NEWBORN INFANTS PREFER ATTRACTIVE FACES
Alan Slater Charlotte
Von der Schulenburg
Elizabeth
Brown
Marion
Badenoch
University
ofExeter
George Butterworth Sonia Parsons University
of Sussex
Curtis Samuels University
of New England
Several previous experiments
have found that infants 2 months of age and older will spend more time
looking at attractive faces when these are shown paired with faces judged by adults to be unattractive. Two experiments are described whose aim was to find whether the “attractiveness effect” is present soon after birth. In both, pairings of attractive and unattractive female faces (as judged by adult raters) were shown to newborn infants (in the age range 14-151
hours from birth), and in both the infants looked
longer at the attractive fat-es. These findings can be interpreted either in terms of an innate perceptual met-hdnism thdt detec-ts dnd responds specifically to faces, or in terms of rapid learning about faces soon after birth.
newborns
face perception
It has been known for some time that infants 2 months of age 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
l
Alan Slater,
Dcpartmcnt
of Psychology, University
attractiveness
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) of Exeter, Wdshington
Singer Laboratories,
Exeter EX4 4QG;
e-mail:
[email protected]. INFANT
BEHAVIOR
& DEVELOPMENT
21 (21, 1998,
Copyright 0 1998 Ablex Publishing Corporation
pp. 345354
ISSN 0163-6383 All rights of reproduction in any form reserved.
34h
(Langlois, Ritter, Roggman. & Vaughn, 1991; Langlois et al., 1987; Samuels & Ewy, 1985); babies also preferred attractive to symmetrical faces when these two dimensions were varied independently (Samuels, Butterworth. Roberts, & Graupner, 1994). There are, as yet, no reports of infants younger than 2 months being tested with such face pairings but it is known that even newborn infants quickly learn about faces. Bushnell, Sai, and Mullin (1989) Field, Cohen, Garcia, and Greenberg (1984). and Pascalis, de Schonen, Morton, Deruelle. and Rabre-Grenet (1995), reported statistically reliable preferences for the mother’s face, compared with that of a female stranger. at 49 hrs. 45 hrs, and 78 hrs from birth, respectively. Walton, Bower, and Bower ( 1992) reported that infants aged between 12 and 36 hrs from birth produced more sucking responses in order to see a videotaped image of their mother’s face, as opposed to an image of a female stranger‘s face. Walton and Bower (1993) reported that newborn infants who were shown four faces for a total looking time of less than one minute extracted a facial prototype in that they subsequently looked more at a composite of the previously seen faces than at a composite of previously unseen faces, where the composites were derived by pixel-averaging the features of the exemplar faces. More recently Walton, Armstrong, and Bower (1997) have demonstrated that newborns recognize a learned face over the three transformations of photonegative transformation, size change, and rotation in the third dimension of visual space. These findings indicate that learning about faces, and the formation of a representation of faces, can be extremely rapid in the newborn period. In addition to the evidence suggesting rapid learning about faces in the newborn period, several lines of evidence converge to suggest that the newborn infant may come into the world with some innately specified representation of faces. Goren, Sarty, and Wu (1975) reported that their newborn infants, who averaged 9 minutes from birth at the time of testing, turned their heads more to follow (i.e., track) a
INFANT
BEHAVIOR
& DEVELOPMENT
Vol. 21, No. 2, 1’2’)1{
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 (199 I ) who use this and other evidence to argue for the existence of “Conspec.” which constitutes some sort of template consisting essentially of three dots, which serves to direct the newborn’s visual attention to faces. Other evidence suggests that the newborn’s hypothesized innate facial representation may be more detailed than the hypothetical Conspec. In particular it has been demonstrated on many occasions that newborn (and older) infants will imitate a variety of facial gestures they see an adult model performing (Mel&off & Moore, 1977, 1984, 1992, 1994). Meltzoff (1995) suggests that newborns “begin life with some grasp of people” (p. 43) and that their ability to recognize when their facial behavior is being copied implies that “there is a representation 01 their own bodies” (p. 53). Some clarification of our understanding of infants’ knowledge about faces, and of the origins and development of the “attractiveness effect” is likely to emerge when infants younger than 2 months are tested in the “attractive/less-attractive” preferential looking paradigm, and this paper describes two such studies. In both of them newborn infants were the participants.
EXPERIMENT
1
In this experiment pairs of attractive/ less-attractive faces were presented to newborns. These faces had been used in an earlier study
with
months Graupner.
infants
(Samuels.
in the age range Butterworth,
4 to 15
Roberts,
&
1994).
Method Participants Sixteen newborn infants, 7 girls and 9 boys, the ages of 14 and IS I hrs (M = 70
between
Newborns
Prefer Attractive
047
faces
hrs) were the participants, and throughout testing they remained in the behavioral state of alert inactivity (Ashton, 1973). Twelve additional infants began the experiment but did not complete testing because of crying and fussing, and their data were not used.
Stimuli and Apparatus Sixteen black-and-white photographs of Caucasian women’s faces, previously used by Samuels et al. (1994) were the stimuli. They were all in the age range 18-23 years, and photographed full-face. These photographs had been selected from a larger set by Samuels et al. on the basis of adults’ ratings of attractiveness: 8 were judged to be attractive, 8 to be unattractive. These ratings were independently confirmed for the purposes of the present study: 14 adult raters judged the faces on a 7-point scale of attractiveness with 7 being “very attractive.” The attractive faces had a mean rating of 4.9 (range 3.5 to 5.7), and the unattractive faces had a mean rating of 2.4 (range 1.8 to 3.4). The faces were either smiling or had a neutral expression (for the attractive faces 3 were smiling showing teeth, 3 smiling with lips closed, and 2 had neutral expressions; for the unattractive faces these were 4, 3, and 1, respectively). The photographs showed the faces from the crown of the head to the neckline below the jaw and ranged in height from 12 to 16 cm, and in width from 8 to 11 cm, and subtended visual angles of 22-28’ (height) and 15-20” (width) at the viewing distance of 30cm. The stimuli ranged in contrast from 50% to 75%, from hair color (dark) to skin color. The stimuli were mounted on cards and presented against a matte-white screen which measured 61 x 45 cm, and the sides of the viewing chamber were hung with matte-white curtain material. The stimuli were illuminated by two strip lights placed behind and to both sides of the infant. On each paired stimulus trial (see below) the two photographs were equidistant from center and separated by 7.5 cm.
Procedure and Design Each infant was brought to the experimental room in the maternity ward of the hospital, and seated upright on one experimenter’s knee with his/her eyes 30 cm (* 2 cm) from the center of the screen, and given 8 preferential looking trials, in each of which one of the attractive stimuli was shown paired with one of the unattractive ones. At no time could the experimenter who held the babies see the stimuli being presented: this was achieved by having this experimenter look away from the stimulus screen once the baby was appropriately positioned. Over the 8 trials, half of the attractive stimuli were shown on the left and half on the right, the order being randomly determined. Across participants the order was counterbalanced so that each stimulus was shown an equal number of times to the left and to the right. Each trial continued until a total of 20 s of looking (at one or both stimuli) had accumulated. Thus, each infant was shown all 16 stimuli and accumulated a total of 160 s of looking. It is often the case that when preferences for one of two paired stimuli are strong many infants look only at the preferred stimulus (Slater, 1995), and on a number of trials in this experiment infants looked only at one stimulus of the pair shown. It can be argued that on these trials it is not clear that the infants had the opportunity to compare both stimuli, and the results were analyzed both with and without trials included (see these “non-compared” below). The 8 photographs judged attractive were numbered from Al to A8, and the 8 judged to be unattractive were numbered from U9 to U16. There was no obvious reason why any one of the attractive stimuli (Al-A8) should be paired with any one of the unattractive ones (U9-U16) and the infants were tested in four groups of four where, for each group, the pairings of attractive and unattractive stimuli were different: stimuli Al through A8 were paired, respectively, with U9 through U16 (Group l), U16 through U9 (Group 2), U13 through U16
148
and U9 through Ul2 (Group 3), and U12 through U9 and U 16 through U 13 (Group 4). Either one or two experienced observers recorded the infants’ fixations of the stimuli from peepholes behind and to the left and right of the viewing chamber. At no time were these observers visible to the infants and on all trials the observers did not know whether the attractive (or unattractive) stimulus was on the right or left of the screen. Two observers recorded looking for 40 paired stimulus trials: the co-author EB was always one of the observers, and CS and SP acted as the second observer). The interobserver agreement, measured as the correlation between the observers’ looking times to both sides, was high, Pearson Y(38) = 0.851,~ <.OOl
RESULTS AND DISCUSSION The report of results is organized into two sections. The first section focuses on whether the infants display an “attractiveness effect.” and presents analyses of individual infants’ preferences for attractive and unattractive faces, and the preferences for the individual facial stimuli. The second considers the effects of facial expression and stimulus contrast on the preferences found.
The stimuli rated as attractive and unattractive accumulated, respectively, 61.5% and 38.5% of the total looking time. Each newborn infant gave data for 8 attractive/unattractive pairings, giving 128 trials in all: for 100 of these trials the infants looked at both stimuli of the pair and for 28 (“non-compared” trials) they looked only at one. The data can be analyzed in two main ways: (I) overall preferences for the individual infants, combining data for each infant across all stimuli shown: and (2) preferences for the individual facial stimuli. combining the looking times for each stimulus from all infant\. These preferences are given in Tables I and 2, respectively. In these tables the data foi
INFANT
IIEHAVIOK
& DEVELOPMENT
Vol. 21, No. 2, 1WH
all 128 trials are given, and also with the “non-compared” trials (those where the infant looked only at one stimulus of the pair shown) removed. It can be argued that infant looking times, and percentage looking times, are often not normally distributed, and that removing trials (for the analyses where the “non-compared” trials are removed) can result in unbalanced distributions. For these reasons nonparametric statistical tests are used to analyze the data. From the findings with all trials included, 12 of the 16 infants looked longer, overall, at the attractive than the unattractive stimuli (Table I ) and these stimuli attracted 61 .S% of the looking time. This preference for attractiveness Wilcoxon was significant, signed-ranks test, T (16) = 18, p <.005. From the findings with the 28 “noncompared” trials removed (Table I ) it can be seen that the attractiveness preference is slightly lower at 58. I%, but 13 of the 16 infants looked longer at the attractive stimulus and the preference foi attractiveness was significant, Wilcoxon test, T (16)= ll,/~<.OOS. From the findings with all trials included, 7 of the 8 stimuli rated as attractive were looked at for longer than the “unattractive” ones they were paired with (the exception is stimulus A4). and 6 of the 8 stimuli rated as unattractive were looked at for less time than their “attractive” pairs (the exceptions are stimuli UIO and U 13) (Table 2). This difference in looking time between the attractive and unattractive stimuli was significant. Mann-Whitney U test. I/ (N, = N, = 8) = 3, 11 <.()()I. From the findings with the “non-compared” trials removed (Table 3) the stimulus preferences are slightly reduced (i.e., move closer to 50%‘). but the difference in looking times between attractive and unattractive stimuli is significant. Mann-Whitney IT test, I/ (N,
= N, = 8) = 9.5. p <.()I.
mentioned earlier, the face stimuli differed in adulta’ ratings of attractiveness from I .X (least attractive) to S.7 (most attracAs was
tive).
The correlation
ness
rating
Ggnificant. /’ < .005 1.
and
between infants’
Speariiian‘~
adult
attractive-
prefei-em33 r./lo
was
( 16) = 0.726.
349
Newborns Prefer Affrdctive Face5
TABLE 1 Experiment
1: Infants’
preferences
for attractive
faces on paired-stimulus
trials
Attrdcfiveness preference” Noncornpdred /flfdflt
SW
Age (hour6
All fridh
removed**
No. tndlr removed
1
F
124
68.5
64.0
1
2
F
119
86.0
62.8
5
3 4 5 6 7
M
89
48.2
48.2
n ‘I
M
38
72.3
55.7
M
1on
62.6
62.6
M
151
76.0
76.0
M
68
57.0
50.9
1
F
23
74.0
65.3
2
no
3
72.6
70.2
4
IO
F
119
46.4
57.2
.5
11
F
49
5 3.7
5 3 .7
M
12
F
50
37.0
49.3
13
M
21
40.1
45.Y
14
F
29
70.4
60.6
15
M
18
61.9
56.4
16
M
14
57.0
50.9
70
61 .5
‘is.1
(1 .3.Y)
(8.41
Mem
(42.2)
(SD)
L
211
TABLC 2 Experiment
1: Stimulus
differences
in percentage
preferences
for attractive
(A) and unattractive
(U) faces
Al
i5.4
4’1.1
40.5
A2
.59.9
5n.4
50.0
5.3.8
A.1
67.6
55
18.0
22.8
A4
45.6
1’9.1
57.2
55.0
A.5
62.0
53.8
$‘I.0
47.‘)
A6
77.6
71 .J
2.3.4
22.8
A7
61.2
61 .L
41 .(I
47.7 46.0
‘1
40.7
A8
62.4
66.6
.~7.8
M&In
61 ..5
56.9
.18.5
42.1
(SD)
(8.6)
(9.4)
(I 1 .II)
(11 .Y)
Facial Expression andStimulus
Contrast
Two potentially confounding variables are the facial expression displayed by the facial stimuli and the contrast of the stimuli, both of which are known to affect young infants’ preferential looking (Nelson, 1987: Slater, 199.5).
While the experiment was not intended to explore these variables in detail it is possible to analyze the data to find if they had a confounding effect on the measures of preferential looking to attractiveness. One possibility, for example, (pointed out by a reviewer of this paper). is that if attractive
150
smiling faces had inadvertently been paired more often with unattractive unsmiling faces. preferences could have emerged due to smiling rather than attractiveness. It was mentioned above that each facial stimulus displayed one of three expressions: ST, smiling showing teeth; SC, smiling with lips closed; NE, neutral expression. The infants were tested in 4 groups of 4, and for each group the pairings of the 8 attractive stimuli with the 8 unattractive ones were different, giving 32 different pairings in all. For 14 of these pairings the expressions of each member of the pair were matched (i.e., ST paired with [pw] ST, SC pw SC, NE pw NE). For the 18 mismatched pairs the overall average attractiveness preference was 62.8% which is very similar to the attractiveness preference of 61.5% for all infants (Table 1). The attractiveness preferences for the mismatched pairs, with the attractive stimulus given first, were as follows: (a) ST pw SC, N = 4, 63.6%; (b) ST pw NE, N = 2,48.2; (c) SC pw ST, N = 6, 59.9%; (d) NE pw ST, N = 4, 67.4%; (e) NE pw SC, N = 2, 75.2%. For purposes of statistical analysis the data for (a) and (b) were combined. and also those for (d) and (e). A Kruskall-Wallace one-way analysis of variance by ranks was nonsignificant, H (N, = N2 = N,= 6) = 0.8 1, p >.05. Thus, it appears that, for this experiment, differences in facial expression within pairings had no appreciable effect on the attractiveness effect. As mentioned above the facial stimuli ranged in contrast from 50% to 75%. Three adult observers rank ordered the stimuli from most to least contrast. Four attractive and 4 unattractive stimuli were above average contrast and 4 of each were below average, so that there was no systematic variation in contrast between the two sets of stimuli. There was no systematic effect of contrast on infants’ preferences (for the attractive stimuli rho (8) = 0.19, and for the unattractive stimuli rho (8) = 0.20. both ps >.05). These findings suggest that newborn infants will look longer at photographs of faces that adults rate as attractive than photographs rated as unattractive, and that the findings hold
INFANT
BEHAVIOK
& DEVELOPMENT
Vol. 21, No. 2, 1908
across both participant and stimulus variations. Additionally, the preference for attractive stimuli was not confounded with preferences for facial expression or for contrast. Thus, in the present experiment. the “attractiveness effect” is not caused by extreme scores of a few participants, or strong preferences for just a few of the stimuli, or by a preference for a particular facial expression, or by the contrast variations in the stimuli. These findings are discussed after description of the next experiment.
EXPERlMEN
T 2
In Experiment I there were small variations in the size and contrast of the stimuli, although these differences did not differ significantly between the attractive and unattractive stimuli (see Samuels et al., 1994) and did not appear to affect the preferences observed in Experiment I. However, given that newborn infants detect and respond to these stimulus variations (Slater, 1995), Experiment 2 was carried out as a partial replication of the first experiment using facial pairings that were equated for size, brightness and contrast.
Method Participants and Stimuli Nineteen newborn babies, 7 girls and 12 boys, between the ages of 18 and 129 hrs (mean = 64 hrs) were the participants. The stimuli were 12 computer print-outs of faces taken from photographs of Caucasian women’s faces, half chosen to be attractive and half chosen to be unattractive, none of which had been used in Experiment 1. As a confirmation of the selections, the faces were judged by 10 adults on a 7-point scale of attractiveness, with 7 being “very attractive:” 6 were judged to be attractive (mean rating = 5.34, range 4.79-6.14) and 6 were judged to be unattractive (mean rating = 1.75, range 1.21-2.29). The face pairs were matched for facial expression:
Newborns
Prefer Attractive
Faces
one pair were smiling with teeth showing, the other faces all displayed neutral expressions. The faces were put into 6 attractive/unattractive pairs, the members of each pair equated as closely as possible for size, brightness, contrast, and hair style. In order to facilitate the matching the photographs were digitally scanned into a PC using a videocamera and Core1 “photopaint” software. They were then scaled and the contrast manipulated to ensure that skin color was identical for each member of a pair, and that each had similar style, color and amount of hair. When appropriate matching was completed they were printed. Pairs 1 through 4 were black and white, and pairs 5 and 6 were color photographs. The photographs were full-face and showed the faces from the crown of the head to the jaw, and each photograph was 19 cm (+ 0.5 cm) in height and 14 cm (+ 0.5 cm) in width, subtending visual angles of 32O (height) and 25” (width) at the viewing distance of 30 cm. The maximum contrast of each pair was 60%, from hair color (dark) to skin color.
Procedure
and Design
Each infant was given a maximum of 6 paired-stimulus preferential looking trials, each continuing until 20 s of looking had accumulated. In order to minimize participant loss the infants were not required to complete all 6 trials, and testing continued until 12 sets of data had been collected for each stimulus pair. With this constraint, stimulus presentation was in a random order, with the left-right positioning of attractive/unattractive stimuli counterbalanced across infants and stimuli. For 29 paired trials, from 9 infants, 2 independent observers recorded looking (EB was always one observer and two others acted as second observer), and the interobserver agreement was high, r (27) = 0.881, p <.OOl). Other details of testing are as described for Experiment 1.
351
RESULTS AND DISCUSSION For each of the 6 attractive/unattractive pairings there were data for 12 trials, giving 72 trials in all. For 61 of these trials the infants looked at both stimuli of the pair, while the other 11 trials were “non-compared’ in that the infants looked only at one stimulus of the pair. The results are analyzed with and without the “non-compared’ trials included. In the next two paragraphs the attractiveness preferences of the individual infants are presented, followed by the preferences for the different stimulus pairs. From the findings with all trials included, 15 of the 19 infants looked longer at the attractive than at the unattractive faces (Table 3). Each infant’s score was represented as the percentage of the total looking time spent viewing the attractive faces: the mean was 67.9% (SD = 16.5) and this preference for attractiveness was significant, Wilcoxon test, T (19) = 16, p c.005. The number of trials completed by the infants ranged from 1 to 6, and there was a significant correlation between number of trials completed and attractiveness preference, rho (19) = -0.639, p c.01, one-tailed, indicating that those infants who completed fewer trials gave the strongest attractiveness preference: this finding is most readily interpreted as habituation over trials to the variables of attractive/unattractive for the longer looking infants. From the findings with the 11 “non-compared’ trials removed 14 of 18 infants looked longer at the attractive stimulus (Table 3): the mean attractiveness preference of 61.8% (SD = 17.1) was significant, Wilcoxon test, T (18) = 31,p <.Ol. The percentage of time spent looking at the attractive member of the stimulus pairs averaged 64.3% (range 50.9 to 78.9) with all trials included, and 60.3% (range 50.9 to 68.3) with the “non-compared’ trials removed (Table 4). Thus, for all six stimulus pairs the attractive face attracted more looking than the unattractive one. For both “all trials” and “non-compared removed” conditions the difference in
INFANT
3.52
Experiment
2: Infants’
preferences
IJEHAVIOR
for attractive
12’1
i8.7
1
tIo.0
70.0
1
60
86.0
8 I ..3
1
f16..5
Ti’i.3
1
42.0
GL .o
61 .tl
61 .II
101 44 ‘1 11
100.0
2
104
8.3.8
8’1.8
62.3
24.7
fh.‘,
66.5
31
67.0
50.5
1
64.8
57.13
1
114.8
84.8 if,.6
1 .3
T; 2
14
60
if, 6
15
32
36.‘)
If,.‘)
10
OH
4Y.8
4’J.H
17
51
85.0
8.5.0
18
50
7 I ..3
71 ..<
1Y
‘38
7i.5
7i i
MWfl
6.3.8
07.‘)
01 .ti
(3’3.0)
(SD)
3
23 111 126
IL
trials
411.9
43
10
faces on paired-stimulus
1 II 101
8
Vol. 21, No. 2, lYYt<
XI DEVELOPMENT
(17.11
(IO.‘,)
TABLE 4 Experiment
2: Percentage
preferences
time
member
60.3
1
78.‘) 50.8
52.8
JJ 4
611.7 50.9
62.4
5
60. i
fIfI.2
0
64.2
61 .O
04. I (8.‘))
GO.3
was significant,
Wilcoxon
test, T
(6) = 0, p 425.
50.9
(6.4)
can discriminate faces
pretation
DISCUSSION
infants cognitive
The findings newborn
from these two studies
infants,
less than a week
show that from birth,
attractive
and that, like older
attractive
GENERAL
of the facial pairings
2
Mrm (SD)
looking
for attractive
faces.
A frequently
of the attractiveness averaging
unattractive
process.
they
expressed effect
is in terms of prototype
computer-averaged typically
from infants,
prefer inter-
in oldel
formation When
the resulting
seen as more attractive
and a
faces
are
prototype
is
than the indi-
Newborns Prefer Attrdctive Faces
vidual faces that are averaged together and, therefore, averageness has been claimed to be an important ingredient of attractiveness (Langlois & Roggman, 1990). 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 one minute (Walton & 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 & Roggmann, 1990, p. 119). An alternative interpretation of the findings is in terms of an innate perceptual mechanism that detects and responds specifically to faces. According to this interpretation the newborns look more at the attractive faces because they match most closely 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 & 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.
353
This research was supported by Grant R000235288 from the Economic and Social Research Council to the first author. We are indebted to the staff of the Maternity Ward, Royal Devon and Exeter Hospital, Heavitree, Exeter, for their support, and thank Claire Brummitt who assisted in the data collection.
Acknowledgments:
REFERENCES Ashton, R. (1973). The state variable in neonatal research: A review. Merrill-Palmer Quarterly, 19, 3-20. Bushnell, I. W. R., Sai, F., & Mullin, J. T. (1989). Neonatal recognition of the mother’s face. British Journal of Developmental Psychology, 7, 3-15. Field, T. M., Cohen, D., Garcia, R., & Greenberg, R. (1984). Mother-stranger face discrimination by the newborn. Infant Behavior and Development, 7, 19-25. Goren, C. C., Sarty, M., & Wu, P. Y.K. (1975). Visual following and pattern discrimination of face-like stimuli by newborn infants. Pediatrics, 56. 544-549. Johnson, M. H., & Morton, J. (1991). Biology and cognitive development. Oxford, UK: Blackwell. Langlois, J. H., Ritter, J. M., Roggman, L. A., & Vaughn, L. S. (1991). Facial diversity and infant preferences for attractive faces. Developmental Psychology, 27, 79-84. Langlois, J. H. & Roggman, L. A. (1990). Attractive faces are only average. Psychological Science, I, 115-121. Langlois, J. H., Roggman, L. A., Casey, R. J., Ritter. J. M., Rieser-Danner, L. A., & Jenkins, V .Y. (1Y87). Infant preferences for attractive faces: Rudiments of a stereotype? Developmental Psychology. 23, 363-369. Meltzoff. A. N. (1995). Infants’ understanding of people and things: From body imitation to folk psychology. In J. L. Bermudez. A. Marcel, & N. Eilan (Eds.), The body and the .se(f(pp. 43-69). Cambridge, MA: MIT. Meltzoff, A. N., & Moore, M. K. (I 977). Imitation of facial and manual gestures by human neonates. Science, 198, 75-78. Meltzoff, A. N., & Moore, M. K. (1984). Newborn infants imitate adult gestures. Child Development. 54.702-709.
& DEVELOPMENT
3.54
Meltzoff. A. N., & Moore, M. K. (1992). Early imitation within a functional framework: The importance of person identity, movement, and development. Infant Beha\aior and Development, IS, 479-50s. Meltzoff. A. N.. & Moore. M. K. (1YY4). Imitation. memory, and the representation of persons. Infiult BehuiYor und Developmrnt. 17, 83-99. Nelson, C. A. (lY87). The recognition of facial expressions in the first two years of life: Mechanisms of development. Child Del~elopment, 58. 889-909. Pascalis, 0.. de Schonen, S., Morton, J.. Deruelle. C.. & Rabre-Grenet, M. (1995). Mother’s face recognition by neonates: A replication and an extension. Infant Behavior and Development, 18. 79-85. Samuels, C. A.. Butterworth. G., Roberts. T.. & Graupner, L. (1994). Babies prefer attractiveness to symmetry. Perception. 23, 823-83 I.
21, No.
19’18
Samuels, C. A., & Ewy, R. (1985). Aesthetic perception of faces during infancy. British Journal of Developmental P.swholog~, 3, 22 I-228 Slater. A. (I 995). Visual perception and memory at birth. In C. Rovee-Collier & L.P.Lipsitt (Eds.), Advuncrs in infancy research (Vol. 9, pp. 107-162). Norwood, NJ: Ablex. Walton, G. E., 6t Bower, T. G. R. (I 903). Newborns form “prototypes” in less than I minute. Psychologicul Science. I, 203-205. Walton, G. E., Armstrong, E.S., 6t Bower, T. G. R. (1997). Faces as forms in the world of the newborn. ZtzfiuztBelmvior and Drveloprrwnt, 20, 537-543. Walton, G. E., Bower, N. J. A., & Bower, T. G. R. (1992). Recognition of familiar faces by newborns. hfunt Brha\~ior und Development, 15. 265-269.
20 August
1996; Revised
02 September
1997 n