The method, in one map
Fifty methods, rules and inferences — from the origin of the problem to the architecture that closes it. Each cell links to the passage of the work that establishes it; each institution opens its record in the diagram.
This map exists first as a printable plate, composed on 14 July 2026. It is remade here for the screen: its positions are no longer paragraph numbers — which change with every new edition — but the structure of the work as this site serves it.
Who this map is for: anyone who wants to audit the approach — to find, for each rule or inference, the passage that establishes it. It is not a general explanation of the book. New here? Start with the argument in one page.
How to read this map
Each cell is a method, a rule or an inference from the work. It sits on the chapter where it is established, and it links back there: click it to read the passage. The colour states the status of what is asserted — the nomenclature is that of the evidential statuses, §1.10, taken word for word.
A — ORIGIN & QUESTIONwhere the problem comes from, how it is framed
- 1Practitioner’s origin (2022-2025)4 instruments (2022) → Humanity Test (2025) → the chain
- 2Humanity Test — the entry filterfive questions put to any decision — box opposite
- 3Definition: invisible externalitiescosts imposed on third parties, without consent or redress
- 4The accountability deficit, defineddocumented, but neither qualified, nor priced, nor contestable
- 55 questions, 7 contributionswhy this void · which functions · AI · adoption
B — METHOD & EPISTEMIC DISCIPLINEthe rules the work imposes on itself
- 6Method: design science researchproblem → mapping → requirements → tested architecture
- 7Every assertion carries a statusfact ≠ programme ≠ design ≠ value: 4 burdens of proof
- 8Refutability stated up frontwhat would refute the work is published; 2 global refutations
- 9Institutional theory (§1.4)7 bodies of theory + 5 causal hypotheses — box opposite
- 103 levels of proof, never conflated① accounting ② attributable (range) ③ causal — here: ① + ②
- 113 kinds of cost, never added togetherA market · B welfare · C transfers — box opposite
- 123 registers kept apartcash disbursed ≠ revalued debt ≠ output forgone
- 13Netting the cyclemoney that came in on the way out does not offset the damage on the way back
- 14No global figureno extrapolation to 80 countries without published data
- 15Mapping under protocolcorpus frozen (12/07/2026), coding published, ambiguous items logged
C — EVIDENCE & DIAGNOSISthree corpora, one quantified case, 80 years of precedents
- 16The magnitude, from third-party sourcesfossiles ~7 400 · illicites ~1 000 · gap ODD ~4 000 Md$/an
- 17Taxonomy: 5 modes × 20 families5 channels × 20 families — the two boxes opposite
- 18One case dissected: 15 mechanisms15 channels: A debt · B price (at zero) · C real · O gains
- 19Quantified on Tunisia 2022-2026gross ~15 bn TND; attributable net €0.76-1 bn; output −€0.9 bn
- 20Corpus 1 — the United Nations101 texts: 12 binding · 7 advisory · 82 declaratory
- 21Corpus 2 — Sevilla130 initiatives: 71 finance, 3 rate, 0 adjudicate, 0 bind
- 22Converging recognition at the IMFthe Fund measures, acknowledges — and asks the victim to adjust
- 23Precedent 2009: debtdebt had its chain; compensation has no holder
- 242009 → 2025: the remedy degradessame authors: from a payment (2009) to an analysis (2025)
- 2580 years of degraded mechanisms = a law of attritionwhat binds no one survives; what binds someone dies
D — DOCTRINE — INTERMEDIATE INFERENCESwhat the evidence warrants, provisionally
- 26Inference 1 — an architectural deficitneither information nor money: each tool holds ONE link
- 27Inference 2 — a holder is missingthe spillover chain has never even been proposed
- 28Inference 3 — Korinek turned aroundefficient perhaps; the ledger remains incomplete
- 29Classify before pricing: S0-S67 classes (box); only 2 open a correction
- 30Candidate accountable party: the central bankout of its profits — never by the taxpayer of the country affected
- 31THE thesis: the holderneither the diagnosis nor the remedy: who carries each link
- 326 convictions → 7 functions“the invisible underwrites injustice” → measure, and so on
- 339 design principles (P1-P9)the core: P3, nothing binds without consent (table)
E — ARCHITECTURE — THE 7 + 3 CHAINseven derived institutions, three instruments
- 34Comparison of 4 candidate architectures8 criteria + human costs; the 5-body variant, the only serious rival
- 35Derivation: 1 function = 1 holderone mission each + what the others may not do (negative mandate)
- 36HuIB — measurepublishes harms in ranges; neither rates nor prices
- 37GIRA — raterates severity; never paid by the entity it rates
- 38GFEA — pricesets the consequence (correction or equity); does not collect
- 39BFM — adjustapplies the price at the crossing — to the process, never to the country
- 40GSF — financeunconditional emergency support + grants; never austerity
- 41IPDA — certifycertifies algorithms — the process, never the content
- 42Humanity Court — adjudicatereviews the legality of the six — without commanding them
- 43The integrity of the interfacesevery figure travels with its case file; counted once
- 44Who pays for measurement: pooledlevy by size: the assessed does not pay its assessor
F — INSTRUMENTS · AI · ADOPTION · RESILIENCEhow it is adopted and how it holds
- 45The 3 instrumentsthe Compact organises · the Charter grounds · the Call legitimises
- 46AI-native — the human decidesAI flags, detects, alerts — a named human signs
- 47Entry routes, without an initial treatymeasure and publish first; binding through law (to be validated)
- 48The objections, facedworld government, capture, §51…: sourced answers
- 49The cost of inaction & resiliencethe invisible debt accumulates; rapid modes pre-committed
- 50Limits & open research programme13 work strands: data, double coding, memoranda, pilots
Seven functions, seven holders, three instruments
This map does not propose seven institutions because seven were needed: it derives seven functions from the diagnosis, and assigns each to a single holder. The negative mandate is half of the design — it states what the holder may not do, failing which the chain closes back on itself.
| Function | Holder | What it does | Negative mandate | What it rests on |
|---|---|---|---|---|
| measure | HuIB — Human Impacts Board | publishes the measurement of harms, in ranges | does not rate, does not price, does not adjudicate | IPCC (Nobel 2007) · Beyond-GDP · SNA 2025 · Sen · Ibn Khaldūn |
| rate | GIRA — Global Impact Rating Authority | converts measurement into a severity rating | no price; never social scoring | credit rating agencies inverted — Abdelal · UNCTAD · UNDP · Akerlof · Spence |
| price | GFEA — Global Financial Equity Authority | sets the consequence: correction or equity | does not collect — immunity untouched | OECD Pillar Two · UN tax convention · DICE — Nordhaus · FSIA · SIA · CMF |
| adjust | BFM — Border Fairness Mechanism | applies the price at the border | no tax doctrine; never nationality | CBAM (EU) · WTO, US — Shrimp · HS customs · Pauwelyn |
| finance | GSF — Global Solidarity Fund | parametric emergency support + recovery in grants | does not tax, does not rate; never austerity | Global Fund · loss & damage · parametric insurance |
| certify | IPDA — Inter-Parliamentary Digital Authority | certifies algorithms democratically | the process, never the content | UIP 1889 · UNCITRAL model laws · vTaiwan · AI Act (EU) |
| adjudicate | Humanity Court | settles challenges, reviews legality | legality, never expediency; enforces nothing itself | CAS + New York 1958 · ICJ · ICC · regional courts · ARSIWA 2001 |
The three instruments that ground them
- Humanity Compact — links the seven — interfaces, no joint liability between bodies
- Carthage Charter — grounds it in law — parliaments, internal primacy, solidarity clause
- Call of Carthage — legitimises — citizen adoption → Assembly of Peoples
What the argument rests on — nine families
Each method on the map carries the family or families of references that support it. The detail of each reference is on the page References.
| Family | What it says | Methods | References |
|---|---|---|---|
| F1 | The official diagnosis exists everywhere; constraint begins nowhere. | 20 | GIEC · Pact 2024 · Sevilla 2025 · UNCTAD · PNUD · FACTI · CIJ 2025 · OCDE · BRI · World Bank · OMS · UIP |
| F2 | The Fund measures the spillover — and addresses the remedy to the victim. | 8 | External Sector Report 2023 · Black et al. 2025 · Basu 2020 · Aiyar 2023 · RST 2022 |
| F3 | Eighty years of remedies proposed then degraded, for want of a holder. | 4 | Keynes 1944 · Stiglitz Commission 2009 · Palais-Royal 2011 (Volcker, Camdessus…) · Rajan 2015 · ICE 2025 |
| F4 | Floating does not insulate (Rey); the centre makes the cycle, the periphery bears it (Ocampo). | 12 | Rey 2015 · Miranda-Agrippino & Rey · Bruno & Shin · Gopinath · Korinek · Ocampo · Abdelal · Rodrik · Grabel |
| F5 | Accountable = informing a forum that can sanction (Bovens); an institution is designed (Koremenos). | 12 | Koremenos · Bovens · Grant & Keohane · Koppell · Kingsbury (GAL) · Barnett & Finnemore · Mearsheimer · Stone |
| F6 | The process can be regulated at the border (US — Shrimp); immunity does not bar recording. | 7 | US — Shrimp (WTO) · Howse & Regan · Pauwelyn · FSIA · NML 2011 · UK SIA · CMF · ARSIWA · New York 1958 · CBAM |
| F7 | The magnitude is quantified by third parties; extra-financial ratings diverge (0.38-0.71). | 11 | Swiss Re 2025 · WEF 2025 · GFI · TJN · Oxfam · Maddison · Rennert 2022 · Berg 2022 · IOSCO · Zucman 2024 |
| F8 | Private cost ≠ social cost (Pigou); without a forum, no exchange (Coase); fairness makes rules last (Ostrom). | 17 | Ibn Khaldūn · Pigou · Coase · Arrow · Sen · Ostrom · Nordhaus · North · Buchanan · Akerlof · Tirole · Stern |
| AUT | 4 | Yaiche 2022 (ERF-IMF) · Yaiche 2023 (Tunisia) · Yaiche 2025 (TSYP13) |
The theories drawn on, one sentence each
Each name opens its entry on the page References.
- Pigou (1920)
- the externality: the decision-maker’s private cost diverges from the social cost.
- Coase (1960)
- without a forum and without title, the market for harm cannot form.
- Korinek (2016)
- a spillover calls for correction only if three conditions hold.
- Rey (2015)
- a dilemma, not a trilemma: even floating, no one is insulated from the global cycle.
- Bovens (2007)
- accountability: an actor informs a forum able to attach a consequence.
- Koremenos et al. (2001)
- the features of an institution are rational design choices.
- Ostrom (1990)
- rules perceived as inequitable are not complied with — and collapse.
- Abdelal (2007)
- the authority of the rating agencies was not seized: it was delegated.
Text equivalent of the map
The fifty methods, in lane order, with their status, their families of sources, the thirty-one links of the plate and the place in the work where they are established.
| N° | Lane | Method | What it does | Status | Families | Follows from | Leads to | In the work |
|---|---|---|---|---|---|---|---|---|
| A — ORIGIN & QUESTION | ||||||||
| 1 | A | Practitioner’s origin (2022-2025) | 4 instruments (2022) → Humanity Test (2025) → the chain | Externally verifiable | AUT | — | 2 | §4.1 |
| 2 | A | Humanity Test — the entry filter | five questions put to any decision — box opposite | Design proposal | AUT | 1 | 29 | §4.1 |
| 3 | A | Definition: invisible externalities | costs imposed on third parties, without consent or redress | Design proposal | F8, AUT | — | 17 | §2.2 |
| 4 | A | The accountability deficit, defined | documented, but neither qualified, nor priced, nor contestable | Design proposal | F5, F1 | — | 26 | §2.5 |
| 5 | A | 5 questions, 7 contributions | why this void · which functions · AI · adoption | Design proposal | — | — | — | §1.9 |
| B — METHOD & EPISTEMIC DISCIPLINE | ||||||||
| 6 | B | Method: design science research | problem → mapping → requirements → tested architecture | Design proposal | — | — | — | §1.10 |
| 7 | B | Every assertion carries a status | fact ≠ programme ≠ design ≠ value: 4 burdens of proof | Design proposal | — | — | — | §1.10 |
| 8 | B | Refutability stated up front | what would refute the work is published; 2 global refutations | Design proposal | — | — | — | §4.4 |
| 9 | B | Institutional theory (§1.4) | 7 bodies of theory + 5 causal hypotheses — box opposite | Externally verifiable | F5 | — | — | §1.4 |
| 10 | B | 3 levels of proof, never conflated | ① accounting ② attributable (range) ③ causal — here: ① + ② | Design proposal | F4, F2 | — | 19 | §9.2 |
| 11 | B | 3 kinds of cost, never added together | A market · B welfare · C transfers — box opposite | Design proposal | F7, F8, F1 | — | 16 | §2.4 |
| 12 | B | 3 registers kept apart | cash disbursed ≠ revalued debt ≠ output forgone | Design proposal | F2, F1 | — | 19 | §9.3 |
| 13 | B | Netting the cycle | money that came in on the way out does not offset the damage on the way back | Design proposal | F4, F2 | — | 29 | §9.8 |
| 14 | B | No global figure | no extrapolation to 80 countries without published data | Normative commitment | — | — | — | §9.3 |
| 15 | B | Mapping under protocol | corpus frozen (12/07/2026), coding published, ambiguous items logged | From the programme | F1 | — | — | §3.4 |
| C — EVIDENCE & DIAGNOSIS | ||||||||
| 16 | C | The magnitude, from third-party sources | fossiles ~7 400 · illicites ~1 000 · gap ODD ~4 000 Md$/an | Externally verifiable | F2, F1, F7 | 11 | — | §2.4 |
| 17 | C | Taxonomy: 5 modes × 20 families | 5 channels × 20 families — the two boxes opposite | From the programme | F7, F1, F8 | 3 | 18 | §2.3 |
| 18 | C | One case dissected: 15 mechanisms | 15 channels: A debt · B price (at zero) · C real · O gains | From the programme | F4, F8, F1 | 17 | 19 | §9.1 |
| 19 | C | Quantified on Tunisia 2022-2026 | gross ~15 bn TND; attributable net €0.76-1 bn; output −€0.9 bn | From the programme | AUT, F2, F1 | 10 12 18 | — | §I |
| 20 | C | Corpus 1 — the United Nations | 101 texts: 12 binding · 7 advisory · 82 declaratory | From the programme | F1 | — | 26 | §3.1 |
| 21 | C | Corpus 2 — Sevilla | 130 initiatives: 71 finance, 3 rate, 0 adjudicate, 0 bind | From the programme | F1 | — | 26 | §3.3 |
| 22 | C | Converging recognition at the IMF | the Fund measures, acknowledges — and asks the victim to adjust | Externally verifiable | F2, F4 | — | 30 | §3.1 |
| 23 | C | Precedent 2009: debt | debt had its chain; compensation has no holder | Externally verifiable | F3 | — | 24 | §3.1 |
| 24 | C | 2009 → 2025: the remedy degrades | same authors: from a payment (2009) to an analysis (2025) | Externally verifiable | F3, F4 | 23 | 25 | §12.7.2 |
| 25 | C | 80 years of degraded mechanisms = a law of attrition | what binds no one survives; what binds someone dies | From the programme | F3, F4, F5 | 24 | 27 31 33 | §2.7 |
| D — DOCTRINE — INTERMEDIATE INFERENCES | ||||||||
| 26 | D | Inference 1 — an architectural deficit | neither information nor money: each tool holds ONE link | From the programme | F1, F5 | 4 20 21 | — | §2.5 |
| 27 | D | Inference 2 — a holder is missing | the spillover chain has never even been proposed | From the programme | F5, F4 | 25 | — | §2.6 |
| 28 | D | Inference 3 — Korinek turned around | efficient perhaps; the ledger remains incomplete | From the programme | F4, F8, F2 | — | 29 | §9.6 |
| 29 | D | Classify before pricing: S0-S6 | 7 classes (box); only 2 open a correction | Design proposal | F4 | 2 13 28 | 30 | §9.7 |
| 30 | D | Candidate accountable party: the central bank | out of its profits — never by the taxpayer of the country affected | Design proposal | F3, F4 | 22 29 | 38 | §9.7 |
| 31 | D | THE thesis: the holder | neither the diagnosis nor the remedy: who carries each link | Design proposal | — | 25 | 35 | §4.7 |
| 32 | D | 6 convictions → 7 functions | “the invisible underwrites injustice” → measure, and so on | Normative commitment | F8 | — | 33 | §4.1.1 |
| 33 | D | 9 design principles (P1-P9) | the core: P3, nothing binds without consent (table) | Design proposal | F5 | 25 32 | 35 45 | §4.2 |
| E — ARCHITECTURE — THE 7 + 3 CHAIN | ||||||||
| 34 | E | Comparison of 4 candidate architectures | 8 criteria + human costs; the 5-body variant, the only serious rival | Design proposal | F5 | — | 35 | §4.5 |
| 35 | E | Derivation: 1 function = 1 holder | one mission each + what the others may not do (negative mandate) | Design proposal | F5, F8 | 31 33 34 | — | §4.7 |
| 36 | E | HuIB — measure | publishes harms in ranges; neither rates nor prices | Design proposal | F1, F8, F5 | — | 37 | §5.2 |
| 37 | E | GIRA — rate | rates severity; never paid by the entity it rates | Design proposal | F4, F1, F8, F7 | 36 | 38 | §5.3 |
| 38 | E | GFEA — price | sets the consequence (correction or equity); does not collect | Design proposal | F1, F8, F6, F7 | 30 37 | 39 | §5.4 |
| 39 | E | BFM — adjust | applies the price at the crossing — to the process, never to the country | Design proposal | F6, F7 | 38 | 40 | §5.5 |
| 40 | E | GSF — finance | unconditional emergency support + grants; never austerity | Design proposal | F1, F8 | 39 | — | §5.6 |
| 41 | E | IPDA — certify | certifies algorithms — the process, never the content | Design proposal | F1, F8, F6, F7 | — | 46 | §5.7 |
| 42 | E | Humanity Court — adjudicate | reviews the legality of the six — without commanding them | Design proposal | F6 | — | — | §5.8 |
| 43 | E | The integrity of the interfaces | every figure travels with its case file; counted once | Design proposal | F5, F8 | — | — | §5.9 |
| 44 | E | Who pays for measurement: pooled | levy by size: the assessed does not pay its assessor | Design proposal | F6, F7 | — | — | §5.10 |
| F — INSTRUMENTS · AI · ADOPTION · RESILIENCE | ||||||||
| 45 | F | The 3 instruments | the Compact organises · the Charter grounds · the Call legitimises | Design proposal | F1, F6 | 33 | — | §6 |
| 46 | F | AI-native — the human decides | AI flags, detects, alerts — a named human signs | Design proposal | F8, F6 | 41 | — | §7 |
| 47 | F | Entry routes, without an initial treaty | measure and publish first; binding through law (to be validated) | Design proposal | F5, F8, F1 | — | — | §12.1 |
| 48 | F | The objections, faced | world government, capture, §51…: sourced answers | Design proposal | F5, F4, F7, F8, F1 | — | — | §13.2 |
| 49 | F | The cost of inaction & resilience | the invisible debt accumulates; rapid modes pre-committed | Externally verifiable | F7, F8, F2, F1 | — | — | §13.3 |
| 50 | F | Limits & open research programme | 13 work strands: data, double coding, memoranda, pilots | From the programme | F7 | — | — | §14.1 |