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

net/publication/327989904

Theory of Mind

Chapter · October 2018

DOI: 10.1007/978-3-319-16999-6_2376-1

CITATIONS READS

7 19,979

2 authors, including:

Evan Westra
Purdue University
22 PUBLICATIONS 292 CITATIONS

SEE PROFILE

All content following this page was uploaded by Evan Westra on 10 October 2018.

The user has requested enhancement of the downloaded file.

T

Theory of Mind theory-of-mind abilities, which are common in

autism spectrum disorder (Tager-Flusberg 2007),
Evan Westra1 and Peter Carruthers2 typically correlate with broader difficulties in
1
Univeristy of Rochester, Rochester, NY, USA social understanding.
2
University of Maryland, College park, MD, USA Many scholars have suggested that theory of
mind is an innate adaptation for social cognition,
emerging very early in development and playing a
Synonyms crucial role in social learning and the acquisition
of language (Baron-Cohen 1997). However,
Folk psychology; Mentalizing; Mental-state attri- others have argued that theory of mind is the
bution; Mind reading product of largely domain-general learning pro-
cesses and is acquired gradually over the course
development through social experience (Wellman
Definition 2014). A third view argues that humans possess
two systems for theory of mind: an innate,
The capacity to predict and interpret behavior by domain-specific “implicit” system and a learned,
using representations of hidden, causally effica- domain-general “explicit” system (Apperly and
cious mental states. Butterfill 2009).

Introduction The False-Belief Task Controversy

“Theory of mind” consists in the ability to use The classic experimental paradigm for testing
concepts of intentional mental states, such as theory-of-mind abilities in children is the false-
beliefs, emotions, intentions, goals, and percep- belief task (Wimmer and Perner 1983). In the
tual states, in order to predict and interpret behav- most common version of this task, participants
ior. Functional magnetic resonance imaging observe an agent place an object in a box and
studies have revealed a distinctive network of then leave the room. While the first agent is
neural regions that is active during theory-of- gone, a second agent moves the object to another
mind tasks, including the temporal-parietal junc- hidden location. When the first agent returns, the
tion, the posterior superior temporal sulcus, the participant is asked to predict where the agent will
medial prefrontal cortex, the precuneus, and the look for the object (or in some versions of the task,
temporal poles (Van Overwalle 2009). Deficits in to say where she thinks the object is). The correct
# Springer International Publishing AG 2018
T.K. Shackelford, V.A. Weekes-Shackelford (eds.), Encyclopedia of Evolutionary Psychological Science,
https://doi.org/10.1007/978-3-319-16999-6_2376-1
2 Theory of Mind

answer is “in the box.” In order to pass the task, Two Views on the Evolution and
the participant must represent that the agent has a Development of Theory of Mind
belief about the location of the object that differs
from what is really the case and use that represen- In the background of the nativist approach is a
tation to accurately predict what she will do next. theoretical commitment concerning the impor-
Similar tasks test the capacity to reason about tance of theory of mind in the evolution of
other mental states, including desires, states of human social intelligence. On this view, innate
knowledge, perceptual access, and emotions adaptations for theory of mind are thought to
(Wellman and Liu 2004). have emerged early in the hominid line. In the
Typically, children fail the false-belief task highly social environments of ancestral hominids,
until after their fourth birthday, as do many adults accurate prediction and interpretation of the
with autism spectrum disorder (Wellman et al. behavior of conspecifics would have been crucial
2001). Since these results were first discovered, for survival. Accounts of the earliest emergence of
there has been controversy over their interpreta- theory-of-mind abilities in primates emphasize
tion. “Constructivists” have argued that children’s the adaptive importance of theory of mind in
shifting performance on this task reflects the social competition, deception, and manipulation,
emergence of a new concept of belief, amounting as individuals sought to gain mating opportunities
to a fundamental change in children’s intuitive and to improve their status within their group’s
theory of the social world, analogous to theoreti- social hierarchy (Byrne and Whiten 1988). This
cal changes brought on by scientific discovery “Machiavellian intelligence” hypothesis is
(Gopnik and Wellman 1992). Children, in other supported by evidence that modern great apes
words, are learning what beliefs are and the cir- seem to demonstrate at least simple forms of
cumstances in which people have them. “Nativ- theory-of-mind abilities in competitive but not
ists,” in contrast, have pointed to the selective cooperative experimental contexts (Call and
deficits on the false-belief task displayed by peo- Tomasello 2008).
ple with autism spectrum disorder as evidence of Accounts of the evolution of more highly
an underlying theory-of-mind module, which is developed, and perhaps specifically human,
selectively impaired in autism (Baron-Cohen theory-of-mind abilities, in contrast (including
1997). Nativists also argue that younger children’s capacities to reason about the false beliefs of
difficulties on the false-belief task don’t reflect the others), tend to place greater emphasis on the
absence of a concept of belief but rather a perfor- cooperative functions of theory of mind, particu-
mance error due the immaturity of their domain- larly when coordinating multiple agents in the
general executive capacities. According to this pursuit of mutually shared goals, such as group
view, the concept of belief is innate, but children hunting and foraging (Tomasello 2014). Infer-
can fail to deploy it successfully in certain exper- ences about beliefs and intentions are also thought
imental contexts. Importantly, the nativist view to have played a crucial role in the emergence of
doesn’t have to claim that theory of mind is unaf- early systems of gestural communication, such as
fected by experience and individual learning. pointing and pantomime, which require infer-
Rather, the claim can be that these changes are ences about mental states on the part of both the
both constrained and facilitated by a domain- communicator and the audience (Scott-Phillips
specific, innately channeled, learning mechanism 2014). These cooperative environments are
that comes equipped with a few basic mental-state thought to have created an adaptive feedback
concepts (Carruthers 2015). loop, where selection pressures for more complex
theory-of-mind abilities led to more complex
forms of cooperation, which created further selec-
tion pressures on our theory of mind. Thus, theory
of mind is thought to have fueled the evolution of
highly complex forms of mutualistic cooperation,
Theory of Mind 3

in addition to supporting the development of these with the nativist’s performance-error account of
same social abilities in ontogeny. children’s performance on the false-belief task,
Constructivists typically accept that human suggesting instead that children’s social and lin-
beings possess some special adaptations for social guistic environment plays an important role in
intelligence but generally deny that these amount determining their theory-of-mind abilities.
to a genuine theory of mind. For example, some In response to these findings, some nativists
constructivists acknowledge that neonates are have argued that they reflect the fact that children
innately biased to attend to faces and eyes and must learn how to apply their innate theory-of-
are innately disposed to engage in imitative mind abilities in different social and linguistic
behavior (e.g., Meltzoff 2007). According to con- contexts and that variations in social environment
structivist accounts, these low-level, noncognitive impact this learning process (Westra 2017). In
mechanisms serve as a scaffold for children’s effect, the suggestion is that younger children
early theory-of-mind development by directing fail at the tasks because their grip on discourse
their attention toward socially relevant phenom- pragmatics is weak, leading them to misunder-
ena. The latter then serve as inputs for domain- stand the point of the questions they are asked,
general learning procedures, such as statistical and it is this that is impacted by social experience.
learning (Ruffman et al. 2012). Many constructiv- In support of the nativist view, a large body of
ists also believe that children’s acquisition of a evidence has emerged more recently, demonstrat-
mature theory of mind depends on exposure to ing a range of theory-of-mind abilities in children
specific linguistic inputs, such as clausal comple- in the first 2 years of life, well before they pass
mentation syntax or mental-state vocabulary (e.g., verbal forms of the false-belief task. For example,
de Villiers and Pyers 2002). Thus, while nativists a number of studies have adapted the classic false-
posit that the ancestral emergence of domain- belief task to make it suitable for infants. Impor-
specific cognitive adaptations for theory of mind tantly, these tasks never ask children to make
made complex forms of cooperation and linguistic explicit, verbal predictions. Instead, they rely on
communication possible, constructivists hold that children’s spontaneous behaviors to measure their
the existence of complex cooperative environ- theory-of-mind abilities. For example, one such
ments lay the developmental foundation for chil- task presented 15-month-olds with a scene in
dren to acquire a theory of mind via individual which an experimenter hid an object in one of
learning, which implies that it is an evolutionarily two boxes and then left. While the experimenter
recent, culturally dependent phenomenon (Heyes was absent, the object was moved to the other box.
and Frith 2014). When the experimenter returned, infants either
saw her reach toward the first, empty box, or the
second box, where the object then was. Results
Does Theory of Mind Emerge Early or show that infants look far longer when the exper-
Late in Development? imenter reaches for the second box, suggesting
that they find this behavior surprising (Onishi
Support for the constructivist view comes from and Baillargeon 2005). The infants were seem-
evidence that theory-of-mind development is ingly expecting the agent to reach into the box
influenced by both social experience and lan- where she believed the goal object to be.
guage. For instance, having older siblings tends Researchers have also designed ways of testing
to lead to earlier success on the false-belief task, as early false-belief competence by exploiting the
do greater amounts of parental mental-state dis- fact that young children are highly motivated to
course (Ruffman et al. 2012). Strikingly, deaf help other people achieve their goals (Buttelmann
children who do not learn sign language until et al. 2009). In these tasks, 18-month-olds observe
later in life also have persistent difficulties on the an experimenter place a toy in one of two boxes
false-belief task, even in adulthood (Pyers and and close the lid. The experimenter then leaves the
Senghas 2009). These findings seem incompatible room, and the children see a second experimenter
4 Theory of Mind

move the toy from the first box to the second box. striking results. In one study, the experimenters
Then the first experimenter returns and begins to showed chimpanzees, bonobos, and orangutans
struggle with the lid of the first box. The authors videos depicting an interaction between an actor
predicted that if children understood that the dressed as a zookeeper and an actor dressed as a
experimenter desired the toy but had a false belief gorilla. The videos began by showing the gorilla
about its location, they should respond by retriev- hitting the zookeeper and then running into one of
ing the toy from the second box. In a true-belief two haystacks arrayed at either side of the screen.
control task in which the first experimenter While the zookeeper turned away to fetch a stick
observes the location change, in contrast, children (which, in familiarizations, was used to hit the
should instead help the experimenter open the first haystack containing the gorilla), the gorilla
box, presuming that she must want something moved from one haystack to the other. Then the
inside it. Indeed, this is what they found. zookeeper turned around with the stick raised,
The interpretation of these results, like those of poised to attack. The researchers used an eye
the original false-belief task, has been a subject of tracker to measure whether the apes would look
great controversy. While nativists have used in anticipation toward the actual location of the
infant false-belief paradigms to support the claim gorilla or to the location where the zookeeper will
that the capacity for representing beliefs is innate think the gorilla is hiding (i.e., anticipating and
(Baillargeon et al. 2010), critics of these para- reasoning from the zookeeper’s false belief). The
digms have offered various alternative, low-level results showed that all three species of ape looked
explanations of the same findings. According to reliably more toward the false-belief location,
some of these interpretations, infants’ successful suggesting that they were indeed tracking the
performance on these tasks may not reflect an zookeeper’s beliefs (Krupenye et al. 2016). In
abstract concept of belief but rather the tracking another study, researchers adapted the active help-
of statistical regularities in observable behavior ing design of Buttelmann et al. (2009) (see above)
(Ruffman et al. 2012). Nativists have responded for use with chimpanzees, bonobos, and orangu-
to these criticisms by pointing out that these alter- tans. They, too, found that great apes seemed to
native explanations have tended to be post hoc use information about false beliefs in order to help
and have failed to issue in new data (Scott 2014). an experimenter retrieve an object from a locked
The nativist framework, in contrast, has generated box (Buttelmann et al. 2017).
a continuous stream of new results, employing an Some critics have argued that these results, like
ever-widening set of experimental paradigms. the infant false-belief-task results, should not be
However, these debates are currently ongoing. interpreted in rich, mentalistic terms; instead, we
should prefer explanations that only attribute to
apes the ability to make predictions about behav-
Can Apes Pass False-Belief Tasks? ior based on low-level, observable regularities,
such as the appearance and disappearance of the
We noted in the section “Two Views on the Evo- colored shirt of the zookeeper (Heyes 2017). To
lution and Development of Theory of Mind” that test this hypothesis, Krupeneye and colleagues
there is evidence that other great apes have at least designed a control task that matched their original
a simple form of theory of mind, one that allows anticipatory looking paradigm but replaced the
them to track and reason about the goals and states actors with inanimate colored shapes (Krupenye
of knowledge or ignorance of other agents. But et al. 2017). They found that, in contrast to the
until recently, all tests for false-belief understand- task involving actors, the participants were no
ing in other primates had proven negative (Call more likely to look toward either target location
and Tomasello 2008). Recently, however, when observing the same interaction between
researchers have successfully adapted for use inanimate shapes. This provides compelling evi-
among primates some of the methods for testing dence that the apes’ predictive gaze behavior was
false-belief competence in young children, with specifically sensitive to the social nature of the
Theory of Mind 5

stimuli, as opposed to its low-level properties. All automatic, and largely encapsulated from execu-
of these results are, however, quite recent, and tive systems; it also persists unchanged into adult-
further research is necessary before conclusions hood and is shared with our nearest primate
can confidently be drawn. relatives. Thus, in many respects this implicit
Thus, two recent studies seem to show that mindreading system resembles the kind of
three other species of great ape are capable of domain-specific adaptation posited by nativists.
passing false-belief tasks. This result, if it is However, two-systems theorists hold that due to
valid, has striking implications for the debate its automatic and encapsulated architecture, the
about the language dependence of theory of implicit mindreading system is subject to “signa-
mind. If nonlinguistic primates are able to repre- ture limits.” In particular, while it can represent
sent beliefs, this casts serious doubt on the claim beliefs about the locations of objects (e.g., “Bill
that such an ability requires linguistic experience, believes that the apple is in the box”), it cannot
or is uniquely human. These findings also suggest handle beliefs about the identity of objects (e.g.,
that belief representation may in fact be evolution- “Bill believes that the apple is really a pear”). This
arily quite ancient and potentially shared by the is because the implicit system is thought to track
last common ancestor of humans, chimpanzees, beliefs and other mental states without
bonobos, and orangutans. representing them as such and, in particular, with-
Such inferences should be drawn cautiously, out representing the aspectual nature of belief
however: the presence of robust theory-of-mind states. (Famously, one can believe that Jocasta is
ability in modern great apes does not necessarily beautiful without believing that one’s mother is
imply that it is an innate adaptation. For it is beautiful, even though Jocasta is in fact one’s
possible that these apes acquired their abilities mother. Here one and the same person is thought
through individual learning (and hence that ances- about under two different aspects.) The implicit
tral apes might have done so as well). In other system cannot, therefore, fully capture the rich-
words, even if great apes do reason about false ness and flexibility of the mature concept of belief.
beliefs, this doesn’t necessarily provide support In humans, one of the functions of the implicit
for nativism about theory of mind in human system is to scaffold the acquisition of a second,
beings. What it does do, however, is undermine parallel, explicit theory-of-mind system that
the claim that false-belief understanding is specif- develops much more gradually. This system is
ically dependent on either linguistic experience or said to be slow, effortful, and heavily reliant
experience of distinctively human forms of col- upon executive resources, such as working mem-
laboration and joint action. ory. It develops gradually via domain-general
learning in response to social and linguistic input
and only emerges after children’s fourth birthday
The two-Systems View (thus enabling them to pass the explicit false-
belief task). Unlike the implicit system, this one
Another set of theorists has attempted to resolve isn’t subject to signature representational limits.
the dispute between nativists and constructivists However, because it relies heavily on working
by proposing an ecumenical solution that incor- memory, it must be directed by top-down goals
porates both constructivist and nativist elements and is compromised under cognitive load. The
(Apperly and Butterfill 2009). According to these explicit mindreading system closely resembles
“two-systems” accounts, infants’ early compe- the conception of mindreading posited by con-
tence on “implicit” false-belief tasks (and by par- structivists, albeit supported by a relatively more
ity of reasoning, the recent successful complex set of domain-specific cognitive adapta-
performance of other great ape species on similar tions. Thus, by adulthood, human beings are said
tasks) does reflect a domain-specific adaptation to possess two parallel theory-of-mind systems,
for tracking mental states. This “implicit” theory- each with a distinct information-processing
of-mind system is said to be fast, effortless, profile.
6 Theory of Mind

To test the prediction that the implicit unclear whether this evidence truly supports a
mindreading system is subject to signature limits, two-systems view. Nevertheless, two-systems
two-systems theorists have compared adults’ and accounts of theory of mind have significantly
young children’s implicit theory-of-mind predic- influenced debates about the cognitive architec-
tions in scenarios where agents have false beliefs ture of our social cognition abilities.
about either an object’s location or its identity
(Low and Watts 2013). To test participants’ ability
to track false beliefs about identity, the experi-
Conclusion
menters constructed a scenario in which partici-
pants were familiarized with an agent who
As can be observed in the debate about the false-
demonstrated a consistent preference for a certain
belief task, there is considerable disagreement in
color, always reaching for blue items rather than
the literature about the role of theory-of-mind
red items, for example. Next, out of sight of the
abilities in socio-cognitive development: while
agent, participants were familiarized with a paper
nativists believe that theory-of-mind is a basic
cutout figure that appeared as a red robot from one
adaptation that plays an important role in early
side and as a blue robot on the other. They then
social learning, constructivists hold that theory of
watched as the agent observed what she would
mind is itself the product of social learning. These
have seen as a blue robot enters one of two boxes.
views correspond to distinct narratives about the
Next, unbeknownst to the agent, the figure rotated
emergence of theory of mind in the hominid line:
180 and moved to the second box while pre-
if the nativist view is right, then theory of mind is
senting to the agent as a red robot. Because the
evolutionarily ancient; if the constructivist view is
agent didn’t know that the red and blue robots are
correct, then it is likely to be a far more recent
the same individual, she should believe that there
phenomenon. The two-systems view attempts to
was still a blue robot in the first box. Thus, if
carve a middle ground between these two camps,
participants were tracking beliefs about identity,
allowing that some aspects of theory of mind are
they should expect the agent to reach toward the
indeed evolutionarily ancient and early develop-
first box. However, neither adults nor children
ing, while others are evolutionarily novel and
reliably looked in anticipation toward the first
acquired via social learning. However, the evi-
box, suggesting that they were insensitive to the
dence in favor of this view is itself a source of
agent’s beliefs about object identity. Meanwhile,
controversy.
both groups showed correct anticipatory looking
on the control task that only involved beliefs
about the object’s location.
These results provide support the claim that Cross-References
while implicit theory of mind accurately tracks
beliefs about the locations of objects, it doesn’t ▶ Autism
track beliefs about the identities of objects. How- ▶ Communication and social cognition
ever, critics of this study and others like it have ▶ Cultural Intelligence Hypothesis
pointed out that the object-identity task, which ▶ Does the chimpanzee have a theory of mind?
involves effortful forms of mental rotation, is ▶ Intentional stance, the
likely to place additional demands on executive ▶ Michael Tomasello
function that are not present in the object-location ▶ Nativism
control task. This makes it unclear whether the ▶ Simon Baron-Cohen
relevant failure is due to signature limitations on ▶ Social intelligence hypothesis
implicit theory of mind or rather limitations on ▶ Theory of mind and nonhuman intelligence
working memory (Carruthers 2015). It is therefore ▶ Theoy of mind and evidence of brain
modularity
Theory of Mind 7

References Low, J., & Watts, J. (2013). Attributing false-beliefs about

object identity is a signature blindspot in humans’
Apperly, I., & Butterfill, S. A. (2009). Do humans have two efficient mindreading system. Psychological Science,
systems to track beliefs and belief-like states? Psycho- 24(3), 305–311.
logical Review, 116(4), 953–970. Meltzoff, A. N. (2007). “Like me”: A foundation for social
Baillargeon, R., Scott, R. M., & He, Z. (2010). False-belief cognition. Developmental Science, 10(1), 126–134.
understanding in infants. Trends in Cognitive Sciences, Onishi, K. H., & Baillargeon, R. (2005). Do 15-month-old
14(3), 110–118. infants understand false beliefs? Science, 308(5719),
Baron-Cohen, S. (1997). Mindblindness: An essay on 255–258.
autism and theory of mind. Cambridge, MA: MIT Pyers, J. E., & Senghas, A. (2009). Language promotes
Press. false-belief understanding: Evidence from learners of a
Buttelmann, D., Carpenter, M., & Tomasello, M. (2009). new sign language. Psychological Science, 20(7),
Eighteen-month-old infants show false belief under- 805–812.
standing in an active helping paradigm. Cognition, Ruffman, T., Taumoepeau, M., & Perkins, C. (2012). Sta-
112(2), 337–342. tistical learning as a basis for social understanding in
Buttelmann, D., Buttelmann, F., Carpenter, M., Call, J., & children. The British Journal of Developmental Psy-
Tomasello, M. (2017). Great apes distinguish true from chology, 30(Pt 1), 87–104.
false beliefs in an interactive helping task. PLoS One, Scott, R. M. (2014). Post hoc versus predictive accounts of
12(4), e0173793. children’s theory of mind: A reply to Ruffman. Devel-
Byrne, R. W., & Whiten, A. (1988). Machiavellian intelli- opmental Review, 34(3), 300–304.
gence: Social expertise and the evolution of intellect in Scott-Phillips, T. (2014). Speaking our minds: Why human
monkeys, apes, and humans. Oxford: Clarendon Press. communication is different, and how language evolved
Call, J., & Tomasello, M. (2008). Does the chimpanzee to make it special (Vol. 3). Basingstoke: Palgrave
have a theory of mind? 30 years later. Trends in Cog- Macmillan.
nitive Sciences, 12(5), 187–192. Tager-Flusberg, H. (2007). Evaluating the theory-of-mind
Carruthers, P. (2015). Two systems for mindreading? hypothesis of autism. Current Directions in Psycholog-
Review of Philosophy and Psychology, 6, 2. ical Science, 16(6), 311–315.
de Villiers, J. G., & Pyers, J. E. (2002). Complements to Tomasello, M. (2014). A natural history of human think-
cognition: A longitudinal study of the relationship ing. Cambridge, MA: Harvard University Press.
between complex syntax and false-belief- Van Overwalle, F. (2009). Social cognition and the brain:
understanding. Cognitive Development, 17, A meta-analysis. Human Brain Mapping, 30(3),
1037–1060. 829–858.
Gopnik, A., & Wellman, H. M. (1992). Why the child’s Wellman, H. M. (2014). Making minds: How theory of
theory of mind really is a theory. Mind & Language, 7 mind develops. Oxford: Oxford University Press.
(1–2), 145–171. Wellman, H. M., & Liu, D. (2004). Scaling of theory-of-
Heyes, C. (2017). Apes submentalise. Trends in Cognitive mind tasks. Child Development, 75(2), 523–541.
Sciences, 21(1), 1–2. Wellman, H. M., Cross, D., & Watson, J. (2001). Meta-
Heyes, C., & Frith, C. D. (2014). The cultural evolution of analysis of theory-of-mind development: The truth
mind reading. Science, 344(6190), 1243091–1243091. about false belief. Child Development, 72(3), 655–684.
Krupenye, C., Kano, F., Hirata, S., Call, J., & Tomasello, Westra, E. (2017). Pragmatic development and the false
M. (2016). Great apes anticipate that other individuals belief task. Review of Philosophy and Psychology, 8(2),
will act according to false beliefs. Science, 354(6308), 235.
110. Wimmer, H., & Perner, J. (1983). Beliefs about beliefs:
Krupenye, C., Kano, F., Hirata, S., Call, J., & Tomasello, Representation and constraining function of wrong
M. (2017). A test of the submentalizing hypothesis: beliefs in young children’s understanding of deception.
Apes’ performance in a false belief task inanimate Cognition, 13(1), 103–128.
control. Communicative & Integrative Biology, 10,
e1343771.

View publication stats