Resources / Institutional research

Institutional Pathologies

A Framework for Examination and Correction

Runcible Research · 19 pathologies across six institutional functions · Version 1.0 · 15 November 2026

Contents
  1. Abstract
  2. Scope
  3. Complete matrix: 19 pathologies
  4. Diagnostic distinctions
  5. Power and intent
  6. Enterprise AI liabilities
  7. Completed correction
  8. Implementation implications

Abstract

Institutions can lose control of their purpose while every office continues to follow procedure. A department can meet its target by exporting costs. A committee can prevent an objection from becoming an appealable decision. A faction can gain control of the rules used to judge its own conduct. A collection of individually authorized projects can promise more than the institution can deliver.

This paper presents a diagnostic framework of nineteen institutional pathologies organized into six functions. Each entry identifies a mechanism of failure and the evidence needed to examine it. The inventory supports diagnosis and repair; it does not establish the prevalence of a pathology in any particular organization or claim to exhaust every possible institutional failure.

Runcible examines claims, proposed actions, and their grounds through constructive falsification. Oversing connects institutional work to roles, authority, commitments, records, and consequences. Their combined relevance extends beyond checking the final AI answer to examining the conditions under which matters reach a decision, authority is exercised, and correction becomes possible.

What the framework distinguishes

The six categories identify the institutional function that fails. The nineteen pathologies distinguish how it fails. A single episode can exhibit several mechanisms, and an observed loss can have more than one cause. Classification directs an inquiry toward discriminating evidence; it does not replace that inquiry.

A pathology can persist without dishonest participants. A reward can favor harmful conduct, a handoff can leave an obligation ownerless, or separately reasonable commitments can compete for the same capacity. Replacing an individual does not repair an arrangement that gives the next person the same incentives, permissions, and information.

A departure from an institutional purpose must be examined against the institution’s actual mandate and applicable obligations. Political, professional, or cultural commitments are not themselves proof of misconduct. The inquiry asks whether the exercise of discretion is warranted by the governing purpose and whether the institution applies the standards it claims to apply.

The nineteen-pathology matrix

The entries are arranged by their primary institutional function. Their numbering is an index for this edition, not a ranking of severity.

1. Purpose and incentives

When people succeed by the institution’s measures and rewards, do their actions advance its purpose after costs to others and consequences over time are counted?

PathologyHow failure occursWhat must be examined
1. Myopia and cost shiftingA unit improves its own results by transferring costs or risks to another unit, an affected party, or a future period. Purchasing earns credit for a lower price while maintenance absorbs a higher lifetime cost.Whose objective improved? Who bore the additional cost, over what period, and did the decision still advance the institutional purpose after those costs were counted?
2. Goal displacementStaff learn to produce the target, complete the procedure, or satisfy the audit while the underlying purpose goes unmet. A service desk meets its closure target by closing cases customers must reopen.What outcome was the measure meant to indicate? Can staff improve the measure while degrading that outcome? Do records of subsequent events reveal the divergence?
3. Malincentives and moral hazardThe institution rewards conduct whose foreseeable costs fall elsewhere. A sales team earns commission at signature while delivery bears the cost of promises made to close the sale. Replacing the salesperson leaves the inducement intact.What conduct earns the reward, when does it vest, and who bears the downside? Can the party carrying the risk refuse or reprice it? How does the reward compare with the value ultimately realized?

2. Information and judgment

Can responsible parties obtain the evidence, see material disagreement, and apply a consistent standard to comparable claims?

PathologyHow failure occursWhat must be examined
4. Information hoardingA person or office preserves its indispensability by controlling access to records, procedures, credentials, or case history that others need to perform their duties.Who depends on the information? Who controls access, on what grounds, and through what route can a legitimate user obtain access or challenge a denial?
5. Information distortion and preference falsificationParticipants sanitize forecasts, omit objections, or publicly assent to judgments they privately dispute. The resulting record can be accessible to everyone and still manufacture the appearance of consensus.What did the original evidence, forecasts, and objections say? Who revised them and why? Could participants register dissent before the outcome was known, and did material disagreements survive into the decision record?
6. Selective judgment, including political biasA decision-maker changes the burden of proof, the weight assigned to harm, or the availability of an exception according to whose interests are at stake. The preference may be political, departmental, professional, or personal.In materially comparable cases, would the reasoning survive an exchange of the parties’ identities or affiliations? Where treatment differed, did a recorded difference warrant it under the institution’s mandate?

3. Distribution of resources and power

Who receives resources, who obtains offices, who sets the rules of allocation and succession, and what causes delegated powers to expire?

PathologyHow failure occursWhat must be examined
7. Resource captureA person, faction, or supplier directs institutional money, labor, assets, or opportunities toward private objectives under an institutional justification.Who proposed and authorized the allocation? Who benefited? What institutional return was promised, what alternatives were available, and what return was realized?
8. Factional advancement and patronageA coalition advances its allies through appointments, promotions, assignments, invitations, or disciplinary exceptions. Each appointment can increase the coalition’s influence over the next one.Who nominated, selected, promoted, and granted exceptions? What criteria were announced beforehand? How did appointees compare with eligible alternatives? Shared affiliation alone cannot establish favoritism.
9. Institutional captureA coalition acquires durable control over appointments, rulemaking, oversight, appeals, or succession. It can determine the standards by which its own conduct is judged and reproduce its authority after its founders leave.Who can change the mandate, appoint successors, select reviewers, hear appeals, and reverse decisions? Can an affected party obtain review independent of the coalition whose conduct is disputed?
10. Mandate creep and persistence of exceptional authorityAn office receives power for a bounded purpose or exceptional condition. The condition passes, but the permission persists. Authority accumulates through successive delegations even without factional coordination or private gain.What authorized the power, what activates it, and what terminates it? When its warrant expires, which permissions, rules, and standing decisions must be withdrawn or reauthorized?

4. Access to decision and review

Which matters receive scrutiny, which wait or disappear, and who can compel a timely decision or challenge an intervention?

PathologyHow failure occursWhat must be examined
11. Bureaucratic obstructionAn office preserves its role by making its assent indispensable. It can delay, return, or veto work at little cost to itself while applicants bear the cost of waiting. The office’s interventions then serve as evidence that the office is needed.What defect did the intervention identify? What would resolve it? Who bears the cost of delay, who can settle the matter, and which interventions demonstrably prevented a loss?
12. Agenda suppression and strategic inactionA matter is defeated before it receives an appealable decision. A complaint goes unregistered, a proposal never reaches the committee, or an application remains pending until the opportunity expires.What was submitted or raised? Who controlled admission and scheduling? What deadline or opportunity passed, what disposition was recorded, and who could require a decision in time for it to matter?
13. Selective enforcementComparable cases receive consistent judgments once heard, but some parties are routinely inspected, reported, or referred while others rarely face scrutiny. An audit of decided cases can therefore find perfect consistency.Of all materially comparable cases eligible for scrutiny, which were examined, by whom, and on what grounds? Where the eligible population is unknown, what conclusion can the available evidence support?

5. Work, coordination, and capacity

Does work serve a necessary end, does responsibility transfer successfully at each handoff, and can the institution fulfill its commitments together?

PathologyHow failure occursWhat must be examined
14. Expansion of workWork expands to fill available time, staff, or budget. Once an activity recurs, its volume becomes the justification for staffing it.What outcome warrants the activity? If its frequency, staffing, or scope were reduced, what obligation would go unmet or what measurable loss would follow?
15. Coordination and handoff failureEach participant performs an assigned duty, yet work falls between roles. The sender considers transfer complete on transmission. The receiver assumes responsibility only on acceptance.What event transfers responsibility? Did the recipient accept it? Until acceptance, who owns the matter, and who can see that an obligation remains outstanding?
16. Overcommitment of shared capacitySeveral individually warranted decisions make incompatible claims on the same people, funds, assets, or time. Each project is feasible in isolation, but the institution cannot honor all its promises together.At authorization, what capacity was already committed for the relevant period? Can the obligations be discharged together? If circumstances force a choice, who has authority to set precedence?

6. Accountability, correction, and continuity

Who answers for the result, what happens when evidence defeats the original warrant, and does the institution retain the competence to recognize and repair failure?

PathologyHow failure occursWhat must be examined
17. Accountability evasionThe proposer, approver, and executor can each discharge a narrow duty while nobody bears responsibility for the outcome or its correction.Who warranted the premises, who authorized the decision, who undertook execution, and who accepted both the duty and the authority to intervene if the result departed from the warrant?
18. Failure to correctEvidence defeats a premise or reveals a failed outcome, yet the decisions, policies, and programs that depend on it continue by default.What evidence warranted reconsideration? Which standing decisions depend on the failed premise? Who must reopen them, and what observable result would establish that the correction worked?
19. Erosion of institutional capabilityCurrent output conceals a loss of competence. An institution continues receiving acceptable work from an expert, vendor, or AI system while losing the ability to produce, verify, challenge, or recover that work independently.Which capabilities does the institution require itself to retain? Who last demonstrated each one, under what conditions, and can anyone still exercise it when the usual provider is unavailable?

Distinctions that change the examination

Local optimization and goal displacement

Myopia can occur while every measure is accurate: purchasing lowers its costs while maintenance bears a larger lifetime expense. Goal displacement concerns a defective relationship between a measure and its purpose: a service desk improves its closure count while customers must reopen cases. The first requires accounting across affected parties and time. The second requires testing whether the reported measure still tracks the intended result.

Resources, offices, and the rules of succession

Resource capture concerns who gets the resources. Factional advancement concerns who gets the offices. Institutional capture concerns who gains the power to decide both questions in the future. These require different records: allocations, selections, and changes in authority itself. Fully documented and individually authorized decisions can still participate in capture.

Judgment and selection for judgment

Consistent treatment of decided cases does not establish consistent access to scrutiny. Selective enforcement requires a denominator: the comparable cases eligible for examination, including those never referred. Agenda suppression requires a record of matters submitted, delayed, excluded, or allowed to expire. When that evidence is unavailable, the examination must leave the selection question open.

Information access and information integrity

An open record can still contain a manufactured consensus. Information hoarding concerns access. Distortion and preference falsification concern what survives into the record: original evidence, changes to forecasts, qualifications, objections, and the ability to register dissent before an outcome is known.

Permission and joint feasibility

An authorized commitment can conflict with other authorized commitments. Permission to promise work does not establish that the institution can fulfill every promise together. Examination must reconcile the proposed claim with outstanding obligations against the same people, funds, assets, and time.

Traceability and correction

A record can expose a failure without producing a remedy. Suppression requires someone with standing to raise the matter, responsibility to pursue it, and authority to obtain correction. A completed correction requires an observable result, not merely a new policy, task, or explanation.

Power, consequences, and intent

Political or factional advantage can cut across all six categories. Record it as a separate dimension. An observed benefit, a pattern of unequal treatment, and an inference of deliberate coordination are different claims requiring different evidence.

DimensionQuestionExample
MechanismWhat institutional act or omission produced the disparity?Comparable complaints against one faction were repeatedly left off the agenda until their deadlines expired.
Distribution of consequencesWho benefited, who bore the cost, and what comparison establishes the disparity?The faction’s officeholders retained their positions without review. Complainants lost access to a remedy available in comparable cases.
Evidence of intent or coordinationWhat supports an inference that someone sought this result?A contemporaneous instruction to delay those complaints supports an inference of intent. A pattern of delay alone warrants examining the process and its effects.

Repeated affiliation alone establishes neither favoritism nor motive. The inquiry can establish a failure of process or an unequal consequence while leaving the question of intent unresolved.

How institutional pathologies become enterprise AI liabilities

AI introduces new exposures and can amplify existing institutional failures. Compliance exposure, shadow IT expansion, siloed optimization, and unauthorized commitments are useful commercial examples. The six-function framework also reaches failures in agenda access, rulemaking, joint capacity, correction, and retained competence.

Institutional functionEnterprise AI exposureExamination and control requirement
Purpose and incentivesMisaligned optimization and transferred liabilityAI activity, speed, or departmental savings can be rewarded while costs and risks move elsewhere. Test the reward against the institutional result and the interests of affected parties.
Information and judgmentUnreliable evidence and selective standardsGenerated summaries can omit qualifications or dissent, and automated judgments can apply inconsistent thresholds. Preserve source evidence, revisions, objections, and the grounds for different treatment.
Distribution of resources and powerCaptured or persistent authorityControl over model configuration, allocation rules, appointments, or delegation can reproduce an interest’s power. Record who may change those controls, who can review them, and when permissions must expire.
Access to decision and reviewBlocked or selective scrutinyA system can examine only the cases someone allows it to see. Preserve submitted matters, the eligible population, admission and scheduling decisions, elapsed time, and an effective route to review.
Work, coordination, and capacityUnowned handoffs and incompatible commitmentsAgents can create work faster than receiving teams can accept or deliver it. Require explicit acceptance and reconcile authorized tasks with outstanding claims on shared capacity.
Accountability, correction, and continuityUnowned outcomes and loss of capabilityA technically completed task can lack an owner for its consequences. Assign monitoring and correction duties, reopen decisions whose premises fail, and retain the ability to challenge or recover AI-assisted work.

From diagnosis to completed correction

Diagnosis should lead to a specific repair. Changing an incentive, narrowing a veto, restoring an appeal route, withdrawing an expired permission, or reserving shared capacity addresses different mechanisms. A generic demand for more transparency cannot substitute for selecting the relevant repair.

  1. Matter and purpose. Record the proposal, complaint, obligation, or exception, including matters that have not yet received a decision. State the purpose and the applicable scope.
  2. Evidence and affected interests. Retain provenance, material qualifications, dissent, available alternatives, and the distribution of costs, risks, benefits, and consequences over time.
  3. Authority and capacity. Identify the source, scope, conditions, and termination of delegated power. Reconcile the proposed action with outstanding commitments and available capacity.
  4. Disposition and responsibility. Record permission, denial, deferral, or an unresolved question. Assign both execution responsibility and the duty to monitor the result. Preserve a route for affected parties to obtain review.
  5. Outcome and correction. Compare the result with the original grounds. Identify dependent decisions when a premise fails, assign corrective work, and retain the evidence that shows whether the repair worked.

Implications for Runcible, Oversing, and implementation partners

The framework specifies what an institution needs to make examinable. Detecting a particular failure depends on the records actually available. A Decidability Record about a submitted decision cannot establish which eligible cases were never submitted, whether a handoff was accepted, or whether a permission should have expired unless the relevant institutional evidence is supplied.

Runcible and Oversing provide complementary roles in this architecture. Runcible examines the candidate account and exposes defects, unresolved conditions, and possible repairs. Oversing supplies the institutional context and governs the authorized continuation of work. Implementation must establish the relevant records, review rights, delegation boundaries, and corrective workflows for the chosen use case. This paper does not report that all nineteen examinations have been automated or independently validated.

For an implementation partner, the practical deliverable is a workflow whose decisions, omissions, commitments, and consequences remain answerable to institutional purpose. The institution must be able to find the responsible authority, obtain a timely examination, and follow a repair through to an observed result.