Cognitive Development and the Innateness Issue - Evolutionary Psychology PDF

5 Cognitive Development and the Innateness Issue Key Concepts nativism empiricism constructivism epigenetic landscape domain specificity modularity imprinting critical period sensitive period theory of mind autism Williams syndrome neuroconstructivism...

5 Cognitive Development and the Innateness Issue Key Concepts nativism empiricism constructivism epigenetic landscape domain specificity modularity imprinting critical period sensitive period theory of mind autism Williams syndrome neuroconstructivism A common belief about evolutionary psychology is that it assumes that a large number of psycho- logical abilities are innate in the sense that they are present at birth. Indeed, in their earlier writing Leda Cosmides and John Tooby (e.g. 1992) imply this in their discussion of ‘innate mental mod- ules’. But innateness can have more than one meaning and evolutionary psychology does not require everything to be present at birth in order to be a viable way of understanding human psychology. Many abilities might have evolved and be under genetic control but not present at birth. Perhaps the most obvious is the ability to reproduce, something that is largely under the guidance of genes but not present in babies or children. In this and the next chapter we discuss the innateness issue in some depth and explore how something can develop as part of a genetic programme but still be sensitive to environmental influence. Another relevant issue is that traditional forms of developmental psy- chology (for example that of Piaget) view child cognition as an incomplete or inadequate version of adult cognition. In fact, what evolutionary psychology teaches us is that babies and children have their own ecological problems to deal with which are sometimes very different from those of adults and solving these requires different psychological abilities from those of adults. To begin with, however, we need to investigate some of the more traditional ways that innateness has been discussed and evaluated and also investigate the extent to which innate psycho- logical ‘modules’ (see Chapter 1) exist, and how these might develop. Nature, Nurture and Evolutionary Psychology One of the central debates of developmental psychology is the so-called nature versus nurture de- bate. This asks to what extent human behaviour is the result of environmental factors (nurture) and to what extent it is the result of innate biological factors (nature). This question has a long history, starting at least as early as the Ancient Greek philosophers in the fourth and fifth centuries BCE and has been revisited by a variety of thinkers ever since. Throughout history the pendulum of opinion has swung in favour of one or other of these forces as new theories are developed and evidence accumulated. Recently the evolutionary approach has led to a renaissance in nativism as a means of explaining human behaviour and, in particular, that development involves domain-specific mental modules (more of which later). 114 Cognitive Development and the Innateness Issue Innate Similarities and Innate Differences The nurture position was perhaps most famously summarised by the behaviourist psychologist J. B. Watson in the early part of the twentieth century who claimed: Give me a dozen healthy infants, well-formed, and my own specified world to bring them up in and I’ll guarantee to take any one at random and train him to become any type of specialist I might select: doctor, lawyer, artist, merchant-chief, and, yes, even beggar man and thief, regardless of his talents, penchants, tendencies, abilities, vocations, and race of his ancestors. (Watson, 1925, 82) The above quotation seems to be a fairly unequivocal statement of a belief that innate mental faculties do not exist and, were it true, might be seen as sounding the death knell for nativism in general and evolutionary psychology in particular. In fact, taken in isolation, all Watson’s quotation suggests is that innate mental faculties play no role in the formation of individual differences (see Chapter 13). This is important because current thinking in evolutionary psychology draws a distinction between the claim that differences between individuals are innate and the claim that similarities between indi- viduals are innate which are completely independent claims. To see why consider the following exam- ple (adapted from Block, 1995). It is well established that the fact that most of us are born with five digits on one hand (rather than four or six) is entirely due to the genetic programme that dictates the development of the individual’s hand. Conversely, that some of us might have fewer than five digits is almost totally the result of the environment, for example a finger might have been lost in an accident, or some toxin might have interfered with the individual’s growth during prenatal development (as was the case with the morning sickness medication Thalidomide which was prescribed in the 1950s). This means that environmental factors account for almost all of the differences between individuals in digit number, even though genetic factors account for the majority of us having five digits. Incidentally this means that if we were to calculate the heritability of having five fingers it would be zero (see Chapters 2 and 6 for more on heritability). This is because heritability is a measure of the variation of some trait due to genes, not the extent to which something is ‘inherited’. Since the variation of finger number is almost entirely environmental, there is little or no variation due to genes and so the heritability would be close to zero. A point worth bearing in mind and one that we will return to in the next chapter. Look at Watson’s argument again. Even if he were correct his claim does not necessarily undermine either nativism or the claims made by evolutionary psychology. For example, some theo- rists have suggested that natural selection has endowed us with an innate ability to acquire language (see Pinker and Bloom, 1990; Pinker, 1994; see also Chapter 10). Using the same argument as we did for the number of digits it is entirely possible that the ability to speak a language is due to the presence of an innate ‘language organ’ as it is sometimes called, whereas individual differences in language skills could be entirely due to environmental factors such as educational opportunities, brain damage and so on. So even if Watson could train his infants into a career of his choosing it does not rule out the notion of evolved learning mechanisms. As we shall see in Chapters 6 and 12 the evidence from behavioural genetics research shows that there are strong genetic influences on individual differences too, so Watson was also wrong in this regard. Beyond Nature versus Nurture Evolutionary psychology has a strong association with nativism (see Chapter 1), but the problem is that words such as ‘nativism’ and ‘innate’ are taken to mean very different things by different people.  Nature, Nurture and Evolutionary Psychology 115 or understand language. So when Noam Chomsky or Steven Pinker argues that language is innate they are not arguing that babies are born with full knowledge of language; they mean that children are born with domain-specific predispositions that enable them to acquire language efficiently (see Chapter 10). MENTAL MODULES AND DOMAIN SPECIFICITY As we discussed in Chapter 1, the form of evolu- tionary psychology that began in the 1990s with the Adapted Mind (Barkow, Cosmides and Tooby, 1992), sometimes known as the Santa Barbara School, placed great emphasis on modularity. To summarise, this notion was originally proposed by the philosopher Jerry Fodor (1983) and suggests that the mind contains mental modules. Modules are basically processing devices and, of their many properties, the ones most relevant to this chapter is that they are thought to be at some level innate and domain-specific. This differs from many previous theories which propose domain-general pro- cessing. It is worth here reflecting further on the notion of domain-specific versus domain-general processes. An analogy might be with Lego bricks. In the early days, a Lego set largely consisted of the familiar bricks of different sizes from which you could make a range of objects: buildings, cars, aeroplanes, rockets; whatever your imagination could come up with. You could even make a passable attempt at a person albeit a pretty unrealistic block-person, such as you might expect to see in the computer game Minecraft. Old-style Lego bricks are domain-general: you can make a range of different kinds of objects (or ‘domains’) using essentially the same materials. Later on, the Lego company started to manufacture ‘bricks’ that were specific shapes to enable a more realistic model to be made more quickly. So there are now types of bricks that are specifically designed to build people (heads, arms, legs), cars (wheels, windscreens, doors) and so on. These are domain-specific, meaning that each kit is designed for a specific domain, or type of object. The advantage of these domain-specific kits is that you can produce a better model more quickly, the disadvantage is that the set of things that you can build is much more limited than for old-school ‘domain-general’ Lego. (You may have already realised that there are degrees of general- ity/specificity: modelling clay is extremely domain-general, if you are suitably skilled you can make realistic models of anything, whereas some Lego kits are extremely specific, only allowing you to make a model of a specific object, for example, Yoda; see Figure 5.1.) To return to domain generality and specificity in learning and development. Domain- general learning is like old-style Lego, you have a set of simple elements and processes that can be used to learn potentially anything: language, how to cooperate with others, how to recognise faces, how to solve moral problems and so on. An example of a domain-general learning mechanism is the ability to form associations between different objects, as proposed by the behaviourists (see Chapter 1). Just like with old-school Lego, domain-general processes can be used to learn, but not very quickly and probably not very well. Domain-specific learning (or domain-specific processes) is like modern Lego: you can learn something very quickly and very well, but such learning only really works for the specific thing (or domain) it is designed to be used for. So, if we were to argue (as many have) that language is learned using domain-specific processes it means that there are features of that learning that work for language and for language alone. You could not use them for morality any more than you could build a car out of a Lego set designed to build people. If you wanted to learn about morality, you had either to fall back on the slow and imperfect domain-general mecha- nisms (old-school Lego) or you used knowledge and processes specific to the domain of morality. A claim of the Santa Barbara school was massive modularity, which simply means that there are many domain-specific modules for a range of human activity. Modules that are dedicated to understanding people, physics, language, biology or language. The reason for proposing massive 116 Cognitive Development and the Innateness Issue (a) (b) Figure 5.1 (left) ‘Old-school’ and (right) ‘modern’ Legos used to create the Star Wars character Yoda. It should be obvious that the Yoda on the right would be vastly easier and faster to build as it contains ‘domain-specific’ pieces (specific to the domain of Yoda), whereas the ‘domain-general’ bricks on the left could be used to make many things other than Yoda. modularity is that it was believed that whenever a species encountered the same problem repeated- ly, natural selection would tend to favour a domain-specific approach over a domain-general one, because the former is more efficient. A child who, at a young age, can communicate with others, recognise its family and show fear in dangerous situations will be more likely to survive than a child who develops these abilities much more slowly. What does all this have to do with innateness? To begin with it is important to note that no psychologist proposes that humans have no innate competencies whatsoever. Even the high-priest of blank-slateism B. F. Skinner believed that animals are born with the ability to form associations (for example learning that pushing a lever leads to a reward), the ability for prior associations to be extinguished (if the lever leads to no food on sufficient occasions then stop pressing it) and the ability for behaviour to be modified by rewards and punishments. These innate competencies are domain-general, Skinner believed that the same processes can account equally for pigeons pecking to get food and humans learning language. Obviously to suggest that we have domain-specific competencies means that there must be more innate knowledge but there are debates about the extent of this innateness. Some, as we shall see, suggest that innate knowledge is relatively low level, dealing with basic perceptual information, whereas others propose that we are born with detailed knowledge of domains such as physics, face recognition and theory of mind. The Early Emergence of Specific Competencies One of the obvious tests for the innateness of an ability is whether it occurs at birth or soon after. Human newborns seem so helpless that it is easy to believe that, apart from a few simple behaviours  Piaget’s Developmental Theory 117 and reflex actions, they seem to have no real cognitive abilities whatsoever, and that all of the com- petencies that they have as older children are slowly acquired by interacting with the environment. This perspective was advanced by the most famous developmentalist of all: the Swiss psy- chologist Jean Piaget (1896–1980). He saw his view as an alternative to the extreme environmental- ism of the philosopher John Locke (and later Watson and Skinner; see Chapter 1) and the extreme nativism of Descartes (and later Chomsky and Fodor; see Chapters 1 and 10). Rather, Piaget saw development as a process whereby the child actively constructs its understanding of the world. Piaget’s Developmental Theory One of Piaget’s most important contributions was his suggestion that there was much in common between psychological development and biological development. Modern geneticists do not see genes as determining physical and behavioural development (see Chapter 2); rather they are seen as guiding the process of development in a probabilistic manner. So behavioural geneticists discuss a ‘developmental timetable’, which involves a series of likely milestones (such as when the first clum- sy steps will be taken or the utterance of the first spoken words). Such a developmental sequence does not proceed along a fixed pathway but depends on a large amount of environmental feedback. Likewise, Piaget (see Figure 5.2) believed that psychological development was neither the result of impressions made via the senses (as Locke suggested), nor the execution of some innate plan. Experiences matter, but that which is learned from experience is constrained by what was already present in the infant’s pre-existing mental structures. Figure 5.2 Jean Piaget. 118 Cognitive Development and the Innateness Issue Box 5.1 Stage Theories of Development A number of theories, but in particular Piaget’s theory, postulate that development progresses in a stage-like manner. It is important to know what this means since it can be somewhat different from what people mean colloquially when they say that a child ‘is going through a stage’. Stage theories usually specify that the child’s knowledge changes drastically at one or more points in development. Piaget’s theory suggests large qualitative changes in knowledge, rather than changes which are merely quantitative, such as learning a new fact about the world or a set of new facts. Here ‘qualitative’ means a change in quality rather than a change in degree, Piaget believed that a child in one stage had a fundamentally different way of thinking about the world than a child in a different stage. Such changes often have a marked effect on behaviour, for example, in Piaget’s theory the ability to adopt the perspectives of others (but see later in this chapter). Piaget’s stage theory is domain-general, so an underlying change affects many domains (e.g. person perception, language, physical reasoning) rather than being specific to a particular domain. Stage theories require changes to be rather sudden; the child’s understanding of the world changing noticeably over a few weeks or months, followed by a more gradual increase in knowledge. The Epigenetic Landscape One question that Piaget sought to address was the observation that although children have very dif- ferent experiences, they nevertheless tend to develop in a very similar way, for instance, reaching the major developmental milestones at approximately similar ages. In order to explain this phenomenon Piaget drew upon the work of geneticist C. H. Waddington (1975), who proposed that development could be thought of in terms of an epigenetic landscape. Figure 5.3 shows such a landscape. The ball is at the top end of a valley that subdivides a number of times, and the path of the ball represents Figure 5.3 The epigenetic landscape.  Piaget’s Developmental Theory 119 the particular developmental trajectory taken by an individual child. As the ball rolls through the landscape, different environmental conditions might lead to perturbations in the ball’s trajectory, but it will tend to return to its original path. Only extreme environmental activity would cause the ball to switch pathways and go down a different valley (depicted by the yellow arrows). Therefore, although environmental conditions can affect the path that the ball takes, these effects are always constrained by the structure of the landscape. To relate this back to human development it suggests that although experiences can affect psychological development, the nature and size of the effect is constrained by the structure of that which is already present (the topography of the epigenetic landscape). These constraints serve as a buffer and children will tend to turn out similar to each other despite often having quite different learning experiences (we return to similar accounts in Chapters 6 and 14). Learning about the Physical World Like all other animals, humans exist in a world that obeys the laws of physics in which unsupport- ed objects fall to the ground, solid objects cannot pass through one another, and getting from A to B requires one to pass through all of the points in between. Understanding these and many other physical principles is essential if the individual is to negotiate the world in an effective manner. If evolution has shaped the way that the mind works then we might expect such understandings to emerge early in life. The Piagetian infant has no understanding of physics. A cynic might counter this by saying that nor do most high school students. This is, however, incorrect as all clinically normal children have an intimate understanding of the physical world. Not the school physics of inclined planes, frictionless surfaces and masses which take up no space, but the common-sense physics that enables us to move around the world, throw things, use containers for liquid and reason about the behaviour of objects. It has been suggested (McCloskey, 1983; Slotta et al., 1995) that it is our deep understanding of common-sense physics that makes learning school physics so difficult. The real world does not have frictionless surfaces; air resistance is a factor in most interactions and concepts such as heat and energy have formal interpretations very different from their mean- ings in common language. Learning school physics is hard because we first need to unlearn our intuitive physics. One of the most basic assumptions of adult physical knowledge is that objects endure in- dependent of our ability to perceive them, something that was investigated by Piaget and is known as object permanence. Object Permanence Piaget believed that the infant has to learn practically everything about the physical world, even that objects continue to exist when they are out of sight. Piaget was led to this position partly because his theoretical starting point was to assume as little innate knowledge as possible, and partly because he believed that he had evidence for it. In one of his studies an infant (younger than ten months) is seated opposite a desirable object such as a soft toy. Under normal circumstances the child shows interest in the toy and attempts to reach towards it. If, however, an opaque screen is placed between the infant and the toy, blocking its view, the infant quickly loses interest in the obscured object, treating it as if it had ceased to exist. (Older children behave differently; they simply reach round the screen to retrieve the toy.) 120 Cognitive Development and the Innateness Issue Piaget’s interpretation of these results is that, to the infant, the object no longer exists which is clearly very different from adults’ understanding of the physical world. We know, or at least believe, that objects continue to exist independently of our ability to perceive them. Of course, over a longer time scale, we adults often place objects so that they are out of sight and subsequently forget that they exist. The banknote that we placed in the back pocket of our trousers and is ruined by a journey through the washing machine; the key that we left on top of the table that we thought was in our purse; and the piece of wedding cake that we find in the pocket of our suit-jacket the next time we use it, often at another wedding. When this sort of thing happens the discovery of the object will often bring to mind the occasion of its placement, or, if we cannot remember, we assume that something must have led to its appearance. We put forward such explanations exactly because of the way we assume the world to work: things do not spontaneously pop into existence. We also assume when we search for something that is no longer where we put it, that the object must be somewhere: things do not just pop out of existence either. These assumptions are fundamental to the adult’s intu- itive theory of physics, but not, according to Piaget, the intuitive physics of the infant. Box 5.2 Habituation Procedures Habituation procedures allow us to have insight into the minds of infants by capitalising on the fact that infants, just like adults, eventually find repetition boring. If the same thing happens repeatedly then, no matter how interesting it was initially, the infant slowly loses interest, and indicates this lack of interest for all to see by averting its gaze from the repetitive stimulus. We can therefore think of habituation as learning not to attend to a stimulus. If we present a stimulus – say, a short animation – repeatedly to two groups of infants they will eventually show signs of boredom. We then present each group with a different animation – call them animation A and animation B – and measure the amount of interest (how much they are looking at the animation, how long it takes them to become bored). Suppose that the infants who had animation A showed significantly more interest than those receiving animation B. We might therefore assume that animation A was seen by the infants as being more different from the original animation than animation B. We can therefore obtain some measure of how infants categorise the world. Of course, in the study described perhaps animation A was just more interesting to all concerned than animation B, so the study will have to have various control conditions to rule out such possibilities. The problem with Piaget’s interpretation of the results is that they do not demonstrate con- clusively that the infant has no understanding of the permanence of objects. For example, the hidden object might simply have become less salient, the infant simply forgets that it is there. Alternatively, maybe the infant knows that the toy is still there, but does not know how to go about retrieving it. Piaget’s original study is therefore inconclusive since there are a number of alternative ac- counts other than the radical one that the infant believes the object has disappeared from existence. One of the big problems in research on infancy is that we cannot simply ask infants what they think as we might with adults, this is why Piaget used infants’ searching behaviour as an index of what was going on in their minds. Recently techniques have been developed which enable us to probe infants’ thought processes in a more systematic way using techniques such as habituation procedures (see Box 5.2).  Studying Object Permanence Using Habituation Studies 121 Experimental Condition Contro Condition (with box) (without box) Habituation Events Test Events Impossible Event 180º Event Possible Event 112º Event Figure 5.4 The apparatus used by Baillargeon. Studying Object Permanence Using Habituation Studies A number of studies has used this habituation procedure on the object permanence problem (Bower, 1974), but the most widely cited and elegant study was conducted by Renée Baillargeon (Baillar- geon, 1987; 1991) using infants of around four months of age. Figure 5.4 shows an outline of the equipment used. The infant is seated opposite a device in which a screen rotates through 180° away from the infant. Video cameras are focused on the infant’s head to enable the time the infant spends looking at the screen to be measured. The screen moves through its cycle over and over again until the infant shows signs of bore- dom by looking away from the screen. This initial habituation phase is then followed by one of four alternatives depending on the experimental condition being run; for the moment we will just focus on the two critical ones. Following habituation, a block is placed behind the screen of a height such that it is clearly visible to the infant as the screen starts to move up, but becomes obscured when the screen is at about 60° through its rotation. Once the block is in position the screen is again rotated, in one condition the screen stops at the point at which it would make contact with the (obscured) block at 112°, in the other condition the screen continues to rotate as if the block were no longer there. In both conditions, interest is measured by amount of looking. Think of what Piaget’s theory would predict. The Piagetian infant, remember, believes that once something is out of sight it no longer ex- ists. If this were the case, then the infant should not be at all surprised by the screen rotating through 180° in the test phase because as soon as it is obscured by the screen it ceases to exist and cannot therefore stop the screen. Indeed, the Piagetian infant should find the condition where the screen stops more interesting since this is more different from the habituation phase because the screen moves through a smaller angle and is therefore perceptually more distinct. In fact, Baillargeon found 122 Cognitive Development and the Innateness Issue Experiment 1 60 Experimental Condition (with box) 40 Impossible Event 20 Looking Time (sec) Possible Event 0 60 Control Condition (without box) 40 112° Event 20 180° Event 0 –6 –5 –4 –3 –2 –1 1 2 3 4 Habituation Trials Test Trials Figure 5.5 Data from Baillargeon (1987) showing the responses of infants (mean age 4 months 14 days) to the apparatus shown in Figure 5.4. (left) Habituation (looking times decrease as the babies become bored). (right) Looking times in the test phase. The upper-right panel shows that babies look longer at impossible events (where the box apparently disappears) than possible ones (where the screen is stopped by the box). that infants attended longer to the 180° condition than to the condition where the screen was stopped by the obscured block, indicating that this was surprising to them. Baillargeon interprets these re- sults as suggesting that, just like adults, infants expect objects to endure even though they are out of sight – something that is in sharp contrast to Piaget’s prediction. Box 5.3 Other Physical Principles Held by Infants Using habituation techniques Elizabeth Spelke and colleagues (Spelke et al., 1992) have suggest- ed that by 4 months of age infants have a number of expectations of the physical world which she refers to as principles. These principles state the following: Objects move as wholes on continuous paths. In other words, objects maintain their shape and if going from A to B have to pass through all the points between. (This could be called the ‘No Teleportation’ principle.) Objects cannot pass through one another or occupy the same point in space or time. (This could be called the ‘No Ghosts’ principle.) Objects cannot act on each other unless they come into contact. (This could be called the ‘No Telekinesis’ principle.)  Studying Object Permanence Using Habituation Studies 123 Box 5.3 (cont.) Infants are not born with all the principles that they will eventually possess; some learning must take place. For example, Spelke et al. (1992) found that six-month-olds, having been habituated to an object falling onto a platform, were surprised when on a test trial the object subsequently came to rest inches above the table. Four-month-olds, however, showed no such surprise, indi- cating that their knowledge of gravity was not fully formed. As we discussed above, this may suggest that babies are born with domain-specific rules to learn about the physical world which requires a degree of input, rather than being born with a complete knowledge of how the physical world operates. It is worth mentioning the two other conditions. In these conditions the screen behaved in exactly the same way as it did in the other two conditions, but no block is used. These conditions were to ensure that it wasn’t simply something about the movement of the screen but was rather due to the interaction between the block and the screen. In these two conditions infants showed more attention to the condition where the screen stops at 112°, than when it carries on through its full rotation, presumably because the screen stops for no apparent reason. This early knowledge of how the physical world operates is by no means a talent unique to human infants. Psychologists Marc Hauser and Susan Carey (Hauser and Carey, 1998) have shown that adult cotton-top tamarinds (new world primates) have similar assumptions about the physical world to those of a five-month-old human infant. They also share with human infants basic number skills such as the knowledge that one plus one is two, two minus one is one, and one plus two is more than two (Wynn, 1993). The similarities between non-human primates and human infants should not be too surprising since the survival of both animals depends upon an understanding about the way that physical objects behave and is therefore likely to have been influenced by natural selection. Human infants therefore may well be born with certain expectations about the way that the physical world operates. This makes evolutionary sense. Imagine a population of infants who have no fear of heights. Now imagine a genetic mutation arises in one individual in this population which provides that individual with a seed of a fear of heights. If having this fear of heights increased the infant’s chances of surviving to maturity compared to its peers then it can be seen that such a muta- tion would, over a number of generations, spread rapidly throughout the population. Should we be surprised by the talents of young infants? In one sense we should not be. The physical world is very stable, and objects have obeyed the same rules for billions of years. In evolutionary terms it would make sense for people to be given a head start in acquiring these rules by endowing them with psychological constraints that make learning easier. What is learned as a result of these constraints is a common-sense theory of physics. Is This Evidence for Innate Knowledge? The early emergence of knowledge of the physical world might be taken as supporting the existence of an innate physics module. But as we have seen, many of the physical principles (e.g. gravity) seem to emerge in the first six months, rather than them being present from birth (recall, how- 124 Cognitive Development and the Innateness Issue ever, the point made above regarding puberty: the fact that some ability is not present from birth doesn’t mean it is not innate). Carey and Spelke (1994) suggest that babies are born with a set of domain-specific innate physical principles that serve to guide their developing knowledge of the physical world based on their experience. In the next section we present a more detailed theory of how nature and nurture interact in this way when we discuss how babies learn about people around them. As for whether it is evidence for innate modules, as some evolutionists have claimed, we dis- cuss that later in this chapter. Recognising Conspecifics: A Comparative Perspective The word ‘conspecific’ means ‘members of the same species’; a dog’s conspecifics are other dogs and likewise a person’s conspecifics are other people. People are very important to babies: they provide nurturance and protection; they also represent the things that infants will ultimately have to get along with when they are older. Ethologists have noted that this is also the case for many species of social animal. Nobel Prize–winning Austrian ethologist Konrad Lorenz (Figure 5.6) is famed for uncovering an early attachment process which he called imprinting. In Lorenz’s notion of im- printing, following birth, precocial animals (those that are born in a mature state such as geese and lambs) enter a critical period during which they rapidly learn the visual (and other) characteristics of their mother. At around 25 hours following hatching domestic chickens and geese, for example, will have learned the specific characteristics of their mother and from this point on will avoid other adults of their species. Lorenz saw this special learning process as the outcome of natural selection since those individuals that rapidly became attached to their mother would be more likely to benefit from the protection that she provides and thereby pass on their genes for this specific early learning process. He also suggested that the young emerge without a preference as to what the imprinting object might be since in the natural world the object will almost certainly be their mother (Konrad Lorenz imprinted young geese on himself; one of the most endearing pictures of Lorenz shows him swimming in a lake surrounded by his own clutch of goslings). Subsequent ethologists have demonstrated that this early learning period is less circum- scribed than Lorenz imagined. Cambridge ethologists Pat Bateson and Robert Hinde (Bateson and Hinde, 1987), for example, discuss imprinting as occurring during a more plastic sensitive period when chicks are more likely to learn the characteristics of their mother. In their view this period depends very much on what sort of external input the chicks receive. Being kept in the dark and viewing an imprinting object later may extend this period for up to several days after hatching. Other researchers have also modified our notion of specificity of the imprinting object; Johnson and Bol- huis (1991) have, for example, demonstrated that chicks, given the choice, will develop a preference for a hen-like object. Despite modifications to Lorenz’s original idea of imprinting, present-day ethologists would agree that he was broadly correct in his assertion that its primary function is to enable recognition of, and attachment to, the mother by the young individual and that such a process is a product of natural selection (Workman, Adam and Andrew, 2000). Evolution has therefore endowed some non-human animals with a ‘quick and dirty’ ap- proach to recognising their parents which basically says ‘attach yourself to whatever you perceive during a very early sensitive period within certain broad constraints’. In the natural environment this works very well since the first thing that a newly hatched chick will see is likely to be its brooding mother, it only tends to go wrong in unnatural environments where chicks are raised in incubators.  Recognising Conspecifics: A Comparative Perspective 125 Figure 5.6 Konrad Lorenz and his ‘children’. To return to human development, clearly the ability to recognise conspecifics is crucial for a baby’s survival and future development; perhaps even more so than it is for chicks, given the importance of social interaction in human life. How, then, do babies discriminate people from other objects, and how do they discriminate between different people? Evidence suggests that, just like adults, babies use faces to inform them whether or not something is a person, and which particular person it is. Infants’ Preferences for Faces in General There has long been evidence that face-like stimuli are interesting to infants. Fantz (1961) compared newborn infants’ looking times at a variety of different visual stimuli; each presented to the infant in pairs. The stimuli consisted of patches of colour, patterns such as bull’s-eyes and checkerboards, and stimuli which looked like simple faces. Fantz found that infants spent significantly more time looking at the stimuli that resembled faces than any of the others. Fantz’s results might be taken as suggesting that infants are born with some knowledge of what a human face looks like, but how detailed is this knowledge? Do infants need to learn any- thing about faces apart from being able to discriminate among them (e.g. who is ‘mother’, who is ‘father’?). As an analogy, you might prefer red wine to blackcurrant juice, but this does not make you a wine connoisseur, because the stimuli are radically different. Similarly, preferring faces to chessboards does not tell us how detailed this knowledge is. To investigate this, Johnson and Morton (1991) conducted a similar study, but they were interested in whether Fantz’s study indicated that infants were born with a fully functioning face identification system, or whether they were born 126 Cognitive Development and the Innateness Issue Realistic face Primitive face Scrambled face Complex pattern Figure 5.7 Stimuli used by Johnson and Morton. with only sketchy knowledge of faces, the details being filled in by experience of faces in the world. Johnson and Morton, like Fantz, used a preferential looking paradigm in which the infant was pre- sented with stimuli shown in Figure 5.7. As can be seen, one of the stimuli is a simple checkerboard pattern, to control for stimulus complexity. Another stimulus was a ‘realistic’ face containing all of the components of a real face (eyes, nose, mouth) depicted in a relatively realistic way. Another was a primitive face, just dark patches in the same configuration as a normal face. The final stimulus contained the same compo- nents as the realistic face but scrambled up into a non-face-like configuration. This condition was to ensure that any preference for the realistic face was due to the configuration of the components, not simply the appearance of the components themselves (a similar control was used by Fantz). The subjects for this experiment were newborns and 4-month-old infants. Johnson and Morton’s results show that for newborns both the primitive face and the real- istic face elicit approximately equal looking times, indicating that they are equally interesting, and both of these are preferred to the other stimuli. The results are rather different for the four-month- olds, although the primitive face is still more interesting than the checkerboard pattern and the scrambled face, the realistic face elicits much longer looking times. Johnson and Morton interpret these results as suggesting that infants are born with some preference for face-like stimuli, but it is by no means fully specified; experience is still needed to flesh out the details. They propose that learning about faces involves two processes which they call Conspec and Conlern, Conspec, it is suggested, is an innately specified, subcortical set of princi- ples that are responsible for directing attention towards stimuli that resemble human faces, but this knowledge is crude and cannot distinguish between primitive and more detailed face-like stimuli. Perhaps the most important thing in learning is that you attend to the appropriate information; con- spec ensures that this happens. Guided by Conspec, Conlern fleshes out the primitive representation based on experience of looking at faces and forms a more realistic representation of what a face is like. This fleshing out process can be illustrated by research showing that at six months human infants can discriminate monkey faces as well as human faces but by nine months this ability to dis- criminate monkey faces disappears (Pascalis et al., 2002) unless exposure to monkey faces contin- ues (Pascalis et al., 2005). Similar results were found for Japanese macaques (Macaca fuscata) who learn to prefer monkey to human faces when given experience of monkey faces or prefer human to monkey faces when given experience with human faces. This further suggests that the mechanisms that underpin conspecific recognition in humans are considerably older than our species, being found as it is in species who share a common ancestry with humans 25 million years ago. Whether these mechanisms exist in more distantly related primates or outside the primate order remains an open question (Johnson, Senju and Tomalski, 2015).  Mind-Reading: The Development of a Theory of Mind 127 Recognising Specific People The research above shows that even newborn babies are social animals in that they are motivated to look at face-like stimuli. Moreover, within this general preference for faces, it appears that some faces are preferred over others. Research suggests that newborns show a preference for their moth- er’s face (Bushnell et al., 1989; Walton et al., 1992) rather than the face of a stranger. It should be pointed out here that there is almost certainly nothing special about the mother, the preferential allocation of attention seems to be towards the person with whom the neonate has had the closest contact in the hours following birth, this is usually (but not always) the biological mother. Babies’ knowledge of faces seems to be quite sketchy however, as the preference disap- peared if the mother and the non-mother were wearing identical wigs, suggesting that the infant was using cues such as the outline of the face more than the internal features (see Figure 5.7). These results showing a preference for significant people are consistent with the predic- tions of modern evolutionary theory that organisms able to recognise close relatives and distinguish them from non-relatives will derive a selective advantage (as we saw with imprinting above). Care must be taken when interpreting the results of these studies. They do not suggest that newborns in any way recognise their mother in the way that adults do. Face recognition is a complicated process that involves, among other things, a knowledge that the face denotes a unique individual, some knowledge about that person and often some emotional response to them. What we are seeing here is just a preference for one stimulus over another, and a relatively small preference at that. There is no reason to assume that newborns’ preference for their mother’s face indicates that they assume that this mother is a unique individual. The mother’s face could merely be a collection of stimuli that is attractive to the infant as a result of repeated exposure. Moreover, and alluded to above, the preference is a small one, as research by Bushnell et al. (1989) shows that it is fragile and can be disrupted by superficial changes. It takes children many months before they really respond to familiar faces in a predictable way. Neville et al. (1993) conducted brain-imaging studies on 6- and 12-month-old-infants’ face-processing abilities. It was found that 6-month-olds used a variety of brain regions across both hemispheres for faces, whereas at 12 months processing had become more localised in the right hemisphere, more like that of adults. This suggests that although infants are competent face processors, this ability undergoes much refinement over the months as a result of experience. (It also shows that face recognition is lateralised; see Chapter 11.) Whatever is happening here, it is clear that the mechanism is more complex and extended than one of simply imprinting on an object in the way that chicks do. Despite these differences, the processes share more similarities with imprinting in domestic chicks than most people would have expected. In both humans and domestic fowl there is evidence of an inherent preference which may serve to orient the individual towards members of their own species in general and parents in particular (Dudai, 2002). Furthermore, in both there is now evidence of quite specific early gazing preferences towards the head or facial region of the parent. Acquiring knowledge of what members of your own species look like is only one step on the way to complex social interaction; in the next section we review evidence of how infants begin to learn about people as psychological entities who have emotions, thoughts, goals and desires. Mind-Reading: The Development of a Theory of Mind Humans live, and have always lived, in complex social groups (see Chapters 7 and 8). One advan- tage of this is that groups provide safety from predators, but they also enable collective action that 128 Cognitive Development and the Innateness Issue can achieve so much more than we can when acting alone. Hunting, for example, is one instance of such coordinated action; acting in groups, humans can successfully kill and carry home a larg- er animal than if they were alone. But such coordinated action brings additional problems. Once prey has been killed, how should it be distributed? Some equitable way of sharing is of benefit if group members are to continue to participate in the hunt. But if every member always gets an equal portion, what is to stop a certain member putting a little less effort into the hunt? Hunting is often dangerous and expends a lot of energy and it makes sense for an individual to hold back a little if they are guaranteed the same return as their more enthusiastic companions. Living in social groups, therefore, has a particular set of selection pressures that act on the group members. Not only do in- dividuals need to understand the physical environment, but it would benefit them enormously were they to evolve a way of understanding the actions, intentions and beliefs of others. If an individual could read the minds of others and discover what they were thinking it would give it and its offspring a tremendous competitive advantage. Among other things, it would be able to detect when it was being deceived, and also be able to deceive and manipulate others with less chance of detection. This ability to manipulate and deceive others has been labelled Machiavellian intelligence by pri- matologists Andrew Whiten and Richard Byrne of the University of St Andrews (Whiten and Byrne, 1988). They first used this term to describe the way that chimpanzees were observed to manipulate others to achieve their own ends. Deception and manipulation are, cognitively speaking, complex and demanding processes to engage in successfully. Being honest is rather simple; all you need to do is to report what you know to be the case about any particular state of affairs. Successful deception, on the other hand, requires you to take into account the mental states of the person or people you are attempting to deceive. The child with a chocolate-smeared face who steadfastly maintains that she has not been raiding the biscuit tin fails to deceive because she fails to take into account what her accusing father can see to be the case. Of course, true mind-reading does not and probably could never exist, but approximations of other people’s mental states can be made. The child above might make the assumption that because her face is covered in chocolate and because her father is looking at her face, he therefore has the knowledge that she has been eating chocolate. Facial expressions also leak otherwise hidden emotions and inten- tions into the outside world. Social animals, particularly humans, have learned to pick up on these cues to enable them to represent mentally what is going on in the minds of others, an ability that psycholo- gists refer to most commonly as theory of mind or sometimes mind-reading. The picture of humanity painted above seems rather negative, focusing as it does on decep- tion, exploitation and the ability to detect deception. Negative or not these are important considera- tions in group living. Nevertheless, there are important positive consequences of theory of mind too. Empathy is an important one: if one can appreciate at a visceral level that someone is suffering, one can, should one be sufficiently motivated, act to help them. So theory of mind is important for moral behaviour and is something that permits societies to function as they do. We discuss this further in Chapter 6 when we consider morality. Theory of Mind and False Beliefs The acid test for the presence of a theory of mind is the false belief test. This is simply the knowl- edge that someone can hold a belief which is either different from current reality or different from what you yourself believe to be true. If, and only if, someone passes this test can we be sure that they have an understanding of the mental states of others. The other person’s belief has to be different from our own because otherwise we could pass the test simply by reporting what we knew to be true,  Mind-Reading: The Development of a Theory of Mind 129 A number of tests has been developed to demonstrate that people can deal with false be- liefs, one of these is the deceptive-box test developed by Perner et al. (1987). A child is shown a Smartie container (a tube that usually contains chocolates) and is asked what she thinks is in there, to which the child invariably answers ‘Smarties’. The tube is then opened to reveal that it unexpect- edly contains pencils. The child is then asked what someone else who had not seen the contents of the box would think was inside. The average four- or five-year-old would say, like us, that another person would think that there were Smarties in the tube; however, a younger child (usually under four years of age) would say that the other person would believe that there were pencils in the box. Even though it is made clear that the other person had no prior knowledge of what was in the box. Furthermore, if after the unexpected contents have been revealed, the younger child is asked what she initially thought was in the box she will often say that she thought that there were pencils in the box, even though only a few seconds earlier she had clearly ventured that she thought that it would contain Smarties! Younger children therefore not only have difficulty reporting other people’s thoughts they also have difficulty reporting what they initially thought when it is different from their current knowledge-state. Research has also ruled out alternative interpretations of this phenomenon: it does not seem to be a simple failure of memory, since children failing the Smarties test have no problems with state-change tasks where no false beliefs are involved (Gopnik and Astington, 1988). How Does Theory of Mind Develop? Other similar tests for theory of mind such as the unexpected transfer test (Wimmer and Perner, 1983) and the Sally-Anne test (Baron-Cohen et al., 1985) are all passed at roughly the same age, which some have taken as suggesting that theory of mind develops in a stage-like manner (Box 5.1). However, others have criticised this interpretation (Mitchell, 1996), suggesting that since the test is all-or-none (you either pass it or you don’t), it has the effect of making what might be a gradual de- velopmental profile look like a sudden stage-like shift. An analogy for this might be a driving test in which all you found out was whether you passed or failed, you were given no feedback on how many errors you had made. Imagine that you failed three times before passing on the fourth attempt. Does this necessarily mean that your driving suddenly and drastically improved in a stage-like fashion between the third and fourth attempts? No. More than likely your driving gradually improved over all four attempts, but because you either pass or fail the test, this gradual improvement was hidden until you reached the pass threshold on the fourth attempt. Similarly, because children of four years tend to pass standard theory of mind tests it does not imply that children younger than four, who fail the test, have no theory of mind. As with the driving test example, it is possible that younger children have some knowledge of false beliefs, but find performing the task too demanding cognitively to reveal their nascent understanding. Evidence in support of this interpretation comes from Baron-Cohen (1995) who used the Sally-Anne task. In this task a doll (named Sally) hides a marble in a box and then leaves. Unknown to Sally, a second doll, Anne, comes in and moves the marble from the box to a basket alongside it. The child is then asked where Sally, on her return, will look for the marble – in the box or in the basket. Consistent with the finding from the deceptive-box test, children of about four years will correctly report that Sally will look in the box because, after all, this is where she last saw it. Younger children, however, report that Sally will look in the basket where they know the marble currently resides. However, careful observation of the child’s eyes reveals that they consistently look at the box before reporting the incorrect location. This has been taken as suggesting that younger children have some under- standing of false belief, but the current location, perhaps because of its greater salience, wins out and they answer the question incorrectly. 130 Cognitive Development and the Innateness Issue It is probably a mistake to think that passing the false belief task represents the child cross- ing the rubicon where they go from not having to having a theory of mind. Even before children are able to pass the false belief test they make many attributions that relate to mental states. For exam- ple, a 3-year-old, on being told that a character is looking for a cat and looks under the table, will explain the action by saying that the character wants the cat and thinks that it might be under the ta- ble (Wellman, 1988). Furthermore, there is still much to learn about people and their states of mind once one passes the false belief test. The ‘second order’ false belief task developed by Baron-Cohen (1989) introduces another actor into the frame and asks the child what he or she thinks that per- son X would think person Y would do is passed at an older age than the standard task. Other tests suggest individual differences in theory of mind ability. The ‘reading the mind in the eyes test’ (or just ‘the eyes test’) (Baron-Cohen et al., 1997) requires participants to judge people’s mental states from pictures of their eye regions (e.g. whether the person is ‘aghast’, ‘fantasising’, ‘impatient’ or ‘alarmed’). This test reveals a number of individual differences, perhaps the most interesting being that females are slightly better at it than males. Box 5.4 People with Autism or Autistic People? In earlier versions of this chapter a number of references were made to ‘autistic people’. In revising this chapter for the fourth edition this seemed curiously out of date and maybe even offensive. After all the person first language movement recommends that we don’t identify a per- son with their particular disease or disability. Thus, and quite reasonably, we are asked to refer to ‘people with epilepsy’ rather than ‘epileptics’, ‘people with cancer’ rather than ‘cancer patients’ and ‘people with a disability’ rather than ‘disabled people’. Somewhat surprisingly ‘people or person with autism’ is not the accepted term for the autism community who prefer the ‘identity first’ term ‘autistic person or people’. Their argument is interesting and illustrative. Most people with cancer consider their cancer as a disease; it affects and afflicts them, but it is not part of their identity; it is not who they are. In contrast the argument made by the autism community is that autism is indeed part of their identity it is part of what makes them who they are (perhaps a cen- tral part to some) and is not considered an affliction. You can remove the cancer from the person and they are the same person (barring, obviously their experience with cancer) but to remove the autism from the person would result in a different person. For this edition we follow this identity first recommendation of autistic people or person. We hope this little box-sized digression serves more than an explanation of what might seem, to some, as our use of unfortunate and even insulting language, because the arguments about what we call people strike at the heart as to how people define themselves. You can read more on this debate at http://autisticadvocacy.org/about-asan/identity-first- language/. Is Theory of Mind Modular? The Case of the Autism Spectrum In 2013 the American Psychiatric Association produced a major update of their manual of men- tal illnesses, DSM-5. Somewhat controversially they removed autism and the related condition of Asperger syndrome from its classification system and replaced them both with Autism Spectrum  Mind-Reading: The Development of a Theory of Mind 131 Disorder, ASD for short. There were two main reasons for this. First, it was in recognition that men- tal illnesses are not things that you either have or do not have but exist on a continuum with some people having more extreme versions and others milder forms. Second, that one could have some symptoms of autism and not others and still attain a level where it could be considered a disorder in that it significantly impairs an individual’s ability to function. Although autism and Asperger syndrome were first identified in the 1940s, it wasn’t until the 1980s that headway was made into understanding their underlying cognitive profiles with work by the British psychologists Simon Baron-Cohen and Alan Leslie proposing that some symptoms might be caused by a deficit in theory of mind. A wealth of research has consistently demonstrated that sufferers from autism have diffi- culty passing standard theory of mind tests when compared both to clinically normal children and children who have Down syndrome and Williams syndrome (Baron-Cohen et al., 1985). Children with Down syndrome, for example, seem to have no more difficulty passing a standard test such as Sally-Anne (see above) than normal children when matched for a variety of measures includ- ing mental age (usually verbal ability is used). Autistic children (see Box 5.4), on the other hand, matched using the same criteria will often fail such a task. So a Down syndrome child with a mental age of 5 or 6 will pass standard false belief tests whereas an autistic child is likely to fail them. Autistic children who systematically fail tests of theory of mind can show reasonable per- formance on other tasks such as perspective taking and other tests of visuo-spatial ability. They also seem to have reasonably good physical reasoning ability. This is shown by their good performance on embedded figures tasks (see Figure 5.8) where the goal is to find the target shape embedded in a more complex shape (Jolliffe and Baron-Cohen, 1997). Not all autistic people fail simple false belief tests, many pass them – particularly those with good general intelligence. However, as pointed out earlier, tests such as Sally-Anne are all-or- none tests, you either pass them or you don’t, and they are thus not very sensitive to more subtle variations in theory of mind. All they tell you is that the participant has a theory of mind equivalent to that of a clinically normal four-year-old. More sensitive tests reveal that even high-functioning Autistic people show impaired theory of mind. There is therefore evidence that people with autism and Asperger’s syndrome have im- pairments to their ability to understand the mental states of other people, with some other cognitive Figure 5.8 Embedded figures tests. Can you find the shape on the left in the figure on the right? 132 Cognitive Development and the Innateness Issue abilities apparently spared (this is true especially for people with high-functioning autism). Further support for the modularity hypothesis could be found if there are people who appear to suffer from the mirror image of autism, that is they have preserved theory of mind with impairments in spatial ability. In technical language finding a disorder and its mirror opposite is called a double dissoci- ation. It has been claimed that sufferers of another developmental disorder – Williams syndrome – might represent such a case. Williams Syndrome: The Mirror Image of Autism? Williams syndrome is a comparatively rare developmental disorder affecting around 1 in 10,000 live births. The symptoms include hypersensitivity to noise, premature ageing of the skin and cardio- vascular problems caused by a narrowing of the major arteries, specifically the aorta, which, unless treated, can lead to premature death. Perhaps the most obvious physical characteristic, however, is their facial appearance. People with Williams syndrome tend to have a broad nose, full lips, large ears and sometimes a star-shaped patterning on the iris which have given rise to their colloquial name of ‘pixie people’. Psychologically, they are even more interesting with a profile of impairments characterised by peaks and troughs – some abilities being impaired and others being relatively spared. General IQ is low, a study by Mervis and Becerra (2007) on 306 children with Williams syndrome found IQ scores between 40 to 112 with a mean of 69. Recall that IQ is calculated so that the population aver- age is always 100. In particular, spatial ability is severely compromised with many sufferers finding it difficult to perform everyday tasks such as finding their way around their own house, retrieving a number of items from a cupboard, or tying their shoes. On the other hand, language tends to be well developed compared with their non-verbal IQ, although there is often a preference for ornate and florid words. One individual, when asked to list some animals, included ibex, pteranadon and yak as well as species new to science such as brontosaurus rex (Bellugi et al., 1990). Perhaps most strikingly, particularly in the light of what we have learned about autism, people with Williams syn- drome seem to have highly developed social skills compared with their general intelligence. For this reason, many have described Williams syndrome as the mirror image of autism. The Cause of Williams Syndrome Williams syndrome is a genetic disorder caused by missing genes on one particular chromosome. Chromosomes, recall (see Chapter 2) are just long strands of DNA located in the cell nuclei. During meiosis, they take on the shape of two sausages connected by a narrow region called the centromere. The arms are of different lengths, the long one being known as the q arm, the short one known as the p arm. In Williams syndrome some of the genetic material on the long arm of one of the chromo- some 7 pairs has been deleted in the 11.23 region. These microdeletions, as they are known, mean that some important protein-assembling genes are absent. Currently, it is estimated that at least 20 genes are deleted. These include ELN which codes for the structural protein elastin which, as the name suggests, gives elasticity to many of the body’s organs (including the skin and major blood vessels). The deletion of ELN – on one of the chromosome 7 pairs, there is a working version of ELN on the other copy of chromosome 7 – leads to a deficiency in elastin which explains the pre- mature ageing of the skin and some of the cardiovascular problems suffered in Williams syndrome. Other deleted genes include GTF2I which may affect IQ, CYLN2 which is expressed in the cere- bellum and LIMK1 which codes for protein, LIM-Kinase 1 which is associated with visuo-spatial ability (Frangiskakis et al., 1996).  Williams Syndrome: The Mirror Image of Autism? 133 Williams Syndrome: A Modular Account A simple modular account might propose that sufferers of Williams syndrome have selective dam- age to the mental module(s) responsible for spatial cognition, with little or no impairment to those involved in social cognition (such as theory of mind). There is some evidence to support this. For instance, Karmiloff-Smith (1997) found no difference between Williams syndrome and clinically normal individuals on tasks measuring face-processing ability. Tager-Flusberg et al. (1998) found that Williams syndrome sufferers perform better on ‘the eyes test’ than people with Prader–Willi syndrome matched for mental age. (Prader–Willi syndrome is a developmental disorder character- ised by low IQ but no specific cognitive deficits.) This last study is important because it suggests that people with Williams syndrome are better at mental state judgements than would be expected given their mental age. Sadly for those who like things nice and simple the true picture seems to be rather more complicated. For instance, research by Tager-Flusberg and Sullivan (2000) suggests that although people with Williams syndrome seem to have preserved social perception abilities they appear to have impairments of social cognition. So although they are good at recognising faces, and identify- ing expressions of emotional and other mental states (such as in the eyes test), they are no better than would be expected given their mental age at reasoning about mental events and causes. For example, one task asked children to explain why a character was crawling around wearing a leopard suit and growling, children with Williams syndrome were no more likely to get the correct answer, that he was pretending to be a leopard, than mental age-matched controls. Such evidence has been used to suggest that the social module might consist of a number of sub-modules, one of which deals with social perception and is preserved in Williams syndrome, another of which deals with social cognition that is impaired. This view is entirely consistent with the view of evolutionary psychologists (see Barkow et al., 1992, 599). Evidence against a Modular Account of Williams Syndrome There is other evidence, however, which appears to contradict the theory of innate modules. Re- search suggests that although children with Williams syndrome are good at recognising faces, they seem to do this in a different way to clinically normal individuals. It is well known that clinically normal individuals are particularly bad at recognising upside down faces, unless the face has some- thing very distinctive about it such as a beard or an unusual hairstyle. This seems to be something peculiar to faces, since we have much less trouble identifying other inverted objects such as types of car or breeds of dog. Karmiloff-Smith (1997) found that people with Williams syndrome were better at identifying inverted faces than were clinically normal controls. Further research suggests that this is because sufferers from Williams syndrome use a different strategy for identifying faces than non-sufferers do. When clinically normal people recognise a face they seem to process it ho- listically – somehow the identity of the face emerges as a gestalt – whereas people with Williams syndrome appear to process faces on a feature-by-feature basis. Using imaging techniques Karmi- loff-Smith et al. (1998) found that individuals with Williams syndrome used different brain regions when processing faces than normals. A similar story can be told for the spared language processing ability of people with Wil- liams syndrome. Karmiloff-Smith et al. (1998) found that, as well as the preference for baroque words, they have other linguistic abnormalities. For example, they find it difficult even to repeat sentences with embedded relative clauses such as ‘the boy the dog chases is big’ (where ‘the dog 134 Cognitive Development and the Innateness Issue that Williams syndrome is characterised by a preserved language module. Karmiloff-Smith argues that this research shows that development is a more complex process than is given in the modular account and advances an alternative account known as neuroconstructivism. Is Theory of Mind an Innate Module? Is It Domain- Specific? As we discussed above the word innate can be interpreted in different ways. If we assume that innate means ‘present at birth with no experience necessary’ then theory of mind is not innate in this sense. For example, we know that theory of mind develops over time and experience seems to be important in the way that it develops. Ruffman et al. (1998) present evidence that children with older siblings tend to pass false belief tests at a younger age than either children without siblings or children with younger siblings. A possible explanation for this finding is that children with older siblings might be forced into situations where there is more need to think about the mental states of others. Younger siblings are often in competition for resources such as food, toys and parental attention, and the abil- ity to get what you want by deception or other forms of manipulation is likely to be advantageous. Furthermore, Heyes and Frith (2014) discuss evidence that there may be two types of theory of mind, implicit theory of mind refers to the ability of individuals to automatically recognise that others can have beliefs that are different from their own, in one study, using a preferential looking task, as young as 7 months (Kovács, Téglás and Endress, 2010). Explicit theory of mind, on the other hand, is the ability to think about, reflect upon and discuss the thoughts, feelings and beliefs of others which is largely what is being picked up in the false belief tests described above. The explicit part of theory of mind, in particular, seems to be sensitive to environmental and specifically cultural conditions. Explicit theory of mind seems to develop later in those cultures such as Samoa where discussing the ins and outs of other people’s mental lives is considered discourteous (Mayer and Träuble, 2013). Rather like Johnson and Morton’s Conspec and Conlern systems, Heyes and Frith suggest that the early-developing implicit theory of minds is produced by genetically specified neurocognitive ma- chinery that is specific to humans and contains domain-specific rules. This acts as a ‘start-up-kit’ that allows explicit theory of mind to develop, but, crucially, they suggest, explicit theory of mind develops through discussing mental states with others through a process of domain-general learning. In fact, Heyes and Frith even go as far as to describe developing explicit theory of mind as akin to the process of learning to read, slow, inefficient and requiring lots of practice: a product of culture rather than one of evolution, albeit with an evolutionary basis (implicit theory of mind). So despite the initial excitement that autism and Williams syndrome evidence for the se- lective impairment of domain-specific, innate modules the weight of evidence currently seems to be that they do not, at least not in the sense of the massive modularity favoured by the Santa Barbara school. Neither is the evidence consistent with a wholly domain-general approach with minimal innate knowledge. The true answer may lie somewhere between these two extremes, with one pos- sibility being neuroconstructivism. Rethinking Modularity and Domain Specificity – Neuroconstructivism and Domain-Relevant Learning The developmental psychologist Annette Karmiloff-Smith points out that much of the evidence for modularity is based on the abilities of either adults or older children; we find that an individual can’t perform a particular task and assume that an innate module must be impaired. As we have seen,  Summary 135 however, experience is essential for the development of a fully functioning theory of mind and the same point could be made for abilities such as language and vision. This is not an argument for old- style Blank-Slateism, however: the wiring diagram of the adult human brain contains nearly 100 bil- lion neurones and 100 trillion synaptic connections, which is impossible to completely specify with a mere 20,000 genes or so. Instead of innate, domain-specific processes Karmiloff-Smith (1994; 2015) suggests that there may be domain-relevant processes, but these are low level, pertaining to perception rather than high level and become increasingly domain-specific through experience. To understand this, think of Conspec and Conlern: Conspec is low level, subcortical, and innate and by directing attention towards face-like stimuli enables Conlern to acquire more detailed knowledge of faces. Conspec, she argues, is not truly domain-specific, remember that infant humans are equally interested in human faces and monkey faces and vice versa so Conspec is relevant to faces but not specific to human or monkey faces in either species. Only through experience does Conspec become domain-specific with the help of Conlern and, so long as infants and monkeys are seeing members of their own species, they quickly learn to express a preference. While this view might undermine the notion of massive modularity, it does not undermine the notion that behaviours such as those discussed above evolved. In a sense it doesn’t matter how you arrive at your destination (e.g. being able to understand language, or recognise faces) what mat- ters is that you arrive there. Having domain-specific may well allow you to arrive at your destination more quickly but there are benefits in the more flexible approach suggested by neurocontructivism. Anthropologist Eric Smith (2007) argues that our ancestors faced radically changing environments throughout their evolutionary history including many ice ages. These environments changed so quickly (relatively speaking) that it was impossible for natural selection to keep pace with them. Instead natural selection favoured neuronal flexibility rather than a fixed architecture to enable our ancestors to change their behaviour in order to survive and reproduce. This is a topic that we explore further in Chapter 14 when we discuss culture. Summary Infants seem to have certain expectations of the way that the physical world operates. For example, infants of at least three months of age have the knowledge that objects exist independently of their ability to perceive them. A number of other physical principles seems to develop in the first few months as if supported by innate learning mechanisms. Infants have preference for face-like stimuli from birth and learn the details of human faces rap- idly. Again, it seems that this is the result of innate learning mechanisms rather than innate knowl- edge per se. Young children have an understanding of the role of mental states as a cause of behaviour, despite the fact that these states cannot be seen. This skill becomes more sophisticated as children devel- op. This ability, known as theory of mind or mindreading, is measured by a number of tasks such as false belief tasks and more recently, the eyes test, in which participants are required to judge how people feel from looking at their eyes. Some argue that disorders such as autism and Williams syndrome, which have cognitive profiles characterised by peaks and troughs, provide evidence for selective impairment of mental modules. Further research shows that such selective impairments are not as simple as they first appear. 136 Cognitive Development and the Innateness Issue For example, people with Williams syndrome, although appearing to have intact face-processing abilities, seem to process faces in a different way from clinically normal people. Karmiloff-Smith proposes a theory of neuroconstructivism that emphasises the importance of learning and a gradual process of modularisation rather than the existence of innate mental modules. Although denying the existence of innate modules, this approach is certainly not incom- patible with an evolutionary approach to cognitive development. Questions 1. How might our ‘intuitive physics’ make learning school physics difficult? How might we over- come these difficulties? 2. To what extent does research showing that non-human primates are as good at recognising human faces as faces of their own species and vice versa suggest that the mechanisms for face recognition evolved in a common ancestor, rather than being uniquely human? 3. To what extent are the claims that differences between people are innate independent from the claim that similarities between people are innate? 4. According to research what is the role of experience in developing a fully functioning theory of mind? Think for yourself about the kinds of experiences which might support this development. Further Reading Baron-Cohen, S. (1995). Mindblindness: Essays in Autism and Theory of Mind. Cambridge, MA: MIT Press. An excellent introduction to the study of autism as a deficit of theory of mind. Gopnik, A., Meltzoff, A. N. and Kuhl, P. K. (1999). How Babies Think. London: Weidenfeld and Nicolson. A good, popular introduction to recent research on infants. Karmiloff-Smith, B. A. (1995). Beyond Modularity: A Developmental Perspective on Cognitive Science. Cambridge, MA: MIT Press. Ridley, M. (2003). Nature via Nurture: Genes, Experience, and What Makes Us Human. London: Fourth Estate. A popular introduction to the genome and way that it interacts with the environment through development.