Perception

Parents of new babies cannot help wondering what their children experience—how much they can see, how well they can hear, whether they connect sight and sound (as in our opening vignette), and so on. William James, one of the first psychologists, believed that the world of the newborn is a “big blooming, buzzing confusion.” Because of remarkable advances in the study of early sensation and perception, modern researchers do not share his view. They have demonstrated that infants come into the world with all their sensory systems functioning to some degree and that subsequent development occurs at a very rapid pace. Sensation refers to the processing of basic information from the external world by the sensory receptors in the sense organs (eyes, ears, skin, and so forth) and the brain. Perception is the process of organizing and interpreting sensory information about the objects, events, and spatial layout of the world around us. In our opening example, sensation involved light and sound waves activating receptors in Benjamin's eyes, ears, and brain; an instance of perception involved, for example, his experiencing the visual and auditory stimulation provided by the crashing goblet as a single coherent event.

In this section, we devote the most attention to vision, both because of its fundamental importance to humans and because so much more research has been conducted on vision than on the other senses. We will also discuss hearing and, to a lesser degree, taste, smell, and touch, as well as the coordination between these multiple sensory modalities. Although these abilities often seem commonplace to us as adults, they are actually some of the most remarkable achievements attained during the first year of life.

Vision

Humans rely more heavily on vision than most species do: roughly 40% to 50% of our mature cerebral cortex is involved in visual processing (Kellman & Arterberry, 2006). As recently as a few decades ago, it was generally assumed that newborns' vision was so poor as to be barely functional. However, once researchers started carefully studying the looking behavior of newborns and young infants, they discovered that this assumption was incorrect. In fact, newborns begin visually exploring the world minutes after leaving the womb. They scan the environment, and when their gaze encounters a person or object, they pause to look at it. Although newborns do not see as clearly as adults do, their vision improves extremely rapidly in their first months. And as you will learn, recent studies have revealed that despite their immature visual systems, even the youngest infants have some surprisingly sophisticated visual abilities.

The evidence that enables us to say this so confidently was made possible by the invention of a variety of ingenious research methods. Because young infants are unable to understand and respond to instructions, investigations of infant abilities required researchers to devise methods that are quite different from those used with older children and adults. The first breakthrough was achieved with the preferential-looking technique, a method for studying visual attention in infants. In this technique, pioneered by Robert Fantz (1961), two different visual stimuli are typically displayed on side–by–side screens. If an infant looks longer at one of the two stimuli, the researcher can infer that the baby is able to discriminate between them and has a preference for one over the other. Fantz established that newborns, just like everyone else, would rather look at something than at nothing. When a pattern of any sort—black and white stripes, newsprint, a bull's–eye, a schematic face—was paired with a plain surface, the infants preferred (i.e., looked longer at) the pattern.

Another method that is used to study sensory and perceptual development in infants is habituation, which you encountered in Chapter 2 as a research tool used in studying fetal development. This procedure involves repeatedly presenting an infant with a particular stimulus until the infant's response to it habituates, that is, declines. Then a novel stimulus is presented. If the infant's response increases, the researcher infers that the baby can discriminate between the old and new stimulus. Despite their simplicity, habituation and preferential–looking procedures have turned out to be enormously powerful for studying infants' perception and understanding of the world.

Visual Acuity

FIGURE 5.1 Testing infants' visual acuity Paddles like those depicted here can be used to assess young infants' visual acuity. Two paddles are shown to the infant simultaneously, one with stripes and one in plain gray. If the infant can detect the contrast difference between the black and white stripes, the infant's gaze should, because of infants' preference for a patterned visual field over a plain one, become oriented toward the striped paddle. The ophthalmologist or researcher presents the infant a succession of paddles with increasingly narrow stripes, with increasingly narrow gaps between them, until the infant can no longer distinguish between the striped paddle and the plain gray one. The degree of grating on the last paddle discriminated provides an indication of the infant's visual acuity.

COURTESY GOOD–LITE COMPANY

The preferential–looking method enables researchers (and eye–care professionals) to assess infants' visual acuity, that is, to determine how clearly they can see. This method builds on research showing that infants who can see the difference between a simple pattern and a solid gray field consistently prefer to look at the pattern (Figure 5.1). By varying the patterns and assessing infants' preferences, researchers have learned a great deal about not only infants' early visual abilities but also about their looking preferences. For example, young infants generally prefer to look at patterns of high visual contrast—such as a black–and–white checkerboard (Banks & Dannemiller, 1987). This is because young infants have poor contrast sensitivity: they can detect a pattern only when it is composed of highly contrasting elements.

The blurred image on the right is roughly what a 1–month–old infant would perceive. The infant's relatively low level of visual acuity leads some features of the image to pop out—those with higher contrast (e.g., the woman's eyes and hairline).

BOTH: GOODSHOOT RF/THINKSTOCK

One reason for this poor contrast sensitivity is the immaturity of infants' cones, the light–sensitive neurons that are highly concentrated in the fovea (the central region of the retina) and are involved in seeing fine detail and color. In infancy, the cones have a different size and shape and are spaced farther apart than in adulthood (Kellman & Arterberry, 2006). As a consequence, newborns' cones catch only 2% of the light striking the fovea, compared with 65% for adults (Banks & Shannon, 1993). This is partly why in their first month, babies have only about 20/120 vision (a level of acuity that would enable an adult to read the large E at the top of a standard eye chart). Subsequently, visual acuity develops so rapidly that by 8 months of age, infants' vision approaches that of adults, with full adult acuity present by around 6 years of age (Kellman & Arterberry, 2006).

Another restriction on young infants' visual experience is that, for the first month or so, they do not share adults' experience of a richly colorful world. At best, they can distinguish some shades from white (Adams, 1995). By 2 or 3 months of age, infants' color vision is similar to that of adults (Kellman & Arterberry, 2006). Indeed, it is similar to the extent that 4– and 5–month–olds prefer (look longest at) the same basic colors that adults rate as most pleasant—red and blue (Bornstein, 1975). They also perceive the boundaries between colors in more or less the same way as adults do: they respond equivalently to two shades that adults label as the same color (e.g., “blue”), but they discriminate between two shades that adults refer to with different color names (e.g., “blue” and “green”) (Bornstein, Kessen, & Weiskopf, 1976).

Visual Scanning

FIGURE 5.2 Visual scanning The lines superimposed on these face pictures show age differences in where two babies fixated on the images. (a) A 1–month–old looked primarily at the outer contour of the face and head, with a few fixations of the eyes. (b) A 2–month–old fixated primarily on the internal features of the face, especially the eyes and mouth. (From Maurer & Salapatek, 1976)

As noted, newborns start visually scanning the environment right away. From the beginning, they are attracted to moving stimuli. However, they have trouble tracking these stimuli because their eye movements are jerky and often do not stay with whatever they are trying to visually follow. Not until 2 or 3 months of age are infants able to track moving objects smoothly, and then they are able to do so only if an object is moving slowly (Aslin, 1981). This developmental achievement appears to be less a function of visual experience than of maturation. Preterm infants, whose neural and perceptual systems are immature, develop smooth visual tracking later than full–term infants do (Strand–Brodd et al., 2011).

Another limitation on young infants' visual experience of the world (and therefore on what they can learn) is that their visual scanning is restricted. With a simple figure like a triangle, infants younger than 2 months old look almost exclusively at one corner. With more complex shapes, they tend to scan only the outer edges (Haith, Bergman, & Moore, 1977; Milewski, 1976). Thus, as Figure 5.2 shows, when 1–month–olds look at a line drawing of a face, they tend to fixate on the perimeter—on the hairline or chin, where there is relatively high contrast with the background. By 2 months of age, infants scan much more broadly, enabling them to pay attention to both overall shape and inner details (see Box 5.1).

Pattern Perception

Accurate visual perception of the world requires more than acuity and systematic scanning; it also requires analyzing and integrating the separate elements of a visual display into a coherent pattern. To perceive the face in Figure 5.2, as 2–month–olds apparently do, they must integrate the separate elements.

BOX 5.1: a closer look INFANTS’ FACE PERCEPTION

A particularly fascinating aspect of infant perception has to do with the reaction of human infants to that most social of all stimuli—the human face. As we have noted, infants are drawn to faces from birth, leading researchers to ask what initially attracts their attention. The answer, it seems, is a very general bias toward configurations with more elements in the upper half than in the lower half—something that characterizes all human faces (Macchi Cassia, Turati, & Simion, 2004; Simion et al., 2002) (see the images in the first column). Evidence in support of a general bias to attend to facelike stimuli comes from studies showing that newborn humans are equally interested in human faces and monkey faces—as long as they are presented right–side up (Di Giorgio et al., 2012).

When presented with each of these three pairs of stimuli, newborns look longer at the image in the left–hand column, revealing a general preference for top–heavy stimuli that contributes to their preference for human faces (Macchi Cassia et al., 2004; Simion et al., 2002). Notice that this simple preference is all that is needed to result in newborns spending more time looking at their mother's face than at anything else. By 3 months of age, however, infants no longer discriminate between the faces in the middle pair of pictures, suggesting that their visual attention is no longer guided by a general top–heavy bias (Macchi Cassia et al., 2006).

© 2006 ELSEVIER

From paying lots of attention to faces, infants very quickly come to recognize and prefer their own mother's face. After exposure to Mom over the first few days after birth, infants look longer at her face than at the face of another woman, even when controlling for olfactory cues (a necessary step because, as discussed in Chapter 2, newborns are highly attuned to their mother's scent) (Bushnell, Sai, & Mullin, 2011). Over the ensuing months, infants develop a preference for faces depicting the gender of the caregiver they see most often, whether female or male (Quinn, et al., 2002).

With exposure to many different faces over their first months, infants gradually develop a well–organized perceptual prototype for human faces. The formation of this detailed face prototype then facilitates discrimination between different faces. Evidence for the formation of a general face prototype in the first year comes from an intriguing study of infants' and adults' ability to discriminate between individual human faces and individual monkey faces. Adults, 9–month–olds, and 6–month–olds can all readily discriminate between two human faces. However, adults and 9–month–olds have a great deal of difficulty telling the difference between one monkey face and another (Pascalis, de Haan, & Nelson, 2002). Surprisingly, 6–month–olds are just as good at discriminating between monkey faces as they are at discriminating between human faces.

The researchers concluded that the 9–month–olds and adults rely on a detailed prototype of the human face to discriminate between people, but this prototype does not help them tell the difference between monkeys. The fact that the 6–month–olds discriminated among monkey faces just as well as they discriminated among human faces suggests that these younger infants have not yet developed a tightly organized prototype for human faces. While 6–month–olds are certainly knowledgeable about faces, they do not yet privilege the details of human faces over the details of monkey faces.

Consistent with this account is research showing effects of experience on face recognition. In one study, from the age of 6 months to 9 months, infants were shown a set of pictures of monkey faces on a regular schedule for 1 to 2 minutes. When they were tested at 9 months of age, they demonstrated that they had retained their ability to distinguish between monkey faces, unlike a control group of 9–month–olds who had not had the exposure to monkey faces (Pascalis et al., 2005).

Another type of experience that shapes infant face perception is exposure to individuals of different races. The other race effect (ORE) is a well–established finding, initially observed in adults, in which individuals find it easier to distinguish between faces of individuals from their own racial group than between faces from other racial groups. It was later determined that the ORE emerges in infancy. Whereas newborns show no preference for own–race faces over other–race faces, 3–month–old White, African, and Chinese infants prefer own–race faces (Kelly, Liu et al., 2007; Kelly et al., 2005). Over the second half of the first year, infants' face processing continues to become more specialized, as shown by the emergence of the ORE; by 9 months of age, infants have more difficulty discriminating between other–race faces than between own–race faces (Kelly, Quinn et al., 2007; Kelly et al., 2009).

177

Do the photographs of the men show the same person or different people? How about the two monkey photos? As an adult human, you no doubt can tell the two men apart quite easily, but you may still not be sure whether the two monkey photos are of different individuals. (They are.)

COURTESY OF OLIVIER PASCALIS

What drives these effects is not the infant's own race per se but, rather, the features of individuals in the infant's immediate environment. For example, 3–month–old African emigrants to Israel who were exposed to both African and White caregivers showed equal interest in African and White faces (Bar–Haim et al., 2006). Further evidence of effects of visual experience on face perception comes from a study suggesting that the facial–scanning abilities of biracial infants—who are exposed to the facial features characteristic of two races in the home—are more mature than those of monoracial infants (Gaither, Pauker, & Johnson, 2012).

One of the most intriguing aspects of infants' facial preferences is the fact that, along with all the rest of us, babies like a pretty face. From birth, infants look longer at faces that are judged by adults to be highly attractive than at faces judged to be less appealing (Langlois et al., 1991; Langlois et al., 1987; Rubenstein, Kalakanis, & Langlois, 1999; Slater et al., 1998, 2000).

Older infants' preference for prettiness, like adults', also affects their behavior toward real people. This was demonstrated in a study in which 12–month–olds interacted with a woman whose face was either very attractive or very unattractive (Langlois, Roggman, & Rieser–Danner, 1990). The first key feature of this study was that the attractive woman and the unattractive woman were one and the same! This duality of appearance was achieved through the use of extremely natural–looking professional masks that were applied before the woman interacted with the infants. On a given day, the young woman who would test the babies emerged from her makeup session looking either fabulous or not so fabulous, depending on which mask she was wearing. The masks conformed to what adults judge to be a very attractive face and a relatively unattractive one.

When interacting with the woman, infant participants behaved differently as a function of which mask she was wearing. They were more positive, became more involved in play, and were less likely to withdraw when she was wearing the attractive mask than when she had on the unattractive one. This study was particularly well designed because the young woman never knew on any given day which mask she had on. Thus, the children's behavior could not have been cued by her behavior; it could only have been due to her pretty or homely appearance.

FIGURE 5.3 Subjective contour When you look at this figure, you no doubt see a square—what is called a subjective contour, because it does not actually exist on the page. Seven–month–olds also detect the illusory square. (From Bertenthal et al., 1980)

A striking demonstration of integrative pattern perception in infancy comes from research using the stimulus shown in Figure 5.3. When you look at it, you no doubt perceive a square, even though no square actually exists. This perception of subjective contour results from your active integration of the separate elements in the stimulus into a single pattern. If you simply looked at the individual shapes in turn, no square would pop out. Like you, 7–month–olds perceive the subjective square in Figure 5.3 (Bertenthal, Campos, & Haith, 1980), indicating that they integrate the separate elements to perceive the whole. Even newborns can do so if motion cues are added to the display, such as arranging it so that the illusory square appears to move back and forth (Valenza & Bulf, 2007).

Infants are also able to perceive coherence among moving elements. In research by Bertenthal and his colleagues (Bertenthal, 1993; Bertenthal, Proffitt, & Kramer, 1987), infants watched a film of moving points of light. Adults who watch this film immediately and confidently identify what they see as a person walking; the moving lights appear to be (and are) attached to the major joints and head of an adult. Five–month–olds apparently see the same thing; they look longer at the point–light displays that suggest human movement than at ones that do not. As with the research on newborns' response to the illusory square in Figure 5.3, recent studies have confirmed that even newborns show a preference for a moving–lights depiction of biological motion over one of nonbiological motion (Bardi, Regolin, & Simion, 2011). Taken together, these results suggest that despite their limited acuity and lack of visual experience, newborns are already attentive to the configurations of elements in their visual world.

Object Perception

One of the most remarkable things about our perception of objects in the world around us is how stable the world appears to be. When a person approaches or moves away from us, or slowly turns in a circle, our retinal image of the person changes in size and shape, but we do not have the impression that the person changes in size and shape. Instead, we perceive a constant shape and size, a phenomenon known as perceptual constancy. For a good demonstration of size constancy, look in the mirror and notice that the image of your face seems to be the normal size of a face. Then steam up the mirror and trace the outline of your face on the mirror. You will find that the outline is actually a great deal smaller than your face. But because of perceptual constancy, you perceive the image in the mirror as being the same size as any other adult face.

FIGURE 5.4 If this infant looks longer at the larger but farther–away cube, researchers will conclude that the child has size constancy.

COURTESY OF ALAN SLATER

The origin of perceptual constancy was a traditional component in the debates between empiricists and nativists. Briefly, empiricists maintain that all knowledge arises from experience, whereas nativists hold that certain aspects of knowledge are, in fact, innate, or hard–wired. Thus, empiricists argue that our perception of the constant size and shape of objects develops as a function of spatially experiencing our environment, whereas nativists argue that this perceptual regularity stems from inherent properties of the nervous system.

The nativist view is supported by evidence of perceptual constancy in newborns and very young infants. In a study of size constancy (Slater, Mattock, & Brown, 1990), newborns were repeatedly shown either a large or a small cube at varying distances. While the cube's actual size remained the same, the size of the retinal image projected by the cube changed from one trial to the next (see Figure 5.4). The question was whether the newborns would perceive these events as multiple presentations of the same object or as presentations of similar objects of different sizes.

To answer this question, the researchers subsequently presented the newborns with the original cube and a second one that was identical except that it was twice as large. The crucial factor was that the second cube was located twice as far away as the original one, so it produced the same–size retinal image as the original. The infants looked longer at the new cube, indicating that they saw it as different in size from the original one. This, in turn, revealed that they had perceived the multiple presentations of the original cube as a single object of a constant size, even though its retinal size varied. Thus, visual experience is not necessary for size constancy (Granrud, 1987; Slater & Morison, 1985).

FIGURE 5.5 Object segregation Infants who see the combination of elements in (a) perceive two separate objects, a rod moving behind a block. After habituating to the display, they look longer at two rod segments than at a single rod (b), indicating that they find the single rod familiar but the two segments novel. If they first see a display with no movement, they look equally long at the two test displays. This result reveals the importance of movement for object segregation. (From Kellman & Spelke, 1983)

Another crucial perceptual ability is object segregation, the perception of the boundaries between objects. To appreciate the importance of this ability, look around and try to imagine that you are seeing the scene and the objects in it for the first time. How can you tell where one object ends and another begins? If the objects are separated by a gap, the boundaries between the objects are obvious. But what if there are no visible gaps? Suppose, for example, that as baby Benjamin watches his parents washing dishes, he sees a cup sitting on a saucer. An adult would perceive this arrangement as two distinct objects, but will Benjamin? Lacking experience with china, Benjamin may be unsure: the difference in shape suggests two objects, but the common texture suggests only one. Now suppose that Ben's mother picks up the cup to dip it in the suds, leaving the saucer on the table. Will he still be uncertain? No, because even infants treat the independent motion of cup and saucer (or any objects) as a signal that they are separate entities. Is this knowledge innate, or do infants acquire it from observing everyday events in their environment?

The importance of motion as a cue indicating the boundaries between objects was initially demonstrated in a classic experiment by Kellman and Spelke (1983). First, 4–month–olds were presented with the display shown in Figure 5.5a. This display could be perceived either as two pieces of a rod moving on each end of a block of wood or as a single rod moving back and forth behind the block. Importantly, adults perceive displays of this type the latter way. After habituating to the display, the infants were shown the two test displays in Figure 5.5b: a whole rod and a rod broken into two pieces. The investigators reasoned that if the infants, like adults, assumed that there was a single intact rod moving behind the block during habituation, they would look longer at the broken rod because that display would be relatively novel. And that is exactly what the babies did.

What caused the infants to perceive the two rod segments presented during habituation as parts of a unitary object? The answer is common movement, that is, the fact that the two segments always moved together in the same direction and at the same speed. Four–month–olds who saw a display that was the same as the one in Figure 5.5a, except that the rod was stationary, looked equally long at the two test displays. In other words, in the absence of common movement, the display was ambiguous.

Common movement is such a powerful cue that it leads infants to perceive disparate elements moving together as parts of a unitary object. It does not matter if the two parts of the object moving behind the block differ in color, texture, and shape, nor does it make much difference how they move (side to side, up and down, and so forth) (Kellman & Spelke, 1983; Kellman, Spelke, & Short, 1986). For infants, common motion may have this effect, in part, because it draws their attention to the relevant aspects of the scene—the moving pieces rather than the block (S. P. Johnson et al., 2008). Strikingly, however, even this seemingly very basic feature of visual perception must be learned. Newborn infants, tested using displays similar to those described above and shown in Figure 5.5, do not appear to make use of common motion as a cue to object identity (Slater et al., 1990, 1996). Only at 2 months of age do infants show any evidence that they use common motion to interpret the occluded rod as a single object (S. P. Johnson & Aslin, 1995). Thus, as powerful a cue as common motion may be, infants must develop the ability to exploit it.

As they get older, infants use additional sources of information for object segregation, including their general knowledge about the world (Needham, 1997; Needham & Baillargeon, 1997). Look at the displays shown in Figure 5.6. The differences in color, shape, and texture between the box and the tube in Figure 5.6a suggest that there are two separate objects, although you cannot really be sure. However, your knowledge that objects cannot float in midair tells you that Figure 5.6b has to be a single object; that is, the tube must be attached to the box.

FIGURE 5.6 Knowledge and object segregation (a) It is impossible to know for sure whether what you see here is one object or two. (b) Because of your knowledge about gravity and support, you can be sure that this figure is a single (albeit very odd) object. (From Needham, 1997)

Like you, 8–month–olds interpret these two displays differently. When they see a hand reach in and pull on the tube in Figure 5.6a, they look longer (presumably they are more surprised) if the box and tube move together than if the tube comes apart from the box, indicating that they perceive the display as two separate objects. However, the opposite pattern occurs in Figure 5.6b: now the infants look longer if the tube alone moves, indicating that they perceive a single object. Follow–up studies using the displays in Figure 5.6 with younger infants suggest that younger infants (4½–month–olds) exhibit the adultlike interpretation of these displays, but only when they have been familiarized previously with the box or the tube (Needham & Baillargeon, 1998). Thus, it appears that experience with specific objects helps infants to understand their physical properties. We will return to this idea later in this chapter when we discuss the implications of motor development for infants' knowledge about objects (particularly with respect to reaching).

Depth Perception

To navigate through our environment, we need to know where we are with respect to the objects and landmarks around us. We use many sorts of depth and distance cues to tell us whether we can reach the coffee cup on our desk or whether the approaching car is far enough away that we can safely cross in front of it. From the beginning, infants are sensitive to some of these cues, and they rapidly become sensitive to the rest.

One cue that infants are sensitive to very early on is optical expansion, in which the visual image of an object increases in size as the object comes toward us, occluding more and more of the background. When an image of an approaching object expands symmetrically, we know that the object is headed right for us, and a sensible response is to duck. Babies cannot duck, but they can blink. Timing this blinking response is critical; if infants blink too soon or too late, they risk having the oncoming object hit their open eye. If you think about it, though, it's not at all obvious how infants would know how to correctly time a blink. Doing so requires infants to rapidly exploit information present in the visual image looming before them, including how rapidly the image is expanding and amount of the visual field taken up by the image. Rather remarkably, infants as young as 1 month blink defensively at an expanding image that appears to be an object heading toward them (Ball & Tronick, 1971; Náñez & Yonas, 1994; Yonas, 1981). Preterm infants show a delayed developmental pattern of blinks to looming objects, suggesting that maturation, and not solely postnatal visual experience, is crucial for this developmental achievement (Kayed, Farstad, & van der Meer, 2008).

Another depth cue that emerges early is due to the simple fact that we have two eyes. Because of the distance between them, the retinal image of an object at any instant is never quite the same in both eyes. Consequently, the eyes never send quite the same signal to the brain—a phenomenon known as binocular disparity. The closer the object we are looking at, the greater the disparity between the two images; the farther away the object, the less the disparity. In a process known as stereopsis, the visual cortex computes the degree of disparity between the eyes' differing neural signals and produces the perception of depth. This form of depth perception emerges quite suddenly at around 4 months of age and is generally complete within a few weeks (Held, Birch, & Gwiazda, 1980), presumably due to maturation of the visual cortex.

At around 6 or 7 months of age, infants begin to become sensitive to a variety of monocular depth cues (so called because they denote depth even if only one eye is open) (Yonas, Elieff, & Arterberry, 2002). These cues are also known as pictorial cues, because they can be used to portray depth in pictures. Three of them, including relative size, are presented in Figure 5.7.

FIGURE 5.7 Pictorial cues This Renaissance painting contains multiple examples of pictorial cues. One is interposition— nearer objects occlude ones farther away. The convergence of lines in the distance is another. To appreciate the effectiveness of a third cue—relative size—compare the actual size of the man on the steps in the foreground to the actual size of the woman in the blue dress.

EMPEROR AUGUSTUS AND THE SIBYL, 1535 BY PARIS BARDONE (1500–1571)
PUSHKIN MUSEUM, MOSCOW, RUSSIA/BRIDGEMAN ART LIBRARY

FIGURE 5.8 Monocular depth cues This 7–month–old infant is using the monocular depth cue of relative size. Wearing an eye patch to take away binocular depth information, he is reaching to the longer side of a trapezoidal window. This behavior indicates that the baby sees it as the nearer, and hence more readily reachable, side of a regular rectangular window. (Yonas et al., 1978)

BOTH: ALBERT YONAS

In one of the earliest studies of infants' sensitivity to monocular depth cues, Yonas, Cleaves, and Pettersen (1978) capitalized on the fact that infants will reach toward whichever of two objects is nearer. The investigators put a patch over one eye of 5– and 7–month–olds (so binocular depth information would not be available) and presented them with a trapezoidal window with one side considerably longer than the other (Figure 5.8). (When viewed by an adult with one eye closed, the window appears to be a standard rectangular window sitting at an angle with one side closer to the viewer.) The 7–month–olds (but not the younger babies) reached toward the longer side, indicating that they, as you would, perceived it as being nearer, providing evidence that they used relative size as a cue to depth. (Box 5.2 reviews research on infants' perception of pictures.)

Auditory Perception

Another rich source of infants' information about the world is sound. As we discussed in Chapter 2, fetuses can hear sufficiently well to learn basic features of their auditory environment (their mother's heartbeat, the rhythmic patterns of her native language, and so forth). At birth, the human auditory system is well developed relative to the visual system. That said, although the inner ear structures appear to be mature and adultlike, the conduction of sound through the outer parts of the ear is inefficient (Keefe et al., 1993). Over the course of infancy, there are vast improvements in sound conduction from the outer and middle ear to the inner ear. Similarly, over the first year, auditory pathways in the brain mature significantly. Taken together, these developments in the ear and in the brain greatly improve the infant's ability to respond to, and learn from, sound.

Other factors add to infants' auditory improvement. One example involves auditory localization, the perception of the spatial location of a sound source. When they hear a sound, newborns tend to turn toward it. However, newborns and young infants are far worse at determining the spatial location of a sound than older infants and toddlers are. To localize a sound, listeners rely on differences in the sounds that arrive at both of their ears: a sound played to their right will arrive at their right ear before reaching their left ear, and will be louder at their right ear than at their left ear, thereby signaling the direction the sound is coming from. Young infants may have more difficulty using this information because their heads are small, and thus the differences in timing and loudness in information arriving at each ear are smaller for infants than for toddlers and children with larger heads. Another reason that this information may be difficult for infants to use is that the development of an auditory spatial map (that is, a mental representation of how sounds are organized in physical space—right versus left, up versus down) requires multimodal experiences, through which infants become able to integrate information from what they hear with information from what they see and touch. The development of an auditory spatial map must therefore await the improvements in visual and motor skills that emerge later in infancy (Saffran, Werker, & Werner, 2006).

BOX 5.2: a closer look PICTURE PERCEPTION

A special case of perceptual development concerns pictures. Paintings, drawings, and photographs are ubiquitous in modern societies, and we acquire an enormous amount of information through them. When can infants perceive and understand these important cultural artifacts?

Even young infants perceive pictures in much the same way that you do. In a classic study, Hochberg and Brooks (1962) raised their own infant son with no exposure to pictures at all: no art or family photos; no picture books; no patterns on sheets, clothing, or toys. They even removed the labels from canned foods. Nevertheless, when tested at 18 months, the child readily identified people and objects in photographs and line drawings. Later research established that infants as young as 5 months old can recognize people and objects in photographs and drawings of them (e.g., DeLoache, Strauss, & Maynard, 1979; Dirks & Gibson, 1977), and even newborns can recognize two–dimensional versions of three–dimensional objects (Slater, Morison, & Rose, 1984).

These 9–month–old infants—two from the United States and two from West Africa—are responding to pictures of objects as if they were real objects. They do not yet know the true nature of pictures. (From DeLoache et al., 1998)

ALL: COURTESY KAREN ADOLPH

Despite their precocious perception of pictures, infants do not understand their nature. The four babies shown here—two from the United States and two from a rural village in West Africa—are all manually exploring depicted objects. Although these 9–month–old babies can perceive the difference between pictures and objects, they do not yet understand what two–dimensionality means; hence, they attempt to treat pictured objects as if they were real objects—with an inevitable lack of success. By 19 months of age and after substantial experience with pictures, American infants no longer manually investigate pictures, apparently having learned that pictures are to look at and talk about, but not to feel, pick up, or eat (DeLoache et al., 1998; Pierroutsakos & DeLoache, 2003). In short, they have come to understand the symbolic nature of pictures and appreciate that a depicted object stands for a real object (Preissler & Carey, 2003).

Whereas most Western infants live in environments filled with pictured objects, infants in other cultures often lack experience with such images. Fascinating cross–cultural research suggests that, in fact, infants who grow up in homes and communities without pictured objects do not show the same trajectory of understanding that pictures are representations of real objects. In one study, Canadian toddlers and preschoolers outstripped their peers from rural India and Peru in their ability to match line drawings of objects to toy objects (Callaghan et al., 2011). Similarly, toddlers from rural Tanzania, who had no prior exposure to pictures, had greater difficulty than did North American toddlers in generalizing the names of objects in color photographs to the objects themselves (Walker, Walker, & Ganea, 2013). These studies suggest that understanding the relationship between 2D images and 3D objects requires experience with pictorial media.

Infants are adept at perceiving patterns in the streams of sound they hear. They are remarkably proficient, for example, at detecting subtle differences in the sounds of human speech, an ability we will review in detail in our discussion of language development in Chapter 6. Here we will focus on another realm in which infants display an impressive degree of auditory sensitivity—music.

Music Perception

Infants are sensitive to music, as shown by the fact that caregivers around the world sing while caring for their infants (Trehub & Schellenberg, 1995). In the United States, for example, 60% of parents sing or play music to their children every day (Custodero, Britto, & Brooks–Gunn, 2003).

When adults sing to their infants, they do so in a characteristic fashion which, like the infant–directed speech register we will discuss in Chapter 6, tends to be slower and higher–pitched, and to suggest more positive affect, than does singing directed toward adult listeners. Perhaps because of these characteristics, infants prefer infant–directed singing over adult–directed singing (Masataka, 1999; Trainor, 1996). Indeed, infant–directed singing even appears to trump infant–directed speech as a preferred stimulus, as suggested by a study in which 6–month–olds were more attentive to videos of their own mother singing than to videos of her speaking (Nakata & Trehub, 2004).

Beyond their interest in music, infants are also able to remember what they hear, recognizing musical excerpts several weeks after first being exposed to them (Saffran, Loman, & Robertson, 2000; Trainor, Wu, & Tsang, 2004; Volkova, Trehub, & Schellenberg, 2006). These memories are surprisingly detailed, and include aspects of the pitch, timbre, and tempo of the original performances. For example, when 7–month–olds were tested on songs that they had heard in a particular key two weeks earlier, they listened longer when the same songs were sung in a new key than when they were sung in the original key (Volkova et al., 2006). This indicates that infants not only discriminated between performances of the same song in two different keys but also continued to remember the original key of the song two weeks after they had last heard it sung.

In many ways, infant music perception is adultlike. One well–studied example is the preference for consonant intervals (e.g., octaves, or perfect fifths like the opening notes of the ABCs song) over dissonant intervals (e.g., augmented fourths like the opening of “Maria” from the musical West Side Story, or minor seconds like the theme from the film Jaws). From Pythagoras to Galileo to the present day, many scientists and scholars have argued that consonant tones are inherently pleasing to human ears, whereas dissonant tones are unpleasant (Schellenberg & Trehub, 1996; Trehub & Schellenberg, 1995). To see if infants agree, researchers employ a simple but reliable procedure. They draw infants' attention toward an audio speaker by using a visually interesting stimulus (e.g., a flashing light) and then play music through the speaker. The length of time infants look at the speaker (actually, at the visual stimulus located in the same position as the speaker) is taken as a measure of their interest in, or preference for, the music emanating from the speaker.

Studies have shown that infants pay more attention to a consonant version of a piece of music, whether a folk song or a minuet, than to a dissonant one (Trainor & Heinmiller, 1998; Zentner & Kagan, 1996, 1998). A study by Masataka revealed that even 2–day–old infants show this pattern of preference (Masataka, 2006). This study is particularly notable in that it was conducted with hearing infants whose mothers were deaf, making it unlikely that the infants would have had prenatal exposure to singing. These results suggest that preferences for consonant music as opposed to dissonant music are not due to musical experience. Indeed, other species (including chicks, macaque monkeys, and chimps) also show preferences for consonant music, supporting the view that preferences for consonance over dissonance are unrelated to musical experience (e.g., Chiandetti & Vallortigara, 2011; Sugimoto et al., 2010).

In certain other aspects of music perception, infants diverge markedly from adult listeners. One of the most interesting differences is in the area of melodic perception, in which infants can make perceptual discriminations that adults cannot. In one set of studies, 8–month–old infants and adults listened to a brief repeating melody that was consistent with the harmonic conventions of Western music. Then, in a series of test trials, they heard the melody again—but with one note changed. On some trials, the changed note was in the same key as the melody; on others, it fell outside the key. Both infants and adults noticed changes that violated the key of the melody, but only the infants noticed the changes that stayed within the key of the melody (Trainor & Trehub, 1992). Does this mean that infants are more musically attuned than adults? Probably not. What appeared to be a heightened musical sensitivity in the infant participants was more likely a reflection of their relative lack of implicit knowledge about Western music. Because it takes years to acquire culture–specific familiarity with musical key structures, the within–key and out–of–key changes were equally salient to the infant listeners (Trainor & Trehub, 1994). For adults, years of hearing music makes it very difficult to detect note changes that stay within a key.

In a similar way, infants are also more “sensitive” to aspects of musical rhythm than adults are. Musical systems vary in the complexity of their rhythmic patterning; the rhythms of Western music, for example, are relatively simple compared with those of some cultures in Africa, India, and Europe. Hannon and Trehub (2005a, 2005b) tested adults and 6–month–olds on their ability to detect meter–disrupting changes in simple rhythms versus complex rhythms. Notably, some of the adults lived in the Balkans, where the local music contains complex rhythmic patterns, and others lived in North America, where popular music is characterized by simpler rhythmic patterns. The results revealed that all groups detected changes in the simple rhythms, but only the North American infants and the Balkan adults detected changes in the complex rhythms. Thus, North American 6–month–olds outperformed North American adults on this task. A follow–up study asked whether North American 12–month–olds and adults could be trained to detect such changes in the complex rhythms. After 2 weeks of exposure to the Balkan rhythms, the 12–month–olds were able to detect changes in complex rhythms, but the adults still failed to do so.

These examples from the musical domain suggest that, with experience, there is a process of perceptual narrowing. Infants, who are relatively inexperienced with music, can detect differences between musical stimuli that adults cannot. Developmental changes in which experience fine–tunes the perceptual system are observed across numerous domains. Indeed, you saw this process of perceptual narrowing in our discussion of face perception in Box 5.1, and you will see the same pattern of development when we examine intermodal aspects of speech perception and, quite prominently, when we take up language acquisition in Chapter 6. Across all these examples and in other domains, experience leads the young learner to begin to “lose” the ability to make distinctions that he or she could make at earlier points in development. In each case, this perceptual narrowing permits the developing child to become especially attuned to patterns in biological and social stimuli that are important in their environment.

Taste and Smell

As you learned in Chapter 2, sensitivity to taste and smell develops before birth, and newborns prefer sweet flavors. Preferences for smells are also present very early in life. Newborns prefer the smell of the natural food source for human infants—breast milk (Marlier & Schaal, 2005). Smell plays a powerful role in how a variety of infant mammals learn to recognize their mothers. It probably does the same for humans, as shown by studies in which infants chose between the scent of their own mother and that of another woman. A pad that an infant's mother had worn next to her breast was placed on one side of the infant's head and a pad worn by a different woman was placed on the other side. Two–week–old infants turned more often and spent more time oriented to the pad infused with their mother's unique scent (MacFarlane, 1975; Porter et al., 1992).

Touch

Initially, every object that a baby can pick up gets directed to his or her mouth for oral exploration—whether or not it will fit. Later, infants are more inclined to explore objects visually, thereby showing an interest in the object itself.

GARETH BROWN/CORBIS
SERGEJS NESCERECKIS/ALAMY

Another important way that infants learn about the environment is through active touch, initially through their mouth and tongue, and later with their hands and fingers. Oral exploration dominates for the first few months, as infants mouth and suck on their own fingers and toes, as well as virtually any object they come into contact with. (This is why it is so important to keep small, swallowable objects away from babies.) Through their ardent oral exploration, babies presumably learn about their own bodies (or at least the parts they can get their mouths on), as well as about the texture, taste, and other properties of the objects they encounter.

From around the age of 4 months, as infants gain greater control over their hand and arm movements, manual exploration increases and gradually takes precedence over oral exploration. Infants actively rub, finger, probe, and bang objects, and their actions become increasingly specific to the properties of the objects. For example, they tend to rub textured objects and bang rigid ones. Increasing manual control facilitates visual exploration in that infants can hold interesting objects in order to examine them more closely, rotating the objects to view them from different angles and transferring them from hand to hand to get a better view (Bushnell & Boudreau, 1991; Lockman & McHale, 1989; Rochat, 1989; Ruff, 1986).

Intermodal Perception

Most events that both adults and infants experience involve simultaneous stimulation through multiple sensory modalities. In the crystal–goblet–falling–on–tile–floor event witnessed by Benjamin, the shattering glass provided both visual and auditory stimulation. Through the phenomenon of intermodal perception, the combining of information from two or more sensory systems, Ben's parents perceived the auditory and visual stimulation as a unitary, coherent event. It is likely that 4–month–old Ben did, too.

According to Piaget (1954), information from different sensory modalities is initially separate, and only after some months do infants become capable of forming associations between how things look and how they sound, taste, feel, and so on. However, it has become abundantly clear that from very early on, infants integrate information from different senses. Research has shown, for example, that very young infants link their oral and visual experiences. In studies with newborns (Kaye & Bower, 1994) and 1–month–olds (Meltzoff & Borton, 1979), infants sucked on a pacifier that they were prevented from seeing. They were then shown a picture of the pacifier that had been in their mouth and a picture of a novel pacifier of a different shape or texture. The infants looked longer at the pacifier they had sucked on. Thus, these infants could visually recognize an object they had experienced only through oral exploration.

When infants become capable of exploring objects manually, they readily integrate their visual and tactile experience. In one study, for example, 4–month–olds were allowed to hold and feel, but not see, a pair of rings that were connected by either a rigid bar or a string. When the babies were shown both types of rings, they recognized the ones they had previously explored with their hands (Streri & Spelke, 1988).

Researchers have also discovered that infants possess a variety of forms of auditory–visual intermodal perception. In studies of this mode of perception, infants simultaneously view two different videos, side by side, while listening to a soundtrack that is synchronized with one of the videos but not the other. If an infant responds more to the video that goes with the soundtrack, it is taken as evidence that the infant detects the common structure in the auditory and visual information.

In a classic study using this procedure, Spelke (1976) showed 4–month–olds two videos, one of a person playing peekaboo and the other of a hand beating a drumstick against a block. The infants responded more to the film that matched the sounds they were hearing. When they heard a voice saying “Peekaboo,” they looked more at the person, but when they heard a beating sound, they looked longer at the hand. In subsequent studies, infants showed finer discriminations. For example, 4–month–olds responded more to a film of a “hopping” toy animal in which the sounds of impact coincided with the animal's landing on a surface than they did to a film in which the impact sounds occurred while the animal was in midair (Spelke, 1979). At this age, infants can also draw more abstract connections between sights and sounds. For example, 3– to 4–month–olds look longer at visual displays in which dimensions in each modality are congruent, such as a ball rising and falling at the same rate as a whistle rising and falling in pitch (Walker et al., 2010).

A setup like this one enables researchers to study auditory–visual intermodal perception. The two monitors display different films, one of which is coordinated with a soundtrack. The video camera records the infant's looking toward the two screens.

Similar studies have found that infants are especially sensitive to the relation between human faces and voices. Between 5 and 7 months of age, infants notice the connection between emotional expressions in faces and voices (Soken & Pick, 1992; Walker–Andrews, 1997). When infants hear a happy voice, they look longer at a smiling face, and they look longer at an angry face when they hear an angry voice. Infants are also attuned to the match between faces producing speech and the sounds of speech. When 4–month–olds are shown side–by–side films of a person talking while they are listening to a soundtrack that matches one of the films, they look longer at the face whose lip movements are synchronized with the speech they hear (Spelke & Cortelyou, 1980; Walker–Andrews, 1997). Four–month–olds even detect the relation between specific speech sounds, such as “a” and “i,” and the specific lip movements associated with them (Kuhl & Meltzoff, 1982, 1984).

However, the processes of perceptual narrowing that we have noted elsewhere also occur in intermodal perception. Young infants can detect correspondences between speech sounds and facial movements for nonnative speech sounds (those not present in their native language), but older infants cannot (Pons et al., 2009). Similarly, young infants can detect the correspondence between monkey facial movements and monkey vocalizations, but older infants are unable to do so (Lewkowicz & Ghazanfar, 2006). Experience thus fine–tunes the types of intermodal correspondences that infants detect.

Infants can do more than detect relationships between information across modalities: they can use information in one modality to interpret ambiguous information in another modality. In an ingenious series of experiments, 7–month–olds listened to a musical rhythm that was ambiguous and could be interpreted in either duple or triple time (Phillips–Silver & Trainor, 2005). While infants were listening, they were bounced up and down at a rate matching either a duple– or triple–time interpretation of the ambiguous rhythm. When tested, infants preferred to listen to the version of the rhythm that fit the pattern in which they were bouncing. These results indicate that infants readily integrate vestibular information with auditory information: how infants were bounced altered how infants interpreted what they were hearing.