Skip to content

Object Permanence

Prime #
1029
Origin domain
Psychology Neuroscience Cognition
Subdomain
cognitive science → Psychology Neuroscience Cognition

Core Idea

Object permanence is the structural pattern of representing unobserved entities as continuing to exist with definite properties. The defining commitment is that a system's internal model maintains entity-existence and entity-identity across periods when the entity is not directly observable. A ball rolled behind a couch still exists in the model; its position, velocity, and identity are propagated through the unobservable interval and re-bound to the same entity when it re-emerges. The commitment is sharper than memory. Memory persists past observations; object permanence is the structural commitment to ongoing existence between observations. The system treats the unobserved interval not as a gap in which entity-status is undefined but as a continuation of the entity's history — a stretch of time over which the entity has state even though no observation is arriving to confirm it.

That distinction is the whole content of the prime, and it carries a precise architecture. There is an entity treated as persistent; an alternation of observable and unobservable intervals; a persistent-state representation carried through the unobservable stretch; a prediction mechanism that propagates that state in the absence of observation; a re-binding mechanism that matches new observations to the same entity on re-emergence; and an underlying commitment — a design choice, not a given — to model this particular entity as persistent rather than ephemeral. The pattern is substrate-independent because none of these roles names a medium. Wherever a system maintains persistent state for unobservable entities, explicitly modeling them as continuing to exist with definite properties during the unobserved interval and re-binding observations on re-emergence to the same entity, the same role structure operates and the same interventions become available. Its formalization is well known: the Kalman filter, with its predict step that propagates latent state between measurements and its update step that re-binds incoming observations, is engineered object permanence, and state-space modeling generally is the latent-state grammar this prime names.

How would you explain it like I'm…

Still There When Hidden

When a ball rolls behind the couch, it's still there — it didn't disappear just because you can't see it. Your mind keeps a little picture of it: where it is, where it's going, that it's the SAME ball. So when it rolls back out, you're not surprised. Things keep existing even when you're not looking.

The Ball Behind the Couch

Object permanence is treating something you can't see right now as still being there, with real properties, even while it's hidden. If a ball rolls behind a couch, your mind keeps track that it still exists, roughly where it is, and which ball it is, so when it comes out the other side you know it's the same ball. This is more than just remembering: remembering is about past sightings, while object permanence is about the thing continuing to exist in the gap between sightings. So the hidden time isn't treated as a blank where the ball's status is unknown; it's treated as the ball quietly carrying on.

Existence Between Glimpses

Object permanence is the pattern of representing unobserved things as continuing to exist with definite properties. The defining commitment is that a system's internal model keeps an entity's existence and identity alive across the times it can't be seen: a ball rolled behind a couch still exists in the model, its position and velocity are carried through the hidden stretch, and it's re-matched to the same entity when it reappears. This is sharper than memory, which only persists past observations; object permanence is the commitment to ongoing existence between observations. So the unobserved interval isn't a gap where the entity's status is undefined, it's a continuation of the entity's history, a stretch of time over which it still has a state even though nothing is confirming it. That single distinction is the whole content, and it relies on a precise set of parts: a persistent entity, alternating visible and hidden intervals, a carried-through state, a prediction step, and a re-matching step on reappearance.

 

Object permanence is the structural pattern of representing unobserved entities as continuing to exist with definite properties. The defining commitment is that a system's internal model maintains entity-existence and entity-identity across periods when the entity is not directly observable: a ball rolled behind a couch still exists in the model, and its position, velocity, and identity are propagated through the unobservable interval and re-bound to the same entity when it re-emerges. The commitment is sharper than memory — memory persists *past* observations, whereas object permanence is the commitment to *ongoing existence between* observations. The system treats the unobserved interval not as a gap in which entity-status is undefined but as a continuation of the entity's history: a stretch of time over which the entity has state even though no observation is arriving to confirm it. That distinction is the whole content of the prime, and it carries a precise architecture: an entity treated as persistent; an alternation of observable and unobservable intervals; a persistent-state representation carried through the unobservable stretch; a prediction mechanism that propagates that state in the absence of observation; a re-binding mechanism that matches new observations to the same entity on re-emergence; and an underlying design choice — not a given — to model this particular entity as persistent rather than ephemeral. The pattern is substrate-independent because none of these roles names a medium. Its formalization is well known: the Kalman filter, with a predict step that propagates latent state between measurements and an update step that re-binds incoming observations, is engineered object permanence, and state-space modeling generally is the latent-state grammar this prime names.

Structural Signature

the entity treated as persistent — the alternation of observable and unobservable intervals — the persistent-state representation carried through the gap — the prediction mechanism propagating it — the re-binding mechanism on re-emergence — the design-choice commitment to model it as persistent

The pattern is present when each of the following holds:

  • An entity treated as persistent. Some object, account, agent, or quantity is modeled as continuing to exist with definite properties.
  • An alternation of observable and unobservable intervals. The entity is sometimes directly observable and sometimes not; there is a gap during which no observation confirms its state.
  • A persistent-state representation. The model carries the entity's state — position, velocity, identity, properties — through the unobservable interval, treating it as ongoing existence rather than an undefined gap.
  • A prediction mechanism. Some process propagates the carried state forward in the absence of observation (a Kalman predict step, a physics extrapolation, a theory-of-mind forecast).
  • A re-binding mechanism. When the entity re-emerges, incoming observations are matched to the same persistent entity rather than treated as new — the operation whose failure is an identity switch.
  • A permanence commitment. Modeling this entity as persistent is a design choice, not a given; it costs storage, prediction, and re-binding machinery, and can be right or wrong.

The commitment is sharper than memory: memory persists past observations, while object permanence commits to ongoing existence between observations. These compose into a latent-state grammar — carry state through the gap, predict where the entity should be, re-bind the next observation to the same continuing thing.

What It Is Not

  • Not memory. associative_memory and memory generally persist state past observations; object permanence commits to ongoing existence between observations — modeling the entity as continuing to have state during the gap, not merely retaining its last-seen value.
  • Not determinism. determinism is the claim that state evolves by fixed law; object permanence is the representational commitment that an unobserved entity persists — orthogonal to whether its evolution is deterministic or stochastic.
  • Not a mental model. A mental_model is a general internal representation of how something works; object permanence is the narrower commitment to entity-existence and identity across unobserved intervals, with prediction and re-binding machinery.
  • Not belief formation. belief_formation is how a system arrives at what it holds true; object permanence is the prior structure that an entity continues to exist between observations, which belief updating then operates over.
  • Not pattern completion. pattern_completion fills in missing parts of a currently-presented pattern; object permanence carries a whole entity's state through time when no observation is present at all.
  • Common misclassification. Satisfying a permanence requirement with mere memory — storing the last observed state without propagating it as a continuing entity, so the system "remembers" where the entity was but is surprised when it re-emerges elsewhere.

Broad Use

  • Cognitive development (origin) — Piaget's infant studies established object permanence as a milestone acquired in the first year; later violation-of-expectation work pushed the onset earlier, but the structural achievement (treat occluded objects as persisting) is the same.
  • Computer-vision multi-object tracking — a tracker must hold identity for objects that occlude one another or briefly leave frame; Kalman filtering, re-identification, and track management are the field's formalization, and loss of permanence shows up as ID switches.
  • Database theory and accounting — entities (customers, assets, accounts) persist across periods of no transactional activity, with the store maintaining state through silent intervals and re-binding to the same entity on the next transaction; inventory audits reconcile a period's closing stock with the next period's opening stock.
  • Latent-variable econometrics — unobservable quantities (potential GDP, NAIRU, underlying inflation) are modeled as persisting between data releases, with state-space models and Kalman filters formalizing permanence for latent state.
  • Physics — mass-balance and energy-balance reasoning is object permanence at the substrate level: matter and energy persist through unobserved intervals, and atoms do not vanish when one stops looking.
  • Theory of mind — representing other agents as having persistent beliefs, desires, and intentions across periods of no interaction is interpersonal object permanence applied to mental states.
  • Operating systems — suspended VMs, idle connections, and background processes persist with state through inactivity, and the OS guarantees re-binding of identity on reactivation.

Clarity

Naming object permanence clarifies a load-bearing representational commitment that is usually left implicit. Many design choices become tractable once "do we treat this entity as persisting?" is asked explicitly: what to track, what to forget, when to re-bind, how to detect identity loss. The commitment to persistence is not free — it costs storage, prediction, and re-binding machinery — so making it explicit turns it into a decision that can be defended where the entity really is persistent and questioned where it is not. The reframing also exposes a class of permanence failures that are otherwise invisible as a category: ID switches in trackers, account collisions in databases, identity loss on re-meeting people after long absence, organizational forgetting of dormant capabilities. These share a structural shape — loss of identity binding across an unobserved interval — and a shared remediation: more reliable re-binding cues, whether re-identification features, primary keys, or contextual recall. Seeing them as instances of one failure rather than unrelated bugs is what the prime buys.

Manages Complexity

The pattern compresses a wide family of "represent unobserved entities as continuing" phenomena into one diagnostic family with a reusable intervention structure. Cross-substrate, the moves sort into three: what cues are used for re-binding (visual features, account IDs, names, mass balance); what guarantees track continuity (Kalman prediction, database keys, succession plans, conservation laws); and what detects identity failure (track-switch metrics, collision detection, identity disputes). The toolkit of better re-id features, key uniqueness, succession clarity, and contextual recall aids carries across substrates because each addresses the same underlying role. Beyond this, the prime supplies the latent-state grammar in its full generality: unobservable things have ongoing state that the system models, and observations update the model without constituting the entity's existence. That grammar is the common spine under state-space modeling, Bayesian filtering, accounting, theory of mind, and developmental cognition, which is why a method developed in one of these can be recognized and borrowed in another rather than reinvented.

Abstract Reasoning

Recognizing object permanence as a structural pattern enables several reasoning moves. The hidden-state grammar: any system reasoning under partial observability admits the questions of what the persistent state is, how observations update beliefs about it, and what the prior is between observations — the state-space grammar in formal terms, the theory-of-mind grammar in interpersonal terms, the latent-variable grammar in statistics. The re-binding problem: when an entity re-emerges, how is it recognized as the same entity, a question with structural commonality across re-identification features, primary keys, facial recognition, mass balance, and signature retrieval. The permanence-versus-ephemerality choice: some entities should be modeled as persistent (people, accounts, conserved quantities) and others as ephemeral (transient signals, one-off transactions, dissipating heat), with design errors flowing from getting this wrong in either direction — false alarms from treating the ephemeral as persistent, identity loss from treating the persistent as ephemeral. The developmental or training trajectory: object permanence is acquired, whether by infants, trackers, or organizations building process maturity, so understanding what cues support its acquisition (trajectory continuity, partial occlusion, re-id features) is itself a substrate-independent design problem. And the permanence-failure mode: systems that once had permanence can lose it under load, fragmentation, or disruption, with a common diagnostic and remediation across cognition, databases, and institutions.

Knowledge Transfer

The role mappings are stable: the entity maps onto the object, account, agent, or quantity treated as persistent; the unobservable interval onto any stretch without direct observation; the persistent state onto the model's representation carried through that stretch; the prediction mechanism onto whatever propagates state (a Kalman predict step, a physics extrapolation, a theory-of-mind forecast, latent-state dynamics); the re-binding mechanism onto whatever matches re-emergent observations to the persistent entity; and the permanence commitment onto the structural decision to model the entity as persisting at all. With these fixed, both the latent-state grammar and the re-binding toolkit transfer, and the historical record shows the transfers happening explicitly. Piaget's developmental insight — model unobservable entities as persisting with definite state — moved into robotic perception via Kalman filtering and SLAM, so that modern visual SLAM is in effect engineered object permanence for the environment. Computer-vision tracking's findings about re-identification cues transferred into customer-identity and know-your-customer systems. State-space models from econometrics inform cognitive-neuroscience accounts of perceptual inference, where hidden-state inference and predictive coding are object permanence for perceptual causes. Theory-of-mind research informs conversational AI that maintains user-model state across sessions. And mass-balance reasoning is the foundation of process-engineering material and energy balances. A self-driving car holding a representation of a cyclist occluded behind a truck, an infant reaching behind a couch for a vanished ball, a database re-binding a dormant account on next login, and an economist treating underlying inflation as a persistent latent variable between releases are all doing the same structural work: carry state through the gap, predict where the entity should be, and re-bind the next observation to the same continuing thing.

Examples

Formal/abstract

Consider a Kalman filter tracking an aircraft on radar — the prime's exact mathematical formalization. The entity treated as persistent is the aircraft, modeled as having a definite state vector (position, velocity) at all times. The alternation of observable and unobservable intervals is the gap between radar sweeps: a measurement arrives every few seconds, and between sweeps no observation confirms where the aircraft is. The persistent-state representation carried through the gap is the filter's state estimate \(\hat{x}\) together with its covariance \(P\), propagated continuously even when no measurement is available. The prediction mechanism propagating it is the predict step: \(\hat{x}_{k|k-1} = F\hat{x}_{k-1|k-1}\), advancing the state through the unobserved interval using the motion model \(F\), while the covariance grows (\(P_{k|k-1} = FP_{k-1|k-1}F^\top + Q\)) — the model's uncertainty about the unobserved entity increases but the entity is never treated as gone. The re-binding mechanism on re-emergence is the update step: when the next radar return arrives, the filter computes the innovation (observed minus predicted), gates it to decide whether this return belongs to this track, and if so folds it in — re-binding the new observation to the same persistent entity rather than spawning a phantom. The design-choice commitment to model it as persistent is explicit: the engineer chose a state-space model that carries the target through dropouts, paying storage and compute, precisely because aircraft do not blink out of existence between sweeps. The prime's signature failure mode — the identity switch — appears concretely when two aircraft cross and the gating mis-assigns returns, swapping the two tracks: a re-binding failure, remedied by richer discriminating features or tighter gating, exactly the prime's prescription.

Mapped back: The aircraft is the persistent entity, the inter-sweep gap the unobservable interval, the state estimate the persistent representation, the Kalman predict step the prediction mechanism, and the gated update the re-binding — the filter is engineered object permanence, with track-swap the permanence-failure mode.

Applied/industry

Consider a relational database maintaining customer accounts, alongside the structurally identical case of an infant reaching behind a couch for a vanished ball — two genuine domains sharing the latent-state grammar. In the database case the entity treated as persistent is a customer account; the alternation of observable and unobservable intervals is the stretch of inactivity between transactions, during which the customer takes no action the system can see. The persistent-state representation carried through the gap is the stored account record — balance, status, history — which the database holds intact through arbitrarily long silent intervals rather than discarding when no activity arrives; this is the prime's sharper-than-memory commitment, modeling ongoing existence between observations, not merely retaining past ones. The prediction mechanism is the set of rules that propagate state through the gap (accruing interest, applying scheduled fees, aging a balance) without any new input. The re-binding mechanism on re-emergence is the primary key: when the customer logs in again after months, the incoming session is matched to the same account via the key, not treated as a new customer — and the prime's permanence-failure mode appears as an account collision or a duplicate record when re-binding fails (a stale cookie, a mistyped identifier), the database analogue of an ID switch. The permanence commitment is a deliberate, costly design choice: storing dormant accounts consumes space and reconciliation effort, justified because customers genuinely persist. The infant case maps role-for-role — the ball is the persistent entity, the occlusion the unobservable interval, the maintained representation the carried state, trajectory extrapolation the prediction, and recognizing the re-emerging ball as the same ball the re-binding. A database re-binding a dormant account on next login and an infant reaching for an occluded ball do the same structural work: carry state through the gap and re-bind the next observation to the same continuing thing.

Mapped back: The account (or ball) is the persistent entity, the inactive stretch (or occlusion) the unobservable interval, the stored record (or mental representation) the persistent state, the primary key (or visual recognition) the re-binding mechanism, and storing dormant accounts (or expecting the ball) the permanence commitment.

Structural Tensions

T1 — Permanence versus Ephemerality (sign/direction). The permanence commitment is a choice, and the same machinery can be wrongly applied in either direction. Modeling a genuinely ephemeral entity as persistent generates phantom state — false alarms, tracks for objects that are gone, accounts for customers who never existed. Modeling a genuinely persistent entity as ephemeral discards state that should have carried — identity loss, forgotten dormant capabilities, re-meeting a known entity as a stranger. Diagnostic: ask whether the entity actually continues to exist between observations, independent of whether observation is convenient — treating the transient as permanent and the permanent as transient are opposite errors of the same misjudged commitment, and only the entity's real ontology, not the observation pattern, settles which applies.

T2 — Carried State versus Growing Uncertainty (temporal). The prediction mechanism propagates state through the gap, but the model's uncertainty about that state grows the longer observation is absent — the Kalman covariance inflates between measurements. The failure mode is treating the carried state as if it retained its original confidence: acting on a long-extrapolated position as though freshly observed, when the entity may have diverged arbitrarily far from the prediction. Diagnostic: ask whether the age of the last observation is tracked alongside the state estimate — a system that carries the mean but not its growing variance will over-trust stale predictions, and the longer the unobserved interval, the more the propagated state is a guess wearing the costume of a measurement.

T3 — Re-binding Correctly versus Identity Switch (scopal). Re-binding matches a re-emergent observation to the same persistent entity — but the matching can fail two ways, and they are opposite. Under-binding spawns a phantom: the returning entity is treated as new (duplicate account, lost track, forgotten acquaintance). Over-binding collides identities: a different entity is matched to the existing one (ID switch when two aircraft cross, account collision, mistaken recognition). The failure mode is tuning the gate for one error and inflating the other. Diagnostic: ask whether re-binding failures present as duplicates (gate too tight, under-binding) or collisions (gate too loose, over-binding) — these demand opposite adjustments to the matching threshold, and a single "improve re-id" framing hides which way the gate must move.

T4 — Permanence versus Memory (scopal). Object permanence is sharper than memory: memory persists past observations, while permanence commits to ongoing existence between them. The boundary matters because they call for different machinery. The failure mode is satisfying a permanence requirement with mere memory — storing the last observed state without propagating it as a continuing entity, so the system "remembers" where the ball was but does not model it as having moved while hidden, and is surprised when it re-emerges elsewhere. Diagnostic: ask whether the system models the entity as having continued to have state during the gap (permanence) or merely retains its last seen state (memory) — a static last-known-value is memory masquerading as permanence and will fail exactly when the entity changed unobserved.

T5 — Permanence Cost versus Tracking Budget (measurement). Modeling an entity as persistent is not free — it costs storage for the carried state, compute for prediction, and machinery for re-binding. The tension is that permanence is valuable per-entity but the budget is finite across all entities. The failure mode is committing to permanence for too many entities and exhausting the tracking budget, so that genuinely persistent entities lose state because the system is busy maintaining ephemeral ones. Diagnostic: ask whether the count of entities held as persistent fits the storage-and-compute budget, and whether low-value persistent commitments are crowding out high-value ones — permanence is a paid choice, and an unbudgeted one degrades the tracking of exactly the entities that most needed it.

T6 — Model Persistence versus Substrate Persistence (sign/direction). Object permanence is a property of the model, not of the world — the system commits to the entity continuing, and that commitment can diverge from what actually happened. The failure mode is conflating the two: assuming that because the model carries the entity, the entity is truly there (the cyclist the model still tracks may have turned off; the account the database holds may belong to a closed business). Diagnostic: ask what would falsify the persistence assumption and whether the system checks for it — a model that only ever propagates and re-binds, never expiring or disconfirming, will hold ghosts indefinitely, so genuine permanence requires both the commitment to continuation and a mechanism to retire entities the world has actually removed.

Structural–Framed Character

Object permanence is a mixed-structural prime, sitting just on the structural side of the structural–framed spectrum. Its skeleton is a latent-state grammar — an entity modeled as persisting through unobservable intervals, with a prediction step propagating its state and a re-binding step matching new observations back to the same entity — and that pattern has an exact formalization in the Kalman filter and state-space modeling generally. It recurs in database records held between accesses, in trackers maintaining identity across occlusion, and in any system carrying hidden state. The Piagetian developmental name is what keeps it in from the bare end.

The diagnostics read structural with one translatable seam. The pattern carries no evaluative weight: maintaining persistent state for an unobserved entity is neither good nor bad — holding "ghosts" of removed entities is a named failure mode, not a moral one. It is not human-practice-bound (human_practice_bound 0): a Kalman filter's predict step propagates latent state between measurements with no human in the loop, and any state-space estimator instantiates the persist-and-rebind structure in pure computational substrates. And invoking it largely recognizes a latent-state commitment already present — that the system models the unobserved interval as a continuation rather than a gap is a fact about its architecture, read off rather than imported. What pulls it to the center is the home vocabulary: "object permanence" arrives from developmental psychology and must be translated into latent-state terms when the substrate is a filter or a database (vocab_travels and import_vs_recognize each 0.5, institutional_origin 0.5 for the field of origin). The unobservable-state-tracking core is substrate-free — the Kalman filter is its engineered form — and the Piagetian label is a thin overlay, exactly the mixed-structural reading the aggregate of 0.3 records.

Substrate Independence

Object permanence is a strongly substrate-independent prime — composite 4 / 5 on the substrate-independence scale. On domain breadth, the maintain-persistent-state-for-unobservable-entities pattern recurs across cognitive development (its origin — the infant treating an occluded object as persisting), computer-vision multi-object tracking (holding identity across occlusion, with ID switches as the failure mode), database theory and accounting (entities persisting across silent intervals, reconciled on the next transaction), latent-variable econometrics (potential GDP, NAIRU modeled as persisting between releases), physics (mass- and energy-balance conservation), theory of mind (persistent beliefs across non-interaction), and operating systems (suspended VMs, idle connections) — a wide span earning a 5 on breadth. On structural abstraction, the latent-state grammar (carry state through the gap, predict where the entity should be, re-bind on re-emergence) is medium-neutral and has an exact engineered form in the Kalman filter's predict-and-update steps, running with no human in the loop; the Piagetian developmental vocabulary needs translating to a filter or database, holding abstraction at 4. On transfer evidence, the prime scores a 5 — the ports are explicit and documented: Piaget's developmental insight moved into robotic perception via Kalman filtering and SLAM (modern visual SLAM is engineered object permanence), state-space models from econometrics inform cognitive-neuroscience accounts of perceptual inference, and CV re-identification cues transferred into customer-identity systems. The strong concrete transfer lifts the composite to a robust 4.

  • Composite substrate independence — 4 / 5
  • Domain breadth — 5 / 5
  • Structural abstraction — 4 / 5
  • Transfer evidence — 5 / 5

Neighborhood in Abstraction Space

Object Permanence sits among the more crowded primes in the catalog (17th percentile for distinctiveness): several abstractions describe nearly the same structure, so a description that fits it will tend to fit its neighbors too — transporting it usually means disambiguating within this family rather than landing on it exactly.

Family — Generative Rules & Stage-Wise Change (19 primes)

Nearest neighbors

Computed from structural-signature embeddings · 2026-06-14

Not to Be Confused With

The most important confusion is with associative_memory — memory in general — because both involve a system holding onto something across time. The prime itself insists the distinction is sharper than it looks. Memory persists information past observations: it retains what was seen, a record of the last state. Object permanence commits to ongoing existence between observations: it models the entity as continuing to have state — moving, changing, accruing — during the very interval when no observation is arriving. The difference is the whole content of the prime. A system with memory but not permanence stores where the ball was last seen; a system with permanence models the ball as having rolled onward while hidden and predicts where it should now be. The failure of conflating them is concrete: satisfying a permanence requirement with a static last-known-value (memory masquerading as permanence) produces exactly the surprise the prime warns of — the entity re-emerges somewhere the stored value never anticipated, because the system retained the past instead of propagating a continuing present.

It must also be distinguished from determinism, its nearest embedding neighbor, with which it is easily entangled because both concern what happens to a system over time. Determinism is a claim about dynamics: that the next state follows from the current one by fixed law, with no randomness. Object permanence is a claim about representation: that an entity continues to exist with definite properties while unobserved, regardless of whether its evolution is deterministic or stochastic. The two are orthogonal — a Kalman filter embodies object permanence precisely while modeling stochastic dynamics with growing uncertainty, and a deterministic system may make no permanence commitment at all. The error of merging them is to assume that carrying an entity through the gap requires predicting it exactly; in fact permanence is compatible with — and usually accompanied by — inflating uncertainty about the propagated state, which determinism would deny.

A third confusion is with belief_formation, since both concern how a system represents things it is not directly seeing. Belief formation is the process of arriving at and updating what the system holds true given evidence. Object permanence is the prior structural commitment that a particular entity continues to exist between observations — the scaffolding that belief updating then operates over. When a re-emergent observation arrives, belief formation decides how to fold it in; object permanence is what made there be a persistent entity to fold it into. Treating permanence as just another belief misses that it is a design-level decision about the system's ontology — which entities get persistent state at all — rather than a content-level judgment about any particular fact.

For a practitioner these distinctions decide the architecture. A memory frame stores last-seen values and fails on entities that moved while hidden; a determinism frame over-trusts exact propagation and ignores growing uncertainty; a belief-formation frame updates facts without committing to which entities persist. Object permanence's contribution is the full latent-state grammar — carry state through the gap, propagate it with honest growing uncertainty, and re-bind the next observation to the same continuing entity — together with the explicit, costly, and falsifiable choice of which entities to model as persistent in the first place.

Solution Archetypes

No catalogued solution archetypes reference this prime yet.