The cultural evolution of mind reading

Abstract

It is not just a manner of speaking: “Mind reading,” or working out what others are thinking and feeling, is markedly similar
to print reading. Both of these distinctly human skills recover meaning from signs, depend on dedicated cortical areas, are
subject to genetically heritable disorders, show cultural variation around a universal core, and regulate how people behave.
But when it comes to development, the evidence is conflicting. Some studies show that, like learning to read print, learning
to read minds is a long, hard process that depends on tuition. Others indicate that even very young, nonliterate infants are
already capable of mind reading. Here, we propose a resolution to this conflict. We suggest that infants are equipped with
neurocognitive mechanisms that yield accurate expectations about behavior (“automatic” or “implicit” mind reading), whereas
“explicit” mind reading, like literacy, is a culturally inherited skill; it is passed from one generation to the next by verbal
instruction.

Full text

DOI: 10.1126/science.1243091
, (2014);344 Science
Cecilia M. Heyes and Chris D. Frith
The cultural evolution of mind reading
This copy is for your personal, non-commercial use only.
clicking here.colleagues, clients, or customers by , you can order high-quality copies for yourIf you wish to distribute this article to others
here.following the guidelines can be obtained byPermission to republish or repurpose articles or portions of articles
): August 16, 2014 www.sciencemag.org (this information is current as of
The following resources related to this article are available online at
http://www.sciencemag.org/content/344/6190/1243091.full.html
version of this article at: including high-resolution figures, can be found in the onlineUpdated information and services,
http://www.sciencemag.org/content/suppl/2014/06/19/344.6190.1243091.DC1.html
can be found at: Supporting Online Material
http://www.sciencemag.org/content/344/6190/1243091.full.html#related
found at: can berelated to this article A list of selected additional articles on the Science Web sites
http://www.sciencemag.org/content/344/6190/1243091.full.html#ref-list-1
, 13 of which can be accessed free:cites 51 articlesThis article
http://www.sciencemag.org/cgi/collection/psychology
Psychology subject collections:This article appears in the following
registered trademark of AAAS. is aScience2014 by the American Association for the Advancement of Science; all rights reserved. The title CopyrightAmerican Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005.
(print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by theScience
on August 16, 2014www.sciencemag.orgDownloaded from on August 16, 2014www.sciencemag.orgDownloaded from on August 16, 2014www.sciencemag.orgDownloaded from on August 16, 2014www.sciencemag.orgDownloaded from on August 16, 2014www.sciencemag.orgDownloaded from on August 16, 2014www.sciencemag.orgDownloaded from on August 16, 2014www.sciencemag.orgDownloaded from on August 16, 2014www.sciencemag.orgDownloaded from

RESE ARCH
20 JUNE 2014 • VOL 344 ISSUE 6190 1357SCIENCE sciencemag.org
PHOTO: TETRA IMAGES/ALAMY
BACKGROUND: We use “theory of mind”
or “mind reading” to understand our own
thoughts and feelings and those of other
agents. Mind reading has been a focus of
philosophical interest for centuries and of
intensive scientific inquiry for 35 years. It
plays a pivotal role in human social inter-
action and communication. Mind reading
allows us to predict, explain, mold, and
manipulate each other’s behavior in ways
that go well beyond the capabilities of other
animals; therefore, mind reading is crucial
to understanding what it means to be hu-
man. In many respects, the capacity to read
minds is like the capacity to read print: It
involves the derivation of meaning from
signs, depends on dedicated brain mecha-
nisms, is subject to specific developmental
disorders, shows cultural variation as well
as cultural commonality, and has both in-
terpretive (reading) and regulative (writ-
ing) functions. However, recent studies of
mind reading in young infants suggest that,
unlike print reading, mind reading devel-
ops very early in human ontogeny.
ADVANCES: In nonverbal tests of mind
reading, infants’ eye movements have been
taken as evidence that infants expect an
agent to reach toward a location where he or
she believes a desirable object to be hidden,
even when the agent’s belief is false. This
“implicit” mind reading could indicate that
humans genetically inherit the specialized
neurocognitive mechanisms used for “ex-
plicit,” verbally mediated mind reading in
adulthood. However, recent research with
adults shows that, unlike explicit mind read-
ing, implicit mind reading does not make
demands on executive function. This indi-
cates that, although they may be genetically
inherited, the mechanisms that mediate
implicit mind reading, whether specialized
or general-purpose, are distinct from those
controlling explicit mind reading. Further-
more, studies of twins, people with hearing
impairments, and children from non-West-
ern cultures, as well as typically developing
Western children, suggest that, like print
reading, explicit mind reading is culturally
inherited. Rather than being constructed by
simulation or theory-testing, explicit mind
reading is a skill that is passed from one
generation to the next by verbal instruction.
Most, possibly all, human neurocogni-
tive skills are shaped by culture and many
are culturally inherited, but the parallels
between mind reading and print reading
are extraordinary. In contrast, whereas lin-
guistic communities vary in the ways that
they categorize colors, color perception is
not culturally inherited in the same way as
print reading. Unlike print reading, color
perception is rooted
in highly specialized,
genetically inherited
mechanisms that hu-
mans share with other
species. Though cul-
tural input adjusts
these mechanisms, it does not make them
into a whole new neurocognitive system.
OUTLOOK: The cultural evolutionary ac-
count of mind reading does not imply that
mental states are mere fictions, but it does
suggest that any aspect of mind reading—
even those relating to knowledge and pri-
mary emotions—could show substantial
cultural variation. More cross-cultural stud-
ies using sensitively translated test proce-
dures are needed to chart extant variation.
Similarly, although the cultural evolutionary
account suggests that humans do not genet-
ically inherit mechanisms that are special-
ized for the representation of mental states,
it assumes that, as in the case of print read-
ing, many of the neurocognitive raw mate-
rials for explicit mind reading are inborn.
Therefore, priorities for future research
are to identify the genetic “start-up kits”
for both implicit and explicit mind reading
and to find out exactly how the products of
the former contribute to the development
of the latter. Our view suggests that, like
print reading, mind reading is a culturally
inherited skill that facilitates the cultural
inheritance of other, more specific skills;
mind reading is a cultural gift that keeps
on giving.
The cultural evolution
of mind reading
SOCIAL COGNITION
Cecilia M. Heyes1* and Chris D. Frith2
1All Souls College and Department of Experimental Psychology,
University of Oxford, Oxford OX1 4AL, UK. 2Wellcome Trust
Centre for Neuroimaging, University College London, London
WC1N 3 BG, UK.
*Corresponding author. E-mail: cecilia.heyes@all-souls.ox.ac.uk
Cite this a rticl e as C. M. H eyes, C. D. Frit h, Science 344,
1243091 (2014). DOI: 10.1126/science.1243091
Read the full article
at http://dx.doi
.org/10.1126/
science.1243091
ON OUR WEBSITE
Learning to read minds is like learning to read print. The acquisition of explicit mind reading
is a slow, effortful process in which a novice develops an important, culture-specific skill through
expert tuition. Experts facilitate development by directing the novice’s attention to signs that the
novice is on the edge of understanding, as well as by explaining in conversation how these signs
relate to their meaning.
REVIEW SUMMARY
Published by AAAS

REVIEW
◥
SOCIAL COGNITION
The cultural evolution of mind reading
Cecilia M. Heyes
1
*and Chris D. Frith
2
It is not just a manner of speaking: “Mind reading,”or working out what others are thinking
and feeling, is markedly similar to print reading. Both of these distinctly human skills
recover meaning from signs, depend on dedicated cortical areas, are subject to genetically
heritable disorders, show cultural variation around a universal core, and regulate how
people behave. But when it comes to development, the evidence is conflicting. Some
studies show that, like learning to read print, learning to read minds is a long, hard process
that depends on tuition. Others indicate that even very young, nonliterate infants are
already capable of mind reading. Here, we propose a resolution to this conflict. We
suggest that infants are equipped with neurocognitive mechanisms that yield accurate
expectations about behavior (“automatic”or “implicit”mind reading), whereas “explicit”
mind reading, like literacy, is a culturally inherited skill; it is passed from one generation to
the next by verbal instruction.
“Why did she do that?”Neuroscience
tries to answer this kind of question
by referring to neural processes,
butineverydaylifeweinvokethe
actor’s thoughts and feelings. We
predict and explain behavior by ascribing be-
liefs, desires, and intentions to the actor; for ex-
ample, “She destroyed the evidence because she
thought it was incriminating,”or “She nodded at
the auctioneer because she wanted to bid for the
Miró.”Our ability to ascribe mental states to our-
selves and others is known as “theory of mind,”
“mentalizing,”“folk psychology,”or “mind read-
ing.”It has been a major focus of philosophical
investigation for centuries and of scientific en-
quiry for 35 years (1). Mind reading has been
studied intensively because it is thought to play
a pivotal role in human social interaction and
communication. Mind reading allows us to predict,
explain, mold, and manipulate each other’sbe-
havior in ways that go well beyond the capa-
bilities of other animals. Understanding mind
reading is therefore crucial to understanding
what it means to be human.
Until now, the term “mind reading”has been
used in a casual, eye-catching way to suggest
kinship with telepathy or to highlight a super-
ficial resemblance between the ascription of
mental states and print reading; both involve
the derivation of meaning from signs. In print
reading, the signs are marks on paper and their
meaning relates to objects and events in the
world. In mind reading, the signs are facial ex-
pressions, body movements, and utterances, and
their meaning relates to the actor’s mental states.
In this article, we argue that the resemblance
goes much deeper (Table 1). We suggest that
mind reading, like print reading or literacy, is
culturally inherited.
Literate people have neurocognitive mecha-
nisms that are specialized for reading. If literacy
were an ancient feature of human life, it would
be natural to assume that these mechanisms
evolved by genetic means. However, because lit-
eracy originated only 5 to 6 thousand years ago,
these specialized neurocognitive mechanisms are
thought to be products of cultural inheritance or
cultural evolution. Cultural evolution has been
modeled in a number of ways (2). The basic idea is
that the reading and writing systems used across
the world today have developed, diversified, and
been honed by nongenetic evolutionary processes.
These processes are evolutionary because they
producechangethroughheritablevariationinfit-
ness. However, the inheritance occurs via social
learning rather than genetic mechanisms, and
fitness is defined by the number of individuals or
groups who adopt the system through social
learning, rather than by the number of biological
offspring. On this kind of evolutionary analysis,
when people are trained to read and write, they
culturally inherit from their teachers a body of
knowledge (for example, about rules governing
character inscription and pronunciation) and a
set of skills (for example, in distinguishing words
from other objects) that are embodied in spe-
cialized neurocognitive mechanisms. These mech-
anisms are constructed or “recycled”(3)from
genetically evolved mechanisms in the course of
literacy training, but they have not been direct
targets of gene-based natural selection.
We begin this Review by discussing features of
mind reading—its neural bases, developmental
disorders, cultural variation, and functions—that
might be thought to show that mind reading de-
pends on genetically evolved neurocognitive mech-
anisms. In each case, we point out that print
reading has similar features. Because it is known
that print reading is culturally rather than genet-
ically inherited, these similarities between mind
reading and print reading clear the way for us to
consider the possibility that mind reading is also
culturally inherited. In the second section, we dis-
cuss recent studies indicating that infants are
capable of mind reading. These studies apparently
support the view that mind reading depends on
genetically evolved mechanisms because there is
verylittleopportunityforculturalinheritancein
the first few months of life. However, we will
suggest that, when they seem to be mind reading,
infants are not using the mechanisms that control
“full-blown”or “explicit”mind reading in adults—
mechanisms that allow us to deliberate and talk
about mental states. In the third section, we turn to
direct evidence that explicit mind reading is cul-
turally inherited. This research indicates that, like
learning to read print, learning to read minds in an
explicit manner is a slow process that depends on
instruction (Fig. 1).
Mind reading and print reading
Neural bases
Neuroimaging has shown that adults have cor-
tical circuits specialized for mind reading. These
circuits, which include the medial prefrontal cortex,
temporo-parietal junction, and precuneus (Fig. 2),
aremoreactivewhenpeoplearethinkingabout
mental states than when they are performing
similar tasks that do not involve thinking about
mental states (4). It is tempting to assume that
this sort of cortical specialization is due to genetic
evolution. However, similar specialization has been
found in the case of print reading. For example, in
literate adults an area of the occipito-temporal
cortex known as the “visual word form area”is
more active during the presentation of words than
comparable nonword stimuli (5).
Developmental disorders
Many people with autism spectrum conditions
(ASCs) have a specific impairment in mind read-
ing. Compared with IQ (intelligence quotient)–
and language-matched controls, people with ASCs
have difficulty working out what others are think-
ing and feeling (6). There is a substantial genetic
contribution to the heritability of ASCs, which
might suggest that neurocognitive mechanisms
specialized for mind reading are usually geneti-
cally inherited and that people with ASCs geneti-
cally inherit atypical versions of these mechanisms.
But dyslexia reminds us that this is only one of
many candidate explanations for the mind-reading
impairment associated with ASCs. Dyslexia is a
genetically heritable developmental disorder that
interferes with the acquisition of a culturally inher-
ited skill: learning to read print (7).
Cultural variation
At first glance, it seems that print reading is
characterized by cross-cultural diversity and mind
reading by cross-cultural commonality. In the print
domain, there is certainly a rich diversity of scripts,
and the size of the speech units that are mapped
onto printed units varies from whole words in
Kanji, to syllables in Japanese Kana, to phonemes
in alphabetic writing systems (8). Furthermore, in
the mental domain, it is commonly assumed that
RESEARCH
1
All Souls College and Department of Experimental
Psychology, University of Oxford, Oxford OX1 4AL, UK.
2
Wellcome Trust Centre for Neuroimaging, University College
London, London WC1N 3BG, UK.
*Corresponding author. E-mail: cecilia.heyes@all-souls.ox.ac.uk
SCIENCE sciencemag.org 20 JUNE 2014 •VOL 344 ISSUE 6190 1243091-1

the members of all cultures ascribe thoughts and
feelings and understand these states to be related
to behavior.
However, recent research suggests that print
reading is more (and mind reading is less) uniform
across cultures than was previously supposed. For
example, featural analysis of 115 writing systems,
contemporary and historical, has shown that most
characters are formed by three strokes (9). In a com-
plementary way, psychological experiments are
beginning to reveal substantialcross-culturalvar-
iation in the development of mind reading. For
example, children in Australia and the United
States understand that different people can have
different opinions (“diverse beliefs”)beforethey
understand that a person can be knowledgeable
or ignorant about a particular fact (“knowledge
access”). In contrast, children in China and Iran,
where individualism and self-expression are less
important, understand knowledge access before
they understand diverse beliefs (10). Research of
this kind supports earlier ethnographic data sug-
gesting that cultures vary widely in the impor-
tance they assign to mental states, rather than
situations, as causes of behavior; whether they
take mental states to reside inside or outside the
body; and the extent to which they regard mental
states as subject to natural rather than super-
natural laws (11). This research also converges
with literary historical studies indicating that,
in Western cultures, ideas about the mind and
mental disorders have changed radically since
ancient times (12,13).
Interpretation and regulation
Literacy has both interpretative (reading) and reg-
ulative (writing) aspects. As children are tutored
in the writing conventions of their language, they
learn not only to decode printed words (i.e., to
relate them to spoken words and to meaning) but
also to produce printed words that obey the con-
ventions of their writing system and can therefore
be decoded by others. Research on mind reading
has emphasized its interpretive role: the way that
the attribution of mental states allows us to ex-
plain and predict behavior. However, “mental lit-
eracy”also has an important regulative role (14).
Novice mind readers learn that not only can
behavior be produced by rational interactions
amon g beliefs and desires, but also that it should
be, and they are encouraged to make their own
behavior obey these conventions. Recent research
in social psychology and cognitive neuroscience
has provided clear evidence in support of this reg-
ulative hypothesis. When adults are discouraged
from believing in free will—for instance, when
they are told that deterministic neurological
processes, rather than mental states, control
behavior—there is a weakening of neural signals
Table 1. Parallels between mind reading and print reading.
Feature Mind reading Print reading
Meaning from signs Mental states from situations and behavior Referents from inscriptions
Neural specialization E.g., medial prefrontal cortex and temporo-
parietal junction (4)
E.g., occipito-temporal
“visual word form area”(5)
Developmental disorders Autism spectrum conditions (6) Dyslexias (7)
Cultural commonality Thoughts and feelings can influence
behavior (11)
Number of strokes per character
and degree of redundancy (9)
Cultural variation Importance, range, location, and causation
of thoughts and feelings (10–13)
Importance, range, and medium of print;
print-to-pronunciation rules (8)
Interpretation Mental states from situations and behavior Reading
Regulation Promotes socially acceptable behavior
(15,16)
Writing and spelling
Implicit competence Accurate expectations about behavior
(17–19,25–29)
Pictorial “reading”
Slow development Specialization continues into late
adolescence (32–34)
Specialization continues into
late adolescence (31)
Cognitive effort Uses executive processes (26) Uses executive processes
Individual differences Due to environmental rather
than genetic factors (35)
Due to environmental rather
than genetic factors (41)
Development depends on
tuition: engineering
Desires and emotions before
beliefs and knowledge (43–45)
Easy words before hard words
Development depends on
tuition: instruction
Causal-explanatory relations
among mental states, situations,
and behavior (38–41,43–48)
Rules relating print to referents
and pronunciation
Fig. 1. Learning to read minds is like learning to read print. It is a slow, effortful process in which a
novice develops an important, culture-specific skill through expert tuition. Experts facilitate the devel-
opment of explicit mind reading by directing the novice’s attention to signs [(A) behaviors, (B)words]
that the novice is on the verge of understanding. Experts also explain in conversation how these signs
relate to their meaning [(A) mental states, (B) objects and events].
1243091-2 20 JUNE 2014 •VOL 344 ISSUE 6190 sciencemag.org SCIENCE
RESEARCH |REVIEW

associated with action planning (15), and they
exhibit more antisocial cheating and aggressive
behavior (16).
Implicit mind reading in infants
and adults
In contrast, studies of infants suggest that the
development of mind reading is very different
from that of print reading. Western children do
not typically acquire the skill of print reading
until they are 5 or 6 years old, but infants as young
as 7 months old seem to be capable of mind read-
ing (17). The most notable results suggest that
infants ascribe false beliefs to agents; they under-
stand that an agent can believe something that
does not match the physical reality of the situation.
For example, in a typical experiment (18), infants
first watch an adult placing a toy in one of two
boxes, green rather than yellow, and subsequently
reaching repeatedly for that box. Then they see the
toy moved from the green box to the yellow box
when the adult is present (true-belief condition) or
absent (false-belief condition). In the final phase,
infants in the true-belief condition show more
surprise, measured by looking time, when the
adult reaches for the green box rather than the
yellow box, whereas infants in the false-belief con-
dition are more surprised when she reaches for
the yellow box. In both cases, the infant “expects”
a reach toward the box where the adult believes
the toy to be located.
Research of this kind, which infers the oc-
currence of mind reading from nonverbal be-
havior such as looking time, is said to provide
evidence of “implicit”mind reading. Implicit mind
reading can be interpreted in three ways. The
continuity interpretation suggests that it is con-
trolled by the same neurocognitive mechanisms
that mediate explicit mind reading in adults (19).
The two-systems interpretation, favored by Frith,
suggests that implicit and explicit mind reading
arise from different neurocognitive mechanisms,
but both systems are specialized for thinking
about mental states (20,21). This is supported by
the case of autism, in which there is evidence of
specific failures in mind reading, whereas other
domains remain intact (22). The implicit system
develops early and tracks mental states in a fast
and efficient way, whereas the explicit system
develops later, operates more slowly, and makes
heavier demands on executive functions such as
working memory and inhibitory control. In com-
mon with the two-systems model, the submental-
izing interpretation, favored by Heyes, assumes
that explicit mind reading depends on specialized
mechanisms that are late developing, slow, and
effortful. However, the submentalizing view suggests
that implicit mind reading depends on domain-
general neurocognitive mechanisms, rather than
mechanisms that are specialized for thinking
about mental states (23,24).
The case of autism suggests that implicit
and explicit mind reading depend on different
mechanisms because explicit mind reading can
be achieved in spite of continuing problems with
implicit mind reading (25). Further evidence for
this dissociation is found in studies of neuro-
typical adults. In tasks in which adults make
verbal judgments about other people’sthoughts
and feelings (explicit mind reading), judgment
accuracy is impaired by concurrent performance
of an executive function task (26). In contrast,
concurrent demands on executive function do
not interfere with implicit mind reading (27).
This dissociation is hard to reconcile with the
continuity hypothesis but is compatible with the
two-systems and submentalizing interpretations
of implicit mind reading.
Evidence that implicit mind reading is
automatic—that it does not make demands on
executive function—comes from a “dot perspec-
tive task”in which participants see pictures of
a room (Fig. 3). There is a human figure (an
“avatar”) in the middle of the room, facing to
the right or left, and dots on the walls. The par-
ticipant is required to report the total number of
dots in each picture. Implicit mind reading is
inferred from the fact that responses are slower
and less accurate when some of the dots are
behind the avatar than when they are all in front
(28). On the two-systems account, this “self-other
consistency”effect is due to automatic calcula-
tion by a specialized system of the number of
dots that the avatar can see (i.e., of the content
of the avatar’s mental state). This automatic
60
40
20
0
-20
-40
-100 -80 -60 -40 -20 0 20 40 60 60 40 20 0 -20 -40 -60 -80 -100
AB
Fig. 2. Brain regions activated by tasks involving mind reading (4). (A) A region located on the
border of the posterior temporal sulcus and temporo-parietal junction is shown on a diagram of the
lateral surface of the right hemisphere embedded in Talairach coordinates. (B) Two regions, the medial
prefrontal cortex (left) and the precuneus (right), are shown on a diagram of the medial surface of the
right hemisphere embedded in Talairach coordinates.
Fig. 3. Stimuli used in the dot perspective task. Adults are asked to report the number of dots
they can see. Providing evidence of implicit mind reading, the adults make a faster judgment when all
dots are in front of the avatar (A) than when some dots are behind the avatar (B). This difference
also occurs when the central object is an arrow [(C) versus (D)], suggesting that it may be due to
domain-general neurocognitive mechanisms that mediate automatic attentional orienting (29).
SCIENCE sciencemag.org 20 JUNE 2014 •VOL 344 ISSUE 6190 1243091-3
RESEARCH |REVIEW

calculation gives a different result from the cal-
culation of the number of dots the participant
can see and thus causes interference. On the
submentalizing account, the effect is due to
automatic attentional orienting to the gaze di-
rection of the avatar (29);thesamedomain-general
process that allows inanimate stimuli, such as
arrows, to direct attention (30).
According to both accounts, however, explicit
mind reading—the kind that allows us to delib-
erate about mental states and to express our
thoughts about mental states in words—is like
print reading: it develops slowly and is cogni-
tiv ely de manding. The slow development of print
reading is reflected in the fact that event-related
potentials indicative of reading fluency are not
quiteasfastin16yearoldsasinmatureadults
(31). Similarly, it is now known that in Western
cultures, the neural systems implicated in expli-
cit mind reading (Fig. 2) are among the last to
reach maturity (32), and performance on explicit
tests of mind reading (including perspective-
taking, emotion recognition, and detection of pre-
tense and irony) continues to improve between
adolescence and adulthood (33,34).
Learning to read minds
Further evidence that the development of expli-
cit mind reading depends on a slow process of
learning, rather than the maturation of geneti-
cally inherited neurocognitive mechanisms, comes
from a major twin study (35). When more than one
thousand twin pairs were given a comprehensive
battery of explicit mind-reading tests at 5 years old,
the correlation in performance within pairs was
the same (0.53) for nonidentical twins, with an
average of 50% of genes in common, and for iden-
tical twins, who have all of their genes common.
This indicates a “substantial shared environmental
influence but negligible genetic influence on in-
dividual differences in theory of mind”(35).
However, by itself, this twin study does not tell
us about the nature of the environmental influ-
ence or the kind of learning involved in the de-
velopment of explicit mind reading. There are
two well-established accounts of mind reading:
simulation theory and theory-theory. Although
they were previously treated as competing ap-
proaches to the explanation of mind reading, it
is now widely accepted that these are comple-
mentary accounts with theory-theory having a
role in explicit high-level mind reading, whereas
simulation theory is more relevant to implicit
low-level mind reading (36). Theory-theory pro-
poses that children learn like scientists, developing
explicit theories of how mental states cause be-
havior through observation and hypothesis test-
ing (37). In this case, the ideas are not only honed
and tested, but also generated by the novice mind
reader herself. In contrast, our cultural evolution-
ary view suggests that the novice’sideasabout
the mind are derived primarily not from obser-
vation but from instruction, from what the nov-
ice is told about the mind by the expert mind
readers in her social world.
A“natural experiment”has provided compelling
evidence that language-based tuition is crucial in
the development of mind reading. This natural
experiment was the emergence in the 1970s of
a new sign language (NSL) among deaf people in
Nicaragua. Pyers and Senghas compared false-
belief understanding in two adult cohorts of NSL
users. They found that the first cohort, who learned
NSL when it was still a rudimentary language,
was less able to understand false beliefs than the
second cohort, who learned NSL ~10 years later
when it included many more signs for mental
states. This suggests that although adults in the
first cohort were 10 years older—and therefore
had 10 years more opportunity to reflect on their
own mental states and to watch the behavior of
social partners—their understanding of false be-
lief lagged behind that of adults in the second
cohort because, lacking mental-state vocabul-
ary, they had not been able to receive instruction
about the mind (38). These results are supported
byarecentstudyshowingthatchildreninSamoa,
where it is considered improper to talk about
mental states, develop an understanding of false
belief at ~8 years of age, 4 or 5 years later than
children in Europe and North America (39).
In combination with other research involving
typically and atypically developing children (40,41),
the NSL and Samoan studies indicate that we
learn about the mind through conversation about
the mind. In principle, the appropriate conversa-
tional experience could come from eavesdropping;
that is, by listening to what expert mind readers say
when they have no intention of teaching a novice.
However, many studies suggest that experts,
especially mothers, tailor or “epistemically engi-
neer”(42) their conversation about the mind so
that it helps children to learn (43–45). For ex-
ample, in conversation with their 15 month old
children, mothers make proportionally more ref-
erences to desires and emotions than to thoughts
and knowledge, and the frequency of desire/
emotion talk selectively predicts the children’s
explicit mind-reading performance at 24 months.
In contrast, at 24 months, mothers talk propor-
tionally more about thoughts/knowledge than
desires/emotions, and it is the frequency of ref-
erences to thoughts/knowledge that predicts ex-
plicit mind reading at 33 months (43). Desires/
emotions are easier to understand because infants
and children are constantly attempting to fulfil
their desires, and emotions are often reflected in
distinctive facial expressions. Therefore, these
findings suggest that, just as novice print readers
are introduced to easy words before hard words,
novice mind readers are introduced to easy states
before hard states, and this epistemic engineer-
ing of the child’s learning environment pro-
motes the development of mind reading.
The development of explicit mind reading is
notonlypredictedbythefrequencywithwhich
mothers use mental state terms—such as “think,”
“want,”and “happy”—in conversation with their
children; it is also predicted by the frequency of
mothers’“causal-explanatory”statements about
the mind, specifying relations between situations,
mental states, and observable behavior. For ex-
ample, “Sheissmilingbecausesheishappy,”and
“Shethinksthetoyisinthegreenboxbecause
she didn’t see it moved to the yellow box”(41). In
addition, Lohmann and Tomasello (46)found
that training sessions in which 3 and 4 year olds
conversed with an adult about deceptive objects
(e.g., a pen that looks like a flower) were especially
effective at improving false-belief understand-
ing when the adult used sentential complement
syntax: sentences such as “What do you think it
is?”and “You say it is a flower,”which take a full
clause as their object complement. Research on
causal-explanatory statements and sentential com-
plements (47) suggests that conversation about
the mind does not merely teach children labels
for mental states. It also teaches them mental
state concepts—what it is to “think”or to “feel”
something, to be “happy”or “doubtful”—and gives
them a format in which to represent these con-
cepts. In other words, children culturally inherit
from their parents, and other mind-reading ex-
perts (48), mechanisms that are specialized for
the representation of mental states.
In emphasizing the importance of conversation
and the possibility of marked cross-cultural var-
iation, our view is similar to social constructivist
accounts of mind reading (49). However, unlike
many social constructivists, and in common with
theory-theory, we assume that the processes in-
volved in the development of explicit mind read-
ing are broadly rational and yield conceptual
knowledge about the mind. It is likely that some
elements of mind reading, such as those relating
to knowing or to primary emotions, will be found
to show less cross-cultural variation than others,
but our framework allows substantial variation,
even in the “heartland”of mind reading. For ex-
ample, given the structure of the human nervous
system, it is likely that, in all cultures, looking is
predictive of behaviors that, in many cultures,
are understood to indicate knowing. But there
could still be substantial variation in beliefs about
the relation between seeing and knowing (50). Dis-
tinctive assumptions about the origin, location,
and functioning of mental states will lead the mem-
bers of different cultures to attend to different
aspects of observable behavior, and because mind
reading has regulative as well as interpretive
functions, distinctive beliefs about the mind will
lead to different patterns of observable behavior.
Start-up kit
In contrast with nativist theories of mind reading,
our cultural evolution hypothesis suggests that
humans do not genetically inherit neurocognitive
mechanisms that are specialized for the devel-
opment of explicit mind reading. Nonetheless,
we naturally assume that genetically evolved
mechanisms provide much of the raw material
for the construction of explicit mind reading: the
mechanisms that become specialized for repre-
sentation of mental states and the processes that
make cultural inheritance possible. Our perspec-
tive implies that a priority for future research is
to identify the genetic “start-up kit”(22)forthe
cultural inheritance of mind reading. No one has
made a concerted attempt to train chimpanzees
to read minds, but few, if any, researchers would
expect this enterprise to be a success. Why? What
1243091-4 20 JUNE 2014 •VOL 344 ISSUE 6190 sciencemag.org SCIENCE
RESEARCH |REVIEW

are the genetically inherited neurocognitive re-
sources that enable humans, but not chimpan-
zees, to culturally inherit mind reading?
Implicit mind reading
The healthy human infant comes into the world
armed with a number of brain mechanisms that
have a critical role in learning about the behav-
ior of others. Some of these mechanisms are
primitive and held in common with many other
species. However, their presence in humans, in
conjunction with dramatically enhanced learn-
ing processes, creates the potential for acquiring
powerful social abilities. From birth, human in-
fants attend to face-like stimuli and, by the end
of their first year, can use faces to recognize iden-
tity, race, emotion, and gaze direction (51). Human
infants also show a preference for observing
moving limbs (52) and, by the age of 12 months,
show predictive eye movements when observing
the actions of others (53). In other words, they
can predict where a movement will end, given
how it starts. Infants are also attending to their
own movements, thereby creating the link be-
tween action observation and action production
that generates the motor resonance fundamental
to accounts of low-level mind reading associated
with simulation theory (54).
Much further work is needed to discover how
and to what extent these low-level observational
mechanisms contribute to the development of
explicit mind reading. Again, the analogy with
printreadingmaybehelpful.Throughlistening
to speech, the infant discovers statistical regu-
larities that define components such as words
and syllables (55). The implicit knowledge of
such components is a vital precursor for learn-
ing how to map sounds onto strings of letters.
Likewise, there are detectable statistical regu-
larities associated with body movements derived,
in part, from the physical constraints created by
gravity and the construction of mammalian bodies
(56), which define a vocabulary of action com-
ponents. At the same time, infants, through the
experience of their own instrumental actions,
are learning to categorize actions in terms of
features such as success and efficiency (57). This
implicit knowledge not only provides a basis for
expectations about the outcomes of observed ac-
tions (58), but also provides predetermined tar-
gets for instructions about how mental states
are related to actions.
Explicit mind reading
Implicit mind reading emerges from observing
the behavior of others. This is a one-way process.
The learner observes the actor, who need not be
aware that he or she is being observed. In con-
trast, explicit mind reading emerges from the
deliberately instructive behavior of others. This
is essentially a two-way process. The behavior of
the actor is designed to help the observer learn,
and both actor and observer are actively engaged
in a communicative process. Components of a
start-up kit for such processes might include the
preference of newborn infants for faces making
eye contact and the preference of very young
infants for objects that respond to their own ac-
tions with high contingency. By 4 months, gaze
following is much more likely to occur if it is
preceded by direct eye contact (59). Between 12
and 18 months, infants can use eye contact and
pointing behavior as signals that distinguish
between communicative and noncommunica-
tive interactions (60,61). This phenomenon of
joint attention, in which both parties are aware
that they are jointly attending, seems to be dis-
tinctly human (62).
Conclusions and future directions
We have argued that explicit mind reading is
culturally inherited in the same way as print
reading. The neurocognitive mechanisms that
allow us to deliberate and talk about mental
states are constructed, or recycled (3), from mech-
anisms that evolved genetically to fulfill more
general functions (to parse and predict dynamic
sequences of events and to get information from
others), and the construction process depends
on tuition. Expert mind readers communicate
ment al-state concepts, and ways of representing
these concepts, to novices. As the present gener-
ation of novices becomes expert, it passes on the
knowledge and skill of mind reading to the next
cultural generation.
Most, possibly all, human neurocognitive skills
are shaped by culture, and many are culturally
inherited, but the parallels between mind read-
ing and print reading are extraordinary. For ex-
ample, linguistic communities vary in the ways
that they categorize colors, but color perception
is not culturally inherited in the same way as
print reading. Unlike print reading, color percep-
tion is rooted in highly specialized, genetically
inherited mechanisms that humans share with
other species. Although cultural input adjusts
these mechanisms, it does not make them into
a whole new neurocognitive system.
The analogy between print reading and mind
reading is not perfect. “Sound symbolism”shows
that the relations between inscriptions and their
corresponding speech sounds and referents of-
ten depend on structural features of the nervous
system (63), but it is likely that these relations
are more arbitrary than the relations between
observable behavior and mental states. In this
respect, numeracy may be a better analog than
literacy. However, explicit mind reading is com-
parable to print reading, not only in terms of
its weak dependence on specialized genetically
inherited mechanisms and strong dependence
on tuition, but also in the shape and size of the
cultural legacy. Like print reading, explicit mind-
reading mechanisms represent representational
relations—between mental states, behavior, and
events in the world—and allow the mind reader
to regulate and interpret a virtually limitless range
of mental contents. Consequently, the print read-
ing and mind reading legacies are also alike in
promoting the cultural inheritance of other, more
specific skills. Reading and writing enable us to
teach others almost anything we know through
the written word, and mind reading improves the
efficiency of all teaching by allowing the expert
to represent what the novice does and does not
already understand. Like print reading, mind read-
ing is a cultural gift that goes on giving, a cul-
turally inherited skill that facilitates the cultural
inheritance of an enormous range of other, more
specific skills (64).
We have suggested that in infancy, when the
enculturation process is only beginning, implicit
mind-reading mechanisms produce, under some
circumstances, accurate expectations about the
behavior of agents. A priority for future research
is to find out whether these implicit mechanisms
are specialized for mind reading, or whether im-
plicit mind reading is mediated by domain-general
mechanisms. In either case, it is clear that implicit
mind-reading mechanisms continue to operate
throughout the life cycle, enabling swift social
coordination of behavior when time is short and
other demands on the neurocognitive system are
heavy. It is possible that the outputs of these im-
plicit mechanisms also contribute to the devel-
opment of explicit mind reading by, for example,
segmenting the stream of observable behavior
into units that can subsequently be aligned with
mental categories. However, in our view, implicit
mind reading is radically insufficient for the de-
velopment of explicit mind reading. Our view
suggests that no amount of individual learning—
implicit mind reading, introspection, and watch-
ing the behavior of others—would be enoughfor
the development of explicit mind reading. If a
group of human infants managed to survive on
a desert island, in a cruel Lord of the Flies–like
experiment, they would be no more likely to de-
velop a theory of mind and become explicit mind
readers than to develop a writing system and be-
come literate print readers.
REFERENCES AND NOTES
1. D. Premack, G. Woodruff, Does the chimpanzee have a theory
of mind? Behav. Brain Sci. 1, 515–526 (1978). doi: 10.1017/
S0140525X00076512
2. R. Boyd, P. J. Richerson, J. Henrich, The cultural niche: Why
social learning is essential for human adaptation. Proc. Natl.
Acad. Sci. U.S.A. 108 (suppl. 2), 10918–10925 (2011).
doi: 10.1073/pnas.1100290108; pmid: 21690340
3. S. Dehaene, L. Cohen, Cultural recycling of cortical maps.
Neuron 56, 384–398 (2007). doi: 10.1016/j.
neuron.2007.10.004; pmid: 17964253
4. F. Van Overwalle, Social cognition and the brain: A meta-
analysis. Hum. Brain Mapp. 30, 829–858 (2009). doi: 10.1002/
hbm.20547; pmid: 18381770
5. S. Dehaene, L. Cohen, The unique role of the visual word form
area in reading. Trends Cogn. Sci. 15, 254–262 (2011).
doi: 10.1016/j.tics.2011.04.003; pmid: 21592844
6. U. Frith, Mind blindness and the brain in autism. Neuron 32,
969–979 (2001). doi: 10.1016/S0896-6273(01)00552-9
pmid: 11754830
7. S. Paracchini, T. Scerri, A. P. Monaco, The genetic lexicon of
dyslexia. Annu. Rev. Genomics Hum. Genet. 8,57–79 (2007).
doi: 10.1146/annurev.genom.8.080706.092312;pmid:17444811
8. S. Dehaene, Reading in the Brain: The New Science of How We
Read (Penguin Books, New York, 2009).
9. M. A. Changizi, Q. Zhang, H. Ye, S. Shimojo, The structures of
letters and symbols throughout human history are selected to
match those found in objects in natural scenes. Am. Nat. 167,
E117–E139 (2006). doi: 10.1086/502806; pmid: 16671005
10. A. Shahaeian, C. C. Peterson, V. Slaughter, H. M. Wellman,
Culture and the sequence of steps in theory of mind
development. Dev. Psychol. 47, 1239–1247 (2011).
doi: 10.1037/a0023899; pmid: 21639620
11. A. Lillard, Ethnopsychologies: Cultural variations in theories
of mind. Psychol. Bull. 123,3–32 (1998). doi: 10.1037/0033-
2909.123.1.3; pmid: 9461850
SCIENCE sciencemag.org 20 JUNE 2014 •VOL 344 ISSUE 6190 1243091-5
RESEARCH |REVIEW

12. W. Harris, Mental Disorders in the Classical World (Brill, Leiden,
Netherlands, and Boston, 2013).
13. C. Burrow, Epic Romance: Homer to Milton (Clarendon Press,
Oxford, 1993).
14. V. McGeer, “The regulative dimension of folk psychology.”in
Folk Psychology Re-Assessed, D. Hutto, M. M. Ratcliffe, Eds.
(Springer, Dordrecht, Netherlands, 2007), pp. 137–156.
15. D. Rigoni, S. Kühn, G. Sartori, M. Brass, Inducing disbelief in
free will alters brain correlates of preconscious motor
preparation: The brain minds whether we believe in free will
or not. Psychol. Sci. 22, 613–618 (2011). doi: 10.1177/
0956797611405680; pmid: 21515737
16. R. F. Baumeister, E. J. Masicampo, C. N. Dewall, Prosocial
benefits of feeling free: Disbelief in free will increases
aggression and reduces helpfulness. Pers. Soc. Psychol. Bull.
35, 260–268 (2009). doi: 10.1177/0146167208327217
pmid: 19141628
17. Á. M. Kovács, E. Téglás, A. D. Endress, The social sense:
Susceptibility to others’beliefs in human infants and adults.
Science 330, 1830–1834 (2010). doi: 10.1126/science.1190792;
pmid: 21205671
18. K. H. Onishi, R. Baillargeon, Do 15-month-old infants
understand false beliefs? Science 308, 255–258 (2005).
doi: 10.1126/science.1107621; pmid: 15821091
19. R. Baillargeon, R. M. Scott, Z. He, False-belief understanding in
infants. Trends Cogn. Sci. 14, 110–118 (2010). doi: 10.1016/j.
tics.2009.12.006; pmid: 20106714
20. I. A. Apperly, Mindreaders: The Cognitive Basis of “Theory of
Mind”(Psychology Press, Hove, UK, and New York, 2011).
21. J. Perner, “Who took the cog out of cognitive science?
Mentalism in an era of anti-cognitivism.”in Cognition and
Neuropsychology: International Perspectives on Psychological
Science, P. A. Frensch, R. Schwarzer, Eds. (Psychology Press,
London, 2010), vol. 1, chap. 15, pp. 241–261.
22. U. Frith, Why we need cognitive explanations of autism.
Q. J. Exp. Psychol. 65, 2073–2092 (2012). doi: 10.1080/
17470218.2012.697178; pmid: 22906000
23. C. Heyes, False belief in infancy: A fresh look. Dev. Sci. (2014).
doi: 10.1111/desc.12148; pmid: 24666559
24. C. Heyes, Submentalizing: I am not really reading your mind.
Perspect. Psychol. Sci. 9, 131–143 (2014). doi: 10.1177/
1745691613518076
25. A. Senju, V. Southgate, S. White, U. Frith, Mindblind eyes: An
absence of spontaneous theory of mind in Asperger syndrome.
Science 325, 883–885 (2009). doi: 10.1126/science.1176170;
pmid: 19608858
26. R. Bull, L. H. Phillips, C. A. Conway, The role of control
functions in mentalizing: Dual-task studies of theory of mind
and executive function. Cognition 107, 663–672 (2008). doi:
10.1016/j.cognition.2007.07.015; pmid: 17765214
27. A. W. Qureshi, I. A. Apperly, D. Samson, Executive function is
necessary for perspective selection, not Level-1 visual
perspective calculation: Evidence from a dual-task study of
adults. Cognition 117, 230–236 (2010). doi: 10.1016/j.
cognition.2010.08.003; pmid: 20817158
28. D. Samson, I. A. Apperly, J. J. Braithwaite, B. J. Andrews,
S. E. Bodley Scott, Seeing it their way: Evidence for rapid and
involuntary computation of what other people see. J. Exp.
Psychol. Hum. Percept. Perform. 36, 1255–1266 (2010).
doi: 10.1037/a0018729; pmid: 20731512
29. I. Santiesteban, C. Catmur, S. Coughlan Hopkins, G. Bird,
C. Heyes, Avatars and arrows: Implicit mentalizing or domain-
general processing? J. Exp. Psychol. Hum. Percept. Perform.
40, 929–937 (2014).pmid: 24377486
30. J. Tipples, Orienting to counterpredictive gaze and arrow cues.
Atten. Percept. Psychophys. 70,77–87 (2008). doi: 10.3758/
PP.70.1.77; pmid: 18306962
31. S. Brem et al., Evidence for developmental changes in the
visual word processing network beyond adolescence.
Neuroimage 29, 822–837 (2006). doi: 10.1016/j.
neuroimage.2005.09.023; pmid: 16257546
32. S.-J. Blakemore, The social brain in adolescence. Nat. Rev.
Neurosci. 9, 267–277 (2008). doi: 10.1038/nrn2353;
pmid: 18354399
33. I. Dumontheil, I. A. Apperly, S. J. Blakemore, Online usage
of theory of mind continues to develop in late adolescence.
Dev. Sci. 13, 331–338 (2010). doi: 10.1111/j.1467-
7687.2009.00888.x; pmid: 20136929
34. N. C. Vetter, K. Leipold, M. Kliegel, L. H. Phillips, M. Altgassen,
Ongoing development of social cognition in adolescence.
Child Neuropsychol. 1–15 (2012). pmid: 22934659
35. C. Hughes et al., Origins of individual differences in theory of
mind: From nature to nurture? Child Dev. 76, 356–370 (2005).
doi: 10.1111/j.1467-8624.2005.00850_a.x; pmid: 15784087
36. A. I. Goldman, Simulating Minds: The Philosophy, Psychology,
and Neuroscience of Mindreading (Oxford Univ. Press,
Oxford, 2006).
37. A. Gopnik, H. M. Wellman, Reconstructing constructivism:
Causal models, Bayesian learning mechanisms, and the theory
theory. Psychol. Bull. 138, 1085–1108 (2012). doi: 10.1037/
a0028044; pmid: 22582739
38. J. E. Pyers, A. Senghas, Language promotes false-belief
understanding: Evidence from learners of a new sign language.
Psychol. Sci. 20, 805–812 (2009). doi: 10.1111/j.1467-
9280.2009.02377.x; pmid: 19515119
39. A. Mayer, B. E. Träuble, Synchrony in the onset of mental state
understanding across cultures? A study among children in
Samoa. Int. J. Behav. Dev. 37,21–28 (2013). doi: 10.1177/
0165025412454030
40. M. Meristo, E. Hjelmquist, G. Morgan, “How access to language
affects theory of mind in deaf children.”in Access to Language
and Cognitive Development, M. Siegal, L. Surian, Eds.
(Oxford Univ. Press, Oxford, 2012), p. 44.
41. V. Slaughter, C. C. Peterson, “How conversational input
shapes theory of mind development in infancy and early
childhood.”in Access to Language and Cognitive Development,
M. Siegal, L. Surian, Eds. (Oxford Univ. Press, Oxford, 2012),
pp. 3–22.
42. K. Sterelny, The Evolved Apprentice (The MIT Press,
Cambridge, MA, 2013).
43. M. Taumoepeau, T. Ruffman, Stepping stones to others’minds:
Maternal talk relates to child mental state language and
emotion understanding at 15, 24, and 33 months. Child Dev.
79, 284–302 (2008). doi: 10.1111/j.1467-8624.2007.01126.x;
pmid: 18366424
44. M. Taumoepeau, T. Ruffman, Mother and infant talk about
mental states relates to desire language and emotion
understanding. Child Dev. 77, 465–481 (2006). doi: 10.1111/
j.1467-8624.2006.00882.x; pmid: 16611184
45. E. Meins, “Social relationships and children’s understanding of
mind: Attachment, internal states, and mind-mindedness.”
in Access to Language and Cognitive Development, M. Siegal,
L. Surian, Eds. (Oxford Univ. Press, Oxford, 2012), pp. 23–43.
46. H. Lohmann, M. Tomasello, The role of language in the
development of false belief understanding: A training study.
Child Dev. 74, 1130–1144 (2003). doi: 10.1111/1467-
8624.00597; pmid: 12938709
47. P. A. de Villiers, J. G. de Villiers, Deception dissociates from
false belief reasoning in deaf children: Implications for the
implicit versus explicit theory of mind distinction. Br. J. Dev.
Psychol. 30, 188–209 (2012). doi: 10.1111/j.2044-
835X.2011.02072.x; pmid: 22429041
48. K. O’Brien, V. Slaughter, C. C. Peterson, Sibling influences on
theory of mind development for children with ASD. J. Child
Psychol. Psychiatry 52, 713–719 (2011). doi: 10.1111/j.1469-
7610.2011.02389.x; pmid: 21418062
49. J. I. M. Carpendale, C. Lewis, Constructing an understanding of
mind: The development of children’s social understanding
within social interaction. Behav. Brain Sci. 27,79–96
(2004). doi: 10.1017/S0140525X04000032;
pmid: 15481944
50. A. Seeger, Nature and Society in Central Brazil: The Suyá
Indians of Mato Grosso (Harvard Univ. Press, Cambridge,
MA, 1981).
51. M. C. Frank, E. Vul, S. P. Johnson, Development of infants’
attention to faces during the first year. Cognition 110,
160–170 (2009). doi: 10.1016/j.cognition.2008.11.010;
pmid: 19114280
52. N. F. Troje, C. Westhoff, The inversion effect in biological
motion perception: Evidence for a “life detector”?Curr. Biol. 16,
821–824 (2006). doi: 10.1016/j.cub.2006.03.022
pmid: 16631591
53. T. Falck-Ytter, G. Gredebäck, C. von Hofsten, Infants predict
other people’s action goals. Nat. Neurosci. 9, 878–879 (2006).
doi: 10.1038/nn1729; pmid: 16783366
54. V. Gallese, A. Goldman, Mirror neurons and the simulation
theory of mind-reading. Trends Cogn. Sci. 2, 493–501 (1998).
doi: 10.1016/S1364-6613(98)01262-5; pmid: 21227300
55. P. W. Jusczyk, How infants begin to extract words from
speech. Trends Cogn. Sci. 3, 323–328 (1999). doi: 10.1016/
S1364-6613(99)01363-7; pmid: 10461194
56. S. M. Thurman, H. Lu, Physical and biological constraints
govern perceived animacy of scrambled human forms. Psychol.
Sci. 24, 1133–1141 (2013). doi: 10.1177/0956797612467212;
pmid: 23670885
57. M. Schlesinger, J. Langer, Infants’developing expectations
of possible and impossible tool‐use events between ages 8
and 12 months. Dev. Sci. 2, 195–205 (1999). doi: 10.1111/1467-
7687.00068
58. A. E. Skerry, S. E. Carey, E. S. Spelke, First-person action
experience reveals sensitivity to action efficiency in
prereaching infants. Proc. Natl. Acad. Sci. U.S.A. 110,
18728–18733 (2013). doi: 10.1073/pnas.1312322110;
pmid: 24167247
59. G. Csibra, Teleological and referential understanding of action
in infancy. Philos. Trans. R. Soc. London Ser. B 358, 447–458
(2003). doi: 10.1098/rstb.2002.1235; pmid: 12689372
60. D. A. Baldwin et al., Infants’reliance on a social criterion for
establishing word-object relations. Child Dev. 67, 3135–3153
(1996). doi: 10.2307/1131771; pmid: 9071774
61. T. Behne, M. Carpenter, M. Tomasello, One-year-olds
comprehend the communicative intentions behind gestures in
a hiding game. Dev. Sci. 8, 492–499 (2005). doi: 10.1111/
j.1467-7687.2005.00440.x; pmid: 16246240
62. M. Carpenter, J. Call, “How joint is the joint attention of apes
and human infants?”in Agency and Joint Attention, J. Metcalfe,
H. S. Terrace, Eds. (Oxford Univ. Press, New York, 2013),
pp. 49–61.
63. O. Ozturk, M. Krehm, A. Vouloumanos, Sound symbolism in
infancy: Evidence for sound-shape cross-modal
correspondences in 4-month-olds. J. Exp. Child Psychol. 114,
173–186 (2013). doi: 10.1016/j.jecp.2012.05.004
pmid: 22960203
64. C. Heyes, Grist and mills: On the cultural origins of cultural
learning. Philos. Trans. R. Soc. London Ser. B 367, 2181–2191
(2012). doi: 10.1098/rstb.2012.0120; pmid: 22734061
ACKNO WLED GMEN TS
We thank I. Apperly, J. Call, M. Eimer, U. Frith, A. Gopnik,
D. Sperber, and two anonymous referees for their comments on
the manuscript (especially the most challenging); V. McGeer,
J. Nagel, and N. Shea for valuable discussion in the early
stages; and I. Court, T. Court, and C. Catmur for the
photographs in Fig. 1.
10.1126/science.1243091
1243091-6 20 JUNE 2014 •VOL 344 ISSUE 6190 sciencemag.org SCIENCE
RESEARCH |REVIEW