There are 4 theoretical positions explaining the developmental data below: the "theory of mind" or metarepresentation hypothesis, simulation theory or disengagement hypothesis, executive function & conditional reasoning as common functional components of false belief tasks and the modular hypothesis.

(n.b. theory-theory and modular hypothesis, simulation theory sees it more as offline decision-making with pretend inputs)

Belief

A belief is representation of the world, a picture in the head. If this representation matches up with the world, then the belief is true. False beliefs misrepresent the world.

Sense

Sense is what the world is represented as

Referent

What is represented is the referent (the world).

Propositional attitude

Mental states are attitudes that agents have towards propositions. E.g. consider the proposition, “it’s raining”. I could, hope that…, believe that…+proposition.

Mental states

Mental states are desires, percepts, beliefs, knowledge, thoughts, intentions, feelings, etc. They play a causal role in behaviour; I don’t go to get some chicken because it’s in the fridge, but because I believe it’s in the fridge and desire it. I’ll still go to the fridge if the chicken’s not there.

Sally puts an object in box X and then departs. Someone else moves the object to box Y during Sally’s absence. When Sally returns, the question to the child is: Where will Sally search for the object-in X or in Y? Most 4 year olds say X, 3 year olds say Y. It seems to be quite a profound problem, e.g. even if Sally says, “I think it’s in X”, they still respond she thinks it’s in Y. In general there’s a developmental transition around 3/4.

If given the same task but solely in verbal form (i.e. the child does not see where objects are hidden or moved to but is merely told so), 3 year olds succeed. If given the task purely in visual form with no verbal input –e.g., in a silent film- then 4 and 5 year olds find the task difficult and it is only at close to 6 years that they succeed.

An experimenter shows a 5-year-old a smarties tube with pictures of smarties on it and asks her what she thinks is in it. "Smarties," she replies. Then the child gets to look inside and discovers that it actually contains pencils, not smarties. The experimenter then asks her what another child who had not yet seen inside the tube would think it contained. "Smarties," the child answers. The experimenter tries the same procedure with a 3-year-old. The response to the initial question is the expected "smarties," but the response to the second is "pencils." Even more surprising is that in response to further questioning, the 3-year-old claims that she had initially thought that there were pencils in the tube and had even said that there were. You get the same results with autistic people. They not only have problems with other people’s false beliefs, but also their own.

There is some correlational evidence that these distinctions tend to develop together; that is, young children who perform well on appearance-reality tasks also tend to perform well on false-belief tasks, visual perspective-taking tasks, and other conceptually related measures. However, exactly what false-belief and appearance-reality tests measure remains the subject of considerable controversy; do they measure the child's developing understanding of mental representation?

Overestimating how much you know and underestimating the other’s

Two children were placed facing each other on opposite sides of a table. A box was placed in the middle of the table between the two children. The questions were: "Does other child know what is in the box?" and "Do you know what is in the box?" Before the questions were asked, either the other or the subject had access to the content of the box. Because the two children were facing each other the subject was fully aware of the informational conditions the other child was exposed to. The older children (5-year-olds) gave correct answers. But younger children (3-year-olds) did not. The most frequent error was denial of the other child's knowledge when the other child had looked into the box or was informed by the experimenter, saying that the other did not know what was in the box. This kind of error was absent in children's assessment of their own knowledge when they’d looked in the box. In another experiment, designed to be sure that the younger children were aware the other child had looked in the box, the subjects were asked both whether the other child had looked in the box and whether the other child knew what was in the box. The children consistently responded affirmatively to the look-question but frequently responded negatively to the knowledge question.

perspective-taking

3 year olds tend to fail and 4-5 year olds tend to pass tests of perceptual appearance of something from one position versus another. They showed a picture flat on the table to a child and the 4-year-old child knew the experimenter saw the picture upside down.

But 3-year olds can tell they can’t see an object on the other side of the wall, but the experimenter can because he’s there.

From 4 to 6, children's understanding of false belief comes with an understanding that perceptual access is important for knowing. However, for them the importance of perceptual access seems to override other sources of knowledge like inference. As a consequence, they fail to see the point in memory cues because such cues enable retrieval of knowledge without direct perception. 3-year-old children also don't appreciate that different properties are gained through different sense modalities, e.g., that you need your eyes to learn what colour an object is, but your hands in order to learn how heavy it is. This understanding develops between 4 and 6 years.

Appearance-reality distinction

Subjects are presented with a sponge made to look like a rock. After discovering each object's true identity or property, subjects are asked how the object currently appears to their eyes (rock) and how or what it really and truly is (sponge). The usual finding is that 3-year-olds tend to give the same answer to both questions, reporting either the appearance twice or the reality twice, as though they do not distinguish conceptually between the misleading perceptual appearance and the underlying reality. In contrast, children of age 4 and older typically can distinguish between appearance and reality. 3-year olds can tell that you shouldn’t drink a glass of milk because it used to have a cockroach in it.

Deception

Deception in stories seems to be understood as—or slightly after—children master the false belief test. It is true that in some situations children do behave deceptively by the age of three. E.g. when accused of some wrongdoing they may deny having done it with a firm, "No", because "no" had beneficial effects in the past.

The knee-jerk test

There is a correlation between false-belief task performances and realising their reflexive knee movement is involuntary.

Preference for faces

Infants develop considerable skill in discriminating different facial expressions over the first 2 years of life.

Attention

Eye direction detection and joint attention

At six months infants can detect eye-direction. From 8-16 months you get joint attention and social referencing.

In subsequent years, children go on to acquire the following four facts about attention.

1. Attention is selective; people do not attend to everything that is in their field of vision or within earshot.
2. Attention entails constructive processing of what has been attended to; different people may mentally represent the same perceptual input differently.
3. Attention is limited; people can attend to only a very limited number of things at the same time.
4. Stimuli can be responded to at different levels of attention or awareness, from unconscious to conscious.

Agency

Infants respond differently to people than they do to objects and seem to expect people to behave differently than objects do. 5-8-week-old babies imitate mouth openings and tongue protrusions produced by an adult but not similar-looking behaviours produced by an object. From 2-4 months the infant seeks face-to-face interaction and turn-taking starts. Infants try to retrieve a just-disappeared object by reaching toward its place of disappearance but try to retrieve a just-disappeared person by vocalizing to the person. They act more surprised when an inanimate object seems to move entirely on its own, with nothing pushing it, than when a person does.

Intentionality

Infants do a variety of things that reflect a dawning awareness of intentionality, using communicative gestures to get you to do something and checking to see whether their attempts have succeeded, e.g. looking at, pointing to, holding up, or vocalizing about an object, following another person's direction of gaze.

At 5 months infants who were habituated to a person grasping a particular object dishabituated more strongly to a change in goal object than to a change in the grasping movement.

Between 9 and 12 months children expect a person looking at an object to manipulate that object and not another object. 12-month-olds expect a person to reach for an object that the person is looking at with positive affect rather than for another one to which the person is not attending.

Fairly sharply around 18 months some drastic changes occur in children's understanding of goals and desires. Children observing a failed action, e.g., trying to pull one of the balls off a small dumbbell, then imitate the intended successful action, not the actually observed action. They must have inferred the intended goal. At 18 months you also get proto-declarative pointing.

By age 3, children may have some ability to distinguish intended actions from non-intentional behaviours such as reflexes and mistakes. For example, when 3-year-olds tried to repeat a tongue twister (e.g. "She sells sea shells by the sea shore") but made errors, they reported that they did not mean to say the sentence wrongly. 3-year olds, who are quite proficient in judging accidents as unintentional, have a hard time realising that their reflexive knee movement is involuntary.

By age 4 or 5 they are able to distinguish intentions from desires or preferences and from the outcomes of intentional actions. E.g. unlike 3-year-olds, they recognize that a person who tried to get object A but chanced to obtain the more desirable object B instead nevertheless originally intended to get A rather than B.

When or how children develop triadic gaze into mind reading skills

Cartoon Larry goes shopping with his mum for a birthday present. In a shop, Larry is described as deciding what he wants, and is pictured signalling by means of eye gaze at 1 of some toys or pets. The children were asked, "What does Larry want?” "Where is he looking?" and "What is he looking at?" 3 year olds could cope with what and where questions, but had difficulty with the want questions, while 4 year olds performed better.

In another study, the children were required to identity the direction of pointing or head direction or eye gaze and infer what was wanted. There was an additional condition in which the character was linked to an object by an arrow. 3 year olds were more accurate using the pointing or head direction cues in inferring what was wanted. There was no general difficulty in linking nonverbal cues to an individual's mental state... rather, the 3 year olds failed to use the gaze or arrow cues, showing a preference for a gestural cue instead of an abstract or symbolic cue. When cues were conflicting, pointing was more salient than eye gaze, but eye gaze was more salient than head direction. Young children can make use of eye gaze for desire inference but they are reliant upon other nonverbal cues such as pointing or head direction.

Linguistic capabilities

Infants are highly attentive to voices from the beginning and can distinguish one voice from another. Newborn infants can distinguish their mother's voice from another woman's based on prenatal, intrauterine auditory experience with her voice. Young babies also have an unlearned ability to hear fine differences between consonant sounds and to perceive them categorically. Infants communicate in a rudimentary way; they smile and coo.

Imitation

Newborns seem able to perceptually represent and imitate another person's movements. They lose the ability and it takes a few months to come back. E.g. the neonate will imitatively stick out its tongue after it has seen an adult do this. Older infants apparently can tell when they are being imitated and prefer to attend to adults who imitate them. This suggests a rudimentary sense of self and other.

Intermodal association

Babies are also capable of other feats of intermodal perceptual representation involving people. By the middle of the first year of life they can match a happy voice with a happy face, and a parent's voice with that same parent's face.

Infant non-egocentricism

18 month olds were given the choice between a tasty cracker and a repulsive piece of broccoli. The infants observed one of the experimenters express the opposite preference supported by clear vocal and facial expressions (e.g., "Ugh, this cracker is yucky! Mmm, this broccoli is yummy."). Children were then requested by that person to hand her something to eat. Most of the younger half of 18 month olds handed her the crackers, whereas most of the older half handed her the broccoli.

Pretend play

From 1½ to 3 years, children's pretence becomes more sophisticated in terms of tracking the counterfactual consequences of pretence. E.g. when naughty Teddy is pretending to have tea in his cup and tilts the cup over a piece of chocolate children are able to state that the piece of chocolate is now wet.

2-year-olds and even older children may lack a fully mentalistic conception of pretence as well as belief. Children were presented with a doll named Moe who knows nothing at all about rabbits but happens to be hopping like one. The children were then asked if Moe was or was not pretending to be a rabbit. Before the age of 4 or 5 years they say that Moe is pretending to be a rabbit, despite having agreed that she did not know how rabbits hop. Young children conceive of pretence in a non-representational way, as “acting like”. 3-4 year olds classify pretence with physical activities, such as clapping one's hands, rather than with mental activities, such as thinking.

Having said that, if a character that had a boot on his fishing line was pretending that he had caught a fish, children agreed (pointed to the relevant think bubble) that he was thinking of a fish rather than thinking of a boot.

The Moe finding might not have picked up so much on a problem about pretence but on a problem about understanding mental states. By the age of 7 or 8 most children responded like adult subjects to such tasks, insisting that someone who has a mental state about something (e.g. pretending to be or wanting a kangaroo) has to know what it (kangaroo) is.

Word reference

Infants develop the ability to learn what an object is called by reading the adult's attentional focus when the adult labels it. E.g. Infants of 19-20 months seem to recognize that it is the adult's attentional focus rather than their own that gives clues as to the adult's referential intent.

Social referencing

By 11 months, infants socially reference; they will look to the caregiver if a stranger walks into the room. By 12 months infants are capable of understanding that the adult's behaviour is about the object the adult is attending to when expressing the positive or negative affect (rather than a sympathy reaction). This implies they know the caregiver’s perception of the event is important.

Manipulation of others

Older infants sometimes appear to be trying to manipulate other people's emotional responses. Even toddlers occasionally seem to try to change other people's feelings, or at least change their affective behaviour.

Empathy & desire to irritate

From 1-2 months you get primary circular reaction; if one baby starts crying so does the next. In the second year of life, they begin to comfort younger siblings in distress by patting, hugging, or kissing them, and they may even bring a security blanket to an adult in pain. Young children sometimes tease or otherwise annoy siblings, as though hoping to frustrate or anger them. Such behaviours, positive or negative, suggest that young children are beginning to identify the conditions that elicit or change emotions or behaviours.

Language to describe mental states

Around 1.5 to 2 years of age, children use words that refer to internal states you can’t see. The states most commonly talked about at this early age relate to perception ("I see a car"), desires ("Want juice"), and emotions ("Those scare me"). By 3 years they’ll use cognitive terms like “think” and “know”.

About perception

3 year olds understand that a person will see an object if and only if the person's eyes are open and aimed in the general direction of the object and if there are no vision-blocking obstacles interposed between the person and the object. If you give them a task where one person can see the front of a picture and another person can only see the back, they can say who knows what the picture is. So they know 2 people can have different knowledge.

Later in the preschool period, they go on to recognize that the same thing may present different visual appearances to two people if they view it from different positions.

About desires

Children seem to show some awareness of the mental state of desire by the end of infancy.

By age 3 they grasp causal relations between desires, outcomes, emotions, and actions. This suggests they are developing something like an implicit theory. E.g. they seem to recognize that people will feel good if they get what they want and feel bad if they do not, and they seem to understand that people will quit searching if they find the desired object they have been looking for but keep searching if they do not.

About emotions

Young preschoolers evidence an understanding of emotions v/s actions (e.g., hitting) and expressions (e.g., smiling) that emotions cause, and they distinguish between the different emotions of different individuals.

In subsequent years children come to understand, e.g. that people do not always really feel what they appear to feel, that people's emotional reactions to an event may be influenced by earlier experiences or by mood, and that people can experience two conflicting emotions simultaneously.

About knowledge

By the end of the preschool period, they realize that the word know expresses more certainty than think or guess and is a surer guide to the true state of affairs.

4- and 5-year-olds claim that they have always known information that they have just learned during the experimental session. While young elementary school children know how and when they came to know some recently acquired fact.

In the late preschool and middle-childhood periods, children discover that to acquire knowledge perceptual information has to be adequate as well as present. They realize that one cannot know an object's colour merely by feeling the object, that one often cannot be certain of an object's identity when only a little bit of it is visible, and that one's interpretation of impoverished/ambiguous perceptual input may be influenced by biases/expectations.

About thinking

Preschoolers understand that thinking is an animates’ in-the-head activity that can take as its objects non-present and non-real things. Finally, they have some ability to infer the presence of thinking in another person provided that the cues are very strong and clear.

Difficulties with others’ thinking

They are unaware of the stream of thinking in people who are conscious. E.g. preschoolers do not consistently attribute any mental activity at all to a person who just sits quietly. They do not assume that something must be going on in a person's mind, even when they know that the person is looking at or listening to something, reading, or talking to another person.

Younger children are inclined to attribute self-awareness and decision-making abilities to an unconscious person, e.g. most 5-year-olds, and few 8-year-olds and adults, claim that people know they are asleep while they are deeply asleep and not dreaming. When a videotaped sleeping person stirs in response to a light touch but does not wake up, older children and adults believe that the sleeper sort of felt it but did not consciously think that he or she had been touched; in contrast, preschoolers tend not to make this distinction, usually saying that the person would experience the conscious thought as well as the low-level feeling.

They are poor at determining when a person (self or other) is thinking and also what the person is and is not thinking about, even when the evidence is very clear.

Difficulties with own thought

When preschoolers are asked to report their own mental activity they also have difficulties. A group of 5-9 year olds were asked to sit in the special “Don’t Think chair” and they were instructed to not think for a while. After about 20 seconds they were allowed to move over to the normal chair and were asked: "While you were sitting over there in that Don’t Think chair, you tried not to have any thoughts. What happened? Did you have no thoughts at all or did you have some thoughts anyway?" Very few 5-year olds but most 8 year olds and adults admitted to having had some thoughts.

They tend to be very poor at recalling or reconstructing both the fact and the content of their own recent or present thinking, even in situations especially designed to make such introspection extremely easy. They seem largely unaware of their own on-going inner speech and may not even know that speech can be covert.

Higher order false beliefs

E.g. when John thinks that Mary mistakenly thinks… are understood nearer 10 years of age. Children cannot really distinguish more complicated speech acts, like irony from lies. Lies as well as irony are false statements, and they are intended to be false by the speaker. The difference emerges at the second order level: irony is not intended to be believed. Faux pas are understood in ones teens.

1-year-old infants may not be aware that people have seeing experiences and mentally attend to objects, although behaviours may constitute developmental stepping-stones to an eventual awareness of mental states. Infants would not necessarily need to be aware of people's mental states to do many of the social-cognitive things that they do. E.g. in cases of social referencing, they only have to read caretaker expressions of fear as meaning, "this object is dangerous" learned through conditioning. Infants could be predicting people's behaviour from their gaze direction. They would not actually have to be aware that the people are also having perceptual and cognitive experiences as their eyes move about.

Children of a mind-reading species are likely to show precursors

Older human infants who look as though they are attributing seeing and other mental states to us are undeniably going to be making genuine mental state attributions in a few months. Given that fact, it is not unreasonable to suppose that they might be doing some precursor or early version of the same thing. Theory theorists, modularity theorists, and simulation theorists have all made the argument that it would be hard to imagine how infants could learn to make mental state attributions later if wholly incapable of anything like it earlier.

The tests may be artificial and confusing

Subjects who look as though they lack a certain understanding in task situations (e.g. of false belief) may nevertheless possess that understanding. The tests may be artificial and confusing.

Infants show a range of suggestive behaviours

Older infants do not do just one or two things suggestive of a mentalistic conception of others. Rather, they do a variety of things, all of which point to the dawning of some such conception: For example, around this age important developments take place in areas as diverse as pretence, self-recognition, imitation, empathy and internal state language.

1.     Information processing perspectives

Some investigators argue that young children's failures on false-belief and other theory-of-mind tasks are due to more domain-general information processing problems. Examples are limited memory abilities and the inability to inhibit a dominant, ready-to-go response.

Episodic memory

The difference between free recall and cued recall is sensitive to the availability of episodic traces, and this correlates with children's ability to, e.g., explain why they know what is in a box ("because you've shown me", or "because you told me") or understand that they need to look inside a tunnel to find out what colour the object has but use their hand to find out how heavy it is.

Executive inhibition and inhibitory control

3 to 5 ½ year olds were tested on false belief, forward digit span (taxing the auditory loop) and backward digit span (taxing the central executive). Backward digit span was a significant predictor of false belief understanding. False belief understanding correlates with dual task performance: naming three objects while counting them, e.g., "one is a doll, two is a car, three is a spoon") or naming the objects while finger tapping. This suggests that the relationship between false belief understanding and working memory hinges on the central executive component of working memory.

Luria's hand game

After imitating the experimenter's hand shape they had to switch to doing the opposite shape.

Experimenter: flat hand Ò fist Ò flat hand Ò ...

Child: fist Ò flat hand Ò fist Ò ...

The natural tendency to imitate interferes with what one is supposed to do. In order to inhibit this natural tendency it is not sufficient to just concentrate more on one’s proper task because that task involves concentrating on what the experimenter does and that enhances the interfering natural tendency to imitate. Automatic inhibition (of irrelevant stimuli) is thwarted and executive inhibition is needed. Executive inhibition consists of representing the existence of the interfering action tendency in order to inhibit that tendency. The false belief task is mastered at the same time as executive inhibition tasks. It’s been suggested that the false belief task itself is an executive function task requiring the inhibition of a predominant response, namely to answer with the object's real location. This however is unlikely, because in the "explanation version" of the false belief task the child observes Maxi look in the wrong cupboard and is asked to explain why he did so. Children who do not understand false belief tend to say nothing, which means there is no prepotent answer strategy that needs to be inhibited.

Windows task

The ability to give correct answers in the false belief task correlates with the ability to suppress pointing to where an object is in favour of pointing to where it is not. Children were trained to point to one of two opaque containers one of which contained a sweet. If it contained the sweet the opponent consumed it. If the container was empty the child got to eat the sweet. This taught the child to point to the empty container. After 15 such trials containers with windows on the child's side were introduced, so that the child, but not the opponent, could see which container was baited. Most 3 year olds pointed to the full container. Most of them kept pointing there for up to 20 test trials. They do better if they’re asked to mislead the other person (autists don’t), or use an arrow.

ToM necessary for executive function? or Executive function necessary for ToM?

3½ years or 4½ years children with autism are not particularly executive function impaired in relation to an age and IQ matched control group. Over the next years the control group gets better at executive function while the children with autism do not improve so that by 5½ years there is a clearly identifiable executive function gap between controls and autists. Other disorders associated with executive dysfunction (obsessive-compulsive disorder, Tourette's syndrome, attention deficit with hyperactivity disorder…) do not necessarily result in autism. This implies theory of mind is either a necessary ingredient for executive control or common brain regions. There are also cases where children did better on executive function than on false belief.

counterfactual Conditional reasoning

Maybe the correlation between the false belief task and executive function tasks is due to a common underlying conditional reasoning structure? The false belief task is a reasoning task. The Maxi story requires deduction because it never mentions where Maxi thinks the chocolate is (it’s not a recall task). If (this weren’t true), then what? Theory of mind tasks (e.g., false belief task) correlate with tasks that require embedded conditionals, e.g. Maxi who had put his chocolate into location A didn't see how his mother used some of the chocolate for baking a cake and then putting the rest of the chocolate into location B. Children's ability to answer the false belief question, "Where will Maxi look for his chocolate?" correlated highly with their ability to answer the counterfactual conditional, "If mother hadn't baked a cake, where would the chocolate now be?"

If-if-then

By the age of 2½ years children can sort according to a single criterion, e.g., put things that look the same together. However, they have problems sorting according to two rules, e.g., "put the car to target A and the flowers to target B". By about 3 years children can do that. But not until about 4 years can they switch sorting rules, e.g.

Sort by colour: the blue flower goes with the blue car and the green car goes with the green flower

Sort by shape: the blue flower goes with the green flower, and the green car goes with the blue car

These two tasks also qualify as executive function tasks since the antecedent condition established under setting 1 interfere with the links required under setting 2.

Against executive function & conditional reasoning as a common functional component

This data speaks against executive components in ToM tests and doubly embedded conditionals as accounts for the observed developmental relationship between ToM and executive function performance.

The appearance-reality distinction

Remember the sponge that looks like a rock? The appearance-reality distinction task correlates with the false belief task, and a metarepresentational account explains this quite well, while a counterfactual reasoning account can’t.

The knee-jerk test

There is a strong correlation between false-belief task performance and realising their reflexive knee movement is involuntary. The knee-jerk reflex task does not contain any obvious executive component or any conditional reasoning requirements.

2.     Theory-Theory

The child as scientist

The theory-theory maintains that prediction, explanation and interpretation exploit an internally represented theory or knowledge structure - a folk psychology. Theory theorists view the child as a scientist, and argue that folk psychology is theoretical in nature. There are parallels between the maturation of theories in science and the child’s developing theory of mind. In science there are paradigm shifts, e.g. from classical to relativistic mechanics. 3 and 4 year olds differ markedly on a range of tasks. But children’s theories are different from scientific theories in that they’re implicit, not conscious.

There are some distinctions to be drawn among different types of theory-theories. Until recently, most models posited internally represented knowledge structures that invoked explicit rules or explicit sentence-like principles. But there has been a growing dissatisfaction with sentence-based and rule-based knowledge structures, e.g. connectionist models store the knowledge used in making predictions in the connection strengths between the nodes of a network. In many of these systems it is difficult to view the network as encoding a set of sentences or rules.

Three properties of a theory

To constitute such an informal theory, a body of knowledge must have three properties. Theory of mind satisfies all three of these conditions.

Specific ontology

First, it must specify a set of entities or processes (an ontology) that are found in its domain of application and not in other domains. Entities or processes such as beliefs, desires, and thinking are found only in the domain of the mental and thus satisfy the ontological criterion.

Causal principles

Second, it must use causal principles that are likewise unique to the theory's domain. The reason that people define anger in terms of causes and effects is because the role of anger in a conceptual causal network is its meaning. Psychological causality (she tried to get it because she wanted it and thought she could get it, etc) is also found only in the domain of the psychological; physical objects are not caused to move by such mental states.

Interrelated concepts

Finally, the body of knowledge must comprise a system of interrelated concepts and beliefs rather than just a collection of unrelated contents, e.g. what we perceive influences what we believe, what we believe may bias what we perceive, beliefs and desires may lead to intentions which may lead to goal directed actions; the success or failure of these actions will cause emotional reactions…

Characteristic intermediate processes

Theory theorists believe that experience provides young children with information that cannot be accounted for by their present theory of mind, information that will eventually cause them to revise and improve that theory. The role of experience is viewed as similar to that in Piaget's equilibration theory: Experience engenders disequilibrium and eventually a new theory.

Denial

Scientists stick to favoured theories in the face of contradictory evidence before abandoning them. The initial reaction of a theory to counterevidence may be denial. The interpretive mechanism of the theory may treat the counterevidence as noise, mess, not worth attending to, e.g. a 3 year old will be adamant they said there were smarties in the tube.

Auxiliary hypotheses

At a later stage the theory may develop ad hoc auxiliary hypotheses designed to account superficially for the counterevidence. But such auxiliary hypotheses undermine the theory's coherence. The theory gets ugly and messy. The preference for simple theories over complex ones plays an additional major role in theory change.

Appearance of an alternative

The next step requires an alternative model to the original theory.

Intense experimentation & observation

A final important feature of theory formation is a period of intense experimentation and/or observation.

Three-step developmental sequence

There’s evidence for the following three-step developmental sequence of steps or milestones toward the adult theory of mind.

2 year old desire psychology

Around age 2, children acquire a desire or internal-state psychology. This psychology includes a simple conception of desires, emotions and perceptual experience or attention. The conception is simple in that although mentalistic, it is non-representational. Desire is an intentional construct, but the ‘object’ of a desire is the object in the world that the desire is about, rather than a representation of the object in the head of the person. Beliefs are necessarily representational, as the ‘object’ of a belief is a proposition about an object, event or state of the world. 2 year olds fail in cases that require the inferrer to attribute internal states whose ‘objects’ are not in the external world. Two year old children can predict the subsequent emotion or action of a story character who either finds, fails to find, or finds a substitute for a desired object, e.g. the character will stop looking if they find the object, or feel sad if they fail to find it. A desire psychology will fail to offer an explanation of why two people with the same desire or the same person at different times may act in different ways, of why a person will act in a way which in fact frustrates their own desires, and will also fail to recognise surprise (rather than happiness) when a desire is satisfied.

3 year old desire-belief psychology

By 3, children are 'reality psychologists', having a non-representational understanding of beliefs. Three year olds view believability and desirability as an objective feature of the world. Around age 3, children begin to talk about beliefs and thoughts as well as desires, and they seem to understand that beliefs are mental representations that can be false as well as true and can differ from person to person. However, at this age they continue to explain their own and other people's actions by appeal to desires rather than beliefs. They’re not interested in beliefs. They’ll understand that John has gone to the fridge because he wants some milk. 3 year olds don’t appreciate that people will act on their beliefs, even when they’re false, just if they’re true.

They’re copy theorists, because they believe their beliefs copy the world. They have a copy understanding of representations. Representations are understood as derived directly from the object, exact copies of the reality that provided the representation. Three year olds are therefore able to predict and explain actions which result from either ignorance or incomplete knowledge, as with the former children think that there is no belief and thus no copy of reality, and in the latter that the person has a copy of reality which whilst incomplete is still a copy of the reality the child perceives.

4 year old belief-desire psychology

At about age 4, children begin to understand that what people think and believe, as well as what they desire, crucially affects how they behave. 4 year olds pass the false belief task because they understand that people’s beliefs interpret the world, rather than are direct copies. They have an interpretive understanding of representations. An interpretive understanding recognises representations as interpretations of reality, which may differ from person to person despite similarities in the reality from which the representations are derived.

two phases in development of children’s theory of mind

An alternative account of the development of children’s theory of mind claims 2 phases, one corresponding to children’s failing the false belief task, and a phase corresponding to children’s passing the false belief task.

Situation theorists

When they fail the false belief task they’re situation-theorists; similar to the desire-theorists mentioned earlier because they know that if Amy’s thinking of books on the shelf she’ll act towards the books on the shelf.

Representation theorists

At 4 they become representation theorists because then beliefs are representations of the world. False beliefs are misrepresentations. At 3, children have a problem with the concept of representation, that’s pictures, language, photographs… not having a complex understanding of representations they can’t possibly understand false beliefs.

deaf children and those with autism

Children have problems with non-mental representations. Children see a Polaroid being taken of big bird on a bed. While it’s developing big bird moves to a chair. 3 year olds say big bird will be in the chair in the photo.

The fact that deaf children and those with autism were seen to be as adept as normal 4-year-olds at understanding the representational properties of photographs despite their poorer than normal comprehension of false belief suggests that the growth of representational understanding may be a domain-specific process, and raises questions about the kinds of experiences that foster the growth of knowledge in the physical and people domains. By watching a camera take a picture, then viewing the emergent photograph, it is feasible for a child to gain an understanding of the representational outcomes of photography in a completely nonverbal way, through direct observation. In a similar way, a conversationally restricted deaf or autistic child might be able by simply watching people and conversing about tangible objects to develop a simple "photographic" working model of the human mind as a storehouse of purely reality-based knowledge that has been postulated by some researchers to be a precursor in normal 3-year-olds to a full-fledged theory of mind. But in order to fully appreciate false belief and other mental states like fantasy, deception and forgetfulness in which a person's thoughts are in direct conflict with known reality, it may well be necessary to engage in a more sophisticated form of conversational exchange. Most parents of deaf children report that communication with their deaf children is limited to topics with a visual reference. Such a limitation would selectively preclude the mother's sharing with her deaf child any information about her false beliefs and other intangible mental states while still enabling dialogue about simple perception, as well as other shared experiences like drawing pictures or taking photographs.

3.     Modularity Theory

Another theory postulates innate or early maturing modular mechanisms dedicated to mental state computations. Beliefs are propositional attitudes, not representations. Modularity theorists believe that young children are not acquiring a theory about mental representations at all. Rather, they acquire through neurological maturation of a succession of three domain-specific and modular mechanisms for dealing with agents versus non-agent objects. Although experience may be necessary to trigger the operation of these mechanisms, it does not determine their nature.

Neural substrate

Active in theory of mind

The brain regions active in theory of mind are identified are part of the prefrontal cortex and the border between prefrontal and premotor cortex, i.e., regions that have traditionally been linked with the control of voluntary action. The most rostral part of the frontal lobe, the extensive prefrontal cortex, is well developed only in primates. The left medial frontal cortex (Brodmann area 8 and partly 9) is responsible for understanding mentalising stories (involving misunderstandings, double bluff, deception…) in contrast to stories about physical events.

Initiating Joint Attention

Initiating Joint Attention skill appears to be associated with a complex system of brain activity involving the left medial frontal cortex in toddlers from 14 to 18 months of age.

Putting action offline after generating visuomotor imagery

A group of patients with prefrontal lesions present the so-called "utilization behaviour". These patients will compulsively imitate gestures or even complex actions performed in front of them by an experimenter. Similarly, when faced with usual objects (e.g., a glass and a bottle of water) they will not be able to refrain from using these objects in a compulsive way (pouring water in the glass and drinking large quantities of water, etc). This behaviour can be explained by an impairment of the inhibitory control normally exerted on the elementary motor schemas. In addition, it stresses the role of prefrontal cortex in organizing behavioural inhibition. One possible consequence of this impairment is that frontal patients presenting this syndrome should be unable to generate motor imagery without immediately transferring the imagined action into motor output.

Goal directed action, or generating motor imagery

A few experimental results suggest that neurons in the monkey prefrontal cortex might have properties relevant to the function of not being influenced by completion of the intermediate steps of the action (e.g., by activation of elementary schemas), but continuing firing until the final goal has been reached. This sustained activity would represent the reference (the goal) to which the current state of execution of the action would be compared. Accordingly, these neurons would remain activated as long as the represented action would not be completed, including in situations where the execution would be blocked. In this case, where the action would not take place, the sustained discharge would be interpreted centrally as a pure representational activity and would give rise to mental imagery. The prefrontal areas where these neurons are located are reciprocally connected with abundant projections to and from the posterior parietal cortex and the premotor area. It is thus possible that the prefrontal cortex for achieving the selected action plan can gate the elementary schemas available in the posterior parietal and the premotor areas.

Elementary schemas = subgoals =visuomotor imagery

Brodmann’s areas 6&7

If-if-then Conditional reasoning

Studies in macaques show that Brodmann area 8 is necessary for conditional reasoning that requires a decision on which object to manipulate. The task for which area 8 is required can be described as a double conditional, e.g. if discriminatory stimulus, then if round, then look for peanut there.

Modulating direction of gaze

Area 8 is anterior of the premotor cortex. It facilitates eye movements and is involved in visual reflexes as well as pupil dilation and constriction. Brodmann area 8 helps orchestrate saccadic eye movements. It contains neurons coding a memory map of visual space. Lesions provoke disturbances in memory-guided saccades. It’s implicated in the conscious circuit, transducing amygdalar (emotional) reactions into modulations of direction of gaze.  Destruction causes the patient not to be able to voluntarily deviate his eyes, but he can track an object.

Selecting objects based on one's past response or order of experience

Area 9 in conjunction with area 46 plays a role in the control of delayed responding. Removal of areas 9 and 46 destroyed the ability to “move that object you haven’t moved before”, i.e., action depending on memory of the object and one's action. It didn’t impair “act on that object you haven’t seen before”. PET scans show the same areas to be important in, “point to objects in any order but never to the same object twice”. Removal of 9 and 46 impaired monkeys' relearning of the ability to repeat a number of presses on one key on another key, while removal of area 46 alone enabled quick relearning. Area 9 is involved in selecting objects based on one's past response or order of experience, e.g. "move that object you haven't moved before".

Mental words

The right orbito-frontal cortex (Brodmann areas 10-14) is activated when listening to mental words (e.g., know, want, remember...) in contrast to action words or body parts (e.g., hand, move, tooth...). Brodmann Areas 9, 10, and 11 are involved in cognitive processes like reasoning and judgement, which may be collectively called biological intelligence.

Working memory

Working memory -tasks produce non-lateralised activation in Brodmann areas 4, 6, 9, 10, & 46 independent of sensory modality.

Asperger syndrome

Asperger syndrome is a mild variant of autistic disorder. The Asperger participants correctly answered most of the test questions of the mentalising stories and also read these stories slightly faster than the physical stories. The Asperger group did not activate Brodmann’s area 8/9 during ToM stories. They activated a neighbouring area, Brodmann’s area 9/10, when processing ToM stories. Controls also activated this area, but to a far lesser extent than area 8/9. The results suggest that the Asperger subjects’ mentalising performance was subserved by a brain system in which one key component was missing.

Cognitive mechanisms

Leslie

In Leslie’s model three cognitive mechanisms help the child’s developing concepts of agency.

Theory of Body mechanism distinguishes agents

The first mechanism, called Theory of Body mechanism, develops early in the first year. The Theory of Bodies mechanism enables the infant to distinguish agents from other objects on the basis of the perceptual nature of their mechanical interactions with other objects. The infant is predisposed to expect certain types of movement characteristic of agents. The more an object changes motion state by itself and not as a result of external impact, the more likely it is that it is an agent. Once the agent v/s non-agent distinction is made, agent-specific conceptual structures can develop.

Theory of Mind mechanism system 1 distinguishes goal-directed actions

"ToMM system1" informs the child’s understanding of agents and action. This first Theory of Mind mechanism, which comes into play later in the first year, allows the infant to construe people and other agents as perceiving the environment and as pursuing goals.

Theory of Mind mechanism system 2 distinguishes propositional attitudes

Finally, a second Theory of Mind mechanism begins to develop during the second year of life. "ToMM system2" informs the child’s understanding of agents and attitudes, or the ability to represent the mental states of agents. This third mechanism allows children to represent agents as holding attitudes toward the truth of propositions. Equipped with this mechanism, children are able to compute propositional attitudes such as Jane is pretending that this empty cup is filled with tea, Peter thinks that this candy box contains candy. Children do not have a problem with representation, as shown by pretend play…a type of propositional attitude.

So why do they fail the false belief task?

Leslie claims we’re born with a theory of mind module, and coupled with it is a selection processor. When there are tasks with two response options you’ve got to select the right one. As adults we suppress where the object actually is in the Maxi-task and say where Maxi thinks it is. Children can’t ignore what’s staring them in the face.

Baron-Cohen

Baron-Cohen has offered another similar model. As in Leslie’s model, the origin of concepts specific to the domain of theory of mind lies in mechanisms that allow agents to be distinguished. In both models, the distinction is as much constructed as perceived.

Eye Direction Detector

In this model, the crucial perceptual feature used by the infant to distinguish agents is the presence of eyes. The "Eye Direction Detector" computes the direction of the agent’s gaze.

Shared Attention Mechanism

The "Shared Attention Mechanism", comes on line later in development, & serves to direct the infant’s attention to the object of the agent’s intentions and provides the means by which the child can develop representations of the agent’s mental contents.

Pretence & ToM

Leslie's decoupling explanation argues that pretence and mentalising abilities require the functioning of the same innate module. The maturation of Theory of Mind mechanism system 2 distinguishes propositional attitudes and mediates the ability to understand pretence and the ability to understand false-belief and other mental states. Pretend-play skills develop during early childhood. This modular theory-of-mind mechanism permits the 18-24 month old to engage in pretend play and to understand as pretence the pretend actions of others. This metarepresentational capacity prevents the child from being confused when someone pretends that a banana is a telephone. It does so by decoupling the temporary pretend identity of the banana (telephone) from its permanent real identity (banana). The child can then compute the relation: "This person is pretending that this banana is a telephone." Evidence: social flexibility in pretend play correlates with ToM performance.

Arguments against pretence being related to ToM

One argument against this claim is the roughly two-year age gap between children's comprehension of pretence and their comprehension of false-belief, deception, appearance-reality, and different perceptual perspectives. If understanding of pretence and understanding of false belief are both mediated by Leslie's theory of mind mechanisms, then why does the former appear so much earlier in childhood than the latter? Leslie and others have argued that performance obstacles explain the late display of false-belief understanding on standard tests, but it is hard to believe that 2-year-olds really do understand false belief.

Evidence for propositional attitude understanding being distinct from representational understanding

Mainstream and autistic kids got the Maxi- and photo tasks:

Leslie uses the fact that subjects with autism typically perform poorly on the false-belief task, which requires the representation of counterfactual belief, and the fact that they perform as well as control subjects on Zaitchik’s "picture" task, which requires the representation of a counterfactual pictorial, rather than mental, representation, to argue that representation of beliefs is subserved by a distinct mechanism that constructs not merely representations, but "about mental states-representations", which represent only propositional attitudes of agents and not the representational content of photographs.

We exploit such meta-representations when we think that

Maxi believes that the chocolate is in the box

or that

Maxi's brother wants the chocolate

or that

Mommy is pretending that the banana is a telephone.

The root of the problem may be an inability to use meta-representations. If this were true, it would explain autists’ difficulty with pretend play.

Problems with Leslie’s account

The story is consistent with the findings, but isn’t it just redescribing the findings? Also if the photo task and the false belief task both share the selection processor, shouldn’t they correlate more? Is pretence a propositional attitude? Couldn’t it just mean “acting as if?”

4.     Simulation Theory

According to simulation theory, children are introspectively aware of their own mental states and can use this to infer the mental states of other people through a kind of role-taking or simulation process. E.g. in the false-belief task, children could predict what a naive other child would think the candy box contained by imagining or mentally simulating what they themselves would think if they were in his or her shoes. This whole process may be largely unconscious. It may be that all you are aware of is the prediction itself. There is no need for a folk psychological theory. Rather, you are using your own cognitive mechanism as a model for another’s. What develops is the ability to make increasingly accurate simulations. Though people also resort to theories in predicting and explaining behaviour, mental-simulation processes are important in the acquisition of social-cognitive knowledge and skills. Like theory theorists, simulation theorists also assume that experience plays an important formative role, in that it is through practice in role taking that children improve their simulation skills.

How you make a decision based on your beliefs and desires

You have a store of beliefs and desires. Some of the beliefs are derived from perception, others from inference. Some desires (e.g. desire to get a drink) arise from bodily states, others (e.g. desire to go to the kitchen) are sub-goals generated by the decision-making system. The decision-making system, which takes your beliefs and desires as input, generates sub-goals and comes up with a decision about what to do. That decision is then passed on to the mechanisms responsible for sequencing and coordinating the behaviour. 

How you predict another’s actions

The Simulation Theory says we predict behaviour by using a skill that has two components: the capacity for making decisions and the capacity to introduce pretend facts and values into one's decision-making.

Assumption 1: you can feed in pretend beliefs and desires

Suppose you could feed the decision-making system some pretend beliefs and desires. If the hypothetical beliefs and desires you've fed into your system are close to the ones that someone else has, then the decision that your system generates will be similar to the one that their system generates.

Assumption 2: you can make a decision without acting on it

It is possible to take the decision-making system off-line by disengaging the connection between the system and the action controllers. You could then generate decisions without acting on them.

How you explain another’s actions

To explain a behaviour we see if we can find some hypothetical beliefs and desires which, when fed into our decision mechanism, will produce a decision to perform the behaviour we want to explain.

To explain the Maxi task results

The child has developed a decision-making system for making on-line decisions on the basis of actual beliefs and desires. But by itself it provides the child with no way of predicting anyone else's behaviour.

You need 2 abilities to make predictions about other people's behaviour

1)      To treat decisions as predictions or expectations, rather than simply feeding them into the action controlling system.

2)      To provide the system with input other than her own actual beliefs and desires. She must be able to supply the system with pretend input so that she can predict the behaviour of someone whose beliefs and desires are different from her own. She follows Maxi as he goes outside, and thus fails to "see" what subsequently happens inside.

if the child acquires both abilities at the same time

We expect two developmental stages. In the first the child can make no predictions. In the second she can make a full range of predictions about people whose beliefs and desires are different from her own.

The child first acquires the ability to take the output off-line, and then acquires the ability to provide the system with pretend input.

We expect three developmental stages. In the first, the child can make no predictions. In the second, she can only make predictions about her own behaviour or about the behaviour of people whose beliefs and desires are identical to hers. In the third, she can make the full range of predictions.

The child first acquires the ability to provide the system with pretend inputs, and then acquires the ability to take the output off-line.

In this case, too, we would expect three developmental stages. . In the first, the child can make no predictions. In the 2^nd stage the child can play-act but not make predictions. In the third, she can make the full range of predictions.

It appears that 3-year-old children have acquired the ability to take the output off-line, but not the ability to provide the system with pretend input. They can feed pretend desires into the decision-making system, but not pretend beliefs. The theory is compatible with the observed developmental pattern.

Pretence and ToM

According to mental simulation theory a faulty capacity for pretence would degrade a person's capacity to ascribe mental states.  Autistic children suffer a deficit in the capacity for pretend-play. They fail to treat others as having points of view distinct from their own. If the off-line simulation theory is right, predicting the behaviour of people whose beliefs differ from our own requires an ability to provide our own decision making system with pretend input. This ability would also play a central role in pretend play.

Simulation and counterfactual conditionals

Conditionals concerning our own actions under hypothetical or counterfactual conditions are related to simulation, e.g. predict what actions one would take upon hearing footsteps coming from the basement.

Simulation theory and metarepresentation

The recognition and understanding of pretence might involve metarepresentation, but not necessarily the production of pretence.

Simulation and eye direction detection/Imitative Behaviour

There is subliminal muscular mimicry of the bodily postures and especially facial expressions.  Where the other's face bears an expression of emotion, adoption of a similar expression tends to produce a similar emotion in oneself.  Even when it does not produce an emotional response, it allows recognition of the other's emotion.  Human beings have an automatic tendency to direct their eyes toward the target of another’s gaze.  This mechanism turns one's own attention from the other's response to the object of the other's attention or emotion or the aim/goal of the other's action.  This emerges in the first year.  If psychological competence depends on a capacity to simulate others, these imitative mechanisms are important.

Evidence for simulation theory

Deception has no effect on autists

Autistic children fail tests of strategic deception, primarily due to their inability to disengage from focus and attention on objects. If the "theory of mind" or metarepresentation hypothesis is correct, subjects should perform better if the element of deception is removed in the windows task. If the disengagement hypothesis is correct, removal of the deceptive element should not change the difficulty of the task for autistics. The removal of an opponent did not improve autist performance.

Young children can’t take decisions off-line

If Fred asks Janet to go swimming with him, and Janet says she can’t because she’s studying, but Janet finishes early and goes to the swimming pool to meet Fred, young children think Fred’s broken a promise if he’s not there. As if saying something means you’ll do it.

Pretence is important in counterfactual conditional performance

Young children do much better at, “if he were at A, what would he do?” tasks if you ask, “pretend he’s at A. what would he do?” This doesn’t help autists.

Arguments against simulation theory

Simulation theory doesn’t explain how we develop an understanding of belief and desire

Simulation-based prediction, explanation and interpretation all seem to require that the person doing the simulating must already understand intentional notions like belief and desire.

Explanatory power

In the domains of language, the behaviour of middle-sized physical objects, judgements about mathematical problems, ability to play chess, etc., the theory-theory is the only game in town. The off-line simulation story makes no sense as an account of our ability to judge grammaticality, or of our ability to predict the behaviour of projectiles.

The following seem likely: (a) that development builds on some innate or early maturing people-reading capacities; (b) that we have some introspective ability that we can and do exploit when trying to infer the mental states of other people; (c) that our knowledge of the mind can be characterized as an informal theory; (d) that improved information-processing and other abilities (e.g. linguistic skills) enable and facilitate theory-of-mind development and help children show what they know on theory-of-mind tasks; and (e) that a variety of experiences serve to change children's conceptions of the mental world and their ability to use these conceptions in predicting and explaining their own and other people's behaviour.

Intracultural Differences

Social experiences appear to foster theory-of-mind development. Preschoolers who have more siblings to interact with perform better on false-belief tasks than those who have fewer. Deaf children whose hearing parents are not fluent in sign language perform much more poorly on a false-belief test than deaf children of fluent-signing deaf parents. The most striking intracultural differences, however, are seen in the pronounced deficits in theory-of-mind development of autistic people.

Orally taught deaf children (who have a good education but because they are not taught sign language are delayed in their language development) trail in their mastery of tasks like the false belief task by about 3 years.

They do not find some normal adults who lack a concept of false belief. However, normal adults differ from one another in their naive theories and knowledge regarding themselves and other people. For example, some people think of intelligence as a fixed, uncontrollable trait or entity, and others think of it as a malleable, controllable quality that can be improved with effort and training.

Intercultural Differences

An important review of the existing evidence - mostly from ethnographic studies - suggests that there are important differences among cultures in adult theories of mind. Important similarities also appear to exist across cultures and languages in theory-of-mind development.

Interspecies Differences

Chimp A observed Chimp B acting as though no food were available at a feeding hopper, although there really was food there. Then Chimp A appeared to depart but actually hid behind a nearby tree and watched until Chimp B took the food, whereupon Chimp A emerged from hiding and snatched it from him! Although such observations may seem persuasive, chimps may have a behaviouristic rather than mentalistic conception of seeing. Although they follow a person's gaze, they do not seem to understand that the person sees and knows about things as a consequence of directing his or her gaze at them.

[ Home ] [ Up ]