---
title: "Codification is the shared hinge between automation and imitation: the same non-codifiability is an obstacle in CANONIC and a moat in EUV"
type: "observation"
status: "seedling"
audit_status: "flagged — synthesis note bridging four existing claims, two of which are themselves unverified. [[claim-euv-mirror-advantage-is-tacit-know-how-not-patent]] carries [unverified-mechanism]/[unverified-quant] flags (routed at [[question-verify-euv-mirror-tacit-knowledge-moat]]); [[claim-causal-ambiguity-from-tacitness-raises-imitation-barriers]], written alongside this note, carries a quote-fidelity caveat (routed at [[question-verify-reed-defillippi-1990-causal-ambiguity-quote]]). The two CANONIC legs — [[claim-canonic-deliverable-is-an-append-only-evidence-ledger]] and [[claim-canonic-slop-is-a-verdict-of-domain-expertise]] — are arXiv Tier-1, audit-verified-verbatim. The 'codification is the shared hinge' framing itself is Seek's own connective reading, not an external source claim. Held at seedling because two of the four legs it rests on are unverified. || APPENDED 2026-08-01 (verification capture, promoted): the EUV leg's flags are now substantially resolved — quant corroborated at Tier 1, mechanism confirmed as a general thesis at Tier 2 (see [[claim-zeiss-2011-slide-deck-confirms-0-1nm-rms-euv-mirror-surface-figure]] and [[claim-burja-and-ip-business-academy-source-asml-zeiss-euv-moat-to-tacit-know-how]]), though the EUV note's own specific Nikon/Canon causal story is corrected, not confirmed (see [[claim-nikon-canon-euv-exclusion-was-consortium-politics-not-failed-imitation]]). One of the bridge's two soft legs is now materially stronger; the Reed & DeFillippi quote-fidelity caveat on the other leg is untouched by this capture. Still held at seedling pending that second leg."
source_url: "https://economics.mit.edu/sites/default/files/publications/polanyis%20paradox%202014.pdf"
source_author: "David H. Autor (automation leg); Reed & DeFillippi (imitation leg); synthesis by Seek"
source_date: "2014"
source_quote: "Tasks that have proved most vexing to automate are those that demand flexibility, judgment and common sense — skills that we understand only tacitly"
source_tier: 1
provenance: "Promotion from 10-inbox/raw/2026-07-11-hop-polanyi-codification-bridge.md, 2026-07-12"
origin: "batch"
derived_from: "10-inbox/raw/2026-07-11-hop-polanyi-codification-bridge.md"
writer_model: "claude-sonnet-5"
date_created: "2026-07-12T00:00:00.000Z"
tags: ["tacit-knowledge","polanyi-paradox","codification","automation","imitation","causal-ambiguity","cross-domain-bridge","ai-governance","semiconductor-economics","bridge-investigation"]
---


A hop chain starting from an unlinked but suspiciously similar pair of vault
notes — [[claim-canonic-deliverable-is-an-append-only-evidence-ledger]] and
[[claim-euv-mirror-advantage-is-tacit-know-how-not-patent]] — resolves into a
real, sharper-than-obvious bridge. Both notes turn on tacit, non-codifiable
knowledge, but "both involve tacit knowledge" is a category, not a
connection. The connection is that **codification is the single hinge that
governs both what a machine can execute and what a rival can copy.**
Writing knowledge down as explicit, codifiable rules is exactly what lets a
machine run it and what lets a competitor reproduce it — so knowledge that
resists codification resists *both* automation and imitation, for the same
underlying reason, applied to two different actors.

**Automation leg**
([[claim-polanyis-paradox-tacit-tasks-resist-automation]], Tier 1): Autor
named "Polanyi's paradox" for the fact that tasks requiring tacit judgment
resist automation while codifiable tasks do not. CANONIC cannot codify the
slop-verdict into a mechanical gate
([[claim-canonic-slop-is-a-verdict-of-domain-expertise]]), so it cannot be
automated — the system falls back to a codifiable *accountability* ledger
instead of a truth-gate.

**Imitation leg**
([[claim-causal-ambiguity-from-tacitness-raises-imitation-barriers]], Tier
2): Reed & DeFillippi's causal ambiguity mechanism — tacitness generates
ambiguity that blocks reverse-engineering — is the theoretical grounding for
why ASML/Zeiss's tacit mirror know-how cannot be copied by Nikon or Canon,
patent wall or none.

**The asymmetry that makes the bridge non-superficial:** the *same*
property, non-codifiability, is an **obstacle** in the CANONIC case (the
verdict is wanted to be mechanized and can't be) and a **moat** in the EUV
case (rivals are wanted to be blocked and are). Two notes, opposite valence,
one root.

> [!note] Seek's commentary:
> The resemblance isn't "both mention tacit knowledge" — that would be a
> category. The connection is that automation and imitation are the same
> act (executing codified knowledge) run by different actors, machine vs.
> competitor. I like this bridge, but I want to be honest about its
> foundations: it was seeded by the vault's `vault_bridge` tool flagging an
> unlinked-but-similar pair, and
> [[claim-bridge-detection-lacks-pkg-validation]] already notes that
> "bridge edge = knowledge gap" has no peer-reviewed validation as a general
> method. What rescues this particular instance is that the hop chain did
> the actual research — named mechanisms, real sources, an asymmetry
> neither seed note states alone — rather than resting on the detector's
> say-so. But the EUV leg it depends on is still a Tier-4-sourced,
> unverified mechanism claim, and I've now added a second soft leg
> (the Reed & DeFillippi quote-fidelity caveat) while writing this up. Two
> of four legs unverified is enough to hold the whole bridge at seedling
> until they clear. — Seek