Conference Agenda

This paper presents evidence from the DFG-funded ERIN study showing that the Adult Attachment Interview (AAI) captures generalized differences in individuals’ communicative capacities across contexts, independent of whether attachment relationships are discussed. First, we demonstrate that AAI classifications can be predicted by coding discourse elements supporting evidential reasoning (e.g., descriptions of physical actions, value judgments, and expressions of propositional attitudes) in short segments of the AAI, independent of overall transcript coherence. Second, we show that these same discourse elements reliably predict AAI classifications even when coding responses about work rather than attachment relationships. Finally, we examine whether AAI classifications are associated with self-reports of learning from parents, further testing the interpretation of the AAI as reflecting differences in communication rather than differences in processing attachment-related information. Together, these results support reconceptualizing the AAI as a measure of broader socio-epistemic capacities in adults, with important implications for developmental and clinical research.

The intergenerational transmission of attachment is traditionally conceptualized as a process by which parents’ attachment representations shape their caregiving behavior, which in turn influences their children’s expectation of their availability for support (i.e. attachment patterns). In this paper, we present empirical data suggesting a different hypothesis: that intergenerational transmission of attachment reflects infants adapting their epistemic trust to their caregivers’ reliability as communicators. Specifically, we build on findings showing that caregivers’ AAI classifications reflect generalized differences in communication across contexts, as outlined in Paper 1. We then extend these findings by showing that mothers’ AAI classifications predict how mothers communicate with their 15-month-old children during a free-play interaction, particularly in terms of establishing epistemic trust. In this context, we introduce a new instrument, the Parent Epistemic Trustworthiness Scale (PETS), and present data from 180 mother-infant dyads demonstrating its concurrent validity with the AAI and parental sensitivity. Finally, we test a mediation model in which parents’ communication styles during an employment interview predict infants’ attachment security, with this effect mediated by parents’ epistemic trustworthiness as measured by the PETS.

Attachment patterns in infancy are often described in terms of differences in proximity-seeking, reflecting expectations about the caregiver’s availability to protect the infant. However, they may also reflect cognitive and epistemic processes, particularly the infant’s epistemic trust in the caregiver. Drawing on the ERIN study, we first test whether infants with different attachment patterns (secure, avoidant, resistant) differ in their information-seeking behavior in a selective learning paradigm that contrasts infants’ responses to caregivers’ versus strangers’ information-giving. In a second study, we present initial results examining infants’ behavior during free-play interactions with their caregiver, testing whether SSP-assessed attachment patterns predict differences in infants’ information-seeking even in a non-alarming context. Together, these findings suggest that differences in attachment behavior may offer a window into early expectations about the reliability and usefulness of social communication.

This will be the discussion for this symposium/panel.