Exp Brain Res (1998) 123:90–94 © Springer-Verlag 1998 Alan Slater · Rachel Kirby Innate and learned perceptual abilities in the newborn infant Abstract From research carried out over the last few years, it has become apparent that the visual world of the newborn baby (0–7 days from birth) is highly organised. It is also clear that the newborn infant is an extremely competent learner. These themes are illustrated with respect to two areas of research, face perception and intermodal learning. Evidence is presented suggesting that the human face is “special” in that newborns respond to them as faces, rather than merely collections of stimulus elements. Additional evidence is presented which demonstrates that newborns can form auditory-visual associations after only a short exposure to the stimulation. These lines of evidence suggest that innate capacities, or modules, facilitate and direct early learning in order to allow newborn infants to understand their newly encountered world. Key words Innate perceptual abilities · Learned perceptual abilities · Newborn perceptual abilities Introduction Research carried out over the last 30 years has changed the traditional view of the young infant’s perceptual world from one of “incompetence” to one of “competence”. While experience is of great importance, infants display considerable perceptual competence at an early age, and even the newborn baby perceives an organised and structured world. Several organisational principles, such as size and shape constancy, the ability to distinguish between 2- and 3-dimensional stimulation and to detect motion, and rudimentary form perception, are innately provided to newborn infants. These help them to parse the visual world and help them to create order out of what would otherwise be chaos (Slater 1995 gives a detailed account of visual perception and memory at A. Slater (✉) · R. Kirby Department of Psychology, Washington Singer Laboratories, University of Exeter, UK birth). In addition to these inborn abilities, it has been suggested that newborn infants have innate representational abilities, and it is also apparent that they engage in very rapid learning about the visual world. This article presents evidence relating to these topics with respect to two areas of research, face perception and intermodal learning. Innate representations and early learning: the case of face perception The human face is one of the most complex visual stimuli encountered by the human infant. It moves, is three-dimensional, has areas of both high and low contrast, and contains features that can appear both in changing (perhaps with changes of expression) and invariant (the positions of eyes, mouth, nose, hair, etc.) relationships. In addition, individual faces are all unique and differ from one another often in subtle ways. A major task confronting the newborn infant, therefore, is to make sense of “faces in general” and also to distinguish between different individual faces. We know that a considerable period of learning is required in order to achieve some degree of competence in these areas. However, it is clear that the newborn infant has something of a “head start”, as shown both by rapid learning about faces and by the likelihood of an innately provided representation of the human face. Early learning about faces The visual information detected by newborn infants is poor, but Hainline and Abramov (1992) point out that, “while infants may not, indeed, see as well as adults do, they normally see well enough to function effectively in their role as infants” (p. 40). For newborn infants, the visual stimuli likely to be of most relevance are people who interact with them and are therefore likely to be close to them. Figure 1 gives an indication of how a face 91 Fig. 1 A face as it might appear to us (right), and to a newborn infant (left) might look to newborn infants at a distance of about a foot from their eyes and how it would look to us. Although the image is degraded and unfocused for the newborn, enough information is potentially available for the infant to learn to recognise the mother’s face and to discriminate her from others. Indeed, it seems that newborns quickly learn about, and show a preference for, their mother’s face (Bushnell et al. 1989; Field et al. 1984; Pascalis et al. 1995; Walton et al. 1992). We do not know precisely what aspects of the face newborn (and older) infants use in making these discriminations. However, it seems that they are attending both to external and internal aspects of the face. Pascalis et al. (1995), in their experiment 1, confirmed earlier findings that newborn infants (their babies averaged 4 days from birth) reliably discriminate and prefer their mother’s face to that of a stranger, when all the facial information is present. However, when mother and stranger both wore scarves around their heads (in experiment 2), the babies were no longer able to make the discrimination. Their interpretation of this is “that what (newborns) have learned has to do with the external features of the face rather than the inner features” (1995, p. 84). However, Walton and Bower (1993) demonstrated that newborns also detect internal facial features. In their experiment, newborn infants who were shown four faces for a total looking time of less than 1 min extracted a facial prototype, or averaged version, of the four in that they subsequently looked more at a composite of the previously seen faces than at a composite of four faces that had not been seen earlier: it would be difficult for the newborns to make these discriminations without attending to the internal features of the faces. In a more recent study (Walton et al. 1997), newborn infants (averaging just over 1 day from birth) recognised a face to which they had previously been trained or familiarised across three facial transformations: (1) a photonegative transformation; (2) a size change (the transformed face was smaller than the original); (3) rotation in the third dimension (the training face was facing straight ahead and the transformed one was half profile). Newborns succeeded with all three transformations in that they sucked more to see the transformed familiar face. These preferences for a “familiar” composite of faces (Walton and Bower 1993) and for the “familiar” transformation (Walton et al. 1997) are further evidence that learning about faces, and the formation of a representation of faces, can be extremely rapid in the newborn period and that what is learned is robust in that it holds across transformations of the original. Is there an innate representation of the human face? Several lines of evidence converge to suggest that newborn infants come into the world with some innately specified representation of faces. Goren et al. (1975) reported that their newborn subjects, who averaged 9 min from birth at the time of testing and had never seen a human face, turned their heads more to follow (i.e., track) a two-dimensional schematic face-like pattern than either of two patterns consisting of the same facial features in different arrangements. A replication of Goren et al.’s study was reported by Johnson and Morton (1991): the stimuli they used were three of those used by Goren et al. and are shown in Fig. 2. Johnson and Morton use these findings and other evidence to argue for the existence of an innate face-detecting device they call “Conspec” (short for conspecifics), which “perhaps comprises just three dark patches in a triangle, corresponding to eyes and mouth” (Pascalis et al. 1995, p. 80) and which serves to direct the newborn infant’s visual attention to faces. Imitation Other evidence suggests that the hypothesized innate facial representation might be more detailed than simply a template that matches three dots. In particular, it has been demonstrated that newborn (and older) infants will imitate a variety of facial gestures they see an adult model performing (e.g., Field et al. 1982; Maratos 1973; 92 Fig. 2 The three stimuli used by Johnson and Morton (1991) in their replication of Goren et al. (1975): newborn infants tracked the Face in preference to Scrambled and Blank Meltzoff and Moore 1977, 1984, 1992, 1994, 1997; Reissland 1988). Meltzoff (1995) suggests that “newborns begin life with some grasp of people” (p. 43) and that their ability to recognise when their facial behaviour is being copied implies that “there is some representation of their own bodies” (p. 53). Infants prefer attractive faces It has been known for some time that infants 2 months and older will spend more time looking at attractive faces, when these are shown paired with less attractive ones. In the studies reported to date, the face pairings have been equated for age, gender, and ethnicity, and the “attractiveness effect” seems to be robust in that it is found for stimulus faces that are infant, adult, male, female, and of two races (African-American and Caucasian) (Langlois et al. 1987, 1991; Samuels and Ewy 1985; Samuels et al. 1994). In two recent experiments (Slater et al. 1998), pairings of attractive and unattractive female faces (as judged by adult raters) were shown to newborn infants (in the age range 14–151 h from birth), and in both the infants looked longer at the attractive faces. The findings from these studies show that newborn infants, less than 1 week from birth, can discriminate attractive from unattractive faces and that, like older infants, they prefer attractive faces. A frequently expressed interpretation of the attractiveness effect in older infants is in terms of prototype formation and a cognitive averaging process. When faces are computer-averaged, the resulting prototype is typically seen as more attractive than the individual faces that are averaged together. Therefore, averageness has been claimed to be an important ingredient of attractiveness (Langlois and Roggman 1990), although it has been claimed that “very attractive faces are not average” (Alley and Cunningham 1991, p. 123). According to this interpretation, attractive faces are seen as more “face-like”, because they match more closely the prototype that infants have formed from their experience of seeing faces. As was mentioned earlier, it is known that newborn infants can form prototypes of faces they have seen in the laboratory in less than 1 min (Walton and Bower 1993), and that they recognise a learned face over transformations (Walton et al. 1997). One interpretation of the present findings, therefore, is that the newborns’ preference for attractive faces is a preference for an image similar to a composite of the faces they have seen in the few hours from birth prior to testing: thus, “infants may prefer attractive or prototypical faces because prototypes are easier to classify as a face” (Langlois and Roggmann 1990, p. 119). An alternative interpretation of the findings is in terms of an innate perceptual mechanism, which detects and responds specifically to faces. According to this interpretation, the newborns look more at the attractive faces because they most closely match the innately provided face template. Several authors have speculated that infants may be born with an innate built-in specification of the face or face module (e.g., Johnson and Morton 1991; Meltzoff 1995; Walton et al. 1997), and Langlois and Roggman (1990) discuss the possibility of an innate account for attractiveness preferences. The results from these studies do not allow us to rule out either of these alternative interpretations. We do not know precisely what features, or combination of features, the newborns (or the older infants of other studies) were using in order to make their preferential choices. Nevertheless, the findings strongly suggest that preferences for attractive faces are present soon after birth and are not the result of a long period of observing faces. Overview It seems now to be reasonably well agreed that “there does seem to be some representational bias... that the neonate brings to the learning situation for faces” (Karmiloff-Smith 1996, p. 10). This representational bias (resulting perhaps from a “face module”) might simply be a tendency to attend to stimuli that possess three blobs in the location of eyes and mouth, or it might be something more elaborate: it is possible that evolution has provided the infant with a more detailed blueprint of the human face. The latter possibility is suggested by newborn infants’ ability to imitate the facial gestures produced by the first face they have ever seen (Reissland 1988) and also, perhaps, by newborn infants’ preferences for attractive faces. The roles of experience and learning: the case of intermodal learning While the newborn and young infant display competence in face perception, their world is very different from ours: “It must certainly lack associations, meaning and familiarity...: (they) see everything, but nothing makes sense” (Gordon 1997, pp. 82–83). Nevertheless, the newborn infant’s ability to learn about the visual world is remarkable, and, in this section, rapid learning by the 93 newborn infant is illustrated by an experiment on intermodal learning. Most of the objects and events we experience are intermodal, in that they provide information to more than one sensory modality. Such intermodal information can be broadly categorised into two types of relation, amodal and arbitrary. Amodal perception is where two (or more) senses provide information that is equivalent in one or more respects, and many types of amodal perception have been demonstrated in early infancy. Newborn infants reliably turn their heads and eyes in the direction of a sound source, indicating that spatial location is given by both visual and auditory information (Butterworth 1983; Muir and Clifton 1985; Wertheimer 1961). One month olds demonstrate cross-modal matching by recognising a visual shape (a pacifier) that had previously been experienced tactually by sucking, indicating that the shape is coded both tactually and visually (Meltzoff and Borton 1979). By 4 months, infants are sensitive to temporal synchrony specified intermodally, in that they detect the common rhythm and duration of tones and flashing lights (Lewkowicz 1986). Four month olds also detect and appropriately match the sounds made either by a single unitary element or by a cluster of smaller elements (Bahrick 1987, 1988). Thus, there is evidence that infants, from birth, perceive a wide range of invariant amodal relations. It is quite likely that the ability to detect amodal relations is innately given to the infant, and, hence, minimal learning is required in their detection. However, many of the intermodal relationships that we perceive appear to be quite arbitrary. For example, there is no information specifying a priori that a particular voice has to be associated with a particular face, that a particular animal makes a certain sound, or that an object makes a certain sound on making contact with a certain surface. Research suggests that many arbitrary intermodal relations are learned in infancy: Spelke and Owsley (1979) found that 31/2 month olds had learned to associate the sound of their mother’s voice with the sight of her face, and Hernandez-Reif et al. (1994) reported that 6 month olds learned the face-voice pairings of same-sex female strangers; Reardon and Bushnell (1988) reported that 7 month olds were able to learn the association between the colour of a container and the taste of the food it contained. There is evidence that newborn babies are also able to learn arbitrary intermodal relations. Slater et al. (1997) familiarised 2-day-old infants, on successive trials, to two visual-auditory stimuli. One visual-auditory combination (“red-mum”) was a red vertical line, and when presented it was accompanied by the sound “mum” spoken in a male voice at a rate of once per second; the other combination (“green-teat”) was a green diagonal line accompanied by the sound “teat” in a female voice. One important characteristic of the familiarisation trials is that the sound was only presented when the infant was judged to be looking at the visual stimulus, meaning that the infants were provided with amodal information, in that there was temporal synchrony of onset and offset of both stimuli. On the post-familiarisation test trials, the infants were presented with one familiar combination (either “red-mum” or “green-teat”), which was alternately presented with a novel combination (“red-teat” or “green-mum”). On the test trials, the infants gave a strong preference for the novel combination, giving a clear demonstration that newborn infants can learn arbitrary visual-auditory combinations. Note that these novelty preferences could not have resulted if the newborns had processed the separate properties of the stimuli presented on the familiarisation trials, because the novel pairings consisted of visual and auditory stimuli that had been presented earlier. Many intermodal events give both amodal and arbitrary information. For instance, when a person speaks, the synchrony of voice and mouth provides amodal information, whereas the pairing of the face and the sound of the voice is arbitrary. In several publications, Bahrick (i.e., 1987, 1988, 1992; Bahrick and Pickens 1994) has provided strong evidence that learning about arbitrary intermodal relations is greatly assisted if there is accompanying amodal information: “detection of amodal invariants precedes and guides learning about arbitrary object-sound relations by directing infants’ attention to appropriate object-sound pairings and then promoting sustained attention and further differentiation” (Bahrick and Pickens 1994, p. 226). In Slater et al.’s (1997) experiment, amodal information was present, and it is likely that this facilitated the newborns’ learning of the arbitrary intermodal relations. It is of interest to note that, when the mother speaks to her infant, the amodal information of temporal synchrony of voice and lips is quite likely to facilitate learning the association of her face and voice, and it seems likely that this learning occurs very soon after birth. Conclusions The newborn baby enters the world visually naive, but possesses a number of means with which to begin the business of making sense of the visually perceived world. Newborn babies have, literally, a head-start, in that it is now clear that they have some innate representational bias to attend to human faces, and perhaps some innate knowledge of the face. These conclusions result from studies of newborn infants’ tracking of face-like patterns and their ability to imitate adult facial gestures. Perhaps, too, the newborn’s preference to look at attractive faces might be guided by an innate facial representation that matches the attractive faces. The nature of infants’ innate facial representation is not well understood and is an intriguing area for future research. It is reasonable to claim that the newborn and young infant are “competent”, in that they perceive the world in an organised manner. Thus, their world is not the “blooming, buzzing confusion” described by William James (1890, p. 488), but it is clearly not the same as 94 ours. At birth, “visual processing begins with a vengeance” (Karmiloff-Smith 1996, p. 10), and this processing can be seen in the rapidity with which newborn infants learn about faces, about other types of stimuli and events (which are not discussed here), and in their ability to learn arbitrary intermodal relations. Evolution has provided the newborn infant with the means to organise and begin to make sense of the visual world and with an innate representation of the human face, which ensures that the infant attends to and learns about faces. Clearly, as infants develop, their perception of the world changes, and infants from birth onwards possess considerable learning ability to ensure rapid learning about the world. Acknowledgements The author’s research reported in this chapter was supported by Research Grant R000235288 from Economic and Social Research Council. References Alley TR, Cunningham MR (1991) Averaged faces are attractive, but very attractive faces are not average. Psychol Sci 2: 123–125 Bahrick LE (1987) Infants’ intermodal perception of two levels of temporal structure in natural events. Infant Behav Dev 10: 387–416 Bahrick LE (1988) Intermodal learning in infancy: learning on the basis of two kinds of invariant relations in audible and visible events. Child Dev 59:197–209 Bahrick LE (1992) Infants’ perceptual differentiation of amodal and modality-specific audio-visual relations. J Exp Child Psychol 53:197–209 Bahrick LE, Pickens JN (1994) Amodal relations: the basis for intermodal perception and learning in infancy. In: Lewkowicz DJ, Lickliter R (eds) The development of intersensory perception: comparative perspectives. Erlbaum, Hillsdale, pp 205–233 Bushnell IWR, Sai F, Mullin JT (1989) Neonatal recognition of the mother’s face. Br J Dev Psychol 7:3–15 Butterworth G (1983) Structure of the mind in human infancy. In: Lipsitt LP, Rovee-Collier CK (eds) Adv Infancy Res 2:1–29 Field TM, Woodson RW, Greenberg R, Cohen C (1982) Discrimination and imitation of facial expressions by neonates. Science 218:179–181 Field TM, Cohen D, Garcia R, Greenberg R (1984) Motherstranger face discrimination by the newborn. Infant Behav Dev 7:19–25 Gordon IE (1997) Theories of visual perception, 2nd edn. Wiley, New York Goren CC, Sarty M, Wu PYK (1975) Visual following and pattern discrimination of face-like stimuli by newborn infants. Pediatrics 56:544–549 Hainline L, Abramov I (1992) Assessing visual development: is infant vision good enough? In: Rovee-Collier C, Lipsitt LP (eds) Adv Infancy Res 7:39–102 Hernandez-Reif M, Cigales M, Lundy B (1994) Memory for arbitrary adult face-voice pairs at six-months of age. Paper presented at the International Conference on Infant Studies, June 2–5, Paris, France James W (1890) The principles of psychology, vol. 2. Holt, New York Johnson MH, Morton J (1991) Biology and cognitive development. Blackwell, Oxford Karmiloff-Smith A (1996) The connectionist infant: would Piaget turn in his grave? SRCD Newslett (Fall issue) 1–10 Langlois JH, Roggman LA (1990) Attractive faces are only average. Psychol Sci 1:115–121 Langlois JH, Roggman LA, Casey RJ, Ritter JM, Rieser-Danner LA, Jenkins VY (1987) Infant preferences for attractive faces: rudiments of a stereotype? Dev Psychol 23:363–369 Langlois JH, Ritter JM, Roggman LA, Vaughn LS (1991) Facial diversity and infant preferences for attractive faces. Dev Psychol 27:79–84 Lewkowicz DJ (1986) Developmental changes in infants’ bisensory response to synchronous durations. Infant Behav Dev 9: 335–353 Maratos O (1973) The origin and development of imitation during the first 6 months of life. Unpublished doctoral dissertation, University of Geneva, Switzerland Meltzoff AN (1995) Infants’ understanding of people and things: from body imitation to folk psychology. In: Bermúdez JL, Marcel A, Eilan N (eds) The body and the self. MIT Press, Cambridge, Mass, London, pp 43–69 Meltzoff AN, Borton RW (1979) Intermodal matching by human neonates. Nature 282:403–404 Meltzoff AN, Moore MK (1977) Imitation of facial and manual gestures by human neonates. Science 198:75–78 Meltzoff AN, Moore MK (1984) Newborn infants imitate adult gestures. Child Dev 54:702–709 Meltzoff AN, Moore MK (1992) Early imitation within a functional framework: the importance of person identity, movement, and development. Infant Behav Dev 15:479–505 Meltzoff AN, Moore MK (1994) Imitation, memory, and the representation of persons. Infant Behav Dev 17:83–99 Meltzoff AN, Moore MK (1997) Explaining facial imitation: a theoretical model. Early Dev Parenting 6:179–192 Muir DW, Clifton R (1985) Infants’ orientation to the location of sound sources. In: Gottlieb G, Krasnegor N (eds) The measurement of audition and vision during the first year of life: a methodological overview. Ablex, Norwood, pp 171–194 Pascalis O, de Schonen S, Morton J, Deruelle C, Rabre-Grenet M (1995) Mother’s face recognition by neonates: A replication and an extension. Infant Behav Dev 18:79–85 Reardon P, Bushnell EW (1988) Infants’ sensitivity to arbitrary pairings of color and taste. Infant Behav Dev 11:245–250 Reissland N (1988) Neonatal imitation in the first hour of life: observations in rural Nepal. Dev Psychol 24:464–469 Samuels CA, Ewy R (1985) Aesthetic perception of faces during infancy. Br J Dev Psychol 3:221–228 Samuels CA, Butterworth G, Roberts T, Graupner L (1994) Babies prefer attractiveness to symmetry. Perception 23:823–831 Slater A (1995) Visual perception and memory at birth. In: RoveeCollier C, Lipsitt LP (eds) Adv Infancy Res, vol. 9. Ablex, Norwood, pp 107–162 Slater A, Brown E, Badenoch M (1997) Intermodal perception at birth: newborn infants’ memory for arbitrary auditory-visual pairings. Early Dev Parenting 6:99–104 Slater A, Schulenburg C von der, Brown E, Badenoch M, Butterworth G, Parsons S, Samuels C (1998) Newborn infants prefer attractive faces. Infant Behav Dev 21:345–354 Spelke ES, Owsley CJ (1979) Intermodal exploration and knowledge in infancy. Infant Behav Dev 2:13–27 Walton GE, Bower TGR (1993) Newborns form “prototypes” in less than 1 minute. Psychol Sci 4:203–205 Walton GE, Bower NJA, Bower TGR (1992) Recognition of familiar faces by newborns. Infant Behav Dev 15:265–269 Walton GE, Armstrong ES, Bower TGR (1997) Faces as forms in the world of the newborn. Infant Behav Dev 20:537–543 Wertheimer M (1961) Psychomotor coordination of auditory and visual space at birth. Science 134:1692