---
title: "There Is No Exception Handler"
subtitle: "When every office does its job and the problem still disappears"
author: Elias Kunnas
description: "Every office can handle its part while an unmatched problem disappears. The missing exception handler owns the no-match state until handoff or closure."
canonical: https://kunnas.com/articles/there-is-no-exception-handler
url: https://kunnas.com/articles/there-is-no-exception-handler.md
date_published: 2026-08-31
date_modified: 2026-09-01
corpus_frame_url: https://kunnas.com/articles/how-to-read-this.md
---
## How to read this corpus

The corpus applies one lens to many domains: what mechanisms produce the outcome? It shares four methodological commitments and one explicit directional commitment. Each linked page argues for its part; the links are derivations and disputes, not evidence inherited by every page. The directional commitment does not by itself settle system boundary, distribution, sacrifice, or institutional authority.

1. **Mechanisms are what act.** Incentive gradients, selection pressures, feedback loops, and capital stocks produce the distribution of outcomes. Intentions, labels, official categories, and stated values are evidence about mechanisms, or are themselves coordination mechanisms. They are not causal substitutes. — [Mechanism Realism](https://kunnas.com/articles/mechanism-realism.md) · [Only Selection](https://kunnas.com/articles/only-selection.md)
2. **The reference telos is sustained flourishing.** The broadest achievable adaptive safety margin over deep time — not the continuity of any incumbent state, coalition, institution, or doctrine. A mechanism's own stated goal can still serve as a local proof obligation — showing that its incentives defeat even the purpose it claims is a bounded finding — but meeting that goal establishes nothing about the margin. — [Flourishing Is Maximum Safety Margin](https://kunnas.com/articles/flourishing-is-maximum-safety-margin.md)
3. **Law, rights, legitimacy, democracy, markets, and sovereignty are mechanisms under evaluation.** They are constraints, carriers, or proxies inside the analysis. None is a terminal value or a boundary of what is real. Treating one as terminal ends the mechanism search before it starts. Evaluation carries current function, replacement cost, path dependence, uncertainty, capture risk, reversibility, and who bears model error into the ledger. — [The Stack](https://kunnas.com/articles/the-stack.md) · [Mechanism Space](https://kunnas.com/articles/mechanism-space.md)
4. **Optimization is a system function.** A civilization has to build, exercise, and revise metamechanisms that search mechanism-space, discard dominated options, install, observe effects, and repair under uncertainty. Not running that loop leaves margin unrealized, and that is itself the failure. No single component — analyst, model, or institution — is presumed to contain a global optimum; the capacity is a property of the system. — [Telic Systems](https://kunnas.com/articles/telic-systems.md) · [The Three-Layer Architecture](https://kunnas.com/articles/three-layer-architecture.md)
5. **Uncertainty is preserved, not spent.** Partial orders, binding constraints, unknowns, and residuals stay explicit. An unmeasured effect is not a favorable default. — [The Compression Paradox](https://kunnas.com/articles/compression-paradox.md) · [Cargo Cult Epistemology](https://kunnas.com/articles/cargo-cult-epistemology.md)

*Each essay bears its own evidence. Links carry definitions, derivations, applications, and disputes; they do not transfer proof. Criticism is answered on its substance.*

Canonical: <https://kunnas.com/articles/how-to-read-this.md>

---

# There Is No Exception Handler

*When every office does its job and the problem still disappears*

Elias Kunnas

## Thesis {#thesis}

Every office can complete its local role—publish, acknowledge, forward, or correctly decline—while the unmatched object disappears. Bills, applications, cases, and programme proposals have ordinary owners; an input that fits no case type, spans several mandates, or loses its decisive feature when classified does not. Local handlers can all return success while the no-match condition produces no system-level fault.

The sender then becomes the memory, router, retry policy, and escalation path. If the sender stops, the exception is swallowed. An institutional exception handler preserves the unmatched object, records the failed routes, assigns temporary progression ownership, and ends in an accepted handoff or a reasoned terminal state. It owns that progression, not the object’s truth or importance.

## Standard objections addressed in this essay

- “Government already has appeals, auditors, coordination bodies, and escalation routes.” — §§IV–V (components compose only where no-match creates an owned handoff)
- “Everyone will call their own idea exceptional.” — §II (exception describes failed matching, not merit)
- “Offices owe outsiders no answer beyond their mandate.” — §VII (a correct local stop still leaves the conjunction unowned unless someone issues a disposition)
- “A general handler would drown in spam.” — §§VII–VIII (triage and inspectable termination remain handling)
- “This is a technocratic super-agency.” — §VI (the handler owns progression, not substantive closure)
- “Unknown exceptions cannot be designed for in advance.” — §VI (the fallback handles failed matching, not future content)
- “The exception handler can itself be captured.” — §VIII (audit, appeal, sunset, and replacement apply recursively)

---

## I. Every Handler Returned Success {#i-every-handler-returned-success}

A consultation publishes the comment. A legislative office identifies a working group and a later window. A policy unit accepts a version rewritten in its native frame and says the matter can proceed through them. Each step is successful local handling: receipt, classification, referral, or publication. None of those local successes, by itself, creates an externally inspectable case whose next transition anyone has accepted responsibility for.

The sender remains the person who knows whether the object is alive, dead, waiting for a window, or never entered a durable queue. The next visible move is still theirs: another translation, another named person, another institutional calendar. Internal consideration, deferment, and non-acquisition can share that same external representation.

Every participant may have performed their local role correctly. One published the submission. One identified a plausible route. One accepted the native-form artifact. One promised internal routing. Nothing in the visible chain had to report failure.

Nothing in the visible chain became responsible for the conjunction:

substantively understood

+plausible destination found

+native artifact supplied

+no observable institutional state follows

Every local handler returned success. The global exception remained unhandled.

---

## II. An Exception Is Not a Special Claim {#ii-an-exception-is-not-a-special-claim}

In software, an exception is an event that disrupts the assumptions of the normal execution path. The runtime preserves information about the failure and searches for a handler capable of deciding what follows. Value, correctness, and moral importance remain separate questions. When no handler exists, the failure becomes visible instead of reporting ordinary completion.

Public processes meet the same event. **An institutional exception is an input for which the ordinary handling assumptions fail.** The input may be a proposal, anomaly, warning, cross-domain dependency, implementation failure, or dispute. It

becomes an exception when at least one of the following holds:

- **No case type.** No recognized case type can carry it.
- **Incomplete mandate.** No mandate covers the complete object.
- **Fragmented ownership.** Several actors own fragments, but none owns progression.
- **Authority gap.** The actor that understands the problem lacks authority over the next transition.
- **Referral loop.** Each referral returns responsibility to the original sender.
- **Feature deletion.** The normal process can classify the input only by deleting the feature that made it consequential.

Importance is a separate question. Most unmatched inputs may be low-value. A false theory can expose a real routing exception. An important proposal can fit an ordinary grant, legislative, judicial, or administrative route and create no exception at all.

**An institutional exception handler is the fallback arrangement for what happens when no ordinary process can own the next transition.** It assigns a named actor or office to carry the no-match state until handoff or termination.

Ordinary processes already have at least three possible outcomes:

- handled
- rejected under the process's own rule
- referred to another recognized process

Institutional systems often permit a fourth, which is one state with three features:

- acknowledged locally
- globally unhandled
- reported nowhere as failure

That is a swallowed exception. The remaining question belongs to the conjunction of those local answers:

> **When every ordinary recipient can correctly say “not mine,” what process becomes responsible for the fact that nobody’s process fits?**

If the answer is still the sender, the system has no exception handler.

---

## III. The Sender Becomes the Runtime {#iii-the-sender-becomes-the-runtime}

The missing handler transfers the work to the person who noticed the exception.

The sender preserves the object across conversations. The sender remembers who has seen which version. The sender translates it into a magazine article, a one-page memorandum, a consultation response, a committee question, a grant application, or a software demonstration. The sender searches for adjacent mandates, asks for introductions, waits for institutional windows, follows up, and decides whether silence means rejection or merely delay.

The effective runtime is a loop the sender runs until they stop:

while the sender still has energy:

- preserve context
- identify the next handler
- translate to native form
- request a handoff
- wait
- retry

if the sender stops:

- ordinary completion

The person is performing several system functions at once:

- **memory:** preserving the object and its history
- **routing:** finding the next plausible jurisdiction
- **translation:** compiling the same object into each recipient’s grammar
- **retry policy:** deciding when silence warrants another attempt
- **escalation:** locating a higher or cross-boundary forum
- **continuity ownership:** keeping the matter alive across staff, elections, and windows
- **error logging:** retaining the fact that all previous routes failed

Curiosity, conscience, obsession, professional identity, or expected social value may justify the effort for the sender. Public architecture that depends on that private energy remains unreliable.

The incentives point the other way for everyone downstream. The public benefit of a better institution is diffuse. The cost of acquiring an unaffiliated proposal is concentrated: attention, reputational exposure, coordination, opportunity cost, and the risk of championing something that fails. Ignoring the proposal creates almost no attributable cost. Successful adoption may credit the institution or politician; failed adoption creates a visible sponsor to blame.

“Caring” therefore decomposes into several different states:

| State | What it means |
|----|----|
| **Epistemic interest** | This may be true. |
| **Personal concern** | This seems important. |
| **Procedural responsibility** | I must issue a disposition. |
| **Strategic incentive** | My role or interests depend on the result. |
| **Operational capacity** | I can authorize, fund, or implement a next step. |

Outreach can reach the first two and produce no externally different result from rejection. Durable institutional consequence usually requires operational capacity joined to procedural responsibility or strategic incentive; exceptional people sometimes substitute for one or both.

Where there is no exception handler, the person who noticed the problem becomes the handler. When that person stops, the system records normal completion.

---

## IV. The No-Match State Is Its Own Failure {#iv-the-no-match-state-is-its-own-failure}

Several neighboring failures can produce silence, and they remain distinct.

| Failure | What already exists | What is missing |
|----|----|----|
| **No router** | a competent evaluator | a path that gets the object there |
| **No procedural object** | a competent process | an artifact the process can receive and act on |
| **No response authority** | an evaluator and a finding | a duty or incentive that turns the finding into a countermove |
| **Structural residue** | a locally valid refusal | ownership of the consequence that survives the refusal |
| **No exception handler** | several normal processes, each locally coherent | a fallback for the conjunction of their no-match results |

The distinction matters because each repair fails when applied to the wrong state.

A better introduction does not help when the destination does not exist. A more polished memorandum does not help when no mandate can own the object. A mandatory answer by the first recipient does not solve a cross-boundary case if that recipient lacks authority and no accepted referral follows. Naming one owner can create capture when the real need is temporary coordination among several authorities.

The no-match state is higher-order: it concerns the composition of normal handlers.

Analysis may exist in one office, legal authority in another, agenda access in a third, and implementation capacity in a fourth. Unless a failed match creates an accepted, stateful handoff among them, the capabilities coexist without composing. “All the functions exist somewhere” does not answer who owns their conjunction. An exception path begins only when that conjunction has an owner.

The same distinction applies one stage upstream to knowledge. Relevant findings may exist across several fields without any stable decision object for an institution to match. [*Non-Compilation*](non-compilation.md) names that earlier failure. The exception handler begins once there is enough of an object to preserve and test; it cannot be a universal substitute for performing every missing cross-field synthesis.

A department may correctly say that programme design is political. A party may correctly say that statutory institutional design belongs to government and the legislature. A legislative committee may correctly say that no bill is before it. An audit office may correctly say that the proposed institution does not fit its audit mandate. A research funder may correctly say that implementation is outside a research call.

Every answer can be locally valid. Their conjunction can still leave a potentially consequential function unowned, with no system state for that fact.

A neighbouring failure, not the clean no-match specimen, is the 2026 Southport Inquiry: no organization or multi-agency arrangement took ownership of the accumulating risk, with persistent uncertainty over the lead agency, while the report also identified missed duties and inconsistent decisions. Local errors become harder to integrate and correct when no process owns the accumulating whole. The ministerial statement compressed the same finding as a failure that belonged to everyone and therefore to no one.

The exception and the proposal remain distinct. Even if the proposal is wrong, the repeated no-match can reveal that the institutional taxonomy has no way to evaluate or reject that class of object as a whole.

This is the live counterpart of [Civilization’s Blind Spot](civilizations-blind-spot.md): a correction that never becomes a case may later disappear from the historical denominator. [Optionality Has No Router](optionality-has-no-router.md) describes a signal that cannot reach a potentially interested evaluator. [The Procedural Object](the-procedural-object.md) describes the form an artifact needs when the process already exists.

[The Mandate Trap](the-mandate-trap.md) explains why locally coherent mandate boundaries can leave an upstream function unowned. [Structural Residue](structural-residue.md) tracks what remains causally live after a gate correctly refuses. The exception handler owns the composition point: what the system does when those local boundaries jointly return no match.

---

## V. Bounded Exception Handlers Already Exist {#v-bounded-exception-handlers-already-exist}

Institutions have built exception handlers where the object class and failure cost became sufficiently legible. The examples below are bounded intakes for recognized anomaly, recommendation, and incident classes, not the general no-match handler of §II. They are partial architecture.

### Aviation safety reporting {#aviation-safety-reporting}

NASA’s Aviation Safety Reporting System accepts confidential reports from pilots, controllers, mechanics, cabin crew, and other frontline participants. Reports are read by experienced analysts, hazards can generate alerts to the appropriate authority, and de-identified records enter a database used for broader learning. The reporting system preserves the anomaly, classifies it, routes it, and makes recurrence visible, leaving command with airlines and regulators.

### Safety recommendations {#safety-recommendations}

The United States National Transportation Safety Board issues recommendations to named recipients and asks them to respond within ninety days, or thirty days for urgent recommendations. Federal law separately requires the Secretary of Transportation to give a formal written response within ninety days, stating whether the recommendation will be adopted in whole, adopted in part, or refused, with a timetable or reasons. The Board then maintains public status classifications and can keep a recommendation open while action remains insufficient.

The NTSB makes acceptance, alternative action, refusal, non-response, reconsideration, and closure different states; implementation remains with the named recipient.

### Unified command {#unified-command}

The Incident Command System creates a temporary structure when an incident crosses jurisdictions or agencies. Under Unified Command, actors with different legal authorities can coordinate a common response without surrendering their own authority or accountability. The exceptional event creates a process for assembling ownership rather than requiring one permanent super-agency to possess every power in advance.

### Petitions and reasoned denial {#petitions-and-reasoned-denial}

Administrative law contains smaller examples. Under 5 U.S.C. § 555(e), denial of a written application, petition, or other request made in connection with an agency proceeding ordinarily requires prompt notice and brief grounds: a bounded way to prevent a recognized procedural object from disappearing into undifferentiated silence.

### Workflow exception handling {#workflow-exception-handling}

Workflow research distinguishes expected exceptions from unexpected ones and proposes handlers at task, case, and process-specification levels. The important move is dynamic selection: when the predefined path fails, context and prior exception knowledge help select a recovery process rather than forcing the case through the nearest normal route.

These systems show that no-match conditions can be represented, assigned, escalated, resolved, and learned from without granting one actor substantive authority over everything. Aviation systems still begin with recognized safety reports and accidents. Incident command begins after an incident has been declared. Administrative petitions begin inside a known agency and legal object class. The remaining problem is the slow, cross-mandate input whose first unresolved question is which public process, if any, should own it.

---

## VI. What the Handler Owns {#vi-what-the-handler-owns}

The named actor or office inside the handler owns progression after ordinary matching fails: the next transition, not the object’s truth. For readability, “the handler” below is shorthand for that accountable role and the process supporting it. Importance, adoption, and implementation remain with the legislature, court, ministry, regulator, profession, or political party that must eventually make the substantive decision.

Its minimum operations are narrower.

### 1. Preserve the object {#preserve-the-object}

The original claim, evidence, requested change, and declared uncertainty receive a stable identity. Translation may add representations; it may not silently replace the object with a more familiar but materially different one.

### 2. Record the failed matches {#record-the-failed-matches}

The case retains which normal routes were attempted, what each route could and could not own, and why the match failed. “Outside scope” becomes evidence about system topology rather than a reason to erase the trace.

### 3. Assign temporary progression ownership {#assign-temporary-progression-ownership}

An actor owns the next transition while the final substantive owner remains unresolved. This is continuity ownership, not policy authority.

### 4. Triage materiality and public standing {#triage-materiality-and-public-standing}

Most unmatched inputs should terminate cheaply. The handler can classify them as unsupported, duplicative, immaterial, private, outside any legitimate public role, or insufficiently specified. Handling includes a bounded no.

### 5. Escalate scope rather than forward sideways forever {#escalate-scope-rather-than-forward-sideways-forever}

Repeated no-match results should reach a forum authorized to resolve jurisdiction, assemble a temporary cross-boundary process, or make the explicit political decision that no public route should exist.

### 6. Require an accepted handoff {#require-an-accepted-handoff}

Forwarding transfers responsibility only when the receiving body acknowledges that it owns the next transition. Until then, progression remains with the handler.

### 7. Terminate inspectably {#terminate-inspectably}

Valid terminal states include:

- reasoned rejection
- no legitimate public route
- bounded deferral with a stated trigger
- accepted handoff to a normal process
- temporary cross-boundary process opened

Silence is not a disposition. Inspectable means visible to the authorized accountability chain; privacy, security, or commercial confidentiality may justify limited or delayed public disclosure.

### 8. Compile recurrence into the normal architecture {#compile-recurrence-into-the-normal-architecture}

A fallback that repeatedly handles the same class has discovered a missing ordinary route. The system should then modify a mandate, create a procedure, define a new case type, or make an explicit decision to leave that class outside public ownership.

The state machine is:

normal match→ordinary process

no match→stable exception case→bounded triage→accepted handler, temporary forum, or reasoned terminal state

recurrent exception class→review the normal routing architecture

The handler is a progression owner of last resort: it can force the system to state who decides, and it cannot decide in their place.

---

## VII. The Right to Stop {#vii-the-right-to-stop}

Finite institutions cannot investigate every claim that someone describes as important.

A party programme team must prioritize among thousands of possible commitments. A ministry has statutory duties and an existing queue. A researcher is entitled to choose a field. A reader is entitled not to spend an hour on an unsolicited theory. An editor’s obligation can end at publication. None becomes blameworthy merely because a sender wanted another transition.

The required distinction is among at least four states:

- **substantive rejection:** the distinct object was understood and rejected;
- **scope rejection:** this process is not competent, and a named accepted route exists or the case returns to the handler;
- **public non-route:** no legitimate public process should own this class of input;
- **defer:** no substantive verdict is claimed, but a trigger or expiry determines whether the case returns.

For low-materiality inputs, these states need not produce a bespoke essay. Rules, machine-readable categories, batch dispositions, and sampled audit can keep handling costs bounded. The requirement is that “nobody acquired it” not count as evidence that the object was considered and found wanting.

A correct terminal answer may be that no public institution should take responsibility for this proposal. That is still an answer about the routing architecture. It is different from a chain in which every office can say “not ours” and no one owns whether the conjunction leaves a consequential function permanently unexamined.

---

## VIII. The Handler Can Become the Failure {#viii-the-handler-can-become-the-failure}

An exception handler can become a universal complaint warehouse, a lobbying portal, a censor, a procedural moat, or a new mandate organ that survives by accumulating unresolved exceptions.

The risks are structural.

- **Denial of service.** Cheap submissions can consume all attention. The handler needs escalating specificity, materiality thresholds, rate limits, and cheap terminal states.
- **Strategic exception claiming.** Every advocate can redescribe an ordinary loss as evidence that the system lacks a category. The sender may allege no match; the handler must independently test whether an ordinary route actually exists.
- **Agenda capture.** Whoever defines materiality can decide what government is allowed to notice. Criteria, decisions, samples of rejected cases, and error audits must be public enough to challenge patterned exclusion.
- **Centralization.** A handler that acquires every substantive power becomes the super-agency the design was meant to avoid. Temporary coordination and accepted handoff are preferable to permanent jurisdiction wherever ordinary owners can act.
- **Bureaucratic storage.** Assigning identifiers and statuses can simulate handling while cases accumulate without movement. Time limits, expiry, explicit defer states, and outcome measures must distinguish storage from progression. A catch-all that absorbs every failure and returns a generic status is error suppression.
- **Permanent exceptionalism.** The fallback can become the normal route for politically inconvenient classes, allowing ordinary institutions to avoid adapting. Recurrent classes must trigger mandate or process review.
- **False closure.** A reason code can become a more respectable form of silence. Random retrospective review, procedural appeal, and comparison of later outcomes against dispositions are needed to estimate false negatives.
- **Handler capture.** The fallback itself can preserve its mandate, protect its preferred frames, or suppress anomalies. It needs recusal, adversarial access, independent review, periodic reauthorization, and a path to replacement or abolition.

There is no infinite tower of perfect handlers. At the top, failure must at least become a visible political or constitutional fault rather than disappear. Software without a matching handler terminates. Public systems may choose to continue, but the unhandled state should survive as an attributable record of what remained unresolved.

The test is whether the handler makes no-match states distinguishable, reduces sender-dependent continuity, improves accepted handoffs, and causes recurrent exception classes to update the normal architecture at acceptable cost.

If it only manufactures tickets, it has swallowed the exception in a more elaborate form.

---

## IX. The Missing Error Signal {#ix-the-missing-error-signal}

A system cannot learn that its taxonomy is incomplete when everything outside the taxonomy disappears without generating an error.

Normal institutions should continue handling normal cases. Most unusual inputs should still be rejected, deferred, returned to private life, or assigned low priority. The system can disbelieve outsiders and still need to know when no ordinary process has accepted the object for handling, rejected it on the merits, deferred it under a trigger, or owned it at all. Rejection is allowed; a challenge that never becomes a case, or never receives a disposition, is not.

The difference is small in procedure and enormous in consequence: *not ours*, plus *not theirs*, plus *no accepted handoff*, is not the same state as *nothing failed*.

The absence of an exception handler is itself a rule. It assigns the cost of unmatched cases to the sender and to whoever later bears the uncorrected consequence. It makes persistence, social access, free time, and personal obsession part of the public selection criterion. It lets agreement, rejection, internal consideration, and oblivion produce the same observable output.

> **An institution may correctly say “not ours.” A civilization cannot let the conjunction of those answers mean “nothing failed.”**

Where there is no exception handler, the sender becomes the handler. When the sender stops, the exception is swallowed.

---

**Related:**

- [Non-Compilation](non-compilation.md) — why relevant findings can exist without a reusable decision object for any institutional process to receive.
- [Civilization’s Blind Spot](civilizations-blind-spot.md) — why corrections lost before case formation disappear from the historical denominator.
- [Optionality Has No Router](optionality-has-no-router.md) — the upstream case where a potentially valuable signal never reaches an evaluator.
- [The Procedural Object](the-procedural-object.md) — the form an artifact needs when a competent receiving process already exists.
- [The Mandate Trap](the-mandate-trap.md) — why locally coherent mandates leave cross-boundary functions globally unowned.
- [Structural Residue](structural-residue.md) — how consequences remain live after a gate correctly refuses a claim.
- [Powerless Intelligence](powerless-intelligence.md) — why accurate findings remain inert without a response path.
- [Corrective Closure Ownership](corrective-closure-ownership.md) — who must reopen and repair a mechanism after later failure.
- [When Ownership Is the Wrong Repair](when-ownership-is-the-wrong-repair.md) — why naming an owner can create capture, theatre, or overload instead of progression.
- [The Fourth Branch](fourth-branch.md) — institutional form for persistent correction architecture for persistent mechanism-lifecycle assurance.
- [The Corrigible State](the-corrigible-state.md) — the complete correction lifecycle; exception handling owns only the no-match ingress condition

## Sources and Notes

### The software allusion {#the-software-allusion}

- Oracle, [“What Is an Exception?”](https://docs.oracle.com/javase/tutorial/essential/exceptions/definition.html) and [“The Catch or Specify Requirement”](https://docs.oracle.com/javase/tutorial/essential/exceptions/catchOrDeclare.html) — exceptions disrupt normal flow; the runtime searches the call stack for a matching handler, and an unhandled exception terminates visibly.
- The title also lightly alludes to [“Dath Ilan’s Views on Stopgap Corrigibility”](https://www.lesswrong.com/posts/eS7LbJizE5ucirj7a/dath-ilan-s-views-on-stopgap-corrigibility), which reproduces a *Planecrash* passage asking what happens when an objective omits something reality later makes important: “Where’s the exception handler?” The present essay transfers that question from omitted objectives to omitted institutional object classes. No familiarity with the source is required.

### Exception handling in workflows {#exception-handling-in-workflows}

- Michael Adams, Arthur H. M. ter Hofstede, David Edmond, and Wil M. P. van der Aalst, [“Dynamic and Extensible Exception Handling for Workflows: A Service-Oriented Implementation”](https://doi.org/10.1007/978-3-540-76848-7_8) (2007) — distinguishes expected and unexpected exceptions at task, case, and process levels and supports context-sensitive handler selection.
- Mark Klein and Chrysanthos Dellarocas, [“A Knowledge-Based Approach to Handling Exceptions in Workflow Systems”](https://doi.org/10.1023/A:1008759413689) (2000) — treats prior exception cases and reusable resolution knowledge as infrastructure rather than one-off improvisation.

### Bounded public exception handlers {#bounded-public-exception-handlers}

- NASA Aviation Safety Reporting System, [“Report Processing”](https://asrs.arc.nasa.gov/overview/report.html), [“Program Outputs”](https://asrs.arc.nasa.gov/overview/outputs.html), and [programme overview](https://asrs.arc.nasa.gov/overview/summary.html) — confidential anomaly intake, expert analysis, hazard alerts, de-identification, and a durable learning database without direct operational command.
- National Transportation Safety Board, [“Responding to our Safety Recommendations”](https://www.ntsb.gov/investigations/Pages/RecipientTipRecommendations.aspx) — asks recipients to respond within ninety days, or thirty for urgent recommendations, and expects continuing updates while recommendations remain open.
- United States Code, [49 U.S.C. § 1135](https://www.law.cornell.edu/uscode/text/49/1135) — requires the Secretary of Transportation to state within ninety days whether a recommendation will be adopted, partly adopted, or refused, with timetables or reasons.
- National Transportation Safety Board, [recommendation status definitions](https://www.ntsb.gov/investigations/Pages/field-descriptions.aspx) — makes response, alternative action, non-response, reconsideration, supersession, and closure separately observable.
- Federal Emergency Management Agency, [“NIMS Management Characteristic: Unified Command”](https://emilms.fema.gov/is_0200c/content/298.html) — Unified Command assembles temporary cross-jurisdiction coordination while preserving each agency’s authority, responsibility, and accountability.
- United States Code, [5 U.S.C. § 555(e)](https://www.law.cornell.edu/uscode/text/5/555) — written applications and petitions denied by an agency receive prompt notice and brief grounds, a bounded reason-giving rule rather than a universal right to adoption.

### Cross-boundary responsibility {#cross-boundary-responsibility}

- U.S. Government Accountability Office, [*2026 Annual Report: Opportunities to Reduce Duplication, Overlap, and Fragmentation*](https://www.gao.gov/products/gao-26-108505) (GAO-26-108505, 12 May 2026) — adjacent literature on fragmented, overlapping, or duplicative federal programmes; not a specimen of locally successful no-match handling.
- The Southport Inquiry, [*Phase 1 report*](https://www.gov.uk/government/publications/the-southport-inquiry-phase-1-report) (13 April 2026) — found a failure by any organization or multi-agency arrangement to take ownership of the risk and persistent uncertainty about the lead agency, together with missed duties and inconsistent decisions; used here as a boundary case, not a pure locally-correct no-match.
- UK Home Office, [“The Southport Inquiry: Phase 1 report”](https://www.gov.uk/government/speeches/the-southport-inquiry-phase-1-report) (13 April 2026) — ministerial statement compressing the finding as a failure that belonged to everyone and therefore to no one.
- Dennis F. Thompson, [“Responsibility for Failures of Government: The Problem of Many Hands”](https://doi.org/10.1177/0275074014524013) (2014) — distributed action makes retrospective attribution difficult and can block institutional correction; prospective responsibility and oversight matter.
- Naomi Aoki, Melvin Tay, and Stuti Rawat, [“Whole-of-government and joined-up government: A systematic literature review”](https://doi.org/10.1111/padm.12949) (2024) — reviews 340 studies of cross-boundary governmental coordination and integration.
- Richard C. Feiock, [“The Institutional Collective Action Framework”](https://doi.org/10.1111/psj.12023) (2013) — explains coordination dilemmas created by fragmented authority among public organizations.
- Karl E. Weick and Kathleen M. Sutcliffe, *Managing the Unexpected* (3rd ed., 2015) — high-reliability organizing emphasizes preoccupation with small failures, reluctance to simplify, resilience, and deference to relevant expertise.

### Field specimen and scope {#field-specimen-and-scope}

The opening sequence in §I is a composite public-administration pattern: consultation, legislative referral, and native-form submission, each completing locally without creating an owned next state. It is not a claim about named recipients, internal work, or the merits of the unmatched object. GAO’s fragmentation reports appear only as adjacent literature.

The essay proposes a diagnostic primitive and minimum fallback semantics. It does not claim that every institution lacks escalation, that every unmatched input deserves public attention, or that one centralized exception office is the correct implementation. Its strongest empirical claim is narrower: ordinary object classes and mandate-specific processes can leave a no-match condition without a stable, owned, auditable state.

---