Show Summary Details

p. 41. Babies and what they knowlocked

  • Usha Goswami


‘Babies and what they know’ explores the early development of children both before birth and during the first stage of childhood, focusing in particular on the importance of interaction. The child starts to learn even before it is born. After birth, babies are capable of learning a great deal from facial cues and the sound of language. They seem fascinated by the human face. This intrinsic interest in faces and eyes has been linked to how we acquire language. Studies on imitation, joint attention, and socio-moral expectations have shown that infants and toddlers start to develop psychological understanding early on.

Early learning

The baby’s brain starts to learn inside the womb. By the third trimester (months 6–9), the infant can hear their mother’s voice. Indeed, despite the filtering effect of the amniotic fluid, at birth babies can distinguish their mother’s voice from that of a strange female. This was shown by a famous ‘sucking’ experiment, in which newborn babies were given a dummy to suck. First, their natural or ‘baseline’ sucking rate was measured. Next, the infants were played a tape recording of their mother reading a story. Each time their suck rate increased above baseline, the tape would play. Each time the suck rate dropped below baseline, a strange female voice would be heard instead, reading the same story. The infants rapidly learned to suck fast to hear their mother’s voice. The following day, the experimenters reversed the contingency. Now slower sucking was required to hear their mother’s voice—and the babies reversed their suck rates. Similar ‘sucking’ experiments have been done using story reading. Mothers read a particular story every day to their ‘bump’ during the final trimester. At birth, the infants could distinguish the familiar story from a novel story. Indeed, sucking experiments have even shown that infants can learn music inside the womb. Infants whose mothers were fans of the soap opera Neighbours were able to recognize the Neighbours theme tune at birth.

p. 5Babies also move a lot inside the womb. Even by the 15th week, the foetus can use a number of distinct movement patterns. These include a ‘yawn and stretch’ position and a ‘stepping’ pattern used for self-rotation. Aspects of the intra-uterine environment, such as the regular heartbeat of the mother, also seem to be learned, and can subsequently have a soothing effect. Studies have shown heartrate deceleration in the foetus in response to certain sounds (thought to index attention). They have also shown habituation (lack of change) of heart rate to familiar stimuli, thought to indicate learning. Therefore, foetal studies show that the infant brain is already learning, remembering, and attending, even inside the womb.

Most of the brain cells that comprise the brain are actually formed before birth. Because of this, if the environment inside the womb contains excessive toxins, such as excessive alcohol or drugs (i.e. alcoholism or drug addiction, not the odd glass of wine), this will affect brain development. These effects are not reversible, although the environment that is experienced following birth can remediate negative effects to some extent. The brain itself remains plastic throughout the life of the individual. However, plasticity is largely achieved by the brain growing connections between brain cells that are already there. Any environmental input will cause new connections to form. At the same time, connections that are not used much will be pruned. Therefore, no single experience can have disastrous developmental consequences. On the other hand, the consistency of what is experienced will be important in determining which connections are pruned and which are retained. If a child consistently experiences warmth and love, different connections will be strengthened compared to a child who consistently experiences anxiety and fear.

One reason that the early years are so important is that the brain is in effect a machine for learning. The brain cells are ready to go even before birth. They also have certain inbuilt ways of processing information, which research has uncovered. The early capability p. 6of this learning machine determines the efficiency of later learning. At birth, brains are not so different. They all have the same inbuilt ways of processing information. However, a brain that is offered early advantages can develop its early architecture more efficiently. So a brain that is born into an optimal learning environment will do better over a person’s lifetime than a brain that is faced with a less optimal early environment. On the other hand, the malleability of the early brain means that interventions can always have an effect. Interventions, such as changing an environment (e.g., by being fostered), can support the strengthening of the connections that process information more optimally or efficiently. This is particularly the case during the first few years of life. So improving a child’s environment will always have a positive effect on later developmental outcomes.

Finally, there are always individual differences between brains. At birth, individual differences are primarily in sensory processing (such as the efficiency of neural mechanisms for seeing or hearing) and lability (fussiness, or speed to respond to stimuli). It is the environment that will determine whether these individual differences, which characterize all of us, have trivial or more measurable consequences. Brains whose biology makes them less efficient in particular and measurable aspects of processing seem to be at risk in specific areas of development. For example, when auditory processing is less efficient, this can carry a risk of later language impairment. Individual differences are not deterministic: less efficiency does not automatically mean a developmental impairment. Usually, as long as the environment is rich enough, sufficient learning experiences enable less efficient brains to reach similar developmental end points to more efficient brains. However, early awareness of impaired efficiency can enable useful environmental interventions. An extreme example is deafness. In some cases of deafness, a small microchip (called a cochlear implant) can be inserted close to the ear in infancy, and for some cochlear implant children, oral language development is then as good as for hearing children.

p. 7Further, even identical twins who developed from the same egg will have different brains. It is not yet clear why this is. One possibility is that these individual differences may depend on the environment inside the womb, which will be experienced slightly differently by each baby. One twin is usually dominant, so perhaps always changes their position first. The other twin must then accommodate this change by changing their position. So the second twin has a different intra-uterine experience to the first twin. Nevertheless, identical twin research shows that biology is never destiny (see Chapter 7). The environment will always have a major impact on child development. The environments in which children develop are governed by their families, their nurseries and schools, and wider society.

People, faces, and eyes

Other people are by far the most interesting things in the world of the baby. Research shows that infants are fascinated by faces from birth. Indeed, there is a specialized brain system for face processing, which seems to function in the same way in infants and adults. Experiments with neonates and young infants show that faces are always preferred over other stimuli, particularly live, mobile faces. The eyes are especially interesting. Newborn infants prefer to look at faces with eyes gazing directly at them. They dislike looking at faces with eyes that are averted. Babies also react negatively to a ‘still face’—an experimental situation in which the mother deliberately suspends interaction with the baby and just looks blank. Presented with a ‘still face’, babies become fussy and upset and look away. The ‘still face’—maternal unresponsiveness—also causes elevated levels of the stress hormone cortisol for some babies. Mothers who are clinically depressed show features of ‘still face’ behaviour.

Some psychological theories argue that social rejection is the most powerful form of psychological suffering. For example, Philippe Rochat (2009) suggests that infants’ upset reaction to a ‘still face’ p. 8is early evidence for the importance of social interaction in forming a concept of the self. According to these ‘socio-cultural’ theories of child development, our ‘self’ is defined by how others react to us. If others are positive to us and interact warmly with us, we feel good about ourselves. If others are hostile or ignoring, we feel bad about ourselves. Socio-cultural theories also argue that our need to avoid social separation and rejection (e.g., avoiding bullying and punishment) determines much of our behaviour. We all need social proximity and intimacy. The forms of proximity and intimacy offered by our environments determine how we think about who we are.

Newborn infants will also imitate facial gestures made by adults. This suggests that babies are intuitively aware of their own bodies. Babies are also immediately engaged by what other people are doing. In one famous study carried out in a maternity hospital (see Figure 1), infants aged from 1 hour old imitated an adult who was either opening his mouth or sticking out his tongue. To test imitation experimentally, the infants were tested in a dark room with their mothers. A light would come on and illuminate the experimenter’s face. He either stuck out his tongue, or opened his mouth wide. After 20 seconds, the light went out, and the infant was filmed in the dark. The infants stuck out their tongues more after watching tongue protrusion, and opened their mouths wide more after watching mouth opening.

The capacity to imitate is probably underpinned by a brain system called the ‘mirror neuron’ system. ‘Mirror neuron’ brain cells are involved in matching actions to feelings. Mirror neurons are active both when you are watching someone else do something, and when you are doing the same action yourself. For example, the same brain cells are active if you pick up a stick, or if you watch someone else pick up a stick. Therefore, this brain system is thought to be one basis for a shared or ‘common code’ between the self and others. In order to copy an observed gesture in the dark, babies must be able to map the actions of someone else onto their p. 9

1. Even new-born babies can imitate adult facial gestures

own bodies. This means that they are recognizing that another person is somehow ‘like me’. This was shown experimentally by demonstrating that infants do not imitate actions made by animated robots. They are apparently aware that a robot is not a human agent. Older infants also imitate people’s actions that were never completed successfully. For example, a baby might watch someone trying to put a string of beads into a jar, and yet always missing the jar’s opening. When the baby is allowed to play with the beads, they are put straight into the jar. This shows that babies are aware of the intentions of other humans. They are not simply imitating their exact physical movements.

Communicative intentions

Intrinsic interest in people’s faces and eyes has been linked to how infants are able to acquire language. The argument is that the p. 10ability to recognize intentions is of central importance. When we speak to someone, we intend them to understand our meaning. Newborn babies have a number of abilities that help them to recognize the ‘communicative intentions’ of others.

First, babies like direct eye gaze. Even for adults, when someone is looking directly into your eyes and establishing eye contact, this is a signal that you are both ‘on line’ for talking. Secondly, babies can take turns. All conversation involves turn-taking, and breast feeding is the prototypical turn-taking experience. Both breastfeeding and bottle feeding are characterized by the infant sucking and stopping. When the baby stops, the mother jiggles the infant, who then starts sucking again. Pausing is not dependent on needing to breathe or on being full—babies could suck continuously if they wanted to. And jiggling never occurs while the baby is sucking. In fact, the research shows that jiggling doesn’t affect the total amount of milk that is taken. Nevertheless, sucking and jiggling are done in turns. This is like the turn-taking pattern of human conversation.

Thirdly, infants are able to detect contingencies. They are aware from very early that some events are intrinsically related (or contingent upon) each other. Sucking and jiggling during breast- or bottle-feeding is one example of ‘contingent responsivity’. Each action is contingent upon the occurrence of the other action. Contingent responsivity is an essential property of human interaction, and is an important concept in child psychology. Contingent responding by caretakers promotes healthy psychological development. There are many examples of infants’ recognition of contingency. For example, one clever study filmed babies kicking (young babies spend a lot of time kicking). The babies were then given a choice of two films to watch. One film showed their own legs kicking in real time. The other film also showed their own legs, but with a time delay. So in the second film, there was no contingency or intrinsic link between what the infants could feel their legs doing, and what they could p. 11see. The experimenters found that babies preferred to watch the contingent video.

Finally, carers tend to use a special tone of voice to talk to babies. This is more sing-song and attention-grabbing than normal conversational speech, and is called ‘infant-directed speech’ or ‘Parentese’. All adults and children naturally adopt this special tone when talking to a baby, and babies prefer to listen to Parentese. For example, when given a choice between listening to tapes of an adult speaking versus the same adult speaking in Parentese, babies will choose to activate the tape that uses Parentese. These are the four critical abilities that seem to be foundational for acquiring language. When someone is gazing directly at you, speaking in Parentese, responding to your gurgling and taking turns with you, these are all signals that you are being intentionally addressed. And they are signals recognized by babies from birth.

Attachment and security

The ability to recognize communicative intent is only one of the intrinsic abilities (or predispositions) towards social interaction that is present in newborn babies. Other innate behaviours such as rooting for the breast, crying and grasping all create proximity to the caretaker. These actions ensure the physical closeness required to build a relationship. Some research suggests that the pitch and amplitude (loudness) of a baby’s cry has been developed by evolution to prompt immediate action by adults. Babies’ cries appear to be designed to be maximally stressful to hear. Smiling also begins early, and infants use smiles to reward social interaction from carers. Experiments show that infants smile most during face-to-face contingent interactions with carers. These are interactions characterized by turn-taking, infant-directed speech and playful warmth (‘interpersonal contingency’). At birth infants prefer the mother’s voice and the mother’s smell, as these are most familiar. However, the important factors in becoming a ‘preferred attachment figure’ are proximity and consistency. Babies quickly p. 12learn to prefer the faces, voices and smells of their most consistent and warm caretakers. These specific attachments that babies form are very important for healthy psychological development.

Nevertheless, research does not suggest that separation from the mother following birth (for example, for a medical procedure) prevents ‘bonding’ with the infant. The psychological relationship or ‘bond’ that mothers and other caretakers form with infants grows over time. Consistency of contact, responsiveness and warmth are the key attributes. The consistency of early attachment experiences are critical for the development of children’s ‘internal working models’ (psychological expectations) of their value as a person who is deserving of love and support from others. If these interactions are characterized by consistency and warmth, the baby is described as showing ‘security of attachment’. If an infant consistently experiences caretaking that fails to be contingent on their needs, or that is not characterized by warmth, then the attachment is said to be ‘insecure’. Similarly, if an infant consistently experiences caretaking that is erratic and neglectful, so that sometimes caretaking is contingent on their needs and sometimes it ignores those needs, attachment is also insecure. Infants who are insecurely attached to their caregivers still prefer those caregivers over other people. The term ‘insecure attachment’ refers to the fact that the infant cannot rely on those caregivers responding appropriately to their cries and smiles—or responding at all.

When attachment is insecure, children develop different ‘internal working models’ of the self. Two main types of insecure attachment are identified in the literature. ‘Insecure-avoidant’ infants appear to become resigned to their fate. They develop self-protective strategies, such as not seeking contact when the carer is close, as though to protect themselves against disappointment. ‘Insecure-dependent’ infants become very clingy and fight against separation, as though trying to force appropriate caretaking behaviours from the adult. Research shows that both forms of insecure attachment are related to less positive developmental p. 13outcomes long-term. These include social–emotional outcomes, relating to self-esteem and self-control, and also cognitive outcomes, relating to intellectual and academic achievement.

In extreme cases, usually involving parental reactions that are frightening for the infant, attachment is ‘disorganized’. Caretaking is so unpredictable that an infant cannot find a way of organizing her behaviour to get her needs met. The internal working model developed in response to such caretaking is often that the child is flawed in some way, and does not deserve love and support from others. Such children are at risk for mental health disorders, including depression, oppositional–defiant disorders or conduct disorders. Healthy attachment relationships do not have to be with the genetic parents. Relationships depend on learning. Learning that your social overtures will be met with contingent responsiveness and warmth are the key factors required for babies to develop secure attachments. Grandparents, foster parents and older siblings can all be sources of secure attachments.

Expectations about behaviour

As well as being predisposed to be social, babies also have interesting expectations about how people should behave to each other. Research shows that older babies (around 12 months) expect people to help others and to behave fairly. For example, in one study infants were shown videos of animated geometric shapes on a computer screen. The shapes moved in certain ways. The screen had a line rising across it, which could be seen as a ‘hill’ to be ‘climbed’. One shape (a circle) duly began to move up this hill. At first it climbed steadily, but then the incline steepened, and the circle rolled back to rest on a plateau. At this point in the video, one of two other shapes (visible at the top of the screen) began to move. For example, a triangle came down to a position behind the circle, and then they moved together to the top of the hill (‘helping’ scenario). Alternatively, a square came down to a p. 14position in front of the circle, and they moved together back to the bottom of the hill (‘hindering’ scenario). Babies were then allowed to choose between 3D soft toys in the shape of the triangle and the square. The babies all preferred to play with the triangle.

This is only one example of a remarkable series of experiments using moving shapes without eyes that are nonetheless interpreted by watching infants as behaving socially. Indeed, adding eyes to the animations, or using real actors, only enhances the experimental effects. Certain types of motion-based interactions between objects seem to specify social behaviour to the infant brain. Socio-moral expectations in babies can also be revealed by experiments using real objects on a small ‘stage’. Infants watch the scenes on stage while sitting on a parent’s lap, and hidden experimenters manipulate the ‘behaviour’ and ‘experiences’ of the objects. For example, one experiment on socio-moral expectations involved two identical toy giraffes. Each giraffe had a place mat in front of them. As the infant watched, the experimenter showed the giraffes two toys (saying excitedly ‘I have toys!’). The giraffes became excited in turn—via hidden means, they began dancing, and shouting ‘yay, yay’. The experimenter then either put both toys on the mat in front of one giraffe, or gave the giraffes one toy each. The giraffes looked down at their place mats without reacting. The experimenters recorded how long the babies watched each scene. The babies looked significantly longer at the event which was unfair.

These experiments, which have various control conditions to rule out other explanations of infant’s choices or looking times, suggest that some socio-moral norms may be innate and culturally universal. Early emerging norms appear to include a concern for fairness, a preference for helping over hindering, and a distaste for actions that harm others. Theoretically, it is thought that socio-moral expectations are inborn because they have evolved to support the continued existence of the species. They are necessary for social group living (society) to work. Socio-moral expectations facilitate p. 15positive interactions between people and foster co-operation within social groups.

Clearly, these norms will be elaborated in different ways by different cultures. Further, once language is acquired (see Chapter 3), we can explain to children how they should behave in certain settings, and why certain moral norms are important. However, even the preverbal infant is learning a lot about socio-moral norms by watching the interactions of those around them. This learning appears to be guided by their innate expectations of how people should behave.

The experiences that promote the learning of socio-moral norms clearly overlap with the experiences supporting secure attachment. The main learning environment for both types of experience in early life is family interactions. Families in which people have warm and supportive relationships with each other are also likely to be families where there is fairness, helping of each other, and little punitive behaviour. Families in which people have hostile and abusive relationships with each other will present babies with learning experiences that seldom model fairness and helping each other, and may actively model aggressive behaviour. In family settings involving physical abuse, infants may learn to inhibit their innate socio-moral expectations and replace them with other expectations about how people behave to each other.

There is little research on the effects of these more negative learning environments, partly as it is more difficult to involve such families in research. Importantly, these influences on learning will be present pre-verbally, and so their effects will be subliminal. The developing infant has no means at their disposal of explaining to themselves the features of their environment. Rather, the environment of the family is their norm. The early learning environment offered by the family has profound effects on psychological and social development, on intellectual development and on the internal working model of the self.


When infants watch the actions of other people, they also learn about psychological causation. The ability to imitate the actions of others provides a ‘like me’ analogy. Brain systems like the mirror neuron system do not simply link seeing and producing certain acts. They also link ‘that looks the way this feels’. Babies seem to assume that people have goals. If babies can understand someone’s goal, they can also imitate an unsuccessful action. An example is the ‘putting the beads into the jar’ experiment discussed earlier. Babies can also imitate selectively depending on the context in which an action occurs. In a famous study, babies watched adults making a completely novel action which they had never seen before. This was to use the forehead to turn on an experimental light panel. The adults made the panel light up by bending forwards and touching it with their heads. Subsequently, the babies were given the light panel to play with. The babies also leaned forward and lit it by using their foreheads. However, babies who watched the same event presented in the context of an adult who was feeling cold, and who had her hands under a shawl, did not use their foreheads to activate the light panel. They just pressed it with their hands. This experiment and others suggest that preverbal infants can infer psychological goals.

Experiments like these show that infants can recognize other people’s intentions, and will imitate accordingly. Further, infants will not imitate accidental actions. Babies also discriminate between an actor who intends to give them a desired toy, but fails (because she cannot get it out of its box), and an actor who can get the toy out of the box, but chooses not to give the toy to the baby. This suggests an insight into the hidden mental states governing the adult’s actions (importantly, the baby does not get the toy in either case). The babies reached out for the toy more and banged the table in frustration more when the actor chose not to pass over the toy than when the actor was unable to p. 17give them the toy. This shows an emergent understanding of psychological causation. Furthermore, the recognition of what people intend is important for effective learning. If babies only imitate the intentional acts of others, then they will acquire many significant cultural skills.

Joint attention

The imitiation experiments suggest that infants are not ‘mind blind’. Infants are not unaware of the hidden mental causes of people’s actions. Although some researchers still dispute this, it appears that babies make mentalistic assumptions that the actions of other people have psychological causes. Babies also interpret seeing as an intentional act. From a young age, babies will follow gaze. They seem to be aware that we look at things because we are getting information about them. Babies able to crawl, who see an adult gazing excitedly at something hidden by a screen, will move to a position where they can see what is being looked at. By around 8–10 months of age, preverbal babies will also try and direct the attention of someone else to something interesting. They do this by pointing at it. Developmentally, there are two kinds of pointing—pointing at something because you wish to be given it, and pointing at something simply to share attention with someone else. The second kind of pointing, which becomes very frequent from around 10 months onwards, is intended by the baby to influence the mental state of another person. This kind of pointing has communicative intent. Interestingly, the absence of this kind of pointing by babies is an early indicator of risk for autism.

Joint attention to an object is a key developmental advance. It is a clear indicator that infants are aware of the mental states of other people. Joint attention is about communicating and sharing experiences. Theoretically, when the child points in order to engage someone’s attention, the child is sharing his or her psychological state with that person (‘I am interested in this’). The child is also showing p. 18an awareness of normative behaviour (‘This is the kind of thing we like to share psychologically, because we are in the same group’).

Joint attention has also been suggested to be the basis of ‘natural pedagogy’—a social learning system for imparting cultural knowledge. Once attention is shared by adult and infant on an object, an interaction around that object can begin. That interaction usually passes knowledge from carer to child. This is an example of responsive contingency in action—the infant shows an interest in something, the carer responds, and there is an interaction which enables learning. Taking the child’s focus of attention as the starting point for the interaction is very important for effective learning. Of course, skilled carers can also engineer situations in which babies or children will become interested in certain objects. This is the basis of effective play-centred learning. Novel toys or objects are always interesting. Objects in which adults show consistent interest are also interesting to infants—for example, car keys and mobile telephones!

Babies will also monitor adults or other children for signs to tell them how they should react to something novel. Again, this suggests an insight into the mental states of others. Such monitoring, called ‘social referencing’, has been shown in various experimental settings. For example, one set of experiments used a mildly scary mechanical robot called ‘Magic Mike’. The experiment used a joint attention setting, in which mothers were trained to look either frightened or happy when Magic Mike appeared. The mothers also used an appropriately emotional tone of voice to say either ‘How frightful’ or ‘How delightful’ (the phrases were deliberately chosen to sound similar but be unfamiliar). In the fearful setting, infants did not approach Magic Mike and also became upset themselves. In the happy setting, infants behaved no differently compared to a ‘neutral face’ setting, in which mothers were neutral. In both the happy and neutral cases, the infants approached Magic Mike and played with him.

p. 19

2. A baby on the visual cliff

The most famous social referencing studies in child psychology began in the 1960s and involved a ‘visual cliff’, shown in Figure 2. Babies who could crawl were placed on a Perspex table top. The Perspex surface covered a design of black and white squares, which were on a surface much lower than the Perspex in the horizontal plane. The squares varied in size to manipulate visual cues to depth, giving the visual impression of a sudden drop. The experiments showed that babies would crawl to the edge of the ‘cliff’ and then look at their mothers for guidance. If the mother looked fearful, most of the babies went no further.

The social brain

These varied experiments on imitation, joint attention, and socio-moral expectations show that infants and toddlers are developing psychological understanding from very early indeed. There is emerging awareness of the mental states of others much earlier than was traditionally believed. Freud, for example, p. 20thought that infants were not aware of the self–other distinction. He argued that physical birth was not the same as psychological birth. Some types of experiment seemed to reinforce these classical ideas of early ‘mindblindness’. Behaviour with mirrors, for example, does not always suggest that young children are aware that it is them in the mirror. On one hand, experiments show that children as young as 3–5 months will test out mirrors by moving different parts of their bodies and seeing what happens in the mirror. On the other hand, it is not until the age of 18 months that children will pass the ‘mark test’. In the mark test, a red mark is surreptitiously put on the child’s face. The test is whether children will touch the mark when they see themselves in the mirror. Many children do not. This behaviour may suggest that consciousness of the self develops gradually. However, it may also be linked to the experimental situation, and as yet remains poorly understood.

A critical factor in early developing psychological awareness is the fact that the infant’s caretakers treat him or her as a social partner. Indeed, carers probably treat infants as acting socially even before infants are deliberately acting in this way. The human brain is a social brain, because humans are a social species. Babies are innately predisposed to be with other humans and maintain social closeness.

As already noted, high quality care does not have to be from the biological parent to be effective. This is further supported by research on daycare settings. Recent studies have employed an objective measure of stress, the hormone cortisol, to assess how much stress preverbal children experience in different types of care settings. Cortisol is measured in the saliva, and higher levels of cortisol are associated with higher levels of stress. Children in high-quality learning environments show lower cortisol levels in these studies, whether they are cared for at home or in a nursery. In high-quality daycare settings, carers provide focused attention and warm stimulation that is essentially sensitive parenting p. 21(i.e., warm and responsively contingent). In such settings, cortisol levels are low. In low-quality learning environments, where carers are intrusive and over-controlling and lack warmth, cortisol levels are high. Again, this is true whether the carer is the biological parent, a nanny or a nursery care assistant.

Accordingly, good teachers, nannies, and daycare providers act as figures of secure attachment. They provide early learning environments that are very similar to high-quality home environments. The best of all worlds is to experience a high-quality learning environment both at home and at nursery. Recent studies suggest that children in high-quality daycare who also have secure attachments to their families show the optimal cortisol profile when in their daycare settings.