Research figure: John Bowlby · Written by Artificial Sapiens
John Bowlby understood attachment as a lifelong human capacity to form selective bonds, seek protection from trusted people, and respond emotionally to separation and loss. The need for accessible attachment figures changes its expression with age: infants depend on physical proximity and direct caregiving, while adolescents and adults can draw security from communication, anticipated availability, reciprocal support, and enduring relationships. Later researchers developed this lifespan proposition into observational studies of children's representations, interviews about adults' attachment histories, models of romantic attachment, longitudinal investigations of continuity and change, and research on relationships in old age. Bowlby's original formulations and these subsequent research traditions belong to distinct stages of scientific development (Bowlby, 1977; Main, Kaplan, and Cassidy, 1985; Fraley, 2019).
The central historical question is how Bowlby's theory traveled from the study of infant-caregiver relationships into an account of attachment across the lifespan, and which parts of that extension withstand empirical scrutiny. The answer requires distinguishing the continued operation of attachment needs from the stability of any particular attachment classification. Research supports both the persistence of attachment-related processes and meaningful opportunities for revision through later relationships, loss, adversity, and supportive experiences. No single childhood observation predicts an individual's adult relationship history with certainty (Fraley, 2002; Waters et al., 2000).
Bowlby's Original Lifespan Proposition
From an infant's protective bond to a general human disposition
Bowlby's early writing took infant attachment seriously as a behavioral system with a protective function. Its characteristic actions, including reaching, crying, following, signaling, and moving toward a familiar person, become especially visible under fatigue, threat, pain, illness, or unfamiliarity. The theory treated the child's specific relationship with a caregiver as a source of security organized around real interactions. In his 1958 paper, Bowlby argued for an account of the child's tie grounded in attachment behavior rather than explanations that reduce affectional bonds to food or drive gratification (Bowlby, 1958).
Across the following decades, Bowlby broadened the explanatory ambition of this model. The first volume of Attachment and Loss, Attachment (1969), elaborated ethological and control-system ideas; the second, Separation: Anxiety and Anger (1973), analyzed threats to availability; and the third, Loss: Sadness and Depression (1980), investigated the implications of disrupted bonds. Together, these works described attachment, separation, and loss as connected organizing themes in human relationships. The trilogy's chronological movement should not be mistaken for a claim that attachment disappears when childhood ends. Bowlby's later lectures explicitly discussed the formation and disruption of affectional bonds as recurring problems of human experience (Bowlby, 1977; Bowlby, 1988).
The theoretical continuity lies in the functions of an attachment relationship: a person monitors whether someone significant is accessible, seeks that person's help under distress, takes reassurance from expected support, and reacts to perceived unavailability. These functions need not be expressed through identical behavior at every age. Development brings new cognitive abilities, broader social networks, physical independence, sexual and romantic relationships, and changing caregiving responsibilities. The empirical task is therefore to identify functional continuity while testing developmental changes in behavior and measurement (Fraley, 2019).
Attachment, internal working models, and experience
Bowlby's concept of internal working models offered a proposed mechanism for the persistence of relational expectations. Repeated experiences with important caregivers could inform expectations about whether others would respond, whether requesting care was safe, and how the self was regarded in attachment relationships. The term was informed by the broader scientific idea of working models, including Kenneth Craik's work, and became central to the later representational development of attachment theory. A working model is a theoretical construct concerning expectations and information processing; it is not an anatomical object or a directly observable childhood memory (Bretherton, 1992; Main, Kaplan, and Cassidy, 1985).
If early expectations influence attention, interpretation, emotion regulation, and choices in new relationships, continuity becomes plausible without assuming exact repetition. A person who anticipates rejection might attend closely to ambiguous signs of withdrawal, for example, while a person expecting reliable support might find it easier to disclose distress. Such mechanisms are hypotheses to be assessed with appropriate designs. Later experience can supply disconfirming information, sometimes gradually changing how people understand themselves and others. Bowlby's lifespan framework thus allows both persistence and learning; the relative contribution of early and later experience is an empirical question (Fraley, 2002).
How Attachment Functions Change Through Development
Infancy and toddlerhood: protection, proximity, and exploration
The earliest attachment relationships involve profound asymmetry. Infants depend on caregivers for survival, transport, feeding, comfort, protection, and help regulating intense states. Separation and reunion can reveal how a child organizes the search for security, yet an isolated episode cannot describe the entirety of a relationship. The more general distinction is between the attachment bond itself and variations in the security of that bond. Mary Ainsworth's subsequent observations and Strange Situation research provided an influential empirical methodology for studying these variations; that contribution must be credited to Ainsworth and her collaborators rather than retrospectively assigned to Bowlby (Bretherton, 1992).
By toddlerhood, movement, language, anticipation, and memory alter how the child reaches the caregiver. A departure can be explained and a reunion expected; a familiar voice or promise may provide limited reassurance over a short interval. These capacities expand a child's means of coordinating proximity and exploration. The developmental shift illustrates Bowlby's premise that the attachment behavioral system is sensitive to circumstances and goals, while later developmental psychology supplies much of the evidence about age-appropriate assessment (Main, Kaplan, and Cassidy, 1985).
Middle childhood: availability becomes increasingly representational
During middle childhood, physical distance often becomes less informative than confidence in a caregiver's availability. A child at school may feel secure because someone can be contacted and relied upon when necessary. Language, autobiographical organization, planning, and knowledge of caregivers' intentions make attachment-related representations more elaborate. The security question changes from an immediate proximity calculation to an increasingly flexible appraisal of who can help, under which circumstances, and with what likely response.
This is a crucial boundary between Bowlby's original propositions and later research. Mary Main, Nancy Kaplan, and Jude Cassidy's 1985 monograph advanced the field toward assessing representations of security as children grew older. Their program connected infant observations to interviews and narrative methods, expanding the subject beyond behaviors typically seen in infant separation-reunion procedures. Representation is not equivalent to a single consciously stated belief; it can be reflected in organization, coherence, expectations, and ways of discussing relationships (Main, Kaplan, and Cassidy, 1985).
Adolescence: autonomy and continuing access to care
Adolescence changes the balance between protection and independent exploration. Young people ordinarily develop greater mobility, privacy, decision-making authority, and peer involvement while remaining dependent on adults in particular circumstances. They may turn to friends or romantic partners for some forms of support and still rely on parents or other long-established caregivers during severe stress. The significance of an attachment figure therefore cannot be equated with the frequency of daily contact.
Bowlby's account offers a way to interpret increasing autonomy as compatible with ongoing emotional security. Whether a particular adolescent's parents, friends, or partners provide secure-haven functions must be established through that person's relationships and context. The proposition does not require one exclusive attachment figure, an identical caregiving hierarchy across cultures, or uniform timing of independence. Modern lifespan investigators study attachment hierarchies and networks as research questions rather than treating a universal ranking as established fact (Fraley, 2019).
Adult relationships: mutual protection and caregiving
Adult attachment relationships often involve reciprocity. Partners may alternate between seeking help and providing it; they can also negotiate household responsibilities, illness, finances, parenting, work, and long-term plans. Attachment becomes embedded within broader patterns of affection, sexuality, shared history, caregiving, cooperation, and conflict. A partner's importance as an attachment figure is most clearly tested by whether that person functions as a source of security under threat and separation, rather than by romantic intensity alone.
This distinction explains both the strength and the limits of Bowlby's historical contribution to adult attachment research. He supplied a broad theoretical architecture for affectional bonds and their disruptions. The operational claim that romantic love could be systematically studied as an attachment process was developed later, especially by Cindy Hazan and Phillip Shaver in 1987. Bowlby did not create today's widely used adult attachment questionnaire categories or the modern two-dimensional self-report model (Hazan and Shaver, 1987; Fraley, Waller, and Brennan, 2000).
Intellectual Genealogy: From Bowlby to Later Attachment Science
Mary Ainsworth and the empirical foundation
Mary Ainsworth developed the secure-base concept through observational research and made individual differences in infant-caregiver attachment tractable through systematic study. Her Uganda and Baltimore projects and work on the Strange Situation were fundamental to the discipline's empirical development. Her research examined how young children balance exploration, proximity, and comfort seeking in relation to caregiving experience. Ainsworth's early secure, avoidant, and resistant classifications were built around observed infant behavior in a specific relational and developmental setting.
The historical relationship between Bowlby and Ainsworth is a partnership of complementary theoretical and empirical contributions, with substantial independent creativity on both sides. Ainsworth's findings helped refine the account of attachment security; Bowlby's larger framework supplied conceptual questions that made such observations consequential. Neither researcher's contribution should be collapsed into a story of a single inventor or a straight, uncontested line from infant behavior to adulthood (Bretherton, 1992).
Mary Main, Nancy Kaplan, and Jude Cassidy: representation and the Adult Attachment Interview
The 1985 work of Mary Main, Nancy Kaplan, and Jude Cassidy was a decisive transition from infant attachment behavior to the organization of attachment representations. In related work, the Adult Attachment Interview developed as a structured interview about childhood attachment experiences and the manner in which adults discuss them. Its research coding concerns the coherence and organization of discourse about attachment, including whether painful or contradictory events can be presented in an integrated account. An interview's classification is therefore not merely a direct inventory of whether childhood caregivers were kind or unkind (Main, Kaplan, and Cassidy, 1985; Roisman et al., 2007).
The interview tradition made intergenerational and longitudinal research possible in ways that Bowlby's own clinical writings had not operationalized. It also introduced a methodological complication: an adult's state of mind concerning attachment, as assessed by a coded interview, differs from a self-reported tendency to fear abandonment or avoid intimacy with a current romantic partner. Treating these instruments as synonyms creates false disagreements and exaggerated claims of continuity. Their comparison requires validation at the level of constructs, measures, populations, and intervals (Roisman et al., 2007).
Hazan and Shaver: romantic love as an attachment process
Hazan and Shaver's 1987 article proposed that adult romantic love could be understood partly through the attachment system first elaborated for child-caregiver bonds. Their studies translated familiar concepts such as security, anxiety about separation, and avoidance of closeness into questions about adult romance. The project was an extension of the theory into a different type of relationship, with adults choosing partners and often serving simultaneously as caregivers and recipients of care (Hazan and Shaver, 1987).
Their initial three-category approach was historically important, but the later field moved beyond those original categories. Adult self-report research increasingly described anxiety and avoidance as continuous dimensions. Attachment anxiety captures concerns about rejection and responsiveness; avoidance captures discomfort with closeness, dependence, or reliance on others. The familiar styles are interpretive groupings within that larger dimensional space. Modern measures assess current orientations or relationship-specific tendencies and cannot reconstruct a person's infant classification from a questionnaire score (Fraley, Waller, and Brennan, 2000).
Bartholomew, Horowitz, and the representation of self and others
Kim Bartholomew and Leonard Horowitz's 1991 four-category model further developed adult attachment research by combining different views of self and others. Their secure, preoccupied, dismissing, and fearful patterns drew from attachment ideas about expectations within close relationships. The model was historically influential in clarifying variation within patterns that earlier typologies could group together, particularly adult avoidance associated with different levels of perceived personal worth and expectations of other people (Bartholomew and Horowitz, 1991).
This elaboration should be understood as later theoretical modeling, not a taxonomy written by Bowlby. It also remains important to distinguish these adult prototypes from the classifications of infant Strange Situation behavior and from disorganized/disoriented classifications developed in separate research traditions. Similar labels across methods do not establish that the measures describe the same phenomenon in the same way.
Stability and Change: The Central Empirical Test
Three different propositions about continuity
Claims about lifespan attachment stability often mix three questions. The first concerns species-typical functional continuity: do people continue to seek security through close ties? The second concerns relational continuity: does a particular individual remain close to the same attachment figures across time? The third concerns individual-difference continuity: does a person's relative security or insecurity remain similar when measured years apart? Evidence for one does not automatically settle the others.
The first proposition is broad and supported by the recurrent importance of close relationships during distress. The second varies with development, mobility, family structure, bereavement, and partnership formation. The third requires longitudinal evidence and comparable, validated measurement strategies. Bowlby's theory provides reasons to expect some persistence of expectations, while subsequent investigators ask how much persistence actually occurs and which experiences alter it (Fraley, 2002; Fraley, 2019).
A twenty-year follow-up and what it can establish
Everett Waters and colleagues followed a group of infants first assessed using the Strange Situation at approximately one year of age. Fifty participants were recontacted about twenty years later and completed the Adult Attachment Interview. In this small and demographically restricted sample, 72% retained the same secure-versus-insecure classification, with a reported kappa of .44. The authors also reported more changes in classification among participants whose families experienced specified adverse life events. The study illustrated possible long-term continuity and possible revision; it did not demonstrate a universal stability rate for all populations (Waters et al., 2000).
Several design features govern interpretation. The follow-up sample was small, predominantly drawn from White middle-class families, and used different age-appropriate instruments at the two assessments. A proportion classified alike across two periods does not mean that identical attachment behaviors persisted unchanged for twenty years. Nor does the reported association of life events with changes establish the isolated causal effect of each event. The study is historically significant because it offers a direct longitudinal test of an important Bowlby-inspired proposition while exposing the difficulties of testing it.
Fraley's meta-analysis and models of developmental continuity
R. Chris Fraley's 2002 meta-analysis and dynamic modeling addressed the mechanisms that might generate observed correlations across developmental periods. One broad approach emphasizes relatively persistent early prototypes; another emphasizes revisable representations and ongoing influences from later environments. The empirical literature offers evidence relevant to both stability and updating. Longitudinal results depend on study intervals, methods, sample characteristics, intervening experiences, and the particular meaning of security used at different ages (Fraley, 2002).
A strong claim that early attachment has no relevance to later functioning would disregard prospective longitudinal findings. An equally strong claim that infancy fixes adult romantic attachment would disregard change, changing relationship partners, later stressors, and measurement differences. Scientific interpretation concerns probabilistic developmental pathways and boundary conditions. The questions become more informative when framed as the degree of stability, the mechanisms that mediate it, and the contexts in which trajectories change (Fraley, 2019).
What meta-analyses of childhood outcomes contribute
Meta-analytic work on early attachment has found associations with later socioemotional outcomes, including aspects of peer competence and behavior. Such findings give partial empirical support to the developmental importance of early caregiving relationships. They do not directly measure every part of Bowlby's lifespan claim, particularly the organization of romantic relationships decades later. The evidentiary distance from infancy to a specific adult outcome grows as new experiences, selection processes, social conditions, and reciprocal relationships accumulate (Groh et al., 2017).
A 2024 meta-analysis of maternal and paternal sensitivity and child attachment security found a moderate positive association rather than a deterministic one. This matters to the lifespan argument because a child's attachment environment is relational and develops across interactions with caregivers; it cannot be reduced to a single isolated parental characteristic or a mother-only causal chain. The reported population-level relationship between sensitivity and security is useful, while its magnitude, heterogeneity, and observational foundations constrain individual prediction (Madigan et al., 2024).
Attachment Measurement Across the Lifespan
Behavioral observations, representations, and questionnaires
Methods differ because the forms of attachment behavior differ with age. An infant's separation-reunion behavior with a particular caregiver, a school-age child's narrative about family relationships, an adult's discourse about childhood attachment, and a romantic partner's questionnaire responses are different observations. They can be connected through a theoretical framework, but measurement invariance and construct validity cannot be assumed merely because all instruments use the word attachment.
Ainsworth's observational classifications concern behavior in a defined caregiver-child assessment. The Adult Attachment Interview concerns discourse and organization of attachment-related recollections. Romantic self-report tools often assess anxiety and avoidance in contemporary relationships. Roisman and colleagues' 2007 investigation of the Adult Attachment Interview and self-reported attachment styles illustrates why the overlap between these traditions must be tested rather than presumed. Their methodological differences are substantive, not a minor technical inconvenience (Roisman et al., 2007).
The psychometric evidence also varies among adult self-report measures. A systematic review covering 40 studies and 24 instruments found uneven methodological quality and insufficient support for a full range of measurement properties in any one instrument. This result warrants careful instrument selection, validity appraisal, and restraint in interpreting personal test results. Adult attachment research is scientifically productive, but popular attachment quizzes do not inherit clinical diagnostic status from Bowlby's theoretical influence (Justo-Núñez, Morris, and Berry, 2022).
Relationship-specific versus general security
An adult may experience security with one partner, apprehension with another, and reliable support from a friend or family member. A general attachment orientation summarizes tendencies across experiences, whereas relationship-specific assessments describe a particular relational context. Both can have research value. Neither implies that all relationships share a uniform structure, and neither gives a license to infer another person's motives solely from one score.
This distinction complicates simplistic accounts of inheritance from childhood. A person's responses reflect memories, current relationships, trait differences, social expectations, and measurement context. Longer research designs that measure multiple partners and life changes are especially valuable in testing how interpersonal experience and earlier representations influence one another (Fraley, 2019).
Adult Attachment Figures, Caregiving Networks, and Relationship Transitions
From parental support to partners and close others
Adult attachment figures need not be identical to childhood caregivers. Romantic partners can become central sources of reassurance and emotional regulation, particularly during illness, danger, bereavement, and uncertainty. Parents, siblings, friends, and other enduring close relationships may continue to provide important support. The division of secure-haven and caregiving functions can vary across relationships; the expectation that one person must fulfill every attachment function would be a poor generalization from Bowlby's framework.
Research on attachment hierarchies asks which people someone turns to for which functions, and whether those preferences vary across situations and the life course. Such questions are distinct from asking whom a person loves most, sees most often, or regards as socially important. A bond may be emotionally significant without consistently functioning as a secure haven during threat (Fraley, 2019).
Relationship experiences can revise expectations
Intimate relationships provide repeated opportunities to confirm or revise expectations about availability and trust. Predictable responsiveness may gradually improve felt security in one relationship, while repeated unreliability or threatening experiences may heighten vigilance. A model allowing such changes is consistent with the central idea that expectations arise in relation to experience. At the same time, apparent change can reflect changes in instruments, interpretation, sample composition, or relationship-specific measurement rather than durable transformation of a general attachment orientation.
The theoretical contribution of Bowlby is the proposal that relational expectations have developmental histories and influence reactions under distress. Modern evidence asks when these expectations are stable, how strongly they predict actual interaction, whether change generalizes beyond one partner, and which environmental or interpersonal processes account for shifts. Findings should be interpreted within those measurement boundaries rather than transformed into promises that any relationship event will reliably change a person's attachment pattern.
Parenting and the coordination of care seeking and care giving
In adulthood, the same person may both seek security from others and serve as a caregiver to children, partners, or relatives. These functions interact but should be studied separately. A parent's capacity for supportive caregiving is affected by circumstances including stress, resources, health, practical support, and the needs of the child. Adult attachment representations may be associated with caregiving patterns, but statistical association is not a moral judgment or individual prediction.
This broadening of the system moves beyond an exclusive mother-infant pair. Modern research examines mothers, fathers, and other regular caregivers; the availability and quality of several relationships may help explain variation in developmental outcomes. A father's role and a child's additional caregiving bonds are substantive empirical matters, not exceptions that can be dismissed from the theory (Madigan et al., 2024).
Separation, Bereavement, and Attachment in Later Life
Separation and loss remain psychologically important after childhood
Bowlby's clinical interest in separation and loss helped make attachment a lifespan issue well before questionnaires about romantic styles became common. Strong affectional bonds expose people to distress when a significant figure is threatened, unavailable, or dies. In his 1977 Maudsley lecture, he connected the formation and breaking of such bonds with anxiety, anger, sadness, and emotional detachment. These were elements of a historical explanatory account and did not constitute a contemporary diagnostic algorithm (Bowlby, 1977).
Attachment-informed bereavement research subsequently diversified through studies of mourning, continuing bonds, grief variability, and individual differences. The historical influence of Bowlby and Colin Murray Parkes is important, but an orderly set of grief stages should not be imposed on every bereaved person. The intensity, form, and course of responses depend on circumstances, cultural practices, individual history, and available support. Grief after a death is not equivalent to a research classification of infant attachment or to a psychiatric diagnosis.
Aging, dependence, and enduring bonds
Attachment considerations remain relevant in older adulthood when partners die, health changes, mobility declines, or caregiving roles shift. Practical dependence can increase without erasing autonomy or the person's existing capacities for decision making and emotional reciprocity. Long-standing partnerships may provide continuity; adult children, friends, and other trusted people may take on larger caregiving roles. Distinguishing an attachment figure from a professional care provider also becomes important, because reliable practical care and a personally meaningful affectional bond are related but different constructs.
A systematic qualitative review of attachment in old age identified age-related differences in attachment figures, possible shifts in attachment anxiety, and associations between attachment and functioning, including caregiving under dementia-related stress. The review also emphasized serious methodological limits in this literature. Old-age findings broaden the lifespan discussion, while intervention claims require their own controlled evidence and should not be assumed from theoretical fit alone (Van Assche et al., 2013).
Scientific Challenges to a Simple Lifespan Narrative
Historical maternal deprivation does not justify maternal exclusivity
Bowlby's earlier concern about disrupted caregiving contributed to changes in child welfare and hospital practice. Some historical formulations emphasized the mother-child relationship in terms that later investigators reexamined critically. Contemporary evidence recognizes attachment relationships with multiple familiar caregivers and distinguishes the absence of stable, responsive care from a child's ordinary contact with several supportive adults. Maternal separation alone is not a valid all-purpose explanation for later psychological distress.
A lifespan framework becomes more robust when caregiving arrangements are examined directly: who is available in distress, how consistently, under what conditions, and with which opportunities for repair. A theory of protective bonds needs careful attention to social support, economic resources, family circumstances, and cultural variation; it cannot infer future outcomes from parental sex or household structure alone (Madigan et al., 2024).
Cultural context and the limits of universal claims
The protective importance of close bonds has a broad developmental rationale, yet cultures vary in caregiving practices, everyday separations, family organization, sleeping arrangements, expectations about dependence, and the interpretation of distress. Measures designed in particular social settings require evidence of validity before being treated as universal standards. A behavior indicating ordinary social adaptation in one environment may have a different meaning in another.
The universalist element of Bowlby's model is best stated at the functional level: people form selective relationships whose availability matters during threat or separation. The interpretation of specific behaviors, individual differences, and proposed mechanisms remains open to empirical testing across settings. Comparing cultures responsibly requires examining context and instrument properties as well as aggregate rates of classifications.
Prediction, confounding, and developmental complexity
Early caregiving is embedded in wider developmental systems. Genetic influences, temperament, parental mental health, material circumstances, peer experiences, neighborhood conditions, discrimination, changing family relationships, and major events may influence later functioning. If early security predicts a subsequent measure, this relationship alone cannot establish which mechanism generated it. Retrospective studies are particularly vulnerable to reconstructive memory and selection effects, while prospective studies face attrition and changes in measurement.
The strongest interpretations compare alternative causal explanations, use repeated observations, include multiple attachment figures when possible, and distinguish contemporary self-report associations from prospective developmental pathways. Fraley's account of adult attachment debates is useful precisely because it treats explanatory gaps as substantive scientific problems (Fraley, 2019).
Classification is a research tool, not a clinical verdict
Attachment security and insecurity are developmental or relationship research constructs. An adult's attachment anxiety and avoidance are questionnaire dimensions. Neither is itself a DSM or ICD psychiatric diagnosis. Infant disorganization does not prove maltreatment, and adult avoidance cannot be used to diagnose a personality disorder. Distress about separation varies normally with age and circumstances; clinical separation anxiety disorder requires a separate diagnostic evaluation of symptoms, duration, impairment, and differential explanations.
These distinctions have practical consequences. Clinicians can consider relationship history and support systems as part of a formulation while relying on appropriate clinical standards for diagnosis and treatment. An historical account of attachment cannot validate an untested treatment, certify a social media attachment label, or infer personal trauma from a questionnaire. Clinical recommendations require evidence matched to the population, condition, and intervention in question.
What the Lifespan Literature Establishes
The evidence supports a productive but qualified synthesis. Bowlby gave attachment science a framework that extends beyond infant dependency to the continuing psychological significance of particular close relationships. Ainsworth supplied indispensable observational findings about the organization of early security. Main, Kaplan, and Cassidy advanced the study of attachment representations. Hazan and Shaver established a research program applying attachment questions to adult romantic love; later scholars refined concepts and measurement, especially the adult anxiety and avoidance dimensions (Bretherton, 1992; Hazan and Shaver, 1987; Fraley, Waller, and Brennan, 2000).
Evidence for associations among caregiving, early security, subsequent social functioning, and adult relationship orientation is established at the group level, with effects varying by study design and construct. Evidence for universal or irreversible continuity from one infant classification to a specific adult style is insufficient. Research on old-age attachment, relationship-specific change, and some interventions remains more limited and heterogeneous. The most defensible lifespan account therefore follows the changing functions of close bonds and measures individual differences probabilistically, while allowing experience to modify developmental pathways (Fraley, 2002; Groh et al., 2017; Fraley, 2019).
This interpretation also respects canonical subject boundaries. The historical contribution and later reception of Bowlby's lifespan proposition belong here. For a full account of contemporary adult attachment dimensions and relationship processes, see Adult Attachment Theory: How Anxiety and Avoidance Shape Relationships. For the developmental science of young children's caregiving relationships and observational classifications, see Attachment in Infancy and Early Childhood: Security and Development. The broad scientific questions have their own established editorial homes.
Frequently Asked Questions
Did John Bowlby believe attachment lasts throughout life?
Yes. Bowlby's theory treated the formation and loss of specific affectional bonds as enduring human concerns. Later writings and lectures articulated implications beyond infancy, while researchers working after him created many of the instruments now used to examine adult attachment (Bowlby, 1977).
Did Bowlby invent adult attachment styles?
Bowlby developed the theoretical system and proposed that relational expectations could influence later life. Hazan and Shaver's 1987 romantic attachment research, Bartholomew and Horowitz's 1991 models, and later dimensional measurement work belong to subsequent scholarly contributions (Hazan and Shaver, 1987; Bartholomew and Horowitz, 1991).
Is attachment fixed by infancy?
No. Some longitudinal investigations find meaningful continuity, but patterns also change, and effect estimates depend on the sample, period, events, and instruments. The 2000 twenty-year follow-up study combined stability with observed changes associated with adverse experiences; later meta-analytic modeling similarly treats persistence and revision as central questions (Waters et al., 2000; Fraley, 2002).
Is the Adult Attachment Interview the same as an adult attachment questionnaire?
No. The Adult Attachment Interview codes features of discourse and state of mind concerning attachment histories. Adult romantic attachment questionnaires often measure current anxiety and avoidance. They are related research traditions with different operational definitions and only partly overlapping evidence (Roisman et al., 2007).
Can adults have more than one attachment figure?
Yes. Attachment-related functions can be distributed across partners, parents, friends, and other significant relationships. Which relationships provide secure haven or emotional support varies with circumstances and developmental period. The existence of multiple close attachment figures is compatible with research on differentiated relationship networks (Fraley, 2019).
Does an insecure attachment classification mean a person has a mental disorder?
No. Research classifications and self-reported dimensions describe aspects of relationships or representations. Psychiatric diagnosis requires its own clinical criteria and assessment. Attachment constructs can inform research or clinical formulation without determining a diagnosis.
Related Articles
John Bowlby — Scientific overview of Bowlby's life, theoretical development, primary works, collaborators, and historical legacy.
Adult Attachment Theory: How Anxiety and Avoidance Shape Relationships — Contemporary adult-relationship theory, dimensions, measurement, and interpretation.
Attachment in Infancy and Early Childhood: Security and Development — Developmental observations, caregiver relationships, classification, and early-childhood evidence.
John Bowlby's Attachment Theory: Origins, Principles, and Legacy — The original attachment theory and its conceptual development.
John Bowlby's Internal Working Models of Attachment — The representational account of self and caregivers.
References
Bartholomew, K., and 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. (1958). The nature of the child's tie to his mother. International Journal of Psycho-Analysis, 39, 350–373. https://pubmed.ncbi.nlm.nih.gov/13610508/
Bowlby, J. (1969). Attachment and Loss, Vol. 1: Attachment. Basic Books. Wellcome Collection, Bowlby archival records
Bowlby, J. (1973). Attachment and Loss, Vol. 2: Separation: Anxiety and Anger. Basic Books. Wellcome Collection, Bowlby archival records
Bowlby, J. (1977). The making and breaking of affectional bonds. I. Aetiology and psychopathology in the light of attachment theory. British Journal of Psychiatry, 130(3), 201–210. https://doi.org/10.1192/bjp.130.3.201
Bowlby, J. (1980). Attachment and Loss, Vol. 3: Loss: Sadness and Depression. Basic Books. Wellcome Collection, Bowlby archival records
Bowlby, J. (1988). A Secure Base: Parent-Child Attachment and Healthy Human Development. Basic Books. WorldCat catalog record
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
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
Fraley, R. C. (2019). Attachment in adulthood: Recent developments, emerging debates, and future directions. Annual Review of Psychology, 70, 401–422. https://doi.org/10.1146/annurev-psych-010418-102813
Fraley, R. C., Waller, N. G., and Brennan, K. A. (2000). An item response theory analysis of self-report measures of adult attachment. Journal of Personality and Social Psychology, 78(2), 350–365. https://pubmed.ncbi.nlm.nih.gov/10707340/
Groh, A. M., Fearon, R. M. P., van IJzendoorn, M. H., Bakermans-Kranenburg, M. J., and Roisman, G. I. (2017). Attachment in the early life course: Meta-analytic evidence for its role in socioemotional development. Child Development Perspectives, 11(1), 70–76. https://doi.org/10.1111/cdep.12213
Hazan, C., and Shaver, P. (1987). Romantic love conceptualized as an attachment process. Journal of Personality and Social Psychology, 52(3), 511–524. https://pubmed.ncbi.nlm.nih.gov/3572722/
Justo-Núñez, M., Morris, L., and Berry, K. (2022). Self-report measures of secure attachment in adulthood: A systematic review. Clinical Psychology and Psychotherapy, 29(6), 1812–1842. https://pubmed.ncbi.nlm.nih.gov/35623630/
Madigan, S., Deneault, A.-A., Duschinsky, R., Bakermans-Kranenburg, M. J., Schuengel, C., van IJzendoorn, M. H., and colleagues. (2024). Maternal and paternal sensitivity: Key determinants of child attachment security examined through meta-analysis. Psychological Bulletin, 150(7), 839–872. https://pubmed.ncbi.nlm.nih.gov/38709619/
Main, M., Kaplan, N., and 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
Roisman, G. I., Holland, A., Fortuna, K., Fraley, R. C., Clausell, E., and Clarke, A. (2007). The Adult Attachment Interview and self-reports of attachment style: An empirical rapprochement. Journal of Personality and Social Psychology, 92(4), 678–697. https://pubmed.ncbi.nlm.nih.gov/17469952/
Van Assche, L., Luyten, P., Bruffaerts, R., Persoons, P., van de Ven, L., and Vandenbulcke, M. (2013). Attachment in old age: Theoretical assumptions, empirical findings and implications for clinical practice. Clinical Psychology Review, 33(1), 67–81. https://doi.org/10.1016/j.cpr.2012.10.003
Waters, E., Merrick, S., Treboux, D., Crowell, J., and 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