The fourth page of the institutional archive no longer records only events; it begins to record people: "reliable user rating" becomes a new field. A standard compresses recurring problems into judgments that can be held in common. Once dimensions, procedures, times, and minimum conditions are made explicit, people need not renegotiate from scratch on every occasion. Precisely because a standard reduces uncertainty, it also easily acquires authority beyond its original problem: a person who does not pass under a certain standard gradually comes to be described as unfit overall.
Imagine a fictional community tool room. Residents may reserve the cutting, grinding, and measuring equipment. The managers set entry rules: complete the basic orientation, demonstrate the steps for shutting down the equipment, and log any abnormality after use. The rules initially existed to reduce equipment misuse and handover risk.
After half a year of operation, the tool room adds further scores: speed of completion, tidiness of the workspace, frequency of asking questions, and whether one takes part in community events. The scores are combined into a "reliable user rating" that determines booking duration, equipment access, and eligibility to attend meetings. A standard system that once handled a specific risk begins to evaluate members' cooperative attitudes and community value.
This page of the archive opposes neither standardization nor the safety rules of real equipment. The tool room and all its metrics are a thought experiment. What the cross-section is meant to show is how a standard's object, evidence, consequences, and period of validity can remain aligned; and how, when it extends from behavior to personality and from minimum conditions to the whole of opportunity, perfect obedience acquires a quantified appearance.
What a Standard Must First Answer
"Being able to shut down the equipment safely" addresses a concrete operational problem. "Completing projects quickly" describes output performance. "Being willing to help others" evaluates cooperative contribution. "Being worthy of trust" may become a still broader judgment of personality. The four can be correlated, yet they cannot substitute for one another.
That someone operates slowly does not mean he does not know the safety steps; that he often asks questions does not mean he lacks responsibility. Conversely, a person with fluent movements may still neglect the log. If multiple objects are to be combined into a single rating, one must explain why they are combined and what decisions the combined result can support.
A minimum standard usually answers whether one may enter a given activity: for example, only after one can identify the necessary steps may one use the equipment independently. A comparative evaluation answers how limited opportunities are to be ranked. The two institutions rest on different warrants: those who pass the minimum line may all possess basic qualification, yet still need to book because resources are finite.
Resource allocation in turn does not automatically yield personal worth. Failing to obtain a peak-hour slot may be the result of a lottery or of rotation, not an indication that the participant is inferior. If managers package allocation outcomes as ability evaluations, contingency and institutional choice enter into identity.
RC's limited order regards rules as a collection of certainties within a particular social cooperation. "Limited" requires that a standard state which kind of interdependence it handles, rather than automatically covering the whole of members' lives merely because it belongs to order.
Observable Behavior and Proxy Indicators
Wanting to know whether members hand over conscientiously, the managers use "whether the log is complete" as the indicator. The log is checkable, yet it is only one manifestation of conscientious handover. Someone promptly explains an urgent abnormality verbally but forgets to fill in the form; someone fills in every column while the actual content is of no help to the next user.
Proxy indicators are not deception. Complex qualities often cannot be observed directly and must be approached through limited manifestations. The problem is that the relation between indicator and target should be subject to the test of consequences, and must not permanently represent the target merely because it is convenient to record.
When scores carry important opportunities, participants learn to satisfy the visible requirements. That forms grow ever tidier does not show that handover information has become more effective. If managers look only at completion rates, the standard may shift from measuring the activity into the object the activity pursues.
For this one cannot solely blame members for "gaming the rules". If the institution rewards form and ignores actual consequences, acting according to form is predictable. Responsibility falls simultaneously on indicator design, resource pressure, and individual choice, and should not be pinned only on those being evaluated.
Improvement requires keeping itemized results and raw material. When it is discovered that the log is complete yet the malfunction was still not conveyed, one should be able to revise the indicator rather than use a high completion rate to deny the malfunction. Standards serve their object; they should not require the object to serve the score.
Why the Same Rule Produces Different Burdens
The tool room requires everyone to pass an online text-based test. The rule is uniform in form, yet it may let those familiar with the terminology finish more easily. Another resident operates stably in practice but reads slowly. Whether the test measures equipment knowledge, or simultaneously measures some capacity for expression, requires examination.
A difference in form does not automatically prove unfairness. A text test may conveniently cover important information and leave a record that can be re-examined. The question is whether abilities unrelated to the target have become an unnecessary threshold, and whether the same target could be checked instead by demonstration, verbal explanation, or supervised practice.
Alternative methods also add cost; one cannot demand that any standard impose no burden on any individual. The reasonable requirement is that those who set the standard state which burdens are necessary for risk control and which merely continue what is convenient, and that they reassess when concentrated exclusions appear.
"Everyone is the same" shows only that the wording of the rule is uniform; it cannot complete a judgment of consequences. "Everyone is different" likewise cannot prove that common standards are impossible. Institutions must build a connection between common objects and relevant differences.
A common rule has value because it lets participants know what conditions suffice. If every person were judged by the manager on the spot, however nominally more flexible this appears, preferences might become harder to examine. Critique of rigid standards therefore cannot simply retreat into unfettered discretion.
The Evidential Basis of a Standard Expires
After the tool room replaces its equipment, the old test still asks about steps that have been canceled. The managers hold that those who passed the old test have performed well over a long period, so the standard need not change. Past correlation is treated as permanent validity for the future.
Environment, tasks, and participants all change. That a standard was initially valid does not show that it will need no re-examination later. Conversely, the existence of change does not mean that all rules should immediately be abolished. One must see whether the change affects the relation between the standard and its target.
A standard can set a re-examination date, trigger conditions on anomalies, and version records. New equipment, repeated misjudgments, or changes in important consequences can all become reasons for re-checking. Without trigger conditions, revision may occur only when it suits those in power.
Modifying a standard must also handle the dependencies already formed. Members who completed training under the old rules cannot be declared never to have qualified without any transition; yet the old qualification also need not permanently cover new risks. Supplementary explanation, a bounded transition period, or renewed demonstration are all options requiring concrete comparison.
Processual completeness requires that cognition retain its value through continuous correction. For standards, revision is not an admission that all past judgments were meaningless, but a statement of the conditions under which the old judgments are no longer sufficient.
An Exception Does Not Mean the Rule Has Failed
A visitor with relevant experience arrives for a short period without completing the regular course. Does the tool room allow him to work under supervision? If the rule specifies alternative proofs and conditions of supervision, this is not necessarily a breach of fairness but another path that handles the same risk.
Exceptions can reveal situations the initial rule omitted, and they can also become privilege for acquaintances. Judgment must look at whether the reason for the exception is related to the target, whether similar cases can receive similar treatment, whether consequences are recorded, and whether the general rule accordingly needs revision.
Strictly forbidding any exception may reduce discretionary bias, yet it may cause the standard to produce unnecessary blockage in unforeseen situations. Allowing managers unlimited exceptions makes the written rule bind only ordinary members. Neither end can be settled by the word "fairness" alone.
The right of exception especially requires scope. The authority to trigger an emergency stop of equipment does not naturally extend into exemption from all of one's own logging; professional experience may substitute for part of the introductory training, yet it does not automatically substitute for knowledge of the specific local rules.
A reviewable exception leaves its reason and its term, without unduly exposing private sensitive information. Publicity of principle and protection of individual boundaries can coexist; the degree of specification should be proportionate to the consequences.
From Competence to Excellence: What Happens to the Scale
The tool room at first judged only whether one could work independently; later it began to select the "best member". Excellence can encourage contribution, yet it must state what it evaluates. Completing many projects, helping newcomers, saving materials, and trying new methods may conflict with one another.
If all values are compressed into a single total score, the weights complete the choice of values on the participants' behalf. Can high speed offset one serious omission in handover? Does time spent helping others lower one's own output? These questions do not receive answers merely because a computation exists.
Multi-scale evaluation lets different contributions remain separately visible, without requiring everyone to compete in the same direction. It may still need to choose among limited rewards, but the chooser should state which goals take priority, rather than claim that the total score has discovered a person's objective value.
RC's critique of evaluation alienation targets precisely the degeneration of evaluation power from public deliberation into one-way certification. But multiple scales can also be misused: the more indicators there are, the wider the information managers collect, and the harder members find it to know which behaviors are being judged. Adding scales does not equal adding the power of subjects. Whether evaluation therefore improves depends on whether participants can understand, question, and influence the scales, whether the underlying differences remain inspectable, and whether the results are used only within the agreed scope.
How Standards Expand into Personal Standing
The managers claim that, since the safety standard is accepted by all, the tool room's complete rating system also represents common consensus. Those who oppose activity points are described as wanting only to enjoy the equipment while refusing shared responsibility. Consent to the rules of minimal risk is converted into authorization of the entire evaluation power. This inference borrows the real necessity of cooperation. Equipment use does affect others, and members cannot invoke individual freedom to refuse all common constraints. But "some safety standard is needed" supports only the corresponding scope; it does not prove that every tool-room-related requirement is necessary.
Expansions of authorization often proceed through the continuity of words. Safety requires reliability, reliability requires active participation, active participation requires identification with community values, and thus operational steps finally connect with attitudes of mind. Each step seems weakly related, yet the accumulated result departs far from the initial object.
Theoretical dimensional reduction requires that interpretation mark its boundaries of application. Standards should undergo the same review: proving that a behavior is related to equipment risk does not prove that it is related to personality, loyalty, or eligibility for all resources.
The critique cannot merely say that expansion proceeds from malice. Managers may sincerely hope to establish a culture of responsibility. Intentions can bear on judgments of responsibility, but they cannot answer for consequences and for the boundaries of authorization.
A member is labeled low-reliability for one tool-room violation. When he applies to join the community garden, the managers share the rating, on the ground that reliability has universal significance. Records from a specific scene begin to decide entry into other activities.
Cross-domain use is sometimes reasonable. Certain serious and relevant behaviors may affect multiple shared spaces, but the relevance must be stated. A mistake in operating one piece of equipment cannot be automatically connected to tending plants, joining discussions, or receiving basic community services merely because all of these belong to "public activities".
Dark structures lower the cost of description through a global identity. Once a person is classified as "unreliable", the various departments no longer check the material of their own domain. The original label replicates itself across scenes; opportunities in new scenes diminish, which in turn makes it difficult for him to produce contrary records.
Data sharing also enlarges the initial consent. That members submit information in order to book tools does not mean they consent to its becoming a personality file of the community. Even if the data are accurate, the change of purpose still requires justification and appropriate boundaries. Deleting the label is also insufficient for full repair. Opportunities, relationships, and time lost to the label may not return automatically. An error-correction institution needs to distinguish the correction of records from the remedy of consequences.
The public rule states that completing training suffices to book, while the backend keeps an undisclosed threshold: asking too many questions, having raised objections to rules, or cooperating with low-rated members all lower one's ranking. Participants act according to the public standard yet still cannot know why their opportunities shrink.
Some systems need to prevent manipulation and cannot disclose every detail of safety checks. But the principal objects affecting basic eligibility, the channels of appeal, and the means of correcting errors cannot all vanish under "confidentiality". The scope of secrecy needs to be proportionate to the specific risk.
Invisible thresholds hand the burden of prediction one-way to members. The manager retains a margin for interpreting outcomes after they arrive, while members can only keep adjusting themselves. If he asks the reason, the asking itself is recorded as insufficient cooperativeness, and the standard forms a self-confirming loop. Such a structure can produce perfect obedience without issuing any explicit command. To avoid unknown punishments, people may reduce questions, relationships, and attempts. Control derives from the persistent presence of uncertain consequences, not from every behavior being directly forbidden.
In critiquing this path we cannot offer an operational manual for optimizing hidden control. What needs to be exposed are the identifiable conditions: public standards long at variance with actual outcomes, appeals unable to obtain objectified reasons, questioning itself entering into punishment, and thresholds never allowed to be proven wrong.
The tool room administrator can verbally allow an acquaintance to skip booking while requiring ordinary members to comply strictly. He explains that he must handle matters flexibly according to the whole, and the whole is invisible to ordinary members. The right of exception gains permanent exemption from inspection through informational advantage.
A managerial position does require some discretion. Temporary equipment failure, urgent public projects, or sudden gaps in the schedule cannot all be covered in advance by fixed rules. The question is whether the discretion revolves around the target, whether it leaves a record, and whether it accepts subsequent review.
Dark usage declares any review a weakening of efficiency. When an exception succeeds, it proves the administrator's acumen; when it fails, the cause lies in the unpredictability of special circumstances. No outcome can reduce his authority. Meanwhile, reasonable deviations by ordinary members are treated as disciplinary problems. The upper and lower levels use different moral vocabularies for uncertainty: at the top it is called flexibility, below it is called violation; at the top, strategic adjustment, below, lack of stability. Language makes a difference of authority look like a difference of ability.
Power hierarchy does not say that every asymmetry should disappear; it requires that evaluation power remain connected with responsibility. Those who can make exceptions should bear a higher duty of justification, not merely acquire more room for action.
A member made many mistakes while learning early on, and his operation has since become stable. The rating system still ranks him by the average of his entire history, and he remains below newcomers with fewer records. Past data prolong the force of the old state.
History has predictive value and cannot be deleted entirely merely because the person concerned wishes to start over. Serious and relevant problems may require a longer period of observation. But different records lose representativeness with time, learning, and changing circumstances; their weights and periods of validity should have reasons.
If the errors of high-rated members are explained as accidental while the successes of low-rated members are explained as temporary, the data serve only to confirm existing identities. The same event acquires different meanings according to its label, and the standard no longer evaluates independently.
A darker structure would use the allocation of opportunity to maintain difference. Low-rated members can use only simple equipment and therefore have no chance to demonstrate complex ability; the lack of demonstration in turn becomes the ground for continued restriction. Risk control and the freezing of identity need to be distinguished through explicit advancement paths.
Allowing the renewal of qualification does not guarantee that anyone will finally pass. It requires that new material be able to change the conclusion, and it states what material is needed. A perpetual "continue observing" without conditions of completion has the same dominating effect as a secret threshold.
The managers say the ratings merely reflect members objectively and should not be questioned — as if, once indicators, weights, and data have been computed, their makers withdrew from the result. The standard acquires the appearance of speaking for itself. In fact, choosing what to record and what not to record, deciding the time window and the handling of errors, all contain judgment. Pointing out the process of construction does not declare the result false; it returns responsibility to the positions of making and using.
If RC's concept of the convergence of observation is used to reinforce a standard, a dangerous claim can result: the mass of data has already formed a stable consensus, and the score is the member's objective reality. Here machine aggregation, social recognition, and multi-level observational consensus are conflated.
That participants object to a score does not make all behavioral facts disappear; that a standard gains majority acceptance does not make the omitted consequences cease to exist. Evaluation can be a real and useful secondary construction, yet it must still be tested under specific targets.
If a standard forever classifies counterexamples as anomalies of the object, it will never be corrected by the object. A member successfully completes a task deemed impossible, and the system says this was accidental; failure then proves the rating correct. So-called objectivity becomes the name for a model's refusal to learn.
An Anti-Standard Stance Can Also Produce Private Privilege
Before discussing the anti-standard stance, we must also address conflicts among standards. The tool room simultaneously requires members to shut down the machine whenever an anomaly appears, to complete projects on time, and not to waste reserved slots. When a noise that is hard to judge appears, the member may violate one of the requirements whether he shuts down or continues.
Multiple goals are unavoidable in reality, and conflict does not prove that the institution deliberately sets traps. But those who make the rules should state the relations of priority, permit the reporting of uncertainty, and let subsequent evaluation consider the information available at the time. If which rule matters more is announced only after outcomes appear, members cannot act on it.
Dark usage deliberately or persistently retains requirements that cannot be satisfied simultaneously. Whoever shuts down is called inefficient, whoever continues is called careless of safety, and whoever asks for instructions is called lacking in judgment. Any choice can produce a negative record, and the evaluator thus permanently retains grounds for punishment. This structure is harder to rebut than the secret threshold, because each standard looks reasonable on its own. Domination occurs in the complete upward concentration of the power to handle conflicts: the evaluated must bear the prediction in advance, while the evaluator may select afterward the most advantageous item from the set of standards.
To recognize it, one must examine how similar conflicts were handled, whether priorities change with persons, whether voluntary reporting is genuinely protected, and whether the evaluator's decisions leave records. A single dilemma proves no control; a repeatedly one-way use of dilemmas constitutes important structural material.
Repair does not require writing out a mechanical ordering for every case. The institution can make safety bottom lines explicit, prescribe provisional authority for when conflicts arise, and improve afterward in light of new material. The key is that members have a path of good-faith action that does not necessarily bring punishment.
Some criticize all standards for compressing the individual and demand that each time they themselves explain whether they qualify. This may protect difference, but it may also give those with the greatest expressiveness, relationships, or time a larger space. The absence of a common line does not mean the absence of power.
The flexible judgment of an acquaintance society can attend to concrete circumstances, yet it may also leave newcomers not knowing how to obtain qualification. Written standards at least provide predictable conditions; the question is whether they agree with actual enforcement.
Others emphasize exceptions only when the standard disadvantages them, and demand strict uniformity when they hold the advantage. Critique should check symmetrically: whether similar cases are handled by similar reasons, and whether differential treatment can be shown to relate to the target.
That standards cannot express all values does not mean that every comparison is illegitimate. Shared equipment is finite and must be scheduled; high-risk operations may require qualification. Abolishing all judgment would hand the cost over to on-the-spot conflict and informal status. What this chapter therefore pursues is standards with scope, reviewability, and revisability, not the absence of standards. The support that rules give to the capacities of subjects also needs to be preserved.
What a Qualification Decision Should Leave Behind
The tool room can separate its evaluations: safety qualification rests only on demonstrations relevant to operational risk; booking follows public rotation or explicit priorities; contribution records serve the corresponding thanks, without automatically altering basic eligibility; community discussion does not let tool ratings decide the right to speak. This is not the only institutional scheme; it merely shows that different authorizations need not be merged. If a record genuinely requires cross-domain use, its makers should state its object, term, access, and error correction, rather than rely on the blanket word "reliability".
When a member does not pass, he should know which item he failed, on what material, and how to demonstrate it anew. For details that cannot be disclosed, procedures sufficient to correct errors should still be provided. If a standard gives only conclusions and no path that could change the conclusion, it comes closer to a verdict of identity.
Exception decisions need to leave their scope and reason, and the manager himself remains bound by the core rules. Emergency discretion may act first, but it cannot cancel the subsequent record. Review does not demand that every outcome be reversed; it ensures that errors can produce actual change.
The relation between standard and target should be checked periodically, with particular attention to misjudgments that are few in number but concentrated in consequence. A good overall accuracy rate does not mean that individual harm needs no remedy, nor that every minority objection is correct.
Standards Cannot Complete Value Judgments for People
Return to the tool room. Standards can help everyone share dangerous equipment, protect handover, and schedule resources. This coordinative capacity deserves recognition in itself. The problem arises when standards move from finite problems to infinite personality, while their makers no longer admit that they have made choices of value.
Speed, tidiness, participation, and questioning can all be meaningful in certain tasks. They do not naturally combine into a person's public value. Even when a composite decision must be made, one should state why certain scales are preferred here, and when the decision lapses.
Perfect obedience is not everyone acting according to one clear rule. It appears more plausibly where the boundaries of rules dissolve, exceptions open only upward, actual thresholds are unknowable, and every counterexample reinforces the original rating. A person is continuously rewritten under each local standard, until finally he can prove himself qualified to evaluate only through conformity to the evaluation.
To resist such a structure does not require retreating from common life into immediate struggle. It requires keeping standards in their instrumental identity: serving definite objects, using relevant evidence, bearing actual consequences, and preserving entries for revision. The more a standard can determine people's opportunities, the more the positions that make and interpret it need to accept public constraint.
A qualified system can say "this operation has not yet passed" without readily saying "this person does not deserve to be trusted"; it can say "this exception has these reasons" without declaring that the manager naturally sees the whole; and it can admit that a certain number is useful while refusing to let the number settle in advance all of a person's possibilities.
A single pass should likewise not be interpreted in reverse as the member's consent to all future changes of the standard. A qualification confirms the version then in force and its corresponding object; once the rules are updated, the organization must state anew the added obligations. Otherwise "you once qualified" becomes evidence that "you already obeyed in advance", and the time boundary of the standard vanishes once more.