Mary Main
Mary Main studied under Mary Ainsworth at Johns Hopkins and spent the majority of her career at UC Berkeley, where her contributions reshaped attachment research in three decisive ways. First, with Judith Solomon she identified the disorganized attachment category in 1986, expanding Ainsworth's original three-category classification and opening the research field to the specific pattern of children whose caregivers had been simultaneously source of comfort and source of fear. Second, she developed the Adult Attachment Interview — a semi-structured protocol that assesses attachment organization not by what adults say about their childhoods but by how they structure the telling. Third, her research on adults who had achieved secure attachment despite difficult childhood histories established the concept of earned security, demonstrating that internal working models formed in childhood could be revised through later relational experience. For the AI transition, Main's work is the empirical foundation of genuine hope: the working models disrupted by technological change can be rebuilt.
In the AI Story
Main's graduate training with Ainsworth positioned her at the center of attachment research during its most productive period. Her initial work extended Ainsworth's Strange Situation by examining the small number of infants whose behavior did not fit the three existing categories. The disorganized category she and Solomon identified captured infants whose behavior showed incompatible strategies — approaching and freezing, moving toward the caregiver while looking away — reflecting the irreconcilable conflict produced by caregiving that was simultaneously source of comfort and source of fear.
The Adult Attachment Interview, developed over the 1980s, represented a methodological breakthrough. Main's insight was that attachment organization in adults could be assessed linguistically — through the coherence, balance, and flexibility of the person's narrative about her childhood attachment experiences. Adults who could describe difficult childhoods with coherent acknowledgment of both pain and growth showed the cognitive markers of secure attachment. Adults whose narratives were dismissive, incoherent, or emotionally flooded showed the markers of insecure patterns.
The earned security finding emerged from this methodology. Some adults reported difficult childhoods but showed secure linguistic markers. These adults had not simply grown out of difficulty — their secure functioning was not a denial of their histories but an integration of them. Through some process (often therapy, sometimes transformative relationships), they had revised their working models while remaining fully aware of the originating experiences. The revision was real, the security was functionally equivalent to continuous security, and the research methodology could demonstrate both.
For the AI moment, Main's research specifies the conditions under which adaptation is possible. Her work establishes that working model revision is real and achievable, that it requires specific relational conditions, and that the outcome can be stable and robust. It also establishes that the revision takes time — measured in years, not weeks — and that it cannot be accomplished in isolation. The implications for institutional practice during the AI transition follow directly.
Origin
Mary Blanchard Main was born in Amherst, Massachusetts, in 1943. She earned her PhD in developmental psychology from Johns Hopkins University in 1973 under Mary Ainsworth's supervision.
She spent her academic career primarily at UC Berkeley (1973 until retirement), where she directed the Social Development Project. She married psychologist Erik Hesse, with whom she collaborated extensively on attachment research. She died in Berkeley in 2023.
Key Ideas
Disorganized attachment. The fourth category added to Ainsworth's original three, capturing infants whose caregivers were simultaneously source of comfort and source of fear.
Adult Attachment Interview. The semi-structured protocol that assesses attachment organization through the coherence of adult narratives about childhood.
Linguistic markers of security. Attachment organization shows in how people talk about their histories — the structure of the telling, not merely its content.
Earned security. The empirical demonstration that working models can be revised across the lifespan through specific relational conditions.
Foundation for AI-era hope. Main's research provides the empirical basis for the claim that adaptation to AI-driven disruption is genuinely possible, though the conditions for it are specific and demanding.
Appears in the Orange Pill Cycle
Further reading
- Mary Main and Judith Solomon, 'Discovery of an Insecure-Disorganized/Disoriented Attachment Pattern' (Affective Development in Infancy, 1986)
- Mary Main, 'Cross-Cultural Studies of Attachment Organization' (Human Development, 1990)
- Mary Main et al., 'Adult Attachment Classification System' (unpublished manuscript, 2002)
- Erik Hesse, 'The Adult Attachment Interview' (in Handbook of Attachment, 2008)