Nathan Goodale & William Andrefsky 2015

Lithic Technological Systems and
Evolutionary Theory
Stone tool analysis relies on a strong background in analytical and meth-

odological techniques. However, lithic technological analysis has not been
well integrated with a theoretical approach to understanding how humans
procured, made, and used stone tools. Evolutionary theory has great poten-
tial to fill this gap. This collection of essays brings together several different
evolutionary perspectives to demonstrate how lithic technological systems
are a byproduct of human behavior. The essays cover a range of topics,
including human behavioral ecology, cultural transmission, phylogenetic
analysis, risk management, macroevolution, dual inheritance theory, cla-
distics, central place foraging, costly signaling, selection, drift, and various
applications of evolutionary ecology.
Nathan Goodale is Assistant Professor of Anthropology at Hamilton College.

He is the author of articles and book chapters dealing with lithic technology
and evolutionary theory in several journals and edited volumes, including
Evolution: Education and Outreach, American Antiquity, Journal of Archaeological
Science, Complex Hunter-Gatherers (2004), and Lithic Technology (Cambridge
University Press, 2008).
William Andrefsky, Jr., is Edward R. Meyer Distinguished Professor of

Anthropology and Dean of the Graduate School at Washington State
University. He is the author of several books dealing with stone analy-
sis, including Lithics (Cambridge University Press, 1998 and 2004), Lithic
Debitage (2001), and Lithic Technology (Cambridge University Press, 2008).
Lithic
Technological
Systems and
Evolutionary
Theory
Editedby
Nathan Goodale
Hamilton College
William Andrefsky, Jr.

Washington State University
32 Avenue of the Americas, New York, NY 10013-2473,USA
Cambridge University Press is part of the University of Cambridge.

It furthers the Universitys mission by disseminating knowledge in the pursuit of
education, learning, and research at the highest international levels of excellence.
www.cambridge.org
Information on this title: www.cambridge.org/9781107026469
Cambridge University Press2015
This publication is in copyright. Subject to statutory exception
and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without the written
permission of Cambridge University Press.
First published2015
Printed in the United States of America
A catalog record for this publication is available from the British Library.
Library of Congress Cataloging in Publicationdata
Lithic technological systems and evolutionary theory / [edited by] Nathan Goodale (Hamilton College),
William Andrefsky, Jr. (Washington State University).
pages cm
This volume is an outgrowth of a symposium organized for the 74th Annual Society for American
Archaeology meeting in Atlanta, Georgia, titled Evolutionary Approaches to Understanding Stone
Technologies as a Byproduct of Human BehaviorContents page.
Includes bibliographical references and index.
ISBN 978-1-107-02646-9 (hardback)
1. Stone implements Analysis Congresses. 2.Tools, Prehistoric Analysis. 3. Human evolution
Philosophy. 4. Social archaeology. 5. Human behavior History. 6. Human ecology History.
I. Goodale, Nathan, 1977 II. Andrefsky, William, 1955 III. Society for American Archaeology.
Annual Meeting (74th : 2009 : Atlanta, Ga.)
cc79.5.s76l 5775 2015
930.1dc23 2014032390
ISBN 978-1-107-02646-9 Hardback
Cambridge University Press has no responsibility for the persistence or accuracy of URLs
for external or third-party Internet Web sites referred to in this publication and does not
guarantee that any content on such Web sites is, or will remain, accurate or appropriate.
Contents
List of Figures and Tables page vii

Contributors xiii
Acknowledgments xvii
Preface xix
Part I Lithic Technological Systems and

Evolutionary Theory
1 Interpreting Lithic Technology under the
Evolutionary Tent 3
William Andrefsky, Jr., and Nathan Goodale
Part IICulture History and Phylogenetic Evolution

2 Graphing Evolutionary Pattern in Stone
Tools to Reveal Evolutionary Process 29
R. LeeLyman
3 Theory in Archaeology: Morphometric
Approaches to the Study of Fluted Points 48
MichaelShott
4 Innovation and Natural Selection in
Paleoindian Projectile Points from the
American Southwest 61
Todd L. VanPool, Michael J. OBrien, and R. LeeLyman
Part IIIApplications of Behavioral Ecology to

Lithic Studies
5 A Case of Extinction in Paleoindian
Archaeology 83
Charlotte Beck and George T.Jones
6 The North China Nanolithic 100
Robert L. Bettinger, Christopher Morgan, and Loukas Barton
v
vi Conten t s
7 When to Retouch, Haft, or Discard?

Modeling Optimal Use/Maintenance
Schedules in Lithic Tool Use 117
Chris Clarkson, Michael Haslam, and Clair Harris
8 Procurement Costs and Tool Performance
Requirements: Determining Constraints on
Lithic Toolstone Selection in Baja California Sur 139
Jennifer M. Ferris
9 A Model of Lithic Raw Material Procurement 156
Raven Garvey
10 Artifacts as Patches: The Marginal Value
Theorem and Stone Tool Life Histories 172
Steven L. Kuhn and D. Shane Miller
11 Signals in Stone: Exploring the Role of
Social Information Exchange, Conspicuous
Consumption, and Costly Signaling Theory
in Lithic Analysis 198
Colin P.Quinn
Part IVCultural Transmission and Morphology

12 An Analysis of Stylistic Variability of
Stemmed Obsidian Tools (Mataa) on Rapa Nui
(Easter Island) 225
Carl P. Lipo, Terry L. Hunt, and Brooke Hundtoft
13 Cultural Transmission and the Production
of Material Goods: Evolutionary Pattern
through Measuring Morphology 239
Nathan Goodale, William Andrefsky, Jr., Curtis Osterhoudt,
Lara Cueni, and IanKuijt
14 What Steward Got Right: Technology, Work
Organization, and Cultural Evolution 253
Nathan E. Stevens
15 Evolution of the Slate Tool Industry at
Bridge River, British Columbia 267
Anna Marie Prentiss, Nathan Goodale, Lucille E. Harris, and
Nicole Crossland
Index 293
Figures and Tables
Figures
2.1. The transformational model and the variational (Darwinian)
model of evolution page 31
2.2. Darwins (1859) model of evolutionary pattern 32
2.3. Fred Plogs (1973) seriogram graph of continuous cultural change 33
2.4. Two illustrations of the relationship between projectile point forms
and the stratigraphy of Mummy Cave 36
2.5. Percentage stratigraphy graph of 27 projectile point types across 9
stratigraphic units at Mummy Cave 37
2.6. Clade-diversity graph for the Mummy Cave projectile points 38
2.7. Measurement values for each of five variables for all individual
points regardless of type per stratigraphic unit at Mummy Cave 39
2.8. Central-tendency graph of the mean for all points regardless of type
per stratigraphic unit at Mummy Cave 41
2.9. Central-tendency graph of the mean (vertical line) and one standard
deviation (box) for all points regardless of type per stratigraphic unit
at Mummy Cave 42
2.10. Coefficient of variation per attribute for all points regardless of type
per stratigraphic unit at Mummy Cave 43
3.1. Regression residual of lnLength upon principal component 1,
plotted against reduction measure lnLT in Folsom replicas 57
4.1. Models of stimulated variation resulting from (a) increased
interaction among members of two or more previously distinct
cultural systems and (b) a rapidly shifting selective environment 63
4.2. The influence of stabilizing selection on variation of a culture trait
over time 64
4.3. The influence of directional selection on variation of a culture trait
over time 64
4.4. The influence of disruptive selection on variation of a culture trait
over time 65
4.5. The influence of a shifting selective environment on variation of a
culture trait within a population 67
4.6. Development of adaptive peaks resulting from selection operating
on increased variation associated with stimulated variation 68
vii
viii Figures a n d Ta bl e s
4.7. Expectations of the model of initial stimulated variation and

subsequent reduction of variation applied to Paleoindian
projectile points 70
4.8. Cumulative corrected coefficients of variation for point length and
maximum width for Blackwater Draw projectile points 72
4.9. Illustration of the dimensions and attributes recorded for the points
in the Eichenberger cast collection 76
4.10. Cumulative corrected coefficients of variation for the eight metric
attributes recorded for Paleoindian points represented in the
Eichenberger cast collection 77
5.1. Model of proposed movements of Western Stemmed (from west to
east) and Clovis (south to west and north) populations 86
5.2. Measurements, attributes, and landmarks of Clovis blades 88
5.3. Distribution of prismatic blades 89
5.4. Distribution of Clovis caches 90
5.5. The relationship between the time spent in the manufacture of a
tool and its utility 93
5.6. Curve-estimate model for finding time thresholds at which an
optimal forager will switch to a different technological alternative 93
5.7. Locations of high-quality toolstone sources on the Great Plains 95
6.1. Relationship between two technologies 103
6.2. Relationship between two mutually viable technologies 104
6.3. Relationship between manufacturing time and return rate 105
6.4. Location of the Dadiwan site in relation to the five early millet
farming complexes of North China 107
6.5. Stratigraphic distribution of major Dadiwan technologies by density
per cubic meter 109
6.6. Flake-and-shatter quartz technology 110
6.7. Microblades 111
6.8. Microblade cores showing all specimens recovered from site 112
6.9. Height (platform to base) of complete cryptocrystalline microblade
cores 113
6.10. Relationships between size and cryptocrystalline fraction of lithic
assemblages 114
7.1. Examples of the experimental tools used in the experiments 120
7.2. Experimental results showing the asymptotic nature of the declining
gain curve over 10,000 strokes for all three experimental tool types 122
7.3. Confidence intervals for gain rate for each tool type over the first
2000 strokes 123
7.4. Relative performance declines for each tool type at 200-stroke
intervals 124
7.5. Model showing the effect of different manufacturing time (T) on
overall gain rate 125
7.6. Model predictions for when to discard each tool type given
different known manufacturing times 126
7.7. The effects of maintenance time as well as manufacturing time
on gain rate and overall efficiency as represented by the slope of the
tangent 127
Figures and Ta bl e s ix
7.8. Mean cumulative weight of wood removed per 1000 strokes

(left y-axis), and mean cumulative weight lost from unretouched
flakes per 1000 strokes (right y-axis) 130
7.9. Mean cumulative rate of increase in step terminated scars for the
3cm used edge (left y-axis) and mean edge rounding rank for the
utilized edge (right y-axis) 130
7.10. Average increases in edge angle (in degrees) over the course of the
experiment for retouched and unretouched edges 131
7.11. Comparison of edge rounding (upper) and stepped scar formation
(lower) on unhafted (broken line) and hafted (solid line)
unretouched scrapers 132
8.1. Map of Baja California peninsula 142
8.2. Map of Espritu Santo Island 143
8.3. Bar chart displaying percentages of flake types for rhyolite
(type 1)and chert/quartzite (type 2) 146
8.4. Line graph of complete flake size grade percentages 147
8.5. Line graph of complete flake reduction trajectory 148
8.6. Line graph displaying proportions of edge damage patterns for
utilized flake tools by material type 150
8.7. Line graph displaying microchip configuration proportions for edge
damage of utilized flake tools by material type 151
9.1. Basic model for establishing the critical use time 163
9.2. The Atuel River drainage, Mendoza Province, Argentina 164
10.1. The marginal value theorem in graphic form 175
10.2. Range of hypothetical artifact utility trajectories 179
10.3. Reformulated MVT 180
10.4. Optimal number of uses after which an artifact should be
abandoned, as a function of maximum potential yield and
artifact cost. (a) Artifact cost = 10. (b) Artifact cost = 25. (c) Artifact cost =
50. Criterion value for abandonment = average potential yield over entire
potential lifetime of artifact (20 uses)-cost 182
10.5. Optimal number of uses after which an artifact should be
abandoned, as a function of maximum potential yield and artifact
cost. (a) Artifact cost = 10. (b) Artifact cost = 25. (c) Artifact cost = 50.
Criterion value for abandonment = average potential yield over first 10 uses
of artifact-cost. 183
10.6. Plots of length versus body width for complete fluted points from
Tennessee 188
11.1. Signaling theory, the fitness continuum, and the relationship
between costly and non-costly signals 206
11.2. A general framework for studying costly signaling behavior with
material culture 208
12.1. The Pacific Islands, showing Rapa Nui on the remote southeastern
edge 226
12.2. Examples of mataa from Rapa Nui assemblages 227
12.3. Location of mataa assemblages on Rapa Nui used in this analysis 229
12.4. Mataa measurements and class divisions 230
12.5. Mataa class dimensions 231
x Figures a n d Ta bl e s
12.6. Seriation solution for mataa classes comprised of stem length/

width ratios and shoulder angle measures 232
12.7. Seriation groups for mataa classes comprised of stem length/width
ratios and shoulder angle measures 233
12.8. Seriation solution for classes of mataa constructed with measures
of stem length and width 233
12.9. Seriation groups for classes of mataa constructed with measures of
stem length and width 234
12.10. Seriation solution for qualitative classes of mataa consisting of stem
shape and shoulder shape dimensions 234
12.11. Seriation groups for qualitative classes of mataa consisting of stem
shape and shoulder shape dimensions 235
12.12. Spatial distributions of the mataa seriation groups on Rapa Nui 236
13.1. Dalton point reduction through use, resharpening, and repair 241
13.2. Map of the southern Levant and early Neolithic sites 242
13.3. An example of an el-Khiam notched point 243
13.4. Direct measurements taken for the NPMI 244
13.5. Image J software plug-ins for NPMI programming 245
13.6. Hierarchical cluster analysis results 246
13.7. Several of the statistically significant clusters 247
13.8. Projectile points made by Ishi 248
14.1. Proposed relationships among behavior, technology, and tradition 257
14.2. Locations of California Central Coast archaeological sites 258
14.3. Proportions of multifunctional tools in California Central Coast
assemblages 258
14.4. Changes in California Central Coast ground stone technology
throughout the Holocene 259
15.1. Major archaeological sites in the Middle Fraser Canyon 272
15.2. Bridge River site with excavation grid superimposed 273
15.3. History of housepit occupations at the Bridge River site 274
15.4. Stratigraphic profile of Area 1 in Housepit 54 (Stratum V = roofs,
III = rim, II = floors) 275
15.5. Housepit 24 stratigraphic profile (V = roof, III = rim, II = floor) 275
15.6. Three (left) and four (right)-sided slate tools from Bridge River 277
15.7. Ratio of total slate tools (TST) to excavated volume (V)
(Table15.2 volume/10,000) 281
15.8. Percentages of sawed and chipped tools from BR 2 and 3 contexts
at Bridge River 281
15.9. Percentages of ground (G) and not ground (NG) tools during BR
2 and 3 occupations at Bridge River 281
15.10. Total sawed and ground slate tools (TSGST) by volume (V)
(Table15.2 volume/10,000) 282
15.11. Number of slate tools (N) per unit of excavated sediment 282
15.12. Percentages of sawed and not sawed tools during BR 2 occupations
at Bridge River 283
15.13. Ratio of total sawed edge (TSE) to total edge (TE) for all slate
tools in BR 2 occupations 283
Figures and Ta bl e s xi
15.14. Number of slate tools (N) per unit of excavated sediment (V)
(Table15.2 volume/10,000) 283
15.15. Percentage of sawed and not sawed tools from BR 3 occupations at
Bridge River 283
15.16. Percentages of ground (G) and not ground (NG) tools from BR 3
occupations 284
15.17. Ratio of sawed and ground slate tools (SGST) to total slate tools
(TST) in BR 3 housepits 284
15.18. Ratio of sawed edge length (TSE) to number of tools (N) with
sawed margins 285
15.19. Ratio of total sawed edge (TSE) to total edge (TE) for all slate
tools in BR 3 housepits 285
15.20. Ratio of total sawed edge (TSE) to total edge of slate tools only
(TEST) 285
15.21. Change in percentages of three- and four-sided tools across BR 2
and 3 occupations 286
15.22. Percentages of three- and four-sided tools in BR 3 occupations at
Bridge River 287
Tables
2.1. Frequencies of projectile points used in analyses and age per stratum
at Mummy Cave 34
4.1. Summary information for point length and maximum width for
Blackwater Draw projectile points 72
4.2. Cultural-historical types and provenience locations for Paleoindian
points in the Eichenberger cast collection 73
4.3. Characters and character states used in the paradigmatic
classification 74
4.4. Summary information for the metric attributes of Paleoindian
points in the Eichenberger cast collection 75
6.1. Dadiwan site components 110
7.1. Details of individual specimens used in the experiment 120
8.1. Proximal flake cortex frequency 145
8.2. Flake type frequency 146
8.3. Tool categories included in the richness index 149
10.1. Results of Pearsons correlations between length and body width
for six Paleoindian point types from Tennessee 187
10.2. Descriptive statistics for basic shape measurements for six
Paleoindian point types from Tennessee 189
11.1. Variables that archaeologists can study within the generalized
framework to identify and explain material culturebased costly
signaling behavior in the past 208
15.1. Slate tool data (counts based on manufacture attributes) 280
15.2. Slate tool data (summed margin length measurements [cm]) and
excavated volume (cubic cm) 280
Contributors
William Andrefsky, Jr.

Dean of the Graduate School
Department of Anthropology
Washington State University
Pullman, WA
Loukas Barton
University of Pittsburgh
Pittsburgh, PA
Charlotte Beck
Anthropology Department
Hamilton College
Clinton, NY
Robert L. Bettinger
University of California, Davis
Davis, CA
Chris Clarkson
School of Social Science
The University of Queensland
Brisbane, Qld
Nicole Crossland
Independent Researcher
Wenatchee, WA
Lara Cueni
Hamilton College
Clinton, NY
Jennifer M. Ferris
Cardno Entrix
Seattle, WA
Raven Garvey
University of Michigan
Ann Arbor, MI
xiii
xiv Contribu t o r s
Nathan Goodale
Hamilton College
Clinton, NY
Clair Harris
School of Social Science
The University of Queensland
Brisbane, Qld
Lucille E. Harris
Applied Archaeological Research, Inc.
Portland, OR
Michael Haslam
Research Laboratory for Archaeology and the History of Art
University of Oxford
Oxford, UK
Brooke Hundtoft
Pima County Community College, East Campus
Department of Humanities, Arts, and Fitness
Tucson, AZ
Terry L. Hunt
Dean of the Robert D. Clark Honors College
University of Oregon
Eugene, OR
George T. Jones
Hamilton College
Clinton, NY
Steven L. Kuhn
School of Anthropology
University of Arizona
Tucson, AZ
Ian Kuijt
University of Notre Dame
Notre Dame, IN
Carl P. Lipo
Department of Anthropology and the Institute for Integrated Research on Materials,
Environments and Society (IIRMES)
California State University, Long Beach
Long Beach, CA
R. Lee Lyman
University of Missouri
Columbia, MO
Contributo r s xv
D. Shane Miller
Department of Anthropology and Middle Eastern Cultures
Mississippi State University
Mississippi State, MS
Christopher Morgan
University of Nevada, Reno
Reno, NV
Michael J. OBrien
Arts and Science Deans Office
Columbia, MO
Curtis Osterhoudt
Independent Researcher
Anchorage, AK
Anna Marie Prentiss
The University of Montana
Missoula, MT
Colin P. Quinn
Museum of Anthropological Archaeology
University of Michigan
Ann Arbor, MI
Michael Shott
Department of Anthropology and Classical Studies
The University of Akron
Akron, OH
Nathan E. Stevens
Far Western Anthropological Research Group, Inc.
Davis, CA
Todd L. VanPool
Columbia, MO
Acknowledgments
This volume has had a long gestation period. We appreciate all the contribu-
tors to this volume for sticking with this effort. We are grateful to the editors,
production staff, and copy editor at Cambridge University Press, as well as
those at their affiliates who guided this project to publication. Thanks go to
three anonymous peer reviewers whose comments greatly improved drafts of
the chapters included in this volume.
The editors would like to acknowledge and thank the late George H. Odell,
an old friend and inspiration to researchers studying lithic technological sys-
tems around the globe.
xvii
Preface
This volume is an outgrowth of a symposium organized for the 74th Annual

Society for American Archaeology meeting in Atlanta, Georgia, titled
Evolutionary Approaches to Understanding Stone Technologies as a Bybroduct
of Human Behavior. The purpose of the symposium and this volume is to
demonstrate the connection between lithic analysis and a body of theory to
guide interpretations of past human behavior in studies of lithic technological
systems. The hope we had for this volume stemmed from the original sympo-
sium and to capture the state of the field of lithic technological organization
incorporating a body of theory for guiding interpretation.We view evolution-
ary theory very broadly and understand that others may have a much narrower
view. With this in mind we invited scholars with diverse perspectives on evo-
lutionary thought who also used lithic technological systems as a medium of
analysis. Our vision was to begin a conversation about interpreting past human
behavior derived from lithic artifacts interpreted through a very wide variety
of evolutionary approaches. In doing so we hope that the diverse perspectives
on evolutionary thought might be viewed as compatible or complementary
rather than exclusionary.
The authors of the various chapters in this volume represent some of the
most respected scholars as well as many young contributors to the field of
lithic analysis and evolutionary archaeology. We selected this field of scholars
in hopes of bringing different perspectives from existing researchers together
under one cover and simultaneously adding new opinions on lithics and evo-
lution from an up-and-coming generation of archaeologists.
This book contains many of the same papers that were presented in the
original symposium. Although we lost a few authors along the way, we also
gained new participants during the journey toward publication.We would like
to thank all of the participants in that session and especially those who con-
tributed their ideas, methodologies, and interpretations to be included in this
volume.
xix
PartI
Lithic Technological Systems and

Evolutionary Theory
One
Interpreting Lithic Technology

under the EvolutionaryTent
William Andrefsky, Jr., and Nathan Goodale
An underlying theme of this volume is that lithic technological analysis is not

well integrated with a theoretical approach, and that evolutionary theory has
great potential to fill this void. This is not to say that evolutionary concepts
and models have not been used by archaeologists who have been working
with lithic technological data. In fact a number of recent volumes have been
published recently that are dedicated to the application of specific evolution-
ary concepts to lithic data. Surovells (2009) book, Toward a Behavioral Ecology
of Lithic Technology, is a good example. The edited volume by Michael OBrien
(2008), Cultural Transmission and Archaeology, draws on many lithic data case
studies. A number of highly regarded and well cited journal articles that have
applied specific evolutionary approaches to lithic technology (Beck etal. 2002;
Bettinger and Eerkens 1999; Brantingham 2003; Mesoudi and OBrien 2008;
OBrien etal. 2001; and others).This volume attempts to bring together several
different evolutionary perspectives and lithic technology. We invited research
contributions from a number of scholars who have been standing on differ-
ent sides of a theoretical fence at one time or another, but have all embraced
Darwinian evolutionary approaches and in this case use lithic technology in
that effort.
The chapters included in this collection use lithic artifacts or artifact char-
acteristics as an empirical proxy for past human land-use strategies and/or past
human behaviors that apply an evolutionary theoretical foundation to help
interpret those data. Even though all of the chapters in this volume emphasize
3
4 Wi ll i am A n d r e fs ky, Jr . , a n d N at h an G o o d al e
evolutionary approaches and lithic technological systems, the amount of

theoretical diversity within the volume is quite striking. The chapters cover a
range of topics beneath a broad evolutionary umbrella including but not lim-
ited to human behavioral ecology (HBE), cultural transmission, phylogenetic
analysis, risk management, macroevolution, dual inheritance theory, cladistics,
central place foraging, costly signaling, selection, drift and various applica-
tions of evolutionary ecology. Some of these evolutionary approaches have not
always completely agreed with one another. However, we believe that within
this group of studies there is a strong common ground for multiple approaches
to Darwinian thinking. In some chapters we see an intentional blending of
multiple evolutionary approaches towards the study of lithic technology. In
other chapters we intentionally point out areas that we believe represent like-
minded thinking from different evolutionary models, even if authors may not
have intentionally made such linkages.
This assemblage of chapters is structured in a way that segregates the vol-
ume contributions into three very broad thematic topics: phylogenetic evo-
lution, HBE, and cultural transmission. However, many of the chapters in this
volume could have been placed into more than one of these themes and we
hope authors and readers are comfortable with our distribution. The fact that
so many of the chapters could be included in multiple sections again points
to our underlying impression that there is increasingly more common ground
rising under the evolutionary umbrella in archaeology. It became evident to
us that a number of crosscutting issues and data sets joined chapters from dif-
ferent themes. Four of the chapters (Beck and Jones; Kuhn and Miller; Shott;
and VanPool etal.) explored evolutionary applications with North American
Paleoindian projectile technology. Four chapters (Bettinger etal.; Kuhn and
Miller; Stevens; Goodale etal.) examined retouch intensity in some form or
another.Two of the chapters used experimental replication of artifacts to assess
evolutionary models (Clarkson etal.; Goodale etal.). Four chapters focused on
lithic raw material provenance in some form (Beck and Jones; Bettinger etal.;
Ferris; Garvey). Of course, all chapters use evolutionary approaches along with
some aspect of lithic technology.
We also hope this volume will inspire lithic researchers to apply their data,
whether generated experimentally, collected from region surveys, or excavated
from detailed stratigraphy, to more problem oriented approaches to analysis
and interpretation.
We feel that the context of an archaeological study (particularly lithic study)
is extremely important for understanding the kinds of activities that have
occurred at a particular location or within a particular region. However, the
value of that specific context can often be measured only by the extent to
which it is abstracted to more generalized interpretations. In some lithic stud-
ies, strict emphasis on context provides little more than a detailed description
Interpreti n g L i t hi c T e chn o l o g y u nd e r th e E v o l u ti o n aryTe n t 5
of artifacts and their associations with one another and their environment.
In other studies the lack of context and emphasis on abstract associations of
data result in little more than untested hypotheses and speculations about
what could or might have happened in past times on sites and within regions.
We believe this volume emphasizes both ends of this spectrum and hope our
examples show how lithic technological data can be tied to evolutionary the-
ory to build stronger interpretations of past human activities.
Culture History, Lithic Data, and Phylogenetic Evolution

If we acknowledge that evolution is defined simply as descent with modifica-
tion (Lyman and OBrien 1998), and that evolutionary approaches deal with
historical phenomenon (Boyd and Richarson 1992; Jones etal. 1995; Lipo etal.
2006), then cultural-historical studies associated with lithic assemblages pro-
vide a common heritage for the various components of evolutionary thought
in archaeology and lithic studies. Archaeologists have been arranging artifact
types and assemblages into chronologies since before the use of radiocarbon
dating (Krieger 1944; McKern 1939; Ritchie 1944; Witthoft 1949) and the
practice continues today (Beck and Jones 2010; Ramenofsky 2009; Sellet etal.
2009).The structuring of lithic types and assemblages into historical sequences
based on similarities of form and compositions, respectively, is a form of phylo-
genetic analysis not substantially different from what takes place in paleoecol-
ogy. Early chronological studies of stone tool assemblages were explicit about
the relationships between different types over time. There was an attempt to
show that similarity of form represented lineal descent with modification.This
is evident in Jesse Jennings discussion of the Plano big game hunting tra-
dition. He notes (1968:123), If typological evidence is to be accepted, one
can see a continent-wide dispersal of Big Game Hunters by, or earlier than
10,000 B.P....In all areas, however, the tradition of the lanceolate blade or
point, fluted or unfluted, first coexists with, and finally becomes part of, the
next widespread and long-lived stage called the Archaic. That similarity of
artifact formover time and space represents common ancestry is an evolution-
ary notion. As noted by Neiman (1995:31), Culture history was grounded in
the interpretation of the record in terms of homologous similarity.
Cultural chronologies of this kind were swept into the evolutionary litera-
ture in archaeology under the wing of the selectionist movement (also identi-
fied as evolutionary archaeology) that can be equated roughly with the work
of Dunnell (1978, 1980, 1982) and his followers (Jones etal. 1995; Leonard and
Jones 1987; Lyman and OBrien 1998; OBrien and Holland 1990, 1992; OBrien
and Lyman 2000; OBrien etal. 1998).They define evolutionary archaeology as
change in the composition of a population over time. In evolutionary archae-
ology, the population is artifacts, which are viewed as phenotypic features, and
it is the differential representation of variation at all scales among artifacts for

which it seeks explanations (Lyman and OBrien 1998:616). Evolutionary
archaeology involves (1) measuring variation that is, dividing it into dis-
crete sets of empirical units...; (2) tracking variation through time and across
space to produce a historical narrative about lineages or particular variants;
and (3) explaining the differential persistence of individual variants compris-
ing lineages in particular time-space contexts (OBrien etal. 1998:487). The
selectionist paradigm takes the work of culture historical archaeologists and
applies heritable continuity to the temporal sequence of artifacts. They, like
paleobiologists they emulate, attempt to distinguish between analogous and
homologous characteristics to assess degree of relatedness.
Some of the early research in this area dealing with lithic technology can be
seen in the scraper study by Meltzer (1981). He attempted to separate aspects
of stylistic variability from functional variability with the underlying notion
that stylistic variability is viewed as nonselective or homologous (see Dunnell
1978:199). His study recognized scraper characteristics on stone tools for times
and places around the world that had little possibility of heritable linkages. In
doing so, he was able to establish those characteristics as functional attributes of
the tools. So far as I can tell, given the variables I selected, the sample size, and
the particular time/space coordinates of data, there is no stylistic component
in the morphology of the tools examined (Meltzer 1981:326). The separation
of style and function in materials is a fundamental distinction for the selection-
ist approach in archaeology. Those units that are functional will be sorted by
natural selection; those that are stylistic will be sorted by the vagaries of trans-
mission (OBrien etal. 2003:576).
The integration of stone tool analysis within the evolutionary framework of
selectionism increased in frequency with the adoption of systematic measures
of phylogenetic analysis known as cladistics. Put rather simply, cladistics is a
form of phylogenetic mapping that uses derived characteristics to construct
phylogenies (Mayr 1982). Such analysis is often displayed in the form of a
branching tree or cladogram. In a cladogram taxa are organized into groups or
clusters based on shared derived characters. Any taxon in the population that
does not share a derived character is graphed alone as an out group. In this way
the cladogram shows the historical relationship of taxa and identifies the attri-
butes or characters that link the various taxa (Buchannan and Collard 2008).
Foley used cladistics on stone tool assemblages to establish relatedness among
early hominids (Foley 1987; Foley and Lahr 2003). Lyman and OBrien (2000)
applied clade-diversity approaches to understanding projectile point variation
from Gatecliff Shelter in Nevada. Their analysis showed that projectile point
diversity at the site may have resulted from an increase in the number of
weapon delivery systems. Others using different kinds of lithic analysis suggest
the same results (Beck 1995; Hughes 1998). This type of analysis was applied
to Paleoindian projectile technology from the southeastern United States to

establish relationships among Paleoindian technologies and later Archaic tech-
nologies (Darwent and OBrien 2006; OBrien etal. 2001). The Paleoindian
example was expanded to explore human peopling of North America using
cladistics (Buchanan and Collard 2007, 2008; Buchanan and Hamilton 2009).
Others use cladistic approaches to assess phylogenetic relationships between
bow and arrow technology and dart technology (Lyman etal. 2008, 2009).
There have been many critics of the selectionist position with regard to
using artifacts as phylogenetic markers in the same way that paleontologists
use fossil bones to reconstruct phylogenetic trees of ancient members of the
animal kingdom (e.g. Bamforth 2002; Boone and Smith 1998; Fitzhugh 2001;
Gabora 2006; Shennan 2002), and there has been ample reply to such criticism
(OBrien and Lyman 2002; OBrien et al. 2003). Though exploring differ-
ences and similarities between various ideological camps under the evolution-
ary umbrella is outside the scope of this book, we do think there has been an
increasing amount of common ground between camps. For instance, Bamforth
(2002) argued that variation in material culture (artifacts) may be conditioned
by a number of different agencies, such as culture and human behavior. He
suggested that not all variation in human artifacts over time may be represent-
ing evolutionary trends in the same way that paleontologists see evolutionary
trends in ancient fossils.We feel that some archaeologists who use phylogenetic
analysis of artifacts also embrace this position or have come to embraceit.
Chapter2 by Lyman explores graphic representation of artifact variation over
time to help illuminate evolutionary processes. He demonstrates several impor-
tant characteristics of graph styles. For instance, he graphs projectile point data
to show relative abundance of types (richness) over time (displayed by strata) is
a good reflection of the Darwinian variational model of evolution. That model
shows changes in frequencies of types over time and not changes in types.When
variation in attributes of point types is displayed over time we can see how for-
mal variation of the population is being altered or incorporated into the types.
Indeed, graphic styles show important and distinct aspects of artifact variation.
However, our take away point here is Lymans recognition of different pro-
cesses associated with different aspects of lithic artifacts. He emphasizes that
graphed patterns and their inferred processes depend on the classificatory units
used in the analysis. He notes, ...those units of measurement, that are graphed,
whether types of points, length of points measured in centimeters or millime-
ters, or neck width measured in millimeters or tenths of millimeters. Not just
knowing the identity of the graphed units, but understanding what those units
actually are, would seem to be a critical step in the production of graphs that
are correctly perceived and subject to a minimum of misinterpretation (or mis-
perception). In our opinion, this is what Bamforth (2002:448) was advocating
for with regard to variation in artifact form in stating,...I have argued here that
archaeologys essentially universal reliance on aggregate data sets that represent

the activities of human groups whose familial and reproductive relations are
unknown currently precludes us from making such a contribution. It may be
possible to develop modes of analysis that allow us to surmount this problem,
but we have certainly not yet accomplished this. We think Lymans study goes
a long way towards understanding and developing such modes of analysis. As a
result we see some common groundhere.
Another aspect of the Lyman chapter we think is critical here especially
with regard to lithic studies and phylogenetic analysis is the recognition of
what we call context of variation. Lyman correctly notes, A graphed tem-
poral sequence of archaeological data does not necessarily imply evolution,
regardless of pattern or process. This is echoed in Chapter3 by Shott, which
recognizes that projectile points change as a result of multiple processes (use,
functional requirements, human situational needs). These sources of morpho-
logical variation need to be understood before practitioners of phylogenetic
lithic analysis graph or even select artifact attributes for phylogenetic study.
Cladistic analysis may plot the sequence of change, but only detailed con-
textual study can explain it (Shott 2008:150). We could not agree more with
Shott (and by extension Lyman) on this issue. If archaeologists are interested
in characterizing evolutionary trends such as descent with modification in
artifact forms it is critical that we select the appropriate attributes to show
phylogenetic relationships. It may not be appropriate simply to use whatever
attributes are available.
Not all attributes or types produced from attributes represent lineal decent.
It is important to understand some of the production, use, maintenance and
reuse processes that influence the morphological variability found in stone
tools before plugging tool attribute variability into clustering algorithms. For
instance, phylogenetic projectile point typologies are meant to show char-
acter states that are the result of shared ancestry derived from the ancestral
state for the type. This is why we can effectively use projectile point typolo-
gies to describe cultural-historical sequences. However, if the projectile point
typologies are built or assessed by morphological characteristics that do not
vary by descent and are not derived from an ancestral state, there is a good
chance we will be barking up the wrong phylogenetic tree. This is relatively
easy to visualize with morphological characters associated with phenomena
we understand well. If we were interested in describing the phylogenetic
history of Alaskan Dall sheep (Ovis dalli dalli) based on skeletal remains we
probably would not measure horn curl length, knowing that (in male sheep)
it correlates positively with the age of the individual animal and is directly
related to the life history of the individual organism. We know this through
observations of contemporary Dall sheep and through studies charting the
growth of horn curl and age at time of death. Foot structure and overall
body size have more to do with historical lineages of the species than horn
curl length. In the same way, we know that some types of projectile points
have blades that are altered and changed throughout the period of time they
are used by ancient humans. Figure 13.1 in Chapter 13 by Goodale et al.
shows variation in blade shape reflected in stages of projectile point produc-
tion and use, taken from Al Goodyears (1974) study of Dalton Points from
the Brand Site. This was among the first studies to demonstrate how blade
shape and size on projectile points were reduced from use and resharpen-
ing. Others have more recently demonstrated such morphological changes
on a variety of projectile point styles using both experimental resharpening
studies and analysis of allometric characteristics from excavated collections
(Ahler and Geib 2000; Andrefsky 2006; Bement 2002; Kuhn and Miller this
volume; Shott and Ballenger 2007; Truncher 1990). If projectile point blade
elements change size and shape during their use-life it is not reasonable to
use this characteristic of projectile points to chart decent. Such measurements
are akin to charting Dall sheep lineages based on horn curl length without
knowing that horn curl length changes during the lives of individual sheep.
Projectile points are not the only stone tools that undergo changes during
their use lives. Stone scrapers, knives, and blades have been shown to change
morphology as a result of use and resharpening (Goodale etal. 2010; Hiscock
and Attenbrow 2003; Hiscock and Clarkson 2007; Clarkson 2002). As Lyman
(Chapter2 and preceding text) notes, it is important to understand the units
we are measuring. It is little wonder that Shott (Chapter3) when referring to
projectile point characteristics used in phylogenetic analysis says, The phy-
logenetic method used, common in cladistic studies, produced parsimonious
cladograms that matched none of the outcomes predicted by any hypothesis,
even the one favored.
Chapter3 by Shott has been mentioned several times in this section. His
contribution emphasizes details that are worth considering in phylogentic
analysis of lithic artifacts. However, he does more than identify problem areas.
He suggests that archaeology needs to embrace a new theoretical perspective
and suggests another evolutionary approach used in the biological sciences,
morphometrics. Shott describes how morphometrics can overcome many of
the analytical problems associated with other phylogentic strategies when deal-
ing with lithic technology. He also eloquently advocates for an archaeological
theory that focuses on form and pattern of material culture: one that explains
variance and change, and allows for an explanation of mode, rate and causes
of change in our materials. We feel Shotts ideas are perfectly aligned with the
challenges of lithic technology and fit well under the umbrella of evolutionary
thought.
Human Behavioral Ecology, Tool Use-Life,

and Raw Material Provenance
Roughly simultaneous with the selectionist genre of evolutionary approaches
in lithic technological studies was the adoption of evolutionary ecology or
behavioral ecology. Evolutionary ecology attempts to explain cultural and
behavioral change as forms of phenotypic adaptation to varying social and
ecological conditions (Boone and Smith 1998:141). Evolutionary ecologists
assume that natural selection has designed organisms to respond to local con-
ditions in ways that increase their fitness (Winterhalder and Smith 1992). Some
archaeologists separate evolutionary ecology and behavioral ecology, where
Behavioral ecology is that subset of evolutionary ecology concerned with
accounting for the evolution and adaptive character of behavior (Fitzhugh
2001:129). In either case, phenotypic variability (including behavior) is con-
strained by natural selection to seek fitness propagating solutions. Models of
behavior (fitness maximizing behavior) are then developed in local ecolog-
ical contexts and are tested against the archaeological record (Boone 1992;
OConnell 1995).
The lithic technological literature is full of such evolutionary ecologi-
cal approaches dealing with risk (Bousman 2005; Clarkson 2008; Fitzhugh
2001; Shott 1996; Torrence 1983), production strategies (Andrefsky 1994;
Brantingham etal. 2000; Jeske 1989; Clarkson 2008), optimization (Bamforth
1986; Bleed 1986; Goodale etal. 2008; Kelly 1988;Tomka 2001), and residential
mobility (Brantingham 2006; Lurie 1989; Parry and Kelly 1987; Shott 1986).
Much of the early and contemporary evolutionary ecology research dealing
with lithic technology used fairly informal modeling that stresses the associa-
tion of two or more variables. For instance, many studies emphasize lithic raw
material transport costs as an independent parameter for or against a depen-
dent variable such as stone tool technology (Bamforth and Becker 2000; Kuhn
2004). Other studies emphasize the relationship between technology and rela-
tive residential sedentism (Kelly and Todd 1988; Wallace and Shea 2006). Such
simplistic modeling has been criticized as nonevolutionary on the grounds
that it does not reference evolutionary forces to explain change (Abbott etal.
1996). However simplistic the modeling, such studies attempt to show causal
relationships between two or more factors and they tend to place their studies
within a historical context to explain change or stasis over time. Explanations
of phenomena do not need to be posed in evolutionary contexts to be related
to the processes of evolution. Bettinger and Richarson provide a good exam-
ple of just such a case (1996:224):
Thus the question posed to a physiologist, Why is this dog panting? Is
more appropriately and directly answered by saying To regulate its body
temperature, than by a protracted explanation involving the evolution

and natural history of dogs and warm-bloodedness. In responding with-
out direct reference to evolutionary processes, the physiologist does not
question that this panting is the result of a long evolutionary history.
The point here is that explanations may be only functional do not mean
they are not useful in an evolutionary context or understanding an evolution-
ary process. The evolutionary biologist, Ernst Mayr (1982:8990) was clear
about this when he applied Allens rule to explain size difference in ravens from
the Arctic and equatorial zones. Body size is larger and extremity size is smaller
in colder than in warmer climates. This is a functional explanation associating
climate with body size characteristics. Mayr does not explain the process of
natural selection within each environment as it relates to the ravens circula-
tory system.
Several contributions to this volume use formal and less formal models of
evolutionary ecology to address lithic artifact data. One of the most ingenious
applications of evolutionary ecology to lithic data is the Kuhn and Miller
(Chapter 10) study. They actually attempt to unite the field of lithic tech-
nological organization and evolutionary ecology. Kuhn and Miller use the
patch choice model developed by Charnov (1976) and apply the Marginal
Value Theorem (MVT) to stone tool data in an effort to determine when
stone tools should be discarded or abandoned (see also Surovell 2009). In this
study they apply the MVT to projectile point life histories. They essentially
conceptualize lithic artifacts as patches of utility. The amount of utility con-
tained in each artifact is limited and utility for many artifacts should decline
over time as the artifacts are increasingly used and wornout.
The model is applied to a set of Paleoindian projectile points from Tennessee.
Utility is measured simply by the amount of correlation between projectile
point type lengths and widths based on the assumption that newly manufac-
tured points begin their use-lives with fairly standardized shapes. As blades and
tips are resharpened or refreshed after use and damage, the types should show
less correlation between the two variables. Results of their study show that
discard patterns of Paleoindian points in Tennessee changed from the earlier to
the later times. The MVT model suggests that projectile points were discarded
later in their use-lives because of increased cost of replacement or because of
a decline in average return from use of points. Both explanations conform to
the model expectations. The second possibility is unexpected given traditional
interpretations of lithic technological organization and suggests that formal
modeling of stone tools may be a productive direction for lithic analysis to
help explain patterning in the record.
Chapter7 by Clarkson etal. also adopts optimality modeling using the MVT.
Again in this case, lithic tools are used as patches of utility. And again, utility is
contrasted with the amount of retouch. However, in the Clarkson etal. study,
utility is empirically calculated on stone tool scrapers by the amount of wood
removed by each scraper in a series of experiments.Three kinds of wood scrap-
ing tools are assessed for utility: (1) unhafted unretouched flake tools, (2) hafted
unretouched flake tools, and (3) unhafted retouched flake tools. Surprisingly,
unretouched and unhafted flake tools are found to be the most efficient scrap-
ing tools when compared to the other two forms.The authors state,The main
conclusion this study reached is that prehistoric tool users should in many cases
only have retouched their woodworking toolkits when replacement material
was scarce and/or unpredictable or when manufacturing costs were high (e.g.
hafting). Interestingly, use of this model in association with experimental data
directly measuring tool efficiency has confirmed some of the less formalized
evolutionary models often associated with lithic technological studies such
as the Parry and Kelly (1987) model of expedient and formalized tools. Both
Clarkson etal. (Chapter7) and Kuhn and Miller (Chapter10) have extended
implications of less formalized technological modeling by using more formal-
ized models associated with evolutionary ecology.
Chapter6 by Bettinger etal. takes optimality modeling a step farther and
models technological investment as a relationship between tool manufacturing
time plus resource procurement time against resource procurement rate (based
on previous work of Bettinger etal. 2006 and Ugan etal. 2003). Simply put, the
model predicts that when resources are abundant, time spent in procurement
will be low, and the less costly technology is superior. This conforms to what
Clarkson etal. (Chapter7) found with regard to unretouched versus retouched
or hafted tools. The less costly technology was more effective and should have
been selected, given all else was equal. However, Bettinger etal. also predict
that when resources are scarce, time spent in procurement will be high, and
the more costly technology will be superior. In other words, groups under the
most resource stress will display the most costly refinements of the most costly
technology. Bettinger et al. apply their model predictions to the stone tool
industry of millet farming aboriginal peoples of a remote section of North
China. Among other things, they conclude that millet farming was introduced
to this marginal environment by a migrant population into the region and that
the costly microlithic or nanolithic technology was extremely costly. Again,
formalized models associated with optimal foraging are used to help explain
not only aspects of technological differences, but also how behavioral, eco-
nomic, or subsistence variability may be related to the technological shifts.
One area of lithic technology and evolutionary ecology that has received
considerable attention lately involves lithic raw material selection, use, and
discard. This is partially due to the fact that lithic raw material source loca-
tional studies can provide some reliable measure of the circulation range of
tool makers and users. Beck etal. (2002) adopt an optimal foraging model to
Interpreti n g L i t hi c T e chn o lo gy u nd e r th e E v ol u ti o n ary T e n t 13
assess transport and quarry behaviors in the Great Basin. Similar to many of
the models noted in the preceding text, they assess efficiency but this time
they predict when tools should be made at raw material source locations ver-
sus transporting raw material to the residential location for later production
into tools. Figure5. 3 in their study illustrates the point.The x-axis to the right
represents sequential stages of biface manufacture. The x-axis to the left shows
travel distance from residence to quarry. A line tangent to the curve predicts
the cost-effective travel distance for production of a biface to a particular stage.
Ultimately their study shows that distance to lithic raw material sources from
residential locations has a significant impact on the extent to which tools are
shaped at the source areas.
Other studies have used lithic source locational data to assess travel routes
and forager ranges (Daniel 2001; Feblot-Augustins 2009; Jones et al. 2003).
Still other studies show that lithic source distances play a role in the extent to
which stone tools are reduced, modified, and recycled (Andrefsky 2008; Dibble
1991; Hiscock 2009;Terry etal. 2009). Some archaeologists have explored lithic
technological characteristics with models that hold raw material availability
and location neutral in an attempt to understand behavioral factors that may
influence stone tool production and consumption (Brantingham 2003; Feblot-
Augustins 1997). Holding raw material procurement neutral, Brantingham
(2006) was able to model forager mobility patterns using a random walk model
to separate out information about organizational parameters such as risk sen-
sitivity, timeenergy optimization, and levels of planning. Many of the expec-
tations derived from less formal models of forager mobility were consistent
with Brantinghams random walk models. In particular, the Levy mobility
model suggests that greater mean and maximum stone transport distances may
indeed reflect increases in planning depth, greater optimization of mobility,
and greater risk sensitivity (Brantingham 2006:449).
Beck and Jones (Chapter5) extend their optimal foraging model to help
explain the spread of Paleoindian lithic technology on the North American
continent. They suggest that locational factors of high chipping quality lithic
raw materials were important for the spread of Clovis blade technology, and
to some extent this is evident from tool caches. Their model also suggests
that Clovis technology probably originated in the southeastern United States
and spread to the north, west, and east from this origin. Of course, this model
requires further testing (see Beck and Jones 2010) but it does contradict the
assumptions of VanPool
etal. (Chapter4), which adopts the north to south
and east migration of Clovis technology through the ice-free corridor of west-
central Canada. Beck and Jones further suggest that the Pacific Northwest was
originally colonized by aboriginal populations using stemmed points and that
Clovis technology came into the interior Pacific Northwest after stemmed
point technology was already in place.
Garvey (Chapter 9) also adopts optimality modeling for explanations of

lithic raw material procurement. Garvey also acknowledges the less formalized
land-use models often associated with lithic technological studies as evolution-
ary but notes that lithic technological analysis has not been easily translated
into fitness measures and she feels it has great promise in pursuing such a path.
Garveys model predicts lithic raw material procurement decisions based on
the assumption that lithic raw materials are ranked according to their quality
and that high-quality materials improve return rates for stone tool activities.
A version of the Bettinger etal. (2006) technological intensification model is
adopted (see also Bettinger etal., Chapter6), with slightly modified param-
eters. Garveys model requires procurement and manufacturing costs, measures
of raw material quality, rates of return from tools of a given material type, and
tool use time. This model is applied to sparse archaeological data from the
Middle Holocene of Mendoza, Argentina and has generated a number of test-
able hypotheses about human land-use practices.
A slightly different approach to optimality models and lithic raw material
procurement is presented in Chapter 8 by Ferris. Here lithic raw material
proximity is inferred to explain tool production behavior from lithic debitage
assemblages. Essentially, lithic raw material provenance is definitively unknown
but optimality models indicate that proximity (Beck and Jones, Chapter 5;
Beck et al. 2002) and quality (Brantingham et al. 2000; Garvey, Chapter 9)
should guide stone tool production decisions. However, Ferris shows that other
factors are at play in her study area. Specifically, it is shown that activity type
or artifact function may be linked to differences in lithic raw material type
preferences. This is similar to results obtained by Braun etal. (2009) that show
lithic raw material quality may be defined more broadly than simply chip-
ping quality and may extend to other characteristics of the raw material such
as durability for performance of certain tasks and edge sharpness for specific
functions. This suggests that models need to be crafted with these contexts in
mind. Raw material quality may be gauged by homogeneity of structure and
brittleness in some situations but other situations may link raw material quality
with durability or shape.
Lithic raw material provenance has great potential to generate extremely
reliable information about aboriginal land-use practices and/or aboriginal
exchange networks because locationally diagnostic sources of stone can be
mapped against tool use and depositional locations. Unfortunately, most tool-
stone found on archaeological sites worldwide is composed of cryptocrys-
talline silicates such as chert and flint. Unlike obsidian and other fast cooling
igneous rocks (Eerkens et al. 2007; Shackley 2005), chert has been difficult
to assess for provenance macroscopically or geochemically. In their study of
Scandinavian chert Hogberg and Olausson (2007) attempted to character-
ize cherts by macroscopic characteristics such as color, structure, translucency,
and cortex condition. They also attempted various geochemical techniques to

establish chert provenance. Unfortunately, they found as much within-source
heterogeneity as between-source homogeneity with regard to chert character-
istics. They discovered that Scandinavian chert has the same kinds of problems
with diagnostic provenance as most other cherts in all parts of the world it
is impossible to determine small-scale locational differences. Unfortunately,
archaeologists conducting technological studies and those applying evolution-
ary models to chert locations have generally ignored this situation and have
assumed provenance of cherts by some unknown or unexplained reasons. We
feel this can create significant interpretive problems and we believe that there
are new techniques and data showing that not all cherts are immune to diagnos-
tic provenance studies. Some progress is being made in the area of authigenic
biogenic mineral formation analysis in cherts that is promising for provenance
(Foradas 2003; Hughes etal. 2010). There have also been some luminescence
analysis of cherts, particularly fluorescence emission analysis, that is effective
for chert provenance (Akridge and Benoit 2001; Lyons etal. 2003). There is
also evidence that not all chert and flint were formed under deep-sea subma-
rine contexts and that some cherts may have formed in sedimentary contexts
associated with fissure eruptions of lava or volcanic venting, creating diagnos-
tic trace elements for very restricted ranges of chert outcrops (Andrefsky etal.
2010; Orr etal. 1999). We feel it is time for lithic researchers to embrace the
lithic raw material provenance challenge in both informal models and formal
models of evolutionary ecology.
Another characteristic of evolutionary ecology and particularly optimal for-
aging models is an underlying assumption that optimal food gathering strat-
egies or foraging efficiency or production strategies is a proxy for fitness. In
other words, the most optimal production or subsistence strategies correlate
with the most fit individuals. Stated another way, optimal foraging theory,
...presumably implies that the variables analyzed in place of fitness for
example, foraging efficiency and caloric intake vary predictably with fitness
and might even imply that this can be, or has been documented empirically
(Bamforth 2002:439). Unfortunately, this has not been demonstrated. We have
no evidence that actually demonstrates efficient food collection and consump-
tion strategies with greater fitness. Similarly, there is substantial evidence that
problems associated with adaptive strategies typically have many local optima
(Bettinger etal. 1996:149). As such, if there is more than one local optimum,
populations may reach and maintain those different optimal solutions depend-
ing on where their starting points are with regard to a particular problem.This
suggests that optimal solutions are multiple and depend on the context of the
situation. Looking at this from a more contemporary example may reveal how
complicated this situation can be. When duck hunting, it may be most effec-
tive to shoot tungsten-loaded shot, because it has a higher density than lead
or steel, and as such, it carries farther down range and impacts targets with
more energy. However, it may not be the most optimum shot because it is not
easy to find in stores, it is more expensive than alternates, and it is too dense
to use in many older make shotguns. Loads made with steel shot might be
more optimal because they are more readily available, less expensive, and easily
shared among hunters in the field even if they are not the most effective for
bringing home ducks. It is important to remember that there may be multiple
local optima when modeling complicated data sets such as stone tools. Again,
context of study can play an important role in understanding the relevance of
model parameters.
Chapter11 by Quinn could be discussed in several sections of this overview
focused on costly signaling theory and its role in lithic technological systems.
His chapter takes on issues related to both cultural transmission and mod-
els of optimality. Quinns contribution highlights the need for methodologi-
cally sound models to incorporate new theoretical toolkits to interpret lithic
technological systems. We chose to end the HBE section with this chapter
because we feel that costly signaling theory rightly belongs within the larger
theory of HBE. Costly signaling approaches are common in anthropological
studies dealing with subsistence data such as meat procurement and sharing
(Bliege Bird and Smith 2005; Hawks and Bliege Bird 2002). Fewer studies have
focused on archaeological studies and fewer yet have focused on lithic tech-
nology and costly signaling theory (McGuire and Hildebrandt 2005; Mithen
2008). Quinns chapter adds to that small but growing assemblage of archaeo-
logical studies dealing with costly signaling theory.
Lithic Technology, Neutral Variables, and Cultural

Transmission
When Dunnell (1978) theoretically separated style from function in archae-
ological materials and Meltzer (1981) applied it to a class of stone tools they
equated these traits as selectively neutral and selected upon, respectively.
Meltzers study concluded that scrapers had functional characteristics and were
selected upon. It was Neimans studies of architecture and ceramics (1990, 1995)
that operationalized stylistic variation in an evolutionary context. Through a
series of mathematical models he showed that stylistic variation (exterior lip
decoration on Woodland cooking pots) was selectively neutral and that vari-
ation within this class of decoration was introduced as a result of drift and/or
cultural transmission in the Boyd and Richarson (1985) and Cavalli-Sforza and
Feldman (1981) evolutionary genre. He emphasized that the Markovian struc-
ture of drift makes it likely that isolated groups will tend to diverge from one
another when considering stylistic traits and that under these circumstances
only some form of cultural transmission among groups shall increase similarity
of stylistic characters (Neiman 1995:31). His work with cultural transmission

and ceramics was corroborated by others investigating aspects of neutral evo-
lution and drift in ceramics (Lipo etal. 1997; Shennan and Wilkinson 2001).
It was not long before cultural transmission studies were directly applied to
lithic technological assemblages (Bettinger and Eerkens 1997, 1999). Aspects of
cultural transmission theory such as the origins of material variation and influ-
ences of copying error were explored with stone tool examples (Buchanan
and Hamilton 2009; Eerkens 2000; Eerkens and Lipo 2005, 2007; Hamilton
and Buchanan 2009).
In our opinion one of the classic studies of cultural transmission using stone
tool technology was the investigation of the spread of the bow and arrow
in the Great Basin (Bettinger and Eerkens 1997, 1999; Eerkens etal. 2006). Here
they explore possible explanations as to why some Elko points from Nevada
some Rosegate points from California are misclassified. After controlling for
age, a series of metric attributes from the two point types from each of the two
regions were investigated. Their analysis shows that misclassified Elko points
from Nevada may be attributed to multifunctional properties of darts. These
are believed to have been used as projectiles and as cutting tools that required
resharpening. However, misclassification of Rosegate points from California
(based on base width measurements) cannot be attributed to resharpening
and instead was the result of differences in cultural transmission and regional
adoption of the bow and arrow. They demonstrate that adoption of the bow
and arrow in central Nevada was probably a result of indirect bias transmission
where point makers acquired multiple aspects of this technology as a complete
package. The adoption of the bow and arrow in eastern California, however,
was probably a result of guided variation where there was a great deal of
experimentation resulting in more variation in Rosegate point characteristics
(Bettinger and Eerkens 1999:236237).
Two chapters in the volume utilize novel measurement techniques to
examine evolutionary patterns. Lipo et al. (Chapter 12) and Goodale et al.
(Chapter13) both examine attributes associated with the haft element or base
of particular types of stone tools. Lipo etal. use the results to produce a seria-
tion and then discuss cultural inheritance in terms of geographic proximity.
Goodale et al. (Chapter 13) extends lithic technological organization by
examining projectile points in Southwest Asia and experimentally manufac-
tured points produced by Ishi, a member of the Yahi/Yana indigenous peoples
of north central California. They develop a technique to characterize projec-
tile point notching styles and use clustering techniques to isolate small groups
of similar specimens (presumably made by individual artisans).They argue that
the high morphological variation across the early Neolithic landscape may sig-
nal that the el-Khiam point was invented and spread through informal infor-
mation exchange without specific student to teacher learning. Their chapter
also emphasizes the importance of considering the measurements analysts take

on stone tools. Their argument stems from realizing which attributes reflect
original production and which reflect tool life history; the former rather than
the later are representative of evolutionary patterns concerning information
exchange and how to produce material culture.
Another chapter (Stevens, Chapter 14) dealing with cultural transmission
theory attempts to combine aspects of HBE with dual inheritance theory
(DIT) or cultural transmission, to investigate stone tool technology.This chap-
ter credits Julian Stewards brand of cultural ecology as the middle ground
between HBE and DIT. It uses the shift in relative proportions of multifunc-
tional tools to specialized tools over the past 10,000years in California to dem-
onstrate how the blended model can work. Stevens suggests that HBE explains
how subsistence changes appear while DIT provides a plausible evolutionary
mechanism for culture change given rules on how information is transmitted.
He says, HBE highlights the economic factors conditioning technological
change while DIT helps explain why technological changes might spread even
if specific groups are resistant. The interesting point of the chapter is that
emphasis is put on local contexts for any study (similar to what Steward would
emphasize). This suggests that any attempt to model individual decision mak-
ing should consider the context of the task, the available technology, and work
organization. We too feel these are important factors in any evolutionary
modeling program or lithic technology study and are too often overlooked
either by generalized models or detailed data analysis, respectively.
The Prentiss etal. chapter (this volume) does not use formal modeling and
might be considered a detailed contextual study of the slate tool industry at one
site.The chapter discusses and attempts to explain the evolution of the slate tool
industry in a complex hunting and gathering economy. They show that slate
technology can be viewed at multiple scales of artifact evolution (micro and
macro). They also show that at the micro evolutionary scale slate tools appear
to have a general trend toward increased production levels over time, suggesting
a process of selection for the tool. At the same time, overall, there is no indica-
tion of increased stylistic diversity or adoption of more slate tools with higher
levels of production effort in general at the site. However, they demonstrate that
more affluent households tended to have higher frequencies of the more costly
produced tools and the same households showed a greater frequency of stylistic
variability in slate tools. The authors suggest that the slate tool industry at the
Bridge River Site shows evolutionary change at multiple scales indicating that
group selection may be at play. They note, Membership in groups with integrated
socioeconomic and political strategies and with-group dominated transmission systems
may have offered stronger impacts on fitness than idiosyncratic tactics associated with (and
artifacts used by ...) individuals. Anti-conformist transmission is suggested as a
cause of differential stylistic markers for particular households. We believe this
chapter covers a great deal of ground from different scales of evolution to group
selection to models of neutrality related to stylistic differences. It too empha-
sizes context of study within an evolutionary perspective.
We have included the VanPool etal. chapter (Chapter4) within the discus-
sion associated with cultural transmission. But it could very well be included
within the section on culture history, where we discuss selectionist approaches
to evolutionary process. However, the emphasis of the culture historical sec-
tion is with stone tools and phylogenetic analysis. We feel the VanPool etal.
chapter emphasizes aspects of transmission associated with stimulated varia-
tion and reduced variation in populations even if much of the discussion and
diagrams relate to natural selection. This chapter clearly cross-cuts both evo-
lutionary approaches in the way Chapter14 by Stevens attempts to link HBE
and DIT. They used metric data from Paleoindian period sites from across the
Southwest and also examined single-site metric data from Blackwater Draw
in New Mexico. Both data sets show a bulge or increased variation during the
Late Paleoindian period with regard to projectile point attributes. This pattern
coincides nicely with the evolutionary notions of innovation (increased varia-
tions) in times of stress followed by selective forces to decrease variation.
We believe Chapter 4 by VanPool et al. is another example of common
ground gathering under the larger Darwinian umbrella, particularly in the
area of lithic technological studies. We interpret this chapter as one that
emphasizes human choice and ingenuity to stimulate variation in technology
when needed.They show stimulated variation is associated with climatic stress,
changes in residential sedentism, and other shifts. Those shifts are reflected as
human innovations and choices in technological variations, whether they be
horizontal via new group interactions or from other sources. The authors
note, Inventions ... can result from transmission errors, novel combinations
of previously existing variants, intentional efforts to improve the efficiency
of some technology, and a host of other factors. Perhaps a decade ago some
of the authors of this chapter would not have made such statements, instead
preferring to focus on the more concrete details of analyzing the historical
patterns of differential trait representation in the archaeological record. In any
event, this chapter shows concrete steps toward integrating Darwinian selec-
tion with aspects of cultural transmission using stone tool data and we think
it goes a long way toward solidifying lithic technological analysis within an
evolutionary framework.
Summary
We have chosen to use the term lithic technological system throughout this vol-
ume because we believe that it can be easily defined as a concept in which
stone and stone tools are under the adaptive umbrella that influences fitness
and reproductive success in individuals who use stone to make a living. As we

believe readers of this volume will come to realize, there is great potential for
interpreting stone tool assemblages that also extends to providing a theoretical
perspective that allows us to deal with time depth in the archaeological record
that other subfields of anthropology are able to avoid (Shott, Chapter3).
This volume stems from both the recent use of evolutionary theory in
lithic studies but also from the lack of theory generally used in lithic studies.
Historically, lithic studies have been focused on method building and analyt-
ical means partially because of access to ever advancing technology. One very
apparent example is the use and application of portable X-ray fluorescence
technology (pXRF) in sourcing studies which has allowed the attainment of
elemental chemistry much more efficiently. The important link here is that
data gained from pXRF or XRF technology in general are well suited to
applying concepts from behavioral ecology such as optimality models. This is
not to say this technology is without fault, because there are still challenges
on the horizon for integrating this and other technologies into lithic studies
(Goodale etal. 2011; Shackley 2010).
Recently there have been great advances toward understanding concepts
such as curation (e.g., Andrefsky 2008 and references therein) and its useful-
ness as a conceptual tool in lithic studies. In this volume we have tried to bring
together authors with specialties that can aid us in both using these recent
conceptual ideas such as artifact life-history and use them to move to the
theoretical level and apply concepts from evolution to understanding lithic
technological systems. This volume represents a culmination of those efforts.
We think this collection goes a long way toward merging evolutionary theory
with the interpretation of lithic technological systems.
References
Abbott, Alysia, L., Robert D. Leonard, and Andrefsky William, Jr. 2006. Experimental and
George T. Jones. 1996. Explaining the Change Archaeological Verification of an Index of
from Biface to FlakeTechnology:A Selectionist Retouch for Hafted Bifaces. American Antiquity
Application, in Darwinian Archaeologies, edited 71:743759.
by Herbert D. G. Maschner, pp. 3342, Plenum Andrefsky William, Jr. 2008. Projectile Point
Press, NewYork. Provisioning Strategies and Human Land-
Ahler, Stanley, and Phil R. Geib. 2000. Why Use. In Lithic Technology: Measures of Production,
Flute? Folsom Point Design and Adaptation. Use and Curation, edited by William Andrefsky,
Journal of Archaeological Science 27:799820. Jr., pp. 195214. Cambridge University Press,
Akridge, D. Glen, and Paul H. Benoit. 2001. Cambridge.
Luminescence Properties of Chert and Andrefsky, William, Jr., Jennifer M. Ferris,
Some Archaeological Applications. Journal of Justin P. Williams, Nathan B. Goodale, and
Archaeological Science 28:143151. George T. Jones. 2010. Geologic Contexts
Andrefsky William, Jr. 1994. The Geological and Diagnostic Provenance of Cherts, Paper
Occurrence of Lithic Material and Stone presented at the 2010 Annual Meeting of
Tool Production Strategies. Geoarchaeology: An the Society for American Archaeology, St.
International Journal 9:345362. Louis,MO.
Bamforth, Douglas B. 1986. Technological Bettinger, Robert L., Bruce Winterhalder, and
Efficiency and Tool Curation. American Richard McElreath. 2006. A Simple Model
Antiquity 51:3850. of Technological Intensification. Journal of
Bamforth, Douglas B. 2002. Evidence and Archaeological Science 33:538545.
Metaphor in Evolutionary Archaeology. Bleed, Peter. 1986. The Optimal Design of
American Antiquity 435452. Hunting Weapons: Maintainability or
Bamforth, Douglas B., and Mark S. Becker. 2000. Reliability. American Antiquity 51:737747.
Core/Biface Ratios, Mobility, Refitting, and Bliege Bird, Rebecca, and Eric Alden Smith.
Artifact Use-Lives: A Paleoindian Example. 2005. Signaling Theory, Strategic Interaction
Plains Anthropologist 45:273290. and Symbolic Capital. Current Anthropology
Beck, Charlotte. 1995. Functional Attributes and 46(2):221248.
the Differential Persistence of Great Basin Boone, James L. 1992. Conflict, Competition, and
Dart Forms. Journal of California and Great the Emergence of Hierarchies. In Evolutionary
Basin Anthropology 17:222243. Ecology and Human Behavior, edited by Eric
Beck, Charlotte, and George T. Jones. 2010. Clovis Alden Smith and Bruce Winterhalder, Aldine
and Western Stemmed: Population Migration Transaction, New Brunswick,NJ.
and the Meeting of the Two Technologies in Boone, James L., and Eric Alden Smith. 1998. Is
the Intermountain West. American Antiquity It Evolution Yet? A Critique of Evolutionary
75:81116. Archaeology. Current Anthropology
Beck, Charlotte, Amanda K. Taylor, George T. 39(Supp):141173.
Jones, Cynthia M. Fadem, Caitlyn R. Cook, Bousman, C. Britt. 2005. Coping with Risk:
and Sara A. Millward. 2002. Rocks Are Later Stone Age Technological Strategies
Heavy: Transport Costs and Paleoarchaic at Blydefontein Rock Shelter, South
Quarry Behavior in the Great Basin. Journal Africa. Journal of Anthropological Archaeology
of Anthropological Archaeology 21:481507. 24:193226.
Bement, Leland. 2002. Pickin Up the Pieces: Boyd, Robert, and Peter J. Richarson. 1985.
Folsom Projectile Point Resharpening Culture and the Evolutionary Process. University
Technology. In Folsom Technology and Lifeways, of Chicago Press, Chicago.
edited by J. E. Clark and M. B. Collins, pp. 135 Boyd, Robert, and Peter J. Richarson. 1992.
140, Lithic Technology Special Publication How Microevolutionary Processes Give Rise
No.4. to History. In History and Evolution, edited by
Bettinger, Robert L., Robert Boyd, and M. H. Nitecki and D. V. Nitecki, pp. 179209,
Peter J. Richardson. 1996. Style, Function, SUNY Press, Albany.
and Cultural Evolutionary Process. In Brantingham, P. Jeffrey. 2003. A Neutral Model of
Darwinian Archaeologies, edited by Herbert Stone Raw Material Procurement. American
D. G. Maschner, pp. 133164, Plenum Press, Antiquity 68:487509.
NewYork. Brantingham, P. Jeffrey. 2006. Measuring Forager
Bettinger, Robert L., and Eerkens, Jelmer W. Mobility. Current Anthropology 47:435459.
1997. Evolutionary Implications of Metrical Brantingham, P. Jeffrey, John W. Olsen, Jason
Variation in Great Basin Projectile Points. In A. Rech, and Andrei I. Krivoshapkin. 2000.
Rediscovering Darwin, edited by C. M. Barton Raw Material Quality and Prepared Core
and G. Clark, pp. 177191. Archaeological Technologies in Northeastern Asia. Journal of
Papers of the American Anthropological Archaeological Science 27:255271.
Association No. 7, Arlington,VA. Braun, David R., Thomas Plummer, Joseph
Bettinger, Robert L., and Eerkens, Jelmer W. V. Ferraro, Peter Ditchfield, and Laura C.
1999. Point Typologies, Cultural Transmission, Bishop. 2009. Raw Material Quality and
and the Spread of Bow and Arrow Technology Oldowan Hominin Toolstone Preferences:
in the Prehistoric Great Basin. American Evidence from Kanjera South Kenya. Journal
Antiquity 64:231242. of Archaeological Science 36:16051614.
Buchanan, Briggs, and Mark Collard. 2007. Middle Paleolithic Assemblage Variability.
Investigating the Peopling of North America In Raw Material Economies among Prehistoric
through Cladistic Analysis of Early Paleoindian Hunter-Gatherers. edited by A. Montet-
Projectile Points. Journal of Anthropological White and S. Holen, pp. 3348. University
Archaeology 26:366393. of Kansas Publications in Anthropology, 19,
Buchanan, Briggs, and Mark Collard. 2008. Lawrence,KS.
Phenetics, Cladistics, and the Search for the Dunnell, Robert C. 1978. Style and Function: A
Alaskan Ancestors of the Paleoindians: A Fundamental Dichotomy. American Antiquity
Reassessment of Relationships among the 43:192202.
Clovis, Nenana, and Denali Archaeological Dunnell, Robert C. 1980. Evolutionary Theory
Complexes. Journal of Archaeological Science and Archaeology. Advances in Archaeological
35:16831694. Method and Theory 3:3599.
Buchanan, Briggs, and Marcus J. Hamilton. 2009. Dunnell, Robert C. 1982. Science, Social
A Formal Test of the Origin of Variation in Science, and Common Sense: The Agonizing
North American Early Paleoindian Projectile Dilemma of Modern Archaeology. Journal of
Points. American Antiquity 74:279298. Anthropological Research 38:125.
Cavalli-Sforza, L. L., and Marc Feldman. Eerkens, Jelmer W. 2000. Practice Makes within
1981. Cultural Transmission and Evolution: A 5% of Perfect: The Role of Visual Perception,
Quantitative Approach. Princeton University Motor Skills, and Human Memory in Artifact
Press, Princeton,NJ. Variation and Standardization. Current
Charnov, Eric L. 1976. Optimal Foraging: The Anthropology 41:663668.
Marginal Value Theorem. Theoretical Popu Eerkens, Jelmer W., Robert L. Bettinger, and
lation Biology 9:129136. Richard McElreath. 2006. Cultural Trans
Clarkson, Chris. 2002. An Index of Invasiveness mission, Phylogenetics, and the Archaeological
for the Measurement of Unifacial and Bifacial Record. In Mapping Our Ancestors: Phylogenetic
Retouch: A Theoretical, Experimental Approaches in Anthropology and Prehistory, edited
and Archaeological Verification. Journal of by C. P. Lipo, M. J. OBrien, S. J. Shennan, and
Archaeological Science 29:6575. M. Collard, pp. 169183, Aldine Transaction,
Clarkson, Chris. 2008. Changing Reduction New Brunswick,NJ.
Intensity, Settlement and Subsistence in Eerkens, Jelmer W., Jeffery R. Ferguson, Michael
Wardaman Country, Northern Australia. In D. Glascock, Craig E. Skinner, and Sharon
Lithic Technologies: Life-Cycles of Production A. Waechter. 2007. Reduction Strategies and
and Retouch, edited by William Andrefsky, Jr., Geochemical Characterization of Lithic
pp. 286316, Cambridge University Press, Assemblages: A Comparison of Three Case
Cambridge. Studies from Western North America.
Daniel, I. Randolf, Jr. 2001. Stone Raw Material American Antiquity 72:585597.
Availability and Early Archaic Settlement in Eerkens, Jelmer W., and Carl P. Lipo. 2005.
the Southeastern United States. American Cultural Transmission, Copying Errors, and the
Antiquity 66:237266. Generation of V ariation in Material Culture
Darwent, John, and Michael J. OBrien. 2006. and the Archaeological Record. Journal of
Using Cladistics to Construct Lineages of Anthropological Archaeology 24:316334.
Projectile Points from Northeastern Missouri. Eerkens, Jelmer W., and Carl P. Lipo. 2007.
In Mapping Our Ancestors: Phylogenetic Cultural Transmission Theory and the
Approaches in Anthropology and Prehistory, edited Archaeological Record: Providing Context
by C. P. Lipo, M. J. OBrien, M. Collard, and to Understanding Variation and Temporal
S. J. Shennan, pp. 185208, Aldine Transaction, Changes in Material Culture. Journal of
New Brunswick,NJ. Archaeological Research 15:239274.
Dibble, Harold L. 1991. Local Raw Material Feblot-Augustins, Jehanne. 1997. Middle and
Exploitation and Its Effects on Lower and Upper Paleolithic Raw Material Transfers in
Western and Central Europe: Assessing the Goodyear, Albert C. 1974. The Brand Site: A
Pace of Change. Journal of Middle Atlantic Techno-Functional Study of a Dalton Site in
Archaeology 13:5790. Northeast Arkansas. Arkansas Archaeological
Feblot-Augustins, Jehanne. 2009. Revisiting Survey Publications on Archaeology, Research
European Upper Paleolithic Raw Material Series7.
Transfers: The Demise of the Cultural Hamilton, Marcus, and Briggs Buchanan.
Ecological Paradigm? In Lithic Materials and 2009. The Accumulation of Stochastic
Paleolithic Societies, edited by Brian Adams and Copying Errors Causes Drift in Culturally
Brooke Blades, pp. 2546, Wiley-Blackwell, Transmitted Technologies: Quantifying
Oxford. Clovis Evolutionary Dynamics. Journal of
Fitzhugh, Ben. 2001. Risk and Invention in Anthropological Archaeology 28:5569.
Human Technological Evolution. Journal of Hawkes, Kristen, and Rebecca Bliege Bird.
Anthropological Archaeology 20:125167. 2002. Showing Off, Handicap Signaling, and
Foley, Robert. 1987. Hominid Species and Stone the Evolution of Mens Work. Evolutionary
Tool Assemblages: How Are They Related? Anthropology 11(2):5867.
Antiquity 61:380392. Hiscock, Peter. 2009. Reduction, Recycling,
Foley, Robert, and Marta M. Lahr. 2003. On and Raw Material Procurement in Western
Stony Ground: Lithic Technology, Human Arnhem Land, Australia. In Lithic Materials and
Evolution, and the Emergency of Culture. Paleolithic Societies, edited by Brian Adams and
Evolutionary Anthropology 12:109122. Brooke Blades, pp. 7894, Wiley-Blackwell,
Foradas, James G. 2003. Chemical Sourcing of Oxford.
Hopewell Bladelets: Implications for Building Hiscock, Peter, and Val Attenbrow. 2003. Early
a Chert Database for Ohio. In Written in Stone: Australian Implement Variation: A Reduction
The Multiple Dimensions of Lithic Analysis, edited Model. Journal of Archaeological Science
by P. Nick Kardulias and Richard W. Yerkes, 30:239249.
pp. 87112, Lexington Books, Lanham,MD. Hiscock, Peter, and Chris Clarkson. 2007.
Gabora, Liane. 2006. The Fate of Evolutionary Retouched Notches at Combe Grenal
Archaeology: Survival or Extinction? World (France) and the Reduction Hypothesis.
Archaeology 38:690696. American Antiquity 72:176190.
Goodale, Nathan, David G. Bailey, George T. Hogberg, Anders, and Deborah Olaussen. 2009.
Jones, Catherine Prescott, Elizabeth Scholz, Scandinavian Flint: An Archaeological Perspective.
Nicholas Stagliano, and Chelsey Lewis. Ashus University Press, Denmark.
2011. pXRF: A Study of Inter-instrument Hughes, Richard E., Anders Hogberg, and
Performance. In press Journal of Archaeological Deborah Olausson. 2010. Sourcing Flint
Science 39:875883. from Sweden and Denmark: A Pilot Study
Goodale, Nathan B., Ian Kuijt, Shane MacFarlan, Employing Non-destructive Energy
Curtis Osterhoudt, and Bill Finlayson. 2008. Dispersive X-ray Fluorescence Spectrometry.
Lithic Core Reduction Techniques: A Model Journal of Nordic Archaeological Science 17:1525.
for Predicting Expected Diversity. In Lithic Hughes, Susan S. 1998. Getting to the Point:
Technology: Measures of Production, Use, and Evolutionary Change in Prehistoric Weaponry.
Curation. edited by William Andrefsky, Jr., Journal of Archaeological Method and Theory
pp. 317336, Cambridge University Press, 5:345408.
Cambridge. Jennings, Jesse D. 1968. Prehistory of North America.
Goodale, Nathan, Heather Otis, William 2nd ed. McGraw-Hill, NewYork.
Andrefsky Jr., Ian Kuijt, Bill Finlayson, and Jeske, Robert J. 1989. Economies in Raw Material
Ken Bart. 2010. Sickle Blade Life-History Use by Prehistoric Hunter-Gatherers. In
and the Transition to Agriculture: An Early Time, Energy, and Stone Tools, edited by R.
Neolithic Case Study from Southwest Asia. Torrence, pp. 3445, Cambridge University
Journal of Archaeological Science 37:11921201. Press, Cambridge.
Jones, George T., Charlotte Beck, Eric E. Jones, Lyman, R. Lee, and Michael J. OBrien. 2000.
and Richard E. Hughes. 2003. Lithic Source Measuring and Explaining Change in Artifact
Use and Paleoarchaic Foraging Territories in Variation with Clade-Diversity Diagrams.
the Great Basin. American Antiquity 68:538. Journal of Anthropological Archaeology 19:3974.
Jones, George T., Robert D. Leonard, and Alyssa Lyman, R. Lee, Todd L. VanPool, and Michael J.
L. Abbott. 1995. The Structure of Selectionist OBrien. 2008. Variation in North American
Explanations in Archaeology. In Evolutionary Dart and Arrow Points When One or Both
Archaeology: Methodological Issues, edited by P. Are Present. Journal of Archaeological Science
A. Teltser, pp. 1332, University of Arizona 35:28052812.
Press, Tucson. Lyman, R. Lee, Todd L. VanPool, and Michael
Kelly, Robert L. 1988. The Three Sides of a J. OBrien. 2009. The Diversity of North
Biface. American Antiquity 53:717734. American Projectile-point Types, Before
Kelly, Robert L., and Lawrence C. Todd. 1988. and After the Bow and Arrow. Journal of
Coming into the Country: Early Paleoindian Anthropological Archaeology 28:113.
Hunting and Mobility. American Antiquity Lyons,William H., Michael D. Glascock, and Pete
53:231244. Mehringer. 2003. Silica from Sources to Sites:
Krieger, Alex D. 1944. The Typological Concept. Ultraviolet Fluorescence and Trace Elements
American Antiquity 9:271288. Identify Cherts from Lost Dunes, Southeastern
Kuhn, Steven L. 2004. Upper Paleolithic Raw Oregon, USA. Journal of Archaeological Science
Material Economies at Ucagizli Cave, 30:11391159.
Turkey. Journal of Anthropological Archaeology Mayr, Ernst. 1982. Systematics and the Origin of
23:431448. Species, Columbia Classics in Evolution Series.
Leonard, Robert D., and George T. Jones. 1987. Columbia University Press, NewYork.
Elements of an Inclusive Evolutionary Model McGuire, Kelly R., and William R. Hildebrandt.
for Archaeology. Journal of Anthropological 2005. Re-Thinking Great Basin Foragers:
Archaeology 6:199219. Prestige Hunting and Costly Signaling during
Lipo, Carl P., M. Madsen, Robert Dunnell, the Middle Archaic Period. American Antiquity
and Terry Hunt. 1997. Population Structure, 70(4):695712.
Cultural Transmission, and Frequency Meltzer, David J. 1981. A Study of Style and
Seriation. Journal of Anthropological Archaeology Function in a Class of Tools. Journal of Field
16:301333. Archaeology 8:313326.
Lipo, Carl P., Michael J. OBrien, Mark Mesoudi, Alex, and Michael J. OBrien. 2008.
Collard, and Stephan J. Shennan. 2006. The Cultural Transmission of Great Basin
Cultural Phylogenies and Explanation: Why Projectile-Point Technology:An Experimental
Historical Methods Matter. In Mapping Simulation. American Antiquity 73:328.
Our Ancestors: Phylogenetic Approaches in McKern, William C. 1939. The Midwestern
Anthropology and Prehistory, edited by C. P. Taxonomic Method as an Aid to Archaeological
Lipo, M. J. OBrien, M. Collard, and S. J. Culture Study. American Antiquity 4:301313.
Shennan, pp. 318, Aldine Transaction, New Mithen, Steven. 2008. Whatever Turns You On:
Brunswick,NJ. A Response to Anna Machin, Why Handaxes
Lurie, Rochelle. 1989. Lithic Technology and Just Arent That Sexy. Antiquity 82:766769.
Mobility Strategies: The Koster Site Middle Neiman, Fraser. 1990. An Evolutionary Approach
Archaic. In Time, Energy, and Stone Tools, to Archaeological Inference: Aspects of Architectural
edited by R. Torrence, pp. 4656, Cambridge in the 17th-Century Chesapeake. Ph.D. disserta-
University Press, Cambridge. tion, Yale University. University Microfilms,
Lyman, R. Lee, and Michael J. OBrien. 1998. Ann Arbor.
The Goals of Evolutionary Archaeology: Neiman, Fraser. 1995. Stylistic Variation in
History and Explanation. Current Anthropology Evolutionary Perspective: Inferences from
39:615662. Decorative Diversity and Interassemblage
Distance in Illinois Woodland Ceramic Measures of Production Use and Curation,

Assemblages. American Antiquity 60:736. edited by William Andrefsky, Jr., pp. 150174.
OBrien, Michael J., ed. 2008. CulturalTransmission Cambridge University Press, Cambridge.
and Archaeology: Issues and Case Studies. SAA Ramenofsky, Ann F., Fraser Neiman, and
Press, Washington,DC. Christopher D. Pierce. 2009. Measuring Time,
OBrien, Michael J., John Darwent, and R. Population, and Residential Mobility from
Lee Lyman. 2001. Cladistics Is Useful for the Surface at San Marcos Pueblo, North
Reconstructing Archaeological Phylogenies: Central New Mexico. American Antiquity
Palaeoindian Points from the Southeastern 74:505530.
United States. Journal of Archaeological Science Ritchie, William. 1932. The Lamoka Lake Site.
28:11151136. Researches and Transactions of the New York
OBrien, Michael J., and Thomas D. Holland. State Archaeological Association,Vol. 7, No. 4,
1990. Variation, Selection, and the Rochester.
Archaeological Record. Archaeological Method Sellet, Frederic, James Donohue, and Matthew
and Theory 2:3179. G. Hill. 2009. The Jim Pitts Site: A Stratified
OBrien,Michael J.,andThomas D.Holland.1992. Paleoindian Site in the Black Hills of South
The Role of Adaptation in Archaeological Dakota. American Antiquity 74:735758.
Explanation. American Antiquity 57:359. Shackley M. Steven. 2010. Is There Reliability
OBrien, Michael J., and R. Lee Lyman. 2000. and Validity in Portable X-Ray Fluorescence
Applying Evolutionary Archaeology: A Systematic Spectrometry? SAA Archaeological Record
Approach. Kluwer Academic/Plenum Press, 10:1720.
NewYork. Shackley M. Steven. 2005. Obsidian: Geology and
OBrien, Michael J., and R. Lee Lyman. 2002. Archaeology in the North American Southwest.
Evolutionary Archaeology: Current Status and University of Arizona Press, Tucson.
Future Prospects. Evolutionary Anthropology Shennan, Stephen J. 2002. Population, Culture
11:2636. History and the Dynamics of Culture Change.
OBrien, Michael J., R. Lee Lyman, and Robert Current Anthropology 41:811835.
D. Leonard. 1998. Basic Incompatibilities Shennan, Stephen J., and James R. Wilkinson.
between Evolutionary and Behavioral 2001. Ceramic Style Change and Neutral
Archaeology. American Antiquity 63:485498. Evolution: A Case Study from Neolithic
OBrien, Michael J., R. Lee Lyman, and Robert Europe. American Antiquity 66:577593.
D. Leonard. 2003. What Is Evolution? A Shott, Michael J. 1986. Settlement Mobility and
Response to Bamforth. American Antiquity Technological Organization:An Ethnographic
68:573580. Examination. Journal of Anthropological Research
OConnell, James F. 1995. Ethnoarchaeology 42:1551.
Needs a General Theory of Behavior. Journal Shott, Michael J. 1996. Innovation and Selection
of Archaeological Research 3:205255. in Prehistory:A Case Study from the American
Orr, Elizabeth L., William N. Orr, and Ewart M. Bottom. In Stone Tools: Theoretical Insights into
Baldwin. 1999. Geology of Oregon. Kendall/ Human Prehistory, edited by G. H. Odell, pp.
Hunt, Dubuque,IA. 279314. Plenum Press, NewYork.
Parry, William J., and Robert L. Kelly. 1987. Shott, Michael J. 2008. Darwinian Evolutionary
Expedient Core Technology and Sedentism. Theory and Lithic Analysis. In Cultural
In The Organization of Core Technology, edited Transmission and Archaeology: Issues and Case
by J. K. Johnson and C. A. Morrow, pp. 285 Studies, edited by M. J. OBrien, pp. 146157,
304. Westview Press, Boulder,CO. SAA Press, Washington,DC
Quinn, Colin Patrick, William Andrefsky, Jr. Ian Shott, Michael J., and Jesse A.M. Ballenger. 2007.
Kuijt, and Bill Finlayson. 2008. Perforation Biface Reduction and the Measurement of
with Stone Tools and Retouch Intensity: Dalton Curation: A Southeastern Case Study.
A Neolithic Case Study In Lithic Technology: American Antiquity 72:153175.
Surovell, Todd A. 2009. Toward A Behavioral Truncer, James J. 1990. Perkiomen Points:A Study
Ecology of Lithic Technology: Cases from in Variability. In Experiments and Observations
Paleoindian Archaeology, University of Arizona on the Terminal Archaic of the Middle Atlantic
Press, Tucson. Region, edited by Roger W. Moeller, pp.162.
Terry, Karisa, William Andrefsky, Jr., and Archaeological Services, Bethlehem,CT.
Mikhail V. Konstantinov. 2009. Raw Material Ugan, Andrew, Jason Bright, and Alan Rogers.
Durability, Function, and Retouch in the 2003. When Is Technology Worth the
Upper Paleolithic of the Transbikal Region. Trouble? Journal of Archaeological Science
In Lithic Materials and Paleolithic Societies, 30:13151329.
edited by Brian Adams and Brooke Blades, pp. Wallace, Ian J., and John J. Shea. 2006. Mobility
256269, Wiley-Blackwell, Oxford. Patterns and Core Technologies in the
Tomka, Steve A. 2001. The Effect of Processing Middle Paleolithic of the Levant. Journal of
Requirements on Reduction Strategies and Archaeological Science 33:12931309.
Tool Form: A New Perspective. In Lithic Winterhalder, Bruce D., and Eric A. Smith. 1992.
Debitage: Context Form Meaning, edited by Evolutionary Ecology and the Social Sciences.
William Andrefsky, Jr., pp. 207225, University In Evolutionary Ecology and Human Behavior,
of Utah Press, Salt LakeCity. edited by E. A. Smith and B. D. Winterhalder,
Torrence, Robin. 1983. Time Budgeting and pp. 323, Aldine de Gruyter, NewYork.
Hunter-Gatherer Technology. In Hunter- Wittohft. John. 1949. An Outline of Pennsylvania
Gatherer Economy in Prehistory, edited by G. Indian History, Pennsylvania History, Quarterly
Bailey, pp. 1122. Cambridge University Press, Journal of Pennsylvania Historical Association,Vol.
Cambridge. 16, No. 3, pp. 315, Harrisburg.
PartII
Culture History and Phylogenetic

Evolution
Two
Graphing Evolutionary Pattern

in Stone Tools to Reveal
Evolutionary Process
R. LeeLyman
An old clich is that a picture is worth a thousand words. In disciplines that

seek to understand empirical phenomena that is, in the sciences the pic-
tures tend to be graphs of myriad kinds of data. Graphs that are well designed
will display relationships of variables patterns in such a way as to prompt
understanding of the structural, functional, or processual nature of the rela-
tionships with a minimum of verbal explanation (Collier 2008 and refer-
ences therein). Further, among well-designed graphs, graph perception the
visual decoding of information encoded in graphs (Cleveland and McGill
1984:531) will not be influenced by the structure or mode of presentation.
The importance of graphs for not only summarizing data and analytical results,
but also for detecting patterns of relationships, cannot be overstated. It is likely
because of this importance that over the past few years several investigators
have found it necessary to study how graphs are designed and how they are
interpreted.This is particularly evident in evolutionary biology where the label
phylogenetic tree has been applied to a large variety of graphs with minimal
explanation on the part of the graphic design artist and a concomitant lack of
understanding on the part of the person studying the graph (e.g., Baum etal.
2005; Catley and Novick 2008; Crisp and Cook 2005; Hunt 2008; Novick
and Catley 2007). In many cases in biology perception of graphs of evolving
phenomena has resulted in misinterpretation of evolutionary process because
of a poorly designed graph of evolutionary pattern. An historical example in
American archaeology was Julian Stewards (1944) and Harold S. Gladwins
29
30 R. LeeLy m a n
(1936) misinterpretation of the taxonomic architectural patterns of particular

classification systems as depicting the descent with modification process of a
phylogeny when in fact both taxonomies they criticized instead represented
static aggregative hierarchies (Valentine and May 1996) rather than a mecha-
nistic descent process.
Virtually ever since archaeology became a scholarly pursuit, practitioners
have sought to not only document but also graphically illustrate cultural change
or evolution (see reviews in Lyman 2009; Lyman etal. 1998). Toward that end,
they invented various styles of graph to illustrate cultural evolution, including
E. B. Sayless (1937) interpretive diamond graph and Jim Fords (1952, 1962)
empirical battleship (frequency distribution) curves. Similarly, beginning in the
late nineteenth century paleontologists and evolutionary biologists worked to
document biological evolution and developed various graph styles to illustrate
change from one animal form to another. Some of the most famous illustra-
tions were by Owen Marsh (1879, 1892) and showed the evolution of North
American horses as a phyletic lineage implying anagenetic or linear evolution.
The new notion in the early twentieth century that the individual organisms
of a species were morphologically variable resulted in a shift in graph style to
illustrate the variation in the form of a frequency distribution of phenotypic
variants; a series of histograms showed the variants and each histogram repre-
sented a different geological stratum or temporal period (e.g., Newell 1947).
Despite the common analytical goals of paleontology and archaeology, there
has been minimal cross-disciplinary pollination with respect to graph styles
(see Lyman and OBrien [2000] for an exception). This is obvious from the
distinct graph styles used to illustrate variability through time in each disci-
pline; archaeologists use centered bar graphs whereas paleontologists use a
chronologically ordered series of histograms. That training influences prefer-
ence for graph style is apparent from the fact that at least some paleontologists
find the archaeological centered bar graph difficult to interpret (Lyman 2009).
As an archaeologist I find a standard paleontological series of histograms inter-
pretable but only with more effort than centered bar graphs that show exactly
the same information.
The goal I originally set for this chapter was to demonstrate that different
graph styles highlight different evolutionary patterns or attributes thereof, and
thereby to underscore the fact that different graph styles should be not be
viewed as all of a piece, but rather distinct yet complementary. Further, I origi-
nally hoped to show that use of multiple graph styles to illustrate evolutionary
pattern and also as explicit analytical techniques should help us avoid the pitfall
of presuming that a particular pattern illustrated by a graph correlates directly
and perfectly with a particular evolutionary process. Although I did not attain
my original goals, I believe some important lessons are learned along the way.
One of those lessons concerns the fact that although I use lithic artifacts as
Graphi ng Evo l u t i o n a ry Pat t e r n i n S to n e T o o l s 31
my data source, the techniques and principles underpinning the graph styles I
present are applicable to cultural phenomena of any type and scale, in so far as
the data that are used are quantitative (metric [size] or meristic [frequency]).
I begin with a bit of background on how the modern theory of biological
evolution conceives the patterns and processes of evolution. I then introduce
the data set used to construct the graph styles illustrated here. Next, seven
graph styles, most of which should be familiar to archaeologists, are described.
Whether or not each graph style illustrates an evolutionary pattern is assessed,
and what each might signify about evolutionary process is suggested. I con-
clude with a few suggestions about the future use of particular graph styles.
A Bit of Background
The fundamental tenet of modern Darwinian evolution is that a sorting
mechanism natural selection, drift, whatever works on a diversity of vari-
ants, resulting in some variants replicating more frequently than others. Add
a new variant or two every so often (the source of the novel variant is not
important at this very general level) and a continuous evolutionary process is
the result. Darwinian evolutionary change is, in this case, defined as changes
in frequencies of specimens of kinds and not changes in the kinds, although
evolutionary turnover means that kinds A, B, C, and D will be represented
early in the sequence; kinds C, D, E, and F will be represented in the middle of
the sequence; and kinds E, F, G, and H will be represented late in the sequence.
This variational theory of evolution is easily depicted and contrasted with the
Midas-touch transformational theory in which all individuals are identical and
are replaced simultaneously with a different form (Figure2.1).
A B B B B C B C C B D D C D D D E E
B A C B B A B C D C C D C D D D E E
B B A B C C B C C C C D D E E D E F
Variational (Darwinian) model
time
Transformational model
A A A B B B C C C E E E F F F G G G
A A A B B B C C D E E E F F F G G G
A A A B B B C C C E E E F F F G G G
2.1.The transformational model and the variational (Darwinian) model of evolution. Conceive
of each individual letter as an individual organism, and each kind of letter as a unique species
(A, species 1; B, species 2; etc.).
32 R. LeeLy m a n
a 10 f 10 m 10
a9 f9 m9
a8 f8 k8 l8 m8
a7 f7 k7 l7 m7
a6 f6 k6 m6
a5 d5 k5 m5
a4 d4 i4 m4
a3 3i m3
a2 s2 m2
a1 m1
2.2. Darwins (1859) model of evolutionary pattern. Although inexplicit in Darwins (1859) dis-
cussion, time passes from the bottom to the top of the graph. Each letter represents one taxon
(e.g., genus), and each number represents a finer level taxon (species) in the sequence.The graph
has been said to denote linear (anagenetic) evolution, and in a sense it does (the a lineage, the
f lineage, the m lineage, and so on), although some branching (cladogenesis) is apparent.
The graph also suggests that few of many forms (each represented by a branch or twig) replicate
successfully over the long term.(Redrawn from Darwin 1859.)
Darwin (1859) chose to illustrate evolution with a model that is readily cat-
egorized as a phylogeny (Figure2.2). It does not come close to depicting the
variational model. It does, however, do two things. First, it reveals the scale at
which Darwin was thinking the individual species or kind, rather than the
attribute states held by an individual, though we must not lose sight of the fact
that it is variation in attributes or character states that prompts us to place dis-
similar specimens in different categories (Lyman and OBrien 2002; OBrien
and Lyman 2002). Second, as advocates of punctuated equilibria are fond of
pointing out, Darwins graph of evolutionary pattern implies the process of
anagenesis, or phyletic or linear evolution as opposed to cladogenesis or diver-
sifying or branching evolution (Eldredge and Gould 1972; Stanley 1975).These
two basic modes of evolution anagenesis and cladogenesis are sometimes
the source of the problems with graph perception that occur today in biology.
Favoring one might cause the viewer to suggest that a particular graph pattern
illustrates that favored mode (process) of evolution (e.g., Fortey 1985). Fred
Plog (1973) recognized a similar problem in archaeology 35years ago when he
(e)
(d)
(c)
(b)
(a)
2.3. Fred Plogs (1973) seriogram graph of continuous cultural change. Each lettered poly-
gon represents a different type of artifact; polygon width represents the frequency of speci-
mens of the type. Time can pass from bottom to top, or top to bottom.(Redrawn from Plog
1973.)
pointed out that our chronologies of cultures cause us to think of time as a

series of successive [static] units rather than as a continuous flow (p.189). The
way to avoid the resulting graph (mis)perception, in Plogs (1973:191) view,
was to graph change in artifacts in the form of what he called a seriogram
(Figure2.3). Such a graph style epitomizes the Darwinian variational model of
evolution (Figure2.1). It does so because it graphs the history of the frequen-
cies of specimens in each of several types, and with times passage on the verti-
cal axis, at least some of those types overlap (occur in) contiguous time periods.
Such overlap is explained in Darwinian terms as reflecting heritable continuity,
or in more modern lingo, cultural transmission. Not all graph styles produced
by archaeologists imply the Darwinian variational model, largely because they
lack one or more of the necessary features of that model.
Data
I illustrate several graph styles here using quantitative data for a collection of
stone projectile points. Most of the graphs should be familiar to archaeologists.
The data were originally collected by Susan Hughes from projectile points
recovered from Mummy Cave in northwestern Wyoming. The collection is
made up of more than 300 points and spans 9200 radiocarbon years distrib-
uted across more than two dozen strata. Hughes presented graphic depictions
of various attribute combinations and ratios in her 1998 publication Getting
34 R. LeeLy m a n
table2.1. Frequencies of projectile points used in analyses

and age per stratum at MummyCave
Stratum No. of points Age (radiocarbon years BP)

1 11 370
2 11 ?
3 64 1230
45 35 20502820
67 72 40904420
89 79 46405390
1013 26 56007600
1418 12 81008500
1912 20 85009250
Source: Data from Hughes (1998, and 2007 personal
communication).
to the Point; her analytical goal in that paper was to distinguish points that
tipped thrusting spears from those that tipped atlatl darts, and these two from
points that tipped arrows. To make the distinctions, she focused on derived
measures made up of combinations of attributes (e.g., width: length ratio) and
produced one basic kind of graph with two variants. One variant showed the
mean and range of a derived measure for all points per stratum; the other var-
iant showed values for individual points per stratum.
I do not question Hughess (1998) interpretations of the shift from darts to
arrows and the like. I have a different question in mind. I am interested in what
different graph styles and graphed patterns reveal about the evolution of pro-
jectile point form as evidenced by the Mummy Cave stone projectile points. I
take an unabashedly inductive, empirical approach to the exploration of differ-
ent graph styles. The process creating the pattern shown by a particular graph
style is an inference and thus different inferences derived from different graph
styles might be used to cross-check each other, thereby avoiding the potentially
skewed influence of one illustrated pattern on inferences of process. For my
purposes here, I have lumped some sets of adjacent strata with small collections
to reduce the effects of small samples on graphed patterns (Table2.1).The result
is nine temporally distinct sequent assemblages that I refer to as a stratum or
stratigraphic assemblage for convenience. Combining or lumping strata is no
cause for concern given that the individual stratigraphic units in the cave were
variously aggregated by the original investigators (including Hughes).
Graph Styles
It is important to realize up front that evolution in biology occurs at two general
scales of inclusiveness and specificity. At the more inclusive and more general
level, macroevolution occurs at the scale of taxa above the species level. Shifts
in the identities of species in a lineage or clade are the subject of concern. At

a less inclusive and more specific level microevolution concerns turnover in the
individuals making up a population of a species such that one or more central
tendencies of attributes of the population (pelt color, body length, number of
teeth) change over time, perhaps eventually leading to change in the identity
of the species represented.Two scales of cultural evolution as manifest archaeo-
logically are also easily identified for my purposes. One concerns turnover in
the types (taxa) of discrete objects we call artifacts; the other concerns changes
in the values, magnitudes, or states of attributes of discrete objects. The first is
more general and inclusive whereas the latter is more specific and exclusive.
Graph styles designed by both paleobiologists and archaeologists have involved
one or the other scale of inclusiveness, and I treat them separatelyhere.
At the Scale of Discrete Object

Turning to the Mummy Cave data, I begin with what is only loosely con-
strued as a graph in the usual sense although it does display the relationship
between projectile point form and stratigraphy or time. The two published
manifestations of the illustration differ slightly in detail but the style of the
illustrations is identical (Husted and Edgar 2002; Wedel et al. 1968). Both
manifestations show the chronological sequence of point forms, but whether
or not any evolution in the sense of sociocultural transmission accompanied
by change is shown is not clear (Figure2.4). A general process of the evolu-
tion of point form might be inferred from some of the similar forms of points
in chronologically adjacent strata, and if so, then a form of phyletic or linear
evolution would seem to be implied as, for example, lanceolate points shift
from having concave bases to having straight bases, and parallel sided stems
shift (evolve) to distally expanding stems. In all fairness to the authors of
these illustrations, they seem to have been mostly interested in documenting
the deep and essentially continuous stratigraphic record of artifacts at the site
and the chronological implications of the different point forms rather than
documenting any kind of evolution in point form (Husted and Edgar 2002;
Wedel etal. 1968).
The stratigraphic sequence illustration plots data at the scale of discrete
object, in this case, types of points. This style of graph should be familiar to
archaeologists.Another graph style that should also be familiar to archaeologists
also concerns data at the scale of discrete objects or types of points. This other
graph style constitutes what Gordon Willey and others referred to as percent-
age stratigraphy more than 60years ago (Lyman etal. 1998). Figure2.5 shows
the relative frequency of point types per stratum at Mummy Cave. Some over-
lap of point types from one stratum to the next is indicated, and this overlap
indicates continuity, heritability, and (vertical) cultural transmission over time.
36 R. LeeLy m a n
m
tu
ra
st
1 370 90 1
2 2
3
3 1230 110
4
4 2050 150
5
5 2820 135
6 4090 140 6
7 4420 150
7
4640 140
8 8
5250 140
9 5390 140
10 5610 280
9
11 5800 120
10
12 6400 75
6780 130
11
13 7630 170
12
14
13
14
15 7970 210
15
16
17
16 8100 130 18
19
17 8136 90 20
18 21
19 8307 78
20 8430 90 22
8465 90
21 23
22
24
9230 150
23 9248 79
24
10,890 120
2.4.Two illustrations of the relationship between projectile point forms and the stratigraphy of
Mummy Cave. Column on left is redrawn from Wedel etal. (1968); column on right is redrawn
from Husted and Edgar (2002). In the original, the right column also had radiocarbon dates,
and additional stratigraphic units were labeled (e.g., strata 3A, 8A, 8B, 8C, and others). Note the
stylized projectile points on the left, and the more realistic points (e.g., broken, incomplete) on
the right.
This graph effectively and clearly illustrates change in the frequency of variants
or kinds as a result of differential replication the archetypical definition of
Darwinian evolution. It is a seriogram in the sense of Plog (1973; Figure2.3).
There is one discontinuity or lack of overlap between adjacent strata and it is
between stratum 45 and stratum 3 if the single, perhaps intrusive single spec-
imen of type 1 in stratum 89 is ignored. Otherwise, the general process of
variational evolution is shown by this graph; no evolutionary process is impli-
cated or implied other than (vertical) transmission, though one might suggest
1
2
3
STRATUM
45
67
89
1012
1317
1823
1 2 3 45 6 7 8 910 11 12 13 14 16 17 1820 22 24 26 27
0 50 15 19 21 23 25
TYPE
Relative Abundance
2.5. Percentage stratigraphy graph of 27 projectile point types across 9 stratigraphic units at
Mummy Cave. Some types overlap or occur in multiple contiguous strata, suggesting cul-
tural transmission from one time period to the next. There is no overlap from stratum 45 to
stratum3.
the discontinuity could represent immigration, a standard culture-historical

interpretation of such discontinuities (e.g., Thompson 1956, 1958).
A third graph style plotting data at the scale of types of discrete objects
involves what is known as a clade-diversity graph. Although not frequently
used in archaeology, it is a common graph type in paleontology (references in
Lyman and OBrien 2000). It illustrates the waxing and waning of what are
supposed to be monophyletic groups (or clades), in this case, types of projec-
tile points. It has not been empirically shown that the projectile points from
Mummy Cave in fact represent a clade (OBrien and Lyman 2003), but it
seems reasonable to assume that they share a common ancestor in the form
of the first flaked stone projectile point and thus represent a clade. Figure2.6
is a clade-diversity graph for the Mummy Cave projectile points; the width
of a bar signifies the number or richness of types per stratum. I note that the
stratum-specific richness values do not correlate with sample size per stratum
(Pearsons r = 0.39, p = 0.3). The graph displays the history of the origin
of new types and the extinction of old types in only a very general way; the
percentage stratigraphy graph displays the same thing but with the additional
information of when individual types first appear and when they disappear.
The times of origination and extinction of particular taxa are not apparent
in a clade diversity graph such as Figure 2.6; thus we cannot surmise from
the graph alone when a particular type appears in a sequence or disappears
from it. Figuring out why the richness of kinds should increase or decrease
irrespective of the history of particular types the historical pattern of rich-
ness alone may reveal something of evolutionary process (e.g., Lyman and
OBrien 2000), though theory to explain such is only slowly being developed
in archaeology (see discussion in Lyman etal. 2009). Finally, no evolutionary
38 R. LeeLy m a n
45
STRATUM
67
89
1012
1317
1823
0 5
N of Types
2.6. Clade-diversity graph for the Mummy Cave projectile points. This same information can
be derived from Figure2.5, but here it is much clearer. Assuming an evolutionary connection
between strata, the appearance and extinction of types is implied, but which types appear or go
extinct when is obscure. Compare with Figure2.10.
change in the form of overlap of types across contiguous strata or in the form
of changing frequencies of variants is shown; a cladistic or seriation analysis
that suggests some evolutionary or social transmission connection across the
strata is required to show that this is in fact a clade-diversity graph and not just
a plot of variant richness or diversity.
The clade-diversity graph does provide an indication of the history of for-
mal diversity within a general category of artifacts, in this case, projectile points.
In so far as we may have a hypothesis about causes of shifts in the diversity
of a general category of artifacts, a clade-diversity graph provides a visually
commonsensical means to illustrate those shifts. Not to be confused with a
percentage stratigraphy graph (e.g., Gould etal. 1987), the Mummy Cave pro-
jectile point clade-diversity diagram indicates greatest typological richness in
stratum 89, or about 4600 to 5400 RCYBP. Recent empirical research by
Ken Ames and colleagues, together with a theoretical notion of Todd VanPool
(2001), provides a hint as to why this might be the case. Stratum 89 may well
be when the bow and arrow first appear in the area (Ames et al. 2010). Artisans
would be experimenting, modifying atlatl point types to make an effective
and efficient arrow point (Lyman etal. 2008, 2009). Subsequently, less effective
arrow-point types would no longer be replicated while more effective arrow-
point types would be replicated, resulting in a reduction in typological rich-
ness.This indeed seems to be the case at the scale of type, but it is important to
recognize that the 27 types represented here may not measure equal amounts
1
2
3
STRATUM
45
67
89
1012
1317
1823
200 300 400 500 600 700 800 100 200 300 20 40 60 80
LENGTH WIDTH THICKNESS
1
2
3
STRATUM
45
67
89
1012
1317
1823
50 100 150 200 250 50 100 150 200 250

NECK WIDTH BASE LENGTH
2.7. Measurement values for each of five variables for all individual points regardless of type per
stratigraphic unit at Mummy Cave. All points in strata 1, 2, and 3 are typologically arrow points;
all other points (with one exception) are typologically dart points. Some values do not show
because of overlap in strata 3, 67, and 89. All measurements are in tenths of millimeters.
of formal variation; that is, one type may be quite variable whereas another
type may not include much variability at all. To determine more accurately
how much variation is present and how that variation changes over time, we
must shift to the scale of attributes of points.
At the Scale of Attributes of Discrete Objects

Metric attributes of projectile points can be examined for evidence of evolu-
tion in several ways. One graphic technique used by some archaeologists (e.g.,
Hughes 1998) and also by some paleontologists (e.g., Gingerich 1980) is to plot
values of a variable for each specimen against time. This graph style is shown
using the Mummy Cave projectile point data in Figure2.7 in the form used by
paleontologists. Note that there are no distinctions of projectile point types but
40 R. LeeLy m a n
rather all measurable points regardless of type are plotted by variable length,
width, and so forth. The graph for each variable is quite busy and in some
cases it is difficult to detect trends because of the plethora of data points (some
of which do not show as a result of replication or overlap). In these cases one
of two options is typically followed. Occasionally a best-fit regression line is
superimposed on the point scatters to assist with the identification of an evo-
lutionary trend; more often the graph is simplified to show only the mean, the
mean and one standard deviation, or the mean, one standard deviation, and the
range. I think of these latter graph styles as central tendency graphs (Lyman 2009).
In the context of this chapter it is important to note that evolution in the sense
of shifts in frequencies of variants or types is not implied by the graph style
showing variable values for all specimens, nor in fact is any form of heritabil-
ity, continuity, or transmission from one era to the next shown or implied in
Figure2.7. Again, such an evolutionary connection from one population to the
chronologically subsequent population must come from some other form of
analysis (e.g., Figure2.5).
Central tendency graphs mimic a graph style used by paleontologists, par-
ticularly those interested in whether biological evolution in a particular instance
more closely approximated the phyletic gradualism model or the punctuated
equilibrium model (e.g., Gould and Eldredge 1977).A similar form of graph was
used by Peter and Rosemary Grant (2002, 2006) to illustrate the year-to-year
change in values of variables displayed by populations of Darwins finches on
the Galapagos Islands.The simplest style of central tendency graph shows a plot
of the mean of individual variables against time, as exemplified in Figure2.8.
Again note that strictly speaking, no evolutionary continuity is shown or
implied in the graph, though it might be presumed to exist based on a different
kind of analysis. Note as well that typological distinctions are ignored, but that
when multiple variables are included, patterns of mosaic evolution are revealed
(assuming evolutionary connection between strata has been established) where
mosaic evolution involves the independent evolution of multiple variables. The
graph for Mummy Cave projectile points indicates a tendency toward diminu
tion in all variables over time; size of all variables increases with age (for simple
best-fit linear regression, slopes are significantly > 1.0 and Pearsons r > 0.13
[r > 0.27 for all but length]).There also seems to be a weak correlation between
Width, Thickness and Neck Width as all decrease abruptly from stratum 45 to
stratum 3, until recently (Ames et al. 2010) the time when it was thought the
bow and arrow first appeared. (All possible variable pairs are significantly corre-
lated; r > 0.34, p < 0.0001 for all pairs.)
In paleontology, the graph of means of variables is sometimes modified to
include both the mean and a measure of variation such as the standard devia-
tion (e.g., Sheldon 1987). This graph style is exemplified in Figure2.9. It adds
a bit of information to the means only graph style in Figure2.8, though yet
1
2
3
STRATUM
45
67
89
1012
1317
1823
200 300 400 500 100 200 30 40 50 60 50 100 150 0 100 200
Length Width Thickness Neck Width Base Length
VARIABLE
2.8. Central-tendency graph of the mean for all points regardless of type per stratigraphic unit
at Mummy Cave. This graph depicts mosaic evolution (the co-evolution of more than one
attribute), assuming that an evolutionary connection between stratum-specific assemblages has
been established.
again an evolutionary connection is neither shown nor implied. Instead what

is shown in particular in Figure2.9 is gradual evolution of most variables most
of the time, again with the exception of the abrupt shift in Width, Thickness,
and Neck Width from stratum 45 to stratum 3. Gradual evolution is implied
by the overlap in variable values from one stratum to the next if heritable con-
tinuity has been shown by a different kind of analysis. No evolution is implied
by Figure2.9, but instead only change in values of variables is indicated. In
Figure2.9, for each variable plotted, I note that evolution is much more than
simple temporal differences in the length or width or thickness of collections
of specimens of different ages. If indeed evolution is being shown in Figure2.9
(again, demanding other sorts of analysis), then one might argue for stabilizing,
diversifying, or directional selection for different sections of graphs for partic-
ular variables.
Yet another way to graph changes in values or states of attributes is exem-
plified in Figure2.10, where changes in the coefficients of variation (CVs) of
individual attributes over time are plotted (see Eerkens and Lipo [2005] for an
alternative way to plot CVs). In general, some variation from one time period
to the next is expected simply given the imperfect fidelity of transmission
(Eerkens and Lipo 2005). But yet again, evolution should not be inferred from
this graph as it shows no heritable continuity from one stratum to the next, nor
does it capture the essence of change in variants; rather it merely illustrates shifts
in the amount of variation present in five variables over nine temporal periods.
Why might there be what seem to be major jumps in variation in all variables
during stratum 3 times? Again, this could well be the time when the bow and
arrow first appear and artisans found that they had to experiment that is,
42 R. LeeLy m a n
1
2
3
STRATUM
45
67
89
1012
1317
1823
200 300 400 500 100 200 30 40 50 60 50 100 150 0 100 200
VARIABLE
2.9. Central-tendency graph of the mean (vertical line) and one standard deviation (box) for all
points regardless of type per stratigraphic unit at Mummy Cave. Mosaic evolution is implied if
an evolutionary connection between stratum-specific assemblages has been established.
generate more than normal variation to find an effective point to serve to

tip an arrow (Ames et al. 2010; Lyman etal. 2008, 2009). Or perhaps the fluc-
tuations are reflections of different durations of each stratum (Sheldon 1993)
or the degree of variation reflects population size (Neiman 1995). Clearly,
central-tendency style graphs must be preceded by establishment of evolu-
tionary continuity between the temporally distinct data that are plotted in the
graphs. Once that is done, then the central tendency graphs can be interpreted
in evolutionary that is, descent with modification terms.
Discussion
Two key questions here are: Have the illustrated graph styles provided insights
to the evolution of projectile point form at Mummy Cave, either in terms of
one graph style relative to another, or in terms of what we think we know
about projectile point evolution? And, what can we say about evolutionary
processes from the illustrated and presumed evolutionary patterns shown by
the graphs? Presuming that evolutionary continuity or connection (effected
by cultural transmission) has indeed been demonstrated for the entire series of
points (and it seems to have been demonstrated to some degree in Figure2.5,
but only in that figure), then the first question should be answered in the affir-
mative.The overall tendency is for points on average, regardless of type, to have
become smaller in all measured dimensions over time. Further, the diversity of
point forms seems to be greatest near the middle of the sequence, in stratum
89 (Figure 2.6), but the morphological variation of points was greatest in
strata 1317, 89, and 3 (Figure2.10). These blips in the degree of variation are
likely related to the appearance of a new weapon delivery system, particularly
1
2
3
45
STRATUM
67
89
1012
1317
1823

0 10 50 VARIABLE
Coefficient of Variation
2.10. Coefficient of variation per attribute for all points regardless of type per stratigraphic unit
at Mummy Cave. Assuming an evolutionary connection between strata, a different evolutionary
process of innovation and sorting is suggested by this graph than that suggested in Figure2.5.
the bow and arrow. But perhaps equally importantly, the apparent lack of cor-
respondence between type richness as indicated in the clade-diversity diagram
(Figure2.6) and in the graph of magnitude of attribute variation as reflected
by CVs (Figure2.10) suggests that different types incorporate quite different
amounts of formal variation. As paleontologists have found, the relationship
between morphological evolution and taxonomic evolution is a field ripe for
detailed study (e.g., Foote 1993, 1996). It may be the same for archaeology,
though I suggest that if this avenue of research is pursued, it should first be
shown that the definitive attributes of types capture equivalent amounts of
variation and that differences between types in morphological variation con-
cern nondefinitive attributes of the included types.
It warrants emphasis that only one graph, the percentage stratigraphy graph
(Figure2.5), actually implies evolution in the sense of connection via transmis-
sion between temporally sequent assemblages and thus shifts in frequencies of
specimens of various types or the variational model of evolution (Figure2.1).
And notice also that even that graph does not necessarily imply a particular
mode of evolution. That is, the graph does not clearly suggest that evolu-
tion was anagenetic (linear) or that it was cladogenetic (branching). In some
ways, both processes or modes might be said to be implied by the graph in
Figure 2.5. Anagenetic evolution is implied by the long-term continuity of
some point forms across multiple contiguous strata; cladogenetic evolution
might be inferred if we knew the typological identities of the point types.
For example, type 6 (Elko corner notched) in Figure2.5 is often suggested to
have evolved into type 1 (Rosegate) (e.g., Bettinger and Eerkens 1999). This
could be taken as evidence that evolution was linear if (and this is critical)
44 R. LeeLy m a n
type 6 does not occur coincident with type 1; this in fact seems to be the case
(Figure2.5). Had specimens of type 6 occurred in the same strata as type 1 (say,
both occurred in strata 2 and 3), then that would be evidence of cladogenetic
evolution (presuming there is good evidence that stratigraphic mixing is not
the cause of the stratigraphic association of the two types).
Conclusions
The styles of graph used by paleobiologists and by archaeologists are critically
important both as analytical devices and as illustrative tools that summarize
either data or interpretations thereof (e.g., Lyman etal. 1998). Different graph
styles have sometimes been misinterpreted by professionals in each discipline
such that a particular graph is interpreted to represent a particular process when
in fact the inferred process is not the one intended by the original analyst. In
such cases graph perception (Cleveland and McGill 1984) is a serious prob-
lem of which we must constantly be aware. The graph styles exemplified here
represent some of the ones commonly used by paleobiologists and archaeolo-
gists to present their research. Importantly, as stand-alone graphs, only one of
them (Figure2.5) can be interpreted as implying the Darwinian variational
model of evolution (Figure2.1). If other evidence suggests that the sequence of
points in fact represents an evolutionary sequence in the sense of a phylogeny,
whether linear or branching, then some of those other graphs imply aspects
of some evolutionary processes. Fluctuations in type richness (Figure2.5) sug-
gest one history of evolutionary innovation and sorting whereas fluctuations
in attribute variation (Figure2.10) suggest a different history of evolutionary
innovation and sorting. This is in part a result of a scale shift from discrete
object type to attributes of discrete objects, but it quite likely is also a function
of different amounts of formal variation being incorporated into individual
types.The evolutionary significance of this difference is a unique insight to the
projectile point sequence from Mummy Cave and requires additional scrutiny.
I consider my analysis a success because of this insight.
The preceding is significant regarding graph styles used to summarize evo-
lutionary pattern and process. Graphed patterns and their attendant inferred
processes depend, as has been argued several times before, on the classificatory
units used (e.g., Lyman and OBrien 2002; OBrien and Lyman 2002; see also
Lyman and OBrien 2006). It is, after all, the classificatory units, those units of
measurement, that are graphed, whether types of points, length of points mea-
sured in centimeters or millimeters, or neck width measured in millimeters or
tenths of millimeters. Not just knowing the identity of the graphed units, but
also understanding what those units actually are, would seem to be a critical
step in the production of graphs that are correctly perceived and subject to a
minimum of misinterpretation (or misperception).
But what of the driving force behind this discussion? That is, what of graph-
ing the pattern of evolution and inferring the particular responsible process(es)
from that graph? That is the issue with which I began. Truthfully, I went in
blindly, figuring that I would graph some patterns evidenced by some data and
then infer some evolutionary processes. However, although I certainly did the
first, I quickly came to duh the realization that graphed data, even if tem-
porally sequent, do not necessarily imply evolution, regardless of process. All
I can do is claim a synapse or two misfired somewhere early in my reasoning,
probably because I was intellectually lazy and in a hurry. In hindsight I can
see that I started with a goal that I could not attain by describing the graphs
that I wanted to describe. But I also learned a lesson that I think is valuable. A
graphed temporal sequence of archaeological data does not necessarily imply
evolution, regardless of pattern or process.
So, in my own face-saving opinion, in a way I attained my original goal,
though via a circuitous route. Graphs may or may not reveal patterns, and dif-
ferent graph styles imply different processes, or perhaps no (causal) process at
all (other than the passage of time). Let me end, then, with this: In our (my?)
rush to make sense of complex evolutionary patterns and to detect particular
evolutionary processes as evidenced by the archaeological record, we (I?) must
not fall prey to inaccurate graph perception because of poorly conceptualized,
designed, and described graphs. Keeping the fundamentals of graph design and
the evolutionary process in mind at all times should help us attain the analyti-
cal goals weseek.
Acknowledgments
A request from William Andrefsky and Nathan Goodale resulted in my writ-
ing this chapter. Collaborations with Michael J. OBrien and Todd L.VanPool
helped me sharpen my thinking about the issues discussed here. Much as I
would like to, I cannot blame any of these people for the misfiring of my
synapses.
References
Ames, Kenneth M., Kristen A. Fuld, and Sara the Spread of Bow-and-Arrow Technology in
Davis. 2010. Dart and Arrow Points on the the Prehistoric Great Basin. American Antiquity
Columbia Plateau of Western North America. 64:231242.
American Antiquity 75: 287326. Catley, Kefyn M., and Laura R. Novick. 2008.
Baum, David A., Stacey DeWitt Smith, and Seeing the Wood for the Trees: An Analysis of
Samuel E. Donovan. 2005.The Tree-Thinking Evolutionary Diagrams in Biology Textbooks.
Challenge. Science 310:979980. BioScience 58:976987.
Bettinger, Robert L., and Jelmer Eerkens. 1999. Cleveland, W. S., and R. McGill. 1984. Graphical
Point Typologies, Cultural Transmission, and Perception: Theory, Experimentation, and
46 R. LeeLy m a n
Application to the Development of Graphical Gould, Stephen J., and Niles Eldredge. 1977.
Methods. Journal of the American Statistical Punctuated Equilibria: The Tempo and Mode
Association 79:531554. of Evolution Reconsidered. Paleobiology
Collier, B. A. 2008. Suggestions for Basic Graph 3:115151.
UseWhen ReportingWildlife Research Results. Gould, Stephen J., N. L. Gilinsky, and R. Z.
Journal of Wildlife Management 72:12721278. German. 1987. Asymmetry of Lineages and
Crisp, Michael D., and Lyn G. Cook. 2005. Do the Direction of Evolutionary Time. Science
Early Branching Lineages Signify Ancestral 236:14371441.
Traits? Trends in Ecology Evolution 20:122128. Grant, Peter R., and Rosemary B. Grant. 2002.
Darwin, Charles. 1859. On the Origin of Species Unpredictable Evolution in a 30-Year Study
by Means of Natural Selection or the Preservation of Darwins Finches. Science 296:707711.
of Favored Races in the Struggle for Life. John Grant, Peter R., and Rosemary B. Grant. 2006.
Murray, London. Evolution of Character Displacement in
Eerkens, Jelmer W., and Carl P. Lipo. 2005. Darwins Finches. Science 313:224226.
Cultural Transmission, Copying Errors, and the Hughes, Susan S. 1998. Getting to the Point:
Generation of Variation in Material Culture Evolutionary Change in Prehistoric Weaponry.
and the Archaeological Record. Journal of Journal of Archaeological Method and Theory
Anthropological Archaeology 24:316334. 5:345408.
Eldredge, Niles, and Stephen Jay Gould. 1972. Hunt, Gene. 2008. Gradual or Pulsed Evolution:
Punctuated Equilibria: An Alternative to When Should Punctuational Explanations Be
Phyletic Gradualism. In Models in Paleobiology, Preferred? Paleobiology 34:360377.
edited by T. J. M. Schopf, pp. 82115. Freeman, Husted, Wilfred M., and Robert Edgar. 2002.
Cooper, San Francisco. The Archeology of Mummy Cave, Wyoming: An
Foote, Michael. 1993. Discordance and Introduction to Shoshonean Prehistory. Special
Concordance between Morphological and Report No. 4, National Park Service and
Taxonomic Diversity. Paleobiology 19:185205. Midwest Archeological Center, Lincoln,NE.
Foote, Michael. 1996. Perspective: Evolutionary Lyman, R. Lee. 2009. Graphing Evolutionary
Patterns in the Fossil Record. Evolution Pattern and Process: A History of Techniques
50:111. in Archaeology and Paleobiology. Journal of
Ford, James A. 1952. Measurements of Some Human Evolution 56:192204.
Prehistoric Design Developments in the Southeastern Lyman, R. Lee, and Michael J. OBrien. 2000.
States. Anthropological Papers 44(3). American Measuring and Explaining Change in Artifact
Museum of Natural History, NewYork. Variation with Clade-Diversity Diagrams.
Ford, James A. 1962. A Quantitative Method Journal of Anthropological Archaeology 19:3974.
for Deriving Cultural Chronology. Technical Lyman, R. Lee, and Michael J. OBrien. 2002.
Bulletin No. 1. Pan American Union, Classification. In Darwin and Archaeology: A
Washington,DC. Handbook of Key Concepts, edited by John P.
Fortey, Richard A. 1985. Gradualism and Hart and John E. Terrell, pp. 6988. Bergin &
Punctuated Equilibria as Competing and Garvey, Westport,CT.
Complementary Theories. In Evolutionary Lyman, R. Lee, and Michael J. OBrien. 2006.
Case Histories from the Fossil Record, edited by Seriation and Cladistics: The Difference
J. C. W. Cope and P. W. Skelton, pp. 1728. between Anagenetic and Cladogenetic
Palaeontological Association, Special Papers in Evolution.In Mapping OurAncestors:Phylogenetic
Palaeontology No.33. Approaches in Anthropology and Prehistory, edited
Gingerich, Philip D. 1980. Evolutionary Patterns by Carl P. Lipo, Michael J. OBrien, Stephen J.
in Early Cenozoic Mammals. Annual Review Shennan, and Mark Collard, pp. 6588. Aldine
of Earth and Planetary Sciences 8:407424. Transaction, NewYork.
Gladwin, Harold S. 1936. Methodology in the Lyman, R. Lee, Steve Wolverton, and Michael
Southwest. American Antiquity 1:256259. J. OBrien. 1998. Seriation, Superposition,
and Interdigitation: A History of Americanist Sayles, E. B. 1937. Stone: Implements and Bowls.
Graphic Depictions of Culture Change. In Excavations at Snaketown: I. Material Culture,
American Antiquity 63:239261. by H. S. Gladwin, E.W. Haury, E. B. Sayles, and
Lyman, R. Lee, Todd L. VanPool, and Michael J. N. Gladwin, pp. 101120. Medallion Papers
OBrien. 2008. Variation in North American No. 25. Gila Pueblo, Globe,AZ.
Dart Points and Arrow Points When One Sheldon, Peter R. 1987. Parallel Gradualistic
or Both Are Present. Journal of Archaeological Evolution of Orodovician Trilobites. Nature
Science 35:28052812. 330:561563.
Lyman, R. Lee, Todd L. VanPool, and Michael Sheldon, Peter R. 1993. Making Sense of
J. OBrien. 2009. The Diversity of North Microevolutionary Patterns. In Evolutionary
American Projectile-Point Classes, Before Patterns and Processes, edited by D. R. Lees
and After the Bow and Arrow. Journal of and D. Edwards, pp. 1931. Linnaean Society
Anthropological Archaeology 28:113. Symposium,Vol. 14. Academic Press, London.
Marsh, Owen C. 1879. Polydactyle Horses: Stanley, Steven M. 1975. Macroevolution: Pattern
Recent and Extinct. American Journal of Science and Process. W. H. Freeman, San Francisco.
and Arts 17:499505. Steward, Julian H. 1944. Re: Archaeological Tools
Marsh, Owen C. 1892. Recent Polydactyle and Jobs. American Antiquity 10:99100.
Horses. American Journal of Science 43:339355. Thompson, Robert S. 1956. An Archaeological
Neiman, Fraser D. 1995. Stylistic Variation in Approach to the Study of Cultural Stability. In
Evolutionary Perspective: Implications for Seminars in Archaeology: 1955, edited by Richard
Middle Woodland Ceramic Diversity. American B. Woodbury, pp. 3157. Society for American
Antiquity 60:736. Archaeology Memoirs No. 11.Washington,DC.
Newell, Norman D. 1947. Infraspecific Categories Thompson, Robert S. (editor). 1958. Migrations
in Invertebrate Paleontology. Evolution in New World Culture History. University of
1:163171. Arizona Social Science Bulletin27.
Novick, Laura R., and Kefyn M. Catley. 2007. Valentine, James W., and Cathleen L. May. 1996.
Understanding Phylogenies in Biology: The Hierarchies in Biology and Paleontology.
Influence of a Gestalt Perceptual Principle. Paleobiology 22:2333.
Journal of Experimental Psychology 13:197223. VanPool, Todd L. 2001. Style, Function, and
OBrien, Michael J., and R. Lee Lyman. 2002. Variation: Identifying the Evolutionary
The Epistemological Nature of Archaeological Importance of Traits in the Archaeological
Units. Anthropological Theory 2:3756. Record. In Style and Function: Conceptual Issues
OBrien, Michael J., and R. Lee Lyman. 2003. in Evolutionary Archaeology, edited by Teresa D.
Cladistics and Archaeology. University of Utah Hurt and Gordon F. M. Rakita, pp. 119140.
Press, Salt LakeCity. Bergin and Garvey, Westport,CT.
Plog, Fred T. 1973. Diachronic Anthropology. Wedel, Waldo R., Wilfred M. Husted, and John
In Research and Theory in Current Archeology, H. Moss. 1968. Mummy Cave: Prehistoric
edited by Charles L. Redman, pp. 181198. Record from Rocky Mountains of Wyoming.
John Wiley & Sons, NewYork. Science 160:184186.
Three
Theory in Archaeology:
Morphometric Approaches
to the Study of Fluted Points
MichaelShott
Consider two aquatic metaphors. First, imagine a horizontal cylinder of ice.

Maintained below freezing, it lacks motive power or inherent dynamics. To
characterize the cylinder efficiently requires knowing only its shape, mass, and
volume. Now imagine a stream of water flowing in a slightly inclined channel.
It too is a continuous, unbroken mass. But the stream is in constant motion.
Although it too can be characterized by shape, mass, and volume between
any two points, these properties are not sufficient descriptions of the stream.
Besides them, the stream is characterized by its dynamic flow, explained by
gravity and the gradient of the stream bed. Description and explanation both
are required to characterize the stream.
Lengths of static ice or flowing water may be subdivided, either arbitrarily
or by perception of change. How do you cut pieces from a cylinder of ice? You
cut them, simply. Each piece forms a sufficient unit separate from the rest. How
do you cut pieces from a flowing stream?
Anthropology as Archaeologys Place?

American archaeology is steeped in anthropology.Whatever theory guides most
modern anthropology, its subjects are people and their cultures. Archaeological
data rarely are linked to individuals, and the cultures we define are constructed
by chains of inference from material evidence. Obviously, archaeology has no
direct access to ancient people or their cultures. We do, however, enjoy direct
48
Th eory i n Ar cha e o l o gy 49
access to the material record of ancient people and cultures; artifacts and their
contexts are our units of observation. We track changes in both through time,
and from the resulting patterns trace changes in ancient cultures. Much of
archaeology is the compilation of sequences of change in artifacts, contexts,
or places.
We can lament our indirect access to behavior or celebrate our direct access
to its products.American archaeology has spent a century lamenting rather than
celebrating.This choice condemns us to ancillary status within a discipline that
lacks a theory of long-term change commensurate with archaeologys time
scale and units of observation. As a result, archaeology suffers from critical defi-
ciencies: (1) its approach is overly typological; (2) it lacks methods of resolving
past time commensurate with both its chronological scope and the nature and
rate of variation in its units of observation; and (3) it lacks a coherent theory
to explain long-term change in artifacts and higher-order units. The deficien-
cies are related, not separate. For instance, resolving typological units into their
continuously varying constituents at once offers the prospect of resolving past
time in the continuous terms commensurate with its nature (e.g., Braun 1985),
not as discrete intervals separated by episodes of abrupt change.
Time Views in Archaeology

Archaeology celebrates its long time perspective. In practice, however, it slices
the continuous flow of past time into discrete pieces, each of which it reduces
to average characteristics for description. The result is less a smooth flow of
past time and the change that characterized it than a string of dammed pools.
The effect is to minimize change within discrete intervals (the pools) and
exaggerate it between them (the dams and waterfalls that separate them).
Artifacts are crucial to our methods of time control, because we use them to
construct the sequences that reduce continuous time to series of discrete inter-
vals. Particular types of stone tools, pottery, etc. serve us as markers of intervals
of time. Holding tool types constant, we characterize each interval by its aver-
age or modal characteristics (e.g., of population, economy, labor organization,
land use). Then we contrast the modal characteristics of successive intervals, in
the process treating change that sometimes is continuous as always episodic.
Finally, we explain sequences in terms of the behavior of analytical individuals,
whether people or abstracted cultures.
These scarcely are original observations (e.g., Frankels 1988:41metastable
equilibria; Plog 1973), and they echo somewhat differently in the Darwinian
archaeology that came to prominence in the past 20years (e.g., OBrien and
Lyman 2003). But the critique of pooling and homogenizing has been ignored
more than heeded, and archaeology has not grasped the implication that our
methods of time control and measurement must be commensurate with the
50 M i chael S ho tt
rate and nature of change in material culture. Among other things, we fail
to explain how and why types change within or between intervals. We use
artifacts to order past time so that we can study change, but lack a theory of
change in those things we use to study change.
A Place Where Archaeology ShouldBe

My subject, then, is the urgent need for changing views of time and for unique
archaeological theory, along with a brief sketch of one small part of the latter.
Archaeological theory explains change (and stasis) in material culture.
We may lack the mature archaeological theory that we need, but decades
ago Clarke (1978) provided the framework. At levels from attribute through
artifact and assemblage to technocomplex, Clarke spoke of the time-pattern
regularities that characterize the material record over long periods. By this tur-
gid expression Clarke meant both the trajectories of change through time that
archaeology can describe, not in inferred cultures but in observed artifacts, and
that it can describe if it has suitable theory. Clarke sketched the scope of the
theory that archaeology needs; it remains for us today to implement his vision.
Theory of Change in Material Culture

Steeped as we are in anthropology, the notion of a theory of material culture
may seem either trivial or a glorified radical empiricism that treats artifacts as
sufficient unto themselves. Advocating it certainly is a self-conscious act, part
awkward groping in the absence of precedent or foundation and part strong
conviction that archaeology has lost its way as stepchild of an anthropology
that, increasingly, shows little interest in the problems that engage us. Only the
briefest outlines of a theory of material culture can be sketched here, more as
a call to action than as a mature, fully developed body of thought.
At the highest level, we do not know why reconstructed cultures or tradi-
tions last as long (or short) as they do and whether their trajectories involve
rapid growth and long decline or the reverse. We do not know if a cultures or
traditions longevity is related to its population size or trends, its environmental
setting, its sociopolitical organization.
At an intermediate level, we do not know if there are regularities in the
patterns of replacement of, for instance, stone or pottery by metal in different
times and places. Do similar types of stone tools or pottery disappear at similar
rates, do similar types persist, and do similar types change in design or use as
metal comes into use in different areas?
We define types and chart their time distribution in relatively small areas so
as to control for spatial variation. Of course we use these types to mark past
intervals of time and, in some cases, to measure rate and pattern of inferred
interaction. But we do not ask why some types persist longer than others, or
explain the shape of type frequencies through time (e.g., symmetrical or asym-
metrical, the latter either toward early or late segments of the time range).
Using stone tools as another example, we do not ask if Type 1 is replaced by
Type 2 or changes by degree into Type 2 nor have we developed the criteria
by which to distinguish these possibilities (Shott 2003). We do not ask why
Type 1 lasted as long (or short) as it did or why it lasted for longer or shorter
than did Type 2. We do not ask if change is faster or likelier in, say, haft or
blade elements, in technological attributes or in metric dimensions. We do
not contemplate the possibility that change is faster or slower in proportion
to degree of elaboration or technical difficulty of the production process. We
do not consider that structural constraints of social value, technology, material,
size, or form canalize artifact change in certain directions.We do not construct
theoretical morphospaces of tools and then chart the degree and pattern of
their filling across space and time. In all of these respects and many more, we
lack a theory of change in artifacts.
To some, this theory risks reduction to artifact physics (Watson 1986:445).
But the risk is worth taking, the benefits greater than the remote danger
entailed. As Watson argued, to so much as define an object as an artifact engages
a chain of behavioral inference.Yet artifacts can be units of historical analysis as
well as observation.There is no artifact physics in an artifact-centered archaeo-
logical theory.
New Approaches: History from Artifacts

Replace aquatic metaphors with a set of stone tools distributed continuously in
a long time series. Each specimen differs slightly from the next.Viewing two or
three in succession, their similarity is clear. But compare the first tool to the last,
and they appear very different. How do you cut pieces from this time series?
From a typological perspective, two or three in succession might be placed in the
same type. But specimens separated by 10 places might appear as different types.
From a typological perspective, it is imperative to define types and the
boundaries that separate them, and to assign specimens to types. Then each
type can be described by its average properties. Within the` type, departures
from average are trivial noise or error, while differences between types are
inherently meaningful. A typological approach treats the time series like a static
cylinder of ice, to be cut into integral pieces. But if the time series is like a
flowing stream, how do you cut pieces fromit?
One way does not abandon typology but modifies it (Shott 2003). Types can
be useful constructs, just not ends in themselves. For some purposes, better is to
regard sequences of types as arbitrary subdivisions of continuous chronotypes,
equivalent to paleobiologys chronospecies (Stanley 1998:13).Types then become
flowing streams, not rigid blocks of ice. From the flow of types, perhaps we can
read history better than in the archaeological equivalent of a set of ice blocks.
History fromStone
The challenge of reading history is great enough in materials like ceramics, which
have the advantage of plasticity and therefore a wide range of final products, and
makers complete control over size, form, and decoration. Even the best knapper
only approximates intended form in aplastic stone, cobble dimensions and frac-
ture mechanics constrain knappers options, and resharpening allometry reduces
size and changes shape in complex ways. History is written in stone less than it is
hinted at among the myriad constraints and sources of variation in lithicdata.
Combined with the natural limitations of stone, the slight degree of modifica-
tion that most stone tools undergo makes them poor candidates for detailed his-
torical inference. Retouched flakes are similar the world over, detailed attempts
at culture-historical classification of their kaleidoscopic form notwithstand-
ing. Their form, size, and technological attributes at discard arise from fracture
mechanics, the kind and amount of resharpening they experience, and perhaps
hafting constraints. With rare exceptions, retouched flakes, hafted hidescrapers,
and most other stone-tool types are poor candidates for historical inference.
Yet points may be different. Most are bifacial, so underwent extensive modifi-
cation over both faces. Most were hafted in use, and therefore subject to both pre-
hensile and functional (e.g., penetration, durability, aerodynamics) requirements
(Hughes 1998) that rewarded close attention to the details of production. Point
production was a learned tradition subject to transmission within or between
generations (MacDonald 1998). Empirically, points also have the advantage of
their abundance, certainly in the Americas. No surprise, therefore, that a measure
of historical relatedness is evident to even casual inspection, or that archaeologists
recently applied the transmission models that gained currency in the past two
decades in an effort to detect and explain historical descent in points (OBrien
et al. 2012; OBrien and Lyman 2003). However muted the historical signal may
be amidst complicating sources of variation, it is worth seeking in points.
History from Points

If developing archaeological theory partly begins with points then, from a
New World perspective, it is fitting to begin at the beginning.The Paleoindian
point sequence starts with Clovis and diversifies into Goshen and Folsom on
the Plains, to Gainey and Barnes/Cumberland in the Midwest, perhaps directly
to Dalton in the mid-South, to Suwannee and other unfluted lanceolates
in the deep South, and to Fluted Fishtails in Latin America. Sequences are well
described and, as descriptive labels, types are convenient markers for complex
patterns of variation. But we must attend to the properties of the stream, not
the ice blocks. We do not know if Clovis morphed by continuous degree
into subsequent types (phyletic gradualism), if types branched off fully formed
while Clovis persisted (cladogenesis), or if types are distinct units that replaced
Clovis. In this state of ignorance, archaeologists have begun to test methods to

distinguish among the possibilities.
Culture historys partial restoration in Darwinian archaeology is attended
by much closer attention to type definition, transmission, and the rate and
direction of change in points, while engaging some of the approachs origi-
nal shortcomings. Cladistic methods are especially popular in recent studies.
Biological anthropology passed through a phase of enthusiastic reception to
cladistics before more sober judgments began to prevail (e.g., Curnoe 2003).
Archaeology repeats that process today.
Cladistics
OBrien and Lyman (2003; see also Beck and Jones 2007; OBrien et al. 2012)
conducted the first major phylogenetic study of points, using cladistic methods
to infer patterns of descent among midcontinental North American fluted points
reduced to sets of discrete characters.The studys pioneering nature commends it
to serious but not uncritical regard (see Shott 2008:148150 for detailed discus-
sion). OBrien and Lyman concluded with a descriptive cladogram; they did not
explain the result in theoretical terms or relate the inferred sequence of discrete
character changes to adaptation, function, or morphological canalization.
Buchanan and Collards (2007) cladistic analysis used two-dimensional land-
mark dimensions of fluted points to identify historical dimensions of shape
variation. They controlled allometric resharpening as a confounding process
by regressing dimensions upon the first component of a principal-compo-
nent analysis, identified as size, and taking residuals as size-free shape measures.
Values then were reduced to discrete states by gap-weighting. States of most
characters were treated as linearly ordered (i.e., capable of changing by one
increment per evolutionary step such that if Character X contains States a, b,
and c, and a specimen possesses State a, then in a single evolutionary step it
could change only to b, not c) and freely reversing (i.e., b could change either
to a or c). Buchanan and Collard tested several colonization and other hypoth-
eses (e.g., variation by site type, by geographic proximity, by environment),
concluding that shape variation in points was mostly historical and consistent
with a rapid colonization process originating from the ice-free corridor.
Critique of Cladistics
Cladistic and transmission studies are honest efforts to broaden the scope of
historical inference from points.Yet a demurral, on several grounds. Following
theoretical preconceptions toward paradigmatic classification and the example
of ceramic analysis (e.g., Neiman 1995), cladistics reduces the complex, con-
tinuous variation that resides in points to categorical or ordinal scales of mea-
surement. This is both questionable at face value points cannot be modeled
adequately as sets of discrete traits (e.g., Shott 2008) and unnecessary both
on theoretical and empirical grounds (MacLeod 2002). Cladistics also assumes
strict branching divergence, can overestimate rate of branching when terminal

taxa are not contemporaneous, can conflate proximity in time and historical
relatedness, and can yield statistically ambiguous parsimonious phylogenetic
solutions (Lipo 2006). Finally, cladistic studies do not sufficiently control for
the effects on point form of resharpening allometry.
Buchanan and Collards assumption about ordering of change may be suit-
able in the living world, where saltation is impossible, but questionable in
stone tools because people could radically change their designs from week to
week, letalone from generation to generation. Function or activity as a source
of variation was rejected because dimensions did not pattern with three site
types defined: cache, kill, and habitation. Environment as a source of varia-
tion was rejected because point dimensions did not pattern by zones defined
by gross ecological measures. The phylogenetic method used, common in cla-
distic studies, produced parsimonious cladograms that matched none of the
outcomes predicted by any hypothesis, even the one favored.
Yet even sober judgments can overreach by denying the possibility of per-
ceiving historical signals of branching descent amidst the loud noise of complex
reticulate patterning (e.g.,Tmkin and Eldredge 2007). Criticism of cladistics is
legitimate; dogmatic rejection isnot.
Morphometrics
Another approach is morphometric, involving landmarks and dimensions
defined by distance and angle between them. Like cladistics, morphometrics is
more advanced in biological sciences than in archaeology. It does not reduce
dimensions to discrete characters, assume branching divergence, or require pro-
duction of cladograms.Yet even morphometric approaches deserve evaluation,
not unqualified acceptance. In particular, it may be better suited to coarser
scale or higher level variation.
Using the same data as Buchanans earlier study, Buchanan and Hamilton
(2009) attributed fine-scale variation in Clovis-point dimensions to drift.
They eliminated material as an explanation by comparing mean dimensions
in chert and obsidian points; environment by correlating mean dimensions
per assemblage with net primary productivity; and resharpening allometry by
finding geographic variation in the mean value of a size measure, attributed
to proportion of cache points across regions. Then they measured point shape
as the zero-centered residuals of dimensions regressed upon within-group
principal-component 1 (i.e., a size measure) scores of zero-centered log-trans-
formed mean dimensions per assemblage. Having eliminated material, envi-
ronment, and allometry, they attributed patterning in this variable to copying
error accumulating overtime.
Eliminating materials ignores the very high probability that chert sources
themselves vary in salient mechanical qualities. Eliminating environment
assumes that net primary productivity is the chief environmental variable to
which point form adapts, and that it does so by gross biome, not continuously
across them. Eliminating allometry ignores the possibility that difference in
proportion of (unresharpened) cache points across regions is itself a strong
indication of allometry, and the considerable evidence for resharpening allom-
etry documented generally in fluted points (Buchanan 2006; Shott etal. 2007)
and in one of their larger assemblages, Debert (Ellis 2004).
Given our imperfect control over these manifest sources of variation, it
seems questionable to posit a historical signal in relatively slight, subtle var-
iation expressed in morphologically similar specimens distributed closely in
time. Aggregating specimens by assemblage and characterizing assemblages
by mean values emphasize intrassemblage over interassemblage variation, and
ignore the possibility that the range of variation between assemblages is dupli-
cated within them. Characterizing Naco, for instance, by its mean values and
then comparing it to other assemblages for differences of possible historical
origin ignores the possibility that Naco itself encompasses the full range of
variation found in larger samples.
Similarly, failure to attribute variation to other factors does not mean that
history explains it.That is, we cannot rely on arguments of elimination of an If
not A, then necessarily B nature. In this regard, it seems particularly question-
able to calculate average values by very broad biome or environmental zone; it
is highly unlikely that prehistoric knappers finely adjusted the dimensions of
their stone tools to net above-ground productivity. Hunting technology does
not adapt to gross biome or ecological measures, but to complex conditions
that aggregate measures are unlikely to represent.
Allometry in Cladistic and Morphometric Studies

Resharpening allometry and its effects on size and form of maintained tools like
points has been well plumbed in recent years (e.g., Andrefsky 2008; lovita 2009,
2011; Shott 2005). To control for it, Buchanan and Collard (2007:372) created
what they called size-free variables by regressing each log-transformed dimen-
sion upon principal-component 1 (PC1) score of multivariate morphometric
analysis. This treats PC1 as a metavariable that expresses the multidimensional
quality of size (Buchanan and Collard 2007:3737), and assumes that regression
residuals are independent of size. (As above, all of this occurred before they
reduced these continuous variables to discrete character states.) This approach
acknowledges the allometry problem, but does not necessarily solveit.
Size independence of residuals can be gauged in a data set of conventional
orthogonal dimensions (e.g., length, width) used in a study of the allometry
of stone-tool reduction (Shott et al. 2007). Here, the key point is that the
specimens were Folsom replicas used, damaged, resharpened and reduced, then
used again for several cycles. This experimental control reveals resharpening
allometry (i.e., how and to what degree dimensions and proportions change as
a function of reduction). The original studys main conclusion was that simple
ratios were valid measures of degree of reduction; the ratio of maximum length
to maximum thickness (LT) correlated best with independent measures of
reduction that experimental control provided.
Following Buchanan and Collards procedure, I regressed log-transformed
dimensions (maximum length, width, and thickness, along with mass and a
shape variable, the distance from the base to the point of maximum width) upon
PC1 of principal-component analysis of the covariance matrix. Unstandardized
residuals to the regression are the size-free variables that should be indepen-
dent of degree of reduction experienced by specimens.
In these data LT measures degree of reduction, so size-free variables should
be independent of LT. Some, but not all, are. The residual of lnLength, for
instance, correlates with the natural log of reduction measure LT in the first
three of five use/resharpening stages studied, and nearly attains significance
in later stages. Figure 3.1 shows the cross-plot for the first stage (r = 0.77,
p < 0.01). It is no surprise that even the residual of lnLength might correlate
with lnLT, a ratio that incorporates length. But dimensions in Buchanan and
Collard (2007:Fig. 3) like overall length, mid-line length, tip to base length,
blade length, edge boundary length also are likely to vary with resharpening.
Because the cases differ in salient respects (e.g., Buchanan and Collards dimen-
sions were interlandmark distances on Clovis-affinity points, Shott and col-
leagues orthogonal dimensions of Folsom replicas, and perhaps Buchanan and
Collards residuals were standardized) and because PC1 values would differ if
other variables were analyzed, the size dependence of some size-free variables
in one case do not prove the same for the other. They should, however, give
cause to reconsider whether shape is easily isolated from size in lithic analysis.
Given unresolved questions about how to measure and control for resharp-
ening allometry, we could confine analysis to haft elements. Although these
too can be modified during use lives, blades undergo much more resharpening
(lovita 2011; Thulman 2012). Analysis confined to haft elements would moot
the allometric problem.
Morphometrics at the TypeLevel

Rather than examine the fine-scale variation within gross taxa like Clovis,
morphometric analysis might better describe the gross variation between
Clovis and descendant taxa. Morrow and Morrows (1999) study of Clovis and
Fluted Fishtail points is a classic example that resolved apparent boundaries
between distinct types into morphological continua in orthogonal dimensions.
Smith and Smallwood (2011) plotted geographical variation in fluted-point
outline form analyzed morphometrically, and interpreted resulting patterns
partly in terms of descent from Clovis origins. Whites (2006, 2013) attribute
study of Indiana fluted points was paradigmatic because it defined boundaries
by gaps in the distribution of continuous dimensions or proportions of fluted
points. In the process, White recast the Midwestern fluted-point sequence
0.15000
0.10000
REGRESSION RESIDUAL
0.05000
0.00000
0.05000
0.10000
0.15000
2.20 2.30 2.40 2.50 2.60 2.70 2.80

InLT
3.1. Regression residual of lnLength upon principal component 1, plotted against reduction
measure lnLT in Folsom replicas. (Data from Shott etal. 2007.)
from a series of discrete types to a complex, continuous process of change in

haft and blade dimensions within an historicaltype.
Morphometrics from landmark data identifies progressive deformations of
an original form (equivalent to the out-group of cladistics) in descendants using
multivariate methods and thin-plate spline graphs. As in cladistics, deformation
by adaptation, resharpening allometry, or other factors must be controlled before
variation can be attributed to historical design change. Degree and pattern of
deformation are amenable to mathematical description and model fitting.
Cardillo (e.g., 2006), Castieira etal. (2007), and Thulman (2006, 2012) con-
ducted morphometric analysis of Fluted Fishtail and other points. Burnett and
Otrola-Castillo (2008) modeled variation in Elko Corner-Notched replicas,
admittedly not Paleoindian. Variation among types was depicted by thin-plate
spline techniques and principal components/relative warps cross-plots. Vector
plots mapped the degree and pattern of departures from consensus or average
forms among the replicas. Burnett and Otrola-Castillo (2008:13) examined
allometry by plotting centroid size again Procrustes distance, showing at once
that resharpened points were more variable than pristine ones in form (no sur-
prise) and that, as allometry suggests, size and shape are not always independent.
A complementary morphometric approach involves laser scanning of three-
dimensional (3D) landmark dimensions, which facilitates measurement of point
area, volume, cross-section size and form, and other morphological attributes
difficult to measure manually. Digital imagery also is more efficient and accu-
rate than manual techniques in the measurement of innovative, nonorthogonal
dimensions (e.g., Buchanan and Collard 2007).
I am collecting 3D digital images of fluted and other late Pleistocene lan-

ceolate points by scanner. One focus is the study of resharpening allometry in
three dimensions, using the Folsom replicas noted in the preceding text (Shott
etal. 2007). Another is to follow Buchanans pioneering approach, mindful of
limitations noted previously.
Questions to consider include mode of change from one recognized type to
another. Is it, for instance, a process of shift in discrete characters best modeled
by cladistics, or a morphometric process of differential change in proportions
of tool dimensions? Can changes from type to type be explained in functional
terms (e.g., Shott 1990) or by canalization of form, such that size and form of
later types is constrained by the range of variation that resides in earlierones?
In this way, we can begin to repair the deficiencies in archaeological thought
identified above. We can resolve differences between types into sometimes
subtle continuous variation in metric dimensions. We can morph one type to
the next, in the process transcending our preoccupation with typology, resolv-
ing past time and other sources of variation in complex, continuous terms, and
identifying and modeling various rates and modes of change in points. Inspired
by this different way of viewing variation in continuous time series of objects,
we can begin to develop the coherent theory that can explain the major pat-
terns of long-term change in artifact series. We can, in a sense, replace discrete
blocks of ice with continuously flowing streams.
Conclusion
It is the twenty-first century, and archaeology continues to lack a distinct body
of theory for its distinct subject and time scale. So much is fine for a field that
stands as mere an adjunct to anthropology. But this condition leaves archaeol-
ogy in an unresolved position, a self-conscious field of study that lacks its own
theoretical core. That position continues to be mapped and critiqued from
different perspectives (e.g., Gillespie and Nichols 2003; Killick and Goldberg
2009). On the ground, the great tragedy of archaeologys place is our failure to
develop an authentic body of archaeological theory.
Whatever archaeologys place, we must ask questions about form and pat-
tern in the material record.We must explain change in the material record, not
just describe it or use it to measure and control the passage of time to construct
static, homogenizing scenarios of the past separated from successive scenarios
by brief intervals of sharp change. That is, we must develop a uniquely archae-
ological theory that includes but is not limited to explanations of change in
material culture, its mode, rate, and cause. We must attend to the properties of
artifact sequences that flow through time.Whatever its place, archaeology can-
not be its own discipline until it develops its own body of theory to explain
its unique subjects in their own terms. Morphometric analysis of stone tools is
one small step toward thatgoal.
References
Andrefsky, William. 2008. Lithic Technology: Segundo Congreso Argentino, pp. 17. Comisin
Measures of Production, Use and Curation. Nacional de Energia Atmica, Buenos Aires.
Cambridge University Press, Cambridge. Clarke, David L. 1978. Analytical Archaeology.
Beck, Charlotte, and George T. Jones. 2007. Columbia University Press, NewYork.
Early Paleoarchaic Point Morphology and Curnoe, D. 2003. Problems with the Use of
Chronology. In Paleoindian or Paleoarchaic? Cladistic Analysis in Palaeoanthropology.
Great Basin Human Ecology at the Pleistocene/ Homo 53:225234.
Holocene Transition, edited by K. Graf and D. Ellis, Chris. 2004. Understanding Clovis
Schmitt, pp. 2341. University of Utah Press, Fluted Point Variability in the Northeast: A
Salt LakeCity. Perspective from the Debert Site, Nova Scotia.
Braun, David P. 1985. Absolute Seriation: A Canadian Journal of Archaeology 28:205253.
Time-Series Approach. In For Concordance in Frankel, David. 1988. Characterising Change in
Archaeological Analysis: Bridging Data Structure, Prehistoric Sequences: A View from Australia.
Quantitative Technique, and Theory, edited by C. Archaeology in Oceania 23:4148.
Carr, pp. 509539. Westview, KansasCity. Gillespie, Susan D., and Deborah L. Nichols, eds.
Buchanan, Briggs. 2006. An Analysis of Folsom 2003. Archaeology Is Anthropology.Archeological
Projectile Point Resharpening Using Papers of the American Anthropological
Quantitative Comparisons of Form and Association, No.13.
Allometry. Journal of Archaeological Science Hughes, Susan S. 1998. Getting to the Point:
33:185199. Evolutionary Change in Prehistoric Weaponry.
Buchanan, Briggs, and Mark Collard. 2007. Journal of Archaeological Method and Theory
Investigating the Peopling of North America 5:345408.
through Cladistic Analysis of Early Paleoindian Killick, David, and Paul Goldberg. 2009. A
Projectile Points. Journal of Anthropological Quiet Crisis in American Archaeology. SAA
Archaeology 26:366393. Archaeological Record 9(1):610,40.
Buchanan, Briggs, and Marcus J. Hamilton. 2009. lovita, Radu. 2009. Ontogenetic Scaling and
A Formal Test of the Origins of Variation in Lithic Systematics: Method and Application.
North American Early Paleoindian Projectile Journal of Archaeological Science 36:14471457.
Points. American Antiquity 74:279298. lovita, Radu. 2011. Shape Variation in Aterian
Burnett, Paul, and Erik Otrola-Castillo. 2008. Tanged Tools and the Origins of Projectile
Knapping on an Idea: Experimental and Technology: A Morphometric Perspective
Archaeological Projectile Point Morphological on Stone Tool Function. PLoS ONE 6(12):
Variability Using Geometric Morphometrics. e29029. doi:10:1371/journal.pone.0029029.
Paper presented at the 73rd Annual Meeting Lipo, Caro. 2006. The Resolution of Cultural
of the Society for American Archaeology, Phylogenies Using Graphs. In Mapping Our
Vancouver. Ancestors: Phylogenetic Approaches in Anthropology
Cardillo, Marcelo. 2006. Temporal Trends and Prehistory, edited by C. Lipo, M. OBrien,
in the Morphometric Variation of the M. Collard, and S. Shennan, pp. 89107. Aldine
Lithic Projectile Points during the Middle Transaction, New Brunswick,NJ.
Holocene of Southern Andes (Puna Region): MacDonald, Douglas H. 1998. Subsistence,
A Coevolutionary Approach. Ms. on file, Sex, and Cultural Transmission in Folsom
Depto. De Investigaciones Prehistricas Culture. Journal of Anthropological Archaeology
y Arqueolgicas (DIPA-IMHICIHU), 17:217239.
CONICET, Buenos Aires. MacLeod, Norman. 2002. Phylogenetic Signals
Castineira, C., M. Cardillo, J. Charlin, J. C. in Morphometric Data. In Morphology,
Fernicola, and J. Baeza. 2007. Anlisis Shape and Phylogeny, edited by N. MacLeod
Morfomtrico de los Cabezales Lticos Cola and P. Foray, pp. 100138. Taylor & Francis,
de Pescado del Uruguay. In Arqueometra, London.
Morrow, Juliet E., and Toby A. Morrow. 1999. Transmission and Archaeology: Issues and Case
Geographic Variation in Fluted Projectile Studies, edited by M. OBrien, pp. 146157.
Points: A Hemispheric Perspective. American Society for American Archaeology Press,
Antiquity 64:215230. Washington,DC.
Neiman, Fraser D. 1995. Stylistic Variation in Shott, Michael J., David A. Hunzicker,
Evolutionary Perspective: Inferences from and Bob Patten. 2007. Pattern and
Decorative Diversity and Interassemblage Allometric Measurement of Reduction in
Distance in Illinois Woodland Ceramic Experimental Folsom Bifaces. Lithic Technology
Assemblages. American Antiquity 60:736. 32(2):203217.
OBrien, Michael J., Briggs Buchanan, Mark Smith, Heather and Ashley Smallwood. 2011.
Collard and Matthew T. Boulanger. 2012. A Geometric Morphometric Exploration
Cultural Cladistics and the Early Prehistory of Fluted Point Shape. Paper presented at
of North America. In Evolutionary Biology: the 76th Annual Meeting of the Society for
Mechanisms and Trends, edited by P. Pontarotti, American Archaeology. Sacramento, CA,
pp. 2342. Springer, Berlin. USA, 1 April.
OBrien, Michael J., and R. Lee Lyman. 2003. Stanley, Steven M. 1998. Macro-evolution: Pattern
Cladistics and Archaeology. University of Utah and Process. Johns Hopkins University Press,
Press, Salt LakeCity. Baltimore, MD, USA.
Plog, Fred. 1973. Diachronic Anthropology. In Tmkin, Ilya, and Niles Eldredge. 2007.
Research and Theory in Current Archaeology, Phylogenetics and Material Culture Evolution.
edited by C. Redman, pp. 181198. John Wiley Current Anthropology 48:146153.
& Sons, NewYork. Thulman, David K. 2006. A Reconstruction of
Shott, Michael J. 1990. Stone Tools and Paleoindian Social Organization in North
Economics: Great Lakes Paleoindian Central Florida. Unpublished Ph.D. disser-
Examples. In Early Paleoindian Economies of tation, Department of Anthropology, Florida
Eastern North America, edited by B. L. Isaac State University, Tallahassee.
and K. B. Tankersley, pp. 343. Research in Thulman, David K. 2012. Discriminating
Economic Anthropology, Supplement 5. JAI Paleoindian Point Types from Florida Using
Press, Greenwich,CT. Landmark Geometric Morphometrics. Journal
Shott, Michael J. 2003. Time as Sequence, Type of Archaeological Science 39:15991607.
as Ideal: Whole-Object Measurement of Watson, Patty J. 1986. Archaeological
Biface Size and Form in Midwestern North Interpretation, 1985. In American Archaeology
America. In Multiple Approaches to the Study Past and Future: A Celebration of the Society
of Bifacial Technology, edited by M. Soressi for American Archaeology 19351985, edited
and H. Dibble, pp. 251271. University of by D. Meltzer, D. Fowler, and J. Sabloff, pp.
PennsylvaniaMuseum Press, Philadelphia. 439457. Smithsonian Institution Press,
Shott, Michael J. 2005. The Reduction Thesis Washington,DC.
and Its Discontents: Review of Australian White, Andrew. 2006. A Model of Paleoindian
Approaches. In Lithics Down Under: Hafted Biface Chronology in Northeastern
Australian Perspectives on Lithic Reduction, Use Indiana. Archaeology of Eastern North America
and Classification, edited by C. Clarkson and 34:2959.
L. Lamb, pp. 109125. British Archaeological White, Andrew. 2013. Functional and Stylistic
Reports International Monograph Series Variability in Paleoindian and Early Archaic
1408, Oxford, Archaeopress. Projectile Points from Midcontinental
Shott, Michael J. 2008. Darwinian Evolutionary North America. North American Archaeologist
Theory and Lithic Analysis. In Cultural 34:71108.
Four
Innovation and Natural Selection

in Paleoindian Projectile Points
from the American Southwest
Todd L.VanPool, Michael J. OBrien, and R. LeeLyman
Although the tempo and mode of evolution can vary both in concert and inde-
pendently, descent with modification of cultural variants affected by selection
produces predictable patterns in the archaeological record (Rogers and Ehrlich
2008). These patterns include stability, in which similar variants are transmitted
from generation to generation through time and across space, as well as rapid
change, during which previously dominant variants are eliminated in favor of
alternatives. The adoption of cultural variants is a two-step process of invention
and innovation, the latter encompassing the replication of some or all of the
variants that are invented (OBrien and Shennan 2010). Inventions are common
in cultural frameworks. They can result from transmission errors, novel combi-
nations of previously existing variants, intentional efforts to improve the effi-
ciency of some technology, and a host of other factors (Eerkens and Lipo 2005;
Lipo and Eerkens 2008; Schiffer 2008; Stone 2008;VanPool and VanPool 2003).
Any number of variables can affect the rate of invention, including popula-
tion size, the amount and nature of interaction with people outside ones fam-
ily, and even a cultural systems accommodation of cultural variation (Henrich
2010; Neiman 1995; Palmer 2010). Depending on our scale of analysis, the rate
can appear fairly steady, but there are times when even at a macroscale it is
evident that the rate of invention has suddenly accelerated, creating in its wake
a wide array of new variants what Schiffer (2005) labels the cascade effect.
The impetus for this effect is the desire of individuals to address perceived
deficiencies in a technological system. As inventors and assemblers are stimu-
lated to action hence Schiffers (1996) term stimulated variation change
61
62 VanPoo l e t a l .
in various linked technological systems accelerates as improvements in one

system lead to changes/improvements in other systems. Stimulated variation
likely reflects humans intentionally seeking to address technological inade-
quacies, but it is equally true that peoples efforts to improve technological
systems and increase the general performance/efficiency of the tasks they
perform typically do not result in stimulated variation. The empirical record
indicates that the amount of variation that can be sustained in a cultural sys-
tem is limited to far below that associated with periods of stimulated variation
(e.g., Basalla 1988).Thus, human intentions to improve a technological system
perceived as inadequate are likely a necessary but not wholly sufficient factor
for explaining stimulated variation and subsequent cascade events.
Here we propose that the rapid increase in invention rates associated with
stimulated variation can be tied to evolutionary forces operating in at least two
general contexts. The first, which we mention here only briefly, is when two
previously more or less distinct cultural traditions become intermingled, as when
two groups begin to share the same landscape and interact intensively. Under
such circumstances, recombination of cultural traits characteristic of each group
will create novel cultural variants. Further, metatraditions cultural practices
that help insure the faithful transmission of cultural traits while simultaneously
discouraging high levels of cultural innovation (Palmer 2010) in any or all
of the cultures may break down, such as in the case of migration associated
with a loss of political and economic power and the disruption of family units
(e.g., refugees and large-scale diasporas). Increased intergroup interaction will
be characterized by a pattern of initially limited invention in the groups before
their interaction, greatly increased rates of cultural invention shortly follow-
ing the increased interaction, and then the eventual return to a rate of cultural
invention consistent with the general population level and the degree of cul-
tural conservativism of the newly emerged cultural system (Figure4.1a).
The second context leading to stimulated variation, and the focus of our
discussion here, is associated with rapid changes in a groups selective environ-
ment. This pattern is characterized by limited variation before the selective
shift, substantial variation during and following the shift, and then the return of
decreased variation as selection eliminates many (or possibly all) of the variants
(Figure4.1b). Although we apply this model to early projectile points from the
American Southwest and adjacent regions, we propose that it is applicable to
other cultural traits in similar periods of changing selective contexts.
Selection and Artifact Variation

The suitability of Darwinian evolution as an explanatory framework for the
archaeological record has been established elsewhere (Dunnell 1980; Leonard
2001; OBrien and Lyman 2000; Shennan 2002). Of interest here are aspects of
I nnovation a n d N at u r a l S e l e ct i o n i n Pal e o i n d i an P ro j e c ti l e P o i n ts 63
(a)
Older
Initial Variation Before Increased Interaction
Time
Stimulated Variation Caused by Increased Invention

Rates Resulting from Increased Interaction
Return to Invention Rates Typical of a Given

Cultural Context
Younger
Morphological Variation
(b)
Older
Stable Selective Environment
Time
Stimulated Variation Caused by Change in Selective

Environment
Reduction of Variation Resulting from Stable

Selective Environment
Younger
4.1. Models of stimulated variation resulting from (a) increased interaction among mem-
bers of two or more previously distinct cultural systems and (b) a rapidly shifting selective
environment.
the operation of selection as it impacts cultural traits reflected in artifact mor-

phology. Selection is, of course, only one of a number of evolutionary mecha-
nisms that affect the transmission of cultural traits. Other mechanisms include
the generation of new variation (invention, often resulting from the recombi-
nation of cultural traits into new variants) and various forms of evolutionary
drift such as hitchhiking, the founder effect, and replicative (transmission) error
(Hurt etal. 2001; Lipo and Madsen 2001; OBrien and Leonard 2001; OBrien
and Lyman 2003). However, natural selection is the only creative evolution-
ary force in the sense that it can result in adaptations through the selective
elimination of variation (OBrien and Holland 1992;VanPool 2002).
The operation of natural selection typically takes one of three forms: sta-
bilizing selection, directional selection, and diversifying selection (Endler
1986; Rogers and Ehrlich 2008; VanPool 2001). Stabilizing selection is
Selection Selection Selection Selection
Passage of
Frequency
Frequency
time
Unspecified Cultural Trait Unspecified Cultural Trait

4.2.The influence of stabilizing selection on variation of a culture trait over time.The horizon-
tal axis represents taxonomic and/or morphometric variation. The vertical lines represent limits
of variation upon which selection does or does not operate.
Selection Selection
Passage of
Frequency
Frequency
time

4.3.The influence of directional selection on variation of a culture trait over time. The hori-
zontal axis represents taxonomic and/or morphometric variation. The vertical line represents a
particular variant for reference purposes only.
created when variation in a population is squeezed around an optimal

range or mean value, resulting in either a stable or decreasing range of
variation through time (Figure 4.2). Archaeologically, this will appear as
morphological continuity of (presumably) the best adapted variants; that is,
a particular set of variants will be replicated over time. Directional selec-
tion is characterized by directional change in the optimal mean value as it
consistently gets larger or smaller. Selection will limit the variation in the
variable around the optimal mean at any given time but will create a pat-
tern of considerable intergenerational variation through time (Figure4.3).
Further, the variation at any given point in time will likely be skewed in the
direction of change. Morphological continuity is at best short term. Finally,
disruptive selection will be created by selection against some portion of the
center of a distribution characterizing a population relative to variation on
either side. This will lead to the creation of two or more peaks or modes,
Selection Selection
Passage of
Frequency
Frequency
time

4.4.The influence of disruptive selection on variation of a culture trait over time. The horizon-
tal axis represents taxonomic and/or morphometric variation. The vertical lines represent limits
of variation upon which selection does or does not operate.
each with limited variation (Figure4.4). The population as a whole, though,

will reflect considerable variation.
Given enough time, the peaks created through disruptive selection may
develop into distinct adaptations, and we may choose to subdivide the popula-
tion into two or more populations. Diversifying selection will then take the
form of stabilizing or directional selection within the newly defined popula-
tions until diversifying selection operates again. Morphological continuity is
again short term at most. Although two of the three selective patterns (direc-
tional and diversifying selection) are characterized by substantial variation in
a population through time, selection, by definition, eliminates some portion
of the variation in a population. As a result, at any given time the amount of
variation in a population affected by selection will be less than the potential
variation that would be present in its absence. The patterns in artifact mor-
phology created by the elimination of the variation will be distinctive in all
three cases (Rogers and Ehrlich 2008).
The limits on variation created by the operation of selection will correspond
to the performance constraints necessary for creating the adaptively favored
performance characteristics, that is, the traits being selected for (VanPool
2001). As many ways as there are to make tools that work, there are many
more ways to make them that will result in their not working. The effects of
selection operating on a cultural trait will cause the frequency of tool variants
that do not work those that do not confer some selectively important per-
formance characteristic to decrease, generally to the point of extinction.The
result is that the range of variation in tool forms after the operation of selec-
tion is always less than it could be. However, the amount of variation tolerated
by selection may vary among cultural traits. The variation in some attributes
may be tightly constrained, with even minor differences creating significant
impacts on the presence of requisite performance characteristics. The range of
selectively successful variation in other attributes may be greater if selection is

more tolerant of change before a given performance characteristic is adversely
impacted. Thus, even though selection always eliminates variation, the amount
of variation that continues to be transmitted can vary from trait to trait and/
or from attribute to attribute (an issue of scale) and also vary through time
depending on the strength of selection and the requirements it places on the
artifacts performance characteristics.
Selection is also far more likely to operate on and eventually eliminate
cultural inventions from a population than it is to cause the inventions to be
replicated and become innovations. This is because new variation is unlikely
to increase the fitness associated with an adaptation. After all, the adaptation
was created through selection against alternate states of the trait, necessitating
that most of the possible variation in functional attributes is selected against as
a matter of course. The likelihood of new variation that actually improves the
fitness associated with an adaptation is consequently quite slim. This nega-
tive view of natural selection, based on the premise that phenotypic changes
are more likely to decrease fitness as opposed to increasing it, is also common
when dealing with genetic evolution (e.g., Kimura 1992). It is likely appli-
cable to most cases of cultural evolution, and is certainly expected to hold in
stable selective environments, as selection creates adaptations from the available
variation.
However, we know that the selective environment impacting the replica-
tion of cultural traits can change, sometimes quickly and profoundly, just as the
selective environment affecting genetic traits does. Although not an exhaustive
list, changes can be caused by shifts in the natural environment (e.g., rainfall
patterns), changes in the social environment (e.g., interaction with new social
groups or changes in warfare goals and intensity), and migration of people
into new environmental zones (e.g., agriculturalists moving into neighboring
regions occupied by hunter-gatherers). Our discussion here focuses on the
development and evaluation of a model concerning what happens in cases of
extreme shifts of these sorts.
Shifting Selective Environments and Stimulated Variation

A substantial change in the selective environment is expected, by definition,
to result in selection against some variants that previously enjoyed replicative
success. Variation previously resulting in comparatively high fitness levels will
no longer do so relative to alternatives, which in turn releases the selective
constraints around the previously successful variants. Extant variation within
a population likely will initially be selected against in a relatively uniform
manner, with new variants under no more intense negative selection than the
previously favored variants (Figure 4.5). The result of the uniform negative
Selection Selection Selection Selection Selection
Shift in
Frequency
Selective
Frequency
Environment

Passage of
time
Stimulated Variation
(intentional and unintentional increased variation)
Frequency
Unspecified Cultural Trait

4.5.The influence of a shifting selective environment on variation of a culture trait within a
population.The horizontal axis represents taxonomic and/or morphometric variation.The ver-
tical lines represent boundaries of variation upon which selection does or does not operate.
selection against effectively all variants in a population is that selection will not
limit variation but instead will suppress the general population level. If strong
enough, this uniform negative selection can lead to the elimination (extinc-
tion) of a population. This is, of course, a general pattern; it is possible that a
selective shift might cause selection to more strongly favor some portion of
the variation that was already favored, thereby further decreasing variation or
perhaps producing directional selection. Still, we postulate as a general propo-
sition that significant shifts will correspond with selection against the previ-
ously favored variants.
As mentioned earlier, cultural systems are prone to the introduction of new
variation through novel invention, recombination of previously existing cul-
tural traits, or transmission error (Eerkens and Lipo 2005). As many ways as
there are to make tools (and otherwise behave) the same as others, there are
more ways to do things differently (Eerkens 2000). The release of previous
selective constraints allows newly invented traits to continue at a rate of repli-
cation comparable to the previously favored variants, which in turn causes the
variation in cultural traits to increase rapidly. In the case of lithic technology,
increasing variation could be caused by several factors, including the inten-
tional modification of current designs because of their perceived inadequacies
(e.g., human agency and reflexivity), incomplete cultural transmission, random
Emerging adaptations may or may not be

Stimulated Variation functionally equivalent
(intentional and unintentional increased variation)
Establishment
of Stable
Selection Selection Selection
Selective
Environment
Frequency
Frequency
4.6. Development of adaptive peaks resulting from selection operating on increased variation
associated with stimulated variation. The horizontal axis represents taxonomic and/or morpho-
metric variation. The vertical lines represent limits of variation upon which selection does or
does not operate.
differences resulting from raw materials and skill levels, and chance invention
(OBrien and Shennan 2010).The exact source is incidental with respect to the
selective success of the variant. The point is that variation will expand signifi-
cantly through the process of stimulated variation (Figure4.5).
As variation increases, the adaptive space comprising the plausible vari-
ation (given the social, technological, and environmental context) will be
increasingly well explored (Lake and Venti 2010; Lyman and OBrien 2000;
Lyman etal. 2008, 2009). As this happens, any inventions that result in higher
fitness levels, that is, that correspond with superior performance character-
istics, will be favored by selection, thereby becoming innovations. Through
time, the operation of selection will form adaptations as the variation around
these variants is selected against and thus suppressed (Figure4.6). There may
be many or a few adaptive peaks, depending on the nature of the selective
environment.
Each innovation potentially may continue, leading to several distinct adapta-
tions, which may or may not be functionally equivalent. Further, as increasingly
successful adaptations form, one or a few of the adaptive peaks may result in
substantially higher fitness levels, which in turn will result in selection against
the alternative, lower peaks relative to the more optimal adaptation (Henrich
and Boyd 1998; Mesoudi 2008; Mesoudi and OBrien 2008). Regardless, all
adaptations are created and maintained by the reduction of variation. The
result of the process is the relatively rapid generation of considerable variation
at the onset of the shift in the selective environment followed by its subsequent
winnowing (Figure4.1b). The rate at which variation is eliminated will reflect
the strength of the negative selection and should be reflected taxonomically
in the number of classes or types represented and in the degree of variation
represented by morphometric attributes of the artifacts.
The model presented in the preceding text is general and should be

broadly applicable to periods of drastic shifts in the selective environment.
The archaeological record is replete with such cases, but we evaluate the
models applicability to one of the more far-reaching and widespread shifts,
the PleistoceneHolocene shift in the New World. In particular, we evaluate
whether this pattern is applicable to the shifting selective environment during
the Paleoindian time period as it is reflected in projectile-point morphology.
Paleoindian Points and Stimulated Variation

Human populations migrated from Asia to the New World at the end of the
Pleistocene, perhaps around 13,500 calendar years before present (cal BP). Once
they moved south through the ice-free corridor in what is today west-central
Canada, they populated much if not all of the United States within a few hun-
dred years (Buchanan and Collard 2007; Hamilton and Buchanan 2009). The
earliest occupations are characterized by Clovis points, which were used to
hunt mammoth, bison, and presumably other large animals. Clovis points first
appeared ca. 13,340 cal BP and disappeared ca. 12,850 cal BP (Holliday 2000),
if not even earlier. They were followed in the Southwest and on the Plains by
Folsom points (ca. 12,85011,900 cal BP) (Haynes etal. 1992; Holliday 2000;
Taylor et al. 1996) and, toward the end of the Folsom range, by Plainview
points (Holliday etal. 1999).
The PleistoceneHolocene transition (PHT) was a time of dynamic
environmental flux. Mammalian communities fragmented and reorganized
into biotas that were taxonomically less rich than previously (FAUNMAP
Working Group 1996), at least some large mammal species such as mam-
moth (Mammuthus sp.) that likely had been exploited by Paleoindians became
extinct at this time (Grayson 2006), and climate fluctuated unpredictably on
century and decadal scales (Roy etal. 1996). It is probable that the environ-
mentally dynamic PHT era prompted Paleoindians to alter their adaptations
by changing such things as mobility and settlement patterns, prey choice,
and weaponry, including projectile point variation. We anticipate that the
changing environment would have so greatly altered the selective pressures
affecting projectile-point manufacture and use that it would have stimulated
variation. We expect that variation that was initially constrained during the
Early Paleoindian period would have increased greatly, with more variation
following the significant selective shifts during the Late Paleoindian period.
This stimulated variation should be reflected both in taxonomic diversity and
in metric variation among points. Subsequent morphological variation then
would begin to be reduced as a stable selective environment allowed more
optimal adaptations to form during the Late Paleoindian period and/or the
Early Archaic period (Figure4.7).
Older Early Paleoindian (Late Pliestocene) Selective

Time Environment
Late Paleoindian period associated with stimulated

variation caused by changing climate, extinction of
prey animals, decreased mobility, and other shifts in
the selective environment.
End of Late Paleoindian/Early Archaic period
occupation associated with the reduction of
variation reflecting the establishment of an
increasingly stable Early Holocene climatic and
Younger social selective environment.
4.7. Expectations of the model of initial stimulated variation and subsequent reduction of varia-
tion applied to Paleoindian projectile points.
Here we use two strategies to evaluate our expectations; the first evaluates
morphological patterns at a single deeply stratified site, and the second employs
a wider, regional perspective. Focusing on a single site, in this case Blackwater
Draw, located in eastern New Mexico, is useful because it allows us to trace
changes in morphological variation in points through time in a single location.
Given that there are significant environmental similarities (e.g., access to lithic
raw materials, basic topographic setting) and likely historical continuities dur-
ing the sites occupation (Boldurian and Cotter 1999), stimulated variation, if
present, should be clearly reflected.The regional perspective will then allow us
to evaluate whether the patterns evident at Blackwater Draw are typical of the
greater American Southwest.
Blackwater Draw produced a wide range of artifacts and paleofaunal remains
(Boldurian 2008; Boldurian and Cotter 1999; Hester etal. 1972), most of which
are dated stratigraphically to the Early Paleoindian period (pre-11,500 cal BP),
the Late Paleoindian period (11,5009500 cal BP), and the Early Archaic period
(post-9500 cal BP). Of interest here, the site produced an extensive projectile-
point assemblage from the three periods. The Early Paleoindian occupation was
characterized by 119 Clovis, Folsom, and Plainview points; the Late Paleoindian
Portales Phase occupation by 45 Agate Basin, Angostura, Eden, Meserve,
Milnesand, Plainview, Portales, Scottsbluff, and untyped points that do not fit
in standard named types; and the Early Archaic occupation by 15 stemmed points
(Hester etal. 1972). We note that Plainview points reflect morphological conti-
nuity between the Early and Late Paleoindian periods (Buchanan and Collard
2010), which is consistent with the suggestion of at least some historical continu-
ity in the occupation of the settlement. Of course, our model does not preclude
morphological continuity during the period of stimulated variation but instead
suggests that it will be augmented with an increasing range of variation.
An increase in the number of named types, from three during the Early
Paleoindian period to eight during the Late Paleoindian period, suggests there
is increased morphological diversity as we predicted, especially given that there
are more than twice as many Early Paleoindian points. A larger sample size is
more likely to include rare or unusual point types relative to smaller samples,
so the increased number of types in the much smaller Late Paleoindian point
assemblage strongly indicates that these points do reflect more morphological
diversity, although cultural-historical types used to classify the points do not
reflect consistent amounts of morphological variation. We return to the point
regarding the difficulty of using cultural-historical types to measure taxonomic
diversity presently when discussing our use of paradigmatic classification for
the regional study, but we note here that cultural-historical types are defined
based on morphological variation. Thus we suspect that the number of cul-
tural historical types should correlate with taxonomic variation quantified
using other means. This in turn suggests that the Late Paleoindian points do
reflect more taxonomic diversity than the more numerous Early Paleoindian
points, with the comparatively smaller assemblage of Early Archaic points again
reflecting less variation, as predicted by our model (Figure4.7).
Again using published data from Hester etal. (1972), we evaluated the metric
variation reflected in point length and maximum width. We excluded broken
and unfinished points from our analysis, and note that point length and max-
imum width are not ideal variables for quantifying morphological diversity,
given that they can change during a points use-life.Variation in these attributes
could reflect variation in point use and repair as opposed to initial morphology.
Nevertheless, we again see increased variation as reflected in the cumulative
corrected coefficients of variation (CCVs) during the Portales phase relative to
the Early Paleoindian period (Figure4.8 and Table4.1).Variation then decreases
in the Early Archaic points, again as expected. Differences in variation are most
pronounced in point length but are also present in maximum point width.
Evidence indicates greater taxonomic and metric variation in projectile
points during the Late Paleoindian period relative to the Early Paleoindian, as
we expected. To evaluate whether the patterns seen at Blackwater Draw are
reflected elsewhere in the region, we analyzed casts of 92 Paleoindian points
from the American Southwest and adjacent regions in the Eichenberger cast
collection housed at the University of Missouri Museum of Anthropology
(Table 4.2). The casts are detailed replicas of individual points with known
proveniences and provide an excellent assemblage to study morphometric
diversity in the points from the Early and Late Paleoindian time periods, given
that they were not selected with a bias toward any particular morphological or
metric characteristics (e.g., large, perfectly formed points) and in several cases
(e.g., the assemblage from the Plainview site, with 17 Plainview points) broadly
reflect the parent assemblages.
table4.1. Summary information for point length and maximum width for Blackwater Draw
projectile points (mm)
Maximum length Maximum width

Time period Count Average Std. dev. CCV Count Average Std. dev. CCV
Early Paleoindian 28 36.0 10.3 29 36 19.3 3.2 17
Portales phase 34 49.5 22.5 45 45 19.9 4.0 20
Early Archaic 7 31.2 10.0 33 7 19.2 3.2 17
70
60
Cumulative Corrected CV (%)
50
40 Max. Width
30 Max. Length
20
10
0
Early Archaic Portales phase Early Paleoindian
(Folsom)
Time Period
4.8. Cumulative corrected coefficients of variation for point length and maximum width for
Blackwater Draw projectile points.
Published data limited us to using cultural-historical types to measure tax-

onomic (typological) diversity and only two metric dimensions during our
analysis of the Blackwater Draw points, but the casts allow us to design a more
useful classification and to measure several additional dimensions. We used a
paradigmatic classification that had previously proved useful for classifying
Early Paleoindian points from the American Southeast (OBrien etal. 2001).
Unlike culture-historical types, paradigmatic classes are based on uniformly
applied sets of attributes (characters) with defined character states (Lyman
and OBrien 2002; OBrien and Lyman 2002). Thus the classes reflect an
explicitly defined range of variation that is held constant for all of the clas-
sified artifacts, making it possible to specify both the range of variation that
is observed within an assemblage and that is not. This in turn makes it possi-
ble to directly compare the variation between assemblages. Our paradigmatic
classification is based on the eight characters listed in Table4.3 and illustrated
in Figure4.9. We also recorded the eight metric attributes shown in Table4.4
and Figure4.9.
An increase in morphological variation in the Late Paleoindian assem-
blage is reflected in the number of cultural-historical types (two types during
table4.2. Cultural-historical types and provenience locations for Paleoindian points in the
Eichenberger cast collection
Period Cultural-historical Site Count

type
Early Paleoindian Clovis Blackwater Draw, New Mexico 4
Dent, Colorado 2
Escapule, Arizona 2
Lehner, Arizona 13
Murray Springs, Arizona 15
Naco, Arizona 9
Schalduck-Bates, Arizona 1
SW of Clovis, NM 2
Folsom Blackwater Draw, New Mexico 1
Folsom, New Mexico 3
Near Clovis, NM 1
SW of Clovis, NM 3
Lindenmeier, Colorado 1
Late Paleoindian Agate Basin Agate Basin,Wyoming 7
Beck Forest Lake SW of Clovis, NM 1
Frazier, Colorado 3
Eden Beck Forest Lake SW of Clovis, NM 1
Claypool, Colorado 1
Jurgens, Colorado 1
Meserve Panhandle of northwestern Texas 1
Plainview Blackwater Draw, New Mexico 1
Plainview,Texas 17
Scottsbluff Scottsbluff, Colorado 1
Jurgens, Colorado 1
the Early Paleoindian occupation compared to five types during the Late
Paleoindian period), but it is also reflected in the frequencies of paradigmatic
classes when controlled for differences in sample sizes. Thirty paradigmatic
classes were represented among the 39 Early Paleoindian period points that
were complete enough to allow their classification, whereas 23 classes were
represented among the 25 Late Paleoindian points that could be classified.
Although the 30 point classes in the larger Early Paleoindian point assemblage
are numerically greater than the 23 classes of the smaller Late Paleoindian
assemblage, the latter actually have greater diversity given their sample size.
This is illustrated in ratios of the number of classes to sample size. The ratio
for the Early Paleoindian period is 0.77 compared with the Late Paleoindian
periods ratio of 0.92, a 20percent increase in variation when sample size is
taken into account.
A similar increase in variation is evident in the metric attributes of the points
as reflected in Table4.4 and the cumulative CCVs reported in Figure4.10. Six
of the eight metric attributes reflect greater variation in the Late Paleoindian
table4.3. Characters and character states used in the paradigmatic classification
Character/character state Character/character state

Character I: Location of maximum Character V: Outer tangangle
bladewidth 1. 93115 degrees
1. Proximal quarter 2. 8892 degrees
2. Second most proximal quarter 3. 8187 degrees
3. Second most distal quarter 4. 6680 degrees
4. Distal quarter 5. 5165 degrees
6. Less than 50 degrees
7. Tangs absent
Character II: Baseshape Character VI:Tang-tipshape
1. Arc-shaped 1. Pointed
2. Normalcurve 2.Round
3.Triangular 3.Blunt
4. Folsomoid 4.Tangs absent
5. Straight/pointed
Character III: Basal indentation Character VII: Fluting
1. No basal indentation 1. Present
2. 0.900.99 (shallow) 2. Absent
3. 0.800.89 (deep)
Character IV: Constrictionratio Character VIII: Max. length/width ratio
1.1.00 1. 1.001.99
2. 0.900.99 2. 2.002.99
3. 0.800.89 3. 3.003.99
4. 0.700.79 4. 4.004.99
5. 0.600.69 5. 5.005.99
6. 0.500.59 6. Greater than 6.00
period, with attributes related to the hafting elements such as basal indenta-
tion width and depth, which are least likely to change during use and repair
relative to other attributes showing the greatest increase in variation. Again,
these results are consistent with those predicted by our model. Both taxo-
nomic and metric variations are greater in Late Paleoindian points than in
Early Paleoindian points.
Conclusions
The preceding arguments are in large part a continuation of those presented
in Lyman et al. (2008, 2009). The application of evolutionary theory to the
archaeological record requires an understanding of processes that influence
the generation of variation (invention processes) as well as those that influence
the transmission of variation (the processes leading to innovation). The gen-
eration of variation is a consistent feature of all cultural systems, but the rate
of invention can vary widely. During extreme cases, cascade events created by
stimulated variation can result in comparatively brief periods of high invention
table4.4. Summary information for the metric attributes of Paleoindian points in the Eichenberger cast collection (mm)
Time Statistic Max. length Max. Max. Basal indentation Basal indentation Width at Width Medial
period width thickness depth width midpoint at base length
Early Average 73.37 28.17 7.49 3.91 13.69 26.60 24.25 69.87
St. dev. 22.50 4.671 1.86 1.41 4.96 4.76 3.85 23.11
Count 42 52 53 47 46 42 48 39
CV 30.85 16.66 24.97 36.36 36.39 17.98 15.97 33.28
Late Average 76.19 23.38 6.68 1.20 8.99 22.35 19.29 75.19
St. dev. 32.32 1.97 1.03 1.49 10.21 2.38 4.42 33.60
Count 26 35 35 30 30 27 30 25
CV 42.84 8.489 15.49 124.73 114.53 10.74 23.08 45.14
75
E D
E D
C
A A
C
A A
B B
B B
F
F
D D
E C
G
E G C
H
Landmark Characteristics Base shapes

A-A = maximum blade width Arc-shaped
B- B = minimum blade width
C-C = height of maximum blade width Normal curve
D- D = medial length
Triangular
E- E = maximum length
F = outer Tang Angle
Folsomoid
G = tang tip
H = flute Straight
4.9. Illustration of the measurements recorded for the points in the Eichenberger cast
collection.
rates that far exceed the typical rates, generally determined by population size,
transmission mechanisms, and various other factors. Human intentionality cer-
tainly is a contributing factor to the generation of this extraordinary amount
of variation, but it is not in and of itself a sufficient explanation, given that
humans often seek to improve the efficiency of technological systems without
creating such high levels of invention.
We propose that the explanation of stimulated variation is contingent on
evolutionary factors that underlie the generation of variation. We identify two
causes: cultural transmission among previously (more-or-less) distinct cultural
traditions and a rapid shift in the selective environment. Our discussion focuses
450
400 Medial length
Cumulative Corrected CV (%)
350 Width at Base
300 Width at Midpoint
250 Basal Indentation Width
200 Basal Indentation Depth
150 Max. Thickness
100 Max. Width
50 Max. Length
0
Early Late
Time Period
4.10. Cumulative corrected coefficients of variation for the eight metric attributes recorded for
Paleoindian points represented in the Eichenberger cast collection.
on the second of these, which we propose accounts for increased morpholog-

ical variation in Late Paleoindian projectile points. According to our model,
the shifting natural and social environments during the PleistoceneHolocene
transition are expected to have initiated stimulated variation as selection began
to operate against the previously existing range of variation. Simply put, vari-
ants that had worked well earlier were no longer as effective as they previously
were. We expect there was a period of increased invention, likely in part a
result of intentional experimentation, during and shortly after the end of the
Pleistocene as people sought to produce projectile points that worked well in
the changed selective environment. As the morphometric range of possible var-
iation was more completely explored, variants that did result in higher fitness
levels became innovations that then formed the basis of additional evolutionary
development of Late Paleoindian/Early Archaicperiod projectile points.
Here we evaluated whether the predicted Late Paleoindianperiod increase
in variation was present using projectile points from Blackwater Draw, New
Mexico, and from throughout the Southwest and adjacent regions. In both cases,
we found increased taxonomic and metric variation among Late Paleoindian
points, as expected. Although our discussion focuses on projectile points, we
believe our model can be applied generally to other instances of substantial
changes in selective environments.
References
Basalla, George. 1988. The Evolution of Technology. Boldurian, Anthony T., and John L. Cotter. 1999.
Cambridge University Press, Cambridge. Clovis Revisited: New Perspectives on Paleoindian
Boldurian, Anthony T. 2008. Clovis Type-Site, Adaptations from Blackwater Draw, New Mexico.
Blackwater Draw, New Mexico: A History, University of Pennsylvania Museum of
19292009. North American Archaeologist Archaeology and Anthropology, Monograph
29:6589. 103. Philadelphia.
Buchanan, Briggs, and Mark Collard. 2007. Innovation in Cultural Systems: Contributions from
Investigating the Peopling of North America Evolutionary Anthropology, edited by Michael J.
through Cladistic Analyses of Early Paleoindian OBrien and Stephen J. Shennan, pp. 99120.
Projectile Points. Journal of Anthropological MIT Press, Cambridge,MA.
Archaeology 26:366393. Henrich, Joe, and Robert Boyd. 1998. The
Buchanan, Briggs, and Mark Collard. 2010. A Evolution of Conformist Transmission and the
Geometric Morphometrics-Based Assessment Emergence of Between-Group Differences.
of the Utility of Blade Shape for Classifying Evolution and Human Behavior 19:215241.
Paleoindian Projectile Points. Journal of Hester, James J., Ernest L. Lundelius, and Roald
Archaeological Science 37:350359. Fryxell. 1972. Blackwater Locality No. 1: A
Dunnell, Robert D. 1980. Evolutionary Theory Stratified Early Man Site in Eastern New Mexico.
and Archaeology. Advances in Archaeological Fort Burgwin Research Center, Southern
Method and Theory 3:3599. Methodist University, Rancho de Taos,NM.
Eerkens, Jelmer W. 2000. Practice Makes within Holliday, Vance T. 2000. The Evolution of
5% of Perfect: The Role of Visual Perception, Paleoindian Geochronology and Typology on
Motor Skills, and Human Memory in Artifact the Great Plains. Geoarchaeology 15:227290.
Variation and Standardization. Current Holliday, Vance T., Eileen Johnson, and Thomas
Anthropology 41:663668. W. Stafford, Jr. 1999. AMS Radiocarbon
Eerkens, Jelmer W., and Carl P. Lipo. 2005. Dating of the Type Plainview and Firstview
Cultural Transmission, Copying Errors, and the (Paleoindian) Assemblages: The Agony and
Generation of Variation in Material Culture the Ecstasy. American Antiquity 64:444454.
and the Archaeological Record. Journal of Hurt, Teresa D., Todd L. VanPool, Gordon F.
Anthropological Archaeology 24:316334. M. Rakita, and Robert D. Leonard. 2001.
Endler, James A. 1986. Natural Selection in the Wild. Explaining the Co-Occurrence of Traits
Princeton University Press, Princeton,NJ. in the Archaeological Record: A Further
FAUNMAP Working Group. 1996. Spatial Consideration of Replicative Success. In Style
Response of Mammals to Late Quaternary and Function: Conceptual Issues in Evolutionary
Environmental Fluctuations. Science Archaeology, edited by Teresa D. Hurt and
272:16011606. Gordon F. M. Rakita, pp. 5167. Bergin and
Grayson, Donald K. 2006. Late Pleistocene Garvey, Westport,CT.
Faunal Extinctions. In Handbook of North Kimura, Motoo. 1992. Neutralism. In Keywords in
American Indians Volume 3, Environment, Origins, Evolutionary Biology, edited by Evelyn F. Keller
and Population, edited by Douglas H. Ubelaker, and Elisabeth A. Lloyd, pp. 225230. Harvard
pp. 208218. Smithsonian Institution Press, University Press, Cambridge,MA.
Washington,DC. Lake, Mark W., and Jay Venti. 2009. Quantitative
Hamilton, Marcus J., and Briggs Buchanan. Analysis of Macroevolutionary Patterning in
2009. The Accumulation of Stochastic Technological Evolution: Bicycle Design from
Copying Errors Causes Drift in Culturally 1800 to 2000. In Pattern and Process in Cultural
Transmitted Technologies: Quantifying Evolution, edited by Stephen J. Shennan, pp. 146
Clovis Evolutionary Dynamics. Journal of 161. University of California Press, Berkeley.
Anthropological Archaeology 28:5569. Leonard, Robert D. 2001. Evolutionary
Haynes, C. Vance, Roelf P. Beukens, A. J. T. Jull, Archaeology. In Archaeological Theory Today,
and Owen K. Davis. 1992. New Radiocarbon edited by Ian Hodder, pp. 156171. Unwin
Dates for Some Old Folsom Sites: Accelerator Hyman, London.
Technology. In Ice Age Hunters of the Rockies, Lipo, Carl P., and Jelmer W. Eerkens. 2008.
edited by Dennis Stanford and Jane S. Day, pp. Culture History, Cultural Transmission, and
83100. Denver Museum of Natural History. Explanation of Variation in the Southeastern
Henrich, Joseph. 2010. The Evolution of United States. In Cultural Transmission and
Innovation-Enhancing Institutions. In Archaeology: Issues and Case Studies, edited by
Michael J. OBrien, pp. 120131. Society for OBrien,Michael J.,andThomas D.Holland.1992.

American Archaeology, Washington,DC. The Role of Adaptation in Archaeological
Lipo, Carl P., and Mark E. Madsen. 2001. Explanation. American Antiquity 57:359.
Neutrality, Style, and Drift: Building OBrien, Michael J., and Robert D. Leonard.
Methods for Studying Cultural Transmission 2001. Style and Function: An Introduction.
in the Archaeological Record. In Style and In Style and Function: Conceptual Issues in
Function in Archaeology, edited by Teresa D. Evolutionary Archaeology, edited by Teresa
Hurt and Gordon F. M. Rakita, pp. 91118. D. Hurt and Gordon F.M. Rakita, pp. 123.
Bergin and Garvey, Westport,CT. Bergin and Garvey, Westport,CT.
Lyman, R. Lee, and Michael J. OBrien. 2000. OBrien, Michael J., and R. Lee Lyman. 2000.
Measuring and Explaining Change in Artifact Applying Evolutionary Archaeology: A Systematic
Variation with Clade-Diversity Diagrams. Approach. Kluwer Academic/Plenum,
Journal of Anthropological Archaeology 19:3974. NewYork.
Lyman, R. Lee, and Michael J. OBrien. 2002. OBrien, Michael J., and R. Lee Lyman. 2002.
Classification. In Darwin and Archaeology: A The Epistemological Nature of Archaeological
Handbook of Key Concepts, edited by John P. Units. Anthropological Theory 2:3756.
Hart and John E. Terrell, pp. 6988. Bergin OBrien, Michael J., and R. Lee Lyman. 2003.
and Garvey, Westport,CT. Style, Function,Transmission:An Introduction.
Lyman, R. Lee, Todd L. VanPool, and Michael J. In Style, Function, Transmission: Evolutionary
OBrien. 2008. Variation in North American Archaeological Perspectives, edited by MichaelJ.
Dart Points and Arrow Points when One OBrien and R. Lee Lyman, pp. 132.
or Both Are Present. Journal of Archaeological University of Utah Press, Salt LakeCity.
Science 35:28052812. OBrien, Michael J., and Stephen J. Shennan. 2010.
Lyman, R. Lee, Todd L. VanPool, and Michael Issues in Anthropological Studies of Innovation.
J. OBrien. 2009. The Diversity of North In Innovation in Cultural Systems: Contributions
American Projectile-Point Types, before from Evolutionary Anthropology, edited by
and after the Bow and Arrow. Journal of Michael J. OBrien and Stephen J. Shennan,
Anthropological Archaeology 28:113. pp. 317. MIT Press, Cambridge,MA.
Mesoudi, Alex. 2008. An Experimental Palmer, Craig T. 2010. Cultural Traditions and the
Simulation of the Copy-Successful- Evolutionary Advantages of Noninnovation.
Individuals Cultural Learning Strategy: In Innovation in Cultural Systems: Contributions
Adaptive Landscapes, ProducerScrounger from Evolutionary Anthropology, edited by
Dynamics, and Informational Access Costs. Michael J. OBrien and Stephen J. Shennan,
Evolution and Human Behavior 29:350363. pp. 161174. MIT Press, Cambridge,MA.
Mesoudi, Alex, and Michael J. OBrien. 2008. Rogers, Deborah S., and Paul R. Ehrlich. 2008.
The Cultural Transmission of Great Basin Natural Selection and Cultural Rates of
Projectile-PointTechnology I:An Experimental Change. Proceedings of the National Academy of
Simulation. American Antiquity 73:328. Sciences of the USA 105:34163420.
Neiman, F. D. 1995. Stylistic Variation in Roy, Kaustuv, James W.Valentine, David Jablonski,
Evolutionary Perspective: Inferences from and Susan M. Kidwell. 1996. Scales of Climatic
Decorative Diversity and Interassemblage Variability and Time Averaging in Pleistocene
Distance in Illinois Woodland Ceramic Biotas: Implications for Ecology and Evolution.
Assemblages. American Antiquity 60:736. Trends in Ecology and Evolution 11:458463.
OBrien, Michael J., John Darwent, and R. Schiffer, Michael B. 1996. Some Relationships
Lee Lyman. 2001. Cladistics Is Useful for between Behavioral and Evolutionary
Reconstructing Archaeological Phylogenies: Archaeologies. American Antiquity 61:643662.
Palaeoindian Points from the Southeastern Schiffer, Michael B. 2005. The Devil Is in the
United States. Journal of Archaeological Science Details: The Cascade Model of Invention
28:11151136. Processes. American Antiquity 70:485502.
Schiffer, Michael B. 2008.Transmission Processes: VanPool, Todd L. 2001. Style, Function, and
A Behavioral Perspective. In Cultural Variation: Identifying the Evolutionary
Transmission and Archaeology: Issues and Case Importance of Traits in the Archaeological
Studies, edited by Michael J. OBrien, pp. Record. In Style and Function: Conceptual Issues
102111. Society for American Archaeology, in Evolutionary Archaeology, edited by Teresa D.
Washington,DC. Hurt and Gordon F. M. Rakita, pp. 119140.
Shennan, Stephen J. 2002. Genes, Memes, and Bergin and Garvey, Westport,CT.
Human History. Thames and Hudson, London. VanPool, Todd L. 2002. Adaptation. In Handbook
Stone, Tammy. 2008. Social Innovation and of Concepts in Modern Evolutionary Archaeology,
Transformation during the Process of edited by John P. Hart and John E. Terrell, pp.
Aggregation. In Cultural Transmission and 1528. Bergin and Garvey, Westport,CT.
Archaeology: Issues and Case Studies, edited by VanPool,Todd L., and Christine S.VanPool. 2003.
Michael J. OBrien, pp. 158163. Society for Agency and Evolution: The Role of Intended
American Archaeology, Washington,DC. and Unintended Consequences of Action.
Taylor, Richard E., C.Vance Haynes, Jr., and Minze In Essential Tensions in Archaeological Method
Stuiver. 1996. Clovis and Folsom Age Estimates: and Theory, edited by Todd L. VanPool and
Stratigraphic Context and Radiocarbon Christine S.VanPool, pp. 89113. University of
Calibration. Antiquity 70:515525. Utah Press, Salt LakeCity.
PartIII
Applications of Behavioral Ecology

to Lithic Studies
Five
A Case of Extinction in Paleoindian

Archaeology
Charlotte Beck and George T.Jones
Blade technologies have traditionally been viewed as inherently superior to

flake technologies because, among other reasons, they are believed to provide
a more efficient use of toolstone. These technologies have long been identi-
fied with the Upper Paleolithic of the Old World and assumed to have con-
ferred a profound advantage on the modern humans who used them. More
recent work, however, has shown that blades were in use well before the Upper
Paleolithic, not only in Europe but also in Africa and Western Asia. Blades are
associated with sites in South Africa, for example, such as Border and Blombos
caves, dating to ca. 90,000 BP. In the Levant these artifacts have been found in
the Amudian layers of Tabun Cave, which may date to over 300,000 BP (Bar-
Yosef and Kuhn 1999:325).
These early appearances are sporadic, however, and consistent use of this
technology truly does not occur until the Upper Paleolithic; but its use is
not ubiquitous, as it is confined primarily to North Africa, Asia, and Europe
(Bar-Yosef and Kuhn 1999:330; Bettinger etal. Chapter6, this volume). The
waxing and waning of blade use over time and its presence in some
regions but not others suggests that, even though blades may represent a
superior technology in some sense, the advantages are only situationally rel-
evant (Bar-Yosef and Kuhn 1999:330). In this chapter we examine one such
situation: the use of prismatic blades during Clovis times and their ultimate
disappearance.
83
84 Charlo t t e B e ck a n d G e o r ge T. Jo n e s
Blade Technology
In a general sense, a blade is defined as any detached piece with parallel
or subparallel edges that is twice as long as it is wide (Andrefsky 1998:xxii;
Crabtree 1972:42), but of course such a product can result unintentionally
from the reduction of a flake core; the latter is referred to as a blade-like
flake. Thus, Tankersley (2004), among others, states that evidence of a true
blade technology can be demonstrated only by the presence of blade cores.
Bar-Yosef and Kuhn (1999) point out that there are a number of techniques
by which blades can be produced, but it is the prismatic technique that is of
interesthere.
In classic prismatic blade production, one or more long ridges are pre-
pared on the face of the core by bifacial flaking, creating the character-
istic crested blade (lame crte). A series of blades is then detached along
part or all of the cores perimeter. (Bar-Yosef and Kuhn 1999:323)
A number of potential advantages have been attributed to prismatic blade

versus flake technology. First, prismatic blade technology results in an efficient
use of toolstone, as the end products maximize the amount of cutting edge
obtained from a given mass (Bar-Yosef and Kuhn 1999:324; Collins 1999:10;
Hayden etal. 1996:19). Second, the shaping and use of prismatic cores allows
more control over size and shape of blanks than does reduction of a flake core,
producing more uniform end products. Such uniformity would be important
when manufacturing replaceable components of composite tools. But also, as
Collins (1999:9) notes, blades are structurally quite strong, which, in combina-
tion with their uniformity in size and shape, allows versatile applications. Not
only can these products be used as they come off the core, either in modified
or unmodified form, but they can also be segmented for further modifica-
tion. A final advantage of blades is that they facilitate multiple resharpenings
for tools made on distal or proximal ends such as burins, endscrapers, borers,
piercers, drills and points (Hayden etal. 1996:37; see also Bar-Yosef and Kuhn
1999:324).
In general, blade technologies provide an extremely reliable toolkit: as
highly standardized products, blades and blade tools are relatively interchange-
able and replaceable, and their use helps insure a consistent and predictable
outcome, reducing risk of failure (Parry 1994:94). But blade technologies
have limitations as well. For example, Hayden etal. (1996:18) point out that
even though they can be seen as efficient and conservative of raw material,
blade technologies can also be wasteful. Systematic production of prismatic
blades requires considerable preparation of the core, resulting in the removal of
a good bit of unusable material (see, e.g., description by Collins 1999:1925).
Further, because considerable skill is necessary for the systematic removal of
A Case of Ex t i n ct i o n i n Pa l e o i n d i a n Arc h a e o l o g y 85
blades from the cores, the risk of ruining the core (and thus wasting material) is
high (Bar-Yosef and Kuhn 1999:324; Hayden etal. 1996:18). Finally, the man-
ufacture of blade cores requires more specific sizes, shapes, and quality of tool-
stone than necessary for flake cores, which may be difficult to find (Hayden
etal. 1996:18). Because of these limitations, Bar-Yosef and Kuhn (1999) state
that to argue that use of this technology provided an advantage, it must be
demonstrated how its properties would have been locally beneficial. With this
in mind, we turn to a discussion of Clovis technology. But before we do so, we
deviate with a short discussion of how we came to investigate this issue and lay
out some of our assumptions.
Context for Discussion

Over the past 20+ years we have been working on the Paleoindian record of
the Intermountain West, and more specifically, the Great Basin. During that
time we have come to the conclusion that fluted point technology did not
have the kind of impact in this region as it seems to have had elsewhere. The
early record here is overwhelmingly dominated by a technology focused on
long, contracting stemmed points, referred to as Western Stemmed. Over the
last 10years we have made the argument that fluted points date later in this
region than they do on the Plains, and possibly the Southeast; it is Western
Stemmed technology that we believe to represent the earliest human record in
the Intermountain West (Beck and Jones 2010, 2013).
We have argued, along with many others, that what we know as Clovis
technology, especially the Clovis fluted point, is earliest in the Southeast and/
or southern Plains and that populations carrying this technology moved out-
ward from these regions (Beck and Jones 2010, 2012, 2013) (Figure5.1). One
obvious route, based on the distribution of Clovis sites, lies along the eastern
Rocky Mountain front, from Texas to Montana. These mountains posed a
significant barrier to human migration, which is what we believe accounts
for this spatial pattern as well as the low frequency of fluted points in the
Intermountain West and Pacific coastal regions. Clovis populations did, how-
ever, eventually penetrate this barrier, as evidenced by the Fenn, Simon, and
East Wenatchee Clovis caches. But when they arrived on the Columbia
Plateau, they encountered a resident population who utilized a different
technology.
The pattern of northward Clovis migration is supported by the radiocarbon
record from the Plains; overall, dates get progressively younger from south to
north. In tracing this movement we also noted a change in the Clovis toolkit,
from one that contained a prismatic blade component in the south to one
without this component in the north. It was the investigation of this pattern
that led us to the discussion we presenthere.
86 Ch arlo t t e B e ck a n d G e o r g e T. Jo n e s
N
0 200 400 600
km
5.1. Model of proposed movements of Western Stemmed (from west to east) and Clovis (south
to west and north) populations.
Clovis Technology
In all areas where remains of the Clovis toolkit have been found it is clear
that the biface is the central component of this technology. Although the
ultimate aim was the production of the Clovis fluted point (Wilke et al.
1991), Huckell (2007:194) argues that the most important feature of the
Clovis biface industry is that the biface was not viewed simply as a blank
or preform, but as a fully functional implement. The reduction sequence
for the Clovis biface has been studied by a number of lithic analysts and all
agree that it is highly distinctive and recognizable in the archaeological rec-
ord (e.g., Bradley et al. 2010; Huckell 2007; Wilke etal. 1991). One of the
most distinctive aspects of this reduction system is the regular and controlled
use of overshot flaking, in which a thinning flake is struck from one margin,
travels across the face of the biface, and frequently removes a portion of the
opposite edge. In the early stages of thinning these overshot flakes are large
and often were used to make other unifacial and bifacial tools (Bradley et al.
2010; Huckell 2007). Thus, in the early stages of manufacture, Clovis bifaces
also served as cores.
A second component that has been recognized in many Clovis assem-

blages is the prismatic blade. Although blades were noted by Green (1963)
at the Blackwater Draw site, only fairly recently have archaeologists focused
much attention on them. Struck from large prismatic cores, Clovis blades
range between 50 and 160mm in length, and have small platforms, flat bulbs,
smooth interiors, and marked curvature (Collins and Lohse 2004:176). Collins
and Lohse (2004:176) argue that this constellation of traits is so distinctive
that such blades are almost as diagnostic as are Clovis projectile points. Some
blades were used as they came directly from the core while others were broken
up and refashioned into tools, such as scrapers (Collins 1999:17).
Collins states that prismatic blades were struck from both conical and wedge-
shaped cores, with the former being the most common. He (1999:8587) uses
13 measures and indices to define criteria for the identification of Clovis blades
(Figure5.2). For example, one of the most readily observable attributes of the
Clovis blade is its profound curvature; another is its flat bulb. Using these crite-
ria Collins reviews reported finds of prismatic blades, identifying those that fit
these criteria and those that do not. We have since added to this inventory and
our results can be seen in Figure5.3. There are two notable factors concern-
ing the distribution of sites represented in Figure5.3. First, this figure shows
locations for which at least one prismatic blade has been reported; it does not
indicate the number of blades at those locations. But there are sizeable differ-
ences in the frequency of blade occurrence from region to region. For exam-
ple, in the area centering on central Tennessee and Kentucky, several quarry
sites have yielded hundreds of blades and blade cores. The Adams site is one
of four sites situated along the Little River in western Kentucky (Figure5.3).
Blades and blade cores have been recovered from all four of these sites; 140
cores and core fragments have been collected from Adams alone and at least
30 have been collected from another, the Ezell site (Yahnig 2004:112). Farther
to the west, in the Western Valley of Tennessee, is the Carson-Conn-Short site,
which has yielded 226 blade cores and 1,956 blades and tools made on blades
(Stanford etal. 2006:145). In stark contrast, the sites in the Intermountain West
are blade-poor, where numbers range from one to five.
The second notable point concerning the locations in Figure5.3 is that
not all of the reported blades fit Collinss criteria for Clovis blades. This
is especially true in the northeastern United States. Collins goes so far as
to state that although many of these reported artifacts fit the definition of
blades, they are not part of a robust blade technology. These blades are gen-
erally smaller and have less curvature than Clovis blades. For example, the
mean curvature index for Clovis blades, as measured by Collins (1999) on
blades from 12 Clovis sites, ranges from 14.4 for the five blades from the
East Wenatchee Cache in Washington state to 5.8 for the 33 blades from
Pavo Real in Texas. Blades and blade tools are rarely shown in profile in the
ma
xim
um
left edge thi
T ckn
margin distal end es
s
interior platform juncture
exterior
edge modification right edge
margin
errailure scar proximal end platform

angle maximum
width
W
prior
maximum length L
ripple
a blade
marks
b scars
fissures
platform
depth
platform
lb
Curvature Index
bu
width b
x 100
a
5.2. Measurements, attributes, and landmarks considered by Collins (1999) in his study of Clovis
blades. (From Collins 1999:80:87, fig.5.3.)
publications for the northeastern Paleoindian sites, but two limaces made on
blades (Grimes and Grimes 1985) from the Michaud site in southern Maine
are so pictured (Speiss and Wilson 1987: Plate 3.12). One is broken but the
other yielded a curvature index value of 0.78, considerably lower than the
mean for the Pavo Real blades.
N
blades that conform to Collinss (1999)
criteria for Clovis blades 0 200 400 600
blades that conform only to some of Collinss km
(1999) criteria for Clovis blades
5.3. Distribution of prismatic blades.
The overall pattern seems to be that Clovis blades are most numerous in
the southern Plains and the mid-South. They decline noticeably outside of
those regions. Although blades are represented in the Northeast, they do not
appear to be Clovis blades; we return to this point later. In the Intermountain
West blades occur sporadically and do not always occur with fluted points.
For example, two small prismatic blade cores and several blades have been
recovered from the Old River Bed in western Utah; although more than
2,100 artifacts have been collected thus far, no fluted points have been
found. Diagnostics are, instead, Western Stemmed points. In fact, in 4 of the
9 Intermountain West cases, stemmed points are present in the assemblage
containing blades.
A similar pattern in Clovis blade distribution is evident in the content of the
artifact caches believed to be associated with Clovis (Figure5.4). The spatial
distribution of 25 of the 30 caches follows that of the dated Clovis sites along
the eastern Rocky Mountain front and then into the Columbia Plateau. The
East Wenatchee
Beach Pelland
Anzick
Simon Crook
County
Rummels-Maske
Franey
Fenn Watts Carlisle
Drake
Mahaffy Busse Mckinnis
CW
Sailor-Helton
JS
Cedar Anadarko
Green
Creek Garland
Dickenson
Evant
Kevin Davis
Crocket Gardens
De Graffenried
Eisenhauer Commanche
Hill
Hogeye
0 200 400 600
km
5.4. Distribution of Clovis caches (caches locations from Collins 1999 and Kilby 2008).
content of these caches changes from south to north (Kilby 2008). All but two
of the most southerly located caches (de Graffenried and Hogeye) contain
only blades and/or blade cores. The next set of caches, with the exception of
Sailor-Helton, contain bifaces and flakes but still have significant numbers of
blades and/or blade cores. Sailor-Helton, however, contains no bifaces, but 8 of
10 cores are conical blade cores and most of the remaining 116 items are blades
or blade-like flakes. Farther north, blades virtually disappear from the caches,
except for Pelland, which contains eight. The CW, Drake, Crook County, and
Simon caches have none, while the Anzick and Fenn caches contain one and
the East Wenatchee cache contains four.These caches are dominated by bifaces
and finished points.
Although only two of the most southerly caches contains bifaces, these tools
dominate non-cache assemblages.They are relatively small, however, compared
with their northern counterparts. Huckell (2007:194) states those found at
Murray Springs would have made excellent hand-held cutting tools and could
be easily resharpened. Its hard to imagine anyone using most of the extremely
large bifaces in the northern caches as hand-held cutting tools. The bifaces
in these latter caches are also more elaborate. The finished points in the East
Wenatchee cache, for example, are extraordinary in size and execution. An
additional change is the appearance of red ochre, present in JS, Busse, Crook
County, Anzick, Fenn, Simon, and East Wenatchee.
Thus, the trend appears to be that caches to the south are blade-oriented
while those to the north are biface-oriented. This may actually relate to a
change in the function of caches. There are a number of opinions concerning
their function, one of which is that they served as insurance gear for a migrat-
ing population moving quickly across uncharted territory (Meltzer 2002).
Another suggestion by Lahren and Bonnichsen (1974) is that they served a rit-
ualistic function. Anzick, for example, was associated with a burial; Wilke etal.
(1991) state that this cache was likely never meant to be retrieved. Perhaps both
explanations are correct the caches served as insurance gear in the southern
Plains and took on additional ritualistic significance as people moved north-
ward. Whatever the case, blades become fewer in caches increasingly distant
from the southern Plains.
In summary, although blades never constitute a large portion of non-cache
Clovis assemblages except at some quarry sites, they appear to comprise an
important component of the technology in the southern Plains and the mid-
South.As populations moved out of these areas, however, this technology seems
to have become less important, except in the Northeast. As mentioned earlier,
non-Clovis blades are prevalent in the Northeast, associated with Paleoindian
sites that date slightly later than Clovis in the southern Plains. The diagnos-
tic fluted points from these sites are not Clovis, but Gainey and Bull Brook
(Deller and Ellis 1992). How might we explain the seeming expansion in the
use of biface technology in Clovis at the expense of blade technology, which
ultimately resulted in the disappearance of the latter? Was the demise of blades
the result of loss of the function(s) to which they were put or did a less costly
substitute replacethem?
When to Invest in a More Expensive Alternative

Our intuition (biased as it is by modern economic thinking) tells us that
when we have a choice between alternative technologies that differ in cost,
we should select the more expensive one when doing so gives a potential for
higher payback than the cheaper one. For instance, we may wear an expensive
suit rather than blue jeans if by doing so we will impress an individual who
might be considering us for a job. But if our chance of actually encountering
those potential employers is low, is there a payoff for dressing up, or can we
meet our needs for warmth and modesty more cheaply? Ugan etal. (2003)
take up this question when does it pay to invest in technology? with a
formal model derived from behavioral ecology. They observe that the cost
of technology is often treated as a fixed component of the handling costs of
resources; what is permitted to vary in foraging models usually is the density
of resources, and hence the costs associated with finding those resources, or in
transport models, the distances over which resources are conveyed (e.g. Beck
et al. 2002). But they wonder: when should investments in technology be
increased or decreased? As we know from experience, a function often can be
met by technological alternatives, some crude and laborious to use, some ele-
gant, but overkill in most circumstances, and some just right.You can drive
a nail with a rock, for instance, or you can buy a nail gun to do the same. But if
you only need to drive a few nails, your best choice might be a hammer. What
Ugan etal. show graphically and computationally is that the decision to invest
in technology, that is, whether or not to commit more effort to improve the
design or reliability of a tool, or to make a different and more effective tool, is
dependent on the time that will be spent in activities that involve thattool.
In brief, Ugan et al. (2003:1321) observe that investment in technology
should go up as total handling time increases. In the typical case, incremental
investments in technology are met by substantial gains in productivity (gain
function) initially, but these productivity gains begin to slow in the fashion of
a diminishing return function. As this upwardly bowing curve flattens out, a
point is reached when further investment in improving the technology will
produce no additional gains. But far before that point is reached, usually, deci-
sions will be made whether to invest more in improvements or whether that
time would be better spent in some other activity.
Bettinger etal. (2006), in reviewing the Ugan etal. model, do not dispute
this argument; they do, however, question the appropriateness of the model
to represent the relation between manufacture cost and tool efficacy for all
alternatives. They distinguish between choices within the same category, such
as between a less costly and more costly fish hook, and choices in different
categories, say between a fish hook and a gill net (Bettinger etal. 2006). They
argue that a more appropriate approach is to model each technological choice
with its own costbenefit curve (see Figure 5.5) (Bettinger etal. 2006:541).
Bettinger etal. argue that the choice between two technologies depends on
three factors: the length of time the tool will be in use (referred to as use time),
the length of time it takes to manufacture the tool, and the return rate using
the tool. For example, say it takes one hour to manufacture tool a, which has
a return rate of x, and five hours to manufacture tool b, with a return rate of
2x. If your work requires the use of the tool only a few times, it is not worth
the extra effort to manufacture tool b because your return advantage is mini-
mal. But if your work, instead, requires frequent heavy use of the tool, then
an investment in manufacturing tool b may be worth the effort. But how do
we know when the threshold to change from one to the other is reached?
Bettinger etal. (2006:541) state that when the time spent in procurement using
either technology produces equivalent returns (referred to as the critical use
time), it is more beneficial to switch to the more costlytool.
f (m)
V1 V0
Vxy m
5.5.The relationship between the time spent in the manufacture of a tool and its utility. (After
Bettinger et al. 2006:540, fig.1.)
C
f (m)
A
A
Vab Vac a c2 c1 b
Vxy m
5.6. Curve-estimate model for finding time thresholds at which an optimal forager will switch
to a different technological alternative. (After Bettinger etal. 2006:543, fig.3.)
If the costbenefit curves of these two technologies can be described pre-

cisely, the critical use time (vab) is defined by the model shown in Figure5.6.
A line is drawn in such a way that it lies tangent to both curves; the critical use
time then is the distance from the origin to the intercept of the line and the
x-axis. Should the time spent fall to the right of the intercept, option A should
continue to be used. If the time spent falls to the left of the intercept, however,
option B should replace A in that activity. When a third option is included,
one of the three will always be less optimal unless all three fall along the same
return function.
Although Bettinger etal. take this argument a little further, we need not do
so for the purposes here. The essential point to take from this discussion is the
implication of these relationships made earlier by Ugan etal. As the actual (or
anticipated) effort in some activity increases, it becomes more likely that the
more costly technological alternative will be selected over the less effective, less
costly one. Thus it makes sense to invest when plans call for a commensurate
investment in a procurementtask.
When Is the Time Invested in Manufacture TooMuch?

Such an elegant model invites attempts to apply it to empirical cases, but it
does not take much reflection to recognize the scope of this task.We must first
define the costs incurred by adopting either a blade-based or a flake-based
technology. The first is the time spent in the manufacture of the blade itself
and the tools made from it. A tool such as a simple flake reaches its maximum
utility with virtually no time investment, so long as raw material for its man-
ufacture is readily available. This holds especially true for flakes coming from
the thinning of bifacial cores, as has already been discussed for Clovis. More
complicated bifacially flaked tools, like Clovis points, while not particularly
expensive from an energetic standpoint, do have opportunity costs given the
time required for their manufacture. Similar investment must be made in the
preparation and rejuvenation of a blade core, but additional investment must
be made in the systematic removal of blades from this core. As noted earlier,
considerable skill and care are required in the latter activities; a misstep can
render the core useless without extensive reworking.
Although the manufacture time for blades is probably somewhat greater
than that invested in a bifacial core, probably a larger share of the investment
in blade tool manufacture lies in material acquisition. As mentioned earlier,
the manufacture of large prismatic blade cores such as were used in Clovis
technology requires more specific sizes, shapes, and quality of toolstone. In
other words, good quality material must occur in large packages and blocky
shapes to accommodate core manufacture. Even if some of these costs are
shared with foraging activities, they are undoubtedly large if significant time
must be expended to quarry and assay raw material, especially at sources of
variable quality.
The third and final cost component is transport, which is likewise shared
with other components of the transported load. As evidence of the significance
of this cost, we find that there is a good deal of actual manufacture of blades
(and other tools) at quarries.This is done, presumably, to eliminate a large share
of low utility waste before the load is assembled, thus reducing the per unit
transport cost. However, transporting cores from two technologies bifaces
for the manufacture of fluted points and flake-based tools and blades for the
specific manufacture of scrapers increase costs considerably.
The task of estimating these costs is a bit daunting. We might gain some
insight about manufacturing costs from experimental studies. The task of esti-
mation grows more difficult, however, when we turn to material acquisition
and transport costs. These components of manufacturing cost are dependent
KRF
WRS NEH
SHJ
ALB
TEC
EDP
N
0 200 400 600
km
5.7. Locations of high-quality toolstone sources on the Great Plains. (After Bamforth 2009:144,
fig. 1.) ALB (Alibates agate), EDP (Edwards Plateau chert), KRF (Knife River Flint), NEH
(Nehawka chert), SHJ (Smoky Hill jasper), TEC (Tecovas jasper), WRS (White River
silicates).
on situational contexts, the distribution of geological sources of raw material,

and the patterns of movement followed by Clovis groups.
But even if we cannot estimate values for these variables, the model does
provide some predictions as to how the record of blade manufacture and use
might differ across the Clovis landscape. We just alluded to the first of these
predictions. Because acquisition and transport costs are such a large component
of the cost of manufacturing blades, their production would have been greatest
in those landscapes where suitable raw material was common and where pat-
terns of movement did not exceed the bounds of the resupply range. If Clovis
people were moving beyond resupply zones, a strategy for decreasing acquisi-
tion costs might involve distributing raw materials across resupply outposts, or
in other words, to supply caches (Bamforth 2009; Meltzer 2002, 2007).
These circumstances seem to describe the blade record of the southern Plains,
where Edwards Plateau chert outcrops across a very large area (Figure5.7). We
note that the cached materials, comprised almost completely of small numbers
of blades and/or blade cores, extend the effective distribution of this material
type. If, as Faught (2009) suggests, the earliest colonization of the southern
region was via the Gulf coast, we might surmise that if we could inventory the
now flooded coastal plain, it would also contain Clovis caches of blades and
blade cores of Edwards Plateau chert. Certainly there is evidence of Clovis on
the continental shelf as Meltzer (1993:296) notes that numerous Clovis points
have been found along the beaches of the Texas Gulf Coast, where they wash
up from offshore, drowned sites.
Moving farther northward, the distribution of high quality lithic sources
grows more spotty (Figure 5.7). The areal extent for most of these is small,
especially compared with that of Edwards Plateau. In addition, many of these
sources may not have been known to a colonizing population, rendering the
landscape source poor. It is notable that the caches to the north are more
diverse and contain a larger number of items in sharp contrast to the small
blade/blade core caches to the south. As these caches were likely created to
serve as alternatives to outcrops, the larger number and greater diversity of
items within them suggests that expectations for encountering a raw mate-
rial outcrop were lower than they were to the south. The consequence of this
seeming paucity of raw material sources would be an increase in the costs of
blade manufacture, and thus the drop in blade frequency, both in cache and
non-cache settings. But, you may say, Clovis people continued to manufacture
their costly bifaces.Wouldnt this manufacturing technology also be influenced
by effectively diminished availability of high quality raw material? Certainly
bifaces, just like blades, require large packages of isotropic raw material; how-
ever, they also permit a greater range of package geometries than do Clovis
blade cores. Simply, bifaces can be made from sheet-like cherts, which are
less suitable for blade core manufacture. The other part of the answer to this
question is that bifaces will be made whether or not other tools are made on
blade or flake blanks. That flake blanks can be struck from bifacial cores and
prismatic blades cannot be would seem to favor flake blanks in those circum-
stances where transport of a potentially redundant technology becomes pro-
hibitively costly.
So, this seems to us to be part of the answer for the question as to why the
contents of Clovis caches change from south to north. Blades proved to be an
increasingly more costly alternative to flake blanks. Their selective advantage
to their users was realized in places where raw material was abundant and
foraging ranges were restricted. They grew less common where raw material
sources were less abundant or where foraging ranges were large. As a corol-
lary, we might anticipate that the range of uses to which blades were put also
narrowed.
There is a second element, of course, that plays a part in the choice of
technological alternatives the commitment of effort to a tool-using activ-
ity. Clovis blades were used both in unmodified and modified form; when
modified, they were almost always fashioned into end or side scrapers.
Use-wear on unmodified specimens often suggests scraping activities as well
(e.g., Green 1963). The choice to use these tools should scale approximately to
the importance of hunting and hide-working. There can be little doubt that
hunting played a role in the Clovis subsistence strategy, but there is a good
deal of debate over its importance. For our purposes here it would be ideal
if we could estimate the contribution of hunting over the geographic range
of Clovis, but so few faunal and botanical remains have been found, this is
not possible. On the face of it, the toolkit suggests that hunting was a signif-
icant aspect of the Clovis adaptation in the southern Plains. We find this to
be credible if the circumstances of Clovis arrival here is that of a colonization
coastal population who, as Kelly and Todd (1988) hypothesize, find that hunt-
ing is most immediately effective when knowledge of other resource sets is less
secure (see also Meltzer 2007).
To the degree that Clovis is an exploratory-colonizing strategy, hunting
was probably always important. The fall-off in blade use, as discussed earlier,
is, we believe, a response to raw material availability, mobility, and possibly a
change in hunting strategy, but it is not a measure of the relative importance of
hunting. Interestingly, where blade use continues in a significant way is in the
Great Lakes region and the Northeast, where populations are believed to have
become relatively specialized in the hunting of caribou. Also notable for sites
in these regions, scrapers of many different forms dominate the tool assem-
blages, at some sites overwhelmingly, and here as in Clovis assemblages, it is
scrapers that are most often made on blades.
Conclusion
As noted earlier in the chapter, the Intermountain West is blade poor when
compared with most other regions, and in 4 of the occurrences blades do not
fit Collinss Clovis criteria. A Clovis population coming into the Columbia
Plateau from the Plains would have encountered a quite different lithic ter-
raine, especially with movement southward into the Great Basin. Although
chert outcrops occur on the Plateau, areas to the south are dominated by
obsidian and fine-grained volcanics. Although obsidian is a preferred material
for projectile points, it is not for scrapers because it is too brittle (Beck and
Jones 1990, 2009). Scrapers are sometimes made from fine-grained volcanics
because this is a tough material like chert, but the control and precision needed
for the manufacture of Clovis blade cores and removal of blades from those
cores would prohibit its consideration.The fact that two Clovis blade cores and
a small number of Clovis blades have been recovered from sites in the Great
Basin suggests the persistence of a functionally circumscribed activity still rel-
evant in some contexts or they are the tail end of a disappearing technological
lineage, where the frequency of blades more strongly reflects sampling pro-
cesses of drift than to selection effects.
In addition, Clovis populations entering the Intermountain West would
have encountered people utilizing a very different technology, one likely well
suited to the region and its lithic terraine and would have been less costly
overall than the more efficient but higher cost Clovis technology.Whatever the
case, Clovis, although recognized in the Intermountain West, never had much
of a presence there.
References
Andrefsky, William, Jr. 1998. Lithics. Macroscopic A View from the North American West. In
Approaches to Analysis. Cambridge University Paleoamerican Odyssey, edited by Kelly E. Graf,
Press, Cambridge. Caroline V. Ketron, and Michael R. Waters,
Bamforth, Douglas B. 2009. Projectile Points, pp. 273291. Center for the First Americans,
People, and Plains Paleoindian Perambulations. Austin.
Journal of Anthropological Archaeology Beck, Charlotte, Amanda Taylor, George T. Jones,
28:142157. Cynthia M. Fadem, Catlyn R. Cook, and Sara
Bar-Yosef, Ofer, and Steven L. Kuhn. 1999. The A. Millward. 2002. Rocks Are Heavy:Transport
Big Deal about Blades: Laminar Technologies Costs and Paleoarchaic Quarry Behavior
and Human Evolution. American Anthropologist in the Great Basin. Journal of Anthropological
101:322338. Archaeology 21:481507.
Beck, Charlotte, and George T. Jones. 1990. Bettinger, Robert L., Bruce Winterhalder, and
Toolstone Selection and Lithic Technology in Richard McElreath. 2006. A Simple Model
Early Great Basin Prehistory. Journal of Field of Technological Intensification. Journal of
Archaeology 17:283299. Archaeological Science 33:538545.
Beck, Charlotte, and George T. Jones. 2009. The Bradley, Bruce A., Michael B. Collin, and Andrew
Archaeology of the Eastern Nevada Paleoarchaic, Hemmings. 2010. Clovis Technology. International
Part 1: The Sunshine Locality, with contri- Monographs in Prehistory, Archaeological Series 17,
butions by Jack M. Broughton, Michael Ann Arbor.
D. Cannon, Amy Dansie, Amy M. Holmes, Collins, Michael B. 1999. Clovis Blade Technology.
Gary A. Huckleberry, Philip W. Hutchinson, University of Texas Press, Austin.
Stephanie D. Livingston, and Donald R.Tuohy. Collins, Michael B., and Jon C. Lohse. 2004. The
The University of Utah Anthropological Nature of Clovis Blades and Blade Cores. In
Papers No.126. Entering America. Northeast Asia and Beringia
Beck, Charlotte, and George T. Jones. 2010. Clovis before the Last Glacial Maximum, edited by
and Western Stemmed: Population Migration David B. Madsen, pp. 159183. The University
and the Meeting of Two Technologies in of Utah Press, Salt LakeCity.
the Intermountain West. American Antiquity Crabtree, Don E. 1972. An Introduction to
75:81116. Flintworking. Occasional Papers of the Idaho
Beck, Charlotte, and George T. Jones. 2012. The State University Museum No. 28. Pocatello.
Clovis-Last Hypothesis: Investigating Early Deller, D. Brian, and Christopher J. Ellis. 1992.
Lithic Technology in the Intermountain West. Thedford II. A Paleo-Indian Site in the Ausable
In Meetings at the Margins. Prehistoric Cultural River Watershed of Southwestern Ontario.
Interactions in the Intermountain West, edited by Memoirs of the Museum of Anthropology,
David Rhode, pp. 23-46. University of Utah University of Michigan No.24.
Press, Salt Lake City. Faught, Michael K. 2008. Archaeological Roots
Beck, Charlotte, and George T. Jones. 2013. of Human Diversity in the New World:
Complexities of the Colonization Process: A Compilation of Accurate and Precise
Radiocarbon Ages from Earliest Sites. American Meltzer, David J. 2007. Modeling the Initial
Antiquity 73:670698. Colonization of the Americas. Issues of Scale,
Green, F. E. 1963. The Clovis Blades: An Demography, and Landscape Learning. In
Important Addition to the Llano Complex. The Settlement of the American Continents. A
American Antiquity 29:145165. Multidisciplinary Approach to Human Biogeography,
Grimes, John R., and Beth G. Grimes. 1985. pp. 123137. The University of Arizona Press,
Flakeshavers: Morphometric, Functional and Tucson.
Life-Cycle Analyses of a Paleoindian Unifacial Parry, William J. 1994. Prismatic Blade
Tool Class. Archaeology of Eastern North America Technologies in North America. In The
13:3557. Organization of North American Prehistoric
Hayden, Brian, Nora Franco, and Jim Spafford. Chipped Stone Tool Technologies, edited by Philip
1996. Evaluating Lithic Strategies and Design J. Carr, pp. 8798. International Monographs
Criteria. In Stone Tools. Theoretical Insights into in Prehistory No. 7, Ann Arbor.
Human Prehistory, edited by George H. Odell, Spiess, Arthur E., and Deborah Brush Wilson.
pp. 945. Plenum Press, NewYork. 1987. Michaud. A Paleoindian Site in the New
Huckell, Bruce B. 2007. Clovis LithicTechnology: England-Maritimes Maritimes Region.Occasional
A View from the Upper San Pedro Valley. In Publications in Maine Archaeology No. 6.
Murray Springs. A Clovis Site with Multiple Augusta.
Activity Areas in the San Pedro Valley, Arizona, Stanford, Dennis J., Elmo Leon Canales, John B.
edited by C. Vance Haynes, Jr. and Bruce B. Broster, and Mark R. Norton. 2006. Clovis
Huckell, pp. 170213. Anthropological Papers Blade Manufacture: Preliminary Data from
of the University of Arizona, No. 7. University the Carson-Conn-Short Site (40Bn190),
of Arizona Press, Tucson. Tennessee. Current Research in the Pleistocene
Kelly, Robert L., and Larry C.Todd. 1988. Coming 23:145147.
into the Country: Early Paleoindian Hunting Tankersley, Kenneth B. 2004. The Concept of
and Mobility. American Antiquity 53:231244. Clovis and the Peopling of the New World.
Kilby, J. David. 2008. An Investigation of Clovis In The Settlement of the American Continents:
Caches: Content, Function, and Technological A Multidisciplinary Approach to Human
Organization. Unpublished Ph.D Dissertation. Biogeography, edited by C. Michael Brton,
Department of Anthropology, University of Geoffrey A. Clark, David R. Yesner, and
New Mexico, Albuquerque. Georges A. Pearson, pp. 4963.The University
Lahren, L., and Robson Bonnichsen. 1974. of Arizona Press, Tucson.
Bone Foreshafts from a Clovis Burial in Ugan, Andrew, Jason Bright, and A. Rogers.
Southwestern Montana. Science 186:147150. 2003. When Is Technology Worth the
Meltzer, David J. 1993. Is There a Clovis Trouble? Journal of Archaeological Science
Adaptation? In From Kostenki to Clovis. Upper 30:13151329.
Paleolithic-Paleo-Indian Adaptations, edited by Wilke, Philip J., Jeffrey J. Flenniken, and T. L.
Olga Soffer and N. D. Praslov, pp. 293310. Ozbun. 1991. Clovis Technology at the Anzick
New York: Plenum Press. Site, Montana. Journal of California and Great
Meltzer, David J. 2002. What Do You Do When Basin Anthropology 12:242272.
No Ones Been There Before? Thoughts on Yahnig, Carl. 2004. Lithic Technology of the
the Exploration and Colonization of New Little River Clovis Complex, Christian
Lands. In The First Americans: The Pleistocene County, Kentucky. In New Perspectives on the
Colonization of the New World, edited by N. G. First Americans, edited by Bradley T. Lepper
Jablonski, pp. 2758. Memoirs of the California and Robson Bonnichsen, pp. 111117. Texas
Academy of Sciences No. 27. San Francisco. A&M University Press, College Station.
Six
The North China Nanolithic
Robert L. Bettinger, Christopher Morgan, and Loukas Barton
This chapter presents a very simple argument: that technology in general, and
lithic technology in particular, can shed critical light on conditions surround-
ing and contributing to major behavioral innovations, in this case the origin
of agriculture. There are probably as many views on the subject as papers,
but there is a fairly clear divide between those who argue that agriculture
evolves under conditions of scarcity among them Binford (1968), Bar-Yosef
(1998), Childe (1951), and others (e.g., Moore et al. 2000)), and those who
argue that it evolves under conditions of plenty (Braidwood and Howe 1960;
Price and Gebauer 1995:79). The conditions of plenty view is prominent
in discussions of the emergence of millet agriculture in North China (Barton
2009) and specifically the argument that, in common with nearly all early
experiments with food production, millet farming developed among complex,
affluent hunter-gatherers living in large, permanent settlements in highly
productive riparian and lacustrine settings that offered a rich variety of wild
plants and animals (Crawford 2006:91; Smith 1995). This view portrays exper-
iments with millet farming as solidifying a position of strength, increasing
the yield and reliability of an already important staple in an already intensive
and highly successful hunting-and-gathering economy (Smith 1995:136137).
Lu (2006), on the other hand, advocates the alternative conditions of scar-
city view. Observing no archaeological evidence that Chinas first farmers
were sedentary and that, in contradiction to the abundance argument, agri-
culture arrived relatively late in the areas of greatest natural plant and animal
100
Th e North Chi n a N a n o l i t hi c 101
productivity (e.g., South China, Yangzi basin), Lu (2006:146149) argues that

sedentism and food production were both responses to declining wild resource
return rates resulting from population growth leading to overharvesting and
territorial circumscription.
Scarcity versus Abundance: Test Implications

In theory, it should be easy to devise tests permitting a clear choice between
these alternative evolutionary scenarios for the origins of millet farming. The
diet breadth model drawn from optimal foraging theory, for example, pro-
vides formal predictions connecting subsistence choice to resource abun-
dance (Bettinger 2009). When resources are abundant relative to demand, diet
breadth should be narrow; more formally, the marginal rate of return (i.e., the
cost threshold below which resources should be ignored) will be high and
foragers should pursue only high-ranked resources, that is, resources that are
relatively easy to acquire and process, producing returns greater than or equal
to the marginal rate of return. As resources grow scarce, diet breadth should
widen; again, more formally, the marginal rate of return drops and consum-
ers should become less selective, adding low-ranked resources, which return
increasingly less food value per unit of procurement and processing time. The
overall state of subsistence (abundance relative to demand), then, is measured
by the marginal rate of return and indexed by the handling time of the costly
(i.e., lowest ranked) resource in the diet.Thus, the conditions of plenty argu-
ment would have agriculture evolving in concert with subsistence patterns,
and consequently archaeobotanical and archaeofaunal assemblages, dominated
by high ranking taxa; the conditions of scarcity argument would have just
the reverse, assemblages dominated by low-rankingtaxa.
Unfortunately, the distinction between high- and low-ranked species is
not always clear, certainly not in the archaeological record. Body size may
(Broughton 1999) or may not (Bird, etal. 2009) index prey rank, for example.
Too, in most instances, rank is often more a function of context than species a
diseased deer sighted on the far side of a steep canyon will probably rank lower
than a fat rabbit with a broken leg sighted a few yards away on this side. These
and many additional complications that come readily to mind (e.g., differential
butchering and preservation) prevent any simple reading of archaeological diet
breadth from floral or faunal assemblages.
Although procurement technology complicates resource ranking (netted
and speared salmon ranking differently, for instance), it affords more consistent
and direct insights about relative resource abundance than judgments about
the marginal rate of return drawn from diet breadth subsistence remains. This
is because the rational tool maker will fashion all but the most expedient tools
to optimize expected future returns, that is, in accord with projections about
102 Bett i nge r e ta l .
resource conditions and some central tendency (mean, minimum, maximum,

etc.) of the marginal rate of return likely to obtain over some future period,
presumably the life of the tool being made. Although actuated via decisions
informed by essentially the same projections, the sum total of foraging that
actually occurs during this interval, and thus the archaeological residue gener-
ated, will vary stochastically in response to real resource encounters. In this
sense, the diet breadth model incorporates the marginal rate of return in a
form of environmental possibilism (Bettinger 1991). Diet breadth cannot pre-
dict what foraging will occur, because resource encounters are random; diet
breadth merely predicts what foraging will not occur, specifically for resources
with handling returns below the marginal rate of return. How closely actual
foraging behavior indexes the marginal rate of return is an empirical problem,
depending on sample size (number of resource encounters) and the probability
of encounters with resources with handling return rates immediately above
the marginal rate of return. Technology does not suffer this defect. Both for-
aging and tool making require judgments about resource supply and demand
(i.e., the marginal rate of return), but tool design is determined by these judg-
ments, foraging merely limited bythem.
Macrofossils representing on-the-ground subsistence behavior are obviously
important, but so long as demand for, and supply of, resources does not vary
dramatically over the use lives of the implements employed in their procure-
ment, technology provides a more reliable index of forager judgments about
these conditions.The clearest predictions about technology in relation to these
basic relationships is the simple model of technological investment developed
by Bettinger etal. (2006) from the more complex version originally presented
by Ugan etal. (2003). Space permits only brief explication of this model here,
a more detailed treatment, complete with exercises, being presented elsewhere
(Bettinger 2009).
Technological Investment
In simplest form, the technological investment model examines the replacement
of alternate procurement technologies characterized by two key variables:
mi manufacturing time time spent making a particular kind of procurement technol-
ogy i, expressed here in hours(hr)
ri procurement rate the rate at which a resource is obtained using technology
i, expressed here in kilocalories per hour (kcal/hr)
In Figure6.1, for example, technology 1 and 2 are characterized by their

manufacturing times m1 and m2, by the rate at which kilocalories are procured
using them r1 and r2, and by the relationship of these variables r1 / m1 and r2 / m2,
the rate at which manufacturing time increases rate of procurement, that
m2 > m1
r 2 > r1
r 1 / m 1 > r2 / m 2
2
r2
return rate r
1
2
r1 /m
r2
1
/m
r1
m1 m2
manufacturing time m
6.1. Relationship between two technologies characterized by manufacturing times m1 and m2,
by the rate at which kilocalories are procured using them r1 and r2, and by the relationship of
these variables r1 / m1 and r2/m2.
is, investing in m1 produces a return rate of r1 and investing in m2 produces

a return rate of r2. Here technological evolution is treated as a competition
between technologies characterized by different combinations of ri, mi, and ri /
mi. In the simplest case there are two alternatives. For either to be competitive
in the presence of the other, it must satisfy one of two conditions.
1. The more costly technology must produce a higher rate of return. Formally, if
m2 > m1, then r2>r1
2. The lower return technology must produce a rate of return per unit of
manufacturing time that is at least equal to that of the technology with the
higher return. Lower returns can be justified only by costs low enough to make
the lower return technology at least equivalent to its higher return alternative in
terms of rate of return per unit of manufacturing time. Formally, if r1 < r2, then
r1/m1 r2/m2
Meeting both these conditions, the alternative technologies 1 and 2 depicted

in Figure6.1 are mutually viable; the more costly technology produces a higher
rate of return (r2 > r1); and the lower return technology a greater rate of return
per unit of manufacturing time (r1/m1 > r2/m2).
With only these variables (manufacturing time and return rate) in play, of
course, it will never pay to invest in the more costly technology 2; technology
1 is superior in all respects, both cheaper (m1 < m2) and producing a better
2
)

+s 1
return rate r
2
+s 2
) / (m
r2
(m 2
)=
r2/ 2
s )< +s 1

+ 2 1
/ (m
1
/ (m
1
r1 ) r1
+s
1
m2
/(
r2
)>
+s
1
m1
s2 s12 s1 /(
r1
procurement time s manufacturing time m

time
6.2. Relationship between two mutually viable technologies. The more costly technology 2
produces a higher rate of return (m2 > m1 and r2 > r2) and the less productive technology 1 pro-
duces a higher rate of return per unit of manufacturing time (r1 < r2 and r1 / m1 > r2 / m2), mak-
ing technology 1 superior when procurement time is less than s12 and technology 2 superior
when procurement time is greater thans12.
payoff per unit of manufacturing time(r1/m1 > r2/m2). The more costly tech-
nology 2 may make sense, however, if procurement returns are measured not
just in relation to time spent making tools but in relation to time spent making
and using tools, in this case, the sum of manufacturing time m and procure-
ment time s, formally definedas
s the total amount of time expended in procurement of a resource, again expressed
in hours(hr).
When procurement time s is included, the object is to maximize return

rate ri relative to the sum of manufacturing time and procurement time, that
is, maximize r i /(m i + s) rather than r i /m i. Thus even when r1/m1 > r2/m2,
because r2 > r1, if s is large enough r2/(m2+s) > r1/(m1+s). This is shown by the
solid line in Figure6.2, where the return rate of technology 1 and 2 are just
equal at s12, which is the procurement time switching point between the
cheaper technology 1 and more costly technology 2. When procurement time
is greater than s12, the costlier technology produces a higher rate of return.
When it is less than s12, the cheaper technology produces a higher rate of
return.
In addition to the kind of replacement discussed in the preceding text,
technological intensification can also take the form of refinement, reflected in
increasingly more effective and costly versions of a basic design, for exam-
ple, fishhooks. As when two qualitatively distinct technologies are involved
(e.g., nets vs. hooks), refinement increases procurement returns at a decreasing
rate, in this case, however, as a continuous function, increasing procurement
return rate r
4
s4 3
s3 s2 s1 m1 m2 m3 m4
procurement time s manufacturing time m

time
6.3. Relationship between manufacturing time and return rate as a single technology is increas-
ingly refined, increased manufacturing cost producing higher returns at a steadily decreasing
rate. Incremental increases in procurement time warrant incremental technological refinement.
time warranting increasingly more productive but costly refinements. Critical

procurement time the procurement time required to warrant a particular
degree of refinement is given by the x-axis intersection of tangents to the
curve describing change in procurement rate as a function of manufactur-
ing time (Figure6.3). Refinement of the technology depicted in Figure6.3
will begin when procurement time reaches s1 and proceed continuously with
increasing procurement time. The point labeled 1 represents the technology
in unrefined form, the point labeled 4 its most refined form, and points 2
and 3 intermediate degrees of refinement warranted when procurement time
reaches exactly s2 and s3 respectively.
Whether dealing with two qualitatively distinct but mutually viable tech-
nologies or differentially refined versions of a single technology, the choice
between less expensive and more costly alternatives hinges on the marginal
rate of subsistence return (the rate at which resources are acquired relative to
procurement and manufacturing time) and changes as a function of procure-
ment time (manufacturing time being fixed). If resources are abundant (i.e.,
the marginal rate of return is high), time spent in procurement will be low, and
the less expensive technology superior. If resources are scarce (the marginal
return rate is low), time spent in procurement will be high, and the more costly
technology superior. For this reason, the cost of the most expensive technology
in use at a given time will vary inversely with the overall rate of subsistence
return, that is, resource abundance relative to demand. Further, among groups
fielding the same suite of technologies in roughly the same environments for
roughly the same purposes, those under the greatest resource stress as measured
by resource supply relative to demand, will display the most costly refinements
of the most costly technology.
Early millet farming is particularly well suited to this sort of analysis because
traditional wisdom has it evolving from a well-defined techno-complex (Clark
1968) the intensive late Pleistocene hunter-gatherer adaptation termed here
the North China Microlithic, whose hallmark microblade technology was cer-
tainly among its most costly, requiring more skill and higher quality raw mate-
rials than any earlier North Asian stone tool technology. The archaeological
record of this microblade-using hunter-gatherer to millet-growing farmer
connection is poorly documented, however, and as a result the early agricul-
tural revolution in north China is not as well understood as those that have
occurred in other parts of the world.
Origins of Millet Farming In NorthChina

Early millet farming is represented in at least five geographically separate but
roughly contemporaneous cultural complexes (Laoguantai, Peiligang, Cishan,
Houli, and Xinglongwa) distributed over an environmentally diverse area
stretching 1500 km from the northeast China Plain to the western Loess Plateau
(Figure6.4).The published literature, however, does not present even one con-
tinuous stratigraphic sequence or statistically convincing seriation connect-
ing any of the cultural complexes representing north Chinas preagricultural
hunter-gatherers to any of the five representing its early millet farmers. Among
the many possible scenarios, the traditional view that millet farming evolved
from the North China Microlithic (Barnes 1993; Bettinger etal. 2007; Elston
etal. 1997; Lu 1998, 1999, 2006; Madsen etal. 1996) is certainly the most plau-
sible yet highly problematic because the hallmark microblades and microblade
cores common in North China Microlithic sites (e.g., Xiachuan, Xueguan,
Shizitan, Shayuan, Hutouliang) are rare or absent in nearby early millet farm-
ing sites (e.g., Cishan and Peiligang). Indeed, Dadiwan lies well outside of
the main distribution of the North China Microlithic (Gai 1985), the nearest
representatives of which are directly north on the upper Yellow River and
adjacent deserts. In any event, prior to the data presented here not one strati-
fied assemblage or series of related assemblages had bridged this technological
gulf from microblade-using forager to millet-raising farmer. As we show in this
chapter, this disconnect is not an artifact of archaeological bias, preservation,
or sampling. Rather, it is the signature of an agricultural transition authored
by hunter-gatherer systems operating at the very edge of their natural range,
using a microlithic technology at the very edge of its natural range, the com-
bination defining what was quite clearly a marginal environment. To support
this argument, we report recent work at the Dadiwan site, which provides the
104 108 112 116 120 124
0 500 km
Xinglongwa
42 42
.
ns
Mt
North
Tengger
Helan
Ordos Plateau China Bohai

Desert
Plain
38 38
r
ve
Ri
llow
Ye
Liu
Cishan
Loess Plateau Houli
pan
M
Dadiwan P.R.C.
tn
s.
34 Peiligang 34
Laoguantai
104 108 112 116 120 124
6.4. Location of the Dadiwan site in relation to the five early millet farming complexes of
North China.
first complete archaeological sequence from north China to record behavioral

variation across the transition to agriculture.
Dadiwan
The Dadiwan site (105.904E, 35.015N) in Shao Dian Village, Qinan County,
Gansu Province, PRC, is the oldest known example of the Dadiwan (or
Laoguantai) cultural complex, which is the westernmost expression of early
millet agriculture in North China (Figure6.4).The original excavations (1978
1984) revealed a multiphase cultural sequence beginning with a Dadiwan
phase (79007200 BP) occupation representing primitive or low-level millet
farming (Gansu 2003, 2006). As with all early millet-based agricultural sites in
North China, no evidence of a preagricultural hunting and gathering occu-
pation was found. However, surveys near the Dadiwan site in 2002 recorded
a late Pleistocene archaeological culture we have called Zhuang Lang-Tong
Xin (32,00018,000 cal BP) (Barton etal. 2007; Bettinger etal. 2005), and test
excavations in 2004 hinted that this Zhuang Lang-Tong Xin assemblage and a
microblade technology related to the preagricultural North China Microlithic
were both present at Dadiwan itself. Work in 2006 and 2009 confirmed the
presence of, and demonstrated the stratigraphic and developmental relation-
ship between, these complexes and the earliest farming phase at Dadiwan.
Although Dadiwans status as a protected national landmark limited the
scope of our work, we excavated to a depth of 10 m, removing 29.14 m3 of
deposit in 10cm arbitrary levels, screening 23.2 m3 of that through 3.0mm
mesh. With the exception of single flake found in situ between 9.6 and 9.8 m,
the cultural assemblage is confined to the upper 7.1 m of deposit, artifact dis-
tributions justifying division of the dividing the relatively undisturbed lower
6.6 m of that into six components (Figure6.5; Table6.1).
Almost all sherds (98%) recovered in the undisturbed deposit were from the
top two components (56), which match ceramic occupations recognized in
previous excavations: early incipient Dadiwan farmers (Component 5)and later
intensive Late Banpo farmers (Component 6).The preceramic Components 1,
2, 3, and 4 were separated on the basis of lithic technology described in brief
in the next section.
Lithic Technology
There are few formal or retouched tools among the 1165 pieces of chipped
stone.The dominant technology, which we call flake-and-shatter, includes simple
flake tools, angular shatter blocks, and core fragments produced by direct and
bipolar, hard-hammer percussion, predominantly on locally abundant massive
quartz river cobbles (Figure6.6), this material accounting for 95 percent of
this assemblage. This is the same percussion quartz technology that defines
the late Pleistocene Zhuang Lang-Tong Xin complex previously documented
for the western Loess Plateau (Barton etal. 2007; Ji etal. 2005). Our Dadiwan
data show that this technology is much older and persisted much later than we
previously thought, beginning by at least 60,000 (and possibly 80,000) years
and lasting until after 7000 BP. Flake-and-shatter quartz technology dominates,
indeed makes up the entirety of, the assemblages from Components 1 and 2,
and almost all of what little there is of Component 3, during which Dadiwan
appears to have been largely abandoned probably in response to climatic
deterioration during the Last Glacial Maximum (LGM).
Tiny microblades (Figure6.7) and microblade cores (Figure6.8) predomi-
nantly (98percent) fashioned from small pieces of nonlocal cryptocrystallines
(e.g., chalcedony) identify a second, very different lithic technology that is
obviously derived from the North China Microlithic. The microblade cores
are predominantly boat-shaped or pebble type (Chen 1984; Elston and
Brantingham 2002) noted elsewhere in northern China, are triangular in cross
section, and vary in taper to take maximum advantage of raw material. The
microblades at Dadiwan are correspondingly small, and although they are
mainly core preparation and maintenance waste, those taken away and used
could not have been much larger. The original excavation at the Dadiwan site
produced fewer than a dozen microblades, all larger, of different materials, and
from post-Laoguantai components. It did, however, yield several thin, tabular
bone handles with lateral margins finely slotted to accept very small micro-
blade insets of the scale we recovered (Gansu 2006). While concentrated in
y
og
ol
hn
y
c
og
Te
ol
r
hn
te
at
c
Te
Sh
e
d-
ad
an
bl
ds
ro
ke
er
ic
la
Sh
M
0 F
00
0
0
0
0
0
24
48
72
96
12
50
10
15
40
80
12
16
0
0
050
5060
6070
7080 Component 6
8090
90100
100110
110120
120130
130140
140150 Component 5
150160
160170
170180
180190
190200
200210
210220
220230
230240 Component 4
240250
250260
260270
270280
280290
290300
300310
310320
320330
330340
Component 3
340350
350360
360370
370380
380390
390400
400410
410420
420430
430440
440450
450460 Component 2
460470
470480
480490
490500
500510
510520
520530
530540
540550
550560
560570
570580
580590
590600
600610
610620 Component 1
620630
630640
640650
650660
660670
670680
680690
690700
700710
6.5. Stratigraphic distribution of major Dadiwan technologies by density per cubic meter.
table6.1. Dadiwan site components
Component Cultural complex cal BP Depth (m)

6 Late Banpo 57007000 0.90.5
5 Dadiwan (Laoquantai) 700013,000 1.90.9
4 Preceramic Microlithic 13,00020,000 2.71.9
3 LGM (abandoned?) 20,00033,000 2.73.9
2 33,00042,000 5.13.9
Zhuang Lang-Tong Xin
1 42,00060,000 7.15.1
0 ? 60,00080,000 9.97.1
6.6. Flake-and-shatter quartz technology, showing all specimens recovered from a single 10cm
level (160 to 170cm) in a 2.0 m2 unit (DDW03).
the early millet farming Dadiwan (or Laoguantai) Component 5 (Figure6.7),

this microlithic technology is clearly earlier than that, its presence in situ in
Component 4 documenting a post-LGM, preceramic occupation quite clearly
connected to the North China Microlithic (Gai 1985).
6.7. Microblades, showing all specimens recovered from a single 10cm level (160 to 170cm)
in a 2 m2 unit (DDW04). The last three in the bottom row are white chalcedony, the rest black
chalcedony.The first specimen in the top row, and fourth and fifth specimens in the bottom row
are core preparation or rejuventation spalls.
In agreement with our extensive surface surveys and conversations with

local farmers that failed to locate any sources, several lines of evidence estab-
lish that the fine-grained cryptocrystallines essential to microlithic technol-
ogy were scarce or entirely absent in the western Loess Plateau, suggesting
the technology itself is exotic. Were cryprocrystallines at all locally abundant,
for example, one would expect evidence of at least their casual use of off and
on throughout the occupation of the site. Instead, they are essentially absent
prior to the appearance of microblade technology. Of course, the most obvi-
ous evidence for this raw material scarcity is the miniaturization of Dadiwans
microblade technology: cores were saved and used well beyond the normal
point of discard. The 54 microblades and 10 microblade cores in our Dadiwan
samples are among the smallest on record anywhere in the world. With the 27
complete specimens averaging just 9.13mm in length (std. dev. = 1.56mm),
3.64mm in width (std = 0.55mm), and 0.94mm in thickness (std = 0.34mm),
these microblades are only two-thirds the size of microblades found in North
China Microlithic sites on the upper Yellow River (Elston and Brantingham
2002), and less than half the size of those from the North American Arctic
and Subarctic (Anderson 1970; Coupland 1996; McGee 1970; Sanger 1968).
6.8. Microblade cores showing all specimens recovered from site. Top row left: black chal-
cedony, 100110 cm; top row right: white chalcedony, 120130 cm; second row left: black
chalcedony, 140150 cm; second row right: black chalcedony, 140150 cm; third row left:
white chalcedony, 160170 cm; third row right: white chalcedony, 200210cm; bottom row left:
white chalcedony, 200210cm; bottom row right: quartz, 220230cm.
Similarly, the nine complete microblade cores average just 10.18 mm from
height (platform to base), 10.34mm in width (side to side), and 18.40mm in
length (front to back), less than half the size of the smallest Neolithic speci-
mens reported from deserts to the north (Chen 1984) and from the North
American Arctic and Subarctic (Sanger 1968). Diminutive to begin with, this
technology seems to have become ever smaller through time at least as mea-
sured by microblade core height, which diminishes as one moves higher in
12.0
11.0
Microblade Core Height (mm)
10.0
9.0 R 2 = 0.374
8.0
7.0
6.0
100 125 150 175 200 225
Depth Below Surface (cm)
6.9. Height (platform to base) of complete cryptocrystalline microblade cores.
the deposit (Figure6.9). Its size justifies the term suggested to us by our col-
league Richard Klein, nanolithic, referring to a formal prismatic core technol-
ogy scaled to produce blades with a mean length of less than 1cm, an order of
magnitude smaller than the upper limit for microblade length (10cm).
An alternate perspective on this raw material scarcity is provided by the rela-
tionship between the size and raw material composition of lithic assemblages.
Here assemblage size can be thought of as a proxy for stone tool demand.
Further, because microlithic technology is 98percent cryptocrystallines, flake-
and-shatter technology 95percent quartz, the material composition of lithic
assemblages (i.e., cryptocrystalline fraction) measures the relative contribu-
tion of microlithic and flake-and-shatter technology in supplying stone tool
demand. Microlithic technology was clearly the preferred means of satisfy-
ing increasing lithic demand; assemblage size is positively correlated with the
cryptocrystalline fraction (rsize-cryptocrystalline% = 0.70) and negatively correlated
with the quartz fraction (rsize-quartz% = 0.66); large assemblages are richer in
cryptocrystallines and poorer in quartz. Given this, the scarcity of the fine-
grained cryptocrystallines needed for microblade production becomes appar-
ent when the cryptocrystalline fraction is plotted against assemblage size. In
the 57 samples representing an individual 10cm level with lithics present, the
cryptocrystalline fraction increases to an asymptote, approaching but never
exceeding 60 percent (Figure 6.10). Obviously, Dadiwans knappers wanted,
but had trouble getting, more cryptocrystallines, a proposition that matches
predictions about technological intensification via refinement. Like the cur-
vilinear function depicted in Figure6.3, the sizecomposition relationship in
Figure 6.10 tracks technological refinement, lithic technology progressively
70%
60%
50%
% Cryptocrystallines
40%
30%
20%
10%
0%
0 50 100 150 200
Lithic Assemblage Size
6.10. Relationships between size and cryptocrystalline fraction of lithic assemblages.
shifting from the less costly flake-and-shatter to the more costly microblade
technology in keeping with increasing stone tool demand, that is, use time.
That the cost of microblade use was very high from the very start is indicated
by the miniaturization of this technology. The asymptote of 60percent simply
represents the point at which the costs of further reliance on microblade tech-
nology were no longer justified at existing levels of demand.
Discussion
That microblade technology appeared so abruptly at Dadiwan (Figure 6.5),
outside the main distribution of the North China Microlithic in a region
lacking the fine-grained raw material essential for its production, makes it an
almost foregone conclusion it was an exotic import, representing an influx
of North China Microlithic populations from the north, where the tech-
nology and necessary raw materials are common. That this microblade tech-
nology appears before, but becomes most prominent during, the Laoguantai
Phase 700013,000 BP further implies that these migrants (not long-term
locals) were responsible for the development of Dadiwans millet agriculture.
Whatever the function of these diminutive microlithics (use in clothing manu-
facture seems most likely;Yi et al. 2013), they were clearly considered important
enough to justify major expenditures of time and effort in acquiring the raw
material and substantial sacrifices in functional efficiency as the technology
was scaled down to suit raw material shortages. Dadiwans microblade tech-
nology was clearly more costly than versions found further north in the heart
of the North China Microlithic, making it the most costly form of the most
costly lithic technology in post-LGM North China. This makes it plain that at
least at Dadiwan, millet agriculture evolved in what was quite clearly a mar-
ginal environment, if not in terms of subsistence resources then certainly in the
availability of the raw materials required for everydaylife.
References
Anderson, D. D. 1970. Microblade Traditions Transition to Agriculture in Northwestern

in Northwestern Alaska. Arctic Anthropology China. In Late Quaternary Climate Change and
7(2):216. Human Adaptation in Arid China, edited by D.
Bar-Yosef, O. 1998. The Natufian Culture B. Madsen, F. H. Chen, and X. Gao, pp. 83101.
in the Levant, Threshold to the Origins Developments in Quaternary Science, Vol. 9.
of Agriculture. Evolutionary Anthropology Elsevier, Amsterdam.
6:159177. Bettinger, R. L., B. Winterhalder, and R.
Barnes, G. 1993. China. In People of the Stone Age, McElreath. 2006. A Simple Model of
edited by G. Burenhult, pp. 134137. Harper, Technological Intensification. Journal of
San Francisco. Archaeological Science 33(4):538545.
Barton, L. 2009. Early Food Production in Chinas Binford, L. R. 1968. Post-Pleistocene Adaptations.
Western Loess Plateau. Ph.D. dissertation, In New Perspectives in Archaeology, edited by S.
University of California, Davis. R. Binford and L. R. Binford, pp. 313341.
Barton, L., P. J. Brantingham, and D. X. Ji. Aldine, Chicago.
2007. Late Pleistocene Climate Change and Bird, D. W., R. B. Bird, and B. F. Codding. 2009.
Paleolithic Cultural Evolution in Northern In Pursuit of Mobile Prey: Martu Hunting
China: Implications from the Last Glacial Strategies and Archaeofaunal Interpretation.
Maximum. In Late Quaternary Climate Change American Antiquity 74(1):329.
and Human Adaptation in Arid China, edited by Braidwood, R. J., and B. Howe. 1960. Prehistoric
D. B. Madsen, X. Gao, and F. H. Chen, pp.105 Investigations in Iraqi Kurdistan. Studies in
128. Developments in Quaternary Science, Ancient Oriental Civilization 31. University
Vol. 9. Elsevier, Amsterdam. of Chicago Oriental Institute, Chicago.
Bettinger, R. L. 1991. Hunter-Gatherers: Broughton, J. M. 1999. Resource Depression
Archaeological and Evolutionary Theory. and Intensification during the Late Holocene,
Interdisciplinary Contributions to San Francisco Bay. University of California
Archaeology. Plenum Press, NewYork. Anthropological Records32.
Bettinger, R. L. 2009. Hunter-Gatherer Foraging: Chen, C. 1984. The Microlithic in China. Journal
Five Simple Models. Eliot Werner Publications, of Anthropological Archaeology 3:79115.
Clinton Corners,NY. Childe,V. G. 1951. Man Makes Himself. C. A.Watts
Bettinger, R. L., L. Barton, P. J. Brantingham, & Co., London.
and R. G. Elston. 2005. Report on 2004 Clark, D. L. 1968. Analytical Archaeology. Methuen,
Archaeological Fieldwork at the Dadiwan London.
Site, Shao Dian Village, Gansu Province, Coupland, G. 1996. The Early Prehistoric
PRC. Prepared for the Pacific Rim Research Occupation of Kitselas Canyon. In Early
Program. http://www.anthro.ucdavis.edu/ Human Occupation in British Columbia, edited
card/usprc/publications.htm. by R. L. Carlson and L. D. Bona, pp. 159
Bettinger, R. L., L. Barton, P. J. Richerson, R. 166. University of British Columbia Press,
Boyd, H. Wang, and W. Choi. 2007. The Vancouver.
Crawford, G. W. 2006. East Asian plant domesti- Significance in the Transition to Neolithic.
cation. In Archaeology of Asia, edited by M. T. Asian Perspectives 37(1):84112.
Stark, pp. 7795. Blackwell Studies in Global Lu, T. L. D. 1999. The Transition from Foraging
Archaeology, L. Meskell and R. A. Joyce, gen- to Farming and the Origin of Agriculture in
eral editor. Blackwell Publishing, Oxford. China. BAR International Series 774. British
Elston, R. G., and P. J. Brantingham. 2002. Archaeological Reports, Oxford.
Microlithic Technology in Northeast Asia: Madsen, D. B., R. G. Elston, R. L. Bettinger,
A Risk Minimizing Strategy of the Late C. Xu, and K. Zhong. 1996. Settlement
Pleistocene and Early Holocene. In Thinking Patterns Reflected in Assemblages from the
Small: Global Perspectives on Microlithization, Pleistocene/Holocene Transition of North
edited by R. G. Elston and S. Khun, pp. Central China. Journal of Archaeological Science
103116. Archaeological Papers of the 23:217231.
American Anthropological Association. Vol. McGee, R. 1970. A Quantitative Comparison
12. American Anthropological Association, of Dorest Culture Microblade Samples. Arctic
Washington,DC. Anthropology 7(2):8996.
Elston, R. G., C. Xu, D. B. Madsen, K. Zhong, R. Moore, A. M. T., G. C. Hillman, and A. J. Legge.
L. Bettinger, J. Li, P. J. Brantingham, H. Wang, 2000. Village on the Euphrates: From Foraging
and J. Yu. 1997. New Dates for the North to Farming at Abu Hureyra. Oxford University
China Mesolithic. Antiquity 71(274):985993. Press, London, NewYork.
Gai, P. 1985. Microlithic Industries in China. In Price, T. D., and A. B. Gebauer. 1995. New
Paleoanthropology and Paleolithic Archaeology in Perspectives on the Transition to Agriculture.
the Peoples Republic of China, edited by R. Wu In Last Hunters, First Farmers: New Perspectives
and J. W. Olsen, pp. 225241. Academic Press, on the Prehistoric Transition to Agriculture, edited
NewYork. by T. D. P. a. A. B. Gebauer, pp. 319. 1st ed.
Gansu, Sheng Wenwu Kaogu Yanjiusuo (Gansu School of American Research Advanced
Provincial Institute of Cultural Relics and Aeminar Series. School of American Research
Archaeology). 2003. Gansu Qinan dadiwan Press, Santa Fe,NM.
yizhi yangshao wenhua zaoqi juluo fajue jianbao Sanger, D. 1968. Prepared Core and Blade
(Excavation of the settlement of early Yangshao Traditions in the Pacific Northwest. Arctic
culture on the Dadiwan site in Qinan County, Anthropology 5(1):92120.
Gansu). Kaogu (Archaeology) 6:1931. Smith, B. D. 1995. The Emergence of Agriculture.
Gansu, Sheng Wenwu Kaogu Yanjiusuo (Gansu Scientific American Library, NewYork.
Provincial Institute of Cultural Relics and Ugan, A., J. Bright, and A. Rogers. 2003. When
Archaeology). 2006. Qinan Dadiwan xinshiqi Is Technology Worth the Trouble? Journal of
shi dai yizhi fa jue baogao (Dadiwan in Qinan: Archaeological Science 30(10):13151329.
Report on excavations at a Neolithic site). Cultural Yi, M., L. Barton, C. Morgan, D. Liu, F.
Relics Publishing House, Beijing. Chen, Y. Zhang, S. Pei, Ying Guan, H.
Lu, T. L.-D. Just clarifying that Lu, T. L.-D is a Wang, X. Gao, and R. L. Bettinger. 2013.
different author than Lu, T. L. D. 2006. The Microblade Technology and the Rise of Serial
Occurrence of Cereal Cultivation in China. Specialists in North-Central China. Journal of
Asian Perspectives 45(5):129158. Anthropological Archaeology 32:212223.
Lu, T. L. D. 1998. The Microblade Tradition
in China: Regional Chronologies and
Seven
When to Retouch, Haft, or Discard?

Modeling Optimal Use/Maintenance
Schedules in Lithic ToolUse
Chris Clarkson, Michael Haslam, and Clair Harris
Utilized and retouched stone flakes are found in varying proportions in all
lithic assemblages, with some artifacts exhibiting signs of hafting and exten-
sive resharpening. Various theories offer explanations for patterns in lithic
reduction, hafting, curation, and discard, including the abundance, proximity,
and opportunities to acquire replacement raw material, and past attempts to
increase functional efficiency. This chapter asks whether differences in tool
efficiency occur for unretouched, retouched, and unretouched hafted arti-
facts for scraping wood after manufacture and maintenance costs are factored
in. Because wood and plant working are common activities identified on
stone tools, and especially scrapers and flakes, in studies of past stone tool
use (Anderson-Gerfaud 1990; Beyries 1988; Dominguez-Rodrigo etal. 2001;
Hardy 2004, 2009; Robertson etal. 2009), no doubt owing to the necessity
of manufacturing wooden tools for subsistence activities and self-defense, this
research is of importance in understanding the selective pressures operating
on lithic technologies and the organization of technology throughout human
evolution.
Retouched stone artifacts, especially those thought to have been hafted,
typically take pride of place in archaeological analyses and illustrations of lithic
assemblages. Their rise in frequency at certain times in the past often serves as
a marker of industrial change and is usually interpreted as improved techno-
logical efficiency, more specialized activities, or increased cultural complexity.
The significance attributed to retouched artifacts often rests in the belief that
117
118 Clarks o n e t a l .
they were shaped to a specific design, were targeted toward and modified
by specific functions, and also formed important markers of ethnic identity
and cultural sophistication (Hiscock 1988). Retouched artifacts also certainly
required greater investment in time and labor to produce than unretouched
flakes, and hence improved efficiency might be predicted from recent evolu-
tionary models that suggest that greater investment in extractive technology
or tech time should result in higher payoffs in subsistence returns (Bright etal.
2002; Ugan etal. 2002).
The significant role of retouched and standardized lithic tools has also fea-
tured heavily in theorizing about appropriate technological responses to risk
and mobility, as seen particularly in discussions of reliable versus maintainable
toolkits (Bleed 1986; Clarkson 2007; Hiscock 2005; Kuhn 1995; Myers 1989;
Nelson 1991). Likewise, hafting a stone tool is commonly argued to increase
the efficiency and precision of the work while also allowing smaller artifacts
to be used (Keeley 1982). The likely greater reliance on hafting in the later
Paleolithic is one of the major observations to be made about changes in lithic
technology in human evolution (Clark 1968).
Some or all of these propositions concerning the important role of retouched
and hafted stone artifacts are likely to be true in certain cases. However, tech-
nological analyses of the last few decades have also shown that retouched forms
can be highly mutable and multifunctional, and retouch intensity appears to be
as responsive to raw material availability, mobility, and economic risk, as it is to
functional and ethnic concerns or greater investments in tech time to secure
higher subsistence returns (Brantingham 2003; Clarkson 2005, 2007; Dibble
1995; Gordon 1993; Hiscock and Attenbrow 2003; Morrow 1997; Neeley and
Barton 1994; Nejman and Clarkson 2008; Shott 1989).
Ethnoarchaeology has also eroded the sense that retouched artifacts are
necessarily the most desirable or significant objects in an assemblage, pro-
viding many cases (particularly in Australia and New Guinea) where unre-
touched flakes were preferentially selected and rapidly discarded with little
concern for retouching or imposing specific shapes on artifacts (Cane 1988;
Hayden 1979; Shott and Sillitoe 2005;White 1969;Wright 1977).These obser-
vations of expedient use puzzled early typologists (Hayden 1977; Hiscock
1998; Holdaway 1995; Wright 1977), and led to a view that perhaps stone tool
use in post-contact times reflected a loss of skill or knowledge about lithic
technology, or that the conditions of use were unlike those of pre-contact
times.
A number of significant usewear and residue studies have also shown that
although certain retouched tool types are traditionally seen as designed for
specific tasks, tight formfunction relationships are typically illusory or non-
existent and that past tool-users, as well as those documented in ethnographic
times, were often agnostic about the choice of artifact type for a given task
Wh en to Re t o u ch, H a ft, o r Di s ca rd ? 119
(Anderson-Gerfaud 1990; Beyries 1998; Clarkson and Connell 2011; Hardy

2004; Hayden 1979, 1985; Robertson etal. 2009; White 1969; Wright 1977).
Again, although some of these statements about a close fit between tool
efficiency and form could be true in certain cases, we set out here to show that
some of our ideas about the relative significance and functional efficiency of
stone tools, particularly retouched and hafted tools, may be unfounded.
Experiments conducted with student volunteers reveal some surprising
results in terms of the differential efficiency of flakes modified and held in dif-
ferent ways. The results of these experiments inform the way we conceive of
the principles governing the organization of lithic technology in past societies,
and make sense of stone tool use among those traditional societies observed
in recent times. In particular, the results of this study reveal that retouching
and hafting are likely to be efficient strategies only in particular contexts, and
that heavy reliance on unretouched toolkits will be the most efficient solution
in many (but not all) cases. This argument is not new, and has been a fea-
ture of theoretical statements about the organization of technology for some
decades (Nelson 1991; Parry and Kelly 1987), but empirical demonstration of
the mechanical principles that determine when to employ different strategies
offers new insight into the choices people might make about when to employ
certain strategies.
The Experiments
The experimental study models the declining rate of wood removed using a
flake in a scraping motion.The experiments entailed extended use of 15 speci-
mens: 5 unretouched, 5 retouched, and 5 hafted flakes (Table 7.1). The tools
were all used to scrape medium hard wood staves for 10,000 strokes with a
stroke length of 30cm each. The wood used was spotted gum (Eucalyptus mac-
ulate) with an air-dry density of 970kg/m3. In each case, experimenters drew
the tool toward the user while holding the edge at a steep angle to the staff.
The wooden staff was weighed after every 50 strokes of roughly 30cm length,
and the time taken to perform each 50 strokes was recorded.
Grams lost from the staff after each 50 strokes is used as the measure of
gain that is, the amount of utility gained by the user for a particular task in
a fixed period. From this we can also formulate a measure of gain rate, being
the rate of gain or loss in tool performance over a number of episodes of tool
use. Gain rate can be used to determine whether a tool improves or declines
in efficiency over time. The 10,000 stroke limit employed in this experiment
equates to an average of about 2 hours of continuous use, or 3 km of continu-
ous scraping.
All stone tools were made from high-quality Bergerac flint from France
(Figure 7.1). Flint was chosen rather than an Australian stone material to
table7.1. Details of individual specimens used in the experiment
Specimen Weight Edgeangle Final weight Final edge Total grams

(degrees) angle removed
Flake 1 78.0 48 77.76 64 562
Flake 2 84.3 50 83.78 75 910
Flake 3 84.8 42 84.68 58 1059
Flake 4 53.8 54 53.26 71 1337
Flake 5 49.8 36 49.52 75 540
Retouched 1 116.0 71.5 115.9 75 679
Retouched 2 162.2 75 161.7 83 574.6
Retouched 3 113.9 61 113.7 63 390
Retouched 4 120.4 63 119.9 64 601
Retouched 5 185.8 55 185.3 74 379
Hafted 1 186 NA NA NA 243.9
Hafted 5 210 NA NA NA 350
All specimens are made of flint. Edge angles could not be measured on hafted scrapers because of
the resin mastic interfering with the measurement.
Experimental Tools
(Bergerac Flint, France)
Retouched flake after Unretouched flake Unretouched hafted flake

10,000 strokes after 10,000 strokes after 10,000 strokes
7.1. Examples of the experimental tools used in the experiments.
maximize global relevance, though in fact any high-quality cryptocrystalline

silicate rock would likely perform in similar ways. The hafted specimens were
set in a very tough spinifex resin (Triodia sp.) from north Queensland, Australia,
and the hafts were made from pine or hardwood. All flakes were large enough
to hold in the hand with ease, and had edge angles ranging from 45 to 65
degrees.
The experiments involved people using tools to scrape hard wood, aim-
ing to do so in the most efficient way possible (i.e., adjusting angle of use
and posture to achieve the best results possible). Collins (2008) recently
performed similar wood working experiments measuring weight lost for a
given number of strokes using a highly controlled and mechanized experi-
mental design to test the differing efficiency of three different edge profiles.
Although we did not mechanize our experiment, choosing to replicate real
human motions instead, many variables were kept constant such as the spe-
cies of wood and its provenience, stroke length, raw material type, and use
action. The main disadvantage of this approach is that the specific inter-
actions between variables cannot be isolated and only the overall patterns
can be observed. The main advantage is that the actual motor habits of real
people are replicated and individual variation can be examined. Our use
of cylindrical staves meant that edge profiles tended to be less important
in our experiment than would be the case in working flat planks of wood
(cf. Collins 2008), and slightly concave, straight, and convex edges were all
capable of making sufficient contact with the worked surface. In any case,
concave edge profiles soon appeared on most tools (Figure7.1). Edge angles
for the retouched flakes were slightly higher on average (65 7 degrees)
than for unretouched flakes (46 7 degrees), as often occurs when flakes
are retouched (see Clarkson 2005; Hiscock and Attenbrow 2005). Because
flakes with the same range of edge angles as those used in the unretouched
experiments were chosen for retouching, our retouched population should
accurately capture the results of retouching those flakes.
To ensure the same portion of the edge was used throughout the experiment,
a 3cm length of edge was marked, and in some cases colored black with ink
to help locate the used portion of the edge and to monitor wear and flake scar
accumulations. This 3cm length of edge was used exclusively for the duration
of the experiment. Edge angles and edge profiles were measured at the comple-
tion of every 1000 strokes, and the 50 stroke intervals were timed and recorded
from time to time.To examine the changes to the morphology of the edge, and
their relationship with efficiency, use wear was also recorded under low mag-
nification (6.7) by the authors (M. H. and C. H.), and tool edges were pho-
tographed under low magnification on both dorsal and ventral surfaces after
each 1000 strokes. Usewear analysis consisted of counting the number of scars
and the number of step and hinge scars, as well as employing a 4 rank system
of edge rounding (0 no rounding, 1 light rounding, 2 moderate rounding,
3 heavy rounding). Edge angles could not be measured on hafted tools owing
to the mastic interfering with the measurement; however, the range of initial
edge angles was the same as that for unretouched flakes.
To ensure the drying of the wood or removal of soft bark were not major
factors driving dropping efficiency in tool use, all staves were air dried for at
least a week before use and the soft outer bark was removed before commenc-
ing the experiment.
500
Unretouched
Asymptote
400
Retouched
Gain rate (g lost)
Hafted
300
Unretouched
200
100
0 10 20 30 40 50 60 70 80 90 100110
Minutes of Use
7.2. Experimental results showing the asymptotic nature of the declining gain curve over 10,000
strokes for all three experimental tool types. Times were calculated using an average of 1.3
strokes per second (45 strokes per minute). Asymptotes were not calculated.
Results
The first experimental finding in this study is that declines in tool efficiency
over the long term are asymptotic for all three kinds of tool use (Figure7.2).
This means that although all tools initially had a high rate of wood removal
after each 50 strokes, the rate of removal declined through time, but at no time
ceased to function entirely. The only exception was for hafted scrapers when
the resin hafting broke and the tool critically failed. This rarely happened dur-
ing scraping and the tools were rehafted and the experiment continued. The
asymptote in Figure7.2 is shown by the sloping line that comes close to inter-
secting the gain curve. Because the rate of gain continues to decline over the
course of the experiment, and perhaps infinitely, the asymptote never actually
intersects the curve, but gets closer with each observation.
An asymptote for wood working lithic tools means that scrapers could
in theory continue to be used to the point were virtually no wood could
be removed any longer, yet the tool could still be considered functional
(albeit very inefficient), because some very small amount of wood could still
be removed. This is important for modeling tool performance for scrapers
because frequent critical failures (i.e., tool breakage) would drastically change
the nature of tool use and replacement. For the purposes of this study, an
asymptotic decline in tool efficiency means that decisions must be made by
the user about when a tool is no longer functional and should be replaced,
rather than the tool suddenly failing and necessitating replacement for the
activity to continue. We are not suggesting that tools of this kind will never
fail, only that this kind of activity is more likely to have to lead to decisions
about retooling, and this makes modeling the tradeoffs between retooling and
maintenance costs worthwhile.
300 Experiment
Unretouc
Hafted U
95% CI Cumulative Weight Loss (g)
250 Retouche
Unretouc
Hafted U
200 Retouche
150
100
50
200 400 600 800 1000 1200 1400 1600 1800 2000
Number of Strokes
7.3. Confidence intervals for gain rate for each tool type over the first 2000 strokes.
Our results closely mirror those of Collins (2008); however, she used only
unretouched flakes and continued her experiments only for 2400 strokes too
soon to establish the asymptotic relationship. Nevertheless, the similarity of our
results gives us confidence in the merits of our approach given that Collins
experiments were mechanized and highly controlled.
The second and perhaps most significant finding is that retouched flakes
are in fact far less efficient than unretouched flakes, both at the outset and
over the longer use-life of the tool. This is still more surprising because the
retouched artifacts had higher edge angles than unretouched flakes, a con-
dition that is usually seen as increasing the efficiency of hard wood scrap-
ing (Wilmsen 1968:159). This is demonstrated by the initial steeper rate of
gain for unretouched flakes and the steeper asymptote, meaning that even as
gain approaches zero, it does so at a slower rate than for retouched or hafted
scrapers. The differences in slope shown in Figure7.2 are also supported by
the average differences in total wood removed by each tool type after 10,000
strokes. Unretouched flakes removed an average of 881.6 g, retouched flakes an
average of 524.7 g, and hafted scraper an average of 267.3g.
Finally, hafted tools are found to be far less efficient than unhafted tools,
and are much less efficient than unretouched flakes. This is a very surprising
conclusion given that increased leverage and grip strength afforded by a solid
handle should increase the force that can be exerted on the worked material
and the precision with which it can be applied (Keeley 1982). Hafted tools
have a much flatter rate of gain as well as a flatter asymptote, meaning that less
material is removed at early and late stages of thework.
Figure7.3 shows the confidence intervals for each tool type over the first
2000 strokes. The differences in efficiency are most pronounced in the first
Number of Strokes
200
400
12.50 600
800
1000
1200
10.00
Mean mass removed (g)
7.50
5.00
2.50
0.00
Hafted Unretouched Retouched Unretouched
Experiment
7.4. Relative performance declines for each tool type at 200-stroke intervals.
1000 strokes and then begin to flatten off. We can see from relative perfor-
mance declines that retouched and hafted scrapers do slightly better at first
(Figure7.4) but that both kinds of tools dull more quickly than unretouched
flakes. Unretouched flakes therefore maintain higher levels of performance
over longer periods than either retouched or hafted tools.
Modeling Optimality in Scraper Use and Discard

It is possible predict the point at which maximum efficiency is reached for
each type of tool, and hence the point at which it should either be replaced or
rejuvenated by use of a model derived from the marginal value theorem and
central place foraging models of Charnov (1976) and Orians and Pearson
(1979) that are used extensively in evolutionary ecological studies. The same
model has been used in archaeology to examine cases of field processing in
which resources are located in a different place to where they were to be con-
sumed (Beck etal. 2002; Bettinger etal. 1997; Jones and Madsen 1989; Metcalfe
and Barlow 1992; Rhode 1990).
The model as presented here factors in manufacturing time for the arti-
fact, such that any continued use of the tool would result in declining yields
when compared to the cost of procuring a new tool. The point of optimal
tool replacement/rejuvenation can be derived by fitting a tangent to the gain
curve, in this case a diminishing gain curve over time (Figure7.5). The model
t1
en
Manufacture Time Use Time
ng
Ta
t2
en
ng
Ta Diminishing
Gain Curve
Optimal
Point of
Discard for
Tangent 2
Optimal
Point of
Discard for
Tangent 1
Time
Long Manufacture Time Short Manufacture Time

7.5. Model showing the effect of different manufacturing time (T) on overall gain rate. Factoring
in manufacturing time (or tech time), enables the point at which maximum productivity has
been reached (m0), such that continuing to use the tool would result in declining yields when
compared to the cost of procuring a newtool.
takes into account both manufacturing time (to the left of the perpendicu-
lar line) and use time of the tool (to the right of the line). The longer the
manufacturing time, the lower the angle of the tangent will be, as shown in
Figure7.5, and thus the longer the tool should be used to recoup the costs of
initial manufacture. Hence, the tool with the short manufacture time inter-
sects the gain curve much earlier in time, and while the tool is still operating
at a higher rate of productivity (Tangent 1)than the tool with a much longer
manufacturetime.
In the case of scraper use, the model shown in Figure7.5 indicates that a tool
with a shorter manufacturing time and higher gain rate should be discarded
more frequently, while at the same time, a shorter manufacture time and higher
gain rate will mean the use of a more efficient technology overall. The bigger
the fall off in gain rate (i.e., the more curvature in the gain curve), the more
frequently tools should be replaced, because greater losses will be sustained by
continuing to use the tool rather than procuring a newone.
The model predicts that unretouched flakes, which in our study all had a
very short manufacture time (average 2 minutes) and a rapid rate of increase
as well as a rapid decline in gain rate, are most efficient for the first 270 strokes
(compare also with Collins (2008) 200 strokes for peak efficiency), or around 6
minutes of use (Figure7.6). Using timed activity data, we know that retouched
flakes take longer to make than unretouched flakes (an average of 4 minutes)
Manufacture Time Use Time Unretouched

& Unhafted Retouched
300
250
Gain Curve (g lost)

200
150 Hafted & Unretouched
100
50
0 10 20 30 40 50 60
tes
inutes
tes
0 s es)
s)
)
es
ke
nu
2 minu
ok
ok
tro
tes 0 str
str
mi
4m
20
50
(27
(45
(13
es
tes
t
inu
nu
nu
mi
6m
mi
10
30
7.6. Model predictions for when to discard each tool type given different known manufactur-
ing times.
and have a lower rate of gain, meaning optimal use is reached at around 10
minutes, or 450 strokes. At this point it is more efficient to procure another
tool, or resharpen the tool than continue using it. Hafted tools obviously take
the most time to make, as resin must be heated, shaped, and cooled before use,
and the flake may also need shaping to fit the haft. This longer manufacture
time (20 minutes), and lower gain rate, means that hafted tools are the least
efficient and should be used for around 30 minutes. Tools should be used for
this longer period to recoup the greater costs of manufacture, because dis-
carding a tool soon after manufacture would mean spending significant time
manufacturing another tool that could have been spent using the previous tool
with no loss in overall efficiency. This situation invokes the concept of sunk
costs, whereby having invested significant time and energy in making a tool,
it is worth accepting this as a start-up cost and continuing to use the tool for
as long as possible rather than make a new one or attempt to procure a more
efficient solution (such as obtaining fresh flakes).
The long manufacture time partly drives down efficiency in hafted scrapers,
but the lower gain rate also requires explanation.We suggest that the extra force
exerted on the edges of hafted flakes causes them to crumble more quickly,
reducing efficiency. This proposition is examined further in this chapter.
ng
ni
pe
g
ing
ar
nin
600
sh
ool
r pe
Re
Ret
ha
&
ng
s
500
Re
fti
ha
Re
Gain (Weight Lost (g))
400
After 30 minutes use:
300 Retooling = 500 g

Resharpening = 350 g
200 Rehafting &
Resharpening = 200 g
100
Manufacturing 0 10 20 30 60 120
Time Minutes of Use
7.7.The effects of maintenance time as well as manufacturing time on gain rate and overall
efficiency as represented by the slope of the tangent.
We can take this modeling approach one step further by examining the
effects of maintenance time on tool performance as well. Here the time taken
to retool with a fresh flake, in the case of unretouched flakes, or resharpen in
the case of retouched flakes, or to rehaft and resharpen in the case of hafted
flakes, is factored into the model as flat spots representing the time taken to
maintain the tool. A short initial manufacture time of only 10 minutes for
the hafted tool was employed here (based on activity data from Australian
Aboriginal examples of stone tool use recorded by Hayden in 1979), as in most
cases people in Haydens study tended to reuse a haft rather than make a new
one each time. Maintenance episodes were repeated five times for each tool
type. After each maintenance episode the tangent was refitted to the cumula-
tive gain curve to determine the point at which the next retooling, resharpen-
ing, or rehafting episode should take place.
Figure7.7 indicates that tool efficiency is heavily affected by retooling and
resharpening, with the differences between overall gain (in terms of wood
removed from a hardwood shaft) differing by up to 300 g after half an hour of
work. Importantly, the overall slope of the line fitted to the gain curve for both
unretouched and hafted unretouched flakes is to the right of the gain curve,
meaning that continued use of this strategy actually results in increasing effi-
ciency, because continued retooling or rehafting will continue to increase the
rate of gain (i.e., push the line to the left) and hence overall efficiency of the
tool.This makes sense particularly for the hafted scrapers given that continuing
to use the tool will recoup the costs of making the haft and will improve the
efficiency of the tool.The line fitted to the retouched gain curve, on the other
hand, intersects to the right of the gain curve, meaning that this strategy will
continue to become less efficient with time as continued use will continue to
push the tangent to right, hence lowering overall gainrate.
Given that Australasian ethnographic accounts (one of the few such accounts
we have, see also Gould 1980; Sillitoe 1988) indicate that wood working with
stone tools to make spears, bowls, clubs, and hafts can take between 2 and 20
hours (Hayden 1979), such differences in efficiency would have huge effects
on overall work time. According to the model shown in Figure 7.6, hafted
scrapers should be replaced roughly every 30 minutes (determined from the
intersection of the tangent with the gain curve in Figure7.6), whereas unre-
touched flakes should be replaced every 6 minutes or so. An hour of work
would therefore mean the most efficient strategy in terms of time (i.e., using
fresh unretouched flakes) would also be the least efficient in terms of raw
material use, consuming nearly 12 times the number of flakes as if the same
flake were constantly resharpened!
Ethnographic Comparisons
If we compare the results of this experiment to ethnographic data, we find
a close fit between model predictions and some real examples of stone tool
use-life. Examining Haydens (1980) data for tool use in the Australian Central
Desert, for instance, we see that his informants chose to discard their unre-
touched flakes after an average of about 6 5 minutes of use.This corresponds
very well with the predictions made in this study for optimal retooling after
about 6 minutes. Furthermore, Hayden found that people retouched their
hardwood scrapers after 6.9 11 minutes on average, again showing that
tool efficiency dropped noticeably around this time. Most importantly, hafted
retouched tools were used and resharpened for a mean overall use-life of
24.7 22.9 minutes, suggesting that people greatly extended the use-life of
these tools rather than incur the costs of rehafting.
Shott and Sillitoes (2004, 2005) recent comparison of Wola unretouched
flake use-life in the highlands of PNG and hafted end scraper (flake shaver)
use-life in an Upper Palaeolithic site using survivorship profiles also sits well
with this model. The unretouched flakes were all discarded early in the overall
potential use-life of the artifact (around 611 minutes), while the hafted end
scrapers were more likely to be completely used up (i.e., retouched to the
point of exhaustion). That pattern can be explained as more frequent retool-
ing to maintain high efficiency for low manufacture and maintenance costs, in
contrast with prolonged use and curation of expensive hafted technologies to
recoup large manufacture and maintenance costs.
Data of a similar kind also exist for Ethiopian hide scrapers. These data
show that resharpening was very frequent for hide scrapers (between 100
and 300 strokes) (Gallagher, 1977; Weedman 2000). However, because skin
scraping is a very different activity with likely different rates of attrition of
tool edges and a different use action, these data are unlikely to be directly
comparable to our own experiments. Future experiments using a range of
flake types to work hides will explore whether similar relationships hold for
different activities.
Usewear Patterns
The use wear data obtained during this study offer substantiation of and expla-
nation for dropping rates of efficiency in stone tool use. Figure7.8 shows both
the gradual mean cumulative weight of wood removed from the staves (square
symbols) as well as the mean cumulative weight of stone lost from the edge of
the unretouched flakes (diamond symbols). These two curves closely mirror
one another and show that increasing wear on the tools edge directly affects
the rate of gain for scrapingwood.
The type of damage to the edge is also important in determining the loss
in efficiency to the tool edge. Figure7.9 plots the mean cumulative accumu-
lation of step terminated scars (square symbols) and edge rounding (diamond
symbols) on the edges of the unretouched flakes. The two are also closely
correlated, showing that step terminated scars increase first, stabilizing edge
loss, followed by an increase in edge rounding. Both stepped scars and edge
rounding increase together dramatically in the final few thousand strokes. In
other words, as edges crush and become stabilized rather than continue to
chip away, the stabilized edges begin to round, further reducing the efficiency
of the tool. The same patterns are seen for retouched and hafted tools, but are
not presented here. Retouched and unretouched tools also show substantial
increases in edge angle as a result of edge attrition (Figure7.10). Edge angles
could not be measured on hafted scrapers owing to the resin mastic interfering
with the measurement.
Finally, our usewear data may shed light on the reason for the poor per-
formance of the hafted unretouched scrapers as against the unhafted ones,
given hafting should increase leverage and force. Figure7.11 shows the rates of
increase in edge rounding and step terminated scars on hafted and unhafted
unretouched flakes. The graph on the left shows that edge rounding is slower
to form on hafted scrapers, whereas the graph on the right shows that step
scarring accumulates more quickly on hafted scrapers, as the edge crumbles
more quickly, perhaps due to the increased force that can be brought to bear
on the hafted tool. As the edge is crumbling quickly over the 4000 strokes,
edge rounding is unable to form, whereas it is able to form on unhafted
450 0.4
400 0.35
Weight of World Removed 350
Weight Lost from Tool

0.3
300
0.25
250
0.2
200
0.15
150
100 0.1
50 0.05
0 0
00
00
00
00
00
00
00
00
00
0
00
10
20
30
40
50
60
70
80
90
10
Number of Strokes
Cumulative Weight of Wood Removed

Cumulative Weight Lost from Tool
7.8. Mean cumulative weight of wood removed per 1000 strokes (left y-axis), and mean cumu-
lative weight lost from unretouched flakes per 1000 strokes (right y-axis).
35 2
1.8
Number of Step Terminated Scars
30
1.6
Edge Rounding Rank
25 1.4
20 1.2
1
15
0.8
10 0.6
0.4
5
0.2
0 0
00
00
00
00
00
00
00
00
00
00
10
20
30
40
50
60
70
80
90
,0
10
Number of Strokes
Steps
Edge Rounding
7.9. Mean cumulative rate of increase in step terminated scars for the 3 cm used edge (left
y-axis) and mean edge rounding rank for the utilized edge (right y-axis).
scrapers. Stepped scarring drops off on hafted scrapers after 4000 strokes and
edge rounding beings to climb steeply.These data suggest that the greater force
exerted on hafted tools causes them to fail more quickly and hence to have a
lower rate of gain when compared to unretouched flakes.
30
25
Average edge angle increase
20
15
10
0
00
00
00
00
00
00
00
00
00
00
10
20
30
40
50
60
70
80
90
,0
10
Number of Strokes
Retouched
Unretouched
7.10. Average increases in edge angle (in degrees) over the course of the experiment for
retouched and unretouched edges.
Discussion
The question arises that if retouch and hafting are so inefficient, and techni-
cal systems concerned with subsistence returns such as making weapons and
domestic tools of wood are likely to be under heavy selection, why do it at all?
The model here predicts that people should retouch only if they have insuf-
ficient raw material to resupply themselves constantly with fresh sharp flakes,
remembering that maintaining the most efficient use of unretouched flakes
means most likely consuming around 12 times the amount of raw material per
hour. It probably also makes sense to haft if you cannot hold the flake effec-
tively in your hand. All the flakes in our experiments were of large enough size
to be easily held in the hand. It would be worthwhile testing whether using
very small flakes results in big declines in efficiency that might be offset by
hafting. It is expected that this would be thecase.
There is no doubt that retouching an artifact once it becomes dulled
increases its efficiency, at least temporarily. If new flakes are not available, then
retouching an implement makes sense, even though a retouched flake always
performs more poorly than a fresh flake. Retouched flakes also dull faster than
unretouched flakes, and the option to retouch means one must continue to
2.5
Edge Rounding Rank 2
1.5
0.5
0
00
00
00
00
00
00
00
00
00
00
10
20
30
40
50
60
70
80
90
,0
10
Number of Strokes
35
Number of Step Terminated Scars
30
25
20
15
10
0
00
00
00
00
00
00
00
00
00
00
10
20
30
40
50
60
70
80
90
,0
10
Number of Strokes
7.11. Comparison of edge rounding (upper) and stepped scar formation (lower) on unhafted
(broken line) and hafted (solid line) unretouched scrapers.
retouch frequently to maintain high efficiency levels. Our experiments also

revealed that retouching an artifact attached to a handle with very strong spi-
nifex resin almost always resulted in damage to the hafting. Hayden (1979) and
others have made the same observation. Retouching a hafted scraper in the
haft therefore may also entail undesirable rehafting costs, further increasing
maintenance time and driving down tool efficiency and further increasing the
value of prolonging the life of thetool.
There are of course many situations in which hafting is vital or dramati-
cally changes what is possible with stone tools, but scraping wood when large
unretouched flakes are available is probably not one of them. Drill technology,
projectiles, and very fine engraving work involving lithic tool bits all likely
require hafting to functionwell.
Our results therefore suggest that retouching is likely to be an important
strategy when raw material is scarce or resupply is unpredictable, and that use
of unretouched flakes will be the most efficient strategy when raw materials
are locally abundant or at least consistently restocked. Uncertainty over raw
material supply may pertain when raw material is simply rare in the landscape,
or when people are highly mobile and cannot carry large quantities of raw
material around with them. In other words, increasing mobility and uncer-
tainty over opportunities to reprovision should mean that people must curate
the toolkits they carry with them, and as flake tools dull quickly, resharpening
is the best option for maintaining efficiency. Because toolkits often need to be
small and portable during periods of high mobility, hafting would be an effec-
tive means of employing small tools, and of transporting them (Kuhn 1995).
Conserving tools, however, means accepting drops in efficiency.
Alternatively, we might also predict that when raw material supply is lim-
ited or unpredictable, the continued manufacture of small-sized flakes from
cores and larger flakes will also be an efficient strategy, rather than extensively
retouching scrapers. Such small artifacts appear to predominate at certain times
and places, such as in the Mousterian (Dibble and McPherron 2006), and in
assemblages dominated by small bipolar artifacts (Hiscock 1996). We might
also expect to find more signs of hafting given that the loss in efficiency that
stems from greater difficulty in gripping the tool may be compensated for by
investing greater time in manufacturing and maintaining a haft. The conclu-
sions drawn from this study suggest that greater efficiency in wood working
is obtained from frequent retooling with fresh sharp flakes. If these can be
obtained only by removing usable flakes from small cores and flakes, then this
may be the most efficient strategy for maintaining a supply of highly effective
tools when raw material supply is limited. Such an approach might explain the
preponderance of technologies such as truncated faceted and kombewa flake
reduction in the Mousterian and Oldowan (Dibble and McPherron 2007), and
may justify the sometimes extreme reduction of small freehand and bipolar
cores.
This rapid depletion of tools and the rapid drops in efficiency that accrue
from conserving them when equipped only with small hafted retouched tools
also provide a good reason to schedule heavy woodworking tasks to periods of
down-time in base camps that are stocked with raw materials (Binford 1980;
Torrence 1983, 1989). This suggests that provisioning frequently and predict-
ably used places with stockpiled raw materials would be an effective strategy
for increasing the efficiency of wooden implement manufacture and main-
tenance within high-mobility land use systems. This is exactly what land use
and provisioning models predict (Kuhn 1995; Nelson 1991; Parry and Kelly
1987), based on observations of hunter-gatherer behavior and archaeological
assemblage variability.
Finally, shifts from so-called expedient flake assemblages to highly retouched
and curated ones would be expected on the basis of results obtained in this
study to correspond to increasing uncertainty about raw material supply (or

changes in task that required specialized hafted tools), such that small tools had
to be curated during more frequent periods of high mobility and uncertainty
about opportunities to reprovision. In short, maximizing efficiency through
the use of many unretouched flakes for short periods would be increasingly
sacrificed for the security of a portable supply of small hafted, but less efficient,
tools.
This work may also have implications for our understanding of scraper typol-
ogy and assemblage formation. It would seem that initially retouching a flake
for hard wood scraping is a bad idea, if the edge is already suited to the work.
Although there may be cases in which retouching an edge to say, change the
edge angle, may provide benefits above those obtained by using a fresh sharp
edge, our experimental results would seem to suggest that in most cases scrap-
ing hard woods should begin with unretouched edges and proceed to retouch-
ing only when raw material must be conserved. The notion that people might
retouch a flake to turn it into a scraper before use would therefore seem con-
trary to efficient tool use, and would also go against ethnographic observations
of people starting out wood working with a fresh unretouched flake.This would
suggest that reduction continuums in scrapers should begin with unretouched
flakes, and that unused portions of the edge might be utilized before beginning
retouching the edge. This is an easily tested argument and would bear consid-
eration for future examination of the relationship between tool reduction and
function (see Connell and Clarkson 2009). However, we also note that heavy
use wear on our scrapers often resembled light retouch, and discriminating
between retouch and use wear may not be easy, even microscopically.
Conclusion
Our experiments would suggest that correctly tailoring use-maintenance
schedules was likely an important issue in prehistoric economies, and should be
an important concern for a wide range of subsistence technologies. As Frison
commented after performing experiments using composite spears armed with
Clovis points to inflict lethal wounds on freshly culled African elephants:
raw-material procurement, manufacture and maintenance of weaponry...
are more time consuming than most investigators realize, but their impor-
tance cannot be minimized in hunting societies. Failure of Clovis hunters
to maintain weaponry in top condition would have negatively affected
not only the economic process but would have increased the probabilities
of self injury and/or death.(Frison 1989:783)
This statement captures two important points made in this chapter: (1) that
tech time can be considerable, and if not properly managed can adversely
affect the efficiency of resource procurement, and (2) that failing to properly
maintain tools can potentially have dire effects in risky situations and may lead
to decisions to forgo a certain amount of efficiency for increased reliability
(e.g., retouching, hafting, overdesigning, use of redundant parts, etc.) (Bleed
1986). Both statements should be true for subsistence technologies of all kinds,
including processing technologies such as grinding stones; primary extrac-
tion tools such as digging sticks, spears, and traps; as well as those tools used
to make extractive tools. Even manufacturing, hafting, and resharpening the
simplest of tools stone wood scrapers can be considerable, as demonstrated
by this study, and archaeologists should begin to factor such considerations
into reconstructions and explanations of past decisions about whether to haft,
retouch, or retool, as this may have major implications for the choice of strat-
egies at different times in different places in thepast.
The main conclusion this study has reached is that prehistoric tool users
should in many cases have retouched their woodworking toolkits only when
replacement material was scarce and/or unpredictable or when manufacturing
costs were high (e.g., hafting). The exceptions would be in cases in which the
task could not be carried out except by hafting (e.g., drills, projectiles, delicate
engraving, adzing, etc.). Hafting probably offered a solution to transporting
small tools and making them effective but offered few other advantages, at
least for wood scrapers used in this study. This supports existing theoretical
notions about the optimal organization of technology. Increasing mobility may
therefore provide a better explanation for the transition to small, retouched,
and hafted toolkits in many contexts than other explanations. We can predict,
therefore, that when replacement raw material is available and need not be
conserved, we should find many minimally used and discarded unretouched
flakes and few retouched flakes. When raw material conservation is a priority,
we should expect to see many retouched flakes, with the degree of use-life
dependent on the severity of raw material restriction. We should rarely expect
to see discarded hafted unretouched flakes in any context, unless they are for
very specific functions or very small in size, as this would be the least effi-
cient form of tool use of all. We should also expect to see signs that people
have made use of most or all of the useable unretouched portions of a tool
edge before proceeding to retouch the artifact. Because retouching will remove
prior signs of use in many cases, specimens would have to be carefully chosen
to test this hypothesis.
Although the power to generalize from limited wood working experi-
ments of this kind has its limits, archaeological investigations of the relation-
ship among manufacture, maintenance, and use have enormous potential to
develop these hypotheses further and test them against real assemblages. Sadly,
the unretouched component of most assemblages is rarely examined for signs
of use, and it may be difficult at present to determine the extent to which
past tool users made decisions about whether to replace or extend their sup-
ply of tools. Connell and Clarksons (2011) recent analysis of scraper use in
northern Australia demonstrates that past tool users were keenly aware of the
functional proclivities of their tools, and often adjusted task associations to fit
the changing nature of the tool edge as resharpening continued. If subtle dif-
ferences in the efficiency of scraper edges of different kinds for different tasks
could be detected by past tool users, then it is likely that past foragers also
made calculated decisions about toolkit design and use-maintenance sched-
ules. Examination of the relationship among raw material procurement, tool
size, mobility, and reduction intensity should therefore continue to play a fun-
damental role in understanding the dynamics of past tool use and provisioning
strategies.
Acknowledgments
We are grateful to all those who have participated in rocks and sticks over
the years, especially Angelo Bellas, Elena Piotti, Joe McCullen, James Smith,
Angela Spitzer, Kate Connell, and Lorien Perchard. This chapter benefited
from discussions with Richard Fullagar, Kate Connell, and Ian Clarkson. Our
thanks also to the Brisbane Fire Brigade for rushing to put a stop to our
spinifex heating experiments twice! Michael Haslam would like to thank
the European Research Council grant (PRIMARCH, grant no. 283959) for
funding his research.
References
Anderson-Gerfaud, P. 1990.Aspects of Behaviour by H. Dibble and A. M. Montet-White, pp.

in the Middle Paleolithic: Functional Analysis 213223. University of Pennsylvania Museum,
of Stone Tools from Southwest France. In The Pennsylvania.
Emergence of Modern Humans: An Archaeological Binford, Lewis R. 1980.Willow Smoke and Dogs
Perspective, edited by P. Mellars, pp. 389418. Tails: Hunter-gatherer Settlement Systems
Edinburgh University Press, Edinburgh. and Archaeological Site Formation. American
Beck, Charlotte, Amanda K. Taylor, George T. Antiquity 45:420.
Jones, Cynthia M. Fadem, Caitlyn R. Cook, Bleed, Peter. 1986. The Optimal Design of
and Sara A. Millward. 2002. Rocks are Heavy: Hunting Weapons: Maintainability or
Transport Costs and Paleoarchaic Quarry Reliability. American Antiquity 51:737747.
Behaviour in the Great Basin. Journal of Brantingham, P. Jeffrey. 2003. A Neutral Model of
Anthropological Archaeology 21, 481507. Stone Raw Material Procurement. American
Bettinger, Robert L., R. Mahli, and H. McCarthy. Antiquity 68:487509.
1997. Central Place Models of Acorn and Bright, Jason, Andrew Ugan, and Lori Hunskar.
Mussel Processing. Journal of Archaeological 2002. The Effect of Handling Time on
Science 24:887899. Subsistence Technology. World Archaeology
Beyries, S. 1988. Functional Variability of Lithic 34:164181.
Sets in the Middle Paleolithic. In Upper Cane, Scott B. 1988. Written on Stone: A
Pleistocene Prehistory of Western Eurasia, edited Discussion on Ethnographic and Aboriginal
Perspection of Stone Tools. In Archaeology with pp. 7590. Cambridge Scholars Publications,
Ethnography: An Australian Perspective, edited Cambridge.
by Betty Meehan and Rhys Jones, pp. 8893. Dominguez-Rodrigo, M., J. Serrallongo, J.
Department of Prehistory, Australian National Juan-Tresserras, L. Alcala, and L. Luque. 2001.
University, Canberra. Woodworking Activities by Early Humans: A
Charnov, E. L. 1976. Optimal Foraging: The Plant Residue Analysis on Acheulian Stone
Marginal Value Theorem. Theoretical Population Tools from Peninj (Tanzania). Journal of Human
Biology 9:129136. Evolution 40:289299.
Clark, Graham. 1968. World Prehistory: A New Gallagher, J. P. 1977. Contemporary Stone Tools
Outline. Cambridge: Cambridge University in Ethiopia: Implications for Archaeology.
Press. Journal of Field Archaeology 4:407414.
Clarkson, Chris. 2005. Tenuous Types Scraper Gordon, David. 1993. Mousterian Tool Selection,
Reduction Continuums in the Eastern Reduction, and Discard at Ghar, Israel. Journal
Victoria River Region, Northern Territory. of Field Archaeology 20:205218.
In Lithics Down Under: Australian Approaches Gould, Richard A. 1980. Living Archaeology.
to Lithic Reduction, Use and Classification, edited Cambridge University Press, Cambridge.
by Chris Clarkson and Lara Lamb, pp. 2134. Hardy, Bruce. 2004. Neanderthal Behaviour and
British Archaeological Reports International Stone Tool Function at the Middle Palaeolithic
Monograph Series S1408. Archaeopress, Site of La Quina. Antiquity 78:547565.
Oxford. Hardy, Bruce. 2009. Mesolithic Stone Tool
Clarkson, Chris. 2007. Lithics in the Land of the Function and Site Types in Northern
Lightning Brothers: 15,000 Years of Technological Bohemia, Czech Republic. In Archaeological
and Cultural Change in Wardaman Country, Science Under a Microscope: Studies in Residue
Northern Territory. Terra Australis No. 25. ANU and Ancient DNA Analysis in Honour of Thomas
E-Press, Canberra. H. Loy. Terra Australis 30, edited by Michael
Collins, Sophie. 2008. Experimental Investigations Haslam, Gail Robertson, Alison Crowther,
into Edge Performance and Its Implications Sue Nugent, and Luke Kirkwood, pp. 159
for Stone Artefact Reduction Modeling, 174. ANU E Press, Canberra.
Journal of Archaeological Science 35:21642170. Hayden, Brian. 1977. Stone Tool Function in
Connell, Kate, and Chris Clarkson. 2011. the Western Desert. In Stone Tools as Cultural
Scraper Reduction Continuums and Markers: Change Evolution and Complexity,
Efficient Tool Use: A Functional Analysis of edited by Richard V. S. Wright, pp. 178188.
Scrapers at Different Stages of Reduction. Humanities Press, Atlantic Highlands,NJ.
In Keeping Your Edge: Recent Approaches to the Hayden, Brian. 1979. Paleolithic Reflections: Lithic
Organization of Stone Artefact Technology, edited Technology and Ethnographic Excavations among
by Ben Marwick and Alex Mackay, pp.4556. Australian Aborigines. Australian Institute of
Archaeopress, Oxford. Aboriginal Studies, Canberra.
Dibble, Harold. 1995. Middle Paleolithic Scraper Hiscock, Peter. 1996. Mobility and Technology
Reduction: Background, Clarification, and in the Kakadu Coastal Wetlands. Bulletin of the
Review of Evidence to Date. Journal of Indo-Pacific Prehistory Association 15:151157.
Archaeological Method and Theory 2:299368. Hiscock, Peter. 1998. Revitalising Artefact
Dibble, Harold L., and Shannon P. McPherron. Analysis. In Archaeology of Aboriginal Australia:
2006. The Missing Mousterian. Current A Reader, edited by Tim Murray, pp. 257265.
Anthropology 47:777803. Allen and Unwin, St Leonards.
Dibble, Harold L., and Shannon P. McPherron. Hiscock, Peter. 2005. Blunt and to the Point:
2007. Truncated-faceted Pieces: Hafting Changing Technological Strategies in Holocene
Modification, Retouch, or Cores? In Tool v. Australia. In Archaeology in Oceania: Australia and
Core: New Approaches in the Analysis of Stone Tool the Pacific Islands, edited by Ian Lilley, pp. 6995.
Assemblages edited by Shannon P. McPherron, Blackwell, Oxford.
Hiscock, Peter, and Val Attenbrow. 2005. Australias Rhode, David. 1990. On Transportation Costs of
Eastern Regional Sequence Revisited:Technology and Great Basin Resources: An Assessment of the
Change at Capertee 3. Archaeopress, Oxford. Jones-Madsen Model. Current Anthropology
Holdaway, Simon. 1995. Stone Artefacts and the 31:413419.
Transition. Antiquity 69:784797. Robertson, Gail, Val Attenbrow, and Peter
Jones, Kevin T., and David B. Madsen. Hiscock. 2009. Multiple Uses for Australian
1989. Calculating the Cost of Resource Backed Artefacts. Antiquity 83:296308.
Transportation: A Great Basin Example. Shott, Michael J. 1989. On Tool-class Use
Current Anthropology 30:529534. Lives and the Formation of Archaeological
Keeley, Laurence H. 1982. Hafting and Retooling: Assemblages. American Antiquity 54:930.
Effects on the Archaeological Record. Shott,Michael J.,and Paul Sillitoe.2004.Modeling
American Antiquity 47:798809. Use-life Distributions in Archaeology Using
Kuhn, Steven L. 1995. Mousterian Lithic Technology. New Guinea Wola Ethnographic Data.
Princeton University Press, Princeton,NJ. American Antiquity 69:339355.
Metcalfe, Duncan, and K. Renee Barlow. 1992. Shott, Michael J., and Paul Sillitoe. 2005. Use Life
A Model for Exploring the Optimal Trade- and Curation in New Guinea Experimental
off between Field Processing and Transport. Used Flakes. Journal of Archaeological Science
American Anthropologist 94:340356. 32:653663.
Morrow,Tony A. 1997. End Scraper Morphology Sillitoe, Paul. 1988. Made in Niugini: Technology
and Use-life: An Approach for Studying in the Highlands of Papua New Guinea. British
Paleoindian Lithic Technology and Mobility. Museum Press, London.
Lithic Technology 22:5169. Torrence, Robin. 1983. Time Budgeting and
Myers, Andrew. 1989. Reliable and Maintainable Hunter-Gatherer Technology. In Hunter-
Technological Strategies in the Mesolithic of Gatherer Economy in Prehistory edited by Geoff
Mainland Britain. In Time, Energy and Stone Bailey, pp. 1122. Cambridge University Press,
Tools, edited by Robin Torrence, pp. 7891. Cambridge.
Cambridge University Press, Cambridge. Torrence, Robin. 1989. Re-tooling: Towards a
Neeley, Michael P. and C. Michael Barton. Behavioral Theory of Stone Tools. In Time,
1994. A New Approach to Interpreting Late Energy and StoneTools edited by R.Torrence, pp.
Pleistocene Microlith Industries in Southwest 5766. University of Cambridge, Cambridge.
Asia. Antiquity 68:275288. Ugan, Andrew, Jason Bright, and Alan Rogers.
Nejman, Ladislav, and Chris Clarkson. 2008. 2003.When Is Technology Worth the Trouble?
Flake Reduction in the Late Middle and Early Journal of Archaeological Science 30:13151329.
Upper Palaeolithic Assemblages of Central Weedman, Kathryn J. 2002. On the Spur of the
Europe. Lithic Technology 33:1633. Moment: Effects of Age and Experience on
Nelson, Margaret C. 1991. The Study of Hafted Stone Scraper Morphology. American
Technological Organization. Archaeological Antiquity 67:731744.
Method and Theory 3:57100. White, Carmel, and Nicholas Peterson. 1969.
Orians, Gordon H., and N. E. Pearson. 1979. Ethnographic Interpretation of the Prehistory
On the Theory of Central Place Foraging. of Western Arnhem Land. Southwestern Journal
In Analysis of Ecological Systems, edited by D. of Anthropology 25:4567.
J. Horn, R. D. Mitchell, and G. R. Stairs, pp. Wilmsen, E. N. 1968. Functional Analysis of Flaked
154177. Ohio State University, Columbus. Stone Srtefacts. American Antiquity 33:156161.
Parry, William J., and Robert L. Kelly. 1987. Wright, Richard V. S. 1977. Introduction and
Expedient Core Technology and Sedentism. Two Studies. In Stone Tools as Cultural Markers:
In The Organization of Core Technology, edited Change, Evolution and Complexity, edited by
by Jay K. Johnson and C. A. Morrow, pp. 285 Richard V.S. Wright, pp. 13. Humanities
304. Westview Press, Boulder. Press, Atlantic Highlands,NJ.
Eight
Procurement Costs and Tool

Performance Requirements:
Determining Constraints on
Lithic Toolstone Selection in Baja
CaliforniaSur
Jennifer M. Ferris
Stone artifacts represent varied human strategies related to how people orga-
nized themselves within their landscape with regards to toolstone procure-
ment, tool manufacture, use, maintenance, and discard. Archaeologists typically
study the treatment of tools, how they were used, for how long, for what task,
and so forth. Often, though, toolstone procurement is largely ignored despite
evidence that the variation we see in lithic technological systems directly links
to the economic decisions people made when selecting their toolstone. Others
have shown the importance of toolstone geological occurrence in relation to
tool production strategies (Andrefsky 1994a, 1994b; Bamforth 1991; Geneste
1985 in Dibble 1991; Kuhn 1991; Nelson 1991; Odell 2000), but an important
aspect these studies present is the absolute knowledge of material source loca-
tions. However, if geochemical sourcing of toolstone is not available, relative
measures may be useful, albeit rarely used, for determining proximate locals.
In this chapter, I focus on how hunter-gatherers selected lithic toolstone and
how this selection was constrained by procurement costs and tool performance
requirements. Procurement costs relate to toolstone acquisition, including
location, abundance, and transport.Tool performance requirements entail both
raw material type and quality, and the types of tasks to be performed. Although
these two types of constraints can work in tandem, often one is more influ-
ential in raw material selection than the other. Expectations are framed using
central place foraging theory and functional studies to evaluate raw material
selection constraints in the archaeological record. These expectations are, in
139
140 Jenni fer M . F e r r i s
turn, applied to the debitage and tool populations from site J69E on Espritu
Santo Island in Baja California Sur to elucidate the driving force behind tool-
stone selection at this site. Four debitage characteristics sensitive to procure-
ment costs are used to assess approximate distance from provenance including
dorsal cortex, flake type, flake size, and reduction trajectory. Tool performance
requirements are assessed through two diversity measures of tool populations,
and flake tool edge damage types and distribution.
Central Place Foraging Theory

Central place foraging (CPF) theory can offer much to the study of lithic
technology, as it is drawn from the larger paradigm of optimal foraging theory
(OFT). OFT provides analytical models from which to test hypotheses about the
archaeological record. Models in optimal foraging theory are based on optimal
solutions for problem solving, born out of rational choice theory (Winterhalder
and Smith 2000). Although these models are not meant to surmise that humans
always act in an optimizing fashion, the use of optimal parameters provides an
avenue for comparing complex behaviors. These models seek to understand the
constraints shaping human behavior in relation to resource selection, time allo-
cation, and habitat movement in order to achieve a goal (usually optimized net
caloric gain) (Kelly 1995;Winterhalder and Smith 2000:54). CPF models primar-
ily assume that foragers will make economically efficient decisions in relation to
field processing and transport costs (Beck 2008; Beck etal. 2002:486; Kelly 1995;
Metcalfe and Barlow 1992). Following Metcalf and Barlow (1992), field pro-
cessing entails separating the high-utility from the low-utility components of a
resource package at the procurement site (e.g., quarry) before transporting the
package to the central place (i.e., habitation site).Trade-offs occur between pro-
cessing time and time spent pursuing other activities, transport weight and travel
distance, and reduction of package size and quantity of toolstone needed (Beck
etal. 2002). In general, it is expected that the farther a central place is from a
procurement site, package reduction (low-utility component removal) becomes
more cost effective and more field processing will occur before transport (Beck
etal. 2002; Kelly 1995; Kuhn 1994; Metcalf and Barlow 1992).
The applicability of CPF theory to lithic studies can be illustrated through
a brief discussion of the Beck and colleagues (2002) study on biface reduc-
tion stages. Their study used CPF theory to evaluate how travel distance from
quarry sites can influence the stage of bifacial reduction from two residential
sites in Nevada. Their study posited that the selection of raw material pack-
ages was predicated on the costs and benefits of transporting the packages and
the time spent processing the materials in the field before they were brought
back to the residential site. Less processed raw material packages require more
energy to transport than more processed packages, and the decision to process
Procurem en t C o s t s a n d T o o l P e r fo r man c e R e q u i r e me n ts 141
in the field is largely a factor of distance but also related to how much tool-
stone is needed at home and what other activities could be pursued instead.
They found that travel distances greatly influenced the decision to process
packages in the field and less field processing occurred when the quarry was
located closer to the residentialcamp.
Procurement Costs and Task Requirements

Understanding whether procurement or performance requirements are
responsible for shaping lithic assemblages archaeologists study is imperative
to comprehending the larger organization of lithic technology. Both of these
constraints may work together to influence selection, although the interplay
between the two is probably best viewed as a continuum.
Procurement costs are a part of all acquisition strategies, which are contin-
gent on landscapes and the human choices made in relation to their surround-
ing environment. Toolstone procurement is often embedded within the larger
subsistence and land-use patterns of hunter-gatherers, but can also occur as
a means unto itself (e.g., Andrefsky 2009; Binford 1980; Nelson 1991; Shott
1986). The geologic occurrence of toolstone will govern whether people pro-
cure at all, how they procure (direct or indirect), and how much they procure
(quantity, shape, and size). Implicit to acquisition strategies is the transportation
of materials. Unless people are living adjacent to a material source, they will
most certainly have to carry lithic packages away from quarries.Transport costs
can greatly determine how much and how far materials are carried, a relation-
ship that is explained through central place foraging theory.
Task functional requirements are dependent on two related factors, includ-
ing the jobs to be performed and the quality of the toolstone. Functional dif-
ferences in toolstone have been documented in other studies (e.g., Andrefsky
1994b, 2005, 2009; Daniel 2001; Ingbar 1994; Jeske 1989; Kuhn 1991; Wenzel
and Shelly 2001). Toolstone differences are largely a function of material qual-
ity, where finer tasks would likely be completed with tools made on finer-
grained or glassy materials and heavier duty tasks would require tools made
from coarser, less brittle materials. People in the past undoubtedly actively
selected different materials for particular tasks. This requirement would have
shaped the selection of toolstone, particularly in environments with abundant
toolstone (Andrefsky 1994b).
Assessment
The focus of this study is the lithic assemblage from the 2006 excavation of site
J69E on Espritu Santo Island in Baja California Sur (Figures8.1 and 8.2). Site
J69E is a shell midden habitation site that dates between the Late Pleistocene/
CALIFORNIA
SAN DIEGO
MEXICALI
TIJUANA ARIZONA
BAJA CALIFORNIA
SONORA
SEA OF
CORTEZ
BAJA CALIFORNIA SUR
PACIFIC OCEAN
ESPIRITU
SANTO
N ISLAND
W E LA PAZ
S
CABO SAN LUCAS
8.1. Map of Baja California peninsula.

LA PARTIDA
ISLAND W E
SEA OF CORTEZ
LA BALLENA J69E
COMPLEX
ESPIRITU SANTO ISLAND
LA PAZ BAY
8.2. Map of Espritu Santo Island.

Early Holocene (Davis 2005). The lithic assemblage consists of 5757 artifacts,
82percent of which are rhyolite (Ferris 2008). Rhyolite is the dominant lithol-
ogy of the island and various qualities are located immediately adjacent to
the site. Rhyolite ranges in quality and texture from porphyritic-aphanitic
with large crystal phenocrysts to a very fine aphanitic texture that appears
almost glassy. All types of rhyolite were used in tool production at site J69E
(Ferris 2008).The second most abundant toolstone types are chert and quartz-
ite, which both have glassy, microcrystalline textures, and each make up 4per-
cent of the assemblage. The remaining toolstone types include other extrusive
igneous rock, namely basalt and andesite, and vein quartz. Given appropriate
sample sizes, only the three most abundant types, rhyolite, chert, and quartzite,
were used in this study.
The high proportion of rhyolite at the site as compared to chert and
quartzite is likely a factor of toolstone selection. This difference may be
partially due to variations in the abundance of the toolstones, where pos-
sibly only smaller amounts of chert and quartzite were available across the
landscape. Alternatively, chert and quartzite source locations may have been
located farther from the site than rhyolite. If indeed there were differences
in geologic occurrence, then those materials located farther away probably
had higher procurement costs, ultimately resulting in their lower represen-
tation at the site. During fieldwork, I observed possible rhyolite sources
adjacent to the site but did not see any potential chert nor quartzite source
locations (Ferris 2008). As no geochemical sourcing was available at the
time of this study, there was no absolute way to determine if the local rhyo-
lite was also used as toolstone. Chert and quartzite may have been available
farther away or potentially in limited abundance. Alternatively, toolstone
location and abundance may have not been a significant constraint to the
sites inhabitants. Instead, the different lithic raw materials may have been
a more important factor in selection to accomplish certain tasks at the site.
The varying qualities of rhyolite may have been more accommodating for a
range of tasks, resulting in a much higher proportion of rhyolite represented
at thesite.
Procurement
Whereas the study of Beck and colleagues used biface reduction stages to test
expectations about procurement strategies, the lithic assemblage at site J69E
contained very few bifaces (Ferris 2008). Instead, the debitage population was
large (n = 5301), permitting the use of this often overlooked artifact type
(Ferris 2008). Tool production is a reductive process. Given that debitage is
rarely removed from the site at which it was detached from the core, it provides
a more complete look at reduction sequences than other artifact types that are
more likely to be removed from sites, such as bifaces. Some flakes may be used
Table8.1. Proximal flake cortex frequency
Dorsal cortex
Material 0% 050% 5099% 100% Total
Rhyolite 2973 (87.0) 236 (6.9) 154 (4.5) 53 (1.6) 3416
Chert/quartzite 283 (91.0) 12 (3.9) 12 (3.9) 4 (1.3) 311
Total 3256 248 166 57 3727
Chi-square x2 = 4.9, d.f. = 3, p = 0.1793
Note: Relative proportions calculated within material type noted in italicized
parentheses.
as tools, effectively removing them from the debitage population (e.g., Beck
etal. 2002); however, as debitage populations can be very large, the removal of
some flakes may not cause much effect.
The four debitage characteristics used here are dorsal cortex amounts, flake
types, flake sizes, and reduction trajectories. The following tests on these char-
acteristics were designed to answer whether tool users at site J69E processed
material packages in the field at quarry locations or brought rough raw materi-
als back to J69E for processing. These characteristics, in turn, illuminate which
toolstone types might have been located closest to the site. If transporting
toolstone was a high cost to the tool users of site J69E (e.g., distantly located),
they probably would have practiced field processing. If so, the debitage should
be represented in more reduced forms rather than early stage reduction forms.
Further, if debitage from different toolstone types exhibit different stages of
reduction, it may support that there was differential access to raw material
sources potentially related to source distances, smaller package sizes, or differ-
ent levels of abundance.
Field processing would effectively reduce transport weight by removing
the low-utility portions of toolstone, such as dorsal cortex. The amount of
dorsal cortex on complete flakes can provide rough estimates to the degree of
reduction that has occurred, provided that cortex cover occurs on the original
objective pieces (Andrefsky 2005). As such, the first expectation was that chert
and quartzite debitage would contain a lower proportion of flakes with cortex
than rhyolite if it were transported farther distances and processed in the field.
Dorsal cortex was recorded along an ordinal scale of 0%, 149%, 5099%, and
100%. Chert and quartzite flakes do indeed show slightly lower frequencies
than the rhyolite flakes in all categories of cortex cover, but the distribution is
not significant (x2 = 4.9, d.f. = 3, p = 0.1793, Cramers V = 0.04) (Table8.1).
This suggests that all toolstone was either trimmed in the field before trans-
port, or very little cortical cover existed on the cores to beginwith.
The types of flakes, including proximal flakes, flake shatter, and angular shat-
ter, can inform about reduction stages. Kooyman (2000) notes that angular
shatter is typical during early stages of core reduction. If one toolstone type
Table8.2. Flake type frequency
Flake type
Material Proximal flake Flake shatter Angular shatter Total
Rhyolite 3417 (78.8) 202 (4.7) 719 (16.6) 4338
Chert/quartzite 310 (76.2) 24 (5.9) 73 (17.9) 407
Total 3727 226 792 4745
Chi-square x2 = 1.936, d.f. = 2, p = 0.3798
Note: Relative proportions calculated within material type noted in italicized parentheses.
100
Flake Type
Proximal
Flake Shatter
75 Angular Shatter
Bars show percents

Percent
50
n = 3417 n = 310
25
n = 202 n = 24
n = 719 n = 73
0
1 2
Material Type
8.3. Bar chart displaying percentages of flake types for rhyolite (type 1) and chert/quartzite
(type2).
has a greater proportion of angular shatter than another, it may suggest that
earlier reduction occurred at the site rather than at the source location. An
examination of the proportion of flake types between materials shows that
their distribution is not significantly different between materials (x2 = 1.936,
d.f. = 2, p = 0.3798, Cramers V = 0.02) (Table8.2 and Figure8.3). This sug-
gests that all toolstone had similar proximities to the site and were likely treated
in a similar manner.
Debitage sizes are often associated with reduction stage.Typically, the largest
flakes are removed first, and as the core gets smaller, the flakes do too. If raw
material field processing occurred to reduce transport weight for one lithic
material type, the proportion of large flakes to small flakes should be smaller
than that of the other material. The debitage populations were aggregated by
material class into size grades in 10-mm intervals, where class 1 is less than
10mm, class 2 is 1020mm, and so forth based on maximum linear dimen-
sion. The materials show similar patterns, where there are fewer large flakes
Material Type
Rhyolite
30.0 CCS/chert/QZT
20.0
Percent
10.0
0.0
1 2 3 4 5 6
Size Class
8.4. Line graph of complete flake size grade percentages.
than smaller flakes, following the notion that reduction is a subtractive process
(Figure8.4). Chert and quartzite may have undergone some field processing, as
their proportion of large flakes is less than the proportion for rhyolite, but this
is more likely due to core morphology than reducing transport costs.
The last proxy for procurement and reduction is flake reduction trajectory.
This was assessed by plotting size grade against weight for complete flakes.
Flakes that are larger in both weight and maximum linear dimension are typi-
cally removed during the early stages of lithic reduction (Kooyman 2000). If
field processing occurred for chert and quartzite as the previous analysis of
size grade percentages tentatively suggested, then chert and quartzite should
have a shorter reduction trajectory exhibited at site J69E than the reduction
trajectory for rhyolite. However, the materials were shown to have similar tra-
jectories, indicating that field processing occurred with similar frequency for
all three stone types (Figure8.5).
The four debitage characteristics evaluated do not support that procure-
ment costs were a significant constraint in toolstone selection. It is likely that
abundance and location were not influencing acquisition strategies. Instead,
task requirement and material quality may have been more important for tool-
stone selection than procurement costs.
Tool Functional Requirements

Toolstone material qualities are likely chosen for specific tasks, and the require-
ments of these tasks in turn influenced toolstone selection. Task requirement
and material quality were evaluated by assessing tool population diversity for
rhyolite and chert and quartzite, as measured by richness and evenness. The
Material Type
40.0 1
2
30.0
Weight (g)
20.0
10.0
0.0
1 2 3 4 5 6
Size Class
8.5. Line graph of complete flake reduction trajectory.
types and distribution of flake tool edge damage were evaluated to determine
whether flake tools of different materials were utilized for different tasks. Chert
and quartzite were combined as they both have a limited range in texture as
compared to rhyolite, which ranged in various textures and qualities.
Richness is measured by the number of tool types represented while even-
ness is calculated as the frequency of each type within the population (Ames
1988; Chatters 1987). Rhyolite should be richer than chert and quartzite as their
range in toolstone quality is conducive to completing many different tasks.
Chert and quartzite should then be more even than rhyolite because they may
be used in a more restricted or specialized set of tasks given their limited range
in texture. The types of tools included in the richness analysis include projec-
tile points, other bifaces, unidirectional cores, multidirectional cores, unifaces,
scrapers, drills, spokeshaves, unimarginal modified informal flake tools, bimar-
ginal modified informal flake tools, and unimarginal and bimarginal modified
informal flake tools. Rhyolite consists of 12 different tool types, while chert
and quartzite only have eight different tool types (Table8.3). Rhyolite follows
the expectation that it is richer than the non-igneous materials, which could
be a factor of sample size.Typically, the larger an assemblage is, the greater vari-
ety it should contain (Rhode 1988). On the other hand, the rhyolite qualities
probably sufficiently met functional requirements at site J69E. However, an
upper limit was placed on the richness of either population because the arti-
fact classes were collapsed to just 12 types. Chert and quartzite are less rich,
which could be either a result of abundance or not meeting the requirements
for certain tasks. Chert and quartzite only occur in very limited frequencies
(n= 22), which may explain why only some classes are represented.Alternatively,
certain tasks, such as sharpening, shaving, and whittling, may be best accom-
plished by less brittle material, such as rhyolite. For example, no spokeshaves
Table8.3.Tool categories included in the richnessindex
Material type
Tool type Rhyolite Chert/quartzite Total
Projectile points 10 2 12
Knives 2 0 2
Other bifaces 10 1 11
Unidirectional cores 36 1 37
Multidirectional cores 147 5 152
Unifaces 2 2 4
Scrapers 51 1 52
Drills 3 0 3
Spokeshaves 4 0 4
Unimarginal flake tools 93 8 101
Bimarginal flake tools 19 0 19
Uni-/bimarginal flake tools 13 2 15
Total count 390 22 412
Total types 12 8 20
were made out of non-igneous materials, while four were made from rhyolite
toolstone (Table8.3).
Evenness, as mentioned previously, is the number of items in each tool
category for each population, which is measured with the evenness index
(Andrefsky 2005; Chatters 1987; Pielou 1966 in Ames 1988). The evenness
index is calculatedby:
n n
i log i
E= n n
ns
where ni is the number of artifacts for each tool type, n is the number of artifacts
for all types within the population, s is the total number of artifact types, and log
is calculated on base 10 logarithms (Drennan 1996; Shennan 1997). The index
ranges from 0 to 1.0 and provides a value for the spread of the number of items
in each tool category for each population. A value of 0.0 means the specimens
in the population are represented by only one type, while a value of 1.0 means
the population is maximally even all the types are equally represented. Chert
and quartzite were found to have an evenness index of 0.853 and rhyolite has
an index value of 0.716. This finding shows that chert and quartzite were more
evenly distributed than rhyolite in terms of tool function, suggesting that peo-
ple were actively selecting different materials for different tasks. This difference
is likely due to the types of toolstone materials, and not abundance. The pro-
duction of tools was most likely the result of the tasks that tools were used for,
which is independent of toolstone distribution (see Kuhn 1991:85).
70.0 Material Type

Rhyolite
60.0 CCS/Quartzite
50.0
40.0
Percent
30.0
20.0
10.0
0.0
FD FD FD SM FD ST FD ST FD ST
MOON MOON SM
Edge Damage Pattern
8.6. Line graph displaying proportions of edge damage patterns for utilized flake tools by mate-
rial type. Edge damage pattern: (FD) feathered; (MOON) half moon; (SM) smoothed; (ST)
stepped.
Two characteristics of flake tool edge damage were assessed to further deter-
mine differential use between material types. The first measure recorded the
types of removals from the tool edge including feathered and stepped micro-
chips, half moon removals, and edge smoothing. The second measure of dam-
age was the distribution of the microchips (how they relate to one another),
including scattered, continuous, overlapping, and superposed (Richards 1988).
Edge damage on all utilized flakes was assessed with a 10 hand lens. Only
utilized flakes were used for the following analyses in efforts to minimize
edge damage that was not caused by use (e.g., retouch). While a suite of lit-
erature exists regarding identifying the origins of edge damage (see Amick
and Mauldin 1997; Andrefsky 2005; Ferris and Andrefsky 2007; Keeley 1980;
Kooyman 2000; Odell 1981; Richards 1988), the purpose here is to show dif-
ferent patterns of use-wear and not to assume specific tasks. In addition, the
types of materials worked and the intensity of the task would alter the patterns
of edge damage on utilized flakes. Essentially, edge damage patterns should
differ between utilized flake tools made on the different types of toolstone.
Given the microcrystalline texture of chert and quartzite, they are likely more
conducive to finer tasks and the range of edge damage should be constrained
to certain types. Conversely, owing to the range of rhyolite qualities, a wide
range in edge damage patterns should be present for the rhyolite flake tool
population.
Figure 8.6 displays the results for the first measure of edge damage. The
igneous material contained a much higher proportion of feathered remov-
als than the non-igneous material, which contained a higher proportion of
feathered and stepped microchip removals. Feathered removals are typically
50.0 Material Type

Rhyolite
CCS/Quartzite
40.0
30.0
Percent
20.0
10.0
0.0
CL CL CL CT CT CT CT CT
SG SP CL CL SG SP
SG
Microchip Distribution
8.7. Line graph displaying microchip configuration proportions for edge damage of utilized
flake tools by material type. Microchip distribution: (CL) overlapping; (SG) scattered; (SP)
superposed; (CT) continuous.
associated with working softer materials and harder materials generally result
in stepped removals (Andrefsky 2005; Odell 1981). Although a broad general-
ization, these results do suggest that different toolstones were used for different
tasks. However, the texture of the toolstone themselves will affect the types of
edge damage and it is possible that similar materials were worked with all the
toolstone.
Figure8.7 depicts the distribution of microchip removals. The most com-
mon distribution types were continuous for rhyolite and a combination of
continuous and superposed for chert and quartzite. In general, continuous
microchip removals are often caused by less intense activities, such as whittling
(Richards 1988), while superposed and overlapping removals typically result
from excessive action, such as sawing or working harder material (Lawrence
1979; Richards 1988). Again, these data suggest differential use between tool-
stone raw materials.
Discussion and Summary

Stone tool assemblages are the direct result from decision-making practices.
These practices are the byproduct of human behaviors, which are well suited to
assessments through evolutionary theoretical models. One such model drawn
from CPF theory was utilized in the present study to frame expectations about
toolstone selection. CPF theory attempts to explain the behaviors of foragers
who consume resources at some central place in relation to where its acquired.
Such a model is aptly suited to understanding lithic material procurement, as
more often than not, toolstone geologic provenances are not the same location
as a residential camp. People would leave their central camps and go out on
logistical forays for resources needed, including toolstone. Once they arrived at
a provenance, decisions had to be made in regards to transporting the materials
home, including how much to carry, what to carry, what to not carry, what to
take part of the way, what to stash for later, and so forth. Stone tool users had
to weigh the costs and benefits of each decision; essentially, each decision was
a trade-off between field processing and transport costs (Metcalfe and Barlow
1992).
Based on analyses of cortex, flake types, size grades, and reduction trajec-
tories, the site J69E debitage population does not support that the stone tool
makers elected to field process their raw materials. It seems that the lithic
materials used for tool production were all available close enough to the site,
eliminating high raw material transport costs. However, transport costs may
still have been low even if materials were carried from the opposite side of the
island, given that Espritu Santo Island measures approximately 20 km north/
south and 5 km east/west. Often local sources are delineated as 5 km away or
less and distant sources are those that are beyond 20 km (Brantingham 2003),
which would mean that the current study area, Espritu Santo, falls somewhere
in between local and not so distant. Moreover, the proxies assessed for rhy-
olite, chert, and quartzite reduction stage all follow similar patterns. As a result,
it is highly likely that these toolstone types had comparable contexts for their
availability across the landscape. No one type was farther away than another
despite the inability to relocate chert and quartzite during field reconnaissance.
Variation in toolstone abundance largely accounts for the large proportional
differences between rhyolite and chert and quartzite.The bedrock of the island
is primarily comprised of thick Miocene rhyolite flows (Carreo and Helenes
2002), providing ample material for stone tools.
Given that lithic toolstone location and abundance did not constrain selec-
tion, it is likely instead that selection was based on the functional demands of
tasks.The implementation of a performance based procurement strategy at site
J69E is supported by the analyses of tool diversity and flake tool edge damage.
Raw material quality played largely into tool functional requirement, where
finer tasks required finer grained toolstone and larger grained materials were
better suited for coarser tasks. Such performance based selection is not directly
accounted for in the CPF model, although it can play a small role in field pro-
cessing decisions where time spent reducing one material type takes away from
time spent looking for other materials.
It is important at this juncture to address one matter inherent to the present
study. Here I have worked under the proviso that people actively sought their
lithic materials.This is in somewhat stark contrast to other studies that approach
toolstone procurement from either a completely opportunistic or a neutral
platform (see Brantingham 2003). Opportunistic procurement strategies are

based simply on toolstone abundance where those types in higher abundance
have a greater likelihood of being picked up (Brantingham 2003:489). Such
an interpretative tool could be used above where I note the abundance of
naturally occurring rhyolite on Espritu Santo and its effects on the assem-
blage. However, to conclude this study by stating that more rhyolite artifacts
were recovered because more rhyolite occurs naturally would be a disservice.
Instead, I assessed multiple proxies for reduction and production in efforts to
tease out the possible behaviors responsible for the observed phenomena.
In addition, although neutral models can provide us with a minimum base-
line, it is the deviations from those baselines that we find interesting as archae-
ologists. To get at these deviations, we must look at the biases that are inherent
in human decisions. Two such biases assessed here were procurement costs and
performance needs. However, other constraints may play roles in toolstone
selection, such as sociopolitical and ethnic boundaries. Determining which
constraints are influencing toolstone selection is necessary to gaining a more
complete understanding of lithic technological organizations.
Lastly, applying geochemical sourcing analyses could greatly strengthen
conclusions about the availability and treatment of toolstone. However, when
such analyses are not available, as in the present study, using characteristics that
are sensitive to field processing and source proximity can greatly add to our
understanding of toolstone procurement.When these characteristics are drawn
from higher theoretical paradigms, such as CPF theory, greater understanding
of behavior is afforded.
Acknowledgments
This research was supported by Instituto Nacional de Antropologa e Historia,
Oregon State University, and Washington State University. I thank Harumi
Fujita, Loren Davis, the Jacksons, and the rest of the 2006 field crew for their
assistance in conducting the fieldwork. I also thank William Andrefsky, Jr. and
Nathan Goodale for their constructive reviews of a draft of this chapter.
References
Ames, Kenneth M. 1988. Early Holocene Andrefsky, William, Jr. 1994a. The Geologic
Forager Mobility Strategies on the Southern Occurrence of Lithic Material and Stone
Columbia Plateau. In Early Human Occupation Tool Production Strategies. Geoarchaeology: An
in Far Western North America:The Clovis-Archaic International Journal 9:375391.
Interface, edited by J. A.Willig, C. M. Aikens, and Andrefsky, William. 1994b. Raw Material
J. L. Fagan, pp. 325371. Nevada State Museum, Availability and the Organization of
Anthropological Papers No. 21, CarsonCity. Technology. American Antiquity 59:2135.
Amick, Daniel S., and Raymond P. Mauldin. 1997. Andrefsky, William. 2005. Lithics: Macroscopic
Effects of Raw Material on Flake Breakage Approaches to Analysis, 2nd ed. Cambridge
Patterns. Lithic Technology 22:1832. University Press, Cambridge.
Andrefsky, William. 2009. Analysis of Stone Tool and S. Holen, pp. 3348. University of Kansas
Procurement, Production, and Maintenance. Publications in Anthropology 19, Lawrence.
Journal of Archaeological Research 17:65103. Drennan, Robert D. 1996. Statistics for
Bamforth, Douglas B. 1991. Technological Archaeologists: A Commonsense Approach.
Organization and Hunter-Gatherer Land Plenum Press, NewYork.
Use: A California Example. American Antiquity Ferris, Jennifer M. 2008. Lithic Technological
56:216234. Organization of Site J69E, Espritu Santo
Beck, Charlotte, Amanda K. Taylor, George Island, Baja California Sur. Unpublished
T. Jones, Cynthia M. Fadem, Caitlyn R. masters thesis, Department of Anthropology,
Cook, and Sara A. Millward. 2002. Rocks Washington State University, Pullman.
Are Heavy: Transport Costs and Paleoarchaic Ferris, Jennifer M., and William Andrefsky, Jr.
Quarry Behavior in the Great Basin. Journal of 2007. Transport Damage and Lithic Analysis:
Anthropological Archaeology 21:481507. New Insights. Paper presented at the New
Beck, Kelly R. 2008. Transport Distance Ground Australasian Archaeology Conference,
and Debitage Assemblage Diversity: An Sydney, NSW, Australia. Manuscript on file,
Application of the Field Processing Model to Department of Anthropology, Washington
Southern Utah Toolstone Procurement Sites. State University.
American Antiquity 73:759780. Ingbar, Eric E. 1994. Lithic Material Selection
Binford, Lewis R. 1980.Willow Smoke and Dogs and Technological Organization. In The
Tails: Hunter-Gatherer Settlement Systems Organization of North American Prehistoric
and Archaeological Site Formation. American Chipped Stone Tool Technologies, edited by P. J.
Antiquity 45:420. Carr, pp. 4556. International Monographs in
Brantingham, P. Jeffrey. 2003. A Neutral Model of Prehistory, Archaeological Series 7, Ann Arbor.
Stone Raw Material Procurement. American Jeske, Robert. 1989. Economies in Raw Material
Antiquity 68:487509. Use by Prehistoric Hunter-Gatherers. In
Carreo, Ana Luisa, and Javier Helenes. 2002. Time, Energy, and Stone Tools, edited by Robin
Geology and Ages of the Islands. In A New Torrence, pp. 3445. Cambridge University
Island Biogeography of the Sea of Cortez, edited Press, Cambridge.
by Ted J. Case, Martin L. Cody, and Exequiel Keeley, Lawrence H. 1980. Experimental
Ezcurra, pp. 1440. Oxford University Press, Determination of Stone Tool Uses: A Microwear
NewYork. Analysis. University of Chicago Press,
Chatters, James C. 1987. Hunter-Gatherer Chicago.
Adaptations and Assemblage Structure. Journal Kelly, Robert L. 1988. The Three Sides of a
of Anthropological Archaeology 6:336375. Biface. American Antiquity 53:717734.
Daniel, I. Robert. Jr. 2001. Stone Raw Material Kelly, Robert L. 1995. The Foraging
Availability and Early Archaic Settlement in Spectrum: Diversity in Hunter-Gatherer
the Southeastern United States. American Lifeways. Smithsonian Institution Press,
Antiquity 66:237266. Washington,DC.
Davis, Loren G. 2005. Preliminary Report of the 2004 Kooyman, Brian P. 2000. Understanding Stone
Archaeological Investigations at Site J69E, Espritu Tools and Archaeological Sites. University of
Santo Island, Baja California Sur. Submitted to Calgary Press, Calgary, Canada.
the Consejo de Arqueologa, Mexico City, Kuhn, Steven L. 1991. Unpacking Reduction:
Mexico, March 15,2005. Lithic Raw Material Economy in the
Dibble, Harold L. 1991. Local Raw Material Mousterian of West-Central Italy. Journal of
Exploitation and its Effects on Lower and Anthropological Archaeology 10:76106.
Middle Paleolithic Assemblage Variability. Kuhn, Steven L. 1994. A Formal Approach to
In Raw Material Economies Among Prehistoric the Design and Assembly of Mobile Toolkits.
Hunter-Gatherers, edited by A. Montet-White American Antiquity 59:426442.
Lawrence, Robert A. 1979. Significance of Richards, Thomas H. 1988. Microwear Patterns on

Attributes Used in Edge-Damage Analysis. In Experimental Basalt Tools. BAR International
Lithic Use-Wear Analysis,edited by Brian Hayden, Series 460, Oxford.
pp. 113131. Academic Press, NewYork. Shennan, Stephan. 1997. Quantifying
Metcalfe, Duncan, and K. Renee Barlow. 1992. Archaeology, 2nd ed. University of Iowa Press,
A Model for Exploring the Optimal Trade- Iowa City.
off between Field Processing and Transport. Shott, Michael J. 1986. Technological
American Anthropologist 94:340356. Organization and Settlement Mobility:
Nelson, Margaret C. 1991. The Study of An Ethnographic Examination. Journal of
Technological Organization. In Archaeological Anthropological Research 42:1551.
Method and Theory,Vol. 3., edited by M. B. Schiffer, Wenzel, Kristen E., and Phillip H. Shelly. 2001.
pp. 57100. University of Arizona Press,Tucson. What Put the Small in the Arctic Small
Odell, George H. 1981. The Mechanics of Tool Tradition: Raw Material Constraints
Use-Breakage of Stone Tools: Some Testable on Lithic Technology at the Mosquito Lake
Hypotheses. Journal of Field Archaeology Site, Alaska. In Lithic Debitage: Context, Form,
8:197209. Meaning, edited by William Andrefsky Jr.,
Odell, George H. 2000. Stone Tool Research at pp. 106123. University of Utah Press, Salt
the End of the Millennium: Procurement and LakeCity.
Technology. Journal of Archaeological Research Winterhalder, Bruce, and Eric A. Smith. 2000.
8:269331. Analyzing Adaptive Strategies: Human
Rhode, David. 1988. Measurement of Behavioral Ecology atTwenty-five.Evolutionary
Archaeological Diversity and the Sample-Size Anthropology 9(2):5172.
Effect. American Antiquity 53:708716.
Nine
A Model of Lithic Raw Material

Procurement
Raven Garvey
Raw material acquisition is fundamental to any technology. Modern engineers

carefully consider the costs and advantages of potential building materials
before ground is ever broken, weighing budgetary constraints against structural
integrity, for example. People of the prehistoric past were faced with similar
decisions both within and apart from formal economies. This chapter consid-
ers lithic raw material procurement from an evolutionary perspective, using a
model that predicts the amount of tool use necessary to warrant investment in
hard to obtain, high-quality raw materials when local but less-good ones are
available.
Models that use objective scales of optimal behavior have been applied
to archaeological records with appreciable success. A majority of these have
focused on aspects of optimal foraging, predicting subsistence behaviors and
their attendant patterns of mobility given certain environmental and, to a
lesser degree, social or technological parameters (e.g. Basgall 1987; Bettinger
and Baumhoff 1982; Bettinger etal. 1997; Broughton 1997; Hildebrandt and
McGuire 2002; Jones 2004; Madsen and Schmitt 2003;Waguespack and Surovell
2003). Although economic models have provided some areas of archaeological
research with fresh interpretations and falsifiable hypotheses, studies of stone
technology have not made extensive use of such models (for important excep-
tions, see Beck etal. 2002; Brantingham 2003; Jeske 1992; Surovell 2009). This
is not to say that modern lithic analysis lacks reference to economic decisions.
Indeed, as Brantingham (2003:504) observes, many lithic studies assume that
156
A M odel of L i t hi c R aw M at e r i a l P r o c u r e m e n t 157
optimal foraging strategies must influence and, therefore, be diagnosed by

stone raw material procurement patterns. Research that addresses the relation-
ship between technology and subsistence has made important contributions to
our understanding of toolkit composition and raw material use. Nonetheless,
and perhaps because lithic technology is not easily translated into fitness (cf.
Dunnell 1980), there are still few applications of neo-Darwinian principles that
assess the direct costs and payoffs associated with stone procurement.
Binfords (1979) insightful discussion of lithic procurement as an embedded
activity changed the way archaeologists think about the economics of stone
use. Raw material collection was described as an ad hoc part of subsistence
forays because the wise (optimizing) forager would not make express and exclu-
sive trips to stone sources except under extraordinary circumstances (Binford
1979:259; italics in original). That is, if everything goes well, there are few or
no direct costs accountable for the procurement of raw materials (Binford
1979:259). Obtaining stone has since been regarded a largely opportunistic
endeavor: foragers, finding themselves close to a source and with the free time,
energy, and space for carrying it, will collect stone rather than return home
empty-handed. Treating lithic procurement as an embedded activity shifts the
focus from the getting of stone to the economizing of stone once it is gotten.
Early descriptions of curated gear (Binford 1973), which solved the prob-
lem of spatial incongruence between stone and food resources, and of embed-
ded procurement (Binford 1979), which effectively liberated hunter-gatherers
from purposive toolstone excursions, inspired a number of studies designed to
interpret the composition of archaeological toolkits. These analyses explore
economizing strategies, assessing the degree of reduction and preparation per-
formed at stone sources, whether certain materials were used expediently or
reserved for formal tools, whether stone tool users were trying to maximize
the use-lives of some or all of the tools in a toolkit, and how these decisions
were influenced by subsistence resource types and their timing (e.g., Andrefsky
1994; Bamforth 1986, 1990; Bleed 1986; Goodyear 1989; Jeske 1992; Kelly 1988;
Kuhn 1991, 1992, 1994; Torrence 1989).
A second major area of lithic research considers the presence of particular
stone types in archaeological assemblages. These studies explore variables such
as raw material richness, evenness, and relative size, and distances from sites
to sources to gauge a groups degree of mobility, estimate foraging radii and
assess stone conservation (e.g. Bamforth 1990; Basgall 1989; Eerkens etal. 2007;
Eerkens etal. 2007; Jones etal. 2003; McGuire 2002; cf. Brantingham 2003).
Research of this ilk has been especially fruitful where available raw materi-
als have distinct and identifiable chemical signatures, as with obsidians in the
western United States.
These two lines of inquiry reflect the influence that embeddedness has
had on lithic studies in that the behavioral interpretations they afford begin
158 Raven G arve y
p ost-acquisition. Hunter-gatherers are assumed to have been primarily

concerned with subsistence resource procurement and only secondarily with
stone procurement; stone was collected when convenient. However, the casu-
alness of embedded procurement seems fundamentally at odds with the care-
ful planning implied in curation and stone economizing. Further, embedding
lithic procurement reduces the opportunity cost of getting stone but is possible
only when food and stone are spatially congruent, in which situation the pres-
sure to budget time and energy carefully that is, to embed is minimal. A
third body of lithics literature brings this paradox into sharper focus.
Goodyear (1989) and others (e.g., Funk 1972; Hester and Grady 1977; Kelly
and Todd 1988) consider the special case of hypermobile Paleoindian groups,
who are noted for their almost exclusive use of high-quality cryptocrystalline
silicates. Sources of these materials are often located considerable distances
from the sites where manufacturing byproducts and tools were discarded. This
phenomenon has been attributed to the premium that Paleoindians placed on
mobility; obtaining high-quality stone was worth the effort because its physi-
cal properties complemented the mobile lifestyle (Goodyear 1989). Coupled
with this is the idea that the paleoenvironment was teeming with high-ranked
subsistence resources, unlike later periods when a changed proportion of
humans to resources governed mobility. During the Paleoindian period, lithic
raw material procurement may have been perfectly embedded in the subsis-
tence system and Paleoindians could have relied on high-quality stone because
it was, in essence, always local. In fact, one might argue that stone was the
limiting resource for Paleoindians and that, because food resources were ubiq-
uitous, Paleoindians embedded hunting in their pursuit of stone rather than
the other way around (cf. Gardner 1977).
Under the current framework, then, choices regarding when and from
where to procure stone appear to require at least three distinct explanations.
The first treats scenarios in which food resources were ubiquitous, as we pre-
sume was true of the Paleoindian period, and high mobility ensured high
encounter rates with both game and stone (Kelly 1988). That is, embedding
lithic procurement was possible. A second explanation addresses the scenario
in which hunter-gatherers were faced with a less abundant resource base and,
therefore, potential disjunctions between food and high-quality stone. People
experiencing food resource stress may have preferred embedding because it
reduced the cost of stone procurement, but resource distributions may have
made this strategy untenable. These hunter-gatherers may have made due
with lower-quality, local raw materials when getting high-quality ones was too
costly. A third explanation deals with stone procurement among groups whose
resources were tightly time constrained and required a reliable toolkit (Bleed
1986; Torrence 1989), which may have necessitated deliberate, costly treks to
high-quality stone sources.
The implication that stone procurement decisions are driven by the relative
abundance of food resources is environmentally deterministic (Gould and
Saggers 1985). Stone technologies, including the procurement decisions at their
base, are seen as adaptations to prevailing environmental conditions. Absent
detailed paleoenvironmental data, explanations of the relationships among
environment, prey, and technology tend to be circular. Further, if cultures are
in homeostasis, perfectly and perpetually adapted to their environments, cul-
ture change requires an external catalyst, frequently taken to be a change in the
environment (Bettinger 1991). When the environment is seen as the sole cause
of cultural change, archaeology ceases to be the study of human behavior and
becomes a chronicling of environmental change.
An alternative assessment of raw material procurement, one that avoids envi-
ronmental determinism and accounts for all three of the scenarios described
previously, uses a simple economic model with a fitness-based explanation of
procurement decisions. Because natural selection favors behaviors that maxi-
mize somatic maintenance and reproductive success, which are largely invisible
archaeologically, proxy measures (e.g., caloric return rates) are used to gauge
prehistoric fitness (Bettinger 1991; Kelly 1995; Winterhalder and Smith 1992).
The model presented here predicts raw material procurement decisions based
on the assumptions that lithic materials are ranked according to their qual-
ity and that high-quality materials improve return rates for the activities they
are used to perform. Importantly, this model can be combined with others to
make and test predictions about complex human behaviors.
TheModel
A model of technological intensification, recently described by Bettinger etal.
(2006; see also Ugan et al. 2003) predicts the amount of time that must be
devoted to a subsistence activity before a tool user will achieve a higher rate of
return from a more costly technology relative to a less costly one. Any amount
of time less than this critical use time ensures a lower rate of return from the
more costly technology because the time and energy required to produce
such a tool negates the energetic payoff of the resources procured with it.
Given this relationship, as increased time is devoted to these [particular] sub-
sistence activities it pays to invest more in technologies that increase their rate
of return (Bettinger etal. 2006:538).
This model can be used to predict lithic procurement decisions when the
currencies and constraints are redefined. Just as technologies differ with respect
to manufacturing costs and return rates, stone types differ in the costs associ-
ated with their procurement and reduction, as well as the benefits they afford
when used. The model parameters necessary to predict when certain materials
will be favored include procurement and manufacturing costs, measures of raw
160 Raven G arve y
material quality and rates of return from tools of a given material type, and
tool use time.
Procurement and ManufacturingCosts

The time or energy required to travel to a stone source, locate, and extract a
portion of suitable size and quality from within the source, and manufacture
a tool from it are the costs associated with stone tool production. Because
stone tool use-lives are understood to be highly variable (Andrefsky 1998),
manufacturing costs the time or energy required to produce usable tools
must be assessed taking a groups technological system into account. Alternate
versions of this costbenefit calculus could include maintenance (the ease of
rejuvenation relative to the frequency with which it is required) or other vari-
ables, determined by the materials record and the nature of the inquiry. A
more complex equation could be drafted to calculate procurement costs when
materials are obtained through trade.
Differences in direct procurement costs will often overwhelm those associ-
ated with tool production, which may be minimal once a suitable piece of
stone is selected. The total time or energy expended in traveling to the more
distant of two stone sources will generally be greater than that expended in
producing a tool from the inferior, closer source. However, although differ-
ences in manufacturing costs may seem negligible, they ought not be wholly
dismissed. The swamping effect may be moderated if procuring the inferior
material requires more within-source searching for a suitable piece and involves
a higher rate of failed production attempts, which would lessen the difference
between the two sources procurement costs and accentuate the difference in
their manufacturing costs.
Raw Material Quality and Rates of Return

Lithic materials differ in their suitability for tool production. Properties that
affect this include isotropism (lack of internal directionality), brittleness (ten-
dency to fracture rather than deform under stress) and homogeneity (Andrefsky
1998; Cotterell and Kamminga 1979; Whittaker 1994). High values of these
properties permit more efficient tool production and extend use-lives because
reduction and rejuvenation are more easily controlled, resulting in less waste
and more use-edge per unit weight.
Crystalline structure also influences toolstone workability. Amorphous
and cryptocrystalline materials are preferred for their predictability, supe-
rior brittleness, and sharp edges (Whittaker 1994). Generally, the larger the
crystal structure, the more difficult (because less predictable and controllable)
the knapping, the less sharp the flaked edge, and the less easily reworked the
resulting tool. Noncrystalline materials such as obsidian and opal are more
easily reduced and maintained than meso- and macrocrystalline materials such
as quartzite and basalt.
To reduce subjectivity and allow for comparison between assemblages,
Brantingham et al. (2000) describe a means of quantifying the properties
described above. Their measures include percent crystallinity, average and
range of crystal size, and abundance of impurities... (Brantingham et al.
(2000:257). These variables have clear advantages, but it may not always be
possible or practical to incorporate them, and analysts may lack the geological
training that would ensure that these measures are consistent across analyses.
When circumstances preclude this level of detail, the analyst might refer to
Callahan (1979:16), who presents a general classification system for lithic raw
materials. Frequently, assemblages will consist of only a low or moderate num-
ber of raw material types (Brantingham 2003) and the differences in their qual-
ity may be readily distinguishable by more conventional means of assessment.
Tool return rates are directly related to raw material quality for the reasons
described earlier, and can be calculated in a variety of ways. For example,
return rates might be conceptualized in terms of a tools total usable hours or
the calories generated by its use. Determining an appropriate return rate will
also depend on the material record and the nature of the inquiry.
UseTime
The amount of time that must be devoted to a particular activity before a tool
user will achieve a higher rate of return from a more costly item relative to a
less costly one is referred to as the critical use time (Bettinger etal. 2006). The
critical use time that should trigger a switch from one lithic raw material type
to another is determined by the costs and benefits associated with the compet-
ing materials, described previously.
In the graphical representation of this model (Figure9.1a), the dimension
time, along the x-axis, is divided into time spent obtaining a raw material
and crafting a useable tool from it, and time spent using the tool.Time of either
kind increases with increased distance from the origin. Return rates lie along
the y-axis. The hypothetical relationship in the figure illustrates that material A
is characterized by a modest procurement-plus-manufacturing time, perhaps
because this material source is located close to the site where it is used, but the
return rate is also relatively low, because it has a large crystal structure or numer-
ous imperfections. Material B has a higher rate of return given its superior qual-
ity, but its procurement-plus-manufacturing cost is also higher, perhaps because
the source is located at a considerable or difficult distance from the site. The
relationship between these two material types, their respective returns relative
to the time it takes to procure each and craft a tool from it, defines the critical
162 Raven G arve y
use time, C, or the amount of time that must be devoted to a task to make
higher-quality material B viable in the presence of lower-quality but more
accessible material A. As a result, any amount of use time less than the critical
use time favors using material A because, at this level of use, the return rate for A
is higher than the return rate for B. Conversely, any amount of use time greater
than the critical use time favors using material B because the return for B is
higher than that for A (Figure9.1b). This relationship makes it such that, as use
time increases, it pays to invest in the material that increases returns.
It is important to note that although this model acknowledges that techno-
logical decisions can be guided by subsistence needs, they are not necessarily so
guided because increased tool use for any purpose (ceremonial, for instance)
influences raw material selection. That is, at a certain level of use, having the
most effective tool (i.e., the one with the highest rate of return, however
returns are measured) outweighs the cost of procuring hard-to-obtain raw
materials.
Model Predictions
Based on the models parameters, at least two scenarios could effect a change
in the selection of raw materials. The first involves an increase or decrease in
tool use time, as described earlier and depicted in Figures9.1a and 9.1b. Tool
use times in excess of the critical use time favor higher-quality raw materials
such that resource intensification, for example, should make procuring them
worth it once the use time threshold is breached. Conversely, decreased time
at a task should favor the use of lower quality, local materials.
A second factor that could influence stone selection is a change in the costs
associated with procurement or manufacturing. Procurement costs could
change if a groups mobility pattern changes, bringing the group closer to or
drawing it farther from a source. Procurement costs might also change as stone
sources are used intensively or over long periods, making it more difficult to
locate portions of suitable size and quality within a source. Figure 9.1c illustrates
the scenario in which changed mobility has brought a group closer to a high-
quality and previously hard to obtain source, B. The resultant reduction in the
cost of obtaining material B redefines the critical use threshold that would trig-
ger a change in procurement. In this case, we would expect a higher frequency
of this material at a lower intensity of tool use.
An Application
This model offers a compelling explanation for changed patterns of stone use
observed in the archaeological record of the Atuel River drainage in south-
ern Mendoza Province, Argentina (Figure 9.2). Local archaeologists report
(a) R
E
T
U
R
N
use B C use A
(b)
C C
(c)
B B
C C
use procurement & manufacturing
TIME
9.1. (a) Graphical depiction of the procurement model showing the critical use time, C, beyond
which threshold a tool user will achieve a higher return rate by switching from lower-cost but
lower-quality material A to higher-cost, higher-quality material B.The dimension time is divided
into procurement and manufacturing time to the right of the origin, and use time to the left of
the origin. Return rates lie along the y-axis. (b) Use times in excess of the critical use time will
always favor material B. (c) The hypothetical scenario in which changed mobility has brought a
group closer to high-quality and previously hard to obtain material B, reducing its procurement
and manufacturing costs (from B to B). The changed procurement and manufacturing costs of
material B redefine the critical use time (from C to C) that would effect a change in material
procurement. (Adapted from Bettinger et al. 2006.)
163
164 Raven G arve y
72 W 70 W 68 W
32 S
Mendoza
Santiago
34 S
San Rafael
36 S
N Pacific
Ocean Chile
0 km 100
Argentina
9.2.The Atuel River drainage, Mendoza Province, Argentina.
a dramatic reduction in archaeological sites in parts of southern Mendoza

during the middle Holocene, between 8000 and 4000 BP (Gil et al. 2005).
Paleoenvironmental data from a number of regions worldwide indicate that
this same four thousand year period was a time of variable and generally more
arid climatic conditions, during which encounter rates with high-ranked sub-
sistence resources would have been reduced (e.g., Antevs 1948; Grayson 1993;
Meltzer 1999; Sheehan 1994, 2002). Lithic collections from the few known
sites with middle Holocene deposits indicate a dramatic shift in raw mate-
rial use from basalts in earlier levels to obsidians in later levels (Garvey 2012;
Neme 2007; Neme et al. 2011), and the change appears to track the large-
scale climate shifts of the Holocene. Here, data derived from middle Holocene
sites in southern Mendoza are compared to the predictions of the stone use
model, and these predictions are combined with those of the marginal value
theorem (Charnov 1976) to tease apart complex behavioral responses to mid-

dle Holocene climatic stimuli.
Argentine researchers have obtained 93 radiocarbon dates associated with
human activity in southern Mendoza (Gil etal. 2005). Although the middle
Holocene interval accounts for 36 percent of the time since initial occupa-
tion, only 13 percent (N = 12)of the radiocarbon dates fall between 8000 and
4000 BP. It remains unclear whether the paucity of middle Holocene sites
in Mendoza indicates a dramatic population reduction, whether sites of this
period are simply not well preserved or visible, or whether people reorganized
their settlement patterns to exploit different biotic communities.
In simplest form, the marginal value theorem (Charnov 1976) predicts that,
because the total amount of energy available in a given patch diminishes as
foragers catch and consume available resources, beyond a critical threshold the
optimizing forager should leave the patch or face starvation.The optimal point
of departure is determined by the amount of energy available in the envi-
ronment as a whole and the distribution of patches on the landscape. When
resources are relatively abundant in an environment, optimizing foragers will
make correspondingly frequent moves to new resource patches. When, due
to environmental or demographic changes, the number of available resource
patches is reduced (and holding all other factors constant), optimizing foragers
should remain in a given patch longer and extract a greater proportion of its
resources before moving to a new one (Bettinger 1991; Charnov 1976).
The behaviors predicted by this model may account for the sparsity of mid-
dle Holocene sites in Mendoza Province. Prior to 8000 BP, when population
densities appear to have been low, foragers may have been highly mobile, tar-
geting only the highest-ranked species, moving frequently in pursuit of them,
but not having to move far to find the next resource patch. If patches were
centered on water resources (e.g., springs or river margins) as they are likely to
have been in this semi-arid environment, and the number of available patches
was reduced during middle Holocene droughts, foragers may have responded
by moving less often but to more distant patches, exploiting resources around
water sources until within-patch foraging returns were low enough to neces-
sitate incurring inter-patch travel costs. Thus, the observed decrease in archae-
ological sites may not be a product of reduced populations, per se, but of a
reduction in the overall number of sites, each occupied for a longer duration.
This scenario can be elaborated to include other parameters that likely con-
tributed to differences in settlement patterns before, during, and after the dry-
ing trend (i.e., the early, middle, and late Holocene). Biotic zones assume two
basic configurations in southern Mendoza. In the Andes and, to a lesser degree,
adjacent foothills, resource zones are vertically stratified and change quickly
with elevation. In the piedmont and plains east of the Andes, resource zones
are arranged horizontally. All of the middle Holocene sites known to date
are located between 1500 and 2500 m above sea level, in the upland valleys
166 Raven G arve y
of major rivers including the Atuel and its tributaries. If middle Holocene
Mendocinos were bound to water resources as people of other regions appear
to have been during this time (e.g., in the Great Basin; Jones et al. 2003),
then it is possible that their settlement followed a roughly west-to-east pat-
tern, tracking seasonally available resources from the upland river valleys in
the west to the rivers lower courses to the east. The vertical arrangement of
Mendozas western biomes may have made foraging nearer obsidian sources in
the Andes unnecessary. So, in the early Holocene, when population levels were
low and resources abundant, foragers moved often, but not far. In the middle
Holocene, when populations were still low (but not necessarily lower than
before) and resources scarce, foragers moved less often but longer distances
and the moves may have been tied to water resources. In the late Holocene,
resources rebounded, but human populations also increased, outstripping local
resource availability and forcing people to move into previously unoccupied
patches and to intensify subsistence resources.
The predictions of the lithic procurement model can be coupled with
this hypothesis for a more complete interpretation of behavioral responses to
middle Holocene climatic events. In southern Mendoza, the two most fre-
quently occurring raw material types, basalt and obsidian, are distinct in their
associated costs and returns. Much of the underlying geology in Mendoza
is basaltic (Rodrguez and Ragairaz 1972) and the ubiquity of basalts rela-
tive to archaeological sites makes their procurement cost low. However, many
Mendozan basalts are grainy, producing serviceable tools but ones that may
be relatively hard to craft and maintain, for example. Obsidians, on the other
hand, make fine tools that are easily maintained, but their procurement cost
is higher because sources are localized, many of them high in the Andes, and
often distant from known archaeological sites. The relationship between these
two material types, their respective return rates relative to the time it takes to
procure them and craft tools from them, defines the critical use time, the amount
of time that must be devoted to a task to make obsidian viable in the presence
of lower quality but more readily available basalt.
Because there is a measurable difference in the return rates associated with
Mendozan basalts and obsidians, their relative abundance in archaeological
deposits may serve as a proxy measure for other behaviors. Given the more
restricted movement predicted by the marginal value theorem, we should
expect local stone to dominate middle Holocene lithic assemblages in the Atuel
drainage for two reasons. First, with restricted movement, both the opportu-
nity and the absolute costs of resource excursions to the north and south of the
Atuel River Valley may have been too high to justify trips to distant obsidian
sources. Second, restricted movement implies a widened diet breadth, incor-
porating more varied resources including small game, seeds, and plant foods
(Jones etal. 2003). Accordingly, procurement was likely too generalized during
the middle Holocene to warrant use of the high-cost, high-return obsidians

one suspects were needed for specialized tools (Bettinger etal. 2006).
If populations grew after the environment stabilized in the late Holocene,
people would have been forced into previously unoccupied areas and foraging
radii may have extended farther into the Andes where high-quality obsid-
ian sources are located, thereby decreasing obsidian procurement costs, as in
Figure9.1c. That is, obsidian procurement could have become embedded in
food resource forays. A shift in stone use might also reflect a change in sub-
sistence. After the middle Holocene, expanded human populations may have
outstripped rebounding resources, thereby increasing resource competition,
requiring technological intensification and triggering a shift from basalt to
obsidian. Thus, changing settlement patterns that reduced obsidian procure-
ment costs, intensified food resource procurement that heightened the need for
high-quality raw materials, or some combination of these two factors would
have favored the use of obsidian after the middle Holocene, but not sooner.
Discussion and Conclusions

Granting that the data available for this analysis are currently few, using an evo-
lutionary framework to assess the scarcity of middle Holocene sites in southern
Mendoza provides empirically testable predictions and offers an alternative to
deterministic accounts of prehistory. The environments in which people lived
are sure to have influenced their behaviors, and the present application clearly
has an environmental component. Significantly, however, the models applied
to the Mendoza scenario are based on the assumption that human action
and decision-making are important and powerful counters to environmentally
induced resource fluctuations. It also bears repeating that this model of raw
material procurement is indifferent to the purpose a tool is put to; increased
ceremonial use should have the same influence on raw material selection as
increased subsistence use.
Some might argue that viewing stone technologies in terms of evolutionary
fitness trades environmental for biological determinism. Evolutionary explana-
tions need not be deterministic, however. Evolutionary anthropologists build
tractable models as though people are purely rational with the expectation that
ones observations will deviate from the models prediction. It is by this method
that we hope to learn about behavioral variation and our reductionism is an
intentional and appropriate methodology designed to isolate variables that are
central to a particular outcome (Friedman 1953:36), reducing the unnumbered
complexities of reality to a tractable number of abstractions (Winterhalder and
Smith 1992).
Studies of embeddedness and stone economizing behaviors are clearly
important for understanding landscape and resource use. Nonetheless, the
168 Raven G arve y
model outlined here offers some potential advantages. First, it eliminates the
need for multiple descriptions of stone procurement that require detailed
knowledge of subsistence behavior and prevailing environmental conditions.
This model of stone procurement can also generate hypotheses regarding the
relative amounts of time people were devoting to particular activities, which
can be combined with other models predictions to interpret complex archae-
ological records. Finally, understanding the relationship between use times and
returns for particular materials should improve our use of optimal foraging
models since resource ranking can change dramatically with changes in tech-
nology (Bettinger 1991, 1999).
Truly informative models of behavior are those that are as applicable to
hunter-gatherers as to more complex societies, to prehistoric knappers as well
as modern engineers. No one model can usefully address all behaviors, but a
given model, if it is to successfully articulate empirically observed phenomena
with more general theories of human behavior, should address all phenomena
of a particular kind. The model presented here is one such attempt to under-
stand material procurement decisions because embedding, although perhaps
ideal, can happen only under certain circumstances.
Acknowledgments
This paper benefited greatly from the comments of Robert Bettinger,
Jelmer Eerkens, Mark Aldenderfer and Bruce Winterhalder. Research abroad
was funded by the J. William Fulbright Foundation, the National Science
Foundation and the Department of Anthropology, University of California,
Davis. Permission to study collections housed at the Museo de Historia Natural,
San Rafael was kindly granted by Humberto Lagiglia. Many thanks to Adolfo
Gil and Gustavo Neme, who provided invaluable advice and logistical support
during my stays in Argentina.
References
Andrefsky, William, Jr. 1994. Raw Material Bamforth, Douglas B. 1990. Settlement, Raw
Availability and the Organization of Material, and Lithic Procurement in the
Technology. American Antiquity 59:2135. Central Mojave Desert. Journal of Anthropological
Andrefsky, William. 1998. Lithics: Macroscopic Archaeology 9:70104.
Approaches to Analysis. Cambridge University Basgall, Mark E. 1987. Resource Intensification
Press, Cambridge. among Hunter-Gatherers: Acorn Economies
Antevs, Ernst. 1948. Climatic Changes and Pre- in Prehistoric California. Research in Economic
white Man. The Great Basin, with Emphasis Anthropology 9:2152.
on Glacial and Postglacial Times. University of Basgall, Mark E. 1989. Obsidian Acquisition and
Utah Bulletin 30(20):168191. Use in Prehistoric Central Eastern California:
Bamforth, Douglas B. 1986. Technological A Preliminary Assessment. In Current Directions
Efficiency and Tool Curation. American in California Obsidian Studies, edited by R.
Antiquity 51:3850. Hughes, pp. 111126. Contributions of the
University of California Archaeological Broughton, Jack M. 1997. Widening Diet

Research Facility48. Breadth, Declining Foraging Efficiency, and
Beck, Charlotte, Amanda K. Taylor, George Prehistoric Harvest Pressure: Ichthyofaunal
T. Jones, Cynthia M. Fadem, Caitlyn R. Evidence from the Emeryville Shellmound,
Cook, and Sara A. Millward. 2002. Rocks California. Antiquity 71:845862.
Are Heavy: Transport Costs and Paleoarchaic Callahan, Errett. 1979. The Basics of Biface
Quarry Behavior in the Great Basin. Journal of Flintknapping in the Eastern Fluted Point
Anthropological Archaeology 21:481507. Tradition: A Manual for Flintknappers and
Bettinger, Robert L. 1991. Hunter-Gatherers: Lithic Analysts. Archaeology of Eastern North
Archaeological and Evolutionary Theory. Plenum America 7(1):1180.
Press, NewYork. Charnov, Eric L. 1976. Optimal Foraging: The
Bettinger, Robert L. 1999. What Happened in Marginal Value Theorem. Theoretical Population
the Medithermal? In Models for the Millennium: Biology 9:129136.
Great Basin Anthropology Today, edited by C. Cotterell, Brian, and Johan Kamminga. 1979.
Beck, pp. 6274. University of Utah Press, Salt The Mechanics of Flaking. In Lithic Usewear
LakeCity. Analysis, edited by B. Hayden, pp. 97112.
Bettinger, Robert L., and Martin A. Baumhoff. Academic Press, NewYork.
1982. The Numic Spread: Great Basin Dunnell, Robert C. 1980. Evolutionary
Cultures in Conflict. American Antiquity Theory and Archaeology. In Advances in
47:485503. Archaeological Theory and Methods,Vol. 3, edited
Bettinger, Robert L., Ripan Malhi, and Helen by M. B. Schiffer, pp. 3599. Academic Press,
McCarthy. 1997. Central Place Models of NewYork.
Acorn and Mussel Processing. Journal of Eerkens, Jelmer W., Jeffrey R. Ferguson, Michael
Archaeological Science 24:887899. D. Glascock, Craig E. Skinner, and Sharon
Bettinger, Robert L., Bruce Winterhalder, and A. Waechter. 2007. Reduction Strategies and
Richard McElreath. 2006. A Simple Model Geochemical Characterization of Lithic
of Technological Intensification. Journal of Assemblages: A Comparison of Three Case
Archaeological Science 33:538545. Studies from Western North America.
Binford, Lewis R. 1973. Interassemblage American Antiquity 72:585597.
Variability the Mousterian and the Eerkens, Jelmer W., Amy M. Spurling, and
Functional Argument. In The Explanation Michelle A. Gras. 2007. Measuring Prehistoric
of Culture Change: Models in Prehistory, edited Mobility Strategies Based on Obsidian
by C. Renfrew, pp. 227254. Duckworth, Geochemical and Technological Signatures
London. in the Owens Valley, California. Journal
Binford, Lewis R. 1979. Organization and of Archaeological Science, DOI: 10.1016/j.
Formation Processes: Looking at Curated jas.2007.05.016.
Technologies. Journal of Anthropological Research Friedman, M. 1953. Essays in Positive Economics.
35:255273. University of Chicago Press, Chicago.
Bleed, Peter. 1986. The Optimal Design of Funk, Robert E. 1972. Early Man in the
Hunting Weapons: Maintainability or Northeast and the Late Glacial Environment.
Reliability. American Antiquity 51:737747. Man in the Northeast 4:739.
Brantingham, P. Jeffrey. 2003. A Neutral Model of Gardner, William. 1977. The Flint Run Paleo-
Stone Raw Material Procurement. American Indian Complex and Its Implications
Antiquity 68:487509. for Eastern North American Prehistory.
Brantingham, P. Jeffrey, John W. Olsen, Jason In Amerinds and Their Paleo-environments
A. Rech, and Andrei I. Krivoshapkin. 2000. in Northeastern North America, edited by
Raw Material Quality and Prepared Core W. Newman and B. Salwen, pp. 257263.
Technologies in Northeast Asia. Journal of Annals of the New York Academy of
Archaeological Science 27:255271. Sciences288.
170 Raven G arve y
Garvey, Raven. 2012. Human behavioral responses Kelly, Robert L., and Lawrence C. Todd. 1988.
to middle Holocene climate changes in north- Coming into the Country: Early Paleoindian
ern Argentine Patagonia. Dissertation on file, Hunting and Mobility. American Antiquity
University of California-Davis. 53:231244.
Gil, Adolfo F., Marcelo A. Zrate, and Kuhn, Stephen L. 1991.Unpacking Reduction:
Gustavo A. Neme. 2005. Mid-Holocene Lithic Raw Material Economy in the
Paleoenvironments and the Archaeological Mousterian of West-Central Italy. Journal of
Record of Southern Mendoza, Argentina. Anthropological Archaeology 10:76106.
Quaternary International 132:8194. Kuhn, Stephen L. 1992. Blank Form and Reduction
Goodyear, Albert C. 1989. A Hypothesis of the as Determinants of Mousterian Scraper
Use of Cryptocrystalline Raw Materials Morphology. American Antiquity 57:115128.
among Paleoindian Groups of North America. Kuhn, Stephen L. 1994. A Formal Approach to
In Eastern Paleoindian Lithic Resource Use, the Design and Assembly of Mobile Toolkits.
edited by C. J. Ellis and J. C. Lothrop, pp. 19. American Antiquity 59:426442.
Westview Press, Boulder,CO. Kuhn, Stephen L. 1995. A Perspective on
Gould, Richard A., and Sherry Saggers. 1985. Lavelleois from a Non-Lavellois: Assemblage:
Lithic Procurement in Central Australia: the Mousterian of Grotts di Sant Agostino.
A Closer Look at Binfords Idea of In The Definition and Interpretation of Lavellois
Embeddedness in Archaeology. American Technology, edited by H. Dibble and O. Bar-
Antiquity 50:117136. Yosef, pp. 157170. Prehistory Press, Madison.
Grayson, Donald K. 1993. The Deserts Past: A Madsen, David B., and Dave N. Schmitt. 2003.
Natural Prehistory of the Great Basin.Smithsonian Mass Collecting and the Diet Breadth Model:
Institution Press, Washington,DC. a Great Basin Example. Journal of Archaeological
Hester, James J., and James Grady. 1977. Paleo- Science 25:445455.
Inidan Social Patterns on the Llano Estacado. McGuire, Kelly. 2002. Obsidian Production
In Paleo-Indian Lifeways, edited by E. Johnson. in Northeastern California and the
The Museum Journal 17:7896. Northwestern Great Basin: Implications for
Hildebrandt, William and Kelly McGuire. 2002. Land Use. In Boundary Lands: Archaeological
The Ascendance of Hunting During the Investigations Along the California-Great Basin
California Middle Archaic: An Evolutionary Interface, edited by K. McGuire, pp. 85103.
Perspective. American Antiquity 67:231256. Anthropological Papers No. 24. Nevada State
Jeske, Robert J. 1992. Energetic Efficiency and Museum, CarsonCity.
Lithic Technology: An Upper Mississippian Meltzer, David J. 1999. Human Response to
Example. American Antiquity 57:467481. Middle Holocene (Altithermal) Climates on
Jones, Emily L. 2004. Dietary Evenness, the North American Great Plains. Quaternary
Prey Choice and Human-Environment Research 52:404416.
Interactions. Journal of Archaeological Science Neme, Gustavo. 2007. Cazadores-Recolectores
31:307317. de Altura en los Andes Meridionales: El Alto
Jones, George T., Charlotte Beck, Eric E. Jones, Valle del Ro Atuel. British Archaeological
and Richard E. Hughes. 2003. Lithic Source Reports International Series, 1591.
Use and Paleoarchaic Foraging Territories in Neme, G. A. Gil, R. Garvey, C. Llano, A.F.
the Great Basin. American Antiquity 68:238. Zangrando, F. Franchetti, C. deFrancesco and
Kelly, Robert L. 1988. The Three Sides of a C. Michieli. 2011. El registro arqueolgico
Biface. American Antiquity 53:717734. de la Gruta de El Manzano y sus implican-
Kelly, Robert L. 1995. The Foraging cias para la arqueologa de nordpatagonia.
Spectrum: Diversity in Hunter-Gatherer Magallana 39:243265.
Lifeways. Smithsonian Institution Press, Rodrguez, Eduardo J., and Alberto C. Regairaz.
Washington,DC. 1972. Resumen geolgico de la provincia de
Mendoza. In Revista de la Sociedad Argentina de Ugan, Andrew, Jason Bright, and Alan Rogers.
Botnica, Number 13 (supplement), pp. 513. 2003. When Is Technology Worth the
Mendoza. Trouble? Journal of Archaeological Science
Sheehan, Michael S. 1994. Cultural Responses to 30:13151329.
the Altithermal: The Role of Aquifer-Related Waguespack, Nicole M., and Todd A. Surovell.
Water Resources. Geoarchaeology 9:113137. 2003. Clovis Hunting Strategies, or How to
Sheehan, Michael S. 2002. Dietary Responses Make Out on Plentiful Resources. American
to Mid-Holocene Climate Change. North Antiquity 68:333352.
American Archaeologist 23:117143. Whittaker, John C. 1994. Flintknapping: Making
Surovell, Todd A. 2009. Toward a Behavioral and Understanding Stone Tools. University of
Ecology of Lithic Technology: Cases from Texas Press, Austin.
Paleoindian Archaeology. University of Winterhalder, Bruce, and Eric A. Smith. 1992.
Arizona Press, Tucson. Evolutionary Ecology and the Social Sciences.
Torrence, Robin. 1989. Retooling: Towards a In Evolutionary Ecology and Human Behavior,
Behavioral Theory of Stone Tools. In Time, edited by E.A. Smith and B. Winterhalder, pp.
Energy and Stone Tools, edited by R. Torrence, 324. Aldine de Gruyter, NewYork.
pp. 5766. Cambridge University Press,
Cambridge.
Ten
Artifacts as Patches: The Marginal

Value Theorem and Stone Tool Life
Histories
Steven L. Kuhn and D. Shane Miller
Human Behavioral Ecology, Technological Organization,

and the Study of Lithic Technology
North American researchers interested in explaining technological varia-
tion and change as a consequence of adaptive problem solving have gravi-
tated toward two conceptual approaches: what is broadly termed, following
Nelsons (1991) terminology, the study of Technological Organization
(TO); and Human Behavioral Ecology (HBE). Although they have some-
what different intellectual foundations, the study of technological organiza-
tion and human behavioral ecology actually share many fundamental goals
and presuppositions.
Both seek to understand human behavior in terms of economic constraints and
payoffs, costs and benefits.
Both assume that behavioral alternatives that are closer to optimal will tend to
become more common over time, holding conditions constant (all other things
being equal).
Both focus on energy or time as key currencies in understanding technological
behavior.
Both assume that variation in technological behavior is not directly subject to
natural selection but reflects flexibility in the human behavioral phenotype. HBE
goes a step further, assuming that the cognitive apparatus that governs decision
making and variation in the behavioral phenotype is under selection, and that
172
Arti facts as Pat che s 173
selection favors cognitive mechanisms that more effectively arrive at optimal

solutions (see Shennan 2008), a proposition also known as the phenotypic gambit
(Grafen 1984; Smith and Winterhalder 1992:33). TO is silent on this matter: it
assumes economic rationality but does not inquire as to its origins.
Both have addressed questions of artifact design and technological investment.
There are some differences between these approaches as well, but these are
relatively unimportant compared to the commonalities. Researchers work-
ing with concepts from HBE tend to couch models in terms of fitness ben-
efits, whereas studies of TO are more explicitly concerned with time, energy,
and risk. However, as energetic or temporal efficiency normally substitute for
direct measures of fitness in HBE studies this is of comparatively little conse-
quence. HBE has also tended to favor quantitative models, whereas TO studies
rely more on narrative models (but see Brantingham and Kuhn 2001; Kuhn
1994). This is more a matter of style than a reflection of deep intellectual dif-
ferences. With specific reference to lithic technology, researchers interested in
HBE have focused mainly on artifact design and technological investment
(e.g., Bettinger et al. 2006; Bright et al. 2002; Ugan et al. 2003). Studies of
technological organization have addressed a wider range of phenomena. In
addition to artifact design and investment in technology,TO has been particu-
larly concerned with raw material economy and artifact life histories. HBE
models have been infrequently applied to the latter issues (cf. Beck etal. 2002;
Metcalfe and Barlow 1993).
One limitation of HBE-based studies of technology to date has been this
tendency to focus on artifact design and technological investment. Although
this is an area ripe for theorization, its applications to the archaeological
record are limited. First, it is no simple task to determine the true costs of
prehistoric artifacts. Yet the costs of even very effective tools, such as nets
(Lupo and Schmidt 2002; Ugan etal. 2003), may be high enough to make
them disadvantageous in many conditions. It is even more difficult to assess
artifacts contextual advantages. Too often, researchers rely on the danger-
ous rule of thumb that later-developing forms of technology are necessarily
more effective. Thus, for example, it has long been assumed that prismatic
blade technologies are more efficient than other methods of blank produc-
tion, and that this explained why blades became the dominant tool blanks in
the Eurasian Upper Paleolithic. However, comparative experiments suggest
the differences in at least some measures of efficiency are small or negligible
(Eren etal. 2008; see also Prasciunas 2007 concerning bifaces and informal
flake cores).
More importantly, the appearance of a novel technology or a new arti-
fact form requiring radically different levels of investment is a rare event in
prehistory. The farther one goes back into the past, the rarer these events
174 Steven L . Ku hn a n d D . S ha n e M i l l e r
become (and the more difficult it is to assess the costs and benefits of
technological alternatives).Thus, although models of investment and artifact
design may be very useful for explaining major technological transitions
and the spread of artifact complexes such as the bow and arrow or whal-
ing boats, there are long periods of time and large areas where they simply
are not relevant. In other words, they are best applied at particular scales of
analysis.
Whereas the lithic archaeological record is not necessarily rich in new
artifact designs, it is very rich in debris generated from the production, main-
tenance, and discard of artifacts. At many temporal scales, and in many con-
texts, the empirical strength of the record is in what it tells us about the
ways people managed the utility of raw material and artifacts as they foraged,
moved about the landscape, and conducted different activities. Researchers
working on technological organization have been pursuing these questions
for two decades or more. The challenge is to adapt the rich and productive
corpus of models from HBE to these most robust dimensions of the avail-
able archaeological data. Admittedly, we cannot demand of all theory that it
account for the most abundant elements of the record. On the other hand,
having an abundant and varied material record makes it all the better for test-
ing and refining theories.
This chapter is a preliminary attempt to close some of the gaps between
HBE and TO. It applies a model widely used in Behavioral Ecology (the
Marginal Value Theorem [MVT]) to stone tools, using it to develop qualitative
predictions about the life histories of artifacts under different conditions. The
MVT replicates some common notions from research on TO, but also makes
some novel predictions. These expectations are then applied to Paleoindian
projectile points from Tennessee.
TheMVT
The Marginal Value Theorem (or MVT) was originally introduced into behav-
ioral ecology by Charnov (1976). It has been employed widely in studies of
animal and, to a lesser extent, human behavior. The MVT seeks to under-
stand how long foragers should continue to feed from the same patch of
food, and when they should abandon the patch to search out another feed-
ing opportunity. A patch can be as small as a single food item, or as large as a
habitat zone. The MVT has been much elaborated, particularly with respect
to the information criteria that animals actually use to make decision (e.g.,
see Brown 1988; Brown and Mitchell 1989; McNair 1979, 1982; Nonacs 1981;
Pyke 1984; Stephens 2008; Stephens and Krebs 1986; Thiel and Hoffmeister
2004;Wajnberg etal. 2000;Wilke 2006:56102). Nonetheless, it remains ...the
dominant paradigm in predicting patch use... (Nonacs 2001:71).
energy
t
time to new patch time in patch
10.1.The marginal value theorem in graphic form. The x-axis indicates time values increase in
both directions from the intercept with the y-axis, which is a measure of energy returns. The
upward convex curve to the right of the y-axis represents instantaneous energy returns from a
patch over time. The tangent line that intercepts the x-axis represents the average instantaneous
energy returns for all patches in the environment. The x-intercept is equivalent to the time cost
of moving between patches.
The most widely used form of the MVT makes several assumptions:
The forager can operate in single patch at any one time (although this assumption
is modifiable).
Travel and foraging are mutually exclusive activities, so there is a direct
opportunity cost to moving off in search of a new patch.
Instantaneous returns from feeding diminish with increasing time spent in a
patch.
The forager can accurately assess instantaneous return rates.
The forager knows the average return rates for all patches in the environment.
The forager knows the average cost of getting to the next patch.
A commonly used and familiar graphic representation of the MVT is shown

in Figure10.1. The x-axis represents time, and the y-axis energy. The upward
convex curve on the right of the vertical axis represents instantaneous return
rates from foraging in a particular patch. The straight diagonal line represents
average returns from other patches: its slope is equivalent to average return
rates from foraging, while the x-intercept is the time cost of moving between
patches. This instantiation of the model optimizes for time: it results in a solu-
tion that produces the greatest returns per unit of foraging and travel time.The
optimal patch residence time is indicated by the point where the diagonal line
is tangent to the upward convex curve.
Even in its simplest instantiation, the MVT makes a number of predic-
tions relevant to understanding the foraging behavior of animals. It predicts
that foragers should give up/leave a patch when the instantaneous returns
drop below the average return rates for other patches, adjusted for the cost
of traveling to a new patch. As the average time needed to move to a new
patch increases, foragers should stay longer in patches. The third prediction
is somewhat less obvious: holding interpatch distances constant, as average
returns increase, patch residence time should decrease. This third prediction
has received less attention, perhaps because it is subject to the limiting assump-
tion that the shape of the time-decay curve for patch residence stays constant
even as average yields increase.
To date, the MVT has been employed successfully to analyze the feeding
behavior of animals exploiting spatially delimited resource patches, ranging
from clusters of resources to single trees to individual prey items. Modifications
to the model have examined such factors as the ways foraging animals can
evaluate instantaneous returns from a patch and the effects of being able
to exploit multiple patches simultaneously (e.g., McNair 1979, 1982, 1983;
Stephens and Krebs 1986). Despite many questionable assumptions, including
those about the quality of forager knowledge of the environment (Pyke 1984),
variants of the MVT are nonetheless quite successful at predicting the behavior
of invertebrates, birds, and mammals, at least at a semiquantitative level. The
MVT and other patch choice models have not been widely applied to human
foraging decisions (Kaplan and Hill 1992:178184; Wilke 2002; cf. OConnell
and Hawkes 1984; Sosis 2002; Thomas 2007). In part, this may be due to the
difficulty of obtaining long-term observations of human individuals or groups
moving among food patches.
Stone Tools as Patches

The MVT has been employed to study the extraction of energy from food
items or clusters of such items. An alternative perspective that has more imme-
diate relevance to lithic studies is to consider artifacts tool, cores, or pack-
ages of raw material as patches. They are not patches of energy or other
nutrients, but of utility. This utility comes from the fact that artifacts provide
a mechanical advantage that makes certain tasks possible or makes them more
time or energy efficient. In other words, although artifacts do not directly
supply energy or nutrients to users, by making work more efficient a usable
implement or weapon has the potential to produce a net energy gain for a
tooluser.
Treating artifacts as patches of utility conforms with many of the assump-
tions of the MVT. The amount of utility contained in a single artifact is finite,
and for many artifacts utility should decline over time as the artifacts are used
up, wear out, or approach a point of failure. The sizes and shapes of artifacts
that are modified through use or resharpening change over time, which may
render them less effective. Artifacts that are not modified with use may still
accumulate microscopic damage through use (structural fatigue), increasing
the likelihood of their failing catastrophically. Assumptions about knowledge
actually fit better with people using tools than they do with animals forag-
ing in patches. People do know how well their tools are working, and their
experience tells them how long the tools are likely to last. People also know
how much better a new artifact would work, and how much time or effort is
needed to produceone.
So, if we accept the notion that artifacts can be viewed as discrete patches
of utility, the MVT can certainly be applied to them. The question here is
how long people should use a core or tool of declining effectiveness or utility
before they replace it (equivalent to moving on to another patch). In this way,
the MVT and modified versions of it can be used to make predictions about
artifacts life histories and the conditions that influencethem.
Different Contexts, DifferentCosts

The typical application of the MVT seeks solutions that optimize foraging
returns over time.The cost of moving on to a new patch is quantified in terms
of travel or search time. This makes sense under the assumption that travel and
foraging are mutually exclusive, and foraging is more or less continuous: time
spent in searching for and traveling to a new food patch has a direct opportu-
nity cost in lost foraging opportunities, and vice versa.
This opportunity cost/time optimizing version of the MVT can be applied
directly to contexts where artifacts lose utility continuously, and where use
events are prolonged, and may even extend beyond the lifetime of a single
implement. In such cases, there is a direct trade-off in time costs between
continuing to use an artifact that provides less and less mechanical advantage
and taking time to search out or make a new one. Typical situations would be
extended bouts of activities such as butchering, hide processing, woodworking,
or stone knapping that cause artifacts to wear out rapidly. Surovell (2003, 2009)
has addressed just this kind of situation with great insight.
In other contexts there is a less direct trade-off between making and using
a tool. Some artifacts are used occasionally but intensively. They may fail or
wear out, but their entire life history will seldom play out in a single episode.
Examples would include many hunting weapons, and other artifacts that are
part of sporadically used, transported personal gear or similar kinds of toolkits.
Because episodes of use are discontinuous and brief, and because consider-
able time may elapse between them, replacing these kinds of artifacts does not
have the same sort of opportunity cost as in the first example. Gear such as
projectile points can be replaced during down time, rather than in the heat of
a hunt, avoiding opportunity costs directly related to the activity in which they
are employed. That does not mean that making a new artifact is without cost,
but it does imply that a currency other than time may be more appropriate.
If time is not limiting when it comes to artifact manufacture we have the
option of considering the cost of switching between artifacts/patches in terms
of energy. The energetic costs of stone artifacts are not necessarily high, but
they are significant. They derive from the effort expended in making artifacts,
but more importantly, from the effort expended in procuring appropriate raw
materials. Costs of obtaining materials, whether by exchange or direct procure-
ment, would be especially significant in applications, such as the manufacture
of large bifacial points, where stones with particular functional properties are
needed (Beck and Jones 1990; Goodyear 1989; Kelly and Todd 1988; Surovell
2009:128133). For artifacts that are parts of composite implements, such as
arrowheads, spearheads, or inset barbs, the costs of reworking or replacing
binding materials should be figured into costs or replacement (Keeley 1982).
In fact, both of these perspectives on optimizing cost are implicit in the
original MVT model. The energetic cost of switching patches contributes to
the y-intercept of the tangent line. It is just that in most cases, the y-intercept is
estimated as a function of distance between patches and average returns, rather
than some othercost.
A Modified MVT Model (Applied to Paleoindian Spear Points)

We argue that an energy-based variant of the MVT is the more appropriate
one for many types of artifacts that were used regularly but in brief, intense
applications. In the remainder of this chapter, we explore the model and apply
it to Paleoindian spear points from the Southeastern United States, an artifact
class that would fit these criteria. The assumptions of the model are similar to
those of the originalMVT.
Hunters effectively use a single point at one time, OR the size of the toolkit
does not change.The ability to exploit several patches simultaneously does affect
predictions about residence time (McNair 1979) but if the number of artifacts
(patches) is held constant it should not matter.
The utility of artifacts (their effectiveness, or the probability that they can be
employed successfully) declines with successive uses.
People monitor artifact effectiveness or condition.
The cost of replacing artifacts is a combination of raw material costs and
production effort, and tool users are aware of those costs.
Each of these assumptions is reasonable, but each could equally be chal-

lenged.The second in particular deserves further discussion.That tools can lose
utility as they are used over and over is immediately evident for artifacts that
change shape through wear or through resharpening. It is reasonable to assume
a
Instantaneous return
Time in Use/Number of Uses

10.2. Range of hypothetical artifact utility trajectories, showing different trajectories for loss in
utility of effectiveness with time in use or number ofuses.
that as projectile points are resharpened, as edges and tips become blunter, they
would be less useful as hunting weapons. However, not all artifacts show pro-
gressive damage or are reworked. Even here, however, the probability of fail-
ure may still increase over time as structural defects accumulate. In fact, many
artifact forms show increasing probability of failure with age due to material
fatigue (Kurtz 1930; Shott and Sillitoe 2004, 2005).1
Unfortunately few data are available on failure rates in stone projectile
points. A recent study (Cheshier and Kelly 2006) of arrowheads fired into a
flesh target shows equal probability of failure, and very high rates of failure,
with successive shots. However, these are small objects fired at comparatively
high speeds, so they are more likely to fail catastrophically when their overall
tolerances are exceeded. For the sake of the present argument, we assume that
larger Paleoindian spear points, which impacted targets at slower speeds, and
that frequently do show signs of reworking, show a more gradual loss of utility
after everyuse.
Although the utility of many artifacts decays or declines with successive
uses, the shapes of the decay curves may vary.These shapes can be grouped into
different families (Shott and Sillitoe 2004), but for the purpose of this discus-
sion they can also be treated as ends of a continuum. At one extreme are arti-
facts that exhibit gradually accelerating loss of utility or failure rates, such that
utility is highest when the artifact is new, but decreases at an accelerating rate
as the artifact ages.This produces an upward convex utility curve (Figure10.2).
At another extreme would be artifacts that tend to fail frequently early on, or
E
Instantaneous return
t1 t2
Time in Use/Number of Uses
10.3. Reformulated MVT. The two solid lines are different hypothetical curves for artifact
instantaneous yields from Figure10.2. E = criterion for artifact abandonment = expected yields
for replacement artifacts minus amortized replacement costs, divided by number of use events.
Optimal points of artifact abandonment (t1, t2) occurs when immediate returns drop below
expected yields for replacement artifacts.
that lose much of their utility in the first few applications. This kind of infant
mortality characterizes some living systems and material objects subject to
hidden manufacturing defects. In the middle we can imagine a straight curve
that utility/increasing probability of failure over time (Figure10.2). Whether
this curve describes actual artifacts is an open question. However, it does repre-
sent the middle of the range of possible life-history curves. Artifacts may also
exhibit sigmoid utility loss curves, in which a sharp initial drop in utility is
followed by a period of decreased but gradually accelerating decline in effec-
tiveness or usefulness. However, it is important to emphasize that the general
predictions of the model are not affected so long as the relationship between
instantaneous yield (or utility) and time is always negative.
In this reformulation of the MVT, we seek to calculate the optimal point
to abandon an artifact, given the number of times it has already been used, or
its resulting condition. Figure10.3 is a graphic representation of the model.
The three curves represent instantaneous energy returns from use of the arti-
fact, following different artifact life-history trajectories. The criterion value E
is equivalent to the average expected instantaneous returns for other similar
artifacts, adjusted for the cost of producing or procuring a new artifact. The
value of E increases as the applications in which the artifacts are used become
more profitable, or as production costs decline. Conversely, the value of E
declines as profitability decreases or costs increase. Cost is calculated in energy

rather than time, and time is calibrated in terms of the number of uses.This last
modification makes sense for stone tools at least, which do not decay unless
used, and has the added benefit that it may be archaeologically measurable.
For simplicity, the maximum life histories of artifacts have been made equiva-
lent nothing lasts more than 20 uses, so that the only difference is the rate at
which utility is lost. The model has also been instantiated in numerical form
in Microsoft Excel.
The optimal point at which to abandon an artifact would be when its
instantaneous yield or potential utility drops below the criterion value E, the
average yield/utility of other available artifacts, adjusted for the cost of produc-
ing or obtaining a replacement, or the point on the x-axis where the hori-
zontal dashed lines intersect the various curves.2 In the original application
of the MVT, decisions to abandon patches were based on average yields for
all patches in the environment, but here we make it an independent variable.
The reason is that humans do not choose new tools at random, but presum-
ably select artifacts in good condition. We might choose to use the average
yield for artifacts across their entire lifetimes, but we could use another value.
We could also set this criterion at the instantaneous yield of a new artifact,
minus cost of production, although this seems unrealistically high and carries
the added disadvantage of confining analysis to situations in which production
costs are recouped in a single use. Alternatively, we can set the criterion at an
arbitrary number of uses. Finally, we can calculate an intrinsic optimal point
of abandonment for any specific utility curve. Fortunately, all four alternatives
have similar consequences the general behavior of the model they differ only
in the specific predictions for the optimal numbers of uses. As these predic-
tions are purely hypothetical, and concern relative effects of different variables
on optimal artifact use lives, the differences are not important for the present
discussion.
Sample results from the Excel model are presented in Figures10.4 and 10.5,
utilizing two discard criteria, one based on whole-life average returns and the
second on the average for the first 10 uses. Some of the results are obvious.
As costs increase (from top to bottom in the three plots), the optimal number
of uses increases, meaning that more expensive artifacts should be kept in use
longer. Faster loss of utility (linear vs. upward convex curves) also favors earlier
abandonment of artifacts. Both conclusions conform to common sense views
of how people should treat their tools.
The third result is less intuitively obvious. For both utility curves, the opti-
mal point to abandon an artifact declines as average returns increase. In other
words, as optimizing tool users stand to gain more from the use of particular
artifacts, they should replace them with new ones more often. Conversely, if
average yield from application of the tools goes down, people should hold on
(a) 17.5
Optimal point to abandon 15.0

(No. of uses)
12.5
10.0
7.5
5.0
10 15 20 25 30
(b) 17.5
15.0
Optimal point to abandon
(No. of uses)
12.5
10.0
7.5
5.0
10 15 20 25 30
(c) 17.5
15.0
(No. of uses)
12.5
10.0
7.5
5.0
10 15 20 25 30
Maximum Potential Yield
10.4. Optimal number of uses after which an artifact should be abandoned, as a function
of maximum potential yield and artifact cost. (a) Artifact cost = 10. (b) Artifact cost = 25.
(c) Artifact cost = 50. Criterion value for abandonment = average potential yield over entire
potential lifetime of artifact (20 uses)-cost.
(a) 17.5
15.0
(No. of uses)
12.5
10.0
7.5
5.0
10 15 20 25 30
(b) 17.5
15.0
(No. of uses)
12.5
10.0
7.5
5.0
10 15 20 25 30
(c) 17.5
15.0
(No. of uses)
12.5
10.0
7.5
5.0
10 15 20 25 30
Maximum Potential Yield
10.5. Optimal number of uses after which an artifact should be abandoned, as a function
of maximum potential yield and artifact cost. (a) Artifact cost = 10. (b) Artifact cost = 25.
(c) Artifact cost = 50. Criterion value for abandonment = average potential yield over first 10
uses of artifact-cost.
to artifacts longer. The effects of return rates on optimal artifact lifetimes are
more apparent with higher artifact costs.
Although this third implication may initially be counterintuitive, it makes
sense. The cost of an artifacts failure is proportionate to the benefits gained by
its use. For activities with very high returns, the potential cost of failure is high
compared to the cost of replacing artifacts as they wear out and become less
effective or reliable. Conversely, as potential gains decline, the cost of replac-
ing tools eventually begins to outweigh the cost of their potential failure. In
fact, the original MVT has similar implications, in that it predicts that patches
should be abandoned earlier as overall yields increase.
In sum, this reformulation of the MVT to consider artifacts as patches of
utility has three principal implications. First, increasing cost increases optimal
time to abandonment. Second, the faster artifacts lose utility with successive
uses, the earlier they should be abandoned. Finally, all other things being equal,
increasing average yields favors earlier abandonment of tools that wear out
gradually. In sum, we can expect artifact abandonment to be conditioned by
cost of replacement, but also by potential gains. As returns decline, optimiz-
ing tool users ought to retain artifacts longer, and abandon them at a more
advanced stage of reduction.
We emphasize that these conclusions refer only to artifacts used in the same
or similar activities, with the same kinds of returns and the similar patterns in
loss of effectiveness. In other words, the model can be used to analyze the life
histories of projectile weapon tips or woodworking tools under different con-
ditions, but it would not be useful for comparing spear points with stone axes,
or flake scrapers with cores.
Background to Southeastern Paleoindian Point Database

Eastern North America has a long tradition of collaboration between avoca-
tional and professional archaeologists in the identification and recording of
Paleoindian bifaces, beginning in the 1940s with Ben McCarys (1947) inven-
tory of Paleoindian bifaces in Virginia. Subsequently, Anderson (1990a, 1990b)
began integrating several regional fluted point surveys into a centralized data-
base, which he used as the foundation for his staging area model of colo-
nization (Anderson 1995, 1996). The Paleoindian Database of the Americas
(PIDBA) is a continuation of this data set and can be easily downloaded via the
Internet (http://pidba.utk.edu) (Anderson etal. 2005, 2009).
Although the PIDBA database is best known as the basis for a series of artifact
distribution maps (Anderson and Faught 1998, 2000; Anderson etal. 2005), one
of the more underutilized aspects of this online resource is the metric attribute
data for more than 15,000 artifacts.This database incorporates several statewide
surveys that, as a rule, contain information such as maximum length, width,
and thickness; raw material; county-level provenience; and whether or not the
artifact is complete or broken. In many cases, researchers also include photo-
graphs or illustrations of the artifact, many of which are now available online as
well. One of the more successful examples, the Tennessee Fluted Point Survey,
began with Morse and colleagues (1964) survey of Paleoindian points in Smith
County and Guthes (1966a, 1966b, 1966c, 1983) subsequent efforts. Since 1988,
John Broster and Mark Norton of the Tennessee Division of Archaeology have
continued to enlarge the survey, which now consists of information on more
than 5400 bifaces from the Early Paleoindian through Early Archaic periods
(Broster and Norton 1996; John Broster personal communication).
In the southeastern United States, the Early Paleoindian period is marked
archaeologically by the presence of Clovis-type bifaces. These are large,
parallel-sided lanceolate bifaces with slightly concave bases and single or mul-
tiple flutes that rarely extend more than a third of the body (Howard 1990;
Justice 1995:1721; Sellards 1952). The Middle Paleoindian period (12,900
12,600 cal BP) is documented by much more regional variability in projectile
point styles (Anderson 2001:155; Anderson and Faught 1998). In the western
United States, Clovis bifaces are replaced by Folsom points, whereas in the
eastern United States point types such as Redstone, Cumberland, Suwannee,
and Gainey occur. In the mid-South, Redstone and Cumberland bifaces are
believed to immediately post-date Clovis. However, there is no single site
that clearly underwrites this proposed sequence (Anderson, etal. 1996:911;
Goodyear 1999:435). The Redstone type is defined by an overall triangular
form that is widest at the base. These points often have indented bases and
more extensive fluting than Clovis bifaces (Goodyear 2006; Justice 1995:22).
Cumberland bifaces are identified by their narrow, deeply fluted, and slightly
waisted appearance. In addition, the bases are slightly concave and often have
faint ears (Justice 1995:2527; Lewis 1954). The Quad and Beaver Lake point
types follow Cumberland in the chronological sequence. Beaver Lake points are
slightly waisted lanceolates with faint ears, slightly concave bases, and moderate
basal thinning (Cambron and Hulse 1975; Justice 1995: 3536). Quad points
have distinct ears, a concave base, and pronounced basal thinning (Cambron
and Hulse 1975; Justice 1995:3536).
The Late Paleoindian period is signaled by the appearance of the Dalton
point type (Goodyear 1982:390). This sequence is supported by excavations at
Dust Cave in northwestern Alabama, where Quad and Beaver Lake components
were found stratigraphically below the Dalton components (Sherwood, etal.
2004). Dalton bifaces typically begin life as lanceolates with concave bases and
serrated edges. Often the basal margins are parallel to slightly incurvate, while
the blade portion is initially excurvate (Justice 1985:4042). Several studies have
shown that the blade margins transition from excurvate to incurvate through
repeated resharpening (Goodyear 1982; Shott and Ballenger 2007). Another
Late Paleoindian point type, Greenbriar, is described as a lanceolate-bladed

expanding stem form that shares characteristics with Daltons and later notched
point types (Justice 1995:42). This includes resharpening patterns character-
istic of Dalton bifaces, although Greenbriar points also feature shallow side
notches.
The working hypothesis for this exploratory study of Paleoindian spear
points from Tennessee is that external conditions, either the cost of replacing
points or the yields from using them, changed over the Paleoindian period, and
that this will be reflected in the conditions of points at discard, in conforma-
tion with the model presented in the preceding text. The null hypothesis is
that there is no change over time, that all types show similar patterns of dis-
card. Because radiocarbon dates are so scarce in the region, we use the rough
temporal sequencing described previously, with Clovis earliest, followed by
Cumberland and Redstone, and with Beaver Lake, Dalton, and Greenbriar as
the most recent varieties.
Results
For this study we used only points classified as whole, meaning that they
retained both tip and an intact base. This does not mean that the specimens
are unmodified. In fact many of the points recorded had clearly been used,
resharpened, and/or reworked, sometimes extensively. It is clear both from
measurements and illustrations, as well as past studies, that some Paleoindian
points were in fact reworked or resharpened many times. This is famously
the case with Dalton points (e.g., Goodyear 1982; Shott and Ballenger 2007),
but many other forms show signs that they were modified over the course of
their times in use (e.g., Breitburg and Broster 1995). What is important for the
purposes of this study is that the specimens retain all their parts, such that
they are at least potentially serviceable. What prehistoric hunters considered
serviceable is another issue. We confined analysis to six types in the Tennessee
database that afforded sufficiently large samples of whole points: these are
Clovis, Cumberland, Redstone, Beaver Lake, Dalton, and Greenbriar.These six
types span the Early, Middle, and Late Paleoindian periods.
There are many ways to measure the effects of reworking and resharpening
on bifacial points (e.g., Andrefsky 2009; Buchanan 2006; Shott and Ballenger
2007). However, although a large number of measurements are available
for some artifacts in the PIDBA database, only a few are provided for most
specimens. To maximize sample size we use a very simple approach here. The
assumption is that spear points of a particular type will begin life with fairly
standardized shapes, so that there should be a high correlation between lengths
and widths of new points. However, as they are used, reworked, or resharpened,
these ratios should change. Because most wear and damage occurs at the tip,
table10.1. Results of Pearsons correla-

tions between length and body width for six
Paleoindian point types from Tennessee
Clovis 115
r2 0.508
p <0.001
Slope 2.61
Cumberland 46
r2 0.502
p <0.001
Slope 4.94
Beaver Lake 46
r2 0.057
p 0.07
Slope 1.23
Dalton 55
r2 0.027
p 0.160
Slope 0.66
Greenbriar 160
r2 0.833
p <0.001
Slope 0.057
Redstone 44
r2 0.237
p <0.001
Slope 2.335
length should decrease while width remains fairly constant: this effect will be
particularly strong if points are reworked in the haft. Thus, categories of point
that are more regularly and extensively reworked before discard should show
lower correlations between length and width.
Table 10.1 shows Pearsons correlations between length and width for
all six point varieties: corresponding scatterplots appear in Figure 10.6.
Table 10.2 shows summary statistics for three of the most commonly
recorded measurements of shape in the database: total length, body width,
and basal width. As expected, coefficients of variation for width measure-
ments are always lower than those for length. This is consistent with the
notion that length changes more over the use lives of artifacts than does
width, and that widths were more standardized, probably a consequence
of basic hafting requirements (e.g., Hoffman 1985; Shott 1997; Shott and
Ballenger 2007).
The strengths of the correlation vary markedly among point types. Clovis
and Cumberland show the strongest correlations between length and width
(Table10.1; Figure10.6a, b). Among three of the later types, correlations are
(a) 60 (b) 40
50
30
Body width
Body width
40
30
20
20
10 10
0 50 100 150 200 0 50 100 150 200
Length Length
(c) 35 (d) 35
30
30
Body width
Body width
25
25
20
15 20
30 40 50 60 70 80 90 100 30 40 50 60 70 80 90
Length Length
(e) 35 (f) 40
30
35
Body width
Body width
25
30
20
25
15
10 20
30 40 50 60 70 80 90 100 110 50 60 70 80 90 100 110 120 130
Length Length
10.6. Plots of length versus body width for complete fluted points from Tennessee. (a) Clovis;
(b) Cumberland; (c) Beaver Lake; (d) Dalton; (e) Greenbriar; (f) Redstone.
nonexistent (Beaver Lake, Dalton, Greenbriar) (Table 10.1, Figure 10.6ce).

There is a statistically significant correlation between length and width for
Redstone points, but the r2 value is substantially lower than for Clovis and
Cumberland. Moreover, the correlation appears to be largely the result of a
few very large specimens (Figure10.6f).
Table10.2. Descriptive statistics for basic shape measurements

for six Paleoindian point types from Tennessee
Max. length Body width Basal width

Clovis
Mean 76.9 29.0 25.9
Std. dev. 24.9 6.8 5.1
CV 0.32 0.23 0.20
Cumberland
Mean 80.2 23.6 21.1
Std. dev. 23.2 3.4 2.7
CV 0.29 0.14 0.13
Beaver Lake
Mean 62.1 23.5 22.9
Std. dev. 13.8 3.2 3.2
CV 0.22 0.13 0.14
Dalton
Mean 53.4 25.5 27.5
Std. dev. 11.0 3.1 3.0
CV 0.20 0.12 0.11
Greenbriar
Mean 63.7 22.2 26.1
Std. dev. 11.4 3.6 3.2
CV 0.18 0.16 0.12
Redstone
Mean 80.2 28.7 28.3
Std. dev. 18.8 4.1 3.8
CV 0.23 0.14 0.14
The statistics for basic dimensions in Table10.2 provide a complementary

perspective on artifact life histories. Interestingly, the coefficients of variation
(CVs) of all measurements are slightly higher for Clovis and Cumberland than
for other types, even though the correlations between length and width tend
to be stronger. Moreover, all of the later types, except Redstone, are much
shorter on average than Clovis and Cumberland points. In aggregate, the
descriptive statistics of shape and correlation results suggests that all Paleoindian
point types in the database were sometimes reworked or resharpened, but that
Beaver Lake, Dalton, and Greenbriar points tended to be more extensively
modified, and abandoned closer to a uniform point of exhaustion, regardless of
their original size. The Clovis and Cumberland points show a greater variety
of lengths, but greater uniformity in overall shape, suggesting that they were
made in a range of sizes but tended to retain more of their original proportions
at the time of discard.
We may safely reject the null hypothesis for the six point types examined
here. Early and Middle Paleoindian points (Clovis and Cumberland) as a
group seem to have been treated differently, to have undergone life histories,
than later Beaver Lake, Dalton, and Greenbriar points. The case of Redstone
points remains ambiguous the correlation results at least place them midway
between the other two groups. To the extent that these point types represent
a chronological sequence, it would be safe to conclude criteria for discard of
points changed over the Paleoindian period in Tennessee.
Why would there have been changes in the life histories of Paleoindian
spear points over time? The modified MVT suggests two alternative (and non-
exclusive) working hypotheses.
The tendency for later Paleoindian point types to be discarded further along
in their use lives could reflect increased cost of replacement. It is unlikely that
manufacture costs per se differed much between these different varieties of
fluted points, and there is certainly no evidence that manufacture becomes
more elaborate over time. However, access to raw material affects costs too.
Greater population packing, with smaller and more rigidly bounded territo-
ries, could have restricted access to raw materials with limited distributions. As
human populations began to fill in the landscape subsequent to initial colo-
nization, at least some later Paleoindian groups would have been confined to
territories farther from sources of high-quality stone suitable for making large
bifacial points. This hypothesis has one clear test implication: if access to raw
materials is behind the differences in conditions of points at discard, we would
expect the general temporal trends described here to differ across the region
according to availability of raw materials. More specifically, the outcomes of
differential levels of resharpening should be more evident in areas with little or
no access to high-quality stone than in areas where suitable stone was abun-
dant and easy to comeby.
A second explanation for changing artifact use-lives stemming from the
modified MVT would be a decline in average return from use of points. Two
linked factors could have contributed to diminished gains from hunting with
spears in the Late Pleistocene and Early Holocene. One is the extinction of
many species of large herbivore are the end of the Pleistocene, which should
have produced a decrease in average returns from successful hunts. The sec-
ond is a shift to smaller game species with higher relative search and handling
costs, whether due to loss of large game or to increased accessibility of smaller
animals. Sites such as Dust Cave in Alabama, as well as several other cave and
rockshelter sites with Late Paleoindian components, are noted for their rel-
atively high diversity of faunal species, including migratory birds and even
squirrels (Walker 2007). The model predicts that if expected returns from use
of spear points declined significantly, there would be longer intervals between
replacements, and people would allow artifacts, on average, to progress fur-
ther along the gradient of declining utility before replacing them. Given the
paucity of stratigraphically sealed Paleoindian sites with well-preserved fauna
in the Southeast, it is difficult to derive useful test implications for the region.
However, looking more broadly, if hunting returns did have the expected
effects on human decisions about retaining or discarding spear points, we
would expect different trajectories of change over time in areas such as the
High Plains, where at least one large herbivore (bison) persisted through
the Holocene, and the Great Basin, where diets began to broaden quite early
on (see Cannon and Meltzer 2008; Elston and Zeanah 2002:105, and refer-
ences therein).
Conclusions
This study uses a modified version of the Marginal Value Theorem a
mainstay of Behavioral Ecology to make predictions about the life his-
tories of artifacts that exhibit declining utility with successive uses. Some
predictions for example, that artifacts should be retained longer as costs
increase conform well with previous thinking falling under the rubric of
technological organization. Other predictions, namely that declines in aver-
age yields from use of artifacts should result in their being retained longer,
are novel. The application of the models general predictions to Paleoindian
projectile points from Tennessee, using data from the PIDBA project, is
explicitly exploratory in nature. The study uses coarse measures of projectile
point resharpening to take advantage of the existing database. Moreover, we
have comparatively little independent data on conditions such as access to
raw material, or yields from spear hunting, that might have affected deci-
sions about when to discard stone tools. What the model does provide is
a set of alternative hypotheses to account for apparent changes over time
in life histories of Paleoindian points. Our principal aim is to show how
these kinds of models can be used in analyses of archaeological stone arti-
facts, and not to provide a complete explanation for the life histories of
Paleoindian points in Tennessee.
In fact, many variables will contribute to variation in the forms of projectile
points: the challenge is in teasing out their effects. In a recent paper, Buchanan
and colleagues (2011) evaluate whether the designs of Paleoindian projectile
points were responsive to the most common type of game targeted. They use
a much more sophisticated approach to examine variation in the sizes and
shapes of Folsom and Clovis Paleoindian projectile points associated with dif-
ferent prey species (mammoth and bison). Their results are equivocal: whereas
the dominant prey could account for differences in the shapes of Folsom and
Clovis points, the shapes of Clovis points associated with different types of large
game do not differ significantly. However, Buchanan etal. explicitly focused
on artifact design as expressed in overall shape. It would be interesting to see
whether shape variables more indicative of resharpening, as well as variance in
artifact size, were related to the kinds of game animals targeted.
Surovell (2003, 2009) has also developed an application of the MVT to

archaeological lithic assemblages. Surovells approach conforms more closely
to Charnovs version of the MVT, in which both time and energy appear
as currencies. The modified version used here employs only energy to assess
both costs and returns from artifacts. As discussed in the preceding text, the
application of these alternative models will vary contextually. If taking time
to replace tools means less time to use them, then Surovells version would be
more appropriate: if not, then this version might be more relevant. For exam-
ple, long-lasting tool forms such as Dalton adzes and ground-stone artifacts
probably conform better with the modified version of the MVT proposed
here. Also, although we have applied the predictions of the modified MVT to
changes over time in the treatment of projectile points, it might be equally, if
not more, useful for examining spatial variation in artifact life histories within
time periods. Geographic information system (GIS)-based data on least-cost
distances to raw material sources and reconstructed levels of environmental
productivity could be used to test both alternative explanations for artifact dis-
card strategies simultaneously.
A central assumption of any application of the MVT to stone tools is that
they lose utility or functionality gradually over time.This is likely to be true of
spear points or any other kind of tool that actually wears out or accumulates
structural damage, although additional experimental work is certainly needed.
It is manifestly true of cores, which inevitably get smaller as they are used,
and yield fewer and smaller products as a consequence. Typically, the level of
reduction of cores is linked to the cost or difficulty of obtaining raw materi-
als (e.g., Andrefsky 1994; Beck etal. 2002; Kuhn 1995), consistent with this
model (and many others). However, where the degree of core exhaustion is
independent of raw material cost when, for example, where it varies within
a single archaeological sequence with no evidence for changes in raw mate-
rial availability the model presented in this chapter provides an alternative
explanation. The returns realized from the blanks produced from cores might
vary according to the functions of flake tools, with resulting implications for
the extent of core reduction. It would be interesting to examine differential
reduction of stone cores across sequences where alternative types of working
edges, such as polished stone or metal tools, became more readily available
to a population. If the ranges of activities involving stone tools declined over
time, and in particular, if stone tools we increasingly relegated to marginal
activities, we might expect the cores still in use to be more heavily reduced.
It is important to point out that the models predictions are expected to
hold true only if a range of assumptions (outlined earlier) is met. Like any
formal construction of optimal behavior, it does not predict that people will
always behave optimally it only describes what optimal behavior ought to
be. It is entirely possible people did not effectively optimize stone tool discard
rates, either because they lacked requisite information, because gains were too
marginal to be effectively measured, or because conflicting interests with larger
fitness consequences weighed more heavily in their decisions. Failure of the
model to predict past behavior points toward new lines of inquiry that reex-
amine these preconditions and assumptions.
Beyond providing a provisional approach to studying the results of decisions
to abandon or retain artifacts, a second aim of this chapter is to show how for-
mal models from Behavioral Ecology can be applied to questions and data sets
that are normally considered the purview of studies of technological orga-
nization (see Surovell 2009 for a more wide-ranging study). As discussed in
the introduction, HBE and TO share many assumptions and goals, even if they
employ rather different theoretical sources and styles of argument. Moreover,
manufacture debris, used, reworked and discarded artifacts, the normal focus
of studies of TO, are the most abundant elements in almost any lithic archaeo-
logical record. With proper reconceptualization of the models developed by
behavioral ecologists, these abundant and ubiquitous artifacts provide a read-
ily accessible and robust testing ground for a range of ideas originating in the
biological and behavioral sciences.
Of course, the simple fact that a category of evidence is very common does
not mean we have to depend on it. James OConnell often cites the metaphor
of the drunken man who looks for his keys under a streetlamp, even though
he dropped them somewhere else, simply because the light is better there.
His aim is to remind archaeologists that we should not confine our thinking to
the evidence that happens to be most abundant. However, although his point
is well taken, the metaphor can also be reversed. After all, we do not really
know where we lost our keys. If there are several sets of keys on the ground
around the streetlamp it is a good idea to pick them up and see what doors
theyopen.
Acknowledgments
We should first thank Todd Surovell for his healthy and constructive skepti-
cism about the ideas leading to this chapter. He may not agree with everything
(or anything) we say, but it would not have taken the form it has without his
extensive critical input. Students in S. K.s (2009/2010) graduate seminar at
the University of Kln and colleagues at the Max Planck Institute for Human
Evolutionary Studies in Leipzig also provided useful critical comments on the
ideas contained in this chapter. We also thank John Broster and Mark Norton
for making data from the Tennessee Fluted Point Survey publically available
for other researchers touse.
Notes
1 For the calculations used here, increasing prob- example, we can hypothesize a type of artifact
ability of failure and gradual loss of effective- that, barring catastrophic breakage, can be used
ness through reworking or wear have identical and reworked an average of 25 times before its
implications for declining utility. In fact, the utility drops to zero. If catastrophic breakage
expected utility of an artifact that does not wear (due to hitting a rock, being stepped on by a
out gradually, but has an increasing probability wounded mammoth, etc.) occurs in 1 in 10 use
of failure, is an inverse function of its proba- events, only 20% of artifacts will ever surpass 15
bility of failure in the next use. What would uses. The resulting assemblage would consist of
differ are options for assessing artifact utility. In 4 broken points for every partially worn one. If
the former case, tool users could simply mon- we increase the frequency of catastrophic fail-
itor the conditions of artifacts. In the second, ure to 15%, fewer than 10% of points survive to
they would have to keep track of the number the same level of reduction.
of times something had been used. Paleoindian 2 This graphic model superficially resembles
(and other) points often show signs of rework- one proposed by Bousman (2005). However,
ing, but in many (most?) assemblages, broken they differ in several respects, most notably
points far outnumber whole ones. This does that the criterion variable in Bousmans model
not substantially affect the case for declining is a cumulative value, whereas the descend-
utility. Points will break for a variety of rea- ing curve is an instantaneous return. For this
sons long before they wear out or are reduced reason the two models have very different
to a near-useless state. To take a hypothetical implications.
References
Anderson, D. G. 1990a. A North American Points: Update 1998. Archaeology of Eastern

Paleoindian Projectile Point Database. Current North America 26:163187.
Research in the Pleistocene 7:6769. Anderson, D. G., and M. K. Faught. 2000.
Anderson, D. G. 1990b. The Paleoindian Paleoindian Artifact Distributions: Evidence
Colonization of Eastern North America: A and Implications. Antiquity 74:507513.
View from the Southeastern United States. Anderson, D. G., D. S. Miller, D. T. Anderson,
In Early Paleoindian Economies of Eastern North S. J. Yerka, J. C. Gillam, E. N. Johanson, and
America, edited by K. B. Tankersley and B. L. A. Smallwood. 2009. Paleoindians in North
Isaac. Research in Economic Anthropology 5, America: Evidence from PIDBA (Paleoindian
pp.163216. JAI Press, Greenwich,CT. Database of the Americas). Poster presented at
Anderson, D. G. 1995. Recent Advances and the Annual Meeting of the Society for American
Paleoindian and Early Archaic Period Research Archaeology, Atlanta, GA, April 24,2009.
in the Southeastern United States. Archaeology Anderson, D. G., D. S. Miller, S. J. Yerka, and M.
of Eastern North America 23:145176. K. Faught. 2005. Paleoindian Database of the
Anderson, D. G. 1996. Models of Paleoindian Americas: 2005 Status Report. Current Research
and Early Archaic Settlement in the Lower in the Pleistocene 22:9192.
Southeast. In The Paleoindian and Early Archaic Anderson, D. G., L. D. OSteen, and K. E.
Southeast, edited by D. G. Anderson and K. Sassaman. 1996. Environmental and
Sassaman, pp. 2957. University of Alabama Chronological Considerations. In The
Press, Tuscaloosa. Paleoindian and Early Archaic Southeast, edited
Anderson, D. G. 2001. Paleoindian Interaction by D. G. Anderson and K. E. Sassaman, pp.
and Mating Networks: Reply of Moore and 315. University of Alabama Press, Tuscaloosa.
Mosely. American Antiquity 65:4366. Andrefsky, W. 1994. Raw-Material Availability
Anderson, D. G., and M. K. Faught. 1998. The and the Organization of Technology. American
Distribution of Fluted Paleoindian Projectile Antiquity 59:2134.
Andrefsky, W. 2009. The Analysis of Stone Tool Buchanan, B., M. Collard, M. J. Hamilton, and M.
Procurement, Production, and Maintenance. J. OBrien. 2011. Points and Prey:A Quantitative
Journal of Archaeology Research 17:65103. Test of the Hypothesis that Prey Size Influences
Beck, C., and G. T. Jones. 1990. Toolstone Early Paleoindian Projectile Point Form. Journal
Selection and Lithic Technology in the of Archaeological Science 38:852864.
Early Great Basin. Journal of Field Archaeology Cambron, J. W., and D. C. Hulse. 1975. Handbook
17:283299. of Alabama Archaeology, Part I, Point Types.
Beck, C., A. K. Taylor, G. T. Jones, C. M. Fadem, Revised Edition. Archaeological Research
C. R. Cook, and S. A. Millward. 2002. Rocks Association of Alabama, Tuscaloosa.
Are Heavy: Transport Costs and Paleoarchaic Cannon, M. D., and D. J. Meltzer. 2008. Explaining
Quarry Behavior in the Great Basin. Journal of Variability in Early Paleoindian Foraging.
Anthropological Archaeology 21:481507. Quaternary International 191:517.
Bettinger, R. L., B. Winterhalder, and R. Charnov, E. L. 1976. Optimal Foraging, the
McElreath. 2006. A Simple Model of Marginal Value Theorem. Theoretical Population
Technological Intensification. Journal of Biology 9:129136.
Archaeological Science 33:538545. Cheshier, J., and R. L. Kelly. 2006. Projectile Point
Bousman, C. B. 2005. Coping with Risk: Later Shape and Durability:The Effect of Thickness:
Stone Age Hunter-Gatherers at Blydefontein Length, American Antiquity 71:353364.
Rock Shelter. Journal of Anthropological Elston, R. G., and D. W. Zeanah. 2002. Thinking
Archaeology, 24:193226. Outside the Box: A New Perspective on Diet
Brantingham, P. J., and S. Kuhn. 2001. Constraints Breadth and Sexual Division of Labor in the
on Levallois Core Technology:A Mathematical Prearchaic Great Basin. World Archaeology
Approach. Journal of Archaeological Science 34:103130.
28:747762. Eren, M. I., Greenspan, A., and C. G. Sampson.
Breitburg, E., and J. B. Broster. 1995. The Clovis 2008.Are Upper Paleolithic Blade Cores More
and Cumberland Projectile Points ofTennessee: Productive than Middle Paleolithic Discoidal
Quantitative and Qualitative Attributes and Cores? A Replication Experiment. Journal of
Morphometric Affinities. Current Research in Human Evolution 55:952961.
the Pleistocene 12:46. Goodyear,A. C. 1982.The Chronological Position
Bright, J., Ugan, A., and L. Hunsacker. 2002. of the Dalton Horizon in the Southeastern
The Effects of Handling Time on Subsistence United States. American Antiquity 47:382395.
Technology. World Archaeology, 34:164181. Goodyear, A. C. 1989. A Hypothesis for the Use
Broster, J. B., and M. R. Norton. 1996. Recent of Crypto-crystalline Raw Materials among
Paleoindian Research in Tennessee. In The Paleoindian Groups of North America. In
Paleoindian and Early Archaic Southeast, edited Eastern Paleoindian Lithic Resource Use, edited
by D. G.Anderson and K. E. Sassaman, pp. 288 by C. J. Ellis and J. C. Lothrop, pp. 19.
297. University of Alabama Press, Tuscaloosa. Westview Press, Boulder,CO.
Brown, J. S. 1988. Patch Use as an Indicator Goodyear, A. C. 1999. The Early Holocene
of Habitat Preference, Predation Risk, and Occupation of the Southeastern United
Competition. Behavioral Ecology and Sociobiology States: A Geoarchaeological Smarmy. In Ice Age
22:3747. Peoples of North America: Environments, Origins,
Brown, J. S., and W. A. Mitchell. 1989. Diet Selec- and Adaptations of the First Americans, edited by
tionon Depletable Resources. Oikos 54:3345. R. Bonnichsen and K. L. Turnmire, pp. 432
Buchanan, B. 2006. An Analysis of Folsom 481. Oregon State University Press, Corvallis.
Projectile Point Resharpening Using Goodyear, A. C. 2006. Recognizing the
Quantitative Comparisons of Form and Redstone Fluted Point in the South Carolina
Allometry. Journal of Archaeological Science Paleoindian Point Database. Current Research
33:185199. in the Pleistocene 23:112114.
Grafen, A. 1984. Natural Selection, Kin Selection Lewis, T. M. N. 1954. The Cumberland Point.
and Group Selection. In Behavioural Ecology, Bulletin of the Oklahoma Anthropological Society
2nd ed., edited by J. R. Krebs and N. B. Davies, 11:78.
pp. 6284. Blackwell Scientific, Oxford. Lupo, K. D., and D. N. Schmitt. 2002. Upper
Guthe, A. K. 1966a. Paleo-Indian Points from Paleolithic Net-hunting,Small Prey Exploitation,
Hardin County. Tennessee Archaeologist and Womens Work Effort: A View from the
22:8183. Ethnographic and Ethnoarchaeological Record
Guthe, A. K. 1966b. Paleo-Indian Points from of the Congo Basin. Journal of Archaeological
Middle Tennessee. Tennessee Archaeologist 22:80. Method and Theory 9:147178.
Guthe, A. K. 1966c. Tennessees Paleo-Indian. McCary, Ben C. 1947. A Survey and Study of
Tennessee Archaeologist 22:6777. Folsom-Like Points Found in Virginia, Nos.
Guthe, A. K. 1983. Comments Regarding the 1131. Quarterly Bulletin, Archeological Society of
Inventory of Fluted Points in Tennessee. Virginia(2)1.
Archaeology of Eastern North America 11:2324. McNair, J. N. 1979. A Generalized Model of
Hoffman, C. M. 1985. Projectile Point Optimal Diets. Theoretical Population Biology
Maintenance and Typology: Assessment with 15:159170.
Factor Analysis and Canonical Correlation. McNair, J. N. 1982. Optimal Giving Up Times
In For Concordance in Archaeological Analysis: and the Marginal Value Theorem. American
Bridging Data Structure, Quantitative Technique, Naturalist 119:511529.
and Theory, edited by C. Carr, pp. 516612. McNair, J. N. 1983. A Class of Patch-Use
Westport Press, Kansas City,MO. Strategies. American Zoologist 23:303313.
Howard, C. D. 1990. The Clovis Point: Metcalfe, D., and R. K. Barlow. 1993. A Model
Characteristics and Type Description. Plains for Exploring the Optimal Trade-off between
Anthropologist 35:255262. Field Processing and Transport. American
Justice, N. 1995. Stone Age Spear and Arrow Anthropologist 94:340356.
Points of the Midcontinental and Eastern United Morse, D.F., P. A. Morse, and J. Waggoneer, Jr.
States: A Modern Survey and Reference. Indiana 1964. Fluted Points from Smith County,
University Press, Bloomington. Tennessee. Tennessee Archaeologist 20(1):1634.
Kaplan, H., and K. Hill. 1992. The Evolutionary Nelson, M. 1991. The Study of Technological
Ecology of Food Acquisition. In Evolutionary Organization. Archaeological Method and Theory
Ecology and Human Behavior, edited by E. A. 3:57100.
Smith and B. Winterhalder, pp. 167202. Nonacs, P. 2001. State Dependent Behavior
Aldine de Gruyter, NewYork. and the Marginal Value Theorem. Behavioral
Keeley, L. H. 1982. Hafting and Retooling: Effects Ecology 12: 7183.
on the Archaeological Record. American OConnell, J. F., and K. Hawkes. 1984. Food
Antiquity 47(4):798809. Choice and Foraging Sites among the
Kelly Robert L., and Lawrence C. Todd. 1988. Alyawara. Journal of Anthropological Research
Coming into the Country: Early Paleoindian 40:504535.
Hunting and Mobility. American Antiquity Praciunas, M. M. 2007. Bifacial Cores and Flake
53(2): 231244. Production Efficiency: An Experimental Test
Kuhn, S. 1994. A Formal Approach to the of the Technological Assumptions. American
Design and Assembly of Transported Toolkits. Antiquity 72:334348.
American Antiquity 59:426442. Pyke, G. 1984. Optimal Foraging Theory: A
Kuhn, S. 1995. Mousterian Lithic Technology: An Critical Review. Annual Review of Ecology and
Ecological Perspective. Princeton University Systematics 15:523575.
Press, Princeton,NJ. Sellards, E. H. 1952. Early Man in America.
Kurtz, E. B. 1930. Life Expectancy of Physical University of Texas Press, Austin.
Property Based on Mortality Laws. Ronald, Shennan, S. 2008. Evolution in Archaeology.
NewYork. Annual Review of Anthropology 37:7591.
Sherwood, S. C., B. N. Driskell, A. R. Randall, Surovell, T. 2003. The Behavioral Ecology of

and S. C. Meeks. 2004. Chronology and Folsom Lithic Technology. PhD dissertation,
Stratigraphy at Dust Cave, Alabama. American Department of Anthropology, University of
Antiquity 69:533534. Arizona, Tucson.
Shott, M. J. 1997. Stones and Shafts Redux: The Surovell, T. 2009. Toward a Behavioral Ecology
Metric Discrimination of Chipped-stone of Lithic Technology: Cases from Paleoindian
Dart and Arrow Points. American Antiquity 62: Archaeology. University of Arizona Press,
86101. Tucson.
Shott, M. J., and J. A. M. Ballenger. 2007. Biface Thiel, A., and T. F. Hoffmeister. 2004. Knowing
Reduction and the Measurement of Dalton Your Habitat: Linking Patch-encounter
Curation:A Southeastern Case Study. American Rate and Patch Exploitation in Parasitoids.
Antiquity 72:153175.. Behavioral Ecology 15:419425
Shott, M. J., and P. Sillitoe. 2005. Use Life and Thomas, F. 2007. The Behavioral Ecology of
Curation in New Guinea Experimental Shellfish Gathering in Western Kiribati,
Used Flakes. Journal of Archaeological Science Micronesia, 2: Patch Choice, Patch Sampling,
32:653663. and Risk. Human Ecology 35:515526.
Shott, M. J., and P. Sillitoe. 2004. Modeling Use- Ugan, A., J. Bright, and A. Rogers. 2003. When
life Distributions in Archaeology Using New Is Technology Worth the Trouble? Journal of
Guinea Wola Ethnographic Data. American Archaeological Science 30:13151330.
Antiquity 69:339355. Wajnberg, E., X. Fauvergue, and O. Pons. 2000.
Smith, E. A., and B. Winterhalder. 1992. Natural Patch Leaving Decision Rules and the
Selection and Decision-making: Some Marginal Value Theorem: An Experimental
Fundamental Principles. In Evolutionary Analysis and a Simulation Model. Behavioral
Ecology and Human Behavior, edited by E. A. Ecology 11:577586
Smith and B. Winterhalder, pp. 2560. Aldine Walker, R. 2007. Hunting in the Late Paleoindian
de Gruyter, NewYork. Period: Faunal Remains from Dust Cave,
Sosis, R. 2002. Patch Choice Decisions among Alabama. In Foragers of the Terminal Pleistocene,
Ifaluk Fishers. American Anthropologist edited by R. B. Walker and B. N. Driskell,
104:583598. pp. 99115. University of Alabama Press,
Stephens, D. W. 2008. Decision Ecology: Tuscaloosa.
Foraging and the Ecology of Animal Decision Wilke, A. 2006. Evolved Responses to an
Making. Cognitive and Affective Behavior in the Uncertain World. PhD dissertation,
Neurosciences 8:475484. Fachbereich Erziehungswissenschaft und
Stephens, D. W., and J. R. Krebs. 1986. Foraging Psychologie, Free University of Berlin.
Theory. Princeton University Press, http://www.diss.fu-berlin.de/diss/receive/
Princeton,NJ. FUDISS_thesis_000000001936
Eleven
Signals in Stone: Exploring the

Role of Social Information
Exchange, Conspicuous
Consumption, and Costly Signaling
Theory in Lithic Analysis
Colin P.Quinn
Stone tools are most often seen and studied as utilitarian objects. However, as
with other types of material culture, lithics may have played an important role
in the creation of identity and the negotiation of interpersonal relationships
for people in the past. As a result, archaeologists interested in lithic technologi-
cal organization should consider not just the function of stone tools to cut,
scrape, or pierce; but also their social function. One evolutionary approach that
links stone tool production, use, and discard with the potential social meaning
and information that these tools have shared is costly signaling theory.
Originating in biology and human behavioral ecology, costly signaling the-
ory is concerned with wasteful and uneconomic displays that impact repro-
ductive fitness and has been used successfully by several researchers to explain
behaviors in humans and nonhuman animals. However, some researchers have
identified existing archaeological applications of costly signaling theory as
just-so-stories (Codding and Jones 2007), a crutch for inexplicable phe-
nomena or seemingly illogical behavior in the archaeological record. Although
costly signaling theory is a fertile theoretical approach, it is important that
archaeological applications of costly signaling theory are grounded in the the-
ory as well as the archaeological data. To make costly signaling theory a sound
scientific approach, we must build models, develop hypotheses, and test them
using the archaeological record.
In this chapter I will not solve all the problems of applying costly sig-
naling theory to lithic data sets, nor will I provide the answer for how to
198
Si gnals i nS t o n e 199
identify costly signaling with lithics. I will, however, explore the theoretical
foundations of costly signaling theory and social information exchange, illu-
minate the potentials and pitfalls of this theoretical perspective for those con-
cerned with lithic analysis, provide a general framework to approach signaling
in the archaeological record, highlight several lines of archaeological evidence
that may be used to identify the presence of costly material displays in stone
in the past, and examine a few existing archaeological examples that may be
appropriate for applying signaling approaches in the future. Together, I will
weigh the strengths and weaknesses of using signaling theory in archaeolog-
ical contexts, explore ways in which it has been used by archaeologists in the
past, and provide a potential roadmap for incorporating it into studies of lithic
technological organization.
Social Information Exchange, Material Culture, and Lithic

Technological Organization
Material culture items convey information and play an important role in the
negotiation of social relationships in human societies (Earle 2004; Preston
2000; Robb 1998, 1999; Wobst 1977, 1999). Some of that information is
actively displayed and received, some is passively conveyed and received, and
some information is not received by peoples observing the material culture
items (Sackett 1982, 1985; Wiessner 1983, 1985). Although lithics have not
often been held up as a primary means of social information exchange (often
that distinction is given primarily to adornment items, dress, and ceramics),
researchers have viewed lithics (particularly diagnostics and tools) as convey-
ing information about past peoples (e.g., Barton 1997; Bernstein 1984; Gero
1989; Holm 1994).
When examining lithic assemblages, archaeologists often first address issues
of tool function, retouch patterns, and procurement strategies with an explicit
focus on the utility of stone tools in important subsistence acquisition strat-
egies. This is not a bad place to start, but it often leaves unanswered or unad-
dressed questions about the social implications of the use and discard of stone
tools. Examining the role of lithics in utilitarian and functional contexts is a
necessary task for archaeologists, and I do not intend to suggest otherwise. I do,
on the other hand, think that lithics can and should be studied as other types
of materials in the archaeological record are studied; with an eye toward their
role in contexts of information exchange in thepast.
There are reasons why stone may have been not only a good vehicle for
social information exchange in the past, but also a valuable source of infor-
mation on dynamic social processes for archaeologists. First, there are sev-
eral characteristics about stone and stone tools (e.g., raw material availability,
source type, and production costs) that make them good candidates as costly
200 Coli n P. Q u i n n
signals. Second, as an artifact type that is often among the best preserved in
the archaeological record, lithics provide a source of data that if properly situ-
ated within discussions of information exchange can prove valuable for testing
models and hypotheses about social information exchange in thepast.
Theoretical Foundations of Costly Signaling Theory

Costly signaling theory combines concepts of costly behavior and public gen-
erosity (Fried 1967; Mauss 1924; Veblen 1994) as forms of social competition
that provide a way to articulate the notion of intangible social benefits that
can be gained through symbolic representations of self with more materialist
notions of individuals as self-interested but socially embedded decision mak-
ers (Bliege Bird and Smith 2005; Quinn 2006). Others have defined costly
signaling as something that increases the fitness of an individual by altering
the behavior of recipients of the signal (Dawkins and Krebs 1978; Hasson
1994; Krebs and Dawkins 1984; Maynard Smith and Harper 1995). The signal
must be beneficial for reproductive fitness in the given information exchange
between individuals, yet costly in other contexts (Hasson 1994; Maynard Smith
and Harper 1995). In addition, the signal must be an honest representation of
an individuals underlying reproductive fitness predicated on the fact that the
signal is so costly that it is impossible to possess without the characteristics or
qualities being signaled.
Costly signaling theory has its roots in biology and the handicap principle.
Zahavi (1975) originally explored costly signaling while attempting to under-
stand why animals would engage in costly and extravagant displays.The handi-
cap principle is the hypothesis that costly physical or behavioral characteristics
will be inherently reliable signals of reproductive success. Males who possess
these characteristics will be selected as mates by females who are trying to
find mates with the best genotypes (Zahavi 1975). For example, the handicap
principle explains the practical inefficiency of peacocks tails. Peacocks require
the necessary genetic ability to fight parasites and invest energy into the tail,
which means that a large and healthy tail would signal to peahens that those
individuals had good genes. Grafen (1990) reformulated the handicap prin-
ciple into a series of models, including costly signaling, where individuals used
visual displays at various costs to signal their quality as mates, and emphasized
that signaling was an evolutionarily stable strategy. Anthropologists then took
this concept and attempted to apply costly signaling theory to conspicuous
consumption and uneconomical displays in ethnographic contexts (Hawkes
1990, 1991, 1993; Kaplan and Hill 1985;Veblen 1994).
Two conditions are required for the evolutionary stability of costly signal-
ing (Grafen 1990; Zahavi 1975). First, signals must convey reliable informa-
tion about variation in the underlying quality being advertised, involving such
aspects as resource control and competitive ability (Quinn 2006). Second, the
signal must impose a cost on the signaler that is directly linked to the qual-
ity being advertised. The payoff to the signaler comes from being chosen as a
mate or ally or deferred to as a dominant in mating, cooperative, or competi-
tive contexts (Smith and Bliege Bird 2000). The payoff to the recipient comes
from the usefulness of the information being signaled to evaluate the signalers
quality as a competitor, mate, orally.
Costly Signaling and Human Behavior

Human cultural behavior often acts in ways not seen in biological systems. To
go beyond qualitative metaphorical comparisons between the two, researchers
must examine whether there is predictive substance in the implication that
social systems satisfy similar principles and constraints as biological systems
(Bettencourt etal. 2007).This has a profound effect on the use of evolutionary
approaches, such as costly signaling theory, taken from biology. Archaeologists
need to adequately explore the ways in which signaling in cultural contexts
differs from biological contexts to appropriately use costly signaling theory to
explain human behavior in thepast.
Behaviors and material culture need not be directly genetically linked to
affect reproductive fitness. Instead, behaviors and material culture are often
seen as part of the extended phenotype (the outward expression of the geno-
type) of an individual (Boone and Smith 1998; Dunnell 1980; Kantner 2003).
There has been much debate about the concept of an extended phenotype
and the role of natural selection operating on the phenotype (interactors)
(Dunnell 1980; Lyman and OBrien 1998, 2001) or on individuals (replicators)
(Boone and Smith 1998; Kantner 2003; Smith and Winterhalder 1992). The
selective process occurs on the individual, as selection of particular material
culture items that make up an extended phenotype is governed by individ-
ual decision making and cultural transmission (Boone and Smith 1998; Boyd
and Richerson 1985; Durham 1991; Kantner 2003; Kelly 2000; Smith 2000;
Smith and Winterhalder 1992). Social information embedded within material
culture is often associated with markers of reproductive fitness as well as the
individual themselves. Just as there is genotypic plasticity that affects biological
phenotypic expression (such as genes for height being affected by an individ-
uals nutrition while young), I believe that phenotypic plasticity can be seen
in material culture. Phenotypic variability is not generated randomly, even in
biological contexts (Dawkins 1987; Rindos 1989; VanPool and VanPool 2003).
For example, artifacts that represent reproductive fitness for an individual at
one time in his or her life may not be reproductively fit at another time in
his or her life. Fitness is, therefore, mutable and fluid. Although not geneti-
cally linked, material culture items associated with an individual often link to
attributes such as age, gender, status, wealth, access to resources, knowledge,

and knowledge about access to up-to-date information, thereby becoming
phenotypically expressed. Because phenotypes affect reproductive success, var-
iability in behavior or material culture that causes an individual to have more,
higher quality offspring than other individuals will be subject to selection pro-
cesses similar to those guiding natural selection.
People who invest in costly signals will do so for a perceived benefit (i.e.,
reproductive fitness, increased social status, increased quantity and quality of
allies) that may or may not translate into actual benefits. There are several rea-
sons that signaling based on perceived benefits would not result in actual ben-
efits for those who invest in the costly signals. People are not always rational
or able to adequately measure the costs and benefits of an activity. A decision
based on perceived benefits may actually result in a low return, one that may
have been higher if the individual had access to more accurate and up-to-date
information about dynamic social, economic, and political conditions. Changes
in these conditions will result in shifting ratios of costs and benefits that may
not be taken into account by individuals or groups when they choose to invest
in costly signals. In addition, there are unpredictability and unintended con-
sequences in human action that limit the benefits associated with a particular
signal. Although actual benefits may not always match perceived benefits, for
costly signaling to be an evolutionary stable strategy it must confer actual ben-
efit often enough to become a shared evolutionary strategy. For archaeologists,
it is fortunate that signaling decisions are made based on perceived benefits.
Because real costs and benefits will be highly fluid and dynamic (based on
changing social, economic, and political milieus), it would be more difficult to
identify constantly changing signaling strategies in the archaeological record.
Perceived benefits, on the other hand, will likely produce more stable signal-
ing strategies (due to a lag between the perceived and actual benefits) that will
produce more identifiable archaeological signatures.
Units of Analysis
The models of costly signaling, like most models that come from evolutionary
ecology and optimal foraging theory, have several significant characteristics
that make uncritical adoption inappropriate. First, these models assume that
the individual is the only relevant unit. In many non-Western contexts, other
units of analysis, such as the household or the community, may be more impor-
tant entities within the social system than individuals (Gintis 2000; Gintis etal.
2003; Wilk 1989, 1996). In these cases, it is quite possible that individuals defer
decision making to household leaders and not act to optimize their repro-
ductive fitness absent a consideration of the household. In addition, in certain
contexts the success of a household can be more important than the success of
an individual (whose success is predicated on group membership and success

of those groups). It is important to not overemphasize the autonomy of indi-
viduals in these contexts and therefore our models must be plastic enough to
shift with the different units that are socially relevant.
In addition to the considerations of the role of the individual in past societ-
ies, archaeologists must deal with another problem when employing models
from human behavioral ecology: the individual is almost always invisible in the
archaeological record. For a theoretical approach designed around individu-
als, individual decision-making processes, and measuring reproductive suc-
cess, direct correlates are often absent from the archaeological record. Even in
most mortuary contexts, where we have an individual and an individual burial
event, we must acknowledge that the individual was buried by others and the
body treatment and artifacts in this context may or may not reflect the status
or identity of the buried individual (Parker-Pearson 1999). The palimpsest of
behavior that forms the archaeological record requires a refocus away from the
individual, individual decision making, and ways of measuring reproductive
fitness that are more available to biologists and ethnographers today.
Although there are pitfalls to employing costly signaling to the archaeolog-
ical record, the archaeological record has the potential to inform our under-
standing of costly signaling in unique ways. First, in the archaeological record
we can most easily identify activities and behaviors that are culturally shared
and repeated (or are substantially durable). Costly signals must be recognized for
their links to reproductive fitness, and as such, must be perceived by the group.
In this way, idiosyncratic materials or behaviors may not be easily recognized
by the potential audience as conveying information related to reproductive fit-
ness (and therefore would not be a costly signal). This is not to say that no idi-
osyncratic behaviors or objects can be costly signals, just that as archaeologists
we will be hard pressed to find evidence of these behaviors and objects and
justify calling them costly signals. Second, the diachronic perspective offered
by archaeological analyses can highlight ways in which material culture use
changed through time at a scale not easily accessed to ethnographic or biologi-
cal researchers. Long-term stability or instability of signaling strategies informs
our understanding of the role of costly signaling within a particular cultural
system as well as our understanding of how objects become and cease to be
costly signals. While archaeologists cannot recognize all costly signaling in the
past, the archaeological record provides enough evidence to identify some
costly signaling behaviors and objects. More importantly, archaeologists can see
patterning and changes in behaviors and material culture use through time and
space that enlighten reconstructions of past lifeways.
To use costly signaling models within anthropological and archaeological
contexts, we must acknowledge the diversity within human social systems
by allowing for other units of analysis and other considerations beyond a
specific individuals own reproductive fitness. These views do not necessitate

an avoidance of models that were initially developed to deal with fitness-
optimizing individuals; instead, they require (1) a consideration of multiple
identities and levels at which a humans reproductive fitness is impacted (such
as households and communal membership); (2) linking the evidence of (a) the
relevant unit of analysis, (b) the relevant reproductive fitness characteristics, and
(c) the costly signals that link to those characteristics for the specific cultural
context being investigated; and (3) an acknowledgment of the limitations of
the archaeological record for seeing individuals, individual reproductive suc-
cess, and individual actions in thepast.
Concept of Costly Signaling in Archaeology

The most common application of costly signaling theory in anthropology
has been associated with identifying costly behaviors (e.g., Bliege Bird and
Smith 2005; Bliege Bird et al. 2001, 2002; Gintis et al. 2001; Hawkes 1990,
1991; Hawkes and Bliege Bird 2002; Henrich 2007; Ohtsubo and Watanabe
2009; Smith 2004; Smith and Bliege Bird 2000; Sosis 2003). In archaeology,
the focus on behaviors as costly signals has also received attention. McGuire
and Hildebrandt have employed costly signaling theory to explain big game
hunting as a prestige acquisition strategy among Great Basin hunter-gather-
ers (Hildebrandt and McGuire 2002, 2003; McGuire and Hildebrandt 2005;
McGuire et al. 2007). Although not universally accepted (see Codding and
Jones 2007), the work by McGuire and Hildebrandt has been among the most
visible incorporation of costly signaling theory in the archaeological literature.
The debate about archaeological evidence of big game hunting as a prestige
acquisition strategy is plagued by limitations of the archaeological record to
preserve evidence of these behaviors (Codding and Jones 2007), necessitating
a greater emphasis on theoretical and methodological approaches for seeing
costly signaling behavior in the archaeological record. Although the potential
of studying costly behaviors in the archaeological record is far from conclusive
or extensive, there is another aspect of costly signaling that is as useful, if not
more, for archaeologists: studying costly signals in material culture.
The perspective that material culture can be costly signals is not a new
idea in archaeology, though often it has not been explicitly linked with costly
signaling theory. Kohn and Mithen (1999) have argued that the production of
large quantities of symmetrical Acheulean handaxes was a way for individuals
to show off their skill and attract mates (through Zahavis handicap principle)
(see critique and reply: Hodgson 2009; Machin 2008; Mithen 2008). Neiman
(1997) has argued for monument construction as a form of conspicuous con-
sumption and wasteful advertising, which would produce distinctive material
correlates as costly signals. Plourde (2008) has argued that prestige goods (such
as objects of personal adornment and other costly objects) evolved as costly

honest signals conveying the quality of an individual who manufactured or
possessed these goods to others to gain advantages of prestige, increased access
to resources, and deference in social relationships.
However, not all costly signals are materialized. For example, Smith and
Bliege Bird (2000) have argued that the Meriam Islanders hunt turtles as a
form of costly signaling. However, the successful turtle hunts and hunters do
not have any material markers of their success. We can imagine that a turtle
shell would be an easy material into which information about turtle hunting
prowess can be embedded, but this is not the case, as turtle shells are discarded
after processing. Thus, it is necessary to examine both where and when costly
signals become materialized, as well as what information is actually being con-
veyed. In the preceding text I have laid out an initial roadmap for the identi-
fication of costly signals in lithics, but this is only a start, as more contextually
specific sets of archaeological correlates must be developed and compared with
real datasets.
There have been only a handful of studies that have explicitly linked costly
signaling theory and lithic technological organization.Waguespack etal. (2009)
have raised the possibility that all stone projectile technology may be costly
signals. Based off of experimental studies, the researchers argue that wooden-
tipped projectiles were functionally equivalent to stone-tipped projectiles. If
stone points provide no functional advantage and impart a cost (access to raw
materials, time and skill of production) that is greater than wooden points,
then as Waguespack et al. suggest there may be a social benefit to the use
of stone points for expressing identity and manipulating social relationships
(Waguespack etal. 2009).
Combining Costly Signaling Theory and Lithic Analysis

Although signaling theory is an overarching theory of social information
exchange through material culture, not all signals interact with reproductive
fitness in the same way. Some signals are more directly related to reproduc-
tive fitness than others, while some signals may not impact reproductive fit-
ness at all. The relationship between signals and reproductive fitness can be
conceptualized as a continuum; on one end are signals that do not affect
reproductive fitness and on the other are signals that play a significant role in
not only broadcasting, but also enhancing, reproductive fitness (Figure11.1).
Links between the material signals and the continuum are fluid and dynamic
based on the relationship between the variables that affect reproductive fitness.
Costly signals contain information about reproductive fitness and affect the
negotiation of interpersonal relationships by imposing an immediate cost (loss
of resources) for a potential deferred benefit (increasing reproductive fitness).
Signaling Theory
Contextually Specific Reproductive Fitness

Fitness Continuum
Low Signaling power High
Non-Costly Signals Costly Signals
11.1. Signaling theory, the fitness continuum, and the relationship between costly and non-
costly signals.
Non-costly signals can still convey information, such as age, gender, and eth-
nicity; however, because they do not incur a significant cost borne by the sig-
naler their costliness does not directly affect active negotiation of reproductive
fitness. Most artifacts associated with lithic technology fall on the non-costly
signal end of the signaling power continuum.They (1) do not convey informa-
tion about one person to another or (2) do not pose a substantial cost to the
individual or group who acquiresthem.
Reproductive fitness is contextually specific. Any given social, economic,
political, or environmental setting will have differing individual attributes
that enhance an individuals reproductive fitness. The characteristics that will
be selected for by others in mating and alliance formation situations will be
directly linked to the attributes that make an individual fit within that con-
text. Hence, certain characteristics may be signaled in one context due to their
desirability for mates or allies, while those characteristics may not be as impor-
tant in the selection of mates or allies in a different cultural context. In addi-
tion, a given artifact may be an accurate signal of reproductive fitness in certain
contexts while not a fitness signal in others based on that artifacts link to
honest representations of highly desired fitness characteristics.Therefore, proxy
markers of reproductive fitness are not universal. The fitness of individuals will
be based on the cultural (including environmental) contexts, and individuals
who signal attributes that are the most directly linked to reproductive fitness
in those contexts will likely be selected as mates or allies over individuals who
do not signal fitness within that cultural context. As archaeologists, therefore,
we cannot simply classify particular items or artifact classes as costly signals
or non-costly signals. Instead, arguments for the past existence of costs and
signaling power must be developed on a case-by-case basis to both predict
expected signals in a given context as well as evaluate signaling power models.
To this end, we need a framework for approaching information exchange and
signaling in the past that can examine the signaling power of lithics in various
cultural contexts by looking at individual artifacts, entire assemblages, and the

way in which those objects were made, used, and discarded in thepast.
Although costly signaling theory has been previously introduced into
archaeological and lithic analysis literature as a potential explanation for seem-
ingly wasteful behavior, there still is much work to be done to link the theo-
retical approach to the archaeological record. In the next section, I provide a
framework (specific to material culture signaling) for investigating costly sig-
naling using objects made of stone in the archaeological record. This approach
is meant to complement, not supplant, investigations into costly behaviors such
as big game hunting. Instead, the framework will allow for rigorous evaluations
of costly signaling as an explanation of material culture use in the past. It is in
the operationalizing of costly signaling models that we can begin to take appli-
cations of costly signaling theory from just-so-stories (Codding and Jones
2007) to fully developed archaeological theory that is ripe with predictive and
testable models.
Costly Signaling: A General Framework

I present a multilevel and nested framework for assessing the signaling power
of lithics. In Figure11.2, the first-order variables as well as their relationships
with each other and their contributions to signaling power are displayed. All of
the variables in Figure11.2 are amalgamations of several lower order variables
that contribute to the primary variables (Table11.1). This framework has the
benefit of providing a common set of variables that are potentially relevant to
the design of data collection strategies, the collection of data, and the analysis
of lithic objects and their signaling power. Similar frameworks have been pro-
posed for the study of other multivariate, dynamic, and complex social sys-
tems (such as socioeconomic systems Ostrom 2009) primarily because this
approach allows for the recognition of complexity, the potential to quantify
aspects of social systems, and the ability to generate and test hypotheses about
human behavior.
This general framework of signaling with lithics builds on previous work
with costly signaling theory, lithic technological organization, and social infor-
mation exchange, and identifies several primary variables that contribute to
the signaling power of an object. This important step in approaching costly
signaling in lithics is the recognition of (1) the constellation of variables that
contribute to the signaling power of an object, (2) the fact that the variables
are interrelated, and (3) the dynamic nature of the connections among the
variables and signaling power based on the abundances of the materialized
signals in circulation.
The four first-order variables are (1) the audience of the signal, (2) the
contexts of use and deposition of the object, (3) the physical attributes of
Social Setting
Audience Attributes
Context Signal Production

(Information)
Ab e
un nc
da da
n un
ce Ab
Materialization
of Signal
Material Culture
Signaling Power
Low High
11.2. A general framework for studying costly signaling behavior with material culture. The
reproductive fitness characteristics in a particular social setting influence, and are influenced
by, the variables that contribute to the costs and benefits of materializing signals. The more
abundant the materialized signals are in circulation, the lower the signaling power of that signal.
There is a dynamic and complex relationship among these variables, the fitness information, and
material culture that requires a holistic approach to material culture information exchange to
identify and explain costly signaling in thepast.
table11.1.Variables that archaeologists can study within the generalized framework to iden-
tify and explain material culturebased costly signaling behavior in thepast
First-order variables Variables that contribute to the cost and benefit of signals
Social organization Units of analysis
Cultural context
Artifact attributes Size
Color
Sheen
Raw material (distance to source; distribution/access)
Production Skill level required
Time to manufacture
Producer/consumer relationship
Cost of maintenance
Audience Relationship between signaler and recipient
Population size/density
Artifact attributes
Context Use context (daily life; ritual/event)
Deposition context (caches, graves, votive offerings; loss; discard)
the object itself, and (4) the structure and organization of production and
acquisition of an object. The relationships of these first-order variables to the
signal, the materialization of the signal, signaling power, and each other can
be seen in Figure11.2. The signaling power of an object is diminished by the
increasing abundance of objects of that type in circulation as the potential for
cheating and dishonest signals increases (which also affects all of the first-order
variables). Each of the first-order variables are affected by several other second-
order variables (seen in Table11.1), which are variables that can potentially be
evaluated qualitatively or quantitatively using archaeological data. When com-
bined, the connections between the second-order and first-order variables can
be incorporated into an analysis of signaling power as a means of evaluating
whether a given object is a costly signal ornot.
When attempting to identify archaeological correlates of costly signals, we
must recognize that there is no single variable that can mark an item as a costly
signal, and therefore no single archaeological correlate of an items use as a costly
signal. There is a complex suite of relationships among these variables and the
archaeological record. This complexity results in a nebulous conceptualization
of signaling power, materiality, and costly signaling behaviors where all variables
contribute and are affected by social, demographic, and technological organiza-
tion. As a result, there is no easy set of boxes to check or trait list to examine to
determine whether or not an object or artifact type was used as a costly signal.
My discussion of the interrelatedness and interconnectivity of variables and their
archaeological correlates will be inadequate to capture the entirety of the system,
but it does provide a starting point for researchers interested in these issues.
The framework begins with an assessment of the social setting. Next, the
characteristics that are associated with success and fitness in that context
become the central focus of signaling. In the process of materializing a signal,
attributes of the artifact and elements of the production system impose costs
on the signaler while the benefits are amplified by the object attributes, the
production system, the audience, and the use context of the object. These
materialized signals are either costly signals (if they are directly related to fit-
ness enhancing variables for the individual or group) or non-costly signals
(if they do not directly impact fitness). This evaluation, however, is not stable.
The circulation of objects that are materialized costly signals is necessarily
restricted to the people who can incur the costs of acquisition or produc-
tion. However, with more and more objects entering circulation, the signaling
power of all of those objects will be diminished owing to lower acquisition
costs and decreased recognition of uniqueness.This feedback loop between the
abundance of materialized signals and the variables that contribute to signaling
power impacts the diachronic stability of a system and can move objects that
were once costly signals lower on the signaling power continuum to the point
where they are non-costly signals (and vice versa).
Assessing Costly Signaling in Lithic Technology

Utilizing the general framework, we can begin to evaluate the utility of costly
signaling theory in studies of lithic technological organization. The social
context in which fitness relationships are negotiated is an extremely impor-
tant factor in the presence or absence and type of costly signaling. The social
organization of the group in which the information exchanges take place
impacts the majority of primary variables that contribute to signaling power
(such as audience size and structure, production organization, and the contexts
of interaction). The materialization of signals in egalitarian hunter-gatherer
groups will differ significantly from that in more complex stratified societies.
In hunter-gatherer communities the needs of mobility limit the size of lithics
that can be transported (potentially resulting in less ostentatious signals).
One of the benefits of information exchange using material culture is the
quick nonverbal conveyance of information to people who would otherwise
be unable to know; including information about reproductive fitness. With
smaller group sizes and fewer social roles (when compared with more com-
plex societies) this can limit the payoff of costly signals in band level societies.
On the other hand, periodic aggregations of normally distant groups provide a
unique opportunity for signaling, mate acquisition, and alliance formation. In
addition, the important role of lithics in the material culture assemblage as well
as the economic lifeways of most hunter-gatherer groups throughout prehis-
tory may have made it an important vehicle for social information exchange.
In more complex societies, the number of social roles and identities that can be
conveyed through material culture can be large. The payoff of a signal in these
contexts can be higher owing to larger audience sizes and the more active use
of material culture items (and their collection) to convey information about
fitness to others. However, the diminishing role of lithics as a result of techno-
logical innovations (such as metallurgy) can also impact their signaling power.
In different social settings there are different emphases on certain identities
that may change the locus of information exchange. An individuals reproduc-
tive success may be directly related to his or her own ability to signal or it
may be the household that is the relevant unit of identity; and the success of
the household may be placed over the success of an individual (though it is
expected that the individual will receive a real or perceived benefit from asso-
ciation with the household). In these cases, it is likely that signals that are asso-
ciated with lineage membership and increase the status of the household will
be emphasized. The differences among different levels of social organization
are expected to be reflected in the form of different materializations of costly
signals as well as the majority of variables that contribute to their signaling
power. For this reason, it is necessary to evaluate any potential signaling within
the specific social and cultural context in which the objects wereused.
The structure and organization of stone tool production also affects the
signaling power of lithics. Formalized tools are more likely than expedient
tools to contain social information. In addition, any restrictions on access to
raw materials and in the role of producers could have significant ramifications
for the utility of lithics as costly signals. With higher degrees of specialization
comes the ability for individuals to control production as well as the byprod-
ucts of production. Other restrictions on the production of particular stone
tools arise from the difficulty of working the material, the skill required to
make a tool, and time investment in production.The more bounded, restricted,
and costly production of stone tools is, the more likely the tools are to carry
information that can impact reproductive fitness. Archaeologically, the pres-
ence of production areas, raw material caches, and stores of finished items may
indicate that there are restrictions on who has access to raw materials and fin-
ished products that can turn the artifacts into honest markers of this potentially
fitness enhancing characteristic.
Characteristics of artifacts can impact the visibility of a signal the larger
and more visible the artifact (perhaps based on size, color, or sheen), the higher
the payoff in terms of numbers of people who receive the information. Lithic
raw material sources that are spatially discrete or distant are easier for individu-
als or small groups of individuals to create control over. Exotic raw materials
that are difficult to acquire may be more associated with differential access
that may also be linked with other useful characteristics (i.e., alliances, wealth,
status). Raw materials that are costly to work (i.e., require high degrees of skill,
difficult to knap) also may have high signaling power. Exotic materials and
rare items are more likely to be used as costly signals than local and abundant
resources. Spatially distinct sources or trade routes may produce assemblages
that show differential access to raw materials, such as significant differences in
exotics or different raw material types in different houses across asite.
The audience to which a signal is displayed is affected by multiple variables.
The overall audience size (potentially measured as population size or population
density) will impact the signaling power of a signal (the larger the audience, the
higher the signaling power). In different cultural contexts, there may be differ-
ent audiences of a signal. For example, hunter-gatherer bands may have lower
audience sizes in daily contexts but also may have important audiences during
times of aggregation or interaction with neighboring communities. In addi-
tion, in more populous social settings (such as sedentary villages), there may
be more people who will see a signal, which will increase the benefit to the
signaler when compared with signalers in contexts of low population density
(even with the same cost). The audience of a signal is dependent on the vis-
ibility of the signal (the more visible the lithic object, the more people will see
it) and the context in which it is displayed (the more people who are around
when a lithic object is used/displayed, the more people will see it). Lithics used
in a common daily activity will be seen by a more restricted (and likely highly
related) audience while lithics used in highly visible ritual events (such as feasts
and funerary events) will have a larger and more diverse audience.
Factors in assessing the use of an item as a costly signal in the archaeo-
logical record are its use and depositional context. The same object found in
different contexts may have had very different roles in the social exchange of
reproductive fitness information. One way in which context matters is if there
is evidence that the object is used in different ways that are not explained as
functional. Costly signals would likely have had different patterns of use-wear
and retouch than non-costly lithics. Objects used in ritual or ceremonial per-
formances may have been stored in locations associated with other parapher-
nalia. Similarly, iconography or other evidence that may hint at the role of
lithics within ritual, feasting, or other highly attended and visible events can
help determine whether or not lithics were used as costly signals.
Another way in which context matters is the way in which an artifact is
deposited. Caching and hoarding behavior may be associated with conspicuous
consumption and destruction of artifacts that may be interpreted as a costly
behavior. Mortuary contexts are also an area where the destruction of wealth,
rare items, and signs of status may indicate costly signaling, though it is impor-
tant to recognize that a variety of other information (from group membership
to individual identity) can be encoded in grave goods. With both the use and
discard of an artifact, it is not where it was found or how much use-wear that
matters; it is the deviation from the expected evidence of active decisions by
people in the past to use the artifact as a vehicle for information exchange
particularly as a fitness enhancing signal that makes it a costly signal.
When combined, the assessments of costliness and payoffs from signaling
displays for each of these variables can inform our understanding of whether
or not a signal is a costly signal or a non-costly signal. However, this assessment
is dynamic and mutable depending on artifact abundance, stability in the social
setting, and the changes in the relationship of the variables throughtime.
Potential Applications of Costly Signaling in Lithic

Technology
As previously mentioned, Kohn and Mithen (1999) and Waguespack et al.
(2009) have attempted to link social information exchange and lithics in the
context of sexual selection and costly signaling theory.Although these research-
ers have not systematically considered all of the variables that impact signaling
power, their work highlights existing attempts to utilize costly signaling theory
to explain some phenomena observed in the archaeological record. Acheulean
handaxe production and the use of stone as a projectile are not the only con-
texts (if they are contexts for this at all) where we anticipate seeing evidence
of costly signaling strategies. In this section I would like to provide a brief and
broad survey of archaeological contexts and examples where costly signaling
may be applied. This is not meant to argue that all of these cases are evidence
of costly signals in lithics, only that these are the kinds of attributes in artifacts,
assemblages, and contexts that are more likely than others to be venues for
wasteful display and conspicuous consumption involving lithics.
Paleoindian Caches: Objects and Deposition

Clovis points, with their distinct basal fluting, were likely more costly to man-
ufacture than other point forms. However, given their widespread distribution,
it may not be appropriate to consider all fluted points as costly signals. In some
cases, however, there does appear to be a distinct social function based on
their attributes, production, and depositional context. In 1987, a cache of fluted
Clovis points from an undisturbed context were discovered in East Wenatchee,
Washington (the Richey-Roberts Clovis Cache; Mehringer and Foit 1990).
These points were distinctly larger in size than most Clovis points, were made
of very high quality raw materials, showed little evidence of retouch, were
found in a high quantity and density that is extremely rare, and were found in
association with ochre (Mehringer and Foit 1990). Together, these suggest that
these artifacts may have been used to communicate social information rather
than as utilitarian implements. Other cases of ritual fragmentation and deposi-
tion of pristine Paleoindian bifaces in the Great Lakes region has been argued
to represent ritual activity (Ellis 2009). The link between ritual performance
and lithic technologies suggests that stone objects may have been embed-
ded with important individual or communal information that was negotiated
through the use of material culture. Individuals or groups that controlled access
to these objects or to the ritual performance may have benefited from these
instances of conspicuous consumption. Larger considerations of stone point
artifacts, their relationship to other Paleoindian lithic technologies around
North America, the social setting of Paleoindian communities, and reproduc-
tive fitness characteristics for the people making and using these points must
be undertaken before they are interpreted as costly signals.
Biface Caches: Mound 72 at Cahokia

During excavations at Cahokias Mound 72, a series of hafted biface caches
were recovered (Fowler etal. 1999). Two caches (caches 1550 and 1551)were
in direct association with burials while a third (cache 1970) appeared to be
deposited later than the other caches without being associated with a burial
(Ahler 1999:102103). The caches were made of several hundred bifaces, many
of which were not heavily curated and numerous ones that were made of
exotic raw materials. Previous work has documented a significant amount of

typological, depositional, raw material, morphological, and contextual variabil-
ity among the different caches (Ahler 1999). This variability has been inter-
preted to imply that the different caches represented different social functions
for their deposition (Ahler 1999:112). The diversity, use of exotic materials,
limited retouch, and association with funerary performance, especially in the
1550 and 1551 caches, is suggested to represent larger audience sizes, repre-
sentations of larger exchange networks, nonutilitarian uses of the lithics, and
costly investment in acquiring and depositing points made of exotic raw mate-
rials. The performance of deposition of these points was likely part of a series
of large-scale public events that would have brought together many com-
munities, suggesting that there would have been large audience sizes (Brown
2006; Pauketat and Alt 2004). Given the abundance of lithics in the mortuary
events and other ritual activities at Mound 72, it may be possible that they
were utilized in conspicuous consumption events. Similar hafted biface caches
have been reported from other Mississippian sites, such as the Spiro site (Ahler
1999), which may suggest a larger regional pattern of signaling with stone
points. Given the variability among caches and the deposition of points in
non-cache contexts during the Mississippian period, it is likely inappropriate
to discuss point use in a general sense as an exhibition of costly signals. Instead,
contextually specific examinations of artifact attributes, contexts, production
organization, and audiences may elucidate where and when hafted bifaces
were used as costly signals during the Mississippian period.
Kimberly Projectile Points: Display and Fitness

In pre- and postcolonial contact periods in northwest Australia, Aboriginal
communities manufactured Kimberley points (Harrison 2002). Kimberley
points possess a distinctive pressure-flaked margin that results in denticulate
or serrate margins (Harrison 2002). These points have been argued to play
important roles in functional resource extraction (Akerman 2008), demon-
strating masculinity (Harrison 2002), displaying social identities (Harrison
2002), and negotiating colonial interactions (Harrison 2006).The utilization of
these objects in a variety of contexts, their unique manufacturing process, and
in some cases the large size and use of exotic raw materials for their manu-
facture may suggest that these objects were used to convey social information.
For these objects to be costly signals, however, they must convey information
that is linked to reproductive fitness in these social contexts. If these points
embodied a desired quality of masculinity, they may have been important tools
in displaying that quality. Work to evaluate costliness, audience variability, and
contextual utility of Kimberley points may provide evidence to link this point
technology to either costly or non-costly signaling.
Kintampo Points: Formalization among Informal Assemblages

The Ceramic Late Stone Age Kintampo Complex lithic technology from
Ghana, West Africa is characterized by expedient and informal tools produced
through bipolar reduction (Casey 1998). Not all tools in these assemblages
were informal and expedient; the people who produced the Kintampo assem-
blages did produce highly formalized hafted biface points. The co-occurrence
of large quantities of informal tools (presumably used for the majority of utili-
tarian tasks) and only one type of highly formalized points suggests that the
points may have been used to transmit socially relevant information (Casey
1998). Given the relatively higher investment in time to produce these formal-
ized tools, their potentially visible role associated with hunting forays, and the
increase in audience sizes with sedentism during the Ceramic Late Stone Age,
it is possible that these points were utilized as costly signals of manufacturing
skill and hunting ability.
Northwestern California Bifaces: Exotics and Performance

In ethnographic and archaeological contexts in northwestern California,
obsidian bifaces were considered important markers of wealth and social
rank (Hughes 1978; Kroeber 1905; Rust 1905; Thompson 1916). Obsidian was
an important tool stone in much of California before and at contact with
Europeans; however, the lack of sources in northwestern California meant
that in this region the stone could be acquired only through long-distance
exchange networks or long logistical forays to source sites (Hughes 1978).
In these communities, large obsidian bifaces (also referred to as blades in
the literature) served an important role in the construction and maintenance
of individual and communal identity and were often a direct embodiment
of fitness-related characteristics of wealth and status. These important mark-
ers were used in a variety of contexts, and their value was based off of their
sources, their size, their color, their completeness, and their display (hafted or
worn from necklaces) (Hughes 1978). Early ethnographic work by Horatio
Rust, who made an attempt to acquire several of these items from indigenous
Californians, provides an incredible insight into the social function of these
objects, their use and depositional patterning, and the awareness by these local
communities of the link between the material objects and the information
that they signal:
These obsidian blades pass from father to son, with hereditary rank, and
are retained with pride as heirlooms; consequently it was only by much
persuasion and considerable expenditure that they could be obtained. In
several instances the Indians regarded the blades as tribal property, and in
one case I found it impossible to persuade the holder to part with the one
in his possession at any price.
One old Indian, living alone in abject poverty, exacted a promise that I
would not tell his neighbors that I had bought his blade. He said: Now
they call me rich. If they know I sell him, they say He poor Indian-no
account. The promise was given and his reputation for wealth and
honor saved. (Rust 1905:688689)
This quote highlights the blurred lines between the information (in this case,
wealth and honor) and the material object used to signal. Obsidian bifaces
were only brought out for dances and during normal daily events they were
stored with much care (Rust 1905), which highlights their use as signals at
times with the largest and most attentive audiences. The deposition of these
objects in mortuary contexts may mark the end of their use by the interred
individual as a costly signal, or might mark a conspicuous consumption event
by the people burying the deceased. The use of these stone points made of
difficult to acquire and costly raw materials in highly visible displays during
periods of aggregation and ritual suggest that these objects may have been
materialized signals of reproductive fitness characteristics of exchange network
access, wealth, and status.
Discussion
Costly signaling is a strategy of resource consumption that people in the past
utilized to increase social capital and reproductive fitness, and the identification
of this phenomenon is only the first step in realizing the explanatory power
of this theoretical perspective. There is important evidence of synchronic pat-
terning in the materialization and utilization of costly signals that can provide
insight into communal social dynamics. Measures of differential access and uti-
lization across a community provide evidence of different signaling strategies
being employed. Diachronic patterning in materialized costly signaling can
also provide evidence of stability or dynamism in signaling strategies. Changes
in the information being conveyed, the characteristics that impact reproduc-
tive fitness, or the material culture assemblages of a group can potentially be
used to reconstruct signaling practices and complex behaviors in thepast.
Just because a raw material is exotic, rare, spatially discrete, or difficult to
work does not automatically make it a costly signal. Information must be
encoded in the material by the individuals making, using, and discarding the
artifacts. Materials that have these qualities, however, are more likely to be
used as costly signals than materials that do not fit these qualifications. In lithic
assemblages where these material conditions are observed, it may be appro-
priate for archaeologists to apply models and hypotheses from costly signaling
theory.
Costly signaling theory is not intended to explain all or even most behavior
in the past. This is particularly true for its utility in examining the relation-
ship among information exchanges, people, and lithics. Its power as a theo-
retical approach lies in accurate application to the archaeological record in
cases where it is appropriate. With costly signaling theory, it is possible to (1)
track the development of social practices in which people embed lithics with
information, (2) develop hypotheses about the stability or dynamism of sig-
naling practices using material culture, and (3) reconstruct the role of lithics in
the construction, modification, or negotiation of identities and interpersonal
relationships.
There must be careful considerations of the archaeological record from both
directions: the soundness of applying the theory to a particular set of research
questions as well as the suitability of the data set for addressing these issues. I
encourage researchers who are looking to costly signaling theory as a mecha-
nism for explaining seemingly wasteful behavior in the archaeological record
to do so carefully, with diligence to the theory as well as the data. Although
its applicability in studies of lithic technological organization is likely not as
wide reaching as its use in studying other types of archaeological materials
(such as personal adornment items), there are still particular criteria that some
archaeological cases meet where not only does the use of lithics in negotiating
social relationships lend itself to costly signaling theory, but also perhaps costly
signaling theory can best explain these phenomena.
Evolutionary theory can inform our models and approaches to lithic tech-
nological organization in the past. Models and approaches derived from evolu-
tionary theory must be critically incorporated to anthropological archaeological
investigations of the past, modified to realize the specifics of human cultural
behavior and the structure of the archaeological record. When done well, as
evidenced by this volume, these approaches can inform our understanding of
past human activities and the complex relationship between people, their envi-
ronment, and the materialized manifestations of their cultural behavior.
Conclusion
In this chapter I have attempted to explore the utility and method of applying
costly signaling theory to studies of lithic technological organization. In the
process, I have highlighted the potential utility of looking at fitness-related
information embedded and signaled in stone objects. I have also warned about
the dangers of inappropriately explaining past behavior as costly signaling
when certain social and material criteria are not met. Studying costly signaling
in lithics is a difficult task, made even more difficult by the multiple lines of
evidence required and the ways in which other social phenomena and differ-
ent types of artifacts can affect the presence of costly signaling and the way in
which it was materialized. Although I urge caution in applying this concept to

the archaeological record to avoid perpetuating a series of just-so-stories, the
causal mechanisms, evolutionary stability, and explanatory power of costly sig-
naling theory are a valuable approach that can provide new insights into lithic
technological organization.
Acknowledgments
I thank N. Goodale and W. Andrefsky Jr. for their invitation to contribute
to this volume and their patience and guidance. Thanks are in order for
J. Speth, H. Wright, I. Kuijt, C. Barrier, A. Wright, A. Compton, T. Landvatter,
R. Hughes, M. Conway, and A. Lemke for their input throughout the devel-
opment, writing, and editing of previous drafts of this chapter. All errors that
remain are mine. I also thank J. OShea, C. Sinopoli, and the rest of the faculty,
students, and staff at the University of Michigan Museum of Anthropology.
References
Ahler, Steven R. 1999. Projectile Point Bliege Bird, Rebecca, Eric Alden Smith, and
Caches. In The Mound 72 Area: Dedicated and Douglas W. Bird. 2001.The Hunting Handicap:
Sacred Space in Early Cahokia, edited by M. Costly Signaling in Male Foraging Strategies.
Fowler, pp. 101115. Illinois State Museum, Behavioral Ecology and Sociobiology 50:919.
Springfield. Bliege Bird, Rebecca, Douglas W. Bird, Eric Alden
Akerman, Kim. 2008. Missing the Point or Smith, and Geoffrey C. Kushnick. 2002. Risk
What to Believe the Theory or the Data: and Reciprocity in Meriam Food Sharing.
Rationales for the Production of Kimberley Evolution and Human Behavior 23:297321.
Points. Australian Aboriginal Studies 2:7079. Boone, James L., and Eric Alden Smith. 1998. Is
Barton, C. Michael. 1997. Stone Tools, Style, and It Evolution Yet? A Critique of Evolutionary
Social Identity: An Evolutionary Perspective Anthropology. Current Anthropology
on the Archaeological Record. In Rediscovering 39:S141S173.
Darwin: Evolutionary Theory in Archaeological Boyd, Robert, and Peter J. Richerson. 1985.
Explanation, edited by G. Clarke and M. Culture and the Evolutionary Process. University
Barton, pp. 141156. Archaeological Papers of of Chicago Press, Chicago.
the American Anthropological Association,7. Brown, James A. 2006. Wheres the Power in
Bernstein, David J. 1984. Utilitarian Lithics as Mound Building? An Eastern Woodlands
Prestige Items: A Preliminary Examination Perspective. In Leadership and Polity in
of Some Lower Central American Mississippian Society, edited by B. M. Butler
Mortuary Practices. BAR International Series and P. D. Welch, pp. 197213. Center for
226:203219. Archaeological Investigations, Southern
Bettencourt, Luis M.A., Jose Lobo, Dirk Helbing, Illinois University.
Christian Kuhnert, and Geoffrey B.West. 2007. Casey, Joanna. 1998. Just a Formality: The
Growth, Innovation, Scaling, and the Pace of Presence of Fancy Projectile Points in a Basic
Life in Cities. PNAS 104(17):73017306. Tool Assemblage. In Gender in African Prehistory,
Bliege Bird, Rebecca, and Eric Alden Smith. edited by S. Kent, pp. 83104. AltaMira Press,
2005. Signaling Theory, Strategic Interaction Walnut Creek,CA.
and Symbolic Capital. Current Anthropology Codding, Brian, and Terry Jones. 2007. Man
46(2): 221248. the Showoff? Or the Ascendance of a
Just-So-Story: A Comment on Recent Signaling. Journal of Theoretical Biology

Applications of Costly Signaling Theory in 213:103119.
American Archaeology. American Antiquity Gintis, Herbert, Samuel Bowles, Robert Boyd,
72(2):349357. and Ernst Fehr. 2003. Explaining Altruistic
Dawkins, Richard. 1987. The Blind Watchmaker. Behavior in Humans. Evolution and Human
W. W. Norton. Behavior 24:153172.
Dawkins, Richard, and John R. Krebs. 1978. Grafen, Alan. 1990. Biological Signals as
Animal Signals: Information or Manipulation. Handicaps. Journal of Theoretical Biology
In Behavioural Ecology: An Evolutionary 144(4):517546.
Approach 1st Edition, edited by J. R. Krebs and Harrison, Rodney. 2002. Archaeology and the
N. B. Davies, pp. 282309. Oxford: Blackwell Colonial Encounter: Kimberley Spearpoints,
Scientific. Cultural Identity, and Masculinity in the
Dunnell, Robert C. 1980. Evolutionary Theory North of Australia. Journal of Social Archaeology
and Archaeology. Advances in Archaeological 2(3): 352377.
Method andTheory,Vol. 3, edited by M. B. Schiffer, Harrison, Rodney. 2006. An Artefact of
pp. 3599. Academic Press, NewYoek. Colonial Desire? Kimberley Points and
Durham, William H. 1991. Coevolution: Genes, the Technologies of Enchantment. Current
Culture, and Human Diversity. Stanford Anthropology 47(1):6388.
University Press, Stanford. Hasson, Oren. 1994. Cheating Signals. Journal of
Earle, Timothy. 2004. Culture Matters: Why Theoretical Biology 167:223238.
Symbolic Objects Change. In Rethinking Hawkes, Kristen. 1990. Why Do Men Hunt?
Materiality: The Engagement of Mind with Some Benefits for Risky Strategies. In Risk
the Material World, edited by E. DeMarrais, and Uncertainty in Tribal and Peasant Economies,
C. Gosden, and C. Renfrew, pp. 153165. edited by E. Cashdan, pp. 145166. Westview
McDonald Institute Monograph, Cambridge. Press, Boulder,CO.
Ellis, Christopher. 2009. The Crowfield and Hawkes, Kristen. 1991. Showing Off: Tests of
Caradoc Sites, Ontario: Glimpses of Palaeo- an Hypothesis about Mens Foraging Goals.
Indian Sacred Ritual and World View. In Ethology and Sociobiology 12:2954.
Painting the Past with a Broad Brush: Papers in Hawkes, Kristen. 1993. Why Hunter-Gatherers
Honour of James Valliere Wright, edited by D. L. Work: An Ancient Version of the Problem
Keenlyside and J. Pilon, pp. 319352. Canadiam of Public Goods. Current Anthropology
Museum of Civilization, Gaineau. 34:341361.
Fowler, Melvin L., Jerome Rose, Barbara Vander Hawkes, Kristen, and Rebecca Bliege Bird.
Leest, and Steven R. Ahler. 1999. The Mound 2002. Showing Off, Handicap Signaling, and
72 Area: Dedicated and Sacred Space in Early the Evolution of Mens Work. Evolutionary
Cahokia. Illinois State Museum, Springfield. Anthropology 11(2):5867.
Fried, Morton H. 1967. The Evolution of Political Henrich, Joseph. 2007. The Evolution of
Society. Random House, NewYork. Costly Displays, Cooperation, and Religion:
Gero, Joan M. 1989. Assessing Social Information Inferentially Potent Displays and Their
in Material Objects: How Well Do Lithics Implications for Cultural Evolution. Papers on
Measure Up? In Time, Energy, and Stone Tools, Economics and Evolution 721:150.
edited by R. Torrence, pp. 92105. Cambridge Hildebrandt, William R., and Kelly R. McGuire.
University Press, Cambridge. 2002.The Ascendance of Hunting During the
Gintis, Herbert. 2000. Strong Reciprocity and California Middle Archaic: An Evolutionary
Human Sociality. Journal of Theoretical Biology Perspective. American Antiquity 67:231256.
206:169179. Hildebrandt, William R., and Kelly R. McGuire.
Gintis, Herbert, Eric Alden Smith, and Samuel 2003. Large-Game Hunting, Gender-
L. Bowles. 2001. Cooperation and Costly Differentiated Work Organization, and the
Role of Evolutionary Ecology in California Machin, Anna J. 2008. Why Handaxes Just Arent
and Great Basin Prehistory: A Reply to That Sexy: A Response to Kohn and Mithen
Broughton and Bayham. American Antiquity (1999). Antiquity 82:761766.
68:790792. Mauss, Marcel. 1924. The Gift: Forms and Functions
Hodgson, Derek. 2009. Symmetry and Humans: of Exchange in Archaic Societies. London: Cohen
Reply to Mithens Sexy Handaxe Theory. andWest.
Antiquity 83: 195198. Maynard Smith, J., and D. G. C. Harper. 1995.
Holm, Lena. 1994. Stone Artefacts as Transmitters Animal Signals: Models and Terminology.
of Social Information: Towards a Wider Journal of Theoretical Biology 177:305311.
Interpretation with a North Swedish Example. McGuire, Kelly R., and William R. Hildebrandt.
Current Swedish Archaeology 2:151158. 2005. Re-Thinking Great Basin Foragers:
Hughes, Richard E. 1978. Aspects of Prehistoric Prestige Hunting and Costly Signaling during
Wiyot Exchange and Social Ranking. The the Middle Archaic Period. American Antiquity
Journal of California Anthropology 5(1):5366. 70(4):695712.
Kantner, John. 2003. Biological Evolutionary McGuire, Kelly R., William R. Hildebrandt,
Theory and Individual Decision-Making. and Kimberly L. Carpenter. 2007. Costly
In Essential Tensions in Archaeological Method Signaling and the Ascendance of No-Can-Do
and Theory, edited by T. L. VanPool and C. S. Archaeology: A Reply to Codding and Jones.
VanPool, pp. 6788. University of Utah Press, American Antiquity 72(2):358365.
Salt LakeCity. Mehringer, Peter J., Jr., and Franklin F. Foit, Jr.
Kaplan, Hillard, and Kim Hill. 1985. Food 1990.Volcanic Ash Dating of the Clovis Cache
Sharing Among Ache Foragers: Tests of at East Wenatchee, Washington. National
Explanatory Hypotheses. Current Anthropology Geographic Research 6(4):495503.
26(2):223245. Mithen, Steven. 2008. Whatever Turns You on:
Kelly, Robert L. 2000. Elements of a Behavioral A Response to Anna Machin, Why Handaxes
Ecological Paradigm for the Study of Just Arent That Sexy. Antiquity 82:766769.
Prehistoric Hunter-gatherers. In Social Theory Neiman, Fraser. 1997. Conspicuous
in Archaeology, edited by M. B. Schiffer, pp. 63 Consumption as Wasteful Social Advertising:
78. University of Utah Press, Salt LakeCity. A Darwinian Perspective on Spatial Patterns
Kohn, Marek, and Steven Mithen. 1999. in Classic Maya Terminal Monument Dates.
Handaxes: Products of Sexual Selection? In Rediscovering Darwin: Evolutionary Theory in
Antiquity 73:518526. Archaeological Explanation, edited by G. Clarke
Krebs, John R., and Richard Dawkins. and M. Barton, pp. 267290. Archaeological
1984. Animal Signals: Mind-Reading and Papers of the American Anthropological
Manipulation. In Behavioural Ecology: An Association,7.
Evolutionary Approach 2nd Edition, edited by Ohtsubo, Yohsuke, and Esuka Watanabe. 2009.
J. R. Krebs and N. B. Davies, pp. 380402. Do Sincere Apologies Need to Be Costly?
Oxford: Blackwell Scientific. The Test of a Costly Signaling Model of
Kroeber, Alfred L. 1905. Notes. American Apology. Evolution and Human Behavior
Anthropologist 7:690695. 30:114123.
Lyman, R. Lee, and Michael OBrien. 1998. Goals Ostrom, Elinor. 2009. A General Framework for
of Evolutionary Archaeology: History and Analyzing Sustainability of Social-Ecological
Explanation. Current Anthropology 39:615652. Systems. Science 325:419422.
Lyman, R. Lee, and Michael OBrien. 2001. Parker-Pearson, Michael. 1999. The Archaeology of
On Misconceptions of Evolutionary Death and Burial. Sutton.
Archaeology: Confusing Macroevolution Pauketat, Timothy R., and Susan M. Alt. 2004.
and Microevolution. Current Anthropology The Making and Meaning of a Mississippian
42:408409. Axe-Head Cache. Antiquity 78:779797.
Plourde, Aimee. 2008. The Origins of Prestige Sosis, Richard. 2003. Why Arent We All
Goods as Honest Signals of Skill and Hutterities? Costly Signaling Theory and
Knowledge. Human Nature 19:374388. Religious Behavior. Human Nature-an
Preston, Beth. 2000. The Function of Things: Interdisciplinary Biosocial Perspective 14(2):
A Philosophical Perspective on Material 91127.
Culture. In Matter, Materiality, and Modern Thompson, Lucy. 1916. To the American Indian.
Culture, edited by P. M. Graves-Brown, pp. Cummins Print Shop, Eureka.
2249. London: Routledge. VanPool,Todd L., and Christine S.VanPool. 2003.
Quinn, Colin P. 2006. Exotics, Exchange, and Agency and Evolution: The Role of Intended
Elites: Exploring Mechanisms of Movement and Unintended Consequences of Action.
of Prestige Goods in the Interior Northwest. In Essential Tensions in Archaeological Method
Journal of Northwest Anthropology 41:207224. and Theory, edited by T. L. VanPool and C. S.
Rindos, David. 1989. Symbiosis, Instability, and VanPool, pp. 89114. University of Utah Press,
the Origins and Spread of Agriculture: A New Salt LakeCity.
Model. Current Anthropology 21:751772. Veblen, Thorstein. 1994 [1899]. The Theory of the
Robb, John E. 1998.The Archaeology of Symbols. Leisure Class. Dover, NewYork.
Annual Review of Anthropology 27:329346. Waguespack, Nicole M., Todd A. Surovell, Allen
Robb, John E. 1999. Material Symbols: Culture and Denoyer, Alice Dallow, Adam Savage, Jamie
Economy in Prehistory.Center for Archaeological Hyneman, and Dan Tapster. 2009. Making a
Investigations, Occasional Paper 26, Southern Point:Wood- versus Stone-Tipped Projectiles.
Illinois University Press, Carbondale. Antiquity 83:786800.
Rust, Horatio N. 1905. The Obsidian Blades of Wiessner, Polly. 1983. Style and Social Information
California. American Anthropologist 7(4):688695. in Kalahari San Projectile Points. American
Sackett, James R. 1982. Approaches to Style in Antiquity 48(2): 253276.
Lithic Archaeology. Journal of Anthropological Wiessner, Polly. 1985. Style or Isochrestic
Archaeology 1:59112. Variation? A Reply to Sackett. American
Sackett, James R. 1985. Style and Ethnicity in Antiquity 50(1): 160166.
the Kalahari: A Reply to Wiessner. American Wilk, Richard R. 1989. Decision Making and
Antiquity 50(1): 154159. Resource Flows Within the Household:
Smith, Eric Alden. 2000. Three Styles in the Beyond the Black Box. In The Household
Evolutionary Analysis of Human Behavior. In Economy: Reconsidering the Domestic Mode of
Adaptation and Human Behavior, edited by L. Production, edited by R.R. Wilk, pp. 2354.
Cronk, N. Chagnon, and W. Irons, pp. 2746. Westview, Boulder,CO.
Aldine de Gruyter, NewYork. Wilk, Richard R. 1996. Economies and Cultures:
Smith, Eric Alden. 2004.Why Do Good Hunters Foundations of Economic Anthropology.Westview,
Have Higher Reproductive Success? Human Boulder,CO.
Nature 15(4):343364. Wobst, Martin. 1977. Stylistic Behavior and
Smith, Eric Alden, and Rebecca L. Bliege Bird. Information Exchange. Anthropological Papers
2000.Turtle Hunting and Tombstone Opening: of the University of Michigan. 61:31742.
Public Generosity as Costly Signaling. Wobst, Martin. 1999. Style in Archaeology, Or
Evolution and Human Behavior 21:245261. Archaeologists in Style. In Material Meanings:
Smith, Eric Alden, and Bruce Winterhalder. Critical Approaches to the Interpretation of Material
1992. Natural Selection and Decision Making: Culture, edited by E.S. Chilton, pp. 118132.
Some Fundamental Principles. In Evolutionary University of Utah Press, Salt LakeCity.
Ecology and Human Behavior, edited by E. A. Zahavi, Amotz. 1975. Mate Selection: A Selection
Smith and B. Winterhalder, pp. 2560. Aldine for Handicap. Journal of Theoretical Biology
de Gruyter, Hawthorne,NY. 53(1):205214.
PartIV
Cultural Transmission and

Morphology
Twelve
An Analysis of Stylistic
Variability of Stemmed Obsidian
Tools (MATAA) on Rapa Nui (Easter
Island)
Carl P. Lipo, Terry L. Hunt, and Brooke Hundtoft
Easter Island, or Rapa Nui as it is known in modern traditional terms, is a small

(164 km2), remote island situated in the southeastern Pacific (Figure12.1). The
island is the easternmost edge of Polynesia, reflecting the origin, culture, and
language of the islands native population. Rapa Nui is most famous for its
hundreds of megalithic statues (moai) that were carved and transported, some
over several kilometers of rough terrain, reaching nearly every part of the
island. The scale of effort, and the investment represented, has baffled visitors
and researchers for centuries. Thus, emerged the so-called mystery of Easter
Island: How and why did people on such a remote and impoverished island
carve and transport these giant statues? A persistent and related theme has
been the islands prehistoric ecological devastation a dimension that added to
the mystic of enormous investment in the statues. The deforestation led early
visitors and modern writers alike to speculate that human-induced ecologi-
cal change led to population (and cultural) collapse. Jared Diamond (2005:6)
called it ecocide for ecological suicide committed by the prehistoric island-
ers. Although not the subject of this chapter, we have shown that ecocide
conflates the historically separated consequences of prehistoric deforestation
with post-European disease-induced population collapse (see Hunt and Lipo
2007, 2009). Part of the collapse narrative that developed, beginning with early
visitors, was that the island suffered constant conflict, violence, and intergroup
225
226 Li po et a l .
1500'0''W 1400'0''W 1300'0''W 1200'0''W 1100'0''W
M a r q u e s a s
Nuku Hiva
Hiva Oa
100'00''S Fatu Hiva
100'0''S
Tu a m o t u s
Manihii Napuka
Mataiva Puka Puka
S o c i e t y South Pacific Ocean
Maupiti
Bora-Bora Moorea
Raiatea Hao
Tahiti Reao
200'00''S 200'0''S
Maturei Vavao
Rurutu
Tubai Mangareva
Raivavae G a m b i e r Henderson
Ducie
Pitcairn
A u s t r a l s
Rapa
Rapa Nui
N
300'00''S 300'0''S
W E
S
Kilometers
0 250 500 1,000 1,500 2,000
1500'0''W 1400'0''W 1300'0''W 1200'0''W 1100'0''W
12.1.The Pacific Islands, showing Rapa Nui on the remote southeasternedge.
warfare leading to the imagined pre-European decimation of the population

(e.g., Flenley and Bahn 2003).
In this chapter, we examine mataa, a class of lithic artifacts found in abun-
dance on Rapa Nui. We consider how to build descriptions that enable us
to study patterns of transmission among prehistoric populations. Such pat-
terns of inheritance track social relations with implications for the evolution
of groups, competition, and scale of sociopolitical organization, in this case
among the prehistoric Rapa Nui population. Although these kinds of studies
are often conducted using assemblages of decorated ceramics (e.g., Lipo etal.
1997), here we show how analysis of variability unconstrained by performance
(style) allows us to measure aspects of inheritance related to the manufacture of
these artifacts. In the case of mataa from Rapa Nui, we demonstrate that it is
possible to reach falsifiable conclusions about the evolutionary dynamics that
shaped the archaeological record on this island.
Mataa are distinguished as stemmed obsidian artifacts with sharp edges, but
highly variable shapes (Figure12.2). Given their similarity to large spear-points
found elsewhere in the world based on the stem for hafting and sometimes
vaguely (incidentally) pointed shape mataa have been called weapons and
then cited as the evidence for warfare among prehistoric populations of this
An Analysi s o f S t y l i s t i c Va r i a b i l i t y o f S te mme d O b s i d i an T o o l s 227
12.2. Examples of mataa from Rapa Nui assemblages.
island. However, these artifacts are not analogous to lithic lancelets. Instead of
a thin bifacial profile created with careful percussion flaking, mataa are made
from large, thick flakes with the stem flaked out of one side relative to the bulb
of percussion.
Found in large numbers across the landscape of Rapa Nui, these artifacts
have been noted since the earliest European visitors described the island. For
example, Johann Georg Adam Forster, the son of a German naturalist who
accompanied Captain Cook on his second voyage to the Pacific, reported that
on Rapa Nui individuals had lances or spears made of thin ill-shaped sticks,
and pointed with a sharp triangular piece of black glassy lava (von Saher
1990:35).
A great diversity of shape has remained one particularly puzzling aspect of
mataa. Mataa shapes are highly inconsistent and vary from rounded to sub-
angluar to angular to complex. The earliest researchers assigned mataa shape
variation to what they conceived as ethnographic categories based on Rapanui
words (Routledge 1919). Later attempts to construct systematic classifications
228 Li po et a l .
have also focused on identifying types based on characterizations of overall

shape. None of these classification efforts produced useful categories.
After extensive analysis William Mulloy (1961:151), for example, concluded
that no significant clustering or correlations could be extracted.... the mate-
rial represents a continuous range of variation without objective natural order,
and that the only classification possible must involve the subjective selection
of ideal types from infinite series of possibilities, and the arbitrary reference of
intermediate for to one or another of these. As Mulloy astutely concluded,
manufacturing procedures dictated shapes, and differences in overall shape of
mataa were best explained by chance.
Overall shape is rarely a useful dimension for problem-oriented classifi-
cation. The forms of objects are limited by technological constraints of the
material, performance aspects that depend on the range of environments in
which the object is used, and simple idiosyncratic variability related to the
manufacturer and the process of production. In the case of mataa much of
the variability in the overall blade shape can be explained by the idiosyncratic
stages of manufacture (Bolt et al. 2006) plus differences structured by func-
tional variation related to the range and kinds of activities for which the tool
was primarily used: cutting, scraping, or some combination (as shown in use-
wear studies; see Church and Rigney 1994; Church and Ellis 1996). Focusing
our measurements on variability in the stem portion of the mataa potentially
avoids these problems. We make this case because the hafting portion need
only provide a means for attaching the object to a shaft, and then the formal
dimensions of the shaft are free to vary to some degree. If the instructions
for making mataa are inherited by individuals, then some of the variance in
the stem shape potentially reflects idiosyncratic preferences of manufacturers
passed from one person to another. Given that there are cognitive limits to
copying, small errors should be introduced into the instructions resulting in
drift over time (Eerkens and Lipo 2005).
Thus, we can derive sets of expectations for variability explicable in terms of
cultural inheritance. First, we can define something as potentially heritable if we
see similarity that occurs with a greater than random chance of being present.
Archaeologically we recognize heritability as continuity. Continuity can be
observed in two ways. First, it can be discerned by examining the distributions
of each measurement class through time (i.e., no random gaps). In addition, we
should see continuous distributions through space for assemblages in which
inheritance is a potential explanation.
Second, as Neiman (1995) has demonstrated, variability explicable by cul-
tural inheritance should also exhibit stochastic changes through time given
the effects of drift. This requirement means that there should be no single
direction of change, and that the rate of change should be scaled to popula-
tion size and copying error. This property of cultural inheritance is central to
explaining why frequency seriation works as it does. However, unimodality

is just one example of stochastic change, and is the easiest one to recognize.
In this sense, unimodality can be viewed as sufficient (i.e., it represents sto-
chastic change) but not necessary (i.e., other patterns of stochasticity are also
possible).
Analytic Procedures
To evaluate hypotheses for cultural transmission through time and across space,
we photographed mataa previously collected from a number of different parts
of Rapa Nui, including Vinapu, Rano Kau, Ahu Tautira, Orito, and Orongo
(Figure12.3). These mataa were collected by multiple field workers over the
past 40years and held in the P. Sebastian Englert Museum on Rapa Nui. In
addition, we photographed mataa we identified during pedestrian surveys of
11 parcels (parcelas) that are divided into two groups: Te Kahurea (18) and Te
Miro Oone (911).These surveys were conducted with University of Hawaii
and California State University Long Beach field school students in collabo-
ration with staff from the Englert Museum. Our survey of these parcels was
undertaken as a voluntary part of historic preservation and planning efforts
made in conjunction with land repatriation from the Chilean government to
native recipients of the Rapanui community. In total, our assemblages com-
prised photographs of 447 mataa.
Following a procedure used by OBrien and colleagues (2003) for measuring
fluted bifaces, we used Canvas (http://www.acdsee.com), a computer-based
Ahu Tautira
Parcela Te Kahurea (18)
Parcela Te Miro Oone (911)
Orito
Orongo Vinapu 0 4 16 Kilometers
Rano Kau
12.3. Location of mataa assemblages on Rapa Nui used in this analysis.
230 Li po et a l .
Blade Shape:
Flat
Curved
Pointed
Irregular
Shoulder Angle:
Curved
Angular
Stem Base:
Pointed
Flat
Subrounded
Rounded
Left Shoulder Angle

Right Shoulder Angle
Right Stem Length
Left Stem Length
Minimum
Stem
Width
Maximum
Stem
Width
12.4. Mataa measurements and class divisions.
drafting program, to make metric measurements from the scaled photographs.

Our measurements consisted of left and right stem length, minimum and max-
imum stem width, and right and left shoulder angles. We also made qualita-
tive assessments of the shape of the stem (e.g., square, sub-rounded, rounded,
pointed, broken/missing), shape of the shoulder angle (e.g., curved, angular),
and the shape of the blade (pointed, flat, curved, irregular, broken/missing)
(Figures12.4 and 12.5).
Classifications
Once we tabulated class occurrences using the measurements and observations
generated from the photographs of the mataa, we constructed deterministic
frequency seriations for the results of each classification. Differences in the
Stem Length Shoulder Angle

Class Value Class Value
A 2 cm 1 122.0
B 2.13 cm 2 122.0125.0
C 3.14 cm 3 125.0
D 4.1 cm
Stem Width Stem Length/Width Ratio

a 1.5 cm 1 Length > Width
b 1.62.0 cm 2 Length Width
c 2.12.5 cm 3 Length < Width
d 2.63.0 cm
e 3.13.5 cm
Stem Shape Shoulder Shape

P Pointed C Curved
R Round A Angular
S Square
SR Subrounded
12.5. Mataa class dimensions.
frequencies of classes between assemblages that are caused by differences in

sample size are evaluated using confidence values of = 0.01 calculated using
the beta distribution.
We constructed the first seriation based on a classification that consisted
of classes of quantitative dimensions for the ratio of the maximum length
and width of the stem and the maximum size of the shoulder angle mea-
sured relative to the stem. Stem measurements were divided into three clas-
ses: length > width, length width (values within 1cm), length < width.
To determine class divisions for shoulder angles, we first measured shoulder
angles for all mataa and determined the average value, 123.5 degrees. We
then arbitrarily divided angles into classes that consisted of values that were
larger than the average value plus and minus 1.5 degrees. In addition, Parcels
1, 4, and 5 had so few members that, given proximity, they were grouped
with Parcel6.
Analytic Results
The seriation analysis resulted in the generation of groups of assemblages
that reveal their spatial coherence (Figure12.6). Consistent with our expecta-
tions for cultural inheritance, the spatial coherence of these groups appears to
232 Li po et a l .
1AA 1AO 1AR 1OA 1OO 1OR 1RA 1RO 2AA 2AO 2OA 2OO 2RO 3AA 3AO 3OA 3OO
Parcel 11
Parcel 9
Parcel 10
Parcel 7
Parcel 6
Parcel 8
Orongo
Rano Kau
Orito
Vinapu
Ahu Tautira
12.6. Seriation solution for classes comprised of stem length/width ratios and shoulder angle
measures (right and left shoulder angle, respectively. O: > 125; R: 122125; A. < 122).
represent interaction among a population that decays with distance; that is,
the more distant the location, the more distinct the combination of class fre-
quencies. If spatial variation (distance decay) did not structure the frequencies
of classes, one would expect a solution in which assemblages formed a single
temporal order, with no necessary spatial patterning (Lipo 2001; Lipo et al.
1997).
Consistent with our model of cultural transmission, the seriation groups can
be viewed as aggregates of events between which spatial variability is greater
than the variability within individual groups (Figure12.7). Within each seria-
tion solution, however, the assemblages show stochastic continuity indicative
of inheritance. Given the chronology of the archaeological record on Rapa
Nui, the within-group assemblage-scale relations, however, are probably over-
lapping and thus represent contemporary assemblages to some degree.There
is little reason to suppose that each assemblage represents discrete, nonover-
lapping distributions in time. However, further analysis is required to better
address this concern.
As an additional analysis to test for congruence in seriations, we built a sec-
ond classification that employed length and width of the stems as dimensions
(Figure12.8). The results of this seriation are remarkably similar to the spa-
tial patterns observed with the classification that employed stem length/width
ratio and shoulder angles (Figure12.9). Although some aspects of the congru-
ence relate to physical constraints inherent in mataa, the degree of similarity
in the solutions supports the notion that the variability in stems is structured
by cultural transmission and by the preferences of individuals.
Finally, we constructed a third seriation based on the classification of the
qualitative measurements comprising stem shape and shoulder angle shape
Ahu Tautira
Orito
Rano Kau
12.7. Seriation groups for mataa classes comprising stem length/width ratios and shoulder angle
measures.
Aa Ab Ac Ba Bb Bc Bd Cb Cc Cd Db Dc Dd De
Parcel 7
Parcel 8
Parcel 6
Parcel 9
Parcel 11
Parcel 10
Ahu Tautira
Rano Kao
Orongo
Orito
Vinapu
12.8. Seriation solution for classes of mataa constructed with measures of stem length and
width.
(Figure12.10). Once again the pattern of the grouping of assemblages is con-

sistent with the other seriations (Figure12.11). This result also points to the
scale of transmission occurring as the set of instructions necessary to replicate
at least the entire stem portion of the mataa.
Significantly, we could not produce a seriation of any size that met the
requirements of the method when we used classes that included blade shape
234 Li po et a l .
Ahu Tautira
Orito
Rano Kau
12.9. Seriation groups for classes of mataa constructed with measures of stem length and
width.
PC RA RC SA SC SRC
Parcel 9
Parcel 11
Parcel 10
Parcel 7
Parcel 8
Parcel 6
Vinapu
Orongo
Orito
Rano Kau
Ahu Tautira
12.10. Seriation solution for qualitative classes of mataa consisting of stem shape and shoulder
shape dimensions.
as a dimension. Our inability to use blade shape in these analyses supports

Mulloys (1961:152) observation of the idiosyncratic nature for overall shape
of mataa, as well as the observation that the shape of the blades was not
transmitted with the instructions used to create the other portions of these
artifacts. The lack of continuous and stochastic distributions for classes that
Ahu Tautira
Orito

Rano Kau
12.11. Seriation groups for qualitative classes of mataa consisting of stem shape and shoulder
shape dimensions.
measure blade shape suggests that the overall form was not inherited along
with the other steps involved in mataa manufacture. Thus, we can suggest
that the shape of the blade portion of mataa is simply an outcome of flake
formation and that shape had neither cultural nor functional importance.
This observation is consistent with the use-wear that mataa were used as
cutting implements and that they only required sharp edges without regard
to shape.
Conclusions
Previous studies of overall shape have failed to produce discernible patterns
among stemmed obsidian artifacts (mataa) from Rapa Nui. We can now see
that this failure is not due to a lack of structure among the variability in the
artifact classes, but is largely a failure to construct explicit theory-based clas-
sifications. Classifications are the key to the process of explanation because
it is from these conceptual structures that meaning of measurements can be
specified. In this case, by using a model of cultural inheritance to structure
classes, make measurements, and generate empirical expectations, we are able
to produce results that can then be explained in terms of the model or not
explained, as seen in variability in blade shape.
Our results allow us to posit two substantive conclusions about Rapa Nui
prehistory. First, information about mataa manufacture was differentially
shared (transmitted) on a local scale distinguished in the portions of the
236 Li po et a l .
Ahu Tautira
Orito

Rano Kau
12.12. Spatial distributions of the mataa seriation groups on RapaNui.
island analyzed. In each of our stem classifications, spatial division of the

assemblages was required to produce a successful seriation order.
Particularly noteworthy is the observation that groupings were consistent
in each independent seriation. This result reveals a pattern in which prehis-
toric populations interacted on this small island, not as a large-scale social
network, but in relatively smaller social groups (Figure12.12). This conclusion
aligns with an emerging picture for the prehistoric populations of Rapa Nui as
small-scale groups organized more or less independently across the island (e.g.,
Lipo and Hunt 2005), in contrast to an island-wide chiefdom, as some have
speculated (e.g., Kirch 1984, 2000:275;Van Tilburg 1994:96).
Second, we posit that the scale of transmission among makers of mataa
comprised an instruction set for producing the stem portion of the artifact, but
not the entire object. Blade shape, it appears, freely varied and was probably
conditioned by the happenstance of the flake selected for the artifact in the
first place. This pattern supports the idea that mataa were not used extensively
as weapons in warfare, given the lack of a specific pattern in shape that would
be driven by the function of stabbing. This conclusion leads us to suggest that
whatever roles mataa played in prehistoric subsistence and settlement systems,
shape was not inherited and these tools performed multiple jobs (or no spe-
cific job at all). We suggest that mataa were used for generalized tasks (e.g.,
cutting and scarping of plant materials) and perhaps opportunistically in sym-
bolic displays related to nonlethal competition (i.e., they would not have been
effective weapons).
Of course, these conclusions should be examined with additional measures

of mataa including larger collections covering a greater portion of the island.
The relatively limited regional distribution of the assemblages (the south and
southwestern portions of the island) and restricted sample size (a total of 447
mataa) make it difficult to draw particular conclusions about the spatial extent
of individual groups. Consequently our findings remain tentative. Nonetheless,
our preliminary results are consistent with a model in which the scale of social
organization was localized and relatively small. This observation has implica-
tions for the evolution of competition and cultural elaboration on this remote
and relatively resource-poor island.
Acknowledgments
We thank Francisco Torres Hochstetter, Director of the P. Sebastian Englert
Museum, Rapa Nui, Chile for his support and assistance in providing us access
to the collections of mataa housed at the museum. Support for this research
was provided to Brooke Hundtoft through a program supporting undergradu-
ate research that was initiated by Dr.Karen Gould, Provost of California State
University Long Beach.
References
Bollt, R., J. E. Clark, P. R. Fisher, H. K. Flenley, J., and P. Bahn. 2003. The Enigmas
Yoshida. 2006. An Experiment in of Easter Island: Island on the Edge. Oxford
the Replication and Classification of University Press. Oxford.
Easter Island mataa. Rapa Nui Journal Hunt,T. L., and C. P. Lipo. 2007. Chronology,
20:125133. Deforestation, and Collapse: Evidence
Church, F., and J. G. Ellis. 1996. A Use- vs. faith in Rapa Nui Prehistory. Rapa
wear Analysis of Obsidian Tools from Nui Journal 21(2):8597.
an Ana Kionga. Rapa Nui Journal Hunt, T. L., and C. P. Lipo. 2009. Revisiting
10:8188. Rapa Nui (Easter Island) Ecocide.
Church, F., and J. Rigney. 1994.A Microwear Pacific Science 63(4):601616.
Analysis of Tools from Site 10241, Easter Kirch, P. V. 1984. The Evolution of the Poly-
Island an Inland Processing Site. Rapa nesian Chiefdoms, Cambridge University
Nui Journal 8:101105. Press, Cambridge.
Diamond, J. 2005. Collapse: How Societies Kirch, P. V. 2000. On the Road of The
Choose to Fail or Succeed, Viking, Winds: An Archaeological History of the
NewYork. Pacific Islands before European Contact, Uni-
Eerkens, J. W., and C. P. Lipo. 2005. Cultural versity of California Press, Berkeley.
Transmission, Copying Errors, and the Lipo, C. P. 2001. Science, Style and the Study
Generation of Variation in Material of Community Structure: An Example from
Culture and the Archaeological Record. the Central Mississippi River Valley. British
Journal of Anthropological Archaeology Archaeological Reports, International
24(4):316334. Series, No. 918, Oxford.
238 Li po et a l .
Lipo, C. P., and T. L. Hunt. 2005. Mapping Neiman, F. 1995. Stylistic Variation in
Prehistoric Statue Roads on Easter Island, Evolutionary Perspective: Inferences from
Antiquity 79:158168. Decorative Diversity and Interassemblage
Lipo, C., M. Madsen, R. Dunnell, and Distance in Illinois Woodland Ceramic
T. Hunt. 1997. Population Structure, Assemblages. American Antiquity 60(1):736.
Cultural Transmission, and Frequency OBrien, Michael J., and R. L. Lyman. 2003.
Seriation, Journal of Anthropological Cladistics and Archaeology. University of
Archaeology 16:33. Utah Press, Salt LakeCity.
Mulloy, W. 1961. Ceremonial Center of Routledge, K. 1919. The Mystery of Easter
Vinapu. In Archaeology of Easter Island, Island, Sifton, Praed & Co., London.
edited by T. Heyerdahl and E. N. J. von Saher, H. 1990. Some Details from
Ferdon, pp. 93161. The School of the Journal of Captain Bouman on the
American Research and the Museum of Discovery of Easter Island. Rapa Nui
New Mexico, SantaFe. Journal 6:3439.
Thirteen
Cultural Transmission and the

Production of Material Goods:
Evolutionary Pattern through
Measuring Morphology
Nathan Goodale, William Andrefsky, Jr., Curtis Osterhoudt,

Lara Cueni, and IanKuijt
Studies of lithic technological systems are traditionally focused on reduction

sequences (Bleed 2001; Goodale etal. 2008), tool use-life (Andrefsky 2008 and
references therein; Eren etal. 2005; Kuhn 1990), settlement patterns and raw
material acquisition (Beck et al. 2002; Brantingham 2003), and typological
construction (Andrefsky 1986; Goodyear 1974) (for a detailed overview of the
contributions of all of these facets to lithic technology see Andrefsky 2007).
Although we have many methodological means in which to assess patterns
associated with how humans utilized stone to make a living in the past, we
have minimally incorporated theory to interpret those patterns. Although this
chapter will not solve this problem, we take an important step in that direc-
tion, by examining one commonly overlooked problem in stone tool studies
related to the measurements we take and how well they perform in describing
morphological variation. We argue here this is a vital consideration before we
can discuss actual evolutionary patterns based on direct measurementdata.
Projectile points play a fundamental role as time sensitive indicators as well
as culture change (Andrefsky 1986). For projectile points to serve this role
archaeologists have more or less relied upon the haft element (Andrefsky 1986;
Bacon 1977; Binford 1963; Corliss 1972) ultimately because unlike the blade
portion of the tool, the haft element does not (in most cases) change during its
use-life due to use, resharpening, or repair (Andrefsky 1983a; Goodyear 1974,
although see Flenniken and Raymond 1986 for arguments against this in one
specific case study). For other parts of the world, others such as Harper and
239
240 Gooda l e e t a l .
Andrefsky (2008) and Andrefsky (1986) have found that the haft portion of a
notched projectile point is the only portion that does not consistently change
through the use-life of thetool.
To this end, the haft element is essentially the only aspect of the projec-
tile point that contains the original blueprint of what the maker envisioned
the tool to look like and that blueprint was influenced in most cases by the
path of cultural transmission (formal versus informal learning, i.e., Boyd and
Richerson 1985).
Recently there have been several studies focused on examining variation
in stone tools to interpret evolutionary patterns (Shott 2008) focusing on the
invention and adoption of technology such as the bow and arrow (Bettinger
and Eerkens 1999; Lyman et al. 2008), settlement patterns and raw material
acquisition (Brantingham 2003), migration and colonization (Buchanan and
Collard 2008), style and functional variation (Eerkens and Bettinger 2008),
cultural transmission (Buchanan and Hamilton 2009) and predicting the ratio
of producers to consumers based on raw material quality and quantity, and
the amount of variation within a given lithic technological system (Goodale
etal. 2008). Many of these studies utilize various direct measurement data from
projectile points to suggest how cultural transmission may be directly measur-
able in the archaeological record (Buchanan and Collard 2008; Mesoudi and
OBrien 2008a, b). Our goal here is similar in nature. Specifically, we aim to
provide morphometric data where some variables have a long history of use
in projectile point characterization and others that have not been used in
the analysis of notched points. Specifically, we suggest that this method can
provide advances toward the goals of understanding evolutionary process, and
more specifically cultural transmission and the implications of understanding
human behavior through stone tools.
Morphometrics and Evolutionary Patters

Morphometric analysis is a tool developed in evolutionary biology used to
extract information in order to understand organisms and evolutionary process
(Roth and Mercer 2000). Morphometric analysis combines some quantitative
means to express morphology and usually incorporates statistical analysis to
examine variability within some population (Monteiro etal. 2002; Roth and
Mercer 2000). To examine variability, the analyst chooses a series of landmarks
associated with discrete features of the specimens in question (Bacon 2000).
In this study, morphometric shape analysis is used to characterize the bases
of notched pointed tools. Twelve landmarks were established to measure the
shape of the bases of the stones tools. Importantly, these landmarks are all
located on the haft element, an area that should not change through the use-
life of the tool (Andrefsky 1986; Quinn et al. 2008). As an example of this,
Cultural Tr a n s m i s s i o n a n d t he P r o d u c ti o n o f M ate r i al G o o d s 241
(a) (b) (c) (d) (e) (f) (g) (h)
13.1. Dalton point reduction through use, resharpening, and repair. (a, b) Performs; (c, d) initial
use stage; (e, f) advance use stage; (g, h) final use stage. (Adapted from Goodyear 1974.)
Figure13.1 demonstrates the modification of Dalton points from the Brand

site. One can easily recognize performs, initial use stage, advanced use stage,
and the final stage as defined by Goodyear (1974). It is easily argued by this
example that measures such as weight, or maximum width and length would
not be appropriate for conducting evolutionary analyses with techniques
derived from cladistics. The reasons for this are due to the fact that these char-
acteristics are not directly associated with cultural transmission and how these
points were produced; rather, the morphological characteristics of everything
above the haft element are a representation of use, resharpening, and repair.
The Production of Notched Points in the Early Neolithic

We now turn to our case study of el-Khiam notched point production dur-
ing the early Neolithic in Southwest Asia (Figures13.2 and 13.3). The notched
point first appears sometime shortly after 12,000 calibrated years ago (cal BP)
in this the southern Levant (Gopher 1994; Nadel 1997). There are no other
notched points in the area before this time so it is an interesting case for exam-
ining the invention of a particular technology. This is one of the primary tool
types that is used to detect the start of the Pre-Pottery Neolithic A (PPNA)
time period (starting ca. 11,700 cal BP) suggesting that there was potentially
a rapid adoption of this point technology across the landscape (Kuijt and
Goodale 2009). It is unclear at this point what the function of the el-Khaim
point was and if it was used consistently for the same purposes across the early
Neolithic landscape. Based on extensive microwear analysis, Smith (2005) and
Goodale and Smith (2001) found that the el-Khiam point may have had a vari-
ety of functions and may not have been used for the same purposes at different
early Neolithic sites. To understand el-Khiam point life history, Quinn etal.
(2008) developed the el-Khiam Curation Index (EKCI). Based on macrowear
and breakage patterns they demonstrate that the likely function of el-Khiam
points at the site of Dhra, Jordan were used to puncture soft materials such as
hide.Through the EKCI Quinn etal. (2008) conclude that the only portion of
the el-Khiam point that changes during its use-life is the blade portion of the
13.2. Map of the southern Levant and early Neolithic sites.
tool. Based on the extensive examination of el-Khiam points, author N. G. has

noted that no points exhibit evidence of consistent resharpening of the haft
element and most exhibit only retouch to produce the notches and concave
bases.
The sample size used in this study consisted of 288 el-Khiam points from
the site of Dhra, Jordan. An additional 92 points were used from 8 other Pre-
Pottery Neolithic A communities in the southern Levant. High-resolution
13.3. An example of an el-Khiam notched point.The tool is made by unifacially retouching the
distal end into a point, producing a pair of notches, and producing a concave base. Quinn etal.
(2008) demonstrate that the blade is the portion of the tool that changes through the use-life
of thetool.
photographs were used in the analysis of the Dhra tools whereas the points
from the other eight sites were collected from the published literature. Points
from these eight sites consisted of high-resolution scans of artist illustrations.
We used all of the points we could find in the literature that had all of the land-
marks for the morphometric analysis as well as a scale bar in the illustration.
Notched Point MorphometricIndex

Twelve landmarks were chosen to measure the shape of the base of the el-
Khiam points. For the index we take a series of direct measurement data
not uncommon in archaeological attempts at classifying projectile points
(Andrefsky 1986; Binford 1963) including the neck width, the distance from
the bottom of the notch to the base of the tool and the width of the base.
We then draw the best fit ellipse into the notches as well as the concave por-
tion of the base. This provides values for the semi-major and semi-minor axes
of the best fit ellipse as well as a value for (the tilt of the semi-major axis
in relationship to the base of the point) (Figure13.4). The shape of the base
of the point is therefore described as the distance and orientation of three
13.4. Direct measurements taken for the NPMI include the neck width, proximity from the
bottom of the notch to the bottom of the base and the base width. Shape measures include
values for the semi-major and semi-minor axes of both notches and the base as well as or the
angle of tilt of the best fit ellipse in relationship to the base of the point.
ellipses in relation to each other. Although direct measurement data have

been utilized extensively to characterize the bases of notched points, to our
knowledge shape fitting has never been applied. As discussed later, we believe
this to be a very important variable when trying to characterize the shape of
the haft element.
All analysis was conducted in the computer program Image J, a freeware
program designed for image analysis. The program is designed to easily write
Java-based script to perform a variety of measurements. Our code first elimi-
nates all of the image apart from the outline of the artifact, finds all 12 land-
marks, and takes the basic measures of neck width, base of the notch to the
bottom of the base on each side. Finally, the notches and base are isolated
and the best fit ellipse is placed inside. Once the measurements are taken all
of the data are directly imported into a spreadsheet, eliminating any transcrip-
tion error. The program also proved to be efficient where all 380 points were
analyzed in approximately 12 hours. With conventional methods using cali-
pers and a notch fitting diagram that the authors originally attempted, it took
months to obtain all of thedata.
Statistical Analysis
First we examined the performance of the Notched Point Morphometric
Index (NPMI) and its ability to characterize the shape of the haft element
in an intuitive and logical manner. In other words, it paired similar-looking
bases with each other. To test this we performed hierarchical cluster analy-
sis. Cluster analysis was performed on x86 PCs, in Mathematica 7.0, with
the HierarchicalClustering package. The metric chosen (DistanceFunction) was
EuclideanDistance, though other metrics gave nearly identical results, and the
dissimilarity linkage chosen was Wards minimum variance. The dendrogram
in Figure 13.6 depicts the black nodes coded for the points from the early
Neolithic in Southwest Asia. The nodes coded in gray are from an evolution-
arily unrelated case discussed in the text that follows.
We calculated a corrected coefficient of variation (Haldane 1955) and used
t-tests to see how many statistically significant breaks there were within the
graph. This analysis revealed 162 statistically significant groups. We think that
this probably represents a high number of people making these tools as well
as probably informal teaching and learning. With a short survey of some of
the clusters around the dendrogram we can see that the NPMI characterizes
the hafts in a manner that appears to make visual sense (Figures13.5 and 13.6).
13.5. Image J software plug-ins for NPMI programming.

Cluster 124
Cluster 56
Cluster 162
Cluster 5
Dhra,Iraq ed Dubb, Wadi Faynan 16

Netiv Hagdud, Ain Darat, Gesher
Salibia IV, Hatoula, el-Khaim Terrace,
Mureybet
Ishi
24
Cluster
13.6. Hierarchical cluster analysis results.
The NPMI groups points that have long bases with slightly concave bases and
small notches (Cluster 5), short flat bases with deep notches (Cluster 56), long
necks with deep notches and concave bases (Cluster 124), and those with very
robust bases (Cluster 162) (Figure 13.7). Although this is a short survey, we
found that the NPMI groups similar-looking points with each other through-
out the dendrogram.
To verify the performance of the NPMI we included another evolution-
arily unrelated tradition of side-notched point manufacture.We conducted the
same analysis on the points the Native American Ishi produced during his time
at the University of California, San Francisco, the logic being if the NPMI
is performing well it should pull out an unrelated projectile point tradition.
The measured side-notched points are those depicted in Shackley (2000). As
Figure13.6 indicates by the gray-colored nodes, Ishis points were pulled out as
almost completely separate. There are a couple of early Neolithic points (all of
13.7. Several of the statistically significant clusters (based on the CCV and t-test analyses). Note
that the NPMI characterizes like with like, making intuitively logical patterns. From left: Cluster
5, Cluster 56, Cluster 124, Cluster162.
which end up being from Dhra with the highest sample size) that fall in line
with Ishis points but the overall trend is very apparent (Figure13.8).
Interestingly, if the data associated with the ellipse fitting are taken out, Ishis
points break up and occur all over the dendrogram, suggesting to us that this
shape-fitting data are probably very important for characterizing the morphol-
ogy of the base of the notched points. In other words, by taking the ellipse fit-
ting data out and running hierarchical cluster analysis, it is not apparent that Ishi
had a separate point making tradition from that of people 11,000years ago half a
world away. To us this speaks to the necessity that we evaluate the performance
of the measurements archaeologists take, and without this, trying to examine
evolutionary patterns through morphometrics would be extremely difficult.
Discussion: Evolutionary Patterns through

Morphometrics
In this study we have attempted to build an index allowing us to gather data
that successfully group morphologically similar side-notched point haft ele-
ments. We have stressed the importance of either measuring characteristics of
13.8. Projectile points made by Ishi. Note in the cluster diagram Ishis points are separated in
one section of the dendrogram (Cluster24).
248
stone tools that do not change during their use-lives or taking into account
effects of use, resharpening, and repair. There are several indices that have
been developed to take into account use-life history in projectile points (e.g.,
Andrefsky 2006; Clarkson 2002; Quinn etal. 2008), and their application has
not been considered thus far in trying to extract patterns related to cultural
transmission from those related to distance from source, intensity of use, or
relative retouch intensity. It is important to note that morphological charac-
teristics of stone tools reflecting the original idea that the manufacturer had
intended the end product to resemble is coded in aspects that do not relate to
resharpening, use, or repair. Thus, the aspects that do not change through tool
use-life will be reflective of evolutionary patterns derived by different modes
of cultural transmission (Boyd and Richerson 1985).
Ongoing experimental work by the authors is focused on examining mechani-
cal causes for morphological variation in the haft element.Though space restricts
a thorough treatment of the subject, our preliminary results suggest several fac-
tors that play a role. More obvious are issues of the original blank morphology
yet less obvious is that the knapping tools play a fundamental role in notch mor-
phology. Nicely sharpened pointed tines produce a notch morphology differ-
ently than a dull pointed tine. In addition, shovel-shaped tines produce different
notches than pointed tines.This may suggest that the knapping tools people used
in prehistory to make stone tools, such as projectile points, are part of the cultural
transmission package that ultimately produces morphological variation.
Lipo et al. (Chapter 12) also make similar arguments in evaluating their
measurements to aid in examining evolutionary patterns, and more explicitly,
provide seriation solutions for mataa tools on Rapa Nui. Lipo etal. also come
to a similar conclusion that measurements traditionally taken on mataa do not
successfully seriate the tools.
We find that although archaeologists can measure as many characteristics
that they can on stone tools, we have to be reminded of (1) different mor-
phological features are a result of different processes and (2) not all of the
measurements we take may be particularly useful in trying to interpret evolu-
tionary patterns. For example, Binford (1963) proposed an attribute list of all
the features an analyst can measure on a projectile point. There is, however, no
evaluation of any of the measurements, what processes their variation may be
a byproduct of, or why we should take the time to conduct the analysis of all
the features. Not all lithic studies are fraught with this issue where several indi-
ces directly try to test the curation of certain tool classes (Andrefsky 2008 and
references therein), yet we conclude in this study that not all measurements
may be useful in detecting evolutionary patterns. As we look into the future of
applying evolutionary theory into interpreting lithic technological systems we
must be aware of measurement performance and how data relate to interpret-
ing evolutionary patterns.
Acknowledgments
Thanks go to M. Steven Shackley, who provided comments on an earlier draft
of this chapter and original photographs of Ishis points for analysis. Thanks
to Mathew Eichenfield, who wrote the code for Image J used to analyze the
points in this chapter and to Alissa Nauman for reading and commenting on
a draft of this chapter. Although we had outside comments, any flaws are ours
alone.
References
Andrefsky, William, Jr. 1983. Experimental the Spread of Bow-and-Arrow Technology in

Archaeology and Lithic Assemblage Analysis. the Prehistoric Great Basin. American Antiquity
Proceedings of the 1983 Middle Atlantic Archaeological 64:231242.
Conference, edited by J. Evans, pp.4456. Binford, Lewis R. 1963. A Proposed Attribute
Rehoboth Beach, DE. 1986. Numerical Types List for the Description and Classification
and Inspectional Types: Evaluating Shape of Projectile Points. In Miscellaneous Studies
Characterization Procedures. North American in Typology and Classification, edited by A. M.
Archaeology 7:95111. White, L. R. Binford, and M. L. Papworth,
Rehoboth Beach, DE. 2006. Experimental pp. 193221. Anthropology Papers No.
and Archaeological Verification of an Index 19. University of Michigan Museum of
of Retouch for Hafted Bifaces. American Anthropology, Ann Arbor.
Antiquity 71:743758. Bleed, Peter. 2001. Trees or Chains, Links
Rehoboth Beach, DE. 2007. The Application or Branches: Conceptual Alternatives for
and Misapplication of Mass Analysis in Lithic Consideration of Stone Tool Production
Debitage Studies. Journal of Archaeological and Other Sequential Activities. Journal of
Science 34:392402. Archaeological Method and Theory 8:101127.
Rehoboth Beach, DE. 2008. Lithic Technology: Boyd, R., and P. J. Richerson. 1985. Culture and
Measures of Production, Use, and Curation, the Evolutionary Process. University of Chicago
edited by William Andrefsky, Jr. Cambridge Press, Chicago.
University Press, Cambridge. Brantingham, P. Jeffrey. 2003. A Neutral Model of
Bacon, Anne-Marie. 2000. Principal Compo- Stone Raw Material Procurement. American
nents Analysis of Distal Humeral Shape in Antiquity 68:487509.
Pliocene to Recent African Hominids: The Buchanan, Briggs and Mark Collard. 2008.Testing
Contribution of Geometric Morphomet- Models of Early Paleoindian Colonization
rics. American Journal of Physical Anthropology and Adaptation Using Cladistics. In Cultural
4:479487. Transmission and Archaeology: Some Fundamental
Bacon, Willard. 1977. Projectile Point Typology: Issues and Cast Studies, edited by Michael J.
The Basic Base. Archaeology of Eastern North OBrien, pp. 5976. Society for American
America 5:107122. Archaeology Press.
Beck, Charlotte, Amanda K. Taylor, George T. Buchanan, Briggs, and Marcus J. Hamilton.
Jones, Cynthia M. Fadem, and Caitlyn R. 2009. A Formal Test of the Origin of
Cook. 2002. Rocks Are Heavy: Transport Variation in North American Early
Costs and Paleoarchaic Quarry Behavior Paleoindian Projectile Points. American
in the Great Basin. Journal of Anthropological Antiquity 74:279298.
Archaeology 21:481507. Clarkson, Chris. 2002. An Index of Invasiveness
Bettinger Robert L., and Jelmer Eerkens. 1999. for the Measurement of Unifacial and Bifacial
Point Typologies, Cultural Transmission, and Retouch: A Theoretical, Experimental
and Archaeological Verification. Journal of School of Oriental Research. Eisenbrauns

Archaeological Science 29:6575. Winona Lake, Indiana.
Corliss, D.W. 1972. Neck Width of Projectile Points: Haldane, J. B. S. 1955. The Measurement of
An Index of Cultural Change and Continuity. Variation. Evolution 9:484.
Occasional Papers 29. Pocatello: Idaho State Lyman R. Lee, Todd L. VanPool, and Michael J.
University Museum. OBrien. 2008. Variation in North American
Eerkens, Jelmer, and Robert L. Bettinger. 2008. Dart Points and Arrow Points When One
Cultural Transmission and the Analysis of or Both Are Present. Journal of Archaeological
Stylistic and Functional Variation. In Cultural Science 35:28052812.
Transmission and Archaeology: Issues and Case Mesoudi, Alex, and Michael J. OBrien. 2008a.
Studies, edited by Michael J. OBrien, Pp. 21 The Cultural Transmission of Great Basin
38. The SAA Press, Washington,DC. Projectile-point Technology I: An experi-
Eren Metin I., Manual Dominguez-Rodrigo, mental Simulation. American Antiquity
Steven L. Kuhn, Daniel S. Adler, Ian Le, and 73:328.
Ofer Bar-Yosef. 2005. Defining and Measuring Mesoudi, Alex, and Michael J. OBrien. 2008b.
Reduction in Unifacial Stone Tools. Journal of The Cultural Transmission of Great Basin
Archaeological Science 32:11901206. Projectile Point Technology II: An Agent
Flenniken, J. Jeffery, and Annan W. Raymond. Based Computer Simulation. American
1986. Morphological Projectile Point Antiquity 73:627644.
Typology: Replication Experimentation and Monteiro, Leandro R., Jose Alexandre F. Diniz-
Technological Analysis. American Antiquity Filho, Sergio F. Reis, and Edilson D. Araujo.
51:603614. 2002. Geometric Estimates of Heritability in
Kuhn, Steven L. 1990. A Geometric Index of Biological Shape. Evolution 56:563572.
Reduction for Unifacial Stone Tools. Journal Nadel, Dani. 1997. The Chipped Stone Industry
of Archaeological Science 17:583593. of Netiv Hagdud. In An Early Neolithic Village
Kuijt, Ian, and Nathan Goodale. 2009. Daily in the Jordan Valley: Part I: The Archaeology of
Practice and the Organization of Space at the Netiv Hagdud, edited by O. Bar-Yosef and
Dawn of Agriculture: A Case Study from the A. Gopher, pp. 71150. American School of
Near East. American Antiquity 74:403422. Prehistoric Research Bulletin 43. Harvard
Goodale, Nathan B., Ian Kuijt, Shane Macfarlan, University, Boston.
Curtis Osterhoudt, and Bill Finlayson. Quinn, Colin Patrick, William Andrefsky,
2008. Lithic Core Reduction Techniques: Jr., Ian Kuijt, and Bill Finlayson. 2008.
Modeling Expected Diversity. In Lithic Perforation with Stone Tools and Retouch
Technology: Measure of Production, Use and Intensity: A Neolithic Case Study. In Lithic
Curation, edited by William Andrefsky, Jr., Technology: Measures of Production Use and
pp. 317336. Cambridge University Press, Curation, edited by William Andrefsky, Jr.,
Cambridge. pp. 150174. Cambridge University Press,
Goodale, N. B., and S. Smith. 2001. Pre-Pottery Cambridge.
Neolithic A Projectile Points at Dhra, Jordan: Roth, Louise V., and John M. Mercer. 2000.
Preliminary Thoughts on Form, Function Morphometrics in Development and
and Site Interpretation. Neo-Lithics: Southwest Evolution. American Zoologist 40:801810.
Asian Lithic Research 2(1):26. Shackley, M. Steven. 2000. The Stone Tool
Goodyear,Albert C. 1974. The Brand Site:A Techno Technology of Ishi and the Yana of
Functional Study of a Dalton Site in Northeast North Central California: Inferences for
Arkansas. Publications on Archaeology, Hunter-Gatherer Cultural Identity in
Research Series 7. Fayetteville: Arkansas Historic California. American Anthropologist
Archaeological Survey. 102:693712.
Gopher, Avi. 1994. Arrowheads of the Neolithic Shott, Michael. 2008. Darwinian Evolutionary
Levant. Dissertation Series 10 American Theory and Lithic Analysis. In Cultural
Transmission and Archaeology: Issues and Case from Wadi Faynan 16 and Dhra: Implications for
Studies, edited by M. J. OBrien, pp. 146157, the Description and Interpretation of Early Neolithic
SAA Press, Washington,DC. Chipped Stone Variability. Ph.D. dissertation in
Smith, Samuel. 2005. A Comparative Analysis of the School of Human and Environmental
the Form and Function of Chipped Stone Artefacts Science, University of Reading.
Fourteen
What Steward Got Right:

Technology, Work Organization,
and Cultural Evolution
Nathan E. Stevens
Tradition makes the man, by circumscribing his behaviour within certain

bounds; but it is equally true that man makes the traditions (Childe
1936:238).
Evolutionary and optimization approaches to archaeology, and to lithic
technology specifically, are a positive, if not novel, development (Beck etal.
2002; Bettinger and Eerkens 1999; Bettinger etal. 2006; Elston 1992; Hughes
1998; Isaac 1972; Jeske 1992; Kuhn 1994; OBrien etal. 2001; Pitt-Rivers 1906
[1875]; Torrence 1983, 1989; Ugan etal. 2003; Wright 1994). The application of
evolutionary concepts has varied widely in these studies, alternately including
ideas traceable to Darwin, Marx, or Spencer, as well as other biological and
social theorists (Bettinger 1991; Dunnell 1980). Currently, most archaeologi-
cal applications of evolutionary theory align themselves with either human
behavioral ecology (HBE; Bird and OConnell 2006; Smith and Winterhalder
1992a; Winterhalder and Smith 2000) or dual inheritance theory (DIT; Boyd
and Richerson 1985; Cavalli-Sforza and Feldman 1981; Henrich and McElreath
2008), which are rooted variously in economics, evolutionary ecology, and
population genetics. Somehow, the contributions of Julian Steward (1936,
1938, 1955) to the understanding of both evolution and ecology as applied to
humans have been almost forgotten in the enthusiasm for these productive
approaches. I will briefly discuss why Stewards concept of cultural ecology
is still relevant and how it may be incorporated into current evolutionary
approaches in anthropology.
253
254 Nathan E . S t e ve n s
Cultural Ecology and Current Evolutionary Approaches

Stewards development of cultural ecology was a reaction to what he saw as
competing, but equally flawed, visions of how cultures change stemming from
environmental determinism and historical particularism (Steward 1955:35).
Rather than attributing recurrent cultural patterns solely to environmental set-
ting or to culture history, Stewards solution was to treat the interaction of the
two as a third, uniquely creative force (Steward 1955:34). This force was medi-
ated by what Steward called the culture core, which continues to be one of
the most appealing and confounding aspects of cultural ecology as developed
by Steward. While Steward never rigorously defined the culture core, he did
suggest it would consist of ...the constellation of features which are most
closely related to subsistence activities and economic arrangements (Steward
1955:37). And that, Cultural ecology pays primary attention to those features
which empirical analysis shows to be most closely involved in the utilization
of environment in culturally prescribed ways (Steward 1955:37).
In Stewards (1955) formulation, the culture core includes aspects of a cul-
ture that influence how people interact with the environment, but importantly,
it is not seen as determined by the environment. Instead it is inherited histori-
cally along with many other aspects of culture that characterize a region. In
this way, Steward wisely separated the vagaries of the environment from the
inner workings of a culture that occupied a given environment. Viewed over
the long term, this means that environment and culture are on more or less
separate evolutionary tracks and that the ability of one to influence the other
is dependent on how each is structured.
Two problems with cultural ecology that contributed to the development of
current evolutionary approaches are (1) the ambiguous concept of the culture
core (Harris 1968; Kelly 1995; Moran 1984) and (2) explicit or implicit appeal
to adaptation and selection at the group level to explain patterns (Bettinger
1991:50; Smith 1984). Within HBE, the concept of culture core has been
replaced by the belief that behaviors most affecting fitness (e.g., subsistence
behavior) are of primary importance (Smith 2000), and the modeling of how
individual decisions influence group-level outcomes has replaced group selec-
tion (Smith and Winterhalder 1992b), at least of the Wynne-Edwards (1962)
variety. Researchers employing DIT emphasize how cultural evolution can
produce both adaptive and maladaptive behaviors but also maintain that the
propensity for culture itself is the product of natural selection (Henrich and
McElreath 2003). Also, DIT has found ways to model how cultural group
selection could work (e.g., Soltis etal. 1995).
Despite these changes, the materialist underpinnings and general subject
matter of both HBE and DIT, especially as applied to archaeological problems,
are not far off the mark from what Steward advocated. Unfortunately, although
Wh at Stewar d G o t Ri g ht 255
many of Stewards insights into the nature of humanenvironment interactions

have been retained in modern evolutionary approaches, his emphasis on tech-
nology and on the social organization of work has faded. In fact, at their
extremes, HBE and DIT often resemble environmental determinism and his-
torical particularism, respectively, with HBE frequently relying on external
factors such as climate change to explain cultural change, and DIT empha-
sizing the potential for disconnects between culture and fitness. Are we once
again in need of a theoretical middle ground and are the ideas of Steward a
good place to go looking forit?
Individual Behavior, Work Organization, and

Technological Tradition
Although it could be argued that evolutionary theory as applied to lithic tech-
nology is just another case of archaeologists discovering a fruitful research
avenue late in the game, such studies hold great potential for providing the
wider anthropological community with empirical data on long-term culture
environment interactions. Anthropologists have long noted that certain items
of material culture are highly conserved across space and time while others
change frequently. Materialists among us might explain this as a function of
the articulation of base and superstructure (Harris 1968), core and noncore
(Steward 1955), technomic and ideo-technic (Binford 1962), or fitness-related
(functional) and neutral (stylistic) traits (Dunnell 1978). In each case, the result
is the same: material culture related to subsistence activities is often redundant
over space and time while material culture related to other softer aspects of
life varies more widely.
More recently, this dichotomy has been explored by anthropologists and
archaeologists applying phylogenetic approaches to cultural change (see
Borgerhoff Mulder 2001; Cavalli-Sforza etal. 1992; Guglielmino etal. 1995;
Lipo etal. 2006; Mace and Holden 2005; OBrien etal. 2001). Studies inves-
tigating the degree of vertical versus horizontal transmission of ethnographic
traits (e.g., Jordan and Shennan 2003, 2009; Moylan etal. 2006; Tehrani and
Collard 2009) have attempted to identify core traditions (see Boyd et al.
1997) made up of traits that tend to agglomerate and are less affected by out-
side influences (e.g., descent and marriage).
Perhaps the core traditions of the phylogeneticists and Stewards culture core
bear more than superficial similarity. Rather than thinking of the culture core
as static, I suggest a more appropriate definition is that the culture core, like all
culture, changes, but that it changes more slowly than other, more peripheral,
aspects of culture (presumably because it is tied to fitness). Lithic technol-
ogy changes slowly, is closely connected with work organization, and seems
to be largely (but by no means exclusively) vertically inherited. Therefore,
more than other types of archaeological data, studying lithic technology, and
its organization, may provide important insights into how human adaptations
are structured. Among lithic technologies, those that are most closely related to
subsistence practices, and that characterize large regions of otherwise culturally
distinctive populations seem good candidates for core traditions that may resist
rapid change due to connections with fitness and work organization.
The relationship of technology to work organization brings up a poten-
tially important point to those of us who are interested in formal models of
human behavior. Current evolutionary approaches are concerned with model-
ing individual decisions, but the types of tools used by a culture may affect the
fit of such models because the behavioral options of individuals were likely
limited to varying degrees by different types of technologies.1 As technologi-
cal traditions evolved, the tools available to an individual at any point in time
would constrain his or her behavior into culturally agreed-upon tasktool
combinations.
Steward asked why human adaptations looked as they did in a given envi-
ronment with a given technology. More specifically, he was also concerned
with how much behavioral latitude was possible given a particular environment
and technology (Steward 1955:36). Considering these questions with some
time depth, one might also ask, In a changing environment, how much does
work organization have to change to produce an archaeologically detectable
change in technology? Given the proposed relationships between technol-
ogy and work organization, the answer is that it probably depends on the
technology.
If potential behavioral strategies are conditioned by both the available tools
and the social and environmental context of the work (see Schiffer and Skibo
1997), the behavioral repertoire of any individual should be related to (1) the
tools in the technological tradition and (2) the culturally learned ways to use
those tools to exploit the local environment. Likewise, the tools in the tech-
nological tradition must also be influenced by the tool-use decisions of indi-
viduals over time (see Figure14.1). In this way, lithic technology can be seen
as reflecting the intersection between individual behavior, work organization,
and cultural tradition.
The glib contention that people will make new tools if necessary or that
new technologies are always available through borrowing or invention does
not square with the reality that over the short term, procurement and manu-
facturing of tools are often embedded within, and dependent on, the coordi-
nated activities of others (Binford 1979). Over the long term, making changes
to existing technologies, and developing or adopting new technologies is as
much a social problem as it is an engineering problem (Bettinger 1999; Fitzhugh
2001; Richerson and Boyd 2001; Rosenberg 1994). In Stewards language, the
configuration of the culture core might prevent new technologies from being
Individual
Behavior
Work
Context Available Organization
Tools
Environmental Division of labor
Social Mobility
Technological
Tradition
14.1. Proposed relationships among behavior, technology, and tradition. Individual behavior
is both constrained by and influenced by context, available tools, and work organization. The
technological tradition results from these interactions, but also influencesthem.
adopted or developed. In other words, interdependencies between techno-

logical tradition, work organization, and individual behavior may restrict both
short-term and long-term behavioral options. Even minor changes to lithic
technologies may therefore reflect important behavioral changes.
Lithic Technological Change in California

Evidence for incremental changes in lithic technology is not difficult to come
by. The difficulty is in assigning causality to any particular variable related to
technological change. Equally valid cases can be made for the role of increasing
human population pressure, decreasing mobility, decreasing foraging efficiency,
environmental change, and other factors. A thorough consideration of these
factors is beyond the scope of this chapter; however, it is useful to point out
some salient patterns and their implications for those who apply evolutionary
theory to the archaeological record.
First, a general pattern of fewer multifunctional tools and more specialized
tools through time is evident among flaked stone assemblages from Californias
central coast (Figure14.2). Evidence from use-wear analysis shows this pattern
as an increasingly steep reduction in the proportion of multifunctional tools
through time, in particular after ca. 3000 BP (Figure14.3).
Individual artifact classes from the region also support the idea that tools
became more specialized through time. For example, during the Early (ca.
50002500 BP) and Middle (ca. 25001000 BP) periods, points of the cen-
tral coast stemmed series (sensu Jones 1993) are the dominant projectile point
form while after 1000 BP, smaller leaf-shaped points dominate the projectile
class. Morphometric comparisons as well as use-wear data suggest the earlier
stemmed points were multifunctional, exhibiting projectile use in addition
to use as knives, while later small leaf-shaped points exhibit only projectile
CA-SBA-755
CA-SBA-990
P a c
CA-SBA-935
CA-SBA-2696
i f i c
CA-SBA-1010
Vandenberg
Air Force Base
O c e
a n
CA-SBA-246
CA-SBA-677
CA-SBA-530
CA-SBA-1823
CA-SBA-212
CA-SBA-503
0 2 4 8
N Miles
0 3 6 12
Kilometers
14.2. Locations of California Central Coast archaeological sites with use-wear data used to cre-
ate Figure14.3.
0.5
0.4
Proportion Multifunctional Tools
0.3
0.2
0.1
0
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10,000
Years BP
14.3. Proportions of multifunctional tools in California Central Coast assemblages2 plotted
against calibrated years before present. Log regression r2 = 0.61; p < 0.001.
Acorns Small Seeds
Late Holocene
Early Holocene
Time
Acorns and Small Seeds
14.4. Changes in California Central Coast ground stone technology throughout the
Holocene.
use (Stevens and Codding 2009). Presumably, after the adoption of small leaf-
shaped points, other tools such as bifacial knives or utilized flakes stepped in to
fulfill cutting tasks for which stemmed points were formerly employed. Again,
the overall pattern is one of decreasing multifunctionality and increasing num-
bers of specialized tools.
If the functional latitude of tools narrowed through time, we might expect
design changes reflecting this. Ground stone tools illustrate this progression par-
ticularly well because the form of ground stone tools is more plastic and much
less constrained by physics when compared to flaked stone tools. Ground stone
tools from the early Holocene consist of millingslabs and handstones that are
thought to have been used to process a wide variety of vegetal foods including
small seeds, fibrous roots, and acorns. Between about 5000 and 3000 BP, however,
the importance of the acorn as a staple foodstuff increased greatly and stone mor-
tars and pestles first appear in the archaeological record in large numbers (Basgall
1987; Glassow 1996b). Interestingly, the older millingslab/handstone technology
was not abandoned, but appears to have been retained for use with small seeds
and other resources that mortars cannot process as efficiently (Figure14.4).
Changes in form within each ground stone tool class also reflect this shift.
Millingslabs progress from shaped and intensively used basin forms to flat,
less intensively used forms sometime after the beginning of the Early period.
Similarly, handstones from Millingstone period contexts (ca. 10,0005000 BP)
are mostly shaped and heavily used while handstones found in later contexts
are frequently unshaped and only ephemerally used. The opposite pattern
is found among mortars and pestles, with earlier mortars consisting largely
of globular unshaped or minimally shaped forms and later forms exhibiting

greater degrees of shaping and thinner walls. Pestles first appear as slightly
shaped squat, cylindrical forms and then later evolve into highly shaped tapered
forms (Glassow 1996b; Stevens etal. 2004).
It is probably not coincidental that the first millingslabs and first mortars
are similar to each other in form while later versions of each become increas-
ingly differentiated, reflecting greater functional specialization (see Eerkens
and Lipo 2007:262 for discussion of artifact divergence). A similar argument
has been advanced to account for changes in ground stone tools in the pre-
historic Southwest (Adams 1993; Glassow 1996b; Martin and Plog 1973:216;
Mauldin 1993; Plog 1974:139) although, interestingly, it is the forms of mill-
ingslabs rather than mortars that become most specialized, probably reflecting
the increasing importance of maize agriculture and seed grinding, rather than
balanophagy and acorn pounding (cf. Adams 1999).
Discussion
Human populations along the central California coast underwent several
important changes during the Late Holocene including decreases in mobil-
ity (Bamforth 1986; Lebow etal. 2007), increasing population (Glassow 1999;
Lebow et al. 2007), and increasing diet breadth (Glassow 1996a; Jones et al.
2008). Concurrent environmental changes include the stabilization of sea level
and a cooler and more variable climate (Atwater etal. 1977; Barron etal. 2003;
Kennett and Kennett 2000). Human behavioral ecologists might interpret the
above technological patterns as related to declining foraging efficiency, per-
haps brought on by environmental change and/or increased population density,
spurring increased investment in more specialized tools (compare to Broughton
1994, 1997; Glassow 1996a; Kennett 2005). A relevant model incorporating DIT
might view environmental change as the catalyst for human population growth,
but consider historically inherited technology and work organization a limiting
factor. Local populations undergoing social or technological intensification may
subsequently alter the carrying capacity of the local environment and experi-
ence further population growth, enabling them to outcompete other groups
in the area (compare to Bettinger and Baumhoff 1982; Richerson etal. 2001).
This second model is closer to Stewards concept of cultural ecology because
the technological tradition is seen as mediating the creative interplay between
environmental changes on one hand and human behavioral responses on the
other. Each of these perspectives, however, contributes an important piece of
our understanding of long-term cultural change. HBE explains how subsistence
changes should look given changes in other variables while DIT provides a
plausible evolutionary mechanism for culture change given certain rules about
how cultural information is inherited. HBE highlights the economic factors
conditioning technological change while DIT helps explain why technological

changes might spread even if specific groups are resistant.
Whichever general explanation is used, the specifics of the observed archae-
ological patterns at the assemblage level can be seen as multifunctional tools
being partially or completely replaced with specialized tools, thus increasing
the number of tool types while at the same time narrowing the functional
latitude of each tool type. The result of this progression is that the relationship
between a given tool type and its intended use becomes stronger over time.
This makes sense in the context of more rigid Holocene land-use systems,
where increasingly, people found themselves compelled to be in a particular
location at a particular time where a specific activity/tool combination may
have precluded other behavioral options. The effect of this may have been less
room for individual learning in tool making and using decisions and overall
less behavioral latitude at the level of the group. This suggests that any attempt
to model individual decision making should consider the context of the task,
the available technology, and work organization.
In addition, if the fit between tool morphology and tool use is variable over
time, then ideas about performance characteristics and use lives of specific arti-
fact classes are likely not as straightforward as have been previously portrayed
(Bettinger etal. 2006; Fitzhugh 2001; OBrien etal. 1994; Schiffer and Skibo
1997; Shott and Sillitoe 2004; Ugan etal. 2003). Importantly, this could also
affect the archaeological visibility of large-scale behavioral changes if a given
group of technologies permitted a wide latitude of behavioral change with
minimal or no tool morphological change or replacement. This sort of behav-
iorally variable, but archaeologically subtle adaptation might help explain the
apparent lack of culture change throughout the California Early (ca. 11,500
7000 BP) and Middle (ca. 70004000 BP) Holocene, a long interval character-
ized by considerable climatic variability.
If the relationships among lithic technology, human behavior, environment,
and the culture core (if such an entity exists) are similar to my broad sketch,
then we should not be surprised that the archaeological record is full of exam-
ples of similar environmental or climatic events resulting in dissimilar reactions
by human populations.This is because the important relationship is not simply
between human behavior and the environment, but how core adaptive tradi-
tions influence this relationship. This was Stewards key insight and it is one
that studies of lithic technology have the potential to explore further.
Acknowledgments
The ideas in this chapter were influenced by conversations with Bruce
Winterhalder, Robert Bettinger, Jelmer Eerkens, Richard McElreath, Brian
Codding, and Raven Garvey. Data from Vandenberg AFB were generously
provided by Douglas Harro and Chris Ryan. Jelmer Eerkens and Brian
Codding read a previous version of the manuscript and greatly improved its
quality.
Notes
1 The words technology and tools are AU5 (Lebow et al. 2006), SBA-935 (Harro
admittedly imprecise, but are used here in the et al. 2000), SBA-990 (Lebow et al. 2005b),
sense that implements suitable for perform- SBA-1010 AUs 16 (Lebow et al. 2005c),
ing different tasks are variable in time and SBA-1823 (Harro etal. 2001), SBA-2696 AU3
space, as are the knowledge and materials to (Colten et al. 1999). Multifunctional tools
make them. Fitzhughs (2001:128) definition were defined as those that showed evidence
of technology as including material, practi- of either concurrent use for multiple actions
cal, informational, and purposive components is (e.g., scraping, cutting) and/or multiple con-
usefulhere. tact materials (e.g., bone, wood), or sequent
2 Sites/components plotted in Figure 14.3 use (e.g., a core recycled as a plane). Ages of
include CA-SBA-212 AUs 1 and 2 (McKim all components plotted are based on associated
et al. 2007), SBA-246 Locus B (Lebow et al. radiocarbon dates except for SBA-1823, which
2001), SBA-503 AUs 2 and 3 (Lebow et al. was plotted at 8000 BP based on Millingstone
2005a), SBA-530 AUs 3, 4, and 6 (Lebow etal. period assemblage. If this component is omit-
2007), SBA-677 (Lebow etal. 1998), SBA-755 ted, r2 = 0.57; p < 0.001.
References
Adams, J. L. 1993.Technological Development of Are Heavy: Transport Costs and Paleoarchaic

Manos and Metates on the Hopi Mesas. Kiva Quarry Behavior in the Great Basin. Journal of
58:331344. Anthropological Archaeology 21(4):481507.
Adams, J. L. 1999. Refocusing the Role of Food- Bettinger, R. L. 1991. Hunter-Gatherers:
grinding Tools as Correlates for Subsistence Archaeological and Evolutionary Theory. Plenum
Strategies in the U. S. Southwest. American Press, NewYork.
Antiquity 64(3):475498. Bettinger, R. L. 1999. From Traveler to Processor:
Atwater, B., C. Hedel, and E. Helley. 1977. Late Regional Trajectories of Hunter-Gatherer
Quaternary Depositional History, Holocene Sea Sedentism in the Inyo-Mono Region,California.
Level Changes, and Vertical Crustal Movement, In Settlement Pattern Studies in the Americas: Fifty
Southern San Francisco Bay, California. US Years Since Viru, edited by B. R. Billman and G.
Geological Survey Professional Paper, No. M. Feinman, pp. 3955. Smithsonian Institution
1014. US Government Printing Office, Press, Washington,DC.
Washington,DC. Bettinger, R. L., and M. A. Baumhoff. 1982.
Bamforth, D. B. 1986. Technological Efficiency The Numic Spread: Great Basin Cultures in
and Tool Curation. American Antiquity Competition. American Antiquity 47:485503.
51(1):3850. Bettinger, R. L., and J. W. Eerkens. 1999. Point
Barron, J. A., L. Heusser, T. Herbert, and M. Lyle. Typologies, Cultural Transmission, and the
2003. High-resolution Climatic Evolution of Spread of Bow-and-Arrow Technology in the
Coastal Northern California during the Past Prehistoric Great Basin. American Antiquity
16,000 Years. Paleoceanography 18(1):114. 64(2):231242.
Basgall, M. E. 1987. Resource Intensification Bettinger, R. L., B. Winterhalder, and R.
among Hunter-Gatherers: Acorn Economies McElreath. 2006. A Simple Model of
in Prehistoric California. Research in Economic Technological Intensification. Journal of
Anthropology 9:2152. Archaeological Science 33:538545.
Beck, C., A. K. Taylor, G. T. Jones, C. M. Fadem, Binford, L. R. 1962.Archaeology as Anthropology.
C. R. Cook, and S. A. Millward. 2002. Rocks American Antiquity 28.(2):217225.
Binford, L. R. 1979. Organization and Dunnell, R. C. 1980. Evolutionary Theory and

Formation Processes: Looking at Curated Archaeology. In Advances in Archaeological
Technologies. Journal of Anthropological Research Method and Theory, Vol. 3, edited by M.
35(3):255273. B. Schiffer, pp. 3599. Academic Press,
Bird, D. W., and J. F. OConnell. 2006. Behav- NewYork.
ioral Ecology and Archaeology. Journal of Eerkens, J. W., and C. P. Lipo. 2007. Cultural
Archaeological Research 14:143188. Transmission Theory and the Archaeological
Borgerhoff Mulder, M. 2001. Using Record: Providing Context to Understanding
Phylogenetically Based Comparative Variation and Temporal Changes in Material
Methods in Anthropology: More Questions Culture. Journal of Archaeological Research
than Answers. Evolutionary Anthropology 15:239274.
10:99111. Elston, R. G. 1992. Economics and Strategies
Boyd, R., M. Borgerhoff Mulder, W. H. of Lithic Production at Tosawihi. In
Durham, and P. J. Richerson. 1997. Are Archaeological Investigations at Tosawihi, a Great
Cultural Phylogenies Possible? In Human by Basin Quarry, edited by R. G. Elston and C.
Nature, edited by P. Weingart, S. D. Mitchell, Raven, pp. 775801. Prepared for the Bureau
P. J. Richerson, and S. Maasen, pp. 355386. of Land Management, Elko Resource Area,
Lawrence Erlbaum Associates, Mahwah,NJ. by Intermountain Research Reports, Silver
Boyd, R., and P. J. Richerson. 1985. Culture and City,NV.
the Evolutionary Process. University of Chicago Fitzhugh, B. 2001. Risk and Invention in
Press, Chicago. Human Technological Evolution. Journal of
Broughton, J. M. 1994. Late Holocene Resource Anthropological Archaeology 20(2):125167.
Intensification in the Sacramento Valley, Glassow, M. A. 1996a. Purisimeno Chumash
California: The Vertebrate Evidence. Journal of Prehistory: Maritime Adaptations along the
Archaeological Science 21:501514. Southern California Coast. Harcourt Brace
Broughton, J. M. 1997. Widening Diet Breadth, College, Fort Worth.
Declining Foraging Efficiency, and Prehistoric Glassow, M. A. 1996b. The Significance to
Harvest Pressure: Icthyofaunal Evidence California Prehistory of the Earliest Mortars
from the Emeryville Shellmound, California. and Pestles. Pacific Coast Archaeological Society
Antiquity 71:845862. Quarterly 32(4):1426.
Cavalli-Sforza, L. L., and M. Feldman. 1981. Glassow, M. A. 1999. Measurement of Population
Cultural Transmission and Evolution. Princeton Growth and Decline during California
University Press, Princeton,NJ. Prehistory. Journal of California and Great Basin
Cavalli-Sforza, L. L., E. Minch, and J. L. Mountain. Anthropology 21(1):4566.
1992. Coevolution of Genes and Languages Guglielmino, C. R., C.Viganotti, B. Hewlett, and
Revisited. Proceedings of the National Academy L. L. Cavalli-Sforza. 1995. Cultural Variation in
of Science of the U S A 89:56205624. Africa: Role of Mechanisms of Transmission
Childe,V. G. 1936. Man Makes Himself. Watts and and Adaptation. Proceedings of the National
Company, London. Academy of Sciencesof the U S A 92:75857589.
Colten, R. H., C. G. Lebow, C. Denardo, R. L. Harris, M. 1968. The Rise of Anthropological Theory.
McKim, D. R. Harro, C. H. Miksicek, and B. Crowell, NewYork.
Bowser. 1999. Hunter-Gatherer Land Use in Harro, D. R., C. G. Lebow, and R. L. McKim. 2001.
the San Antonio Creek Drainage: Archaeological Archaeological Investigations at CA-SBA-1823,
Investigations at CA-SBA-2696. California Vandenberg Air Force Base, Santa Barbara County,
State Water Project, Coastal Branch Series, California. Applied EarthWorks, Inc., Fresno,
Paper No. 11. San Luis Obispo County CA, for Tetra Tech, Inc., Santa Barbara, CA.
Archaeological Society, San Luis Obispo. Submitted to 30 CES/CEV, Vandenberg Air
Dunnell, R. C. 1978. Style and Function: A Force Base, Santa Barbara,CA.
Fundamental Dichotomy. American Antiquity Harro, D. R., C. G. Lebow, R. L. McKim, C.
43(2):192202. Ryan, and C. Denardo. 2000. Eligibility Testing
at CA-SBA-935, -2321, and -2345 for El Nino Kennett, D. J., and J. P. Kennett. 2000. Compet-
Related Road Repairs Vandenberg Air Force Base, itive and Cooperative Responses to Climatic
California. Applied EarthWorks, Inc., Fresno, Instability in Coastal Southern California.
CA, Submitted to 30 CES/CEV, Vandenberg American Antiquity 65(2):379395.
Air Force Base, Santa Barbara,CA. Kuhn, S. L. 1994. A Formal Approach to the
Henrich, J., and R. McElreath. 2003. The Design and Assembly of Mobile Toolkits.
Evolution of Cultural Evolution. Evolutionary American Antiquity 59(3):426442.
Anthropology 12:123135. Lebow, C. G., D. R. Harro, R. L. McKim, and C.
Henrich, J., and R. McElreath. 2008. Dual Denardo. 2001. Archaeological Excavations at CA
Inheritance Theory:The Evolution of Human SBA 246, An Early Holocene Site on Vandenberg
Cultural Capacities and Cultural Evolution. Air Force Base, Santa Barbara County, California.
In Oxford Handbook of Evolutionary Psychology, Applied EarthWorks, Inc., Fresno,CA, for Tetra
edited by R. Dunbar and L. Barrett, pp. 555 Tech, Inc., Santa Barbara,CA. Submitted to 30
570. Oxford University Press, Oxford. CES/CEV,Vandenberg Air Force Base,CA.
Hughes, S. S. 1998. Getting to the Point: Lebow, C. G., D. R. Harro, R. L. McKim, C.
Evolutionary Change in Prehistoric Weaponry. Denardo, J. L. Gerber, and C. Ryan. 1998.
Journal of Archaeological Method and Theory Archaeological Investigations at CA SBA 671,
5(4):345408. -677, and -2961, Vandenberg Air Force Base,
Isaac, G. L. 1972. Chronology and Tempo of Santa Barbara County, California. Applied
Cultural Change during the Pleistocene. In EarthWorks, Inc., Fresno, CA, for Tetra Tech,
Calibration of Hominoid Evolution, edited by Inc., Santa Barbara, CA. Submitted to 30
W. W. Bishop and J. A. Miller, pp. 381430. CES/CEVPC, Vandenberg Air Force Base,
Scottish Academic Press, Edinburgh. Santa Barbara,CA.
Jeske, R. J. 1992. Energy Efficiency and Lithic Lebow, C. G., D. R. Harro, R. L. McKim, A. M.
Technology:An Upper Mississippian Example. Munns, C. Denardo, J. Onken, and R. Bury.
American Antiquity 57(3):467481. 2005a. The Archaeology and Rock Art of Swordfish
Jones, T. L. 1993. Big Sur: A Keystone in Central Cave (CA-SBA-503), Vandenberg Air Force Base,
California Cultural History. Pacific Coast Santa Barbara County, California. Applied
Archaeological Society Quarterly 29(1):178. EarthWorks, Inc., Fresno, CA. Submitted
Jones,T. L., J. F. Porcasi, J. Gaeta, and B. F. Codding. to 30 CES/CEVNC, Vandenberg Air Force
2008. The Diablo Canyon Fauna: A Coarse- Base,CA.
grained Record of Trans-Holocene Foraging Lebow, C. G., L. Haslouer, J. M. Fancher, N. E.
from the Central California Mainland Coast. Stevens, and A. M. Munns. 2005b. Archaeological
American Antiquity 73:289316. Investigations Supporting Consultation with the
Jordan, P., and S. Shennan. 2003. Cultural Trans- State Historic Preservation Officer for the Heritage
mission, Language, and Basketry Tradi- Launch Program Demolition on Vandenberg Air
tions amongst the California Indians. Journal Force Base in Santa Barbara County, California.
ofAnthropological Archaeology 22(1):4274. Applied EarthWorks, Inc., Submitted to: 30
Jordan, P., and S. Shennan. 2009. Diversity in CES/CEVNC,Vandenberg Air ForceBase.
Hunter-Gatherer Technological Traditions: Lebow, C. G., R. L. McKim, Douglas R. Harro,
MappingTrajectories of Cultural Descent with C. M. Hodges, and A. M. Munns. 2005c. Large
Modification in Northeast California. Journal Game Hunting and Other Paludal Adaptations
of Anthropological Archaeology 28(3):342365. at Barka Slough: Excavations at CA-SBA-1010,
Kelly, R. L. 1995. The Foraging Spectrum: Diversity Vandenberg Air Force Base, Santa Barbara
in Hunter-Gatherer Lifeways. Smithsonian County, California. Applied EarthWorks, Inc.,
Institution Press, Washington,DC. Fresno, CA. Submitted to 30 CES/CEVNC,
Kennett, D. J. 2005. The Island Chumash: Behav- Vandenberg Air ForceBase.
ioral Ecology of a Maritime Society. University of Lebow, C. G., R. L. McKim, D. R. Harro, and
California Press, Berkeley. A. M. Munns. 2006. Prehistoric Land Use in
the Casmalia Hills throughout the Holocene: OBrien, M. J., J. Darwent, and R. L. Lyman.
Archaeological Investigations along Combar 2001. Cladistics Is Useful for Reconstructing
Road, Vandenberg Air Force Base, Santa Barbara Archaeological Phylogenies: Palaeoindian
County, California. Applied EarthWorks, Inc., Points from the Southeastern United States.
Fresno, CA. Submitted to 30 CES/CEVNC, Journal of Archaeological Science 28:11151136.
Vandenberg Air ForceBase. OBrien, M. J., T. D. Holland, R. J. Hoard, and
Lebow, C. G., R. L. McKim, D. R. Harro, G. L. Fox. 1994. Evolutionary Implications
A. M. Munns, and C. Denardo. 2007. of Design and Performance Characteristics
Littoral Adaptations Throughout the Holocene: of Prehistoric Pottery. Journal of Archaeological
Archaeological Investigations at the Honda Beach Method and Theory 1(3):259304.
Site (CA-SBA-530), Vandenberg Air Force Base, Pitt-Rivers, A. L. 1906. The Evolution of Culture
Santa Barbara County, California. Applied and Other Essays. Clarendon Press, Oxford.
EarthWorks, Inc., Fresno, CA. Submitted to 30 Plog, F. 1974. The Study of Prehistoric Change.
CES/CEVNC,Vandenberg Air ForceBase. Academic Press, NewYork.
Lipo, C. P., M. J. OBrien, M. Collard, and S. J. Richerson, P. J., and R. Boyd. 2001. Institu-
Shennan, eds. 2006 Mapping Our Ancestors: tional Evolution in the Holocene: The Rise
Phylogenetic Approaches in Anthropology of Complex Societies In The Origin of Human
and Prehistory. Aldine Transaction, New Social Institutions, edited by W. G. Runci-
Brunswick,NJ. man, pp. 197204. Proceedings of the British
Mace, R., and C. J. Holden. 2005. A Phylogenetic Academy110.
Approach to Cultural Evolution. Trends in Richerson, P. J., R. Boyd, and R. L. Bettinger.
Ecology and Evolution 20(3):116121. 2001. Was Agriculture Impossible during
Martin, P. S., and F. Plog. 1973. The Archaeology the Pleistocene but Mandatory during the
of Arizona. Natural History Press, Garden Holocene? A Climate Change Hypothesis.
City,NY. American Antiquity 66(3):387411.
Mauldin, R. P. 1993. The Relationship between Rosenberg, M. 1994. Pattern, Process, and
Ground Stone and Agricultural Intensification Hierarchy in the Evolution of Culture. Journal
in Western New Mexico. Kiva 58:317330. of Anthropological Archaeology 13:307340.
McKim, R. L., C. G. Lebow, D. R. Harro, and A. Schiffer, M. B., and J. M. Skibo. 1997. The
M. Munns. 2007. CA-SBA-212: Sea Mammal Explanation of Artifact Variability. American
Hunting and Other Late Holocene Littoral Antiquity 62(1):2750.
Adaptations, Vandenberg Air Force Base, Santa Shott, M. J., and P. Sillitoe. 2004. Modeling Use-
Barbara County, California.Applied EarthWorks, Life Distributions in Archaeology Using New
Inc., Fresno, CA. Submitted to 30 CES/ Guinea Wola Ethnographic Data. American
CEVNC,Vandenberg Air ForceBase. Antiquity 69(2):339355.
Moran, E. F. 1984. Limitations and Advances Smith, E. A. 1984. Explanatory Limitations of
in Ecosystems Research. In The Ecosystem the Ecosystem Concept. In The Ecosystem
Concept in Anthropology, edited by E. F. Moran, Concept in Anthropology, edited by E. F. Moran,
pp. 332. AAAS Selected Symposium 92. pp. 5185. AAAS Selected Symposium 92.
Westview Press, Boulder,CO. Westview Press, Boulder,CO.
Moylan, J. W., M. Borgerhoff Mulder, C. M. Smith, E. A. 2000. Three Styles in the Evolu
Graham, C. L. Nunn, and N. T. Hakansson. tionary Analysis of Human Behavior.
2006. Cultural Traits and Linguistic Trees: In Adaptation and Human Behavior: an
Phylogenetic Signal in East Africa. In Anthropological Perspective, edited by L. Cronk,
Mapping Our Ancestors: Phylogenetic Approaches N. Chagnon, and W. Irons, pp. 2746. Aldine
in Anthropology and Prehistory, edited by C. de Gruyter, Hawthorne,NY.
P. Lipo, M. J. OBrien, M. Collard, and S. J. Smith, E. A., and B. Winterhalder, eds. 1992a.
Shennan, pp. 3352. Aldine Transaction, New Evolutionary Ecology and Human Behavior.
Brunswick,NJ. Aldine de Gruyter, NewYork.
Smith, E. A., and B. Winterhalder. 1992b. Ethnology Bulletin 120. Smithsonian

Natural Selection and Decision Making: Institution Press, Washington,DC.
Some Fundamental Principles. In Steward, J. H. 1955. Theory of Culture Change.
Evolutionary Ecology and Human Behavior, University of Illinois Press, Urbana.
edited by E. A. Smith and B. Winterhalder, Tehrani, J. J., and M. Collard. 2009. The
pp. 2560. Foundations of Human Behavior, Evolution of Material Culture Diversity
S. B. Hrdy, general editor. Aldine De Gruyter, among Iranian Tribal Populations. In Pattern
NewYork. and Process in Cultural Evolution, edited by
Soltis, J., R. Boyd, and P. J. Richerson. 1995. S.Shennan, pp. 99111. University of Califor-
Can Group-Functional Behaviors Evolve by nia Press, Berkeley.
Cultural Group Selection? An Empirical Test. Torrence, R. 1983. Time Budgeting and Hunter-
Current Anthropology 36(3):473. Gatherer Technology. In Time Budgeting
Stevens, N. E., and B. F. Codding. 2009. Inferring and Hunter-Gatherer Economy in Prehistory: A
the Function of Projectile Points from the European Perspective, edited by G. Bailey, pp.
Central Coast of Alta California. California 1122. Cambridge University Press, NewYork.
Archaeology 1(1):728. Torrence, R. 1989. Time, Energy and Stone Tools.
Stevens, N. E., R. T. Fitzgerald, N. Farrell, M. Cambridge University Press, NewYork.
A. Giambastiani, J. M. Farquhar, and D. Ugan, A., J. Bright, and A. Rogers. 2003. When
Tinsley. 2004. Archaeological Test Excavations Is Technology Worth the Trouble? Journal of
at Santa Ysabel Ranch, Paso Robles, San Luis Archaeological Science 30(10):13151329.
Obispo County, California. Report prepared Winterhalder, B., and E.A. Smith. 2000.Analyzing
for Weyrich Development Company, LLC. Adaptive Strategies: Human Behavioral
Cultural Resource Management Services, Ecology at Twenty-five. Evolutionary
Paso Robles. Anthropology 9:5172.
Steward, J. H. 1936. The Economic and Social Wright, K. I. 1994. Ground-Stone Tools and
Basis of Primitive Bands. In Essays in Hunter-Gatherer Subsistence in Southwest
Anthropology Presented to A. L. Kroeber, edited Asia: Implications for the Transition to
by R. H. Lowie, pp. 331350. University of Farming. American Antiquity 59(2):238263.
California Press, Berkeley. Wynne-Edwards, V. C. 1962. Animal Dispersion
Steward, J. H. 1938. Basin-Plateau Aboriginal in Relation to Social Behavior. Oliver & Boyd,
Sociopolitical Groups. Bureau of American London.
Fifteen
Evolution of the Slate Tool

Industry at Bridge River, British
Columbia
Anna Marie Prentiss, Nathan Goodale, Lucille E. Harris, and Nicole

Crossland
The study of artifact variation with concepts from evolutionary theory has
been front and center throughout nearly all of the published works in evolu-
tionary archaeology. Since the early formulations examining the dichotomy
of style and function (Dunnell 1978), researchers have developed a variety of
sophisticated approaches to identifying and explicating variability in the his-
tories of artifact traditions. Underlying much of this research has been the
Neo-Darwinian assumption that evolutionary processes act on populations
of the smallest underlying characters (genes in biology; variously, memes or
traits in culture) and their phenotypic manifestations (e.g., artifacts in cultural
contexts).Within this framework larger scale cultural phenomena identified by
anthropologists as technological systems, subsistence strategies, and social orga-
nizations are viewed as ephemera, byproducts of more fundamental processes
of change. Thus, artifact performance characteristics and style are considered
markers of that fundamental process. Performance characteristics changed by
virtue of a selection-like process while stylistic evolution is viewed as a product
of drift (e.g., OBrien and Lyman 2000). However, research in anthropolog-
ical and linguistic phylogenetics (e.g., Mace and Holden 2005) point to the
importance of evolutionary entities organized and evolving on higher scales.
Macroevolutionary archaeologists (Rosenberg 1994; Spencer 1997) make simi-
lar arguments.The latter scholars argue that the evolution of functional designs
may be affected by the role of those designs in wider cultural matrices for
example, harpoon designs in the Arctic or cornet designs in nineteenth and
267
268 Prenti ss e t a l .
early twentieth centuries (Eldredge 2009; Mason 2009). If this is the case then
it is possible that artifact evolution could be overridden by processes of change
on the scale of socioeconomic and political strategies. Likewise in the realm of
style, evolution of style may be affected by cultural matrix as human groups use
styles as markers of identity and success (Bettinger etal. 1996). In this chapter
we outline a study of slate tools from a prehistoric housepit village in British
Columbia with the goal of examining the impact of different scales of evo-
lution on the history of a single lineage of artifacts. Results suggest that the
evolutionary process was complicated with tool performance characters and
social matrix playing equal roles.
Theories of Artifact Evolution

As Eldredge (2000, 2009) has noted, artifact evolution is not a simple process.
Multiple factors could affect the development, persistence, and extinction of
particular forms, and unlike biology, extinction is not forever. Consequently,
pulling apart the various cultural and biological strands influencing change
in artifacts can be a challenging process. In the following discussion, we
briefly outline some of the major evolutionary factors required for artifact
evolution.
Cultural Microevolution
At the most simple level, artifacts change as a consequence of simple trial and
error experimentation. However, as noted by Boyd and Richerson (1985),
trial and error is generally an expensive way to go and people commonly
rely on some form of cultural transmission when making decisions about
tool manufacture. A large literature has developed concerning artifacts and
cultural transmission (e.g., Bettinger and Eerkens 1999; Eerkens and Lipo
2007; Henrich 2001; Jordan and Shennan 2003; Mesoudi and OBrien 2008;
OBrien 2008 and references therein; OBrien and Lyman 2003 and refer-
ences therein). We have learned that there are a variety of decision rules taken
by people and these can affect the direction of artifact evolution. Cultural
transmission can be structured by conformist, prestige seeking and content
biases (Boyd and Richerson 1985, 2005a). Critically, biological process is not
required for cultural transmission though over the long term some pathways
can have fitness implications. But even without fitness implications, human
preference alone can have significant impacts on the histories of artifacts.
However, as outlined by Eerkens and Lipo (2007), the impact of transmission
will vary with information transmitted, context of transmission, and nature of
transmission process.
Evolution of the Slate Tool Industry at Bridge River, British Columbia 269
Cultural Macroevolution
As proponents of evolutionary archaeology have argued for some time now
(e.g., OBrien and Lyman 2000), archaeologists rarely have access to data ade-
quate for the study of cultural microevolution. Consequently, much of our
research is better termed cultural macroevolution and emphasizes variability in
the tempo and mode of artifact evolution over long time spans. Much of the
research of evolutionary archaeology relies on the approach known as organis-
mic macroevolution (Eldredge 1989) or the slow accumulation of phenotypic
traits. Lyman and OBrien (1997), for example, argue that Darwinian evolu-
tion is continuous and gradual, a literal extension of the microevolutionary
process. From this perspective, it is impossible to define particular artifact types
other than as arbitrary constructions (e.g., artifact analogues of chrono-species)
because change is continuous. Artifact evolution is therefore understood in a
strict Neo-Darwinian framework as a consequence of either selection or drift
gradually affecting variability in population level characters of artifacts across
multiple generations. Evolutionary archaeology has generated a number of
studies to support this framework. Many implicate the economic payoffs of
artifacts and conclude a selection-like process drove change (e.g., Abbott etal.
1996; OBrien etal. 1994). Others explain change as a consequence of neu-
tral sorting or drift-like processes (e.g., Neiman 1995; Shennan and Bentley
2008). A fundamentally important aspect of this work is the implication that
artifacts are evolving for their own specific characters. They are not recog-
nized as members of larger sets, subject to top-down evolutionary process (e.g.,
Eldredge 1985).
However, there is good reason to believe that artifact evolution may be con-
ditioned by an array of more complex factors. Although evolutionary archae-
ology was not initially receptive, some researchers raised the possibility that
higher level processes or those associated with social and cultural dynam-
ics outside the fitness potential of the artifacts themselves could affect the
evolution of those artifacts (Neff 1992). Two groups of scholars subsequently
promoted the idea that cultural evolution could act on multiple scales from
artifact up through the more complex forms of culture. Macroevolutionary
archaeologists argued for evolutionary process acting on these higher scales
often independent of action on the level of artifact (Prentiss and Chatters 2003;
Rosenberg 1994; Spencer 1997). Phylogenetics researchers independently
argued that whether data are drawn from languages or other cultural charac-
ters, evolution clearly acts on a variety of scales (e.g., Boyd etal. 1997; Holden
and Shennan 2005; Jordan and Mace 2005; Mace and Holden 2005).
Although relatively little empirical archaeological work has been pursued since
Neff s original arguments, new studies are offering support for the hierarchical
model. Spencer and Redmond (2001; see also Spencer 2009) demonstrated
coevolution of traits on several simultaneous scales during the emergence of
the Zapotec state. Mason (2009) points out that elements of harpoon form and
function evolved in the Bering Strait in unison with wider socioeconomic and
political changes particularly associated with the rise of the Punuk and Thule
whalers. Most critically, he points out that harpoon evolution was predicated on
change at those higher scales of integration. Chatters (2009) links the fluores-
cence of Pacific Northwest geophyte pit-cooking coincided with the Middle
Holocene organizational revolution associated with the emergence of collector
economic strategies. He points out that it was not the fitness potential per se of
the pits themselves but their role within human socioeconomies that changed.
Kuijt and Prentiss (2009) also examine the role of wider sociocultural develop-
ments in conditioning a variety of changes in artifact inventories during the
late Epipaleolithic of the Near East. Eldredge (2009) demonstrates that cornet
evolution was often affected by wider social trends in late nineteenth and early
twentieth century western European and American society regardless of vari-
ability in performance characteristics of the horns themselves.
Biological Impacts on Artifact Frequencies

Much of the now classic transmission theory literature in Darwinian anthropol-
ogy is founded on the assumption that human decision making, and thus trans-
mission strategies, have fitness implications (e.g., Boyd and Richerson 2005b).
As noted by Bettinger (2008), cultural transmission evolved in humans for its
adaptive value in an environment that was neither perfectly static nor constantly
changing. As he argues (see also Richerson etal. 2005), perfect homogeneity
would favor genetics as the cheapest alternative; regular change would favor pro-
pensity for individual learning. But cultural transmission is most cost effective
where there is a substantial degree of unpredictability. Biologically, this meant
that during human evolution, there must have been significant fitness payoffs
associated with the evolution of the cognitive structures that would permit
these behaviors and also for variation in the nature of those practices. Extending
this logic one step further, it means that we should be able to model variation
in transmission expecting fitness payoffs for optimal behavior (Bettinger 2009).
If artifacts are at least partial products of transmission processes, then we could
expect fitness of their users to play at least some role in their history.
The issue of biological fitness in cultural evolution is complicated on at least
two levels. It is possible that artifacts may simply have never (or at least rarely)
made a significant impact on individual fitness. Bamforth (2002) argues that
archaeologists have never demonstrated a significant relationship between arti-
fact histories and variability in human reproductive fitness. Indeed it is likely
that the vast majority of artifacts have a very negligible impact on fitness. But
this does not mean that we should give up all hope of considering relationships
between biology and culture. Bettinger (2009; see also Soltis etal. 1995) points
to group selection and cultural macroevolution as a more likely candidate.
Membership in groups with integrated socioeconomic and political strategies
and within-group dominated transmission systems may have offered stronger
impacts on fitness than idiosyncratic tactics associated with (and artifacts used
by ) individuals. Spencer (2009; Spencer and Redmond 2001) demonstrates
rising population in the Valley of Oaxaca associated with success of the emer-
gent Zapotec state. Prentiss (2009) recognizes short-lived population increases
associated with each of several changes in broad scale socioeconomic strate-
gies during the Middle to Late Holocene of the Pacific Northwest. Prentiss
and Lenert (2009) outline socioeconomic strategies that permitted lower lat-
itude populations to become culturally and biologically adapted to Arctic
environments.
Researchers in evolutionary archaeology solved some of these problems by
drawing on the theoretical tenets of Dawkins (1976, 1990) and assuming that
artifacts are simply extensions of the human phenotype (OBrien and Lyman
2000). Human artifacts are thus seen as equivalent to beaver dams and wasp
nests. From this standpoint, just as beaver dam evolution would probably mark
evolution of the beaver itself, human artifact evolution should be an indica-
tor of human biological evolution. Critiques of this argument (e.g., Gabora
2006) emphasize the logical disassociation between the objects of evolution
(artifacts) and the appropriate replicators. If artifacts are markers of biological
evolution then they must evolve through a genetic process but they cannot
because they must be made by people and no one has yet conclusively dem-
onstrated a direct link between genes and artifact types. Although this takes us
back to the fundamental importance of transmission theory and macroevolu-
tionary cultural modeling, it does not fully exhaust the possibility of biological
impacts on artifact histories (Mesoudi and OBrien 2008).
Macroevolutionary research into human biological history clearly demon-
strates artifact evolution during speciation events. No one doubts that there were
significant artifact changes after the emergence of Homo ergaster, and later, H.
heidelbergensis, H. neandertalensis, and H. sapiens. And this means that our propen-
sity for artifact manufacture and use is tied to our cognitive abilities. We have
not demonstrated, however, that change in artifacts in any simple and direct way
mark the biological process outside of that which is associated with speciation.
The Slate Industry at the Bridge RiverSite

The Bridge River site is a village of approximately 80 semi-subterranean
housepits located in the Middle Fraser Canyon area of British Columbia
(Figures15.1 and 15.2). Extensive mapping, test excavations, and radiocarbon
15.1. Major archaeological sites in the Middle Fraser Canyon.
dating at the site have revealed that the village was initiated at approximately
1800 BP1 with a limited number of houses but subsequently grew, such
that by ca. 12001300 BP at least 29 houses were simultaneously occupied
(Prentiss etal. 2008) (Figure15.3). After a period of apparent abandonment
after ca. 1000 BP, the village was reestablished between 400 and 100 BP This
study will focus on the earlier village, particularly Bridge River (BR) 2 (ca.
15501325 BP) and BR 3 (ca. 12751100 BP). Recent excavations have been
designed to examine the development of interhousehold ranking in the vil-
lage. Excavations, conducted in 2008 and 2009, centered on large (>15 m in
diameter) and small (<15 m) housepits. As outlined in Prentiss etal. (2008),
housepit stratigraphy at Bridge River is highly variable and can be very
complex. Individual house strata generally include floor deposits made from
transported clay and silt mixed with small gravels, capped by roof depos-
its, typically also clay but with much higher numbers of larger clasts. Roof
deposits are typically burned and when clay content is high, are red-brown
in color, in stark contrast to the gray-brown floors. Floors are also evidenced
39
37
38
333
0
338
33
337
78 77 47
65
220 34
45
35 64 48
66
335
33
33
52
6
54 53 46
200 50
31 32 49
55 56
71
33 44
1 30 57
180
33
29 28 70 41
4
27
33
51
0
26 59 69 43
67
160 25 63
80 60 42
24 58 14
33
2
329
36
62
33
140
3
23 5
21 61
20
22 19
15
120 13
32
12
33
9
16
1
17 330
4
100 3 1 11
18 79
2 7
8 10
6
80
32
9
N
9
W E
32
8
S
60
Bridge River Site
EeR14
Site Datumn 40
40
Excavation Unit
20
Survey Baseline 0
Reference Grid 20 20
External Pit Features 40
Housepit Rim 60
Contours
Index
0 5 10 20 30 40
Supplemental Meters Matt Hogan
15 February
15.2. Bridge River site with excavation grid superimposed.
by artifacts and faunal remains lying in horizontal positions and frequent

presence of hearth and cache pit features. Houses excavated in 20082009
were often characterized by stratified deposits including multiple floors and
burned roofs (Housepits 11, 16, 20, and 54; Figure15.4). Houses with single
floor and roof deposits were far less common (Housepit 24; Figure 15.5).
Artifacts used in this study were derived from floor, roof, and sometimes rim
(redeposited roof material) contexts, as long as they could confidently be
assigned to a particular occupation period.
The goal of the excavations, conducted in 20082009, was to reconstruct
the patterns of changing socioeconomic and political organization within the
Bridge River Site

British Columbia, Canada
EeR14
Bridge River 1 Bridge River 2
38
78
35 64 48
52
54
50
32 56 49
71
28
26 27 26
25
58
23
20
15
3 11
6
9
Bridge River 3 Bridge River 4

39
37
38
77
65
66 34
33
54 54
50
31 55
29 28 70 57
51 59
25 63
24 58 60 60
62
61 5
20 20
19 22
12
17 16 4 4
3 1 3
18 18
8 10
Housepit Rim N
Contours W E
0 20 40 60 80 100 120
Index S
Matt Hogan
Meters 19 September 2010
15.3. History of housepit occupations at the Bridge River site.
village. Because funding was not available for complete household excavations,
houses were sampled by excavating trenches in three activity sectors of each
house. Sectors were identified using applied archaeo-geophysics, particularly
magnetic susceptibility, to identify likely clusters of features (especially cache
pits identified as strong negative anomalies) associated with domestic activity
or kitchen areas (Prentiss etal. 2009). This procedure was highly successful,
as out of 15 BR 2 and 3 areas sampled, only one strong negative anomaly failed
to generate an activity area with the requisite features and associated artifacts.
15.4. Stratigraphic profile of Area 1 in Housepit 54.
15.5. Housepit 24 stratigraphic profile.
In most contexts we located multiple floors including more than one kitchen
related activity area, for example, in Housepit (HP) 54 where areas produced
more than a dozen stratified floors, most of which contained cache pits and/
or cooking hearths. Given the high degree of consistency in our results we are
confident that our assessments of interhousehold variability in artifacts are not
simply measuring variation in activity area function. Nearly all derive from
domestic activity zones. Artifacts from roof deposits are redeposited and have
been effectively randomized before excavation, meaning that they have been
cleared from primary use context and discarded into secondary or tertiary
contexts within trash dumps on theroof.
Analyses of lithic artifacts, faunal remains, and feature volume have gen-
erated provocative results (Prentiss etal. 2009, 2012). In a multivariate analy-
sis of relative cache pit volume, fire-cracked rock (FCR) counts (controlled
for excavation volume), mammalian predation signatures (mammal index; see
Prentiss etal. 2007, 2012), prestige item, and nonlocal raw material item counts
(again controlled for excavated volume) we demonstrated that during BR 2
times there was little evidence for significant interhousehold variation in these
measures. In contrast, during BR 3 times it was apparent that one large house
(24) featured excessively high numbers of prestige items, nonlocal raw materi-
als, and mammal remains (deer and dog), along with a relatively high score on
cache pit volume. We concluded that, based on the current evidence, wealth-
based household ranking2 had not developed until BR 3 times. Further, house-
pit size did not guarantee a strong signal of superior wealth and status because
while HP 24 scored strongly, similarly large HP 20 faltered. Indeed, a dramatic
decline in FCR scores at HP 20 may signal demographic decline (e.g., Ames
2006). For purposes of this study, we do not recognize any BR 2 assemblages
(HP 11, 20, and 54)as marking particularly high ranked groups. In contrast, we
assume that HP 24 was a high ranked household during BR 3 while all others
(HP 16, 20, and 54)were lower ranked.
Excavations at the Bridge River village revealed a unique slate tool indus-
try (Mandelko 2006). Indeed, as far as we know it is not replicated at any other
Plateau villages in North Americas Pacific Northwest. Slate and a very similar
material, known as silicified shale, expose locally in the Bridge River Valley and
can be found within a relatively short walking distance from the Bridge River
village. Bridge River occupants took a variety of approaches to manufacturing
tools from this material but tools were rarely made as finely as those of many
contexts on the nearby Northwest Coast (e.g., Graesch 2007). Nearly all slate
tools from Bridge River have chipped or snapped edges. At a bare minimum,
pieces of slate were simply shaped by snapping one or more margins to form a
desired shape and immediately used. Many have combinations of chipped and
snapped edges. Chipped edges often form concavities in the margins resembling
broad notches, though we have found relatively little evidence of hafting wear
and conclude that most of these tools were hand-held. A far smaller percentage
(20%) of the tools were shaped by sawing, a laborious task requiring cutting the
slate using a sandstone saw or quartz sand and hemp-string (two sawed edges on
triangular tool in Figure15.6). Preliminary experimentation in the laboratory
has demonstrated that it could take as much as 56 hours to cut approximately
20cm with hemp string and fine sand. Sawing strengthens working edges and
provides more solid margins for hafting or grasping by hand. A far higher per-
centage (3050%) of the tools received facial grinding (ground face is evident
on the lower portion of the four-sided tool in Figure15.6), a laborious (though
less so than sawing) procedure designed to strengthen the working edges of the
tool (Graesch 2007). Tool shapes are somewhat variable3 but generally divide
into two sets, three and four sides (Figure15.6). Slate tool function is relatively
consistent; approximately 80% are scrapers with a consistent pattern of intense
rounding with deep striations (perpendicular to the working edge) and medium
15.6.Three (left) and four (right)-sided slate tools from Bridge River. The three-sided tool has
two sawed edges. The four-sided tool has been ground on its lowerface.
to dull polish. At present4 we believe these to be tools employed in hide scrap-

ing, though some exhibit intensive edge damage consist with heavier duty work
such as wood shaving or bark removal.The other 20% of tools are primarily slate
knives and feature similar use-wear, though striations run parallel to the working
edge. Fewer than 3% of the slate tools were adze slugs or adze preforms. It is still
premature to state with any confidence that specific use-patterns changed over
time or varied by household. However, preliminary use-wear results do suggest
the possibility that tool functions may have diversified somewhat by BR 3 times.
However, this would not be surprising, as a statistical effect of the rise in numbers
of tools is documented for this period. Given their relatively frequent presence
in Bridge River site deposits and the high degree of variability in manufacturing
investment, the slate industry provides an excellent data set for exploring issues
in artifact evolution.
Studying Evolutionary Process at BridgeRiver

The Bridge River site provides us with the opportunity to conduct an ini-
tial examination of evolutionary processes associated with slate tools. As out-
lined previously, Eerkens and Lipo (2007) note that cultural transmission can be
affected by information content, transmission context, and transmission mode.
To understand the evolutionary process it is important to control for as many
of these factors as possible. By controlling for some variables we can more effi-
ciently examine the variable effects of others on the history of tool production.
Constants
We are able to hold information transmission relatively constant because our
study examines a specific tool form, produced using a limited array of tech-
niques on a single raw material type. Bridge River slate tools were produced,
used, and discarded within the housepits at the site. Surface collections and
extramural test excavations have failed to produce slate tools outside household
contexts. Abundant slate debitage on house floors and in roof and rim deposits
backs up our assertion that tools were produced within the houses. Occupation
of the Bridge River village during periods 13 was constant.There is no direct
evidence for abandonments and reoccupations. Nor is there direct evidence
for change in broader cultural practices (whether measured as artifact styles or
approaches to socioeconomy [they remained winter village collectors]) of the
inhabitants. Based on these patterns we conclude that there was probably no
change in the sociolinguistic or ethnic makeup of the Bridge River peoples
during the period of ca. 18001100 BP. This means that there is no question
as to the makers and the tradition of manufacture; our data are not influ-
enced by cultural replacements as may be the case in some other evolutionary
archaeological studies (e.g., Lyman etal. 2009). Transmission mode is hard to
evaluate with archaeological data. However, given the likelihood that learning
occurred within house groups and that, from ethnographic studies among the
Sttimc (Mathewson 2008; Teit 1906), important household information was
generally passed down to younger persons by older people from the extended
family, we can be clear that transmission was most likely a many to one model.
Because many to one transmission is generally associated with conservatism
(MacDonald 1998; Shennan 2002), it is not surprising (at least at this level of
resolution) that change in Bridge River slate tools across periods BR 2 and 3
only manifested as variability in item frequencies.
Variants
Research at Bridge River (Prentiss etal. 2008, 2009) has demonstrated signifi-
cant change across the BR 2 to 3 periods. There was a near doubling of village
size during this time, as radiocarbon dating of house floors suggests a rise from a
maximum of 18 BR 2 houses to at least 29 houses occupied during BR 3. FCR
densities also rise in most houses during BR 3 times and assuming that more
FCR primarily results from more people to be cooked for (Prentiss etal. 2007)
we can conclude that household membership may have also increased. However,
some houses show the opposite pattern; Housepit 20 produced a dramatic drop
in FCR between the BR 2 and 3 floors. Zooarchaeological data indicate a pro-
cess of resource extensification and the development of resource depression (e.g.,
Broughton 1994) at least in some important species between BR 2 and 3.This is
particularly evident in ungulate assemblages where by BR 3 times, remains from
most houses consist almost entirely of lower limbs and fragmentation is extreme
implying longer hunting trips and increasingly intense processing activities.
Finally, although houses were constructed in a wide range of sizes throughout
the village history, it is only during BR 3 that we find indicators of wealth-based
interhousehold ranking. Nineteenth and early twentieth century ethnographic
research (Teit 1906) documented a complex society in the Mid-Fraser villages

that was characterized by hereditary ranking, potlatching, and competitive feast-
ing. Much archaeological work in the Mid-Fraser villages has focused on devel-
oping an understanding of the evolution and organization of this pattern (e.g.,
Hayden 1997; Prentiss etal. 2003, 2007, 2008). Among BR 3 houses at Bridge
River, only HP 24 stands out as a potentially high-ranking household.
Hypotheses and Expectations

Given data limitations we cannot explore all of the interesting questions about
evolutionary processes and the Bridge River slate industry. For example, it is
impossible at this time to formally evaluate the possible effects of intra- versus
interhousehold transmission on slate tool manufacture. This will require more
exacting data on tool morphologies. However, we are in a position to offer
preliminary evaluations of several more fundamental hypotheses.
1. If slate tools evolved purely for performance value (the basic selectionist model)
then we would expect change in frequencies of tools across the site without
significant variability between houses (assuming regular intercommunication
between house groups). If sawed and ground tools offer better designs than
purely chipped items then they could also be expected to rise in frequency
over time given the fact that they significantly enhance the ability of the tool to
withstand application to other materials.
2. In contrast, if change in household social standing, demographics, and/or economy
affected tool form or frequencies beyond the intrinsic properties of the tools
themselves then we should expect significant variability on an interhousehold basis.
3. If artifact style (measured as tool shape) has no adaptive value at any level (simple
neutral model) and if we assume regular interhousehold communication, but no
sharing of ideas between villages (as Bridge River appears to have been the only
village employing this technology to any significant degree), then, drawing from
Neiman (1995) and Shennan and Bentley (2008), we would expect to see low
rates of innovation and high degrees of similarity between houses.
4. If stylistic form is under control of some other factor (e.g., anti-conformist
transmission bias per Boyd and Richerson 1985; see also Shennan and Wilkinson
2001) then it should be in evidence on an interhousehold basis. In this case we
might expect to see stylistic approaches unique to particular houses.
Data Collection
To evaluate the hypotheses we collected a variety of basic artifact data on
slate tools judged to be complete (not small fragments) (Table15.1). To assess
investment in the enhancement of performance characteristics we recorded
table15.1. Slate tool data (counts based on manufacture attributes)
Housepit Ground Not Sawed Chipped Sides Sawed and ground

Ground
Three Four
24 (BR3) 8 12 7 13 14 6 5
20 (BR3) 1 4 2 3 1 4 1
20 (BR2) 4 2 1 5 0 6 1
54 (BR3) 8 21 7 22 6 23 2
54 (BR2) 2 0 1 1 0 2 1
11 (BR2) 1 6 1 6 3 4 0
16 (BR 3) 4 11 1 14 5 10 0
Table15.2. Slate tool data (summed margin length measure-

ments [cm]) and excavated volume (cubiccm)
Housepit Sawed edge Total edge Excavated volume

24(BR3) 71.1 455.1 1,843,750
20 (BR3) 4.8 95.6 1,819,750
20(BR 2) 9.8 100.4 480,000
54 (BR3) 46.9 683.2 2,713,130
54 (BR2) 9.7 55.7 676,960
11 (BR2) 2.3 180.4 3,900,750
16 (BR3) 6.8 398.5 2,472,500
data on edge sawing (presence/absence, linear length of sawed edge, and linear
length of total tool margin) (Table15.2), presence/absence of facial grinding,
and basic artifact dimensions. All measurements were obtained with vernier
calipers. Edge lengths were obtained with relatively limited error given angu-
larity of the tool margins. Although measuring stylistic variability in slate tools
at Bridge River could be facilitated by mathematically rigorous and complex
algorithms (e.g., Fourier transforms), time has not yet permitted this level of
data collection. For this study we simply draw a distinction between tools with
three versus four sides (Figure15.6).
Analysis
Sawed Margins and Ground Faces
As illustrated in Figures 15.715.9, there is a dramatic increase in numbers
of slate tools between BR 2 and 3. However, there is relatively little change
in percentages of sawed edge tools compared to those with chipped edges.
Surprisingly, percentages of tools with ground faces reverse between BR 2 and
3, suggesting the possibility that over time fewer people were investing the extra
effort to produce such robust tools. However, when we look at frequencies of
0.1
0.08
0.06
TST/V
0.04
0.02
0
BR2 BR3
15.7. Ratio of total slate tools (TST) to excavated volume (V) (Table15.2 volume/10,000).
100
80
60 BR3
40 BR2
20
0
sawed not sawed
15.8. Percentages of sawed and chipped tools from BR 2 and 3 contexts at Bridge River.
80
60
40 BR2
BR3
20
0
G NG
15.9. Percentages of ground (G) and not ground (NG) tools during BR 2 and 3 occupations at
Bridge River.
tools with both ground faces and sawed margins, we still recognize a dramatic
increase between BR 2 and 3 but one that simply parallels the rise in numbers
of slate tools overall (Figure15.10). Thus, the only evident trend is increased
rates of production but without dramatic change in the nature of those tools.
Next we need to evaluate if these trends are even across the village or if they
are driven by actions in particular houses.
Bridge River 2 assemblages were relatively small, so to have enough tools
to see patterns we had to aggregate the data from two smaller houses (11 and
54)for comparison with those from a single larger house (20). Figure15.11
demonstrates that even controlling for volume of excavated sediment,
the larger housepit produces far more slate tools than either of the others.
0.012
0.01
0.008
0.006
TSGST/V
0.004
0.002
0
BR2 BR3
15.10.Total sawed and ground slate tools (TSGST) by volume (V) (Table15.2 volume/10,000).
0.16
0.14
0.12
0.1
0.08
0.06 N/V
0.04
0.02
0
20 11+54
15.11. Number of slate tools (N) per unit of excavated sediment (V) (Table15.2 volume/10,000)
during the BR 2 occupations at Bridge River.
Although there is no difference in the percentages of sawed tools between

houses (Figure 15.12), residents of Housepit 20 are clearly making a greater
investment in sawed edges (Figure15.13).
Analysis of potential BR 3 household ranking suggests that HP 24 ranked
highest followed by 54 and then 16 (Prentiss etal. 2009, 2012). Housepit 20
falls in the HP 54 range during BR 3 despite its large size but contributed
too few slate tools for analysis. As illustrated in Figure 15.14, Housepits 24
and 54 contributed similar numbers of slate tools, dominating that of HP
16. Ground and sawed margin tools are consistently more common in HP
24 (Figures15.1515.17). Stronger evidence for variability in household tool
production is evident in the data on lengths of sawed edges in BR 3 houses.
Given good sample sizes from the three houses, we were able to measure
length of sawed edges in three ways (length of sawed edge/slate tool count
per house; length of sawed edge/total edge length all slate tools per house; and
length of sawed edge/total edge of all sawed slate tools). Results were similar
100
80
60
20
40
11+54
20
0
sawed not sawed
15.12. Percentages of sawed and not sawed tools during BR 2 occupations at Bridge River.
0.12
0.1
0.08
0.06 TSE/TE
0.04
0.02
0
20 11+54
15.13. Ratio of total sawed edge (TSE) to total edge (TE) for all slate tools in BR 2 occupations
at Bridge River.
0.12
0.1
0.08
0.06 N/V
0.04
0.02
0
16 54 24
15.14. Number of slate tools (N) per unit of excavated sediment (V) (Table15.2 volume/10,000)
for BR 3 occupations at Bridge River.
100
80
60
sawed
40
not sawed
20
0
16 54 24
15.15. Percentage of sawed and not sawed tools from BR 3 occupations at Bridge River.
80
70
60
50
40 G
30 NG
20
10
0
16 54 24
15.16. Percentages of ground (G) and not ground (NG) tools from BR 3 occupations at Bridge
River.
0.3
0.25
0.2
0.15
SGST/TST
0.1
0.05
0
16 54 24
15.17. Ratio of sawed and ground slate tools (SGST) to total slate tools (TST) in BR 3
housepits.
in each case (Figures15.1815.20), demonstrating consistently highest scores

associated with Housepit24.
These results suggest that while numbers of slate tools rose from BR 2 to
3 times, relative percentages of more heavily modified tools did not signifi-
cantly change. Further, there is no significant change in the overall dimensions
(Length [L] Width[W] Thickness[Th]) of the tools across these periods
either (Students t = 1.995, p = 0.418, at 68 d.f.). Given the rise in overall num-
bers of tools it is tempting to conclude that these items were being simply
selected for on a pan-household or village scale, perhaps in relation to other
kinds of scraping and cutting tools. Ground and sawed tool data further back
up this conclusion because they essentially demonstrate temporal consistency,
at least from the standpoint of data sets drawn from all excavated houses com-
bined. However, these conclusions appear premature when interhousehold
data are considered. House size plays a partial role in the patterns of inter-
household variation. During the BR 2 occupation, Housepit 20 clearly pro-
duced the most tools and the greatest length of sawed margins per capita. The
pattern is less consistent during BR 3, where one large house (24) dominated
14
12
10
6
TSE/N
4
0
16 54 24
15.18. Ratio of total sawed edge length (TSE) to number of tools (N) with sawed margins.
0.18
0.16
0.14
0.12
0.1
0.08 TSE/TE
0.06
0.04
0.02
0
16 54 24
15.19. Ratio of total sawed edge (TSE) to total edge (TE) for all slate tools in BR 3 housepits.
0.6
0.5
0.4
0.3
TSE/TEST
0.2
0.1
0
16 54 24
15.20. Ratio of total sawed edge (TSE) to total edge of slate tools only (TEST).
80
70
60
50
40 BR3
30 BR2
20
10
0
3 sides 4 sides
15.21. Change in percentages of three- and four-sided tools across BR 2 and 3 occupations at
Bridge River.
the others in all ground and sawed tool indices, while the other large house
(20) produced so few tools it could not even be considered. This implicates
factors beyond simple house size and suggests that demographics and social
status may have played arole.
Household group size and status are generally correlated in Mid-Fraser
housepit villages, whether viewed through ethnographic or archaeological
lenses (Hayden 1997; Kennedy and Bouchard 1978; Prentiss et al. 2007; Teit
1906;). Slate tool frequencies also to match these relationships given their rel-
atively high frequency in houses with high FCR counts (HP 20 during BR
2 and HPs 24 and 54 in BR 3). It is evident, however, that it was not simply
demographics that drove these patterns. If that had been the case then, all things
equal, the BR 2 occupation of Housepit 20 should have produced as many or
more slate tools than the BR 3 occupations of at least Housepit 54 and pos-
sibly 24. This implicates household status ranking as a driving force behind
variation in the slate tool data and suggests that the explosion in BR 3 slate
tools was dependent on labor in select houses that could be prevailed upon to
produce not only the simple slate tools but also the far more time-consuming
ground and sawed tools. If household ranking after ca. 1300 BP was in any way
similar to that of the historic period, households could have included proper-
ty-owning elite families along with nonproperty-owning people and slaves.
Evidence for slavery is rare and ambiguous in the archaeological record (e.g.,
Donald 1997). But, frequent sawing of stone with pre-European technologies
in wealthy households could be an indicator.
Slate ToolShape
Percentages of three-sided versus four-sided tools do not change appreciably
between BR 2 and 3 on a pan-household or village scale (Figure15.21).Three-
sided tools are consistently and substantially fewer in number than those with
four sides.There is no significant change in the general size (L W Th) of the
tools over time (for three-sided tools: Students t = 2.068, p = 0.895, 23 d.f.; or
90
80
70
60
50
three
40 four
30
20
10
0
16 54 24
15.22. Percentages of three- and four-sided tools in BR 3 occupations at Bridge River.
for four-sided tools: t = 2.01, p = 0.411, 52 d.f.). This implies the possibility that
production of three- versus four-sided shapes offered little to no specific advan-
tage to anyone and thus varied at low levels throughout the village across time.
However, interhousehold analysis demonstrates some underlying patterning. As
shown in Figure15.22, only Housepit 24 produces a stylistic signature unique in
comparison with the rest of the village. Here, three-sided tools heavily dominate
and imply a house-specific preference for these shapes. This does not appear to
be related to tool function, as use-wear patterns do not vary with shape and
size is not statistically significant between shapes during BR 2 (Students t =
2.641, p = 0.48, 13 d.f.) or 3 (t = 2.986, p = 0.993, 73 d.f.). Although highly
preliminary, these data imply that select houses may have maintained distinct
stylistic traditions of slate tool manufacture. It is interesting and perhaps not
surprising to find that Housepit 24 had a unique tradition of tool manufacture
given its apparent high rank during BR 3 times. Given the houses position of
importance at Bridge River, household elders could have sought to maintain
a uniquely standardized tradition of tool manufacture as perhaps one of many
markers of household membership. In different words, it could also be seen as
a relatively short-lived (HP 24 was likely occupied for no more than 150years,
compared to HPs 20 and 54, whose occupations spanned up to about 500years)
effect of strong intragroup anti-conformist transmission.
Discussion
The Bridge River slate tool industry is characterized by a general trend toward
increased production levels.To us this implies that there may have been a simple
selection-like process at work favoring this tool form. We do not think it is
simply increased need (e.g., more wood working later in time) for this kind
of tool because other lithic tools could do the same job (indeed dacite wood
working tools are not common at Bridge River compared to similar nearby
contexts such as the Keatley Creek village). Plus, there is preliminary evidence
that tool-use practices may have actually diversified during the BR 3 occupa-
tions. But the picture is complicated by that fact that the most efficient tools
from a performance standpoint (with ground faces and sawed edges) were never
favored to any significant degree. It may be that the energetic expense required
for production typically outweighed any payoffs of these investments. It is not
surprising therefore that we would see the highest incidence of sawing and
grinding in the most densely populated and affluent households in the village.
We recognize a similarly complex history in reference to stylistic investment
measured as tool shape. In this case we recognize a slight increase in three-
sided tools but overall no major change through time. The limited variation
across the village and through time implicates the predictions of the neutral
model. But select households apparently did invest in a different way in style,
at least from a quantitative standpoint. This implies general background pat-
tern of regular interhousehold interaction, but layered over with occasional
examples anti-conformist trends.
In the larger picture, we recognize some degree of evolution favoring sim-
ple target values manifested as rising numbers of slate tools over time. But
the general trend could apparently be overridden through the development
of cultural practices associated with rise of ranked house groups. Ability to
attract labor determined that certain tool production systems could flourish
at greater frequencies. In evolutionary terms we argue that this is an example
of evolution acting on multiple scales. Thus we see the evolution of emergent
phenomena such as the organizational principles behind wealth-based social
inequality (e.g., Prentiss 2011; Rosenberg 2009) which override the evolution
of particular tool systems, in this case favoring the expansion of some only
in elite households. This provides a finer grained example of a process recog-
nized by others on courser scales (Eldredge 2009; Mason 2009; Spencer 2009;
Spencer and Redmond 2001).
These data implicate a complex process of interaction and information
sharing that changed subtly over time. Future investigations need to use more
precise and meaningful measures to study variability in interhousehold inter-
action, particularly emphasizing more complex measures of manufacture and
style. One promising approach to this could be through the use of approaches
outlined in Chapter13 by Goodale etal.
Acknowledgments
We thank Bill Andrefsky and Nathan Goodale for organizing the SAA session and
inviting our participation. Excavations at the Bridge River site were conducted
in partnership with the Bridge River (Xwisten) Band and funded by generous
grants from the National Science Foundation (Grants BCS-0313920 and BCS-
0713013). A large number of students participated in the field and laboratory
phases of this project and we thank in particular Lisa Fontes, Rachel Horowitz,
Tosh McKetta, Lee Reininghaus, and Maggie Schirack for their contributions.
Notes
1 All dates are calibrated. 3 Actual slate tool shapes are highly variable and
2 Ethnographic research in the Mid-Fraser doc- include a variety of triangles, squares, rect-
umented a complex hunter-gatherer-fisher angles, trapezoids, and other more amorphous
society known today as the Upper Lillooet or forms. Despite this variation, they can be most
Sttimc Nation (Teit 1906). Mid-Fraser soci- simply subdivided into three- and four-sided
ety was stratified with hereditary chiefs and items. There is significant potential for more
chiefly families, common people, and slaves. rigorous quantification and analysis of tool
Households in villages were ranked with the forms.
chief of the top ranking house having great- 4 Research into tool function and use history
est status and influence. Practices, better known is ongoing and will eventually include both
from the nearby Northwest Coast, such as elite use-wear and residue analyses. However, these
exchange networks, competitive feasting, and results are not yet fully available.
potlatching, were common.
References
Abbot, Alysia, A., Robert D. Leonard, and Between Biology and the Social Sciences edited by
George T. Jones. 1996. Explaining the Change P. Weingart, S. D. Mitchell, P. J. Richerson, and
from Biface to Flake Technology. In Darwinian S. Maasen, pp. 355384. Lawrence Erlbaum
Archaeologies, edited by Herbert D. G. Maschner, Associates, Mahwah,NJ.
pp. 3342. Plenum Press, NewYork. Boyd, Robert, and Peter J. Richerson. 1985.
Bamforth, Douglas B. 2002. Evidence and Culture and the Evolutionary Process. University
Metaphor in Evolutionary Archaeology. of Chicago Press, Chicago.
American Antiquity 67: 435452. Boyd, Robert, and Peter J. Richerson. 2005a.
Bettinger, Robert L. 2009. Macroevolutionary Not By Genes Alone: How Culture Transformed
Theory and Archaeology: Is There a Big Picture? Human Evolution. University of Chicago Press,
In Macroevolution in Human Prehistory: Evolutionary Chicago.
Theory and Processual Archaeology, edited by Anna Boyd, Robert, and Peter J. Richerson. 2005b.
Marie Prentiss, Ian Kuijt, and James C. Chatters, The Origin and Evolution of Cultures. Oxford
pp. 275296. Springer, NewYork. University Press, Oxford.
Bettinger, Robert L., Robert Boyd, and Peter Broughton, Jack. 1994. Late Holocene Resource
J. Richerson. 1996. Style, Function, and the Intensification in the Sacramento Valley,
Cultural Evolutionary Process. In Darwinian California: The Vertebrate Evidence. Journal of
Archaeologies, edited by Herbert D. G. Maschner, Archaeological Science 21: 501514.
pp. 133164. Plenum Press, NewYork. Chatters, James C. 2009. A Macroevolutionary
Bettinger, Robert L., and Jelmer Eerkens. 1999. Perspective on the Archaeological Record of
Point Typologies, Cultural Transmission, and North America. In Macroevolution in Human
the Spread of Bow and Arrow Technology in Prehistory: Evolutionary Theory and Processual
the Prehistoric Great Basin. American Antiquity Archaeology, edited by Anna Marie Prentiss,
64: 231242. Ian Kuijt, and James C. Chatters, pp. 213234.
Boyd, Robert, M. B. Mulder,William H. Durham, Springer, NewYork.
and Peter J. Richerson. 1997. Are Cultural Dawkins, Richard. 1976. The Selfish Gene. Oxford
Phylogenies Possible? In Human by Nature: University Press, Oxford.
Dawkins, Richard. 1990. Extended Phenotype: Cultural Evolution? In The Evolution of Cultural
The Long Reach of the Gene (new ed.). Oxford Diversity: A Phylogenetic Approach, edited by R.
University Press, Oxford. Mace, C. J. Holden, and S. Shennan, pp. 1129.
Donald, Leland. 1997. Aboriginal Slavery on the UCL Press, London.
Northwest Coast of North America. University of Jordan, P., and T. Mace. 2005. Tracking Culture-
California Press, Berkeley. Historical Lineages: Can Descent with
Dunnell, Robert C. 1978. Style and Function: A Modification Be Linked to Association by
Fundamental Dichotomy. American Antiquity Descent? In Mapping our Ancestors: Phylogenetic
43: 192202. Approaches in Anthropology and Prehistory, edited
Eerkens, Jelmer W., and Carl P. Lipo. 2007. Cultural by C. L. Lipo, M. J. OBrien, M. Collard, and S.
Transmission Theory and the Archaeological J. Shennan, pp. 149168. Aldine/Transaction,
Record: Providing Context to Understanding New Brunswick,NJ.
Variation and Temporal Changes in Material Jordan, P., and S. Shennan. 2001. Cultural
Culture. Journal of Archaeological Research 15: Transmission, Language, and Basketry
239274. Traditions amongst the California Indians.
Eldredge, Niles. 1985. Unfinished Synthesis: Journal of Anthropological Archaeology 22: 4274.
Biological Hierarchies and Modern Evolutionary Kennedy, Dorothy I., and Randy Bouchard.
Thought. Oxford University Press, NewYork. 1978. Fraser River Lillooet: An Ethnographic
Eldredge, Niles. 1989. Macroevolutionary Dynamics: Summary. In Reports of the Lillooet Archaeological
Species, Niches, and Adaptive Peaks. McGraw Project Number 1. Introduction and Setting, edited
Hill, NewYork. by Arnoud Stryd and Stephen Lawhead, pp.
Eldredge, Niles. 2000. Biological and Material 2255. National Museum of Man, Mercury
Cultural Evolution: Are There Any True Series, Archaeological Survey of Canada Paper
Parallels? Perspectives in Ethology 13: 113153. No. 73, Ottawa.
Eldredge, Niles. 2009. Material Cultural Kuijt, Ian, and Anna Marie Prentiss. 2009.
Macroevolution. In Macroevolution in Human Niche Construction, Macroevolution, and
Prehistory: Evolutionary Theory and Processual the Late Epipaleolithic of the Near East. In
Archaeology, edited by Anna Marie Prentiss, Macroevolution in Human Prehistory: Evolutionary
Ian Kuijt, and James C. Chatters, pp. 297316. Theory and Processual Archaeology, edited by
Springer, NewYork. Anna Marie Prentiss, Ian Kuijt, and James C.
Gabora, Liane. 2006. The Fate of Evolutionary Chatters, pp. 253274. Springer, NewYork.
Archaeology: Survival or Extinction? World Lyman, R. Lee, and Michael J. OBrien.
Archaeology 38: 690696. 1997. The Concept of Evolution in Early
Graesch, Anthony P. 2007. Modeling Ground Slate Twentieth Century Americanist Archeology.
Knife Production and Implications for the Study In Rediscovering Darwin: Evolutionary Theory
of Household Labor Contributions to Salmon in Archeological Explanation, edited by C.
Fishing on the Pacific Northwest Coast. Journal Michael Barton and G. A. Clark, pp. 2148.
of Anthropological Archaeology 26: 576606. Archeological Papers of the American
Hayden, B. 1997. The Pithouses of Keatley Creek. Anthropological Association No.7.
Harcourt Brace College, Fort Worth. Lyman, R. Lee, Todd L. VanPool, and Michael
Henrich, Joseph. 2001. Cultural Transmission J. OBrien. 2009. The Diversity of North
and the Diffusion of Innovations: Adoption American Projectile Point Types, Before
Dynamics Indicate the Biased Cultural and After the Bow and Arrow. Journal of
Transmission Is the Predominate Force in Anthropological Archaeology 28:113.
Behavioral Change. American Anthropologist MacDonald, D. H. 1999. Subsistence, Sex, and
103: 9921013. Cultural Transmission in Folsom Culture.
Holden, C. J., and Stephen Shennan. 2005. Journal of Anthropological Archaeology 17:
Introduction to Part 1: How Tree-like Is 217239.
Mace, R., and Holden, C. J. 2005. A Phylogenetic Archaeological Perspectives. University of Utah
Approach to Cultural Evolution. Trends in Press, Salt LakeCity.
Ecology and Evolution 20: 116121. Prentiss, Anna Marie. 2009. The Emergence of
Mandelko, Sierra A. 2006.The Slate and Silicified New Socioeconomic Strategies in the Middle
Shale Industry Recovered from the Bridge and Late Holocene Pacific Northwest Region
River Site (EeRl4) British Columbia. M.A. of North America. In Macroevolution in Human
Thesis, Department of Anthropology, The Prehistory: Evolutionary Theory and Processual
University of Montana. Archaeology, edited by Anna Marie Prentiss,
Mason, Owen K. 2009. The Multiplication Ian Kuijt, and James C. Chatters, pp. 111132.
of Forms: Bering Strait Harpoon Heads as Springer, NewYork.
a Demic and Macroevolutionary Proxy. In Prentiss, Anna Marie. 2011. Social Histories
Macroevolution in Human Prehistory: Evolutionary of Complex Hunter-Gatherers: Pacific
Theory and Processual Archaeology, edited by Northwest Prehistory in a Macroevolutionary
Anna Marie Prentiss, Ian Kuijt, and James C. Framework. In Hunter-Gatherer Archaeology
Chatters, pp. 73110. Springer, NewYork. as Historical Process, edited by Kenneth E.
Mattewson, Lisa. 2008. Psychology. In Wencw Sassaman and Donald H. Holly, Jr., pp.1733.
Iz: Sqwqwels sLaura True Stories by Laura University of Arizona Press, Tucson (in press).
Thevarge, edited by Lisa Matthewson, Prentiss, Anna Marie, Eric Carlson, Nicole
Christiana Christodoulou, John Lyon, and Crossland, Hannah Schremser, and Lee
Martin A. Oberg. University of British Reininghaus. 2009. Report of the 2008
Columbia Working Papers in Linguistics Vol. University of Montana Investigations at the
22.Vancouver. Bridge River Site (EeRl4). Report on file at
Mesoudi, Alex, and Michael J. OBrien. 2008. the National Science Foundation and Bridge
The Cultural Transmission of Great Basin River Band Office, Lillooet, British Columbia
Projectile Point Technology: An Experimental (online version accessible at www.cas.umt
Simulation. American Antiquity 73:328. .edu/grants/bridgeRiver/data/default.php;
Neff, Hector. 1992. Ceramics and Evolution. In accessed October 1, 2014).
Archaeological Method and Theory, Vol. 4, edited Prentiss, Anna Marie, Guy Cross, Thomas A. Foor,
by Michael B. Schiffer, pp. 141193. University Dirk Markle, Mathew Hogan, and David S.
of Arizona Press, Tucson. Clarke. 2008. Evolution of a Late Prehistoric
Neiman, Fraser D. 1995. Stylistic Variation in Winter Village on the Interior Plateau of
Evolutionary Perspective: Inferences from British Columbia: Geophysical Investigations,
Decorative Diversity and Interassemblage Radiocarbon Dating, and Spatial Analysis of the
Distance in Illinois Woodland Ceramic Bridge River Site. American Antiquity 73: 5982.
Assemblages. American Antiquity 60:736. Prentiss,Anna Marie,Thomas A. Foor, Guy Cross,
OBrien, Michael J., ed. 2008. CulturalTransmission Lucille E. Harris, and Michael Wanzenried.
and Archaeology: Issues and Case Studies. SAA 2012.The Cultural Evolution of Wealth-Based
Press, Washington,DC. Inequality at Bridge River, British Columbia.
OBrien, Michael J.,Thomas D. Holland, Robert American Antiquity 77:542564.
J. Hoard, and G. L. Fox. 1994. Evolutionary Prentiss, Anna Marie, and Michael Lenert. 2008.
Implications of Design and Performance Cultural Stasis and Change in Northern North
Characteristics of Prehistoric Pottery. Journal America: A Macroevolutionary Perspective. In
of Archaeological Method and Theory 1: 259304. Macroevolution in Human Prehistory: Evolutionary
OBrien, Michael J., and R. Lee Lyman. 2000. Theory and Processual Archaeology, edited by
Applying Evolutionary Archaeology. Kluwer Anna Marie Prentiss, Ian Kuijt, and James C.
Academic/Plenum Press, NewYork. Chatters, pp. 235252. Springer, NewYork.
OBrien, Michael, and R. Lee Lyman, eds. Prentiss, Anna Marie, Natasha Lyons, Lucille
2003. Style, Function, Transmission: Evolutionary E. Harris, Melissa R. P. Burns, and Terrence
M. Godin. 2007. The Emergence of Status Shennan, Stephen J., and R. Alexander Bentley.
Inequality in Intermediate Scale Societies: A 2008. Style, Interaction, and Demography
Demographic and Socio-Economic History among the Earliest Farmers of Central Europe.
of the Keatley Creek Site, British Columbia. In Cultural Transmission and Archaeology: Issues
Journal of Anthropological Archaeology 26: and Case Studies, edited by Michael J. OBrien,
299327. pp. 164177. SAA Press, Washington,DC.
Prentiss,William C., and James C. Chatters. 2003. Shennan, Stephen J., and J. R. Wilkinson. 2001.
Cultural Diversification and Decimation in Ceramic Style Change and Neutral Evolution:
the Prehistoric Record. Current Anthropology A Case Study from Neolithic Europe. American
44: 3358. Antiquity 66: 577594.
Prentiss, William C., Michael Lenert, Thomas A. Soltis, J., Robert Boyd, and Peter J. Richerson.
Foor, Nathan B. Goodale, and Trinity Schlegel. 1995. Can Group Functional Behaviors Evolve
2003. Radiocarbon Dating at Keatley Creek: by Cultural group Selection? An Empirical
The Chronology of Occupation at a Complex Test. Current Anthropology 36: 473494.
Hunter-Gatherer Village. American Antiquity Spencer,Charles S.1997.Evolutionary Approaches
68: 719736. in Archaeology. Journal of Archaeological Research
Rosenberg, Michael. 1994. Pattern, Process, and 5: 209264.
Hierarchy in the Evolution of Culture. Journal Spencer, Charles S. 2009. Testing the
of Anthropological Archaeology 13: 307340. Morphogenesist Model of Primary State
Rosenberg, Michael. 2009. Proximate Causation, Formation:The Zapotec Case.In Macroevolution
Group Selection, and the Evolution of in Human Prehistory: Evolutionary Theory and
Hierarchical Human Societies: System, Processual Archaeology, edited by Anna Marie
Process, and Pattern. In Macroevolution in Prentiss, Ian Kuijt, and James C. Chatters, pp.
Human Prehistory: Evolutionary Theory and 133156. Springer, NewYork.
Processual Archaeology, edited by Anna Marie Spencer, Charles S., and Elsa M. Redmond. 2001.
Prentiss, Ian Kuijt, and James C. Chatters, pp. Multilevel Selection and Political Evolution
2325. Springer, NewYork. in the Valley of Oaxaca 500100 B.C. Journal
Shennan, Stephen J. 2002. Genes, Memes, and of Anthropological Archaeology 20: 195229.
Human History: Darwinian Archaeology and Teit, James. 1906. The Lillooet Indians. Memoirs of
Cultural Evolution. Thames and Hudson, the American Museum of Natural History, Jesup
London. North Pacific Expedition 2: 193300.
Index
abundance, 7, 52, 100, 101, 105, 117, 139, 144, 102, 106, 107, 115, 116, 124, 136, 156,
145, 147, 148, 149, 152, 153, 159, 161, 159, 161, 163f9.1, 165, 166, 167, 168,
166, 209, 212, 214,226 169, 173, 195, 240, 250, 251, 253, 254,
allometric, 9, 53,56 256, 260, 261, 262, 265, 268, 270,
allometry, 52, 54, 55, 56, 57, 58,195 271,289
Anderson, 111, 115, 117, 119, 136, 184, 185, bifacial flaking,84
194,195 Binford, 100, 115, 133, 136, 141, 154, 157, 169,
Andrefsky, 3, 9, 10, 13, 15, 20, 22, 23, 25, 26, 170, 239, 243, 249, 250, 255, 256,262
45, 55, 59, 84, 98, 139, 141, 145, 149, Bird, 16, 21, 23, 101, 115, 200, 201, 204, 205,
150, 151, 153, 154, 155, 157, 160, 168, 218, 219, 221, 253,263
186, 192, 194, 195, 218, 239, 240, Blackwater Draw, 19, 70, 71, 72, 72f4.8,
243, 249, 250, 251,288 72t4.1, 73t4.2, 77, 87,90
Argentina, 14, 162, 164f9.2, 168,170 Bleed, 10, 21, 118, 135, 136, 157, 158, 169,
arrow, 7, 17, 38, 39f2.7, 40, 41, 43, 174, 239,250
240,251 Boone, 7, 10, 21, 201,218
artifacts, 3, 4, 5, 7, 9, 11, 18, 30, 33, 35, 38, 49, Bow, 21, 24, 45, 47, 79, 250, 262, 289,290
50, 51, 66, 68, 70, 72, 83, 87, 89, 117, Boyd, 5, 15, 16, 21, 68, 78, 115, 201, 218, 219,
118, 133, 139, 144, 149, 153, 173, 174, 240, 249, 250, 253, 255, 256, 263, 265,
176, 177, 178, 179, 180, 180f10.3, 266, 268, 269, 270, 279, 289,292
181, 184, 186, 187, 190, 191, 192, 193, Brantingham, 3, 10, 13, 14, 21, 108, 111, 115,
194n. 1, 201, 203, 206, 207, 211, 212, 116, 118, 136, 152, 153, 154, 156, 157,
213, 216, 217, 226, 227, 235, 267, 161, 169, 173, 195, 239, 240,250
268, 269, 270, 271, 273, 274,275 Braun, 14, 21, 49,59
Atuel River, 162, 164f9.2,166 Bridge River, 18, 267, 271, 273f15.2,
274f15.3, 276, 277, 277f15.6, 278,
Baja California, 139, 140, 141, 142f8.1,154 279, 280, 281, 281f15.8, 281f15.9,
Bamforth, 7, 10, 15, 21, 25, 95, 95f5.7, 139, 282f15.11, 283f15.12, 283f15.13,
154, 157, 168, 260, 262, 270,289 283f15.14, 283f15.15, 284f15.16,
Barton, 21, 100, 107, 108, 115, 118, 138, 199, 286f15.21, 287, 287f15.22, 288,291
218, 220,290 British Columbia, 115, 267, 268, 271,291
Bar-Yosef, 83, 84, 85, 98, 100, 115, 170,251 Buchanan, 7, 17, 22, 23, 53, 54, 55, 56, 58, 59,
Beck, 3, 4, 5, 6, 12, 13, 14, 21, 24, 53, 59, 69, 70, 78, 186, 191, 195, 240,250
73t4.2, 83, 85, 92, 97, 98, 124, 136,
140, 144, 154, 156, 169, 170, 173, 178, caches, 13, 85, 89, 90, 90f5.4, 91, 95, 96,
192, 195, 239, 250, 253,262 211,213
Behavioral ecology, 10,263 Cavalli-Sforza, 16, 22, 253, 255,263
Bettinger, 3, 4, 10, 12, 14, 15, 17, 21, 22, 43, central place foraging, i, 4, 124, 138,
45, 83, 92, 93, 93f5.6, 98, 100, 101, 139,141
293
294 I ndex
central tendency, 40, 42,102 directional selection, vii, 41, 63, 64f4.3,
chert, 14, 54, 95, 95f5.7, 97, 144, 145, 146f8.3, 65,67
147, 148, 149, 150, 151,152 disruptive selection, 64, 65, 65f4.4
China, 12, 100, 106, 107, 107f6.4, 108, 111, dual inheritance, i, 4, 18,253
114, 115,116 Dunnell, 5, 6, 16, 22, 24, 62, 78, 157,
clade diversity,37 169, 201, 219, 238, 253, 255, 263,
Cladistics, 22, 25, 46, 47, 53, 60, 79, 267,290
238,265
cladogram,6,53 Easter Island, 225,238
Clarke, 50, 59, 218, 220,291 edge angle, ix, 129, 131f7.10,134
Clarkson, 4, 9, 10, 11, 12, 22, 23, 60, 117, 118, Eerkens, 3, 14, 17, 21, 22, 41, 43, 45, 46, 61,
119, 121, 134, 136, 137, 138, 249,251 67, 78, 79, 157, 168, 169, 228, 237,
Clovis, 13, 21, 22, 23, 52, 54, 56, 59, 69, 70, 240, 250, 251, 253, 260, 261, 262,
73t4.2, 77, 78, 80, 83, 85, 86, 86f5.1, 263, 268, 277, 289,290
87, 88f5.2, 89, 90f5.4, 91, 94, 95, 96, el-Khiam, 17, 241, 242, 243, 243f13.3
97, 98, 99, 134, 153, 171, 185, 186, evenness, 147, 148, 149,157
187, 188f10.6, 189, 191, 195, 196, evolutionary approaches, 3, 4, 5, 10, 19, 201,
213,220 253, 254, 255,256
coefficient of variation,245 Evolutionary Archaeology, 21, 23, 24, 25, 47,
Collard, 6, 7, 22, 24, 46, 53, 54, 55, 56, 57, 78, 79, 80, 220, 289, 290,291
59, 69, 70, 78, 195, 240, 255, 265, Evolutionary Ecology, 21, 26, 171, 196, 197,
266,290 220, 221, 265,266
Collins, 21, 84, 87, 88f5.2, 97, 98, 121, 123, evolutionary models, 4, 12, 15,118
125,137 Evolutionary Pattern, 29, 46,239
Conspicuous Consumption, vi,198 evolutionary process, 11, 19, 29, 30, 31, 36,
Costly Signaling, 24, 198, 200, 201, 204, 205, 37, 43f2.10, 45, 240, 268, 269,277
207, 210, 212, 218, 219, 220,221 evolutionary trends,7,8
crested blade,84 experimental, 4, 9, 12, 55, 94, 119, 120f7.1,
Crossland, vi, 267,291 121, 122, 122f7.2, 134, 138, 192, 205,
Cueni, vi,239 249,251
Cultural Evolution, 78, 219, 253, 264, 265, extinction, 37, 38f2.6, 65, 67, 190,268
266, 290, 291,292
Cultural Transmission, 3, 16, 21, 22, 24, 25, Ferris, 4, 14, 20, 139, 144, 150,154
45, 46, 60, 78, 79, 80, 223, 238, 239, fitness, 10, 14, 15, 18, 19, 66, 68, 77, 157, 159,
250, 251, 252, 262, 263, 289, 290, 167, 173, 193, 198, 200, 201, 202, 203,
291,292 204, 205, 206, 206f11.1, 208f11.2,
Culture Change, 25, 47, 169,266 209, 210, 211, 212, 213, 214, 215, 216,
Culture History, v, 5, 25, 27, 47,79 217, 254, 255, 268, 269,270
culture trait, 64f4.2, 64f4.3, 65f4.4, 67f4.5 Fluted Point, 59, 98, 169, 185, 193,195
Fluted Points, 48,196
Dadiwan, 106, 107, 107f6.4, 108, 109f6.5, Foley,6,23
110t6.1, 111, 113, 114, 115,116 Folsom, 20, 21, 52, 56, 57f3.1, 58, 59, 60, 69,
Darwinian archaeology, 49,53 70, 73t4.2, 78, 80, 98, 185, 191, 195,
Darwinian evolution, 31, 36, 62,269 196, 197,290
Darwinian evolutionary,60
descent, 5, 8, 30, 42, 52, 53, 54, 61, 255,264 Garvey, 4, 14, 46, 47, 78, 79, 80, 156,261
Dibble, 13, 22, 60, 118, 133, 136, 137, 139, Gatecliff Shelter,6
154,170 Goodale, 3, 4, 9, 10, 17, 20, 23, 45, 218, 239,
diet breadth, 101, 166,260 240, 241, 251, 267, 288,292
I ndex 295
Goodyear, 9, 23, 157, 158, 170, 178, 185, 186, 157, 166, 169, 170, 178, 195, 198, 204,
195, 239, 241, 241f13.1,251 207, 219, 220, 250, 257, 260, 262,
Grayson, 69, 78, 164,170 264,289
Great Basin, 13, 17, 21, 24, 45, 59, 79, 85, 97,
98, 99, 136, 138, 154, 166, 168, 169, Kelly, 10, 12, 24, 25, 97, 98, 99, 119, 133,
170, 191, 195, 204, 220, 250, 251, 262, 138, 140, 154, 157, 158, 159, 170,
263, 289,291 178, 179, 195, 196, 201, 219, 220,
254,264
haft, 17, 51, 56, 57, 126, 127, 128, 131, 132, Kuhn, 4, 9, 10, 11, 12, 24, 83, 84, 85, 98, 118,
133, 135, 187, 239, 240, 242, 244, 245, 133, 138, 139, 140, 141, 149, 154, 157,
247,249 170, 172, 173, 192, 195, 196, 239, 251,
haft element, 239,241 253,264
hafted, 12, 52, 117, 118, 119, 120, 120t7.1, Kuijt, 23, 25, 218, 239, 241, 251, 270, 289,
121, 122, 123, 124, 126, 127, 128, 129, 290, 291,292
132, 132f7.11, 133, 134, 135, 138, 213,
215,250 landmarks, 54, 88f5.2, 240, 243,244
hafting, 12, 52, 74, 117, 118, 119, 122, 129, land-use, 3, 14,141
131, 132, 133, 135, 187, 226, 228,276 Levant, 26, 83, 115, 241, 242,
Hamilton, 7, 17, 22, 23, 54, 59, 69, 78, 83, 242f13.2,251
195, 239, 240, 250,267 life history,8
Harris, 117, 254, 255, 263, 267,291 Lipo, 5, 17, 22, 24, 41, 46, 54, 59, 61, 63,
Haslam, vi, 117,137 67, 78, 79, 225, 226, 228, 232, 236,
Hawkes, 23, 176, 196, 200, 204,219 238, 249, 255, 260, 263, 265, 268,
Hayden, 84, 99, 118, 119, 127, 128, 132, 137, 277,290
155, 169, 279, 286,290 Lithic Analysis, vi, 23, 25, 60, 154, 198,
Hiscock, 9, 13, 23, 118, 121, 133, 137,138 205,252
histogram,30 Lithic Raw Material, vi, 154, 156,170
history, 5, 8, 11, 18, 19, 20, 33, 37, 38, 44, 52, lithic studies, 4, 5, 8, 20, 140, 156, 157,
53, 55, 177, 180, 240, 241, 249, 254, 176,249
268, 270, 271, 277, 278,288 Lithic Technology, 3, 16, 20, 21, 23, 24, 25,
Holocene, 14, 59, 69, 77, 98, 115, 116, 137, 26, 59, 60, 98, 99, 108, 137, 138, 153,
144, 153, 164, 165, 166, 167, 170, 190, 155, 172, 195, 196, 197, 210, 212, 250,
195, 259, 259f14.4, 260, 261, 262, 251,264
263, 264, 265, 270, 271, 289,291 low ranked,101
housepit, x, 268, 272, 274f15.3, 276, 281,286 Lyman, 5, 6, 7, 8, 9, 24, 25, 29, 30, 32, 35, 37,
Huckell, 86, 90, 99 38, 40, 42, 44, 46, 47, 49, 53, 60, 61,
Hughes, 6, 15, 23, 24, 33, 34, 34t2.1, 39, 62, 68, 72, 74, 79, 201, 220, 238, 240,
46, 52, 59, 169, 170, 215, 218, 220, 251, 265, 267, 268, 269, 271, 278,
253,264 290,291
Human Behavioral Ecology, 10, 155,
172,266 Macroevolution, 47, 220, 269, 289, 290,
Hundtoft, vi, 225,237 291,292
hunt, 69, 177,205 manufacturing costs, 12, 14, 94, 135, 159, 160,
163f9.1
industry, 12, 18, 86, 251, 276, 279,287 manufacturing time, 12, 102, 103, 104,
Ishi, x, 17, 246, 247, 248f13.8, 250,251 104f6.2, 105, 105f6.3, 124, 125,
125f7.5, 127f7.7, 161, 163f9.1
Jones, 4, 5, 13, 14, 20, 21, 23, 24, 53, 59, 83, Marginal Value Theorem, 11, 22, 172, 174,
85, 97, 98, 124, 136, 137, 138, 154, 156, 191, 195, 196,197
296 I ndex
mataa, 227f12.2, 228, 229f12.3, 232f12.6, Paleoindian, 4, 7, 11, 13, 19, 21, 22, 24, 25, 26,
233f12.7, 233f12.8, 234f12.9, 52, 57, 59, 60, 61, 69, 70, 70f4.7, 71, 72,
234f12.10, 235f12.11, 237, 72t4.1, 73, 73t4.2, 75t4.4, 77, 77f4.10,
236f12.12,249 78, 83, 85, 88, 91, 98, 99, 158, 170, 174,
Meltzer, 6, 16, 24, 60, 91, 95, 96, 97, 99, 164, 178, 179, 184, 185, 186, 189, 190, 191,
170, 191,195 194, 194n. 1, 195, 196, 197,213
Mendoza, 14, 162, 164f9.2, 165, 166, patches, 11, 165, 166, 175, 175f10.1, 176, 177,
167,170 178, 181,184
Mesoudi, 3, 24, 68, 79, 240, 251, 268, phenotype, 172, 201,271
271,291 phenotypic, 5, 10, 30, 66, 173, 201, 267,269
microblade, 106, 107, 108, 111, 113, 113f6.9, phylogenetic analysis, i, 4, 5, 6, 7, 8,9,19
114,116 Phylogenetic Evolution, v,5,27
Microevolution, 220,268 phylogenetic tree,8,29
Microlithic, 106, 107, 108, 110t6.1, 111, 113, Prentiss, 18, 267, 269, 270, 271, 272, 274,
114, 115,116 275, 278, 282, 286, 288, 289, 290,
Miller, 4, 9, 11, 12, 172, 194,264 291,292
millet, 12, 100, 101, 106, 107, 107f6.4, prismatic blade, 84, 85, 87, 89, 94,173
110,114 procurement, 12, 13, 14, 16, 92, 94, 101,
Morgan, v,100 102, 104, 104f6.2, 105, 105f6.3, 134,
morphology, 6, 9, 63, 65, 69, 71, 121, 138, 135, 136, 139, 140, 141, 144, 147, 151,
147, 240, 247, 249,261 152, 153, 156, 157, 158, 159, 160, 161,
Morphometric, 48, 55, 59, 99, 195, 240, 162, 163f9.1, 166, 167, 168, 170, 178,
243,257 199,256
Morrow, 25, 56, 60, 99, 118,138 Procurement Costs, vi, 139,141
procurement rate, 12, 102,105
nanolithic, 12,113 production, 7, 8, 10, 13, 14, 15, 18, 44, 51, 52,
natural selection, 6, 10, 11, 19, 31, 63, 66, 54, 84, 86, 95, 100, 113, 114, 139, 144,
159, 172, 201,254 149, 152, 153, 160, 173, 174, 178, 180,
neck width, 7, 44, 159, 243, 244, 181, 198, 199, 204, 205, 209, 210, 211,
244f13.4 212, 213, 214, 228, 241, 250, 277, 281,
Neiman, 5, 16, 24, 25, 42, 47, 53, 60, 61, 79, 282, 287,288
204, 220, 228, 238, 269, 279,291 projectile points, 8, 9, 11, 17, 33, 34, 34t2.1,
Nelson, 118, 119, 133, 138, 139, 141, 155, 36f2.4, 37, 38, 38f2.6, 39, 40, 62,
172,196 70f4.7, 71, 72f4.8, 72t4.1, 77, 87, 97,
Neolithic, 17, 23, 25, 112, 116, 241, 242, 148, 174, 177, 179, 191, 192, 239, 240,
242f13.2, 245, 246, 251, 252,292 243, 249,250
notched points, 240, 241, 244, 246,247 Provenance, 10,20
OBrien, 3, 5, 6, 7, 22, 24, 25, 30, 32, 37, 44, quartzite, 144, 145, 146f8.3, 147, 148, 149,
45, 46, 47, 49, 53, 60, 61, 62, 68, 72, 78, 150, 151, 152,161
79, 80, 195, 201, 220, 229, 238, 240, Quinn, vi, 16, 25, 198, 200, 201, 221, 240,
251, 252, 253, 255, 261, 265, 267, 268, 241, 243f13.3, 249,251
269, 271, 290, 291,292
OConnell, 10, 25, 176, 193, 196, 253,263 Rapa Nui, 225, 226, 226f12.1, 227, 227f12.2,
Obsidian, 25, 169, 170, 215, 216, 221,225 229, 229f12.3, 232, 236, 237, 236f12.12,
optimal foraging, 12, 13, 15, 101, 140, 156, 238,249
168,202 raw material acquisition, 239,240
Optimal number, 182f10.4, 183f10.5 raw material availability, 13, 97, 118,
Osterhoudt, vi, 23, 239,251 192,199
I ndex 297
raw material procurement,14 132, 136, 137, 138, 152, 174, 181, 191,
raw material quality, 14, 160, 161,240 192, 198, 199, 211, 240, 249, 251,
regression, 40, 56, 258f14.3 259,260
resharpen, 126,127 stratigraphy, 4, 35, 36f2.4, 37, 37f2.5, 38,
retouch, 4, 12, 25, 118, 131, 134, 135, 150, 43,272
199, 212, 213, 214, 242, 249, 250,251 Stylistic, vi, 25, 47, 60, 79, 221, 225, 238,291
rhyolite, 144, 145, 146f8.3, 147, 148, 149, 150, Surovell, 3, 11, 26, 156, 171, 177, 178, 192,
151, 152,153 193, 197,221
Richerson, 115, 201, 218, 240, 249, 250, 253,
256, 260, 263, 265, 266, 268, 270, technological organization, 11, 17, 138, 172,
279, 289,292 173, 174, 191, 193, 198, 199, 205, 207,
risk, 4, 10, 13, 51, 84, 118, 173,195 209, 210,217
Technological Tradition,255
selection, 4, 10, 12, 18, 19, 61, 62, 63, 64f4.2, Toolstone, 21, 98, 139, 141, 147, 154,195
65, 65f4.4, 66, 67f4.5, 68, 68f4.6, 77, Torrence, 10, 23, 24, 26, 133, 138, 154, 157,
98, 131, 137, 139, 140, 141, 144, 147, 158, 170, 219, 253,266
151, 152, 153, 162, 167, 172, 201, 206, transmission, 4, 6, 16, 17, 18, 19, 33, 35,
212, 228, 254, 267, 271,287 37f2.5, 38, 40, 41, 42, 43, 52, 53, 61,
Sellet,5,25 62, 63, 67, 74, 76, 201, 226, 229, 232,
seriation, 17, 38, 106, 229, 231, 232, 233, 236, 233, 236, 237, 240, 241,
236f12.12,249 249, 251, 255, 264, 268, 270, 271,
Shennan, 7, 17, 22, 24, 25, 46, 59, 61, 62, 68, 277, 279,287
78, 79, 80, 149, 155, 173, 196, 255, 264, Transport Costs, 21, 98, 154, 169,195
265, 266, 268, 269, 278, 279, 290,292 Typological,24
Shott, 4, 8, 9, 10, 20, 25, 48, 51, 53, 55, 56,
57f3.1, 58, 60, 118, 128, 138, 141, Ugan, 12, 26, 91, 92, 93, 99, 102, 116, 118,
155, 179, 185, 186, 187, 197, 240, 252, 136, 138, 159, 171, 173, 195, 197, 253,
261,265 261,266
shoulder angle, 230, 231, 232f12.6, unretouched, 12, 117, 118, 119, 121, 123,
233f12.7,232 124, 125, 127, 128, 129, 130f7.8, 131,
signaling, 4, 16, 198, 200, 201, 202, 203, 204, 131f7.10, 132, 132f7.11, 133, 134, 135,
205, 206, 207, 208f11.2, 208t11.1, use time, 14, 92, 93, 114, 125, 159, 160, 161,
209, 210, 211, 212, 214, 216,217 162, 163f9.1,166
Sillitoe, 118, 128, 138, 179, 197, 261,265 use wear, 121, 129,134
slate, 18, 268, 276, 277, 277f15.6, 279,
280, 281, 281f15.7, 282, 282f15.10, VanPool, 4, 13, 19, 24, 38, 45, 47, 61, 63, 65, 78,
282f15.11, 283f15.13, 283f15.14, 284, 79, 80, 201, 220, 221, 251,290
284f15.17, 285f15.19, 286, 287,288 Variation, 21, 22, 23, 24, 25, 46, 47, 57, 59,
Social Information Exchange, vi, 198,199 60, 62, 66, 69, 71, 78, 79, 80, 98, 99,
southwest, 19,138 152, 221, 251, 263, 290,291
Spatial distributions, x, 236f12.12
stemmed, 13, 70, 85, 89, 226, 235,257 Western Stemmed, viii, 21, 85, 86f5.1,
Stevens, 4, 18, 19, 253, 259, 260, 264,266 89,98
Steward, 18, 29, 47, 253, 254, 255, 256, 260, Winterhalder, 10, 21, 26, 98, 115, 140, 155,
261,266 159, 167, 168, 169, 171, 173, 195, 196,
stimulated variation, 19, 61, 62, 63f4.1, 68, 197, 201, 221, 253, 254, 261, 262,
68f4.6, 69, 70, 70f4.7, 74,76 265,266
stone tools, 6, 8, 11, 13, 16, 17, 18, 19, 25, 49, work organization, 18, 255, 256, 257, 257f14.1,
50, 51, 52, 54, 55, 58, 117, 119, 128, 260, 261

Nathan Goodale &amp; William Andrefsky 2015

Diunggah oleh

Informasi Dokumen

Hak Cipta

Format Tersedia

Bagikan dokumen Ini

Bagikan atau Tanam Dokumen

Opsi Berbagi

Apakah menurut Anda dokumen ini bermanfaat?

Apakah konten ini tidak pantas?

Hak Cipta:

Format Tersedia

Nathan Goodale &amp; William Andrefsky 2015

Diunggah oleh

Hak Cipta:

Format Tersedia

Lithic Technological Systems and

Stone tool analysis relies on a strong background in analytical and meth-

Nathan Goodale is Assistant Professor of Anthropology at Hamilton College.

William Andrefsky, Jr., is Edward R. Meyer Distinguished Professor of

William Andrefsky, Jr.

Cambridge University Press is part of the University of Cambridge.

List of Figures and Tables page vii

Part I Lithic Technological Systems and

Part IICulture History and Phylogenetic Evolution

Part IIIApplications of Behavioral Ecology to

7 When to Retouch, Haft, or Discard?

Part IVCultural Transmission and Morphology

4.7. Expectations of the model of initial stimulated variation and

7.8. Mean cumulative weight of wood removed per 1000 strokes

12.6. Seriation solution for mataa classes comprised of stem length/

William Andrefsky, Jr.

This volume is an outgrowth of a symposium organized for the 74th Annual

Lithic Technological Systems and

Interpreting Lithic Technology

William Andrefsky, Jr., and Nathan Goodale

An underlying theme of this volume is that lithic technological analysis is not

evolutionary approaches and lithic technological systems, the amount of

Culture History, Lithic Data, and Phylogenetic Evolution

it is the differential representation of variation at all scales among artifacts for

to Paleoindian projectile technology from the southeastern United States to

archaeologys essentially universal reliance on aggregate data sets that represent

Human Behavioral Ecology, Tool Use-Life,

temperature, than by a protracted explanation involving the evolution

Garvey (Chapter 9) also adopts optimality modeling for explanations of

and cortex condition. They also attempted various geochemical techniques to

Lithic Technology, Neutral Variables, and Cultural

of stylistic characters (Neiman 1995:31). His work with cultural transmission

also emphasizes the importance of considering the measurements analysts take

and reproductive success in individuals who use stone to make a living. As we

Distance in Illinois Woodland Ceramic Measures of Production Use and Curation,

Culture History and Phylogenetic

Graphing Evolutionary Pattern

An old clich is that a picture is worth a thousand words. In disciplines that

(1936) misinterpretation of the taxonomic architectural patterns of particular

Variational (Darwinian) model

pointed out that our chronologies of cultures cause us to think of time as a

table2.1. Frequencies of projectile points used in analyses

Stratum No. of points Age (radiocarbon years BP)

in the identities of species in a lineage or clade are the subject of concern. At

At the Scale of Discrete Object

the discontinuity could represent immigration, a standard culture-historical

50 100 150 200 250 50 100 150 200 250

At the Scale of Attributes of Discrete Objects

again an evolutionary connection is neither shown nor implied. Instead what

generate more than normal variation to find an effective point to serve to

Length Width Thickness Neck Width Base Length

Consider two aquatic metaphors. First, imagine a horizontal cylinder of ice.

Anthropology as Archaeologys Place?

Time Views in Archaeology

A Place Where Archaeology ShouldBe

Theory of Change in Material Culture

New Approaches: History from Artifacts

History from Points

Clovis. In this state of ignorance, archaeologists have begun to test methods to

strict branching divergence, can overestimate rate of branching when terminal

Allometry in Cladistic and Morphometric Studies

Morphometrics at the TypeLevel

2.20 2.30 2.40 2.50 2.60 2.70 2.80

from a series of discrete types to a complex, continuous process of change in

Nathan Goodale & William Andrefsky 2015

Nathan Goodale & William Andrefsky 2015