Cognition

Clearly, infants are capable of learning in a variety of ways. But do they actually think? This is a question that has intrigued parents and developmental psychologists alike. Baby Benjamin's parents have no doubt looked with wonderment at their child, asking themselves, “What is he thinking? Is he thinking?” Developmental scientists have been working diligently to find out to what extent infants engage in cognition (knowledge, thought, reasoning). The resulting explosion of fascinating research has established that infants' cognitive abilities are much more impressive than previously believed, although the nature and origin of these impressive skills is a matter of considerable debate. Theorists of cognitive development vary with respect to the relative roles they attribute to nature and nurture, especially in terms of whether development is guided by innate knowledge structures and special–purpose learning mechanisms or by general learning mechanisms relevant to experiences in all domains.

So once again, the primary debate is between nativists and empiricists. Some nativists argue that infants possess innate knowledge in a few domains of particular importance (Carey & Spelke, 1994; Gelman, 2002; Gelman & Williams, 1998; Scholl & Leslie, 1999; Spelke, 2000; Spelke & Kinzler, 2007). As you will see in Chapter 7, for example, these nativists maintain that infants are born with some knowledge about the physical world, such as the fact that two objects cannot occupy the same space, and that physical objects move only if something sets them in motion. They also propose that infants possess rudimentary understandings in the domains of biology and psychology. Other nativists emphasize specialized learning mechanisms that enable infants to acquire this kind of knowledge rapidly and efficiently (Baillargeon, 2004; Baillargeon, Kotovsky, & Needham, 1995). According to empiricists, infants' mental representations of the physical world are gradually acquired and strengthened through the general learning mechanisms that function across multiple domains (Munakata et al., 1997). The details of this debate are examined in detail in Chapter 7 with respect to conceptual development. In the following sections, we examine findings regarding infants' cognitive abilities and limitations, explanations for which both nativists and empiricists are working to pin down.

Object Knowledge

A large part of what we know about infant cognition has come from research on the development of knowledge about objects, research originally inspired by Piaget's theory of sensorimotor intelligence. As you learned in Chapter 4, Piaget believed that young infants' understanding of the world is severely limited by an inability to mentally represent and think about anything that they cannot currently see, hear, touch, and so on. His tests of object permanence led him to infer that when an infant fails to search for an object—even a favorite toy—that has disappeared from sight, it is because the object has also disappeared from the infant's mind.

A substantial body of research has provided strong support for Piaget's original observation that young infants do not manually search for hidden objects. However, as noted in Chapter 4, skepticism gradually arose about his explanation of this fascinating phenomenon, and an overwhelming body of evidence has established that young infants are in fact able to mentally represent and think about the existence of objects and events that are currently out of sight.

The simplest evidence for young infants' ability to represent an object that has vanished from sight is the fact that they will reach for objects in the dark, that is, they reach for objects they cannot see. When young infants are shown an attractive object and the room is then plunged into darkness, causing the object (and everything else) to disappear from view, most babies reach to where they last saw the object, indicating that they expect it to still be there (Perris & Clifton, 1988; Stack et al., 1989).

Young infants even seem to be able to think about some characteristics of invisible objects, such as their size (Clifton et al., 1991). When 6–month–olds sitting in the dark heard the sound of a familiar large object, they reached toward it with both hands (just as they had in the light); but they reached with only one hand when the sound they heard was that of a familiar small object.

FIGURE 5.16 Possible versus impossible events In a classic series of tests of object permanence, Renée Baillargeon first habituated young infants to the sight of a screen rotating through 180 degrees. Then a box was placed in the path of the screen. In the possible event, the screen rotated up, occluding the box, and stopped when it reached the top of the box. In the impossible event, the screen rotated up, occluding the box, but then continued on through 180 degrees, appearing to pass through the space where the box was. Infants looked longer at the impossible event, showing that they mentally represented the presence of the invisible box. (From Baillargeon, 1987a)

The majority of the evidence that young infants can represent and think about invisible objects comes from research using the violation-of-expectancy procedure. The logic of this procedure is similar to that of the visual–preference method we discussed earlier. The basic assumption is that if infants observe an event that violates something they know about the world, they will be surprised or at least interested. Thus, an event that is impossible or inconsistent with respect to the infant's knowledge should evoke a greater response (such as longer looking or a change in heart rate) than does a possible or consistent event.

The violation–of–expectancy technique was first used in a classic series of studies designed by Renée Baillargeon and her colleagues (Baillargeon, Spelke, & Wasserman, 1985) to see if infants too young to search for an invisible object might nevertheless have a mental representation of its existence. In some of these studies, infants were first habituated to the sight of a solid screen rotating back and forth through a 180–degree arc (Figure 5.16). Then a box was placed in the screen's path, and the infants saw two test events. In one, the possible event, the screen rotated upward, occluding the box as it did so, and stopped when it contacted the box. In the impossible event, the screen continued to rotate a full 180 degrees, appearing to pass through the space occupied by the box (which the experimenter had surreptitiously removed).

Infants as young as 3½ months of age looked longer at the impossible event than at the possible one. The researchers reasoned that the full rotation of the screen (to which the infants had previously been habituated) would be more interesting or surprising than the partial rotation only if the infants expected the screen to stop when it reached the box. And the only reason for them to have had that expectation was if they thought the box was still present—that is, if they mentally represented an object they could no longer see. The results also indicate that the infants expected the box to remain in place and did not expect the screen to be able to pass through it.

Other studies have shown that young infants' behavior in this situation is influenced by some of the characteristics of the occluded objects, including height (Baillargeon, 1987a; Baillargeon, 1987b). They expect the screen to stop sooner for a taller object than for a shorter one. Thus, research using two very different assessments—reaching in the dark and visual attention—provides converging evidence that infants who do not yet search for hidden objects nevertheless can represent their continued existence and some of their properties.

Physical Knowledge

FIGURE 5.17 Infants' developing understanding of support relations Young infants appreciate that an object cannot float in midair, but only gradually do they come to understand under what conditions one object can be supported by another. (Adapted from Baillargeon, 1998)

Infants' knowledge about the physical world is not limited to what they know and are learning about objects. Other research has examined what they know about physical phenomena, such as gravity. Even in the first year of life, infants seem to appreciate that objects do not float in midair, that an object that is inadequately supported will fall, that a nonround object placed on a stable surface will stay put, and so forth. For example, in a series of studies in which infants observed a ball being released on a slope, 7–month–olds (but not 5–month–olds) looked longer when the ball moved up the slope than when it moved down, indicating that they had expected the ball to go down (Kim & Spelke, 1992). Similarly, they looked longer at an object that traveled more slowly as it rolled down a slope than at one that picked up speed.

Infants also gradually come to understand under what conditions one object can support another. Figure 5.17 summarizes infants' reactions to simple support problems involving boxes and a platform (Baillargeon, Needham, & DeVos, 1992; Needham & Baillargeon, 1993). At 3 months of age, infants are surprised (they look longer) if a box that is released in midair remains suspended (as in Figure 5.17a), rather than falling. However, as long as there is any contact at all between the box and the platform (as in Figure 5.17b and 5.17c), these young infants do not react when the box remains stationary. By approximately 5 months of age, they appreciate the relevance of the type of contact involved in support. They now know that the box will be stable only if it is released on top of the platform, so they would be surprised by the display in Figure 5.17b. Roughly a month later, they recognize the importance of the amount of contact, and hence they look longer when the box in Figure 5.17c stays put with only a small portion of its bottom surface on the platform. Shortly after their 1st birthday, infants also take into account the shape of the object and hence are surprised if an asymmetrical object like that shown in Figure 5.17d remains stable.

Infants presumably develop this progressively refined understanding of support relations between objects as a result of experience. They observe innumerable occasions of adults placing objects on surfaces, and once in a while, as in the crashing crystal observed by baby Benjamin, they see the consequences of inadequate support. And, of course, they collect additional data through their own manipulation of objects, including lots more evidence than their parents would like about what happens when a milk cup is deposited on the very edge of a high–chair tray.

Social Knowledge

In addition to acquiring knowledge about the physical world, infants need to learn about the social world—about people and their behavior. An important aspect of social knowledge that emerges relatively early is the understanding that the behavior of others is purposive and goal–directed. In research by Amanda Woodward (1998), 6–month–old infants saw a hand repeatedly reach toward one of two objects sitting side by side in a display (see Figure 5.18). Then the position of the two objects was reversed, and the hand reached again. The question was whether the infants interpreted the reaching behavior as directed toward a particular object. They did, as shown by their looking longer when the hand went to the new object (in the old place) than when it reached for the old object it had reached to before. Thus, the infants apparently interpreted the reaching behavior as directed toward a particular object. However, this was true only for a human hand; another group of infants did not react the same way when a mechanical claw did the reaching. (This study may remind you of the one by Meltzoff [1995b] in which older infants imitated the actions of a human but not of a mechanical device.) Shown the same training event (Figure 5.18a), slightly older infants (11–month–olds) were able to correctly predict what the human hand would do next, moving their eyes to the goal object in the test display before the hand actually moved to the goal object (Cannon & Woodward, 2012). Again, though, they did not have the same expectations for the claw as they had for the human hand.

FIGURE 5.18 Infants were habituated to the event shown in (a), a hand repeatedly reaching for a ball on one side of a display. When tested later with displays (b), (c), and (d), infants who saw the hand reach for the other object looked longer than did those who saw it reach for the ball (regardless of the ball's position). The pattern of results indicates that the babies interpreted the original reaching as object–directed. (Adapted from Woodward, 1998)

Other research by Sommerville, Woodward, and Needham (2005) established that infants' understanding of the goal–directed nature of another's actions is related to their own experience achieving a goal. Three–month–olds, who were not yet able to pick up objects on their own, were fitted with Velcro “sticky mittens” (like those described earlier in this chapter and in Chapter 4) that enabled them to capture Velcro–patched toys. Their brief experience successfully “picking up” objects enabled them to interpret the goal–directed reaching of others in the procedure in Figure 5.18 a few months earlier than they would otherwise have been able to do.

Further understanding of intentionality is revealed by studies showing that older infants even attribute intentions and goals to inanimate entities if the objects seem to “behave” like humans. In research by Susan Johnson, 12– and 15–month–olds were introduced to a faceless, eyeless blob that “vocalized” and moved in response to what the infant or experimenter did, thus simulating a normal human interaction (Johnson, 2003; Johnson, Slaughter, & Carey, 1998) (see Figure 5.19). Subsequently, when the blob turned in one direction, the infants looked in that direction. Thus, they seemed to be following the blob's “gaze,” just as they would do with a human partner, assuming that the person had turned to look at something. They did not behave this way when the blob's initial behavior was not contingently related to their own.

FIGURE 5.19 When this amorphous blobby object “responds” contingently to infants, they tend to attribute intention to it.

COURTESY OF SUSAN JOHNSON

Older infants even interpret quite abstract displays in terms of intention and goal–directed action (Csibra et al., 1999, 2003; Gergely et al., 2002). For example, 12–month–olds saw a computer animation of a ball repeatedly “jumping” over a barrier toward a ball on the other side. Adults interpret this display as the jumping ball's “wanting” to get to the other ball. So, apparently, did the infants. When the barrier was removed, the infants looked longer when they saw the ball continue to jump, just as it had done before, than when they saw it move straight to the second ball.

Even younger infants seem to attribute intention with respect to simple displays involving small objects. In a study that used a ball, a cube, and a pyramid, all with “googly” eyes attached, 10–month–olds watched as the ball—the “climber”—repeatedly “attempted” to climb up a hill, each time falling back to the bottom (Hamlin, Wynn, & Bloom, 2007; see Figure 5.20). Then the climber was alternately bumped up the hill by the pyramid or pushed back down by the cube. On the subsequent test event, the infants observed the climber alternately approach the “helper” triangle or the “hinderer” cube. The infants looked longer when the climber approached the “hinderer,” indicating by their surprise not only their understanding of the “intentions” of all three objects but also their understanding of what the “climber's” response to the “helper” and “hinderer” might be expected to be.

FIGURE 5.20 Viewers of the “climber” event described above—infants and adults alike—readily interpret it in terms of intentional action. First they see the ball as “trying” to move up a hill, but then rolling back down, thereby “failing” to achieve its goal of reaching the top. On some trials, after the ball starts to roll back down, a triangle appears below the ball and seems to “push” it upward, “helping” it get to the top. On other trials, a cube appears in front of the ball and “hinders” it by seeming to “push” it down the hill.

HAMLIN, WYNN, & BLOOM, 2007/RIGHTS MANAGED BY NATURE PUBLISHING GROUP

Infants go beyond attributing intentions to others based on their actions: they exhibit preferences for particular individuals and objects based on the individuals' and objects' actions. Earlier in this chapter, we described research focused on infants' visual preferences (Box 5.1). Infants also exhibit social preferences, as evidenced by their desire to engage with some individuals over others. In one of the first studies to demonstrate early social preferences (Kinzler, Dupoux, & Spelke, 2007), U.S. and French 10–month–olds saw alternating life–sized video projections of two individuals speaking to them, one in English and one in French. They then saw another life–sized video of the same two individuals standing side by side behind a table, both holding an identical plush toy. Silently and simultaneously, they smiled at the infant, then at the toy, then at the infant again, and then leaned forward, holding the toys out as though giving them to the infant. The moment the toys disappeared from view on the screen, they appeared (through researchers' magic) on a table in front of the infant, creating the impression that they had come directly from the individuals in the video. The infants' responses suggested a social preference for the individual who had spoken their native language: English–learning infants chose the toy offered by the English speaker, whereas French–learning infants chose the toy offered by the French speaker. Crucially, because the toy was offered in silence, these social preferences were attributable to a preference for the individual who shared the infant's language, not for the language itself.

Similar findings emerged in a food–choice paradigm, in which infants were more likely to choose a food offered by a speaker of their language than by a speaker of another language (Shutts et al., 2009). Indeed, even objects similar to those depicted in Figure 5.20 evoke social preferences (Hamlin et al., 2007). In a variation of the “climber” procedure described above when infants as young as 6 months were presented with the objects they had just observed bumping the “climber” object up the hill or pushing it down the hill, they tended to choose the “helper” object. The social preferences exhibited in studies like these can be quite nuanced. In one recent study that used puppets rather than objects, 5–month–olds uniformly preferred characters who were positive toward “helpers,” whereas 8–month–olds preferred characters who were positive toward “helpers” and negative toward “hinderers” (Hamlin et al., 2011).

These and related studies indicate that well before their 1st birthday, infants have already learned a great deal about how humans behave and how their behavior is related to their intentions and goals. Infants and young children can also draw inferences about other people's knowledge states. For example, 15–month–olds can make inferences about what a person will do based on their knowledge of what the person knows (Onishi & Baillargeon, 2005). In a visual–attention version of the false–belief task (discussed in Chapter 4), infants seem to keep track of what information an adult has about the location of an object. If the object is moved to a new location while an infant—but not the adult—witnesses the move, the infant expects the adult to subsequently search for the object in its original location. That is, the infant expects the adult to search where he or she should believe the toy to be, rather than in the location where the infant knows it actually is. This interpretation is based on the fact that the infants looked longer when the adult searched the object's current location than they did when the adult searched its original location. Thus, this study indicates that 15–month–olds assume that a person's behavior will be based on what the person believes to be true, even if the infant knows that the belief is false. This result suggests that there may be very early precursors of a theory of mind.

Looking Ahead

The intense activity focused on cognition in infancy has produced a wealth of fascinating findings. This new information has not, however, resolved the basic issues about how cognition develops in infancy. The evidence we have reviewed reveals a remarkable constellation of abilities and deficits. Infants can be both surprisingly smart and surprisingly clueless (Keen, 2003; Kloos & Keen, 2005). They can infer the existence of an unseen object but cannot retrieve it. They appreciate that objects cannot float in midair but think that any kind and amount of contact at all provides sufficient support. The challenge for theorists is to account for both competence and incompetence in infants' thinking.