HookThe baby monkeys chose comfort over food
In the late 1950s Harry Harlow gave infant rhesus monkeys a cruel choice between two surrogate 'mothers': one made of bare wire that dispensed milk, and one covered in soft cloth that gave nothing but comfort. If attachment were really about food, the babies should have clung to the wire mother. They did not. They spent almost the whole day — around 22 hours — pressed against the cloth mother, darting to the wire one only to feed before rushing back. When something frightened them, they ran to cloth, not milk.
Harlow's monkeys demolished the tidy idea that a baby loves whoever feeds it — 'cupboard love' — and pointed instead at contact comfort and an emotional bond as the foundation of attachment. This topic follows that bond from its first flickers in caregiver-infant interaction, through the theories that explain it and the famous method that measures it, to what happens when it is broken — and how the very first relationship becomes a template for every relationship that follows. Every essay here turns on one question: how much does that first attachment shape the whole life?
ModelThe first conversations — reciprocity, synchrony and Schaffer's stages
Attachment begins in the fine-grained back-and-forth between caregiver and infant. Reciprocity is a two-way, turn-taking exchange in which each partner responds to and elicits a response from the other — Brazelton likened it to a 'dance', with babies signalling readiness through alert phases. Interactional synchrony is caregiver and infant mirroring each other's actions and emotions in a coordinated, simultaneous way; Meltzoff and Moore (1977) found infants as young as two weeks imitated adults' facial expressions and gestures, and Isabella et al. (1989) found that higher synchrony predicted stronger, better-quality attachment.
Schaffer and Emerson (1964) studied 60 Glasgow babies monthly and mapped four stages of attachment: the asocial stage (first few weeks — similar responses to people and objects), indiscriminate attachment (2–7 months — a clear preference for people but comfort accepted from anyone), specific attachment (around 7 months — one primary attachment figure, with stranger and separation anxiety appearing; the mother in 65% of cases), and multiple attachments (by around one year — secondary attachments to other familiar people). On the role of the father, fathers were the sole first attachment in only 3% of cases but a joint first object in 27%; Grossmann (2002) found fathers' play-based interaction predicted later attachment quality, and Field (1978) found primary-caregiver fathers showed the same sensitive responsiveness as mothers — suggesting the role depends on responsiveness, not gender.
CaseGeese and monkeys — what animal studies reveal
Two animal studies anchor this topic. Lorenz (1935) divided a clutch of greylag goose eggs: half hatched with the mother, half in an incubator where Lorenz was the first moving thing they saw. The incubator group followed Lorenz everywhere — imprinting — and it occurred only within a narrow critical period of hours to a couple of days. Imprinting was irreversible and shaped later mate choice (sexual imprinting): the birds later directed courtship at the species, or object, they had imprinted on. Harlow (1958) gave infant rhesus monkeys the wire and cloth surrogate mothers described above; the babies overwhelmingly chose contact comfort over the milk-giving wire mother. The long-term damage was severe: monkeys reared on surrogates grew up socially and sexually abnormal, were aggressive, and became neglectful or abusive mothers themselves, with a critical period of about 90 days after which the harm was hard to reverse.
For AO3, generalising from birds and monkeys to humans is risky because human attachment involves far richer emotional and cognitive processes, and Harlow's studies carry heavy ethical costs that must be weighed against their real influence on how we understand child care.
ModelTwo explanations — cupboard love versus Bowlby's monotropy
Learning theory (Dollard and Miller, 1950), nicknamed 'cupboard love', says attachment is learned through association and reward. By classical conditioning, the caregiver (a neutral stimulus) is repeatedly paired with food (an unconditioned stimulus that produces pleasure) until the caregiver alone becomes a conditioned stimulus producing pleasure and comfort. By operant conditioning, a baby's crying is positively reinforced when it brings food and comfort, while the caregiver is negatively reinforced because attending to the baby stops the crying; food is a primary reinforcer and the caregiver becomes a secondary reinforcer.
Bowlby's monotropic theory rejects this, arguing attachment is innate and evolved to aid survival. Monotropy is the idea of one special attachment, more important than all others, and the more consistent time spent with this figure the better (the law of continuity and the law of accumulated separation). Social releasers — a baby's innate cute behaviours such as smiling, cooing and gripping — unlock adult caregiving. Attachment must form within a critical period of around 2.5 years (a sensitive period up to about 5) or it may not form well at all. The first attachment builds an internal working model, a mental template for what relationships are, which shapes later friendships, romantic bonds and one's own parenting — the continuity hypothesis.
A model AO3 paragraph evaluating learning theory shows how to turn a criticism into a full argument: 'A major weakness of learning theory is that it is contradicted by animal research. If attachment formed simply because the caregiver is associated with food, Harlow's infant monkeys should have attached to the wire mother that fed them; instead they spent almost all their time clinging to the cloth mother that provided only contact comfort. This undermines the central claim that food is the basis of attachment and suggests that comfort and emotional security matter more. It also has real-world weight, because if attachment depended on feeding we would expect babies always to attach to whoever feeds them, yet Schaffer and Emerson found many infants formed their primary attachment to a parent who was not their main feeder.' The paragraph names a specific study, states exactly what its result predicts under the theory, explains why the mismatch is damaging, and then reinforces it with a second, converging piece of evidence. That build — evidence, implication, corroboration — is precisely the sustained line of reasoning the top AO3 band rewards, rather than a one-line 'Harlow disproves this'.
CaseMeasuring the bond — the Strange Situation and cultural variation
Ainsworth's Strange Situation (1970) was a controlled observation of around 100 middle-class American infants across eight three-minute episodes, measuring exploration and safe-base behaviour, stranger anxiety, separation anxiety and — most tellingly — reunion behaviour. It produced three attachment types. Secure (Type B, about 66%): the infant explores happily using the caregiver as a safe base, is moderately distressed on separation and easily comforted at reunion. Insecure-avoidant (Type A, about 22%): explores freely, shows little stranger or separation anxiety, and avoids contact at reunion. Insecure-resistant (Type C, about 12%): clingy and reluctant to explore, highly distressed on separation, and resists comfort at reunion — seeking and rejecting at once.
Van IJzendoorn and Kroonenberg (1988) meta-analysed 32 studies across 8 countries (about 1,990 infants). Secure was the most common type everywhere, but variation within a country was about 1.5 times greater than between countries; Germany showed more avoidant infants (a culture valuing independence) and Japan and Israel more resistant infants (where infants are rarely separated from mothers). This suggests a universal, innate basis to attachment with cultural shading on top. For AO3, the Strange Situation may be an imposed etic — a US-designed tool applied to other cultures — and may measure one specific relationship rather than a fixed trait of the child.
MechanismWhen the bond breaks — deprivation and institutionalisation
Bowlby's maternal deprivation hypothesis claims that continual disruption of the attachment during the critical period causes lasting harm — a lowered IQ and, emotionally, affectionless psychopathy (an inability to feel guilt or empathy). His 1944 '44 thieves' study found that of 44 juvenile thieves, 14 showed affectionless psychopathy and 12 of those had experienced prolonged early separation from their mothers, versus almost none in a non-criminal control group. (Note the distinction Rutter later stressed: deprivation is losing an attachment that had formed, whereas privation is never forming one — and the severest effects tend to follow privation.)
Studies of institutionalisation sharpened the picture. Rutter's ERA study (2011) followed 165 Romanian orphans adopted into British families: those adopted before 6 months later had a mean IQ around 102, against roughly 86 for those adopted between 6 months and 2 years and 77 for those adopted after 2 years, and late-adopted children often showed disinhibited attachment (over-friendly, attention-seeking behaviour towards strangers). Zeanah et al. (2005), in the Bucharest Early Intervention project, found 74% of a control group were securely attached against just 19% of the institutionalised group. Effects include physical underdevelopment, disinhibited attachment and intellectual delay — much of it recoverable if adoption happens early. For AO3, this research has real-world value (it reshaped care towards key workers and earlier adoption), but natural studies carry confounding variables such as poor pre-adoption conditions.
MechanismThe template for life — early attachment and later relationships
The mechanism linking infancy to adulthood is the internal working model: the first attachment sets an expectation of what relationships offer, which is carried forward. In childhood, securely attached children tend to form better friendships and are less caught up in bullying — Myron-Wilson and Smith (1998) found insecure-avoidant children were more likely to be victims and insecure-resistant children more likely to be bullies. In adult romantic relationships, Hazan and Shaver's (1987) 'love quiz' in an American newspaper drew about 620 replies and found that securely attached adults reported happier, more trusting and longer-lasting relationships, avoidant adults feared intimacy, and resistant adults were preoccupied and prone to jealousy — mirroring the infant types. In parenting, Bailey et al. (2007) studied 99 mothers and found the attachment type they had experienced with their own mothers tended to be passed on to their babies, evidence that the internal working model shapes how we care for the next generation.
For AO3, this research is correlational — attachment type and later outcomes may share a third cause such as temperament — the love quiz relied on retrospective self-report open to recall bias, and it is overly deterministic to imply an insecure infant is doomed, since many go on to form healthy adult bonds. Attachment is a probabilistic influence, not a life sentence.
VocabularyKey terms the mark scheme pays for
TrapsMisconceptions that cost marks
ExamWhat examiners want
AQA marks AO1 (describe), AO2 (apply to the stem) and AO3 (evaluate); 16-mark essays are 6 AO1 and 10 AO3. Studies are your currency here — name Ainsworth's three types with their percentages, quote van IJzendoorn and Kroonenberg's finding that within-culture variation was about 1.5 times greater than between cultures, and cite Rutter's IQ figures for the different adoption ages. Precise study detail is what separates detailed AO1 from vague description.
For AO3, the strongest evaluation is usually methodological — the validity of the Strange Situation as an imposed etic, the risk of extrapolating from Lorenz's geese and Harlow's monkeys to humans, and the correlational nature of later-relationship research — each elaborated into a consequence rather than stated and dropped. Guard the deprivation/privation distinction: it is a favourite discriminator, and mislabelling the 44 thieves study as privation (or vice versa) signals shaky understanding. On application and 'role of the father' questions, lift specific details from the stem and name the process (interactional synchrony, sensitive responsiveness, the internal working model) that fits — a scenario answer that never returns to the scenario cannot reach the top band.