Cognitive Substrate ↔ NSM — Comparison and Bridge Analysis

Part I — Comparison

The two primitive sets side by side

Same name "Evaluation" / "Evaluative" on both sides, and it does the same structural work in both. But the rest is not a direct alignment, and that's the structural finding.

What they're actually about (the level-of-description distinction)

The two analyses are at different levels of description of the same broader phenomenon — they describe complementary aspects, not the same thing at different resolutions. (Compare with the NSM-internal multi-resolution view: ~7 substrate / ~16 surface / ~65 sub-level all describe NSM at different zoom levels of the same analytical surface. The cognitive-substrate / NSM relationship is different — they're at different layers of the realization chain, not different zooms on the same layer.)

Cognitive substrate answers: what operations does cognition perform? Its primitives are information-processing operations — representing, categorizing, associating, sequencing, symbolizing, evaluating. These are the OPERATIONS the substrate runs.

NSM answers: what semantic content does cognition produce? Its primitives are content-shaped — references, mental predicates (as content, not as operations), actions, states, spatiotemporal positioning, modals, evaluations. These are the OUTPUTS the substrate emits as meaning units.

The user's framing — "ours fits into the information substrate, theirs is focused more on language" — is exactly right. The cognitive substrate is at the information-processing layer; NSM is at the semantic-content layer. The two are connected by a bridge (the W2 analysis below).

Where alignments DO appear

Several primitives have natural couplings across the two sets — these are the bridge edges:

Cognitive Ev ↔ NSM Ev. Both are evaluation. The cognitive substrate's Ev is the operation of assessing representations against goals; NSM's Ev is the content of evaluative judgments (GOOD, BAD, KIND OF, LIKE). Same primitive name, different aspect: operation on one side, content on the other. The cognitive substrate's Ev produces NSM's Ev as semantic output.

Cognitive Sy ↔ NSM Rf, Mn, all primes. Cognition's Symbolization (the arbitrary sign-meaning mapping operation) is the source of NSM primes. When cognitive Sy operates at Sy3+ (arbitrary symbols), it produces lexicalizations that get fixed cross-linguistically as NSM-like primes. NSM primes ARE the content that cognitive Sy produces.

Cognitive Ct ↔ NSM Rf (especially substantives/determiners/quantifiers). Cognitive Categorization (operation of grouping representations) produces NSM's Reference-shaped content (categories that referents fall into). NSM's "KIND OF" and "PART OF" are the semantic content of cognitive Ct's outputs.

Cognitive As ↔ NSM Md (logical connectives). Cognitive Association (operation of connecting representations by relations) produces NSM's Modal-shaped content (BECAUSE, IF, LIKE, MORE). NSM's logical connectives are the semantic content of cognitive As's outputs.

Cognitive Sq ↔ NSM Sp (time portion). Cognitive Sequence (operation of temporal ordering) produces NSM's temporal primitives (WHEN, NOW, BEFORE, AFTER). Sequence-as-operation on the cognitive side; sequence-as-content on the NSM side.

Cognitive Rp ↔ NSM Mn, Ac, St (most others). Representation (the most general operation) produces most NSM content categories: mental predicates (the cognitive substrate represents mental states), action descriptions (it represents events), state descriptions (it represents persistent conditions). NSM's substantive-rich vocabulary derives from cognitive Rp.

Where alignments DON'T appear (or where they get awkward)

NSM has no clean equivalent to cognitive Sy as an operation. NSM treats symbols as primitive units; it doesn't have a primitive for "the operation of making something a symbol." This is reasonable for NSM (which is about content, not operations) but it means the cognitive substrate's Sy is invisible from NSM's level — Sy is "behind" the NSM primes, producing them but not appearing as one.

Cognitive substrate has no clean equivalent to NSM Md. Modal/logical primitives in NSM (NOT, MAYBE, CAN, IF) emerge from a high-level composition of cognitive primitives (Rp at Full level + As at Full level + Sy at recursive level), but they're not a separate cognitive operation. NSM treats them as primitive because at the semantic-content level they are; the cognitive substrate treats them as emergent because at the information-processing level they decompose.

NSM lacks the operation/content distinction cognitive substrate has. The cognitive substrate cleanly distinguishes the operation of representing (Rp) from the content being represented. NSM doesn't make this distinction because its primitives ARE content. This means NSM is "flatter" — it has fewer ontological categories at its level of analysis.

Verdict — parallel structures at different layers

The cognitive substrate and NSM are parallel structures at adjacent layers of the realization chain. The cognitive substrate is the information-processing layer; NSM is the semantic-content layer immediately above. The bridge between them is the W2 analysis in Part II.

The two primitive sets do not conflict. They describe different aspects of the same broader phenomenon (cognition + meaning). The differences in primitive count (6 vs 7), in core triad ({Rp, Ct, Sy} vs {Rf, Mn, Ev}), and in which primitives appear as "primitive" vs "emergent" are exactly what the methodology's level-relativity discipline (M2/R12) predicts when comparing analyses at adjacent layers.

This is an instance of the methodology's structural pattern: adjacent layers see different primitive sets, and the differences are diagnostic. The cognitive substrate ↔ NSM relationship parallels:

• Biology substrate ↔ organism architecture (substrate operations vs functional capabilities)
• Entity-system substrate ↔ application-architecture (substrate primitives vs feature space)
• Neural hardware ↔ cognitive substrate (neural operations vs cognitive operations)

The bridge between cognitive substrate and NSM should look structurally similar to those other bridges.

Part II — W2 Bridge Analysis

What the bridge is

A bridge between two layers extracts the machinery by which the lower layer's outputs become the upper layer's inputs. For cognitive substrate ↔ NSM:

• Lower layer (source): cognitive substrate, with its six primitives and their outputs (categorized representations, learned associations, sequenced events, symbolic encodings, evaluated content).
• Upper layer (target): NSM, with its seven substrate-level primitives organized as semantic content units (references, mental predicates, actions, states, spatiotemporal positioning, modals, evaluations).
• Bridge primitives: the machinery that turns cognitive substrate outputs into NSM-shaped semantic content units.

Step 1 — Information gathering

Sources for the bridge analysis:

• Wierzbicka's empirical work on how NSM primes are tested and added/removed (cross-linguistic translation as the validation).
• Cognitive linguistics literature on lexicalization, semantic universals, the relationship between conceptual structure and lexical structure (Talmy, Jackendoff, Langacker).
• Developmental psychology on how children acquire NSM-like vocabulary (Bowerman, Choi, Slobin on cross-linguistic acquisition).
• The cognition arrangement's existing bridges, especially analysis-neural-to-cognitive-bridge.md, as the closest structural precedent.

Step 1b — Domain type declaration

Bridge domain. R11 prediction: ~12 bridge primitives, hub structure, ~25-40% filter stringency (bridges typically fall in this band).

Step 1c — Level of description declaration

Level: the lexicalization/crystallization layer between cognitive operations and stable semantic-content units.

Floor: the candidate bridge primitives (~12 expected per R11). These are the operations that translate cognitive substrate outputs into NSM-prime-shaped content.

Background: the cognitive substrate's six primitives (treated as input source); NSM's seven substrate-level primitives (treated as output specification); the cultural-transmission mechanisms that propagate primes (treated as downstream).

Step 2 — Landscape analysis

What kinds of "cognitive output → semantic content" mechanisms are there?

• Word learning in childhood — how a representation (Rp) + category (Ct) gets a lexical handle (a candidate NSM-like prime).
• Cross-linguistic empirical iteration — how candidate primes survive or fail when translated.
• Cognitive frame-building (Fillmore) — how event structure gets organized into lexicalizable units.
• Semantic-role assignment (theta-role theory) — how participants in events get mapped to argument positions.
• Metaphor extension (Lakoff-Johnson) — how primes' meaning extends to non-literal uses.
• Cultural conventionalization — how a candidate prime stabilizes across a speech community.

Each of these is a candidate bridge mechanism. The methodology's job is to extract the irreducible primitives among them.

Step 3 — Bridge primitive extraction

Candidate bridge primitives (first-pass — needs iteration):

1. Lexicalization (Lx) — the operation of attaching a stable arbitrary sign to a cognitive output (category, representation, or associative pattern). The cognitive substrate's Sy primitive at Sy3+ provides the capacity; Lx is the bridge primitive that exercises the capacity to produce a candidate semantic-content unit. Levels: Lx0 = no lexical handle; Lx1 = ad-hoc tag; Lx2 = stable individual lexicalization; Lx3 = community-shared lexicalization; Lx4 = cross-cultural stabilization.

2. Reification (Rf) — the operation of treating a cognitive output as a discrete thing that can be referred to. Cognitive Rp represents states; Rf packages a representation into a referenceable unit. The bridge between cognitive representation-as-state and NSM Reference-as-thing. Levels: Rf0 = no reification; Rf1 = perceptual entity; Rf2 = abstract entity; Rf3 = referent of discourse; Rf4 = formal object.

3. Conventionalization (Cv) — the cross-individual stabilization that turns an individual's lexicalization into a community-shared form. Without Cv, every person has private vocabulary; with Cv, lexical mappings are shared and meanings can be jointly accessed. Levels: Cv0 = private; Cv1 = paired; Cv2 = small-group; Cv3 = community-wide; Cv4 = cross-community.

4. Universalization (Un) — the further cross-community stabilization that makes a lexicalization show up in every language. This is what makes a candidate prime an NSM prime (per Wierzbicka's empirical claim). Levels: Un0 = absent; Un1 = found in some languages; Un2 = found in most languages; Un3 = found in all studied languages; Un4 = required for any communication (the universal-prime claim).

5. Combinability (Cb) — the grammatical capacity to combine primes productively into composite meanings. Cognitive Sq + Sy at Sy4+ provides the capacity; Cb is the bridge primitive that produces grammatically combinable semantic units. Levels: Cb0 = no combination; Cb1 = parataxis (lists); Cb2 = predication; Cb3 = embedding; Cb4 = recursive (NSM's paraphrase grammar).

6. Reference resolution (Rr) — the operation that maps a symbolic handle (lexicalization output) back to a substrate-level representation. NSM-prime use in context requires resolving what specifically the prime refers to. Levels: Rr0 = no resolution; Rr1 = deictic; Rr2 = anaphoric; Rr3 = constructed-context; Rr4 = full discourse.

7. Translation (Tr) — the operation of mapping between lexicalizations across different languages while preserving the underlying semantic content. NSM's discipline of cross-linguistic translation tests this. Levels: Tr0 = no cross-language equivalence; Tr1 = approximate; Tr2 = paraphrasable; Tr3 = paraphrase-preserving (the NSM standard); Tr4 = lossless.

8. Decomposability (Dc) — the bridge's ability to break complex concepts into prime-level paraphrases. The reverse of combinability. NSM's paraphrase discipline IS decomposability. Levels: Dc0 = atomic only; Dc1 = approximate decomposition; Dc2 = paraphrase using non-primes; Dc3 = full prime-only paraphrase; Dc4 = recursive decomposition to primes.

9. Crystallization (Cr) — the phase transition by which a candidate prime becomes irreducible at the NSM level. Distinct from conventionalization (Cv) and universalization (Un): crystallization is the structural stabilization where the prime can no longer be reduced to other primes' combinations within the metalanguage. Levels: Cr0 = reducible; Cr1 = strongly preferred; Cr2 = entrenched; Cr3 = irreducible-in-tested-languages; Cr4 = irreducible-in-principle.

10. Type abstraction (Ta) — the bridge primitive that turns specific cognitive categories into prime-shaped types. Cognitive Ct produces categories; Ta packages them as types that NSM treats as referenceable. Levels: Ta0 = concrete only; Ta1 = ad-hoc abstraction; Ta2 = formal categories; Ta3 = type-of-type; Ta4 = recursive abstraction.

11. Modal projection (Mp) — the bridge primitive that turns cognitive As (associations) into NSM Md (modal/logical content). Specifically: cognitive associations of cause-effect get projected as BECAUSE; conditional associations as IF; possibility associations as MAYBE. Levels: Mp0 = no modal content; Mp1 = epistemic only; Mp2 = causal; Mp3 = counterfactual; Mp4 = full modal logic.

12. Speech-act packaging (Sp) — the bridge primitive that turns cognitive outputs into utterance-ready packages. Cognitive substrate produces internal content; speech-act packaging produces externally-utterable content with illocutionary force (statement, question, command). Levels: Sp0 = internal only; Sp1 = vocalized; Sp2 = informative; Sp3 = directive; Sp4 = performative.

Count check: 12 bridge primitives. Matches R11's prediction for bridge domains (~10-12).

Step 4 — Bridge dependencies

Lx → Sy (cognitive symbolization capacity)
Rf → Rp (cognitive representation)
Cv → Lx (need lexicalizations to conventionalize)
Un → Cv (need conventionalization before universalization)
Cb → Sy + Sq (need symbols + sequence capacity)
Rr → Rf + Rp (resolve reference to a representation)
Tr → Cv + Dc (need shared lexicalization + decomposability)
Dc → Cb (decomposition requires composition rules)
Cr → Lx + Cv + Un + Dc (crystallization is the conjunction)
Ta → Ct + Lx (turn category into typed lexicalization)
Mp → As + Lx (turn association into modal expression)
Sp → Cv (need community-shared lexicalization for speech acts)

Hub: Lx (Lexicalization) — depends only on cognitive Sy; many others depend on Lx (Cv, Cr, Ta, Mp, Sp). Lexicalization is the bridge's root.

Step 5–6 — Bridge pair-relationships (first-pass)

C(12,2) = 66 pairs. Selective load classification of the heaviest:

~10 heavy pairs out of ~30 selectively classified ≈ in the 25-40% load-bearing range expected for bridges.

Step 7 — Coherent sub-lattice (informal)

The dependency structure prunes substantially: Lx is the root, several primitives chain from it (Cv → Un → Cr in series), and many primitives presuppose multiple inputs (Cr requires four predecessors). Estimated filter stringency: ~15-25%, in the bridge band.

Step 9 — Bridge core triad candidates

{Lx, Cv, Un} — the lexicalization → community-stabilization → universalization chain. This is the operational core of how NSM primes get produced. Without Lx, no symbols. Without Cv, no shared symbols. Without Un, no cross-cultural prime candidates.

{Cb, Dc, Tr} — the paraphrase-discipline triangle. Combination + Decomposition + Translation is the NSM methodology's empirical surface. Without any one, NSM's discipline collapses.

Hub candidate: Lx (six heavy pairs).

Step 10 — Bridge emergent properties (with M3 derivations)

All nine derive cleanly from the bridge primitives. M3 passes for the bridge.

Step 11 — Cross-bridge pattern observation

Comparing this bridge to existing bridges in the corpus:

• Biology-to-organism bridge has ~12 developmental mechanisms.
• Neural-to-cognitive bridge has ~12 bridge primitives (per analysis-neural-to-cognitive-bridge.md).
• Entity-system substrate-to-application-architecture has 11 substrate-bridge extensions.
• Cognitive-substrate-to-NSM (this analysis) has 12 bridge primitives.

The ~10-12 bridge-primitive count is consistent across substrate→surface and substrate→content bridges. This is a recurrent pattern.

The bridge's hub primitive (Lx) parallels biology's Cell (Cl) as bridge hub, the entity system's Type (T) as substrate-bridge hub. Hubs at bridges tend to be the packaging operation — the primitive that turns substrate output into surface-consumable units.

Step 12 — Literature alignment

NSM literature is the upper-layer source. Wierzbicka and Goddard's work IS the empirical surface for Cv, Un, Cr, and Tr — they're the discipline's instruments. Cognitive linguistics (Talmy, Jackendoff, Langacker, Fillmore) provides Lx, Rf, Ta, Mp content. Developmental psychology (Bowerman, Choi, Slobin) provides the acquisition trajectory.

The bridge as proposed here has not been formalized in any single literature. NSM treats its primes as primitive; cognitive linguistics treats lexicalization as a research subject; developmental psychology studies prime acquisition. The methodology's contribution is naming the bridge as a structural unit and extracting its irreducible primitives.

Part III — Findings worth recording

1. The cognitive substrate and NSM are parallel structures at adjacent layers. They do not conflict; they describe complementary aspects (information-processing operations vs semantic content). The user's framing is confirmed.

2. The bridge exists and is analyzable. ~12 bridge primitives extracted, structurally similar to the biology-to-organism and neural-to-cognitive bridges. Hub: Lexicalization (Lx). Core triads: {Lx, Cv, Un} (the prime-production chain) and {Cb, Dc, Tr} (the NSM-discipline triangle).

3. NSM's empirical discipline IS the bridge's evaluator. Cv, Un, Cr, Tr are the operations Wierzbicka's empirical method uses. NSM treats them as methodology; the bridge analysis treats them as bridge primitives. The convergence is informative — what NSM calls "testing" is structurally the bridge's operational behavior.

4. The realization chain is sharpened. Adding the cognitive-substrate → NSM bridge to the cognition arrangement gives a fuller picture:

neural hardware
  ↓ (neural-to-cognitive bridge, ~12 primitives)
cognitive substrate (Rp, Ct, As, Sq, Sy, Ev)
  ↓ (cognitive-to-NSM bridge, ~12 primitives) [NEW]
NSM substrate-level (Rf, Mn, Ac, St, Sp, Md, Ev)
  ↓ (NSM-to-cultural-ecosystem bridge — not yet analyzed)
cultural ecosystem

The chain now has explicit bridges at every step, and the semantic-content layer (NSM) is positioned as a substrate of its own right rather than as "the surface of cognitive substrate" (which was the prior framing).

5. M3 passes on a third domain. The bridge analysis's emergent properties derive cleanly from the 12 bridge primitives. M3 now has three independent confirmations (entity system, biology, NSM) plus a fourth on the bridge.

6. The methodology's level-relativity discipline (M2/R12) is empirically valuable. Without explicit level-of-description declaration, the cognitive substrate ↔ NSM comparison would have looked like a conflict (different primitive counts, different core triads). With R12, the comparison is clear: they're at adjacent layers, the differences are diagnostic, and the bridge is the natural follow-up. This is concrete evidence that M2 was worth adding.

7. A finding for the NSM tradition. The bridge analysis suggests that what NSM calls "prime status" (Wierzbicka's universality + irreducibility claim) is actually the joint output of four bridge primitives: Cv (conventionalization), Un (universalization), Cr (crystallization), and Dc-failure (decomposition fails to reduce the prime to others). NSM's empirical discipline is implicitly running this conjunction; making it explicit would clarify what the universality claim actually requires.

Part IV — Open questions

1. NSM-to-cultural-ecosystem bridge. The chain is incomplete on the upward side. The cultural ecosystem (analysis-cultural-ecosystem.md, primitives Cd/Tr/...) consumes NSM content. What's the bridge?

2. The W2 bridge as written is first-pass. The 12 bridge primitives are candidates; the 3/3b iteration loop hasn't been run. Some primitives may merge or split.

3. The Tr (translation) primitive overlaps with Cv (conventionalization) in some readings. Cross-linguistic stabilization is what makes a lexicalization a prime; intra-linguistic stabilization is what makes a lexicalization a community word. These may be levels of the same primitive rather than two primitives.

4. The relationship between Cr (crystallization) and Un (universalization). Crystallization is structural irreducibility; universalization is cross-linguistic presence. Are they independent, or is one a consequence of the other? Empirical question.

5. Connecting back to Paper 11. When Paper 11 is updated (deferred), the cognition arrangement's full chain (now with NSM and the bridge) is a clean cross-domain example for the methodology's range. Worth including.

6. A note on the "Sy" name collision. The cognitive substrate's Sy = Symbolization (the operation). The bridge's Sp = Speech-act packaging (separate primitive). NSM's Rf = Reference (another primitive). Several "S"-named primitives across the arrangement; worth verifying no semantic collision causes confusion downstream.