structure

Structure note — step 4 (4b) — COVID origins

generated: 2026-07-23 04:14 CEST

Counts

• H handed in (step-3 total): 40 (H-1…H-40)
• active after 4b: 10 — the cluster members (5 from papers + 5 minted here)
• merged: 28 (4a’s 4 + 4b’s 24)
• dropped (stragglers): 7
• clusters: 3
• minted at 4b: 5 H (3 constructed members H-41/H-42/H-43, 2 residuals H-44/H-45) + 3 HC
• total H files on disk: 45 (nothing deleted; ≥ 40)

Cluster inventory (the top-level map)

1. HC-1 — Origin of SARS-CoV-2 (natural zoonosis vs research-related incident). The main question. Members: H-41 natural zoonotic spillover, no research · H-42 research-related leak of an unmodified natural virus (zoonotic collection) · H-43 research-related origin of a laboratory-manipulated virus (engineered lab leak) · H-44 residual. All object-level empirical evidence (genome, geography, phylodynamics, institutional) attaches here.
2. HC-2 — Number of independent introductions of SARS-CoV-2 into humans. Members: H-5 two or more separate introductions · H-16 a single introduction. exhaustive_by_construction: true (a count of 1 vs ≥2; no residual). depends_on HC-1.
3. HC-3 — Causal role of the Huanan Seafood Market. Members: H-14 primary wildlife-to-human spillover origin · H-15 externally-seeded amplifier (infected people or cold-chain goods) · H-19 detection/ascertainment artifact, not causally special · H-45 residual. depends_on HC-1 and HC-2.

Member composition (what each member absorbed)

• H-41 (natural zoonosis, constructed): H-4, H-12 (+H-13), H-28 (natural-origin verdicts); H-31, H-32, H-33, H-36, H-37, H-38 (natural-genome / natural-adaptation object-level claims).
• H-42 (natural-virus lab leak, constructed): H-3 (Rootclaim zoonotic-collection).
• H-43 (engineered lab leak, constructed): H-1, H-2, H-17, H-27 (engineered/lab verdicts); H-22, H-34 (+H-35), H-39, H-40 (engineering-mechanism claims).
• H-5 (≥2 introductions): absorbs H-6 (intermediate haplotypes did not genuinely circulate — a premise about the count, not a separate answer).
• H-14 (market spillover origin): 4a had absorbed H-21, H-24; 4b adds H-10 (agnostic epicenter — logical disjunction of the spillover/amplifier answers), H-23 (raccoon-dog intermediate host), H-25 (not pangolin/bat), H-26 (viable wildlife route).
• H-15 (amplifier): absorbs H-20 (cold-chain introduction — a second external seed route).

Verdict / testimony resolution (trap 2)

The bottom-line conclusions were merged into the object-level origin member each asserts, never placed as co-members (that would double-count and break MECE): Rootclaim’s four scenarios → H-1/H-2→H-43, H-3→H-42, H-4→H-41; Judge Eric H-12/H-13 → H-41; Andersen H-31 → H-41; Weissman synthesis H-17 → H-43; FBI/DOE H-27 → H-43; IC-plurality H-28 → H-41. Provenance is preserved in each survivor’s additional_sources, and 53 arguments had their affects_hypotheses repointed from the absorbed nodes to the survivors (e.g. the IC-split argument A-75 now bears on both H-43 and H-41).

Genome “engineered vs natural” is NOT a separate cluster (trap 1)

Carving H-31/32/33/34/36/37/38/39/40 as their own cluster would re-partition HC-1 (engineered ≈ H-43, natural ≈ H-41∪H-42) and double-count the genome evidence. Instead the engineered claims merged into H-43 and the natural ones into H-41; the genome observations (O-58…O-77, O-34…O-38) discriminate the engineered member from the natural members inside HC-1 at step 5. H-31/36/37/38 also support H-42 (shared natural genome); because step 5 scores each observation per-member, their bookkeeping home (H-41) does not deny H-42 that support.

Cross-cluster dependence (depends_on)

• HC-2 → HC-1: two-or-more introductions essentially requires a repeatable zoonotic source (very unlikely under one leak), so it is strongly correlated with H-41; a single introduction is neutral. Kept separate (distinct phylodynamic evidence) rather than merged.
• HC-3 → HC-1: a primary wildlife spillover at the market (H-14) essentially entails natural zoonosis; the “artifact” answer (H-19) is the position associated with a non-market/research origin. Strong correlation, kept separate (distinct geospatial/environmental/wildlife evidence).
• HC-3 → HC-2: under zoonosis both inferred introductions are placed at the market, so H-14 co-varies with the ≥2-introductions answer; mediated by the shared zoonosis picture.

HC-1’s depends_on is empty: it is the root question, and the reverse dependences are recorded on the dependent clusters per the brief.

Stragglers dropped as “clusters nowhere” (traps 3 & 4)

Seven hypotheses assert no distinct MECE answer to any cluster’s sub-question and were dropped as stragglers; each is on-topic base-rate / precedent / timing input whose evidential content lives in observations that attach to the clusters directly at step 5 (not off-topic):

• H-18 — numeric prior/base-rate parameter for a research origin; feeds step 7’s origin prior, not a member.
• H-7 — cryptic-circulation timing (no substantial pre-December-2019 circulation); evidence O-6/O-7.
• H-11 — early-transmission fact (H2H by mid-December 2019); evidence O-7/O-9/O-10.
• H-8, H-9 — Dutch-mink host-competence / spillback precedent (a different host system); evidence O-11…O-15.
• H-29, H-30 — SARS-civet market-amplification / host-role precedent (a different virus); evidence O-54…O-57.

Deviations from the brief’s illustrative carving

• HC-3 members. The brief illustrated “H-14 vs H-15 vs H-10 (unspecified epicenter)“. H-10 is logically the disjunction of the spillover and amplifier answers, so it cannot be a mutually-exclusive co-member; it is merged into H-14. The genuine third answer is H-19 (the market is a detection artifact — the lab-leak-side stance), which is promoted to a member, giving a cleaner MECE partition {spillover-origin, amplifier, artifact, residual}.
• Extending 4a’s merges (H-10, H-23, H-25, H-26 → H-14; H-20 → H-15). 4a’s note on H-14 had deliberately left H-23/H-10 unmerged on the dedup discriminability test (correct for dedup). For MECE clustering these are sub-claims of the wildlife-spillover answer, not separate answers; a standalone intermediate-host cluster would double-count the species-correlation evidence (O-40/O-50/O-51). They are therefore folded into H-14, and that evidence discriminates H-14 vs H-15 inside HC-3.