars from flakes
130

Previous Research on the Determination of Debitage

Fig. 5:1. Outline drawing of marginal (on the edge) and non-marginal (behind
the edge) flaking and typical flakes.

removed from the opposite margin. It is reasonable to suspect that this


decision to make the flakes longer, in order to reach into at least the
middle zone of the biface will affect the appearance of the flakes in
a way that will distinguish them from flakes removed during the previous stage. For instance, in order to produce long, thinning flakes, it is
important that the point of impact shall be located on the margin of
the preform (Fig. 5:1), and this will require extensive platform preparation. As a result, the flake will, hypothetically, have a thin, small, platform remnant and display platform preparation on the dorsal face of
the platform remnant. In this way it can be argued that the production
debitage will change in character when the mental template of the flint
worker alters and a change in manner is applied. It is, however, crucial
that the variables that reveal these changes shall be identified and defined in the debitage.
If the change in manner that occurs at the transition between two
stages also involves a change in technique, for instance, a change from a
hard hammerstone of quartzite to a antler billet, the flakes from each
stage will theoretically be distinguished from each other. It will be possible to distinguish them not only by their shape and size but also by
the attributes visible on the platform remnants and ventral surfaces of
the flakes, owing to the different characteristics that the two hammers
produce (pronounced bulb/lipped flake, ring cracks etc). It is important to remember that the flint worker does not have to be conscious of
the stage concept. In prehistory in general, and perhaps also during the
Late Neolithic in southern Scandinavia, it can be suspected that knapping
skills were reproduced down the generations as technological conventions
(K. Knutsson, in prep.). The stage concept should be regarded as an
131

Jan Apel
analytical tool developed by modern flint workers, in order to formalise
the learning process, and by lithic analysts, in order to tidy up the debris
on prehistoric knapping floors for the purpose of studying cognition or
technological specialisation (K. Knutsson, in prep., 15 ff.). In this
context, the discussion as to whether or not every flint worker is aware
of production stages is not relevant. Stages, like typologies, are often
partly subjectively constructed in order to solve specific questions.
In my experimental study of the type-IV daggers, it has been of vital
importance to evaluate how lithic-production stages relate to the production debitage. If this relationship can be established, it will open up
possibilities of studying potential production sites and discussing the
degree of technological specialisation and organisation of dagger production in relation to the organisation of society in general: ... the
flintknapping techniques are culturally learned behaviors. They are highly
variable through time and space, since specific lithic reduction processes
are affected by different variables in response to cultural needs (Yerkes
& Kardulias 1993:93).

Earlier attempts
The idea that flint tools were produced in natural stages is old and has
been used in archaeological interpretations since the late 19th century.
The American archaeologist W. H. Holmes conducted flint-working
experiments in order to find ways of distinguishing between implements (finished tools) and preforms in the Virginia Tidewater quarries
(Johnson 1978:340). Holmes used five criteria for the identification of
implements: (1) degree of elaboration, (2) indications of specialisation,
(3) signs of use, (4) manner of occurrence and (5) association of other
articles (Holmes 1894:123). These criteria had to be taken into consideration before an artefact could be regarded as an implement rather
than as some kind of preform.
Holmes suggested that, in order to evaluate prehistoric stone tools,
especially those found in quarries, their history had to be studied through
the process employed in their manufacture (Johnson 1978:341). In
fact, as early as 1894, he had argued for an approach similar to the
operational-chain studies of today (Holmes 1894:122, Sellet 1991).
He was able to demonstrate that there were significant similarities between crude implements and early preforms for more elaborate tools
and presented different production stages from blank to finished pro132

Previous Research on the Determination of Debitage


jectile points (Johnson 1978:342): It will be seen that the course of
procedure in the simplest shapes is repeated more or less closely in the
earlier steps of the more elaborate shapes and that the divergence of
specialisation takes place towards the end of the process in each case
(Holmes 1894:128 f.).
Holmes used this argument and was able to demonstrate that crude,
handaxe-like bifaces were discarded preforms for more elaborate artefacts. Working within an evolutionary framework, he was convinced
that the preforms must thus be later than the European handaxes and
that they did not constitute evidence for Palaeolithic activities on the
American continent a subject that was discussed at the time. It had
already earlier been recognised, by experimental archaeologists, that the
various edge treatments on knapped artefacts were not necessarily
decorations or functional edges on finished implements but might be
by-products from the manufacture of implements (Johnson 1978:339).
Thus, the presence of retouch on the edges of preforms could be seen as
the result of edge preparation and not as evidence of finished implements. The most important implication of Holmess studies was that
the length of American prehistory, based on the available archaeological
evidence, was considerably cut down since the arguments for the presence
of humans on the continent during the Lower and Middle Palaeolithic
were dismissed (Willey & Sabloff 1973:58 f.). The fact that Holmes
worked within a strict evolutionary perspective seems to have deprived
him the attention that he deserved.
In 1929, the Swedish archaeologist Folke Hansen argued, in a similar way to Holmes, that a particular kind of almond-shaped artefact,
earlier interpreted as Palaeolithic handaxes, were in fact preforms for
Neolithic bifaces, such as daggers, spearheads and sickles. The main
reason for this interpretation was that the crude bifaces were found at
stra torp in Skne where obvious preforms for Neolithic artefacts,
such as daggers and spearheads were also found (Hansen 1929:248 f.).
Hansen suggests that the site was a flint workshop dating from the later
part of the Stone Age. Observations like these made it apparent that the
production of complicated stone artefacts was conducted in stages that
could be recognised, rather than as a fluid continuum, and that discarded
preforms for more elaborate, bifacial tools could be similar to cruder
tools. It was also recognised that flint-knapping required a great deal of
cognition and, as a consequence, an ability to think in several stages in
133

Jan Apel
order to convert a complicated mental template into flint.
In 1975, Bradley (1975:7) argued that carefully controlled, experimental analogies were probably the most useful way of gaining understanding of lithic-reduction sequences for different, prehistoric-tool types.
Such an experimental procedure would accordingly involve a careful
examination of the production debitage, the techniques and methods
used, and the sequence of reduction. Bradley suggested that, once the
reduction sequence had been reconstructed, it could be divided into
stages and steps of manufacture (ibid.). These stages were (1) preliminary
modification, (2) blank and (3) preform. He maintained that the most
difficult task for the analyst was to demonstrate intention, and suggested
that the modified, non-implement, non-core category was to be used
in this process, i.e. the production debitage.
Within this category there may be typological groups identifiable by
clustering of morphological and technological attributes. Once these
groups are defined, the nonmaterial evidence (recovery associations)
may be examined to determine whether or not there is a high correlation between the two. In an archaeological context this is most likely to
be encountered in the form of caches or closely associated (spatially)
groupings. If there is a statistically valid correlation, it may be inferred
that these groups represent intended stages. These stages may then be
classified as preliminary modification, blank, or preform (Bradley
1975:7).

Bradleys suggestion, that it was possible to demonstrate the intention of the prehistoric flint worker if the right variables were documented in the production debitage, was followed by several archaeologists during the following years. Many of these analyses concerned
worked, bifacial-reduction sequences, and I shall recapitulate some of
these in the following pages.

Recent attempts
Skilled flint workers and lithic analysts can recognise and differentiate
between bifacial-thinning flakes and flakes produced during unifacial
reductions. It is also possible to argue that certain flakes stem from
early or late production stages. This way of classifying debitage stems
from experience and is thereby intuitive and not formalised. It can be
considered as emanating from an emic approach in which reason is based
134

Previous Research on the Determination of Debitage


on experience (Yerkes & Kardulias 1993:94). An excellent example of
the intuitive classification of production debitage from bifacial reductions
is the system worked out by Jeffrey Flenniken for the Flint-knapping
Field School. This system has been used on lithic assemblages from
western North America and Australia (Flenniken 1984) and is presented
in Fig. 5:2 and Table 5:1. Using this system, it is possible to assign
flakes stemming from bifacial-reduction sequences to the reduction
stages presented in Chapter 2. Primary and secondary decortication flakes
can arguably be assigned to the rough-out stage, early-stage, bifacialthinning flakes to primary thinning, and secondary thinning flakes to
the secondary or tertiary thinning stage. Pressure flakes, in turn, stem
from the last stages when the blade surface and the edges of the blade
are given their final form.
During the 1970s and 1980s the emic approach, based on experience rather than on scientific experiments, was heavily criticised by
processual archaeologists. It was argued that assumptions regarding an
archaeological observation had to be objectively evaluated, that is; the
arguments used for warranting the meaning given to observations must
be intellectually independent of the argument being evaluated through
an appeal to the meaning of observation (Binford 1982:53). It was,
according to Binford, not sufficient to understand lithic materials by
experience, as such an understanding is exclusive to those who have a
similar degree of experience and has to be taken for granted by others. It
was argued that such folk knowledge (Binford 1982:48) would only
be apparent to those sharing the same frame of reference. In responding
to these new ideas, lithic analysts felt the need to formalise their emic
knowledge. Working within the new academic environment, it became
crucial to be able to quantify any results. This was achieved by controlled
experiments and measurements explicitly described. Measurements of
flakes were indeed intellectually independent of the idea that staging
could be recognised in lithic debris. Even though this development
was not entirely positive, it forced the analysts away from a Yuppie/
Guppie perspective in which lithic specialists were able to see the truth
solely by being the most experienced (Binford 1989:6 f.). Instead, a
modernist, anti-authoritative perspective, in which experiments and
measurements could be replicated and tested by others, was preferred.

135

Jan Apel

Primary Decortication Flakes

Alternate Flake

Edge Preparation Flake

Secondary Decortication Flakes

Early Bifacial Thinning Flakes

Interior Flakes

Late Bifacial Thinning Flakes

Margin Removal Flakes


Early

Late
Pressure Flakes

Fig. 5:2. Some examples of technological debitage categories concerning


bifacial-reduction sequences after categories developed by Jeffrey Flenniken at
the Flintknapping Field School. Drawing by Lotta Sundeberg (after Yerkes &
Kardulias 1993:96, Fig. 2).

Newcomer 1971
In 1971, the results of quantitative experiments in Palaeolithic, handaxe
manufacture were presented by Newcomer in World Archaeology. One
of the goals of these experiments was to discuss the morphology of the
waste flakes produced during handaxe manufacture in relation to stages.
Initially, Newcomer subjectively defined three different production
stages:
1 A rough-out stage in which the nodule is bifacially flaked in order to
obtain an edge all round. This is achieved by using the flake scar
136

Previous Research on the Determination of Debitage


Flake type
Primary decortication flake
Secondary decortication flake
Interior flake
Early-stage, bifacial-thinning flake

Late-stage, bifacial-thinning flake

Alternate flake
Edge-preparation flake
Margin-removal flake
Early-stage, pressure flake

Late-stage, pressure flake

Definition
Produced during the initial removal of cortex. These flakes have
cortex over their entire dorsal surfaces.
Produced during final stages of cortex removal. Cortex is present on
part of their dorsal surfaces.
Flakes removed from the interior of the core or cobble. There is no
cortex on their surface.
Produced by percussion during the primary reduction of a biface (as
a result of making the biface symmetrical). They have few dorsalsurface scars, are slightly curved or twisted in longitudinal section,
have single-faceted or multi-faceted platforms and are usually the
largest thinning flakes produced during biface manufacture.
Produced by percussion during the final stages of roughing out a
biface. They have numerous scars on their dorsal surfaces are almost
flat in longitudinal section, usually exhibit a feathered termination
and have multifaceted platforms.
Produced as a result of the creation of a bifacial edge from a square
edge on a piece of stone. They are much wider than they are long and
are triangular in cross-section.
Removed from the edge of a flake blank in order to prepare the blank
for further reduction. The dorsal surface of the flake blank serves as
the platform for these flakes.
Half-moon-shaped fragments of bifacial edge produced when the
biface is struck too hard and too far from the margin.
The first series of pressure flakes removed when bifaces are being
finished. These small flakes have multiple scars on their dorsal
surfaces and are twisted in longitudinal section, and their platforms
form an oblique angle with the long axis of the flake.
Produced during the final episode of pressure flaking on a tool. They
are small, parallel-sided and slightly twisted in longitudinal section
and have multifaceted, abraded platforms and a single dorsal arris.

Table 5:1.Verbal description of some technological-debitage categories concerning bifacial reduction developed by Jeffrey Flenniken at the Flintknapping
Field School (after Yerkes & Kardulias 1993:97, Table 1).

after each removal as a platform for the next. No attempt is made to


thin the nodule at this stage and large areas of the nodule may therefore have remains of cortex on both surfaces.
2 A thinning and shaping stage, in which the goal is to remove long
flakes travelling at least halfway across the surface of the preform
and thereby to thin it down and remove any remaining cortex and
bumps.
3 A finishing stage, in which the edge of the preform is finally shaped
and straightened. During this stage, the goal is to remove small, thin
flakes which usually go no further than halfway across the surface.
Newcomer subjectively described the flakes from the first stage as
being relatively consistent in their appearance; they tended to be thick
and had varying amounts of cortex on their dorsal surfaces. The platform remnants (Newcomer refers to them as butts) were wide and
plain (i.e. not facetted). Bulbs and percussion cones on the ventral sides
of the flakes were well developed.
137

Jan Apel
Flakes from the second production stage were regarded as characteristic, and Newcomer also pointed out that flakes like these are known
in numerous examples from sites belonging to Acheulian and Mousterian
traditions. They were described as typically thin with all edges feathered. Any undulations on the ventral surfaces are poorly marked, and
they also display negative flake scars of other, similar flakes on their
dorsal surfaces. Newcomer also suggests that flakes from this stage, if
they were removed with deer antlers, would display a lip on the ventral
side of the platform remnant (Newcomer 1971:88 f.).
Flakes from the third production stage closely resemble flakes from
the second stage, but are smaller. Newcomer points out that the flakes
from the second and third stages have often been lumped together as
hand-axe finishing flakes or eclats de taille in earlier literature. This
type of flake has been referred to as thinning flakes (Crabtree 1972) or
bifacial-thinning flakes (Flenniken 1985). It seems as if Newcomer was
of the opinion that the flakes from the second and the third stages
could not be distinguished in any other way than by size. This implies
that the change in manner is not significant enough to show in the
qualitative variables that he used. It should be emphasised that these
variables are not accounted for in the article. In other words, there is no
way of judging their credibility.
During the experiments, three handaxes were made and each flake
removal was weighed. This data was processed and the results revealed
that flakes gradually became smaller as the reduction sequence continued. Newcomer also concludes: As indicated in the figures, these curves
seem to be divided into three distinct areas, each roughly corresponding
to one of the stages of manufacture subjectively defined above
(Newcomer 1971:93).
The results of the experiments suggested that the three different stages
inherent in Newcomers production of handaxes could be defined in
the debitage in two ways: by qualitative classifications of flake attributes,
and by quantifying the weight of each principal flake. In this context, a
principal flake is a flake that is the result of planned action, i.e. cases in
which the knapper knows beforehand where to strike and what will be
achieved by this stroke. Edge-preparation flakes can seldom be considered
as principal flakes, since they are not predicted in the same exact way as
principal flakes are. However, it should be stressed once again that the
flakes from stages 2 and 3 could not be distinguished by any other
138

Previous Research on the Determination of Debitage


means than by weight. Since the weight of the flakes from different
stages is dependent on the size of the preform, it is not a good variable
to use on prehistoric materials from sites where it is suspected that
multiple reduction sequences are present.

Nichols and Callahan 1980


In 1980, Nichols and Callahan published the results of experiments
aimed at demonstrating production stages in a quantitative analysis of
debitage from three, bifacially-flaked, Lecroy projectile point replicas.
Even though artefacts of this kind are significantly smaller than daggers
or handaxes, the problem of staging in relation to debitage is similar to
that with larger artefacts and no less complicated. In order to succeed,
the flint worker reduces the blank or flake in stages similar to those of
larger artefacts and thereby adjusts his aims and techniques accordingly.
During the experiments, Callahan worked in natural stages. The
transition between the stages was defined as a point in the production
where a significant objective (mental template) was achieved that in
itself provided the necessary conditions for the next stage. According to
the figures in the article, five natural stages were defined. These stages
are not verbally defined, but, judging by the figures, they seem to coincide
roughly with the five stages involved in the production of fluted points
presented by Callahan in The Basics of Biface Knapping in the Eastern
Fluted Point Tradition (1979 & 1996).
In this paper, it is proposed that patterns related to stages in the
production of bifacial projectile points can be recognised in the debitage
according to the following fact: There is a necessary sequence of manufacture in a complex tool: the edge must be prepared, the biface must
be thinned, etc. (Nichols & Callahan 1980:12). Since an important
objective was to investigate whether these stages could be defined in the
debitage, the length and width of each major flake with a length over 2
mm were measured. Two tests were then conducted on the data. Mean
values of the variables were subjected to a standard chi-square test, in
order to reveal whether there were any significant differences related to
the stages in the debitage or whether the data merely reflected a continuum. After this initial investigation, the stage hypothesis was tested
by an investigation of the standard deviation in relation to the central
tendency of the flakes in each stage. Such a procedure will define any
stages by their relative homogeneity (Nichols & Callahan 1980:13).
139

Jan Apel
The tests were apparently guided by earlier empirical observations made
by the authors. They knew that flake sizes were repeated within stages,
not stage to stage, and we proposed that it would be possible to
predictively assign individual flakes to stages by their sizes, and to
characterise stages by the redundancy of these flakes (Nichols &
Callahan 1980:13).
The tests suggested that there were statistically significant differences
between the mean values of the length and weight variable from the
different stages. A decreasing variability per reduction stage was also
discovered, i.e. the flakes tended to be more and more regular in size
(and shape) per reduction stage. It was also concluded that there was a
non-repetition of flake size within the experiments, i.e. the flakes became
smaller.

Magne and Pokotylo 1981


In 1981, Magne and Pokotylo published the results of a set of experiments aimed at determining whether or not experimentally defined,
lithic-reduction stages could be inferred from debitage attributes (Magne
& Pokotylo 1981:34). They remarked that earlier studies had demonstrated that exhaustive variable lists tend to be redundant, and that
multivariate statistical techniques can be used to isolate a few independent variables. They also concluded that these earlier studies were not
sufficient, since the variables used had been obtained from, and then
applied to, archaeological materials. The fact that these earlier studies
had been conducted purely within static archaeological contexts had
prevented the authors from studying the dynamic behaviour behind
the static remains. Magne and Pokotylo realised the need for results
created under controlled forms, in a systemic context, where flakes could
be directly linked with the actions that caused them: What is presently
desired is the objective testing of a limited number of variables in a
controlled experimental situation, where debitage variability can be
directly related to reduction stages (Magne & Pokotylo 1981:34).
A model of the ideal mental template of each manufacturing stage
was created through the study of fourteen bifaces and biface fragments
collected from known prehistoric sites in Upper Hat Creek Valley. Some
of these were obviously preforms of different stages, while others were
complete artefacts. This seems to be a dubious way of working, since
abandoned preforms found on a site can hardly be interpreted as ideal
140

Previous Research on the Determination of Debitage


production stages. The simple fact that they were abandoned implies
that they did not meet the specific demands that were required if the
preforms were to be considered for further reduction. As a preparation
for the experiments, 16 spalls, suitable as blanks for bifaces, were knapped
from a basalt nodule. One of these blanks was reduced to a biface with
properties similar to the prehistoric bifaces. All the flakes were saved
and documented (Magne & Pokotylo 1981:35). The problem with
the study is that the stages were defined within the archaeological material
with all the complications that this leads to (see above). The only
reasonable way of demonstrating possible stages in experiments is for
the flint worker to work from a finished implement and then to study
how stages can be defined within this work.
Initially, Magne and Pokotylo registered five continuous variables in
their experimental debitage: weight, maximun length, maximum width
perpendicular to maximum length, dorsal scar count, and degree of
cortex cover. In addition, three ordinal variables that could only be
registered on debitage with platforms, so-called platform-remnant-bearing (PRB) flakes, were used: platform width, dorsal-flaking angle (angle between platform remnant and dorsal face), and platform-scar count
(i.e. the number of facets on the platform remnant). The data were
analysed by multi-statistical methods. The results indicated that length,
width, platform width and dorsal angle were without significance in an
analysis that also included the following and subsequently more important variables: weight, dorsal-scar count, platform-scar count, and
degree of cortex cover. The experimental debitage was then subjected to
a statistical test, aimed at evaluating the utility of the last four variables,
i.e. to see if these variables would be sufficient to reveal sequential
patterning in the debitage. It was concluded that these four variables
could indeed be used to describe sequential patterning within the
debitage. In accordance with these results, Magne and Pokotylo suggested that one of the most important findings from the experiments
was that different kinds of production debitage might be distinguished
on the basis of dorsal-scar counts. Another finding was that size and
scar counts of all debitage are useful in trying to infer lithic-debitage
patterning ( Magne & Pokotylo 1981:38). The article demonstrated
that it was possible to describe debitage variability in terms of reduction sequences with few variables.

141

Jan Apel

Holm 1985 & 1991


Holm approaches debitage studies from a slightly different angle than
the scholars referred to in the previous pages. Her aim was to investigate whether and how the lithic debitage from Grsvattnet VI, a mountain site in northern Sweden, varied over time. Radioacarbon datings of
the site range between 6990115 BP and 3090180 BP, suggesting
that it was inhabited from the Mesolithic up to the Bronze Age (Holm
1985:304). What concerns us here is how flakes from the production
of bifacial arrowheads from the Late Neolithic and Early Bronze Age
differ from flakes stemming from Mesolithic technologies, such as flakes
from the production of handle cores. The raw materials used on
Grsvattnet VI were of local origin. However, the raw material that was
the subject of Holms investigation, Breccia quartz, is a kind of plastic,
quartzite-like material with flint-like properties. A random flake sample
of 10% (2 241) was collected from six different levels of the cultural
layer on the site (Holm 1985:305). The following investigation was
initially not based on her own experiments, but she stated that it relied
on the present knowledge of lithic technology (Holm 1985:303).
However, in later works, the results of experiments conducted by an
experienced flint knapper, Harm Paulsen, were compared with the
archaeological results (Holm 1991:71).
The following variables were documented on the flakes:
1
2
3
4

Flake form: (a) platform flakes, (b) no-platform flakes and (c) others.
Flake width
Flake thickness
Cortex flakes: (a) primary cortex flakes, (b) secondary cortex flakes
and (c) no-cortex flakes.
5 Platform length
6 Platform width
7 Lip: presence or absence of lip on platform remnant.
Holm concludes that a clear pattern is discernible in the material
from Grsvattnet VI (Holm 1985:307 ff.). The width-thickness relationship between flakes from the upper part of the stratigraphic sequence (the A-horizon), on the one hand, and the lower part of the
stratigraphic sequence (the cultural layer), on the other, differs consid-

142

Previous Research on the Determination of Debitage


erably. The flakes in the A-horizon have a higher width-thickness ratio
(6.5) than those in the cultural layer (between 4 and 3.1). The flakes
from the A-horizon tend to have lips (the result of a soft reduction
technique) and are significantly thinner than flakes from the cultural
layer. The end shape of the flakes in the A-horizon tends to be featherlike, while the end shape of the flakes in the lower levels is not so
characteristic in the cultural layer. The flakes in the A-horizon display a
low degree of cortex, while such flakes are more frequent in the cultural
layer.
Holm thus concludes: In the A-horizon, dated to the Bronze Age,
the bifacial technique produced thin and wide flakes with small platforms. The production was extensive with a selected raw material, decorticated outside the site. There was a relatively large element of soft
technique in the thinning phase (Holm 1985:309).

Amick et al. 1988


An ambitious attempt to relate debitage attributes to the sequence of
debitage removal in an experimental context was published by Amick
et al. in 1988. The experiments consisted of the production of a bifacial
core. In line with processual archaeology in general, they argued that
the traditional methods of inference used by archaeologists had been
intuitive and implicit, and that many papers dealing with debitage analysis presented unsupported inferences regarding activities etc: The long
history of flintknapping experimentation has produced a considerable
amount of folk knowledge about the process of lithic tool manufacture. While this folk knowledge can guide experimental studies designed
to evaluate inferences, it cannot be used to describe the archaeological
record until rigorously verified (Amick et al. 1988:26). The data used
in the study consisted of 325 flakes and flake fragments created by 101
percussion blows during the reduction of a single nodule of high-quality Georgetown chert. The nodule was, according to the mental template of the flint worker, reduced to a bifacial core. In Table 5:2, the 13
variables used are presented. These data were initially subjected to different univariate and bivariate statistical tests in order to determine which
variables were appropriate for a multivariate, discriminant analysis. The
reduction sequence was divided into four stages with even numbers of
flake removals in each: stage 1 = 1-25, stage 2 = 26-50, stage 3 = 51-76,
and stage 4 = 76-101 (Amick et al. 1988:27). The grounds for this
143

Jan Apel
division are not presented in the article; however, it is obvious that
these stages do not reflect the intention of the knapper. It is an arbitrary
division and consequently not related to intention. This can arguably
be considered as a major weakness in the study, since changes in the
production debitage as a reflection of a change in manner cannot be
studied. On the other hand, owing to the fact that the subjective element of the experienced knapper or analyst has largely been omitted,
this kind of study will perhaps be the most suitable method of approaching an archaeological material. Of course, this method will have
to be accompanied by a limited technological evaluation of the material that will indeed be subjective and that will stem from folk knowledge. In order to work, this evaluation will have to include a wellargued distinction between flakes removed in bifacial-reduction sequences
and flakes removed in other kinds of reduction sequences.
In the study, flakes with cortex-covered, dorsal surfaces were divided
into two sub-categories: primary decortication flakes (51 - 100% dorsal
cortex cover) and secondary decortication flakes (0 - 50% dorsal cortex
cover). The results of the initial statistical treatment of the cortical flakes
revealed that primary decortication flakes were significantly more
frequent in the first stage than expected and that they were significantly
less frequent in the last production stage. In the middle stages, no
significant differences were discovered. It was concluded that the cortex
appeared to have been removed rapidly and that the variable was
diagnostic only in the earliest stage of reduction (Amick et al. 1988:28).
Only a small number of flakes possessed more than two dorsal-flake
scars. The dorsal-directionality variable displayed a similar distribution.
This suggested that the importance of these variables in an analysis aiming
at recognising bifacial-production stages might be limited. It should be
noted that the experimental reduction sequence was being taken through
only to a point in the reduction that would be somewhat equal to
secondary thinning. Because these variables were closely correlated, it
was decided that dorsal directionality should be dropped at this point
of the investigation. In the following univariate analysis, it was concluded that flakes with higher than expected, dorsal-scar counts tended
to occur late in the reduction sequence, while flakes with lower than
expected values tended to occur early in the reduction. It was also
concluded that the intermediate- and late-stage debris exhibited similar
distributions, while only the first stage was distinct.
144

Previous Research on the Determination of Debitage


Variable

Description

Event

Ordered sequence of percussion blows

Sequence

Unique specimen number for all flakes

Hammer

Percussor type: 1=large hard, 2=small hard, 3=large soft, 4=small soft

Cortex

Cortex cover estimated in ordinal categories of cortex present*

Scars

The number of scars greater than 2 mm long on the dorsal surface*

Direction

The number of directions, recorded on an 8-coordinate system, which have scars present on the
dorsal surface*

Weight

To the nearest 0.1 g*

Length

To the nearest 0.1 cm*

Mid-width

Midpoint width to the nearest 0.1 cm*

Max-width

Maximum width to the nearest 0.1 cm*

Distance

Distance from the platform to the maximum width recorded for complete flakes only.
Measurement to the nearest 0.1 cm

P-width

Platform width. Recorded to the nearest 0.1 cm on all platform-remnant-bearing flakes (PRBs)

P-thick

Platform thickness. Recorded to the nearest 0.1 cm on all PRBs

* Recorded on all debitage.

Table 5:2. The variable coding format in Amick et al. 1988.

The authors stressed that the different hammer types had a significant effect on the morphology of the flakes. This is demonstrated by
the different morphological traits displayed by the flakes produced with
different hammers (Amick et al. 1988:29). Here, the authors face
another crucial problem. The only way to evaluate whether a certain
hammer type is responsible for certain morphological traits in the production debitage is by conducting experiments in which all possible
variables, except the type of tool used, are held reasonably constant.
This is not done in the article and it means that all the differences that
Amick et al. present as due to different hammers may be regarded as the
result of other phenomena. In fact, it is likely that these differences may
be the result of significant changes in the knappers manner of organising the flaking. Earlier studies have indeed demonstrated that the
significance of the hammer types related to traits in the production
debitage has been slightly exaggerated (Patterson & Sollberger 1978:107
f.). However, it is concluded that the large hammer flakes are the largest
in all dimensions, except distance and length, where soft hammer flakes
are largest. This demonstrates, according to the authors, that large
hammer flakes are thicker relative to their length than soft hammer
flakes (Amick et al. 1988:29). It is hereby suggested that these differences may also be the result of on-the-edge or behind-the-edge flaking
strategies.
145

Jan Apel
Even though weight seems to be correlated with almost all the other
variables, it is suggested that the variable has considerable value in
debitage studies (Amick et al. 1988:29). It is further concluded that
flake weight reflects flake thickness, that distance is a function of length,
and that maximum width and medium width are closely correlated
variables.
After conducting a multivariate analysis, based on the variables described above, it was concluded that the reduction stages could be reliably predicted after edge-trimming and platform-preparation flakes had
been identified and omitted from the investigation. These kinds of flakes
are manufactured throughout the whole reduction sequence and change
little in morphology and, because of this, they tend to confuse attempts
to define relative reduction sequences within the debitage (Amick et al.
1988:32). It was further concluded that the counting of dorsal scars
and cortex-cover percentages could be used to define the early stages of
the production. However, it proved difficult to define the intermediate
stages on the basis of the chosen variables. These stages are very important in the production of a biface, since a large proportion of the flakes
from a production will be debris produced during bifacial trimming
and thinning. The authors suggest that alternative variables may improve the chances of recognising the intermediate stages (Amick et al.
1988:34).

Odell 1989
In 1989, Odell published a paper in which he sets out to investigate
three questions regarding un-retouched debris:
1 Is it possible to accurately distinguish flake-core and bifacial technologies?
2 Is it possible to accurately distinguish the various stages of a bifacial
technology?
3 Is it possible to discern chronological differences in the debitage?
To answer these questions, Odell conducted a set of controlled experiments, together with a competent flint worker, Greg Thomas. We
had determined that the metric and discrete attributes of reduction flakes
appeared to contain sufficient internal variation and intrasite homogeneity
to warrant trying to discriminate among technologies and stages by
recording and analyzing attributes (Odell 1989:165). Odell stressed
146

Previous Research on the Determination of Debitage


Variable

Description

1. Weight

To the nearest 0.01 g, if possible.

2. Striking-platform width

Point-to-point measurement of maximum distance.

3. Striking-platform thickness

Point-to-point measurement taken perpendicular to the


S.P.W.

4. Maximum length

Measurement taken perpendicular to the axis of the


S.P.W.

5. Maximum Width

Point-to-point measurement taken at the locus of


maximum extension perpendicular to the M. L. line.
Record to the nearest 0.05 mm.

6. Quarter-widths (three measurements)

Measurements taken perpendicular to the M.L. line at


distances of , , and from the striking platform to
the distal end.

7. Quarter-thickness (three measurements)

Thickness measurements taken perpendicular taken at


the same locations as the Q-W measurements-

Table 5:3. The continuous variables used in Odell (1989).


Variable

Description/comment

1. Bulbar scar

To be recorded, at least one scar on the ventral surface


near the bulb of percussion must be present. Coding:
0 if absent, 1 if present.

2. Distal termination

This variable includes only the feather, hinge and


reverse hinge categories. Coding: 1 = feather, 2 =
hinge, 3 = reverse hinge.

3. Flake twist

Twist refers to the behaviour of the distal end of the


flake relative to its proximal end. Left twist is counterclockwise and right twist is clockwise rotation of the
distal end relative to the P.F.W. line.

4. Lipping

Measurement with reflective light microscope and


reticle. A lip is considered to be an extension from the
platform that is over 0.3 mm. Coding: 0 = extension
of platform less than 0.3 mm, 1 = extension equal to
or greater than 0.3 mm.

5. Modification of striking platform

Coding: 1 =cortex, patina and/or fracture plane only


(i.e. no traces of faceting, battering or grinding), 2 =
one facet, 3 = 2 or more facets, 4 = battered, 5 =
ground, 6 = faceted and battered, 7 = faceted and
ground, 8 = battered and ground, 9 = faceted,
battered and ground.

6. Dorsal perimeter scarring

Scarring damage on the dorsal surface below its


juncture with the striking platform. The variable does
not include traces of grinding and battering.

7. Dorsal surface removals

Coding: 1 = 1-4 removals, 2 = 5-9 removals, 3 =


10 or more removals.

Table 5:4. The nominal and ordinal variables used in Odell (1989).

147

Jan Apel
the importance of an objective way of recording the variables. This was
done to ensure that any variable used had to be replicable by another
person to 95% reliability. In the study each production stage was defined
as a sequence in which the knapper mentioned that he was progressing
to a different phase of operation (Odell 1989:164). Thus, a form of
natural staging was used in the experiments, but since an explicit,
technological definition of each stage is missing, this should be
considered to be a methodological weakness in the study.
The experiments included a flake core knapped in one stage, a large
biface knapped in one stage, the first two stages of both a Snyders and
a Hardin, bifacial-projectile point, stages 3 and 4 of the Snyders sequence, and stages 3 to 6 of the Hardin sequence (Odell 1989:170).
The variables that were recorded in the production debitage from these
stages are presented in Tables 5:3 and 5:4.
The data obtained from the variables were subjected to several series
of bi- and multivariate statistical tests. It was regarded as important to
analyse the nominal, ordinal and interval variables in separate groups, in
order to be able to use the methods most appropriate to each group
(Odell 1989:170). This procedure must be regarded as crucial, especially since multivariate statistics are often used uncritically in archaeology.
In order to answer the first question whether there was a possibility of distinguishing between cores of biface debitage the five nominal variables were evaluated in a series of chi-square tests, in which the
debitage from both core-reduction and different, biface-reduction stages
participated. The result of this test was that flake-core debitage was not
easily distinguished by with this data but that it displayed significantly
more bulbar scars than the other stages. Odell suggests that the relative
lack of platform preparation in the reduction of the flake core is responsible for this difference. It was further concluded that one of the
ordinal variables was better suited to discriminate between debitage from
flake-core and biface-reduction sequences than any nominal variable.
This was scarring on the dorsal perimeter of the striking platform
(Odell 1989:172). Odell was able to demonstrate that the flake-corereduction debitage had substantially less platform-preparation damage
than the flakes from the bifacial stages. As Odell himself points out,
this is not a surprising conclusion, but he concludes that ... it is
comforting that, at least in this data set, dorsal perimeter scarring can be
148

Previous Research on the Determination of Debitage


employed to discriminate core reduction flakes from all other production units (Odell 1989:174). And as a result of the study, it is concluded that core-reduction flakes can be distinguished from biface-reduction sequences. Since it is obvious that dorsal-perimeter scarring
reflects a decision to place the point of impact on the edge, one can
simplify Odells conclusion by stating that flakes with prepared striking platforms originate from core reductions.
The second question that the article set out to test if the various
stages of a bifacial technology can be accurately distinguished was
more difficult to answer. Striking-platform modification proved to be
the most consistent discriminator among the nominal variables (Odell
1989:176). It was concluded that the initial stage of the reduction sequences displays very little platform modification. During the second
stage, platform preparation increases and this manifests itself as more
facets and a high degree of battering (Odell 1989:177). Another nominal variable that proved to be disciminatory, was the type of distal
termination. The debitage from the final stages of both bifacial sequences
displayed almost exclusively feathered terminations, while the third stage
of the Snyders sequence contained the highest percentage of hinge
terminations.
The ordinal variables indicated that the average proportion of the
dorsal perimeter of the striking platform covered with removals increased per reduction stage. The opposite trend was true of the dorsal
cortex-cover variable. In this case, the variable decreased per reduction
stage. Odell points out that this variable turned out to be less discriminatory than expected, owing to the small sample size in some debitage
groups. Common sense will, however, inform anyone that it is very
likely that dorsal cover will occur on debitage from early stages, so it
seems that this fact should not have bothered Odell so much. It was
also possible to distinguish one of the stages well by the relative quantity of dorsal removal facets (flake scars). In the middle stage of the
Hardin sequence, two-thirds of the complete flakes had between five
and nine dorsal removals (Odell 1989:178).
According to Odell, the result of the investigation suggests that there
is a possibility that individual stages can be accurately distinguished from
each other, although some distinctions was achieved (Odell 1989:183).
Odell finally proposes that the following variables were the best overall
discriminators among the variables used in the study: striking-platform
149

Jan Apel

Fig. 5:3. A cumulative diagram showing a combination of attributes in relation


to flakes from production experiments with a flint dagger and two squareshaped axes (from Hgberg 1998: 9).

width, striking-platform thickness, striking-platform modification,


thickness, and weight. Odell also stresses that width at the mid-section
of the flake is an effective, internal, biface-reduction discriminator (Odell
1989:185).

H gberg 1998
In order to create a reference collection of flint flakes from the production of different kinds of Neolithic flint artefacts, a flint-knapping session in which several axes and one flint dagger were made was organised
by the Malm Museum in Malm, Sweden, in the summer of 1997.
An experienced flint-knapper, Thorbjrn Petersen, made the artefacts,
and the debitage from the experiments was analysed by Anders Hgberg,
of the Malm Museum (Hgberg 1998 and 1999). One of the objects
of this investigation was to see whether axe and dagger production could
be distinguished from each other in the production debitage (Hgberg
1998:1). A large number (26) of conceivable attributes were documented
on the flakes. According to Hgberg, none of the variables was discriminatory by itself. Although a certain attribute may be considerably
over-represented within the production of one artefact category, it will
150

Previous Research on the Determination of Debitage


always be present on, at least, some of the flakes from the other category
and vice versa (Hgberg 1998:8). It became clear, after subjecting the
recorded variables to statistical tests, that the following variables, in
combination with each other, can be used successfully to separate flakes
from square-shaped-axe production, on the one hand, and bifacial-dagger
production, on the other (Hgberg 1998:16 ff.):
1 Degree of angle between platform remnant and dorsal side of the
flake. It is concluded that square-axe-body flakes tend to have a relatively high angle (c. 90 degrees) while bifacial-thinning flakes have a
relatively low angle (c. 45 degrees).
2 Profile of bulb. Qualitative estimation of whether the flake has a
marked or a diffuse impact bulb on its ventral side. Square-axe-body
flakes tend to have pronounced-bulb profiles, while bifacial-thinning flakes tend to have diffuse-bulb profiles.
3 Flake curvature. Whether or not the profile of the flake is curved or
straight. Square-axe-body flakes generally have straight flake profiles, while dagger flakes have curved profiles.
4 Type of impact point. Whether the point of impact was near or on
the edge (marginal flakes) or behind the edge (non-marginal flakes)
of the platform remnant. Square-axe-body flakes tend to be detached
behind-the-edge while dagger flakes are detached on-the-edge.
5 Platform-remnant morphology. Whether the platform is faceted or
whether it is ground or prepared in some other way. Square-axebody flakes have faceted platform remnants while dagger flakes tend
to have small, preparedplatform remnants.
A cumulative diagram (Fig. 5:3) demonstrate how different combinations of these variables destinguish square-axe-body flakes from
bifacial-thinning flakes. In Fig. 5:4, typical flakes made by these two
reduction methods are presented.

Concluding discussion
In Chapters 6 and 7, flakes have been defined, besides my own experiments and observations, in accordance with the findings of this chapter.
As I see it, those who have worked with emic categories based on
generalisation from experiments (for instance, Flenniken and Hgberg)
have constructed flake categories that are easy to use but may be hard to
151

Jan Apel
distinguish for those who lack experience of materials like these.
However, in the cases in which processual archaeologists have been able
to define flake categories in an independent way, these definitions tend
to confirm the obvious. Thus, my own classification system is based
mainly on experience, although I have tried to formalise it whenever it
has been possible to do so. The following flake categories will be used
in order to classify the flint debitage on sites in southern Scandinavia
and eastern central Sweden in Chapters 6 and 7.
(Sub-category 1.) On-the-edge flakes removed by a soft technique. These
flakes are often thin and display a carefully performed, platform
preparation that is necessary in order to prevent the platform from
breaking during impact. Bifacial-thinning flakes, which, in this context, are the most common flake type within this category on Late
Neolithic sites in Scandinavia, often have a platform angle of c. 45
degrees.
(Sub-category 2.) Behind-the-edge flakes removed by a soft technique.
These flakes tend to be thicker than flakes of the former category.
The striking platforms are wide and deep have not been prepared.
This means, as has been suggested by several of the authors presented earlier in this chapter, that the relative thickness may be a
variable that can be used to separate on-the-edge flakes from behind-the-edge flakes. It has also been pointed out that the relative
thickness of different flake categories may reflect weight differences
as well. Consequently, in Chapter 6, the flakes from the Myrhj
site, which have been classified as Sub-category 1 and 2, have been
weighed in order to see whether weight is an independent variable
that can be used to distinguish these two flake types or not.
(Sub-category 3.) Bipolar flakes. This flake type has not been discussed
earlier in this chapter and it is very rare in the settlements that will be
discussed in the following chapters.
(Sub-category 4.) Behind-the-edge flakes removed by a hard technique.
(Sub-category 5.) Pressure flakes. I have used Flennikens definition, although these flakes are not divided further into primary and secondary pressure flakes.
(Sub-category 6.) Flakes that have been difficult to classify. In the Myrhj
investigation, presented in Chapter 6, this sub-category contains larger,
152

Previous Research on the Determination of Debitage

Fig. 5:4. Typical flake from squareshaped axe productio (upper flake) and
dagger production (lower flake) (from
Hgberg 1998:16 and 1999:85 f.).

153

Jan Apel
fragmented flakes without preserved striking platforms. In the investigations carried out in Chapter 7, the sub-category will contain
splinters and other fragments of flakes.
Our knowledge of the ability to infer production stages in the debitage
from bifacial-reduction sequences has increased dramatically during the
last 20 years. The summarised studies that have demonstrated that variables such as weight and length are useful analytical units. As we have
seen, the following variable has been successfully used in order to separate flakes from unifacial- or quadrifacial-reduction sequences from those
produced during bifacial-reduction sequences.
1. Flake curvature: bifacial flakes have curved profiles while core and
axe flakes tend to have straight profiles.
2. Bulb profile: Axe and core flakes tend to have more pronounced
bulbs than bifacial flakes.
3. Distal termination: bifacial flakes tend to have feathered distal terminations while the higher frequency of axe and core flakes have
abrupt distal terminations.
4. Flake thickness: Core and axe flakes tend to be thicker than bifacial
flakes.
5. Flake weight: The relative weight of bifacial flakes tend to be lower
than the weight of core and axe flakes. Flake thickness and flake
weight is most probably dependent variables.
6. Striking platforms width: Axe and core flakes tend to have wider
striking platforms than bifacial flakes.
7. Platform preparation (dorsal perimeter scarring): Bifacial flakes generally display signs of extensive platform preparation while core and
axe flakes generally are in devoid of any platform preparation.
8. Type of impact point: If the point of impact has been positioned
behind the edge or on the edge. Bifacial flakes are on-the-edge flakes
while core and axe flakes are behind-the edge-flakes.
9. Dorsal flaking angle: Bifacial flakes will have dorsal flaking angle of
approximately 45 degrees while axe flakes will tend to have a dorsal
flaking angle of 90 degrees. Most platform core flakes will have a
dorsal flaking angle that fall somewhere in between.

154

Previous Research on the Determination of Debitage


As regards variables that can be used to distinguish between flakes from
different production stages of bifacial artefacts, the following might be
considered:
1. Degree of dorsal cortex: Flakes from early stages will tend to have
more cortex on their dorsal surface than flakes from later stages.
2. Dorsal flake scar count: Flakes from early stages will tend to have
less flake scars on their dorsal surfaces than flakes from later stages.
3. Weight: Flakes from early stages will be relatively heavier than flakes
from later stages.

155

Chapter 6

Flint technology in Southern


Scandinavia

In this chapter, the archaeological contexts related to the production of


Late Neolithic flint daggers are presented. Most of these contexts were
recovered near the natural flint resources and, thus, this chapter is mainly
concerned with southern Scandinavian sites. In fact, even though there
certainly are dagger-production sites in Skne, the Danish contexts are
emphasised in this chapter. This is partly dependent on prehistoric
circumstances but is more likely to reflect the fact that more Danish,
Late Neolithic sites with dagger-production debitage have been recovered
and published than Swedish. Since it is not my main objective to present
all known contexts with evidence of flint-dagger production, this bias
is of less importance in this connection. The purpose of this chapter is
rather to see whether an archaeologically based understanding of the
organisation of the flint-dagger production can reveal any information
on the social system in which it functioned. Different kinds of sites,
such as flint mines, flint-extraction sites and settlements with evidence
of bifacial-flint working, will be analysed in detail, in order to try to
shed some light on the spatial organisation of the flint-dagger technology. The number of dagger-production sites dated to LN I outnumbers those dated to LN II and the Early Bronze Age. This is, in fact, an
interesting pattern in itself and will be discussed in social terms when
the chapter is concluded.

The Limfjord area


Most of the Danish settlements containing debitage from the production of flint daggers have been recovered in the Limfjord area in northern
Jutland (Fig. 6:1). These sites are usually dated to LN I and share a few
common features that reveal the important information about the
conventions that guided the architecture of the settlements. The houses
157

Jan Apel

3
2
4

8 7
11

9
10
13
1514
16

12

Fig. 6:1. The geographical distribution of the sites discussed in this chapter.

generally lie in the east-west direction and have two aisles divided by a
row of roof-bearing posts. The eastern part of the houses often has a
sunken floor that reveals itself as a darker area during the excavations. In
connection with these depressions, post-holes from roof- and wallbearing posts are regularly found. The depressions vary in depth between
0.25 and 0.6 m and in size between 9 and 14 m in length and between
5 and 7 m in width (Simonsen 1985:86).
This architectural tradition has its roots in the Single-grave Culture
in Jutland, where several houses of this type have been recovered and
dated to MN B (see, for instance, Hvass 1977, 1986). The excavation
of a house belonging to the Swedish-Norwegian Battle-Axe Culture in
eastern central Sweden indicated that this tradition was widely spread
even as early as the later part of the Middle Neolithic period (Hallgren
2000:19 ff.). The tradition was in use in virtually all Scandinavian areas
where agriculture was practised through the Late Neolithic and well
158

Flint technology in Southern Scandinavia


into the Early Bronze Age, when it was replaced by a three-aisle
construction, with two rows of roof-bearing posts without sunken floors
(Boas 1980).

Gug (1)
The first Late Neolithic house to be recovered in Jutland was the traces
of a building at Gug, south of the town of Aalborg, in the province of
Himmerland, near the Limfjord (Fig. 6:1). It was found by two amateur
archaeologists in the early 1950s and has never been properly published
(Simonsen 1982:86). A large number of flint implements were collected
during the investigation, including preforms, nearly finished and
complete flint daggers (Brndsted 1957:311). The house was found
when the remains of a sunken floor measuring 4 m x 2 m and filled
with flint waste, were discovered. The fact that large amounts of flint
were found led Brndsted to interpret the house as a flint-workers hut
(Brndsted 1957:312). Among the flint debitage covering the floor,
the following artefacts were found: 10 dagger and spearhead preforms,
25 sickle preforms, four arrowheads, one flake scraper, one drill, one
blade scraper and pottery sherds (Brndsted 1957:312).
In 1982, Simonsen pointed out that the flint debitage filling up the
depression probably represented a secondary filling of the pit (Simonsen
1982:88). From the ethnographical point of view this interpretation
makes sense, since it has been demonstrated that natural depressions,
abandoned structures and pits often became disposal locations in all
types of settlements (Schiffer 1987:61 f.). As regards contexts like these,
which are likely to have been used over a long period, it is almost
impossible to distinguish between primary and secondary depositions
(OConnell 1987). Different methods are needed to allow for alternative
interpretations, and it is likely that a commonsense inference based on
empirical observations alone will be wrong. It is important to use
formation theory based on first-hand observations formalised, in order
to allow for general application (as an example dealing with square-axe
production, see Hansen & Madsen 1983; as for daggers, see Callahan
& Apel, ms). Until this crucial step is taken, different views on how to
interpret materials like these can only be vague hypotheses. In the case
of the house at Gug, the less specific interpretation presented by
Simonsen seems more likely than Brndsteds more specific
interpretation.
159

Jan Apel

Fig. 6:2. Excavation plan of the Myrhj site (from Jensen 1972).
160

Flint technology in Southern Scandinavia


As regards chronological questions, it may be possible to draw
interesting conclusions from sites like Gug without having to conduct
extensive, middle-range analyses linking the static archaeological pattern
with a dynamic behaviour that explains it. If the flint waste is regarded
as a secondary deposition, a classification of daggers and the dagger
preforms in these can be used to date the structure. It may, at least, be
possible to make a morphological classification of the dagger material
to see if any chronologically dependent differences can be observed (i.e.
if the daggers are of types I-III or IV-VI; see Chapter 8). Since the site
has not been properly published, neither a more extensive analysis nor a
simple chronological one can be made without an examination of the
original material. Debbie Olausson has investigated the flint material
from Gug. According to her, the lithic debitage indicates that the
production was carried out by individuals of varying levels of skill
(Olausson 2000:129).

Myrh j (2)
The largest known settlement dating from the early Late Neolithic period
northern Jutland is Myrhj. The site is located only a couple of
kilometres from the shore of the Limfjord (Fig. 6:1) and was discovered by archaeologists from Aalborg Historical Museum during the
investigation of a row of stones dated to the Late Bronze Age. The site
was eminently well preserved, with sealed layers without archaeological
materials from earlier or later periods (Jensen 1972:74). Features
forming three houses (Fig. 6:2) were recovered, together with numerous
artefacts, many of which were typical of the early part of LN I. A
proportion of the pottery consisted of Bell-Beaker types. This is regarded
as a strong indication that western-European influences affected northern
Jutland during the late Single-grave Culture and the early Late Neolithic
period (Jensen 1973, Lomborg 1977, Vandkilde 1996). The site was
C14-dated to 3860 BP, which, according to the calibration curve
presented by Pearson and Stuiver (1993), gives an interval of 24642142 cal BP (1 sigma).
The Myrhj settlement and the LN-I period in Scandinavia are
contemporary with the Dutch Veluwe-Beaker phase. Several sites dated
to this period have produced Beaker pottery, especially in the Limfjord
area (Vandkilde 1996:281, Fig. 289). It deserves to be mentioned that
only a small percentage of the pottery on these sites is of typical Bell161

Jan Apel
Beaker forms. Thus, ordinary types of pottery, often plain and coarse,
predominate. Accordingly the occurrence of fine Beakers in settlements
as well as burials suggests that Beaker pottery served as a kind of table
ware, which was used for feasting and drinking rather than ordinary
household work (Vandkilde 1998). On the continent as well as in the
British Isles, the Beakers of this phase are regularly associated with
triangular, bifacially-shaped arrowheads in flint and stone wristguards
with two or four holes (Harrison 1982:26). A fragment of a stone
wristguard was indeed found at the Myrhj site and all three houses
contained triangular, bifacial arrowheads in flint. In addition, ten
arrowhead preforms were also recovered, indicating that arrowheads
were actually produced at the site (Jensen 1973:74 ff.). Type-I daggers
are also frequently associated with Beaker pottery (Vandkilde 1998).
Thus, in the early phase of LN I, there is a close connection between
northern Jutland and the late, western, Bell-Beaker horizons, as they are
manifested in the Rhine delta and in Wessex. As we shall see in Chapter
9, a considerable number of Scandinavian flint daggers have been found
in the Netherlands (Bloemers 1969), but only a few examples have
been recovered in association with bell Beakers. Harrison (1982:26),
for instance, mentions one context only in which a Scandinavian flint
dagger was recovered, together with a bell Beaker: a barbed-wire Beaker
from the Dutch Early Bronze Age, which is contemporary with LN II
in Scandinavia. Unfortunately, Harrison does not present any
information on what dagger type it was. The Veluwe phase represents
the first archaeological indications of metalworking in the Rhine delta
(Harrison 1982:27). The emergence of local coppersmiths may explain
why imported Scandinavian daggers were not considered to be grave
goods and were thus not found in this period. In earlier Beaker phases
in western Europe, for example, in the AOO (All-Over-Ornamented)
Beaker phase, flint daggers of imported, honey-coloured GrandPressigny flint were frequently used as grave goods (Harrison 1982:21).
Seed impressions in the pottery and the presence of harvesting knives
made of blades struck from conically-shaped cores suggests that crop
agriculture was practised by the inhabitants of Myrhj. Animal teeth
from domestic cattle were also recovered. Concentrations of mussel
shells were present in all three houses, indicating that mussels were
collected on the shores of the Limfjord 15 km to the west. It is likely
that this work was done at locations such as Rnbjerg Strandvolde, a
162

Flint technology in Southern Scandinavia


site that will be described below, where Late Neolithic artefacts have
been recovered. Jensen suggests that the shells might have been collected chiefly as raw material for making the chalk powder to temper
the pottery (Jensen 1973:78).
During the excavation of Myrhj, a considerable amount of flint
artefacts, including 14 187 flakes, was recovered. This number does
not include splinters and greatly fragmented flakes. A large part of the
un-retouched material consisted of wide flakes, sometimes wider than
long, many of them with large bulbs and narrow angles between the
platforms and the ventral surfaces. It is suggested that these flakes were
waste products from the making of axes, chisels, daggers, etc. (Jensen
1972:79 f.). Unfortunately, no further technological information is
presented about this production debitage. As we have seen in the previous
chapter, more detailed information on certain flake attributes, regarding
such things as platform morphology, measurements, weight, etc., would
increase the possibility of presenting a substantial interpretation of the
flint technology present at the site. Evidence of behind the edge flaking, in the form of flakes with thick and un-retouched, platform
remnants, well into a reduction sequence, would suggest that the finished
product was some kind of square axe. On the other hand, evidence of
edge-flaking, in the form of small and carefully prepared, platform
remnants, points to bifacial-reduction sequences. In order to investigate
this matter further, I subjected a proportion of the flint flakes from the
Myrhj site to a technologically orientated classification during a fiveweek visit to the Historical Museum in Moesgrd in the summer of
1999. Some preliminary results of this investigation will be presented
in the following pages.
According to Jensen, two qualities of flint are represented at Myrhj.
One is grey and calceous and generally gives a coarse impression; the
other is black and glossy and appears to be of a very good quality. These
two qualities correspond to the terms danian and senonian flint, as
defined by Becker (1988:46 f.). Jensen notes that the black flint was
predominantly used to make cutting and scraping tools, such as scrapers,
daggers and arrowheads, while the grey flint was used for chopping
tools, i.e. chisels and axes (Jensen 1973:80). Thus, it seems that it is in
the black-flint waste at the Myrhj site that we should expect to find
dagger-production debitage. In this context, Beckers emic categorisation
of flint types are used (Becker 1988). Therefore, the terms senonian
163

Jan Apel

10 cm
Fig. 6:3. Fragments with negative-flake scars displaying a parallel-flaking
strategy from Myrhj (drawings by Lotta Sundberg).

and danian flint might not be correct from the geological point of view
(see Thomsen 2000). However, from the flint knappers point of view,
these terms are relavant.
Most of the flint nodules recovered at Myrhj were obviously
collected as beach pebbles and not mined. This conclusion is further
substantiated by the fact that more than one-third of the flakes display
cortex on their dorsal sides (Jensen 1973:80). In this context, it is
suggested that these pebbles and flakes represent a local and domestic
production of everyday items that is probably present in all the Late
Neolithic hamlets in the area. It is likely that this domestic production
was aimed at producing scrapers, since a large quantity of scrapers was
recovered in each of the houses. Of these, 80% display cortex,
something that indicates that they were made from beach pebbles and
that a large part of the flakes from the site should be related to the
production of scrapers. The debitage stemming from the production
164

Flint technology in Southern Scandinavia


of everyday items is of a very generalised character and these materials
will thereby be present on sites from different Stone Age periods. It is
hereby suggested that a certain proportion of the flakes from the Myrhj
site can be connected with this type of debitage. However, the following
investigation will mainly concern flakes that can be tied to the production
of core tools, such as axes and daggers.
Fourteen fragments of flint daggers were found at Myrhj, and they
formed the basis on which the site was placed in its chronological and
cultural framework. All these fragments are of lancet-shaped daggers
and date the site to LN I. Some fragments display parallel flake scars
(Fig. 6:3). This, together with the observation that there are large pressure
flakes (Fig. 6:6) among the production debitage from the site, indicates
that daggers of type I C were made here. However, there is no indication
of the type-I preforms being ground before parallel flaked. Generally
most parallel-flaked dagger blades were grounded before parallel-flaked.
Another dagger fragment had been extensively retouched (Jensen
1973:88). In addition to the dagger fragments found at the sites, several
finds of other artefact categories in flint were also recovered during the
excavation: seven borers, five edge stickles, a strike-a-light made of a
long flake fragment, two harvesting knives with gloss on their edges
and made of blades, several, complete, thick-butted axes and numerous
preforms for thick-butted axes and adzes, one chisel and a preform for
a chisel, 15 bifacial arrowheads and 10 preforms for bifacial arrowheads
and five transverse arrowheads (Jensen 1973:82 ff.).
Jensen concludes that the daggers and other products, such as bifacial
arrowheads, were produced on the site. Thus, the Myrhj settlement
may be the earliest flint-dagger production site in Scandinavia (Nielsen
1981:138). However, at the time when Jensen wrote his report, there
was no way of investigating this idea in a clearly argumentative way.
The large amounts of flint debitage indicate that flint work was carried
on at the site; whether daggers actually were produced remains to be
seen. If Myrhj was a workshop, the retouched fragment indicates that
the settlement not only represents a link in the flint-dagger life chain,
but that examples were utilised on the site or that utilised examples
were brought there to be re-sharpened by the producer. Thus, a closer
analysis of the production debitage from Myrhj may shed some light
on the social context of the fabrication process involved in the production
of the lancet-shaped, flint daggers of LN I. In accordance with the main
165

Jan Apel
aim of this thesis, it is suitable to start this investigation here.
The flakes recovered at Myrhj belong to four main discrete contexts; flakes were recovered in the depressions of all three houses found
on the site (designated D, EAB and GAB), as well as in an area without
any structures located immediately to the east of house D, designated
VfD. The basis of the following analysis is a database consisting of
1 896 documented flakes. These flakes make up 13% of the total
number of flakes from the site. However, flakes were sampled from
only three of the four discrete contexts. Since house D was the first Late
Neolithic structure to be found during the excavation, it was not as
extensively documented as the other houses. It was also stated in the
report that only 26.8% of the flakes from house D were so-called wide
flakes, presumably stemming from the production of core artefacts,
such as axes and daggers, while a higher proportion of the flakes from
the other houses were flakes of this kind (Jensen 1973:80). For these
two reasons, in combination with the fact that I had a limited amount
of time at my disposal, the investigation was restricted to flakes from
EAB, GAB and VfD. In Table 6:1, the actual relation between the total
number of flakes from the contexts and the sampled numbers is
presented.
The ratio between the black- and grey-flint variants in the total flake
material is about 2:1. In each of the three contexts that concern us here,
the relationship between these two flint types is expressed in Table 6:2.
It is apparent that, at least as regards this aspect, the variability of the
original contexts has been covered in EAB and GAB where the
relationship between the sample and the total amount coincide perfectly.
In VfD, however, there is not such a good correlation. In order to
understand this discrepancy, a discussion of the sampling strategy, in
relation to the anticipated characters of the different contexts, is needed.
In the museum collection, the flakes from each square metre are contained
in brown-paper bags. During the classification of the flakes, I simply
worked my way through the bags, from each investigated context, in a
random fashion. This means that a formalised sampling strategy, aimed
at covering a certain area, was not used in this case. For practical purposes,
this way of working was convenient, since bags containing flakes from
certain square metres did not have to be located. This somewhat
primitive, sampling strategy was used because it was anticipated that

166

Flint technology in Southern Scandinavia


the flakes in the house depressions were secondary refuse deposits and
not primary knapping floors, i.e. that they had been dumped in the
depressions after the houses were abandoned (for ethno-archaeological
evidence of such behaviour, see Schiffer (1987:61 f.). Concerning the
flint in the depressions in the two houses, this view is strengthened by
the fact that the samples are representative of the total amounts.
However, in area VfD, the agreement between the sample and the total
amount of flakes does not coincide quite as neatly. It may be that this
area, to a greater extent than the houses, contains material that has been
left in primary positions. This actually makes sense, since this is an area
between the houses where several different tasks may have been
performed. Consequently, flakes of different raw materials were deposited
at different locations within the area. In order to cover the variability
within this context, a sampling strategy based on spatial considerations
would have to be used. For the purposes of this chapter, however, the
chosen strategy is considered to be sufficient.
T otal no. of flakes
Sam ple size
Sam ples size in %

D
5856
0
0

fD
1203
414
34%

EAB
5290
1109
21%

GAB
1838
507
28%

Table 6:1. Total number of flakes in each of the four main contexts at Myrhj
in relation to the sample size.
F lin t ty pe
S en o n ian
D an ian

V fD
77
23

Sample
69
31

EAB
70
30

Sample
68
32

GAB
55
45

Table 6:2. Percentages of the two flint types in the relevant contexts in
relation to the samples.

The sampled flakes were classified in the following sub-categories,


based on the concluding discussion in Chapter 5:
1. On-the-edge flakes removed by a soft technique. These flakes are
often thin and display the careful platform preparation that is necessary
in order to prevent the platform from breaking during impact.
2 Behind-the-edge flakes removed by a soft technique.
3. Bipolar flakes.

167

Jan Apel
4. Behind-the-edge flakes removed by a hard technique.
5. Pressure flakes.
6. Larger, fragmented flakes without preserved platform remnants.
In this context, small-sized, production debitage, such as small flake
fragments and splinters, will not be discussed. However, samples
including such flake fragments were also documented and I hope to be
able to present this data in the near future. In order to give a more
detailed, technological meaning to the flake categories presented above,
some generalising arguments about the categories that will be used to
separate axe-production debitage from dagger-production debitage will
be presented. Sub-category 1 flakes are consistently produced during
the thinning stages of bifacial-reduction strategies. These flakes have a
combination of attributes that are consistent with Flennikens or
Hgbergs definitions of bifacial-thinning flakes and dagger production
debitage that were presented in Chapter 5 (compare the flakes in Figs.
5:3 and 5:5 with the flakes of sub-category 1 in Fig. 6:4). A large amount
of the middle-stage, production debitage from dagger-making will be
of this type. However, these flakes will also be produced in some stages
of square-axe production, mainly during the construction of the cutting
edge, when on-the-edge flakes will be removed from the edge and
into the interior parts of the axe. For instance, it has been estimated that
these flakes make up only c. 1 % of the total amount of flakes from the
production of squared-shaped axes in stone (Sundstrm & Apel
1998:160 ff.). The investigation of an axe-production site in southwestern Skne reveals that 6 % (nine of 150) of the flakes from the
production of a Neolithic, square-sectioned axe in flint were of a typical
bifacial character (Hgberg 1999:86, Fig. 86). Generally speaking, the
flakes produced during the shaping and preparation of the cutting edge
on an axe will tend to be smaller than flakes from the thinning stages in
the production of bifacial artefacts, if the size is the same. This is due to
the fact that the former flakes are struck in order to form the edge and
do not have to reach into the central area of the axe side. Bifacial-thinning
flakes, on the other hand, must reach into the central area of the artefact
in order to be overlapped by flakes struck from the opposite margin,
thereby making the blade thinner. It must also be mentioned that onthe-edge flakes, with prepared platforms, will be produced during the

168

Flint technology in Southern Scandinavia

10 cm

Fig. 6:4. Examples of sub-category-1 flakes from the Myrhj site. These flakes
are the result of typical bifacial-reduction strategies. In this context, this flake
category is interpreted as production debitage from the later stages in the
production of flint daggers. Drawings by Lotta Sundberg.

169

Jan Apel

10 cm

Fig. 6:5. Examples of sub-category-2 flakes from Myrhj. In this context, this
flake category is interpreted as being related to the production of squareshaped, flint axes. Drawings by Lotta Sundberg.
170

Flint technology in Southern Scandinavia

10 cm
Fig. 6:6. Examples of sub-category-5 flakes from the Myrhj site. In this
context, this flake category is interpreted as being related to the parallelflaking of flint daggers of type I C. Drawings by Lotta Sundberg.

production of certain blade types as well. These can be distinguished


from bifacial-thinning flakes by their having a relatively higher platform
angle (c. 90 degrees) than the bifacial flakes (c. 45 degrees) (Hgberg
1998). At Myrhj, where there is scarce evidence of blade production,
most of the flakes of this type are likely to have originated from the
thinning stages in the production of bifacial artefacts. Sub-category-2
flakes are commonly produced during the middle and later stages in the
production of square-shaped axes (Hansen & Madsen 1983).
The three flake types, 1, 2 and 5, presented in Figs. 6:4-6:6 have
been classified according to certain criteria presented in Chapter 5. Thus,
qualities such as the presence of the point of impact on or behind the
platform edge, the platform angle and the curvature of the flake profile
have been used to distinguish between the flakes of these different types.
But there is also a more formalised way of distinguishing these flake
171

Jan Apel
70
60
50
40
30
20
10
0

sc1 a (100 flakes)


70
60
50
40
30
20
10
0

00, 0,0
00
0
1- 09
0, 0,0
00
0
2- 19
0, 0,0
00
02
39
0, 0,0
00
0
4- 39
0, 0,0
00
0
5- 49
0, 0,0
00
0
6- 59
0, 0,0
00
06
79
0, 0,0
00
0
8- 79
0, 0,0
00
0
9- 89
0,
00
99
0,
01
-

0,
01

0,

00

0-

0,
00
1- 09
0,
0,
00
00
19
20,
0,
00
00
29
30,
0,
00
00
39
40,
0,
00
00
49
50,
0,
00
00
59
60,
0,
00
00
69
70,
0,
00
00
79
80,
0,
00
00
89
90,
00
99

70
60
50
40
30
20
10
0

sc2 a (271 flakes)

00, 0,0
00
0
1- 09
0, 0,0
01
00
9
20, 0,0
0
00
3- 29
0, 0,0
0
00
4- 39
0, 0,0
0
00
5- 49
0, 0,0
05
00
9
60, 0,0
06
00
9
70, 0,0
07
00
9
80, 0,0
08
00
9
90,
00
99
0,
01
-

00, 0,0
00
0
1- 09
0, 0,0
00
0
2- 19
0, 0,0
00
02
39
0, 0,0
00
0
4- 39
0, 0,0
00
0
5- 49
0, 0,0
00
0
6- 59
0, 0,0
00
06
79
0, 0,0
00
0
8- 79
0, 0,0
00
0
9- 89
0,
00
99

0,
01
-

70
60
50
40
30
20
10
0

sc4 a (388 flakes)

sc5 a (93 flakes)

Fig. 6:7. The distribution of the density (measured as the average weight per mm2) of flakes
from sub-categories 1, 2, 4 and 5.

categores. During behind-the-edge flaking, a lot of flake mass can be


found in the platform area (Fig. 6:5). During on-the-edge flaking, on
the other hand, there is less mass to be found near the platform (Fig.
6:4). This means that sub-category 1 flakes tend to be thinner than subcategory 2 flakes. It is also a fact that pressure-flaking tends to make less
of a bulb on the ventral face of the flake than percussion does. Thus, it
should be possible to distinguish between on-the-edge flakes released
by percussion from on-the-edge flakes released by pressure. This means
that on-the-edge percussion flakes (1) should generally be thicker than
on-the-edge pressure flakes (5, see Fig. 6:6). Since, as we saw in Chapter
5, differences in the relative thickness of flakes equal differences in weight,
it should be possible to distinguish the flakes by their relative density.
Thus, the following simple formula was used to express this idea:
Flake weight per mm2 (density) = Weight (g) / (length (mm) x width (mm)).

In Fig. 6:7, diagrams showing the distribution of the density of the


flakes of the four flake sub-categories 1, 2, 4 and 5 are presented. As we
172

Flint technology in Southern Scandinavia


can see, the 1 flakes are generally less dense than 2 flakes. This means
that, in circumstances such as those on many of the sites presented in
this chapter, data on the length, width and weight of all or a sample of
the flakes from a site may be a substantial help in making technological
interpretations. As yet this is only a suggestion but, as soon as I have the
opportunity, I shall try this method of making flakes in modern
experiments with full control of the tool that is produced. The pressure
flakes (5) generally have the lowest density and the sub-category-4 flakes,
which can be connected with a hard, direct, behind-the-edge, knapping
strategy, aimed at turning small beach nodules into everyday items, such
as scrapers, have the highest density of the four categories of flakes.
Thus, the density of complete flakes may be a useful tool in
distinguishing flakes produced by different knapping strategies. The
frequencies of axe and dagger flakes (between sub-categories2 and 1 is)
from the different contexts are presented in Fig. 6:8.
It is apparent in Table 6:3 that flakes of sub-category 2 show more
signs of cortex than those of sub-category 1. While only 15 % of the
sub-category-1 flakes in senonian flint display corteon their dorsal sides,
33 % of the flakes of sub-category sub-category 2 do so. Of the subcategory-4 flakes in senonian flint, 42 % show signs of cortex. The fact
that 80 % of the scrapers found at Myrhj had cortex may indicate that
these scrapers were produced by a hard-hammer, platform strategy and
that a large proportion of the sc-4 flakes were waste products of this
strategy. It is interesting to note that, while 15 % of the sub-category-1
flakes in senonian flint display cortex on their dorsal faces, only 5 % of
the sub-category-1 flakes does so. If the majority of the bifacial flakes
in danian flint originate from the forming of the cutting edges of squareshaped axes, this pattern actually makes sense. The edge on such axes is
Flake and flint type
sc-1A
sc-1B
sc-2A
sc-2B
sc-4A
sc-4B
sc-5A
sc-5B

No cortex
85%
95%
68%
84%
58%
73%
96%
95%

< 50 %
14%
5%
25%
11%
27%
20%
3%
5%

> 50 %
1%
0%
8%
5%
15%
7%
1%
0%

Table 6:3. The degree of cortex coverage on the dorsal faces of the flake subcategories presented in the text (A = senonian, B = danian flint).

173

Jan Apel
19%

49%

51%

EAB flakes of flint quality a

81%
EAB flakes of flint quality b
6%

27%

73%
GAB flakes of flint quality a

94%
GAB flakes of flint quality b
13%

38%

62%

WfD flakes of flint quality a

87%
WfD flakes of flint quality a

Fig. 6:8. The relationship between sub-category-1 flakes (lighter area) and subcategoryc-2 flakes (darker area). The first flakes are produced mainly during
dagger making while the latter flakes are produced during square-axe production.

generally shaped when the body has reached its squared form. This means
that there is less of a chance that any cortex will still be present on the
axe body when the flakes are detached. During bifacial-reduction
strategies, however, the knapper will try to avoid detaching flakes from
areas that are flat or concave, since such flakes will tend to be abruptly
terminated in hinge fractures. Consequently, in areas such as these, cortex
is liable to be untouched well into the middle stages of the production.
If the bifacial, subcategory-1 flakes in danian flint actually are production
174

Flint technology in Southern Scandinavia

Fig. 6:9. The width (horisontal axis)-length (vertical axis) relationship of the sc1 flakes from Myrhj. Danian flint (circles), senonian flint (dots).

debitage from the making of square axes, they will tend to be shorter
than the bifacial flakes created during the production of, for instance,
daggers. In the diagram presented in Fig. 6:9, we can see that this was
actually the case on Myrhj.
The investigation of the flakes from the Myrhj site suggests that a
large part of the material was the result of the production of everyday
items, especially scrapers. However, traces of more sophisticated
knapping strategies are also present in the flakes. The presence of
subcategory-2 flakes, together with a large number of preforms, indicate that thick-butted axes were made on the site. These axes were preferably made of grey, calceous, danian flint, even though senonian flint,
to a lesser extent, was also used for this purpose. Bifacial-thinning flakes
and large pressure flakes in senonian flint reveal that the later stages in
the making of type-I flint daggers were also performed on the site.
Since large, early-stage flakes, from axe as well as dagger production, are
lacking on Myrhj, it may very well be that these early stages were
performed in more secluded places near the raw-material sources.
As we have seen in Chapters 2 and 4, production stages that require
a large degree of know-how are sometimes performed on full display in
tribal societies, in order to bring a message of authority to the members
of the community. In Fig. 6:10, a flow chart of different aspects of the
flint industry at Myrhj in relation to their inherent degrees of the
175

Jan Apel

Fig. 6:10. Flow chart of the flint industry at Myrhj related to the theoretical concepts of
the present and the past in the present presented in Chapter 4.

theoretical terms the present andthe past in the present (see Chapter 3) is
presented.

R nbjerg Strandvolde (3)


Not far away from Myrhj, in the province of Himmerland on the
shore of the Limfjord, c. 8 km north of the famous, Late Mesolithic,
Erteblle site, worked flint appeared when an area near the shoreline
was ploughed (Skousen 1998). This site was, on typological grounds,
dated mainly to the Erteblle phase of the Late Mesolithic. However,
there was also evidence of Neolithic activities, in the form of decorated
pottery and different kinds of flint artefacts. The Neolithic flint artefacts
apparently lacked the kind of patina present on the larger part of the
Mesolithic flints (Skousen 1998:41). In this context I shall concern
myself only with the Late Neolithic artefacts recovered at the site. The
decorated pottery was of the same type as the Bell-Beaker-influenced
pottery found at Myrhj and can therefore be dated to an early part of
the Late Neolithic (Skousen 1998:47). Skousen (1998:49) also suggests
that fragments of thick-butted flint axes and a preform for a bifacial
arrowhead also belong to the early phase of LN I or possibly to the
latter part of MN B. A couple of bifacial sickles and a strike-a-light in
flint were also recovered. It is pointed out that these may have originated
from the Late Neolithic but that artefact categories such as these were
176

Flint technology in Southern Scandinavia


also used during the Bronze Age (Skousen 1998:49). It is concluded
that the Late Neolithic activities at Rnbjerg Strandvolde, in
combination with the fact that shells collected in the Limfjord were
recovered on the inland Myrhj site, demonstrate that it is meaningful
to study inter-site differentiation during the Late Neolithic period
(Skousen 1998:65).

Stendis (4)
The settlement at Stendis in north-western Jutland resembles Myrhj
in many ways. The house construction is similar to those of the three
houses at Myrhj. Bell-Beaker pottery and large amounts of flint were
recovered at both sites. At Stendis, 1 445 pieces of flint were found. Of
these, 24 were secondarily worked (Skov 1982:40). There was a general
tendency for the un-retouched flakes to be short and wide, just like the
Myrhj flakes. In his report, Skov does not mention anything about
the platforms on the flakes, and, as we have seen above, this makes it
difficult to give a technological interpretation of the material on the
basis of the written report. However, as at Myrhj, some of these flakes
may very well have been debitage from bifacial-production sequences.
No lancet-shaped daggers were found at the site, but three bifacial
arrowheads and one preform for an arrowhead were recovered (Skov
1982:40).
Skov mentions three hammerstones in quartzite with battered ends,
which, he suggests, were used for finer flint work, such as the making
of bifacial arrowheads and daggers (Skov 1982:40). However, it seems
unlikely that the stones were used for this purpose. Quartzite would
make a hard hammer and this in turn would indicate behind-the-edge
flaking, which requires a thick platform (Callahan 1979:67). This
technique, if used at all in biface work, would be preferred only in the
early stages of production. On the other hand, if the flakes at the site
stem from square-axe production, the hammerstones may be related to
this activity.

Tastum (5)
At Tastum, another typical, Late Neolithic, house construction with a
sunken floor was recovered. Bell-Beaker pottery was also found here. A
layer in the house contained a fragment of a dagger handle and a bifacialarrowhead preform (Simonsen 1985:85). This layer is interpreted as a
secondary filling of the floor pit that was filled in after the house had
177

Jan Apel
been abandoned (Simonsen 1985:82). Simonsen writes that the material
confirms the exclusion of a dating of the house to the period previous
to LN (Simonsen 1985:87). This is a curious statement, since the
artefacts that were actually found in the depression were interpreted as
secondary filling. However, by analogy with the sites mentioned above,
it seems likely that the site should be dated to an early part of the Late
Neolithic. Apparently, it was not possible to classify the dagger fragments
according to any of Lomborgs types (Simonsen 1985:86).

Late Neolithic sites in the province of Thy


The Thy project, which began in 1990, was aimed at conducting a
long-term investigation of a remote region facing the stormy North
Sea. Within a diachronic perspective stretching from the early Neolithic
period to the end of the Bronze Age, economic variation, local settlement patterns and their translocal integration into a political economy
were to be related to models originating from earlier research by the
participants (Earle et al. 1998:2). As a part of the project, fieldwork
was conducted in two of Thys over 60 parishes: Snderh and Heltborg.
Archaeological materials, interpreted as reflecting prehistoric settlement
sites, were collected during field-walks. A combination of block-walking,
shovel testing, phosphate testing and plough-zone excavations was also
used. Finally, a sample of Late Neolithic and Early Bronze Age houses
was excavated. According to the project members, this micro-regional,
household archaeology allowed us to document an emerging economic
specialization as part of the intensification that supported the articulation of Thy with a translocal political economy (Earle et al.1998:2).
In the following pages, the Late Neolithic sites that were investigated in
this project will be described and discussed. As we shall see, these particular sites do not seem to have been directly involved in the production of flint daggers, as were the majority of the other sites discussed in
this chapter. However, the conclusions that the participants in the project
draw from these investigations need to be discussed in relation to my
own findings.
Three sites dated to the Late Neolithic and four sites dated to the
Late Neolithic or Early Bronze Age periods were investigated and will
be briefly described in the following pages. All these sites are located in
Snderh parish, which covers an area of c. 12 km2 and is situated in the
central parts of Thy. According to the excavators, these sites can be
178

Flint technology in Southern Scandinavia


divided into two distinct types: (1) small clusters of sunken-floor houses
or pit-houses and (2) lithic scatters with some post-holes but with no
defined house structures (Earle et al. 1998:8). The presence of ard marks
and cereal grains, mostly of barley but to a lesser extent also of emmer
and spelt wheat, indicates the presence of crop agriculture. A high
frequency of scrapers (compare with Myrhj), probably for processing
hides, and evidence of ceramic sieves for the preparation of cheese
indicate that animal husbandry was practised as well (Earle et al. 1998:8).

Thy 2758 (6)


This site is located in the northern part of Snderh. A Bronze Age
barrow was situated on the site and probably helped to protect the site
from being destroyed by the modern, deep ploughing. Features forming
three distinct houses were recovered during excavation (Steinberg
1997:139). Extensive plough marks, found beneath the barrow, suggest
that the area around the site was farmed in the Late Neolithic. Since
refuse had been secondarily deposited in one of the house pits, it was
assumed that the settlement had been inhabited for some time (Steinberg
1997:139). House 1, the best-preserved structure at Thy 2758, was
13.7 m long and 3.7 m wide, in a ship-shaped depression. Artefacts
were recovered mainly along the deepest, central axis of the house and
the excavators seem to suggest that they were deposited in primary
positions, where movement would have been easiest (Earle et al.
1998:9). The formal artefact categories recovered during the excavation
of the site included several daggers, blades and scrapers.
In addition to the excavation, a plough-zone test was conducted on
the site, in order to form the basis for a lithic site signature of the site
that could be compared with similar signatures from the other sites
(Steinberg 1998). Thirty-four plough-zone samples, distributed over
the site, were sieved and the lithic material was classified (Table 6:4).
An estimate of the real amount of lithic artefacts deposited on the site
during prehistory was made by a statistical calculation that will not be
commented in this context (for a description, see Steinberg 1998).

Thy 2756 (6)


This site was located 500 m from Thy 2758. No coherent structures
were recognised below the plough layer. However, a few features, in the
form of a thin cultural layer and a few post holes or pits, were
documented (Steinberg 1997:146). The site covered an area of 2 519
179

Jan Apel
Category
Flakes
Cores
Preforms
Retouched frost flakes
Hammerstones
Burnt flint
Scrapers
Blades
Burins
Borers
Knives
Strike-a-lights
Arrowheads
Axes
Daggers
Sickles

Thy 2758
1211
8
5
16
0
242
5
0
2
0
2
0
3
1
1
1

Thy 2756
91
0
0
0
0
30
0
0
0
0
0
0
0
0
0
0

Thy 2757 Thy 2920


669
150
8
0
5
0
9
0
0
0
133
0
2
1
0
2
2
0
0
0
0
0
0
0
1
0
3
0
2
0
0
0

Thy 2922
470
1
1
0
0
0
4
10
0
0
0
0
0
0
2
0

Thy 2923 Thy 2924


129
129
1
0
0
1
1
3
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0

Table 6:4. Flint artefacts recovered in plough-zone samples taken at Late Neolithic sites
investigated by the Thy project (from Steinberg 1997:144 ff.).

m2 and seven plough-zone samples distributed over the site area were
sieved. Only 91 flakes and 30 pieces of burnt flint were recovered (Table
6:4). Fortunately, the site could be dated to the Late Neolithic period
on account of a dagger found during field-walking, an axe of a Late
Neolithic type found in the excavation backfill and a bifacial arrowhead
found during the excavation of the cultural layer (Steinberg 1997:150).

Thy 2757 (6)


This site was located in an area where a Bronze Age, long barrow was
found in 1877. Since then, diagnostic artefacts from all periods,
including fragments of polished axes, 11 flint daggers and 5 sickles from
the Late Neolithic period, have been recovered in this area (Steinberg
1997:153). Thirteen plough-zone samples were taken from the 12 829m2 area making up the site. The result of this investigation is presented
in Table 6:4.
In addition to these three Late Neolithic sites, four sites were
investigated that are dated either to the Late Neolithic or to the Early
Bronze Age.

Thy 2920 (6)


This site was recognised during field-walks in which 2 fragments of
polished axes, 2 sickles, 2 pressure-flaked tools and 13 scrapers were
found. The site was never excavated, but 14 plough-zone samples were
taken in an area covering 20 103 m2. These samples indicate that the
180

Flint technology in Southern Scandinavia


frequency of flakes on the site was comparatively low, since only 7
flakes per square metre were recovered. This number should be compared
with, for instance, the 22 flakes per square metre recovered at Thy 2758.
Apart from flakes, only three formal artefacts were recovered in the
samples: a scraper and two blades (Table 6:4).

Thy 2922 (6)


This site is the largest of the four Late Neolithic/Bronze Age sites
investigated in the Thy project. It was investigated by surface collecting
and by taking 23 plough-zone samples and covered an area of 20 428
m2. Consequently, no features were recorded. According to Steinberg,
the presence of two dagger fragments and an abundance of pressureflaked tools may suggest that this site should actually be regarded as a
Late Neolithic rather than as a Late Neolithic/Early Bronze Age site
(Steinberg 1997:166). The artefacts recovered in the plough-zone
samples from the site are presented in Table 6:4.

Thy 2923 (6)


Very few artefacts, apart from pottery sherds, were recovered during
field-walks at Thy 2923, thus making it one of the smaller Late Neolithic
sites to be investigated in the project, covering an area of 9 926 m2
(Steinberg 1997:170 ff.). Nine plough-zone samples were taken and
one scraper and a fragment of a polished axe were found (Table 6:4).
No features were recovered.

Thy 2924 (6)


This is the smallest Late Neolithic site investigated in the Thy project.
Only one formal artefact was recovered: a fragment of a polished axe.
According to Steinberg (1997:174), the flake signature of the site has a
Late Neolithic character. However, the site is surrounded by Bronze
Age barrows, thus assigning it to Steinbergs Late Neolithic/Early Bronze
Age site category. The site covers an area of 4 950 m2 (Steinberg
1997:174).

The unproductive flint daggers of Thy


Steinberg concludes that his investigation of the production debitage
from the investigated sites revealed an interesting pattern. The flake
profiles established on each site suggested that, in this respect, the Early
Neolithic TRB sites were very homogeneous and the Late Neolithic

181

Jan Apel
sites fairly homogeneous, while the Bronze Age sites were heterogeneous.
Consequently, Steinberg suggests that there was no intra-site
specialisation during the Early and Late Neolithic and that specialisation
occurred in the Bronze Age. This investigation thus support Kristiansens
view that specialisation, and therefore also chiefdom complexity, occurred
during the Bronze Age. Unfortunately, I have not had time to visit the
Thy museum in order to investigate the archaeological materials from
the Thy project personally. However, after consulting Jens-Henrik Bech,
the head of the Thy museum, by e-mail, it became apparent that no
dagger-production sites had been investigated in the Thy region. The
only specialised production sites that were recovered in the project were
Bronze Age sites with evidence of bifacial-sickle production. The fact
that no specialised Late Neolithic production site was investigated within
the Thy project was not, as I see it, a result of the fact that there is no
technological specialisation was present during this time. On the contrary,
it is a strong indication of intra-site specialisation during the Late
Neolithic. The geographical scale of the Thy project was simply not
large enough to contain the complexity of Late Neolithic craft
specialisation in southern Scandinavia. It is quite possible that the results
of Steinbergs investigation would have been different if the surveys
had been carried out in the parishes of Thy, such as Thisted, where Late
Neolithic flint mines have been found. It is likely that more specialised
production sites were located in this area. Steinbergs conclusion would
certainly have been different if he had considered the Late Neolithic
sites in Himmerland and Djursland, such as Myrhj, that are presented
in this chapter. It seems as if the investigation of the Late Neolithic sites
in Snderh parish managed to map only the domestic aspects of flint
work, simply because the specialised production took place elsewhere.
As will be evident in the conclusion of this chapter, my opinion is
that it is impossible to understand the political landscape of the Thy
region without considering an area covering, at least, the central and
northern parts of Jutland. The specialisation of the Late Neolithic period,
that Steinberg (1997) and Earle et al. (1998) failed to recognise, is
evident on considering a larger geographical area, including sites with
evidence of flint extraction and dagger production. When Steinberg
proposes that the flint daggers were unproductive artefacts in the exchange
networks of prestige objects, as is implied by the following quote, he is
simply ignoring plain facts: The Late Neolithic is characterized by a
182

Flint technology in Southern Scandinavia


decreased environmental exploitation, little monumentality, and unproductive daggers (Steinberg 1997:176). This conclusion differs
radically from the results presented in this thesis (see Chapter 10).
Scandinavian flint daggers were exchanged within a network covering a
region including the Rhine delta, northern Germany, the north-western
part of Poland, Denmark and large parts of the Scandinavian peninsula.
A few daggers even reached areas located far away from the production
sites, such as Austria and the northern parts of Scandinavia (see Chapter
9). I would even go so far as to suggest that it is impossible to understand
the political landscape of the southern Scandinavian Late Neolithic
without relating it to a larger region covering the north-western and
central parts of Europe, as well as the central and northern parts of
Scandinavia.

Eastern and southern Jutland and northern


Germany
A number of settlements dated to the Late Neolithic and Early Bronze
Age have been found in the eastern parts of Jutland, mainly on the
Djursland peninsula (Fig. 6:1).

Egeh j (7)
At Egehj in Djursland, three houses with rows of single, roof-bearing,
post holes were recovered not far from the Egehj barrow. This barrow
has been excavated and contained a burial dated to the Early Bronze
Age (Boas 1983:90). One hundred and twenty kilograms of flint was
found at the site. More than 20 000 waste flakes and many artefacts
were excavated. Bifacial arrowheads with hollowed bases make up the
biggest find category, a total of 90 examples being found. Twenty-one
daggers and dagger preforms were recovered (Boas 1983:95). Two types
of daggers are represented in the material Lomborgs type V and
type VI. The type VI daggers are all under 13 cm, which, according to
Lomborgs classification, would make them strike-a-lights of type A
or B (Lomborg 1973:27 f., 1959:160). Three of these dagger/strike-alights are complete and show no wear pattern and this may mean that
they were produced on the site. Apart from the fact that one would not
expect used tools on a production site, Lomborgs length definition is
hard to accept. Further down the chain of consumption, only distinct
wear patterns or contexts (hearths, etc.) would permit a strike-a-light
183

Jan Apel
interpretation before a dagger interpretation. However, it seems to be
clear that these specimens should be dated to the Early Bronze Age.
The flint at Egehj is of a very good quality. Only 18 % of the
scrapers at Egehj have traces of cortex on their dorsal surfaces. This
should be compared with the Myrhj site, where 80 % showed signs of
cortex. Only 1 % of the flint debitage consists of short and wide flakes.
Boas suggests that this kind of flake resulted from axe manufacture and
that the higher figure at Myrhj (26 %) is explained by the large number
of axes that were found there (Boas 1983:95). Axes were not found at
the Egehj site. This reasoning implies that Boas was acquainted with
axe-production technology and was able to recognise the morphological
attributes of its debitage. If this was the case, it is not apparent in the
site report. One way of confirming this hypothesis would be to reexamine the Myrhj material, classify the flakes by flint and see in which
category the majority of the flakes fall. Good-quality flint would imply
cutting tools and poor quality flint would imply chopping tools.
Another way to verify the hypothesis would be to make a technological
analysis of the flakes from Myrhj and compare the material with the
results of square-axe debitage (Hansen & Madsen 1984) and with the
flint-dagger debitage, as it is presented in this volume. It is important
to remember that other reduction methods than square-axe production
produce short and wide flakes. Flakes with attributes like these are
produced in the making of most bifacial-tool types and, as was the case
with the Myrhj material, one would have to consider a range of
technological information, i.e. types of platforms, etc., before a
judgement could be made.
The distribution of the flint material allows Boas to make a detailed
interpretation of different activity areas and deposits in the settlement.
He divides the flint material into three categories in order to examine
specific activity areas (Boas 1983:97): (1) everyday items such as simple
flake tools, (2) finished pressure-flaked tools, and (3) production
debitage and preforms in flint (Boas 1983:97). The last group is seen as
indicating specific activity areas connected with the production of bifacial
arrowheads. Boas concludes that this, therefore, seems to be an example
of specialized arrowhead production, with rough shaping taking place
near the first roof-bearing post, finer working between this post and
the gable wall, and the apparently rejected examples of rough-outs being

184

Flint technology in Southern Scandinavia


disposed of against the end wall itself (Boas 1983:97). Boas was also
able to detect another production area similar to the above-mentioned
and, in addition to this, he suggests that the east end of one of the
houses was a specific activity area reserved for the production of daggers
and sickles (Boas 1983:98). It is thus possible to distinguish specific
activity areas by a simple analysis of distribution (Boas 1983:99). I do
not agree with Boas on this point. The arguments that he presents in
favour of his ideas are of a clearly post hoc character (Binford 1989:8).
On a site like this, that may arguably have been used over a long
period of time, it is unwise to talk about specific activity areas without
a proper discussion of cultural and non-cultural, transformation processes
and the way in which these are liable to affect the deposits (Schiffer
1987). It is especially important to evaluate the cultural processes (ctransformations) and their relation to a site of this kind. It is likely that
continuous maintenance of the site during its use would affect the
distribution of the artefacts within the habitation area in ways that would
make it impossible to present reliable and specific interpretations, as
Boas does (OConnell 1987). However, he seems to have been aware
of these problems (Boas 1983:99) and, in spite of these remarks, one
cannot argue with the fact that it is likely that the settlement represents
a production site. To make a more specific interpretation, i.e. to say
whether it is possible to distinguish between primary and secondary
deposits, a more thorough analysis must be undertaken. This would
involve not only a simple analysis of distribution but also a thorough
analysis of distribution patterns based on the results of a technological
classification of the waste material.

Hemmed church and plantation (8)


These two sites lay approximately 150 m apart, and traces of 12 houses
were recovered. The sites are dated from the transition between the
Middle and the Late Neolithic to the Early Bronze Age (Boas 1991:119).
As at Myrhj, wind-transported deposits in ancient times sealed the
sites, and this has preserved the cultural layers in very good condition.
Numerous finds were made in the excavations. I shall restrict myself to
the flint finds. At Hemmed church, 106 kg of flint waste, 88 flint
cores and 492 flint implements were found (Boas 1991:119). At
Hemmed plantation, 107 kg of flint waste, 66 flint cores and 443 flint
implements were recovered (Boas 1991).
185

Jan Apel
Boas was able to establish a chronology of the houses that stretches
from the Middle Neolithic to the Late Bronze Age. This scheme also
includes one of the houses at Egehj (Boas 1983). It is interesting to
note that the earliest Late Neolithic house, House III at Hemmed
plantation, contained fragments of type-I daggers, that the latest Late
Neolithic house, House I at Hemmed plantation, contained the hilt of
a type-IV dagger and that the earliest Bronze Age house, House III at
Hemmed church, contained a type-V dagger (Boas 1991:125 ff.). This
confirms that lancet-shaped daggers date to LN I and that daggers with
fishtailed handles date to LN II or the Early Bronze Age. In fact, these
datings also support Lomborgs chronological distinction between types
IV and V. For a thorough discussion of the chronological value of the
dagger typology, see Chapter 8.
Most of the artefacts at Hemmed church were recovered in house
III. This applies to the flint waste as well. Boas does not interpret any
activity area in connection with the waste and implements of flint. One
would suspect that a technological examination of the waste material in
house III would give us information as to which flint-knapping was
carried on and, if that was the case, what reduction strategy was in use.
However, This has not been done.

Svapk rret and Diverh j (9)


These two Late Neolithic sites are situated in Djursland, the eastern
part of Jutland (Fig. 6:1). The Svapkrret site was found during the
investigation of a barrow that had been ploughed. The excavation was
conducted in 1968. The report that is referred to in the following text
is Boass re-interpretation of the site, published in 1986. The barrow
contained burials dated to the Late Neolithic and Early Bronze Age
(Boas 1986: 320). In one of the Late Neolithic graves, a type-I dagger
and three amber beads were recovered. In the filling between the stones
constituting the grave, charcoal, flint waste, ten flint drills, nine flint
scrapers, fragmented pottery, and a couple of amber beads were found
during the excavation (Boas 1986:320). Traces of ploughing done by a
simple wooden ard were recovered underneath the burial, together with
three post holes. A concentration of flint waste was found adjacent to
the burial, together with a stone the size of a human head with crush
marks on top. A couple of flint arrowhead preforms, a dagger and an
axe preform were recovered in the post holes.
186

Flint technology in Southern Scandinavia


Boass interpretation of the site is based on the presence of the post
holes. He suggests that they formed the western part of a two-aisle
house, like those at Myrhj, and that the sunken floor may be found
east of the excavated area (Boas 1986: 321). Among the flint waste
from a cultural layer surrounding the flint concentration and from the
filling of the grave, six fragments that seem to derive from type-I daggers
were registered (Boas 1986: 321). Boas does not discuss the amount or
the quality of the flint waste recovered at the site.
Another Late Neolithic site was also detected, owing to the investigation of a Bronze Age barrow Diverhj near Tustrup (Fig. 6:1).
Among other things, an Early Bronze Age burial in an oak coffin was
found in the earliest feature of the barrow a stone cairn. The cairn
was surrounded by traces of plough marks. Boas suggests that they
belonged to a field that was in use immediately before the construction
of the cairn (Boas 1986: 318). A lancet-shaped, flint dagger was found
in the secondary filling of the grave. The dagger, together with other
recovered artefacts, such as flint arrowheads and pieces of amber, may
belong to a burial dated to LN I. The filling of the cairn also contained
a compact mixture of flint waste, pottery sherds and charcoal. Post
holes from three houses were recovered beneath and beside the barrow.
In one of the houses, a depression measuring 3 m x 3 m x 0.3 m was
located in the eastern part.
In spite of the extensive cultivation that seems to have taken place
on the site in the later part of the Late Neolithic and the Early Bronze
Age, a cultural layer belonging to the period when the houses were
inhabited still remained (Boas 1986:319). Among the artefacts collected
in this layer were flint waste (c. 100 kg), finished tools and preforms in
flint, pottery sherds, and some pieces of amber. All in all, 16 type-I
daggers and preforms of type-I daggers were recovered, together with
10 flint arrowheads and 20 arrowhead preforms (Boas 1986:320).
Nothing is mentioned about the technology or the quality of the flint
waste. I tried to contact Boas in order to arrange a meeting during a
five-week visit to Jutland in the summer of 1998 but was unsuccessful.
Thus, I have not been able to study the production debitage from any
of the previous sites in person. However, I find it likely that daggerproduction debitage may be present on all three sites. I hope to be able
to return to and investigate the flint from these sites more closely in the
future.
187

Jan Apel

Vejlby (10)
This site is located on a slope in the Eg valley at Vejlby parish in the
Hasle district, north of the town of rhus (Fig. 6:1). At the site, a pit
(4.5 m x 4 m) was excavated and was interpreted as the sunken, eastern
end of a Late Neolithic/Early Bronze Age house, such as the ones
discussed above (Jeppesen 1984:101 f.). Four post holes were also
revealed and they seemed to constitute part of the slightly rounded,
western end of the house (Jeppesen 1984:102). If the house
interpretation is correct, Jeppesen suggests that we are dealing with a
house at least 9.5 m long and 5 m wide. The pit itself contained charcoal,
numerous small fragments of fire-cracked stone, flint objects, flint waste
and pottery (Jeppesen 1984:99). I have personally inspected the flint
material from the excavation and there are no signs of the production
of bifacial tools on the site.

Forn s (11)
Most of the dagger-production sites that have been discussed so far
have been found in settlement contexts. There are, however, also
production sites with no evidence of structures, such as houses or huts.
Forns in Djursland is such a place. The site is located on the beach just
beside Sangstrups Klint, a white limestone rock that dives into the
Kattegatt. The beach is littered with nodules that have fallen down
from the cliff (Glob 1951:24). There are three types of raw flint material
at Forns; the first is dark and the second is lighter (white-brownish
or white-grey). The Late Neolithic flint-knappers used both of these
types extensively. The third flint type is grey and was used almost exclusively for axe preforms (Glob 1951:26). These flint types correspond
to the senonian and danian flint types mentioned earlier. As was the
case at Myrhj, the better-quality, senonian flint was used for more
complicated flint work, such as dagger production, while the dryer,
danian flint was used for axe production. This may reflect a convention. It seems clear that axes were seldom made of the finest quality
flint, not even at places where the raw material was abundant. During
the Late Neolithic period, large nodules of finer quality flint were exclusively used in the production of bifaces. This coincides with
observations that Errett Callahan has made in Danish museum
collections: Danish flint daggers are almost exclusively made of senonian
flint. This is, however, a rule with exceptions. Flint daggers were indeed
188

Flint technology in Southern Scandinavia


made of rougher materials. There are daggers made in danian and
Kristianstad flint and in Norway there is at least one example of a typeI dagger made of quartzite (Scheen 1979).
During the excavation of the site, several heaps of flint waste were
found (Fig. 6:11). These were interpreted as primary deposits the
results of flint-knapping (Glob 1951:26). At one place, the points of
six flint daggers were found in the same heap, and the associated handles
were found in another heap 2-3 m away. Glob suggests that the flint
knapper broke the daggers while making them (Glob 1951:26). This is
a reasonable explanation, but when Glob continues by interpreting these
dagger fragments as evidence of a fast working pace, namely that the
flint-smith produced them in one day and that he probably also
produced several finished daggers as well, he is indulging in pure
speculation. This pattern is more likely to be interpreted as evidence for
the structure of the production site and probably represents a period of
time. This phenomenon has been described as a form of contextual
contemporaneousness (Ahlbeck 1995:42). The term pinpoints the fact
that, when the structure of a site or an activity area is maintained and
preserved over a period, archaeologists are liable to recognise them and,
just because of this, they often infer that they are the result of a short
period of habitation. In addition to this, it is unlikely, on the basis of
the technological experiments and the time studies (Callahan & Apel,
MS), that one single flint-knapper produced the six daggers and more
in one day as Glob suggests (Glob 1951:27).
It is interesting to note that the site does not contain used flint
implements or large amounts of pottery. Except for some hearths, no
features were found at the site. This suggests that Forns was a specialised
production site focused exclusively on the extraction and initial processing
of the raw material. A large part of the artefacts found consists of
preforms. This is indeed an indication that the site manifests the early
production stages. In all, 663 artefacts and fragments of artefacts were
recovered (Table 6:5). As we can see, the grey, dry, danian flint was used
only for axes, strike-a-lights and blade scrapers, while the better quality
flints, especially the black variant, was used for a variety of artefacts
with cutting edges. Most of the axe preforms are of poor quality and
they obviously represent rejects (Glob 1951:31). The biface preforms
are fairly small, between 10 and 18 cm long, and Glob suggests that
they were sorted out from the piles of preforms that were exported
189

Jan Apel
Artefact categories
Axe preforms, complete
fragments
Bifacial preforms, whole
fragments
Sickle Preforms, whole
fragments
Axes
Daggers, whole
handles
blades
cracked
Sickles, whole
fragments
Spearheads
Arrowheads
Strike-a-lights
Blade scrapers
Borer
Hammerstones, flint
stone

Black flint
17
16
66
72
27
51
6
2
38
15
5
15
31
3
4
8
17
2

W/G flint
20
17
1
4
4
1
7
7
1
1
1
1
8
11
7
3

H/B flint Grey flint


20
12
17
13
9
14
2
1
1
1
2

2
1

Total
69
63
76
90
33
52
8
2
46
24
6
16
32
4
12
21
25
5
6
65

Table 6:5. Breakdown of totals of flint objects at the Forns site (from Glob
1951:30, Fig. 8).

(Glob 1951:32). This is a fascinating idea that needs further


investigation. It implies that the Forns site represents a site for the
processing of early-stage preforms for axes and bifacial tools (maybe the
first two stages in Callahan 1979, see also Chapter 2) and that the
subsequent production stages were performed elsewhere.
Glob does not discuss the waste material found. An analysis of this
material (such as the one proposed in this book) would make it possible
to distinguish early production stages from later. If this were done, it
would be interesting to compare the Forns site with a potential
production site, maybe one of the sites mentioned earlier. The waste
material seems to be of key importance when it comes to understanding
the organisation of the Late Neolithic, flint-dagger industry. If it is
possible to map the production chain, then this would open the door
for economic and social interpretations of Late Neolithic society in
general.
The daggers and the preforms are of types I-III. This implies that
the site dates from LN I like most of the Late Neolithic sites in Jutland.
Table 1 shows that 46 dagger handles were found but that only 24
dagger blades were observed. As Glob points out, it is likely that the
rejected flat blades were used as preforms for spearheads (Glob 1951:33).
He also suggests that some of the handle parts could have been used as
strike-a-lights (Glob 1951:34). From the typological point of view, it
190

Flint technology in Southern Scandinavia


is important to note that there were type-II daggers at the site. As we
will see, the distribution maps of type-II daggers clearly suggests that
they were produced predominantly on the eastern Danish islands and
in the south-western part of Skne (Chapters 8 & 9). Thus, it may be
wise not to take generalisations too far.
Debbie Olausson (1997 & 2000) has investigated the material from
Forns that is kept at the archaeological museum at Moesgrd in Jutland.
She maintains that the flint-knapping at Forns was of a low quality
(Olausson 1997 & 2000:129. Obvious knapping errors, indicating
moderate knapping skills and a poor technological know-how, were
frequently recognised in the material from the site. After having inspected
the Forns material personally, I agree with Olaussons conclusions.
The excavators saved only preforms and cores. Unfortunately, the
production debitage, in the form of complete and fragmented flakes,
was not collected and cannot therefore be studied (Fig. 6:11).
When interpreting the site one must bear in mind that the discarded
and broken preforms were left on the site because they were considered
to be useless. Another important aspect worth considering before
presenting an interpretation of this context is that the daggers, that
evidently had been used, were found at the site. Presumably, it is these
examples that Olausson refers to when she states that there also
pieces showing good controle and high workmanship in the collections
from the site (Olausson 2000:129). I observed that these daggers are
made in other flint qualities than those of the preforms. They were
brought to the site used and deposited there. Thus, it can be stated that
only early- to middle-stage preforms were made at Forns. It may very
well be that the few daggers of higher quality from the site served as a
kind of model visualising the size and character of the desired preforms.
If this assessment is correct, the desired preforms may very well have
been taken from the secluded Forns site in order to be refined into
finished daggers elsewhere. In Chapter 2, it was suggested that the early
stages in the production of the daggers were carried out by younger,
inexperienced apprentices on their way to becoming artisans. This scenario
would in fact explain the apparently low quality of the knapping in
evidence at the site. Young apprentices were sent here to gather suitable
flint nodules and turn them into preforms with the desired qualities
that could be refined into finished daggers by artisans and masters in
public at settlement sites such as Myrhj, Gug and Diverhj.
191

Jan Apel

Tegelbarg (12)
Not far from the town of Flensburg in northern Germany on a former
shoreline of the Baltic Sea, a Late Neolithic site was excavated by Volker
Arnold at the beginning of the seventies. Considerable flint waste was
recovered at the site. All artefacts were plotted within a quarter-metre
grid, and this accuracy made it possible to distinguish between
concentrations of flint waste that could be analysed. A large proportion
(75 %) of flint artefacts and production debitage recovered on the site
was found in a two square meter area. Arnold is of the opinion that the
heap did not represent a primary deposition but that it should be
regarded as a secondary deposition of material from an as yet unrecognised
knapping site (Arnold 1990). In cases like this, in which a secondary
deposition has been found, it may be possible to recognise a nearby
primary deposition by taking samples with a probe every metre to detect
micro debitage that, owing to its size, missed the maintenance of the
activity area (af Geijerstam 1996). Of course, if a blanket or some other
kind of foundation had been in use during the production, this would
not have produced such results, but then the micro-debitage could have
been detected in the secondary deposition (Thorsberg 1984:47 ff.,
Schiffer 1987:63). If so, one way to define an activity area of this kind
would be to make a spatial interpretation of the documented scatter
pattern. In Callahan & Apel (ms), the scatter patterns documented during dagger-production experiments will be presented and discussed.
From the method of refitting, Arnold concluded that the heap
contained debitage from the production of at least four different flintknives. Among them, two seemed to have been lancet-shaped daggers
and one a rough-out, perhaps for another dagger. Only a few flakes
could be assigned to the fourth knife (Arnold 1990). The waste of the
other three specimens consisted of 80 flakes from a broken, unfinished
Lancet-shaped dagger, 87 flakes from a dagger rough-out, and 155 flakes
that revealed an almost complete, dagger-reduction sequence. The original
blank for this piece could be completely reconstructed (Arnold 1990).
The refitted part is like a cast, and the lancet-shaped-dagger interpretation
is based on the shape of the missing piece in the middle of the blank.
The produced dagger was not recovered at the site.
The successful refitting of this dagger allowed for some interesting
technological conclusions. Early in the production sequence, hardhammer percussion had been used. This activity is interpreted as having
192

Flint technology in Southern Scandinavia


been performed elsewhere, because the flakes resulting from such activity
are lacking in the flint heap, and the activity (hard-hammer percussion)
revealed itself only as two opposite bulb scars on the blank (Arnold
1990:213). After this initial treatment of the blank, Arnold suggests
that the preform was brought to the knapping floor. It may be suggested
that these initial missing flakes were detached somewhere nearer the
raw-material source and left there. Another possibility is that the large,
initial flakes were themselves used as raw material for the production of
other artefacts, such as spoon-shaped scrapers or small bifacial sickles.
Anyway, from now on, subsequent treatment exclusively involved soft
percussion with either an antler or a soft hammerstone: During the
following finishing a gradual improvement of techniques can be
registered (Arnold 1990:213). This statement implies that Arnold did
not notice or try to detect traces of different production stages in the
material. The debitage shows evidence of a smooth grinding or abrading
that was performed to strengthen the platforms before each detachment (Arnold 1990:214).

The Danish isles and Skne


Drenge s at Sejer (13)
On the island of Sejer off the north-western coast of Zealand, several
knapping floors have been located on beach walls were large numbers
of rolled flint nodules are easily obtained (Kempfner-Jrgensen &
Liversage 1985:18). In order to learn more about the extent to which
these flint-knapping workshops were used, The National Museum
conducted an excavation on one of these sites, a beach wall called
Drenges, from 1982 to 1984. This site was found during the 1950s,
when the local people collected preforms for bifacial sickles and daggers
in a ploughed field situated on the beach wall. The workshop at Drenges
covered an area of 600 m2. The flint at the site stems from two
chronologically separated layers, from both of which bifacial sickles
and flint daggers have been produced. The older layer, which presumably
dates from LN I, contained white-patinated flint. In this layer, a type-II
B dagger (Fig. 6:11), a small bifacial sickle, a preform for a broad-edged
axe (all of flint) and a polished axe (of slate) were recovered. At least
ten, failed, sickle preforms, four dagger preforms and two axe preforms
were also found here (Kempfner-Jrgensen & Liversage 1985:22). This
layer was sealed off by sand, presumably as the result of a storm. In the
193

Jan Apel

Fig. 6:11. Two flint daggers found in the separate layers at Drenges. On the
left, a type-III A dagger from the earlier layer and, on the right, a broken
preform of a type-VI dagger from the later layer (from Kempfner-Jrgensen &
Liversage 1985:23 f., Figs. 8 & 10).

later layer, which was characterised by flint without secondary patina,


the largest number of flint artefacts was recovered. Among other things,
it contained a type-VI-dagger preform (Fig. 6:11) and pottery dating to
the Early Bronze Age. Over 100 sickle preforms, twelve axe preforms,
four dagger preforms and three preforms for bifacial arrowheads were
also recovered. In addition to these, a number of simpler artefacts, such
as scrapers, borers and knives were also found in the later layer.
The site was interpreted as a flint workshop that had been visited for
short periods of time. It is suggested that it was used primarily to produce
sickles that could be used as harvest knives (Kempfner-Jrgensen &
Liversage 1985:25). Three hearths were the only features recovered during
the excavation. On account of the large amounts of failed preforms and
production debitage, it has been estimated that between 450 and 670
flint sickles were produced at the site on the six to ten occasions when it
was in use (Kempfner-Jrgensen & Liversage 1985:26). The raw material
194

Flint technology in Southern Scandinavia


was obtained from the beach wall and the general quality of this flint is
poor, a fact which help to explain the high ratio of failed preforms.
I have not had the opportunity of studying the material from
Drenges in person, but fortunately, Olausson has studied this material
as well. Her investigation reveals that there is a high degree of knapping
errors in the material, something that partly can be explained by the
quality of the raw material (Olausson 2000:129).

Lindebjerg (14) and R jle Mose (15)


At Lindebjerg and Rjle Mose on northern Fyn, two early Bronze Age
settlements have been excavated (Jger & Laursen 1983). On both
sites, the remains of some sort of building were partially recovered.
Flint daggers of type VI were found in both settlements and date the
settlements to the early part of the Bronze Age (Jger & Laursen
1983:114 f.). In the 1983 report by Jger & Laursen, it is not possible
to discern how many daggers and dagger fragments were found in the
settlements. In the List of Finds (109, Table 1), daggers are grouped
together with strike-a-lights and spearheads. At the Lindebjerg site, these
find categories consisted of a total of eight, and at the Rjle site, of 46
examples. A dagger preform, found at Rjle Mose, indicated that daggers
were produced at the site. However, this preform, judging by the
published drawing, is very crude and relatively thick. It is unlikely that
it could have resulted in a finished dagger of the same quality as the
other daggers found on the site. The specimen may have been a reject,
but my own personal experience of bifacial flint-knapping indicates
that it was a less experienced flint-knappers best try. In cases like these,
Pelegrin has suggested that it may be justifiable to infer the obvious
lack of know-how as, for instance, the result of the activity of (young?)
individuals during a period of acquisition of know-how (Pelegrin
1989:121). Callahan refers to this stage in the development of a
flintknapper as Learning Phase A (1979:35). This observation seems
to coincide with comments regarding the flint technology at the
settlements. It is stated that the waste material consists of irregular flakes,
most of them probably resulting from a direct, hard hammer technique
(Jger & Laursen 1983).
There is also evidence of smaller, thin flakes with very small, striking
platforms. These may have been the result of pressure-flaking. At Rjle
Mose, 52 arrowheads of flint were recovered. All of these are bifacially
195

Jan Apel
flaked. Judging from the small pressure flakes, some of the arrowheads
may have been produced on the site. In general, a lot more flint tools
were found at Rjle Mose than at Lindebjerg.

Str by (16)
In 1934, Therkel Mathiassen published an article on the Neolithic flint
trade. The article was initiated by numerous flint artefacts discovered in
an area surrounding Strby Ladeplads at Stevns on eastern Zealand (Fig.
6:1). Here, collectors had found hundreds of artefacts, mainly preforms
for different tool types, made when the area was exploited. Mathiassen
suggested that the prehistoric peoples of this area exported preforms for
artefacts that were finished in other places (Mathiassen 1934:22). Most
of the finished artefacts that were recovered were axes or axe preforms,
but some of the square-shaped preforms were so thin that they were
more likely dagger preforms. If this was the case, there seem to have
been at least two fundamentally different methods of creating preforms
suitable for daggers, the bifacially almond-shaped preform and the square
preform. Maybe these two methods could be spatially distinguished
from each other (Jutland/Zealand?). In any event, a finished flint dagger
has been discovered among the preforms, but Mathiassen does not discuss its morphology (Mathiassen 1934:20).
The danian flint at Stevns is tannish-grey and calceous and is not as
brittle and easily worked as the senonian flints of Stevns, Falster/Lolland
or northern Jutland. Thus, it is not quite suitable for advanced dagger
production (Callahan, personal communication). As we have seen above,
the dagger-production centre moved from northern Jutland to the Danish
isles in the transition between LN I and LN II (Vankilde 1989). The
not-yet-found production sites of large, prestigious, type-IV daggers
are therefore likely to be found on Lolland/Falster and perhaps Zealand.
It is possible that they may be found as the result of a specific survey
aimed at detecting sites like these. The uniformity that these daggers
show indicates that they may have been the products of a single production site with quality control (Callahan, personal communication).
This is obviously not the case with the earlier types. An increased specialisation that occurred during LN II may therefore explain why production sites of this phase are scarce.

196

Flint technology in Southern Scandinavia

Sk ne
Skne, the southernmost part of Sweden, should be mentioned in this
context. This part of Sweden has more in common with Denmark
than with other parts of the country. For instance, this is the only area
in Sweden where flint is abundant. In 1951, Mrta Strmberg wrote
that there were approximately 60 known settlements with material from
the Late Neolithic period in Skne. However, almost all these sites were
mixed settlements with material from other periods (Strmberg
1951:72). The Late Neolithic artefacts on these sites were few, compared
to the materials from other periods.
On some of these sites, especially around the Barsebck area and in
the Malm area, numerous flint daggers have been recovered (Althin
1954: sites 55, 56 and 56a). It is likely that there are many production
contexts to be found in these parts of Skne. However, for the time
being, as regards dagger-production sites, I shall restrict myself mainly
to the data I have gathered from Denmark.

Late Neolithic flint mines


It was Sophus Mller who first suggested that flint had most probably
been prospected for in mines during the southern-Scandinavian Neolithic
(Mller 1897:167). In 1906, the first Scandinavian flint mine was
recovered by the Swedish geologist Nils-Olof Holst at Sdra Sallerup
outside Malm in Scania (Rudebeck, in press). Almost 50 years later,
the first Danish flint mines were found at Skovbakken near Aalborg
(Becker 1951a:107 ff., 1951b:135 ff., Grantzau 1954:30 ff. & Becker
1959:87 ff.). During the 1950s and 1960s, systematic investigations of
natural chalk deposits revealed flint mines at Thy in the western part of
the Limfjord area as well (Becker 1980:456 ff. & 1993:111). Most of
the Scandinavian flint mines can be dated to the Early Neolithic period
(Becker 1993:111, Rudebeck, in press). However, in this context, only
mines dated to the Late Neolithic period will be considered. Since there
is no evidence for Late Neolithic mining at Sdra Sallerup (Rudebeck,
in press), the Late Neolithic mines in the Limfjord area will be described
in the following paragraphs.
The Skovbakken mines (Fig. 6:12) are all dated to the Late Neolithic
period. In fact, the first mine was originally discovered when a cache,
consisting of 19 flint daggers, was found in one of the mine shafts

197

Jan Apel
(Becker 1993:111). At Skovbakken, the flint was mined in two different ways. In cases in which the flint was available close to the surface,
pits were dug into the chalk from the ground surface (Rudebeck, in
press). When the flint layer was situated at a deeper level, shafts were
dug sometimes up to 7 m into the chalk. From these shafts, galleries
were dug in several directions when the flint layers were reached (Becker
1993: 111 f., Rudebeck, in press). In 1993, Becker pointed out that no
Late Neolithic settlements had been found in the vicinity of the
Skovbakken mines (Becker 1993:112). However, during the last couple
of years, an excavation conducted by the Aalborg Historical Museum at
Bejsebakken, located c. 500 m from the mines at Skovbakken, has
produced several Late Neolithic houses and large numbers of formal
artefacts and debitage in flint. It is likely that the population that used
the flint mines lived here. The flint nodules that were extracted from
the Skovbakken mine-shafts were quite small. Consequently, only daggers of up to 20-25 cm could be made from these nodules (Becker
1993:112).
At Hov, near Hillerslev in the province of Thy in the western part of
the Limfjord area, smaller mines, dated to the Late Neolithic, were
recovered in 1957 (Becker 1980, Rudebeck, in the press). The mining
seems to have been carried on in a similar way to the mining at
Skovbakken. Circular shafts with a diameter of about 4 m were dug
into the ground. When a flint layer was found, galleries were extended
in different directions.
As was the case at Skovbakken, the Late Neolithic mines at Hov did
not, according to Becker, produce large nodules. Consequently, it was
concluded that the large daggers of type I C, which undoubtedly
originated from the Limfjord area, must have been made of flint from
mines that had not yet been recovered (Becker 1993:121). I have
personally collected large flint nodules at a present day chalk mine at
Hillerslev near Hov and must agree with Becker on this point. However,
I do not agree with his casual remarks on the origins of some of the
other dagger types. Becker does not acknowledge, as is suggested in this
book, that certain dagger types may have been produced almost
exclusively on the eastern Danish islands. Even though Becker was aware
of the fact that the majority of the caches of certain types of daggers, for
instance, types I D, II and III, were located on the eastern Danish islands,

198

Flint technology in Southern Scandinavia

Fig. 6:12. Cross-section and horizontal plan of three Late Neolithic mines at Skovbakken
near Aalborg (from Becker 1993:112, Fig. 1).

he insisted that they should not be thought of as eastern Danish variants


(Becker 1993:122). I do not agree with this conclusion.

Conclusion
Some of the investigated sites display production debitage that contain
a high proportion of knapping errors. Interestingly, these sites were
located to secluded places near the extraction area (for instance Forns
and Drenges). On the settlement sites however, such as Gug and
Myrhj, evidence of high quality knapping as well as knapping of lower
quality has been recognised. A closer examination of the flint production
debitage from Myrhj revealed that debitage from both the production
of everyday items and debitage from the production of more elaborate
artefacts, such as thick-butted axes and daggers, are present on the site.
It is suggested in this context that the range in the quality of the flint
industry at the Gug, investigated by Olausson (1997 & 2000), may be
explained in a similar way.
199

Jan Apel
Thus, the most important result in this chapter is that the production of flint daggers usually took place on at least three different types
of sites. In the following paragraphs, I shall relate these sites to some
theoretical issues, regarding the social organisation of complex
technologies in traditional societies, discussed in Chapter 4. First of all,
the flint was extracted at certain locations where the raw material was
readily available. Obviously, the flint mines at Hov and Skovbakken
represent important links in the chain of production. However, certain
beaches, such as those at Forns and Drenges, where flint nodules were
washed clean by the water, were probably important as extraction sites
as well. In fact, since the Late Neolithic farming methods were based
on the use of ards, there is no reason to believe that flint nodules were
ploughed up to the surface in the way that happens with modern farming
techniques. Thus, on Lolland and Falster, where large flint nodules of
good quality can be picked up in the ploughed fields of today, suitable
nodules would have had to be prospected for in prehistory, just as chunks
of obsidian are located with iron rods and dug up by hand at Flint
Butes in present-day Oregon (Callahan, personal communication).
Alternatively, nodules could be collected on the beaches, where they
were washed clean from soil and were readily available for collection. It
has been suggested that the limited availability of high-quality flint
nodules for the production of prestige artefacts, such as axes and daggers,
during the Neolithic might have contributed to the fact that a large
part of the production can often be found outside the settlements
(Knarrstrm 2000a:163). However, in the following conclusion, I shall
focus mainly on the relation between the intra-site, spatial distribution
of the dagger technology and the theoretical issues discussed in Chapters
2-4.
The early stages of the dagger production were, owing to a relatively
low degree of know-how, conducted at secluded and private places,
preferably near the natural flint resources. Just as regards traditional ironworking in Kenya, where the smelting is surrounded by prohibitions
and taboos, these stages were shared with other people apart from those
directly involved in the production. The low degree of know-how did
not guarantee that the know-how could be kept exclusive by within the
group if the technology was conducted openly. The symbolic capital
that was connected with the craft could only remain a valuable asset as
long as the tricks involved in the production could be kept secret.
200

Flint technology in Southern Scandinavia


Consequently, the sites where the early stages of the production took
place were located far from settlements and other, more public places.
The final stages of the production were, on the other hand, conducted on the settlement sites. Myrhj is an example of a settlement
where the final stages of dagger production took place, side by side
with domestic and more expedient technologies, such as the making of
scrapers for everyday use. In this context, the final stages of daggermaking were put on display, in much the same way as the traditional
blacksmiths of the eastern and central parts of Africa practised the forging
of iron into tools in public places. Since the know-how involved in
these stages could not be imitated, but that it was adjusted to the muscle
memory of the flint-knappers body through years of practice, the
exclusiveness of the technology was not threatened by performing it in
public places. On the contrary, the direct confrontation with the public
ensured that the formalised, technological language could bring forward
its authoritative message of the past in the present to the members of
the community. In the following chapter, it will become evident that it
was not only the population of southern Scandinavia that knew of the
powers of the flint-knappers. Indeed, the flakes from the final percussion
stages of dagger production had a symbolic value in all agrarian
communities in the southern and central parts of Scandinavia. The lack
of known dagger-production sites dating to LN II suggests that an increased specilisation took place around 1950 cal BC.

201

Chapter 7

Flint Technology
in Eastern Central Sweden

The title of this chapter may seem to be a contradiction in terms to


anyone familiar with the Late Neolithic settlements of eastern central
Sweden. It is justifiable to question the notion that it is meaningful to
use the flint artefacts from this area to demonstrate the presence of any
long-lived, technological tradition, since they tend to be of low frequency and of poor quality. In addition, the few, formally defined artefacts that have been found are most likely to be regarded as imported.
Nevertheless, in this chapter, an attempt will be made to understand
how flint was treated and perceived in an area with no natural flint
resources. The results will be related to the knowledge that has been
gained regarding the advanced flint technology of southern Scandinavia
during the same period (see Chapter 6). The incorporation of this study
in the present thesis is justified by the fact that it will pinpoint how the
production of flint daggers in the southern parts of Scandinavia was
known and respected in other parts of Scandinavia. In fact, the social
message, or the mythological story, of the daggers forced the agrarian
populations in Sweden and Norway to adhere to a very formalised
technological tradition that, from the purely practical point of view,
did not suit them very well.
Technology is in this context regarded as a skill that can be reproduced from generation to generation (see Chapter 2 and Apel 2000a &
b). As we have seen in earlier chapters, this view inevitably forces the
discussion towards the social conditions in which the technology was
embedded. In order to use the definition successfully, it is of vital
importance that the recipes for action and the practical movements that
allowed the technology to be reproduced can be recognized in the
archaeological material. If not, the idea of a conscious and defined
technology will lack an empirical foundation and, in the processual
203

Jan Apel
perspective, it will then not be possible to incorporate the static archaeological material into a systemic context (Binford 1983:19 f.).
In the following pages, flint artefacts from 16 Late Neolithic and
Early Bronze Age sites in eastern central Sweden are discussed (Fig. 7:1).
In this context, eastern central Sweden consists of the provinces of
Uppland, Vstmanland and Sdermanland. The majority of the sites
have been interpreted as settlements. However, in order to broaden and
deepen the context of the production and consumption of flint in the
area, the flint materials from a Late Neolithic mortuary (Skillsta 2:15,
Skogs-Tibble parish in the province of Uppland), a ritual deposition of
thin-edged, broad-bladed, flint axes (Granath Zilln 1999) and four
Gallery-Graves have also been studied. The geographical distributions
of the sites and the flint artefacts are presented in Fig. 4:1 and Table 4:1.
The main idea behind this study is to establish how flint, in the form
of flakes and more elaborate objects, was used in social strategies to gain
power and prestige, similar to those used in the Late Neolithic
communities of southern Scandinavia.
The sites included in this chapter are not the only ones known
from the period in this part of Sweden. For an extensive presentation of
the archaeological sites dating to the Late Neolithic period and excavated
by the National Heritage Board (UV-mitt and UV-Uppsala) see Holm
et al. (1997). For a presentation of the Gallery-Graves of this region,
dating to the Late Neolithic and Early Bronze Age periods, see Apel
(1991 & 1992) and Holm et al. (1997). However, in my opinion the
information to be gained from the flint materials in these contexts would
not change the conclusions of this chapter in any significant way. On
the contrary, I believe that the 16 sites included in this study are representative of the flint industry in the currently known, Late Neolithic
occupation of this region. This is indicated by the repeated occurrence
of certain features of the flint industry on the sampled sites. The Late
Neolithic, flint materials in eastern central Sweden have been discussed
briefly in Holm et al. who concluded that there are few traces of a local
reduction in the settlements and that there is no evidence of a bifacial
technology. This conclusion is not entirely consistent with my own
experience of excavating the Late Neolithic sites of Fgelbacken in the
province of Vstmanland (Lekberg 1996) and Lugnet in the province
of Uppland (Lekberg & Apel, ms). I believe that it is possible to extend
this first impression given by Holm et al. in a way that may lead to a
204

Flint Technology in Eastern Central Sweden

Fig. 7:1. Late Neolithic sites in eastern central Sweden mentioned in the text.
Settlements are marked by circles, gallery graves by triangles and the ritual
deposition by a square. The small dots mark the parishes where flint daggers
have been found. The sites are projected on a map showing the coastline at
3600 BP (c. 2000 B.C. cal.), made by Leif Andersson at the Swedish Geological
Survey (SGU) in Uppsala.

deeper knowledge of the flint industry and its context and thus of the
role that flint artefacts may have played in the formation and
reproduction of the social order in eastern central Sweden during the
Late Neolithic period.

The sites
In this section, a selection of the Late Neolithic and Early Bronze Age
settlement sites with flint material treated in the following analysis is
briefly presented in order to inform the reader of the contexts in which
flint is to be found. Some of the sites included in this study have not
been properly published. This is mainly due to the fact that they have
205

Jan Apel

Fig. 7:2. A fragment of the handle of a type IIA


flint dagger recovered at Fgelbacken (ra
163). Drawing by Jan Jger (from Lekberg
1996).

been excavated recently. In these cases, the excavators have been kind
enough to let me present the flint materials, but any further presentation will have to wait until the post-excavation work is finished and
published. However, there is one exception. The Piparberget site was
excavated in 1998 and the excavators have kindly provided me with
additional data prior to the publishing of the report, thus making it
possible to include a brief description of the site in this context.

F gelbacken, ra 163, 160 and 162 (Hubbo parish in the province of V stmanland)
These three sites were discovered during salvage-excavations of a
prehistoric complex with several archaeological sites dating from the
Early Neolithic to the Viking Age (Lekberg 1997).
Ra 163 was located on a sandy plateau on the western shore of a
bog, c. 43 m above the present sea-level. Only two features were recovered during the investigation, a hearth and a pit with fire-cracked stone.
The artefacts consisted of pottery, a hammerstone, a grindstone and
some stone flakes, a core and a scraper in quartz. Twenty-seven flint
artefacts were recovered. Nine of these were complete flakes, while 16
were fragments of complete flakes. The formal artefacts consisted of
the tip of a bifacial arrowhead and a fragment of the handle of a type
IIA flint dagger (Fig. 7:2). Two samples, charcoal from the hearth and a
burnt hazelnut shell, were C14-dated to 2140 70 BP (Ua-10 000)
and 4420 90 BP (Ua-4091) respectively (1 sigma, Stuiver & Pearson
1986). Thus, the Late Neolithic dating of the site was based solely on
typology. The flint dagger seemed to be of an older type; the site was
therefore dated to LN I (2350-1950 cal BC).
206

Flint Technology in Eastern Central Sweden

Fig. 7:3. Arrowhead of porphyrite from Fgelbacken (ra 160). Drawing by


Jan Jger (from Lekberg 1996).

Ra 160 was located on a sandy plateau approximately 44 m above


the present sea-level and only c. 100 m east of ra 163. Two concentrations of fire-cracked stone and nine dark features, initially regarded as
potential post holes but later dismissed as of natural origin, were recorded
during the fieldwork. The recovered artefacts consisted of pottery, burnt
daub, a bifacial arrowhead of porphyrite (Fig. 7:3), 11 flakes or
fragmented flakes of quartz and quartzite, a drilling-centre core from
the shaft hole of a stone axe and burnt bone and hazelnut shells. Eleven
complete and fragmented flakes made of southern Scandinavian flint
and four complete and fragmented flakes of Kristianstad flint were also
recovered. A hazelnut shell was C14-dated to 335 85 BP (Ua-4092)
and the dating of the site to the Late Neolithic period therefore rests
entirely on typological reasoning. The pottery proved crucial in this
respect, especially as some of the recovered potsherds had a typical offset
beneath the rim and were decorated with cord-stamps, thus dating them
to the Late Neolithic (Lekberg 1996: Chapter 8).
Ra 162 was situated on the north shore of a bog on a sandy slope
approximately 42-46 m above the present sea-level, only about 100 m
north of ra 160. More than 60 features were recovered on the site,
including post holes, hearths, pits, concentrations of fire-cracked stone
and knapping floors where quartz had been worked. The recovered
artefacts consisted of pottery, two amber beads, burnt daub, clay figurines, burned human bones and quartz. One complete and nine fragmented flakes of southern Scandinavian flint and one flake of Kristianstad
flint were recovered. No C14-samples were analysed and the dating of
the site rests solely on the typology. The pottery belonged to two
207

Jan Apel
chronologically separated horizons: an Early Iron Age phase and a
Neolithic phase. The latter horizon, which concerns us here, consisted
of potsherds from at least 35 identified vessels that were classified as
Battle-Axe pottery of Malmers types Ad H and J (Lekberg 1996:
Chapter 9, Malmer 1962:8 ff. & 1975:26). These types are usually
dates to the later part of the Middle Neolithic period and the flint
material from this context would therefore appear to be of less interest
in this study. However, Battle-Axe pottery of the J type, now and then,
occurs in Late Neolithic contexts (M. Olausson 1995, Holm et al.
1997) and there is a strong possibility that this pottery was also used
during the early part of the Late Neolithic. Thus, the flint from ra 162
is also included in the following analysis.
The three sites at Fgelbacken were tentatively interpreted as making
up different parts of the periphery of a large, Late Neolithic, settlement
complex only partly affected by the salvage excavation (Lekberg 1996).
Thus, the flint material is treated as coming from a single site. During
the Late Neolithic period, when the sea-level was situated about 25-30
m above the present one, the Late Neolithic parts of the Fgelbacken
complex were on an inland site.

Lugnet ( sterunda parish in the province of Uppland)


The site is located on a sandy plateau about 45 m above the present sealevel. The remaining part of the site is now located in an enclosed pasture
that has not been cultivated. However, parts of the original settlement
were destroyed by gravel extraction in the northern part of the
investigation area and by extensive cultivation of a field adjacent to the
enclosed pasture. Berndt Andersson, a farmer and amateur archaeologist and farmer living nearby, has collected large numbers of Late
Neolithic artefacts in the cultivated field near the gravel pit. It was his
collection of artefacts, with detailed descriptions as regards where they
had been collected, that initiated archaeological excavations in the area
in 1998 and 1999 (Lekberg & Apel, ms).
During the excavation, some features in the form of a post hole, a
layer of clay, interpreted as part of a clay floor, and a pit containing
burned bones and potsherds (possibly the remains of an Iron Age
inhumation grave) were recovered. The find material from the site is
presented in Table 7:1. The site was dated on typological grounds. Two
bifacial arrowheads with concave bases and a certain kind of pottery
208

Flint Technology in Eastern Central Sweden


Find Material
Flint
Stone
Burned bone
Iron
Slate
Quartz
Sandstone
Porphyrite
Pottery
Burned daub
Quartzitic sandstone
Porcelain

Number
54
148
205
6
4
20
116
1
163
236
3
1

Weight (g)
45,4
2027,3
34,8
66,4
2,2
42,5
1238,2
1,1
164,6
351,6
71
0,9

Table 7:1. The find material recovered at the Lugnet site in the province of
Uppland.

with an offset beneath the rim date the site to Late Neolithic or Early
Bronze Age (Lekberg & Apel, ms). Finds of later dates were also found
(Table 7:1).
All but five pieces of the flint artefacts were unmodified, complete
or fragmented flakes. However, two fragmented arrowheads, two flakes
with retouches (interpreted as arrowhead preforms) and one gun flint
were also found at the site.

Piparberget, ra 240 (Lunda parish in the province of Uppland)


The excavation at Piparberget concerned an area cleared of stone in
relation to some stone cairns situated on a small ridge (Beronius
Jrpeland, ms). A small quartz outcrop was located within the investigation area, and a few features, in the form of pits and concentrations
of fire-cracked rock, were also found in an area below the outcrop. The
recovered artefacts consisted of quartz flakes, pottery and four flint
artefacts: a bifacial arrowhead, a retouched flake and two un-retouched
flakes. The quartz outcrop was situated c. 35 m above the present sealevel and the area where the find material was encountered was c. 30 m
above the present sea-level. The bifacial arrowhead dates the site to the
Late Neolithic or the Early Bronze Age.

Gr ndal, ra 192 and 241 (Husby- rlinghundra and Lunda


Parishes in the province of Uppland)
This site was situated between 27 and 33 m above the present sea-level.
It was located on the southern, eastern and northern slopes of a moraine
hill close by the Late Neolithic seashore. The excavated area was 24 000
m2. The site was dated by the C14 method and by artefact typology.
209

Jan Apel
According to the report, most of the remains belonged to two, separate,
chronological phases: one was the Late Neolithic and the other was
oldest phase of the Iron Age (Appelgren et al. 2000:4). During the
excavation, 232 features were revealed. Of these, 22 were hearths and
numerous post holes were also documented. Two areas with post holes
and cultural layers were tentatively interpreted as the remains either of
pit houses from the Iron Age or of depressions belonging to Late
Neolithic long houses (see Appelgren et al. 2000:20 f.).
Numerous artefacts of different kinds were recovered during the
investigation. This material consisted of, among other things, several
stone axes and axe fragments, grindstones for the processing of vegetables or crops, pottery, worked quartz and burned and unburned bone
(Appelgren et al. 2000:13). The bones of sheep or goats, pigs, horses
and cattle were recovered on the site (Appelgren et al. 2000:17). The
flint material from the site is presented in Table 7:2.

Annelund/Stenvreten ra 17 (Enk ping parish in the province of


Uppland)
This site was situated on a southerly slope c. 25-30 m above the present
sea-level. It was located in the terrain between the low mountains and
moraine hills in the north, east and west and a lower, flat, cultivated
area in the south. The ground was characterized by glacial clay with
patches of sand and moraine (Fagerlund & Hamilton 1995:128). The
area affected by the salvage excavation amounted only to a small part of
a larger area with previously known, ancient monuments dating from
the Late Neolithic to the Iron Age.
More than 800 features were recorded on the site. Of these, 101
were interpreted as post holes and the outlines of three houses were
visible in the western part of the excavated area. One house was also
recovered in the eastern part of the excavated area. The site was extensively dated by the analysis of 19 C14-samples. The results suggest that
the site was in use from the Late Neolithic (c. 2100 cal BC) to the end
of the Bronze Age (c. 500 cal BC) and that it was used sporadically
during the Iron Age and into historical times (Fagerlund & Hamilton
1995:116). One of the houses in the western part of the investigation
area was dated to the Late Neolithic period (Fagerlund & Hamilton
1995:116). Apart from the C14-dates, the site could also be archaeologically dated to the Late Neolithic period
210

Flint Technology in Eastern Central Sweden


Finds in the form of pottery, burnt daub, quartz, burnt bone, several fragments of ground objects in different kinds of rock, two axes,
two hammerstones and four grindstones in stone and flint artefacts
were recovered. The flint material consisted of 46 pieces, including
formally defined objects and production debitage. The former category
included a fragment of an undefined, bifacial object, three fragments of
bifacial sickles, a bifacial arrowhead, seven retouched flakes, a strike-alight and one fragment of a chisel or an axe. The debitage amounted to
32 flakes.

Conclusion
Eastern central Sweden has been severely affected by land uplift in the
wake of the latest ice age. As a result, the sea-level during the Late
Neolithic period was approximately 25-30 m above the present sealevel. Among the settlements presented above, there are both inland
and coastal sites. Annelund and Grndal (both located c. 27 meters
above present sea level) were situated on the coast, while, for instance,
Lugnet and the Fgelbacken sites, located between 40 and 50 m above
the present sea-level, were inland sites. Several of the sites are mixed, in
the sense that they contain artefacts or ecofacts from other periods. In
some cases, such as Fgelbacken, older materials are present, while in
most cases Iron Age materials predominate.

The flint material


In all, 458 flint artefacts are included in this study. With some exceptions, I have studied these materials personally. The exceptions include
flint artefacts from three of the Gallery-Graves, those from the mortuary
and settlement at Skillsta, and the two axes and the arrowhead from
Rssberga. In the cases of the Gallery-Graves and the Rssberga deposit, this does not pose a great problem, since these contexts generally
produced formal artefact categories, in which descriptions are accompanied by pictures or drawings. It is not likely that a personal inspection of these artefacts would have changed the interpretation in any
significant way. I have not been able to study the material from the
mortuary and the settlement at Skillsta but rely on information published in three reports and one article. However, this means that the
material from Skillsta is not included in Table 7:2 but is only mentioned in the text.
211

Jan Apel
The material was initially divided into two groups: formally defined artefacts and production debitage (Fig. 7:4). The first group consists
of the following artefact categories:
(a) daggers
(b) sickles
(c) bifacial arrowheads
(d) square-shaped axes/chisels
(e) unspecified bifacial objects
(f) scrapers
(g) strike-a-lights
(h) retouched flakes
The second group, the production debitage, consists of what appear
to be unmodified, complete or fragmented flakes and cores. Since this
group must be understood in detail if the aim of this paper is to be
fulfilled, the material was subjected to a detailed analysis.
Consequently, the production debitage was related to a couple of
important analytical concepts in the field of lithic analysis, namely those
of technique and method (see Chapter 5). These two concepts, together
with the natural constraints of the flint, determine the appearance of a
flake. Technique refers to the way in which the power is transmitted
from the knapper to the raw material. Method, on the other hand,
refers to the choices of techniques, impact angles and working positions,
as well as the intentions of the knapper during the production process.
The relative importance of these concepts, as regards the appearance of
the production debitage, has been extensively discussed in the literature
(Bradley 1978, Hayden & Hutchings 1989, Dibble & Pelcin 1995,
Pelcin 1997 and Bradbury & Carr 1999). In recent years, it has become
apparent that the choice of method plays a greater role in this process
than previously realized. It has been acknowledged that experienced flint
knappers can produce flakes with similar attributes but by different
techniques (Callahan, personal information). The obvious conclusion
of these findings is that the intentions of the flint knapper, which are a
vital element in the method concept, will determine the morphology
of the flakes (Pelcin 1997, Bradbury & Carr 1999). I shall not discuss
this problem any further in this paper. It will be sufficient to conclude
212

Flint Technology in Eastern Central Sweden

sc4

sc5
sc3

h
c

10 cm
Fig. 7:4. Drawings of different types of production debitage and formally
defined artefacts in flint recovered at Late Neolithic sites in eastern central
Sweden. The artefacts have been classified according to the system described
in the text. Drawings by Alicja Grenberger.

213

Jan Apel
that the production debitage from the Late Neolithic sites of eastern
central Sweden has been assigned to categories that are based on choices
of technique and/or method (see Chapter 5).
The production debitage was primarily divided into the following
categories, which are discussed in more detail in Chapter 5:
1. Complete or fragmented flakes with lips on the inside of the platform
remnant.
These flakes lack signs of crushing on the platform remnant and
pronounced bulbs on the inside.
2. Complete or fragmented flakes with pronounced bulbs on the inside
and no lipping at the edge of the platform remnant.
These flakes often bear signs of crushing on the platform and/or on
the distal end and sometimes display lateral fractures from the platform remnant downwards.
3. Small (less than 15 mm in long), complete or fragmented flakes
produced with lipped platform remnants.
These flakes are distinguished by their small size as well as by the
fact that the platform remnant is in mint condition.
The three categories coincide in some parts coincide with the terminology based on the technique that was formerly used to classify lithicproduction debitage, i.e. the flakes would have been regarded as detached with soft, hard or pressure technique respectively.

Flakes of the first category


The flakes in the first category were based on qualities that are encompassed by the term method concept, divided into two sub-categories:
on-the-edge or marginal flakes, and behind-the-edge or non-marginal
flakes (Bradley 1978, Callahan 1987). Flakes of the first sub-category
are distinguished by the fact that they have been detached on, or close
to, the edge of the platform remnant. If a flake of this kind is anticipated
by the knapper, the platform has to be prepared in order to prevent it
from collapsing during impact. Evidence from the accumulated
experience of modern flint-knapping indicates that it is easier to produce
successfully flakes of this kind with a hammer made of a softer material
than the flint, for instance, a deer or elk antler. These flakes are typical
214

Flint Technology in Eastern Central Sweden


of the bifacial method, since this method, if it is to produce satisfactory
results, requires a consistent, on-the-edge, flaking strategy. They will
often display small and thin platform remnants with traces of platform
preparation and have a relatively low angle between the platform remnant
and the dorsal side (Hgberg 1997). Flakes of the second sub-category
(sc2) are distinguished by the fact that the impact point has been placed
somewhat behind the edge of the platform remnant. Thus, these flakes
will have deep and wide, platform remnants. In certain cases, it seems as
if they have been detached with a soft and indirect technique with a
punch. This is typical of sub-category 2-flakes with facetted, divided,
platform remnants, a type that is common in the production debitage
from the third production stage of Neolithic squared axes from southern
Scandinavia (Hansen & Madsen 1983).

Flakes of the second category


The flakes in the second category were also divided into two subcategories: bipolar flakes and cores (sub-category 3) and complete and
fragmented platform flakes (sub-category 4). Flakes of the latter subcategory are distinguished from 1 and 2 by their appearance. They all
have some of the following traits: pronounced bulbs, bulb scars, ring
cracks around the point of impact, and radial fractures from the point
of impact downwards.

Flakes of the third category


The third group consists of small, nail-shaped, platform flakes with
small or no bulbs and with lipped platform remnants (sub-category 5).
These flakes have complete, undamaged, platform remnants and can
be distinguished from the other categories by their small size.

Remaining flakes
Pieces of flint that cannot be assigned to any of the categories proposed
above have been classified in sc6.

Conclusions
In Table 7:2, all formal objects and production debitage in flint from
the sites included in the investigation are presented according to the
described classification system. From these facts, it may be concluded
that complete, formally defined artefacts, with the exception of the
retouched flakes, were not deposited in settlements if they were not
broken. Accordingly, it seems as if formally defined artefacts were
215

Jan Apel
primarily deposited in graves (Table 7:2). However, there are some
minor exceptions to this rule. In the settlement at Annelund/Stenvreten,
two complete, bifacial arrowheads were found, and at the Grndal
settlement several, complete, flint artefacts were recovered. In the latter
case, a source-critical explanation can be put forward. Grndal is a large
site and it has been extensively excavated by hand in a way in which no
other settlement site included in the study has been excavated. This
procedure may have had the effect that scarce find categories not usually
found in settlements were recovered here (Appelgren et al. 2000). This
knowledge should influence the choice of excavation methods used on
Late Neolithic settlement in the future, at least if the artefact assemblage
and its composition are expected to offer important contributions to
the interpretation of the sites. From the general point of view, it must
be mentioned that almost all the settlement sites included in this study
are problematic in the source-critical perspective. In many cases, the
sites are of mixed ages, and structures, in the form of houses etc., are
unusual. In this context, I shall not discuss this problem further but
may return to this subject in the future.
Production debitage is not usually found in graves but predominates in the settlements. When production debitage is recovered in graves,
it is usually in the form of disc-shaped flakes that are excellent as blanks
for bifacial arrowheads. Two of the four gallery-graves included in this
study contained production debitage in flint. In the gallery-grave
at Sjtorpet, one flake was found and in the Annelund gallery-grave,
three (Fig. 7:5). Two of the flakes in the Annelund grave were produced
by a bifacial-reduction method and were thereby classified as sc1-flakes.
It is likely that they were deposited as arrowhead blanks. This interpretation is further enhanced by the fact that some of the retouched flakes
found in the Annelund grave should most probably be regarded as
arrowhead preforms. Curiously, a bipolar flint core was also found in
the Annelund grave. A retouched, disc-shaped flake was also recovered
in the gallery-grave at Dragby. This flake is a good example of a preform
for a bifacial arrowhead. The flake found in the gallery-grave at
Sjtorpet has not been studied by me personally and is therefore assigned
to sub-category 6. However, it would not surprise me if this flake was
in fact disc-shaped and stemmed from a bifacial-reduction sequence.

216

Flint Technology in Eastern Central Sweden

An interpretation of the flint technology


The analysis resulted in the following interpretations. Certain formal
artefacts, such as daggers, spoon-shaped scrapers and sickles, were
imported into eastern central Sweden, presumably from areas where
natural flint, and thus production sites, occur, such as the southern parts
of Scandinavia or possibly the Swedish west coast. These artefacts were
not locally produced and this is apparent from the absence of the large
large concentrations of production debitage that would have been found
on the sites if this had been the case. It is also, on theoretical grounds,
unlikely that complicated flint technologies like these could have been
maintained in areas without suitable raw material (Apel 2000a & b).
The artefacts were, however, consistently sharpened in eastern central
Sweden by persons who lacked some of the knowledge and know-how
required to sharpen such objects in the proper way, i.e. in a way that
kept the cutting edge within strict, edge-angle values. Small, nail-shaped,
platform flakes without a pronounced bulb (sub-category 5-flakes)
indicate this activity. Such flakes are created when an indirect, pressureflaking method is used on created or re-sharpened edges, and they are in
evidence in several of the settlements in the area of study (Table 7:2).
An indirect observation that also strengthens this interpretation is that
almost all stray finds of flint daggers in eastern central Sweden display
traces of extensive sharpening (compare Chapter 9). The edges of the
blades of such daggers are not as effective as edges re-sharpened by knowledgeable knappers, something that it is important to take into account
if it is believed that the daggers were ever put to practical use. Independently of whatever practical or ritual use the daggers might have had,
these facts indicate that it is likely that the artefacts were re-sharpened in
eastern central Sweden and most likely by people with restricted
experience of the true bifacial-knapping method. On account of these
facts, it seems as if the imported artefacts were seldom deposited in the
settlements unless they were damaged. The complete artefacts were
primarily deposited in hoards or in graves in different stages of their
working lives (see Chapter 9).
Apart from the sharpening of the edges of imported flint tools, the
most significant flint production present on the sites in eastern central
Sweden seems to have been the production of bifacial arrowheads in
flint. These arrowheads were made of thin and wide flakes shaped into

217

Jan Apel
Formal artefacts
Site

Alby (grave), ra 96, Botkyrka sn, Sdermanland 0 0 2,1 0 0


1 0 1 0 0
Annelund (grave), ra 17, Enkping, Up
1 0 0 0 0
Bjurhovda (grave), ra 459, Badelund sn, Vsml
1 0 2 0 0
Dragby (grave), ra 86c, Blinge sn, Up
0 0 1 2 0
Rssberga (offering), ra 403, Odensala sn, Up
0 0 1 0 0
Sjtorpet (grave), ra 288, Odensala sn, Up
0 3 1 1 1
Annelund (boplats), ra 17, Enkping, Up
Fgelbacken, ra 160, 162, 163, Hubbo sn, Vsml 1 0 1 0 0
1,2 0 1,3 4 1
Grndal, ra 192, 241, rlinghundra/Lunda sn,
Up
0 0 0 2 0
Karleby, ra 17, stertlje sn, S
0 0 2 0 0
Lugnet, N. Bngsbo, sterunda sn, Up
0 0 3 0 0
Maren, ra 138, Husby-rlinghundra sn, Up
0 0 0 1 0
Mrby 5:1, ra 473, S. Turinge sn, S
0 0 1 0 0
Piparberget, ra 240, Lunda sn, Up
0 0 0 0 0
Ribby, Vsterhaninge sn, S
0 1 0 0 1
brunna 1:1, ra 201, sterhaningen sn, S
0 0 1 0 0
stertlje 1:15, stertlje sn, S

0
0
0
0
0
0

0
0
0
0
0
0

2
11
2
1
0
0

0
2
0
0
0
0

0
0
0
0
0
0

1
1
0
0
0
0

0
0
0
0
0
0

0
0
0
0
0
0

12
0
0
0
0
1

18
16
3
4
3
2

0
1
1

1
0
0

7
2
17

3
1
10

1
7
0

6
2
8

10
17
38

5
11
11

7
13
25

46
56
122

1
0
1
0
0
1
0
1

1
1
1
2
0
0
2
0

0
2
0
1
1
1
3
7

0
1
1
2
0
0
4
0

0
2
0
0
0
0
5
3

1
3
0
0
0
0
3
0

2
5
0
0
1
0
12
5

2
13
0
0
0
0
1
3

3
20
26
0
1
1
5
25

12
49
32
6
4
3
37
45

57

24

18

25

90

46

139

458

Production debitage

d e

tot.

21 10 3

sc1 sc2 sc3 sc4 sc5 sc6 tot.

Table 7:2. The flint material grouped by site and by the classification system used in the text.
Formal artefacts, except retouched flakes (cat. h) that are complete, are marked by bold type.
The first five sites consist of the four gallery graves and a ritual deposit, while the last ten
sites have been interpreted as settlements. Further information on the sites can be found in
the following literature: Andersson 2001 (Alby), Andersson & Hjrthner-Holdar 1989,
Andersson, Fagerlund & Hamilton 1990, Fagerlund & Hamilton 1995 (Annelund/Stenvreten),
Borenius-Jrpeland MS (Piparberget), Eklf 1999 (stertlje 1:15), Appelgren et al. 2000
(Grndal), Gejvall 1963 (Dragby), Granath Zilln 1999 (Rssberga), Hallesj & Hamilton
1996 (Maren), Jaanusson 1969 (Bjurhovda), Jaanusson & Silvn 1962 (Dragby), Lekberg (ed.)
1996 (Fgelbacken), Lekberg & Apel MS (Lugnet), Olausson 1995 (Sjtorpet), Strucke ms
(Mrby), Strucke 1998 (Ribby).

the proper form by a simple kind of pressure technique (Fig. 7:6). It is


important to note that the production did not involve a true, bifacialcore industry, such as those used in similar and more or less contemporary traditions among hunter-gatherers in present-day Dalarna and the
northern parts of Scandinavia and, in agrarian communities in the
southern parts of Scandinavia. Owing to their complexity, bifacial-core
industries that required direct percussion evolved almost exclusively in
areas where suitable raw materials were abundant. In northern Scandinavia, local materials, such as quartzite and porphyrite, were used and
in southern Scandinavia flint. There was simply not enough flint in
eastern central Sweden to maintain the know-how of a bifacial-core
industry. However, it is intriguing to consider that there were other raw

218

Flint Technology in Eastern Central Sweden

10 cm
Fig. 7:5. Retouched flake (designated scraper in the report) from the
Annelund gallery grave. This flake is, in this context, regarded as a preform for
a bifacial arrowhead. Drawing by Lotta Sundberg.

materials available in this area that would have been suitable for the
production of bifacial-core artefacts. For instance hlleflinta is likely to
work fairly well, but for some reason the population of eastern central
Sweden preferred to make their arrowheads of flint from southern Scandinavia and thereby also relied on a simpler production method. In this
way, values and ideas from the southern parts of Scandinavia were
enhanced and expressed in contrast to those of southern Norrland and
Dalarna. This is a pattern that is also reflected in the choice of gallerygraves in which to bury the dead and in the architecture of the settlements.
Local raw materials were used to produce bifacial arrowheads in northern
Sweden and Dalarna (personal observations at the National Museum in
Stockholm, the Gvleborgs County Museum in Gvle and the
Jmtlands County Museum in stersund). A glance through
Lannerbros catalogue of archaeological materials from Dalarna reveals
only two arrowheads in southern Scandinavian flint, while there are
numbers of arrowheads made of local, plastic, rock types (Lannerbro
1991 & 1992). On the sites discussed in this paper, only one bifacial
arrowhead not made in flint was recovered: an arrowhead of blybergs
219

Jan Apel

Fig. 7:6. A fundamental drawing showing the outline of a bifacial-thinning flake


(1) in relation to the outlines of a preform for a bifacial arrowhead (2) and a
complete, bifacial arrowhead with concave base (3). Drawing by Lotta
Sundberg.

porphyry found on Fgelbacken (Lekberg 1996). This arrowhead was


carefully made, using a bifacial-core strategy. This raw material has its
origin in lvdalen in southern Norrland but can also be found in eskers
in central Sweden. I maintain, however, that it is likely that this arrowhead was not locally produced but reached Fgelbacken as an exchange
gift from the north.
The differences in the choices of raw material and the methods of
production between eastern central Sweden and Dalarna and Norrland
are further reinforced by the economic differences between these areas.
There is some evidence of cultivation along the northern Swedish coast
during the Late Neolithic and the Bronze Age (Knigsson 1989:140,
Welinder 1998:104). However, most of the bifacial artefacts in Norrland
were made by a mobile hunter-gatherer groups (Forsberg 1985, 1988
& 1992, Holm 1985 & 1991, Willemark 1992, Bolin 1999). In
contrast, the economy of eastern central Sweden also encompassed
animal husbandry and cultivation (Welinder 1998:104 f., 188 f.). These
cultural differences may have originated as far back as the Mesolithic
(Hallgren 1998). They are also evident during the Early Neolithic, when
the population of eastern central Sweden suddenly, and in sharp contrast
to the hunter-gatherer groups of Dalarna and Norrland, appropriated
the material and economic expressions of the Neolithic way of life. It
220

Flint Technology in Eastern Central Sweden


has been proposed that this pattern was created when the peoples of
two language groups met and, as a result, there was a strong need to
signal group identity (Hallgren 1998). In such a scenario, it is interesting to note that the few, formal, Late Neolithic, flint artefacts of southern
Scandinavian origin in Norrland reached this region by way of a western
exchange route. This route involved northern Jutland, the Norwegian
west coast and Trndelag (Becker 1952, Apel 2000a & b). The presence
of this exchange network may suggest that Norrland had closer contacts with the Norwegian west coast and Jutland than with eastern central
Sweden. This observation will in fact be strengthened later on in this
thesis (compare Chapter 9).
The bifacial arrowheads of eastern central Sweden are found in settlements and in graves. The majority of the settlement finds are fragmented. Usually, the tip has been broken off by a bending/compression fracture, suggesting that the damage occurred during use. Arrows
with undamaged shafts and broken heads were brought back to the
settlements, perhaps in order to have the arrowheads replaced. In some
cases, parts of animals in which arrows were embedded may have been
taken to the settlements and butchered, and the arrowheads fell off
during the process. This interpretation is further enhanced on considering
the distribution of flint artefacts at the Lugnet settlement in Uppland.
Here, an area with production debitage was spatially separated from an
area where two, fragmented, bifacial arrowheads were found (Fig. 7:7).
Whichever explanation is preferred, the presence of fragmented
arrowheads is not direct evidence for the production of such artefacts in
the settlements themselves. The flint debitage has to be studied in order
to reveal whether this was the case. In this context, it should be
mentioned that the bifacial arrowheads could also have been used in
war. This artefact type was an important part of the warriors outfit,
with a similar appearance in large areas of central, western and northern
Europe during the period in question.
There are several ways of obtaining suitable blanks for arrowheads
of the type produced in eastern central Sweden. The most suitable blanks
from the technical point of view were the bifacial thinning flakes
produced in using true, bifacial-core strategies (Fig. 7:4: sub-category
1). In the literature, this type of flake has been named handaxe-finishing
flake, clats de taille (Newcomer 1971), thinning flake (Crabtree
1972) and bifacial-thinning flake (Flenniken 1985). These flakes were
221

Jan Apel
probably hard to come by in eastern central Sweden. The bifacial-thinning
flakes stem from the middle and late stages in the production of large
bifaces, such as daggers and sickles, and they had to be imported into
eastern central Sweden. It is therefore likely that they were considered
to be valuable and even precious. Large, bifacial-thinning flakes with
initial pressure scars around the margins found in Late Neolithic gallerygraves, such as those at Annelund and Dragby, suggest that this was the
case. I have personally noticed, during visits in the Archaeological
Department of the Gothenburg City Museum (GAM) and the Lund
University Historical Museum (LUHM), that large, bifacial-thinning
flakes, often with traces of some preliminary shaping in the form of
pressure-flake scars, are common as grave goods in Gallery-Graves in
the western and southern parts of Sweden as well. However, they have
generally been classified as retouched flakes or scrapers. In this context,
it may be appropriate to mention a deposit from Fjlkestad in Skne,
containing 16 flint artefacts and 8 of these were preforms for heartshaped arrowheads (Malmer 1955:6 ff., Karsten 1994:95). On account
of contextual evidence, this deposition was interpreted as an offering.
This interpretation is further strengthened by the fact that at least two
preforms for bifacial arrowheads have been found in wetlands in Skne.
According to Karsten, the fact that arrowhead preforms have been
recovered in wetland contexts may indicate that arrowheads in general
that are found in wetlands can be regarded as offerings and not as arrowheads lost during hunting trips (Karsten 1994:95).
So one way of producing arrowheads in eastern central Sweden was
to import suitable blanks from the southern parts of Scandinavia. However, simpler flake types used for the same purpose were produced locally.
These flakes had to be relatively wide and thin in order to serve as
blanks. To achieve these formal criteria, the flakes had to be removed
from slightly convex, almost flat surfaces without pronounced ridges.
One way of making such flakes is to use the so-called combewa-method
in which flakes are removed from the ventral surfaces of larger flakes.
Combewa flakes tend to be wide and thin just because they are detached
from slightly convex and flat surfaces. In Sweden, this strategy was first
observed on the Middle Neolithic site at Kusmark in the province of
Vsterbotten (K. Knutsson 1986). At Kusmark, the method was used
to produce blanks for transverse arrowheads in flint. Since this is a simple
but effective method, it should theoretically have been suitable for areas
222

Flint Technology in Eastern Central Sweden

Fig. 7:7. The distribution of flint artefacts at the Lugnet site in Uppland
(from Lekberg & Apel, ms).

with no large amounts of raw material available but where blanks of


this type were used. Subsequently, it was anticipated that the method,
as it was now described, would be recognized in archaeological materials from other parts of Scandinavia as well (K. Knutsson 1988:194).
Since the idea was published, the combewa-method has indeed been
discovered on Early Neolithic Funnel-Beaker sites in eastern central
Sweden, where it was used to produce blanks for the production of
transverse arrowheads (Apel et al. 1997:21 f.). And it is not surprising
that at least one of the flakes found at Lugnet was produced by this
method. However, it should be emphasized that the combewa-method
was only one among several relatively simple methods within a pragmatic
repertoire on the Late Neolithic settlements in eastern central Sweden.
It is therefore not surprising that the bipolar technique was also present
on the sites, together with a simple, hard percussion-technique. A
common feature in the ways in which blanks were produced in eastern
central Sweden is that a careless and hard technique was used and in
turn that the debitage was characterized by radial fractures and secondary,
223

Jan Apel
high-speed fractures. It is obvious that the local methods used to produce
blanks for arrowheads also produced large amounts of debitage.
The shaping of the blank into an arrowhead was performed in the
simplest way. A pressure flaker, presumably with a handle of wood and
a tip of antler, bone, copper or flint, was used to press small, nail-shaped
flakes (sub-category 5-flakes) from the edges on both sides of the blank.
If a bifacial-thinning flake was used as a blank, the tip of the arrowhead
often coincided with the platform remnant of the flake, in order to
take advantage of the natural shape of the flake. The majority of the
bifacial arrowheads from eastern central Sweden exhibit the original
surface of the flake in their central parts on both sides (see the arrowheads in Fig. 7:2). The pressure flaking was never used to thin the arrowhead down by detaching flakes from two sides that overlap at the
centre a procedure that is common in the production of similar
arrowheads in Norrland. In eastern central Sweden, the pressure technique was used only to shape the arrowhead according to formal conventions. Experimental work revealed that all the stages in the production of bifacial arrowheads belong to the lowest grade of difficulty
involved in the production of daggers (Callahan, personal information).
The most difficult stage of the production was to get hold of or to
make blanks.

The value of disc-shaped flakes


In the previous section, it was suggested that flakes used as blanks for
bifacial arrowheads in eastern central Sweden, in the form of disc-shaped
flakes, were both imported and locally made. Further, it was proposed
that the imported flakes, the bifacial-thinning flakes, were ideal as blanks
and that the home-made flakes were of a poorer quality. As a result, it is
now possible to discuss the possible symbolic value of these flakes, as it
is revealed in graves and hoards.
In 1980, Ebbesen published an article in which he suggested that
flakes of flint found in different Scandinavian contexts should be regarded as offerings. The starting-point of Ebbesens study was 35 flakes
found in a Late Neolithic gallery-grave at Sneverholt on Jutland. The
flakes had been deposited, together with a flint dagger of type VA and
a barrel-shaped clay vessel, as part of a secondary burial dated to LN II
(Ebbesen 1980:147 ff.). They were positioned in a way that suggested
that they had originally been lying in some kind of box or bag (Ebbesen
224

Flint Technology in Eastern Central Sweden


1980:147 ff.). Fifteen of the flakes had retouched edges. Ebbesen pointed
out that the Sneverholt grave was unique in Denmark because flakes
had been used as burial gifts. He proceeded to relate the find to a series
of hoards containing flint flakes and sometimes also flint daggers, sickles
or spoon-shaped scrapers found in northern Jutland, on the Swedish
west coast and in the southern and western parts of Norway (Ebbesen
1980:151 ff. & Fig. 7). During the excavation of a cult house dated to
LN II at Sunnan outside Malm in Skne, several bifacial-thinning
flakes in flint, not produced at the location, were recovered, together
with, among other artefacts, flint daggers and bifacial sickles (Steineke
et al., in prep.). Consequently, it seems that bifacial-thinning flakes
were treated with respect in the production area, as well as in other parts
of Scandinavia.
Similar circumstances may also account for some of the large flint
deposits in northern Sweden, for instance at Bjurselet. In connection
with the hoards of thick-butted, flint axes at Bjurselet, five flakes
stemming from the production of such axes were also recovered (K.
Knutsson 1988:52). One of these flakes was found in connection with
a star-shaped, axe deposit. The flake in question stems from the third
production stage of a square-shaped axe (Hansen & Madsen 1983:45
f.). The find association indicated that the flake was deposited together
with the axes and several other flakes (K. Knutsson 1988:77). Knutsson
suggested that axe-production flakes from Bjurselet originated from an
axe production in southern Scandinavia. A large part of the squared
axes in the northern-Swedish, flint deposits are of the Hornby or Kregme
types (Ebbesen 1975b, Bertvall 1985:13) and there are some examples
of broad- and thin-bladed axes with curved edges (Siemen 1981:172).
The Hornby and Kregme types are generally dated to a later part of the
Middle Neolithic period, while the thin-bladed axe belongs to an early
phase of the Late Neolithic period (LN I). Consequently, the flint
deposits of northern Sweden seem to belong to an early part of the Late
Neolithic period (K. Knutsson 1989:256, see also the extensive
discussion of axe typology on pages 252 ff.). Such a dating would in
fact also coincide with the C14-datings of the first cereal horizon in
pollen diagrams from the area (Knigsson 1989).
Flakes stemming from the production of square-shaped axes have
also been found on some of the sites in eastern central Sweden discussed in this paper. A scraper made of a stage-three, axe-production
225

Jan Apel
flake was recovered at stertlje and several such flakes were found at
Skillsta (Ekmyr-Westman 1991:18). It has been suggested that the
production of square-shaped axes in flint diminished at the transition
between LN I and II (1950 cal BC), when a local production of flanged
bronze axes was introduced into southern Scandinavia (Vandkilde
1996:270 ff.). Thus, it is possible that sites with axe flakes in eastern
central Sweden should be dated to LN I (2350-1950 cal BC).
In contrast to the situation in Denmark, flint flakes were fairly
common as grave gifts in Sweden during the Late Neolithic period
(Spetz 1994:45 ff.). The frequency of flint flakes with retouched edges
in Late Neolithic graves, in relation to all flint artefacts deposited in
graves, tends to increase along the south-north axis. A study of unretouched flakes in Late Neolithic graves reveals two distinct areas. In
the western area, consisting of Halland, Smland, Vstergtland,
Bohusln and Dalsland, 59% of the flint objects are un-retouched flakes;
in the eastern area, Nrke, stergtland, land, Gotland and Blekinge,
only 34% of the flint objects are un-retouched flakes (Spetz 1994:49).
It is likely that the larger amount of un-retouched flakes in the graves of
western Sweden is related to the presence of the hoarded flakes in the
same area.
According to Ebbesen, the flakes that were deposited as offerings
stemmed from the reduction of a core that might have been reduced on
the site. However, I do not agree with this conclusion. It is my impression that the Norwegian hoards, for instance, the one at Hauske on
Karmya in Rogaland (inv. no. S. 3 513, Archaeological Museum in
Stavanger), consist of flakes selected on the grounds of their formal
qualities and not all the flakes from one or several reduction sequences.
The flakes may in some cases come from a single reduction sequence,
but I maintain that they do not originate from platform cores. In my
opinion, the flakes are debitage from the production of large and
regularly shaped, core artefacts, such as daggers and sickles and in some
cases square-shaped axes. Recently, similar flakes were reported to have
been found in the Oslo area in Norway as well (Berg 1998).
I have previously suggested that the production of flint daggers in
southern Scandinavia was organized in regional apprenticeship systems
(see Chapter 2 and Apel 2000a & b), i.e. kinship-based systems, in
which the craft was passed on from one generation to the next. This
reasoning implies that the division of labour was institutionalised during
226

Flint Technology in Eastern Central Sweden


the Late Neolithic. In the light of this reasoning, it is not surprising
that different regional styles should occur in the areas where flint daggers
were produced. Different workshops developed different dagger styles
and today these local styles are recognized as types, i.e. types that
correspond to normative, technological traditions. It may even be that
in these regional traditions certain parts of the recipes for action were
deliberately kept secret, presumably in order to maintain or improve
the social position of the members (Chapter 2 and Apel 2000a & b). I
would like to return to the discussion of the disc-shaped flakes above in
order to mention some thoughts. These flakes were valuable, not only
because they were excellent as blanks for bifacial arrowheads, but also
because they were an integral part of the production of prestige items.
Objects such as daggers, sickles and axes could be seen as insignia for
certain segments of the population (see the further discussion in Chapter
10) and it is hereby suggested that the flakes were so too. The fact that
these flakes were excellent as blanks for bifacial arrowheads contributed
to their symbolic value. Bifacial arrowheads and disc-shaped, flint flakes
must be regarded as having been as much symbols for important
connections with foreign areas as they were functional objects.
The tradition of depositing blanks and preforms for bifacial arrowheads in graves can be traced back to the Bell-Beaker tradition in western Europe. A newly excavated example is the Bell-Beaker burial at
Wellington Quarry, Marden in Herefordshire, England (Harrison et al.
1999). The burial consisted of an unmarked, flat grave that may have
been timber-lined (Harrison et al. 1999:1). The grave goods consisted
of a fragmented, tanged, copper dagger, a polished stone wristguard,
four barbed and tanged, bifacial arrowheads, three arrowhead blanks,
three flint knives, two triangular points or small daggers and four flint
flakes. The three arrowhead preforms are thin, disc-shaped flakes with
an initiated retouch around the edges, just like the similar flakes found
in the gallery-graves of Annelund and Dragby. It is possible that the
four un-retouched flakes were also meant to be arrowhead blanks, but
unfortunately they are not sufficiently described in the article. It is not
entirely unlikely that the flint objects that are classified as knives may in
fact be arrowhead preforms too. The pottery in the grave in Wellington
Quarry can on typological grounds be dated to the interval 2750-2500
cal BC (Harrison et al. 1999:11). This means that it is several hundred
years older than the earliest Scandinavian sites with evidence of the pro227

Jan Apel
duction of bifacial arrowheads and daggers for that matter, for instance,
the Myrhj site in Jutland, C14-dated to 2350 cal BC (Jensen 1973).
Bifacial arrowheads in flint or in other plastic rock types have also
been recovered in graves all over Europe from the Middle Neolithic
period up to the Bronze Age, for instance at Mycanae, where 29 and 17
bifacial arrowheads of flint and obsidian were found in two of the grave
shafts (Mylonas 1973: Plate 72, 123). Since these graves are dated to
the interval 1700-1500 BC, they are contemporary with the first period
of the Scandinavian Bronze Age.
Another parallel to the treatment of disc-shaped, flint flakes in eastern
central Sweden during the Neolithic and Early Bronze Age has been
presented by Helena Knutsson (see Chapter 3 and H. Knutsson 1996,
1999 & 2000). She has been studying a certain kind of flint blades that
are recovered from Battle-Axe graves in Scandinavia. She relates these
findings to Haydens theory of prestige technologies (Hayden 1995:68;
Olausson 1998:90). In this context, a prestige technology does not, or
not only, have a practical function, but is also used in power games by
aggrandizers in order to achieve and maintain their social positions
through their connections with the blades and their makers (H. Knutsson
2000). This type of technology can be traced over large regions encompassing areas in northern, central and southern Europe, as well as the
western parts of Asia (Knutsson 2000). It is likely that the production
and distribution of disc-shaped flakes were organized in a smaller but
similar fashion.

Conclusions
Large flint artefacts, such as flint daggers, sickles and axes, were imported from areas with natural flint resources, but they were re-sharpened in eastern central Sweden. The flint-knapping on the sites studied
in this paper was directed to the sharpening of large imported, edged
tools but also to the production of bifacial arrowheads. Thin, discshaped flakes were used as blanks and they were either imported or
made locally. The ideal blanks for the production of bifacial arrowheads were bifacial-thinning flakes and such flakes were subsequently
imported into eastern central Sweden, together with large artefacts. The
imported and locally produced blanks were then shaped into arrowheads by a simple, pressure technique. Disc-shaped flakes of a poorer
quality were produced locally by simpler methods.
228

Flint Technology in Eastern Central Sweden


It is apparent that the imported disc-shaped flakes also had a great
symbolic value. This value is reflected in the fact that they are found in
Gallery-Graves in the southern and central parts of Sweden and in hoards
and votive offerings in Norway and southern Sweden and on the Swedish
west coast. It is also of interest to note that at least one Danish GalleryGrave (Sneverholt) also produced finds of this kind. This may indicate
that even parts of the population of southern Scandinavia that were not
involved in the production of large bifaces thereby came to treasure and
value these flakes. In this context, it is intriguing to consider that it is
the flakes from the difficult, and hence public, stages of dagger production
(see Chapters 2 and 4) that have a great symbolic value, especially in
areas quite far away from where the daggers were produced. In other
words, it is apparent that these flakes, as well as other southern
Scandinavian artefacts in flint, such as daggers, carried a strong and
normative past in the present message to the Late Neolithic population
of the agrarian communities of Scandinavia.
In contrast to the northern parts of Scandinavia, where bifacial arrowheads were produced in local raw materials, the population of eastern
central Sweden chose to make their arrowheads of southern Scandinavian
flint. I suggest that this was a reflection of a deliberate attachment to
southern Scandinavian values and an equally deliberate repudiation of
the values of the northern hunter-gatherers. This distinction is also
apparent in the choices of different economies in the two areas.
The flint working that was practised on the Late Neolithic settlements of eastern central Sweden was dictated by strict conventions
regarding the form and the choice of southern-Scandinavian flint as the
raw material. This resulted from the fact that the area could not maintain
the level of craftsmanship present in areas immediately to the north.
From the purely practical point of view, this makes no sense, and, consequently, we can expect that this pattern will be explained by the fact
that artefacts of southern-Scandinavian types and raw materials had a
great symbolic value for the population of eastern central Sweden.

229

Chapter 8

The Classification System


of the Scandinavian Flint Daggers

My interest in the social reproduction of the Late Neolithic has, so far,


led me to investigate the flint industries in a production as well as a
consumption area. It has been suggested that the dagger technology was
reproduced within corporate lineage-based groups. The degree of knowhow is reflected in the quality of the finished daggers and the exclusiveness of the knowledge the recipes for action may be reflected in the
formal properties of the flint daggers. Thus, it is now time to investigate the products in detail.
This chapter forms the foundation for the various investigations
presented in the following chapter, in which the Scandinavian and the
northern European dagger materials are presented. Thus, the purpose
of this chapter is to describe the principles that have guided the classification of the Swedish flint-dagger material in this book. Definitions of
the traditional types and subtypes of Scandinavian flint daggers are presented in the first section of the chapter. This section is followed by
comments on some general, chronological trends in the material and a
more detailed, research history, aiming at an understanding of the
chronological and geographical implications of the typology.

The flint-dagger typology


One of the classical archaeological topics of debate is the question
whether types are natural and exist a priori or whether they are merely
invented by the archaeologists (see, for instance, Jevons 1874, Brew
1946, Ford 1954, Malmer 1962 & 1963, Grslund 1974 & 1996,
Spaulding 1982). In this context it is stressed that all practical, archaeological typologies are to a certain extent natural, in the sense that they
reflect prehistoric circumstances, and to a certain extent artificial, in
that they are created today in order to answer to specific needs (Grslund
231

Jan Apel
1996:17). This means that they are partly discovered and partly invented
by the archaeologist (Willey & Phillips 1958:13). The following
quotation clarifies this view coherently: We can say as a general rule
that nature (or culture) creates modalities, but we, the typologists, choose
among them those that are informative for any particular purpose of
our own (Adams & Adams 1991:68). According to William and Ernest
Adams, it is the combinations of natural and artificial boundaries that
have made it difficult to end this debate (Adams & Adams 1991:68). It
seems as if we shall have to accept that each nomothetic typology in
itself is a combination of natural (or given) and artificial aspects (Grslund
1996:16 ff.). In this respect, types can be regarded as partly essential
and partly instrumental, since they are determined partly by what they
are and partly by what they do (Sokal 1977:187 ff.). This conclusion,
however, does not justify the fact that types sometimes are poorly defined in the archaeological literature (Malmer 1963).
As regards the Scandinavian flint-dagger types, it is apparent that
they were discovered rather than constructed. The boundaries between
the different main types were so sharp and intuitively obvious that some
of them may very well be a priori evidence of a prehistoric reality. It is
also worth considering the fact that the dagger as an object is still an
institutional fact today. We shall therefore recognise a dagger when we
see one if it fulfils the dagger norm. Ironically, since traditional
archaeology was concerned mainly with the chronological aspects, the
obvious patterning that the different dagger types displayed was
interpreted almost exclusively in chronological terms rather than as the
result of, for instance, different geographical distributions or as reflecting
different social strata. Ebbe Lomborgs thorough typological treatment
of the Danish dagger material was, in many respects, instrumental and
to some parts artificial in its attempt to formalise the dagger typology
to meet the typological standards of the 1970s. As we shall see in the
definitions of the different subtypes, Lomborg used the technical aspects,
the measurements and the design to formalise the typology. Through
classifying Danish and Norwegian flint daggers, I have been able to
compare my classifications with Lomborgs (1973) and Scheens (1979)
classifications and I did not encounter any important discrepancies in
our classifications. From this it may be concluded that the type
descriptions presented by Lomborg are sufficient.
The fact that the main, Scandinavian, flint-dagger types, with minor
232

The Classification system of the Scandinavian Flint Daggers


alterations, were defined in the latter part of the 19th century, and in
detail during the course of the 20th century, means that different regional
studies stand a chance of being reasonably calibrated (see Chapter 9).
For this reason alone, it is worth using the existing typology without
proposing any major changes to it. However, in the case of the
Scandinavian flint daggers, this is not the sole, or the most important,
reason for accepting the typology as it was presented most recently by
Ebbe Lomborg (1973). As we shall see, the major types, and some of
the subtypes, have been defined in a way that makes it possible to interpret
them as coinciding with prehistoric distinctions, i.e. distinctions based
on regional, normative, technological traditions. The fact that some
major trends reflected in important elements of the typology can be
traced to certain periods and regions will in fact make it possible to
simplify Lomborgs definitions in order to suit the questions that will
be asked in this book. However, in the database that forms the basis of
the investigation carried on in the following chapters, all the daggers
have been defined according to Lomborgs typology. The daggers that
form the empirical basis of this book have, for this reason, been classified
according to the typology presented in Lomborg (1973). As we shall
see, this does not mean that I have adhered to all of Lomborgs chronological conclusions.
The following description of the flint-dagger types is basically a repetition of the type definitions presented by Lomborg (1973:32 ff.). I
have tried to make the definitions as short and distinctive as possible.
Some of the type definitions include the way in which the blade has
been prepared. For this purpose, Lomborg used two categories that
were defined by Sophus Mller (1902), namely blades with glatte
Retusche and blades with parallel Retusche. A blade of the former category
displays a surface with heterogeneously shaped, flake scars while a blade
of the latter category display homogeneously shaped oblong flake scars
positioned side by side over the blade (Fig. 8:1). In this context, a technological definition is used to pinpoint these different, blade-finishing
strategies; consequently, blades with heterogeneous, flake-scar patterns
are referred to as percussion-flaked blades, while homogenous, flakescar patterns are referred to as parallel, pressure-flaked blades. Before
presenting the type definition, it remains to be said that Scandinavian
flint daggers are distinguished from other knives and daggers in flint in
being oblong, bifacially knapped artefacts with pronounced handles and
233

Jan Apel

Fig. 8:1. Lomborgs six main dagger types (assembled from various pictures in
Lomborg 1973).
234

The Classification system of the Scandinavian Flint Daggers

Fig. 8:2. Type-I subtypes (from Lomborg 1973:32, Fig. 9).

two cutting edges. Considering the first type, however, this definition
has to be complemented, since the majority of the type-I daggers do
not display pronounced handles (see below).

Type I
Lancet-shaped daggers without handles, or with marginally marked
handles that display a lenticular cross-section (Fig. 8:1). There is no
difference in thickness between blade and handle. The type is distinguished from other oblong and thin, bifacial objects, such as lance- and
spearheads, by the fact that the edges of the handle taper towards the
base and that the base tends to be convex or sometimes even pointed.
The edges of spearheads, for instance, do not taper towards the base;
instead they tend to have straight or concave bases (see Mller 1902,
Lomborg 1973).

Type I A
Percussion-flaked, type-I daggers with their greatest width nearer the
tip than the middle of the dagger (Fig. 8:2). The blades of these daggers
tend to appear as very wide, an impression that is enhanced by the fact
that the widest part of the dagger is found near the tip. In Denmark,
this subtype is most common around the Limfjord in northern Jutland
(Lomborg 1973:36, Fig. 12).
235

Jan Apel

Type I B
Percussion-flaked, type-I daggers with their greatest width at the middle or below the middle of the dagger, but always nearer the middle
than the tip (Fig. 8:2). In Denmark, this subtype, like the previous
type, is most common in the Limfjord area in northern Jutland
(Lomborg 1973:36, Fig. 13). In cases in which the blade has been extensively re-sharpened, in a way that has distorted its original outline,
the two previous subtypes cannot be distinguished from each other.
Daggers of this kind are classified as type I A/B.

Type I C
Parallel-flaked, type-I daggers with symmetrically curved and concaveshaped, handle edges (Fig. 8:2). In Denmark, this subtype is almost
exclusively concentrated to the Limfjord area in northern Jutland
(Lomborg 1973:37, Fig. 14). To a greater extent than the other type-Idagger subtypes, it is common in LN I graves in this area (Rasmussen
1990:34 ff.), and the type has been connected with the west European
Beaker influences that affect this area during the first half of LN I
(Vandkilde 1996:281).

Type I D
Type-I daggers with concave-shaped, handle sides; a trait that enhances
the impression of these daggers as having pronounced handles (Fig. 8:2).
This type is reminiscent of type I B but has a more pronounced handle
than the other type-I daggers. On certain examples, the handles tend to
be somewhat thicker than the blades and, according to Lomborg, the
transition from type I D to type II A is indeed continuous (Lomborg
1973:41, 79). In Denmark, the hoards, as well as single finds containing
this type, are concentrated to eastern Danish islands (Lomborg 1973:37,
Fig. 15).

Type I E
A type-I dagger with systematically flaked notches on the edges at the
transition between blade and handle. This is a problematic type, since
the notches were most likely created in order to keep the shaft or handle in place. Subsequently, the retouched notches may not have had
anything to do with the production of the dagger but may have been
the solution of a practical problem that occurred at a later stage when
the handle was being adjusted to the dagger base. This is also indicated
236

The Classification system of the Scandinavian Flint Daggers


by the fact that daggers of several subtypes display these notches (see
Lomborg 1973:41, Fig. 17).

Type I x
Extensively re-sharpened, type-I daggers that, on this account, cannot
be assigned to any of the subtypes defined above. Curiously, Lomborg
includes dagger preforms in this subtype as well, provided that they can
be distinguished as type-I preforms (Lomborg 1973:44). In my classification of the Swedish daggers that will be presented in the next chapter, I have not included preforms in this category. Preforms are instead
discussed as a separate category.

Type II
Daggers with lancet-shaped blades and pronounced handles that are at
least 2 millimetres thicker than the blades. The handles display oval or
lenticular cross-sections (Fig. 8:1). This type is divided into two subtypes,
defined according to the width/thickness relationship of the handle. In
fact, as we shall see, it is only this relationship that so far makes it
possible to differentiate between daggers of types I D, II A and B, and
III A. I have usually defined these types impressionistically by sight,
which means that there may be some in-consistencies involved in the
classification. However, as we shall see below, in many of the analyses
that will be made in Chapter 9, these types, together with the other
type-III daggers, are regarded as a homogenous group and will in most
cases be merged. This is due to the fact that I regard the boundaries
between these types as a construction by the archaeologist and not relevant to any prehistoric norm, just as we have seen that Lomborg himself did (Lomborg 1973: 44 ff.). These daggers represent one and the
same regional and temporal tradition that is rooted on the eastern Danish
islands and in the south-western parts of Skne in Sweden. Arguments
in support of this opinion will be presented later on in this chapter.

Type II A
Type-II daggers with handles that display a width/thickness relationship of between 2:1 and 3:2 (Fig. 8.3).

Type II B
Type-II daggers with handles that display a width/thickness relationship of between 3:2 and 1:1 (Fig. 8:3).

237

Jan Apel

Fig. 8:3. Type II A and B (after Lomborg 1973:44, Fig. 21).

The majority of the type-II daggers display percussion-finished blades,


however, there are a few parallel, pressure-flaked daggers of this type as
well. In Denmark, such examples are found mainly in Jutland. However,
the general geographical distribution of the type-II daggers in Denmark
is concentrated to the eastern Danish islands, where 11 out of 14 hoards
also were recovered (Lomborg 1973:44 ff., Figs. 22 & 23).

Type III
Daggers with small handles with parallel sides, rhombic, roundedrhombic or rectangular cross-sections and evenly shaped outlines (Fig.
8:1). This type, like the earlier types, has lancet-shaped blades. With the
notable exception of subtype A, type-III daggers regularly exhibit pressure-flaked seams on the edges and sides of the handles. The majority
238

The Classification system of the Scandinavian Flint Daggers


of the type-III daggers have flat or straight, handle bases: a typological
element considered by Lomborg as being introduced when the production of type-III daggers began (Lomborg 1973:47). Rasmussen has
pointed out that the type-III daggers are more evenly distributed than
the earlier types. This may be the case if only daggers found singly are
considered; however, if we choose to look at the distribution of hoards,
it is apparent that this is not the case. Of the 15 hoards containing typeIII daggers in Denmark, 13 are located on the eastern Danish islands
(Lomborg 1973:50 f., Figs. 25-27).

Type III A
Type-III daggers whose handles have a width/thickness relationship of
1:1 and square cross-sections (Fig. 8:4). This subtype does not have
pressure-flaked seams on either side of the handle. Occasionally, though,
a pressure-flaked seam is visible on the handle edges.

Type III B
Type-III daggers with handles that have a width/thickness relationship
of 1:1 and square cross-sections (Fig. 8:4). This subtype has pressureflaked seams on both sides, as well as on the edges of the handles.

Type III C
Type-III daggers with handles that have a width/thickness relationship
of 1:1 and square cross-sections (Fig. 8:4). This subtype has pressureflaked seams on one side, as well as on the edges of the handle.

Type III D
Type-III daggers whose handles have a width/thickness relationship of
1:1 and square cross-sections (Fig. 8:4). However, in comparison with
the other type-III daggers, the handle is turned through 45 degrees, so
that two sides of the squared handle are parallel to the blade. This subtype
has four pressure-flaked seams. Only three examples of this subtype are
known in Scandinavia (Montelius 1917:421, Mller 1902:129,
Lomborg 1973:49).

Type III E
Type-III daggers with pressure-flaked seams on both sides, as well as on
the edges of the handles (Fig. 8:4). The width of the handles is always
greater than the thickness and, consequently, they display rectangular
cross-sections.
239

Jan Apel

Fig. 8:4. Type-III subtypes (from Lomborg 1973:48, Fig. 24).

240

The Classification system of the Scandinavian Flint Daggers

Type III F
Type III daggers with pressure-flaked seams on one side and on both
edges of the handle. The width of the handles is always greater than
their thickness and they have rectangular cross-sections.

Type IV
Daggers with fishtail-shaped handles with rhombic, rounded, triangular or rectangular, cross-sections (Fig. 8:1). All the subtypes, except F,
display pressure-flaked seams in the middle of the handles on at least
one side (Fig. 8:1). Subtypes A and B have blades that are almost lancet-shaped but are often slightly wider nearer the blade-handle transition than in the middle of the blade. Subtypes D-F display blades with
a rounded, triangular outline with their greatest width nearer the bladehandle transition than the blade (Fig. 8:5). In Denmark, the geographical distribution of the type-IV daggers varies between the subtypes.
However, the four known hoards containing this dagger type were found
on the Danish islands, three of them on Fyn (Lomborg 1973:55, Figs.
30-34). This may indicate that the type has predominantly, an eastern
Danish and south-western Swedish origin. A microscopic investigation
of the pressure-flaked seams of nine type-IV daggers revealed traces of
what appeared to be metal on the stitched seam (Stafford 1998:342).
This may mean that copper-tipped pressure flakers were needed in order
to make these small and elaborate seams.
Most of the parallel-flaked blades on the type-IV daggers were
ground before the final pressure-flaking commenced. However, there
are examples of large, parallel-flaked, type-IV daggers that were not, for
example two daggers from Igellsa parish in the province of Smland
(Fig. 8:6) (Hofrn 1925:32 f., Gurstad-Nilsson 1999:191). These were
found together in a square-shaped, small, construction made of sandstone slabs that had been dug into the clay.

Type IV A
Type-IV daggers with handles that are thick and display rhombic crosssections and have straight edges that broaden evenly towards the base
(Fig. 8:5). The handles thicken towards the base and display pressureflaked seams on the back and front sides, as well as on the edges. The
type can be distinguished from type III E only in that the handles become wider towards the base. The blades are lancet-shaped, although
they tend to be widest near the blade-handle transition. In Denmark,
241

Jan Apel

Fig. 8:5. Type-IV subtypes (from Lomborg 1973:53, Fig. 29).

242

The Classification system of the Scandinavian Flint Daggers

Fig. 8:6. Two type-IV flint daggers (32 and 30 cm long) from Igellsa parish in
the province of Smland. These daggers were not grinded before parallel
flaked (photo: the Kalmar County Museum).

243

Jan Apel
their distribution is concentrated to the eastern Danish islands, mainly
to Sjlland, and the sole known Danish hoard containing daggers of
this type was also recovered here (Lomborg 1973:54, Fig. 30).

Type IV B
Type-IV daggers that have thick handles with rhombic cross-sections
and concave edges that broaden towards the base (Fig. 8:5). The handles thicken towards the base and display pressure-flaked seams on the
back and front sides, as well as on the edges. The blades are lancetshaped, although they tend to be widest near the blade-handle transition. In Denmark, this type is evenly distributed over the northern and
eastern parts of Jutland, as well as over the eastern Danish islands
(Lomborg 1973: 54, fig. 31).

Type IV C
Type-IV daggers with pressure-flaked seams on the back and front of
the handle (Fig. 8:5). The handles have flat, rhombic cross-sections and,
in contrast to the two preceding types, they are evenly thick. The blades
are relatively wider than the blades of the two preceding types. The
blades in type IV C (as well as in other, later, dagger types) usually
display a rounded, triangular outline, widest near the blade-handle transition. In Denmark, the daggers of this type found singly seem to be
evenly distributed over the northern and western parts of Jutland, as
well as over the islands. However, the two hoards containing this type
were recovered on Fyn and Sjlland respectively (Lomborg 1973:55,
Fig. 32).

Type IV D
Type-IV daggers with pressure-flaked seams on the front, but not on
the back, of the handles (Fig. 8:5). The back of the handle and its base
are slightly curved. The blade on these daggers is usually rounded, triangular, with the widest part near the blade-handle transition. In Denmark, this type is concentrated mainly to the southern parts of Jutland
and to the island of Fyn (Lomborg 1973:55, Fig. 33).

Type IV E
Type-IV daggers with pressure-flaked seams on the front, but not on
the back, of the handles (Fig. 8:5). The back of the handle is flat or
rounded and the back of the base is straight. The blades usually have a
triangular outline, with the edges rounded, and they are widest near the
244

The Classification system of the Scandinavian Flint Daggers

Fig. 8:7. Type V A and B (from Lomborg 1973:58, Fig. 35 ).

blade-handle transition. In Denmark, this type is common in the southwestern parts of Jutland, and, to some degree, on Fyn (Lomborg
1973:57, Fig. 34).

Type IV F
Type-IV daggers with handles that have rectangular cross-sections with
the broad sides positioned parallel to the blade (Fig. 8:5). Two pressureflaked seams are positioned on the edges of the handle. However, no
seam is present in the middle of the blade. This type is very unusual;
Lomborg lists only five examples from Denmark and only two of them
have information on provenience (Lomborg 1973:58).

Type V
Daggers with handles that widen at the base and display lenticular crosssections (Fig. 8:1). The handles do not have pressure-flaked seams on
either side but may display seams on the edges and on the base. In
Denmark, the type-V daggers are evenly distributed and, of the six
known hoards, three were recovered in the northern parts of Jutland
and three on the eastern Danish islands one on Fyn and two on Sjlland
(Lomborg 1973:59, Figs. 36 & 37). The type is divided into two
subtypes according to the form of the base of the handles.
245

Jan Apel

Fig. 8:8. Subtypes of type VI (from Lomborg 1973:40 & 41).

246

The Classification system of the Scandinavian Flint Daggers

Fig. 8:9. British flint dagger dated to the late


Bell Beaker or the early Beaker phase (left)
and a Scandinavian flint dagger of type I A
(right) . Drawings made by Lotta Sundberg
from original pictures published in Green et
al. (1982) and in Lomborg (1973).

Type V A
Type-V daggers with straight or curved bases similar to the bases of the
type-III and IV daggers (Fig. 8:7).

Type V B
Type-V daggers on which pointed ridges run across the base (Fig 8:7).

Type VI
Daggers with handles that have straight, parallel-sided edges (Fig. 8:7).
The handles display lenticular or oval cross-sections and are often unevenly shaped. However, the base is never the widest part of the handle.
These daggers, according to Lomborgs definition, always have
percussion-finished blades. In Denmark, the type-VI daggers found singly are evenly distributed. Of the five hoards containing daggers of typeVI A, two were recovered in Jutland, north of the Limfjord, one on
Fyn and two on Sjlland. The two hoards containing type VI B daggers
were both found in Jutland (Lomborg 1973).
247

Jan Apel

Type VI A
Type-VI daggers with handles that are thicker than the blades (Fig. 8:8).

Type VI B
Type-VI daggers with handles that are not thicker than the blades (Fig.
8:8).

Type VI C
Type-VI daggers with details that have obviously been copied from
daggers and swords in metal (Fig. 8:8).

Some chronological trends in the Scandinavian


flint-dagger material
As we shall see in the next section, Lomborgs detailed, chronological
interpretation of the dagger typology has rightly been questioned.
However, there are two major chronological trends reflected in the flintdagger typology that have been firmly established by analogies to dated
assemblages on the British Isles and the Continent:
1 The general outlines of the blades of type-I to type-IV A-C daggers,
on the one hand, and type-IV D to type-VI daggers, on the other,
are chronologically dependent.
2 The introduction of the fishtail handle on type-IV and type-V daggers is also dated.
I shall briefly outline these chronological reference points in the following paragraphs.
During the Scandinavian Middle Neolithic period, immediately before
the production of Scandinavian flint daggers began around 2350 cal
BC, there were several dagger types circulating in northern Europe.
Tanged, copper flat daggers were deposited in the graves in north-western
parts of Europe (Harrison 1982). A dagger of this kind has actually
been recovered at Nttraby in the Province of Blekinge in south-eastern
Sweden (Cullberg 1960:35, Abb. 1). Another tanged, flat dagger in
copper is known from Scandinavia. It was found in Aalborg County,
close by the Late Neolithic flint mines at Skovbakken. This dagger
probably orginates from the British Isles (Vandkilde 1996:180 f.).
Bifacially-shaped daggers, made of honey-coloured flint from GrandPressigny in the Loire valley were distributed over large areas of north248

The Classification system of the Scandinavian Flint Daggers


western Europe, including the north-western parts of Germany (Lanting
& van der Waals 1976:13 f.). The way in which these daggers were
made differs in important aspects from the Scandinavian flint-dagger
tradition described in Chapter 2. Large blades, struck from so-called
Livre-de-buerre cores, were bifacially worked into daggers. In SchleswigHolstein and Mecklenburg in the northern parts of Germany, variants
of these daggers were produced in local flints during the same period
(Struve 1955), and a few of these pseudo Grand Pressigny daggers
have been recovered in the south-eastern parts of Denmark as well (VangPetersen 1993:124).
According to Lomborg (1973:87 ff.), however, the prototypes of
the earliest Scandinavian daggers are to be found on the British Isles.
Here, bifacial daggers, made by a core method similar to that of the
Scandinavian dagger tradition, with their widest part below the middle
(Fig. 8:9), were produced during the later part of the Scandinavian
Middle Neolithic period. The distribution maps of the British flint
daggers (Grimes 1932:343) seem to indicate that there was a centralised production of flint daggers in the vicinity of the flint mines of East
Anglia and Sussex (Green et al. 1982:500). On the basis of a comparison
of the outlines of British flint-dagger blades (Grimes 1932:341) and
bronze-dagger blades (Gerloff 1975: Pls. 2-4) it has been suggested that
the prototype of the British flint daggers was most probably the
Butterwick dagger in bronze (Green et al. 1982:499). On these bronze
daggers, the blade is fitted to the handle with rivets, as opposed to the
earlier, tanged, copper daggers from Bell-Beaker contexts. David Clarke
has concluded that the British flint daggers were consistently associated
with beakers dated to the interval 1850-1400 bc (uncalibrated) contemporary with the LN I period (Fig. 8:18) but that daggers have been
found in graves from Lanting & van der Waals step 2, which is contemporary with the Upper Grave phase on Jutland (MN B III). Hence,
the majority of British flint daggers are contemporary with the lancetshaped daggers from the Scandinavian LN I phase, except that there are
earlier examples as well. The early British flint daggers may have been
influenced, not only by Butterwick daggers, but also by tanged, BellBeaker daggers in copper, such as the one from Nttraby mentioned
above, especially if only the form of the blade is considered, and not the
differences between the tangs on bronze versus flint daggers, since they
would obviously be covered by the handle. If this was the case, the
249

Jan Apel

Fig. 8:10. A Scandinavian flint dagger of type IV E and atriangular bronze


dagger from the classical phase of the Unetice Culture. Drawings from Divac
& Sedlcek (1999) and Montelius (1917).

250

The Classification system of the Scandinavian Flint Daggers


production of British flint daggers can be dated to a late Bell-Beaker
horizon in the British Isles that is contemporary with the MN B-phase
on Jutland. Since the earliest production of Scandinavian type-I daggers
can be dated to around 2350 cal BC, and since the production took
place in the western parts of Denmark, it is likely that the British flint
daggers were prototypes of the daggers with lancet-shaped blades. The
development of the more pronounced dagger handle that occurs from
type I D onwards is, on the other hand, a phenomenon that is exclusive
to Scandinavia (Mller 1902:131).
From type VI D onwards the outline of the dagger blades is distinctly changed. As we have seen earlier in this chapter, they become
significantly wider in relation to their length and the widest part of the
blade lies near the blade-handle transition, so that the blade displays a
triangular shape with rounded corners. There is a general consensus that
the prototypes of, or templates for, this new shape were the triangular
bronze daggers (Fig. 8:10) made in Bohemia-Moravia during the classical
phase of the netice Culture (Mller 1902, Forssander 1936, Lomborg
1973, Vandkilde 1996). The triangular bronze daggers have handles
with expanded bases. The fact that a large proportion of the Scandinavian
flint daggers with triangular blades (type IV A onwards) also display
fishtail handles strengthens the idea that the triangular flint daggers were
the prototypes of the later Scandinavian flint daggers. The triangular
bronze daggers in the netice area were produced during the classical
phase of the netice culture, which extended from 1950 to 1700 cal
BC. Accordingly, the Scandinavian LN II phase, which is defined by
the production of daggers with fishtail handles and rounded-triangular
blades, can be dated to the same period.
It can thereby be concluded that there are strong reasons for suggesting that Scandinavian flint daggers with lancet-shaped blades and nonexistent, or straight-edged, handles were produced during a phase
contemporary with the pre-Wessex-Culture, Beaker phase in the British Isles dated to 2350-1950 cal BC. There are also strong reasons for
accepting that Scandinavian flint daggers with fishtail handles and triangular blades with rounded corners should be dated to the interval
1950-1700 cal BC. I now proceed to a history of the research that will
shed some light on the chronological and geographical implications of
the traditional dagger typology.

251

Jan Apel

Fig. 8:11. Mller and Forssanders flint-dagger typologies (from Becker 1964).

The flint-dagger typology and its relevance


The Danish archaeologist Sophus Mller was the first to treat the
Scandinavian flint-dagger material in a systematic way (Mller 1888 &
1902). He suggested a division of the Danish dagger material into five
types (Fig. 8:11). He also maintained that these types reflected a chronological sequence. In order to prove this theory, the typology was tested
on archaeological-find contexts from Denmark. Evidence in the form
of stratigraphic observations and find-combination data from graves
and hoards was presented in its favour (Mller 1902:130 ff.). The time
span of the prehistoric, flint-dagger-production period was already
delimited both backwards and forwards in time when Mller started
his work. Danish archaeologists already knew that megalithic passage
graves often contained flint daggers (Mller 1898:259), and in the middle of the 19th century, the Danish archaeologist Jens Jacob Asmussen
Worsaae demonstrated that some of the Gallery-Graves containing flint
252

The Classification system of the Scandinavian Flint Daggers


daggers were to be dated to the Early Bronze Age (Worsaae 1860:17 f.,
Grslund 1974:125). The archaeological contexts suggested that the
flint daggers belonged to the time between the passage graves and the
Early Bronze Age and that they consequently belonged to a Late Stone
Age period. Later on, when it was realised that the Danish passage graves
were built during a fairly short period, it was suggested that the flint
daggers in them belonged to secondary burials (Rosenberg 1929:256
ff.). Thus, the introduction of the earliest daggers was dated to when
they replaced battle axes as male prestige gifts in the graves of the Jutlandic
variant of the Corded Ware tradition, the Single Grave Culture
(Lomborg 1973:158).
The obvious analogy to metal daggers from later times helped Mller
to define and delimit in time flint daggers with an evident handle.
However, like the archaeologists before him, he had problems in defining oblong bifaces without an evident handle. In earlier works, these
forms were thought of as lance- and spearheads (Worsaae 1860:15,
Madsen 1868: Taf. 34-37, Montelius 1872-74:12 ff.). Mller solved
this problem by dividing the material into two forms, an earlier and a
later. The earlier form encompassed oblong bifaces with an edge covering
the whole long side and with straight bases, i.e. different types of sharp,
oblong artefacts that could not with certainty be classified as daggers.
The chronological and geographical aspects of lance- and spearheads of
the older form were later dealt with by Carl-Johan Becker (1957). The
palstaves from the Early and Middle Neolithic, Funnel Beaker Culture, another artefact category from the earlier form group, has recently
been treated by Klaus Ebbesen (1994).
Mllers later form group consisted of examples that today are
thought of as flint daggers. But, among the later forms, there were still
types that lacked evidence of a handle, namely those that were later to
form the first and consequently the oldest daggers in the present flintdagger typology. Mller is somewhat vague on how he distinguished
these daggers from the earlier form group. In 1888, he suggested a criterion based on size: examples with lengths between 15 and 20 cm were
classified as daggers, while longer and shorter examples were classified
as lance- and spearheads. Later on, he abandoned this metrical definition and claimed that the early daggers, in contrast to other, bifacial,
lance- and spearheads with straight bases, had curved long sides and
thin, convex or pointed bases (Mller 1902:126). He also claimed that
253

Jan Apel
the later dagger types (i.e. those with a pronounced handle) were exclusive to Scandinavia, while the earlier types were common European
forms (Mller 1902:131). We can assume that in this context he was
referring to the resemblance with daggers made of the French honey
coloured Grand-Pressigny flint found in Bell Beaker graves in Western
Europe and flint daggers made on the British Isles. Mller also points
out that the positioning of the earlier dagger types in Late Neolithic
graves confirms the dagger interpretation; they are often found beside
the hip as if they had been attached to a belt. Some 30 years later Mllers
idea was confirmed when a Type-I dagger in a leather sheath and belt
was found in a bog in Wiepenkathen near Hamburg in Germany (Cassau
1935). A wooden handle covered the base of the dagger and, in order to
make it fit, a piece of cloth had been inserted between the handle and
the base. This little cloth, by the way, is the earliest evidence of wool in
northern Europe (Jrgensen 1992). Microscopic use-wear investigations
of flint daggers from late Bell Beaker and Beaker contexts in Britain
have strengthened the dagger interpretation further. These daggers have
revealed a patina created by the fact that a form of handle had covered
the base and that the daggers were kept in leather sheaths (Green et al.
1982, Grace 1990:11). According to Green et al. (1982:495), the
repeated sheathing and unsheathing of a flint dagger would tend to
remove any micro-wear polishes left by daily use, simply because dry
hide/leather polish is a very abrasive type of micro-wear polish.

Sophus M llers typology 1888-1902


In his publication, Flintdolkene i den nordiske stenalder (The Flint
Daggers in the Nordic Stone Age) (1902) Mller presented his final
dagger typology. The types were defined from the form, design and
cross-section of the handle. The first type is lancet-shaped and lacks a
pronounced handle. The second type is lancet-shaped with a handle
that is thicker than the blade. The third type has a thick handle with a
square cross-section. The fourth type has a thick handle with concave
edges that widens towards the base. The handle has a rhombic-square
cross-section. The fifth and last type has a thick handle with concave
sides that broadens at the base. The handle has a triangular cross-section.
Mller also considered to some extent the form of the blade in his
division of the dagger material. He suggested that the earlier types had
254

The Classification system of the Scandinavian Flint Daggers


narrow blades with their maximum width near the middle of the blade,
while the later types had broader blades with their maximum width
near the handle. He also pointed out that the blades of the later types
were generally longer than the blades of the earlier types (Mller
1902:127 f.). Even though Mller did take into account the form of
the blade in his typological reasoning, it is evident that he used the
handle characteristics to the uttermost in his practical work. The fact
that he never defined a late dagger type with a straight handle with oval
cross-section and a blade with its widest part near the blade-handle
transition (type VI) indicates that the blade was not considered as
important as the handle (Bondesson 1980:2 f.). It is as if Mller
understood the importance of a typology that would work, in spite of
continuous sharpening of the blade. In 1912, Sune Lindqvist suggested
that the dagger typology should be enlarged with a type like the one
mentioned above and that it should be dated to the Early Bronze Age
(Lindqvist 1912:20). However, he lacked explicit evidence for this
dating. It was probably based on a German book about the Bronze Age
in Schleswig-Holstein, in which a dagger of this kind was pictured in a
Bronze Age context (Splieth 1900:14, Lomborg 1973:15).
Oscar Montelius dagger typology included both Mllers earlier
and later forms. Seven different types were defined and designated A to
G. In this typology, Mllers fourth and fifth types are regarded as one
and the same form type F. Apparently, Montelius did not agree with
Mller on the chronological significance of the cross-sections of the
fishtail handles. Instead of this, Montelius distinguished a dagger type
with a fishtail handle without seams on the sides and the edges type
G. This reasoning seems to be well founded, considering Lindqvists
definition of the Bronze Age dagger type (above). In this way, the
development of the dagger typology reflects the cornerstone in
Montelius thinking on evolutionary laws and material culture: that
the development of forms proceeds from simple to complex and then
degenerates before the production ceases. In this context, Montelius
own definition of type G daggers speaks for itself: Degenerate
descendants of type F. The handle is less carefully worked and the blade
is wider than the handle. The whole dagger reveals, in form and execution
alike, a certain slackness, compared with the earlier type (Montelius
1924:49, authors translation). It should be mentioned here that Mller
himself recognised and reproduced this type in his publication in 1902
255

Jan Apel

Fig. 8:12. Photographs showing the Skogsbo gallery grave during excavation in
1913 (Photos: ATA, Stockholm).

but that he, with hesitated to define these daggers as an independent


type (Mller 1902:129).

John Elof Forssanders typology 1936


In 1936, John Elof Forssander, incorporated Mllers and Montelius
typological opinions into his own typology in Der Ostskandinavische
Norden whrend der ltesten Metallzeit Europas. In this book,
Forssander uses Mllers distinction between daggers, on the one hand,
and lance- and spearheads, on the other. He also accepts Montelius
revision of the typology, and one could say that the book is partly an
attempt to prove his view empirically. With references to Lindqvist
(1912) and Splieth (1900), Forssander accepts Lindqvists idea of a late
dagger type belonging to the Early Bronze Age (i.e. D VI). Thus, a flintdagger typology consisting of six types is created (Fig. 8:11). Forssander
claims that the typology has chronological relevance, in that every single
type corresponds to a delimited phase of the Late Neolithic and Early
256

The Classification system of the Scandinavian Flint Daggers

257

Jan Apel
Bronze Age. He presents descriptions of graves from northern Germany
containing type-VI daggers that confirm the dating of this type to the
Early Bronze Age (Forssander 1936:127 f.). His sole argument for the
chronological order of the other types is based on stratigraphic evidence
from a Late Neolithic Gallery-Grave at Skogsbo in the province of
Vstergtland (Fig. 8:12) (Forssander 1936:121). It was excavated by
Sune Lindqvist in 1913 and contained 17 flint daggers of different types.
Later, the typology was criticised by archaeologists (Strmberg 1975).
It is obvious that those who had the greatest problem with the practical
application of the typology were working with materials recovered in
areas far from the flint source. The typology was regarded as less useful
in these areas, because the flint daggers that are found often show signs
of extensive re-sharpening. In order to test the typology on an
archaeological material, Lili Kaelas classified flint daggers from two
Swedish provinces, Vstergtland and Bohusln, kept in the collection
of the Archaeological Museum in Gothenburg (GAM). She concluded
that only 28% of the daggers from Bohusln and 38% of the daggers
from Vstergtland could be classified according to Forssanders typology (Kaelas 1964:135). Unfortunately, this statement is rather difficult
to judge, since all daggers with extensively re-sharpened blades were
sorted out, because they were regarded as impossible to classify. This is
rather curious, since, in most cases, there is sufficient evidence to classify
the dagger type by handle characteristics only. In 1975, Ebbe Lomborg
stated that he had no problem in classifying daggers that had been found
far away from the production area according to his revised version of
the Mller-Forssander typology (Lomborg 1975:116). My personal
opinion is that the typology can be used on daggers from northern and
central Sweden without any major problems, as long as the shape of the
handle is allowed to guide the classification. However, a more serious
problem with Kaelas investigation is that she used a metrical definition
of type-I daggers. Only those examples with lengths over 18 cm were
considered to be daggers, while shorter examples were regarded as arrowand spearheads (1964:145, note 16). This mixture of type definition,
on the one hand, and her personal interpretation of what the object was
actually used for during prehistory, on the other, confuses the
argumentation.
In her article, Kaelas also comments on the fact that Forssanders
idea of the chronological significance of his typology rests solely on his
258

The Classification system of the Scandinavian Flint Daggers


stratigraphic interpretation of the Skogsbo stone cist (Kaelas 1964:136).
By referring to the original excavation report, she is able to demonstrate
that the stratigraphic state of affairs regarding this stone cist was not as
clear as Forssander had intimated. Mrta Strmberg shared Kaelas
doubts regarding Forssanders dagger typology in an article on Late
Neolithic settlements in Skne published in 1971.

Ebbe Lomborgs typology 1973


In 1973, Ebbe Lomborg published a refined dagger typology, together
with a presentation of the Danish flint-dagger material. One of the
aims of this work was to demonstrate that the dagger typology had
chronological relevance (Lomborg 1973:158). In the book, the Danish
dagger material is classified according to Forssanders six types with some
minor alterations. In addition, Lomborg defines between two and six
subtypes of each main type. Some of these subtypes are regarded as
having chronological relevance, others as local forms. Herein lies the
main problem with Lomborgs typology. It is obvious that his ambition
was to classify all Danish daggers. In order to achieve this goal, he was
forced to use different kinds of criteria for different subtypes. The most
important subtypes are delimited within type I, where two
contemporary subtypes are defined I C and I D. With the help of
distribution maps showing grave, stray and hoard finds, Lomborg assigns
the type-I C daggers to northern Jutland and the type-I D daggers to
the eastern Danish islands (Lomborg 1973:39 ff.). Some minor
corrections regarding the definitions of types II and III are also made. A
dagger form that, according to Forssanders typology, would belong to
type II is here transferred to the subtype III A.
When Lomborg treats the Early Bronze Age dagger material (type
VI), he is faced with another problem concerning type definition. During the Early Bronze Age, the production of dagger-shaped strike-alights began. These are virtually impossible to distinguish from type-VI
daggers, considering the typological elements only (for a definition of
these strike-a-lights, see Lomborg 1959:160) and even in cases in which
they show signs of wear, it is difficult to distinguish a dagger from a
strike-a-light, because examples of all dagger types were often used as
strike-a-lights and therefore show similar wear patterns. In order to solve
this problem, Lomborg entered a metrical definition: type-VI daggers
that are under 13 cm in length are regarded as strike-a-lights, while
259

Jan Apel
examples that are 13 cm or longer are regarded as daggers (Lomborg
1973:27 f.). This is a very unfortunate definition, since daggers that are
found far from the production area have more frequently been resharpened and consequently tend to be shorter (Scheen 1979:58). In
cases like these, it is important to classify the objects according to context
and/or wear patterns. If artificial boundaries like these are used, the
defined types will risk reflecting only groups of more or less used daggers.
Lomborg was of the opinion that the six main types formed a chronological sequence. Thus, types I-V represented five, separate, chronological phases of the Late Neolithic, and type VI was still fixed firmly in the
Early Bronze Age. However, according to Lomborg, it would not be
meaningful to divide the Late Neolithic period into five phases, because
there were no other Late Neolithic artefacts that could be classified in a
way that would support such a detailed chronology. Instead, he chose
to divide the Late Neolithic into three phases designated LN A, LN B
and LN C. Daggers of type I define LN A, types II and III define LN B,
and types IV and V define LN C (Lomborg 1973:158). Evidence, in
the form of stratigraphic observations and find-combination data, was
presented to support this tripartition. These studies give, according to
Lomborg, ein berraschend eindeutiges bild. The diagram in Fig. 8:13
shows that none of the main types has been found underneath an older
type. However, it is worth mentioning that 8 of the 67 examples in the
diagram are the results of what Lomborg labels horizontal stratigraphy;
i.e. when two or more daggers have been found in a passage grave, for
instance, those nearest the entrance are considered to be later than those
in the chamber. It is unwise to use these observations in the way that
Lomborg does and he subsequently received criticism for this reasoning.
There are no facts that justify the claim that a dagger found near the
entrance belong to a later burial. On the contrary, it could be argued
that these daggers were older and that they were cleared out of the chamber
in order to make room for later burials. The diagram in Fig. 8:14 shows
that (almost) all daggers found in find combinations are combined either with the same type or with the nearest type (Lomborg 1973:67).
The two cases that do not fit the picture are explained away as having
been recovered during earlier, poorly conducted excavations and therefore not safe to trust. In this context, one must ask whether Lomborg
was suspicious also of old excavation reports in which his chronology
was confirmed?
260

The Classification system of the Scandinavian Flint Daggers

Fig. 8:13. Stratigraphical finds of flint


daggers in Denmark, according to
Lomborg (1973).

Fig. 8:14. Closed find combinations


containing flint daggers in Denmark,
according to Lomborg (1973).

In addition to his contributions to the Late Neolithic chronology,


Lomborg also presents interesting regional differences within Denmark
during the period in question. After an extensive analysis of sites and
artefacts, mainly graves, metal artefacts, and flint daggers, he is able to
divide Denmark into two zones during the Late Neolithic period. These
two zones correspond to the two cultural zones that are detectable later
on during the Bronze Age in southern Scandinavia (Lomborg 1969:94
ff.). Zone I consists of the northern part of Jutland, the eastern Danish
islands, and Skne. Zone II consists of the central and southern part of
Jutland and of Schleswig-Holstein in northern Germany. According to
Lomborg, the daggers were produced mainly in zone I. This area had
contacts with metal-producing areas, such as the British Isles (Wessex)
and the netice Culture in Bohemia and Moravia. Lomborg also notes
that there are considerable regional differences within zone I that are
detectable in the grave material. In northern Jutland, the dominant burial
forms are Late Neolithic barrows and secondary burials in stone cists of
the Middle Neolithic Single Grave Culture. On the islands and in Skne,
the dominant burial forms are inhumation graves dug in flat ground
an influence from the netice area and burials in Late Neolithic
stone cists (Lomborg 1973:132). According to Lomborg, the western
part of central Sweden have contacts with northern parts of Jutland
(Lomborg 1973:132).
261

Jan Apel
The typology has been used to classify Scandinavian flint daggers all
over northern Europe (see Chapter 9). During the 1970s and 1980s,
Kaelas initial critique of Forssanders typology was complemented with
a critique of Lomborgs refined typology. As we have seen earlier,
Lomborg does not trust a purely typological reasoning to sustain his
chronological results and therefore tries to strengthen his typology with
reference to find-combination data and stratigraphic observations. A
great deal of the criticism that has been published has concentrated on
this issue.

The debate in N.A.R. 1975


In 1975, Lomborg summarised his thesis in the Norwegian Archeological Review. Four scholars, experts on the Late Stone Age and Early
Bronze Age of northern Europe, supplied comments on the book. I
would briefly like to refer briefly to their critique and to Lomborgs
answers before continuing any further.
Niels Bantelmann pointed out that it might have been useful to try
to define the dagger types with the help of a detailed metrical analysis
(Bantelmann 1975:103). Lomborg replied that the frequent and extensive re-sharpening of dagger blades would make this difficult. He
stressed that the only meaningful measurements would concern the
width/thickness indexes of the hilts (Lomborg 1973:21 ff. & 1975:116).
Jay J Butler argued against the view that western European influences, mainly from the British Isles, triggered the production of the
earliest daggers in Scandinavia: The first flint daggers in British Beaker
graves occur in their step 5 (1800-1650 bp) [here Butler refers to Lanting
&Van der Waals beaker chronology for the British Isles published in
1973], whereas the beginning of the Late Neolithic in Jutland goes
back to about 1900 bp... (Butler 1975:105). Because of the relatively
late dating of the earliest graves with flint daggers in the British Isles,
these daggers could hardly be prototypes of the Scandinavian type-I
daggers. In his reply, Lomborg demonstrates that Butler was obviously
not aware of all the closed-find contexts with daggers in Britain when
he formulated this critique. As an example of an earlier flint-dagger
context in Britain, he referred to a grave with a flint dagger published
by David Clarke in 1970 that is dated to Lanting & Van der Waals step
2, contemporary with the Upper Grave phase (MN B III) of the Single
Grave Culture in Denmark (Lomborg 1975:121).
262

The Classification system of the Scandinavian Flint Daggers


Klaus Ebbesen put forward an extensive criticism on Lomborgs
ideas of the transition from the Middle Neolithic to the Late Neolithic
but this will not be dealt with further in this book. More important in
this context was Ebbesens critique of Lomborgs typology. He stated
that Lomborg had an inexact definition of the word type. The term
type is employed sometimes for a whole group of forms, sometimes
for part of such group, and sometimes even for single, often unique,
daggers. This impedes understanding and undermines the authors otherwise rigorous and well-documented dagger typology (Ebbesen
1975:107). In addition to this, I would like to add that the inconsistent
typological treatment of the daggers seems to stem from the fact that
Lomborg had two incompatible purposes with his dagger typology.
On the one hand, he wanted to strengthen Mller and Forssanders
opinion that the typology reflected a chronological sequence and on
the other, he wanted to formulate a typology in which every conceivable type would be included. Obviously he was aware that some of the
subtypes reflected merely regional types or even single, unique daggers.
However, one has to bear this in mind while working with the typology.
Another important aspect of Ebbesens comments was concerned
with Lomborgs treatment of Late Neolithic dagger hoards. Ebbesen
neatly demonstrates that hoards from LN A (Type I daggers) are concentrated to northern Jutland and hoards from LN B (type-II and typeIII daggers) are concentrated to the south-eastern part of the Danish
islands (Ebbesen 1975a:108, Fig. 1). Ebbesen claimed that, by failing
to observe these patterns, Lomborg deprived himself of an opportunity
to reflect on the organisation of trade and exchange in southern Scandinavia during the Late Neolithic period. Mrta Strmberg put forward
a similar critique when, in line with Lomborgs own observations (1973:
132), she argued that it is possible to identify several local regions within
zone I in Denmark, and she concluded that this area cannot be regarded
as uniform during the Late Neolithic period. As we shall see, other
archaeologists were to develop these ideas in a way that in the end was
to lead to the questioning of the chronological relevance of the typology.
Ebbesen also suggested that it might be fruitful to explain the transition between the Middle and the Late Neolithic and the changes accompanying it in technological terms, instead of by the appearance of
flint daggers (Ebbesen 1975:111). In line with this reasoning, the bifacial
263

Jan Apel
technology would, in itself, have created new technological opportunities that opened doors to a greater exploitation of natural resources. In
this context, Ebbesen pointed out that the most important tools to
study would be flint sickles and other practical tools and not flint daggers. If I have understood Ebbesen correctly, he maintained that the
bifacial sickles had technological advantages that in some way would
improve, for instance, harvest capacity or the collecting of leaves as
animal fodder. Perhaps one should look here for the causes of the sudden
disappearance of the old cultures, just as this circumstance may be behind
the enormous economic growth in population in the heart of South
Scandinavia together with a great expansion to the north (Ebbesen
1975:111). In my opinion this is an interesting idea but it needs to be
confirmed with experiments.
Lomborgs reply regarding the inconsistent typology was that he
had just followed old conventions regarding typology and that it is
perfectly clear from the type descriptions whether a type is regarded as
reflecting a chronological sequence or regional patterning (Lomborg
1975). When it came to the question of the regional differences in the
material, Lomborg avoided the principal discussion by referring to his
distribution maps.

A revision of the chronology


In 1978, Ann Segerberg published an article in which she applied a
source-critical view of the flint-dagger typology and its chronological
relevance. She thereby checked the accuracy of the arguments in favour
of the chronology by studying the original contexts, i.e. excavation
reports on graves, hoards, etc. She thus demonstrated that two of the
three stratigraphic evidences that Sophus Mller published in 1902,
cases in which an earlier dagger type was found under a later dagger
type, might have been the result of the finding of stray daggers in burial
mounds (Segerberg 1978:179). In this context, it is important to
remember that Mller based his chronological interpretations partly on
unpublished Danish materials of stratigraphic observations from graves
(Lomborg 1973:17). Segerberg also concluded that Kaelas (1964) suspicions regarding Forssanders interpretation of the stratigraphy in the
Skogsbo cist were justified. Apparently, Forssander considered only those
daggers in the cist that confirmed his chronology. As a result of this, he
gave an account of only 10 of the 17 daggers in the cist (Segerberg
264

The Classification system of the Scandinavian Flint Daggers


1978:181) (Fig. 8:15). Segerberg reconstructed the stratigraphic sequence
in the cist with the help of a drawn cross-section made in 1950 by
Sverker Jansson (one of the excavators). She was thereby able to show
that type-II daggers occurred from the bottom to the top of the filling
in the cist, that type-III daggers occurred near the topsoil and in connection with type-II daggers, that type-IV daggers were found a couple
of centimetres above type-V daggers, and that type-V daggers were found
at the same level as the highest type-III dagger (Segerberg 1978:182). In
other words, this context could not confirm Forssanders chronological
interpretations.
Segerberg also made a source-critical evaluation of nine of the contexts that Lomborg used and concluded that only three of these had an
unambiguous stratigraphy (Segerberg 1978:184). For instance, Lomborg
did not take into account archaeological evidence in which two separate
graves with the same type of dagger are present. In such cases, the dagger
typology was used to date the graves and they were consequently
interpreted as being contemporary. Segerberg concluded that the relative chronology of the Late Neolithic period, based on the flint-dagger
material, was in no way decided by the flint-dagger typology as it was
presented by Lomborg (Segerberg 1978:185).
Torsten Madsen was next in line to question the dagger chronology
(Madsen 1978) and it is fair to say that his article was a corner stone in
the revision of the Late Neolithic chronology that archaeologists from
Aarhus University conducted during the 1980s and 1990s. He used the
dagger typology in a fundamental discussion on artefacts and their suitability for chronological studies. The main point in his reasoning was
that a seriation based on find-combination data or typology that seemed
to display chronological sequences could in fact mirror not only chronological but also spatial differences in the material. Minor differences in
the spatial distribution of an artefact type could result in more or less
separate groups of combinations that were not necessarily are chronologically dependent (Madsen 1978:54). This form of Doppler effect
had already been observed in an American debate on the problems of
artefact seriation (Deetz & Dethlefsen 1965).
The majority of the closed contexts that were used by Lomborg to
create the seriation diagram were hoards. It is therefore of great importance, in order to be able to give meaning to the geographical distribution, to reach an understanding of what these hoards represent, in this
265

Jan Apel

Fig. 8.15. Daggers and spearheads in flint from the Skogsbo gallery grave
(From Forssander 1936: Taf. XXIII-XXIV).
266

The Classification system of the Scandinavian Flint Daggers


context, the spatial separation of hoards with different dagger types. A
prerequisite for this interpretation was that the hoards were deposited
near the production sites. Malmer has suggested that the size and frequency of hoards with Late Neolithic or Early Bronze Age flint sickles
are dependent on the distance from the natural flint sources (Malmer
1988:91). Whether one chooses to regard them as caches or as offerings, the flint-dagger hoards from southern Scandinavia follow the same
pattern; the large hoards were deposited at, or close by, the raw-material
sources.
In line with Ebbesen (1975:107 f.), Madsen concluded that daggers
of types I and II are unevenly distributed in Denmark. According to
Madsen, it is likely that these regional differences caused the clear-cut
division between the two types in Lomborgs seriation diagram (Fig.
8:14). Madsen pointed out that the two stratigraphic contexts that
Lomborg presented in favour of the chronological interpretation (Fig.
8:13) could be rejected because they concerned daggers found in stone
chamber graves and because one was the result of a horizontal,
stratigraphic interpretation (Madsen 1978:55). It is important to remember that four or five contexts, showing the same stratigraphic evidence, are required before it can be inferred that two types are not of
the same age (Malmer 1963:33, Segerberg 1978:184, Madsen 1978:55).
The result of this analysis shows that there is no reason to believe that
the differences between types I and II in Lomborgs diagram reflect a
chronological sequence. On the contrary, they seem to reflect spatial
differences. Madsen was also able to demonstrate that these two types
overlap each other chronologically, although the type-II daggers, together with daggers of type I D, were introduced later than the other
type-I daggers and that they are the result of a local production on the
Danish islands. Madsen used information provided in Lomborgs thesis and, with the help of a statistical analysis, he was able to predict the
ways in which the differences between the types were due to chronology (Madsen 1978:55 ff.). The result of this investigation is presented
in Fig. 8:15. It is worth noting that Madsen retained the tripartition of
the period, even though Lomborgs chronological interpretation of the
dagger typology was revised.
Madsen suggested that the variation of types in the Danish flintdagger material could have been the result of different, regional-workshop traditions. Such a scenario would naturally cause problems in us267

Jan Apel
ing closed hoards in seriation tables based on find combinations. If the
hoards represent workshop stocks or, for that matter, offerings deposited near the workshop, their distribution would reflect the production
of a single workshop. The result of this is that the hoard content will be
homogenous and the chronological differences, compared with other
dagger hoards, may be greater than expected (Madsen 1978:57).
Lisbeth Wincentz Rasmussen draws the obvious conclusions from
Madsens revised dagger chronology (Rasmussen 1990). She presents a
new division of the Late Neolithic period into an earlier and a later
phase, instead of Lomborgs and Madsens tripartition. In this new
division, daggers of types I and II represent the early phase and daggers
of types III-V the later phase. Daggers of type I A-C are assigned to a
delimited period in northern Jutland when this part of Scandinavia was
influenced by the western European, Bell-Beaker and Beaker traditions.
The area around the Limfjord was, according to Rasmussen, the centre
of the dagger production during the early phase. Daggers of types I D
and II are seen as local variants from the eastern Danish islands. This
area, in turn, shows signs of contact with the netice area during the
early phase (Rasmussen 1990:38).
The Beaker influences can be noted on the sites of the period in the
form of typical pottery from settlements and in graves in north-western Jutland. There are, for example, seven graves in which a type-I dagger
has been found with Bell-Beaker pottery (Rasmussen 1990:41, note
16). There are several graves in this area in which type-I daggers are
combined with typical, Bell-Beaker, artefact categories, such as V-incised,
conically-shaped buttons in amber and heart-shaped, bifacial arrowheads
in flint (Rasmussen 1990:35). A fragment of a wrist guard in slate has
also been found in the early Late Neolithic settlement of Myrhj in
northern Jutland (see Jensen 1973 and Chapter 6 in this book). This
artefact category is typical of the western European Bell-Beaker tradition
(Rasmussen 1990:37). Fragments of type-I daggers were also found on
the site (Jensen 1973:88) and a considerable proportion of the flint
debitage suggests that the later stages dagger-production was actually
conducted on the site (Chapter 6). Rasmussen concludes that the
population on the eastern Danish islands, where daggers of types I D, II
and III were made, had contacts with the netice area in central Europe.
This is due to the fact that metal prototypes of needles and other artefacts in bone found in eastern Danish graves were made in the netice
268

The Classification system of the Scandinavian Flint Daggers

Fig. 8:16. (a) The relative chronology of the flint-dagger types, based on the
numbers of daggers, the regional differences and the find combinations. (b)
The Late Neolithic phases in accordance with a. (c) The Late Neolithic phases
in accordance with Lomborg (1975) (from Madsen 1978).

269

Jan Apel
area (Rasmussen 1990:37). According to Rasmussen, the regional differences between western and eastern Denmark ceased when the production of type-III daggers started. This conclusion, as we have seen
earlier, was not drawn on account of any major changes in the distribution of hoards, since the type-III hoards, like the type-II hoards, are
concentrated to the eastern Danish islands (Lomborg 1973:50 f., Figs.
25-28). Instead, Rasmussen argues that the general distribution of daggers
implies that they were more widely consumed. At the transition
between the early and the late period, the core area of the flint-dagger
production was moved from northern Jutland to the islands. This state
of affairs was strengthened when the production of dagger with fishtail
handles commenced (Rasmussen 1990:37).
In the late 1980s, Helle Vandkilde presented a division of the Late
Neolithic into two phases that, in some minor but important details,
differ from Rasmussens (Vandkilde 1989 & 1996). Here, Lomborgs
SN A and SN B are incorporated into an earlier Late Neolithic phase
(LN I) while Lomborgs LN C (that corresponds chronologically to
the classical phase of the netice Culture) is renamed LN II. Thus,
Vandkilde, as opposed to Rasmussen, consider that the daggers of type
III belong to the earlier phase. In the previous section of this chapter,
the argumentation behind this decision has been explicitly described.
Nevertheless, it may also be appropriate to mention that Vandkildes
division does not proceed primarily from the dagger typology. It is
based on a thorough, typological investigation of the metal hoards from
the Late Neolithic and the Early Bronze Age in Scandinavia (Vandkilde
1996). Vandkilde presents a typology of the bronze axes in the hoards
and is able to define two phases in this material. A seriation diagram
based on find combinations of different artefact categories shows that
an early metal horizon coincided with the later Late Neolithic period,
LN II, and that two metal horizons coincided with the first two periods of the Early Bronze Age. These results are formalised by the application of a statistical method (correspondence analysis) that confirms
the seriation. The development through time in the chemical composition of the early metal objects and their alloying techniques also confirms
the chronology (Vandkilde 1996:136 ff.). Vandkildes results, as regards the production of flint daggers, are presented in Fig. 8:17. The
local production of metal artefacts replaced the production of flint daggers during the latter part of the Late Neolithic. The technologically
270

The Classification system of the Scandinavian Flint Daggers

Fig 8:17. A quantitative analysis of the relationship between the numbers of


metal artefacts and flint daggers produced during the Late Neolithic periods.
The differences in length between the Late Neolithic phases have been taken
into consideration. The LN I phase lasted for 400 years, which means that
0.13 recovered metal artefacts and 7 recovered flint daggers were produced
per year. The LN II phase lasted for 250 years, which means that 0.97 recovered metal artefacts and 1.7 recovered flint daggers were produced per year
(from Vandkilde 1989).

advanced type-IV daggers may, seen from this angle, have been the result
of a last effort by the flint knappers to challenge the new metal
technology.
An alternative interpretation of the flint dagger has been presented
by Gundela Lindman (1987) and has been further developed by Eva
Weiler (1994). On the basis of Kaelas (1964) and Segerbergs critique
of the dagger typology, they suggest that it lacks chronological value
(Lindman 1987:128). According to Lindman and Weiler, the typology
primarily reflects social hierarchies in the Late Neolithic population.
Lindman claims that the type-IV daggers form one group and all the
other dagger types form another group. It is further stated that the
type-IV daggers signified high status and therefore belonged to chiefs.
The other daggers are, on the other hand, regarded as common types.
Lindman admits that, although type-IV daggers signified high status,
there are indications that this type was produced during a relatively
short period. However, Lindman avoids this possibility in the following paragraph: But this reasoning creates problems. First one has to
explain why the leaders all appeared during a very short time period in
the later part of Late Neolithic. Where did they dwell during the other
271

Jan Apel
parts of the period? (Lindman 1987:133 f.). The chronological discussion, earlier in this chapter, indicate that it might pose great sourcecritical problems to regard daggers from LN I and II as contemporary
types in social analyses. However, these ideas may be worth discussing
in areas far away from the natural flint sources, where older types might
have been in circulation for greater lengths of time (see Weiler 1994).

Conclusion
The Scandinavian flint daggers appear in distinct, morphological groups.
This fact has tempted archaeologists, since the 19th century, to formulate flint typologies that would reflect a chronological sequence. During the last two decades, a revision of the detailed chronological results
that Lomborg presented in 1973 has been conducted. The chronological relevance of the flint-dagger typology has received critique from
two directions. Some scholars have argued that the closed-find contexts
and stratigraphic observations that make up the empirical basis of
Lomborgs chronology can be questioned (Segerberg 1978). Likewise,
it has been maintained that Mller and Forssanders typology was difficult to use on materials far from the production areas, owing to the
extensive re-sharpening these daggers had been subjected to (Kaelas
1964). Others accept the data but are of the opinion that the typology
does not constitute a chronological sequence in which each type follows the other (Ebbesen 1975, Madsen 1978, Rasmussen 1990). This
critique is based on the regional distribution of hoards containing different dagger types and on the assumption that the dagger hoards reflect the work of local or regional workshops.
In 1989, a scheme (Fig. 8:18) was published that has resolved a
major part of the discussion on the Late Neolithic chronology (Vandkilde
1989). This scheme is based on almost 300 c14-samples from southern
Scandinavia and a central and western Europe. For the first time, relative
chronologies from different areas in this large region can be compared.
The beginning of LN I is set at 2350 BC, the transition to LN II is set
at 1950 BC and the transition to the Early Bronze Age at 1700 BC
(Vandkilde 1989:29). These results make the flint-dagger typology
superfluous as a detailed chronological tool.
If we accept the critique of Lomborgs chronological conclusions, it
seems that the daggers were produced in two core areas during LN I: a

272

The Classification system of the Scandinavian Flint Daggers

Fig. 8:18. A comparative chronological scheme based on evaluations of archaeological


finds and contexts and over 300 c14 dates (from Vandkilde 1996).

western area around the Limfjord in Jutland and an eastern area including the eastern Danish islands and south-western parts of Skne in
Sweden. Type-I daggers, except for subtype I D, were produced in the
western area, and subtype I D, type II and III daggers were produced in
the eastern area (Fig. 8:19)
The Late Neolithic period in Scandinavia was about 650 years long.
Flint daggers were produced for at least 800 years. We know that the
daggers with lancet-shaped blades were produced during LN I and that
the core production area during this phase was the Limfjord area in
northern Jutland. The impulses that triggered the production seem to
have come from western Europe, most probably from the British Isles.
It is likely that the prototypes of the lancet-shaped daggers were tanged
copper daggers found in western European, Bell-Beaker contexts and
British flint daggers found in Beaker contexts contemporary with the
Late Single Grave phase on Jutland. During LN II (250 years), daggers

273

Jan Apel

Fig. 8:19. The distribution of Danish hoards with daggers from LN I. Type-I
daggers, except subtype I D (circles), subtype I D, type II and type III daggers
(triangles), based on information from Lomborg (1973).

with pronounced handles and triangular blades with rounded corners


were being produced. The core production area during this phase had
moved to the eastern Danish isles, even though ordinary types were
produced. During this phase, influences came from the central-European, netice Culture, and it is likely that the prototypes of the daggers
of types III-V were bronze daggers with handles and wide hilts made in
Bohemia and Moravia. These metal daggers have ready-made handles
and these handles are what the flint knappers tried to duplicate during
this time.
It is importance to note that the production of flint daggers decreased significantly with the transition to LN II. At the same time the
local production of metal objects increased dramatically in southern
274

The Classification system of the Scandinavian Flint Daggers


Scandinavia. However, the quality of the daggers produced during LN
II is higher than that of the earlier daggers, at least if the technical
difficulties involved in the production are considered. This may mean
that the daggers lost part of their role in society with the introduction
of metal objects. A parallel to this phenomenon can be studied in the
Dutch Bell-Beaker culture. In the early phases, flint daggers made of
French Grand-Pressigny flint are often found in the graves in this culture. When the local metal production starts in the Rhine delta, these
daggers are no longer a part of the burial gifts, being replaced by metal
objects and tools used in the production of metal objects (Harrison
1980:17 ff.). It is too early to say whether this was the case in Scandinavia. I would like to return to this subject at some time in the future.

275

Chapter 9

The Scandinavian Flint


Daggers

In traditional societies, specialised craft production and long-distance


trade are two ways in which tangible objects are charged with intangible,
supernatural powers that give their owners honour and power (Helms
1998). It is therefore crucial for the purpose of this book to understand
the geographical distribution of the Scandinavian flint daggers and the
principles that guided the exchange. I shall therefore now put into play
the theoretical ideas presented in Chapters 2-4 and further discussed, in
relation to other types of archaeological materials, in Chapters 6 and 7.
The most important thing to be investigated is whether the complexity
of the exchange networks, in which flint daggers were distributed, in
some way measures the complexity of the social organisation of Late
Neolithic Scandinavia. If flint daggers were exchanged from hamlet to
hamlet, in the form of a down-the-line system, this would reflect a
segmentary social organisation. However, if the daggers were involved
in a more direct exchange from producers through central places to
consumers this would indicate a chiefdom level of complexity, in
which local chiefs participated in peer-exchange networks in which
prestigious artefacts were distributed.
This investigation will include studies of the relative frequencies of
flint-dagger types, the relative lengths of daggers and the degree to which
the daggers were re-sharpened in different regions of Scandinavia. It is
anticipated that the results of this study will shed further light on the
degree of social complexity in the Late Neolithic societies of Scandinavia.
I shall also present the Swedish flint-dagger material and relate it
mainly to the Danish and Norwegian materials (as presented by Lomborg
1973 and Scheen 1979). In order to be able to discuss exchange networks
and social organisation, it is also of great importance to delimit the area
in which these kinds of daggers were utilized. Thus, the published
277

Jan Apel
Scandinavian flint-dagger materials of the northern and central parts of
Europe parts of the exchange network are also included in the
study (Clark 1932, Meinander 1954, Bloemers 1969, Wyszomirski
1974, Wojciechowski 1976, Khn 1979, Bantelmann 1982, Agthe
1989a & b, Trnka 1991 & MS, Rassmann 1993, Beuker & Drenth
1995).
Only daggers with at least parish provenience are used in this study.
Presumably, several thousands of daggers have been excluded because
they lack information on where they were found. Not all museum
collections, and no private collections, have been visited during the
registration of the data.1 In all, 4 454 daggers have been registered. This
may seem like a small number, especially since Malmer registered 6 515
flint daggers with known find locations in Skne only (Malmer
1962:699). This is partly explained by the fact that Malmer visited
more museums and private collections than I have done. However, it
may also to some extent be due to the fact that we used different
typological criteria. Malmer did not make a distinction between lanceand spearheads and palstaves in flint, on the one hand, and flint daggers,
on the other (Malmer, personal communication). Even though it is
highly unlikely that these differences make up for the discrepancy in
the numbers of daggers from Skne, they certainly contributed. It is
assumed that the collected material is representative of the flint daggers
that were deposited in prehistory. This is likely, since the discerned
pattern, as regards the frequencies and different types, coincides with
investigations conducted with a smaller material base (see Apel 2000a
& b). Thus, it is anticipated that the present material will be sufficient
to reveal important and general aspects of the production and
consumption of flint daggers in Sweden.

1 Flint daggers from the following Swedish museum collections are included in this study: the
Gvleborg County Museum, the Archaeological section of the Gothenburg City Museum, the
Halland County Museum, the Jmtland County Museum, the Lund University Historical
Museum, Gustavianum Museum at Uppsala University, the National Museum in Stockholm,
the Uppland County Museum and the Vstmanland County museum.

278

The Scandinavian Flint Daggers

Representativity
Archaeological investigations based on stray finds are subject to severe
source-critical problems. However, this should not discourage
archaeologists from working with stray finds. As I see it, those who
choose to do so have two options. Either they may set out to record
every conceivable problem related to the use of stray finds in an
archaeological analysis or they may collect a material that is large enough
to allow of general and large-scale interpretations, in order to limit the
effect of any source-critical problems. To discuss and evaluate every
possible source-critical problem is a never-ending task that, if it is to be
convincingly pursued, will demand at least a book in itself, such as the
books on archaeological-formation processes (Schiffer 1976 & 1987,
Kristiansen 1985). In this context, therefore, the former strategy has
been chosen. It is apparent that the distribution of the Scandinavian
flint daggers, as regards the frequencies of the different types, forms
patterns that cannot be explained by post-depositional, cultural and
non-cultural, formation processes but that reflect the prehistoric reality
(see Apel 2000a & b).

Swedish flint daggers


I have personally registered the flint daggers found in Sweden and used
in this study. As regards the type classification of the Norwegian flint
daggers, I have relied on Rolf Scheens dissertation of 1979. The distribution of daggers in Norway, presented in Figs. 9:2-9:10, is also based
on information in Scheens publication. However, the Norwegian daggers have been measured and classified according to the degree of resharpening by myself.2
In Fig. 9:1, all the Swedish parishes where flint daggers have been
recovered are plotted. In Table 9:1, the material has been classified in
provinces according to Lomborgs main types. Flint daggers are scarce
north of a line running from the northern part of the province of Uppland
west towards Norway. In the following text the daggers from the
following northern provinces have been merged under the designation
Norrland: Dalarna, Gstrikland, Hlsingland, Hrjedalen, Jmtland,
2 This work was done in 1998 while I was visiting the following museums: the Historical
Museum in Bergen, the Archaeological Museum in Stavanger, the Oldsagsamling of Oslo University
and the Museum of Natural History and Archaeology in Trondheim.

279

Jan Apel
Province
Blekinge
Bohusln
Norrland
Dalsland
Gotland
Halland
Nrke
Skne
Smland
Sdermanland
Uppland
Vrmland
Vstergtland
Vstmanland
land
stergtland
Total

Type I
10
89
9
81
9
71
15
262
11
12
12
1
140
5
6
22
755

Type II
12
60
1
36
8
63
11
372
22
9
7
4
66
6
7
24
708

Type III Type IV Type V


23
13
22
100
14
32
6
1
2
107
25
30
15
8
21
116
24
57
16
5
13
315
69
191
31
16
32
16
11
8
18
7
16
10
1
3
144
43
63
11
3
10
13
3
12
30
13
27
971
256
539

Type IV
6
72
5
34
14
78
18
193
19
16
14
8
57
8
13
35
590

Unclassified
8
9%
60
14%
8
25%
95
23%
8
10%
70
15%
13
14%
192
12%
16
11%
9
11%
16
18%
7
21%
96
16%
8
16%
4
7%
25
14%
635
14%

Total
94
427
32
408
83
479
91
1594
147
81
90
34
609
51
58
176
4454

Table. 9:1. The Swedish flint-dagger material divided into Lomborgs main types and
registered by province.

Lappland, Medelpad, Norrbotten, Vsterbotten and ngermanland.


These provinces cover two-thirds of Sweden and only 32 daggers are
known from this vast area. Most of them originated from Jutland and
reached this area via the Norwegian west coast (Fig. 9:1). This may be a
general pattern that governs several different artefact types dating from
the Neolithic period. One observation that supports this idea is that
Neolithic flint artefacts from southern Scandinavia tend to be scarce in
the southern parts of Norrland, while they are more common in the
northern parts (Danielsson 2000:12). Thus, other southern-Scandinavian
artefact categories may have reached the northern parts of Norrland via
the Norwegian west coast and Trndelag. This will have to be evaluated
in future research, however. As regards the flint daggers, the conclusion
based on observations presented in this chapter and in earlier articles
(Apel 2000a & b) is that the overwhelming majority of the groups of
people who used flint daggers lived in the southern and central parts of
Sweden. Incidentally, this is the area that the first Scandinavian farming
communities inhabited (c. 4000 cal BC).
Fourteen per cent of the Swedish daggers would not, for some reason,
be assigned to any of the main types (Table 9:1). The majority of these
was fragmented daggers that had been severely re-sharpened or had been
turned into other tool types, such as strike-a-lights or axes, adzes and
chisels. However, a small number consisted of complete daggers that
had been formed in an unconventional way that made it impossible to
assign them to any of Lomborgs main types. It is apparent that a larger
280

The Scandinavian Flint Daggers

Fig. 9:1. The


distribution of
flint daggers in
Sweden. All the
parishes where
flint daggers have
been recovered
are marked by a
dot. The daggers
are projected on
a map showing
the coastline at
3600 BP (c. 2000
cal BC) made by
Leif Andersson at
the Swedish
Geological
Survey (SGU) in
Uppsala.

281

Jan Apel

1
2-3
4-5
6Fig. 9:2. The geographical distribution
of type-I flint daggers
in Sweden and
Norway. Information
on the Norwegian
daggers has been
extracted from
Scheen (1979). The
daggers are projected
on a map showing
the coastline at 3600
BP (c. 2000 cal BC)
made by Leif
Andersson at the
Swedish Geological
Survey (SGU) in
Uppsala.

282

The Scandinavian Flint Daggers

Fig. 9:3. The geographical distribution


of subtype-I C flint
daggers in Sweden
and Norway. Information on the
Norwegian daggers
has been extracted
from Scheen (1979).
The daggers are
projected on a map
showing the coastline
at 3600 BP (c. 2000
cal BC) made by Leif
Andersson at the
Swedish Geological
Survey (SGU) in
Uppsala.

283

Jan Apel

1
2-3
4-5
Fig. 9:4. The geographical distribution
of subtype-I D flint
daggers in Sweden
and Norway. Information on the
Norwegian daggers
has been extracted
from Scheen (1979).
The daggers are
projected on a map
showing the coastline
at 3600 BP (c. 2000
cal BC) made by Leif
Andersson at the
Swedish Geological
Survey (SGU) in
Uppsala.

284

The Scandinavian Flint Daggers

1
2-3
4-5
6Fig. 9:5. The geographical distribution
of type-II flint daggers
in Sweden and
Norway. Information
on the Norwegian
daggers has been
extracted from
Scheen (1979). The
daggers are projected
on a map showing
the coastline at 3600
BP (c. 2000 cal BC)
made by Leif
Andersson at the
Swedish Geological
Survey (SGU) in
Uppsala.

285

Jan Apel

1
2-3
4-5
6Fig. 9:6. The geographical distribution
of type-III flint
daggers in Sweden
and Norway. Information on the
Norwegian daggers
has been extracted
from Scheen (1979).
The daggers are
projected on a map
showing the coastline
at 3600 BP (c. 2000
cal BC) made by Leif
Andersson at the
Swedish Geological
Survey (SGU) in
Uppsala.

286

The Scandinavian Flint Daggers

1
2-3
4-5
Fig. 9:7. The geographical distribution
of type-IV flint
daggers in Sweden
and Norway. Information on the Norwegian daggers has
been extracted from
Scheen (1979). The
daggers are projected
on a map showing
the coastline at 3600
BP (c. 2000 cal BC)
made by Leif
Andersson at the
Swedish Geological
Survey (SGU) in
Uppsala.

287

Jan Apel

1
2-3
4-5
6Fig. 9:8. The geographical distribution
of type-V flint
daggers in Sweden
and Norway. Information on the
Norwegian daggers
has been extracted
from Scheen (1979).
The daggers are
projected on a map
showing the coastline
at 3600 BP (c. 2000
cal BC) made by Leif
Andersson at the
Swedish Geological
Survey (SGU) in
Uppsala.

288

The Scandinavian Flint Daggers

1
2-3
4-5
6Fig. 9:9. The geographical distribution
of type-VI flint
daggers in Sweden
and Norway. Information on the
Norwegian daggers
has been extracted
from Scheen (1979).
The daggers are
projected on a map
showing the coastline
at 3600 BP (c. 2000
cal BC) made by Leif
Andersson at the
Swedish Geological
Survey (SGU) in
Uppsala.

289

Jan Apel

1
2-3
4-5
6Fig. 9:10. The
geographical distribution of flint-dagger
preforms in Sweden
and Norway. Information on the
Norwegian preforms
has been extracted
from Scheen (1979).
The daggers are
projected on a map
showing the coastline
at 3600 BP (c. 2000
cal BC) made by Leif
Andersson at the
Swedish Geological
Survey (SGU) in
Uppsala.

290

The Scandinavian Flint Daggers

5
4

3
6
7
9
2
1

10

Fig. 9:11. Different regions of Scandinavia discussed in the text. (1) eastern
Denmark, (2) western Denmark, (3) eastern Norway, (4) western Norway, (5)
Norrland, (6) eastern central Sweden, (7) western central Sweden, (8) southeastern Sweden, (9) south-western Sweden and (10) Skne.

291

Jan Apel
share of the daggers found far away from the natural flint sources are
difficult to classify (Table 9:1). This may be explained by the fact that
the flint was more valuable in these areas.

Scandinavia
For practical purposes, Scandinavia will now be divided into regions on
the basis of the homogeneity of the flint-dagger material (Fig. 9:11).
Denmark is divided into an eastern and a western part because of the
two main sources of good-quality flint, regarded as being crucial for the
maintenance of the flint-dagger technology. The following parishes
belong to the eastern region: Bornholm, Copenhagen, Fredriksborg,
Holbk, Maribo, Odense, Prst, Sor and Svendborg and the
following belong to the western region: Haderslev, Hjrring, Randers,
Ribe, Ringkbing, Skanderborg, Snderborg, Thisted, Tnder, Vejle,
Viborg, benr, lborg and rhus. Thus, the eastern area comprises
the Danish islands east of Jutland, while the western region comprises
the Danish part of the Jutlandic peninsula (Fig. 9:11).
Norway has also been divided into an eastern and a western region.
This division is based on obvious differences in the Norwegian
archaeological materials dating from the Late Neolithic that are due to
both cultural and natural circumstances (Apel 2000a & b). The eastern
area consists of the provinces of Akershus, Aust-Agder, Buskerud,
Hedmark and Opland, Telemark, Vestfold and stfold and the western
area consists of Finnmark, Hordaland, Mre og Romsdal, Nordland,
Nord-Trndelag, Rogaland, Sogn og Fjorda, Sr-Trndelag, Troms
and Vest-Agder,
The following division of the southern and central parts of Sweden
is based on the data presented in Figs. 9:2 to 9:10 and in Table 9:1. This
area has been divided into five regions:
Eastern central Sweden, consisting of the provinces of Uppland,
Vstmanland, Nrke, Sdermanland and Vrmland.
Western central Sweden, consisting of the provinces of Bohusln,
Dalsland and Vstergtland,
South-eastern Sweden, consisting of the provinces of Gotland, Smland,
land and stergtland
South-western Sweden, consisting only of the province of Halland.
292

The Scandinavian Flint Daggers


Because of the large dagger material from Skne in southern Sweden,
this province is also treated separately.
I have earlier suggested that daggers from the two production areas
were distributed to two different consumption areas during LN I (Apel
2000 a & b, Apel 2001). This pattern still holds up and is most visible
in the distribution of the two subtypes I C and I D (Figs. 9:3 & 9:4).
As we have seen in Chapter 8, these two types were made at the same
time in the two production areas. The parallel-flaked, type-I C daggers
were deposited in graves in northern Jutland during LN I. Debitage
from the production of this type has been recovered on bell-beaker sites
in Himmerland, such as Myrhj (see Chapter 6). Daggers of type I D
were at the same time produced in the eastern area on the Danish isles
and in the south-western parts of Skne. The outlines of these daggers
are similar to those of type I C, but, instead of having a pressuredflaked finish, the blades of type I D were finished by percussion. The
geographical distribution of these daggers reflects two production areas
that were connected with two consumption areas (Figs. 9:3 & 9:4).
The two consumption areas are also reflected in the distribution of
type-I daggers (Fig. 9:2), on the one hand, and that of type-II daggers
(Fig. 9:5), on the other. The type-III daggers (Fig. 9:6) are more evenly
distributed, though we should remember that the hoards of these daggers
were found mainly in the eastern production area (see Fig. 8:18).
During LN II, the overall production of flint daggers decreased.
However, within some production fraternities, especially in the eastern
area, this period saw an increased specialisation. The triangular bronze
daggers of Bohemia and Moravia (Divac & Sedlcek 1999) inspired
the production of fishtail daggers in flint. Type IV (Fig. 9:7) may be
regarded as a prestigious type, even though there are several examples of
poorly made copies all around Scandinavia. During the same period,
the less elaborately made, type-V daggers (Fig. 9:8) were also produced.
Finally, during the Early Bronze Age, small, type-VI daggers were
made and used, among other things, as strike-a-lights. Most daggers of
this type are small and the blades tend to have a rougher texture,
compared with earlier types. However, there are also large, sword-like
examples, such as the ones depicted in Chapter 8 (see Fig. 8:7). Like the
other dagger types, these are obvious copies of bronze objects, in this
case, different kinds of swords. While specialised production fraternities
produced sword-like, type-VI daggers, the production of simpler and
293

Jan Apel
smaller type-VI daggers may have been of a more general character,
maybe carried on at the household level.

Schleswig-Holstein, Germany
Over 2 000 flint daggers of Scandinavian types have been recovered in
Schleswig-Holstein in Germany (Table 9:2). This area consists of the
southern part of the Jutlandic peninsula and consequently borders on
Denmark. As we saw in Chapter 6, there are known production sites
with defined, bifacial-production debitage in the northern parts of this
area (Arnold 1990). The dagger preforms are concentrated mainly in
the eastern coastal area of Schleswig-Holstein (Khn 1979, Karte 2).
The large amount of preforms found here, in combination with the
fact that production debitage has been tied by refitting to daggerproduction sites, suggests that Schleswig-Holstein should be regarded
as a production rather than as a consumption area. However, this
production was probably not extensive. There are only three areas in
the northern parts of Germany where primary flint resources have been
located: on the islands of Helgoland off Holstein and Rgen off the
coast of the eastern part of Mecklenburg. The ice-transported flint in
northern Germany is not suitable as raw material for large tools and
blades (Hoika 1998:2). There is also a primary flint source further south
at Hemmoor in the Lneburg district of Germany (Beuker & Drenth
1998:8). Therefore we have to assume that the large, prestigious daggers
of Schleswig-Holstein were made of flint taken from the primary flint
sources of northern Jutland or the islands of Falster and Lolland, while
small daggers may have been produced in local flint. The majority of
the preforms that have been assigned to types have incipient handles
(18 examples) while only eight examples of type-I-like, lancet-shaped
preforms are present. Type-I daggers are not represented. The relative
distribution of dagger types (Fig. 9:12, Table 9:2) indicates that, from
this point of view, this area may have had closer contacts with the eastern
production area of southern Scandinavia than with the western
production area of northern Jutland. Khn argues that the flint-dagger
types of Schleswig-Holstein, like those of the Danish dagger material,
form a chronological sequence. However, the two diagrams presented
in favour of this idea do not constitute convincing evidence (Khn
1979:47, Abb. 7 & 8). For instance, of 12 closed-find combinations,

294

The Scandinavian Flint Daggers


Area

Austria
British Isles
Eastern central Sweden
Eastern Denmark
Eastern Norway
Finland
Mecklenburg
The Netherlands
Norrland
North Rhine-Westphalia
Thringen and Saxony
Poland
Schleswig-Holstein
Skne
South-western Sweden
South-eastern Sweden
Western central Sweden
Western Denmark
Western Norway
Total
%

Type I

Type II

Type III

Type IV

Type V

Type VI

Total

0
0
45
503
144
2
619
28
9
7
130
22
373
262
71
58
310
1511
456
4550
35

0
0
37
252
4
3
31
11
1
7
7
1
82
372
63
73
162
84
8
1198
9

0
0
71
252
111
0
214
17
6
3
21
5
393
315
116
112
351
210
118
2315
18

2
0
27
126
10
2
22
3
1
4
14
3
145
69
24
53
82
168
17
772
6

0
1
50
168
31
1
147
3
2
2
7
10
269
191
57
114
125
252
45
1475
11

3
4
64
252
145
0
235
8
5
1
34
6
807
193
78
87
163
420
353
2858
22

5
5
294
1553
445
8
1268
70
24
24
213
47
2069
1402
409
497
1193
2645
997
13168
100

Table 9:2. The Scandinavian, flint-dagger material of northern Europe divided into
Lomborgs main types.

only three combined two different dagger types: type V was once
combined with type IV and type V is twice combined with type VI.

Mecklenburg, Germany
Mecklenburg is situated immediately to the south-east of SchleswigHolstein along the Baltic shoreline. In the east, Mecklenburg borders
on Pomerania in the north-western part of Poland. In this area, 1 268
Scandinavian flint daggers have been registered (Rassmann 1993:283
ff.). The relative distribution of types in this material differs rather
curiously from the Schleswig-Holstein daggers. Here, type-I daggers
are predominant, while the other types are quite scarce (Table 9:2). In
this respect, the type profile of this area shows striking similarities
with the western production area in northern Jutland (Fig. 9:12).
According to, Rassmann, flint daggers were produced in this area too
(Rassmann 1993:68 ff.). However, there is no study in which flakes
from dagger-reduction sequences have been defined. The interpretation
rests solely on findings of dagger and bifacial-sickle preforms. Therefore,

295

Jan Apel
70%

70%

60%

60%

50%

50%

40%

40%

30%

30%

20%

20%

10%

10%
0%

0%
T ype I

T ype II

Type I

T ype III T ype IV T ype V T ype VI

Type II

Type III

Thringen, Saxony and Sachsen-Anhalt


(205 daggers)

Poland (47 daggers)

70%

70%

60%

60%

50%

50%

40%

40%

30%

30%

20%

20%

10%

10%

0%

Type IV

Type V

Type VI

Type IV

Type V

Type VI

Type IV

Type V

Type VI

Type V

Type VI

0%
Type I

Type II

Type III

Type IV

Type V

Type VI

Type I

Type II

Type III

Schlesvig-Holstein (2069 daggers)

Jutland (2 645 daggers)

70%

70%

60%

60%

50%

50%

40%

40%

30%

30%

20%

20%

10%

10%

0%

0%
Type I

Type II

Type III

Type IV

Type V

Type VI

Type I

Type II

Type III

Eastern Danish islands (1 553)

Mecklenburg (1 268 daggers)

70%

70%

60%

60%

50%

50%

40%

40%

30%

30%

20%

20%

10%

10%

0%

0%
Type I

Type II

Type III

Type IV

The Netherlands (57 daggers)

Type V

Type VI

Type I

Type II

Type III

Type IV

Western Norway (997 daggers)

Fig. 9:12. The frequencies of Lomborgs six main dagger types in different
regions of northern Europe (in percentages).

296

The Scandinavian Flint Daggers


70%

70%

60%

60%

50%

50%

40%

40%

30%

30%

20%

20%

10%

10%

0%

0%
Type I

Type II

Type III

Type IV

Type V

Type VI

Type I

Type II

Type III

Type IV

Type V

Eastern Norway (445 daggers)

South-western Sweden (409 daggers)

70%

70%

60%

60%

50%

50%

40%

40%

30%

30%

20%

20%

10%

10%

0%

Type VI

0%
Type I

Type II

Type III

Type IV

Type V

Type VI

Type I

Type II

Type III

Eastern Central Sweden (295 daggers)

Skne (1 402 daggers)

70%

70%

60%

60%

50%

50%

40%

40%

30%

30%

20%

20%

10%

10%

0%

Type IV

Type V

Type VI

Typ IV

Typ V

Typ VI

Type IV

Type V

Type VI

0%
Type I

Type II

Type III

Type IV

Type V

Type VI

Typ I

Typ II

Typ III

Western Central Sweden (1 193 daggers)

Norrland (24 daggers)

70%

70%

60%

60%

50%

50%

40%

40%

30%

30%

20%

20%

10%

10%

0%

0%
Type I

Type II

Type III

Type IV

Type V

South-eastern Sweden (497 daggers)

Type VI

Type I

Type II

Type III

North Rhein-Westphalia (24 daggers)

297

Jan Apel

Fig. 9:13. The distribution of Grand Pressigny blade daggers (black dots) and
locally made, pseudo Grand Pressigny daggers made in local flint types (from
Struve 1955: Tafel 33).

in this study, this area is regarded as a consumption area. However, it


may be useful to remind the reader that the so-called, pseudo Grand
Pressigny, flint daggers were produced in local flint in this area in the
latter part of the Middle Neolithic period (Fig. 9:13). This dagger type
consists of local copies of the honey-coloured, Grand Pressigny, blade
daggers made in the Loire valley in present-day France. This area
evidently has a long tradition of making dagger-like objects in flint
and, consequently, smaller examples of Scandinavian flint daggers may
have been produced here as well.

Poland
The distribution of Scandinavian flint daggers in Poland is almost
exclusively concentrated in Pomerania, i.e. the north-western province
of the country (Wojciechowski 1976: 72, Mapa. I). There are only a
few flint daggers in evidence east of the river Vistula (Wyszomirski
1974:77, Fig. 1) and only a handful in the south-western parts of the
country (Wojciechowski 1976: 72, Mapa. I). Wojciechowski points
out that it is difficult to determine whether some of the Polish flint
298

The Scandinavian Flint Daggers


daggers should be regarded as lance- and spearheads or whether they are
Scandinavian dagger types (Wojciechowski 1976:38 ff.). This is especially
true of the examples that disappeared during the Second World War
and are poorly described in the remaining registers. He also mentions
that secondary re-sharpening of the edges also complicates the
classification. However, 47 Polish flint daggers have been positively
classified as Scandinavian types, and another 21 examples have, with
some hesitation, been classified as such (Wojciechowski 1976:49 ff.). A
slightly modified typology, compared with Lomborgs (discussed in
Chapter 8), is used (for details, see Wyszomirski 1977). However, as
regards the questions discussed in this context, it is perfectly compatible with the other dagger materials presented in this chapter.
According to Wojciechowski (1976:65), the earliest, Scandinaviantype, flint daggers in Poland appear together with undecorated pottery
belonging to the second, Polish, corded ware phase, which is
contemporary with the later part of the MN B phase in Scandinavia.
The majority of the Polish daggers should, however, be dated to the
classical phase of the netice culture, 1950-1700 BC cal. and the typeVI daggers to the phase immediately before the emergence of the Lausitz
culture in the area, i.e. around 1500 BC cal. The early types (I, II and
III) are concentrated in Pomerania, while the later types can also be
found further south in the provinces of Great Poland and Silesia
(Wojciechowski 1976, Mapa. 1 and 2). Wojciechowski also remarks
that especially type-V daggers also occur in Bohemia and Moravia in
the present day Czech and Slovak republics. i.e. in the core area of the
netice culture (see Figs. 9:14 & 9:15). This strengthens the hypothesis
presented in Chapter 8, in which type-IV and type-V daggers were dated
to the period between 1950 and 1700 cal BC on account of their
similarity with the triangular bronze daggers of the classical phase of the
netice culture (see Chapter 8).
Wojciechowski concludes that the flint daggers of Scandinavian types
found in Poland are made of flint from Jutland and not from the nearest,
primary flint source, at the island of Rgen. Apparently, the indigenous
flints of Poland, such as the chocolate flint from the Radom area, the
grey, white-spotted flint from the Annopol area and the banded flint
from the Ostrowiec Swietokrzyski region (for further information on
these flint types, see Domanski & Webb 2000) during the whole
Neolithic period never reached Pomerania (Wojciechowski 1976,
299

Jan Apel

Fig. 9:14. The geographical distribution of the netice culture (dots) and the
Scandinavian flint daggers (circles) in Bohemia (from Agthe 1989:50, Abb. 24).

Fig. 9:15. The geographical distribution of the proto-netice culture (dots) and
the proto-netice flint-dagger graves (circles with dot) in Moravia (Agthe
1989:51, Abb. 25).

Czerniak & Kabacinski 1997, Hoika 1998). The north-western part of


Poland was apparently not closely connected with the areas in south
eastern Poland, where high-quality flint was extracted, but instead
displays Scandinavian influences in a pattern that goes back to the
Mesolithic. Peter Bogucki, for instance, has suggested that Pomerania
may have been part of a social territory that included Mecklenburg,
Schleswig-Holstein, Denmark and southern Sweden, i.e. the Maglemose
tradition (Bogucki 1987:43, Fig. 3.2). It is intriguing to consider that a

300

The Scandinavian Flint Daggers


similar pattern may have been present during the Late Neolithic period
as well.

Thringen, Saxony and Sachsen-Anhalt, Germany


Marcus Agthe has published the flint daggers of the Scandinavian types
found in the north-western part of the netice area, namely the parishes
of Cottbus, Dresden, Erfurt, Gera, Halle, Karl-Marx-Stadt and Leipzig
and those found in the southern parts of Magdeburg (Agthe 1989a &
b). This is an area situated just north of the present German border
with the Czech republic and 340 flint daggers have been found here.
Of these, 127 could not be assigned to the typology (Agthe 1989b:305).
The type profile of the remaining 213 examples is presented in Table
9:2 and Fig 9:12. Agthe shares Lomborgs (1973) and Khns opinion
that the typology reflects a chronological sequence (1979:305 f.).
It is concluded that the earliest type-I daggers found in graves in this
investigation should be dated to the older phase of the netice culture,
while the latest belong to the hocker-phase of the netice culture (Agthe
1989b:308). However, it should be pointed out that, in neighbouring
areas, Scandinavian daggers have been found in graves of the Bell Beaker
culture (Schwarz-Mackensen 1985: 28 ff., Agthe 1989b:308) and this
suggests a dating prior to 2350 cal BC. The later dagger types of the
investigation cannot be dated. However, Moucha has suggested that
type-IV daggers in Bohemia are associated with the classical phase of
the netice culture (1963:143, Abb. 87). Therefore, Agthe maintains
that type-I daggers, on the one hand, and the later types, on the other,
were not used during the same period in his investigation area. However,
the relative lack of type-II and type-III daggers may suggest that the
netice area had few contacts with the eastern production area during
LN I. This would in fact also explain the type profile of this area.

North Rhine-Westphalia, Germany


Bantelmann (1982) has presented the flint daggers from the area around
Dortmund in the western part of central Germany. He has divided this
material into nine form groups. Flint daggers of Lomborgs Scandinavian
types belong to groups 5-7. Within these groups, the daggers are divided
into Lomborgs types, which makes the material compatible with the
other materials presented in this chapter. However, I would like to point

301

Jan Apel
out that Bantelmanns form group 8, should most probably be regarded
as extensively re-sharpened daggers of the Scandinavian type I. In this
respect, they are not, as Bantelmann suggests (1982:44), flint copies of
tanged, bell-beaker daggers in bronze. They have gained their
characteristic form from the fact that the blade has been extensively resharpened in a way that makes it look like a tang. This is a problem that
has also been recognised by Beuker & Drenth (1999:33) in their reassessment of the flint-dagger material from the Drenthe province of
the Netherlands. In this context, I include these daggers in Bantelmanns
form group 5, which contains daggers of Lomborgs type I. Accordingly, the Scandinavian daggers in this area consist of 7 daggers of type
I, 7 of type II, 3 of type III, 4 of type IV, 2 of type V and 1 of type VI.
Only 24 Scandinavian daggers have been recovered in this area. For
obvious reasons, this makes it difficult to judge the type profile presented
in Fig. 9:12 and Table 9:2.

The Netherlands
The daggers of Scandinavian types in the Netherlands were published
1969 by Bloemers. In all, 99 daggers are included in his article. Of
these, 17 are regarded as uncertain and 12 are regarded as being of
Scandinavian origin but impossible to assign to any of Forssanders six
types. Bloemers also presents six examples of strike-a-lights. In this
context, they have been assigned to types V and VI, on account of the
published drawings and the published information. This means that 70
Scandinavian-type daggers are known from the Netherlands. All of
Forssanders (and Lomborgs) six main types are represented in this
material. The majority of the daggers are concentrated in the northeastern part of the country, especially in the province of Drenthe. Very
few daggers have been recovered south and west of Veluwe (Bloemers
1969:47, Abb. 2). There are no primary-flint sources in the Netherlands
and all flint is imported. Beuker & Drenth (1999:33) suggest that all
flint daggers arrived as finished products, since only one preform is
known. As regards the chronological aspects, the flint daggers of the
Netherlands cannot contribute to further understanding. Only one
dagger of Scandinavian type has been found in a grave (Harrison
1980:26). Therefore, the chronological knowledge of the Scandinavian
flint daggers in the Netherlands relies on evidence from Denmark and

302

The Scandinavian Flint Daggers


Germany. However, all Dutch specimens belong to the period between
2200 and 1500 cal BC. Curiously, it has been suggested that a proportion
of the Dutch daggers were made in flint from the island of Helgoland
off the coast of Germany (Beuker & Drenth 1998:33). This is not
unlikely, since Helgoland has primary-flint sources. Let us hope that
future research will investigate this further. Daggers from LN I
predominate in the Netherlands (Fig: 9:12 and Table 9:2).
It is interesting to note that the average length of the Scandinavian
flint daggers in the Netherlands is 164 mm (calculation based on
Bloemers 1969:106). This is a fairly high figure, compared with those
in the Swedish dagger material presented in Table 9:3, and may reflect a
direct exchange with flint-dagger-producing areas (see below).

Finland and the Baltic countries


In connection with his treatment of the Late Neolithic Kiukais culture
in Finland, Meinander mentions 12 flint daggers of Scandinavian types
found in Finland (Meinander 1954). One of these, a type-III dagger
found at Saltvik in the land archipelago, is nowadays considered to
have reached Saltvik as part of the ballast of a ship of southern Scandinavia
that came here during historical times. Thus, 11 Scandinavian flint
daggers have been found in Finland (Table 9:2). According to Meinander
the majority of the flint objects from the Finnish Late Neolithic,
including these daggers, originate from southern Scandinavia. The
geographical distribution is confined to the south-western parts of the
country (Wyszomirski 1974:88, Fig. 3).
According to Mark Wyszomirski (1974:90), nine, flint daggers of
Scandinavian types have been found in the Baltic countries: three in
Estonia, two in Latvia and two in Lithuania (Table 9:2). This means
that only 24 Scandinavian-type flint daggers have been recovered in this
vast geographical area, covering the eastern parts of Poland, the Baltic
countries and Finland, in other words, the area east of the river Vistula
and east of Sweden.

Austria
Five, Scandinavian, flint daggers have been found in Austria (Trnka
1991 and ms). According to Gerhard Trnka, these are currently the
southernmost finds of Scandinavian flint daggers (Trnka, ms). Their

303

Jan Apel
geographical distribution is limited to the eastern part of Austria (Lower
and Upper Austria as well as Styria). The finest example is a type-IV
dagger from Gusen-Berglitz in Upper Austria. Apart from this nice
specimen, Trnka presents four fragmented daggers of Scandinavian types.
All these are examples of later types and it may be that the Scandinavian
daggers that reached as far south as the core area of netice culture, and
beyond, did so during the classical phase of the netice culture, 19501700 cal BC.

The British Isles


Scandinavian-type flint daggers have also found their way to the British
Isles. Graham Clark mentions at least six examples (Clark 1932:186
ff.). According to the published drawings, these flint daggers are from
the LN II and early Bronze Age periods: one is of type V B and three,
possibly four, are of type VI (Clark 1932:188 f.). The remaining
specimen is a handle fragment that seems to be impossible to classify
according to the typology. As regards these daggers, Clark mentions
that from the normal British types they stand out as uncompromisingly alien (Clark 1932:190).

General distribution of the Scandinavian flint


daggers
The distribution of the Scandinavian flint daggers is mainly concentrated in an area from the central parts of Germany (Thringen, Saxony
and North Rhine-Westphalia) in the south to the northern parts of the
Norwegian west coast in the north and from the Rhine delta in the
west to the north-western parts of Poland in the east. Outside this area,
a small number of daggers have also been recovered in Austria, in the
British isles, in the Baltic countries and Finland, in the northern parts
of Sweden and in the central parts of the netice culture in Bohemia
and Moravia. However, I suggest that, within the main area of
distribution, daggers were valued according to the same cultural
principles, while this was not necessarily the case of the few daggers that
were deposited outside the main area.
We have already seen that Pomerania in Poland and the provinces of
Mecklenburg and Schleswig-Holstein shared, as early as the Early/Middle
Mesolithic phase, important traits with the southern parts of Scandinavia
(Bogucki 1987). It has further been suggested that the Late Mesolithic,
304

The Scandinavian Flint Daggers


Erteblle complex included not only the southern parts of Scandinavia,
Schleswig-Holstein and Mecklenburg/Pomerania but also the northwestern parts of Germany and the Rhine delta (Czerniak & Kabacinski
1997), i.e. the main distribution area of the Late Neolithic flint daggers.
It has also been demonstrated, by an analysis of the distribution of
characteristic tools of different flint types, that the same area was
connected during the Early Neolithic and the early Middle Neolithic
periods as well (Hoika 1998). Thus, this area had shared certain materialcultural traits for a long period at the time when the production of flint
daggers commenced. It is therefore suggested, in this context, that the
Late Neolithic exchange network for prestigious objects included the
southern and central parts of Scandinavia, the Norwegian west coast,
the north-eastern part of the Netherlands, the northern parts of Germany
and the north-western part of Poland.
During LN I the western production area had close contacts with
western Norway and indirectly with the northern part of Sweden (Fig.
9:16 & 9:17). The eastern parts of Germany and Pomerania in the
north-western part of Poland also seem to have been influenced mainly
by the western production area in northern Jutland. As regards SchleswigHolstein and the Netherlands, the composition of the LN I dagger
material suggests that these areas were connected with the eastern
production area. In this context, we also have to bear in mind that there
may have been an indigenous production of flint daggers along the
east-coast of Schleswig-Holstein and that parts of the Dutch dagger
material might originate from this area as well. We know from the
presence of bell-beaker pottery and other typical bell-beaker artefacts
on LN I sites on northern Jutland that the western production area had
extensive contacts with the Rhine delta during LN I.
During LN II the presence of fishtail-shaped daggers in Bohemia
and Moravia (and even the eastern parts of Austria) shows that contacts
with the classical phase of the netice culture were established.

Down-the-line or direct exchange


One way of gaining information on the social complexity of a traditional society is to investigate how prestige objects were distributed. In
most segmental, tribal societies, it is thought that prestige objects were
made in one place, often in the vicinity of the raw-material source, and
then distributed from hamlet to hamlet, by exchange. The archaeological
305

Jan Apel

Fig. 9:16. The composition of the flintdagger types during


LN I in different
regions of northern
Europe.Type-I
daggers from the
western production
area (white) and type
II and III daggers from
the eastern production area (black).

306

The Scandinavian Flint Daggers

Fig. 9:17. Exchange


routes from the two
main production
areas to different
parts of northern
Europe during LN I.

307

Jan Apel
indication of such a down-the-line, distribution system is that the number
of artefacts declines as a function of the distance from the production
area. However, since this is a general pattern it may also apply to redistribution based on exchange, as well as to organised trade; it has to
be complemented. A more discriminating pattern for the existence of
down-the-line exchange is that artefacts found in areas far from where
the production took place would tend to circulate for a longer period
of time before being deposited than artefacts near the production area.
Thus, if an artefact type that has been continuously re-sharpened, such
as flint axes or daggers, is found far from the area of production, it
would tend to have been more re-sharpened than examples found nearer
the production area (Hodder & Lane 1982, Olausson 1983, Sundstrm 1992). Hence, the degree to which the average length varies
between the Swedish provinces, as well as the degrees to which daggers
from different provinces have been re-sharpened, will be investigated
below. If there is a significant decrease in the average length of the daggers
in relation to the distance from the production area, it is possible to
argue that the daggers were exchanged in a down-the-line network. If
the daggers are more frequently re-sharpened, the further away they are
from the production area, there will equally be a chance of finding
arguments for a down-the-line, exchange network. Such results will
point to a society without centralised power on the regional level.
In more complex traditional societies the exchange of prestige objects
was organised in a different way. Here, local aristocracies often
belonging to the lineage with the closest genealogical ties to the mythical
lineage ancestor had direct contact with other aristocracies and
participated in an exchange network in which prestige objects circulated.
These aristocracies had exclusive control over the influx of prestige objects
and other foreign materials into their sphere and, by redistributing these
objects to the other segments of society, they were able to collect returns
in the form of physical support, surplus food and daily goods.
Consequently, a prerequisite for a redistribution system of this kind is
the presence of regional, administrative centres from which foreign and
valuable objects are distributed locally (Renfrew & Bahn 1991:310).
The archaeological traits indicating a distribution organised in this way
are more difficult to comprehend. It has been suggested that significant
concentrations of artefacts, within the consumption area, may indicate
local re-distribution centres (Renfrew & Bahn 1991). However, as I see
308

The Scandinavian Flint Daggers


it, the best indication of a more complex, social structure would be that
no evidence of down-the-line exchange can be brought forward. Hence,
the archaeological material in all regions participating in the network
would have been of a similar, generalised nature on the regional level.
Before continuing any further, I would like to recapitulate an
investigation, concerning the distribution of Early Neolithic, thin-butted
axes, conducted by Lars Sundstrm (1992). This investigation is for
several reasons of great concern in relation to the similar investigations
that will be presented in the following pages. First of all, these axes were
also made in southern Scandinavia and were distributed over a larger
area, similar to the area to which flint daggers were distributed. Secondly,
since these axes are made of flint, their formal properties will, like the
formal properties of the daggers, change considerably as the result of resharpening.
Sundstrm started out by assuming, in line with Hodder & Lane
(1982) and Olausson (1983), that the archaeological signal indicating
down-the-line exchange would be that axes would gradually became
shorter as a function of the distance from the production area. The
presence of such a pattern would indicate that the Early Neolithic FunnelBeaker society in Scandinavia was of a non-centralised character and
that axes were exchanged from farm to farm. Thus, Sundstrm set out
to measure all stray finds of thin-butted axes from the provinces of
southern and central Sweden. This investigation showed that the average
length of the axes in each province increased with the distance from the
production area and was actually the opposite of what had been expected.
However, at the same time, the standard deviation of the length
distribution of the axes in each province demonstrated that the axe
population in each province became more heterogeneous, the further
away from the production area one looked. This turned out to be due
to the fact that two groups of axes were in circulation: (1) working axes
that were continuously re-sharpened and thereby became shorter as a
function of distance from the production area, and (2) long, prestigious
axes that were exchanged without being re-sharpened. The combination
of the length distributions of these two categories of axes created a pattern
that was at first difficult to understand. However, in the new perspective,
it was clear that the distribution of thin-butted axes in Scandinavia
should be regarded as the result of down-the-line exchange.

309

Jan Apel

Degree of sharpening
There has been no systematic treatment of the problems related to the
degree to which flint daggers were sharpened. However, virtually all
those who have worked on flint daggers have pointed out the
methodological problems created by the continuous re-sharpening of
dagger blades in prehistoric times. For instance, and as we have seen
earlier, Lomborg introduced the subtypes I A/B and I x in order to be
able to classify extensively sharpened daggers that lacked a pronounced
handle. Regarding the rest of the dagger types, Lomborg was of the
opinion that the characteristics of handles would make it possible to
classify them according to the typology, even in cases in which the blade
had been sharpened to the point at which it lost its original outline.
All finished flint daggers have been sharpened at least once. Sharp
edges were apparently one of the characteristics that proper flint daggers
should have. The following description is based on information acquired
from Errett Callahan during our joint efforts to document the
production of type-IV daggers (Callahan & Apel, ms).
From the technical point of view, the sharpening of the edges of a
bifacial flint artefact, such as a flint dagger, will take place when the
blade has reached its intended outline during the production. The
sharpening process can be divided into two separate acts, each relating
to a repertoire of distinct gestures. (1) The entire edge of the blade is
retouched. Flakes are there evenly removed by pressure technique from
the blade-handle transition on one edge, down to the tip and to the
blade-handle transition on the other side. These flakes should be released
so that the flake scars penetrate well into the middle zone of the blade
(Callahan 1996: vii). This will give the blade the desired, evenly lenticular
cross-section. When the first stage of the retouch is finished, the edge
will display a denticulate outline. In fact, earlier on, bifacial sickles with
a denticulate edge were regarded as saw blades. Nowadays, we have to
accept that they may just as well be sickles that have not been subjected
to the second phase of the sharpening process. However, as regards certain
types of lance- and spearheads, the denticulate edges were no doubt
anticipated as the end result. (2) During the second phase of the sharpening process, the tip of each toothcreated during the first stage is
removed by a very light-handed, pressure technique performed with a
small pressure flaker (Callahan & Apel, ms). The edge that is created by
a properly conducted, sharpening process is almost knife-sharp and will
310

The Scandinavian Flint Daggers


endure a lot of cutting (of soft materials) before it has to be sharpened
again. It is indeed intriguing to consider that the Scandinavian flint
daggers were not, like their counterparts in bronze on the continent,
easy to use as stabbing devices, especially since the actual tips of the
Scandinavian daggers were often quite blunt. If the blade of the dagger
was intended to answer any practical need, apart from the obvious
symbolic meaning that an elaborate artefact of this kind undoubtedly
had, it was as a cutting knife. Practical experience has shown that blades
of this kind are excellent butchering knives. Errett Callahan and other
flint-knappers in America regularly sell bifacially shaped, flint knives to
hunters, who use them to butcher game. In this context, it is also worth
mentioning that knives with bifacially shaped blades were used as
ceremonial sacrificial knives by Maya priests in Meso-America (see, for
instance, Sievert 1990). It has been demonstrated that the thickness of
the flint-dagger blade, in combination with bifacially retouched edges,
makes for a superior cutting performance during butchering (Walker
1978:712). The reason for this is that the edge of a bifacial blade tends
to follow the path of least resistance between the skin and the body of
the game without destroying the hide. Obviously, since a bifacial, flintdagger blade is brittle and non-elastic, bones and other hard parts of the
material have to be avoided, since the flint dagger, unlike modern, steelbladed knives, cannot be used as a bending device. However, it has been
suggested that the sharp dagger blades did not actually serve any practical
purpose but should be regarded as playing a purely symbolic role in
certain rituals. If this was the case, the extensive and continuous
sharpening of the blade would also have to be regarded as part of this
ritual (Stenskld, ms). In any event, it has been of interest to record the
degree to which each dagger has been subjected to re-sharpening.
In order to establish the degree of sharpening of the blades on the
Swedish and Norwegian flint daggers, a qualitative judgement of each
individual dagger, as regards the status of the blade, and particularly its
edges, has been commenced. Thus, each dagger has been classified in
one of the following categories:
1. Daggers with blades that have been subjected to a final retouch,
conducted by an experienced flint-knapper who understood the
principles of retouching bifacial edges. The edges are in pristine
condition and do not appear to have been used. They may, however,
311

Jan Apel
display damage that happened during the discovery or in the
museums. Dagger blades of this category display a evenly curved,
lenticular cross-section.
2. Daggers with blades that have been re-sharpened by a person who
understood the principles of retouching bifacial edges. The scars of
the pressure flakes that indicate the first stage of the retouching process
run smoothly well into the blade, so that the edge angle is not
significantly decreased. Dagger blades of this category display a
lenticular cross-section (Fig 9:18). The blade surface on daggers resharpened in this way is clean and smooth, without major hinges
and other problems related to insufficient know-how.
3. Daggers with blades that have been re-sharpened by a person who
did not understand the principles of biface retouching. Usually, this
inexperience results in short, retouch, flake scars and, as a consequence
of this, higher edge angles, and a cross-section of such blades will
reveal an uneven, lenticular cross-section (Fig 9:18). The blade surface
on daggers of this kind will frequently exhibit abrupted, negative,
flake scars and hinges.
4. Daggers that have been turned into other kinds of artefacts by
additional knapping or grinding. Usually, it was broken daggers that
were transformed. It was for instance, quite common for handles of
types III and IV to be ground into axes and chisels. Dagger tips, on
the other hand, were turned into spear- and arrowheads.
In this context, it is important to acknowledge that the difference in
the degree of sharpening between categories 2 and 3 is qualitative in its
nature. This means that a dagger with a blade of category 2 may very
well have been extensively sharpened, so that it displays a significantly
altered, width-thickness relationship. However, the way in which this
sharpening has been executed will result in an evenly lenticular crosssection, as opposed to the uneven cross-sections of blades of category 3.
In certain cases, dagger blades of category 3 are not extensively sharpened,
although they were sharpened by someone who lacked the proper
knowledge and know-how needed to produce functional blades. This
means that the qualitative classification system presented above is needed
in order to complement any measurements taken on the daggers.

312

The Scandinavian Flint Daggers

Fig. 9:18. Dagger blade, cross-sections showing the different re-sharpening


abilities of categories 2 and 3 in the text.

Measurements
Unfortunately, very few of those who have studied Scandinavian flint
daggers have bothered to publish data on the measurements. As we
shall see, measurements of daggers can indeed play an important role in
determining the social and political aspects of the Late Neolithic societies
in question. In this book, the measurements are also used in an
independent way, in discussing the degree of sharpening of daggers that
in some respect complements the qualitative classification system
described above. Consequently, the great majority of the Swedish flint
daggers included in my database have been measured. The three most
important measurements, used in Chapter 9, are as follows:
1. Maximum length, taken from the middle of the base to the tip of
the dagger.
2. Maximum width of blade, taken perpendicularly to the length on
the widest part of the blade.
3. Blade thickness, taken on the widest part of the blade.

The length of the Scandinavian flint daggers


The length distribution of the Swedish flint daggers forms a bell-shaped
curve (Fig. 9:19). On this general level, a division of this material into
the two categories that Sundstrm was able to define among thin-butted
axes cannot be applied to the dagger material. In fact, if we look at the
length distribution of the daggers in each province, this division is not
apparent here either (Fig. 9:20). A way of investigating whether this
conclusion is reliable is to look at the standard deviation of the length
distributions in each province (Table 9:3). There is no general trend in

313

Jan Apel

079
80
10 -99
01
12 19
01
14 3 9
01
16 59
01
18 79
01
20 99
02
22 19
02
24 39
02
26 59
02
28 79
029
9
30
0-

900
800
700
600
500
400
300
200
100
0

Fig. 9:19. Length distribution of the Swedish flint daggers.

this material that relates directly to the distance from the production
areas. It should be noted that the dagger populations in some provinces,
such as Skne and Vstergtland, have higher standard deviations,
indicating that these materials are more heterogeneous, as regards the
length distribution, than others. However, this pattern cannot be related
to the distance from the production area. It can thus be concluded that
no fall-off pattern can be recognised from the length distribution of the
Swedish flint-dagger material. As regards length, the composition of
the dagger materials in the different provinces does not vary in a
significant way.
Now we have to bear in mind the specific way in which flint daggers
are re-sharpened. Since the edges are located mostly on the sides of the
daggers, from the theoretical point of view, a re-sharpened dagger may
not lose very much length, compared with an axe. This means that,
instead of investigating length, it may be fruitful to investigate whether
or not a fall-off tendency can be discerned, as regards the width of the
blades. In Fig 9:21, the distributions of the blade-width values of the
Swedish dagger material are presented. These distributions also take on
an almost bell-shaped form. In order to see if the dagger blades became
narrower, the further they were from the production area, the thickness
of the blade (mm) was divided by the maximum width of the blade
(mm). If the result was a relatively low value, it would mean that the
dagger blades are wide in relation to their thickness, while a relatively
high value would mean that they are narrow in relation to their thickness.
314

The Scandinavian Flint Daggers


Province
Gotland
land
Uppland
Blekinge
Smland
Halland
stergtland
Bohusln
Skne
Vstergtland
Nrke
Norrland
Sdermanland
Dalsland
Vrmland
Vstmanland

Average Std. dev.


155
152
151
151
148
147
147
146
145
145
145
144
143
139
138
137

42
32
33
31
47
36
37
43
39
42
34
29
36
30
27
49

Table 9:3. The average length (mm)


and the standard deviation of the
lengths of flint-dagger populations in
the Swedish provinces.

Province
Skne
Uppland
Gotland
Vstergtland
Blekinge
stergtland
Smland
Vrmland
Nrke
land
Sdermanland
Halland
Dalsland
Bohusln

Average Std. dev.


0,37
0,34
0,36
0,36
0,36
0,35
0,36
0,35
0,34
0,34
0,34
0,33
0,33
0,32

0,17
0,11
0,14
0,13
0,11
0,11
0,11
0,12
0,11
0,15
0,10
0,12
0,09
0,10

Table 9:4. The average thicknesswidth relationship and the standard


deviation of the thickness-width
relationships of the blades of the
Swedish flint daggers.

The mean values and the standard deviation in the distribution of


thickness-width values in the different provinces do not display any
variation that can be interpreted as a function of the distance from the
production areas (Table 9:4). This means that daggers were re-sharpened
as much in Skne as they were in the northern Mlaren valley (the
provinces of Uppland and Vstmanland).
I now turn to the qualitative classification of the degree of resharpening presented above. The data presented in Tables 9:5 and 9:6
demonstrate that a qualitative difference can be discerned between the
re-sharpening done in southern Scandinavia and the re-sharpening done
in eastern central Sweden. These differences are further illustrated in
Fig. 9:22. It is apparent that the greater part of the flint daggers in the
production areas, in this case Jutland and Skne, were re-sharpened by
persons who had mastered this craft. For instance, these qualitative
differences indicate that a smaller part of the dagger-owning population
in eastern central Sweden had the opportunity to have their daggers resharpened by a knowledgeable craftsman, compared to southern
Scandinavia, where 60 to 80 % had this opportunity. However, it is
striking that 34 % of the daggers from eastern central Sweden were
new or had been re-sharpened by an artisan. Either this means that the
315

Sdermanland

316

Uppland

26

0-

27

30

0-

0%

10%

0%

23

20%

10%

0-

30%

20%

22

40%

30%

40%

19

26

22

18

0-

27

23

19

30

0-

0-

0-

0%

0-

10%

0%

18

20%

10%

30%

20%

15

40%

30%

0-

40%

14

26

22

18

14

0-

27

23

19

15

30

0-

0-

0-

0-

0%

10%

0%

11

20%

10%

0-

30%

20%

10

40%

30%

40%

11

26

22

18

14

10

0-

27

23

19

15

11

30

0-

0-

0-

0-

0-

0%

0-

079

10%

0%

15

10

0-

27

23

19

15

30

0-

0-

0-

0-

11

20%

10%

0-

Skne
079

26

22

18

14

0-

30%

20%

0-

27

23

19

15

30

0-

0-

0-

0-

11

40%

30%

11

Halland

079

26

22

18

14

0-

10

079

40%

0-

0-

27

23

19

15

30

0-

0-

0-

0-

11

Dalsland

079

26

22

18

14

0-

10

079

Blekinge

14

0-

27

23

19

15

11

30

0-

0-

0-

0-

0-

10

079

30

0-

-2
79

-2
39

-1
99

-1
59

-1
19

26
0

22
0

18
0

14
0

10
0

079

79
10 -99
01
12 1 9
01
14 3 9
015
16 9
01
18 7 9
01
20 99
02
22 19
02
24 39
02
26 59
02
28 79
029
9
30
0-

0-

80

40%
35%
30%
25%
20%
15%
10%
5%
0%

10

26

22

18

14

10

079

Jan Apel

40%

30%

20%

10%
0%

Bohusln

Gotland

Nrke

Smland
4
3
3
2
2
1
1

Vstmanland

Vstergtland
40%
40%

30%
30%

20%
20%

10%
10%

0%
0%

200

150

100

50

Fig. 9:21. The blade-width frequencies of the Swedish flint daggers.

317

26
0

22
0

18
0

14
0

0-

30

0-

-2
79

-2
39

-1
99

-1
59

-1
19

079

30

-2
79

-2
39

-1
99

-1
59

-1
19

0%

10
0

10
0

26
0

22
0

18
0

14
0

10
0

10%

88

0-

27

23

19

15

11

30

0-

0-

0-

0-

0-

079

20%

079
80
10 -99
01
12 19
01
14 39
01
16 59
017
18 9
01
20 99
02
22 19
02
24 39
025
26 9
02
28 79
029
9
30
0-

26

22

18

14

10

079

30%

81

079
80
10 -99
01
12 19
013
14 9
01
16 59
017
18 9
01
20 99
021
22 9
02
24 39
025
26 9
02
28 79
029
9
30
0-

40%

70

63

56

49

42

35

28

21

14

The Scandinavian Flint Daggers

35%
30%
25%
20%
15%
10%
5%
0%

Vrmland

land

40%

30%

20%

10%

0%

stergtland

Fig. 9:20. The length distribution of the flint-dagger populations of the Swedish province.

Jan Apel
Provinces
Blekinge
%
Bohusln
%
Norrland
%
Dalsland
%
Gotland
%
Halland
%
Nrke
%
Smland
%
Sdermanland
%
Uppland
%
Vstergtland
%
Vrmland
%
stergtland
%
land
%
Skne
%
Total

Cat. 1
7
8%
48
13%
3
14%
29
8%
8
10%
60
13%
5
6%
10
7%
3
4%
4
5%
74
13%
2
6%
14
8%
8
15%
184
16%
459

Cat. 2
37
40%
151
41%
6
29%
128
37%
27
34%
186
41%
30
34%
65
46%
23
32%
20
25%
146
26%
7
22%
92
53%
26
47%
522
44%
1466

Cat. 3
47
51%
151
41%
11
52%
157
46%
41
51%
187
42%
46
53%
63
45%
42
58%
51
63%
276
50%
20
63%
64
37%
19
35%
437
37%
1612

Cat. 4
1
1%
15
4%
1
5%
28
8%
4
5%
16
4%
6
7%
3
2%
5
7%
6
7%
61
11%
3
9%
3
2%
2
4%
33
3%
172

Total
92
365
21
342
80
449
87
141
73
81
557
32
173
55
1176
3709

Table 9:5. The Swedish flint daggers divided into the four re-sharpening
categories defined in Chapter 8.
Regions
Eastern central Sweden
%
Western central Sweden
%
South-eastern Sweden
%
Skne
%
Halland
%
Norrland
%
Western Norway
%
Eastern Norway
%
Jutland (Moesgrd)

318

1
14
5%
151
12%
47
9%
184
16%
60
13%
3
14%
153
16%
40
17%
16
20%

2
80
29%
425
34%
247
46%
522
44%
186
41%
6
29%
295
31%
64
27%
47
59%

3
159
58%
584
47%
234
43%
437
37%
187
42%
11
52%
451
48%
131
55%
17
21%

4 Total
20 273
7%
89 1249
7%
13 541
2%
33 1176
3%
16 449
4%
1
21
5%
40 939
4%
4
239
2%
0
80
0%

Table 9:6. The


Scandinavian flint daggers
divided into the four resharpening categories
defined in Chapter 8.

The Scandinavian Flint Daggers

nd
Sw
ed
en
c

rla

ay

N
or

N
or

Sw
c

ed

en

ay
w

en

.N
or

ed

en
e

Sw

S-

Sw
ed

Sk
n
e

Sw

Ju
tla
nd

90
80
70
60
50
40
30
20
10
0

Fig. 9:22. Curve showing the relative frequencies of sharpening categories 1


and 2 in different regions in Scandinavia.

owners of these daggers could afford to send them off to be re-sharpened


or that they could afford to bring an artisan up to eastern central Sweden
to re-sharpen daggers. The conclusion of this is that only the lite in
eastern central Sweden could afford to wear daggers with blades of good
quality. In the production area, we must assume that other types of
foreign and prestigious objects functioned in the same way as the daggers
did in eastern central Sweden and other areas far away from the
production areas.
It is also interesting to see that a larger part of the daggers from
northern Sweden (43 %) than those from eastern central Sweden have
been re-sharpened by artisans. As I see it, this further indicates that the
Norwegian daggers reached the northern parts of Norrland by the way
of Trndelag and the Norwegian west coast and not by the way of
eastern central Sweden.

The character of the distribution


The distribution of flint daggers in Scandinavia does not appear to have
been the result of down-the-line exchange. As we have seen, the length
distributions of the materials from each Swedish province, as well as
those from Norway and the Netherlands, suggest that all take on a
more or less bell-shaped curve, culminating somewhere around 140/
150 mm. The average length does not decline as a function of the distance
from the production area. Neither does the degree of re-sharpening, at
least when measured, independently, as the thickness-width ratio of the
319

Jan Apel
blades. In fact, the distribution of the width of the Swedish dagger
blades, in all Swedish parishes, also takes on a bell-shaped appearance.
These conclusions are supported by the fact that there is no fall-off
tendency in the standard deviations of either length or thickness-width
ratio. In these respects, the distribution of flint daggers is contrasted by
the distribution pattern of the thin-butted axes in the same area during
the Early Neolithic period (Sundstrm 1992). These differences reflect
serious changes in the scale of the political landscape from that of the
Early Neolithic TRB communities to that of the Late Neolithic
communities in the same area. In the latter case, the lack of evidence for
a down-the-line exchange may indicate that there were local centres,
based on hierarchical differentiation, conducting directional exchange
with other local centres. Bergljot Solberg (1994) has indeed presented
convincing evidence for directional exchange between the Thy district
in Jutland and central places, such as those at Jren and North Karmy,
on the Norwegian west coast during the Late Neolithic and the Early
Bronze Age. This conclusion is supported by my own observations in
Norwegian museum collections: the dagger materials from the
Norwegian west coast are quite homogeneous, as regards size and quality.
In Sweden studies have been made that indicate the presence of
centralised places of local or regional authority. A study of the Late
Neolithic and the Early Bronze Age in Vstergtland certainly suggests
hierarchical differentiation on the local level, among other things, as
reflected in the different qualities of grave goods in the numerous gallery
graves of this region (Weiler 1994:54 ff.). Investigations of the centralised
versus local production of different qualities of Late Neolithic, shafthole axes, utilized in different social contexts, also point to hierarchical
differentiation on the local level during the Late Neolithic (Lekberg
1997, 2000a, b & c). Studies conducted in the Mre region in the
province of Smland also suggest a political landscape that was organised
according to hierarchical principles. Here, the utilization of prestige
objects made of flint, in the Late Neolithic, and metal, in the Early
Bronze Age, indicates that similar structural principles were prevalent
in this period (Gurstad-Nilsson 1999:191). A spatial analysis of Late
Neolithic artefacts from two areas in the Mre region suggests that
there was an unequal distribution of valuable objects in the landscape,
indicating regional hierarchies (Hagberg 1999). This study is actually

320

The Scandinavian Flint Daggers


based on the classification systems of shaft-hole axes presented by
Lekberg (1997, 2000a, b & c) and flint daggers presented by Weiler
(1994). Incidentally, there are concentrations of flint daggers in the
southern parts of Vstergtland, as well as in the Mre region (see
distribution maps presented in Figs. 9:2-9:11). Future research will no
doubt uncover further regional differences that can be interpreted as
mirroring hierarchical differentiation in other parts of Scandinavia as
well.

321

Chapter 10

The Formation, Reproduction and


Reformation

This chapter starts with a brief rsum of the previous chapters. Further
on, a more detailed interpretation of the agrarian Late Neolithic societies
in Scandinavia, based on the conclusions of the previous chapters, is
presented. An interpretation of the social setting in which the flintdagger technology was embedded will be presented and the disappearance
of the technology will be discussed in social terms.

Rsum
The aim of this dissertation was to investigate how far the organisation
of a traditional technology corresponds to the degree of social complexity
in a sedentary, agrarian society. Thus, I argue that the social structure of
the Late Neolithic societies in Scandinavia was more complex than has
previously been thought. An examination of the production of flint
daggers indicates the presence of formal apprenticeship systems based
on corporate descent groups. The selection of apprentices was based on
hereditary principles. In other words, the craft appears to have been
institutionalised, and this suggests a fairly complex, social structure.
This conclusion is related to the theory of symbolic assets presented by
Bourdieu and to the idea of the Aggrandizer presented by Hayden.
Accordingly, the main social function of the Late Neolithic, descent
groups in the southern and central parts of Scandinavia was to reproduce
their access to collective assets throughout the generations by following
strict, hereditary principles. In reaching these conclusions, I have concentrated on two main subjects:
(1) The ways in which the division of labour and the use of prestige
technologies were related to the social complexity in kin-ordered or
segmentary societies in general. This discussion includes theoretical issues,
323

Jan Apel
as well as the questions put and the observations made during controlled,
flint-working experiments.
(2) The role that the production of elaborate flint tools played in
the transformation of segmentary tribal societies into chiefdoms that
allegedly occurred during the Late Neolithic and Early Bronze Age in
Scandinavia. For this purpose, the contexts in which Late Neolithic,
Scandinavian, flint daggers were produced and consumed have been
investigated.
Technology is here defined as a coherent system of artefacts, behaviours and knowledge that can be handed down from one generation to
another. Since this definition concentrates on the reproduction of
technology, it recognises that technology is integrated with other aspects
of society. A holistic, supra-individualistic perspective is regarded as
especially important in this context, since flint daggers were made over
a period of c. 800 years.
My understanding of the flint technology is based on my collaboration with an experienced flint-knapper and archaeologist, Errett
Callahan. During the course of two summers, at the Historical Research
Centre in Lejre, Denmark, and at Uppsala University in Sweden, several
production stages were defined as the result of a dialectic research strategy
in which archaeological evidences and the experimental findings were
confronted and compared (Callahan & Apel, MS). This resulted in
some understanding of the degree of knowledge, skill and time that
had to be achieved in order to make prestigious flint daggers.
After being defined, the flint-dagger technology was subjected to an
operational-chain analysis. This method is rooted in Durkheimian
sociology and, consequently, gestures, even those that seem to reflect
the biological aspects of humanity, are regarded as social phenomena
that are learned in social contexts. This is an important observation,
since the implication is that the study of gestures concerns subjects such
as anthropology, archaeology and sociology. Two important concepts
form the basis of my investigation of the social aspects of the flintdagger technology: (1) knowledge (connaissance) and (2) know-how
(savoir-faire). These terms define two fundamental elements of a
distinctly, neuro-psychological nature that are involved in the execution
of all gestures. Knowledge has an explicit and declarative character and
can be communicated to others; it can be passed from teacher to pupil
by word of mouth, signs or written language. Know-how is an
324

The Formation, Reproduction and Reformation


unconscious memory that springs from practical experience only. It is
intuitive, connected with body movements and can only be learned by
practical repetition.
During the experimental work, the gestures involved in each of the
defined, dagger-production stages were graded according to their relative
degrees of knowledge and know-how. Some stages, like the early ones
in which raw material was collected and when crude preforms were
shaped, were based on simple knowledge, in the form of recipes for
action, and a very low degree of know-how. Other stages, like those in
which the dagger blade was formed and thinned down by percussion
with a deer- or elk-antler billet, demanded a fair proportion of knowledge
and an enormous amount of know-how. It takes years of practical training
to co-ordinate muscles and applied force in a way that secures the desired
outcome.
This suggests that the craftsmanship involved in the production of
flint daggers was handed down the generations by a form of apprenticeship system based on hereditary principles, i.e. that the trade was inherited
within clans or lineage groups. The logic behind this reasoning is twofold.
First, in such a system, the time needed to transmit the skills in order to
be able to produce flint daggers through the generations made the simple
principle of kinship the most convenient mechanism for recruitment.
Second, flint and manufacturing skills were valuable assets that
stimulated some form of limited access and thus regulations of group
membership. Accordingly, fixed social institutions must have been
present in the Late Neolithic communities of Scandinavia. This
interpretation is, in turn, related to the complexity debate in archaeology
and anthropology, in which formal institutions are important indications
of a fairly high level of social complexity. The flint-dagger technology
therefore entailed a highly developed craftsmanship, and flint daggers
were distributed over large areas of northern and central Europe in
exchange networks that were controlled either by corporate groups or
by regional and local lites. This interpretation is also related to an explicit theory of practice suggested by Pierre Bourdieu. Consequently,
terms such as symbolic capital and habitus have been used to give social
meaning to the reproduction of the technology and its role, actively
and metaphorically, in the reproduction of the Scandinavian, Late
Neolithic communities in general.
The theoretical concepts, formulated in Chapters 2-4, are put into
325

Jan Apel
play in the last part of Chapter 4 in the form of an ethno-archaeological
example (traditional iron-working in Kenya). Here, strict hereditary
principles decide who will have the opportunity of becoming a smith
and who will not. It is suggested that those stages of iron-working that
are based mostly on knowledge (recipes for action) and do not demand
a great deal of know-how, such as smelting skills, are surrounded by
taboos and are conducted at secluded locations. However, stages that
require a high degree of know-how, such as forging, are conducted in
public, often at a central place of the hamlet. The point is made that the
recipes for action in iron-working are protected by the smiths at the
same time as they actively exhibit their practical know-how. The practical
know-how is exclusive, because it takes years to master, and thereby
gives the smiths both power and prestige when demonstrated in public.
Chapter 5 is concerned with a research history aimed at understanding
how flint debitage from bifacial-reduction sequences can be defined.
This work is needed in order to proceed with Chapters 6 and 7, in
which the flint debitage from Late Neolithic sites in eastern central
Sweden and Jutland is investigated and discussed.
In Chapter 6, the flint industry of southern Scandinavia is investigated. Several sites with evidence of flint-dagger production, mainly in
Jutland, form the basis of the investigation. These sites include
settlements, extraction sites and mines. It was concluded that the
different stages in the production of daggers took place in different
locations. The extraction sites of each production fraternity mines
and certain beaches were, for natural reasons, kept secret and were
guarded by taboos and esoteric regulations. No one from outside the
production fraternity was allowed to participate in the early production
stages, since this would threaten the exclusiveness of the technology.
The young, inexperienced apprentices collected suitable nodules for
dagger-making. It was also their task, together with the journeymen, to
make the early-stage preforms. This activity took place at secluded spots
near the raw-material sources and was also surrounded by esoteric
regulations. Thus, each corporate lineage had its own outcrop. The Late
Neolithic flint mines of Thisted and Himmerland are examples of such
raw-material sources. Places for the extraction of suitable raw material
have also been found, for instance, at the beaches of Fornaes and
Drenges. The flint debitage on these sites often displays obvious
knapping errors, and the sites are littered with broken preforms.
326

The Formation, Reproduction and Reformation


The final and most difficult stages, from the practical point of view,
were carried out by the master knappers on full public display in the
hamlets. In this way, the performance of these stages functioned as a
vehicle for the dominant social structure of the society, i.e. these stages
involved a high proportion of repetitive aspects and thereby served to
maintain a strong, past-in-the-present consciousness among the
population. That is, the exclusiveness of the technology was actively
used to bring forward an authoritative, conservative, social message.
Such sites have been found, for instance, at Gug and at Myrhj in
Himmerland.
In Chapter 7, the Late Neolithic, flint industries of eastern central
Sweden were investigated as an example of a consumption area with no
natural flint sources. Here, flint was extremely valuable. Large artefacts
in flint, such as daggers, sickles, spoon-shaped scrapers and axes were
imported from flint-rich areas. The flint industries of these sites were
directed to the re-sharpening of large, imported, edged tools but also to
the production of bifacial arrowheads. Thin, disc-shaped flakes were
used as blanks and they were either imported or made locally. The ideal
blanks for the production of bifacial arrowheads were bifacial-thinning
flakes, created during the later stages of dagger-making (the display
stages), and such flakes were consequently imported into eastern central
Sweden, together with large artefacts. The imported and locally produced blanks were then shaped into arrowheads by a simple, pressure
technique. Disc-shaped flakes of a poorer quality were produced locally
by simpler methods. The evidence, in the form of cores and flakes both
on settlements and in gallery graves, suggests that one important local
way of producing blanks for bifacial arrowheads was the bipolar method.
It is apparent that the imported, disc-shaped flakes also had a great
symbolic value. This value is reflected by the fact that they are found in
gallery graves in the southern and central parts of Sweden. They are also
found in hoards and votive offerings in Norway, in southern Sweden and
on the Swedish west coast. In this context, it is intriguing to consider that
it is the flakes from the difficult and hence public stages of dagger
production (see Chapters 2 and 4) that have a great symbolic value,
especially in areas quite far away from where the daggers were produced.
In other words, it is apparent that these flakes, as well as the flint daggers,
carried a strong and normative, past in the present message to the Late
Neolithic population of the agrarian communities of Scandinavia as well.
327

Jan Apel
The flint-working that was practised in the Late Neolithic settlements of eastern central Sweden was dictated by strict conventions
regarding the form and the choice of southern-Scandinavian flint as the
raw material. This resulted from the fact that the area could not maintain
the level of craftsmanship present in areas immediately to the north
where bifacial arrowheads were made in local raw materials at the same
time. From the purely practical point of view, it makes no sense to use
only flint as raw material in eastern central Sweden when there are local
raw materials, such as hlleflint , available. Consequently, we may expect
that this pattern will be explained by the fact that artefacts of southernScandinavian types and flint had a great symbolic value for the agrarian
population of eastern central Sweden.
In Chapters 8 and 9, aspects of the Scandinavian flint daggers are
investigated. It is concluded here that several of the different dagger
types and subtypes that have been defined by earlier scholars can be
regarded as reflecting regional, or even local, workshops. These regional
production areas distributed flint daggers to different consumption areas.
The general distribution of the Scandinavian flint daggers covered a
large region in northern Europe from the Rhine delta in the west to
Pomerania in the east, and from central Germany in the south to the
Norwegian north-west coast in the north. Outside this area, a small
number of daggers have been recovered in the British Isles, in Bohemia
and Moravia, in Austria, in Finland and the Baltic countries and in the
northern parts of Sweden.
Finally, an investigation of the distribution of flint daggers in
Scandinavia reveals that they were distributed by direct exchange.
Exchange systems of this kind require redistribution centres on the local
or regional level. This presupposes hierarchical settlement patterns on
the local and regional level and is an indication of chiefdoms. In this
context, it is appropriate to mention that institutionalised heredity, such
as has been proposed for the dagger production, is another indication
of a chiefdom level of society. It can thus be concluded that there are
strong archaeological arguments in favour of regarding the Late Neolithic
societies as having had a social structure at the chiefdom level (as defined
in Chapter 4).

328

The Formation, Reproduction and Reformation

The social structure of the Late Neolithic


In recent years most scholars who have discussed the social structure of
the Neolithic and the Early Bronze Age (Prescott & Walderhaug 1995,
Vandkilde 1996, Gurstad-Nilsson 1999, Goldhahn 1999) have related
their ideas to Kristian Kristiansens model of the formation of tribal
systems in later, northern European prehistory (Kristiansen 1982 &
1987). This model assumes that there was a fluctuating, uneven
development towards the full-scale, Bronze-Age chiefdoms of south
Scandinavia during the whole Neolithic period.
During the Early Neolithic (4000-3400 cal BC), a rapid population
expansion led to large-scale forest clearings. During this period, there is
evidence of a settlement stabilisation and ritual manifestations in the
form of earthen long barrows and the prestige exchange of long, polished
axes in flint (Kristiansen 1982:257 f.).
During the early part of the Middle Neolithic (3400-3200 cal BC),
the social system of the Early Neolithic culminated in the building of
thousands of megaliths during a short period. In line with Colin
Renfrews interpretation of Neolithic Wessex (1973), Kristiansen
suggests that southern Scandinavian society during this period was
organised as a chiefdom: At the local level megaliths served common
functions as ritual central places for ancestor worship and burial places
of chiefly lineages (Kristiansen 1982:259).
During the Middle Neolithic period (MN A and B, 3200-2350 cal
BC) these chiefdoms gradually disintegrated. The importance of
cultivation declined and a more mobile pastoral economy was preferred.
In eastern central Sweden, Pitted-Ware-culture sites suggest that huntergatherer strategies were preferred during the greater part of the Middle
Neolithic period. From LN II onwards, the proposed development is
presented in Fig. 10:1.
According to Charlotte Damm (1991:15), because of his interest in
Marxist theories concerning the reproduction and transformation of
society, Kristiansen has discussed the change and development in social
organisation rather than the relationship between material culture and
social system. While Kristiansen regards the appearance of the MN B,
Single-Grave Culture as the result of a migration from the south, Damm
considers this material expression as being the result of an active choice.
Vandkilde interprets the emergence of the Bell-Beaker tradition in

329

Jan Apel
1900 B.C.

Small hamlets of long-houses. Metal trade in simple tools

Establishing long-distance

Clans

alliances

1500 B.C.

Chiefly barrows Extended

Chiefly control of metalwork. New international status

exploitation and control of

Small chiefly retinue.

ideology. Rise of chiefly elites.

First crisis in metal supplies

Chiefly consolidation and

land
1100 B.C.

Intensification of production.
Settlement clustering.

800 B.C.

500 B.C.

competition

Increasing degradation of

Legitimation crisis. Increased Regional chiefs and settlement

land. Decline of production.

investment in ritual.

hierarchy.

Larger hamlets of small

Revolts and reorganization.

Collapse of international trade

houses.

Local migrations.

& chiefly superstructures.

Fig. 10:1. Development and transformation of the southern Scandinavian societies from the
beginning of LN II to the end of the Bronze Age (after Kristiansen 1991:32, Fig. 2:4).

northern Jutland from a similar perspective (1998). She maintains that


material culture can be regarded as reflecting different temporal structures.
Thus, she explains the general, Corded-Ware/Bell-Beaker phenomena
as a form of large-scale longue dure and argues that regional groups of
people will express local and regional identities by varying certain aspects
of their material culture. Olausson regards the Early Bronze Age society
as the result of a competition between two separate systems during the
Late Neolithic: one which used flint daggers as prestige items and another
that used metal items, mainly bronze daggers, as prestige gifts. The old
lite tried to maintain and expand their power during LN II and the
Early Bronze Age by increased specialisation and control of the
production of bifacial flint items. The new lite was based on control
over the influx of metal and operated by controlling the recipes for
metalworking (Olausson 2000:132). According to Olausson, the
production of daggers was attempt to gain social control through a
prestige technology, in a similar way to that of the Early Neolithic
production of thin-butted axes. However, she argues that this attempt
failed since it was impossible to keep the technology exclusive (Olausson
2000:128). I agree with Olausson that peoples outside of the production
fraternities may have made simpler flint daggers. However, I maintain
that such daggers were easy to recognise and were valued accordingly.
As we have seen in Chapter 9, the thin-butted axes and the flint daggers
were exchanged in networks of completely different natures.

330

The Formation, Reproduction and Reformation

The formation
In this context, the formation of the dagger technology is related to the
anti-intellectualistic theory of culture change discussed in Chapter 4.
The main proposition, in line with the ideas presented by Durkheim
and Binford, was that technological specialisation and social complexity
may be the results of an increased moral density rather than of the
conscious decisions made by individual agents. One way of detecting
moral density in prehistory is to find arguments for an increased
population density creating a population pressure. Thus, it may be
possible to consider the appearance of the flint-dagger technology as
the result of a packing problem created when domestic units were
marginalized from rich areas and were thereby forced to develop new
technologies in order to maintain their traditional social structure.
What would be the archaeological consequences of a packing problem of the kind that Durkheim and Binford imagined? One of the
characteristic features of the Late Neolithic period in Scandinavia is
that there are signs of an increased population density. There is evidence
of a massive relocation of domestic units to remote areas where
previously there were no signs of Neolithic populations. This process
manifests itself, perhaps most clearly, in areas such as the provinces of
Vrmland, Dalsland and Smland in Sweden where gallery graves were
erected and Neolithic technologies, including cultivation, are introduced
during this period (Johansson 1961, Heimann 2000, Lagers 2000).
In eastern central Sweden settlements tended to be localized to the heavy
clay soils, rather than, as during earlier periods, to glacial eskers and
other well-drained soils. This means that, in this respect, the Late
Neolithic settlement pattern is similar to that of the Bronze and Early
Iron Ages. It is, in fact, more than likely that the Late Neolithic
settlements in these areas are hidden in the archaeological remains of
later periods. In areas with rich archaeological remains from later periods,
such as eastern central Sweden, Late Neolithic graves and settlements
tend to be covered by later settlements and graves (Holm et al. 1997:253,
Andersson 2001:81). The gallery graves of Alby, Annelund and Dragby,
for instance, have all been covered in later times. It has been suggested
that the Bronze Age population of eastern central Sweden, by covering
the earlier collective graves, buried an old, collective ideology that was
seen as a threat to the social values of the stratified, Bronze Age society

331

Jan Apel
(Andersson 2001:82). However, the forested areas of Smland and
Vrmland, which have few traces of agricultural practices after the Late
Neolithic/Bronze Age transition, can still be regarded as fossilised
remains of the Late Neolithic, cultural landscape, a landscape that was
never covered by Bronze Age graves and settlements. The tendency to
exploit the heavier clay soils during the Late Neolithic has, in earlier
research, been related to the land-upheaval process in eastern central
Sweden during the Neolithic. The main argument was that valleys with
clay soils were continuously exposed as the sea withdrew and that people
moved in because of the increased population pressure. However, the
tendency to use clay soils was not confined only to areas where landupheaval processes had changed the landscape. A similar tendency can
be discerned in Denmark as well (Douglas Price, personal
communication). Thus, it can be concluded that the choices of where
to live and where to bury the dead are cognitive aspects that the Late
Neolithic population of the southern and central parts of Scandinavia
shared with the peoples of later times.
As we have seen, there are archaeological signs indicating an increased
population density during the Late Neolithic period in the southern
and central parts of Scandinavia and this may have led to the exploitation
of new environments. But, in order to argue for an anti-intellectualistic
explanation of the emergence of new technologies related to craft
specialisation that is implied by the flint-dagger technology, we need to
gather more evidence. We need to investigate the area of the first
development of the Scandinavian dagger technology, i.e. the situation
in Jutland prior to the introduction of the flint dagger at the transition
between the MN B and the LN I around 2350 cal BC.
The Middle Neolithic period in southern Scandinavia is divided
into two distinct phases; MN A and B. In this context, we shall be
concerned only with the latter phase. In Jutland, MN B is defined by
the appearance of a new type of grave. The collective chamber-tombs
of the MN-A phase were replaced by individual or occasionally
double inhumation graves in small mounds. These graves make up
the majority of the archaeological remains from the Single-Grave
Culture of Jutland. The tripartition of MN B into the three sub-phases,
I, II and III, originates from Sophus Mllers definitions of the three
grave types commonly found in these mounds: lower, middle and upper
graves (Mller 1889, Damm 1991:83). Since these graves corresponded
332

The Formation, Reproduction and Reformation

Fig. 10:2. The geographical distribution of battle axes from the early phase of
MN B (a) and battle axes from the latest phase of MN B (b). The distribution
indicates the primary habitation area of the Single-Grave Culture during the
earliest and latest phases of MN B respectively (from Struve 1955).

to a stratigraphical sequence with chronological significance, the SingleGrave period was later divided into a lower, middle and upper subphase (Glob 1944) and this division corresponds to the MN B I, II and
III phases that are used today (Damm 1991:83). However, it has been
suggested that it would correspond better to the actual archaeological
material to divide MN B into an earlier and a later sub-phase (Damm
1991:81). The transition between MN B and LN I is set to the times
when daggers replaced battle axes in stone as the prestige gift in the
male single graves. Thus, it was within the context of the late SingleGrave-Culture in northern Jutland that the production of flint daggers
commenced at the transition between MN B and LN I.
The geographical distribution of the early phase of the Single-Grave
culture is illustrated in Fig. 10:2a. The grave mounds were situated in
groups of 10 to 20 along streams in the central part of Jutland. The
distribution of Single-Grave-Culture graves from the early MN B phase
333

Jan Apel
coincides with that of Jutlands most fertile soils light diluvial sand,
moraine clay or sandy moraine (Damm 1991:81). These kinds of soil
are suitable for crop agriculture with a wooden ard and for grazing as
well. Excavations of Single Grave sites on Jutland, such as those at
Hemmed on Djursland (Boas 1993) and Vorbasse near Esbjerg (Hvass
1977, 1986:327 ff.), have demonstrated that long houses no less
elaborate, as regarded construction details, than the houses of the FunnelBeaker Culture in Denmark and Sweden were built. The battle axes
from the early phase of MN B which reflect the distribution of
inhumation graves are concentrated to the central parts of Jutland
(Fig. 10:2a).
During the course of MN B the material manifestations of the SingleGraves culture expands to the north as well as to the Danish islands in
the east. This tendency is best illustrated by the distribution of the later
battle-axe types found in the upper graves (Fig. 10:2b). Thus, during
the final Middle Neolithic phase, the geographical distribution of battleaxes also included the provinces in the Limfjord area, such as
Himmerland and Thisted, and it was here that the first daggers were
produced somewhere around 2350 cal BC. It is also in this area that the
majority of the Danish, Late Neolithic settlements have been recovered
(Bjrhem & Sfvestad 1989:77, Fig. 73). However, this was apparently
not because the Limfjord area had a high population density at this
time (Damm 1991, Vandkilde 1996). The reason for the high frequency
of Late Neolithic settlements here is in my opinion, only because of
their archaeological visibility. That is to say, the flint industry of these
sites, including broken dagger preforms and thousands of easily
recognisable, flake types from bifacial-reduction sequences, tends to
amplify them and make them easy to find, as well as to date. While
Late Neolithic sites in other areas of Scandinavia have a tendency to get
lost in the archaeological remains of later periods, the sites near the
natural flint sources, are simply easier to spot. In fact, Jensens calculations
of the population density of Denmark, based mainly on grave
frequencies, clearly suggests that central Jutland was the most populated
area of Jutland during the Bronze Age (Jensen 1979). Thus, we may
assume that central Jutland had the highest population density in this
region during the Late Neolithic as well. It is therefore not unlikely
that the highly visible sites in the Limfjord area are the remains of a
population that was marginalised from central Jutland in the course of
334

The Formation, Reproduction and Reformation


MN B, owing to population pressure. These domestic units eventually
formed new clans and lineages and took advantage of the high-quality,
Limfjord flint in the production of prestige goods in order to maintain
the social structure to which they were accustomed. Since the term
social structure is an abstraction, it may be appropriate to specify what
kind of social practises and institutions that the marginalised population
of northern Jutland had if choosing not to specialise. The chances of
gaining continuous access to good, traditional food (now obtained
through the exchange networks rather than produced locally), prestigious objects and nice houses, to which one was accustomed, may have
triggered the specialisation. It is possible that the opportunity to maintain
a traditional, sedentary way of life, through specialisation, was more
tempting than the prospect of having to become more flexible as regards
subsistence strategies. The opportunity to commence costly religious
rituals in the traditional way might also have contributed to the
introduction of the flint dagger technology.
Since there was a demand for prestige objects that could be exchanged
between the lite fractions of northern Europe, the production of lancetshaped, flint daggers was an immediate success. The increased
specialisation was triggered by an attempt to maintain a certain social
structure. Ironically, the specialisation, in itself, led to the formation of
a new social order that was different from the old one.
During the early part of the Late Neolithic, there was a great deal of
variation in the grave types in the Limfjord area, especially in the province
of Himmerland where the earliest dagger-production sites are located
(Vandkilde 1996). This may have been connected with a turbulent social
situation in which new lineages were formed and people felt the need
to express their identities through grave types. It is interesting to consider
that there are no obvious prestige items made of flint to be found in the
graves of the preceding Danish, Single-Grave culture. In fact, the mining
of flint in the Limfjord area, in both Thisted and Himmerland, is dated
to the Early and Late Neolithic only. There is no evidence for mining
during the Middle Neolithic period (Becker 1993). We know that large,
prestigious, thin-butted axes in flint were produced in northern Jutland
during the Early Neolithic and that prestigious daggers were produced
during the Late Neolithic. However, no prestigious type of flint artefact
was produced here during the Middle Neolithic. Thus, it may be
assumed that mining during the Neolithic was intimately connected
335

Jan Apel
with the organised production of prestige goods. The connection between
outcrops and the mining of lithic materials, on the one hand, and the
production of prestige items, on the other, is a general pattern in the
greater part of Europe during the entire Neolithic period (see H.
Knutsson, in prep.).

The reproduction
The Middle Neolithic, Bell-Beaker communities of Europe were familiar with metallurgy, wore woollen clothes, practised horse-riding
and held ceremonies, including feasting and the drinking of alcohol
beverages. This is a context in which an idealised, male-warrior role was
formed and projected (Sherratt 1987:88, Vandkilde 1998). The copper
and flint daggers that appear during this period in large parts of Europe
form part of the symbolic language that was related to this new, warrior
identity. From now on, up to the Late Bronze Age, this new identity
characterised large areas of central and northern Europe. In Euro-Asia,
the dagger is traditionally thought of as a male weapon and symbol. In
the earliest, European, written sources, the daggers are part of a warrior
outfit dated to the Bronze Age. In areas of the Near East, such as Yemen,
metal daggers formed an important part of a respectable mans wardrobe
up to the present (Cammann 1977).
In Scandinavian prehistory, flint daggers seem to replace the battleaxe of stone as the male prestige gift in the upper graves of the Jutlandic
Single-Grave Culture. These daggers were made under influences from
western and central Europe. There are ethnographic examples of how
the spectacular outfits worn by warrior Age-Sets came to form the ideal
of a larger region, including several other ethnic groups. Among the
pastoral Masaais of Kenya, for instance, males were not allowed to
marry until they had passed their thirties. Because of this, the large
group of unmarried young males formed a powerful warrior class that
raided neighbouring territories for cattle and were generally feared and
respected. Eventually, the respect that this outfit gave its bearers tempted
the younger men of neighbouring tribes to adopt certain, typical, Masaaiwarrior features, such as the typical head-gear with feathers and the long
spears, thus forming their own warrior age-classes (Hkansson, personal
communication, see Muriuki 1979:98 f., Spear 1997:239). It is
intriguing to consider that in this example the young males from different
ethnic groups shared a similar interest in representing a warrior ideal. It
336

The Formation, Reproduction and Reformation


is not unlikely that the appearance of the flint daggers in Scandinavia
mirrored a similar situation, when a male warrior ideal, formed in the
southern and central parts of Europe and in which the metal dagger
formed an important part, was taken up into new ethnic contexts.
It may be appropriate to present an ethnographic example to illustrate how the dagger has been used as a manhood and prestige symbol,
as well as reflecting regional identity. In the Yemen Arab republic, a
proportion of the adult males still wear the traditional Arabian dagger
(jambiya) (Cammann 1977:27). These daggers have curved, doubleedged, steel blades with handles usually made of horn. They are worn
in different positions that relate to the status of the owner.
Consequently, city commoners and tribesmen wear their daggers in one
way, while the daggers of men of high status are adjusted to the belt in
a slightly different fashion (Cammann 1977:31). In these communities,
such subtle material signals mirror the status differentiation among the
dagger-bearers that is objectively present in the social life. This
observation should be compared with Chapter 9, where it was concluded
that the distribution of daggers of different qualities may reflect
hierarchical differentiation on a local geographical level.
Striking regional differences, as regards those concerning other
weapons, types of headgear and ways of dressing, also applied to the
Yemeni daggers. Consequently, silver dagger-hilts, such as those on the
Hodeia-dagger type, are common on the Tihama plain, while another
type of metal hilt was common in the region surrounding Zabid. In
other districts, metal hilts were not in evidence. Hence, the daggers also
mirrored regional traditions (Cammann 1977:32). After the revolution
in 1962, when Yemen became a republic and old hereditary imams
were removed from power, the daggers seem to have been used only as
a mark of manhood, or as a kind of status symbol. However, in the
old days, the daggers were also used in combat or for defence.
Cammann suggests that, in the old traditional society, boys were allowed
to use these daggers after a coming-to-age initiation rite: There is ample
reason to suppose that in the past a Yemeni male would be presented
with a jambiya, or at least given the right to wear one, in the course of
some coming-of-age ceremony. In other Muslim lands, boys of twelve
to fourteen years traditionally underwent ritual circumcision, after they
were considered to have attained manhood and could assume the right
to wear a mans headgear and to carry a sword or rifle. Probably the
337

Jan Apel
right to wear a dagger was formally granted at the same time (Cammann
1977:33).
I have suggested that the dagger technology was complex and most
likely transmitted vertically within lineage groups or clans. It is likely
that the apprenticeship started at a fairly early age, since this would
simplify the learning process and thus make it easier to teach those who
were supposed to learn the craft independently of their natural abilities.
As we have seen, this interpretation also makes sense from the sociological
point of view. It would therefore have been convenient if the apprentice
was actually brought up with, or close to, the teacher. In the following
discussion, it is anticipated that the Scandinavian, Late Neolithic, flint
daggers were used as status insignia by an Age-Grade of warriors.
Among the traditional Kabyle farmers of Algeria, the rifle is a symbol
of nif (the personal qualities that define a respectable, honourable male).
The rifle functions as a form of symbolic embodiment of the nif within
the male agnatic group of these patrilinear farmers (Bourdieu 1977:61).
It is tempting to regard the flint dagger as functioning in similar fashion
as a symbol either of adult males or of an age-class of males in the Late
Neolithic communities of Scandinavia. This is actually very likely, since
the different qualities of bronze daggers and flint daggers differ radically,
thus making a purely functional interpretation highly unlikely (see
Chapter 9).
If the flint daggers were used as male prestige objects, it is not entirely
unlikely that the flint smiths who produced the flint daggers were males.
In the following discussion, it is thus anticipated that the flint-smiths
were males. It has been proposed that the flint-dagger technology was
most likely reproduced vertically within family units (see Chapters 24). It has also been suggested that the apprenticeship started at a fairly
early age, i.e. around puberty. However, even if the formal introduction
to the apprenticeship occurred at puberty, it would clearly be an
advantage if children intended to become apprentices were brought up
in an environment in which the craft was practised. In other words, it
would be a practical advantage if they were brought up among artisans
and master knappers. Such a scenario also makes sense if we consider
Haydens theory, in which the domestic unit tries to maintain and enhance
its status and power through child-growth strategies, including letting
children participate in costly education (see Chapter 4). Some form of
residence that maintained long contact between older males (masters)
338

The Formation, Reproduction and Reformation


and young males (apprentices) would have been in existence. Either
virilocal residence, where fathers transmitted skills to sons, or avunculocal
residence, where mothers brothers were the teachers, may have solved
this. Both matrilineal and patrilinear descent systems are connected with
virilocal residence in the ethnographic literature, so this condition does
not tell us anything about what type of unilineal descent system was
practised (Thomas Hkansson, personal comment). When discussing
matrimonial strategies as involved directly in the social reproduction of
the social order it is once again important to recognise the differences
between the ideal marriage (proclaimed by structure) the actual marriages
(practice) that do not always, and sometimes seldom, coincide.
According to Bourdieu (1977:59) it is simply not enough to characterise
marriage rules in geneological terms. One has to regard them as:
similar to a card game, in which the outcome depends partly on the
deal, the cards held (their value itself being defined by the rules of the
game, characteristic of the social formation in question), and partly on
the players skill: that is to say, firstly on the material and symbolic
capital possessed by the families concerned, their wealth in instruments
of production and in men, considered both as productive and
reproductive power and also, in previous state of play, as fighting strength
and hence symbolic strength; and secondly on the competence which
enables the strategists to make the best use of this capital, practical
mastery of the (in the widest sense) economic axiomatics being the
precondition for production of the practices regarded as reasonable
within the group and positively sanctioned by the laws of the market in
material and symbolic goods (Bourdieu 1977:58)

It is interesting to consider Sahlins (1968) remark that lineages that


are closely related to the mythical ancestor sometimes practise endogamy.
Presumably, such unnatural practices are solidifying the minimal
lineage as a corporate group in factional struggle as Fredrik Barth
suggested was the leading role of the parallel-cousin marriage in southern
Kurdistan (Barth 1953, Bourdieu 1977:32).
For instance, among the Kabyle farmers, the ideal marriage is between
parallel cousins. In such a case, ego marries one of his fathers brothers
children. It has been established that these marriages are especially
important under circumstances in which the lineage cohesion is
vulnerable and threatened by an internal power struggle that may arise,
339

Jan Apel
for instance, as the result of the death of the male elder. Such an event
will tend to provoke internal struggles between brothers or cousins, i.e.
the sons or grandsons of the elder. The more vulnerable a family, the
more nif it must possess to defend its sacred values. There are two reasons
why the organisation of a parallel-cousin marriage will maintain the
cohesion of the group. (1) The two brothers will, through the marriage
of their children, be closely united and thereby less inclined to engage in
a power struggle that would lead to a fraction within the lineage. (2)
The community as a whole will appreciate the occasion, since it honours
the tradition. This may be why the women in these communities are
less inclined to instigate parallel-cousin marriages; usually the elder male
within the lineage does it, often without consulting the women and
actually trying to arrange it over their heads. Usually, under circumstances in which the lineage is not threatened, women tend to actively
participate in matrimonial negotiations, since they are free to negotiate
without losing face. The matrimonial strategies used by women aim at
incorporating more women from their original groups through marriage,
since such strategies will increase their own social power (Bourdieu 1977).
This is actually a nice illustration of a continuous dialectic relationship
between social structure and practice. The parallel-cousin marriages are
necessary in order to maintain the cohesion of the lineage, while exogamy
is necessary, not only for purely biological reasons, but also because it
provides the necessary contacts used to construct exchange networks.
Since it is likely that the flint-dagger technology was organised
according to hereditary principles, i.e. that it was institutionalised, we
must also accept that other aspects of the society were also organised in
this fashion. What we are dealing with is a social structure that revolves
around a ranked and segmented, common-descent group.

The reformation
The dominant social structures of the Late Neolithic period were upheld
primarily by male elders and organised as unilineal descent groups. The
chiefly lineages had control over the most productive land and were,
through their close genealogical ties to the clan mythical ancestor and
through their external exchange contacts, distinguished from the general
populace. In marginal areas, lineage-groups specialised in the production
of prestige items in order to maintain their traditional way of life. These
prestige items circulated in an exchange network controlled by the male
340

The Formation, Reproduction and Reformation


elders. It should be mentioned that the women most probably also had
a large exchange network in which other types of goods were distributed.
In trobriands, for instance, a male exchange sphere was complemented
with a female exchange sphere.
The corporate descent groups that produced prestigious flint daggers may have been organised as a form of attached specialists (see the
definition of this term in relation to the production of, among other
things, flint daggers in Olausson 1993 & 2000:123). However, it is
also possible that they controlled the distribution of daggers themselves.
As we have seen, the specialisation that occurred in the marginal areas of
Scandinavia led to a diversification, as regards, for instance, grave traditions. Thus in Himmerland near the Limfjord, there is a great variation
in grave types. The flint daggers gave their owners prestige and linked
them to the esoteric powers, owing to the fact that they were produced
by a group of skilled artisans.
Carefully directed exogamous marriages, in order to establish and
maintain important contacts with other peer groups, was arranged over
an area covering the main distribution of flint daggers. Women usually
negotiated and arranged marriages in times when there was no direct
threat to the lineage cohesion. Since the women were not as affected by
and dependent on the kind of official prestige that a strict observation
of the tradition gave the men, they were not inclined to follow the
marriage rules proposed by the social structure. Instead, the women
strove to bring more women from their own original lineage into their
household, since this would enhance their own influence in the hamlet.
The dynamic relationship between the dominant social structure (elder
men) and practices with regard to matrimonial strategies for a long
period conserved the Late Neolithic social structure. However, the influx
of metal objects and techniques, which was controlled by the chiefly
lineages within the clans, coupled with the relative simplicity of the
bronze technology made the flint specialists in the Limfjord area more
or less obsolete around 1950 cal BC. Eventually, because of the lack of
objective foundation that the traditional order encountered as a result
of these events, the discrepancy between structure and practice became
unbearable. The internal conflict between structure and practice could
be upheld only as long as the dominant elders of the lineages guaranteed
the relative wealth that the flint-dagger production gave them. When
the chiefly lineages started to exchange metal, instead of flint objects,
341

Jan Apel
the producers of prestige items in flint could no longer maintain the
authority of the social structure. Certain lineages in the eastern production area solved the problem by increasing the quality of their product,
hence the production of the prestigious, fishtail, parallel-flaked daggers
of LN II. However, in the Limfjord area, the craft diminished.
Only agnatic relatives in the lineages were buried in the collective
gallery graves. Others, mostly women who had married into the lineage
but also males who were excluded from the right to inherit a position
in the corporate groups because of limited resources, were instead buried
in individual inhumation graves. This is supported by a significant
imbalance between the sexes in the Late Neolithic collective graves.
Thus, in the gallery graves of eastern central Sweden, there are 2,3 men
buried for every woman (Apel 1992) and from a larger material base, it
has been concluded that there are 1,7 buried men for every women
during the same period (Welinder 1979). The decline of the lineage
society is in fact signalled by the individual burial of the Early Bronze
Age, which should be connected with the single burials of women and
outcast men during LN II.
Within the consumption area, the daggers increased in value as a
function of the distance from the production areas. In general, they
were not very valuable in the production areas, simply because of the
large number of daggers in circulation. In these areas, it was the prestige
objects that were exchanged for daggers and that came from far away
that were valuable. In eastern central Sweden, the flint daggers, as well
as the debitage from the production stages that required a high degree
of know-how and were performed in public on the production sites,
were extremely valuable. As Mary Helms (1998) has pointed out,
excellent craftsmanship and long-distance trade are two strategies by
which tangible objects may become charged with intangible, esoteric
powers that are derived from cosmological realms and that give their
owners power and prestige. The chiefly lineages of eastern central Sweden
thereby actively used the flint artefacts and flakes that they obtained
through external contacts to maintain their power.

342

References

Adams, W. Y. & Adams, E. W., 1991. Archaeological Typology and Practical Reality. A
Dialectic Approach to Artifact Classification and Sorting. Cambridge University
Press. Cambridge.
Aghte, M., 1989. Bemerkungen zu Feuersteindolchen im nordwestlichen
Verbreitungsgebiet der Aunjetitzer kultur. Arbeits- u. Forschungsber. z. schs.
Bodenkmalpflege 33: 15-113.
1989. Bemerkungen zu Feuersteindolchen im nordwestlichen Verbreitungsgebeit
der Aunjetitzer Kultur. Praehistorica XV-XIV. Internationales symposium,
Univerzitas Karlova, Praha 1989: 305-309.
Ahlbeck, M., 1995. Kvartsiell Affasi. Den gropkeramiska boplatsen Bollbackens
stenmaterial ur ett processuellt perspektiv. BA-thesis, Department of Archaeology,
Uppsala University.
Althin, C. A., 1954. The Chronology of the Stone Age Settlements of Scania, Sweden, I.
Acta Archaeologica Lundensia. Series in 4, No 1. Lund.
Amick, D. S., Mauldin, P., and Tomka, S. A., 1988. An Evaluation of Debitage
Produced by Experimental Bifacial Core Reduction of a Georgetown Chert
Nodule. Lithic Technology, Vol. 17, No 1:26-36.
Andersson, A. & Andersson, S., 1991. Frord. I: Tre klassiska texter. Bokfrlaget
korpen. Gteborg.
Andersson, K. & Hjrthner-Holdar, E., 1989. Annelund - ett senneolitiskt
bebyggelsekomplex i sydvstra Uppland. Fornvnnen 1988/3:209-215.
Andersson, K., 2001. En hllkista vid Alby i Botkyrka socken, Sdermanland. I:
Bolin, H., Kaliff, A. & Zachrisson, T. (eds.), Mellan sten och brons.
Uppdragsarkeologi och forskning kring senneolitikum och bronslder. OPIA 27 &
Stockholm Archaeological Reports No. 39, 2001:74-84.
Andersson, K., Fagerlund, D. & Hamilton, J., 1990. Annelund - en senneolitisk
bondgrd. Arkeologi i Sverige 1987. Stockholm.
Apel, J., 1991. Hllkistor i stra Mellansverige. BA-thesis, Department of
Archaeology, Uppsala University.
1992. Aspekter p senneolitikum i norra Mlardalen eller slkten framfr allt
Fjlnir 11/1:5-21.

343

Jan Apel
2000a. Flint Daggers and Technological Knowledge. Production and
Consumption during LN I. In: Olausson, D. & Vandkilde, H. (eds.), Form,
Function and Context. Material culture studies in Scandinavian archaeology.
Almqvist & Wiksell International. Lund:135-154.
2000b. Kunskap, handlag och rmaterial en diskussion kring produktionen
och konsumtionen av senneolitiska flintdolkar i Skandinavien. Tor 30: 35-82.
2000c. Strict Conventions and Limited Ability Late Neolithic Flint
Technologies in Eastern Central Sweden. In: Hjrthner-Holdar, E. & Risberg, C.
(eds), Hantverkets roll i samhllet produktion och reproduktion. Rapport frn den
tredje workshopen om hantverksspecialisering Uppsala 18-19 mars 1999.
Riksantikvariembetet, UV GAL, Forskningrapport R0004: 3-16.
Apel, J., Bckstrm, Y., af Geijerstam, M., Hallgren, A.-L. and Kritz, A., 1996.
Valhalla, en mesolitisk lgerplats vid Lingbo, Ockelbo sn, Gstrikland. Slutrapporter
frn Arkeologikonsult AB 1996:6. Upplands Vsby.
Apel, J., Bckstrm, Y., Hallgren, F., Knutsson, K., Lekberg, P., Olsson, E., Steineke,
M. & Sundstrm, L., 1995. Fgelbacken och trattbgarsamhllet.
Samhllsorganisation och rituella samlingsplatser vid vergngen till en bofast
tillvaro i stra Mellansverige. Tor 27.
Apel, J., Hadevik, C. & Sundstrm, L., 1997. Burning down the house. The
transformational use of fire and other aspects of an Early Neolithic TRB site in
eastern central Sweden. Tor 29: 5-47.
Appelgren, K., Eklf, N. & Evanni, L., 2000. En senneolitisk boplats vid Grndal.
UV mitt, Dokumentation av fltarbetsfasen 2000:8. Arkeologisk underskning.
Riksantikvariembetet, Avdelningen fr arkeologiska underskningar. Stockholm.
Arnold, D. E., 1985. Ceramic Theory and Cultural Process. Cambridge University
Press. Cambridge.
Arnold, J., 1993. Labor and Rise of Complex Hunter-Gatherers. Journal of
Anthropological Archaeology 12:75-119.
1996. The Archaeology of Complex Hunter-Gatherers. Journal of Archaeological
Method and Theory, vol. 3, no. 2:77-126.
Arnold, V., 1990. Refitting of Waste Material from Dagger Production of Site
Tegelbarg. In: Cziesla, E., Eickhoff, N., Arts, N. & Winter, D. (eds.), The Big
Puzzle. International Symposium on Refitting Stone Artefacts, Mon Repos, 1987.
Holos, Bonn.
Artursson, M., (ed.) 2000. Stngby stationssamhlle. Boplats- och bebyggelselmningar
frn senneolitikum till yngre jrnlder Skne, Vallkrra sn, vg 930. UV syd,
rapport 2000:79. Arkeologisk frunderskning och underskning. Malm.
Asplund, J., 1970. Om undran infr samhllet. Argos. Uppsala.
Bakka, E., 1976. Fire vestnorske kystbuplaser fr yngre steinlder. Iskos 1:
Bantelmann, N., 1982. Endneolithische Funde im rheinisch-westflischen Raum. Karl
Wachholtz Verlag. Neumnster.
Barth, F., 1953. Principles of Social Organisation in Southern Kurdistan. Universitets
Ethnografiske Museum Bulletin, No. 7. Oslo.
1981. Process and Form in Social Life. Selected Essays of Fredrik Barth: Volume 1.
Routledge & Keegan Paul. London.

344

References
Becker, C.-J., 1951a. Flintgruberne ved Aalborg. Nationalmuseets Arbejdsmark 1951:
107-112.
1951b. Late-Neolithic Flint Mines at Aalborg. Acta Archaeologica, vol. XXII: 135152.
1952. Die Nordschwedischen flintdepots. Ein beitrag zur geschichte des
neolithischen fernhandels in Skandinavien. Acta Archaeologica XXIII: 31-79.
1957. De stortandade flint-spyd og deres alder. Aarbger for nordisk oldkyndighed
og historie 1956: 205-220.
1959. Flintmining in Neolithic Denmark. Antiquity, vol. 23: 87-92.
1980. Hov, Gem. Sennels, Amt Thisted, Jtland. In: Weisgerber, G. (ed.), 5000
Jahre Feuersteinbergbau. Deutsche Bergbau Museum, Bochum:457-464.
1988. Sydskandinavisk flint i nordsvenske fund fra mellem- og sen-neolitisk tid.
Iskos 7: 43-52.
1993. Flintminer og flintdistribution ved Limfjorden. Limfjordsprojektet. Rapporter
nr. 6: 111-134.
Berg, E., 1998. Intensjonelle nedlegginger av flintgjenstander i Rygge og Rde,
stfold. Beskrivelse og analyse av tre funn og funnomstndigheter. Universitetes
Oldsaksamling 1997/1998:43-62.
Bernstein, R. J., 1987., Bortom objektivism och relativism. Rda Bokfrlaget.
Gteborg.
Bertalanffy, L. von, 1968. General System Theory. New York.
Bertvall, C., 1985. En studie av flintdeperna i Vsterbotten. BA-thesis. Ume
University.
Beuker, J. R. & Drenth, E., 1998. Scandinavishe dolken in Drenth. Van
Rendierjager tot Ontginner, Nieuwe oudheidkundige ontdekkingen in Drenthe XLIV:
3-33.
Binford, L. R., 1968. Post-pleistocene adaptations. In: Binford, L. & Binford, S.
(eds.), New Perspectives in Archaeology. Chicago:313-341.
1972. An Archaeological Perspective. Academic Press. New York.
1982. Objectivity-Explanation-Archaeology-1981. In: Binford, L., 1983.
Working at Archaeology. Academic Press. New York, pp. 45-56.
1983a. In Pursuit of the Past. Decoding the Archaeological Record. Thames and
Hudson. New York.
1983b. Working at Archaeology. Academic Press. New York.
1989. Debating Archaeology. Academic Press. New York.
Bjrhem, N. & Sfvestad, U., 1989. Fosie IV. Byggnsadstradition och
bosttningsmnster under senneolitikum. Malmfynd 5. Malm.
Bloch, M., 1974. Symbols, song, dance and the features of articulation: is religion an
extreme form of traditionary authority? Archives Europeenes de Sociologie 15:55-81.
1989. Ritual, History and Power. Selected Papers in Anthropology. Monographs on
Social Anthropology 58. Th Athlone Press. London and Atlantic Highlands, N.J.
Bloemers, J. H. F., 1969. Flintdolche vom skandinavischen Typus in den
Niederlanden. Berichten van de Rijksdienst voor het Oudheidkundig
Bodemonderzoek 18: 47-100.

345

Jan Apel
Boas, F., 1955. Primitive Art. New York.
Boas, N. A., 1980. Egehjbopladsen fra aeldste bronzealder. Thrane, H., (ed.),
Bronzealderbebyggelsen i Norden. Odense:103-121.
1983. Egehj. A Settlement from the Early Bronze Age in East Jutland. Journal of
Danish Archaeology vol 8, 1989: 88-107
1986. Tidlige senneolitiske bosttelser p Djursland. I: C. Adamsen & K Ebbesen
(eds.), Stridsxetid i Sydskandinavien. Arkeologiske skrifter 1: 318-324.
Kpenhamn.
1991. Bronze Age Houses at Hemmed Church, East Jutland. Journal of Danish
Archaeology vol 8, 1989: 88-107
1993. Late Neolithic and Bronze Age Settlements at Hemmed Church and
Hemmed Plantation, East Jutland. Journal of Danish Archaeology vol 10, 1991:
119-135.
Bogucki, P. I., 1987. Forest Farmers and Stockherders. Early Agriculture and its
Consequences in North-Central Europe. New Studies in Archaeology. Cambridge.
Bolin, H., 1999. Kulturlandskapets korsvgar. Mellersta Norrland under de tv sista
rtusendena f. Kr. Stockholm Studies in Archaeology 19. Stockholm.
Bondesson, B., 1980. Sknska flintdolkar. En jmfrelse av det sknska och det danska
flintdolksmaterialet med utgngspunkt frn E Lomborgs Die Flintdolche Dnemarks.
BA-thesis, Univerity of Lund.
Bonnichsen, R, 1977. Models for Deriving Cultural Information from Stoone Tools.
National Museum of Man. Archaeological Survey of Canada. Mercury Series 60.
Ottawa.
Borenius-Jrpeland. ms. Piparberget. Rapportmanuskript frn Riksantikvariembete
UV Mitt. Stockholm.
Boserup, E., 1966. The Conditions of Agricultural growth. London.
Bourdieu, P., 1958. Sociologie de lArgerie. Oue sais-je?; 802. Paris.
1977. Outline of a Theory of Practice. Cambridge University Press. Cambridge.
1986. Kultursociologiska texter. Salamander frlag. Liding.
1991. Kultur och kritik. Daidalos. Gteborg.
1996. Homo Academicus. Brutus stlings Bokfrlag. Stockholm/Stehag.
1999a. Praktiskt frnuft. Bidrag till en handlingsteori. Daidalos. Gteborg.
1999b. Den manliga dominansen. Daidalos. Gteborg.
Bradbury, A. P. & Carr, P. J., 1999. Examining Stage and Continuum Models of
Flake Debris Analysis: An Experimental Approach. Journal of Archaeological
Science 26:105-116.
Bradley, B., 1975. Lithic Reduction Sequences: A Glossary and Discussion. In:
Swanson, E. (ed.), Lithic Technology: Making and Using Stone Tools. Mouton,
The Hague:5-14.
1978. Hard Hammer Soft Hammer: an Alternate Explanation. Flintknappers
Exchange 1 (2):8-10.
Braidwood, A. P. & Howe, B., 1960. Prehistoric Investigations in Iraqi Kurdistan.
Oriental Institute Studies in Ancient Oriental Civilization. Chicago.

346

References
Brew, J. O., 1946. Archaeology of Alkali Ridge, Southeastern Utah. Papers of the
Peabody Museum of Archaeology and Ethnology. Harvard University.
Broady, D., 1991. Sociologi och epistemologi. Om Pierre Bourdieus frfattarskap och den
historiska epistemologin. Andra korrigerade upplagan. HLS Frlag. Stockholm.
2000. Frord. In: Bourdieu, P., Konstens Regler. Det litterra fltets uppkomst och
struktur. Brutus stlings Bokfrlag. Stockholm/Stehag.
Broady, D. & Palme, M., 1986. Utgivarnas frord. I: Bourdieu, P., 1986.
Kultursociologiska texter. Salamander frlag. Liding.
Brown, J., 1995. Traditional Metalworking in Kenya. Oxbow Monographs 44.
Cambridge. Monographs in African Archaeology 38. Oxford.
Brumfield, E. M. & Earle, T. K. (eds.) 1987. Specialization, exchange, and complex
societies. Cambridge Universitey Press.
Brndstedt, J., 1957. Danmarks Oldtid I. Stenalderen. Kpenhavn.
Burstrm, M., 1999. Cultural Diversity in the Home Ground. How Archaeology
Can Make the World a Better Place. Current Swedish Archaeology vol. 7:21-26.
Butler, J. J., 1975. Comments on the Flint Daggers of Denmark. Norwegian
Archaeology Review vol. 8, no. 2:
Callahan, E., 1979. The Basic of Biface Knapping in the Eastern Fluted Point
Tradition. Archaeology of Eastern North America 7:1-180.
1987. An Evaluation of the Lithic Technology in Middle Sweden during the
Mesolithic and Neolithic. Aun 8. Uppsala.
1996. The Basics of Biface Knapping in the Eastern Fluted Point Tradition: A
Manual for Flintknappers and Lithic Analysts (second edition). Archaeology of
Eastern North America 7:1-180.
Callahan, E. & Apel, J., ms. The Production of the Prestigious Flint Daggers of Type IV
Flint Daggers.
Cammann, S. V. R., 1977. The Cult of the Jambiya: Dagger Wearing in Yemen.
Expedition. The magazine of Archaeology/Anthropology Vol. 19, No. 2: 27-34.
Campbell, T., 1986. Sju teorier om samhllet. Akademilitteratur. Stockholm.
Cassau, A., 1935. Ein Feuerstaindolch mit Holzgriff und Lederschiede aus
Wiepenkathen, Kreis Stade. Mannus XXVII:199-209.
Childe, V.G., 1928. The Most Ancient East: The Oriental Prelude to European
Prehistory. London.
Childs, T. S., 1999. After All, a Hoe Bought a Wife: The Social Dimensions of
Ironworking among the Toro of East Africa. In: Dobres, M-A. & Hoffman, C. R.
(eds.), The Social Dynamics of Technology. Smithsonian Institution Press.
Washington and London.
Clark, J. E. & Blake, M., 1996. The Power of Prestige: Competitive Generosity and
the Emergence of Rank Societies in Lowland Mesoamerica. In: Preucel, R. W. &
Hodder, I. (eds), Contemporary Archaeology in Theory - a reader. Blackwell. Oxford:
258-182.
Clark, J. E. & Parry, W. J., 1990. Craft Specialization and Cultural Complexity.
Research in Economic Anthropology, Vol. 12:289-346.

347

Jan Apel
Clark, J. E., 1986. Craft specialization and cultural complexity. Unpublished paper,
Department of Anthropology, University of Michigan.
Clark, J. G. D., 1932. Scandinavian Flint Daggers in Britain. Man 1932, no.
223:186-190.
Coser, L., 1984. Introduction. In: Durkheim, E., The Division of Labour in Society.
MacMillan Press ltd. London.
Crabtree, D. E., 1972. An Introduction to Flintworking. Occasional Papers of the
Idaho State University Museum, Number 20. Pocatello, Idaho.
Cullberg, C., 1960. Der Dolch von Ntraby. Meddelanden frn Lunds universitets
historiska museum 1960: 34-42.
Czerniak, L. & Kabacinski, J., 1997. The Erteblle Culture in the Southern Baltic
Coast. In: Krl, D. (ed.), The Built Environment of Coast Areas during the Stone
Age. A Symposium at the Centenary of Archaeological Excavations at Rzucewo 4-9
October 1994. Rzucewo-Gdansk.
Damm, C. B., 1991a. Continuity and Change. An Analysis of Social and Material
Patterns in the Danish Neolithic. Ph.D. Thesis. Dpt. of Archaeology, University of
Cambridge.
1991b. The Danish Single Grave Culture Etnic Migration or Social
Construction? Journal of Danish Archaeology, Vol. 10:199-205.
Danielsson, T., 2000. Norr om neoliticum? Neolitiska fynd i Dalarna. Unpublished
BA-thesis, Inst. fr Humaniora, Mitthgskolan. stersund.
Darwin, C. R., 1868. The Variation of Animals and Plants under Domestication.
Murray. London.
1976 [1859]. Om arternas uppkomst genom naturligt urval, eller de bst utrustade
rasernas bestn i kampen fr tillvaron. Natur och kultur. Stockholm.
Davy, G., 1960. . Durkheim. Revue francaise de sociologie I:3-24.
Deetz, J. & Dethlevsen, E., 1965. The Doppler Effect and Archaeology. A
Consideration of the Spatial Aspects of Seriation. Southwest Journal of Archaeology
vol. 21.
Dibble, H. L. & Pelcin, A., 1995. The effect of hammer mass and velocity on flake
mass. Journal of Archaeological Science 22: 429-439.
Divac, G. & Sedlcek, Z., 1999. Hortfund der Altbronzezeitlichen dolche von Praha 6Suchdol. Museum Nationale Pragae. Prag.
Dobres, M-A. & Hoffman, C. R. (eds.), 1999. The Social Dynamics of Technology.
Smithsonian Institution Press. Washington and London.
Domanski, M. & Webb, j. A., 2000. Flaking properties, petrology and use of Polish
flint. Antiquity 74:822-832.
Durkheim, E., 1978 [1895]. Sociologins metodregler. Bokfrlaget Korpen. Gteborg.
1984 [1893]. The Division of Labour in Society. Macmillan. London.
1993 [1897]. Sjlvmordet. Argos. Lund.
1995 [1912]. The Elementary Forms of Religious Life. The Free Press. New York.
Durkheim, E. & Mauss, M., 1963. Primitive Classification. The University of
Chicago Press. Chicago.

348

References
Earle, T., 1977. A reappraisal of redistribution: complex Hawaiian chiefdoms. In
Earle, T. K. & Ericson, J. (eds), Exchange Systems in Prehistory. Academic Press,
New York: 213-232.
1978. Economic and Social Organisation of a Complex Chiefdom: the Halelea
District, Kauai, Hawaii. Museum of Anthropology, University of Michigan,
Anthropological Paper 63. Ann Arbor.
1991. The Evolution of Chiefdoms. In: Earle, T. (ed.), 1991. Chiefdoms: Power
and Economy, and Ideology. Cambridge University Press. Cambridge.
(ed.), 1991. Chiefdoms: Power and Economy, and Ideology. Cambridge University
Press. Cambridge.
Earle, T., Bech, J-H., Kristiansen, K., Aperlo, P., Kelertas, K. and Steinberg, J., 1998.
The Political Economy of Late Neolithic and Early Bronze Age Society: the Thy
Archaeological Project. Norwegian Archaeological Review, Vol. 31, No. 1:1-28.
Ebbesen, K., 1975a. Comments on the flint daggers of Denmark. Norwegian
Archaeological Review vol. 8. nr 2:107-111.
1975b. Die jngere Trichterbecherkultur auf den dnischen inseln. Arkaeologiska
studier II. Kpenhavn.
1980. Flintafslag som offer. Kuml 1980:147-157.
1994. Tragtbaegarkulturens dolkstave. Aarbger for nordisk oldkyndighed og
historie 1992:103-136.
Eklf, N., 1999. stertlje 1:15. Rapport, Riksantikvariembetet UV mitt.
Stockholm.
Ekmyr-Westman, E., 1991. Upl. Skogs-Tibble sn, Skillsta 2:15, ra 138. Rapport
1991. Ume.
Elster, J., 1990. Vetenskapliga frklaringar. Bokfrlaget Korpen. Gteborg.
Englund, L-E., 2000. Smeder och makt. I: Hjrthner-Holdar, E. & Risberg, C.
(eds.), Hantverkets roll i samhllet produktion och reproduktion. Rapport frn den
tredje workshopen om hantverksspecialisering Uppsala 18-19 mars 1999.
Forskningsrapport R0004, Riksantikvariembete UV GAL. Uppsala.
Ericson, J. E. & Earle, T. K. (eds.) 1982. Contexts for Prehistoric Exchange. Academic
Press. New York.
Eriksen, T. H., 1993. Ethnicity and Nationalism. Anthropological Perspectives. Pluto
Press. London.
Ethelberg, G., 2000. Det snderjyske landbrugs historie: sten- og bronzealder . Haderslev
Museum: Historisk Samfund for Snderjylland. Haderslev.
Fabech, C., 1988. Storstenkisten fra Blaere. Kuml 1988: 45-75.
Fagerlund, D. & Hamilton, J., 1995. Arkeologi p vg. En hllkista och bebyggelse frn
frn senneolitikum och bronslder. Riksantikvariembetet UV Uppsala, rapport
1995:13. Uppsala.
Featherstone, M., 1994. Kultur, kropp och konsumtion. Symposion. Stockhom/Stehag.
Fields, K., 1995. Religion as an Eminently Social Thing. Translators Introduction in:
Durkheim, E., 1995. The Elementary Forms of Religious Life. The Free Press. New
York.
Firth, R., 1964. Essays on Social Organization and Values. The Athlone Press. London.

349

Jan Apel
Flenniken, J., 1985. Stoone Tool Reduction Techniques as Cultural Markers. In: Plew,
M. G., Woods, J. C. and Pavesic, M. G. (eds), Stone Tool Analysis: Essays in Honor
of Don E. Crabtree. Albuquerque: University of New Mexico Press:265-276.
Flenniken, J., Patterson, L. W. & Hayden, B., 1979. More on Staging. Flintknappers
Exchange 2 (1):26
Ford, J. A., 1954. The Type Concept Revisited. American Anthropologist 56 (1):4254.
Forsberg, L., 1985. Site Variability and Settlement Patterns. An Analysis of the
Hunther-Gatherer Settlement System in the Lule Valley, 1500 BC-BC/AD.
Archaeology and Environment 5. Ume.
1988. Bosttningsmnster vid Lule och Ume lv under bronslder och
frromersk jrnlder. Arkeologi i norr 1: 51-101.
1992. De norrlndska hllristningarna sociala kontext - Alternativa tolkningar.
Arkeologi i norr 3, s. 55-70.
Forssander, J. E., 1936. Der Ostskandinavische Norden whrend der ltesten Metallzeit
Europas. Skrifter utgivna av Kungl. Humanistiska Vetenskapssamfundet i Lund
XXII. Lund.
Geertz, C., 1957. Ritual and Social Change: A Javanese Example. American
Anthropologist 59:32-54.
1973. The Interpretation of Cultures. Selected Essays by Clifford Geertz. Hutchinson
of London. London.
Geijerstam, M. af, 1996. Mikrodebitageanalys som metod vid arkeologiska
underskningar. En inledande metodutvrdering med den neolitiska lokalen Ra
162 vid Fgelbacken som analysomrde. In: Lekberg, P. (ed.), 1996. Fgelbacken
- ett fornlmningskomplex i stra Vstmanland. Del II. Lmningar frn
mellanneolitikum, senneolitikum och jrnlder underskta 1994. Rapport till
lnsstyrelsen frn Arkeologikonsult AB i samarbete med Arkeologiska
institutionen vid Uppsala universitet.
Gejvall, N., 1963. Skelettmaterialet i Dragby hllkista. Tor IX: 85-122.
Geneste, J-M., 1989. Les Industries de la Grotte Vaufrey: Technologie du Dbitage,
Economie et Circulation de la Matiere Premire. In: Rigaud, J. P. (ed.), La Grotte
Vaufrey. Mmoires de la Socit Prhistorique Francaise, vol. 19:441-517.
Gerloff, S., 1975. The Early Bronze Age Daggers in Great Britain, and a
Reconsideration of the Wessex Culture. Prhistorische Bronzefunde. VI.2. Mnchen.
Giddens, A., 1978. Durkheim. Fontana. Glasgow.
1984. The Constitution of Society. Outline of the Theory of Structuration. Polity Press.
Cambridge.
1987. Social Theory and Modern Society. Polity Press. Cambridge.
1990. The Consequences of Modernity. Stanford university press. Stanford.
Glob, P V, 1951. En flintsmedie p Forns. Kuml 1951: 23-39
1945. Studier over den jyske Enkeltgravskultur. Aarbger for Nordisk oldkyndighed
og historie 1944: 1-23.
Godelier, M., 2000. Is Social Anthropology Still Worth the Trouble? A Response to
Some Echoes from America. Ethnos, vol. 65:3:301-316.

350

References
Goldhahn, J., 1999. Sagaholm. Hllristningar och gravritual. Studia Archaeologica
Umensis 11. Jnkpings Lns Museums Arkeologiska Rapportserie 41.
Jnkping.
Grace, R., 1990. The Limitations and Applications of Use Wear Analysis. In:
Grslund, B., Knutsson, H., Knutsson, K. & Taffinder, J. (eds.), The
interpretative possibilities of microwear studies. Proceedings of the International
conference on lithic use wear analysis, 15th-17th February 18989 in Uppsala,
Sweden. Aun 14, Uppsala.
Granath Zilln, G., 1999. En deposition i ett senneolitiskt landskap. I: Olausson, M.
(red.), Spiralens ga Tjugo artiklar kring aktuell bronsldersforskning.
Riksantikvariembetet, UV mitt. Stockholm:321-346.
Green, S., Houlder, C. H. & Keeley, L. W., 1982. A Flint Dagger from Ffair Rhos,
Ceredigion, Dyfed, Wales. Proceedings of the Prehistoric Society 48: 492-501.
Grimes, W. F., 1932. The Early Bronze Age Flint Dagger in England and Wales.
Proceedings of the Prehistoric Society East Anglia 6:340-355.
Grslund, B., 1974. Relativ datering. Om kronologisk metod i nordisk arekologi. Tor
XVI. Uppsala.
1996. Arkeologisk datering. Studentlitteratur. Lund.
Guneriussen, W., 1997. Aktr, handling och struktur. St