See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/298522361

Homo Symbolicus: The dawn of language, imagination and spirituality

Book · October 2011

DOI: 10.1075/z.168

CITATIONS READS

71 10,837

2 authors:

Christopher S Henshilwood Francesco d'Errico

University of Bergen CNRS-University of Bordeaux
167 PUBLICATIONS 11,462 CITATIONS 436 PUBLICATIONS 20,257 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Christopher S Henshilwood on 13 April 2016.

The user has requested enhancement of the downloaded file.

Homo Symbolicus
Homo Symbolicus
The dawn of language, imagination and spirituality

Edited by

Christopher S. Henshilwood
University of Bergen & University of the Witwatersrand

Francesco d’Errico
University of Bordeaux I & University of Bergen

John Benjamins Publishing Company

Amsterdam / Philadelphia
 TM
The paper used in this publication meets the minimum requirements of
8

American National Standard for Information Sciences – Permanence of

Paper for Printed Library Materials, ansi z39.48-1984.

Library of Congress Cataloging-in-Publication Data

Homo symbolicus : the dawn of language, imagination and spirituality /

edited by Christopher S. Henshilwood, Francesco d’Errico.
p. cm.
Includes bibliographical references and index.
1. Symbolism (Psychology) 2. Human behavior. 3. Language and languages--Origin.
4. Psycholinguistics. 5. Biolinguistics. I. Henshilwood, Christopher Stuart. II.
D’Errico, Francesco.
BF458.H65   2011
155.7--dc23 2011031012
isbn 978 90 272 1189 7 (Hb ; alk. paper)
isbn 978 90 272 8409 9 (Eb)

© 2011 – John Benjamins B.V.

No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any
other means, without written permission from the publisher.
John Benjamins Publishing Co. · P.O. Box 36224 · 1020 me Amsterdam · The Netherlands
John Benjamins North America · P.O. Box 27519 · Philadelphia pa 19118-0519 · usa
 Table of contents

Editors’ introduction vii

Acknowledgements xi
chapter 1
Pan Symbolicus: A cultural primatologist’s viewpoint 1
William C. McGrew
chapter 2
The evolution and the rise of human language: Carry the baby 13
E. Sue Savage-Rumbaugh & William M. Fields
chapter 3
The origin of symbolically mediated behaviour: From antagonistic
scenarios to a unified research strategy 49
Francesco d’Errico & Christopher S. Henshilwood
chapter 4
Middle Stone Age engravings and their significance to the debate on the
emergence of symbolic material culture 75
Christopher S. Henshilwood & Francesco d’Errico
chapter 5
Complex cognition required for compound adhesive manufacture
in the Middle Stone Age implies symbolic capacity 97
Lyn Wadley
chapter 6
The emergence of language, art and symbolic thinking: A Neandertal test
of competing hypotheses 111
João Zilhão
chapter 7
The human major transition in relation to symbolic behaviour, including
language, imagination, and spirituality 133
David Sloan Wilson
chapter 8
The living as symbols, the dead as symbols: Problematising
the scale and pace of hominin symbolic evolution 141
Paul Pettitt
 Table of contents

chapter 9
Biology and mechanisms related to the dawn of language 163
George F.R. Ellis
chapter 10
The other middle-range theories: Mapping behaviour
and the evolution of the mind 185
Benoît Dubreuil
chapter 11
Metarepresentation, Homo religiosus, and Homo symbolicus 205
Justin L. Barrett
Index 225
 Editors’ introduction

Few of us take the time to reflect on the role that symbols play in our everyday
lives, both on a conscious and sub-conscious level, and how these symbols have
become an intricate part of our humanity. Symbols can be inscribed on our bodies
from birth through to later life, for example, in the form of circumcision, scarifica-
tion, mutilation and tattoos. We also cover our bodies with symbols each day with
the application of make-up, donning of clothing and accessories and the way we
dress our hair. The artificial environments in which we live at home and at work
are permeated with symbols and we attribute symbolic meanings to the natural
world as well, including far away planets, stars and galaxies. A myriad of symbols
are created and stored in our minds; we establish dialogues among them within
our heads and are able to transmit these representations to others despite the fact
they exist only in our imagination.
Symbols can act as powerful weapons in war and peace and we think here of
insignia, flags and mottos. We recycle symbols used by others and give them a dif-
ferent, possibly opposite meaning. Swastikas were important symbols in various
ancient civilizations and remain widely in use in Indian religions. The adoption
of the swastika by the Nazi party in Germany after 1920 led to an association of
this symbol by some with fascism and white supremacists, the antithesis of its
earlier meaning. Language, writing, religion, science, institutions and laws could
not possibly exist without our capacity to master the creation, transmission, and
disposal of symbols. We also take our symbols to the grave and after death we may
ourselves become regarded by others as a symbol.
Some researchers believe that substantial progress has been made in under-
standing the functioning of various categories of symbols and how they condi-
tion our thoughts, speech and actions today, but our knowledge of how and when
these aptitudes first emerged within our lineage is imprecise and speculative. If we
adopt a Darwinian perspective, we could argue that the first innovative introduc-
tion of symbols to mediate behaviour must stem from exaptation and long term
selective pressures. For decades, Palaeolithic archaeologists have suggested that
symbols first appeared in association with the material culture of Homo sapiens in
Germany or France about 40 000 years ago. Recent archaeological discoveries in
Africa support a much earlier emergence for the production and use of symbols.
These finds have set the stage to re-open and re-examine theoretical perspectives
and opened the door for constructive interdisciplinary collaborations. The more
 Homo symbolicus

research data we receive, the more apparent it becomes that the use and applica-
tion of symbols, at least to some degree, is a constitutive component of animal cog-
nition. It is therefore counter-intuitive to suggest that exploring only how we and
our ancestors used symbols can provide the key to understanding the very nature
of this phenomenon. The net needs to be cast more widely. Early pre-adaptations
to symbolic use must lie therefore within our own and other species and without
these roots, the more recent imposing achievements associated with Homo sapiens
could not have evolved.
Over the millennia, hominins have built on this potential to embark on a
symbolically laden adventure, filled at times with ambushes and failures, but also
successes. Only now are archaeologists starting to discover some of the trials of
these early communities. It has been a to-and-fro trial with balances and coun-
terbalances and the costs at times have exceeded the advantages, leading, in some
cases, to the return of old ways and the maintenance of the same “solutions” for
long periods of time. On occasion it has also led to the introduction of even more
innovative solutions. How this process led in the end to the cultural complexity
with which we are now familiar is one of the unanswered yet fascinating questions
of our evolutionary history. In this volume, we offer an opportunity for interdis-
ciplinary collaboration between scholars from a range of disciplines in an attempt
at addressing this question and bridging the gap between raw archaeological evi-
dence and human behaviour. We bring together new perspectives from the fields
of archaeology, anthropology, philosophy, evolutionary biology, religion, palaeo-
anthropology and linguistics with the aim of linking genetic, neural, cognitive,
and behavioural development. We believe that this approach opens novel path-
ways for understanding the ways in which the human mind evolved. In particular,
we look at the emergence of symbolism, imagination and spirituality from diverse
perspectives. The volume starts by examining the communication patterns and
lives of apes in habitats that resemble those in which the last common ancestor
of Homo and Pan resided, accepting that apes raised in a Pan/Homo environment
acquire significant aspects of human language and human culture. Since language
fluency was involved in archaic Homo sapiens becoming modern Homo sapiens,
and as some members of Pan paniscus have also acquired symbolic language, they
deserve the appellation Pan symbolicus. We subsequently explore the puzzling gap
of more than 80,000 years that exists in the archaeological record between the first
anatomical appearance of the species we call Homo sapiens sapiens and the appear-
ance of behaviours which we associate with our own behaviour today. We evaluate
archaeological evidence, including burial practices, the use of pigment and sym-
bolic artefacts, the manufacture of complex adhesives and the innovations that set
the stage for the later routine incorporation of material culture within a symbolic
system that mediated human behaviour. In the volume, we discuss the adaptations
 Editors’ introduction 

that brought about the dispersal of modern humans out of Africa, into Asia,
Australasia and Europe. We ask too, what are the main factors that account for the
emergence of key behavioural innovations in Africa and Eurasia between 200,000
and 30,000 years ago. Do we see significant differences between modern humans
in Africa and contemporaneous archaic populations living outside of this conti-
nent? Did climatic changes influence the behavioural differences between Nean-
dertals and Homo sapiens? A central focus of the volume is the development
of the ability to deliberately consider, ponder, or evaluate mental representa-
tions and how this may have lead to the development of language, spirituality
and symbolic behaviour. To address this, we ask to what extent activities such
as language, decoration, adornment, trade and exchange, and imagery are really
predicated on the incorporation of symbols within a behavioural repertoire. At
first sight the chapters in this book may appear eclectic, but with further reading
it becomes clear that there is a subtle thread that links these ideas together – we
believe that this approach of presenting highly diversified perspectives has pro-
vided an insightful and innovative account into the whys and wherefores of later
hominin evolution.
 Acknowledgements

Financial support was provided to the editors of this volume by the European Research
Council Advanced Grant TRACSYMBOLS (FP7 No. 249587), and the PROTEA
France - South Africa exchange programme. Christopher S. Henshilwood was also
supported by a National Research Foundation/Department of Science and Tech-
nology funded Chair at the University of the Witwatersrand, South Africa and
the Norwegian Research Council. The John Templeton Foundation and the Uni-
versity of Bergen, Norway, have contributed to establishing the interdisciplinary
research network that culminated in 2009 in a conference in Cape Town, South
Africa titled ‘Homo symbolicus: The Dawn of Language, Imagination and Spiri-
tuality’. The authors acknowledge the financial contributions of these two orga-
nizations with special thanks to Mary Ann Meyers, Paul Wason, Anna Gro Vea
Salvanes, Sigmund Grønmo, Svenn-Åge Dahl and Jan-Petter Myklebust. Apprecia-
tion is extended to Laura van Niekerk and Petro Keene who worked on earlier and
final versions of the texts. Anke de Looper and Patricia Leplae of John Benjamins NL
kindly guided us through the publishing process.
 chapter 1

Pan symbolicus
A cultural primatologist’s viewpoint
William C. McGrew
Department of Archaeology and Anthropology, University of Cambridge

Traditional views of non-human primate communication held that it was ­

hard-wired, involuntary, and emotionally expressive only. All of these
conclusions have been proven wrong: chimpanzees use natural vocalisations to
signal the nature and location of foodstuffs, that is, spontaneous referentiality.
Gibbons signal anti-predator alarm by using song-elements in rule-governed
combinations. Chimpanzees show tactical deception in their recruitment screams,
tailoring these to the audience. Pant-hoots show local variation that if found in
human language, would be called dialects. Males devise individualised courtship
routines. Some of these signals become group-typical, and such gestures are
neither iconic nor onomatopoeic, but instead are arbitrary. Primatological studies
of apes behaving spontaneously show us repeatedly that their communicative and
cognitive capacities exceed our anthropocentric expectations.

1. Introduction

Short of inventing a time-machine, we will never know directly the acts of our
extinct ancestors and their ilk. So, we must model what is missing, by one of three
methods: either we must use the discovered products of behaviour, that is, arte-
facts, to infer the missing behaviour of their use in antiquity, via archaeology; or,
we must seek to put flesh on the bones of the fossil record, and via functional
morphology, infer the whole, behaving organism from palaeontology; or, we must
scrutinise the behaviour of other living forms, on the grounds that their actions
will be more informative of the evolutionary process that led to us, than are studies
of ourselves, as the end-product (McGrew 2010).
Seeking such a proxy is not straightforward, and there is lively debate about
how to choose such a model (Sayers & Lovejoy 2008; Lovejoy 2009). Some evo-
lutionary anthropologists favour referential models, that is, choosing a living
taxon (usually but not always another species), as the nearest approximation to
an extinct hominin. Others favour non-referential models, that is, constructing
a hypothetical model based on evolutionary and ecological principles, which is
 William C. McGrew

not bounded by the limited choice available amongst extant taxa. Some favour
some kind of combination of the two (McGrew 2004). (Not seriously considered
here are more fanciful models based on speculative surmise, e.g. the Aquatic Ape
Theory, ­Morgan, 1997. Such ‘evolutionarios’ may be good fun to play with, but
unless they are testable, they take us nowhere scientifically).
One basis for choosing a referential model is phylogenetic proximity, that
is, concentrating on humankind’s nearest living relative, on the grounds that
maximally shared genes should be manifest in maximally similar behaviour, all
other things being equal. This sends us to the two living species of the genus Pan:
c­ himpanzee (P. troglodytes) and bonobo (P. paniscus). Each is equally related
to anatomically and behaviourally modern Homo sapiens, as the two ape taxa
­separated after the Pan line separated from our own. Thus, these African apes
are the living forms with whom we last shared common ancestry, albeit it about
7 million years ago. Genotypically, it is clear from the sequencing of the human
and chimpanzee genomes, that we share more than 99% of our genes. Phenotypi-
cally, there is more similarity than meets the eye (that is, hairy quadrupeds versus
mostly hairless bipeds): Pan and Homo blood is virtually interchangeable, and any
one of us could donate a life-saving transfusion to an ape, provided we correctly
matched the ABO blood groups.
Another basis for choosing a referential model is ecological similarity. That is,
we can focus on organisms that survive and thrive in environments similar to those
in which humankind emerged. We should expect such creatures to show similar
adaptations in response to similar natural selection pressures, whether these be
from predators, prey, climate, vegetation, etc. Finally, we might choose a refer-
ential model on the basis of behavioural performance. The archaeological record
of the evolutionary origins of human technology usually is based on persisting
lithics, as opposed to organic raw materials that perished long ago (McPherron
et al. 2010). But we cannot ignore the elementary technology of capuchin mon-
keys, especially Cebus libidinosus (for a recent review, see Ottoni & Izar 2008) or
long-tailed macaques (Gumert et al. 2009). These monkey species are but distantly
related to us by comparison with the African apes, but they do impressive things
with stones. Capuchin monkeys use stones in communication, agonism, and sub-
sistence, e.g. they are the only non-humans known to use stones as trowels to dig
up the underground storage organs of plants (Moura & Lee 2004). Long-tailed
macaques use different types of stone hammers for cracking open animal versus
plant prey. When the monkeys show such percussive technology, the modified
lithics are there to be recovered and analysed.
Here, in tackling the evolutionary origins of language, imagination and spiri-
tuality, I will focus on Pan, and whenever possible, on apes living in habitats that
resemble those in which the Last Common Ancestor of Homo and Pan resided.
 Chapter 1. Pan symbolicus 

For as much as possible, I will refer to observational studies of apes in nature,

rather than experimental studies of apes in captivity. I can only hope that what
is gained in socio-ecological validity from studying the organism in situ counter-
balances what is lost in the absence of control of key variables. Why choose a refer-
ential model? Because it gives us real behaviour in the real world. Why not choose
between chimpanzee and bonobo? Because the fossil record for post-Miocene
African apes is so sparse that we cannot say whether the Last Common Ancestor
more closely resembled one kind of Pan or the other.
Finally, I make explicit, a methodological and epistemological bias: I am an
ethologist, trained to record the behaviour of organisms and to try to infer their
intentions from their actions. The rationale for this approach is the same, whether
the subject of study is ant or aunt. Unobservable thoughts and feelings are thus
opaque, in any species, including our own; only their ‘reflections’ can be noted,
e.g. in autonomic responses, facial expressions, etc. Any such inference is an exer-
cise in probability, based on the consequences of behaviour or other recordable
phenomena, e.g. blushing. A person can no more readily penetrate the ‘black box’
that is the mind of a chimpanzee than it can that of a fellow human being. In
fact, we are bound to be more sceptical of behavioural data (including verbal self-
report) from human than from non-human animals, given the former’s capacity
for deception and especially self-deception. This may make the remainder of this
essay disappointingly conservative, but I hope not.
In tackling the prospect of Pan symbolicus, I make four basic assertions, striving
to develop each with regard to our current state of knowledge regarding Pan:

1. Communication does not equal language.

2. Language does not equal speech.
3. Symbol-use does not equal language.
4. Non-human language does not equal human language.

In all cases, I shall call upon the published literature in behavioural primatology,
as manifest in anthropology, biology and psychology, based on natural or natura-
listic processes. For that reason, experimental studies of ‘pongo-linguistics’ are
omitted (Savage-Rumbaugh et al. 1998). Similarly, emerging studies of the neuronal
substrates for communication, as studied comparatively (e.g. Taglialatela et al. 2008)
cannot be done justice here.

2. Communication does not equal language

If communication is defined minimally by the sequence of ‘sender-message-

receiver’, and if language is defined simply as a socially-acquired system of
 William C. McGrew

semantic and syntactic processing, then the two phenomena obviously are not
the same. Neither one is a necessary nor sufficient condition for the other. The two
need not even overlap, that is, one can imagine an organism that communicates
by pheromones, and thinks in terms of sentences but never transmits those
thoughts. Lots of organisms, even one-celled ones, communicate, but few
(apparently) use language to do so.
However, while language is undoubtedly useful in thinking, and most human
thought is hard to imagine without it, in everyday human life, language in what-
ever modality is a subset of communication. That is, humans habitually communi-
cate verbally (words), but also non-verbally (facial expressions), and para-verbally
(inflection, accent).
Chimpanzees communicate in all sensory modalities: Visually, they show a
wide range of facial expression, gesture, posture, and locomotion. For example,
invitation to play may take the simultaneous form of a ‘play-face’, slapping the
ground, quadrupedal bouncing, and a syncopated gallop (Goodall 1986). Acousti-
cally, chimpanzees drum with hands and feet on the resonating buttress roots of
trees, and each individual has a distinctive drumming ‘signature’ (Arcadi et al.
2004). Olfactorally, cycling females signal their changing reproductive states over
the menstrual cycle via perineal odour that is monitored by males. Tactilely, as
well as visually, the relative positions of the hands in the grooming hand-clasp (see
below) indicate the relative social dominance ranks of the two participants, with
the subordinate providing the support for the upraised arms while the dominant
is supported (McGrew 2004).
Is any of chimpanzee communication linguistic? Traditional views of nonhu-
man primate communication held that it was instinctive, involuntary, and emo-
tionally expressive only. Each of these conclusions since has been proven wrong:
Chimpanzees use natural ‘rough grunt’ vocalisations to signal the nature and loca-
tion of foodstuffs, that is, their communication shows spontaneous referentiality
(Slocombe & Zuberbühler 2005). (This confirms, in everyday life, what Savage-
Rumbaugh et al. 1978 have shown experimentally with an artificial language
system.) Gibbons also show referential signalling in their songs that signal anti-
predator alarm, by emitting song elements in rule-governed combinations (Clarke
et al. 2006). These are the roots of semantics and syntax.
Are such vocal emissions intentional? Yes, even monkeys show situation-
ally contingent vocal interaction: Langur monkey males give anti-predator alarm
calls only if there is an audience; moreover they continue calling until all group
members have responded, thus showing the ability to monitor and to recall the
vocal replies of their group mates (Wich & de Vries 2006). More impressively,
chimpanzees show tactical deception in their recruitment screams, when they
are aggressed again: depending on the composition of the audience, that is, their
 Chapter 1. Pan symbolicus 

social rank relative to that of the persecutor, they may exaggerate their screams.
This shows a working knowledge of so-called ‘triadic awareness’, indicating
understanding of third-party relationships (Slocombe & Zuberbühler 2007;
­Zuberbühler 2006). Thus, the modern view of non-human primate communica-
tion is that it is learned, voluntary, and intentionally referential and rule-governed.
Language as communication necessarily implies sociality, or in humans,
collectivity. Hundreds of human languages exist that clearly demarcate societal
boundaries and as a consequence, signal group identity, that is, ‘us versus them’.
(In contrast, solitary creatures avoid interaction and may reduce communication
to isolated ‘calling cards’, e.g. scent-marks left in the environment). To what extent
is there evidence of group-specific communication in apes? Marshall et al. (1999)
found that chimpanzees brought together in captivity from a variety of sources
showed a wide range of variants in their performance of the species’ long-call
vocalisation, the pant-hoot. However, over time these individual variants changed
to converge on a single version that became characteristic of the group.
In nature, different chimpanzee populations perform pant-hoots that show
local variation; the calls have the same elements but show nuanced differences that
if found in human language, we would call dialect (Mitani et al. 1992).

3. Language does not equal speech

Speech (spoken language) is neither a necessary nor a sufficient condition for

language. Signed language need not involve sound, and echolalic speech may be
nonsense. Extensive study of standard sign languages (as used by the deaf) seems
to indicate that it is not ‘second-class’ speech, and because of its visual format, it
may be even more efficient or expressive than speech. Speech by talking parrots,
mynah birds, etc. was once dismissible as mindless mimicry, but recent studies
of an African gray parrot may shade this black and white distinction as gray
(Pepperberg 1999).
More complicatingly, many vocalisations that are uttered are not verbal,
whether these are sighs or cries, and so need not be linguistic. Their non-linguistic
status is abundantly clear when they are involuntary, e.g. gasp, sob, yet they may be
highly communicative and arresting in their attention-getting power. Furthermore,
mechanical sounds may function linguistically, even if their perceived input is
non-verbal, e.g. Morse code. Finally, in principle, other sensory-motor modalities
than sound can function linguistically, but actual working examples are usually
secondary and derived, e.g. visual printing, tactile Braille.
Non-vocal communication in apes varies at the individual or group level
(Pollick & de Waal 2007). Individual male chimpanzees devise courtship ‘routines’
 William C. McGrew

that communicate to females their desire to mate; these sometimes incorporate

novel elements, e.g. an everted lower lip, that are added to behavioural sequences
by association with more unmistakeable sexual signals, e.g. an erect penis. At
Mahale, such signals are group-typical: males noisily tear to pieces leaf after leaf,
seeking to draw the attention of females to their sexual arousal. Such gestures
are not iconic or onomatopoeic; leaves have nothing to do with copulation per se
(Nishida et al. 2010).
The grooming hand-clasp (GHC) is a striking behavioural variant in chim-
panzee social grooming: Two individuals sit facing one another and each fully
extends the right or left arm overhead, where they clasp hands (McGrew & Tutin
1978). This symmetrical, mutual posture, in ‘A-frame’ configuration, occurs in
some populations of chimpanzees, but not in others. Detailed study shows inter-
group variation in GHC, even within the same population, e.g. in some groups, the
clasp is palm-to-palm, in others palm-to-wrist. In effect, the style of performance
seems to be an indicator of natal group identity, and immigrants modify their
GHC, but may retain it when grooming with others who originated in the same
natal group (McGrew 2004). If we saw such signals in two neighbouring human
groups, we would interpret these as ethnic markers.

4. Symbol-use does not equal language

If symbols are defined as mental (conceptual) representations of phenomena,

whether concrete or abstract, then they must be distinguished from signs, which
are more basic perceptual signals. The two are often confused. The colour of a t­ raffic
light (green or red) arbitrarily signifies permission to proceed or obligation to halt,
but no mental representation is necessary for the system to function. Instead, a
learned stimulus-response association will suffice. This distinction may be crucial
in distinguishing between human and non-human language (see below).
Even if symbols were never used in communication, they seem to be
essential to human thought. Despite the Sapir-Whorf hypothesis, that language
determines culture, non-human cultures seem to exist without language
(McGrew 2004; cf. Washburn & Benedict 1979). Mathematics is the prototypical
symbol system, and even if some numerical abilities are widespread across species
(Gelman & Gallistel 2004), some aspects of it, e.g. algebra, constitute invented
symbol systems. Comparative data do not always show human superiority:
chimpanzees are better at remembering strings of numerals presented briefly
than are adult humans (Matsuzawa 2009). However, like many other forms of
code, e.g. computer machine code, symbols may function non-linguistically,
even if devised by linguistic creatures.
 Chapter 1. Pan symbolicus 

On the other hand, symbol-use seems to be a necessary condition for language,

as defined above, even if it is not a sufficient one. All human languages seem to be
symbolic, although for how long in human evolution this has been the case is hard
to say, given that only written language persists.

5. Non-human language does not equal human language

The absent adjective ‘human’ is almost always assumed when the noun ‘language’ is
used. This is understandable, given our species’ anthropocentrism, but it need not
be the case. Decades ago, the dance language of honey bees (von Frisch 1927) was
recognised to have many linguistic features (e.g. Hockett 1960). Does this mean
that if a non-human language satisfied 9 of 10 criteria on some checklist of features,
then that phenomena could be considered 90% human language? Such musing
becomes more than speculation when contemplating the communicatory abilities
of dolphins and whales, whether via echo-location or full-blown singing. Some of
these issues are merely ones of sensory capacity, e.g. the vocal communication of
bats (ultra-high-frequency) or of elephants (ultra-low-frequency). Even humbler
creatures, that is, ones with relatively small brains, show unexpected complexity
in their language-like signals, e.g. meerkats simultaneously encode referential and
affective information in their alarm calls (Manser et al. 2002).
And what are we to do with human beings that are mewling infants or
drooling elders? If they produce utterances, however unintelligible, must they
be classed as language-users? Just because if they mature normally, the former
will become linguistic, and the latter once were so? Or, being pre-verbal or post-
verbal, either by immaturity or senility, are they in some way equivalent to the
less-than-complete language produced by other species? (The preceding questions
apply only to normal ontogeny, but become even more vexed when abnormality is
involved; at what point on the scale does diminished linguistic ability become no
different from absent ability?).

6. Pan symbolicus

To what extent can living Pan enlighten us about language, imagination, and spiri-
tuality? Several studies have already been mentioned above, especially with regard
to complex communication that may share some features with full-blown human
language, and more continue to be added. For example, female chimpanzees
manipulate their emission of copulation calls, increasing their calling frequency
in the presence of high-ranking males (i.e. potential mates), while decreasing their
 William C. McGrew

frequency in the presence of high-ranking females (i.e. competitors) (Townsend

et al. 2008). Restrictive generalisations that limited non-human communication to
involuntary expressions of emotion are no longer tenable.
Similarly, prior categorical dicta that had apes incapable of vocal imitation
look increasingly shaky. In addition to earlier studies of bonobo vocalisations (see
Savage-Rumbaugh et al. this volume), a recent report has an orang-utan showing
spontaneously whistling in mimicry of its caretaker (Wich et al. 2008).
But what of symbolic communication in nature? Two reports stand out, but
neither is conclusive. Boesch (1991) reported for the chimpanzees of the Tai For-
est, Ivory Coast, symbolic drumming. That is, behavioural responses to drumming
indicated that information about the drummer in time and space was being trans-
mitted, e.g. ‘we are resting here for a certain period, then will move off in a particular
direction’. However, once the main drummer died, the behaviour disappeared, and
so the phenomenon could not be studied further. Savage-Rumbaugh et al. (1996)
reported that bonobos at Wamba, Democratic Republic of Congo, left messages for
others trailing behind, in the form of bent-over vegetation which communicated
their direction of travel. However, these data were gathered over only two days, and
so constitute only a pilot study that needs to be followed up.
What can be said about the imagination of apes in nature? The natural his-
tory and ethnography of wild chimpanzee populations, especially reports from
Gombe, Tanzania, are rich in content and potential. Goodall (1986) has variously
described the ‘rain dance’, a wild and woolly social display, done especially by adult
males, that occurs at the outset of thunder storms. This has been interpreted as
some sort of statement of collective defiance at the sound and fury of nature, and
has been likened to human rituals in response to sensorally overwhelming natural
forces. However, rain dancing never has been analysed in detail, despite its exis-
tence being known for more than 40 years.
Similarly, chimpanzees may show prolonged and acrobatic ‘waterfall displays’
at the 10-m-high waterfall at the head of Kakombe Valley (Goodall 1986). This is
often done by solitary individuals, so that it can have no communicatory signifi-
cance, and thus it has been likened to an individual expression of awe or perhaps
superstition. What makes this more impressive than the rain dance is that it is
done predictably, that is, the site and nature of the waterfall is constant and well-
known. So, why should an ape be drawn to engage in such an energetic (and even
dangerous, as it involves swinging on vines high above a rocky streambed) activity,
as no reaction is possible from absent peers, predators, or prey? This mystery has
suggested deeper motives, perhaps to do with the evolutionary roots of animism,
to some commentators.
In fact, we have no more idea what is going on in the minds of ape rain­­dancers,
or waterfall displayers, than we can know what motivated the individuals who
 Chapter 1. Pan symbolicus 

c­ reated startling depictions of animals on Upper Palaeolithic cave walls or crafted

three-dimensional images of sumptuous women (Conard 2009). Even more tenta-
tive are inferences drawn about symbol-use based on simpler artefacts, at least as
it appears to this simple-minded primatologist. For example, from finding a con-
centration of snail shells, each having a single, similar hole, it is inferred that these
make a necklace, and that a necklace necessarily means deliberate ornamentation,
which amounts to evidence of symbolic capacity. If we grant the existence of the
still-hypothetical string that would be necessary to infer even the most minimal
necklace, there are other reasons to string items, as any fisherman knows. (There
is even an English vocational surname, Stringer, to remind us of this). Even if we
grant the putative necklace status as a manufactured ornament, this still does not
make it symbolic, as any ambitious bowerbird knows (Madden 2008). Perhaps
there is still a place for Occam’s razor in interpreting the significance of artefacts,
no matter how attractive the more complex explanation seems to be.
With the living apes, we can note certain individual variables, e.g. age, sex,
rank, etc. of the performers, as well as some contextual ones, e.g. whether alone or
in company, and if the latter, with whom; what activities both precede and follow
the event; what is the state of arousal of the performer, either before, during or
after the performance, etc. (Having said this, the published scientific literature is
disappointing. Primatologists mostly have been content to report even provoca-
tive events as anecdotes, but usually have not given them systematic study).
None of this tells us anything directly about the mental representations that
may or may not underlie the performance. The behaviour may appear compelling,
even ritualised, in terms of its form, tempo, and complexity, but this is true of
many displays performed by a variety of animals, even invertebrates, and we do
not seek to invoke imagination to explain their causation. In the uncontrolled
circumstances of the field, I can think of no way to penetrate the ape’s ‘black box’,
in order to ascertain the extent to which, if at all, imagination underlies such
intriguing behaviour.
If imagination is a problem, what then to do with spirituality? If the latter
phenomenon is notionally defined (crudely) as belief in supernatural forces, such a
definition stumbles already on its non-operationality: belief is problematic enough
to infer, but the supernatural by definition is inaccessible to science, as science is
bounded by natural law. If, on the other hand, natural but yet unexplained forces of
certain types are enough to infer spirituality, then we have to find a way to measure
them. I confess to not being able to comprehend how spirituality can be validated
even in Homo sapiens, given that verbal professions of faith are vulnerable to
deception, either intentional or otherwise. The unintentional form of self-deception
is especially problematic, even if one invokes the aid of psycho-physiology, e.g.
polygraph. Much as I would like to present apes as models for the origin and
 William C. McGrew

evolution of religion, I do not see an empirical way to do so. This is frustrating. (Of
course, if apes gathered in a forest clearing and knelt collectively to face the setting
sun, while vocalising rhythmically in unison, as it sank beneath the horizon, we
might be tempted to infer something potentially interesting in this regard. Then we
would have candidate behaviour to record and to analyse. Such a tableau is not as
far-fetched as it seems, as anyone who has seen groups of ring-tailed lemurs pose
with widespread arms to greet the morning sun can testify).
In summary, primatological studies of apes behaving spontaneously have
shown us that their communicatory and cognitive capacities exceed what even
recently was thought to be present. For example, a male chimpanzee surreptitiously
made and stored stone missiles to be used later as projectiles (Osvath 2009); such
acts done by humans would be labelled unhesitatingly as future planning. Fewer
and fewer boundaries between human and non-human hold up to empirical scru-
tiny, and the gap between linguistic and non-linguistic communication seems ever
narrower. These questions, at least, are amenable to scientific investigation, in both
field and laboratory. Questions of cognitive functioning, including imagination,
would seem to be accessible to well-designed experiments (see Matsuzawa 2009),
but not in nature, where too few variables can be controlled. Questions regarding
spirituality take us into areas that primatologists are ill-equipped to tackle, at least
with their current tools.

References

Arcadi, A.C., Robert, D. & Mugurusi, F. 2004. “A comparison of drumming by male chimpan-
zees from two populations.” Primates 45: 135–139.
Boesch, C. 1991. “Symbolic communication in wild chimpanzees.” Human Evolution 6: 81–90.
Clarke, E., Reichard, U. & Zuberbühler, K. 2006. “The syntax and meaning of wild gibbon
songs.” PLoS ONE 1 (1): e73.
Conard, N.J. 2009. “A female figurine from the basal Aurignacian of Hohle Fels Cave in south-
western Germany.” Nature 459: 248–252.
Gelman, R. & Gallistel, C.R. 2004. “Language and the origin of numerical concepts.” Science
306: 441–443.
Goodall, J. 1986. The Chimpanzees of Gombe. Cambridge MA: Harvard University Press.
Gumert, M.D., Kluck, M. & Malaivijitnond, S. 2009. “The physical characteristics and usage
patterns of stone axe and pounding hammers used by long-tailed macaques in the A
­ ndaman
Sea region of Thailand.” American Journal of Primatology 71: 594–608.
Hockett, C.F. 1960. “The origin of speech.” Scientific American. 203: 88–96.
Lovejoy, C.O. 2009. “Reexamining human origins in light of Ardipithecus ramidus.” Science 326
(5949): 74e1–74e8.
Madden, J.R. 2008. “Do bowerbirds exhibit cultures?” Animal Cognition 11: 1–12.
Manser, M.B., Seyfarth, R.M., & Cheney, D.L. 2002. “Suricate alarm calls signal predator class
and urgency.” Trends in Cognitive Sciences 6 (2): 55–57.
 Chapter 1. Pan symbolicus 

Marshall, A.J., Wrangham, R.W. & Arcadi, A. 1999. “Does learning affect the structure of vocal-
izations in chimpanzees?” Animal Behaviour 58: 825–830.
Matsuzawa, T. 2009. “Symbolic representation of number in chimpanzees.” Current Opinion in
Neurobiology 19: 1–7.
McGrew, W.C. 2004. The Cultured Chimpanzee. Reflections on Cultural Primatology. Cambridge:
Cambridge University Press.
McGrew, W.C. 2010. “In search of the last common ancestor: new findings on wild chimpanzees.”
Philosophical Transactions of the Royal Society B 365: 3265–3276.
McGrew, W.C. & Tutin, C.E.G. 1978. “Evidence for a social custom in wild chimpanzees?” Man
13: 234–251.
McPherron, S.P., Alemseged, Z., Marean, C.W., Wynn, J.G., Reed, D., Geraads, D., Bobe, R. &
Bearat, H.A. 2010. “Evidence for stone-assisted consumption of animal tissues before 3.39
million years ago at Dakika, Ethiopia.” Nature 466: 857–860.
Mitani, J., Hasegawa, T., Gros-Louis, J., Marler, P., & Byrne, R.W. 1992. “Dialects in wild
chimpanzees?” American Journal of Primatology 27: 233–243.
Moura, A.C.A. & Lee, P.C. 2004. “Capuchin stone tool use in a Caatinga dry forest.” Science
306: 1909.
Nishida, T., Zamma, K., Matsusaka, T., Inaba, A. & McGrew, W.C. 2010. Chimpanzee Behavior
in the Wild. An Audio-Visual Encyclopedia. Tokyo: Springer.
Osvath, M. 2009. “Spontaneous planning for future stone throwing by a male chimpanzee.”
­Current Biology 19 (5): R190-191.
Ottoni, E.B. & Izar, P. 2008. “Capuchin monkey tool use: Overview and implications.”
­Evolutionary Anthropology 17: 171–178.
Pepperberg, I.M. 1999. The Alex Studies. Cognitive and Communicative Abilities of Grey Parrots.
Cambridge, MA: Harvard University Press.
Pollick, A.S. & de Waal, F.B.M. 2007. Ape gestures and language evolution. Proceedings of the
National Academy of Sciences 104: 8184–8189.
Savage-Rumbaugh, E.S., Rumbaugh, D.M. and Boysen, S. 1978. “Symbolic communication
between two chimpanzees.” Science 201: 641–644.
Savage-Rumbaugh, S., Williams, S.L., Furuichi, T. and Kano, T. 1996. “Language perceived:
Paniscus branches out.” In Great Ape Societies, W.C. McGrew, L. F. Marchant & T. Nishida
(eds), 173–184. Cambridge: Cambridge University Press.
Savage-Rumbaugh, S., Shanker, S.G. and Taylor, T.J. 1998. Apes, Language, and the Human
Mind. New York: Oxford University Press.
Sayers, K. & Lovejoy, C.O. 2008. “The chimpanzee has no clothes: A critical examination of Pan
troglodytes in models of human evolution.” Current Anthropology 49: 87–114.
Slocombe, K.E. & Zuberbühler, K. 2005. “Functionally referential communication in a chim-
panzee.” Current Biology 15: 1779–1784.
Slocombe, K.E. & Zuberbühler, K. 2007. Chimpanzees modify recruitment screams as a function
of audience composition. Proceedings of the National Academy of Sciences 104: 17228–17233.
Taglialatela, J.P., Russell, J. L, Schaeffer, J.A. & Hopkins, W.D. 2008. “Communicative signalling
activates Broca’s homolog in chimpanzees.” Current Biology 18: 343–348.
Townsend, S.W., Deschner, T. & Zuberbühler, K. 2008. “Female chimpanzees use copulation
calls flexibly to prevent social competition.” PLoS ONE 3 (6): e2431. doi:10.1371/journal.
pone.0002431.
von Frisch, K. 1927. (1954). The Dancing Bees. London: Methuen. (English translation).
Washburn, S.L. & Benedict, B. 1979. “Non-human primate culture.” Man 14: 163–164.
 William C. McGrew

Wich, S.A. & de Vries, H. 2006. Male monkeys remember which group members have given
alarm calls. Proceedings of the Royal Society B 273: 735–740.
Wich, S.A., Swartz, K.B., Hardus, M.E., Lameira, A.R., Stromberg, E. & Shumaker, R.W. 2008. “A
case of spontaneous acquisition of a human sound by an orang-utan.” Primates 50: 56–64.
Zuberbühler K. 2006. “Language evolution: The origin of meaning in primates.” Current Biology
16 (4): R123–125.
 chapter 2

The evolution and the rise of human language

Carry the baby

E. Sue Savage-Rumbaugh & William M. Fields

Great Ape Trust/Simpson College

Is it the case that a complete psyche discontinuity exists between ape and man,
as a function of human consciousness and human grammatical abilities? The
genetic evidence makes this seem unlikely. We offer an alternative explanation,
based on the plasticity of neuronal development and the discontinuity between
infant clinging and infant carriage in ape and human. Human infants not only
fail to cling, they display innate motor patterns of rotational hand-waving and leg
kicking. These patterns (absent in ape infants) prevent clinging. In their place,
ape infants display an elaboration of the moro-reflex, which enables them to cling
with all four limbs shortly after birth. The absence of clinging in the human infant
is not due to loss of hair, but to the presence of motor patterns incompatible with
clinging. The neuronal- developmental consequences of these contrasting innate
patterns form the substrate for the emergence of human/ape differences.

1. Introduction

Although the classic man/animal dichotomy continues to undergird Western

scientific thought, evidence from molecular genetics reveals that humans are
more closely related to chimpanzees and bonobos than scientists could have
­surmised, from anatomical evidence, a few decades ago. The degree of DNA
­similarity between humans, bonobos and chimpanzees now classifies us as
­sibling species. This means that, as strange as it seems, we are more closely related
to each other than either of us is, to a gorilla (Wildman et al. 2002). The behav-
ioural ­discrepancies, and to a lesser degree the morphological ones, are larger
than DNA differences readily explain. DNA differences of similar magnitude
between other closely related ­species produce only minor changes in anatomi-
cal and cognitive traits. These puzzling facts have led biologists, psychologists,
­primatologists, anthropologists and philosophers to assert that, at some point
in the Homo lineage, a total discontinuity, at the psyche level, emerged between
 E. Sue Savage-Rumbaugh & William M. Fields

ape and man. The epicenter of this discontinuity is taken to be the appearance
of human language. Consequently great effort has been devoted to determining
both the evolutionary and the genetic basis of language. Concomitant with this
search are the following implicit assumptions regarding language: (a) that it is
absent in all other life forms; (b) that it allows for the appearance of reason, self-
reflection and moral agency; (c) that it makes possible complex planning; and
(d) that material artefacts in the archaeological record which indicate art (such
as ochre, markings that have the appearance of intentional symbols, carvings,
etc.) are coincident with, as well as sufficient and necessary evidence for, the
appearance of human language.
These genetic and philosophical issues will be considered below, in view of
the following established facts: (a) bonobos reared in a complex bispecies Pan/
Homo environment acquire, through observation, most components of human
language; (b) they demonstrate complex self-reflection; (c) they are able to reason;
(d) they exhibit moral agency; and (e) they comment upon the past, the future and
events displaced in time and space. Importantly, they display these abilities in ways
that are distinctly nonhuman, because they are reared – not in a human world as
human children – but in a Pan/Homo world (Fields et al. 2007). Their world has
provided an enculturation experience unique in evolutionary history. From the
cultural fusion produced by the diverse Pan and Homo worlds around them, these
bonobos have developed utterances which seem to function as a pigeon language
(Savage-Rumbaugh et al. 2004). It is precisely the nonhuman categories and quali-
ties of this language that render it difficult for un-acculturated humans to under-
stand and accept the validity and the reality of their expressions. It is also these
nonhuman qualities that lead to subtle differences in the Pan/Homo construction
of moral agency, but not to the absence of moral agency, as often assumed. Fol-
lowing human cultural norms, these bonobos make judgments regarding moral
agency; but following bonobo cultural norms, they reliably take the welfare of the
group into account. For example, if the caloric intake of some bonobos is restricted
by staff at a facility, the other bonobos will hide their own food to give to those
who are restricted when the staff is not watching.
The Great Ape Trust bonobos represent the only group of apes reared in a
Pan/Homo world. Thus, it is logical to expect their linguistic and cultural abilities
to differ from those of other living apes, captive or wild. The recurrent failures
to find simple protohuman language skills in other apes (such as pointing) are
understandable. The deficiency of nonverbal skills in other groups is a direct func-
tion of rearing. Other apes have been reared (a) as human children (that is, in a
completely human world); (b) as zoological specimens (where minimal cultural
information is transmitted because apes are relocated for genetic reasons); or (c)
as free-ranging apes. None of these rearing environments would lead to the mental
 Chapter 2. The evolution and the rise of human language 

construction of a Pan/Homo world, as experienced and as expressed by Kanzi and

his family. Enculturation into a human world would lack the rich array of rapid
sounds, movements and physical experiences present in a bonobo world. These
presumably serve an important role in initiating and guiding the development
of both the neurological “software” and “hardware” of the infant bonobo brain,
enabling it to move toward its full functional capacity. The bicultural rearing of
Kanzi and Panbanisha has affected their anatomical structure and their behaviour
patterns. They are neither bonobo, nor human. They are neither protohuman, nor
childlike. They are bicultural. They can, and do, shift strongly from one pole in
their bicultural world, to another, depending upon their environmental circum-
stances. The language they comprehend and employ is influenced by the culture
Matata brought with her from Wamba.
The complex behaviours of great apes suggest that nonhuman forms of
language are waiting to be revealed among free-ranging groups. They have likely
gone unnoticed because scientists have searched only for utterances that possess
a distinctly human flavour. That is, they have searched for human phonemes,
human syntax, human sensibilities toward object names, a human sense of self-
agency, a human tendency to evaluate the truth-value of propositional statements
and a human kind of vocal fluency that reflects the safety afforded by a home base
(Bickerton 2009). Were scientists to search instead for a quiet, symbolic, highly
coded system, with intense cognitive loading, their results would, and will in the
future, be different (Savage-Rumbaugh et al. 1996).
We suggest that human language is not special because of syntax or semantics,
Lana demonstrated both abilities by 4 years of age (Savage-Rumbaugh & Fields, in
prep). Human language is special because it portrays the world through linguistic
devices that insist upon a fundamental distinction between I and me. These lin-
guistic devices cause the human speaker to bifurcate his or her self-agency into two
roles, that of the doer and the observer. Both of these conscious entities eventually
come to co-reside in the adult human mind, at all times. This bifurcation of human
consciousness into an ‘I’ – who becomes a separate entity from the ‘me’ – repre-
sents the quintessential trait we call humanness. Through the structural vehicle of
human language, our particularly human flavor of consciousness is thus passed on
anew to each generation. Consequently, it is the structure of our language which
­significantly defines and delimits that which we identify as human conscious-
ness, and that which we take to be distinct form animal consciousness. All human
­languages and all human cultures seem to encode this bifurcation between the ‘I’
and the ‘me’, though many indigenous cultures place far less emphasis on it than do
modern cultures (Heider 2007, personal observation).
In this bifurcation of consciousness resides the ability of a postulated
observer – living in the same body as the active doer, to self-reflect upon his
 E. Sue Savage-Rumbaugh & William M. Fields

or her own actions, as though they were being performed by another. Out of
this ever-present bifurcation and ensuing constant self-reflection arises the emer-
gent ability (followed by a nascent sense of responsibility) to make judgments
regarding the intent and the normativity of all of one’s own actions, as well as
­judgments about the actions of others. Also out of this ever-present bifurcation
of the self, arises the sense of freewill. For as the observer begins to consider all
the actions the doer might have done or might do – and thus to look forward and
backward in time – that which might be done takes on a living visualized form
in the consciousness of the observer. Thus does the self-reflective observer begin
to assume increasingly greater control over the active (and sometimes thought-
less) doer. Nonetheless, the doer cannot and does not go away easily, for most
immediate action must occur without reflection, as the speed at which we engage
our social and physical world demands a quickness that the reflective observer
cannot match. The reflective mode constructs its own version of space and time
(Castro-Tejerina & Rosa 2007; Calvin 1996, 2000). As a result, we may find we
have considerable difficulty in explaining our actions, even to ourselves. However,
even when we encounter anomalous circumstances, like those faced by split brain
patients, in which the left hemisphere does not know what the right h ­ emisphere
sees, the observer is nonetheless able to deftly confabulate an acceptable justifica-
tion for the action of the doer (Slunecko & Hengl 2007; G ­ illespie 2007), though
it be completely false.
In addition, whenever we enter a state of hypnosis or heightened ­suggestibility,
we abdicate our state of bifurcated consciousness, and give the ability of the
observer (to evaluate and reflect) away to the hypnotist, or to others around us.
As we do so, our perceived freewill vanishes, not by design or desire, but rather
through our willingness to move into a conscious state which absolves us of the
need to self-reflect, to choose our own course action, or to justify to ourselves and
to others, the logical and normative (or moral) reasons for our actions.
It is the structural nature of the linguistic separation of the I from the me,
­coupled with the recursion embedded in human language, which makes possible
the construction of questions such as “Why did I do that?” In linguistic construc-
tions, the person raising the question is the ‘I’ and the person committing the action
is viewed as the ‘me’, but it is still one person. This is a fundamental structural
property of human language, and it is from this structural component of human
­language that the bifurcation of conscious into the primary-self and the reflective-
self arises. This becomes the child’s way of perceiving causality as he or she acquires
the social vehicle of language. Thus, does the underlying grammatical structure of
human language frame our perception of causal reality. Or to put it more pointedly,
the personal ego is constructed as we learn to speak because our language requires
 Chapter 2. The evolution and the rise of human language 

the construction of a personal ego. Since our language, once acquired, becomes our
main vehicle of thought and ­reason, we have no ready means of conceptualizing
non-ego based realities and/or ­languages, and we cannot recall what our life and/or
reality was like prior to language acquisition.
Previous failures to find language in free-living apes, represent a failure to
locate a specifically human language, that is one with recursive structure and the
primary I/me division. Locked into our own perceptual frameworks, we fail to
realize that our language is directly derived from, bounded by, and enmeshed
within, our particular infant rearing patterns, our object gathering and arranging
patterns and our home base spatial mapping patterns. These patterns define the
basic categories of our world and our linguistic expressions of that world. They
impose particular qualities upon the neurological substrate of our infants in ways
we have not previously cognized. Yet the neural substrate that defines language is
open to virtually any kind of organizational mapping, it only becomes defined to
certain functions by the narrow constraints of human language.
Rearing behaviours are heavily cultural in both us and nonhuman apes.
Therefore, human-like rearing patterns could quickly be brought into play, in
living bonobos, with the appearance of infants less able to cling. Bonobos are
prone to adopt bipedal gaits, both in the trees and on the ground. Infants less able
to cling would derive rapid morphological adaptation for stable bipedalism, social
reorganization with an increased dependence on the role of the male in child
rearing, changes in diet and foraging patterns that would necessitate decreased
dependence on arboreal feeding (because infants who don’t cling fall out of trees)
and possibly the eventual adoption of a ground-based life style with multiple
home-bases and inverted nests for protection from rain and predators. Bonobos,
in contrast to chimpanzees, already display tendencies in these directions (Kano
1992; Savage-Rumbaugh et al. 1996).
The discovery of Ardipithecus suggests that bipedalism goes back further
than previously thought (White et al. 2009). All apes have forms of locomotion
distinctly unlike those of monkeys. Knuckle-walking is not the obvious “next
step” away from the quadruped gait of monkeys, who employ the inner surface
of their fingers, which are splayed backwards during travel. Gibbons (lesser apes
who are presumed to be the first to have evolved away from the monkey lineage)
are completely bipedal when they move on the ground. Yet they can travel only
short distances before they must stop and rest and they never employ their hands
for walking. Orangutans walk on their fists. Like the bipedal gibbon, they cannot
travel long distances on the ground and they never employ their fingers for walk-
ing. They travel too slowly on the ground to avoid predators, and are mostly arbo-
real in the wild, unless they are reintroduced to the wild following captive rearing.
 E. Sue Savage-Rumbaugh & William M. Fields

Only the African apes that have become specialized in the art of knuckle-walking,
are able to travel long distances on the ground, with some speed.
Bonobo and chimpanzee infants are not initially adapted to walk upon their
knuckles. The changes in the hand and wrist, required to support knuckle-walking,
appear only with use. Some chimpanzee and bonobo infants try to become bipedal
during the first few years of life. Wild bonobos walk bipedally in clearings, when
travelling arboreally on large branches, in swamps and when the ground is wet.
When ancient environmental conditions required a bipedal ape to travel farther
and farther on the ground with some degree of rapidity, the ability of the infant
to cling to a bipedal mother, without losing its grip, rapidly became the limiting
factor for survival, not the efficiency of the gait. Any bipedal ape that could bend
over, walk on its knuckles and employ its thighs to aid an infant trying to cling,
would be able to travel further and faster. But if the infant were supported by the
mother’s hands and her bipedal stride became proficient, then travel distance
would be limited only by the ability of the mother, or others who might carry
the infant, not by the strength of the infant to cling, with the intermittent aid of
the mother’s thighs (Savage-Rumbaugh 1994).
If knuckle-walking and many other Homo traits are derived from a common
bipedal ancestor, then the relationship we share with chimpanzees and bonobos is
distinct from that which we share with other apes. Our DNA reveals that bonobos,
chimpanzees and humans are sibling species; and more closely related to each other
than to any other apes. At some point we must question whether it is scientifically
appropriate to classify bonobos and chimpanzees as animals, and ourselves alone
as “human.” This classification, once seemingly correct, does not fit comfortably
with the new genetic knowledge. Our seeming differences may reflect the special
lifestyles we adopted in the relatively recent past, and these lifestyles, in turn, affect
our gene expression profiles. Most scientists view chimpanzees and bonobos as
“relics” of an earlier time, long before traits of humanness appeared. We suggest
that this view is the product of culturally biased ways of thinking about human/
ape differences. Modern life is so far removed from life in the forest that scientists
tend to overlook the deeper kinship, hidden by differences in gene expression.
Human/bonobo rearing differences begin before birth and are a partial function
of the epigenetic effects of arboreal locomotion and its effect on pre and postna-
tal gene expression. In coming face-to-face with bonobos or chimpanzees while
constructing shared linguistic lives on a daily basis, the superficial anatomical dif-
ferences become meaningless, and we are left with the reality of similar minds
encountering one another on essentially equal footing. When social interchanges
takes place at this level, the human/animal boundary is an outmoded category.
In our reality of shared social consciousness, members of both groups constantly
experience a linguistically co-constructed continuity and flow of life.
 Chapter 2. The evolution and the rise of human language 

2. Biological humanness

The biological relationship between human and chimpanzee DNA, coupled with
the chromosomal fusion or fission events that have occurred in the human/ape
lineage, imply that factors other than selection for large brains, upright posture,
hunting, throwing, tool manufacture, and the ability to plan for the future resulted
in the appearance of the Homo lineage. The DNA data reveal that epigenetic ­factors,
and seemingly random chromosomal rearrangements, affected and are continuing
to affect the Pan and Homo lines in ways yet to be understood. For example, it is
puzzling that the human genome is far less variable than that of ­living apes. This
implies that apes evolved rapidly while Homo remained stable, and that all current
living human beings descended from a relatively small population (Alon et al. 2009;
Caswell et al. 2008).
When the chimpanzee genomes was first sequenced, it was thought that the
greater number of duplications, inversions, and copy-number variations found,
when better studied, would elucidate what made chimpanzees and bonobos
different from ourselves. The opposite occurred. The more human genomes that
have been sequenced, the more it has become apparent that the original view that
only .01% of the human genome differing among humans was incorrect. Humans
have copy number variations, duplications, inversions single nucleotide differences
that vary from person to person to a considerable degree. Only as genomes of more
apes and more humans are sequenced will anything like an accurate picture of
evolution have the potential to emerge.
Reconciling the different lines of data produced by genetics and the fossil
record is proving increasingly difficult with each new find in either area. The rise
in discrepancies suggests that a complete revision of our view of human origins is
drawing near. Let us take, by way of example, our most prized human trait, the large
brain. According to interpretations of the fossil record, an increase in cranial size has
been the major dimension of change and the driving factor of selection and specia-
tion toward humanness. But DNA data suggest that the genes which have changed
the fastest since the human/chimpanzee split are ones that code for inflammatory
responses and cell proliferation (Perry et al. 2008), not brain size. Chimpanzees
have improved their ability to resist microbial diseases and pathogenic organisms.
Humans, on the other hand, have increased their susceptibility to cancer, an odd
adaptation unless, as some speculate, by allowing apoptosis we are enabling the
brain to grow larger as well. In addition, there are segments of the human genome
that seem to be limiting change, while similar chimpanzee segments are continuing
to evolve. Science cannot, at this time, tell us how precisely the genome is accessed
or what kind of form will be created, as epigenetic factors act upon development
from meiosis forward. Fossil data definitively reveal that many species, displaying
 E. Sue Savage-Rumbaugh & William M. Fields

characteristics of both apes and humans, existed in the past but the clocks utilized
by either approach are subject to error.
On the bases of a genome wide comparison of human and chimpanzee DNA,
Cheng et al. (2005) suggest that most of the asymmetrical increase of duplicated
DNA in the chimpanzee lineage has emerged as a mechanistic consequence of
changes in the chromosome structure and is not due to selection. Most of the
difference between human DNA and chimpanzee DNA is due to duplications. This
view is echoed by Feuk et al. (2005) who note that the recent genetic evidence on
deletions, inversions and copy number variations should ‘cause people to rethink
their ideas about how species evolved’ (Howard Hughes Medical Institute, 2005).
The characteristics which taxonomists have traditionally employed as indica-
tive of as morphological speciation, do not always correspond to the genetic
differentiation being discovered as the genomes of ever more species become
sequenced. Species which appear to have different morphologies but similar DNA,
express their DNA differently, due to epigenetic factors. Many of these factors are
not yet known, but they can include heat shock, population size, magnetic fields,
dietary restrictions, and cultural variables. Epigenetic variables can cause changes
in gene expression in a single generation. Copy number variations, deletions,
insertions and transposons can similarly occur in a single generation, and at high
rates. This ever-mushrooming new information regarding the mechanisms under-
lying genetic selection and expression are quietly rewriting the Darwinian view of
evolution and ‘natural selection.’ While there is no doubt that selection can and
has occurred, it is no longer clear that natural selection is playing the major role in
morphological change across time.
Darwin developed the hypothesis of natural selection, based on what had
been achieved through the selective breeding of dogs and other domestic animals.
He extended the idea of intentional selection to include what nature, operating
blindly, could produce and reasoned that an intelligent designer was not required
for a similar process of selection to operate. All that was required was a sufficiently
constant criteria, held in place long enough by any means, to produce change
across time (for example larger beaks, thinner beaks, thicker beaks, were the result
of feeding constraints and food preferences of various species of bird, etc.). Dar-
win extended this principle much further, suggesting that, given sufficient time,
these cumulative changes would eventually lead to speciation events, brought
about by selective pressures acting upon any new traits that proved valuable to
the individual. It is remarkable that Darwin saw these kinds of connections even
before there was an understanding of genetics. It is more remarkable that his views
have stood the test of time, even as the field of genetics evolved.
But as science increasingly documents the external forces acting upon the
genome, the theory of natural selection will face major challenges. To maintain
 Chapter 2. The evolution and the rise of human language 

that environmental events are both (a) causing genes to express particular traits
and (b) also selecting for the genes that carry those traits, is not a theoretically
viable concept. If flexible gene expression is a feature of selection over the long
dure, then the organism would benefit by keeping its genome relatively intact
across vast changes in geological time by simply coming up with a way to access
its potential differently in new conditions. If the organism adopted such a
strategy the genome would function more like a tool-kit than a map, in that it
would contain an organism’s building blocks which could be employed in many
different ways, depending upon the environmental conditions in place at the time.
Some organisms with large genomes, such as corn, appear to have adopted the
strategy of utilizing a very large genome to cope with environmental variation.
Other organisms, with very small genomes seem to have acquired the ability to
use the genes they have in new ways by adding epigenetic information. The latter
process is not a matter of selection it is an intelligent biological response to rapid
environmental change.
Support for this view comes from the observation that large-scale human
alterations of the natural environment can result in extremely rapid changes in
species characteristics. Unlike intentional selection, these events, such as habitat
homogenization, are accidental, but nonetheless affect living organisms across
the population. For example, in response to the building of dams, salmon have
begun wintering over rather than migrating to the sea and their morphology has
changed as well, in historical time. Also, when farmed salmon breed with wild
salmon, hybridization interacts with the environment, in some manner other than
selection, to produce a gene expression profile in the offspring which differs from
both parents, more than the parents differ from each other. Such observations
reveal that organisms have ways of responding to their environments, through gene
expression, that are not the products of population selection. These mechanisms
allow for rapid adaptive response to broad-scale habitat change. Traditionally, any
genetic change not due to population selection has been attributed to random
mutations. These ‘mutations’ theoretically became fixed in a population, only
when they had a special survival value.
Often overlooked is the simple fact that the lifestyle choices of any mobile
organism determine what it will need its genes to do. The assumption that genes
can somehow unwittingly determine the organism’s lifestyle choices, while alluring,
goes beyond the current biological explication of what genes actually do. While
correlational data suggest the existence of genes for everything from ‘shopping
behaviour’ to ‘language;’ the simple fact is that our current understanding of genes
is limited to the proteins they code for and the developmental pathways that turn
them off and on. We have no way to determine, given current scientific methods,
how any gene would or could code for the complex behaviours represented in
 E. Sue Savage-Rumbaugh & William M. Fields

the daily lifestyle choice made by primates. Correlational data which suggest that
genes determine these kinds of complex behaviours look only at the relationship
between a portion of the genome and behaviour, they do not explain anything at
a mechanistic level.
The impact of the ‘new genetics’ upon our understanding of human origins,
is just beginning to emerge. With the aid of modern molecular phylogenetics, we
currently trace the origin of Homo sapiens sapiens to central Africa between 150,000
and 200,000 years ago (Liu et al. 2006). We also believe that our first ancestors
moved out of Africa and into Asia and Europe between 60,000 and 80,000 years
ago and that probably only a small group left Africa (possibly as few as 150 people)
(Liu et al. 2006; Mancia et al. 2007). This suggests that the Homo erectus and Homo
habilis populations found outside Africa are not direct ancestors of modern humans
living in those locations today. After Homo sapiens arrived on the scene, more
than 80,000 years elapsed (the time period between the anatomical appearance of
the species we call Homo sapiens sapiens and the appearance of behaviours which
are clearly associated with symbolic processes) before anything is found in the
archaeological record to clearly indicate symbolic ability (Lock 2000; Mithen 1996;
Yellen et al. 1995). If we trust the archaeological record and assume that this gap is
real, rather than the result of inadequate preservation (or current lack of discovery
of appropriate remains), then we must conclude that modern human language
and symbolic behaviour arose between 40,000 and 100,000 years ago. Before that
time, no examples of representational art, fossilized forms of bodily decoration
(such as ochre, beadwork, clothing, etc.), hearths for cooked food or indications
of widespread trading have been identified, at sites associated with Homo sapien
remains. Thus, for a significant period of our existence, we probably lived in forest
conditions similar to those inhabited by modern day bonobos. If so, we must have
employed a dramatically distinct lifestyle from that which was to emerge later on.
For this reason, the designation Homo symbolicus has arisen. Its usage implies
the arrival of a unique kind of human entity on the scene – one which differed
behaviourally, but not anatomically (at least as far as skeleton remains indicated)
from the humans who preceded it.
It has been argued the change began with a single point mutation in one gene,
the FoxP2, which enabled the appearance of human language (Enard et al. 2002;
Varga-Khadem et al. 2005). When the FOXP2 gene undergoes mutation in modern
humans, we manifest a language deficit known as developmental verbal dyspraxia
(Vargah-Khadem et al. 2005). This is a disruption of the motor coordination and
control required not for speech per se, but for fluent speech. Interestingly this
deficit leaves cognition relatively intact. Mice that are genetically engineered to
contain a copy of the human version of FOXP2 turn out to be more vocal than
normal mice. They also display increased synaptic plasticity and dendrite length
in the basal ganglia (Enard et al. 2009; Lieberman 2009) supporting the view that
 Chapter 2. The evolution and the rise of human language 

this gene acts to increase the tendency to make vocal noise. Whether or not this
‘noise’ becomes language depends upon many other factors. Further support for
this hypothesis is provided by the discovery that the up-regulation of FOXP2 is
related to song acquisition in birds (Haesler et al. 2004).
The chimpanzee version of the FOXP2 gene differs from the human version
by only two amino acids. Such a small difference lends credence to the view that
humans may have suddenly gained fluent speech (and thus become Homo sym-
bolicus) with the incorporation of a single point mutation (Enard et al. 2002).
However expression of the FOXP2 gene is regulated by other genes (Carroll
2005; King & Wilson 1975; Scharff & White 2004), which complicates the picture
because we still do not understand how gene regulation leads to many human/
ape differences. Nonetheless, the fact that such a small mutation could alter vocal
fluency implies that the linguistic ability of chimpanzees and bonobos has prob-
ably been underestimated, because their lack of speech possibly results from a
motor inhibition of vocal behaviour (influenced by FOXP2), rather than a cogni-
tive deficit as often assumed. It is logical that their vocal fluency would be kept in
check by the high rate of infant predation they experience in the wild. Were this
constraint lifted, as humans lifted it with the construction of home bases, infant
apes that vocalized frequently and loudly could survive.
A troubling aspect of the Pan-Homo fossil record has always been the lack
of fossil apes. This has caused White et al. (2009) to suggest that all modern apes
are relics, cul de sacs of an earlier period that existed more or less unchanged for
more than 6.5 million years. However, if this is so, then we would expect ape DNA
to have changed far less than ours as well, but exactly the opposite is the case
(Bakewell et al. 2007). Ape DNA has altered nearly twice as rapidly as ours, sug-
gesting that they, not we, are the ones who have undergone environmental pressure
for rapid change. But if White et al. (2009) are correct, then our ancestral form was
much more apelike – and it is we, not they, who have changed morphologically
and behaviourally. If we accept White’s view that all previous “ape-like” beings are
protohominins not proto-apes, our DNA should have altered four or five times
as rapidly as chimpanzee DNA. Many species of yeast differ by a greater genetic
degree than do humans and chimpanzee, and most of the differences between
human and chimpanzees lie in noncoding (Pollard 2009).
However it happened, what we do know is that a very special relationship
exists between humans, bonobos, and chimpanzee and humans; for ‘there are
widespread regions of the genome where chimpanzees and bonobos are less
closely related to each other than any of them are to humans’ (Caswell et al. 2008).
These findings do not fit well within in any version of current Darwinian Theory.
This close relationship is manifest not only in our DNA, but also in the structural
and organizational components of that DNA, as expressed in its location on the
chromosomes themselves. The human chromosome number two is hypothesized
 E. Sue Savage-Rumbaugh & William M. Fields

to have resulted from the fusion of chromosomes twelve and thirteen in an ances-
tral ape, reducing the total number of human chromosomes from 48 to 46 (Yunis
& Prakash 1982; Fan et al. 2002), as chromosome two is essentially identical to
chromosomes twelve and thirteen in living chimpanzees. This fusion event, specu-
lated to have occurred 15 mya ago, has resulted in either similar changes or no
changes, in chromosome two for humans and chromosomes twelve and thirteen
in apes and chimpanzees across 15 mya. Given the rapid fossil evolution docu-
mented in just the last 6.5 mya for apes, it seems odd that changes on the chromo-
somes which fused would remain essentially identical in the living representatives
of both species (humans and chimpanzees) for 15 million years.
Inversion coupled with fusion is an uncommon, and typically unstable event,
in primate evolution. The possibility exists that, rather than a fusion event, the
common ancestral form contained 46 chromosomes and that the modern chim-
panzee/bonobo, gorilla and orangutan are the result of at least one and possibly
more chromosomal fission events. If so, this would explain why there are no fossil
apes (McBrearty & Jablonski 2005), and why each of the great apes shares charac-
teristics with humans that the others do not share, for example:

1. orangutans do not walk on their knuckles but African apes do.

2. the teeth of orangutans are well enamelled as are human teeth, but not the
teeth of African apes.
3. only chimpanzee and bonobo females exhibit sexual swellings.
4. only bonobos and humans lack striking sexual dimorphism.
5. only bonobo and human females are sexually receptive throughout their cycle.
6. only chimpanzees, bonobos and humans live in large mixed social-sexual
groups.

Some of these traits, such as knuckle-walking, do not exist elsewhere in the order
Primates. If knuckle-walking was an intermediary stage of locomotion between
monkeys and Homo, it would likely be found in some monkeys, in ­gibbons,
siamangs or orangutans, however, it is not. This makes it more reasonable to
assume that knuckle-walking represents an adaptation away from bipedalism.
Recent reports that chimpanzees in some areas use sticks for walking uphill also
support this view (Ben Beck, personal communication, 2010) as does the recent
report that an adult male gorilla in a zoo in England suddenly began walking
upright.
Another clue to the mysterious origin of modern apes may come from a
chromosomal rearrangement that is not uncommon in the human population, the
trisomy of Down’s syndrome resulting in 47 chromosomes. The morphological and
neurological changes which manifest as Down’s Syndrome include: short stature,
decreased life span, speech and cognitive difficulties, larger than normal span
 Chapter 2. The evolution and the rise of human language 

between the first and second toes, a flat nasal bridge, shorter limbs, and differences
in affective expression. Similar kinds of differences (short stature, decreased life
span, speech and cognitive differences, larger than normal span between the first
and second toes, shorter limbs and differences in affective expression) also exist
between bonobos and humans, though they manifest differently, and no one
would confuse a Down syndrome individual with an ape.
Still, the range of traits affected by Down syndrome reveals that fission
events which result in the formation of two new chromosomes would affect a
broad range of physical characteristics. Whether two of the 48 ape chromosomes
fused to produce 46 human chromosomes, or one of the 46 human chromo-
somes fissioned to produce 48 ape chromosomes, cannot be discerned from the
fossil record, or from DNA. What is known is that once the lines leading to the
African apes and humans began to separate, interbreeding between these lines
continued for nearly 1.2 million years, (Patterson et al. 2009), because the X
chromosome is younger that the rest of the genome. This suggests that either the
differing number of chromosome (46 and 48) had no effect on interbreeding or
that during the period of interbreeding the chromosomal number was the same
and only changed later. If the latter were the case, then the current suggested
date of the fusion at 15mya cannot be correct. If the former is the case, then
interbreeding should currently be possible between chimpanzees and humans.
Additionally, since the interbreeding that took place after the split lasted for
1.2 million years, ostensibly with Pan males interbreeding with Homo females,
(since the X chromosome is younger than the rest of the genome), the question
of how interbreeding could have occurred for 1.2 million years, between indi-
viduals with different numbers of chromosomes has to be raised.
Should modern apes be the result of multiple fission events (rather than a
single fusion event) it also becomes easier to understand (a) why their genes might
have evolved at a more rapid rate than our own (to cope with the changes needed
to adapt to forest life), (b) why there are no fossil apes, and (c) why there are anom-
alies in all living apes, such as the presence or absence of knuckle-walking, tooth
enamelisation, female sexual swelling, great sexual dimorphism, blue colouration
of teeth in the orangutan, etc.
Additionally we might predict that if modern apes are derived from a human
ancestral form, they could retain human cognitive characteristics, such as the
capacity for some forms of language and self-awareness, but perhaps have lost the
capacity for fluid control of the tongue and diaphragm. Although Lovejoy (2009)
does not discuss living bonobos, the traits which are said to have placed Ardipithe-
cus on its uniquely human journey (reduced canine size, increased bipedality, and
decreased sexual dimorphism), are all extant in modern day bonobos, indicating
that these features of hominin evolution are highly plastic, subject to epigenetic
modification, and possibly recently derived.
 E. Sue Savage-Rumbaugh & William M. Fields

3. Ape language and ape culture

Whether the explanation for the appearance of Homo symbolicus lies in the FOXP2
gene, the gradual appearance of our large brain, our capacity for joint regard and
imitation, or in some puzzling recombination of chimpanzee genes, it remains the
case that human language undergirds all the ways in which we think about ourselves,
explain ourselves, plan for our future, conduct our science, devise our mathematics,
construct our societies according to the rule of law, attempt to understand the
reasons for our existence, fight our wars, plot against our fellow human beings and
protect our property rights and individually accumulated wealth. Without human
language we could do none of these things. Therefore, it is self-evident that any
serious consideration of human origins must either place ape-language studies at
the epicenter of that stage, or attempt to discredit them.
Robert Yerkes (1929) was among the first to speculate that, given the other
obvious dimensions of chimpanzee intelligence, chimpanzees ought to be able
to acquire human language. His views were put to the test by the work of the
­Kellogg’s in the 30’s with the co­rearing experiment of Donald (their son) and
Gua (his c­himpanzee sister). The results were sufficiently startling as to cause
­Kellogg to state that the answer was clear even before Gua began to speak. The
study was c­ oncluded early when Donald began demonstrating chimpanzee behav-
iour, including delayed language onset, and Gua began demonstrating human
­behaviour, including comprehension of spoken language and increased b ­ ipedalism
(Kellogg & Kellogg 1933).
In the 1970’s breakthroughs arrived with the use of non-speech modalities
which produced the first widely heralded evidence for talking apes. These results
were quickly disputed by claims that the chimpanzees did not know what they
were doing. It was thought that they were producing behaviour that merely “aped
language,” through means of imitation, cueing, conditioned discriminations,
or chaining (Savage-Rumbaugh et al. 2009). But when two young chimpanzees
­(Sherman and Austin) began using symbols to communicate with one another, the
realization arose that inculcating language also engendered properties of behaviour
that altered the way apes communicated with other apes. For example, Sherman and
Austin began employing gestures, glances and joint regard while communicating
with symbols. They also negotiated outcomes and through such means, ­corrected
symbolic mistakes. They shared food and objects calmly and deliberately. They
formed novel two word utterances to communicate with the researchers and with
each other. Their utterances, much like those employed by children, constantly
reflected an understanding of the mutual knowledge shared by speaker and listener
alike; knowledge which is inherent in every linguistic exchange, but which differs
from one context to another ­(Greenfield & ­Savage-Rumbaugh 1984). None of these
 Chapter 2. The evolution and the rise of human language 

behaviours were shaped. Such o ­ bservations awakened the scientific community to

the fact that language was about something more than just symbols and grammar
(Savage-Rumbaugh 1986). They suggested the existence of a deeper nature to
language, one that arises out of the way in which it integrates social interaction
into normative planned behaviour across time. The interpretations Sherman and
Austin attributed readily to each other’s utterances implied the existence of a large
body of shared knowledge about the history, and intent, of their past exchanges.
To equate such behaviours with conditioned discriminations, as critics were
initially prone to do, was to miss the essence of how it is that language builds
new interpretations and creative combinations through shared exchanges across
historical time.
This work was followed by a project designed to rear a bonobo (Kanzi) in a
world lacking didactic language tutoring. In its place, language was constantly
employed for normal communicative ends, as happens with human children.
This world was intentionally composed of both bonobo and human beings. To
rear a bonobo apart from others of its species is to deprive it of the normal social-
ization experiences it requires to become a bonobo. Sherman and Austin’s world
had been composed of apes and human beings, but Sherman and Austin did not
have a chimpanzee mother present. Kanzi’s mother was present and participated
in his rearing alongside human companions who permitted themselves to love
and care for Kanzi in the same way that she did. Kanzi’s mother, Matata, was
born and reared in the Congo. Any language or socio-cultural ways of being
which she acquired while growing up were also components of Kanzi’s world.
No previous ape language investigation attempted to include a mother as part of
the social dynamic of the group, much less a wild-reared ape. This component
of the research has not been replicated to date. Across time, the bonobo group
expanded from Matata and Kanzi to a family of eight. As the size of the group
grew, the complexity of the Pan/Homo world, which was established to rear Kanzi
and Panbanisha as Pan/ Homo bonobos, also began to grow exponentially and
the number of bonobos in the Pan/Homo community began to vastly outnumber
the human members (­ Savage-Rumbaugh et al. 1986; Lyn and Savage-Rumbaugh
2000; Segerdahl et al. 2005; Savage-Rumbaugh et al. 2007).
The inclusion of bonobo infants into the true sociolinguistic dynamic
of the human world as meaningful individuals with expectancies, rights and
­responsibilities rather than as subjects, was sufficient to result in the spontaneous
comprehension of English grammar and the production of a proto-grammar at
the lexical level (Savage-Rumbaugh et al. 1993; Greenfield & Lyn 2007; ­Greenfield
et al. 2008). By nine years of age, they readily decoded novel spoken English
­sentences which included recursions, pronomials, and other syntax dependent
utterances (Can you put the can of coke in the trash can, Give the dog a hotdog)
 E. Sue Savage-Rumbaugh & William M. Fields

as well as concepts of metaphor (for example, Kanzi began to employ, on his own,
the colour term ‘yellow’ for cowardly actions), moral agency, and pretense (Can
you make the toy snake bite the toy dog?), (Lyn et al. 2008; Lyn et al. 2006; Savage-
Rumbaugh et al. 1998). They also communicated representational information to
one another through their own vocal channel on a voluntary basis (Taglialatela
et al. 2003; Savage-Rumbaugh et al. 2004). Since apes are thought to be unable
to vocalize except when the limbic system is emotionally aroused (Zuberbühler
2003), the existence of this ability in the Pan/Homo group implies that human
language, acting as an epigenetic variable, has the capacity to override this
neurological constraint.
This bicultural world of bonobos and human beings, sought to incorporate
as many aspects of free-ranging bonobo culture as possible, therefore spoken
­language, the lexigram keyboard, and daily travel in a 50 acre forest environment
constituted the world and mode of daily life for the experimental group – Kanzi,
Panbanisha, Nathan, and Nyota. As infants and juveniles, these bonobos had no
sense of being caged or confined, and no sense that bonobos and humans were
members of different species. Their group was always composed of bonobos and
humans living, interacting and travelling freely together as a unit in the forest.
Close relatives of both species were always present. Conversations with humans (or
other bonobos) were ubiquitous throughout the day and composed of c­ ontextually
embedded normative gestural, vocal, and lexical communications about an unlim-
ited range of topics directly relevant to all ongoing activities (Benson et al. 2002;
Benson & Greaves 2005; Pederson & Fields 2009; Lyn et al. in press). By ­contrast a
control group, consisting of Tamuli, Neema, Maisha, and Elykia were reared by the
wild born parents, Matata and P-suke. The interactions of the control group with
human beings were limited to those that took place behind wire, and humans and
bonobos were clearly treated as different species in this case.
The experimental and control groups lived in different buildings but bono-
bos of both groups were provided with frequent opportunities to interact and to
form social bonds. Any vocal language which Matata might have brought with
her from the wild was present in her behaviour, and thus available to the entire
group. This may have played a key role in Kanzi’s acquisition of lexigrams and
his understanding of human speech. To the degree that Matata’s vocal world was
symbolic, Kanzi was a member of a bilingual world from birth. Children reared
in bilingual homes pick up both languages. The early observation that Kanzi or
Panbanisha would, when asked to “Tell Matata X or Y” – turn toward Matata and
begin to vocalize – suggested an awareness of, and a competency for, a unique
bilingual world. Upon moving to Iowa, both groups were housed together from
2006 to 2011. Together they were maintained in the manner originally accorded
the control group and most human members of the bicultural group were absent.
The effects of this ­common housing and the separation of the bonobos from the
 Chapter 2. The evolution and the rise of human language 

human members of their group were not investigated. The focus of the research
turned toward determining whether or not students and investigators who were
basically unknown to the bonobos could collect data about them.
The original rearing environment created a cultural group that was distinctly
different from bonobos in zoos or in the wild. These bonobos employed ­referential
gestures (Pederson et al. 2009) demonstrated ToM in the classic linguistic “Sally
Ann” test setting (Savage-Rumbaugh 1997), engaged in conversations about com-
plex events, including those removed in space and time, and were cognizant of
moral rules of engagement and normative action (Savage­-Rumbaugh & Fields
2007). They responded to questions regarding their social life and their captive
conditions, and reported upon past events of their lives (Savage-Rumbaugh 1999;
Savage-Rumbaugh & Fields 2000). They were aware of who they are and mani-
fested internal personal and historical narratives of their lives. They functioned
as moral agents within their world-view and behaved in socially appropriate and
morally responsible ways as adults (Fields 2008).
Even though bonobos comprehend human language, they do not behave as
human beings. Instead, they are constantly blending aspects of human culture with
those of bonobo culture. To take a small example, Panbanisha began to ­regularly
illustrate where she wanted to travel outdoors, by drawing lexigram-like con-
structions on the floor with chalk, then adding a line that went from the ­lexigram
directly toward the door, in the direction to be travelled in the forest. Human
beings drew lexigrams on the floor, and in the wild bonobos drag branches in a
line indicative of their travel plans. Panbanisha combined these two processes into
one – of her own accord.
A high level of linguistic self-control has come only with adulthood. Adulthood
has also brought with it an ever-increasing sense of moral agency and e­ vermore
reliable moral responsibility. Previous ape language studies never extended close
human/ape relationships into adulthood, for fear of physical harm. Thus ape
morality, from the standpoint of language related behaviour, was never investi-
gated. Kanzi, Panbanisha and Nyota not only ever harm those human beings whom
they have known all of their lives, but they protect them without fail. Given that
other researchers have not replicated the Pan/Homo world which produced Kanzi,
Panbanisha and Nyota, they should not expect other bonobos to display the cogni-
tive and linguistic abilities reported for this group. Nor should they be skeptical of
the results or assert that experimenters are over-interpreting the data – until they
are willing to invest effort to determine the adequacy of those inter­pretations from
within the normative framework of a Pan/Homo culture. ­Language always requires
interpretations that take into account cultural categories, common knowledge, and
historical context. Therein lies the power of language. Speakers of any language,
who lack the capacity to make rich interpretations of the utterances of others, are
unable to engage in normal social dialogue.
 E. Sue Savage-Rumbaugh & William M. Fields

A traditional scientific response would be an attempt to replicate the method

which produced this unique group of bonobos, rather than to claim that the
researchers are intentionally fraudulent (Sebeok & Umiker-Sebeok 1980; Wynn
2008); that the bonobos’ linguistic abilities lack some key aspect of human
­language (Tomasello 2008; de Waal 2002); or to confuse the issue with so called
“language” studies that employ didactic training paradigms intentionally designed
to eliminate communication (Matzuzawa 2001; Premack 2004; Povinelli & Vonk
2003; Shumaker & Beck 2003). The eagerness of scholars outside the field to accept
superficial critiques as valid (Pinker 1994; Dunbar 2009), without addressing even
minimal attention to the unique reality of a Pan/Homo world, illustrates the degree
to which human egocentrics views act to maintain the politically correct view of
apes as brutish, dull and incapable of behaviour worthy of serious scientific study
(Lovejoy 2009; Savage-Rumbaugh et al. 2009).

4. M
 aternal infant carriage and interaction as substrate
of human agency

Mammals, apart from primates, do not carry their offspring. Their infants either
remain cached and in contact with siblings until they become mobile (wolves,
lions, rodents, etc.) or they are able to walk within hours after birth (antelope,
elephants, horses, etc.). Primates changed this basic pattern by bearing an infant
that was immobile, but not cached. Their infants survived by clinging to moth-
ers who travel constantly while foraging for fruit in trees. Humans changed the
basic mammalian pattern yet again by bearing infants unable to cling or to remain
immobile. The distinct Homo/Pan styles of infant care (carriage versus clinging)
sets the development of the Homo/Pan infant neurological systems on very dif-
ferent trajectories. As we shall see, minor changes in these initial set points, lead
invariably to salient behavioural differences between humans and apes.
Before birth all mammals begin life as an integral part of the mother’s body.
During pregnancy, every sound the mother hears, every move her body makes
and every stress she experiences impinges upon, and has the potential to change,
the developing embryo within. Truly the baby and the mother are one at this
time. Although we know that epigenetic factors, such as drugs, can result in fetal
malformation, we have yet to understand the myriad of internal and external
prenatal factors that act to set the sensitivities of the nervous system during
development. For example, it has only recently been determined that the language
a baby experiences before birth, affects the pattern of its crying. The pattern of the
infant’s cry also affects the mother and constitutes the first vocal communicative
component of cultural exchange between the dyad (Wermke & Mende 2006;
Wermke et al. 2007).
 Chapter 2. The evolution and the rise of human language 

Bonobo babies are always in full body contact with a mother or relative.
Consequently they initially experience the world, with and through, the actions
of the mother; rather than through their own actions as intentional free-willed
agents. Such is true of human infants as well, but for a much briefer period.
The amount of time human mothers spend carrying babies varies widely from
indigenous to modern cultures. Cultures which extend this period tend toward
greater co-dependence between adults. Trevarthen (1989) has characterized the
experienced consciousness of the infant at this stage as ‘Mommy and I are one’
(Trevarthen & Hubley 1978). Human infants who do not achieve the state of joint
consciousness (as well as those who fail to make a smooth transition out of this
state) often become less dependent (Acquarone 2007).
For the first several weeks of life, bonobo infants are unable to support their
own weight by clinging to the mother’s body for more than five to ten ­minutes.
Bonobo mothers must therefore aid newborn babies by travelling with a crouched
posture whilst supporting the baby with their thighs, or with one hand under as
they move tripedally (Bolser & Savage-Rumbaugh 1989). The fact that a bonobo
baby accounts for 2% or less of the mother’s body weight, makes these patterns
of locomotion possible. However the human infant comprises 8 to 10% of the
mother’s weight (Brakke & Savage-Rumbaugh 1990), and the mother must either
support the baby with both hands or utilize a sling. If bonobo infants were to
have body fat at birth they could not cling and the modes of transport utilized
by their mothers would be untenable. By one month of age bonobo infants (who
remain very thin and lightweight) are able to actively maintain contact with the
mother by clinging on their own for up to 20 to 30 minutes. By 3 months of
age they require only intermittent support as the mother travels. Bonobo infants
cling reflexively from birth. The engagement of the clinging reflex inhibits the
rotational hand-waving and foot kicking, so prevalent in human babies. Thus
bonobo babies do not have any substantial periods of time during which they
move their hands and feet, while observing themselves visually as they lie supine.
Human babies tend to spend lots of time observing the movements of their own
limbs and, as objects are placed in their hands, the causal effects of shaking and
moving objects.
The need to remain with the mother – clinging as she travels – requires
essentially all of the bonobo baby’s attention, as well as its constant mental and
physical effort. It also ensures that the baby encounters its early world primarily
through the lens of its mother’s bodily reactions to the events occurring around
them both. The world of a bonobo baby is one of constant motion, constant sound,
and constant reactivity to ongoing occurrences. In such a world, bonobos infants
receive far less conscious visual attention than would a human baby. They do
not cry or fuss for attention. They rarely vocalize except when separated, or to
accompany sounds produced by their mothers. Mother-infant vocal dialogues do
 E. Sue Savage-Rumbaugh & William M. Fields

not occur. Essentially all of the physical and emotional needs of the infant are
mediated through the vehicle of constant contact.

Figure 1. Gorilla mother with newborn. This gorilla mother (left) is supporting her newborn
infant with both hands, just as would a human mother. However, when she stands, she can no
longer use her hands to support her infant, as do human mothers. Therefore her infant must
quickly become able to support itself by clinging. To be able to do so it must mature rapidly at
first and stay lightweight. Human infants (right) can afford to mature slowly and to add body
fat because they do not need to cling to stay with the mother. The mother will keep them with
her and will carry them as she travels

Figure 2. A human mother lays her infant down to free her hands. The chimpanzee infant
clings to the mother, while being supported by her thighs, leaving the chimpanzee mother’s
hands free
 Chapter 2. The evolution and the rise of human language 

Human infants encounter a radically different trajectory of maternal care.

They are born unable to cling even if their mother had hair. They are too fat, and
too weak. Their heads are too heavy and their necks are not able to support the
head. They have reflex patterns of waving the hands and feet that interfere with
clinging. Because of this the human mother, rather than the baby, must make
every effort to hold the infant or to watch it carefully if she puts it down. The baby
not only fails to support any of its own weight, it manifests behaviours which inter-
fere with clinging, and it cries easily, which can attract the attention of predators.
The added weight of the human infant results from the increase in head size and
body fat, rather than an increase in overall body length. Body fat allows human
infants to regulate their body temperature more effectively than ape infants,
thereby enabling humans to occupy colder climates and to put their infants down.
All human babies, regardless of climate, continue to show rapid increase in body
fat after birth. They can easily afford to do so because they do not support their
weight by clinging. They also need to do so because they are placed in positions
where they are out of contact with the mother’s body and thus must regulate their
body temperature on their own while still immobile. Even African climates can
chill infants during the rainy season. The common view that the human baby is
too “immature” to cling is incorrect. As the baby matures, it never reaches a point
where it can support its own weight while the mother travels. Likewise, the com-
mon view that the ape baby is more mature at birth is also incorrect. Ape babies
do not cling because they are mature. They cling because they come equipped with
a different set of reflexes, a small head and no body fat. If any of these factors are
altered, ape babies do not cling, regardless of their level of ‘maturity.’
Because the human baby does not cling, the mother must decide what to do
with it the moment it arrives. She must know how to support the large head, weak
neck and mobile spine in a manner that fosters survival and she must either know
how to keep the baby quiet or be in the midst of a protective group. Should human
mothers wish to use their hands for something other than supporting the baby,
they must either sit and put the baby in their lap or place it on a flat solid substrate,
or in some object that will support its head and spine. Ape mothers do not have
to ‘think about what the baby needs’ – they need only offer it brief support if they
perceive it starting to fall.
The far-reaching effects of the lack of clinging upon the developing nervous
system have been overlooked. When a non-mobile human infant is placed on a
substrate away from the mother’s body, a very different visual, auditory, tactile,
and kinesthetic world impinges upon it, than when it is in full body contact with
the mother. Instead of looking at a body or a breast, such a baby is suddenly look-
ing at the world around it. Often this includes mother’s face and a mouth which
is moving in synchrony with the sound that is coming to its ears as she speaks
(Figure 3).
 E. Sue Savage-Rumbaugh & William M. Fields

Figure 3. The human mother (left) engages in frequent eye contact with her infant who is
often placed in a position which makes this possible, because it cannot cling. The baby, at
these times, becomes the focus of attention for the mother and the mother becomes the focus
of attention for the baby. The chimpanzee infant (below right) is rarely in a position to focus
its attention on the mother’s eyes because it is clinging to her. It tends to focus its attention
on the aspects of the external world that attract the attention of its mother. Because of these
­differences in carriage patterns, the consciousness of the chimpanzee infant remains in the
“Mommy and I are one” stage for a much longer period than that of the human infant

When human babies develop appropriate species affect, they tend to become
visually entranced with the mouth movements, facial expressions and sounds
of others. They then engage in facial-vocal dialogues with caretakers by two to
three months of age (Stern 1971, 1977; Trevarthen 1989, 1998). These cooing dia-
logues bear strong rhythmic similarities to later true conversations that emerge
between ages one and two. Newborn infants also imitate facial expressions such
as tongue protrusions and smiles from a few days after birth (Meltzoff 1996;
Myowa-Yamakosi et al. 2004). The early onset of these behaviours, suggests they
are being guided by the mirror neuron system, which is also intimately tying
them to their caregivers from birth forward through rhythmic facial/vocal inter-
play (Rizzolatti et al. 2001).
But a human mother must do more than make decisions about how to ­support
and transport a heavy immobile baby. When the mother sets the baby down, even
 Chapter 2. The evolution and the rise of human language 

for a short time – things can happen to the baby that could not happen to her.
Human maternal care requires a continuous conscious awareness and monitoring
of the difference between what is (or could be) happening to the mother herself,
as contrasted with what is (or could be) happening to the baby. To take a simple
example, if a human baby is exposed to sun, rain, smoke, insects, wind or preda-
tors, it can be harmed quickly by events that would be of little consequence to
the mother herself. The human mother must realize that heat from the sun or
insects or predators, which will not harm her, can quickly harm the baby. She must
monitor the baby’s world from the baby’s vantage point, in a way not required for
nonliving objects. The ability of the human mother to create in her mind, a con-
stant scenario of the baby’s needs and points of view is demanding for many new
mothers.

Figure 4. The spine and head of the human infant require support during carriage which
the mother must attend to carefully while the infant is very young because the weight of the
infant is such that it cannot, until more fully developed, manage to keep the head and the
spine aligned easily in all positions. The small head and lack of body fat enable the chimpanzee
infant to manage its own postural adjustments of the head and spine from birth, freeing the
mother’s attention to watch for predators without worrying about the infant

Their feeling about this effort, and the baby’s receptivity to it, affects the
content and style of the linguistic dialogue which then accompanies the mothers
actions toward the baby (Oh, you’re all wet, You are getting sleepy, Now smile, don’t
be fussy, What is bothering you?). Such dialogue carries a quality of “aboutness”
toward the baby and the degree of “aboutness” expressed corresponds closely to
the degree of differentiation taking place between mother and infant (Brigaudiot
et al. 1996; Savage-Rumbaugh 1990).
Clearly, in the order Primates, something unique happened when babies who
could not cling appeared. What are the implications of this for the Pan/Homo
 E. Sue Savage-Rumbaugh & William M. Fields

c­ ulture? Bonobo infants who are carried by human mothers quickly begin to use
their hands for something other than clinging and they are exposed to languaged
comments about their actions from birth forward. They reach for objects and
manipulate them much sooner, and more frequently, than bonobo-only reared
infants. In addition they habituate to being placed on the ground and do not cry
or show distress when bodily contact is broken gently and slowly.

Human invention of support devices, preoccupation with objects and

extension of consciousness enables baby to focus on self

Mother monitors the world

Mother monitors baby’s of her group as first priority,
world as it cannot locomote self and group overlap

Once baby can locomote it

monitors its own world

Baby begins to check mother’s Once it locomotes, the

monitoring, cries for attention,
long before it can locomote.
When mother’s consciousness is
constantly extended to the
baby, fussing is prevented
baby checks mother’s
position, cries only if mother is
not near. Baby does not fuss
as its consciousness is
constantly extended to the mother

Figure 5. Differences in the bodily location of a newborn infant, relative to the mother,
place specific constraints on the role of the mother. They also place specific demands on the
infant. The mother must monitor any infant that is not in contact with her body. She must
also monitor the external surround and any task she is doing. She must shift her attentional
priorities between all three. A human infant that is not held (and that is also immobile)
needs the ability to cry and to alert the mother that it is in pain or danger, in the event that
she has placed her attention on other things. The chimpanzee infant, clinging tightly to its
mother, experiences only the same dangers which might affect her. The ability to make noise
must become available to the infant even before the infant has the cognitive ability to discern
whether or not the situation is one in which a noise might attract a predator and thus prove to
be dangerous for the group. The consciousness of the human infant thus becomes separated to
communicate to the mother information about states that it experiences and she does not

When placed on a solid surface, they orient toward, and respond appropri-
ately to the facial expressions and the rhythmic conversational engagements of
 Chapter 2. The evolution and the rise of human language 

the human mother. In bonding emotionally to a human mother, as well as to a

bonobo mother, they begin a neurodevelopmental trajectory that carries them
irrevocably along an extraordinary bicultural developmental path into human
language. In such a rearing paradigm, human contact alone is not a sufficient
variable. A mutual emotional attachment must take place as well. Without it, the
bonding that is required to enable the infant to visually, and tactilely engage the
rhythms of the human world, fails to take place. Ape infants raised with human
contact in zoological nurseries are not the same beings as apes reared in a bicul-
tural world. Ape infants raised in nurseries or as pets in human homes, typically
manifest autistic behaviours such as rocking and visual detachment (personal
observation). These infants experience human contact and care but lack true
mothering, bonding and nurturing. They carry the scars of cognitive detachment
throughout their lives.
From the different maternal rearing patterns, emerge the styles of conscious-
ness which have long interested philosophers (Kitcher 2006; Lock 1690/1959;
Savage-Rumbaugh et al. 2005). The consciousness of self-as-agent is emphasized
early in ontogeny with the human pattern of rearing. Lengthy en face social engage-
ment, visual/vocal exchanges, object manipulation with freed hands, and attuned
linguistic overlay all emerge when babies do not cling, but are carried instead
(Bullowa 1979; Halliday 1979; Lock 1978). The baby’s attention is focused early
upon the results of its actions toward objects or the effects of its social facial/vocal
displays toward others (Papousek 2007). Instead of ‘self-agency’ – consciousness
of co-agency is emphasized early with the bonobo pattern of rearing. By the time
the bonobo infant is locomoting on its own and beginning to develop an aware-
ness of self-agency, its primary focus of attention has already become fixed upon
group cohesion and group movement through the terrain and it is highly attentive
to the slightest alteration in that terrain. When the wild-reared bonobo baby begins
to manipulate objects in an exploratory manner, the cognitive information that is
accrued is immediately linked into the previously mapped knowledge of the land-
scape and how that landscape can be utilized in the service of survival. Thus the
solid dense gnarled knobby handled sticks which bonobos employ for digging are
not carried from place-to-place, but hidden at useful locations for later use (per-
sonal observation, first author 1996). Wild bonobos are very quiet on the ground
and extremely noisy in the tree-tops. Bonobo infants are unable to reliably discern
which situations are safe for sound and which are not, thus they must be quiet at all
times and are unable to cry and/or to make voluntary sounds. The ability to ‘speak
at will’ does not appear in bonobos until adolescence or adulthood, therefore this
skill, and that of object manipulation, fails to undergo the great freedom of expres-
sion during the critical developmental period of childhood. The development of a
bonobo baby is constrained by the need to cling and to be quiet.
 E. Sue Savage-Rumbaugh & William M. Fields

Ape infants quickly loose their insistance on clinging if they are

always supported and put down only when comfortable

They begin to be self-aware and

to attend to the gaze of the mother
and objects around them before
they are able to locomote

Figure 6. When an ape infant is reared, as though it were a human infant, it begins to
monitor its world and to vocalize for attention. It loses its seemingly constant need to cling and
its attention begins to focus on objects at a much earlier age, while it is still relatively immobile.
It is often in a position to engage in eye contact and facial gazing and the human mother’s face
becomes the focus of its attention and the infant’s face and eyes become the focus of the human
mother’s attention enabling the emergence of vocal/facial dialogues. Apes raised as pets, or in
entertainment do not experience this kind of human maternal care. Additionally once apes
reared as pets are able to cling they are often not supported by human caretakers. They are
allowed to cling and support themselves, thereby eliminating, during a critical developmental
period, the infant’s developing manual object manipulative skills, and vocal/facial exchanges,
thereby forcing the chimpanzee to follow the same developmental path it would follow if it
were being reared by a chimpanzee mother. Thus large differences in vocal/facial dialogue,
object manipulatory skills and emotional development emerge between so called “human
reared” apes

The human child’s attention is, by contrast, drawn toward object play and
observation of social interactions. Its hands are free to manipulate objects even
before it can sit up and in so doing its actions become landscape independent,
and its visual and auditory system becomes human. Mothers provide babies with
objects, such as rattles, that they can grasp and move, enabling babies to experience
their own capacity to produce noise and movement, at an early age. Mothers over-
lay the baby’s world with maternal comments which cast the baby in the role of self
as agent and scaffold causal interactions between the baby and themselves (Wolff
1987). Even actions as simple as lifting the baby, offering it food and ­changing its
position are overlaid with causal language (Why are you fussing? Can you hold
 Chapter 2. The evolution and the rise of human language 

still, Do you want some of this?). The highly patterned repetitive interactions of
social exchange imprint the infant brain and form a trajectory that manifests (in
its adult form) as the sociolinguistic interchange patterns of the group (Lock 2000;
Lock & Peters 1996; Rogoff 1990; Vandell & Wilson 1988).
Most behaviours said to characterize the differences between humans and
apes derive directly from these early mother-infant rearing patterns and the
ensuing alterations in the hands, the gaze patterns, the locomotor patterns, the
facial expressions, and the conversational style (Thibault 2004; Taglialatela et al.
2004). Because they occur so early in the development of mind, they set the form
and tone for all future social-physical interaction in distinct ways. Sociality is a
prerequisite for the emergence of language (Steels 2009), and the nature of that
sociality defines the nature of the language which arises and rides upon it. The
central dimension of all human solutions to infant rearing is that they are actions
extended to the baby by the caregiver. Because it is not the baby’s job to maintain
contact with the mother, but rather the mother’s job to maintain contact with the
baby, self-agency becomes dimensionally different in ape and man.

5. Self-agency and the duality of consciousness

Most psychologists attribute some level of self-agency to apes, because of their

capacity to recognize themselves in mirrors. We suggest that touching a mark on the
body (the typical test for self recognition) can be accomplished by understanding
that mirrors reflect images. However, awareness of the self, as distinct from being
self-aware, requires a bifurcation of consciousness. When individuals seek out
mirrors to investigate their appearance because they have intentionally self-altered
that appearance, then consciousness has begun to bifurcate into an observer and
a doer. Sherman, Austin, Kanzi, Panzee and Panbanisha all painted their faces,
put on wigs, shawls, and monster masks, and tried to blow bubbles with bubble
gum while using mirrors to check out their altered appearance. Sometimes they
practiced display behaviours by adding fur capes while swaggering in front of
a mirror. Only bonobos and chimpanzees who have achieved some degree of
human language competency display mirror behaviour of this sort (Menzel et al.
1985; Savage-Rumbaugh 1986).
Human language is, after all, a system designed to maximize self-reflection.
Its comprehension and use require increasing bifurcation of consciousness. This
process is a natural extension of the discovery of self-agency which begins with
early human infant care patterns. As the dualistic self-agency begins to operate,
there emerges, within a single brain and body, the capacity to simultaneously sepa-
rate the imaged self into a doer of actions and an observer of those same actions
 E. Sue Savage-Rumbaugh & William M. Fields

(Bates 1990). This duality of consciousness permits the organism to think about
what it is doing, the appearance of its actions, and how the actions will be perceived
by others, at the same time. It is like a seed, which grows slowly, enabling steadily
increased reflective control over the actions of the organism by the observer. This
capacity to reflect inevitably results in a wish that certain actions had not been
taken, either because they did not accomplish the organism’s goals, or because
they could not be justified to others (Lewis 1991, 1995; Henriques 2004). With the
realization that some actions are judged (by one’s companions) as more correct
than others, moral agency obligatorily comes into focus.
When the bifurcation of consciousness is in its initial stages, the doer considers
only the potential outcomes for himself and only for short spans of time ahead.
With increased sociolinguistic experience, the doer and the observer (reflected
linguistically as the ‘I’, who is an agent, and the ‘me’ who is the recipient of the agency
of others) bifurcate more fully, and for longer periods of time (Smiley & Johnson
2006). Finally there arises a sense of permanence to the externally objectified
conscious experience of, and narrative about, the self. Alternative outcomes of
one’s actions can then become recognized as things which are inevitably shared
by the group as a whole not just by the doer (Brownell & Carriger 1990). When
this realization arises in the mind of the doer, it becomes possible to consciously
and reflectively behave in a manner that is constructive or destructive for the
larger social group in which one is enmeshed. Such is the sine quo none of human
intelligence. While other organisms may benefit from altruism, only the human
linguistic organism can consciously practice altruism with the full awareness that
constructive action which benefits the group, inevitably benefits the self. From
this conscious realization, all the greater principles of what is termed ‘religious
awareness’ (for example, “do unto other as you would have them do unto you”)
inevitably arise. Thus the human trajectory is, from birth forward, focused upon
the growth of conscious awareness. The characteristics of the human infant,
demand that it be cared for in ways unique within the order mammlia. All of
its survival needs, when met appropriately, allow for the very early emergence of
self-awareness. Self-awareness is quickly followed by other-awareness, language
and bifurcation of consciousness into the “I” versus the “me” – as well as the “me”
versus “you.” Much of the rest of each human beings life is spent learning how to
resolve this bifurcation.
This capacity for dual consciousness lies at the root of Cartesianism. However,
there is no need to reduce the bifurcation of consciousness to mind/body dualism.
When human consciousness (as an emergent electronic property of the neural sys-
tem) reaches a certain level of awareness of the self as causal agent (accomplished
through the vehicle of human language) it bifurcates and gains the ability to reflect
upon itself. As this occurs, consciousness can be thought of as splitting itself into
 Chapter 2. The evolution and the rise of human language 

two parts, rather like a soap bubble which divides in the middle. One half of the
consciousness bubble specializes in guiding the immediate actions of the organ-
ism and the other half specializes in reflecting upon those actions. The body and
brain are quite competent to engage in many complex behaviours, that to all outer
appearances, seem to be conscious intentional actions, but without any conscious
evaluation of the adequacy of the action, much as occurs in states of dreaming or
hypnosis. For shorter periods of time, we all hand over conscious evaluation of
the adequacy of our actions to others during social encounters and self-agency
becomes blurred.

6. Conclusion

Thus the more advanced properties of human language arise directly from the
simple structural properties required to coordinate and order attention and to
explore self-agency. These behaviours, which begin at birth, order our perceptions
of time and space, our construction of reality, and the construction of our personal
narrative. This organization of consciousness occurs so early that we lack any
awareness that many aspects of ourselves could ever be fashioned differently. There
is sufficient commonality among all human cultures, driven by the exigencies of
caring for our immobile heavy infants, as to enable major dimensions of human
behaviour to self-manifest in a comprehensible fashion across our different
cultures.
Social exchanges occur, however, at a pace too rapid for any extensive
­conscious on-line self-reflection. Consequently, even highly self-reflective beings
depend mainly upon the real-time sociolinguistic rules of engagement which
are culturally co-constructed through social interaction during their childhood
(Gladwell 2005; Wilson 2002). We know very little about the effects of how we
actually interact, and less about what it is we really do during “en face” interactions,
because these skills are acquired as we acquire language and function always in
concert with it, except in written texts. Our means of social interactions are not
merely imprinted upon our brains they become the structural substrate of our
brains. This is as true for bonobos as it is for human beings, but the two species
learn to do this differently. Both species take great infractions of expected rules
of social engagement. This is because the bifurcated mind, once having accepted
a distinction between the person who is acting, and the mind controlling the
intention behind the action, forever after will always interpret the behaviour of
other members of its species as manifestations of intentional acts. Consequently
any lack of conformity to group norms is taken as an intentional affront to the
larger group values (Shotter 1993). Every action must be considered intentional
 E. Sue Savage-Rumbaugh & William M. Fields

and every action must be interpreted within the confines of the larger nomethetical
context. This is the real dividing line we draw between human behaviour and
animal behaviour. Human behaviour is deemed intentional and humans are held
responsible for their acts. Animals are not. Childhood, among human beings, is
about acquiring the knowledge and skills to intentionally control ones behaviour
according to the normative standards of ones culture. Without human language
this is not possible to do. One can however, imagine a language that has grammar
and semanticity, but no I/Me distinction and which requires no explanation or
justification of why one does or says any particular thing. Such a language seems
to be the province of wild bonobos and is, we suggest, delimited by the fact that
they do not “carry the baby.”

References

Acquarone, S. 2007. Signs of Autism in Infants: Recognition and Early Intervention. London:
­Karnac Books.
Alon, K., Mullikin, J.C., Patterson, N., Reich, D. 2009. “Accelerated genetic drift on chromosome
X during the human dispersal out of Africa.” Nature Genetics 41: 6–70.
Bakewell, M., Shi, P. & Zhang, J. 2007. “More genes underwent positive selection in ­chimpanzee
evolution than in human evolution.” Proceedings of the National Academy of Sciences 104:
7489–7494.
Bates, E. 1990. “Language about me and you: pronominal reference and the emerging concept
of self.” In The Self in Transition, C. Cicchetti, and M. Beeghly (eds), 165–182. London: The
University of Chicago Press.
Benson, J.D., Fries, P., Greaves, W.S., Iwamoto, K., Savage-Rumbaugh, S. & Taglialatela, J.P.
2002. “Confrontation and support in bonobo-human discourse.” Functions of Language
91: 1–38.
Benson, J.D. & Greaves, W.S. 2005. Functional Dimensions of Ape-Human Discourse. London:
Equinox.
Bickerton, D. 2009. Adam’s Tongue: How Humans Made Language, How Language Made
Humans. New York: Hill and Wang.
Bolser, L. & Savage-Rumbaugh, E.S. 1989. “Periparturitional behavior of a bonobo.” American
Journal of Primatology 17: 93–103.
Bullowa, M. 1979. Before Speech: The Beginning of Interpersonal Communication. Cambridge:
Cambridge University Press.
Brakke, K. & Savage-Rumbaugh, E.S. 1990. “Comparative motor and manipulatory ­development:
Behavioral growth in infant human and apes.” Poster presented at the 11 Biennial C ­ onference
on Human Development Richmond Virginia.
Brigaudiot, M., Morgenstern, A. & Nicolas, C. 1996. “Guillaume i va pas gagner, c’et d’abord
maman: Genesis of the first person pronoun.” In Children’s Language (9), C. Johnson and J.
Gilbert (eds), 105–116. Mahwah, NJ: Lawrence Erlbaum Associates.
Brownell, C. & Carriger, M. 1990. “Changes in the cooperation and self-other differentiation
during the second year.” Child Development 61: 1146–1174.
Calvin, W.C. 1996. The Cerebral Code: Thinking a Thought in the Mosaics of the Mind. Cambridge:
MIT Press.
 Chapter 2. The evolution and the rise of human language 

Calvin, W.C. 2000. Lingua ex Machina: Reconciling Darwin and Chomsky with the Human Brain.
Cambridge: MIT Press.
Carroll, S.B. 2005. “Evolution at Two Levels: On Genes and Form.” PLoS Biol. 3 (7): e245.
doi:10.1371/journal.pbio.0030245. PMC 1174822. PMID 16000021.
Castro-Tejerina, J. & Rosa, A. 2007. “Psychology within Time: Theorizing about the Making
of Socio-Cultural Psychology.” In The Cambridge Handbook of Sociocultural Psychology,
J. Valsiner and A. Rosa (eds), 62–81. Cambridge: Cambridge University Press.
Caswell, J.L., Mallick, S., Richter, D. J, Neubauer, J., Schirmer, C., Gnerr, S. & Reich, D. 2008.
Analysis of Chimpanzee History Based on Genome Sequence Alignments. PLoS Genet
4 (4): e1000057. doi:10.1371/journal.pgen.1000057.
Cheng, Z., Ventura, M., She, X., Khaitovich, P., Graves, T., Osoegawa, K., Church, D., DeJong,
P., Wilson, R.K., Pääbo, S., Rocchi, M. & Eichler, E.E. 2005. A genome-wide comparison of
recent chimpanzee and human segmental duplications. Nature 437: 88–93.
de Waal, F.B.M. 2002. Tree of Origin: What Primate Behavior Can Tell Us about Human Social
Evolution. Cambridge, MA: Harvard University Press.
Dunbar, R. 2009. “Why only humans have language.” In The Prehistory of Language, R. Botha &
C. Knight (eds), 12–35. Oxford: Oxford University Press.
Enard, W., Przeworski, M., Fisher, S., Lai, C., Wiebe, V., Kitano, T., Monaco, A. & Pääbo
S. 2002. “Molecular evolution of FOXP2, a gene involved in speech and language.” Nature
418 (6900): 869–872.
Enard, W., Gehre, S., Hammerschmidt, K., Holter, S., Blass, T., Somel, M., Bruckner,
M., ­S chreiweis, C., Winter, C., Sohr, R., Becker, L., Wiebe, V., Nickel, B., Giger, T.,
Müller, U., Groszer, M., Adler, T., Aguilar, A., Bolle, I., Calzada-Wack, J., Dalke, C.,
Ehrhardt, N., Favor, J., Fuchs, H., Gailus-Durner, V., Hans, W., Hölzlwimmer, G.,
Javaheri, A., Kalaydjiev, S., Kallnik, M., Kling, E., Kunder, S., Mobrugger, I., Naton,
B., Racz, I., Rathkolb, B., Rozman, J., Schrewe, A., Busch, D., Graw, J., Ivandic, B.,
Klingenspor, M., Klopstock, T., Ollert, M., Quintanilla-Martinez, L., Schulz, H.,
Wolf, E., Wurst, W., Zimmer, A., Fisher, S., ­Morgenstern, R., Arendt, T., Hrabé de
Angelis, M., Fischer, J., Schwarz, J. & Pääbo, S. 2009. “A humanized version of Foxp2
affects Cortico-Basal Ganlia Circuits in Mice.” Cell 137: 961–971.
Fan, Y., Newman, T., Linardopoulou, E. & Trask, B.J. 2002. “Gene content and function of the
ancestral chromosome fusion site in human chromosome 2q13-2q14.1 and paraloguous
regions.” Genome Research 12: 1663–1672.
Feuk, L., MacDonald, J.R., Tang, T., Carson, A.R., Li, M., Rao, G., Khaja, R. & Scherer, S.W.
2005. “Discovery of Human Inversion Polymorphisms by Comparative Analysis of
Human and Chimpanzee DNA Sequence Assemblies.” PLoS Genet 1 (4). doi:10.1371/
journal.pgen.0010056.
Fields, W.M., Segerdahl, P. & Savage-Rumbaugh, E.S. 2007. “The Materical Practices of Ape
Language Research.” In The Cambridge Handbook of Sociocultural Psychology, J. Valsiner
and A. Rosa (eds), 164–202. Cambridge: Cambridge University Press.
Fields, W.M. 2008. “Il Kanzi etnografico contro il Kanzi empirico: sulla distinzione tra “Casa”
e “Laboratorio” nella vita delle antropomorfe acculturate.” In Rivista di Analisi del Testo
Filosofico, Letterario e Figurativo 8, R. Mascialino (ed.), 171–207. Padova: Cleup.
Gillespie, A. 2007. “The Social Basis of Self-Reflection.” In The Cambridge Handbook of
­Sociocultural Psychology, J. Valsiner & A. Rosa (eds), 678–691. Cambridge: Cambridge
University Press.
Gladwell, M. 2005. Blink: The Power of Thinking Without Thinking. New York: Little, Brown and
Company.
 E. Sue Savage-Rumbaugh & William M. Fields

Greenfield, P.M. & Savage-Rumbaugh, E.S. 1984. “Perceived variability and symbol use: A
­common language-cognition interface in children and chimpanzees.” Journal of C ­ omparative
Psychology 98: 201–218.
Greenfield, P.M. & Lyn, H. 2007. “Symbol Combination in Pan: Language, Action and ­Culture.” In
Primate Perspectives on Behavior and Cognition, D. Washburn (ed.), 255–267. ­Washington
D.C.: American Psychological Association.
Greenfield, P.M., Lyn, H. & Savage-Rumbaugh, E.S. 2008. “Protolanguage in ontongeny and
phylogeny: Combining deixis and representation.” Interaction Studies 9: 34–35.
Halliday, M.A.K. 1979. “One child’s protolanguage.” In Before Speech: The Beginning of
­Interpersonal Communication, M. Bullowa (ed.), 171–190. Cambridge: Cambridge Univer-
sity Press.
Haesler, S., Wada, K., Nshdejan, A., Morrisey, E.E., Lints, T., Jarvis, E.D. & Scharff, C. 2004.
“FOXP2 expression in avian vocal learners and non-learners.” Neuroscience 24: 3164–3175.
Henriques, G.R. 2004. “The development of the unified theory and the future of psychotherapy.”
Psychotherapy Bulletin 39: 16–21.
Kano, T. 1992. The Last Ape. Palo Alto: Stanford University Press.
Kellogg, W.N. & Kellogg, L.A. 1933. The Ape and the Child: A Study of Environmental Influence
upon Early Behavior. New York: McGraw-Hill.
King, M.C. & Wilson, A.C. 1975. “Evolution at two levels in humans and chimpanzees.” ­Science.
188 (4184): 107–16.
Kitcher, P. 2006. “Two normative roles of consciousness.” In The Missing Link in Cognition, H.S.
Terrace & J. Metcalfe (eds), 174–187. Oxford: Oxford University Press.
Lewis, M. 1991. “Ways of knowing.” Developmental Review 11: 231–243.
Lewis, M. 1995. Shame, The Exposed Self. New York: The Free Press.
Lieberman, P. 2009. FOXP2 and human cognition. Cell 137: 800–802.
Liu, H., Prugnolle, F.A., Manica, A. & Balloux, F. 2006. “A geographically explicit genetic
model of worldwide human-settlement history.” American Journal of Human Genetics 79:
230–237.
Lock, A. 1978. Action, Gesture and Symbol: The Emergence of Language. London: Academic Press.
Lock, A. 2000. “Phylogenetic Time and Symbol Creation: Where Do Zopeds Come From.”
­Culture and Psychology 6: 105–129.
Lock, A.J. & Peters, C. (eds). 1996. Handbook of Human Symbolic Evolution. Oxford: ­Clarendon
Press.
Lock, J. 1690/1959. An Essay Concerning Human Understanding. 2 Vols. A.C. Fraser (ed.). Oxford
UK: Clarendon. Reprinted 1959. New York: Dover.
Lovejoy, O.C. 2009. “Re-examining Human Origins in Light of Ardipithecus Ramidus.” Science
326: 74.
Lyn, H. Greenfield, P.M. & Savage-Rumbaugh E. S. 2006. “The development of r­ epresentational
play in chimpanzees and bonobos: Evolutionary implications, pretense and the role of
inter-species communication.” Cognitive Development 21: 199–213.
Lyn, H. & Savage-Rumbaugh, E.S. 2000. “Observational word learning in two bonobos (Pan
paniscus): ostensive and non-ostensive contexts.” Language & Communication 20:
255–273.
Lyn, H., Franks, B. & Savage-Rumbaugh, E.S. 2008. “Precursors of morality in the use of
the ­symbols “good” and “bad” in two bonobos (Pan paniscus) and a chimpanzee (Pan
­troglodytes).” Language and Communication 28: 213–224.
Lyn, H., Greenfield, P.M. & Savage-Rumbaugh, E.S. in press. “Semiotic Combinations in Pan:
A Comparison of Communication in a Chimpanzee and Two Bonobos.” First Language.
 Chapter 2. The evolution and the rise of human language 

Matsuzawa, T. 2001. Primate Origins of Human Cognition and Behavior. Tokyo: Springer-Verlag.
McBrearty, S. & Jablonski, N.J. 2005. “First fossil chimpanzee.” Nature 437: 105–108.
Meltzoff, A.N. 1996. “The human infant as imitative generalist: a 20 year progress report on
infant imitation with implications for comparative psychology.” In Social Learning in
Animals: The Roots of Culture, C.M. Heyes and B. G. Galef (eds), 347–370. NY: Academic
Press.
Menzel, E., Savage-Rumbaugh, E.S. & Lawson, J. 1985. “Chimpanzee (Pan troglodytes) spatial
problem solving with the use of mirrors and televised equivalents of mirrors.” Journal of
Comparative Psychology 99 (2): 177–185.
Mithen, S. 1996. The Prehistory of the Mind. New York: Thames and Hudson.
Myowa-Yamakoshi, M., Tomonaga, M., Tanaka, M. & Matsuzawa, T. 2004. “Imitation in
­neonatal chimpanzee (Pan troglodytes).” Developmental Science 7: 437–442.
Patterson, N., Richter, D.J., Gnerre, S., Lander, E.S. & Reich, D. 2009. “Genetic evidence for
­complex speciation of humans and chimpanzees.” Nature 441: 1103–1108.
Papousek, M. 2007. “Communication in early infancy: An arena of intersubjective learning.”
Infant Behaviour and Development 30: 258–266.
Pederson, J. & Fields, W.M. 2009. “Aspects of Repetition in Bonobo Human Conversation:
Creating Cohesion in a Conversation Between Species.” Integrative Psychological and
Behavioral Science 43: 22–41.
Pederson, J., Segerdahl, P. & Fields, W.M. 2009. “Why apes point: Indexical pointing in
­spontaneous conversation of language-competent Pan/Homo bonobos.” In Primatology:
Theories, Methods and Results, E. Potocki and J. Krasiñski (eds), 53–74. Hauppauge, NY:
Nova ­Science Publishers.
Perry, G.H., Yang, F., Marques-Bonet, T., Murphy, C., Fitzgerald, T., Lee, A.S., Hyland, C., Stone,
A.C., Hurles, M.E., Tyler-Smith, C., Eichler, E.E., Carter, N.P., Lee, C. & Redon, R. 2008.
“Copy number variation and evolution in humans and chimpanzees.” Genome Research 18
(11): 1698–1710.
Pinker, S. 1994. The Language Instinct. New York: Morrow.
Pollard, K. 2009. “What Makes Us Human.” Scientific American 300: 44–49.
Povinelli, D.J. & Vonk. J. 2003. “Chimpanzee minds, suspiciously human?” Trends in Cognitive
Sciences 7 (4): 157–160.
Premack, D. 2004. “Is language the key to human intelligence?” Science 303: 318–320.
Rizzolatti, G., Fogassi, L. & Gallese, V. 2001. “Neurobiological mechanisms underlying the
understanding and imitation of action.” Nature Review Neuroscience 2: 661–700.
Rogoff, B. 1990. Apprenticeship in thinking: Cognitive development in the social context. New
York: Oxford University Press.
Savage-Rumbaugh, E.S. 1986. Ape Language: From Conditioned Response to Symbol. New York:
Columbia University Press.
Savage-Rumbaugh, E.S. 1990. “Language as a Cause-Effect Communication System.”
­Philosophical Psychology 90: 55–77.
Savage-Rumbaugh, E.S. 1994. “Hominid evolution: Looking to modern apes for clues.” In
­Hominid Culture in Primate Perspective, D. Quiatt & J. Itani (eds), 7–49. Niwot, CO:
­University Press of Colorado.
Savage-Rumbaugh, E.S. 1997. “Why are we afraid of apes with language?” In The Origin and
Evolution of Intelligence, A.B. Scheibel & J. W. Schopf (eds), 43–69. Sudbury, MA: Jones &
Bartlett.
Savage-Rumbaugh, E.S. 1999. “Ape Language: Between a rock and a hard place.” In The Origins
of Language, Barbara King (ed.), 115–188. Sante Fe: School of American Research Press.
 E. Sue Savage-Rumbaugh & William M. Fields

Savage-Rumbaugh, E.S., McDonald, K., Sevcik, R.A., Hopkins, W.D. & Rubert, E. 1986. “Sponta-
neous symbol acquisition and communicative use by pygmy chimpanzees (Pan paniscus).”
Journal of Experimental Psychology: General 115: 211–235.
Savage-Rumbaugh, E.S., Murphy, J., Sevcik, R.A., Brakke, K.E., Williams, S.L. & Rumbaugh,
D.M. 1993. “Language Comprehension in Ape and Child.” Monographs of the Society for
Child Development 58: 1–256.
Savage-Rumbaugh, E.S., Williams, S., Furuichi, T. & Kano, T. 1996. “Language Perceived, Pan
paniscus branches out.” In Great Ape Societies, W.C. McGrew, L.F. Marchant, & T. Nishida
(eds), 173–184. New York: Cambridge University Press.
Savage-Rumbaugh, E.S., Shanker, S.G. & Taylor, T.J. 1998. Apes, Language and the Human Mind
Oxford: Oxford University Press.
Savage-Rumbaugh, E.S. & Fields, W.M. 2000. “Linguistic, Cultural and Cognitive Capacities of
Bonobos (Pan paniscus).” Culture and Psychology 6: 131–153.
Savage-Rumbaugh, E.S., Fields, W.M. & Spircu, T. 2004. “The emergence of knapping and vocal
expression embedded in Pan/Homo culture.” Biology and Philosophy 19: 541–575.
Savage-Rumbaugh, E.S., Segerdahl, P. & Fields, W.M. 2005. “Individual differences in l­anguage
competencies resulting from unique rearing conditions imposed by different first
­epistemologies.” In Symbol Use and Symbolic Representation, L.L. Namy (ed.), 199–219.
Mahwah: Lawrence Earlbaum Associates.
Savage-Rumbaugh, E.S., & Fields, W.M. 2007. “Rules and tools: Beyond anthropomorphism.” In
The Oldowan: Case Studies into the Earlies Stone Age, N. Toth & K. Schick (eds), ­223–241.
Gosport: Stone Age Institute Press.
Savage-Rumbaugh, E.S., Wamba, K., Wamba, P. & Wamba, N. 2007. “Welfare of apes in a captive
environment: Comments on and by, a specific group of apes.” Journal of Applied Animal
Welfare Sciences 10: 7–19.
Savage-Rumbaugh, E.S., Rumbaugh, D. & Fields, W.M. 2009. “Empirical Kanzi: The Ape
­Language Controversy Revisited.” Skeptic 15 (1): 25–33.
Scharff, C. & White, S.A. 2004. “Genetic Components of vocal learning.” Annals of the New York
Academy of Sciences 1016. 325–347.
Sebeok, T.A. & Umiker-Sebeok, J. 1980. Speaking of Apes: A Critical Anthology of Two-Way
­Communication with Man. New York: Plenum Press.
Segerdahl, P., Fields, W.M. & Savage-Rumbaugh, E.S. 2005. Kanzi’s Primal Language: The
­Cultural Initiation of Primates into Language. New York: Palgrave Macmillian.
Shotter, J. 1993. Conversational Realities: Constructing Life Through Language. London: Sage
Publications.
Shumaker, R. & Beck, B. 2003. Primates in Question: The Smithsonian Answer Book. ­Washington,
D.C.: Smithsonian Institution Press.
Sluneko, T. & Hengle, S. 2007. “Language, Cognition and Subjectivity: A Dynamic C ­ onstitution.”
In The Cambridge Handbook of Sociocultural Psychology, J. Valsiner and A. Rosa (eds),
40–61. Cambridge: Cambridge University Press.
Smiley, P.A. & Johnson, R.S. 2006. “Self-referring terms, event transitivity and development of
self.” Cognitive Development 21: 266–284.
Steels, L. 2009. “Is sociality a crucial prerequisite for language?” In The Prehistory of Language:
Studies in the Evolution of Language, R. Botha and C. Knight, (eds), 36–57. Oxford: Oxford
University Press.
Stern, D. 1971. “A micro-analysis of mother-infant interaction: Behavior regulation and social
contact between a mother and her 3 1/2 month-old twins.” Journal of the American ­Academy
of Child Psychiatry 12: 402–421.
 Chapter 2. The evolution and the rise of human language 

Stern, D. 1977. The First Relationship. London: Fontana/Open Books.

Taglialatela, J., Savage-Rumbaugh, E.S. & Baker, L.A. 2003. “Vocal Production by a Language-
Competent Pan Paniscus.” International Journal of Primatology 24: 1–15.
Taglialatela, J., Savage-Rumbaugh, E.S., Rumbaugh, D.M., Bensen, J. & Greaves, W. 2004.
“Language, Apes & Meaning Making.” In The Development of Language: Functional
­Perspectives on Species and Individuals, G. Lukin and A. Williams (eds), 91–111. New
York: Continuum.
Thibault, P. 2004. Agency and Consciousness in Discourse: Self-other Dynamics as a Complex
System. London: Continuum.
Tomasello, M. 2008. Origins of Human Communication. Cambridge, MA: MIT Press.
Trevarthen, C. 1989. “Sharing makes sense: Intersubjectivity and the making of an infant’s
­meaning.” In Language Topics: Essays in Honor of Michael Halliday, R. Steele and T. Thread-
gold (eds), 177–191. Amsterdam: John Benjamins.
Trevarthen, C. 1998. “The concept and foundation of infant intersubjectivity.” In ­Intersubjectivity,
Communication and Emotion in Early Ontogeny, S. Braten (ed.), 183–189. Cambridge:
Cambridge University Press.
Trevarthen, C. & Hubley, P. 1978. “Secondary intersubjectivity: confidence, confiding and acts
of meaning in the first year.” In Action, Gesture and Symbol: The Emergence of Language,
A. Lock (ed.), 183–229. London: Academic Press.
Vandell, D.L. & Wilson, K.S. 1988. “Infants’ interactions with mother, sibling, and peer: C­ ontrasts
and relations between interaction systems.” Child Development 48: 176–186.
Vargah-Khadem, F., Gadian, D.G., Copp, A. & Mishkin, M. 2005. “FOXP2 and the n ­ euroanatomy
of speech and language.” Nature Review of Neuroscience 6: 131–138.
Wermke, K., Leising, D. & Stellzig-Eisenhauer, A. 2007. “Relation of melody in infant cries to
language outcome in the second year of life.” Clinical Linguistics & Phonetics 21: 961–973.
Wermke, K. & Mende, W. 2006. “Melody as a primordial legacy from early roots of language.”
Behavioral and Brain Sciences 29: 300.
White, T., Asfaw, B., Beyene, Y., Haile-Selassie, Y., Lovejoy, C.O., Suwa, G. & WoldeGabriel,
G. 2009. “Ardipithecus ramidus and the Paleobiology of Early Hominids.” Science 236: 64.
Wildman, D., Grossman L. I. & Goodman, M. 2002. “Functional DNA in humans and chimpan-
zees shows they are more similar to each other than either is to apes.” In Probing Human
Origins, M. Goodman & A. S. Moffat (eds), 11–32. Cambridge, MA: American Academy
of Arts and Sciences.
Wilson, T. 2002. Strangers to Ourselves: Discovering the Adaptive Unconscious. Cambridge, MA:
Harvard University Press.
Wolff, P.H. 1987. The Development of Behavioral States and the Expression of Emotions: Early
Infancy: New Proposals for Investigation. Chicago: University of Chicago Press.
Wynn, C. 2008. “Aping Language: A skeptical analysis of the evidence for nonhuman primate
language.” Skeptic 13: 10–14.
Yellen, J.E., Brooks, A.S., Cornelissne, E., Mehlman, M.J. & Stewart, K. 1995. “A Middle Stone
Age bone industry from Katanda, Upper Semliki Valley, Zaire.” Science 268: 553–557.
Yerkes, R. 1929. The Great Apes: A Study of Anthropoid Life. New Haven, CT: Yale University Press.
Yunis, J.J. & Prakash, O. 1982. “The Origin of Man: A Chromosomal Pictorial Legacy.” Science
215: 1525–1530.
Zuberbühler, K. 2003. “Referential signaling in non-human primates: Cognitive precursors
and limitations for the evolution of language.” In Advances in the Study of Behavior, P.P.
Slater, J. Rosenblatt, C. Snowdon, T. Roper, & M. Naguib (eds), 265–307. New York:
­Academic Press.
 chapter 3

The origin of symbolically mediated behaviour

From antagonistic scenarios to a unified
research strategy
Francesco d’Errico1,2 & Christopher S. Henshilwood2,3
1CNRS-UMR 5199 PACEA, Université de Bordeaux/2Institute for Archaeology,
History, Culture and Religion, University of Bergen/3Institute for Human
Evolution, University of the Witwatersrand

The aim of this paper is to summarize what we do know about the origin of
symbolic material cultures in Africa and Eurasia, and explore paths that could
allow us to move from a situation in which the same evidence is accounted for
by antagonistic scenarios to a research strategy that may produce, in the end,
a unified theory for the emergence of this innovation. Our review highlights
the need to develop an integrated research strategy in which assumptions on
cognition based on taxonomic affiliation will play no role. The key tools to
address this topic should be archaeology, palaeoenvironmental studies and
new methods to integrate results from these disciplines. This appears to be one
positive way to understand the mechanisms that governed cultural transmission
and social learning between 160 and 30 ka.

1. Introduction

A fundamental change in human behaviour occurred when symbolism became

inherent in material culture. This innovation, which demonstrates the ability for
sharing, storing and transmitting coded information within and across groups,
has played a crucial role in creating, maintaining and transmitting the social con-
ventions, beliefs and identities that characterise all known human societies. The
origin of symbolic material culture, long associated with a rapid cognitive change
in Europe 40 ka ago, is now considered by many to have gradually emerged in
Africa, in conjunction with the origin of our species. Others stress the use of sym-
bolic artefacts by Neandertals, and their burial practices, to challenge the idea that
symbolic behaviour is peculiar to our species.
The question of the emergence of symbolically mediated behaviour has
stimulated heated debates in the last decade for a number of good reasons. The view
one adopts on this issue has implications for our vision of the factors that have driven
 Francesco d’Errico & Christopher S. Henshilwood

cultural evolution in our lineage (genetic or cultural drift, ­climate, demography),

the way they have interacted, and the timing and mode of the emergence of other
crucial human quintessential features such as modern ­cognition, working memory,
language, religious beliefs, culturally transmitted social ranking systems, artificial
memory systems and so forth. The aim of this paper is to summarize what we
do know at present about the origin of symbolic material cultures in Africa and
Eurasia, and explore paths that could allow us to move from a situation in which the
same evidence is accounted for by antagonistic scenarios to a research strategy that
may produce, in the end, a unified theory for the emergence of this innovation, and
a better comprehension of the ­mechanisms that have stimulated the emergence,
maintenance and transmission of this unique human adaptation.
We have long known that symbolic thinking, defined here as the capacity to
attribute specific meaning to conventional signs, is not peculiar to our species
(McGrew, this volume; Savage-Rumbaugh & Fields, this volume). We share that
capacity with a growing number of primates and non-primate species. The key
criterion is not the capacity for symbolic thought but the creation of symbolic
material culture that unequivocally integrates socially shared symbolic meaning.
It is through the existence of these symbols that humans are able to condition their
and others’ actions and choices.
Until very recently the emergence of symbolism was considered coincident
with the arrival of Homo sapiens in Europe at 40 ka. This view, commonly
known as the “Human Revolution” scenario (Mellars 1996; Klein 2008), has been
replaced in the last decade by the “out of Africa” model (McBrearty & Brooks
2000; Mellars 2006, 2007). This latter model argues that symbolically mediated
behaviour developed in Africa and is linked with the emergence of Homo
sapiens in that continent (Henshilwood & Marean 2003, 2006; Henshilwood &
Dubreuil 2011). According to this view the appearance of modern cognitive
behaviours (syntactical language, advanced cognition, symbolic thinking) in
Eurasia, associated with Homo sapiens, is possibly the result of the dispersal of
an already symbolic species at c. 60 ka or earlier. This spread would have led
to rapid extinction of pre-modern human populations in Asia and Europe with
little or no biological exchange and limited cultural interaction. However, an
additional model that we may term the ‘Symbolic Neandertal’, suggests that the
material culture associated with Neandertals in Europe before the first arrival
of Homo sapiens was associated with behaviour that was mediated by symbols
(Zilhão & d’Errico 1999; d’Errico 2003; Zilhão et al. 2010; d’Errico & Stringer
2011). If correct, then it seems possible that there was a parallel evolution towards
behavioural modernity and the use of symbols on two continents by two or more
distinct populations. At this stage the effect of genetic exchange between Homo
sapiens and Neandertals on their respective evolution of material culture and
 Chapter 3. The origin of symbolically mediated behaviour 

innovations is not well understood (Green et al. 2010). The tenets of the parallel
evolution model favours a ­scenario in which demographic changes triggered by
environmental variation may have stimulated, but not necessarily synchronically,
the emergence of ­ innovations among both modern humans in Africa and
Neandertals in E ­urasia. Acquisition of innovations from a neighbouring
population is also a possibility, though difficult to test outside the Near East and
Europe, the areas where we know Homo sapiens and Neandertals came into direct
contact. If demonstrated, the acquisition of innovation through contact would
support rather than c­ ontradict the ‘Symbolic Neandertal’ scenario since it would
show, as is the case in many historical instances, that both populations were able
to assimilate cultural innovations created by other groups.

2. First instances of symbolic material culture and their implications

Deliberate inhumation and special treatment of the dead, such as secondary

burial, cremations, intentional curation of the skeleton or selected body parts
and the conscious use of human bone as a raw material to produce artefacts, are
generally regarded as quintessential features of modern humanity (Pearson 1999;
McBrearty & Brooks 2000; Henshilwood & Marean 2003; d’Errico et al. 2003). The
oldest modern human burials date back to at least 100–130 ka (Grün et al. 2005).
Neandertal burials are as old or might be older if one accepts the most ancient
date for the Tabun C1 burial (Bar-Yosef & Callender 1999; Grün & Stringer 2000;
Mercier et al. 2000).
The oldest burials in Africa, which may date to 60–76 ka, are those from Nazlet
Kather (Vermeersch et al. 1984; Crevecoeur 2006) Qena, Egypt (Vermeersch
et al. 1998) and Border Cave (Millard 2006; Bird et al. 2003). The intentional
character and symbolic significance of burials prior to 30 ka, especially those of
Neandertals, has been the subject of intense debate over the past decades. Most
Neandertal burials were excavated long ago and available reports often lack the
information now required to assess the anthropogenic origin of the inhumation
and whether items interpreted as grave goods were deliberately interred. There
is nevertheless enough information on Neandertal burials to accept that in some
areas of Europe and the Near East these populations systematically buried their
dead (Vandermeersch et al. 2008). Grave goods consisting of stone tools, bone
retouchers, engraved bone and a rock slab engraved with cupules are also reported
at La Ferrassie, La Chapelle-aux-Saint, Le Moustier in France, and Amud and
Dederiyeh in the Near East.
The earliest abstract designs, engraved on bone and ochre, are found in South
Africa (Figure 1) and are dated to ca 100 ka (Henshilwood et al. 2002, 2009; see
 Francesco d’Errico & Christopher S. Henshilwood

Henshilwood & d’Errico, this volume). Examples are the complex ­geometric
patterns on ochre from 100 to 70 ka levels at Blombos Cave ­(Henshilwood et
al. 2002, 2009) and from MSA layers at Klein Kliphuis in the Western Cape
(Mackay & Welz 2008), ca 73 ka old notched and engraved bone from B ­ lombos
and ­Klasies (d’Errico et al. 2001; d’Errico & Henshilwood 2007). Abstract designs
on artefacts seem to disappear in southern Africa between 70 ka and c. 65 ka
after which they reappear at Diepkloof Rock Shelter in the form of engraved
ostrich eggshells (Rigaud et al. 2006; Texier et al. 2010; H
­ enshilwood & d’Errico,
this volume). Evidence from the Near East includes an engraved ­cortex dated at
c. 50 ka from the Mousterian site of Quneitra (Figure 2), that could be a­ ssociated
with Homo sapiens or Neandertals (Goren-Inbar 1990; ­Marshack 1996) and
an engraved lithic core from c. 90 ka levels at Qafzeh (Hovers et al. 1997).
A ­number of objects bearing putative engravings have been reported from Lower
and Middle Palaeolithic sites in Europe (Lorblanchet 1999; Bahn 1996). Some
of these ‘engravings’ result from natural phenomena and the marks left by
carcass processing (d’Errico & Villa 1997). A number can be interpreted as
­deliberate engravings (Soressi & d’Errico 2007) but few show complex s­ tructured
designs.

a d

e
c

scale = 1cm

Figure 1. Material culture from the c. 78–72 ka Still Bay levels at Blombos Cave, South Africa:
(a) bifacial foliate point in silcrete, (b) bone tool, (c) engraved ochre, (d) Nassarius kraussianus
shell beads, (e) engraved bone. (a: photo Henshilwood; b: photo d’Errico/Henshilwood;
c: photo Henshilwood; d: photo d’Errico et al. 2005; e: photo d’Errico/Henshilwood)
 Chapter 3. The origin of symbolically mediated behaviour 

Convincing evidence for the use of personal ornaments, consisting of

­perforated marine shell belonging, at each site, to a single species, is found
at sites from South Africa (Henshilwood et al. 2004; d’Errico et al. 2005),
North Africa (Bouzouggar et al. 2007; d’Errico et al. 2009), and the Near East
­(Vanhaeren et al. 2006) dated to between 100 and 70 ka. At Blombos Cave forty
nine d ­ eliberately perforated Nassarius kraussianus shell beads (Figure 1) with
clear evidence of ­use-wear and some bearing traces of ochre come from 75 ka old
­levels (­ Henshilwood. 2004; d’Errico et al. 2005). Possible beads made of Littorina
africana marine shells were found in the upper Still Bay layers of Sibudu Cave
(d’Errico et al. 2008). The perforated Conus shell from Border Cave, associated
with the burial of a young individual may be as old as 76 ka according to the
recent chronological attribution of this burial (Millard 2006). Perforated Nassa-
rius gibbosulus shells (Figure 2) were recovered at Skhul from c. 100 ka levels that
include eleven Homo sapiens burials (Vanhaeren et al. 2006). Perforated shells of
the same species (Figure 2) showing traces of intentional modifications, possible
deliberate heating to change the colour of the bead, use-wear, and traces of red
ochre were recovered from the Aterian site of Oued Djebbana, Algeria, and the
80–70 ka levels at Grotte des Pigeons, Rhafas, Ifri n’Ammar, and Contrebandiers
in Morocco (d’Errico et al. 2009).
Other marine shells interpreted as beads (Figure 2) come from the c. 90 ka
Mousterian levels XXI–XIV at Qafzeh Cave in Israel (Bar-Yosef Mayer et al. 2009).
These consist of ten naturally perforated Glycymeris insubrica shells, some with
traces of red pigments. The only Neandertal site predating the arrival of modern
human populations in Europe that have yielded possible evidence for the use of
beads by Neandertals is Cueva de los Aviones, southern Spain (Zilhão et al. 2010).
The Mousterian layers of this site, dated to ca 50–45 ka, contained marine shell
assemblages including three valves of Acanthocardia and Glycymeris shells bearing
natural perforations (Figure 2). One of the latter contained a residue of red pig-
ment identified as haematite.
Beads disappear in Africa and the Near East between 70 ka and 40 ka
(d’Errico & Vanhaeren 2007; d’Errico et al. 2009) and reappear almost every-
where in Africa and Eurasia after this time span. In Europe 40 ka beads are
associated with both Neandertals and Homo sapiens (d’Errico et al. 2003).
These beads differ from their 100–70 ka antecedents in that they take the form
of hundreds of discrete types that identify regional patterning (Vanhaeren &
d’Errico 2006). At 40 ka, beads in Africa are made from ostrich eggshells and
only later are a diverse range of raw materials introduced for bead manufacture
(Figure 2).
 Francesco d’Errico & Christopher S. Henshilwood

b c

f g
e

Scale = 1cm

Figure 2. (a) engraved cortex from the site of Quneitra (ca 50 ka), Israel; (b) Nassarius
­gibbosulus shell beads from Skhul, Israel (ca 100 ka); (c) Oued Djebbana, Algeria (>40 ka);
(d) Grotte des Pigeons, Taforalt, Morocco (ca 82 ka) respectively; (e) perforated Glycimeris
from Qafzeh, Israel (ca. 90 ka); (f) perforated Glycimeris from Cueva de los Aviones, Spain
(ca 50 ka); (g) Ostrigh eggshell beads from Mumba cave (ca 40 ka). (a: photo d’Errico/Nowell;
b–d: photo d’Errico/Vanhaeren; e: modified after Bar-Yosef et al. 2010; f: by courtesy of
N. Conard, ­University of Tübingen)

In south east Asia, the oldest documented ornament is a perforated tiger

shark tooth found in New Ireland, New Guinea at a site dated between 39.5 and
28 ka (Leavesley 2007). The earliest evidence for bead use in Australia comes
from the site of Mandu Mandu, Cape Range of Western Australia, where 22
Conus shell beads were recovered in a layer dated to ca. 32 ka (Morse 1993). In
addition, ten Dentalidae shell beads are reported from the 30 ka old layers of Riwi
in the Kimberly region of Western Australia, a site located 300 km inland (Balme
& Morse 2006).
The systematic use of red pigments (Figure 3) is evident in Africa at archae-
ological sites dated to 160 ka (Marean et al. 2007) and possibly at sites dated
 Chapter 3. The origin of symbolically mediated behaviour 

to 280 ka (McBrearty & Brooks 2000). In the Near East the oldest evidence for
­systematic use of pigments dates to ca 100 ka (Hovers et al. 2003; d’Errico et al.
2010). P­ igments, mostly black, have been sporadically used by N­ eandertals in
Europe since 300 ka (Marshack 1981) but their use only becomes ­systematic
after 60 ka (d’Errico 2003; Soressi & d’Errico 2007; Zilhão et al. 2010). The
last ­Neandertals in France and Italy made intensive use of both black and red
­pigments just before, and at the moment of contact with Homo sapiens in Europe
(Salomon 2009; R­ onchitelli et al. 2009).

Scale = 1cm

Figure 3. Early instances of pigment use in the Near East, Africa, and Europe; (a) Skhul cave,
Israel (100 ka); (b) Blombos Cave, South Africa, Still Bay layers, ca 75 ka; (c) Pech-de-L’ Azé
I, Dordogne, France, Mousterian of Acheulean tradition, ca 50 ka; (d) Klasies River, South
Africa, Howieson Poort layers, ca 60. ka. (a: photo d’Errico et al. 2010; b–d: photos F. d’Errico)
 Francesco d’Errico & Christopher S. Henshilwood

Different interpretations have been placed on early pigment use. Some

­archaeologists have interpreted pigments as proof of a growing use of symbols
(Knight et al. 1995; Watts 1999, 2002; McBrearty & Brooks 2000; Henshilwood et
al. 2002; Henshilwood & Marean 2003; Hovers et al. 2003; d’Errico 2003; d’Errico
et al. 2009; Zilhão et al. 2010; others stress the deliberate choice of pigments for
their intense red hues (Watts 1999); a preference for pigments from far away
sources (Zilhão et al. 2010); the deliberate heating of pigments to change their
colour (Hovers et al. 2003); the deliberate use of pigment to stain one side only
of an object (Zilhão et al. 2010); to colour shell beads (Henshilwood et al. 2004;
d’Errico et al. 2005, 2009); and as decoration on bodies or on clothing. It is also
suggested that early pigment use may have been utilitarian, for example as skin
protection from sun or insects, as a medicine, for tanning hides, or as a binding
agent to facilitate hafting (Velo 1984; Klein 1999; Wadley 2001; Klein & Edgar
2003; Wadley et al. 2004, 2009; Lombard 2007).
Figurative representations (Figure 4), consisting of painted, engraved and
carved animals, appear much later, at c. 28 ka in Africa, at Apollo 11 shelter (Wendt
1976), Namibia, and at 32 ka in Europe, for example at Chauvet, Fumane and in
southern Germany (Clottes 2003; Conard et al. 2004). The oldest carved musical
instruments, consisting of flutes made of bird bone and mammoth ivory decorated
with notches, are found in Europe and date to 32 ka (d’Errico et al. 2003; Conard
et al. 2004). No convincing musical instruments are associated with Neandertals
(d’Errico & Lawson 2006).
Inferring symbolic capacity from bone and stone tools is a notoriously
tricky endeavour. If we accept, with some caveats (Wiessner 1983), the definition
of style as an over - determination of form imposed by cultural standards that
identifies ethnically bounded populations (Byers 1994, 1999; Chase 2006) then
convincing instances of such behaviour can be found in southern Africa stone tool
assemblages. Examples include the stone bifacial points from the c. 75 ka Still Bay
levels at Blombos Cave (Villa et al. 2009), in the Near East Mousterian from 100 ka
(Hovers & Belfer-Cohen 2006) and in Europe with regional traditions appearing
in the Mousterian at c. 60 ka (Delagnes & Meignen 2005). Recent research has
shown that such trends do not follow a progressive development. In South Africa,
as in Europe, techno-complexes consisting of highly developed tool kits alternate
with periods in which stylistic intent, if present, becomes difficult to perceive
(Villa et al. 2009).
Fully shaped bone tools, some of which are decorated with engraved lines and
notches (spear points, awls, spatulas, harpoons) occur in Africa from c. 75 ka and
possibly at 90 ka (Henshilwood et al. 2001; d’Errico & Henshilwood 2007; Yellen
et al. 1995; Backwell et al. 2008). The careful deliberate polishing of the ca 75 ka
Blombos bone projectile points has no apparent functional purpose and, rather,
 Chapter 3. The origin of symbolically mediated behaviour 

seems a technique used to give a distinctive appearance and/or an ‘‘added value’’

to this category of artefacts. This may imply that symbolic meaning was attributed
to bone tools. Bone tools disappear periodically after 70 ka in Africa and are only
in common and continued use in the final periods of the Later Stone Age. Nean-
dertals produced complex bone tools, some of which are decorated, only after 40
ka (d’Errico et al. 2003; d’Errico et al. 2011). From 35 ka, bone tools are ubiquitous
in European tool kits associated with Homo sapiens.

a e g

f i

b c

Figure 4. First examples of figurative representations from Africa and Europe: (a) painted
slab from Apollo 11 Cave, Namibia (ca 28 ka BP), (b) figurine from Straitzig/Krems-Rehberg,
­Austria (ca 32 ka BP), (c–d) wall fragments from Fumane, Italy (ca 32 ka BP), (e) wall
paintings from Chauvet Cave, France (ca 32 ka BP), (f) human with a lion head from
Hohlenstein-Stadel, Germany (ca 32 ka BP), (g–i) mammoth, lion and lion head from
Vogelherd, ­Germany (ca 32 ka BP); (j) female figurine from Hohle Fels, Germany (ca 32 ka
BP), (k) Sculpted vulvas from the Aurignacian layers of La Ferrassie rock shelter, France
 Francesco d’Errico & Christopher S. Henshilwood

Our succinct review of the evidence for symbolic material culture (1)
contradicts the idea that the production of symbolic material culture is the result
of a sudden change in human cognition occurring in Europe or in Africa after
50–40 ka; (2) suggests the presence of symbolic material culture in Africa by at
least 150 ka, in the Near East after 100 ka and probably by at least 60–50 ka in
Europe; (3) indicates that engravings, pigments, personal ornaments, formal bone
tools and burial practices do not appear in the archaeological record as a single
package; (4) re-emphasizes, the equivocal status of pigment and more strongly
of stone and bone tool stylistic trends as hallmarks of symbolic behaviour; (5)
reveals that during c. 150 ka and 30 ka these material culture innovations appear,
disappear and reappear in different forms suggesting major discontinuities in
cultural transmission; (6) contradicts, if one accepts reported instances of symbolic
artefacts associated with Neandertals, the assumption that ground breaking
innovations can be associated only with Homo sapiens.

3. Demography as a triggering factor?

The complex patterns associated with innovations in material culture that we

describe above cannot be explained solely by speciation. The evidence shows
that it is discontinuous in time and space, and the commonalities found in both
hemispheres indicate that once a cognitive threshold is crossed, local conditions
must have played a role in the emergence, diffusion and the eventual disappearance
of crucial innovations. These local conditions must have been closely linked to the
size and organisation of cultural systems and ecological settings in which these
populations evolved.
Can selective pressure eliminate cultural complexity and refined innovations
when they are deemed unfit for use within a group or population? What are the
conditions under which this can happen? Archaeological data and ethnographic
and historical accounts show that cultural groups can lose innovations. At the
beginning of the 20th century William Rivers (1912: 123) speculated on the rea-
sons that led to the loss of seaworthy canoes in Oceania. His interpretation was that
“social and magico-religious, as well as material and utilitarian factors should be
taken into account”. Homo sapiens who colonised Tasmania at c. 35 ka, took with
them a number of cultural innovations that they transmitted and implemented for
25 ka. After 10 ka, with the flooding of the Bass Strait, Tasmania was transformed
into an island. The population was reduced to a few thousand and the technology
employed by these people became very simple compared to that employed on the
Australian mainland (Diamond 1978).
One research direction has been to try and create models of cultural
transmission and then to link these models to models of genetic evolution. It is
 Chapter 3. The origin of symbolically mediated behaviour 

a tradition for which most archaeologists and palaeoanthropologists have paid

little attention. Examples of this research are the seminal papers and book by­­
Cavalli-Sforza and Feldman (1981), followed by Boyd and Richerson (Boyd &
Richerson 1985; Richerson et al. 2005, 2009). Only recently have scholars tried
to apply or adapt these models to account for the emergence and loss of cultural
innovations. Recent papers (Shennan 2001; Henrich 2004; Powell et al. 2009;
Premo & Kuhn 2010) have explored this approach by modelling the role of
demography in the emergence and loss of cultural innovations. Shennan examines
the vertical and oblique cultural transmission of craft-techniques in small size
populations and highlights the crucial role of population size in the spread and
maintenance of innovations. He supports the assumption that the emergence of
modern human culture can be associated only with Homo sapiens and that even
though there is sporadic ­evidence of this before 70 ka, symbolically mediated
culture is prevalent only after 50 ka. Henrich’s model is based on the observation
that humans have a propensity to imitate skilfull individuals. The combination of
skill-biased transmission and learning errors would create a cumulative adaptive
evolution. In this ­process the number of social learners appears as the key factor
and this would account for the loss of technological innovations in Tasmania.
Powell et al. (2009) extend Henrich’s model by partitioning, in two regions,
different population densities, cultural exchange rates and migratory activities.
Premo and Kuhn (2010) reach a comparable conclusion for cultural evolution in
the Early Stone Age. Powell et al’s ­conclusion is that demography and migration are
the key factors influencing cultural exchange. Theoretically their model accounts
for the appearance and disappearance of innovations and for the delay between
the origin of anatomically modern Homo sapiens and that of the first evidence for
modern cultural traits. Their results are certainly significant as they provide a sound
explanation for the emergence and loss of innovations without invoking speciation
as the prime mover, but they nevertheless leave open a number of questions. Like
Shennan (2001), these authors do not consider the possibility that their model
accounting for the emergence of symbolic culture in Homo sapiens could be equally
applied to Neandertals. Following the logic of Powell et al. (2009), one may argue
that differences between Neandertals and Homo sapiens with regard to cultural
modernity may have depended on variations in group size and the rate of cultural
exchange rather than on hard wired differences in cognition.
Using a modelling approach to highlight the potential role of demographic
­factors in the emergence of cultural modernity does not provide absolute proof that
this was the necessary path. Models are models and even when they seem to broadly
account for reality they need to be tested against the empirical evidence. The ­question
is how to test these models and what do we accept as proof that they account for past
reality. Each appearance of a cultural innovation in the MSA, or of the spread of
innovations associated with migratory activities such as the ­colonisation of Europe
 Francesco d’Errico & Christopher S. Henshilwood

and Asia, are interpreted by Powell et al. (2009) as archaeological reflections of an

increase in population size. Conversely, cultural discontinuities are interpreted as
signs of demographic demises. Evidence supporting these links is however tenuous.
Through genetics alone we are unable to reconstruct the timing of repeated bottle-
necks or their impact on population dynamics, for example at c. 70 ka (Rampino
& Self 1993; Ambrose 1998). Powell et al. (2009) evoke the climatic deterioration
during Marine Isotope Stage (MIS) 4 as a possible factor leading to the decline,
­fragmentation and range contraction of human populations but they propose no
clear mechanisms to explain this. Invoking climatic change as the prime mover
equates with attributing these changes to a single speciation event. It creates a “deus
ex machina” that answers the question without solving the problem.

4. Climate as a triggering factor

One paradox of research into the evolution of humans is the observation that
climate change was the motor behind speciation and behavioural change. Only
coarse tools however have been developed that integrate the relevant climatic
data with human evolution, and the dialogue between archaeologists and
palaeoclimatic researchers is limited. For example, what is the state of our current
knowledge for the period 160–20 ka, the time during which symbolic material
culture emerges?
Geological research records indicate that during the last million years the Earth
underwent repeated changes between glacial periods, in which large continental
ice sheets developed in the northern hemisphere, and interglacial periods with
greatly reduced global ice volume. These changes link with changes in the amount
of solar insolation received at high northern latitudes during the summer, with
dominant cyclicities at 23, 41 and 100 ka reflecting shifts in the Earth’s orbital
geometry, and the potential for the development of significant ice sheets (Imbrie
1992, 1993; Milankovitch 1969). Within this framework, major interglacials occur
approximately every 100 ka, with periods of subtler warming, known as interstadials,
corresponding to periods of increased northern summer insolation at c. 23 ka
intervals. These large-scale cycles of climate change are well-documented in both
northern and southern hemisphere polar ice cores (Dansgaard et al. 1993; Jouzel
et al. 2007), and understood to define global climates during the late Pleistocene.
During the last 15 years, concerted research efforts have focused on rapid
or millennial-scale climate changes occurring during the last glacial cycle
(c. 80–10 ka to present). Detailed climate records (Figure 5) from the Greenland
ice cores and North Atlantic marine cores reveal that the last glacial period was
not ­uniformly cold, but was punctuated by abrupt atmospheric warming events,
 Chapter 3. The origin of symbolically mediated behaviour 

known as D ­ ansgaard-Oeschger (D-O) events (Bond et al. 1993). The amplitude

of these warming events has been estimated to be c. 10 0 Cover Greenland; the bulk
of the warming for each event typically occurred within a few decades, and a D-O
event lasted between 500 and 2000 years.
During certain Greenland cold phases, marine sediments of the North Atlantic
contain abundant ice-rafted detritus (IRD) in the form of sand-sized mineral
grains that have been transported from the adjacent continents. These layers
of IRD are accompanied by high percentages of the polar foraminifer species
N. pachyderma in the sediments, reflecting episodes of massive iceberg discharge
and strong oceanic cooling in the North Atlantic Ocean known as Heinrich events
(Heinrich 1988). There were seven such episodes during the last climatic cycle,
between 70–10 ka. The enormous amount of fresh water introduced to the North
Atlantic during these events is thought to have had a profound influence on climate
through a shut-down of North Atlantic Deep Water production, and a restriction
of oceanic heat transport to the North Atlantic.
GISP2 d18 O(‰)

–36 Greenland 19 20 21
A 4 8 12 14 17
–38
–40

Byrd d18 O(‰)

–42 B Heinrich Event –34
Antarctica H3 H4 H5 H5a H6 A7 –36
A1 A2 A3 A4 A6
280 A5 –38
260 C –40
CO2 (ppm)

240
220 Atmospheric CO2
D
200 700
CH4 (ppb)

600
180 Atmospheric CH4
500
400

10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90
GISP Age (ka)

Figure 5. Atmospheric CO2 composition and climate during the last glacial period. (A)
Greenlandic temperature proxy, δ18Oice. Red numbers denote DO events. (B) Antarctica
temperature δ18Oice. A1 to A7, Antarctic warming events, (C) Atmospheric CO2
concentrations. Vertical blue bars, timing of Heinrich events (H3 to H6). Brown dotted
lines, abrupt warming in Greenland. Blue bars indicate DO events and MIS5 (modified
after Ahn et al. 2008)

While ice and marine archives provide detailed records of D-O climate
variability, it has been for long problematic, in spite of 40 years of research on
 Francesco d’Errico & Christopher S. Henshilwood

long pollen records, to link specific events in the ice and marine domains with
vegetation changes, mostly because the duration of individual events is often
less than the chronological uncertainties involved in the dating of these records.
A major advance in this field has been introduced by the analysis of pollen and
microcharcoal deposited in marine environments. Transported by winds and river
currents into the ocean, these continental marine proxies sink rapidly through
the water column as a part of the marine snow along with marine and ice volume
climatic indicators, such as foraminifera on which stable isotope measurements
are performed. In this way vegetation changes can be directly correlated with SST
changes and ice dynamics. The analysis of cores strategically located along the
Iberian and French margin has allowed researchers to document how vegetation
and fire regime are affected by the millennial scale climatic variability of MOS 2
and 3 at different latitudes (Sanchez-Goñi et al. 2008).
Over the past decade and beyond, concentrated efforts by archaeologists,
climatologists, geographers and environmental scientists are contributing to a fine
tuning of the climatic variations in southern Africa after the end of the previous
interglacial (Marine Isotope Stage 6/5e (MIS)), (e.g. Grindley 1969; Deacon &
Lancaster 1988; Van Andel 1989; Thackeray 1992, 2007; Avery 1995; Partridge 1993,
1997; Lambeck & Chappell 2001; Blunier & Brook 2001; Bateman et al. 2004; Carr
et al. 2006; Scholz et al. 2007; Roberts et al. 2008; Henshilwood 2008). One example
is the IGCP 437 (International Geological Correlation Program) that examines
coastal environmental change during sea-level high stands (Mastronuzzi et al.
2005). There remains, however, significant discordance and debate among records
and researchers as to how regional environments have changed as a response to
past cycles of global warming and cooling as registered in polar ice cores.
Attempts to understand the impact of environmental variability on human
adaptation have used a variety of datasets and methods and have primarily
focused on the archaeological records of Africa and Europe. In Africa, research has
attempted to model the impact of environmental changes on hominin evolution,
dispersal, and cultural evolution (Potts 1996; deMenocal 2004; McBrearty &
Brooks 2000; Sepulchre et al. 2006; Henshilwood 2008). Some researchers have
focused on the Toba super-eruption a c. 73 ka, and its possible role in creating
a genetic bottleneck (Ambrose 1998; Rampino & Self 1993). Research projects
have recently been targeted at South Africa during MIS 6–2 ( Henshilwood 2008;
Henshilwood & Marean 2006; Jacobs et al. 2006; Marean et al. 2007; Jacobs &
Roberts 2008). These pilot studies have reached differing conclusions concerning
the influences of climatic variability on human behaviour, and it is clear that no
consensus yet exists on how best to integrate the wide variety of datasets from
multiple disciplines into a single analytical approach. This lack of consensus has
led researchers to propose a direct link between ice core records, regional climates
 Chapter 3. The origin of symbolically mediated behaviour 

and population dynamics (e.g. Jacobs & Roberts 2008). This approach may identify
general trends but it is clear that a better knowledge of the impact of climate on
continental ­ecosystems at a regional scale is necessary to model those processes
(c.f. Chase & Meadows 2007).
A similar situation exists in Europe where a variety of records have been
used to examine climate human interaction. One of the better known European
endeavours was the interdisciplinary Stage 3 Project (van Andel & Davies 2003),
but its results were hampered by the fact that the scholars relied only on t­ errestrial
climatic records and employed simplified palaeoclimatic reconstructions. Others
have performed similar research at both continental (Bocquet-Appel & Demars
2000; Bocquet-Appel et al. 2005) and regional scales (Blockley et al. 2006;
d’Errico & ­Sanchez-Goñi 2003; d’Errico et al. 2006; Finlayson 2004; Finlayson &
­Carrion 2007; Gamble et al. 2004; Stringer 2006; Sepulchre et al. 2007). In some
cases these research projects have involved more precise and up-to-date climate
models ­taking in due account the marine and terrestrial record (Sepulchre et al.
2007). In the framework of the ESF-funded RESOLuTION project, marine and
terrestrial palaeoclimatic records have been examined in order to reconstruct
palaeoenvironments more precisely (Ampel et al. 2008; Sanchez-Goñi et al.
2008; Fletcher & Sanchez-Goñi 2008) with an eye on how they impacted human
populations, but such research has concentrated exclusively on MIS 3–2.

5. Discussion

We should now move forward and find a consensual way to make progress in this
field. Our review highlights the need to develop an integrated research strategy
in which assumptions on cognition based on taxonomic affiliation will play no
role. The key tools should be archaeology, palaeoenvironmental studies and
new methods to integrate results from these disciplines. This appears the only
way to reconstruct the timing and mode of the emergence of key innovations
in material culture in Europe and southern Africa, to identify whether and how
climatic changes have influenced Neandertals and Homo sapiens’ distributions and
behavioural patterns, and to understand the mechanisms that governed cultural
transmission and social learning during this crucial time.
We argue that this requires (1) new archaeological excavations and dating
of archaeological sites in Africa and Europe located in areas associated with the
emergence and loss of cultural innovations. This need is more cogent in Africa where
some cultural phases are only represented by a few sites but is also necessary in
Europe where many key sites are undated; (2) the application of innovative methods
to the analysis of early symbolic and complex material culture of Homo sapiens and
 Francesco d’Errico & Christopher S. Henshilwood

Neandertals, including burials, abstract engravings, pigments, personal ornaments

and stylised bone and stone tools. A large portion of this crucial archaeological
material has never been properly analysed and one can reasonably predict that a
closer analysis may strengthen the symbolic interpretation, for example in the case
of pigments; (3) reconstruct climate, vegetation, fire regime and faunal changes in
Europe and southern Africa by combining the analysis of multiple proxies from
marine and terrestrial archives with high resolution palaeoclimatic simulations
and up-to-date analyses of faunal assemblages; (4) bring together archaeological
and palaeoenvironmental data, a coupled atmosphere-ocean General Circulation
Model (GCM) and subsequently derived high resolution atmospheric simulations,
and apply to them new tools that allows for the reconstruction, quantification and
comparison of the ecological niches exploited by human populations within each
climatic phase; (5) run realistic multi-agent based models to explore the impact
of the reconstructed climatic changes on population dynamics at a regional scale.
Such an integrated approach is now within our reach. Recent advances in
biodiversity studies have developed tools for estimating ecological niches of spe-
cies and predicting their responses to environmental changes. It has been demon-
strated that tools such as GARP (Genetic Algorithm for Rule Set Prediction) and
Maxent have considerable potential for reconstructing eco-cultural niches of past
human populations, defined as the potential range of environmental conditions
within which a human adaptive system can exist, and identifying the environmen-
tal parameters that define those niches (Banks et al. 2008, 2009). To do so, GARP
identifies palaeoclimatic features shared by archaeological sites with contempo-
raneous cultural levels, dated to a specific climatic event, and then finds other
regions with the same palaeoclimatic parameters. Through hundreds of error-
correcting iterations, it obtains an optimal eco-cultural niche reconstruction.
GARP also allows for a reconstructed eco-cultural niche for one climatic phase
to be projected onto the climatic conditions of a different one in order to evaluate
whether an eco-cultural niche has remained stable, expanded, contracted, or even
disappeared. This approach detects changes in niches and their respective ranges,
and identifies the environmental and cultural mechanisms behind them. GARP
has been already successfully applied for identifying and quantifying the mecha-
nisms behind human-environment interactions in Europe during the Last Glacial
Maximum and the Middle-to-Upper Palaeolithic transition (Banks et al. 2008).
If applied to the question at hand, this approach can potentially verify whether
the disappearance of key cultural innovations was linked to environmental shifts.
If so it may be possible to identify the underlying environmental factors and cul-
tural mechanisms that lie behind stable or shifting niches. If climate turns out
to have been a key factor in affecting the evolution and devolution of cultural
 Chapter 3. The origin of symbolically mediated behaviour 

i­nnovations in the past this will serve to remind us that cultural innovations are
not inscribed in our genes – climate change does not only affect what we can or
cannot do but it also specifically defines who we were and what we have become.

Acknowledgements

Financial support was provided to the authors by the European Research Council Advanced
Grant TRACSYMBOLS (FP7 No. 249587), and the PROTEA French-South Africa exchange
programme. CSH was also supported by a National Research Foundation/Department of Sci-
ence and Technology funded Chair at the University of the Witwatersrand, South Africa, the
Norwegian Research Council and the University of Bergen.

References

Ahn, J. & Brook, E.J. 2008. “Atmospheric CO2 and Climate on Millennial Time Scales During
the Last Glacial Period.” Science 322 (5898): 83–85.
Ambrose, S. 1998. “Late Pleistocene human population bottlenecks, volcanic winter, and
­differentiation of modern humans.” Journal of Human Evolution 34 (6): 623–651.
Ampel, L., Wohlfarth, B., Risberg, J. & Veres, D. 2008. “Paleolimnological response to m­ illennia
l and centennial scale climate variability during MIS 3 and 2 as suggested by the diatom
record in Les Echets, France.” Quaternary Science Reviews 27: 1493–1504.
Avery, D.M. 1995. “Physical environment and site choice in South Africa.” Journal of Archaeo-
logical Science 22: 343–353.
Backwell, L., d’Errico, F. & Wadley, L. 2008. “Middle Stone Age bone tools from the
Howiesons Poort layers, Sibudu Cave, South Africa.” Journal of Archaeological Science 35:
566–1580.
Bahn, P. 1996. “New developments in Pleistocene art.” Evolutionary Anthropology 4: 204–215.
Balme, J. & Morse, K. 2006. “Shell beads and social behaviour in Pleistocene Australia.” ­Antiquity
80 (310): 799–811.
Banks, W.E., d’Errico, F., Townsend Peterson, A., Masa Kageyama, M., Sima, A. & S­ anchez-Goñi,
M.F. 2008. “Neanderthal Extinction by Competitive Exclusion.” PLoS ONE 3 (12): 1–8.
Banks, W., Zilhão, J., d’Errico, F., Kageyama, M., Sima, A. & Ronchitelli, A. 2009. “Investigating
Links between Ecology and Bifacial Tool Types in Western Europe during the Last Glacial
Maximum.” Journal of Archaeological Science 36: 2853–2867.
Bar-Yosef Mayer, D., Vandermeersch, B. & Bar-Yosef, O. 2009. “Shells and ochre in Middle
Paleolithic Qafzeh Cave, Israel: indications for modern behaviour.” Journal of Human
­Evolution 56: 307–314.
Bar-Yosef, O. & Callender, J. 1999. “The woman from Tabun: Garrod’s doubts in historical
­perspective.” Journal of Human Evolution 37: 879–885.
Bateman, M.D., Holmes, P.J., Carr, A.S., Horton, B.P. & Jaiswal, M.K. 2004. “Aeolianite and
­barrier dune construction spanning the last two glacial–interglacial cycles from the
­southern Cape coast.” South Africa Quaternary Science Reviews 23: 1681–1698.
 Francesco d’Errico & Christopher S. Henshilwood

Bird, M.I., Fifield, L.K., Santos, G.M., Beaumont, P.B., Zhou, Y., di Tada, M.L. & Hausladen, P.A.
2003. “Radiocarbon dating from 40 to 60 ka BP at Border Cave, South Africa.” Q ­ uaternary
Science Review 22: 943–947.
Blockley, S.M., Donahue, R.E., Lane, C.S., Lowe, J.J. & Pollard, A.M. 2006. “The chronology of
abrupt climate change & Late Upper Palaeolithic human adaptation in Europe.”
­Journal of Quaternary Science 21: 575–584.
Blunier, T. & Brook, E.J. 2001. “Timing of millennial-scale climate change in Antarctica and
Greenland during the Last Glacial period.” Science 291: 109–112.
Bocquet-Appel, J.-P. & Demars, P.-Y. 2000. “Population kinetics in the Upper Paleolithic of
Western Europe.” Journal of Archaeological Science 27: 551–570.
Bocquet-Appel, J.-P., Demars, P.-Y., Noiret, L. & Dobrowsky, D. 2005. “Estimates of Upper
Paleolithic Meta-Population Size in Europe from Archaeological Data.” Journal of Archaeo-
logical Science 32: 1656–1668.
Bond, G., Broecker, W.S., Johnsen, S., McManus, J., Labeyrie, L.D., Jouzel, J. & Bonani, G. 1993.
“Correlations between climate records from North Atlantic sediments and ­Greenland ice.”
Nature 365: 143–147.
Bouzouggar, A., Barton, N., Vanhaeren, M., d’Errico, F., Collcutt, S., Higham, T., Hodge,
E., Parfitt, S., Rhodes, E., Schwenninger, J.-L., Stringer, C., Turner, E., Ward, S., Moutmir,
A. & Stambouli, A. 2007. 82 000-year-old shell beads from North Africa and implica-
tions for the origins of modern human behaviour. Proceedings of the National Academy of
­Sciences 104: 9964–9969.
Boyd, R. & Richerson, P.J. 1985. Culture and the Evolutionary Process. Chicago: University of
Chicago Press.
Byers, A.M. 1994. “Symboling and the Middle-Upper Palaeolithic transition: a theoretical and
methodological critique.” Current Anthropology 35: 369–399.
Byers, A.M. 1999. “Communication and Material Culture: Pleistocene Tools as Action Cues.”
Cambridge Archaeological Journal 9: 23–41.
Carr, A.S., Thomas, D.S.G. & Bateman, M.G. 2006. “Climatic and sea level controls on Late
Quaternary aeolian activity on the Agulhas Plain.” South Africa Quaternary Research 65:
252- 263.
Cavalli-Sforza, L.L. & Feldman, M. 1981. Cultural Transmission and Evolution. Princeton: Princ-
eton University Press.
Chase, B.M. & Meadows, M.E. 2007. “Late Quaternary dynamics of southern Africa’s winter
rainfall zone.” Earth Science Reviews 84 (3–4): 103–138.
Chase, P.G. 2006. The Emergence of Culture: the Evolution of a Uniquely Human Way of Life. New
York: Springer.
Clottes, J. 2003. Return to Chauvet Cave: Excavating the Birthplace of Art: the First Full Report.
London: Thames & Hudson.
Conard, N.J., Malina, M., Münzel, S.C. & Seeberger, F. 2004. Eine Mammutelfenbeinflöte aus
dem Aurignacien des Geissenklösterle: neue Belege für eine musikalische Tradition im
Frühen Jungpaläolithikumauf der Schwäbischen Alb. Archäologisches Korrespondenzblatt
34: 447–462.
Crevecoeur, I. 2006. Etude anthropologique des restes humains de Nazlet Khater (Paléolithique
supérieur, Egypte). Thesis (Ph.D.). University of Bordeaux 1.
d’Errico, F. 2003. “The invisible frontier. A multiple species model for the origin of behavioural
modernity.” Evolutionary Anthropology 12: 182–202.
d’Errico, F. & Villa, P. 1997. “Holes and grooves: the contribution of microscopy and t­ aphonomy
to the problem of art origins.” Journal of Human Evolution 33: 1–31.
 Chapter 3. The origin of symbolically mediated behaviour 

d’Errico, F., Henshilwood C. & Nilssen, P. 2001. “An engraved bone fragment from 70,000
­ka-year-old Middle Stone Age levels at Blombos Cave, South Africa: implications for the
origin of symbolism and language.” Antiquity 75: 309–318.
d’Errico F. & Sanchez-Goñi M. F. 2003. “Neandertal extinction and the millennial scale ­climatic
variability of the OIS 3.” Quaternary Science Reviews 22 (8–9): 769–788.
d’Errico, F., Henshilwood, C., Lawson, G., Vanhaeren, M., Soressi, M., Bresson, F., Tillier,
A.-M., Maureille, B., Nowell, A., Backwell, L., Lakarra J. A. & Julien, M. 2003. “The search
for the origin of symbolism, music and language: a multidisciplinary endeavour.” Journal
of World Prehistory 17: 1–70.
d’Errico, F., Henshilwood, C., Vanhaeren, M. & van Niekerk, K. 2005. “Nassarius kraussianus
shell beads from Blombos Cave: evidence for symbolic behaviour in the Middle Stone Age.”
Journal of Human Evolution 48: 3–24.
d’Errico, F. & Lawson, G. 2006. “The sound paradox. How to assess the acoustic significance
of archaeological evidence?” In Archaeoacoustics, C. Scarre & G. Lawson (eds), 41–57.
­Cambridge: McDonald Institute Monographs.
d’Errico, F., Sanchez-Goñi, M.F. & Vanhaeren, M. 2006. “L’impact de la variabilité climatique
rapide des OIS3-2 sur le peuplement de l’Europe.” In L’homme face au climat, E. Bard (ed.),
265–282. Paris: Odile Jacob.
d’Errico, F. & Henshilwood, C. 2007. “Additional evidence for bone technology in the ­southern
African Middle Stone Age.” Journal of Human Evolution 52: 142–163.
d’Errico, F. & Vanhaeren, M. 2007. “Evolution or revolution ? New evidence for the origin of
symbolic behaviour in and out of Africa.” In Rethinking the Human Revolution, P. Mellars,
K. Boyle, O. Bar-Yosef & C. Stringer (eds), 275–286. Cambridge: McDonald Institute
Monographs.
d’Errico, F., Vanhaeren, M. & Wadley, L. 2008. “Possible shell beads from the Middle Stone Age
layers of Sibudu Cave, South Africa.” Journal of Archaeological Science 35: 2675–2685.
d’Errico, F., Vanhaeren, M., Barton, N., Bouzouggar, A., Mienis, H. & Richter, D. 2009.
­“Additional evidence on the use of personal ornaments in the Middle Paleolithic of North
Africa.” PNAS 106 (38): 16051–16056.
d’Errico, F., Salomon, H., Vignaud, C. & Stringer, C. 2010. “Pigments from the Middle
­Palaeolithic levels of Es-Skhul (Mount Carmel, Israel).” Journal of Archaeological Science
37: 3099–3110.
d’Errico, F., Borgia, V. & Ronchitelli, M.-G. 2011. “Uluzzian bone technology and its i­ mplications
for the origin of behavioural modernity.” Quaternary International. doi: 10.1016/
j.quaint.2011.03.039.
d’Errico, F. & Stringer, C. 2011. “Evolution, Revolution or Saltation scenario for the emergence
of modern cultures?” Philosophical Transactions of the Royal Society B: Biological Sciences
366: 1060–1069.
Dansgaard, W., Johnsen, S.J., Clausen, H.B., Dahl-Jensen, D., Gundestrup, N.S., Hammer,
C.U., Hvidberg, C.S., Steffensen, J.P., Sveinbjörnsdottir, A.E., Jouzel, J. & Bond, G. 1993.
­“Evidence for general instability of past climate from a 250 kyr ice core record.” Nature
364: 218–220.
Deacon, J. & Lancaster, N. 1988. Late Quaternary palaeoenvironments of southern Africa. Oxford:
Clarendon Press.
Delagnes, A. & Meignen, L. 2005. “Diversity of lithic production systems during the M ­ iddle
Paleolithic in France.” In Transitions before the transition evolution and stability in the
Middle Paleolithic and Middle Stone Age, E. Hovers and S. L. Kuhn (eds), 85–107. USA:
Springer.
 Francesco d’Errico & Christopher S. Henshilwood

deMenocal, P.B. 2004. “African climate change and faunal evolution during the Pliocene-­
Pleistocene.” Earth and Planetary Science Letters 220: 3–24.
Diamond, J. 1978. “The Tasmanians: the longest isolation, the simplest technology.” Nature 273:
185–6.
Finlayson, C. 2004. Neanderthals and Modern Humans. An Ecological and Evolutionary
Perspective. Cambridge: Cambridge University Press.
Finlayson, C. & Carrión, J.S. 2007. “Rapid ecological turnover and its impact on Neanderthal
and other human populations.” Trends in Ecology and Evolution 22: 213–222.
Fletcher, W.J. & Sánchez-Goñi, M.F. 2008. “Orbital- and sub-orbital climate impacts on
­vegetation of the western Mediterranean basin over the last 48,000 yr.” Quaternary Science
Reviews 70: 451–464.
Gamble, C., Davies, W., Pettitt, P. & Richards, M. 2004. “Climate change and evolving human
diversity in Europe during the last glacial.” Philosophical Transactions of the Royal Society
of London B359: 243–254.
Goren-Inbar, N. 1990. Quneitra: A Mousterian Site on the Golan Heights. Jerusalem: Qedem 31.
Green, R.E., Krause, J., Briggs, A., Maricic, T., Stenzel, U., Kircher, M., Patterson, N., Weiwei
Zhai, H., Hsi-Yang Fritz, M., Hansen, N., Durand, E., Malaspinas, A., Jensen, J., Marques-
Bonet, T., Alkan, C., Prüfer, K., Meyer, M., Burbano, H., Good, J., Schultz, R., Aximu-
Petri, A., Butthof, A., Höber, B., Höffner, B., Siegemund, M., Weihmann, A., Nusbaum, C.,
Lander, E., Russ, C., Novod, N., Affourtit, J., Egholm, M., Verna, C., Rudan, P., ­Brajkovic,
D., Kucan, Z., Gušic, I., Doronichev, V., Golovanova, L., Lalueza-Fox, C., Rasilla, M., Fortea,
J., Rosas, A., Schmitz, R., Johnson, P., Eichler, E., Falush, D., Birney, E., Mullikin, J., Slatkin,
M., Nielsen, R., Kelso, J., Lachmann, M., Reich, D. & Pääbo, S. 2010. “A draft sequence of
the Neandertal genome.” Science 328 (5979): 710–722.
Grindley, J. 1969. “Quaternary marine palaeoecology in South Africa.” South African Archaeo-
logical Bulletin, Quaternary Studies in South Africa 24 (95/96): 151–157.
Grün, R. & Stringer, C.B. 2000. “Tabun revisited: revised ESR chronology and new ESR and
U-series analyses of dental material from Tabun C1.” Journal of Human Evolution 39 (6):
601–612.
Grün, R., Stringer, C., McDermott, F., Nathan, R., Porat, N., Robertson, S., Taylor, L., Mortimer,
G., Eggins, S. & McCulloch, M. 2005. “U-series and ESR analyses of bones and teeth relat-
ing to the human burials from Skhul.” Journal of Human Evolution 49: 316–334.
Heinrich, H. 1988. “Origin and consequences of cyclic ice rafting in the NE Atlantic Ocean
­during the past 130,000 years.” Quaternary Research 29: 142–152.
Henrich, J. 2004. “Demography and cultural evolution: How adaptive cultural processes can
produce maladaptive losses – the Tasmanian Case.” American Antiquity 69 (2): 197–214.
Henshilwood, C.S. 2008. “Winds of change: palaeoenvironments, material culture and human
behaviour in the Late Pleistocene (c. 77–48 ka) in the Western Cape Province, South Africa.”
South African Archaeological Society Goodwin Series 10: 35–5.
Henshilwood, C.S., d’Errico, F., Marean, C., Milo, W. & Yates, R. 2001. “An early bone tool
­industry from the Middle Stone Age at Blombos Cave, South Africa: implications for
the origins of modern human behaviour, symbolism and language.” Journal of Human
­Evolution 41: 631–678.
Henshilwood, C.S., d’Errico, F., Yates, R., Jacobs, Z., Tribolo, C., Duller, G.A.T., Mercier, N., Sealy,
J.C., Valladas, H., Watts I. & Wintle, A.G. 2002. “Emergence of modern human behavior:
Middle Stone Age engravings from South Africa.” Science 295: 1278–1280.
Henshilwood, C.S. & Marean, C.W. 2003. “The origin of modern human behaviour: a review
and critique of models and test implications.” Current Anthropology 44: 627–65.
 Chapter 3. The origin of symbolically mediated behaviour 

Henshilwood, C.S., d’Errico, F., Vanhaeren, M., Van Niekerk, K. & Jacobs, Z. 2004. “Middle
Stone Age shell beads from South Africa.” Science 304: 404.
Henshilwood, C.S. & Marean, C.W. 2006. “Remodelling the origins of modern human
­behaviour.” In The prehistory of Africa: tracing the lineage of modern man, H. Soodyall (ed.),
31–46. Johannesburg: Jonathan Ball Publishers.
Henshilwood, C.S., d’Errico, F. & Watts. I. 2009. “Engraved ochres from the Middle Stone Age
levels at Blombos Cave.” South Africa Journal of Human Evolution 57: 27–47.
Henshilwood, C.S. & Dubreuil, B. 2009. “Reading the artefacts: Gleaning language skills from
the Middle Stone Age in southern Africa.” In The Cradle of Language, R. Botha & C. Knight
(eds), 41–60. Oxford: Oxford University Press.
Henshilwood, C.S. & Dubreuil, B., 2011. “The Still Bay and Howiesons Poort, 77–59 ka:
­Perspective-taking and the evolution of the modern human mind during the African
­Middle Stone Age.” Current Anthropology.
Henshilwood, C.S. & d’Errico, F. this volume. “Middle Stone Age engravings and their s­ ignificance
to the debate on the emergence of symbolic material culture.” In ‘Homo Symbolicus’: The
Dawn of Language, Imagination, and Spirituality, C. Henshilwood & F. d’Errico (eds) 75–96.
Amsterdam: John Benjamins.
Hovers, E., Vandermeersh, B. & Bar-Yosef, O. 1997. “A Middle Palaeolithic engraved artefact
from Qafzeh Cave, Israel.” Rock Art Research 14: 79–87.
Hovers, E., Ilani, S., Bar-Yosef, O. & Vandermeersch, B. 2003. “An Early Case of Color ­Symbolism.
Ochre Use by Modern Humans in Qafzeh Cave.” Current Anthropology 44 (4): 492–522.
Hovers, E. & Belfer-Cohen, A. 2006. “Now you see it, now you don’t – modern human b ­ ehavior
in the Middle Paleolithic.” In Transitions before the transition: evolution and stability in the
Middle Paleolithic and Middle Stone Age, E. Hovers & S. L. Kuhn (eds), 205–304. New York:
Springer.
Imbrie, J. 1992. “On the structure and origin of major glaciation cycles, 1. Linear responses to
Milankovitch forcing.” Paleoceanography 7: 701–738.
Imbrie, J. 1993. “On the structure and origin of major glaciation cycles, 2. The 100,000-year
cycle.” Paleoceanography 8: 669–735.
Jacobs, Z., Duller, G.A.T., Wintle A. G. & Henshilwood. C. 2006. “Extending the chronology of
deposits at Blombos Cave, South Africa, back to 140 ka using optical dating of single and
multiple grains of quartz.” Journal of Human Evolution 51: 255–273.
Jacobs, Z. & Roberts, R.G. 2008. “Testing times: old and new chronologies for the Howieson’s
Poort and Still Bay Industries in environmental context.” South African Archaeological
S­ociety Goodwin Series 10: 9–34.
Jouzel, J., Masson-Delmotte, V., Cattani, O., Dreyfus, G., Falourd, S., Hoffmann, G., Minster,
B., Nouet, J., Barnola, J.M., Chappellaz, J., Fischer, H., Gallet, J.C., Johnsen, S., Leuenberger,
M., Loulergue, L., Luethi, D., Oerter, H., Parrenin, F., Raisbeck, G., Raynaud, D., Schilt,
A., Schwander, J., Selmo, E., Souchez, R., Spahni, R., Stauffer, B., Steffensen, J.P., Stenni, B.,
Stocker, T.F., Tison, J.L., Werner, M. & Wolff, E.W. 2007. “Orbital and millennial ­Antarctic
climate variability over the past 800,000 years.” Science 317: 793–797.
Klein, R.G. 1999. The human career. 2nd edition. Chicago (IL): Chicago University Press.
Klein, R. 2008. “Out of Africa and the evolution of human behaviour.” Evolutionary ­Anthropology
17: 267–281.
Klein, R. & Edgar, B. 2003. The Dawn of Human Culture. New York: John Wiley & Sons.
Knight, C., Powers, C. & Watts, I. 1995. “The human symbolic revolution: a Darwinian account.”
Cambridge Archaeological Journal 5 (1): 75–114.
 Francesco d’Errico & Christopher S. Henshilwood

Lambeck, K. & Chappell, J. 2001. “Sea level change through the Last Glacial cycle.” Science 292:
679–686.
Leavesley, M.G. 2007. “A shark-tooth ornament from Pleistocene Sahul.” Antiquity 81: 308–315.
Lombard, M. 2007. “The gripping nature of ochre: the association of ochre with Howiesons
Poort adhesives and Later Stone Age mastics from South Africa.” Journal of Human
­Evolution 53: 406–419.
Lorblanchet, M. 1999. La naissance de l’art. Paris: Errance.
Mackay, A. & Welz, A. 2008. “Engraved ochre from a Middle Stone Age context at Klein Kliphuis
in the Western Cape of South Africa.” Journal of Archaeological Science 35: 1521–1532.
Marean, C.W., Bar-Matthews, M., Fisher, E., Goldberg, P., Herries, A.I. R., Jacobs, Z., Jerardino,
A., Karkanas, P., Minichillo, T., Nilssen, P.J., Thompson, E., Watts I. & Williams, H.M. 2007.
“Early human use of marine resources and pigment in South Africa during the ­Middle
Pleistocene.” Nature 449 (7164): 905–908.
Marshack, A. 1981. “On Paleolithic ochre and the early uses of color and symbol.” Current
Anthropology 22: 188–191.
Marshack, A. 1996. “A Middle Palaeolithic symbolic composition from the Golan Heights: the
earliest known depictive image.” Current Anthropology 37: 357–365.
Mastronuzzi, G., Sansò, P., Murray-Wallace, C.V. & Shennan, I. 2005. “Quaternary coastal
morphology and sea-level changes – an introduction.” Quaternary Science Reviews 24
(18–19): 1963–1968.
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. this volume. “Pan Symbolicus: A Cultural Primatologist’s Viewpoint” In ‘Homo
Symbolicus’: The Dawn of Language, Imagination, and Spirituality, C. Henshilwood and F.
d’Errico (eds) 1–12. Amsterdam: John Benjamins.
Mellars, P. 1996. The Neandertal legacy: an archaeological perspective from western Europe.
­Princeton (NJ): Princeton University Press.
Mellars, P. 2006. “Why did modern human populations disperse from Africa ca. 60,000 years
ago? A new model.” Proceedings of the National Academy of Sciences 103: 9381–9386.
Mellars, P. 2007. “Rethinking the revolution: Eurasian and African perspectives” In Rethinking
the revolution: new behavioural and biological perspectives on the origin and dispersal of
modern humans, P. Mellars, K. Boyle, O. Bar-Yosef and C. Stringer (eds), 1–14. Cambridge:
McDonald Institute Monographs.
Mercier, N., Valladas, H., Froget, L., Joron, J.L. & Ronen, A. 2000. “Datation par la­thermolumi-
nescence de la base du gisement paléolithique de Tabun (Mont Carmel, Israël).” Comptes
Rendus de l’Académie des Sciences 330: 731–738.
Milankovitch, M.K. 1969. Canon of insolation and the Ice Age problem. Serbian Academy Beorg.
Special Publication 132.
Millard, A.R. 2006. “Bayesian analysis of ESR dates, with application to Border Cave.” ­Quaternary
Geochronology 1 (2): 159–166.
Morse, K. 1993. “Shell beads from Mandu Mandu Creek rock-shelter, Cape Range peninsula,
Western Australia, dated before 30,000 ka BP.” Antiquity 67: 877–883.
Partridge, T.C. 1993. “Warming phases in Southern Africa during the last 150,000 years – an
overview.” Palaeogeography, Palaeoclimatology, Palaeoecology l01: 237–244.
Partridge, T.C. 1997. “Cainozoic environmental change in southern Africa, with special empha-
sis on the last 200 000 years.” Progress in Physical Geography 21 (1): 3–22.
Pearson, M.P. 1999. The Archaeology of Death and Burial. Phoenix Mill: Sutton.
 Chapter 3. The origin of symbolically mediated behaviour 

Potts, R. 1996. Humanity’s descent: the consequences of ecological instability. New York: William
Morrow and Co.
Powell, A., Shennan, S. & Thomas, M. 2009. “Late Pleistocene demography and the a­ ppearance
of modern human behavior.” Science 324 (5932): 1298–1301.
Premo, L. & Kuhn S. 2010. “Modeling Effects of Local Extinctions on Culture Change and
Diversity in the Paleolithic.” PLoS ONE 5 (12): 15582. doi:10.1371/ journal.pone.0015582.
Rampino, M.R. & Self, S. 1993. “Climate – Volcanism feedback and the Toba eruption of ca.
74000 years ago.” Quaternary Research 40: 269–280.
Richerson, P., Bettinger, R. & Boyd, R. 2005. “Evolution on a Restless Planet: Were E­ nvironmental
Variability and Environmental Change Major Drivers of Human ­Evolution?” In Handbook
of Evolution, Vol. 2: The Evolution of Living Systems (Including Hominids), F. Wuketits and
F. Ayala (eds), 223–242. Weinheim: Wiley.
Richerson, P., Boyd, R. & Bettinger, R. 2009. “Cultural Innovations and Demographic Change.”
Human Biology 81 (2–3): 211–235.
Rigaud, J.-P., Texier, P.-J., Parkington, J. & Poggenpoel, C. 2006. “Le mobilier Stillbay et
­Howiesons Poort de l’abri Diepkloof: la chronologie du Middle Stone Age Sud-Africain et
ses implications.” Comptes Rendus, Palevol 5: 839–849.
Rivers, W. 1912. “The disappearance of useful arts.” Festschrift Tillagnad Edvard Westermarck,
Helsingfors 117–130.
Roberts, D.L., Bateman, M.D., Murray-Wallace, C.V., Carr, A.S. & Holmes, P.J. 2008. “Last inter-
glacial fossil elephant trackways dated by OSL/AAR in coastal aeolianites, Still Bay, South
Africa.” Palaeogeography, Palaeoclimatology, Palaeoecology 257 (3): 261–279.
Ronchitelli, A., Boscato, P. & Gambassini, P. 2009. “Gli ultimi neandertaliani in Italia: aspetti
culturali.” In La lunga storia di neandertal. Biologia e comportamento, F. Facchini &
M. Belcastro (eds), 257–287. Milano: Jaca Book.
Savage-Rumbaugh, E.S. & Fields, W.M. this volume. “The Evolution and the Rise of Human
­Language.” In ‘Homo Symbolicus’: The Dawn of Language, Imagination, and Spirituality,
C. Henshilwood and F. d’Errico (eds) 13-47. ­Amsterdam: John Benjamins.
Salomon, H. 2009. Les matières colorantes au début du Paléolithique supérieur: sources,
­transformations et functions. Thesis (Ph.D.). University of Bordeaux.
Sanchez-Goñi, M.F., Landais, A., Fletcher, W.J., Naughton, F., Desprat, S. & Duprat, J.
2008. “Contrasting impacts of Dansgaard-Oeschger events over a western European
latitudinal transect modulated by orbital parameters.” Quaternary Science Reviews 27:
1136–1151.
Scholz, C.A., Johnson, T.C., Cohen, A.S., King, J.W., Peck, J.A., Overpeck, J.T., Talbot,
M.R., Brown, E.T., Kalindekafe, L., Amoako, P.Y., Lyons, R.P., Shanahan, T.M.,
Castañeda, I.S., Heil, C.W., Forman, S.L., McHargue, L.R., Beuning, K.R., Gomez, J. &
Pierson, J. 2007. East African megadroughts between 135 and 75 thousand years ago and
bearing on early-modern human origins. Proceedings of the National Academy of Sciences
104: 16416–16421.
Sepulchre, P., Ramstein, G., Fluteau, F., Schuster, M., Tiercelin, J-J. & Brunet, M. 2006. “­ Tectonic
Uplift and Eastern Africa aridification.” Science 313 (5792): 1419–1423.
Sepulchre, P., Ramstein, G., Kageyama, M., Vanhaeren, M., Krinner, G., Sánchez-Goñi, M.F.
& d’Errico, F. 2007. H4 abrupt event and late Neanderthal presence in Iberia. Earth and
Planetary Science Letters 258: 283–292.
Shennan, S. 2001. “Demography and cultural innovation: a model and its implications for the
emergence of modern human culture.” Cambridge Archaeological Journal 11: 5–16.
 Francesco d’Errico & Christopher S. Henshilwood

Soressi, M. & d’Errico, F. 2007. “Pigments, gravures, parures: les comportements symboliques
controversés des Néandertaliens.” In Les Néandertaliens. Biologie et cultures, B. Vandermeersch
and B. Maureille (eds), 297–309. Paris: Éditions du CTHS.
Stringer, C. 2006. Homo Britannicus: The incredible story of human life in Britain. London:
Penguin.
Texier, J.-P., Porraz, G., Parkington, J., Rigaud, J.-P., Poggenpoel, C., Miller, C., Tribolo, C.,
­Cartwright, C., Coudenneau, A., Klein, R., Steele, T. & Vernai. C. 2010. A Howiesons Poort
tradition of engraving ostrich eggshell containers dated to 60,000 years ago at ­Diepkloof
Rock Shelter, South Africa. Proceedings of the National Academy of Sciences doi: 10.1073/
pnas.0913047107.
Thackeray, J.F. 1992. “Chronology of Late Pleistocene deposits associated with Homo sapiens at
Klasies River Mouth, South Africa.” Palaeoecology of Africa 23: 177–191.
Thackeray, J.F. 2007. Sea levels and chronology of Late Pleistocene coastal cave deposits at
­Klasies River in South Africa. Annals of the Transvaal Museum 44: 219–220.
Van Andel, T.H. 1989. “Late Pleistocene sea levels and the human exploitation of the shore and
shelf of southern South Africa.” Journal of Field Archaeology 16: 133–154.
Van Andel, T.H. & Davies, W. (eds). 2003. Neanderthals and modern humans in the European
landscape during the last glaciation: archaeological results of the Stage 3 Project. Cambridge:
McDonald Institute for Archaeological Research, University of Cambridge.
Vandermeersch, B., Cleyet-Merle, J.-J., Jaubert, J., Maureille, B. & Turc, A. 2008. Première
humanité, gestes funéraires des Néandertaliens. Paris: Réunion des Musées Nationaux.
Vanhaeren, M. & d’Errico, F. 2006. “Clinal distribution of personal ornaments reveals the
­ethno-linguistic geography of Early Upper Palaeolithic Europe.” Journal of Archaeological
Science 33 (8): 1105–1128.
Vanhaeren, M., d’Errico, F., Stringer, C., James, S.L., Todd, J. & Mienis, H.K. 2006. “Middle
­Palaeolithic shell beads in Israel and Algeria.” Science 312 (5781): 1785–1788.
Velo, J. 1984. “Ochre as medicine: a suggestion for the interpretation of the archaeological
record.” Current Anthropology 25 (5): 674.
Vermeersch, P.M., Gijselings, G. & Paulissen, E. 1984. “Discovery of the Nazlet Khater man,
Upper Egypt.” Journal of Human Evolution 13: 281–286.
Vermeersch, P.M., Paulissen, E., Stokes, S., Charlier, C., Van Peer, P., Stringer, C.B. & Lislay, W.
1998. “A Middle Palaeolithic burial of a modern human at Taramsa Hill, Egypt.” Antiquity
72: 475–484.
Villa, P., Soressi, M., Henshilwood, C.S. & Mourre, V. 2009. “The Still Bay points of Blombos
Cave (South Africa).” Journal of Archaeological Science 36 (2): 441–460.
Wadley, L. 2001. “What is cultural modernity? A general view and a South African perspective
from Rose Cottage.” Cambridge Archaeological Journal 11 (2): 201–221.
Wadley, L., Williamson, B.S. & Lombard, M. 2004. “Ochre in hafting in Middle Stone Age
­southern Africa: a practical role.” Antiquity 78: 661–675.
Wadley, L., Hodgskiss, T. & Grant, M. 2009. Implications for complex cognition from the­h­afting
of tools with compound adhesives in the Middle Stone Age, South Africa. P ­ roceedings of the
National Academy of Sciences 106: 9590–9594.
Watts, I. 1999. “The origin of symbolic culture.” In The evolution of culture, R. Dunbar, C. Knight
and C. Power (eds), 113–146. Edinburgh: Edinburgh University Press.
Watts, I. 2002. “Ochre in the Middle Stone Age of southern Africa: ritualized display or hide
preservative?” South African Archaeological Bulletin 57: 1–14.
Wendt, W.E. 1976. “‘Art mobilier’ from the Apollo 11 Cave, South West Africa: Africa’s oldest
dated works of art.” South African Archaeological Bulletin 31: 1170–1199.
 Chapter 3. The origin of symbolically mediated behaviour 

Wiessner, P. 1983. “Style & social information in Kalahari San projectile points.” American
Antiquity 48: 253–276.
Yellen, J.E., Brooks, A.S., Cornelissen, E., Mehlman, M. & Stewart, K. 1995. “A Middle Stone Age
worked bone Industry from Katanda, Upper Semliki Valley, Zaire.” Science 268: 553–556.
Zilhão, J. & d’Errico, F. 1999. “The chronology and taphonomy of the earliest Aurignacian and its
implications for the understanding of Neandertal extinction.” Journal of World ­Prehistory
13 (1): 1–68.
Zilhão, J., Angelucci, D., Badal-Garcíac, E., d’Errico, F., Daniel, F., Dayet, L., Douka, K., Higham,
T., Martínez-Sánchez, M., Montes-Bernárdez, R., Murcia-Mascarósj, S., Pérez-Sirventh, C.,
Roldán-Garcíaj, C., Vanhaeren, M., Villaverde, V., Wood, R. & Zapata, J. 2010. ­“Symbolic use
of marine shells and mineral pigments by Iberian Neandertals.” PNAS 107 (3): 1023–1028.
 chapter 4

Middle Stone Age engravings

and their significance to the debate
on the emergence of symbolic material culture
Christopher S. Henshilwooda,b & Francesco d’Erricoc,a
aInstitute
for Archaeology, History, Culture and Religion, University of Bergen/
bInstitute
for Human Evolution, University of the Witwatersrand/CCNRS-UMR
5199 PACEA, Université de Bordeaux

Archaeological evidence associated with modern cognitive abilities provides

important insights into when and where modern human behaviour emerged.
Modern human behaviour here means the thoughts and actions spontaneously
shaped by minds equivalent to those of Homo sapiens today. Key among these is
the use of symbols. Three models for the origins of behavioural modernity are
current: (i) a late and rapid appearance at ~ 40–50 ka associated with the European
Upper Palaeolithic and the Later Stone Age (LSA) of sub-Saharan Africa, (ii) an
earlier and more gradual evolution rooted in the African Middle Stone Age (MSA
~ 300–40 ka), (iii) a discontinuous evolution rooted in both the African Middle
Stone Age and the Mousterian of Eurasia. Material evidence for modern behaviour
before 40 ka in Africa was, until a few years ago, relatively rare and often regarded
as ambiguous compared to evidence from the Upper Palaeolithic or LSA. However,
in sub-Saharan Africa archaeological evidence for changes in technology, economy
and social organization and the emergence of symbolism in the Middle Stone Age
contradicts the first model. Examples of these changes include standardized formal
lithic tools, shaped bone implements, innovative subsistence strategies, evidence
for personal ornaments and the deliberate engraving of abstract designs on ochre,
ostrich eggshell and bone. In this chapter we review the earliest evidence for
purposely made engravings from southern Africa and discuss their significance
for arguments favouring early advances in human cognition.

1. Introduction

In past decades, anthropologists, archaeologists, linguists, and cognitive scien-

tists assumed that the production of symbolic material culture, and its presumed
involvement in the mediation of social behaviour, was a recent innovation in
human history and the probable outcome of a stochastic event that took place after
c. 40 ka. Instances of non-utilitarian behaviours older than 40 ka from ­Neandertal
 Christopher S. Henshilwood & Francesco d’Errico

sites in Eurasia, and Homo sapiens sites in Africa were considered ambiguous with
respect to their significance and/or their anthropogenic nature. Some archaeolo-
gists now argue that ‘symbolically mediated behaviour’ (for definition see Wadley
2001; Henshilwood & Marean 2003, 2006; Henshilwood & Dubreuil 2009) evolved
gradually in Africa, and link its origins with the emergence of Homo sapiens in
that continent (e.g. Knight et al. 1995; Watts 1999, 2009; McBrearty & Brooks
2000; Henshilwood et al. 2002, 2004, 2009; Henshilwood & Dubreuil, 2011) define
symbolically mediated culture as “one in which individuals understand that arte-
facts are imbued with meaning and that these meanings are construed and depend
on collectively shared beliefs. This criterion is crucial. It explains how human norms
and conventions differ from the ritualized behaviours found in nonhuman primates”.
The tenets of the ‘gradual emergence in Africa model’ are that the appearance of
modern cognitive behaviours in Eurasia, associated with Homo sapiens, is the
end result of an ‘out of Africa’ dispersal of an already symbolic species at c. 60 ka
or ­earlier (e.g. Henshilwood & Marean 2003, 2006; Forster 2004; Mellars 2006).
Other authors argue that Neandertal behaviour may have independently evolved
along similar lines and favour demographic events triggered by climate change
instead of a speciation event as the main factor accounting for the emergence of
symbolic mediated behaviours in Africa and Eurasia (Zilhão 2001; d’Errico 2003;
d’Errico et al. 2003; see d’Errico & Henshilwood, this volume).
New evidence for symbolic behaviour from MSA sites in Africa and
­Mousterian sites in Europe and the Near East is challenging these models. The
evidence for the archaeological expression of symbolically mediated behaviours
now includes geometric and iconographic representations, elaborate burials, and
personal ornaments and, to some extent, pigment use and complex bone and
stone tools (see d’Errico & Henshilwood, this volume for a review of the ­evidence).
In this chapter we focus on the earliest evidence for deliberate e­ ngravings from
southern Africa, explore their symbolic nature, and discuss their significance for
the debate on the origins of cultural modernity. Before presenting the archaeo-
logical evidence and its context, we address the question of why engravings are
important for the symbol related debate and what information they may yield
that is relevant to this debate.
In many respects engravings represent a special category of material ­culture.
Pigments (colloquially called ochre by most archaeologists in Africa) found at
archaeological sites are generally the by-product of a sequence of actions but
­recreating these sequences is problematic. It is one of the main reasons for the
interpretation of pigments being a controversial topic. Different actions, for exam-
ple grinding pigment, that were made with or without symbolic intent may well
produce a similar end product. Personal ornaments found at archaeological sites,
for example shell beads, could be described as the disposed elements of a string
 Chapter 4. Middle Stone Age engravings and their significance 

of signs that once was meaningful. How these ornaments were worn or arranged
on the human body is now irretrievable. Perishable ornaments, perhaps in wood
or on leather, have not survived to help us with our reconstruction of past mate-
rial culture. Burials, virtually unknown for the MSA in Africa, may help with the
reconstruction of funerary practices in the past (Duday 2009) but we know from
ethnographic accounts of burials that archaeologists are only able to retrieve a
fraction of the whole picture from site remains. Technological complexity and
morphological standardisation in stone and bone tool production help with
assessing the cognitive abilities of our ancestors, but again disentangling stylistic
and symbolic elements from purely functional actions is often difficult.
An engraving is the end product of incising or pecking a design onto a
­material, such as pigment or stone with the aid of a tool. Engravings are perhaps
the only category of potentially symbolic early material culture that still reflects
the complete set of cutting actions performed by the artist. As is the case with
drawing, engraving reflects deep unconscious feelings of self (Freeman & Cox
1985; Thomas & Silk 1990) while, at the same time, engravings “organise” a shared
visual culture (Cox et al. 2001). Contrary to other early symbolic manifestations,
engravings can be formally described, compared and their differences measured
from a variety of perspectives (Bosinski et al. 2001). The microscopic analysis
of the lines that make up engraved patterns, including junctions and crossings,
provide information on the chronology of gestures employed, the use of a single
vs multiple tools, engraver laterality, and possible changes in the orientation of
the engraved surface during the production of incisions. It can also indicate the
handedness and relative location of the engraver to the piece at the time of the
incisions. Experimental reproduction of these processes helps to understand
these past actions in fine detail (d’Errico 1996; Fritz 1999). By combining those
data it is possible to evaluate the technical competence of the engraver, and,
to some extent, to establish if the engraver followed a mental template during
production.
A brief summary of the evidence for engravings in the MSA shows that
­overall, only a few isolated pieces are reported and that very few sites have more
than one or two examples. An engraved stone with a cross-hatched pattern was
recovered from presumed Howiesons Poort levels at the Kleinkliphuis River
site, Western Cape (Mackay and Welz, 2008; Figure 1a); in 1963 an engraved
pebble was found by Wolfgang Sydow at a site containing only MSA lithics near
­Palmenhorst, Namibia (Wendt 1975; Figure 1b). Parallel line engravings on stone
are reported from Cape Hangklip (Heese n.d.; Figure 1c) and from an undated
MSA context at Bushman Rock Shelter, Mpumulanga (Watts 1998). Dendritic/
convergent line motifs were found in MSA II and Howiesons Poort (HP) contexts
at Klasies, E
­ astern Cape (Knight et al. 1995; Watts 1998) and from c. 100 ka ­levels
 Christopher S. Henshilwood & Francesco d’Errico

at Pinnacle Point, W ­ estern Cape (Watts, 2010). Similar parallel lines engraved
on bone f­ ragments are found at Klasies and at Blombos Cave (BBC) (d’Errico &
Henshilwood 2007). A striking collection of engraved items comes from BBC
and Diepkloof Rock Shelter (Figures 2, 3, 4 and 5). At the former site at least 14
pieces of ochre engraved with deliberate abstract patterns come from c. 75 ka
and 100 ka levels (Henshilwood et al. 2002, 2009). At the latter site 270 engraved
ostrich e­ ggshell fragments were found in HP levels with an age of c. 60 ka (Texier
et al. 2010). In this chapter we focus on these two sites.

a b c

scale=1cm

Figure 1. (a) engraved pebble from reportedly Howiesons Poort levels at Kleinkliphuis River,
Western Cape, South Africa (after Mackay & Welz, 2008); (b) decorated pebble from MSA
site near Palmenhorst, Namibia recovered by Wolfgang Sydow in 1963; (c) worked ochre
pencil from the Still Bay site at Cape Hangklip, Western Cape described by Frans Malan
(after Heese n.d., papers at the Iziko-South African Museum. The location of this artefact is
currently unknown)

2. Howiesons Poort techno-tradition

2.1 Diepkloof Rock Shelter

2.1.1 Archaeological context and dating
Diepkloof Rock Shelter (DRS) is a quartzitic sandstone shelter located 180
km north of Cape Town and 18 km from the Atlantic (Figure 2c). It overlooks
the ­Verlorenvlei River. The site was excavated in the 1990s by a team from the
­University of Cape Town. A Franco-South African team have excavated the site
since 1998. The DRS stratigraphic sequence consists of laminations of anthropo-
genic material, mostly composed of burnt and non-burnt organic residues, ash
associated with detrital sands and diagenetic components, and rich faunal and
lithic assemblages (Texier et al. 2010; Figure 2b). The 2.5 m deep MSA sequence
is classified, from top to bottom, as Post-Howiesons Poort, Howiesons Poort
(HP), Still Bay (SB), and Pre-Still Bay. Thermoluminescence (TL) and optically
a b c
Blombos Cave Diepkloof Rock Shelter
290 BP 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180190 200
K6 L6 M6
LSA 50 50
2 ka
60 60
Claude
Hiatus DUN

POST
70 70
70 ka MSA
LSA
phases 80 80 HP

CA
90 Darryl 90
CB M1 Deon
CC
c. 72 ka 100 100
CCC Ester Ester
CD
CDA 110 110
Eve
M2 Frans
120 Fred 120
CFA upper
c. 78 ka Frank

EOES ZONE
130 130
CFB/CFC Fox
Fox Fiona
CGAA M2 140 Fiona 140
CGAB lower
CGAC Governor
c. 85 ka 150 150

HOWIESONS POORT
CH John
M3 160 160
CI upper Jeff Jeff
c. 100 ka 170 170
CJ Joy
Jack
180 180

50 cm

Figure 2. (a) Blombos Cave stratigraphy and ages; (b) Diepkloof Rock Shelter stratigraphy; (c) location of Blombos Cave (BBC) and Diepkloof Rock
Shelter (DRS)
Chapter 4. Middle Stone Age engravings and their significance

 Christopher S. Henshilwood & Francesco d’Errico

s­ timulated luminescence (OSL) techniques provide an age for these sequences that
range from c. 130 ka to 45 ka.
TL dates of 74–60 ka (Parkington 1999) are provided for the HP at DRS but it
was thought these dates may be too old (Rigaud et al. 2006). A subsequent set of
ages for the HP levels, calculated by the TL method, range from c. 96 ka – 60 ka
­(Tribolo et al. 2009). These are not concordant with Jacobs et al.’s (2008) OSL ages
of 58.1 ± 1.9 ka to 63.3 ± 2.2 ka (also Jacobs et al. 2008). The reason for the differ-
ences in the ages reported by Jacobs et al. (2008) and Tribolo et al. (2009) has not
been resolved. In summary, the most consistent ages for HP in southern Africa
sites fall within the range of 65–59 ka.

3. Diepkloof engraved ostrich eggshells

Ostrich eggshell fragments are found in all the MSA levels although ostrich
bones are absent, despite excellent preservation of faunal remains. Ostrich eggs
are a regular dietary item at Diepkloof, as is the case at most other MSA sites
in southern Africa (e.g. Henshilwood et al. 2001a; Klein et al. 2004). About 270
ostrich eggshell fragments from the upper levels of the HP (c. 60 ka) are engraved
with abstract patterns (e.g. Figure 3) and three fragments have punctured open-
ings on the apical part of the egg (e.g. Figure 3a). One of these has engraved lines
that diverge from the aperture. The fragmentation of the e­ ggshells into pieces
20–30 mm in size is attributed to post depositional processes. Reconstructing
the original decoration on a whole egg is thus problematic but four distinct
motifs, across the range of eggs, have been identified (Figure 3): (i) two parallel
lines intersected at right angles by regularly spaced lines forming ladder-like
patterns (Figure 3 b,c,d,e,f); (ii) a series of deeply engraved, straight, sub-parallel
lines(Figure 3 g,h); (iii) slightly curved lines crossing a central line (Figure 3i);
(iv) cross-hatched lines (Figure 3a).
The first of these motifs is recorded on 69 fragments that come only from
the lower HP levels. The second motif is most common in the upper levels and
the third and fourth motifs are present on one piece each. The third motif comes
from a level in the middle of the HP and the provenance of the fourth motif is
not reported. (Texier et al. 2010). Microscopic analysis of the crossing of the lines
on the ladder motif pieces suggests to Texier et al. (2010) that the production of
this pattern was standardised as the engraving of the long parallel lines was done
first, then the shorter lines were made that cross over the longer lines. There is
considerable variability in the spacing between the hatched lines and the angles
of intersection.
 Chapter 4. Middle Stone Age engravings and their significance 

a b c

1 cm

d e f

g h i

Figure 3. Engraved ostrich eggshell from the c. 60 ka Howiesons Poort levels at Diepkloof
Rock Shelter, southern Cape

Texier et al. (2010) compared the HP evidence for ostrich eggshell use
with that of historical hunter-gatherers, mainly Khoesan, who used the eggs
as flasks to store and transport water. Many of these were also engraved. The
conclusion of the team’s study is that the engraved fragments from the HP levels
probably came from eggs used as storage containers. Considering the diversity
of motifs, their stylistic variability and the numbers of fragments, the authors
suggest that the eggshell pieces came from at least 25 whole eggs. They argue
that the signature of the designs on the shells is different to those on engraved
ochre and bones from other MSA sites; specifically that the motifs served no
technical purpose; that they are standardized and systematized; and that this
 Christopher S. Henshilwood & Francesco d’Errico

favoured the development of ­specific geometric rules. The authors state further
that as the engraved pieces come from 18 stratigraphic levels and as there are
changes in engraved patterns across these levels, that the engraving of eggs
can be interpreted as evidence for a tradition that lasted for several thousand
years. They conclude that these engraved eggs provide compelling evidence for
communication using symbols.

4. Still Bay techno-tradition

4.1 B
 lombos Cave
4.1.1 Archaeological context and dating
Blombos Cave is situated on the southern Cape coast, 300 km east of Cape Town
(Figure 2c). The MSA levels at the site are divided into three phases M1, M2 and M3
(Henshilwood et al. 2001a) (Figure 2a). The five uppermost layers below a s­ terile
level named BBC Hiatus are assigned to the M1 phase. The M1 and Upper M2
phase lithics are typified by Still Bay type bifacial foliate points, the fossile directeur
of the SB Industry (Goodwin and van Riet Lowe 1929; Villa et al. 2009) and­­
end- and side- scrapers (Henshilwood et al. 2001a). Engraved ochre and bone
(d’Errico et al. 2001; Henshilwood et al. 2002, 2009; Watts 2009), ­formal bone tools
(Henshilwood et al. 2001b) and forty nine Nassarius shell beads (Henshilwood et al.
2004; d‘Errico et al. 2005) come from this phase. The lower M2 phase is t­ ypified by
carbonised deposits, and includes shellfish and fauna, but these layers are ­generally
low intensity deposits. High density shellfish deposits, hearths, faunal remains,
many lithics and ochre pieces dominate most levels in the M3 phase.
The SB levels have been dated using a number of methods (Jones 2001;
Jacobs et al. 2003a,b, 2006; Tribolo et al. 2006). The Hiatus level composed of
undisturbed aeolian sand above the M1 phase is dated by OSL to 69 ± 5 ka and
70 ± 5 ka (Henshilwood et al. 2002: Jacobs et al. 2003 a,b; Jacobs et al. 2006)
(Figure 2a) and provides a minimum age for the SB deposits at the site. An OSL
age of 72.7 ±3.1 ka was obtained for the upper part of the SB M1 phase (Jacobs et
al. 2003 a,b). Thermoluminescence (TL) dates for the M1 and Upper M2 phase
indicate that 74 ± 5 and 78 ± 6 ka are the likely ages for these SB levels (Tribolo
et al. 2006). OSL dates for the lower M2 phase fall between 84.6 ± 5.8 ka to 76.8
± 3.1 ka (Jacobs et al. 2006) (Figure 2). Similar ages were obtained using the
electron spin resonance method (Jones 2001). The lower M2 phase (CG levels)
does not contain artefacts associated with the SB. The upper M2 phase does
­contain these SB markers and the inference drawn is that the age of 76.8 ± 3.1 for
the CF level (Jacobs et al. 2006) (Figure 2a) should be regarded as the terminus
 Chapter 4. Middle Stone Age engravings and their significance 

post quem for the SB levels at BBC. Using single aliquots, an OSL age of 98.9 ±
5.5 ka was obtained for the upper portion of level CI, representing the upper
M3 phase. The level below, CJ, has provided a date of 143.2 ± 5.5 ka by the same
method (Jacobs et al. 2006); however, this came from a sub-unit of sterile sand
within CJ and cannot, therefore, be treated as a reliable date for the a­ rchaeology.
The ­stratigraphic integrity of artefacts recovered from these levels has been
­demonstrated and there is minimal evidence for movement of artefacts between
the MSA phases (Jacobs et al. 2003 a, b; Henshilwood 2006; H ­ enshilwood et al.
2009).

5. Blombos engraved ochre

In total 1,534 pieces of ochre ≥10mm in length with a net weight of 5,581 g have
been recovered from the MSA levels at BBC (Henshilwood et al. 2009). Most of
the ochre comes from the M3 phase (78.6% of pieces, 82% of mass) and over-
all the quantity of ochre from BBC is considerably greater than that recorded for
most southern African MSA sites (Henshilwood et al. 2001a, 2009; Watts 2009).
Within an approximately fifty kilometre radius of BBC, the most likely sources
of ochre are outcrops of Bokkeveld Group deposits (predominantly comprised of
shale and siltstone) (Henshilwood et al. 2001a; Watts 2009). At present the nearest
such ­outcrops are adjacent to the Goukou and Duiwenhoks Rivers (respectively 19
km east and 21 km west of BBC), but even a small drop in sea level (as occurred
following the interglacial transgression of MIS 5e), may have exposed Bokkeveld
shales, now buried under marine deposits and aeolian sands, within c. 3–5 km of
the site (Watts 2009; see also Rogers 1988: 411).
Modification of ochre was primarily directed at producing a powder, either
by grinding pieces across an abrasive surface or by scraping them using a stone
tool. Grinding results in flat or slightly convex facets covered by multiple ­fusiform
striations (d’Errico & Nowell 2000; d’Errico & Backwell 2003). Scraping ­produces
parallel striations of different width and depth resulting from projections on the
working surface of the implement. Repeated applications of the scraping tool,
in the same area, result in wide grooved or concave areas with fringed ends
­corresponding to individual exit points.
Some pieces recovered at BBC bear striations inconsistent with a natural,
­incidental or “technical” interpretation and, we have recently argued ­(Henshilwood
et al. 2009), represent instances of unequivocally intentional engraving (Figures 4
and 5; Table 1). None of the grooves present on the BBC pieces fit the c­ riteria for
geological or animal markings. Incidental markings left by trampling, common
on the engraved ostrich eggshells from Diepkloof, result in randomly oriented
 Christopher S. Henshilwood & Francesco d’Errico

straight or slightly curved individual striations and smoothing of the exposed areas.
The incised BBC ochres are well preserved and show no evidence of t­aphonomic
damage including striations indicative of natural abrasion. The incisions ­cannot
be interpreted as the outcome of cutting motions. A flat stone used as a base
on which to cut soft material will need to be of a size compatible with the task
(Binford 1978). This requirement eliminates the likelihood that the pieces of
ochre from BBC could have been used as cutting-boards. Lines left on cutting-
boards are generally straight, sub-parallel, overlapping, and oriented on the long
axis of the board, a feature not recorded at Blombos. Microscopic analysis of the
BBC ­incisions shows that they were made by the point of a stone tool and not by
a ­cutting edge (d’Errico 1995; Fritz 1999; d’Errico et al. 2001; Alvarez et al. 2001).
The tip of a stone tool is not effective for repeated cutting as it blunts rapidly, yet it
is effective to precisely engrave lines. Tool sharpening does not account for any of
the ­incisions on the ochre pieces as the knapped stone tools from the MSA layers of
BBC were not sharpened by grinding. The small size of the incisions on the ochre
pieces indicates that they are not compatible with arrow/spear shaft straightening.
Notched bones used to scrape soft material such as vegetable or hide are found in
the MSA levels at Klasies River (d’Errico & Henshilwood 2007). However, none
of the incised ochre pieces shows a wear pattern consistent with this kind of use.
If ochre pieces were deeply incised or sawn to assist in splitting them there is no
evidence of these techniques having been applied as the recorded lines are too
­shallow for this purpose.
Testing an ochre piece for powder quality or shade only requires one or two
strokes, however while multiple lines are present on most of the incised BBC
ochres. Further, in a number of cases incisions were made on a ground surface,
making further testing unnecessary. To obtain a useable quantity of powder
from an ochre piece requires grinding across an abrasive surface or scraping it
repeatedly with a tool. On some of the pieces examined there is evidence of these
modifications (e.g. M1/1, M1/3, M1/5, M1/6, M3/2, M3/6, M3/9, M3/3 and M3/8
conjoins: Figures 4 and 5), but these are not consistent with the production of
ochre powder. Many incisions are narrow and superficial and would have resulted
only in a tiny quantity of powder being produced. In some cases (M1/4, M3/1,
M3/10: Figures 4 and 5) these narrow incisions are the only modification visible
on the surface of the ochre, which contradict this interpretation. Further, many of
the lines interpreted as deliberate engravings (M1/5, M1/6, M3/1, M3/6, M3/10:
Figures 4 and 5) are different to those made when scraping ochre for powder
extraction. The latter typically consists of a single wide deep groove with fringed
ends, created by a relatively unrestrained back and forth motion, or multiple­
sub-parallel deep grooves produced by repeated motions in one direction within
the same plane.
 Chapter 4. Middle Stone Age engravings and their significance 

a b

1 cm
d
1 cm
c

1 cm
1 cm
e

1 cm

1 cm

Figure 4. Engraved ochre from the M1 phase (78–72 ka) at Blombos Cave
 Christopher S. Henshilwood & Francesco d’Errico

a b

1 cm

c d

1 cm

e
1 cm
f

1 cm

g h
1 cm

Figure 5. Engraved ochre from the M2 and M3 phases (c. 85–100 ka) at Blombos Cave
 Chapter 4. Middle Stone Age engravings and their significance 

Table 1. Description of Blombos engraved ochre pieces

Number & provenience Brief description Type of engraving

of piece (see Henshilwood et al. 2009)

M1 phase
(quadrant and layer)
M1-1 (E5a BZ) Flake off a larger piece Two groups of incisions, one on the centre
(Figure 4) of ochre and one close to the edge. In the centre two
joining lines form a ‘Y’ that is crossed by a few
perpendicular parallel lines. Three incisions
cross these lines.
M1-2 (F6a, BZB) Fragment of a larger Two lines that cross perpendicularly on the
(Figure 4) piece of ochre top right margin. Converging lines produced
with a single lithic point.
M1-3 (G6a, CAB) Ochre flake Retains only a small area of the original
(Figure 4) engraved pattern. Three straight oblique lines
incised on the top left with two sinuous lines
that cross them.
M1-4 (E10d, CA); Ochre flake reduced by Three distinct sets of lines engraved on a
(Figure 4) knapping natural surface. Piece was then knapped and a
part of the engraving removed.
M1-5 (E6a, CC) Ochre slab ground on A group of sinuous lines engraved on one
(Figure 4) both main faces and face. The opposite face is highly scraped and
three edges engraved with a cross-hatched pattern.
M1-6 (H6a, CD) Relatively large Cross-hatched pattern incised on one
(Figure 4) rectangular piece of long edge.
ochre (166.6 g)
M2 Phase
M2-1(G5a, CFB/CFC) Tabular piece of ochre One face is crossed vertically by deep striae
(Figure 5) incised from top to bottom that converge at
the base.
M3 Phase
M3-1(F5c, CH/CI) Tiny shard of ochre A remnant of the original engraved surface is
(Figure 5) fractured off a larger retained on one edge. The engravings consist
piece of two parallel lines crossed obliquely by three
lines.
M3-2 (E5b, CH/CI), M3–5 Three conjoining pieces Four deep parallel single-stroke lines engraved
(F5c,CH/CI), M3-11 from a larger ochre slab using the same stone tool.
(F5b, CJ)
(Figure 5)
M3-4(F5c, CH/CI) Small abraded piece of Engraved with four single stroke lines.
(Figure 5) ochre
M3-6 (F5c, CH/CI) Eroded fragment of Two sets of parallel incisions aligned at
(Figure 5) ochre right-angles to each other are engraved.
M3-7(G6a, CH/CI) Flat fragment of ochre On one surface a group of conjoining lines are
(Figure 5) incised. This dendritic pattern is flanked by a
single deep line on the right.

(Continued)
 Christopher S. Henshilwood & Francesco d’Errico

Table 1. Description of Blombos engraved ochre pieces (Continued)

Number & Provenience Brief description
of piece
M3-8 (E4, CH/CI), M3-3 Three conjoined pieces
(E5b, CH/CI) of ochre that come
(Figure 5) from an originally
thick tabular piece
M3-9 (E4, CH/CI) A relatively large
(Figure 5) rectangular slab of
ochre
M3-10 (E5a, CJ) Pinkish-grey rounded
(Figure 5) slab of siltstone
Type of engraving
(see Henshilwood et al. 2009)
Apparent crenellated pattern made with
different lithic tools. It is difficult to
interpret the result as having been made
with deliberate intent.
One edge is engraved with a cross-hatched
pattern. There is a possible dendritic motif
on one side.
One surface displays a long pair of sub-
parallel single stroke lines. Four shorter
convergent single stroke lines are engraved
on the left and follow the same direction.
Two of these lines end at the base of a
slight, but steep rise in the surface.

On a number of pieces (M1/2, M1/5, M1/6, M3/1, M3/6, M3/10; Figures 4

and 5) regularity in the profile and outline of incisions indicates precise neuro-
motor control. Good examples are M1/6 and M3/1 (Figures 4 and 5). On the
former the engraver filled in a blank space by incising two lines to complete the
symmetry of the pattern; in the latter case the start and end points of some lines
were placed to exactly conjoin with existing lines.
Clearly, the engraved lines that form the motifs were carefully controlled by
the makers during their execution and this action now assists in gauging the intent
of the artist. Microscopic analysis of the engraved patterns allows us to identify
consistency in the engraving techniques, the direction in which the lines were
incised, and their chronological order (M1/3, M1/5, M1/6, M3/1, M3/6, M3/10).
In sum, although a degree of ambiguity is certainly implied in our interpretation
of some pieces, the results indicate that clear evidence of unambiguous engraved
patterns are present in the MSA levels at BBC. Conservatively, this conclusion can
be extended to include at least seven pieces coming from different layers (M1/2,
M1/4, M1/5, M1/6, M3/1, M3/6, M3/10, and possibly the conjoins M3/3, M3/8:
Figures 4 and 5).

6. Contrasting the significance of early engravings

The discovery of tool made markings on ostrich eggshells and on ochre pieces
at MSA sites raises a number of questions including: (i) were the markings on
these objects produced with the intent of embodying material culture with sym-
bolic meaning?; (ii) were the DRS and BBC engravings perceived as ­symbolic by
 Chapter 4. Middle Stone Age engravings and their significance 

the inhabitants of these sites at the time of their production or perhaps just the
­artisan; (iii) what was the main purpose of these markings and how may they
have functioned with a symbolically mediated society – if this was the case?
Definitions of symbols, and how they act as behavioural mediators when they
are recorded on a variety of media, varies according to disciplines, epistemological
stands, and individual scholars (see for example, Jung & De Laszlo 1958; Savage-
Rumbaugh 1986; Savage-Rumbaugh et al. 1986; Savage-Rumbaugh, this volume;
McGrew, this volume; Tomasello 2008). A ‘broad’ definition for a symbol is
something that represents something else by association, resemblance, or
convention. A more ‘strict’ definition, that we adopt here, denotes a sign that has
no natural connection or resemblance to its referent (Peirce 1998).
Archaeologists are unable to formally demonstrate that in prehistoric
human societies meaning was associated to what we now perceive to be abstract
depictions (Hodder 1982; Stark et al. 1993; Robb 1998). Arguments in favour
of a symbolic hypothesis for these depictions rely on, and often try to combine,
observations and inferences of various natures and heuristic power. Evaluating
the pertinence of those inferences is central to understanding what we can
say (and cannot say) about possible early instances of symbolically mediated
behaviours.
Various explanations are suggested to support the symbolic interpretation
of early engravings: (i) an absence of obvious functional reasons behind the
production of the engravings; (ii) consistencies in the media on which the
engravings are made; (iii) the preparation of the surface prior to engraving;
(iv) the degree of neuromotor control inferred from the analysis of each line;
(v) the type of tool used; (vi) the use of the same tool for the production of
the entire pattern; (vii) the consistent organisation of the sequence of motions
articulating the marking action; (viii) the regularity of the resultant pattern; (ix)
the presence of engravings on a number of objects rather than on a single one;
(x) the repetition of the same motif on more than one object; (xi) variations
within what is perceived as the same basic motif; (xii) the production of a variety
of different motifs; (xiii), temporal continuity in the production of engravings on
the same media; (xiv) persistence or change in the production of motifs through
time; (xv), production of similar engravings on the same media at a number of
sites; and (xvi), similarity in the media used for engraving by prehistoric, extant
and/or historically known groups. Although none of these arguments proves
that the engravings or the objects carried a specific meaning for past artisans and
their associated band or group that may differ to our current interpretation, we
can draw on these arguments to help rule out alternative hypotheses, highlight
possible inconsistencies in our interpretations, and help in making balanced
analyses.
 Christopher S. Henshilwood & Francesco d’Errico

At BBC and DRS there is continuity, over time, in the engravings being made
on the same types of raw material. At both sites the chaine opératoire employed
for engraving on ochre or eggshell shows that the artisans were deliberately
producing engraved designs – in other words they were not accidental. But just
how ­deliberate these designs were, and how much planning went into executing
them, and whether they conformed to personal or social templates, remains
elusive. We previously argued against the “doodling” and the “notational”
interpretation in our discussion of the BBC engravings on ochre and bone
(Henshilwood et al. 2009). Doodling is broadly defined as unfocused drawing
made while a person’s attention is otherwise occupied. Incising lines on hard
ochre requires focused attention so that the right pressure is applied to keep the
depth of the incision constant. Both hands are required to perform this action,
one to hold the pieces, which are small in the case of BBC, and one to engrave.
A “notation” can be defined as a ­marking system specifically conceived to record,
store and recover information outside the physical body (d’Errico 1995, 2001).
Such a system must allow for a clear d ­ istinction between the marks that carry
specific information. This is not present in most of the BBC engravings as the
individual marks cannot be visually identified as ­discrete signs. Also, no evidence
exists of sequential markings produced by different tools that may be interpreted
as a notation based on an accumulation of information over time. At both sites
each engraving was made in a single session, the makers had a clear goal when
producing an engraved pattern, and they showed interest in the end result of
their design. For DRS the conclusions could be tempered by the limited number
of published pieces and by the very few pieces that have been technologically or
microscopically analysed.
At both sites templates are apparent on a number of engraved patterns, for
example the BBC cross-hatched designs and the DRS ladder-like motifs. The
latter, however, present a large range of morphological variation that apparently
shows no obvious conformity. In some engravings the “rungs” only cover three
quarters of the space between the “members” of the ladder leaving a large empty
space, in others they overstep them, there are three or more “members” instead
of two, or some have intermediate members between “ladders” with no rungs.
Texier et al. (2010) attribute this diversity to personal variability in the production
of the same motif. It is possible that these engravings belong either to a more
discrete, still not categorised template, or that the template was not actually a
single category in the minds of the artisans. This means there could have been a
communication system in place in which different engravings did not function
as distinct s­ymbols. We suggest that the whole object could have conveyed a
­symbolic message, i­ rrespective of the type of engraving.
 Chapter 4. Middle Stone Age engravings and their significance 

Choosing between alternative hypotheses is complicated by the few

­published images of the engraved pieces of eggshell and by inconsistencies in
assigning motifs to pieces. For example, the more common pattern, the ladder
motif, a­ pparently appears only on 69 of the 270 engraved fragments (Texier et al.
2010: 3; SOM table S1–2). No count is given for the fragments with the second
most f­requent pattern, the parallel lines motif. As many of the pieces are very
small, it may hamper attributing a particular motif to a specific piece. For the two
remaining motifs, the line intersecting two sets of “s” shaped parallel lines and
the criss-cross pattern, there is only one fragment for each, but the latter is not
depicted in the publication.
A consensual exploration of the number and proportion of motifs depicted on
eggshell containers requires comprehensive publication of all the engraved fragments
from DRS, including their contextual information, photographs and technological
analysis. This would also allow evaluation of some of the other arguments proposed
by the authors of the study to support their symbolic interpretation. The apparently
high number of engraved eggshells at DRS can be used to argue in favour of
engraving being a consistent behaviour at the DRS site compared to the only 8
engraved ochre pieces from BBC. However, most DRS fragments are very small
and a dozen or more could come from one egg. The method used by the authors for
estimating that the recovered fragments come from at least 25 eggs is not explained.
The surface of an ostrich egg (Paganelli & Olszowka 1974) and that of the recovered
fragments can easily be calculated to establish the minimum number of eggs that
produced the fragments. Even if 25 is the correct number for the number of ostrich
eggs that were decorated in the HP levels then this sample size is not significantly
different from that of the engraved ochres at BBC. The ‘consistent behaviour’
argument is therefore not, in our opinion, stronger for the DRS engraved shells
than could be made for the engraved BBC ochres.
The similarity that the DRS engraved shells have with eggshell containers
used by extant or historical Khoesan is cited by Texier et al. (2010) as strong
support for symbolic meaning once being attached to the HP engraved eggs.
But, it is also apparent that in recent contexts, engravings could also simply
signal ownership and may not be symbolic. The dangers associated with using
the ethnographic present to explain the prehistoric past abound. Binford (1978)
invoked “Middle Range Theory” to provide explicitly scientific explanations for
the patterning of faunal remains in the Nunamiut archaeological record. There is
wide ranging agreement that Binford’s approach and particularly linking middle
range theory to a body of general theory about human behaviour is problematic
as there is no consensus for the latter within archaeology (e.g. Hodder, 1982;
Gould and Yellen, 1987). Do the similarities of the Khoesan engraved eggs to
 Christopher S. Henshilwood & Francesco d’Errico

those in the HP at DRS allow for an argument suggesting cultural continuity or

symbolic links? We question this hypothesis.
Labels such as “ownership” or “context” have rich and complex meanings in
ethnographic societies. Such meanings and their relationship with the production
and use of the engravings need to be explored in depth before any generalisation
is proposed. If the ownership function is correct then one would expect to find a
variety of visually distinct markings identifying different owners. This is not the
case at DRS if the categorisation of motifs is correct, since apart from two exceptions
only two main categories of patterns are present. Perhaps the owner/maker of an
egg could be identified by group members based on the variation within one large
“motif ”, and also that a particular “motif ” reflected more the cultural affiliation of
the group than of a particular individual. If correct, this supports our contention
that the categories of motifs are too flexible to be meaningful with regard to ‘marks
of ownership’.
The argument for continuity of the HP engraving traditions with those of the
present Khoesan requires that engraved eggs are found at sites from 60 ka up to
the ethnographic present. This is not the case as engraved eggshells are not found
at southern African archaeological sites until at least 40 ka years later. In our view
one explanation is that the engraved eggshell tradition identified at DRS could
be regarded as a cultural innovation independent from more recent instances of
engravings on the same media. In this respect the DRS evidence supports and
shares a strong similarity with cultural innovations recorded at other MSA sites.
An example is the marine shell beads found within layers older than 70 ka at
archaeological sites in North and South Africa, and in the Near East (see d’Errico
and Henshilwood, this volume).
In conclusion, formally demonstrating that the engravings from DRS and BBC
were made with symbolic intent is problematic, but we highlight in this chapter a
number of consistencies in these engraved pieces that point the evidence in that
direction. The main difference between the engraved records from these two sites is
that the DRS eggshells, contrary to the engraved ochres at BBC were clearly made
for visual display and were probably recognized as such by the members of the
group in which they were used, and perhaps by the regional bands or groups that
shared these symbolic meanings. On the other hand the symbolic codes expressed
by engravings on BBC ochre are discreet, almost secret, and may be better explained
as being a part of communication that involved only few individuals – perhaps even
only at the level of the exchange of a gift between two people. What is clear, as more
archaeological evidence becomes available, is that methods of communication
and the use of symbols to mediate behaviour in the southern African MSA was
complex, with perhaps regional variation, and that our understanding of the
symbolic behaviours of these early Homo sapiens is still rudimentary.
 Chapter 4. Middle Stone Age engravings and their significance 

Acknowledgements

Financial support was provided to the authors by the European Research Council Advanced
Grant TRACSYMBOLS (FP7 No. 249587), and the PROTEA French-South Africa exchange
programme. CSH was also supported by a National Research Foundation/Department of
Science and Technology funded Chair at the University of the Witwatersrand, South Africa a
joint Norwegian Research Council/South African National Research Foundation Grant and the
University of Bergen.

References

Alvarez, M., Fiore, D., Favret, E. & Guerra C. R. 2001. “The use of lithic artefacts for m
­ aking rock
art engravings: observation and analysis of use-wear traces in experimental tools through
optical Microscopy and SEM.” Journal of Archaeological Science 28: 457–464.
Binford, L.R. 1978. Nunamiut ethnoarchaeology. New York: Academic Press.
Bosinski, G. d’Errico, F. & Schiller, P. 2001. Die gravierten Frauendarstellungen von G ­ onnersdorf
(Der Magdalonien-Fundplatz Gonnersdorf Vol. 8). Stuttgart: Franz Steiner.
Cox, M.V., Koyasu, M., Hiranuma, H. & Perara J. 2001. “Children’s human figure drawings in the
UK and Japan: The effects of age, sex and culture.” British Journal of Developmental Psychology
19 (18): 275–292.
d’Errico, F. 1995. L’art gravé azilien. De la technique à la signification. XXXIème. Supplément à
Gallia-Préhistoire. Paris: CNRS Editions.
d’Errico, F. 1996. “Image analysis and 3-D optical surface profiling of Upper Palaeolithic
­mobiliary art.” Microscopy and Analysis 39: 27–29.
d’Errico, F. 2003. “The invisible frontier. A multiple species model for the origin of behavioural
modernity.” Evolutionary Anthropology 12: 182–202.
d’Errico F. & Nowell A. 2000. “A new look at the Berekhat Ram figurine: implications for the
origins of symbolism.” Cambridge Archaeology Journal 10 (1): 123–167.
d’Errico F., Henshilwood, C.S. & Nilssen, P. 2001. “An engraved bone fragment from ca. 75
kyr Middle Stone Age levels at Blombos Cave, South Africa: implications for the origin of
­symbolism.” Antiquity 75: 309–18.
d’Errico, F. & Backwell, L.R. 2003. “Possible evidence of bone tool shaping by Swartkrans early
hominids.” Journal of Archaeological Science (30) 12: 1559–1576.
d’Errico, F., Henshilwood, C., Lawson, G., Vanhaeren, M., Soressi, M., Bresson, F., Tillier, A.M.,
Maureille, B., Nowell, A., Backwell, L., Lakarra, J.A. & Julien, M. 2003. “The search for the
origin of symbolism, music and language: a multidisciplinary endeavour.” Journal of World
Prehistory 17 (1): 1–70.
d’Errico, F., Henshilwood, C., Vanhaeren, M. & van Niekerk, K. 2005. “Nassarius ­kraussianus
shell beads from Blombos Cave: Evidence for symbolic behaviour in the Middle Stone
Age.” Journal of Human Evolution 48: 3–24.
d’Errico, F. & Henshilwood, C. 2007. “Additional evidence for bone technology in the ­southern
African Middle Stone Age.” Journal of Human Evolution 52 (2): 142–163.
d’Errico, F., Henshilwood, C.S. this volume “The origin of symbolically mediated behaviour:
from antagonistic scenarios to a unified research strategy.” In ‘Homo Symbolicus’: The
 Christopher S. Henshilwood & Francesco d’Errico

Dawn of Language, Imagination, and Spirituality, C. Henshilwood and F. d’Errico (eds),

49–73. Amsterdam: John Benjamins.
Duday, H. 2009. The Archaeology of the dead. Oxbow Books, Oxford.
Forster, P. 2004. Ice ages and the mitochondrial DNA chronology of human dispersals: a review.
Philosophical Transactions of the Royal Society of London B 359, 255–264.
Freeman, N.H. & Cox, M.V. 1985. Visual order: the nature and development of pictorial
­representation. Cambridge: Cambridge University Press.
Fritz, C. 1999. La Gravure Dans l’art Mobilier Magdalénien. Du Geste à la Représentation. Paris:
Documents d’Archéologie Française 75, Editions de la Maison des Sciences de l’Homme.
Goodwin, A.J.H. & van Riet Lowe, C. 1929. The Stone Age cultures of South Africa. Annals of
the South African Museum 27: 1–289.
Gould, R. & Yellen, J. 1987. “Man the hunted: determinants of household spacing in desert and
tropical foraging societies.” Journal of Anthropological Archaeology 6: 77–103.
Heese, C.H.T. n.d. Collected papers of C. H. T. D. Heese, on file at Iziko, South African Museum,
Cape Town.
Henshilwood, C.S. 2006. “Stratigraphic integrity of the Middle Stone Age levels at Blombos
Cave.” In From tools to symbols. From early hominids to modern humans, L. Backwell and F.
d’Errico (eds), 441–458. Johannesburg: Witwatersrand University Press.
Henshilwood, C.S., Sealy, J.C., Yates, R.J., Cruz-Uribe, K., Goldberg, P., Grine, F.E., Klein, R.G.,
Poggenpoel, C., van Niekerk, K.L. & Watts, I. 2001a. “Blombos Cave, southern Cape, South
Africa: Preliminary report on the 1992–1999 excavations of the Middle Stone Age levels.”
Journal of Archaeological Science 28 (5): 421–448.
Henshilwood, C.S., d’Errico F., Marean, C., Milo, R. & Yates, R. 2001b. “An early bone tool
­industry from the Middle Stone Age at Blombos Cave, South Africa: implications for
the origins of modern human behaviour, symbolism and language.” Journal of Human
­Evolution 41: 631–678.
Henshilwood, C.S., d’Errico, F., Yates, R., Jacobs, Z., Tribolo, C., Duller, G.A.T., Mercier N.,
Sealy, J.C., Valladas, H., Watts, I. & Wintle, A.G. 2002. “Emergence of Modern Human
Behaviour: Middle Stone Age engravings from South Africa.” Science 295: 1278–1280.
Henshilwood, C.S. & Marean, C.W. 2003. “The origin of modern human behaviour: A review
and critique of models and test implications.” Current Anthropology 44 (5): 627–651.
Henshilwood, C.S., d’Errico, F., Vanhaeren, M., van Niekerk, K. & Jacobs, Z. 2004. “Middle
Stone Age shell beads from South Africa.” Science 384: 404.
Henshilwood, C.S. & Marean, C.W. 2006. “Remodelling the origins of modern human
­behaviour.” In The prehistory of Africa: tracing the lineage of modern man, H. Soodyall (ed),
31–46. Johannesburg: Jonathan Ball Publishers.
Henshilwood, C.S., d’ Errico, F. & Watts, I. 2009. “Engraved ochres from the Middle Stone Age
levels at Blombos Cave, South Africa.” Journal of Human Evolution 57: 27–47.
Henshilwood, C.S. & Dubreuil, B. 2009. “Reading the artefacts: Gleaning language skills from
the Middle Stone Age in southern Africa.” In The cradle of language, R. Botha and C. Knight
(eds), 41–60. Oxford: Oxford University Press.
Henshilwood, C.S. & Dubreuil, B. June 2011. “The Still Bay and Howiesons Poort, 77–59ka:
Symbolic Material Culture and the Evolution of the Mind during the African Middle Stone
Age.” Current Anthropology (52) 3: 361–400.
Hodder, I. (ed.). 1982. Symbolic and Structural Archaeology. Cambridge: Cambridge University
Press.
Jacobs, Z., Duller G.A.T. & Wintle, A.G. 2003b. “Optical dating of dune sand from Blombos
Cave, South Africa: II – single grain data.” Journal of Human Evolution 44: 613–625.
 Chapter 4. Middle Stone Age engravings and their significance 

Jacobs, Z., Wintle, A.G. & Duller, G.A.T. 2003a. “Optical dating of dune sand from Blombos
Cave, South Africa: I—Multiple Grain Data.” Journal of Human Evolution 44: 599–612.
Jacobs, Z., Duller, G.A.T., Henshilwood, C.S. & Wintle, A.G. 2006. “Extending the chronology of
deposits at Blombos Cave, South Africa, back to 140 ka using optical dating of single and
multiple grains of quartz.” Journal of Human Evolution 51: 255–273.
Jacobs, Z., Roberts, R.G., Galbraith, R.F., Deacon, H.J., Grun, R., Mackay, A., Mitchell, P., Vogel-
sang, R. & Wadley, L. 2008. “Ages for the Middle Stone Age of Southern Africa: ­Implications
for Human Behavior and Dispersal.” Science 322 (5902):733–735.
Jones, H.L. 2001. Electron spin resonance dating of tooth enamel at three Palaeolithic sites. M.Sc.
dissertation, McMaster University.
Jung, C.G. & De Laszlo, V.S. 1958. Psyche and Symbol: A Selection from the Writings of C.G. Jung.
New York: Doubleday.
Klein, R.G., Avery, G., Cruz-Uribe, K., Halkett, D., Parkington, J., Steele, T., Volman, T. & Yates,
R. 2004. “The Ysterfontein 1 Middle Stone Age Site, South Africa, and Early Human
Exploitation of Coastal Resources.” Proceedings of the National Academy of Sciences 101
(16): 5708–5715.
Knight, C., Powers, C. & Watts, I. 1995. “The human symbolic revolution: a Darwinian account.”
Cambridge Archaeological Journal 5 (1): 75–114.
Mackay, A. & Welz, A. 2008. “Engraved ochre from a Middle Stone Age context at Klein
Kliphuis in the Western Cape of South Africa.” Journal of Archaeological Science 35 (6):
1521–1532.
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. this volume. “Pan symbolicus. A cultural primatologist’s viewpoint.” In ‘Homo
Symbolicus’: The Dawn of Language, Imagination, and Spirituality, C. Henshilwood and
F. d’Errico (eds), 1–12. Amsterdam: John Benjamins.
Mellars, P. 2006. “Why did modern human populations disperse from Africa ca. 60,000
years ago? A new model.” Proceedings of the National Academy of Sciences 103 (25):
9381–9386.
Paganelli, C.V. & Olszowka, A.A. 1974. “The avian egg: surface, area, volume, and density.” The
Condor 76: 319–325.
Parkington, J. 1999. “Western Cape landscapes.” Proceedings of the British Academy, Oxford:
Oxford University Press.
Peirce, C.S., 1998. The essential Peirce: selected philosophical writings. Bloomington: Indiana
­University Press.
Rigaud, J–P., Texier, P–J., Parkington, J. & Poggenpoel, C. 2006. “South African Middle Stone
Age Chronology: New Excavations at Diepkloof Rock Shelter; Preliminary Results.”
­Palaeovol 5: 839–849.
Robb, J. E.1998. “The archaeology of symbols.” Annual Review of Anthropology 27: 329–346.
Rogers, J. 1988. “Stratigraphy and geomorphology of three generations of regressive sequences
in the Bredasdorp group, southern Cape Province, South Africa.” In ­Geomorphological
studies in southern Africa, G.F. Dardis and B.P. Moon (eds); 407–433. Rotterdam:
Balkema.
Savage-Rumbaugh, S. 1986. Ape language: from conditioned response to symbol. New York:
Columbia University Press.
Savage-Rumbaugh, S., McDonald, K., Sevcik, R.A., Hopkins, W.D. & Rubert, E. 1986.
­“Spontaneous symbol acquisition and communicative use by pygmy chimpanzees (Pan
paniscus).” Journal of Experimental Psychology: General 115 (3): 211–235.
 Christopher S. Henshilwood & Francesco d’Errico

Savage-Rumbaugh, S. & Fields, W.M., this volume. “The evolution and the rise of human lan-
guage.” In ‘Homo Symbolicus’: The Dawn of Language, Imagination, and Spirituality, C.
Henshilwood and F. d’Errico (eds), 13–47. Amsterdam: John Benjamins
Stark, M.P., Yoffe, A. & Sherrat, A. 1993. Archaeological theory: who sets the agenda? ­Cambridge:
Cambridge University Press.
Texier, J.-P., Porraz, G., Parkington, J., Rigaud, J.-P., Poggenpoel, C., Miller, C., Tribolo, C.,
Cartwright, C., Coudenneau, A., Klein, R., Steele, T. & Vernai, C. 2010. “A Howiesons
Poort tradition of engraving ostrich eggshell containers dated to 60,000 years ago at Diep-
kloof Rock Shelter, South Africa.” Proceedings of the National Academy of Sciences 107
(14): 6180–6185.
Thomas, G.V. & Silk, A.M.J. 1990. An introduction to the psychology of representation in ­children.
New York: Harvester Wheatsheaf.
Tomasello, M. 2008. Origins of human communication. Cambridge, Mass.: MIT Press.
Tribolo, C., Mercier, N., Selo, M., Joron, J–L., Reyss, J–L., Henshilwood, C., Sealy, J. & Yates,
R. 2006. “TL dating of burnt lithics from Blombos Cave (South Africa): further evidence
for the antiquity of modern human behaviour.” Archaeometry 48 (2): 341–357.
Tribolo, C., N. Mercier, H. Valladas, J.L. Joron, P. Guibert, Y. Lefrais, M. Selo, P.J. Texier, J.P. Rigaud
& G. Porraz. 2009. “Thermoluminescence dating of a Stillbay–Howiesons Poort sequence
at Diepkloof Rock Shelter (Western Cape, South Africa).” Journal of ­Archaeological Science
36 (3): 730–739.
Villa, P., Soressi, M., Henshilwood, C.S. & Mourre, V. 2009. “The Still Bay points of Blombos
Cave (South Africa).” Journal of Archaeological Science 36 (2): 441–460.
Wadley, L. 2001. “What is cultural modernity? A general view and a South African perspective
from Rose Cottage Cave.” Cambridge Archaeological Journal 11: 201–221.
Watts, I. 1998. The origins of symbolic culture: The southern African Middle Stone Age and Khoisan
ethnography. Ph.D. thesis: University of London.
Watts, I. 1999. “The origin of symbolic culture.” In The Evolution of Culture, R. Dunbar, C.
Knight & C. Power (eds), 113–146. Edinburgh: Edinburgh University Press.
Watts, I. 2009. “Red ochre, body-painting, and language: interpreting the Blombos ochre.” In
The Cradle of Language, R. Botha & C. Knight (eds), 62–92. Oxford: Oxford University
Press.
Watts, I. 2010. “The pigments from Pinnacle Point Cave 13B, Western Cape, South Africa.”
­Journal of Human Evolution 59 (6): 641–656.
Wendt, W.E. 1975. “Notes on some unusual artefacts from South West Africa.” Cimbebasia Ser.
B 2 (6), 180–185.
Zilhão, J. 2001. Anatomically archaic, behaviorally modern: the last Neanderthals and their
­destiny. Amsterdam: Amsterdams Archeologisch Centrum.
 chapter 5

Complex cognition required for compound

adhesive manufacture in the Middle Stone
Age implies symbolic capacity
Lyn Wadley
Institute for Human Evolution and School of Geography, Archaeology and
Environmental Studies, University of the Witwatersrand

Compound adhesives were made in the Middle Stone Age (MSA) of southern
Africa using a complex process in order to attach stone tools to shafts.
Glue-makers must switch attention between fire control, measuring ingredients,
and assembling compound tools. No rehearsals or recipes guarantee success;
artisans must multi-task (a characteristic of modern human minds involving
cognitive fluidity) and think abstractly about properties of glue ingredients.
Mental rotation, an ability implying advanced working memory capacity, is
needed to place stone inserts in various positions to create novel weapons and
tools. The manufacturing process is sufficiently intricate that the early artisans
must have had minds with abilities that overlapped those of people living today;
such minds are capable of symbolic thought and action.

1. Introduction

The name Homo symbolicus describes us well; almost everything we do has sym-
bolic undertones. It is therefore not surprising that we expect people like us in the
past to have depended equally on symbolism, which would have demanded the
use of some form of language. It is an entirely reasonable assumption. Archaeolo-
gists like to use symbolic material culture as a marker in the deep past for behav-
iour that overlaps with that of people living today. Nonetheless, few archaeologists
agree on definitions of symbolism, symbolic material culture or even of attributes
that imply modern behaviour (Wadley 2001; Henshilwood & Marean 2003). The
presence of art is often taken to be an indicator of symbolism in the past, yet art
historians themselves are not agreed on what is or is not art. Personal ornaments
are thought to be markers of symbolism because they imply marking of personal
or group identity. Shell beads with ages of between ~80 and 70 ka have been found
 Lyn Wadley

at three Middle Stone Age (MSA) sites from distant corners of Africa: Taforalt,
Morocco (Bouzouggar et al. 2007), Sibudu, KwaZulu-Natal (d’Errico et al. 2008)
and Blombos, Western Cape (d’Errico et al. 2005). Most archaeologists consider
these beads as persuasive evidence for early symbolism, but Wynn et al. (2009)
represent the perspective that these archaeological traces are not smoking guns
for concept formation and do not require that their makers had language. Thus,
­notwithstanding the consensus view that symbolism is represented in items of
material culture by at least 80 ka, this position remains controversial in archaeo-
logical and other scientific circles.
While few archaeologists define symbolism as it applies to items of material
culture, an even smaller number study the kind of mental architecture that is
required for symbolic innovations (Coolidge & Wynn 2005). This omission creates
serious interpretive problems for, as Barnard (2010a) points out, we cannot create
behavioural theory without employing theories of mental architecture.
Here, I draw on the savoir faire of archaeologists, linguists, psychologists and
cognitive scientists to create an original way to recognise minds that have some
resemblance to our own in the deep past without relying on the traditionally
accepted (and disputed) indicators for early symbolism. This does not mean that
I do not think symbolism important; it is important, but in this paper I wish to
avoid the contention that surrounds recognition of symbolism in material culture
items found in MSA sites. I shall therefore approach symbolism indirectly by
examining unequivocal evidence for behaviour in the past that can be linked to
human cognition like our own. In borrowing theories from psychologists and
cognitive scientists, I recognise that these can also be controversial and short-lived
(Barnard 2010b), and that my interpretations based on such theories will, at best,
be superseded in the near future.

2. What is complex cognition?

The type of cognition attributed to people who think like us must minimally
involve language (in itself symbolic), abstract thought, and capacity for novel,
sustained multilevel operations. Language requires not only symbolism, but also
abstraction (the mental ability to recognise common attributes or regularity
amongst diverse objects or behaviours). Yet language also permits people to think
and talk about behavioural variation (Reuland 2010). All complex cognitive
processes have abstraction as a common denominator; abstraction would not
have been possible prior to the advent of people with thought processes much
like ours (Barnard 2010a). Furthermore, Barnard suggests that only the most
advanced mental architecture can simultaneously control multilevel operations
 Chapter 5. Complex cognition required for compound adhesive manufacture 

like talking, thinking and walking. The capacity for novel, sustained multilevel
actions could have arisen from neural connectivity in part of the prefrontal
cortex (Amati & Shallice 2007), a development that may have been favoured
in modern humans by a relatively early increase in overall brain size (Aboitiz
2001). Mithen (1996) talks about ‘cognitive fluidity’; the term is compelling
because it implies the ease with which modern humans generate and employ
innovative thoughts. Many neurons are needed to resolve intricate behavioural
challenges and the idea of a neural mutation giving rise to complex cognition
of the kind used by modern people was introduced years ago (for example,
Coolidge & Wynn 2005; Klein & Edgar 2002; Mithen 1996). Neurons need, and
possess, numerous transmitters, maybe because many nerve terminals synapse
on a single neuron (Deutch & Roth 2003). There are critical periods in animal
and human maturation processes (required for bird song or human speech) that
are greatly affected by the development of neural pathways. Neurons select their
permanent range of inputs during childhood, when the capacity for adjustment
is greater than is possible after the achievement of sexual maturity (Knudsen
2003). The opportunity for neural change tends to ‘close’ at sexual maturity,
irrespective of experience obtained, thus long childhood and late onset of sex
hormones allow for extended learning which is beneficial for the development
of complex capabilities among humans (Knudsen 2003: 570).

3. What gave rise to complex cognition?

The prefrontal cortex seems to have a fundamental role in cognitive control.

­Prefrontal cortex neurons participate in cognitive control – they can sustain their
activity to keep task-relevant information online during tasks (Miller & Wallis
2003). Neural storage in the form of ‘working memory’ can hold things in mind
(in the sense discussed earlier) at the same time as manipulating information.
Working memory capacity correlates with fluid intelligence (an ability to solve
novel problems) and enhanced working memory is an important prerequisite for
the attributes of modern cognition mentioned in the previous section: language,
creation of abstract ideas and the planning required for manufacturing tools and
other artefacts. The enhancement of working memory that occurred in the rela-
tively recent human past, most likely after the first appearance of anatomically
modern humans, may have marked the final stage in the evolution of reason and
the development of language and culture (Coolidge & Wynn 2005). Enhanced
working memory enables innovation by making thought experiments possible.
These work through heuristic action and an ability to make predictions into the
future. Relatively modern innovative technologies that provide good cases in point
 Lyn Wadley

for enhanced working memory include alloying metals, and the production of
kiln-fired ceramics (Coolidge & Wynn 2006; Wynn & Coolidge 2007a,b).
Working memory capacity can predict higher order cognitive tasks, such
as attention, language comprehension and production, reasoning, and general
and fluid intelligence (novel problem solving abilities) (Engle & Kane 2004;
Kane & Engle 2002). One of the ways in which psychologists predict cognitive
performance on higher order cognitive tasks is through reliable and valid
psychometric tests involving complex span tasks (Kane et al. 2004). Amongst
these tasks are spatial ones requiring a person, for example, to rotate a letter
mentally or decide whether a figure is symmetrical around a vertical axis (Kane
et al. 2004). Thus technology that requires mental rotation is a good indicator of
complex cognition.
The appearance of enhanced working memory, most likely the result of
a ­ relatively simple, genetically inherited neurological advance, resulted in
­cognitive abilities called ‘executive functions’ that depend on frontal-lobe-linked
abilities (Wynn & Coolidge 2003). Complex goal-directed actions, fl ­ exibility in
problem-solving, innovative solutions to problems, analogical reasoning and
planning over long distances or time are amongst executive functions (Wynn
& Coolidge 2003). The ability to switch attention is a marker of the kind of
thought process enabled by modern executive control. Switching attention
requires ­people to move their focus of attention: it involves scheduling as well
as ­inhibiting processes (Smith & Jonides 2003). One test for the ability to switch
attention involves using the Wisconsin Card Sort Task in which subjects are
asked to sort cards by colour. Then they are asked to switch attributes and sort
cards, for instance, by the designs on them. People with frontal lobe damage
cannot do this (Smith & Jonides 2003). A more complex task involves switching
between objects in working memory (such as between equations and words).
Here the subject needs to switch between updating working memory and some
other task (Smith & Jonides 2003).
Generally I should agree with Wynn and Coolidge (2007b) that technology
and subsistence are spheres of action that rely not on complex mental models,
but rather on well-learned procedural routines that can be taught during
apprenticeship. Most subsistence activities can be learned by observation and
training in a principally non-verbal way and can be classified as belonging to
the procedural type of long-term memory. Nevertheless, I suggest here that the
capacity for novel, sustained multilevel operations can be predicted from some
types of technology and I use compound adhesive manufacture as an example.
In order to demonstrate that technology can be a proxy for complex cognition,
I must make it clear that some cognitive executive steps cannot be taken without
drawing on mental abilities such as abstraction, recursion, multilevel thought
and cognitive fluidity.
 Chapter 5. Complex cognition required for compound adhesive manufacture 

4. The archaeological evidence

Pigments, including red ochre, have a long history of use in Africa, where they
may even have been used by archaic Homo sapiens (d’Errico 2008). Thus the use
of red ochre per se cannot be used as a marker of behaviour associated exclusively
with anatomically modern humans. My own study was inspired by Middle
Stone Age (MSA) stone tools called segments. Segments (sometimes known in
the literature as crescents or lunates) are defined as a portion of a circle with a
curved, abruptly-blunted back and a straight, sharp cord (Deacon 1984) (Figure
1). Each end of a segment is pointed. The abrupt blunting (which is faceted in a
manner that archaeologists call backed retouch) on the convex edge of segments
probably assists their hafting by creating an area of friction for the firm attachment
of adhesives (Lombard 2007; Nuzhnyi 2000; Phillipson 1976). Residue analysis
supports the suggestion that backed edges were designed to facilitate hafting
because plant gum/resin, sometimes mixed with ochre or other substances, was
found on the backed edges of segments and other backed tools from the Howiesons
Poort Industry with ages of between 64 ka and 61 ka (Jacobs et al. 2008) at Rose
Cottage (Gibson et al. 2004) and Sibudu (Delagnes et al. 2006; Lombard 2006,
2007, 2008) (Figure 1).
Ochre-stained segments were also found in Kenya at Enkapune Ya Muto,
dated to between 50 ka and 40 ka (Ambrose 1998). At Sibudu, carefully-mapped
positions of glue residues show that segments could have been hafted in a vari-
ety of positions (Lombard 2007, 2008). By rotating a segment so that its cutting
edge is horizontal, vertical or diagonal, it can be hafted to form one of a variety of
tools, weapons such as arrowheads or spearheads, or weapon components, such as
barbs that would be placed below the arrow tip or spearhead. Segments, because
of their half-moon shape, with the straight cutting edge along the full length of
the tool, cannot be bound with twine (the cutting edge would sever the twine)
and the hafting of these stone inserts must have depended entirely on adhesive.
Plant gum used alone is effective glue – Acacia gum, for example, is designed to
seal wounds on the tree – though it often shatters like glass on direct impact. Such
brittle glue is only advantageous when weapon tips are intended to shatter inside
prey (Clark 1975) where the sharp, broken fragments of stone would cause severe
haemorrhaging that would kill the animal and leave a trail for hunters to follow.
Adding ochre or some other aggregate to plant gum creates an adhesive that is
workable, plastic and with a good consistency that hardens to a robust product
(Wadley 2005; Wadley et al. 2009). This is the kind of glue that would be needed
to prevent the stone tip of a spear from splitting from its shaft during repeated
thrusts at prey. Some glue on ancient stone tools contains red, yellow or orange
ochre, while other specimens have no ochre at all. Instead, some glue seems to
contain plant gum mixed with fat, and perhaps sand or charcoal (it is difficult to
 Lyn Wadley

know whether charcoal and sand are post-depositionally incorporated into plant
gum residues on stone tools). Thus, it looks as if artisans in the past intentionally
selected t­ ask-appropriate recipes for their adhesives and that they were fully aware
of the properties of the required ingredients. Replications of compound glues and
their use with experimental tools demonstrate that chemical, mechanical and
possibly electrostatic changes are affected by the addition of ochre to plant gum
(Wadley et al. 2009), so it seems that ochre additions have technical applications,
whether or not they also have symbolic meaning. People may have coloured their
weapons with red ochre for symbolic purposes, for example, to signify the blood
of prey and success in the hunt, but this explanation cannot be the only one for
the mixing of ochre into glue. First, there are the technical issues already men-
tioned. Then, the coloured stains are present only on those parts of stone tools
once attached to hafts (Figure 1); MSA weapons were not entirely covered with
ochre, as has been the case in recent ethnographic examples in Australia (Wallis
& O’Connor 1998). Importantly, not all MSA adhesives contain red ochre and this
would be expected if the colourant was considered essential for hunting fortune.
Whether or not ­symbolism played a role in the placement of red ochre in the adhe-
sives, the process of making compound adhesives is so complex that it could only
have been undertaken by people whose minds had abilities that overlapped with
our own. I shall demonstrate this point now.

Ochre Backing

Cutting edge
10 mm

Figure 1. Segment from Sibudu Cave showing red ochre traces that were part of the
­compound ­adhesive used to attach it to a shaft or handle. The segment comes from a
layer with an age of about 65 ka. The line drawing of another segment from Sibudu shows
the relative positions of the backed and cutting edges
 Chapter 5. Complex cognition required for compound adhesive manufacture 

5. Replicated compound adhesive manufacture: Methods

To explore the efficacy of ochre-loaded glue, and the complex thought processes
required by its manufacture, replicated tools were hafted with compound adhesive
made from red ochre mixed with Acacia karroo gum (see Figure 2 for an example
of one such tool). Only products and methods available in the MSA were used,
except where synthetic iron oxide was attempted for comparison with the geologi-
cal products. Yellow and red ochres were collected from a Snuffbox Shale quarry in
the vicinity of Sibudu Cave, but one sample of red ochre came from the ­Waterberg,
Limpopo. Ochre was powdered by rubbing nodules on coarse, flat, sandstone slabs
(Sibudu roof-spall) (Figure 3) because such slabs with ochre stains have been found
at Sibudu. Three main adhesive recipes were used: (1) Acacia gum alone, (2) ­Acacia
gum mixed with either natural ochre powder or synthetic haematite powder and,
(3) Acacia gum mixed with natural ochre powder and melted beeswax. All replica-
tions were conducted with open wood fires whose temperatures varied depending
on the type and amount of wood used, but did not exceed 726°C. Detailed method-
ology is recorded elsewhere (Wadley 2005; Wadley et al. 2009).

1 cm

Figure 2. Replicated compound adhesive using Acacia gum and red ochre. The glue was used
to attach a replicated stone flake to a wooden handle. The tool was subsequently successfully
used to chop bark from branches

The partially dehydrated adhesive was used to mount the stone insert to its
wooden haft (for example, Figure 2) and the hafted tool was placed near the fire
(60oC–100oC) for four hours to dry and harden the adhesive. Each tool was rotated
regularly to prevent adhesive dripping from the shaft, and each was moved closer
 Lyn Wadley

or farther from the fire depending on its heat. The entire process was i­ntricate
and demanded full attention, requiring the manufacturer to hold many things
in mind simultaneously. No set recipe or procedure could be used for any single
tool or compound adhesive because of the variability amongst ­ingredients such
as the plant gum. Natural products and wood fires have unpredictable v­ ariables.
Artisans in the past were obliged to deal with these just as I was, and they had
to compensate for the difficulties presented by their ingredients. For example,
Acacia gums from different trees or different seasons have changeable viscosity,
thereby requiring varying amounts of loading agent. Acacia gum that I worked
with ­fluctuated between being highly viscous, and dry and crystalline, and its pH
varied between 3.0 and 4.4. Thus, the chemical reaction between the gum and
ochre differs between adhesive mixtures with different attributes. Ochre pieces
can also be dissimilar in respect of their quantities of iron and other minerals and
also in respect of particle size. Some ochre pieces grind into coarser powders than
others, depending on their silica contents and also on the composition of the stone
on which they were ground.
Ancient glues were almost certainly heated near a fire because, without
dehydrating the replicated glues, they take six days or more to dry. Experimentally
produced tools separated from their shafts when the glue was not properly dry.
Camp fire temperatures depend on the wood type used, the amount of fuel,
ambient humidity and wind speed, so artisans would have needed to control fire
temperature to avoid spoiling the adhesives by burning them or causing them to
boil. Thus, compound adhesives of the kind described here could not have been
made using fixed recipes as we do for making cakes. On-going assessments and
on-the-spot adaptations, requiring cognitive fluidity, must have been made during
adhesive production. It is apparent that people in the MSA were sufficiently
skilled to realise early in the amalgamation of ingredients whether adhesive had
the right consistency and workability. If necessary, workability could be improved
by adding more aggregate such as ochre, sand and/or small amounts of fat or
beeswax. The early artisans must have made choices and measurements based
on the ability of certain combinations of ingredients to create reliable adhesives.
Successful compound adhesive manufacture using ochre and plant gum seems to
have been achieved when there was (1) a chemical change from acid to less acid
pH, (2) a transformation in mechanical properties to one of workability, and (3)
slow dehydration of wet adhesives near open fires.

6. Compound adhesive manufacture as a proxy for modern cognition

Creating compound adhesive from disparate ingredients may have been regarded as
symbolic in the past, but this is a difficult interpretation to make. From the ­evidence
 Chapter 5. Complex cognition required for compound adhesive manufacture 

presented here, it is easier to claim that people in the MSA were conversant with
some aspects of chemistry and pyrotechnology. While they are unlikely to have
known pH as we understand it, or have known how to gauge the percentages
of iron or silica in rocks, they did seem to understand the importance of
complexing iron minerals or colloids with acid plant gum and dehydrating them
near controlled, open fires. Furthermore, some of the steps required for making
compound adhesive seem impossible without the ability to abstract relevant
features of the products they worked with. Qualities of gum, such as wet, sticky
and viscous, must have been mentally abstracted and these meanings weighed
against ochre powder properties such as dry, loose and dehydrating. It is not
easy to imagine how an expert glue-maker could train an apprentice to make
compound adhesives without language that could explain, in abstract terms,
attributes and conditions such as stickiness, viscosity, workability, consistence,
plasticity, texture, particle size, temperature, concretization, water-solubility,
hydroscopic, dehydration, reversible process, irreversible process, shrinkage,
homogeneity, creep and shrinkage. These concepts probably had to be explained
using language as we appreciate it, in other words incorporating recursion,
abstraction and concepts about the past and the future.
The glue makers needed to keep in mind work already done in order to carry
out those tasks still needed for the successful completion of the glue and the
mounting of the composite parts of weapons or tools. In addition, the artisans
had to be equipped to alter ingredients and adjust fire temperatures on-the-spot;
concurrently they had to think about the appropriate angles for placing stone
inserts on their hafts. They were able to perform mental rotation of segments or
other tools in order to create weapons or other implements in an assortment of
designs for different objectives. They also needed to think abstractly about the
qualities of their segments in order to envisage their use after they were rotated
at different angles. As mentioned earlier, successful mental rotation requires
advanced working memory capacity (Kane et al. 2004) and, in turn, complex
cognition.
Some birds (such as swallows that build mud nests) and wasps also create
forms of compound adhesives, but they do so instinctively with simply coded
operational sequences, in which the distance between problem and solution is
far smaller than that demonstrated by the human action of making a composite
hunting weapon (Haidle 2010). One obvious difference in human manufacture
of compound glue is the use of pyrotechnology. Temperature control depends on
understanding wood types, their moisture or resin contents and their propensity
to form long-lasting coals. Vigilance is essential because my adhesives burned, or
boiled to form air bubbles, when they were too close to the fire. Over-dehydration
caused loss of cohesiveness and a crumbly product, while adhesive that boiled lost
homogeneity and consequently became unstable.
 Lyn Wadley

Figure 3. An ochre nodule used to create red ochre powder by grinding it on a sandstone slab

The glue-maker needs to pay careful attention to the condition of ingredients

before and during the procedure and must be able to switch attention between
aspects of the methodology. To hold many courses of action in mind involves
multi-tasking, which is one trait of modern human minds (Barnard et al. 2007) as
well as mental flexibility. Capacity for multilevel operations, abstract thought and
mental rotation are all required for the process of compound adhesive manufacture
and hafting of tools. Although fully modern behaviour seems recognizable only
relatively late in the MSA (Klein 2008), the circumstantial evidence provided here
from experimental work with compound adhesives suggests that people who made
them in the MSA shared some advanced behaviours with their modern successors.
While compound adhesives have presently only been recorded on stone tools that
are about 70 000 years old, this date is likely to be pushed farther into the past
when residue analysis is conducted on older tools.
In conclusion, I contend that compound adhesive manufacture in the MSA
required complex cognition of the kind that intersected with our own. People
who make compound adhesives from natural products such as plant gum and
geologically variable iron oxides must be able to multi-task because they must
simultaneously think, mix glue (changing quantities at short notice) and maintain
fire temperature. Since the intention is always to use the glue to create composite
tools, the artisan must also be able mentally to rotate stone tools and visualise
 Chapter 5. Complex cognition required for compound adhesive manufacture 

the finished tool or weapon. Some of the stages involved in making compound
adhesives, and their use with composite tools, entail not only multi-tasking, but
also mental abstraction, switching attention, recursive behaviour and cognitive
fluidity. Accordingly, it seems justifiable to attribute advanced mental abilities to
people who lived in the African MSA. Although symbolism has not been ­studied
directly in this paper, complex cognition of the kind described here is most likely
to have co-occurred with a capacity for symbolic thought and behaviour. So many
variables and unpredictable actions are involved in the production of compound
adhesive and in the hafting of segments that the process seems to require
­language for instructing apprentices. I suggest that my circuitous route for finding
­symbolism in the archaeological record specifically that of the MSA of Africa is
theoretically satisfying. Furthermore, the approach avoids making a leap of faith
between ancient items of material culture and their meaning.

References

Aboitiz, F. 2001. “Size and complexity of the brain in human evolution.” In Humanity from
African naissance to coming millenni, P. V. Tobias, M.A. Raath, J. Moggi-Cecchi &
­
G.A. Doyle, (eds), 355–359. Firenze: Firenze University Press.
Amati, D. & Shallice, T. 2007. “On the emergence of modern humans.” Cognition 103:
358–385.
Ambrose, S.H. 1998. “Chronology of the Later Stone Age and food production in East Africa.”
Journal of Archaeological Science 25: 377–392.
Barnard, P.J. 2010a. “From executive mechanisms underlying perception and action to the
­parallel processing of memory.” Current Anthropology 51 (1): S39–54.
Barnard, P.J. 2010b. “Current developments in inferring cognitive capabilities from the
­archaeological traces left by stone tools: caught between a rock and a hard inference.” In
Stone tools and the evolution of human cognition, A. Nowell and I. Davidson (eds), 207–226.
Boulder, CO: Colorado University Press.
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.
Bouzouggar, A., Barton, N., Vanhaeren, M., d’Errico, F., Collcutt, S., Higham, T., Hodge,
E., Parfitt, S., Rhodes, E., Schwenninger, J.-L., Stringer, C., Turner, E., Ward, S., Moutmir, A.
& Stambouli, A. 2007. 82,000-year-old shell beads from North Africa and ­implications for
the origins of modern human behaviour. Proceedings of the National Academy of ­Science,
USA 104: 9964–9969.
Clark, J.D. 1975. “Interpretations of prehistoric technology from ancient Egyptian and other
sources. Part II: Prehistoric arrow forms in Africa as shown by surviving examples of the
traditional arrows of the San Bushmen.” Paléorient 3: 127–150.
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. 2006. “Recursion, phonological storage capacity, and the evolution
of modern speech.” A paper presented to the Cradle of Language Conference, November
2006, Stellenbosch, South Africa.
 Lyn Wadley

Deacon, J. 1984. The Later Stone Age of southernmost Africa. Cambridge Monographs in African
Archaeology 12. Oxford: British Archaeological Reports.
Delagnes, A., Wadley, L., Villa, P. & Lombard, M. 2006. “Crystal quartz backed tools from the
Howiesons Poort at Sibudu Cave.” Southern African Humanities 18: 43–56.
d’Errico, F. 2008. “Le rouge et le noir: implications of early pigment use in Africa, the Near
East and Europe for the origin of cultural modernity.” South African Archaeological Society
Goodwin Series 10: 168–174.
d’Errico, F., Henshilwood, C., Vanhaeren, M. & van Niekerk, K., 2005. “Nassarius kraussianus
shell beads from Blombos Cave: evidence for symbolic behaviour in the Middle Stone Age.”
Journal of Human Evolution 48: 3–24.
d’Errico, F., Vanhaeren, M. & Wadley, L. 2008. “Possible shell beads from the Middle Stone Age
layers of Sibudu Cave, South Africa.” Journal of Archaeological Science 35: 2675–2685.
Deutch, A.Y. & Roth, R.H. 2003. “Neurotransmitters.” In Fundamental Neuroscience, S­ econd
Edition, L.R. Squire, F.E. Bloom, S.K. McConnell, J.L. Roberts, N.C. Spitzer and M.J.
­Zigmond, (eds), 163–196. San Diego: Academic Press.
Engle, R.W. & Kane, M.J. 2004. “Executive attention, working memory capacity, and a ­two-factor
theory of cognitive control.” In The psychology of learning and motivation, B. Ross (ed),
145–199. New York: Academic Press.
Gibson, N.E., Wadley, L. & Williamson, B.S. 2004. “Microscopic residues as evidence of hafting
on backed tools from the 60 000 to 68 000 year-old Howiesons Poort layers of Rose Cottage
Cave, South Africa.” Southern African Humanities 16: 1–11.
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) 1: S149–166.
Henshilwood C. & Marean C. 2003. “The origin of modern human behavior: critique of the
models and their test implications.” Current Anthropology 44: 627–651.
Jacobs, Z., Roberts, R.G., Galbraith, R.F., Deacon, H.J., Grün, R., Mackay, A.M., Mitchell,
P., Vogelsang, R. & Wadley, L. 2008. “Ages for the Middle Stone Age of Southern Africa:
Implications for human behavior and dispersal.” Science 322: 733–735.
Kane, M.J. & Engle, R.W. 2002. “The role of prefrontal cortex in working memory capacity,
executive attention, and general fluid intelligence: An individual-differences perspective.”
Psychonomic Bulletin and Review 9: 637–671.
Kane, M.J., Hambrick, D.Z., Tuholski, S.W., Wilhelm, O., Payne, T.W. & Engle, R.W. 2004.
“The generality of working memory capacity: A latent-variable approach to verbal and
­visuospatial memory span and reasoning.” Journal of Experimental Psychology: General
133: 189–217.
Klein, R.G. & Edgar, B. 2002. The dawn of human culture. New York: Wiley and Sons.
Klein, R.G. 2008. “Out of Africa and the evolution of human behavior.” Evolutionary ­Anthropology
17: 267–281.
Knudsen, E.I. 2003. “Early experience and critical periods.” In Fundamental Neuroscience,
­Second Edition. L. R. Squire, F.E. Bloom, S.K. McConnell, J.L. Roberts, N.C. Spitzer and M.
J. Zigmond, (eds), 555–573. San Diego: Academic Press.
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: 57–67.
Lombard, M. 2007. “The gripping nature of ochre: the association of ochre with Howiesons
Poort adhesives and Later Stone Age mastics from South Africa.” Journal of Human
­Evolution 53: 406–19.
 Chapter 5. Complex cognition required for compound adhesive manufacture 

Lombard, M. 2008. “Finding resolution for the Howiesons Poort through the microscope:
­micro-residue analysis of segments from Sibudu Cave, South Africa.” Journal of Archaeo-
logical Science 35: 26–41.
Miller, E.K. & Wallis, J.D. 2003. “The prefrontal cortex and executive brain functions.” In
­Fundamental Neuroscience, Second Edition. L.R. Squire, F.E. Bloom, S.K. McConnell,
J.L. Roberts, N.C. Spitzer & M.J. Zigmond, (eds),1353–1376. San Diego: Academic Press.
Mithen, S. 1996. The prehistory of the mind. London: Thames and Hudson.
Nuzhnyi, D.Y. 2000. “Development of microlithic projectile weapons in the Stone Age.” Anthro­
pologie et Prehistoire 111: 95–101.
Phillipson, D.W. 1976. The prehistory of Eastern Zambia. Nairobi: Memoir 6 of the British
­Institute in Eastern Africa.
Reuland, E. 2010. “Imagination, planning and working memory: the emergence of language.”
Current Anthropology (51) 1: 99–110.
Smith, E.E. & Jonides, J. 2003. “Executive control and thought.” In Fundamental Neuroscience,
Second Edition. L.R. Squire, F.E. Bloom, S.K. McConnell, J.L. Roberts, N.C. Spitzer &
M.J. Zigmond, (eds), 1377–1394. San Diego: Academic Press.
Wadley, L. 2001. “What is cultural modernity? A general view and a South African perspective
from Rose Cottage Cave.” Cambridge Archaeological Journal 11: 201–221.
Wadley, L. 2005. “Putting ochre to the test: replication studies of adhesives that may have been
used for hafting tools in the Middle Stone Age.” Journal of Human Evolution 49: 587–601.
Wadley, L., Hodgskiss, T. & Grant, M. 2009. “Implications for complex cognition from the h ­ afting
of tools with compound adhesives in the Middle Stone Age, South Africa.” P ­ roceedings of
the National Academy of Science, USA 106: 9590–9594.
Wallis, L. & O’ Connor, S. 1998. “Residues on a sample of stone points from the west Kimberley”
In A closer look: recent Australian studies of stone tools, R. Fullagar (ed), 1­ 49–178. Sydney:
Sydney University Press.
Wynn, T. and Coolidge, F.L. 2003. “The role of working memory in the evolution of managed
foraging.” Before Farming 2003/2 (1): 1–16.
Wynn, T. & Coolidge, F.L. 2007a. “Executive functions, working memory, and the evolution of
the modern mind.” Paper presented to AAAS, March, 2007.
Wynn, T. & Coolidge, F.L. 2007b. “A Stone Age meeting of minds.” American Scientist 96: 44–51.
Wynn, T., Coolidge, F.L. & Bright, M. 2009. “Hohlenstein-Stadel and the evolution of human
conceptual thought.” Cambridge Archaeological Journal 19: 73–83.
 chapter 6

The emergence of language, art

and symbolic thinking
A Neandertal test of competing hypotheses
João Zilhão
ICREA Research Professor SERP – Department of Prehistory, Ancient
History and Archaeology,
University of Barcelona/Institució Catalana de Recerca i Estudis Avançats (ICREA)

There is a widespread understanding that the personal ornaments of the African

Middle Stone Age and the animal and human figurines of the Aurignacian of
southern Germany provide the earliest evidence of the possession of “modern”
cognitive capabilities, ones that appeared for the first time in human evolution
as a result of the speciation of Homo sapiens and that would explain its rapid
expansion from Africa into Eurasia and the attendant extinction of coeval archaic
humans (such as the Neandertals). The archaeological facts contradict this view,
since there is abundant evidence for the existence of such “modern” capabilities in
non-sapiens populations, and that language, “symbolic thinking” by definition, is
probably as old as the human genus. Therefore, the explanation for the emergence
of body ornamentation and figurative art must be sought not in the realm of
cognition but in that of history, with demographic growth and the intensification
of social interaction networks playing a primary role in the process.

1. Introduction

The last fifty years of scientific research established beyond reasonable doubt that
the earliest human ancestors appeared in Africa some time around two million
years ago. Soon after, these Homo erectus people expanded into Eurasia. By one
and a half million years ago, they had already reached the Indonesian island of
Java, but it would take a bit longer for Europe to be stably settled (Dennell &
Roebroeks 2005).
The earliest evidence comes from Iberia, where the so-called Homo antecessor
fossils from Atapuerca date to about one million years ago (Bermúdez de Castro
et al. 2004; Carbonell et al. 2008). Coeval African fossils are scarce, but, altogether,
the evidence suggests that a trend towards increased brain size and correlated
changes in the shape of the skull was under way at this time throughout the entire
 João Zilhão

Old World (McHenry & Coffing 2000; Lee & Wolpoff 2003). Geneticists have
related these changes to a second Out-of-Africa expansion, represented, archaeo-
logically, by the spread of the Acheulian technocomplex, whose iconic stone tool
is the handaxe or biface (Templeton 2002, 2005).
Subsequent geographic isolation led to the differentiation of these Acheu-
lian populations into two lineages. In Europe and western Asia, Homo erectus
became Neandertal man sometime around 500 ka. At the same time, in Africa,
Homo ­erectus became Homo sapiens (or “modern humans”), and, some 50 ka, in
the framework of a third Out-of-Africa event, spread into Eurasia, Australia, and,
eventually, the Americas (Trinkaus 2005).
Over the last quarter of a century, it has become clear that early African sapi-
ens are ancestral to all present-day living humans but no agreement exists where
Eurasian Neandertals are concerned. The level of their separation in taxonomy, the
extent of their differences in biology, behaviour and culture, and their ultimate fate,
remain to this day among the hottest topics in human evolution studies. That Nean-
dertals are no more is uncontroversial, but when, why and how were they replaced?
These questions have fundamental implications for the understanding of the
emergence of art, language and symbolic thinking in the human lineage. The
long lasting geographical segregation of the two palaeontological taxa, Homo
­neanderalensis and Homo sapiens, and the ultimate replacement of the former
by the l­ atter are widely assumed to imply that they were truly d­ ifferent ­biological
species. And, as textbook definitions require species to differ in behaviour as
much as in morphology, the corollary expectation is that significant behav-
ioural differences, with attendant cognitive implications, must have separated
“anatomically modern” people from coeval “archaic” humans, namely the
­
­Neandertals (Henshilwood & Marean 2003).
The notion that such a separation existed at biospecies level dovetails with
speculations that certain features of complex human culture, which are undoc-
umented in the archaeological record of Homo erectus and other early humans
such as art, or ritual burial, must have emerged as a by-product of the processes
involved in the speciation of the African sapiens. Under this “Human Revolution”
hypothesis (Mellars & Stringer 1989; Mellars et al. 2007), the absence of those
features reflects the lack of the required cognitive capabilities. In this view, it is
only after the acquisition of such capabilities by the first “modern humans” that
the corresponding behavioural correlates could be and indeed were externalized
in archaeologically visible ways.
Material evidence consistent with this paradigm – marine shell beads, inter-
preted as personal ornaments and, in some instances, associated with s­keletal
remains of early modern humans – has been obtained over the last decade at
archaeological sites dated to between 75 ka and 100 ka in Palestine, Morocco
 Chapter 6. The emergence of language, art and symbolic thinking 

and South Africa (d’Errico et al. 2005; Vanhaeren et al. 2006; Bouzouggar et al.
2007) (Figure 1). In the ethnographic present, personal ornaments play the role
of ­conveyors of the social identity of persons – group membership, gender, and
individual life-history characteristics (age, marital status, etc.). Working with such
symbolic systems of personal presentation and re-presentation implies language
and requires cognitive capabilities unknown among our closest living relatives, the
chimpanzees. Although we have to bear in mind that, prior to the invention of writ-
ing systems, the evidence for language can only be indirect, all of this still is rather
uncontroversial. But was the emergence of such capabilities in the human lineage as
recent a phenomenon as postulated by the “Human Revolution”?

Figure 1. Middle Stone Age Nassarius kraussianus shell beads from Blombos Cave, South
Africa (size range: 7–10.5 mm) (photo courtesy of C. Henshilwood and F. d’Errico)

A major empirical hurdle faced by this paradigm is that the putative speciation
event leading to Homo sapiens occurred between 150 ka and 200 ka (Lahr & Foley
1998), which begs an obvious question: If symbolic thinking and modern cognition
are a simple by-product of the biological processes involved in that speciation
event, why did it take at least 50 000 years for manifestations of those capabilities
(such as the African shells beads) to appear in the ­archaeological record? And why
was it that yet another 50 000 years were necessary for the emergence of figurative
art, the earliest examples of which are the cave paintings of Chauvet, in France
(Clottes et al. 1995), and the animalistic, anthropomorphic and therianthropic
ivory figurines from different caves of the Swabian Jura, in ­Germany, dated to
about 35–37 ka (Conard 2009) (Figure 2)?
 João Zilhão

From the ethnographic record, we also know that the visual display of
objects conveying information on the personal and social identity of the indi-
viduals carrying such objects is targeted at encounters with strangers or people
infrequently met. There are two rather good reasons for this: firstly, without
some prior experience of interaction, the meaning of the visual symbols would
be opaque to the viewer; secondly, identifying one’s affiliation or identity to
­family and immediate acquaintances does not require material symbols (Kuhn
et al. 2001). The possibility exists, therefore, that the appearance of ornaments
in the archaeological record reflects the crossing of demographical thresholds,
above which long-distance interaction networks involving alliance, exchange or
­mating were necessary.

Figure 2. Sculpted ivory figurines from the Late Aurignacian of southern Germany, dated to
~35–37 ka. Left: “Venus” pendant from Hohle Fels. Right: horse (top) and lion (bottom) from
Vogelherd (photos by H. Jensen, courtesy of N. Conard, University of Tübingen)

If so, then the absence of evidence for “modern” cognition prior to about 100 ka
would not be evidence for its absence among anatomically “modern” humans prior
to that time; it would simply mean that, in those days, the social life of humans had
not yet effected the “release from proximity” (Gamble 1998) that eventually gener-
ated the need to have symbolic identities and ways of displaying it. But once we
admit that the emergence of the earliest material e­ vidence of ­“modern” cognition
can relate to social, not biological processes, we have no choice but to ask ourselves
whether the same does not apply as well to ­earlier humans, namely Homo erectus.
Given that brains do not fossilise and that the evidence for language in Palaeolithic
times is indirect, could it be that, ­cognitively, these earlier ancestors were also fully
human, i.e. gifted with such behavioural features as language and all its palaeon-
tologically invisible neurological correlates? Put another way, could it be that, as
argued by Deacon (1997), language and ­symbolic thinking appeared in the earliest,
 Chapter 6. The emergence of language, art and symbolic thinking 

not the latest stages of the evolution of humans, but did not ­externalize in ways
amenable to preservation in the material record of the prehistoric past until much
more recently?
Given the genetic and palaeontological evidence that the European lineage
leading to Neandertals had already branched off of the African stem by half a
­million years ago, the archaeology of the Neandertals provides the ideal test-
ing ground for the different views of the emergence of language and “modern”
­cognition. If the “Human Revolution” is right, then neither personal ornaments
nor art should be found among the Neandertals. If either is found, then the
“Human Revolution” is refuted and we must look for alternative ways to explain
the emergence of those behaviours in the archaeological record.

2. Neandertal-ness

Neandertals are named after a skeleton found in 1856 at the Kleine Feldhoffer
cave, in the Neander valley, near Düsseldorf. Today’s scientists, however, are not
the only group of people for whom “Neandertal” has a well-defined meaning.
The word is also used in common language to disqualify dislikeable individuals,
including political opponents. Opening any dictionary immediately brings up
these alternative meanings. The Cambridge On-line, for instance, gives the follow-
ing: (1) “relating to a type of primitive people who lived in Europe and Asia from
about 150,000 to 30,000 years ago” (2) said “of people or beliefs very old-fashioned
and not willing to change” (3) said “(of people) rude or offensive.”
In order to understand the Neandertals’ enduring bad reputation, we have
to bear in mind that, in the mid-nineteenth century, Evolution was conceived in
the framework of a progressivist mindset – the directional development of ever
more complex and sophisticated forms of life from a simple, primitive common
ancestor, with humans sitting at the top of the ladder. Evolution also implied,
as Darwin eventually made explicit, that humankind had ape-like ancestors.
In this context, two things were, in retrospect, entirely predictable: firstly, a
predisposition to interpret any intermediate fossil forms as “part-ape/part-man”
in both morphology and behaviour; secondly, because things are what they are
only trough their opposition with what they are not, a predisposition to imagine
the “animality” of those ancestors as consisting of features representing the exact
opposite of “humanity” as Victorians perceived it.
To make things worse, progressivist preconceptions were compounded by
scientific error (Trinkaus & Shipman 1993). One of the first Neandertal articulated
skeletons to be found, 100 years ago now, was that from the cave site of La Chapelle-
aux-Saints. The famous French physical anthropologist Marcelin Boule made a
 João Zilhão

classical and in many respects paradigmatic description of this fossil. Unfortunately,

he also mistook for the normal Neandertal condition numerous and major
pathological, arthritic malformations developed late in the ontogeny of the elderly
subject of his study. As a result, both popular and scientific opinion converged in
considering Neandertals as a side branch, a dead-end of Evolution, both distinct
from and inferior to true humans. As late as 1953, Neandertals were still portrayed
as the archetypal “half-man/half-beast” of a famous Hollywood film (Figure 3).

Figure 3. The original advertisement for the 1953 movie “The Neanderthal Man” as
­reproduced on the cover of a recent DVD edition

In the 1960s, this prevailing view was challenged. Boule’s error was exposed,
and greater emphasis was placed on the significance of skeletons found in ­Palestine
in the 1930s. These fossils, recovered at two nearby cave sites in Mt. Carmel,
­seemingly displayed an intermediate anatomy, prompting suggestions that the
Near East had functioned as a zone of admixture between European Neandertals
and early A­ frican sapiens (McCown & Keith 1939). Moreover, there was g­ rowing
­recognition at this time that, archaeologically, the two groups had been doing
 Chapter 6. The emergence of language, art and symbolic thinking 

pretty much the same thing throughout the period between 100 ka and 50 ka. Their
stone tools were indistinguishable, and they had buried their dead, a ­practice that
implies world views and religious beliefs (Leakey & Lewin 1977). In sum, the two
lineages behaved in ways whose level of complexity required the use of language
and ­symbols, as should be expected from cranial capacity – Neandertal brains were
in fact larger than ours.
These developments led many scholars to wonder whether Neandertals,
instead of an unrelated side-branch, could have been a regional variant of a single
evolving human species and, as such, the direct ancestors of today’s Europeans. In
this view, called the Multiregional Hypothesis (Wolpoff 2002; Thorne & Wolpoff
2003), present racial diversity would be the outcome of a deep-rooted continuity
between today’s populations and those of the remotest past. There would have
been one and only one Out-of-Africa event, modern Asians and Europeans would
be the descendants, through a series of convergent changes in morphology, of the
first Homo erectus settlers, and features such as the big noses of Europeans would
be an example of the persistence of Neandertal “blood” in living humans.
The 1980s saw the birth of an entirely new line of inquiry, human genetics,
which eventually questioned this 1960s view of the Neandertals as fundamentally
human. The study of variation in the mitochondrial DNA of extant people led
to the conclusion that we are all very closely related, implying a very recent last
common ancestor, one who would have lived in Africa some 150 ka (Cann et al.
1987). Because mtDNA is inherited through the maternal line, this genetic ances-
tor was called Eve. In the Eve scenario, her children would subsequently replace
the Neandertals and other coeval archaic forms of Eurasian humans, which would
all become extinct without descent. This view was supported by genetic inferences
derived from the fossil mtDNA successfully extracted from the original Neander-
tal specimen, in 1997, followed, a decade later, by preliminary sequencing results
for the entire nuclear genome of another Neandertal individual from the Croatian
cave site of Vindija (Krings et al. 1997; Green et al. 2006; Noonan et al. 2006).
The weight of scientific opinion (e.g. Klein 2003) saw in these results support
for the notion that the Neandertals were phylogenetically distant and belonged
in an altogether different species. Unaffected by the “Human Revolution”, they
must have lacked language, or have only had an exceedingly primitive version of
it. Moreover, no division of labour and no form of social organization beyond that
required by the group’s need to reproduce would have existed, and the so-called
Neandertal burials probably were nothing more than simple body disposal w ­ ithout
religious meaning. In these circumstances, the outcome of contact situations could
only have been total replacement with no admixture. “Humans” would have seen
“Neandertals” as unsuitable non-human mates, and the cognitive superiority of
our ancestors meant that they would inevitably have prevailed in the competition
for territory and resources.
 João Zilhão

3. Paradigm lost

The empirical evidence generated by the last decade of research has falsified the
behavioural tenets of the “Human Revolution”. Ironically, at the same time as
archaeologists working in Africa were uncovering evidence supporting “­ modern”
cognition tens of millennia prior to the Out-of-Africa of Eve’s children (and
conceivably explaining it) (McBrearty & Brooks 2000), archaeologists ­working
in Europe were also uncovering evidence that, contrary to the postulates of
the “Human Revolution”, complex and sophisticated cognitive and intellectual
­capabilities were also apparent in the material culture of Eurasian Neandertals
(d’Errico et al. 1998, 2003a; Zilhão and d’Errico 1999; Zilhão 2001, 2006a, 2006b,
2007; d’Errico 2003; Conard 2008).
For instance, excavations carried out between 1994 and 1997 in a ­German
brown coal mine near Schöningen yielded three 400,000-year-old wooden
­artefacts of great significance (Thieme 1997). Long and pointed, they were made
from the base of individual spruce trees, with the maximum thickness and weight
at the front and long tails that taper towards the proximal end. In all these respects,
they resemble the javelins of field-and-track competitions, suggesting use as
­projectile weapons rather than thrusting spears. They further imply that the laws
of ­ballistics underlying the shape of modern javelins had already been mastered by
the ­founding fathers of the Neandertal lineage.
Further evidence for sophisticated craftsmanship comes from Neumark-
Nord, another German brown coal mining site. Chemical analysis of organic
­residue adhering to a flint flake recovered in levels dated to more than 100 ka
showed it to be an extract of oak bark macerated in water, of a kind used
until the ethnographic present in the tanning of hides for the manufacture of
­water-proof clothing and shoe wear (Meller 2003). In the 1930s, a nearby site,
the Ilsenhöhle rock shelter, had already yielded a few bone awls from the time
of the latest Neandertals, around 40 ka (Hülle 1977). Combined, this evidence
­suggested a long tradition of hide working for the manufacture of clothes and
other equipment.
This should come as no surprise. Good-quality artificial insulation was
a pre-requisite for survival in Ice Age central Europe, where, considering the
wind-chill effect, average winter temperatures ranged between –20 and –30ºC.
Thermoregulatory models (Aiello & Wheeler 2003) show that the minimum
external temperature Neandertals would have been able to support if dressed in a
modern business suit was –24ºC. In the absence of even such basic level of clothing,
only a thickness of body fat below the skin in excess of 3 cm could have provided
equivalent protection. The weight of such fat, however, would be of some 50 kg, an
amount that would leave an 80 kg Neandertal very little left for muscle, bone and
 Chapter 6. The emergence of language, art and symbolic thinking 

other tissue; or, if added to the 80 kg of a lean, muscular body, would transform the
average 1.65 m tall male Neandertal into the archetypal obese, unable to procure
his own subsistence in a society that lacked cash, automobiles, and shopping malls.
The implication is clear: like present-day subarctic peoples, Neandertals must have
had good quality clothing as well as all the other gear without which survival in
such environments is impossible.
Further and more telling evidence that Neandertals were quite good at
Chemistry came from yet another German brown coal mining site, Königsaue.
In ­1963–64, salvage excavations yielded two fragments of birch bark pitch used
for stone tool hafting, one of which still bore a fingerprint of the Neandertal that
manipulated it. These items have since been directly dated by radiocarbon to more
than 45 ka, and a study of their composition (Koller et al. 2001) showed that they
are not unmodified natural products, such as the bitumen used in the Near East
for the same hafting purposes since at least 200 ka. In fact, they are a synthetic
raw material, the first ever in human history. They were produced through a
several hour-long smouldering process requiring a strict manufacture protocol:
under exclusion of oxygen, and at tightly controlled temperatures, between 340
and 400°C.
This evidence suggests that, in fact, Neandertals were cognitively as well
endowed as we are. But what about their biology? Were they in fact a separate
species? And was the reason for their eventual extinction somehow related to their
biological separateness?
Careful consideration of the mtDNA evidence shows that such notions
have little basis. Given that contemporary populations of chimpanzees are
more diverse than all living humans and Neandertals put together (Gagneux
et al. 1999), the parsimonious interpretation of the genetic evidence is that, by
Primate standards, present-day humans are abnormally homogeneous, not that
Neandertals were a different species. In fact, the most recent synthesis of the
history of modern human dispersals based on mtDNA places the immigration
of the bearers of the oldest extant European variants (haplotypes H, I and U) at
some 30 ka (Forster 2004). Since anatomically modern people are documented
in Europe since at least 40 ka (Zilhão et al. 2007), it follows that the mtDNA
variants characteristic of those original settlers must be as extinct as those of the
Neandertals. Obviously, this does not mean that such early European moderns
belonged to a species different from our own, and one that went extinct without
descent. By the same token, no such inferences can be made for the Neandertals
on the basis of the absence of their mtDNA among living humans. The take-
home message is that the mtDNA of present-day Europeans reflects recent
demographic history, not the remote interactions (or lack thereof) between
African sapiens and ­non-African archaics.
 João Zilhão

In short: The pattern of mtDNA homogeneity among extant humans is

consistent with a recent origin for modern humans, but it does not rule out the
possibility that Neandertals and other archaic groups contributed to the gene pool
of subsequent populations. That such contributions indeed occurred is otherwise
apparent in the nuclear genome, where as much as 80% of its loci carry evidence
of the assimilation of genetic material from non-African archaic people (Eswaran
et al. 2005). One example is the microcephalin gene, involved in the control of
brain size during development, whose adaptive allele, which occurs in 70% of
today’s humans, seems to have introgressed from an archaic lineage, most probably
the Neandertals, sometime around 37 ka (Evans et al. 2006). Most recently, the
results of the Neandertal genome project also brought additional support to these
findings by showing that Neandertals contributed something like 1–4% of the
nuclear DNA of present-day humans, implying significant interbreeding at the
time of contact (Green et al. 2010).
These developments contradict the notion that Neandertals were a different
species and show that, even if they had been, they were so close that admixture
at the time of contact was inevitable, and did happen. In fact, a recent study
of ­species intersterility versus time of divergence (Holliday 2006) suggests that
the whole debate concerning the taxonomic status of Neandertals is a lot like
the proverbial Byzantine argument about the sex of the angels. The study shows
that, among the many lineages of mammals for which fossil or molecular data
are available, 1.4 ­million years is the minimum amount of time for reproductive
separation to emerge between two branches splitting from a recent ­common
ancestor. This minimum was observed among gazelles. Among hominins,
­however, the interval between generations is at least four times longer. The
implication is one of no reproductive isolation between contemporary lineages
of hominins separated by less than five to six million years of divergence. Such
a length of time corresponds to the entire evolutionary life span of the hominin
family, and is at least ten times the duration of the interval separating the
Neandertal/­modern splitting event from the period of contact in Europe. By
Mammal standards, therefore, Neandertals were not, and could not have been, a
different biological species.

4. Paradigm found

Until about ten years ago, the presence of ornaments at late Neandertal sites was
acknowledged by supporters of the “Human Revolution” but disregarded as a
by-product of “imitation without understanding” of modern human behaviours
observed in contact situations (Stringer & Gamble 1993). The following analogy
 Chapter 6. The emergence of language, art and symbolic thinking 

was famously proposed by a distinguished British archaeologist to make the point:

“if a child puts on a string of pearls, she is probably doing this to imitate her mother,
not to symbolize her wealth, emphasize her social status, or attract the opposite
sex” (Mellars 1999).
Research carried out since 1998 on the Châtelperronian culture of France
and northern Spain has dramatically changed the picture. At the Grotte du
Renne, in France, the Châtelperronian levels yielded bone awls identical to those
from the Ilsenhöle, but with three differences (d’Errico et al. 1998, 2003b): they
came in larger numbers; some featured regular decorative patterns; and they
were ­associated with body ornaments (Figure 4). These finds were published in
the early 1960s ­(Leroi-Gourhan 1961, 1964), but their significance was impaired
by doubts on the authorship of the Châtelperronian. In 1979, however, a
Neandertal skeleton was found in a Châtelperronian context at the French site of
St.-Césaire (Lévêque & V ­ andermeersch 1980), and, in 1996, the association was
confirmed for the fragmentary remains recovered at the Grotte du Renne itself
(Hublin et al. 1996). Eventually, it became clear that the Châtelperronian, with
its suite of personal ornaments, was an independent Neandertal development
predating modern human immigration by several millennia (Zilhão 2001, 2006b,
2007; d’Errico 2003). The conclusion that Neandertal society was symbolically
organized is further strengthened by results from use-wear analyses of hundreds
of chunks of black pigment from another and even earlier French cave site, the
Pech de l’Azé. These analyses concluded that they were pencils used for body
painting (Soressi & d’Errico 2007).

c
a b
d

1 cm

Figure 4. Pierced and grooved pendants made of animal bone and teeth, the most common
personal ornaments of Europe’s late Neandertals, all from basal Châtelperronian level X of
the Grotte du Renne (Arcy-sur-Cure, France): (a-b) fox canines; (c) bison incisor; (d) lateral
phalange of reindeer
 João Zilhão

In short, late Neandertals had attained a level of cultural achievement i­ dentical

to that documented among their African contemporaries. What happened when
these two fully symbolic cultural traditions eventually met should be treated,
therefore, without preconception. Did they exchange genes and memes? Or was
mutual avoidance the rule, resulting in the extinction of one of the sides?
The answers must be sought in the biology and culture of the post-contact
populations, those of the earliest modern humans of Europe (Trinkaus 2007).
If we find no Neandertal contributions in those post-contact populations, then
we must conclude that interaction and admixture were trivial or non-existent.
But, if ­Neandertal contributions are indeed apparent, then we must conclude
that s­ ignificant interaction and admixture occurred, regardless of whether such
­contributions were or were not subsequently lost.

Figure 5. The Lagar Velho child skeleton

In 1998, the discovery and excavation of the 30,000-year-old burial of a five

year old child in the Lagar Velho rock shelter, Portugal, provided hard evidence
 Chapter 6. The emergence of language, art and symbolic thinking 

for a hypothesis that a group of scholars had been entertaining for many years:
that populations of the African lineage spreading into Europe would have inter-
bred with the local Neandertals, whose disappearance would have been largely a
process of assimilation, not extinction without descent (Zilhão & Trinkaus 2002).
In fact, this child (Figure 5) featured an anatomical mosaic mixing characteris-
tics for the most part modern, such as a clear, prominent chin and a high cranial
vault, with characteristics reminiscent or even distinctive of the Neandertals and
other archaic Eurasian populations, such as the robusticity of the leg bones, the
arctic, cold-adapted body proportions, and several minor features in the skull, the
mandible and the dentition. These features are known to be genetically inherited,
so their presence indicates a part-Neandertal ancestry for the child. Soon after
the Lagar Velho discovery, in 2003–2005, the Romanian cave of Oase was to pro-
vide additional evidence – the mandible of a young adult and the near complete
cranium of an adolescent, dated to 40 ka and at present Europe’s earliest modern
human fossils (Trinkaus et al. 2003; Rougier et al. 2007) – in support of this notion
(Figure 6).
The archaeological evidence supports this scenario. The Protoaurignacian
culture of western and central Europe is contemporary with the Oase fossils
and, as such, the first cultural entity that reliably can be assigned to European
early moderns. The personal ornaments of the Protoaurignacian are consistent
with this notion. For the most part, they are made of the same small marine
shell beads of diverse taxonomy but identical basket-shaped morphology found
among modern human cultures of the Near East and Africa, where they go back
to some 100 ka (cf. Figure 1). By comparison with these cultures, however, the
Protoaurignacian also features some novelties, such as pierced animal teeth,
namely of red deer and fox. These kinds of pendants are completely unknown
in Africa and the Near East prior to the time of contact. But they are precisely
the types of ornaments that characterize such pre-contact ­Neandertal-associated
cultures as the Châtelperronian (cf. Figure 4). The parsimonious explanation for
these elements of discontinuity with the African/Levantine tradition of p ­ ersonal
ornamentation can only be that modern humans acquired them from the
indigenous, Neandertal societies where such novel elements originally emerged
(Zilhão 2007).

5. Conclusion

The implication of these finds is that, in their strict, original formulations,

Multiregional Evolution and Mitochondrial Eve are now both obsolete
views of the tempo and mode of human evolution. The palaeontological and
archaeological evidence favours Assimilation models and refutes the notion that
 João Zilhão

Neandertals were a different species (Figure 7). Even for hard-line supporters
of the Neandertal’s fundamental separateness, the evidence still carries the twin
implications that (1) archaeologically visible manifestations of fully symbolic
sapiens behaviour emerged independently among different human species and,
(2) that the biological/genetical foundations for that behaviour must therefore
have existed in the human genus prior to the split between the African and
European lineages.

Figure 6. The Oase fossils. Top: the Oase 2 cranium. Bottom: the Oase 1 mandible

So, even in the framework of multi- rather than single-species views of human
evolution, the corollary of the last decade of empirical discoveries is that explana-
tions of the emergence of “behavioural modernity” as a simple by-product of a
putative speciation event in the late Middle Pleistocene of Africa are refuted – the
“hardware” requirements for symbolic thinking must have been in place before half
 Chapter 6. The emergence of language, art and symbolic thinking 

a million years ago, when the Neandertal lineage began to diverge. This c­ onclusion
has two additional corollaries: firstly, that the search for the genetic and cognitive
processes underlying the emergence of language and symbolism in the human
­lineage needs to be refocused on aspects of the differentiation and e­ volution of
Homo erectus people between two and one million years ago, s­ econdly, that the
much later appearance of personal ornaments represents a qualitative leap in
­culture, reflecting the operation of demographic and social factors (Gilman 1984;
Shennan 2001; Powell et al. 2009).

Multiregional Mitochondrial Assimilation

Evolution Eve Model
Time t2 Time t2

Time t1 Time t1

Time t Time t
Single species Multiple species Single species
Anagenetic evolution Cladogenesis, range with geographic
expansions, subspecies
extinctions Diffusion, admixture,
Inconsistent with Inconsistent with local population
the chronological the paleontological extinctions
evidence! evidence!

Figure 7. Different models for the explanation of the replacement of Neandertals by modern
humans in Europe

The commonplace notion that the first modern humans in Europe were
“astonishingly precocious artists” (Sinclair 2003: 774) whose superior cognition
sufficed to explain the demise of the Neandertals is also in contradiction with the
facts. The documented artistic skills of the earliest European moderns are identical
to those documented in late Neandertal cultural contexts, and consist simply of
patterned markings applied to bone tools with decorative or functional purposes.
The earliest figurative art (the cave paintings and ivory figurines from France and
southern Germany), in fact, post-dates by five millennia the first archaeological or
palaeontological indicators of modern human immigration (Zilhão 2007). Much
as with personal ornaments, the explanation for these novel developments must
therefore be sought in transformations occurring at that time in European society,
not in the human brain.
 João Zilhão

The ethnographic record abundantly documents that rock art primarily

functions as a way of embodying places with economic, ideological or social
significance, and the thousands of open air petroglyphs of the Côa Valley, in
Portugal (Zilhão et al. 1997; Baptista 2009), show that the same holds true for
the Palaeolithic period (Figure 8).

Figure 8. Two examples of the 17 km-long Palaeolithic rock art complex of the Côa
Valley, Portugal: top, Rock no. 6 of Penascosa, bottom, Rock no. 1 of Canada do Inferno
(after ­Baptista 2009, photos by P. Guimarães, courtesy of Parque Arqueológico do
Vale do Côa)
 Chapter 6. The emergence of language, art and symbolic thinking 

Thus, the parsimonious explanation for why art only appears in the archaeo-
logical record around 35 ka is that only then did the need arise for systems of social
identification/differentiation extending beyond the individual to include the land-
scapes and resources claimed as territory by the different groups to whom people
advertised their allegiance through the use of body ornaments. Sculpted figurines,
in turn, are likely to have represented manifestations of the same phenomenon in
the personal and domestic arenas of behaviour.
The need for such systems can easily be explained as a consequence of
adaptive success, with technological innovation leading to demographic growth
and implying both increased inter-group competition and increased regulation
of that competition. In such a context, it is easy to understand the adaptive value
of the emergence of ceremonial behaviours addressing issues of property and
rights over resources, and of the development of myths and religious beliefs
relating such rights to real or ideal ancestors. Therein lies the origins of art, not
in an evolutionarily late mutation endowing modern humans with the capacity
for symbolic thinking. The corresponding “hardware”, in fact, must have been in
place as soon as the size and shape of the brain case entered modern ranges of
variation and the cultural record documents behaviours that require language,
i.e. symbolic thinking by definition. The palaeontological record concurs in
suggesting that such a Rubicon had already been crossed by half a million years
ago. The rest is History.

References

Aiello, L. & Wheeler, P. 2003. “Neandertal Thermoregulation and the Glacial Climate.” In
Neanderthals and modern humans in the European landscape during the last glaciation,
T. H. van Andel & W. Davies (eds), 147–166. Cambridge: McDonald Institute for Archaeo-
logical Research.
Baptista, A.M. 2009. O Paradigma Perdido. O Vale do Côa e a Arte Paleolítica de Ar Livre em
Portugal. Vila Nova de Foz Côa/Porto, Parque Arqueológico do Vale do Côa/Edições
Afrontamento.
Bermúdez De Castro, J.M., Martinón-Torres, M., Carbonell, E., Sarmiento, S., Rosas, A., Van Der
Made, J. & Lozano, M. 2004. “The Atapuerca Sites and Their Contribution to the Knowledge
of Human Evolution in Europe.” Evolutionary Anthropology 13: 25–41.
Bouzouggar, A., Barton, N., Vanhaeren, M., d’Errico, F., Collcutt, S., Higham, T., Hodge,
E., Parfitt, S., Rhodes, E., Schwenninger, J.-L., Stringer, C., Turner, E., Ward, S., Moutmir,
A. & Stambouli, A. 2007. “82,000-year-old shell beads from North Africa and implications
for the origins of modern human behavior.” Proceedings of the National Academy of Sciences
104: 9964–9969.
Cann, R.L., Stoneking, M. & Wilson, A.C. 1987. “Mitochondrial DNA and human evolution.”
Nature 325: 31–36.
 João Zilhão

Carbonell, E., Bermúdez De Castro, J.M., Parés, J.M., Pérez-González, A., Cuenca-Bescós,
G., Ollé, A., Mosquera, M., Huguet, R., Van Der Made, J., Rosas, A., Sala, R., Vallverdú,
J., García, N., Granger, D.E., Martinon-Torres, M., Rodríguez, X.P., Stock, G.M., Vergès,
J.M., Allué, E., Burjachs, F., Caceres, I., Canals, A., Benito, A., Diez, C., Lozano, M., Mateos,
A., Navazo, M., Rodríguez, J., Rosell, J. & Arsuaga, J.L. 2008. “The first hominin of Europe.”
Nature 452: 465–470.
Clottes, J., Chauvet, J.­‑M., Brunel­‑Deschamps, E., Hillaire, Ch., Daugas, J.­‑P., Arnold, M.,
­Cachier, H., Evin, J., Fortin, Ph., Oberlin, Ch., Tisnerat, N. & Valladas, H. 1995. “Les
­peintures paléolithiques de la Grotte Chauvet­‑Pont d’Arc (Ardèche, France): datations
directes et indirectes par la méthode du radiocarbone.” Comptes Rendus de l’Académie des
Sciences de Paris 320, IIa:1133­‑1140.
Conard, N.J. 2008. “A Critical View of the Evidence for a Southern African Origin of ­Behavioural
Modernity.” South African Archaeological Society Goodwin Series 10: 175–179.
Conard, N.J. 2009. “A female figurine from the basal Aurignacian of Hohle Fels Cave in
­southwestern Germany.” Nature 459: 248–252.
Deacon, T. 1997. The Symbolic Species: The Coevolution of Language and the Brain New York: W.
W. Norton & Co.
d’ Errico, F. 2003. “The Invisible Frontier. A Multiple Species Model for the Origin of Behavioral
Modernity.” Evolutionary Anthropology 12: 188–202.
d’Errico, F., Zilhão, J., Baffier, D., Julien, M. & Pelegrin, J. 1998. “Neanderthal Acculturation in
Western Europe? A Critical Review of the Evidence and Its Interpretation.” Current Anthro-
pology 39: 1­‑44.
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. 2003a. “Archaeological
Evidence for the Emergence of Language, Symbolism, and Music -An Alternative Multidis-
ciplinary Perspective.” Journal of World Prehistory 17 (1): 1–70.
d’Errico, F., Julien, M., Liolios, D., Vanhaeren, M. & Baffier, D. 2003b. “Many awls in our argu-
ment. Bone tool manufacture and use in the Châtelperronian and Aurignacian levels of
the Grotte du Renne at Arcy-sur-Cure.” In The Chronology of the Aurignacian and of the
Transitional Technocomplexes. Dating, Stratigraphies, Cultural Implications, J. Zilhão and
F. d’Errico (eds), 247–270. Lisboa: Trabalhos de Arqueologia 33, Instituto Português de
Arqueologia.
d’Errico, F., Henshilwood, C., Vanhaeren, M. & Van Niekerk, K. 2005. “Nassarius kraussianus
shell beads from Blombos Cave: evidence for symbolic behaviour in the Middle Stone Age.”
Journal of Human Evolution 48: 3–24.
Dennell, R. & Roebroeks, W. 2005. “An Asian perspective on early human dispersal from Africa.”
Nature 438: 1099–1104.
Eswaran, V., Harpending, H. & Rogers, A.R. 2005. “Genomics refutes an exclusively African
origin of humans.” Journal of Human Evolution 49 (1): 1–18.
Evans, P.D., Mekel-Bobrov, N., Vallender, E.J., Hudson, R.R. & Lahn, B.T. 2006. “Evidence that
the adaptive allele of the brain size gene microcephalin introgressed into Homo sapiens
from an archaic Homo lineage.” Proceedings of the National Academy of Sciences 103 (48):
18178–18183.
Forster, P. 2004. “Ice Ages & the mitochondrial DNA chronology of human dispersals: a review.”
Philosophical Transactions of the Royal Society London B 359: 255–264.
 Chapter 6. The emergence of language, art and symbolic thinking 

Gagneux, P., Wills, C., Gerloff, U., Tautz, D., Morin, P.A., Boesch, C., Fruth, B., Hohmann, G.,
Ryder, O.A. & Woodruff, D.S. 1999. “Mitochondrial sequences show diverse e­ volutionary
histories of African hominoids.” Proceedings of the National Academy of Sciences 96:
5077–5082.
Gamble, C. 1998. “Palaeolithic Society and the Release from Proximity: A Network Approach to
Intimate Relations.” World Archaeology 29 (3): 426–449.
Gilman, A. 1984. “Explaining the Upper Palaeolithic Revolution.” In Marxist Perspectives in
Archaeology, M. Spriggs (ed.), 115­‑126. Cambridge: Cambridge University Press.
Green, R.E., Krause, J., Ptak, S.E., Briggs, A.W., Ronan, M.T., Simons, J.F., Du, L., Egholm,
M., Rothberg, J.M., Paunovic, M. & Pääbo, S. 2006. “Analysis of one million base pairs of
­Neanderthal DNA.” Nature 444: 330–336.
Green, R.E., Krause, J., Briggs, A.W., Maricic, T., Stenzel, U., Kircher, M., Patterson, N., Li, H.,
Zhai, W., Fritz, M. H.-Y., Hansen, N.F., Durand, E.Y., Malaspinas, A.-S., Jensen, J., Marques-
Bonet, T., Alkan, C., Prüfer, K., Meyer, M., Burbano, H.A., Good, J.M., Schultz, R., Aximu-
Petri, A., Butthof, A., Höber, B., Höffner, B., Siegemund, M., Weihmann, A., Nusbaum, C.,
Lander, E.S., Russ, C., Novod, N., Affourtit, J., Egholm, M., Verna, C., Rudan, P., Brajkovic,
D., Kucan, Ž., Gušic, I., Doronichev, V.B., Golovanova, L.V., Lalueza-Fox, C., De La Rasilla,
M., Fortea, J., Rosas, A., Schmitz, R.W., Johnson, P.L. F., Eichler, E.E., Falush, D., Birney, E.,
Mullikin, J.C., Slatkin, M., Nielsen, R., Kelso, J., Lachmann, M., Reich, D. & Pääbo, S. 2010.
“A Draft Sequence of the Neandertal Genome.” Science 328: 710–722.
Henshilwood, C. & Marean, C. 2003. “The Origin of Modern Human Behavior. Critique of the
Models and Their Test Implications.” Current Anthropology 44 (5): 627–651.
Holliday, T.W. 2006. “Neanderthals and modern humans: an example of a mammalian
­syngameon?” In Neanderthals Revisited: New Approaches and Perspectives, K. Harvati and
T. Harrison (eds), 289–306. New York: Springer.
Hublin, J.-J., Spoor, F., Braun, M., Zonneveld, F. & Condemi, S. 1996. “A late Neanderthal
­associated with Upper Palaeolithic artefacts.” Nature 381: 224–226.
Hülle, W.M. 1977. Die Ilsenhöhle unter Burg Ranis/Thüringen. Eine paläolitische Jägerstation.
Stuttgart: Gustav Fischer.
Klein, R.G. 2003. “Whither the Neanderthals.” Science 299: 1525–1527.
Koller, J., Baumer, U. & Mania, D. 2001. “High-tech in the middle Palaeolithic: Neandertal-
manufactured pitch identified.” European Journal of Archaeology 4 (3): 385–397.
Krings, M., Stone, A., Schmitz, R.W., Krainitzki, H., Stoneking, M. & Pääbo, S. 1997. “­ Neandertal
DNA Sequences and the Origin of Modern Humans.” Cell 90: 19­‑30.
Kuhn, S.L., Stiner, M.C., Reese, D.S. & Güleç, E. 2001. “Ornaments of the earliest Upper
­Paleolithic: New insights from the Levant.” Proceedings of the National Academy of Sciences
98 (13): 7641–7646.
Lahr, M.M. & Foley, R. 1998. “Towards a Theory of Modern Human Origins: Geography, Demog-
raphy, and Diversity in Recent Human Evolution.” Yearbook of Physical ­Anthropology 41:
37–176.
Leakey, R. & Lewin, R. 1977. Origins. London: Mcdonald and Jane’s.
Lee, S.-H. & Wolpoff, M.H. 2003. “The pattern of evolution in Pleistocene human brain size.”
Paleobiology 29 (2): 186–196.
Leroi-Gourhan, A. 1961. “ Les fouilles d’Arcy-sur-Cure.” Gallia Préhistoire 4: 3–16.
Leroi-Gourhan, A. 1964. Les religions de la Préhistoire. Paris: Presses Universitaires de France.
 João Zilhão

Lévêque, F. & Vandermeersch B. 1980. “Découverte de restes humains dans un niveau

castelperronien à Saint-Césaire (Charente-Maritime).” Comptes rendus de l’Académie des
Sciences de Paris 291D: 187–189.
McBrearty, S. & Brooks, A. 2000. “The revolution that wasn’t: a new interpretation of the origin
of modern human behavior.” Journal of Human Evolution 39: 453–563.
McCown, T.D. & Keith, A. 1939. The Stone Age of Mount Carmel. Vol. 2. The fossil human remains
from the Levalloiso-Mousterian. Oxford: Clarendon Press.
McHenry, H.M. & Coffing, K. 2000. “Australopithecus-to-Homo: Transformations in Body and
Mind.” Annual Reviews of Anthropology 29: 125–146.
Mellars, P.A. & Stringer, C.B. (eds). 1989. The Human Revolution. Edinburgh: Edinburgh
University Press.
Mellars, P.A. 1999. “The Neanderthal Problem Continued.” Current Anthropology 40 (3): 341–350.
Mellars, P. Boyle, K., Bar-Yosef, O. & Stringer, C. (eds). 2007. Rethinking the Human Revolution.
Cambridge: McDonald Institute for Archaeological Research.
Meller, H. 2003. Geisteskraft. Alt- und Mittelpaläolithikum. Haale (Saale): Landesmuseum für
Vorgeschichte.
Noonan, J.P., Coop, G., Kudaravalli, S., Smith, D., Krause, J., Alessi, J., Chen, F., Platt, D., Pääbo,
S., Pritchard, J.K. & Rubin, E.M. 2006. “Sequencing and Analysis of Neanderthal Genomic
DNA.” Science 314: 1113–1118.
Powell, A., Shennan, S. & Thomas, M.G. 2009. “Late Pleistocene Demography and the
Appearance of Modern Human Behavior.” Science 324: 1298–1301.
Rougier, H., Milota, S., Rodrigo, R., Gherase, M., Sarcină, L., Moldovan, O., Zilhão, J., Constantin,
S., Franciscus, R.G., Zollikofer, C.P. E., Ponce De León, M. & Trinkaus, E. 2007. “Peşteracu
Oase 2 and the cranial morphology of early modern Europeans.” Proceedings of the National
Academy of Sciences 104 (4): 1165–1170.
Shennan, S. 2001. “Demography and Cultural Innovation: a Model and its Implications for the
Emergence of Modern Human Culture.” Cambridge Archaeological Journal 11 (1): 5–16.
Sinclair, A. 2003. “Art of the ancients.” Nature 426: 774–775.
Soressi, M. & d’Errico, F. 2007. “Pigments, gravures, parures: les comportements symboliques
controversés des Néandertaliens.” In Les Néandertaliens. Biologie et cultures, B. van der
Meersch & B. Maureille (eds), 297–309. Paris: Documents Préhistoriques (23), Éditions
du CTHS.
Stringer, C. & Gamble, C. 1993. Search of the Neanderthals. London: Thames and Hudson.
Templeton, A. 2002. “Out of Africa again and again.” Nature 416: 45–50.
Templeton, A. 2005. “Haplotype Trees and Modern Human Origins.” Yearbook of Physical
Anthropology 48: 33–59.
Thieme, H. 1997. “Lower Palaeolithic hunting spears from Germany.” Nature 385: 807–810.
Thorne, A. & Wolpoff, M. 2003. “The multiregional evolution of humans.” Scientific American
Special Issue 46–53.
Trinkaus, E. 2005. “Early Modern Humans.” Annual Reviews of Anthropology 34: 207–230.
Trinkaus, E. 2007. “European early modern humans and the fate of the Neandertals.” Proceedings
of the National Academy of Sciences 104 (18): 7367–7372.
Trinkaus, E. & Shipman, P. 1993. The Neandertals. New York: Knopf.
Trinkaus, E., Moldovan, O., Milota, S., Bîlgăr, A., Sarcina, L., Athreya, S., Bailey, S.E., Rodrigo,
R., Mircea, G., Higham, Th., Bronk Ramsey, C.H. & Plicht, J.V.D. 2003. “An early modern
human from the Peştera cu Oase, Romania.” Proceedings of the National Academy of
­Sciences 100: 11231–11236.
 Chapter 6. The emergence of language, art and symbolic thinking 

Vanhaeren, M., d’Errico, F., Stringer, C., James, S.L., Todd, J.A. & Mienis, H.K. 2006. “Middle
Paleolithic Shell Beads in Israel and Algeria.” Science 312: 1785–1787.
Wolpoff, M. 2002. Human Paleontology. Ann Arbor: University of Michigan.
Zilhão, J. 2001. Anatomically Archaic, Behaviorally Modern: The Last Neanderthals and Their
Destiny. Amsterdam: Stichting Nederlands Museum voor Anthropologie en Praehistoriae.
Zilhão, J. 2006a. “Genes, Fossils and Culture. An Overview of the Evidence for Neandertal-­
Modern Human Interaction and Admixture.” Proceedings of the Prehistoric Society 72: 1–20.
Zilhão, J. 2006b. “Neandertals and Moderns Mixed, and It Matters.” Evolutionary Anthropology
15: 183–195.
Zilhão, J. 2007. “The emergence of ornaments and art: an archaeological perspective on the­
­origins of ‘behavioural modernity.’ ” Journal of Archaeological Research 15: 1–54.
Zilhão, J., Aubry, Th., Carvalho, A.F., Baptista, A.M., Gomes, M.V. & Meireles, J. 1997. “The
Rock Art of the Côa Valley (Portugal) and its Archaeological Context.” Journal of European
Archaeology 5 (1): 7­‑49.
Zilhão, J. & d’Errico, F. 1999. “The chronology and taphonomy of the earliest Aurignacian
and its implications for the understanding of Neanderthal extinction.” Journal of World
­Prehistory 13 (1): 1–68.
Zilhão, J. & Trinkaus, E. (eds). 2002. Portrait of the Artist as a Child. The Gravettian Human
Skeleton from the Abrigo do Lagar Velho and its Archeological Context. Lisboa: Trabalhos de
Arqueologia 22, Instituto Português de Arqueologia.
Zilhão J., Trinkaus, E., Constantin, S., Milota, S., Gherase, M., Sarcina, L., Danciu, A., Rougier,
H., Quilès, J. & Rodrigo, R. 2007. “The Peştera cu Oase people, Europe’s earliest mod-
ern humans.” In Rethinking the Human Revolution, P. Mellars, K. Boyle, O. Bar-Yosef and
C. Stringer (eds), 249–262. Cambridge: McDonald Institute for Archaeological Research.
 chapter 7

The human major transition in relation

to symbolic behaviour, including language,
imagination, and spirituality
David Sloan Wilson
Departments of Biology and Anthropology, Binghamton University

Human evolution can be described in terms of three C’s: Cognition, Culture, and
Cooperation. Cognition includes the capacity for symbolic thought that lies at the
heart of both language and spirituality. Culture includes the capacity to transmit
information, both horizontally and vertically, leading to cumulative behavioural
change and rapid adaptation to local environments. Cooperation includes the
capacity to engage in prosocial behaviours far beyond one’s circle of genealogical
relatives and narrow reciprocators. The three C’s all have precursors in nonhuman
species, but they are vastly elaborated in our species.
In what sequence did the 3 C’s of human evolution arise and how are they
related to each other? A commonly invoked scenario is that the first step was
the evolution of advanced cognition, often called “theory of mind (ToM)”,
which enabled widespread cooperation and culture (e.g. Tomasello 1999). More
recently, a consensus is forming around a second scenario. The first step in
human evolution was a major evolutionary transition, which enabled within-
group cooperation to take place much more strongly than before (e.g. Boehm
1999; Wilson 2002, 2006, 2007; Tomasello et al. 2005). The major transition took
place without a prior advance in cognitive ability and was a pre-requisite for the
advanced forms of human cognition that we associate with language, symbolic
thought, and spirituality. Moreover, much simpler adaptations were required
as prerequisites for the advanced forms. The entire package of traits that make
humans so distinctive are forms of teamwork that require interactions among
trustworthy social partners. The first C to evolve was cooperation and the other
two C’s are forms of cooperation.

1. What is a major evolutionary transition?

Until the 1970’s, natural selection was thought to take place entirely by small
mutational change – individuals from individuals. Then, the cell biologist Lynn
Margulis (1970) proposed that nucleated (eukaryotic) cells are highly integrated
 David Sloan Wilson

communities of bacterial (prokaryotic) cells-individuals from groups. Her

­symbiotic cell theory was regarded as radical at the time but has since become
widely accepted. Cell organelles such as mitochondria and chloroplasts are clearly
derived from older species of microbes that led a more autonomous existence in
the deep past.
In the 1990’s, John Maynard Smith and Eors Szathmary (1995, 1999) gener-
alized Margulis’s theory to include other major events during the history of life,
such as the first cells, multi-cellular organisms, social insect colonies, and even
the origin of life itself as groups of cooperating molecular reactions. In each case,
natural selection can take place within a given unit, which tends to favour traits
that humans associate with cheating, free-riding, and exploitation. Or, natural
selection can take place between units, favouring traits that humans associate with
cooperation and prosociality. The balance between levels of selection is not static
but can itself evolve. A major transition occurs when selection between units dom-
inates selection within units, causing the unit to become a functionally adaptive
“superorganism.”
As an example, consider a proto-cell containing genes that can replicate
independently. Some genes are “solid citizens” that participate in the economy
of the whole cell, such as gathering resources or defense, all of which would be
recognized in human terms as public goods. Other cells “cheat” by using the
common resources of the cell to replicate faster than the “solid citizen” genes. It is
difficult to describe the scenario without resorting to the human lexicon of social
behaviours! The “cheat” genes have the highest relative fitness within the cell. They
will increase in frequency and will ultimately replace the “solid citizen” genes
unless something more can be added to the story. That “something” is natural
selection among cells. Cells with a high frequency of solid citizens will survive and
divide faster than cells with a high frequency of cheats. What evolves in the total
population depends upon the balance between opposing levels of selection, within
and among groups.
Against this background, consider the evolution of chromosomes, which binds
the genes within a cell into a single structure that must replicate as a unit. The
­chromosome prevents differential replication within groups, causing b
­ etween-group
selection to become the dominant evolutionary force. Natural selection now takes
place primarily on the basis of teamwork within the cell, turning the cell into a
superorganism. Similar scenarios can be described for the origin of life, multicel-
lular organisms, and social insect colonies.
There are three hallmarks associated with a major evolutionary transition:

1. They are rare events in the history of life. Evidently, it is not easy for
­between-group selection to dominate within-group selection.
 Chapter 7. The human major transition in relation to symbolic behaviour 

2. Once a major transition occurs, the new “superorganisms” become ecologically

dominant and diversify into many species. For example, eusociality in insects
is estimated to have originated less than 20 times, but these gave rise to many
thousands of eusocial insect species that account for over half of the biomass
of all insects (Holldobler & Wilson 2008).
3. The transition is never complete. Within group “cheating” is only suppressed,
never entirely eliminated. Even single cells and multicellular organisms, the
paradigm of internally harmonious organisms, have a disturbing number of
elements that benefit at the expense of the organism (Burt & Trivers 2006).

The concept of major evolutionary transitions is one of the most important devel-
opments in the history of evolutionary theory. It is so recent that the generalized
version didn’t appear in book form until the 1990’s!

2. Human evolution as a major transition

Although Maynard Smith and Szathmary were bold about expanding the concept
of major transitions, they were timid about applying it to human evolution,
restricting themselves to the genetic basis of language. Now it appears likely that
human evolution was a full-fledged major transition. The reason that we are so
unique among primates is because our ancestors became the primate equivalent
of a single organism or a social insect colony (Boehm 1999; Wilson 2007; Wilson
et al. 2008).
Recall that the key ingredient of a major transition is the suppression of
­fitness differences within groups, causing between-group selection to become the
­primary evolutionary force. In most primate species, including our closest ances-
tors, intense within-group competition limits the opportunities for cooperation
among members of the group. This is in contrast to extant human hunter-gatherer
­societies, which are fiercely egalitarian. What accounts for this shift and when did
it occur in human evolution?
Humans are incomparably better at throwing projectiles than other primates,
an ability that required whole-body anatomical changes and evolved early in the
hominin lineage. Although the original purpose of throwing was presumably
to deter predators and competing scavengers, it could also be used to suppress
bullying and other domineering behaviour within-groups (Bingham 1999). This is
a specific version of a more general hypothesis of guarded egalitarianism, advanced
by Boehm (1993, 1999) on the basis of the egalitarian nature of most extant hunter-
gatherer societies. However it was accomplished, guarded egalitarianism provides
the key ingredient of an evolutionary transition.
 David Sloan Wilson

In retrospect, human evolution has all the hallmarks of a major transition.

It was a rare event, occurring only once among primates. It had momentous
consequences; cooperation enabled our ancestors to spread over the planet,
­eliminating other hominins and many other species along the way. We also diver-
sified to occupy all climatic zones and hundreds of ecological niches, although
by cultural evolution rather than genetic evolution. The advent of agriculture
enabled us to increase the scale of society by many orders of magnitude by a
process of cultural multilevel selection (e.g. Turchin 2005). Finally, the transi-
tion was not complete. Within-group selection still takes place and is merely
­suppressed compared to between-group selection.

3. C
 ognitive teamwork and simple forms
that preceded more advanced forms

Physical forms of teamwork, such as hunting, gathering, defense, warfare, and

childcare, have always been emphasized in scenarios of human evolution. It is
somewhat less obvious that our mental capacities for symbolic thought on which
language and spirituality are based, along with our capacity for cultural transmis-
sion of traits, are forms of cognition cooperation that can only take place among
trustworthy social partners. Moreover, the advanced forms of cognition and
culture associated with modern humans required the prior evolution of much
simpler adaptations. The reason that these adaptations did not evolve in other pri-
mates is not because they were too complex, but because they were too helpful to
be advantageous in groups compromised by within-group selection. Examples of
such simple adaptations include the following.

1. Eyes as organs of communication. We are the only primate species with white
sclera and other features that provide information to our social partners about
direction of gaze, emotional state, and more. According to the “cooperative
eye” hypothesis, this is because this degree of helpfulness was not favoured in
any other primate species (reviewed by Tomasello et al. 2005).
2. Pointing. Even though pointing appears simple to us and appears early infancy,
apes evidently do not point things out to each other or do anything comparable
in natural environments. Apes raised in human social environments point to
what they want but seldom point to help out their human social partners.
What seems to be lacking is a sense of what others might want and/or
the motivation to help by pointing. In another telling comparison, dogs
comprehend pointing much better than wolves, even when wolves are raised
in human social environments. The mental ability of dogs evidently evolved,
 Chapter 7. The human major transition in relation to symbolic behaviour 

by genetic evolution, over 100,000 years of domestication (Tomasello et al.

2005; Hare et al. 2002).
3. Shared social awareness in human infants. From an amazingly early age,
human infants have an ability to adopt the perspective of others and to help by
pointing, joining a task, or otherwise coordinating activities. In short, humans
have instincts for teamwork that appear extremely early in life and are not
manifested in apes at any age (Tomasello et al. 2005).
4. The evolution of laughter. Converging lines of evidence strongly suggest that
our ancestors were laughing long before they were talking. Moreover, laughter
has functions that are clearly related to teamwork, rapidly coordinating mood
and facilitating playful activities during short periods of safety and satiety
(Gervais & Wilson 2005).

The precursors for these relatively simple adaptations were present in our ape
ancestors, and merely needed a more cooperative social environment to be
favoured and elaborated by between-group selection.

4. Language, imagination, and spirituality

According to Terrence Deacon (1998) in his book The Symbolic Species, humans
are unique in their capacity for symbolic thought. Yet, symbolic thought does not
require a bigger brain, or even a different brain, than possessed by our closest
ape relatives and even other non-human species. The reason that the capacity
for symbolic thought is absent in most species is not because it is too complex,
but because it isn’t adaptive. To understand the evolution of symbolic thought in
humans, we must search for an environmental context that made it adaptive.
What distinguishes symbolic thought from other forms of cognition, such as
associative learning? With associative learning, mental representations are learned
when they correspond directly to the environment and are unlearned when the
correspondence ceases (e.g. the sound of a bell when paired with food in Pavlovian
learning). With symbolic thought, mental representations take on a life of their
own, even when they don’t directly correspond to anything in the environment
(e.g. the word and concept of “ghost”). Symbolic thought can be hugely adaptive
because of the behaviours that they motivate in the real world, even though they
don’t necessarily correspond directly to features of the real world.
One implication of Deacon’s thesis is that non-human species can be
taught to think symbolically, more like us than their own kind. This retrospec-
tive ­“prediction” is confirmed by examples such as Kanzi the bonobo and Alex
the African Grey Parrot. The question is then to understand how the social
 David Sloan Wilson

e­ nvironment of proto-humans provided something equivalent to the lengthy

training procedure that enabled Kanzi and Alex to think (and speak, in the case
of Alex!) symbolically. It would need to be a cooperative social environment.
Thus, Deacon’s thesis fits very nicely within the new paradigm of human evolu-
tion as a major transition.
Language, imagination, and spirituality had at least three profound effects on
human evolution. First, they enabled a higher degree of social control (e.g. through
gossip), suppressing cheating and favouring teamwork of all kinds. ­Second, they
provided a new source of behavioural variation, upon which natural selection
can select. Third, they enabled the most successful behavioural alternatives to
be transmitted, creating a new mechanism of inheritance. These abilities enabled
our ancestors to achieve their ecological dominance and planetary diversity while
remaining a single species.

5. Testable predictions

The scenario outlined above leads to numerous predictions that can be tested in the
hominin fossil and early archeological record. The ability to throw stones requires
many skeletal changes that can be dated during hominin evolution. S­ uppression of
competition within groups includes competition among males for females, p ­ erhaps
reducing the degree of sexual dimorphism (although b ­ etween-group conflict can
favour large males). Even the evolution of the whites of our eyes can potentially be
genetically dated, once we identify the associated genes. Other ­testable predictions
are possible with sufficient ingenuity.

References

Bingham, P.M. 1999. “Human Uniqueness: A general theory.” Quarterly Review of Biology
74: 33–169.
Boehm, C. 1993. “Egalitarian society and reverse dominance hierarchy.” Current Anthropology
34: 227–254.
Boehm, C. 1999. Hierarchy in the Forest: Egalitarianism and the Evolution of Human Altruism.
Cambridge, Mass: Harvard University Press.
Burt, A., & Trivers, R. 2006. Genes in conflict. Cambridge, MA: Harvard University Press.
Deacon, T.W. 1998. The Symbolic Species. New York: Norton.
Gervais, M. & Wilson, D.S. 2005. “The Evolution and Functions of Laughter and Humor:
A Synthetic Approach.” Quarterly Review of Biology 80: 395–430.
Hare, B., Brown, M., Williamson, C. & Tomasello, M. 2002. “The domestication of social
­cognition in dogs.” Science 298: 1634–1636.
Holldobler, B. & Wilson, E.O. 2008. The Superorganisms. New York: Norton.
 Chapter 7. The human major transition in relation to symbolic behaviour 

Margulis, L. 1970. Origin of Eukaryotic cells. New Haven: Yale University Press.
Maynard Smith, J. & Szathmary, E. 1995. The major transitions of life. New York: W.H. Freeman.
Maynard Smith, J. & Szathmary, E. 1999. The origins of life: from the birth of life to the origin of
language. Oxford: Oxford University Press.
Tomasello, M. 1999. The cultural origins of human cognition. Cambridge, MA: Harvard Univer-
sity Press.
Tomasello, M., Carpenter, J., Call, J., Behne, T. & Moll, H. 2005. “Understanding and sharing
intentions: the origins of cultural cognition.” Behavioral and Brain Sciences 28: 675–735.
Turchin, P. 2005. War and Peace and War. New York: Pi Press.
Wilson, D.S. 2002. Darwin’s Cathedral: Evolution, Religion and the Nature of Society. Chicago:
University of Chicago Press.
Wilson, D.S. 2006. “Human groups as adaptive units: toward a permanent concensus.” In The
Innate Mind: Culture and Cognition, P. Carruthers, S. Laurence and S. Stich (eds), 78–90.
Oxford: Oxford University Press.
Wilson, D.S. 2007. Evolution for Everyone: How Darwin’s Theory Can Change the Way We Think
About Our Lives. New York: Delacorte.
Wilson, D.S., Van Vugt, M., & O’Gorman, R. 2008. “Multilevel selection and major evolutionary
transitions: implications for psychological science.” Current Directions in Psychological
Science 17: 6–9.
 chapter 8

The living as symbols, the dead as symbols

Problematising the scale and pace
of hominin symbolic evolution
Paul Pettitt
Department of Archaeology, University of Sheffield

The ‘symbolic capacity’ has come to be seen as a core trait of anatomically modern
humans, and probably separates them cognitively eand behaviourally from all
other hominins. While archaeologists agree on what aspects of the archaeological
record constitute evidence of symbolism, such as burials, use of pigments, and
personal ornamentation, only generic concepts of ‘symbolism’ are invoked from
these, resulting in a simplistic discourse about its origins. I try to problematise
the concept of symbolism, using these archaeological categories, breaking each
down into differing levels of symbolic sophistication. Following this, I try to
link these to Dunbar’s levels of intention, and explore how one might identify
these from the archaeological record. I conclude by making a necessarily coarse
comparison of Neandertals and modern humans in terms of the expression of
these characteristics.

1. Introduction: Living symbols, dead symbols

Palaeoanthropology is unique in providing insights into the long-term evolution

of human behaviour. To a certain extent it should also provide unique insights
into the cognitive capacities which underlie and facilitate certain behaviours. It
should, at least in theory, provide middle range bridging between the modern
human mind and that of our closest evolutionary relatives, the great apes. From
what scholars can tell, a vast cognitive gulf separates the two; and at the heart
of this difference apparently lies symbolism. If one accepts in a broad sense that
religion is itself symbolic, then it is clear that a symbolic capacity and religious
imperative is a fundamental part – perhaps an inevitability (Boyer 2008) – of being
human. It is no surprise, therefore, that documenting the emergence of symbolism
has in recent years become central to palaeoanthropology. One should add it to
the major events of human social development that traditionally involved broad
spectrum economies, agriculture, writing and state societies. In fact, it would be
difficult to conceive of these without symbolic underpinning.
 Paul Pettitt

Archaeologists have, however, approached the archaeological record in rela-

tively simple ways, focussing on simple objects as being simply indicative of sym-
bolism. Thus, the recovery of a used fragment of ochre becomes proxy evidence
of symbolism. Such an approach does not take us very far, and certainly does
not allow us to explore nuanced cognitive development among the later homini-
nae. We should think of symbols not as just material cultural objects or things in
the mind, but as ways of engaging with the world. Following on from Piercean
­semiotics, it becomes clear that what potentially makes signs symbolic is not
­necessarily inherent in the object itself, but is derived from its context, particularly
how the sign and its signifier are regarded in relation (Schults 2009). Thus context
is all-important, and, as Sloane Wilson (this volume) has noted, to understand the
development of s­ ymbolism we must search for the context that made it adaptive.
Here, I try to widen a contextual approach to the long-term evolutionary emer-
gence of hominin symbol use, deconstructing what we mean by symbolism.

2. P
 roblematising the archaeological debate: Symbolic revolutions
that were or were not

The emergence of ‘symbolism’ over the course of hominin evolution has in the past
few years become arguably the most important object of archaeological study in
the quest for defining what makes us ‘behaviourally modern’ humans. Palaeolithic
archaeologists seem to agree that the ‘symbolic capacity’ or ‘symbolically-mediated
behaviour’ is a defining – perhaps the defining – behavioural trait of Homo sapiens,
but debate has begun in the last few years as to whether ‘pre-modern’ hominins pos-
sessed symbolic capacities and, if so, to what extent (see, for example, Mellars 1991,
1995; McBrearty & Brooks 2000; Wadley 2001; Henshilwood & Marean 2003; d’Errico
et al. 2003 and particularly d’Errico 2003). Despite this, it is surprising that there
is still no agreement on a definition of symbolism (Wadley, this volume). Instead,
the debate so far has often centred upon a ‘trait list’ of behaviours that appear in
the late Middle Pleistocene and Upper Pleistocene that were apparently novel to the
hominin repertoire, although, as d’Errico (2003: 199) has succinctly observed, these
traits “are no more than a list of the major archaeological features that c­ haracterise
the Upper Palaeolithic in Europe”. Similarly, Henshilwood and Marean (2003) have
rightly emphasised that the trait list is Eurocentric in origin and thus of q
­ uestionable
relevance to behaviours which apparently emerged first elsewhere, and have conse-
quently proposed new traits with which to distinguish behavioural ‘modernity’. This
has given the debate a new, at present Afrocentric, bias.
Whether African, European or, for that matter, Asian (a land mass far larger
than Europe and the archaeologically explored areas of Africa put together, for
 Chapter 8. The living as symbols, the dead as symbols 

which relatively little is known) (Dennell 2008), implicit in all perspectives is the
progressivist notion of an accumulation of novel behaviours that include symbolism.
As with other aspects of the palaeoanthropological record (notably the geographical
dispersal of hominins), archaeologists tend to assume that from their point of
emergence, hominins and behaviours were present continuously from then on;
dots signifying new behaviours are placed on time charts and the dots are then
joined up, creating an impression of gradually increasing behavioural complexity.
This is apparent, for example, in McBrearty and Brooks’ (2000) reorientation of the
development of behavioural modernity to Africa and to a long-term gradualism.
They see a “fitful expansion” that was “built incrementally” (ibid.: 531) over the long
duration “since at least 250 ka” (ibid.: 532). With regard to symbolism, they discuss:
special treatment of the dead (no evidence before the Upper Pleistocene and, in fact,
no uncontroversial evidence for the African Middle Stone Age (MSA)); beads and
ornaments (again no evidence before the Upper Pleistocene – see discussion below);
and use of pigments (no figurative art until the Later Stone Age (LSA), the earliest
date usually quoted being 26–28 ka BP for the Apollo 11 Cave plaques which are not
reliably dated and may be much younger; but recovery of pigments from various
African MSA sites spans the last 250,000 years). Leaving aside controversially dated
examples, the evidence for early use of pigments comprises less than one dozen sites,
far lower than the number of European Middle Palaeolithic sites with the same.
Of the African examples, however, a small group of sites cluster around 100 ka
BP, a handful of other sites may furnish much older examples, and others post-
date 80 ka BP. The earliest manifestations are few and are hampered by imprecise
dating. ‘Red stained earth’, numerous haematite fragments, ochre and grinding
stones occur in a context older than 250 ka BP at Kapthurin; natural stone balls
coloured with manganese were recovered from Olorgesailie site BOK1 dating to
>340 ka BP; and at site GOK1 a stone block coloured with ochre and bearing
grinding marks was recovered from a context dating to 290–493 ka BP (Brooks
& Yellen 2009). At least 150,000 years later than this group, a second comprises
red ochre fragments at Klasies River Mouth >100 ka (although the majority of
pieces are younger than 80 ka); haematite pencils “throughout the MSA sequence
at ­Border Cave” (McBrearty & Brooks 2000: 528), the base of which has been
estimated at >100 ka; grinding slabs (not pigment crayons) from Porc Epic, ESR
dated to 121 ± 6 ka BP; and, most remarkable of the group by far, the numerous
ochre fragments in all of the main (M1, M2 and M3) levels of Blombos Cave,
South Africa, at least 18 of which bear engravings which can arguably be grouped
into symbolic ‘traditions’ (Henshilwood et al. 2009). It was generally agreed in the
Homo symbolicus workshop that it is difficult to interpret the Blombos examples
as anything but symbolism. This may now be considered a robust archaeological
record, but we may not be justified in drawing a continuous line from Kapthurin
 Paul Pettitt

to the LSA. Well over 100,000 years separate Kapthurin from the earliest cluster
of sites ~100 ka BP, and twenty thousand years or more separate this cluster from
more numerous examples <80 ka BP (Henshilwood and d’Errico 2009). 240,000
years separate the use of pigments at Kapthurin and the earliest known appearance
of figurative art after (possibly well after) 40,000 years in Africa. With a burgeoning
number of well-excavated MSA sites, we cannot simply assume this is a factor
of recovery, and I shall forward below a falsifiable hypothesis to account for this
pattern.
‘Time depth’ is implicit in gradualist models of the emergence of human
­behaviour, but was it a significant aspect of symbolic behaviour? Perhaps we
archaeologists implicitly assume it is. Once a novel behaviour appears, particularly
one as profound as the symbolic capacity, it seems logically inconceivable to us that
it might disappear again. Gaps in the chronological representation of these behav-
iours are written off as deriving simply from the lack of excavated sites, justifying
the drawing of solid lines. But why should the appearance of these behaviours have
been cumulative? Broad surveys of the Eurasian Middle Palaeolithic and African
MSA show how the appearance and chronological trajectories of many behaviours
varied considerably, region to region (even within France, for example), often lacked
clear trajectories, and in Europe did not inevitably lead to the Upper Palaeolithic.
‘Recursive change’, whereby novel traits appear in a region, rise in f­ requency, then
disappear can be observed in the Levant and may have been widely characteristic
of pre-modern behaviour (see papers in Hovers & Kuhn 2006). As d’Errico and
Henshilwood (2009) have noted, there is no continuity in the expression of pigment
use and personal ornamentation. As this is so, then why might a recursive nature
not apply for symbolism or, for that matter, religion (Pettitt in press)? Chase (e.g.
1999, 2006) has specifically drawn attention to the recursive nature of symbolism,
arguing that its manifestation should vary depending upon specific behavioural
contexts, and thus that we should expect its appearance in the archaeological record
to vary geographically and chronologically.
If one can challenge the ‘out of Africa 1’ model on palaeontological and
archaeological grounds (Dennell & Roebroeks 2005) there is no reason why we
should not be critical of the currently favoured African model for the origins
of behavioural modernity (Pettitt 2007). If regional trajectories of change and
‘recursion’ are ­characteristic of pre-LSA/Upper Palaeolithic hunter-gatherers (see
papers in Hovers & Kuhn 2006), then the African MSA may not have been so
critical to the emergence of modern behaviour after all (Pettitt 2007). In this light,
the ­multiregional, multispecies model for the emergence of modern behaviour
­proposed by d’Errico (2003) becomes highly feasible; ‘ “modern” [behavioural]
traits may have appeared in different regions and among different groups of
humans, much as happened later in history with the inventions of agriculture,
writing, and state society’ (ibid. 200).
 Chapter 8. The living as symbols, the dead as symbols 

For the purposes of this paper I therefore assume a null hypothesis that there
was no single centre of emergence of symbolism among hominin societies, or at
least that such a single centre will not be recognisable archaeologically. Instead of
trying in vain to identify origins, I shall instead attempt to deconstruct the notion
of ‘symbolism’ as used by palaeoanthropologists, and suggest a more fragmented
way in which it may have arisen among hominin groups, both in the long and
short terms. I begin by elaborating an heuristic scheme using a relatively well
documented and debated source of data – personal ornamentation – and then
proceed to deconstruct another aspect of behaviour seen as behaviourally ‘mod-
ern’ by many – the special treatment of the dead. Finally, I try to integrate these
with Dunbar’s suggestions about the cognitive development of levels of intention
over the course of hominin evolution to show how symbolism can occur at many
organisational and cognitive levels.

3. Material culture symbols among the living

Symbols only function as such when both a writer and a reader are in accord. In
archaeology, one tends to focus on the writer (i.e. through their non-perishable
creations recovered through archaeology), and assume that all persons who came
into contact with these material culture creations were informed readers (i.e. they
could decode the intentional messages these were created to contain). This need not
necessarily have been the case, and while an object can be considered as a symbol
if it was created to function as such, it does not necessarily follow that it was widely
or universally successful in that functioning. A shell pendant might therefore have
functioned symbolically among the conspecifics of whichever erstwhile occupant
of Blombos Cave made it around 80 ka BP (Henshilwood et al. 2004), but its status
as a symbol may, or may not, have disappeared if it were viewed by other African
Homo sapiens populations, or, for that matter, by an archaeologist eighty thousand
years in the future. The only guide as to the efficiency of a symbol’s agency in
the past might therefore be an abundance of that symbol, not only on one site,
but among several sites of the same broad time period. Until we have such an
archaeological record it might be rash to argue from a handful of sites of widely
different ages that symbolism was widespread among groups and geographical
regions, let alone endemic to the species.
This caution might be applied to a small group of artefacts that are often
forwarded as potentially very early examples of symbolism. Three examples of
these pierres figures are known; natural stone cobbles that fortuitously resemble
the human form, a resemblance which was accentuated by restricted use of delib-
erate engraving. Two derive from the Lower Palaeolithic (from the Levallois-rich
site of Berekhat Ram, Israel, probably 350–500 ka BP; and from Acheulian d ­ eposits
 Paul Pettitt

at Tan-Tan, Morocco, around 400 ka BP), and one from the Middle Palaeolithic
(from the late Middle Palaeolithic cave site of La Roche Cotard, France ~32 ka
[14C] BP). Respectively, these take the form of a pebble of basaltic tuff resembling
a human torso and head, intentionally modified with grooves around its ‘neck’
and ‘sides’ (d’Errico & Nowell 2000); a quartzite cobble reminiscent of a human
body modified with eight grooves and with the addition of red pigment (Bednarik
2003); and a flint beach cobble, around the periphery of which several flakes have
been removed, through which a natural perforation runs and into which a bone
splinter has been wedged – the overall effect resembling a human face (Marquet
& Lorblanchet 2003). While a sample of three, widely spaced in time, is hardly
grounds for robust interpretation of pierres figures as unambiguous indicators of
early symbolism, we should not write them off as casual “lithic doodles,” as ­Dennell
(2008: 285) has noted. Instead, he argues that, like the appearance of precocious
lithic technologies in the Lower Palaeolithic such as end-scrapers and burins at
Berekhet Ram, symbolism (and by extension, perhaps ritual) drifted in and out of
use. In this case, “rather than dismissing these objects as non-symbolic that would
be regarded as symbolic if found in later contexts, it might be advisable to consider
instead why they are so rare, and under what circumstances they might occur”
(ibid.: 285). Indeed, why are these figures not more common in the archaeological
record? This cannot be due to recovery bias as one might argue for figurative art,
so we may presumably conclude that their occurrence was genuinely rare, evi-
dence of the regionally (or perhaps culturally) varied expression of early symbolic
systems. But what kind of symbolism? The process begins with a reader, as the
process is predicated on the initial recognition of the human form in a natural
object. Thereafter the reader becomes the writer, making artificial modifications
of the natural object to enhance its resemblance to the human form. The creative
process therefore relates to the conscious removal of ambiguity in the symbol’s
reading. Whether or not it was subsequently ‘read’ by its discoverer/creator alone
or by others, the object must in any general use of the term be considered to be
symbolic, because it carries within itself an explicit reference to the human body.
But that is all; we can infer nothing further from its message: ‘I look like a human’.
While we cannot, of course, rule out that the pierres figures symbolised a lot more
(e.g. ‘I represent my creator, his agency in the group while he is not present, and
the shared social norms that keep us together’) there is a considerable conceptual
gap between the two forms of symbolism. The simplest way in which these pierres
figures may have been used (if of course they were ‘used’ at all) cannot be regarded
as cognitively sophisticated as the latter example. We need to deconstruct what we
as archaeologists mean by symbolism, and I attempt to do so here with reference
to personal ornamentation. I include pigment use in this argument, for which rea-
sons should become apparent.
 Chapter 8. The living as symbols, the dead as symbols 

Whatever our opinion of the robustness of the earliest evidence for pigment
use, some have seen the recovery of pigment ‘crayons’ from sites such as Twin
Rivers (Zambia) and Kapthurin (Kenya) as “…convincing proof of the symbolic
use of pigments…” by 200 ka years ago (d’Errico et al. 2003: 4 my emphasis). No
criteria, however, have been proposed that allow us to identify from the archae-
ological record exactly how pigments were used (see the useful discussion in
­Henshilwood et al. 2009 & Wadley 2005). The apparent selection of a small selec-
tion of colours from a wider variety of those available (e.g. Barham 1998) and
the selection of highly saturated reds in both South Africa (Watts 1999) and in
Qafzeh Cave, Israel (Hovers et al. 2003) does at least suggest a symbolic func-
tion, but the problem here is that archaeologists assume a broad interpretation of
­‘symbolism’, namely that if pigments were in use, then whatever their specific use
was it possessed a symbolic dimension. This is not simply a semantic problem;
scales of symbolism vary from the simple to the complex, and archaeologists tend
to assume only the latter. Simple ornaments, in the form of natural shells pierced
for suspension, are known from secure contexts in Blombos Cave, South Africa for
which a date of ~75 ka BP is usually cited (Henshilwood et al. 2004); Skhul Cave,
Israel in a horizon dated to 100–135 ka BP; Qafzeh Cave, Israel around 90–100 ka
BP (Vanhaeren et al. 2006); the Grotte des Pigeons, Taforalt, Morocco p ­ ossibly
around 82 ka BP (Bouzouggar et al. 2007, although one would like to see the OSL
dates on which this is based backed up by other methods); Üçağizli Cave, ­Turkey
around 39–41 ka BP and probably the same broad age at Ksar Akil, ­Lebanon
(Kuhn et al. 2001); and >35 ka BP from Oued Djebbana, Algeria (Vanhaeren
et al. 2006).1 Taking due consideration of chronometric imprecision, none of these
need ­pre-date 100 ka BP. Taking the mean ages at face value, 30 000 years or more

. The dates given are those generally cited in the literature. For Skhul, the ornaments were
recovered from Layer B, for which ESR and U-series dates indicated an age range of 43–134
ka BP, ‘but recent ESR and U-series analyses, including direct dating of a [human] molar from
the Skhul II skeleton, indicate ages between 100 and 135 ka’ (Vanhaeren et al. 2006, 786).
There are, however, large errors associated with the coupled ESR/U-series dates from this level
­(including that on Skhul II at 116 +43/-24 ka BP) and the best estimate of the age of the Skhul
II and IV is 98 +19/-10 ka BP (Grün et al. 2005) which could therefore be as young as 78 ka BP
at 2 Sigma. Phase M1 at Blombos, from which 41 perforated tick shells derive, has been dated
by OSL to 75.6 ± 3.4 ka BP and by TL to 77 ± 6 ka BP (Henshilwood et al. 2004, 304), thus
could be younger than 70 ka BP at 2 Sigma. With one infinite conventional radiocarbon date
for the open air site of Oued Djebbana, I consider it undated even though it has an Aterian
attribution ­(Vanhaeren et al. 2006). Although the age of the ornaments from Üçağizli Cave are
usually cited at 39–41 ka BP, the age range of five 14C dates for Layer H reported by Kuhn et al.
(2001) could indicate an age as young as 35 ka BP at 2 Sigma, and ages for the relevant contexts
at Ksar Akil are based on poor chronometric dates and estimated sedimentation rates.
 Paul Pettitt

separate Blombos backwards to Skhul, and forwards to Ksar Akil and Üçağizli1.
We should t­herefore be cautious about inferring that “the initial appearance of
Upper ­Palaeolithic ­ornament technologies was ­essentially simultaneous on three
continents” (Kuhn et al. 2001: 7641) or that these simple points in time reflect “a
long-lasting and widespread beadworking tradition [that] existed in Africa and the
Levant” ­(Vanhaeren et al. 2006: 1788). A simultaneous emergence and ­continuous
tradition may, of course, eventually be demonstrated beyond reasonable doubt,
but for now the archaeological record does not demonstrate this and, I suggest, we
should conceive of a null hypothesis – there for elimination – that the appearance
of traditions of personal ornamentation varied from region to region and, like the
hominin populations ­themselves, were by no means continuous.
If discontinuity and regional variation in use was the rule, it follows that
­symbols need not have functioned in the same ways among different groups, or
for that matter between different individuals. Assuming that pigment colourants
and perforated shells and beads were used to ornament the body, one can conceive
of different levels of use, from the simplest – what one might argue to be non-
symbolic decoration – through to concept-mediated symbolism. Such a scheme
of different symbol use need not be cumulative. I suggest, for example, that one or
several of the following could be in operation at any one time or place:
Decoration: the employment of colouring/ornamentation for visual effect with no
associated symbolic meaning, or the uninformed reading of an otherwise sym-
bolic code (‘I wear red because I like red’).
Enhancement: the use of colouring/ornamentation/modification to bring out a
simple (symbolic) message by enhancing existing clues (‘I wear red as I know you
will read it as a sign of my strength or be impressed by it’).
Accessorization: the use of colouring/ornamentation/modification to make a more
subtle or specific statement than enhancement by acting as a material cultural
accessory to message (‘I wear red as I know you will recognise it as the regalia of
our clan and infer from it that were are culturally the same’).
Full symbolism: the use of colouring/ornamentation/modification to make an
explicit statement by acting as a full material cultural symbol from which a reader
can decode complex messages from (‘I wear red as, like you, I am a successful hunter
and have killed an adult eland; it is my right to wear this colour and I ­therefore
­command respect from all’).
Time/space-factored symbolism: the incorporation of temporal and spatial dimen-
sions into full symbolism, e.g. beliefs, myths and stories, object biographies and
histories (‘I wear red only at a specific time, marking the time of the year when the
ancestors created this land, in honour of the creation myths and to mark out that I
am the bearer of this knowledge’).
 Chapter 8. The living as symbols, the dead as symbols 

A fanciful set of examples, but the problem is real: at which of these levels were
the tick shells and engraved ochre fragments from Blombos Cave f­unctioning?
One, or multiple? As d’Errico and Henshilwood (this volume) have noted, pigment
fragments are notoriously ambiguous as they ‘do not represent the direct outcome
of past symbolic behaviours’. All one can do is apply a logical approach to intuitive
interpretation. One might rule out simple decoration, given that one can observe
redundancy in the selection of specific shells and creation of specific engravings
(suggesting that each had specific meaning), but how might we confidently infer
the full symbolism or time/space-factored symbolism usually assumed by archae-
ologists from them? The problem is particularly acute on sites with the recovery of
non-engraved ­pigment crayons. Leaving aside the often intractable arguments that
pigments could be used for more prosaic purposes (well demonstrated by ­Wadley
2005, although cf. Watts 1999 & Hovers et al. 2003), we cannot eliminate the
hypothesis that the recovery of pigment ‘crayons’ alone refers only to decoration or
enhancement. On the other side of the coin of course a sceptic might argue that if
one was to interpret Leonardo’s workshop on the grounds of his pigments alone, we
might reduce him to decorated dunce, but one might propose that we should only
confidently interpret pigment crayons alone as being indicative of full ­symbolism
when they occur at times and places where figurative art or artificial memory
­systems are also known. In this case, context would be critical to the e­ lucidation
of specific symbolic systems. Interpreting the data presented by McBrearty and
Brooks (2000) in this light, I s­ uggest the following (I hope falsifiable) hypothesis for
the emergence of these aspects of behavioural modernity in Africa:

1. Simple pigment crayons and pigment processing found before 100 ka BP and
back, perhaps, to 250 ka BP or beyond, represent little more than decoration
or enhancement. The recovery of large numbers of fragments and, particu-
larly, engravings on fragments or other engraved objects in association with
pigments, would allow us to reject this hypothesis.
2. The floruit of personal ornamentation after 100 ka BP, including traditions of
engraved designs on ochre, suggests levels of accessorisation or full symbol-
ism was in operation around 100 ka BP and after 80 ka BP.
3. True time/space-factored symbolism, in which figurative art often plays a role,
did not emerge in Africa until after 40 ka years ago.

By this argument, I would also suggest that the Eurasian Middle Palaeolithic record
(at least 40 Neandertal sites in Europe and a handful stretching back to the Lower
Palaeolithic) (d’Errico 2003) shows that some groups of Homo neandertalensis
and Homo heidelbergensis engaged in decoration or enhancement. Of interest here
would be the items of personal ornamentation recovered from a C ­ hâtelperronian
 Paul Pettitt

context in Layer X of the Grotte du Renne at Arcy-sur-Cure, given that the

­association of symbolic items with the body would be strong indication of symbol-
ism of at least the level of enhancement (d’Errico & Henshilwood, this ­volume).
It should at least be obvious by now that simply referring to these examples as
‘symbolic’ results in the meaningless attribution of ‘part behavioural modernity’ to
the Neandertals. Such attributions take us nowhere; we need to define symbolism
by deconstructing it; and we need to develop heuristic schemes that allow us to
explore how we might identify levels of symbolism from the archaeological record
(and, perhaps, primatological world – a difficult task, see McGrew, this volume).

4. The dead as symbols

The ‘trait list’ approach treats burial of the dead as an aspect of ‘modern’ behaviour,
although proponents of the importance of burial do not define why it deserves to be
on the list. How, for example, does placing a corpse in a shallow grave really differ
from a female chimpanzee carrying around the body of her dead child for several
days or, in one case, an entire month (e.g. Goodall 1986; Matsuzawa 2003; Pettitt
2011)? As with personal ornamentation, there is no reason why the treatment of the
dead could not have differed in its symbolic function over the course of hominin
evolution. Elsewhere, I argue that the social interaction of the living with the dead
has a very long evolutionary history, beginning with the intellectual interest in the
corpse (which I refer to as morbidity) that can be observed among extant primates.
This became elaborated through the deliberate deposition (mortuary caching) of
the dead at certain parts of the natural landscape, until features for caching were
deliberately created (burials) and locales were given specific symbolic meaning as
places for the dead (Pettitt 2011). Three examples widely separated in time and
space serve as examples, the first perhaps most controversially.
The 3–3.5ma old locality AL-333/333w at Hadar (Ethiopia) lies on a steep hill
slope, and yielded >200 hominin fossils, representing nine adults, two ­juveniles
and two infants (MNI=13) assigned to Australopithecus afarensis within a small
area (Aronson & Taieb 1981; Johanson et al. 1982). These stand out against a poor
background of mammalian fauna at the site, and seem to have been c­ overed by
sediments fairly rapidly. The lack of palaeontology suggests that there was ­little
activity in this point of the landscape. The site stands out from other Hadar
localities as it is, as Johanson and Shreeve (1989: 87) note, “…just hominids
littering a hillside”. The question as to how the hominin accumulation formed has
attracted considerable debate. A dynamic event such as a flood can be ruled out
on sedimentological grounds, and lack of carnivore modifications of the bones
rules out predation; furthermore it is difficult to see how an entire group could
 Chapter 8. The living as symbols, the dead as symbols 

become bogged down on a wet plain to die together on a hill. To my knowledge,

no one has advanced a hypothesis that sees australopithecines as the active agents
of accumulation. I propose that at least thirteen dead individuals came to lie on
the hill within a short space of time, because they had been deliberately placed
there by their conspecifics. The locale seems to have been a relatively quiet area
on a dynamic and dangerous landscape, perhaps given simple meaning as bodies
could be placed in the long grass, minimizing the possibility that carnivores would
scavenge from them. One needs invoke no specific meaning to this, further than
the desire to protect corpses from scavenging, or even just to remove them from
sight, but it is easy to see how, at some cognitive stage in hominin evolution, such
places might begin to acquire meaning, and in such a case one might see this as
relatively simple symbolism.
Secondly, the accumulation of the (complete) bodies of at least 32 individuals
assigned to Homo heidelbergensis in the Sima de los Huesos (‘Pit of the Bones’) at
Atapuerca, Spain, offers the earliest intriguing example of the use of a particular
place for mortuary disposal. Between 400–500 ka BP thousands of bones accumu-
lated in the 13m deep pit, mainly comprising bears (Ursus deningeri MNI=166),
several felids and canids, and the hominins, the latter heavily skewed towards
prime adults (Arsuaga et al. 1997; Bischoff et al. 2003). Lack of decent degrees of
carnivore gnawing show that they were not responsible for the deposition of the
hominin bodies, and degrees of articulation, and lack of damage and considerable
mixing of the hominin bones, suggest that they are either in situ or have not moved
far. The consensus seems to be that they were deliberately placed here, perhaps at
the top of the shaft which may have been open to the air at the time (Arsuaga et al.
1997, 2003). The lack of any archaeology in the pit save for one Acheulian biface
(ibid.) suggests, further, a non-prosaic nature of the accumulation. It is difficult to
see this accumulation as anything other than the deliberate caching of the dead
at this one place, and if this were so, then it must have been given specific mean-
ing as a place of the dead, another example of simple association of a place in the
landscape with death.
Finally, a parsimonious reading of the Middle Palaeolithic record shows that
between 30–40 simple inhumations of Homo neandertalensis are known, with more
inclusive estimates approaching 60 (see Pettitt 2002 and references therein). These
burials, all without the inclusion of grave goods, span the period from 80–34 ka
BP (possibly a little earlier), and overlap with the dates for the earliest burials of
Homo sapiens at Qafzeh and Skhul in Israel and Mungo, Australia. Both young
and old Neandertals were buried, and examples fall into distinct regional groups,
notably SW France, Germany, and the Levant. Given the relatively low number of
burials known despite a rich Middle Palaeolithic record, one should not simply
conclude from this that ‘Neandertals buried their dead’; it may be more apposite
 Paul Pettitt

to conclude that some Neandertals buried some of their dead, some of the time,
(i.e. what we might call cultural variation). While it is unclear whether these simple
inhumations were emplaced for prosaic reasons, the use of sites for multiple burials
and the possible use of grave markers might suggest that some underlying belief
accounts for the burials. The representation of multiple Neandertals among frag-
mentary remains at several sites is intriguing: at least 25 individuals at Krapina; 20
at L’Hortus, France, among which young adults dominate; at least 22 at La Quina,
France; seven at La Ferrassie of which two are juveniles and three foeti/neonates;
at least seven at Shanidar cave, Iraq; and two in the Feldhoffer cave in the Neander
Valley; and in Amud and Tabun caves, Israel. At La Ferrassie, several of the grave
pits, those of children, seem to have been covered with large boulders, one of which
bore ‘cup marks’ (Peyrony 1934). It is tempting to view the latter as specific grave
markers, and if they are, they are at least simple symbols (message: the dead lie
below here). In this light, Dunbar’s inclusion of Neandertals into the ‘fourth level
intentionality’ may further support the notion that by the Late Middle Palaeolithic,
at least an incipient ritual had emerged.
Three points in time, showing widely different funerary practices, which
might have operated at different points on the following example scale:

1. Simple (non-symbolic) observation: little activity beyond morbidity; investiga-

tion of the corpse, establishment that it is dead, and renegotiation of society
now that a member has dropped out: (‘It is dead, I am confused’).
2. Emotive (non-symbolic) interaction: the living interact with the dead; their
emotional response affects certain simple behaviours of disposal. (‘It is dead,
I am mourning; hide the corpse away from activity’).
3. Associative (symbolic) interaction: the deceased is associated with a specific
­activity at a specific place; the place symbolises the dead. (‘He is dead; he must
be ­disposed of at a recognized place’).
4. Time/Space-factored associative interaction: the agency of the dead is recog-
nized in mortuary treatment (who gets special treatment, where and when)
and mortuary activity is organized in time and space according to social rules.
(‘He is dead; he was an elder in life and has earned the right to be buried at the
place of the elders’).

One must of course remember that, presumably, most human mortuary activity
in prehistory is invisible to archaeology, but this at least shows that even for
archaeologically observable mortuary activity one cannot simply argue that it
is ‘symbolic’ in any straightforward way. I propose that simple observation and
morbidity has very deep evolutionary roots (at least back to Miocene apes); that
emotive interaction with the dead might have originated in the earliest hominin
 Chapter 8. The living as symbols, the dead as symbols 

communities; that associative interaction with the dead (and thus a degree of
symbolism) originated at least among Homo heidelberbergensis populations in
the Middle Pleistocene and became more elaborate with Homo neandertalensis
and early Homo sapiens, but that the earliest true time/space-factored associative
interaction can as yet only be recognized among European Mid Upper Palaeolithic
(Gravettian) burials, which represent a highly-redundant (and often pathological)
sub-section of society in which social differentiation seems to have been one of the
criteria governing the disposal of the dead (e.g. Zilhão & Trinkaus 2002; Zilhão
2003; Pettitt 2006; Formicola 2007).

5. The evolution of Homo symbolicus: Gradual, abrupt, or fragmentary?

Despite progressivist narratives the story of hominin evolution is in a sense largely

one of failure: multiple dispersals from Africa (and one assumes elsewhere) when
climatic and environmental circumstances allowed, most of which resulted in
local extinctions as environments shut down in response to climatic downturns
associated with Heinrich events. In the long-term context, as discussed above,
behavioural repertoires of the MSA and Middle Palaeolithic waxed and waned,
were situationally dependent, and, beyond drawing upon general repertoires,
regionally independent (see Chase 1999, 2006 and papers in Hovers & Kuhn 2006).
Such attenuated and regionally-differing dispersal and behavioural ­trajectories
provide an appropriate context for the evolution of the symbolic capacity and
its expression, which itself, I argue, should have had a recursive and interrupted
developmental pattern. While I have concentrated upon specific examples to try
to deconstruct what archaeologists mean by ‘symbolism’, a general evolutionary
context has been provided by Dunbar, into which, I suggest, symbolic evolution
might fit.
Dunbar (2003) has interpreted brain evolution in terms of intentional
states – reflexive sequences of belief states – which range from one (‘I believe
that…’), through typical human functioning of three, to the normal human limit
of four. Although it became apparent in the Homo symbolicus workshop that there
is c­ onsiderable debate as to whether one can pigeonhole intentional states in this
way, and whether such a classificatory system is of heuristic use for cognitive
evolution, it at least forms a useful framework within which to conceptualise
symbolic evolution. To Dunbar, Theory of mind, which in modern humans
emerges between 4–5 years of age, requires level 2 intention (‘I believe that you
believe…’). Requiring individuals to conform to social norms requires three
levels of intention (‘I want you to believe that you must behave how we want’),
whereas religion, at least as we conceive of it, requires level four intention (‘I have
 Paul Pettitt

to believe that you suppose that there are supernatural beings who understand that
you and I desire that things ­happen in a certain way’). Dunbar has suggested that
levels of intentionality increased over the course of hominin evolution, and can
be equated with increasing brain size, group size and grooming time. This would
grant australopithecines approaching two levels of intentionality (thus a theory of
mind), archaic Homo such as Homo erectus and Homo heidelbergensis three levels,
and four levels to Neandertals and anatomically modern humans. In light of the
latter it is perhaps not surprising that it is with both Neandertals and modern
humans that burial of the dead was from time to time practised. Might Dunbar’s
conclusions be relevant to the origins of symbolism?
While systems of decoration and enhancement I introduced above could
function with two levels of intentionality (‘I know that you will be impressed’) ‘true’
symbolism should require three (for a full symbol to work ‘I need to know that you
understand this symbol/place means this’). By extension, however, giving symbolic
integration to people and places requires four (‘I need to know, that you under-
stand, that this place gives meaning to this thing/person/act’). It follows that the
evolution of the symbolic capacity (although not always accompanied by sym-
bolic expression of that capacity) should have paralleled cognitive evolution, and
it is the task of the archaeologist to identify the explicit way in which this might
have occurred. In Table 1, I try to place the examples discussed above in Dunbar’s
­context of cognitive evolution.

Table 1. Scales of the evolution of symbolism using Dunbar’s (2003) concept of evolution
of the social brain and personal ornamentation and mortuary activity as examples
Intentional Hominin Dunbar’s Personal Mortuary activity
level grade example ornamentation
1 Pre- I believe that… Decoration Simple
australopithecines observation
& (I believe that
australopithecines you are dead)
2 Australopithecines I believe that Enhancement (I Emotive
you believe… know you will be interaction
impressed) (I empathise
that you are dead)
& simple mortuary
caching
3 Archaic Homo I want you to True symbolism Associative
believe that (I know that symbolic
you must you understand caching (I know
behave how we that this symbol you must be
want means this…) deposited at a
specific place)

(Continued)
 Chapter 8. The living as symbols, the dead as symbols 

Table 1. (Continued)
Intentional Hominin Dunbar’s Personal Mortuary activity
level grade example ornamentation
4 Homo I have to Time/space- Time/space-
neandertalensis, believe that factored factored
Homo sapiens you suppose symbolism (I associative
that there are know, that you symbolism
supernatural understand, that (Because of your
beings who this place gives agency, you must
understand meaning to this be disposed of
that you and thing/person/act) in this way, by
I desire that this method,
things happen at this place, as
in a certain recognised by our
way social rules)

The difficulty, of course, will be to develop specific predictions of the

archaeological record that we have a chance of addressing. With an archaeological
record that is still overwhelmingly poor, particularly for Africa, and largely non-
existent for much of the Old World, all we can do for now is to adopt parsimony
in our interpretations. Thus I suggest that, in situations where we have only
pigment fragments, we need only infer decoration or enhancement; with simple
burials perhaps only indicative of emotive interactions with the dead. As symbolic
systems are elaborated, and in particular when they are employed in combination,
one might reasonably infer more sophisticated forms of symbolism. I try to
outline a set of archaeological predictions based on this notion of parsimonious
interpretation in Table 2.

Table 2. Potential archaeological signatures of developing levels of symbolism, based on

various combinations of pigment use, engraving, personal ornamentation, burial and art
Pigments Personal Figurative Burial Parsimonious Chronology
ornamentation art symbolic
function
Ochre Decoration Intermittent
fragments/ from the
processing Middle
Pleistocene
Personal Decoration Intermittent
ornamentation from >100 ka
Personal Enhancement, Intermittent
ornamentation possibly from >100 ka
(selection of accessorisation
restricted shell
taxa & colouring
by burning)
(Continued)
 Paul Pettitt

Table 2. Potential archaeological signatures of developing levels of symbolism, based on

various combinations of pigment use, engraving, personal ornamentation, burial and art
(Continued)

Pigments Personal Figurative Burial Parsimonious Chronology

ornamentation art symbolic
function
Simple Emotive Intermittent
inhumation interaction (probably
or rare) from
deposition >100 ka
of the body
Ochre Enhancement Intermittent
fragments/ or from >100 ka
processing, accessorisation
selection
of certain
colours/
saturated
hues
Ochre Personal Combination Intermittent
fragments ornamentation suggests from >100 ka
likelihood of
enhancement or
accessorisation
Multiple Associative Intermittent
inhumation interaction from >100 ka
Ochre Personal Accessorisation Intermittent
fragments ornamentation or full from >100 ka
with symbolism
engraved
traditions
Multiple Time/space- Intermittent,
inhumation factored probably only
& adjunct associative from ~30 ka
material interaction
culture (e.g.
pigments,
grave
goods)
Ochre Personal Art Full space/ Intermittent,
fragments ornamentation time- factored probably only
with symbolism from ~35 ka
engraved
traditions

One cannot, of course, rule out that sophisticated symbolic systems might
underlie simple manifestations of the data of concern. After all, they often do in
 Chapter 8. The living as symbols, the dead as symbols 

the modern world, although one can at least use such an heuristic to evaluate
observable levels of symbolic sophistication, if not the symbolic capacity itself.
From Tables 1 and 2, I would infer that Homo heidelbergensis was capable of
employing emotive interaction with the dead and perhaps decorative interaction
with the l­ iving; Homo neandertalensis and early populations of Homo sapiens were
capable of associative interaction with the dead and decorative, enhanced and
accessorized interaction with the living; and only later populations of Homo sapiens
were capable of full symbolic interaction with the living (perhaps intermittently
after 80 ka BP); and that full time/space-factored interaction with both the living
and the dead emerged relatively late, that is, after 35 ka BP and was even then
intermittent, probably for the duration of the Palaeolithic.

100 kyr
Homo sapiens
90 kyr Homo
neandertalensis
80 kyr

70 kyr

60 kyr

50 kyr ?
B NE
40 kyr

30 kyr B Eur PO Euo

20 kyr
PO B B Eur
NE

Figure 1. Manifestation of personal ornamentation (PO) and burials (Near East: B NE and
Europe: B Eu) for Homo neandertalensis and Homo sapiens. Black bars indicate presence;
open bars indicate absence during the known presence of the species in the region

As the cognitive and behavioural capacities of Homo sapiens and Homo

neandertalensis are major research interests for current palaeoanthropology, it is
worth comparing the two in terms of some of the symbolic behaviours discussed
above. Figure 1 shows the presence or absence of personal ornamentation and burial
for each taxon, further divided into two regions, the Near East and Europe, and
spanning the period 100–10 ka BP. The use of continual bars to denote the presence
of these phenomena contains considerable imprecision of dating methods, and thus
should not necessarily be taken to represent continuity within these periods. The
figure is simply a very coarse reflection of the current state of knowledge (see Pettitt
 Paul Pettitt

2011 for a more contextualised discussion). A few very preliminary observations can
be made from this. First, for Homo sapiens, there is a correlation between the presence/
absence of burials in both Europe and the Near East and personal ornamentation;
i.e. in periods where burial was practised, personal ornamentation was produced
too, and this pattern holds for populations in both Europe and the Near East. The
picture for the Neandertals is not as clear, although Near Eastern burials appear
around the same time as those in Europe, but are truncated earlier, almost certainly
because Neandertals became extinct in the region earlier than they did in Europe. A
clear contrast with modern humans is the lack of personal ornamentation, the sole
exception being the jewellery from the Grotte du Renne at Arcy-sur-Cure, assuming
it was made by Neandertals. A final observation is that there is an inverse correlation
between the two taxa: in the period that some Neandertals were burying some of
their dead in Europe and the Near East, we have no examples for Homo sapiens,
who are instead practising burial and personal ornamentation before and after the
Neandertal practise of these phenomena. One would not yet want to place too much
emphasis on this relatively poor record, but such regional trajectories may repay the
effort of further study if and when the record improves.

6. C
 onclusion

This is by necessity a speculative paper, although I draw on the current state of

knowledge for archaeological phenomena that are usually thought to be indicative
of symbolic thought. It may be full of holes: the data I draw upon as examples are
of course open to other interpretations; my reading of the existing archaeological
record may be overly-critical, and the capacity for ‘symbolism’ may be heuristically
divided in many other ways. It is also certain that the archaeological record will
change as new discoveries are made, particularly in Africa and Asia. I hope at
least, however, that the paper will stimulate discussion among Palaeolithic
archaeologists as to how to develop our concepts of symbolism; after two decades
of debate in which symbolism has emerged as the human capacity, we still use
rudimentary concepts of what it is and how it is to be recognised archaeologically.
These discussions began at the Cape Town workshop and, I hope, will continue to
run. To be successful we have to problematise what we mean by symbolism. I’ve
made a start here and throw down a friendly gauntlet.

Acknowledgements

I am much indebted to The Templeton Foundation, particularly Mary Ann Meyers, for providing
the support for the Homo symbolicus symposium; to Christopher Henshilwood for being such a
 Chapter 8. The living as symbols, the dead as symbols 

pleasant and welcoming host, and Karen van Niekerk for her friendly organisation of the event.
Christopher Henshilwood and Francesco d’Errico very kindly gave up time to show us Blombos
Cave. The symposium provided a splendid opportunity to debate many issues ‘out of the box’
in pleasant surroundings with colleagues from very different disciplines. I very much benefited
from discussion with all of the participants.

References

Aronson, J.L. & Taieb, M. 1981. “Geology and palaeography of the Hadar hominid site,
Ethiopia.” In Hominid Sites: Their Geologic Settings, G. Rapp & C. F. Vondra (eds), 165–195.
Westview: Boulder.
Arsuaga, J.L., Martínez, I., Gracia, A., Carretero, J.M., Lorenzo, C. & García, N. 1997. “Sima
de los Huesos (Sierra de Atapuerca, Spain), the site.” Journal of Human Evolution
3: 109–127.
Arsuaga, J.L., Carbonell, E., & Bermúdez de Castro, J.M. 2003. The First Europeans: Treasures
from the Hills of Atapuerca. New York: American Museum of Natural History/Junta de
Castilla y León.
Barham, L. 1998. “Possible early pigment use in south-central Africa.” Current Anthropology
39: 703–710.
Bednarik, R. 2003. “A figurine from the African Acheulian.” Current Anthropology 44 (3): 405–412.
Bischoff, J.L., Shamp, D.D., Aramburu, A., Arsuaga, J.L., Carbonell, E. & Bermudez de Castro, J.M.
2003. “The Sima de los Huesos hominids date to beyond U/Th equilibrium (>350Ka) and
perhaps to 400–500 ka: new radiometric dates.” Journal of Archaeological Science 30: 275–280.
Boyer, P. 2008. “Religion: bound to believe?” Nature 455: 1038–1039.
Bouzouggar, A., Barton, N., Vanhaeren, M., d’Errico, F., Collcutt, S., Higham, T., Hodge,
E., Parfitt, S., Rhones, E., Schwenninger, J.-L., Stringer, C., Turner, E., Ward, S., Moutmir, A.
& Stambouli, A. 2007. “82,000 year-old shell beads from North Africa and implications for
the origins of modern human behavior.” Proceedings of the National Academy of Sciences.
104 (24): 9964–9969.
Brooks, A. & Yellen, J. 2009. “Was Homo symbolicus also Homo sapiens? A palaeoanthropologi-
cal perspective.” Pre-circulated paper for the Homo symbolicus symposium, Cape Town.
Chase, P.G. 1999. “Symbolism as reference and symbolism as culture.” In The Evolution of ­Culture,
C. Power, R. Dunbar & C. Knight (eds), 34–49. Edinburgh: Edinburgh ­University Press.
Chase, P.G. 2006. The Emergence of Culture: the Evolution of a Uniquely Human Way of Life. New
York: Springer.
Dennell, R. 2008. The Palaeolithic Settlement of Asia. Cambridge: Cambridge University Press.
Dennell, R. & Roebroeks, W. 2005. “An Asian perspective on early human dispersal from Africa.”
Nature 438: 1099–1104.
d’Errico, F. 2003. “The invisible frontier: a multiple species model for the origin of behavioural
modernity.” Evolutionary Anthropology 12: 188–202.
d’Errico, F. & Nowell, A. 2000. “A new look at the Berrekhat Ram figurine: implications for the
origins of symbolism.” Cambridge Archaeological Journal 10 (1): 123–167.
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 multidisci-
plinary perspective.” Journal of World Prehistory 17 (1): 1–70.
 Paul Pettitt

d’Errico, F. & Henshilwood, C. this volume. “The origin of symbolically mediated behaviour:
from antagonistic scenarios to a unified research strategy.” In ‘Homo Symbolicus’: The Dawn
of Language, Imagination, and Spirituality, C. Henshilwood and F. d’Errico (eds), 49–73.
Amsterdam: John Benjamins.
Dunbar, R.I.M. 2003. “The social brain: mind, language, and society in evolutionary p ­ erspective.”
Annual Review of Anthropology 32: 163–181.
Formicola, V. 2007. “From the Sunghir children to the Romito dwarf: aspects of the Upper
­Palaeolithic funerary landscape.” Current Anthropology 48 (3): 446–453.
Goodall, J. 1986. The Chimpanzees of Gombe: Patterns of Behaviour. Cambridge MA: Harvard
University Press.
Grün, R., Stringer, C., McDermott, F., Nathan, R., Porat, M., Robertson, S., Taylor, L., Mortimer,
G., Eggins, S. & McCulloch, M. 2005. “U-series and ESR analyses of bones and teeth
relating to the human burials from Skhul.” Journal of Human Evolution 49 (3): 316–334.
Henshilwood, C. & Marean, C. 2003. “The origin of modern human behaviour: critique of the
models and their test implications.” Current Anthropology 44 (5): 627–651.
Henshilwood, C.S., d’Errico, F., Vanhaeren, M., van Niekerk, K. & Jacobs, Z. 2004. “Middle
Stone Age shell beads from South Africa.” Science 304: 404.
Henshilwood, C.S., d’Errico, F. & Watts. I. 2009. “Engraved ochres from the Middle Stone Age
levels at Blombos Cave.” South Africa Journal of Human Evolution 57: 27–47.
Henshilwood, C. & d’Errico, F. this volume. “Middle Stone Age engravings and their ­significance
to the debate on the emergence of symbolic material culture.” In ‘Homo Symbolicus’: The
Dawn of Language, Imagination, and Spirituality, C. Henshilwood and F. d’Errico (eds),
75–96. Amsterdam: John Benjamins.
Hovers, E., Ilani, S., Bar-Yosef, O. & Vandermeersch, B. 2003. “An early case of colour symbolism:
ochre use by modern humans in Qafzeh Cave.” Current Anthropology 44 (4): 491–522.
Hovers, E. & Kuhn, S. 2006. Transitions before the Transition: Evolution and Stability in the
­Middle Palaeolithic and Middle Stone Age. New York: Springer.
Johanson, D.C., Taieb, M. & Coppens, Y. 1982. “Pliocene hominids from the Hadar formation,
Ehtiopia (1973–1977): stratigraphic, chonological, and paleoenvironmental contexts, with
notes on hominid morphology and systematics.” American Journal of Physical ­Anthropology
57: 373–402.
Johanson. D. & Shreeve J. 1989. Lucy’s Child, The Discovery of a Human Ancestor. New York:
Early Man Publishing, Inc.
Kuhn, S.L., Stiner, M.C., Reese, D.S. & Güleç, E. 2001. “Ornaments of the earliest Upper
­Paleolithic: New insights from the Levant.” Proceedings of the National Academy of Sciences.
98 (13): 7641–7646.
Marquet, J.-C. & Lorblanchet, M. 2003. “A Neanderthal face? The proto-figurine from La
­Roche-Cotard, Langeais (Indre-et-Loire, France).” Antiquity 77: 661–670.
Matsuzawa, T. 2003. Jokro: the Death of An Infant Chimpanzee (DVD film with associated
­leaflet). Kyoto: Primate Research Insititute.
McBrearty, S. & Brooks, A. 2000. “The revolution that wasn’t: a new interpretation of the o ­ rigin
of modern human behaviour.” Journal of Human Evolution 39: 453–563.
McGrew, W. this volume. “Pan symbolicus: a cultural primatologist’s view.” In ‘Homo Symbolicus’:
The Dawn of Language, Imagination, and Spirituality, C. Henshilwood & F. d’Errico (eds)
1–12. Amsterdam: John Benjamins.
Mellars, P. 1991. “Cognitive changes and the emergence of modern humans in Europe.”
­Cambridge Archaeological Journal (1): 63–76.
 Chapter 8. The living as symbols, the dead as symbols 

Mellars, P. 1995. The Neanderthal Legacy: An Archaeological Perspective from Western Europe.
Princeton: Princeton University Press.
Pettitt, P.B. 2002. “The Neanderthal Dead: exploring mortuary variability in Middle P ­ alaeolithic
Eurasia.” Before Farming 2002/1 (http://www.waspress.co.uk/journals/beforefarming/
journal_20021).
Pettitt, P.B. 2006. “The living dead and the dead living: burials, figurines and social performance
in the European Mid Upper Palaeolithic.” In Social Archaeology of Funerary Remains,
R. Gowland & C. Knüsel (eds), 292–308. Oxford: OxBow.
Pettitt, P.B. 2007. “Trajectories before the transition and revolutions that were or were not (lead
review).” Journal of Human Evolution 53: 755–759.
Pettitt, P.B. 2011. The Palaeolithic Origins of Human Burial. Abingdon: Routledge.
Pettitt, P.B. in press. “Religion and ritual in the Lower and Middle Palaeolithic.” In The Oxford
Handbook of the Archaeology of Ritual and Religion, T. Insoll and R. Maclean (eds). Oxford:
Oxford University Press.
Peyrony, D. 1934. “La Ferrassie: Moustérien – Périgordien – Aurignacien.” Préhistoire 3: 1–92.
Schults, F.L. 2009. Homo symbolicus: science, theology and the limits of semiotic engagement.
Pre-circulated paper for the Homo symbolicus symposium, Cape Town.
Sloane Wilson, D. this volume. “The human major transition in relation to symbolic ­behaviour,
including language, imagination and spirituality.” In ‘Homo Symbolicus’: The Dawn of
­Language, Imagination, and Spirituality, C. Henshilwood and F. d’Errico (eds) 133–139.
­Amsterdam: John Benjamins.
Vanhaeren, M., d’Errico, F., Stringer, C., James, S.L., Todd, J.A. & Mienis, H.K. 2006. “Middle
Palaeolithic shell beads in Israel and Algeria.” Science 312: 1185–1188.
Wadley, L. 2001. “What is cultural modernity? A general view and a South African perspective
from Rose Cottage Cave.” Cambridge Archaeological Journal 11 (2): 201–221.
Wadley, L. 2005. “Putting ochre to the test: replication studies of adhesives that may have
been used for hafting tools in the Middle Stone Age.” Journal of Human Evolution 49:
587–601.
Wadley, L. this volume. “Complex Cognition Required for Compound Adhesive Manufacture
in the Middle Stone Age Implies Symbolic Capacity.” In ‘Homo Symbolicus’: The Dawn
of Language, Imagination, and Spirituality, C. Henshilwood & F. d’Errico (eds) 97–109.
Amsterdam: John Benjamins.
Watts, I. 1999. “The origin of symbolic culture.” In The Evolution of Culture, R. Dunbar,
C. Knight and C. Power (eds), 113–146. Edinburgh: Edinburgh University Press.
Zilhão, J. 2003. “Burial evidence for the social differentiation of age classes in the Early Upper
Palaeolithic.” In Comportements des Hommes du Paléolithique Moyen et Supérieur en
Europe: Territoires et Milieux, D. Vialou, J. Renault-Miskovsky and M. Patou-Mathis (eds),
231–241. Liège: ERAUL 111.
Zilhão, J. & Trinkaus, E. 2002. “Social implications.” In Portrait of the Artist as a Child: the
­Gravettian Human Skeleton from the Abrigo do Lagar Velho and its Archaeological ­Context,
J. Zilhão & E. Trinkaus (eds), 519–541. Lisbon: Ministério da Cultura and Instituto
­Português de Arquelogia.
 chapter 9

Biology and mechanisms related

to the dawn of language
George F.R. Ellis
Mathematics Department, University of Cape Town

The purpose of language is prediction and problem solving in the social

and ecological context. Its implementation is enabled by the combination of
bottom-up and top-down causation in the brain. Visual thinking develops basic
abilities of pattern recognition and classification, but music and play can more
effectively develop the key feature of recursion. The motivational impulse for
language use starts with mother-child nurturing but moves on to the joy of
imaginative play, which develops theories of other minds and basic symbolic
abilities; it also easily includes music and song. Key steps towards language
could be changes in neural connectivity allowing recursion to emerge, in the
context of developing technological needs.

1. The functional and structural context

The emergence of language is partly a socially and environmentally based

­phenomenon, but equally it has a biological base. As well as social and functional
needs, evolutionary arguments should take this link to biology into account.
In biological terms, structure enables function, and hence study of language
­evolution must consider the neurological structure of the brain in relation to
how it ­functions. Of course the nature of language itself is also central to this
­analysis, so key questions are what are the crucial features of language that enable
its e­ xtraordinary representative and communicative abilities? And how are they
realised in neurological terms?
Like all truly complex systems, language has both a modular hierarchical
structure (Simon 2001; Booch 2007) and an evolutionary origin (Christiansen &
Kirby 2005). It is clear that social development requires language: without
communication society cannot exist. As social cooperation greatly increases
survival prospects, evolutionary processes will favour capacities that lead to
language development (Burling 2007). The purpose of it all is prediction and
problem ­solving in the social and ecological context. The functional basis is the
embodied brain and its biological context. Its mode is via the nature of semiotic
 George F.R. Ellis

representation. The key functional element in representing complex aspects

of reality is a modular hierarchical structure, with a class structure involving
inheritance and the capacity for recursion. It is realised through grammar that is
usage-based and embodied (Bergen & Chang 2003; Feldman 2006), rather than
being rule-based.
The crucial biological feature whose emergence enabled the flowering of
human language, separating us from all other animal species, is still unclear. A
key question here is, was the advent of language essentially due to some change
in brain structure, which then inevitably led to social changes? If so, what was
that change? (Striedter 2005). Or was it rather basically due to ecological and
social changes, which then led to changes in brain structure? If so, what were
these key changes?
This chapter first looks at foundational issues underlying the discussion,
and then at some evolutionary issues and possible archaeological implications. It
should be emphasized that this is to some extent an incursion of mine into another
field than my primary academic discipline; so the reader is asked to excuse some
infelicities in presentation in exchange for some provocative ideas that may be
productive.

2. The nature of language: Crucial features

Language has crucial features related both to its semiotic function and to its
embodied nature.

2.1 An embodied symbolic system

Language is a symbolic system (Deacon 1997) with a semiotic function (Trask
2007): its purpose is to convey meaning, emotions, facts, and concepts in a social
context through systematic use of symbols. It represents the world of objects,
actions, feelings, and qualities, as well as relationships, ideas, and theories. This
representational function involves naming, indexing, and use of metaphor (Trask
2007; Lakoff & Johnson 1980). Facts represented are both contingent (historical,
geographical, and other specific features of the world and of narratives) and
generic (universal patterns characterising the way it all works in general). The
relation between these two features (concrete/specific and abstract/generic) is a
key aspect of thought and of language, involving development of classes of entities
and classification of specific instances.
While language has an abstract character, it is embodied via an equivalence
class of physical representations; in particular it has spoken and written forms.
Its existence enables the cumulative building up of understandings and ideas in
 Chapter 9. Biology and mechanisms related to the dawn of language 

individuals and in society, and (through technological means) their long-distance

sharing in geographic terms as well as their storage and preservation in various
media, enabling their propagation over long times. This storage of ideas in external
media (ranging from diaries and memos to encyclopaedias and wikipedia) greatly
facilitates our mental powers and underlies an exponential growth of knowledge
(Clark 2008). Language attains its social power through enabling mediation of
social interactions on a small scale, and through enabling the utility of mass
media and communication systems on a large scale. These enable widespread
dissemination of facts, ideas, and meanings, extending the cognitive web from
local communication to a global system. Minds therefore cannot be understood
on their own (Donald 2001): they are part of a society that is in turn part of a
global intercommunication network.

2.2 Equivalence class of representations and embodiment

Physical realisation of language can be neural (in an individual’s brain), spoken
(sound), written (visual), electronic (digital), or in visually transmitted sign
patterns (sign languages). The same structural patterns are embodied in these
different representations. They are all enabled by the physical structure of the
brain, which is hierarchically structured so as to enable an interplay of sensory
interpretation and prediction, based on pattern recognition, classification,
memory, and extrapolation (Hawkins & Blakeslee 2004). Meaning is embodied
in an equivalence class of such surface representations: it is independent of
whether language is spoken, written or signed, and of language family, dialect/
pronunciation, and font. A profound ability of the mind, underlying the flexibility
of language usage, is to recognize them all as functionally equivalent. The concepts
represented are recognized as entities that exist in their own right, which can be
labeled and represented in many different ways.

2.3 Key features of language

The major function of language is its labelling of specific and generic objects and
instances, as well as abstract entities, through use of indices and symbols. Via
recursion (Deacon 2003; Trask 2007), this referential and representational nature
can allow reference to itself, and hence disjunction from physical referents. Related
key design features of language are (Trask 2007):
Arbitrariness: the absence of any necessary connection between the form of a word
and its meaning. This uncoupling of the signifier from the signified is c­ rucial to
representing objects symbolically (Striedter 2005). It allows equivalence classes of
representations to exist: the same meaning can be expressed in different ­symbolic
forms and systems.
 George F.R. Ellis

Duality of patterning: a type of structure, encoded in a grammar with a ­particular

syntactical structure, in which a small number of meaningless units are combined
to produce a large number of meaningful (semantic) units. This ­introduces the
crucial feature of discreteness of structural units (the quantum principle), which
can then be repeatedly combined as recognised and named higher level units, giv-
ing hierarchy and allowing arbitrary complexity of combinations to be built up
through recursion. Its constituent structure is based on this principle (Trask 1999).
Stimulus-freedom: our ability to say anything at all in any situation, so enabling
discourse that is freed from the immediate situation and stimuli. This enables us
to think off-line, i.e. without having to immediately act on what is thought about;
hence it enables us to reflect on the past and plan for the future (Bickerton 2001).
This facilitates the crucial feature of suppressing stimulus bound behaviours and
replacing them with less determinate voluntary acts (Striedter 2005).
Displacement: the ability to speak about things other than here and now: the future,
the past, the possible, even the impossible. This enables us to analyse t­heoretical
situations and so indulge in imagination and abstract analysis based in past mem-
ories and future possibilities.
Open-endedness: the ability of language to say new things, virtually without limit.
This enables language to be a vehicle whereby creativity can emerge.
Redundancy: the full message is entailed by part of the given text/message, hence
one can determine the full message by partial information (if the context is
known). This enables effective communication: we can understand the message in
the presence of distortion or noise, and can predict what is meant on the basis of
only a part of the text – an essential aspect of listening and reading.
A key feature of the way language functions is the use of metaphor, which
plays a major role in cognition and meaning-making (Lakoff & Johnson 1980).
This is enabled by the features listed above, and takes place in the context of
­conceptual schemas and cultural frames (Feldman 2006), which are the context of
our understandings.

3. Hierarchical structuring

These features in turn are based in the functional structure of language, enabled by
the physical structuring of the brain.

3.1 Functional structure

Language has a modular hierarchical structure (Booch 2007) that enables its com-
pletely flexible representational and social function. This structure is bound by
 Chapter 9. Biology and mechanisms related to the dawn of language 

strict semiotic requirements (Deacon 2003), leading to a necessary set of implicit

rules, but with a great variety of possible realisations (different languages/dialects).
Its hierarchical character is enabled by a class structure with inheritance, embody-
ing the crucial feature of recursion (the occurrence in a sentence of a syntactic
category containing within it a smaller version of the same category, based on
chunking and labelling). “The recognition of a suitable set of syntactic categories
allows us to analyse all the sentences of a language as being built up, by means of
a fairly small set of rules allowing recursion, from just these few categories” (Trask
1999: 288). This is a very general way of handling complexity: recursion occurs
whenever you break up a complex task into simpler tasks that are completed first
(Hofstadter 1980).
Reflecting and enabling the modular hierarchical structure of language, the
brain also has a modular hierarchical structure (Beer 1981; Hawkins 2004), in
turn based in the hierarchical structure of its physical constituents (Scott 1995),
(see Table 1; for a more detailed description of this hierarchical structure, see
http://www.mth.uct.ac.za/~ellis/cos0.html). Major issues arise in the linkages
between different scales in the brain – the relation of molecules to neuronal
activity; of neuronal connectivity to brain function; of brain regions to cognitive
function; of the brain as a whole to individual psychology; and of the individ-
ual to the social and ecological environment. A further major issue is how this
physical hierarchical structure underlies and enables the hierarchical functional
structure of language. This is part of the larger issue: how does brain structure
underlie the complex functioning of the mind? (Koch 2004).

Table 1. The hierarchy of structure and causation. A simplified representation

of the hierarchy of levels of causation in human beings, in terms of the corresponding
academic subjects
Level 8 Sociology/Economics/Politics
Level 7 Psychology
Level 6 Physiology
Level 5 Cell biology
Level 4 Biochemistry
Level 3 Chemistry
Level 2 Atomic Physics
Level 1 Particle physics

3.2 Bottom-up and top-down causation

Emergent behaviour is enabled by the combination of bottom-up and top-down
causation in the hierarchy of complexity and causation (Ellis 2008). This feature
 George F.R. Ellis

occurs at all levels; in particular, between micro and macro aspects of the brain
itself, in the relation of the brain to psychology, and in the relation of the individual
to society (note that this applies to all animal brains, not just human). It is enabled
by the existence of equivalence classes of lower level states that correspond to
single higher level states (Auletta et al. 2008).
The interaction between these two types of causation in the functioning of
the human cortex is emphasized by Feldman (2006) and Hawkins and Blakeslee
(2004), while the way it takes place as regards individuals in the context of society
is nicely illustrated in Berger (1963), Berger and Luckmann (1967), and Donald
(2001). Top-down causation has crucial effects both on causal relations, and on the
way we understand them. It is also a key feature in both language production and
understanding, where holding context in mind is crucial to understanding both
speech and writing (Smith 1976); key processes in language production, reception,
and learning are therefore top-down driven. Indeed the way meaning is embodied
in the hierarchical structure of language is context-dependent all the way down:
the individual units at each level (sentences, phrases, words, phonemes) only
attain their meaning and function, and even pronunciation, in the larger context
of the whole meaningful situation (Krashen & Terrell 1983; Goodman 2005).
The key element allowing new ideas and information to come into being
is adaptive selection, which is a specific form of top-down causation guided by
higher-level selection criteria (Ellis 2008) which enable multi-level selection
to take place (Okasha 2006). The top-down nature of this causation is crucial
to this process, enabling new kinds of entities to come into being. Indeed,
David Sloan Wilson (2011) remarks that the transition from bottom-up to top-
down dominated causation in the relation of mind to the society in which it
is imbedded is a major evolutionary transition in the historical development
of humanity, resulting in the emergence of the social order as a higher level
entity in its own right, and a consequent change in the nature of the evolutionary
processes at work. However, adaptive selection also occurs on functional and
developmental timescales.
Production: Speaking and Writing: Speaking requires the physical ability to talk
(vocal chords), and writing requires hands that can clasp and move delicately,
plus a bipedal posture that allows them the needed freedom to operate. These are
characteristically human features, although some animals do have some of these
attributes (Striedter 2005). But one also needs the ability to form thoughts in the
context of the current situation; this is where top-down causation occurs from
that overall context to detailed thought processes. One then requires the ability
to turn developing thoughts into a meaningful grammatical form, on the basis of
past patterns experienced and usage-based language skills. This involves top-down
causation from past experiences of linguistic patterns, as discussed below.
 Chapter 9. Biology and mechanisms related to the dawn of language 

Reception: Listening and Reading: Listening and reading are based on seeing and
hearing, which are very ancient attributes in the animal kingdom, exquisitely
developed in many higher animals as well as humans. Decoding syntax and
grammar involves top-down causation from past experiences of linguistic
patterns, similar to the processes involved in speech production. However,
listening and reading are essentially based in processes of understanding,
involving expectation and prediction in the face of partial information and
noise (Smith 1976). These expectations are based on current context and
memory, so this is also a form of top-down causation from those higher level
concepts, based on similar processes to those occurring in general cognition
(Hawkins & Blakeslee 2004).
Learning: Language is learnt by experiential processes in a meaningful
context (Smith 1976), supplemented by a combination of formal and informal
instruction. This learning is an individual process of experimental development:
­hypothesis formation, trial and error, and imitation takes place in a social ­context
of observation and experience. Like all human capacities, it is then embodied in
a set of hierarchically structured automatized skills that are constructed in and
constantly revised by consciousness (Donald 2001). This process is driven by the
need to understand and predict (an example of top-down causation), in the overall
context of the search for understanding and for meaning (Hawkins & Blakeslee
2004). Thus it is based – through experience and usage – firstly on intention-
reading and cultural learning; secondly on schematization and a­nalogy;
thirdly on entrenchment and competition; and fourthly on functionally based
distribution analysis (Tomasello 2003). This learning is enabled by Latent
Semantic Analysis (Landauer et al. 1998), resulting in an Embodied Construction
­Grammar (Bergen & Chang 2003; Feldman 2006). Spoken and written language
­abilities are developed by variants of the same experiential process: meaningful
­experiences drive the process in a top-down way (Smith 1976). Formal t­ eaching
is needed to supplement and systematise this learning process, but cannot
replace it.

4. Language modules and the development of language

The brain has developed through evolutionary processes that have led to the devel-
opment of the conscious mind and language.

4.1 Evolutionary development

Evolutionary processes apply both to living beings, and to culture (Richerson &
Boyd 2005; Mesoudi et al. 2006). Language develops over time through ­adaptation
 George F.R. Ellis

in a social context, occurring via variation and then selection of language f­ eatures
according to communication utility (Burling 2007). The key point here is that
according to standard genetic theory, nothing you understand or think ­influences
the genes you pass on to your progeny, for they are fixed the day you are born.
Hence any intellectual understandings you may develop cannot be directly
selected for in biological terms, as they do not influence your genetic inheritance
(although they do influence the likelihood of your passing on that unchanged
genetic inheritance).
Rather, genetically based development results in general purpose intelligence
(pattern recognition, classification, prediction) linked to memory (Hawkins &
Blakeslee 2004), together with basic emotional modules (Panksepp 1998) that are
effective in guiding the development in the individual of analytic ability in general,
and language ability in particular (Greenspan & Shanker 2004). This takes place
in his/her emotional context (Damasio 2000), with the social context crucially
shaping individual minds in terms of understanding and language (Berger &
Luckmann 1967; Donald 2001). That social context then reciprocally shapes
the development of language itself: a culturally based dual evolutionary process
(Richerson & Boyd 2005). This whole process favours genetic developments that
lead to the propensity to develop language, but cannot lead to specific genetically
determined language modules, for a variety of reasons I now briefly discuss.

4.2 Language modules

A large literature, largely based in linguistic analysis (for example, Pinker 1994),
suggests that language processing is based in genetically determined language
modules in the brain. Four problems make this extremely unlikely. First, it is
developmentally unlikely that detailed connections in the cortex can be genetically
determined, even though this does occur in other parts of the nervous system.
Second, even if this was possible, there is not enough genetic information avail-
able to carry out this task; essentially, this is a key finding of the human genome
project. Thirdly, as just mentioned, one can’t directly select DNA to promote such
connections because of the central dogma of molecular biology: the DNA one
passes on to ones progeny is not affected by any of one’s intellectual activities.
And fourthly, the kinds of specific issues involved in detailed language processing
would have to compete with numerous other factors affecting survival, and it is
highly unlikely these specific items would dominate over all the rest and so result
in genetic processes determining specific language modules, as originally envis-
aged by Chomsky.
An alternative view suggests language ability develops by conversion of
genetically shaped general-purpose brain domains and capacities to language use.
 Chapter 9. Biology and mechanisms related to the dawn of language 

The way language is developed then is through the process of usage-based language
acquisition (Tomasello 2003; Feldman 2006). As explained by these authors,
these processes undermine the Poverty of Stimulus argument: sufficient depth
of experience to carry out this process is provided by everyday life (Tomasello
2003). The universal underlying features of all languages (sometimes interpreted
as resulting from the deep structure of a generic Universal Grammar) can then
be understood as resulting from universal constrains intrinsic to their semiotic
function (Deacon 2003).
Thus it is highly plausible that functional modular structures of language
do not arise genetically, but rather have the restricted forms they do because of
semiotic constraints. Then what is genetically determined in each individual is not
specific language modules, but rather the capacity to develop language obeying
these constraints. Hence in searching for the dawn of language, we do not have
to explain genetically determined modules, but rather evolution of capacities that
lead to the development of language. This has two aspects: the requisite intellectual
capacities, and the needed emotional drivers; we discuss each below.

5. Patterns and symbols: Vision and music

What intellectual capacities and problem solving abilities underlie the develop-
ment of language, and what hints of their presence could we hope to find in the
archaeological record? Can we identify the intellectual capacities, and then pro-
pose other ways they may display their presence?

5.1 Pattern recognition and classification

A first such key feature is the recognition of spatial and temporal patterns (Burling
2007: 82–84), and then categorisation: classifying and naming them, as discussed
by Tomasello (2003), in the context of memory of past events and prediction of
future happenings (Hawkins & Blakeslee 2004). Developing some form of con-
ceptual structure is natural in higher animals (Burling 2007), indeed it is neces-
sary for their functioning and survival, so they form proto-propositions (Hurford
2007). Classification skills allow organisms to categorize different aspects of their
world into a manageable number of kinds of things and events (Tomasello 2003).
Humans are able to go on to verbally name objects and events (labelling them
symbolically), which develops naturally out of this process: the essence of lan-
guage is its symbolic dimension, with grammar being derivative (Tomasello 2003).
It is suggested by Hawkins and Blakeslee (2004) that the basic purpose of intel-
ligence is memory-based prediction: “the brain is not a computer, but a memory
 George F.R. Ellis

based system that stores experiences in a way that reflects the true structure of the
world, remembering sequences of events and their nested relationships and mak-
ing predictions based on those memories. It is this memory-prediction system
that forms the basis of intelligence, perception, creativity, and even consciousness”.
The use of symbolism and language crucially helps these purposes, their utility
being based in the core features of language discussed above. These abilities are
based in the nature of local brain connections adapted to prediction and problem
solving. How the core process of naming is implemented via neural connections is
discussed by Hawkins and Blakeslee (2004); this is the basis of symbolic systems.
A key question is why these kinds of inter-level connections came into being in the
cortex. Understanding experiences, as mediated by the various senses, must have
been the core driver, and in particular, interpreting visual images in conjunction
with sounds may have been crucial in leading to their development.

5.2 Visual thinking

Vision is one of the first detailed sensory capacities to have been developed,
and is so effective and important that there have been various independent
origins of eyes (Conway Morris 2003; Parker 2004). It has been suggested in a
provocative analysis by Rudolf Arnheim (1969) that visual thinking is primary in
the development of intelligence, with abstract thinking being secondary to it. The
description of his book states: “He shows that even the fundamental processes of
vision involve mechanisms typical of reasoning … our perceptual response to the
world is the basic means by which we structure events, and from which we derive
ideas and therefore language”. He suggests the importance of the capacity to first
represent relationships geometrically, identifying them as geometrical patterns,
and then to express them abstractly, identifying them as abstract patterns. This is
the start of symbolism, when one conjoins the ability to recognise both individual
objects and classes of objects with the ability to associate some kind of symbol to
each class of objects. This implies that the starting point for symbolic thinking is
recognising patterns in space, hence the Blombos kind of pattern experiments are
a starting point in such understanding. Additionally, vision involves the same kind
of “filling in” of information that is a characteristic feature of listening and reading,
as emphasized above (Koch 2004; Purves 2010).
Hence vision strongly develops core pre-language abilities. However these
are shared with all the higher animals: it is not clear that human visual under-
standing is greatly different from that of other animals. It is true that symbols are
expressed in spatial or temporal relationships and patterns, and hence decoding
them involves recognising such patterns; but it seems unlikely that such pattern
recognition in humans is beyond that of many other higher animals, who through
 Chapter 9. Biology and mechanisms related to the dawn of language 

categorical perception can acquire a wide range of abstract concepts (Hauser et al.
2002). The progression from pictorial representation through icons to hierarchi-
cally structured recursive symbol systems (see Figure 1) is, in intellectual terms,
a natural progression (Burling 2007). Nevertheless visual thinking per se only
weakly, if at all, develops the key feature of recursion: it only develops when writ-
ing develops, long after spoken language has evolved. While fractal patterns occur
in nature and in some art, such as Escher’s work, they are not a central feature of
vision and visual understanding. Recursion is possible in sign language, but again
is technically demanding and therefore not particularly natural and easy. Some-
thing else must be involved in addition to vision.

Pictorial representations → icon → indices and symbols

→ Hierarchically structured symbolic systems
→ Hierarchically structured recursive symbolic systems

Figure 1. Evolution of key characteristics of written language. A natural progression

occurs in visual symbolic systems (Burling 2007); a parallel process should occur in the
case of written words

5.3 Grammatical forms and music

Music and language each have complex hierarchical structures with an associated
syntax. Together they are universal in human societies, and are unique to our spe-
cies (Levitin 2007). They can be claimed to have interesting similarities in terms
of their “syntactic architecture” (Patel 2008: 267), and to share a number of basic
processing mechanisms with an overlap in the neural areas and operations that
provide the resources for syntactic integration (Patel 2008). Singing and instru-
mental playing help refine motor skills needed for speech production and writing.
Listening to music involves processes of expectation, prediction, and perceptual
completion (Levitin 2007), just as listening and reading do.
Thus music development would have encouraged the same intellectual abili-
ties and motor skills as are needed for language, and may have played a more
specific role in language development than vision. It would have had a deep effect
on evolutionary psychological development because of the strong emotional
power of music and dance, ensuring they played a major role in social bonding
and cohesion, which thereby justifies its evolutionary importance. In particular,
dance and music are associated with play, which is important for higher-level
integrative processes (Levitin 2007) and specifically in language development
(see the following section).
 George F.R. Ellis

Most important of all, music commonly entails recursive structures (Patel

2008), which are key also in the structure of language, in both cases allowing its
generative nature. The progression shown in Figure 1 is much more easily attained
in vocal than visual form, because hierarchy and recursion are naturally developed
through music and song.
Development of Recursion: A key feature in the development of language is the
emergence of recursion in symbolic systems (Deacon 2003), enabling abstract
thought patterns to emerge. “Recursion is pervasive in the grammars of the lan-
guages of the world, and its presence is the chief reason we are able to produce a
limitless variety of sentences of unbounded length just by combining the same few
building blocks” (Trask 1999: 244). Thus a crucial aspect of language is identify-
ing as a single unit and naming compound experiences or concepts, thus allowing
hierarchical structuring and recursion (building up patterns of patterns). This key
ability for language (Hauser et al. 2002; Fitch et al. 2005; Burling 2007) and com-
puter languages (Roberts 2009) must be based in specific aspects of local neural
connections, involving specific kinds of connections enabling naming procedures
to be applied to names themselves (perhaps developing specifically in Wernicke’s
area: see Ramachandran 2011: 188–189). This type of neural connectivity, presum-
ably involving links from higher levels of structure to lower levels, should be char-
acterisable in the same kind of way that Hawkins and Blakeslee (2004) identify the
neural bases of naming.
Consequently, one might envisage the key step in language development as
being the development of neural connections allowing recursion, either directly
as the result of some advantageous genetic mutation, or as a by-product of some
other-directed advantageous mutation (Hauser et al. 2002; Fitch et al. 2005), per-
haps related to the use of tools (Ramachandran 2011). In any event, as soon as
the utility of this new ability was realised in developmental terms, perhaps first in
relation to imaginative play, it would have become so important in mental life as
to have demanded massive new physical resources, thus requiring major expan-
sion of the cortex. Thus on this view, that crucial p­ hysical development (Striedter
2005) is regarded as being a result of, rather than the cause of, increased pro-
cessing needs. A modest selective advantage in terms of allowing more effective
cooperation between individuals will allow language and grammar to evolve by
a conventional Darwinian process (Dunbar 2005). The analogue of Figure 1 can
then occur in spoken language, and much later get embodied in written language.
Overall, I suggest that generic sensory interpretation, and visual thinking
in particular, provided the broad basis for development of the abilities needed
for language development and usage, with social interactions and play naturally
­developing the basis of symbolic reference through pointing and embryonic words.
 Chapter 9. Biology and mechanisms related to the dawn of language 

The specific capacities needed for language utility were greatly enhanced by the
development of musical activities, evolving into singing and dancing. This ­provided
a considerable part of the basis for language because of a deep c­ onnection between
musical and linguistic syntax in the brain (Patel 2008), and in particular the use of
recursion in musical syntax (Patel 2008).
If this view is true, it should have several testable consequences:
First, the initial variation leading to recursive type structures would have been
random (as there is no way it could have been directed), but it would have had a
genetic basis, as otherwise it would not have been a heritable quality; hence there
should be some determinable genetic linkage to this property. Either an identifi-
able set of genes, or some identifiable epi-genetic processes, should be the key to
making it possible.
Second, it should result in some specific type of neural connectivity whereby recur-
sive naming takes place, of the generic nature identified by Hawkins and Blakeslee
(2004). This type of structure should eventually be identifiable.
Thirdly, the resulting ability for recursive thought should then be manifested more
or less simultaneously in various intellectual domains, particularly in language
and play, but also in others such as patterns of music and art, and in technology.

6. The importance of emotions in this development

A key issue is what motivational structure drove these developments. Intelligent

life is not a purely rational affair, as some analysts suggest (e.g. Gintis 2007). Rather
emotional drivers largely determine to what use intellect is put (Damasio 1995).

6.1 The motivational impulse for language use

The motivation for language use is based in the functional importance of
emotions at the psychological level, associated at the neuronal level with non-
local neuronal connections such as the monoamine systems (Kingsley 2000).
Particularly important here is the mother-child emotional connection, this is the
key to early language development (Schore 1994; Greenspan & Shanker 2004),
and the developmental significance of play in learning (Vygotsky 1978) and
specifically in language (Paley 2004), a key to later language development. These
features occur in the ambient social and ecological environment, where top-down
action from the social level to individuals provides the integrated context for
emergence of language, and plays a significant role in determining the outcome.
The development of language is crucially dependent on the emotional drivers
 George F.R. Ellis

that power the desire to communicate with others and assure the individual of
their place in society. This leads to a key feature in language acquisition, namely
intention reading (Tomasello 2003).
These drivers are based in the genetically determined primary emotional
systems that are our evolutionary heritage from our distant ancestors (Panksepp
1998, 2008). According to Stevens and Price (2000), two key such systems are the
Affiliation/Belonging system, and the Rank/Status system. These systems must
therefore be important in communication, symbolism, and language development.
This is confirmed by much evidence, the importance of the affiliation/belonging
system in language development of children being emphasized by Greenspan
and Shanker (2004), while that of the Rank/Status is emphasized by d’Errico and
Henshilwood (this volume). Initial language development is based on the primary
emotional bond of an infant with its mother/primary carer (Greenspan & Shanker
2004; Schore 1994) but then is developed through social interaction, particularly
within the family context and peer group situations. These both underlie the
development of joint attention and imitation, which are key features of human
language development (Burling 2007).
A further key emotional system associated with language development is the
Play system, discussed by Jaak Panksepp (1998, 2008). This system, with its associ-
ated feelings of joy and fun and behavioural patterns of laughter, is a key founda-
tion for learning and the development of imagination, which is why we share it
with all our mammalian relatives, in particular the primates (Pellegrini & Smith
2005). However a probable major feature here was the change from rough and
tumble play to imaginative play (Paley 2004; Smith 2005; Toronchuk & Ellis 2005),
which probably played a significant role in developing theories of other minds.
Evidence for this is the fact that weak functional and pretend play is a contributor
to autism (Charman 2003).

6.2 The basis in developmental biology and evolution

By what kind of mechanism can brain plasticity lead to the formation of these
effective modules in this adapted way? This has to be an adaptive process, of the
kind labelled ‘Neural Darwinism’ by Gerald Edelman (1989), implemented by
neuromodulators such as dopamine broadcast to the neocortex from the limbic
system (thereby forming Edelman’s “value system” that guides the direction
of plasticity). But the limbic system is the seat of emotional processes, so the
fitness function guiding these adaptive processes is provided by the genetically
determined primary emotional systems of the kind examined in detail by
Panksepp (1998). Hence, evolutionary pressures in the ancestral environment
developed various psychological traits that are experienced by us as emotions and
 Chapter 9. Biology and mechanisms related to the dawn of language 

feelings, which result in behaviour enhancing our evolutionary adaptation to the

ancestral environment. Thus evolutionary processes enable brain development
through genetically determined primary emotional modules (Panksepp 1998) that
guide cortical development via a process of Neural Darwinism responding to daily
experiences and events (Edelman 1989), with a crucial affective nature (Ellis &
Toronchuk 1995). Note that this process in humans is just a further development
of the same process occurring in vertebrates and primates. There is a discernable
continuity of mechanism with our evolutionary predecessors that gives much
evidential support for this proposal (Panksepp 1998).
What is inherited, then, are basic cognitive abilities rather than specific cog-
nitive modules, plus the basic sensory and emotional systems that guide the use
of cognition. Any effective cognitive modules that result develop from interac-
tions of these systems with the social and physical environment, with the salience
of reactions guided by the inherited emotional systems (Ellis 2008a) (Figure 2).
Overall this process is of Darwinian rather than Lamarckian nature, because it
does not propose genetically determined modules with specific cognitive con-
tent, but rather genetically determined emotional systems that guide cognitive
development. The behaviour that gets inbuilt in effective folk modules will be
suitably tuned ab initio to the culture in which the individual lives, because they
are created through interaction with that culture. This experiential shaping of
these systems to fit the local environment is an aspect of the crucial feature of
brain plasticity (Donald 2001).

Evolution → genetically determined emotional systems

Emotional systems + experience → effective folkpsychology behaviour
Effective folkpsychology behaviour → learning effects

Figure 2. Evolution and effective psychological modules. The way emotion underlies the
­existence of effective psychological modules

7. Key steps towards language

Many of the basic cognitive capacities that allow language development (Toma-
sello 2003) are shared between us and other higher mammals. The dramatic
increase of brain size in relation to body weight (which may have been either the
cause or the result of other changes) enables greatly increased memory and pro-
cessing c­ apacity. Together with physiological improvements in terms of speech
production, these make speech physiologically possible. However this is not
sufficient: intellectual development is crucial also. I suggest on the basis of the
 George F.R. Ellis

above discussion that two crucial emergent features distinguishing human beings
from other animals in both developmental and evolutionary terms are changes in
the emotional system underlying language use, and changes in neural functional
capacity so as to allow recursive thought.

7.1 Changes in physiology

The physiological changes mentioned above are crucial for language development,
in particular, improved physical ability to form sounds and a greatly enlarged
brain and associated greater memory capacity are certainly needed. But it is not
clear if these are cause or effect. What caused them to happen? Brain size is crucial
in terms of increasing both total computing power and memory capacity, and also
because it necessarily promotes modularity (Striedter 2005), but this is not enough
by itself (for there are other animals with larger brains!). We need something more
than just brain size to make language fly.

7.2 Changes in the social and ecological context

Language development takes place in a social context where symbols are used for
communication of social and environmental understandings so as to form a dis-
tributed cognitive network (Donald 2001). Thus its development needs the right
social context of a society with common needs and purposes. As emphasized by
David Sloan Wilson (this volume), there will be particular environmental situa-
tions where these will be stronger drivers for language development than others.
Crucial drivers were the need for social communication to underlie coping and
cooperation, enabled by development of a theory of mind and intention reading
(Tomasello 2003), together with the child-caring and raring needs due to the long
period between birth and adulthood. But no obvious change has been identified
causing much greater need in these regards than in primates: there seems a con-
tinuity with them rather than discontinuity, except perhaps in regard to develop-
ment and use of technology.
It is certainly a possibility that development of technology was a major driver
of the need for the kind of better communication afforded by language. According
to Jacob Bronowski (1976: 64, 74) “The largest single step in the ascent of man is
the change from nomad to village agriculture … settled agriculture creates a tech-
nology from which all physics, all science takes off ”. But again it is not clear if that
is the driver or the result of language change. The features of language listed above
would have been very helpful in developing technology; perhaps one can make the
case that language and technology co-evolved as equal partners, the rapid devel-
opment at later times simply being the result of the essential nature of exponential
growth. The idea that development of technology was the key contextual partner
 Chapter 9. Biology and mechanisms related to the dawn of language 

in language development seems a very viable proposal (see the articles in d’Errico
& Backwell 2005 and particularly those by Tobias and by Parkington et al.). But
what was the crucial threshold leading to the take-off of the language-technology
partnership? It is argued above that this would have been when this co-evolution
led to the discovery of the principle of recursion in symbolism.
Perhaps times of change associated with new environments are particular
stressors demanding more effective communication, and so driving commu-
nication and language development. Could the move out of Africa, between
50,000 and 100,000 years ago, perhaps have been one such stressor? The need to
­communicate is much greater in a strange environment than in a known environ-
ment, so this move may have been the kind of change that strongly encouraged
development of better communication abilities. This would have been greatly
assisted by the associated freeing-up from debilitating diseases that resulted
from the move out of Africa (Reader 1998), allowing the mind to flourish in an
unprecedented way. This was of course recent in evolutionary terms, but perhaps
gene-culture co-evolution (Richerson & Boyd 2005) could have occurred and
allowed this event to become of evolutionary significance for the present day.
Again this would have been related to development of technology, related both
to travel and to adapting agriculture to new contexts. This might seem a promis-
ing context for a leap forward in language.

7.3 Emotional development: The nature of play, other minds

It has been emphasized above that mother/child bonding is crucial ingredient in
language development, but this is essentially the same as in other higher animals –
apart from the use of language, which is what we are trying to explain! There is no
obvious discontinuity here between pre-humans and humans. But the transition
from rough-and-tumble play to imaginative play, based in the effects of non-local
neural connections and resulting in displaced symbolic activity, could be a key
development in terms of freeing up creativity and encouraging development of a
theory of mind, as well as basic communication skills. This is a promising avenue
for promoting language abilities, particularly because music and song are natural
aspects of play. This kind of powerful joyful activity develops pre-language abilities
in terms of listening, production of sounds, and theory of mind, as well as devel-
oping basic grammatical skills.
Intellectual development: Emergence of Recursion: It has been argued above (in
agreement with Hauser et al. 2002; Fitch et al. 2005) that the single most important
feature in the emergence of language is the development of recursion in symbolic
systems, enabling abstract thought patterns to emerge. I suggested this is enabled
by specific classes of neural connections, initially emerging through genetic chance
 George F.R. Ellis

but then rapidly spreading because they provide a key aspect of the development
of higher level thought and language.
One should recognise here the multimodal nature of meaning-making and
cognition: each of the modes of communicating has a semiotic nature, and they
support each other (Kress 2000; Kress & van Leeuwen 2001). However it has been
suggested above that technology, music, and recursion may have been the links
through to developing language capacity.
Archaeological traces: Might any of the features discussed here perhaps leave some
traces in the archaeological record? There are many such links to the above narra-
tive, most already being explored; the point here is that this search can be related
to aspects of how the brain underlies language, as outlined above, which might
possibly help focus the search.
There are of course records of group membership, related to belonging/ cohe-
sion, and symbols of social status, related to rank/power; these are a small step
along the symbolic route. One might search further for records of music and danc-
ing, argued above to be key players in the development of language, seeing if any
traces of recursive patterning occur in these records. One can search for symbols
and artefacts in the archaeological record associated with play: toys as symbols
used in imaginative play, playing boards, scoreboards, and so on. Could any of
the Blombos patterned artefacts be of this nature? And in particular do any show
signs of recursive features, such as fractal-like patterning? One might ask if there
is any way of identifying traces of recursive thought in any other artefacts, for
example those that are purely decorative. And finally the link to technology is
crucially important, as evidenced in other articles in this book. Perhaps recursive
patterning can be found in tool design, implying it is embodied in manufacturing
techniques. If so, that would be a significant discovery as regards thought patterns
essential to language. However it has to be admitted they will individually be sug-
gestive rather than decisive. A powerful case will only be made by a collection of
evidence of all kinds, of such a nature as to provide a cumulative case.

References

Arnheim, R. 1969. Visual Thinking. Berkeley: University of California Press.

Auletta, G., Ellis, G.F.R. & Jaeger, L. 2008. “Top-Down Causation: From a Philosophical Problem
to a Scientific Research Program.” J R Soc Interface 5: 1159–1172.
Beer, S. 1981. Brain of the Firm. Chichester: John Wiley.
Bergen, B.K. & Chang, N. 2003. “Embodied Construction Grammar in Simulation Based
Language Understanding.” In Construction Grammars(s): Cognitive and Cross-Language
Dimensions, J–O. Östman & M. Fried (eds), 147–190. Amsterdam: John Benjamins.
Berger, P. 1963. Invitation to Sociology: A Humanistic Perspective. New York: Doubleday.
 Chapter 9. Biology and mechanisms related to the dawn of language 

Berger, P. & Luckmann, T. 1967. The Social Construction of Reality: A Treatise in the Sociology of
Knowledge. New York: Anchor.
Bickerton, D. 2001. Language and Human Behaviour. Seattle: University of Washington Press.
Booch, G. 2007. Object-Oriented Analysis and Design with Applications. Menlo Park:
Addison-Wesley.
Bronowski, J. 1976. The Ascent of Man. London: The BBC.
Burling, R. 2007. The Talking Ape: How Language Evolved. Oxford: Oxford University Press.
Charman, T. 2003. “Theory of mind and the early diagnosis of autism.” In Understanding other
minds: Perspectives from developmental cognitive neuroscience, S. Baron Cohen, H. Tager-
Flusberg & D. J. Cohen (eds), 422–441. Oxford: Oxford University Press.
Christiansen, M. & Kirby, S. (eds). 2005. Language Evolution. Oxford: Oxford University Press.
Clark, A. 2008. Supersizing the Mind: Embodiment, Action, and Cognitive Extension. Oxford:
Oxford University Press.
Conway Morris, S. 2003. Life’s Solution: Inevitable Human Beings in a Lonely Universe.
­Cambridge: Cambridge University Press.
Damasio, A. 1995. Descartes’ error: Emotion, reason, and the human brain. New York: Avon.
Damasio, A. 2000. The feeling of what happens: Body, emotion and the making of consciousness.
London: Vintage.
Deacon, T. 1997. The symbolic species: The co-evolution of language and the brain. London:
­Penguin Books.
Deacon, T. 2003. “Universal Grammar and Semiotic Constraints.” In Language Evolution,
M. Christiansen & S. Kirby (eds), 111–139. Oxford: Oxford University Press.
d’Errico, F. & Backwell, L. (eds). 2005. From Tools to Symbols: From Early Hominids to Modern
Humans. Johannesburg: Witwatersrand University Press.
d’Errico, F. & Henshilwood, C. this volume. “The origin of symbolically mediated ­behaviour:
from antagonistic scenarios to a unified research strategy.” In ‘Homo Symbolicus’: The
Dawn of Language, Imagination, and Spirituality, C. Henshilwood and F. d’Errico (eds).
49–73. ­Amsterdam: John Benjamins.
Donald, M. 2001. A mind so rare. New York: Norton.
Dunbar, R.I.M. 2005. “The origin and subsequent evolution of language.” In Language Evolution,
M. Christiansen & S. Kirby (eds), 219–234. Oxford: Oxford University Press.
Edelman, G. 1989. Neural Darwinism: The theory of group neuronal selection. Oxford: Oxford
University Press.
Ellis, G.F.R. 2008a. “Commentary on ‘An Evolutionarily Informed Education Sience’ by David C
Geary.” Educational Psychologist 43 (4): 1–8.
Ellis, G.F.R. 2008. “On the nature of causation in complex systems.” Transactions of the Royal
Society of South Africa 63: 69–84.
Ellis, G.F.R. & Toronchuk, J. 1995. “Neural development: Affective and immune system influ-
ences.” In Consciousness and emotion: Agency, conscious choice, and selective perception,
R. D. Ellis & N. Newton (eds), 81–119. Amsterdam: John Benjamins.
Feldman, J. 2006. From Molecule to Metaphor: A Neural Theory of Language. Cambridge, Mass:
MIT Press.
Fitch, W.T. Hauser, M.D. & Chomsky, N. 2005. “The evolution of the language faculty:
­clarifications and implications.” Cognition 97: 179–210.
Gintis, H. 2007. “A framework for the unification of the behavioral sciences.” Behavioral and
brain sciences 30: 1–61.
 George F.R. Ellis

Goodman, K. 2005. What’s Whole in Whole Language: 20th Anniversary Edition. Muskegon, MI:
RDR Books.
Greenspan, S. & Shanker, S. 2004. The first idea: How symbols, language, and intelligence evolved
from our primate ancestors to modern humans. Cambridge, Mass: Da Capo Press.
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.
Hawkins, J. & Blakeslee, S. 2004. On Intelligence. New York: H Holt & Co.
Hofstadter, D.R. 1980. Gödel, Escher, Bach: An Eternal Golden Braid. Harmandsworth: Penguin
books.
Hurford, J.R. 2007. The Origins of Meaning. Language in the Light of Evolution. Oxford: Oxford
University Press.
Kingsley, R. 2000. Concise text of neuroscience. Philadelphia: Lippinscott, Williams and Wilkins.
Koch, C. 2004. The quest for consciousness: A Neurobiological approach. Englewood, Co: Roberts
and Company.
Krashen, S.D. & Terrell, T.D. 1983. The natural approach: Language acquisition in the c­ lassroom.
San Francisco: Alemany Press.
Kress, G. 2000. “Multimodality.” In Multiliteracies: Literacy and design of social futures, B. Cope &
M. Kalantzis (eds), 182–202. London: Routledge.
Kress, G. & van Leeuwen, T. (2001). Multimodal discourse: The Modes and Media of C ­ ontemporary
Communication. London: Arnold.
Lakoff, G. & Johnson, M. 1980. Metaphors we live by. Chicago: University of Chicago Press.
Landauer, T.K., Foltz, P.W. & Laham, D. 1998. “An Introduction to Latent Semantic Analysis.”
Discourse Processes 25: 259–284.
Levitin, D.J. 2007. This is your brain on music: the science of a human obsession. London: Plume.
Mesoudi, A., Whiten, A. & Laland, K.N. 2006. “Towards a unified science of cultural E ­ volution.”
Behavioral and brain sciences 29: 329–383.
Okasha, S. 2006. Evolution and the levels of selection. Oxford: Oxford University Press.
Paley, V.G. 2004. A Child’s Work: the Importance of Fantasy Play. Chicago: University of Chicago
Press.
Panksepp, J. 1998. Affective neuroscience: The foundations of human and animal emotions.
Oxford: Oxford University Press.
Panksepp, J. 2008. “The Power of the Word May Reside in the Power of Affect.” Integr Psych
Behav 42: 47–55.
Parker, A. 2004. In the blink of an eye: how vision kick-started the big bang of evolution. London:
Free Press.
Patel, A.D. 2008. Music, Language, and the Brain. Oxford: Oxford University Press.
Pellegrini, A.D. & Smith, P.K. 2005. The Nature of Play: Great Apes and Humans. New York: The
Guilford Press.
Pinker, S. 1994. The Language Instinct. London: Penguin Books.
Purves, D. 2010. Brains: How they seem to work. Saddle River, New Jersey: Pearson Education.
Ramachandran, V.S. 2011. The Tell-Tale Brain: Unlocking the Mystery of Human Nature. London:
William Heinemann.
Reader, J. 1998. Africa: A Biography of the Continent. London: Penguin Books.
Richerson, P.J. & Boyd, R. 2005. Not by Genes Alone: how culture transformed human evolution.
Chicago: University of Chicago Press.
Roberts, E.S. 2009. Java: An introduction to computer science. Boston: Addison Wesley.
 Chapter 9. Biology and mechanisms related to the dawn of language 

Schore, A.N. 1994. Affect regulation and the origin of the self: The neurobiology of emotional
development. Hillsdale, NJ: Erlbaum.
Scott, A. 1995. Stairway to the mind. New York: Springer Verlag.
Simon, H.A. 2001. The sciences of the artificial. Cambridge, Mass: MIT Press.
Smith, F. 1976. Reading. Cambridge: Cambridge University Press.
Smith, P.K. 2005. “Social and pretend play in children.” In The Nature of Play: Great Apes and
Humans, A.D. Pellegrini & P. K. Smith (eds), 173–209. New York: The Guilford Press.
Stevens, A. & Price, J. 2000. Evolutionary psychiatry: A New beginning. New York: Taylor and
Francis.
Striedter, G.F. 2005. Principles of Brain Evolution. Sunderland, Mass: Sinauer Associates.
Tomasello, M. 2003. Constructing a language: A usage-based theory of language acquisition.
Cambridge, Mass.: Harvard University Press.
Toronchuk, J.A. & Ellis, G.F.R. 2005. Affective neuronal group selection: The nature of the p­ rimary
emotional systems. Unpublished manuscript, University of Cape Town. Available at http://
www.mth.uct.ac.za/~ellis/AND%20II.doc.
Trask, R. L 1999. Key Concepts in Language and Linguistics. Abingdon: Routledge.
Trask, R.L. 2007. Language and Linguistics: The Key Concepts. Abingdon: Routledge.
Vygotsky, L.S. 1978. Mind in Society: The development of higher psychological processes.
­Cambridge, Mass: Harvard University Press.
Wilson, D.S. this volume. “The human major transition in relation to symbolic behaviour,
including language, imagination, and spirituality.” In ‘Homo Symbolicus’: The Dawn of
­Language, Imagination, and Spirituality, C. Henshilwood and F. d’Errico (eds). ­Amsterdam:
John Benjamins.
 chapter 10

The other middle-range theories

Mapping behaviour and the evolution of the mind
Benoît Dubreuil
Philosophy Department,Université du Québec à Montréal

Archaeologists have long been aware of the need to make explicit the
middle-range theories that they use to bridge the gap between raw archaeological
evidence and human behaviour. In this paper, I discuss the difficulties in
proposing such theories when dealing with long-extinct hominin populations,
among which significant biological differences may have existed. I emphasize how
far we remain from a general theory that would link genetic, neural, cognitive,
and behavioural evolution, and how difficult it is to go beyond trivial explanations
of the evolution of the human mind and behaviour. I argue that students of
human cognitive evolution should move away from general mental phenomena,
which they have focused on in recent years, and turn their attention to
increasingly precise cognitive mechanisms realized in specific neural structures.

1. Introduction

Lewis Binford’s theoretical work in archaeology is most famous for his call for the
development of a middle-range theory, that is, for his emphasis on the need to
develop procedures and frameworks to bridge the gap between raw archaeologi-
cal evidence and dynamic human behaviour (Binford 1977, 1981). Over the years,
archaeologists have followed this suggestion, developing diverse strategies to guide
and constrain their interpretation of the archaeological record, and to reconstruct
the behaviour of ancient human populations. To be sure, bridging the gap between
material culture and behaviour is an unending endeavour. New technologies
­produce new kinds of data, which in turn create new problems of interpretation.
Moreover, as archaeological data are by nature limited and more often than not
compatible with various behavioural patterns, there is sometimes simply no way of
arriving at a satisfactory middle-range theory, and large parts of our past are thus
condemned to remain unexplained. These problems are well known and inescap-
able. They represent the frontiers that archaeology faces as a discipline.
 Benoît Dubreuil

Nevertheless, as an outsider to archaeology, I must admit that I have always

been impressed by the amount of information that archaeologists can extract from
scant evidence. Although they often disagree on their general interpretation of
the archaeological record, they mostly agree on what can in principle be learned
from various types of evidence, and how. This is true, at least, for what concerns
recent human prehistory. The interpretation of older periods of human prehistory
raises more acute problems. The reason is that the human species has evolved,
and therefore the deeper we look into our past, the more likely we are to find
a significant gap between the mind and behaviour of modern humans and that
of extinct hominins. Human biological evolution thus adds several layers of
complexity to our inquiry.
Anthropologists, sociologists, and other social scientists usually take for
granted that any variation in behaviour between two human populations results
from cultural or ecological factors. Sharing this assumption, students of our recent
archaeological past have made extensive use of anthropological and sociological
theories to constrain and regiment their interpretation of material culture. But
expanding the scope of inquiry over hundreds of thousands of years makes this
assumption questionable. The fact of human evolution dramatically alters our
quest for a middle-range theory. Cognition and biology become additional vari-
ables that must be taken into account as soon as we consider behaviours spanning
into our distant past. For such an inquiry it is insufficient for archaeologists to take
inspiration from anthropology and sociology; ethnology, cognitive neuroscience
and evolutionary biology must kick in.

2. A research agenda for the next two centuries

As the number of disciplines and methods expand, however, the risk of students of
human evolution being submerged by new theories and data becomes real and, to
tell the truth, it is far from obvious that the challenges of interdisciplinary research
will be met successfully. Indeed, working at several different explanatory levels at
once is not only beyond the reach of most scholars, but inaccessible even to the best
among them. There is simply too much information to process. A comprehensive
theory of human evolution must start from the archaeological record and explain
how the evolution of human behaviour relates to underlying cognitive, neural,
and genetic evolution (see Figure 1). Fortunately, changes at one level are linked
to changes at the other levels (e.g. genetic evolution is linked to neural evolution
which, in turn, is linked to cognitive evolution). What we know at one level can
thus in principle be used to guide and constrain hypotheses about what is going
on at the other levels. Unfortunately, there is a wide gamut of ways in which the
 Chapter 10. The other middle-range theories 

various levels can relate to one another. What is worse, we know close to nothing
about how genetic evolution affects neural evolution, or how neural evolution
affects cognitive and behavioural evolution.

Material culture

Behaviour

Cognition

Natural Cultural
Neurons
environment environment

Genes

Natural selection

Figure 1. Levels of explanations in human evolution

General theories covering the whole problem of evolution from natural

selection to the archaeological record will probably remain beyond our reach
for several generations. Moreover, given the intrinsic complexity of the biolog-
ical world, the best we can hope is probably to produce patchy explanations,
with piecemeal integration between different levels of explanation (Mitchell
2003; Craver 2007). In the case of human evolution, to my knowledge, the only
hypothesis that ever attempted to cover the whole ladder of complexity was the
so-called FOXP2 hypothesis. According to this hypothesis, the mutation of a
specific gene (FOXP2), expressed in the basal ganglia, has been selected for in
modern Homo sapiens because of its impact on language. The enhanced speech,
motor control and recursive syntax made possible by this mutation would in
turn explain the evolution of “modern behaviour,” including symbolic culture
(Corballis 2004; Lieberman 2007). But nearly every step in this analysis can be
questioned. FOXP2 apparently took its modern form long before the evolu-
tion of modern humans (Krause et al. 2007) and its impact on language and
­cognition hardly explains the kinds of behaviours associated with Homo sapiens
(Dubreuil 2008).
 Benoît Dubreuil

Despite its limited plausibility, the FOXP2 hypothesis has generated much
enthusiasm in the literature on human cognitive evolution. This is probably
because it offered, for the first time ever, a complete explanatory scenario, drawing
a connection between natural selection and the archaeological record, with at
least some test implications. The FOXP2 hypothesis was ambitious, probably too
ambitious, but it is definitely the kind of hypotheses that we would like to find. To
understand how far we remain from such comprehensive theories, we should keep
in mind how difficult it is to link only one level of explanation with another. Consider
for instance the well-known difficulty in linking the archaeological record with
behaviour, that is, of producing middle-range theories in Lewis Binford’s classical
sense. We still have no consensus, for instance, on the behavioural implications of
the use of symbolic artefacts by Neandertals, a question of the utmost importance
for human cognitive evolution. Current evidence is compatible with various
behavioural interpretations, each having different implications at the lower levels
of cognitive, neural and genetic evolution (Mellars 2005; Zilhaõ 2006). Choosing
among these interpretations is currently impossible and will remain so until new
findings make it possible to reconstruct more precisely the evolution of symbolic
behaviour in Eurasia and Africa.
As a philosopher working in the cognitive science and human behaviour,
it is quite unlikely that I will contribute to these findings. Fortunately that is
not my goal. The question that I would like to ask instead is the following one:
is it possible to inquire into the lower levels of complexity before settling the
most crucial issues at the higher levels? Can we discuss the genetic, neural,
and cognitive foundations of symbolic behaviour before we determine if, and
for what purpose, Neandertals created and used symbolic artefacts? In other
words, can we devise middle-range theories linking behaviour to cognition, or
cognition to neurons and genes, without first settling the hottest issues at the
behavioural level?
If it is impossible, then specialists in cognitive neuroscience, genetics, and
evolutionary biology might do well to avoid discussing the evolution of the human
mind for some time, and to return to it once major debates among archaeologists
have been settled. But the contribution of non-archaeologists might not be totally
useless, even in this context of uncertainty. At the end of the day, archaeological
data, fossils and artefacts, will be deciding, but this does not preclude going beyond
the construction of middle-range theories in Lewis Binford’s classical sense, that
is, theories linking the archaeological record to behaviour. We can also propose
other middle-range theories at the lower level of genes, neurons, and cognitive
functions. After all, our mind might have evolved following various scenarios, and
yet all scenarios are not equally likely.
 Chapter 10. The other middle-range theories 

3. What middle-range theories for non-archaeologists?

A peculiar feature of non-archaeological research on human evolution is that

the most advanced research has probably been achieved at the lowest level of
complexity, that of natural selection, thanks to the work done over the years by
evolutionary biologists. There are now very efficient theories on the market to
model how specific cognitive or behavioural traits might have been selected for
in our ancestral environment. This is truly helpful, especially since evolutionary
biologists have learned to work hand in hand with evolutionary psychologists to
determine how selective pressures might have led to the evolution of properly
human dispositions.
The counterpart of this progress, obviously, is that it does not tell us when,
and in which contexts, the specific features of our mind evolved in the human
lineage. It is one thing to recognize that the modern mind has a genetic basis that
has been selected for in our evolutionary past; it is another to determine what
genes have been selected for, and when and what was the precise impact of this
selection on our ancestors’ brains and behaviour. Very few genes can be directly
and confidently linked to cognition and behaviour. In fact, higher-level cognitive
processes and behaviours mostly have a polygenetic basis that makes a direct
inquiry into the genetic foundations of the modern mind problematic (Schaffner
1998; Wynn & Coolidge 2007).
The other feature that I want to highlight is the difficulty that we have had until
now in formulating nontrivial theories linking neural, cognitive, and behavioural
evolution. Many proposals have not gone very far beyond what I would call the
“bigger brain-greater intelligence hypothesis”. Encephalisation is beyond doubt the
most noticeable fact in the evolution of the human brain. It would be astounding
if this process were not causally linked specifically with human creativity and
culture. But the “bigger brain – greater intelligence hypothesis” has no clear test
implications, because there is an infinite variety of ways in which encephalization
can be related to cognitive and behavioural evolution.
Science does not progress by explaining a complex phenomenon by an equally
complex phenomenon. The creativity of modern Homo sapiens behaviour, for
instance, remains entirely unexplained if it is simply related, with no further spec-
ification, to the evolution of modern intelligence or language. The complexity of
the explanans (modern intelligence or language) being equivalent to that of the
explanandum (modern culture), it is not clear what such an explanation is doing
apart from rephrasing the problem. I think that I am not too far from the truth if I
say that archaeologists have sometimes been tempted by such trivial explanations.
It is certainly not my objective to criticize them for that, since in the context of an
 Benoît Dubreuil

interdisciplinary division of labour the construction of nontrivial cognitive and

neural theories to account for behavioural evolution in extinct hominins is not
strictly their responsibility.
The reason why archaeologists often have recourse to uninformative
cognitive and biological explanations might be because such explanations are
not at the centre of their interest, and often intervene in their accounts of past
behaviour as an explanation of last resort. A behavioural change is explained
biologically when other non-biological variables – such as culture, demography,
or climate – have exhausted their explanatory potential. This is not totally
unreasonable, because non-biological variables are likely to impact on the
presence of the behaviours we are interested in, be it the production of symbolic
artefacts or other innovations. It is thus reasonable to control for these variables
before looking to biology for an explanation. But in the context of human
evolution, using biology as a variable of last resort also creates some problems.
Consider, for instance, the way Richard Klein frames the current alternatives in
the explanation of modern sapiens behaviour:
“Most palaeoanthropologists now accept Out of Africa and argue mainly about
the extent to which non modern Eurasians and modern African immigrants may
have interbred. Most authorities further agree that a behavioural transformation
underlay the modern African expansion, but they presently divide between a
majority who believe that advanced behaviours accumulated gradually between
perhaps 120 and 50 ka and a minority who believe they appeared abruptly
about 50 ka. Those who favour gradual development usually attribute it to long-
term social, demographic, or economic shifts, while those who perceive abrupt
development, mainly the present writer, suggest that it was prompted by genetic
change.” (Klein 2008: 267, emphasis added)

One problem with this proposition is that it is difficult to see how the value
of the biological explanation can be weighed against the alternatives. To assess the
explanatory value of genetic evolution, controlling for other non-biological vari-
ables is certainly necessary but it is not on its own sufficient. Genetic or (more
generally) biological changes are not all equally likely and they do not all explain
equally well various behavioural changes. Thus, biology cannot simply enter as a
“get out of jail free card” when other explanations become implausible. One also has
to present scenarios that link plausible genetic, neural, and cognitive changes to the
behavioural evolution to be explained. Those are precisely the nontrivial middle-
range theories that we need. By contrast to Lewis Binford’s classical ­middle-range
theories – linking the archaeological record to behaviour – they can only emerge
from the cooperation of philosophers, psychologists, linguists, and social scientists
with a strong background in cognitive neuroscience.
 Chapter 10. The other middle-range theories 

4. Current middle-range theories

Valuable proposals have already attempted to go beyond the “bigger brain-greater

intelligence hypothesis”. In this section, I discuss the limitations of three influential
and thought-provoking middle-range theories: (1) Derek Bickerton’s syntactic
hypothesis, (2) Robin Dunbar’s social intelligence hypothesis, and (3) Frederick
Coolidge and Thomas Wynn’s working memory hypothesis. In the following
section, I argue that the common problem with these middle-range theories lies
with their focus on complex mental phenomena, rather than on the cognitive and
neural mechanisms that produce these phenomena.
Bickerton (2003) has gone beyond the trivial linguistic explanation by
emphasizing the role of syntactic recursion. Recursion is a formal property of
human language that makes it possible to embed clauses within clauses. There
is no question that recursion (or something like it) is instrumental in expressing
in language the kind of metarepresentations that humans are capable of. There
are nevertheless important problems with this proposal. The first is that recursion
is a formal property of language and not a cognitive mechanism. Unfortunately,
there is no consensus on its cognitive (not to say neural) foundations in humans
(Calvin & Bickerton 2000; Lieberman 2007; Wynn & Coolidge 2007). Another
problem is that, while some contend that recursion is proper to humans, others
doubt whether it is even a true property of natural human language (Hauser et al.
2002; Suddendorf & Corballis 2007; Marcus 2008). Another problem is that the
behavioural implications of the evolution of linguistic recursion are far from clear.
I have raised doubt elsewhere about the capacity of recursion to account for the
diversity of behaviours that has been associated with modern sapiens culture and
even to explain the emergence of symbolic behaviours (Dubreuil 2008).
Another challenging middle-range theory has been advanced by Robin
Dunbar, who has argued over the years in favour of a “social brain hypothesis”
(Dunbar 2003, 2007). Dunbar explains the evolution of modern sapiens culture by
the evolution of “theory of mind” or the capacity to ascribe mental states to oneself
and others. Psychologists still debate in what way(s) human theory of mind differs
from that of nonhuman primates, but there is little doubt that differences exist
and must have played a role in the evolution of language, culture, and imagination
(Call & Tomasello 2008). One problem with the social brain hypothesis, however,
is that the concept of “theory of mind”, although is it not as general as “language”
or ‘intelligence”, is still very general. Many specific cognitive mechanisms might be
responsible for the evolution of specifically human theory of mind. Dunbar (2004,
2007) points toward the existence of a correspondence between brain growth and
the evolution of various levels of intentionality, but this move is not motivated at
 Benoît Dubreuil

the cognitive or neurological levels. There are simply no data that could allow us to
establish a correlation between brain size and levels of intentionality.
Psychologist Frederick Coolidge and archaeologist Thomas Wynn (Coolidge &
Wynn 2001, 2005; Wynn & Coolidge 2007) have proposed a third hypothesis that
I want to mention. Their proposal is similar to earlier ones that explain properly
human behaviours by the evolution of the executive functions of the brain (Mithen
1996; Deacon 1997). Coolidge and Wynn’s hypothesis is centred on the concept
of working memory, which is defined as the capacity to hold in mind recently
­processed information and to keep it available for further processing. They build
on the model of working memory developed over the years by psychologist Alan
­Baddeley. According to Baddeley (Baddeley & Hitch 1974; Baddeley 2001), the
­system of working memory is composed of a central executive, responsible for the
control and regulation of information, and of two “slave systems”, the phonological
loop and the visuospatial sketchpad. The fact that modern humans perform ­better
than nonhuman primates in different working memory tasks suggests that the
system evolved in the human lineage (Read 2008). This evolution, according to
Coolidge and Wynn, is likely to explain the appearance of more flexible behaviours
in recent human evolution, as well as the difference between modern Homo sapiens
and Neandertals (Coolidge & Wynn 2004; Wynn & Coolidge 2004). The central
problem with the concept of working memory, even as specified in B ­ addeley’s
model, is that it is still very general. At the neural level, working memory tasks
activate widely distributed networks in the prefrontal, parietal, and temporal
cortices. Moreover, it is difficult to establish exactly what is specific to the modern
human working memory system. Although the human working memory system
certainly differs from that of nonhuman primates, the different components of
working memory are already present in our closest relatives. But if the gap between
humans and nonhuman primates is more quantitative than qualitative, how are
we to determine how much working memory is needed to produce modern
sapiens behaviour? How much working memory was there in Neandertal or Homo
heidelbergensis? Before coming back to these questions, I want to discuss what I see
as a common limitation of these three hypotheses.

5. What’s wrong with current theories?

The middle-range theories that we need must offer a link between behaviour
and cognitive evolution. This is basically what the hypotheses discussed above
are doing. Recursive syntax, theory of mind, and working memory are cognitive
concepts; that is, they refer to the ability of the human mind to process some sort
of information. Their common limitation, however, comes from the fact that they
 Chapter 10. The other middle-range theories 

refer to complex cognitive tasks, which are realized by largely overlapping cognitive
and neural mechanisms. This, I will argue, makes it impossible to draw distinct test
implications from these hypotheses.
Let’s begin with working memory and theory of mind. At the cognitive level,
they depend on similar cognitive functions. The capacity to inhibit prepotent
responses, for instance, is needed to succeed in complex theory of mind tasks
(Carlson et al. 2004), but it is also a function of the central executive in Baddeley’s
model of working memory (Jurado & Rosselli 2007). The same can be said of
selective attention and attention switching, which the central executive is thought
to perform, but which are also essential to succeed in complex theory of mind tasks
(Gerrans & Stone 2008). But if theory of mind and working memory depend on
the same cognitive mechanisms (inhibitory control and selective attention), how
are we to distinguish between a behavioural change related to (1) the evolution
of theory of mind, (2) the evolution of working memory, and (3) the evolution of
inhibitory control or selective attention?
A similar problem exists with the syntactic hypothesis. De Villiers and
Pyers (2002), for instance, have shown that the mastery of embedded clauses in
language development was predictive of success in complex theory of mind tasks,
suggesting the presence of a common underlying mechanism. What then are the
distinct implications of the syntactic hypothesis? Things risk getting even more
confused when syntactic recursion is equated with other complex cognitive tasks.
Suddendorf and Corballis (2007), for instance link recursive syntax not only to the
evolution of theory of mind, but also to that of episodic memory and mental time
travel. Coolidge and Wynn (2007), for their part, link the evolution of recursion
to that of working memory and, more particularly of the phonological storage
capacity. Here again, it is difficult to see how these various hypotheses have distinct
test implications.
The picture does not get clearer if one shifts the focus from the cognitive to
the neural level. Disentangling the different hypotheses remains difficult because
syntax, theory of mind, and working memory activate overlapping brain areas.
This is most clearly the case for theory of mind and working memory. Both of
them, for instance, activate areas of the prefrontal cortex, associated with goal-
maintenance and inhibition (Costa et al. 2008; Edin et al. 2009). Complex syntax,
for its part, does activate brain areas that are not associated with theory of mind or
working memory, such as Broca’s area in the inferior frontal gyrus. Nevertheless,
the production of complex hierarchical sentences also depends on long-distance
connections with more posterior regions of the brain such as Wernicke’s area
(Friederici 2009), which is located in the temporoparietal junction, an area
associated with both theory of mind and selective attention (Saxe & Kanwisher
2003; Decety & Lamm 2007).
 Benoît Dubreuil

The presence of such overlapping areas creates an interesting conundrum for

students of human evolution. If the evolution of language was related to a change
in the temporoparietal junction, it might have coincided with the evolution of
theory of mind and working memory. By contrast, if it was related to a change in
Broca’s area, it is unlikely to have coincided with the evolution of these abilities,
but it might have coincided with the evolution of other abilities that depend on
hierarchical action sequencing (Koechlin & Jubault 2006; Stout & Chaminade
2009). Similarly, if the evolution of theory of mind was related to a change in the
dorsal prefrontal cortex, it is likely to have coincided with a change in working
memory, but not if it was related to the reorganization of a region that is not
typically linked to working memory (such as the medial prefrontal cortex, for
instance).
In sum, although different hypotheses have been proposed to account for the
evolution of the modern mind, these hypotheses do not have distinct implica-
tions because they refer to abilities realized by overlapping cognitive and neural
mechanisms.

6. From mental phenomena to mental mechanisms

In a book on explanations in cognitive neuroscience, philosopher William Bechtel

discusses epistemological problems that have appeared in research on memory and
that, I suggest, are also relevant for research on the evolution of the human mind.
Bechtel (2008) argues that psychologists working on memory first attempted to
distinguish different types of memory. They established distinctions, for instance,
between long-term and short-term memory, semantic and episodic memory,
implicit and explicit memory, etc. Although they managed to identify distinct and
robust phenomena at the psychological level, problems appeared when they tried
to identify the cognitive and neural mechanisms that produced these phenomena.
The different types of memory appeared not only to activate largely overlapping
brain areas, but also to activate areas related to other processes, such as perception
or motor control. Bechtel argues that the decomposition of memory processes
have pointed toward a characterization of memory that is largely orthogonal to
the phenomenal distinctions established by psychologists:
“Although the view that memory involves storing information in a location
separate from where it might be processed (as references to memory stores
suggest) seems compelling from the perspective of devising an artificial mind, it
looks increasingly less plausible when considering the range of areas that seem to
be activated in memory tasks as well as in other cognitive activities. An alternative
view is that remembering is simply one aspect of various mental phenomena, that
 Chapter 10. The other middle-range theories 

as the mind performs the operations that generate those phenomena, it is altered
in ways that alters future performance and in some cases allows it to remember
what has happened to it.” (Bechtel 2008: 83)

What parallel can be drawn between research on memory and research on the
evolution of the human mind? As with early memory research, the study of
cognitive evolution has been largely centred on distinctive and well-studied mental
phenomena. Recursive syntax, working memory, and theory of mind, I suggest,
have received so much attention because they refer to well-studied cognitive
tasks in which humans’ performance is robustly different from that of nonhuman
primates. As in the case of memory, however, the clear distinction between these
abilities is blurred when we consider how they are produced at the cognitive and
neural levels. As explained above, recursive syntax, theory of mind, and working
memory activate largely overlapping brain areas. To be sure, some areas are not
activated by all three abilities, but these areas also perform operations that go well
beyond the specific ability that the theories appeal to. Broca’s area, for instance, is
involved in the production of complex sentences, but also in other behaviours that
require hierarchical sequencing.
The lesson to be drawn from memory research is, I think, straightforward.
Recursive syntax, theory of mind, and working memory refer to phenomena at a
much too high level to be useful for explanation. We need to move toward more
specific cognitive mechanisms as they are realized in specific brain areas. This
objective, however, is not without its own difficulties, which we must be aware of.
Memory research has been able to move toward finer-grained accounts of memory
because it had access to the powerful tools of cognitive neuroscience. Above all,
it could build on imaging techniques that make it possible to localize cognitive
functions in brain structures. Students of human evolution must learn from
cognitive neuroscience, but they also have to be aware that the architecture of the
modern brain could have evolved along various pathways and that it is impossible
to decide between them on the sole basis of what we know – or what we may know
– of brain structures and cognitive functions in normal functioning modern Homo
sapiens. Neuroimaging studies, for instance, are of no help in deciding if Broca’s or
­Wernicke’s area took their modern form in Homo erectus or Homo heidelbergensis.
To produce such inferences we need different tools, tools which are more adapted
to our object, but which come with their own limitations.

7. Toward palaeocognitive neurosciences

Students of human cognitive evolution will never directly benefit from the ­powerful
imaging techniques that have revolutionized cognitive neurosciences. But that is
 Benoît Dubreuil

not to say that there is no way to support specific inferences about the evolution of
the brain. In this section, I present the three most relevant tools for our purpose.

7.1 Human and nonhuman primate comparative neuroscience

Although it is impossible to scan the brain of extinct hominins, it is still possible
to compare cognitive functions and neural structures in humans and nonhuman
primates. Comparative neuroscience can provide hints as to which cognitive
and neural mechanisms evolved in the human lineage. Comparative lesion and
imaging studies, for instance, indicate that visual working memory pathways
in the parietal and temporal cortices have been displaced in the course of
human evolution, leaving more space in- between for the parietotemporal areas
dedicated to high-level cognitive abilities (Ungerleider et al. 1998). Comparative
neuroanatomic studies have also yield interesting results. They have shown, for
instance, that the human frontal lobe is not unexpectedly large for a primate of our
brain size, and thus it did not benefit more form encephalization than other parts
of the neocortex (Semendeferi et al. 2002). They have also highlighted the need
for understanding the role of white matter in encephalization and its importance
in establishing connections both between close and distant areas of the brain
(Schenker et al. 2005; Schoenemann et al. 2005; Rilling et al. 2008). The obvious
limitation of comparative neuroscience is that it says nothing about when different
mechanisms might have evolved in the human lineage, or how they evolved. For
this, we need to turn to other tools.

7.2 Developmental neuroscience

Development studies are both essential and dangerous for understanding the
evolution of the human mind. On the one hand, it is often reasonable to assume
that cognitive and neural mechanisms that develop later in childhood evolved
more recently. This is especially the case when the later-developing mechanisms
somehow depend on the normal functioning of the earlier-developing ones. It is
not accidental, for instance, that the multimodal association areas develop after
the different sensory cortical areas. Without an input from unimodal areas, there
would simply be no multimodal information to process. On the other hand, there
are well known caveats to the idea that ontogeny recapitulates phylogeny. One of
them is that it is not strictly impossible for an early-developing trait or mechanism
to evolve later, if it functions in a more or less modular fashion. Another is that,
given the plasticity of the brain, later-evolving mechanisms can evolve first and
then see their functioning altered by the later evolution of an early-developing
mechanism. Despite these limitations, developmental neuroscience can function
as a useful heuristic. It makes sense, for instance, to argue that the basic motivations
 Chapter 10. The other middle-range theories 

to communicate and to share attention with others evolved before more complex
features of language, because they are at the foundation of language acquisition
in modern humans (Tomasello 2008). At the same time, it also makes sense to
argue that metalinguistic abilities evolved late in the human lineage, after recursive
syntax, because they also develop later in human children (Dubreuil 2008). As for
comparative neuroscience, however, developmental neuroscience is not likely to
indicate when different mechanisms evolved in the human lineage.

7.3 P
 alaeoneurology
To address the when question, we have to turn to the study of endocasts. To be
sure, cognitive functions are not written in endocasts, and inferences about brain
evolution can be drawn only indirectly. Moreover, the paucity of the fossil record
for crucial phases of human evolution (and especially for the evolution of modern
Homo sapiens) often condemns researchers to base their interpretation on a few
incomplete specimens. But we should not minimize the information that endocasts
can convey. They are likely to provide indication about relative and absolute brain
size, about the relative size and shape of different lobes, about gyrification, and,
potentially, about the pace of brain growth in children. The study of endocasts
alone, for instance, gives no principled way to decide if encephalization must
be considered in absolute or relative terms (total brain mass or encephalization
quotient). It indicates, however, that encephalization in early members of the
genus Homo was more significant in absolute than in relative terms (given the
larger body size of Homo ergaster and Homo erectus), while encephalization
in Homo heidelbergensis was significant in both absolute and relative terms
(Rightmire 2004). If encephalization was selected for its impact on cognition, as
we can reasonably assume, then middle-range theories based on encephalization
have more weight for Homo heidelbergensis than for early members of the genus
Homo. The same principle applies to modern Homo sapiens. Different arguments
have been made to link the evolution of the modern cranium with a relative
expansion of the temporal and parietal lobes (Liebermann et al. 2002; Bruner et
al. 2003). Although changes in behaviour are not yet unambiguously connected to
the evolution of the cranium, palaeoneurology adds weight to explanations that
link the evolution of the modern mind to the evolution of these lobes.
Comparative neuroscience, developmental neuroscience, and palaeoneurology
are useful tools, but their real strength becomes apparent only when they are
used together to explain behavioural transitions that can be documented in the
archaeological record. Christopher Henshilwood and I (Henshilwood & Dubreuil
2009, 2011), for instance, have used these different tools to link the transition to
“modern behaviours” to a change in the temporoparietal areas of the brain and
 Benoît Dubreuil

to the evolution of attentional abilities underlying high-level perspective taking

and theory of mind. This hypothesis provides a parsimonious account of the
evolution of material culture (including the appearance of increased formalism, of
symbolism, and style) and it can find support in the different tools described above.
We know, for instance, that there are significant differences in theory of mind and
perspective-taking abilities in humans and nonhuman primates and that there are
structural and functional differences in the parietotemporal areas (comparative
neuroscience). High-level theory of mind and perspective-taking abilities develop
relatively late in infancy, significantly later than joint attention and syntactic
recursion, and depend on the maturation of areas dedicated to cognitive control
and multimodal association in the prefrontal, temporal, and parietal cortices
(developmental neuroscience). Finally, the evolution of the modern cranium was
not linked to a general encephalization process, but to a specific expansion of the
temporal and parietal lobes (palaeoneurology).
Future research in archaeology and neuroscience may invalidate this hypothesis
or parts of it, but it may also strengthen it or push us toward new refinements. As a
middle-range theory, however, I would like to stress that it does not face the same
difficulty as the three hypotheses discussed in the previous section. By contrast
with them, it is not focused on a complex mental phenomenon such as theory of
mind or working memory, but on an attentional mechanism that is involved in
different tasks (including theory of mind, perspective taking, and categorisation
tasks) and that is realized in a relatively precise area of the brain, one which is
likely to have undergone recent evolution in the human lineage.

8. Conclusion and caveats

In this chapter, I have made the case that students of human cognitive evolution
should move away from the mental phenomena that they have focused on
in recent years and turn their attention to more precise cognitive mechanisms
realized in specific neural structures. As became apparent in the last section, I am
optimistic that significant progress can be made if we learn to make optimal use of
the appropriate neurological tools. In conclusion, I want to mention two reasons
for not being too optimistic about the kind of middle-range theories that can be
proposed.
The first is that moving beyond coarse middle-range theories will appear as
a lasting challenge. Archaeology and palaeoneurology, as well as comparative
and developmental neuroscience provide useful guidance, but they are consis-
tent with several different evolutionary histories in the fine-grained study of the
­cognitive and neural (not to say genetic) organisation of the brain. It is one thing,
 Chapter 10. The other middle-range theories 

for instance, to say that the temporoparietal areas are likely to have undergone
evolutionary changes along with the evolution of modern Homo sapiens, but it
is another to explain with sufficient detail what changes they underwent or how
these areas were organized in closely related human populations such as Homo
heidelbergensis or Neandertals. This problem, however, is not proper to human
evolution. Even high-resolution neuroimaging techniques are often insufficient to
detail the functional organization of specific brain areas (including the temporo-
parietal junction). We have no alternative but to learn the optimal use of available
tools, while remaining conscious of their limitations.
The second reason has to do with causality. Humans are complex organisms
who live in rich and changing social environments. Because of this, it is not
obvious how we are to determine the properly causal role of genetic, neural, and
cognitive factors in human evolution. It is one thing to infer that significant neural
or cognitive changes occur at different times in human evolution, but it is another
to say that these changes caused behavioural evolution. The problem is obviously
that, in Baldwinian evolutionary processes, changes in social environments are
also likely to cause biological evolution. As a consequence, biology cannot be
straightforwardly taken as the prime mover of human evolution. This problem
must also be considered in the wider context of human neural and cognitive
plasticity, which prevents us from concluding that causation goes from genes, to
neurons, to cognition, and to behaviour. For the time being, and maybe forever,
students of cognitive evolution should be content with establishing patchy (though
reliable) correlations between neural, cognitive, and behavioural evolution,
without claiming too much about causality.
To conclude on an additional note of caution, it should be remembered that
the validity and the relevance of middle-range theories depend on their ability to
fit archaeological data. Non-archaeologists can become quite enthusiastic about
their evolutionary scenarios and models. This is understandable, given the kind
of emotional processes that give us the motivation to engage in foundational
research, but we have to bear in mind that the validity of our scenarios and models
always depend on archaeological evidence, which is, and will for the most part
remain, open to several conflicting interpretations.
Most authors who have dealt with the evolution of the human mind have
tended to adopt very standard – not to say dominant – interpretations of the
archaeological record. Most evolutionary scenarios have been based on the
assumption that modern Homo sapiens is the only human population to have
ever evolved modern-like cognition and culture, following a cognitive and bio-
logical revolution that occurred about 200–50 ka in Africa, and was followed
by modern humans’ expansion around the globe. Although this interpretation
is supported by much of the data, we must remain conscious of the numerous
 Benoît Dubreuil

uncertainties that remain to be clarified, including many that might have a deci-
sive impact on the middle-range theories that we can devise in order to link
neural, cognitive, and behavioural data. Did Neandertals use symbolic artefacts
and in what context? Did “modern behaviour” in Africa appear as a package or
gradually? Are early occurrences of ochre in the Middle Stone Age evidence of
symbolic behaviour? Is the modernisation of Homo sapiens’ cranium concomi-
tant with the emergence of modern sapiens behaviour, whatever that means?
Although we all have our preferred answers to these questions, we must admit
that they remain, for the most part, open. In this context, it might not be a bad
idea to develop not one, but many middle-range theories to link neural, cogni-
tive, and behavioural evolution, in order to test them afterwards against new
archaeological findings.

Acknowledgements

I thank Luc Faucher, as well as the participants at the symposium Homo ­symbolicus in Cape
Town, for comments on an earlier version of this chapter. I am grateful to the Templeton
Foundation for the invitation to the symposium and to Chad Horne for comments and
editorial support. This work was supported by a postdoctoral fellowship from the Fonds
québécois de la recherche sur la société et la culture.

References

Baddeley, A.D. 2001. “Is working memory still working?” American Psychologist 56 (11):
851–864.
Baddeley, A.D. & Hitch, G. 1974. “Working memory.” In The psychology of learning and
­motivation: Advances in research and theory, G. H. Bower (ed.), 8: 47–89. New York:
­Academic Press.
Bechtel, W. 2008. Mental mechanisms: philosophical perspectives on cognitive neuroscience. New
York: Routledge.
Bickerton, D. 2003. “Symbol and structure: A comprehensive framework for language
evolution.” In Language evolution, M. H. Christiansen & S. Kirby (eds), 77–93. Oxford:
Oxford University Press.
Binford, L.R. 1977. “Introduction.” In For Theory Building in Archaeology: Essays on Faunal
Remains, Aquatic Resources, Spatial Analysis, and Systemic Modeling, L.R. Binford (ed.),
1–10. New York: Academic Press.
Binford, L.R. 1981. Bones: Ancient Men and Modern Myths. New York: Academic Press.
Bruner, E.M. G., Manzi, G. & Arsuaga, J.L., 2003. “Encephalization and allometric trajectories in
the genus Homo: Evidence from the Neandertal and modern lineages.” Proceedings of the
National Academy of Sciences 100: 15335–15340.
Call, J. & Tomasello, M. 2008. “Does the chimpanzee have a theory of mind? 30 years later.”
Trends in Cognitive Science 12 (5): 187–192.
 Chapter 10. The other middle-range theories 

Calvin, W.H. & Bickerton D. 2000. Lingua ex machina: reconciling Darwin and Chomsky with the
human brain. Cambridge, Mass.: MIT Press.
Carlson, S.M., Mandell, D.J. & Williams, L. 2004. “Executive function and theory of mind:
­stability and prediction from age 2 to 3.” Developmental Psychology 40: 1105–1122.
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. 2004. “A cognitive and neuropsychological perspective on the
­Chatelperronian.” Journal of Anthropological Research 60: 55–73.
Coolidge, F.L. & Wynn, T. 2005. “Working memory, its executive functions, and the ­emergence
of modern thinking.” Cambridge Archaeological Journal 15 (1): 5–26.
Coolidge, F. & Wynn, T. 2007. “The working memory account of Neandertal cognition-How
phonological storage capacity may be related to recursion and the pragmatics of modern
speech.” Journal of Human Evolution 52 (6): 707–710.
Corballis, M.C. 2004. “The origins of modernity: Was autonomous speech the critical factor?”
Psychological Review 111: 543–552.
Costa, A., Torriero, S., Olivieri, M. & Caltagirone, C. 2008. “Prefrontal and temporo-parietal
involvement in taking others’ perspective: TMS evidence.” Behavioral Neurology 19
(1–2): 71–74.
Craver, C.F. 2007. Explaining the brain: mechanisms and the mosaic unity of neuroscience. New
York: Oxford University Press; Oxford: Clarendon Press.
De Villiers, J.G. & Pyers, J.E. 2002. “Complements to cognition: a longitudinal study of
the relationship between complex syntax and false-belief understanding.” Cognitive
Development 17: 1037–1060.
Deacon, T.W. 1997. The symbolic species: the co-evolution of language and the brain. New York:
W.W. Norton & Co.
Decety, J. & Lamm, C. 2007. “The Role of the Right Temporoparietal Junction in Social
Interaction: How Low-Level Computational Processes Contribute to Meta-Cognition.”
Neuroscientist 13: 580–593.
Dubreuil, B. 2008. “What do modern behaviors imply for the evolution of language?” In The
Evolution of Language, A.D.M. Smith, K. Smith & R. Ferrer i Cancho (eds), 99–106.
Proceedings of the 7th International Conference (Evolang 7), Singapore: World Scientific.
Dunbar, R.I.M. 2003. “The social brain: mind, language and society in evolutionary per­
spective.” Annual Review of Anthropology 32: 163–181.
Dunbar, R.I.M. 2004. The human story: A new history of mankind’s evolution. London: Faber
and Faber.
Dunbar, R.I.M. 2007. “The Social Brain and the Cultural Explosion of the Human Revolution.”
In Rethinking the Human Revolution, P. Mellars, K. Boyle, O. Bar-Yosef and C. Stringer.
(eds), 91–98. Cambridge: McDonald Institute Monographs.
Edin, F., Klingberg T., Johansson, P., McNab, F., Tegnér, J. & Compte, A. 2009. “Mechanism for
top-down control of working memory capacity.” Proceedings of the National Academy of
Sciences 97 (7): 3573–3578.
Friederici, A.D. 2009. “Pathways to language: fiber tracts in the human brain.” Trends in Cognitive
Sciences 13 (4): 175–181.
Gerrans, P. & Stone, V.E. 2008. “Generous or Parsimonious Cognitive Architecture? Cognitive
Neuroscience and Theory of Mind.” British Journal for the Philosophy of Science 59
(2): 121–141.
 Benoît Dubreuil

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.
Henshilwood, C.S. & Dubreuil, B. 2009. “Reading the artifacts: Gleaning language skills from
the Middle Stone Age in southern Africa.” In The Cradle of Language, Volume 2: African
perspectives, R. Botha & C. Knight (eds), 41–61. Oxford: Oxford University Press.
Henshilwood, C.S. & Dubreuil, B., 2011. “The Still Bay and Howiesons Poort, 77–59 ka:
­Symbolic material culture and the evolution of the mind during the African Middle Stone
Age.” C ­ urrent Anthropology. 52 (3): 361–400.
Jurado, M.B. & Rosselli, M. 2007. “The elusive nature of executive functions: a review of our
current understanding.” Neuropsychological Review 17 (3): 213–233.
Klein, R.G. 2008. “Out of Africa and the evolution of human behavior.” Evolutionary Anthropology
17 (6): 267–281.
Koechlin, E. & Jubault, T. 2006. “Broca’s area and the hierarchical organization of human
­behavior.” Neuron 50 (6): 963–974.
Krause, J., Lalueza-Fox, C., Orlando, L., Enard, W., Green, R., Burbano, H., Hublin, J–J., Ha¨nni,
C., Fortea, J., de la Rasilla, M., Bertranpetit, J., Rosas, A. & Pa¨a¨bo, S. 2007. “The derived
FOXP2 variant of modern humans was shared with Neandertals.” Current Biology 17: 1–5.
Lieberman, P. 2007. “The Evolution of Human Speech: Its Anatomical and Neural Bases.”
­Current Anthropology 48 (1): 39–66.
Lieberman, D.E., McBratney, B.M. & Krovitz, G. 2002. “The evolution and development of c­ ranial
form in Homo sapiens.” Proceedings of the National Academy of Sciences 99: 1134–1139.
Marcus, G.F. 2008. Kluge: The haphazard construction of the human mind. Boston: Houghton
Mifflin.
Mellars, P. 2005. “The impossible coincidence. A single-species model for the origins of modern
human behavior in Europe.” Evolutionary Anthropology 14: 12–27.
Mitchell, S.D. 2003. Biological complexity and integrative pluralism. Cambridge, UK: New York,
N.Y.: Cambridge University Press.
Mithen, S. 1996. The Prehistory of the Mind. London: Thames & Hudson.
Read, D.W. 2008. “Working memory: A cognitive limit to non-human primate recursive
­thinking prior to hominid evolution.” Evolutionary Psychology 6 (4): 676–714.
Rightmire, G.P. 2004. “Brain size and encephalization in Early to Mid-Pleistocene Homo.”
American Journal of Physical Anthropology 124: 109–123.
Rilling, J.K., Glasser, M.F., Preuss, T.M., Ma, X., Zhao, T., Hu X. & Behrens, T.E. J. 2008. “The
evolution of the arcuate fasciculus revealed with comparative DTI.” Nature Neuroscience
11: 426–428.
Saxe, R. & Kanwisher, N. 2003. “People thinking about thinking people: the role of the
­temporoparietal junction in theory of mind.” NeuroImage 19: 1835–1842.
Schaffner, K. 1998. “Genes, behavior, and developmental emergentism: One process, n ­ divisible?”
Philosophy of Science 65: 209–252.
Schenker, N.M., Desgouttes, A.M. & Semendeferi, K. 2005. “Neural connectivity and cortical
substrates of cognition in hominoids.” Journal Human Evolution 49: 547–569.
Schoenemann, P.T. & Sheehan, M.J. 2005. “Prefrontal white matter volume is disproportionately
larger in humans than in other primates.” Nature Neuroscience 8: 242–252.
Semendeferi, K., Lu, A. Schenker, N. & Damasio, H. 2002. “Humans and great apes share a large
frontal cortex.” Nature Neuroscience 5 (3): 272–276.
Stout, D. & Chaminade, T. 2009. “Making tools and making sense: complex, intentional
­behaviour in human evolution.” Cambridge Archaeological Journal 19 (1): 85–96.
 Chapter 10. The other middle-range theories 

Suddendorf, T. & Corballis, M.C. 2007. “The evolution of foresight: What is mental time travel
and is it unique to humans?” Behavioral and Brain Sciences 30 (3): 299–313.
Tomasello, M. 2008. Origins of human communication. Cambridge, Mass.: MIT Press.
Ungerleider, L.G., Courtney, S.M. & Haxby, J.V. 1998. “A Neural System for Human Visual
Working Memory.” Proceedings of the National Academy of Sciences 95 (3): 883–890.
Wynn, T. & Coolidge, F.L. 2004. “The skilled Neanderthal mind.” Journal of Human Evolution
46: 467–487.
Wynn, T. & Coolidge, F.L. 2007. “Did a small but significant enhancement in working-memory
capacity power the evolution of modern thinking?” In Rethinking the Human Revolution,
P. Mellars, K. Boyle, O. Bar-Yosef, & C. Stringer (eds), 79–90. Cambridge: McDonald
Institute Monographs.
Zilhão, J. 2006. “Neandertals and Moderns Mixed, and It Matters.” Evolutionary Anthropology
15: 183–195.
 chapter 11

Metarepresentation, Homo religiosus,

and Homo symbolicus
Justin L. Barrett
Centre for Anthropology & Mind Institute of Cognitive & Evolutionary
Anthropology, University of Oxford

What cognitive adaptation enabled humans to become the distinctively symbolic

species that we are? Drawing upon insights from the cognitive sciences and
evolutionary psychology, research in cognitive science of religion (CSR)
converges on the claim that the ability to form mental representations about
mental representations (metarepresentation) is a key factor enabling and
encouraging religious expression. Such metarepresentation may also be the key
to symbolic behaviour – including linguistic expression – more generally. If so,
then the same cognitive equipment that underwrites symbolism also gave rise to
religion, and the two could have evolved concurrently.

1. Introduction

Cognitive science of religion (CSR) has drawn from evolutionary psychology and
cognitive development to generate a number of theories regarding the ­cognitive
architecture necessary for various religious ideas and practices. A review of
these theories suggests that one common factor that propels proto-religious
thoughts and impulses into the full-blown religion that we see in modern
humans is m ­ etarepresentation – the ability to form mental representations of
mental representations. For instance, it may be that basic belief in the afterlife
is a byproduct of two different cognitive systems with different evolutionary
and developmental histories – one that deals with bodies and one that deals
with minds. But m ­ etarepresentation may be necessary for afterlife beliefs to
transform into a­ ncestor-spirit cults. Similarly, an agency detection device that
is tuned to be hypersensitive may lead to the postulation of intentional beings
with supernatural properties, but metarepresentation may be required to support
beliefs and practices of interactions with these supernatural beings.
Metarepresentation may likewise serve as the critical factor in transforming
other cognitive capacities into more sophisticated systems that can sustain cumu-
lative cultural traditions including symbolic ones, and may be a distinctive feature
of modern humans.
 Justin L. Barrett

In this essay I begin by considering current theories of the evolution of

religious cognition in humans and argue for the importance of metarepresentation
as a critical capacity in religious thought. In the following section I extend the
discussion to other cultural domains in which metarepresentation seems to
play a key role, including cultural expression that appears to require symbolism.
I conclude by discussing what the proposed account of religious and symbolic
evolution might mean for archaeological studies.
Throughout I adopt the standard practice in CSR of treating ‘religion’ as ideas,
practices, and commitments that are distributed across individuals, which pertain
to the affirmed existence of at least one form of minded agency that deviates to
some degree from the natural intuitions delivered by ordinary human cognitive
systems. That is, religion concerns counterintuitive intentional agency. Such a
definition is close to the common view that ‘religion’ has something to do with
gods, spirits, ghosts, and the like. I recognize that such a definition carries the
liability of being narrower than the myriad phenomena scholars may include
under the ‘religion’ umbrella. It carries the strength, in this occasion, of picking
out a set of phenomena that may be causally and explanatorily coherent from a
cognitive perspective. I hope this working definition and its virtues in this context
will become clearer below.

2. Metarepresentation and religious evolution

Metarepresentation, the ability to deliberately consider, ponder, or evaluate

mental representations, that is, thinking about thoughts, may be the lynch-pin
that transforms many other cognitive capacities into those attributes that consti-
tute human distinctiveness. Current scholarship in evolutionary and cognitive
studies of religion recurrently implicates or presumes a critical role for metare­
presentation in explaining the cross-cultural recurrence of religious beliefs and
practices.
Darwinian evolution raises at least four alternative possibilities concerning
when and how religion might have arisen, when Homo might have become Homo
religiosus: (1) the Standard View, that there is nothing in human nature that
predisposes us or specially accounts for religious thought and behaviour; (2) the
Adaptation Hypothesis that religion is an evolved trait on top of and later than
the evolution of whatever it is that makes us behaviourally modern humans; (3)
the Pre-Human Religion Hypothesis that first religion evolved as an adaptation
and then later behaviourally modern humans evolved; and (4) the Concurrence
 Chapter 11. Metarepresentation, Homo religiosus, and Homo symbolicus 

Hypothesis, that whatever it is that makes us behaviourally modern humans is the

same thing that makes us religious.1

3. The Standard View

The Standard View is the view that there is nothing about human nature that
makes us prone to be religious or to believe in gods. Religion is something of a
historical-cultural accident analogous to the invention of modern science or the
game of chess. Though we are capable of these forms of cultural expression, they
might have never happened. Had someone not invented chess, there would not
be any chess players. Had no one invented religion, people wouldn’t be religious.
On this view, religion does not have any special place in human nature and it
surely is not an adaptation. The Standard View is rooted in the idea that human
cognition provides a bias-free clean slate for experience – including cultural expe-
riences – to write on unhindered. The environment largely dictates who and what
we are. In some of their writings, Richard Dawkins (Dawkins 2006) and Nicholas
Humphrey (Humphrey 1998) seem to express this Standard View, suggesting that
if children were not indoctrinated to believe in religion it would disappear.2
I regard the Standard View as problematic. Its presumption that human minds
can be likened to blank slates has been demonstrated to be erroneous many times
over. It is simply not the case that humans can learn any information equally well,
or are equally likely to acquire any ideas or practices they might be exposed to.
Such a simplistic view of human cognition and behaviour, though still
common in neighbouring fields, has held no serious place in psychological
sciences since the so-called cognitive revolution that began in the 1950s and by
the 1970s had overthrown behaviourism’s empiricist regime. Though there is still
lively disagreement about just what sorts of biases human minds naturally have,
how they develop, and so forth, that ordinary human biological endowment plus
ordinary features of the world yields a predictable, natural mental tool kit, is not

. Strictly speaking, it could be that becoming religious and becoming modern humans
­occurred simultaneously due to the co-evolution of two different traits. That is, religion might
be an adaptation that happened to evolve at the same time as other adaptations that make
modern humans what they are. For the sake of this presentation, I consider such a possibility
as a limiting case of both the Adaptationist and the Pre-Human Religion Hypotheses.
. Dawkins (2006), Chapter 9, especially seems to suggest that the only critical disposition
toward religious belief is evolved gullibility. In other parts of his book such as Chapter 5,
however, he does take a more receptive stance toward religion having some natural anchors.
 Justin L. Barrett

controversial. Rather than a sponge that soaks in anything or a blank slate ready
for any scribbling, the human mind is more usefully likened to a Swiss Army knife
possessing numerous functional units for solving particular problems.3
The Standard View also looks suspect given some striking divergences from
other ‘arbitrary’ non-natural forms of cultural expression such as modern science
or playing chess. Cultural expression that is largely unbuckled from natural
biases or predilections tends to be extremely variable within and across cultures.
So people and some cultures simply do not do modern science or play chess. In
contrast, religious thought and action-perhaps particularly belief in some kind
of supernatural agency – is ubiquitous across cultures (Atran 2002). But it isn’t
just the greater recurrence of religious thought (as compared with chess playing
or modern science) that suggests it is a different kind of cultural expression. We
might also point to its relative ease of transmission even to very young children,
and its resilience in the face of deliberate attempts to eliminate it.
Arbitrary cultural forms tend to be acquired later in life than the more
naturally anchored sort of cultural expression. Doing algebra or modern science,
playing chess or riding a bicycle all are forms of cultural expression that generally
arrive late enough in our lifespan (if at all) that we remember not doing them.
Walking, speaking, and entertaining religious thoughts, in contrast, can be picked
up so early in childhood that we might not remember not doing them. Related,
arbitrary cultural expression generally requires explicit instruction to acquire.
Intensive training is usually required to do science, whereas doing religion appears
to be almost caught like the common cold. We can just pick it up from observing
those around us, much like how we just pick up our native language.
These sorts of considerations regarding the resilience, the ubiquity, and the
ease with which even young children acquire religion has prompted many scholars
to explore why it is that humans seem prone to religious expression (Atran 2002;
Barrett 2000, 2004; Boyer 2003; Pyysiäinen 2001; Tremlin 2006). Why is it so
common? Why is it so easy for children to acquire? In short, why does it seem that
religion is a part of human nature? That scholars are making progress on these
sorts of questions is evidence in itself that the Standard View is mistaken.

4. The Adaptation Hypothesis

The Adaptation Hypothesis is that religious belief and practice amount to an

evolved propensity that arose after our ancestors had become behaviourally

. Steven Pinker gives a good overview of this more content-rich view of the human mind
(Pinker 1997).
 Chapter 11. Metarepresentation, Homo religiosus, and Homo symbolicus 

modern humans. That is, religion is something extra added that was added because
it helped our ancestors survive and reproduce better than competing modern
humans lacking religion.
Adaptationist arguments tend to focus on intra-group cooperation in one of
two ways: the ability for morally interested supernatural agents to police human
action in social arrangements (Bering & Johnson 2005), and the role of rituals in
stimulating intra-group trust (Alcorta & Sosis 2005; Wilson 2002).4 Both types of
argument begin with some common premises. It is an advantage for members of
a social group to share and trade. Sometimes our hunting, gathering, or farming
is more successful than other times. If we share with each other then we level out
boom-and-bust cycles for everyone. Likewise, if I am a great hunter but you are
a great clothing maker, it might be best for both of us for me to give you meat in
exchange for clothing. But such resource sharing typically requires a degree of
trust. Will you really give me meat if I fail to find game when you are successful?
Or will you hide what you caught and not tell me about it? Or knowing that you
are better at gathering than I am and a generous sharer, why should I work so
hard? Maybe I should do just enough work so you don’t think I am a free-loader.
Cooperation has potential costs, so trust is important. I need to know that you
won’t cheat me.
One way of handling the free-rider problem is to punish cheaters when
they are caught. The downsides are that you have to expend energy catching and
punishing cheaters, and cheaters could simply get better at not getting caught.
Further, some people might not punish cheaters because it is too costly to them
so then non-punishers would have to be punished for not punishing and so on.
Enter the gods.
Dominic Johnson and Jesse Bering have suggested that gods help arrest
this punishment regress (Bering & Johnson 2005). If gods do the punishing,
we humans don’t have to do so much. Another up-side of gods policing human
behaviour is that gods – by sake of being invisible, mind-reading, all-knowing,
omnipresent, or able to see through walls – know with greater fidelity who is
a cheat and who isn’t. They have what Pascal Boyer has called full access to
strategic information about humans (Boyer 2001, 2003). People that believe in a
morally interested, full-access agent such as a god are less likely to cheat on you.

. Harvey Whitehouse has considered religious ritual participation as a means for forming
coalitional commitment and trust with highly emotional, rarely performed rituals (such as
initiation) leading to small, intensely cohesive groups, and repetitive, lower-arousal rituals
leading to larger communities (Whitehouse 2000; Whitehouse 2004). Note, however, that
Whitehouse does not regard these dynamics as evidence of religion being an adaptation, but
emphasizes its role in cultural evolution.
 Justin L. Barrett

So, they are more reliable sharing and trading partners and more people will
want to share and trade with them. Hence, they have better access to resources
than less trustworthy individuals. This resource advantage will make them fitter
to survive and reproduce, and so, their genes that make them god-believers will
be reinforced through natural selection.
Similarly, it may be that even aside from policing supernatural agents, the
existence of a transcendent realm or of gods that our group recognizes can serve
as the hub for ritual performances. These ritual performances, because they
do not obviously have the direct utility of ordinary actions, can serve to signal
commitment to the in-group. One’s participation in a ‘costly’ ritual display may
signal to others that one is a true believer in the group’s values and practices
(Alcorta & Sosis 2005; Ruffle & Sosis 2007; Sosis 2003). These signalling displays
may then encourage trust and cooperation, leading to fitness benefits. The
underlying gene-culture complex then would be selectively reinforced. A variation
of the Adaptation Hypothesis has been advanced by David Sloan Wilson (Wilson
2002). He argues that religious groups out compete non-religious groups, and so
the genes of groups with a propensity toward religious belief and action will be
selectively promulgated.
Two sorts of evidence would be helpful in supporting variants of the Adapta-
tion Hypothesis: evidence that there is a particular gene that promotes religiousness
and, more importantly, evidence that being religious really is a­ daptive. Is there such
evidence?
On the first score-is there a religion-or god-gene?-, sensational titles and
headlines aside, the evidence to date is pretty thin. Certainly there appears some
heritable component in what has been labelled ‘religiosity’, but whether this is a
disposition toward believing in gods or ritualism or in-group conformity or some
other feature of modern religiosity in contemporary western societies is unclear.
In his book, The God Gene, Dean Hamer admits that the gene that he has identified
that is related to some aspects of religious identification manages to explain only a
very small fraction of the variance (Hamer 2004). That is, this ‘god gene’ is a long
way from accounting for human religious propensities. So far then, evidence for
a particular genetic adaptation that makes people religious (who otherwise would
not be) is far from compelling.
The Adaptation Hypothesis does much better with regard to the second type of
evidence. We do have reason to believe that being religious is adaptive, particularly
that it may promote group cohesion and cooperation. Anthropologist Richard
Sosis has produced several studies showing that religious communities often out-
survive non-religious ones, and are marked by higher levels of cooperation. For
instance, in one study Sosis compared religious and secular kibbutz communities
in Israel. Members of religious kibbutzim had higher levels of trust on economic
 Chapter 11. Metarepresentation, Homo religiosus, and Homo symbolicus 

games, particularly if they were regular participants in religious ­rituals (Sosis &
Ruffle 2003). Likewise, religious communities appear to have greater survival as
communities than comparable secular ones, particularly if there are ‘ritualized’
behaviours associated with membership (Sosis & Bressler 2003). Psychologists
have recently shown that Belgian university students who were subliminally
exposed to religious words acted with more pro-social intention (picking up more
brochures to publicize a charitable organization) and were more forgiving of a
harsh critic than students not so primed (Pichon et al. 2007; Saroglou et al. 2009).
Suppose we accept that belief in gods and the actions that these beliefs motivate
are adaptive, it does not follow that religion is an adaptation. Lots of things are or
could be adaptive without being adaptations. Yoga may reduce stress and improve
fitness, thereby promoting survival and reproductive success, but surely yoga is not
a direct product of genetic selection. Having community fire brigades or rescue
corps may improve the fitness of a community but they surely are not adaptations.
Even something as old and as widespread as clothing making-clearly adaptive for
living in cold climates – is an implausible candidate for an adaptation. The idea
that behaviours such as clothing making are the result of a genetic mutation when
our ancestors already possessed the ability to use tools, detect patterns, reason
causally, and solve problems, seems implausible and unparsimonious. Rather,
clothing-making was likely an innovative product of other capacities that then
has been passed on through cultural transmission. Similarly, it seems unlikely that
all that captured by ‘religion’ here is the result of a genetic adaptation. Consider
what that would mean. Belief in gods, belief in the afterlife, reasoning as if gods
are morally interested in the activity of humans and that they act on that interest
to punish or reward, seeing natural phenomena as indicators of divine action, and
socially organizing around these beliefs all would be products of the same genetic
modification under the same selection pressure.

5. Possible but improbable

If the selective pressure was toward intra-group trust and cooperation for sharing
resources, why not a much simpler mutation that made us all hyper-vigilant
do-gooders? Or how about some kind of gentle paranoia that somehow whenever
we are bad someone is bound to catch us? Introducing religion to solve this
cooperation problem seems a little like trying to kill a gnat with a hand grenade.
For these reasons, though religion may be adaptive, I suspect we cannot conclude
that all of religion is an adaptation. Nevertheless, the possibility that a host of
adaptations providing fragmentary preludes to religion were conjoined by a single
modest adaptation remains plausible. I consider such a possibility below.
 Justin L. Barrett

6. The Pre-human Religion Hypothesis

The Pre-human Religion Hypothesis is that at some point before becoming

modern humans, our ancestors acquired religion as an adaptation through natural
selection. Then, at a later date, they acquired additional adaptations that made
them behaviourally modern humans. Some archaeologists argue that Neandertals
were religious (Tattersall 1998). This claim is based on evidence that Neandertals
occasionally buried their dead with grave goods in a way that suggested ritualized
treatment of the remains and (perhaps) a belief in the afterlife. Maybe some
Neandertals did have a belief in the afterlife that motivated the seemingly
“ritualized” burial practices, or perhaps they simply regarded the bodies and
possessions of deceased loved ones as somehow special and uncanny without any
worked out ideas about what happens after death. Either way, we don’t have strong
evidence that Neandertals had the kind of fully-developed religious beliefs and
practices that modern humans have been expressing for thousands of years.
Another non-human species that has been cited as possibly having some kind
of religious sentiment is the chimpanzee. Male chimpanzees have been known to
make dominance displays of the sort they use in confrontation with other males in
response to waterfalls, thunderstorms, and even motorized vehicles (Whiten et al.
2001). Does this mean chimpanzees are pan-omorphizing the natural world much
like humans anthropomorphize the natural world in religious contexts? It is hard
to know. It could be that chimpanzees have a behavioural routine (the dominance
display) that usually gets triggered only in the presence of competing males
but occasionally natural (or human made) sights and sounds satisfy the ‘input
conditions’ to get these routines going. Likely, these ‘rain dances’ as they have been
called, are more similar to horses running from harmless plastic grocery sacks
blowing in the wind than to humans regarding hurricanes as divine judgement
for misdeeds.
Further, as an adaptationist account, the Pre-human Religion Hypothesis also
shares the same liabilities as the Adaptation Hypothesis, and so, for lack of evi-
dence and plausibility, I suspect that the Pre-human Religion Hypothesis is false.

7. The Concurrence Hypothesis

The Concurrence Hypothesis suggests that whatever it was that made behav-
iourally modern humans, happened to appear at the same time as whatever it
was that gave humans a religious disposition. More precisely, those capacities
that prompt us to be religious are the same capacities that make modern humans
cognitively and behaviourally distinct. Becoming modern humans and becoming
Homo religiosus are one and the same. On the surface, this seems an extremely
 Chapter 11. Metarepresentation, Homo religiosus, and Homo symbolicus 

unlikely proposition, but let me try to persuade you that the Concurrence Hypoth-
esis is a real possibility and perhaps even a more likely possibility than the other
hypotheses I offered above.
To build my case, I will quickly sketch three accounts for why humans are
naturally disposed to be religious developed by cognitive scientists of religion.
What these accounts have in common is that instead of proposing special genetic
mutations that give rise to religiousness, they observe that features of ordinary
human cognitive systems that might have evolved for entirely different reasons in
our ancestral past may encourage the religiousness of modern humans.

8. Anthropomorphism, agency detection, and gods

The cognitive area’s earliest account, developed by anthropologist Stewart Guthrie,

is that because of the importance of detecting predators and prey, friends and foe,
we have evolved a tendency to over-detect human-like beings in our environment.
That is, we have a tendency to anthropomorphize objects and states of affairs in the
natural world. Guthrie argues that such a tendency leads to postulating minded,
non-human agents as being responsible for various events in the environment;
beings more commonly known as ‘gods’.
Considerable research has been devoted to understanding the human Agency
Detection Device (ADD) as it has sometimes been called (Barrett 2004). This
ADD is regarded as being evolved to detect the presence of minded, intentional
beings even under conditions in which evidence is scant or inconclusive. The
­evolutionary argument for this hair-trigger detection is that the cost of failing
to detect a predator or prey, a friend or foe, would be so great that the ADD
is b
­ etter tuned to generate false positives-occasionally detecting agents when
there are none-than false negatives-missing agents when they are around. And
certainly experimental evidence suggests that infant and adult humans can be
led to regard even geometric shapes as minded entities (Rochat et al. 1997). All it
takes are c­ ertain seemingly non-inertial, goal-directed movements such as when
an object launches itself toward another object. So the Agency Detection Device
may be hypersensitive under certain conditions, hence it is sometimes called the
Hypersensitive Agency Detection Device or HADD.
But this tendency to find goal-directed agents in the environment readily and
rapidly is a trait shared by many non-human species. Mammals and birds can
readily be tricked into acting as if a non-living object such as a blowing plastic
bag or a remote-controlled vehicle is some kind of agent. Maybe they too have
some kind of primitive HADD, but this kind of HADD, that only detects objects
as potential agents, is insufficient for supporting a relationship with gods. HADD
must also register states of affairs as the products of agency even when these states
 Justin L. Barrett

are not the agents. That is, HADD must see sounds or traces or movement of
objects as intentionally brought about.
If I wander through an allegedly haunted house and I hear chains rattling,
I might regard the sound of chains as produced by a ghost. The chains are the prod-
uct of intentional action but are not the agent. Without this capacity to perceive
states of affairs as intentionally caused by an unseen agent, people could not regard
events such as thunderstorms or good crops as the gift or curse of a god. Of course,
modern humans do have such a capacity and take advantage of it to interact with
other people who are not immediately present.
This sort of agency detection is much less clear in non-humans. Perhaps dogs
regard the sound of their unseen human masters’ whistle as evidence of intentional
agency, but perhaps they more simply develop a behavioural response to the w ­ histle
that brings reward. Perhaps animals are shy of human buildings and roads not
because they understand them to be signs of human agency, but perhaps instead
they just learn to associate these strange things with danger and do not entertain the
contents of human mental states.5 Even if we give non-human animals the ­benefit
of the doubt in these situations, humans can do something more complex still. Not
only do we regard states of affairs as products of the intentional activity of an unseen
agent, but then speculate what the intentions of the unseen agent might have been.
Solid evidence that any non-humans have these cognitive capacities is yet to be
found. Even if members of Pan and extinct Homo had such capacities, these are
still insufficient to be counted as Homo religiosus. The ability to suppose that the
­thunderstorm was caused by a god and even perhaps because the god is angry is not
quite sufficient for religion. Something else is needed. I’ll turn to the something else
after sketching more cognitive accounts of religious predisposition.

9. Intuitive dualism and spirits

Another cognitive account is offered by Yale psychologist Paul Bloom (Bloom 2004;
Bloom 2007). He notes that developmental psychology has produced e­ vidence
that from early in childhood we have two different conceptual systems that both

. Perhaps the best evidence are reports that chimpanzees break branches or twigs to signal
their travel route to other chimpanzees who then successfully use those signs. But even here
it is possible to regard these behaviours on the part of chimpanzees in another way, however.
Perhaps the following chimpanzees do not see the ‘signs’ as intentionally produced by other
chimpanzees but have developed a behavioural routine such that they simply follow these
markers to find other chimpanzees much like following the tracks or droppings of an animal
or even an object that might have rolled away.
 Chapter 11. Metarepresentation, Homo religiosus, and Homo symbolicus 

have to be used to make sense of the behaviour of humans. One system deals
with the properties and motion of physical bodies. And one system, called ‘theory
of mind’, tries to explain and predict behaviour based on mental states: beliefs,
desires, goals, feelings, dispositions, and so forth. The theory of mind system (or
ToM) is much newer in terms of both evolution and development. These two
systems have to be held together to make sense of human behaviour, but their
tenuous relationship leads to a sort of intuitive dualism. That is, we tacitly assume
there is some minded part of an individual that is separable from the body. Given
this intuitive dualism, the idea that some part of a person can continue existing
and even acting after death is natural, says Bloom, and so the slightest encouraging
evidence (perhaps provided by HADD) or corroboration from trusted others may
be enough to get afterlife beliefs up and running.6 Similarly, Pascal Boyer has
proposed that human corpses are perceived in a unique way because of conflicts
between several specialized inference systems – notably the animacy system and
the person-file that holds details pertaining to a specific known person. Seeing the
corpse of a family member prompts uneasiness because the animacy system tells
us that the person is not living but the person-file (with the aid of ToM) refuses to
stop producing inferences about the dead person’s wants, needs, and beliefs (Boyer
2001). Bloom’s and Boyer’s accounts help explain why the idea that some human
spirit survives death and continues to act in the world may be the most widespread
and oldest religious concept.
But only believing that the ancestors can continue to act in the world and
speculating about their goals, wants, and beliefs does not get us all the way to
religion. Something else is still needed, but before I get to that something, allow
me one more set of cognitive dynamics that may encourage belief in gods.

10. Naïve creationism and intuitive theism

Psychologist Deborah Kelemen has suggested that children may be intuitive theists
because of a different set of cognitive dynamics (Kelemen 2004). She has conducted
dozens of experiments with children and adults that show a strong tendency to
see natural things such as rocks, plants, and animals as purposefully designed.
Her research and that of others has also shown that young children assume that

. Psychologist Jesse Bering has argued something similar, noting the great difficulty people
have in imagining their own lack of all epistemic states. This inability to simulate not having
any thoughts or consciousness at all may encourage belief in afterlife, argues Bering. He calls
this account the simulation constraint hypothesis (Bering 2006).
 Justin L. Barrett

purposeful design requires a designer – indeed, until Darwin essentially everyone

assumed design requires a designer – and so children are prone toward believing
in some kind of intelligent being or beings that account for the natural world.
Children quickly learn that humans lack the ability to account for the apparent
design in the natural world, so children are prepared to believe the suggestion that
some kind of god or gods account for the natural world.
These three accounts-Guthrie’s anthropomorphism account, Bloom’s intuitive
dualism account, and Kelemen’s intuitive theism account – each have evidence in
their favour and they could be combined. For instance, intuitive dualism might lead
us to suppose that some spirit continues after death, the Hypersensitive Agency
Detection Device might then encourage us to detect evidence of that spirit’s activity,
and the tendency to see the world as designed might encourage us to believe that
the ancestor-spirits create and sustain the natural world. Other cognitive factors,
too, have been scientifically developed that might buttress religiousness (Barrett
2007). These three accounts all make use of ordinary cognitive architecture that
likely evolved for reasons entirely unrelated to religious belief and practice, and
indeed, likely evolved, at least in large part, in pre-human species. We would not
be too surprised if Homo erectus had a Hypersensitive Agency Detection Device,
some kind of basic ToM system that enabled speculation about the mental states
of others and leading to intuitive dualism, and even a bias to understand the
natural world in terms of purpose and design, perhaps leading to speculations that
someone intentionally caused elements of the natural world to be what they are.
All three cognitive accounts, as I have presented them so far, also lack the
same extra factor that would produce what we identify as religion in humans. If
Homo erectus did have the cognitive capacities identified above, Homo erectus was
still not Homo religiosus.

11. Metarepresentation and Homo religiosus

Recall that to be Homo religiosus what is needed is to be able to have a joint rela-
tionship with a god or gods. Homo religiosus must be able to act contingently in
response to what they think the god’s mental states might be, and they must be
able to believe that another shares a similar relation with the same god, otherwise
we would have lots of individual, idiosyncratic beliefs and practices instead of faith
communities.
The lynch-pin that makes this kind of relational religion possible is the
cognitive capacity known as higher-order ToM or, as I prefer, metarepresenta-
tional ToM: the ability deliberately and reflectively to think about the contents
of other’s thoughts. I call this ‘metarepresentational ToM’ because it is the ability
 Chapter 11. Metarepresentation, Homo religiosus, and Homo symbolicus 

to form mental representations about the contents of mental representations.

One could think about another’s thoughts about thoughts, including my own
thoughts (e.g. “She knows that I know she knows what I want”). Metarepresen-
tational ToM opens new doors for thinking about and relating to gods above and
beyond ­simpler ToM abilities. Consider the following:

a. The god has mental states (but I don’t know what they are) that motivate the
god’s actions.
b. The god does not like my behaviour and expressed that dislike by sending
illness.
c. The god does not like my behaviour and sent illness so that I would know that
the god does not like my behaviour and I would change.
d. You know that the god does not like my behaviour and sent illness so that
I would know that the god does not like my behaviour and would change.

Sentence (a) expresses a simple awareness that others have mental states that
­ atter in some way to their actions. This represents the most basic mentalizing
m
activity and seems to be a capacity that infants and perhaps non-human primates,
dogs, and elephants may possess.
Sentence (b) expresses an understanding of the contents of another’s mental
states and how those states are connected to an action. Two year olds and apes may
have this ability.
Sentence (c) expresses awareness that other’s mental states may include
­representations of my mental states, as in ‘He wants me to know…”, and some
sense that actions can be motivated by a desire to change other’s mental states and
not merely their behaviours. Now we are in range of drawing complex relation-
ships among mental states and their relations to actions. It is not clear that any
non-human species alive today have this capacity or that humans exhibit it before
age three or four.
Sentence (d) takes things one step further still. All that is expressed in (c) is now
attributed as knowledge of another. This is very high order ToM but ­qualitatively
may not differ from (c). If I can think about your thoughts about my thoughts, it
may be that I have the capacity to think about your thoughts about my thoughts
about your thoughts. I just need to be able to represent a thought-about-a-thought
as a thought. Even if (d) represents an extension of the same cognitive capacity as
in (c), it surely does require more conscious attention and effort. No wonder we
have no good evidence that (d)-type thinking appears in any non-humans and is
challenging even for adult humans in many situations.
It is (d)-type thinking, being able to understand that others share the same
perspective that the gods want to change our thinking and behaving, that is required
 Justin L. Barrett

for Homo religiosus. With this metarepresentational, ToM humans can acknowledge,
worship, and serve the same god and know that they are worshipping the same
god. Metarepresentational ToM, then, allows the cognitive accounts of religion to
do more than explain why people postulate gods. They may begin to explain why
people collectively and jointly can and do interact with the gods in relationship.
Metarepresentational ToM would also allow for religious beliefs to facilitate intra-
group cooperation and trust: we can know that each other both believe that certain
behaviours are punishable by the gods (Dunbar 2004). Similarly, being able to
speculate about the mental states, particularly the intentions, of another is essential
for reading natural states of affairs as communicative acts. When people regard a
famine or storm as an act of a god (perhaps requiring a ritual response), they are
speculating that an unseen god has intended to act through these natural events
for a particular reason. Such speculation is metarepresentational and skirts very
near symbolism as discussed below. Given the pivotal role of metarepresentational
theory of mind to religion, it is no surprise that ToM is cited as a critical cognitive
component of religion by many cognitive scientists of religion (Barrett 2004; Boyer
2001; McCauley & Lawson 2002).
Metarepresentation in the domain of thinking about mental states may be
the extra something that made our ancestors Homo religiosus. I regard metarep-
resentational abilities as a more plausible adaptation than ‘religion’ as a whole,
because metarepresentation is an incremental extension of regular, lower-order
ToM capacities.

12. Metarepresentation and cumulative culture

The reason I regard this possibility as distinct from the Adaptation Hypothesis is
that adding metarepresentation to the existing hominin cognitive tool kit would
not only produce religion. Together with pre-existing capacities, metarepresenta-
tion (particularly in the ToM realm) would yield many abilities and activities that
we consider indicative of behaviourally modern humans. The ability to form rep-
resentations of representations or thoughts about thoughts has been suggested as
the secret ingredient, not just for religion, but cumulative culture more generally.
By cumulative culture, I refer to socially-acquired ideas or practices that are read-
ily expanded or modified by subsequent generations. Taking a useful tool such as a
hand axe and hafting it to make it even more versatile is an example of cumulative
culture or what Michael Tomasello calls the ‘ratchet effect’ because such cultural
evolution seems to require a holding mechanism to prevent slippage ‘relatively
faithful social transmission’ and then a mechanism for innovation and creativ-
ity that moves the same goal forward (Tomasello 1999). The proposed holding
 Chapter 11. Metarepresentation, Homo religiosus, and Homo symbolicus 

mechanism needed to facilitate strong, relatively error-free cultural transmission

is joint attention, the ability for me to know that you know that we are both attend-
ing to the same thing. ‘Attending’ here does not merely mean ‘oriented toward’
or even ‘looking at’ but something more like ‘holding in mind’. If we jointly hold
in mind the same thing, then I can knowingly communicate information about
it. Otherwise, I cannot be confident that any utterances or gestures I make in an
effort to communicate about the thing to which I am attending will be understood
as bearing on that thing. Likewise, unless you know you are attending to the same
thing I am attending to, you would not be able to be confident of the target of my
communication. For this reason, Tomasello argues that such joint attention is a
prerequisite for human-type language. Note that this sort of joint attention is an
instance of metarepresentation.
A more sophisticated form of joint attention may also be part of the mech-
anism that moves cultural innovation forward: shared intentionality. By shared
intentionality, I mean knowing that another is sharing a particular intention with
you. If we both know that we both want to build the same hut, we have shared
intentionality. This requires a more challenging form of joint attention because
the object of our intention has to be imagined; it is not yet realized. This shared
intentionality enables us to work together toward a single goal, vastly improving
our capacity for cultural innovation. Here we are dealing with a form of metarep-
resentation that may be absent in all non-humans.
Cognitive anthropologist Dan Sperber has similarly argued that the criti-
cal adaptation for creative cultural evolution is a ‘module of metarepresentation’
(Sperber 1994). This mental device allows us to take a representation (e.g. as of a
plant, animal, tool, or problem) and perform mental operations on it (e.g. what
if my staff had a sharp end?). Cognitive archaeologist Steven Mithen similarly
proposes that the modern human mind with its capacity for cumulative cultural
evolution arose because of increased cognitive fluidity, made possible because of
an adaptation enabling various specific mental devices to communicate with each
other in a metarepresentational ‘space’ (Mithen 1996).

13. Metarepresentation and Homo symbolicus

The power of metarepresentation to transform preceding capacities into quali-

tatively different abilities of cultural expression, including religion, may be
clearer if I briefly consider the role of metarepresentation in symbolism and
language.
Non-human primates, cetaceans, canids, honey bees, and other animals
seem able to signal. Signals can be utterances, gestures, or complex dances. What
 Justin L. Barrett

signalling does is prompt a behavioural reaction in the observers. For instance,

cries may be uttered to prompt others to stop moving, to run, to climb, to draw
near, or to attend to something or someone. Sophisticated animals can learn large
numbers of signals and even combine context cues with signals to produce more
nuanced signal-behaviour mappings. But signalling (as I am using the term here)
falls short of symbolism.
I take symbolism to refer to the cognitive process whereby an object of
perception triggers formation of a private mental representation that goes
beyond generating an identification of the object and its ontological properties,
and beyond the formation of a behavioural scheme for motor reaction to the
object of perception. If I see a sphere, I may instantaneously represent it as a
bounded physical object (with all the properties that go with such objects), as an
artefact, and as having certain affordances for interaction (e.g. I might be able
to pick it up and throw it). Such a set of representations would not amount to
symbolism. To understand it as a symbol (say, of sovereign power), I would need
to be able to represent the intentions of its maker or user. I would need to be able
to wonder “what is it intended for?” much in the same way that religious thought
often includes what signs and portents are intended for. Such cognitive activities
are metarepresentational.
What metarepresentation seems to add is more flexible signalling still as well
as symbolically triggering mental states (thoughts, ideas, affective states, etc.)
instead of only triggering behavioural routines. When talking to a border collie
(a very clever animal), a gesture toward ducks in a pond and the utterance ‘ducks’
will likely trigger the dog to attempt herding the ducks. When talking to a human
child the same gesture and utterance is likely to trigger an epistemic state: ‘those
things are called ducks,’ or ‘there are ducks in the pond.’ Because of metarepresen-
tational abilities, humans can consider what mental state or intention is behind
an utterance or gesture, including the possibility that the intention was to change
a mental state, not just a behaviour. For these reasons cognitive scientist Michael
Tomasello has argued that it is this higher-order ToM that makes human language
possible. No metarepresentation, no true language. It may be that metarepresenta-
tion (especially in the context of ToM) is the capacity that changes signalling into
linguistic communication and symbolism more generally.
If this analysis is right, it may be that when metarepresentation was added to
our ancestors’ cognitive tool kits, they quite suddenly became capable of cumula-
tive culture, symbolism, and language all at the same time, all because of the addi-
tion of one incremental change. If so, then the tool kit that makes Homo religiosus,
makes Homo symbolicus and behaviourally modern humans as well. The three are
identical and evolved concurrently.
 Chapter 11. Metarepresentation, Homo religiosus, and Homo symbolicus 

 n evidence for the evolution of metarepresentation, Homo religiosus

14. O
and Homo symbolicus

If metarepresentation endows the possessor with just one of the adaptive capacities
sketched above (cumulative culture, symbolism, language, or religion), then that
ability could be sufficiently adaptive to drive the evolution of metarepresentation.
If indeed metarepresentation has the whole suite of knock-off effects suggested
above, then the selective advantage for the Homo that acquired it would be great.
As metarepresentational ToM appears to be an extension of lower-level ToM, ToM
is the likely evolutionary predecessor. But when did this metarepresentational
variant of ToM emerge? Unfortunately, the picture is unclear.
One difficulty in determining when and how metarepresentational ToM and
its alleged products such as religiousness might have evolved in humans is that
they do not necessarily leave the kind of material traces that archaeologists use
as evidence. I can believe in the existence and activity of forest spirits without
leaving any material trace. If I pray devoutly five times a day, no distinctive mate-
rial trace need be left behind. If I have a rich belief in afterlife and reincarnation
that prompts me to burn dead bodies, the bodies and the evidence of the ritual is
unlikely to be found. Symbolism presents similar difficulties. If I wear shell beads
around my neck, I might be symbolizing my clan membership, my group rank,
or my marital status. But I might just find them aesthetically pleasing and am
encouraged to wear them by the positive attention they garner. If my symbolism is
manifest through utterances or gestures, they may leave no material trace. Artefact
evidence, then, may not be sufficient for determining the evolution of metarepre-
sentation, symbolism, and religion.
Fortunately, other strategies are available. One way of estimating how old a
cognitive mechanism is is to look for evidence in other species. If chimpanzees
and gorillas share a particular cognitive mechanism with humans, it may be safe
to suppose that our common ancestor with chimpanzees shared the mechanism.
In the case of metarepresentational theory of mind, no non-human species have
yet unambiguously exhibited it, with even the most generous of attributions being
credibly limited to members of Pan (Call & Tomasello 2008; Penn & Povinelli
2007). Ignoring the possibility of convergent evolution, such evidence from
comparative cognitive studies seems to set the bounds for the emergence of this
higher-order theory of mind at the common ancestor of Pan and Homo, but more
probably after the split.
Additionally, on a gross level, because certain cognitive mechanisms may be
associated with changes in brain size or shape, cognitive evolution can be informed
by skeletal remains. For instance, evolutionary anthropologist Robin Dunbar has
 Justin L. Barrett

noted a linear relationship between brain cortex size and behavioural evidence
of increasingly complex ToM in living species. Applying this linear relationship
to the fossil record (treating cranial capacity as a proxy for brain size), Dunbar
tells us we can be pretty confident that Homo erectus did not have enough brain-
power to exhibit higher-order metarepresentational ToM (Dunbar 2004). Such an
ability likely evolved later, likely not until (or after) the appearance of anatomically
modern humans.

15. C
 onclusion

Though we cannot yet be certain when in our evolutionary history metarepresen-

tation emerged and gave rise to religion and symbolism, if the analysis presented
above is correct, we can be confident that when evidence of either symbolic or
religious activity is discovered, capacity for the other is present. Further, evidence
of cumulative cultural evolution suggests the presence of metarepresentation,
the factor that makes both religion and symbolism possible. Metarepresentation
transforms a host of disparate conceptual biases and predilections into a natural
propensity toward religious thought and expression. Likewise, metarepresentation
transforms gestures, utterances, and other signs into symbolism and language. If
metarepresentation is the lynch-pin that holds together these forms of cultural
expression, then its evolution, even if only a small modification on previous theory
of mind capacities, could lead to a radical break in behavioural possibilities for its
possessors.

References

Alcorta, C. & Sosis, R. 2005. “Ritual, emotion, and sacred symbols: The evolution of religion as
adaptive complex.” Human Nature 16: 323–59.
Atran. S. 2002. In Gods We Trust: The evolutionary landscape of religion. Oxford: Oxford
­University Press.
Barrett, J.L. 2000. “Exploring the natural foundations of religion.” Trends in Cognitive Sciences
4: 29–34.
Barrett, J.L. 2004. Why would anyone believe in God? Walnut Creek, CA: AltaMira Press.
Barrett, J.L. 2007. “Cognitive Science of Religion: What is it and why is it?” Religion Compass 1
(6): 768–786.
Bering, J.M. 2006. “The folk psychology of souls.” Behavioral and Brain Sciences 29: 453–62.
Bering, J.M. & Johnson. D.D.P. 2005. “’O Lord … You Perceive my Thoughts from Afar’:
­Recursiveness and the Evolution of Supernatural Agency.” Journal of Cognition and Culture
5: 11842.
 Chapter 11. Metarepresentation, Homo religiosus, and Homo symbolicus 

Bloom, P. 2004. Descartes’ Baby: How child development explains what makes us human. London:
William Heinemann.
Bloom, P. 2007. “Religion is natural.” Developmental Science 10: 147–51.
Boyer, P. 2001. Religion Explained: The Evolutionary Origins of Religious Thought. New York:
Basic Books.
Boyer, P. 2003. “Religious thought and behavior as by-products of brain function.” Trends in
Cognitive Sciences 7: 119–124.
Call, J. & Tomasello, M. 2008. “Does the chimpanzee have a theory of mind? 30 years later.”
Trends in Cognitive Sciences 12: 187–192.
Dawkins, R. 2006. The God Delusion. London: Bantam Press.
Dunbar, R.I.M. 2004. The Human Story. London: Faber and Faber.
Hamer, D. 2004. The God Gene: How faith is hardwired into our genes. New York: Doubleday.
Humphrey, N. 1998. “What shall we tell the children?” Social Research 65: 777–805.
Johnson, D.D.P. 2008. “Gods of War: The Adaptive Logic of Religious Conflict.” In The Evolution
of Religion: Studies, Theories, and Critiques, J. Bulbulia, R. Sosis, C. Genet, R. Genet, E. Harris
& K. Wyman (eds), 111–117. Santa Margarita, CA: Collins Foundation Press.
Kelemen, D. 2004. “Are children “intuitive theists”? Reasoning about purpose and design in
nature.” Psychological Science 15: 295–301.
McCauley, R.N. & Lawson, E.T. 2002. Bringing Ritual to Mind: Psychological Foundations of
­Cultural Forms. Cambridge: Cambridge University Press.
Mithen, S. 1996. The Prehistory of the Mind: A search for the origins of art, religion and science.
London: Thames and Hudson Ltd.
Penn, D.C. & Povinelli, D.J. 2007. “On the lack of evidence that chimpanzees possess anything
remotely resembling a ‘theory of mind.’” Philosophical Transactions of the Royal Society,
B 362: 731–744.
Pichon, I., Boccato, G. & Saroglou, V. 2007. “Nonconscious influences of religion on ­prosociality:
A priming study.” European Journal of Social Psychology, 37: 1032–1045.
Pinker, S. 1997. How the Mind Works. New York: W. W. Norton & Company.
Pyysiäinen, I. 2001. How Religion Works: Towards a New Cognitive Science of Religion, Cognition
and Culture Book Series, 1. Leiden: Brill.
Rochat, P., Morgan, R. & Carpenter, M. 1997. “Young infants’ sensitivity to movement infor­
mation specifying social causality.” Cognitive Development 12: 537–561.
Ruffle, B. & Sosis, R. 2007. “Does it pay to pray? Costly ritual and cooperation.” The B. E. Journal
of Economic Analysis and Policy 7: 1–35 (Article 18).
Saroglou, V., Corneille, O. & Cappellen, P. 2009. “ ‘Speak, Lord, your servant is listening’: Religious
priming activates submissive thoughts and behaviors.” International Journal for the Psychology
of Religion 19: 143–154.
Sosis, R. 2003. “Why aren’t we all Hutterites? Costly signaling theory and religious behavior.”
Human Nature 14: 91–127.
Sosis, R. & Bressler, E. 2003. “Cooperation and commune longevity: A test of the costly signaling
theory of religion.” Cross-Cultural Research 37: 211–239.
Sosis, R. & Rouffle, B. 2003. “Religious Ritual and Cooperation: Testing for a Relationship on
Israeli Religious and Secular Kibbutzim.” Current Anthropology 44: 713–722.
Sperber, D. 1994. “The modularity of thought and the epidemiology of representations.” In
Mapping the Mind: Domain Specificity in Cognition and Culture, L. Hirschfeld & S. Gelman
(eds), 39–67. Cambridge: Cambridge University Press.
 Justin L. Barrett

Tattersall, I. 1998. Becoming Human: Evolution and Human Uniqueness. New York: Harcourt
Brace.
Tomasello, M. 1999. The Cultural Origins of Human Cognition. Cambridge. MA: Harvard
­University Press.
Tremlin, T. 2006. Minds and Gods: The cognitive foundations of religion. New York: Oxford
­University Press.
Whitehouse, H. 2000. Arguments and Icons: Divergent modes of religiosity. Oxford: Oxford
­University Press.
Whitehouse, H. 2004. Modes of Religiosity: a cognitive theory of religious transmission. Walnut
Creek, CA: AltaMira Press.
Whiten, A., Goodall, J., McGrew, W.C., Nishida, T., Reynolds, V., Sugiyama, Y., Tutin, C.E.
G., Wrangham, R.W. & Boesch, C. 2001. “Charting cultural variation in chimpanzees.”
­Behaviour 138: 1481–1516.
Wilson, D.S. 2002. Darwin’s Cathedral: Evolution, Religion, and the Nature of Society. Chicago:
University of Chicago Press.
 Index
A Afrocentric 142
abnormality 7
abstract 6, 51–52, 64, 75, 78,
80, 89, 98–99, 105–106,
164–166, 172–174, 179
abstract ideas 99
abstract thinking 172
Acacia 101, 103–104
acanthocardia 53
Acheulean 55
Acheulian technocomplex 112
acoustically 4
adaptation 17, 19, 24, 50, 62, 133,
169, 177, 205–212, 218–219
Adaptation Hypothesis 206,
208, 210, 212, 218
adaptive 21, 59, 64, 120, 127,
134, 137, 142, 168, 176,
210–211, 221
adaptive allele 120
adaptive selection 168
ADD 32, 141, 213, 220
adhesives 97, 101–102, 104–107
adolescent 123
advanced cognition 50, 133
advanced working
memory 97, 105
aeolian 82–83
Affiliation/Belonging
system 176
Africa 22, 49–58, 62–65, 75–78,
80, 92–93, 97–98, 101, 107,
111–113, 117–118, 123–124,
142–144, 147–149, 153, 155,
158, 179, 188, 190, 199–200
African 2–3, 5, 18, 24–25, 33,
75, 78, 83, 92–93, 107,
111–113, 115–116, 119–120,
122–124, 137, 142–145, 190
African apes 2–3, 18, 24–25
African gray parrot 5
African Grey Parrot 137
African sapiens 112, 116, 119
African/Levantine
tradition 123
afterlife 205, 211–212, 215, 221
Agency Detection Device 205,
213, 216
agriculture 136, 141, 144, 178–179
alarm calls 4, 7
Alex 137–138
Algeria 53–54, 147
allele 120
alloying metals 100
Americas 112
Amud 51, 152
analogical reasoning 100
anatomical 13, 15, 18, 22,
123, 135
anatomical changes 135
anatomical structure 15
anatomically modern
humans 99, 101, 141,
154, 222
ancestor-spirit 205
ancient human
populations 185
animacy 215
animal 2, 13, 15, 18, 42, 83,
99, 101, 111, 121, 123, 164,
168–169, 214, 219–220
animal consciousness 15
animal markings 83
animalistic 113
animality 115
animism 8
antagonistic scenarios 49–50
Antarctic 61
Antarctica 61
anthropocentric 1
anthropocentrism 7
anthropogenic 51, 76, 78
anthropological 186
anthropologist 115, 210, 213,
219, 221
anthropology 1, 3, 133, 186, 205
anthropomorphic 113
anthropomorphism 213, 216
anthropomorphize 212–213
anti-predator alarm 1, 4
ape 2, 8–9, 13–14, 18–19, 23–27,
29, 33, 37–39, 115, 137
ape DNA 23
ape infants 13, 33, 37–38
ape language 26–27, 29
ape like ancestors 115
ape rain dancers
apes 1–3, 5, 8–10, 14–15, 17–20,
23–28, 30, 37–39, 136–137,
141, 152, 217
Apollo 11, 56–57, 143
Aquatic Ape Theory 2
arboreal 17–18
arboreal feeding 17
arboreal locomotion 18
archaic Homo sapiens 101
archaics 119
Arcy-sur-Cure 121, 150, 158
Ardipithecus 10, 17, 25
arrow tip 101
arrowheads 101
art 14, 18, 22, 97, 111–113, 115,
125–127, 143–144, 146, 149,
155–156, 173, 175
art historians 97
arthritic malformations 116
artificial language 4
artisans 89–90, 97, 102,
104–105
Asia 22, 50, 54, 60, 112, 115, 158
Asian 142
Assimilation models 123
associative 137, 152–157
associative (symbolic) 152
associative interaction 152–153,
156–157
associative learning 137
Atapuerca 111, 151
Aterian site 53
Atlantic 60–61, 78
Atomic Physics 167
Aurignacian 57, 111, 114
Australia 54, 102, 112, 151
australopithecines 151, 154
 Index
Australopithecus
afarensis 150
Austria 57
autism 176
autonomic responses 3
awls 56, 118, 121
B biological species 112, 120
backed 101–102, 147
bacterial (prokaryotic) cells 134
Baddeley, Alan 192–193
Barbs 101
basal ganglia 22, 187
basket-shaped
morphology 123
Bass Strait 58
bats 7
BBC 78–79, 82–84, 88, 90–92
BBC Hiatus 82
beads 52–54, 56, 76, 82, 92,
97–98, 112–113, 123, 143,
148, 221
beadworking 148
Bechtel, William 194
behavioural 2–3, 6, 8, 13, 30,
50, 60, 63, 75, 89, 98–99,
112, 114, 118, 124, 133, 138,
142–144, 149–150, 153, 157,
176, 185, 187–191, 193, 197,
199–200, 212, 214, 220, 222
behavioural modernity 50, 75,
124, 143–144, 149–150
behavioural patterns 63,
176, 185
behavioural traits 144, 189
behavioural
transformation 190
behavioural variation 98, 138
belief 9, 152–153, 205,
207–208, 210–212,
215–216, 221
bell 137
Berekhat Ram 145
between-group 134–138
between-group conflict 138
between-group selection
134–137
Bickerton, Derek 191
bicultural 15, 28, 37
biface 112, 151
bifacial 52, 56, 82
big noses 117
bigger brain-greater
intelligence 189, 191
Binford, Lewis 91, 185, 190
biochemistry 167
biodiversity 64
biological 19, 21, 50, 112–114,
119–120, 124, 163–164, 170,
185–187, 190, 199, 207
biological differences 185
biological/genetical
foundations 124
biomass 135
biospecies level 112
bipedal 17–18, 168
bipedalism 17, 24, 26
bipeds 2
birch bark 119
bird bone 56
birth 13, 18, 28, 30–31, 33–36,
40–41, 117, 178
bison incisor 121
bitumen 119
Blakeslee 165, 168–172, 174–175
Blombos Cave 52–53, 55–56,
78–79, 82, 85–86, 113, 143,
145, 147, 149, 159
blood 2, 102, 117
blood groups 2
Bloom, Paul 215-216
body 15, 27, 30–33, 35–36,
39–41, 51, 77, 90–91, 111,
117–119, 121, 123, 127, 135,
146, 148, 150, 156, 177,
197, 215
body ornaments 121, 127
body weight 31, 177
Bokkeveld 83
bond 61, 176
bonding 37, 173, 179
bone 51–52, 56–58, 64, 75–78,
82, 90, 118, 121, 125, 146
bone awls 118, 121
bone retouchers 51
bone splinter 146
bone tool 52, 56–58, 77, 82, 125
bonobo 2–3, 8, 14–15, 18, 24,
27–29, 31, 36–37, 137
Border Cave 51, 53, 143
Boule, Marcelin 115-116
Boyer, Pascal 209, 215
Braille 5
brain 15–16, 19, 26, 39, 41,
99, 111, 120, 125, 127, 137,
153–154, 163–172, 175–178,
180, 189, 191–199, 221–222
brain connections 172
brain growth 191, 197
brain plasticity 176–177
brain size 19, 99, 111, 120, 154,
177–178, 192, 196–197,
221–222
brittle glue 101
Broca’s area 193–195
Brooks 50–51, 55–56, 62, 76,
118, 142–143, 149
burial 49, 51, 53, 58, 112, 122,
150, 154–158, 212
burial practices 49, 58, 212
buried their dead 51, 117,
151, 212
burning 104, 155
Bushman Rock Shelter 77
Byzantine 120
C
Cambridge On-line 115
Canada do Inferno 126
Cape Hangklip 77–78
Cape Range 54
Cape Town 78, 82, 158, 163, 200
captive 14, 17, 29
capuchin monkeys 2
carbonised 82
carnivore 150–151
carved 56
causal 16, 31, 38, 40, 168, 199
causal reality 16
causal relations 168
cave paintings 113, 125
Cebus libidinosus 2
cell 19, 133–134, 167
cell biologist 133
cell biology 167
cell organelles 134
ceramics 100
chaine opératoire 90
charcoal 101–102
Chase 56, 63, 144, 153
Châtelperronian 121, 123, 149
Chauvet 56–57, 113
chemical analysis 118
chemistry 105, 119, 167
child skeleton 122
childcare 136
child-caring 178
childlike 15
chimpanzee 2–6, 8, 10, 18–20,
23–24, 26–27, 32, 34–36,
38, 150, 212

chloroplasts 134
Chomsky 170
chromosomes 23–25, 134
chunking 167
climate 2, 33, 50, 60–61, 63–65,
76, 190
climate change 60, 65, 76
cling 13, 17–18, 30–35, 37–38
clinging 13, 30–34, 36
Clottes 56, 113
Côa Valley 126
coded information 49
cognition 22, 49–50, 58–59,
63, 75, 97–100, 104–107,
111, 113–115, 118, 125, 133,
136–137, 166, 169, 177, 180,
186–189, 197, 199, 206–207
cognitive 1, 10, 13, 15, 23–25,
29, 36–37, 49–50, 58,
75–77, 97–100, 104, 107,
111–113, 117–118, 125, 133,
136, 141–142, 145, 151,
153–154, 157, 165, 167,
177–178, 185–200, 205–207,
213–221
cognitive abilities 75, 77, 100,
177, 196
cognitive adaptation 205
cognitive architecture 205, 216
cognitive control 99, 198
cognitive development 142,
145, 177, 205
cognitive evolution 153, 154,
185–186, 188, 192, 195,
198–199, 221
cognitive fluidity 97, 99–100,
104, 107, 219
cognitive function 10, 167, 188,
193, 195–197
cognitive neuroscience 186,
188, 190, 194–195
cognitive science 188, 205
cognitive science of
religion 205
cognitive sciences 205
cognitive systems 205–206, 213
cold-adapted body
proportions 123
colourant 102
coloured stains 102
common attributes 98
communication 1–8, 10, 24, 30,
82, 90, 92, 136, 163, 165–166,
170, 176, 178–179, 219–220
communicative abilities 163
comparative
neuroscience 196–198
complex 7, 9, 14–15, 21–22, 29,
41, 52, 57–58, 63, 76, 92,
97–100, 102–103, 105–107,
112, 115, 118, 126, 136–137,
147–148, 163–164, 167, 173,
189, 191, 193, 195, 197–199,
210, 214, 217, 219, 222
complex capabilities 99
complex cognition 97–100,
103, 107
complex cognitive
processes 98
complex cognitive tasks 193
complex hierarchical
sentences 193
complex planning 14
compound 97, 100, 102–107,
174
compound adhesive
manufacture 97, 100,
103–104, 106
compound adhesives 97, 102,
104–107
compound tools 97
computer languages 174
computer machine code 6
Conard 9, 54, 56, 113–114, 118
conceptual 6, 146, 166, 171,
214, 222
conceptual schemas 116
conceptual structure 171
Concurrence Hypothesis 206,
212–213
Congo 8, 27
conscious mind 169
consciousness 13, 15–16, 18,
31, 34, 36–37, 39–41, 169,
172, 215
Conus 53–54
convex edge 101
Coolidge and Wynn 192–193
Coolidge, Frederick 191–192
cooperative eye hypothesis 136
copulation calls 7
core 52, 62, 141, 172
cortex 52, 54, 99, 168, 170, 172,
174, 193–194, 222
cortical development 177
cortices 192, 196, 198
counter intuitive intentional
agency 206
Index 
courtship 1, 5
cracking open 2
cranial capacity 117, 222
cranium 123–124, 197–198, 200
crayons 143, 147, 149
creativity 166, 172, 179, 189, 218
crescents 101
cries 5, 33, 220
Croatian cave site 117
cross-cultural 206
cross-hatched 77, 80,
87–88, 90
CSR 205–206
Cueva de los Aviones 53–54
cultural 1, 14, 17, 20, 27, 29–30,
49–51, 56, 58–60, 62–65,
76, 92, 122–123, 125, 127,
136, 142, 148, 152, 166,
169, 186, 205–209, 211,
218–219, 222
cultural drift 50
cultural evolution 50, 59,
62, 136, 209, 218–219,
222, 136
cultural expression 206–208,
219
cultural frames 166
cultural learning 169
cultural modernity 59, 76
cultural multilevel
selection 136
cultural systems 58
cultural variation 152
culture 6, 15, 26, 28–29,
36, 42, 49–52, 58–60,
63, 75–77, 88, 97–99, 107,
112, 118, 121–123, 125,
133, 136, 145, 156, 169,
177, 179, 185–187,
189–191, 198–200, 210,
218, 220–221
cumulative behavioural
change 133
cup marks 152
cupules 51
cutting 77, 84, 101–102
cutting edge 84, 101, 102
D
d’Errico, Francesco 159
dance 7–8, 173
dance language 7
dancing 8, 175, 180
Dansgaard-Oeschger 61
 Index
Darwin 20, 115, 216
Darwinian 20, 23, 174,
177, 206
dating 62–63, 78, 82, 143,
147, 157
dawn of language 163, 171
Deacon’s thesis 137–138
Dean, Hamer 210
decorated 56–57, 78, 91, 149
decoration 22, 56, 80, 148–149,
154–155
decorative 121, 125, 157, 180
decorative interaction 157
decorative patterns 121
Dederiyeh 51
defense 134, 136
deliberate heating 53, 56
Democratic Republic of
Congo 8
demographic changes 51
demographical thresholds 114
dendritic 77, 87–88
Dentalidae 54
dentition 123
designs 51–52, 75, 81, 90, 100,
105, 149
developmental 13, 21–22,
37–38, 153, 168, 174–176,
178, 196–198, 205, 214
developmental
neuroscience 196–198
developmental verbal
dyspraxia 22
dialects 1, 167
Diepkloof Rock Shelter 52,
78–79, 81
diet 17
differences 5, 13–14, 18–19, 23,
25, 30, 34, 36, 38–39, 59,
77, 80, 112, 121, 135, 185,
191, 198
direction of gaze 136
dispersal 50, 62, 76, 143, 153
diversity 81, 90, 117, 138, 191
DNA 13, 18–20, 23, 25, 117,
120, 170
D-O 61
dogs 20, 136, 214, 217
dolphins 7
domestication 137
dominance 4, 138, 212
domineering behaviour 135
doodling 90
dopamine 176
Dordogne 55
dorsal prefrontal cortex 194
Down’s syndrome 24
drawing 19, 29, 77, 90, 100,
102, 143–144, 153, 188,
205, 217
DRS 78–80, 88, 90–92
drumming 4, 8
drumming signature 4
duality of patterning 166
Duiwenhoks 83
Dunbar, Robin 141, 145,
152–154, 191, 221–222
E 90
Early Stone Age 59
Eastern Cape 77
echolalic speech 5
echo-location 7
eco-cultural 64
ecological 1–3, 58, 64, 136,
138, 163–164, 167, 175,
178, 186
ecological dominance 138
ecosystems 63
egalitarian 135
egalitarianism 135
eggshell 54, 75, 78, 80–81,
90–92
eggshell containers 91
eggshells 52–53, 80, 83, 88,
91–92
ego 16–17
electron spin resonance 82
electrostatic 102
elephants 7, 30, 217
Elykia 28
embodied brain 163
Embodied Construction
Grammar 169
emergence of language 39, 111,
115, 125, 163, 175, 179
emotional 32, 37–38, 136, 152,
170–171, 173, 175–179,
199, 209
emotional drivers 171, 175
emotional power 173
emotional state 136
emotions 164, 175–176
emotive 152, 154–157
emotive (non-symbolic) 152
emotive interaction 152,
154–157
emotive non-symbolic 152
Enard 22–23
encephalisation 189
endocasts 197
end-scrapers 146
England 24
engraved 51–52, 54, 56, 77–78,
80–83, 85–88, 90–92, 149,
156
engraved bone 51, 52
engraved eggs 82, 91–92
engraved lines 56, 80, 88
engraved ochre 52, 81–83,
85–86, 88, 91, 92, 149
engraved patterns 77, 82, 88,
engraved pebble 77, 78
engraver 77, 88
engraver laterality 77
engravings 52, 58, 64, 75–77,
84, 87–92, 143, 149
enhanced speech 187
Enkapune Ya Muto 101
environmental variation 21,
51
epigenetic 18–21, 25, 28, 30
episodic memory 193–194
epistemological 3, 89, 194
ESR 143, 147
Ethiopia 150
ethnographic 58, 77, 91–92,
102, 113–114, 118, 126
ethnographic present 91, 92,
113, 118
ethnology 186
ethologist 3
eukaryotic 133
eukaryotic cells 133
Eurasia 49–51, 53, 75–76,
111–112, 188
Eurasian 112, 117–118, 123, 144,
149
Eurasian humans 117
Eurasian Middle
Palaeolithic 144, 149
Eurasian Neandertals 118, 112
Eurocentric 142
Europe 22, 49–53, 55–59,
62–64, 76, 111–112, 115,
118–123, 125, 142, 144, 149,
157–158
European 57, 63, 65, 75, 93,
115–116, 119, 123–125,
142–143, 153
eusocial insect species 135

eusociality 135
Eve 117–118, 123
everted lower lip 6
evolution 7, 10, 13, 19–20, 24–
25, 49–51, 58–60, 62, 64,
75, 97, 99, 111–112, 115–116,
123–125, 133–138, 141–142,
145, 150–151, 153–154, 163,
171, 173, 176–177, 179,
185–200, 206–207, 209,
215, 218–219, 221–222
evolution of reason 99
evolutionary 1–2, 8, 14, 120,
133–135, 141–142, 150,
152–153, 163–164, 168–170,
173, 176–179, 186, 188–189,
198–199, 205–206, 213,
221–222
evolutionary biologists 189
evolutionary biology 186, 188
evolutionary development 169
evolutionary psychological
development 173
evolutionary psychologists 189
evolutionary psychology 205
evolutionary theory 135
evolutionary transition 133–135,
168
evolutionary transitions 135
excavations 63, 118–119
executive functions 100, 192
experiential processes 169
experimental studies 3
explanandum 189
explanandum (modern
culture) 189
explanans 189
explanans (modern
intelligence or
language) 189
explicit memory 194
extended learning 99
extinct 1, 117, 119, 158, 185–186,
190, 196, 214
extinction 50, 111, 119,
122–123
eyes 34, 38, 136, 138, 172
F funerary practices 77, 152
facial expressions 3–4, 34,
36, 39
facial-vocal dialogues 34
farming 209
fat 31–33, 35, 101, 104, 118
Index 
fauna 82, 150 GCM 64
faunal 64, 78, 80, 82, 91 gender 113
Feldhoffer 115, 152 gene 18, 20–23, 26, 120, 179,
fetal malformation 30 187, 210
figurative art 111, 113, 125, gene expression 18, 20, 21
143–144, 146, 149, 155–156 gene pool 120
fingerprint 119 genealogical 133
fingers 17 genealogical relatives 133
fire 62, 64, 97, 103–106, 211 General Circulation Model 64
fire control 97 generic 141, 164–165, 171,
fists 17 174–175
fitness differences 135 genes 2, 19, 21–23, 25–26, 65,
flake 87, 103, 118 122, 134, 138, 170, 175,
flakes 146 188–189, 199, 210
flasks 81 genetic 13–14, 18, 20–21, 23, 50,
flint 118, 146 58, 62, 64, 115, 117, 119–120,
flint flake 118 125, 135–137, 170, 174–175,
flutes 56 179, 185–190, 198–199,
Foley 113 210–211, 213
foliate 52, 82 Genetic Algorithm for Rule Set
foliate point 52 Prediction Gathering 64
food 14, 20, 22, 26, 38, 137 genetic change 190, 21
foodstuffs 1, 4 genetic evolution 58, 125, 136,
foraging patterns 17 137, 186–188, 190
foraminifera 62 genetic inheritance 170
forest 8, 10, 18, 22, 25, 28–29, 221 genetic selection 20, 211
formal 58, 75, 82, 169, 191 genetically inherited 100, 123
formal bone tools 58, 82 geneticists 111
formal lithic tools 75 genome 19–23, 25, 117,
fossil record 1, 3, 19, 23, 25, 120, 170
197, 222 genomes 2, 19–21
fourth level intentionality 152 geographic isolation 112
fox canines 121 geographical 112, 143, 145, 164
FOXP2 22–23, 26, 187–188 geographical segregation 112
fractal patterns 173 geological 21, 60, 62, 83, 103
France 51, 55, 57, 113, 121, 125, geometric 52, 76, 82, 213
144, 146, 151–152 geometrical pattern 172
Franco-South African 78 Germany 56–57, 111, 113–114,
free-ranging apes 14 125, 151
freewill 16 gestures 1, 6, 26, 29, 77, 219,
frequency 7–8, 134, 144 221–222
friction 101 gibbons 1, 4, 17, 24
frontal 100, 193, 196 glacial 60–61, 64
frontal lobe 100, 196 global warming 62
frontal lobe damage 100 glue 97, 101–106
Fumane 56–57 glue makers 97, 105
functional modular glue residues 101
structures 171 Glycimeris 54
gnawing 151
goal 90, 100, 188, 193, 213,
G 218–219
GARP 64 goal directed actions 100
gaze 39, 136 goal maintenance 193
gazelles 120 god-gene 210
 Index

gods 206–207, 209–211, 213, half-man/half-beast 116 Homo erectus 22, 111, 112, 114,
215–218 Hamer Dean 210 117, 125, 154, 195, 197, 216
Gombe 8 hand 4, 6–7, 9, 13, 18–19, Homo ergaster 197
Goodall 4, 8, 150 31, 41, 64, 92, 189, 196, Homo habilis 22
gorillas 221 211, 218 Homo heidelbergensis 149,
Goukou 83 hand axe 112 151, 154, 157, 192, 197, 199
grammatical 13, 16, 168, hand waving 13, 31 Homo neandertalensis 149,
173, 179 haplotypes 119 151, 153, 155, 157
grammatical abilities 13 harpoons 56 Homo religiosus 206, 212, 214,
grammatical structure 16 Hawkins 165, 167–172, 216, 218, 220
grave 51, 150–152, 156, 212 174–175 Homo sapiens 2, 9, 22,
grave goods 51, 151, 156, 212 Hawkins and Blakeslee 168, 50–53, 55, 57–59, 63,
grave markers 152 171–172, 174–175 75, 76, 92, 101–113, 128,
grave pits 152 hearths 22, 82 142, 145, 151, 153, 155,
Gravettian 153 heating 53, 56 157–158, 187, 189, 192,
Great Ape Trust 13–14 Heinrich events 61, 153 195, 197, 199, 200
Greenfield 26–27 hide 14, 84, 118, 152, 209 Homo sapiens sapiens 22
Greenland ice cores 60 hide working 118 Homo symbolicus 22, 23, 26,
grinding 76, 83–84, 106, 143 hides 56, 118 97, 143, 153, 158, 200, 220
grinding marks 143 hierarchical structuring 166, hormones 99
grinding slabs 143 174 horse 114
grinding stones 143 high 7–8, 20, 23, 29, 60–62, 64, Howieson Poort 55
grooming 4, 6, 154 82, 91, 123, 134, 195–196, HP 77–78, 80–81, 91–92
grooming hand-clasp 4, 6 198–199, 217 hues 56, 156
grooved 83, 121 high cranial vault 123 human 1–8, 10, 13–42,
grooved pendants 121 higher 100, 138, 166, 168–169, 49–51, 53, 57–60,
Grotte des Pigeons 53–54, 147 171–174, 177, 179–180, 62–64, 75–77, 89, 91,
Grotte du Renne 121, 150, 158 188–189, 210, 216, 97–99, 105–106, 111–115,
group 1, 4–6, 14, 22, 27–30, 220–222 117–121, 123–125, 133–138,
33, 36–37, 39–41, 58–59, higher level states 168 141, 144–147, 152–153, 158,
83, 87, 89, 92, 97, 113, 115, higher order 100, 216, 220, 164, 167–170, 172–173,
117, 123, 127, 133–138, 143, 221, 222 176, 178, 185–192,
145–146, 150, 154, 176, 180, highly coded system 15 194–199, 206–210,
209–211, 218, 221 high-resolution neuroimaging 212–217, 219–221
group identity 5, 6, 97 techniques 199 human agency 213, 214
group membership 113, 180 Hohle Fels 57, 114 human genome project 170
group size 59 Hollywood film 116 human genus 111, 124
group specific hominin 1, 25, 62, 120, human nature 206–208
communication 5 135, 138, 141–142, Human Revolution 50,
grunt 4 145, 148, 150–155, 112–113, 115, 117–118, 120
Gua 26 185, 218 human species 117, 124, 133, 137,
guarded egalitarianism 135 hominin lineage 135 186, 212, 213, 216, 217, 221
gum/resin 101 homininae 142 human working memory
Homo 2, 9, 13–15, 18–19, system 192
H 22–30, 35, 50–53, 55, human/ape differences 13,
habitat homogenization 21 57–59, 63, 75–76, 92, 97, 18, 23
Hadar 150 101, 111–114, 117, 125, hunt 102
HADD 213–215 142–143, 145, 149, 151, hunter 81, 135, 144, 148, 209
haematite 53, 103, 143 153–155, 157–158, 187, hunter gatherer societies 135
haematite pencils 143 189, 192, 195, 197, hunter gatherers 81, 144
hafted 101, 103 199–200, 205–206, hunting 19, 102, 105, 136, 209
hafting 56, 101, 106–107, 212, 214, 216, 218–222 Hypersensitive Agency
119, 218 Homo antecessor 111 Detection Device 213, 216
hafts 102, 105 Homo bonobos 27 hypnosis 16, 41

I intuitions 206
Iberia 111
Iberian 62
ice 60–62, 118
Ice Age 118
ice cores 60, 62
ice sheets 60
iceberg 61
iconographic 76
icons 173
Ifri n’Ammar 53
IGCP 437 62
Ilsenhöhle rock shelter 118
imaginative play 163, 174, 176,
179–180
imitation 8, 26, 120,
169, 176
implements 75, 105
incising 77, 88, 90
incisions 77, 84, 87–88
incisor 121
Indonesian island of Java 111
infant 13, 15, 17–18, 23, 30–40,
176, 213
infant carriage 13
infant clinging 13, 36
inferior frontal gyrus 193
in-group conformity 210
inheritance 138, 164, 167, 170
inhibitory control 193
inhumation 51, 156
innate patterns 13
innovation 49–51, 59, 75, 92,
99, 127, 218–219
innovations 51, 58–59, 63–65,
92, 98, 190
innovative technologies 99
insects 35, 56, 135
intellectual capacities 171
intentionality 152, 154, 191–192,
219
intention-reading 169
interaction 4–5, 27, 30, 39,
41, 50, 63, 111, 114, 122,
150, 152–154, 156–157, 168,
176–177, 220
interbred 123, 190
interbreeding 25, 120
interglacial 60, 62, 83
interglacial transgression 83
inter-group competition 127
inter-level connections 172
International Geological
Correlation Program 62
Index 
127, 133, 135–138, 163–180,
intuitive dualism 214–216 187, 189, 191, 193–194, 197,
involuntary expressions 8 208, 219–222
Iraq 152 language like signals 7
iron oxide 103 language modules 169, 170, 171
Israel 53–55, 145, 147, language technology 179
151–152, 210 Langur monkey 4
Italy 55, 57 Last Common Ancestor 2–3,
ivory 8, 56, 113–114, 125 117
Ivory Coast 8 Last Glacial Maximum 64
Iziko-South African Late Aurignacian 114
Museum 78 Late Middle Palaeolithic 146,
152
J Latent Semantic Analysis 169
Java 111 Later Stone Age 57, 75, 143
javelins 118 laughter 137, 176
juveniles 28, 150, 152 Le Moustier 51
leather 77
K Lebanon 147
Kakombe Valley 8 leg bones 123
Kanzi 15, 27–29, 39, 137–138 leg kicking 13
Kanzi’s mother 27 Lemurs 10
Kapthurin 143–144, 147 Leroi-Gourhan 121
Kenya 101, 147 lesion 196
Khoesan 81, 91–92 Levallois-rich 145
kiln-fired ceramics 100 Levant 144, 148, 151
Kimberly 54 limbic system 28, 176
kinship 18 linguistic 4–7, 10, 14–18,
Klasies 52, 55, 77–78, 84, 143 23, 26, 29–30, 35, 37,
Klein Kliphuis 52 40, 168–170, 175, 191,
Kleine Feldhoffer 115 205, 220
knapped 84, 87 linguistic analysis 171
knapping 87 linguistic constructions 16
knuckles 18, 24 linguistic patterns 168, 169
Königsaue 119 linguistic recursion 191
Krapina 152 lion 57, 114
Ksar Akil 147–148 lithic 52, 75, 78, 87–88, 146
KwaZulu-Natal 98 lithic doodles 146
Littorina Africana 53
L locomotion 4, 17–18, 24, 31
L’Hortus 152 long-tailed macaques 2
La Chapelle-aux-Saint 51 Lovejoy 1, 10, 25, 30
La Chapelle-aux-Saints 115 lower level states 168
La Ferrassie 51, 57, 152 Lower Palaeolithic 145–146,
La Quina 152 149
La Roche Cotard 146 LSA 75, 143–144
ladder-like patterns 80 lunates 101
Lagar Velho 122–123 macaques 2
Lahr 113
Lamarckian 177 M
Lana 15 Madden 9
language 1–7, 13–17, 21–23, magico-religious 58
25–30, 37–42, 50, 97–100, Mahale 6
105, 107, 111–115, 117, 125, Maisha 28
 Index
malformation 30
mammoth 56
man/animal dichotomy 13
mandible 123–124
Mandu Mandu 54
Manufacturing 97, 99, 180
Manufacturing process 97
Manufacturing tools 99
mapping 17, 185
Margulis, Lynn 134
marine 53, 60–64, 83, 92,
112, 123
marine cores 60
marine deposits 83
Marine Isotope Stage 60, 62
marine shell 53
marine shell beads 92, 112, 123
markings 14, 83, 88–90,
92, 125
Matata 15, 27-28
mate 6
material symbols 114
maternal line 117
mathematics 6, 26, 163
mating 114
Maxent 64
Maynard Smith 134–135
McBrearty 24, 50–51, 55–56,
62, 76, 118, 142–143, 149
meaning-making 166, 180
meat 209
medial prefrontal cortex 194
meerkats 7
memes 122
memory 50, 97, 99–100, 105,
149, 165, 169–172, 177–178,
191–196, 198
memory based prediction 171
memory prediction system 172
menstrual 4
mental 6, 9, 14, 31, 77, 97–98,
100, 105–107, 136–137, 165,
174, 185, 191, 193–195, 198,
205–207, 214–220
mental architecture 98
mental mechanisms 194
mental phenomena 185, 191,
194, 195, 198
mental powers 165
mental representations 9, 137,
205, 206, 217
mental rotation 97, 100,
105, 106
mental template 77
metalinguistic 197
metarepresentation 205–206,
216, 218–222
mice 22
microcephalin gene 120
microcharcoal 62
Mid Upper Palaeolithic 153
Middle 41, 52, 64, 75, 80, 91,
97–98, 101, 111, 113, 124,
141–144, 146, 149, 151–153,
155, 185–186, 188–192,
197–200
Middle Palaeolithic 52
Middle Range Theory 91, 185,
186, 191, 198
Middle Stone Age 75, 97–98,
101, 111, 113, 143, 200
Middle to-Upper
Palaeolithic 64
middle-range theories 185,
188–192, 197–200
migration 59
mimicry 5, 8
mind tasks 193
minded agency 206
Miocene apes 152
MIS 60, 62–63, 83
Mithen 22, 99, 192, 219
mitochondria 134
mitochondrial DNA 117
Mitochondrial Eve 123
modern 2, 5, 15, 18, 22–25, 31,
50–51, 53, 59, 75–76, 97,
99–101, 104, 106, 111–115,
117–123, 125, 127, 136,
141–142, 144–145, 150,
153–154, 157–158, 186–187,
189–192, 194–195, 197–200,
205–210, 212–214,
218–220, 222
modern behaviour 65, 75, 97,
106, 144, 187, 197, 200
modern cognitive
capabilities 111
modern cultural traits 59
modern human 22, 50, 51,
53, 59
modern human
immigration 121, 125
modernity 50, 59, 75–76, 124,
142–144, 149–150
modification 25, 83–84, 148,
211, 222
modified lithics 2
modular hierarchical
structure 163–164,
166–167
molecular 13, 22, 120, 134, 170
molecular biology 170
molecular data 120
molecular genetics 13
molecular phylogenetics 22
molecules 167
monoamine systems 175
moral agency 14, 28–29, 40
Morocco 53–54, 98, 112,
146–147
moro-reflex 13
mortuary 150–152, 154–155
mortuary activity 152, 154, 155
mortuary caching 150, 154
MOS 2 62
mother-child 163, 175
mother-child nurturing 163
motif 80, 88–92
motor 5, 13, 22–23, 60, 88, 173,
187, 194, 220
motor control 88, 89, 187, 194
motor inhibition 23
motor patterns 13, 39
motor skills 173
Mousterian 52–53, 55–56,
75–76
Mpumulanga 77
MSA 52, 59, 75–78, 80–84,
88, 92, 97–98, 101–107,
143–144, 153
MSA adhesives 102
Mt. Carmel 116
mtDNA 117, 119–120
multicellular organisms 134–135
multilevel 98–100, 106, 136
multilevel actions 99
multilevel operations 98, 100,
106
multiple 17, 25, 62, 64, 77,
83–84, 149, 152–153, 156
multiple burials 152
multiple dispersals 153
Multiregional 117, 123, 144
Multiregional Evolution 123,
125
Multiregional Hypothesis 117
multispecies 144
Mumba cave 54
Mungo 151
music 163, 171, 173–175, 179–180
musical instruments 56

mutation 22–23, 99, 127, 174,
187, 211
mutational change 133
myths 127, 148
N neuronal activity 167
N. pachyderma 61
Namibia 56–57, 77–78
Nassarius 52–54, 82, 113
Nassarius gibbosulus 53, 54
Nassarius kraussianus 52, 53
natural 1–4, 8–9, 20–21, 39,
52–53, 83–84, 87, 89,
103–104, 106, 119, 133–134,
136, 138, 143, 145–147, 150,
171, 173, 179, 187–189, 191,
206–208, 210–213, 215–216,
218, 222
natural abrasion 84
natural phenomena 52, 211
natural selection 2, 20, 133,
134, 138, 187, 188, 189, 210,
212
Nazlet Kather 51
Neander Valley 115, 152
Neandertal 50–51, 53, 75–76,
111–112, 115–125, 149, 158,
192
Neandertal burials 51, 117
Neandertal ness 115
Neandertal skeleton 121
Neandertals 49–53, 55–59,
63–64, 111–112, 115–125,
141, 150–152, 154, 158, 188,
192, 199–200, 212
Near East 51–53, 55–56, 58, 76,
92, 116, 119, 123, 157–158
necklace 9
Neema 28
neocortex 176, 196
nervous system 30, 33, 170
nests 17, 105
Neumark-Nord 118
neural 17, 40, 99, 163, 165,
172–179, 185–196,
198–200
neural connectivity 99, 163,
174–175
Neural Darwinism 176–177
neural evolution 186, 187
neural mechanisms 191, 193,
194, 196
neural mutation 99
neuroanatomic 196
neuroimaging 195, 199
neurological structure 163
neuromodulators 176
neuromotor 89
neuronal 3, 13, 167, 175
neuronal connections 175
neuronal connectivity 167
neuronal developmental 13
neurons 99, 188, 199
neuroscience 186, 188, 190,
194–198
New Guinea 54
New Ireland 54
nomad 178
non-burnt 78
non-human 1, 3, 5–8, 10, 117,
137, 212–214, 217, 219, 221
non-human language 3, 6, 7
non-human primates 217, 219
non-human species 137, 212,
213, 217, 221
non-linguistic 5, 10
non-primate 50
non-referential models 1
non-religious 210
non-utilitarian 75
non-vocal communication 5
North Africa 53
North Atlantic 60–61
notches 56
nuanced cognitive
development 142
nuclear genome 117, 120
nucleated (eukaryotic)
cells 133
numerical abilities 6
Nunamiut 91
O para-verbally 4
oak bark 118
Oase 123–124
Oceania 58
ochre 14, 22, 51–53, 75–76,
78, 81–84, 86–88, 90–92,
101–106, 142–143, 149,
155–156, 200
ochre stained segments 101
odour 4
Old World 111, 155
optically stimulated
luminescence 78
orange ochre 101
orang-utan 8
Index 
organic 2, 78, 118
organic raw materials 2
organic residue 78, 118
organism 1, 3–4, 21, 40–41, 135
organs of communication 136
ornaments 53, 58, 64, 75–77,
97, 111–115, 120–121, 123,
125, 127, 143, 147
OSL 80, 82–83, 147
ostrich 52–53, 75, 78, 80–81,
83, 88, 91
ostrich eggshell 52, 53, 75, 78,
80, 81, 83, 88
ostrich eggshells 52, 53, 80,
83, 88
Oued Djebbana 53–54, 147
out of Africa 22, 50, 76, 144,
179, 190
P
P. paniscus 2
P. troglodytes 2
palaeoanthropology 141, 157
palaeoclimatic 60, 63–64
palaeocognitive 195
palaeocognitive
neuroscience 195
palaeoenvironmental 49, 63–64
Palaeolithic 9, 52, 64, 75, 114,
126, 142–146, 148–149,
151–153, 157–158
palaeoneurology 197–198
palaeontology 1, 150
Palestine 112, 116
Palmenhorst 77–78
Pan symbolicus 1, 3, 7
Pan/Homo 14–15, 27–30, 35
Pan/Homo bonobos 27
Panksepp 170, 176–177
Parkington 80, 179
parrot 5, 137
part- 115, 123
part-ape/part-man 115
part-Neandertal ancestry 123
past behaviour 190
pattern recognition 163, 165,
170–172
patterning 53, 91, 166, 180
patterns 13, 15, 17, 31, 33–34, 37,
39, 52, 58, 63, 77–78, 80,
82, 88, 90, 92, 121, 164–165,
168–169, 171–176, 179–180,
185, 211
 Index
Pavlovian 137
pecking 77
Penascosa 126
pencil 78
pendant 114, 145
perforated 53–54, 147–148
perineal odour 4
personal 15–17, 24, 29, 37, 41,
53, 58, 64, 75–76, 90, 97,
111–115, 121, 123, 125, 127,
141, 144–146, 148–150,
154–158
personal ego 16, 17
personal ornamentation 123,
141, 144, 145, 146, 148, 149,
150, 154, 155, 156, 157, 158
personal ornaments 53, 58, 64
personal presentation 113
petroglyphs 126
phalange 121
pheromones 4
phonemes 15, 168
phylogenetic proximity 2
phylogenetics 22
phylogeny 196
pictorial representation 173
pierced 121, 123, 147
pierced animal teeth 123
pierres figures 145–146
Pigeons 53–54, 147
pigment 53, 55–56, 58, 76–77,
121, 143–144, 146–149,
155–156
pigment fragments 149, 155
pilot study 8
Pinnacle Point 78
Pit of the Bones 151
planetary diversity 138
plant 2, 101–102, 104–106, 219
plant gum 101, 102, 104, 105, 106
plant prey 2
plastic 25, 101, 212–213
plasticity 13, 22, 105, 176–177,
196, 199
play 2, 4, 17, 38, 49, 63, 113, 163,
173–176, 179–180, 206, 208
play face 4
Pleistocene 60, 124, 142–143,
153, 155
pointing 14, 136–137, 174
points 30, 35, 56, 82–83, 88, 98,
148, 152, 191
polar ice cores 60, 62
polishing 56
pollen 62
polygenetic 189
polygraph 9
pongo-linguistics 3
Porc Epic 143
Portugal 122, 126
post-contact populations 122
Post-Howiesons Poort 78
post-Miocene 3
posture 4, 6, 19, 31, 168
Poverty of Stimulus 171
powder 83–84, 103,
105–106
Pre australopithecines 154
pre-contact 123
predation 23, 150
predators 2, 8, 17, 33, 35,
135, 213
prefrontal cortex 99, 193–194
Pre-Human Religion
Hypothesis 206, 212
pre-language 172, 179
pre-modern 50, 142, 144
Pre-Still Bay 78
primate 1, 4–5, 24, 50, 119,
135–136, 196
primates 22, 24, 30, 35,
50, 76, 135–136, 150,
176–178, 191–192,
195–196, 198,
217, 219
primatological 1, 10, 150
problem-solving 100
prokaryotic 134
prokaryotic cells 134
prominent chin 123
prosocial behaviours 133
prosociality 134
Protoaurignacian 123
proto-cell 134
protohuman language 14
proto-religious 205
P-suke 28
psyche level 13
psychological traits 176
psychology 3, 167–168, 205, 214
psychometric tests 100
psycho-physiology 9
punctured openings 80
Q 208, 220, 222
Qafzeh 52–54, 147, 151
Qena, Egypt 51
quadruped gait 17
quadrupedal 4
quadrupedal bouncing 4
quadrupeds 2
qualitative leap 125
quantum principle 166
quartzite 146
quartzitic 78
Quneitra 52, 54
R
racial diversity 117
rain 8, 17, 35, 212
rain dance 8
Rank/Status system 176
rapid expansion 111
Reading 146, 148, 151, 158, 166,
169, 172–173, 176, 178,
209, 218
recruitment screams 1, 4
recursive 17, 107, 144, 153,
173–175, 178, 180, 187,
192–193, 195, 197
recursive change 144
recursive syntax 187, 192, 193,
195
red 6, 53–56, 61, 101–103, 106,
123, 143, 146, 148
red deer 123
red ochre 53, 101, 102, 103,
106, 143
red pigments 53, 54, 55
red stained 143
referential 1–5, 7, 29, 165
referential models 1
referential signalling 4
reflections 3, 60
reflective mode 16
regional variant 117
reindeer 121
religion 10, 49, 75, 141, 144, 153,
205–216, 218–219, 221–222
religiosity 210
religious 40, 50, 58, 117, 127,
141, 205–216, 218, 220, 222
religious beliefs 50, 117, 127,
206, 212, 218
religious communities 210, 211
religious expression 205, 208
religious ritual 209, 211
religious thought 205, 206,
remains 22, 26, 34, 62, 77, 80,
82, 90–91, 98, 112, 121, 152,
189, 193, 211–212, 221

reproductive separation 120
residue 53, 101, 106, 118
residue analysis 101, 106
resin 101, 105
retouch 101
Rhafas 53
rhythmic facial/vocal
interplay 34
ring-tailed lemurs 10
rituals 8, 209, 211
Riwi 54
rock art 126
Romanian cave 123
Rose Cottage 101
rotational hand-waving 13, 31
rough 4, 176, 179
rough and-tumble 179
rough grunt vocalisations 4
rule-governed 1, 4–5
S shell pendant 145
sandstone 78, 103, 106
sapiens 2, 9, 22, 50–53, 55,
57–59, 63, 75–76, 92, 101,
111–113, 116, 119, 124, 142,
145, 151, 153, 155, 157–158,
187, 189–192, 195, 197,
199–200
Sapir-Whorf hypothesis 6
scavenge 151
scavengers 135
scavenging 151
scent-marks 5
schematization 169
Schöningen 118
scrapers 82, 146
screams 1, 4–5
sculpted ivory figurines 114
sea level 83
segment 101–102
selective 20, 58, 174, 189, 193,
211, 221
selective attention 193
selective pressures 20, 189
self 3, 9, 14–16, 25–26, 29,
37–41, 60, 62, 77
self agency 15
self reflection 14, 16, 39, 41
semantic 4, 147, 166, 169, 194
semiotic 163–164, 167,
171, 180
semiotic function 164, 171
semiotic representation 5
sender-message-receiver 3
sensory 4–5, 7, 165, 172, 174,
177, 196
sensory motor modalities 5
sex hormones 99
sexual 6, 24–25, 99, 138
sexual dimorphism 24, 138
sexual maturity 99
sexual signals 6
sexual swellings 24
shafts 97, 104
shale 83, 103
Shanidar cave 152
shard 87
shared beliefs 76
sharing 49, 165, 186, 209–211,
219
shell 52–54, 56, 76, 82, 92, 97,
112–113, 123, 145, 155, 221
shell beads 52–54, 56, 92, 113,
123, 221
shellfish 82
shelter 52, 56–57, 77–79, 81,
118, 122
Sherman and Austin 26–27
shoe wear 118
short-term memory 194
sibling species 13, 18
Sibudu 53, 98, 101–103
side branch 116
side-scrapers 82
sighs 5
sign 5, 89, 142, 148, 165, 173
sign language 5, 165, 173
sign patterns 165
signal 1, 4–5, 91, 210, 214,
219–220
signalling 4, 210, 220
signed language 5
signifier 142, 165
silcrete 52
siltstone 83, 88
Sima de los Huesos 151
simple burials 155
singing 7, 173, 175
single 5, 9, 19–20, 22–23,
25, 39, 53, 58, 60, 62, 77,
83–84, 87–90, 99, 104, 117,
124, 134–135, 138, 145, 168,
174, 178–179, 211, 219
single cells 135
single package 58
single species 53, 124, 125, 138
skeletal remains 112, 221
Index 
skeleton 22, 51, 115, 121–122, 147
Skhul 53–55, 147–148, 151
skull 111, 123
slave systems 192
Smith Maynard 134-135
smouldering process 119
snail shells 9
sob 5
social 4–6, 8, 16–18, 24, 27–29,
37–41, 49–50, 58–59, 63,
75, 90, 111, 113–114, 117,
121, 125–127, 133–138, 141,
146, 150, 152–155, 163–170,
173–178, 180, 186, 190–191,
199, 209, 211, 218
social bonding 173
social control 138
social display 9
social dominance 4
social identification/
differentiation 127
social identity 113, 114
social insect colonies 134
social insect colony 135
social intelligence
hypothesis 191
social interactions 38, 41,
165, 174
social order 168
social rank 5, 50
social reorganization 17
social scientists 134
sociality 5, 39
societal boundaries 5
socio-ecological 3
solid citizen genes 134
song 1, 4, 23, 99, 163, 174, 179
song elements 1, 4
Sosis Richard 210
sounds 5, 15, 31, 34, 37, 172,
178–179, 212, 214
South Africa 51–53, 55–56, 62,
65, 78, 92–93, 113, 143, 147
south east Asia 54
southern Africa 52, 56, 62–64,
75–76, 80, 97
southern Cape 81–82
Spain 53–54, 121, 151
span tasks 100
spatial 17, 100, 148, 171–172
spatulas 56
speaking 168, 207–208
spear points 56
spearheads 101
 Index
speciation 19–20, 58–60, 76,
111–113, 124
species intersterility 120
speech 3, 5, 22–26, 28, 99,
168–169, 173, 177, 187
spirits 206, 214, 216, 221
spirituality 2, 7, 9–10, 133,
136–138
split brain patients 16
spoken language 5, 26, 28,
173–174
spontaneous referentiality 1, 4
spruce trees 118
SST 62
St.-Césaire 121
stable isotope 62
Stage 3 Project 63
stain 56
Standard View 206–208
state society 144
Stewart Guthrie 213
Still Bay 52–53, 55–56, 78, 82
Still Bay techno-tradition 82
stimulus 6, 166, 171
stimulus bound behavior 166
stimulus freedom 166
stimulus response 6
stone 2, 10, 51, 56–59, 64,
75–77, 83–84, 87,
97–98, 101–106, 111–113,
117, 119, 143, 145, 193, 200
stone hammers 2
stone inserts 97, 101, 105, 97,
101, 105,
stone missiles 10
stone tool 51, 56, 64, 76, 84, 87,
101, 102, 106, 112, 117, 119
stone tool hafting 119
stone tools 51, 56, 64, 76, 84,
97, 101, 102, 106, 117
storage containers 81
Straitzig/Krems-Rehberg 57
stratigraphic 78, 82–83
striations 83–84
Stringer 9, 50–51, 63, 112, 120
style 6, 17, 35, 39, 56, 198
stylistic variability 81
sub-Saharan Africa 75
subsistence 2, 75, 100, 119
supernatural 9, 154–155, 205,
208–210
superorganism 134
suppression of fitness
differences 135
survival 18, 21, 33, 37, 40,
118–119, 163, 170–171, 211
SW France 151
Swabian Jura 113
Sydow Wolfgang 77
symbiotic cell theory 134
symbolic 7–9, 15, 22, 26, 28,
49–51, 56–60, 63–64,
75–77, 88–92, 97–98, 102,
104, 107, 111–114, 122, 124,
127, 133, 136–137, 141–144,
146–150, 152–158, 163–165,
171–174, 179–180, 187–188,
190–191, 200, 205–206, 222
symbolic ability 22
symbolic artefacts 49, 58, 188,
190, 200
symbolic behaviour 22, 49, 58,
76, 92, 133, 144, 149, 157,
188, 200, 205
symbolic capacity 9, 56, 141,
142, 144, 153, 154, 157
symbolic culture 59, 187
symbolic evolution 141, 153, 206
symbolic hypothesis 89
symbolic innovations 98
symbolic intent 76, 92
symbolic material 49–51, 58,
75, 97
Symbolic Neandertal 50–51
symbolic thinking 50, 111–113,
124, 127, 172
symbolically mediated
behaviour 49–50, 76
Symbolism 49–50, 75, 97–98,
102, 107, 125, 141–151,
153–156, 158, 172, 176, 179,
198, 205–206, 218–222
symbols 6, 14, 26–27, 50, 56,
75, 82, 89–90, 92, 114, 117,
141–142, 145, 148, 150, 152,
164–165, 171–172, 178, 180
symbol-use 3, 6–7, 9
syncopated gallop 4
syntactic 4, 167, 173, 191, 193, 198
syntactic architecture 173
syntactic category 167
syntactic hypothesis 191, 193
syntactic integration 173
syntactic processing 4
syntax 4, 15, 27, 169, 173, 175,
187, 192–193, 195, 197
synthetic iron oxide 103
Szathmary 134–135
T
Tabun 51, 152
Tabun C1 burial 51
Taforalt 54, 98, 147
Tai Forest 8
talk 98, 168
talking apes 26
Tamuli 28
tanning of hides 118
Tan-Tan 146
Tanzania 8
Taphonomic 84
task-appropriate recipes 102
Tasmania 58–59
taxonomic status 120
teamwork 133–134, 136–138
techno-complexes 56
techno-tradition 78, 82
teeth 24–25, 121, 123
temporoparietal 193–194, 197,
199
Terrence Deacon 137
terrestrial 63–64
Texier 52, 78, 80–81, 90–91
The Symbolic Species 137
theory of mind 133, 153–154,
178–179, 191–195, 198, 215,
218, 221–222
therianthropic 113
therianthropic ivory
figurines 113
thermoluminescence 78, 82
thermoregulatory models 118
Thomas Wynn 191–192
thought experiments 99
tiger shark tooth 54
time depth 144
time/space-factored 148–149,
152–153, 155–157
TL 78, 80, 82, 147
Toba super-eruption 62
Tobias 179
ToM 29, 133, 215–218, 220–222
Tomasello 30, 89, 133, 136–137,
169, 171, 176–178, 191, 197,
218–221
tool kit 207, 218, 220
tools 10, 49, 51, 56–58, 60,
63–64, 75–77, 82, 84, 88,
90, 97, 99, 101–107, 117,
125, 174, 195–199, 211
tooth enamelisation 25
top-down causation 163,
167–169
 Index 

toys 180 Venus pendant 114 white matter 196

traces 53, 98, 102, 180, verbal self-report 3 white sclera 136
214, 221 Verlorenvlei River 78 Wisconsin Card Sort
transition 31, 64, 133–136, 138, vertebrates 177 Task 100
168, 179, 197 Vindija 117 within-group 133–136
Trevarthen 31, 34 vision 49, 171–173 within-group
triadic awareness 5 visual thinking 163, 172–174 competition 135
Trinkaus 112, 115, 122–123, 153 visuospatial 192 within-group
Tübingen 54, 114 vocal 4–5, 7–8, 15, 22–23, 28, cooperation 134
Turkey 147 30–31, 34, 37–38, 168, 174 within-group selection 134,
Twin Rivers 147 vocal chords 168 136
twine 101 vocal fluency 15, 23 Wolfgang Sydow 77–78
vocal imitation 8 wolves 30, 136
U von Frisch 7 wooden artefacts 118
Ucagizli Cave 147 working memory 50, 97,
Upper Palaeolithic 9, 64, 75, W 99–100, 105, 191–196, 198
142, 144, 148, 153 Wamba 8, 15 working memory
Upper Pleistocene 142–143 warfare 136 hypothesis 191
upraised arms 4 warming 60–62 writing 113, 141, 144, 168, 173
Ursus deningeri 151 water-proof clothing 118 Wynn and Coolidge 100
usage-based language weapons 97, 101–102,
skills 168 105, 118 Y
use-wear 53, 121 Wernicke’s area 174, 193, 195 yellow 28, 101, 103
Western Australia 54 Yerkes 26
V Western Cape 52, 77–78, 98
Varga-Khadem 22 whales 7 Z
vegetable 84 whistling 8 Zambia 147
Venus 114 white 5, 17, 23, 136, 196 zoological specimens 14
View publication stats