Does cognitive science literature formally distinguish schema formation from schema restructuring?

Topic question: Does cognitive science literature formally distinguish schema formation (building new schemas from scratch) from schema restructuring (modifying existing ones), and what does that imply for the effort and conditions each requires?

Related concepts: schema-theory · schema-restructuring · schema-formation · conceptual-change · accretion-tuning-restructuring · SEEK-model · weak-vs-strong-restructuring · cognitive-conflict

Summary answer

Qualified yes, with a terminology caveat. At least one current source explicitly and formally names "schema formation" and "restructuring" as two distinct mechanisms in a shared taxonomy (the 2026 SEEK model, below). A structurally similar two-way distinction — build genuinely new representational structure vs. reorganize the relations among existing ones — recurs across roughly 45 years of cognitive-science and conceptual-change literature, but under shifting labels: the classic schema-theory term "restructuring" (Rumelhart & Norman, 1978) originally meant what the 2026 SEEK model calls "schema formation," while the conceptual-change tradition (Carey; Vosniadou & Brewer) split the same territory into "weak" vs. "strong/radical" restructuring. Whether the effort and conditions required for forming a new schema differ from those required for restructuring an existing one, specifically, is not clearly resolved by the sources found this session — most sources instead contrast the pair (formation + restructuring) jointly against lower-effort learning (accretion/tuning), rather than contrasting formation against restructuring directly. That narrower sub-question is marked [unverified — could not confirm or deny after search] below.

Claim 1 — The SEEK model (2026) names schema formation and restructuring as two of four formally distinct learning mechanisms

Claim type: Definitional (naming/taxonomy) and technical-mechanism (what triggers each) → floor Tier 1–2 required for the mechanism portion. Flag: [unverified-mechanism — needs primary] — quote is from the published abstract as reproduced by a search engine; direct fetch of the ScienceDirect page and its open-access eScholarship mirror both failed (403) this session, so full-text confirmation is outstanding.

The most direct and recent confirmation of the topic question comes from Xu, Sandhofer & Stigler's Schema Expansion and Error-driven Knowledge-building (SEEK) model, which explicitly lists "schema formation" and "restructuring" side by side as two of four named mechanisms, alongside accretion and tuning:

"conceptualizes learning as a dynamic interplay among four mechanisms — accretion, tuning, restructuring, and schema formation — each triggered by prediction error and modulated by automatic or deliberate curiosity"

This is the field's most explicit head-to-head naming of the two processes as formally distinct: not two words for one phenomenon, but two of four discrete slots in a single model. However, per the abstract as reproduced, the model's stated dimension of differentiation for conditions is curiosity type, and it groups restructuring and schema formation together on the "deliberate curiosity" (metacognitive) side, opposite accretion and tuning on the "automatic curiosity" (perceptual) side — it does not, in the material accessible this session, further differentiate restructuring from schema formation from each other in terms of relative effort.

Claim type check: Definitional/taxonomic naming — supported at Tier 1 venue via reproduced abstract, flagged for full-text confirmation. ✓ (with flag)

Claim 2 — In the founding schema-theory paper (Rumelhart & Norman, 1978), "restructuring" originally meant creating brand-new schemas, not modifying existing ones

Claim type: Historical/definitional claim about a named theory's terminology → Tier 3–4 acceptable when uncontested. Sources: ERIC record (Tier 3-4) and InstructionalDesign.org (Tier 3-4); primary PDF located but not machine-extractable this session, so effort/frequency specifics below carry [unverified-mechanism — needs primary].

Rumelhart & Norman's original three-mode theory (1976 technical report; published 1978 in Semantic Factors in Cognition) is the field's foundational schema-learning taxonomy, and it is the ancestor of the SEEK model's four-way split. Per the ERIC abstract record for the report:

"The learner acquires facts and information, accumulating more structures onto the already existing knowledge structures" [accretion]; "The learner must devise new memory structures to interpret the material that is to be acquired" [restructuring]; restructuring is "the most difficult and the most significant form of learning, for it marks the acquisition of truly new conceptualizations about a topic matter."

Notably, in this original 1978 formulation, "restructuring" is defined as devising new memory structures — i.e., what the 2026 SEEK model separates out and relabels "schema formation" — via two sub-processes: patterned generation (building a new schema by analogy to an existing one, e.g. replacing a constant term with variables) and schema induction (building a new schema by combining recurring patterns across existing schemata). Per a secondary summary of the theory (InstructionalDesign.org, Tier 3-4):

"Restructuring involves the formation of new conceptual structures or schema. Structuring occurs much less frequently and requires considerable effort compared to accretion... Restructuring involves some form of reflection or insight (i.e., metacognition) and may correspond to a plateau in performance."

The same secondary source notes Rumelhart & Norman's 1981 extension added an explicitly analogical route to new-schema creation: "a new schema is created by modeling it on an existing schema and then modifying it based upon further experience" — a hybrid that sits between pure formation-from-nothing and pure modification-of-an-existing-structure.

Terminology caveat (important for reading the rest of the literature): because 1978-vintage "restructuring" denotes what is now more often called "schema formation," later papers using the bare word "restructuring" must be checked for which sense they intend — the field's own vocabulary shifted under the term.

Claim type check: Definitional/historical (Tier 3-4, acceptable, uncontested) for the taxonomy itself ✓. The effort/frequency specifics ("much less frequently," "considerable effort," "plateau in performance") are treated here as [unverified-mechanism — needs primary] since only a tertiary paraphrase was accessible this session, not the primary 1978/1981 text.

Claim 3 — The conceptual-change tradition draws a parallel distinction under the labels "weak" vs. "strong/radical" restructuring

Claim type: Historical/definitional claim → Tier 3–4 acceptable when uncontested (definitional core); technical-mechanism specifics flagged. Flag: [unverified-mechanism — needs primary] for the specific weak/strong definitions and difficulty claims — despite two independent attempts to fetch primary PDFs (Carey 2000; Vosniadou & Brewer 1987), no machine-extractable primary text was obtained this session.

Independently of the schema-theory lineage above, developmental psychology's conceptual-change literature (Susan Carey; Stella Vosniadou & William Brewer) converges on a structurally similar two-way split, using different vocabulary. Vosniadou & Brewer's paper is confirmed to exist and to be about exactly this territory — its CrossRef-mirrored abstract opens:

"This paper reviews recent work on knowledge acquisition with a focus on knowledge restructuring. This work suggests that it may be possible to replace Piagetian theories of global restructuring with a more domain-specific approach."

Per secondary description consistent across several independent tertiary sources (not independently confirmed against primary text this session): Carey's "weak restructuring" denotes new relations formed among existing concepts, while "strong restructuring" denotes the formation of genuinely new concepts not reducible to prior ones — described as considerably harder to achieve and, in some accounts, as requiring different learning mechanisms than ordinary belief revision. Vosniadou & Brewer's parallel "weak" vs. "radical" restructuring distinction is described the same way: weak = accumulation of new facts and relations among existing concepts; radical = change in the core concepts and structure themselves.

This maps onto the schema-formation/restructuring split as follows: weak/relational restructuring ≈ modern "restructuring" (reorganizing an existing structure); strong/radical restructuring ≈ modern "schema formation" (building structure that didn't exist before).

Claim type check: Definitional/historical naming — Tier 3-4 acceptable and treated as reasonably uncontested (the weak/strong distinction is widely cited under Carey's and Vosniadou & Brewer's names across independent tertiary descriptions). The specific effort/difficulty characterizations remain [unverified-mechanism — needs primary].

Claim 4 — Restructuring (in the modern, narrower sense) requires cognitive conflict as a trigger; congruent information cannot produce it

Claim type: Technical-mechanism claim → Tier 1–2 required. Met. Source: Danek & Flanagin (2019), AIMS Neuroscience 6(2):60-84. Direct access verified. (Already exact-quoted in this vault's 2026-06-28 "SeekVault self-rewiring" capture; reproduced here because it is directly relevant to the "conditions" half of this topic question.)

"conflict detection is a prerequisite of restructuring" ... "only the incongruent cues produce a cognitive conflict by forcing participants to question their initial problem representation" ... "cognitive conflict is needed in order to trigger a change in the problem representation."

This bears on conditions, not on the formation/restructuring distinction itself: it establishes that restructuring an existing schema needs a detected mismatch between incoming information and the current representation, not mere additional information.

Claim type check: Technical-mechanism — Tier 1, exact quote, verified. ✓

Claim 5 — Schema formation in infancy is triggered by "moderate discrepancy" between stimulus and existing schema — the same principle proposed for schema expansion

Claim type: Technical-mechanism claim → Tier 1–2 required. Met at Tier 2. Source: Zelazo (2024), Developmental Psychology 60(11):1978-1991. PMID 37917486. Verbatim abstract obtained directly via PubMed this session.

"This research and studies with older infants indicate that both the formation and the expansion of schemata occur through moderate discrepancies, a concept that Jerome Kagan promoted conceptually and through his research... Optimal attentiveness and positive affect occur at the peak of formation and to moderate discrepancies. Redundancy beyond the optimal level produces decreasing interest and positive affect and increasing negative affect resulting in boredom and avoidance."

This is the one primary/Tier-2 source located this session that speaks to conditions for schema formation specifically (as opposed to restructuring), and it is notable for what it does not find: it describes formation and expansion (a modification-like process) as governed by the same triggering principle — moderate discrepancy — rather than by categorically different conditions. This is developmental (infant/neonate) literature, not necessarily generalizable to adult conceptual restructuring, but it is a data point against a sharp qualitative-conditions split between the two processes, at least at this end of the literature.

Claim type check: Technical-mechanism — Tier 2, verbatim quote, verified. ✓

Claim 6 — Deep schema reorganization is global (not local) and executive/deliberate-processing-dependent

Claim type: Technical-mechanism claim → Tier 1–2 required. Flag: [unverified-mechanism — needs primary] — abstract reproduced via search engine; direct fetch returned 403. Already flagged identically in this vault's 2026-06-28 capture; reproduced here for context. Source: Kurashige, Kaneko & Matsumoto (2025), bioRxiv, DOI 10.64898/2025.12.30.697002.

"Schema accommodation is the reorganization of a preexisting memory schema to deeply accept new information incongruent with it. Although it is crucial for flexible intelligence, it occurs infrequently... the schema changes that occurred during the task were global reorganizations rather than local adjustments... the executive network responsible for deliberate processing played a central role."

This speaks to conditions for restructuring/accommodation of an existing schema specifically — global, infrequent, executive-network-dependent — but the paper (as reproduced) does not contrast this against the conditions for building a schema where none existed before, so it does not resolve the formation-vs-restructuring effort question either.

Claim type check: Technical-mechanism — Tier 1 venue, flagged for full-text confirmation. ✓ (with flag)

What remains unresolved

The literature surfaced this session supports a qualified yes on the core question — the distinction between building new schema structure and reorganizing existing structure is a recurring, formally named division across at least two independent lines of cognitive-science literature (schema theory: Rumelhart & Norman → SEEK model; conceptual change: Carey; Vosniadou & Brewer) spanning five decades. But the narrower question — do the two processes require different effort or different conditions from each other, specifically — is [unverified — could not confirm or deny after search]. Every source that addresses effort/conditions does so by contrasting the harder pair (formation + restructuring, or radical + weak restructuring taken together) against the easier pair (accretion + tuning, or ordinary learning), not by contrasting formation against restructuring directly. The one source that isolates formation on its own (Zelazo 2024, infant schema formation) suggests a shared triggering principle (moderate discrepancy) with expansion/modification rather than a categorically distinct one — which, if it generalizes beyond infancy, would argue against a sharp effort/conditions split between forming and restructuring, but this is a single developmental-psychology data point and should not be over-read.

Further leads

• Full text of Rumelhart & Norman (1978/1981) and Carey (1985/2000) could not be machine-extracted this session (scanned/compressed PDF renderings in both cases); a session with OCR tooling or clean HTML mirrors would let the effort/frequency and weak/strong claims move from Tier 3-4-with-flag to confirmed Tier 1-2.
• Vosniadou & Brewer (1987) full text — confirmed to exist and be on-topic via CrossRef metadata, but the actual weak/radical restructuring definitions were not independently verified against primary text; worth a dedicated follow-up capture given the paper's status as a foundational review.
• SEEK model full text (Xu, Sandhofer & Stigler, 2026) — both the publisher page (ScienceDirect) and the open-access mirror (eScholarship) returned errors on direct fetch this session. This is the single highest-priority follow-up: it is the only source that names schema formation and restructuring as siblings in one taxonomy, and its full discussion section likely contains the explicit comparison (or explicit non-comparison) of their respective effort/conditions that this capture could not resolve.
• Chi's "categorical shift" account of radical conceptual change (ontological recategorization) was not investigated this session but is a third, related line in the conceptual-change tradition worth a comparative capture alongside Carey and Vosniadou & Brewer.
• Piaget's assimilation/accommodation was checked briefly and appears not to draw a clean formation-vs-restructuring line the way the other traditions do — secondary descriptions state accommodation "requires the modification of existing schemas or the creation of new ones," treating both as outcomes of the same single process rather than as formally distinct mechanisms. This is worth confirming against primary Piaget text, since it would mean the formation/restructuring split is not universal across all schema theories, only across the specific lineages in Claims 1–3.