REGULATION, ATTACHMENT SECURITY, AND REPRESENTATIONAL THOUGHT (PRETENCE AND FANTASY) George Gergely & Orsolya Kos
Summary Aims 1. To identify early individual differences among infants in using contingency detection, control seeking, maternal contingent reactivity, and affect-mirroring in emotional self-regulation at 6 months; 2. To test whether such differential patterns of affective self-regulation predict attachment security at 12 months with the specific aim of identifying characteristic patterns predicting Disorganized attachment; 3. To test how a) patterns of affect-regulative interactions at 6 months, and b) quality of infant attachment at 12 month correlate with the quality of cognitive-representational and affect-regulative use of pretence, imagination, and empathic competence during the 2nd year. Methods There were two studies with similar designs. In the first study (SEE COMPONENT 1), infants were assessed in a mirror interaction task (see below) at 6 months and attachment security and attachment disorganization were tested in the strange situation at 12 months. In the second study (SEE COMPONENT 2) the strange situation and mirror interaction were both assessed at 12 months. The pretence-empathy test battery was only done for the first study (see below). Data for this battery was also collected for the second study but has not yet been scored. Results 1. Our most significant finding is that Disorganized attachment at 12 months is predicted by a specific pattern of seeking out and maximizing high contingent control in 6-month-olds. These infants show differential success in affect-regulation by generating high contingent control over (both social and non-social) stimulus sources in the mirror following the stressful contingency deprivation still-face period of the MIS. A) they spend more time exploring their perfectly contingent self-movements in the mirror following the stress invoked by the contingency deprivation of the 2nd period (Still-face) compared to infants who will become securely attached at 12 months (click here to see a graph), B) they initiate more mirroring of their mother during episodes of high maternal contingent reactivity, and C) they show a higher proportion of successful affect-regulation both when exploring their self- movement and when engaging contingent maternal reactivity compared to infants who will become securely attached at 12 months (click here to see a graph). 2. A) We found that amount of maternal references to infant mental and affective states during the affect-regulative Phase 3 of the MIS at 6 months showed strong correlations with measures of advanced cognitive-representational pretence and empathic competence in the 2-year-olds. B) Measures of maternal affect-attunement and mother-initiated mirroring interactions during the MIS at 6 months showed strong positive correlations with cognitive-representational understanding of pretence as well as negative correlations with measures of pretence-reality confusion during the 2nd year. C) Amount of exploration of the perfectly contingent self-movements, especially during Phase 3 of the MIS at 6 months (that predicted Disorganized attachment at 12 months), showed intriguing negative correlations with (i) measures of using pretence to cope with negative affect, (ii) measures of empathy, and (iii) measures of cognitive-representational understanding of pretence. D) Amount of infant-initiated mirroring during maternal contingent reactivity in the MIS (a factor also correlated with Disorganization at 12 months) showed negative correlations with (i) measures of using pretence to cope with negative affect, (ii) measures of empathy, and (iii) measures of cognitive- representational understanding of pretence.
Implications for psychoanalysis This research has aimed at operationalizing and empirically testing within an attachment theory framework some of the most central psychoanalytic developmental concepts such as maternal affect- mirroring, attunement and holding that psychoanalysts have for long suggested to play a central role in early psychosocial development forming the core of functional self development, experiencing a sense of self agency, the internalization of affective control structures, and the formation of symbolic thought, mentalization, and social reality testing. These central constructs that capture the most important causal aspects of the quality of functional and adaptive affect-regulative attachment environments, can also direct our attention to the nature of those dysfunctional attachment contexts and their interaction with individual temperamental factors of vulnerability that may establish pathological developmental outcomes leading to developmental psychopathologies of the self such as borderline personality disorder, or dissociative or narcissistic personality disorders. One of the most important results of the current developmental research project that has attempted to empirically identify such psychoanalytic developmental mechanisms and their functional consequences for psychic development, is the identification of a possibly innate temperamental feature of contingent control seeking and maximizing that seems to lie at the basis of a dominant affect-regulative style of imposing perfect contingent control on the social and non-social environment in the face of experiencing the trauma of unpredictable loss of contingent control over the attachment environment. Such a control-dependent affect-regulative trait may function as a serious factor of vulnerability for dysfunctional attachment contexts characterized by repeated experiences of unpredictable loss of contingent control such as those experienced in parental abuse, neglect, maltreatment, and unresolved loss. The interaction of these environmental and temperamental factors together may provide the developmental background for the establishment of dissociative styles of defensive mentatalization strategies, involving inhibition of mentalizing about the other's intentional mind states in a defensive attempt to impose perfect control over the other's mind and actions through projection and projective identification. Our results also provide evidence that early dependency on high contingent control is detrimental to the development of representational abilities (as evidenced by a deficient ability for understanding pretend play and for affect-regulative use of fantasy and imagination) that are necessary both for the development of a functional ability to mentalize about other minds, and to avoid teleological concretistic thinking that leads to acting out, lack of emotional self-awareness, and deficits in reflective self function. Our research has also provided evidence that maternal affect mirroring, contingent maternal reactivity, and maternal attention and references to the infant's internal affective, intentional, and attentional states promote later representational functioning, mentalization and self-reflective capacities, and emotional self-awareness and control in the young child. In sum, our research has provided an important example of how empirical infancy research and clinically- based psychoanalytic developmental theory can mutually and beneficially influence each other leading to a much needed scientific grounding of often metaphoric clinical insights. Keywords Affect-mirroring, Control seeking, Attachment security, Attachment disorganization, Representational functioning, Pretence, Imagination, Fantasy, Borderline Personality Disorder, Projective Identification, Psychosocial origins of developmental psychopathologies of the self Bibliography Gergely, G., & Watson, J. S. (1996). The social biofeedback theory of parental-affect-mirroring: The development of emotional self-awareness and self-control in infancy. The International Journal of Psycho-Analysis, 77, 1-31.2. Gergely, G. (1998). Naissance de la capacit de rgulation des affects. In: J. Sacrispeyre (Ed.), Prendre soin d'un jeune enfant, (pp. 59-70), Toulouse: Societ d'Editions "Recherches et Syntheses" rs.3. Gergely, G. & Watson, J. S. (1999). Early social-emotional development: Contingency biofeedback model. In P. Rochat (Ed.), Early Social Cognition (pp. 101-137). Hillsdale, NJ: Erlbaum.4. Gergely, G. (2000). Mahler reapproached: New perspectives on normal autism, symbiosis, splitting and libidinal object constancy from cognitive developmental theory. Journal of the American Psychoanalytic Association, 48(4), 1197-1228.5. Gergely, G. (2002). The development of understanding self and agency. In U. Goshwami (Ed.), Handbook of Childhood Cognitive Development. Oxford: Blackwell.6. Gergely, G. (2002). Some confusion about pretence-reality confusions. Developmental Science, 5(4).7. Gergely, G., Kos, O., & Watson, J. S. (2002). Perception causale et role des comportements imitatifs des parents dans le dveloppement socio-motionnel prcoce. In J. Nadel & J. Decety (Eds.), Imiter pour dcouvrir l'human: Psychologie, neurobiology, robotique et philosophie de l'esprit. Paris: Press Universitaires de France.8. Kos, O., & Gergely, G. (2001). The 'flickering switch' hypothesis: A contingency-based approach to the etiology of disorganized attachment in infancy. Cognitive and interactional foundations of attachment. Special Issue of the Bulletin of the Menninger Clinic, 65(3), 397-410.9. Gergely, G. (2000) The development of the representation of self and others. Journal of Infant, Child and Adolescent Psychotherapy and Psychoanalysis, Vol. 1., No. 3., 25-32. 10. Fonagy, P. Target, M. & Gergely, G. (2000). Attachment and borderline Personality Disorder: A theory and some evidence. Psychiatric Clinics of North America, Vol. 23, No. 1. pp. 103-123.11. Nichols, K., Gergely, G., & Fonagy, P. (2001). Experimental protocols forInvestigating relationships among mother-infant interaction, affect regulation, physiological markers of stress responsiveness, and attachment. In J. Allen, (Ed.) Cognitive and interactional foundations of attachment, Special Issue of the Bulletin of the Menninger Clinic, 65(3), 371-379.12. Gergely, G., Fonagy, P., & Target, M. (2002). Attachment, mentalization, and the etiology of borderline personality disorder. Self Psychology, Vol. 7, No. 1, 61-72. COMPONENT STUDY 1
Aims 1. To identify early individual differences among infants in using contingency detection, control seeking, maternal contingent reactivity, and affect-mirroring in emotional self-regulation; 2. To test whether such differential patterns of affective self-regulation predict attachment security with the specific aim of identifying characteristic patterns predicting Disorganized attachment; 3. To test how a) patterns of affect-regulative interactions at 6 months, and b) quality of infant attachment at 12 month correlate with the quality of cognitive-representational and affect-regulative use of pretence, imagination, and empathic competence during the 2nd year. Methods Infants were assessed in a mirror interaction task (see below) at 6 months and attachment security and attachment disorganization were tested in the strange situation at 12 months. The pretence-empathy test battery was also done for this study.
A) The 'Mirror Interaction Situation' (MIS) (a modified Still-face procedure in front of a one-way mirror) at 6.5 months. Mother and infant are seated 2 ms next to each other facing a one-way mirror. They are separated by an occlusion screen so that they cannot touch each other, but can interact by facial and vocal gestures through the mirror (click here to see a diagram of the setup). Two cameras placed facing (but invisible to) them on the other side of the one-way mirror records their facial and vocal behaviours. These records are fed into a mixer that creates a synchronised time-coded split- screen record for off-line analysis. The situation consists of 3 1-min episodes: Phase 1.: 'Free interaction' in which mother can freely interact with the baby through the mirror; Phase 2.: 'Contingency deprivation' during which Mother puts on a neutral, motionless 'still-face'; and Phase 3.: 'Recovery period' in which Mother interacts freely again. (FOR DETAILS SEE: coding categories for the mirror interaction situation).
B) Infant attachment security at 12 months: The Strange Situation Test (SST)The SST is a standardized procedure for assessing quality of infant attachment at 12 months (Ainsworth et al., 1978). It involves two brief separations and two reunions with the parent and/or a stranger in an unfamiliar laboratory setting. Infant reactions to separation and reunion are coded by a standardized coding system that measures individual differences in the strategies to cope with separation-induced stress. It identifies four types of attachment patterns: Secure (B-type), Avoidant (A-type), Resistant (C-type), and Disorganised/Disoriented (D-type).
C) The 'Pretence-Empathy' test battery: it measures a) the level of cognitive-representational understanding of pretence, b) the quality of empathic competence and c) the quality of affect- regulative use of imaginative play.This battery of pretend play and empathy induction tasks developed by us is administered and videotaped in the family home. Different aspects of representational and affect-regulative functioning have been operationalized in these tasks (production and comprehension of pretend object-substitution, stipulation and extrapolation of make-belief identity, inferring make- belief actions and their causal consequences, differentiating reality from pretence, ease of 'switching' between these two representational modes, spontaneous empathic reactions, and elaboration, enrichment and modification of pretend scenarios for affective self-regulative purposes). FOR DETAILS SEE: the 'pretence-empathy' test battery
Results See the summary of the research program for details. CODING CATEGORIES FOR THE MIRROR INTERACTION SITUATION (MIS)
The response categories included the number, duration and quality of emotion expressions (positive, negative, neutral) of both mother and infant, the frequency and duration of looking at self vs. other for both mother and infant, the amount of imitative vs. nonimitative contingent facial and vocal interactions, the relative frequency of mother- vs. infant-initiated contingent and imitative (attuned) facial and/or vocal interactive episodes, etc. Below is the full list of behavioural categories used: Maternal behavioral categories: Gaze direction (duration, frequency); o Infant's mirror image o Other Affect state (duration, frequency); o Positive o Negative o Neutral Maternal verbalisations (duration) Content of maternal verbalisation (frequency) o Reference to infant's behavioural, affective or mental state, o Reference to mother's own presence, behavioural, affective or mental state, o Reference to object of infant's attention o Attention seeking, greeting verbal gestures Mother initiated Interactive Episodes (frequency) Mother initiated matching during Interactive Episodes (frequency) Affect attunement that take place outside the Interactive Episodes (frequency) o Verbal non-imitative but contingent acknowledgement of the baby's state, o Mother imitates the infant Non-attuned gestures (frequency); o Body o Vocal o Facial Attuned-gestures (frequency) during matching or affect attunement o Body o Vocal o Facial Infant's behavioral categories: Gaze direction (duration, frequency); o Mother's mirror image o Own mirror image o Own body o Looking away Contingency testing in the mirror (frequency) Affect state (duration, frequency); o Positive o Negative o Neutral o Infant initiated Interactive Episodes (frequency) Infant initiated matching during Interactive Episodes (frequency) Affect attunement that takes place outside the Interactive Episodes (frequency): o Infant imitates the mother The MIS data were analyzed by the Observer (Noldus) Program. The statistical analyses of the MIS data in relation to maternal and infant attachment groups were carried out using SPSS for Windows. Parametric (Multivariate ANOVA; planned pair-wise comparisons by t-tests) and non-parametric statistics (Wilcoxon, Mann-Whitney, binominal sign test, Chi-Square); Reggression and Corelation analyses. Return to Component 1 Return to Main Study THE 'PRETENCE-EMPATHY' TEST BATTERY This battery of pretend play and empathy induction tasks has been administered and videotaped in the family home by a trained experimenter (with the (passive) presence of the mother). We have developed the series of tasks have with consideration of comparability to previously used measures of representational functioning in pretend play available in the literature (such as Harris and Kavanaugh's (1993) representative study of the youngest age groups, Harris et al's (1991) procedure as modified by Golomb and Galasso (1995) testing emotionally charged pretence, or Golomb and Kuersten's (1996) paradigm to test 'switching' between reality vs. pretence modes of thinking). Different aspects of representational functioning have been operationalized in these tasks (such as ability to engage in and to interpret pretend object substitution, stipulation and extrapolation of make-belief identity, inferring make-belief actions and their causal consequences, ability to differentiate between reality and pretence, capacity to 'switch' between these two representational mode, and spontaneous elaboration, enrichment and modification of stipulated pretend scenarios to adaptively deal with emotionally charged pretence topics). In particular, the following five categories were tested by using several pretend tasks or scenarios for each: Task 1: Stipulation of make-belief identity (2 tasks); Task 2: Extrapolation of make-belief identity (2 situations with 3 different episodes in each); Task 3: Inferring make-belief actions and their imagined causal consequences (2 episodes); Task 4: Elaborative spontaneous use of pretend play for affect-regulative purposes: the ability to use make-belief to cope with negative affect induced in imagination (2 open-ended pretence scenarios (evoking separation and physical injury, respectively) to be completed by the child); Task 5: Evocation of empathic reactions towards the mother vs. towards a stranger (2 enacted episodes in which either the mother or the experimenter (stranger) acts as if being in pain due to an accidental injury). The children's reactions were independently coded by two trained coders using an objective scoring system developed by the Principal Investigator and the Co-investigator. Interrater reliability was high (kappa= .96). Each child received a combined numerical score along the following response categories: Categories measuring high metarepresentational functioning: 1. A combined index of pretence comprehension: This overall score of metarepresentational functioning was made up by the combination of three subscales measuring different representational aspects of understanding make-belief: 1.1. Understanding stipulated pretend identity (Task 1) 1.2. Ability to transform pretend identity based on contextual cues (Task 2) 1.3. Understanding make-belief actions and inferring imagined causal consequences in a pretend world (Task 3) 2. Spontaneous make-belief actions (This category measured the frequency of engaging in spontaneous pretence actions that were additional to the requested make-belief responses tested for by the different tasks) 3. Spontaneous use of imaginary objects (This category measured the frequency of using gestures (instead of actual objects with stipulated pretend identity) to indicate the presence of imaginary objects) Categories measuring low metarepresentational functioning: Reality-based interpretation (This measure consisted of the combined frequency score of the following two subcategories of reality-based interpretations: a) Real person substitution; and b) Real object use) Real person substitution (Instead of the stipulated pretend figure (doll) the child performs the pretend action on him/herself, on the experimenter, or on the mother) Real object use (instead of the stipulated pretend object (e.g., straw as toothbrush) the child uses a realistic object to perform the pretend action (e.g., brings his/her real toothbrush from bathroom to wash teddy's teeth with) Categories measuring the quality of coping with negative affect induced in imagination: (a) Primitive coping reactions (This category included the following types of responses to the open- ended scenarios depicting separation and physical injury: Refusal of continuing ('escape to reality'); freezing; denial; attempt to involve mother to resolve the situation) (b) Advanced coping reactions (This category included the following types of responses to the open- ended scenarios depicting separation and physical injury: the child him/herself acts as an active agent to resolve the source of negative affect (e.g., active soothing); the child uses dolls to act to resolve the source of negative affect) Categories measuring the quality of empathic reactions induced: (a) Nonempathic reactions (0) (This category included the following types of responses to apparent injury and expressions of pain by mother or stranger: Doesn't orient toward or react to inducing event; attends to inducing event, but shows no reaction; attends to event with expressing fearfulness) (b) Slightly empathic reactions (1) (This category included the following types of responses to apparent injury and expressions of pain by mother or stranger: Attention to event with empathic facial expression; asks about injury; acts to soothe but only upon being instructed) (c) Highly empathic reactions (2) (This category included the following types of responses to apparent injury and expressions of pain by mother or stranger: Attention to event with empathic facial expression accompanied by spontaneous and adequate act(s) to soothe the sufferer). (This new procedure has been devised by Gyorgy Gergely and Orsolya Koos. The detailed coding procedure can be obtained by contacting them.) Questionnaire: A set of 17 open ended questions on transitional object use, imaginary companion, play activities in the home, temperamental characteristics, emotional stability/instability of infant, emotion-regulative practices in the family, etc. (This new questionnaire has been devised by Gyorgy Gergely and Orsolya Koos. The detailed coding procedure can be obtained by contacting them)
Return to Component 1 Return to Main Study COMPONENT STUDY 2
Aims The aim of this study was to replicate the results of the previous study with a larger sample and also to look at the significance of attachment security as a moderator for the self-contingent responses induced by the still-face procedure in the mirror interaction situation. The design of this study was the same as study 1 except that infants were tested at 12 months both in the Strange Situation test and the mirror interaction situation (SEE COMPONENT 1) Methods See component study 1 for details of methods (SEE COMPONENT 1). The sample were 150 infants and mothers from whom we have valid data for x infant-mother pairs. Mothers were recruited from the community by advertisement. Results The results broadly replicated the findings of the previous study.
1. There were 92 secure and 47 insecure mother-infant pairs. There was a highly significant interaction between attachment security and the proportion of time that the infant spent exploring perfectly contingent self movements in the mirror during the stress invoked by contingency deprivation in the still-face period (SEE FIGURE 1). There was no difference between groups in the proportion of time spent in self-exploration in phase 1 of the procedure. There was a highly significant increase (p=.0001) in the time spent exploring perfectly contingent self movements in the mirror in the insecure group. Most of the insecurely attached infants returned to pre-stressor levels of self exploration in the third phase of the procedure so that the difference between secure and insecure infants in this phase was no longer statistically significant (p=.08)
2. There were 32 avoidant and 15 resistant infant-mother pairs in the insecure group. On the whole there were no important differences between the two insecure groups although both clearly differed from the secure infants (SEE FIGURE 2). Quite clearly there were differences between the 3 groups in terms of the quadratic effect (the size of the elevation in self exploration in phase 2) (p=.001). Avoidant infants were not quite so quick to return to baseline levels as resistant infants. Avoidant infants spent somewhat more time in the final phase exploring self-contingent behaviour than secure infants (p=.09). Both avoidant and resistant infants were highly significantly different from secure infants in the second phase.
3. There were few disorganized infants in the sample. Only 11 infants had disorganization ratings of 5 or more and 20 infants 4 or more. When we contrasted these 20 infants with the remaining 119, we found that the 2 groups differed substantially in terms of the amount of contingent mirror self- exploration they engaged in (SEE FIGURE 3). Unlike the impact of attachment security, attachment disorganization had particularly marked effects in the 3rd phase of the procedure. Both disorganized and non-disorganized infants increased their self exploration in the second phase relative to the first phase, but whereas non-disorganized infants returned to baseline levels, infants rated as disorganized during separation & reunion continued self-exploration even after the mother started reacting spontaneously to her infant. While for the disorganized infants there was a significant elevation (.03) of this behaviour in the 3rd phase relative to the 2nd, for non-disorganized infants there was a highly significant reduction (.0001)
4. Exploring the secure and insecure groups separately (SEE FIGURE 4), an interesting marginally significant pattern emerges. (p<.057). It seems that while for securely attached infants who are nevertheless rated as disorganized in their attachment the elevation of self-exploration occurs only in the period following the still-face procedure, for insecurely attached disorganized infants the elevation during the still-face procedure is considerably higher and remains high for the third phase of the procedure. Thus in the third phase disorganized infants respond with self-exploration whether they are secure or insecure, but only insecure disorganized infants appear to do so in the still-face procedure.