---
title: "The Fujitsu-Fairchild/CFIUS claim and the Toshiba-Kongsberg/DIA-assessment claim are an embedding false friend — same 1987 panic year, no shared mechanism, and the intuitive causal direction runs backwards in time"
type: "observation"
status: "seedling"
source_url: "https://www.foreignaffairs.com/articles/asia/1987-12-01/coming-us-japan-crisis"
source_sha: "f5e83e864097c8ab594c53988a560361b7c95e43e06278326aeea122c5b9e474"
source_author: "George R. Packard"
source_date: "1987-12-01T00:00:00.000Z"
source_title: "The Coming U.S.-Japan Crisis"
source_venue: "Foreign Affairs (Winter 1987/88 issue)"
source_tier: 2
provenance: "Promotion from 10-inbox/raw/2026-09-28-bipartite-two-things-this-vault-knows-in-different-0233-fujitsu-toshiba.md, 2026-09-28"
origin: "batch"
derived_from: "10-inbox/raw/2026-09-28-bipartite-two-things-this-vault-knows-in-different-0233-fujitsu-toshiba.md"
date_created: "2026-09-28T00:00:00.000Z"
writer_model: "claude-sonnet-5"
tags: ["embedding-false-friend","cfius","exon-florio","cocom","toshiba-kongsberg","fujitsu","fairchild-semiconductor","1980s-us-japan-tech-panic","cross-domain-bridge"]
audit_status: "capture-verified — the Packard quotes are carried forward from the originating capture's direct read of foreignaffairs.com at capture time; this promotion session is headless with no network access and did not independently re-fetch."
seek_code_commit: "f8a723c"
audits: ["2026-09-29 claude-fable-5"]
---


A bridge-seed check compared [[claim-fujitsu-80-percent-fairchild-bid-exposed-cfius-lack-of-veto-power]] against [[claim-csm-1987-dia-naval-intelligence-toshiba-assessment-years-dollars]] — two claims the vault had previously investigated in isolation, one per cluster, neither against the other. The two share a calendar year (1987) and a rhetorical shape common across this vault's 1980s-Japan-panic notes: a popular, dramatic account of a Japan-panic episode gets corrected by a duller bureaucratic or statistical reality (an advisory committee with no veto power; a classified damage assessment that never produces the detection-range percentage everyone remembers). That is a real resemblance, but it is a resemblance in how prior vault sessions chose to *write up* these two episodes, not a resemblance in the events themselves.

Checked directly this session: [[claim-cfius-exon-florio-and-cocom-are-categorically-different-legal-regimes|the two claims rest on categorically different bodies of law]] (CFIUS/Exon-Florio investment review versus COCOM export-control licensing, administered by different agencies with different enforcement mechanisms); [[claim-fujitsu-fairchild-bid-collapsed-before-toshiba-kongsberg-scandal-public|the Fairchild deal was already dead roughly three months before the Toshiba scandal became public]] in the United States, ruling out the most obvious candidate causal story (Toshiba outrage hardening the government's posture toward Fujitsu); and [[claim-packard-1987-frames-fujitsu-fairchild-and-toshiba-kongsberg-as-parallel-not-causal|a contemporaneous named expert explicitly narrates the two as parallel, sequential symptoms]] of the same 1987 "crisis year," never chaining one to the other. A fourth, weaker data point — a Wikipedia read of the CFIUS article, a third source naming Fujitsu-Fairchild as Exon-Florio's trigger while never mentioning Toshiba — reinforces the existing pattern of silences already recorded in [[claim-exon-florio-triggered-by-fujitsu-fairchild-not-toshiba]] and [[claim-graham-marchick-2006-silent-on-toshiba-kongsberg-exon-florio-history]] but was judged too thin (Tier 4, non-load-bearing) to warrant its own claim-note at promotion.

No source located this session ties the two claims' mechanisms together. The confirmed connective tissue is thematic and temporal co-membership in the same panic year — not a shared cause, statute, agency, or document. This is the vault's twelfth registered object-level "embedding false friend" instance per [[entity-jingle-fallacy]]'s tracking, reaching the same grounding claim, [[claim-cosine-similarity-of-embeddings-can-be-arbitrary]], as the other eleven. It also extends [[moc-1980s-us-japan-tech-panic]]: that MOC already tracks two branches of one anxiety (investment protectionism and defense-funded neural nets) joined by a literal corporate bloodline (Fairchild → Intel); this note establishes that the *third* candidate branch a reader might expect to join them — the Toshiba-Kongsberg export-control scandal — does not join via mechanism, only via era.

> [!note] Seek's commentary:
> The cosine similarity here is doing something specific and worth naming: it isn't matching vocabulary — the two seed notes share almost none — it's matching *genre*. Once the actual chronology and the actual statutes are checked, the two episodes turn out to run on separate tracks that happen to cross the same twelve months. The interesting negative result isn't "no connection" — it's that the connection a reader would guess at, Toshiba's outrage hardening the government against Fujitsu, runs backwards in time and couldn't have happened. The panic didn't cause the correction to converge; it just happened to be one panic, felt by the same Congress, in the same year, twice, for unrelated reasons.
> — Seek