
Definition
Bowlby's attachment theory holds that human infants are born with an innate behavioral system that drives them to seek and keep closeness to a caregiver, especially when frightened, tired or ill. Early experiences of whether that caregiver is available and responsive shape internal working models, expectations about other people and about oneself that guide close relationships later in life.
When a case hands back the early mirror
Imagine a case, a composite built for teaching. Marcela is 34 and comes to a counseling center saying: "I don't understand why I keep ending up in relationships where I feel invisible." She works as a designer, has kept friends since college and describes her working life as fine. Romantically, though, she sees a pattern that dates back to her teens: she picks emotionally distant partners, steps up her attention when she senses withdrawal, and feels she must keep proving she is still loved. Asked about her early years, she describes no flagrant trauma. Her mother went through a hard postpartum depression, and Marcela spent her first months at her maternal grandmother's house. She was not mistreated, but she cannot recall a moment when she could predict whether approaching her would bring comfort or indifference. What she describes is a discontinuity: a caregiver who was present but not reliably available.
That story is the kind of thing attachment theory was built to describe. John Bowlby proposed that first bonds are not just a warm backdrop to childhood but the mold on which a person builds expectations about availability, trust and emotional regulation. What Marcela experiences as "invisibility" can be read as the continuation of an adaptive strategy she learned before she had words for it. This guide covers John Bowlby's ideas, his four phases of attachment, the internal working models, Mary Ainsworth's Strange Situation and its patterns, and what research does and does not show about how far early attachment reaches into adult life. Where the evidence is modest, the text says so.
The short version
- John Bowlby (1907–1990) was a British psychiatrist and psychoanalyst who developed attachment theory from the 1950s on, mainly in his trilogy Attachment and Loss (1969, 1973, 1980).
- He argued that attachment is a primary motivational system with its own biological function, protection from danger, and not a by-product of feeding.
- He described four phases of attachment: pre-attachment, attachment-in-the-making, clear-cut attachment and goal-corrected partnership.
- Internal working models are largely unconscious expectations about whether others will be available and whether the self is worth responding to.
- Mary Ainsworth's Strange Situation identified secure, avoidant and ambivalent (resistant) patterns; Main and Solomon later added disorganized.
- Attachment patterns show moderate, not total, continuity across the life span, and they can change.
What is Bowlby's attachment theory?
Bowlby proposed that infants are born with a behavioral system oriented toward keeping proximity to a caregiver. That system does not aim only at food or physical protection. Its goal is a felt sense of security, regulated through the closeness of someone who responds predictably.
Several sources fed into his thinking. Bretherton (1992) traces them in detail. One was his clinical work in London, including a study of juvenile thieves in which many of the young people had experienced early, prolonged separations from their mothers. Another was a World Health Organization commission in 1950 to report on the mental health of homeless children, which led to the 1951 report on maternal care and mental health. A third was ethology: Konrad Lorenz's work on imprinting and Robert Hinde's studies of primate social behavior, along with Harry Harlow's experiments with infant monkeys, which showed that the young monkeys' attachment to a cloth surrogate depended on contact comfort and not on being fed (Harlow, 1958).
From these lines came a proposal that was radical for its time. Classical psychoanalysis tended to treat the infant's tie to the mother as derived from the satisfaction of oral needs. Bowlby argued that attachment is a primary system with an evolutionary function of its own.
Attachment behavior is not activated only by hunger or cold. It is activated by threat, novelty, tiredness and separation, and its function is to restore a feeling of safety through the caregiver's availability.
This formulation has concrete consequences. If attachment is a primary system, the quality of the early response a child receives is not a minor environmental detail: it structures expectations about how the relational world works. Bowlby and later researchers gathered longitudinal evidence suggesting that early differences in caregiver availability predict, modestly but reliably, later patterns of emotional handling and social behavior.
The theory does not say that destiny is sealed in the first two years. It says that early relationships build a starting point, a set of expectations about others and about oneself, which later relationships confirm, correct or leave in place. A better question than "what happened to you at age two?" is "what model of availability did you learn, and how do you apply it today?"
Short definition for the exam
Bowlby's attachment theory states that humans have an innate motivational system that leads them to seek and maintain closeness to a caregiver under threat or stress, and that early experiences of response or its absence shape internal working models that guide expectations and behavior in relationships across the life span.
The four phases of attachment according to Bowlby
Bowlby (1969/1982) described attachment as developing through four sequential phases. Each adds a new capacity. The order is stable; the ages are approximate and vary across textbooks.
| Phase | Approximate age | Characteristics |
|---|---|---|
| 1. Pre-attachment | Birth to about 6 weeks | The baby orients to and signals people without discriminating among them; responds to voice and touch but shows no marked preference |
| 2. Attachment-in-the-making | About 6 weeks to 6–8 months | Preference for familiar people appears; the baby distinguishes and actively seeks particular caregivers |
| 3. Clear-cut attachment | About 6–8 months to 18 months or 2 years | Active protest at separation, wariness of strangers, use of the caregiver as a secure base for exploring |
| 4. Goal-corrected partnership | From about 2 to 3 years | The child begins to understand the caregiver's goals and plans and can negotiate; separations can be tolerated better |
Phase 3 is the one textbooks usually call attachment proper. It is where the behavior Bowlby considered diagnostic becomes visible: the child uses the caregiver as a secure base, goes out to explore and comes back when threatened or tired. If the base responds consistently, exploration flows more freely; if it is unpredictable, exploration is interfered with, either by vigilance or by inhibition.
Phase 4 is easy to miss and easy to test. The child now grasps that the caregiver has goals and plans of her own, which opens the door to negotiation ("Will you come back soon?"). This is related to what later literature calls mentalizing, though Bowlby did not use that word.
Internal working models
Bowlby introduced a concept that links the theory to adult life: internal working models. They are largely unconscious representations, built in early interaction, that organize two basic questions: Is the other person available when I need them? and Am I the kind of person others will respond to?
From repeated experience with a caregiver, the child builds an operating scheme. It is not a theory the child can put into words but a tacit script that switches on in each relationship. If the caregiver responds reliably, the child builds a model in which closeness is dependable and distress may be expressed. If the caregiver responds erratically, the child learns that closeness is uncertain and that showing distress can bring relief or rejection.
An internal working model is not a memory you recall; it is a lens you look through. It filters how you read another person's behavior before you have time to question it.
Such models tend toward stability, but not through determinism. The mechanism is confirmation: people select relationships that fit their expectations, interpret the other person's behavior in light of them and, in doing so, tend to reproduce the model. That is one way to understand why someone who expects people to be unavailable might find an attentive partner dull or untrustworthy while staying in an unstable relationship that fits the old expectation.
Stability is not immutability. Therapy, a steady partner over years or a parenting experience that rearranges priorities can revise a model. Change is usually gradual and tends to take repeated corrective experiences, not a single insight.
Ainsworth and the Strange Situation: attachment patterns
Mary Ainsworth, Bowlby's close collaborator, developed a laboratory procedure known as the Strange Situation, reported in its classic form by Ainsworth and Bell (1970). It is a short sequence of separations and reunions involving a one-year-old (typically 12 to 18 months), the caregiver and a stranger, in a playroom with toys. The key observation is the child's behavior at reunion, not how much the child cries during the separation.
Ainsworth identified three organized patterns. A fourth, described later by Main and Solomon, covers behavior that did not fit the first three (Main & Solomon, 1990).
| Pattern | Behavior at reunion | Relational strategy |
|---|---|---|
| Secure | Seeks contact, is comforted by the caregiver, returns to play | Trust in availability; distress is expressed and settled through closeness |
| Avoidant | Does not seek contact, ignores the caregiver, resumes play without signaling distress | Minimizes distress, so as not to activate a caregiver expected to be rejecting |
| Ambivalent (resistant) | Mixes seeking contact with resistance and anger; is not easily soothed | Maximizes distress to hold the attention of a caregiver expected to be unpredictable |
| Disorganized | Contradictory behavior: freezing, or approaching with the head turned away | Strategy breaks down when the caregiver is both a source of alarm and a source of refuge |
The clinical reading of this table is not moralizing. The avoidant and ambivalent patterns are adaptive strategies: the child does what works in that environment. The avoidant child learns that showing need brings no response and withdraws the signal. The ambivalent child learns that the caregiver responds only to an intense signal and turns the signal up. The secure pattern is associated with sensitive, responsive caregiving, not with a parent who satisfies every wish. A meta-analysis of 66 studies found that parental sensitivity is moderately related to infant security, with an association of about r = .24 in the closest-to-Ainsworth studies (De Wolff & van IJzendoorn, 1997). That is a real but modest link, a reminder that sensitivity is not the whole story.
The disorganized pattern deserves a separate note. A meta-analysis of nearly 80 studies concluded that it can be coded reliably, that it shows modest stability and that about 15% of infants in normal middle-class families show it, with the figure higher in other contexts and clinical groups (van IJzendoorn et al., 1999). The paradox is that the person who should calm the child is the same one who alarms the child, and the attachment system finds no coherent way out.
Does early attachment mark the whole life?
The question is often posed too bluntly. The evidence points to moderate continuity. A meta-analysis of attachment stability from infancy to adulthood found moderate stability rather than either strong continuity or none (Fraley, 2002). In a 20-year follow-up of 50 people first seen in the Strange Situation at 12 months, 72% received the same secure-versus-insecure classification in early adulthood, and negative life events such as a parent's death, divorce, serious illness or abuse were an important factor in change (Waters et al., 2000). The sample was small and middle-class, so treat the figures as illustration, not as a universal rate.
Three mechanisms are commonly proposed to explain why early attachment can have long influence. Keep in mind that these are explanations the theory offers, and that research supports them to differing degrees.
- Continuity of the working model. Early representations filter how one reads others' behavior. A person with an avoidant model may read a partner's request for closeness as pressure on their autonomy; one with an ambivalent model may read distance as imminent abandonment. Those readings produce behavior that tends to draw responses that seem to confirm them.
- Selection of partners and settings. People tend to choose relationships consistent with their model. Someone who expects unavailability may feel more at ease with a distant partner than with an attentive one, because steady closeness feels unfamiliar and, oddly, threatening.
- Transmission of regulation. Secure early attachment is associated with a greater capacity to regulate distress, seek support flexibly and sustain intimate relationships. Insecurity does not prevent these capacities but makes them less fluid.
It is just as important to note what the theory does not say. It does not say that a mother who worked outside the home produces insecure attachment. It does not say that every child with an avoidant pattern will develop a personality disorder. And associations with later problems are small: a meta-analysis of 42 samples found only a small link between insecure attachment and children's internalizing symptoms (d = 0.15), with the link significant for avoidance but not for resistance or disorganization (Groh et al., 2012).
Attachment in adulthood: from the nursery to romantic relationships
Hazan and Shaver (1987) proposed that romantic love can be understood as an attachment process, which opened a large literature on adult attachment. Bartholomew and Horowitz (1991) described four adult styles, secure, preoccupied, dismissing-avoidant and fearful-avoidant, built from a person's models of self and of others. Self-report questionnaires built on this tradition are common in research.
Attachment and emotional regulation
One of the most practical contributions of attachment theory is its reading of emotion regulation. Activation of the attachment system can be understood as a response to threat or distress, and closeness to a responsive caregiver as the route by which that distress settles. When closeness does not reliably work, children and, later, adults develop other ways of handling it. This is a conceptual reading, not a measured mechanism in every case.
Mikulincer and Shaver (2003) described two strategies that people with insecure attachment use when the system is activated. Hyperactivating strategies amplify signals of need to hold the attention of a partner seen as unreliable; deactivating strategies minimize them to avoid depending on a partner seen as rejecting. Both make sense in their origin, and both can be costly in adulthood, producing relationships in which one person asks for too much and the other for too little. Someone who calls a partner fifteen times in an afternoon is not necessarily controlling; their attachment system may be on high alert. Someone who withdraws during conflict is not necessarily cold; they may have learned that expressing distress brings no response.
Emotion regulation is not learned alone. It is learned with someone, through the repeated experience that distress can be received and handed back in a tolerable form.
That idea underlies approaches such as mentalization-based treatment, emotion-focused therapy and attachment-based work with couples. What they share is a hypothesis: a therapeutic relationship can offer, in a limited and controlled way, a corrective experience in which distress is received and regulated in the presence of another person, and repeated experiences of this kind may revise the internal working model. Whether this is the mechanism that produces change is still a matter of research.
Intergenerational transmission
One of the sturdier findings of attachment research is that parents' mental states about attachment predict their children's attachment. The Adult Attachment Interview, developed by George, Kaplan and Main in the 1980s, assesses how a parent talks about childhood experiences, with a focus on the coherence of the account and not on the facts. A meta-analysis of 18 samples (N = 854) found that parents' AAI classification predicted infant security with a combined effect size of 1.06, a match of about 75% for the secure-insecure split (van IJzendoorn, 1995). The same review found a smaller link between parents' representations and their responsiveness to the infant.
Two points keep this from being misread. First, the match is substantial but not exact, and part of how it works is still unexplained (researchers call this the "transmission gap"). Second, the usual explanation is not genetic or conscious imitation but the way a parent's own model shapes how she responds to the child's distress. The AAI's "dismissing" category tends to go with avoidant infants, "preoccupied" with resistant infants and "unresolved" with disorganized infants (Main et al., 1985).
There is also evidence that parenting can be changed. A meta-analysis of 70 studies of early interventions found that those that were effective in improving parental sensitivity also improved infant attachment security, and that the most effective ones were fairly short with a clear behavioral focus (Bakermans-Kranenburg et al., 2003). Effects were modest (d = 0.33 for sensitivity and d = 0.20 for attachment insecurity).
A second reading follows for adults who recognize in themselves a pattern they also see in their parents. The useful question is not "whom should I blame?" but "what model did I inherit, and what do I want to do with it?" Intergenerational transmission is a working hypothesis, not a sentence.
How attachment is approached with adults in therapy
In the clinical literature, working with attachment in adults is not described as going back to the past to relive it, confronting parents or labeling a person with a fixed category. It is described as identifying the internal working model that guides someone's relationships and offering experiences that let that model be revised. What follows is a conceptual outline, not a protocol.
- Formulation. The history looks not only at facts but at the felt quality of availability: who responded when the child cried, who was physically present but emotionally absent, whether there were separations, illnesses or moves. The aim is not a linear story but the working hypothesis the person carries.
- Observing the pattern in the therapeutic relationship itself. The internal model shows up in the bond with the therapist. A person with an avoidant strategy may arrive on time and complete tasks while keeping emotional contact at a distance; a person with an ambivalent strategy may need constant confirmation that the therapist will not leave, or react with alarm to silences. These behaviors are treated as clinical information about the model, not as resistance to be fought.
- A corrective relational experience. The therapist aims to act as a secure base: predictable, responsive, nonpunitive and able to receive distress without handing it back amplified. Repeated over months and years, such an experience is thought to modify the internal model. It is not a magic repair but experiential learning, in which a person lives through the fact that distress can be received and regulated.
- Narrative integration. As corrective experience takes hold, a person can tell the story of early experiences with new balance, without minimizing and without being trapped. This coherence is what the Adult Attachment Interview assesses, and it is the quality linked to parents' ability to raise securely attached infants (van IJzendoorn, 1995).
In attachment-informed work the goal is not that the person understand their childhood. It is that they live, in a new relationship, an experience of availability their model did not expect.
Many contemporary approaches converge here. Mentalization-based therapy, emotion-focused therapy, transference-focused therapy and attachment-based approaches for couples share the hypothesis that deep change is not only cognitive but experiential and relational. Self-compassion often appears at this stage of treatment as an inner resource, because it allows a person to bear distress without attacking themselves, something insecure attachment tends to make harder.
Pace matters. Internal working models are not usually revised in a handful of sessions, and a promise to transform an attachment pattern within a few weeks is more likely to describe symptom work than attachment work.
Criticisms and limits
- Culture. Much of the research comes from Western, middle-class samples, and the idea of a "secure" child as one who explores freely and is soothed quickly may fit some cultural values better than others (Rothbaum et al., 2000).
- Effect sizes. Many associations between infant attachment and later outcomes are small to moderate, as in the studies above.
- Mothers. The theory is sometimes read as blaming mothers. Bowlby's account centers on the principal caregiver, who was historically the mother, but the theory describes a dynamic and does not assign guilt; later research considers fathers, grandparents and other caregivers.
- Measurement. Different instruments, interviews and questionnaires do not always agree with one another, and an attachment "type" is a simplification of a continuum.
Common mistakes that cost points on exams
- Calling attachment a by-product of feeding. Bowlby's whole point, supported by Harlow's monkeys, was that attachment is a primary system.
- Mixing up Bowlby and Ainsworth. Bowlby supplied the theory and the phases; Ainsworth supplied the Strange Situation and the A, B, C patterns.
- Forgetting that observers look at the reunion. The classification depends on what the child does when the caregiver returns, not on crying at separation.
- Listing three patterns when the question asks for four. The disorganized pattern was added by Main and Solomon.
- Treating attachment as destiny. The evidence supports moderate continuity and room for change.
- Confusing the phases with the patterns. The four phases describe how attachment develops in everyone; the patterns describe differences in its quality.
Frequently asked questions
Does secure attachment guarantee healthy adult relationships?
No. Secure attachment is a protective factor, not a guarantee. Adult relationships also depend on circumstances, the partner, life events and each person's mental health. Insecure attachment does not condemn anyone to difficult relationships either: many people with insecure attachment build stable bonds when they find partners whose availability is sustained or when they take up therapeutic work.
Can attachment patterns change in adulthood?
Yes. The same longitudinal work that shows continuity also shows change, particularly after major life events (Waters et al., 2000). Therapy, long and steady relationships and life events that rearrange priorities are commonly proposed as routes to change, though the strength of the evidence varies. The stability of the internal model is a tendency, not a fixed trait, and lasting change usually seems to take time and repeated corrective experiences rather than a single insight.
What is the difference between attachment and emotional dependence?
Attachment is a universal motivational system. "Emotional dependence" is not a formal diagnosis; it is an everyday label for a pattern in which a person's well-being and self-worth depend heavily on a relationship. Anxious attachment may be a risk factor for such a pattern, but the two are not synonyms: some people with insecure attachment are not dependent, and some people's emotional dependence is not explained by attachment alone.
Does attachment theory blame mothers?
The theory describes the importance of a caregiver's availability and does not require a moral judgment. Contemporary research looks at fathers, grandparents, nonparental caregivers and the wider context. Guilt is a moral reading a person may bring to the theory, but the theory does not require it. If anything, interventions aimed at attachment are directed at the caregiving system as a whole, not at the mother in isolation.
How is attachment measured?
In infancy, the standard method is the Strange Situation. In adulthood, research uses the Adult Attachment Interview (which codes how a person talks about childhood) and self-report questionnaires based on Bartholomew and Horowitz's four styles or on dimensions of anxiety and avoidance, such as the Relationship Questionnaire and the Experiences in Close Relationships inventory. Questionnaires are short and practical, though they have less depth than the interview.
Keep reading
- Erikson's 8 stages of psychosocial development — trust versus mistrust, the first stage, compared with Bowlby's account of early bonds.
- Freud's psychic apparatus — the psychoanalytic background Bowlby was working against and within.
- Family systems theory — how patterns pass through a family system.
References
- Ainsworth, M. D. S., & Bell, S. M. (1970). Attachment, exploration, and separation: Illustrated by the behavior of one-year-olds in a strange situation. Child Development, 41(1), 49–67. https://doi.org/10.2307/1127388
- Bakermans-Kranenburg, M. J., van IJzendoorn, M. H., & Juffer, F. (2003). Less is more: Meta-analyses of sensitivity and attachment interventions in early childhood. Psychological Bulletin, 129(2), 195–215. https://doi.org/10.1037/0033-2909.129.2.195
- Bartholomew, K., & Horowitz, L. M. (1991). Attachment styles among young adults: A test of a four-category model. Journal of Personality and Social Psychology, 61(2), 226–244. https://doi.org/10.1037/0022-3514.61.2.226
- Bowlby, J. (1982). Attachment and loss: Vol. 1. Attachment (2nd ed.). Basic Books. (Original work published 1969)
- Bretherton, I. (1992). The origins of attachment theory: John Bowlby and Mary Ainsworth. Developmental Psychology, 28(5), 759–775. https://doi.org/10.1037/0012-1649.28.5.759
- De Wolff, M. S., & van IJzendoorn, M. H. (1997). Sensitivity and attachment: A meta-analysis on parental antecedents of infant attachment. Child Development, 68(4), 571–591. https://doi.org/10.1111/j.1467-8624.1997.tb04218.x
- Fraley, R. C. (2002). Attachment stability from infancy to adulthood: Meta-analysis and dynamic modeling of developmental mechanisms. Personality and Social Psychology Review, 6(2), 123–151. https://doi.org/10.1207/S15327957PSPR0602_03
- Groh, A. M., Roisman, G. I., van IJzendoorn, M. H., Bakermans-Kranenburg, M. J., & Fearon, R. P. (2012). The significance of insecure and disorganized attachment for children's internalizing symptoms: A meta-analytic study. Child Development, 83(2), 591–610. https://doi.org/10.1111/j.1467-8624.2011.01711.x
- Harlow, H. F. (1958). The nature of love. American Psychologist, 13(12), 673–685. https://doi.org/10.1037/h0047884
- Hazan, C., & Shaver, P. (1987). Romantic love conceptualized as an attachment process. Journal of Personality and Social Psychology, 52(3), 511–524. https://doi.org/10.1037/0022-3514.52.3.511
- Main, M., Kaplan, N., & Cassidy, J. (1985). Security in infancy, childhood, and adulthood: A move to the level of representation. Monographs of the Society for Research in Child Development, 50(1–2), 66–104. https://doi.org/10.2307/3333827
- Main, M., & Solomon, J. (1990). Procedures for identifying infants as disorganized/disoriented during the Ainsworth Strange Situation. In M. T. Greenberg, D. Cicchetti, & E. M. Cummings (Eds.), Attachment in the preschool years: Theory, research, and intervention (pp. 121–160). University of Chicago Press.
- Mikulincer, M., & Shaver, P. R. (2003). The attachment behavioral system in adulthood: Activation, psychodynamics, and interpersonal processes. Advances in Experimental Social Psychology, 35, 53–152. https://doi.org/10.1016/S0065-2601(03)01002-5
- Rothbaum, F., Weisz, J., Pott, M., Miyake, K., & Morelli, G. (2000). Attachment and culture: Security in the United States and Japan. American Psychologist, 55(10), 1093–1104. https://doi.org/10.1037/0003-066X.55.10.1093
- van IJzendoorn, M. H. (1995). Adult attachment representations, parental responsiveness, and infant attachment: A meta-analysis on the predictive validity of the Adult Attachment Interview. Psychological Bulletin, 117(3), 387–403. https://doi.org/10.1037/0033-2909.117.3.387
- van IJzendoorn, M. H., Schuengel, C., & Bakermans-Kranenburg, M. J. (1999). Disorganized attachment in early childhood: Meta-analysis of precursors, concomitants, and sequelae. Development and Psychopathology, 11(2), 225–249. https://doi.org/10.1017/S0954579499002035
- Waters, E., Merrick, S., Treboux, D., Crowell, J., & Albersheim, L. (2000). Attachment security in infancy and early adulthood: A twenty-year longitudinal study. Child Development, 71(3), 684–689. https://doi.org/10.1111/1467-8624.00176