239 87 9MB
English Pages 136 [144] Year 2009
NEW DIRECTIONS IN ARCHAEOLOGY
Time, energy and stone tools EDITED BY ROBIN TORRENCE
THE LIBRARY ST, MABY4S COLLEGE OF MARYLAND ST.
-
MARY’S CITY,
MARYLAND
20605
NEW DIRECTIONS IN ARCHAEOLOGY Editors
Richard Bradley Professor of Archaeology, University of Reading
Timothy Earle Professor of Anthropology, University of California, Los Angeles
Ian Hodder Lecturer in Archaeology, University of Cambridge
Colin Renfrew Disney Professor of Archaeology, University of Cambridge, and Master of Jesus College
Jeremy Sabloff University Professor of Anthropology and the History and Philosophy of Science, University of Pittsburgh
Andrew Sherratt Department of Antiquities, Ashmolean Museum, Oxford
TIME, ENERGY AND STONE TOOLS
** ■
TIME, ENERGY AND STONE TOOLS EDITED BY ROBIN TORRENCE
"V
§§f
M i j
i
3
3
i wk i i
i i;
0
The right of the University of Cambridge to print and sell all manner of books was granted by Henry VIII in 1534. The University has printed and published continuously since 1584.
CAMBRIDGE UNIVERSITY PRESS CAMBRIDGE LONDON
NEW YORK
MELBOURNE
SYDNEY
NEW ROCHELLE
For Lewis R. Binford
Published by the Press Syndicate of the University of Cambridge The Pitt Building, Trumpington Street, Cambridge CB2 1RP 32 East 57th Street, New York, NY 10022, USA 10 Stamford Road, Oakleigh, Melbourne 3166, Australia © Cambridge University Press 1989 First published 1989 Printed in Great Britain at the University Press, Cambridge British Library cataloguing in publication data Time, energy and stone tools 1. Prehistoric stone tools I. Torrence, Robin II. Series 621.9080901
Library of Congress cataloguing in publication data Time, energy and stone tools. (New directions in archaeology) Bibliography. Includes index. 1. Man, Prehistoric-Tools. 2. Stone implements. I. Torrence, Robin. GN799.T6T55 1989 930.T028 88-25712
ISBN 0 521 25350 0
AL
CONTENTS
List of contributors Preface 1 Tools as optimal solutions Robin Torrence
vi v>i
1
2 From chopper to celt: the evolution of resharpening techniques Brian Hayden
7
3 The occupational history of sites and the interpretation of prehistoric technological systems: an example from Cedar Mesa, Utah Eileen Camilli
17
4 Trade or embedded procurement?: a test case from southern Illinois Carol A. Morrow and Richard W. Jefferies
27
5 Economies in raw material use by prehistoric hunter-gatherers RobertJeske
34
6 Lithic technology and mobility strategies: the Koster site Middle Archaic Rochelle Lurie
46
vi
Contents
7 Re-tooling: towards a behavioral theory of stone tools Robin Torrence
57
8 A cost-benefit study of functionally similar tools Roger A. Boydston
67
9 Reliable and maintainable technological strategies in the Mesolithic of mainland Britain Andrew Myers
78
10 Assessing social information in material objects: how well do lithics measure up? Joan M. Gero
92
11 Optimization and stone tool studies: problems and potential Michael A. Jochim
106
References
112 121
Index
CONTRIBUTORS
Roger Boydston, Department of Anthropology, Northwestern University, Evanston, Illinois Eileen Camilli, 3100 9th NW, Albuquerque, New Mexico Joan M. Gero, Department of Anthropology, University of South Carolina, Columbia, South Carolina Brian Hayden, Department of Archaeology, Simon Fraser University, Burnaby, British Columbia Richard W. Jefferies, Department of Anthropology, University of Kentucky, Lexington, Kentucky Robert Jeske, Department of Anthropology, Northwestern University, Evanston, Illinois Michael A. Jochim, Department of Anthropology, University of California, Santa Barbara, California Rochelle Lurie, Department of Anthropology, Northwestern » University, Evanston, Illinois Carol A. Morrow, Department of Anthropology, Southern Illinois University, Carbondale, Illinois Andrew Myers, Department of Archaeology and Prehistory, Sheffield University, Sheffield Robin Torrence, Department of Archaeology and Prehistory, Sheffield University, Sheffield
Vll
PREFACE
Nearly ten years ago Robin Derricourt of Cambridge University Press invited me to put together an edited book on stone tools, one
tool form, manufacture, use, maintenance and discard. Here, then, was the core of what I felt was a highly produc¬
that would combine innovative theoretical approaches with good
tive debate: because the issues transcended stone tools, the results
solid case studies. I set out enthusiastically to find potential contri¬
could lead to major developments in theory building relevant to
butors who were pursuing thoughtful research on lithics, but the
the discipline as a whole. For once attempts were being made to
project turned out not to be so simple as I had naively believed it
look at the general causes lying behind the variability in stone tool
would. At that time it was (and still is) the case that the major de¬
form and production, so well replicated and documented in the
velopments in research on lithics focused on methodological issues,
past but still so poorly understood. Perhaps now stone tools might
such as use-wear analysis or conjoining, and the detailed recon¬
be worth snatching from the grasp of the specialists who control
struction of behavior at single sites. Ethnoarchaeological studies
their study and restored to their rightful place in the center of
were yielding tantalizing data that presented some puzzles for
archaeological studies of past human behavior.
archaeological inferences, but almost no one appeared to be wrest¬
Consequently, in 1982 Robert Jeske and I organized a ses¬
ling with the theoretical issues raised by this work. All the excite¬
sion at the 27th Annual Meeting of the Society for American
ment in archaeology had shifted from artifacts to ecofacts and di¬
Archaeology in Minneapolis so that we could all share our various
rect evidence for subsistence.
competing ideas with others grappling with similar issues. A big
Disappointed in the response to my enquiries, I shelved the
thank you is owed to Bob for his instrumental role in getting this
project until 1981 when I had the good fortune to spend five
project back off the ground. Most of the papers in this volume
months teaching at Northwestern University. Here I discovered a
were first presented at the meeting, although they were all exten¬
core of graduate students, led by James Brown and Robert Vierra,
sively revised subsequently in the light of our fruitful discussions
who were attempting to apply Binford’s theoretical principles
and the insightful comments of our discussant, Larry Keeley, to
about hunter-gatherer subsistence and settlement to stone tools.
whom we are most grateful. Unfortunately, Peter Woodman’s
In addition, we shared an interest in neo-Darwinian theory in
paper could not be included here but those by Joan Gero and An¬
ecology concerning the role of optimization in shaping behavior.
drew Myers, which were solicited later, help broaden the theoreti¬
What provided much of the fun during our heated arguments,
cal and geographical coverage of this volume. We all owe a huge
however, was that no one agreed about which of the several likely
debt to Michael Jochim, who quite late in the day kindly agreed to
currencies was most relevant for understanding variability in stone
take on the task of evaluating the overall success of the contribu-
Preface
viii
tions from the perspective of someone sympathetic to the use of
Undeniably optimization theory is a new direction for lithic
optimization theory in archaeology but not a specialist in stone
studies. Yet, all such innovations occur in a specific academic con¬
tools. Although his insightful criticisms have certainly highlighted
text and one especially important source of inspiration is worth
problem areas where further thinking or research will be needed
special mention. Most of the authors have had little direct contact
and have identified potential limitations in inferences based on
with Lewis Binford, but all have been heavily influenced by his
lithics, his creative and tantalizing use of ideas put forward here in order to reinterpret changes during the Mesolithic of southwestern
highly creative attempts to explore the factors which structure the organization of hunter-gatherer technology. Binford’s (1971;
Germany is a real tribute to the potential of this theoretical ap¬
1973; 1983) attacks on traditional thinking about Mousterian
proach and in many ways vindicates the entire book.
variability provided crucial first glimpses into new ways of concep¬
For various reasons the book has taken much longer to
tualizing stone tools, although he had already addressed the ques¬
come to fruition than was originally intended. I would like to ex¬
tion of the meaning of style when designing a typology for projec¬
press my deepest gratitude to all the contributors for their kind pa¬
tile points (Binford 1963). He then went on to seriously widen the
tience and understanding during some difficult times. That their
cracks in established approaches; his seminal studies of tech¬
work is still as important and timely as it was in 1982 speaks a
nological organization among the Nunamiut set the scene for the
great deal for the high level of creativity shown here and for the
issues raised in this book (e.g. Binford 1977; 1978a; 1979; 1983).
poverty of theory in archaeology relevant to the organization of
Such concepts as curated and expedient technologies, embedded
technology. This is all the more striking given the continued ad¬
procurement of raw materials, or forager and collector subsis¬
vances in methods for describing stone tools and their uses.
tence-settlement systems form the backbone of many of these stu¬
Various people at Cambridge University Press have also de¬
dies or at the very least provide a framework against which to
monstrated admirable tolerance and been very supportive: Robin
build a different way of thinking. It therefore seems highly appro¬
Derricourt, Claire Davies-Jones, Kate Owen, Peter Richards, and
priate that in recognition of his many fundamental contributions
the always helpful and unflappable Sylvia Christiansen. Dorothy
to the development of both method and theory in lithics studies,
Cruse retyped several long manuscripts; Rochelle Lurie, Peter
this volume be dedicated to Lewis R. Binford.
White, Julian Thomas, and especially Andy Myers deserve thanks for their critical assistance in the final stages of production.
Robin Torrence
1
Chapter 1
Tools as optimal solutions Robin Torrence
Tool-using is a fundamental attribute of human behavior and as such
scriptions have been made for their own sake and items collected
deserves to be studied in its own right rather than simply as a reflec¬
to enrich museum basements. The current lack of interest has been
tion of other activities. In order to accomplish this task, archaeolo¬
well portrayed by Oswalt (1976, pp. 214-15) who notes that
gists should conceive of technology as one way in which people solve
introductory texts in anthropology rarely devote more than one or
problems posed both by external, environmental factors and by in¬
two pages to the whole area of tool use. Part of the explanation for
ternal social needs. The papers in this volume attempt to achieve this
this oversight is obvious: indigenous technologies are one of the
goal by using insights from evolutionary theory to propose how tech¬
first components to change with contact. Ethnographers may feel
nology can be optimized. Although there are disagreements among
there is little value in studying modern material culture when other
the authors about which currency is most appropriate and whether
traits such as social and ritual behavior appear to survive more
emphasis should be placed on the initial design of tools or how tech¬
intact.
nology adapts to other aspects of behavior, the detailed case studies,
Sadly, one potential consequence of this neglect is the fur¬
which range extensively in time and space, demonstrate the enor¬
ther widening of the gap between the sub-disciplines of archae¬
mous potential of building theory around the concept that tools are
ology and ethnology, since for the majority of human prehistory
optimal solutions.
by far the largest class of data available in the archaeological record is comprised of stone artifacts. On the other hand, archae¬
Theory of lithic studies
ologists have been notoriously poor at producing their own theo¬
Tool-using has often been cited as one of the major distinc¬
ries for behavior and have depended largely on borrowing from
tions in behavior between what is quintessentially human as op¬
anthropology and ecology, with, it must be admitted, mixed re¬
posed to animal: in other words, as one of the fundamental traits
sults. As a result, prehistorians are presently either simply ignoring
in what sets culture apart from nature. Certainly, there can be no
the bulk of their evidence in an attempt to mimic ethnographic
doubt that in the contemporary world technology plays an in¬
studies of subsistence, settlement, and exchange systems or worse
creasingly dominant role in everyday social life. But what can be
yet are conducting their research in a theoretical vacuum. For ex¬
said about the function and importance of tool-using in prehis¬
ample, most recent achievements in lithics studies involve the cre¬
toric and pre-modern societies? Unfortunately, anthropology has
ation of ever more sophisticated methods for studying stone tools
had little to offer in finding answers to these questions, since it has
(e.g. use-wear and residue analysis, fracture mechanics, spatial
long ignored technology and material culture except so far as de¬
patterning), but the results of analyses using this battery of tech-
Robin Torrence
2
As pointed out by Jochim in the concluding chapter, there
niques rarely contribute to our understanding of human behavior because the work is not specifically guided by questions significant
are both advantages and limitations associated with the accep¬
to the field as a whole. In the end, both archaeology and anthro¬
tance of optimization theory. On the positive side, the approach
pology have failed to develop a theoretical basis capable of ac¬
allows a large number of different forms of behavior to be con¬
counting for the wide range of tool-using observed among past
ceived of and treated in the same way. Technology, then, can be
and present societies and have therefore failed to address one of the most important aspects of human evolution.
incorporated into a broader view of behavior and studied along¬
Tools obviously play a part in the way people adapt to their
side and in the same way as subsistence, settlement, or social or¬
surroundings but they operate alongside and in conjunction with a
ganization. We can therefore begin to understand how various human strategies operate with respect to each other to achieve de¬
number of strategies to ensure the continuity of the social group as
sired goals rather than seeing them as totally separate entities. As
a whole. How do tools contribute to social reproduction in so¬
a result, stone tools have the potential to make a significant contri¬
cieties with varying degrees of complexity and how do they relate
bution to the study of human behavior. This is very good news for
to other forms of behavior such as exchange, ritual, or symbolic
the archaeologist and ultimately for all students of human behav¬
systems? How can we explain variability in the way people manu¬
ior.
facture and use tools as well as differences in the tools themselves?
On the other hand, optimal behavior in one sphere may be
Since they are fundamental for the understanding of human be¬
clearly defined, but for various reasons, such as incompatibility
havior and the origin and development of culture, such issues
between the goals of different forms of behavior, it may never
should be central to anthropology and archaeology. A primary
occur. Lack of fit between predictions and reality, therefore, is not
aim of this book, therefore, is to refocus the interest of archaeol¬
unexpected (cf. Foley 1985). Consequently, optimal models can
ogists and anthropologists on the study of tool-using by illustrating
never be tested, but are most profitably used to compare and con¬
that theory for explaining human strategies of tool manufacture
trast different forms of behavior, e.g. to evaluate changes in be¬
and use can and is being developed. Although none of the authors
havior through time or in different environments. A further limitation is that evolutionary theory strictly ap¬
of the papers that follow would claim to have a completely satis¬ factory and comprehensive theory about tool-using, all have at
plies only to types of behavior which have an effect on reproduc¬
least recognized the current need for theoretical approaches and
tive success. In other cases there is no reason to expect that selec¬
have each suggested innovative ways to look at the factors con¬
tion will produce an optimal outcome. Most authors here have
ditioning tool production and use and/or the role tools play in
therefore concentrated their attention on technology used in ac¬
human adaptive strategies. What, then, are the sources of variability in prehistoric
tivities which can be directly linked to reproductive success, as for example in the procurement and processing of food. Despite this
tools? When the book was first conceived, the idea which we set
limitation, Gero’s study of Peruvian stone tools demonstrates
out to explore was in what ways and to what degree are principles
that, as in economic studies in general, the basic concepts of cost/
derived from evolutionary theory relevant for understanding tool
benefit maximization have proved to be applicable to an even
production, use, maintenance, and discard. In particular, the
wider set of behavior. Her work demonstrates that when com¬
assumption that initially oriented most of these studies was that
pared to other forms of material culture, the investment of energy
tool-using, as for many other forms of behavior, was carried out in
required to carry desired quantities of information does not pay
such a way as to optimize the expenditure of time and energy.
off in certain circumstances. The issue of whether one should re¬
Since tools are created and employed to satisfy a perceived need
strict optimization solely to evolutionary theory or alternatively
and to accomplish tasks which would themselves be susceptible to
adopt the assumption of economic choice theory - that people will
selective pressures, then an optimal technology would be favored
always maximize some commodity (i.e. “utility”) - is raised here
and would persist. These concepts are operationalized by means of
but has not been adequately dealt with. It will certainly require
the relationship between the costs of adopting a form of behavior
greater consideration if further work along these lines is under¬
and the resulting benefits which accrue from it. The strategy which
taken.
produces the most favorable ratio of benefits to costs, calculated
Although the very basic concepts underlying optimization
in terms of the chosen currency (e.g. time, energy), is then the one
theory in ecology are generally shared by all the contributors, it is
which is defined as “optimal.”
interesting to see how many different directions in theory building
It is important to note that by adopting optimization
and in the associated case studies have been taken by the authors.
theory, the authors do not also assume that “progress” in tech¬
Beyond the very basic agreement that tools represent some form
nology is inevitable. Hayden, Boydston and Torrence make it very
of optimal solution, there are major disagreements about both the
clear that earlier notions in which tools somehow evolved apart
choice of currency (i.e. what problems are being addressed by or
from their users and that the change from primitive to sophisti¬
otherwise influence technological behavior) and whether the func¬
cated, inefficient to efficient, or simple to complex, for example, is
tion of the tool or the way technology adapts itself to external con¬
unproblematic are completely unacceptable. In contrast the pos¬
tracts should have primacy in constructing an optimal model. It
ition clearly taken by the contributors is that technology must be
is worth exploring these differences in approach and emphasis
understood as a particular adaptation created by the operation of
since they highlight the issues raised by the volume as a whole and
general principles of optimizing which were working within
also represent areas where increased attention to theory building is
specific local conditions.
needed. Not surprisingly, many of the controversies raised by
Tools as optimal solutions
3
comparing the various articles are also considered to be problema¬
optimization of the energy invested in lithic provisioning as the
tical in optimization studies in general and therefore reinforce the
limiting factor. He himself argues that technologies are devised in
potential contribution of stone tool studies for the wider study of
order to minimize expenditure of raw material because like food,
human behavior.
stone is a rare resource in the environment. The generality of Hay¬ den’s assumption is questioned in the final summing up by Jochim
Currencies
and the role of raw material is considered at length in my own
Most ecological studies of human behavior have utilized en¬
paper where, taking an opposing view, I prefer to see it as one of
ergy as the currency to be optimized, although calories, time, pro¬
the constraining factors on behavior rather than as a currency
tein, vitamins, trace elements, uncertainty, risk, and security have
which is itself optimized for its own sake. Whether raw material
also been suggested as at least as important, if not more so, par¬
should take precedence over other currencies or should be recast
ticularly in certain circumstances (e.g. Foley 1985, p. 229, p. 233;
in terms of energy is an unresolved issue which deserves further
Winterhalder and Smith 1981; Jochim 1983; Smith 1986; Winter-
discussion and debate since many archaeologists share Hayden’s
halder 1986). Not surprisingly, energy is also most commonly
emphasis on raw material conservation (e.g. see discussion in Ed¬
utilized here; however, many studies do not explicitly describe
monds 1987; Johnson and Morrow 1987).
their currency because, as noted below, technology is viewed more
A third currency is introduced by Gero. She stresses the im¬
as responsive to other forms of behavior, rather than as the focus
portance of information, although she does not explicitly assume
for optimizing. Energy forms the basic currency in the studies by Camilli,
that all groups will necessarily optimize the amount of infor¬
Morrow and Jefferies, Jeske, and Lurie, although Camilli and Jeske mention time as also important. For all these authors tech¬
to carry meaningful messages which take an active part in the process by which social relationships are negotiated, created, and
nological behavior is envisaged as the outcome of adjustments
maintained. Following on from perceptive observations by Wobst
made because the overall energy budget in the society is assumed
(1977) about the role of “style” and from Hodder’s (1982)
to be limited. Since this is the case, if there are additional require¬
proposals that material culture plays a dynamic, causative role in
ments placed on one aspect of behavior, such as a high demand for
social strategies, Gero examines the way lithic artifacts were used
processing material (Hayden) or increased difficulty in gaining
during the development of social complexity in Peru. What is es¬
access to raw materials for various reasons (Morrow and Jefferies,
pecially fascinating about her article is not so much her clear
Jeske, Lurie), then tool production, use, maintenance, and discard
demonstration of a decline in the degree of style observed on stone
has to alter in order to help process the smaller amount of energy
tools as complexity increases, but rather her hypothesis that lithic
more efficiently. Efficiency in energy processing is assumed by
artifacts can only take a limited part in this behavior because the
these authors not so much because it confers advantages by in¬
nature of stone tool production, a substractive technique, severely
creasing reproductive fitness, but because people have no other
limits the complexity of forms that can be created and, by defini¬
mation processed. Instead, she focuses on the ability of stone tools
choice. Accordingly, emphasis is placed on the constraints im¬
tion, the amount of information that can be encoded. In contrast,
posed on behavior rather than on how the currency can be opti¬
the potential of pottery for carrying complex messages is greatly
mized. The major criticism that can be directed at using energy as a
a theory predicting not only the degree of style which we would ex¬
currency, not only for modelling optimal technologies but for
pect to be expressed on material items, but also the optimum
other types of behavior as well, is whether energy is sufficiently
medium for achieving this goal. Certainly, it would be worth pur¬
scarce in most environments as to impose selective pressures
suing the relationship between information processing and stone
(Foley, 1985, p. 229). I also raise energy scarcity in my paper as
tool technology along the lines illustrated by Gero’s Peruvian
one of the reasons for questioning its utility; in addition, I ques¬
study.
tion the ability of most simple technology to process energy ef¬
enhanced because it is a plastic medium. Here we have the germ of
One approach to information which is not considered with¬
ficiently in the first place. In response, Torrence (1983), Myers
in this book concerns “uncertainty”, or simply lack of information
(1986) and Boydston in this volume have suggested that time is
about the environment defined in both physical and social terms.
more likely'to be a limiting factor and therefore is a more appro¬
Part of the reason why behavior designed to minimize uncertainty
priate currency for studying technology.
has been neglected may be that stone tools play only a minor role
At this point it is worth considering the role of raw
in this regard, although other forms of material culture can be in¬
materials in optimization models. For Morrow and Jefferies and
strumental, such as many types of facility (Torrence 1983) or spe¬
for Jeske, who deal specifically with the procurement of raw
cial purpose tools like calendrical devices for monitoring seasonal¬
materials, it is the energy involved in obtaining the stone that af¬
ity. Nevertheless, the potential of stone tools for participating
fects the subsequent production and consumption rather than the
within an optimal solution to problems posed by uncertainty
quantity of raw material itself. Lurie also considers the “availa¬
should ultimately be specifically considered.
bility of raw material” to be important but her phrase implies
A final issue that can be subsumed under the discussion of
more about the energy required in gaining access to it than simply
appropriate currency forms the basis of my own paper and is fur¬
the abundance of the stones. Only Hayden can be described as
ther illustrated by Myer’s study of British Mesolithic assemblages.
using raw material rather than energy as a currency (cf. Bamforth
In general, until recently, scholars using optimization theory have
1986), although his argument could easily be recast in terms of
assumed that selection would lead to the most efficient forms of
Robin Torrence
behavior. As noted by Boydston and Jeske, for example, efficiency is defined as the ratio of inputs to outputs of time or energy. Near¬ ly all the authors conform to established procedure and have therefore designed their predictions for optimal behavior on the types of procurement, production, etc. that would lead to the most efficient outcome, regardless of whether time or energy was the chosen currency. Ecologists and some anthropologists have begun to ques¬ tion the validity of focusing solely on efficiency (e.g. Stephens and Charnov 1982; Stephens and Krebs 1986; Smith 1986; Winterhalder 1986). An alternative view expressed by myself and Myers (also cf. Edmonds 1987) and referred to by Boydston, is that forms of behavior which deal with the management of risk will be more sensitive to selective pressures and are therefore more likely to approach optimality. It is argued that risk is especially relevant when considering optimal forms of technology because simple tools have only limited ability to increase the overall efficiency of energy capture but they are essential and can operate effectively in gaining access to resources. Instead, I have proposed that tool¬ using, procurement, production, and maintenance are better un¬ derstood as being one of the primary means used by humans to re¬ duce the potential effects of risk. Furthermore, in contrast to other types of behavior, such as mobility or exchange (e.g. Weissner 1977; 1982a; 1982b; Cashdan 1985), tool-using is most effective at reducing risk which occurs because resources are only available for short periods of time. For this reason time becomes not a cur¬
4
materials as well as the manufacture, distribution or exchange, maintenance, consumption, and re-use and recycling of stone tools. Two basic approaches are represented here, but they can most profitably be seen as two ends of the same process of model¬ building, rather than as opposing points of view. When they are fully incorporated, I think they will satisfy Jochim s criticism that stones should not be “isolated from the evidence of prehistoric human behavior” and should be “embedded” within “simul¬ taneous analyses of other classes of behavior.” Nevertheless, I do not feel that all stages of modelling should necessarily occur sim¬ ultaneously,” but rather that some should take preference over others. Until we are fully satisfied that our basic principles are cor¬ rect, we will not make further progress in understanding how and why humans design, create, and use stone tools. A fundamental principle that must not be overlooked is the fact that tools are not ends in themselves and do not dictate the be¬ havior of their makers and consumers. On the contrary, they are employed by humans to accomplish some function. In modelling the outcome of decisions, we must therefore assume that people choose the technology that will best suit their needs. Consequent¬ ly, the first step in building an optimal model of stone tool tech¬ nology must be to identify which among the problems faced by the social group in question can be solved by adopting some form of technology. Such problems may be generated from external causes such as environmental factors (e.g. the abundance of resources or the risk and uncertainty associated with their distribution in space
rency to be optimized for its own sake, but the most relevant attri¬
and time) or internal social variables such as competition for pres¬
bute of the type of risk to which technology provides an optimal
tige or the need for communication. Having isolated the ways in which tools can be solutions, the next task is to predict the optimal technology. This stage of
response. There can be no doubt that all the currencies used within the context of these papers are relevant for understanding human technology because energy, time, information, uncertainty and risk comprise basic and critical components of the environment to which human behavior must adapt. As in ecology at present, the relative importance of the likely currencies is still unknown and re¬ quires detailed research. For archaeology what is unresolved at this stage is the relevance of each of these currencies for modelling variability in stone tools. Do we need to take all factors into ac¬ count equally or, as some authors have argued, are some curren¬ cies simply more appropriate to modelling technology than others? Alternately, do particular environments pose unique types of potential for optimization or do different forms of selection op¬ erate in each case? Not only do we need to know more about the relationships between technology and currencies, but also the na¬ ture and role of selection within the process of evolution is still un¬ clear. Hopefully, additional research by archaeologists along the lines initiated by these studies will contribute to these wider con¬
modelling is considered in detail in the papers in the latter half of the book, i.e. Torrence, Boydston, Myers and Gero. It is not as easy as it sounds. Since people pursue a wide range of goals simul¬ taneously but are ultimately limited in the degree to which all can be achieved, priorities must be established and compromises reached between all the various aims. The result is a complex set of strategies in which the importance of technology will be esta¬ blished by the degree to which goals conflict and the resulting level of priority assigned to the problem which technology plays a part in solving. In other words, the more serious or “important” the problem, the more technology will be used and emphasized. Nevertheless, regardless of their potential to achieve a goal, tools will not be used unless their function is considered necessary. Opti¬ mal solutions are therefore not equivalent to the simple maxi¬ mization of a currency. The complex juggling of priorities which goes on in every cultural setting means that we should not isolate technology from the wider behavioral setting in which it operates. Jochim stresses
cerns.
this point when he suggests that stone tools be “embedded” in
A problem-solving approach
other studies of subsistence, settlement, etc. Although he is un¬
Beyond the choice of currency, there is a further fundamen¬
doubtedly correct to refocus the attention of lithics specialists
tal disagreement among the contributors concerning how to define
away from tools as the center of all behavior and place them
the “problems” that are relevant for modelling for stone tools and
alongside other strategies, I do not think we should then assume
associated technology as optimal solutions. By technology I am
that all goals pursued by individuals can be treated as equivalent
referring to all the activities involved in the acquisition of raw
or that all problems faced are equally serious. In some cases tech-
Tools as optimal solutions
nology will be essential to the successful completion of desired
quirement for raw materials is so flexible that the local scarcity of
strategies and in others it will play a minor role. For example, it
stone resources is irrelevant to the way technology operates. One
may be crucial to acquire certain raw materials in order to
would not be able to judge the importance of raw material before
produce particular critical tools or to invest much time and energy
knowing the technological requirements in the first place.
in making and maintaining tools which are deemed to be impor¬
Rather than begin with the ultimate limits and then work
tant. As a result, other forms of behavior will have to adapt to
backward to what might be a range of adaptations, surely it would
technological needs. For instance, as I have suggested in my
make more sense to begin with identifying the problems faced,
paper, Eskimos live off stored food at certain times of the year in
suggest a range of plausible solutions, then choose the optimal
order to “gear up” equipment which is absolutely essential for
technology from among these, on the basis of what is possible
their survival. In many other settings, however, technology will
given both internal constraints due to tasks with higher priorities
have to adapt to other needs. The outcome will always be com¬
and the external, ultimate limits. Another way to think of this is
plex, but in order to understand variability in technology, we need
that one should move from general concepts of the role of tech¬
to find ways to predict its relative importance in relation to other
nology within adaptive strategies to the specific, local conditions
activities for each particular setting. All these complex decisions are carried out within a physical
theory building should inevitably proceed along all these lines, as
and social environment which obviously determines the costs of
is the case in this volume, the most urgent need at the moment is
certain behavior and may set limits on the time and energy bud¬ gets. Technology like other forms of behavior, will therefore be
within which the basic principles are operationalized. Although
for a better understanding of how technology can be designed to solve problems rather than how it fits into wider behavior or how
situation. For example, in some environments raw materials are
it adapts to external constraints. These latter approaches have already received a great deal
scarce, whereas in others they are not. Groups with high mobility
of attention by archaeologists and perhaps are now in danger of
can more easily embed procurement of raw materials in other ac¬
leading us astray and obscuring the importance of studying
tivities than sedentary groups. All these factors combined with the
human technology for its own sake as one of the key elements of
initial tool requirements will shape the eventual outcome in terms
social and cultural behavior. The assumption that lies behind most current thinking seems to be that stone tools are completely adap¬
further adjusted to fit the specific constraints of each particular
of the means used to minimize the expenditure of raw materials within all stages from production through to discard and reuse.
table and therefore reflect all sorts of behavior, except their func¬
The papers in the first part of the book focus on the ways tech¬
tion. Certainly contributing to this view are the methodological
nology must respond to external constraints posed by the distri¬
difficulties faced when attempting to reconstruct past tool uses,
bution of raw materials or limits in the energy budget and make
but archaeologists appear to be ignoring the fact that tools were
predictions based on this aspect of the decision making process.
made to be used (not necessarily just for utilitarian purposes).
These are the papers by Hayden, Camilli, Morrow and Jefferies,
Another factor that has also clouded the issue is the emphasis placed on reconstructing the mobility patterns of hunter-gath¬
Jeske, Lurie. Obviously, the technology utilized by a group is the out¬
erers. Since this has become such an important aim in archae¬
come of decisions concerning firstly the initial design of the tools
ology, lithics experts have focused on the aspects of technology
and secondly the way technology must fit in and adapt to the
that could adapt to the need for various patterns of mobility (cf.
overall goals set by people and the constraints under which they
Camilli, Lurie). Yet, one can question the importance of mobility
inevitably operate. Both aspects are important to the construction
in creating the design of tools relative to the function of the imple¬
of a final optimal model for a technology. Future attempts at con¬
ment or even to third or fourth factors with even higher orders of
structing predictions should attack both aspects rather than focus
priority. Mobility patterns will contribute to the context in which
on one or the other as is presently the case, possibly because the
tools must be acquired and used, but we should be more specific
task is so large and difficult. The papers in this volume indicate
about its direct effects on the choice of tools in the first case.
that already we have many of the appropriate concepts at hand. In
Once again it is the desire to develop better methodologies
order to counter Jochim’s criticism that we have an incomplete un¬
for making inferences about past behavior that has dominated re¬
derstanding of stone tools, we simply need to incorporate all these
search on stone tools. Tools are currently being studied largely in
approaches into a wider view of the total process of problem-solv¬
order to be used to answer questions about other aspects of be¬ havior. Obviously, the need for theory linking artifacts to be¬
ing.
havioral strategies is especially crucial to archaeology since all
Priorities for future research
studies of dynamic phenomena are necessarily dependent on in¬
Regardless of the importance of examining both the role of
ference, which is based on just this type of theory. The significance
technology within the entire range of problem-solving undertaken
of these attempts for what Binford (1983) has called middle range
by past individuals and the ultimate constraints on behavior, the
theory has not been lost on the contributors. Most of the papers
former approach must initially take precedence in making models.
here, and in particular the extremely insightful study by Camilli,
After all, it is ultimately the needs that define what aspects of the
make very important contributions to the methodology by point¬
physical and social environment will be relevant. For example, it
ing to new directions where stone tools can be used to infer about
may be the case that the demand for tools is so low or that the re¬
other types of behavior.
Robin Torrence
Methods are undoubtedly important to archaeology, but I
6
considers the entire history of stone tool-using, the temporal but
feel that in emphasizing them to the degree that is fashionable at
not the spatial scale is decreased in Torrence’s paper which focuses
present, we may lose sight of the potential of stone tool studies for
on a universal phenomenon, but one that occurs during a much
contributing to the development of general theories about human
smaller length of time. We then move gradually down in scale to
behavior. For this reason, it becomes all that much more impor¬
the regional studies of Boydston, Myers, and Jochim and then
tant for further research to direct more attention to the wider
finally to the site as the major focus of study in the work by
issues raised by the studies represented here. What are the roles
Camilli, Morrow and Jefferies, Jeske, and Lurie. Like lithic studies
which technology plays in human adaptive strategies? How and
in general, however, the majority concentrate on hunter-gatherer
why do these vary in specific settings? Under what conditions does
adaptations, although the papers by Hayden, Camilli, and Tor¬
technology take precedence over other strategies and when does it
rence address the changes that occur with the adoption of agricul¬
play a lesser role? What is the relevant currency for modelling
ture and Gero provides a valuable insight into the changing roles
technology and does this differ for other aspects of behavior? To
of stone tools with the development of social complexity.
what extent do external constraints limit technological behavior? By taking up the challenge of these very general and difficult ques¬
To conclude the volume, Michael Jochim was asked to provide an evaluation of the potential of stone tools studies in
tions, the study of lithics could play a much more fundamental
light of the work presented here. Since he is not a lithics expert
part in the general debates about the nature of human behavior
himself, but does have an interest in using evolutionary theory to
than has traditionally been the case. It is hoped that the attempts
understand the development of human behavior (e.g. Jochim
represented here will stimulate a wider audience of archaeologists
1981; 1983), his insights about the general relevance of this work
to direct their efforts to the understanding rather than simply to the
are especially valuable. Although Jochim has pinpointed some real
description, of one of the more neglected facets of human evol¬ ution.
difficulties with the concepts as they stand at present and es¬ pecially with operationalizing them, he shares the view of the authors that general theory relevant to understanding human be¬
Applications
havior is essential for the study of stone tools. Much more hard
Theories rarely develop in isolation from concrete facts and
thinking, carefully controlled experimental work, and detailed
such is the case with the work presented here. All the authors have
archaeological applications are certainly required to fully imple¬
analysed their ideas in the context of specific archaeological cases.
ment the insights of these studies. .Before proceeding further, how¬
One of the real strengths of this collection must be the wide variety
ever, I must stress that we need not limit ourselves to considering
of the settings represented. The broad range in the spatial and
only how stone tools can inform us about aspects of behavior. We
temporal scales of the analyses serves to highlight the generality of
could also significantly increase our knowledge about human be¬
using optimality as a way to approach stone tools. Beginning with
havior if we also recognized and attempted to understand the ac¬
Hayden who applies his ideas on a world-wide scale and who also
tive role that tools and technology play in human strategies.
7
Chapter 2
From chopper to celt: the evolution of resharpening techniques Brian Hayden
Among hunter-gatherers important constraints exist on the procure¬
sequence, billet techniques were added to hard-hammer per¬
ment of lithic materials. When changes in one sector of the lithic
cussion resharpening of core tools. This was followed by an em¬
system increase procurement costs beyond acceptable limits,
phasis on resharpening flake tools, then by the introduction of
compensatory change in other aspects of the system become
pressure flaking for resharpening tools, and finally by edge-grind¬
necessary. The two most variable parts of the system in this regard
ing. In many areas of the world, changes over time in resharpening
involve (1) the quantity of materials processed and (2) the use-lives of tools as determined by resharpening techniques. Largely because
techniques continue to be extremely useful for situating assemb¬ lages in broad developmental schemes. For instance, if an assem¬
cutting requirements increased over time and also because of
blage contains pressure-flaking or edge-grinding, it is unlikely to
advances in wood, hunting, and shelter technology, the resharpening
be more than 20,000 years old no matter where it is found. Simi¬
mode changed from hard-hammer percussion, to soft-hammer
larly, assemblages with billet worked bifaces are unlikely to be
percussion, to the secondary resharpening offtakes, to pressure
more than a million years old.
resharpening, andfinally to edge-grinding. This model posits that as
Surprisingly, even after almost two decades of processual
the conservation of lithic raw material becomes an increasingly
archaeology, no one has attempted to explain why major changes
important consideration, reduction and resharpening techniques
in resharpening techniques should take place. It seems to have
capable of conserving material to greater degrees will be adopted
been largely assumed, tacitly or otherwise, that the various reshar¬
despite higher energy costs in obtaining and preparing special raw
pening modes changed as a natural effect of “progress” or in the
materials and in the manufacturing of tools.
same manner as stylistic elements. For instance, the Acheulian bi¬ face is sometimes viewed in terms of its aesthetic superiority to the
Introduction
Oldowan chopper (Childe 1942, p. 29; Clark 1967, p. 35). How¬
The identification of resharpening techniques has been used
ever, I would argue that there are few aspects of lithic assemblages
for well over a century as an important criterion in the classifica¬
which can be as directly tied to considerations of energy expen¬
tion of major stages in cultural evolution. By resharpening
diture and selection mechansims as resharpening techniques.
technique” I refer to the rejuvenation of dulled edges to prolong
Viewing them in terms of stylistic variation or general notions of
the use-life of the tool, rather than the production of the original
progress are probably the least useful explanatory approaches.
edge except where manufacturing of that edge is intended to “set it up” for future resharpenings. In the classic European Stone Age
The mere fact that changes in resharpening techniques took place over widespread areas, that they persisted for great lengths
Brian Hayden
5
of time, and that they became increasingly labor intensive in most
such as the Australian Aboriginals, the New Guinea Highlanders,
areas, should be a strong indication that they reflected adaptive
and the Maya (Hayden 1979a; Hayden and Nelson 1981; White
responses to powerful material constraints. In this paper I propose
1968, p. 513; 1969). Individuals who continue to make stone tools
an explanation for major shifts that have occurred in resharpening
in traditional ways simply do not have normative ideas of mor¬
modes over the past two million years. Resharpening modes can
phology when they make most tools. Whether expedient or cur¬
be categorized into five strategies, which I will argue are adaptive under different conditions: (1) hard-hammer percussion of core tools; (2) soft-hammer or billet resharpening of core tools; (3) use of resharpened flake and blade tools; (4) use of pressure-flaking; and (5) edge-grinding. A sixth strategy might consist of not shar¬ pening an original edge at all. While the following model can cer¬ tainly accommodate this strategy, it will not be the main focus of this article.
ated, tools are manufactured largely because individuals are faced both with a task to perform and with the necessity of creating an edge which will accomplish it. The problem, then, is simply to de¬ termine what kind of raw material and edge modification will achieve this aim. In this interactive situation, the details of the resharpening strategy are the major determinants of ultimate tool morphology. This is not to say that morphological patterning is lacking from
In some respects, the resharpening strategies that I will be dealing with resemble the “organization” of assemblages that
more complicated affair than the normative view of culture leads
lithic analysts have begun to investigate in recent years (Goodyear
archaeologists to believe. I must also emphasize that while the
1982; Torrence 1983; Shott 1986; Johnson and Morrow 1987).
manufacture of pressure-flaked and edge-ground tools or other
these assemblages, only that explaining these patterns is a much
However, I prefer the term, “strategy” because it more directly im¬
hafted tools does, to some extent, require the existence of “mental
plies a unifying principle that is not static but responds to changes
templates” related to the “finished” object, these templates are
in conditions, and is more intuitively explanatory in nature. I do
geared not to the form of an object per se, but to the need for solid
not purport to deal with overall assemblage organization in this
hafting or for successive resharpenings. That is, such objects are
study, but only those factors that most directly affect resharpening choices. Because the material constraints varied from group to
only shaped initially in order to make them serviceable for hafting and initial use and to set them up for further resharpenings. In this sense they are not really “finished” at all until they are discarded.
group and from time to time, and because several options were
Such tools may only appear toward the end of the Pleistocene and
often available for dealing with particular problems, there was
in environments where it is adaptive to “gear up” for an entire
probably no universal, rigid pattern of succession in the strategies
season’s work, as among the Inuit (Binford 1979; 1980). However,
to be discussed. Yet there are some very broad similarities. To a
Acheulian and Mousterian bifaces probably represent early exam¬
large extent, the following discussions will deal with long-term
ples of such artifacts. Some aspects of the following model of
trends and general evolution - those behaviors which finally em¬
changes in resharpening strategies have been presented at least in
erge as being most adaptive after the vagaries of tradition, poorly-
part elsewhere (Hayden 1977a; 1981, p. 520).
considered short-term solutions, deceptive environmental con¬ ditions, and options with poor future potentials have been elimi¬ nated. On the other hand, it is also possible that some of the rela¬ tionships to be discussed may also operate in a strong manner in the context of very specific, short-term adaptations. It should be emphasized that particularly in the latter strategies, several dif¬ ferent resharpening modes could be combined in a single assem¬ blage to deal with the differing characteristics of resource exploita¬ tion. In the following discussion, I will only discuss tools which were modified to rejuvenate their cutting edges. I will not deal with objects which were pressure-flaked or ground merely to achieve desired shapes, such as ritual or status items or perhaps some arrow points. I do not pretend to provide a model for all classes of stone tools.
Effort and lithic resource procurement Concisely stated, the model I employ posits that as the con¬ servation of lithic raw material becomes an increasingly important consideration, reduction and resharpening techniques capable of conserving material to greater degrees will be adopted despite higher energy costs in obtaining and preparing special raw mat¬ erials and in the manufacturing of tools. Because suitable raw materials for making stone tools are not ubiquitous in the environ¬ ment, there is a definite and usually substantial cost associated with their systematic procurement. In many respects, prehis¬ torians can treat lithic resources like plant food resources (Wintdrhalder and Smith 1981) since they are necessary for existence, occur in localized patches, and involve (1) search costs, (2) pro¬ curement costs, and (3) processing costs. Absolute limits on the
The model Before outlining the details of an explanatory model, it is worth emphasizing that in contrast to traditional interpretations, I view most (but not all) tool morphologies as the result of reshar¬ pening rather than any manufacturing process in which an artisan had a preconceived form in mind, sat down to make a biface, or a scraper, and then tucked it away for vague future uses. I have re¬ jected this latter point of view largely as a result of my own and other archaeologists’ work with contemporary lithic-using groups
amount of lithic material that individuals can carry with them (probably not more than one to two kilograms per family), com¬ bined with steady consumption, means that stone materials have to be replenished at regular intervals and are in regular demand. Although lithic resources may occur in a few areas where food is also found in quantity, in most areas the reliable sources of useful raw material are probably rare enough so that travel to these limited places presents a substantial cost in terms of time, ef¬ fort, and scheduling. From my experience, I would be surprised if
From chopper to celt: the evolution of resharpening techniques
9
Fig. 2.1. Factors affecting the amount of energy spent on lithic provisioning
the area normally utilized by most hunter-gatherers included more
effort must be expended in the yearly provisioning of lithic resour¬
than two or three sources of abundant, high quality lithic material,
ces: (1) the size and shape of tools; (2) the availability of suitable
although other high quality sources with limited quantities and in¬
raw materials for specific tasks; and (3) the frequency of tool re¬
volving high search and procurement costs might occur. Optimal
placements (fig. 2.1).
exploitation of food and lithic resources at specific sites might co¬ occur a few times during the year; however, this could not be ex¬ pected to be the case for the majority of sites. For most locations,
Tool size and shape Variation in tool size can affect the frequency with which
if individuals or groups were to run out of raw material, replenish¬
lithic material must be procured. The smaller the tools and the
ing lithic resources would involve a significant extra expenditure in
more blade-like they are, the more economically a given quantity
time and effort and would create scheduling conflicts. As Lee
of raw material can be used (Sheets and Muto 1972; Byrne 1980).
(1969, p. 60) has emphasized, hunter-gatherers view unnecessary
However, the nature of most tasks usually imposes relatively nar¬
subsistence travel as undesirable, and for good reason. One way of
row ranges on size. These limits can be overcome to a considerable
minimizing the effort incurred by inconvenient procurement trips
extent by the use of blade forms and/or hafting. With hafting, for
would be to store or cache raw materials at regularly used camp¬
example, a small chipped stone adze can perform many of the
sites. This corresponds to Binford’s (1979; 1980) logistically or¬
tasks of a core chopping implement. Reduction in tool size and the
ganized resource strategy. However, this is only a partial solution
use of blades or hafting might help decrease raw material procure¬
to the problem and is primarily useful in emergencies. If this stra¬
ment costs, but such strategies are relatively limited in their capac¬
tegy constituted the major way of coping with the risk of running
ity to offset increasing consumption of stone when compared to
short of lithic material, individuals with significantly increased
other factors, such as resharpening techniques.
cutting requirements would find they were spending inordinate amounts of time and energy replenishing their caches. Because procurement of lithic rescources constitutes a
Availability of suitable raw material The availability of suitable raw material for specific uses is
major cost in terms of time and effort during most of the year,
affected by a number of factors (fig. 2.1) which will be discussed
there are strong reasons to expect that individuals and groups
here only in summary fashion since my main concern is with re¬
would have attempted to minimize unnecessary procurement
sharpening strategies. It must be emphasized from the outset,
costs. Groups which did not do so might well find that their diet
however, that raw material availability may vary according to the
suffered accordingly when they needed to make special trips to re¬
specific task. For instance, vein quartz pebbles which may be per¬
plenish stone materials; in famine years they might find that the re¬
fectly suitable for butchering small or moderate-sized game, are
sulting scheduling conflicts and extra energy expenditure of such
often found widely dispersed over landscapes used by hunter-
trips could have disastrous results. Groups following such non-
gatherers and are therefore easily accessible. On the other hand,
optimal strategies would have strong natural selection pressures
such material would be of little use in heavy woodworking and in
brought upon them to change their behavior. Thus, there are im¬
the butchering of very large animals. Suitable material for these
portant reasons for assuming that lithic procurement costs were
tasks might only be found in a single location in the same landscape.
kept relatively low among hunter-gatherers.
Thus, each task may be associated with a separate strategy for
There are three factors which directly determine how much
minimizing the effort involved in lithic provisioning, and several
Brian Hayden
10
different strategies might be represented in a single assemblage. Of
material processed by hunter-gatherers has changed relatively lit¬
course, in the case of generalized industries with only a few multi¬
tle during the Pleistocene and Holocene. In fact, there is ample ev¬
functional tools, the number of strategies used would be minimal.
idence to indicate that the quantity of organic materials processed
The most obvious factor influencing the availability of suit¬
has changed over time by many orders of magnitude with substan¬
able raw material for a given task would be the distance to the
tial consequences for the lithic provisioning system (Hayden
source. Distance is a measure of the time and effort costs of ac¬
1981).
quiring raw material. In cases where barriers prevent direct pro¬
Finally, and most importantly, the type of resharpening
curement by users, raw materials would have to be obtained via
used on tools has a dramatic effect on the frequency of tool re¬
exchange or trade. In such cases, availability would be measured in terms of trade costs. Substantial barriers to raw material access
placement. Depending on the technique employed, a tool may last
are probably more typical of sedentary communities since they
can be satisfactorily resharpened using the same techniques.
have reduced foraging ranges and more limited movements away
Consequently, the resharpening properties of the available raw materials also affect the frequencies of tool replacements. In most
from residences (Rolland 1981, p. 31, p. 34). A factor which is often overlooked in assessing the availa¬
a few minutes, or several generations. Not all lithic raw materials
areas, materials suitable for the full range of resharpening techni¬
bility of suitable raw materials is difficulty of procurement. Al¬
ques could probably be procured either directly, or indirectly by
though suitable material might have been found in a number of
trade, if the advantage of doing so was perceived as great enough.
different locations, the very sporadic occurrence at these locations of acceptable quality and appropriately sized pieces probably re¬
Variable interactions
sulted in excessive search or even excavation costs. Thus, as with
In the above model, factors affecting the effort expended on
food resources, it was probably more efficient only to exploit
lithic provisioning have been presented separately. However, there
sources where suitable quality and quantity of raw material could
are several additional important features of the model: (1) it is in¬
be obtained rapidly and easily on a regular basis. This does not
teractive, or systemic, and (2) the ranges of values for many of the
imply that individuals would have ignored good pieces of raw
factors differ considerably. When I refer to the interactive nature
material if encountered by chance, or that less desirable sources
of the model, I mean that changes in one aspect resulting in unac¬
would be ignored when need and opportunity dictated. Similarly,
ceptable increases in energy spent on lithic provisioning must be
most foragers do not ignore unexpected encounters with game,
offset by changes in other aspects which will return energy expen¬
and they shift to less cost-efficient foods when necessary. The main
ditures to acceptable levels. Failure to accomplish this can have
point is that for the regular, systematic, dependable lithic require¬
severe survival consequences in times of resource stress. Thus,
ments of hunter-gatherers, sub-optimal lithic sources would not be
where there is insufficient raw material for whatever reason, and
efficient or reliable. Consequently, heavy use or dependence on
assuming a minimal number of cutting/processing operations are
sub-optimal sources would increase lithic procurement costs
necessary for survival, the energy spent on lithic provisioning can
above acceptable limits.
be kept within reasonable limits by either (1) reducing the size of
One other factor affecting procurement costs in temperate
stone tools, (2) obtaining more cutting edge from raw material, (3)
areas, which is often overlooked by prehistorians, is the seasonal
decreasing the frequency of tool replacements (e.g. Walker 1978),
variation of ground conditions and water levels. In areas with sig¬
(4) using a combination of techniques, or (5) using non-lithic sub¬
nificant snow cover, for instance, procurement of raw materials
stitutes such as bone, shell, or bamboo. These will not be pursued
may become prohibitively expensive during the winter months,
here. Similarly, where the amount of organic material processed
thereby requiring stock-piling or careful conservation of raw mat¬
increases the consumption of tools and thus the energy spent on
erials for much of the year (Rolland 1981). Similarly, seasonal flooding may render cobble beds in streams inaccessible, as among
lithic provisioning, such increases can be offset by augmenting availability through trade, by decreasing tool size, by using more
metate makers in Guatemala (Hayden and Nelson 1981). Thus
efficient resharpening strategies, or by using raw materials with
seasonal availability probably influences the strategies used for
more efficient cutting properties. However, if a particular reshar¬
minimizing effort spent on lithic provisioning during each season.
pening technique becomes more efficient, only certain raw mat¬ erials might be suitable, and this in turn affects the availability of
Tool replacement
appropriate raw materials. Thus, the interactive characteristics of
There are several factors which strongly affect the frequency
the model can be relatively complex and a number of alternate
of tool replacement (fig. 2.1). The edge efficiency and rate of edge dulling of lithic materials used in specific contact situations is one
solutions to given cost increases can be envisioned (fig. 2.1). In reality, however, it appears that some of the factors con¬
of the most obvious factors. For example flint edges last far longer
sidered have relatively limited ranges of acceptable values. For in¬
in meat cutting or woodworking than calcareous chert edges.
stance, the difference between the minimal and maximal cutting
Another extremely important factor in determining the frequency
efficiency of lithic materials likely to be used in a given task is
of tool replacement is the frequency and amount of work required,
probably not great. When the efficiency of lithic materials drops
that is, the types and amounts of materials being processed for
below a certain point, they cease to be used and efforts are directed
making tools, facilities, and as food. This is far too often over¬
toward procuring more acceptable raw materials. Alternatively,
looked by prehistorians who tend to assume that the quantity of
the task might simply be dropped from the cultural repertory if it
From chopper to celt: the evolution of resharpening techniques
is not critical, or if alternatives are available. Similarly, ranges in
11
(a)
acceptable and useful tool sizes are probably quite limited. I view the availability of raw materials, the use of hafting, and the extension of cutting edge per kilogram via blade produc¬ tion, as occupying an intermediate range of values. Some groups had more abundant lithic resources than others, and the use of hafts and blade forms could reduce the number of reprovisioning trips to quarries considerably. I hypothesize that the potential for absorbing increasing provisioning costs by these means, although considerable, was still limited in comparison to other factors, such as changing resharpening techniques. Differences in these factors should be relatively easy to monitor archaeologically from assemblage characteristics and climatic and geological studies. Raw material availability in a given area would not be expected to change significantly over time except as new task requirements emerged. The factors which I view as having the greatest latitude in terms of possible values and which changed the most over time are (1) the amount of organic materials processed by groups and (2) the effect that resharpening techniques had on tool use-lives. It is on these two factors that I shall concentrate in attempting to ex¬ plain the broadest technological changes in stone tools during the Pleistocene. Ideally, in determining precisely which resharpening modes would be most economical under given conditions of stone availability, at given levels of cutting requirements, or under other conditions, it would be necessary to estimate the frequency with which lithic tools would have to be replaced in each resharpening mode. While some estimates exist for some of the tools, they tend to be rather subjective and vague. Much better estimates can be obtained in the context of controlled experiments. However, my purpose here is not to make quantified “tests” of this model, but only to present it as a possible explanation for changes in reshar¬ pening modes, and to show how it might be applied to various assemblage types over time. “Tests” of the model in any event have to take into account raw material availability for specific hunter-gatherer groups as well as the other factors mentioned variables which at present can usually only be roughly estimated for most archaeological assemblages. Having presented the general assumptions of the model, it can now be applied in more specific terms, focusing on major changes in resharpening modes as they sequentially appeared over time. Once again, these changes are long-term, general evolutionary trends which may become ob¬ scured in short-term or other specific situations due to the use of
Fig. 2.2. (a) Characteristic resharpening sequence using unifacial hard-hammer percussion. Note that only a few flakes can be re¬ moved before edge angles become too abrupt for use or further resharpening, (b) Bifacial resharpening of core tools using hardhammer percussion often could be used succesfully to extend the use-lives of core tools by enabling a few more flakes to be removed, (c) By carefully preparing platforms and using soft-hammer per¬ cussors, core tools could be resharpened many times and the uselife of core tools could be greatly extended, as in the case of Acheulian bifaces
alternative strategies that might be adequate under most con¬ ditions but become critically deficient during periods of famine or
generally not essential. Resharpening is thus an extremely rapid
other episodically severe conditions. On the other hand, some of
process lasting only a few seconds. Therefore, where economy of
these changes may appear as the most reasonable short-term, spe¬
raw material is not a great concern, resharpening via hard-
cific evolutionary solutions for many groups.
hammer percussion is the strategy to be preferred. The only disad¬ vantages are that using hard-hammer percussion on core tools: (1)
The resharpening modes
is, as a rule, wasteful of raw material since the resharpening flakes
Hard-hammer percussion
tend to be thick, with wide platform (edge) remnants and (2) it
Of all the resharpening modes, hard-hammer percussion
tends to create progressively steeper edge angles with every reshar¬
used on core tools is the easiest and fastest. Since percussors can
pening, thus rapidly creating edge angles too steep for effective use
be obtained anywhere where there are rocks, no special equipment
(fig. 2.2a). However, under conditions where cutting requirements
is required; furthermore, elaborate preparation of platforms is
are extremely low, these considerations will not be important as
Brian Hayden
12
long as a reasonable supply of raw material is accessible. These
smoothing and straightening of spear shafts, (3) more frequent
tools are resharpened simply to avoid the effort of searching for replacements.
kills of large game necessitating heavy tools for some aspects of butchering (Hucknell (1979) has shown that bifaces used in butch¬
This scenario fits well with what we know of early Palaeo¬
ering large game need to be resharpened every four minutes on av¬
lithic assemblages. In the first place, it is difficult to imagine very
erage), or (4) by the addition of other types of wooden tool, such
many cutting requirements at this level of technological evolution.
as bowls or simple throwing sticks.
From a review of the extant literature on stone tool use among
Initially, increased cutting requirements might be accom¬
hunter-gatherers (Hayden 1977b, p. 182; 1979) and from what can be surmised about the nature of resources used at this date, it
would tend to maintain a constant edge angle longer than unifa¬
seems highly unlikely that stone tools would have been required in
cial retouch (fig. 2.2b). More of the perimeter of tools might also
modated by hard-hammer bifacial resharpening techniques which
early Palaeolithic times for procuring or processing plant foods, or
be resharpened. These modifications, however, would only mini¬
for working skins, or for any other task beyond the making of
mally extend the use-life of such tools. The innovation which had
spears, digging sticks and shelters, and the initial gutting of ani¬
a dramatic impact on the use-life of core tools was the use of soft-
mals. In the latter task, bipolar flakes produced from small pieces
hammer percussors. In fact, use-lives of tools resharpened by this
of common raw materials would generally satisfy needs (e.g. Mac-
technique potentially could have been extended to such an extent
Caiman and Grobbelaar 1965). Moreover, the rudimentary spears
that no major change in resharpening techniques was required for
may have been used rather infrequently due to the as yet unperfec¬
nearly a million years.
ted biological adaptation to big game hunting (e.g. Hayden 1981).
The major technical advantage is that soft-hammer per¬
Using data from Australian hunter-gatherers, it can be estimated
cussion enables the stoneworker to remove broad, flat, thin flakes
that these wooden tools would probably only have to be replaced
with narrow platforms or edge remnants. As Hucknell (1979) has
about once a month (see Hayden 1977b, p. 182; 1979), or even less frequently, and spear tips would require correspondingly less re¬
demonstrated, this means not only that many more resharpenings
sharpening.
is possible to maintain the desired edge angle virtually throughout
During the Oldowan, there is little doubt that core tools
can be obtained from a given amount of material, but also that it the entire series of resharpening events (fig. 2.2c). The disadvan¬
were carried about from site to site, or that suitable material for
tages of the technique are first, that a narrower range of lithic mat¬
making core tools was accessible at least a few times within the
erials are appropriate for soft-hammer resharpening. Secondly, it
annual rounds of most foraging groups. At Olduvai Gorge, the
requires special equipment to carry out the resharpening; and, fin¬
nearest sources of raw materials have been reported about ten kil¬
ally, considerable care often has to be taken in the act of reshar¬
ometers from sites (Rensberger 1974, p. 52; Isaac et al. 1971).
pening, including special platform preparation involving small,
Most nomadic groups at this technological level would probably
but structurally important amounts of material. Because really
never have been much farther than ten to twenty kilometers from
good soft-hammers are not readily available in most localities (or
the nearest sources of suitable stone since a wide range of rock
if they are, as in the case of antler or wood, they often require con¬
types would have been appropriate for the simple cutting require¬
siderable effort to be transformed into usable forms), resharpening
ments of this period.
equipment has to be carried with individuals on a regular basis.
I am confident that the cutting requirements postulated above for the Oldowan could be easily accommodated by using core tools replenished every three or four months, and reshar¬
This cost, however, is still less than the cost of making special trips to distant lithic sources to replenish exhausted supplies. Before considering examples of this resharpening mode in
pened by the hard-hammer percussion technique. Similar stra¬
more detail, it should be pointed out that a number of flint knap-
tegies could be expected to occur in environments where cutting
pers have denied that there is any difference between hard-
requirements were similarly restricted. For example, desert en¬
hammer and soft-hammer percussion (Mewhinney 1964). In the
vironments are often characterized by very limited cutting require¬
instances with which I am personally familiar, these claims have
ments due to the great travel distances required and the subse¬
primarily rested on a confusion between percussion using stones
quent limitations placed on wooden material items which families carry with them.
where (Hayden and Hutchings 1986), the important factor is not
and percussion using antler or wood. As I have emphasized else¬ whether stone or organic materials are used as percussors, but the
So ft-hammer percussion When cutting requirements increased to the point that for¬ aging bands regularly began to run short of stone tools before
nature of the area of contact and the hardness of the percussor material (cf. Tsirk 1979). The most commonly cited example of soft-hammer reshar¬
being able to return to lithic sources, it would be entirely expected
pening involves the more developed handaxe industries. I feel that
for groups to begin to employ alternative, less wasteful, resharpen¬
many of these tools, especially the early forms, were primarily
ing techniques. Increased cutting requirements might have emerged because of(l) the ongoing biological adaptation of hominids to hunting resulting in better stalking and throwing skills and the more frequent use of spears, (2) the use of more refined types of spears requiring not only pointing the end of sticks, but also the
used for heavy wood-cutting, although they may have also be¬ come important in butchering and skinning large animals in later times when hunting provided more reliable and regular subsis¬ tence fare (cf. Keeley 1980, p. 161, p. 169). In later time periods, when a variety of activity-specific tools were common, soft-
From chopper to celt: the evolution of resharpening techniques
13
hammer resharpening persisted in certain areas, and may have been useful primarily in the context of butchering (e.g. as bifaces in Palaeo-Indian, numerous American Archaic, and Solutrean complexes). Although no quantified data exist concerning the rela¬ tive increases in cutting capacity per kilogram of tools reshar¬ pened by hard versus soft-hammer percussion, nor concerning the relative costs involved in using each of these resharpening modes, I am confident that soft-hammer percussion is more effort inten¬ sive and considerably more material conservative than hardhammer percussion. Moreover, resharpening with soft-hammer percussion produces many more flakes which can be used for
complete explanation capable of accounting for the European and Asian differences, I would suggest that the problem can be usefully approached in terms of a model of resharpening needs and stra¬ tegies. Additional advantages of using flake tools, especially adzes, are that suitable cutting edges can be pre-selected for transport. Disadvantages include greater loss of pieces, difficulty in trans¬ porting pieces without containers such as bags, and, when using hafts, the effort required to procure suitable lithic and mastic materials for use in hafting, as well as in making and maintaining hafts (Hayden 1977a). Because of these considerations, hafted tools tend to be repeatedly resharpened until their use potential is largely exhausted (Hayden 1979). Rolland (1981, pp. 33-4) has recently argued that the need to economize raw material explains important aspects of the va¬ riability in Mousterian assemblages. I would extend his argument to include the development of blade tools. In addition to provid¬ ing much more cutting edge per kilogram of material, blade tools with working edges at their distal ends (e.g. end scrapers, burins, perforators, borers) could be resharpened over and over again much like repeatedly sharpening a pencil, and they were easily haf¬ ted. Flake tools did not generally have these advantages, although they required much less skill and care to produce. Perhaps the major disadvantage of using blades involves the skill and training required to produce them. Franqois Bordes (personal communica¬ tion) once mentioned that after seasons when he made no blades, it took him about two weeks of practice before he could remove
various other tasks. Flake and blade resharpening
While core tools undoubtedly were used for the manufac¬ ture, shaping, and resharpening of wooden-shaft tools such as spears and digging sticks, as well as for manufacturing bulkier items such as bowls, flake tools constitute alternative tool forms that could be used for many aspects of the manufacture and main¬ tenance of wooden tools. The adoption of flake tools for wood¬ working and other tasks could have had several advantages in terms of reducing energy spent in procuring lithic resources. First, much more usable cutting edge could be obtained per kilogram of raw material (Leroi-Gourhan 1964, pp. 190-97; Butzer 1971, p. 475) and this could be considerably extended via blade technology (Sheets and Muto 1972). Secondly, cutting edges on flakes and blades themselves could generally be resharpened up to two, three or even more times. Finally, if flake tools could be hafted as adzes, the increase in use-life and cutting activity per kilogram of raw material would have increased even more dramatically (Hay¬ den 1979). In Europe, some flake tools had probably always been used for the manufacture or maintenance of different types of perish¬ able tools. However, it was only with the Middle Palaeolithic that pronounced specialization and elaboration of flake tools began to take place. This development was possibly due to the advent of hafted adzes and bifaces (Anderson 1980, p. 33; 1981, pp. 79-80, p. 88; F. Bordes, personal communication) replacing the heavier core tools in many woodworking tasks, as in many parts of Aus¬ tralia. Hafting was, of course, extensively used during the Middle Palaeolithic Aterian of North Africa. The increasing dominance of flake and blade tools may have resulted from a substantial elaboration of the number and di¬ versity of wooden-shaft tools in the technological repertory of the Middle Palaeolithic as well as increased importance of other tasks for which flake tools would be useful, e.g. hide working. Flake tools would have been especially suited to the manufacture of wooden shaft implements including tent poles, hafts, and tools used in snow and ice conditions. A similar increase in importance offtake stone tool also characterized Southeast Asia (Hayden 1977a), and Australia (Lampert 1981, pp. 160-64), although in these areas changes in cutting requirements were undoubtedly of a different magnitude and nature, probably more related to the types of wooden-shaft tools that were being added to various tech¬ nological repertories or improved upon. While I do not have a
good blades consistently. Unfortunately, there is little empirical data than can be used to gauge the relative costs and benefits of the flake (and blade) resharpening strategy. It is more complex than other strategies, due to the frequent co-existence of task-specific tools, and several basic resharpening modes often exist in single assemblages. More¬ over, there is little data on the comparative economy of effort or material involved in using flake tools as opposed to either hardhammer resharpened or soft-hammer resharpened core tools. Good data does exist on the use of adzes (see Hayden 1979), and here it is clear that hafted flake tools use much less material than flaked core tools, and that the manufacture and maintenance of' hafted tools is considerably more costly in terms of time and eff¬ ort. On this basis alone, one would expect core tools to be used where material was abundant and hafted flake tools to be used where suitable lithic resources were scarce. However, other factors also play important roles: (1) amounts of trauma associated with the two tool types; (2) relative amounts of energy expended in work; (3) the amount of control in working wood; and (4) cultural prohibitions against women using hafted or exotic stone tools. Pressure-flaking
At the end of the Palaeolithic, when many activity-specific tools were being developed and various techniques were being de¬ vised to extend the use-life of tools in accordance with the cutting and edge requirements of each task, a fourth resharpening strategy emerged for the first time in both the New and the Old World pressure-flaking.
Brian Hayden
The advantages of resharpening tools (including flakes and
14
entire archaeological epoch, the Neolithic. Yet, the question of
blades) by pressure-flaking clearly include (throughout a series of
why edge-grinding emerged has rarely if ever been addressed.
resharpenings) very little raw material removal; edge angles cap¬
Assumptions have often been made that it reflected a major ad¬
able of being held constant; greater accuracy; and less risk of
vance in the intellect of the human species, that it was especially
breakage than direct percussion resharpening techniques. The re¬
useful in coastal economies, that it was somehow vaguely connec¬
sult is long-lasting, lightweight tools. For any nomadic group
ted to the agricultural revolution, that it diffused into local areas
which had considerable baggage to transport and/or who rarely
from more advanced areas, and even that it was discovered by
had access to good stone sources, pressure-flaking would be highly
some prehistoric genius or committee grinding bones on a beach
advantageous. There is evidence to suggest that it was used dif¬
one sunny day from whence it diffused throughout the world!
ferentially, at least in some locations, according to distance of the
(Clark 1980, 1982; Childe 1925, p. 13; Ritchie 1969). The general
group from quarry sources (Sollberger 1971; Walker 1978). Stem¬
question has been given superficial attention in Australia, where
ming and/or notching indicate that pressure-flaked knives were
edge-ground cutting tools occur in some areas. However, the ex¬
generally hafted. They often resembled or even functioned at times
planations are no more satisfactory than traditional European ex¬
as spear points (Ahler 1971) and would have been most useful in
planations. The Australians largely adhere to diffusionist models,
butchering and skinning tasks as well as light woodworking tasks,
or postulate that edge-ground axes were necessary to make canoes
including resharpening wooden spear tips. In addition, broken
or climb large trees (Helms 1890, p. 274; Tindale 1950, p. 270;
stone spear points could often be retipped by pressure-flaking,
Davidson and McCarthy 1957, p. 427). But to my mind, satisfac¬
thus extending their use-lives.
tory arguments have never been tendered or tested.
While the advantages of pressure-flaking in terms of mat¬
For my part I was convinced that the traditional explana¬
erial conservation are striking, there are also some major disad¬
tions regarding the advances in intellect had little basis in fact
vantages. In the first place, special tools are necessary for carrying
when my three-year-old son sat down one day with several pieces
out resharpenings using the pressure technique. Raw material for
of slate and proceeded to unifacially and bifacially grind some of
flakers must be obtained and then properly shaped. In order to
them, just as many young children often grind stones when play¬
minimize these costs, flakers would have to be transported on a
ing. My son then presented them to me and explained that they
regular basis. In addition, special lithic raw material would have
were to be used as knives. He had seen chipped stone knives be¬
to be obtained for use in the manufacture and maintenance of
fore, but never any ground stone knives; nor had I ever manufac¬
pressure-flaked tools. Generally, only fine-grained, highly iso¬
tured any. I reasoned that if a three-year-old child could develop
tropic stone types can be successfully pressure-flaked. Furthermore,
edge-ground cutting technology, it must have also been within the
whereas other resharpening modes require only a modicum of
capacity of early Homo sapiens, and undoubtedly Homo erectus, if
skill, successful resharpening by pressure-flaking requires con¬
not Homo habilis or Australopithecus. If this was the case, then
siderable training and ability. The production and resharpening of
what made it useful for Neolithic people to resharpen their tools
edges is also time consuming. Because of all these disadvantages, it
by grinding edges but not for Palaeolithic people to do the same?
seems reasonable to expect the resharpening of cutting tools via
Fortunately, there are some observations pertinent to the
pressure-flaking to occur only in very material-strained circum¬
general problem e.g. Davidson and McCarthy (1957). In the first
stances, or situations in which the constraints placed on the
place, maintaining cutting tools by edge-grinding involves a con¬
amount of material baggage are unusually pronounced. I would
siderable amount of work. Some of my Australian informants
argue that these are precisely the conditions under which the
even went so far as to describe their fathers as “always” in the
pressure-flaking of cutting tools can be observed to occur (e.g.
process of resharpening their axes. In the second place, most tradi¬
Palaeo-Indian, American Plains Archaic, Desert Archaic, Eskimo).
tional edge-ground axes are not significantly more efficient in cut¬
Once again, quantitative data is largely absent concerning
ting wood than chipped stone equivalents. Third, edge-ground
the amount of cutting activity that can be performed per kilogram
axes did not occur universally throughout Australia. In fact, their
of raw material using this mode of resharpening, just as data is
distribution at the time of contact is very interesting. Edge-ground
largely absent on the energy expenditures necessary to undertake
axes occurred in those areas of Australia with the most complex
this form of resharpening, although Walker (1978) has performed
set of wooden tools, requiring the greatest amount of heavy wood¬ cutting activity (fig. 2.3).
some relevant experiments comparing the efficiency of bifaciallyflaked tools to that of unretouched flakes. Despite the paucity of
Having noted this relationship, it became clear that in every
data, relative assessments for the variables involved create a con¬
major culture area in the world where edge-grinding adzes or axes
vincing picture of the merits of pressure-flaking as opposed to the
had become important, wood-cutting requirements were unusu¬
other modes discussed thus far.
ally high. For example on the northwest coast of North America, edge-ground tools became established at the same time that hun¬
Edge-grinding
ter-gatherers were settling down into semi-sedentary villages,
Perhaps the clearest example of the importance of cutting
building substantial structures of wood, and undoubtedly ela¬
requirements in determining resharpening strategies involves
borating an entire array of wooden primitve valuables and sculp¬
edge-grinding. Remarkable as it may seem, edge-grinding has ser¬
tured wooden monumental architecture. During the Neolithic in
ved for the better part of a century as the defining attribute for an
Europe, not only the implements used to procure food would have
From chopper to celt: the evolution of resharpening techniques
15
place to such an extent that vast amounts of resources had to be cut and processed, often within very limited time periods. On the northwest coast of North America, enormous numbers of salmon had to be butchered, split and dried within a few weeks. In many places, it would have been extremely inefficient to continuously stop this processing to procure lithic raw materials to replace ex¬ hausted butchering tools; similarly, advance caching of lithic re¬ sources would have involved excessive search and transport time and effort for unreasonably large amounts of material. Instead, ground slate knives that could be resharpened seemingly in¬ definitely were used. The same activities would have consumed enormous amounts of chipped stone materials. Even with metal knives, only about twenty fish can be processed (about an hour’s work) before the knives need resharpening (George MacDonald, personal communication). Eskimo groups which faced similar time constraints in processing large numbers of fish and caribou — — Apparent distribution of throwing clubs boomerangs Distribution of edge ground axes
also developed edge-ground tools (ulus). If we attempt to plot the edge per kilogram of raw material
Fig. 2.3. Australia: the distribution of wooden tools requiring in¬ tensive use of stone tools in their manufacture (throwing clubs and boomerangs) and the distribution of edge-ground axes (from Hay¬
for edge-ground tools, the values would be many orders of mag¬ nitude above any other resharpening mode since it would be
den 1977a)
other resharpening technique that can come anywhere close to
been fashioned out of wood, but permanent or semi-permanent structures would have been built and, most importantly, substan¬ tial numbers of trees would have had to have been felled to create
necessary to count each resharpening as a new edge. There is no such an efficient use of raw material, although the cost in grinding energy can be quite high, especially after edge damage (Dickson 1981).
agricultural plots. The critical factor determining whether or not edge-ground tools were used was not the level of human intell¬ igence, not whether groups were agricultural or not, but the level of their wood-cutting requirements. Under conditions where wood-cutting requirements were
Conclusions Of all the postulated relationships in the above model, the conditions leading to edge-grinding appear to be the most amen¬ able to investigation and testing. Edge-grinding should be found
not extreme, various less efficient resharpening strategies could be
among groups with specialized and intensive cutting requirements.
used to prolong the life of cutting tools until more raw material
The precise nature of these requirements among prehistoric
could be procured. However, where cutting requirements were ex¬
groups can sometimes be a difficult question to deal with, but it is
tremely high, it would simply become too costly in terms of effort,
one that bears at least a good promise of being answered not only
time, and scheduling to have to return to quarry sites at short in¬
for edge-ground technologies, but for other technologies as well.
tervals to replace exhausted tools or materials. It was only under
For instance, Southeast Asia was characterized by a slightly dif¬
relatively extreme conditions that edge-grinding, the most energy
ferent overall progression in resharpening techniques than Eur¬
intensive resharpening technique of all, was viewed as useful. A
ope. In Southeast Asia, the general sequence was as follows: hard-
particularly good example of this development is provided by the
hammer percussion core tools, increased flake tools, and edge-
successive changes of celts and adzes in eastern North America.
ground tools. In contrast, in Europe the sequence was different:
The earliest forms are chipped from siliceous materials, as with the Dalton adze (Morse and Goodyear 1973). These forms are
choppers, soft-hammer percussion bifaces, increased flake tools,
followed by the appearance of non-siliceous tools flaked into
ground tools. What I propose is that these differences can be acc¬
shape but with ground bits, as in the Kirk horizon (Broyles 1971;
ounted for by the nature and the frequency of cutting activities un¬
see also Fitzhugh 1972). This in turn is followed by the appearance
dertaken within each of the areas. European groups engaged in¬
of fully ground tools in the later time periods.
tensively in hunting, hide preparation and softening, and they de¬
I hasten to add that the grinding technique of resharpening
increased blade tools with some pressure-flaked tools, and edge-
veloped appropriately edged blade tools which could be sharpened
tools was adaptive for a wide variety of cutting activities wherever
repeatedly while used in these tasks. Partially due to climatic
and whenever they were carried out at intensive enough levels. For
and faunal differences there may have been no comparable em¬
example, in some American Archaic and even contemporary hun¬
phasis on hunting or the processing of skins in Southeast Asia,
ter-gatherer groups, specialization of food procurement took
and perhaps as a result no elaborate blade technology was em-
Brian Hayden
16
ployed. Similarly, Palaeolithic Europeans, who sometimes needed to harvest large numbers of herd animals, may have found billetflaked core tools, hafted-flake tools, and pressure-flaked cutting implements, such as the Solutrean “points,” to be useful solutions to the problem of conserving raw material. In contrast, since large numbers of animals were not killed or processed in short spans of time in Southeast Asia, the conservation of stone material may not have been as important, thereby obviating the need to develop soft-hammer percussion and pressure-flaking techniques. Alter¬ natively, the lack of billet-flaking may have been due to the em¬ phasis placed on different types of wooden implements in the two areas, or on other factors such as the use of bamboo knives for butchering instead of stone tools, or the greater seasonal avail¬ ability of raw material. The fine details obviously still remain to be worked out. However, it seems likely that at least major parts of the present model closely approximate the underlying causal factors respon¬ sible for some of the more fundamental and pervasive changes in
EFFORT COST PER RESHARPENING (INCLUDING COSTS RELATED TO TOOL PREPARATION AND CREATION OF RESHARPENING TOOLS)
technology. The increasing importance of reduction and reshar¬ pening techniques capable of conserving material to greater de¬ grees as the conservation of lithic material becomes a greater con¬ cern may also be reflected in reduced debitage size as well as in changing tool forms and resharpening techniques (e.g. Matson 1976, p. 185). The resharpening techniques that have been discussed here can be plotted with respect to these variables, as in figure 2.4. In general, increased cutting requirements over time probably con¬
Fig. 2.4. The general relationship between various resharpening techniques, the relative energy costs involved in using the techniques, and the degree of raw material conservation implied by the use of each technique
In spite of the relative uncertainty of flake tool resharpening in this overall scheme, the strategy has been included in the present discussion in order to probe the limits of the model. At the ex¬ tremes, it appears generally sound. It is primarily in the inter¬
stitutes the single most important factor determining changes in
mediate areas that the details appear less certain. Moreover, the
lithic resharpening strategies. The exact status of some of these
approach advocated here promises to provide considerable in¬
strategies, such as the resharpening of flake tools, is, however, still
sights into the overall, and sometimes specific, subsistence activi¬
rather uncertain. It may eventually turn out to be the case that
ties of prehistoric communities, including the intensity in which
other considerations must be taken into account to fully explain
they were performed since each resharpening strategy is tied to
the emergence of each strategy. Here, I am reminded of Tasma¬
specific levels of task intensity and frequency. If the basic model is
nian flake tool assemblages where core tools are not common, and
supported by subsequent work, these considerations will be critic¬
yet where the wooden tool repertory was perhaps the most limited
ally important in our understanding of the evolution of stone- and even metal-using cultures of the past.
of recent hunter-gatherers. Is Jones’ (1977; 1978) argument about the inevitable pauperization of cultures in small isolated habitats correct in regard to Tasmanian wood technology? Or do other fac¬ tors make the emphasis on flake tool resharpening more useful than an emphasis on core tool resharpening? The situation must remain somewhat nebulous for the time being.
Acknowledgments I would like to thank Robin Torrence for her interest and critical support, Jon Driver for help with background data, Chris Ellis for sharing his extensive knowledge, and John Clark, Morley and Anne Eldridge for their interest and helpful comments on an earlier draft of this paper.
17
Chapter 3
The occupational history of sites and the inter¬ pretation of prehistoric technological systems: an example from Cedar Mesa, Utah
Eileen Camilli
Models of technological systems based on ethnographic observation
the archaeological record itself be recognized and taken into ac¬
provide a partial framework for developing general theories of
count. One step towards controlling for some of the variability in
human behavior which can be applied to stone tools. In particular,
the archaeological record is to devise analytical methods which
the understanding of such concepts as personal gear, site-specific
identify the behavioral processes that bring into being the static
facilities and situationally manufactured tools is especially import¬
patterns observed in archaeological remains.
ant. Before these functional classes of tools can be linked with pat¬
The structure of the assemblage of tools in terms of the as¬
terning in the archaeological record, however, the causes of differ¬
sociations and correlations of its contents is one type of such pat¬
ences among stone tool assemblages must be known. The roles of
terning. Redundant kinds of assemblage structure should permit
two factors in determining assemblage variability - a site’s occu¬
the recognition of at least some aspects of the organization of the
pational history and its function
cultural system which produced it. The causes of variability in
-
are explored here by means of
analyses of the spatial structure of sites and the composition of tool assemblages.
assemblage structure among assemblages in a subsistence-settle¬ ment system have generally been assumed to stem from differences in the way places were used. Such differences among locations,
Introduction Most of the contributions to this volume are attempting to
however, could also represent not only the archaeological conse¬ quences of functional differences in activities performed at lo¬
account for the social and economic causes of technological be¬
cations but also variations in the histories of the way sites were
havior. This paper takes a different approach: methods which con¬
used. In other words, archaeological assemblages cannot be
tribute to the evaluation of general theory about stone tool manu¬
treated as the frozen consequences of one episode of activity, as is
facture and use are proposed. A requirement of such methods is
often assumed by researchers guided by ethnoarchaeological ob¬
that they link the concepts and definitions incorporated in ex¬ planations of technological behavior with the remains in the ar¬
servations of simple events. In settlement systems which employ technologies organized
chaeological record. This linking process is critical since it
in an efficient manner, strategies for the production and main¬
provides the only means for evaluating general theory. If evalu¬
tenance of many tools (e.g. the personal gear and site-specific im¬
ation procedures are to be reliable, however, it is essential that the
plements defined by Binford 1977) will vary among functionally
sources of variability among stone tool assemblages stemming
distinct locations, while casually manufactured tools might occur
both from the character of past behavior and from the nature of
in all assemblages. Evaluating expectations for the ways in which
Eileen Camilli
18
tools contribute to assemblage structure at places used for dif¬ ferent purposes therefore requires knowledge of the potential of lithic data for reflecting these different foci of settlement. In ad¬ dition to the nature of stone tool production and maintenance of items, the history of how a location was used establishes the num¬ ber of items present in assemblages as well as the relationship between their relative frequencies. An investigation of the struc¬ ture of functionally distinct assemblages, therefore, also neces¬ sitates lines of inquiry into the occupational character of sites. Further, if this process is to be reliable in a scientific sense, then it must utilize data which are independent of the artifacts them¬ selves. History of site occupation
Fig. 3.1. Comparisons between models of artifact distributions
An occupation has been defined as “the uninterrupted use of
produced by different occupations and the resultant visible arch¬
a place by participants in a cultural system” (Binford 1982, p. 5).
aeological record
Archaeological analysis, however, does not focus directly on occu¬
wet season, residential camps of the Dobe San are rarely reoc¬
pations; instead, the unit of virtually all archaeological analysis is
cupied after several months both because of infestation of old huts
the site. Sites are places made up of spatial associations of artifacts
by insects and because gathering is more difficult in areas
and features which result from one or more uses of a place. Conse¬ quently, their occupational history is not necessarily self-evident
previously exploited. The site of Dobe, the only permanent water source, however, is reoccupied during the dry season. Kua for¬
from an inspection of the contents of the assemblage present or of
agers in the eastern Kalahari have been attributed with similar be¬
the characteristics of the spatial distributions of artifacts. Hence,
havior (Hitchcock 1982, p. 331).
the degree to which assemblages from sites are isomorphic with the remains generated by a single occupation is unknown. Ethnographic examples provide insights into a pattern of
Other ethnographic accounts report reuse in terms of multi¬ ple, overlapping occupations which have been generated during a number of annual cycles. Binford (1978a, p. 491) had described
land use in which residential camps are reutilized for different
occupations of different types in the vicinities of Nunamiut vill¬
special purpose functions. Several patterns of land use have been
ages. This pattern results from residential occupation of the loca¬
recognized (Binford 1980, 1982; Foley 1981). With patterns
tion during several different seasons combined with special pur¬
characteristic of a logistically organized system, a residential group
pose use at other times. Hitchcock (1982) describes the same situa¬
is positioned with respect to known distributions of food, water
tion for the Kua groups who, when conducting hunting trips from
and fuel resources (Binford 1980; 1982). Special purpose task
seasonal camps, will seek out campsite locations in the vicinity of
groups undertake activities which are partitioned into sets of tasks
abandoned residential camps. Gould (1980, pp. 1-28) has also re¬
and carried out at different locations. Economic zones around the
corded a similar pattern of site reuse for the Ngatatjara of western
residential camp are exploited by different mobility strategies. A
Australia. At the waterhole of Tika-tika residential camps were
small area around the residential camp is exploited by means of
repeatedly constructed at different seasons and the site was also
short foraging trips. This inner zone is surrounded by a logistical
used for outlying temporary camps of male visitors and hunting
zone into which special resource procurement trips of longer dura¬
parties. The archaeological manifestation of this behavior is a
tion are made from the base camp. In the inner, foraging zone
number of “weak” artifact clusters over a large area (Gould 1980,
archaeological sites representative of activities involving extrac¬
p. 27).
tion are generated. Types of sites in the logistical zone include field camps, observation stations and caches.
Considering the above examples, the general spatial distri¬ butions of artifacts found on locations with the different occu¬
When residential camps are moved, specific locations for¬
pational histories which are associated with mobile settlement sys¬
merly surrounding the old camp can be reused in different ways
tems can be anticipated (fig. 3.1). Essentially, the manner in which
since the foraging and logistic zones are also shifted. One result of
groups move from place to place insures that some locations will
this land use pattern is that former base camps which now fall into
contain archaeological assemblages that are structurally complex.
the logistic zone of the currently occupied site can be reused by
Since the rate of natural deposition may or may not coincide with
special task groups. Positioning of a residence with respect to
that of cultural discard, however, assemblages resulting from
known distributions of food, fuel and water also raises the possi¬
single events, activities or occupations cannot necessarily be
bility that future residential bases will be located on the sites of
isolated through inspection of horizontal artifact distributions.
present special use camps. Although residential camps can potentially be reused for other functions, researchers have ob¬ served that for mobile systems which move the base camp fre¬
Variation in the structural properties of sites In the following section, the spatial patterning in arch¬
quently, locations occupied in connection with foraging are not
aeological artifact distributions is assessed and expectations are
usually reused. For example, Yellen (1977, p. 67) has noted that
developed for the structural properties of locations with histories
The occupational history of sites
19
Fig. 3.2. Locations of intensively surveyed areas and of the watershed units sampled by the Cedar Mesa Project
of single and multiple use. The discussion of multiple use will be facilitated by two terms. Reuse of a general area but not necess¬ arily reoccupation of a site location itself is one kind of history of
presence of different settlement components which may be func¬ tionally distinct - for example, residential and special purpose camps - or which may comprise different occupational mixes. The site sample was selected from a seven per cent surface
use. The resulting spatial pattern is one of a number of overlap¬ ping distributions of items which have been generated during separate occupations. Multiple occupation is used here to describe this type of spatial pattern. Actual reuse of houses, storage facili¬ ties or processing features at different times is another pattern. Facility reuse tends to focus activities and may, in contrast to mul¬ tiple occupations, result in stratigraphically superimposed artifact and feature distributions. Reoccupation is used to describe the lat¬ ter spatial pattern. In order to illustrate how archaeologists can begin to infer the occupational history of a site, the remainder of the paper will focus on the analysis of sites from Ceda Mesa, a 500 square kilo¬ meter area in southeastern Utah (fig. 3.2). Between 1972 and 1975 mesatop environments of the Ceda Mesa highland were systemati¬ cally surveyed for archaeological surface remains (Matson and Lipe 1975; 1978). Results of previous survey and excavation in¬
survey of five drainage units on Cedar Mesa (fig. 3.2) comprising an area of 12.48 square kilometers. Sites in the seven per cent sur¬ face sample include aceramic lithic scatters with associated fea¬ tures interpreted by Matson and Lipe (1975, p. 3) as either Basketmaker II occupations (AD 200-400) or sites belonging to an un¬ known period of occupation. In order to provide a framework for the examination of variation in site size, sites were initially se¬ gregated into three architectural classes: (1)
sites lacking surface indications of features;
(2)
sites containing only unlined ash hearth and midden feat¬
ures; (3)
sites containing features of sandstone slab, jacal or masonry
construction including hearths and cists. Histograms depicting the distribution of site size in square
dicate that the area was initially occupied between AD 200 and
meters were compiled for each of the classes. Inspection of the re¬
400. Occupation continued until about AD 1250 with an occu¬
sults portrayed in figure 3.3 reveals a striking pattern; each of the
pational hiatus dated to between AD 700 and 1000 (Matson and Lipe 1975, pp. 2-3). The analysis will focus on surface collections
distributions has a tendency toward bimodality. In group I, com¬ prised of sites between one and fifty square meters, site size ap¬
of lithic material lacking associated ceramics. This material is be¬
pears to be normally distributed around a mean of about twenty-
lieved to date to the initial occupation of the region by seasonally
five square meters: above fifty square meters the measurements of
mobile horticulturalists.
size do not mimic a normal distribution. Instead, the histograms for each of the architectural classes are skewed to the right with
Site size Variation in site size forms the basis for interpreting the
modal site area between sixty and eighty square meters: these are labelled group II sites. Visual interpretations of the statistical
occupational character of Cedar Mesa sites. The assumption here
character of histograms for site size can be reinforced through the
is that formal patterning in the extent of a site is indicative of the
use of the Shapiro-Wilk W statistic (Shapiro and Wilk 1965).
Eileen Camilli
20
Group I
Group II
250
200
450
Area (m2)
Fig. 3.3. Group I and II site area distributions for sites in three architectural classes
Small values of W indicate non-normality. The values of W de¬
spatial properties of distributions resulting from different occu¬
rived for group I site area distributions are not significant; that is,
pational histories. The following discussion of the spatial charac¬
they do not indicate non-normality at the 0.05 significance level
teristics of artifact and feature distributions considers the impli¬
(table 3.1). In contrast, all group II distributions were found to be
cations of assemblage size and density given different occupational
non-normal at the 0.01 significance level.
histories.
A normal distribution of site area suggests a regular or modal pattern in the extent of space use, since it indicates that site
Artifact frequency and scatter size
size has an equal probability of falling on both sides of a mean
Multiple occupation of an area for residential and special
area. This pattern may be evidence that group I sites represent the
use functions has been noted as a characteristic of mobile settle¬
area required for single occupations; in other words, small scatters
ment systems. The spatial distributions of artifacts and features
may form basic units which reflect single activity areas. The
formed by the spatial overlap of debris from separate, multiple
skewed form of the group II site area distributions indicates a
occupations will be larger in areal extent than those representing
lower probability for the occurrence of the largest sites in this
single occupations, other things being equal. Since larger, multiple
group. If group II sites represent spatial combinations of debris
occupations could be expected to represent the spatial combin¬
from separate occupations, the form of the distribution may mean
ation of greater numbers of debris scatters, they should contain
that there is less chance for the spatial combination of debris from
larger amounts of material. For sites occupied repeatedly, the site
large, as opposed to small, numbers of single occupations. Inves¬
area and artifact frequency should, therefore, be strongly related.
tigations of the occupational character of group I and group II sites will be approached further below using expectations for the
In the case of reoccupation, the relationship between site area and abundance of debris deposited is dependent on the de¬ gree to which additional space requirements accompany different occupational episodes. With strictly superimposed distributions, the size of the area utilized may not vary while the amount of *
Table 3.1. Site sizes.
debris may increase, and there need be no necessary relationship Architectural category
n
Mean
W
between site size and quantity of debris. In contrast, for reoccu¬
P
area (m2)
pations with additional space requirements, site size and debris frequency should increase together. If, on the other hand, sites re¬
Group I sites
present occupations at which variation in the amount of material
No Features
44
25.27
0.97
0.688
Ash Features
20
25.50
0.95
0.455
Slab Features
15
26.86
0.96
0.625
present is caused by the differences in activities and situational conditions for the performances of those activities, relationships between the areal extent of scatters and artifact frequency would not be direct. All of the above expectations are, of course, depen¬
Group II sites No Features
24
198.20
0.49