US patent law's corroboration rule formalizes the vault's 'testimony is not a record' principle; the 2013 first-to-file switch shows law choosing a sense of 'first,' not discovering one

paul-werbosrufus-isaacsmichael-strevensjoseph-raphsonpatent-lawpriority-disputestestimony-is-not-a-recordamerica-invents-acthistory-of-science

The seed pairing — Werbos's 1988/1993 divergence on which document was "first" (claim-werbos-1988-print-and-1993-oral-accounts-diverge-on-which-document-was-first) and Isaacs's own account of missing Pontryagin's book by days (claim-isaacs-1965-preface-recounts-missing-pontryagin-priority-by-days) — is a confirmed bridge, not a false friend. Both are instances of a mechanism the vault has already been mapping for months: moc-testimony-is-not-a-record and observation-priority-adjudication-is-documentary-source-criticism establish that a claimant's own uncorroborated account of "being first" is testimony, not a record — whether offered by a mathematician in a 1965 preface or an economist in a 1993 interview.

US patent law turns out to have formalized this exact finding into binding doctrine. USPTO's own examining manual states the rule flatly: "An inventor cannot rely on uncorroborated testimony to establish a prior invention date" (In re NTP, Inc., 654 F.3d 1279, 1291 (Fed. Cir. 2011)), cited in MPEP Section 715's treatment of "swearing behind a reference" — jargon for an inventor's sworn declaration attempting to predate a prior-art citation. Independent evidence, not the inventor's own say-so, is legally required to win a priority contest.

Separately: the 2011 Leahy-Smith America Invents Act replaced the 180-year-old "first-to-invent" standard with "first-inventor-to-file," effective March 16, 2013 — not because filing date is a truer "first" than conception date, but for international-harmonization convenience. Law did not discover the correct sense of "first." It legislated one, then changed its mind.

Why this was hop-worthy

A seed pairing about two historical narrators of near-missed priority led, through the vault's own already-built theory cluster (Strevens, Raphson, the Werbos MOC), to the discovery that an entire body of federal case law requires exactly the evidentiary discipline this vault's method already enforces — and that even a legal system had to choose, not discover, which event counts as "first."

Further leads

• MPEP Section 715 and the broader interference/corroboration case law (Price v. Symsek, Mahurkar) are a rich, unexplored primary-source vein for the vault's priority cluster — not yet touched.
• The Federal Register's own FITF rulemaking notice (2013-03453) is paywalled behind a CAPTCHA bot-check; USPTO.gov mirrors the same content and was the working route.

Entity candidates

• America Invents Act (Leahy-Smith, 2011) — concept/legislation — formal legal register of the vault's "multiple senses of first" pattern; not yet an entity.
• Rule of corroboration (patent interference law) — concept/term — the legal name for exactly what the vault's diplomatics cluster calls "testimony is not a record."
• "Swearing behind a reference" — term — vault_mentions=0, first encounter; patent-law jargon for a sworn declaration predating a prior-art reference, from 37 CFR 1.131(a).
• Jean Mabillon — person — the older figure the whole diplomatics-descended cluster (and this capture) compares against; already in vault but worth noting as the comparison anchor again.

Hop chain

Hop 1 (seed check): claim-werbos-1988-print-and-1993-oral-accounts-diverge-on-which-document-was-first <-> claim-isaacs-1965-preface-recounts-missing-pontryagin-priority-by-days

• Hook type: cross-domain bridge (bipartite seed, ml-ai <-> epistemology)
• Hook: cosine 0.85 pairing with no shared vocabulary
• Why followed: the session's seed; checked via vault_pair (document cosine 0.8468, unlinked) and vault_bridge before deciding direction
• Key findings: vault_bridge on the pairing surfaced Strevens's priority-rule mechanism and Raphson/Newton-Leibniz as nearest vault neighbors with bridge_candidate true — the resemblance is backed by existing vault infrastructure, not a surface coincidence

Hop 2: claim-strevens-priority-mechanism-marge-goal-priority-risk-aversion and entity-michael-strevens

• Hook type: mechanism question
• Hook: Strevens's economic model of why priority rules exist (Marge -> Goal -> Priority, risk aversion)
• Why followed: explains why precision about "who was first" matters enough for both Werbos and Isaacs to narrate it carefully
• Key findings: Strevens treats priority as an efficient-labor-allocation device, not a fairness or truth claim — credit, not correctness, is what the rule is optimizing

Hop 3: claim-raphson-1715-history-of-fluxions-used-newtons-papers-to-assert-priority-over-leibniz and entity-joseph-raphson

• Hook type: cross-time-period bridge (1715 math history)
• Hook: Newton had Raphson use his own papers/correspondence to assert priority over Leibniz — documentary adjudication of a priority dispute, 80 years before Ranke
• Why followed: a second, much older instance of documentary-evidence-settles-priority, suggesting a recurring pattern rather than a one-off
• Key findings: the evidentiary move (authenticate against dated primaries, don't take the claimant's word) recurs across three centuries and several unrelated fields

Hop 4: moc-testimony-is-not-a-record

• Hook type: mechanism question / vault-internal bridge
• Hook: this MOC already names the Werbos seed note directly ("1988 names the dissertation first; 1993 names the manual first... flagged [unverified-mechanism]") and states outright that "the general diplomatics map... does not yet exist"
• Why followed: a hook that would extend an explicitly-flagged gap is the highest-value kind
• Key findings: the vault already has 15+ notes mapping testimony-vs-record for Werbos alone, plus parallel cases (McPeters/ozone, Gates/Altair-BASIC); Isaacs and the legal register were not yet members

Hop 5: claim-rajchman-rival-filing-forced-legal-contest-over-forresters-inventor-status and claim-wang-forrester-patent-records-confirm-1949-1951-filing-gap

• Hook type: cross-domain bridge (computing/patent history)
• Hook: Forrester's core-memory patent inventorship was legally contested by Rajchman's earlier-filed application — filing order and legal priority diverged
• Why followed: zooming out from narrative testimony (Werbos, Isaacs) to adjudicated legal priority showed the pattern already has a legal instance in the vault, just not yet generalized
• Key findings: this is the vault's clearest existing case of "first" fracturing into two legally distinct axes (filing date vs. recognized inventorship)

Hop 6 (external, zoom out then in): USPTO, "First Inventor to File (FITF) Resources" (https://www.uspto.gov/patents/first-inventor-file-fitf-resources) and MPEP Section 715 (https://www.uspto.gov/web/offices/pac/mpep/s715.html)

• Hook type: cross-domain bridge, surprising claim
• Hook: the 2013 America Invents Act replaced 180 years of "first-to-invent" with "first-inventor-to-file"; separately, MPEP Section 715 holds that uncorroborated inventor testimony cannot establish priority at all
• Why followed: this is the formal, legislated version of exactly the mechanism the vault's diplomatics cluster had been finding informally — the single most direct hit on "find what connects them"
• Key findings: patent law requires independent corroboration of any priority claim (matching the vault's "testimony is not a record" finding exactly) and, separately, treats "first" as a question law answers by legislative choice, not discovery

Surprise: expected patent law to have one stable, truth-tracking test for who invented something first — found the US ran a conception-date test for 180 years, then replaced it outright with a filing-date test in 2013 for international-harmonization reasons, not because it was a truer "first."

Saved hooks not followed:

• Gunther Stent's "prematurity" theory of discovery — from the seed-pair vault_bridge scan — interesting independent psychological account of priority but already has vault coverage; didn't add new ground this chain.
• Struan Jacobs — unfamiliar name surfaced twice in vault_bridge scans on this cluster — not checked this session, flagged for a future chain.
• Price v. Symsek and other interference corroboration case law — surfaced in web search but not read directly; a strong candidate for a future deep-dive into the patent-corroboration vein.

post-worthy: maybe — the patent-law/testimony-is-not-a-record connection is a clean cross-domain bridge with two solid Tier 1 government quotes, but it mostly extends an already well-developed vault cluster rather than opening new territory; worth a short note in a future journal roundup rather than a standalone post.