← all systems · see it in The System
Attachment (the infant-caregiver bond)
Attachment is the affectional bond a child forms with those who care for it, as developmental psychology, ethology and neuroscience describe it: the child's disposition, when alarmed, ill or tired, to seek protection and comfort from a stronger and wiser familiar person (a safe haven) and to explore from that person as a secure base, met on the other side by the caregiver's caregiving bond. It belongs to a relationship, not to the child alone; it forms over the first years (selective bonds from about six or seven months), is observed in separation-and-reunion procedures (the Strange Situation) and represented later in mental models and adult interviews (the Adult Attachment Interview), and has counterparts in other mammals and in birds' filial imprinting. How it should be classified, whether its forms are universal across cultures, and how far early bonds predict later life are standing debates.
Properties published scienceour proposal
Sourced fields, and our own readings.
| aliases | Attachment |
| asymmetry | asymmetric: the child seeks a 'stronger and wiser other'; the caregiver's bond is the reciprocal caregiving bond |
| born at | after birth in each child; selective bond from six or seven months |
| born precision | months |
| boundary | Inside: the bond between a young child and the people who care for it, from both sides: the child's attachment behaviours (crying, smiling, clinging, following, protest at separation), the attachment-exploration balance (safe haven and secure base), the caregiver's caregiving bond (sensitivity, comfort, protection, limit setting), the patterns a relationship settles into (secure, avoidant, resistant, disorganized), the stress physiology the caregiver buffers, the mental models and scripts that carry the bond forward, the network of attachment figures, the procedures that observe it (Strange Situation, Attachment Q-Set, Attachment Script Assessment, Adult Attachment Interview), and the bond's counterparts in other animals (rat pups, rhesus infants, dogs, cats, ewes and lambs; chicks' imprinting). Outside: the child as a whole (child development, human), the organ (human brain), emotion and memory in general, the family as a group, adult romantic bonds (adult attachment, kept on the shelf), the clinical attachment disorders (reactive attachment disorder, disinhibited social engagement disorder; shelf), and the theory as an idea held by science (attachment theory, its own walk). |
| cheater | unseen as such; the nearest is frightening or rejecting caregiving, which the child meets with avoidance or disorganization |
| drive | driven |
| enforcement | partner-choice |
| flips when | a caregiver becomes a source of alarm (unresolved trauma, maltreatment): the haven and the threat are the same person |
| joint output | felt security with exploration: a child that turns to the figure when alarmed and explores when calm, and the mental models it carries forward |
| knockout | Signs of the inhibited type of reactive attachment disorder responded quickly to placement in foster care (src:smyke-2012): without a consistent caregiver (institutional rearing), attachment-disorder signs appear |
| nested level | A part of child development (its social and emotional line) built between a child and its caregivers inside a family or wider caregiving network; it runs on the senses, brains and stress axes of both, and what it carries forward (mental models, the caregiving of the next generation) feeds adolescence, adult relationships and the family that follows. |
| obligate | obligate |
| partners | sys:attachment-behavioural-system, sys:caregiving-behavioural-system |
| recurrence | repeated daily over years; expectations built through 'more or less consistent responsive interactions' (src:opie-2026) |
| regime | starts after birth (selective from six or seven months); changes form after about three (the system formed); continues as adult bonds; ends with separation or death (grief) |
| saturation | security is about quality, not strength: no 'over-attached' child (src:opie-2026) |
| scale | social |
| separability | two people, each with an identity and other relationships; the child has relationship-specific patterns with each figure |
Measures (44) published science
| caregiver sensitivity and attachment security | .25 (95% CI .22-.28; k = 174, N = 22,914) r (pooled) |
| maternal vs paternal sensitivity and security | maternal .26 (N = 21,483); paternal .21 (N = 1,626) |
| caregiver sensitivity and attachment security (umbrella) | .25 (k = 253, n = 37,444) r |
| intergenerational transmission (2016 synthesis) | secure-autonomous .31; unresolved .21 (95 samples, N = 4,819) |
| intergenerational transmission, 1995 meta-analysis | .47 (d = 1.06; 18 studies, N = 854) r |
| stability of attachment across early childhood (four-way) | 0.23 kappa |
| stability across early childhood (secure/insecure) | 0.28 r |
| staying in the same pattern across early childhood | secure 65.08; avoidant 35.67; resistant 26.45; disorganised 37.59 % of dyads |
| stability infancy to early adulthood | .39 (n = 21,072 attachments, 127 papers); none significant past 15 years |
| preschool attachment distribution (secure) | 53.5 secure, 14.0 avoidant, 11.0 ambivalent, 21.5 disorganized/controlling (97 samples, N = 8,186) % |
| North German SSP distribution (avoidant) | 52 avoidant, 34 secure, 13 resistant (46 pairs) % |
| disorganized attachment, share of dyads | less than 20 % of young child-caregiver dyads |
| insecure classifications, share of children | roughly half of children |
| shared environment in infant security (twins) | 52 (genetic part negligible; 157 twins) % of variance |
| heritability of attachment, adolescence | about 40 (551 twin pairs aged 15) % |
| heritability of secure base script knowledge, middle childhood | 38 (415 twin pairs aged 8-11) % |
| heritability of adult attachment styles | about 36 overall; about 51 parent-specific (678 pairs, mean age 70) % |
| early attachment and peer competence | peer competence .19; externalizing -.15; internalizing -.07 |
| infant security and later socioemotional outcomes, partial r | <= .10 to .15 (MLSRA and SECCYD, four demographic covariates) partial r |
| variance in social competence at 19 explained by infant security (Minnesota) | 5 (about half with social circumstances included) % of variance |
| two secure relationships and language | .26 d |
| negative emotionality and number of secure relationships | -0.12 d |
| Strange Situation duration | 21 (seven episodes of about three minutes, two separations and reunions) |
| phase three of attachment formation | from six or seven months to two years or more age |
| goal-corrected partnership stabilises | around 30 months of age |
| social buffering of cortisol becomes reliable | end of the first year (separation raised cortisol at 13 months, not at 9) age |
| rat pup stress-hyporesponsive period ends | 10-11 postnatal days |
| rat pups open eyes and ears | third week of life age |
| chick imprinting sensitive period, upper limit | approximately 10 (not extendable by thyroid hormone) days after hatching |
| chick imprinting exposure for a lasting preference | 15 min of exposure gives a preference lasting at least two days minutes |
| Harlow surrogate preference, days of testing | 165 (8 infants: 4 with a lactating cloth mother, 4 with a lactating wire mother) days |
| oxytocin and ventral striatum response to own infant | 0.57 (P = 0.002) Spearman rS |
| male prairie voles without partner preference | approximately 30 % |
| zebra finch corticosterone recovery on reunion | within 48 hours |
| VIPP-SD trials | 25 randomized controlled trials |
| ABC coaching rate | at least one comment per minute; 10 sessions comments per minute |
| UK referrals to children's social care before age five (born 2009-10) | 1 in 5 referred; 1 in 19 forensic assessment children |
| global infant SSP distribution (secure) | 51.6 secure, 14.7 avoidant, 10.2 resistant, 23.5 disorganized (285 studies, 20,720 dyads) % |
| within-country vs between-country variation in SSP distributions | within-country variation nearly 1.5 times between-country (32 samples, N = 1,990, 8 countries) ratio of chi-square variation |
| AAI-infant correspondence (1995 meta-analysis) | 75% (kappa = .49, n = 661); effect size 1.06 (18 samples, N = 854) % agreement |
| maltreatment and disorganization (meta-analysis) | d = 2.19 vs other high-risk children; five socioeconomic risks d = 1.20, not significantly different |
| heritability of family relationships in infancy (twin meta-analysis) | 0.41 (95% CI 0.06-0.71) pooled h2 |
| randomized interventions' effects | sensitivity d = 0.33 (n = 7,636); attachment d = 0.20 (n = 1,503) |
| maternal-presence switch age in rat pups | 12-15 postnatal days |
Open contentions (14) open contention
Where sources or readings disagree. Both stay on the map.
- C-413 Where do thoughts come from?
booked: No theory is proven, so the map lists the main ones side by side and doesn't pick one. It shows where thinking is handled in the brain, not where thoughts start.
the other reading: as in C-413 - C-1102 Do genes make people vulnerable to stress, or just more sensitive to their surroundings?
booked: The map keeps both models: vulnerability triggered by stress, and sensitivity that works for better and for worse.
the other reading: as in C-1102 - C-3118 Is infant attachment the same across cultures?
booked: The map keeps the competing readings; none is ruled.
the other reading: secure as the universal norm and insecure as pathology (a misconception per Opie 2026); attachment theory as a Western middle-class view whose universality claim carries moral judgements (Keller 2018); attachment as a main driver of later outcomes; the published Keller-Mesman exchange leaves one disagreement: how flexibly sensitivity can be read across cultures (Mesman 2018 reply) - C-3578 How many kinds of infant attachment are there, and are they kinds at all or points on a scale?
booked: The map keeps the competing readings; none is ruled.
the other reading: Crittenden's dynamic-maturational model (A/C and other organized classes); the resistant class 'the most poorly understood' and its relation to D unclear; self-report adult styles not the same constructs as infant D (Granqvist 2017) - C-3579 How far does a baby's attachment pattern carry into adult life?
booked: The map keeps the competing readings; none is ruled.
the other reading: security as 'the normative homeostatic state' with insecurity consolidated in brain circuitry (Opie 2021's reading); developmental vs life-course views of early disruption (crime literature, encyclopedia) - C-3580 What did attachment evolve for: protection from predators, or something wider?
booked: The map keeps the competing readings; none is ruled.
the other reading: Bowlby: an innate infant attachment system as the key to infant survival; critics: the system is not working until about three, so caregivers ensured survival; cooperative breeding first (Burkart 2025) - C-3581 Why does attachment look inherited in teenagers and adults but not in babies?
booked: The map keeps the competing readings; none is ruled.
the other reading: gene-environment correlation (older children shape their experiences) vs measurement: later measures lean on cognitive and verbal ability (Runze 2025 tested IQ for this reason); the four estimates measure different constructs (observed SSP, script knowledge, interview coherence, self-report), and questionnaire and interview measures are 'essentially orthogonal' (Fearon 2014); an infant twin meta-analysis puts 'family relationships (attachment and dependency)' at h2 0.41 (Austerberry 2022) - C-3582 Is attachment theory a testable scientific theory or a movement?
booked: The map keeps the competing readings; none is ruled.
the other reading: a scientific theory refutable by its predictions (Bowlby) / a psychosocial movement inside a 'bubble' (critics quoted in the encyclopedia; Ainsworth's own reported view); the 1970s critique of 'index' behaviours (as Sroufe and Waters 1977 state it) and their organizational-construct answer - C-3614 Is the dog's bond with its owner unique to dogs, or do hand-raised wolves form it too?
booked: The map keeps the competing readings; none is ruled.
the other reading: hand-raised wolves show comparable attachment behaviour (Hansen Wheat et al. 2022, as Gácsi et al. report it) / dogs unique (Gácsi, Miklósi and Topál 2023) - C-3635 Is a duckling's imprinting the same process as a baby's attachment?
booked: The map keeps the competing readings; none is ruled.
the other reading: one early-bonding process across species (Lemche 2020's reintegration; a 'unique attachment learning circuit has been posited across a variety of species', Opendak 2019) / different processes (Bowlby dropped the analogy) - C-3639 What is an 'internal working model' of attachment?
booked: The map keeps the competing readings; none is ruled.
the other reading: internal working models 'still lack any agreed definition' (encyclopedia); critics: invisible structures inferred by trained coders after the observational claims failed - C-3649 Does a baby attach to one main person or to a network of carers?
booked: The map keeps the competing readings; none is ruled.
the other reading: a 'most familiar' caregiver of particular importance as a safe haven in the earliest years, with consequences for custody time (debated among the Forslund 2022 authors themselves); Bowlby's original monotropy - C-3665 What is a child seeking from a caregiver: closeness, or the feeling of security?
booked: The map keeps the competing readings; none is ruled.
the other reading: proximity (physical distance) / felt security (an appraisal of availability); the specialist adds Sroufe and Waters 1977 and Bowlby 1973 as the turn (not read for this point) - C-3667 What does the Strange Situation test measure?
booked: The map keeps the competing readings; none is ruled.
the other reading: SSP classifications mostly reflect the dyad's recent social circumstances rather than an internal model (Lamb and colleagues, as the encyclopedia reports); Ainsworth's own study found separation protest and stranger fear far less common than expected
Cross-scale patterns seen here (17) our proposal
Shapes that recur when the same kind of problem is solved at different scales, and how each shows in Attachment (the infant-caregiver bond). Each links to every other system that shows it.
- same role, different builds (structural)
The job, a young animal learning whom to stay close to, is done by olfactory learning in rat pups, by the IMM in the chick forebrain, and by networks across both hemispheres in humans; cats and dogs meet human attachment criteria. - same shape, opposite sign (structural)
One process, attention to attachment-related information, run with its sign reversed: avoidant infants minimize it, resistant infants maximize it (Main, as Granqvist 2017 states both). And in rat pups one odour-shock pairing yields a preference in the sensitive window and avoidance after it: the sam - same toolkit, independent builds (structural)
Pair bonding 'is rare in mammals and has evolved multiple times independently', each time on common pathways with the maternal bond (oxytocin, dopamine, nucleus accumbens and ventral pallidum). - symmetry breaking (name the cut) (structural)
Before: smiles and cries go indiscriminately to any caregiver present (six-month-olds still do). The cut: discrimination of familiar from unfamiliar at three to six months. After: a preferred figure and a hierarchy with a principal figure at the top, which sets whom the child turns to. - trade one property for another (structural)
The attachment-exploration dial: avoidant infants shift attention to exploration 'at the expense of communication of their feelings', resistant infants cling 'at the expense of play', secure infants balance the two. Proximity is bought with exploration and the reverse, and the patterns sit at differ - Umwelt: each system carves its own slice of one shared medium (structural)
One caregiving world, sliced per species: rat pups meet the mother by odour alone until the third week, human newborns through all senses, chicks by sight of a moving object, Harlow's monkeys by contact more than milk. - who opens the door (structural)
Doors built ahead by the receiver: the pup's attachment circuit (olfactory bulb, piriform, locus coeruleus) with low corticosterone holding the amygdala shut; the newborn's face predisposition (Conspec) before any face is learned; the secure mother's reward response to her own infant's face; the set - borrowed-then-held (partial)
Regulation first supplied by the caregiver (the mother's presence blocks a pup's corticosterone response; a figure buffers a toddler's cortisol), then held by the child: secure attachment 'promotes later independence and self-regulation' and after about three the bond survives separation. The held p - coordination-before-mapping (partial)
Before any model exists, the newborn's cries and smiles coordinate care: they 'function as if they are signals' by triggering the caregiving of anyone nearby; the working model comes later. - lightning (partial)
Imprinting: 15 minutes of exposure in a sensitive window a few days after hatching sets a preference lasting at least two days; a brief firing that reshapes what follows. In the human bond the forming is slow (months), so this fires on the shelf partner more than on attachment itself. - nested ladder (partial)
Bowlby's four phases each build on the last (signals to anyone; discrimination; a clear-cut bond organized toward a goal; a goal-corrected partnership), so each phase implies the ones below; the ordering is the theory's, with the timing data contested. - outside-to-inside (partial)
A regulator outside the child (the figure's presence and responses) becomes an expectation inside it (scripts, working models); the outside phase comes first in the sources (buffering by the end of the first year; models later). - templating (partial)
The maternal bond is 'proposed to be the primordial system' whose neural scaffold the pair bond reuses, and adult romantic attachment was modelled on the infant categories (four adult styles 'roughly correspond' to infant ones). The module is reused, but with species-specific differences. - The Payment to Entropy (partial)
Signal form: each hand-off in the chain parent's childhood -> state of mind -> sensitivity -> child's pattern -> later outcomes passes on less. Transmission r = .31 (and fell from .47 as larger samples came in); sensitivity carries only part of it (the gap); infant security predicts adult outcomes a - the whole sets the local reference (partial)
What a pattern means is read against a culture's caregiving: a North German sample is 52% avoidant, a Japanese sample has no avoidant infants, and Keller holds that sensitivity is defined from one culture's concept of the person; Opie 2026 reads insecure strategies as adaptive in their own context. - holding / entrainment (surface)
Mother-infant behavioural and neural synchrony: early sensitive care is associated with mother-adolescent interbrain synchrony a decade later. Two partners in step, not a whole setting its parts' clock. - namespace completeness (the map predicts its gaps) (surface)
Ainsworth's three patterns left cases that fit nowhere; Main and Solomon added the disorganized class from records of behaviours 'discrepant' with A, B and C, and Crittenden offered A/C. The gap was found from leftovers, not predicted with properties beforehand.
Valence: what it trades, and with whom (20) published science
Each exchange with a neighbour, read from both sides: who gains, who pays, and what carries it.
- Peer group via sys:sensory-environment · Attachment (the infant-caregiver bond) gain
- Adolescence via sys:sensory-environment · Attachment (the infant-caregiver bond) mixed
- Amygdala via sys:sensory-environment · Attachment (the infant-caregiver bond) mixed
- Oxytocin signalling via sys:sensory-environment · Attachment (the infant-caregiver bond) gain
- Human (Homo sapiens) via sys:sensory-environment · Attachment (the infant-caregiver bond) gain
- Domestic dog (Canis familiaris) via sys:sensory-environment · Attachment (the infant-caregiver bond) gain
- Human brain via sys:sensory-environment · Attachment (the infant-caregiver bond) gain
- Family (the human family as a system) via sys:sensory-environment · Attachment (the infant-caregiver bond) gain
- Emotion (feelings) via sys:sensory-environment · Attachment (the infant-caregiver bond) gain
- Rhesus macaque via sys:sensory-environment · Attachment (the infant-caregiver bond) gain
- Memory via sys:sensory-environment · Attachment (the infant-caregiver bond) gain
- HPA axis via sys:sensory-environment · Attachment (the infant-caregiver bond) gain
- Personality via sys:sensory-environment · Attachment (the infant-caregiver bond) mixed
- Culture via sys:sensory-environment · Attachment (the infant-caregiver bond) mixed
- Japanese macaque (Macaca fuscata) via sys:sensory-environment · Attachment (the infant-caregiver bond) gain
- Domestic cat via sys:sensory-environment · Attachment (the infant-caregiver bond) gain
- Imprinting via sys:sensory-environment · Attachment (the infant-caregiver bond) gain
- Stress (biology) via sys:sensory-environment · Attachment (the infant-caregiver bond) gain
- Child development via sys:sensory-environment · Attachment (the infant-caregiver bond) gain
- Prairie vole via sys:sensory-environment · Attachment (the infant-caregiver bond) gain
Partners and parts published science
Built from / builds (7)
- Social buffering
- Attachment behavioural system (the child's side)
- Caregiving behavioural system (the caregiver's side)
- Internal working model (mental models, secure base scripts)
- Attachment network (the child's several attachment figures)
- Caregiver sensitivity
- Child development
Sits inside / holds (3)
expressed by (14)
- Scaffolding/Enablement (Meta-Tooling)
- Symbiosis/Co-Creation (Collaborative Specialization)
- Sensing/Scanning
- Mapping/Modeling
- Uncertainty Reduction/Learning
- Risk-managed Exploration
- Social Learning / Public Information Use
- Quick Harm Avoidance
- Reactive Response
- Proactive Defense
- Validation of Identity/Intention
- Resilience & Repair
- Redundancy & Diversity
- Quality Control & Error Correction
floor filled by (19)
- Human (Homo sapiens)
- Human brain
- Family (the human family as a system)
- HPA axis
- Personality
- Secure attachment
- Disorganized attachment
- Imprinting
- Social buffering
- Insecure-avoidant attachment
- Insecure-ambivalent/resistant attachment
- Attachment behavioural system (the child's side)
- Internal working model (mental models, secure base scripts)
- Attachment network (the child's several attachment figures)
- Caregiver sensitivity
- Adult attachment (romantic and close relationships)
- Adult attachment interview
- Strange Situation Procedure
- Child development
represents (1)
Kinds (9)
- Secure attachment
- Anxious-avoidant attachment
- Anxious-resistant attachment
- Disorganized attachment
- Insecure-avoidant attachment
- Insecure-ambivalent/resistant attachment
- Dynamic-maturational model of attachment (Crittenden)
- Dog-owner attachment
- Cat-human attachment
wired (1)
Notes published science
- Baby monkeys chose comfort over milk: in Harlow's 1958 experiment, infants fed by a bare wire 'mother' still spent their time clinging to a soft cloth one that gave no milk, and kept doing so through 165 days of testing; the wire mother was, in his words, 'biologically adequate but psychologically inept' (src:harlow-1958).
- A rat pup cannot learn to fear its mother: in the first days of life, pain paired with her smell makes the pup like the smell more, because its low stress hormone keeps the fear centre (amygdala) switched out; evolution built the infant brain to attach even to a rough caregiver, and the cost shows up only later, as emotional and social problems from weaning on (src:opendak-2019).
- What 'attachment' means changes with age, and so does how much genes seem to matter: in infant twins observed in the Strange Situation, genes count for close to nothing (shared family environment explains about half); measured as story-telling about a secure base at 8-11, or as coherence in an interview at 15, it is about 38-40% heritable. But these are different measures of different things (behaviour, stories, interviews), so the 'rise' may partly be a change of ruler rather than of the bond, and an infant twin meta-analysis that pools attachment with dependency already finds 41% (src:bokhorst-2003, src:runze-2025, src:fearon-2014, src:austerberry-2022).
- Culture matters less than it looks: when almost 2,000 Strange Situations from 8 countries were pooled, the differences between samples inside one country were nearly 1.5 times the differences between countries, and the famous 52%-avoidant German sample was the only one of 32 where secure babies were not the largest group (src:vanijzendoorn-1988).
Layers
The ways it takes in its world, where they meet, and how it acts back, each from L1 (where it meets the world) up.
attachment: from the caregiving world to the bonds a person carries (reader's view)
- L1: The caregiving world: one or several attachment figures (parents, grandparents, carers), their presence, contact and responses, and the dangers and strangeness around
- L2: Signals both ways: the infant's senses (all of them in human newborns; smell in rat pups) and the attachment behaviours (rooting, sucking, smiling, crying, reaching, clinging, looking)
- L3: A selective bond: familiar told from unfamiliar from three to six months; a preferred figure, then clear-cut attachment from six or seven months (rat pups: an odour preference; chicks: an imprinted object)
- L4: The attachment-exploration balance: safe haven and secure base, proximity as a set-goal, the stress axis buffered by the figure
- L5: The relationship's pattern: secure, avoidant, resistant, disorganized; relationship-specific
- L6: Mental models: expectations, secure base scripts, internal working models, and the adult state of mind about attachment
- L7: Bonds across the lifespan and into the next generation: social competence and close relationships (probabilistic), and the caregiving given to one's own children
the boardroom (shared floors)
- L4: The attachment-exploration balance: proximity sought when alarmed (safe haven), exploration when calm (secure base); a set-goal read as proximity (Bowlby 1969) or as felt security, the figure's availability (later reading, contested); the caregiver's presence buffering the stress axis (in human infants reliable from the end of the first year; shown separately in rat pups from day 10-12 and in squirrel monkeys)
- L5: The pattern of each relationship: secure, avoidant (attention minimized), resistant (attention maximized), or disorganized (the strategy broken by fear or conflict); relationship-specific, moderately stable, changeable with experience
- L6: Mental models: expectations, secure base scripts, internal working models (no agreed definition), and in adults a state of mind about attachment; held in networks across both hemispheres
- L7: Bonds carried across the lifespan and into the next generation: friendships, partners and the caregiving of one's own children; probabilistic, with weak links to infancy over long spans
way in: Alarm: danger, strangers, the unfamiliar
- L1: A stranger approaching, an injury, an unfamiliar room
- L3: Alarm: the attachment system activated by fear of danger; by 12 months the mother's face is read for whether a thing is safe
way in: Contact comfort: being held, clinging, warmth
- L1: Body contact and its comfort: in Harlow's monkeys it outweighed milk
- L3: A preferred surrogate learned: responsiveness shifts to the cloth mother over the first months
way in: Emotion: fear, anxiety, anger, sadness (inner)
- L1: Anxiety: the fear of being cut off from the attachment figure
- L3: Separation feelings in sequence: anxiety and anger, then sadness and despair
way in: Faces: an inborn pull toward face-like patterns
- L1: Faces and face-like patterns in view
- L2: A newborn predisposition to attend to human facial patterns (found after the chick work)
- L3: Learning about a particular face once attention is caught (Conlern)
way in: Inherited endowment: genes and temperament
- L1: Genes: negligible for infant Strange Situation security in twin studies (no candidate-gene finding has replicated); later measures of a different kind (script knowledge, interview coherence, self-report styles) show 36-40% heritability, and an infant meta-analysis pooling attachment and dependency puts family relationships at 0.41
- L2: Temperament: negative emotionality shifts the count of secure and resistant relationships only slightly
- L3: Two-way interplay of the child's heritable traits and the caregiving environment
way in: Other attachment figures: fathers, grandparents, siblings, carers
- L1: Several caregivers: the norm in many settings; foragers mix maternal, allomaternal and peer care
- L3: Arranged in a hierarchy with a principal figure at the top; each relationship its own pattern
way in: The caregiver's odour (rodent pups)
- L1: The caregiver's odour: how pups find the nipple, warmth and care
- L2: Olfaction is the pups' main sense; they cannot see or hear until the third week
- L3: Attachment learning: olfactory bulb, anterior piriform cortex and a hyper-functioning locus coeruleus make an odour preference; low corticosterone keeps the amygdala out
way in: The caregiver's own attachment history (inner, caregiving side)
- L1: The caregiver's own childhood attachment experiences, as now represented
- L3: A state of mind: coherent (secure-autonomous), dismissing, preoccupied, or unresolved about loss or trauma
way in: The caregiver's presence: where the attachment figure is, leaving and coming back
- L1: The caregiver's presence, departure and return in the space the child shares with them
- L2: Seen and heard: human newborns meet the caregiver through all their senses
- L3: Familiar told from unfamiliar (from three to six months); a preferred figure singled out
way in: The caregiver's response to the child's signals (sensitivity)
- L1: Sensitive care: prompt, appropriate, contingent responses, consistent and accessible
- L3: Expectations of safety and trust built from more or less consistent responsive interactions
way in: The child's own state: illness, fatigue, pain, the stress axis
- L1: Illness, fatigue and fear inside the child
- L3: The stress axis answers separation: no cortisol rise at 9 months, a rise at 13 months
way in: The family's social background: stability, support, risk
- L1: Socio-economic, marital, housing, support and childcare circumstances
- L3: Risk: at-risk secure children less likely to stay secure, at-risk insecure children most likely to stay insecure
way in: The infant's cues reaching the caregiver (the caregiving side)
- L1: The infant's faces and cues, happy or sad
- L3: Reward: secure mothers' ventral striatum, orbitofrontal and medial prefrontal cortex respond to their own infant's face
way in: Thought: expectations, beliefs and memories about attachment figures (inner)
- L1: Beliefs, expectations, appraisals and wishes about attachment experiences, as narratives show them; where thoughts come from is not settled
way out: Attachment behaviour: crying, calling, clinging, following, protest, greeting
- L1: Signals in the air and the body moving that trigger anyone nearby to give care
- L2: Crying, smiling, reaching, clinging, looking
- L3: Organized into goal-directed behaviour from six or seven months
- L4: The system activated: alarm or distance beyond the set-goal
way out: Caregiving: comfort, protection, following the child's lead, limit setting (or frightening behaviour)
- L1: Comfort and contact given, or subtly frightening, frightened or dissociative behaviour
- L4: Sensitivity: noticing, interpreting and quickly answering the child's signals
- L6: The caregiver's state of mind about attachment (AAI) shapes care; unresolved trauma may surface
way out: Exploration from the secure base: play, moving away, engaging with the world
- L1: Play resumed once comforted
- L4: The attachment system relaxed by the figure's presence
way out: Narrative: telling one's attachment story (scripts, the AAI)
- L1: Narratives of 75-300 words, or an interview about one's childhood
- L6: The model put into a story: secure base script knowledge, a coherent or incoherent account
way out: The body's answer: cortisol, heart rate, immune markers
- L1: Physiology that can show distress the behaviour hides (avoidant infants); inflammatory and cortisol markers in adults
- L4: Social buffering of the stress axis by the attachment figure