Skip to content

Procedure-Work Mismatch

Prime #
1088
Origin domain
Organizational & Management Science
Subdomain
safety and reliability → Organizational & Management Science

Core Idea

Procedure-work mismatch is the structural pattern in which a system's prescribed model of how work happens — formal procedures, standard operating procedures, regulations, checklists, runbooks, role definitions — diverges systematically from how work is actually performed under real conditions. The prescription describes a clean, deterministic workflow against an idealised constraint set; the enactment confronts variable inputs, unreliable equipment, time pressure, exceptional cases, and the unwritten know-how operators have developed precisely because the prescription does not survive contact with reality. Hollnagel's "work-as-imagined versus work-as-done" is the canonical articulation, but the pattern recurs whenever a system uses an explicit normative artefact to govern a class of work and the artefact is built at a different granularity than the work itself.

The signature commitment is layered representation: there is the work, and there is a separate, authoritative description of the work that the institution has invested with control authority — and the two slowly diverge. The divergence is not necessarily anyone's fault: procedures lag changes in equipment, inputs, regulation, and demand; the people doing the work patch the gap through informal adaptation; the institution continues to audit, train, certify, and litigate against the formal model. The two views coexist for a long time, frequently without trouble, until an incident, an audit, or a workforce change exposes the gap.

A subtle structural fact is that the gap is usually load-bearing. Many systems work because operators silently bridge the prescription-enactment gap, so suppressing the informal adaptation without fixing the procedure typically degrades performance — the "work-to-rule" slowdown is exactly the experiment of removing the gap-bridging adaptation. Procedure-work mismatch is therefore not simply a defect but a feature of how layered normative systems function, and the question is how it is managed, not whether it can be eliminated.

How would you explain it like I'm…

Grandma's Real Recipe

The recipe says cook for ten minutes, but Grandma knows the real oven runs hot, so she takes the cookies out early and they come out perfect. The written rule and what you really have to do aren't quite the same. People who do the job every day quietly fix the gap so things still work.

Rules Versus Real Work

Procedure-Work Mismatch is when the official instructions for a job — the rules, checklists, and step-by-step guides — don't match how the job actually gets done in real life. The instructions imagine perfect tools, perfect inputs, and no surprises, but real work has broken equipment, weird cases, and time pressure. The people doing the job fill the gap using know-how that isn't written down anywhere. Often nobody is to blame — the rules just fall behind as tools and demands change. The surprising part is the gap is usually helpful: many systems only work because workers quietly bridge it, which is why 'working strictly to rule' on purpose actually slows everything down.

Work-as-Imagined vs. Work-as-Done

Procedure-Work Mismatch is the structural pattern in which a system's prescribed model of how work happens — formal procedures, regulations, checklists, runbooks, role definitions — diverges systematically from how the work is actually performed under real conditions. The prescription describes a clean, deterministic workflow against an idealized set of constraints; the enactment confronts variable inputs, unreliable equipment, time pressure, exceptional cases, and the unwritten know-how operators developed precisely because the prescription doesn't survive contact with reality. The signature commitment is a layered representation: there's the work, and there's a separate authoritative description of the work that the institution has invested with control authority, and the two slowly drift apart. Crucially the divergence is usually nobody's fault — procedures lag changes in equipment and demand while workers patch the gap informally. And the gap is usually load-bearing: many systems work because operators silently bridge it, so suppressing the informal adaptation without fixing the procedure degrades performance, which is exactly what a 'work-to-rule' slowdown demonstrates.

 

Procedure-Work Mismatch is the structural pattern in which a system's prescribed model of how work happens — formal procedures, standard operating procedures, regulations, checklists, runbooks, role definitions — diverges systematically from how work is actually performed under real conditions. The prescription describes a clean, deterministic workflow against an idealized constraint set; the enactment confronts variable inputs, unreliable equipment, time pressure, exceptional cases, and the unwritten know-how operators have developed precisely because the prescription does not survive contact with reality. Hollnagel's 'work-as-imagined versus work-as-done' is the canonical articulation, but the pattern recurs whenever a system uses an explicit normative artefact to govern a class of work and that artefact is built at a different granularity than the work itself. The signature commitment is layered representation: there is the work, and there is a separate, authoritative description of the work that the institution has invested with control authority, and the two slowly diverge. The divergence is not necessarily anyone's fault — procedures lag changes in equipment, inputs, regulation, and demand; workers patch the gap through informal adaptation; the institution keeps auditing, training, certifying, and litigating against the formal model, so the two views coexist for a long time, often without trouble, until an incident, audit, or workforce change exposes the gap. A subtle structural fact is that the gap is usually load-bearing: many systems work because operators silently bridge the prescription-enactment gap, so suppressing the informal adaptation without fixing the procedure typically degrades performance — the work-to-rule slowdown is precisely the experiment of removing the gap-bridging adaptation. The mismatch is therefore not simply a defect but a feature of how layered normative systems function, and the question is how it is managed, not whether it can be eliminated.

Structural Signature

a prescribed model of how work happens — an enacted practice of how work is actually done — a control authority invested in the prescription — a systematic divergence between the two layers — a load-bearing gap bridged by tacit adaptation — a brittleness invariant: deceptive stability with fragility to anything disrupting the informal layer

The pattern is present when each of the following holds:

  • A prescriptive artefact. An explicit normative description of the work — procedure, checklist, regulation, runbook, role definition — built at some granularity and idealized constraint set.
  • An enacted practice. The actual conduct of the work under real conditions: variable inputs, time pressure, exceptional cases, and the unwritten know-how operators develop.
  • Invested authority. The institution has given the prescription control authority — it audits, trains, certifies, and litigates against the formal model rather than the enacted one.
  • A systematic divergence. The two layers drift apart, not randomly but in a sustained direction, as the prescription lags changes in equipment, inputs, demand, and regulation.
  • A load-bearing gap. The system frequently works because operators silently bridge the divergence; suppressing the adaptation without fixing the prescription degrades performance.
  • A brittleness invariant. A large load-bearing gap yields excellent routine performance and concealed fragility — the system fails on whatever disrupts the informal layer: rapid onboarding, work-to-rule, enforcement against informal practice, or literal automation.

The components compose so that the object of management is the relationship between the prescribed and enacted layers, not either alone: the structure replaces "did the operator deviate?" with "was the prescription a recipe the work could actually follow, and which adaptations are protective versus latent risk?"

What It Is Not

  • Not an impedance mismatch. impedance_mismatch_and_coupling_efficiency is a transfer-efficiency loss between two coupled components; procedure-work mismatch is a representational divergence between an authoritative description of work and its enactment — the gap is often load-bearing, not pure loss.
  • Not a formalization failure. formalization is the act of making tacit practice explicit; the mismatch persists after formalization because the formal artefact lags the evolving work it describes.
  • Not failed validation. validation checks whether an artefact meets its spec; here the artefact may be internally valid yet diverge from how work is actually done.
  • Not implicit knowledge as such. implicit_knowledge is the tacit know-how operators hold; procedure-work mismatch is the structural gap that such knowledge bridges — the relationship between a prescription and an enactment, not the knowledge itself.
  • Not a normalization of deviance. normalization_of_deviance is the drift of acceptable practice toward unsafe; the mismatch is broader and morally neutral — much of the gap is protective adaptation that keeps the system running.
  • Common misclassification. Reading every incident as "the operator didn't follow procedure." Catch it by asking whether the procedure was followable and whether the enacted practice everyone tacitly relied on was the one the operator performed; if the book could not be followed, the cause is the gap, not the deviation.

Broad Use

The pattern recurs whenever an institution governs work with an explicit normative artefact. In aviation and nuclear operations, checklists and operating manuals are continuously confronted with edge cases the writers did not anticipate, and "what the book says versus what we actually do" is institutionally recognised and partially formalised. In medicine, clinical guidelines and surgical timeouts confront the actual conduct of cases, and the literature on nursing workarounds and clinicians' adaptation of electronic workflows is exactly this pattern. In finance operations, settlement and compliance procedures are where most operational-risk events originate, and the regulatory regime is in part an attempt to enforce procedure-work alignment. In software and DevOps, runbooks and on-call playbooks are the empirical reason engineers say "documentation is always lying," and blameless postmortems try to surface enacted reality. In public administration, case-handling procedures confront how caseworkers actually triage — "street-level bureaucracy" is the social-science name for the gap as institutionally constitutive of policy. In construction, drawings confront what the field actually builds, showing up as request-for-information traffic and as-built drawings. In education, curriculum confronts what is taught. In law, written law confronts customary practice, and bodies of administrative law are codifications of the gap after the fact.

Clarity

The vocabulary forces a distinction between two questions usually conflated: is the procedure right? — whether the formal model accurately describes how the work needs to be done — and is the work right? — whether the enacted practice is safe, efficient, and producing the intended outcome. These can diverge in all four combinations: a good procedure with bad work (a training failure), a bad procedure with good work (informal adaptation rescuing the system), a bad procedure with bad work (an incident waiting to happen), and a good procedure with good work (the stated goal). Naming the two axes separately is what lets the analyst place a given situation in the right cell rather than collapsing all four into "people aren't following the rules."

The frame also changes who one looks at after an incident. The instinct to blame the operator — "they didn't follow procedure" — presumes the procedure was followable; the procedure-work-mismatch frame asks whether it was, and whether the work-as-done the operator performed was the work-as-done everyone tacitly relied on. That reframe, central to Just Culture and to Safety-II, redirects attention from the operator's compliance to the relationship between the prescribed and enacted layers, which is where the recoverable cause usually lives. The clarifying move is to ask not "did the operator deviate?" but "was the prescription a recipe the work could actually follow?"

Manages Complexity

The pattern compresses a wide and confusing family of safety, compliance, and operational-risk events to one diagnostic frame: identify the prescriptive artefact, identify the enacted practice, characterise the gap, and ask whether the gap is harmless adaptation, latent risk, or active failure. The intervention family is correspondingly compressed into a small menu: re-author the procedure against observed practice; re-train the practice against the procedure; instrument the gap with feedback channels that surface enacted practice to the prescription-owner; design for resilience by making the system safe under the realistic range of enacted practice; or formalise selectively, codifying only what must be standardised.

The compression matters because the wrong intervention makes the problem worse, and the frame says which is wrong when. Re-training without re-authoring drives the adaptation underground, since the operators were bridging a real gap; re-authoring without instrumenting locks in yesterday's adaptations as tomorrow's prescription, recreating the gap on a delay. By naming the gap's load-bearing character explicitly, the frame warns against the two reflexive fixes — discipline the operators, or rewrite the manual to match current practice — and points instead toward instrumenting the gap so that the choice among interventions is made with visibility into which adaptations are protective and which are latent risk.

Abstract Reasoning

Recognising the pattern lets one reason about a system's normative architecture — the relationship between its prescribed and enacted layers — independently of the substrate, through three transferable abstractions. The audit gap: any institution that monitors the prescribed layer (training records, certifications, sign-offs) without instrumenting the enacted layer (telemetry, observation, post-hoc analysis) cannot detect the divergence until it produces an incident. The handoff gap: mismatch is worst where work crosses institutional boundaries — between specialties, shifts, or contractors and operators — because the gap-bridging informal knowledge does not cross the boundary. And the brittleness signature: systems with a large, load-bearing gap are deceptively stable, with excellent routine performance because operators bridge the gap but fragility to anything that disrupts the informal layer — rapid onboarding, work-to-rule, enforcement against informal practice, or automation that follows the procedure literally.

This last is structurally identical to the tacit-knowledge-in-production pattern in economics and the Polanyi paradox in automation: the gap between prescribed and enacted work is exactly the surface across which substituting machinery for skilled humans goes wrong. The reasoning is, however, heavily framed and low in structural abstraction: "procedure," "work," and "normative artefact" are wholly human-institutional categories, the pattern presupposes a recording institution and an enacted practice, and it does not apply to non-human substrates at all. That is why its substrate independence is moderate despite broad domain reach — the pattern travels across every human-institutional field but does not generalise beyond them, and its structural abstraction is the lowest among the batch's primes.

Knowledge Transfer

The pattern travels because its roles map across human-institutional substrates: the prescriptive artefact maps to a checklist, a clinical guideline, a settlement procedure, a runbook, a case-handling procedure, a drawing, a curriculum, or a statute; the enacted practice maps to the cockpit's actual conduct, the nurse's workaround, the trader's exception handling, the engineer's deployment, the caseworker's triage, or the field's build; and the gap-bridging knowledge is the tacit, boundary-bound adaptation in every case. Because the roles correspond, the intervention family — re-author, re-train, instrument, design for resilience, formalise selectively — is the same set of moves everywhere, and choosing wrongly makes the problem worse in the same way everywhere.

The documented transfers are concrete and have shaped real institutions. Crew Resource Management, born from cockpit incidents where the procedure-practice gap killed people, transferred to operating-theatre teamwork and ICU rounding. Nuclear operations' peer-review and operating-experience programs are the structural ancestor of blameless postmortems and shared incident databases in software reliability. Street-level-bureaucracy reasoning transfers directly to the design of decision-support systems for caseworkers, where how much discretion to leave is the procedure-work-gap question formalised as a design parameter. Construction's request-for-information practice against ambiguous specs has an exact analogue in software's architecture-decision-record traffic, with the same management techniques. And Safety-II's argument that procedures are resources for action rather than recipes for action transfers to incident-response playbooks that responders adapt under time pressure. The transfer is genuine — Hollnagel, Lipsky, Reason, and Polanyi each named the same structural shape in their own domain — but it is strongly framed: the pattern does not apply to non-human substrates, so the structural shape travels only across institutions that maintain an authoritative description of work and a workforce that enacts it, which is what places the prime at the framed end of the spectrum.

Examples

Formal/abstract

Hollnagel's work-as-imagined versus work-as-done framework, stated abstractly, treats a work system as two coupled representations: a prescriptive layer \(P\) (the procedure, a finite ordered set of steps defined over an idealized constraint set) and an enacted layer \(E\) (the realized conduct over the actual constraint set, including variable inputs, time pressure, and exceptions). The institution invests \(P\) with control authority: it audits, certifies, and litigates against \(P\), not \(E\). The structural claim is that the divergence \(E \setminus P\) is non-empty and load-bearing — performance depends on the tacit adaptations in \(E\) that \(P\) does not encode. One can make this rigorous by treating \(P\) as a partial function from situations to actions, with gaps wherever the idealized constraint set fails to cover a real situation; \(E\) is a total function operators construct by filling those gaps with know-how. The brittleness invariant follows formally: any operation that forces \(E \to P\) — a strict work-to-rule, literal automation, or rapid onboarding of operators who lack the tacit fill — removes exactly the gap-bridging that made the partial \(P\) executable, so routine performance is excellent while the system fails precisely when the informal layer is disrupted. The diagnostic the framework dictates: instrument \(E\) (observe and record actual practice), not just \(P\), because auditing \(P\) alone cannot detect a divergence until it produces an incident.

Mapped back: The two-layer formalism instantiates every role — prescriptive artefact, enacted practice, invested authority, systematic divergence, load-bearing gap, brittleness invariant — and shows the object of management is the relation between the layers, not either alone.

Applied/industry

In hospital nursing, an electronic medication-administration system prescribes a barcode-scan workflow: scan patient wristband, scan medication, confirm match, administer. The prescriptive artefact is the mandated scan sequence; the enacted practice is what nurses actually do under a full ward with malfunctioning scanners and uncooperative patients — batch-scanning medications at the cart, taping spare wristband barcodes to a clipboard, administering then scanning. The institution audits scan-compliance rates and certifies against the prescribed workflow, yet the ward runs on time only because nurses bridge the gap. The load-bearing character is exposed the moment a strict enforcement drive or a literal scanner lockout removes the workarounds: throughput collapses and, worse, the safety the scan was meant to provide was never actually delivered by the documented path. The prime's intervention menu applies directly — instrument the gap (observe real scanning practice), re-author the workflow against it, and design scanners and exception paths that the realistic range of practice can follow. The identical structure governs software on-call operations: a runbook prescribes incident-response steps that lag the current architecture, so engineers say "the docs are always lying" and respond from tacit system knowledge; the blameless postmortem is the instrument-the-gap move, surfacing enacted reality back to the runbook owner. And in public-benefits administration, a case-handling procedure confronts how caseworkers actually triage caseloads under time pressure — the street-level-bureaucracy gap — where the design question is how much discretion to leave, the procedure-work gap formalized as a design parameter.

Mapped back: Across nursing, software operations, and public administration the same roles recur — an authoritative prescription, a divergent enacted practice, a load-bearing tacit bridge — and the same intervention family transports: instrument the gap before choosing between re-authoring, re-training, or designing for the realistic range of practice.

Structural Tensions

T1 — Protective Adaptation versus Latent Risk (sign/direction). The gap is load-bearing, but not every deviation is protective — some workarounds are the latent failure waiting to surface. The prime says the gap is "managed not eliminated," yet it does not by itself tell which adaptations to keep. The failure mode is blanket tolerance: treating all enacted deviation as protective and absorbing a genuinely dangerous shortcut. The boundary is with near_miss_normalization, where a margin-eroding adaptation reads as competence. Diagnostic: for each adaptation, does it bridge a real gap in the prescription, or consume safety margin the prescription was protecting? Only the former is protective.

T2 — Re-Author versus Drive Underground (coupling). Re-training to the procedure without re-authoring drives adaptation underground; re-authoring to current practice without instrumenting locks yesterday's workaround in as tomorrow's rule. The two reflexive fixes each make it worse, and the frame's value is saying which is wrong when. The failure mode is wrong-lever amplification: picking re-training when the gap was real, or re-authoring when the practice was unsafe. Diagnostic: is the enacted practice safe-and-efficient or unsafe? Re-author the former, re-train the latter — never both with one default.

T3 — Instrumented Visibility versus Observer Effect (measurement). The central remedy is instrumenting the enacted layer, but observing actual practice changes it — operators perform the procedure when watched and revert when not. The failure mode is Hawthorne capture: the instrumented enacted layer shows compliance that the un-watched layer does not, so the audit measures the wrong practice. Shared with shortcut_learning's gap between graded behavior and real behavior. Diagnostic: does the observed enacted practice match the telemetry from unobserved operation? A divergence means the instrument is measuring performance for the observer, not work-as-done.

T4 — Routine Stability versus Brittle Onboarding (temporal). A large load-bearing gap yields excellent routine performance and concealed fragility — the brittleness surfaces only when the informal layer is disrupted. The failure mode is deceptive-stability complacency: years of smooth operation read as a healthy system right up to the rapid-onboarding or automation event that removes the tacit fill. This is the latent-exposure tension. Diagnostic: simulate disruption of the informal layer (work-to-rule drill, new-hire shadowing) and measure performance; routine metrics cannot reveal the brittleness because the bridge is silently holding.

T5 — Standardization versus Local Adaptation (scalar). The frame recommends formalizing selectively — codify only what must be standardized — but the boundary between "must standardize" and "leave to local know-how" is exactly contested, and over-standardizing destroys the adaptive capacity that made the system work. The failure mode is over-codification: a Polanyi-paradox failure where automating or rigidifying the tacit layer removes the discretion performance depended on. Boundary with stovepipe_system's standardization tension. Diagnostic: does the work cross institutional boundaries (where standardization helps) or stay local (where discretion helps)? The handoff gap, not the whole process, is where codification pays.

T6 — Operator Blame versus System Cause (scopal). The frame relocates cause from "the operator deviated" to "the prescription was unfollowable," but over-applying this can absolve genuine operator error, removing accountability where the procedure really was adequate. The failure mode is systemic-cause overreach: treating every incident as a procedure-work mismatch when some are training or conduct failures (the "good procedure, bad work" cell). Boundary with action_bias's evaluation-layer concern. Diagnostic: place the incident in the four-cell grid — was the procedure followable and the enacted practice still wrong? If so, the cause is the work, not the gap.

Structural–Framed Character

Procedure-work mismatch sits firmly at the framed end of the structural–framed spectrum — an aggregate of 0.9, the highest in this batch, with four of five diagnostics at or near the ceiling. There is a thin relational shape underneath (two coupled representations, one invested with authority, diverging over time), but almost everything load-bearing about the prime is human-institutional, and the grade reflects that without exaggeration.

The diagnostics line up overwhelmingly toward framed. Institutional origin, human-practice-bound, import-vs-recognize, and vocabulary-travels all read at the maximum. The prime's three core terms — procedure, work, and normative artefact — are wholly human-institutional categories: a procedure is something an institution writes, audits, certifies, and litigates against, and "work-as-done" is the enacted practice of human operators filling gaps the prescription never covered. The pattern presupposes a recording institution and an enacted practice, and the entry states plainly that it does not apply to non-human substrates at all. Invoking it imports a whole interpretive frame — Safety-II, Just Culture, the four-cell grid of good/bad procedure against good/bad work, the instinct to ask "was the prescription followable?" rather than "did the operator deviate?" — rather than merely recognizing a pattern already present in physics. The vocabulary carries its full home lexicon: checklist, runbook, work-to-rule, street-level bureaucracy, blameless postmortem. Only evaluative weight sits lower, at the midpoint, and rightly so: the prime insists the gap is morally neutral and frequently protective, much of it competent adaptation rather than deviance, which keeps it from a clean evaluative loading.

The prime's own substrate note confirms the placement and even flags that its structural abstraction is the lowest among the batch's primes — the pattern travels across every human-institutional field (aviation, medicine, finance, DevOps, public administration, construction, law) but does not generalize beyond them, so the structural shape moves only among institutions that maintain an authoritative description of work and a workforce that enacts it. That is the defining mark of a heavily framed prime: a genuine but thin relational skeleton, legible only once the institutional practice that constitutes it is carried along.

Substrate Independence

Procedure-work mismatch is a moderately substrate-independent prime — composite 3 / 5 on the substrate-independence scale. Its domain breadth is genuinely wide within the human-institutional world: Hollnagel's work-as-imagined/work-as-done gap recurs with the same structural force across aviation and nuclear operations (checklists meeting edge cases), medicine (clinical guidelines versus nursing workarounds), finance operations (settlement and compliance procedures as the origin of most operational-risk events), software and DevOps (runbooks that "are always lying," surfaced in blameless postmortems), public administration (street-level bureaucracy), construction (drawings versus as-built reality), education, and law (written law versus customary practice). What pulls the structural-abstraction component down to the bottom of the scale is that the prime's three load-bearing terms — procedure, work, and normative artefact — are wholly human-institutional categories: the pattern presupposes a recording institution that writes and audits an authoritative description and a workforce that enacts it, and it does not apply to any physical or biological substrate at all. Transfer evidence is strong but confined to that band: the diagnostic move (ask whether the prescription was followable, manage the gap rather than abolish it) and the Safety-II / Just-Culture remedies travel cleanly across every documented field, yet each destination is itself an institution governing work with a normative artefact. Wide reach across human institutions, capped by the absence of any non-institutional substrate, is exactly what fixes the composite at 3.

  • Composite substrate independence — 3 / 5
  • Domain breadth — 4 / 5
  • Structural abstraction — 2 / 5
  • Transfer evidence — 4 / 5

Relationships to Other Primes

One-hop neighborhood: parents above, mutual partners to the right, children below.Procedure-WorkMismatchsubsumption: Formal vs. Informal StructuresFormal vs. Info…

Parents (1) — more general patterns this builds on

  • Procedure-Work Mismatch is a kind of Formal vs. Informal Structures

    Procedure-work mismatch (work-as-imagined vs work-as-done) is the official-vs-actual divergence specialized to a prescriptive work artefact and its enacted practice, with the gap often load-bearing. is-a the formal/informal-structures distinction applied to procedures. Clean fit.

Path to root: Procedure-Work Mismatch → Formal vs. Informal Structures → Formalization → Transformation

Neighborhood in Abstraction Space

Procedure-Work Mismatch sits in a sparse region of abstraction space (95th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely rather than landing on a neighbor.

Family — Unclustered & Miscellaneous (91 primes)

Nearest neighbors

Computed from structural-signature embeddings · 2026-06-14

Not to Be Confused With

The nearest existing prime by embedding is impedance_mismatch_and_coupling_efficiency, and the confusion is real because both describe two things that "don't fit" and a resulting loss. But impedance mismatch is about energy or signal transfer between coupled components: a source and a load with mismatched impedances waste power at the interface, and the cure is to match the impedances so transfer is efficient. Procedure-work mismatch is about a representational relationship: an authoritative description of work and the actual conduct of work diverge, and — crucially — that divergence is frequently load-bearing rather than pure waste. The system often works because operators silently bridge the gap, so "matching" the procedure to the work can destroy the adaptive capacity the system depended on. Impedance mismatch frames the gap as inefficiency to be eliminated; procedure-work mismatch frames it as a relationship to be managed, with some adaptations protective and some latent risk. A practitioner who imports the impedance intuition will reflexively try to close the gap and trigger exactly the work-to-rule collapse the prime warns against.

A second genuine confusion is with normalization_of_deviance, which sits one file away conceptually and shares the safety-engineering lineage. Normalization of deviance is the temporal drift by which a once-unacceptable practice becomes routine and then invisible — a moralized slide toward the unsafe. Procedure-work mismatch is broader and morally neutral: it names the standing structural divergence between work-as-imagined and work-as-done, of which normalization of deviance is one harmful special case (the subset where the drifting enacted practice is consuming safety margin). Much of the procedure-work gap is the opposite of deviance — it is competent adaptation filling holes the procedure never covered. The distinction is load-bearing because the two prescribe different reflexes: normalization of deviance says "this drift is dangerous, halt it," while procedure-work mismatch says "first ask whether this adaptation is protective or latent risk before deciding to halt it." Treating all mismatch as deviance pathologizes the very adaptations that keep the system running.

A third confusion is with implicit_knowledge. Implicit knowledge is the tacit, hard-to-articulate know-how that operators carry — the skill that resists codification. Procedure-work mismatch is not that knowledge but the gap it bridges: the structural relationship between the formal prescription and the enacted practice, across which tacit knowledge does its work. Implicit knowledge is one of the resources operators deploy to bridge the mismatch; the mismatch is the surface that demands bridging. The distinction matters because the remedies differ: implicit knowledge calls for apprenticeship, mentoring, and tacit-to-explicit transfer; procedure-work mismatch calls for instrumenting the enacted layer and deciding among re-authoring, re-training, and designing for the realistic range of practice. A practitioner who sees only "missing tacit knowledge" will invest in training when the real problem may be a procedure that no amount of skill can render followable.

For a practitioner, the distinctions sort by what is actually wrong. If two components waste signal at an interface, it is impedance_mismatch_and_coupling_efficiency; if a once-forbidden practice has quietly become normal and is eroding margin, it is normalization_of_deviance; if the issue is uncodifiable skill in operators' heads, it is implicit_knowledge; and if an authoritative description of work has diverged from how work is really done — with some adaptations protective and some dangerous — it is procedure-work mismatch, whose distinctive move is to instrument the gap before choosing which way to close it.

Solution Archetypes

No catalogued solution archetypes reference this prime yet.