Developmental Psychology Copyright 1988 by the American Psychological Association, Inc. 1988, Vol. 24, No. 4, 464--469 0012-1649/88/$00.75

Neonatal Imitation in the First Hour of Life: Observations in Rural Nepal

Nadja Reissland Department of Experimental Psychology

University of Oxford, England

Neonatal imitative responses were studied in 12 Maithil neonates (6 boys and 6 girls) during their first hour postpartum. No drugs were administered prior to or during labor, and the delivery was concluded without complications. The neonates observed two modeling conditions: lips widened and lips pursed. It was found that neonates moved their lips significantly more often in accordance with the model's lip position than at variance with the positions. These results suggest, when consid- ered in the light of studies of Caucasian infants of North American and European parentage, that imitative capacity is present at birth.

Studies of neonatal imitation of adult facial expressions have provoked considerable controversy. Whereas some psycholo- gists have obtained evidence of imitative capacity (Field, Wood- son, Greenberg, & Cohen, 1982; Jacobson, 1979; Maratos, 1982; Meltzoff & Moore, 1977, 1983a; Vinter, 1986), others have not (Hayes & Watson, 1981; Koepke, Hamm, Legerstee, & Russell, 1983; McKenzie & Over, 1983a). In part, the discus- sion of these disputed findings has focused on methodological issues. Positive findings have been counterexplained with refer- ence to observer bias (Koepke, Hamm, & Legerstee, 1983) and scoring techniques (Masters, 1979), negative findings, by seat- ing arrangements (Meltzoff& Moore, 1983a), lighting (Meltzoff & Moore, 1983b), and the relevance of the social context in which the neonate is observed (McKenzie & Over, 1983b). The experimentation has also raised profound theoretical issues, so that researchers even disagree over whether the same observed act is, in fact, an act of imitation (De Gelder, 1982; Meltzoff & Moore, 1977; Uzgiris, 1984).

The keen interest in neonatal imitation lies, of course, in the inference that can be drawn about innate cognitive capacity. Yet this inference is not entirely warranted even by those posi- tive reports that do exist. First, many studies purport to be "from bir th" yet they have taken place hours or days (rather than minutes) after birth, by which time the infant has already entered into regular patterns of interaction with parents, rela- tives, and hospital staff, all of whom might influence the neo- nate's response to facial stimuli (the only exception was Melt- zoff& Moore, 1983a, who had in their sample one baby at 42- rain postpartum). Second, research on neonatal imitation has been carried out on Caucasian infants of North American and

I thank P. Bryant, R. Campbell, A. Slater, A. Vinter, P. McLeod, R. Burghart, and P. Harris for their helpful comments in the analysis of the data and P. Harris for his valuable suggestions while I was in Nepal. I am grateful to His Majesty's Government for permission to carry out research at the Janakpur Hospital in Nepal and to the nurses, peons, and parents of the babies for their cooperation in my research. I also thank the three independent judges for their patient coding of the data.

Correspondence concerning this article should be addressed to Nadja Reissland, Department of Experimental Psychology, University of Ox- ford, South Parks Road, Oxford, England OX1 3UD.

464

European cultural background. Whether infants from a differ- ent genetic pool, subject to different prenatal cultural practices, and delivered free of medication are also capable of imitation remains to be tested.

This article reports a study of neonatal imitation in which all subjects were tested during their first postpartum hour. Neona- tal imitation was defined in the following terms proposed by Meltzoff and Moore: "any infant behavior that is produced on the basis of perceiving modeled behaviors, is structurally sim- ilar to the behavior that elicited it, and is not attributable to a chance match between modeled behavior and response" (1985, p. 14). As Lewis and Sullivan (1985) pointed out, it is some- times uncertain whether researchers have counted approxima- tions of the modeled movements or, more narrowly, exact imi- tations of modeled movements. To clarify, the present study fo- cused on two visible muscle movements of the neonate's mouth region: movement in the lateral and inferior plane (i.e., lip wid- ening) and movement in the anterior-posterior plane (i.e., lip pursing). Only the direction of movement, not the completeness of the modeled expression, was noted. The test of imitation lay in whether the subjects moved their lips significantly more in the direction of the modeled lip positions than they did in either a neutral or in the opposite direction.

Me thod

Subjects Twelve neonates, 6 boys and 6 girls, were observed in their first hour

postpartum. All infants were of Maithil parentage (a people of northern Bihar, India, and of the eastern Tarai, Nepal). They were delivered vagi- nally without complication at the Janakpur Hospital, Dhanusa District, Nepal. No drugs were administered to the mothers before or during labor. Under such conditions one would expect the infants to be subject to catecholamine surges, which would have the effect of arousing the neonate, clearing its lungs of fluid, and supplying ample amounts of oxygen to vital organs (Lagercrantz & Slotkin, 1986). Furthermore, all babies were born at term if one takes as index the flection response, that is, arms and legs were flexed and resisted straightening (Amiel-Tison, 1985).

The sample was selected according to the following criteria. Over sev- eral weeks, the experimenter endeavored to film all infants delivered at

NEONATAL IMITATION IN THE FIRST HOUR OF LIFE 465

the hospital between approximately 1 I a.m. and 3 p.m. (i.e., when there was sufficient natural light in the labor room for filming to take place). In practice this meant that. at most, one neonate was filmed per day. All the subjects who began the session also completed it. Abnormal births and abnormal (e.g., breech) deliveries were excluded from the sample at the outset. Apparently normal babies were excluded from analysis in only two cases: In one case, the neonate cried continuousl~ so that no modeling could be attempted, in the other case, two mothers gave birth simultaneously. The hospital staff placed the two babies together, leaving no room for the experimenter to model facial expressions. No subjects were excluded retrospectively from the analysis.

Setting

The neonates ~ere filmed in the labor room of the Janakpur Hospital, where the experimenter attended the final stages of the mothers' labor. As soon as the baby was born, the hospital peon (who serves as a medical auxiliary assisting the nursel placed the baby on a small iron table situ- ated behind the mother and out of her view. The peon perfunctorily wiped the neonate's face and then turned to help the nurse deliver the afterbirth. The mothers were given no anesthetics, and episiotomies were almost routine for primigravidas. Hence, the third stage of labor was in many cases excruciatingly painful and resisted by the mother. It took up to 45 min before the peon was able to return to the baby, wiping it clean with mustard oil and then cutting the umbilical cord.

Until the peon returned to clean the neonate, the baby was left on its own; the hospital was too short-staffed to attend to the child. Mean- while, the relatives of the mother did not touch the baby until it had been cleaned, otherwise they would have become rituall.v polluted by him or her. Nor was the baby given immediately to the mother. In Mith- ila, midwives say that if the mother comes to know that she has given birth to a son, she will be so overjoyed that the afterbirth will not be expelled. Hence, at the very moment of birth, that is when the neonates" eyes were still closed, it x~.as hospital custom to wrap the baby in an old sari and to put it out of view until the afterbirth had also been delivered and the episiotomy sewn. It was during these minutes, when the bab~ was left unattended, that the experiment x~as carried out. Hence, the experimenter was the first person that the baby saw and the first person with whom the baby interacted.

Design and Procedure

On opening their eyes, the neonates were presented with t~o condi- tions: (a) unattended (no interaction with the baby took place) in a non- modeling condition and (b) attended in a modeling condition in which the experimenter's lips were either pursed (lips closed and moved in the anterior-posterior plane, with teeth not visible) or widened (lips moved in a horizontal plane, with edges upturned and teeth ~isible). In the nonmodeling condition the baby ~as left lying on a pillow in a slightly inclined plane ~.ith the head higher than the rest of the torso. In this condition only the baby's face and upper torso were visible on the screen. During the modeling conditions the bab.~ ~as approximately 20 cm from the model's face, I~ing on a pillow in the same slightly inclined plane. The modeling conditions were videotaped, with the child's face and upper torso appearing in the center and left of the screen and the model's mouth movements in the upper-right corner. During the model- ing condition the experimenter's lower arm and hand were visible: dur- ing the nonmodeling condition this was only sometimes the case (e.g., when the model rested her hand at the side of the table and turned to respond to questions from the nurse and peon in the delivery room).

Measures

As soon as the peon placed the neonate on the table, the videocamera was turned on and left running throughout the entire session and con-

tinued after the return of the peon to clean the bab), cut the cord, and swaddle him or her. Filming was stopped with the removal of the baby from the table. The modeling and nonmodeling conditions were enacted between the time when the neonate was placed on the table and the return of the peon to cut the cord. For the 12 neonates this resulted in a mean of 25.08 rain of filmed data, with a range of 9 to 45 min.

The data used in the analysis of the results were selected from this filmed data in a two-stage process. In the first stage, the babies were variously placed in two conditions.

Nonmodehng condition. Seven neonates were subjected toward the end of the session to the nonmodeling condition, as were 5 neonates toward both the beginning and the end; both sessions were used in the analysis. The nonmodeling condition for all 12 neonates lasted a mean of 9.6 min, with a range of 3 to 17 min. if the nonmodeling condition took place at the beginning of the session, its length (M = 6 min, ranging from 2 to l0 min~ was determined by the experimenter, who estimated the elapse of time and made sure to set aside enough time for the model- ing condition. If placed at the end of the session, the nonmodeling condi- tions' length (M = 7.3 rain, ranging from 3 to 14 min) was determined by hospital procedure, that is, by the peon interrupting the condition by cutting the cord and wiping the neonate clean.

Modeling conditions. During the modelingcondition the 12 neonates were presented alternately with exemplars of widened lips and pursed lips. The experimenter initiated each exemplar when the baby was in a calm and attentive state. It was required that the baby imitate the two modeled expressions at the time they saw the expressions. On the one hand. this had the disadvantage of forcing the baby to respond without delay to modeled expressions. On the other hand, an appropriate re- sponse after some elapse of time assumes the existence of memory ca- pacity, and such an assumption is possibly unwarranted in human sub- jects only minutes old. The length of presentation of each exemplar varied. The assumed mouth position was held until the neonate became restless (e.g., b.v arching back~vard, crying, turning of its head) or be- came disturbed b v flies landing on its face. The mean number ofexem- plats of widened lips for each neonate was 9. I (ranging from 6 to 14) and for pursed lips, 8.5 (ranging from 4 to 14). The experimenter's lips were widened for a mean of 4.8 min, ranging from 2.4 to 9.0 min; they were pursed for a mean of 3. I min, ranging from 1.2 to 5.3 min.

Despite on-line efforts to present exemplars only when the baby was attentive, subsequent analysis of the video recordings indicated that these efforts were not always successful. Occasionally; a gesture ~as pre- sented when the baby was not attentive (see below). Moreover, during the presentation of some exemplars, the subjects raised their hands, blocking the view of their mouths. Furthermore, during the nonmodel- ing condition the infants often cried, had their faces averted, or had their eyes closed. Th us, the second stage in the derivation of the data entailed the selection of those episodes in which the neonates were attentive and their mouths were visible. For lip widening, this selection yielded a mean of 3.8 exemplars (ranging from 2 to 6) and for pursed lips, 3.6 (ranging from 2 to 5). Despite intersubjeet variability in the number and length of exemplars, the ensuing means for the total time of each condition were similar for the 12 babies: 73 s (ranging from 41 to 153 s) for the modeling of widened lips, 72 s (ranging from 35 to 155 s) for the modeling of pursed lips, and 76.8 s (ranging from l0 to 188 s) for the nonmodeling condition.

The following nine behaviors (see Figure 1) describe the babies' mouth movements within the technical limits of videorecording.

In the two modeling conditions, the experimenter moved her lips in a lateral or an anterior-posterior plane. Given that the babies were not expected to distinguish in their own mouth movements between lip and tongue movements, the coded behaviors pertaining to the coordinate muscle movements were combined in each of the two modeling condi- tions. Two groups of behaviors were coded. First, the following lateral movements, henceforth called v-wide, were coded: (a) lips broadened laterally and (b) lips moved downward. Second, the following anterior- posterior movements, henceforth called v-pursed, were coded: (e) lips

466 NADJA REISSLAND

LIP MOVEMENTS

1. Lips broadened laterally 2. Lips moved downward 3. Lips pursed and closed or nearly closed 4. Lips pursed and open vertically 5. Lips open and relaxed and not moving in any direction 6. Lips closed and relaxed and not moving in any direction

TONGUE MOVEMENTS

7. Tongue visible between lips 8. Tongue protruding beyond lips 9. Tongue not visible

Figure 1. Behaviors coded to establish whether the 12 neonates showed imitative capacity. (Categories 5 and 6 were not used in the final analysis of the data but were included in the statistical analysis of the reliability of the coding; Category 9 was included for the calculation of Cohen's Kappa and was not used for the analysis reported here.)

pursed or (d) pursed vertically and (e) tongue visible or (f) tongue pro- truding. The presence or absence of each of these behaviors was coded every half second, and the mean frequency of the two groups of behav- iors was calculated for each subject.

ReliabifiO'

.4ttentiveness. Neonatal attentiveness was assessed using an index of a number of separate behaviors that two independent judges coded at time intervals of every half-second from the filmed data (see Figure 2).

During the nonmodeling sessions, the baby was coded as attentive only if(a) the eyes were open and (b) the hands were either resting by the side of the baby or were waving without obstructing the view of the infant's face. During the modeling sessions, the baby was only coded as attentive if(a) the eyes were open and he or she was looking at the experimenter and (b) the hands were either resting by the side of the baby or were waving without obstructing the view of the infant's face.

The reliability of these six behaviors was established using Cohen's Kappa, a statistic correcting for chance agreements. Because the catego- ries tested have to be independent, mutually exclusive, and exhaustive, eye movement and hand movement were tested separately for reliability (Cohen, 1960. p. 38). Fleiss ( 1981 ) characterized kappas of .40 to .60 as fair, .60 to .75 as good, and over .75 as excellent. The agreement between the experimenter and Independent Judge I was good for eye movements (~ = 0.75) and hand movements (K = 0.68); the agreement between experimenter and Independent Judge 2 was excellent lbr eye movements (r = 0.84) and hand movements (~ = 0.83).

The attentiveness of the neonates during the episodes selected for cod- ing was corroborated by counting the neonate's eyeblinks during both types of modeling and during the nonmodeling session. One index of attentiveness (as well as of cognitive activity) is the endogenous eye blink (Stem, Walrath, & Goldstein, 1984). The number of blinks shown by each neonate was counted, and the mean number across all 12 neonates was established (agreement for eye blinks was 97.5% for the experi- menter and Judge I and 98.1% for the experimenter and Judge 3). The mean number of blinks per minute for the 12 babies was 10.7 when the model broadened her mouth, 9.3 when the model pursed her mouth, and 6.5 in the nonmodeling condition. These results lend support to the observation that the neonates were alert and attentive in the modeling and nonmodeling conditions.

Imitative capacitj~ The reliability of the nine behaviors (see Figure I) was established by two independent judges. Again Cohen's Kappa (Cohen. 1960) was used. Because the categories tested have to be inde-

pendent, mutually exclusive, and exhaustive, lip movement was tested for reliability separately from tongue movement.

Coding was done blindly in that the experimenter's face, which was normally visible in the upper-right-hand comer ofthe video screen, was covered. Illustrations of the neonate's facial expressions may be found in Figure 3.

The agreement between experimenter and Independent Judge 3 (on 16% of the sample) was good for both lip movements (r -- 0.75) and tongue movements (K = 0.70); between experimenter and Independent Judge I (on 42% of the sample) the agreement was excellent for both lip movements (r = 0.86) and tongue movements (K = 0.87).

Resu l t s

Behavior During Modeling and Nonmodeling Sessions

For each neonate the mean frequency of his or her mouth movements (v-wide and v-pursed combined) in the two model- ing conditions (M = 18.09, SD = 3.99) was compared with the mean frequency of his or her mouth movements (v-wide and v- pursed combined) in the nonmodeling condition (M -- 8.50, SD = 5.04) using a Wilcoxon matched-pairs signed-ranks test. The 12 neonates produced significantly more mouth move- ments in the modeling as compared with the nonmodeling con- dition, T = 0, N = 12, p < .001, two-tailed.

Neonatal Behavior in the Two Modeling Conditions

The critical test for imitative capacity was the selectivity o f the neonates' behavior in the two modeling conditions. A Wil- coxon matched-pairs signed-ranks test, comparing the mean frequency of v-wide versus v-pursed for each o f the 12 neonates, showed that (a) there were significantly more v-wide behaviors when the model 's lips were broadened (M = 9.09, SD = 5.57) than when her lips were pursed (M = 6.23, SD = 2.68), T = 10, N = I I, p < .025, one-tailed, and (b) significantly more v-pursed behaviors (M = 32.97, SD = 7.27) when the model 's lips were pursed than when they were broadened (M = 26.87, SD = 9.57) T = 13, N = 12, p < .025, one-tailed. Thus, the 12 neonates, from the moment o f opening their eyes minutes after birth, imi- tated the modeled mouth positions by moving their own mouths in either direction.

Neonatal Behavior in Relation to the Model's Behavior

Because the exemplars were not presented according to a strict t ime schedule, the positive results might be explained by the model 's having unwittingly imitated the neonate, rather

EYE MOVEMENTS

1. Eyes open and looking at the experimenter 2. Eyes open and looking elsewhere 3. Eyes closed or closing

HAND MOVEMENTS

4. Hands placed in mouth region 5. Hands or arms waving 6. Hands rest by the side of the body

Figure 2. Behaviors coded to establish attentiveness of the 12 neonates.

NEONATAL IMITATION IN THE FIRST HOUR OF LIFE 467

Ftgure 3. Examples of thcial expressions coded under the t~vo modeling conditions. (A. Neonate's tongue visible between lips; B. Neonate with vertically pursed lips: C. Neonate's tongue protruding beyond lips: D. Neonate's lips laterally broadened; E. Neonate's lips lateralls broadened; and E Neonate's lips moved do~ nward.)

than by the neonates' having imitated the model. One factor speaking against such a h~pothesis is that the timing of each exemplar was determined by the neonate+s attentiveness and that the exemplars were modeled alternately. Nonetheless, in order to test the alternative hypothesis the neonates" lip and

tongue positions were examined approximately I s before the start of each coded episode to determine whether they were in a v-pursed, v-x~.ide, or neutral position. In 91% of the episodes in which the 12 neonates were judged to be attentive, the mouth position of the baby was at variance with that of the model.

468 NADJA REISSLAND

In 36% of the exemplars the neonates' mouth was in a neutral position. In 55% of the exemplars the position of the neonates' mouth was opposite to that which the model assumed. Those 9% of exemplars in which model and neonate exhibited the same direction of lip movement were further examined, and it appeared that the v-pursed or v-wide positions shown by the neonates were typically due to circumstances beyond the con- trol of the experimenter. For example, one baby had just sneezed and showed a slightly pursed lip position; another baby had just ceased crying and still showed a slightly widened lip position. Hence, the counterhypothesis was rejected: The neo- nates did imitate the model's lip position, not vice versa.

Discuss ion

These results confirm the capacity of neonates to imitate fa- cial expressions; they also suggest that the contrary findings of other researchers might be more fruitfully considered as a theo- retical rather than a methodological issue. It may not be fortu- itous that previously, with the exception of Fontaine (1984), who studied infants between birth and 6 months, the positive results of imitation have been obtained from very young sub- jects of the varying ages of mean age 30 min (present study), 32 hr (Meltzoff& Moore, 1983a), 36 hr (Field, Woodson, Green- berg, & Cohen, 1982). 35 to 42 hr (Field, Woodson, Cohen, Greenberg, Garcia, & Collier, 1983), 4 days (Vinter, 1986) and mean age 14 days (Meltzotf& Moore, 1977), whereas the nega- tive results have been obtained from relatively older subjects of 9 to 30 days (McKenzie & Over, 1983a), mean age 21 days (Koepke, Harem, & Legerstee, 1983), and mean age 48 days (Hayes & Watson, 198 I). Such a pattern, whereby a certain be- havior or skill disappears only to reappear at a later age, has been observed in studies of fetal development (e.g., the flection response away from a stimulus; see Hofer, 1981) and of early infant development (e.g., reaching and walking skills; Hofer, 1981; see also Fontaine, 1984, and Maratos, 1982). On reap- pearance, the behavior or skill is more limited, but also more coordinated and it appears almost purposeful in the eyes of the observer.

Having set for neonates the task of imitating an act that they could not have seen in the extrauterine environment before modeling took place, the issue of what level of imitation neo- nates are capable of still remains. Some studies suggest that the level must be rather limited. First, newborns mostly attend to stimuli of high contrast with sharp contours (Fantz & Fagan, 1975; Salapatek, 1975). Second, even 4-month-old infants, who have been habituated to a schematic face, do not notice modi- fications in the mouth region (Caron, Caron, Cadwell, & Weiss, 1973). These findings seem to contradict the claim that newly born infants can reproduce mouth movements such as opening and closing the mouth and protruding the tongue. As Slater (1986) pointed out, however, newborns can detect perceptual invariants such as shape constancy. He suggested that results, such as Caron et al. (1973), are due to the fact that those stimuli that are seen most clearly by the baby will attract most atten- tion. This effect is so powerful that, if one has a pair of stimuli that are clearly different from each other but in which one is more detectable than the other, no amount of familiarization will shift this preference in favor of the less preferred stimulus. Hence, in the present study, in which they did not see any other

competing features to sidetrack their attention from the model's lips, it can be assumed that the neonates could detect the form of the lips as being either "wide," when the model widened her lips, or "narrow," when the model pursed her lips.

In sum, the results ofthis study confirm that neonates in their first hour postpartum are capable of acts of imitation. Because no social interaction took place prior to experimentation, the imitation cannot be explained by extrauterine learning experi- ence. Furthermore, because the subjects were Indo-Aryan in- fants of Maithii parentage, their acts of imitat ionmwhen con- sidered in combination with studies of Caucasian infants of North American or European parentagemconfirm that imita- tive capacity is present at birth in the human species.

References

AmieI-Tison, C. (1985). Pediatric contribution to the present knowl- edge on the neurobehavioral status of infants at birth. In J. Mehler & R. Fox (Eds.), Beyond the blooming buzzing confusion. London: Erlbaum.

Caron, A. J., Caron, R. E, Caldweli, R. C., & Weiss, S. J. (1973). Infant perception of the structural properties of the face. Developmental Ps3rholog3', 9, 385-399.

Cohen, J. A. (1960). Coefficient of agreement for nominal scales. Edu- cational and Psjrhological Measurement, 20, 37-46.

De Gelder, B. (1982). Imitation, knowledge and representation. In B. De Gelder (Ed.), Knowledge and representation. London: Roufledge & Kegan Paul.

Fantz, R.. & Fagan, J. E (1975). Visual attention to size and number of pattern details by term and preterm infants during the first six months. Child Development, 16, 3-18.

Field, T. M., Woodson, R., Cohen, D., Greenberg, R., Garcia, R., & Collier, K. (1983). Discrimination and imitation of facial expressions by term and preterm neonates. Infant Behavior and Development, 6, 485-489.

Field, T. M., Woodson, R., Greenberg, R., & Cohen, D. (1982). Dis- crimination and imitation of facial expressions by neonates. Science, 218, 179-182.

Fleiss, ( 1981 ). Statistical methods for rates and proportions. New York: Wiley.

Fontaine, R. (1984). Imitative skills between birth and six months. In- fant Behavior and Development, 7, 323-333.

Hayes, L. A., & Watson, J. S. ( 1981 ). Neonatal imitation: Fact or arti- fact? Developmental Ps)rholoD: 17. 655-660.

Hofer, M. A. ( 1981 ). The tools of human behavior San Francisco: Jos- sey-Bass.

Jacobson, S. W. (1979). Matching behavior in the young infant. Child Development, 50, 425-430.

Koepke, J. E., Hamm, M., & Legerstee, M. ( i 983). Neonatal imitation: Two failures to replicate. Infant Behavior and Development, 6, 97- 102.

Koepke, J. E., Hamm, M., Legerstee, M., & Russell, M. (1983). Meth- odological issues in studies of imitation: Reply to Meltzoff and Moore. Infant Behavior and Development, 6, I 13-116.

Lagercrantz, H., & SIotkin, T. A. (I 986). The "stress" of being born. Scientific American, 254, 92-102.

Lewis, M., & Sullivan, M. S. (1985). Imitation in the first six months of life. Merrill-Palmer Quarterly, 31, 315-333.

Maratos, O. (1982). Trends in the development of imitation in early infancy, in T. G. Bever (Ed.), Regression in mental development: Ba- sic phenomena and theories. Hillsdale, N J: Erlbaum.

Masters, C. (1979). Interpretative "imitative" responses in early infancy. Science, 205, 215.

McKenzie, B., & Over, R. (1983a). Young infants fail to imitate facial and manual gestures. Infant Behavior and Development, 6, 85-95.

NEONATAL IMITATION IN THE FIRST HOUR OF LIFE 469

McKenzie, B., & Over, R. (1983b). Do neonatal infants imitate? A reply to Meltzoff and Moore. Infant Behavior and Development, 6. 109- i11.

Meltzoff, A., & Moore, M. K. (1977). Imitation of facial and manual gestures by human neonates. Science, 198, 75-78.

Meltzoff, A., & Moore, M. K. (1983a). Newborn infants imitate adult facial gestures. Child Development, 31, 78-84.

Meltzoff, A., & Moore, M. K. (1983b). The origins of imitation in in- fancy: Paradigm, phenomena and theories. In L. P. Lipsett (Ed.), Ad- vances in infancy research (Vol. 2, pp. 265-301). Norwood, NJ: Ablex.

Meltzoff, A., & Moore, K. (1985). Cognitive foundations and social functions of imitation and intermodal representation in infancy. In J. Mehler & R. Fox (Eds.), Beyond the blooming buzzing confusion. London: Erlbaum.

Salapatek, P. (1975). Pattern perception in early infancy. In L. Cohen & P. Salapatek (Eds.), Infant perception. New York: Academic Press.

Slater, A. (1986, April). Visual perception at birth. Paper presented at the 5th Biennial International Conference on Infant Studies, Los An- geles.

Stern, J., Walrath, L. C., & Goldstein, R. (1984). The endogenous eye- blink. Ps)rhopht,sioloho: 21, 22-23.

Uzgiris, I. C. (1984). Imitation in infancy: Its interpersonal aspects. In M. Perlmutter (Ed.), Parent-child interaction and parent-child re- lations in child development (The Minnesota Symposia on Child Psy- chology, Vol. 17, pp. 1-32) Hillsdale, NJ: Erlbaum.

Vinter, A. (1986). The role of movement in eliciting early imitations. ChiM Development, 5 7, 66-7 I.

Received March 2, 1987 Revision received December 10, 1987

Accepted December 23, 1987 �9