System, Authority and Institution

T7 is the seventh independent theory document in the T1–T9 programme. It examines systems, organisations, institutions, mandates, competence, authority, power, legitimacy, decisions, and correction as arrangements within actuality.

T1–T5 established the ground, conditions of knowledge, and the problem of representation. T6 distinguished the person from role, status, and category. T7 carries these distinctions into arrangements that distribute tasks, store knowledge, make decisions, and implement action over time.

The document does not by itself determine whether a particular decision is lawful, professionally correct, or legitimate. Such assessments require the proper sources, jurisdiction, case material, adversarial participation, and competent review.

Actuality does not depend on a system in order to be. A system depends on the people, conditions, material, and purpose it orders.

A system may be formal or informal, technical or social, legal or administrative. It may coordinate, store, standardise, control, and correct. These functions make it real and consequential, but not self-grounding.

A system must remain testable against its stated purpose, mandate, actual operation, rights, effects, and mechanisms of correction.

An organisation is a concrete association of people, roles, resources, procedures, and technical systems. An institution is a more durable social, legal, or professional arrangement with a recognised function.

An organisation may administer an institutional function without being identical with the institution. Criticism of a particular organisation is therefore not automatically criticism of the institution as a whole, and defence of the institution is not automatically defence of every organisation acting in its name.

The distinction prevents institutional legitimacy from being transferred to practices that do not hold, and prevents isolated failure from becoming an unwarranted total judgment of the whole order.

A system must be able to state why it exists, who established it, what task it holds, and what lies outside that task.

A mandate is the formal or expressly assigned task and scope of action of an actor, role, or institution. It must be traceable to law, decision, agreement, delegation, or another valid assignment.

T7 distinguishes stated purpose, actual function, measured result, and actual effect. These may coincide but may also diverge.

• Who granted the mandate?
• What does it cover?
• What lies outside it?
• Who may alter it?
• Who is affected by its exercise?

Competence is the relevant capacity or authority to perform a task or assess a question. It may be professional, legal, procedural, technical, or experiential.

Jurisdiction is the matter, person, place, or professional domain within which authority applies. Competence without the proper jurisdiction does not automatically confer decision power, and legal competence does not confer professional omniscience.

Institutional precision requires visibility about what competence is claimed, what grounds it, what the question requires, and where the boundary lies.

Authority is a recognised or formally assigned reason for being heard, followed, or having decisions accepted within a bounded domain. It may rest on competence, mandate, law, role, trust, or documented integrity.

T7 distinguishes epistemic, professional, legal, institutional, moral, and practical authority. These forms may overlap but must not be transferred automatically from one domain to another.

Authority may give reason to give weight to an assessment or comply with a decision. It does not make the statement true without correspondence and does not make the office-holder incorrigible.

Trust is necessary because no actor can independently inspect every source, method, and decision. Grounded trust may rest on competence, open procedure, documentation, stable practice, independent control, correction history, and visible conflicts of interest.

Trust must not become a barrier to questions. Authority may reduce the need to repeat the entire inquiry, but cannot remove the right to ask what the assessment rests on and where the competence boundary lies.

Where authority itself is used as proof, the source role must be distinguished from the factual ground.

Power is the actual capacity to influence, govern, compel, or allocate. It may be legal, economic, organisational, social, technological, informational, physical, or symbolic.

Power, office, authority, legitimacy, and coercion are not identical. Lawful coercion may be legitimate within mandate and safeguards, while informal power may be extensive without a public responsibility line.

Use of power must be assessed by legal basis, purpose, necessity, proportionality, alternatives, documentation, actual effect, and available control.

Legitimacy is the grounded rightfulness of an arrangement, role, procedure, or decision within its domain. It cannot be reduced to popularity, effectiveness, or formal validity alone.

Relevant grounds may include lawful establishment, democratic mandate, professional competence, procedural fairness, factual ground, safeguards, public reasons, actual effect, and corrigibility.

A process may be lawful but professionally weak. An assessment may be professionally strong but made by the wrong body. An action may be effective but disproportionate. Legitimacy therefore requires a bounded, multi-part assessment.

Institutions use statutes, regulations, guidelines, standards, procedures, practice, custom, and professional discretion. Each form has a different status and degree of binding force.

Rule-governance may strengthen equality and predictability. Discretion allows attention to the concrete case within mandate and competence. Arbitrariness arises where the decision cannot be traced to a relevant rule, material, or assessment.

Discretion is not arbitrariness, and mechanical rule application is not automatically correct application. A standard may hide that a case does not fit the pattern it assumes.

An institutional decision is produced through registration, delimitation, collection of material, adversarial participation, assessment, reasons, decision, notice, implementation, and control.

The procedure should show who did what, what material was available, what was omitted, what rule or standard was applied, what uncertainty remained, and who held competence.

Correct procedure does not always guarantee a true outcome. Without defensible procedure, however, the knowledge base, safeguards, and legitimacy are weakened.

Delegation is the transfer of task, authority, or decision competence from one actor to another under a documented ground. Delegation of work, assessment, decision, and implementation must remain distinct.

Tasks may be delegated without automatically removing responsibility for mandate, control, provider selection, and actual effect. An external expert opinion or technical supplier does not by itself remove the commissioning body’s responsibility.

“The system decided” is not a sufficient account of responsibility where identifiable people or bodies designed, selected, approved, or used the system.

• Who selected the solution?
• Who could override it?
• Who approved it?
• Who implemented it?
• Who could correct it?
• Who bore final responsibility?

Institutions may carry knowledge through archives, registers, records, databases, procedures, professional communities, and documented decisions. This knowledge may exceed what any one person knows.

Institutional memory may also preserve old errors, lost context, outdated premises, misclassifications, and copied suspicions. Repetition across systems does not establish independent sourcing.

Traceable governance requires provenance, version, access, responsible ownership, correction, and withdrawal so that superseded information does not continue to govern new decisions.

Reasons are the visible account of the question decided, facts and rules relied upon, treatment of counter-material, exercise of discretion, and why the outcome follows.

A template or automatically generated text is insufficient where it does not show the actual basis of the decision. Reasons make an assessment reviewable without requiring reconstruction of every internal mental process.

Reasons are not only information for the affected party. They are also a control point for the body itself, the review authority, and public scrutiny.

Adversarial participation means that an affected party has a real opportunity to know the relevant material, correct factual error, submit counter-material, and have the response considered before or through competent review.

It does not mean that all claims receive equal weight. Material must still be assessed by relevance, authenticity, strength, and relation to the matter.

Participation without access to decisive premises, sufficient time, or capacity for actual correction may be formal without being real.

Systems need arrangements for professional disagreement, internal incident reporting, whistleblowing, party objections, and public criticism. These forms have different legal and institutional grounds.

Criticism must not automatically be reclassified as disloyalty, collaboration difficulty, or risk. At the same time, disagreement must be made concrete and handled with regard to source, confidentiality, safety, role, and rights.

A system unable to receive relevant internal criticism loses a central channel of correction and increases the risk of closure.

Reconsideration, appeal, oversight, audit, judicial review, and peer or expert review have different mandates. They must not be conflated or described as stronger review than their competence allows.

Effective control requires sufficient independence, access to material, relevant competence, real power to change, and protection against retaliation.

An explanation without the capacity to change erroneous premises or outcomes is not always effective review. At the same time, no control arrangement is necessarily tasked with reviewing every dimension of a matter.

Openness strengthens review but is not absolute. Privacy, confidentiality, source protection, security, investigations, trade secrecy, and protection of children or third parties may require limits.

Limits should be reasoned, legally grounded, proportionate, time-bounded where relevant, and subject to control. Lawful protection must be distinguished from closure used to avoid responsibility.

Lack of public access does not by itself prove error or abuse. It limits what outsiders can establish and increases the importance of the quality of control arrangements.

Systems are often governed through output targets, indicators, budgets, rankings, deadlines, risk scores, and rewards or sanctions.

Metrics may make responsibility and development visible, but may also substitute for purpose. When indicators become governing objects, the organisation may optimise reporting rather than actual effect.

T7 therefore distinguishes metric from purpose, indicator from what it indicates, and reported result from actual effect.

Institutions use case-management systems, rankings, risk assessment, document generators, surveillance, artificial intelligence, and external providers. These systems enter decision chains and must remain auditable.

Responsibility must be visible for procurement, purpose, data, model choice, logging, override, appeal, correction, and supplier changes.

A human reviewer is not real control where the person lacks access, understanding, time, power to alter the result, or protection against sanction for departure.

System closure arises where an arrangement protects its own representation or decision against relevant correction. This may occur through definitions, procedure, role allocation, data storage, appeal structure, or performance targets.

Indicators may include counter-material being made irrelevant, criticism reclassified as disloyalty, errors unable to change registers or practice, review bodies copying the same premises, or no actor owning the decision.

A single error does not establish system closure. The claim requires a documented structure or pattern in which relevant correction cannot operate.

System capture is a bounded condition in which the system’s own frame, procedure, classification, or self-maintenance takes governing priority over the purpose, material, or actual condition the system was meant to answer to.

Institutional capture applies the same structure where it has become embedded in a durable institutional arrangement or social function.

A strong finding requires an identified system, identified purpose or actuality relation, documented substitution, governing priority, resistance to correction, and concrete consequence.

• The system or frame is identified.
• The purpose or actuality relation is identified.
• The substitution is documented.
• The systemic priority is governing.
• Relevant correction is resisted or neutralised.
• A concrete consequence is documented.

Authority capture is a bounded condition in which a person, title, office, or institutional status governs assessment in place of the relevant ground the authority was meant to carry.

It requires more than following an expert or lawful decision. It must be shown that status replaces relevant testing, competence is transferred too far, or counter-material is rejected because of the source’s role.

The claim must be connected to actual decision effect and resistance to correction, not merely strong respect or hierarchy.

Institutional correction includes correction of individual information and decisions, and change to process, system, data, rule, responsibility structure, and future practice.

Correction must be distinguished from symbolic response. A new routine or software version is not by itself proof that the failure has been corrected.

It must be shown what changed, why, how the change was tested, what earlier matters were affected, and whether actual effects improved.

Crisis may require rapid decisions, limited access, temporary concentration of power, or simplified procedure. Such departures must not be concealed as ordinary practice.

Exceptions should be expressly identified, legally grounded, necessary, proportionate, time-bounded, documented, and reviewable afterwards.

Temporary power must not become permanent merely because the system was built or expanded during crisis.

T7 must not protect itself from criticism through its own concepts. The objections below identify points at which the theory may become too strict, too weak, or self-referential.

T7 enters discussion with social and political philosophy, rule-of-law theory, bureaucracy, professions, institutional and organisational theory, social epistemology, governance, critical theory, and technology and algorithmic governance.

The document agrees with institutional theory that arrangements, roles, and statuses are real and consequential. It agrees with legitimacy and rule-of-law traditions that office must be bounded by competence, procedure, and control. It nevertheless maintains that no institutional status can make a representation true or incorrigible merely by decision.

T7 does not derive one form of government from DET SOM ER. It establishes conditions for assessing authority and institutional practice. Full comparison belongs to T9.

T7 requires later documents to keep system, organisation, institution, role, mandate, competence, authority, power, decision, and responsibility distinct.

If T7 is substantively revised, T8, T9, R1, A5, A6, A7, and institution-directed application documents in particular must be reviewed.

• T8: Systems must be assessed by actual social effect and practice.
• T9: The theory of authority and institutions must be compared and tested.
• R1: Methodology must distinguish mandate, competence, process, decision, and effect.
• A5: System and Institutional Audit receives its theoretical ground.
• A6: Authority Audit receives its theoretical ground.
• A7: Practice Audit tests institutional action against purpose and representation.
• D2, D5, D6, and D8: Law, work, technology, and governance receive a common institutional distinction framework.
• R7: Worksheets keep mandate, role, material, decision, and responsibility separate.