Countermeasures and Detection

This section catalogs the mechanisms the source material proposes against the patterns cataloged in §9.1 to §9.4 and the failure modes enumerated in §9.5. Entries are organized by what a mechanism does rather than by which source names it. §9.6.1 to §9.6.4 treat detection: mechanisms that look for a pattern already running. §9.6.5 to §9.6.10 treat structural resistance: mechanisms that change the arrangement of roles, bodies, or exit rights so that a pattern is harder to sustain. §9.6.11 to §9.6.15 treat practice and culture: mechanisms that act on what members are trained to notice and on what the network records about its own failures. §9.6.16 to §9.6.19 hold four mechanisms added after the first draft of this section: three recovered by the sweep described below, and one the extraction had recorded without the target its source states for it. They are grouped at the end rather than interleaved with the functional groups above. Two of the four are detection mechanisms and belong with §9.6.1 to §9.6.4; one is a set of tracked indicators that no source presents as a countermeasure at all; the fourth is an onboarding measure and belongs with §9.6.11 to §9.6.15. §9.6.20 names the governance instruments the extraction records that belong to chapter 03 rather than to this chapter. §9.6.21 states, for every threat in §9.1.2 and every failure mode in §9.5, what the material proposes against it.

Mechanism names are the sources’ own, on the same basis as the threat names in §9.1: they are retained so that a later reader can trace an entry back to the material that produced it. Five limits apply to everything in this section, and the second governs how §9.6.21 should be read.

1. Nothing here has been built or tested. No mechanism below has been implemented in an operating network, and no source reports an outcome from one. Each is a proposal recorded in a red-team exercise conducted against the model’s documentation. The limits stated in §9 apply to this section as they do to the threat catalog, and the direction of the error is not known: an untested countermeasure may fail against the pattern it names, and it may also work against patterns nobody proposed it for.
2. The threat-to-countermeasure mapping is the extraction’s and this chapter’s, not the sources’. In the source documents the countermeasures appear as lists attached to a response’s overall subject rather than as a mapping onto named threats. Five attachments are made at the source level, counting an attachment as source-level only where a source itself states a mechanism against a named pattern or scenario. Three of the five are the inline Prevention: lines beneath the timeline items in p13r01, Response 1; the fourth is the set of responses the worked scenario document states beneath the case it develops; the fifth is a glossary definition that names its own target. Three of the five land on entries in the catalog, and each is marked where it occurs: the language watchdogs of §9.6.3 against the pattern §9.4.9 names Lexical Drift; the three responses in §9.6.1 and §9.6.10 against the scenario treated in §9.3; and the ethical stress tests noted in §9.6.15 against Mission Creep. The other two, at §9.6.16 and §9.6.19, are attached by their source to scenarios that §9.1.2 does not catalog, so neither supports a row in §9.6.21 by itself; §9.7.7 treats the question they raise. A heading can also frame a whole block of mechanisms around one pattern without stating any of them against it individually — p12r01’s third response does this for Mission Creep — and framing of that kind is not counted here. Every other correspondence recorded in §9.6.21 was drawn in the staged extraction, in the primary-source sweep described below, or in this chapter, and the cells state which. A correspondence marked as inferred is a judgment that a mechanism bears on a threat; it is never a claim that a source says so.
3. The polycentric countermeasure material is general governance content. The extraction records that the pass producing the mechanisms in §9.6.20 largely reproduces general public-administration material rather than mechanisms specific to this model, and that four independent model variants converged on similar generic answers to a prompt asking for novel solutions. That observation is the extraction’s. No claim is made here that any named study or research tradition established these instruments: as with the failure modes in §9.5, the sources for them are model-generated responses citing no author or paper.
4. The mechanisms name no one to run them. Several entries below specify an overseeing role — a watchdog, a parallel board, an analytical system reviewing decisions — without specifying who appoints it, who funds it, or what holds it to account. One source raises the question for its own proposal and supplies no answer. The gap is recorded here and carried forward to §9.7.
5. Several are posed as questions rather than specified. A number of these mechanisms appear in the sources under headings for research directions, and are stated as questions about what could be built or modeled rather than as descriptions of something specified. The entries below mark where that is so. Where a mechanism is posed as a question, the description here reports the question’s subject and does not convert it into a specification the source does not contain.

Provenance of this section’s material. This section was first written from the staged extraction alone, as the rest of chapter 09 was. That extraction was then found to be incomplete against the documents it was drawn from: at least one mechanism carrying a source-level attachment had not been carried into it, and several mechanisms it did record had been separated from threats their own wording bears on. A targeted sweep of the primary sources was run in consequence, reading both adversarial interview passes, the worked scenario document, all four polycentric-governance responses, and both glossary files in full, and reporting only what the extraction had missed or misattributed. Three results follow, and a reader should be able to tell which claims rest on which body of material.

• Three mechanisms are new to this section and came from the sweep rather than from the extraction: §9.6.16, §9.6.17, and §9.6.18. Their entries say so. A fourth, §9.6.19, was in the extraction as a mechanism with no stated target, and the primary source turns out to state one; that entry says so too.
• Five of the threats the sweep examined, each of which the extraction left with nothing, are recorded here as partial: Divide and Conquer, Manufacturing Dissent, Poisoning the Well, Exploiting Blind Faith in Tech, and Hijacking the Support Network. Trojan Horse stands in the same position — the extraction attaches nothing to it either, and its partial status rests on the correspondence §9.6.2 draws — but it was not among the entries the sweep was asked to check. Three of the five moved status in this section — the other two were already partial on the correspondence §9.6.2 draws, and the sweep added a second mechanism to each. Every one of the five rests on an attachment the sweep marks as inferred rather than stated by a source, and every cell in §9.6.21 says so in those terms. None of the five is a source-level attachment, and partial here records that a mechanism bears on the threat, not that anyone proposed it for the threat.
• Six threats were checked against every primary source and confirmed to have nothing directed at them. Four came out of the first sweep: Founder Capture, Talent & Idea Poaching, Greenwashing Through Imitation, and Inaction as a Weapon. A supplementary sweep covered the two the first had passed over, Tyranny of Small Kindness and Weaponized Burnout, and found the same. Their cells in §9.6.21 state the stronger finding this permits — that the sources offer nothing, rather than that the extraction records nothing. Three citation errors in the extraction were corrected at the same time, and are noted at the entries that carry them.

Everything not covered by those three results rests on the staged extraction, on the same footing as the rest of chapter 09.

9.6.1 Ethical Audits Beyond Financial Matters

The mechanism. Audit coverage is extended past financial anomalies to patterns of conduct that work the framework’s gray areas for advantage. The stated object of detection is a pattern rather than an incident, and the source’s framing of the change is a reorientation of the network’s technical capacity toward harmony rather than toward technical optimization alone.

What it detects. Conduct that leaves no financial irregularity because none occurred. The class of patterns this describes is the one §9.3 treats: acts that are lawful, defensible in the network’s own vocabulary, and invisible to an audit that examines whether the figures are correct, because the figures are correct.

What it is directed at. The source names it as one of three possible responses to the framework-turned-against-itself scenario (§9.3.1 to §9.3.5), which it presents as unresolved and does not close with any of the three. It states an explicit limit on the first of the three and states none on this one. The extraction also maps it onto Death by a Thousand Papercuts (§9.3.6), on the basis that both patterns stay below the threshold at which any single instance can be established as a breach; that second correspondence is the extraction’s rather than the source’s.

Sources: ~/code/papr/icn/security-bad-actors-and-mmm.md.

9.6.2 Algorithm Auditing

The mechanism. Standing scrutiny of the network’s decision-support tools, directed at detecting bias in them and at maintaining the condition that they augment rather than replace human judgment and value-based decision-making.

What it detects. Bias in a tool’s output, as distinct from error in it. The entry is one line in a terminology list, and the source states no procedure, no cadence, and no party responsible for conducting the audit.

What it is directed at. The extraction names no threat for this mechanism. The correspondence to the three tool-directed patterns — Trojan Horse (§9.4.5), Poisoning the Well (§9.4.6), and Exploiting Blind Faith in Tech (§9.4.7) — is this chapter’s, and it rests on those entries’ own stated precondition that no periodic bias audit is conducted. The fit is closest for Trojan Horse, where the defect is in the tool. For the other two it is weaker: a mechanism that scrutinizes a tool is not on its face a mechanism that checks the provenance of the data the tool was trained on or the inputs supplied to it at the point of use. §9.6.21 records all three as partial for that reason.

Sources: ~/code/site-provide-coop/!! Localization.md, term 8. The staged extraction attributes this term to ~/code/site-provide-coop/Values--.txt, which is an error. The two files hold the same glossary pass, and the primary-source sweep establishes the difference between them: !! Localization.md carries terms 6 to 20 in three headed groups, and Values--.txt carries only terms 11 to 20, dropping terms 6 to 10 while retaining the section header that belongs above them. Terms 7, 8, and 9 — mandatory psychoeducation, algorithm auditing, and the sanctioned-practice sense of “Trojan Horse” recorded at §9.1.6 — exist only in !! Localization.md. The citation here follows the file the term appears in.

9.6.3 Language Watchdogs and Linguistic Immunology

The mechanism. Two forms appear. In the first, a designated role analyzes patterns in word usage across the network: which terms are falling out of use, and which are acquiring a connotation they did not previously carry. In the second, the analysis is conducted by a system trained to detect shifts in language patterns, examining the rhetorical strategy by which a proposal is framed as well as its content.

What it detects. Change in what the network’s shared vocabulary requires, at a rate slow enough that no single usage is remarkable. The first form detects the change by comparison over time; the second is proposed as a research direction rather than as a specified system, and the source states it as a question about what could be trained rather than as a description of something built.

What it is directed at. Lexical Drift (§9.4.9). This is one of the five attachments made at the source level: the source attaches the watchdog role to the pattern it names “Capturing the Narrative”, which is the pattern §9.1.2 catalogs as Lexical Drift. The same prevention line carries two further measures, recorded here so that the attachment is stated in full: mandatory participation in ethics debate clubs, and the celebration of dissenters in the network’s official history. Neither detects a shift in usage, and the status §9.6.21 records for Lexical Drift rests on the watchdog role.

Sources: p13r01, Response 1 (language watchdogs) and Response 3 (linguistic immunology).

9.6.4 Anomalous-Pattern Review

The mechanism. Two proposals treat an anomaly in the network’s own records as the object of review. In the first, analytical systems examine network data for hidden patterns indicating factional alignment, including alignment among participants who intend no harm; the source adds that such systems are themselves to be kept under scrutiny, so that they do not become an instrument of control in their own right. In the second, members hold positions on the outcomes of pending decisions, and unusual movement on a proposal whose outcome appeared settled is treated as a signal that the decision warrants closer examination. The second is posed in its source as a question about whether such a market would expose manipulation, and no procedure for running one is stated.

What it detects. In the first case, coordination that no participant has declared and that formal records do not show. In the second, a divergence between the expected and the priced outcome of a decision, which is a signal that something not visible in the deliberation is affecting it.

What it is directed at. The extraction records both mechanisms without a stated target, and this chapter does not supply one. Both are listed in the extraction among mechanisms it marks as lower priority. The self-scrutiny condition attached to the first is the point at which one source raises the question of who oversees the overseers, and it names no mechanism for that scrutiny; §9.7 carries the question forward.

Sources: p13r01, Response 3 (analytical review of factionalism patterns) and Response 2 (positions held on decision outcomes).

9.6.5 Rotation as a Principle

The mechanism. No member holds the same role for an extended period. The source names leadership positions, resource-allocation committees, and prominent roles in conflict resolution as within scope, and states two effects: entrenchment is prevented, and the range of skills held across the network widens.

What it prevents. Accumulation of influence in a position over time. The mechanism acts on tenure rather than on conduct, so it does not require that anything be detected before it operates.

What it is directed at. The extraction maps it onto Hidden Oligarchy (§9.2.1), Cult of Personality (§9.2.2), and Founder Capture (§9.2.9). The mapping to Founder Capture is where the mechanism’s limit is clearest, and §9.2.9 states it: rotation reaches formal office, and the authority a founder holds is informal, so a rotation requirement can run in full while the influence the pattern converts is untouched by it. This is the first of the three mechanisms §9.2.9 refers to as general anti-entrenchment mechanisms cataloged here.

Sources: p13r01, Response 1.

9.6.6 Shadow Governance and Shadow Boards

The mechanism. Two forms appear. In the first, units selected at random form temporary parallel structures for consequential decisions, and the recommendations those structures produce are compared against the outcomes the official process reached. In the second, every major decision body has a standing unofficial counterpart whose members are selected at random, whose terms are strictly limited, and which holds veto standing over decisions it judges too risky.

What it detects or prevents. The first form detects divergence: a systematic gap between what a parallel body drawn by lot recommends and what the official body decides is a signal about the official body rather than about any single decision. The second form does not detect, it blocks, and the source states its purpose as disrupting the ability of a faction to hold entrenched control.

What it is directed at. The extraction maps both forms onto Hidden Oligarchy (§9.2.1), Weaponized Consensus (§9.2.3), and Founder Capture (§9.2.9). Against Weaponized Consensus the mapping is partial in a specific way: the pattern in §9.2.3 works by extending deliberation without a terminating condition, and neither form of this mechanism states how a parallel body brings a deliberation to an end. This is the second of the three mechanisms §9.2.9 refers to, and it is the one whose selection by lot §9.2.9 describes.

Sources: p13r01, Response 1 (shadow governance) and Response 3 (shadow boards).

9.6.7 Values-Based Circuit Breakers

The mechanism. Metrics tied to long-term wellbeing carry thresholds, and a breach triggers an audit automatically and mandates specified actions, without regard to the unit’s financial performance. The source names two thresholds: disparity in member access to mental-health resources, and leadership composition skewing toward a single demographic group.

What it detects. A condition rather than an act. The mechanism requires no finding about anyone’s conduct and no judgment that a pattern is under way: the threshold is either breached or it is not, and the response follows from the reading.

What it is directed at. The extraction maps it onto Mission Creep (§9.1.3) and, through the leadership-composition threshold, onto Hidden Oligarchy (§9.2.1).

Sources: p12r01, Response 3.

9.6.8 Breakaway Provisions

The mechanism. Units pre-negotiate the terms on which they may exit, against the case that the wider network departs unacceptably from its stated commitments. The source states two effects: the provision deters, and where deterrence fails, the network fragments into clusters that retain the commitments rather than being held whole under the control of the parties that captured it.

What it prevents. It does not detect. It changes what a capture is worth by limiting what a captured network retains, and the change operates whether or not anyone identifies the capture as it occurs.

What it is directed at. The extraction maps it onto Founder Capture (§9.2.9) and onto the escalation of Mission Zealot (§9.1.5, §9.2.8) and Purity Trap (§9.1.4) — that is, onto the case in which those patterns have already carried the network past a point the exiting unit will accept, rather than onto the patterns as they operate. This is the third of the three mechanisms §9.2.9 refers to.

Sources: p12r01, Response 3.

9.6.9 Controlled Burns and Virtuous Viruses

The mechanism. Two proposals introduce change into the network’s own arrangements on a schedule the network sets rather than in response to an event. Controlled burns inject low-level disruption deliberately — simulated conflicts, temporary limits on access to a resource, the sudden dissolution of an established unit — each followed by a debrief. Virtuous viruses model elements of the network’s operating arrangements on processes that spread, mutate, and recombine, so that no part of the network stays fixed long enough for a faction to hold it; the source poses this one as a question about whether such modeling is possible, and specifies no element to which it would apply.

What it prevents. In both cases, the persistence a capture depends on. Neither mechanism identifies a party or an act; both act on the stability that any pattern requiring accumulated position would need.

What it is directed at. The extraction maps virtuous viruses onto Hidden Oligarchy (§9.2.1), Cult of Personality (§9.2.2), and Founder Capture (§9.2.9). It maps controlled burns onto resilience generally rather than onto a named pattern, and notes that they connect to the antifragility scenario modeling specified in docs/15-advanced/06-systems-simulation-and-stress-testing.md, which §9 cites as existing coverage rather than restating.

Sources: p13r01, Response 1 (controlled burns) and Response 2 (virtuous viruses).

9.6.10 Rule Revision and Redefined Member Wellbeing

The mechanism. Two of the three responses the source names to the framework-turned-against-itself scenario. The first is immediate revision of the rule to close the specific gap that was used. The second redefines what the network counts as member wellbeing so that it includes the psychological cost borne by the members of a unit asked to rationalize conduct they judge harmful, which relocates responsibility for that conduct from market conditions to the network’s own participants.

What they prevent. The first removes one instance and nothing else. The source states its own limit: it buys time, and where it is not accompanied by wider reform it generates resentment. §9.3.5 states a further limit that follows from the vulnerability it names — a closed gap adds a rule, the rule joins the set being optimized against, and the process continues from the revised set. The second changes a measure rather than a procedure, and the source states it as a question about what the network’s wellbeing measures could be made to include rather than as a specification of how.

What they are directed at. The framework-turned-against-itself scenario (§9.3.7), at the source level. The source presents the scenario as unresolved by design and offers none of its three responses as a resolution.

Sources: ~/code/papr/icn/security-bad-actors-and-mmm.md.

9.6.11 Devil’s Advocate Simulation and Mandatory Apprenticeship

The mechanism. Two forms. In the first, a rotating group is assembled on the model of a corporate red team and tasked with devising strategies for turning a live unit proposal to harm, with the exercise directed at morally ambiguous conduct rather than at conduct that is plainly unlawful; the wider network then has to identify the vulnerability and answer it. In the second, a period of service in a designated contrarian role is a precondition of holding a leadership position. The source states the second form’s effect on the leadership pipeline directly: it removes the standing of a leadership group defined by moral purity.

What it prevents. Exposure to the ambiguous case is the stated object in both forms. The first produces a finding about a specific proposal; the second produces a population of leaders who have had to occupy the adversary’s position before holding office.

What it is directed at. The extraction maps the pair onto Mission Creep (§9.1.3), Purity Trap (§9.1.4) through the forced engagement with ambiguity, and Founder Capture (§9.2.9) through the leadership-pipeline requirement. The Founder Capture mapping is subject to the limit §9.2.9 states, since a service requirement conditions formal office and not informal standing.

Sources: p12r01, Response 3 (devil’s advocate simulation); p13r01, Response 3 (mandatory devil’s advocate apprenticeships).

9.6.12 Branching Corruption Exercises

The mechanism. Members work through branching scenarios in which ethical compromises are presented as growth opportunities. Choices carry simulated consequences for network health, reputation, member wellbeing, and financial return. The exercise also records the justification a participant gives for each choice, and the source treats that record, rather than the choice, as the output: it allows common rationalization patterns to be identified across participants and answered through targeted education before they are used in a live decision.

What it detects. Rationalization patterns in the population, as distinct from a disposition in any one participant. The unit of analysis is the reasoning offered, and it is collected under conditions where offering it costs nothing.

What it is directed at. The extraction maps it onto the rationalization patterns of Mission Creep (§9.1.3) generally rather than onto a specific sub-pattern of it. In the source it appears among proposals directed at Mission Creep as a whole.

Sources: p12r01, Response 3.

9.6.13 Contests of “Elegant Evil”

The mechanism. A gamified challenge invites members to design contract gaps that violate the network’s values while remaining likely to survive legal scrutiny, or to devise ways of working a regulatory process. The stated purpose is to have the network find its own pathways to corruption before an outside party does.

What it detects. Vulnerabilities in the network’s own drafting and in its regulatory position, found by directed search rather than by monitoring. The mechanism produces a list of available moves and nothing about whether anyone has made one.

What it is directed at. The extraction maps it onto the framework-turned-against-itself scenario (§9.3), where the object of the exercise and the mechanism of the threat are the same — a gap that is lawful to use — and onto Regulatory Capture (§9.4.1) through the part of the challenge concerning regulatory process. The second mapping is partial in a way worth naming: Regulatory Capture as §9.4.1 states it is conducted by incumbents from outside the network, and an internal contest that maps the network’s regulatory exposure produces knowledge of that exposure rather than a defense against the party exploiting it.

Sources: p12r01, Response 3.

9.6.14 Myth of the Fallen

The mechanism. Where a unit or venture has been captured, the case is written up honestly and shared openly within the network as an educational document. The stated purpose is a standing record that good intentions confer no immunity, together with a catalog of the early indicators the case exhibited.

What it prevents. Nothing directly. Both sources that name it treat it as a cultural mechanism directed at complacency rather than as a detection mechanism, and the extraction records it that way.

What it is directed at. No threat by name. The extraction records it as general.

Sources: p12r01, Response 3; and independently as term 19 in ~/code/site-provide-coop/!! Localization.md and ~/code/site-provide-coop/Values--.txt. Term 19 falls in the range both files carry, and the sweep confirms both citations.

9.6.15 Further Mechanisms Named Without a Stated Target

The extraction records a further set of mechanisms that no source attaches to a named threat, and marks them as lower priority for this chapter. One entry in the list below is an exception and is marked as such, and one that the extraction placed here has been moved out to §9.6.19, its source having stated a target for it. The rest are listed here so that the coverage in §9.6.21 can be read as a statement about the mapping rather than about the inventory: a threat recorded there as having no countermeasure proposed may still fall within the reach of something in this list, and no source says so. Several of the entries below appear in their sources under headings for research directions and are posed as questions, per limit 5 above; the descriptions state what each question is about and do not supply a specification.

• A network immune system: self-correcting mechanisms triggered at thresholds of stress, conflict, or detected deviation, combining human-led intervention with algorithmic detection. The extraction records an unresolved question about this term, since one source attributes the function to a specific entity — term 11 of the same glossary lists an “immune system” for the network among provide.io’s functions — and another poses it as an open research direction; §9.7.4 carries the question forward.
• Mandatory psychoeducation at onboarding, in conflict resolution and communication, stated as preventive rather than remedial. This is term 7, and it exists only in !! Localization.md; the extraction’s citation of Values--.txt for it is the third of the three errors §9.6.2 records.
• Ethical stress tests: simulations exposing vulnerabilities in the model. This entry is one of the five source-level attachments noted in the opening limits, and the only one in this list — the source states their aim as uncovering the potential for gradual erosion of the mission rather than overt ethical failure, which attaches them to Mission Creep (§9.1.3).
• Mandated outsiders: a rotating skeptic drawn from another unit, joining temporarily with the task of identifying potential for conduct contrary to the network’s commitments.
• Red teams assembled before any major initiative, tasked with finding how the plan could be turned to advantage, as a generalization of the mechanism in §9.6.11.
• Real-time dilemma simulation, in which members respond to escalating moral dilemmas without preparation and the responses are examined by the group afterward.
• Public post-mortems on close calls, in which an exposed attempt is written up in anonymized form as an educational document, on the model of §9.6.14 but covering attempts rather than completed captures.
• An adversarial analytical system maintained separately from the network’s collaborative tools and tasked with finding vulnerabilities and running simulated attacks. The source names the attack classes it would simulate, and one of them is a targeted misinformation campaign, which is the mechanism Divide and Conquer (§9.4.4) works by. The extraction’s paraphrase dropped that detail; the sweep restored it, and §9.6.21 records the resulting attachment as inferred.
• Game-theoretic modeling directed at making conduct aligned with the network’s stated values the rational choice. This overlaps the game-theoretic material already specified in docs/15-advanced/06-systems-simulation-and-stress-testing.md, and §9 records that no source connects that material to any threat in this catalog.

Sources: p12r01, Responses 1 to 3; p13r01, Responses 1 to 3; ~/code/site-provide-coop/!! Localization.md (terms 7, 11, 16, 17); ~/code/site-provide-coop/Values--.txt (terms 11, 16, 17 only, per §9.6.2).

9.6.16 Transparency Rotation, Public Decision Dashboards, and Gamified Detection

The mechanism. A single prevention line in the source combines three measures: a mandatory transparency rotation on the committees that allocate resources; public dashboards on which the justification given for a decision is recorded and can be tracked; and gamified challenges in which members analyze the network’s own data to identify resource manipulation presented as legitimate oversight.

What it detects. Manipulation of resource allocation conducted in the form of oversight. The three measures act at different points: the rotation changes who sits on the committee, the dashboard makes the stated justification for a decision available for comparison against the decision, and the challenge distributes the analysis across the membership rather than concentrating it in a review body.

What it is directed at. The source attaches this line to a scenario of its own — a coalition circulating accounts of bias in resource-allocation decisions, algorithmic and human, in order to degrade trust in the network’s leadership. The attachment is source-level, but the scenario it attaches to has no entry in the catalog at §9.1.2; it is one of the two source-level attachments in that position, the other being §9.6.19. The sweep recorded in the provenance note above finds that scenario thematically close to Manufacturing Dissent (§9.4.8) without being identical to it: Manufacturing Dissent works by manipulating what the health-monitoring system displays, and this scenario works by circulating accounts about decisions. The link to Manufacturing Dissent is therefore inferred rather than stated, and §9.6.21 marks it as such.

Sources: p13r01, Response 1. Recorded in the primary-source sweep; absent from the staged extraction.

9.6.17 Early-Warning Flagging of Unusual Activity

The mechanism. Data on communication patterns, resource accumulation, and the justifications recorded for decisions is analyzed to highlight unusual clusters of activity for human examination. The source states a limit as part of the mechanism: the analysis cannot establish intent, and what it produces is a starting point for an investigation rather than a substitute for one.

What it detects. A cluster rather than a conduct. The output identifies where to look, and the source assigns the question of whether anything is there to a person.

What it is directed at. The source lists it in a general prevention block alongside rotation, shadow governance, and controlled burns, and attaches it to no named threat. Its stated limit — that it does not replace human review — bears on the mechanism of Exploiting Blind Faith in Tech (§9.4.7), which is a system’s output being acted on without independent checking. That attachment is inferred by this chapter and by the sweep that supplied the entry; the source names no threat next to the mechanism.

Sources: p13r01, Response 1. Recorded in the primary-source sweep; absent from the staged extraction.

9.6.18 Ethical Health Metrics

The mechanism. A set of tracked indicators of the network’s condition extending past financial data, divided into quantitative measures of resource-distribution equity and qualitative measures drawn from member surveys on belonging and fairness.

What it detects. Nothing on its own. The entry is a definition in a terminology list, and the source does not present it as a countermeasure or state any use to which the indicators would be put.

What it is directed at. No threat. The sweep that supplied this entry rates it the least confident item it found and marks the link doubly inferred: from a glossary term to a countermeasure function, and from that function to a threat. The reasoning it records is that independently tracked measures of resource distribution are the kind of record against which a fabricated distribution claim could be checked, which would bear on Manufacturing Dissent (§9.4.8). That rating is reproduced here as given and is not strengthened, and §9.6.21 does not rest Manufacturing Dissent’s status on this entry alone.

Sources: ~/code/site-provide-coop/!! Localization.md, term 6. Recorded in the primary-source sweep; absent from the staged extraction.

9.6.19 Buddy System, Values Quizzes, and Fallacy Training at Onboarding

The mechanism. A single prevention line in the source combines three measures applied at the point where a unit is formed: mentors for newly formed units drawn from a pool by random assignment; values assessments during onboarding, set as case studies; and early training in identifying the logical fallacies used to justify conduct at odds with the network’s commitments.

What it prevents. Misrepresentation of the network’s arrangements to a unit that has no independent basis for checking the account it is given. The three measures act at different points: random assignment removes a mentor’s ability to select whom they mentor, the case-study assessment gives a new unit a reference against which a mentor’s account can be compared, and the fallacy training acts on the form of a justification rather than on its content.

What it is directed at. A scenario the source states directly above the prevention line: a member with facility in rhetoric presenting as a mentor to newly formed units, misrepresenting the network’s policies and directing those units toward personal connections rather than official resources. The attachment is source-level. The scenario has no entry in the catalog at §9.1.2, so this mechanism supports no row in §9.6.21 by itself, and it is the second of the two source-level attachments in the material that land on nothing the catalog names; §9.7.7 treats the question that raises. No correspondence to a cataloged entry is drawn here.

Sources: p13r01, Response 1. The staged extraction recorded the buddy system among the mechanisms it marks as having no stated target, and dropped the values quizzes, the fallacy training, and the scenario the source attaches all three to; the attachment is restored here from the primary source.

9.6.20 Governance Instruments That Belong to Chapter 03

The extraction records a further set of mechanisms proposed against the polycentric failure modes of §9.5. They are governance design rather than detection or structural resistance: they specify how coordinating bodies are constituted, what agreements between units contain, and how authority over a contested decision is settled. Chapter 03 specifies affirmative governance design for a Humanized Autonomous Organization (HAO) — the network’s coordinating framework — and this chapter cross-references it rather than restating it, per the boundary §9 sets. The instruments are named below with the failure modes the extraction maps them to, and are not specified here.

• Bridging organizations, mapped to the coordination cluster, which §9.5 divides between Absence of Overarching Vision (§9.5.1) and Cross-Boundary Problem Paralysis (§9.5.2).
• Joint governance agreements, mapped to the jurisdictional-overlap cluster, which §9.5 treats as Conflict Deadlock (§9.5.7), and bearing on Cross-Boundary Problem Paralysis (§9.5.2) through the cross-boundary responsibilities they record.
• Equity scorecards paired with redistributive transfers, mapped to the equity and representation cluster, which §9.5 divides between Elite Capture of Governing Units (§9.5.4) and Equity Drift (§9.5.5).
• Citizen audits with open-data tracking of decisions and outcomes, mapped to the accountability and transparency cluster, which §9.5 treats as Accountability Loss Through Opacity (§9.5.8).
• Policy labs and governance sandboxes with sunset clauses, mapped to the entrenchment and path-dependence cluster, which §9.5 folds into Absence of Overarching Vision (§9.5.1).
• Resilience networks, mapped to resilience across the modes rather than to one of them, and overlapping the containment protocols in docs/06-lifecycle/03-collapse-and-containment-protocols.md and the antifragility modeling in docs/15-advanced/06-systems-simulation-and-stress-testing.md.
• Social impact bonds for governance, recorded by the extraction as a funding mechanism directed at no particular failure mode.

The staged extraction omitted the fourth polycentric response, 005-r-pi.md, which states six further instruments. That response differs from the other three in form: each instrument is stated against a named gap in the construction “To address X…”, and the gap terms it names are the ones the same variant supplied to the earlier gaps prompt. The correspondence from an instrument to a gap term is therefore the source’s; the correspondence from that gap term to a failure mode is §9.5’s consolidation, which reads that variant’s terms at §9.5.1, §9.5.2, §9.5.5, §9.5.6, and §9.5.7.

• Knowledge-sharing platforms for data and practice among centers, stated against information asymmetry, and adaptive and participatory monitoring and evaluation, stated against the problem of assessing a polycentric system’s performance. §9.5.6 reads both of those terms for Monitoring and Enforcement Cost at Scale.
• Capacity-building initiatives — training, technical assistance, or financial support directed at less-resourced centers so that they can take part in decisions — stated against capacity and resource constraints, which §9.5 reads for its resource-provision sense at Absence of Overarching Vision (§9.5.1) and for its participation sense at Equity Drift (§9.5.5).
• Mediation and conflict-resolution mechanisms, stated against conflicting interests and values, and the clarification of each center’s roles, responsibilities, and decision-making authority, stated against overlapping jurisdictions and unclear boundaries. §9.5 treats the second of those terms as Conflict Deadlock (§9.5.7).
• Multi-level coordination bodies, stated against scaling up and coordination across scales, which §9.5 treats as Cross-Boundary Problem Paralysis (§9.5.2).

Two of the eight failure modes receive nothing directed at them in this set: Race to the Bottom Between Jurisdictions (§9.5.3), and — except through the resource-allocation material folded into it — the system-wide provision aspect of Absence of Overarching Vision (§9.5.1).

Limit 3 in the opening applies to this subsection in particular. The extraction’s own assessment is that the material it carried largely reproduces general public-administration content rather than mechanisms specific to this model, and that it is more useful for its taxonomy of failure modes than for its mechanism language. The fourth response, recovered here, is of the same kind. §9.5 uses the material for the taxonomy; this subsection records the mechanisms without adopting them.

Sources: ~/code/papr/polycentric-governance/005-r-claude.md, 005-r-gemini.md, 005-r-chatgpt.md, 005-r-pi.md. The last is absent from the staged extraction and is read here from the primary source.

9.6.21 Coverage

The tables below state, for every entry in the catalog at §9.1.2 and every failure mode in §9.5, what the material proposes against it. Status carries three values:

• addressed — one or more mechanisms in this section are directed at the entry in the material.
• partial — a mechanism bears on the entry, but the material directs it at a wider or adjacent target, or the source does not offer its own answer as a resolution. Cells reaching this status on an attachment nobody stated say so in the cell, in the form “attachment inferred”.
• none proposed — nothing in the material is directed at the entry. Where the primary-source sweep checked the entry directly, the cell says so, and the finding is then about the sources rather than about the extraction.

Two properties of the tables should be read together with limit 2 above. First, a status is a statement about the mapping recorded in the material, not about whether a mechanism would work: nothing here has been tested, and addressed records that something was proposed, not that the pattern is answered. Second, every correspondence this chapter or the sweep draws rather than inherits from a source is marked as inferred in the cell where it appears, and the count of them exceeds the count of source-level attachments several times over.

The failure modes are tabulated separately from the threats. §9.1.1 defines the polycentric category as the one requiring no actor, so the column head “Threat” does not apply to them.

What the tables show. Of the 22 cataloged threats, two are recorded as addressed, fourteen as partial, and six as having no countermeasure proposed. Of the eight failure modes, none is recorded as addressed, seven as partial, and one as having none proposed; the seven partial entries all rest on the material §9.6.20 assigns to chapter 03, which is general governance content rather than mechanisms specific to this model.

Fourteen partial entries is a larger figure than the material’s own directedness supports, and the cells state why. Exactly one of the fourteen rests on an attachment a source makes itself: the framework-capture scenario of §9.3, which the source presents as unresolved and answers with three responses it does not offer as a resolution. Six rest on correspondences this chapter or the primary-source sweep drew rather than found — five marked inferred by the sweep, three drawn by §9.6.2, with two rows carrying both. The remaining seven rest on correspondences the staged extraction drew. Three of those seven — Purity Trap, Hidden Oligarchy, and Cult of Personality — rest on the same kind of untagged prevention block that yields partial elsewhere in the table, and are recorded as partial on that basis rather than as addressed. A reader counting what the sources actually direct at a named threat should count the two addressed entries and that single partial, and read the rest as a record of what bears on a threat rather than of what anyone proposed for one.

All six threats with nothing proposed against them were checked directly against every primary source and are empty there. Four came out of the first sweep: Greenwashing Through Imitation and Talent & Idea Poaching, both of which act from outside the network’s boundary against a countermeasure set that is almost entirely internal; Inaction as a Weapon, for which the nearest material is an open research question rather than a proposed mechanism; and Founder Capture. A supplementary sweep covered the remaining two, Tyranny of Small Kindness and Weaponized Burnout, which act on the network’s mutual-support and participation norms. Each is named once in the sources, in a single sentence, with nothing attached to it, and the nearest candidate mechanisms are directed elsewhere: the adversarial-analysis tooling nearest the first is stated generally enough to bear on any threat, and the faction-detection tooling nearest the second acts on factionalism rather than on how work and emotional labor are allotted, which is what the pattern’s own precondition names. §9.7 may state all six at the same strength.

Founder Capture is the finding this section ends on, and the sweep raises it from a property of the extraction to a property of the sources. Every source group was checked for a mechanism directed at the informal legitimacy a founder holds by having founded the network, and none holds one. The worked scenario document reaches the case in its closing line, states that it would require holding members in leadership positions to a higher standard than the one applied generally, and stops without naming a mechanism or invoking any element of the framework. The sources raise the problem and decline to resolve it. That is the whole of the claim, and §9.7 carries it forward in that form.

§9.7 inverts these tables and treats the entries with nothing proposed against them, together with the questions this section leaves open: who conducts the mechanisms in §9.6.1 to §9.6.4 and §9.6.17, which entity performs the network immune-system function named in §9.6.15, and whether the scenarios §9.6.16 and §9.6.19 are attached to warrant catalog entries of their own, since they are the two source-level attachments in the material that land on nothing §9.1.2 names.