# 08 — Country Chapters and Jurisdictional Sovereignty > *A single universal filter cannot survive first contact with national law. Chapter-scoped aggregators preserve global plurality while respecting local obligations.* ## Why this is necessary A global protocol with a single editorial policy will fail at the first legal collision. Germany's StGB §130 criminalises Holocaust denial; the US First Amendment protects most of the same speech. A single filter that suppresses denial content is illegal in the US; a single filter that permits it is illegal in Germany. Either the protocol serves one of these audiences worse, or it handles jurisdictional divergence architecturally. Similar collisions: France's laws against contesting crimes against humanity; India's IT Rules 2021 + takedown mandates; Singapore POFMA; Brazil's Marco Civil + content obligations; the EU DSA Article 16 notice-and-action + Article 22 Trusted Flagger; the UK Online Safety Act 2023; Russia's "fake news" laws (post-2022); Turkey's defamation law applied to foreign publications; Thailand's lèse-majesté. And these are only the headline cases. A protocol that claims to be "neutral infrastructure" while holding one global filter is making a specific editorial choice that privileges one jurisdiction's legal regime over all others. If that choice is US-First-Amendment, the protocol is legally operable in fewer countries than it pretends. If it is EU-DSA-compliant, it exceeds what many jurisdictions require and imports editorial policy that those jurisdictions may reject as imperial. The answer is **country chapters** — per-jurisdiction organisational units that operate aggregators under their own national legal compliance, while participating in the global protocol for the write layer, payment flow, and shared spec evolution. ## Architecture ``` ┌──────────────────────────────────────────────────────────┐ │ GLOBAL PROTOCOL LAYER │ │ - chain-anchored attestation + cascade log (universal) │ │ - write layer is permissionless │ │ - hard-list of narrow operational refusals (universal) │ └──────────────────────────────────────────────────────────┘ ↑ reads chain events ┌──────────────────────────────────────────────────────────┐ │ GLOBAL FOUNDATION │ │ - stewards the specification │ │ - operates the reference chain contracts │ │ - coordinates chapter relationships │ │ - global treasury │ │ - appoints dispute panel │ └──────────────────────────────────────────────────────────┘ ↕ affiliation agreements ┌──────────────┐ ┌──────────────┐ ┌──────────────┐ ┌─────┐ │ CHAPTER DE │ │ CHAPTER US │ │ CHAPTER EU │ │ ... │ │ - DSA compl. │ │ - 1st Am. │ │ - DSA compl. │ │ │ │ - NetzDG │ │ - Section 230│ │ - AI Act Art │ │ │ │ - filter X │ │ - filter Y │ │ - filter Z │ │ │ │ - national │ │ - national │ │ - national │ │ │ │ aggregator │ │ aggregator │ │ aggregator │ │ │ │ - national │ │ │ │ │ │ │ │ NGO funding│ │ │ │ │ │ │ └──────────────┘ └──────────────┘ └──────────────┘ └─────┘ ``` Each chapter: - Is legally independent, incorporated in its jurisdiction. - Signs an affiliation agreement with the global foundation (spec alignment + trademark + good-faith membership). - Operates one or more aggregators serving its jurisdiction. - Publishes its filter rules publicly, with legal rationale. - Has its own national funding mix (NGOs, donations, commercial) and some revenue share from global treasury. - Has a seat on the foundation's global coordination body. ## What is filtered, what is not **Not filtered at chapter level:** - The chain log itself. A claim posted from any jurisdiction is recorded universally. - Chapter filters apply to *display*, not to *write*. - Validators from any jurisdiction can post attestations visible worldwide on the chain, even if a specific chapter does not surface them. **Filtered at chapter level:** - Display of specific claims or attestations in that chapter's aggregator responses. - Serving of specific content through that chapter's edge cache. - Surfacing of validators credentialed in other chapters. **The user experience of a filter:** - Default jurisdiction is auto-detected by client (IP, locale, user preference). - Jurisdiction-default aggregator applies its filter. - UI *explicitly informs the user*: "You are viewing the US-chapter aggregator. 0 items are hidden from this view. [Switch view]" - User can opt to view a different chapter's aggregator with one click; warning is surfaced about jurisdictional implications. - User can run a local aggregator that applies no chapter filter at all (for researchers, lawyers, journalists operating with legal awareness). ## Relationship to CPML CPMLs operate at the consumer layer; chapter filters operate at the aggregator layer. They compose: - A user in Germany with a `cpml:minimal-evidence` profile sees only claims her CPML surfaces, from the German chapter's aggregator, with the chapter filter applied. - The same user can manually switch to the US chapter aggregator and see the same claims with the US filter (less restrictive). She is doing so knowingly, with UX disclosure. Chapter filters are thus a legal-compliance layer on top of CPML personalisation. Neither replaces the other. ## Critical analysis **1 — Balkanisation.** If every country has a different filter, the "plural verdicts" promise weakens — users in different countries see materially different claim sets. Response: the *write* layer remains universal; filters affect only *display*. Researchers and cross-border consumers can explicitly cross chapters. Chapter filters are visible and published, not hidden. **2 — Chapter capture.** A national government pressures the local chapter; chapter complies; citizens of that country see only state-approved claims. Response: chapter governance is independent of the state (foundation + private members + NGO funding); chapter affiliation can be revoked by foundation if it is captured; users can access other chapters' views; the chain log still preserves everything publicly. **3 — Affiliation-agreement disputes.** Foundation and chapter disagree on what constitutes good-faith membership; chapter forks. Response: affiliation agreement has arbitration clauses; forking is possible but the trademark stays with the foundation; a chapter can fork operationally but cannot call itself "Veritas-Germany" without the foundation's assent. **4 — Per-jurisdiction rule explosion.** 195 UN member states × dozens of sub-jurisdictions × varying legal updates = administrative nightmare. Response: chapters are only formed where a jurisdiction has a distinct legal regime that materially affects Veritas operation. Phase II plans 3–5 chapters (US, DE, FR or EU27-as-one, UK, possibly Brazil or India). Many jurisdictions will share the EU27 chapter; others will be served by the global-default aggregator until a local chapter emerges. **5 — Funding asymmetry.** Rich countries form chapters easily; poor countries lack resources. Result: Global South is served only by the global-default. Response: foundation's global treasury funds "seed chapters" in under-resourced regions when NGO + academic partners materialise. NGOs like Africa Check and Chequeado (Latin America) are natural seed-partners. **6 — Legal jurisdiction mismatch.** Content produced in the US by a US validator, served to a German user via the German chapter, filtered in Germany. Whose law applies? Response: aggregators serve their jurisdiction's legal regime. US validators are subject to US law for their posts; German chapter is subject to German law for its serving decisions. Both can be simultaneously correct under respective laws. This is the same legal posture as Wikipedia's national chapters. **7 — Chapter does the wrong thing by its own local standards.** A US chapter filters content that is protected under US law, prompted by political pressure. Response: chapter filter rules are published; any user can see what is filtered; appeals process exists; foundation can revoke affiliation. ## Related work - **Wikimedia Foundation + national chapters.** Wikimedia Deutschland, Wikimedia France, Wikimedia UK — independent legal entities with affiliation agreements to the global foundation. Governance crisis of Wikimedia UK 2012 is instructive. Wikimedia's chapter model has lasted 20+ years and handled substantial jurisdictional stress. - **Creative Commons jurisdictional porting.** CC licences were originally ported per-jurisdiction 2003–2013; CC 4.0 (2013) unified to an international licence. The lesson: porting is useful early, unification may serve better later. Veritas should consider whether chapters are a transitional structure or a permanent one. - **ICANN regional structures (RALOs, ccNSO).** Regional communities under a global root; ccTLDs operated by national bodies. Governance works but is politically charged; lesson: publishing decisions transparently and having formal appeal reduces conflict. - **Internet Society chapters.** Policy advocacy analog; chapter structure is useful for local outreach even where technical protocols are global. - **Mozilla Foundation's national affiliates.** Advocacy and localisation; comparable affiliation model. ## Phased roll-out **Phase II (months 6–18) — Initial chapter structure.** - Global foundation formed; reference chain contracts deployed. - EU chapter established first (single chapter handling DSA + AI Act + national content laws of member states; model on Wikimedia EU coordination). - US chapter established second (Section 230 / 1st Amendment posture). - Affiliation agreements finalised with both. **Phase III (months 18–36) — Second wave.** - UK chapter (Online Safety Act). - One South American chapter (Brazil or via Chequeado network). - One Asian chapter (Japan or Singapore). - India chapter evaluated (IT Rules 2021 + regulatory turbulence; may need to wait). **Phase IV (month 36+) — Long tail.** - Additional chapters based on partner availability and jurisdictional need. - Global-default aggregator continues serving jurisdictions without chapters. ## Open questions - Should chapter affiliation require board seat on global foundation, or is participation sufficient? - What is the foundation's response if a chapter's national law requires filtering content the foundation considers universally important to preserve? - How are cross-border content disputes handled (e.g., a US-originated claim defamatory under UK law, served via UK chapter)? - How does chapter funding interact with chapter editorial independence? Donor-state chapters raise concerns. - What is the minimum viable chapter (population × resources × legal need) below which a chapter is not formed? ## What we'd build - **Foundation affiliation template** — standard agreement between global foundation and national chapter. - **Chapter dashboard** — per-chapter filter rules, funding, board, published decisions. - **Cross-chapter routing API** — aggregator federation enabling users to query across chapters with disclosure. - **Chapter-level treasury interface** — revenue-share from global + local donation + member fees. - **Chapter onboarding toolkit** — legal templates, reference implementations, ops playbook for chapter-establishment. - **Trademark + brand policy** — what chapters can call themselves, what they can publish under the Veritas brand, revocation procedure.