Attachment Theory

What is Attachment Theory?

Attachment theory is a psychological theory, developed by British psychiatrist John Bowlby, which proposes that infants are biologically predisposed to form a close emotional bond with at least one caregiver, and that the quality of that bond shapes how they handle closeness, distress and relationships later on. Research by Mary Ainsworth and others identified four broad patterns in young children: secure, anxious-resistant (ambivalent), avoidant and disorganized.

At the heart of the theory is the idea that early experiences build internal working models: mental expectations about whether other people will be available and responsive, and whether the self is worthy of care. These models are not fixed scripts, but they tend to guide how people seek comfort and respond to closeness and separation. Attachment theory sits within developmental psychology and has strongly influenced research on emotion regulation, parenting, relationships and therapy.

Core principles

• A biological system. Bowlby argued that attachment evolved because staying close to a protective adult increased an infant's chances of survival.
• Safe haven. When frightened, tired or ill, a child turns to the attachment figure for comfort.
• Secure base. When the child feels safe, the attachment figure becomes a base from which to explore the world.
• Internal working models. Repeated experiences with caregivers become expectations about relationships.
• Continuity with change. Patterns tend to show some stability over time, but they can and do change with new experiences.

Origins and Development of Attachment Theory

John Bowlby: the foundation

Bowlby (1907–1990) trained as a psychoanalyst, but he broke with the dominant view that infants love their mothers mainly because mothers feed them. Influenced by ethology, including Konrad Lorenz's work on imprinting in geese and Harlow's findings that infant monkeys preferred a soft cloth "mother" to a wire one that provided milk, he argued that the need for closeness is a primary, built-in motivation. He also emphasized real experiences of care and separation rather than children's fantasies, and he worked with James and Joyce Robertson, whose films of young children separated from parents in hospitals helped change hospital visiting policies. See our guide to Harlow's monkey experiments.

Bowlby described attachment developing in phases: pre-attachment (birth to about 6 weeks, when babies respond to anyone), attachment-in-the-making (about 6 weeks to 6–8 months, with growing preference for familiar people), clear-cut attachment (about 6–8 months to 18–24 months, with separation protest and wariness of strangers) and a goal-corrected partnership (from about age 2, as children begin to understand caregivers' plans and can negotiate). The ages are approximate and vary between children.

Mary Ainsworth: the Strange Situation

Mary Ainsworth (1913–1999) worked with Bowlby in London and then studied mothers and infants in Uganda and in Baltimore. To measure attachment, she and her colleagues created the Strange Situation, a roughly 20-minute laboratory procedure for infants aged about 12 to 18 months. In a series of short episodes, the infant plays in an unfamiliar room with the caregiver, meets a friendly stranger, is briefly left with the stranger, is reunited with the caregiver, is left entirely alone, and is then reunited again.

The key observations are not how upset a baby gets when the caregiver leaves, which varies with temperament, but how the baby behaves at reunion: whether they seek contact, are comforted, and return to play. In her Baltimore sample, Ainsworth classified about two-thirds of infants as secure, about a fifth as avoidant, and about an eighth as resistant. Ainsworth also found that mothers' sensitivity to their babies' signals at home predicted secure attachment, a link later confirmed in meta-analyses, though the association is modest rather than overwhelming.

Later developments

Mary Main and Judith Solomon noticed that some infants did not fit Ainsworth's three categories and showed contradictory or disoriented behavior; they called this disorganized attachment. Main also developed the Adult Attachment Interview (AAI), which classifies adults by how coherently they talk about their childhood relationships. Hazan and Shaver (1987) showed that adult romantic relationships can be studied through an attachment lens, starting a large social-psychology research tradition. Kim Bartholomew (1990) proposed a four-category adult model built on views of self and others.

The Four Attachment Styles

The four styles below were first described in infants. Rough prevalence figures come from normative (non-clinical) samples; a frequently cited cross-national analysis by van IJzendoorn and Kroonenberg (1988) found about 65% secure, 21% avoidant and 14% resistant among children classified using Ainsworth's three categories, and later meta-analyses put disorganized attachment at around 15% in low-risk samples, higher in high-risk ones.

Important considerations

Insecure patterns are not disorders, and they are not "bad." Ainsworth and later researchers viewed them as reasonable adaptations: an infant whose bids for comfort are often rebuffed may learn to minimize them, while one whose caregiver is unpredictable may learn to amplify them. A person can also show different patterns with different people, for example with each parent. And while the styles help explain tendencies, they do not determine anyone's future.

How Attachment Develops

Attachment develops through thousands of everyday interactions in the first years of life. Researchers have highlighted several aspects of caregiving:

• Sensitivity: noticing a baby's signals, interpreting them accurately and responding promptly and appropriately. Ainsworth considered this central.
• Consistency: caregiving that is reasonably reliable over time, so the child can predict what will happen.
• Attunement and repair: no caregiver gets it right every time. Research by Edward Tronick and others suggests that ordinary mismatches followed by repair are normal and may help children learn that disconnection can be fixed.
• Mind-mindedness: Elizabeth Meins's term for a caregiver's tendency to treat the infant as a person with thoughts and feelings, which has been linked to secure attachment.

Children also bring their own characteristics. Temperament, the inborn tendencies toward calm or reactivity, affects how infants behave in the Strange Situation and how easy they are to soothe. Psychologist Jerome Kagan argued that temperament explained much of what attachment research attributed to parenting. Most researchers now see attachment as shaped by both caregiving and child characteristics, along with wider stresses on families such as poverty, illness and conflict.

The attachment behavioral system

Bowlby described attachment as a behavioral system, a bit like a thermostat. It switches on when a child is frightened, tired, ill, in an unfamiliar situation, or when the caregiver seems unavailable, prompting crying, calling, following or clinging. It quiets when the child regains closeness and feels safe, freeing them to explore again. The balance between attachment and exploration is what the Strange Situation is designed to observe.

Biological foundations

Neuroscience research links close relationships to stress-regulation systems such as the hypothalamic-pituitary-adrenal (HPA) axis, which releases cortisol, and to reward circuits involving dopamine. The hormone oxytocin is involved in bonding in many mammals, although popular claims about it as a simple "love hormone" go well beyond the human evidence. Studies of children raised in severely depriving institutions, such as the English and Romanian Adoptees study led by Michael Rutter, show that early deprivation can have lasting effects, and also that many children improve substantially after adoption into caring families.

Adult Attachment

Hazan and Shaver proposed that romantic love is itself an attachment process: partners become each other's safe haven and secure base. Adult attachment is usually measured in one of two ways. Developmental researchers use interviews such as the AAI. Social and personality psychologists use questionnaires such as the Experiences in Close Relationships (ECR) scale (Brennan, Clark and Shaver, 1998), which measures two dimensions:

• Attachment anxiety: worry about rejection and whether partners will be available.
• Attachment avoidance: discomfort with closeness and depending on others.

People low on both dimensions are described as relatively secure; high anxiety corresponds to the "preoccupied" or anxious style, high avoidance to the "dismissing" style, and high scores on both to the "fearful-avoidant" style. These scores describe tendencies, not diagnoses.

How stable is attachment?

Longitudinal studies show that attachment security has moderate stability from infancy into adulthood, with substantial change for many people. A meta-analysis by R. Chris Fraley (2002) found a modest correlation between infant and later attachment security. Change is more likely after significant life events, such as parental divorce, loss or abuse, and after new supportive relationships. Adults whose AAI interviews show coherent, reflective accounts of difficult childhoods are sometimes described as having earned security.

Impact on Relationships

Romantic relationships

Research links greater attachment security with higher relationship satisfaction, more constructive conflict and more effective support-seeking, although the effects are moderate and many other factors matter. People higher in attachment anxiety tend to monitor partners closely for signs of rejection and may escalate conflicts; people higher in avoidance tend to withdraw under stress and minimize emotional discussions. When an anxious partner pursues and an avoidant partner withdraws, each person's response can intensify the other's, a "pursuer-distancer" pattern often discussed by couple therapists.

Recognizing these patterns as learned strategies rather than character flaws can help partners understand each other. Our guide to attachment styles and attachment in dating looks at everyday relationship dynamics in more detail. Online quizzes that sort people into a single "type" should be treated with caution; validated measures and professional assessment are more reliable.

Parenting and intergenerational transmission

A meta-analysis by Marinus van IJzendoorn (1995) found that parents' AAI classifications predicted their infants' Strange Situation classifications substantially better than chance, suggesting that attachment patterns can carry across generations. However, parental sensitivity explains only part of this link, a puzzle known as the transmission gap. Parents' capacity to reflect on their child's mental states (reflective functioning) appears to play a role, and many parents who experienced difficult childhoods raise securely attached children.

Friendships, school and work

Attachment ideas have been applied to friendships, teacher–student relationships and the workplace. Studies suggest that children with secure histories tend, on average, to show somewhat better social competence with peers, and that warm relationships with teachers can support children who have experienced adversity. Research on attachment and leadership or team behavior exists but is less developed, and findings should not be overgeneralized.

Cultural Perspectives and Criticisms

Attachment theory is one of the most researched theories in developmental psychology, but it has important critics.

• Culture. Distributions of attachment patterns vary between countries. Early studies found relatively more avoidant classifications in some German samples and more resistant classifications in some Japanese samples. Critics argue that the Strange Situation reflects Western norms about independence and that brief separations mean different things in cultures where infants are rarely left with strangers. Research in communities where many adults share childcare also challenges the focus on a single mother figure.
• Monotropy. Bowlby initially emphasized one primary attachment figure. Later research, beginning with Schaffer and Emerson's 1964 study of Glasgow infants, showed that babies commonly form multiple attachments, including to fathers, siblings and other caregivers.
• Maternal deprivation. Michael Rutter's 1972 critique, Maternal Deprivation Reassessed, argued that many harms Bowlby attributed to separation reflected broader privation, family discord and poor-quality care. This helped dispel guilt-inducing claims that working mothers or high-quality childcare inevitably harm children.
• Determinism and parent-blaming. Popular accounts can imply that early experience fixes adult personality, or that every relationship difficulty traces back to parents. The evidence shows moderate continuity and plenty of change, and outcomes are shaped by genes, temperament, later relationships and circumstances.
• Measurement. Different measures of adult attachment correlate only modestly, so research findings depend partly on which tool is used.

Changing Attachment Patterns

Attachment patterns can change, and research on "earned security" suggests that later relationships and reflection play a part. Experiences associated with change include:

• Supportive relationships with partners, friends, mentors or other family members who are consistently responsive.
• Reflection on one's relationship history, which can build a more coherent understanding of past experiences.
• Psychotherapy, in which the relationship with the therapist can itself function as a secure base. Research consistently links a strong therapeutic alliance with better outcomes.

Attachment-informed therapies

Several therapies draw on attachment theory. Emotionally focused therapy (EFT), developed by Sue Johnson and Les Greenberg, helps couples identify negative cycles and the attachment needs beneath them; randomized trials support its effectiveness for relationship distress. Mentalization-based treatment and some psychodynamic approaches also draw on attachment concepts. For families, parenting interventions such as video-feedback programs and Attachment and Biobehavioral Catch-up (ABC), developed by Mary Dozier, aim to increase caregiver sensitivity; trials suggest they can improve parenting and child outcomes in some groups. Approaches vary in their evidence base, and people interested in therapy can learn more in our guide to finding a therapist.

Clinical and Practical Applications

Assessment tools

These are research and clinical tools that require training to administer and interpret. They are not diagnostic tests, and a self-report score does not label a person.

Mental health

Research in clinical psychology has found associations between insecure attachment and a higher risk of various difficulties, including depression and anxiety, and between disorganized attachment and later problems with emotion regulation. These are associations across groups, not predictions for individuals, and attachment is one factor among many. It is important not to confuse attachment styles with clinical conditions. The DSM-5-TR does include two rare childhood diagnoses linked to severe neglect, reactive attachment disorder and disinhibited social engagement disorder, which are distinct from ordinary insecure attachment and are diagnosed only by qualified professionals.

Education and childcare

Attachment research has influenced hospital visiting policies, foster care and adoption practice, and the design of early childcare, such as key-person systems in nurseries that give each child a consistent caregiver. In schools, it informs approaches that emphasize predictable routines and warm teacher–student relationships, especially for children who have experienced adversity.