Object Permanence in Infants: When and How It Develops

The companion piece to this article (Object Permanence: What It Is and Why It Matters) covers the practical day-to-day picture for parents — what the milestones look like, how peek-a-boo works, why separation anxiety emerges. This piece looks one layer down: where the research came from, how Piaget got it partly right and partly wrong, and what 70 years of looking-time experiments have added.

If you're more interested in the practical "what should I expect from my 9-month-old?" angle, the partner article is the one to read. If you find the developmental science interesting in its own right — and want to understand why the textbook timeline is what it is — this is the longer-form version.

The Healthbooq app tracks the everyday milestones around this development as they appear, which gives a clearer picture at health visitor checks than recall does.

Piaget's Original Framework

Jean Piaget worked out the structure of infant object permanence in the 1930s and 40s by carefully observing his three children — Jacqueline, Lucienne, and Laurent — through the first two years of life. The methodology was not what we'd now call rigorous (n=3, no controls), but the observations were astonishingly detailed and the framework has held up remarkably well.

Piaget placed object permanence within the sensorimotor stage (birth to ~2 years), and broke that stage into six substages. The relevant ones for object knowledge:

• Substage 1–2 (0–4 months): the baby's world is what's currently being perceived. A toy that leaves the visual field doesn't trigger search. Piaget's term: a "tableau" of disconnected scenes rather than a stable world.
• Substage 3 (4–8 months): "secondary circular reactions" — the baby starts repeating actions for their effects, including looking back to where something disappeared. Some search for partially hidden objects (a toy under a cloth with a corner sticking out).
• Substage 4 (8–12 months): the breakthrough — active search for fully hidden objects. And the famous A-not-B error: a toy hidden repeatedly in location A, then visibly moved to B in front of the baby, is still searched for at A. The baby's plan is "tied to" the prior successful action, not flexibly updated by new information.
• Substage 5 (12–18 months): A-not-B error fades. The baby tracks visible displacements — they can find a toy moved through a sequence of hiding places they've watched.
• Substage 6 (18–24 months): invisible displacement — the baby can mentally simulate a hiding sequence they didn't fully see. By the end of this stage, object permanence is essentially the same as in adults.

Piaget's broader claim was that object permanence is constructed by the baby through interaction with the world: it's not innate, it's something the infant works out through repeated cycles of looking, reaching, dropping, retrieving.

Where the Field Pushed Back

From the 1980s onward, Renée Baillargeon (Illinois) and Elizabeth Spelke (Harvard, then MIT) developed a different methodological approach: looking-time studies. Instead of asking the baby to act (reach, retrieve), they measured how long babies stared at events.

The classic Baillargeon paradigm (1985, Cognition): a baby watches a screen rotate from flat to upright and back, then a block is placed behind the screen. The screen rotates again. In the "possible" condition, it stops where it would hit the block; in the "impossible" condition, it rotates all the way through (the block has been secretly removed). 4½-month-olds — months before they can pass Piaget's reaching test — stare significantly longer at the impossible event. The interpretation: they expect the block to still be there, even though it's hidden.

Baillargeon, Spelke, and others showed similar effects across many paradigms, with infants as young as 2½–3 months. The conclusion: babies have some knowledge of object persistence much earlier than Piaget thought, but they can't yet act on that knowledge.

This led to the modern reconciliation: there are at least two systems at work.

• Implicit object knowledge — present from very early infancy, possibly innate, visible in looking-time and surprise.
• Explicit object permanence — developed across the 4–18-month window, visible in reaching, retrieval, and search behaviour, requiring the integration of memory, motor planning, and inhibition of prepotent responses.

The A-not-B error in particular has been reinterpreted as more about executive function than about object knowledge. Adele Diamond's work in the 1990s showed that the A-not-B error correlates with development of the prefrontal cortex; babies who fail the task often look toward B even as their hand reaches toward A. They know where the toy is — they just can't override the habituated motor plan.

What Parents See, Versus What's Going On Inside

The clean way to hold all of this:

• The behavioural milestones Piaget described are accurate as a map of when the baby can act on object permanence in front of you.
• The cognitive achievement underneath — knowing things persist when hidden — is partly already there much earlier; what develops is the ability to use that knowledge to plan and act.
• The 8–12-month period is when this comes together visibly, which is also when separation anxiety appears, which is also when parents notice their baby has become recognisably "more of a person."

The practical consequence for parents is that there's nothing to "teach." The development happens through ordinary play and ordinary interaction. The way you can support it is to be a reliable, returning person, to play hiding-and-finding games at the level the baby is currently working on, and to give them objects to retrieve, hide, and post into containers.

The Visible Stages, in Plain Terms

Putting Piaget and the looking-time work together, here's what you'll see in your own baby. Individual variation is wide — these are typical ranges, not deadlines.

Toys that move out of the visual field don't trigger search. Faces that disappear don't reliably trigger surprise. The baby may track a moving object with eyes briefly but loses interest when it leaves the field.

Eye-tracking shows the baby looking toward the place an object disappeared. They watch a ball roll behind a cushion and look at the other side, anticipating its emergence. They may pull a partially-hidden toy out from under a cloth (the corner has to be visible).

The classic milestone. A toy is hidden under a cloth in front of them; they lift the cloth and retrieve it. By 9 months, almost all typically-developing babies pass this. This is also when peek-a-boo transitions from a social game to a hide-and-find game.

Toy hidden in location A, retrieved several times. Toy then visibly moved to location B. The baby reaches for A. This is normal. Piaget interpreted it as incomplete object permanence; modern researchers interpret it as a problem of inhibitory control. Either way, it disappears around 12 months.

Watch a toy go under cloth A, then under cloth B (visible to the baby). The toddler will search at B correctly. They can also chain through more complex movements.

The toddler can deduce where a toy must have gone even if they didn't see all the moves. ("It went into the box, the box is now in the cupboard, the toy must be in the cupboard.") This is essentially adult-form object permanence, and it underlies symbolic play and language: a word can stand for an absent thing, a teddy can stand in for a real animal, because the toddler now reliably represents things that aren't currently visible.

Why It Matters Beyond Babies

Object permanence is foundational for several later cognitive abilities:

• Working memory — the ability to hold information in mind that isn't currently in front of you.
• Symbolic thought — words, drawings, and pretend play all depend on representing things that aren't here.
• Planning — anticipating that future situations exist, and acting toward them.
• Theory of mind — eventually, understanding that other people have mental states (representations of things not currently visible to anyone).

The earliest building block for all of those is the 4–18-month object permanence development. Which is why developmental paediatricians take it seriously even though, day to day, it looks like a baby getting weirdly excited that you came back from the loo.

When to Flag Concerns

The full picture is wide. But speak to your health visitor, GP, or paediatrician if:

• 9 months: no searching for a hidden toy at all, no anticipation of objects reappearing, doesn't watch where dropped objects go, no engagement with peek-a-boo even socially.
• 12 months: still no searching, no joint attention (following your gaze or pointing), no babbling, no recognisable response to their name.
• 18 months: no pretend play, no use of objects to "stand for" other things, no single words, very limited eye contact and shared interest.
• At any age: loss of skills that were previously present (a baby who used to search and now doesn't, or who has stopped engaging with hide-and-find games).

These overlap with the developmental review checklists used at the routine 8–12-month and 18–24-month checks in the UK, and with the autism early-signs list. Most concerns turn out to be normal variation — but the early years are when intervention matters most if support is needed.

Object permanence — knowing things go on existing when out of sight — develops gradually between roughly 4 and 18 months. Piaget mapped the visible behavioural milestones in the 1950s; Baillargeon, Spelke, and others have since shown that babies have implicit object knowledge much earlier than they can act on it. The timeline parents will actually see: tracking by 4–6 months, lifting a cloth to retrieve a toy by 7–8 months, the A-not-B error from 8–12 months, and reliable searching through multiple hides by 12–18 months. The same developmental window produces separation anxiety — the two are not coincidental, they are the same cognitive change wearing different costumes.