University of Iowa
Iowa Research Online
Theses and Dissertations
Summer 2016
Human brain activity during stone tool production
: tracing the evolution of cognition and language
Shelby Stackhouse Putt
University of Iowa
Copyright 2016 Shelby Stackhouse Putt
This dissertation is available at Iowa Research Online: http://ir.uiowa.edu/etd/2133
Recommended Citation
Putt, Shelby Stackhouse. "Human brain activity during stone tool production : tracing the evolution of cognition and language." PhD
(Doctor of Philosophy) thesis, University of Iowa, 2016.
http://ir.uiowa.edu/etd/2133.
Follow this and additional works at: http://ir.uiowa.edu/etd
Part of the Anthropology Commons
HUMAN BRAIN ACTIVITY DURING STONE TOOL PRODUCTION:
TRACING THE EVOLUTION OF COGNITION AND LANGUAGE
by
Shelby Stackhouse Putt
A thesis submitted in partial fulfillment
of the requirements for the Doctor of Philosophy
degree in Anthropology in the
Graduate College of
The University of Iowa
August 2016
Thesis Supervisor: Professor Robert G. Franciscus
Copyright by
SHELBY STACKHOUSE PUTT
2016
All Rights Reserved
Graduate College
The University of Iowa
Iowa City, Iowa
CERTIFICATE OF APPROVAL
____________________________
PH.D. THESIS
_________________
This is to certify that the Ph.D. thesis of
Shelby Stackhouse Putt
has been approved by the Examining Committee for
the thesis requirement for the Doctor of Philosophy degree
in Anthropology at the August 2016 graduation.
Thesis Committee:
____________________________________________
Robert G. Franciscus, Thesis Supervisor
____________________________________________
John P. Spencer
____________________________________________
James G. Enloe
____________________________________________
Daniel T. Tranel
____________________________________________
Nicholas P. Toth
____________________________________________
Andrew Kitchen
For Dr. Robert Allen Mahaney
In the beginning was the word.
But if unheard…
This is a crazed interdependence,
so many, many mirrors faced towards each other
reflecting each in each other
forever.
Who knows where the past began
and where it ends?
There is nothing here
and there is everything here.
I am…
…I am here
somewhere.
ii
ACKNOWLEDGEMENTS
I have always found that it is prose in which I am best able to express myself, and
so it is my hope that this string of words may convey the depth of gratitude I feel for the
following individuals and organizations that have been instrumental in the many stages
leading up to the submission of this dissertation. I am most grateful to my advisor, Robert
Franciscus, and committee member and mentor, John Spencer, for their open-mindedness
toward interdisciplinarity above all else. It was Bob’s assurances and willingness to try
something completely new and different that led to the design and realization of this
project. And the project would have ended as soon as it started had John not immersed
me, an oft-bewildered anthropologist, into his neuroscience lab group and graciously
offered the Child Imaging Laboratory in Developmental Science (CHILDS) for my
research purposes.
I owe a huge thanks to the brilliant Sobanawartiny Wijeakumar, who guided me
through much of the nuts and bolts of processing and analyzing NIRS imaging data,
which was no easy feat. I am also grateful to the other members of the University of Iowa
CHILDS lab group for their assistance on different aspects of the project: Bryan Brown,
Nicholas Fox, and Lourdes Delgado Reyes.
There were many others whose contributions aided in the completion of this
dissertation. Alexander Woods served as my expert flintknapper for the instruction videos
that were shown to the study participants. The other members of my committee,
including Daniel Tranel, Nicholas Toth, James Enloe, Russell Ciochon, and Andrew
Kitchen, provided helpful feedback on my prospectus, comprehensive exams, and
iii
dissertation. Thomas Wynn also took the time to read and make suggestions for Chapter
2. Many students devoted their time to assisting with data collection and ensuring that the
study ran smoothly: Danielle Jones, Chloe Daniel, Elizabeth DeForest, Andriana Vega,
Emma Dellopolous, Anna Wells, Emily Hoeper, Graham Brua, Sean Allchin, and
Madison Adams.
I was very fortunate that the Wenner-Gren Foundation, Leakey Foundation,
Sigma Xi, The Scientific Research Society, and the Graduate and Professional Student
Government at the University of Iowa saw the value of this project and provided me with
the funding that paid for all the expenses I accrued while conducting my doctoral
research. I also held an American Fellowship from the American Association of
University Women (AAUW) during my final writing year. I am grateful to Beverly
Poduska and Shari Knight for carrying out the administrative work that came with these
grants.
I am very grateful to those fellow graduate students whose advice and friendship
helped me find my way to the end of this program with a sound mind. There were many,
but I want to especially thank Christina Nicholas and Meredith Wismer. Finally, none of
this would have been possible without the love and support of my parents, Rhonda and
Mark Putt. I hope this makes you proud.
iv
ABSTRACT
This study aims to shed light on how and when mechanisms of the human brain
evolved to support complex cognition and language. The field of evolutionary cognitive
archaeology asserts that prehistoric technologies, as products of past cognition in action,
are informative of the minimum cognitive and linguistic abilities that hominins needed to
possess for their production. Previous researchers attempted to reconstruct the neural
correlates of two Early Stone Age (ESA) tool industries, the 2.6 million-year-old
Oldowan industry and the 0.7 million-year-old late Acheulian industry, by using positron
emission tomography (PET) to observe the functional activation occurring in the brains of
trained and expert stone knappers after making these different tool types. Because of
evidence for overlap between the knapping and language circuits of the brain and
increased anterior frontal activity during Acheulian tool production, these researchers
argued that their results 1) indicate increased cognitive demands for late Acheulian tool
production relative to Oldowan tool production and 2) support a technological origin for
language, meaning that certain language functions co-opted the neural substrate and
functions that were already established for toolmaking and tool use. Because of the
motion limiting aspects of PET, however, these studies were unable to record the
hemodynamic response of naturalistic stone knapping in real-time. They also were unable
to observe the functional activation associated with the earliest stage of learning, which is
likely to differ from late stage learning or expertise. Furthermore, any conclusion
regarding a technological origin for language is problematic if it relies on data obtained
from participants who learned to knap with verbal instruction.
v
To test these two claims, this dissertation utilized a neuroimaging technique called
functional near-infrared spectroscopy (fNIRS) to explore the neural correlates of realtime, naturalistic Oldowan and Acheulian stone knapping at three different points in
learning. Participants in the study were separated into two groups to learn ESA knapping
skills. Both groups watched the same video tutorials that depicted an expert’s hands as he
made stone tools, but those in the verbal group heard spoken instructions, while those in
the nonverbal group watched a version with the sound turned off. Functional brain images
were reconstructed from the digitized landmarks of each participant’s head and from the
optical data. An analysis of variance (ANOVA) revealed a clearer distinction between the
neural processes of Oldowan and Acheulian tool manufacturing tasks than has previously
been demonstrated. Only the Acheulian task recruited a frontotemporal working memory
network. Selection for individuals with increased working memory capacities, which
would have allowed them to make increasingly complex tools to gain access to novel
dietary items, may have spurred the evolution of larger brain size in the genus Homo
during the early Pleistocene. The results also demonstrated that the presence or absence
of language during training dictated which higher-order cognitive areas of the brain
become engaged and at what point in training. Thus, the results of previous
neuroarchaeological studies reflect a very specific condition of stone knapping skill
acquisition that involves linguistic instruction, which may not be analogous to how skills
were transmitted during the ESA. Finally, evidence of overlap between left hemisphere
language and stone knapping circuits among the participants in the nonverbal group lends
additional support for the technological origin for language hypothesis.
vi
PUBLIC ABSTRACT
Some of the biggest questions in human evolution are why we have such large
brains and how our ancestors acquired language and exceptional intelligence. Our
extreme reliance on technology has set us humans and our ancestors apart from other
primates for more than three million years. It is widely thought that tools from the distant
past may hold the clues to answering these questions because they represent all that is left
of ancient minds at work. This study addresses these questions by using brain imaging
technology to determine which areas of the brain of modern-day humans become most
active as they make two types of tools from the past, one from as early as 2.6 million
years ago (Ma) known as the Oldowan industry, and the other from 1.75 Ma known as
the early Acheulian industry. Because it remains unclear whether early humans possessed
language this far back in the past, an instructor taught half of the participants in this study
to make stone tools with language, while the other half learned by nonverbal imitation.
The analysis of the resulting brain imaging data revealed that Acheulian toolmaking
requires higher-order conceptualization than Oldowan toolmaking. Selection for
individuals who could store and manipulate more information and therefore make the
most productive Acheulian tools may have been the prime reason for the evolution of
large brain size in humans. The complex cognition that evolved as a result of such
technology likely provided the framework on which language could build.
vii
TABLE OF CONTENTS
LIST OF TABLES ............................................................................................................ xii
LIST OF FIGURES ......................................................................................................... xiii
LIST OF ABBREVIATIONS .......................................................................................... xvi
CHAPTER 1: INTRODUCTION ........................................................................................1
Rationale for research ...................................................................................................1
Objectives .....................................................................................................................8
Organization of the dissertation ..................................................................................12
Summary .....................................................................................................................15
CHAPTER 2: EVOLUTIONARY COGNITIVE ARCHAEOLOGY: A
HISTORY ..........................................................................................................................18
Introduction.................................................................................................................18
“Fossil intelligence” and “petrified mind” .................................................................19
Towards a hypothetico-deductive approach ...............................................................24
The application of Piagetian theory to the archaeological record ..............................34
“It is time for a ‘cognitive archaeology’”: The search for the hominin mind ............39
A shifting emphasis to the brain and cognition ..........................................................53
Controversies continued in the new century ...............................................................58
How to move past the mistakes of the past.................................................................70
Summary .....................................................................................................................76
CHAPTER 3: THE HUMAN WORKING MEMORY SYSTEM AND ITS
EVOLUTION ....................................................................................................................78
Introduction.................................................................................................................78
viii
Background on working memory as a neural system .................................................80
A comparative approach to the evolution of working memory ..................................81
A developmental approach to the evolution of working memory ..............................87
An archaeological approach to the evolution of working memory ............................92
Summary .....................................................................................................................96
CHAPTER 4: THE CONTRIBUTIONS OF NEUROARCHAEOLOGY TO
HUMAN LANGUAGE ORIGIN THEORIES ..................................................................99
Introduction.................................................................................................................99
Clarifications.............................................................................................................100
A brief survey of the multidisciplinary approach to language evolution .................102
Neuroarchaeology: The wedding of neuroscience and archaeology ........................133
Summary ...................................................................................................................152
CHAPTER 5: MATERIALS AND METHODS .............................................................154
Testable hypotheses ..................................................................................................154
fNIRS as a neuroimaging technique .........................................................................155
Experimental design .................................................................................................160
Experimental methods and procedures .....................................................................162
Data acquisition and processing ...............................................................................183
Data analysis and interpretation................................................................................189
Summary ...................................................................................................................192
CHAPTER 6: THE NEURAL CORRELATES OF THE
OLDOWAN-ACHEULIAN TRANSITION ...................................................................193
Introduction...............................................................................................................193
ix
Results.......................................................................................................................198
Discussion .................................................................................................................206
Summary ...................................................................................................................209
CHAPTER 7: THE FUNCTIONAL NEUROANATOMY OF LEARNING TO
MAKE EARLY STONE AGE TOOLS ..........................................................................211
Introduction...............................................................................................................211
Summary of methods ................................................................................................214
Results.......................................................................................................................218
Discussion .................................................................................................................232
Summary ...................................................................................................................241
CHAPTER 8: A TECHNOLOGICAL EXPLANATION FOR HOMININ BRAIN
EXPANSION IN THE EARLY PLEISTOCENE ...........................................................243
Introduction...............................................................................................................243
The working memory hypothesis for hominin brain expansion ...............................244
Summary ...................................................................................................................253
CHAPTER 9: DISCUSSION AND CONCLUSIONS ....................................................254
Introduction...............................................................................................................254
Findings related to the evolution of hominin cognition............................................256
Findings related to ESA skill transmission and the evolution of human language ..277
Conclusions...............................................................................................................290
APPENDIX A: SUBJECT SELECTION MATERIALS ................................................294
Eligibility questionnaire and Benton Neuropsychology Clinic Handedness test .....294
Psychiatric, neurologic, and drug screening questionnaire ......................................297
x
APPENDIX B: SUPPLEMENTARY METHODS AND RESULTS .............................301
Supplementary materials for Chapter 5 ....................................................................301
Supplementary materials for Chapter 6 ....................................................................305
Supplementary materials for Chapter 7 ....................................................................311
APPENDIX C: SUBJECT INTERVIEWS......................................................................312
REFERENCES ................................................................................................................314
xi
LIST OF TABLES
Table 1. Holloway's (1969) application of Greenberg (1967) and Hockett's (1960)
structure and design features of language to the process of stone tool manufacture. 32
Table 2. Localization of activated parietal, frontal, and temporal cortical clusters
during the execution of stone knapping tasks. ......................................................... 141
Table 3. Summary of task design during neuroimaging sessions. .................................. 162
Table 4. Score system for the Placing Test ..................................................................... 165
Table 5. Source-detector distances corresponding with adult-sized caps ....................... 171
Table 6. Summary of behavioral measures for knapping skill determination during
the third neuroimaging session ................................................................................ 200
Table 7. Regions of significant activation (p < 0.05) as determined by a two-way
ANOVA between Group and Task .......................................................................... 202
Table 8. Total number of significant voxels with a Session effect for each baseline
condition .................................................................................................................. 217
Table 9. Knapping areas active at different points in skill learning ............................... 219
Table A1. Functional near-infrared spectroscopy motion processing
parameters...………………………………………..……………...…………..........301
Table A2. General motor areas involved in knapping ESA tools at different points in
skill learning.........…...…...…...…...…...………………...…...…...……................311
xii
LIST OF FIGURES
Figure 1. Functional subregions of Broca’s area with a posterior-anterior gradient
of abstraction for processing language and tool use. ............................................... 147
Figure 2. A comparison of the mean ratio of flake size to flake mass between the
verbal and nonverbal groups from Putt et al. (2014). .............................................. 150
Figure 3. Median asymmetry index scores for the handaxes produced by the verbal
and nonverbal groups during the fourth and fifth practice sessions from Putt et
al. (2014).................................................................................................................. 151
Figure 4. A schematic representation of light in the NIR range traveling through the
head.......................................................................................................................... 158
Figure 5. Photograph of a subject knapping during a practice session while wearing
safety equipment. ..................................................................................................... 167
Figure 6. Flowchart demonstrating the steps involved in creating and visualizing the
optode geometry design. .......................................................................................... 170
Figure 7. Optode geometry design and digitized points from the same optode
geometry registered onto an adult atlas head........................................................... 171
Figure 8. Sensitivity distributions of left hemisphere channels generated by
Monte Carlo photon migration simulations using the digitized points from the
optode geometry registered to an adult atlas. .......................................................... 174
Figure 9. Coverage of the ROIs in the cerebral cortex with the optode geometry
designed for this experiment.................................................................................... 174
Figure 10. Demonstration of neuroimaging session set-up with fNIRS. ........................ 179
Figure 11. Motor baseline task movements. ................................................................... 181
xiii
Figure 12. A staged demonstration of fNIRS data being collected during a knapping
task. .......................................................................................................................... 183
Figure 13. Oxy-Hb and deoxy-Hb concentration levels corresponding to three
channels in the left frontal cortex during a knapping task ....................................... 185
Figure 14. Digitization of head landmarks and source/detector positions on the head
using the Polhemus Patriot device. .......................................................................... 187
Figure 15. Representative sketches of ESA core artifacts. ............................................. 194
Figure 16. Areas where functional overlap occurs between ESA knapping, languageprocessing, and/or VWM......................................................................................... 196
Figure 17. Coverage of key areas in the frontal, parietal, and temporal cortex with
fNIRS. ...................................................................................................................... 198
Figure 18. Active clusters where an ANOVA task main effect or interaction effect
occurs and its relative spatial relationship to the results of a previous
neuroarchaeological study ....................................................................................... 203
Figure 19. Acheulian activation in the dlPFC revealed in the Task main effect and
these clusters' spatial relationship to language centers, VWM areas of the brain,
and the results of a previous neuroarchaeological experiment ................................ 221
Figure 20. Learning context activation differences as revealed by the Group main
effect and their spatial relationship to language centers of the brain ...................... 223
Figure 21. Significant clusters with an interaction between Group and Task and these
clusters' spatial relationship to VWM areas of the brain and the results of a
previous neuroarchaeological experiment ............................................................... 224
xiv
Figure 22. Left PrG cluster involved in early-stage learning of ESA tool
manufacture.. ........................................................................................................... 225
Figure 23. Significant clusters with an interaction between Task and Session and
their spatial relationship to the results of a previous neuroarchaeological
experiment ............................................................................................................... 227
Figure 24. Significant clusters with an interaction between Group and Session and
these clusters' spatial relationship to language centers of the brain. ....................... 229
Figure 25. Significant clusters with an interaction between Group, Task, and Session
and their spatial relationship to the results of a previous neuroarchaeological
experiment, language centers, and VWM areas of the brain ................................... 231
Figure 26. Language-processing areas that are active during lithic reduction tasks. ..... 282
Figure A1. Subject responses by group to the question, “Did you think with
language while knapping?”…………………………………………………...……312
Figure A2. Percentage of subjects to recognize different goals in silent instruction
videos for Oldowan and Acheulian tasks across neuroimaging sessions………….312
Figure A3. Percentage of subjects to accurately describe different goals in silent
instruction videos for Oldowan and Acheulian task……………………………….313
xv
LIST OF ABBREVIATIONS
AFNI
Analysis of Functional NeuroImages
ANOVA
analysis of variance
BA
Brodmann area
cm
centimeters
DAST-10
Drug Abuse Screen Test
deoxy-Hb
deoxygenated hemoglobin
dlPFC
dorsolateral prefrontal cortex
DTI
diffusion tensor imaging
EEG
electroencephalography
ESA
Early Stone Age
FEF
frontal eye fields
fMRI
functional magnetic resonance imaging
fNIRS
functional near-infrared spectroscopy
GLM
general linear model
h
hours
HRF
hemodynamic response function
Hz
Hertz
IFG
inferior frontal gyrus
IPL
inferior parietal lobule
Ka
thousand years ago
LP
Lower Paleolithic
LSA
Late Stone Age
xvi
Ma
million years ago
MEG
magnetoencephalography
MFG
middle frontal gyrus
MTG
middle temporal gyrus
MMDT
Minnesota Manual Dexterity Test
MNI
Montreal Neurological Institute
MP
Middle Paleolithic
MSA
Middle Stone Age
MSH
Mirror System Hypothesis
NIR
near-infrared
nm
nanometers
OD
optical density
oxy-Hb
oxygenated hemoglobin
PET
positron emission tomography
PoG
postcentral gyrus
PrG
precentral gyrus
ROI
region of interest
SMA
supplementary motor area
SMG
supramarginal gyrus
SPL
superior parietal lobule
STG
superior temporal gyrus
TMS
transcranial magnetic stimulation
UP
Upper Paleolithic
xvii
tPCA
targeted principal components analysis
vlPFC
ventrolateral prefrontal cortex
xviii
CHAPTER 1:
INTRODUCTION
Rationale for research
Our species, modern Homo sapiens, is characterized by cumulative culture (Hill
et al. 2011), which is made possible by our universal capacity for complex cognition and
a sophisticated language system. Perhaps the most obvious physical evidence for
complex culture in today’s world is technology. While tool use, toolmaking, and simple
traditions have been documented in other apes (Boesch 2012), only humans rely on
culture and technology to the extent that they govern nearly every aspect of an
individual’s day-to-day life. There is still little known about how the hominin1 brain
evolved to allow for complex cognition and language, what selective pressures
contributed to its evolution, and the timing of these events. Answering these questions is
critical to understanding and defining what it means to be human.
Just as technology is a clear manifestation of human culture and, by extension,
language and cognition in the present-day, its ubiquity in the archaeological record in the
form of stone tools that date as far back in time as 3.3 million years ago (Ma) may also be
helpful for exploring the evolution of these important features in the past. The earliest
evidence for hominin technology, known as the Lomekwian industry, predates the
appearance of the genus Homo and so far has only been found at West Turkana, Kenya
1
The term ‘hominin’ is used to refer to all habitually bipedal apes that existed after the
lineage that would eventually lead to H. sapiens diverged from a common ancestor
shared with extant Pan species.
1
(Harmand et al. 2015, but see Domínguez-Rodrigo and Alcalá 2016). The Lomekwian
industry involved smashing large cobbles against anvils to produce flakes, probably by
using bipolar and block-on-block methods.
Far more extensively documented is the Oldowan tool industry, the second oldest
technology attributed to hominins, with its earliest occurrence at Gona, Ethiopia around
2.6 Ma (Semaw et al. 1997). It is a quick and expedient method of obtaining a sharp flake
to use as a tool by striking a core with a hard hammerstone by using the knapping gesture
(Toth 1985b), which results in non-standard cores that reflect the original shape of the
stone.
Around 1.75 Ma, a series of technological advances appear in southern Ethiopia
and West Turkana, Kenya, including hierarchical, centripetal flaking and the production
of bifacially-flaked, intentionally shaped, symmetrical tools called handaxes (Lepre et al.
2011; Beyene et al. 2013). This technocomplex is known as the Acheulian industry and is
usually attributed to Homo erectus2. Archaeologists recognize an important transition
between the early and late Acheulian (ca. 0.7 Ma). Handaxes from this later period are
much more refined than earlier handaxes. They are characterized by careful platform
preparation, cross-sectional thinning with the aid of a soft hammer, and threedimensional symmetry (Wynn 2002; Stout 2011). The Acheulian industry therefore
increased in technological complexity over time. Some researchers have claimed that the
appearance of the Acheulian technocomplex in the archaeological record signifies an
increase in cognitive capacity and potentially the introduction of protolanguage in
relation to the earlier Oldowan industry (e.g., Tattersall 2008; Shipton 2010; Arbib 2011).
2
Taxonomic debates will be largely avoided in this dissertation. Fossil hominins that date between 1.8
and 0.5 Ma will be referred to as H. erectus sensu lato, unless otherwise specified.
2
Interestingly, a significant increase in cranial capacity occurred between 1.6 and 1.8 Ma
(Klein 2009; Shultz et al. 2012), around the same time as the advent of the Acheulian
industry. The correspondence between the timing of the appearance of Acheulian
handaxes and this expansion in brain size may be more than a coincidence and will be
discussed in more detail in later chapters.
If it is assumed that, like the functional anatomy of the postcranium (Holliday
2012), human cognition and its underlying neural structures underwent a gradual and
mosaic evolution over time, rather than a saltational evolution as is often speculated, then
there should be evidence for the appearance and accumulation of human-like cognitive
features over time. Unfortunately, the information from fossil crania and endocasts are
limited in that they cannot provide definitive proof for whether a hominin species did or
did not possess a specific feature of cognition, such as inhibition, theory of mind, or
complex working memory. As products of cognition, stone tools may be more
informative about extinct hominin cognition and potentially language than fossil evidence
can currently offer; however, they also can lead to misinterpretations and unsupported
speculation. Historically, archaeologists who were interested in cognition focused their
analyses on the final shape and characteristics of what they determined to be the intended
tools of the toolmaker. This exercise did little more than to highlight the biases and
subjectivity of the archaeologists; thus, little progress was made in identifying the
cognitive processes that were necessary to make the tool. To avoid this kind of
speculation, controlled experiments that focus on the entire process of tool manufacture,
or the chaîne opératoire, rather than just the end product, are required (Sellet 1993).
Neuroimaging methods, when incorporated into experimental archaeology, offer perhaps
3
the most effective means to link up the process of toolmaking with the functional
processes of the brain in order to identify the cognitive features that are necessary for
making certain stone tool types.
The first attempts to map stone knapping to the brain focused on the production of
Oldowan and late Acheulian tool types3 (Stout et al. 2000; Stout 2003; Stout et al. 2006;
Stout and Chaminade 2007; Stout et al. 2008). The studies by Stout and colleagues lend
support to the claims that the Acheulian industry represents a trend towards more
complex cognition and protolanguage, mainly because they report some minor, but
promising, differences between the Oldowan and late Acheulian tasks. They note that
Oldowan toolmaking involves sensorimotor, superior parietal, and premotor areas of the
cortex, and late Acheulian toolmaking involves the right pars triangularis in addition to
these areas. Pars triangularis forms the anterior part of the inferior frontal gyrus (IFG),
which is a supramodal processor for hierarchically structured sequential information
(Fadiga et al. 2009). Pars triangularis is probably best known for its role in language
processing because, as a part of Broca’s area in the left hemisphere, it is thought to be
involved in the integration of semantic and syntactic information (Vigneau et al. 2006).
This difference between the neural correlates of Oldowan and Acheulian toolmaking
could imply the need for more complex, hierarchical processing during handaxe
manufacture than what is required for Oldowan flake production. Stout and Chaminade
(2012) argue that these results indicate that language had a technological origin, such that
it co-opted the hierarchical processing functions of the IFG that were already in place for
carrying out complex actions like tool manufacture. Moreover, they posit that intentional
3
Because the Lomekwian industry was only recently discovered, it has not yet been included in any
neuroarchaeological studies, including the current study.
4
pedagogical demonstration of toolmaking skills could have provided a scaffold for the
evolution of intentional vocal communication.
If handaxe manufacture indeed requires the participation of different neural
structures from those required for simple flake production in modern human subjects,
especially structures that are involved in executive functions, then it is probable that these
differences in neural processing also existed in the past. This could potentially shed light
on the timing of major evolutionary changes in hominin cognition, as well as identify
exactly what these changes were. The neuroimaging studies conducted by Stout and
colleagues offer a promising glimpse into possible differences between Oldowan and
Acheulian toolmaking cognition; however, none of these studies describe functional
neural activation attributed to naturalistic stone knapping at the same time that it is taking
place. For example, most of the participants in these studies knapped outside the positron
emission tomography (PET) scanner during the body’s uptake of the radiological tracer.
This and the fact that PET has very low temporal resolution means that the images taken
during these experiments could not capture functional brain activity in real time.
Furthermore, the sample size of these studies is low, ranging from one to six participants.
Low sample size and scanning the brain only after the task has already been completed
could reduce the likelihood of detecting a true effect. One consequence of this may be the
unintended omission of functional areas that are indeed recruited during the task. It is
therefore important that, in order to test the viability of these results, a study with a larger
sample size be conducted that utilizes a neuroimaging tool that has a high temporal
resolution and allows for hemodynamic activity to be measured during the actual act of
stone knapping.
5
The hemodynamic signal can look very different at various stages in training
while performing the same task (Floyer-Lea and Matthews 2005). Practice can result in
an increase or decrease in neural activation in certain areas, or it can lead to a functional
reorganization of brain activity (Kelly and Garavan 2005). Many times, brain areas
involved in cognitive control are recruited more heavily in the earliest stage of learning
and then subsequently drop out once the task can be performed using procedural memory
(Kübler et al. 2006). This is important to consider when making inferences about past
hominin cognition that are based on functional neuroimaging data from one point in time.
Even if executive functions were needed only during the earliest stage of learning to
make Oldowan and/or Acheulian tools and not after some experience, this would still
indicate that early Homo needed to possess these cognitive abilities to be able to acquire
the skills to complete the task in the first place. This information is lost if an experiment
does not record the hemodynamic response at an early stage in skill acquisition.
Knapping experience is addressed to some extent in the PET study described by
Stout (2003) and Stout and Chaminade (2007), in which they scanned the brains of
participants after they attempted to make simple Oldowan flake tools prior to any training
and then again after four hours of training. Because the participants received no training
whatsoever when they first attempted to produce flakes, it could be argued that their pretraining session is not technically the same task as their post-training session and
therefore is not reflective of an early-stage learning network associated with Oldowan
tool production. Furthermore, Stout and colleagues (Stout et al. 2000; Stout et al. 2006;
Stout et al. 2008) only enrolled subjects with ten or more years of experience for the
studies that focused on late Acheulian handaxe manufacture; thus, there is currently no
6
information on Acheulian early-stage learning networks. To test whether higher-order
cognition areas are ever recruited while learning how to make Oldowan and Acheulian
tools, a study that measures the hemodynamic response at multiple points in training is
necessary.
While it is possible that language in some form may have evolved by the time
early Homo was making handaxes, most researchers agree that there is not enough
evidence at this point to support this notion. Moreover, experimental studies reveal that
skills needed to make Oldowan and Acheulian tools can be learned without the aid of
linguistic instruction, and the flakes produced under verbal and nonverbal learning
conditions display different traits (Putt et al. 2014b; Morgan et al. 2015). If the behavioral
output (i.e., flakes) varies based on a verbal versus nonverbal learning context, then this
likely signifies differences in how the two groups process the task mentally. Whether the
hominins who made Oldowan and Acheulian tools had either language or protolanguage
is a question that remains unresolved. Furthermore, it is possible that learning context
interacts with how stone knapping is learned. For these reasons, language is a variable
that should be controlled to test for this possibility. Stout and co-workers’ neuroimaging
experiments may be the closest science has come to visualizing extinct hominin cognition
in action; however, the participants in these experiments all learned via verbally delivered
instructions at some point in their training. Thus, a study is needed that tests the effects of
both language and non-language based learning modalities during stone knapping skill
transmission on neural activation patterns.
If linguistic instruction does effect change in the neural processing of Early Stone
Age (ESA) toolmaking, especially in the IFG areas that are reportedly active during ESA
7
tasks, then this would call into question the Technological Origin for Language
Hypothesis proposed by Stout and Chaminade (2012). The recruitment of language
processing areas during ESA tool production may simply be the result of the presence of
language in the learning context. On the other hand, if these areas remain active even
when linguistic instruction does not form a part of the learning context (i.e., if
participants in the study learn ESA skills through silent imitation), this would lend
support to the hypothesis that language exapted these brain structures and functions that
were previously in place for complex actions. Therefore, a study that looks at the effects
of training under verbal and nonverbal conditions on ESA skill acquisition can also shed
light on the potential technological origin for language.
Objectives
This dissertation has four main objectives, which are addressed by reporting the
results of a large-scale neuroarchaeological experiment that measured the functional
activation of selected regions of the cerebral cortex during ESA stone tool manufacturing
tasks, using a neuroimaging technique called functional near-infrared spectroscopy
(fNIRS).
1. Determine the neural correlates for Oldowan and Acheulian lithic reduction.
Are Oldowan and Acheulian tool manufacturing tasks separate processes in the
brain? In other words, is there evidence that Oldowan and Acheulian tool production
behaviors activate different neural networks? Preparing an Acheulian handaxe is assumed
to involve more complex behaviors than producing Oldowan flakes, but is this difference
8
reflected in the functional neuroanatomy of these tasks? If a difference between the two
tasks exists in the brain, this should be reflected by the involvement of more anterior
areas of the frontal cortex during handaxe manufacture and more posterior areas of the
frontal cortex during simple flake production. Anterior frontal cortex (i.e. prefrontal
cortex) activation during the Acheulian task would indicate the involvement of executive
functions, such as attention, working memory, impulse control, and planning, among
others.
To address this first objective and answer the questions posed above, an
experiment was designed that used fNIRS to measure real-time changes in the brains of
31 trained human subjects while they participated in naturalistic Oldowan and Acheulian
knapping tasks. fNIRS as a neuroimaging technique measures oxygenated and
deoxygenated hemoglobin (oxy- and deoxy-Hb) levels in specified areas of the cortex.
This is done by placing a series of laser sources and light-detecting optodes on the scalp,
which send near-infrared light into the head and simultaneously detect the level of light
attenuation in the returning signal for each source-detector pair. By positioning these
optodes over known neuroanatomical locations, these signal differences can reveal the
increased participation of one neural region relative to others during a controlled task.
fNIRS is the ideal tool for measuring brain activity during stone tool
manufacturing tasks for several reasons. All it requires is that the subject wear a snug cap
that attaches to the laser sources and detectors. Because of software that can detect and
eliminate motion artifacts after data collection, subjects are allowed much greater
freedom of head, body, and hand movement than what is typical for other common
neuroimaging techniques, such as PET or functional magnetic resonance imaging (fMRI).
9
With a temporal resolution in the millisecond range, fNIRS can detect changes in
functional brain activity as they occur. It is also safe, noninvasive, and offers a more
realistic setting for replicating ESA tools.
2. Investigate the learning networks involved during ESA skill acquisition.
How do the neural networks that are recruited during ESA toolmaking change
over time as a result of training? Do cognitive control areas play a larger role at an earlier
stage of learning than they do at a later stage of learning? Finally, does training lead to
different neural activation patterns as participants learn to knap Oldowan and Acheulian
tools? This study further tests the hypothesis that the brain processes Oldowan and
Acheulian toolmaking differently by observing whether or not they follow the same
general learning trends over time. To address this objective, the hemodynamic responses
of 33 subjects were measured as they participated in Oldowan and Acheulian toolmaking
tasks at three different points in their training, once near the beginning of their training,
once mid-way through their training, and then once more after they had attended all the
required practice sessions and previous neuroimaging sessions.
3. Explore the influence of spoken language on complex motor skill acquisition.
As has been noted previously, participants who learn to knap stone tools under
varying linguistic contexts (e.g. verbal vs. nonverbal instruction) produce different types
of assemblages, as well as flakes of varying shape (Putt et al. 2014b). This suggests that
the context of the transmission of ESA skills may alter the pattern of areas that are
recruited when making stone tools. The employment of different cognitive strategies, for
example, may explain the distinctiveness of the assemblages of verbally versus
nonverbally instructed knappers. Does learning context interact differently with Oldowan
10
and Acheulian tasks? Can it change how knapping skills are learned over time? To
address this third objective and accompanying questions, the participants in the study
were separated into two groups that would determine whether their learning context
included verbal instructions along with a video display of ESA toolmaking skills or
simply a silent video display during the seven practice sessions.
4.
Test the Technological Origin for Language Hypothesis.
Previous research has emphasized a potential co-evolutionary relationship
between stone toolmaking and language, but as Chase (2006:107) asserts, it does not
seem as if “any of the scholars who link stone tool making to language has made a
methodical effort to consider alternative hypotheses [to a technological hypothesis] by
developing and applying test implications that would make it possible to choose among
them.” To address this point, this study explicitly tests the validity of the Technological
Origin for Language Hypothesis by eliminating linguistic instruction as a potential
confounding variable. With its connections to temporal areas and the basal ganglia,
Broca’s area may retrieve lexical and semantic information stored in declarative memory
(Ullman 2006). By eliminating linguistic instruction in the learning context, this reduces
the likelihood that language-sensitive areas of the brain become active during stone tool
manufacture because of their role in retrieving declarative knowledge gained during
practice sessions that included spoken instructions.
11
Organization of the dissertation
The dissertation is organized into nine chapters, including an introductory chapter,
three background chapters, one methods chapter, two stand-alone chapters that present
the results of separate analyses, and two chapters that discuss the broader context of the
results in relation to human evolution.
Chapter 2 establishes the theoretical framework upon which research into the
cognition of extinct hominin species must stand, and it provides the historical context for
the field of evolutionary cognitive archaeology by reviewing the works that have been
most instrumental in establishing it as a viable science in the study of human evolution.
Despite a general curiosity about the origin of the human mind that extends back for
millennia, it was not until the late nineteenth century that scholars made the first explicit
connection between ancient artifacts and their implication for the behaviors and
underlying cognition of the archaic human ancestors who made them. Moreover, it was
not until the late 1970s that archaeologists began to substantiate their claims about the
mental capacities of extinct hominin species with established psychological theory. Since
then, archaeologists whose theories have been grounded in established models from the
cognitive sciences, especially cognitive neuroscience, have found themselves to be in the
preeminent position to investigate the roots of human cognition.
Working memory is a cognitive process and a crucial element that contributes to
general intelligence. Its role in the evolution of modern human cognition and behavior
has become the center of some debate, so it is necessary to explore what is currently
known about its evolutionary past. Chapter 3 provides a background on this by way of
12
three different approaches: a comparative approach, a developmental approach, and an
archaeological approach. Working memory is present in other species and therefore has a
long evolutionary history. By studying the working memory capacities of other primates,
a baseline for early hominin working memory can be established. The developmental
stages of working memory in human children can help substantiate the establishment of
this baseline, as well as identify additional executive function components that older
children and adolescents possess that are absent in younger children and primates, which
can help guide archaeologists in their search of evidence for elements of extinct hominin
cognition in the archaeological record.
Chapter 4 highlights some of the disparate theories on the origin and evolution of
language that have emerged because of the lack of communication and cooperation
between disciplines in the recent past. Unlike working memory, language is a uniquely
human behavior in the animal kingdom; however, certain subcomponents of language
occur in other species, which may indicate that language evolved gradually out of a preexisting system. One such system may have been the neural structure already in place for
manual coordination. If this is the case, then the potential relationship between language
and the complex movements required to make some of the earliest tools cannot be
overlooked. Thus, this chapter focuses on the ideal placement of neuroarchaeology in the
language evolution debate because of its unique, interdisciplinary perspective and access
to multiple modes of relevant data, including artifacts from the archaeological record and
brain-behavior relationships. The chapter concludes by discussing the rationale for
carrying out the current study.
13
The hypotheses, materials, and methods included in this dissertation are detailed
in Chapter 5. Specifically, four hypotheses that directly relate to the four objectives of the
dissertation are proposed. The principles of optical imaging using fNIRS are described,
along with the benefits and drawbacks of this particular neuroimaging technique. This
chapter goes into detail on the subject population used and the overall design of the
experiment. The methods of data collection and post-processing are discussed. Finally,
the statistical analyses used to test the study’s hypotheses are specified.
Chapter 6 is written as a stand-alone article that focuses on what the brain
activation patterns of trained knappers can tell us about the potential Oldowan-Acheulian
cognitive transition and exaptation of toolmaking operations for language processing. It
describes the results of an analysis that compares the brain activation patterns of the
verbally and nonverbally instructed groups during Oldowan and Acheulian tool
production after around eight hours of training. This analysis reveals a clear distinction
between the manner in which Acheulian and Oldowan toolmaking are processed in the
brain. Additionally, this analysis confirms that learning context affects which areas of the
brain become active during the act of stone knapping. Some of the areas identified by
Stout and colleagues (2008) as key to stone knapping are replicated in this study but only
in the group that learned to knap with spoken instructions.
Chapter 7 is also written as a stand-alone article that explores the issue of how
knapping neural networks change over time as a result of one’s learning context and
training. It presents the results of a follow-up analysis that includes data from three
neuroimaging sessions for each participant. These results provide an even clearer
distinction between the cognition required to make Oldowan and Acheulian tools. An
14
investigation into whether these differences are the result of different learning stages
rather than the recruitment of different neural networks and cognitive strategies follows.
Learning differences between the verbal and nonverbal groups are explored and
explicated as well.
Chapter 8 synthesizes these results in order to propose a novel hypothesis, one
that posits that early Pleistocene hominin individuals with larger working memory
capacities and their offspring experienced higher rates of reproductive success because of
their higher likelihood to successfully learn how to make handaxes, which were most
likely used to procure difficult-to-access, calorically dense food items or possibly a wider
diversity of nutrient food resources. Because of evidence that tool use and working
memory correlate with brain size, this chapter makes the claim that the selection for
enhanced working memory in relation to technological change during the early
Pleistocene resulted in the expansion of working memory areas of the brain.
The structure of the final chapter, Chapter 9, is organized by the four objectives
and associated hypotheses of the dissertation, outlined above. Each objective is restated
in the context of the results, and how the dissertation succeeded at addressing each of
these objectives is discussed. The broader implications of these results for human
evolution are explained, and future research avenues are suggested.
Summary
Paleoanthropologists are limited in what they can confidently state about how the
most characteristic features that demarcate humans from other species, complex
15
cognition and language, arose and evolved. This is mainly because these features do not
fossilize. As a product of cognition and culture, technology, in the form of stone tools,
may offer the best evidence for extinct hominin cognition and possibly language because
different tool types demand varying levels of skill, with the least complex tools extending
as far back in time as 3.3 Ma. By combining neuroimaging techniques with experimental
archaeology, the cognitive abilities needed to produce stone tools of varying complexity
can be empirically tested in modern humans and applied via analogy to the extinct
hominins who made these tool types in the past. Previously published studies using these
methods have proposed a technological origin for language and have provided some
tantalizing clues as to possible cognitive differences between the hominins who made the
Oldowan tools and those who made the late Acheulian tools. Because of some limitations
in the sample size, neuroimaging technique, and experimental design used in these
previous studies, it was important that these findings be replicated with a novel
experiment that addresses each of these issues.
This dissertation presents a study that measures the brain activation patterns of
human subjects as they make replicative Oldowan and Acheulian stone tools. The
designing and conducting of an experiment with a large sample size relative to other
neuroimaging studies addresses the limitations of previous neuroarchaeological research.
Additionally, this experiment utilizes a neuroimaging technique that captures the
functional imagery of the brain in real time as the subject knaps over multiple sessions,
and the experiment design controls for language instruction in the learning context. The
four objectives for the dissertation are 1) to determine the neural correlates for Oldowan
and Acheulian lithic reduction, 2) to investigate the learning networks involved during
16
ESA skill acquisition, 3) to explore the influence of spoken language on complex motor
skill acquisition, and 4) to test the Technological Origin for Language Hypothesis. These
objectives are addressed in the following eight chapters. In these chapters, it is argued
that technology may be more than a mere reflection of past cognition; rather, it may be
one of the contributing causes for the evolution of the complex cognition and language
systems that define H. sapiens as a species.
17
CHAPTER 2:
EVOLUTIONARY COGNITIVE ARCHAEOLOGY: A HISTORY
Asking an archaeologist to discuss language is rather like asking a mole to describe life
in the treetops…. And yet the intricate physiological basis of language makes it perfectly
clear that this human ability has deep roots… If the forest has been cut down and all that
remains are the roots, then the mole may not be such an inappropriate consultant.
—Glynn L. Isaac (1976:275)
Introduction
Many view the mind as the very essence of what makes one human, and arguably,
one cannot understand the self without understanding the mind. So, the questions that
researchers across many disciplines have been asking for generations, ‘where do we
come from?’ and ‘how did we get here?’ are really questions of where the human mind
came from and how it has arrived at its current state. In this way, a study of the mind in
the past is of utmost importance. As so many have already pointed out, though, the mind
does not fossilize. Without living people to study or a fossilized mind to exhume, there is
little remaining for the search for the mind in the past, and so has been the general
mentality towards the subject for most of written history. Archaeologists, however, have
come to play an important role in this search.
It is the archaeologist in fact who is in the prime position to study the roots of
human cognition. This chapter explores the history of the academic discipline known as
evolutionary cognitive archaeology, whose charge has been to investigate the process of
18
human mental evolution via interpretations of the past material record of humans and
their ancestors. Specifically, it examines the successes and pitfalls of those works that
have been most instrumental in establishing evolutionary cognitive archaeology as a
viable science in the study of human evolution, beginning from some of the earliest
written records on the human mind to articles in the present day that probe the many
different issues that highlight the diversity of interests of those currently publishing in the
field. The largest hurdle for evolutionary cognitive archaeologists would prove to be,
ironically enough, their adherence to the search for the evolution of the human mind
because the concept of the ‘mind’ is a social construct and therefore unsuitable for
scientific inquiry.
“Fossil intelligence” and “petrified mind”
The study of the evolution of mental processes has recent origins in philosophy
and science. And even some of these earliest scholars of mental evolution thought it was
only natural to incorporate archaeological perspectives into their study of mental
processes in the past. Evolutionary theory, on the other hand, has ancient roots in
Western thought, with ancient Greek philosopher Anaximander (ca. 611-547 BCE) and
Roman philosopher Lucretius (99-55 BCE) as some of the first thinkers to coin the
concept. They argued that living things are related to each other and have changed over
time (Clagett 2001). Despite the idea of evolution being present as early as GraecoRoman times, its application to the origins of the human mind was not seriously
considered until the nineteenth century. The mind prior to this time was viewed as an
19
abstract concept that was often believed to be extrasomatic and therefore separate from
organic processes, such as evolution. Some writers, such as Alcmaeon of Croton (ca. 500
BCE), Hippocrates (ca. 460-377 BCE), and Galen (ca. 129-199 CE), believed the brain to
be the seat of the mind, but it was the nonmaterial theories of the mind, or psyche,
espoused by Plato (429-347 BCE) and Aristotle (348-322 BCE) that would be adopted by
Christianity and thus would have a long influence on the way people thought about the
mind and behavior for the next 2,000 years (Kolb and Whishaw 2009). Descartes (15961650), for example, continued the perpetuation of viewing the mind as an immaterial soul
separate from the body. He did allow, however, that the mind controlled the pineal body
of the brainstem, which, he believed, interacted with the body to make it move and react.
For Descartes, the body was like a machine, whereas the mind was indicated by the
presence of language and reason, which thus meant that all animals lacked a mind to the
exclusion of human beings.
The concept of the mind has been inextricably linked with language throughout
history, and there continues to be an unabated discussion of the relationship between
language and the evolution of the mind even today (e.g. Corballis 2013). Thus, any early
discussions of language origins may reveal how philosophers at the time viewed the
origins of the mind, but like the treatment of the mind by Graeco-Roman and early
Christian writers, the origin and evolution of language received little attention as a
serious line of inquiry because the idea that humans descended from animal ancestors
that lacked language was not widely accepted (Hewes 1993). There were a few
exceptions. Lucretius (ca. 99-55 BCE), in his De rerum natura, presented an account of
primitive humans living like animals without tools, fire, clothing, shelter, or language,
20
and these things arose from nature. An early Christian writer named Caecilius Firminaus
Lactantius (ca. 250-326 CE) believed that the powers of speech were perfected when
humans stood on two feet because it freed the hands to make and use tools (Hewes 1993).
This connection between tools and language might have been simply the observation that
freed hands allow for the expression of language through writing, but the connection is
still interesting to note because of the co-evolutionary link archaeologists would later
apply between tools, language, and the evolving mind in the twentieth and twenty-first
centuries.
By the middle of the nineteenth century, a materialist theory surfaced that asserted
that the nervous system rather than a nonmaterial mind or soul was responsible for
producing rational behavior. Charles Darwin and Alfred Russell Wallace agreed that
common characteristics shared by different species were because of common descent.
Their treatment of the evolution of the human mind, however, differed drastically.
Darwin (1871:38) argued that human intellectual abilities were an expansion of capacities
already present in animal predecessors and explicitly credited the brain as the material
source of intellect: “Little is known about the functions of the brain, but we can perceive
that as the intellectual powers become highly developed, the various parts of the brain
must be connected by the most intricate channels of intercommunication.” Wallace
(1864, 1869), on the other hand, maintained a strict division between the body and mind.
He fervently argued that the human body was the result of adaptive evolution, but he
could not fathom how something as complex and elegant as human rationality, morality,
and other higher faculties could have evolved by natural selection. Instead, he credited a
21
“superior intelligence” for the creation and purpose of the human mind. This was in large
part due to his conversion to spiritualism in the 1860s (Gross 2010).
By this time, the evolution of language had become a popular topic, one that
seemed to invite much speculation, eventually leading to its formal ban by the Linguistic
Society of Paris in 1867. This did not stop Darwin, however, from publishing The
Descent of Man in 1871, in which he presented his own imitative, musical protolanguage
theory. Meanwhile, the antiquity of chipped stone tools was becoming more widely
accepted due to the discovery of Acheulian handaxes in association with extinct
megafauna by John Frere in 1797 and Jacques Boucher de Crèvecoeur de Perthes in the
1830s (Trigger 2006). There were few attempts, however, to promote theories on the
evolution of mental or linguistic abilities based on archaeological finds. One exception
was Ludwig Noiré’s (1880) book on the significance of stone tools in human evolution,
Das Werkzeug und seine Bedeutung fuer die Entwickelungsgeschichte der Menscheit
(The Tool and Its Importance for the History of Human Development). In this work, he
argued that early, pre-linguistic vocalizations developed in tandem with primeval
toolmaking and tool using to become articulate language. Noiré (1877:354) believed that
humans were able to think because of language, stating, “Thought and speech are one;”
thus, because of this connection between language and cognition, this may be one of the
first attempts to describe primitive states of mind from an archaeological perspective.
One of the earliest writers to make the explicit connection between artifacts from
the archaeological record and the evolution of the mind was Henry Drummond
(1894:187): “The flints and arrow-heads, the celts and hammers, of early Man are fossil
intelligence; the remains of primitive arts and industries are petrified Mind.” Drummond
22
argued that stages of mental evolution could be witnessed directly through their
recapitulation in children, as well as through material artifacts, and “the Mind of a
Savage” (Drummond 1894:163). His description of the evolution of mind was entirely
unilineal, which was characteristic of the late nineteenth century regarding the
evolutionary process in general. He argued that stone tools reveal that the mind
progressed from “a very low order existing from an unknown antiquity” and gradually
improved to the present day, with some relapses (Drummond 1894:175). His interests
centered on the universality of a primitive mind as represented by “rough” stone tools all
over the world, rather than it being confined to a few peoples. Although he claimed that
the dawn of mind likely arose with the first stone tools, he granted only instinctual
behavior to these early peoples. The next stage of mental evolution occurred with the
transition to ground stone tools, which was the result of the once absent ability to
innovate, but “his inspiration probably came from nature” and not from some internal,
abstract representation (Drummond 1894:179). Drummond (1894:191) paid particular
attention to the apparently sudden jump in intelligence “at the eleventh hour,” concluding
that this sudden rise in intelligence could only have been brought on by the novel ability
to express one’s mind, or in other words, language.
By the 1930s, the idea that the human mind evolved organically over a long
period of time was generally accepted (e.g. M’Dougall 1925). Robert Schmidt (1934)
made further attempts to decipher elements of the mind from stone tools, artwork, and
other traces in the archaeological record, though these claims had no scientific backing
and only reflected the perspective of the author. Schmidt adhered to a strict
recapitulationist view of the evolution of the mind, looking to developmental stages in
23
children to learn about evolutionary stages but made no effort to look at tool use in
children. He argued that primitive humans lived in a similar state of imagination as
children before they were able to think rationally and logically. Overall, by this point,
scholars were still very limited in what they knew about human evolution, as Homo
neanderthalensis, Pithecanthropus and Sinanthropus (H. erectus) were the only accepted
members of the human lineage. Even though the Taung Child (A. africanus) had been
discovered in South Africa in 1925, most scholars dismissed it as nothing more than a
juvenile ape; therefore, scholars were not even aware of the large role that the African
continent played in the origin of humans. Only once more pieces to the human ancestry
puzzle were put together in the next few decades could a true cognitive archaeology
emerge.
Towards a hypothetico-deductive approach
Although scientists, philosophers, and laypersons alike have long remarked upon
the rather apparent differences between modern H. sapiens and extant non-human
primates in their language abilities, relative brain size, and toolmaking and tool using
behaviors, prior to 1979, there was little effort towards establishing a cognitive
framework to explain how these general features could be interconnected, even by those
paleoanthropologists whose research was so intrinsically tied to understanding the
importance of these three features to human evolution. During the mid-twentieth century,
the mental abilities of extinct hominin species were rarely a consideration when
analyzing the archaeological and human paleontological records. Instead, archaeologists
24
and paleoanthropologists alike emphasized taxonomy and later, ecology. This was largely
in part a result of the development of the “New Archaeology” and “New Physical
Anthropology” of the 1950s and ‘60s.
At a time when most archaeologists followed a typological approach for
understanding material culture, in which case emphasis was placed on artifact classes
rather than the human behaviors and cognition involved in their production, one French
archaeologist, André Leroi-Gourhan, not only developed the concept of chaîne
opératoire in 1952 but also laid the foundations for understanding the evolution of
hominin cognition and especially language from technical procedures in the 1960s
(Leroi-Gourhan 1964; Pelegrin 2009; Trigger 2006). The chaîne opératoire methodology
looks at the operational sequences that occur in the production and life history of
artifacts. Even though the chaîne opératoire approach to studying the production process
rather than the final form of an artifact alone is widely used by archaeologists, especially
by cognitive archaeologists today (e.g. Mahaney 2014b), at the time, Leroi-Gourhan, and
the work of francophone archaeologists in general, was largely ignored by anglophone
archaeologists, but his work would have considerable influence on future French
archaeologists.
American archaeologists independently began to move away from the typological
classification approach of cultural-historical archaeology and towards a reductionist,
neoevolutionary approach, in which cultures were thought to be functionally integrated
components within an ecological system and were subject to a limited number of
historical processes (Caldwell 1959; Trigger 2006). Expounded upon by Lewis Binford in
two formative papers, “Archaeology as Anthropology” (1962) and “Archaeological
25
Systematics and the Study of Culture Process” (1965), this new, or processual,
archaeology interpreted cultural systems as adaptive responses to alterations in the
natural environment or to competition within or between cultural systems.
There was little emphasis on the individuals who produced the archaeological
record, the transmission of knowledge between these individuals, or the degree of
cognitive complexity necessary for the transmission of said knowledge, in the case of
individuals dating to the Paleolithic. “For Binford, the concept of culture signified
primarily the different ways in which groups of human beings adapted to their
environmental settings” (Trigger 2006:396; emphasis mine). This approach is perhaps
best exemplified by the “Mousterian debate” between Binford and Francois Bordes.
Whereas Bordes (1953, 1972) argued that the changes between Mousterian layers in rock
shelters in the south of France marked the ethnic differences between Neandertal groups,
Binford, along with his colleague and wife, Sally Binford (1966), argued that because
these artifact types were intermingled at sites in other regions of Europe and the Middle
East, these tools actually represented different functional activities performed by
potentially the same group of Neandertals. So, whereas Bordes saw the different artifact
shapes in an assemblage as representative of the mental imagery and concepts in a
toolmaker’s head, which were culturally transmitted over generations between
individuals, Binford denied that such capacities even existed during the Middle
Paleolithic (MP), citing the environment as the main determining factor for the shape of
the artifacts (Porr 2005). In other words, the tools were shaped by the material conditions
in which they were to be used, not because of culturally transmitted, stylistic preferences.
This emphasis on the interactions of populations with their environment had a
26
particularly strong influence on physical anthropology as well, changing the way
researchers interpreted the fossil record and how they viewed the course of human
evolution.
There was a turning point in the field of physical anthropology after World War
II. In the early twentieth century, physical anthropologists focused their energies on
taxonomy and racial typology, but there was a shift towards a rather different way of
asking questions in the 1950s and ‘60s, beginning with Sherwood Washburn’s (1951)
article, “The New Physical Anthropology.” In the early twentieth century, physical
anthropology was more of a technique. Everybody essentially used the same methods to
answer different questions, but the field lacked any theory of its own. Washburn’s new
physical anthropology would bring the field to the forefront of modern evolutionary
science. He called for an interdisciplinary approach to studying the process of human
evolution, combining the latest genetic contributions and ideas from the Modern
Synthesis, functional anatomy, paleontology, and primate behavioral ecology (MikelsCarrasco 2012). Washburn (1951:299) also emphasized that, “…selection on the
phenotype, adapting animals to their environment, is the primary cause of alteration in
gene frequencies.” In other words, it is not enough to discuss adaptation alone, as so
many physical anthropologists had done previously; adaptation can only be discussed in
the context of the environment.
The 1960s, thus, was a time of upheaval of old ideas and a transition to a more
modern paradigm for physical anthropology. While this was facilitated by the
introduction of interdisciplinary techniques to paleoanthropology as espoused by
Washburn, several new hominin species also were discovered and dated with radiometric
27
techniques (e.g. Fleischer et al. 1965; Leakey et al. 1961), thus expanding the known
range of phenotypic variation in hominins and providing a clearer focus on the timeframe
of hominin evolution and the first tool industries. The general consensus at the time
regarding hominin origins was that the last common ancestor that humans shared with the
other great apes lived during the early Miocene. The misinterpreted Ramapithecus fossils
dated to 7-14 million years ago (Ma) were thought to belong to the earliest hominin
ancestor (Cartmill et al. 1986; Simons 1961), thus allowing for at least 14 million years
of evolutionary distance between humans and the other apes. It was thought that this long
period of separate evolution could explain the seemingly discontinuous characteristics
that humans possessed and other apes lacked, such as bipedalism, language, and a high
degree of intelligence. This idea was overturned by biomolecular data that demonstrated
a close affinity between humans, chimpanzees, and gorillas and the new evolutionary
timeframe imposed on hominin evolution by Vincent Sarich and Allan Wilson (1967)
using Zuckerkandl and Pauling’s (1962) “molecular clock” method to infer a
phylogenetic split between the three groups around 5 Ma. Additionally, Weiner and
colleagues (1953) proved the forged Piltdown Man to be a hoax with fluorine dating, thus
leading to the dismantling of the long-held belief that large brains evolved early in the
hominin lineage. Finally, the behavioral ecology studies on primates, especially those
reporting on chimpanzee behavior by Jane Goodall (e.g. 1964), established that many
complex behaviors once thought to be unique to humans, such as toolmaking and tool
use, were shared with other primates. Because of these discoveries, which were made
possible by novel techniques introduced from other disciplines, paleoanthropologists
were able to expand their scope of study beyond classification to include evolutionary
28
processes and ecological models. Once the new physical anthropology created a clearer
picture of major human evolutionary trends, hypotheses on cognitive evolution would
soon follow.
Even though some of the first papers on cognitive archaeology were not
enmeshed in the ecology fervor of the time, one could argue that the nascent field of
cognitive archaeology would not have arisen at all without the hypothetico-deductive and
interdisciplinary approaches of processual archaeology and the new physical
anthropology. These new approaches to the archaeological and human paleontological
records provided the theoretical framework on which cognitive archaeology could build,
even if at first those scholars discussing hominin cognition were a small minority and
may have even felt at odds with other researchers studying early human fossils and
archaeological assemblages.
Among this minority was Ralph Holloway (1969), who challenged the popular
view of the time that humans and their artifacts are simply adaptations to their
environment in his Current Anthropology paper, “Culture: A human domain.” Instead, he
gave agency to humans and their toolmaking ancestors rather than the environment by
arguing that human culture by definition is “the imposition of arbitrary form upon the
environment” (Holloway 1969:395). He criticized those who relied upon ecological
models for human evolution and stressed the importance of finding markers for
intelligence, patterns of emotional behavior, and brain physiology and anatomy. And the
only clues to the behaviors and intelligence of early humans are the stone tools they left
behind.
29
Even though Holloway was not the first person to speculate on the potential
evolutionary connection between language and stone tool manufacture, his thesis stating
that language and toolmaking are similar, perhaps even identical, cognitive processes has
withstood the test of time and has only gained traction as more evidence from multiple
fields builds upon his case. He supported his thesis by systematically comparing the
process of toolmaking to Greenberg’s (1967) language structure features and some of
Hockett’s (1960) design features for language (Table 1). Holloway applied this logic to
the archaeological record of the Early Stone Age (ESA) to explore the question of
whether or not australopiths were protocultural because it was assumed at this time that
australopiths were the first toolmakers.4 Based on the similarities of the Oldowan
industry to language described above, he argued that there is no evidence that could rule
out language behavior in australopiths. He concluded that the makers of the Oldowan
possessed the cognitive structure necessary for language and also possessed culture
because they were able to impose their own culturally conceived arbitrary form upon
their natural environment in a way that exceeds the toolmaking abilities of any living,
non-human apes. He also posited that imitation and observational learning would not be
sufficient to explain the time depth and wide geographical distribution of certain tool
4
It is interesting to note that Holloway’s assertion that australopiths were the first hominins to make
intentionally modified stone tools has come full circle since 1969. When Holloway wrote “Culture: A
human domain,” Homo habilis had only recently been discovered, and there was still much contention
about who was responsible for making the Oldowan industry. More recently, the Oldowan industry usually
has been attributed to early Homo because it fits the expectations that the first stone toolmakers should have
larger brains than their predecessors. With the recent discovery of 3.3-million-year-old stone tools from
Lomekwi, Kenya (Harmand et al. 2015), however, the debate may finally be settled that australopiths did
indeed modify stone tools. At present, the oldest fossil hominin attributed to Homo is the recently
published mandibular remains from Ledi-Geraru in the Afar region of Ethiopia dated to 2.8 Ma
(Villmoare et al. 2015).
30
types. For instance, symbolic language would have been necessary by the time handaxes
appeared in the archaeological record.
31
Table 1. Holloway's (1969) application of Greenberg (1967) and Hockett's (1960) structure and design
features of language to the process of stone tool manufacture.
Greenberg and Hockett's
structure and design features
Structure and design features as
applied to language
Phonology
The combination of sequences of
meaningless sound units called phonemes
Grammar
The rules that regulate the order of
elements to be meaningful
Semantics
Duality of patterning
Productivity
The meaning of a linguistic utterance
The organization of language into the
smallest meaningless elements (i.e.
phonemes) that can be combined to form
the smallest meaningful elements (i.e.
morphemes)
The limitless ability to use language to
make novel statements
Arbitrariness
The absence of any natural connection
between a word and its meaning
Traditional transmission
The cultural transmission of language
from generation to generation
32
Structure and design features as applied
to stone toolmaking
The combination of small, meaningless
units related to stone toolmaking, such as
the combination of a hard hammer,
application of force, and a flint core
The combination of smaller, meaningless
unit actions to produce a meaningful end
product, such as a handaxe
The meaning or purpose of a finished tool
and the meaning applied to each unit
action in relation to preceding and future
unit actions
The combination of meaningless unit
actions (e.g. removing a flake, flipping the
piece to set up a new platform, etc.) that
culminate in a meaningful final product
The ability to use the same basic tool types
for multiple purposes
The absence of any resemblance of an end
product stone tool to the original rock
from which it was reduced
Complex organization of sequential units
of stone toolmaking according to
established conventions required that skill
be learned and transmitted across
generations
In spite of the cold reception Holloway met with this paper, he was not alone in
thinking about the importance of cognition in human evolution. Oakley (1969) almost
simultaneously published an article that claimed that the ability to use tools to make tools
requires foresight. In the 1970s, other scholars began to incorporate these ideas in their
interpretation of the evolutionary past. Some examples include Hewes’ (1973) proposal
for a shared visual, motor, and cognitive pathway for toolmaking, tool using, and gestural
language, Montagu’s (1976) claim that the manufacture and transmission of the Oldowan
implies the existence of some type of rudimentary speech, Guilmet’s (1977) discussion of
whether it was possible to discern learning from intentional or unintentional modeling by
observing artifact type variation, and Kitahara-Frisch’s (1978) position on stone tools
being the first unambiguous evidence for a capacity for reflection. Most notably, Glynn
Isaac (1976) addressed the problem that there was no systematic way to measure artifact
complexity in order to gauge the level of cognitive development of the toolmakers. His
paper highlighted evolutionary trends as they are represented in the archaeological record
and offered a somewhat speculative system for measuring cognitive complexity. He
proposed the following as indicators of cognitive development: 1) the number of distinct
artifact classes; 2) the amount of variation between artifacts within the same class; 3) the
number of operations involved in the production of an artifact; 4) whether the artifact is a
composite tool or not; and 5) the extent of regional variation. With this system
established, he concluded that there were three stages of cognitive development. In the
first phase, there was an evolutionary shift towards a protohuman adaptive complex,
which included bipedalism and toolmaking. The second phase established an adaptive
system that selected for enhanced communication and information exchange in relation to
33
more complex tool design. Finally, in the third phase of the Middle Pleistocene, he
credited the “Upper Paleolithic cultural spurt” to the maturation of cultural and linguistic
capabilities (Isaac 1976:286). Even though his argument was rooted in a detailed
understanding of the archaeological record, his account was still admittedly speculative
because it was not based on any empirical evidence that could definitively link these
stages of cognition with these stages of artifact complexity. The speculative nature of
some cognitive archaeology papers has been a lingering issue in the field and a cause for
criticism. There was a clear need for a theoretical framework of cognitive archaeology
that effectively linked the patterns observable in the archaeological record to empirically
based evidence.
The application of Piagetian theory to the archaeological record
In 1979, three independent papers were released that marked the beginning of
contemporary cognitive archaeology from an evolutionary perspective. The first of these
papers was by John Gowlett (1979:14), who published an appeal in Nature to other
scientists to consider hominin intellectual development from an archaeological
perspective, stating, “Man can…investigate the evolution of his present cognitive
capacity through the permanence of the cultural record.” He focused on the Oldowan and
Acheulian tool industries, claiming that the levels of standardization seen in the Oldowan
imply flexibility, stability, and culturally controlled behavior, while the complex process
of Acheulian toolmaking implies forethought, planning, and systematic exploitation of an
environment, this last remark being reminiscent of Holloway (1969). Gowlett stressed the
34
importance of interdisciplinary cooperation between archaeology, hominin neurology,
and psychology for future research, as well as the appropriate use of animal models.
Later that year, two papers by Sue Parker and Kathleen Gibson (1979) and
Thomas Wynn (1979) answered Gowlett’s (1979) call for an interdisciplinary approach
to interpreting evidence for intellectual development from the archaeological record.
They independently converged upon the same Piagetian genetic epistemology to study
the intelligence of early hominins. Jean Piaget was a Swiss psychologist who studied the
origins of thinking at a time when most American psychologists were behaviorists and
avoided mental concepts altogether. He is best known for being the first psychologist to
attempt a systematic study of cognitive development (Piaget 1936), though most modern
developmental psychologists no longer adhere to his ideas. From his naturalistic and
clinical observations of children, he claimed that all humans pass through universal,
developmental stages, which include sensorimotor, preoperational, concrete operational,
and finally formal operational stages of intellectual development (Piaget 1936).
In the sensorimotor stage, infants pass through several substages to learn how to
interact with objects in their environment (Piaget 1937, 1964). For example, a key feature
of the sensorimotor stage is object permanence, meaning that the child understands where
an object is located, even if it is hidden from view. In the preoperational stage, children
progress from focusing on topological relations between objects to Euclidean spatial
notions, and by the end of this period, they can develop true concepts based on reversible
mental operations, and hypothetico-deductive reasoning (Piaget 1936). During the
concrete operational stage, children are capable of logical thought but only if applied to
physical objects (Piaget 1937). Finally, when adolescents enter the formal operational
35
stage, they are able to think abstractly without the crutch of concrete manipulation (Piaget
and Inhelder 1958). Piaget and Inhelder (1971) recognized from these patterns of
intellectual development that intelligence arises out of action, not perception. This is an
interesting point in relation to ideas about the evolutionary development of intelligence in
human ancestors because Piaget essentially provided archaeologists with a smoking gun
in the form of stone tools. Stone tools are the results of an individual’s thought processes
acted out upon the environment, which can, in turn, reflect the minimum stage of
intelligence of that individual. Finally, there was a developmental model by which the
archaeological record could be compared.
Parker and Gibson (1979) took Piaget’s developmental model and went one step
further by applying it to extant primates, and subsequently to early hominins. Their
argument was chiefly based on the recapitulation theory, which, in the case of their
article, stated that human children traverse through successive stages of intellectual
development that reflect the same stages of cognitive evolution of their distant ancestors,
and remnants of these earlier stages of intelligence are therefore recognizable in living
primate species. They surveyed a range of primate species, including prosimians (lemurs,
lorises, and tarsiers), New World and Old World monkeys, and apes; although, the Piaget
system had only been applied to a few species of monkeys and great apes (e.g. Parker
1973; Chevalier-Skolnikoff 1977; Parker and Gibson 1977).
Broadly speaking, Parker and Gibson (1979) found that prosimians are only
capable of the first couple stages of sensorimotor intelligence, lacking object
permanence. They established a phyletic series of stages for the evolution of intelligence
in primates, and they hypothesized that the earliest hominins were capable of the early
36
stages of preoperational intelligence; therefore, the subsequent stages of preoperational,
concrete operational, and formal intelligence must have appeared later in hominin
evolution.
Similarly, Parker and Gibson (1979) also compared the symbolic capacity of the
great apes to young children. Chimpanzees were known at the time to be capable of
intraspecific referential communication, intentional message transmission and reception,
and reversible roles (Savage-Rumbaugh et al. 1978, 1979), and to have a more flexible
gesturing communication system than vocal communication system, similar to the
protolanguage stage of human infants. For these reasons, they also argued that the earliest
hominins probably possessed a protolanguage similar to phase II protolanguage in
children. During this phase of language development, children engage in dialogues and
can refer to objects and events outside their immediate environment. Parker and Gibson
hypothesized that extractive foraging was likely the driving force behind the evolution of
hominin intelligence and language.
Parker and Gibson’s (1979) developmental model for the evolution of hominin
intelligence and language was a definite step in the right direction for the field; however,
critics found several weaknesses with their overall argument. In the open review
immediately following the paper, several authors commented on their overreliance on a
largely discredited recapitulation theory and on Piagetian theory, which had its share of
criticisms as well. For example, some studies have reported that many adults fail at
formal operation tasks (Keating 1979; Dasen 1994), which could be due to some of the
difficult to understand tests that Piaget designed. This could also mean that children’s
abilities were underestimated (Hughes 1975). Piaget’s studies have also been criticized
37
by the fact that he often used small sample sizes, and in his earliest studies, he used his
own children as subjects. Regarding Parker and Gibson’s (1979) application of Piaget’s
model to primates, one could also argue that this presents an anthropomorphic fallacy. It
seems unfair to impose an intellectual development system originally created to explain
human development on other primate species. A reversed model would be preferred that
places humans in the larger context of primate cognitive development, behavior, and
evolution. This problem becomes obvious when taking a closer look at Parker and
Gibson’s proposed phyletic stages represented by primate groups, which are not nearly as
neat as they avow. The cebus monkeys, for example, “like great apes…display the
intellectual abilities of the fifth and sixth stages of the sensorimotor intelligence” (Parker
and Gibson 1979:369), even though they are phylogenetically more primitive than the
macaques, which do not always demonstrate these apelike qualities.
Whereas Parker and Gibson (1979) applied Piaget’s model to the purported
functions of stone tools, Wynn (1979) assessed Acheulian hominin intelligence by
reconstructing the thought processes involved in the manufacture of late Acheulian tools
at the Isimila Prehistoric Site, Tanzania under the umbrella of Piaget’s genetic
epistemology. Because Piaget included spatial and geometric concepts, his model lent
itself well to the study of stone tool artifacts. And while Parker and Gibson (1979)
emphasized the development and early stages of intelligence, Wynn’s (1979) prime
objective was to search for the earliest evidence for adult operational thinking in the
archaeological record.
The two fundamental regulators of operational thinking are reversibility and
conservation. Reversibility is the mental ability to reverse the order of categories to
38
predict the outcome of an event. This is best explained by the statement, “If A=B and
B=C, then A=C.” Conservation is the idea that something is conserved even after a
manipulation, which allows one to pre-correct errors by going back to a starting point
within thought. The main conclusion of his paper was that the Isimila Acheulian tools
represent operational intelligence because the production process could only have been
organized by the employment of reversibility and conservation. Some of the bifaces
produced at the site have a high degree of symmetry and a minimal number of flakes
removed. Wynn argued that these tools could not have been produced by a preoperational
intelligence strategy of trial-and-error; rather, the knapper anticipated the final shape and
knew the exact combination of flakes to remove to reach the end product. This would
signify that operational intelligence had evolved by 300,000 years ago, which would
mean that events further down the line, such as the Upper Paleolithic (UP) Revolution,
were the result of cultural development, not an increase in intelligence. He also applied
Piaget’s genetic epistemology to evaluate the intelligence of Oldowan hominins at
Olduvai Gorge. He found that these hominins approached the task with a trial-and-error
strategy, and they failed to pre-correct errors and control multiple variables. He thus
concluded that they only possessed a preoperational intelligence, not so different from
ape intelligence (Wynn 1981).
“It is time for a ‘cognitive archaeology’”: The search for the hominin mind
By the 1980s, a number of archaeologists began to criticize processual
archaeology and its hypothetico-deductive approach, especially in the context of studying
39
the human mind in the past. Similar to the psychological school of behaviorism, which
only studied the behavioral stimuli and responses of subjects, not the minds of subjects,
processual archaeology also only focused on environmental stimuli and how people in the
past reacted to these external forces. Binford (1965:204), for example, thought “paleopsychology” was too speculative for objective study. The post-processual critique, as it
came to be called, pointed out that processual archaeology was a flawed approach to
interpreting the archaeological record because it could not directly test its claims, only
indirectly through experimental and comparative work, such as ethnoarchaeology
(Abramiuk 2012). More importantly, post-processualism was largely a critique of the
assertion that objective archaeological interpretations were possible through the
application of the scientific method, a key tenet of processualism. Post-processual
archaeologists did not limit themselves to the hypothetico-deductive approach; rather,
they viewed the archaeological record and the thoughts of their long-dead protagonists
from multiple, subjective perspectives. By bringing in many subjective views, they could
then identify the least plausible accounts without limiting their interpretation to one
privileged account. The counter-critique against post-processualism attacked their lack of
a method to actually root through the enumerable interpretations of the past and select the
most plausible explanation. Processual archaeology, with its comparative, scientific
method was still necessary for this task.
As a result of the post-processual critique, two camps of cognitive archaeologists
emerged (Abramiuk 2012), one from the cognitive-processual school that focused on
explaining how people in the past thought (e.g. Renfrew 1983), and the other from the
post-processual school that focused on what people in the past thought about (e.g. Hodder
40
1986). While it is not the goal of this chapter to pit one camp against the other, it should
be noted that a processual approach to the evolution of the mind has obvious benefits
over a post-processual approach. In order to make interpretations of prehistoric
individuals’ thought patterns based on their artifacts, post-processual cognitive
archaeologists must operate under the assumption that people in the past possessed
similar minds to what people have today, but this is problematic when dealing with the
minds of extinct species of human. For this reason, Renfrew’s (1994) processual
approach, which followed Wynn's earlier work and posited that behavioral manifestations
in the archaeological record are contingent on certain cognitive capacities being present,
is ideal when studying the evolution of the mind. So, while a post-processual approach to
cognitive archaeology might be a useful tool for interpreting the mind frames of recent
populations from the archaeological record (i.e. H. sapiens), it would be largely untenable
in the discipline of evolutionary cognitive archaeology. From here on, the narrative will
follow the development of evolutionary cognitive archaeology, not cognitive archaeology
in general, which diverged to the point that processual and post-processual cognitive
archaeologists no longer found the other camp’s research to be complementary to their
own research, and where only recently have attempts been made to treat the two as the
same discipline (Abramiuk 2012).
During the 1980s, evolutionary cognitive archaeology was still in the process of
establishing itself as a legitimate field of inquiry. As Renfrew (1983:13-14) stated,
“cognitive archaeology is a goal, not yet a body of coherent arguments….still regarded
by many social anthropologists as a complete impossibility.” This is also well represented
by Gowlett’s (1984) chapter in Hominid Evolution and Community Ecology. This
41
chapter, which emphasized the importance of human mental evolution, stands alone in a
book of contributed chapters about ecology, taphonomy, and migration patterns, which is
quite telling in itself. It is apparent that at the point that Gowlett was writing, the viability
of the study of human mental evolution still elicited skepticism from most researchers.
For example, Gowlett (1984:168) attempted to convince his readers that “mental
evolution can profitably be investigated in the past, at least in the period for which
cultural evidence survives.” It is worth noting that Gowlett’s (1984) approach to
comparing mental abilities of different groups based on their technologies is by an
“operation chain” approach, or a diagram of procedural steps necessary to produce
certain tool types, which is very similar to the French chaîne opératoire concept. With
this strategy, he came to very different conclusions from Wynn (1979, 1981) on the
mental abilities of the Oldowan toolmakers because a diagram of the operation chain
involved in making an Oldowan tool5 is much longer than any operation chain employed
by a gorilla, for example.
Wynn continued to be a strong supporter of Piagetian theory in the ‘80s, and
psychological models in general, in their application to the archaeological record. He
warned against traditional technological sequences created by archaeologists, where it
was assumed that increasingly complex technologies were the results of increasing
intelligence. According to Wynn (1985:32), these interpretations were “based on
common-sense ideas, whose source is usually self-reflection, rather than rigorously
established theories of intelligence.” Piaget’s model was the most applied theory of
5
Gowlett’s (1984) terminology here was meant to describe a core tool that has had more than one flake
removed. Wynn (1979, 1981, 1989) also focused on Oldowan core tools and retouched flake tools. In this
way, their conclusions are at least comparable, but there is little consideration of the concept that flakes
may have been the intended tools at this point in time.
42
intelligence yet developed and therefore was the best method to objectively address the
question of the evolution of the human mind at the time. Wynn responded to the
recapitulationist critiques against the application of Piaget’s developmental stages to
phylogeny by arguing that it is more of a constructionist approach than a recapitulationist
approach. In other words, each stage cannot begin to develop unless it can build upon the
preceding stage that had already developed, whether in the case of a developing child or
in the case of an evolving population (Wynn 1989). With all this said, Piagetian theory
lent itself well to the exploration of hominin spatial intelligence, but in this way it was
also extremely limiting because it could not address semiotic development, language, or
social organization.
At this time, the most sophisticated approach in evolutionary cognitive
archaeology was to apply a psychological model to archaeological remains as Parker,
Gibson, and Wynn had done. There had yet to be a theoretical, archaeological model
developed out of an established psychological model, that is to say until a team composed
of an archaeologist, Iain Davidson, and a psychologist, William Noble (1989, 1993;
Noble and Davidson 1996), developed their own theory for a depictive origin of language
based on James Gibson’s (1986) ecological psychology. The ecological psychology of
Gibson was unique in cognitive science in that it denied the existence of internal
representations, or mental imagery of any kind; rather, it stressed that perception is
specified by ecological information in an organism’s surrounding environment.
Consequently, Davidson and Noble claimed that without language, humans would not be
capable of abstract thoughts. But how did language evolve in the first place?
43
Davidson and Noble dismissed the importance of stone tools to the discussion of
language evolution and instead turned to an alternative visible behavior in the
archaeological record as evidence for language origins, that being depictive art. They
argued that only through iconic depiction could arbitrary symbols emerge. These
seemingly different behaviors both require reference to the perceived world and
reflection on what has been perceived in order to talk about it in the case of language and
in order to draw its likeness in the case of depictive art. Based on this argument, fully
syntactical language would not have emerged until the first depictive art appeared, which
was 32 thousand years ago (Ka) in Europe6 (Davidson and Noble 1989).
The response Davidson and Noble (1989) received for this theory was mainly
skepticism. One reviewer even went so far as to say, “The paper under discussion caused
me to experience an acute nostalgia for the good old days when the French Academy
would accept no papers on the origin of language because no evidence could, by
definition, be forthcoming” (Black 1989:138). The major points of contention revolved
around Davidson and Noble’s punctuationalist and discontinuous approach to the
evolution of human language and the lack of consideration of neurophysiological and
paleoneurological literature. Up to now, most scholars had agreed that the evolution of
language was likely a gradual process that occurred prior to the emergence of modern H.
sapiens, perhaps as early as the lower Pleistocene (e.g. Falk 1980b; Holloway 1969;
Montagu 1976; Tobias 1981), whereas Davidson and Noble only allowed for a very late
appearance, once brain size had already reached its modern dimensions. This is the
inherent result of a Eurocentric bias that assumes cognitive sophistication and behavioral
6
The dating and location of the earliest depictive imagery have changed since Davidson and Noble (1989)
wrote their article. Using the uranium-thorium dating technique, artwork in an Indonesian cave on the
island of Sulawesi was dated to 40,000 years ago (Aubert et al. 2014).
44
modernity occurred only during the UP Revolution (McBrearty and Brooks 2000). It
results from excluding stone tools from their analysis, as all other nonperishable evidence
of intentional modification behavior is limited to the late Pleistocene. It also stems from
the disregard of paleoneurological studies that, as Falk (1989b) pointed out, are notorious
for their disagreement about many aspects of hominin brain evolution, but not when it
comes to the capacity for language in early hominins (Falk 1983, 1987a, 1987b;
Holloway 1983b; Tobias 1981, 1988). Black’s (1989) critique on the lack of
neurophysiological literature considered in the formulation of Davidson and Noble’s
(1989) argument is an interesting one because it is true that at this point psychological
theory had been employed by evolutionary cognitive archaeologists, but besides the
lateralization studies of the 1970s and 1980s, no one had yet attempted to marry
archaeology with neuroscience to explore any of these pertinent questions. Others were
not so quick to abandon stone tool production as a viable source of information about the
language and cognition of early humans, though it was obvious a shift of focus to the
brain instead of the mind would be necessary.
Davidson and Noble deemphasized the importance of stone tools in the evolution
of cognition and language because of their Gibsonian theoretical stance, meaning that
early stone tools, such as handaxes, could not have existed as ideas in the minds of the
hominins who made them because there is no evidence that they created depictive
images, and thus by extension, they could not have had language. If early Homo did not
have internal representations of the handaxes they made, then handaxes were nothing
more than core products of mindless knapping. This led to what Davidson (2002) termed
the ‘finished artifact fallacy,’ which states that it is fallacious logic to assume that
45
archaeologists today know what were and were not intended tools in the past. Harold
Dibble (1984, 1987; Chase and Dibble 1987) came upon this idea earlier when he noted
that the typological variation among Mousterian scrapers might be the result of repeated
resharpening and modification, not stylistic choice. Chase and Dibble (1987:268) also
commented on how it is “far from proven that the ‘types’ that have been defined for the
Lower and Middle Paleolithic industries demonstrate particular cognitive categories of
the prehistoric hominids who made and used those tools. They certainly do reflect the
cognitive categories of the archaeologists.”
Davidson (2002; Davidson and Noble 1993) expanded on this idea, with many of
the critiques directly aimed at some of the evolutionary cognitive approaches to the
archaeological record to date. He argued that it is fallacious to assume that stone artifacts
are in the form that was intended by their makers. There are many facets to this
argument: 1) archaeologists sometimes unwittingly see patterns in the archaeological
record that simply did not exist in the minds of their past subjects; 2) it cannot be
assumed that the toolmaker intended anything more besides the removal of a flake; 3)
unmodified flakes may have also been useful tools; 4) apparent similarities in tools can
result from similar mechanical principles of knapping and do not always need to imply
similarities in intention; and 5) other constraints may cause what appears like intentional
patterning of the core. Overall, Chase and Dibble (1987) and Davidson and Noble (1993)
criticized the linguistic categories that Holloway (1969, 1981) and Gowlett (1984) placed
on stone tools, using the finished artifact fallacy. Although Chase and Dibble (1987) were
skeptical of archaeologists’ tendency to overvalue stone tools as evidence for prehistoric
culture or behavior, they did suggest, based on the current evidence that the art of
46
toolmaking and language involves similar processes (c.f. Holloway 1966; Leroi-Gourhan
1964; Lieberman 1975), that the neural structures involved in toolmaking could have
served as a preadaptation for language. Because of the uncertainties surrounding the
intentions of ancient toolmakers highlighted by the finished artifact fallacy, it became
increasingly frowned upon to interpret the process of stone tool manufacture as a window
into the mind of extinct hominins, especially regarding linguistic capabilities. For
example, Wynn (1991) argued that the differences between sequential actions involved in
tool manufacture and use and sentence grammar are so great that it is nearly impossible
to believe that they could belong to the same cognitive domain; therefore, based on this
argument, stone tools are unlikely to be informative at all about the origin and evolution
of grammar.
Like Davidson and Noble, Steven Mithen also used an alternative psychological
model to develop an archaeological theory of prehistoric cognition, one that could
address more than just spatial intelligence or language. In his seminal book, The
Prehistory of the Mind, Mithen (1996:11) highlighted the need for continuing exploration
into the minds of extinct hominins using archaeological evidence, stating, “It is time for a
‘cognitive archaeology.’” The popular field of evolutionary psychology heavily
influenced his theoretical position on the prehistory of the mind. Evolutionary
psychologists viewed the mind as a product of biological evolution, or more specifically,
a product of natural selection and not a product of chance. Therefore, according to
evolutionary psychologists, the mind is modularized and comparable to a Swiss army
knife, where each tool represents a different module or intelligence, and the only way to
predict what kinds of tools might be inside that Swiss army knife is to know what kinds
47
of problems Pleistocene hominins faced. While quite taken with the idea of an
encapsulated mind, Mithen found this analogy unsuitable for the creativity and fluidity of
the modern human mind. Rather, he borrowed cognitive scientist Dan Sperber’s (1994)
idea of a “strictly modular but also a highly creative modern mind” (Mithen 1996:59),
which has a specialized ‘module of metarepresentation’ that acts as a clearing house
through which new ideas must pass before they find a home, but they may also continue
to return to this clearing house even after finding a home. This allows for creativity
because different modules, which normally would have little to no interaction, can now
be mixed with other modules in the clearing house, which can create “all sorts of
mischief” (Mithen 1996:60), or in other words, allow for creative thoughts.
By observing human development, Mithen (1996) claimed that the human mind
passes through three architectural stages of development: 1) general intelligence, which
involves general purpose learning and decision-making rules; 2) general intelligence
supplemented by specialized intelligences that are devoted to specific domains and work
in isolation of each other; and 3) the working together of multiple specialized
intelligences, allowing for the flow of knowledge and ideas between behavioral domains.
These specialized intelligences include technical, social, natural history, and perhaps
linguistic.
As one might expect, Mithen applied these same developmental stages to
evolutionary development. Chimpanzees were used as a model for reconstructing the last
common ancestor between chimpanzees and humans at around 6 Ma. According to
Mithen, chimpanzees accomplish most tasks with general intelligence but also have a
developed social intelligence. He claimed that H. habilis set the foundations for technical
48
and natural history intelligences. The Oldowan tool industry surpasses the technological
capabilities of chimpanzees but lacks imposed form, and H. habilis was likely able to
develop hypotheses concerning resource location from visual clues but were limited to a
narrow environmental setting. H. erectus and H. neanderthalensis had advanced technical
intelligence, natural history intelligence, and social intelligence, according to Mithen.
Overall, he argued that there is a lack of evidence for the fluidity of ideas between these
domains. For example, their technologies, while complex, lack several elements,
including temporal and spatial variation, organic materials, such as bone or antler,
composite tools, and design for specific purposes, which indicate an inflexible adaptation
for toolmaking that did not allow any overlap into other social or natural history domains.
Language may have been present by this time; however, Mithen argued that it arose out
of the social domain and remained exclusively a ‘social language’ until the appearance of
modern humans. Cognitive fluidity, or the appearance of the metarepresentation module,
did not occur until after 60 Ka, at different times and in different parts of the world, and
language may have helped bridge the independent domains (Mithen 1996).
Mithen’s theory and supporting evidence, while intriguing, are riddled with
inaccuracies. Firstly, according to his logic, if chimpanzees have a developed social
intelligence, it should be difficult for them to combine the social and non-social realms.
He cites nutcracking as an example of this, as chimpanzees are not known to teach their
young how to do this task. There are several examples of tools being combined with
social displays, however, even at the time Mithen was writing, such as the leaf-clipping
display as an expressive gesture usually in relation to copulation solicitation (Nishida
1980), throwing sticks at conspecifics during play (van Lawick-Goodall 1968, 1970), and
49
throwing sticks, leaves, and stones as a part of male social display (Nishida 1970). These
behaviors undermine the idea of modules working in isolation of each other, and since
Mithen wrote this book, the idea that the brain can be neatly divided into functional
modules has largely been replaced by a model of the brain as a series of dense,
interconnecting networks (e.g. Markov et al. 2013).
Secondly, Mithen argued that modern cognition did not evolve until 60 Ka, and
even then at different times and in different regions. His model is unparsimonious
because it must therefore assume a multiregional, independent evolutionary convergence
upon cognitive fluidity. Why should this be? It would seem more likely that at least a few
modern human populations living in relative isolation until recently would lack the same
cognitive fluidity as seen in other populations, but this certainly is not the case. The
multiregional problem weakened his case for a recent evolution of modern cognition.
Lastly, it is difficult to imagine Mithen’s scenario in which there was no
interaction between social and technical intelligences in the case of learning the Levallois
method, which is generally agreed upon to be one of the most difficult flintknapping
techniques to master. While it has been demonstrated that the Acheulian and perhaps the
Levallois techniques can be learned without spoken language (Ohnuma et al. 1997; Putt
et al. 2014b), even the participants who learned nonverbally in these experiments still
learned in a social environment. This was likely a critical factor because Morgan and
colleagues (2015) and Putt (in progress) have found a very limited success rate among
participants learning how to make simple stone tools and large core bifaces without some
type of social interaction. Based on the model he has presented, Mithen would likely
argue that the results of these latter studies only point to modern humans’ reliance on
50
cognitive fluidity between domains, which is why placing subjects in a non-social
learning environment would result in poorer performance at learning a technical task. In
other words, he would argue that modern humans are a poor analogue for Neandertals
and other early humans because modern humans cannot even begin to fathom how an
early human internally processed the world. This argument could only be made, however,
if Mithen’s definition of modern cognition and language indeed evolved in modern
humans alone. This is the danger of relying upon one psychological model to interpret the
whole of human evolution—one begins to fit whatever evidence can be found to only that
psychological model, seeing patterns that may not actually be occurring and ignoring
other evidence to the contrary, a sort of ad hoc fallacy.
By this time, some French archaeologists were interested in applying cognitive
explanations to ancient material culture. They relied on Leroi-Gourhan’s writings for
theory and emphasized the chaîne opératoire methodological approach for making
inferences about past cognition. One example is Nathan Schlanger’s (1996) refitting work
on Marjorie’s core, a Levallois core from the 250,000-year-old site of MaastrichtBelvédère in the Netherlands. He addressed the issue that certain ideas about the
cognition involved in using the Levallois method had come to be considered truisms, but
these common sense assumptions lacked any scientific backing. In this case, the
supposition that the Levallois method was predetermined and thus reveals cognitive
abilities like conceptualization, abstraction, intelligence, and language actually stemmed
from Bordes’ (1950) initial definition of the Levallois method as ‘predetermination by
special preparation.’ He went on to say, “the men of this time had a perfectly clear mental
image of the object to be made before they set about making it” (Bordes 1968:137). So,
51
this essentially became the mantra whenever discussing the Levallois. By refitting the
core to understand the sequence of actions and corresponding mental events, Schlanger
concluded that the knapper held a plan-like principle in mind in the form of a fluid,
enabling image, rather than the traditional idea that the knapper had a systematic plan.
This fluid image set the objective for the knapper to attain and also allowed him/her to
react to transformations of the rock as they occurred. Schlanger’s thorough analysis of the
refit of Marjorie’s core provided an excellent description of the chain of actions involved
in the Levallois technique but stopped short of addressing psychological theory. Wynn
and Coolidge (2004; Coolidge and Wynn 2009) would later describe Marjorie’s core and
the Levallois technique in general as a product of Neandertal expertise, suggestive of an
enhanced procedural long-term memory relative to earlier hominins but lacking in a
significant advancement in working memory capacity.
Up to this point, evolutionary cognitive archaeologists were in search of the
origins of the human mind in the ancient past, which is always likely to be a fruitless
endeavor on its own because the definition of the mind is ambiguous and inherently
human. This is because it was humans who conceived of the intangible idea of the mind
to describe their own sense of being. Therefore, the mind, which, in the Western world, is
generally thought to be the seat for consciousness, intelligence, and reasoning, is an
unfair social construct to measure mental abilities of other species, living or extinct,
because it is essentially an all-or-nothing definition of humanness. One either has a mind
or has not. This does not translate well to the piecemeal nature of evolution.
52
A shifting emphasis to the brain and cognition
By the late twentieth century, evolutionary cognitive archaeologists were strongly
influenced by the cognitive approach in psychology, which began to replace behaviorism
as the dominant school of thought in the 1950s and 60s (Pinker 2011). The cognitive
approach explores how thoughts are processed and how these processes affect human
behavior. What blossomed out of the cognitive sciences was the field of neuroscience,
which came into its own and established multiple, independent, corroborating techniques
to link the brain to certain behaviors, such as recording single nerve cells in animals,
applying cortical stimulation, linking deficient behaviors with damaged areas of the
brain, and observing brain activation patterns with neuroimaging techniques. It became
clear shortly before the turn of the millennium that archaeologists would need to embrace
these developments in neuroscience to study the evolution of the human brain and brainbehavior relationships, because previously, the only discussion of hominin brain
organization and the behaviors that could be inferred from it revolved around the natural
or artificial endocasts of fossil endocrania, which was problematic because of
disagreements within the field regarding how to interpret sulcal and gyral patterns (Falk
2012).
Mainly, paleoneurology was notorious for its contradictory findings. For example,
Dean Falk and Holloway argued back and forth on whether the brain endocasts of the
gracile and robust australopiths were more ape-like or more human-like in their
organization because of the location of a mere dimple that was thought to mark the
location of the lunate sulcus (Falk 1980a, 1985, 1989a; Holloway 1972, 1983a, 1984;
53
Holloway and Kimbel 1986). The major limitations of studying endocasts is that they
usually produce poor sulcal patterns that are easily misinterpreted (Falk 2012), and while
everyone could agree that Broca’s region was expanded in the endocasts of early
hominins, any claims about whether or not this meant they had language was purely
speculation. For example, this expanded region may simply reflect a tendency toward
right-handedness or a complex network involved in manual motor manipulation. Perhaps
when combined with other lines of evidence, however, such a case for language could be
made.
Based on skull asymmetry and petalias, or uneven cerebral hemispheres (Abler
1976; Holloway 1980, 1981; Holloway and De La Costelareymondie 1982), and the
enlargement of Broca’s area on the left hemisphere present on a myriad of hominin
endocasts, many scholars turned their attention to the probability that the lateralization of
hemispheres has a long evolutionary history, which in turn could indicate that those
behaviors that are lateralized in modern humans may have been present and lateralized in
early hominins as well. Some lateralized behaviors known from split-brain and lesion
studies include language, skilled motor function, and mental imagery (Falk 1987b). The
fact that both language and handedness are lateralized to the left hemisphere in the
majority of people led some scholars to propose a causal relationship between the two
(Bradshaw and Nettleton 1982). Paleoanthropologists then suggested that the
manufacture and use of tools might have been the spark or foundation for the
lateralization of handedness and language (Frost 1980).
Nicholas Toth (1985a) turned to the archaeological record to search for traces of
handedness. From experimental studies, he argued that right-handed individuals prefer a
54
clockwise rotation of stone cores during flaking, which is observable on the flakes. He
found this same pattern of right-handedness, to a certain degree, in stone tools from
Lower and Middle Pleistocene contexts. He concluded that the hominin brain was likely
already lateralized and becoming more specialized for different functions, such as
dextrality and language, by 1.4-1.9 Ma. Toth’s (1985a) findings were welcomed because
of their consistency with paleoneurological studies; however, several experiments since
then have tried to replicate his findings with little success (Patterson and Sollberger 1986;
Pobiner 1999; Rugg and Mullane 2001; Uomini 2001, 2006; Bargalló and Mosquera
2014; Ruck et al. 2015; Daniel et al. 2016).
The shift toward recognizing the importance of neuroscience to evolutionary
cognitive archaeology occurred in 1990, when scholars from a variety of disciplines,
including social and physical anthropology, archaeology, zoology, primatology,
neurology, psycholinguistics, and developmental psychology, convened in Cascais,
Portugal for the Wenner-Gren Foundation International Symposium #110 on Tools,
Language and Intelligence: Evolutionary Implications, with the objective to discuss the
possible relationships between tool use, language, and social behavior (Gibson 1991). An
edited volume (Gibson and Ingold 1993) resulted from this conference, which would set
the stage for the most important questions and approaches of evolutionary cognitive
archaeology in the twenty-first century. The conference concluded that complex
relationships do indeed exist between tool use, language, and social behavior, and an
interdisciplinary approach is essential to clearly define these relationships in order to
reconstruct the evolution of language and cognition.
55
A number of topics were discussed at this Wenner-Gren conference, including
chimpanzee capacities for syntax and tool use, the integral role of imitation in tool use
behaviors, and the hypothesis that a common circuitry in Broca’s area of the brain exists
that processes the hierarchization of both tool using and language, to name a few.
Considerable debate focused on whether or not chimpanzees were capable of producing
Oldowan tools, with Wynn (1993) and William McGrew (1993; see also Wynn and
McGrew 1989) arguing in favor and Toth and Schick (1993) against. It was already
known that a captive orangutan was able to make simple flakes from imitating a human
model (Wright 1972), but chimpanzees and bonobos had not yet been tested, nor were the
orangutan’s stone tools compared to actual Oldowan artifacts. Toth and colleagues (1993;
Schick et al. 1999) would later prove that bonobos are also capable of making
intentionally chipped stone tools but not at the skilled level of the Oldowan toolmakers
with their experiment involving the bonobo, Kanzi.
Additionally, the seminar participants discussed the role of language in the
transmission of tool techniques. Boesch’s (1993) observations of imitation in wild
chimpanzees during nut-cracking activities and Wynn’s (1991) anecdotes about human
apprentices learning to flintknap largely by imitation and trial-and-error emphasized the
potential importance of imitation over language for the transmission of tool use among
early hominins. On the other hand, language and social behavior facilitate tool use when
it requires large bodies of culturally accumulated knowledge and discussions of tool
using and toolmaking tasks, especially those that involve complex long-term planning
(Parker and Milbrath 1993). These considerations indicated that toolmaking and language
have a history of reciprocal causation, and Greenfield (1991) proposed that this reciprocal
56
relationship was caused by a common circuitry between tool use and language in Broca’s
area. This conference made it clear that there was a “need [for] more knowledge of the
neurology and development of human and ape cognition and cognitively-based
behaviors” (Gibson 1991:263). This push would come from the cognitive sciences,
particularly from Philip Lieberman and Patricia Greenfield, who both attended the
conference.
Lieberman is a cognitive scientist who spent much of his career on the topic of
language evolution. He was particularly interested in the relationship between speech,
cognition, and language. In his book, The Biology and Evolution of Language (1984), he
was adamantly opposed to Chomsky’s universal grammar and the idea that language
represents a unique module of the mind that can be localized to a small area of the left
hemisphere. Instead, he proposed that human language derives from general neural
mechanisms that structure other aspects of cognition, and these neural mechanisms are
homologous, and thus continuous, with those of other animals. Specifically, he argued
that the biological substrate for speech, with its special properties for the automatization
of rapid vocal communication, served as the preadaptation for language, which evolved
as small structural changes over a long period of time.
While Lieberman emphasized the close link between speech motor control and
language in the brain, Greenfield (1991) stressed the ontogenetic and phylogenetic
relationship between the hierarchical processing of language and object manipulation,
which questioned the long-held belief that language and object combination (including
tool use) are separate cognitive modules associated with distinct neural structures (Fodor
1983). She found that there is a synchronous, developmental progression toward
57
increasing hierarchical complexity in object combination and language among American
and Zincantecos children in Mexico. Similarly, Broca’s aphasics are also unsuccessful at
hierarchical organization. These lines of evidence pointed to the same neural structure for
processing hierarchical organization involved in tool use and language. She claimed that
evidence from ape behavior supports the idea that object combination and language
evolved together as well. Chimpanzees have difficulty combining two objects to
manipulate a third object and tend to only use combinations of two communicative
gestures sparingly. Greenfield suggested that the neural structure provided for
combinatorial manual activity served as a preadaptation for the combinatorial aspects of
language, which is a conclusion that had also been arrived at by other researchers
previously (Kimura 1979; Lieberman 1990; Reynolds 1976) and has continued to be one
of the dominant hypotheses for the evolution of language in the field of evolutionary
cognitive archaeology.
Controversies continued in the new century
The turn of the millennium brought some important changes for the direction of
the field. The cognitive approach to archaeology was becoming a more widely accepted
tool for interpreting the paleoanthropological record, largely because of the recent
explosion of research in cognitive science. For example, Stanley Ambrose, an
archaeologist who had once operated under the standard ecological model, began to
incorporate the cognitive approach in his research in order to explain the visible
developments evident in the fossil and archaeological records. Ambrose’s (2001) paper in
58
Science marked a turning point for the field of evolutionary cognitive archaeology in the
United States, as it matter-of-factly coalesced archaeological and paleontological
evidence with cognitive theory in an intuitive sequence of events that led to modern
humans.
This was also a turning point for Wynn, for even though he had spent around
fifteen years advocating for a Piagetian approach, he began his “fall from Piagetian
grace” when he recognized that it was too limited in the kinds of interpretations it could
support (Wynn 2002:428). He began exploring other cognitive approaches, which led
him to publish an article in Behavioral and Brain Sciences that focused on spatial
cognition as evidenced by artifactual symmetry and exposed evolutionary cognitive
archaeology to a much wider audience, in a sense, legitimizing the field’s aims and
methods.
It was also in 2000 when Wynn began collaborating with neuropsychologist
Frederick Coolidge. Together, they would incorporate the latest findings in psychology
and neuroscience to develop the ‘enhanced working memory’ model based on Baddeley’s
(1993, 2000, 2001) working memory system. This system involves a central executive
and two slave systems, the phonological loop and visuospatial sketchpad. The enhanced
working memory model proposed that one recent genetic mutation could be responsible
for an enhanced working memory that was then exapted by language (Coolidge and
Wynn 2001, 2005, 2009). They argued that flintknapping only involved the visuospatial
sketchpad, which maintains and integrates visual and spatial information. Because they
did not believe the phonological loop to be engaged during flintknapping, this means that
this technical activity would not involve the most enhanced version of working memory.
59
Rather, facilities, such as traps, and reliable tools like harpoons, which are usually
overworked to ensure functionality, require enhanced working memory. The evidence
indicates that this development in working memory only occurred around 80 Ka. While
Wynn still maintained that stone tools were not reliable indicators for linguistic patterns
with this new model, the timeframe he espoused for the evolution of modern cognition
shifted to a considerably later date than what he originally proposed when employing a
Piagetian framework. Coolidge and Wynn’s work would spur many other archaeologists
to consider the evolution of working memory as well, resulting in another Wenner-Gren
International Symposium #139 dedicated to evolutionary cognitive archaeology, a special
issue in Current Anthropology on “Working Memory: Beyond Language and
Symbolism” and a slew of papers in 2010 (Ambrose 2010; Belfer-Cohen and Hovers
2010; Davidson 2010b; Haidle 2010; Martín-Loeches 2010; Nowell 2010; Wadley
2010a, 2010b).
Around the same time, Dietrich Stout, a student of Toth and Schick, was
developing his own research program that would produce some of the most widely cited
papers in evolutionary cognitive archaeology and would spark the coining of a new term
for the application of neuroscience theory and methods to archaeological questions—
neuroarchaeology (Malafouris 2009; Stout and Hecht 2015). Stout, Toth, and Schick
were strongly influenced by the 1990 Wenner-Gren symposium in Cascais, Portugal.
Stout (2002:693) wrote that it was the “inspiration for many of the questions addressed
and perspectives adopted” in his first papers. Together, they sought to bring a more
empirical approach to the study of human cognitive evolution through largely
experimental and actualistic research. They answered Gibson’s (1991) call for a
60
neurological perspective on the cognitive evolution of humans in their pioneering pilot
study that introduced functional neuroimaging to archaeology (Stout et al. 2000). This
experiment, which used positron emission tomography (PET), was the first attempt to
explore the neurological underpinnings of the cognitive complexity required for the
production of simple stone tools. From this small-scale study, Stout went on to
investigate the neural correlates associated with ESA tool manufacture in several
significant papers that would have important implications for the evolution of cognition
and language (Stout and Chaminade 2007, 2012; Stout et al. 2008; Stout et al. 2011). His
neuroimaging work would inspire a number of other scholars to apply neuroimaging
techniques to archaeological questions (Uomini and Meyer 2013; Putt et al. 2014a;
Williams et al. 2014; Bell 2015). In addition to his work with brain imaging techniques,
Stout, while still under the tutelage of Toth, emphasized the need for actualistic models to
aid in the reconstruction of the cognition and behaviors of extinct hominin species. He
observed skilled Langda adze makers from Irian Jaya in their traditional social context,
during which he found that adze making required well-developed motor coordination and
accuracy, high-level strategic planning and conceptualization, and a lengthy
apprenticeship enmeshed in an elaborate cultural system of meanings to master these
skills (Stout 2002).
The cognitive and empirical approaches of Wynn and Stout, respectively, are
emphasized here because of the profound influence they have had on the kinds of
research conducted since the turn of the millennium. The field of evolutionary cognitive
archaeology has experienced a dramatic growth in the past couple decades, probably in
large part because of the wider audiences it has reached via respected journals and edited
61
volumes (e.g. de Beaune et al. 2009; Nowell and Davidson 2010). It would be nearly
impossible to summarize all the recent work that has accumulated in the past few years;
however, many of the important questions that occupied scholars in the past are still
important today. These questions are outlined below.
Does the Oldowan technocomplex represent a cognitive evolutionary leap or is it simply
an extension of the ape adaptive grade?
The debate over whether the Oldowan toolmakers were more human-like or more
ape-like has a long history. The paleoneurological debate over the lunate sulcus pattern of
australopiths is one example of this. Also, some researchers argued that the skill level
evident in Oldowan tools required mental capacities that surpassed those of modern apes
(Gowlett 1984; Toth and Schick 1993; Toth et al. 1993), and others stressed the
similarities between the behaviors inferred from Oldowan tools and the observable tool
use behaviors of modern apes (McGrew 1993; Wynn 1981; Wynn and McGrew 1989).
Recently, there have been several promising research avenues that have attempted to
answer this question, including archaeological investigations of nonhuman primate tool
assemblages (Haslam et al. 2009), continuing investigations into captive ape knapping
behaviors (Roffmann et al. 2012) and primate tool use in the wild (Byrne 2005), PET
scans of humans knapping (Stout et al. 2000; Stout and Chaminade 2007), and
experimental work with tool behaviors of human and nonhuman apes (Bril et al. 2012).
Together, these research directions have added considerably to the knowledge of tool use
behaviors of other primates and the level of cognitive sophistication necessary to make
62
Oldowan tools; however, there is still disagreement on how to interpret this new evidence
in relation to the cognition of the Oldowan toolmakers.
Among those who have argued more recently for ape-like cognition to be
attributed to the hominins who are responsible for the Oldowan tool industry are Byrne
(2004) and Wynn and colleagues (2011). By looking at skilled manual behavior in
addition to tool use in nonhuman apes, Byrne (2004) observes that many of the motor and
cognitive elements usually attributed to Oldowan toolmaking can also be found among
the great apes, including precision handling, accurately aimed and powerful blows,
bimanual role differentiation, sequential planning, hierarchical organization, and manual
lateralization. Wynn and his collaborators revisit his original claim from 1989 for the
behaviors associated with the Oldowan industry falling into the ape adaptive grade of tool
use behavior and cognition. Even with all the discoveries of additional Oldowan sites,
species now known to make and use tools, and the wider variety of tool types attributed
to other primate species, they remain steadfast in their original claim that the hierarchical
complexity necessary for producing Oldowan tools is no different from the procedures
employed by gorillas and chimpanzees in the wild. Overall, they claim that there are far
more similarities between the behaviors inferred from these early stone tools and
nonhuman ape tool use behaviors than differences.
On the other end, Toth and Schick (2009) and de la Torre (2010) argue that
biomechanical and cognitive restraints inhibit nonhuman apes from achieving the same
level of sophistication and skill that is seen in even the earliest Oldowan tools. In a
comparison of the flintknapping products made by captive-taught bonobos, the 2.6million-year-old artifacts from Gona, Ethiopia, and expert modern humans, Toth and
63
colleagues (2006) find that the bonobos seem to be the outgroup in terms of skill, while
the Gona tools either cluster with the human tools or are intermediate between the tools
produced by bonobos and expert humans. Similarly, de la Torre (2010) argues that by 2.6
Ma, a visible difference in motor control and cognition had evolved. This is based on his
interpretation of the artifacts from the early Oldowan sites of Gona, Lokalalei 2C, and
Omo as being demonstrations of a sophisticated understanding of conchoidal fracture
mechanics, as well as showing evidence for planning, flexibility, and problem-solving.
If the behaviors entailed in knapping Oldowan tools are indeed qualitatively
different from the capacities of modern, nonhuman apes, and presumably the last
common ancestor, then one should assume that the sudden complexity at 2.6 Ma in the
archaeological record implies an evolutionary leap, not only in technology, but also in
cognition, which defies traditional ideas of Darwinian gradualism. This has led several
researchers to propose that hominins had been modifying stone tools prior to 2.6 Ma and
that older artifacts will be found that represent a previous technological phase when
hominins recognized the benefits of sharp tools but were unaware of the mechanisms of
knapping (Dennell, 1998; Panger et al., 2002; Semaw et al., 1997; Putt 2015). Because of
the similarities between the hammer-and-anvil technique used by chimpanzees for nut
cracking and knapping, one popular origin story for the knapping technique is through the
accidental production of flakes during nut cracking activities (Boesch and Boesch 1993;
Davidson and McGrew 2005; Marchant and McGrew 2005; but see Bril et al. 2012; de la
Torre 2010). Others have predicted bipolar and throwing techniques as intermediate steps
for hominins to break rocks prior to the advent of the Oldowan and knapping techniques
(Haslam et al. 2009; Putt 2015). Recently, speculations for an earlier origin of chipped
64
stone technology were proven correct with the discovery of stone tools at the Lomewki
site at West Turkana, Kenya that are dated to 3.3 Ma (Harmand et al. 2015, but see
Domínguez-Rodrigo and Alcalá 2016). Perhaps unsurprisingly, these early stone tools are
thought to have been fractured by less controlled methods such as bipolar and “passive
hammer,” which involves striking a core against an anvil. The large size of the tools and
these uncommon methods of flaking warrant a new name for this industry, the
Lomekwian. This new discovery incites new questions on hominin cognition during the
Pliocene, which will likely provide further fuel for the debate over the cognition of
Oldowan hominins.
Are stone tools informative traces for the evolution of language?
The attitude toward stone tools as reliable indicators of language has vacillated
over the past four decades. The current attitude seems to be in favor of launching new
methods that explore the evolutionary relationship between language and stone tool
use/manufacture. Investigations continue into the potential evolutionary relationship
between speech and language and population-level right-hand dominance. Steele and
Uomini (2009) provide a nice summary of the evidence for and against the Homo loquens
(Talking Man) and Homo faber (Toolmaking Man) scenarios. The first scenario stems
from the observation that there are more people who are left-lateralized for speech
processing than there are people who are right-handed. It predicts that handedness is a
consequence of hemispheric specialization for language processing. This means that any
evidence discovered for right-handedness in the archaeological record would be a good
65
diagnostic marker for language. Alternatively, the second scenario predicts the
opposite—that language developed out of pre-existing adaptations for manual praxis.
They conclude that current evidence supports the Homo loquens scenario because
conclusive evidence has not presented itself for population-level right-handedness in toolusing tasks among the other great apes or small-brained early hominins, which is what
would be expected if the praxis network provided the preadaptation for language.
Because of indirect evidence for right-handedness from the orientation of debitage scatter
and striations on teeth from lateralized eating activities, as well as a lack of evidence for
population-level right-handedness prior to 0.5 Ma, Steele and Uomini argue that language
probably evolved after this date. It should be noted, however, that there has yet to be an
agreed upon method of determining handedness from ESA tool assemblages (Ruck et al.
2015); thus, it is too early to make any hard conclusions regarding which came first.
Despite some of the criticism of Holloway’s (1969) original thesis on the common
cognitive structures shared by language and toolmaking, there has been a resurgence of
support for this idea, thanks in large part to Greenfield’s (1991) study on the ontogeny of
hierarchical processing of language and object manipulation in young children. For
example, Mark Moore (2010, 2011), a student of Iain Davidson, explores the design
space of stone flaking by following Greenfield’s model. He hypothesizes that early stone
flaking should reflect the evolutionary development of subassemblies of increasing
complexity. He demonstrates that knapping can be organized into a tree structure of
motor actions, with the smallest divisible elements of flaking combining to form flake
units, which can thence be combined into flake unit assemblies. Mahaney (2014a) more
explicitly compares Moore’s flaking design space model with language, finding that the
66
organization of action units involved in the thinning phase of Acheulian handaxe
production is as complex as many spoken language utterances. Stout’s work has focused
on the neuroanatomical evidence for the proposed evolutionary relationship between
language and stone toolmaking. He posits that a functional and anatomical overlap for
language processing and toolmaking exists at the site of the anterior ventral premotor
cortex (Stout and Chaminade 2009; Stout 2010).
Is modern cognition the result of slow, gradual evolution or a saltational revolution?
This question is essentially an extension of the behavioral modernity debate,
which has spurred disagreement over the origin and age of modern human behaviors. As
Henshilwood and Marean (2003:627) point out, “Rather than focusing on the
development of theory, many researchers have suggested behavioral traits that are
thought to be modern,” which has resulted in a Eurocentric interpretation of the
archaeological record. Several testable models have emerged for behavioral modernity,
however, such as a punctuated evolution model, a gradual evolution model, and a cultural
evolution model. The punctuated model assumes that some type of recent genetic
mutation occurred in the modern human lineage, leading to advanced cognitive
development that allowed for complex and symbolic behaviors (e.g. Klein 1995, 2009).
There are several variants of the punctuated evolution model, mainly addressing the
timing of modern cognitive development. For example, Klein (2001) supports a Late
Stone Age (LSA) appearance for modern behaviors less than 50 Ka. On the other hand,
Deacon (2001) argues for a punctuated event leading to modern human behavior
67
extending back to the Middle Stone Age (MSA). In stark contrast to these punctuated
hypotheses is the gradualist model, which assumes that the traits commonly associated
with behavioral modernity evolved in a piecemeal fashion (Renfrew 1996; Straus 2012).
Finally, the cultural model proposes that the cognitive prerequisites of modern behavior
were already in place in the ancestors of Neandertals and anatomically modern humans,
and there was a subsequent saltational cultural evolution in the modern human lineage
(Hovers and Belfer-Cohen 2006; d’Errico and Stringer 2011).
It seems only natural that evolutionary cognitive archaeologists should weigh in
on the behavioral modernity debate, as modern behavior is derived from modern
cognition. For the most part, evolutionary cognitive archaeologists agree that the
evolution of modern cognition was likely a gradual process, but perhaps a last stage of
cognitive evolution occurred after 100 Ka. According to Wynn and Coolidge (2011:10),
the “search for the evolution of modern cognition is a fool’s game. The components of
modern cognition, like the components of modern anatomy, evolved at different times for
different reasons.” In other words, the full complement of mental abilities that are
considered uniquely human cannot be attributed to a single mutation or evolutionary
event. Evolutionary cognitive archaeology is in the ideal position for positing
mechanisms for the cognitive evolution that may be the underlying cause for the visible
changes in symbol use, tool kits, landscape use, and other complex behaviors in the
archaeological record in the MSA and LSA. The main point of contention is whether this
mechanism was biological or cultural.
Those who support a biological explanation for behavioral modernity include
Wynn and Coolidge, Ambrose, and Davidson. Wynn and Coolidge (2011) propose that
68
an enhanced working memory may have been the final piece added to an already
advanced labyrinth of cognitive abilities, thus producing a finalized package of
recognizably modern human cognition after 100 Ka. The highly heritable property of
working memory lends itself well to the explanation that working memory underwent
strong, positive selection. Ambrose (2010) believes that the enhanced working memory
model encompasses too many cognitive domains that should be examined separately in
the context of cognitive evolution. He argues that constructive memory may be a better
measure of cognitive modernity, and it likely evolved before the human and Neandertal
lineage divided, as evidenced by composite tools. Davidson (2010a) adopts Barnard’s
(1985; Barnard et al. 2007) 9-subsystem interacting cognitive systems model for
interpreting the evolution of cognition through archaeological remains. He assumes that
the nine subsystems evolved gradually over time from a 6-subsystem model similar to
that of modern apes. Evidence from the archaeological record for the emergence of the
final ninth subsystem is anything that shows objectives that are remote from immediate
action, such as building watercraft, heat treating stones for future flaking, or hafting stone
tools. With all this considered, Davidson argues for fully modern cognition evolving in
the last 100,000 years. Other researchers have also applied Barnard’s system to their
interpretation of archaeological evidence for modern cognition (Wadley et al. 2009;
Wadley 2010a).
On the other end, Ian Tattersall (2008) and Francesco d’Errico and colleagues
(d’Errico et al. 2003; d’Errico and Stringer 2011) maintain that the neural architecture for
modern cognition had already emerged in the ancestor to modern humans and
Neandertals, and what ensued was a saltational cultural evolution in the human lineage.
69
There is evidence for a symbolic capacity among Neandertals (d’Errico et al. 2003).
d’Errico and Stringer (2011) interpret the appearance, disappearance, and reappearance of
complex technologies, engravings, pigments, personal ornaments, and burial practices
among different lineages between 160 Ka and 20 Ka to signify major discontinuities in
cultural transmission influenced by local conditions, not major biological changes.
Tattersall (2008) suggests that the mechanism that allowed for a cultural release of the
symbolic capacity was the invention of language.
How to move past the mistakes of the past
This review has traversed a long history of thought on the evolution of cognition
from an archaeological perspective. In many ways, evolutionary cognitive archaeology
has garnered respect as a modern cognitive science because of its use of prominent
theories and literature from the cognitive sciences and cutting-edge neuroimaging
techniques, while also being the only field that provides a much needed interpretation of
the only material evidence for past cognition. Evolutionary cognitive archaeology and its
current approaches have not been immune to critique. Specifically, concerns have been
raised about archaeologists’ preconceptions of cognitive evolution, the avoidance of
research on brain plasticity, and the general lack of consensus on the most suitable
methodology.
Marco Langbroek’s (2012) main criticism of evolutionary cognitive archaeology
is its use of a unilinear evolutionary model in a time when a branching phylogenetic tree
structure is the standard model for human evolution. He argues that the evolution of
70
cognition is still associated with a progressive line of development from primitive to
sophisticated. This problem is largely the result of the comparative samples used, the
tendency to see brain growth as a constant trend after 2 Ma, the use of Piagetian models,
and generally viewing modern human cognition as the pinnacle of cognitive evolution.
On this first point, the tendency to use modern, non-human apes as representatives of the
ancestral cognition condition and modern hunter-gatherers as representatives of
sophisticated cognition puts these two reference frames in opposition to each other,
which inherently creates a straight line of progression between them. Secondly, a direct
link has been made between brain size and the evolution of cognition (Dunbar 1998).
This increase in brain size is often perceived as linearly progressive, which translates to a
linear evolution of cognition as well. Thirdly, Piagetian models are founded on
development, which emphasizes the progression of cognition through a series of stages.
Langbroek (2012:9) is particularly critical of the preconceptions archaeologists bring to
their interpretations of the cognition of especially later dated hominins, stating, “…the
linear ladder model is structurally immune to testing against the archaeological record, as
it is a model where the structural preconceptions inherent to the model dictate
interpretations of data, instead of data dictating interpretations and the structure of the
model.” This has hampered discussions of Neandertal cognition, for example, because
Neandertal cognition is only compared to modern human cognition when in reality there
may have been two different types of complex cognition. There should be more focus on
identifying unique cognitive patterns in extinct hominins instead of only comparing them
to modern humans or great apes; otherwise, how far has evolutionary cognitive
archaeology really come from the unilinear model of Drummond (1894)?
71
What exactly shapes the brain and cognition? Evidence indicates that the brain is
a highly plastic organ. The brain is not simply shaped by genes, but as increasingly more
studies are demonstrating, by the external environment. Neuroplasticity allows the
developing brain, as well as the adult brain (Blakemore and Choudhury 2006), to modify
its genetically predetermined organization (Pascual-Leone et al. 2005). Amunts and
colleagues (1997), for example, find that prolonged experience playing a musical
instrument leads to enlargement of the motor cortex area that represents the hand.
Similarly, seasoned London taxi drivers have significantly larger hippocampi than control
subjects because of their extensive training and experience navigating the large city
(Maguire et al. 2000). If the brain is undergoing structural change throughout one’s
lifetime as a result of external events, what does this mean for cognition?
The idea that mental processes and the mind are not isolated to the brain alone but
rather extend beyond the brain to include aspects of the external environment and one’s
interactions with that environment is known as extended cognition. In other words, the
mind is embodied, meaning that mental processes are shaped by the body’s sensations
and actions within an interactive environment (Rowlands 2010). Lambros Malafouris
(2004, 2008, 2013) extends this idea to interpretations of the archaeological record and
the evolution of the mind with his material engagement theory. By using the example of
the blind man’s stick (Merleau-Ponty 1962), Malafouris (2008) describes how any
artifact can become an extension of one’s perception and cognition. A blind man’s
walking stick is essentially an extension of his arm. It ceases to be perceived by the man
but instead informs him of incoming sensory information from the external environment,
and consequently allows him to interact with that environment. It is thus also an
72
extension of his perceiving mind because if one were to remove a blind man from his
walking stick, with which he has learned to “see” the world by using it to interact with
the environment, one would effectively be removing or damaging a part of the man’s
brain because this would remove one of his perceptual systems. According to Malafouris’
hypothesis, the assumption that brain anatomy and structure stayed the same after any
major speciation event is not sustainable because it ignores a critical part of normal brain
functioning, those processes of neuroplasticity. The observed changes in the material
record should be seen as indicative of possible plastic effects, rather than cognitive or
genetic change. He argues that the approach of most cognitive archaeologists has been
hampered in that they assume material culture is the product of cognitive change, when in
reality, material culture may be one of the reasons behind cognitive change. For example,
the 76,000-year-old shell beads at Blombos Cave, rather than being manifestations of the
symbolic cognition of their makers, may have actually triggered the emergence of selfawareness because the beads as body ornaments changed the wearer’s perception of
his/her body image and mediated the transformation of the bodily self to a self perceived
by the other (Malafouris 2008).
As Wynn (2009:145) remarks, “Evolutionary cognitive archaeology remains a
largely inchoate amalgam of approaches…an eclectic array of interests, methods, and
theories… They do share one common thread: The conviction that prehistoric minds
structured prehistoric action, and that archaeology has access, albeit limited to those
minds.” Even this last point, that everyone agrees that prehistoric minds were structured
by prehistoric action, may not be entirely accurate, as Malafouris (2008) has
demonstrated in his argument that prehistoric interactions with objects and persons in the
73
environment structured prehistoric minds. So, what binds the eclectic array of research
that is considered evolutionary cognitive archaeology? Currently, it is an interest in
revealing the structure of prehistoric minds through the objects and landscapes with
which they interacted. This definition leaves much to be desired, however.
With the increasing number of scholars who are publishing in evolutionary
cognitive archaeology, it becomes more important than ever for there to be agreed upon
methods so that going forward, evolutionary cognitive archaeology can be a systematic
study of past cognition, where researchers have a set of shared, established methods.
Without these shared, established methods, datasets cannot be compared or replicated,
which is critical when dealing with archaeological sites that may be separated in time by
hundreds of thousands of years or by thousands of miles of distance. Otherwise, each
researcher essentially works within his/her own isolated bubble, bouncing off other
scholars’ bubbles instead of coalescing, resulting in multiple stories that are difficult to
interweave because of their methodological incompatibilities. Interpretations of past
cognition often pick and choose evidence from the archaeological record that fits one’s
theory, leading to the fabrication of scenarios that, while intriguing, are not falsifiable or
scientifically rigorous.
Traditionally, the two methods that have been used the most often by evolutionary
cognitive archaeologists are those that mainly involve interpretations of the final products
of prehistoric activity and those that have focused on a chaîne opératoire approach, or
reconstructing sequences of action (Wynn 2009). There are more weaknesses than
strengths with this first method. It is problematic because of equifinality, the idea that the
same end could have been reached by several different means; thus, archaeologists have
74
to assume the simplest procedure when interpreting final products from the
archaeological record. Assuming minimum competence on the toolmaker’s part,
however, risks underestimating actual abilities. Other pitfalls of this method include the
finished artifact fallacy and the still contentious matter of whether or not hominins
possessed a mental template. The chaîne opératoire approach is not plagued by these
problems and is a more objective means for interpreting prehistoric cognitive abilities
than what can be achieved by focusing on end products alone. But at this point, there is
still no consensus on how to describe and quantify action sequences.
Overall, the greatest methodological weakness of evolutionary cognitive
archaeology has been the near absence of hypothesis-driven experimentation in the work
conducted to date. The archaeological record is incomplete. Archaeologists are often
forced to work with small sample sizes. And once an excavation is complete, it cannot be
repeated. All of this can lead to subjectivity, misinterpretations, and an unclear picture of
the past. Experimental archaeology as a scientific methodology emphasizes research
design and procedure that adhere to the scientific method to ensure that results are
replicable and thus potentially falsifiable; data can be completely recovered and
resampled; and alternate possibilities can be explored (Marsh and Ferguson 2010). Longheld assumptions based on considerations of the archaeological record alone can be
quickly overturned by soundly derived experimental data. Evolutionary cognitive
archaeologists have relied on the cognitive sciences immensely for their interpretations of
cognition in the past, which is why it is odd that so few attempts have been made to
utilize the primary method of cognitive scientists: controlled experimentation. Going
forward, Wynn (2009) has predicted that an understanding of the brain using methods
75
borrowed from cognitive neuroscience will likely be what unites evolutionary cognitive
archaeologists in a coherent discipline in the near future. This will only be possible if
more evolutionary cognitive archaeologists embrace experimental methods.
Summary
While one would normally not think of an archaeologist as being a neuroscientist,
as it turns out, archaeology is the only discipline that allows one to study the evolution of
cognition in the past using material evidence. This approach to the study of the mind was
only possible once Darwin’s explicit materialist theory of natural selection was
introduced and the true antiquity of artifacts in the archaeological record was recognized
in the late nineteenth century.
The year 1979 marked a birth of contemporary evolutionary cognitive
archaeology when two papers were published independently that converged upon a
Piagetian epistemological framework in an attempt to investigate the intelligence of early
hominins. The search for the origin of the human mind continued in the 1980s and 90s,
though researchers tended to disagree on many important questions, such as how humanlike australopith cognition was, whether linguistic capacities left any traces on artifacts,
and when “modern” cognition first appeared. These themes are still disputed today, and
the lack of agreement likely stems from the discordant methodologies evolutionary
cognitive archaeologists continue to employ. By the end of the twentieth century,
archaeologists took note of the cutting edge research methods and models of the
cognitive sciences, especially those of neuroscience, which was making exciting new
76
discoveries in the area of brain and behavior relationships. At the turn of the millennium,
evolutionary cognitive archaeology moved away from the study of the mind, a Western
social construct, and embraced these developments in neuroscience to study the evolution
of the human brain and brain-behavior relationships. As for the future of the field, Wynn
(2009:146) states, “It is in the domain of cognitive neuroscience that [evolutionary
cognitive archaeology] appears most likely to coalesce into a coherent discipline.”
What evolutionary cognitive archaeology has revealed so far is that human
cognition likely evolved in a mosaic fashion, with distinct cognitive abilities evolving at
different times in prehistory (Wynn 2009). And instead of simply adding yet another spin
on the same paleoanthropological scheme, research in this area has demonstrated that
important developments in cognition did not all occur at the transition from the MP to the
UP or with the appearance of a new species. Just like humans’ bony anatomy, the brain’s
current organization and resultant behaviors evolved slowly, with some elements of
human cognition appearing earlier than others. Much is still unclear about the evolution
of human cognition. Going forward, hypothesis-driven, experimental, archaeological
research, especially in the realm of neuroscience appears to be the most promising route
to investigate cognition in the past in order to get closer to answering those questions so
essential to understanding what it means to be human.
77
CHAPTER 3:
THE HUMAN WORKING MEMORY SYSTEM AND ITS EVOLUTION
Introduction
Chapter 2 explored the history of evolutionary cognitive archaeology and
some of the major theories on the evolution of cognition that have been proposed
based on evidence from the cognitive and archaeological sciences. One influential
theory proposed that a recent genetic mutation that brought about an enhanced
working memory system in H. sapiens is responsible for modern human intelligence,
creativity, and perhaps even language (Coolidge and Wynn 2005, 2009; Wynn and
Coolidge 2011). Working memory is a domain-general subsystem of the mind that
allows for one to activate and sustain a set of mental representations for further
manipulation and processing (Carruthers 2013). It allows for one to perform complex,
cognitive tasks, such as reasoning, comprehension, and learning (Baddeley 2010) by
keeping active representations in attention7 but being flexible enough to allow new
items to enter into attention or to discard irrelevant items (Biedowski et al. 2009).
Representations in working memory can be held in active state for as long as active
attention is directed toward them. In this way, working memory is dependent on
attention to operate. Could Coolidge and Wynn be correct in their hypothesis that the
evolution of an enhanced working memory system was the key to the development of
modern human cognition?
7
Attention is defined as a cognitive process that selectively concentrates on specific aspects of
information while ignoring other aspects of information (Anderson 2004).
78
For working memory to evolve, it must be heritable, and for natural selection
or other evolutionary forces to play a role in its evolution, then there must be
variation between individuals in their working memory capacity. Twin studies
demonstrate that working memory task performance measures are indeed heritable
phenotypes (Ando et al. 2001; Luciana et al. 2001; Polderman et al. 2006). They also
establish that the brain structures involved in working memory and the neural activity
associated with working memory tasks are highly heritable (Schmitt et al. 2007;
Blokland et al. 2011). Working memory capacity varies between individuals (Just et
al. 1992; Vogel and Machizawa 2004), and these differences between individuals in
working memory capacity also help explain differences in general fluid intelligence.
Kane and colleagues (2005) report a strong correlation between working memory
capacity and general fluid intelligence. It is therefore crucial to learn how the working
memory system evolved in humans in order to unravel one of the biggest questions in
science, how human intelligence arose.
Despite the importance of this question, there has been relatively little
research dedicated to determining how and when working memory evolved in
humans, as compared to the amount of research that has gone into the evolution of
language, for example. By combining findings from relevant fields, however, such as
comparative primate psychology, developmental psychology, and archaeology, a
clearer picture emerges of the working memory system of the last common ancestor
of chimpanzees and humans and the modifications that must have occurred during its
trajectory from the last common ancestor version to its modern human version.
79
Background on working memory as a neural system
It became evident in the 1960s that short-term and long-term memory systems
were not sufficient to explain the mind’s ability to manipulate information in active
attention. Short-term memory is simply the temporary storage of small amounts of
information in the sensory cortices for a short period of time, usually lasting only two
seconds (Carruthers 2013). Items in short-term memory are generally unconscious,
unless attention is directed towards them before they expire, at which point they enter
working memory. Miller and co-authors (1960) were the first to coin the term
‘working memory’ in their classic book, Plans and the Structure of Behavior. This
term was later adopted by Atkinson and Shiffrin (1968) in their influential paper on
the control processes of a human memory system and by Baddeley and Hitch (1974)
for their multicomponent working memory model. Since this time, the concept of
working memory has become widely used, with thousands of articles dedicated to it.
The most widely accepted model for the operations of working memory is the
Baddeley model, which originally assumed that there were three components to
working memory: a central executive attentional control system, a visuospatial
sketchpad for short-term storage of visual material, and a phonological loop for the
short-term storage of verbal-acoustic material (Baddeley and Hitch 1974). A fourth
subsystem has since been added, called the episodic buffer, which can hold episodes
of multimodal information in temporary storage and allows for working memory to
interact with information from perception and long-term memory (Baddeley 2000,
2010).
80
A number of lesion, neurophysiological, and neuroimaging studies report
increased dorsolateral prefrontal cortex (dlPFC) activation during working memory
tasks in humans, and it is widely believed that the prefrontal cortex is the most
important substrate for working memory (Curtis and D’Esposito 2003). There has
been some debate over whether the dlPFC plays a prominent storage role in working
memory or if it plays the executive control role in a top-down relationship with other
storage regions, selecting and integrating relevant representations to use for the
planning and execution of goal-directed behaviors. There are likely different working
memory networks that include the dlPFC as the executive control and its connections
to premotor, parietal, and temporal cortices and some subcortical regions, which are
selected based on the goal or task at hand. This interpretation of the role of the
prefrontal cortex and subsidiary areas in working memory supports Baddeley’s
model.
A comparative approach to the evolution of working memory
Unfortunately, investigating the evolution of working memory can be as
problematic as investigating the evolution of language. Neither fossilizes, nor do they
leave behind many recognizable traces. Unlike language, however, other mammals
possess working memory. In fact, many studies reveal that non-human primates
possess a homologous working memory system to humans, albeit in a somewhat
more primitive form. The functional neuroanatomy of working memory in nonhuman primates is very similar to humans. Because working memory is a continuous
81
trait and operates in much the same way in the closest relatives to humans, this means
that a great ape baseline for working memory can be established, and the small,
evolutionary changes that inevitably occurred on the path to becoming human can be
identified. Perhaps a comparative approach may also help reveal the selective
pressures that led to an enhanced working memory system in humans.
Some of the paradigms that have been used to study working memory in
primates include immediate serial recall (Botvinick et al. 2009), matching to sample
tasks (Tavares and Tomaz 2002), and temporary monitoring of multiple responses
(Fagot and De Lillo 2011). The working memory capacity of the monkey species
tested thus far appears to be limited to two to three items in a sequence. Macaques,
for example, can recall the first three items in a sequence with fair accuracy, but the
amount of errors increases on the fourth item to the point that their selection is no
different from chance (Botvinick et al. 2009). The retention interval required to recall
these items ranged from four to eleven seconds, which falls within the domain of
working memory (Carruthers 2013). In a similar test, baboons were only capable of
consistently recalling two to three items, in contrast to humans, who consistently
recalled four to five items (Fagot and De Lillo 2011). While monkeys have trouble
recalling multiple items over a period of eleven seconds, Tavares and Tomaz (2002)
discovered that capuchins’ retention performance remains consistently above chance
across delays ranging from fifteen seconds to ten minutes when they only need to
remember two items.
The functional anatomy of the working memory system in monkeys is
homologous to that of humans. Functional neuroimaging and neurophysiological
82
studies report sustained activity in the prefrontal cortex of monkeys, especially the
dlPFC, during working memory tasks (Goldman-Rakic 1995; Fuster 2000; Petrides
2000; Carruthers 2013). A distributed network involving areas beyond the frontal
lobe illustrates that working memory involves parallel, distributed neuronal networks
in monkeys, as well as in humans (Constantinidis and Procyk 2004). Monkeys also
display a similar default system while at rest that involves higher-order association
areas, such as medial and lateral prefrontal and parietal areas, which suggests that,
similar to humans, monkeys may have internal thought processes occurring while at
rest (Kojima et al. 2009).
Compared to human and non-human apes, monkeys appear limited in some of
their executive functions. In a memory task that requires the subject to remember the
locations of hidden, matching images, Washburn and colleagues (2007) note the
tendency of macaques to perseverate to the point that they perform well below
chance. The macaques continue to make the same error repeatedly, dozens of times in
a row, and sometimes have to stop working altogether to break the cycle. Human and
non-human apes, on the other hand, have the ability to inhibit these perseverative
behaviors. Additionally, capuchins’ ability to match static images does not appear to
transfer to dynamic, visual stimuli or from the visual to auditory modality (D’Amato
et al. 1985). These examples demonstrate that while monkeys have a working
memory system very similar to that of humans, it is a more primitive system that was
elaborated upon in the great apes and later in hominins.
The working memory of great apes, especially chimpanzees, shares many
similarities with human working memory. Chimpanzees who participate in an object
83
displacement test, a task that requires the representation of a stimulus to be held in
active attention for more than two seconds in working memory, perform similarly to
human children (Barth and Call 2006). Chimpanzees demonstrate behaviors in
captivity (Osvath 2009; Osvath and Karvonen 2012), in experiments (Mulcahy and
Call 2006; Osvath and Osvath 2008), and in the wild (Sanz et al. 2004) that point to
their ability to mentally rehearse actions and long-term plan. They can make
inferences, reason by exclusion, recall events that they perceived in the past, and
imagine events that could happen in the future (Tomasello and Herrmann 2010).
With all these behaviors that require working memory that are similar to
humans, it should come as no surprise then that the functional anatomy of the
prefrontal cortex in chimpanzees shares many similarities with that of humans as
well. For example, chimpanzees and humans differ from macaques in their delayed
development of prefrontal white matter. By the time chimpanzees and humans reach
the late stage of pre-pubescence (6 years and 10.5 years, respectively), their prefrontal
white matter has not yet reached its full adult volume; this is unlike macaques, whose
prefrontal white matter reaches full adult volume by the time they are still juveniles at
1.9 years old (Sakai et al. 2011). A delay in the development of the prefrontal cortex
is thought to make it more susceptible to external influences. This suggests that
postnatal experiences are important for shaping the functional connections that form
between the prefrontal cortex and other regions of the brain in chimpanzees, similar
to humans.
While non-human great ape working memory is similar in many ways to
human working memory, there are some important differences. Great apes struggle to
84
perform better than chance in a matching task that requires them to locate hidden
image pairs, while humans excel at the task (Washburn et al. 2007), which suggests
that humans have more efficient “cross-talk” between memory systems that process
spatial and visual information than the other great apes. Similar to monkeys,
chimpanzees’ working memory capacity can handle only two or three items, which is
similar to the working memory capacity of a human child but does not rival that of an
adult human (Read 2008). Besides having a larger working memory capacity than the
other great apes, humans are also able to increase the number of items they hold in
working memory by representing information as chunks or groups of related items.
Other primates seem to represent each item individually, regardless of its relative
position or similarity to other items (Fagot and De Lillo 2011). A consequence of
having a working memory limit of two to three items is that chimpanzees are unable
to carry out recursive mental operations (Read 2008). Recursion is a key ability that
humans possess that allows for complex thinking, theory of mind, using a tool to
make a tool, and complex language, among other things. Moreover, many of the
genetic differences between chimpanzees and humans have been found to be directly
or indirectly related to working memory. For example, ASPM and MCPH1 are two
genes that are involved in neural cell proliferation and have undergone accelerated
evolution in the human lineage (Evans et al. 2004a; Evans et al. 2004b; MartínLoeches 2010). This is also the case for NRCAM, a gene that is involved in the
regulation of neuronal connections, and SIGLEC11, a gene that contributes to glial
expression and affects neural metabolism and synaptic processes (Hayakawa et al.
2005).
85
With all this in mind, it can be assumed that the last common ancestor shared
by humans and chimpanzees possessed a working memory capacity of two to three
items and a limited ability for hierarchical and recursive thinking; however, the last
common ancestor was capable of some planning into the future, recalling past events,
problem-solving and making inferences by way of mutual exclusivity, and was likely
able to inhibit perseverative behaviors in order to accomplish a task. By 5-6 Ma, the
lineages leading to Pan and Homo already possessed a sophisticated working memory
system and set of cognitive tools that could be tweaked over time to allow for a larger
working memory capacity, greater working memory stability in the face of
interference, more efficient information chunking mechanisms, and more flexible
rehearsal strategies in the case of humans. Elaboration upon a symbolic
representational medium that was likely present in the last common ancestor may also
have led to greater working memory efficiency. Washburn and co-workers (2007)
demonstrate that both humans and language-trained bonobos make fewer errors
remembering where items are hidden when those items are represented as symbols. In
this way, language may be a derivative of selection for an enhanced working memory
system. Selection for these behaviors would place intense evolutionary pressure on
the development of white matter in the prefrontal cortex to result in increased
connections between the prefrontal cortex and other parts of the brain and their
delayed development.
Chimpanzees have a more derived working memory system than monkeys;
thus, behaviors that chimpanzees and humans share that humans have elaborated
upon may hold the clue as to why working memory received such strong selection in
86
the hominin clade. Both chimpanzees and humans are extremely social apes that
make and use tools far more often and in a wider diversity of contexts than any other
primate species. Chimpanzees, and humans to an even greater degree, depend on
social relationships with individuals in their community to acquire technical skills
starting at a young age. With the right ecological incentives, early hominins may have
employed their flexible, technical skills to solve problems likely related to foraging.
As Haidle (2010) points out, the more flexible tool use employed by a species, the
more decisions must be made to find a solution dependent on an evaluation of the
problem, which would rely on working memory. In order to more easily gain access
to difficult-to-procure food items, selection for enhanced working memory related to
toolmaking cognition could have led to gradually more complex tool types that
required additional intermediate subgoals and vice versa.
A developmental approach to the evolution of working memory
While the development of working memory in human children should not be
assumed to be a direct recapitulation of the evolution of working memory in human
ancestors, the developmental stages that lead to adult working memory may, in some
ways, parallel the evolutionary stages that led to modern human working memory
because of the need for certain functions to build upon previously existing functions.
Developmental studies provide a clear picture of the behaviors and capabilities of
children at different points in their working memory development. They also describe
the neuroanatomical changes that occur in concert with the development of working
87
memory. The combination of primate comparative and developmental research on
working memory may help to shed light on what types of behaviors to search for in
the archaeological record so as to investigate how working memory evolved in the
past.
Even in infancy, there are signs of an operating working memory system. In a
delayed response task carried out by Reznick and colleagues (2004), twenty-six
infants between the ages of 4.5 and 6.5 months were enrolled to watch an
experimenter, standing behind one of two windows, disappear behind a curtain. They
were then distracted for one or two seconds before the curtains lifted to reveal empty
space in both windows. The infants’ eye movements were tracked to determine in
which window they expected to see the experimenter, based on their memory. By 5.5
months of age, infants demonstrate a tendency toward successful performance at this
task, indicative of the beginnings of a working memory system, and it is possible that
traces of working memory could be found in even younger infants with a shorter
delay period.
There remains the question of whether the working memory system that is
present early in development mirrors the structure of adult working memory. In other
words, is there evidence for a visuospatial sketchpad, phonological loop, and central
executive in young children? According to Gathercole and coworkers (2004), the
three main components of the Baddeley model for working memory are in place by
six years of age and remain fairly constant through childhood, except the capacity of
these components increases linearly from the age of four years into adolescence. For
example, visual working memory (VWM) capacity increases from two items
88
remembered in a change detection task at five to seven years old to four items
remembered at eleven years old (Riggs et al. 2006). Executive functioning appears to
be structured similarly in children and adults; however, the different components
develop at different rates (Gathercole et al. 2004; Henry 2012). For instance,
inhibition matures faster than planning (Henry 2012). Similarly, the phonological
loop develops later in childhood than the visuospatial sketchpad. Children who are
four to five years old tend to use visual codes to remember pictures, while older
children are more likely to use verbal coding (Henry 2012). This is because before
seven years of age, children do not have reliable, spontaneous rehearsal in the
phonological loop and therefore tend to rely on the visuospatial sketchpad to support
recall of physical forms of stimuli (Gathercole et al. 2004). After the age of seven
years, the phonological loop develops past the stage of simply acting as a
phonological store and improves memory performance by recoding visual inputs into
a phonological form via rehearsal. This demonstrates that the phonological loop and
visuospatial sketchpad are relatively independent of each other and separable at
different points in childhood (Gathercole et al. 2004).
The functional neuroanatomy of working memory is also very similar between
children and adults. Working memory tasks activate the same prefrontal, parietal, and
occipital areas (Klingberg et al. 2002). Although previous studies associated lower
activity in prefrontal areas with improved performance as a result of training (e.g.
Jenkins et al. 1994), Klingberg and coworkers (2002) find a positive correlation
between brain activity related to working memory and age, which suggests increased
involvement of these areas in working memory tasks as children develop. While
89
training in some tasks can lead to automaticity and decreased frontal activity, working
memory tasks do not become automatic with more training.
It is often assumed that there are few changes made to the working memory
system during adolescence because earlier studies that compared groups of
prepubescent children with young adults concluded that these functions reach
maturity in early adolescence (e.g. Welsh et al. 1991). This seems unlikely, however,
because changes in the brain that are linked to working memory continue to occur
throughout adolescence. The process of myelination is particularly pronounced in the
frontal lobe during adolescence, for example (Paus et al. 2001). Luciana and
coworkers (2005) conducted a study that compared a group of adolescents nine to
seventeen years old with a group of young adults eighteen to twenty years old on a
battery of spatial working memory tests. This study revealed that executive aspects of
spatial working memory continue to develop into adolescence, while other aspects do
indeed mature early, such as recognition memory, which reaches its adult capacity by
the age of nine years or earlier. A spatial self-ordered search task proved to be the
most demanding task administered to the participants in this study. This task
“requires response selection, memory, continuous updating of information, and a high
degree of executive control. It also demands self-monitoring and formulation of a
strategy, a task variable that has been shown to be specifically impaired in
neurological patients with frontal…lobe lesions” (Luciana et al. 2005:708). It is these
skills associated with this demanding task that are not fully developed until the age of
sixteen years.
90
The general timing for the development of certain aspects of working memory
and executive functions fits expectations based on the comparative primate data.
Many aspects of cognition that appear early in human development have a
phylogenetically ancient past, as is evidenced by their presence in non-human apes
and monkeys. A human infant’s working memory capacity of two items is
comparable to the two to three items that monkeys and apes can remember and
manipulate in their working memory. It is interesting to note that the phonological
loop matures later than the visuospatial sketchpad. Monkeys and apes’ phonological
loop component to their working memory system, if they indeed possess one, likely
only acts as a store for acoustic information and probably does not act to recode
visual stimuli into phonological form because they do not possess language and
therefore cannot mentally rehearse information to improve their memory capacity.
Instead, they rely on visual coding of information, similar to young children.
Inhibition, which matures slightly later, around ten years of age (Welsh et al. 1991), is
present in apes, but not monkeys. This combined evidence paints a clearer picture of
the last common ancestor shared by chimpanzees and humans and also highlights the
aspects of higher-order cognition, which do not mature in humans until adolescence,
that likely evolved after the split from chimpanzees, such as a larger working memory
capacity than two to three items, complex planning/strategizing, and mental rehearsal
of information in the phonological loop. It is then the evolutionary cognitive
archaeologist’s task to determine when, in the course of human prehistory, these
abilities evolved by finding archaeological evidence for these capacities.
91
An archaeological approach to the evolution of working memory
As was discussed in Chapter 2, Coolidge and Wynn (2005) introduced the
enhanced working memory hypothesis in an attempt to narrow down the timing of the
evolution of modern human cognition. Even though they remark that it is a fool’s
errand to search for the evolution of modern cognition because it was probably a
mosaic process, they argue that enhanced working memory was likely the final piece
(Wynn and Coolidge 2011). This hypothesis proposes that a recent genetic mutation
that occurred less than 100 thousand years ago (Ka) led to an increase in working
memory capacity in H. sapiens alone, which supported the construction of facilities,
reliable tools, and managed foraging that can be traced back to only ~80 Ka at the
earliest. They argue that stone tools, including Mousterian tools made by Neandertals,
require procedural cognition and long-term memory but not working memory. Only
once an enhanced working memory was in place were humans able to create art forms
like therianthropes, paintings or figurines that combine human and non-human animal
elements, because they require the artist to hold in attention the category of a type of
animal and combine it with the category of a person to create an abstract image that is
counter to reality. The earliest occurrence of therianthropes is not until the Upper
Paleolithic (UP). Thus, they assume that characteristically human working memory
rapidly evolved late in the hominin lineage. This would also mean that any tool
industries occurring prior to the acquisition of this genetic mutation that led to
enhanced working memory could have been learned and improved upon with
procedural cognition alone (Wynn and Coolidge 2004).
92
Many evolutionary cognitive archaeologists argue to the contrary, that instead
of one late-occurring genetic mutation in H. sapiens, there was likely a mosaic
emergence of different aspects of working memory over time (e.g., Davidson 2010a;
Martín-Loeches 2010). There is archaeological evidence in the Levant, Australia,
Europe, and Africa, for example, that support a lengthier, piecemeal evolution of
working memory and other executive functions. To assume that the complex
processes associated with making stone tools need only rely on expertise and
procedural memory also ignores the cognitive processes that were more than likely
needed during the earliest stages of learning these skills, a point that will be returned
to in Chapter 7.
Behaviors that are typically considered to be “modern,” such as burials,
pigments, personal ornaments, and engravings can be found in the Levant during the
Middle Paleolithic (MP). Belfer-Cohen and Hovers (2010) contend that the
occurrence of these behaviors reflects the mental capacity of MP humans for abstract
thinking, generalization, intentionality, and mental plasticity, all hallmarks of
enhanced working memory. If these behaviors do indicate that MP humans possessed
an enhanced working memory, then one would have to assume it arose much earlier
than the MP-UP transition.
The colonization of Australia and the artwork of the earliest Australians also
support an earlier evolution for enhanced working memory. Using the logic of
Coolidge and Wynn, Davidson (2010) argues that the colonization of Australia
around 60 Ka represents the presence of enhanced working memory because the
watercraft that were used to cross the ocean from Sundaland to Sahul were facilities.
93
As facilities, they would have required long-range temporal planning. The apparent
explosion of artwork in Europe around 40 Ka is often cited as evidence for the
transition to behavioral modernity in this region, which could imply a cognitive
revolution as well. Martín-Loeches (2010) notes that there may be other explanations
for this apparent imbalance between Europe and other parts of the Old World, such as
a demographic explosion in Europe or better preservation of artwork because of a
larger number of caves in the region. The sheer number of pieces of art in Europe
cannot hide the fact that similar types of art dated to the same time period can be
found in other parts of the world, such as Australia. And while therianthropes are
mainly found in Europe during the UP, their absence in other regions cannot hold up
as evidence for the absence of modern working memory abilities (Martín-Loeches
2010). If the capacities to produce facilities and artwork were already present in the
first Australian colonists, then the ancestors they shared with the modern, European
colonists in Africa over 100 Ka must also have possessed an enhanced working
memory.
Ambrose (2010) and Wadley (2010a) posit that composite tool manufacture
requires more than simply procedural memory, as Wynn and Coolidge (2004)
suggest. Through experimental replication of composite tools from Middle Stone Age
(MSA) sites in Africa, Wadley argues that the entire process is quite complex.
Creating the glue alone would require complex cognition because it requires the
addition of materials that seem counter-intuitive to the overall task. Ochre, for
example, is not an adhesive substance. Its inclusion in compound glues in the MSA
implies imagination and innovation on the toolmaker’s part. Assembling the tool by
94
combining the stone point to the wooden shaft would have required mental rotation
and planning into the future. And keeping in mind the goals and subgoals of each
subtask in relation to the larger end goal, the composite tool, implies an advanced
working memory capacity. The entire process requires at least four subtasks (locating
materials, knapping a stone point, creating the glue, and notching the shaft) to
complete the overall task, which exceeds the working memory capacity of all nonhuman primates. And because Neandertals in Europe are also known to have made
compound glues and composite tools (Koller et al. 2001), it can be argued that
Neandertals also possessed these same cognitive abilities. If evidence for enhanced
working memory can be found among anatomically modern humans in MSA Africa
and Neandertals in MP Europe, then it must be assumed that the last common
ancestor to H. sapiens and H. neanderthalensis also possessed an enhanced working
memory (Ambrose 2010), pushing its emergence back to at least 550-765 Ka, when
these two populations diverged from each other (Meyer et al. 2016).
While these researchers are most likely correct that the construction of
facilities and composite tools, artwork, burials, and other behaviors require advanced
working memory capabilities, they are only assumptions and disagreement will
continue until these behaviors are found to involve working memory areas of the
brain. Only by combining experimental archaeological methods with the
neuroimaging
techniques
of
cognitive
neuroscience
can
this
be
done.
Neuroarchaeological studies have yet to identify working memory areas in the
prefrontal cortex during the production of replicative Oldowan or Acheulian stone
tools (Stout and Chaminade 2007; Stout et al. 2008); therefore, it is still unknown
95
whether early stone tools require working memory at any stage of learning or if only
procedural memory is necessary to carry out these tasks, as is predicted by Wynn and
Coolidge. If it were discovered that the production of replicative Acheulian and
Oldowan stone tools results in differential activation of working memory areas, this
would imply that working memory has been evolving gradually for potentially
hundreds of thousands of years and likely cannot be attributed to one recent genetic
mutation.
Summary
Working memory is a system of the mind that processes and manipulates
domain-general information by holding it in active attention, which allows one to
perform complex tasks. Multiple lines of evidence demonstrate that the dlPFC is the
most important substrate for working memory, either as a prominent storage center or
as the executive control that has a top-down relationship with other storage regions.
Other animals possess working memory, including non-human primates. Working
memory capacity is heritable, variable between individuals, and positively correlates
with general fluid intelligence. For these reasons, it is possible that it evolved via
natural selection from a more primitive state to its current state. Because it does not
fossilize, we must use other lines of evidence to determine how human working
memory evolved.
This chapter enlisted comparative, developmental, and archaeological
approaches to determine what kind of working memory system the last common
96
ancestor of chimpanzees and humans possessed, what modifications were made to the
human system after this split, and when these modifications most likely occurred.
Monkeys and apes possess a homologous working memory system to that of humans,
though it is a more primitive system. Monkeys have the capacity to hold around two
items in active attention. They have difficulty controlling impulses. On the other
hand, non-human apes are able to hold in mind two or three items, inhibit
perseverative behaviors, plan into the near future, recall past events, problem-solve,
and make inferences. These differences between monkeys and apes may relate to
apes’ larger dependency on social relationships and skill acquisition at a young age.
Non-human apes are unable, however, to remember information as chunks like
humans. And both monkeys and apes appear to have less cross-talk between
subsystems compared to humans. The last common ancestor likely possessed a
working memory system similar to that of chimpanzees.
Developmental research supports this model because the working memory
system in early human childhood resembles that of monkeys and apes. Aspects of
more advanced working memory that develop only during adolescence, such as a
larger storage capacity, complex planning, and mental rehearsal in the phonological
loop extend past the abilities of modern apes, which suggests they evolved after
humans’ divergence from chimpanzees.
The timing of the evolution of modern human cognition remains a contentious
topic in archaeology, but evidence from several continents, including facilities and
artwork among the first inhabitants of Australia and composite tools in the MP/MSA
of Europe and Africa, support a gradual evolution of enhanced working memory that
97
took place over the course of the MP/MSA, or possibly earlier. This can be tested
further by mapping Stone Age behaviors to their neural correlates in the brain using
neuroimaging techniques to determine whether working memory areas play an active
role in technological behaviors that existed prior to the UP/Late Stone Age (LSA).
And because the same cognitive operations appear to be processed by the same
structures in the human and non-human primate brain, it is unparsimonious to assume
that the structures and functions of the modern human brain differ so drastically from
those of extinct hominin species that they can no longer be a useful analog for past
brain operations.
98
CHAPTER 4:
THE CONTRIBUTIONS OF NEUROARCHAEOLOGY TO HUMAN
LANGUAGE ORIGIN THEORIES
Introduction
Unlike working memory, language is a uniquely human behavior. As one of the
most complex behaviors in existence, unraveling the complexities of how language
evolved in humans occupies the time of many scholars from multiple disciplines.
Neuroarchaeologists are newcomers to this field of inquiry. As Bickerton (2007) notes in
his address to newcomers to the field of language evolution, there are several key
questions that have received much attention and debate. 1) Is language derived from
primate communication, or is it a unique development that only occurred in H. sapiens?
2) What was the main driving force for the evolution of language? 3) Did language
evolve gradually or abruptly? 4) How did symbolic units and syntax evolve? 5) Did
language begin as speech or as manual signs? This chapter highlights some of the recent
theories on language evolution that have been expounded by anthropologists,
evolutionary biologists, linguists, evolutionary psychologists, and neuroscientists.
Because of the multidisciplinary nature of the field thus far, there has been little
agreement over any of these questions among researchers. The chapter then focuses on
the unique, interdisciplinary perspective that neuroarchaeological studies, including the
study that is the focus of this dissertation, have to offer to the language evolution debate.
99
Clarifications
Before continuing with a discussion on the different approaches to studying the
evolution of language, there needs to be some clarification on what is even meant by this
term. Until very recently, the study of language evolution has been a multidisciplinary,
rather than an interdisciplinary, affair, meaning that there have been few attempts to
integrate methodologies between disciplines. Perhaps one of the reasons why there has
not been a more cohesive, interdisciplinary effort to tackle this problem in science is
because of researchers from different disciplines talking past each other by using the
same terms in different ways. ‘Language evolution’ can have multiple meanings. For
example, language evolution to a biologist might mean the changes to neural/cognitive
networks involved in the flexible and limitless communication of thoughts to others,
while to a linguist it could mean the change in word morphology, phonology, and syntax
of a language as a result of social pressures. Similarly, the term ‘language’ has many
definitions and is often misused. For example, sometimes an animal communication
system might be referred to as ‘language’ because of its complexity relative to other
animal systems, or instinctual human gestures will sometimes be called ‘body language.’
Clearly, it is important to provide definitions for some of the more ambiguous terms that
will be used throughout the rest of this thesis.
‘Language evolution’ can be interpreted in usually one of two ways, either as the
evolution of modern languages or the biological evolution of language. The former is the
product of culture, while the latter refers to evolutionary forces acting upon the complex
cognitive system underlying language (Fitch 2010). Although the difference between
100
these processes seems obvious today, in the past, historical linguists, such as August
Schleicher (1863), claimed that modern languages are living things and behave and
evolve as such. Chomsky (1986) addresses this issue by naming the languages that are
the products of populations of speakers (e.g. Mandarin, French, and Shona), “external” or
“E” languages and names the underlying neural system, “internal” or “I” language. He
also stresses the importance of focusing on I-language as the basic evolving unit, while
E-languages are simply the output of I-language. Nonetheless, the study of historical
language change remains an important field of inquiry, but to abate any confusion, the
term ‘glossogeny’ has been proposed to describe this epiphenomenon (Hurford 1990).
For the purposes of this dissertation, language evolution will refer to the biological
evolution or origin of I-language, not glossogeny.
All animals communicate, but only humans communicate via language.
Communication, thus, is any signaling system used by an animal to effect a reaction in
other conspecifics or individuals of different species. These signals can be simple or
complex, instinctual or learned; regardless, the scope and flexibility of such signals are
limited. Language can be defined in different ways depending on one’s background, but
Fitch’s (2010) definition will be used in this dissertation, as he has been one of the most
influential proponents of interdisciplinary cooperation in the field of language evolution
(e.g., Fitch 2002c, 2005; Hauser et al. 2002; Fitch et al. 2005). Language is a composite
system made up of multiple subcomponents that alone are insufficient to be called
language but together allow an individual to express any thought to other conspecifics
(who share the same E-language) via a set of mutually comprehensible signals (Fitch
2010).
101
Language is one of several communication systems that humans possess, but it is
different from these other communication systems in that it is discontinuous with animal
communication systems. The culturally universal meanings inferred from human facial
expressions and body movements (e.g., frowning, submissive posture) and the nonlinguistic vocal signal system (e.g., crying, laughter), among other signaling systems, are
continuous with the apes (Preuschoft and van Hooff 1995; Schmidt and Cohn 2001).
Continuity, in this case, refers to a feature of anatomy or behavior that is shared among
two or more living species (Fitch 2010). Because humans are the only extant species that
possesses language and there are no intermediate forms still surviving, it is discontinuous
with primate vocal or manual signals. This does not mean, however, that there is
necessarily an evolutionary discontinuity, or saltational change, between human language
and primate communication systems, as some have suggested (e.g., Bickerton 1990).
There are multiple pre-adaptations for language that primates display, which indicate that
human language and cognition likely evolved from Miocene ape behavior. Nevertheless,
arguments for an evolutionary discontinuity between human language and ape
communication are numerous among researchers from a variety of disciplines.
Few would argue that language is an indivisible whole. Language can be broken
down into multiple, independent subcomponents that have other functions in addition to
their role in language. There are multiple lines of evidence for this interpretation.
Children acquire certain aspects of language before other aspects. A lesion to a specific
area of the brain can cause a deficit in one or more aspects of language, but it will not
completely obliterate language in its entirety. In other words, the lesion may affect one or
more subcomponents of language, while it leaves other subcomponents intact. There are
102
obvious features of language that humans share with other animals, but the possession of
some language features does not mean that animals actually possess language.
Unfortunately, at this point, there is no list of language subcomponents that everyone
agrees upon, and researchers generally approach this problem in different ways.
Linguists, for the most part, emphasize subcomponents that are unique to language and
humans, while psychologists, biologists, and others emphasize the subcomponents of
language and cognition that are shared between humans and other animals. Hauser and
colleagues (2002) have deemed the former as the faculty of language in a narrow sense,
though much more research is needed to determine if there are mechanisms that are truly
unique to humans. The latter is called the faculty of language in the broad sense. There is
also a general consensus that because language is divisible into these multiple,
independently evolving mechanisms that there were likely protolanguage stages of
evolution before fully modern language developed, when certain combinations of
language features evolved in our hominin ancestors to the point that their communication
system was different from any known animal system. Thus, the term, ‘protolanguage’
will be used in this regard and will not refer to parent, or stem, languages of modern
languages (e.g., proto-Indo-European language), as it is sometimes used in glossogenetic
studies. The fact that language has so many subcomponents calls into question whether it
can even be considered a phenomenon, or singular event, and if one can really study the
evolution of language as a whole. At this point, it may be more realistic to study the
evolution of the individual pre-adaptations leading up to language, as has been the trend
more recently (e.g. Givón and Malle 2002; Hurford 2003).
103
A brief survey of the multidisciplinary approach to language evolution
To best demonstrate the multidisciplinary nature of research on language
evolution, the theories of some of the better-known scholars in their respective fields will
be discussed and compared. These theoreticians usually focus on one or more of
Bickerton’s key questions that were mentioned above. And it becomes apparent that even
scholars within the same discipline cannot agree on one theory.
Anthropology
Anthropologists, while eager to join the language origins conversation, represent
an eclectic group in their different approaches to the problem. Some anthropologists
emphasize language traces in the archaeological record; others focus on primate or
modern human behavior; and still others stress comparative anatomy. If everyone would
come to similar conclusions, then one could argue that this eclectic approach to studying
language evolution is a strength of anthropology, but this is not the case. Anthropologists
have not, to date, been able to agree on the functional evolution of language, out of which
modality it originated, whether it evolved gradually or abruptly, or if it is evolutionarily
continuous with primate communication or not. For example, while Philip Lieberman
(2003) argues that language evolved out of brain mechanisms for motor control, Terrence
Deacon (1997) hypothesizes that the driving force was the need for social contracts, as
meat became an indispensable part of the human diet. While Steven Mithen (2005) insists
that language evolved out of musical protolanguage, Gordon Hewes (1973, 1996) is a
104
strong proponent for a gestural origin for language. While William Noble and Iain
Davidson (1996) argue for an evolutionary discontinuity between human language and
primate communication, Robin Dunbar (1996) postulates that human language and
cognition are evolutionarily continuous with our primate ancestors. Overall, when it
comes to answering some of the key questions about the origins and evolution of
language, anthropology is a field divided. Focus will be placed on the research of
Davidson and Dunbar, as their work is representative of this divide in anthropology.
Iain Davidson is an archaeologist who maintains that certain aspects of language
leave traces in the archaeological record. He argues that human language developed very
late in human evolution and that there is an evolutionary discontinuity between the
language capabilities of hominins and their recent ape ancestors. Noble and Davidson
(1996; Davidson 2003) present a four-stage model for language evolution, which focuses
on two of Hockett and Ascher’s (1964) design features of language: productivity and
displacement. Productivity is the ability to create new messages by combining existing
signs, while displacement is the ability to refer to events that are not currently taking
place in the physical environment of the interlocutors. These features are the result of,
and indicate, the use of symbols; thus, the evolution of the capacity to use symbols is a
core emphasis in their works (Davidson and Noble 1989, 1993; Noble and Davidson
1996; Davidson 2003).
Based on the occurrence of certain types of archaeological artifacts, Davidson and
Noble speculate that symbolic communication did not evolve until after the appearance of
H. sapiens. They dismiss evidence for a possible link between stone tool manufacture and
language, which could potentially push the timing of language origins to pre-sapiens
105
Homo. The only archaeological traces that they believe can reliably imply symbolic
communication include man-made watercraft to transport modern humans to Australia,
ochre on the bodies at Blombos Cave, and depictive art; all these traces occurred less than
100 Ka. The first evolutionary stage Davidson (2003) proposes is similar to modern apes
in that early hominins communicated without language. Hominins entered the second
stage when they discovered or invented communication with indexical signs. Because of
evolving cognition and bipedal locomotion that freed up the hands, human ancestors were
able to visually-guide and control movement of their forelimbs, allowing them to indicate
silently by pointing to significant environmental events. This behavior would have had a
selective advantage over non-indexical communication because it would have improved
food foraging and the avoidance of predators by concealing one’s location (Davidson and
Noble 1993). Stage 3 involved working out the implications of communication with
symbols, or consciously learning that signs can have meanings, which could have been
discovered by leaving traces in a pliable medium, such as mud or sand. This would have
led them to realize signs’ malleability and the potential to refer to events not currently
happening. Finally, the fourth stage was the adoption of modern syntactical language.
Davidson and Noble (1989) posit that depictive art is necessarily embedded in a system
of shared meanings, which can only be expressed via syntactically-controlled language.
According to their hypothesis, a relatively recent and rapid evolution of language is
assumed which combined gesture, mimicry, and depiction.
Robin Dunbar, an anthropologist and evolutionary psychologist, takes a quite
different approach to the evolution of language than Davidson. While Davidson (2003) is
critical of anatomical evidence for speech or language capabilities, Dunbar has dedicated
106
much of his time and work to the anatomical evidence for early human group size,
cognitive ability, and language origins (Dunbar 1992, 1996, 1998, 2003). Also unlike
Davidson, Dunbar believes that language evolved in a piecemeal, or continuous, fashion
from ape communication and cognition. He goes on to criticize discontinuists, stating,
“The view that language evolved suddenly with no precursors seems to be based on a
naïve comparison between what humans, on one hand, and monkeys and apes, on the
other, do now, as though neither taxon had an evolutionary history” (Dunbar 1998:104).
In spite of this view, he concludes that language evolved around the same time as H.
sapiens based on his hypothesis that the ecological need for an increasing group size in
early humans required a new behavior, language in this case, for social grooming to
maintain group cohesion.
How language originated remains speculative, but Dunbar (1996, 1998) offers a
well-supported hypothesis that does not assume an anthropocentric evolutionary path. He
believes that it is clear that there was strong selection on language because the extra
neural hardware to support language is so calorically expensive. He builds upon the
‘social brain’ hypothesis that states that the large size of primate brains is a result of
social intelligence, or solving social problems (Brothers 1990; Barton and Dunbar 1997).
Additionally, the average neocortex size of a species correlates with average group size
(Dunbar 1992; Barton 1996). When humans’ neocortex size is included in this regression,
it is estimated that humans should live in groups of 150, whereas chimpanzees usually
have a maximum group size of sixty (Dunbar 1993). Hominins follow a smooth primate
curve; however, H. erectus and later hominins’ group size exceeds any modern primate
example. Other primates spend about 20% of their waking hours participating in social
107
grooming to form strong bonds with other individuals and maintain the cohesion of the
group. Because of a larger group size, humans would have to spend twice this amount of
time grooming if they were confined to the dyadic grooming behaviors present in all
other primates. This would surely reduce their amount of time to forage for food. Dunbar
reasons that language would offer a solution to this problem because one could “socially
groom” at least three other individuals at a time, while also advertising oneself, courting
prospective allies or mating partners, or inquiring about other prospective allies or
partners not currently present. This last behavior would provide a strong selection for
theory of mind, a crucial aspect of language. Aiello and Dunbar (1993) thus propose a
four-stage model for language evolution: 1) individuals kept track of each other through
contact calls similar to those of non-human primates; 2) once group size grew, which
could have been caused by the higher risks of predation when occupying a grassland
niche, the need to out-compete other hominin groups, or the need to maintain cohesive
relationships across a large territory to ensure ready access to water and food sources,
chorusing was adopted for social grooming; 3) to increase the efficiency of social
information transfer, grammatical structures developed; and 4) a fully modern language
developed. Thus, Dunbar’s account describes a gradual evolution of language out of
primate vocal calls and emphasizes the importance of theory of mind over syntax.
While anthropologists have not yet come to a consensus on most major questions
pertaining to language evolution, their continued participation is crucial for the
advancement of the field. It is quite common for researchers in other disciplines to make
large, sweeping generalizations about the archaeological record of stone tool use and the
behaviors of modern humans and other primates. For example, Bickerton (2002, 2003)
108
attempts to support his ‘catastrophic syntax’ idea, that syntax was a recent, rapid
development in H. sapiens, by referencing the stasis of the Acheulian for over one
million years and across geographic space as evidence for stagnant protolanguage and
cognition in H. erectus. This generalization is typical of someone who is unaware of the
Paleolithic literature, which is inundated with examples of Acheulian variation across
time and space (e.g. Movius 1948; Roe 1968; Wynn and Tierson 1990; White 2006;
Nowell and White 2010). Similarly, whereas most psychologists make universal claims
about the language behaviors of humans based on subjects from western, educated,
industrialized, rich, and democratic societies, whom Henrich and colleagues (2010) have
claimed are some of the “weirdest” people in the world, anthropologists recognize the
importance of including data from cultures and populations that do not fall under any of
these categories. Because of the inherent interdisciplinary nature of anthropology, it is
expected that anthropologists will play a key role in anthologizing the data from various
fields into a coherent theory of language evolution, in the future.
Evolutionary biology
Surprisingly few evolutionary biologists have formed theories on the evolution of
language. W. Tecumseh Fitch is the most notable and vocal biologist in the field of
language evolution. Like most biologists, Fitch is a proponent of a comparative approach,
especially when it comes to language. Several of his publications successfully debunk
supposed unique characteristics of human language by using broad comparisons to other
species (Fitch 2000a, 2002c, 2010; Fitch and Reby 2001; Hauser and Fitch 2003; Fitch
109
and Fritz 2006). Additionally, he points out the flaws in other protolanguage theories that
are based on cooperative information sharing explanations and offers his own ‘Mother
Tongues’ Hypothesis to address these problems (Fitch 2004, 2007).
Once a comparative approach is used to study the evolution of language, it is
nearly impossible to state that there is an evolutionary discontinuity between language
and animal communication and cognition, simply because most mechanisms that were
once thought to be unique to human language can be found in some capacity in one or
more other species. This is the main thesis of Fitch’s cumulative work. For example,
because speech is the main mode for the transmission of language, much attention has
centered on the apparent uniqueness of the human vocal apparatus. The larynx has a
permanently lower position in the neck in adults compared to other primates, which
enables humans to produce a variety of discriminable formants, allowing the vocal
intricacy to communicate complex linguistic concepts (Fitch 2000b). The initial descent
of the larynx also occurs by two years of age, around the same time that children begin to
speak (Lieberman et al. 2001). As a result, emphasis has been placed on the position of
the larynx in fossil hominins to infer their speech capabilities (Lieberman and Crelin
1971; Arensburg et al. 1990). Any claims against the possession of speech by another
species of hominin because of larynx position are faulty because many animals, including
primates, have a flexible larynx that retracts during vocalization; therefore, it is possible
that speech would have been achievable in species with a higher-positioned larynx (Fitch
2000b, 2002a). And as it turns out, the low position of the human larynx is not a unique
trait after all; deer, large cats, and even koalas have a permanently lower position of the
larynx (Fitch and Reby 2001; Hauser and Fitch 2003; Fitch 2005), which is thought to be
110
related to vocally communicating to conspecifics that one’s size is larger than it actually
is (Fitch and Reby 2001). Similarly, categorical perception was once thought to be a
unique adaptation for speech perception in humans, but this is not the case, either (Hauser
and Fitch 2003). Most “unique” mechanisms associated with human language are not
unique at all, but rather general sensory, motor, or cognitive mechanisms that are present
in other species.
Fitch argues that the explanation that language evolved for the “obvious” reason
of communication, as Bickerton (1990:156) puts it, is not an obvious explanation at all.
Why would it benefit an individual to share his/her valuable knowledge with someone
else? As Fitch (2004, 2007) explains, if information is being transferred to close kin, such
as one’s child, there is an indirect benefit to the sender because of inclusive fitness. This
explanation is not unique to the communication of humans. Honeybees will “dance” to
inform their hivemates of the location of nectar sources because they are all each other’s
closest relatives (Wilson 1975). Animals are more likely to sound an alarm call if
relatives are nearby. If alone or around non-relatives, they are much less likely to call
(Cheney and Seyfarth 1990). Fitch maintains that kin selection allowed for protolanguage
to evolve so that relatives could exchange complex information; however, it is clear that
language today is not used to share information with kin only. Reciprocal altruism would
have allowed non-related individuals to share information with each other once the
biological adaptations required for honest information sharing were already in place due
to kin selection. Fitch’s Mother Tongues Hypothesis addresses the functional necessity
for language and the continuity between primates and humans, but he neglects to discuss
how specific properties of language evolved, or, in the case of syntax, if it was a product
111
of biological evolution at all. Syntax has been claimed by some to be the result of cultural
evolution, not biological evolution (e.g. Bickerton 1990; Steels 1997; Tomasello 2003;
Arbib 2005).
Linguistics
Curiously, linguists have been reluctant to weigh in on the language evolution
problem until quite recently. As Newmeyer (2003) and Bickerton (2003) note, this is
similar to botanists shying away from discussing the origin of plants or biologists
avoiding the evolution of sex as a suitable research topic, which is certainly not the case
for either field. So, why have linguists been reluctant to join the conversation on the
origin of language? Bickerton (2003) blames the ordinance of the Linguistic Society of
Paris in 1867 that banned discussion on the topic, which led to a vacuum to be filled by
researchers in other fields. Newmeyer (2003) claims that linguists’ reluctance may be due
to the many competing theories and incomplete understanding on what the language
faculty actually is. He also argues that Noam Chomsky (1975) discouraged linguists from
speculating on language origins with his opposition to the idea that universal grammar
was a product of natural selection. Some notable linguists who have taken on the
challenge, despite these barriers, include Derek Bickerton and Ray Jackendoff. While
anthropologists generally focus on fossils and artifacts and biologists compare aspects of
language to animal behaviors and anatomy, linguists have tended to rely on “living
fossils” of modern human language to make inferences about the origin and subsequent
evolution of human language. Linguists focus most of their attention on the evolution of
112
syntax because of its unique complexity and tend to pay little attention to other aspects of
language.
Bickerton has been arguing for a discontinuous evolution of syntactical language
for more than two decades. In his book Language and Species, Bickerton (1990)
proposes a two-stage model for language evolution, an unstructured protolanguage stage
that probably occurred around two million years ago (Ma) with H. erectus and resembles
the linguistic communication abilities of language-instructed non-human apes, and a
‘catastrophic syntax’ stage, contemporaneous with the appearance of anatomically
modern H. sapiens. Because of his rejection of primate vocal calls as precursors for
human language, Bickerton argues that language evolved primarily as a system of
representation and secondarily for communication. He argues that symbolism arose
culturally instead of biologically in hominins because the last common ancestor that
humans shared with apes already had the biological equipment necessary for symbolism
(Bickerton 1990, 2003); therefore, he does not dedicate much discussion or ascribe much
importance to the initial transition to protolanguage. He does, however, emphasize which
elements were present during his protolanguage stage and which elements were necessary
for modern syntax to evolve.
For the most part, Bickerton (1990) rejects the idea that fossils and artifacts can
adequately reconstruct the language capabilities of earlier species. Instead, he argues that
possible living fossils of protolanguage exist in the language behaviors of humans and
non-human apes today. These living fossils include child language, pidgin and creole
languages, trained ape utterances, and the language behavior of Genie, an unfortunate girl
who was locked away with no human contact for most of her childhood. Genie and
113
trained apes, like Kanzi, Nim, and Koko, represent Bickerton’s protolanguage stage
because they are able to acquire and accurately use referential symbols but lack syntax.
He argues that children between the age of one and two have no grammar whatsoever and
thus rapidly acquire grammatical morphemes and phrase structure during their third year.
Similarly, the transition from a grammar-poor pidgin language to a syntactically-rich
creole is also rapid, occurring only in a couple generations. These lines of evidence, he
claims, give reason to suspect that there was also a rapid, or catastrophic, evolution of
syntax. Bickerton (1990, 2002, 2003) claims that the paleoanthropological record gives
credence to his two-stage model. H. erectus seems a likely candidate for protolanguage
because of its large degree of encephalization and the invention of the Acheulian
technocomplex. The supposed stagnation of stone tool technology in pre-sapiens Homo
and the rapid, elaborate toolmaking culture of H. sapiens indicates to Bickerton that the
evolution of language was a punctuated event that occurred only in the human species.
Ray Jackendoff (1999) expands upon Bickerton’s (1990) two-stage model of
language evolution, but his argument differs in several key ways. While Bickerton and
other discontinuists view syntax as an emergent phenomenon unique to humans,
Jackendoff argues that syntactical language evolved in incremental steps from the general
cognitive abilities of earlier primates. Jackendoff (1999:272) speculates that the driving
selective force for language was communication: “…linguistic adaptation arose first in
the interest of enhancing communication and secondarily in enhancing or refining
thought.” He reports that he has no justification for this claim and simply expects it to be
self-evident, that even a few words would be more beneficial than none at all.
114
Similar to Bickerton, Jackendoff (1999) focuses on aspects of modern human
language that he believes represent earlier stages of language evolution. The most
primitive aspect of human language that is directly comparable to the alarm calls of
primates is the use of situation-specific utterances associated with high affect, such as,
‘dammit!,’ ‘wow!,’ or ‘ouch!’ These exclamations have no syntax and are also preserved
in the most severe aphasics. They are often used non-communicatively and involuntarily;
however, there are also similar words in modern language that are situation-specific but
communicative in purpose and voluntary, such as, ‘shh!’ or ‘hey!’ These represent fossils
of the one-word stage, which gradually changed to allow speakers to use symbols nonreferentially. Young children, and language-instructed modern apes to a certain degree,
can refer to many different things with a single utterance. For example, a child who says,
“Kitty,” may be referring to the physical presence of a cat in the room, or the utterance
could have a variety of other meanings, such as, summoning the cat to come closer,
‘where is the kitty?’ or ‘this item looks like a kitty.’ After this stage, Jackendoff argues
that there needed to be two major inventions, a large lexicon and rudimentary grammar.
He turns to the language of Genie as his next living fossil. Genie learned many words
after she was rescued, but was never able to develop a complex grammar, indicating that
the lexicon and grammar are independent of each other. He stresses the importance of
syllables and how phonological structure was a major cognitive advance. The
combination of inherently meaningless phonemes is generative because it systematizes
existing vocabulary items and creates new ones. When adults learn any second language,
they universally demonstrate the following qualities: lexical competence, lack of verb
inflection, lack of relative clauses, and simple word order with the agent first and the
115
focus last. Jackendoff suggests that this pattern could reflect the protolanguage stage
prior to the invention of complex syntax. Finally, there were several intermediate stages
to complex syntax involving word order and the addition of function words, such as
adverbial phrases. As evidenced by his many stages of language evolution, Jackendoff
argues for a gradual evolution of language that is derived from primate calls; however, he
provides little discussion on why there was an adaptive need for any of his proposed
stages.
There are multiple issues with some linguists’ reliance on living fossils as
analogies for the language evolutionary stages of earlier species of human. Firstly, while
it is tempting to link the ontogenetic language stages of children directly to the
phylogenetic stages of language evolution, this logic is prone to flaws. The ontogeny of
humans does not always recapitulate aspects of human evolution from a potentially
chimpanzee-like ancestor (Fitch 2010). Children have evolved to learn language and are
constantly exposed to it before they even begin their one-word stage, which was not the
case for early hominins because they did not possess the genetic and neural mechanisms
for language readiness (Slobin 2002). Secondly, the proposal that the transition from
pidgin to creole represents a rapid evolution of syntax is even weaker because it relies on
the assumption that glossogeny recapitulates phylogeny, an assumption with no
supporting evidence. At this point, there is controversy over the definition of what a
pidgin language even is, to the point that Baker (1995, 1997) has argued that there is no
clear distinction between pidgins and creoles. Pidgins are derived from syntactical parent
languages, and the fact that the transition to creole is so rapid is simply the result of
children being language-ready (Botha 2006). For these reasons, the pidgin-creole
116
transition is a debatable analogy for protolanguage. Lastly, it is unwise to rely on one
isolated child case to make the claims that Bickerton (1990) and Jackendoff (1999) do.
There are historically many cases of “wild children” isolated from society, and their
language-learning behaviors varied widely. It is also difficult to separate out the
psychological effects caused by years of abuse in Genie’s case (Fitch 2010). At this point,
the original mode for language (speech, sign, or song) is still controversial, yet Bickerton
and Jackendoff take for granted that the original signaling system was lexical rather than
gestural without adequate explanation and instead focus on the transition from
protolanguage to modern language.
More recently, linguists and psychologists have extended their research platform
to include computer simulation and mathematical modeling to test language evolution
theories. Language evolution theories are difficult to test empirically and are often vague,
but computer simulation using mathematical modeling offer several advantages: 1) it
allows researchers to run virtual experiments that would be impossible to run in reality;
2) it tests the internal validities of theories and turns out predictions that can be used to
search for new empirical evidence from real data; and 3) it allows researchers to study
language as a hierarchically-organized complex system (Cangelosi and Parisi 2002).
Most simulation models deal with the emergence of syntax (e.g., Christiansen et al. 2002;
Kirby et al. 2008), but they have also been used to address the evolution of signaling
systems (e.g., de Boer 2002; Noble et al. 2002). Computer simulation experiments are
also prone to some limitations, which include the necessity for over-simplification,
subjective and arbitrary choices during the design of a model, and difficulty comparing
the results with real data (Cangelosi and Parisi 2002).
117
Evolutionary psychology
While most linguists avoid terms like ‘adaptation’ or ‘natural selection’ when
theorizing about the evolution of language, one can almost always expect an adaptive
explanation for language from an evolutionary psychologist. This is because they operate
under the assumption that natural selection is the only known process capable of
producing complex physiological or psychological mechanisms, such as language (Buss
1995). Evolutionary psychologists have pushed for the basic tools of behavioral ecology
to be applied to research on language evolution, but it remains a disparate,
multidisciplinary assortment of researchers who cannot even agree on what major
questions should be asked and addressed (Scott-Phillips 2010). Steven Pinker has been
one of the most influential evolutionary psychologists in the field of language evolution
because of his defense of language as an adaptation that evolved by means of natural
selection in Pinker and Bloom (1990). Michael Corballis is also representative of
evolutionary psychologists in that he too has argued for language as an adaptation. As has
been the case for the other disciplines, however, Pinker and Corballis disagree over the
evolutionary continuity of human language and primate communication. While Pinker
emphasizes the unique aspects of human behavior and language, Corballis recognizes the
similarities between primate and human gestural communication, which is also a key
characteristic of the mirror system hypothesis developed by neuroscientists Giacomo
Rizzolatti and Michael Arbib, to be discussed later.
Other researchers have borrowed aspects of Pinker and Bloom’s (1990) views on
language evolution, which defended the theory that language was an adaptation that
118
evolved by means of natural selection. This was an extremely influential publication that
set the groundwork for present interest in the topic (Scott-Phillips 2010). Since this
publication, Pinker (2003, 2010) has attempted to defend the idea that many of the unique
qualities of humans, including language, are the result of occupying a “cognitive niche.”
The cognitive niche implies a need to share information. Hominins entered this niche
because of their increased intelligence, sociality, and communication, which thus allowed
them to manipulate their environment instead of be manipulated by it. In turn, this led to
further adaptations, such as longer childhood and adolescence, increased lifespan,
complex sexual behavior, bi-parenting and multigenerational parenting, and the division
into cultures. Information became a commodity. If one could pass this information along
to kin, they could also benefit from such knowledge. Thus, kin selection is an important
aspect of Pinker’s argument, similar to Fitch’s (2004) Mother Tongues Hypothesis.
Pinker argues that complex syntax evolved for more effective communication, which is
similar to Jackendoff’s (1999) belief that communication would have been enough of a
driving force for language to evolve. How hominins invented the cognitive niche may be
impossible to know for sure, but Pinker hypothesizes that the combination of bipedalism
and prehensile hands was the first tipping point towards a selection for increased
intelligence because of the ability to manipulate tools and the environment. Additionally,
the omnivorous diet of early humans may have selected for cleverness to more effectively
obtain meat, and group living would have selected for social intelligence. He goes on to
hypothesize that language, intelligence, and increased sociality co-evolved in humans,
and kin selection eventually gave way to reciprocal altruism and cooperation between
non-kin.
119
Pinker (2010:8996) rejects non-adaptationist accounts of language evolution,
stating that because of the “undeniable practical advantages of reasoning, cooperation,
and communication, it seems superfluous, when explaining the evolution of human
mental mechanisms, to assign a primary role to macromutations, exaptation, runaway
sexual selection, group selection, memetics, complexity theory, cultural evolution…or
gene–culture coevolution.” The evidence he cites for language being a product of natural
selection includes the recent selective sweep of the human FOXP2 gene, the fact that
there are universal ontogenetic language stages in children, and the dissociation between
general intelligence and language abilities (Pinker 1994), all of which are debated.
Throughout most of his publications, Pinker emphasizes syntax as an “innate” behavior
of humans, but he does not address the origin of symbols in much detail, how these first
signals were transferred, nor if there was a protolanguage stage(s). It is unclear from his
argument whether human language evolved out of a primate communication system or
could only have emerged once hominins occupied a cognitive niche, but the latter is
suspected.
Michael Corballis, on the other hand, offers a theory much like Pinker’s in that
natural selection is the main evolutionary force responsible for language, but addresses
many of the points that are unclear in Pinker’s argument. Essentially, Corballis (2002,
2003a, 2003b) posits that language has deep roots: 1) it evolved gradually out of the
primate gestural system that is still present in modern non-human apes; 2) facial and
vocal gestures were added to manual signals; 3) as the vocal component increased,
language became more lateralized to the left hemisphere, thus leaving a majority of the
population right-handed; and 4) autonomous speech was invented after the modern vocal
120
apparatus was in place. The Gestural Origin of Language Hypothesis has been proposed
multiple times (e.g., Hewes 1973; Givón 1995; Armstrong 1999) and is gaining support
from researchers who study the brain (e.g., Rizzolatti and Arbib 1998).
Corballis (2002, 2003a, 2003b) presents a theory that draws upon interdisciplinary
lines of evidence. He notes that while primates certainly can vocalize and produce
different calls, their calls cannot be combined into sequences or be broken down into
interchangeable parts, and they seem to be involuntary most of the time, driven by
emotion. The gestures of primates, on the other hand, probably are evolutionarily
continuous with human speech and language because modern apes have been observed to
use more than thirty gestures in captivity and the wild, usually in reference to other
individuals that are present (Tanner and Byrne 1996; Tomasello et al. 1997). Apes’
gestures are voluntary, meaning that they will only communicate via gesture when they
know they are being watched, while vocal calls are generally not directed at anyone. The
last common ancestor probably possessed a gestural system similar to that of
chimpanzees and bonobos. The number of signs hominins were able to produce gradually
increased, and gestures incorporated a syntactical structure over time. This idea is
controversial because most of the researchers already discussed argue that syntax was a
cultural invention contemporaneous with the species H. sapiens, while Corballis sees
syntax evolving via natural selection over a period of about two million years.
To explain how gesture gradually transitioned to speech, Corballis looks to
modern sign language users, who often incorporate facial/mouth and head movements
while signing, which carry syntactical information. Adding mouth movements and vocal
sounds would have allowed the application of more than one meaning to an otherwise
121
identical gesture. Humans apparently are still influenced by facial gestures. For example,
there is the McGurk Effect, which occurs when an individual hears a phoneme such as
‘ba’ but sees the lips of someone saying, ‘fa’ and reports that ‘fa’ was actually said. Over
time, facial gestures began to retreat inside the mouth and throat. This idea treats speech
as the product of vocal gestures, rather than abstract phonemes (c.f., Studdert-Kennedy
1998). Corballis gives credit to speech for the left-hemisphere asymmetry of the brain
and right-handedness that are dominant in all human populations. This would push vocal
control back to early Homo because of evidence for left-hemisphere asymmetry in H.
habilis (Holloway 1983b) and possibly right-handed stone tool manufacture (Toth
1985a). Speech probably experienced strong positive selection because it freed up the
hands for other activities, such as toolmaking, but it had to wait for the vocal apparatus to
catch up, which may not have occurred until very recently.
Neuroscience
Traditionally, it was believed that the brain was made up of specifically located,
functional modules, largely in part due to the discovery that lesions to specific areas of
the brain result in specific behavioral deficits. Paul Broca (1861) was the first to
recognize that a behavior can be mapped to a specific locus of the brain when he
observed a lesion on the inferior frontal gyrus (IFG) of the left frontal cortex of a patient
with nonfluent aphasia. Broca’s claim that this area of the brain controlled for expressive
speech stirred other scientists to map functions to the cerebral cortex, including Carl
Wernicke (1874), who localized speech comprehension to the superior, posterior area of
122
the left temporal lobe. The classic Broca-Wernicke-Lichtheim-Geschwind model of
language processing consisted of speech production in the left IFG (Broca’s area), speech
perception in the posterior superior temporal gyrus (STG; Wernicke’s area), and semantic
processing localized to the inferior parietal cortex, all connected by a white fiber tract
called the arcuate fasciculus.
While Broca and Wernicke’s areas of the brain are still considered important
participating regions in speech production and comprehension respectively, there are
several problems with this classic model for language processing. Firstly, Broca’s aphasia
is not caused by damage to Broca’s area defined in the classical sense (Mohr et al. 1978;
Poeppel and Hickock 2004). The symptoms of Broca’s aphasia, agrammatisim, mutism,
and verbal stereotypes are only displayed when a lesion encompasses Broca’s area,
adjacent regions, and the insula (Mohr et al. 1978). Similarly, Wernicke’s aphasia is not
caused by damage to Wernicke’s area, defined classically as the posterior third of the
STG (Dronkers et al. 2000; Poeppel and Hickok 2004). Secondly, it is now recognized
that other parts of the brain also participate in language processing, for example the
supramarginal gyrus (SMG) of the parietal lobe (Homae et al. 2002; but see Binder et al.
1997), the left prefrontal cortex, and lateral and ventral temporal lobe regions (Binder et
al. 1997). Thirdly, Broca and Wernicke’s areas participate in other functions not related
to language. For example, damage to Wernicke’s area causes nonverbal auditory agnosia,
or the inability to comprehend nonverbal sounds (Saygin et al. 2010). There is also ample
evidence for the overlap of distal manual motor and domain-general functions in the
homolog of Broca’s area in non-human primates (Rizzolatti et al. 1981; Kurata and Tanji
1986; Rizzolatti et al. 1988; Hepp-Reymond et al. 1994; Taglialatela et al. 2008) and
123
Broca’s area in a broad sense in modern humans (Heiser et al. 2003; Higuchi et al. 2009;
Fedorenko et al. 2012). Finally, diffusion tensor imaging (DTI) studies reveal that the
arcuate fasciculus is not the only white fiber tract connecting Broca and Wernicke’s
areas to each other; there are actually multiple white fiber pathways that connect these
classic language areas to each other, as well as to other regions involved in languageprocessing (Friederici 2009). With all this considered, the language processing system is
better defined as a neural network with white fiber connections to many different regions
of the brain, and the classic language areas are simply a part of this network that also take
part in other functional neural networks. The domain-general networks in which the
classic language areas actively participate and non-language functional areas that overlap
with the classic language areas may impart important clues as to the evolutionary history
of language in the brain.
One such neural network that may be essential to both language and manual
praxis (i.e., motor planning) is the mirror neuron system. Mirror neurons, or neurons that
have both motor and sensory components, were first discovered by Rizzolatti and
colleagues (Rizzolatti et al. 1988; Gallese et al. 1996). They measured the activity of
individual neurons in area F5 (Brodmann area [BA] 6/BA 44) in macaques and noted that
some F5 neurons discharge during specific goal-directed manual motor acts performed by
the monkey and when the monkey observes a human or conspecific perform the same
acts. They argue that the function of a mirror neuron is to internally represent an action
for motor learning and to understand the intentions of the action. Some F5 mirror neurons
respond to audio input in addition to visual input (Kohler et al. 2002; Keysers et al.
2003). These audiovisual, or “echo” neurons (Tomasello 2002) discharge when the
124
monkey performs a hand action, observes someone else perform the same action, or even
when it hears a sound related to the action being carried out. Mirror neurons are also
found in the rostral portion of the inferior parietal lobule (IPL; Fogassi et al. 1998;
Gallese et al. 2002), which receives input from the superior temporal sulcus and sends
output to the ventral premotor cortex, including F5 (Rizzolatti and Craighero 2004).
Neurons in the superior temporal sulcus respond to observed actions done by others as
well (Perrett et al. 1989; Perret et al. 1990; Jellema et al. 2000), but because these cells do
not have any motor component, they cannot be considered mirror neurons. Therefore, the
rostral portion of the IPL and ventral premotor cortex (F5) make up the mirror neuron
circuit in macaques.
Mirror neurons are active during the executed and observed actions of the hand
and appear to identify and respond to the goal of a motor act. Originally, Gallese and
colleagues (1996) found that macaque mirror neurons in F5 fire only during object
manipulation using the hand. When the monkey observes the human experimenter
manipulate an object with a tool or pantomime the action without the object being
present, this neuronal activity ceases. Since this finding, several studies have discovered
that mirror neurons can respond to goal-oriented actions that involve tools (Ferrari et al.
2005; Umiltà et al. 2008; Rochat et al. 2010). For example, Rochat and others (2010)
demonstrate how mirror neurons in F5 fire during the observed and executed actions
involving tool use to grasp an object. This is the result of prolonged practice with reverse
pliers, which requires the monkey to open its hand in order for the pliers to grasp the
object. Moreover, despite never seeing someone stab an object with a sharpened stick,
some of the monkeys’ F5 mirror neurons respond to this action as well upon their first
125
observation, albeit weakly. Prolonged practice allows mirror neurons to generalize to
goal-oriented actions using different implements. Additionally, the distance between an
observer and the observed action, as well as the observer’s subjective value placed on an
item, may modulate the response of mirror neurons in monkeys (Caggiano et al. 2009;
Caggiano et al. 2011). Assuming that all catarrhines possess mirror neurons, it is not
unreasonable to assume that either a small evolutionary change or perhaps simply a
prolonged period of learning could allow for some of the advanced toolmaking behaviors
visible among the great apes.
While most the research on mirror neurons has focused on the upper sector of F5
where hand movements are represented, Ferrari and colleagues (2003) demonstrate that
mirror neurons in the lateral portion of F5 are related to mouth actions. Mouth mirror
neurons can be divided into two types, ingestive and communicative. Although the
ingestive mirror neurons make up about 80% of the total amount of recorded mouth
mirror neurons in macaques, the communicative mirror neurons are intriguing because of
their increased firing during observation of communicative signals, such as lip-smacking,
though they respond to ingestive signals as well. This may be because the communicative
neurons are derived from ingestive neurons and have not yet been freed from their
original ingestive function (Rizzolatti and Craighero 2004). Mirror neurons provide a
mechanism for directly linking the intentions of a sender of a message to its receiver,
which is a key property necessary for language. Interestingly, there is evidence that area
F5 in macaques is the homolog to the pars opercularis of Broca’s area (BA 44) in
humans (Petrides and Pandya 2002, 2009), though this is still controversial (Grèzes et al.
2003).
126
Direct recordings of spiking activity from single neurons in epileptic patients
confirm that humans indeed have neurons with mirroring properties as well (Mukamel et
al. 2010). Numerous studies had previously provided indirect evidence that a mirror
neuron circuit also exists in humans through functional neuroimaging techniques (e.g.,
Iacoboni et al. 1999; Nishitani and Hari 2002; Grèzes et al. 2003; Koski et al. 2003;
Kilner et al. 2009; Molenberghs et al. 2012). Transcranial magnetic stimulation (TMS)
and functional magnetic resonance imaging (fMRI) have been some of the principal tools
for identifying cortical activation areas during observed actions. For example, Fadiga and
colleagues (1995) report that as subjects observe the experimenter perform an action
during stimulation to their motor cortex with TMS, motor-evoked potentials occur in the
same muscles of their right hand and arm as those used by the experimenter during the
action. Another interesting result of this study is the observation that motor-evoked
potentials occur while the subjects watch both transitive and intransitive acts. The
response to intransitive, or meaningless, actions differentiates mirror neurons of humans
from those of monkeys.
Buccino and others (2001) investigate this issue further in an experiment utilizing
fMRI, in which participants observe both transitive and intransitive object manipulation
actions of the mouth, hand/arm, and foot/leg. They discover that object-related mouth
movements activate the pars opercularis and the inferior and posterior portions of the
IPL; object-related hand/arm movements also activate the pars opercularis and precentral
gyrus (PrG); and object-related foot/leg movements activate the PrG and posterior
parietal lobe. Intransitive actions activate all the same areas, except for the parietal
regions. These results confirm the same circuit seen in the macaque. Indeed, in a brain
127
imaging study using positron emission tomography (PET), Hecht and colleagues (2013)
report that chimpanzees not only demonstrate similar distributions of neural activity in
frontoparietal networks as reported in macaques and humans during observation and
execution of manual grasping action, but these regions also respond to observations of
intransitive actions. This new evidence indicates a gradual evolution of the mirror neuron
system to support novel behaviors.
Because of Mukamel and colleagues’ (2010) direct recordings of 1,177 neurons in
twenty-one epileptic patients, there are now four additional brain regions in humans
known to have mirror neurons, including the supplementary motor area (SMA), the
parahippocampal gyrus, the hippocampus, and the entorhinal cortex. There may be
additional areas containing mirror neurons that are still yet unknown, but the placement
of electrodes in this study was limited to clinical purposes and thus restricted the
researchers to measure the activity of individual neurons in only certain areas. The
responses of the mirror neurons in these new regions indicate that they may play a
mirroring role in many aspects of actions of the self and other (Iacoboni 2012).
While the discovery of mirror neurons and their potential for psychology have
been compared to the discovery of DNA and its contributions to biology (Ramachandran
2000), there are others who are critical of the initial interpretation of the function of
mirror neurons (e.g., Dinstein et al. 2008; Lingnau et al. 2009; Heyes 2010). Hickok
(2009, 2014) has been the most vocal critic of the mirror neuron theory of action
understanding. His main point of criticism lies in the assumption that humans also have
mirror neurons, and these human mirror neurons have different functions than macaque
mirror neurons. Macaques, which are known to possess mirror neurons, do not possess
128
higher order cognitive processes, and humans who possess higher order cognitive
processes are not conclusively known to possess mirror neurons. This logic is based on
current assumptions that mirror neurons exist in humans; they have evolved to support
new functions in humans that they do not support in monkeys; and this evolution has
conserved the same functional properties found in monkeys. He claims that these
assumptions, however, do not stand up to empirical examination because damage to the
IFG in humans does not cause deficits in action understanding. Action production and
action understanding are dissociated, meaning that two independent structures of the
brain support their function. Although the study carried out by Mukamel and others
(2010) is not a well-controlled experiment because the electrodes were placed in epileptic
patients for medical reasons alone, it is nevertheless clear that neurons with action
mirroring properties exist in humans, and they appear to be more distributed than
predicted, which could explain why damage to the IFG does not result in action
understanding deficits.
Imitation is widely accepted as one of the necessary pre-adaptations for language
(Christiansen and Kirby 2003). Imitative behavior in monkeys is largely absent
(Visalberghi and Fragaszy 1994; but see Subiaul et al. 2004); therefore, imitation, here
defined as the causal relationship between observation of a feature of body movement
and the execution of the same feature of body movement by the observer (Heyes 2001), is
a key difference between monkeys and humans. The discovery that mirror neurons fire
when the animal performs a goal-directed action or when it observes another individual
perform a goal-directed action has led to active research into the potential imitative
capabilities of mirror neurons in humans.
129
Iacoboni and colleagues (1999) observe that two main areas of the human brain
are activated during a simple finger-tapping imitation task using fMRI: the left inferior
frontal cortex and the most rostral portion of the right superior parietal lobule (SPL).
Similarly, in an event-related fMRI study of naïve participants learning guitar chords by
imitating an expert, Buccino and colleagues (2004b) conclude that the IPL and the
posterior part of the IFG and adjacent premotor cortex form the underlying circuit during
imitation learning. These studies demonstrate that the neural correlates of imitative
learning coincide with the frontal and parietal centers of the mirror neuron system
(Rizzolatti and Craighero 2004). The mirror neuron system has also been implicated in
humans’ ability for hierarchically complex learning of sequential information (MolnarSzakacs et al. 2006) and the ability to understand the intentions and emotions of others
(Carr et al. 2003; Iacoboni et al. 2005; but see Churchland 2011). Its dysfunction is even
cited as a potential cause for autistic deficits (Iacoboni and Dapretto 2006). Buccino and
colleagues (2004a) also note that in human subjects observing oral communicative
actions, such as silent speech in humans and lip-smacking in monkeys, a small area of
pars opercularis is activated in both hemispheres. This connection between mirror
neurons and communicative signals of the mouth and hands has led some researchers to
suggest an evolutionary relationship between mirror neurons and language (Rizzolatti and
Arbib 1998; Arbib 2005; Corballis 2010; Arbib 2012).
Mirror neurons appear to be involved in aspects of human cognition that were
necessary pre-adaptations to language, such as imitation, processing complex hierarchical
information, and the understanding of other individuals as intentional beings. For these
reasons, it is plausible that Broca’s area evolved atop the mirror neuron circuit located in
130
the IFG; in other words, humans’ capacity for language evolved from the mirror
multimodal system that included manual, vocal, and facial behaviors. Rizzolatti and
Arbib (1998) were the first to present this idea, naming it the Mirror System Hypothesis
(MSH). The presence of a mirror system does not equate to language, however, as
monkeys have mirror neurons but do not have language. The monkey mirror system also
differs from the human mirror system in that it does not respond to imitation or
pantomime (Gallese et al. 1996). Thus, the mirror system underwent several significant
changes during the evolution of hominoids and hominins. The MSH proposes that
primitive mirror neurons that identify actions, as seen in the macaque F5 area, evolved in
several stages to support imitation, pantomime, “proto-signing,” “proto-speech,” and
eventually, vocal language.
The mirror system for grasping probably evolved prior to 25 Ma, before humans
and macaques shared a common ancestor. Primitive mirror neurons are excited by the
interaction of object and hand in a goal-directed action performed by the self or other
individuals within view. These types of mirror neurons do not respond with all-ornothing action potentials; rather, they are trained early in life to respond to certain types
of actions and work together as a population code for understanding actions in relation to
the self (Arbib 2012). During development, infants ineptly swing their hands and arms
around and come into contact with objects and thus learn how to grasp these objects. It is
through experience, not imitation, that human infants and other primates learn how to
grasp objects. And according to Arbib’s (2005) hypothesis, at some point during
hominoid evolution, mirror neurons took on the ability of simple imitation, as is seen in
the apes. An early human ancestor was able to extend imitation to increasingly more
131
complex actions by identifying the sub-goals of observed behaviors and parsing these
complex movements into a composite of familiar actions. This transition was supported
by the selection for over-imitation and a longer childhood. Because human mirror
neurons fire during observation of intransitive actions but macaque mirror neurons do
not, Arbib suggests that pantomime provided the breakthrough from complex imitation to
being able to communicate freely about a huge variety of situations, actions, and objects.
Pantomime was limited, however, in what could be expressed, and this form of
communication would have been costly and inefficient. Arbib (2012) asserts that
pantomime provided the scaffold for proto-sign, a form of communication with
conventionalized gestures that were shared by the community. These signs were likely
iconic in the beginning but then paved the way for abstract gestures.
As with any gestural origin of language theory, one must attempt to explain how
speech became the dominant language domain over gesture. For instance, if audiovisual
mirror neurons exist, why not skip pantomime and proto-sign altogether and build protospeech out of the mirror system for goal-directed action sounds? Arbib (2012) insists this
was not possible because there is no evidence for vocal mimicry among monkeys or apes.
Instead, the evolution of a system for voluntary control of intentional gestural
communication based partly in BA 44 provided the basis for the evolution of more
prominent connections from this area to the vocal apparatus because of pre-existing
connections between the hand and mouth for ingestive behaviors. Several studies
demonstrate that hand/arm and speech gestures are linked in some way or share a neural
substrate (e.g. Gentilucci et al. 2001; Horwitz et al. 2003; Meister et al. 2003;
Komeilipoor et al. 2014). Proto-speech may have then evolved along with proto-signing,
132
and they fed off each other in an expanding spiral (Arbib 2012). This would have applied
the necessary pressures on the evolution of the vocal apparatus to lead to the modern
configuration of the human vocal tract. Language, however, according to the MSH, is not
an innate behavior, but rather, a cultural innovation that occurred only after this “neural
critical mass” was achieved. Arbib (2012) thus rejects Chomsky’s universal grammar and
proposes that proto-sign and proto-speech evolved biologically, but language with fully
modern syntactic and semantic structure evolved culturally once the brain was “languageready.”
Arbib addresses many of the questions posed by Bickerton at the beginning of this
chapter with his MSH; however, it is unclear what main driving force acted to cause the
exaptation of the mirror system for protolanguage, besides clearer communication.
Defining the adaptive need for cognitively complex behaviors, such as imitation,
hierarchical learning of sequential information, and theory of mind may be better
addressed by paleoanthropology, and, more specifically, neuroarchaeology.
Neuroarchaeology: The wedding of neuroscience and archaeology
Despite considerable research in multiple fields (Christiansen and Kirby 2003;
Fitch 2000a, 2005), it is unknown when or how language evolved in the human lineage
because of the lack of direct fossil evidence (Fitch 2000a). The archaeological record of
the production of early stone tool technologies, despite some limitations (Buckley and
Steele 2002), may offer an indirect window into the cognitive and linguistic capabilities
of early humans due to the varying levels of skill required to produce different stone
133
tools. Different aspects of stone tool form and manufacture have historically been linked
with evolving language and cognition (see Chapter 2). Some of these aspects include the
shift to a symmetrical, standardized tool form (Klein 1995; Wynn 2002) that some
scholars have claimed represents the knapper’s mental template on the stone (Gowlett
1984; Tattersall 2010), increasing operational complexity over time (Ambrose 2010), and
the potentially analogous relationship between the grammar of language and the
“grammar” of toolmaking (e.g., Montagu 1976; Greenfield 1991). While each idea has its
critics (e.g., see McPherron 2000; McNabb et al. 2004; Iovita and McPherron 2011), the
latter two ideas have become more sophisticated over time as more cognitive
archaeologists have incorporated the latest findings of neuroscience into their hypotheses
(e.g., Stout and Chaminade 2012). These ideas also reflect archaeologists’ shift of focus
from end product classification to understanding the entire process of technological
procedure.
Tool use behaviors and their implications for cognition
Non-human ape cognition represents the closest analog to the cognitive capacities
of the earliest hominins. Tool use behavior is the manifestation of an organism’s
cognitive capacity on its physical environment; therefore, it is useful to study the tool use
behaviors of apes in the context of human evolution. Free-ranging chimpanzees make and
use tools more often than other non-human apes, and these tools are used for food
acquisition and sometimes as weapons (McGrew 1993). Some patterns of chimpanzee
tool behavior include the use of multiple types of raw material for the same type of
134
functional tool (Matsuzawa 1994), the use of the same type of raw material for making
different types of functional tools (McGrew 1992), the use of the same tool for multiple
functions (Sanz and Morgan 2009), and combining two elements to solve a problem
(Koops et al. 2010). The chaîne opératoire approach applied to chimpanzee nut cracking
reveals a complex operational sequence, involving raw material selection and
transportation, use and reuse, and discard, resulting in diverse stone assemblages between
sites (Carvalho et al. 2008). While apes make and use tools in nature, they do not
intentionally chip stone tools. There are some accounts, however, of apes intentionally
chipping stone tools when taught in captivity (Wright 1972; Toth et al. 1993; Schick et al.
1999). While these apes are capable of removing flakes from a larger core, analyses
reveal that their methods of doing so differ from the Oldowan stone tool industry in
several ways: they do not exploit acute angles on the core to make flakes and they throw
the hammerstone against the core or the core itself (Schick et al. 1999).
It was once believed that there were large differences between ape and hominin
tool use (see Chapter 2). For example, it was thought that only hominins were capable of
tool-assisted hunting, digging for underground storage organs of plants, and cleaver
percussion, but accumulating evidence from wild chimpanzees speaks to the contrary
(Hernandez-Aguilar et al. 2007; Pruetz and Bertolani 2007; Koops et al. 2010). With all
the evidence considered, the Lomekwian and Oldowan technologies fit into the adaptive
grade of ape tool behavior and do not exceed it (Wynn et al. 2011). Therefore, one should
assume that language would not have been necessary for the transmission of the skill and
intention necessary to make these earliest of hominin stone tools, nor would these
135
techniques have exerted strong, selective pressures on the brain to support the cognitive
prerequisites for language that surpass those possessed by non-human apes.
The Oldowan tool industry, dated to as early as 2.6 Ma at Gona, Ethiopia (Semaw
et al. 1997), was once thought to be the oldest technology attributed to hominins, until
recently when 3.3 million-year-old stone tools of a different type were discovered at
Lomekwi, Kenya (Harmand et al. 2015). Because the Lomekwian tools are such a recent
discovery, little has been published on their implications for hominin cognition or
linguistic ability at this point, though many lively debates on these topics will probably
surface soon. The Oldowan, on the other hand, has a rich literature, especially pertaining
to cognition and language. It is a quick and expedient method of obtaining a sharp flake
to use as a tool by striking a core with a hard hammer stone (Toth 1985b), resulting in
non-standard cores that reflect the original shape of the stone. The timing of the advent of
the Oldowan industry corresponds with some modest increases in brain size, from 450
cm3 in Australopithecus garhi up to more than 600 cm3 in H. habilis. Holloway and
coworkers (2004) suggest that this increase in brain size correlates with language
abilities, hunting, increased social complexity, and tool standardization. Several cognitive
abilities have been inferred from these earliest of stone tools as well. The Oldowan
knappers were skilled at controlled flaking; they understood the differences of quality
between raw materials; and the transport of raw materials demonstrates forethought (Toth
and Schick 2009). The process of Oldowan knapping probably required a learned
tendency to select targets and detach flakes based on the position of previously detached
flakes (Stout 2011) but did not necessitate long-term planning or working memory.
136
Around 1.75 Ma, a series of technological advances appear in the archaeological
record, including hierarchical centripetal flaking and the production of bifacially-flaked,
intentionally shaped, symmetrical tools called handaxes. This technocomplex is known as
the early Acheulian and is usually attributed to H. erectus. These toolmakers likely had
the goal of producing sharp flakes for food processing, but they also intentionally shaped
the core into a large cutting tool for various functions, as is evidenced by use wear
patterns and experimentation (Jones 1980; Domínguez-Rodrigo et al. 2001). Some
researchers argue that the early Acheulian industry represents aspects of an increasing
cognitive capacity, such as complex imitation and shared intentionality (Shipton 2010;
Arbib 2011), the ability to create mental representation (Tattersall 2008), and
protolanguage (Arbib 2011).
Traditionally, the Acheulian industry was thought to represent one of the longest
static traditions in prehistory, lasting for one million years without much change (e.g.,
Mithen 1999). This restricted view of the Early Stone Age (ESA) allowed for
speculations about the limited cognitive capabilities of H. erectus; however, this attitude
toward the Acheulian industry has changed more recently as evidence for evolving
cognition and regional and temporal variation have been recorded (Norton et al. 2006;
Lycett and Gowlett 2008; Nowell and White 2010; Goren-Inbar 2011). Specifically,
archaeologists recognize an important transition between the early and late Acheulian (ca.
0.7 Ma). Handaxes from this later period are much more refined than earlier handaxes.
They are characterized by cross-sectional thinning with the aid of a soft hammer,
platform preparation, and three-dimensional symmetry (Wynn 2002; Stout 2011). The
Acheulian, therefore, increased in technological complexity over time, which likely
137
represents evolving cognitive abilities as well (Belfer-Cohen et al. 1994; Stout et al.
2008).
The late Acheulian industry also saw the innovation of the Levallois method
around 0.3 Ma. The Levallois method, which is most commonly associated with
Neandertals, is a reduction strategy that prepares a core for the ultimate removal of one or
more preferential flake blanks, which then may or may not be further modified (Wynn
and Coolidge 2004). The finely made handaxes of the late Acheulian may have allowed
for an easy transition to the Levallois method (Rolland 1995). There is also evidence
from stemmed points (Clark 1992), microwear traces from mounts (Anderson-Gerfaud
1990), and mastic organic residues (Boëda et al. 1996; Lombard 2006) for the first hafted
composite tools, possibly as early as the transition from the Acheulian to the Middle
Paleolithic (MP)/Middle Stone Age (MSA; Ambrose 2001). Although Foley and Lahr
(1997) propose that the Levallois method has its origins in Africa, the dispersal of the
method across Europe, the Middle East, and Africa seems to have occurred
simultaneously, which could be evidence for the transmission of ideas through extensive
social networks (White and Ashton 2003). The combination of the Levallois technique
and the large increase in technological complexity as a result of composite toolmaking
have been linked with higher-order cognitive abilities, such as working memory,
constructive memory, problem-solving, and grammatical language (Wynn and Coolidge
2004; Ambrose 2010; Wadley 2010a), while earlier industries are usually considered to
imply an absence of these cognitive abilities.
138
Application of functional neuroimaging techniques to stone tool manufacture
The pioneering work of Stout and colleagues (Stout et al. 2000; Stout 2003; Stout
et al. 2006) introduced the use of functional neuroimaging to archaeology with two pilot
experiments using PET to explore the cognitive complexity required for the production of
simple stone tools similar to the Oldowan industry and the more complex bifacial tools
similar to the late Acheulian industry. Only one expert subject performed the tasks in
both experiments; however, in the former, the subject knapped while lying prostrate in
the scanner and had to use half the normal force to remove flakes so as to limit head
movement (Stout et al. 2000). This unnatural posture for flaking, combined with the
restraints and constrained striking force the subject could conduct, weakens the analogy
between the experimental task and prehistoric knapping (Stout 2003); thus, in the latter
study, the subject knapped outside the scanner for more than forty minutes before
entering the scanner to create a more naturalistic knapping situation (Stout et al. 2006).
Stout and colleagues (2000) report that the areas of the brain receiving the most
activation for Oldowan tool production are those primary motor and somatosensory
regions around the central sulcus, the cerebellum, and to some degree, the SPL (Table 2).
With these results, they suggest that simple Oldowan reduction most likely requires the
integration of multiple sensory inputs, though it does not appear to recruit any prefrontal
language, planning, or problem-solving regions (Stout 2006). In the second pilot
experiment, Stout and colleagues (2006) observe more bilateral activity during the
Acheulian task, but otherwise, the only differences in whole brain activity between
Oldowan and Acheulian knapping are quantitative differences in intensity, not qualitative
139
differences in pattern. In a slightly larger study, Stout and Chaminade (2007) use PET to
observe the brain activation patterns of six novice flintknappers after four hours of
practice to learn to produce simple flakes, and they observe similar clusters of activation
as what is reported in their pilot studies. In addition to the primary motor and
somatosensory areas, parietal areas, and cerebellar activity observed in the pilot studies,
they also find activation in the SMA and left ventral premotor cortex, which borders the
posterior boundary of Broca’s area (BA 44; Table 2).
140
Table 2. Localization of activated parietal, frontal, and temporal cortical clusters during the execution of stone knapping tasks.
Localization
Hemisphere
BA1
Coordinates
x
y
z
Task2
Source
Parietal cortex
Intraparietal sulcus
Left
7, 40
-34
-56
62
Novice Oldowan pre-practice
Stout and Chaminade (2007)
Intraparietal sulcus
Left
7, 40
-28
-60
60
Novice Oldowan post-practice
Stout and Chaminade (2007)
Intraparietal sulcus
Left
7, 40
-28
-48
52
Expert Oldowan
Stout et al. (2008)
Intraparietal sulcus
Left
7, 40
-28
-48
52
Expert Acheulian
Stout et al. (2008)
Intraparietal sulcus
Right
7, 40
34
-52
60
Expert Oldowan
Stout et al. (2008)
Intraparietal sulcus
Right
7, 40
34
-52
60
Expert Acheulian
Stout et al. (2008)
Intraparietal sulcus
Right
7, 40
22
-54
56
Novice Oldowan pre-practice
Stout and Chaminade (2007)
Intraparietal sulcus
Right
7, 40
22
-54
56
Novice Oldowan post-practice
Stout and Chaminade (2007)
Intraparietal sulcus
Left
3, 40
-40
-38
42
Novice Oldowan post-practice
Stout and Chaminade (2007)
Intraparietal sulcus
Left
3,40
-42
-40
44
Novice Oldowan pre-practice
Stout and Chaminade (2007)
Intraparietal sulcus
Right
3,40
34
-32
46
Novice Oldowan pre-practice
Stout and Chaminade (2007)
Intraparietal sulcus
Right
3,40
34
-32
46
Novice Oldowan post-practice
Stout and Chaminade (2007)
Intraparietal sulcus
Left
-26
-52
62
Stone toolmaking observation
Stout et al. (2011)
Intraparietal sulcus
Right
32
-60
56
Stone toolmaking observation
Stout et al. (2011)
Postcentral gyrus
Left
43
-56
-20
36
Novice Oldowan pre-practice
Stout and Chaminade (2007)
Postcentral gyrus
Left
43
-56
-20
36
Novice Oldowan post-practice
Stout and Chaminade (2007)
Postcentral gyrus
Right
1
-39
-26
56
Expert Oldowan
Stout et al. (2000)
Postcentral gyrus
Right
3
38
-32
62
Novice Oldowan pre-practice
Stout and Chaminade (2007)
Postcentral gyrus
Left
-30
-44
62
Stone toolmaking observation
Stout et al. (2011)
Postcentral gyrus
Right
36
-40
56
Stone toolmaking observation
Stout et al. (2011)
Superior parietal
Left
7
21
-49
56
Expert Oldowan
Stout et al. (2000)
Superior parietal
Right
7
-30
-53
61
Expert Oldowan
Stout et al. (2000)
Superior parietal lobule
Left
7
-24
-62
54
Novice Oldowan pre-practice
Stout and Chaminade (2007)
141
Table 2 – Continued
Localization
Hemisphere
BA
Coordinates
x
y
z
Task
Source
Superior parietal lobule
Left
7
-24
-62
54
Novice Oldowan post-practice
Stout and Chaminade (2007)
Superior parietal lobule
Right
7
24
-60
66
Oldowan
Stout et al. (2008)
Superior parietal lobule
Right
7
22
-62
68
Expert Acheulian
Stout et al. (2008)
Superior parietal lobule
Left
5
-14
-54
70
Expert Oldowan
Stout et al. (2008)
Superior parietal lobule
Left
5
-14
-54
70
Expert Acheulian
Stout et al. (2008)
Supramarginal gyrus
Left
40
-48
-32
40
Expert Oldowan
Stout et al. (2008)
Supramarginal gyrus
Left
40
-48
-32
42
Expert Acheulian
Stout et al. (2008)
Supramarginal gyrus
Right
40
58
-30
36
Oldowan
Stout et al. (2008)
Supramarginal gyrus
Right
40
58
-30
36
Expert Acheulian
Stout et al. (2008)
Supramarginal gyrus
Right
66
-20
36
Stone toolmaking observation
Stout et al. (2011)
Supramarginal gyrus
Left
-64
-20
36
Stone toolmaking observation
Stout et al. (2011)
Frontal cortex
Central sulcus
Right
3, 4, 6
40
-14
50
Novice Oldowan pre-practice
Stout and Chaminade (2007)
Central sulcus
Right
3, 4, 6
40
-16
48
Novice Oldowan post-practice
Stout and Chaminade (2007)
Central sulcus
Left
1,4
33
-26
52
Expert Oldowan
Stout et al. (2000)
Dorsal precentral gyrus
Left
6
-24
-8
58
Expert Acheulian
Stout et al. (2008)
Dorsal precentral gyrus
Right
6
34
-8
58
Expert Acheulian
Stout et al. (2008)
Dorsal precentral gyrus
Left
-32
-8
58
Stone toolmaking observation
Stout et al. (2011)
Dorsal precentral gyrus
Right
30
-8
54
Stone toolmaking observation
Stout et al. (2011)
Dorsal premotor cortex
Left
6
-24
-6
54
Novice Oldowan pre-practice
Stout and Chaminade (2007)
Dorsal premotor cortex
Left
6
-20
-4
60
Novice Oldowan post-practice
Stout and Chaminade (2007)
Inferior prefrontal gyrus
Right
45
60
2
26
Expert Acheulian
Stout et al. (2008)
Lateral precentral gyrus
Left
4,6
-44
-14
46
Expert Acheulian
Stout et al. (2008)
142
Table 2 – Continued
Localization
Hemisphere
BA
Coordinates
Task
Source
x
y
z
-46
-16
44
Expert Oldowan
Stout et al. (2008)
Lateral precentral gyrus
Left
4, 6
Medial premotor cortex
Right
6
6
-2
52
Novice Oldowan pre-practice
Stout and Chaminade (2007)
Medial premotor cortex
Right
6
6
-2
52
Novice Oldowan post-practice
Stout and Chaminade (2007)
Orbital gyrus
Left
11
-24
32
-22
Expert Oldowan
Stout et al. (2008)
Pars opercularis
Right
60
10
26
Stone toolmaking observation
Stout et al. (2011)
Pars opercularis
Left
-54
8
22
Stone toolmaking observation
Stout et al. (2011)
Pars triangularis
Right
50
44
10
Stone toolmaking observation
Stout et al. (2011)
Ventral precentral gyrus
Right
6
60
2
26
Expert Acheulian
Stout et al. (2008)
Ventral precentral gyrus
Left
6
-52
6
28
Expert Oldowan
Stout et al. (2008)
Ventral precentral gyrus
Left
6
-52
6
28
Expert Acheulian
Stout et al. (2008)
Ventral premotor cortex
Left
6
-62
10
32
Novice Oldowan pre-practice
Stout and Chaminade (2007)
Ventral premotor cortex
Left
6
-52
2
30
Novice Oldowan post-practice
Stout and Chaminade (2007)
-48
-50
-20
Stone toolmaking observation
Stout et al. (2011)
Temporal cortex
Fusiform gyrus
Left
Fusiform gyrus
Right
Inferior temporal gyrus
Right
34
20/21/37 52
-48
-46
-50
Inferior temporal gyrus
Left
-64
Where blanks appear, BA was not provided in the original source.
2
Any task that did not involve knapping execution is marked.
1
143
-28 Stone toolmaking observation
-10 Expert Oldowan
-10 Stone toolmaking observation
Stout et al. (2011)
Stout et al. (2008)
Stout et al. (2011)
Because of the level of difficulty and long apprenticeship required to make
refined, late Acheulian handaxes, Stout and colleagues (2008) turned to expert knappers
to map the activation patterns associated with late Acheulian tool manufacture. This
study consists of only three expert subjects, who knapped outside the PET scanner. The
results show increased activation of the right ventrolateral prefrontal cortex vlPFC; BA
45; Table 2), which is important for the hierarchically-organized action sequences
necessary to make a biface (Stout et al. 2008). Similar to their earlier studies, there is no
involvement of the dorsolateral prefrontal cortex (dlPFC), a commonly cited region that
is normally active during working memory tasks. They interpret this absence to mean that
strategic planning does not play a large role in Oldowan or Acheulian tool production. In
a more recent study using fMRI, Stout and coworkers (2015) test whether the dlPFC is
involved in ESA tool manufacture by asking six subjects, who were trained to make late
Acheulian handaxes over a two-year-long period, to predict and make judgments on
toolmaking outcomes based on images of in-progress stone tools. The setup of this task,
asking subjects to imagine operations on an object, is typical of tasks that are designed to
elicit activation in working memory areas (John Spencer, letter to author, September 7,
2015; e.g., Hyun and Luck 2007); thus, it comes as no surprise that they find that a
complex knapping procedure demonstrates increased activity in the dlPFC under these
conditions (Table 2). Without a more direct approach, one that would involve naturalistic
knapping, it is still unclear whether the dlPFC is a necessary part of the network required
for knapping.
Stout and coworkers (2011) also look at neural activation patterns of subjects of
varying knapping expertise during observations of Oldowan and late Acheulian
144
toolmaking instead of execution, and the results of this study are similar to the previous
PET activations during the execution of Paleolithic toolmaking. The Acheulian task
shows increased activation of bilateral pars opercularis (BA 44) and right pars
triangularis (BA 45), which, as they argue, reflects the syntactic processing necessary for
the increasing complexity of hierarchically organized actions during Acheulian
toolmaking (Table 2). The activation of Broca’s area and its right hemisphere analog
during stone toolmaking, as well as the presence of a homolog to Broca’s area during
grasping in macaques (Gallese et al., 1996), has led Stout and other researchers to suggest
that a praxic system (i.e., manual motor control) in Broca’s area was elaborated and later
co-opted for articulatory control required for speech and hierarchical language production
during human evolution (Pulvermüller and Fadiga, 2010; Stout and Chaminade 2012).
Some researchers suggest that the transition from the Oldowan to the Acheulian
represents aspects of an increasing cognitive capacity, such as complex imitation and
shared intentionality (Shipton 2010; Arbib 2011), the ability to create mental
representation (Tattersall 2008), and protolanguage (Arbib 2011). Mirror neurons have
been implicated in several human behaviors that are crucial for language, such as
imitation (Rizzolatti and Craighero 2004), hierarchically complex learning of sequential
information (Molnar-Szakacs et al. 2006) and theory of mind (Carr et al. 2003; Iacoboni
et al. 2005). These behaviors correspond to some of the hypothesized cognitive abilities
of early Homo as inferred from the Oldowan to Acheulian transition. Mirror neurons have
been directly identified in the premotor cortex, SMA, and inferior parietal cortex, among
other areas, and they may also be located in the IFG in humans (Kilner et al. 2009). The
IFG could provide the best evidence for overlap between language and praxis. It is
145
known to function as a supramodal processor for hierarchically structured, sequential
information, and there is evidence for a posterior-anterior hierarchical gradient of
abstraction (Koechlin and Jubault 2006; Badre and Wagner 2007; Race et al. 2009; Stout
and Chaminade 2012). For example, as input comes in, the premotor cortex first
deciphers basic hand-object interactions and processes phonemes; the pars opercularis
processes simple tool-use action sequences and linguistic syntax; and the pars
triangularis processes more complex actions and semantic/syntactic integration (Figure
1). Oldowan toolmaking seems to involve the premotor cortex but not the more anterior
IFG. Late Acheulian toolmaking involves the premotor cortex and also the pars
triangularis, which is associated with more abstract action representation and hierarchical
organization. Because of the evidence for the involvement of the IFG in distal manual
motor functions, observation and execution of stone tool reduction, as well as speech
production (Heiser et al. 2003; Higuchi et al. 2009; Stout et al. 2011), it is reasonable to
hypothesize that certain subcomponents of language, such as complex imitation,
hierarchical processing of sequential information, and shared intentionality, were already
in use for stone tool manufacture and other complex activities to a certain degree before
the existence of language or protolanguage, and these cognitive mechanisms were then
exapted from the same neural substrate for protolanguage (Stout and Chaminade 2012).
This general hierarchical information-processing network may originate from the primate
mirror system. This is known as the Technological Origin for Language hypothesis.
146
Figure 1. Functional subregions of Broca’s area with a posterior-anterior gradient of
abstraction for processing language and tool use.
Stout, Chaminade, and colleagues have contributed much to the current
understanding of the neural correlates of Paleolithic toolmaking; however, it is imperative
that these studies are replicated for several reasons. Firstly, a major weakness of each of
their knapping execution studies is the habitually small sample size (n = 1, Stout et al.
2000; n = 1, Stout et al. 2006; n = 6, Stout and Chaminade 2007; n = 3, Stout et al. 2008),
resulting in very low statistical power. This is a common problem in the neurosciences,
which can lead to a lesser likelihood that a statistically significant result reflects a true
effect and reduces the chance of detecting a true effect in general (Button et al. 2013).
Secondly, none of these experiments imaged the same subjects throughout the early
stages of learning to observe the general effects of learning (but see Stout et al. 2015 for a
longitudinal study over a period of two years). Currently, there is no study to compare
147
neural activation patterns associated with Oldowan and Acheulian tool manufacture
among novices in the early stages of learning because only expert Acheulian toolmaking
is reported. Hecht and colleagues (2015) demonstrate with DTI that intensive training in
handaxe manufacture effects structural changes in the brain. For this reason, it is
important that the brain activity of novice toolmakers be compared to that of other
novices, not experts. And if a comparison is to be made between the neural correlates
associated with Oldowan and Acheulian tool manufacture, then comparisons should be
made between the same subjects. Thirdly, because of the neuroimaging techniques they
used, they were unable to measure the functional activity of the brain while their subjects
actively knapped but instead depended on remnant glucose uptake after a knapping
session in the case of the PET studies and an observational paradigm in the case of the
fMRI studies. Finally, it is still unknown what effect language has on novices as they are
learning to make large core bifaces. Stout and colleagues (2000, 2008, 2011, 2015; Stout
and Chaminade 2007) do not consider language instruction, as opposed to nonverbal
mimicry, as a variable in their experiments, which may confound some of their results
because it is possible, if not probable, that earlier species of humans lacked complex,
spoken language and could not have learned to knap in a similar manner.
To address the issues outlined above, a novel experiment is introduced in this
thesis that uses functional near-infrared spectroscopy (fNIRS) to explore the brain
activation patterns among thirty-three individuals learning to knap, either with or without
language instruction, over the course of several weeks of training. Similar to PET and
fMRI, fNIRS, as detailed in the following chapter, measures the hemodynamic response
of neurons during different tasks but does so with innocuous near-infrared light rather
148
than radioactive tracers injected into the bloodstream (PET), and it allows the subject to
sit comfortably with free range of movement in more realistic test conditions rather than
being immobilized in an extremely loud and confined machine (fMRI).
The modality in which one learns a new task (i.e., verbally vs. nonverbally) can
have profound observable effects in the brain and on the resulting behavior (Hampson
and Kimura 1984; Taylor et al. 1990; Kelley et al. 1998; Narumoto et al. 2000; Putt et al.
2014b). The study that is the subject of this dissertation explores the effect of learning
modality on toolmaking in greater detail and consequently, the co-evolutionary
hypotheses for toolmaking. Based on a co-evolutionary hypothesis for toolmaking and
language, one should expect to see significant activation of Broca’s area and its right
hemisphere analogue, and possibly other areas known to be involved in language
processing, even if the toolmaking skills are learned nonverbally. If toolmaking skills are
learned nonverbally and various language-processing centers of the brain are activated,
this would support the idea that these regions were already in place for complex motor
skills before language evolved. The results of a previous experiment indicate that the presence of spoken
language while learning to knap stone affects the outcome of the debitage production but
not the final shape or quality of the biface (Putt et al. 2014b; Figures 2-3). In other words,
individuals who receive verbal or nonverbal instruction learn the process of stone
knapping in fundamentally different ways but come to the same result. The verbal
participants more faithfully imitate the instructor, to the point of over-imitation, by
devoting more time to setting up ambitious platforms that, in the end, are too difficult to
execute at this early a stage of learning, as is evidenced by their higher frequency of
149
missed hits and thick, stout flakes. Ambitious platform preparation may not be obvious to
the nonverbal participants as a goal to attempt to meet because of the absence of speech;
therefore, they focus more on emulating the process to reach the goal of a large core
biface by detaching small, thin flakes until they are satisfied that their product resembles
those produced by the instructor. Subjects in the nonverbal group produce about twice the
amount of debitage as the subjects in the verbal group. So, spoken language during
transmission of bifacial knapping may aid in more faithful imitation and a better
understanding of goals. If these learning differences are visible in the stone, then it is
likely that inter-group differences also exist in the brain during this learning process, and
a neuroimaging experiment which could identify differences in the brain between
subjects who learn to knap with spoken language and those who learn by observation and
mimicry alone is necessary.
Figure 2. A comparison of the mean ratio of flake size to flake mass between the verbal
and nonverbal groups from Putt et al. (2014), clearly demonstrating a significant
difference between the two groups (error bars represent 95% confidence intervals). A
higher ratio of size to mass is the result of a higher ratio of platform width to platform
thickness.
150
Figure 3. Median asymmetry index scores for the handaxes produced by the verbal and
nonverbal groups during the fourth and fifth practice sessions from Putt et al. (2014),
demonstrating the similarity in biface symmetry between the two groups (error bars
represent 95% confidence intervals).
To address the issues outlined above, a novel experiment is introduced in this
thesis which uses functional near-infrared spectroscopy (fNIRS) to explore the brain
activation patterns among thirty-three individuals learning to knap, either with or without
language instruction, over the course of several weeks of training. Similar to PET and
fMRI, fNIRS, as detailed in the following chapter, measures the hemodynamic response
of neurons during different tasks but does so with innocuous near-infrared light rather
than radioactive tracers injected into the bloodstream (PET), and it allows the subject to
sit comfortably with free range of movement in more realistic test conditions rather than
being immobilized in an extremely loud machine (fMRI).
151
Summary
Language is often considered a defining human feature (Henshilwood and Marean
2003; Coolidge and Wynn 2005; Hill et al. 2009). In fact, Davidson and Noble (1989)
even argue that complex human culture would not be possible without language, yet
when or how this unique communication system evolved remains contentious. Whereas
one camp of researchers argues that language is discontinuous with other animal
communication systems and has a relatively recent origin (i.e. within the last 200,000
years; e.g., Davidson and Noble 1989; Bickerton 1990), the opposing camp contends that
the complexity of language necessitates a gradual evolution from an ancestral primate
communication system (e.g., Aiello and Dunbar 1993; Jackendoff 1999; Corballis 2003;
Fitch 2004). Mirror neurons, which are brain cells that process both sensory and motor
information, have been implicated in the evolution of language and stone tool
manufacture, potentially indicating a gradual co-evolution of these abilities in human
ancestors (Arbib 2005; Stout and Chaminade 2012). Stone tools have often been linked with evolving language and cognition
(Montagu 1976; Gowlett 1984; Greenfield 1991; Ambrose 2010; Tattersall 2010), but
until recently, supporting neurological evidence for these ideas was nearly absent. A
series of neuroimaging experiments demonstrate that Acheulian toolmaking possibly
involves higher-order association areas of the brain than Oldowan toolmaking (Stout et
al. 2000; Stout and Chaminade 2007; Stout et al. 2008; Stout et al. 2011). In particular,
they argue that their results support motor hypotheses of language origins, linking manual
152
coordination with evolving capacities for speech production because of the concentration
of oxygenated blood flow to Broca’s area and its right hemisphere analog.
A multidisciplinary approach to the origins and evolution of language has not yet
come to a consensus on most of Bickerton’s questions, let alone when hominins first
began to speak linguistic utterances or use spoken language as a tool for teaching. This is
why it is important to replicate Stout and coworkers’ results but with a novel
experimental paradigm that emphasizes the differences in the linguistic mode by which
novices learn to knap stone. If these studies’ results are replicable using novice knappers
as subjects who received nonverbal instruction, then a more confident claim could be
made that the language network that involves hierarchical processing in Broca’s area is
an exaptation of a pre-existing praxis network used for such activities as stone knapping. The results of the current study supplement existing findings regarding the neural
correlates of knapping in the modern human brain, specifically Oldowan and Acheulian
tool production, which, in effect, provide insights into the cognition and behaviors of
early Homo. The results of this project add to the current knowledge of how the ancestral,
monkey-like mirror neuron system for grasping and action understanding evolved into
the modern human mirror neuron system that allows for complex imitation, language, and
theory of mind. Additionally, this research applies a rigorous test for the proposed coevolutionary relationship between language and technology (Stout and Chaminade 2012).
As a result, this study aids in the interpretation of fossil and archaeological evidence to
better understand how modern cognition and language evolved in human ancestors.
153
CHAPTER 5:
MATERIALS AND METHODS
Testable hypotheses
This chapter lays out the specific hyopotheses that are tested in this thesis, along
with an explication of the the materials and methodological details that are employed.
This research rests on several assertions that have been generated from the existing
literature: 1) the Acheulian technocomplex from the Early Stone Age (ESA) represents a
more complex technology relative to earlier tool industries, which could, in effect, signify
that H. erectus possessed a more complex cognition compared to earlier hominins, or
possibly possessed protolanguage abilities; 2) the hominins who made the Oldowan and
Acheulian industries were skilled toolmakers; 3) early Homo probably did not possess
modern human language and therefore could not have learned to make stone tools via
verbal, grammatical language instruction; and 4) evidence for a co-evolution of language
and technology should be observable as overlapping areas of activation in the brain. To
test these assertions and link this theoretical framework with the proposed experimental
design, the following explicit hypotheses are tested:
1. If the manufacture of complex, bifacial stone tools similar to the early Acheulian
stone tool industry requires advanced cognition relative to the manufacture of
expedient flakes similar to the Oldowan stone tool industry, then higher order
association structures of the brain (e.g., prefrontal cortex and anterior Broca's
area) should be significantly active only during Acheulian tool production.
154
2. If trained knappers enlist a different set of behaviors or cognitive strategies than
novice knappers, then the knapping tasks should recruit different neural networks
at different stages in the participants’ training.
3. If verbal instruction reorganizes the neural pattern for processing a learned motor
skill, such as Oldowan and Acheulian tool production, then different clusters of
activation should be found among novice knappers who learned via verbal or
nonverbal instruction.
4. If language co-opted the neural networks already in place for tool manufacture
(e.g., the mirror neuron system), then the classic language-processing areas should
be significantly active among novices learning to knap without verbal instruction.
Functional near-infrared spectroscopy (fNIRS) as a neuroimaging technique
fNIRS developed originally in the 1940s out of optical methods used for the
muscle oximeter invented by Glenn Milikan (Ferrari and Quaresima 2012). More
recently, Frans Jöbsis (1977) discovered that brain tissue has a relatively high degree of
transparency in the near-infrared (NIR) range. This discovery led to the non-invasive
detection of hemoglobin oxygenation in real-time in the human brain using
transillumination spectroscopy. The first fNIRS studies were published in 1991 and 1992
and used a single-channel system with low temporal resolution and poor sensitivity. A
channel consists of a light emitting/detecting diode with wavelengths between 650 and
1000 nm, which can record measurements from one small area of the cortex. These
single-channel systems were combined initially to record measurements from multiple
155
regions of the brain, and by the mid-1990s, multi-channel systems were introduced.
fNIRS data have been reliably reproduced (Haeussinger et al. 2011) and are highly
consistent with functional magnetic resonance imaging (fMRI) findings (Steinbrink et al.
2006).
According to Ferrari and Quaresima (2012), fNIRS is based on several principles.
Human tissue is relatively transparent to NIR light, which is either scattered in tissues or
is absorbed by pigmented compounds called chromatophores. Scattering is about 100
times more likely to occur than absorption, which is why NIR light is able to travel
through living tissue. Hemoglobin, a protein in red blood cells that carries oxygen to cells
in the body, is a chromatophore that acts to absorb light in the NIR range. Oxygenated
and deoxygenated hemoglobin (oxy-Hb and deoxy-Hb) possess different extinction coefficients, which can be detected by different wavelengths, oxy-Hb at the lower end and
deoxy-Hb at the higher end of the light spectrum (Wijeakumar et al. 2015); thus, it is
typical for fNIRS systems to measure brain activation in two wavelengths, with one
above and one below 800 nm (Boas et al. 2004). When localized arteriolar vasodilation
occurs, there is increased cerebral blood flow and cerebral blood volume to that area. In
general, the oxygen that is transported to this area usually exceeds what the local neurons
can utilize, resulting in an overabundance of oxy-Hb in active areas and a concomitant
decrease in deoxy-Hb (Ferrari and Quaresima 2012). Because hemoglobin is a
chromatophore that absorbs NIR light and oxy-Hb and deoxy-Hb have different
extinction co-efficients, this means that fNIRS is sensitive to these changes in oxygenated
and deoxygenated blood flow to different regions of the brain in real-time.
156
A number of techniques and systems have been developed for fNIRS. The current
experiment utilizes a Cw6 system made by TechEn in the United States and a continuous
wave technique that is based on constant tissue illumination that measures light
attenuation through the head (Ferrari and Quaresima 2012). It measures oxygenation
departures from the arbitrary baseline of zero, which is calculated using a modification of
Lambert-Beer’s law for a highly scattering medium. The Cw6 is a 24-channel system that
transports NIR light to and from the scalp through source and detector fiber optic
bundles, respectively. Adequate depth of the light can be achieved with a source-detector
distance of about 3 cm, but this depends on the light intensity and wavelength, the age of
the subject, the head region measured, and the size of the head. It is generally agreed that
for this source-detector distance, sensitivity will be found between the source and
detector at a depth of roughly 1.5 cm below the surface of the skin, with a banana-shaped
region of sensitivity extending above and below this depth (Figure 4; Ferrari et al. 2004).
Functional, cortical maps can be reconstructed by projecting the locations of head
landmarks and optodes onto a Montreal Neurological Institute (MNI) atlas head (Custo et
al. 2010).
fNIRS has many advantages over other commonly used neuroimaging techniques.
For the research participants, it is as simple as sitting comfortably in a chair and wearing
a stretchy cap on the head connected to laser optodes, whose weight can be supported by
a hook. fNIRS is safe enough to use on infants and is completely noninvasive. It does not
cause claustrophobia, nor does it create loud sounds. Overall, fNIRS offers a more
naturalistic environment for performing tasks than positron emission tomography (PET)
or fMRI. For the researcher, fNIRS is an inexpensive neuroimaging technique. Signals
157
can be obtained from infants, children, and adults alike. It even has the potential to be a
useful tool for measuring the cortical activity of other primates (Wakita et al. 2010),
which opens the door potentially to many more opportunities to investigate primate
cognition and its evolution. fNIRS allows the opportunity to observe deoxy-Hb
concentration levels in addition to oxy-Hb concentration levels. Compared to the spatial
resolution of electroencephalography (EEG) and magnetoencephalography (MEG) that
operate within the range of centimeters or tens of centimeters, the ~5 mm spatial
resolution of fNIRS is more favorable (Ferrari et al. 2004). The temporal resolution of
fNIRS is similar to EEG, allowing for relatively instantaneous recordings in the
millisecond range, whereas the temporal resolution for fMRI is in the seconds range and
PET is in the minutes range.
Figure 4. A schematic representation of light in the NIR range traveling through the head
(not to scale). Light is emitted from the source (red), scatters through the layers of skin,
bone, cerebrospinal fluid and fibrous membranes surrounding the brain, and gray matter,
and detected by a nearby detector (blue). Penetration depth is thus limited and only
extends to the superficial portions of the cerebral cortex.
158
Despite all the benefits of using fNIRS, there are a few drawbacks, mainly
concerning issues of localizing a region of interest (ROI). Because fNIRS is an in vivo
technique and the precise location of specific brain regions varies between individuals, it
is difficult to be certain of the identification of the brain areas beneath the optodes (Hoshi
2003; Ferrari et al. 2004). Another noteworthy localization drawback of fNIRS is the
limited penetration depth of only a few centimeters that it allows, depending on factors
such as source-detector distance and skull thickness. This means that fNIRS is limited to
only superficial, cortical regions of the brain. This does not pose too much of a
disadvantage for this particular experiment because all of the ROIs are located in the
superficial cortex. It is unknown how much pial cerebral vessels on the brain surface
contribute to the signal (Hoshi 2003) and how blood volume effects changes in the path
length of NIR light through tissues (Ferrari et al. 2004). Physiological noise and motion
artifacts can corrupt the signal (Tak and Chul 2014). The setup can also be quite timeconsuming, especially when the subject has thick or dark hair, which disrupts the path of
the light into the head. There are methods that have been developed recently to address
most of these shortcomings, however, and will be addressed in more detail below in
relation to the current study.
fNIRS has been recruited for a variety of different studies that are relevant to an
experiment on stone knapping and the evolution of language networks. In particular,
there has been an abundance of optical imaging research on complex motor behaviors
(Leff et al. 2011); over sixty studies have been published since 1998 on cortical
activation in the brain’s classic language areas in infants, children, and adults (Quaresima
et al. 2012); and prefrontal regions involved in executive functions have been
159
investigated as well (Masataka et al. 2015). There have been advances in fNIRS
technology such that there are multi-channel systems now that are miniaturized and
wireless and can be worn during normal, everyday activities (Hoshi 2007). This could
have important implications for future neuroarchaeological studies that might require
body positions other than sitting.
Experimental design
This is an experimental study in which two groups of novice knappers (n = 33; 17
females, 16 males; mean ± SD age 23.8 ± 7.9 years), all of whom were healthy and righthanded, learned how to make bifacial stone tools while the oxy- and deoxy-Hb levels in
different regions of their brains were monitored with fNIRS. One group received verbal
language instruction (n = 17; 9 females, 8 males), and the other group received nonverbal
communication instruction only (n = 16; 8 females, 8 males). An a priori power analysis
was performed for sample size determination based on data from a pilot study, comparing
verbal with nonverbal instruction. Beta values from fifteen 20-s intervals of knapping
were extracted from a channel that overlies anterior Broca’s area. The effect size
(Cohen’s d = 1.13) is considered to be large using Cohen’s (1988) criteria. With an alpha
= 0.05 and power = 0.80, the projected sample size needed with this effect size is
approximately 14 subjects per group (Soper 2014). Therefore, a sample size of at least 28
subjects should be sufficient to detect a significant effect between the verbal and
nonverbal groups.
160
The subjects individually attended seven 1-h knapping practice sessions, during
which they learned how to knap stone tools by watching instructional videos featuring an
expert knapper with over twelve years of experience. Both groups watched the same
instruction videos; however, the nonverbal group watched them with the sound turned
off. The instructor's face was not visible in any of the shots so as to eliminate any
linguistic cues that the nonverbal learners may have been able to observe from his mouth
or face. Each practice session introduced a new goal for the subject to meet or reviewed
and refined skills already introduced. The skills and tool types learned during practice
sessions 1 and 2 were comparable to the skills and tool types of Oldowan simple tool
production. Practice sessions 3-7 introduced and reviewed skills involved in the
production of the early Acheulian technocomplex, which involves a more efficient
removal of flakes and the intentional shaping of a large cutting tool (Stout 2011).
Each novice subject underwent three 90-m neuroimaging sessions, usually a week
apart from each other, with identical tasks at different stages in their learning (Table 3).
Each imaging session consisted of a 6-m resting state measure and the same three tasks.
A motor baseline task made up of nine blocks of activity segregated by rest periods was
used to observe activation of the motor cortex while striking rocks together without the
added element of actual knapping. A silent video instructed the subject to make Oldowan
tools from a selection of three cobbles within arm's reach. This task lasted for more than
6 m and was segregated into five 1-m blocks of activity with 15-s resting periods in
between each block. Similar to the Oldowan task, the Acheulian task was introduced by a
silent instruction video. The task lasted approximately 18 m and was broken up into
fifteen 1-m blocks, separated by 15-s rest periods.
161
Table 3. Summary of task design during neuroimaging sessions.
Task
Motor baseline
Oldowan
knapping
Acheulian
knapping
Description
Externally- and internallypaced percussion
Naturalistic knapping focused
on expedient flake production
Naturalistic knapping focused
on core shaping
Block
duration
(s)
Rest
duration
(s)
# times
repeated1
40
20
9
60
15
5
60
15
15
1
The number of times the block was repeated differed between the Oldowan and Acheulian tasks because
the participants exhausted multiple Oldowan cores in five minutes, but more time was needed to complete a
handaxe.
Study setting and schedule of research
All functional neuroimaging took place at the CHILDS fNIRS facility, located in
Spence Labs at the University of Iowa. Practice sessions were held in MacBride Hall at
the University of Iowa. Data collection began in the summer of 2013 and concluded in
the spring of 2015.
Experimental methods and procedures
Subject screening, selection, and withdrawal
Participants were recruited for the study via posted flyers that advertised for
participants interested in learning to make stone tools. Any persons interested in
participating in the study received an online questionnaire that, when completed,
determined their eligibility to participate. They were screened for knapping experience,
162
handedness, neurological, psychiatric, and physical handicaps, and drug use (Appendix
A). Only individuals with no prior experience making stone tools were asked to
participate in the experiment. Because of evidence for abnormal language lateralization in
left-handed and ambidextrous individuals (Szaflarski et al. 2002), the Benton
Neuropsychology Clinic Handedness test was administered during the screening process
to determine the laterality quotient of potential subjects so as to help ensure the accuracy
of ROIs (Oldfield 1971). The handedness assessment included ten questions that asked
the subjects to rate how often they use either hand for specific activities, such as drawing
or throwing. Positive points were assigned for answers involving the right hand; negative
points were assigned for answers involving the left hand. Only subjects who fell within
the range of +75 - +100 points, or extreme right-handedness, were included in the
experiment.
After positively demonstrating right-hand dominance and consenting to
participate, subjects were asked to answer a series of questions adapted from a form by
the National Institute of Health to assess their psychiatric and neurologic history. Persons
who had experienced traumatic brain injury (including stroke, anoxia and hypoxia, brain
tumor, infections of the brain, etc.), loss of consciousness, a history of seizures, or severe
learning disability were not included in the study. Individuals with serious psychiatric
disorders, such as autism, were excluded from the study. Additionally, the Drug Abuse
Screen Test (DAST-10) was included on the questionnaire to quantify the degree of drug
abuse problems of potential subjects (Skinner 1982). Individuals with a recent history of
drug abuse show impairment in cognitive tasks (London et al. 2000). DAST-10 consists
of ten questions with yes/no answer choices. “Correct” choices receive a score of 0.
163
“Incorrect” choices receive a score of 1. Only persons who received a score of 2 or lower
were permitted to participate, as higher scores correspond with the rise in the level of
drug problems reported (Skinner 1982).
Once subjects cleared the screening questionnaires, they were semi-randomly
assigned to an experimental condition group, as determined by their gender and their
score on a dexterity test. Males and females were evenly distributed across the two
groups. The level of manual dexterity subjects possess will more than likely affect their
ability to knap, as it is a skill that requires a high degree of hand-eye coordination and
precision. Although it was not used as a screening filter for participation in the study,
manual dexterity was an important variable to consider in relation to knapping skill. The
Minnesota Manual Dexterity Test (MMDT) assesses the manual dexterity required to
place sixty round pegs with the dominant right hand in specific places on a board
(Yankosec and Howell 2009). While it is often used by physical and occupational
therapists to determine baseline progress data from an injured patient, the MMDT has
also proven to be a reliable and valid method for obtaining measures of manual dexterity
in healthy adults (Desrosiers et al. 1997; Yankosec and Howell 2009; but see Surrey et al.
2003). The MMDT has established norms to aid in the interpretation of individual scores
(Table 4). The nonverbal group averaged 182.4 ± 17.5 s to place all sixty pegs in the
holes on the board in three iterations, while the verbal group averaged 182.7 ± 16.9 s.
There was no significant difference in dexterity between the two groups based on this
assignment (t = 0.06, p = 0.95). Males, who averaged 181.4 ± 14.2 s, and females, who
averaged 183.6 ± 19.5 s, also did not significantly differ in their dexterity scores (t = 0.34, p = 0.74).
164
Table 4. Score system for the Placing Test,
reported in The Minnesota Dexterity Test
Examiner Manual (Lafayette Instrument 1998).
Dexterity
Very High
High
Average
Low
Very Low
Percentile
Rank
100
90
80
70
60
50
40
30
20
10
0
Seconds
for 3 Trials
Standard
Score
138
144
148
152
155
159
162
167
174
6.28
5.84
5.53
5.25
5.00
4.75
4.47
4.16
3.72
Once the consenting subjects passed these screening measures, they were invited
to attend their first practice session. One participant dropped out of the study halfway
through this first session. As mentioned earlier, dark or thick hair can interrupt the NIR
light path into the head, resulting in signals lacking a hemodynamic response function
(HRF). For this reason, four subjects had to be withdrawn from the study during their
first neuroimaging session when it was determined that adequate signals could not be
obtained. Finally, two subjects withdrew from the study before the final neuroimaging
session for personal reasons, resulting in 31 subjects who completed the entirety of the
experiment.
165
Subject safety and ethical considerations
This study was approved by the IRB and Human Subjects Office at the University
of Iowa (IRB ID #: 201304789). For all practice and neuroimaging sessions, subjects
were required to wear safety goggles, leather, work gloves, and lap pads (Figure 5). They
were also given the choice to wear a facemask to block out small particles of airborne
silicates.
This study was conducted in conformance with the principles of the ‘Declaration of
Helsinki.’ All researchers involved in the study were certified in human subject
protections through the Collaborative Institutional Training Initiative program, as
required by the University of Iowa. Precautions were taken to ensure the safety of the
participants. A consent form that describes the study and any potential risks or benefits to
the subject was provided to individuals interested in participating. If they read and signed
the document, they gave their voluntary, informed consent. Participants were assigned a
subject number to be used in all data extraction. All electronic data were stored on a
password-protected computer in a locked office at the University of Iowa.
166
Figure 5. Photograph of a subject knapping during a practice session while wearing
safety equipment.
Materials for stone knapping
Subjects were presented with three or four local, granitic rocks of varying sizes
that were naturally rounded for use as hammer stones. A goal of the training was to
introduce the subjects to different qualities, shapes, and types of rock to fracture so that
they would learn through trial and error to select the blank of highest quality and the most
workable edges from the three choices that they were always provided. Thus, a variety of
unheated cherts from the Midwest, Texas, and California were obtained from collectors
167
in Missouri and Texas, though most of the material was Burlington chert, a fine- to
medium-grained stone that is easy to flake (Hoard and Anglen 2003). Generally, smaller
pre-made spalls of chert with edges of very acute angles were provided in the first two
practice sessions. By the third and fourth practice sessions, they could choose from
medium-sized spalls without cortex that had edges with more difficult angles, as well as
rounded cobbles with cortex but with one or more flakes already removed to help them
get started. A mix of small- to medium-sized spalls and cobbles were available to choose
from for the Oldowan task during the neuroimaging sessions. Larger, more challenging
pieces, many with square edges, were provided for the fifth, sixth, and seventh practice
sessions and the Acheulian task during the neuroimaging sessions. Prior to being made
available for the subjects to knap, each stone was assigned a unique, identifying label,
weighed on a digital scale, and assigned a measurement of volume by the water
displacement method.
Localization of Regions of Interest
The determination and localization of the ROIs and design of the optode geometry
for this experiment was a multi-step process based on a paper on optimal probe design by
Wijeakumar and colleagues (2015) and established protocols originally developed in the
context of EEG (Figure 6). ROIs were determined by the results of three stone knapping
studies that involve either PET or fMRI (Stout and Chaminade 2007; Stout et al. 2008;
Stout et al. 2011). The coordinates of any superficial cerebral cortex regions with
significant activation reported in these studies were recorded. To further investigate the
168
supposed involvement of the ventrolateral prefrontal cortex (vlPFC) during the transition
to bifacial flaking, Table 2 in Badre and Wagner (2007) was consulted, which averages
the coordinates for the vlPFC reported in six other studies (Badre et al. 2005; Badre and
Wagner 2005, 2006; Dobbins and Wagner 2005; Fernandes et al. 2005; Gold et al. 2006).
Similarly, to confirm the lack of involvement of the dorsolateral prefrontal cortex
(dlPFC; i.e., working memory) during ESA tool manufacture, the coordinates of dlPFC
activation in leading papers on working memory were also compiled (Pessoa et al. 2002;
Pessoa and Ungerleider 2004). Talairach coordinates were converted into MNI
coordinates using GingerALE 2.3 software (Lancaster et al. 2007; Laird et al. 2010). A
pairwise comparison of all the compiled MNI coordinates was performed. Any
coordinates that fell within less than 2 cm of each other were grouped together, and each
group was averaged to obtain one set of coordinates for each ROI. The correct anatomical
location for each ROI was then confirmed using the Analysis of Functional NeuroImages
(AFNI) software (Cox 1996; Cox and Hyde 1997).
169
Figure 6. Flowchart demonstrating the steps involved in creating and visualizing the
optode geometry design, adapted from Wijeakumar et al. (2015).
An optode geometry was designed using the EEG 10-20 coordinate system that
would determine the placement of sources and detectors onto an EasyCAP (Brain
Products GmBH, Germany) to be worn by the participants during the neuroimaging
sessions (Figure 7). The 10-20 system is a system of coordinates on the scalp that allows
for a standard and reproducible method of applying electrodes, in the case of EEG, or
optodes, in the case of fNIRS (Reilly 2005). These coordinates can be positioned and
scaled to each individual’s head using anatomical landmarks: inion, nasion, and right and
left pre-auricular points. The accuracy of the 10-20 system to the underlying cortical
170
structures has been well documented (Jasper 1958; Morris et al. 1986; Homan et al.
1987). To achieve the maximum penetrating depth of the NIR light, sources and detectors
were placed at equal distances from each other. This distance depended upon the size of
the cap (Table 5).
Figure 7. Optode geometry design (not to scale; left) and digitized points from the same
optode geometry registered onto an adult atlas head (right). Red circles represent light
sources and blue circles represent light detectors.
Table 5. Source-detector distances corresponding with adult-sized caps.
Head circumference (cm)
60
58
56
54
Cap size
Large
Medium
Small
X-small
Source-Detector distance (mm)
30
29
28
27
The following only describes the right hemisphere placement of optodes using the
10-20 system, but the optode geometry was symmetrical and therefore could be mirrored
on the left hemisphere. Beginning at the anchor points of F8, T8, and P4, the pitch of the
optode grid was moved anteriorly so that the slope of the grid was decreased by elevating
171
T8 and P4 by 20% of the 10-20 system distance. In other words, a drafting compass was
set to 20% of the F8-T8 distance and 20% of the P4-Cz distance, and circles were drawn
around these anchor points. Six sources and eleven detectors were placed on each
hemisphere, with two detectors on the sagittal midline. Three detectors were placed on
the anchor points; one was placed halfway between T8 and P4. A source was placed at
the midpoint between T8 and F8. From there, additional sources were defined by the
intersection of drafting compass rotations around these adjacent detectors. Similarly,
additional detectors were defined by the intersection of compass rotations around
adjacent sources. In order to abate the problem that the fNIRS signal can be contaminated
with interference from the superficial layers of the head, four short source-detector pairs
were included in the optode geometry as regressors. Because of the shorter distance
between these sources and detectors, they were more sensitive to superficial layers and
thus could be used to remove the superficial interference when recovering the cortical
HRF (Gagnon et al. 2012).
After designing the optode geometry on the cap, the cap was precisely aligned to
the 10-20 landmark positions on an adult’s head. To verify the positioning of the cap on
the head, measures were taken from inion to nasion while centering over Cz, and between
the pre-auricular points while centering over Cz and inion. A Polhemus Patriot™ Motion
Tracking System (Colchester, VT) was used to digitize the major landmarks of the head,
including inion, nasion, Cz, and the left and right pre-auricular points, as well as all the
optode positions. A template of the optode geometry obtained from the digitization was
created with SDGUI software in HOMER2 (Huppert et al. 2009). The next step was to
172
transform a representation of the brain to the digitized points, which was done using an
adult atlas available in AtlasViewer GUI in HOMER2.
Monte Carlo simulations (Metropolis and Ulam 1949) were run using
AtlasViewerGUI in HOMER2 to create a sensitivity distribution for each channel that
demonstrated the sensitivity of each channel to detecting changes in absorption of NIR
light (Figure 8). After visually inspecting the results from the Monte Carlo simulations in
Slicer, adjustments were made to the optode geometry to increase the best fit for all
ROIs. Figure 8 demonstrates that large parts of the frontoparietal cortex centered around
the motor strip are covered by the optode geometry. The next step was to quantify the
amount of intersection between the sensitivity distributions and the ROIs. Using the
BRAINSFit registration tool in Slicer, a transformation matrix was generated from the
subject-specific head volume and ROI coordinates. It displayed the ROIs as spheres
within the subject-specific brain volume and sensitivity distribution (Figure 9). Any ROIs
that were more than 2 cm deep or did not receive adequate coverage by the optode
geometry were discarded. For the quantitative results of the intersection between the
sensitivity distributions and ROIs for this study in terms of estimated optical density
(OD) changes, the reader is directed to Table 8 and Figure 9 of Wijeakumar et al. (2015).
173
Figure 8. Sensitivity distributions of left hemisphere channels generated by Monte Carlo
photon migration simulations using the digitized points from the optode geometry
registered to an adult atlas. Sensitivity is scaled logarithmically.
Figure 9. Coverage of the ROIs in the cerebral cortex with the optode geometry designed
for this experiment, using Slicer software. The two most posterior ROIs (superior parietal
lobules) were eliminated because their inclusion would have caused the exclusion of
frontal cortex ROIs, which are more essential to the major hypotheses. S-Superior, PPosterior, I-Inferior, R-Right.
174
Using these methods, the following neural regions were determined and targeted
bilaterally. For each of the listed ROIs, the reader is directed to Table 2 to see the extent
of activation found in these areas by Stout and colleagues during different tasks related to
stone toolmaking (Stout and Chaminade 2007; Stout et al. 2008; Stout et al. 2011). 1) The
primary motor cortex (precentral gyrus [PrG] and anterior paracentral lobule, Brodmann
area [BA] 4) is one of the main contributors in the execution of movement. 2) The ventral
premotor cortex and pars opercularis (inferior frontal gyrus [IFG], BA 6/44) are less than
2 cm apart from each other and are thus considered the same area based on the
methodology already described for determining ROIs. The ventral premotor cortex
participates in goal-directed hand actions (Binkofski and Buccino 2006) and forms part of
the mirror neuron circuit (Rizzolatti and Craighero 2004). The pars opercularis of
Broca’s area is known to participate during tool use (Higuchi et al. 2009), imitation in
motor tasks (Heiser et al. 2003), and verbal syntax (Stout and Chaminade 2012). 3) Pars
triangularis (located between the inferior frontal sulcus, anterior horizontal ramus, and
anterior ascending ramus, BA 45) of Broca’s area is known to process more complex
actions and semantic/syntactic integration (Stout and Chaminade 2012). 4) The dorsal
premotor cortex (dorsal PrG, BA 6) is known to be active during action observation from
multiple studies (Grèzes and Decety 2001; Grafton 2009; Caspers et al. 2010) and
observation of stationary, common tools (Grafton et al. 1997); thus, it is thought to
contribute to action understanding. 5) The lateral premotor cortex (lateral PrG, BA 4/6) is
involved in the recognition and prediction of sequential patterns of external events
(Schubotz and von Cramon 2002). 6) The medial premotor cortex (medial PrG, BA 6)
has been linked to action selection decision-making and motor preparation (Hernández et
175
al. 2002; Simon et al. 2002). 7) The dlPFC (BA 9/46) is known to be involved in
planning and working memory (Curtis and D’Esposito 2003; Hoshi and Tanji 2004);
however, there is no evidence for its activation during the execution of ESA toolmaking
(Stout and Chaminade 2007; Stout et al. 2008), though predictions and judgments made
about strategic late Acheulian toolmaking goals elicit dlPFC activation (Stout et al.
2015). 8) The vlPFC (BA 47) is known to be involved in coordinating flexible, goaloriented behavior (Ridderinkhof et al. 2004). 9) The primary sensory cortex (postcentral
gyrus [PoG], BA 1/2/3) is the main sensory area for touch. 10) The supramarginal gyrus
(SMG; BA 40) is known to be involved in hand-object manipulations (Tunik et al. 2008),
hand gestures (Hermsdörfer et al. 2001), and language perception and processing (Homae
et al. 2002). 11) The intraparietal sulcus (BA 3/7/40) is known to be involved in
sensorimotor integration of manual prehension and in action planning, execution, and
observation tasks (Rizzolatti et al. 1998; Grèzes and Decety 2001; Frey et al. 2005; Stout
and Chaminade 2012). Together with other frontal and parietal regions, the intraparietal
sulcus forms a part of the mirror system (Rizzolatti and Craighero 2004). 12) The
superior temporal gyrus (STG; BA 22) has several functions: it processes object-related
and space-related information and species-specific vocalizations (Karnath 2001), and it
contains Wernicke’s area, which is known to be involved in semantic processing
(Gernsbacher and Kaschak 2003) and possibly other nonlinguistic behaviors
(Bookheimer 2002).
176
Knapping practice sessions
Subjects attended their own individual practice sessions. During each session,
they watched a demonstration video with an angle that displayed only the expert
knapper’s hands. The videos introduced new goals and skills and, if applicable, reviewed
goals and skills from prior sessions. The same demonstration video was played for both
groups to ensure that they learned the same skills at the same time, but the verbal group
watched the video with sound and heard verbal instructions, though they were not able to
see the instructor’s lips moving. The nonverbal group watched the video muted. Each
video clip was ~10 m long, followed by 20 m of individual practice time when the
participants had the opportunity to work with the stone. Each practice session proceeded
in the following order: 1) a 10-m instruction video; 2) 20 m of practice; 3) the same 10-m
instruction video; and 4) 20 m more to practice. They were always presented with a
selection of three cores of chert to choose from and a variety of hammer stones of
different sizes. All the debitage created while knapping fell on a large tarpaulin mat
(Figure 5). After the participants completed a core or core tool and moved on to another
rock, the core/core tool and its corresponding debitage were collected, bagged, and
labeled with the rock number and other pertinent information for further analysis.
During the first practice session, subjects were introduced to the Oldowan
technique. They learned how to recognize ideal striking angles on the raw material and
tried to create flakes. They continued to practice making expedient flakes during the
second practice session. The second video taught them how to recognize the best raw
material for flaking. Subjects learned which materials fracture easily by trial and error.
177
This also was communicated verbally in the verbal group. The third practice session
video featured alternate flaking around a squared edge as the main goal for this session,
which is an important skill for making bifaces. The new instruction video for the fourth
practice session introduced core bifaces, and the instructor in the video demonstrated
biface manufacture at a very slow rate. The next instruction video in the fifth practice
session began to focus more on primary thinning of a piece to remove large convexities.
The sixth instruction video presented information on how to shape and refine a biface by
trimming. Finally, the subjects were presented with an instruction video during the
seventh practice session that focused on the entire process of bifacial reduction so that
they could continue to practice the skills they learned from prior weeks to attempt to
consistently produce refined bifaces.
Neuroimaging sessions
Participants attended three individual neuroimaging sessions of 90-120 m, during
which they were video recorded and their brains were imaged using the TechEn CW6
system with 12 lasers of two wavelengths (690 and 830 nm) and 24 detectors. They sat in
a small, artificial room surrounded by black curtains. The experiment program was
designed with EPrime software. The presentation of stimuli was synchronized with the
Cw6 system. Set-up involved measuring the participant’s head to ensure that the proper
size of cap was assigned to them. Measurements were taken for the correct placement of
the cap on the head. Hair was then gently moved to the side with a bobbing pin under
each hole in the cap until scalp was visible, and the optodes were snapped into place
178
(Figure 10). The landmarks and positions of the sources and detectors on the head were
digitized. The pointer was then mounted on the right forearm with gauze to track arm
movements. Each imaging session consisted of four tasks. In order to eliminate the
possibility of linguistic contamination of the signals, the experiment was designed so that
all instructions were given via silent video and different tones. Subjects were instructed at
the beginning of the experiment to perform the same activity that they viewed in the
instruction videos, which preceded each new task or event. Instructions also included
training on the meanings of different tones they would hear throughout the session that
would signal whether to stop or start an action.
Figure 10. Demonstration of neuroimaging session set-up with fNIRS.
179
The procedures for each imaging session were identical. After being fitted with
the fNIRS cap, fiber optic cables, and Polhemus Patriot motion-tracking device, subjects
were asked to sit quietly for 6 m. This has been demonstrated to be an effective technique
at revealing structural connectivity between neural networks (Greicius et al. 2009).
Furthermore, several studies correlate change in resting state neural activity with learning
(Albert et al. 2009; Lewis et al. 2009). Because the current study observes subjects at
different stages in the process of learning a new motor task, it may offer an important
contribution to the current literature on resting state and learning correlations, though
these data will not be included in this dissertation.
After completion of the resting state, the subjects were instructed to strike two
hammer stones together without trying to create flakes, using three different hand
movements at a pace of 60 beats per minute. The point of this task was to isolate motor
cortex activation. This task lasted for approximately 9 m. Each new hand movement was
demonstrated on a computer screen placed in front of the subject. A 3-s countdown
alerted the subject that the task was about to begin. Then the metronome started and
continued for 10 s, to which the subjects matched their percussion. When the metronome
stopped, they attempted to continue the correct pace for 30 s. A stop tone played, and
they rested for 20 s. This sequence repeated three times for each hand movement, which
included direct percussion, or directly hitting two hammer stones together using both
hands, glancing percussion, or using the dominant hand to direct blows towards the edge
of the rock in the other hand, and grinding, or using the dominant hand to rub one
hammer stone against the other (Figure 11). These latter two movements approximate
common actions used in stone toolmaking.
180
A
B
C
Figure 11. Motor baseline task movements, including the direct percussion motion
(Panel A), glancing percussion movement (Panel B), and grinding movement (Panel C).
Photos by Christopher Vander Linden.
After completion of the motor baseline task, the subjects were instructed to make
Oldowan tools via a video demonstration on the computer screen. They selected a cobble
or blank from three choices on a shelf setting near them. Because of space and equipment
limitations, the participants were provided with flake blanks from a larger cobble, or they
were provided with a small cobble that could be worked on the hand or leg. The
181
participants also needed to decide on which hammer stone to use based on size, and
would continue to make this choice throughout the session as they continued to reduce
the core. This task lasted for more than 6 m and was segregated into five 1-m epochs with
15-s resting periods in between each epoch. The subjects received a signal to rest and to
knap via a stop tone and countdown. If the subjects completed one core, they were
instructed at the beginning of the experiment to not commence working on another until
the time duration for the block had ended. Each finished core and its debitage were
collected during rest periods before the subjects were allowed to begin a new core.
After completion of the Oldowan task, the subjects were instructed to make an
Acheulian tool via video demonstration. This task lasted for 18 m and was broken up into
fifteen 1-m blocks, separated by 15-s rest periods that were communicated to the subjects
via the same tones. If the participants snapped any of the cores before their completion,
they could try to salvage their work. See Figure 12 for a demonstration of the fNIRS data
acquisition during the act of stone knapping.
182
Figure 12. A staged demonstration of fNIRS data being collected during a knapping task.
The model in this photograph is left-handed, unlike the participants in the study. The
participants were also completely surrounded by black curtains while the tasks were in
motion. Photo by Christopher Vander Linden.
Data acquisition and processing
fNIRS data acquisition
fNIRS is a neuroimaging technique that uses near-infrared light (695-1000 nm) to
measure the absorption and scattering of photons as they pass through skull and brain
tissue, which reflects changes in cerebral blood volume and oxygenation, resulting from
functional activation (Buss et al. 2014). fNIRS data were acquired at 25 Hz with the
TechEn CW6 system. Light was delivered to a customized cap via fiber optic cables.
Three signals were computed using the modified Beer-Lambert Law and known
183
extinction co-efficients for oxy-Hb and deoxy-Hb: oxy-Hb, deoxy-Hb, and total
hemoglobin. All three signals were localized to a roughly 1-cm3 voxel on the head.
fNIRS data processing
Most fNIRS studies rely on channel-based analyses; in other words, a group
analysis is usually performed on the mean or median signal amplitudes for each channel,
(source-detector pair; e.g., Holper et al. 2010; Scherer et al. 2011). Using the methods
designed by Wijeakumar and colleagues (in press), one of the goals of this study was to
visualize the functional activation of knapping stone tools by performing image-based
rather than channel-based analyses in order to directly compare the results of this
experiment with previous neuroimaging experiments that observed stone knapping with
PET and fMRI. This was done in three steps: raw optic signal processing, sensitivity
profile creation, and finally image reconstruction.
Hemodynamic signal value extraction
HOMER2 software was employed to demean and convert the data into OD
measures. Using the StimGUI tool in HOMER2, additional stimulus markers were
assigned to every 20 s of knapping activity in the case of the Oldowan and Acheulian
tasks. The stimulus markers for the motor baseline task (one at the beginning and end of
the metronome) had already been synchronized with the Cw6 system during data
acquisition and were not altered any further. A targeted principal component analysis
184
(tPCA) was applied to each assigned stimulus marker in the three tasks mentioned above
to eliminate noise and motion artifacts (Yücel et al. 2014). tPCA targets and eliminates
only the periods of data that were identified to have motion artifacts. Motion artifacts
were identified by predetermined thresholds for changes in the absolute signal amplitude
(0.5), changes relative to the standard deviation of the data (50) within a given period of
time (1 s), and within a time range around the motion artifact (1 s). The data were also
band-pass filtered to remove frequencies lower than 0.016 Hz and higher than 0.5 Hz. For
additional information on the parameters used to process the optical data, see Appendix
B, Table A1. Using a general linear model (GLM), a beta value (β) was obtained for both
oxy-Hb and deoxy-Hb measures in every channel for all conditions in every task for each
subject in all three sessions (Figure 13). The β provides the best fit for the data or, in
other words, it describes the strength of the relationship between time and the change in
hemoglobin concentration in a specific voxel.
Figure 13. Oxy-Hb (solid lines) and deoxy-Hb (dotted lines) concentration levels (left)
corresponding to three channels in the left frontal cortex (right) during a knapping task
after motion artifacts were eliminated from the data.
185
Sensitivity profile creation
Subjects’ heads were digitized at the beginning and end of each neuroimaging
session (Figure 14). The digitizations were projected onto an adult atlas head in
AtlasViewerGUI, where they were visually evaluated, and the session digitization with
the most symmetrical scalp landmarks was selected as the reference. When necessary, a
subject’s digitizations were transformed to fit this reference set of landmarks to correct
for any digitization errors. A sensitivity profile was created for each channel of the
optode geometry for every subject in each session via Monte Carlo simulations, as
described earlier. This procedure defines the probability of the path of photons from the
light source to the detector for each channel and produces a sensitivity profile, which is
the area of cortex that can be reached by each source-detector pair. Subject-specific head
volumes were skull-stripped and transformed to the head volume in the native atlas space
using an affine transform (BRAINSFit in Slicer 3D). The transformation matrix obtained
was applied to the sensitivity profiles to move them to the transformed head volume
space (BRAINSResample in Slicer3D). Sensitivity profiles for each channel were
summed to create a session- and subject-specific mask, and then these masks were
summed across all sessions and subjects. An intersection mask was calculated across
subjects in all sessions to identify the common voxels shared between every subject as
determined by the sensitivity profiles.
186
Figure 14. Digitization of head landmarks and source/detector positions on the head
using the Polhemus Patriot device. Photo by Christopher Vander Linden.
Image reconstruction
To reconstruct an image of the brain with the hemodynamic response, the β value
for each condition in each task and each session was multiplied with the OD value in its
respective sensitivity profile, which created weighted sensitivity profiles. A sum of all the
weighted sensitivity profiles was multiplied with the intersection mask to create
individual β maps for each condition in each task in each session for every subject so that
only the common areas shared between all subjects and sessions were included in
statistical analyses. These individual β maps were then subjected to statistical analysis.
187
Lithic data acquisition and processing
Core tools from the Acheulian task were determined to be bifaces by the presence
of two opposing faces and at least one bifacial edge. The maximum width and thickness
of each biface and partial biface was recorded with digital calipers. All debitage pieces
were collected after the completion of each finished core during both the practice and
neuroimaging sessions. Any debitage that passed through a ¼” screen was discarded. The
remaining pieces were labeled and measured by traditional caliper morphometrics and
other measurements to assess any differences between groups and learning over time.
Each piece was weighed to the nearest tenth of a gram, measured for maximum thickness
to the nearest millimeter, and allocated to a metric size category continuum as defined by
the smallest of a series of nested squares on centimeter graph paper into which the piece
would completely fit (i.e., 1 cm2, 2 cm2, 3 cm2.., etc.).
To discern any differences in skill at raw material selection between the groups,
both raw material quality and the presence of faults were recorded for each piece. The
presence of faults as evidenced by cracks or surficial mineral staining on flat interior
surfaces was noted. Each piece was coded as a flake (either complete, proximal, or
distal), or nonflake debitage shatter (Andrefsky 2005). The presence of an intact striking
platform and a bulb of percussion determined what were considered to be proximal
flakes, while distal flakes were determined by the presence of recognizable ventral and
dorsal surfaces. Any angular pieces with no clear ventral or dorsal surfaces were
classified as nonflake debitage shatter. The maximum width and thickness of striking
platforms were also collected. These measures provided a means to examine several
188
aspects of flake and shatter dimensions, material quality, and platform setup as key
factors in this analysis.
A random sample of 1,389 complete flakes with an area of ≥ 2 cm2 was selected
from 18 participants (9 verbal, 9 nonverbal) during the third neuroimaging session to
measure knapper skill using the ratio of striking platform width to thickness and the ratio
of flake size to mass. Higher ratios describe longer, thinner flakes, which signify greater
platform control on the part of the knapper. Skill was measured by the ratio of platform
width to platform thickness, the ratio of flake size to flake mass, and by the frequency of
missed hits, as evidenced by incipient cones and crushed areas on the flakes and shatter.
Data analysis and interpretation
Analysis #1
The first goal of the analysis described in Chapter 6 was to use fNIRS to define
the neural correlates for Oldowan and early Acheulian tool manufacture in trained
subjects. Data were acquired from 31 participants who completed ~8 h of training. Only
the final neuroimaging session was included in this analysis because it was the first point
at which most of the subjects understood that they were to work towards different goals
during the two knapping tasks, and it was not until this point that all the participants
understood the concept of a handaxe. After completing the final neuroimaging session, 28
of the subjects participated in a short interview about their experiences in the study. By
the second neuroimaging session, only 61% of the subjects noticed a difference in goals
189
between the Oldowan and Acheulian tasks, but by the third session, 93% of the subjects
knew there was a difference between the two tasks (Appendix C, Figure A2), and of
those who noticed a difference, 73% were able to accurately describe the two techniques
in their own words (Appendix C, Figure A3).
The second goal of this analysis was to compare any results to the ROIs that Stout
and colleagues (2008) identify using PET and to language and visual working memory
(VWM) networks described in two published meta-analyses (Vigneau et al. 2011;
Wijeakumar et al. 2015). The hypotheses that were tested include the following: 1)
Acheulian toolmaking engages anterior prefrontal areas relative to Oldowan toolmaking;
and 2) linguistic instruction reorganizes the neural networks involved in learning a new
motor skill, specifically stone knapping.
To achieve these goals, a two-way analysis of variance (ANOVA) was performed
on the whole brain (i.e., the shared regions in the sensitivity profile of the optode
geometry) using only the third session neuroimaging data, by employing the 3dMVM
AFNI program (Chen et al. 2014). Task (Oldowan and Acheulian) and Group (verbal and
nonverbal) were included as factors, producing a main effect for Group and Task, as well
as an interaction effect between Group and Task. A report of the clusters of active voxels
was created using the AFNI program, 3dclust. Using the coordinates for the center of
mass for each main effect, the β values in these areas were extracted for the Oldowan and
Acheulian tasks, and each was compared with the β values in the same regions during the
motor baseline task using the Wilcoxon signed-rank test. This step was included to
identify only the significant clusters that were unique to stone knapping and not simply
contributing areas to a more general motor network.
190
The main effect functional images were combined with Stout and colleagues’
(2008) fronto-parietal ROIs, Vigneau and colleagues’ (2011) language ROIs, and
Wijeakumar and colleagues’ (2015) VWM ROIs that were located in the study’s optode
geometry to examine potential overlaps occurring between these studies. Any of the ROIs
from the previous neuroarchaeological study that overlapped with the current study’s
interaction effect where the verbal group had a higher average β value than the nonverbal
group were interpreted as regions that may be unique to learning to make stone tools with
language instruction and thus may not be an accurate reflection of the neural networks
involved in stone tool manufacture by ESA hominins.
Analysis #2
The goal of the second analysis was to examine the functional changes in the
brain related to stone knapping practice over the course of several weeks. To achieve this
goal, data were acquired from 33 participants at three different points in their training. A
three-way ANOVA was performed on the whole brain using the neuroimaging data from
all three sessions by employing the 3dMVM AFNI program (Chen et al. 2014). Task
(Oldowan and Acheulian), Group (verbal and nonverbal), and Session (1-3) were
included as factors, producing a main effect for Group, Task, and Session, as well as
interaction effects between Group and Task, Group and Session, Task and Session, and
between Group, Task, and Session. A report of the clusters of active voxels was created
using the AFNI program, 3dclust. Using the coordinates for the center of mass for each
main effect, the β values in these areas were extracted for the Oldowan and Acheulian
191
tasks from each session, and each was compared with the β values in the same regions
during the motor baseline task using the Wilcoxon signed-rank test. Like the first
anlaysis, the results were placed in the same space as Stout and colleagues’ (2008) ROIs,
Vigneau and colleagues’ (2011) language-processing ROIs from a meta-analysis, and
Wijeakumar and colleagues’ visual working memory (VWM) ROIs from a meta-analysis.
Summary
This chapter describes the methods and materials used to conduct two analyses
related to mapping the neural correlates of stone tool knapping with the neuroimaging
technique, fNIRS. The first analysis investigates which regions of the brain are most
active during Oldowan and Acheulian tool manufacture in trained knappers and also tests
the effect of linguistic instruction on the involvement of specific brain regions during
these tasks. These results are presented in Chapter 6. The second analysis examines
longitudinal changes in functional activity over time as the participants increased their
skill at knapping (Chapter 7). 192
CHAPTER 6
THE NEURAL CORRELATES OF THE OLDOWAN-ACHEULIAN
TRANSITION
Introduction
The human brain has increased in absolute and relative size in the last 2-3 million
years (Schoenemann 2006), particularly in the prefrontal, temporal, and parietal cortices
(Rilling et al. 2002; Schenker et al. 2005; Schoenemann et al. 2005; Bruner et al. 2011).
This increase in brain size presumably coincided with the evolution of the distinctive
features of modern human cognition and linguistic communication (Sherwood et al.
2008). The evolution of cognition and language in the human lineage, however, is still
poorly understood because of the methodological and theoretical limitations of
paleoneurology and evolutionary psychology, including the inability of paleoneurology to
directly link hominin brain structures to specific functions or behaviors and the reliance
of evolutionary psychology on reverse engineering (Wynn 2009; Falk 2012).
Neuroarchaeology, or the use of neuroscience methods and theories to investigate the
evolution of the brain and cognition from the perspective of past material culture,
provides a potential solution to these problems (Stout and Hecht 2015). For example,
stone tool production was a behavior of some Pleistocene hominins that can still be
performed by humans today. This means that the patterns of brain activity that modern
humans exhibit while making replicative stone tool types from the archaeological record
may reflect the same patterns of brain activity in the hominins who originally made these
193
tools in the past (Stout and Chaminade 2007; Stout et al. 2008; Stout et al. 2011; Uomini
and Meyer 2013; Stout et al. 2015).
Around 1.75 million years ago (Ma), hominins during the Early Stone Age (ESA)
added to their repertoire of simple flake and pebble tools of the Oldowan industry the
more sophisticated, shaped core tools (handaxes and cleavers) of the Acheulian industry
(Figure 15). The apparent differences in technological complexity between these two tool
types may be indicative of a shift in cognition and language abilities from a more ape-like
to human-like state. For example, Acheulian stone tool manufacture is hypothesized to
require increased cognitive control and working memory than Oldowan tool manufacture
(Stout et al. 2015). It has also been argued that a special co-evolutionary relationship
exists between toolmaking and language, where language may have co-opted the praxic
system functions that were already in place for Oldowan technological operations and
subsequently the hierarchical processes of the right inferior frontal gyrus (IFG) that were
necessary for Acheulian technological operations (Stout and Chaminade 2012).
Figure 15. Representative sketches of ESA core artifacts, belonging to (A) the Oldowan
industry and (B) the Acheulian industry. Illustrations by José-Manuel Benito Álvarez.
194
Supporting evidence for these hypotheses is detailed in several recent
neuroarchaeological studies. In a functional magnetic resonance imaging (fMRI) study,
Stout and colleagues (2015) demonstrate that dorsolateral prefrontal cortex (dlPFC) and
the working memory functions associated with this region are recruited when human
subjects make strategic judgments about planned actions on partially completed
Acheulian tools (Stout et al. 2015), but dlPFC participation is surprisingly absent during
Oldowan and Acheulian tool production tasks (Stout and Chaminade 2007; Stout et al.
2008). Others argue that procedural memory rather than working memory is all that is
required to make stone tools (Wynn and Coolidge 2011). It therefore currently remains
unclear whether ESA knapping requires the internal rehearsal and evaluation of action
plans supported by the recruitment of a working memory network. Some of the areas that
do come online during stone tool production, however, are located near languageprocessing regions, including bilateral ventral precentral gyri (PrG) and right IFG (Stout
et al. 2008). This potential overlap between language and toolmaking areas of the brain
has been used to support the claim for a co-evolutionary relationship between the two
behaviors but has yet to be substantiated. By plotting the coordinates of eight significant
clusters from a recent neuroarchaeological study (Stout et al. 2008) in the same space as
the coordinates from a visual working memory (VWM) meta-analysis (Wijeakumar et al.
2015) and a language-processing meta-analysis that includes phonological, lexicosemantic, and sentence processing neuroimaging studies (Vigneau et al. 2011), it
becomes apparent that stone knapping functional activation indeed overlaps with
language centers and also overlaps with the VWM network, a fact that is overlooked in
previous studies (Figure 16).
195
Figure 16. Areas where functional overlap occurs between ESA knapping (Stout et al.
2008; purple), language-processing (Vigneau et al. 2011; red), and/or VWM
(Wijeakumar et al. 2015; green), including (A) right IFG (pars triangularis), (B) bilateral
ventral PrG, (C) left inferior parietal lobule (IPL), and (D) bilateral dorsal PrG. Spheres 8
mm in diameter were constructed from published coordinates. Overlap between spheres
is represented by light blue, mauve, and yellow colors.
Although previous neuroarchaeological studies provide a tantalizing view into the
early evolution of the human brain and toolmaking behaviors, their results may not be
entirely reliable for several reasons. Firstly, participants in these earlier studies learned to
knap via verbal instruction. This likely does not mimic the learning context during the
Pleistocene when hominins probably did not possess language or the cognitive abilities
required for interactive teaching. This is important because key cortical areas implicated
in these studies, such as the IFG, are language-sensitive areas. Thus, IFG activation might
reflect the presence of language instruction in the learning context rather than the
196
necessity of IFG activation for ESA tool production per se. Secondly, previous studies
use positron emission tomography (PET) and fMRI, technologies that provide only a
limited window onto real-time motor learning. For instance, fMRI is quite sensitive to
motion artifacts; thus, it is not possible to have subjects make stone tools while in the
scanning environment, which is key to replicating early hominin neural activation
patterns in modern human participants. Finally, previous studies of ESA toolmaking
enrolled small numbers of participants, resulting in low statistical power.
To overcome these limitations, the current study used functional near-infrared
spectroscopy (fNIRS), a neuroimaging technique that is sensitive to real-time changes in
oxygenated (oxy-Hb) and deoxygenated hemoglobin (deoxy-Hb) in the cortex, to
reconstruct functional images of key cortical sites during naturalistic ESA tool production
by 31 trained participants (Figure 16). Fifteen of the participants learned to knap stone
via verbal instruction by watching videos of a skilled knapper’s actions (the face was not
visible) as he demonstrated and explained how to knap; sixteen of the participants learned
to knap via nonverbal instruction using the same instructional videos, but with the sound
turned off (Appendix B).
Participants in this study attended seven training sessions to learn how to knap
stone in order to make replicative Oldowan and Acheulian tools. They also attended three
neuroimaging sessions at different points in their training (after the first, fourth, and
seventh training sessions). During each neuroimaging session, fNIRS data were recorded
while participants took part in three tasks: a motor baseline task that involved striking
rocks together without attempting to remove flakes; an Oldowan task that focused on
removing simple flakes from a core; and an Acheulian task that centered on the shaping
197
of a bifacial core tool called a handaxe (Appendix B). The neuroimaging data reported
here are from the final testing session.
Figure 17. Coverage of key areas in the frontal, parietal, and temporal cortex with
fNIRS.
Results
A key issue when comparing different groups in neuroimaging studies that
measure changes over learning is that participants might learn at different rates depending
upon their group assignment. To examine this possibility, digitial calipers were used to
take measurements on cores and flake debris from the final neuroimaging session to
determine whether one of the learning groups produced stone tools with greater skill than
the other group (Appendix B). There was no clear difference in skill between the verbal
and nonverbal groups based on measures of their core tools or flake debris (Dibble 1997;
198
Putt et al. 2014; Stout et al. 2014), excepting a significantly higher ratio of flake size to
flake mass in the nonverbal group (Table 6). These results imply that the two groups
learned at similar rates, and any differences in localized activation reflect the
instructional context.
199
Table 6. Summary of behavioral measures for knapping skill determination during the third neuroimaging session.
Verbal
Biface N
Nonverbal
Biface N
Verbal
Total
Nonverbal
Total
Verbal
Proportion
Verbal
Mean
Verbal
SD
Nonverbal
Proportion
Nonverbal
Mean
Nonverbal
SD
Prob1
Proportion of successful bifaces2
24
25
-
-
0.71
-
-
0.68
-
-
-
Ratio of biface width to thickness3
24
25
-
-
-
1.86
0.59
-
1.72
0.08
0.36
16
16
1437
1179
0.91
-
-
0.93
-
-
-
16
16
823
566
-
3.42
1.94
-
3.73
2.45
0.07
16
16
823
566
-
2.34
3.91
-
3.66
5.32
<0.001*
Measurement
Total proportion of flakes4
Ratio of platform width to thickness
Ratio of flake size to mass
5
1
Probability results for equal means via t-tests in the case of normal distributions and equal variances (ratio of biface width to thickness) and the
Kolmogorov-Smirnov test in the case of non-normal distributions and unequal variances (ratio of platform width to platform thickness and ratio of
flake size to flake mass). Significant at *p < 0.05.
2
The proportion of successful bifaces was determined by dividing each group’s total number of successful bifaces by the group’s total number of attempted
bifaces.
3
The ratio of biface width to thickness is informative about the level of biface refinement, such that a refined handaxe should have a larger width relative to
thickness, which would present as a larger ratio (>1.0). This was determined by dividing width by thickness for each attempted biface. Then each of these ratios
were averaged across each group.
4
A random sample of sixteen rocks was analyzed from each group. The total proportion of flakes was determined by dividing the total number of flakes by the
total number of fractured pieces, including flakes and shatter, produced by each group. A higher skill level is associated with the ability to produce a higher
proportion of flakes. The verbal and nonverbal totals include the sum of flakes and shatter produced from this sample of rocks.
5
The ratios of platform width to thickness and flake size to mass were determined from the same random sample of sixteen rocks from each group. Only
complete flakes with intact striking platforms were included from this sample, as can be seen in the verbal and nonverbal total columns. A larger mean ratio for
each of these features would indicate the production of large, thin flakes, and thus, a higher skill level.
200
The hemodynamic responses of the verbal and nonverbal groups and the Oldowan
and Acheulian tasks were compared using a two-way analysis of variance (ANOVA;
Table 7). In order to analyze the highest-order effect in each spatially unique cluster,
main effect areas that overlap with areas where an interaction occurs between Group and
Task were interpreted as the interaction because it subsumes the main effect.
Furthermore, frontoparietal sensorimotor clusters that are active during general arm
movements, as determined by motor baseline activation, were excluded from further
analysis and interpretation. Only those significant clusters where post-hoc tests
determined knapping activation to be significantly higher than the motor baseline
activation were included. Because the motor baseline task did not control for auditory
stimuli while clicking rocks together, active temporal clusters were still included, even if
their signal was not significantly higher than the motor baseline signal. There are
consequently six clusters that represent an effect of Task and four clusters that represent
where Oldowan and Acheulian toolmaking are modulated by the linguistic context of
previous training sessions (Figure 18). Each Group main effect was subsumed by an
interaction.
201
Table 7.Regions of significant activation (p < 0.05) as determined by a
two-way ANOVA between Group (verbal and nonverbal) and Task
(Oldowan and Acheulian).
Localization
MNI Coordinates (mm)
x
y
z
Volume
Mean ± SEM
Task main effect
Left
Superior temporal gyrus*
-60.8
-31.9
17.7
3600
6.18 ± 0.06
Left
Precentral gyrus
-31.7
-4.3
59.7
3584
6.26 ± 0.07
Right
Postcentral gyrus
46
-25.2
62
1688
5.07 ± 0.05
Right
Postcentral gyrus*
58.5
-14.7
32.3
1624
6.55 ± 0.12
Left
Precentral gyrus*
-50.2
5.8
33.5
1104
4.92 ± 0.05
Right
Middle temporal gyrus*
67.7
-33.6
2.8
536
4.39 ± 0.02
Right
Precentral gyrus
61.9
7
28.7
432
5.81 ± 0.20
Left
Supplementary motor area*
-9.9
1.4
75.7
352
4.73 ± 0.07
Left
Postcentral gyrus*
-50.7
-14.2
32.8
320
5.18 ± 0.11
Group main effect
Right
Rolandic operculum
63.4
-12.3
11.6
6904
7.03 ± 0.10
Left
Left
Right
Inferior parietal lobule
Superior frontal gyrus
Postcentral gyrus
-55.2
-22.5
36.3
-31.4
-0.7
-33.1
38.9
65.8
71
6312
5688
328
7.38 ± 0.08
6.48 ± 0.06
5.30 ± 0.13
Group x Task interaction
Right
Temporal pole*
57.3
9.6
-5.8
4968
6.83 ± 0.08
Left
Middle frontal gyrus*
-27.9
-1.4
64.9
4928
8.40 ± 0.13
Right
Supramarginal gyrus
63.7
-26
19.9
4008
6.39 ± 0.07
Left
Supramarginal gyrus
-55.5
-42.6
33
2456
5.13 ± 0.04
Right
Postcentral gyrus
46.7
-32
62.8
1864
6.88 ± 0.13
Right
Postcentral gyrus
60.3
-2
30.2
1192
5.78 ± 0.11
Right
Inferior frontal gyrus*
776
4.79 ± 0.05
51.4
37.2
13.5
Left
Precentral gyrus*
624
5.00 ± 0.06
-40.3
6.5
46.2
*Indicates the highest-order effect where knapping activation is also significantly higher than
motor baseline activation
202
Figure 18. Active clusters where an ANOVA Task Main Effect or Interaction Effect
occurs and its relative spatial relationship to the results of a previous neuroarchaeological
study (Stout et al. 2008) and a VWM meta-analysis (Wijeakumar et al. 2015). Color key:
Red-Task Main Effect or Interaction Effect; Purple-Results from Stout et al. (2008);
Light green-VWM network; Turquois-Overlap between Task Main Effect/Interaction
Effect and Stout et al.; Mauve-Overlap between Stout et al. and VWM network; YellowOverlap between Task Main Effect/Interaction Effect and VWM network; Dark blueOverlap between Task Main Effect/Interaction Effect, Stout et al., and VWM network.
Error bars in bar plots represent 95% confidence intervals.
203
The analysis revealed novel areas of activation associated with Acheulian and
Oldowan toolmaking that have not been reported in previous studies. In addition to
replicating the Acheulian-biased activation in the left ventral PrG from previous research
(Stout et al. 2008), which also overlaps with the MFG in the VWM network (Wijeakumar
et al. 2015), unique areas recruited during the Acheulian task include the supplementary
motor (SMA) and middle and superior temporal areas. The left superior temporal gyrus
(STG) is involved in complex sound processing and auditory short-term memory
(Vouloumanos et al. 2001; Leff et al. 2009), while the SMA forms the cognitive control
center of a medial premotor system of the brain whose primary functions are to plan
future movements based on an internally derived model of the world and to fluently
execute extended sequences of component movements, especially those requiring
bimanual coordination (Goldberg 1985). The STG and SMA are connected via white
fiber tracts that coalesce at the insular cortex (Augustine 1996). While the blood
oxygenation level in the insular cortex is too deep to record with fNIRS, the involvement
of this area during the observation of ESA tool production has been noted previously
(Stout et al. 2011).
Acheulian toolmaking depends on the execution of a skilled striking platform
setup to plan the direction, shape, and size of a series of flakes that will effectively thin
and shape the piece (Stout et al. 2014). Sound discrimination plays an important role in
tool-related planning among primates (e.g., Visalberghi et al. 2009). The selective
activation of the left STG during the Acheulian task may signify that the knappers were
holding in mind the varying sounds of impact that arrived in rapid succession to judge
whether a platform was successfully prepared for removal of a flake of predetermined
204
shape and size. The ability to plan and execute this flexible sequence of actions to meet
the desired outcome of a finished handaxe could be accomplished by integrating the
working memory component of left ventral PrG with the complex motor planning of
SMA and the auditory feedback of STG via the insular circuit.
Unique cortical areas recruited during the Oldowan task include the hand
representation portions of the primary sensorimotor cortex in both hemispheres. The
activation of sensorimotor areas near the lateral sulcus during the Oldowan task may
implicate the involvement of a lateral premotor system, which, unlike the medial
premotor system employed during the Acheulian task, is dependent on external visual
input to recognize and assign significance to external objects (Goldberg 1985). This is
unsurprising, as the only goal of the Oldowan task is to visually identify ideal platforms
and remove flakes until the core is exhausted.
The ANOVA revealed four clusters where a significant interaction exists between
Group and Task, signifying that instruction type received during training sessions had an
effect on the cortical areas recruited during Acheulian and Oldowan toolmaking. A posthoc test identified two areas where the Acheulian task significantly varied by group
(Figure 18). A large cluster that includes the right temporal pole and pars orbitalis is
activated among participants in the nonverbal group and suppressed in the verbal group
(Mann-Whitney U = 55.0; p = 0.009). The right temporal pole is a multimodal
association cortex involved with semantic processing (Wong and Gallate 2012) with
strong connections to pars orbitalis. The right orbital portion of the prefrontal cortex is
known to be involved in decision-making, especially as it pertains to reward-related
information (Rogers et al. 1999). This may indicate that the nonverbal group relied more
205
extensively on auditory and visuospatial feedback while planning actions related to
handaxe production. The verbal group, on the other hand, has significantly higher
activation than the nonverbal group in pars triangularis of the right IFG (Mann-Whitney
U = 198.0; p = 0.001), suggesting that the participants who received verbal instructions
during prior training sessions potentially retrieved lexical and semantic knowledge from
declarative memory to complete the task (Ullman 2006). The subjects’ responses during
an exit interview support this interpretation because individuals in the verbal group
reported higher incidences of inner speech while knapping than those in the nonverbal
group (Appendix C, Figure A1).
A post-hoc test identified two active Oldowan clusters where the task
significantly varied by group (Figure 18). As described earlier, the activation of
sensorimotor areas during Oldowan flake production indicates that this task depends
primarily on a lateral premotor system that detects target platforms on a core for flake
removal based on visual cues. The Oldowan task appears to come under increased
cognitive control, however, when the task has been learned in the absence of verbal
instruction. For example, it is only in the nonverbal group that the left MFG, or frontal
eye field (FEF), is activated (Mann-Whitney U = 33.0; p < 0.001). This area forms part of
the theorized dorsal visual attention network, where FEF controls the location of attention
during visual search by discriminating between targets and distractors and prepares an
appropriate response (Corbetta and Shulman 2002). The recruitment of this visual
attention network only among the participants who learned via silent imitation suggests
that learning to produce simple flakes without language requires increased attention to
visuospatial stimuli that relies upon increased planning.
206
Discussion
How do these novel results address the two hypotheses about the OldowanAcheulian transition that were posed at the beginning of this chapter, relating to working
memory and language evolution? The critical difference between the Oldowan and
Acheulian tasks is that Oldowan tool manufacture relies on the coordination of visual
attention and motor control to successfully remove simple flakes. Acheulian tool
manufacture, on the other hand, requires the integration of higher-order motor planning,
working memory, and auditory feedback mechanisms in order to attend to information
from multiple modalities (visual, auditory, touch), possibly in relation to shifting
attention to different goals involved during handaxe production relative to the one goal of
the Oldowan task. Both tasks, however, are similar to each other in that they employ
cortical areas known to be involved in working memory, a fact that was largely
overlooked in previous neuroarchaeological studies (Stout and Chaminade 2007; Stout et
al. 2008; Stout et al. 2011; see Figures 16 & 18); although, with the recruitment of the
FEF, it appears that the Oldowan task relies more upon selectively processing visual
information (attention) than maintaining the information in an active state (working
memory). Working memory is not a uniquely human feature, but modern humans have
been argued to possess an “enhanced working memory” that did not evolve until the late
Pleistocene, in part because of the apparent lack of evidence for working memory
involvement during ESA tool manufacture (Wynn and Coolidge 2011). It can now be
confirmed that even stone tool industries as ancient as the Oldowan and Acheulian
require the toolmaker to employ a VWM network to some extent; thus, the notion that an
207
enhanced working memory system evolved only after the appearance of modern H.
sapiens needs to be reconsidered.
If language evolved by co-opting the motor areas of the brain that were used first
for ESA stone tool manufacture, then the activation of the right IFG, for example, should
not be contingent upon whether one learned to knap stone tools via language instruction
or via silent imitation. Interestingly, the current study replicated the right IFG activation
patterns that are reported in previous neuroarchaeological studies but only among
participants in the verbal group. This result indicates that the presence of language
instruction while learning to knap indeed affects how the brain processes Oldowan and
Acheulian toolmaking tasks. By extension, previous claims for a technological origin for
language, based primarily on overlapping functional activation in IFG during
semantic/syntactic and ESA toolmaking tasks, lose support because the involvement of
the right anterior IFG during Acheulian handaxe production is the result of a modern
learning condition that may not have existed prior to the evolution of the modern human
species. Researchers should therefore be cautious when making claims about extinct,
potentially pre-linguistic species of hominin, especially regarding their linguistic abilities,
when these claims are based on the results of neuroarchaeological studies that do not
control for language in the learning context.
While the current results do not support a technological origin for language
operations in the right IFG, the selective activation of temporal areas during handaxe
production suggests that a co-evolutionary relationship between language and technology
may exist in this region. The STG has phylogenetically ancient functions related to the
semantic processing of conspecific vocalization features (Gil-da-Costa et al. 2006) and is
208
important for speech comprehension in modern humans (Vouloumanos et al. 2001;
Vigneau et al. 2011). If one’s ability to successfully shape a handaxe depends on the STG
to discriminate between the semantics of similar sounds arriving in rapid succession as
one knaps, then complex stone toolmaking may have played a role in fine-tuning the STG
for this purpose, which would be important for the evolution of speech processing. It is
also interesting to note that the Acheulian technocomplex coincided in timing with the
evolution of a derived middle ear anatomy in Homo that was more attuned to speech
frequencies (Martínez et al. 2013; Quam et al. 2013). The shared function of both
complex toolmaking and language for semantic categorization of auditory streams should
be explored further as a potential clue to the evolution of language.
Summary
The apparent differences in technological complexity between the earlier
Oldowan and later Acheulian industries of the ESA have generated hypotheses from
researchers that relate hominin cognition and language evolutionary change to this
technological transition. Functional neuroimaging during the manufacture of replicative
ESA tools offers insights into the cognitive neural networks that were required by early
humans to carry out these behaviors. This study used fNIRS to measure the brain activity
of two groups of adult human participants as they made Oldowan and Acheulian stone
tools, after attending training sessions that included skill transmission via verbal
instruction or nonverbal imitation. The results reveal that the Acheulian task requires
higher-level motor planning than the Oldowan task, which is accomplished by integrating
209
information from multiple modalities via a cognitive control network. This study
demonstrates that working memory is, in fact, a necessary component of Acheulian
toolmaking, despite claims that ESA tool production can be accomplished by procedural
memory alone (Wynn and Coolidge 2011). Moreover, these data do not support a
technological origin for language operations in the right IFG, but selective activation of
temporal language-processing areas during the Acheulian task lends support to a potential
co-evolutionary relationship between language and technology in the temporal lobe.
210
CHAPTER 7:
THE FUNCTIONAL NEUROANATOMY OF LEARNING TO MAKE EARLY
STONE AGE TOOLS
Introduction
Learning a new skill usually proceeds through two stages, an initial acquisition
stage characterized by rapid increases in performance early in one’s training, followed by
a later stage characterized by slow, incremental gains in performance after several
sessions of practice (Floyer-Lea and Matthews 2005). This transition is generally thought
to reflect a shift from the slow, effortful controlled processing typical of novice
performance to the fast, low-effort automatic processing typical of skilled performance
when the task no longer requires active attention (Kelly and Garavan 2005).
According to the dual processing theory, the domain-general, prefrontal cognitive
control areas would be expected to decrease with experience until they drop out entirely
(Haier et al. 1992; Chein and Schneider 2005). A decrease in functional activation of
prefrontal areas could reflect the recruitment of fewer neurons or neural circuits and
therefore may mean that fewer cognitive strategies are employed after practice (Garavan
et al. 2000). Or it could reflect the development of a more efficient role of the prefrontal
cortex through the fine-tuning of its connections with other areas (Dux et al. 2009).
Alternatively, experience or training may lead to a reorganization of brain circuits that are
recruited for task performance, where the functional neuroanatomy present early in
training is replaced by different neuroanatomy late in training (Kelly and Garavan 2005).
211
This transition reflects the employment of different cognitive processes at different points
in learning. Functional neuroanatomical changes associated with learning may depend on
the complexity of the task being learned, however. In a functional magnetic resonance
imaging (fMRI) study conducted by Garavan and colleagues (2000), participants learned
a visuospatial working memory task over a comparable training period of 8 h. They find
that extensive practice does not alter the cognitive demands of the task, and working
memory areas remain consistently engaged throughout training. These results indicate
that cognitively demanding tasks do not necessarily fit the expectations of the learning
models listed above, which assume that a decline in the activation of cognitive control
areas will occur with continued practice.
There exist only a limited number of studies that look at complex motor learning
over an extended training period. Chapter 6 discussed the results of a neuroimaging study
that investigated verbally-communicated instruction versus nonverbal demonstration of
knapping skills and their effect on the hemodynamic activation of Oldowan and
Acheulian replicative toolmaking after ~8 h of practice. In sum, differential functional
activation was attributed to the two knapping tasks, with the Oldowan task engaging
primary sensorimotor areas and the Acheulian task engaging a frontotemporal cognitive
control network. Participants who learned to knap without language instruction relied
more extensively on a dorsal visual attention network during the Oldowan task and
auditory feedback during the Acheulian task. Conversely, those who learned with
language instruction recalled lexical and semantic information from declarative memory
during the Acheulian task.
212
These results only reflect one point in time, however. Without observing the
functional neuroanatomy of these tasks and conditions over an extended period of time, it
is unclear whether these patterns of functional activation accurately represent Oldowan
and Acheulian toolmaking under verbal and nonverbal conditions in general, or if they
simply represent the processing of these skills during one particular point in training.
Participants in the study spent ~2 h learning how to make the simple flakes that are
characteristic of the Oldowan industry before moving on to learn the skills necessary for
making Acheulian handaxes because skills associated with simple flaking must be
acquired before handaxe manufacture can be attempted. It is possible that the differential
activation patterns associated with Acheulian and Oldowan toolmaking in the final
neuroimaging session simply reflect two different stages of learning, controlled
processing and automatic processing, respectively. The extent of the hand representation
in the primary motor cortex enlarges after extensive practice of complex motor tasks
(Karni et al. 1995; Hund-Georgiadis and von Cramon 1999; Ungerleider 2002). The
occurrence of functional activation of the primary hand motor cortex in the left and right
hemispheres and the general absence of prefrontal engagement during the Oldowan task
may therefore indicate that this task has already transitioned to being processed by
procedural memory by the third neuroimaging session. The Acheulian task, on the other
hand, may not have made this transition to automatic processing by the third session, as is
indicated by its engagement of the supplementary motor area (SMA). Activation in the
SMA during a complex motor task tends to diminish after practice (Hund-Georgiadis and
von Cramon 1999). It is also possible that if handaxe production is a working memory
213
task, then activation in cognitive control areas may not diminish, even after extensive
practice.
To further test the hypothesis that Acheulian stone tool production requires
greater recruitment of executive functions than Oldowan stone tool production, follow-up
analyses were conducted that involved 33 adult subjects learning to produce replicative
Oldowan and Acheulian stone tools over the course of ~8 h, while neural activity
associated with these tasks was simultaneously recorded with functional near-infrared
spectroscopy (fNIRS) at three different points in their training. As a working memory
task, Acheulian toolmaking should consistently elicit activation in working memory
areas, even after several hours of training, while Oldowan toolmaking, if not a working
memory task, should display signs of decreasing involvement of cognitive control areas
over time and an increase in the engagement of primary motor areas. Furthermore, the
effect of linguistic instruction versus nonverbal imitation on the recruitment of the
supposedly domain-general cognitive control network at different stages in learning is
investigated. The implications for hominin brain evolution are then discussed.
Summary of methods
This section provides a short summary of the methods employed to conduct this
study. For an in-depth discussion of the participants, methods, and materials, see Chapter
5 and Appendix B. Thirty-three healthy, right-handed subjects (17 females, 16 males; age
[mean ± SD] 23.8 ± 7.9 years) with no previous training in stone knapping attended
seven 60-m training sessions to learn skills associated with Oldowan and Acheulian tool
214
manufacture. At different points in their learning (after the first, fourth, and seventh
training sessions), participants attended three 90-m fNIRS neuroimaging sessions.
Participants were divided into two groups, which determined their context of Early Stone
Age (ESA) toolmaking skill transmission. One group watched instruction videos in which
the instructor delivered the lessons with verbal instructions (n = 17; 9 females, 8 males).
The other group watched the same videos but with the sound turned off so that skills were
learned via imitation rather than verbal instructions (n = 16; 8 females, 8 males). The
instructor’s face was not visible in any of the videos to eliminate linguistic cues.
Neuroimaging sessions consisted of a baseline motor task, an Oldowan task, and
an Acheulian task. Participants knew to transition from one task to another from a series
of tones that were played through the computer and by watching silent ~30-s clips that
demonstrated how to carry out the task. The motor baseline task consisted of three
different arm movements while holding two hammer stones: direct percussion such that
both hands clicked the rocks together simultaneously, glancing percussion such that the
dominant hand clicked one rock against the other that was held stationary, and a smaller
grinding motion that rubbed one rock against the other. Participants matched their
movement to an externally heard pace of 60 beats per minute and continued to keep this
pace internally once the metronome dropped out after 10 s. The Oldowan and Acheulian
tasks were performed in an alternating task and rest manner such that participants carried
out naturalistic knapping during 60-s intervals followed by 15-s rest intervals.
215
Preliminary analysis
NIRS data were motion corrected and reconstructed in image space within the
brain volume prior to statistical analysis (see Chapter 5 and Appendix B for details on
this process). The purpose of the motor baseline task was to create a functional map for
neural activity associated with general knapping movements but without the added
element of flake production. Those areas where knapping showed significantly higher
activation than the motor baseline were identified as areas of the brain that are unique to
knapping behaviors rather than general motor behaviors.
The idea behind the motor baseline was that there should not be a large degree of
learning across sessions and that its signals should remain stable over time. Initial
analyses of the motor baseline data revealed, however, that performance varied from
session to session. Therefore, a Pace (external, internal) x Session (1-3) analysis of
variance (ANOVA) was performed for each of the three conditions of the baseline task
(Direct, Glancing, and Grinding) to identify which condition has the fewest number of
session-related effects, for the purpose of identifying a stable motor baseline to contrast
with the knapping tasks (Table 8). There was a significant effect of Session for each
condition (F = 3.153, p < 0.05). The glancing condition was selected as the baseline for
this study because its combined significant clusters have the fewest number of significant
voxels of the three conditions in the Session effect, meaning this condition remains the
most stable over time. Two separate ANOVA tests were then performed for internal and
external pacing during the Glancing condition to test for learning changes. The different
pace conditions enlist a similar number of voxels in the Session effect; therefore, both
216
internal and external pacing during glancing percussion were included as part of the
motor baseline.
Table 8. Total number of significant voxels
with a Session effect for each baseline
condition.
Condition
Direct percussion
Glancing percussion
Grinding motion
Voxel count
28,600
10,304
17,336
Statistical analysis
Data from the knapping tasks from all three sessions were analyzed with a Task
(Oldowan, Acheulian) x Group (verbal, nonverbal) x Session (1-3) ANOVA. The internal
and external glancing motor baseline hemodynamic signals were extracted using the
center of mass coordinates from each significantly active knapping cluster. The average
knapping values for each significant cluster were then compared to that region’s averaged
motor baseline values, using the Wilcoxon signed-rank test. Any clusters where knapping
had significantly higher activation than the motor baseline were determined to be
knapping regions of interest and are presented in Table 9. All other significant clusters
with similar or lower activation than the motor baseline, excepting temporal areas, are
presented in Appendix B, Table A2.
217
Results
The three-way ANOVA reveals significant clusters for each of the main effects
and their interactions (Table 9). Results are organized into two sections: the first section
addresses whether the aggregate results included from all three sessions replicate the
activation patterns described for the third neuroimaging session (see Chapter 6); the
second section explores how knapping skills learned over time are reflected in the
functional anatomy of the brain and how this varies based on the modality of instruction
and the type of stone tool being produced.
218
Table 9. Knapping areas active at different points in skill learning.
Localization1
MNI Position (mm)
(x, y, z)
Vol.
Mean ± SEM
21.9, -34.4, 78.7
54.0, 17.8, 16.5
880
248
5.45 ± 0.09
4.87 ± 0.10
-47.9, 16.1, 35.9
-58.9, 22.1, 12.5
5360
480
6.90 ± 0.07
7.04 ± 0.31
-28.1, -3.9, 65.7
-41.6, 9.1, 48.6
-53.9, -7.8, 29.4
4071
1632
328
6.12 ± 0.06
5.63 ± 0.08
5.22 ± 0.13
-23.5, -18.8, 75.3
568
4.05 ± 0.08
36.9, 19.3, 56.6
-49.2, 29.9, 23.3
-29.3, -21.7, 70.9
2856
752
680
4.33 ± 0.04
3.62 ± 0.03
3.76 ± 0.05
36.5, -8.1, 64.3
-37.0, 18.5, 52.2
68.5, -38.6, 6.2
37.1, -33.7, 69.9
3440
1552
400
232
4.34 ± 0.04
3.65 ± 0.03
3.43 ± 0.03
4.00 ± 0.10
Group Main Effect
Right Postcentral gyrus
Right Inferior frontal gyrus
Task Main Effect
Left
Left
Middle frontal gyrus
Inferior frontal gyrus
Group x Task Interaction Effect
Left
Left
Left
Superior frontal gyrus
Precentral gyrus
Postcentral gyrus
Session Main Effect
Left
Precentral gyrus
Group x Session Interaction Effect
Right Middle frontal gyrus
Left Inferior frontal gyrus
Left Precentral gyrus
Session x Task Interaction Effect
Right
Left
Right
Right
Precentral gyrus
Middle frontal gyrus
Middle temporal gyrus
Postcentral gyrus
Group x Session x Task Interaction
Left Supramarginal gyrus
-56.3, -45.0, 34.4
3752
4.47 ± 0.04
Left Superior temporal gyrus
-61.6, -10.6, 5.2
2000
3.68 ± 0.02
Left Postcentral gyrus
-59, -16.3, 31.8
1768
5.09 ± 0.08
Right Superior temporal gyrus
65.3, -10.8, -9.7
1176
3.82 ± 0.03
Right Precentral gyrus
46, -10.2, 53.5
920
3.55 ± 0.03
1
Areas listed include significant clusters (p < 0.05) from the Task x Group x Session
ANOVA that are also significantly higher than the motor baseline.
219
Overall group and task differences
In Chapter 6, the effects of Group and Task and their interaction were explored
during the final neuroimaging session. Many of the effects reported during the final
session are also present in the combined dataset from three points in training (after 1 h,
~5 h, and ~8 h of training). A total of six clusters were identified for the contrast between
the Oldowan and Acheulian tasks. In two of these clusters, neural activity associated with
the Acheulian task was significantly higher than that associated with the motor baseline
task, including the left middle frontal gyrus (MFG; Z = 451; p = 0.002) and left pars
triangularis in the inferior frontal gyrus (IFG; Z = 402; p = 0.030). The Task effect in the
MFG was driven by deactivation during the Oldowan task (Figure 19, Panel A), while in
the IFG, it was driven by activation during the Acheulian task (Figure 19, Panel B). The
left MFG cluster overlaps with the MFG portion of the visual working memory (VWM)
network (Wijeakumar et al. 2015). This cluster is large, extending anteriorly from the
precentral gyrus (PrG) into MFG and inferiorly into IFG, where it also overlaps with a
semantic/syntactic processing language area (Vigneau et al. 2011). As was noted in
Chapter 6, this cluster is relegated to only the PrG during the third session, where it
overlaps with the ventral precentral region reported by Stout and colleagues (2008). The
left IFG cluster overlaps with the dorsolateral prefrontal cortex (dlPFC) portion of the
VWM network (Wijeakumar et al. 2015) and is not represented in the results from the
third neuroimaging session. This decrease in volume in prefrontal areas over time likely
indicates the involvement of fewer neurons and connections as the system becomes more
efficient with learning and practice, but there are no signs that these areas drop out
220
completely. In fact, there is increased Acheulian activation in the left PrG portion of the
MFG cluster during the third session. The active clusters of this ANOVA do not include
the left supplementary motor area (SMA), which forms part of the Acheulian circuit
during the third neuroimaging session. It is even more apparent from this larger sample,
however, that, relative to the Oldowan task, the working memory areas of the left dlPFC
are critically important for the early stages of learning how to make Acheulian handaxes.
Figure 19. Acheulian activation in the dlPFC revealed in the Task main effect (red) and
these clusters’ spatial relationship to language centers (Vigneau et al. 2011; dark green),
VWM areas of the brain (Wijeakumar et al. 2015; light green), and the results of a
previous neuroarchaeological experiment (Stout et al. 2008; purple). Panel A displays the
activation center in the left MFG. Note that this area overlaps with both a VWM and
knapping area from a previous experiment (blue) and overlaps with a language area
(pink). Panel B displays the activation center in the left IFG, which overlaps with a
VWM area (yellow). Error bars represent 95% confidence intervals.
221
Four clusters were identified for the Group main effect. Of these clusters, only
two have significantly higher activation during the knapping tasks compared to the motor
baseline task (Figure 20). These include the right IFG (pars triangularis), where
knapping activation is significantly higher than motor baseline activation in the verbal
group (Z = 24; p = 0.013) and right postcentral gyrus (PoG), where the difference
between the knapping and the motor baseline tasks was significant in the nonverbal group
(Z = 118; p = 0.010). In both cases, the hemodynamic signal is higher among participants
in the nonverbal group. The right IFG cluster also overlaps with a language center that
comes online during syntactic tasks (Vigneau et al. 2011). Neither of these clusters
overlaps with any of the active clusters reported in Chapter 6. And because activation in
these clusters is not modified by Task, this indicates that these two areas probably play a
specialized role either in imitative learning or simply trying to learn a complex motor
skill without the aid of language.
222
Figure 20. Learning context activation differences as revealed by the Group main effect
(red) and their spatial relationship to language centers of the brain (Vigneau et al. 2011;
purple). Panel A displays activation in the right IFG. Note that this area overlaps with a
language center (turquois). Panel B displays activation in the right PoG. Error bars
represent 95% confidence intervals.
There are eight significant clusters where an interaction between Group and Task
occur. Three of these clusters, all occurring in the left hemisphere, have a significantly
higher signal during one of the knapping tasks compared to the motor baseline task.
Oldowan task activation is significantly higher than motor baseline activation in the
nonverbal group for the superior frontal gyrus (SFG; Z = 25; p = 0.026) and in the verbal
group for the PrG (Z = 140; p = 0.003). Verbal group activation during the Acheulian
task is significantly higher in the PoG than motor baseline activation (Z = 122; p =
0.031). Both the SFG and PrG clusters overlap with corresponding active clusters from
the third session and have the same interaction effect. The left SFG is only recruited by
the nonverbal group during the Oldowan task (Figure 21, Panel A), and the left PrG
223
forms part of the Oldowan network but only in the verbal group (Figure 21, Panel B). The
left PoG also has stronger activation during the Oldowan task than the Acheulian task
among participants in the nonverbal group (Figure 21, Panel C). No interaction effect was
detected for the right IFG in this larger sample, as was seen in the results from the third
session (Figure 18).
Figure 21. Significant clusters with an interaction between Group and Task (red) and
these clusters’ spatial relationship to VWM areas of the brain (Wijeakumar et al. 2015;
light green) and the results of a previous neuroarchaeological experiment (Stout et al.
2008; purple). Panel A displays the activation center in the left SFG. Note that this area
overlaps with a VWM cluster (yellow) and a knapping area from a previous experiment
(turquois). Panel B displays the activation center in the left PrG. Panel C displays the
activation center in the left PoG. Error bars represent 95% confidence intervals.
224
Session differences
The results of the ANOVA reveal changes in the functional neuroanatomy
associated with learning to make stone tools over a period of several weeks. There were
three significant clusters identified for the Session main effect. Of these clusters, only the
left PrG has significantly higher activation during the knapping tasks than during the
motor baseline task (Z = 473; p = 0.001). This cluster experiences the most robust
activation during the first neuroimaging session, which subsequently decreases in the
second and third sessions (Figure 22).
Figure 22. Left PrG cluster involved in early-stage learning of ESA tool manufacture, as
determined by the Session main effect. Error bars represent 95% confidence intervals.
The ANOVA identified a total of eight significant clusters where an interaction
between Task and Session occurred. Four of these clusters have a significantly higher
signal during one or more of the knapping tasks compared to the motor baseline task
(Figure 23). In the right PrG, Oldowan activation is stronger than the motor baseline
during the first session (Z = 393; p = 0.016). This cluster overlaps with the dorsal
precentral ROI in a previous neuroarchaeological study (Stout et al. 2008) and displays a
225
pattern of activation associated with the Oldowan task in the first session that then
decreases in strength in later sessions, while Acheulian activation increases from the first
to third session (Figure 23, Panel A). A cluster in the left MFG has significantly higher
Oldowan activation than motor baseline activation during the first session (Z = 407; p =
0.007), as well as higher Acheulian activation during the second session (Z = 400; p =
0.033). This cluster shows a similar pattern of activation to the right PrG (Figure 23,
Panel B). Together, these two areas cover a significant portion of the frontal eye fields
(FEF). The same right middle temporal gyrus (MTG) cluster that was selectively active
during the third session analysis shows an increase in Acheulian activation from the first
to third session in the current analysis (Figure 23, Panel C). Because the activation in this
area becomes increasingly stronger with experience, this explains why the ANOVA
performed on the session 3 data alone identified this cluster, but it did not show up as an
effect of Task when all three sessions were included in the 3-way ANOVA. In the right
PoG, Acheulian activation is significantly higher than the motor baseline during the
second (Z = 413; p = 0.018) and third sessions (Z = 373; p = 0.048). This area is
deactivated during the first session for the Oldowan task but increases in strength by the
second session, while the Acheulian signal declines after the first session (Figure 23,
Panel D).
226
Figure 23. Significant clusters with an interaction between Task and Session (red) and
their spatial relationship to the results of a previous neuroarchaeological experiment
(Stout et al. 2008; purple). Panel A displays the activation center in the right PrG. Note
that this area overlaps with an active cluster from the previous experiment (turquois).
Also pictured are the left MFG (Panel B), right MTG (Panel C), and right PoG (Panel D).
227
There are six significant clusters where an interaction between Group and Session
occurs. Three of these clusters have a significantly higher signal during the averaged
Oldowan and Acheulian knapping tasks than during the motor baseline task in at least
one session, including the right MFG (Z = 106; p =0.049), left pars triangularis in the
IFG (Z = 122; p = 0.005), and left PrG (Z = 131; p = 0.010). There is also an interaction
effect in the right superior temporal gyrus (STG), which is included here despite the
knapping tasks not having a significantly higher signal than the motor baseline task
because auditory stimuli were not controlled in the motor baseline task. The right MFG is
strongly recruited by the nonverbal group during the first session but drops out in
subsequent sessions (Figure 24, Panel A). The right STG comes online only during the
third session in the nonverbal group, which confirms the finding from Chapter 6 that the
nonverbal group is more tuned in to auditory information while knapping than the verbal
group (Figure 24, Panel B). The left IFG cluster, which overlaps with a language center
(Vigneau et al. 2011), activates during the first and second sessions among nonverbally
instructed participants, but it is only active among the verbally instructed participants
during the third session (Figure 24, Panel C). Finally, the left PrG displays a similar
pattern to the right MFG, with activation occurring during the first session, but in the
verbal group rather than nonverbal group (Figure 24, Panel D). These results indicate that
the context in which a new motor skill is learned, either with verbal instruction or
nonverbal imitation, affects the cognitive strategies used to attend to the task. In other
words, differences in motor planning over time may be governed by the varying
recruitment of cognitive control centers due to learning context.
228
Figure 24. Significant clusters with an interaction between Group and Session (red) and
these clusters’ spatial relationship to language centers of the brain (Vigneau et al. 2011;
purple). Panel A displays the activation center in the right MFG. Panel B displays the
activation center in the right STG. Panel C displays the activation center in the left IFG.
Note that this area overlaps with a language center (turquois). Panel D displays the
activation center in the left PrG. Error bars represent 95% confidence intervals.
229
The ANOVA identified a total of five significant clusters where an interaction
between Task, Group, and Session occurred, all of which have a significantly higher
signal during one or more of the knapping tasks compared to the motor baseline task,
excepting a left STG cluster (Figure 25). These include the left supramarginal gyrus
(SMG; Z = 113; p = 0.020), left PoG (Z = 139; p = 0.003), right STG (Z = 134; p =
0.006), and right PrG (Z = 100; p = 0.023). The left SMG overlaps with the VWM
network but shows no clear task or group distinction in its activation (Figure 25, Panel
A). Otherwise, it appears that the primary sensorimotor areas are mainly recruited during
the Oldowan task. The left PoG and right PrG, the latter of which overlaps with the dorsal
PrG cluster reported in a previous neuroarchaeological study (Stout et al. 2008), become
active for the Oldowan task among the participants in the nonverbal group only during
the first session and then fall out in subsequent sessions (Figure 25, Panels B-C). The
temporal areas, which overlap with language-processing centers, are more strongly
activated among participants in the nonverbal group (Figure 25, Panels D-E).
230
Figure 25. Significant clusters with an interaction between Group, Task, and Session
(red) and their spatial relationship to the results of a previous neuroarchaeological
experiment (Stout et al. 2008; purple), language centers (Vigneau et al. 2011; dark
green), and VWM areas of the brain (Wijeakumar et al. 2015; light green). Panel A
displays the activation center in the left SMG. Note that this area overlaps with the VWM
network (yellow). Panel B displays the activation center in the left PoG. Panel C displays
the activation center for the right PrG, which overlaps with an active cluster from a
previous experiment (turquois). Panel D displays the activation center in the right STG,
which overlaps with a language-processing center (pink). Panel E displays the activation
center in the left STG. This area also overlaps with a language-processing center.
231
Discussion
The present study measured the oxygenated hemoglobin (oxy-Hb) levels in the
frontal, parietal, and temporal cortices of subjects’ brains at three different points in their
training as they learned to make replicative Oldowan and Acheulian stone tools after
having received either verbal or nonverbal instructional training. The major difference
between the Oldowan and Acheulian tasks is the recruitment of higher-order cognition
areas of the brain during the Acheulian task. These areas include two clusters within the
left dlPFC that were significantly more active than the motor baseline. The Oldowan task
displays deactivation in the left MFG relative to the Acheulian task. The fact that this
area is deactivated during the Oldowan task probably means that Oldowan tool
production does not require a large degree of deductive reasoning to reach the end goal
(i.e., there is an unambiguous sequence of flakes that can be removed). Oldowan
toolmaking does not require the maintenance of active representations of several
alternative actions nor the integration of information across events. This is because the
goal of the task is to remove flakes from a core until the core is expended, which does not
require the knapper to attend to multiple subgoals simultaneously in order to reach a final
goal, as is most likely the case when making a handaxe (Mahaney 2016). Altogether, the
results of the present study highlight a critical divide between the processing necessary to
carry out Oldowan toolmaking versus Acheulian toolmaking, where Acheulian tool
production requires the recruitment of a working memory network.
The functional map of ESA toolmaking depends on one’s learning context. Two
clusters are significantly higher than the motor baseline in the Group main effect,
232
including the right PoG and pars triangularis. The overall effect of Group, including the
clusters that were not determined to be significantly different from the motor baseline
(see Appendix B), shows increased activity among the nonverbal group. The profusion of
active areas of the brains of participants who learned by imitation alone highlights the
apparent neural and potentially metabolic inefficiency of this mode of learning in the
context of stone knapping in comparison to learning via verbal instruction. In other
words, these areas likely come online in the nonverbal group to compensate for learning
without language rather than representing specialized brain areas for imitative learning.
For example, even though pars triangularis is close in proximity to pars opercularis,
which is known to participate during imitation tasks (Heiser et al. 2003), imitative
learning does not appear to elicit activation in pars triangularis (Molnar-Szakacs et al.
2005). It is possible that the inefficiency of nonverbal, imitative learning in this case is an
artifact of modern humans’ extreme reliance on language for learning new skills,
meaning that the nonverbally-instructed participants in the study may have needed to
devote increased attention to the task to achieve the same end product results as the
verbally-instructed participants. It is difficult to say at this point whether this would have
been the case for pre-linguistic hominins, however.
The number of interactions between Group and Task in this larger sample was the
same as that found for the final session data. Two of the areas in this analysis overlap
with frontal clusters in the prior analysis, including the left SFG and PrG. The mode of
instruction has a strong effect on frontal involvement during Oldowan tool production.
While nonverbal instruction leads to increased participation of the SFG, verbal
instruction results in higher activity in the PrG. The strong activation of the SFG might
233
reflect a higher degree of uncertainty on the part of the nonverbally-instructed knappers,
as activation in this area has previously been described in a study when participants
experience uncertainty (Volz et al. 2005), but it is more likely that it acts as a control
center for directing visual attention, thus guiding behavior in the absence of verbal
instructions. Overall, these significant interactions reveal that the mode of instruction
while learning to knap, and potentially while learning other complex motor tasks as well,
leads to differential involvement of cognitive networks. Thus, caution should be taken
when making interpretations about extinct hominin cognition based on the results of
neuroarchaeological studies that do not control for language instruction.
Early Stone Age learning networks
The data so far indicate that Oldowan and Acheulian tool manufacture enlist
different neural networks after some training, but it is also important to consider the
possibility that these apparently different networks may reflect the engagement of the
same network for both tasks but at different stages in learning. The dual processing
theory, for example, assumes that a set of association areas of the brain mediate the
controlled processing critical for novice performance during the first stage of learning.
Then these areas become less essential after automaticity of the task develops in the
second stage of learning. It is possible that the Oldowan task already underwent a
transition to procedural memory by the third neuroimaging session. The deactivation of
dlPFC and the recruitment of mainly sensorimotor areas during the third session may
234
signify a shift from cognitive control strategies to internally generated representations of
movements based on procedural memory (Grafton et al. 1994).
Participants in the present study learned how to make simple Oldowan flakes
during their first and second practice sessions and thus had ~8 h of practice making flakes
by the final neuroimaging session, while specific skills for handaxe manufacture were not
demonstrated in an instruction setting until the third practice session. So, the participants
only had ~6 h of practice attempting to make Acheulian handaxes by the third
neuroimaging session. Could the differential activation between the Acheulian and
Oldowan tasks by the final neuroimaging session simply reflect two different stages of
learning, controlled processing and automatic processing, respectively? If Oldowan and
Acheulian tool production are operating under these dual learning stages by the final
neuroimaging session, this could explain the evidence for activation of a working
memory network during the Acheulian task and the absence of one during the Oldowan
task. If this is the case, then the Oldowan task should exhibit a similar set of cognitive
control areas during the first neuroimaging session that then declines in strength as
controlled processing transitions to automatic processing over the next two neuroimaging
sessions.
The interaction effect between Task and Session reveals the presence of a visual
attention network of regions involved in Oldowan tool manufacture during the first
neuroimaging session. This network includes bilateral activation in the FEF. The
Oldowan task also weakly recruits the right MTG in the first session, whose signal then
decreases in the next two sessions. By the second session, all of these regions, including
the potential cognitive control areas, drop out entirely. In their place arises the right PoG,
235
a primary sensorimotor area, which reaches its peak activation after the first session. The
Task effect displays deactivation in the left dlPFC during the Oldowan task. The FEF and
auditory sensory MTG become deactivated, while a sensorimotor area becomes
increasingly engaged. This pattern of deactivation of cognitive control areas and
activation of a sensorimotor area may be a product of the transition to automaticity and
decreased attention after 4 h of practice. By this point, the simple removal of flakes
without the added element of shaping a core tool was well rehearsed and therefore did not
demand the active attention of the participants. A number of deeper brain structures are
associated with procedural memory that could not be recorded using fNIRS, including the
cerebellum (Molinari et al. 1997); however, Stout and Chaminade (2007) note cerebellum
involvement during Oldowan toolmaking after 4 h of practice.
Based on the results for the Oldowan task, one might expect to see signs of a
similar decline in control network involvement for the Acheulian task by the third session
because by this point, the participants had a similar number of hours of formal training on
this task as they had on the Oldowan task by the time it reached automatic processing;
however, the dlPFC clusters identified in the Task main effect, discussed above, suggest
that Acheulian tool production may be a cognitively demanding, working memory task.
Garavan and colleagues (2000) demonstrate that cognitive control areas do not decline in
strength or fall out after training if the task is a complex, working memory task; in fact,
some studies report that the activation of prefrontal areas actually increases after practice
during working memory tasks (Olesen et al. 2003; Westerberg and Klingberg 2007). If
the activation pattern differences between the Oldowan and Acheulian tasks are the result
of task complexity rather than learning stage differences, then cognitive areas of the brain
236
involved in Acheulian tool manufacture should remain consistently active or become
increasingly more active during each successive session.
Whereas there is a decrease in the engagement of cognitive control areas, motor
planning, and temporal areas and an increase in primary sensorimotor areas during the
Oldowan task over time, the Acheulian task displays no such trend. Rather, the left MFG,
right premotor area, and right MTG increase with additional practice, and activation in
the right PoG decreases with more training. This pattern of increasing activation of the
control network and coinciding deactivation of a primary sensorimotor area during
Acheulian tool production indicates an emphasis on the employment of cognitive
strategies at all measured stages of learning. Combined with the Task effect, this suggests
that handaxe production is a visuospatial working memory task that consistently engages
working memory areas even after extensive training. Furthermore, these results
strengthen the claim made in Chapter 6 that Oldowan and Acheulian tool manufacture are
separate processes in the brain. There is the possibility, however, that because Oldowanand Acheulian-associated activation patterns were measured from the same participants,
then there is no clear way to distinguish between the effects of these two tasks, as the
participants may have been incorporating similar strategies for both tasks that they would
not have otherwise done had they only learned how to do one type of task. A future study
that compares one group that learns to make Oldowan stone tools to a separate group that
learns to make Acheulian stone tools may provide a clearer picture of hemodynamic
responses associated with each of these tasks.
237
Learning context and neural network change
The context in which participants learn stone knapping skills, whether verbally or
nonverbally, leads to the differential recruitment of certain cortical areas at different
learning stages. In particular, there are prefrontal areas that experience peak activation at
different points in time for the two groups. For example, the group that received
nonverbal instructions engages the right MFG and left IFG (pars triangularis) during the
first and second sessions more than the group that received verbal instructions. By the
third session, the verbal group recruits the IFG cluster more than the participants in the
nonverbal group. Conversely, the verbal group recruits a motor area during the first
session in the left PrG. Awareness and monitoring of one’s movements are implemented
within the same primary sensorimotor areas that are responsible for initiating said
movements (Berti et al. 2005). It is possible that the participants in the verbal group are
more aware of their controlled movements in relation to the task earlier on in the
experiment because of the verbal instructions they received, while the participants in the
nonverbal group are more reliant on cognitive strategies involving MFG and IFG regions
in the earliest stage of learning.
The dual processing theory for learning assumes a domain-general control
network. For example, Chein and Schneider (2005) note that verbal and nonverbal tasks
elicit activation in overlapping cognitive control regions. Under these premises, it should
be assumed that the linguistic context of instruction would have little influence on the
learning networks employed during ESA skill acquisition. This is not the case in this
particular study, however, as the presence or absence of language during training appears
238
to dictate which higher-order association areas of the brain become engaged and at what
point in training, which likely also has top-down effects on lower-level perception and
motor control.
Hominin cognition during the Early Stone Age
By looking at how ESA toolmaking skills are learned over time, the claim that
different cognitive components are involved in producing Oldowan and Acheulian type
tools garners further support. The act of making simple flakes does not appear to
emphasize working memory, problem solving, or decision-making when the skill is first
being learned, though it does enlist the dorsal visual attention network at this point. This
visual attention network selects visual information via eye movements that shift attention
to salient features on the stone. The task is then rapidly automatized after a few hours of
practice, whereas handaxe production relies on a cognitive control network, even after
several hours of training. The multiple layers of goals and subgoals required to make a
handaxe actively engage working memory. This form of hierarchical learning is argued
by several authors to be a necessary precursor for the evolution of language (Greenfield
1991; Christiansen and Kirby 2003; Arbib 2012).
This study demonstrates that, in modern humans, there is a critical difference
between Oldowan and Acheulian toolmaking cognition because of the amount and kinds
of information that have to be stored and manipulated in order to complete these tasks.
Because the Oldowan and Acheulian tasks in this experiment represent different levels of
complexity, it is probable that different cognitive strategies were employed in the past as
239
well while producing these different tool types. While it cannot be confidently claimed
that pre-Acheulian hominins lacked complex cognition, such as enhanced working
memory and problem solving capabilities, based on these results, if the rule of minimum
competence is applied, it must be assumed that hominins who made Acheulian handaxes
possessed these qualities, or otherwise they would not have been successful at preparing
such tools. It is therefore no longer tenable to assume that early Homo mindlessly
reduced cores into standard handaxe shapes time and again without being aware of their
goals and actions and the future consequences of their actions, as Moore suggests (2010,
2011). Although the existence of these cognitive abilities in pre-Acheulian hominins
cannot necessarily be dismissed simply because of their lack of evidence, the sudden
increase in brain size around the time of the advent of the Acheulian industry suggests
that a functional reorganization and resultant structural reorganization of the brain
evolved around the time that H. erectus first appears. If this is the case, then these results
further substantiate a cognitive transition around the time that H. erectus evolved.
It is impossible to know whether the Oldowan and Acheulian knapping networks
identified in modern-day participants as they make ESA stone tools are the same
networks that early Homo employed as they engaged in similar behaviors. On the other
hand, the sulcal/gyral and asymmetry patterns on the endocasts of early Homo resemble
the derived pattern of H. sapiens (Tobias 1987; Balzeau et al. 2012), and chimpanzees
display active default neural networks during a resting state that are similar to those
displayed by humans (Rilling et al. 2007). These examples illustrate that although there
are likely to be some species-specific differences in brain function, overall, the general
functional neuroanatomy of early Homo was probably similar to that of modern Homo.
240
Summary
Neuroimaging experiments that study conceptual-level representations of
acquired tool use skills, such as identifying tools (Martin et al. 1995; Martin et al. 1996)
and generating and identifying action words associated with tools (Grabowski et al.
1998), are common in the neuroscience literature. It is rare, however, to find studies that
use neuroimaging methods to investigate mechanisms involved in producing tool use
skills, most likely because of the constraints of the most commonly used neuroimaging
techniques (Johnson-Frey 2004). By using fNIRS, the production of ESA tools not only
adds to the limited knowledge on the neural correlates of tool use skills in action but also
potentially reveals ancient neural networks that may have been present in human
ancestors while making tools and the cognitive components that were necessary for the
skills associated with Oldowan and Acheulian stone tool manufacture.
This longitudinal study looks at neural activation differences between Oldowan
and Acheulian tool production and verbal and nonverbal ESA skill transmission over the
course of three neuroimaging sessions. With this larger data set, a clearer distinction
between these two stone tool industries emerges. The Acheulian task requires the
involvement of working memory areas in the dlPFC. Oldowan tool manufacture, on the
other hand, shows a clear transition from controlled processing guided by the dorsal
visual attention network in the first session to automatic processing by the second
session. Thus, the activation differences between these two tasks by the third session, as
highlighted in Chapter 6, are the products of differences in task complexity, not
differences in learning stages. Learning context (i.e., verbal instruction versus nonverbal
241
imitation) also leads to differential activation distributed across the frontal and temporal
cortices and affects how ESA skills are learned over time. Verbal instruction results in
fewer active clusters than nonverbal instruction; therefore, the presence of language in
the learning environment may lead to a more efficient way to process tasks related to
stone tool manufacture, at least in modern humans who are accustomed to learning with
language. These results support a cognitive explanation for hominin brain expansion in
the early Pleistocene because of the apparent dichotomy in cognitive resources used to
produce Oldowan and Acheulian type tools.
242
CHAPTER 8:
A TECHNOLOGICAL EXPLANATION FOR HOMININ BRAIN EXPANSION IN
THE EARLY PLEISTOCENE
Introduction
The results of the neuroimaging study described in Chapters 6-7 demonstrate that
different neural networks participate during the execution of Oldowan and Acheulian tool
manufacture. Whereas skills associated with Oldowan toolmaking require visual attention
and simple motor planning at an early stage of learning and procedural memory at a later
stage of learning, the skills associated with Acheulian toolmaking require complex motor
planning and the involvement of working memory to be able to hold in mind the different
subgoals related to handaxe production and flexibly respond to problems that arise during
this process. These results support earlier claims for a posterior-anterior hierarchical
gradient of abstraction associated with the increasing complexity of the transition from
Oldowan to Acheulian toolmaking, but the activation of the supplementary motor area
(SMA) and dorsolateral prefrontal cortex (dlPFC) is a novel finding that indicates that H.
erectus possessed cognition more advanced than previously assumed.
With it established that the Oldowan and Acheulian industries differ from each
other in technological and cognitive complexity, there is now further evidence that an
evolutionary transition in cognition occurred around the time of the invention of the
Acheulian industry. It may not be a coincidence that a step increase in brain size also
occurred around this time (Shultz et al. 2012) that was driven by a disproportionate
243
expansion of the prefrontal and temporal cortices (Bruner and Holloway 2010), the same
areas that are selectively activated by the Acheulian task in the current study. While the
human brain is complex and there are likely many contributing factors to its evolution
that should not be overlooked, this remarkable link between the varying brain activation
patterns during Oldowan and Acheulian stone tool manufacture and the timing of
hominin brain expansion cannot be understated. This link suggests that technology may
have played an important role in the evolution of human cognition and the expanding
brain. This chapter lays down the groundwork for a novel hypothesis, one that proposes
that selection for a working memory network that operates during the early stages of
learning to make Acheulian tools provided the impetus for brain expansion in the genus
Homo during the early Pleistocene.
The Working Memory Hypothesis for Hominin Brain Expansion
The handaxes that are characteristic of the Acheulian tool industry had a lengthy
tenure, lasting over one million years (Lycett and Gowlett 2008). They were widely
distributed across the African and Eurasian continents. After a gap of around 200,000
years, they reappeared in Middle Paleolithic (MP) France as part of the Mousterian
toolkit (Iovita and McPherron 2011). Some variation in form exists at the regional level,
but handaxes of the Lower Paleolithic (LP)/Early Stone Age (ESA) in Eurasia and Africa
share the basic general patterns of the same “bauplan” (Lycett and Gowlett 2008). These
patterns highlight the importance of handaxes as functional tools for the genus Homo
during a large extent of the Pleistocene. There is evidence to suggest that handaxes were
244
multipurpose tools, used for procuring food resources (Jones 1980; Phillipson 1997) or
potentially for crafting wooden tools for similar purposes (Binneman and Beaumont
1992; Domínguez-Rodrigo et al. 2001). Handaxes were potentially the key to accessing
many new food resources that would have been previously unattainable. If handaxes were
such an integral part of the day-to-day lives of LP/ESA hominins for gaining access to
multiple types of food items, then those individuals whose tools were consistently the
most functional would have been some of the healthiest in the population, which would
come with reproductive benefits and advantages for their offspring in the form of food
resource variety provided to them in their early years, as well as an ideal model from
which to acquire this beneficial skill to use into their adulthood. It becomes apparent that
what is being selected for is not the handaxe by itself in isolation, but also the cognition
required to learn how to replicate a model’s successful handaxe. It is possible that once
the skill was learned and mastered, it might have transitioned to become an automatic
behavior, at which point, higher-order cognition would not have played as large a role as
during the initial transmission process. There is some evidence, however, that handaxe
production is, by definition, a working memory task, in which case, the activation of
prefrontal areas would continue to be necessary even after extensive practice, and in
effect, driving selection for enhanced working memory all the more.
The results of this study reveal different neural circuits involved in the process of
Oldowan and Acheulian tool manufacture. Whereas Oldowan tool production appears to
rely on a visual attention network that is later replaced by procedural memory, Acheulian
tool production is a cognitive task that requires goal-directed strategies processed by
working memory. This shift in cognitive strategies may reflect an evolutionary transition
245
in cognition around the time that hominins began to experiment with bifacial edging prior
to the advent of the Acheulian industry. This shift would have been driven by selection
for increased working memory capacities. Working memory, as discussed in Chapter 3, is
not a unique cognitive capacity of humans. All mammals, including the great apes, are
thought to possess a homologous working memory system to humans (Carruthers 2013).
Chimpanzees appear to have a working memory capacity of two to three items based on
data collected from wild nut cracking behaviors and numerical order experiments, which
is similar to the working memory capacity of a human child between the age of 42 and 62
months (Read 2008). By the time humans reach young adulthood, their working memory
capacity increases to three to five items (Halford et al. 2007; Cowan 2010). If it is
assumed that the last common ancestor shared by chimpanzees and humans had a similar
or even reduced working memory capacity to that of extant chimpanzees, then the
increase in working memory capacity between the time of the last common ancestor and
the appearance of behaviorally modern H. sapiens must be explained in evolutionary
terms, and a trigger for the evolution of larger working memory capacities needs to be
identified.
This study supports bifacial toolmaking and the emphasis on handaxe production
during the early Pleistocene as this trigger because of the reproductive benefits that an
enhanced working memory capacity would bring for its bearer and offspring. Handaxe
production requires the ability to hold in mind multiple sub-goals, specifically striking
platform setup, flake size and shape, creating a sharp edge through alternate, bifacial
flaking, thinning and shaping the piece, all for the larger end goal of producing a
functional tool to be used in the future for a completely separate purpose. Those
246
individuals with larger working memory capacities would have been more successful
than those with average or smaller working memory capacities at holding these multiple
sub-goals in mind while working towards the end goal. They would have been more
likely to revert their attention back to these sub-goals after distractions and problem-solve
after the inevitable setbacks that occur when working with stone material. These flexible,
cognitive abilities would be reflected in the production of more functional tools than their
counterparts and an advantage in obtaining higher quality food resources for themselves
and their young.
Around 1.6-1.8 Ma, there was a step increase in hominin brain size associated
with the species H. erectus (Shultz et al. 2012). Prior to this, australopith and
paranthropith brain size was slightly larger than that of extant apes and remained steady.
A consequence of an increase in brain size would be a hike in energy intake requirements
to supply the very calorically expensive extra tissue. The increase in brain size in H.
erectus would have brought along with it an estimated resting metabolic rate that was
1.53 times higher than that of A. afarensis (Aiello and Wells 2002). Compared to the
other great apes, modern humans have a substantially greater total daily energy budget
and resting metabolic rate, which, as Pontzer and colleagues (2016) argue, may have
accommodated growth trends in hominin brain size, lifespan, and reproduction without
the expected energetic tradeoffs. A faster metabolic rate also means a higher likelihood
for energy deficits, however, which could be particularly risky for pregnant females. This
suggests that there was probably strong selection for anatomical or behavioral adaptations
that would help mitigate this risk, such as gut reduction (Aiello and Wheeler 1995),
locomotor efficiency (Pontzer et al. 2014), cooking (Carmody et al. 2011), increased
247
adipose storage, and/or a shift in diet toward more energy-dense foods like meat or tubers
(Pontzer et al. 2016).
It follows then that the costs of having a larger brain would have to be outweighed
by the commensurate benefits that resulted from having a larger brain for this trend in
encephalization to continue. Moreover, it is logical to hypothesize that the driver that led
to brain expansion in Homo already carried some of these benefits that could be
elaborated to counteract the metabolic costs of increasing brain size because this would
have facilitated its evolution. Put in another way, if a prime driver for brain expansion
does not also directly lead to an increase in the acquisition of food resources or in the
diversity of food resources to supply the extra, expensive, caloric demands of additional
brain tissue, then those who possess larger brains would be at a disadvantage to those
individuals with smaller brains at the outset. It is often assumed that mental abilities that
could indirectly lead to incremental calorie intake automatically arrive with an increase in
brain size, but this idea is largely untested.
Several scenarios and potential drivers have been proposed for the evolution of
increasing brain size in hominins, such as climate change and sociality. The expansion of
the savannah as a consequence of a cooler, dryer climate may have forced individuals to
encounter novel habitats that would have altered their diets and behavioral patterns
(Wolpoff 1980). Increased flexibility in behavior and innovation would be useful in the
event of an unpredictable climate with extreme fluctuations or abrupt habitat shifts (Potts
1998; Trauth et al. 2010). Changes in environment and predation may have also led to
larger group size, which would have imposed new cognitive demands as a result of
operating within a coordinated social group (Dunbar and Shultz 2007). Increased
248
sociality may have resulted in selection for cooperation (Dunbar 2011) and/or deceptive
strategies (Whiten and Byrne 1988). Language has also been proposed as a potential
driver for brain expansion (Shultz et al. 2012). In most of these scenarios, increasing
technological complexity would simply be a manifestation of evolving cognition and not
a direct contributor to it.
These scenarios as potential drivers for brain expansion are not convincing on
their own. Shultz and colleagues (2012) find no evidence for a relationship between
periods of accelerated change in brain size and increased climate variability over 100,000
year time blocks, based on analyses of mean sea level and Aeolian dust records, which
indicate continental aridity and variability. Furthermore, if climate were the sole factor,
why would brain expansion be a trend only in the hominin clade and not in the many
other animal species that were also exposed to the same climate variability? It is unclear
which elements of cognition are being selected for in these ecological hypotheses. For
example, Potts (1998:130) states that “the cognitive mechanisms evolved in Pleistocene
humans fostered the input, analysis, and mental representation of highly varied external
information and the output of versatile, novel response,” but neglects to identify the
cognitive mechanisms that would be selected for and elaborated upon by these versatile
environments. Sociality probably played a large part in hominin brain expansion;
however, it could only indirectly lead to the acquisition of calories necessary to
counteract the increasing metabolic costs of a larger brain.
Could selection for an enhanced working memory for learning how to make
complex, bifacial tools have been responsible for the large increase in hominin brain size
between 1.6 and 1.8 Ma? It may not be coincidental that the Acheulian industry makes its
249
first appearance in the archaeological record around 1.75 Ma (Lepre et al. 2011; Beyene
et al. 2013). There is evidence to suggest that tool use and working memory are strongly
correlated with brain size. 1) Reader and Laland (2002) conducted a large literature-based
search for incidences of complex traits associated with behavioral flexibility, such as
innovation, social learning, and tool use, among 116 primate species to test whether
incidences of complex traits occur in species with relatively larger brain volumes. They
find a clear correlation between these traits and brain size; large-brained primate species
innovate, learn from others, and use tools more frequently than small-brained primate
species. 2) At the individual level, Hecht and colleagues (2015) find with diffusion tensor
imaging (DTI) that the brains of human subjects who learned to make complex stone
tools over a two-year period of time underwent plastic structural remodeling of the white
matter tracts of the lateral frontoparietal cortex. Both white and gray matter expanded in
these regions. They argue that similar structural changes would have occurred in smallerbrained, habitually toolmaking hominins. By means of Baldwinian evolution, genetic
variants that enhanced the reliability and efficiency of this plastic remodeling response
would have received positive selection pressure. 3) Posthuma and others (2003) carried
out a large twin study that tested whether white matter volume, gray matter volume, and
total cerebral volume are related to scores on an IQ test that evaluates verbal
comprehension and working memory. The working memory dimension of the IQ test
consistently and significantly correlates to all three measures of brain volume, while
verbal comprehension does not correlate significantly with any of the three brain
volumes. The results of these studies demonstrate a genetic relationship between working
memory capacity and brain size that supports the hypothesis that complex toolmaking
250
was a prime driver for expansion of the hominin brain because it led to the selection for
increased working memory capacity. Language, on the other hand, does not appear to be
genetically correlated with brain size and thus loses its strength as a potential driver for
brain growth.
The working memory hypothesis described here supports technological and
ecological explanations for the evolution of the human brain and intelligence. “Technical
intelligence” hypotheses argue that technology or technical skills drove primate brain
evolution. For example, Passingham (1982) suggests the formation of a feedback loop
between technology and hominin brain size in that an increase in relative brain size
would provide an increase in mental capacity and allow for technological advance. New
inventions would generate new selection pressures for intelligence because of the novel
problems and situations presented in the artificial environments created by tool-wielding
hominins. In turn, these new selection pressures would lead to a further increase in brain
size. The working memory hypothesis operates under the context of skill transmission
related to tool manufacture and potentially other tool-related behaviors and therefore is a
technological explanation for the evolution of hominin brain size and intelligence. One
drawback to Passingham’s technical intelligence hypothesis is that it assumes that an
initial increase in brain size is necessary to give rise to a higher mental capacity that
would allow for the invention of new tools and an advancement of the artificial
environment, but what then caused this initial expansion in brain size if not technology?
The working memory hypothesis solves this issue because complex working memory
was already present in the last common ancestor between humans and chimpanzees and
therefore provided the cognitive and structural framework and the individual variation
251
necessary for selective pressures to act upon working memory capacity that would
eventually lead to increased brain size.
Technological and ecological explanations for hominin brain evolution need not
be mutually exclusive, as the working memory hypothesis also supports Parker and
Gibson’s (1979; Parker 2015) extractive foraging hypothesis. The extractive foraging
hypothesis proposes that brain enlargement is the result of the complex of cognitive
abilities in great apes that arose for the purpose of exploiting high-energy foods that
would normally be unavailable without the assistance of tools by intelligent tool use and
social transmission of tool use skills. This hypothesis can be extended to hominin
cognitive evolution by arguing that hominins expanded upon these abilities by using a
variety of tools on an increasing number of resources. The results of the current study
identify the potential cognitive mechanism—working memory—that received positive
selective pressure because of the elaboration of extractive foraging behaviors in early
Homo that required increasingly complex tools and social transmission. This adaptation
likely played a large role in the expansion of the neocortex 1.6-1.8 Ma. Moreover, the
dietary shift to more energy-dense foods, which was made possible by increasing
technological complexity, also may have facilitated the evolution of a greater resting
metabolic rate. In turn, there was no longer the need for an energy tradeoff between brain
size, longevity, and reproduction, as Pontzer and colleagues (2016) suggest, which could
explain why humans have the largest brains, longest lifespans, and highest reproductive
rates of all the great apes.
252
Summary
Because of the large distribution of Acheulian handaxes across Africa and Eurasia
and their long tenure of use, it is probable that they were one of the most important tools
for gaining access to valuable food resources that might otherwise have been unavailable
without the assistance of advanced tools. Handaxe production recruits a visuospatial
working memory network that is centered in the dlPFC and draws upon connections to
temporal areas for auditory feedback. Those individuals who possessed larger working
memory capacities would have been more successful at learning the crucial skills
associated with handaxe production and thus may have experienced increased
reproductive success over their counterparts because of easier access to calorically dense
and diverse food resources. Tool use and working memory show a positive relationship
with brain size. Thus, selection for enhanced working memory may have led to the
increase in brain size, particularly in the prefrontal and temporal cortices, that occurred
near the beginning of the Pleistocene, which set Homo on the path to becoming human.
253
CHAPTER 9:
DISCUSSION AND CONCLUSIONS
Introduction
The largest goal of this dissertation is to contribute to the currently limited
knowledge on the evolution of complex cognition and language during the course of
human prehistory. These features are perhaps the most defining traits of H. sapiens.
Determining under what circumstances human cognition and language originated and the
timing of these events are critical to understanding humanity’s place in the natural world.
If it can be demonstrated that these features have an ancient evolutionary past, then this
could change how behavioral modernity is defined.
To
address
this
goal,
this
dissertation
presented
the
results
of
a
neuroarchaeological study, which is entrenched in the methods and theories of
evolutionary cognitive archaeology and therefore rests on several key assumptions.
Firstly, prehistoric minds structured prehistoric action. Products of those actions that
preserve over time allow archaeologists access, albeit limited, to those minds. Secondly,
archaeologists can expand upon the number of inferences they are able to make about
past cognition by combining the methods of cognitive neuroscience and experimental
archaeology to conduct experimental, middle range research that identifies the cognitive
functions that are needed to carry out the operational sequence that must occur in order to
produce certain artifact types. Finally, the rule of minimum competence assumes that if
modern human models produce the same types of tools by using the same operational
254
sequence as extinct hominins, then these earlier hominins must have possessed at least
the minimum cognitive operations that modern human models display that are required to
complete the same task. With this theoretical foundation, the current study set out to meet
four objectives: 1) to determine the neural correlates for Oldowan and Acheulian lithic
reduction; 2) to investigate the learning networks involved during Early Stone Age (ESA)
skill acquisition; 3) to explore the influence of spoken language on complex motor skill
acquisition; and 4) to test the Technological Origin for Language Hypothesis.
This chapter summarizes how successful the current study was at addressing each
of these objectives and how the results contribute to the larger goal of determining how
and when human language and complex cognition evolved. The hypotheses for the first
two objectives are restated in the first section, which focuses on the cognitive operations
of ESA tool knapping. These results have significant implications for the expansion of
the neocortex in genus Homo during the early Pleistocene. Thus, this first section also
summarizes the evidence for the Working Memory Hypothesis for Hominin Brain
Expansion. The second section focuses on the evidence that this study provides for how
ESA skills were transmitted and how human language evolved, and the hypotheses that
correspond to the third and fourth objectives are tested.
255
Findings related to the evolution of hominin cognition
The neural correlates of Oldowan and Acheulian lithic reduction
The apparent differences in technological complexity between the Oldowan and
Acheulian industries beg the question of whether the appearance of handaxes in the
archaeological record is indicative of an increase in cognitive abilities in H. erectus
relative to earlier hominin species. There are certainly differences between the form of
the end products of Oldowan and Acheulian tools. Perhaps more importantly, different
strategies were employed during the process of lithic reduction in order to reach these
different end forms. For example, whereas Oldowan toolmaking requires the
identification and selection of ideal striking platforms, which are often created after each
flake is struck, Acheulian toolmaking involves alternate flaking, which creates sharp
edges on the core tool, and the careful setup of striking platforms that allow for control
over the shape and size of flakes that are intended to thin the piece or to produce usable
flake tools. These differences are emphasized to an even greater extent in the more
refined late Acheulian handaxes. Stout and coworkers (Stout et al. 2008; Stout et al.
2011) demonstrate potentially cognitive differences between Oldowan and late Acheulian
toolmaking processes with PET and functional magnetic resonance imaging (fMRI)
techniques, such that the Acheulian task elicits stronger activation in a more anterior area
of the right inferior frontal gyrus (IFG) than does the Oldowan task. This difference in
cortical activation may be indicative of cognitive evolution occurring between 2.6 Ma
and 0.5 Ma. Because the neuroimaging techniques Stout’s team used did not allow for
256
scans to be taken during the act of knapping, it is possible that these results may not
provide the full story for the real-time participation of areas of the brain during these
tasks. The current study therefore sought to replicate these results with functional nearinfrared spectroscopy (fNIRS) by testing the following hypothesis: If the manufacture of
complex, bifacial stone tools similar to the early Acheulian stone tool industry requires
advanced cognition relative to the manufacture of expedient flakes similar to the
Oldowan stone tool industry, then higher order association structures of the brain,
especially those in the prefrontal cortex, should be significantly active only during
Acheulian tool production.
The results reported in Chapters 6-7 support this hypothesis. A complex motor
planning network involving the left supplementary motor area (SMA) and superior
temporal gyrus (STG) comes online only during the Acheulian task in the third
neuroimaging session. The aggregate data from all three sessions reveal a left IFG cluster
that overlaps with the dorsolateral prefrontal cortex (dlPFC) portion of a visual working
memory (VWM) network (Wijeakumar et al. 2015). Multiple lines of evidence point to
the dlPFC as being the most important structure for working memory (Curtis and
D’Esposito 2003). It also plays a pivotal role in planning. During what is known as the
Tower of London task, the left dlPFC becomes more involved during problems with an
ambiguous goal hierarchy (Kaller et al. 2011). Goal hierarchy refers to the degree to
which the sequence of actions that must be executed to reach the final goal state can be
deduced. An unambiguous goal hierarchy provides an explicit sequence of subgoal
actions, which can easily be assumed at any point in the sequence and does not require
holding in mind the beginning state of the problem. On the other hand, an ambiguous
257
goal hierarchy requires one to infer the sequence of subgoals by integrating visuospatial
information across all subgoal states (Kaller et al. 2011). This term, which is used to refer
to problems confronted in the Tower of London task, is relevant to stone knapping as
well.
The fact that the Acheulian task activates left dlPFC suggests that this task has an
ambiguous goal hierarchy. Handaxe production relies on constant monitoring of the
intermediate steps that must be deduced before one can reach the end goal state(s). For
example, when a knapper comes up against a square edge, she must search for an angle of
less than 90°. She must decide on the appropriate amount of force needed to remove
some of the material of the square edge and whether the hammer stone she currently
holds is too large or too small to remove the current flake she has in mind. At the same
time, the knapper must also consider the quality of the material, the presence of any faults
that might lead to the risk of snapping the piece if she strikes this particular spot, whether
it would be safer to tackle the square edge from another direction, and how she will
remove the square edge if she errs in her strike and obliterates the platform by producing
a step fracture. All of this must also be considered in relation to the flakes she removed
prior to the planned flake because they determine the shape of the current platform she is
considering, which she could also decide to alter by removing small flakes around it, thus
giving her more control over the direction and shape of the intended flake. The knapper
also must keep in mind the subgoals she will need to attend to after removing the square
edge, such as thinning the piece, and the final goal of producing a symmetrical, teardropshaped tool with sharp edges that can be used for a certain purpose. She may be
simultaneously entertaining another end goal to produce appropriate flake tools for a
258
certain purpose as she crafts the handaxe, which adds another level to the ambiguity of
the goal hierarchy. This suggests that handaxe production with 8 h or less of training is a
cognitively demanding task that requires multiple representations to be held in active
attention simultaneously. This does not appear to be the case for simple flake production,
however.
The lack of dlPFC involvement during Oldowan simple flake production suggests
that this task does not require intermediate steps to reach the end goal and that the goal
hierarchy is unambiguous. For example, assuming the final goal state is to remove a
sharp flake suitable for a specific task, a visual search for an angle of less than 90° and
the proper determination of force can very quickly lead to the removal of an appropriate
flake and the end goal state reached. If additional flakes are required, the previous flake
removal serves as an explicit guide for the next flake removal, thus eliminating any
ambiguity in the sequence of actions to complete the task. The simplicity of this task is
reflected in the cortical regions that are selectively more active during the Oldowan task,
which include clusters in primary sensorimotor and premotor areas. Thus, it does not
appear that working memory, future planning, and executive functions in general play a
large role during simple flake production to the same extent as they do during early
Acheulian handaxe manufacture.
These results support the hypothesis in that they reveal the involvement of higherorder association structures in the prefrontal cortex during early Acheulian tool
production, but they do not completely replicate the findings of Stout and coworkers
(Stout et al. 2008; Stout et al. 2011). Previous studies emphasize the increased
involvement of the right pars triangularis portion of the IFG during Acheulian
259
toolmaking. In the current study, however, the main effect for Task with all three
neuroimaging sessions included does not demonstrate increased oxygenated hemoglobin
(oxy-Hb) levels in right pars triangularis during the Acheulian task relative to the
Oldowan task. In fact, separate active clusters are associated with Oldowan and
Acheulian tool manufacture in the pars triangularis of the left hemisphere rather than the
right hemisphere. The Oldowan cluster was not reported in Chapter 7, however, because
neither of the knapping tasks elicited significantly higher activation than the motor
baseline in this cluster (Appendix B). Stout and colleagues emphasize the importance of a
right hemisphere network for stone toolmaking (Stout et al. 2008; Stout and Chaminade
2012; Stout et al. 2015), even though this stands in contrast to the left lateralization of
everyday tool use (Johnson-Frey et al. 2005).
There may be several reasons for these somewhat contradictory results. Stout and
colleagues (Stout et al. 2008; Stout et al. 2011) observe geochronologically late
Acheulian handaxe production, while the current study observes geochronologically early
Acheulian handaxe production. If the technological complexity of these two tasks differs
from one another, then one should expect to find differences in the involvement of
cortical areas. Because the late Acheulian is thought to be more technologically complex
than the early Acheulian, as reflected by its more sophisticated handaxes, one should
expect then for late Acheulian toolmaking to employ prefrontal areas like dlPFC to a
larger extent than early Acheulian toolmaking, but this does not appear to be the case.
As has been mentioned, these previous studies may not capture Acheulian dlPFC
activation because of a type II error due to low sample size. The study that looks at the
effects of observational understanding of Acheulian toolmaking on neural activation has
260
a larger sample size than previous studies (n = 26; Stout et al. 2011), but this leads into
the next potential reason that the results of the current study differ from the results of
previous studies. The cognitive strategies used to observe a goal-directed action likely
differ to a certain extent from the cognitive strategies used to physically carry out a goaldirected action. Stout and colleagues find that the results of Acheulian observation
corroborate the results of Acheulian production; however, the temporal resolution of
PET, the neuroimaging technique they employ in their motor execution study, is very low
in comparison to fMRI and optical imaging and thus does not capture the real-time
changes that occur in the brain during the course of carrying out a task like fNIRS does.
Finally, disparate experience levels might also explain these differences. The
participants in Stout and colleagues’ (2008) Acheulian study are expert knappers with
more than ten years of experience, while the participants in the current study are novices
with only a few hours of training. It is possible that the production of a late Acheulian
handaxe can be accomplished with only procedural cognition after years of experience,
whereas, if they had scanned participants at an earlier point in training while making late
Acheulian handaxes, they may have witnessed activity in more prefrontal areas. This
latter explanation will be explored at length later.
Knapping activation and the mirror neuron system
As was discussed in Chapter 4, Arbib (2005, 2012) proposes a scenario based on
the Mirror System Hypothesis (MSH) that explains how multiple stages of evolution
involving the mirror neuron system led to the emergence of complex imitation, which
261
allowed hominins to carry out actions with multiple subgoals, and thenceforth language.
These stages included a primitive stage for simple goal-directed actions like grasping that
evolved prior to 25 million years ago (Ma), followed by a simple imitation stage prior to
the split between humans and non-human great apes, and finally a complex imitation
stage after this split. Arbib (2011) argues that the transmission of skills associated with
the Oldowan industry and the range of Oldowan tool types are similar to the abilities
observed in chimpanzees. This would indicate that the mirror system of the hominins
who lived prior to 1.75 Ma remained in the hominid (great ape) simple imitation stage.
The advent of the Acheulian industry, with its increased depth of action hierarchy, on the
other hand, reflects the transition to the next evolutionary stage for the mirror system, that
which allowed for complex imitation.
It is difficult to test Arbib’s (2011) hypothesis without knowing whether the
anatomy of the mirror system between human and non-human apes differs. For example,
if the human mirror system consists of additional brain areas compared to the chimpanzee
mirror system, then it could be argued that the areas involved in the chimpanzee mirror
system represent a more primitive state, while the addition of novel areas that are only
seen in humans could be argued to represent a derived mirror system. Then it would be
possible to observe whether Oldowan and Acheulian toolmaking differentially activate
these areas and potentially state that the manufacture of these tool types provides
evidence for Arbib’s proposed mirror system evolutionary stages. It is currently known
that a core mirror system exists in macaques, chimpanzees, and humans that consists of
the inferior parietal lobule (IPL) and ventral premotor cortex. Neurons with mirror
qualities were discovered in additional areas in humans, including the SMA,
262
parahippocampal gyrus, hippocampus, and entorhinal cortex (Mukamel et al. 2010), but
whether these areas form part of the mirror system in other primates remains to be
investigated.
The results of this study demonstrate that areas known to contain mirror neurons
are indeed differentially activated depending on whether the task involves Oldowan or
Acheulian toolmaking. The Oldowan task tends to recruit bilateral supramarginal gyrus
(SMG; see Appendix B), while the Acheulian task employs bilateral ventral precentral
gyrus (PrG) and left SMA. These results may lend some support to Arbib’s (2011) MSH,
which assumes that the Acheulian industry ushered in the next evolutionary stage of the
mirror system, but more research is needed to confirm that the activation of ventral PrG
and SMA is because of task complexity requiring more complex imitation, as opposed to
these areas becoming active in response to working memory.
ESA learning networks
Archaeologists often overlook the presence of children in the archaeological
record, especially when it comes to the interpretation of stone tools. In fact, Shea
(2006:212) even goes so far as to ask facetiously, “Were there children in Early
Paleolithic times? At first glance, this seems a stupid question…. Yet in archaeological
models of Paleolithic stone tool variability and assemblage formation processes, children
might as well be invisible.” Children obviously existed and more than likely contributed
to the archaeological record during the ESA, but as Shea notes, archaeologists “have not
looked hard enough or in the right way at the lithic record” to identify children’s
263
activities. Human children in hunter-gatherer societies mimic adult activities through play
at a young age (Park 2006) and begin to learn how to make and repair tools more
seriously later in childhood (Briggs 1970; Keith 2006). Similarly, juvenile chimpanzees
have been observed to learn to crack nuts by watching other chimpanzees (InoueNakamura and Matsuzawa 1997). It should thus be assumed that skill acquisition
associated with stone tool production began during childhood in early Homo as well. As
has already been noted, Acheulian handaxe production relies on a working memory
network in the modern human adults who participated in the current study. Because the
functional neuroanatomy of working memory tasks is demonstrably similar between
adults and children (Klingberg et al. 2002), then it should be assumed that early Homo
juveniles also possessed a similar working memory system to that of the adults and were
therefore capable of acquiring the skills they needed to successfully make a handaxe as
they developed. This also means that the lithic debris of novice adults may be informative
for interpreting children’s activities during the ESA.
Similar to the lack of acknowledgment of children in the ESA archaeological
record, evolutionary cognitive archaeologists also tend to overlook the effect of a
toolmaker’s stage of learning on the cognitive strategies they employ to make a tool.
They also tend to assume that the process of making a certain type of stone tool always
maps neatly onto the same cognitive mechanisms. For example, Coolidge and Wynn
(2005) argue that the Levallois technique provides evidence that Neandertals only
possessed a “long-term working memory,” a type of domain-specific procedural
cognition that is fast and reliable but narrowly focused, while they lacked the flexible,
“enhanced working memory” possessed by modern humans. Even though Wynn and
264
Coolidge admit that the level of expertise exemplified by MP tools would not have been
easily acquired, possibly taking ten years or more to master, there is little consideration
for the initial cognitive strategies required to learn such a difficult task, which, by their
own definition, could not have been long-term working memory.
Most motor learning studies find that the neural networks that are involved while
learning to perform a complex motor activity undergo qualitative or quantitative changes
over time (Kelly and Garavan 2005). Oftentimes, prefrontal areas are recruited when first
learning a task but reduce in strength or drop out entirely after some practice. This could
explain why a working memory network is not observed among the subjects who
received 4 h of training to make simple flakes nor among the expert knappers making
handaxes who participated in Stout and coworkers’ studies (Stout and Chaminade 2007;
Stout et al. 2008). For these reasons, it was important for the current study to test the
following hypothesis: If executive functions, such as planning, problem solving, and
attention play the most predominant role at an early stage of learning to make stone tools,
then anterior prefrontal cortex activation should drop out after some practice.
Overall, learning to make Oldowan and Acheulian technologies leads to a
reorganization of employed neural networks. After about an hour of introduction to stone
knapping, the major area that comes online during the Oldowan task is the FEF in both
hemispheres, which controls eye movements and focuses visual attention (Zhou and
Desimone 2011). This indicates that at this early stage of learning, visual search for
distinguishing features of a target is the principal strategy that guides attention and leads
to planned movements. During the first neuroimaging session, many of the participants
were observed to spend a large portion of their time visually inspecting their cores,
265
turning them over repeatedly before finally attempting to remove a flake; they
presumably were searching for striking platforms with an external angle of less than 90°.
Before the third neuroimaging session, however, this area drops out and is replaced by
primary sensorimotor areas, a result that indicates a transition from cognitive control to
automatic processing. Thus, the Oldowan task results support the hypothesis. The fact
that such a drastic change in neural network recruitment can be observed in such a
limited time span is important to note because it highlights the need for caution when
making conclusions about extinct hominin cognition based on functional brain data from
a single point in time.
The Acheulian task, on the other hand, does not show a trend towards automatic
processing. In fact, the cognitive control area that drops out after several hours of training
during the Oldowan task actually increases in strength over time during the Acheulian
task. This result provides even further evidence that Oldowan and Acheulian toolmaking
form separate processes in the brain, even when first being learned. It is possible that
with extensive training beyond the 8 h provided in this study, the Acheulian task might
eventually transition to procedural cognition, but because handaxe production appears to
rely on a working memory network that only strengthens with more experience, this
suggests that making a handaxe is a working memory task. And as a working memory
task, it should then be expected that cognitive control areas would continue to be
recruited, even after extensive practice. If this is the case, then the expert knappers in
Stout and colleagues’ (2008) study should also have active working memory areas as
they make late Acheulian handaxes, but this is not the case. The lack of dlPFC
266
involvement that they report may have less to do with knapper experience and more to do
with the limitations of the experiment design as described in the prior section.
Implications for the evolution of the human brain and cognition
If there was a transition from ape-like to more human-like cognition between the
advent of the Oldowan and Acheulian stone tool industries, as the results of the current
study indicate, then one should also expect to find evidence for structural changes to the
hominin brain that reflect this new emphasis on certain cognitive features. There was a
step increase in the brain size of Homo occurring around 1.6-1.8 Ma, around the same
time that handaxes began to appear in the archaeological record (Shultz et al. 2012). Also
around this time, Homo evolved a more human-like configuration of Broca’s area and
cerebral asymmetries (Holloway et al. 2009). Was the change in technology somehow
responsible for the evolution of more complex cognition and larger brain size, or was it
simply a byproduct of this change that occurred for other reasons?
The Working Memory Hypothesis for Hominin Brain Expansion, introduced in
Chapter 8 and summarized briefly here, proposes that because of a strong reliance on
technology in hominins for procuring foodstuffs, technology was largely responsible for
the transition to complex cognition and encephalization. Early Homo individuals who
were able to make increasingly, hierarchically-complex tools were selected for because of
the larger diversity of food items that this afforded them and their offspring. Handaxes
were multipurpose tools. Although it is rare to find preserved organic residue or use wear
polish on artifacts originating from ESA contexts, animal fat and plant residues have been
267
found on some handaxes from Acheulian sites in the Levant and Africa (DomínguezRodrigo et al. 2001; Solodenko and colleagues 2015), while use wear on two handaxes
from Wonderwerk Cave in South Africa establishes that they were also used for
woodworking (Binneman and Beaumont 1992). This means that handaxes may have been
used to gain access to food items that were otherwise unavailable to them. They may
have been used for butchery, exhuming or cutting vegetal materials, and possibly shaping
wooden tools, which may also have been used for acquiring difficult-to-access food
items.
Handaxe manufacture is a task that requires working memory. Representations
that gain access to working memory are, by definition, conscious because items
manipulated in working memory require active attention (Carruthers 2013). Thus, claims
that handaxes were the result of the “mindless” flaking of flake unit chains (Moore 2010)
are unfounded. It is more likely that, similar to modern Homo, early Homo toolmakers, at
least in the early stages of learning to make bifacial tools, were conscious of their goal to
reduce a blank into a predetermined shape and the requisite subgoals leading up to this
end stage. To hold in mind the multiple subgoals required to make a bifacial tool,
especially if that tool was then used to craft an additional wooden tool, would have been
too taxing for early hominins that possessed an ape-like working memory capacity of two
or three items. Working memory capacity, however, varies across a population; therefore,
those individuals with a larger working memory capacity would have been more
successful at storing these suboals while manipulating the stone. They would have been
able to consistently make the most functional tools and would have used these tools to
gain access to a wider diversity of food items and to consume more calories than their
268
counterparts with smaller working memory capacities. In this case, the cognitive
mechanism that received positive selection was working memory. Multiple lines of
evidence indicate that tool use and working memory positively correlate with gray and
white matter volumes, which contribute to brain size in human and non-human primates
(Reader and Laland 2002; Posthuma et al. 2003; Hecht et al. 2015). With enough
selection for tool use and working memory over time, this could have led to the
expansion of the hominin brain 1.6-1.8 Ma.
The long tenure of handaxes during the Pleistocene highlights the extent to which
H. erectus relied on them for food acquisition and perhaps for other activities as well.
The Acheulian industry is often mistakenly said to have remained stagnant over a period
of more than one million years. Handaxes did indeed change through time (Vaughan
2001), but these changes are often overlooked because they occurred so gradually.
Similarly, there is a gradual and continuous increase in brain size from the early to midPleistocene (Shultz et al. 2012). This trend complements the evolution of technology
occurring during this time as well, which supports the scenario that working memory
capacity continued to receive positive selection after an initial punctuated event.
With all this in mind, some speculation can be made regarding the requisite
cognition for stone tool industries that came before and after the Oldowan and Acheulian
industries in time. The 3.3 million-year-old flakes and cores found at the Lomekwi site in
Kenya that are being dubbed the Lomekwian industry are larger than most Oldowan
products and would have been extremely difficult to knap by freehand (Harmand et al.
2015). Lomekwian flakes were likely produced using a block-on-block technique where
the edge of the core was struck against a passive anvil, though freehand knapping cannot
269
be ruled out at this point completely. This technique would not have required the fine
motor movements of the hands associated with Oldowan freehand knapping but would
have required some visuomotor skill to aim and apply the correct amount of force to be
effective. It is unclear the extent of motor planning that would be required for the blockon-block method because of the dearth of experiments that have attempted to
systematically study this flaking technique, but it is assumed that the use of a passive
hammer is more of a haphazard method of producing flakes than freehand knapping. It is
therefore probably safe to assume that the cognition associated with Lomekwian
toolmaking was similar to that associated with the Oldowan industry, with some degree
of premotor cortex involvement for motor planning, but the involvement of structures for
coordinating movements and reward centers of the brain may have been sufficient for
carrying out this activity. These ideas remain to be tested, however.
For the time period postdating the Oldowan/Acheulian industries, much debate
surrounds the cognition associated with the Middle Paleolithic (MP) Mousterian stone
tool industry. The Mousterian industry, which is arguably even more complex than the
Acheulian industry to produce, is characterized by handaxes, the hafting of flake tools,
and the Levallois method, a prepared core technique. Although no one has yet reported
the neural correlates of Mousterian toolmaking, it can be assumed, based on the results
from the Acheulian task, that working memory and potentially other executive functions
would have been necessary to produce Mousterian tools, at least in the early stages of
learning. Additionally, because archaeological experiments have demonstrated the
lengthy training time required to master this skill (Eren et al. 2011; Stout et al. 2015), it
270
can be assumed that there would have been increased selection for working memory in
Neandertals.
Coolidge and Wynn introduced their Enhanced Working Memory hypothesis,
which argues that enhanced working memory is not apparent in the archaeological record
until the appearance of facilities, or permanent, immobile constructions, such as traps and
snares, around the end of the Pleistocene. In this model, these types of artifacts are the
manifestation of a recent genetic mutation that occurred less than 100 thousand years ago
(Ka) and led to the rapid spread of increased working memory capacities in modern
human populations (Coolidge and Wynn 2005; Wynn and Coolidge 2011). They argue
that most stone toolmaking, including Mousterian and Acheulian tools, relied upon
procedural cognition and long-term memory (Wynn and Coolidge 2011). The evidence
from the current study weakens the timing aspect of Coolidge and Wynn’s Enhanced
Working Memory hypothesis because it suggests not only an earlier origin for more
human-like cognition, but also a gradual, rather than punctuated, evolution of this feature
after the initial step increase in the early Pleistocene. It is possible that with substantial
practice, any of the stone tool industries discussed so far could eventually be produced
with procedural memory as a result of expertise, similar to what may be occurring in the
Oldowan task of the current experiment, but what Coolidge and Wynn neglect to take
into account is the cognition required for the transmission and initial learning curve
associated with complex stone tool industries.
If there was an initial directional selection towards larger working memory
capacity during the early Pleistocene because of technology, and if this is considered as
the first relatively punctuated event in the evolution of human cognition, then it is
271
certainly possible that a second punctuated event related to technology or culture could
have occurred more recently, perhaps after the appearance of modern H. sapiens as
Coolidge and Wynn propose. Many archaeologists agree, however, that the presence of
composite tools prior to the MP-Upper Paleolithic (UP) transition indicates that this shift
in cognition likely occurred earlier. A way to test this hypothesis is discussed below.
Suggestions for future research in Stone Age cognition
As this dissertation has demonstrated, neuroarchaeology has the ability to
differentiate between the cognitive requirements associated with stone tool industries of
varying technological complexity and age. The insights that are drawn from
neuroarchaeological research are probably the closest scientists will ever come to
observing past hominin cognition in action. Because it is still such a novel field, there is
much yet to be discovered. Presented here are some directions for future research related
to questions on the evolution of hominin cognition.
The marriage of neuroscience and experimental archaeology has the potential to
change common assumptions that archaeologists make about past human skill and
cognition based on stone artifact features. A large number of flakes, cores, and other
waste pieces were produced during this experiment, and each of these pieces possesses
numerous measurable features that can be correlated with the brain activation patterns
that were occurring during the lithic reduction process. This analysis would highlight
which flake and core features are the most informative about different aspects of
272
cognition and skill so that more informed interpretations can be made about the past
humans who produced these features on stone tools found in archaeological assemblages.
The analyses used in this study are limited in that they focus only on the
differences between Oldowan and Acheulian tool production; therefore, any overlapping
areas between these two tasks that are important to knapping in general are overlooked.
Additional analyses that directly contrast each of these tasks with the motor baseline at
different points in training will produce individual maps for the full complement of
neural correlates for Oldowan and Acheulian toolmaking, thus providing the most
complete account currently available for Oldowan and Acheulian neural activity at an
early stage of learning. Additionally, to test the hypothesis put forth here—that handaxe
production is a working memory task that does not transition to automatic processing—a
large sample of expert knappers should be imaged with fNIRS while they make early
Acheulian handaxes.
There is still much to be learned about children and their activities involving
technology in the Paleolithic. It is likely that the skills needed to make stone tools were
learned later in childhood once children had the motor coordination and strength to
remove flakes, while younger children may have only played at making stone tools.
There have been few accounts in the literature of the behaviors and products associated
with modern-day children knapping, and these are usually anecdotal (Hogberg 2002;
Ferguson 2003). Children’s knapping activities probably left visible traces in the
archaeological record, but until systematic, experimental research is carried out among
modern human children, there is only the lithic debris of adult novices to act as an analog
273
for children, which, while better than no evidence at all, may not provide the most
accurate model for archaeological traces of children’s activities.
At this point in time, no one has attempted to map the neural correlates of realtime Lomekwian and Mousterian tool production. Once these studies are carried out, a
more complete model of hominin cognitive evolution, one that spans a large portion of
human prehistory, will be realized. There has been one critique of this type of research
that will continue to plague neuroarchaeology until it is properly addressed: the
functional neuroanatomy of adult modern humans may not be the best analog for extinct
hominin functional neuroanatomy because it is unknown whether culture, language,
development, and evolution have altered the functions of the human brain to the point
that it no longer resembles the organization of prehistoric brains. One theme in this
dissertation has been the importance of comparative research. Pioneering studies already
demonstrate that it is possible to record the functional brain activation patterns of wake
primates (Vanduffel et al. 2002; Orban et al. 2004; Gil-da-Costa et al. 2006; Orban et al.
2006; Taglialatela et al. 2008; Kojima et al. 2009; Peeters et al. 2009; Taglialatela et al.
2009; Hecht et al. 2013), even with fNIRS (Wakita et al. 2010). The discovery of the
Lomekwian industry has introduced a unique opportunity because the methods that were
employed by Lomekwian toolmakers to fracture rocks are within the behavioral
repertoire of modern, non-human apes. If fNIRS can be used to record the neural
activation of both chimpanzees and humans as they make Lomekwian tools, this would
allow for a direct comparison of the functional neuroanatomy of ESA tool manufacture in
chimpanzees and humans. If it could be shown that the task is processed similarly by both
274
species, this would support the findings of other studies, such as this one, that rely upon
modern humans as analogs for extinct hominins.
Upper Paleolithic (UP) behaviors, such as “reliable tool” and facility construction
and therianthropic artwork may have required increased working memory involvement
compared to earlier technological behaviors, according to Coolidge and Wynn. This
hypothesis could be tested by measuring the strength of neural activation in working
memory areas during a uniquely UP behavior that they argue can only be carried out by
using enhanced working memory, such as therianthropic artwork. Therianthropic artwork
involves the combination of human and animal features into one form. This could then be
compared to the neural activity associated with Acheulian, Mousterian, and composite
tool manufacture8. If their hypothesis is correct, then one should expect similar patterns
of prefrontal activation during these three toolmaking behaviors, while the UP behavior
should elicit either stronger activation in these same areas or activation in additional
working memory areas to be able to support the occurrence of a punctuated evolution of
working memory in H. sapiens.
The Acheulian industry did not appear concurrently in different geographic
regions. If the Working Memory Hypothesis for Hominin Brain Expansion is valid, then
there should be variation in brain size among the earliest populations of H. erectus that
relates to the presence of bifacial tools. For example, the artifact assemblage at the 1.7
8
At first glance, it may seem disjunctive to make comparisons between technology on the one hand and a
symbolically-laden domain such as artwork on the other hand, given their inherent differences. According
to Coolidge and Wynn’s model, however, stone tool traditions, even those in the UP, do not require an
enhanced working memory to produce; therefore, the inclusion of an UP stone tool industry instead of
therianthropic artwork would not technically test this hypothesis. Because they allow that reliable tools
require enhanced working memory, something like harpoon construction would be a more appropriate
comparison with earlier stone tool industries. This presents a challenge for an experimental design that
enlists a functional neuroimaging technique, however, because reliable tools may take hours or days to
complete.
275
million-year-old site of Dmanisi, Republic of Georgia, consists of cores, choppers, and
flake tools but no handaxes (Mgeladze et al. 2011). The Homo fossils at the site have
cranial capacities that fall below 800 cm3 (Gabunia et al. 2000). On the other hand,
Acheulian handaxes have been discovered in Bed II at Olduvai Gorge, Tanzania, which
dates to 1.65-1.7 Ma (Rightmire 1979; Diez-Martín et al. 2015). The specimen known as
O.H. 9 from Bed II of Olduvai Gorge has a large cranial capacity of 1,067 cm3 (Holloway
1975). A systematic study carried out on all the early Pleistocene sites that include stone
tools associated with fossil crania with measurable cranial capacities should reveal larger
cranial capacities where bifacial tools are present. Furthermore, the prefrontal and
temporal cortices should drive this larger cranial capacity more than other brain regions,
according to the Working Memory Hypothesis for Hominin Brain Expansion. This could
be achieved by taking endocranial measurements using geometric morphometrics
methods, similar to the work carried out by Bruner (e.g., 2008). Archaeology has a
history of falsely equating minds of less sophisticated intelligence with the production of
simple tool types (e.g., see Tylor 1894 for a discussion of Tasmanians as a modern
example of Paleolithic man), but it is clear now that modern humans and extinct
hominins of sophisticated intelligence often created simple, expedient tools. It is expected
that even large-brained H. erectus individuals who possessed the advanced working
memory capacity necessary to make handaxes may be associated only with simple flake
tools at some sites, but because an enhanced working memory system is the minimum
requirement for handaxe production, it would be surprising to find H. erectus individuals
with relatively smaller brains or more primitive prefrontal or temporal cortex
development at sites containing handaxes.
276
Findings related to ESA skill transmission and the evolution of human language
The effect of learning context on complex motor skill transmission
There has yet to be agreement on the timing of the evolution of language or
protolanguage. While some researchers have argued that hominins living as early as the
late Pliocene or early Pleistocene possessed protolinguistic abilities (e.g., Montagu 1976;
Bickerton 1990), a larger majority believes that these abilities did not develop until the
time of H. heidelbergensis or even later (e.g., Davidson and Noble 1989; Steele and
Uomini 2009). Because of this disagreement, the fact that language and pedagogy tend to
accompany one another, and non-human primates do not possess language nor do they
actively teach their young (but see Boesch 2012), it cannot automatically be assumed that
early hominins transmitted and acquired toolmaking skills via language or via teaching.
Language instruction versus imitative learning leads to differential lithic reduction
processes, as is evidenced by the debitage created during replicative Oldowan and early
Acheulian tool manufacture (Putt et al. 2014b; Morgan et al. 2015). It remains to be seen
whether these differences are also reflected in the motor behaviors that correspond to
flake removal events. Putt and colleagues (2014b) and the current study contend that
spoken language instruction is not necessary to learn how to make a handaxe in modern
humans; therefore, the successful transmission of Acheulian knapping skills between
early Homo individuals also likely did not require language.
With all of this in mind, archaeological experiments that attempt to replicate ESA
behaviors, especially those that include a neuroimaging element, should probably not rely
277
upon data gathered from linguistically instructed knappers alone. There is surprisingly
little research that investigates the effect of language instruction on how motor skills are
learned and processed in the brain. It was therefore important to explore how complex
toolmaking skills are learned, depending on whether the learning context included
linguistic or imitative information transmission. This was done by testing the following
hypothesis: If verbal instruction reorganizes the neural pattern for processing a learned
motor skill, such as Oldowan and Acheulian tool production, then different clusters of
activation should be found among novice knappers who learned via verbal or nonverbal
instruction.
The results of this study support the hypothesis. There are overall differences in
regional activation between the verbally and nonverbally instructed groups that include
classic language-processing areas as well as other motor regions. The nonverbal group
recruits the right postcentral gyrus (PoG) and IFG more than the verbal group. Learning
context also affects how learned knapping skills are processed over time, the main
difference occurring in frontal areas. Early in training, the nonverbal group enlists the
right middle frontal gyrus (MFG) and left IFG, while the verbal group recruits the left
PrG. The same IFG cluster is active late in training as well, but only in the verbal group.
This suggests that the two groups potentially enlist different cognitive strategies to learn
the same task, which explains why there are measurable differences between the two
groups in their stone tool products.
Overall, these results demonstrate that the mode by which one learns a complex
motor skill has an effect on the neural networks that are recruited to carry out the task.
When these results are compared directly with Stout and colleagues’ (2008) results, it is
278
clear that language influenced how the participants in their study processed the task as
well. Most notably, there is stronger activation in the right pars triangularis during the
Acheulian task relative to the Oldowan task after ~8 h of training, which replicates the
pattern of activation they found in this area; however, this pattern only occurs among the
verbally instructed knappers in the current study. In sum, the results of the current study
support the hypothesis that states that the mode by which ESA knapping skills are
transmitted has a differential effect on the cortical areas that are recruited while making
stone tools. This suggests that future neuroarchaeological studies should account for
language in their experimental design.
The current status of the Technological Origin for Language Hypothesis
Stout and Chaminade’s (2012) Technological Origin for Language hypothesis
expands upon Arbib’s Mirror System hypothesis by arguing that a general hierarchical
information-processing network, which originated from the primate mirror system and
was already in place to support complex technological activities like stone tool
manufacture, was exapted by language or protolanguage. This would mean that the
original functions of the neural substrates that participate in language production were for
action understanding, action planning, and hierarchical information processing in relation
to tool manufacture and use. This hypothesis also proposes that intentional pedagogical
demonstration during the transmission of stone knapping skills may have provided the
scaffold for intentional vocal communication leading to speech.
279
Multiple studies demonstrate that Broca’s area and its right hemisphere analog
also have a motor component (e.g., Heiser et al. 2003; Higuchi et al. 2009), as would be
expected if language had a motor origin, but to make the more specific claim that a
“special evolutionary relationship” exists between toolmaking and language would
require a more specific overlap in the neuroanatomical substrates between these two
behaviors (Stout and Chaminade 2012:76). Support for this hypothesis comes from the
activation of the ventral premotor cortex in both hemispheres during Oldowan and
Acheulian lithic reduction and pars triangularis in the right hemisphere during the
Acheulian task (Stout and Chaminade 2007; Stout et al. 2008). It should be noted that
while right pars triangularis does participate in some language functions, it is the areas in
the left hemisphere that handle most language functions. The technological hypothesis
would be more convincing if there were evidence for Broca’s area activation in the left
hemisphere during stone knapping.
Additionally, caution is required when interpreting these results because of the
influence that language instruction has on the neural networks that come online during
lithic reduction tasks, as has been demonstrated. These language-sensitive areas may
have been active during these experiments because language instruction was not
controlled. Because the current study did control for this variable, Stout and Chaminade’s
technological hypothesis can be tested more rigorously. If language co-opted the neural
networks already in place for tool manufacture, then it should be expected that classic
language areas would become active among novices learning to knap without verbal
instruction as well. If these areas become active only among the verbally instructed
280
knappers, this would implicate linguistic instruction as the cause, rather than original
motor functions.
To test this hypothesis, any significant clusters that overlap with areas from
Vigneau and colleagues’ (2011) language-processing meta-analysis where verbal or
nonverbal knapping activation was above the activation threshold were drawn from the
second analysis (see Chapter 7 and Appendix B). A total of four language-processing
areas were identified, including the right and left pars triangularis and two clusters in the
left STG. Interestingly, the nonverbal group has increased activation relative to the verbal
group in all four areas (Figure 26), which lends support to the Technological Origin for
Language hypothesis. This could imply that stone knapping learned by imitation alone
places high demands on all those areas that make stone tool manufacture possible,
including those areas that have taken on new functions like language-processing since
they were used exclusively for other purposes in the distant past.
281
Figure 26. Language-processing areas that are active during lithic reduction tasks.
Displayed are the right IFG from the Group main effect (Panel A), the left IFG from the
Group x Session interaction effect (Panel B), the left STG from the Task x Session
interaction effect (Panel C), and another left STG cluster from the Group x Session x
Task interaction effect (Panel D). Error bars represent 95% confidence intervals.
One caveat to keep in mind is that knapping activation in these IFG areas should
elicit significantly greater activation than general, motor behaviors if one is to argue that
the language functions in these areas are exapted specifically from functions involved in
technological behaviors. Otherwise, a more general motor origin for language cannot be
282
ruled out. Motor baseline activation is actually higher than knapping activation in the
right pars triangularis. Conversely, the left pars triangularis demonstrates a specialized
role for stone knapping. The nonverbal group’s knapping signal is significantly higher
than the motor baseline signal for both the first (Wilcoxon Z = 107; p = 0.044) and
second neuroimaging sessions (Z = 122; p = 0.005), and the verbal group’s knapping
signal is also higher (Z = 100; p = 0.023).
The purpose of the motor baseline task was to establish which regions are unique
to stone knapping and which regions become active during a general, bimanual motor
task. The motor baseline task also involved a timing element, which required the
participants to match their movements to the pace of a metronome, as well as try to keep
this pace when the metronome could no longer be heard. The large motor baseline signal
in the right pars triangularis may reflect this timing element. The language function of
the right pars triangularis involves comprehending affective prosody (Wildgruber et al.
2005), which is the rhythmic pattern of stress and intonation in language. From these
results, it appears that this area plays an important role in translating rhythm into body
movement while knapping as well. Its increased activation during stone knapping in the
current and past studies may simply reflect a tendency to knap to an internal rhythm. In
fact, when asked what she was thinking about while knapping in a concluding interview,
one subject said that she purposely tried to knap to a rhythm in her head to help focus on
the task. Two subjects reported playing songs in their head while knapping.
The activation of the left IFG and STG (i.e., Broca and Wernicke’s areas) among
novice learners in the nonverbal group while stone knapping provides a stronger
argument for the Technological Origin for Language hypothesis than the premotor areas
283
and right IFG put forward previously as evidence. Most functions related to language are
processed in the left hemisphere. The extent that functions are lateralized in the human
brain is unique when compared to other primates. Humans also happen to be the singular
species that possesses language. The lateralization of language and toolmaking in the left
IFG and STG is likely not simply coincidence. If language exapted the functions of Broca
and Wernicke’s areas that were already lateralized for skilled movements and sound
perception involved in tool use and manufacture, then language also would have been
mainly lateralized to the left hemisphere. Archaeological and paleoneurological evidence
suggest that the lateralization of functions in the hominin brain may have evolved at least
1.8 Ma (Toth 1985a; Cornford 1986; Bermúdez de Castro et al. 1988; Holloway et al.
2009). Now neuroarchaeology can provide further evidence to support this claim. So, the
functional networks of language may have essentially “piggy-backed” on the functional
networks of skilled motor processing. This does not necessarily mean that stone tool
production provided the impetus for the evolution of language per se, nor does it
necessarily mean that language existed as early as 1.75 Ma, but it does suggest that the
motor and perception framework upon which language would eventually build was likely
already in place by this point in time.
Once language or protolanguage evolved, it may have formed a co-evolutionary
relationship with technology because they both relied on similar motor networks. Those
participants who learned to knap via imitation recruited a far larger volume of cortex than
those who learned with language instruction to complete the same task. Of the 16,784
voxels where a group effect was identified (including those that overlap with interaction
effects), 13,904 (82.8%) are attributed to the nonverbal group. When neurons in the brain
284
become active, this stimulates cerebral blood flow and metabolic responses, which are
necessary to meet the increased energy requirements of these neurons and the glial cells
that support them (Bélanger et al. 2011). If such a large volume of cortex becomes active
in knappers who learned via imitation, then this may suggest that the task is a larger
energy requirement for them than it is for the knappers who learned via verbal
instruction. These results indicate that those who receive language instructions while
learning to knap stone tools process this task more efficiently than those who learn by
imitation. If the first language-wielding hominins began to incorporate language into the
transmission of complex skills like toolmaking, its potential energy-saving benefits may
have led to selection for increasingly complex language to maximize upon this feature,
thus allowing for more complex technologies to develop as well.
Language probably also facilitated joint attention of two or more individuals on
the same task and would have allowed for communication between individuals related to
the task. Essentially, this could create a situation for extended cognition, where a second
individual acts as an external store of information for the toolmaker. Hypothetically, if
two individuals each have a working memory capacity of three items and have no way to
communicate complex ideas to one another, then they each would only be capable of
producing tools that require holding in mind three subgoals at a time. If, on the other
hand, they possess the same working memory capacity but have the ability to
communicate complex ideas to one another with language, then the toolmaker could
theoretically produce a tool that requires more than three intermediate steps to reach the
end goal because once the toolmaker’s working memory capacity has been reached, she
can then turn to the other individual for externally-stored information. Over time, this
285
relationship between language and technology could have led to increasingly complex
tools, sentence structures, and ideas.
Most of the literature on language evolution that has been explored so far has
focused on Broca’s area and its probable origin as a motor-based structure, and little
attention has been paid to the original functions of Wernicke’s area before it evolved to
become the prime speech perception processor. Wernicke’s area has several languagerelated functions, including semantic processing (Chou et al. 2006) and sentence and
word generation (Lurito et al. 2000; Brown et al. 2006), as well as other more general
auditory functions, including complex sound processing (Mirz et al. 1999), sensitivity to
pitch (Patterson et al. 2002), sound intensity processing (Hart et al. 2003), and rapid
sound detection (Lehmann et al. 2007). As was noted in Chapters 6-7, there is differential
recruitment of Wernicke’s area between Oldowan and Acheulian toolmaking. It is
significantly more active during the Acheulian task but only after several hours of
training. There was no interaction with Group that might suggest higher activity in this
area due to verbal instruction. Thus, this difference in activation in Wernicke’s area must
be explained by the behavioral differences that accompany these two different tasks.
By the third neuroimaging session, most of the participants had a clear notion of
the goal of the Acheulian task and how it differed from that of the Oldowan task. To
achieve the goal of a final handaxe product, the participants needed to use an alternate
flaking technique to remove square edges, and they had to effectively thin the piece by
sending long flakes across the blank. Over time, they learned that the tone of a successful
flake removal sounds different from the tone of an unsuccessful flake removal attempt.
This often had to be determined quite rapidly as they delivered blows one after another in
286
quick succession until hearing the sound of a successful flake removal, at which point
this required the inhibition of further movement, followed by inspection of the result.
While they were probably aware of the sounds associated with flake removals for the
Oldowan task as well, the removal of a flake during this task was not as consequential to
carrying out the rest of the task like it was for the Acheulian task, nor did they need to
concern themselves to quite the same extent with the size or shape of the flake or the
direction it must run on the piece. Acheulian toolmaking thus likely taps the functions of
Wernicke’s area related to complex sound processing, sensitivity to pitch, and rapid
sound detection in order to allow the individual to make the quick decisions that relate
meaning to the sounds of her hammer stone on the blank.
What is described here is analogous to the language functions of Wernicke’s area.
An individual perceives a rapid stream of discrete sounds of varying pitches and must
assign meaning to each sound and to the overall stream of sounds in relation to the
broader context of the situation in order to decide how to respond. This sentence could be
describing Wernicke’s role in speech perception or its role in Acheulian lithic reduction.
Because of this similarity, it is possible that Acheulian tool production helped fine-tune
the ability of Wernicke’s area to rapidly discriminate between similar sounds and assign
semantic meaning, which could have laid the groundwork for speech perception. This
region becomes active in other primate species when they hear vocal signals from
conspecifics (Gil-da-Costa et al. 2006), so it does not seem a far leap to go from
discriminating between the ‘bop bop tink’ sounds of unsuccessful and successful flake
removals to determining the semantics of vocal signals.
287
Interestingly, the anatomy of the middle ear of Australopithecus and
Paranthropus consists of a derived, human-like malleus and primitive, ape-like incus and
stapes bones (Quam et al. 2013). This mosaic anatomy indicates that these genera had
decreased auditory sensitivity to midrange frequencies, which encompass the frequencies
of modern human speech (1-4 kHz). By 500 Ka, however, hominin fossils from the Sima
de los huesos site in Spain had an essentially modern middle ear anatomy with the same
auditory capacities as modern humans (Martínez et al. 2013). Together, fossil and
neuroarchaeological evidence reveal that a major shift in hominin audition and auditory
processing occurred after Homo diverged from Australopithecus and Paranthropus and
before the appearance of H. heidelbergensis. This auditory evolution would eventually
support speech perception, most likely because of selection for aural and neural anatomy
that can adequately capture the frequencies of the human voice and identify meaningful
units within the vocal signal. Flint material can produce a range of audible frequencies
between 2 and 20 kHz (Blake and Cross 2008); therefore, it would have been presumably
easy for early Homo to distinguish between the sounds produced when percussing stone,
even if these hominins were less sensitive to midrange frequencies, similar to other
primates and more primitive hominins. This would support the hypothesis that selection
for the semantic functions of Wernicke’s area occurred prior to the appearance of vocal
language, possibly in part due to making Acheulian tools.
288
Suggestions for future research related to skill transmission and language evolution
There is still much to learn about how one’s learning context affects the
transmission of ESA toolmaking skills. While it is clear that linguistic instruction and
imitation result in differential neural processing of Oldowan and Acheulian tasks and that
this leads to morphological differences in the debitage produced during these tasks, it
remains unclear which visible behaviors exactly are affected by these different neural
activation patterns that lead to measurably different flakes between the two groups. In
other words, the knapping gestures one uses may depend upon one’s learning context. By
using video coding methods similar to those used by Geribàs and coworkers (2010), it
may be possible to discern differences in the knapping gestures used by verbally and
nonverbally instructed participants. A correlation relationship between regional neural
activity and specific knapping behaviors could then be explored.
It is still unknown when in human prehistory pedagogical behaviors evolved.
Now that data exist from an experiment that involved interactive teaching of ESA
knapping skills (Putt et al. 2014b) and the current experiment that involved no interaction
between instructor and student, the data from these two experiments can be compared to
observe whether pedagogical differences are reflected in the stone tool products. If so, it
may be possible to apply these measurements to archaeological assemblages in order to
detect when pedagogical transmission became the norm. This would also signal the
appearance of theory of mind.
One question brought up by the activation of Wernicke’s area during Acheulian
lithic reduction is whether it participates in assigning meaning to the various sounds that
289
arrive in rapid succession as one percusses the core with a hammer stone, similar to its
semantic processing role in language. Because of the excellent temporal resolution of
fNIRS, it is possible to monitor the differential cortical activation associated with
successful versus unsuccessful flake removals, as well as other flaking events that occur
while stone knapping. The results of this study may help determine the role of
Wernicke’s area during Acheulian handaxe production, as well as identify other areas
that participate in the rapid decision-making that occurs after each strike.
Conclusions
Haidle (2010) points out that there has yet to be a proven causal relation between
working memory and the emergence of certain artifacts, making it very difficult to trace
the evolution of executive functions. The study presented here is the first of its kind that
unequivocally demonstrates the need for an advanced working memory system relative to
that of other primates to learn to produce early Acheulian handaxes. The Oldowan and
Acheulian stone tool industries have long been thought to represent two different levels
of technological complexity in the archaeological record, as well as possibly two
different evolutionary stages in hominin cognition. These suppositions are largely
confirmed by the results of this study. Oldowan lithic reduction recruits a dorsal visual
attention network at an early stage of learning. After some practice, this task elicits neural
activity in primary sensorimotor areas. This indicates some motor planning and
coordination of movements but little involvement of cognitive control after some
training. Early Acheulian lithic reduction, on the other hand, recruits cortical areas that
290
are involved in working memory, planning, decision-making, auditory feedback, and
carrying out complex motor sequences.
The novel finding of this research is the discovery that dlPFC and SMA become
active during early Acheulian handaxe production. Handaxe production appears to be a
working memory task that requires the maintenance of robust working memory states
while simultaneously combining multiple forms of information, including 1) the
visuospatial properties of the core tool, 2) a working memory representation of how each
action affects future actions related to the final goal of the task, and 3) an auditory
representation for judging the success of each action. This all suggests that H. erectus
possessed cognition that was not only more sophisticated than that of earlier hominins,
but was also more advanced than was previously assumed by most scientists. Selection
for enhanced working memory capacity and stability across a distributed frontotemporal
network may have been the driving force behind the expansion of frontal and temporal
neocortices that is evident in early Homo and the trend towards a modern reorganization
of the brain.
The number of engaged clusters increases when knapping skills are transmitted
nonverbally via imitation. Verbally and nonverbally delivered instructions during ESA
skill transmission result in differential activation of distributed neural networks in the
frontal, parietal, and temporal cortices, and this difference in learning context continues
to determine how the task is learned by which neural networks are recruited at different
points in training. The fact that there is little agreement on the timing of the evolution of
language and that stone knapping is processed differently in the brain depending on the
presence or absence of language in the learning context implies that caution is needed
291
when interpreting claims about prehistoric hominin cognition in studies that do not
control for language in their experimental design.
Finally, there is evidence to suggest that some features of language may owe
their origin to the motor and perception systems that are also active during skilled stone
tool production. Language and technology may have then formed a mutually reinforcing
co-evolutionary relationship after this point. If language exapted the motor and
perception systems already in place for skilled tool production, this could explain the
occurrence of several phenomena. For example, specialized language and ESA
toolmaking functions overlap with each other in Broca’s area, and these functions are
lateralized to the left hemisphere. Similarly, Wernicke’s area in the left hemisphere is
selectively active during Acheulian tool production because of its potential semantic role
in assigning meanings to the sounds associated with successful platform setup and flake
removal.
Paleoanthropologists have long emphasized the importance of chipped stone
technologies, language, encephalization, complex cognition, and cumulative culture for
defining H. sapiens as a species. Indeed, there has been a coordinated effort among
researchers to find the earliest traces for each of these features. Some of these features,
such as larger cranial capacity and chipped stone tools, leave clear traces behind in the
form of fossil and artifactual evidence. Tracing the evolution of language, cognition, and
culture, however, requires some more creativity. This thesis has embraced
interdisciplinary methods and theory to search for traces of these features’ evolutionary
past in the functioning modern human brain. What has become increasingly clear as a
result of this experiment is that the co-occurrence of this list of features in modern
292
humans is likely not by coincidence. Rather, the hypothesis proposed here states than an
emphasis on bifacial tools for procuring once unobtainable food items led to selection for
individuals with advanced imitative abilities and enhanced cognition in the form of
increased working memory capacity and stability. Additionally, kin selective pressures
would have promoted more efficient skill transmission behaviors, such as theory of mind,
joint attention, and basic teaching. An increase in grey and white matter in involved
neocortical structures would have followed as a result of selection for working memory,
leading to larger brain size in general. These novel, multimodal, cortico-cortical
connections and the increased lateralization of functions of the brain brought on by
skilled motor processing would provide the neural framework upon which language
could build, eventually forming the modern human pattern of technology, language,
intelligence, and brain size.
293
APPENDIX A:
SUBJECT SELECTION MATERIALS
This appendix includes all the materials used for the selection of subjects for
participation in the study. Individuals interested in participating were first given a short
eligibility questionnaire to fill out that included the Benton Neuropsychology Clinic
Handedness test. Those individuals who were determined to be potentially eligible for
participation (see inclusion criteria in Chapter 5) were asked to fill out a form detailing
their psychiatric and neurologic history, which also included a drug screening test called
Drug Abuse Screen Test (DAST-10).
Eligibility questionnaire and Benton Neuropsychology Clinic Handedness test
Thank you for your interest in this study. The answers that you provide on this document
will help determine if you are eligible to participate. Once completed, please return this
document to Shelby Putt in 114 Macbride Hall or by e-mail at [email protected].
Name: __________________________________________
E-mail address: ___________________________________
Phone number: ___________________________________
Age: _____
Sex: _____
1. Are you a student at the University of Iowa?
Yes [ ]
No [ ]
2. If you answered yes to the prior question, what type of student are you
(undergraduate, graduate, professional, etc.)? _______________________
3. If applicable, what is your major? _______________________
294
4. Do you have any experience in making stone tools?
Yes [ ]
No [ ]
If you answered yes to #4, could you describe your experience making stone tools
in more detail (flintknapping class, self-taught, how long ago? etc.)?
________________________________________________________________
5. Is there any reason why you may not have the full function of your arms?
Yes [ ]
No [ ]
If yes, please explain:
__________________________________________________________________
6. Do you have a history of mental illness, learning disability, seizures, autism, or
brain injury (including stroke, anoxia and hypoxia, brain tumor, infections of the
brain, etc.)?
Yes [ ]
No [ ]
7. Do you have a history of drug/alcohol abuse, or are you currently using
recreational or prescription drugs that may alter your state of consciousness?
Yes [ ]
No [ ]
Handedness:
1. Are you
RIGHT handed [ ] LEFT handed [ ]
2. Have you always been right/left handed?
Yes [ ]
No [ ]
295
MIXED handed [ ]
3. Which hand do you use for the following?
Always
right
Usually
right
Either hand Usually left
Always left
Writing?
Drawing?
Throwing a
ball?
Using
scissors?
Brushing
your teeth?
Using a
knife?
Using a
spoon?
Holding the
handle of a
shovel?
Striking a
match?
Twisting the
lid off a jar?
4. Are there any other activities for which you use your least preferred hand?
5. Which foot do you kick with?
Left [ ]
Right [ ]
Both [ ]
6. Familial handedness:
Right
Left
Mixed
Don’t know
Mother
Father
Brothers (how
many? ____)
Sisters (how
many? _____)
Children (how
many? _____)
7. Do you know of any other left-handedness or mixed-handedness in your family?
Yes [ ]
No [ ]
296
Psychiatric, neurologic, and drug screening questionnaire
Thank you for participating in this study. The answers that you provide on this document are
kept strictly confidential and will only be seen by the experimenters of this study. Because this
experiment is about the brain, it is very important that we learn a bit more about yours. Once
completed, please return this document to Shelby Putt in 114 Macbride Hall or by e-mail at
[email protected].
Past Psychiatric History:
Please check if you have experienced or have been diagnosed with any of the following:
___ Attention Deficit Hyperactivity Disorder (ADHD)
___ Anxiety
___ Depression
___ Bipolar Disorder
___ Mania
___ Obsessive Compulsive Disorder (OCD)
___ Tics/ Tourette’s Syndrome
___ Pervasive Developmental Delay
___ Autism
___ Asperger’s Disorder
___ Schizophrenia
___ Losing touch with reality
___ Learning difficulties
___ Mental retardation
___ Post-Traumatic Stress Disorder (PTSD)
List any inpatient psychiatric hospitalizations (dates of hospitalization, reason):
297
Neurological history:
Please check if you have ever had:
___Stroke
___Transient ischemic attack
___Huntington’s disease
___Epilepsy
___Cerebral palsy
___Multiple sclerosis
1. Have you ever been seen by a neurologist or neurosurgeon?
Yes [ ]
No [ ]
2. Have you ever had a head injury involving unconsciousness?
Yes [ ]
No [ ]
If so, how long were you unconscious?
3. Have you ever required overnight hospitalization for a head injury?
Yes [ ]
No [ ]
4. Have you ever had encephalitis or meningitis?
Yes [ ]
No [ ]
5. Have you ever had a cancer other than skin cancer in the past three years?
Yes [ ]
No [ ]
6. Have you ever been resuscitated?
Yes [ ]
No [ ]
7. Have you ever had heart surgery?
Yes [ ]
No [ ]
8. Have you ever had a heart attack?
Yes [ ]
No [ ]
If so, did you have a change in your memory, ability to talk, or solve problems 24 hours
later?
Yes [ ]
No [ ]
298
9. Have you taken any medications for mental or emotional problems in the past five years?
Yes [ ]
No [ ]
10. Have you ever had seizures?
Yes [ ]
No [ ]
11. Have you ever had any illness which caused a permanent decrease in memory and
cognition?
Yes [ ]
No [ ]
12. Have you ever had brain surgery?
Yes [ ]
No [ ]
13. Have you ever been diagnosed as having a brain tumor?
Yes [ ]
No [ ]
14. Have you ever had major surgery with general anesthesia?
Yes [ ]
No [ ]
If so, did you have any change in your memory, ability to talk or solve problems one
week after surgery?
Yes [ ]
No [ ]
Drug use history:
The following questions concern information about your possible involvement with drugs, not
including alcoholic beverages, during the past 12 months. In the statements, “drug abuse” refers
to the use of prescribed or over-the-counter drugs, which may include cannabis (e.g. marijuana,
hash), solvents, tranquilizers (e.g. Valium), barbiturates, cocaine, stimulants (e.g. speed),
hallucinogens (e.g. LSD) or narcotics (e.g. heroin).
1. Have you used drugs other than those required for medical reasons?
Yes [ ]
No [ ]
2. Do you abuse more than one drug at a time?
Yes [ ]
No [ ]
3. Are you always able to stop using drugs when you want to?
Yes [ ]
No [ ]
4. Have you had “blackouts” or “flashbacks” as a result of drug use?
Yes [ ]
No [ ]
5. Do you ever feel bad or guilty about your drug use?
Yes [ ]
No [ ]
299
6. Do your parents (or spouse) ever complain about your involvement with drugs?
Yes [ ]
No [ ]
7. Have you neglected your family because of your use of drugs?
Yes [ ]
No [ ]
8. Have you engaged in illegal activities in order to obtain drugs?
Yes [ ]
No [ ]
9. Have you ever experienced withdrawal symptoms (felt sick) when you stopped taking
drugs?
Yes [ ]
No [ ]
10. Have you had medical problems as a result of your drug use (e.g. memory loss, hepatitis,
convulsions, bleeding, etc.)?
Yes [ ]
No [ ]
300
APPENDIX B:
SUPPLEMENTARY METHODS AND RESULTS
Supplementary materials for Chapter 5
Table A1. Functional near-infrared spectroscopy (fNIRS) motion processing parameters.
Motion Processing
Parameters
tPCA
Description
Identifies motion artifacts in input data
matrix (d). If any active data channel
exhibits a signal change greater than the
STDEVtresh or AMPthresh, then a segment
of data around that time point is marked as a
motion artifact.
tMotion
1.0
Checks for signal change indicative of a
motion artifact over time range (tMotion).
Units of seconds.
tMask
1.0
Marks data over +/- tMask seconds around
the identified motion artifact as a motion
artifact. Units of seconds.
STDEVthresh
50.0
If the signal (d) for any given active channel
changes by more than STDEVthresh x
STDEV(d) over the time interval (tMotion),
then this time point is marked as a motion
artifact. The standard deviation is
determined for each channel independently.
AMPthresh
0.50
If the signal (d) for any given active channel
changes by more than AMPthresh over the
time interval tMotion, then this time point is
marked as a motion artifact.
301
Table A1—Continued
Parameters
Description
nSV
0.97
This is the number of principal components
to remove from the data. If this number is
less than 1, then the filter removes the first n
components of the data that removes a
fraction of the variance up to nSV. Yücel et
al. (2014) use nSV=0.97.
Maxlter
5
Maximum number of iterations. Yücel et al.
(2014) use maxlter=5.
Motion Processing
Motion Artifacts by
Channel
Identifies motion artifacts in an input data
matrix (d). If any active data channel
exhibits a signal change greater than
STDEVthresh or AMPthresh, then a
segment of data around that time point is
marked as a motion artifact.
tMotion
1.0
Checks for signal change indicative of a
motion artifact over time range (tMotion).
Units of seconds.
tMask
1.0
Marks data over +/- tMask seconds around
the identified motion artifact as a motion
artifact. Units of seconds.
STDEVthresh
50.0
If the signal (d) for any given active channel
changes by more than STDEVthresh x
stdev(d) over the time interval tMotion, then
this time point is marked as a motion
artifact. The standard deviation is
determined for each channel independently.
AMPthresh
0.50
If the signal (d) for any given active channel
changes by more than AMPthresh over the
time interval (tMotion), then this time point
is marked as a motion artifact.
302
Table A1—Continued
Motion Processing
Parameters
Stim exclude
tRange
Description
Excludes stimulus markers that fall within
the time points identified as motion artifacts
from the hemodynamic response function
(HRF) calculation.
-1.0-20.0
An array of 2 numbers specifying how
many seconds surrounding motion artifacts
to consider as excluded data and therefore
exclude any stimuli which fall within those
buffers.
Bandpass Filter
Performs a bandpass filter.
hpf
0.016
High pass filter frequency (Hz). Typical
values are 0 to 0.02.
lpf
0.50
Low pass filter frequency (Hz). Typical
values are 0.5 to 3.
OD to Conc
ppf
Block Average
Converts optical densities to concentrations.
6.0-6.0
Partial path length factors for each
wavelength. If there are 2 wavelengths of
data, then this is a vector of 2 elements.
Typical value is ~6 for each wavelength if
the absorption change is uniform over the
volume of tissue measured.
Calculates the block average given the
stimulation conditions in s over the time
range (tRange). The baseline of the average
is set to zero by subtracting the mean of the
average for t<0. If a stimulus occurs too
close to the start or end of the data such that
tRange extends outside of the data range,
then the trial is excluded from the average.
303
Table A1—Continued
Motion Processing
Parameters
Description
trange
0.0-20.0
Defines the range for the block average.
Units of seconds.
Find Outlier Trials
Finds trials that are outliers with respect to
the average HRF. Removes those trials from
the stimulus vector (s) and re-averages the
results. The mean and standard deviation of
the trials are found over the time range
specified by tRange. Outliers are trials with
a mean deviating more than stdThresh
standard deviations from the mean.
tRange
0.0-20.0
The time range over which the mean is
estimated. Units of seconds.
stdThresh
3.0
The number of standard deviations that a
trial must deviate to be considered an
outlier.
minNtrials
5
Only removes outliers if number of trials for
the given condition is equal to or greater
than this limit.
General lineral model
(GLM) HRF
This script estimates the HRF with options
to specify the temporal basis function type
and corresponding parameters, whether or
not to perform simultaneous regression of
short separation channels, drift order, and
whether or not to correct for motion
artifacts.
tRange
0.0-20.0
Defines the range for the block average.
Units of seconds.
glmSolveMethod
1
This specifies the GLM solution method to
use. Uses ordinary least squares (Ye et al.
2009).
304
Table A1—Continued
Motion Processing
Parameters
Description
IdxBasis
2
2 (tau sigma T) applied to both HbO and
HbR of (tau1 sigma1 T1 tau 2 sigma2 T2)
paramsBasis
0.1 3.0,
10.0, 1.8,
3.0, 10.0
Parameters for the basis function depends
on idxBasis.
rhoSD ssThresh
10.0
flagSSmethod
1
Maximum distance for a short separation
measurement. Follows the static estimate
procedure described in Gagnon et al.
(2011).
Performed with the short separation
channel with the greatest correlation.
driftOrder
3
Polynomial drift correction of this order.
flagMotionCorrect
0
Does not correct between motion epochs.
Supplementary materials for Chapter 6
With contributions from Sobanawartiny Wijeakumar (University of East Anglia)
Participants
Thirty-one healthy, right-handed participants (16 females, 15 males; age [mean ±
SD] 24.0 ± 8.1 years) took part in the experiment to learn how to make replicative
Oldowan and Acheulian stone tools while the oxygenated and deoxygenated hemoglobin
(oxy-Hb and deoxy-Hb) levels in different regions of their brains were monitored with
fNIRS. They were segregated into two groups based on their performance during a
manual dexterity test so that dexterity levels were equally distributed. One group received
305
verbal language instruction (n = 15; 8 females, 7 males), and the other group received
nonverbal communication instruction only (n = 16; 8 females, 8 males).
Localization of regions of interest (ROIs)
ROIs were determined by mining the coordinates for superficial cerebral cortex
regions with significant activation reported in three stone knapping studies that involve
either PET or fMRI by Stout and collaborators (Stout and Chaminade 2007; Stout et al.
2008; Stout et al. 2011). To further investigate the supposed involvement of the vlPFC
during the transition to bifacial flaking, Table 2 in Badre and Wagner (2007) was
consulted, which averages the coordinates for the vlPFC reported in six other studies.
Similarly, to test for the involvement of the dorsolateral prefrontal cortex (dlPFC) during
Early Stone Age (ESA) tool manufacture, coordinates for dlPFC activation were
compiled from Pessoa et al. (2002) and Pessoa and Ungerleider (2004).
A custom optode geometry was designed using the EEG 10-20 coordinate system
that would determine the placement of sources and detectors onto an EasyCAP (Brain
Products GmBH, Germany) to be worn by the participants during the neuroimaging
sessions. The locations of the sources and detectors were digitized using a Polhemus
Patriot™ Motion Tracking System (Colchester, VT) and projected onto an adult atlas
head available in AtlasViewer GUI in the HOMER2 software package (Huppert et al.
2009). Final adjustments were made to the optode geometry after performing Monte
Carlo simulations to create a sensitivity distribution for each source-detector pair (i.e., the
sensitivity of each source-detector pair to detecting changes in absorption of NIR light)
306
and visually inspecting the results in Slicer. Further information on these methods is
available in Wijeakumar et al. (2015). The end result was an optode geometry that
provided coverage surrounding the central sulcus, lateral prefrontal, superior temporal,
and inferior parietal regions. The activation of occipital, inferior temporal, posterior
parietal, and prefrontal regions anterior to the posterior dlPFC, such as the frontopolar
cortex, during ESA toolmaking cannot necessarily be excluded because the optode
geometry design for the present study did not record from these regions.
Image acquisition and processing
fNIRS data were acquired at 25 Hz with a 24-channel TechEn CW6 system with
wavelengths of 690 nm and 830 nm located at the University of Iowa. Light was
delivered to a customized cap via fiber optic cables. HOMER2 software
(www.nmr.mgh.harvard.edu/PMI/resources/Homer2) was employed to demean and
convert the data into OD measures. tPCA was applied to each assigned stimulus marker
in the three tasks mentioned above to eliminate noise and motion artifacts (Yücel et al.
2014). Using a GLM, a beta value (β) was obtained for the oxy-Hb measures in every
channel for all conditions in every task for each subject.
The image reconstruction process is summarized briefly here. Additional details
can be found in Wijeakumar et al. (in press). Scalp landmarks from the session that had
the best symmetry were chosen as the reference for each subject. The landmarks from the
other two sessions were transformed (linear) to fit this reference set of landmarks. The
transformation matrices were applied to the corresponding source and detector positions.
307
AtlasViewerGUI (available within HOMER2) was used to project the points onto an
adult atlas using a relaxation algorithm. The projected geometry was used to run Monte
Carlo simulations based upon a GPU-dependent Monte Carlo algorithm (Fang and Boas
2009) for each session and subject. This resulted in sensitivity profiles (100 million
photons) for each channel of the probe geometry. Head volumes and sensitivity profiles
of channels were converted to niftii images. Subject-specific head volumes were skullstripped and transformed to the head volume in the native atlas space using an affine
transform (BRAINSFit in Slicer 3D). The transformation matrix obtained was applied to
the sensitivity profiles to move them to the transformed head volume space
(BRAINSResample in Slicer3D). Sensitivity profiles for all channels (excluding the short
source-detector channels) were thresholded to include voxels with an OD of greater than
0.0001. These profiles were summed to create a session and subject-specific mask, and
then these masks were summed across all sessions and subjects. Those shared voxels
were used to create an intersection mask across participants.
The beta coefficients obtained for each channel, condition (within each task), and
subject for oxy-Hb and total Hb concentration levels were combined with the forward
model results obtained from the Monte Carlo simulations to create voxel-based changes
in oxy-Hb and total Hb concentration.
The relationship between the HRF (beta) in oxy-Hb/total Hb concentration and
that in delta-OD is given by:
%
%
%
!"#$
= ''( % . +. ,-.#
. !-.# + ''( % . +. ,-.0
. !-.0 (1)
where, d is the source-detector distance and ppf is the partial pathlength factor (Li
et al., 2004).
308
Equation (1) can be re-written to accommodate the forward model and betas from
each channel for each wavelength to estimate voxel-wise changes in oxy-Hb and total Hb
concentrations,
()
()
!. $%&'
. *%&' + !. $%&,
. *%&,
$ () ./ ()
= %&'
(((!. $%&' . *%&' + !. $%&, . *%&,
$%&'
./ (-
()
$%&,
./ ()
Δ123456
.
(Δ127456
$%&, ./ (-
(2)
where, F is the channel-wise sensitivity volumes from the Monte Carlo simulations.
ΔHbOvox and ΔHbRvox are voxel-wise relative changes in oxy-Hb and total Hb
concentrations – this is what is estimated in the image reconstruction process. Note that β
and F are obtained for each channel and are represented as arrays within the matrix
above.
Equation (2) can be rewritten as,
! = #. &
(3)
where,
Y=
L=
%&
!"#$
%'
!"#$
%&
!"#$
.) %&
%+
!"#$
.) %+
%&
!"#*
.) %&
%+
!"#*
.) %+
Δ"#$
X = Δ"#(%&' %&'
Inverting L to solve for X results in an ill-conditioned and under-determined
solution that might be subject to rounding errors. An alternative is to use a popular
regularization method called Tikhonov regularization (Tikhonov 1963). In this case, the
above ‘system’ can be replaced by a regularized ‘system’. The solution is given by the
Gauss-Markov equation,
" = (%& % + (. *)-- %& ..(4)
309
where λ is a regularization parameter that determines the amount of regularization and I is
the identity operator.
The solution to (4) can be found by minimizing the cost function (Calvetti et al.,
2000) ,
"#$% min ) = +. )-.
/
+ 1. |)-)3 |/ (5)
where the size of the regularized solution is measured by the norm λ . |X – X0|2. X0 is an a
priori estimate of X, which is set to zero when no priori information is available. Picking
the appropriate regularization parameter is dependent on the trade-off between fitting Y
and maintaining a small residual (if too much regularization is applied) and eliminating
the contributions of data and rounding errors (if too little regularization is applied).
Hence, an L-curve is plotted between the norms of the solution and the residual. The
corner of this L-curve is identified and the corresponding regularization parameter is used
to estimate X.
Here X is determined for each chromophore and condition (3 SS x 2 Trial types x
2 accuracy conditions) separately. Once Equation (5) is solved, there is now a voxel-wise
estimate of the concentration data. Thus, the best estimate of the channel-wise
concentration data for each condition (from the GLM) has been combined with
information from the photon migration results to create an estimate of the voxel-wise
concentration data for each chromophore, for each condition, and for each subject.
The resultant beta maps were intersected with the Intersection mask to restrict
analyses to the voxels that were common to all sessions and subjects. Consequently, beta
maps were obtained for each condition (within each task) and subject for oxy-Hb and
total Hb concentration levels.
310
Supplementary materials for Chapter 7
Table A2. General motor areas involved in knapping ESA tools at different points in skill
learning.
Localization1
Group Main Effect
Left Inferior frontal gyrus
Left Superior parietal lobule
Task Main Effect
Left Inferior frontal gyrus
Right Paracentral lobule
Right Precentral gyrus
Right Postcentral gyrus
Group x Task Interaction
Right Supramarginal gyrus
Right Angular gyrus
Right Precentral gyrus
Right Precentral gyrus
Right Supramarginal gyrus
Session Main Effect
Left Superior frontal gyrus
Left Postcentral gyrus
Group x Session Interaction
Right Superior temporal gyrus
Left Paracentral lobule
Right Precentral gyrus
Task x Session Interaction
Right Inferior parietal lobule
Left Superior temporal gyrus
Left Inferior frontal gyrus
Right Supramarginal gyrus
MNI Position (mm)
(x, y, z)
Vol.
Mean ± SEM
-52.7, 36.9, 0.9
-27.9, -52, 70.7
2096
912
6.08 ± 0.08
5.18 ± 0.06
-47.2, 41, 12.7
7.2, -20.7, 82.8
62.9, 8.5, 29.1
56.3, -20.7, 51.7
1672
392
232
232
4.90 ± 0.03
5.46 ± 0.12
5.08 ± 0.13
5.01 ± 0.12
64.2, -27.4, 20.7
41.2, -56.4, 51.4
-58.6, 7.8, 20.7
39.9, -18.3, 57.9
62.5, -45.4, 33.4
5712
1304
464
304
256
7.18 ± 0.07
5.23 ± 0.05
4.90 ± 0.07
4.85 ± 0.07
4.52 ± 0.04
23.8, -3.8, 65.5
-60.5, -1.9, 26.5
1528
544
3.99 ± 0.04
3.86 ± 0.06
66.6, -14.9, 13
-9.2, 27.9, 77.8
50.4, -4.4, 36.7
792
536
248
4.75 ± 0.13
3.46 ± 0.03
3.44 ± 0.03
55.5, -37.9, 46.7
-58.1, -34.1, 18.7
-61.5 7 15.3
64.1, -19.3, 36.5
2264
680
368
232
3.90 ± 0.04
3.53 ± 0.03
3.70 ± 0.09
3.76 ± 0.10
1
Areas listed include significant clusters (p < 0.05) from the Task x Group x Session ANOVA that were
either significantly lower than the motor baseline or had similar signals to the motor baseline.
311
APPENDIX C:
SUMMARY OF SUBJECT INTERVIEWS
Figure A1. Subject responses by group to the question, “Did you think with language
while knapping?” Subjects’ responses were coded as one of three categories. A
completely negative response to the question was coded as ‘Spatial Thinking.’ Responses
that indicated minimal involvement of inner speech while thinking about the task were
coded as ‘Spatial Thinking with Some Words,’ and participants who emphasized inner
speech as their main mode of thinking or mentioned recalling entire phrases from the
instruction videos were coded as ‘Inner Speech.” Blue = Verbal, red = Nonverbal. Error
bars represent standard error.
Figure A2. Percentage of subjects to recognize different goals in silent instruction videos
for Oldowan and Acheulian tasks across neuroimaging sessions. Any response indicating
that the subject recognized a difference between the two videos, even if inaccurate, was
considered as recognition of two different goals.
312
Figure A3. Percentage of subjects to accurately describe different goals in silent
instruction videos for Oldowan and Acheulian tasks. Of the subjects who recognized
differences between the silent instruction videos that introduced the Oldowan and
Acheulian tasks during the neuroimaging sessions, those who could accurately describe
these differences were included.
313
REFERENCES
Abler, W.L. (1976). Asymmetry in the skulls of fossil man: Evidence of lateralized brain
function? Brain, Behavior and Evolution 13, 111-115.
Abramiuk, M.A. (2012). The Foundations of Cognitive Archaeology. Cambridge: The
MIT Press.
Aiello, L.C., Dunbar, R.I.M. (1993). Neocortex size, group size and the evolution of
language. Current Anthropology 34, 184-193.
Aiello, L.C., Wells, J.C.K. (2002). Energetics and the evolution of the genus Homo.
Annual Review of Anthropology 31, 323-338.
Aiello, L.C., Wheeler, P. (1995). The expensive tissue hypothesis. Current Anthropology
36, 199-221.
Albert, N.B., Robertson, E.M., Miall, R.C. (2009). The resting human brain and motor
learning. Current Biology 19, 1023-1027.
Ambrose, S.H. (2001). Paleolithic technology and human evolution. Science 291, 17481753.
Ambrose, S.H. (2010). Coevolution of composite-tool technology, constructive memory,
and language: Implications for the evolution of modern human behavior. Current
Anthropology 51, S135-S147.
Amunts, K., Schlaug, G., Jäncke, L., Steinmetz, H., Schleicher, A., Dabringhaus, A.,
Zilles, K. (1997). Motor cortex and hand motor skills: Structural compliance in
the human brain. Human Brain Mapping 5, 206-215.
314
Anderson, J.R. (2004). Cognitive Psychology and Its Implications, 6th ed. New York:
Worth Publishers.
Anderson-Gerfaud, P. (1990). Aspects of behavior in the Middle Paleolithic: Functional
analysis of stone tools from Southwest France. In P. Mellars (Ed.), The
Emergence of Modern Humans (pp. 389-418). Edinburgh: Edinburgh University
Press.
Andrefsky, W. (2005). Lithics: Macroscopic Approaches to Analysis, 2nd ed. Cambridge
University Press, Cambridge.
Ando, J., Ono, Y., Wright, M.J. (2001). Genetic structure of spatial and verbal working
memory. Behavior Genetics 31, 615-624.
Arbib, M.A. (2005). From monkey-like action recognition to human language: An
evolutionary framework for neurolinguistics. Behavior and Brain Sciences 28,
105-167.
Arbib, M.A. (2011). From mirror neurons to complex imitation in the evolution of
language and tool use. Annual Review of Anthropology 40, 257-273.
Arbib, M.A. (2012). How the Brain Got Language: The Mirror System Hypothesis. New
York: Oxford University Press.
Arensburg, B., Schepartz, L.A., Tillier, A-M., Vandermeersch, B., Rak, Y. (1990). A
reappraisal of the anatomical basis for speech in Middle Paleolithic hominids.
American Journal of Physical Anthropology 83, 137–146.
Armstrong, D.F. (1999). Original signs: Gesture, sign, and the source of language.
Washington D.C., Gallaudet University Press.
315
Atkinson, R.C., Shiffrin, R.M. (1968). Human memory: A proposed system and its
control processes. In K.W. Spence and J.T. Spence (Eds.), The Psychology of
Learning and Motivation: Advances in Research and Theory, vol 2 (pp. 89-195).
New York: Academic Press.
Augustine, J.R. (1996). Circuitry and functional aspects of the insular lobe in humans.
Brain Research Reviews 22, 229-244.
Baddeley, A.D. (1993). Working memory or working attention? In A.D. Baddeley and I.
Weiskrants (Eds.), Attention: Selection, Awareness, and Control: A Tribute to
Donald Broadbent (pp. 152-170). Oxford: Oxford University Press.
Baddeley, A.D. (2000). The episodic buffer: A new component of working memory?
Trends in Cognitive Sciences 4, 417-423.
Baddeley, A.D. (2001). Is working memory still working? American Psychologist 11,
851-864.
Baddeley, A.D. (2010). Working memory. Current Biology 20, R136-R140.
Baddeley, A.D., Hitch, G.J. (1974). Working memory. In G.A. Bower (Ed.), Recent
Advances in Learning and Motivation, vol. 8 (pp. 47-89). New York: Academic
Press.
Badre, D., Poldrack, R.A., Pare-Blagoev, E.J., Insler, R.Z., Wagner, A.D. (2005).
Dissociable controlled retrieval and generalized selection mechanisms in
ventrolateral prefrontal cortex. Neuron 47, 907–918.
Badre D., Wagner, A.D. (2005). Frontal lobe mechanisms that resolve proactive
interference. Cerebral Cortex 15, 2003–2012.
316
Badre, D., Wagner, A.D. (2006). Computational and neurobiological mechanisms
underlying cognitive flexibility. Proceedings of the National Academy of Sciences
of United States of America 103, 7186–7191.
Badre, D., Wagner, A.D. (2007). Left ventrolateral prefrontal cortex and the cognitive
control of memory. Neuropsychologia 45, 2883-2901.
Baker, P. (1995). Motivation in creole genesis. In P. Baker (Ed.), From Contact to Creole
and Beyond (pp. 3-15). London, University of Westminster Press.
Baker, P. (1997). Directionality in pidginization and creolization. In A.K. Spears and D.
Winford (Eds.), The Structure and Status of Pidgins and Creoles (pp. 91-109).
Philadelphia, John Benjamins.
Balzeau, A., Holloway, R.L., Grimau-Hervé, D. (2012). Variation and asymmetries in
regional brain surface in the genus Homo. Journal of Human Evolution 62, 696706.
Bargalló, A., Mosquera, M. (2014). Can hand laterality be identified through lithic
technology? Laterality 18, 1-27.
Barnard, P.J. (1985). Interacting cognitive subsystems. In A. Ellis (Ed.), vol. 2: Progress
in the Psychology of Language (pp. 197-258). London: Lawrence Erlbaum
Associates.
Barnard, P.J., Duke, D.J., Byrne, R.W., Davidson, I. (2007). Differentiation in cognitive
and emotional meanings: An evolutionary analysis. Cognition and Emotion 21,
1155-1183.
Barth, J., Call, J. (2006). Tracking the displacement of objects: A series of tasks with
great apes and young children. Journal of Experimental Psychology 32, 239-252.
317
Barton, R.A. (1996). Neocortex size and behavioural ecology in primates. Proceedings of
the Royal Society B 263, 173-177.
Barton, R., Dunbar, R.I.M. (1997). Evolution of the social brain. In R. Byrne and A.
Whiten (Eds.), Machiavellian Intelligence, vol. II, (pp. 240-263). Cambridge,
Cambridge University Press.
de Beaune, S.A., Coolidge, F.L., Wynn, T. (2009). Cognitive Archaeology and Human
Evolution. New York: Cambridge University Press.
Bélanger, M., Allaman, I., Magistretti, P.J. (2011). Brain energy metabolism: Focus on
astrocyte-neuron metabolic cooperation. Cell Metabolism 14, 724-738.
Belfer-Cohen, A., Goren-Inbar, N., Gowlett, J.A.J. (1994). Cognition and communication
in the Levantine Lower Palaeolithic. World Archaeology 26, 144-157.
Belfer-Cohen, A., Hovers E. (2010). Modernity, enhanced working memory, and the
Middle to Upper Paleolithic record in the Levant. Current Anthropology 51,
S167-S175.
Bell, C. (2015). What were they thinking? Using electroencephalogram (EEG) to map
brain activations during stone tool manufacture. Paper presented at the 80th
Annual Meeting of the Society for American Archaeology, San Francisco, April
15-19.
Bermúdez de Castro, J.M., Bromage, T.G., Jalvo, Y.F. (1988). Buccal striations on fossil
human anterior teeth: Evidence of handedness in the middle and early Upper
Pleistocene. Journal of Human Evolution 17, 403-412.
318
Berti, A., Bottini, G., Gandola, M., Pia, L., Smania, N., Stracciari, A., Castiglioni, I.,
Vallar, G., Paulesu, E. (2005). Shared cortical anatomy for motor awareness and
motor control. Science 309, 488-491.
Beyene, Y., Katoh, S., WoldeGabriel, G., Hart, W.K., Uto, K., Sudo, M., Kondo, M.,
Hyodo, M., Renne, P.R., Suwa, G., Asfaw, B. (2013). The characteristics and
chronology of the earliest Acheulean at Konso, Ethiopia. Proceedings of the
National Academy of Sciences 110, 1584-1591.
Bickerton, D. (1990). Language and Species. Chicago, The University of Chicago Press.
Bickerton, D. (2002). From protolanguage to language. In T.J. Crowe (Ed.), The
Speciation of Modern Homo Sapiens (pp. 103-120), Oxford, Oxford University
Press.
Bickerton, D. (2003). Symbol and structure: A comprehensive framework for language
evolution. In M.H. Christiansen and S. Kirby (Eds.), Language Evolution (pp. 7793). Oxford, Oxford University Press.
Bickerton, D. (2007). Language evolution: A brief guide for linguists. Lingua 117, 510526.
Biedowski, C., Rahm, B., Rowe, B. (2009). What ‘works’ in working memory? Separate
systems for selection and updating of critical information. Journal of
Neuroscience 43, 13735-13741.
Binder, J.R., Frost, J.A., Hammeke, T.A., Cox, R.W., Rao, S.M., Prieto, T. (1997).
Human brain language areas identified by functional magnetic resonance imaging.
The Journal of Neuroscience 17, 353-362.
Binford, L.R. (1962). Archaeology as anthropology. American Antiquity 28, 217-225.
319
Binford, L.R. (1965). Archaeological systematics and the study of culture process.
American Antiquity 31, 203-210.
Binford, L.R., Binford, S.R. (1966). A preliminary analysis of function variability in the
Mousterian of the Levallois facies. American Anthropologist 68, 238-295.
Binkofski, F., Buccino, G. (2006). The role of ventral premotor cortex in action execution
and action understanding. Journal of Physiology Paris 99, 396-405.
Binneman, J., Beaumont, P. (1992). Use-wear analysis of two Acheulean handaxes from
Wonderwerk Cave, Northern Cape. Southern African Field Archaeology 1, 92-97.
Black, L.T. (1989). Review of The archaeology of perception: Traces of depiction and
language. Current Anthropology 30, 139.
Blake, E.C., Cross, I. (2008). Flint tools as portable sound-producing objects in the
Upper Palaeolithic context: An experimental study. In P. Cunningham, J. Heeb,
and R. Paardekooper (Eds.), Experiencing Archaeology by Experiment (pp. 1-20).
Oxford: Oxbow Books.
Blakemore, S.J., Choudhury, S. (2006). Development of the adolescent brain:
Implications for executive function and social cognition. Journal of Child
Psychology and Psychiatry 47, 296-312.
Blokland, G.A.M., McMahon, K.L., Thompson, P.M., Martin, N.G., de Zubicaray, G.I.,
Wright, M.J. (2011). Heritability of working memory brain activation. The
Journal of Neuroscience 31, 10882-10890.
Boas, D.A., Dale, A.M., Franceschini, M.A. (2004). Diffuse optical imaging of brain
activation: Approaches to optimizing image sensitivity, resolution, and accuracy.
NeuroImage 23, S275-S288.
320
Boëda, E., Connan, J., Dessort, D., Muhesen, S., Mercier, N., Valladas, H., Tisnerat, N.
(1996). Bitumen as a hafting material on Middle Palaeolithic artefacts. Nature
380, 336-338.
de Boer, B. (2002). Evolving sound systems. In A. Cangelosi and D. Parisi (Eds.),
Simulating the Evolution of Language (pp. 79-98). London, Springer-Verlag.
Boesch, C. (1993). Aspects of transmission of tool-use in wild chimpanzees. In K.R.
Gibson and T. Ingold (Eds.), Tools, Language, and Cognition in Human
Evolution (pp. 171-184). Cambridge: Cambridge University Press.
Boesch, C. (2012). Wild Cultures: A Comparison between Chimpanzee and Human
Cultures. Cambridge: Cambridge University Press.
Boesch, C., Boesch, H. (1993). Diversity of tool use and tool-making in wild
chimpanzees. In A. Berthelet and J. Chavaillon (Eds.), The Use of Tools by
Human and Non-Human Primates (pp. 158-168). Oxford: Clarendon Press.
Bookheimer, S. (2002). Functional MRI of language: new approaches to understanding
the cortical organization of semantic processing. Annual Review of
Neuroscience 25, 151–188.
Bordes, F.H. (1950). Principes d’une method d’étude des techniques de débitage et de la
typologie du Paléolithique ancient et moyen. L’Anthropologie 54, 19-34.
Bordes, F.H. (1953). Essai de classification des industries “moustériennes.” Bulletin de la
Société Préhistorique Française 50, 457-466.
Bordes, F.H. (1968). The Old Stone Age. New York: McGraw-Hill.
Bordes, F.H. (1972). A Tale of Two Caves. New York: Harper and Row.
321
Botha, R. (2006). Pidgin languages as a putative window on language evolution.
Language and Communication 26, 1-14.
Botvinick, M.M., Wang, J., Cowan, E., Roy, S., Bestianen, C., Mayo, J.P., Houk, J.C.
(2009). An analysis of immediate serial recall performance in a macaque. Animal
Cognition 12, 671-678.
Bradshaw, J.L., Nettleton, N.C. (1982). Language lateralization to the dominant
hemisphere: Tool use, gesture and language in hominid evolution. Current
Psychological Reviews 2, 171-192.
Briggs, J.L. (1970). Never in Anger: Portrait of an Eskimo Family. Cambridge: Harvard
University Press.
Bril, B., Smaers, J., Steele, J., Rein, R., Nonaka, T., Dietrich, G., Biryukova, E., Hirata,
S., Roux, V. (2012). Functional mastery of percussive technology in nut-cracking
and stone-flaking actions: Experimental comparison and implications for the
evolution of the human brain. Philosophical Transactions of the Royal Society B
367, 59-74.
Broca, P. (1861). Remarks on the seat of the faculty of articulated language, following an
observation of aphemia (loss of speech). Bulletin de la Société Anatomique 6,
330-357.
Brothers, L. (1990). The social brain: A project for integrating primate behavior and
neuropsychology in a new domain. Concepts in Neuroscience 1, 27-51.
Brown, S., Martinez, M.J., Parsons, L.M. (2006). Music and language side by side in the
brain: A PET study of the generation of melodies and sentences. European
Journal of Neuroscience 23, 2791-2803.
322
Bruner, E. (2008). Comparing endocranial form and shape differences in modern humans
and Neandertals: A geometric approach. PaleoAnthropology 2008, 93-106.
Bruner, E., De La Cuétara, J.M., Holloway, R. (2011). A bivariate approach to the
variation of the parietal curvature in the genus Homo. The Anatomical Record
294, 1548-1556.
Bruner, E., Holloway, R.L. (2010). A bivariate approach to the widening of the frontal
lobes in the genus Homo. Journal of Human Evolution 58, 138-146.
Buccino, G., Binkofski, F., Fink, G.R., Fadiga, L., Fogassi, L., Gallese, V., Seitz, R.J.,
Zilles, K., Rizzolatti, G., Freund, H.-J. (2001). Action observation activates
premotor and parietal areas in a somatotopic manner: an fMRI study. European
Journal of Neuroscience 13, 400–404.
Buccino, G., Lui, F., Canessa, N., Patteri, I., Lagravinese, G., Benuzzi, F., Porro, C.A.,
Rizzolatti, G. (2004a). Neural circuits involved in the recognition of actions
performed by non-conspecifics: an fMRI study. Journal of Cognitive
Neuroscience 16, 1–14.
Buccino, G., Vogt, S., Ritzl, A., Fink, G.R., Zilles, K., Freund, H-J., Rizzolatti, G.
(2004b).Neural circuits underlying imitation learning of hand actions: An eventrelated fMRI study. Neuron 42, 323–334.
Buckley, C., Steele, J. (2002). Evolutionary ecology of spoken language: Co-evolutionary
hypotheses are testable. World Archaeology 34, 26-46.
Buss, A.T., Fox, N., Boas, D.A., Spencer, J.P. (2014). Probing the early development of
visual working memory capacity with functional near-infrared spectroscopy.
Neuroimage 85, 314-325.
323
Buss, D.M. (1995). Evolutionary psychology: A new paradigm for psychological science.
Psychological Inquiry 6, 1-30.
Button, K.S., Ioannidis, J.P.A., Mokrysz, C., Nosek, B.A., Flint, J., Robinson, E.S.,
Munafò, M.R. (2013). Power failure: Why small sample size undermines the
reliability of neuroscience. Nature Reviews Neuroscience 14, 365-376.
Byrne, R.W. (2004). The manual skills and cognition that lie behind hominid tool use. In
A.E. Russon and D.R. Begun (Eds.), The Evolution of Thought: Evolutionary
Origins of Great Ape Intelligence (pp. 31-44). Cambridge: Cambridge University
Press.
Byrne, R.W. (2005). The maker not the tool: The cognitive significance of great ape
manual skills. In V. Roux and B. Bril (Eds.), Stone Knapping: The Necessary
Conditions for a Uniquely Hominid Behaviour (pp. 159-170). Cambridge:
McDonald Institute for Archaeological Research.
Caggiano, V., Fogassi, L., Rizzolatti, G., Their, P., Giese, M.A., Casile, A. (2011). Viewbased encoding of actions in mirror neurons of area F5 in macaque premotor
cortex. Current Biology 21, 144-148.
Caggiano, V., Fogassi, L., Rizzolatti, G., Their, P., Casile, A. (2009). Mirror neurons
differentially aencode the peripersonal and extrapersonal space of monkeys
Science 324, 403-406.
Caldwell, J.R. (1959). The new American archeology. Science 129, 303-307.
Cangelosi A., Parisi D. (2002). Computer simulation: A new scientific approach to the
study of language evolution. In A. Cangelosi and D. Parisi (Eds.), Simulating the
Evolution of Language (pp. 3-28). London, Springer-Verlag.
324
Carmody, R.N., Weintraub, G.S., Wrangham, R.W. (2011). Energetic consequences of
thermal and nonthermal food processing. Proceedings of the National Academy of
Science 108, 19199-19203.
Carr, L., Iacoboni, M., Dubeau, M-C., Mazziotta, J.C., Lenzi, G.L. (2003). Neural
mechanisms of empathy in humans: A relay from neural systems for imitation to
limbic areas. Proceedings of the National Academy of Sciences 100, 5497-5502.
Carruthers, P. (2013). Evolution of working memory. Proceedings of the National
Academy of Sciences 110, 10371-10378.
Cartmill, M., Pilbeam, D., Isaac, G. (1986). One hundred years of paleoanthropology.
American Scientist 74, 410-420.
Carvalho, S., Cunha, E., Sousa, C., Matsuzawa, T. (2008). Chaînes opératoires and
resource-exploitation strategies in chimpanzee (Pan troglodytes) nut cracking.
Journal of Human Evolution 55, 148-163.
Caspers, S., Zilles, K., Laird, A.R., Eickhoff, S.B. (2010). ALE meta-analysis of action
observation and imitation in the human brain. NeuroImage 50, 1148–1167.
Chase, P.G. (2006). The Emergence of Culture: The Evolution of a Uniquely Way of Life.
New York: Springer.
Chase, P.G., Dibble, H.L. (1987). Middle Paleolithic symbolism: A review of current
evidence and interpretations. Journal of Anthropological Archaeology 6, 263-296.
Chein, J.M., Schneider, W. (2005). Neuroimaging studies of practice-related change:
fMRI and meta-analytic evidence of a domain-general control network for
learning. Cognitive Brain Research 25, 607-623.
325
Chen, G., Adleman, N.E., Saad, Z.S., Leibenluft, E., Cox, R.W. (2014). Applications of
multivariate modeling to neuroimaging group analysis: A comprehensive
alternative to univariate general linear model. NeuroImage 99, 571-588.
Cheney, D.L., Seyfarth, R.M. (1990). How Monkeys See the World: Inside the Mind of
Another Species. Chicago, Chicago University Press.
Chevalier-Skolnikoff, S. (1977). A Piagetian model for describing and comparing
socialization in monkey, ape, and human infants. In S. Chevalier-Skolnikoff and
F.E. Poirier (Eds.), Primate Biosocial Development. New York: Garland Press.
Chomsky, N. (1975). Reflections on Language. New York, Pantheon.
Chomsky, N. (1986). Knowledge of Language: Its Nature, Origin, and Use. Westport,
CT, Praeger.
Chou, T.L., Booth, J.R., Bitan, T., Burman, D.D., Bigio, J.D., Cone, N.E., Lu, D., Cao, F.
(2006). Developmental and skill effects on the neural correlates of semantic
processing to visually presented words. Human Brain Mapping 27, 915-924.
Christiansen, M.H., Dale, R.A.C., Ellefson, M.R., Conway, C.M. (2002). The role of
sequential learning in language evolution: Computational and experimental
studies. In A. Cangelosi and D. Parisi (Eds.), Simulating the Evolution of
Language (pp. 165-188). London, Springer-Verlag.
Christiansen, M.H., Kirby, S. (2003). Language evolution: The hardest problem in
science? In M.H. Christiansen and S. Kirby (Eds.), Language Evolution (pp 1-15).
Oxford, Oxford University Press.
Churchland, P.C. (2011). Braintrust: What Neuroscience Tells Us about Morality.
Princeton, NJ: Princeton University Press.
326
Clagett, M. (2001). Greek Science in Antiquity. New York: Macmillan.
Clark J.D. (1992). African and Asian perspectives on the origins of modern humans.
Philosophical Transactions of the Royal Society B 337, 201-215.
Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences, 2nd ed.
Hillsdale, NJ: Laurence Erlbaum Associates.
Constantinidis, C., Procyk, E. (2004). The primate working memory networks. Cognitive,
Affective, & Behavioral Neuroscience 4, 444-465.
Coolidge, F.L., Wynn, T. (2001). Executive functions of the frontal lobes and the
evolutionary ascendancy of Homo sapiens. Cambridge Archaeological Journal
11, 255-260.
Coolidge, F.L., Wynn, T. (2005). Working memory, its executive functions and the
emergence of modern thinking. Cambridge Archaeological Journal 15, 5-26.
Coolidge, F.L., Wynn, T. (2009). The Rise of Homo sapiens: The Evolution of Modern
Thinking. Oxford: Wiley-Blackwell.
Corballis, M.C. (2002). From Hand to Mouth: The Origins of Language. Princeton,
Princeton University Press.
Corballis, M.C. (2003a). From hand to mouth: The gestural origins of language. In M.H.
Christiansen and S. Kirby (Eds.), Language Evolution (pp. 201-218). Oxford,
Oxford University Press.
Corballis, M.C. (2003b). From mouth to hand: Gesture, speech, and the evolution of
right-handedness. Brain and Behavioral Sciences 26, 199-260.
Corballis, M.C. (2010). Mirror neurons and the evolution of language. Brain and
Language 112, 25-35.
327
Corballis, M.C. (2013). The Recursive Mind: The Origins of Human Language, Thought,
and Civilization. Princeton, Princeton University Press.
Corbetta, M., Schulman, G.L. (2002). Control of goal-directed and stimulus-driven
attention in the brain. Nature Reviews 3, 201-215.
Cornford, J.M. (1986). Specialized resharpening techniques and evidence for handedness.
In P. Callow and J.M. Cornford (Eds.), La Cotte de St. Brelade 1961-1978:
Excavations by C.B.M. McBurney (pp. 337-414). Norwich, Geo Books.
Cowan, N. (2010). The magical mystery four: How is working memory capacity limited,
and why? Current Directions in Psychological Science 19, 51-57.
Cox, R.W. (1996). AFNI: Software for analysis and visualization of functional magnetic
resonance neuroimages. Computers and Biomedical Research 29, 162-173.
Cox, R.W., Hyde, J.S. (1997). Software tools for analysis and visualization of fMRI data.
NMR in Biomedicine 10, 171-178.
Curtis, C.E., D’Esposito, M. (2003). Persistent activity in the prefrontal cortex during
working memory. Trends in Cognitive Science 7, 415-423.
Custo, A., Boas, D.A., Tsuzuki, D., Dan, I., Mesquita, R., Fischl, B., Grimson, W.E.L.,
Wells, W. (2010). Anatomical atlas-guided diffuse optical tomography of brain
activation. NeuroImage 49, 561-567.
D’Amato, M.R., Salmon, D.P., Colombo, M. (1985). Extent and limits of the matching
concept in monkeys (Cebus apella). Journal of Experimental Psychology (Animal
Behavior) 11, 35-51.
328
Daniel, C., Putt, S., Franciscus, R. (2016). Investigating other causes for stone flake
features attributed to handedness. Poster presented at the 81st Annual Meeting of
the Society for American Archaeology, Orlando, April 6-10.
Darwin, C.R. (1871). vol 1: The Descent of Man and Selection in Relation to Sex. 1st
edition. London: John Murray.
Dasen, P. (1994). Culture and cognitive development from a Piagetian perspective. In
W.J. Lonner and R.S. Malpass (Eds.), Psychology and Culture (pp. 145-149).
Boston: Allyn and Bacon.
Davidson, I. (2002). The finished artifact fallacy: Acheulean hand-axes and language
origins. In A. Wray (Ed.), Transitions to Language (pp. 180-203). Oxford: Oxford
University Press.
Davidson I. (2003). The archaeological evidence of language origins: States of art. In
M.H. Christiansen and S. Kirby (Eds.), Language Evolution (pp. 140-157).
Oxford, Oxford University Press.
Davidson, I. (2010a). The archaeology of cognitive evolution. Wiley Interdisciplinary
Reviews: Cognitive Science 1, 214-229.
Davidson, I. (2010b). The colonization of Australia and its adjacent islands and the
evolution of modern cognition. Current Anthropology 51,b S177-S189.
Davidson, I., McGrew, W.C. (2005). Stone tools and the uniqueness of human culture.
Journal of the Royal Anthropological Institute 11, 793-817.
Davidson, I., Noble, W. (1989). The archaeology of perception: Traces of depiction and
language. Current Anthropology 30, 125-155.
329
Davidson, I., Noble, W. (1993). Tools and language in human evolution. In K.R. Gibson
and T. Ingold (Eds.), Tools, Language, and Cognition in Human Evolution (pp.
363-388). Cambridge: Cambridge University Press.
Deacon, H.J. (2001). Modern human emergence: An African archaeological perspective.
In P.V. Tobias, M.A. Raath, J. Maggi-Cecchi, and G.A. Doyle (Eds.), Humanity
from African Naissance to Coming Millennia: Colloquia in Human Biology and
Palaeoanthropology (pp. 217-226). Florence: University of Florence Press.
Deacon T.W. (1997). The Symbolic Species: The Co-Evolution of Language and the
Brain. New York, W. W. Norton & Company.
d’Errico, F., Henshilwood, C., Lawson, G., Vanhaeren, M., Tillier, A-M., Soressi, M.,
Bresson, F., Maureille, B., Nowell, A., Lakarra, J., Backwell, L., Julien, M.
(2003). Archaeological evidence for the emergence of language, symbolism, and
music—An alternative multidisciplinary perspective. Journal of World Prehistory
17, 1-70.
d’Errico, F., Stringer, C.B. (2011). Evolution, revolution or saltation scenario for the
emergence of modern cultures? Philosophical Transactions of the Royal Society B
366, 1060-1069.
Dennell, R.W. (1998). Grasslands, tool making and the hominid colonization of southern
Asia: A reconsideration. In M.D. Petraglia and R. Korisettar (Eds.), Eary Human
Behaviour in Global Context (pp. 280-303). London: Routledge.
Desrosiers, J., Rochette, A., Hebert, R., Bravo, G. (1997). The Minnesota Manual
Dexterity Test: Reliability, validity and reference values studies with healthy
elderly people. Canadian Journal of Occupational Therapy 64, 270–276.
330
Dibble, H.L. (1984). Interpreting typological variation of Middle Paleolithic scrapers:
Function, style, or sequence of reduction? Journal of Field Archaeology 11, 431436.
Dibble, H.L. (1987). Reduction sequences in the manufacture of Mousterian implements
of France. In O. Soffer (Ed.), The Pleistocene Old World (pp. 33-44). New York:
Planum Press.
Dibble, H.L. (1997). Platform variability and flake morphology: A comparison of
experimental and archaeological data and implications for interpreting prehistoric
lithic technological strategies. Lithic Technology 22, 150-170.
Diez-Martín, F., Yustos, P.S., Uribelarrea, D., Baquedano, E., Mark, D.F., Mabulla, A.,
Fraile, C., Duque, J., Díaz, I., Pérez-González, A., Yravedra, J., Egeland, C.P.,
Organista, E., Domínguez-Rodrigo, M. (2015). The origin of the Acheulean: The
1.7 million-year-old site of FLK West, Olduvai Gorge (Tanzania). Scientific
Reports 5, 1-9.
Dinstein, I., Thomas, C., Behrmann, M., Heeger, D.J. (2008). A mirror up to nature.
Current Biology 18, R13-R18.
Dobbins, I.G., Wagner, A.D. (2005). Domain-general and domain-sensitive prefrontal
mechanisms for recollecting events and detecting novelty. Cerebral Cortex 15,
1768–1778.
Domínguez-Rodrigo, M., Alcalá, L. (2016). 3.3-million-year-old stone tools and butchery
traces? More evidence needed. PaleoAnthropology 2016, 46-53.
331
Domínguez-Rodrigo, M., Serrallonga, J., Juan-Tresserras, J., Alcala, L., Luque, L.
(2001). Woodworking activities by early humans: a plant residue analysis on
Acheulian stone tools from Peninj (Tanzania). Journal of Human Evolution 40,
289-299.
Dronkers, N.F., Redfern, B.B., Knight, R.T. (2000). The neural architecture of language
disorders. In Gazzaniga M.S. (Ed.), The New Cognitive Neurosciences (pp. 949958). Cambridge, MA: MIT Press.
Drummond, H. (1894). The Ascent of Man. The Lowell Lectures. London: Hodder and
Stoughton.
Duffau, H., Capelle, L., Denvil, D., Gatignol, P., Sichez, N., Lopes, M., Sichez, J-P., Van
Effenterre, R. (2003). The role of dominant premotor cortex in language: A study
using intraoperative functional mapping in awake patients. NeuroImage 20, 19031914.
Dunbar, R.I.M. (1992). Neocortex size as a constraint on group size in primates. Journal
of Human Evolution 20, 469-493.
Dunbar, R.I.M. (1993). Coevolution of neocortical size, group size and language in
humans. Behavioral and Brain Sciences 16, 681-735.
Dunbar, R.I.M. (1996). Grooming, Gossip, and the Evolution of Language. London,
Faber & Faber.
Dunbar, R.I.M. (1998). The social brain hypothesis. Evolutionary Anthropology 6:178190.
332
Dunbar, R.I.M. (2003). The origin and subsequent evolution of language. In M.H.
Christiansen and S. Kirby (Eds.), Language Evolution (pp. 219-234). Oxford,
Oxford University Press.
Dunbar, R.I.M. (2011). Evolutionary basis of the social brain. In J. Decety and J.T.
Cacioppo (Eds.), The Oxford Handbook of Social Neuroscience (pp. 28-38).
Oxford, Oxford University Press.
Dunbar, R.I.M., Shultz, S. (2007). Evolution in the social brain. Science 317, 1344-1347.
Dux, P.E., Tombu, M.N., Harrison, S., Rogers, B.P., Tong, F., Marois, R. (2009).
Training improves multitasking performance by increasing the speed of
information processing in human prefrontal cortex. Neuron 63, 127-138.
Eren, M.I., Bradley, B.A., Sampson, C.G. (2011). Middle Paleolithic skill-level and the
individual knapper: An experiment. American Antiquity 76, 229-251.
Evans, P.D., Anderson, J.R., Vallender, E.J., Gilbert, S.L., Malcom, C.M., Dorus, S.,
Lahn, B.T. (2004a). Adaptive evolution of ASPM, a major determinant of
cerebral cortical size in humans. Human Molecular Genetics 13, 489-494.
Evans, P.D., Anderson, J.R., Vallender, E.J., Shim Choi, S., Lahn, B.T. (2004b).
Reconstructing the evolutionary history of microcephalin, a gene controlling
human brain size. Human Molecular Genetics 13, 1139-1145.
Fadiga, L., Craighero, L., D’Ausilio, A. (2009). Broca’s area in language, action, and
music. Annals of the New York Academy of Sciences 1169, 448-458.
Fadiga, L., Fogassi, L., Pavesi, G., Rizzolatti, G. (1995). Motor facilitation during action
observation: a magnetic stimulation study. Journal of Neurophysiology 73, 2608–
2611.
333
Fagot, J., De Lillo, C. (2011). A comparative study of working memory: Immediate serial
spatial recall in baboons (Papio papio) and humans. Neuropsychologia 49, 38703880.
Fang, Q., Boas, D.A. (2009). Monte Carlo simulation of photon migration in 3D turbid
media accelerated by graphics processing units. Optics Express 1, 20178-20190.
Falk, D. (1980a). Hominid brain evolution: The approach from paleoneurology. Yearbook
of Physical Anthropology 23, 93-107.
Falk, D. (1980b). Language, handedness, and primate brains: Did the australopithecines
sign? American Anthropologist 82, 72-73.
Falk, D. (1983). Cerebral cortices of East African early hominids. Science 221:10721074.
Falk, D. (1985). Hadar AL 162-28 endocast as evidence that brain enlargement preceded
cortical reorganization in hominid evolution. Nature 313, 45-47.
Falk, D. (1987a). Brain lateralization in primates and its evolution in hominids. Yearbook
of Physical Anthropology 30, 107-125.
Falk, D. (1987b). Hominid paleoneurology. Annual Review of Anthropology 16, 13-30.
Falk, D. (1989a). Ape-like endocast of “Ape-Man” Taung. American Journal of Physical
Anthropology 80, 335-339.
Falk, D. (1989b). Review of The archaeology of perception: Traces of depiction and
language. Current Anthropology 30, 141-142.
Falk, D. (2012). Hominin paleoneurology: Where are we now? Progress in Brain
Research 195, 255-272.
334
Fedorenko, E., Duncan, J., Kanwisher, N. (2012). Language-selective and domaingeneral regions lie side by side within Broca’s area. Current Biology 22, 20592062.
Ferguson, J.R. (2003). An experimental test of the conservation of raw material in
flintknapping skill acquisition. Lithic Technology 28, 113-131.
Fernandes, M.A., Moscovitch, M., Ziegler, M., Grady, C. (2005). Brain regions
associated with successful and unsuccessful retrieval of verbal episodic memory
as revealed by divided attention. Neuropsychologia 43, 1115–1127.
Ferrari, M., Mottola, L., Quaresima, V. (2004). Principles, techniques, and limitations of
near infrared spectroscopy. Canadian Journal of Applied Physiology 29, 463-487.
Ferrari, M., Quaresima, V. (2012). A brief review on the history of human functional
near-infrared spectroscopy (fNIRS) development and fields of application.
NeuroImage 63, 921-935.
Ferrari, P.F., Gallese, V., Rizzolati, G., Fogassi, L. (2003). Mirror neurons responding to
the observation of ingestive and communicative mouth actions in the ventral
premotor cortex. European Journal of Neuroscience 17, 1703-1714.
Ferrari, P.F., Rozzi, S., Fogassi, L. (2005). Mirror neurons responding to observation of
actions made with tools in monkey ventral premotor cortex. Journal of Cognitive
Neuroscience 17, 212-226.
Fitch, W.T. (2000a). The evolution of speech: A comparative review. Trends in Cognitive
Sciences 4, 258-267.
Fitch, W.T. (2000b). The phonetic potential of nonhuman vocal tracts: Comparative
cineradiographic observations of vocalizing animals. Phonetica 57, 205-218.
335
Fitch, W.T. (2002a). Comparative vocal production and the evolution of speech:
Reinterpreting the descent of the larynx. In A. Wray (Ed.), The Transition to
Language (pp. 21-45). Oxford, Oxford University Press.
Fitch, W.T. (2002b). The evolution of language comes of age. Trends in Cognitive
Sciences 6, 278-279.
Fitch, W.T. (2002c). The evolution of spoken language: A comparative approach. In J.L.
Hansen (Ed.), International Conference on Spoken Language Processing, (pp. 18). Denver, CO: University of Colorado Press.
Fitch, W.T. (2004). Kin selection and ‘mother tongues’: A neglected component in
language evolution. In D.K. Oller and U. Griebel (Eds.), Evolution of
Communication Systems: A comparative approach (pp. 275-296). Cambridge,
MIT Press.
Fitch, W.T. (2005). The evolution of language: A comparative review. Biology and
Philosophy 20, 193-230.
Fitch, W.T. (2007). Evolving meaning: The roles of kin selection, allomothering and
paternal care in language evolution. In C. Lyon, C. Nehaniv, and A. Cangelosi
(Eds.), Emergence of Communication and Language (pp. 29-51). New York,
Springer.
Fitch, W.T. (2010). The Evolution of Language. Cambridge, Cambridge University Press.
Fitch, W.T., Fritz, J.B. (2006). Rhesus macaques spontaneously perceive formants in
conspecific vocalizations. Journal of the Acoustical Society of America 120,
2132-2141.
336
Fitch, W.T., Hauser, M.D., Chomsky, N. (2005). The evolution of the language faculty:
Clarifications and implications. Cognition 97, 179-210.
Fitch, W.T., Reby, D. (2001). The descended larynx is not uniquely human. Proceedings
of the Royal Society London B 268, 1669-1675.
Fleischer, R.L., Price, P.B., Walker, R.M., Leakey L.S.B. (1965). Fission-track dating of
Bed I, Olduvai Gorge. Science 148, 72-74.
Floyer-Lea, A., Matthews, P.M. (2005). Distinguishable brain activation networks for
short- and long-term skill learning. Journal of Neurophysiology 94, 512-518.
Fodor, J. (1983). Modularity of Mind. Cambridge: MIT Press.
Fogassi, L., Gallese, V., Fadiga, L., Rizzolatti, G. (1998). Neurons responding to the
sight of goal directed hand/arm actions in the parietal area PF (7b) of the macaque
monkey. Society for Neuroscience 24:257.5 (Abstr.).
Foley, R., Lahr, M. (1997). Mode 3 technologies and the evolution of modern humans.
Cambridge Archaeological Journal 7, 3–36.
Frey, S.H., Vinton, D., Norlund, R., Grafton, S.T. (2005). Cortical topography of human
anterior intraparietal cortex active during visually guided grasping. Cognitive
Brain Research 23, 397-405.
Friederici, A.D. (2009). Pathways to language: Fiber tracts in the human brain. Trends in
Cognitive Sciences 13, 175-181.
Frost, G.T. (1980). Tool behavior and the origins of laterality. Journal of Human
Evolution 9, 447-459.
Fuster, J.M. (2000). Executive frontal functions. Experimental Brain Research 133, 6670.
337
Gabunia, L., Vekua, A., Lordkipanidze, D., Swisher, C.C. III, Ferring, R., Justus, A.,
Nioradze, M., Tvalchrelidze, M., Antón, S.C., Bosinski, G., Jöris, O., de Lumley,
M-A., Majsuradze, G., Mouskhelishvili, A. (2000). Earliest Pleistocene hominid
cranial remains from Dmanisi, Republic of Georgia: Taxonomy, geological
setting, and age. Science 288, 1019-1025.
Gagnon, L., Cooper, R.J., Yücel, M.A., Perdue, K.L., Greve, D.N., Boas, D.A. (2012).
Short separation channel location impacts the performance of short channel
regression in NIRS. NeuroImage 59, 2518-2528.
Gagnon, L., Perdue, K., Greve, D.N., Goldenholz, D., Kaskhedikar, G., Boas, D.A.
(2011). Improved recovery of the hemodynamic response in diffuse optical
imaging using short optode separations and state-space modeling. NeuroImage
56, 1362-1371.
Gallese, V., Fadiga, L., Fogassi, L., Rizzolatti, G. (1996). Action recognition in the
premotor cortex. Brain 119, 593–609.
Gallese, V., Fogassi, L., Fadiga, L., Rizzolatti, G. (2002). Action representation and the
inferior parietal lobule. In W. Prinz and B. Hommel (Eds.), Attention &
Performance XIX. Common Mechanisms in Perception and Action (pp. 247-266).
Oxford: Oxford University Press.
Garavan, H., Kelley, D., Rosen, A., Rao, S.M., Stein, E.A. (2000). Practice-related
functional activation changes in a working memory task. Microscopy Research
and Technique 51, 54-63.
338
Gathercole, S.E., Pickering, S.J., Ambridge, B., Wearing, H. (2004). The structure of
working memory from 4 to 15 years of age. Developmental Psychology 40, 177190.
Gentilucci, M., Benuzzi, F., Gangitano, M., Grimaldi, S. (2001). Grasp with hand and
mouth: A kinematic study on healthy subjects. Journal of Neurophysiology 86,
1685-1699.
Geribàs, N., Mosquera, M., Vergès, J.M. (2010). What novice knappers have to learn to
become expert stone toolmakers. Journal of Archaeological Science 37, 28572870.
Gernsbacher, M.A., Kaschak, M.P. (2003). Neuroimaging studies of language production
and comprehension. Annual Review of Psychology 54, 91-114.
Gibson, J. (1986). The Ecological Approach to Visual Perception. Hillsdale, NJ:
Lawrence Erlbaum Associates.
Gibson, K.R. (1991). Tools, language and intelligence: Evolutionary implications. Man
26, 255-264.
Gibson, K.R., Ingold, T., eds. (1993). Tools, Language and Cognition in Human
Evolution. Cambridge: Cambridge University Press.
Gil-da-Costa, R., Martin, A., Lopes, M.A., Muñoz, M., Fritz, J.B., Braun, A.R. (2006).
Species-specific calls activate homologs of Broca’s and Wernicke’s areas in the
macaque. Nature Neuroscience 9, 1064-1070.
Givón, T. (1995) Functionalism and grammar. Philadelphia, John Benjamins.
Givón, T., Malle B.F., eds. (2002). The Evolution of Language Out of Pre-Language.
Philadelphia, John Benjamins.
339
Gold, B.T., Balota, D.A., Jones, S.J., Powell, D.K., Smith, C.D., Andersen, A.H. (2006).
Dissociation of automatic and strategic lexical-semantics: Functional magnetic
resonance imaging evidence for differing roles of multiple frontotemporal
regions. Journal of Neuroscience 26, 6523–6532.
Goldberg, G. (1985). Supplementary motor area structure and function: Review and
hypotheses. The Behavioral and Brain Sciences 8, 567-616.
Goldman-Rakic, P.S. (1995). Cellular basis of working memory. Neuron 14, 477-485.
Goodall, J. (1964). Tool-using and aimed throwing in a community of free-living
chimpanzees. Nature 201, 1264-1266.
Goren-Inbar, N. (2011). Culture and cognition in the Acheulian industry: A case study
from Gesher Benot Yaʿaqov. Philosophical Transactions of the Royal Society
B 366, 1038-1049.
Gowlett, J.A.J. (1979). Complexities of cultural evidence in the Lower and Middle
Pleistocene. Nature 278, 14-17.
Gowlett, J.A.J. (1984). Mental abilities of early man: A look at some hard evidence. In R.
Foley (Ed.), Hominid Evolution and Community Ecology (pp. 167-192). Orlando:
Academic Press.
Grabowski, T.J., Damasio, H., Damasio, A.R. (1998). Premotor and prefrontal correlates
of category-related lexical retrieval. NeuroImage 7, 232-243.
Grafton, S.T. (2009). Embodied cognition and the simulation of action to understand
others. Annals of the New York Academy of Sciences 1156, 97–117.
340
Grafton S.T., Fadiga L., Arbib M.A., Rizzolatti G. (1997). Premotor cortex activation
during observation and naming of familiar tools. Neuroimage 6, 231-236.
Grafton, S.T., Woods, R.P., Tyszka, M. (1994). Functional imaging of procedural motor
learning: Relating cerebral blood flow with individual subject performance.
Human Brain Mapping 1, 221-234.
Greenberg, J.H. (1967). The science of linguistics. In M. Fried (Ed.), vol. 1: Readings in
Anthropology (pp. 347-363). New York: Thomas Y. Crowell.
Greenfield, P.M. (1991). Language, tools, and brain: The development and evolution of
hierarchically organized sequential behavior. Behavioral and Brain Sciences 14,
531-595.
Greicius, M.D., Supekar, K., Menon, V., Dougherty, R.F. (2009). Resting-state functional
connectivity reflects structural connectivity in the default mode network. Cerebral
Cortex 19, 72-78.
Grèzes, J., Armony, J.L., Rowe, J, Passingham, R.E. (2003). Activations related to
“mirror” and “canonical” neurones in the human brain: an fMRI study.
NeuroImage 18, 928–937.
Grèzes, J., Decety, J. (2001). Functional anatomy of execution, mental simulation,
observation, and verb generation of action: a meta-analysis. Human Brain
Mapping 12, 1–19.
Gross, C. (2010). Alfred Russell Wallace and the evolution of the human mind. The
Neuroscientist 16, 496-507.
Guilmet, G.M. (1977). The evolution of tool-using and tool-making behaviour. Man 12,
33-47.
341
Haeussinger, F.B., Heinzel, S., Hahn, T., Schecklmann, M., Ehlis, A.C., Fallgatter, A.J.
(2011). Simulation of near-infared light absorption considering individual head
and prefrontal cortex anatomy: Implications for optical neuroimaging. PloS One
6, e26377.
Haidle, M.N. (2010). Working-memory capacity and the evolution of modern cognitive
potential: Implications from animal and early human tool use. Current
Anthropology 51, S149-S166.
Haier, R.J., Siegel, B.V., MacLachlan, A., Soderling, E., Lottenberg, S., Buchsbaum,
M.S. (1992). Regional glucose metabolic changes after learning a complex
visuospatial/motor task: A positron emission tomographic study. Brain Research
570, 134-143.
Halford, G.S., Cowan, N., Andrews, G. (2007). Separating cognitive capacity from
knowledge: A new hypothesis. Trends in Cognitive Sciences 11, 236-242.
Hampson, E., Kimura, D. (1984). Hand movement asymmetries during verbal and
nonverbal tasks. Canadian Journal of Psychology 38, 102-125.
Harmand, S., Lewis, J.E., Feibel, C.S., Lepre, C.J., Prat, S., Lenoble, Boës, X., Quinn,
R.L., Brenet, M., Arroyo, A., Taylor, N., Clément, S., Daver, G., Brugal, J-P.,
Leakey, L., Mortlock, R.A., Wright, J.D., Lokorodi, S., Kirwa, C., Kent, D.V.,
Roche, H. (2015). 3.3-million-year-old stone tools from Lomekwi 3, West
Turkana, Kenya. Nature 521, 310-315.
Hart, H.C., Hall, D.A., Palmer, A.R. (2003). The sound-level-dependent growth in the
extent of fMRI activation in Heschl’s gyrus is different for low- and highfrequency tones. Hearing Research 179, 104-112.
342
Haslam, M., Hernandez-Aguilar, A., Ling, V., Carvalho, S., de la Torre, I., DeStefano,
A., et al. (2009). Primate archaeology. Nature 460, 339-344.
Hayakawa, T., Angata, T., Lewis, A.L., Mikkelsen, T.S., Varki, N.M., Varki, A. (2005).
A human-specific gene in microglia. Science 309, 1693-1693.
Hauser, M.D., Chomsky, N., Fitch, W.T. (2002). The faculty of language: What is it, who
has it, and how did it evolve? Science 298, 1569-1579.
Hauser, M.D., Fitch, W.T. (2003). What are the uniquely human components of the
language faculty? In M.H. Christansen and S. Kirby (Eds.), Language Evolution
(pp. 158-181). Oxford, Oxford University Press.
Hecht, E.E., Gutman, D.A., Khreisheh, N., Taylor, S.V., Kilner, J., Faisal, A.A., Bradley,
B.A., Chaminade, T., Stout, D. (2015). Acquisition of Paleolithic toolmaking
abilities involves structural remodeling to inferior frontoparietal regions. Brain
Structure and Function 220, 2315-2331.
Hecht, E.E., Murphy, L.E., Gutman, D.A., Votaw, J.R., Schuster, D.M., Preuss, T.M.,
Orban, G.A., Stout, D., Parr, L.A. (2013). Differences in neural activation for
object-directed grasping in chimpanzees and humans. The Journal of
Neuroscience 33, 14117-14134.
Heiser, M., Iacoboni, M., Maeda, F., Marcus, J., Mazziotta, J.C. (2003). The essential
role of Broca’s area in imitation. European Journal of Neuroscience 17, 1123–
1128.
Henrich, J., Heine, S.J., Norenzayan, A. (2010). The weirdest people in the world?
Behavioral and Brain Sciences 33, 61-135.
343
Henry, L. (2012). Development of Working Memory in Children. London, Sage
Publications Ltd.
Henshilwood, C.S., Marean, C.W. (2003). The origin of modern human behavior:
Critiques of the models and their test implications. Current Anthropology 44, 627651.
Hepp-Reymond, M-C., Husler, E.J., Maier, M.A., Qi, H.X. (1994). Force related
neuronal activity in two regions of the primate ventral premotor cortex. Canadian
Journal of Physiology and Pharmacology 72, 571–579.
Hermsdörfer, J., Goldenberg, G., Wachsmuth, C., Conrad, B., Ceballos-Baumann, A.O.,
Bartenstein, P., Schwaiger, M., Boecker, H. (2001). Cortical correlates of gesture
processing: Clues to the cerebral mechanisms underlying apraxia during the
imitation of meaningless gestures. NeuroImage 14, 149-161.
Hernandez-Aguilar, R.A., Moore, J., Pickering, T.R. (2007). Savanna chimpanzees use
tools to harvest the underground storage organs of plants. Proceedings of the
National Academy of Sciences 104, 19210–19213.
Hernández, A., Zainos, A., Romo, R. (2002). Temporal evolution of a decision-making
process in medial premotor cortex. Neuron 33, 959-972.
Hewes, G.W. (1973). An explicit formulation of the relationship between tool-using,
tool-making, and the emergence of language. Visible Language 7, 101-127.
Hewes, G.W. (1993). A history of speculation on the relation between tools and
language. In K.R. Gibson and T. Ingold (Eds.), Tools, Language, and Cognition
in Human Evolution (pp. 3-31). Cambridge: Cambridge University Press.
344
Hewes, G.W. (1996). A history of the study of language origins and the gestural primacy
hypothesis. In A. Lock and C.R. Peters (Eds.), Handbook of Human Symbolic
Evolution (pp. 571-595). Oxford, Clarendon Press.
Heyes, C. (2001). Causes and consequences of imitation. Trends in Cognitive Science 5,
253-261.
Heyes, C. (2010). Where do mirror neurons come from? Neuroscience and Biobehavioral
Reviews 34, 575-583.
Hickok, G. (2009). Eight problems for the mirror neuron theory of action understanding
in monkeys and humans. Journal of Cognitive Neuroscience 21, 1229-1243.
Hickok, G. (2014). The Myth of Mirror Neurons: The Real Neuroscience of
Communication and Cognition. New York: W. W. Nortion & Company.
Higuchi, S., Chaminade, T., Imamizu, H., Kawato, M. (2009). Shared neural correlates
for language and tool use in Broca's area. Neuroreport 20, 1376–1381.
Hill, K., Barton, M., Hurtado, A.M. (2009). The emergence of human uniqueness:
Characters underlying behavioral modernity. Evolutionary Anthropology 18, 187–
200.
Hill, K.R., Walker, R.S., Božičević, M., Eder, J., Headland, T., Hewlett, B., Hurtado,
A.M., Marlowe, F., Wiessner, P., Wood, B. (2011). Co-residence patterns in
hunter-gatherer societies show unique human social structure. Science 331, 12861289.
Hoard, R.J., Anglen, A.A. (2003). Lithic analysis. Plains Anthropologist 48, 36-50.
Hockett, C.F. (1960). The origin of speech. Scientific American 203, 88.
345
Hockett C.F., Ascher R. (1964). The human revolution. Current Anthropology 5, 135147.
Hodder, I. (1986). Reading the Past: Current Approaches to Interpretation in
Archaeology. Cambridge: Cambridge University Press.
Hogberg, A. (2002). Playing with flint and learning to knap. Paper presented at the 67th
Annual Meeting of the Society for American Archaeology, Denver, March 20-24.
Holliday, T.W. (2012). Body size, body shape, and the circumscription of the genus
Homo. Current Anthropology 53, S330-S345.
Holloway, R.L. (1966). Cranial capacity, neural reorganization, and hominid evolution: A
search for more suitable parameters. American Anthropologist 68, 103-121.
Holloway, R.L. (1969). Culture: A human domain. Current Anthropology 10, 395-412.
Holloway, R.L. (1972). New australopithecine endocast, SK 1585, from Swartkrans,
South Africa. American Journal of Physical Anthropology 37, 173-186.
Holloway, R.L. (1975). Early hominid endocasts: Volumes, morphology and significance
for hominid evolution. In R.H. Tuttle (Ed.), Primate Functional Morphology and
Evolution (pp. 393-415). Mouton, The Hague.
Holloway, R.L. (1980). Indonesian “Solo” (Ngandong) endocranial reconstructions:
Some preliminary observations and comparisons with Neandertal and Homo
erectus groups. American Journal of Physical Anthropology 53, 285-295.
Holloway, R.L. (1981). Culture, symbols, and human brain evolution: A synthesis.
Dialectical Anthropology 5, 287-303.
Holloway, R.L. (1983a). Cerebral brain endocast pattern of Australopithecus afarensis
hominid. Nature 303, 420-422.
346
Holloway R.L.(1983b). Human paleontological evidence relevant to language behavior.
Human Neurobiology 2, 105–114.
Holloway, R.L. (1984). The Taung endocast and the lunate sulcus: A rejection of the
hypothesis of its anterior position. American Journal of Physical Anthropology
64, 285-287.
Holloway, R.L., Broadfield, D.C., Yuan, M.S. (2004). Vol. 3: The Human Fossil Record:
Brain Endocasts, The Paleoneurological Evidence. Hoboken, NJ: Wiley.
Holloway, R.L., De La Costelareymondie, M.C. (1982). Brain endocast asymmetry in
pongids and hominids: Some preliminary findings on the paleontology of cerebral
dominance. American Journal of Physical Anthropology 58, 101-110.
Holloway, R.L., Kimbel, W.H. (1986). Endocast morphology of Hadar hominid AL 16228. Nature 321, 536-536.
Holloway, R.L., Sherwood, C.C., Hof, P.R., Rilling, J.K. (2009). Evolution of the brain
in humans—Paleoneurology. In M.D. Binder, N. Hirokawa, and U. Windhorst
(Eds.), Encyclopedia of Neuroscience (pp. 1326-1334). Berlin, Springer Berlin
Heidelberg.
Holper, L., Muehlemann, T., Scholkmann, F., Eng, K., Kiper, D., Wolf, M. (2010).
Testing the potential of a virtual reality neurorehabilitation system during
performance of observation, imagery and imitation of motor actions recorded by
wireless functional near-infrared spectroscopy (fNIRS). Journal of
Neuroengineering and Rehabilitation 7, 57.
347
Homae, F., Hashimoto, R., Nakajima, K., Miyashita, Y., Sakai, K.L. (2002). From
perception to sentence comprehension: The convergence of auditory and visual
information of language in the left inferior frontal cortex. NeuroImage 16, 883900.
Homan, R.W., Herman, J., Purdy, P. (1987). Cerebral location of international 10-20
system electrode placement. Electroencephalography and Clinical
Neurophysiology 66, 376-382.
Horwitz, B., Amunts, K., Bhattacharyy, R., Patkin, D., Jeffries, K., Zilles, K., Braun,
A.R. (2003). Activation of Broca’s area during the production of spoken and
signed language: A combined cytoarchitectonic mapping and PET analysis.
Neuropsychologia 41, 1868–1876.
Hoshi, E., Tanji, J. (2004). Area-selective neuronal activity in the dorsolateral prefrontal
cortex for information retrieval and action planning. Journal of Neurophysiology
91, 2707-2722.
Hoshi, Y. (2003). Functional near-infrared optical imaging: Utility and limitations in
human brain mapping. Pyschophysiology 40, 511-520.
Hoshi, Y. (2007). Functional near-infrared spectroscopy: Current status and future
prospects. Journal of Biomedical Optics 12, 062106.
Hovers, E., Belfer-Cohen, A. (2006). “Now you see it, now you don’t”—Modern human
behaviors in the Middle Paleolithic. In E. Hovers and S.L. Kuhn (Eds.),
Transitions before the Transition (pp. 295-304). New York, Springer.
Hughes, M. (1975). Egocentrism in Preschool Children. Ph.D. dissertation. Edinburgh
University.
348
Hund-Georgiadis, M., von Cramon, D.Y. (1999). Motor-learning-related changes in
piano players and non-musicians revealed by functional magnetic-resonance
signals. Experimental Brain Research 125, 417-425.
Huppert, T.J., Diamond, S.G., Franceschini, M.A., Boas, D.A. (2009). HomER: A review
of time-series analysis methods for near-infrared spectroscopy of the brain.
Applied Optics 48, D280-D298.
Hurford J. (1990). Nativist and functional explanations in language acquisition. In I.M.
Rica (Ed.), Logical issues in Language Acquisition (pp. 85-136). Dordrecht, Foris
Publications.
Hurford J. (2003). The language mosaic and its evolution. In M.H. Christiansen and S.
Kirby (Eds.), Language Evolution (pp. 38-57). Oxford, Oxford University Press.
Hyun, J-S., Luck, S.J. (2007). Visual working memory as the substrate for mental
rotation. Psychonomic Bulletin & Review14, 154-158.
Iacoboni, M. (2012). The human mirror neuron system and its role in imitation and
empathy. In F.B.M. De Waal and P.F. Ferrari (Eds.), The Primate Mind: Built to
Connect with Other Minds (pp. 32-47). Cambridge, MA, Harvard University
Press.
Iacoboni, M., Dapretto, M. (2006). The mirror neuron system and the consequences of its
dysfunction. Nature Reviews Neuroscience 7, 942-951.
Iacoboni, M., Molnar-Szakacs, I., Gallese, V., Buccino, G., Mazziotta, J.C., Rizzolatti, G.
(2005). Grasping the intentions of others with one's own mirror neuron system.
Plos Biology 3, e79.
349
Iacoboni, M., Woods, R., Brass, M., Bekkering, H., Mazziotta, J.C., Rizzolatti, G. (1999).
Cortical mechanisms of human imitation. Science 286, 2526-2529.
Inoue-Nakamura, N., Matsuzawa, T. (1997). Development of stone tool use by wild
chimpanzees (Pan troglodytes). Journal of Comparative Psychology 111, 159173.
Iovita, R., McPherron, S.P. (2011). The handaxe reloaded: A morphometric reassessment
of Acheulian and Middle Paleolithic handaxes. Journal of Human Evolution 61,
61-74.
Isaac, Glynn L. (1976). Stages of cultural elaboration in the Pleistocene: Possible
archaeological indicators of the development of language capabilities. Annals of
the New York Academy of Sciences 280, 275-288.
Jackendoff, R. (1999). Possible stages in the evolution of the language capacity. Trends
in Cognitive Sciences 3, 272-279.
Jasper, H.H. (1958). Report of Committee on Methods of Clinical Examination in
Electroencephalography. Electroencephalography and Clinical Neurophysiology
10, 370-375.
Jellema, T., Baker, C.I., Wicker, B., Perrett, D.I. (2000). Neural representation for the
perception of the intentionality of actions. Brain and Cognition 442, 280–302.
Jenkins, L.H., Brooks, D.J., Nixon, P.D., Frackowiak, R.S.J., Passingham, R.E. (1994).
Motor sequence learning: A study with positron emission tomography. Journal of
Neuroscience 14, 3775-3790.
Jöbsis, F.F. (1977). Noninvasive, infrared monitoring of cerebral and myocardial oxygen
sufficiency and circulatory parameters. Science 198, 1264-1267.
350
Johnson-Frey, S.H. (2004). The neural bases of complex tool use in humans. Trends in
Cognitive Sciences 8, 71-78.
Johnson-Frey, S.H., Newman-Norlund, R., Grafton, S.T. (2005). A distributed left
hemisphere network active during planning of everyday tool use skills. Cerebral
Cortex 15, 681-695.
Jones, P.R. (1980). Experimental butchery with modern stone tools and its relevance for
Palaeolithic archaeology. World Archaeology 12, 153-165.
Kaller, C.P., Rahm, B., Spreer, J., Weiller, C., Unterrainer, J.M. (2011). Dissociable
contributions of left and right dorsolateral prefrontal cortex in planning. Cerebral
Cortex 21, 307-317.
Kane, M.J., Hambrick, D.Z., Conway, A.R.A. (2005). Working memory and fluid
intelligence are strongly correlated constructs: Comment on Ackerman, Beier, and
Boyle (2005). Psychological Bulletin 131, 66-71.
Karnath, H-O. (2001). New insights into the functions of the superior temporal cortex.
Nature Reviews Neuroscience 2, 568-576.
Karni, A., Meyer, G., Jezzard, P., Adams, M.M., Turner, R., Ungerleider, L.G. (1995).
Functional MRI evidence for adult motor cortex plasticity during motor skill
learning. Nature 377, 155-158.
Keating, D. (1979). Adolescent thinking. In J. Adelson (Ed.), Handbook of Adolescent
Psychology (pp. 211-246). New York: Wiley.
Keith, K. (2006). Childhood learning and the distribution of knowledge in foraging
societies. Archaeological Papers of the American Anthropological Association
15, 27-40.
351
Kelly, A.M.C., Garavan, H. (2005). Human functional neuroimaging of brain changes
associated with practice. Cerebral Cortex 15, 1089-1102.
Kelley, W.M., Miezin, F.M., McDermott, K.B., Buckner, R.L., Raichle, M.E., Cohen,
N.J., Ollinger, J.M., Akbudak, E., Conturo, T.E., Snyder, A.Z. (1998).
Hemispheric specialization in human dorsal frontal cortex and medial temporal
lobe for verbal and nonverbal memory encoding. Neuron 20, 927-936.
Keysers, C., Kohler, E., Umiltà, M.A., Nanetti, L., Fogassi, L., Gallese, V. (2003).
Audiovisual mirror neurons and action recognition. Experimental Brain Research
153, 628-636.
Kilner, J.M., Neal, A., Weiskopf, N., Friston, K.J., Frith, C.D. (2009). Evidence of mirror
neurons in human inferior frontal gyrus. The Journal of Neuroscience 29, 1015310159.
Kimura, D. (1979). Neuromotor mechanisms in the evolution of human communication.
In H.D. Steklis and M.J. Raleigh (Eds.), Neurobiology of Social Communication
in Primates: An Evolutionary Perspective (pp. 197-220). New York: Academic
Press.
Kirby, S., Cornish, H., Smith, K. (2008). Cumulative cultural evolution in the laboratory:
An experimental approach to the origins of structure in human language.
Proceedings of the National Academy of Sciences 105, 10681-10686.
Kitahara-Frisch, J. (1978). Stone tools as indicators of linguistic abilities in early man.
Annals of the Japan Association for Philosophy of Science 5, 101-109.
Klein, R.G. (1995). Anatomy, behavior, and modern human origins. Journal of World
Prehistory 9, 167-198.
352
Klein, R.G. (2001). Southern Africa and modern human origins. Journal of
Anthropological Research 57, 1-16.
Klein, R.G. (2009). The Human Career: Human Biological and Cultural Origins.
Chicago: The University of Chicago Press.
Klingberg, T., Forssberg, H., Westerberg, H. (2002). Increased brain activity in frontal
and parietal cortex underlies the development of visuospatial working memory
capacity during childhood. Journal of Cognitive Neuroscience 14, 1-10.
Koechlin, E., Jubault, T. (2006). Broca’s area and the hierarchical organization of human
behavior. Neuron 50, 963-974.
Kohler, E., Keysers, C., Umiltà, M.A., Fogassi, L., Gallese, V., Rizzolatti, G. (2002).
Hearing sounds, understanding actions: Action representation in mirror neurons.
Science 297, 846-848.
Kojima, T., Onoe, H., Hikosaka, K., Tsutsui, K., Tsukada, H., Watanabe, M. (2009).
Default mode of brain activity demonstrated by positron emission tomography
imaging in awake monkeys: Higher rest-related than working memory-related
activity in medial cortical areas. The Journal of Neuroscience 29, 14463-14471.
Kolb, B., Whishaw, I.Q. (2009). Fundamentals of Human Neuropsychology. 6th edition.
New York: Worth Publishers.
Koller, J., Baumer, U., Mania, D. (2001). High-tech in the Middle Palaeolithic:
Neandertal-manufactured pitch identified. European Journal of Archaeology 4,
385-397.
353
Komeilipoor, N., Vicario, C.M., Daffertshofer, A., Cesari, P. (2014). Talking hands:
Tongue motor excitability during observation of hand gestures associated with
words. Frontiers in Human Neuroscience 8, 767.
Koops, K., McGrew, W.C., Matsuzawa, T. (2010). Do chimpanzees (Pan troglodytes)
use cleavers and anvils to fracture Treculia africana fruits? Preliminary data on a
new form of percussive technology. Primates 51, 175–178.
Koski, L., Iacoboni, M., Dubeau, M.C., Woods, R.P., Mazziotta, J.C. (2003). Modulation
of cortical activity during different imitative behaviors. Journal of
Neurophysiology 89, 460–71.
Kübler, A., Dixon, V., Garavan, H. (2006). Automaticity and reestablishment of
executive control—An fMRI study. Journal of Cognitive Neuroscience 18, 13311342.
Kurata, K., Tanji J. (1986). Premotor cortex neurons in macaques: Activity before distal
and proximal forelimb movements. Journal of Neuroscience 6, 403-411.
Lafayette Instrument (1998). The Minnesota Dexterity Test: Examiner’s Manual. Indiana.
Laird, A.R., Robinson, J.L., McMillan, K.M., Tordesillas-Gutierrez, D., Moran, S.T.,
Gonzales, S.M., Ray, K.L., Franklin, C., Glahn, D.C., Fox, P.T., Lancaster, J.L.
(2010). Comparison of the disparity between Talairach and MNI coordinates in
functional neuroimaging data: Validation of the Lancaster transform. Neuroimage
51, 677-683.
354
Lancaster, J.L., Tordesillas-Gutierrez, D., Martinez, M., Salinas, F., Evans, A., Zilles, K.,
Mazziotta, J.C., Fox, P.T. (2007). Bias between MNI and Talairach coordinates
analyzed using the ICBM-152 brain template. Human Brain Mapping 28, 11941205.
Langbroek, M. (2012). Trees and ladders: A critique of the theory of human cognitive
and behavioural evolution in Palaeolithic archaeology. Quaternary International
270, 4-14.
Leakey, L.S.B., Evernden, J.F., Curtis, C.H. (1961). Age of Bed I, Olduvai Gorge,
Tanganyika. Nature 191, 478-479.
Leff, D.R., Orihuela-Espina, F., Elwell, C.E., Athanasiou, T., Delpy, D.T., Darzi, A.W.,
Yang, G-Z. (2011). Assessment of the cerebral cortex during motor task
behaviours in adults: A systematic review of functional near infrared spectroscopy
(fNIRS) studies. NeuroImage 54, 2922-2936.
Leff, A.P., Schofield, T.M., Crinion, J.T., Seghier, M.L., Grogran, A., Green, D.W.,
Price, C.J. (2009). The left superior temporal gyrus is a shared substrate for
auditory short-term memory and speech comprehension: Evidence from 210
patients with stroke. Brain 132, 3401-3410.
Lehmann, C., Herdener, M., Schneider, P., Federspiel, A., Bach, D.R., Esposito, F., di
Salle, F., Scheffler, K., Kretz, R., Dierks, T., Seifritz, E. (2007). Dissociated
lateralization of transient and sustained blood oxygen level-dependent signal
components in human primary auditory cortex. NeuroImage 34, 1637-1642.
355
Lepre, C.J., Roche, H., Kent, D.V., Harmand, S., Quinn, R.L., Brugal, J.P., Texier, P.J.,
Lenoble, A., Feibel, C.S. (2011). An earlier origin for the Acheulian. Nature 477,
82-85.
Leroi-Gourhan, A. (1993[1964]). Gesture and speech. A.B. Berger, trans. Cambridge:
The MIT Press.
Lewis, C.M., Baldassarre, A., Committeri, G., Romani, G.L., Corbetta, M. (2009).
Learning sculpts the spontaneous activity of the resting human brain. Proceedings
of the National Academy of Sciences 106, 17558-17563.
Li, A., Zhang, Q., Culver, J.P., Miller, E.L., Boas, D. (2004). Reconstructing
chromosphere concentration images directly by continuous-wave diffuse optical
tomography. Optics Letters 29, 256-258.
Lieberman, D.E., McCarthy, R.C., Hiiemae, K.M., Palmer, J.B. (2001). Ontogeny of
postnatal hyoid and larynx descent in humans. Archives of Oral Biology 46, 117128.
Lieberman, P. (1975). On the Origins of Language. New York: Macmillan.
Lieberman, P. (1984). The Biology and Evolution of Language. Cambridge: Harvard
University Press.
Lieberman, P. (1990). Rule governed processes and brains: Speech, syntax, and tool
making. Paper presented at the Wenner-Gren Foundation for Anthropological
Research International Symposium 110, Cascais, Portugal, March 16.
Lieberman, P. (2003). Motor control, speech, and the evolution of human language. In
M.H. Christiansen and S. Kirby (Eds.), Language Evolution (pp. 255-271).
Oxford, Oxford University Press.
356
Lieberman, P., Crelin, E.S. (1971). On the speech of Neanderthal man. Linguistic Inquiry
2, 203–222.
Lingnau, A., Gesierich, B., Caramazza, A. (2009). Asymmetrick fMRI adaptation reveals
no evidence for mirror neurons in humans. Proceedings of the National Academy
of Sciences 106, 9925-9930.
Lombard, M. (2006). Direct evidence for the use of ochre in the hafting technology of
Middle Stone Age tools from Sibudu Cave. Southern African Humanities 18, 5767.
London, E.D., Ernst, M., Grant, S., Bonson, K., Weinstein, A. (2000). Orbitofrontal
cortex and human drug abuse: Functional imaging. Cerebral Cortex 10, 334-342.
Luciana, M., Conklin, H.M., Hooper, C.J., Yarger, R.S. (2005). The development of
nonverbal working memory and executive control processes in adolescents. Child
Development 76, 697-712.
Luciana, M., Wright, M.J., Smith, G.A., Geffen, G.M., Geffen, L.B., Martin, N.G.
(2001). Genetic covariance among measures of information processing speed,
working memory and IQ. Behavior Genetics 31, 581-592.
Lurito, J.T., Karaken, D.A., Lowe, M.J., Chen, S.H.A., Matthews, V.P. (2000).
Comparisons of rhyming and word generation with fMRI. Human Brain Mapping
10, 99-106.
Lycett, S. J., Gowlett, J.A.J. (2008). On questions surrounding the Acheulean ‘tradition.’
World Archaeology 40, 295-315.
357
Maguire, E.A., Gadian, D.G., Johnsrude, I.S., Good, C.D., Ashburner, J., Frackowiak,
R.S.J., et al. (2000). Navigation-related structural change in the hippocampi of
taxi drivers. Proceedings of the National Academy of Sciences 97, 4398-4403.
Mahaney, R.A. (2014a). Exploring the complexity and structure of Acheulean
stoneknapping in relation to natural language. PaleoAnthropology 2014, 586-606.
Mahaney, R.A. (2014b). Lithic analysis as a cognitive science: A framework. Lithic
Technology 39, 173-189.
Mahaney, R.A. (2016). Cognition and Planning in Paleolithic Technology: Studies in
Experimental Archaeology. Ph.D. dissertation. Indiana University, Bloomington.
Malafouris, L. (2004). The cognitive basis of material engagement: Where brain, body
and culture conflate. In E. DeMarrais, C. Gosden, and C. Renfrew (Eds.),
Rethinking Materiality: The Engagement of Mind with the Material World (pp.
53-62). Cambridge: McDonald Institute for Archaeological Research.
Malafouris, L. (2008). Beads for a plastic mind: The ‘blind man’s stick’ (BMS)
hypothesis and the active nature of material culture. Cambridge Archaeological
Journal 18, 401-414.
Malafouris, L. (2009). “Neuroarchaeology”: Exploring the links between neural and
cultural plasticity. In J.Y. Chiao (Ed.), vol. 178: Progress in Brain Research (pp.
253-261). Elsevier.
Malafouris, L. (2013). How Things Shape the Mind: A Theory of Material Engagement.
Cambridge: The MIT Press.
358
Marchant, L.F., McGrew, W.C. (2005). Percussive technology: Chimpanzee baobab
smashing and the evolutionary modeling of hominin knapping. In V. Roux and B.
Bril (Eds.), Stone Knapping: The Necessary Conditions for a Uniquely Hominin
Behaviour (pp. 341-350). Cambridge: McDonald Institute Monographs.
Markov, N.T., Ercsey-Ravasz, M., Van Essen, D.C., Knoblauch, K., Toroczkai, Z.,
Kennedy, H. (2013). Cortical high-density counterstream architectures. Science
342, 578-591.
Marsh, E.J., Ferguson, J.R. (2010). Introduction. In J.R. Ferguson (Ed.), Designing
Experimental Research in Archaeology: Examining Technology through
Production and Use (pp. 1-12). Boulder: University Press of Colorado.
Martin, A., Wiggs, C.L., Ungerleider, L.G., Haxby, J.V. (1996). Neural correlates of
category-specific knowledge. Nature 379, 649-652.
Martin, A., Haxby, J.V., Lalonde, F.M., Wiggs, C.L., Ungerleider, L.G. (1995). Discrete
cortical regions associated with knowledge of color and knowledge of action.
Science 270, 102-105.
Martínez, I., Rosa, M., Quam, R., Jarabo, P., Lorenzo, C., Bonmati, A., GómezOlivencia, A., Gracia, A., Arsuaga, J.L. (2013). Communicative capacities in
Middle Pleistocene humans from the Sierra de Atapuerca in Spain. Quaternary
International 295, 94-101.
Martín-Loeches, M. (2010). Uses and abuses of the Enhanced-Working-Memory
Hypothesis in explaining modern thinking. Current Anthropology 51, S67-S75.
Masataka, N., Perlovsky, L., Hiraki, K. (2015). Near-infrared spectroscopy (NIRS) in
functional research of prefrontal cortex. Frontiers in Human Neuroscience 9, 274.
359
Matsuzawa, T. (1994). Field experiments on the use of stone tools by chimpanzees in the
wild. In R.W. Wrangham, W.C. McGrew, F.B.M. De Waal, and P.G. Heltne
(Eds.), Chimpanzee Cultures (pp.351-370). Cambridge, MA: Harvard University
Press.
McBrearty, S., Brooks, A.S. (2000). The revolution that wasn’t: A new interpretation of
the origin of modern human behavior. Journal of Human Evolution 39, 453-563.
McGrew, W.C. (1992). Chimpanzee material culture: implications for human evolution.
Cambridge: Cambridge University Press.
McGrew, W.C. (1993). The intelligent use of tools: Twenty propositions. In K.R. Gibson
and T. Ingold (Eds.), Tools, Language and Cognition in Human Evolution (pp.
151-170). Cambridge: Cambridge University Press.
McNabb, J., Binyon, F., Hazelwood, L. (2004). The large cutting tools from the South
African Acheulean and the question of social traditions. Current Anthropology 45,
653-677.
McPherron, S.P. (2000). Handaxes as a measure of the mental capabilities of early
hominids. Journal of Archaeological Science 27, 655–663.
M’Dougall, W. (1925). Mental Evolution. In Evolution in the Light of Modern
Knowledge: A Collective Work. London: Blackie and Son.
Meister, I.G., Boroojerdi, B., Foltys, H., Sparing, R., Huber, W., Töpper, R. (2003).
Motor cortex hand area and speech: Implications for the development of
language. Neuropsychologia 41, 401-406.
Merleau-Ponty, M. (1962). Phenomenology of Perception. London: Routledge.
360
Metropolis, N., Ulam, S. (1949). The Monte Carlo method. Journal of the American
Statistical Association 44, 335-341.
Meyer, M., Arsuaga, J-L., de Filippo, C., Nagel, S., Aximu-Petri, A., Nickel, B.,
Martínez, I., Gracia, A., Bermúdez de Castro, J.M., Carbonell, E., Viola, B.,
Kelso, J., Prüfer, K., Pääbo, S. (2016). Nuclear DNA sequences from the Middle
Pleistocene Sima de los Huesos hominins. Nature 531, 504-507.
Mgeladze, A., Lordkipanidze, D., Moncel, M-H., Despriee, J., Chagelishvili, R.,
Nioradze, M., Nioradze, G. (2011). Hominin occupations at the Dmanisi site,
Georgia, Southern Caucasus: Raw materials and technical behaviours of Europe’s
first hominins. Journal of Human Evolution 60, 571-596.
Mikels-Carrasco, J. (2012). Sherwood Washburn’s new physical anthropology: Rejecting
the “religion of taxonomy.” History and Philosophy of the Life Sciences 34, 79102.
Miller, G.A., Galanter, E., Pribram, K.H. (1960). Plans and the Structure of Behavior.
New York: Holt.
Mirz, F., Ovesen, T., Ishizu, K., Johannsen, P., Madsen, S., Gjedde, A., Pedersen, C.B.
(1999). Stimulus-dependent central processing of auditory stimuli: A PET study.
Scandinavian Audiology 28, 161-169.
Mithen, S. (1996). The Prehistory of the Mind: The Cognitive Origins of Art, Religion
and Science. London: Thames and Hudson.
361
Mithen, S. (1999). Imitation and cultural change: a view from the Stone Age, with
specific reference to the manufacture of handaxes. In H.O. Box and K.R. Gibson
(Eds.), Mammalian Social Learning: Comparative and Ecological Perspectives
(pp. 389-413). Cambridge, MA: Cambridge University Press.
Mithen, S. (2005). The Singing Neanderthals: The Origins of Music, Language, Mind
and Body. London: Weidenfeld & Nicolson.
Mohr, J.P., Pessin, M.S., Finkelstein, S., Funkenstein, H.H., Duncan, G.W., Davis, K.R.
(1978). Broca’s aphasia: pathological and clinical. Neurology 28, 311-324.
Molenberghs, P., Cunnington, R., Mattingley, J.B. (2012). Brain regions with mirror
properties: A meta-analysis of 125 human fMRI studies. Neuroscience &
Biobehavioral Reviews 36, 341-349.
Molinari, M., Leggio, M.G., Solida, A., Ciorra, R., Misciagna, S., Silveri, M.C.,
Pestrosini, L. (1997). Cerebellum and procedural learning: Evidence from focal
cerebellar lesions. Brain 120, 1753-1762.
Molnar-Szakacs, I., Iacoboni, M., Koski, L., Mazziotta, J.C. (2005). Functional
segregation within pars opercularis of the inferior frontal gyrus: Evidence from
fMRI studies of imitation and action observation. Cerebral Cortex 15, 986-994.
Molnar-Szakacs, I., Kaplan, J., Greenfield, P.M., Iacoboni, M. (2006).Observing
complex action sequences: The role of the fronto-parietal mirror neuron system.
NeuroImage 33, 923–935.
Montagu, A. (1976). Toolmaking, hunting, and the origin of language. Annals of the New
York Academy of Sciences 280, 266-274.
362
Morgan, T.J.H., Uomini, N.T., Rendell, L.E., Chouinard-Thuly, L., Street, S.E., Lewis,
H.M., et al. (2015). Experimental evidence for the co-evolution of hominin toolmaking teaching and language. Nature Communications 6, 1-8.
Moore, M.W. (2010). “Grammars of action” and stone flaking design space. In A.
Nowell and I. Davidson (Eds.), Stone Tools and the Evolution of Human
Cognition (pp. 13-44). Boulder: University Press of Colorado.
Moore, M.W. (2011). The design space of stone flaking: Implications for cognitive
evolution. World Archaeology 43, 702-715.
Morris, H.H., Lüders, H., Lesser, R.P., Dinner, D.S., Klem, G.H. (1986). The value of
closely spaced scalp electrodes in the localization of epileptiform foci: A study of
26 patients with complex partial seizures. Electroencephalography and Clinical
Neurophysiology 63, 107-111.
Movius, H. (1948). The Lower Paleolithic cultures of Southern and Eastern Asia.
Transactions of the American Philosophical Society 38, 329-420.
Mukamel, R., Ekstrom, A.D., Kaplan J., Iacoboni, M., Fried, I. (2010). Single-neuron
responses in humans during execution and observation of actions. Current
Biology 20, 750-756.
Mulcahy, N.J., Call, J. (2006). Apes save tools for future use. Science 312, 1038-1040.
Narumoto, J., Yamada, H., Iidaka, T., Sadato, N., Fukui, K., Itoh, H., Yonekura, Y.
(2000). Brain regions involved in verbal or non-verbal aspects of facial emotion
recognition. Brain Imaging 11, 2571-2576.
363
Newmeyer, F.J. (2003). What can the field of linguistics tell us about the origins of
language? In M.H Christiansen and S. Kirby (Eds.), Language Evolution (pp. 5876). Oxford, Oxford University Press.
Nishida, T. (1970). Social behavior and relationship among wild chimpanzees of the
Mahali Mountains. Primates 11, 47-87.
Nishida, T. (1980). The leaf-clipping display: A newly-discovered expressive gesture in
wild chimpanzees. Journal of Human Evolution 9, 117-128.
Nishitani, N., Hari, R. (2002). Viewing lip forms: Cortical dynamics. Neuron 36, 1211–
1220.
Noble, J., Di Paolo, E.A., Bullock, S. (2002). Adaptive factors in the evolution of
signaling systems. In A. Cangelosi and D. Parisi (Eds.), Simulating the Evolution
of Language (pp. 53-78). London, Springer-Verlag.
Noble, W., Davidson, I. (1996). Human Evolution, Language, and Mind: A
Psychological and Archaeological Inquiry. Cambridge: Cambridge University
Press.
Noiré, L. (1877). Der Ursprung der Sprache. Berlag von Bictor v. Zabern.
Noiré, L. (1880). Das Werkzeug und seine Bedeutung für die Entwicklungsgeschichte der
Menschheit. Mainz.
Norton, C.J., Bae, K., Harris, J.W.K., Lee, H. (2006). Middle Pleistocene handaxes from
the Korean Peninsula. Journal of Human Evolution 51, 527–536.
Nowell, A. (2010). Working memory and the speed of life. Current Anthropology 51,
S121-S133.
364
Nowell, A., Davidson, I. (2010). Stone Tools and the Evolution of Human Cognition.
Boulder: University Press of Colorado.
Nowell A., White M. (2010). Growing up in the Middle Pleistocene: Life history
strategies and their relationship to Acheulian industries. In A. Nowell and I.
Davidson (Eds.), Stone Tools and the Evolution of Human Cognition (pp. 67-81).
Boulder, CO, University Press of Colorado.
Oakley, Kenneth P. (1969). Man the skilled tool-maker. Antiquity 43, 222-224.
Ohnuma, K., Aoki, K., Akazawa, T. (1997). Transmission of tool-making through verbal
and nonverbal communication: Preliminary experiments in Levallois flake
production. Anthropological Science 105, 159-168.
Oldfield, R.C. (1971). The assessment and analysis of handedness: The Edinburgh
inventory. Neuropsychologia 9, 97-113.
Olesen, P.J., Westerberg, H., Klingberg, T. (2003). Increased prefrontal and parietal
activity after training of working memory. Nature Neuroscience 7, 75-79.
Orban, G.A., Claeys, K., Nelissen, K., Smans, R., Sunaert, S., Todd, J.T., Wardak, C.,
Durand, J-B., Vanduffel, W. (2006). Mapping the parietal cortex of human and
non-human primates. Neuropsychologia 44, 2647-2667.
Orban, G.A., Van Essen, D., Vanduffel, W. (2004). Comparative mapping of higher
visual areas in monkeys and humans. Trends in Cognitive Sciences 8, 315-324.
Osvath, M. (2009). Spontaneous planning for future stone throwing by a male
chimpanzee. Current Biology 19, R190-R191.
Osvath, M., Karvonen, E. (2012). Spontaneous innovation for future deception in a male
chimpanzee. PLoS ONE 7, e36782.
365
Osvath, M., Osvath, H. (2008). Chimpanzee and orangutan forethought: Self-control and
pre-experience in the face of future tool use. Animal Cognition 11, 661-674.
Panger, M.A., Brooks, A.S., Richmond, B.G., Wood, B. (2002). Older than the Oldowan?
Rethinking the emergence of hominin tool use. Evolutionary Anthropology 11,
235-245.
Park, R.W. (2006). Growing up north: Exploring the archaeology of childhood in the
Thule and Dorset cultures of Arctic Canada. Archaeological Papers of the
American Anthropological Association 15, 53-64.
Parker, S.T. (1973). Piaget’s sensorimotor series in an infant macaque: The organization
of non-stereotyped behavior in the evolution of intelligence. Ph.D. dissertation.
University of California, Berkeley.
Parker, S.T. (2015). Re-evaluating the extractive foraging hypothesis. New Ideas in
Psychology 37, 1-12.
Parker, S.T., Gibson, K.R. (1977). Object manipulation, tool use, and sensorimotor
intelligence as feeding adaptations in cebus monkeys and great apes. Journal of
Human Evolution 6, 623-641.
Parker, S.T., Gibson, K.R. (1979). A developmental model for the evolution of language
and intelligence in early hominids. Behavioral and Brain Sciences 2, 367-408.
Parker, S.T., Milbrath, C. (1993). Higher intelligence, propositional language, and culture
as adaptations for planning. In K.R. Gibson and T. Ingold (Eds.), Tools,
Language, and Cognition in Human Evolution (pp. 300-313). Cambridge:
Cambridge University Press.
366
Pascual-Leone, A., Amedi, A., Fregni, F., Merabet, L.B. (2005). The plastic human brain
cortex. Annual Review of Neuroscience 28, 377-401.
Passingham, R. (1982). The Human Primate. San Francisco: W.H. Freeman.
Patterson, K., Sollberger, J.P. (1986). Comments on Toth’s right-handedness study.
Lithic Technology 15, 109-111.
Patterson, R.D., Uppenkamp, S., Johnsrude, I.S., Griffiths, T.D. (2002). The processing
of temporal pitch and melody information in auditory cortex. Neuron 36, 767-776.
Paus, T., Collins, D.L., Evans, A.C., Leonard, G., Pike, B., Zijdenbos, A. (2001).
Maturation of white matter in the human brain: A review of magnetic resonance
studies. Brain Research Bulletin 54, 255-266.
Peeters, R., Simone, L., Nelissen, K., Fabbri-Destro, M., Vanduffel, W., Rizzolatti, G.,
Orban, G.A. (2009). The representation of tool use in humans and monkeys:
Common and uniquely human features. Journal of Neuroscience 29, 1152311539.
Pelegrin, J. (2009). Cognition and the emergence of language: A contribution from lithic
technology. In S.A. de Beaune, F.L. Coolidge, and T. Wynn (Eds.), Cognitive
Archaeology and Human Evolution (pp. 95-108). Cambridge: Cambridge
University Press.
Perrett, D.I., Harries, M.H., Bevan, R., Thomas, S., Benson, P.J., Mistlin, A.J., Chitty,
A.J., Hietanen, J.K., Ortega, J.E. (1989). Frameworks of analysis for the neural
representation of animate objects and actions. The Journal of Experimental
Biology 146, 87-113.
367
Perrett, D.I., Harries, M.H., Mistlin, A.J., Hietanen, J.K., Benson, P.J., Bevan, R.,
Thomas, S., Oram, M.W., Ortega, J., Brierley, K. (1990). Social signals analyzed
at the single cell level: Someone is looking at me, something touched me,
something moved! International Journal of Comparative Psychology 4, 25-55.
Pessoa, L., Gutierrez, E., Bandettini, P.A., Ungerleider, L.G. (2002). Neural correlates of
visual working memory: fMRI amplitude predicts task performance. Neuron 35,
975-987.
Pessoa, L. Ungerleider, L.G. (2004). Neural correlates of change detection and change
blindness in a working memory task. Cerebral Cortex 14, 511-520.
Petrides, M. (2000). The role of the mid-dorsolateral prefrontal cortex in working
memory. Experimental Brain Research 133, 44-54.
Petrides, M., Pandya, D.N. (2002). Comparative cytoarchitectonic analysis of the human
and the macaque ventrolateral prefrontal cortex and corticocortical connection
patterns in the monkey. European Journal of Neuroscience 16, 291-310.
Petrides, M., Pandya, D.N. (2009). Distinct parietal and temporal pathways to the
homologues of Broca’s area in the monkey. PLoS Biology 7, e1000170.
Phillipson, L. (1997). Edge modification as an indicator of function and handedness of
Acheulian handaxes from Kariandusi, Kenya. Lithic Technology 22, 171-183.
Piaget, J. (1952[1936]). Origins of Intelligence in Children. Margaret Cook, trans. New
York: International Universities Press.
Piaget, J. (1954[1937]). The Construction of Reality in the Child. Margaret Cook, trans.
New York: Basic Books.
368
Piaget, J. (1964). Part I: Cognitive development in children: Piaget development and
learning. Journal of Research in Science Teaching 2, 176-186.
Piaget, J., Inhelder, B. (1958). The Growth of Logical Thinking from Childhood to
Adolescence. New York: Basic Books.
Piaget, J., Inhelder, B. (1971). Mental Imagery in the Child. New York: Basic Books.
Pinker, S. (1994). The Language Instinct: How the Mind Creates Language. New York,
HarperCollin.
Pinker, S. (2003). Language as an adaptation to the cognitive niche. In M.H. Christiansen
and S. Kirby (Eds.), Language Evolution (pp. 16-37). Cambridge, Cambridge
University Press.
Pinker, S. (2010). The cognitive niche: Coevolution of intelligence, sociality, and
language. Proceedings of the National Academy of Sciences 107, 8993-8999.
Pinker, S. (2011). The cognitive revolution. Harvard University.
https://www.youtube.com/watch?v=AeoyzqmyWug, accessed July 24, 2015.
Pinker, S., Bloom, P. (1990). Natural language and natural selection. Behavioral and
Brain Sciences 13, 707-784.
Pobiner, B. (1999). The use of stone tools to determine handedness in hominids. Current
Anthropology 40, 90-92.
Poeppel, D., Hickok, G. (2004). Towards a new functional anatomy of language.
Cognition 92, 1-12.
369
Polderman, T.J.C., Gosso, M.F., Posthuma, D., Van Beusterveldt, T.C.E.M., Heutink, P.,
Verhulst, F.C., Boomsma, D.I. (2006). A longitudinal twin study on IQ, executive
function, and attention problems during childhood and early adolescence. Acta
Neurologica Belgica 106, 191-207.
Pontzer, H., Brown, M.H., Raichlen, D.A., Dunsworth, H., Hare, B., Walker, K., Luke,
A., Dugas, L.R., Durazo-Arvizu, R., Schoeller, D., Plange-Rhule, J., Bovet, P.,
Forrester, T.E., Lambert, E.V., Thompson, M.E., Shumaker, R.W., Ross, S.R.
(2016). Metabolic acceleration and the evolution of human brain size and life
history. Nature 533, 390-392.
Pontzer, H., Raichlen, D.A., Rodman, P.S. (2014). Bipedal and quadrupedal locomotion
in chimpanzees. Journal of Human Evolution 66, 64-82.
Porr, M. (2005). The making of the biface and the making of the individual. In C.
Gamble and M. Porr (Eds.), The Hominid Individual in Context: Archaeological
Investigations of Lower and Middle Palaeolithic Landscapes, Locales and
Artefacts (pp. 68-80). Abingdon, Oxon: Psychology Press.
Posthuma, D., Baaré, W.F.C., Pol, H.E.H., Kahn, R.S., Boomsma, D.I., de Geus, E.J.C.
(2003). Genetic correlations between brain volumes and the WAIS-III dimensions
of verbal comprehension, working memory, perceptual organization, and
processing speed. Twin Research and Human Genetics 6, 131-139.
Potts, R. (1998). Environmental hypotheses of hominin evolution. American Journal of
Physical Anthropology 41, 93-136.
Preuschoft, S., van Hooff, J. (1995). Homologizing primate facial displays: A critical
review of methods. Folia Primatologica 65, 121–137.
370
Pruetz, J., Bertolani, P. (2007). Savanna chimpanzees, Pan troglodytes verus, hunt with
tools. Current Biology 17, 412–417.
Pulvermüller, F., Fadiga, L. (2010). Active perception: sensorimotor circuits as a cortical
basis for language. Nature Reviews Neuroscience 11, 351–360.
Putt, S.S. (2015). The origins of stone tool reduction and the transition to knapping: An
experimental approach. Journal of Archaeological Science: Reports 2, 51-60.
Putt, S.S., Wijeakumar, S. Franciscus, R.G., Spencer, J.P. (2014a). Flintknapping on the
brain: Exploring the verbal and nonverbal learning mechanisms involved in stone
tool manufacture with fNIRS. Paper presented at the 113th Annual Meeting of the
American Anthropological Association, Washington, D.C., December 2-7.
Putt, S.S., Woods, A.D., Franciscus R.G. (2014b). The role of verbal interaction during
experimental bifacial stone tool manufacture. Lithic Technology 39, 96-112.
Quam, R.M., de Ruiter, D.J., Masali, M., Arsuaga, J-L., Martínez, I., Moggi-Cecchi, J.
(2013). Early hominin auditory ossicles from South Africa. Proceedings of the
National Academy of Sciences 110, 8847-8851.
Quaresima, V., Bisconti, S., Ferrari, M. (2012). A brief review on the use of functional
near-infrared spectroscopy (fNIRS) for language imaging studies in human
newborns and adults. Brain & Language 121, 79-89.
Race, E.A., Shanker, S., Wagner, A.D. (2009). Neural priming in human frontal cortex:
Multiple forms of learning reduce demands on the prefrontal executive system.
Journal of Cognitive Neuroscience 21, 1766-1781.
371
Ramachandran, V.S. (2000). Mirror neurons and imitation learning as the driving force
behind “the great leap forward” in human evolution. J. Brockman (Ed.), Edge.
http://edge.org/3rd_culture/ramachandran/ramachandran_index.html, accessed
August 28, 2015.
Read, D.W. (2008). Working memory: A cognitive limit to non-human primate recursive
thinking prior to hominid evolution. Evolutionary Psychology 6, 676-714.
Reader, S.M., Laland, K.N. (2002). Social intelligence, innovation, and enhanced brain
size in primates. Proceedings of the National Academy of Sciences 99, 4436-4441.
Reilly, E.L. (2005). EEG recording and operation of the apparatus. In E. Niedermeyer
and F.H. Lopes da Silva (Eds.), Electroencephalography (pp. 139-159).
Philadelphia: Lippincott Williams & Wilkins.
Renfrew, A.C. (1983). Divided we stand: Aspects of archaeology and information.
American Antiquity 48, 3-16.
Renfrew, A.C. (1994). Towards a cognitive archaeology. In C. Renfrew and E. Zubrow
(Eds.), The Ancient Mind: Elements of Cognitive Archaeology (pp. 3-12).
Cambridge: Cambridge University Press.
Renfrew, A.C. (1996). The sapient behaviour paradox. In P. Mellars and K.R. Gibson
(Eds.), Modelling the Early Human Mind (pp. 11-14). Cambridge: McDonald
Institute Monographs.
Reynolds, P.C. (1976). Language and skilled activity. Annals New York Academy of
Sciences 280, 150-166.
Reznick, J.S., Morrow, J.D., Goldman, B.D., Snyder, J. (2004). The onset of working
memory in infants. Infancy 6, 145-154.
372
Ridderinkhof, K.R., van den Wildenberg, W.P.M., Segalowitz, S.J., Carter, C.S. (2004).
Neurocognitive mechanisms of cognitive control: The role of prefrontal cortex in
action selection, response inhibition, performance monitoring, and reward based
learning. Brain Cognition 56, 129–140.
Riggs, K.J., McTaggart, J., Simpson, A., Freeman, R.P.J. (2006). Changes in the capacity
of visual working memory in 5- to 10-year-olds. Journal of Experimental Child
Psychology 95, 18-26.
Rightmire, G.P. (1979). Cranial remains of Homo erectus from Beds II and IV, Olduvai
Gorge, Tanzania. American Journal of Physical Anthropology 51, 99-116.
Rilling, J.K., Barks, S.K., Parr, L.A., Preuss, T.M., Faber, T.L., Pagnoni, G., Bremner,
J.D., Votaw, J.R. (2007). A comparison of resting-state brain activity in humans
and chimpanzees. Proceedings of the National Academy of Sciences 104, 1714617151.
Rilling, J.K., Seligman, R.A. (2002). A quantitative morphometric comparative analysis
of the primate temporal lobe. Journal of Human Evolution 42, 505-533.
Rizzolatti, G., Arbib, M.A. (1998). Language within our grasp. Trends in Neuroscience
21,188–194.
Rizzolatti, G., Camarda, R., Fogassi, L., Gentillucci, M. (1988). Functional organization
of inferior area 6 in the macaque monkey. II. Area F5 and the control of distal
movements. Experimental Brain Research 71, 491–507.
Rizzolatti. G., Craighero, L. (2004). The mirror-neuron system. Annual Reviews of
Neuroscience 27, 169-192.
373
Rizzolatti, G., Scandolara, C., Gentilucci, M., Camarda, R. (1981). Response properties
and behavioral modulation of ‘mouth’ neurons of the postarcuate cortex (area 6)
in macaque monkeys. Brain Research 225, 421-424.
Rochat, M.J., Caruana, F., Jezzini, A., Escola, L., Intskirveli, I., Grammont, F., Gallese,
V., Rizzolatti, G., Umiltà, M.A. (2010). Responses of mirror neurons in area F5 to
hand and tool grasping observation. Experimental Brain Research 204, 605-616.
Roe, D.A. (1968). British Lower and Middle Paleolithic handaxe groups. Proceedings of
the Prehistoric Society 34, 1-82.
Roffmann, I., Savage-Rumbaugh, S., Rubert-Pugh, E., Ronen, A., Nevo, E. (2012). Stone
tool production and utilization by bonobo-chimpanzees (Pan paniscus).
Proceedings of the National Academy of Sciences 109, 14500-14503.
Rogers, R.D., Owen, A.M., Middleton, H.C., Williams, E.J., Pickard, J.D., Sahakian,
B.J., Robbins, T.W. (1999). Choosing between small, likely rewards and large,
unlikely rewards activates inferior and orbital prefrontal cortex. Journal of
Neuroscience 19, 9029-9038.
Rolland, N. (1995). Levallois technique emergence: Single or multiple? A review of the
Euro-African record. In H.L. Dibble and O. Bar-Yosef (Eds.), The definition and
interpretation of Levallois technology (pp. 333-359). Madison: Prehistory Press.
Rowlands, M. (2010). The New Science of the Mind: From Extended Mind to Embodied
Phenomenology. Cambridge: The MIT Press.
Ruck, L, Broadfield, D.C., Brown, C.T. (2015). Determining hominid handedness in
lithic debitage: A review of current methodologies. Lithic Technology 40, 171188.
374
Rugg, G., Mullane, M. (2001). Inferring handedness from lithic evidence. Laterality 6,
247-259.
Sakai, T., Mikami, A., Tomonaga, M., Matsui, M., Suzuki, J., Hamada, Y., Tanaka, M.,
Miyabe-Nishiwaki, T., makishima, H., Nakatsukasa, M., Matsuzawa, T. (2011).
Differential prefrontal white matter development in chimpanzees and humans.
Current Biology 21, 1397-1402.
Sanz, C., Morgan, D., Gulick, S. (2004). New insights into chimpanzees, tools, and
termites from the Congo Basin. The American Naturalist 164, 567-581.
Sanz, C., Morgan, D. (2009). Flexible and persistent tool-using strategies in honeygathering by wild chimpanzees. International Journal of Primatology 30, 411–
427.
Sarich, V.M., Wilson, A.C. (1967). Immunological time scale for hominid evolution.
Science 158, 1200-1203.
Savage-Rumbaugh, S., Rumbaugh, D., Boysen, S. (1978). Symbolic communication
between two chimpanzees (Pan troglodytes). Science 201, 641-644.
Savage-Rumbaugh, S., Rumbaugh, D., Boysen, S. (1979). Linguistically mediated tool
use and exchange by chimpanzees (Pan troglodytes). Behavioral and Brain
Sciences 1, 539-554.
Saygin A.P., Leech R., Dick F. (2010). Nonverbal auditory agnosia with lesion to
Wernicke's area. Neuropsychologia 48, 107-113.
375
Schenker, N.M., Desgouttes, A-M., Semendeferi, K. (2005). Neural connectivity and
cortical substrates of cognition in hominoids. Journal of Human Evolution 49,
547-569.
Scherer, L.C., Fonseca, R.P., Amiri, M., Adrover-Roig, D., Marcotte, K., Giroux, F.,
Senhadji, N., Benali, H., Lesage, F., Ansaldo, A.I. (2011). Syntactic processing in
bilinguals: An fNIRS study. Brain & Language 121, 144-151.
Schick, K.D., Toth, N., Garufi, G., Savage-Rumbaugh, E.S., Rumbaugh, D., Sevcik, R.
(1999). Continuing investigations into the stone tool-making and tool-using
capabilities of a bonobo (Pan paniscus). Journal of Archaeological Science 26,
821-832.
Schlanger, N. (1996). Understanding Levallois: Lithic technology and cognitive
archaeology. Cambridge Archaeological Journal 6, 231-254.
Schleicher, A. (1863). Die Darwinsche Theorie und die Sprachwis-senschaft. Offenes
Sendschreiben, etc. Weimar, Hermann Böhlau.
Schmidt K.L., Cohn J.F. (2001). Human Facial Expressions as Adaptations:Evolutionary
Questions in Facial Expression Research. American Journal of Physical
Anthropology 33, 3-24.
Schmidt, R.R. (1936[1934]). The Dawn of the Human Mind: A Study of Palaeolithic
Man. R.A.S. Macalister, trans. London: Sidgwick & Jackson.
Schmitt, J.E., Eyler, L.T., Giedd, J.N., Kremen, W.S., Kendler, K.S., Neale, M.C. (2007).
Review of twin and family studies on neuroanatomic phenotypes and typical
neurodevelopment. Twin Research and Human Genetics 10, 683-694.
376
Schoenemann, P.T. (2006). Evolution of the size and functional areas of the human brain.
Annual Review of Anthropology 35, 379-406.
Schoenemann, P.T., Sheehan, M.J., Glotzer, L.D. (2005). Prefrontal white matter volume
is disproportionately larger in humans than in other primates. Nature
Neuroscience 8, 242-252.
Schubotz, R.I., von Cramon, D.Y. (2002). Predicting perceptual events activates
corresponding motor schemes in lateral premotor cortex: An fMRI study.
NeuroImage 15, 787-796.
Scott-Phillips, T.C. (2010). Evolutionary psychology and the origins of language. Journal
of Evolutionary Psychology 8, 289-307.
Sellet, F. (1993). Chaine operatoire: The concept and its applications. Lithic Technology
18, 106-112.
Semaw, S., Renne, P., Harris, J.W.K., Feibel, C.S., Bernor, R.L., Fesseha, N., et al.
(1997). 2.5-million-year-old stone tools from Gona, Ethiopia. Nature 385, 333336.
Shea, J.J. (2006). Child’s play: Reflections on the invisibility of children in the
Paleolithic record. Evolutionary Anthropology 15, 212-216.
Sherwood, C.C., Subiaul, F., Zawadzki, T.W. (2008). A natural history of the human
mind: Tracing evolutionary changes in brain and cognition. Journal of Anatomy
212, 426-454.
Shipton, C. (2010). Imitation and shared intentionality in the Acheulean. Cambridge
Archaeologicial Journal 20, 197-210.
377
Shultz, S., Nelson, E., Dunbar, R.I.M. (2012). Hominin cognitive evolution: Identifying
patterns and processes in the fossil and archaeological record. Philosophical
Transactions of the Royal Society B 367, 2130-2140.
Simon, S.R., Meunier, M., Piettre, L., Berardi, A.M., Segebarth, C.M., Boussaoud, D.
(2002). Spatial attention and memory versus motor preparation: Premotor cortex
involvement as revealed by fMRI. Journal of Neurophysiology 88, 2047-2057.
Simons, E.L. (1961). The phyletic position of Ramapithecus. Postilla 57, 1-9.
Skinner, H.A. (1982). The drug abuse screening test. Addictive Behavior 7, 363-371.
Slobin, D.I. (2002). Language evolution, acquisition, and diachrony: Probing the
parallels. In T. Givón and B.F. Malle (Eds.), The Evolution of Language Out of
Pre-Language (pp. 375-392). Philadelphia, John Benjamins.
Solodenko, N., Zupancich, A., Cesaro, S.N., Marder, O., Lemorini, C., Barkai, R. (2015).
Fat residue and use-wear found on Acheulian biface and scraper associated with
butchered elephant remains at the site of Revadim, Israel. PLoS One 10,
e0118572.
Soper, D.S. (2014). A-priori sample size calculator for Student’s T-tests [Software].
Available from http://www.danielsoper.com/statcalc.
Sperber, D. (1994). The modularity of thought and the epidemiology of representations.
In L.A. Hirschfeld and S.A. Gelman (Eds.), Mapping the Mind: Domain
Specificity in Cognition and Culture (pp. 39-67). Cambridge: Cambridge
University Press.
378
Steele, J., Uomini, N. (2009). Can the archaeology of manual specialization tell us
anything about language evolution? A survey of the state of play. Cambridge
Archaeological Journal 19, 97-110.
Steels L. (1997). The synthetic modeling of language origins. Evolution of
Communication 1, 1-34.
Steinbrink, J., Villringer, A., Kempf, F., Haux, D., Boden, S., Obrig, H. (2006).
Illuminating the BOLD signal: Combined fMRI-fNIRS studies. Magnetic
Resonance Imaging 24, 495-505.
Stout, D. (2002). Skill and cognition in stone tool production: An ethnographic case study
from Irian Jaya. Current Anthropology 43, 693-722.
Stout, D. (2003). Stone Tools and the Evolution of Human Thinking: Cultural, Biological,
and Archaeological Elements in an Anthropology of Human Origins. Ph.D.
dissertation, Department of Anthropology, Indiana University.
Stout, D. (2006). Oldowan toolmaking and hominin brain evolution: Theory and research
using positron emission tomography (PET). In K. Schick and N. Toth (Eds.), The
Oldowan: Case Studies into the Earliest Stone Age, (pp. 267-305). Gosport, IN:
Stone Age Institute Press.
Stout, D. (2010). Possible relations between language and technology in human
evolution. In A. Nowell and I. Davidson (Eds.), Stone Tools and the Evolution of
Human Cognition (pp. 159-206). Boulder: University Press of Colorado.
Stout, D. (2011). Stone toolmaking and the evolution of human culture and cognition.
Philosophical Transactions of the Royal Society B 366, 1050-1059.
379
Stout, D., Apel, J., Commander, J., Roberts, M. (2014). Late Acheulean technology and
cognition at Boxgrove, UK. Journal of Archaeological Science 41, 576-590.
Stout, D., Chaminade, T. (2007). The evolutionary neuroscience of tool making.
Neuropsycholgia 45, 1091-1100.
Stout, D., Chaminade, T. (2009). Making tools and making sense: Complex, intentional
behaviour in human evolution. Cambridge Archaeological Journal 19, 85-96.
Stout, D., Chaminade, T. (2012). Stone tools, language and the brain in human evolution.
Philosophical Transactions of the Royal Society B 367, 75-87.
Stout, D., Hecht, E. (2015). Neuroarchaeology. In E. Bruner (Ed.), vol. 3: Human
Paleoneurology (pp. 145-175). New York: Springer International Publishing.
Stout, D., Hecht, E., Nada, K., Bradley, B., Chaminade, T. (2015). Cognitive demands of
Lower Paleolithic toolmaking. PLoS ONE 10, e0121804.
Stout, D., Passingham, R., Frith, C., Apel, J., Chaminade, T. (2011). Technology,
expertise and social cognition in human evolution. European Journal of
Neuroscience 33, 1328-1338.
Stout, D., Toth, N., Schick, K. (2006). Comparing the neural foundations of Oldowan and
Acheulean toolmaking: A pilot study using positron emission tomography (PET).
In K. Schick and N. Toth (Eds.), The Oldowan: Case Studies into the Earliest
Stone Age, (pp. 321-331). Gosport, IN: Stone Age Institute Press.
Stout, D., Toth, N., Schick, K.D., Chaminade, T. (2008). Neural correlates of Early Stone
Age tool-making: Technology, language and cognition in human evolution.
Philosophical Transactions of the Royal Society B 363, 1939-1949.
380
Stout, D., Toth, N., Schick, K., Stout, J.C., Hutchins, G. (2000). Stone tool-making and
brain activation: Positron emission tomography (PET) studies. Journal of
Archaeological Science 27, 1215-1223.
Straus, L.G. (2012). The emergence of modern-like forager capacities and behaviors in
Africa and Europe: Abrupt or gradual, biological or demographic? Quaternary
International 247, 350-357.
Studdert-Kennedy, M. (1998). The particulate orgins of language generativity: From
syllable to gesture. In J. Hurford, M. Studdert-Kennedy, and C. Knight (Eds.),
Approaches to the Evolution of Language (pp. 202-221). Cambridge: Cambridge
University Press.
Subiaul, F., Cantlon, J.F., Holloway, R.L., Terrace, H.S., (2004). Cognitive imitation in
rhesus macaques. Science 305, 407-410.
Surrey, L.R., Nelson, K., Delelio, C., Mathie-Majors, D., Omel-Edwards N., Shumaker,
J., Thurber, G. (2003). A comparison of performance outcomes between the
Minnesota Rate of Manipulation Test and the Minnesota Manual Dexterity Test.
Work 20, 97–102.
Szaflarski, J.P., Binder, J.R., Possing, E.T., McKiernan, K.A., Ward, B.D., Hammeke,
T.A. (2002). Language lateralization in left-handed and ambidextrous people:
fMRI data. Neurology 59, 238-244.
Taglialatela, J.P., Russell, J.L., Schaeffer, J.A., Hopkins, W.D. (2008). Communicative
signaling activates ‘Broca’s’ homolog in chimpanzees. Current Biology 18, 343348.
381
Taglialatela, J.P., Russell, J.L., Schaeffer, J.A., Hopkins, W.D. (2009). Visualizing vocal
perception in the chimpanzee brain. Cerebral Cortex 19, 1151-1157.
Tak, S., Chul, Y.J. (2014). Statistical analysis of fNIRS data: A comprehensive review.
NeuroImage 85, 72-91.
Tanner, J.E., Byrne, R.W. (1996). Representation of action through iconic gesture in a
captive lowland gorilla. Current Anthropology 37:162–173.
Tattersall, I. (2008). An evolutionary framework for the acquisition of symbolic cognition
by Homo sapiens. Comparative Cognition and Behavior Reviews 3, 99-114.
Tattersall, I. (2010). Human evolution and cognition. Theory in Biosciences 129, 193201.
Tavares, M.C.H., Tomaz, C. (2002). Working memory in capuchin monkeys (Cebus
apella). Behavioural Brain Research 131, 131-137.
Taylor, M.J., Smith, M.L., Iron, K.S. (1990). Event-related potential evidence of sex
differences in verbal and nonverbal memory tasks. Neuropsychologia 28, 691705.
Tikhonov, A. (1963). Solution of incorrectly formulated problems and the regularization
method. Soviet Mathematics-Doklady 5, 1035-1038.
Tobias, P.V. (1981). The emergence of man in Africa and beyond. Philosophical
Transactions of the Royal Society of London B 292, 43-56.
Tobias, P.V. (1987). The brain of Homo habilis: A new level of organization in cerebral
evolution. Journal of Human Evolution 16, 741-761.
Tobias, P.V. (1988). The brain of Homo habilis: A new level of organization in cerebral
evolution. Journal of Human Evolution 16, 741-761.
382
Tomasello, M. (2002). Not waving but speaking: How important were gestures in the
evolution of language? Nature 417, 791-792.
Tomasello, M. (2003). On the different origins of symbols and grammar. In M.H.
Christiansen and S. Kirby (Eds.), Language Evolution (pp. 94-110). Oxford,
Oxford University Press.
Tomasello, M., Call, J., Warren, J., Frost, G., Carpenter, M., Nagell, K. (1997). The
ontogeny of chimpanzee gestural signals: A comparison across groups and
generations. Evolution of Communication 1, 223–259.
Tomasello, M., Herrmann, E. (2010). Ape and human cognition: What’s the difference?
Current Directions in Psychological Science 19, 3-8.
de la Torre, I. (2010). Insights on the technical competence of the early Oldowan. In A.
Nowell and I. Davidson (Eds.), Stone Tools and the Evolution of Human
Cognition (pp. 45-66). Boulder: University Press of Colorado.
Toth, N. (1985a). Archaeological evidence for preferential right-handedness in the Lower
and Middle Pleistocene, and its possible implications. Journal of Human
Evolution 14, 607-614.
Toth, N. (1985b). The Oldowan reassessed: A close look at early stone artifacts. Journal
of Archaeological Science 12, 101-120.
Toth, N., Schick, K. (1993). Early stone industries and inferences regarding language and
cognition. In K.R. Gibson and T. Ingold (Eds.), Tools, Language and Cognition in
Human Evolution (pp. 346-362). Cambridge: Cambridge University Press.
Toth, N., Schick, K. (2009). The Oldowan: The tool making of early hominins and
chimpanzees compared. Annual Review of Anthropology 38, 289-305.
383
Toth, N., Schick, K., Savage-Rumbaugh, S., Sevcik, R., Rumbaugh, D. (1993). Pan the
tool-maker: Investigations into the stone tool-making and tool-using capabilities
of a bonobo (Pan paniscus). Journal of Archaeological Science 20, 81-91.
Toth, N., Schick, K., Semaw, S. (2006). A comparative study of the stone tool-making
skills of Pan, Australopithecus, and Homo sapiens. In N. Toth and K. Schick
(Eds.), The Oldowan: Case Studies into the Earliest Stone Age (pp. 155-222).
Gosport, IN: Stone Age Institute Press.
Trauth, M.H., Maslin, M.A., Deino, A.L., Junginger, A., Lesoloyia, M., Odada, E.O.,
Olago, D.O., Olaka, L.A., Strecker, M.R., Tiedemann, R. (2010). Human
evolution in a variable environment: The amplifier lakes of Eastern Africa.
Quaternary Science 29, 2981-2988.
Trigger, B.G. (2006). A History of Archaeological Thought. 2nd edition. New York:
Cambridge University Press.
Tunik, E., Lo, O-Y., Adamovich, S.V. (2008). Transcranial magnetic stimulation to the
frontal operculum and supramarginal gyrus disrupts planning of outcome-based
hand–object interactions. The Journal of Neuroscience 28, 14422-14427.
Tylor, E.B. (1894). On the Tasmanians as representatives of Palaeolithic Man. The
Journal of the Anthropological Institute of Great Britain and Ireland 23, 141-152.
Ullman, M.T. (2006). Is Broca’s area part of a basal ganglia thalamocortical circuit?
Cortex 42, 480-485.
Umiltà, M.A., Escola, L., Intskirveli, I., Rochat, M.J., Caruana, F., Jezzini, A., Gallese,
V., Rizzolatti, G. (2008). When pliers become fingers in the monkey motor
system. Proceedings of the National Academy of Sciences 105, 2209-2213.
384
Ungerleider, L.G., Doyon, J., Karni, A. (2002). Imaging brain plasticity during motor
skill learning. Neurobiology of Learning and Memory 78, 553-564.
Uomini, N.T. (2001). Lithic Indicators of Handedness: Assessment of Methodologies and
the Evolution of Laterality in Hominids. M.Sc. Thesis. Department of Biological
Anthropology, University of Durham.
Uomini, N.T. (2006). In the Knapper’s Hands: Testing Markers of Laterality in Hominin
Lithic Production, with Reference to the Common Substrate of Language and
Handedness. Ph.D. Thesis. Department of Archaeology, University of
Southampton.
Uomini, N.T., Meyer, G.F. (2013). Shared brain lateralization patterns in language and
Acheulean stone tool production: A functional transcranial Doppler ultrasound
study. PLoS One 8, e72693.
Vanduffel, W., Fize, D., Peuskens, H., Denys, K., Sunaert, S., Todd, J.T., Orban, G.A.
(2002). Extracting 3D from motion: Differences in human and monkey
intraparietal cortex. Science 298, 413-415.
Van Lawick-Goodall, J. (1968). The behavior of free-living chimpanzees in the Gombe
Stream Reserve. Animal Behaviour Monographs 1, 161-311.
Van Lawick-Goodall, J. (1970). Tool-using in primates and other vertebrates. In D.S.
Lehrman (Ed.), vol. 3: Advances in the Study of Behavior (pp. 195-246). New
York: Academic Press.
385
Vaughan, C.D. (2001). A million years of style and function: Regional and temporal
variation in Acheulean handaxes. In T.D. Hurt and G.F.M. Rakita (Eds.), Style
and Function: Conceptual Issues in Evolutionary Archaeology (pp. 141-164).
Westport, CT, Bergin & Garvey.
Vigneau, M., Beaucousin, V., Hervé, P.Y., Duffau, H., Crivello, F., Houdvé, O.,
Mazoyer, B., Tzourio-Mazoyer, N. (2006). Meta-analyzing left hemisphere
language areas: Phonology, semantics, and sentence processing. NeuroImage 30,
1414-1432.
Vigneau, M., Beaucousin, V., Hervé, P-Y., Jobard, G., Petit, L., Crivello, F., Mellet, E.,
Zago, L., Mazoyer, B., Tzourio-Mazoyer, N. (2011). What is right-hemisphere
contribution to phonological, lexico-semantic, and sentence processing? Insights
from a meta-analysis. NeuroImage 54, 577-593.
Villmoare, B., Kimbel, W.H., Seyoum, C., Campisano, C.J., DiMaggio, E.N., Rowan, J.,
Braun, D.R., Arrowsmith, J.R., Reed, K.E. (2015). Early Homo at 2.8 Ma from
Ledi-Geraru, Afar, Ethiopia. Science 347, 1352-1355.
Visalberghi, E., Addessi, E., Truppa, V., Spagnoletti, N., Ottoni, E., Izar, P., Fragaszy, D.
(2009). Selection of effective stone tools by wild bearded capuchin monkeys.
Current Biology 19, 213-217.
Visalberghi, E., Fragaszy, D.M. (1994). Do monkeys ape? In S.T. Parker and K.R.
Gibson (Eds.), “Language” and Intelligence in Monkeys and Apes: Comparative
Developmental Perspectives (pp. 247-273). New York: Cambridge University
Press.
386
Vogel, E.K., Machizawa, M.G. (2004). Neural activity predicts individual differences in
visual working memory capacity. Nature 428, 748-751.
Volz, K.G., Schubotz, R.I., von Cramon, Y. (2005). Variants of uncertainty in decisionmaking and their neural correlates. Brain Research Bulletin 67, 403-412.
Vouloumanos, A., Kiehl, K.A., Werker, J.F., Liddle, P.F. (2001). Detection of sounds in
the auditory stream: Event-related fMRI evidence for differential activation to
speech and nonspeech. Journal of Cognitive Neuroscience 13, 994-1005.
Wadley, L. (2010a). Compound-adhesive manufacture as a behavioral proxy for complex
cognition in the Middle Stone Age. Current Anthropology 51, S111-S119.
Wadley, L. (2010b). Were snares and traps used in the Middle Stone Age and does it
matter: A review and a case study from Sibudu, South Africa. Journal of Human
Evolution 58, 179-192.
Wadley, L., Hodgskiss, T., Grant, M. (2009). Implications for complex cognition from
the hafting of tools with compound adhesives in the Middle Stone Age, South
Africa. Proceedings of the National Academy of Sciences 106, 9590-9594.
Wakita, M., Shibasaki, M., Ishizuka, T., Schnackenberg, J., Fujiawara, M., Masataka, N.
(2010). Measurement of neuronal activity in a macaque monkey in response to
animate images using near-infrared spectroscopy. Frontiers in Behavioral
Neuroscience 4, 31.
Wallace, A.R. (1864). The origin of human races and the antiquity of man deduced from
the theory of ‘natural selection.’ Journal of the Anthropological Society of London
2, clviii-xxxvii.
387
Wallace, A.R. (1869). Geological climates and the origin of species. The Quarterly
Review 126, 359-394.
Washburn, D.A., Gulledge, J.P., James, F., Rumbaugh, D.M. (2007). A species difference
in visuospatial working memory: Does language link “what” with “where”?
International Journal of Comparative Psychology 20, 55-64.
Washburn, S.L. (1951). The new physical anthropology. Transaction of the New York
Academy of Science Series II 13, 298-304.
Weiner, J.S., Oakley, K.P., Le Gros Clark, W.E. (1953). The solution of the Piltdown
problem. Bulletin of the British Museum 2,139-146.
Welsh, M.C., Pennington, B.F., Groisser, D.B. (1991). A normative-developmental study
of executive function: A window on prefrontal functions in childhood.
Developmental Neuropsychology 7, 131-139.
Wernicke, C. (1874). Der aphasische Symptomencomplex; eine psychologische Studie
auf Anatomischer Basis. Breslau: Cohn & Weigert.
Westerberg, H., Klingberg, T. (2007). Changes in cortical activity after training of
working memory—A single-subject analysis. Pysiology & Behavior 92, 186-192.
White M.J. (2006). Axing cleavers: Reflections on broad-tipped large cutting tools in the
British Lower and Middle Palaeolithic. In N. Goren-Inbar and G. Sharon (Eds.),
Axe Age: Acheulian Toolmaking, from Quarry to Discard (pp. 365-386).
Jerusalem, Equinox.
White, M.J., Ashton, N.M. (2003). Lower Palaeolithic core technology and the origins of
the Levallois method in North-Western Europe. Current anthropology 44, 598609.
388
Whiten, A., Byrne, R.W. (1988). The Machiavellian intelligence hypotheses. In R.W.
Byrne and A. Whiten (Eds.), Machiavellian Intelligence: Social Expertise and the
Evolution of Intellect in Monkeys, Apes and Humans. Oxford, Oxford University
Press.
Wijeakumar, S., Huppert, T., Magnotta, V.A., Buss, A.T., Spencer, J.P. (In Press).
Probing the neural basis of visual working memory with image-based fNIRS: An
fMRI validation study. NeuroImage.
Wijeakumar, S., Spencer, J.P., Bohache, K., Boas, D.A., Magnotta, V.A. (2015).
Validating a new methodology for optical probe design and image registration in
fNIRS studies. NeuroImage 106, 86-100.
Wildgruber, D., Riecker, A., Hertrich, I., Erb, M., Grodd, W., Ethofer, T., Ackermann, H.
(2005). Identification of emotional intonation evaluated by fMRI. NeuroImage 24,
1233-1241.
Williams, V.M.E., Burke, A., Lombard, M. (2014). Throwing spears and shooting
arrows: Preliminary results of a pilot neuroarchaeological study. South African
Archaeological Bulletin 69, 199-207.
Wilson, E.O. (1975). Sociobiology. Cambridge, Harvard University Press.
Wolpoff, M.H. (1980). Paleoanthropology. New York, Knopf.
Wong, C., Gallate, J. (2012). The function of the anterior temporal lobe: A review of the
empirical evidence. Brain Research 1449, 94-116.
Wright, R.V.S. (1972). Imitative learning of a flaked stone technology—The case of an
orangutan. Mankind 8, 296-306.
Wynn, T. (1979). The intelligence of later Acheulean hominids. Man 14, 371-391.
389
Wynn, T. (1981). The intelligence of Oldowan hominids. Journal of Human Evolution
10, 529-541.
Wynn, T. (1985). Piaget, stone tools and the evolution of human intelligence. World
Archaeology 17, 32-43.
Wynn, T. (1989). The Evolution of Spatial Competence. Chicago: University of Illinois
Press.
Wynn, T. (1991). Tools, grammar and the archaeology of cognition. Cambridge
Archaeological Journal 1, 191-206.
Wynn, T. (1993). Layers of thinking in tool behavior. In K.R. Gibson and T. Ingold
(Eds.), Tools, Language and Cognition in Human Evolution (pp. 363-388).
Cambridge: Cambridge University Press.
Wynn, T. (2002). Archaeology and cognitive evolution. Behavioral and Brain Sciences
25, 389-438.
Wynn, T. (2009). Whither evolutionary cognitive archaeology? Afterword. In S.A. de
Beaune, F.L. Coolidge, and T. Wynn (Eds.), Cognitive Archaeology and Human
Evolution (pp. 145-149). Cambridge: Cambridge University Press.
Wynn, T., Coolidge, F.L. (2004). The expert Neandertal mind. Journal of Human
Evolution 46, 467–487.
Wynn, T., Coolidge, F.L. (2011). The implications of the working memory model for the
evolution of modern cognition. International Journal of Evolutionary Biology
2011, 1-12.
Wynn, T., Hernandez-Aguilar, R.A., Marchant, L.F., McGrew, W.C. (2011). “An ape’s
view of the Oldowan” revisited. Evolutionary Anthropology 20, 181-197.
390
Wynn, T., McGrew, W.C. (1989). An ape’s view of the Oldowan. Man 24, 383-398.
Wynn T., Tierson F. (1990). Regional comparison of the shapes of later Acheulian
handaxes. American Anthropologist 92, 73-84.
Yankosec, K.E., Howell, D. (2009). A narrative review of dexterity assessments. Journal
of Hand Therapy 22, 258-270.
Ye, J.C., Tak, S. Jang, K.E., Jung, J., Jang, J. (2009). NIRS-SPM: Statistical parametric
mapping for near-infrared spectroscopy. NeuroImage 44, 428-447.
Yücel, M.A., Selb, J., Cooper, R.J., Boas, D.A. (2014). Targeted principle component
analysis: A new motion artifact correction approach for near-infrared
spectroscopy. Journal of Innovative Optical Health Sciences 7, 1350066.
Zhou, H., Desimone, R. (2011). Feature-based attention in the frontal eye field and area
V4 during visual search. Neuron 70, 1205-1217.
Zuckerknadl, E., Pauling, L.B. (1962). Molecular disease, evolution, and genic
heterogeneity. In M. Kasha and B. Pullman (Eds.), Horizons in Biochemistry (pp.
189-225). New York: Academic Press.
391