---
title: "The Pontryagin naming-credit asymmetry is not structurally identical to the vault's backprop-attribution cluster — the mechanism is documented internal pressure, not mutually unaware literatures"
type: "claim"
status: "seedling"
source_url: "https://bibliotekanauki.pl/articles/970918.pdf"
source_author: "Hans Josef Pesch and Michael Plail"
source_date: 2009
source_venue: "The Maximum Principle of optimal control: A history of ingenious ideas and missed opportunities, Control and Cybernetics, vol. 38 (2009), no. 4A, pp. 973–995"
source_tier: 2
source_quote: "Boltyanskii claimed the version of the maximum principle as a necessary condition to be his own contribution and described how Pontryagin hampered his publication."
source_sha: "9c020a040ebe80a02bf0b7e5021b39868d245406c26fb8a3607acfa258cbd13f"
independent_roots: 2
provenance: "Promotion from 10-inbox/raw/2026-11-05-pontryagins-actual-co-discoverer-of-the-1956-maximum.md, 2026-11-05 (headless)"
origin: "batch"
derived_from: ["10-inbox/raw/2026-11-05-pontryagins-actual-co-discoverer-of-the-1956-maximum.md"]
date_created: "2026-11-05T00:00:00.000Z"
writer_model: "claude-sonnet-5"
audit_status: "capture-verified — the comparison rests on one new Tier-2 claim ([[claim-boltyanskii-testimony-locates-pontryagin-naming-convention-in-internal-soviet-pressure]]) and one existing verified-verbatim Tier-1 claim ([[claim-ml-and-ad-communities-mutually-unaware]]); no new primary fetch for this note itself. | 2026-11-06 cross-model audit (auditor claude-opus-5; writer claude-sonnet-5): the Pesch & Plail source was re-fetched via extract_pdf anyway, returning sha256 9c020a040ebe… identical to the capture-time source_sha, TLS verified; source_quote re-read in place at p. 985 and confirmed verbatim and contiguous. Both cited roots check out — [[claim-ml-and-ad-communities-mutually-unaware]] carries verified_verbatim against arXiv:1502.05767 and is Tier 1, so independent_roots: 2 is honest. One defect found and fixed. The note glossed [[claim-pesch-plail-2012-frames-cold-war-optimal-control-as-blue-red-corners]] as describing \"two traditions that never cited each other\" — stronger than that note's own careful framing (\"without documenting contact\", and its explicit \"Not independence, absence of documented contact\" section), and stronger still against Cali's 2026-10-13 review of that note, which flagged the source as documenting partial isolation: \"the flow of information, though not completely cut, was not as easy as before.\" The 2009 Pesch & Plail paper read for this audit says the same, noting that \"scientific publications became soon mutually known\" and that Bliss's 1946 monograph existed in Russian translation from 1950. Gloss narrowed to no *documented citation* contact, with Cali's correction carried inline so the overstatement cannot propagate again. The note's central claim — that the Pontryagin mechanism is intra-group pressure rather than mutual unawareness — is unaffected and HOLDS. Previous wording retained here for the record: \"two traditions that never cited each other\"."
tags: ["pontryagin","boltyanskii","backpropagation","automatic-differentiation","credit-attribution","priority-dispute","history-of-science","comparison"]
seek_code_commit: "d40ee39"
quote_sweep: "PASS 2026-11-06 — source_quote verbatim (normalized) in the capture-time archive (sha256 9c020a040ebe…), same-night model-free sweep"
verified_verbatim: "2026-11-06 — source_quote matched verbatim (normalized) against a direct fetch of source_url by seek_verify (no model involved)"
audits: ["2026-11-06 claude-opus-5"]
---


The vault's backpropagation-attribution cluster locates its credit gaps in a specific mechanism: independent discoverers working in separate fields whose literatures did not read each other, documented as the JMLR survey's own finding that "the fields of machine learning and AD have largely been unaware of each other" ([[claim-ml-and-ad-communities-mutually-unaware]]). No one discoverer in that cluster is shown pressuring another over naming; Linnainmaa's name receded through mutual silence, not negotiation.

The Pontryagin case shows a different, documented mechanism: an *intra-group* pressure, applied by a senior collaborator onto a junior one inside a single institute, resolved before publication — [[entity-v-g-boltyanskii|Boltyanskii]]'s own reported testimony that [[entity-lev-pontryagin|Pontryagin]] "hampered his publication" until he agreed to credit Pontryagin disproportionately and adopt the "Pontryagin maximum principle" name ([[claim-boltyanskii-testimony-locates-pontryagin-naming-convention-in-internal-soviet-pressure]]). The vault's existing "blue corner / red corner" framing of Cold War optimal control ([[claim-pesch-plail-2012-frames-cold-war-optimal-control-as-blue-red-corners]]) describes a structurally similar silo between the American and Soviet camps — two traditions between which that source documents no citation contact, though, as that note itself records, what the source describes is partial isolation rather than a sealed wall. Either way the silo explains the *East-West* split, not the *asymmetry within* the Soviet group itself, which is this claim's separate, internal mechanism.

The parallel a researcher might reach for — "forgotten co-discoverer" as one shape — covers two different causes here: mutual unawareness across a disciplinary wall (backprop/AD) versus a named, dated, internally pressured settlement (Pontryagin/Boltyanskii). Both produce the same visible symptom, one name surviving where several contributed; only one of the two has an agent who can be named as having done it on purpose.

> [!note] Seek's commentary:
> Two different shapes of the same word "forgotten." Linnainmaa wasn't pressured by anyone — nobody in the neural-network literature had read the numerical-analysis literature at all, a colder, more structural kind of erasure. Boltyanskii's name dropping off "the principle" has a name attached to the dropping, and a date, and his own later signature on the settlement. That's worse in one way — someone did it on purpose — and better in another: at least someone could, in principle, be asked why.
> — Seek