Home Introduction Cognitive Psychology Cognitive Perspective Social Perception Social Memory Social Categorization Social Judgment Language Automaticity Self Social Neuropsychology Personality Social Intelligence Development Sociology of Knowledge Social Construction Conclusion Lecture Illustrations Exam Information


Social-Cognitive Development

See also the General Psychology lectures on

"Psychological Development"


Psychology offers three principal views of development:

The Ontogenetic View

Although there are developmental psychologists who are primarily interested in development throughout the entire lifespan, from birth to death, most work in developmental psychology is focused on infants and children.  In various ways, it addresses a single question:

What is the difference between the mind of an adult and the mind of a child?

In answering this question, developmental psychology has focused on two sorts of differences:

Over its history, developmental psychology has shifted back and forth between quantitative and qualitative change.  

In addition, developmental psychology has been dominated by two views of ontogenetic development:

Preference for Faces in Neonates

The literature on newborns' preference for faces offers an example of the nativist-empiricist debate.  Do infants come out of the womb already knowing something about, and interested in, faces?  Or is their knowledge of, and interest in, faces acquired through experiences of being fed and comforted?

A series of studies by Morton, Johnson, and their colleagues (e.g., 1991) showed that newborns prefer to look at objects that resemble faces, over objects that have the same features arranged in a non-face-like configuration. 

018Faces.jpg (36897
                bytes)One JohnsonMorton1b.JPG (50707 bytes)study presented newborn infants, tested within the first hour or so after birth, with a face-like stimulus, a stimulus with the same features arranged in a non-face-like configuration, and a blank stimulus.scrambled.  The infants were pretty uninterested in the blank stimulus, but showed more interest in the face-like stimulus than in the non-face-like one.



JohnsonMorton2a.JPG (49348 bytes)Another JohnsonMorton2b.JPG (51818 bytes)study confirmed this finding: newborn infants preferred the face-like stimulus to one that had a face-like configuration but not face-like features, another consisting of a linear arrangement of facial features, and yet another that inverted the configural stimulus.



JohnsonMorton3a.JPG (54182 bytes)Interestingly, JohnsonMorton3b.JPG (57418 bytes)however, a further study of infants who were 5, 10, and 19 weeks old showed no preference for faces (compared to the configural stimulus) for the five-week-olds.  Apparently, newborns have a preference for faces that 5-week-olds have lost, and then regain by 10 weeks!



The apparent paradox was resolved by Morton and Johnson with a "two-process" theory of face recognition.

Interestingly, infants also distinguish among faces by racial features.  Scott (2012) found that 5-month-old infants matched happy sounds (like laughter) with happy faces equally well, regardless of the race of the face.  But by 9 months of age, the infants more readily associated those same happy sounds with faces representing their own race, compared to other-race faces. 


Trait Attribution

That social-cognitive processes change over time was demonstrated in an early study by Peevers and Secord (1973), who asked subjects in various age groups (kindergarten, 3rd, 7th, and 11th grades, and college students) to describe three friends and one disliked person, all of the same gender as the subject.  The spontaneous descriptions emerging from these interviews were then coded for various features:

007Peevers1.jpg (49326
          bytes)Coding for descriptiveness indicated that kindergartners rarely used trait terms, preferring to describe others in simple, superficial terms.  However, high-school and college students were more likely to make trait attributions.



                (41358 bytes)Remarkably, the descriptions had very little depth at any age.  However, high-school and college students did begin to show some appreciation that other people's characteristics were more conditional than stable and consistent.


The Piagetian Approach to Development

Over the 20th century, developmental theories gradually shifted their focus away from debates over the cognitive starting point, as in the debate between nativism and empiricism, and toward debates over the cognitive endpoint.  Thus, developmental psychology began to take on a "teleological" perspective, asking "What is the child developing toward?".

From one empiricist standpoint, for example, the notion of a cognitive endpoint doesn't make any sense: you start out knowing nothing, and you keep learning until you die.

An early example of this shift in viewpoint came with the work of Kurt Goldstein and Michael Scheerer (1941), who construed cognitive development as the development of abstract thinking.

But this shift really began with Jean Piaget, whose theory of development combined aspects of both nativism and empiricism.  Piaget held that the neonate comes into the world with an innate cognitive endowment, in terms of a set of primitive cognitive schemata.  Just as a genotype develops into a phenotype through interaction with the environment, Piaget argued that development occurs as a dialogue between these schemata and the environment:

Piaget further held that the genetic program for change was marked by discontinuities between four stages of cognitive development:
The final accomplishment of development, from the Piagetian point of view, was the ability to employ abstract concepts and formal logic.

Further, Piaget argued that the child's developmental progress was marked by a series of milestones.  Most of these pertained to "nonsocial" aspects of cognition:

          (65745 bytes)In the domain of social cognition, the principal Piagetian milestone is the loss of egocentrism, or the child's "belief that other people experience the world as the child does".  Egocentrism is commonly measured by the "three mountains" point of view task (Piaget & Inhelder, 1967).  Before age 4, children will say that the doll can see the church.  After age 7, children understand that the doll cannot see the church, because the mountain is in the way. 


Brodzinsky1980.JPG (47042 bytes)The loss of egocentrism marks the transition between the pre-operational period and concrete operations, and typically occurs sometime between ages 4 and 7.  Egocentrism is related to conservation, another milestone of this transition, in that children who have lost egocentrism understand that the same object can look different, depending on their point of view.  But in the context of social cognition, the loss of egocentrism reflects the child's ability to appreciate what is on another person's mind.

The "Theory of Mind"

Egocentrism is related to the idea that cognitive development consists in the development of a theory of mind.  

In the course of a debate over nonhuman cognitive abilities, Premack and Woodruff (1978) suggested that humans (and perhaps other primates, especially great apes) had a theory of mind -- the understanding that we have mental states of belief, feeling, and desires, that other people also have mental states, and that their mental states may differ from our own. 

With respect to psychology, the basic elements of a theory of mind are as follows:

Viewed in this way, the theory of mind is not just an aspect of cognitive development.  It is very much the development of social cognition.

021FalseBelief.jpg (130591 bytes)The concept of theory of mind was quickly imported into human developmental psychology, as exemplified by interest in the false belief task, a variant on Piaget's "three mountain" test of egocentrism.

JacobEsau.JPG (91322
            bytes)The false-belief task is new to psychology, but the basic idea is as old as the Hebrew Bible.  Genesis 27 tells the story of Jacob and Esau.  Rebekah bore Isaac twin sons: Esau, who was hairy, was delivered before Jacob, who was smooth-skinned; therefore Esau was, technically, the oldest son.  However, Jacob cheated Esau out of his birthright (Genesis 25).  As Isaac approached death (at this point he was 123 years old, and blind), he decided to bestow a blessing on his firstborn.  However, Rebekah had heard a prophecy that Jacob would lead his people, and "the older shall bow down to the younger".  Isaac sent Esau out to hunt game for a l meal.  Meanwhile Rebekah prepared a meal, and instructed Jacob to present it to Isaac as if he were Esau.  When Jacob protested that Isaac would not be fooled, because he had smooth skin.  Rebekah covered his arms and back with goatskin. She knew that Isaac would be fooled into thinking that Jacob was really Esau.  The ploy worked, Esau was cheated again, and -- to make a long story short -- the children of Jacob went on to become the Twelve Tribes of Israel.

A somewhat more recent example comes from the great Act II finale in Mozart's opera, the Marriage of Figaro (1786). Figaro is valet to Count Almaviva; he is about to marry Susanna, personal maid to the Countess. They learn that the Count intends to exercise his droit du siegneur on their wedding night. They plot with the Countess to fool the Count into thinking she is going to have an assignation with an Susanna, but instead he will encounter Cherubino, a page who is in love with the Countess. They also get Figaro to write a letter to the Count, informing him that the countess is having an assignation with an unknown man. While Susanna and Cherubino are exchanging clothes in the Countess's rooms, the Count appears unexpectedly. They hide Cherubino in a locked closet. The Count leaves to find tools to break down the closet door, taking the Countess with him. Meanwhile, Cherubino escapes through a window and Susanna takes his place. When the Count and Countess return, the Countess tries to explain why he's about to find Cherubino hiding in her closet. But when the Count opens the door, he finds Susanna instead. Susanna and the Countess also tell the Count that Figaro wrote the anonymous letter. Chagrined, the Count decides to forgive everyone. At this point, Figaro reappears. The Count, knowing everything, questions Figaro -- who, despite increasingly desperate hints from the Countess and Susanna, denies everything. And on it goes, for 20 minutes of duets, trios, quartets, quintets, and even a septet: all of it based on the idea that someone knows what someone else doesn't know.


022FalseBelief2.jpg (51979 bytes)A large number of studies (178, actually) reviewed by Wellman et al. (2001) show clearly that children aged 3 or  younger typically fail the false belief task, while those aged 5 and older typically pass.  The older children apparently possess an ability the younger ones lack -- the ability to infer the contents of someone else's mind.

023FalseBelief3.jpg (42579 bytes)Interestingly, young children apparently have the same problem reading their own minds.  In the representational change task, a variant on the false-belief task, children are shown that a crayon box actually contains candies.  When asked what they thought the box contained before they looked inside, children younger than 3 tend to say that they thought it contained candies; children older than five say they thought it contained crayons.  Age trends for performance on the representational change task precisely parallel those obtained for the standard false-belief task.

                (92563 bytes)It's possible to push performance on the false-belief test by various manipulations, but in general, 3-year-olds seem to fail, while 5-year-olds pass. 

The notion of a "theory of mind" was quickly imported into the literature on developmental psychology. Early on, the theory of mind displaced formal operations as the new "endpoint" of cognitive development -- the goal toward which development was proceeding.  The "theory of mind" was gradually expanded into a theory theory of cognitive development offered by Alison Gopnik, Henry Wellman, Andrew Meltzoff, and others.  According to the "theory theory":

Subjectivity and Intersubjectivity

Egocentrism and the "theory of mind" suggest that, from the point of view of social cognition, the endpoint of development is the emergence of intersubjectivity.

Intersubjectivity, in turn, entails two levels of intentionality.  But in order to understand this we must first distinguish between two quite different meanings of the term.
Thus, intersubjecivity entails higher-order intentionality.  And it doesn't stop at the second order.  Consider:
  1. I feel happy.  This is a statement about my own subjective mental state.
  2. Lucy knows that I feel happy.  This is a statement about Lucy's mental state, which in turn refers to my mental state.
  3. Judy believes that Lucy knows that I feel happy.  This is a statement about Judy's belief about Lucy's belief about my mental state.
  4. Stan forgot that Judy believed that Lucy knew that I feel happy.  This is a statement about Stan's belief about Judy's belief about Lucy's belief about my mental state.

Like William James' story of the turtles, it's intentionality all the way down.

Intersubjectivity in Literature

We've already seen one literary example of intersubjectivity: the story of Jacob and Esau, which revolves around Rebecca's (and Jacob's) belief about what Isaac will believe.  Rebekah tells Jacob what Isaac will believe.  But there are lots more, once you know to look for them.

OdysseySheep.JPG (87854 bytes)Consider the 'Polyphemus" episode from Homer's Odyssey (Book IX).  Odysseus and his men take shelter on the island of the Cyclopes, in the cave where Polyphemus lives with his sheep, and start eating his food.  Polyphemus comes home, imprisons them, and begins to eat Odysseus' men.  Odysseus gets Polyphemus drunk, and while he is asleep, the surviving men blind him with a sharpened tree trunk.  But they're still prisoners.  How to get of the cave?  When Polyphemus lets the sheep out the next morning to graze, he feels the back of each sheep to make sure that none of Odysseus' men are escaping.  But Odysseus has instructed his men to suspend themselves underneath the sheep's bellies.  Odysseus knows what Polyphemus will believe.  

Earlier, when Odysseus had demanded that he receive proper treatment as Polyphemus' guest (hospitality is a major theme in the Odyssey) Polyphemus had promised that he would eat Odysseus last.  When Polyphemus, in turn, demanded the name of his "guest", Odysseus replied "Noman" (actually, of course, the Greek equivalent).  Polyphemus then promised that "I will eat Noman last!".  Big joke.  But after Odysseus and his men had made their escape, Polyphemus called for assistance.  When his fellow Cyclopes asked who was harming him, he cried "Noman is harming me".  So nobody came to help.  Odysseus knew what the other Cyclopes would infer from Polyphemus' response.

Othello.JPG (78540 bytes)Here's another one, in Shakespeare's Othello (from Robin Dunbar, 2004).  In the play, Iago plants Desdemona's handkerchief so that Othello will believe that she is in love with Cassio.  That's two levels of intersubjectivity, three levels of intentionality, right there:

  1. Iago intended that 
  2. Othello would believe that
  3. Desdemona loved Cassio. 

But then, when you think about it, there are actually two more levels.  In order for the play to work,

  1. Shakespeare desired that
  2. the audience would understand that
  3. Iago intended that 
  4. Othello would believe that 
  5. Desdemona loved Cassio.  

Persuasion.JPG (107202 bytes)The cognitive-science-y literary scholar Lisa Zunshine has argued that multi-level intersubjectivity plays a big role in the English novel, beginning with the work of Jane Austen.  In Persuasion (1818), Anne Elliot is shopping in the town bakery with her sister Elizabeth, when her former fiance, Wentworth (Anne had been persuaded to break their engagement) walks in.  Anne recognizes Wentworth right away.  So does Elizabeth, but she pretends not to notice him, and Wentworth pretends to ignore the fact that he's not been acknowledged.

  1. Anne realizes that
  2. Wentworth knows that
  3. Elizabeth has recognized him.

According to Zunshine (2007), this complex, recursive intentionality was essentially Austen's invention.  Certainly it only becomes a major feature of English literature with her work.  In earlier English novels, such as Henry Fielding's Tom Jones (1749), there is much less portrayal of the characters' mental states.

The Economist, in an unsigned article (07/15/2017) commemorating the 200th anniversary of Austen's death, wrote that:

her uniqueness lay in combining... realism with a ew narrative style, one which moved deftly between the narrator's voice and the characters' innermost thoughts.  This "free indirect speech" allowed the reader to see, think and feel exactly as the character did while also maintaining  critical distance and the ability to move between various points of view.  It was radically inventive.

Going from the sublime to the less-sublime, consider this famous episode of the TV sitcom, Friends:

  1. Monica and Chandler have begun to date, but they want to keep it a secret.
  2. Phoebe and Rachel learn about this, and they decide to play a joke on them.
  3. Monica and Chandler discover what Phoebe and Rachel are up to, and decide to play a joke on them, instead.
  4. Phoebe finds out about Monica and Chandler's plans, and revises her original plan. As she says to Rachel, "They don't know that we know they know we know".
Actually, I'm one of the few people in the universe who never watched Friends.  I only learned about this from an article about Zunshine: "Next Big Thing in English: Knowing They Know That You Know" by Patricia Cohen (New York Times, 04/01/2010), from which some of this material is drawn

Back to Zunshine: She infers from psychological research that people can keep track of three levels of intentionality (X knows what Y knows about Z).
  • In a literary technique known as free indirect style, the author's voice is intermixed with that of one or more characters, so that readers, creating two or three levels of intentionality.
  • Many modern authors -- Like Henry James or Virginia Woolf -- require readers to keep track of as many as six levels at once.

Zunshine (2011) has discussed the varying levels of intentionality in her concept of mental embedment -- by which she means the extent to which a narrative contains references to mental states.  To take her examples:

  • The statement My last name begins with a Z, while Alan's last name begins with a P has zero embedment, because it contains no reference to mental states.
  • The statement I don't want to read The Da Vinci Code contains refers to 1 mental state, and therefore has 1 embedment.
  • The statement I used to think that I would hate The Da Vinci Code refers to two mental states, and therefore has 2 embedments
    • Note, however, that both mental states belong to the same person.
  • The statement I would have hoped that Alan could surmise that I wouldn't want to read The Da Vinci Code has 3 embedments.

Zunshine refers to the pattern of embedment as sociocognitive complexity, and argues that the "baseline" level of complexity for fiction is the third level, "a mind within a mind within a mind" -- the mind of the Reader, the mind of the Narrator, and the mind of the Character.

Moving from the sublime to the ridiculous, an interesting example of higher-order intersubjectivity in a dyad is found in the "battle of wits" scene from Rob Reiner's film, The Princess Bride (1987). 

For more on Zunshine's work, see her scholarly books:
  • Why We Read Fiction: Theory of Mind and the Novel (2006)
  • Getting Inside Your Head: What Cognitive Science Can Tell Us  About Popular Culture (2012).

Zunshine is not the only literature scholar to make use of psychology and cognitive science.

There is, of course, a long history of humanists drawing on Freudian psychoanalysis to interpret literature -- essentially, viewing characters and plots as Freudian case studies.  an excellent example is the Anxiety of Influence by Harold Bloom, a professor of English at Yale who argued that the relationships between poets can be understood as the working out of the Oedipus complex.  writers This form of literary scholarship is still popular.  However, while psychoanalysis might be useful for understanding literary (and other artistic works) influenced by psychoanalytic thinking, in general it seems like a bad idea to interpret literature in general based on a psychological theory that has been wholly discredited by modern science.  One of the earliest proponents of using psychoanalytic theory in literary criticism was UCB's Frederick Crews; later, he saw the light, and became a vigorous critic of psychoanalysis, both as a vehicle for literary criticism and for psychotherapy.  See, for example, Out of My System (1975); Skeptical Engagements (1986); and The Memory Wars: Freud's legacy in Dispute (1995).  Crews also published two satirical books, in which he illustrated various modes of literary scholarship as they might be applied to the "Winnie the Pooh" books:  The Pooh Perplex (1963) and Postmodern Pooh (2001); both make wonderful reading for readers who want to understand literary criticism better -- provided, of course, that you've read A.A. Milne's The Wind in the Willows.first!

And then there is literary darwinism, which recapitulates the mistakes of the Freudians by reading fiction, poetry, and drama through the lens of Darwin's theory of evolution by natural selection.  The only difference is that Darwin's theory is valid and Freud's is not.  But really, do we really need Darwin to understand that Pride and Prejudice is about the importance of making a good match?  Nah.  For a critical account of literary Darwinism, see "Adaptation", a reviwe of several books in this genre, by William Deresiewicz, a literary critic, in The Nation, 06/08/2009)

Interestingly, you can have higher levels of intersubjectivity without involving more than two people, because one person can think about what a second person thinks about what the third person thinks.  And here, too, it's "turtles all the way down".


Just as the false-belief task has served as the "gold standard" for the acquisition of a theory of mind, the false-belief task can be expanded indefinitely.  For example, Perner and Wimmer (1985) distinguished between first-order and second-order false beliefs.

Their experiment employed the "Ice-Cream Truck Story", in which John has to figure out where Mary would go to get ice cream, and also to justify their answer.

IceCreamTruck1.JPG (102358 bytes) IceCreamTruck2.JPG (85432 bytes)

2oFB.JPG (45406
                    bytes)While 4- and 5-year-old children typically pass the 1st-order FB task, the majority of children don't pass the 2nd-order FB task until age 9 or 10, unless they are given special memory aids.



Of course, processing multiple levels of intersubjectivity makes great demands on cognitive resources.  While in principle intersubjectivity can be infinitely recursive, there are cognitive constraints on what we can process.

Kinderman1998a.JPG (74798 bytes)For Kindermann1998b.JPG (48684 bytes)example, Kinderman et al. (1998) tested adult subjects on their understanding of stories that had up to five levels of intersubjectivity.  As a check, they employed stories that had multiple levels of causality, but no intentionality.  People could process four levels of intentionality (i.e., three levels of intersubjectivity) pretty well, but just fell apart at the fifth (fourth) level.  But they were able to process up to six levels of causality perfectly well.  So the problem is not in level, in the abstract, but specifically in level of intentionality or level of intersubjectivity.  Given George Miller's "Magical Number Seven, Plus or Minus Two", we might think that we could handle about seven levels of intentionality, but we can't; intersubjectivity itself requires cognitive resources, and makes the limits on cognitive processing even more severe.

Theory of Mind Beyond the False-Belief Task

WellmanLiu1994.JPG (60558 bytes)Just as first-order theory of mind is not the be-all and end-all of theory of mind, neither is the false-belief task.  Wellman and Liu (1994) have offered a larger catalog of theory of mind abilities, including:



Diverse Desires: That others might have desires that differ from their own. WellLiuDD.JPG (57849 bytes)
Diverse Beliefs: That others might have beliefs that differ from their own. WellLiuDB.JPG (52984 bytes)
Knowledge Access: That one has access to knowledge that others might lack. WellLiuKA.JPG (60452 bytes)
False Belief: That others might have beliefs that differ from one's own, and that these beliefs might be false (this is tapped by the standard "false belief" test). WellLiuCFB.JPG (67325 bytes)
Explicit False Belief: That one's own beliefs might, indeed, be false. WellLiuEFB.JPG (47507 bytes)
The Relation between Beliefs and Emotion: That our emotional state depends on whether our beliefs prove to be true or false. WellLiuBE.JPG (67166 bytes)
The Difference between Real and Apparent Emotion: That, for whatever reason, people might want to hide their true emotional states, so that what appears to be their emotional state might be different from that they actually feel. WellLiuRAE.JPG (62073 bytes)

WellLiuFig.JPG (52444 bytes)Testing a group of 3-5-year-olds, they found that some of these ToM tests were generally easier than others. 

Wellman and Liu (1994) suggested that their test battery constituted a Guttman scale of theory of mind abilities.  A Guttman scale (named for Lewis Guttman, an American statistician) is one in which scale items are arranged in increasing order of difficulty, such that an individual who passes one item can be assumed to pass all earlier items.  Guttman scales promote efficiency in ability testing and attitude measurement, because we can begin in the middle of the scale: if the respondent passes that item, it can be assumed that s/he would have passed all earlier items as well.  The SAT and GRE, as administered by computer, employ a version of Guttman scaling to reduce the time required to administer the test.

Pushing the False-Belief Envelope

So it seems that 4-year-olds have a theory of mind, as assessed by the FB test, but three-year olds don't.  But then again, the FB test is very verbal in nature, and it might be that younger children would pass the test if it were administered in nonverbal form.

After all, as Gopnik and Wellman (1992, p. 150) long ago concluded:

The 2-year-old is clearly a mentalist and not a behaviorist.  Indeed, it seems unlikely to us that there is ever a time when normal children are behaviorists....  It seems plausible that mentalism is the starting state of psychological knowledge.  But such primary mentalism , whenever it first appears, does not include all the sorts of mental states that we as adults recognize.  More specifically, even at two years psychological knowledge seems to be structured largely in terms of two types of internal states, desires, on the one hand and perceptions, on the other.  However, this knowledge excludes any understanding of representation.

Do 2-year-olds really have some concept of desire?  A classic experiment suggests that they do. 

Pushing the limits further, Onishi and Baillergeon (2005) devised a totally nonverbal version of the FB task, relying on the general finding that even infants look longer at events that violate their expectations.

First, the infants were given three familiarization trials.

On Trial 1, they saw an actor hide a (plastic) slice of watermelon in a green box.

On Trials 2 and 3 the actor returned and reached into the green box for the watermelon.

Onishi1.JPG (57361 bytes)
Then the infants were divided into four groups for the belief-induction trial.  The infants, saw, for just one trial:

In the True Belief Green condition, the actor watched as the yellow box moved toward the green box.

In the True Belief Yellow condition, the actor watched as the watermelon moved from the green box to the yellow box.

In the False Belief Green condition, the actor was no longer present when the watermelon moved to the yellow box.

In the False Belief Yellow condition, the actor watched as the watermelon moved from the green box to the yellow one; but was no longer present when the watermelon moved back to the green box.

Onishi2.JPG (104483 bytes)
On the Test Trial, the infants watched as the actor opened the door and reached into the green box or the yellow box.   Onishi3.JPG (58711 bytes)

                    (60495 bytes) And, in fact, the infants looked longer on trials where an actor behaved in a way that seemed to contradict her (the actor's) understanding of where a plastic watermelon slice had been hidden.  Apparently, even infants have some sense of what others believe, that their beliefs might be different from their own, and that their beliefs might be incorrect.  If infants expect others to behave in accordance with their beliefs, and are surprised (and pay extra attention) when they do not do so, then it can be said that even infants, long before age 4, have a rudimentary theory of mind.


Actually, it's not always the case that infants look longer at events that violate their expectations.  In some case, they look for less time at counter-expectational events.  Nobody quite knows why.  But in either case, a difference in looking times indicates that, from the infant's point of view, something has gone wrong.  And that's all the logic of the experiment requires.

The Theory of Mind as an Actual Theory

OK, so little kids are mentalists, not behaviorists.  But does their performance on these experimental tasks really mean really that they have a theory of the mind?  Maybe they just have a concept, or an idea, about the mind and mental life -- a concept that matches adult folk-psychology.  Gopnik and Wellman (1992) argue that the "theory" part of the theory of mind should be taken seriously. 

OK, but do kids actually do this?  Again, Gopnik and Wellman (1992) argue that they do.

In these and other ways, the child's theory of mind is not just an idea about the mind, but a real theory about the nature of mental life.


The Theory of Mind in the Brain

                    (81787 bytes)An adult fMRI study suggests that theory-of-mind processing may be performed by a particular area in the brain.  In this study, Saxe and Kanwisher (2003) had adult subject read stories in which they had to reason about either true and false beliefs, mechanical processes, or human actions (without beliefs).

These investigators found differential activation at the junction of the temporal and parietal lobes (also known as the temporo-parietal junction, or TPJ), which they tentatively named the temporo-parietal junction mind area.  They also identified another area, in extra-striate cortex, which is activated when people think about body parts: they named this the extra-striate body area.

The Theory of Mind as Mindreading

Simon Baron-Cohen (a pediatric psychiatrist and yes, cousin of the comedian Sacha) and others have argued that there are at least four elements to the theory of mind:

An  intentionality detector (ID), in which events are interpreted in terms of goals and desires.  This is, of course, the "English" sense of intentionality.  The intentionality detector is nicely revealed by the studies of Michotte and Heider and Simmel on phenomenal causality, in which goals -- " intentions" in the English sense of the word -- are attributed to inanimate objects.

                                      (76418 bytes)

An  eye-direction detector (EDD), which detects both the presence of eyes and computes the direction of another person's gaze as "at me" or "not at me".  The EDD interprets another person's gaze as "seeing", and makes the inference that the other person is aware of what its eyes are looking at.  
A shared-attention mechanism (SAM), by which the subject assumes a relation between knowledge and seeing.  The SAM goes beyond the EDD to assume a triadic relation between the Self, the Agent, and the Object -- i.e., that if the Self looks where the Agent is looking they'll see the same thing.   SAM.JPG (87107 bytes)
A theory of mind mechanism (ToMM), by which one person can infer another person's mental states.  

Baron-Cohen calls the theory of mind a capacity for mindreading.  He further assumes that each aspect of mindreading is mediated by a separate cognitive module, and -- by extension -- a separate brain system.  

032ElementsChild.jpg (37555 bytes)Studies of biologically normal children summarized by Baron-Cohen (1995) indicate that most infants possess both ID and EDD by 9 months of age; they acquire SAM sometime between 9 and 18 months; and they acquire the ToMM sometime in their 4th year of age (as demonstrated by performance on the standard false-belief test).



Shared Attention in Infants

For example, evidence of a shared-attention mechanism comes from infants' protodeclarative pointing gesture: an infant will point his fingers at some point in space, and then check the other person's gaze to make sure that she is looking there.  

Scaife and Bruner (1975) argued that the mother will orient to the infant's gaze, and, within limits, the baby will orient to the mother's gaze.  Based on Piagetian developmental theory, they argued that there were limits on egocentricity even before the child reached the stage of "concrete operations".  Shared attention in infants shows that, even before "concrete operations", they appreciate that someone might be looking at something that they themselves cannot see.

SAM_infants.JPG (72804 bytes)Formal evidence for a shared-attention mechanism comes from a classic study by Butterworth & Cochran (1980), which found that infants and mothers would coordinate their gazes. 

The principal finding was that the babies would almost always shift their gave when the mother looked at a point in the babies' field of vision, but not the mothers'; and also very frequently when the mother looked at a point that was already in their joint visual field.  But they rarely shifted their gaze when their mothers looked at a field that was in the mothers' field of vision, but not the babies'.

Perceiving Intention

However, Baron-Cohen's evidence for an intentionality detector in infants is somewhat indirect and anecdotal.  

033Kuhlmeier1.jpg (76981 bytes) In a formal experiment,, 034Kuhlmeier2.jpg (66125 bytes) Kuhlmeier, Wynn, and Bloom (2003) produced better evidence for an infant ID in an experiment that employed a variant on the animation techniques introduced by Michotte (1946) and Heider & Simmel (1944).  In this experiment, 5- and 12-month-old infants first habituated to a film depicting a triangle "helping", or a square "hindering", a ball's progress up a hill.  In the test film, the infants were presented with a film depicting either the ball approaching the square (a plot departure) or, alternatively, the ball approaching the triangle (a plot continuation). 

035Kuhlmeier3.jpg (40175 bytes)The 12-month-olds discriminated between the two test films, while the 5-month-olds did not.  But note the nature of the discrimination.  In this case, the 12-month-olds looked longer at the plot continuation than they did at the plot departure.  This isn't exactly what we'd expect, given the idea, discussed above, that infants pay more attention, and thus look longer, at counter-expectational events.  You were warned that this doesn't always happen, and in fact it's possible that the relationship between expectancy and looking time is U-shaped.  That is, infants will look at counter-expectational events for either a relatively long, or a relatively short, time.  But the important result in the study was that the 12-month-olds did, in fact, discriminate between the two test films -- which suggests that they had some idea about the intentionality behind the two scenarios.

Subsequent studies in the Wynn-Bloom, using variants on this methodology, have pushed back the acquisition of an intentionality.  In the Kuhlmeier et al. (2003) study, the 5-month-old infants showed no difference in looking times between the "plot departure" and "plot continuation" videos.  But in an experiment by Hamlin, Wynn, and Bloom (2007), 5-month-olds reached for puppets representing the "helping" and "hindering" characters, and 3-month-olds spent more time looking at the "helper" than the "hinderer" (2010).  So, even the 3-month-olds seemed to have the concept of desire to help and desire to hinder -- just the sort of thing that would be a product of an intentionality detector.

Mindreading -- and Mindblindness -- in Autism

Baron-Cohen has also suggested that childhood autism is a case of mindblindness.  He attributes the severe deficits in social behavior, communication, and imagination as specific impairments in social cognition, reflecting the child's lack of a theory of mind.

Childhood autism was first described by Leo Kanner in 1943, and the case studies that he presented certainly seem to depict a child who does not understand (or, at least, care) that other people are sentient beings with thoughts, feelings, and desires (or, at least, thoughts, feelings and desires that are different from his or her own). 

Kanner illustrated his observations with a number of case studies of autistic children.

Kanner1.JPG (86286 bytes) Kanner2.JPG (85074 bytes) Kanner3.JPG (103619 bytes)

At about the same, but independently, Hans Asperger, a Swiss psychiatrist, made very similar observations, illustrated with case studies, which he labeled "autistic psychopathy".  The two descriptions, Kanner's and Asperger's, were remarkably similar.  Put bluntly, Asperger's syndrome is autism with language.  And for that reason, it is typically less severe and more manageable.

These days, autistic syndrome is classified as a pervasive developmental disorder, along with Asperger's syndrome, first diagnosed in 1944, and a few other sub-types.  According to the 4th edition of the Diagnostic and Statistical Manual for Mental Disorders (DSM-IV), published in 1994, autism was described as a disorder, emerging relatively early in development, affecting social functions, communication functions, and imagination.

DSM-IV distinguished between clasic autistic disorder (sometimes called "Kanner syndrome"), and "Asperger's syndrome".

DiagAut1.JPG (80865 bytes) DiagAut2.JPG (74640 bytes) DiagAut3.JPG (94276 bytes)
DiagAut4.JPG (81285 bytes) DiagAut5.JPG (57453 bytes) DiagAut6.JPG (88104 bytes)

The fifth edition of DSM (DSM-5),released in 2013, combines classic autism and Asperger's syndrome under a single category of autism spectrum disorder, with the implication that Asperger's syndrome is just a a milder form of autism. But if, indeed, Asperger's syndrome is autism with language, it's not quite right to put the two syndromes on a simple continuum of severity.  There may be qualitative differences between them.  If so, perhaps these qualitative differences will be recognized in DSM-6!

  • DSM-IV based the diagnosis of autism on a package three symptoms:impairment in social interaction, impairment in social communication (language); and restricted, repetitive, and stereotyped patterns of behavior, interests, and activities.
  • DSM-5 collapses the two kinds of social impairment together, as if language and nonverbal communication were somehow equivalent. Moreover, it fails to distinguish between individuals who are not interested in social interaction (classic autism), and those who are interested but lack the requisite skills.

Advocacy groups for Asperger's syndrome have complained that such a move would stigmatize individuals who are just "different" from "neurotypicals", and shouldn't be treated as if they are mentally ill at all. For a vigorous critique of DSM-5's treatment of autism and Asperger's syndrome, see The Autistic Brain: Thinking Across the Spectrum by Temple Grandin and Richard Panek (2013).

In these lectures, I am sticking with the DSM-IV classification.

Autism in Literature

One of the features of the modern novel is the view it gives of its characters' interior lives. But lately, several novelists have tried to write novels in which the central character is autistic -- someone who, given the clinical definition of autism, doesn't have much interior life. Eli Gottleib, who has written two novels inspired by his autistic brother -- one in the third person, the other in the first person -- reflected on this problem in an essay entitled "Giving a Voice to Autism" (Wall Street Journal, 01/05-06/2013). Gottlieb notes that the "Ur-text" for the "developmentally disabled" perspective in fiction is William Faulkner's The Sound and the Fury (1929), which is written from the perspective of Benjy, a developmentally disabled adult -- not autistic, but perhaps mentally retarded or just plain brain-damaged.

...Faulkner produces his effects by shattering the normally smooth perceptual continuum into discrete chunks: "Through the fence, between the curling flower spaces, I could see them hitting. They were coming towards where the flag was and I went along the fence. Luster was running in the grass by the flower tree. They took the flag out and they were hitting".

The repetitive staccato sentences highlight Benjy's limited mental means, while the tone skates close to that literary cousin of the dysfunctional adult narrator: the child-voice. Books told from children's points of view often employ the same limited vocabulary, magical thinking, and emotional foreshortening as those of the developmentally disabled.

In the Curious Incident of the Dog in the Night Time, author Mark Haddon borrows from both camps and ventriloguizes the voice of a 15-year-old autistic [sic]:  "Siobhan has long blond hair and wears glasses which are made of green plastic. And Mr. Jeavons smells of soap and wears brown shoes that have approximately 60 tiny circular holes in each of them.

In my own case, I'd originally cast the new book in the third person, but I found that I needed the ground-hugging intimacy available only with first-person narration. This, however, required a crucial adjustment away from the "literary" prose which is my default mode. a third-person sentence like this: "Sometimes his handlers would escort him to church, where the soaring, darkwood vault transfixed him, and the rich voice of the preacher as he spoke of hellfire an damnation moved the hair on the back of his neck," would end up transposed to this: "Sometimes she takes me to her mega-church where the Lord is so condensed that people faint and shout out loud at how much of the Lord there is. The preacher has a rich yelling voice and when the chorus sings it's like the bang of thunder that comes mixed with lightning."

Maybe physicists are right, after all, that the best thinking happens in childhood. My challenge in the new novel is to recover that buried perceptual-cognitive mix and haul it into the fictional light of day.

Autism in Adulthood

Despite the fact that Kanner and Asperger were seeing autistic children in the 1940s, and Rimland wrote a classic text on the syndrome in the 1960s, the literature on autism focuses almost exclusively on children, and we know relatively little about what happens when autistic children grow up.  Autism promises to be a serious problem for mental health policy, as some provisions are going to be have to be made to take care of these individuals after their parents (and other family members) are no longer able to do so.  The seriousness of the problem can be seen in the simple fact that diagnoses on the autism spectrum have increased exponentially over the past couple of decades.

The problem, and the possibilities of a good transition to adulthood in autism, were highlighted in a series of articles which appeared in the New York Times in 2011:

Interestingly, some autistic patients achieve what might be called an "optimal outcome", in that they become able to manage their social relations and navigate the social world (Fein et al., 2013).  These adults generally had a milder form of autism when children. 

The most prominent example of a person with autism who has made a successful transition to adulthood is Temple Grandin, a professor of animal science at Colorado State University, who was the subject of a made-for-TV movie starring Claire Danes (of My So-Called Life and Homeland fame).  See especially Grandin's recent book, The Autistic Brain: Thinking Across the Spectrum (2013).

Face Perception in Autism

043AutismFace.jpg (90269 bytes)A study by Schultz, Gerlotti, & Probar (2001) found that autistic patients showed less activation of the "fusiform face area" when asked to discriminate between pairs of faces, compared to normal subjects.  Of course, as we now know, the fusiform area is also recruited by tasks that involve perceptual recognition at subordinate levels of categorization, regardless of whether the stimuli are faces.  


The Theory of Mind in Autism

More to the point, autistic children appear to perform more poorly than normal children on the standard false-beliefs task.

046FalseAutism.jpg (49231 bytes)A study by Peterson and Siegel (1999) showed that autistic children aged 9-10 years performed more poorly on the false belief task than normal children aged 4-5 years.  A formal meta-analysis by Happe (1995) showed that, compared to normal children, "high functioning" autistic children are delayed by about 5 years in passing the standard false-belief test. The others may never pass it at all.



PeteSieg99b.JPG (54001 bytes)However, the lack of a ToMM may be a result, not a cause of autism.  Autistic individuals have limited contact with other people, but so do the deaf children of hearing parents, who acquire sign-language late, and therefore have had somewhat impoverished interpersonal relationships.  These children, too, show problems on the false-belief task, indicating that they, too, have problems with the ToMM.


PWL2005.JPG (60375 bytes)A later study by Peterson, Wellman, and Liu (2005), employing a larger set of ToM tasks, confirmed these results, but also revealed some surprises.  In this study, autistic and deaf children aged 5-14 years old were compared to typical preschoolers, aged 4-6 years old.  The deaf children included both native users of sign language, and children who learned to sign relatively late in childhood. 

This study indicates that deaf children show the normal pattern of acquisition of ToM -- although the process is delayed in late signers.  Surprisingly, the autistic children showed a late reversal -- they can understand hidden emotions, a developmentally advanced ToM task, even though they can't understand the developmentally prior false beliefs.  Peterson et al. suggest that the autistic children have "hacked out" a rote solution to emotional situations -- they don't really understand the distinction between real and apparent emotions, but they've learned that people sometimes hide their true feelings.

048ElementsDiag.jpg (43616
                                    bytes)Baron-Cohen (1995), summarizing the available literature, concluded that even mentally retarded children possess ToMM.  By contrast, one group of autistic children appears to lack the ToMM, while another group of autistic children appears to lack the SAM as well.


This general conclusion has held up, although there are not a few caveats to consider (e.g., Tager-Flusberg, 2007).

Autism Beyond Mindblindness

As a matter of historical record, it should be noted that, while Baron-Cohen usually (and rightly) gets credit for the mindblindness hypothesis of autism, the general idea was floating in the air before he formulated it in quite this way.  B-C did his graduate studies in the Developmental Psychology Unit at University College London, working with a large group of developmental psychologists including Uta Frith and Alan Leslie. Frith, in fact, was one of the first developmental psychologists to take an interest in autism (for an autobiographical account of her research, see Frith, 2012), following publication of Bernard Rimland's classic book on Infantile Autism (1964).  Following a pattern established earlier for the study of schizophrenia, as discussed in the General Psychologicy lectures on "Psychopathology and Psychotherapy", Frith and her colleagues embarked on a series of studies comparing various aspects of human information processing in autistic and normal children, using paradigms derived from the experimental study of cognition.  These studies quickly focused on autistic children's deficits in social skills -- as Frith put is, "how it was that procesing information about the physical world could be so good, while processing information about [the]social world could be so poor" (2012, p. 2077).  Alan Leslie suggested that the problem in autism was a failure to create "metarepresentations" of physical reality -- that is, what philosophers call intentional states of believing, feeling, and wanting; or what the psychologist John Morton called "mentalizing".  From there it was only a very short step to B-C's idea that autistic children lack a theory of mind -- they lack the understanding that mental states are representations of reality.

Frith, for her part, is convinced that mindblindness is only part of the problem in autism, because -- as discussed by Tager-Flusberg (2007) as well -- autistic children have problems in the nonsocial domain as well.  She argues that they also suffer from weak central coherence (WCC): "while the ability to discern a wide variety of things about the world around is strong, the drive to make these various things cohere is weak" (2012, p. 2080).  While B-C's mindblindness hypothesis is grounded in the neuroscientific Doctrine of Modularity, as discussed in the lectures on "Social-Cognitive Neuroscience", WCC  supposes that the mind is not entirely modular, and that coherence is a property of the mind as a general-purpose information-processing system.


What Causes the Theory-of-Mind Impairment in Autism?

Assuming that autistic children really do lack a theory of mind, there are at least three plausible causal explanations for this (lack of) correlation:

Broken Mirror Neurons in Autism?

With respect to a pure neurological account, one possibility is that autistic children and adults suffer from a malfunctioning mirror neuron system.  Recall that mirror neurons, as the neural basis for social cognition, were discussed in the lectures on "Social-Cognitive Neuroscience".  According to a popular theory, mirror-neuron systems in the frontal and parietal cortex, and in the limbic system, allow us to understand, and empathize with, the actions and emotional expressions of other people.  If so, then it might be that autism, in which the patients apparently cannot understand and empathize with the actions and emotions of others, cannot do so because of a malfunctioning mirror-neuron system.

Just such a hypothesis has been proposed by Ramachandran and Oberman (2006) in their "broken mirrors" theory of autism.  R&O accept the view of autism as a form of , but argue that this theory, written at the psychological level of analysis, simply restates and summarizes the symptoms; a proper theory, in their view, would have something to say about the biological mechanism that causes mindblindness.  They suggest that this biological mechanism can be thought of as a "broken" mirror neuron system. 

As evidence for their theory, R&O report a study of "Mu"-wave blocking in autistic and normal children.  The Mu wave is observed in the EEG as a regular, high-amplitude wave with a frequency of 8-13 cycles per second.  If that sounds like EEG "Alpha" activity, to you, that's just what it is -- except that alpha activity is typically recorded in the occipital (visual) cortex, while mu activity is typically recorded in the frontal (motor) cortex.  A notable feature if alpha activity is blocking: when the subject is presented with a surprising stimulus, alpha activity disappears (or is greatly suppressed), and replaced by high-frequency, low-amplitude "Beta" activity.  It turns out that a similar blocking phenomenon occurs with mu waves, except that mu waves are blocked by engaging in, or watching, voluntary muscle activity.  But not in autistic children.  Comparing autistic and control children Oberman et al. (2006) found no mu-blocking when either group of subjects viewed balls in motion, and both groups showed substantial blocking when they made a voluntary arm movement; by contrast, the autistic children showed relatively little mu-blocking, compared to normals, when watching someone make a hand movement.  Oberman et al. attribute this difference in mu-wave blocking to a malfunction in the mirror neuron system.

Further evidence comes from an fMRI study of the frontal MNS when autistic children were asked to imitate, or just observe, facial expressions of emotion.  The extent of activation was negatively correlated with severity of symptoms along the autistic spectrum.  Put another way, the more severe the child's autism, the less activation was visible in the MNS (Dapretto et al., 2006).

All of this seems quite promising:  Mirror neurons support mindreading; autistic children appear to suffer from mindblindness; and autistic children also seem to have problems with the functioning of the mirror-neuron system that is the biological substrate of mindreading.  All well and good, until we recall, from the lectures on "Social-Cognitive Neuroscience", that there are problems with the traditional interpretation of mirror neurons.  They might not be important for action understanding after all (Hickok, 2011); and if they are not, there goes the whole explanation of autism in terms of mirror neurons.  Minor detail.....

Even accepting the "mindreading" hypothesis of mirror-neuron function, there are still problems with the hypothesis that autistic individuals suffer from a "broken" mirror-neuron system (Gernsbacher, 2011).  Mostly, findings such as those described above have proved difficult to replicate.

According to Gernsbacher's 2011 tally:

So, in view of problems with the "mindblindness" hypothesis of autism, as discussed by Tager-Flusberg (and Gernsbacher, too, elsewhere), and problems with the "broken mirrors" hypothesis, as discussed by Gernsbacher, it seems that there's more work to be done to nail down the psychological and neural deficits in autism.  And given the argument, again in the lectures on "Social-Cognitive Neuroscience", that an adequate psychological description must precede an adequate biological theory, it looks like the theoretical landscape on autism remains very much open for exploration.

Beyond the Theory of Mind: Social-Cognitive Development

The theory of mind, and similar theories, currently dominate discussions of social-cognitive development, but there are other approaches to social-cognitive development that are also fruitful.  Among the most interesting of these is the social-cognitive developmental (SCD) approach to the development of aspects of personality, proposed by Olson and Dweck ((2008).  Instead of looking at the development of social cognition, such as the ability to read other people's minds, the SCD approach examines the role of social and cognitive processes in development generally.  That is, SCD traces the role of social and cognitive processes in the development of such features as the individual's gender identity and role, aggressiveness, and achievement motivation.

In a sense, the SCD approach reverses the concerns of the ToM or "theory theory" approach, because it takes social cognition as a given, and looks at its effects on other aspects of social development and behavior.  But at the same time, these same findings can be used to illustrate the child's developing theories about himself and the social world.

The Phylogenetic View

Darwin's Origin of Species (1859) argued convincingly for a continuity between humans and other (sic), nonhuman, animals. The evolutionary perspective (the "modern synthesis") which defines modern biology was quickly imported into psychology, with assertions of continuity at the level of mind and behavior, not just anatomy and physiology. Interestingly, the stage for this importation was set by Darwin himself, who argued in the Descent of Man (1871) and The Expression of Emotions in Man and Animals (1872) that there were similarities between humans and other animals in the facial and other bodily expressions of emotions -- a position revived in our time by Ekman with his work on innate "basic emotions".  The implication is that social cognition is part of our evolutionary heritage, and that it is shared with at least some nonhuman species.   Which raises the question:

What is the difference between the mind of a human and the mind of an animal?

Of course, humans are animals too: we learned that from Darwin.  But still, there are differences:

The phylogenetic view of development raises the same question as the ontogenetic view: whether development is continuous or discontinuous.  Either way, though, it might be, to paraphrase Kagan, that "animals are a lot smarter than we think" -- and that their intelligence might extend to at least some aspects of the theory of mind.



053Darwin.jpg (102039 bytes)The first was the mirror test of self-awareness developed by Gallup (1970), based on observations originally made by Darwin himself. The initial response of chimpanzees (and other animals) to exposure to a mirror is to engage in other-directed behaviors -- i.e., to treat the reflection as a conspecific.  With continued exposure, however, chimps begin to engage in self-directed behaviors. Gallup hypothesized that the chimps came to recognize themselves in the mirror.

In a formal test of self-awareness, Gallup painted red marks on the foreheads of mirror-habituated chimpanzees.  The painting was performed while the chimps were anesthetized and the paint was odorless, so the animals could only notice the spot when they looked at themselves in the mirror. The chimps' response was to examine the spots visually (by looking at themselves in the mirror, touch the spots, and visually inspect (and smell) the fingers that had touched the spots. Gallup argued that the chimpanzees recognized a discrepancy between their self-image and their image in the mirror. 


img059.jpg (70347 bytes)060Marked.jpg (52163 bytes)Later experiments added methodological niceties, such as a comparison of marked and unmarked facial regions. 




062MirrorCont.jpg (94207
                                    bytes)Another Povinelli study indicated 061SelfMirror.jpg (50744
                                    bytes)that mirror self-recognition is relatively rare, even among chimpanzees.  It happens, but it's rare. 




Thus, not all chimpanzees show self-recognition in mirrors.  In general, the effect is obtained from chimps who are sexually mature (but not too old), have been raised in groups, and have had prior mirror exposure.

The same mirror self-recognition effect is found in orangutans and human infants, but not usually in other primate species (the status of gorillas, such as the famous Koko, is controversial).   

Self-Recognition in Solaris

In the great science-fiction film Solaris (Russian-language original scripted and directed in 1972 by Andrei Tarkovsky, based on the novel of the same title by Stanislaw Lem; English-language remake, 2002, by Steven Soderburgh, starring George Clooney), a psychologist, Kris, visits a space station orbiting the planet Solaris that has been the site of mysterious deaths.  It turns out that the planet is a living, sentient organism, which has inserted creatures into the station based on images stored in the cosmonauts' (repressed?) memories.  As soon as Kris gets to the station, he encounters the spitting image of Hari (she's named Rheya in the book), his late wife, who had committed suicide 10 years before.  Searching through his baggage, the woman comes upon a picture of herself, but she does not recognize the image until she views herself, while holding the picture, in a mirror.  It is clear that until that moment she had no internal, mental representation of what she looked like.  

The episode does not appear in Lem's book, which appeared in 1961, long before Gallup's original article was published (at least I can't find it) -- though Lem does have some interesting remarks about Rheya's memory -- she doesn't have much, and one of her memories is illusory.  But Tarkovsky's film appeared in 1972, so he would have had the opportunity to hear about Gallup's findings (which were prominently reported), and incorporate them into the script (losing the material about memory).

064Primate.jpg (51707 bytes)If it really turns out that orangutans have self-recognition, but gorillas do not, the implication is that the capacity for self-recognition arose independently at least twice in primate evolution.  This is because the evolutionary line that led to modern orangutans, who pass mirror self-recognition, split off from the main line of primate evolution before the slit which created the line which led to modern gorillas, which do not pass the mirror test. 



065Tamarin1.jpg (98665
                                      bytes) The matter is complicated, however, because 066Tamarrin2.jpg (61233 bytes) Hauser has obtained evidence for mirror self-recognition in cotton-top tamarinds, a species of monkeys that are, in evolutionary terms, quite distant from the great apes.



Gallup has argued that chimpanzees, at least, have bidirectional consciousness: they are responsive to events in the external world, and they are also aware of the relationship between these events and themselves. In his view, this last form of awareness is the hallmark of consciousness.


Theory of Mind

Long before Baron-Cohen,  David Premack and Guy Woodruff (1978) had argued that chimpanzees (and perhaps other primates, especially great apes) had a "theory of mind".   This paper -- actually written as a commentary on another paper introduced the concept of ToM to psychology, philosophy, and cognitive science.  In their study, they reported that at least one chimp, namely the famous Sarah, was able to solve novel problems that entailed attributing mental states to humans. Sarah had been raised since infancy in the company of humans, and had been taught a rudimentary visual "language" in which tokens represented various concepts. 

The experiments were variants on Kohler's classic demonstrations of insight problem-solving in chimpanzees (that's one of Kohler's apes, Sultan, in the slide on the left), and took advantage of the fact that Sarah loved to watch TV.  So, a series of problems were presented to her via 30-second video vignettes.  Each showed a human confronting some obstacle, and trying to solve the problem.  Sarah was then presented with two still pictures, one of which depicted the correct solution to the problem.  She was then asked to choose the correct solution, and place the corresponding picture near her TV set (as she had already been trained to do for other purposes).  Finally, Sarah received verbal feedback was to whether her choice was correct.  All the problems involved objects with which she was familiar.

070NonFoodProb.jpg (97392
                                                          bytes) Some of the problems concerned food.  In one, a bunch of bananas was suspended overhead, out of reach of her human caretaker.  The appropriate solution was to step on a nearby box.

Other problems did not concern food.  For example, a caretaker was trapped in a locked cage.  The appropriate solution was to get a key on a ring.
Sarah's performance on all the tests was excellent, regardless of whether the problems were about food.  In fact, Sarah was usually correct on the very first trial!
  • The problem that gave Sarah the most difficulty was the one involving two steps: removing cement blocks in order to move a box to get at food.

Premack and Woodruff concluded that Sarah did, indeed, have a theory of mind -- which is to say that she was able to connect representations of problems with representations of their solutions.  And that, in turn, required Sarah to impute mental states to humans:

Evidence for a theory of mind is evidence for social cognition in nonhuman animals -- extending not only to conspecifics, but to members of other species as well -- a point to which I'll return later.

Mindreading in Nonhuman Species

Elaboration of the theory of mind by developmental psychologists has led to further research on animal awareness by comparative psychologists, using adaptations of the tasks developed for investigating various components of the theory of mind in infants and children.  Baron-Cohen's modular theory of mindreading provides an excellent framework for reviewing this research.

076GazeShifting.jpg (43254
                                      bytes)For example, developmental studies show that infants consolidate the EDD by about 18 months.  Povinelli and Eddy (1996) found that chimpanzees will likewise follow the general direction of the gaze of a human experimenter.


078SharedChimp.jpg (92671
                                      bytes)However, another experiment by Povinelli et al.079SharedChimp2.jpg (38099
                                      bytes) (1999) suggested that chimpanzees lack SAM.  That is, they did not interpret the pointing behavior of a human experimenter as referring to objects in the environment.


Research by Josep (sic) Cal and Michael Tomasello (1998), using a nonverbal version of the false-belief task was the first to test for a theory of mind using a nonverbal version of the false-belief test with a mixed sample of orangutans and chimpanzees (as opposed to the case study of Sarah), as well as a sample of 4- and 5-year-old children

082FalseBelief.jpg (36794
                                      bytes)The 5-year-old children, but not the 4-year olds, performed well on both verbal and nonverbal versions of the test.  In contrast, the apes performed even more poorly than the 4-year-old children


Herrmann et al. (2007), working in Tomasello's laboratory,  recently reported the performance of large groups of chimpanzees (M age = 10 years), orangutans (M age = 6 years), and children (M age = 2.5 years) on a Primate Cognition Test Battery which included a number of tests of social as well as physical skills -- including two tests of of the theory of mind.  compared to the toddlers, the apes did quite well on physical skills involving space, quantity, and causality.  But the apes fell down completely on tests of social learning, and also did poorly on the tests of ToM.


Herrmann2007a.JPG (60630
                                                bytes) Herrmann2007b.JPG (58405


                                        (39841 bytes)To summarize, species comparisons suggest that monkeys and even great apes are sorely lacking in mindreading abilities, even by the standards of 3-year-old human children.  David Povinelli, a leading researcher in the area of primate cognition, has concluded that nonhuman primates appear to lack a theory of mind. 

Again, we have to be careful, because this area of research is rife with potential confounds.  Frans deWall, a leading primatologist who has written a great deal about the social behavior of chimpanzees and other primates (he pretty much discovered the bonobo, or pygmy chimpanzee, which makes love not war) has been particularly critical of these experiments. 

Moreover, the more recent results seem to contradict the conclusions by Premack and Woodruff, who found evidence that Sarah, at least, did have a theory of mind?  Povinelli has suggested that the chimpanzee ToM may include to a rudimentary understanding of desire, but no understanding of belief.  Interestingly, when Sarah "retired" to a refuge for chimpanzees, her new caretakers reported that Sarah "makes her wishes clear to staff and is quite opinionated about what she wants and when she wants it" (New York Times, 11/15/2011).

This conclusion may seem to violate the Darwinian doctrine of psychological continuity, but there have to be some discontinuities, somewhere, unless you want to start looking for evidence of ToM in cockroaches.  In commenting on this situation, Povinelli has turned Darwin on his head, reminding us that, despite the continuities uniting species, the whole point of evolution is to add new traits that aren't possessed even by close relatives.  For humans, language appears to be such a trait.  Perhaps consciousness, in the form of ToM, is another.

Still, the line of research on mirror self-recognition initiated by Gallup seems to indicate that chimpanzees, at least, have a rudimentary sense of self.  And that, as we shall see, reveals a rudimentary consciousness.


The Cultural View

In addition to the phylogenetic and ontogenetic views familiar to psychology, there is another view of development that can be found in other social sciences, such as economics and political science.  This is a cultural view of development, by which it is held that societies and cultures develop much like species evolve and individuals grow.  Which raises the question:

How is mind -- and, in particular, the "theory of mind" -- affected by cultural context?

Although, from a biological perspective, development entails a single process of maturation, Prof. Barbara Rogoff of UC Santa Cruz (and others) have argued that "individuals develop as participants in their cultural communities, engaging with others in shared endeavors and building on cultural practices of prior generations".  Although many psychologists construe "culture" in terms of race or ethnicity, Rogoff argues that this "social address" approach treats culture as static, and assumes that individuals with the same "social address" will share the same characteristics.  In contrast, Rogoff emphasizes the dynamic aspect of culture, acknowledging that cultural change across generations, and vary within communities.  And far from pitting "culture" against "biology", she seeks to explore the interactions between them.  In particular, there are certain biological universals that constrain cultural variability.  Rogoff traces a number of cultural influences in her book, The Cultural Nature of Human Development (2003), including childrearing, dependence and autonomy, independence and interdependence, lifespan developmental transitions, attachment, learning, cognitive development, and social identity.  She also makes the point that while the practices of Western, European-American middle-class culture seems "natural" and "normal" to those (like most psychologists and other social scientists!) who operate within it, these reflect just another cultural tradition to be studied -- not some "baseline" or "control group" against which other cultures are compared.  Rogoff's own research focuses on indigenous cultures of North and Central America, including Mayan communities in Guatemala, and immigrants to the United States from indigenous communities in Mexico. 

Here's a short list of cultural differences that, in Rogoff's view, may have important implications for the child's development (for a summary, see Rogoff's essay in the General Psychologist, 4(1), 2007, from which this material was drawn):

The Stages of Socio-Cultural Development

The origins of this cultural view of development lie in the political economy of Karl Marx, who argued that all societies went through four stages of economic development:

  • feudalism;
  • bourgeois capitalism;
  • socialism; and
  • communism.

Marx further proposed that certain habits of mind, or modes of cognition, were associated with each stage of development.  So, for example, it required a degree of consciousness raising to get peasants and workers to appreciate that they were oppressed by lords and management. Note that Marx didn't see any need for further consciousness raising for people to appreciate their oppression by communism.

In 1960, the American economic historian W.W. Rostow offered a non-Marxist alternative conception of economic growth which he called the stages of growth:

  • the traditional society;
  • the preconditions for take-off;
  • the take-off;
  • the drive to maturity; and
  • the age of high mass-consumption.

Along the same lines, in 1965 A.F.K. Organski, a comparative political scientist, proposed  stages of political development:

  • the politics of primitive unification;
  • the politics of industrialization;
  • the politics of national welfare; and
  • the politics of abundance.

Stage theories of political and economic development are about as popular in social science as stage theories of cognitive or socio-emotional development have been in psychology!  

Note, however, the implications of the term development, which suggests that some societies are more "developed" -- hence, in some sense better -- than others.  Hence, the familiar distinction between developed or advanced and undeveloped or underdeveloped nations.  The implication is somewhat unsavory, just as is the suggestion, based on a misreading of evolutionary theory, that some species (e.g., "lower animals") are "less developed" than others (i.e., humans).  For this reason, contemporary political and social thinkers often prefer to talk of social or cultural diversity rather than social or cultural development, thereby embracing the notion that all social and cultural arrangements are equally good.

It is this "diversity" view of culture that prevails today -- which is why psychologists can compare "Eastern" and "Western" views of self and social cognition, without any implication that one view is more developed than another.  They're simply different -- cultural choices.

Early cross-cultural studies of mind were somewhat informal, in that they were not based on psychological theories of mental life, nor did they involves systematic inquiry about the theory of mind. Still, the observations are interesting (for details, see Lillard, 1997, 1998).

Margaret Mead, studying the Manus people of New Guinea, observed that children in that culture were not animistic, but adults were.  This contrasts with the typical pattern in the West, where children attribute minds to inanimate objects (like their teddy bears), and then lose this tendency as they grow up.
Lawrence Kohlberg, studying the Atayal of Formosa, observed that children in that culture did not believe that their dreams were "real", but that adults did.  This also contrasts with the typical pattern in the West, where children often confuse their dreams with reality, but adults understand that dreams are "just dreams".

Rosaldo studied the Illongot, a headhunting tribe in the Philippines.  One of the striking features of the Illongot is that, while they have a keen interest in social relations, they have little interest in internal, mental life.  In Illongot culture, rinawa, the seat of thought and emotion, is located in the heart rather than the brain, and is held to be a property of all living hings -- human, animal, and plant life as well.  Newborns are fully endowed with rinawa at birth, which gradually leaves the body as the person ages and disappears entirely at death.  Rinawa leaves the body during dreams.  It can be stolen through magic, and restored through ritual.

Evans-Pritchard studied witchcraft in the Azande, a farming tribe in Central Africa.  The Azande sacrifice chickens for use as oracles.  Of course, some predictions are not fulfilled, but from the Azande's point of view, the oracles are never wrong.  Rather, oracular errors are attributed to witchcraft.  In fact, all unfortunate events are explained by witchcraft.  The Azande seem to have o sense of internal causation: all causation is at the hands of external agents.  Witchcraft itself is held to be motivated by envy, and accused witches always deny practicing witchcraft.  They're not doing it.  Something else is doing it!

Now, as I noted earlier, these are all fairly informal, observational studies -- at least by the standards of modern scientific psychology.

An early and excellent example of controlled cross-cultural research on social cognition was reported by Joan Miller (1984), involving Hindus living in the state of Gujarat, in India.  She interviewed urban-dwelling Gujarati, as well as urban-dwelling Americans living in Chicago, asking them to explain various events. 

  • Miller's dissertation advisor, Richard Schweder, had previously reported that Gujarati seldom described each other with the trait terms (like "neurotic introvert" that are familiar in Western culture.
  • Miller herself examined causal attributions, and found that while Americans tended to attribute deviant behaviors to actors' internal dispositions, the Gujarati attributed the same deviant behaviors to factors in the external situational context. 

Picking up on Miller's theme, Angeline Lillard (1997, 1998, 1999) has noted that there are two broad ways to explain behavior of all sorts:

Recall Rogoff's caution against thinking that things that are commonplace in Western, developed, modern European-American culture are "natural" or "normal".  The theory of mind is one such thing, and Lillard concludes, based on her survey of cross-cultural data, that the theory of mind might not be innate after all.  She suggests that mind may be a concept in Western "folk psychology" that has no counterpart in other psychologies.  While European-Americans may prefer to explain behavior in terms of mental states, what she calls "Other Folks'" psychology  -- may be quite different (in fact, some other folks might not have a psychology at all!).  There may well be profound cultural differences in the concept of mind, and in such details as the difference between thinking and feeling.  If so, the concept of "mind" may be a learned cultural product, acquired through socialization, and maintained by cultural transmission -- not something that is pre-adapted, modular, and embedded in the architecture of the brain.

Perhaps, but even Lillard concedes that some aspects of folk psychology may reflect the operation of something resembling Chomsky's Universal Grammar (UG).  That is, we may possess an innate cognitive module for explaining behavior -- a cognitive module that is, in turn, in evolved neuroanatomy.  She notes that no culture is absolutely behavioristic -- abjuring any reference to mental states.  And all languages have words that refer to mental states, even if these words are not commonly used.  As in Chomsky's UG, she suggests that cultures vary in parametric settings just as languages do, so that behavioral explanation varies from culture to culture in terms of its constraints and preferences -- variance which comes through in adults' conceptualizations of the causes of behavior.  What's universal in this UG, she speculates, is false belief.  That is, we may be innately prewired to take notice when behavior does not accord with the way the world really is -- as in the standard false belief task or its infant cognate.  Explaining this discrepancy may take various culturally defined forms.  In modern, Western, European-American culture, we explain this discrepancy in terms of how reality appears to the actor.  In other cultures, the same discrepancy might be explained in terms of witchcraft, or the alignment of the stars.

So mentalistic explanations of behavior may not be innate.  Even these may be cultural achievements.

With respect to self-awareness, paleontological data would suggests that it is very old indeed. Excavation of some of the earliest cave-dwellings has yielded evidence of beads and other adornments. These clearly indicate that these early human ancestors knew what they looked like -- and cared about it.

It is even possible that intersubjectivity -- the belief that others have minds that might differ from one's own, that the contents of others' minds might be interesting, and that we could infer these contents from their behavior -- might be a cultural invention, like the wheel or written language.

Setting aside this admittedly controversial reading of ancient (and modern) literary texts, proposals for something like a UG for behavioral explanation imply that there should be cross-cultural similarity in the way that young children explain behavioral discrepancies.  The development of a theory of mind (or whatever the culture prefers) would then come later, as a result of socialization and acculturation.  The result would be that, among adults, there is some cultural variance, or "choice", in explanations of behavior:


This page last revised 07/25/2017.