Professional Documents
Culture Documents
10 - Weinberg - and - Tronick-Stillface Paradigm
10 - Weinberg - and - Tronick-Stillface Paradigm
10 - Weinberg - and - Tronick-Stillface Paradigm
WEINBERG, M. KATHEHINE, and THONICK, EDWARD Z . Infant Affective Reactions to the Resump-
tion of Maternal Interaction after the Still-Face. CHILD DEVELOPMENT, 1996, 67, 905-914. To
investigate infants' reactions to the Face-to-Face Still-Face Paradigm and in particular the re-
union episode, 50 6-month-olds' affective, behavioral, and physiologic reactions were recorded
and analyzed. Infants reacted to the still-face with negative affect, a drop in vagal tone, and an
increase in heart rate. By contrast, they reacted to the reunion episode with a mixed pattern of
positive and negative affect. There was a carryover of negative affect from the still-face, an
increase in fussiness and crying, and a rebound of positive affect. During this episode, the
infants' heart rate and vagal tone returned to initial levels. The data indicate that infant affective
displays are specifically related to different interactive events, but that their physiologic reactions
do not show the same level of specificity. The findings also highlight the complexity of the
affective and reparatory processes that take place in mother-infant interactions.
The Face-to-Face Still-Face Paradigm mond, Langhorst, & Demos, 1982; Gusella,
has been used extensively to evaluate young Muir, & Tronick, 1988; Mayes & Carter,
infants' communicative abilities, sensitivity 1990; Stack & Muir, 1990, 1992; Stoller &
to changes in maternal behavior, and capac- Field, 1982; Toda & Fogel, 1993), infants
ity to regulate affective states (Tronick, typically respond to the still-face episode as
1989). The paradigm confronts the infant contrasted to the first play episode with de-
with three age-appropriate conditions: (1) a creases in smiling and gazing at the mother
face-to-face social interaction with the and increases in motor activity; touching and
mother; (2) a still-face episode during which grasping of the self, clothing, or chair; with-
the mother assumes a still or poker face and drawal; and, according to some but not all
remains unresponsive to the infant; and (3) reports, an increase in negative affective fa-
a reunion episode of face-to-face social inter- cial and vocal displays including grimaces,
action with the mother. Each episode typi- distress brows, and crying.
cally lasts 2 or 3 min, and the paradigm has
been used with infants ranging in age from A number of interpretations have been
2 to 9 months (see, e.g., Toda & Fogel, 1993; advanced to account for the still-face effect.
Weinberg & Tronick, 1994). One interpretation is that the still-faced
mother violates the infant's expectation for
Most research on the paradigm has fo- a normal interaction (Field et al., 1986;
cused on the still-face episode. During this Tronick et al., 1978) or is behaving in a man-
episode, mothers are asked to assume a still, ner too discrepant from the infant's schema
neutral face, to look at the infant, but not of her normal behavior (McCall & JCagan,
touch or talk to the infant. As described by 1967). Another interpretation is that by with-
Tronick (Tronick, Als, Adamson, Wise, & drawing all forms of contact with the infant,
Brazelton, 1978; see also Carpenter, Tecce, the mother no longer provides the infant
Stechler, & Friedman, 1970) in the study with the regulatory input necessary to main-
that originated the paradigm and as repli- tain an organized social and affective state
cated in several other projects (Field, Vega- (Stack & Muir, 1990). Research by Muir (Gu-
Lahr, Scafidi, & Goldstein, 1986; Fogel, Dia- sella et al., 1988; Stack & Muir, 1990, 1992)
This research was supported by grants from NSF (BNS 85-06987) and NIMH (ROl MH
45547 and ROl MH 43398) awarded to Edward Z. Tronick. The authors wish to thank Beatrice
Beebe, Jeff Cohn, and the reviewers for their helpful comments on the manuscript. The authors
also thank Karen Nelson, Margaret Kelsey, Wendy Copes, Rachel Buch, and Michael Shively
for their assistance with coding and data collection, as well as Barry Lester for his analysis of
the heart rate data. We also thank the mothers and infants who made this research possible.
Correspondence concerning this article should be addressed to M. Katherine Weinberg, Child
Development Unit, Children's Hospital, 1295 Boylston Street, Boston, MA 02215.
[Child Development, 1996, 67,905-914. © 1996 by the Society for Research in Child Development, Inc.
All rights reserved, 0009-3920/96/6703-0019$01.00]
906 Child Development
demonstrates that if the mother is allowed sionally followed by the infant arching his
to touch the infant during the still-face, the or her back away from the mother "as if he
infant's negative reaction is attenuated. Al- had not forgiven her the previous insult" (p.
ternatively, Tronick (1980, 1989) has pro- 10). Field (Field et al., 1986) further found
posed that the mother's lack of respon- that infants show distress after the resump-
siveness disrupts the infant's goal for social tion of maternal behavior. She reported that
engagement. This disruption generates neg- infant motor activity, distress brow, and cry-
ative emotional states along with attempts ing increased during the still-face episode
by the infant to regulate these states by dis- and remained high during the reunion epi-
engaging and utilizing self-comforting be- sode. Similarly, Fogel found that infants
haviors. This explanation argues that the in- cried more during the reunion episode than
fant has a goal for relating to people and that during the other two conditions (Fogel et al.,
when there are obstacles to the achievement 1982; Toda & Fogel, 1993) and smiled less
of this goal the infant is stressed and nega- during the reunion episode than in the first
tive emotional states are generated (Trevar- play (Fogel et al., 1982). Other researchers,
then, 1977). however, have not reported a continuation
or increase of affectively negative displays
The reunion episode of normal interac- following the resumption of normal maternal
tion, which follows the still-face, has re- interactive behavior. Gusella, Muir, and
ceived scant attention in the literature. This Tronick (1988) found no evidence that the
is unfortunate because this episode may reunion episode was as or more stressful
highlight the affective and dyadic regulatory than the still-face. Rather they found in-
processes that take place in the mother- creases in infant looking and smiling at the
infant interaction. Tronick and Cohn (1989; mother during the reunion episode. These
see also Beebe & Lachmann, 1994) have de- findings in the literature are consistent with
scribed normal mother-infant interactions as an interpretation that infants experience
characterized by numerous interactive er- both positive and negative affective reac-
rors or mismatches that are quickly repaired. tions when their mother resumes interactive
The interaction moves rapidly from coordi- behavior.
nated to miscoordinated states and back
again. During the reunion episode the
mother and infant must repair the interac- In this study, we microanalytically ex-
tion following a prolonged interactive error amined the differential distribution of 6-
(i.e., the still-face). The reunion episode month-old infants' affective reactions to the
episodes ofthe Face-to-Face Still-Face Para-
therefore presents the infant with an af- digm. In particular, we were interested in
fectively complex and demanding regulatory evaluating the nature ofthe infants' reaction
task. The infant must simultaneously cope to the reunion episode. Most ofthe previous
with the resumption of maternal behavior (in work on the Face-to-Face Still-Face Para-
all likelihood a positive event for the infant) digm, though microanaiytic, has coded a
and cope with the intra- and interpersonal priori clusters of behaviors (e.g.. Monadic
carryover of negative affect from the still- Phases; see Tronick, Als, & Brazelton, 1980)
face. We would expect that these confiicting that have typically grouped facial expres-
demands would be expressed in a mixed pat- sions with other behaviors. In this study, we
tern of positive and negative affective dis- coded infant facial expressions with Izard's
plays that may not be observed in interac- AFFEX system (Izard & Dougherty, 1980)
tions that are not preceded by an interactive separately from the infants' other behaviors
disturbance. (i.e., direction of gaze, vocalizations, ges-
tures, self-comforting, withdra^val, and auto-
The data available in the literature sup- nomic stress indicators), which were coded
port the perspective that the reunion epi- with the Infant Regulatory Scoring System
sode elicits a mixed affective reaction in (IRSS; Tronick & Weinberg, 1990). Many
young infants. Several researchers have de- of these IRSS displays, and in particular
scribed a carryover of negative affect from AFFEX-coded facial expressions, have not
the still-face to this episode (Field et al., been previously examined in this paradigm.
1986; Fogel etal., 1982; Tronick etal., 1978). Additionally, the infants' autonomic reactiv-
For example, in the original paper using the ity as indexed by heart rate and vagal tone
Face-to-Face Still-Face Paradigm, Tronick was assessed during the episodes of the
(Tronick et al., 1978) reported that infants Face-to-Face Still-Face Paradigm. It was ex-
show an initial period of wary monitoring pected that the still-face episode in particu-
and anger when the mother first resumes lar would be an effective stressor that would
normal interactive behavior, which is occa-
' Looking at objects was coded if the infant looked at an object for 2 sec or more. This
coding criterion was used in order to distinguish between sustained object engagement and
scanning of the environment, which was defined as looking at something for less than 2 sec.
Weinberg and Tronick 909
TABLE 1
MEAN PROPORTIONS ( M P ) , STANDARD DEVIATIONS (SD), AND NUMBER OF INFANTS ( N ) W H O DISPLAYED
THE AFFEX-CoDED FACIAL EXPRESSIONS OF JOY, INTEREST, SADNESS, AND ANGER DURING PLAY 1,
THE STILL-FACE, AND REUNION PLAY
FACIAL EXPRESSION MP SD N MP SD N MP SD JV
Joy (.89) 26a .22 45 .06b .09 37 .32c .25 47 39.32**
Interest (,95) 61a ,22 50 .72b .24 50 .51c .25 50 17.33**
Sadness (.82) 00a ,01 8 .05h .11 22 .03b ,06 22 6,33*
Anger (.89) 02a ,07 16 ,07b .17 28 .07b .15 22 3,91*
NOTE,—AFFEX codes are mutually exclusive. Mean interrater reliability for each AFFEX code is in parentheses.
Mean propoitions with different letters differ significantly at j> < ,05,
*p< ,05,
** p < ,01,
face box. This interface box electronically Internal Validity and Manipulation Check
calculated the time between adjacent R- Following Cohn and Elmore (1988) and
waves of ECC and generated a signal whose Toda and Fogel (1993), a manipulation
voltage was proportional to the inter-R-wave check was carried out on the mothers' be-
interval. This signal was then digitized by havior during the still-face episode to deter-
computer (using a DataTranslation DT2821 mine whether the mothers complied with
A/D board) and reconverted into inter-beat- the instructions for maintaining a still-face.
interval (IBI) values measured in millisec- Eighteen mothers smiled during the still-
onds. Sequential IBIs were used to calculate face for a total of 108 sec or a total of 1.8% of
heart rate (HR) values. The HR values were all the still-face data. As in Toda and Fogel's
then analyzed with MXEdit Vagal Tone (VT) study, these smiles lasted an average of 6
Software, and VT estimates were computed sec. Furthermore, as in Cohn and Elmore's
for 10-sec epochs across the entire duration work, the percentage of time that the moth-
of each of the episodes of the Face-to-Face ers smiled during the still-face was signi-
Still-Face Paradigm (for details see Lester et ficantly less than in the Play 1 or reunion
al., 1990). episodes (see Table 1). These results dem-
TABLE 2
MEAN PROPORTIONS (MP), STANDARD DEVIATIONS (SD), AND NUMBER OF INFANTS (A?) W H O DISPLAYED
EACH I R S S - C O D E D BEHAVIOR DURING PLAY 1, THE STILL-FACE, AND REUNION PLAY
BEHAVIOR MP SD MP SD N MP SD N
Look at Mother (.87) 39a ,25 50 .24b .15 50 .53c .25 50 33.41**
Look at Objects (.96) 42a .23 50 ,50a ,21 50 .30b .24 46 14.43**
Scans (.92) 19a ,11 50 .24b ,12 50 ,14c ,10 50 14.63**
Neutral/Positive Vocalizations (.75) 08a .13 42 .05a ,06 37 .18b ,19 43 21.58**
Fussy Vocalizations (.81) 05a ,11 19 ,07a .13 21 .10b .14 31 3,85*
Crying (,77) 00a .01 1 .Ola .04 3 .06b .20 7 5.08*
Pick-Me-Up Gesture (.77) Ola .03 5 ,04b ,07 22 .04b .10 12 3.96*
Gestural Signals (.77) 15a .19 41 .09a ,11 42 ,28b .23 48 19.87**
Mouthing Body Part (1,00) 04 ,09 15 .02 .06 19 ,05 ,13 17 2.37
Mouthing Object (.96) 05 .10 18 .03 .08 15 ,04 ,08 19 1.03
Distancing/Escape (.91) Ola .03 12 .03b .08 19 .03b .08 15 3.18*
Autonomic Stress Indicators (.80) Ola ,03 13 .04b .08 21 .03b ,08 16 3.28*
NOTE,—The categories Look at Mother, Look at Objects, and Scans are mutually exclusive. The other behavioral
categories can co-occur with Look at Mother, Look at Objects, and Scans, and with each other. Mean interrater
reliability for each IRSS code is in parentheses. Mean proportions with different letters differ significantly at p < ,05,
* p < ,05,
**p< ,01,
910 Child Development
onstrate the effectiveness of the still-face least likely to scan during the reunion epi-
manipulation. sode. Thus the infants looked less at the mother
during the still-face and spent a considerable
Results amount of time looking at objects and scan-
ning. During the reunion episode, there was a
Specificity of AFFEX-Coded Facial rebound of looking at the mother above the
Expressions by Condition "baseline" level observed in the first play.
To evaluate the speciflcity of the At the same time, looking at objects and
AFFEX-coded facial expressions and IRSS- scanning decreased to their lowest levels.
coded behaviors by condition, a repeated-
measures, one-way ANOVA with episodes The infants displayed more IRSS-coded
as repeated measures was carried out. Post- behaviors communicating a negative evalua-
hoc tests were further used to explore sig- tion of the interaction in the reunion episode
nificant findings. As can be seen in Table 1, than in the other two episodes. Fussy vocal-
there were main effects of episode for the izations and crying were most likely to occur
AFFEX facial expressions of joy, interest, during the reunion episode. Furthermore,
sadness, and anger. Facial expressions of joy pick-me-up gestures, attempts at physically
occurred significantly more during the two distancing the self by turning and twisting
play episodes than during the still-face and in the seat, and autonomic stress indicators
significantly more during the reunion epi- were unlikely to occur during the first play,
sode than the first play. Facial expressions but increased significantly during the still-
of interest occurred frequently in all three face and remained at this higher level during
conditions but were most likely to occur dur- the reunion episode. There were no signifi-
ing the still-face. Facial expressions of sad- cant differences in the incidence of these be-
ness and anger were unlikely to occur dur- haviors during the still-face and reunion epi-
ing the first play. Sad and angry facial sodes.
expressions, however, increased signifi-
cantly in the still-face and remained at this The infants also displayed more IRSS-
higher level during the reunion episode. coded behaviors communicating a positive
There was no significant difference in the evaluation of the interaction in the reunion
incidence of sadness and anger facial expres- episode than in the other two episodes. Neu-
sions during the still-face and reunion epi- tral/positive vocalizations and gestures other
sodes. These data indicate that infants in- than pick-me-ups were significantly more
hibit facial expressions of joy and display likely to occur during the reunion episode
facial expressions of sadness and anger than during either the first play or the still-
when their mothers act in an unresponsive face. Neutral/positive vocalizations and ges-
manner. The data further indicate that they tures were equally likely to occur during the
respond with a rebounding of facial expres- first play and the still-face. These data indi-
sions of joy above the "baseline " level of the cate that the infants displayed a mixed pat-
first play episode but not a decrease in facial tern of positive and negative IRSS-coded be-
expressions of sadness and anger when the haviors during the reunion episode that was
mothers resume normal interactive be- not observed in either the first play or the
havior. still-face.
displays measured by AFFEX facial expres- infants continued to show facial expressions
sions and IRSS behaviors and the episodes of sadness and anger, pick-me-up gestures,
of the Face-to-Face Still-Face Paradigm was distancing, and autonomic stress indicators
supported. The infants' affective displays at levels not significantly different from
were specifically related to the interactive those observed in the still-face. The infants
event the infant was experiencing. During were also significantly more likely to fuss
the first play episode, the infants were gen- and cry in the reunion episode than in the
erally positive and showed little negative af- still-face. These data are consistent with the
fect. During the still-face episode, as com- interpretation in the literature of a carryover
pared to the first play episode, the infants effect of negative affect from the still-face to
displayed significantly fewer facial expres- the reunion episode and indicate that infant
sions of joy and looked less at their mother. negative affective states are not easily as-
They also showed significantly more facial •^uaged by the resumption of maternal inter-
expressions of interest. This is consistent active behavior (Field et al., 1986; Fogel et
with research by Toda and Fogel (1993) al., 1982; Tronick et al., 1978). It is also pos-
showing that 6-month-old infants react to the sible that the increase in fussiness and cry-
still-face with reduced smiling and gazing at ing reflects the infants' expectation of ma-
the mother and increases in neutral facial ternal intervention to help regulate their
expressions. The infants also showed sig- negative affect.
nificantly more affective displays communi-
cating a negative evaluation of the mothers' On the other hand, the reunion episode,
behavior during the still-face than the first unlike the still-face, was also characterized
play. Facial expressions of sadness and by positive affect. There was a rebound of
anger, scanning, pick-me-up gestures, dis- positive mother-oriented behaviors to levels
tancing by twisting and turning in the seat, even higher than those observed in the first
and autonomic stress indicators such as hic- play. Facial expressions of joy, looking at the
cuping and spitting up were more likely to mother, neutral/positive vocalizations, and
occur during this episode than during the gestural signals directed toward the mother
first play. These data suggest that infants re- were all significantly more likely to occur
act negatively to their mother posing the during this episode than during either the
still-face, although there is relatively little first play or the still-face. Moreover, the in-
fussing and crying. Furthermore, the con- fants were least likely to look at objects and
figuration of increased facial expressions of scan the laboratory environment during the
interest, decreased looking at the mother, reunion episode. These data are consistent
and high level of object engagement sug- with an interpretation that the infants wel-
gests that focusing attention at something comed the resumption of maternail interac-
other than the mother may be a form of dis- tive behavior.
traction, coping, and emotional regulation
employed by the infants during the still-face These findings highlight the importance
(Brazelton, Koslowski, & Main, 1974; Cia- of the reunion episode as a critical but much
nino & Tronick, 1988). neglected component of the Face-to-Face
Still-Face Paradigm. The hypothesis that in-
During the reunion episode, the infants fants would display a mixed pattern of posi-
showed a different pattern of responses than tive and negative affective displays during
during the still-face. On the one hand, the the reunion episode that would not be ob-
reunion episode, like the still-face, was char- served in the other two episodes was sup-
acterized by negative affective displays. The ported. The reunion episode was character-
912 Child Development
ized by a carryover of negative affect from bivalence in the infant. It is possible that
the still-face, an increase in fussiness and the infant has both a goal of repairing and
crying, and a rebound of positive affect. Why resuming the interaction and a goal of with-
might this emotional complexity manifest it- drawing from the interaction as a result of
self in this episode in particular? One possi- the mother's behavior in the still-face. This
bility is that maternal behavior was responsi- emotional ambivalence is reminiscent of
ble for the infants' reaction to the reunion data reported by Campos, Emde, Gaens-
episode. Mayes (Mayes, Carter, Egger, & bauer, and Henderson (1975). They found
Pajer, 1990) has suggested that some moth- that infants express the widest range of af-
ers may not resume their usual interactive fective reactions and display the highest lev-
style after the still-face. She found that moth- els of smiling and sobering in response to
ers who felt uncomfortable during the still- the departure of a stranger. Campos's find-
face returned to the interaction with more ings may be indicative of the infant's con-
soothing comments. In previous research, flicting goals to both interact with and with-
however, we found few changes in maternal draw from the stranger.
behavior between the first play and reunion
episodes (Weinberg, 1992). Further exami- The hypothesis that autonomic reactiv-
nation of this issue is needed. ity would reflect the predominant affective
state associated with each of the episodes
Alternatively, the complexity of the in- of the Face-to-Face Still-Face Paradigm was
fant's reaction to the reunion episode may partially supported. The heart rate and vagal
reflect the affective and dyadic regulatory tone data confirm the stressful nature of the
processes that take place in the mother- still-face episode. Heart rate went up and
infant interaction. Tronick and Cohn (1989) vagal tone went down during the still-face
have characterized normal interactions as a episode compared to the first play episode.
sequence of dyadic matching states associ- This pattern of autonomic reactivity indi-
ated with positive affect and mismatching cates that the infants experienced the still-
states associated with negative affect. The face as stressful and conforms to the primar-
process of transforming mismatching states ily negative affective displays observed
into matching states has been labeled inter- during the still-face. The heart rate and vagal
active reparation. From this perspective, in- tone data during the reunion episode, how-
fants experience positive and negative affect ever, are not so easily interpreted. During
even during well-coordinated interactions. this episode, the infants' heart rate and vagal
However, the predominant affect is positive tone returned to the levels observed during
given the mother-infant dyad's capacity for the first play episode. Were the affective dis-
interactive reparation. During the still-face, plays similar in the first play and reunion
the infant is faced with an irreparable inter- episodes, the data could be interpreted as
active error. Even afler attempts to change the infants' heart rate and vagal tone re-
maternal behavior and repair the interaction, turning to nonstressful levels. However, the
the mother remains unresponsive. The inter- AFFEX facial expressions and IRSS behav-
active error continues, the reparation fails, iors displayed by the infants were very dif-
and, as a result, infant negative affect pre- ferent in these two episodes. In the first
dominates. By contrast, in the reunion epi- play, the infants were generally positive and
sode, the mother and infant can begin the showed little negative affect. In the reunion
process of reparation, which, if successful, episode, the infants displayed a mixed pat-
leads to positive affect. However, infant neg- tern of both positive and negative affect.
ative affect is not easily assuaged during the
reunion episode because carryover effects One interpretation of these data is that
from the still-face make reparation difficult. expressed negative affect and autonomic
This difficult process of reparation is indi- measures may not be tightly coupled. Cun-
cated by the high incidence of fussing and nar (Cunnar, Mangelsdorf, Larson, & Herts-
gaard, 1989) found dissociations between
crying at levels higher than those observed behavioral and physiologic measures of
in the first play or the still-face. Thus the stress. The data reported here are consistent
complexity of the infants' affective experi- with her interpretation that infants may not
ence in the reunion episode reflects the dy- be as physiologically stressed as their af-
ads' attempt to renegotiate their typical in- fective displays suggest. Thus, the data indi-
teraction and to cope with the negative intra- cate that, although the infants' affective dis-
and interpersonal aftermath of the still-face. plays are specifically related to different
Relatedly, one could speculate that the interactive events, their physiologic reac-
reunion episode arouses an emotional am- tions do not show the same level of specific-
Weinberg and Tronick 913
ity. An alternative hypothesis is that positive on the organization of behavior in 3-month-
infant affect in response to the resumption old infants. Infant Behavior and Develop-
of maternal interactive behavior buffers the ment, 11, 493-505.
infant against lingering stress and negative Cohn, J. F., & Tronick, E. Z. (1987). Mother-infant
affect, and that this buffering is indexed by face-to-face interaction: The sequence of dy-
a return to nonstressful levels in autonomic adic states at 3, 6, and 9 months. Develop-
measures. In autonomic terms, this may rep- mental Psychology, 23, 1-10.
resent a balancing of the regulatory input of Field, T., Vega-Lahr, N., Scafldi, F., & Goldstein,
the parasympathetic and sympathetic sys- S. (1986). Effects of maternal unavailability
tems (J. Kagan, personal communication). on mother-infant interactions. Infant Behav-
This balancing would be consistent with the ior and Development, 9, 473-478,
hypothesis that infants experience strong Fogel, A., Diamond, G. R., Langhorst, B. H., &
ambivalent emotions during the reunion ep- Demos, V. (1982), Affective and cognitive as-
isode. pects of the 2-month-old's participation in
face-to-face interaction with the mother. In E.
In sum, the infants' reactions to the epi-
Tronick (Ed.), Social interchange in infancy:
sodes of the Face-to-Face Still-Face Para-
Affect, cognition, and communication (pp.
digm evidence specificity as well as com-
37-57). Baltimore: University Park Press,
plexity in relation to the communicative and
Fogel, A., Nwokah, E., Dedo, J, Y., Messinger,
interactive context of each episode. This
D., Dickson, K. L., Matusov, E,, & Holt, S. A.
complexity is interpretable in terms of the
(1992). Social process theory of emotion: A
meaning of the events to the infant and the
dynamic systems approach. Social Develop-
communicative interactive process context
ment, 1, 122-142.
in which they are embedded. Thus, as we
have argued elsewhere (Weinberg & Tron- Gianino, A., & Tronick, E. Z. (1988). The mutual
ick, 1994; see also Fogel et al., 1992; Trevar- regulation model: The infant's self and inter-
then, 1977), the infants' affective displays active regulation, coping, and defense. In T.
form configurations that are related to the Field, P. McCabe, & N. Schneiderman (Eds,),
infants' affective states and to interactive Stress and coping (pp. 47-68). Hillsdale, NJ:
contexts. Eribaum.
Gunnar, M. R,, Mangelsdorf, S., Larson, M., &
Hertsgaard, L. (1989), Attachment, tempera-
References
ment, and adrenocortical activity in infancy:
Beebe, B,, & Lachmann, F, M, (1994), Representa- A study of psychoendocrine regulation. De-
tion and internalization in infancy: Three velopmental Psychology, 25, 355-363,
principles of salience. Psychoanalytic Psy- Gusella, J. L., Muir, D., & Tronick, E, Z. (1988),
chology, 11, 127-165. The effect of manipulating maternal behavior
Brazelton, T. B., Koslowski, B., & Main, M, (1974), and interaction on three- and six-month-olds'
The origins of reciprocity: The early mother- affect and attention. Child Development, 59,
infant interaction. In M. Lewis & L. A. Rosen- 1111-1124.
blum (Eds.), The effect of the infant on its Izard, C, E, (1977). Human emotions. New York:
caregiver (pp. 49-76). New York: Wiley. Plenum.
Campos, J., Emde, R., Gaensbauer, T., & Hender- Izard, C. E., & Dougherty, L. M. (1980). A system
son, C, (1975). Cardiac and behavioral interre- for identifying affect expressions by holistic
lationships in the reactions of infants to judgements iAFFEX). Newark: University of
strangers. Developmental Psychology, 11, Delaware, Instructional Resources Center.
589-601, Izard, C, E., & Malatesta, C. Z. (1987). Perspec-
Carpenter, G, C , Tecce, J, J., Stechler, G,, & tives on emotional development: 1. Differen-
Friedman, S. (1970), Differential visual be- tial emotions theory of early emotional devel-
havior to human and humanoid faces in early opment. In J. D. Osofsky (Ed.), Handbook of
infancy, Merrill-Palmer Quarterly of Behav- infant development (2d ed., pp, 494-554).
ior and Development, 16, 91-108, New York: Wiley.
Cicchetti, D, V., & Feinstein, A, R. (1990). High Lester, B., Boukydis, Z., McGrath, M., Censullo,
agreement but low kappa: I. The problems of M,, Zahr, L., & Brazelton, T. B. (1990). Behav-
two paradoxes, youmaZ of Clinical Epidemiol- ioral and psychophysiologic assessment of the
ogy, 43, 543-549. preterm infant. In B. M. Lester & E. Z. Tron-
Cohen, J. F, (1960). A coefficient of agreement for ick (Eds.), Stimulation of the preterm infant:
nominal scales. Educational and Psychologi- The limits of plasticity (pp. 155-171). Clinics
cal Measurement, 20, 37-46, in Perinatology, 17(1).
Cohn, J, F., & Elmore, M. (1988). Effect of contin- Mayes, L. C , & Carter, A. S. (1990). Emerging
gent changes in mothers' affective expression social regulatory capacities as seen in the
914 Child Development
still-face situation. Child Development, 61, Penticuff (Eds,), The exceptional infant: Psy-
754-763. chosocial risks in infant environment trans-
Mayes, L. C , Carter, A, S., Egger, H, L., & Pajer, actions (pp, 144-158). New York: Brunner/
K. A, (1990). Reflections on stillness: Mothers' Mazel.
reactions to the still-face situation, Jotirnai of Tronick, E. Z. (1989). Emotions and emotional
the American Academy of Child and Adoles- communication in infants. American Psychol-
cent Psychiatry, 30, 22-28, ogist, 44, 112-126,
McCall, R, B,, & Kagan, J. (1967). Stimulus Tronick, E, Z., Als, H., Adamson, L., Wise, S,, &
schema discrepancy and attention in the in- Brazelton, T, B, (1978). The infant's response
fant./ourna/ of Experimental Child Psychol- to entrapment between contradictory mes-
ogy, 5, 381-390. sages in face-to-face interaction. Journal of
Melson, G, F,, & Fogel, A, (1982), Young chil- the American Academy of Child Psychiatry,
dren's interest in unfamiliar infants. Child 17, 1-13.
Development, 53, 693-700. Tronick, E. Z,, Als, H,, & Brazelton, T. B. (1980).
Stack, D. M,, & Muir, D, W. (1990), Tactile stimu- Monadic phases: A structural descriptive
lation as a component of social interchange: analysis of infant-mother face-to-face interac-
New interpretations for the still-face effect, tion. Merrill-Palmer Quarterly of Behavior
British Journal of Developmental Psychol- and Development, 26, 1-24,
ogy, 8, 131-145, Tronick, E. Z,, & Cohn, J. F, (1989). Infant-mother
Stack, D. M., & Muir, D, W. (1992). Adult tactile face-to-face interaction: Age and gender dif-
stimulation during face-to-face interactions ferences in coordination and the occurrence
modulates five-month-olds' affect and atten- of miscoordination. Child Development, 60,
tion. Child Development, 63, 1509-1525, 85-92.
Stoller, S. A,, & Field, T, (1982), Alteration of Tronick, E. Z., & Weinberg, M. K. (1990). The In-
mother and infant behavior and heart rate fant Regulatory Scoring System (IRSS). Un-
during a still-face perturbation of face-to-face published document. Children's Hospital/
interaction. In T, Field & A, Fogel (Eds,), Harvard Medical School, Boston,
Emotion and early interaction (pp, 57-82). Weinberg, M, K, (1992). Sex differences in 6-
Hillsdale, NJ: Eribaum, m.onth-old infants' affect and behavior: Im-
Toda, S,, & Fogel, A, (1993). Infant response to pact on maternal caregiving. Doctoral Dis-
the still-face situation at 3 and 6 months, De- sertation, University of Massachusetts,
velopmental Psychology, 29, 532-538, Amherst,
Trevarthen, C, (1977), Descriptive analysis of in- Weinherg, M. K,, & Tronick, E, Z. (1994). Beyond
fant communicative behavior. In H, R, Schaf- the face: An empirical study of infant affective
fer (Ed,), Studies of mother-infant interac- configurations of facial, vocal, gestural, and
tion (pp. 227-270). London: Academic Press. regulatory behaviors. Child Development, 65,
Tronick, E. Z, (1980), On the primacy of social 1503-1515.
skills. In D, B. Sawin, L. O, Walker, & J, H,