Synthese
DOI 10.1007/s11229-012-0127-6
The developmental paradox of false belief
understanding: a dual-system solution
L. C. De Bruin · A. Newen
Received: 5 February 2012 / Accepted: 10 May 2012
© Springer Science+Business Media B.V. 2012
Abstract We explore the developmental paradox of false belief understanding. This
paradox follows from the claim that young infants already have an understanding of
false belief, despite the fact that they consistently fail the elicited-response false belief
task. First, we argue that recent proposals to solve this paradox are unsatisfactory
because they (i) try to give a full explanation of false belief understanding in terms
of a single system, (ii) fail to provide psychological concepts that are sufficiently
fine-grained to capture the cognitive requirements for the various manifestations of
false belief understanding, and (iii) ignore questions about system interaction. Second, we present a dual-system solution to the developmental paradox of false belief
understanding that combines a layered model of perspective taking with an inhibition-selection-representation mechanism that operates on different levels. We discuss
recent experimental findings that shed light on the interaction between these two
systems, and suggest a number of directions for future research.
Keywords Theory of mind · False belief understanding in infancy ·
Shared representations · Perspective taking
1 Introduction
In our everyday social interactions we frequently attribute mental states (e.g., beliefs,
desires and intentions) to others in order to predict or explain their behavior. For
example, if John knows that Mary wants the chocolate, and he also knows that Mary
L. C. De Bruin (B) · A. Newen
Institute for Philosophy II GA3/153, Ruhr-University Bochum, Universitätsstr. 150,
44801 Bochum, Germany
e-mail: [email protected]
123
Synthese
believes the chocolate is in the blue cupboard, he can predict that she will search for
the chocolate in the blue cupboard. This works equally well if John wants to explain
Mary’s behavior: Mary searches in the blue cupboard, because she wants the chocolate
and believes the chocolate is in the blue cupboard. It is generally accepted amongst
developmental psychologists that this capacity is enabled by a Theory of Mind (e.g.,
Perner 1991; Baron-Cohen 1995; Leslie 2000).
Over the last couple of decades, most research in Theory of Mind has focused
on the development of false belief understanding. In order to test whether John truly
understands Mary’s behavior from her point of view, we need to make sure that he
does not simply attribute to Mary his own belief about the location of the chocolate.
The False Belief Test (FBT) has been specifically designed to solve this problem by
introducing a condition in which the protagonist has a false belief about some state of
affairs in the world.
In the ‘elicited-response’ version of the false belief test, children are asked to give
a verbal prediction of the protagonist’s behavior based on his or her false belief. In
the ‘unexpected location’ elicited response FBT (Wimmer and Perner 1983; BaronCohen et al. 1985), for example, children observe a protagonist who sees an object
being placed in a certain location. Then the protagonist leaves, and the object is moved
to another location. When the protagonist returns, he mistakenly believes the object
is still in its initial location. At this point, the children are asked to predict where the
protagonist will look for the object. Test results show that 3-year-olds typically give a
wrong answer to this question, while 4-year-olds answer correctly. Findings on other
elicited-response FBTs, such as the ‘unexpected identity’ task (Perner et al. 1987);
Moses and Flavell 1990; Wellman 1990) confirm this picture. Many researchers have
therefore concluded that verbal or ‘explicit’ false belief understanding does not emerge
until the age of four (Wellman 2002; Wellman et al. 2001).
Recently, however, this conclusion has been challenged on the basis of ‘spontaneous-response’ FTBs, which no longer require an explicit answer to a question about the
protagonist’s belief. Instead, children’s understanding of false belief is inferred from
the behavior they spontaneously produce. Employing ‘violation of expectation’ and
‘anticipatory looking’ paradigms, these experiments indicate that false belief understanding emerges at a considerably earlier age, in 25-month-olds (Southgate et al.
2007), 15-month-olds (Onishi and Baillargeon 2005), 13-month-olds (Surian et al.
2007), and even 7-month-olds (Kovács et al. 2010). These findings give rise to a very
interesting and widely debated ‘developmental paradox’: if young infants already
understand false belief, as the spontaneous-response FBT suggests, then why do they
fail the elicited-response FBT?
This article proposes a model that allows us to account for the spontaneous-response
and elicited-response FBT results and shed new light on the developmental paradox.
The article is based on previous work on false belief understanding (De Bruin and
Newen 2012), and attempts to refine our earlier ideas and update them in the light of
new findings.
Section 2 starts with a discussion of a number of important spontaneous-response
FBTs. In Sect. 3, we evaluate recent solutions to the developmental paradox and argue
that they fail because they (i) try to give a full explanation of false belief understanding in terms of a single system, (ii) fail to provide psychological concepts that are
123
Synthese
sufficiently fine-grained to capture the cognitive requirements for implicit and explicit
false belief understanding, or (iii) ignore questions about system interaction. In Sect. 4,
we present a dual-system solution to the developmental paradox of false belief understanding—one that combines an Association Module, which allows for different levels
of perspective taking, with an Operating System, which is essentially an inhibitionselection-representation mechanism. We discuss the interaction between these systems
in the light of recent experimental findings (in particular, studies by Fabricius et al.
(2010) and Kovács et al. (2010)). Section 5 closes with some final comments and
questions for future research.
2 False belief understanding in early infancy
It is well known that the elicited-response FBT places rather strong demands on children’s cognitive capacities (Bloom and German 2000; Carlson and Moses 2001).
Spontaneous-response FBTs have been designed in such a way as to reduce these
demands in order to see whether children are capable of false belief understanding at
an earlier age.
Onishi and Baillargeon (2005), for example, used a ‘violation-of-expectation’ spontaneous-response FBT to investigate whether children would look reliably longer when
another agent acts in a way that is inconsistent with her false belief. First, 15-month-old
infants were familiarized with a protagonist hiding a toy in location A. The protagonist left, and the toy was moved to a different location B without her knowledge.
Then the infants were shown scenes of the protagonist searching for the hidden toy,
either where she falsely believed it to be (location A), or where it was actually located
(location B). Onishi and Baillargeon found that infants looked significantly longer at
those scenes in which the protagonist searched at the correct location B despite her
false belief about where the toy was hidden (location A).
These findings stand in tension with the results of an earlier experiment by Clements
and Perner (1994). They asked children to watch how the protagonist stored an object in location A. While the protagonist was asleep, the object was moved to a
different location B. The experimenters monitored where the children were looking in anticipation of the protagonist’s reappearance, given his false belief about the
object’s location. Clements and Perner contrasted this spontaneous behavior with the
explicit answers children gave to the question where the protagonist would look for
the object. They found that 3-year-olds looked at the correct location (A) when anticipating the protagonist’s return, even when they explicitly made the incorrect claim
that he would go to location B. Clements and Perner labeled this early manifestation of false belief understanding ‘implicit’, because the children were not explicitly
aware of the knowledge conveyed in their correct eye gaze. Importantly, they found
no sign of implicit false belief understanding in 2-year-old children and therefore concluded that children under 3 years of age do not understand false belief (cf. Perner and
Clements 2000; Clements and Perner 2001; Ruffman et al. 2001; Garnham and Perner
2001).
Southgate et al. (2007) have argued that Clements and Perner arrived at this conclusion because their study still included a verbal element: in order to maximize
123
Synthese
the frequency of anticipatory looking at location A, the experimenter said aloud,
‘I wonder where he [i.e., the protagonist] is going to look?’ before asking the question. Southgate et al. (2007) claimed that this primed 2-year-old infants to look at the
incorrect location, and decided to remove the verbal element from the design. In their
own study, they used an eye-tracker to measure anticipatory looking in 25-month-olds.
The infants observed an experimenter who witnessed how a protagonist hid an object in
location A. Then the experimenter became distracted and turned away from the scene.
Meanwhile, the protagonist transferred the object to a different location B. Southgate
et al. (2007) found that most 25-month-olds correctly anticipated the experimenter’s
behavior and looked at the location where she falsely believed the object to be hidden
(location A).
Other spontaneous-response FBTs have found that infants do not only understand
false beliefs about object location, but also about object identity. In a study by Song and
Baillargeon (2008), for example, 14.5-month-old infants witnessed an experimenter’s
hand placing a doll with blue hair and a stuffed skunk with a pink bow on placemats
or in shallow containers during familiarization trials. Afterwards, an agent showed
her preference for either of the two toys by reaching for it. Then, without the agent
watching, the toys were put in two boxes, one of which had a tuft of blue hair on the lid,
suggesting it contained the doll. The doll was placed in the plain box, and the skunk
in the box with the tuft of blue hair on it. When the agent returned and reached for the
box with the tuft of hair after having showed a preference for the skunk, the infants
looked considerably longer than when she reached for the plain box - and conversely.
According to Song and Baillargeon (2008), this meant that the infants expected the
agent to ‘falsely conclude’ that the doll was hidden in the hair box.
In another experiment by Scott and Baillargeon (2009), 18-month-old infants were
familiarized with two toy penguins, one in one piece, the other in two (hollow) pieces
that had to be assembled. The assembled two-piece penguin was indistinguishable
from the one-piece penguin. Each familiarization trial started with a disassembled
two-piece penguin and a complete one-piece penguin. The agent put a key in the twopiece penguin, after which an experimenter’s hand assembled the two-piece penguin
and placed both penguins on platforms or in shallow containers. In the test trial, the
two-piece penguin was assembled by the experimenter and placed under a transparent cover. The one-piece penguin was placed under an opaque cover. None of this
was witnessed by the agent. When she returned with her key—which she apparently
was going to hide in the two-piece penguin—she reached for either the transparent
cover or the opaque cover. When she reached for the transparent cover, infants looked
longer. According to Scott and Baillargeon, this suggested that they expected her: (i) to
falsely believe that the penguin under the transparent cover was the one-piece penguin
(because the two-piece penguin was always disassembled at the start of the familiarization trials), (ii) to falsely believe that the disassembled two-piece penguin was under
the opaque cover (because both penguins were always present in the familiarization
trials), and hence (iii) to reach for the opaque cover.
Table 1 presents an overview of spontaneous-response FBTs (see also Baillargeon
et al. (2010) and Poulin-Dubois et al. (2009)). See Wellman (2002) for a review, and
Wellman et al. (2001) for a meta-analysis of elicited-response FBTs.
123
Synthese
Table 1 An overview of studies on false belief understanding in early infancy
Study
Age
Non-verbal spontaneous response FBTs
Violation of expectation
Object location
Kovács et al. (2010)
7 months
Surian et al. (2007)
13 months
Onishi and Baillargeon (2005)
15 months
Träuble et al. (2010)
15 months
Object identity
Scott and Baillargeon (2009)
18 months
Object appearance
Song and Baillargeon (2008)
14.5 months
Object properties
Scott et al. (2010)
18 months
Southgate et al. (2007)
25 months
Song et al. (2008)
18 months
Scott et al. (2012)
2, 5 years
He et al. (2012)
2, 5 years
Scott et al. (2012)
2, 5 years
Clements and Perner (1994)
3 years
Southgate et al. (2010)
17 months
Buttelmann et al. (2009)
2, 5 years 18 months
Anticipatory looking
Object location
Verbal spontaneous response FBTs
Violation of expectation
Object location
Anticipatory looking
Object location
Active helping
Object location
3 Problematic solutions to the developmental paradox of false belief
understanding
The results of the spontaneous-response FBTs give rise to an important question:
if young infants already understand false belief, then why do they fail the elicitedresponse false belief task? Over the last decade, theorists have proposed various solutions to this developmental paradox. In what follows, we will discuss and evaluate the
main strengths and weaknesses of these solutions.
3.1 Single-system approaches
Baillargeon et al. (2010) have recently put forward two subsystems to account for the
developmental paradox of false belief understanding. Sub-system 1 enables infants to
attribute both motivational and ‘reality-congruent’ informational states to other agents.
This system is hypothesized to be operational by the end of the first year. Baillargeon
et al. (2010) define motivational states in terms of goals and dispositions that specify
the agent’s motivation in a given scene. Reality-congruent informational states are
defined as states that specify what knowledge or accurate information another agent
123
Synthese
possesses about the scene. Sub-system 2 enables infants to attribute ‘reality-incongruent’ informational states. These states specify the inaccurate information another agent
possesses about the scene (i.e. her false beliefs). Sub-system 2 becomes operational
in the second year of life.
According to Baillargeon et al., the main difference between the task requirements
of the spontaneous-response FBT vis-à-vis the elicited-response FBT is that, whereas
the former only involves (i) a process of false belief representation, the latter also
involves (ii) a response selection process (when asked the test question, children must
access their representation of the agent’s false belief to select a response) and (iii)
a response-inhibition process (when selecting a response, children must inhibit any
prepotent tendency to answer the test question based on their own knowledge).
Baillargeon et al. assume that infants are innately predisposed to attribute their own
(true) beliefs to other agents.1 However, this implies that successful performance on
the spontaneous-response FBT already requires infants to inhibit their own true belief
and select a response that is based on the other agent’s false belief. And this, in turn,
makes it very unlikely that infants fail the elicited-response FBT because they have to
jointly activate inhibition, selection and representation processes.
Despite the fact that Baillargeon et al. postulate two sub-systems, only one of them
(sub-system 2) is used to explain false belief understanding. But this does not seem
to be sufficient to account for the differences between the spontaneous-response and
elicited-response FBT findings. To see why Baillargeon et al.’s solution is problematic, it pays to consider an earlier account of false belief understanding proposed by
Leslie (1987, 1994, 2000, 2005).
Leslie initially suggests that successful performance on the elicited-response FBT
depends on two systems: a Theory of Mind Mechanism, which provides infants with
the meta-representational concept of belief (as well as desire and pretense), and a
Selection Processing System, which allows them to inhibit their own true belief, and
select and represent the false belief of another agent. Since the Selection Processing System matures later than the Theory of Mind Mechanism (Leslie takes this last
mechanism to be innate), children fail the elicited-response FBT until the Selection
Processing System is in place—around 4 years of age.
The spontaneous-response FBT results seem to challenge this view. Leslie mentions the findings by Onishi and Baillargeon (2005), and argues that they show that the
Selection Processing System might already be operational in 15-month-old infants
(cf. Leslie 2005). However, this results in a very awkward developmental picture,
according to which the Selection Processing System (a) enables false belief understanding and successful spontaneous-response FBT performance in early infancy, (b)
installs a default attribution strategy at a later stage of development, and (c) overrides
this default strategy in order to enable successful elicited-response FBT performance
around 4 years of age. 2 Nevertheless, Leslie seems to endorse this picture, since he
1 As Leslie points out, belief attribution has a default bias because ‘a belief that misinforms an agent is
a useless, even a dangerous thing: beliefs ought to be true. Therefore, the optimal default strategy for the
belief attributer is to assume that an agent’s beliefs are true’ (Leslie 2000, p. 1242).
2 This developmental picture becomes even more implausible in the light of findings on true-belief versions
of the elicited-response FBT (Fabricius et al. 2010), see also Sect. 4.4).
123
Synthese
speculates that ‘sometime between 15 and 30 months, SP [the Selection Processing
System] learns to make the true-belief attribution the default’ (Leslie 2005, p. 532).
This is puzzling. Why would the Selection Processing System ‘make true-belief attribution the default’ given that infants already have an early understanding of false belief,
and this default strategy has to be overridden again at a later stage of development?
The proposals by Leslie and Baillargeon et al. essentially face the same problem:
if it is granted that infants have a default tendency to attribute their own true beliefs
to others, then we cannot differentiate between the spontaneous-response and elicited-response FBT findings by postulating a single inhibition-selection-representation
system. Although both Leslie and Baillargeon et al. seem to endorse a dual-system
approach, the Theory of Mind Mechanism (Leslie) and Sub-system 2 (Baillargeon et
al.) do no actual work when it comes to explaining the development of false belief
understanding. They merely provide infants with the required innate ‘competence’
to understand false belief; however, this by itself does not explain the differences in
performance on the spontaneous-response versus the elicited-response FBT.
3.2 Fine-grained explanatory concepts
A possible way to address this problem is to allow for a more dynamic understanding of ‘competence’—one that goes beyond the classic conception of false belief as a
propositional attitude. This would be in line with a more general trend to avoid the standard folk psychological concept of belief in the explanation of early emerging sociocognitive capacities. There have been various proposals in this direction, ranging from
the postulation of a Naive, Weak, or Minimal Theory of Mind (respectively Bogdan
2009; Tomasello et al. 2003; Apperly and Butterfill 2009), Perceptual Mindreading
(Bermúdez 2009), to an Early Mindreading System (Nichols and Stich 2003). However, these dual-system approaches usually do not come with psychological concepts
that are sufficiently fine-grained to identify the different cognitive capacities involved
in spontaneous-response and elicited-response FBTs.
Bogdan (2009), for example, postulates a Naive Theory of Mind, understood as
an assembled cluster of abilities that enables the grasping and representing of the
basic psychological states of others, specifically gazing, seeing, and emoting. Bogdan
claims that infants slowly move from initially noticing such things as another’s direction of gaze, bodily posture or movement in purely behavioral ways to being able to
track, register, or represent the target of the ‘purposed aboutness’ of another’s attending. Hutto (2008) coins the term ‘intentional attitudes’ to describe the way in which
infants selectively respond to certain aspects of their environment. Young infants are
able to track and attend to the intentional attitudes of other agents without having to
represent (the content of) their psychological states.
These proposals are helpful insofar they point out the limitations of the traditional
folk psychological concept of belief, but it is doubtful whether they give us enough in
return. Although notions such as ‘purposed aboutness’ and ‘intentional attitude’ may
come with the right philosophical connotations, they do not allow us to distinguish
between the specific cognitive demands of spontaneous-response vis-à-vis elicitedresponse FBTs.
123
Synthese
Consider, for example, the study by Southgate et al. (2007) discussed in Sect. 2. In
this study, understanding the experimenter’s ‘intentional attitude’ or ‘purposed aboutness’ comes down to being able to anticipate her behavior on the basis of her visual
perspective. Now consider a different experiment by Woodward (1998). In this experiment, 5-month-olds watched how an experimenter repeatedly reached for object A
as opposed to object B. When the objects’ positions were reversed, infants expected
the experimenter to reach for object A in its new position, and they looked reliably
longer if the agent reached for object B instead. In contrast to the study by Southgate
et al., there is no need to take into account the visual perspective of the experimenter.
Infants do have to grasp the motivational state of the agent, i.e. her action towards
object A. However, this only requires them to deal with differences in what we call
‘motor perspective’.3
It is not clear how the concepts of purposed aboutness or intentional attitude could
help us to identify the cognitive capacities recruited in these studies. Of course, this
is not a decisive argument against Bodan or Hutto, but it does indicate that we need
to look further for a satisfying solution to the developmental paradox of false belief
understanding.
3.3 System interaction
Apperly and Butterfill (2009) have recently offered an interesting alternative to account
for the development of false belief understanding (see also Apperly 2010). They propose that the results on spontaneous-response FBTs indicate that young infants have a
‘minimal’ Theory of Mind, but still lack a full (propositional) Theory of Mind. Apperly
and Butterfill introduce the technical terms ‘encountering’ and ‘registering’ as alternatives to the classic folk psychological concept of belief. Encountering is defined as ‘a
relation between an individual, an object and a location, such that the relation obtains
when the object is in the individual’s field’ (2009, p. 962). A field is defined, simply, as
a certain region of space around the individual. Building on this, registering is defined
as a slightly more complex psychological relation that obtains between an individual,
an object and a location. An individual is said to register an object at a location when
(a) she encounters the object at the location and (b) has not since encountered it somewhere else. A registering is off-target when the object registered is not located where
it is registered to be. The importance of this concept lies in its connection to actions:
‘One can understand registration as an enabling condition for action, so that registering
an object and location enables one to act on it later […] Further, registration also can
be understood as determining which location an individual will direct their actions to
when attempting to act on that object’ (ibid.). These technical notions place specific
limitations on the socio-cognitive capacities involved in spontaneous-response FBT
performance. Apperly and Butterfill mention a few such ‘signature limits’: registering
does not permit tracking mental states that involve quantifiers or complex combina-
3 Infants are able to register another agent’s motor perspective on the basis of their perception of her action
because action and perception share the same representational space (see Sect. 4.2).
123
Synthese
tions of properties of objects (e.g. their location and color), and it does not support
level 2 perspective taking, i.e. keeping track of modes of presentation.
Despite the fact that Apperly and Butterfill’s account seems to provide a promising way to differentiate spontaneous-response from elicited-response FBT performance, it also prompts a pressing question: how are the two Theory of Mind systems
functionally related? According to Apperly and Butterfill, the studies by Clements and
Perner (1994) and Southgate et al. (2007) suggest that there is no direct interaction
between them. They take this to be consistent with the possibility that a minimal, earlydeveloping Theory of Mind system for tracking ‘belief-like states’ is guiding infants’
eye movements, whereas a later-developing, full Theory of Mind system underlies
their explicit judgments about false beliefs.
One of the main motivations for this dual-system approach stems from considerations about conflicting demands on the cognitive basis of false belief understanding.
Despite the fact that children’s performance on executive functioning tests predicts
their later elicited-response FBT performance, infants with only very rudimentary
executive functioning reliably form correct expectations about another agent’s behavior that are sensitive to the agent’s false beliefs (Apperly 2010). However, the question
is whether this justifies the assumption that there is no system interaction. Following Leslie (see Sect. 3.1), we think it is more likely that these conflicting demands
emphasize the importance of the development of executive functioning for false belief
understanding. Furthermore, the assumption that there is no interaction between the
two systems seems unlikely in light of empirical findings. Recent (longitudinal) studies (Aschersleben et al. 2008, Kristen et al. 2011) indicate that several socio-cognitive
capacities involved in the spontaneous-response FBT (e.g., looking-time patterns, eyedirection detecting, etc.) to some extent predict performance on the elicited-response
FBT. Moreover, denying system interaction also implies that the later-developing,
full Theory of Mind somehow remains invisible during the first years of development
and suddenly becomes fully operational at age four. At least we would expect to see
precursors to such a Theory of Mind system. For example, although Baron-Cohen’s
(1995) Theory of Mind Mechanism can be understood as a separate, late-developing
mechanism, it depends on and receives input from the earlier developing Intentionality
Detector, Eye-Direction Detector, and Shared Attention Mechanism. Despite the fact
that there does not need to be interaction ‘all the way down’ or ‘all the way up’ between
two ToM systems, we think it is safe to assume that a solution to the developmental
paradox not only has to account for the dissociations but also for the continuations in
the development of false belief understanding.
In this section, we have shown that current attempts to solve the developmental
paradox of false belief understanding are lacking insofar they (i) try to give a full
explanation of false belief understanding in terms of a single system (whereas we
need two systems); (ii) fail to articulate psychological concepts that are sufficiently
fine-grained to capture the cognitive requirements for spontaneous-response and elicited-response FBT performance, or (iii) ignore questions about system interaction. In
the remainder of this article, we will present a novel dual-system account that is able
to overcome these problems.
123
Synthese
4 A dual-system account of false belief understanding
Central to our account of false belief understanding are the following two systems: an
Association Module, which provides infants with the capacity to register ‘congruent’
associations between agents and objects, and an Operating System, which allows them
to transform these associations into ‘incongruent’ associations through a process of
inhibition, selection and representation (see also De Bruin and Newen 2012).4 The
interaction between these two systems enables infants throughout social development
to register increasingly complex associations on the basis of another agent’s motor,
visual, and cognitive perspective.
4.1 The association module
The Association Module makes it possible for infants to register associations5 between
other agents and objects in the outside world. In order to register such agent-object
associations, infants need to be able to understand other agents as being directed
towards certain objects in their environments (rather than being accidently in their
vicinity).
The Association Module initially allows infants to register primitive associations
based on the agent’s movement towards a given object. Visual habituation experiments
indicate that this capacity emerges around 5 months, when infants begin to respond
selectively to the goals of other agents rather than the physical details of their actions
(Woodward 1998, 2005).6 Some studies suggest that infants also respond to inanimate entities that behave as goal-directed agents (e.g., Luo and Baillargeon 2005).
This seems to depend on the availability of additional cues such as self-propelled
motion, indicating the animacy of the agent (Biro and Leslie 2007).
At this first stage of development, the Association Module registers associations by
encoding the motor perspective that specifies the goal-directed movement of the agent
towards the object. Around 9 months, however, it also starts taking into account the
looking behavior of other agents. This allows infants to register more ‘distal’ associations on the basis of another agent’s visual perception of the object. Woodward (2003)
conducted a study in which infants of 7–9 months old followed the gaze of another
agent. When the infants noticed that the agent looked at the toy and grasped it, they
did not only look at this toy themselves, but also selectively registered the association
between agent and object over other aspects of the event. However, when the infants
noticed that the agent only looked at the object without touching it, they failed to register such an association. By contrast, 12-month-old infants were capable of registering
an agent-object association solely on the basis of the agent’s gaze towards the object.
4 We use the ‘congruent association’ to refer to those associations infants register on the basis of their own
information about the scene, and ‘incongruent association’ to refer to associations that are incompatible
with their own information about the scene.
5 The term is borrowed from Perner and Ruffman (2005), but our interpretation and use of association
differs considerably.
6 Infants also respond selectively to another agent’s intentional versus accidental actions on objects (Wood-
ward 2005; Hamlin et al. 2008).
123
Synthese
According to Woodward (2003) registering an association on the basis of another
agent’s visual perception is more difficult because there is a lack of physical connection, not only in the sense that agent and object are separated in space, but also in the
sense that gaze has no effect on the object. Furthermore, the consequences of gaze for
another agent are not as readily observed (whereas grasping often involves physical
consequences for the object, which could help infants to understand the agent’s goals).
This could also explain why the ability to register associations between gestures such as
pointing and their specific target objects only emerges towards the end of the first years
(Phillips et al. 2002; Sodian and Thoermer 2004; Woodward and Guajardo 2002).
An important premise of our dual-system account is that the infants’ capacity to
register agent-object associations is grounded in the capacity to register associations
between objects and their own goal-directed behavior (cf. Meltzoff 2004; Tomasello
1999). This does not imply that infants only understand those actions they are familiar
with themselves. Empirical evidence suggests that there is in fact a two-way interaction
between their understanding of self and other, to the extent that: (i) the infant’s own
action towards an object influences her perception of another agent’s goal-directed
behavior towards this object (e.g., Sommerville et al. 2005, 2008), and (ii) the infant’s
perception of another agent’s goal-directed behavior towards an object influences her
own goal-directed action towards this object (Hamlin et al. 2008; Meltzoff 1995).
The infant’s perception of another agent’s goal-directed behavior towards an object
also influences how she herself perceives the object. For instance, it has been shown
that behavioral cues (Biro et al. 2007; Biro and Leslie 2007), infant-directed talk and
eye-contact induce gaze-following towards specific objects (Senju and Csibra 2008),
and the use of specific linguistic labels increases the salience of objects over others
(Xu 2002; Xu et al. 2004). Other studies indicate that pointing behavior promotes
the infant’s perception of the ‘surface properties’ of the object of interest, whereas
reaching behavior promotes the ‘spatiotemporal’ properties of the object (e.g., Csibra
and Gergely 2006).
These findings suggest that the Association Module facilitates a two-way interaction between action observation and action production. Georgieff and Jeannerod
(1998) have introduced the ‘shared representation hypothesis’ to articulate the idea
that action and perception essentially share the same representational space. We propose to understand associations as a special kind of shared representations that specify
the relation between agents and objects. Because the associations registered for oneself have the same representational format as those registered for others (Kovács et al.
2010), it is possible to explain how one’s understanding of the goal-directed actions
of others inform one’s own goal-directed action and vice versa.
This idea fits nicely with the findings of mirror neurons–neurons that fire during both the observation and the production of a specific action (e.g., Rizzolati and
Craighero 2004; Gallese and Sinigalaglia 2011) and show a ‘human bias’, in the sense
that they resonate stronger with perceived actions of human versus non-human agents
(Press et al. 2007; Tsai et al. 2008). Research on newborn imitation has been cited as
evidence for an inherited mirror neuron system (e.g., Iacoboni et al. 1999; Decety et al.
2002; Grezes et al. 2003; Iacoboni 2005; Iacoboni and Dapretto 2006), and a number
of studies have reported mirror neuron activity in young infants (e.g., Kanakogi and
Itakura 2010).
123
Synthese
Recently, however, several critical questions have been raised about the existence
of the mirror neuron system and its role in infant development (Gerson and Woodward
2010; Meltzoff 2006). It is also not clear whether the mirror neuron system should
indeed be seen as an innate system, or rather as a byproduct of associative learning
that is shaped through interaction with others (Heyes 2010). Therefore, it should be
emphasized that mirror neurons are just one way of getting at the more general idea
that action production and action perception can be understood in terms of shared
representations. There are other grounds for supporting this idea as well, e.g. on the
basis of proposals about action coding (e.g., Elsner and Hommel 2001; Prinz 2002)
or findings from imitation studies (Meltzoff 2004, 2006; Meltzoff and Moore 1977,
1994; Meltzoff and Brooks 2001).
4.2 The Operating System
The Operating System allows us to explain how infants become capable of registering
incongruent agent-object associations, i.e. associations between agents and objects that
are different from their own. Like Leslie’s Selection Processing System, the Operating
System functions as ‘de-coupling’ mechanism. However, instead of decoupling false
beliefs, it decouples the congruent associations registered by the Association Module.
First, the Operating System enables infants to register the motor perspective of
another agent, by means of the following three processes: (i) a response inhibition
process (infants have to inhibit their own motor perspective), (ii) a response selection
process (infants have to select the motor perspective of the other agent), and (iii) a
representation process (infants have to represent the incongruent motor perspective
of the other agent). This first-order mode of decoupling congruent motor-based associations explains what happens in experiments such as the one by Woodward (1998)
discussed in Sect. 3.2, in which infants anticipate the behavior of another agent on
the basis of her movement towards the object. During the familiarization trials of the
experiment, the Association Module registers a motor-based association between the
agent and object A—one that is congruent with the infant’s own motor-based association. 7 Then the positions of the objects are reversed. In the next step, the operating
system: (i) inhibits the motor information of the infant’s own associations, and (ii)
selects the motor information that specifies the agent’s movement towards object A.
In the final step, the operating system (iii) represents the association as an expectation
of the agent’s behavior towards object A (looking behavior). If the agent reaches for
object A in its new position, the infant’s expectation matches with the behavior of the
agent and no extra processing is required. However, if the agent reaches for object B,
the infant’s expectation is violated and the association module has to register a new
agent—object association. This arguably requires extra processing and hence a longer
looking time. Notice that we can explain how the infants select the motor information
required for step (ii) by appealing to the shared representation hypothesis: the infant’s
perception of the agent’s action informs her own action.
7 The strength of this association is determined by various factors, such as repetition, communicative cues,
presence of others, salience of the object, working memory etc.
123
Synthese
Second, the Operating System also enables infants to register the perceptual perspective of another agent. Again, we assume that infants by default register congruent
associations on the basis of their own visual perception of the object. Therefore, in
order to anticipate another agent’s behavior on the basis of her visual perception, they
have to take ‘offline’ their own perception-based association. This second-order mode
of decoupling involves three Operating System processes: inhibition, selection and representation. Now we are in a position to explain what happens in spontaneous-response
FBTs that require infants to anticipate the behavior of another agent on the basis of
her visual perception of the object. Take the experiment by Southgate et al. (2007), for
example. In this study, 25-month-old infants observe during the familiarization trials
how an experimenter witnesses a protagonist that hides an object in location A. As
a result, the Association Module registers a congruent perception-based agent-object
association, which is compatible with the infant’s own association.8 Then the experimenter gets distracted and turns away from the scene. Meanwhile, the protagonist
moves the object to a different location B without her knowledge. The Association
Module registers a new association between infant and object—one that is incongruent
with the one registered during the familiarization trials. In the next step, the Operating
System: (i) inhibits the perceptual information of the infant’s own association, (ii)
selects the perceptual information that specifies the experimenter’s previous perception of the object at location A (registered during the familiarization trials), and (iii)
represents this information as an expectation (in terms of anticipatory looking) of the
experimenter’s behavior towards A.
Finally, the Operating System allows infants to register the cognitive perspective
of another agent. The distinction between ‘perceptual’ and ‘cognitive’ perspective is
meant to illustrate the differences between understanding the proximal versus distal
goal-directed actions of other agents. Before they become able to register the cognitive
perspectives of other agents, children are already capable of representing their proximal
goal-directed actions as incongruent perception-based associations with a means-end
structure. This provides them with a basic understanding of the ‘in-order-to’ relations
that are characteristic for goal-directed behavior, e.g., ‘the agent reaches out in order
to grasp the object’. However, the emergence of linguistic competence, long-term
memory and executive functioning more in general allows children to (re)configure
the information encoded in their associations in much more abstract ‘in-order-to’
relations. Following Flavell (1992, 1999), we can explain the difference between perceptual and cognitive perspective registration in more detail by distinguishing between
‘level 1 perspective-taking’, i.e. representing what is visible from another agent’s point
of view, and ‘level 2 perspective-taking’, i.e. representing how a given object is seen or
presented to another agent. The latter mode of perspective registration involves a specification of the object’s ‘mode of presentation’ or ‘aspectual shape’ (Perner et al. 2003;
Searle 1992). This is an important requirement for the elicited-response FBT, which
requires children to deal with symbol-based experimental scenarios, i.e. stories or pictures instead of the real-life agents and objects encountered in spontaneous-response
FBTs. We propose that in order to pass the elicited-response FBT and verbally predict
8 Since the experimenter is wearing a visor, the infants are not able to follow her gaze. However, they are
still capable of perceiving what entities are in the agent’s visual field (Sodian and Thoermer 2004).
123
Synthese
another agent’s behavior on the basis of her cognitive perspective, three Operating
System processes are required: inhibition, selection and representation.
4.3 A solution to the developmental paradox
In the remainder of this paper, we will discuss a number of applications of our dualsystem approach to false belief. First, let us address the central issue of this paper: the
developmental paradox. At this point, we could simply claim that this paradox can be
solved as follows: children (a) fail the elicited-response FBT because they are not yet
able to register the cognitive perspective of other agents, and (b) pass the spontaneous-response FBT because they are capable of registering their motor and perceptual
perspective. This however does not yet adequately explain why children have more
trouble with the elicited-response vis-à-vis the spontaneous-response FBT.
One possibility is that the elicited-response FBT is more difficult because it requires
a stronger form of decoupling, which places increasing demands on executive functioning. Several studies have found robust correlations between elicited-response FBT
performance and response inhibition (e.g., Perner and Lang 1999; Cole and Mitchell
2000; Moses et al. 2005) and working memory (Carlson et al. 2002; Hala et al. 2003a;
Perner et al. 2002). Another possibility is that in particular verbal interaction between
infant and experimenter contributes to the difficulty of the elicited-response FBT.
Many experiments have found strong correlations between linguistic competence and
elicited-response FBT performance (Astington and Jenkins 1999; Gale et al. 1996; De
Villiers and De Villiers 2000; Watson et al. 2002; Farrar and Maag 2002; Hale and
Tager-Flusberg 2003b).
A recent study by Scott et al. (2012) indicates that the right answer may be somewhere in between. The experimenters tested 2.5-year-olds with two verbal spontaneous-response FBTs that imposed significant linguistic demands: a preferential-looking
task in which children listened to a story about false belief while looking at a picture
book (with matching and non-matching pictures), and a violation-of-expectation task
in which children watched an agent answer (correctly or incorrectly) a standard false
belief question. Despite their linguistic demands, positive results were obtained with
both tasks. According to Scott et al. (2012), the difference between these two verbal
spontaneous-response FBTs and the classic elicited-response FBTs is that they do not
require infants to answer a direct question about another agent’s false belief. In terms
of the three Operation System processes, what this shows is that 2.5-year-olds are
already capable of (i) inhibiting the cognitive information of their own associations,
and (ii) selecting the cognitive information that specifies the experimenter’s association. However, they fail to (iii) represent this information as a verbal prediction of the
agent’s behavior. This is what gives rise to the developmental paradox of false belief
understanding.
4.4 Recent experimental findings and their implications for our model
In the previous section we have proposed that the developmental paradox results from
the fact that Operation System process (iii), which enables infants to represent the
123
Synthese
agent-object association, is not yet fully developed. A number of recent false belief
experiments also allow us to shed more light on the development and functionality of
the Operation System processes (i) and (ii), i.e. the inhibition process and the selection
process respectively. Fabricius et al. (2010), for example, conducted a series of studies
to test whether children use a ‘perceptual access approach’ to reasoning about mental
states before they understand belief.
According to the perceptual access hypothesis, children of 4- to 5-year old use two
main principles in their reasoning about mental states: (a) seeing and other forms of
perceptual access lead to knowing, and lacking perceptual access leads to not knowing
and (b) knowing leads to acting correctly, and not knowing leads to acting incorrectly
(Fabricius and Khalil 2003). According to Fabricius et al. (2010), standard versions
of the elicited-response FBT are limited to only two alternatives, and therefore confounded in a way that makes correct responses ambiguous: children may pass the
elicited-response FBT because they understand the protagonist will act on a false
belief, or because they conclude she will be mistaken as a result of ignorance. For
example, it has been assumed that children who pass the classic ‘unexpected location’
elicited-response FBT by stating that the protagonist will look in the original location,
reason that she maintains a false belief (which is acquired during her initial visit).
However, as Fabricius et al. (2010) point out, children might also reason that when
the protagonist returns she will not see where the object is and therefore be ignorant
of the object’s current location and will be mistaken and choose the original location
because it is the only ‘wrong’ alternative. This interpretation applies to other versions
of the elicited-response FBT as well: children may deduce that the ignorant protagonist will be wrong and therefore select the false belief alternative. In line with this
idea, the perceptual access hypothesis predicts a U-shaped developmental pattern of
performance in true elicited-response FBTs, in which 3-year-olds who reason about
reality should succeed, 4- to 5-year-olds who use perceptual access reasoning should
fail, and older children who use belief reasoning should succeed.
Fabricius et al. (2010) tested children in this age range (i.e. 3–6.5 years old) with
novel ‘true-belief’ versions of the ‘unexpected content’, the ‘unexpected location’ and
the ‘unexpected identity’ elicited-response FBT. In the unexpected content experiment,
children were initially deceived by a container that holds an unexpected object (i.e., an
M&M-container that contains a pencil). However, instead of putting the unexpected
content back into the container, as is done in the classic version of this elicited-response
FBT, the experimenter replaced the unexpected content with the expected content (i.e.,
the pencil is replaced with M&M’s) while the child observed. The hypothesis was that
4- to 5-year old children who supposedly used perceptual access reasoning should
reason that the other person will not see the contents, and therefore will not know that
M&M’s are inside and as a result will be mistaken and say the container holds a pencil.
Children using the belief approach, however, should reason that the other person will
think that the container holds M&M’s because it is an M&M-container. Besides this
task, Fabricius et al. (2010) also used two true-belief versions of the unexpected location elicited-response FBT, in which the object either remained in its original location,
or was transferred but in the presence of the protagonist, and an unexpected identity
elicited-response FBT, in which children were asked whether another person would
123
Synthese
be able to distinguish between a real rock and an artificial rock made out of sponge
material.
Fabricius et al. (2010) found that the children’s performance was in line with their
perceptual access hypothesis, that is, the studies yielded a U-shaped developmental
pattern of performance in the true-belief elicited-response FBTs: 3-year-olds succeeded, 4- to 5-year-olds failed, and older children also succeeded. These results have
interesting implications for our model, in particular for the Operating System processes of (i) inhibition and (ii) response selection. The findings in 3-year-olds are
compatible with our model insofar as they suggest that infants simply register the cognitive perspective of the other agent on the basis of their own congruent agent-object
associations.9 But what about the findings in 4- to 5-year-olds, which showed that they
failed the true-belief versions of the elicited-response FBTs? Taking the true-belief
version of the unexpected location elicited-response FBT as our example, we suggest
the following interpretation. First, the child observes how the protagonist hides an
object in one of two locations, say location A. As a result, the Association Module
registers a congruent agent-object association, based on the infant’s own cognitive
perspective (since we are dealing with an elicited-response FBT). Then the protagonist leaves the scene, and the object is removed but then again hidden in location A
(which means that the Association Module does not have to register a new association
between infant and object). When the protagonist returns, the Operating System is
triggered by the difference in perceptual access between the infant and the protagonist
(because of the ‘not seeing -> not knowing’ rule), and (i) inhibits the cognitive information of the child’s own association. From there, we see two possible explanations of
why the infant registers an incongruent agent-object association. First, despite the fact
that the Association Module makes the congruent agent-object association available,
the Operating System might fail to (ii) select it because the selection process is not
yet fully operational. At this stage of development, the Operating System might still
be guided by external conditions. That is, the incongruent agent-object association
might be selected because, given the inhibition of the congruent agent-object association, it is the only option left. Second, the Operating System might fail to (ii) select
congruent agent-object association because it is simply no longer available, perhaps
because of limited memory capacity or other factors influencing the strength of the
associations registered by the Association Module. According to this interpretation,
the selection process is already operational but there is only one association left for it to
process.10
9 Since the studies by Fabricius et al. (2010) are elicited-response FBTs, we have to account for them at
the level of cognitive perspective-taking (or level 2 perspective-taking, see Sect. 4.2). The findings that
4- to 5-year old children fail true-belief versions of the elicited-response FBT are therefore compatible with
other results showing that younger infants already select and represent incongruent agent-object associations
based on the visual perspective of another agent (level 1 perspective-taking).
10 However, this explanation is less plausible when we consider children’s capacity for working memory,
which is already quite advanced around this age (Carlson et al. 2002; Hala et al. 2003a; Perner et al. 2002).
123
Synthese
4.5 Operating System processing and looking times
So far we have simply assumed that the Association Module allows infants to register congruent agent-object associations, which can be transformed into incongruent
agent-object associations through a process of inhibition, selection and representation.
However, we haven’t said anything about whether this process translates into longer
looking times. Recent experiments might shed light on this question. Kovács et al.
(2010), for example, have shown that 7-month-olds registered an agent-object association even if the beliefs of the agent were irrelevant to the task they had to perform.
In their violation of expectation spontaneous-response FBT, infants watched how an
agent placed a ball on a table, which rolled behind an occluder. In the next scene, the
ball either stayed behind the occluder or left the scene. Then, the agent left the scene.
In the agent’s absence, either the ball left the scene, returned behind the occluder,
or did not move. Finally, the agent returned, and the occluder was lowered. In half
of the trials, the infants would see the ball behind the occluder. Kovács et al. (2010)
measured looking times as a function of (a) the infant’s ‘belief’ (ball behind occluder
versus no ball behind occluder) and (b) the agent’s belief (ball behind occluder versus no ball behind occluder). They showed that, when infants register a congruent
association (the ball is not behind the occluder) which is eventually not confirmed (it
turns out the ball is behind the occluder), they look reliably longer compared to when
they register a congruent association which is eventually confirmed (the ball is indeed
behind the occluder). This was true both for the infant’s own associations and the
agent-object associations they registered. Kovács et al. (2010) compared the baseline
condition, where neither the infant nor the agent believed the ball to be behind the
occluder, to a condition in which both the infant and the agent believed that the ball
was behind the occluder. When the occluder revealed no ball, infants looked longer in
this condition than in the baseline condition. Moreover, they also looked longer (compared to the baseline condition) when only the agent believed the ball to be behind the
occluder.
We can explain these longer looking times as follows. In the first condition, the
Association Module registers an agent-object association that is congruent with that
of the infant. When the occluder reveals no ball, there is a ‘double violation of expectation’, i.e. both the infant’s own association and her agent-object association do not
match the expected outcome. In the second condition, there is a single ‘violation of
expectation’, since only the agent-object association does not match the expected outcome. However, whereas the first condition only requires infants to process a congruent
association, the second condition requires them to process an incongruent agent-object
association as well. This involves additional Operating System processes, and could
explain why Kovács et al. (2010) found slightly longer looking times for the second
vis-à-vis the first condition (a difference of approximately 2 s).
5 Closing comments and directions for future research
One of the strengths of the dual-system approach to false belief understanding presented in the previous section is that it does justice to the developmental continuity
123
Synthese
of false belief understanding, and gives a clear explanation of how the interaction
between the Association Module and the Operating System provides infants, at each
stage of development, with more advanced capacities to understand other agents. Furthermore, it offers psychological concepts that are sufficiently fine-grained to identify
the different cognitive capacities involved in the spontaneous- and elicited-response
versions of the FBT. In this way, we are able to avoid the problems mentioned in
Sect. 3.
One might wonder whether the Association Module actually deserves the label
‘system’, and hence whether the account we have presented in the previous sections
is really a dual-system account.11 Since the Association Module only provides infants
with congruent agent-object associations, it could be argued that the Operating System
is doing the ‘real work’, i.e. transforming these congruent agent-object associations
into incongruent agent-object associations (which are required for false belief understanding). However, despite the fact that the Association Module is not a ‘dedicated
system’, on our view it still is a necessary prerequisite for (the development of) false
belief understanding. In order to understand that another agent can have a false belief
about a given object, infants need to be able to form associations between the agent and
the object in question. Of course, this also involves other (perceptual) systems, but the
Operating System only uses the agent-object associations of the Association Module
as direct input for further processing. It is in this sense that we take the interaction
between the two systems to be crucial to (the development of) false belief understanding, and consider our proposal to be a dual system account. The difference with
the proposal by Baillargeon et al. (2010), which we have labeled as a single system
account, is that the latter does not postulate any interaction between sub-system 1 and
2, and only uses sub-system 2 to explain false belief understanding (see Sect. 3.1).
If the Association Module is implemented by a mirror neuron system, as we suggested in Sect. 4.1, then what about the neural substrate that underlies the Operating
System? The Operating System clearly depends on the infants’ capacities for executive functioning. As Saxe et al. (2006) have shown, however, there is more to false
belief understanding than executive functioning (see also Perner et al. 2002). Verbal
and non-verbal false belief attribution not only recruit brain regions associated with
executive functioning (such as the anterior cingulate cortex), but also brain regions of a
‘Theory of Mind’ network, which consists of the anterior cingulate cortex, the temporoparietal junction, the superior temporal sulcus and the temporal poles (Amodio and
Frith 2006; Frith and Frith 2003). An interesting question for future research concerns
the specific interactions between the mirror neuron system, the brain areas recruited
for executive functioning, and the Theory of Mind network during performance on
spontaneous- and elicited-response FBTs.
Another question is how the model we have presented in this article relates to
cognitive computational models such as ACT-R (Anderson 2007; Anderson et al.
2004). Verbrugge (2009) has suggested that the ACT-R architecture could be used
to test processing bottlenecks that might occur between the step from first-order
towards second-order belief attributing around 6–9 years of age. Perhaps the ACT-
11 We thank an anonymous reviewer for bringing this to our attention.
123
Synthese
R architecture could also be employed to investigate possible processing bottlenecks
between the step from spontaneous-response to elicited-response FBT
performance.
More specific questions need to be addressed as well. For example, in Sect. 4.4
we suggested that the 4- to 5-year-olds’ failure on true-belief versions of the elicitedresponse FBT might be due to the fact that the Operating System’s selection process
is not yet fully functional. At this stage, the selection process could be governed by a
very basic principle such as learning by exclusion: if there are only two possibilities
and one of them is inhibited, the other one is automatically selected. We see two ways
in which this could be further investigated. One possibility is to test whether the selection process is indeed guided by something like a primitive exclusion principle by
having children observe a protagonist who interacts with more than two objects. Moll
and Tomasello (2007), for example, have shown by means of a spontaneous-response
FBT that in certain circumstances 12- and 14-months-olds can keep track of which
objects (out of three) are known versus unknown to an experimenter in the sense of
being experienced or not by her in the past (cf. Tomasello and Haberl 2003).
Another possibility is to focus more in particular on the sort of exclusion principle
that is used. Fabricius et al. (2010) assume that children reason about mental states
and exclude certain predictions about another agent’s actions on the basis of principles
that are guided by differences in visual information (‘seeing -> knowing’ and ‘not
seeing -> not knowing’). However, it might well be that this reasoning depends on a
more general set of principles, namely, ‘perceiving -> knowing’ and ‘not perceiving
-> not knowing’. These principles are not limited to the domain of visual perception,
but instead depend on the child’s ability to notice epistemic congruency and epistemic
incongruency. This hypothesis could be tested by means of a ‘litmus test’ experiment,
in which the child is first familiarized with a piece of blue litmus paper that turns red
under acidic conditions. Then, during the test trials, both the child and another agent
are shown some tokens of blue and red litmus papers. Subsequently, they observe how
a blue litmus paper is transferred from box A to box B, while they are both told that
there is transparent acid in box B. Now, when the child is asked ‘What color does
the agent think the paper has’, Fabricius et al. (2010) perceptual access hypothesis
predicts that she will give a wrong answer and say that the agent thinks the paper
will be red—given that both the child and the agent have the same visual information
about the scene. However, according to our alternative hypothesis, the child will give
the correct answer: since there is an epistemic incongruency between what the child
knows vis-à-vis what the agent knows, she will say that the agent thinks the paper will
be blue. These and other experiments could help to further substantiate the account of
false belief understanding we have put forward in this article.
Acknowledgments We would like to thank two anonymous reviewers for their valuable comments, and
Anika Fiebich for her insightful suggestions with respect to the distinction between perceptual and cognitive perspective-taking. We would also like to thank the audience of the workshop ‘Reasoning About Other
Minds’ for their feedback on a related poster presentation.
123
Synthese
References
Amodio, D. M., & Frith, C. D. (2006). Meeting of minds: The medial frontal cortex and social
cognition. Nature Reviews Neuroscience, 7, 268–277.
Anderson, J. R., Bothell, D., Byrne, M. D., Douglass, S., Lebiere, C., & Qin, Y. (2004). An integrated
theory of the mind. Psychological Review, 111(4), 1036–1060.
Anderson, J. R. (2007). How Can the Human Mind Occur in the Physical Universe?. New York: Oxford
University Press.
Apperly, I. (2010). Mindreaders: The cognitive basis of “theory of mind”. Hove: Psychology Press.
Apperly, I., & Butterfill, S. (2009). Do humans have two systems to track beliefs and belief-like
states?. Psychological Review, 116, 953–970.
Aschersleben, G., Hofer, T., & Jovanovic, B. (2008). The link between infant attention to goal-directed
action and later theory of mind abilities. Developmental Science, 11, 862–868.
Astington, J. W., & Jenkins, J. M. (1999). A longitudinal study of the relationship between language
and theory-of-mind development. Developmental Psychology, 35, 1311–1320.
Baillargeon, R., Scott, R. M, & Zijing, H. (2010). False-belief understanding in infants. Trends in
Cognitive Science, 14(3), 110–118.
Baron-Cohen, S. (1995). Mindblindness: An Essay on Autism and Theory of Mind. Cambridge, MH: MIT
Press.
Baron-Cohen, S., Leslie, A. M., & Frith, U. (1985). Does the autistic child have a ”theory of
mind”?. Cognition, 21, 37–46.
Bermúdez, J. (2009). Mindreading in the animal kingdom. In R. Lurz (Ed.), The Philosophy of Animal
Minds. Cambridge: Cambridge University Press.
Biro, S., Csibra, G., & Gergely, G. (2007). The role of behavioral cues in understanding goal-directed
actions in infancy. Progress in Brain Research, 164, 303–322.
Biro, S., & Leslie, A. M. (2007). Infants’ perception of goal-directed actions. Development through
cues-based bootstrapping. Developmental Science, 10(3), 379–398.
Bloom, P., & German, T. (2000). Two reasons to abandon the false belief task as a test of theory of
mind. Cognition, 77, B25–B31.
Bogdan, R. (2009). Predicative Minds: The Social Ontogeny of Propositional Thinking. Cambrdige,
MA: MIT Press.
Buttelmann, D., Carpenter, M., & Tomasello, M. (2009). Eighteen-month-old infants show false belief
understanding in an active helping paradigm. Cognition, 112, 337–342.
Carlson, S., & Moses, L. (2001). Individual differences in inhibitory control and children’s theory of
mind. Child Development, 72, 1032–1053.
Carlson, S., Moses, L., & Breton, C. (2002). How specific is the relation between executive function
and theory of mind? Contributions of inhibitory control and working memory. Infant and Child
Development, 11, 73–92.
Clements, W. A., & Perner, J. (1994). Implicit understanding of belief. Cognitive Development, 9,
377–397.
Clements, W. A., & Perner, J. (2001). When actions really do speak louder than words? but only
implicitly: Young children’s understanding of false belief in action. British Journal of Developmental
Psychology, 19, 413–432.
Cole, K., & Mitchell, P. (2000). Siblings in the development of executive control and a theory-ofmind. British Journal of Developmental Psychology, 18, 279–295.
Csibra, G., & Gergely, G. (2006). Social learning and social cognition: The case for pedagogy. In Y. Munakata & M. H. Johnson (Eds.), Processes of Change in Brain and Cognitive Development. Attention
and Performance (xxi). Oxford: Oxford University Press.
De Bruin, L. C., & Newen, A. (2012). An association account of false belief understanding. Cognition, 123(2), 240–259.
Decety, J., Chaminade, T., Grezes, J., & Meltzoff, A. (2002). A PET exploration of the neural mechanisms
involved in reciprocal imitation. NeuroImage, 15, 265–272.
De Villiers, J., & De Villiers, P. (2000). Linguistic determinism and the understanding of false beliefs. In P. Mitchell & K. J. Riggs (Eds.), Children’s Reasoning and the Mind (pp. 191–
228). Hove: Psychology Press.
Elsner, B., & Hommel, B. (2001). Effect anticipation and action control. Journal of Experimental
Psychology: Human Perception and Performance, 27, 229–240.
123
Synthese
Fabricius, W. V., & Khalil, S. L. (2003). False beliefs or false positives? Limits on children’s understanding
of mental representation. Journal of Cognition and Development, 4, 239–262.
Fabricius, W. V., Boyer, T., Weimer, A. A., & Carroll, K. (2010). True or False: Do five-yearolds
understand belief?. Developmental Psychology, 46, 1402–1416.
Farrar, M. J., & Maag, L. (2002). Early language development and the emergence of a theory of
mind. First Language, 22, 197–213.
Flavell, J. (1999). Cognitive development: Children’s knowledge about the mind. Annual Review of
Psychology 50, 21–45.
Flavell, J. H. (1992). Perspectives on perspective taking. In H. Beilin & P. Pufall (Eds.), Piaget’s theory:
Prospects and possibilities (pp. 107–139). Hillsdale, NJ: Erlbaum.
Frith, U., & Frith, C. D. (2003). Development and neurophysiology of mentalising [Special issue]. In
C. D. Frith & D. Wolpert (Eds.), Philosophical Transactions, Series B, 358. On Mechanisms of
Social Interaction, pp. 459–473.
Garnham, W. A., & Perner, J. (2001). When actions really do speak louder than words? but only
implicitly: Young children’s understanding of false belief in action. British Journal of Developmental
Psychology, 19, 413–432.
Gale, E., deVilliers, P., deVilliers, J., & Pyers, J. (1996). Language and theory of mind in oral deaf
children. Paper presented at the Boston University Conference on Language Development, Boston,
MA.
Gallese, V., & Sinigalaglia, C. (2011). What is so special about embodied simulation?. Trends in
Cognitive Sciences, 15, 512–519.
Georgieff, N., & Jeannerod, M. (1998). Beyond consciousness of external reality: A “Who” system for
consciousness and action and self-consciousness. Consciousness & Cognition, 7, 465–487.
Gerson, S., & Woodward, A. L. (2010). Building intentional action knowledge with one’s hands. In S.
P. Johnson (Ed.), Neo-Constructivism. New York: Oxford University Press.
Grezes, J., Armony, J., Rowe, R., & Passingham, J. (2003). Activations related to ”mirror” and
”canonical” neurones in the human brain: An fMRI study. Neuroimage, 18, 928–937.
Hala, S., Hug, S., & Henderson, A. (2003a). Executive functioning and false-belief understanding in
preschool children: Two tasks are harder than one. Journal of Cognition and Development, 4, 275–298.
Hale, C. M., & Tager-Flusberg, H. (2003b). The influence of language on theory of mind: A training
study. Developmental Science, 6(3), 346–359.
Hamlin, J. K., Hallinan, E. V., & Woodward, A. L. (2008). Do as I do: 8-month-old infants selectively
reproduce others’ goals. Developmental Science, 11, 487–494.
He, Z., Bolz, M., & Baillargeon, R. (2012). 25-year-olds succeed at a verbal anticipatory-looking falsebelief task. British Journal of Developmental Psychology, 30, 14–29. doi:10.1111/j.2044-835X.
2011.02070.x.
Heyes, C. M. (2010). Where do mirror neurons come from? Neuroscience and Biobehavioural Reviews,
34, 575–583.
Hutto, D. (2008). Folk psychological narratives: the socio-cultural basis of understanding reasons.
Cambridge MA: MIT Press.
Iacoboni, M. (2005). Understanding others: Imitation, language, empathy. In S. Hurley & N. Chater
(Eds.), Perspectives on Imitation: From Cognitive Neuroscience to Social Science: Mechanisms of
Imitation and Imitation in Animals (pp. 77–99). Cambridge, MA: MIT Press.
Iacoboni, M., Woods, R. P., Brass, M., Hekkering, H., Mazziotta, J. C., & Rizzolatti, G. (1999). Cortical
mechanisms of human imitation. Science, 286, 2526–2528.
Iacoboni, M., & Dapretto, M. (2006). The mirror neuron system and the consequences of its dysfunction. Nature Reviews Neuroscience, 7, 942–951.
Kanakogi, Y., & Itakura, S. (2010). The link between perception and action in early infancy: From the
viewpoint of the direct-matching hypothesis. Japanese Psychological Research, 52(2), 121–131.
Kovács, A., Teglas, E., & Endress, A. (2010). The social sense: Susceptibility to others’ beliefs in
human infants and adults. Science, 330, 1830–1834.
Leslie, A. M. (1987). Pretense and representation: The origins of ”theory of mind”. Psychological
Review, 94, 412–426.
Leslie, A. M. (1994). ToMM, ToBy, and agency: Core architecture and domain specificity. In L. Hirschfield & S. Gelman (Eds.), Mapping the Mind: Domain Specificity in Cognition and Culture (pp. 119–148). Cambridge: Cambridge University Press.
123
Synthese
Leslie, A. M. (2000). How to acquire a ‘representational theory of mind’. In D. Sperber (Ed.), Metarepresentations: A Multidisciplinary Perspective (pp. 197–223). New York: Oxford University
Press.
Leslie, A. M. (2005). Developmental parallels in understanding minds and bodies. Trends in Cognitive
Sciences, 9(10), 459–462.
Luo, Y., & Baillargeon, R. (2005). Can a self-propelled box have a goal? Psychological reasoning in
5-month-old infants. Psychological Science, 16, 601–608.
Meltzoff A. N. (1995). Understanding the intentions of others: Re-enactment of intended acts by
18-month-old children. Developmental Psychology 31, 838–850.
Meltzoff, A. N. (2004). Imitation and other minds: The “like me” hypothesis. In S. Hurley & N. Chater (Eds.), Perspectives on Imitation: From Neuroscience to Social Science (Vol.
II) (pp. 55–77). Cambridge, MA: MIT Press.
Meltzoff, A. N. (2006). The ”like me” framework for recognizing and becoming an intentional agent. Acta
Psychologica, 124, 26–43.
Meltzoff, A. N., & Brooks, R. (2001). ‘Like me’ as a building block for understanding other minds:
Bodily acts, attention, and intention. In B. F. Malle, L. J. Moses, & D. A. Baldwin (Eds.), Intentions
and Intentionality: Foundations of Social Cognition (pp. 171–191). Cambridge, MA: MIT Press.
Meltzoff, A. N., & Moore, M. K. (1977). Imitation of facial and manual gestures by human neonates. Science, 198, 75–78.
Meltzoff, A. N., & Moore, M. K. (1994). Imitation, memory, and the representation of persons. Infant
Behavior and Development, 17, 83–99.
Moll, H., & Tomasello, M. (2007). How 14- and 18-month-olds know what others have experienced. Developmental Psychology, 43(2), 309–317.
Moses, L. J., & Flavell, J. H. (1990). Inferring false beliefs from actions and reactions. Child Development,
61, 929–945.
Moses, L. J., Carlson, S. M., & Sabbagh, M. A. (2005). On the specificity of the relation between executive
function and children’s theory of mind. In W. Schneider, R. Schumann-Hengsteler, & B. Sodian (Eds.), Young Children’s Cognitive Development: Interrelations Among Executive Working
Memory, Verbal Ability and Theory of Mind (pp. 131–146). Hillsdale, NJ: Erlbaum.
Nichols, S., & Stich, S. (2003). Mindreading. An Integrated Account of Pretence, Self-Awareness, and
Understanding of Other Minds. Oxford: Clarendon Press.
Onishi, K. H., & Baillargeon, R. (2005). Do 15-month-old infants understand false beliefs?. Science,
308(8), 255–258.
Perner, J., Leekam, S., & Wimmer, H. (1987). Three-year-olds’ difficulty with false belief: The case
for a conceptual deficit. British Journal of Developmental Psychology, 5, 125–137.
Perner, J. (1991). Understanding the Representational Mind. Cambridge, MA: Bradford Books.
Perner, J., Brandl, J.L., & Garnham, A. (2003). What is a perspective problem? Developmental issues
in belief ascription and dual identity. Facta Philosophica 5, 355–378.
Perner, J., & Lang, B. (1999). Development of theory of mind and executive control. Trends in Cognitive
Sciences, 3(9), 337–344.
Perner, J., & Ruffman, T. (2005). Infants’ insight into the mind: How deep?. Science, 308, 214–216.
Perner, J., & Clements, W. A. (2000). From an implicit to an explicit “Theory of Mind”. In Y. Rossetti & A. Revonsuo (Eds.), Beyond Dissociation: Interaction Between Dissociated Implicit and
Explicit Processing (pp. 273–294). Amsterdam: John Benjamins Publishing Co.
Perner, J., Lang, B., & Kloo, D. (2002). Theory of mind and self control: More than a common problem
of inhibition. Child Development, 73, 752–767.
Phillips, A., Wellman, H., & Spelke, E. (2002). Infants’ ability to connect gaze and emotional expression
to intentional action. Cognition, 85, 53–78.
Poulin-Dubois, D., Brooker, I., & Chow, V. (2009). The developmental origins of naïve psychology in
infancy. In P. J. Bauer (Ed.), Advances in Child Development and Behavior (Vol. 37) (pp. 55–104). San
Diego, CA: Academic Press.
Press, C., Gillmeister, H., & Heyes, C. (2007). Sensorimotor experience enhances automatic imitation
of robotic action. Proceedings of the Royal Society B: Biological Sciences, 274, 2639–2644.
Prinz, W. (2002). Experimental approaches to imitation. In A. N. Meltzoff & W. Prinz (Eds.), The
Imitative Mind: Development, Evolution, and Brain Bases (pp. 143–162). Cambridge: Cambridge
University Press.
123
Synthese
Rizzolati, G., & Craighero, L. (2004). The mirror-neuron system. Annual Review of Neuroscience, 27, 169–
192.
Ruffman, T., Garnham, W., Import, A., & Connolly, D. (2001). Does eye gaze indicate knowledge of false
belief: Charting transitions in knowledge. Journal of Experimental Child Psychology, 80, 201–224.
Saxe, R., Schultz, L. E., & Jiang, Y. V. (2006). Reading minds versus following rules: Dissociating
theory of mind and executive control in the brain. Social Neuroscience, 1(3-4), 284–298.
Scott, R. M., He, Z., Baillargeon, R., & Cummins, D. (2012). False-belief understanding in 2.5-year-olds:
Evidence from two novel verbal spontaneous-response tasks. Developmental Science, 15, 181–
193. doi:10.1111/j.1467-7687.2011.01103.x.
Scott, R. M., Baillargeon, R., Song, H., & Leslie, A. M. (2010). Attributing beliefs about non-obvious
properties at 18 months. Cognitive Psychology, 61, 366–395. doi:10.1016/j.cogpsych.2010.09.001.
Scott, R. M., & Baillargeon, R. (2009). Which penguin is this? Attributing false beliefs about object
identity at 18 months. Child Development, 80, 1172–1196.
Searle, J. R. (1992). The rediscovery of the mind. Cambridge, MA: MIT Press.
Senju, A., & Csibra, G. (2008). Gaze following in human infants depends on communicative signals. Current Biology, 18, 668–671.
Sodian, B., & Thoermer, C. (2004). Infants’ understanding of looking, pointing, and reaching as cues
to goal-directed action. Journal of Cognition & Development, 5(3), 289–316.
Sommerville, J. A., Woodward, A., & Needham, A. (2005). Action experience alters 3-month-old
infants’ perception of others’ actions. Cognition, 96, B1–B11.
Sommerville, J. A., Hildebrand, E. A., & Crane, C. C. (2008). Experience matters: The impact
of doing versus watching on infants’ subsequent perception of tool use events. Developmental
Psychology, 44, 1249–1256.
Song, H., & Baillargeon, R. (2008). Infants’ reasoning about others’ false perceptions. Developmental
Psycholology, 44(6), 1789–1795.
Song, H., Onishi, K. H., Baillargeon, R., & Fisher, C. (2008). Can an agent’s false belief be corrected by an
appropriate communication? Psychological reasoning in 18-month-old infants. Cognition, 109, 295–
315. doi:10.1016/j.cognition.2008.08.00.
Southgate, V., Senju, A., & Csibra, G. (2007). Action anticipation through attribution of false belief by
two-year-olds. Psychological Science, 18, 587–592.
Southgate, V., Chavalier, C., & Csibra, G. (2010). Seventeen-month-olds appeal to false beliefs to
interpret others’ referential communication. Developmental Science, 13, 907–912. doi:10.1111/j.
1467-7687.2009.00946.x.
Surian, L., Caldi, S., & Sperber, D. (2007). Attribution of beliefs to 13-month-old infants. Psychological
Science, 18, 580–586.
Tomasello, M. (1999). The cultural Origins of Human Cognition. Cambridge: Harvard University Press.
Tomasello, M., & Haberl, K. (2003). Understanding attention: 12- and 18-month-olds know what’s new
for other persons. Developmental Psychology, 39, 906–912.
Tomasello, M., Call, J., & Hare, B. (2003). Chimpanzees understand psychological states: The question
is which ones and to what extent. Trends in Cognitive Sciences, 7, 153–156.
Träuble, B., Marinovic, V., & Pauen, S. (2010). Early theory of mind competencies—Do infants
understand others’ beliefs?. Infancy, 15, 434–444.
Tsai, C. C., Kuo, W. J., Hung, D. L., & Tzeng, O. J. (2008). Action co-representation is tuned to other
humans. Journal of Cognitive Neuroscience, 20(11), 2015–2024.
Verbrugge, R. (2009). Logic and social cognition: The facts matter, and so do computational models. Journal of Philosophical Logic, 38(6), 649–680.
Watson, A., Painter, J., & Bornstein, M. (2002). Longitudinal relations between 2-year-olds’ language
and 4-year-olds’ theory of mind. Journal of Cognition and Development, 2, 449–457.
Wellman, H. M. (1990). The Child’s Theory of Mind. Cambridge, MA: MIT Press.
Wellman, H. M., Cross, D., & Watson, J. (2001). Meta-analysis of theory of mind development: The
truth about false-belief. Child Development, 72(3), 655–684.
Wellman, H. M. (2002). Understanding the psychological world: Developing a theory of mind. In U. Goswami (Ed.), Handbook of Childhood Cognitive Development (pp. 167–187). Oxford: Blackwell.
Wimmer, H., & Perner, J. (1983). Beliefs about beliefs: Representation and constraining function of
wrong beliefs in young children’s understanding of deception. Cognition, 13, 103–128.
Woodward, A. L. (1998). Infants selectively encode the goal object of an actor’s reach. Cognition, 69, 1–34.
123
Synthese
Woodward, A. L. (2003). Infants’ developing understanding of the link between looker and object. Developmental Science, 6, 297–311.
Woodward, A. L. (2005). The infant origins of intentional understanding. In R. V. Kail (Ed.), Advances
in Child Development and Behavior (Vol. 33) (pp. 229–262). Oxford: Elsevier.
Woodward, A. L., & Guajardo, J. J. (2002). Infants’ understanding of the point gesture as an object-directed
action. Cognitive Development, 17, 1061–1084.
Xu, F. (2002). The role of language in acquiring object kind concepts in infancy. Cognition, 85, 223–250.
Xu, F., Carey, S., & Quint, N. (2004). The emergence of kind-based object individuation in infancy. Cognitive Psychology, 49, 155–190.
123