Distributive Equity Through Structure: A Community-Scale Worked Example of Values Stickiness

Structural Distributive Equity: How a Community-Scale Platform Implements Values Stickiness Through a Constitutional Architecture at Sub-Big-Tech Scale.

Author — John Stroh, Director, My Digital Sovereignty Limited, Aotearoa New Zealand ORCID — 0009-0005-2933-7170 DOI — 10.5281/zenodo.19600614 Version — 1.0 (first reviewed edition) Date of first publication — 2026-04-16 Licence — Creative Commons Attribution 4.0 International (CC BY 4.0) Suggested citation — Stroh, J. (2026). Distributive Equity Through Structure: A Community-Scale Worked Example of Values Stickiness. Version 1.0. My Digital Sovereignty Limited, Aotearoa New Zealand. DOI: https://doi.org/10.5281/zenodo.19600614. Published at https://agenticgovernance.digital/whitepapers/distributive-equity.html. ORCID: https://orcid.org/0009-0005-2933-7170. Licensed under CC BY 4.0. Corresponding author — john.stroh@mysovereignty.digital

Status. This paper is not peer-reviewed. The author is not a legal scholar. It is a documentary case study produced by the operator of the platform it describes, written to make the platform’s structural and constitutional commitments legible to a research programme that has been developing analytical tools for the welfare concerns those commitments address.

Abstract

A body of recent legal scholarship argues that digital platforms exercise a distinct form of power — ecosystem power — operating simultaneously through three roles: as gatekeepers to the platform, as legislators of the relationships within their ecosystems, and as contractual actors participating in the transactions they rule on.¹ Adjacent work proposes distributive equity, the fair allocation of welfare across all ecosystem participants, as a candidate additional consideration for antitrust enforcement where those participants are not equally served by traditional competition-law analysis.²

This paper documents a single worked example: a community-scale platform (Village, operated by My Digital Sovereignty Ltd, Aotearoa New Zealand) whose structural commitments are the enactment of a prior theoretical commitment — that the welfare pathology identified in the research programme is best understood as a values drift pathology, and that structural architecture is the mechanism by which a platform’s declared values can be made sticky enough to resist that drift.

The paper situates Village’s structural commitments inside the Tractatus framework that generated them — a constitutional architecture grounded in Wittgenstein’s sayable / unsayable distinction, Berlin’s value pluralism, Ostrom’s polycentric governance, Alexander’s living-systems principles, and Te Ao Māori frameworks of indigenous data sovereignty — and argues that the overlap between Village’s work and the legal-academic research programme is at the values layer, not only at the structural layer. Both are responses to the same concern: that platform power, left to drift, will be exercised against the welfare of ecosystem participants whose welfare the market does not defend. Village’s answer is an architecture in which values are enforced by the platform’s code rather than asserted in the platform’s marketing.

The author is a single-founder company director, not a legal scholar. The paper’s contribution is documentary rather than theoretical: it offers the research programme one primary-source-rich case to assess, critique, extend, or reject.

Section 1 — Frame and scope

The Taming Ecosystem Power of Platforms through Contract and Competition Law research project at the University of Antwerp Faculty of Law, conducted under principal investigator Jan Blockx and funded by the Research Foundation – Flanders, has proposed and developed an ecosystem-based legal model for the distinctive forms of power platforms exercise inside the ecosystems they host.³ The project’s central analytical move is to treat the platform not as a single actor with market power but as an actor discharging three simultaneous and sometimes conflicting functions: as gatekeeper determining access to the ecosystem; as legislator writing the rules that govern participants within the ecosystem; and as a contractual actor party to the transactions those rules govern. The author of this paper understands the three-function model as the paper’s primary analytical lens and acknowledges the Blockx project as the framing scholarship.

A separate but related contribution proposes distributive equity as an analytical extension: the welfare generated by an ecosystem should be evaluated not only for total magnitude but for how it is distributed among the stakeholder groups the ecosystem comprises, with particular attention to groups whose position in the ecosystem is asymmetric and whose welfare is most vulnerable to the platform’s internal pricing and contractual leverages.⁴ That proposal does not prescribe a specific remedy; it identifies a gap in existing competition, contract, and consumer law where internal welfare distribution is not adequately addressed and proposes distributive equity as one candidate consideration for closing it.

This paper’s author is a single-founder company director who has spent the past two years building the platform described in Section 3 and the theoretical framework described in Section 2. The author is not trained in EU competition law and makes no claim to peer standing with the research programme cited above. The paper is a documentary submission: one worked example, written up by the operator of the platform it documents, offered to the legal-academic community on the footing that the community will exercise its own authority in assessing it.

What the paper is. A documentation of one platform’s public structural and constitutional commitments; an argument that those commitments are the enactment of a prior theoretical position about values stickiness in organisational form; a mapping of that position onto the three-function model and the distributive-equity framing; a disclosure of what is not yet enforced and what depends on founder good-faith; an invitation to the legal-academic community to assess whether the documentary approach offers anything useful to the research programme, and if so, what.

What the paper is not. A general theory of platform constraint; a claim that the platform has “solved” distributive equity; a counter-thesis to existing legal scholarship; a marketing piece for the platform, its operator, or any associated commercial interest; a proposal for regulatory enforcement or legislative change; a peer contribution to the research programme it cites.

Method. The paper’s factual claims about the platform are verifiable from the public artifacts cited in Section 8. Its theoretical claims are grounded in the published philosophical foundations of the Tractatus framework⁵⁶ and in the value-drift and mission-drift analyses the operator has published in the AI Governance for Communities article series.⁷⁸ Where the author has relied on AI-assisted drafting, that assistance is disclosed; the author takes full responsibility for every claim and welcomes correction.

Section 2 — Values stickiness: the argument structure is answering

2.1 The seed insight and its intellectual origin

The theoretical position this paper documents did not arise from a single reflection. It emerged over approximately two years of the author’s pre-Village work on organisational form and digital sovereignty, under the working name Sy.Digital, before the Village platform itself existed. Two documents in that corpus anchor the position in dated form. The first, Sy.Digital Core Values and Principles (STR-VAL-0001, 29 March 2025), articulated a single coherent set of organisational values — sovereignty, transparency, community, and progressive implementation among them — along with a governance framework (STR-GOV-0002, 31 March 2025) that attempted to align all organisational activity to that single set. The second, Agentic Organizational Structure: A New Paradigm for Digital Sovereignty (STO-INN-0002, 22 April 2025), took the next step. It argued that “traditional organizational hierarchies were designed around knowledge control as a primary organizing principle”, that “when knowledge is no longer scarce but universally accessible through AI assistance, the fundamental premise of hierarchical organization breaks down”, and proposed a four-quadrant structure organised around time horizons and information persistence rather than knowledge control. Its tenth section was entitled Beyond Bureaucracy.⁹

The author’s summary of the underlying realisation, recorded a year later in an unsent grant application draft in March 2026, reads in full: “The project was inspired by the realisation that AI’s most significant impact on organisations in the short term would mean that Max Weber was no longer relevant. The value of an organisation could no longer rely on hierarchies of knowledge and skill.”¹⁰ The insight was the same insight that STO-INN-0002 had developed a year earlier without naming Weber directly. This paper is the first to make the argument in Weber’s own terms and to place it in a legal-academic frame.

Max Weber’s theory of organisation, most completely articulated in the posthumously published Wirtschaft und Gesellschaft (1922), located organisational legitimacy in the rational-legal division of expertise: the bureaucratic form organises specialised knowledge into hierarchies of authority whose decisions are legitimated by the expertise the hierarchy embodies. Bureaucracy is efficient because the apex sets direction, the middle layers translate direction into procedure, and the operatives execute procedure under supervision. Each level adds value because each level holds knowledge the level below does not. Knowledge asymmetry supplies the coordination that makes hierarchical authority functional. For a century, the Weberian form has been the dominant model of large organisational activity across government, industry, education, and the platform-age corporate structure alike.

Large language models trained on internet-scale text have, in practical effect, substantially collapsed the knowledge asymmetry on which the Weberian form depends. An operative at the base of a hierarchy can now reach, on demand, expert-level content on any domain the hierarchy once organised around. The apex can draft strategy at the speed of a senior operative. The middle layers whose function was to translate direction into procedure are rapidly discovering that much of the translation task is now automatable. The hierarchical ordering of knowledge and skill that Weber identified as the source of bureaucratic legitimacy is no longer doing the coordination work it once did.

Some substitute mechanism must take its place. What traditional organisations have long claimed would take its place — values, mission, purpose, culture — has historically failed to. Declared values drift under personnel changes, market pressure, competitive imitation, and the gradual erosion that occurs when nobody monitors the distance between what the organisation says it is for and what it does. The failure mode is familiar enough in organisational practice to have become proverbial: “in the end they became what they set out to replace.” Village’s own published analyses document this failure mode under the headings Mission Drift Through Technology Adoption¹¹ and Resisting Drift Toward Global-Internet Norms,¹² and locate the same observation in each case: declared values erode under technological and market pressures unless given a mechanism by which they can persist. Weber’s theory does not supply that mechanism; it did not need to, because the knowledge hierarchy supplied it. In the post-knowledge-hierarchy condition, some substitute is required.

2.2 What “values stickiness” means, and the pivot from monolithic to plural values

This paper uses values stickiness to name the property an organisation has when its declared values are architecturally resistant to drift. An organisation has values stickiness if the mechanisms by which it coordinates action, resolves conflict, and holds participants accountable are themselves structurally bound to its declared values, such that drifting away from the declared values would require breaking the structure rather than reinterpreting a policy document. The claim is not cultural — it is not about what the organisation’s members believe — but structural: about what the organisation’s architecture makes easy, hard, and impossible.

The concept is uncontroversial once stated, and its force derives from the observation that most contemporary platforms do not possess it. A platform whose values exist in a marketing document, a code of conduct, or a published mission statement — while its code enforces whatever is optimal for engagement, revenue, or growth — has no values stickiness. The declared values can drift whenever they become commercially inconvenient, and the drift will not be visible to participants until the pathology the research programme diagnoses is already under way.

The pivot from monolithic to plural values. The author’s own earlier work aimed at values stickiness within a different conceptual frame. The March-April 2025 Sy.Digital governance documents cited above sought to secure a single coherent set of organisational values against drift through the mechanism of a values alignment framework — essentially, a rubric that mapped a unitary value-set to observable indicators across organisational activity, so that a single value-framework could be held stable over time. The intuition behind that framework is familiar in late-modern organisational practice, and it draws on a diagnosis that has been developed across a substantial body of scholarship: that the cultural and institutional drift from communal to individualist value frameworks over the past two centuries — what Alasdair MacIntyre describes as the fragmentation of moral discourse under conditions of late modernity,¹³ what Charles Taylor identifies as atomism as a cultural condition rather than a natural one,¹⁴ what Robert Bellah and colleagues document as the tension between individualism and community in late-modern societies,¹⁵ what Robert Putnam demonstrates empirically as the decline of social capital,¹⁶ what Michael Sandel describes as the procedural republic crowding out substantive community goods,¹⁷ and what Thomas Piketty’s work on long-run capital concentration suggests has attendant economic effects¹⁸ — has produced a condition in which the rights and interests of communities, as distinct from the rights and interests of the wealthiest individual participants, have become progressively harder to defend through market or contract mechanisms alone. The author makes no attempt to adjudicate that substantial scholarly debate. This paper treats it as background context to a narrower observation: the early Sy.Digital work attempted to respond to this condition by holding a single organisational value-set stable against drift, as if the correct answer to values erosion were better alignment to a unitary framework.

The middle-of-evolution realisation. The pivotal realisation that eventually reshaped the Tractatus framework was that the single-framework response was itself misconceived. The deeper claim — drawn from Isaiah Berlin’s mature statement of value pluralism, most explicitly in his 1988 lecture The Pursuit of the Ideal¹⁹ — is that the plurality of genuine human values is not an obstacle to a coherent value system but a condition of human life as such. Berlin’s view, developed across Four Essays on Liberty²⁰ and reaffirmed in the late essay, is that the search for a single harmonious value system is both a philosophical error (because genuine goods are sometimes incommensurable) and a historical danger (because monolithic value systems tend, under pressure, toward coercion). John Gray’s interpretive study of Berlin develops this reading: that pluralism for Berlin is not relativism, nor a second-best alternative to the search for a unified moral framework, but the condition under which genuinely human life is possible.²¹ On this view, plural values are not a feature to be accommodated in an organisation designed for something else. They are the substrate of the organisation’s possibility.

The practical consequence for the Tractatus framework was that values stickiness could not be achieved by holding a single value-set stable. It had to be achieved by holding plural values open — by structurally preventing the platform from collapsing the plurality into a single hierarchy, whether through engagement optimisation, procedural homogenisation, or the accumulated pressure of competitive imitation. The architectural task shifted from “how do we stabilise our values?” to “how do we keep plural values genuinely plural over time?” Section 2.3 describes the Tractatus framework’s response to the reformulated problem, and Section 2.4 describes the three-layer constitutional architecture in which the response is implemented.

2.3 The Tractatus framework as values stickiness engineered

Village’s values stickiness is implemented by a constitutional architecture called the Tractatus framework. The name deliberately invokes Wittgenstein’s Tractatus Logico-Philosophicus (1921). The framework has been documented in the operator’s published philosophical materials,²²²³ and its philosophical foundations are drawn from five traditions separated by a century and a hemisphere: the sayable / unsayable distinction from Wittgenstein, the value pluralism of Isaiah Berlin, the polycentric-governance and commons research of Elinor Ostrom, the living-systems pattern-language work of Christopher Alexander, and the Māori data sovereignty frameworks articulated by Te Mana Raraunga and the Global Indigenous Data Alliance.

From Wittgenstein: the epistemic boundary. Proposition 7 of the Tractatus Logico-Philosophicus — “Whereof one cannot speak, thereof one must be silent” — is not a counsel of defeat. It is an epistemic commitment: some things can be systematised and some cannot, and confusing the two produces nonsense. The Tractatus framework inherits this commitment architecturally. Technical optimisations, pattern matching, information retrieval, measurement — these belong to the domain of the sayable, and the platform’s AI systems are permitted to act autonomously within that domain. Value hierarchies, cultural protocols, grief processing, strategic direction, the resolution of incommensurable goods — these belong to the unsayable, and the platform’s AI systems are not permitted to act autonomously on them. The boundary is enforced not by policy documents but by code: a BoundaryEnforcer service classifies every decision type and blocks AI from acting autonomously on anything outside the technical domain.²⁴

From Berlin: value pluralism as the condition of human life. Isaiah Berlin’s central claim, developed across Two Concepts of Liberty (1958), Four Essays on Liberty (1969), and restated most explicitly in The Pursuit of the Ideal (1988), is that genuine human values are plural, sometimes incommensurable, and frequently in conflict, and that the attempt to reduce them to a single harmonious system is both philosophically mistaken and historically dangerous.²⁵²⁶ As the author understands Berlin, and as John Gray’s interpretive study develops the reading,²⁷ value pluralism is not relativism. It is not a second-best compromise reached when a unifying framework proves elusive. It is the anthropological condition under which the choices that make human life recognisably human become intelligible in the first place. A creature for which every value could be weighed against every other on a single scale would not make choices in the sense that humans make them; a life in which no genuine trade-off between goods ever presented itself would not be recognisably a human life. Plural values, on this reading, are what keeps humans being human.

The implication for AI governance is immediate. No objective function resolves conflicts between incommensurable values. Any system that claims to “optimise” across such values is not neutral — it is imposing a hidden hierarchy, and the hidden hierarchy will drift in the direction of whatever is easiest to measure. The Tractatus framework inherits Berlin’s commitment in three specific architectural forms. First, it recognises six irreducibly different moral frameworks — deontological, consequentialist, virtue, care, communitarian, and indigenous relational — and refuses to resolve conflicts between them algorithmically, instead surfacing each conflict to a human decision-maker together with a transparent account of what each framework would recommend and what each choice would sacrifice.²⁸ Second, it applies asymmetric evidence burdens to value-loaded changes: a change that tightens a safety threshold requires only 60% confidence, while a change that loosens one requires 85% confidence, on the ground that the consequences of error are not symmetric across value dimensions and the costs of false negatives exceed the costs of false positives where values are at stake.²⁹ Third, and most importantly for the post-pivot reading sketched in Section 2.2, it treats the preservation of value plurality as itself a Layer 1 invariant — the platform is not permitted to collapse the plurality into a single hierarchy through any means, including the indirect means of optimisation toward engagement, revenue, or growth metrics that would, over time, do the collapsing silently.

From Ostrom: polycentric governance and nested enterprises. Elinor Ostrom’s Nobel-prize-winning research in Governing the Commons (1990) demonstrated that communities govern shared resources effectively through polycentric governance — multiple independent centres of authority operating without hierarchical subordination, with clear boundaries, collective-choice arrangements, monitoring, graduated sanctions, conflict resolution, and nested enterprises.³⁰ The Tractatus framework inherits this commitment by structuring Village’s governance as a three-layer constitutional architecture in which platform-level universal principles, tenant-level community constitutions, and member-level personal preferences each operate under clearly defined authority and nest into each other without either subordinating or erasing the lower layers. Section 2.4 documents the architecture in detail.

From Alexander: structural integrity as values integrity. Christopher Alexander’s work on pattern languages and architectural theory (A Pattern Language, 1977; The Nature of Order, 2002–2004) argues that living systems exhibit structural properties that emerge from attention to how parts relate to wholes, and that these properties cannot be achieved through top-down planning.³¹ Five of Alexander’s principles are codified into the Tractatus framework as named rules: Deep Interlock (components coordinate through mutual validation rather than isolated approval), Structure-Preserving Transformation (changes preserve essential structure), Gradients Rather Than Boundaries (living systems operate on intensity gradients rather than binary switches), Living Process (the framework evolves from operational experience rather than predetermined specification), and Not-Separateness (governance is embedded in architecture, not bolted on as an afterthought).³² The last of these is load-bearing for the values-stickiness argument. Bolted-on governance can be bypassed under pressure; embedded governance cannot, because the structure within which the platform operates is itself the governance. This is values stickiness stated as an architectural principle.

From Te Ao Māori: kaitiakitanga and rangatiratanga. Indigenous data sovereignty frameworks, particularly Te Mana Raraunga’s principles and the CARE Principles for Indigenous Data Governance,³³³⁴ provide a complete account of the relationship between data, community, and authority that the other four traditions do not supply on their own. Data about a community belongs to that community — not to a platform, not to a researcher, not to a government. The community exercises rangatiratanga (self-determination) over its own data; the platform exercises kaitiakitanga (guardianship) — a fiduciary obligation to protect, not to own.³⁵ The Tractatus framework inherits this commitment architecturally: tenant isolation, community-controlled governance, and sovereign hosting on infrastructure outside US-jurisdiction are not engineering choices that happen to align with indigenous data sovereignty. They are implementations of rangatiratanga as a design invariant.

2.4 The three-layer constitutional architecture

The Tractatus framework is implemented at Village through a three-layer constitutional architecture in which each layer is bound to the layer above it and constrains the layer below.³⁶

Layer 1 — Universal platform principles (immutable). Certain commitments are hardcoded and cannot be overridden by any tenant, administrator, or user. They include tenant data isolation enforced at the data-access layer; the right of any member to leave with their data at any time; consent requirements for data use; no imposed value hierarchy across communities; and the Not-Separateness principle itself — governance is embedded in architecture, not applied as a filter. These are not policies that could be changed through a governance process. They are structural constraints that make certain categories of violation architecturally impossible.

Layer 2 — Tenant constitutional principles (customisable within Layer 1). Each Village defines its own constitution within the bounds established by Layer 1: its tone and communication style, its content moderation norms, its decision-making model (consensus, majority, delegated), its privacy and transparency settings, its cultural protocols, its AI-assistance boundaries. This layer embodies Berlin’s value pluralism in practice: different communities have legitimately different values, and the platform accommodates that diversity rather than imposing homogeneity. A family village and a conservation village serve different kinds of community and are constituted differently because their values are different. The platform does not treat that difference as a bug to be resolved; it treats it as the primary distribution of authority in the system.

Layer 3 — Member personal preferences (individual). Individual members configure their own preferences within the bounds of their community’s constitution: notification frequency, language preferences, AI-assistance levels, privacy defaults for their own content. Layer 3 preferences yield to Layer 2 community standards, which yield to Layer 1 universal principles. Layer 3 also accommodates a documented system of thirteen wisdom traditions (Simone Weil on attention, Stoicism, Care Ethics, Confucian, Buddhist, Ubuntu, Jewish, Islamic, Māori, and others) that shape how AI assistance is framed and delivered, without ever overriding the structural protections of Layers 1 and 2.³⁷

The architecture operationalises values stickiness. A drift-inducing pressure — a commercial incentive to reduce privacy, a personnel change replacing a founder with an engagement-optimised successor, or a competitive imitation that nudges the platform toward Silicon Valley defaults — cannot express itself as a policy change that leaves the code untouched. To drift, the platform must modify Layer 1, and Layer 1 is hardcoded. A platform seeking to drift from its commitments has three options: modify the code, fork the code and operate a different platform, or accept the constraint. The first is publicly visible; the second is an exit; the third is the intended outcome. The architecture does not render drift metaphysically impossible. It renders drift visible, costly, and traceable — which is the most any structural commitment can do.

2.5 Why this matters for the research programme

The Blockx research programme diagnoses the welfare pathology that arises when platform power is exercised against the interests of ecosystem participants who have no market defence against it. Its three-function model is a tool for locating where the pathology is being produced: at the gatekeeper function (extractive pricing, lock-in), at the legislator function (unilateral rule-writing with no participant voice), at the contractual-actor function (conflict of interest as both party and rule-setter). Li’s distributive-equity extension asks about the resulting welfare distribution and proposes a candidate additional consideration for antitrust analysis.

This paper’s claim is that the pathology these analyses identify is best understood as the predictable consequence of values drift in a post-Weberian organisational form. The three functions are the three places where drift shows up; the distributive inequality is the phenomenon that drift produces. The research programme is developing an analytical response. Village is developing an architectural response. The two projects share the same concern — that platform power must be constrained by values that go beyond market efficiency — and they are developing different mechanisms for that constraint. The overlap is at the values layer, not only at the structural layer. The structural commitments that Sections 3 (the platform) and 4 (the AI substrate) document, together with the mappings in Sections 6 and 7 and the audit criteria in Section 8, are not the paper’s thesis. They are the enactment of the thesis. The thesis is that values stickiness is achievable as architecture, that Village’s Tractatus framework is one such architecture, and that at sub-Big-Tech community scale the architecture is auditable from primary-source artifacts without the platform disclosing confidential commercial information.

Section 3 — Village as the case

This section documents the platform’s scope, scale, and architectural commitments with the specificity a reader needs to assess the worked example. The commitments described here are the enactment of the Tractatus framework from Section 2. Each is verifiable from the public artifact cited in Section 8.

3.1 Scale, scope, and stage

Village serves a sub-Big-Tech audience by design. Each community is capped at 200 members by architectural ceiling; the starting configuration is 25 members and growth to 200 requires an explicit add-on subscription. Communities beyond 200 members are directed toward bilateral federation with other Villages rather than unlimited growth within a single tenant.³⁸ The paper makes no claim about Village’s applicability at platform scales larger than the 200-member-per-community ceiling; it is a worked example at community scale, not at platform scale in the Big Tech sense. The 200-member cap is itself a values-stickiness commitment: the platform cannot become the hub for a large network because it structurally refuses to scale any single community past that ceiling, and the scale-driven drift pressures that push larger platforms toward extractive practices are structurally unavailable to Village.

The platform currently supports twelve product types — community, family, whānau, governance, committee, membership, business, episcopal, carpool, conservation, diaspora, and clubs — each reconfiguring the interface vocabulary, default governance structures, and feature emphasis via a vocabulary system that operates over a single codebase. The implication for a values analysis is that Village’s distributional commitments are made at the architectural layer, not per product. A whānau village and a conservation village are served by the same Layer 1 invariants, the same flat-pricing commitment, and the same constitutional self-binding. The pluralism is at Layer 2.

The operating company is My Digital Sovereignty Ltd, a single-founder New Zealand private limited company. The founder is 74 years old and has publicly named this as a structural weakness that the company’s planned Charitable Trust — provisionally named Te Puna Rangatiratanga (The Sovereignty Foundation) — is intended to remediate. AI assistance (Claude, from Anthropic) is disclosed as part of the company’s operational capacity and documented on the company’s About page.

3.2 Architectural commitments

Each commitment named here is verifiable from the public artifact cited. The structural audit table in Section 8 formalises the verification methodology.

Architectural tenant isolation. Each Village is architecturally isolated from every other Village. Isolation is enforced at the data-access layer by a tenant-filter plugin applied automatically to every database query. Cross-tenant queries are refused as a design invariant and the refusal is enforced in the code path, not only in policy. This is a Layer 1 universal principle and is not a feature that can be disabled by an administrator or a future owner without a code change visible in the public repository.

Flat per-community pricing with no per-seat extraction. Village charges a flat price per community rather than per member, with a founding programme offering a 50% permanent rate reduction to early communities, contractually bound not to increase. There is no per-seat fee, no per-message fee, and no tier gated behind member count below the design ceiling. The platform’s commercial interest is therefore in member retention at the community level, not in extracting from growth within the ecosystem. Removing the gatekeeper-function incentive to extract from growth is a values-stickiness commitment expressed through the pricing architecture.

Member cap by design; federation for scale. Communities grow to 200 members via add-ons, beyond which expansion occurs through federation rather than monolithic growth. The 200-member cap is a deliberate sub-Big-Tech design choice. Federation between communities is structured as a bilateral contract between the two communities, with the platform providing infrastructure but not a counterparty position.

Vendor sovereignty. Village’s runtime infrastructure is outside United States jurisdiction. Production servers run on OVH (France) for European tenants and Catalyst Cloud (Porirua, New Zealand) for Oceania and Asia-Pacific tenants. The payment provider is Airwallex (NZ) Limited. The operator does not use Stripe, Google Cloud, AWS, Microsoft Azure, Cloudflare, or any other US-domiciled runtime service. The US CLOUD Act extends US-authority jurisdiction to US-owned infrastructure worldwide; by choosing non-US runtime services, the platform places the data it holds outside that jurisdictional reach as a structural matter rather than as a matter of legal argument.

Constitutional self-binding. My Digital Sovereignty Ltd publishes a versioned, multilingual constitution as the platform’s primary self-limiting instrument.³⁹ The current version (1.2.0, effective 2025-11-20) is published in English, German, French, Dutch, and te reo Māori. The constitution makes explicit what the platform commits to and what it refuses to do — including refusals around data sale, model training on member content, behavioural tracking, proprietary lock-in, and content access by platform administrators. A summary of the operator’s six core principles is also published on the Values page,⁴⁰ and a shorter overview of the philosophical positions under the four themes of human agency, data sovereignty, community first, and radical transparency is published on the Philosophy page.⁴¹

Full data portability and exit rights. Members and communities can leave at any time, taking their data with them in open formats. The deletion commitment in the constitution specifies that when content is deleted, it is removed from production databases, backups, and AI systems — not flagged as “deleted” while still persisting somewhere accessible. Exit and deletion are published commitments backed by code that is reviewable in the repository.

3.3 Governance posture and planned commitments

Three commitments are published as intention rather than accomplished fact and are disclosed as such.

Charitable Trust (planned). The operator has reserved the name Te Puna Rangatiratanga and has prepared a constitutional framework for a New Zealand Charitable Trust that would hold the constitution, the Tractatus governance framework, and succession protocols. Formal establishment is contingent on the maturation of relationships that would give the Trust genuine governance depth rather than a legal shell; it is explicitly not yet incorporated.

Technical Advisory Board (planned). The operator has published the mandate for an independent Technical Advisory Board with a commitment that at least 50% of seats will be reserved for indigenous or Global South representation. The Board is in formation; no members have been publicly named, and the operator’s published position is that the Board will be announced only when it has sufficient depth to be credible.

Community governance voice (planned). Mechanisms for Village subscribers to have voice in platform governance, weighted by cumulative subscription contribution, are published as a concept under development and explicitly not yet built. The operator’s current public position is that this mechanism is more likely to be developed in collaboration with Māori governance researchers than in isolation.

Public disclosure of planning status is itself a values-stickiness move: each planned commitment, if built, addresses a Layer 1 concern that the architecture alone cannot yet resolve. Readers are invited to judge Village on both the enacted architecture and the candour of the planned extensions.

Section 4 — The AI substrate: Village AI as a Situated Language Layer

4.1 Why the AI substrate matters for the thesis

The structural commitments documented in Section 3 describe the platform side of Village. A reader might reasonably ask whether the argument stops there. It does not, and cannot, for a reason specific to the present moment: the platform is operated through artificial-intelligence systems, and those systems are themselves an enactment layer at which values can drift or be held sticky. A platform whose constitutional architecture bound its human operators but left its AI substrate unconstrained would be a platform whose values stickiness was partial at best. The question the research programme is developing analytical tools for — whether platform power is exercised against the welfare of structurally dependent participants — is increasingly a question about the AI that mediates the platform’s interaction with its participants, not only about the business logic written in traditional code.

This section documents what Village has done at the AI substrate. It is drawn substantively from Article 5 of the operator’s published AI Governance for Communities series, Village AI as a Situated Language Layer (April 2026),⁴² which sets out the design principles, architecture, governance framework, training methodology, and security posture of the AI component of the platform. The purpose of this section is not to reproduce Article 5 in full but to situate it inside the values-stickiness argument of Section 2 and to make explicit what a reader of the research programme should take from it.

4.2 What a Situated Language Layer is

Article 5 introduces the term Situated Language Layer (SLL) to name a small, locally-trained language model that runs on community-controlled infrastructure. The article is specific about the choice of word: “in philosophy, situated knowledge refers to understanding that emerges from a particular context, shaped by specific relationships, histories, and values. A Situated Language Layer is AI that knows where it is, whom it serves, and what it should not do — because the community that trained it made those decisions explicitly.”⁴³ The qualifier small is likewise deliberate: “a model small enough to run on modest hardware is a model the community can actually control. A model trained on community content, with community consent, under community governance, is a model whose behaviour the community can inspect, adjust, and hold accountable.”

The architecture is described in Article 5 at a governance level rather than a technical one. The relevant elements for the present paper are five.

Open-weight foundation model. The Village AI begins from an open-weight foundation — currently the 14-billion-parameter Qwen2 family from Alibaba, selected after evaluation on the grounds that the model weights can be inspected by auditors, that the model runs on community-owned hardware without dependency on any single vendor, and that the model can be fine-tuned without the developer’s permission or knowledge. Article 5 documents that the choice of foundation model has already been revised once in practice: Village initially used Meta’s Llama family before migrating to Qwen2 on the basis of superior multilingual performance, particularly for te reo Māori and the European languages the platform supports. The choice of foundation is itself a governance decision, and the operator treats it as one.

Parameter-efficient fine-tuning adapters. On top of the open-weight foundation, Village adds thin adapter layers produced through parameter-efficient fine-tuning. Each adapter encodes community values, governance boundaries, and domain knowledge specific to its community type. Article 5 notes three governance advantages of this approach: adapters are dramatically cheaper to train than full models, making community-sovereign AI economically viable at community scale; adapters can be updated when community values evolve without retraining from scratch; and adapters can be reverted instantly if a training run produces undesirable behaviour. Reversibility is itself a values-stickiness property: a platform whose AI cannot be rolled back has an AI that will drift faster than its governance processes can correct.

Per-product-type specialisation. Village does not run a single AI model serving every tenant. Article 5 documents that the platform deploys per-product-type specialised models, each fine-tuned on the specific vocabulary, governance structures, and cultural context of its community type. Nine specialisations are in production at the time Article 5 was published: whānau, episcopal, community, family, business, and four further specialisations triggered only when the first real tenant of that type exists (conservation, diaspora, clubs, alumni). A 14B community model serves as the fallback for any product type without its own specialisation, and routing is handled by an InferenceRouter that selects the correct model based on the requesting tenant’s product type. Article 5 is explicit that this is a governance design, not simply a performance optimisation: “each community gets the model trained on content most like its own, not a one-size-fits-all general assistant.”

Data sovereignty as an architectural constraint. Community content stays on community infrastructure. Training data is drawn from the community’s own content, stored on its own infrastructure. No queries, responses, or usage data are transmitted to external systems. Article 5 treats this not as a policy that could be changed through a settings menu but as an architectural constraint: “the community can verify these claims because the entire system is auditable.”

Graceful degradation. The routing infrastructure supports fallback from the primary GPU endpoint to a CPU-served model at reduced quality rather than failing silently, and the community is told when this happens. Transparency about capability limitations is itself a governance commitment in Article 5’s framing.

4.3 The AI substrate subjected to the Tractatus framework

The structural commitments in Section 2.3 and Section 2.4 apply to the AI substrate as surely as they apply to the platform logic. This subsection makes the correspondences explicit.

Layer 1 hard red lines are enforced in the AI itself. Article 5 lists four hard red lines that are embedded as architectural constraints rather than as guidelines that might be overridden: the AI must not make decisions for people; the AI must not build behavioural profiles of members; the AI must not optimise for engagement; and the AI must not disclose one member’s content to another without authorisation. Each of these maps to a Tractatus Layer 1 invariant documented in Section 2.4 of this paper. The first corresponds to the Wittgensteinian sayable/unsayable boundary (Section 2.3) — values and value-laden decisions are unsayable in the Tractatus sense and therefore cannot be delegated to machines. The second and fourth correspond to the tenant-isolation and no-cross-tenant-surveillance invariants that are also enforced in the platform’s data-access layer. The third corresponds to the operator’s refusal to adopt an engagement objective function, which is a direct consequence of Berlin’s value pluralism — as Section 2.3 describes, a system optimising across incommensurable values is imposing a hidden hierarchy, and Village’s refusal to optimise for engagement is the AI-layer expression of that refusal.

Guardian Agents verify AI output in a different epistemic domain from generation. This is the direct operational consequence of Wittgenstein’s sayable / unsayable distinction as Section 2.3 describes it. The Guardian Agents documented in Guardian Agents and the Philosophy of AI Accountability⁴⁴ verify AI output using embedding similarity, not additional generative inference. Measurement, not classification. The architecture is such that the AI component that generates responses operates in a domain that necessarily touches the unsayable, while the component that verifies those responses operates entirely in the sayable. The verifier is not another speaker — it is a measuring instrument. This is values stickiness at the inference layer.

Per-community adapters are the Layer 2 enactment at the AI substrate. Just as each Village defines its own Layer 2 community constitution within the Layer 1 universal bounds described in Section 2.4, each Village has its own adapter that encodes the community’s values, cultural protocols, and governance boundaries into the AI behaviour. Per-product-type specialisation is pluralism operationalised at the AI substrate: a family village’s AI is not a policy layer on top of a one-size-fits-all model, it is a differently-trained model whose training-time decisions reflect the family village’s values. This is the AI-layer answer to Berlin’s point in Section 2.3 that no objective function resolves value conflicts across incommensurable values: Village does not run one AI with a value hierarchy, it runs several AIs trained on different value sets.

Counter-training against internet-scale bias is values stickiness applied to the training process itself. Article 5 is explicit that base models carry implicit assumptions reflecting the demographics of the internet’s most prolific contributors, and that these assumptions may conflict with community values. The operator’s response, as Article 5 describes it, is explicit counter-training rather than censorship: where the base model treats efficiency as unconditionally desirable, fine-tuning can shift the default so that thoroughness is valued more highly in the community context; where the base model treats direct communication as the preferred register, fine-tuning can shift the default so that indirect approaches are read as respect rather than as evasion. Article 5 frames the underlying commitment as ensuring that the AI reflects the community’s values rather than the internet’s values, which it describes as the values of no community in particular.⁴⁵ The choice of what to counter-train against is itself a governance decision, documented and audited through the community’s processes rather than delegated to the model’s developers.

Consent is opt-in, granular, revocable, and informed. Article 5 describes the consent regime for AI training on member content in exactly these four terms: opt-in (default is exclusion), granular (members can consent to some uses but not others), revocable (withdrawal triggers retraining without that content), informed (clear non-technical explanations of what training means). Each is a values-stickiness property at the member-relationship layer: the platform cannot drift toward assuming consent it does not have, because the code requires explicit per-purpose consent flags before any training run includes the content. The three AI-purpose consent classes currently documented in the operator’s consent model are ai_triage_memory, ai_ocr_memory, and ai_summarization_memory, each of which a member may consent to or decline separately.

4.4 The rate of AI capability change as empirical context

Article 5 includes a section on security in what it calls a post-Mythos world. The reference is to Anthropic’s April 2026 disclosure of a model it chose not to release publicly because, per the disclosure, it can discover software weaknesses at scale across every major operating system and web browser and produce functioning intrusion code against them. The capabilities were offered through a controlled-release programme (Project Glasswing) to approximately forty large technology companies so they could find and patch their own weaknesses in advance of equivalent capabilities proliferating. Article 5 cites these facts and draws one immediate practical implication: the ability to identify and leverage hidden software weaknesses — previously the preserve of nation-state cyber programmes — will, within a year or two, be reachable by anyone with access to a sufficiently capable model. The barrier to entry falls from millions of dollars and years of expertise to a single model prompt.⁴⁶

The point of citing this in the present paper is not to editorialise about the rate of change in AI reasoning capability, nor to predict any specific outcome. The point is narrower: the empirical fact that AI capability is changing rapidly, documented in the primary disclosures Article 5 cites, is part of the context in which the research programme’s analytical work is being conducted. A research programme developing legal and analytical tools for platform power is doing so while the substrate through which platform power is exercised — the AI that mediates between the platform and its participants — is itself a moving target. That fact does not settle any theoretical question, but it does establish that the structural question — who constrains the AI, how, and through what mechanism? — is a present question rather than a future one. The operator’s view, documented across the AI Governance for Communities series, is that architectural answers to this question are in short supply and that values stickiness at the AI substrate is one of the few approaches that can be implemented by a community-scale platform without waiting for regulation or generic safety tooling to catch up.

Article 5 also documents the operator’s specific security-posture response to the post-Mythos context: dependency audits, a 48-hour patch cycle policy, AIDE file-integrity monitoring on both production servers, encryption at rest using AES-256-CBC, and the continuing absence of US-cloud dependencies — which the article argues functions as a security posture in addition to a sovereignty posture, on the reasoning that “a small, well-defended target is not in the blast radius of the mass-exploitation scenarios that Mythos-class capabilities enable.”⁴⁷ These are concrete operational responses documented in the public record, not speculative positions.

4.5 Relevance to the research programme

The research programme analyses platform power. A platform’s AI substrate is increasingly the mechanism through which platform power is exercised over ecosystem participants — the mediation layer between the platform’s rules and the participants’ experience of them. The question “how is this platform’s AI constrained?” is therefore becoming part of “how is this platform’s power constrained?”, and a paper that mapped Village’s structural commitments onto the three-function model while leaving the AI substrate unexamined would be mapping half of the worked example.

Village’s SLL approach demonstrates one architectural answer to the question: the AI is bound by the same constitutional architecture the platform is bound by. The AI’s hard red lines are Layer 1 invariants. The AI’s per-community behaviour is a Layer 2 constitutional enactment. The AI’s verification operates in a different epistemic domain from its generation, preserving Wittgenstein’s boundary. The AI’s training is subject to the community’s consent and governance processes. The AI’s tendencies are actively counter-trained against internet-scale drift. And the whole substrate runs on community-controlled infrastructure that is inspectable, modifiable, and, if necessary, reversible.

For the research programme, this is worth attention for a specific reason: it is an existence proof that the AI substrate can be subjected to the same values-stickiness architecture as the platform, without waiting for external regulation or generic safety tooling to catch up. It does not settle the question of whether that architecture is sufficient. It does establish that the question is not premature. As with every other claim in the paper, the operator’s contribution is documentary rather than theoretical — the research programme is invited to assess, critique, extend, or reject the documented approach on its own terms.

Section 5 — Māori-grounded principles inside the Tractatus pluralism layer

The Tractatus framework’s pluralism commitment (from Berlin, operationalised in the three-layer architecture of Section 2.4) is not decoration on a Western framework. It is a substantive commitment that one of the plural value systems the platform accommodates is a Māori-grounded framework already in active use. This section names the Māori-grounded principles and shows how each is operationalised at Village as a first-class commitment rather than a feature.

Rangatiratanga — authority and self-determination over one’s own domain — is the organising principle of Layer 1 tenant isolation. A community’s data remains under that community’s authority. The platform exercises kaitiakitanga (guardianship), not ownership. Rangatiratanga appears in the paper’s architectural audit as a first-class design invariant, not as a label applied to a pre-existing technical decision.⁴⁸

Whakapapa — relational knowledge that connects people to each other, to their ancestors, and to their land — is operationalised as the mentor-recognition model in Village’s support services architecture. A mentor’s mana (standing, authority, recognition) is visible through the lineage of villages they have helped to establish, not through gamification badges or quantitative metrics. The operator’s published commitment is that recognition is allocated through whakapapa, not through metric.⁴⁹

Whanaungatanga — kinship through shared purpose — is the relational underpinning of the federation model. Federation between villages is structured as a bilateral agreement between communities that have chosen to be in relationship, not as market clearing on a platform-managed marketplace. The platform deliberately does not interpose itself as a transaction counterparty in the relationships it enables.

Kaitiakitanga — guardianship and stewardship — is the ethical frame the operator uses to describe the founder’s relationship to the platform during the pre-Trust period. The founder is the current kaitiaki of the platform, with the published intention that stewardship will transition to the planned Charitable Trust when the Trust has sufficient governance depth to be credible rather than symbolic.

Koha — gift-based reciprocity — is the operator’s published access model for indigenous communities outside Aotearoa in the later phases of its roadmap, and reflects a values commitment that access to the platform’s most significant services will be decoupled from ability to pay at market rate and re-coupled to relational contribution.⁵⁰

They are not decoration on a Western framework; they are a non-Western distributive framework in which welfare flows according to relational obligation rather than market clearing, and in which recognition is allocated through lineage rather than metric. At sub-Big-Tech community scale, a Māori-grounded framework of this kind operationalises many of the distributive goals the legal-academic research programme has been developing analytical tools for, through mechanisms that are structurally embedded in the platform’s Layer 2 constitutional architecture.

The author has not earned authority in this area and makes no claim to cultural expertise. The purpose of this section is to make the framework legible to a legal-academic reader, to cite the academic articulations that do carry subject-matter authority, and to name the platform’s published commitments so that a reader with relevant expertise can assess them.

Section 6 — The three-function model as three places drift happens

The three-function model distinguishes the platform’s role as gatekeeper (controlling access to the ecosystem), legislator (writing the rules for relationships within the ecosystem), and contractual actor (participating in transactions under those rules).⁵¹ This paper’s claim is that each of the three functions is a place where values drift produces the ecosystem-power pathology the research programme diagnoses. Village’s structural commitments are one enactment of a values-stickiness response to each of the three drift sites.

5.1 Gatekeeper function — drift toward extractive access

A platform acting as gatekeeper holds the keys to access: to membership, to functionality, to exit, to data. The research programme’s concern is that platforms can drift toward extractive gatekeeping — lock-in, switching costs, barriers to exit, differential treatment of structurally dependent participants. This drift is not typically the consequence of an explicit decision; it is the accumulated weight of small decisions each of which is individually justifiable on efficiency grounds. The platform adds a “friction” feature to reduce abuse, and the friction becomes a barrier to exit. The platform raises prices in response to cost pressures, and the price becomes a barrier to entry for the members the platform was set up to serve. The drift is lawful, incremental, and, within a Weberian knowledge hierarchy, coordinated through procedural rationality.

Village’s values-stickiness response to the gatekeeper drift site is structural: a flat per-community subscription price that does not increase with member count, a 200-member ceiling beyond which growth is redirected to federation rather than intensified extraction, architectural tenant isolation that removes the cross-tenant data-accumulation pathway that turns gatekeeping into network-effect lock-in, and published commitments on data portability and deletion backed by code. Each is a Layer 1 invariant. Each requires a code change to violate. Each is visible in the repository. The gatekeeper function’s drift has been structurally constrained, not merely policed.

5.2 Legislator function — drift toward unilateral rule-writing

A platform acting as legislator writes the rules that govern participants’ behaviour and relationships within the ecosystem. The research programme’s concern is that platforms write these rules unilaterally, with no participant voice, no external constraint, and no constraint on the platform’s own behaviour as rule-writer. The drift pattern at this function is the progressive reduction of participant voice in rule-setting and the progressive expansion of platform discretion — not through any single decision, but through the accumulation of small rule changes each of which individually seems reasonable.

Village’s values-stickiness response to the legislator drift site is the constitutional self-binding published in the operator’s constitution and the three-layer constitutional architecture that pins platform rule-making behind a layered authority system. The operator’s Layer 1 universal principles are not a policy document that the operator can amend at will. They are the hardcoded invariants of the platform, and any change requires a code change that is visible in the repository. The operator’s pluralism commitment — that different communities have legitimately different values, and that the platform will not impose a single value hierarchy across communities — is itself a Layer 1 invariant, which means that the operator has structurally foreclosed its own capacity to drift toward Silicon-Valley default homogenisation under competitive pressure. The legislator function’s drift has been structurally constrained by binding the legislator to the constitution that generated it.

5.3 Contractual-actor function — drift toward role conflict

A platform acting as contractual actor is simultaneously party to transactions and rule-setter for those transactions. The research programme’s concern is that this creates an unaddressable conflict of interest: the platform can rewrite the rules of transactions it is party to. The drift pattern at this function is the progressive capture of the rule-writing authority by the contractual-actor role — the platform writes rules that favour its own transactions and discloses nothing until the conflict has fully crystallised.

Village’s values-stickiness response to the contractual-actor drift site is to deliberately restrict the platform’s contractual-actor role. The platform’s direct contractual relationships are only two: the subscription relationship between the operator and the community (flat price, founding rate lock, published terms), and the federation relationship between the platform and any federated community the platform itself participates in as a Village (currently none). The platform deliberately does not insert itself as the contractual counterparty in member-to-member or community-to-community interactions. Federation between communities is a bilateral contract between the two communities, with the platform providing infrastructure but not a counterparty position. The drift pattern cannot play out because the role conflict is structurally foreclosed — the platform’s code does not support the transaction patterns that would produce the conflict.

In each of the three functions, the research programme has diagnosed a drift pathology; in each case, Village has architecturally foreclosed the drift pattern by binding the platform’s behaviour to a Layer 1 invariant that it cannot unilaterally amend. The three-function map is therefore not a mechanical correspondence between Village’s structure and the three functions. It is a statement of values alignment: the Blockx research programme and the Village project respond to a shared concern, and differ only in the mechanism of response — the research programme develops analytical and legal instruments, Village develops architectural and constitutional ones.

Section 7 — Distributive equity as a consequence of values stickiness

The distributive-equity framing asks whether welfare generated by a platform’s ecosystem is fairly allocated among the stakeholder groups within the ecosystem.⁵² This paper’s claim is that distributive equity is the welfare shape that a values-sticky platform produces. If a platform’s architecture prevents the gatekeeper-function drift that extracts from growth, the gatekeeper-function welfare flows to participants rather than to the platform. If the architecture prevents the legislator-function drift toward unilateral rule-writing, the legislator-function welfare is distributed across tenants rather than captured by the platform. If the architecture prevents the contractual-actor drift toward role conflict, the contractual-actor welfare is distributed across the participants to the relationships rather than extracted by the platform as rule-setter. Distributive equity is not a separate goal the platform additionally pursues. It is the predictable consequence of the platform’s values stickiness, once the values in question include the welfare of ecosystem participants.

6.1 Stakeholder groups at Village’s scale

Five groups are relevant at Village’s scale and stage:

1. Operator. My Digital Sovereignty Ltd.
2. Communities (tenant villages). The 25-to-200-member units that subscribe to and operate on the platform.
3. Members. The individual people who belong to a community.
4. Moderators and mentors. Community members who carry additional responsibility, and (under the tuakana-teina roadmap) mentors who help other communities establish themselves.
5. Federation partners. Other communities connected via bilateral federation contracts.

6.2 Distributional commitments per group

Operator. The operator’s distributional commitment is the Sustainable Business Model principle of the published constitution: Village charges what it costs to run the service sustainably, plus reasonable profit, and refuses to sell advertising, data, or feature access. Founder compensation, runway, and any cross-customer subsidies are not currently disclosed. The published intention is that at Charitable Trust formation, the operator-level distribution will become subject to trustee governance rather than founder-only control. The current state is stage-appropriate confidentiality disclosed as such; the gap is that operator-level welfare distribution depends on founder good-faith until the Trust is formed.

Communities. Communities receive flat per-community pricing with a permanent founding rate, full control over their own data and governance, the right to exit at any time with data portability, a hard architectural member cap that protects the community’s scale, the pluralism commitment that the operator will not impose values on the community, and the published refusal to sell community data, train models on community content without consent, or use community information to benefit other customers. The gap: communities do not yet have formal voice in platform-level decisions. Community governance voice is published as a concept under development and is not yet built.

Members. Individual members receive no per-head extraction, full data ownership including open-format export, a deletion guarantee covering production, backups, and AI systems, consent-based AI interaction with time-bounded memory retention, protection against behavioural tracking and advertising-driven engagement optimisation, and the right to leave without penalty. Member-level structural commitments are in production today. The author does not identify a specific member-level gap at this stage, though verification of the deletion guarantee is a task for an external auditor rather than the operator.

Moderators and mentors. Moderators operating under the tuakana-teina framework receive recognition through whakapapa (lineage of communities helped), koha-based reciprocity in the later roadmap phases, and a published commitment to a Māori-led professional-services pathway in the final phase of the roadmap. Phase 1 is in production; Phases 2 through 5 are roadmapped but not yet built. The gap: mentor-level welfare distribution currently depends on Phase 1 and on the good-faith of the operator’s commitment to roadmap the later phases.

Federation partners. Federation is structured as a bilateral contract with explicit termination terms, layered consent, and exit rights. The 200-member cap structurally prevents federation asymmetries arising from one community being orders of magnitude larger than another. Formal asymmetric-protection clauses for cases where a federation partner has significantly different resource capacity are not yet present; the cap does most of the work, but the gap is named.

6.3 Where commitment depends on founder good-faith

Three commitments are not yet structurally enforced and depend on the current single-founder governance:

1. Operator-level welfare distribution before Trust formation.
2. Community governance voice in platform-level decisions.
3. Long-term stewardship transition beyond the single founder’s active period.

The author names these explicitly because a values-alignment analysis that ignored them would be incomplete. The published roadmap addresses all three — Charitable Trust formation for (1) and (3), community-governance-voice mechanisms and the Technical Advisory Board for (2) — but none has been completed at the time of writing. The paper does not claim the gaps are solved; it claims they are named, published, and committed to as future work. The disclosure is itself part of the values-stickiness posture: a sticky-values platform discloses its own unfinished work so that the drift from intention to accomplishment can be tracked publicly.

Section 8 — Structural audit criteria

The paper’s core hypothesis is that a values-sticky platform’s commitments can be audited at sub-Big-Tech scale from public primary-source artifacts, without the platform disclosing confidential commercial or financial information. This section formalises the audit modality as a table of checkable claims, each with the public artifact that evidences it and a falsification path a reader could use to test the claim independently.

Each row can be checked by a reader with public-internet access. Rows marked Partially require code review or traffic observation in addition to reading published text; a reader with those capabilities can complete the audit independently, and the operator welcomes third-party verification of any row.

The author’s claim is not that this audit modality replaces all other welfare-assessment modalities at every scale, nor that it constitutes a complete ecosystem-power remedy on its own. The claim is narrower: that at sub-Big-Tech community scale, the values stickiness described in Section 2 is checkable from primary-source artifacts alone, and that the checkability is itself a distributive-equity signal worth the research programme’s consideration.

Section 9 — Gaps

A values-sticky platform must disclose the points at which its declared values are not yet structurally enforced. This section names five such gaps, each of which appears on a public operator page and each of which is the subject of a published remediation commitment.

Gap 1 — Charitable Trust not yet incorporated. The Te Puna Rangatiratanga Trust, which the operator has committed to as the long-term steward of the constitution and the Tractatus governance framework, is not yet incorporated. The operator has reserved the name and prepared a constitutional framework, but the Trust has no deed, no trustees, and no legal existence at the time of writing. The published position is that the Trust will be established when the relationships that would give it governance depth have matured sufficiently. The gap is real; the mitigation is that the intention is named publicly and the criteria for formal establishment are published.

Gap 2 — Community governance voice is aspirational. No formal tenant council, member assembly, or community-representative body exists yet. The cooperative framing in the operator’s values page and the community-governance-voice concept in the plan page are published as concepts under development, not as shipped features. The gap is real; the mitigation is that the concept has been published with enough specificity for readers to hold the operator accountable to future implementation, and that the Technical Advisory Board is published as a separate accountability channel.

Gap 3 — Operator-level distribution is not publicly audited. Founder compensation, runway, and any cross-customer subsidy flows are not currently disclosed, audited, or governed by any body other than the single founder. The operator’s published position is that this is stage-appropriate: early-stage company confidentiality under New Zealand company law is the norm, and the company’s audit modality is expected to transition to Trust governance at Trust formation. The gap is real; the mitigation is stage-appropriate confidentiality plus a published transition intention.

Gap 4 — Single-founder plus AI succession risk. The founder is 74. The operator’s published position is that this is a structural weakness; the remediation is the Charitable Trust formation plus the Technical Advisory Board, neither of which is yet established. The paper’s author is also the founder and takes the view that naming this gap publicly on the plan page and in this paper is part of the stewardship commitment. A reader assessing the platform’s values-stickiness posture should give weight to the fact that the gap is named rather than concealed.

Gap 5 — Tuakana-teina Phases 2 through 5 are not shipped. Phase 1 (four support channels: AI help widget, owner-and-moderator briefing, feedback channel, introductory video session with the founder) is in production. Phases 2 through 5 (village-to-village mentoring, mentoring network with registered expertise, extension to indigenous communities beyond Aotearoa, Māori-led professional services) are roadmapped but not yet built. The distributive commitment to mentors therefore currently rests on Phase 1 and on the published intention to proceed with the later phases. The gap is real; the mitigation is that Phase 1 is shippable today and can be inspected, and the later phases are documented with enough specificity to hold the operator accountable.

None of these five gaps is concealed. Each appears on a public operator page. Public naming of unfinished commitments is itself a values-stickiness signal: a platform whose architecture makes drift visible also makes the gap between declared intention and current enactment visible. The reader is invited to judge the platform on both the enacted architecture and the candour of the gap disclosure.

Section 10 — Open research questions

The paper’s worked example is offered to the legal-academic community as a documentary submission. The questions below are the ones the author believes the community is best placed to assess, and they are written in a form that aims for concreteness.

1. Is values stickiness the right concept? The paper uses values stickiness to name the property an organisation has when its declared values are architecturally resistant to drift. Is this a useful concept, and does the research programme already have a term for it that the author should adopt? Where the term is wrong or misleading, what is the better phrasing?

2. Is the structural audit modality sufficient at sub-Big-Tech community scale? At the scale and stage described in this paper, is the set of commitments listed in Section 8 sufficient to establish a meaningful distributive-equity posture, or is it an insufficient substitute for modalities that become applicable at larger scale?

3. What additional structural criteria would strengthen the audit? Are there structural commitments the Section 8 table omits and that a reader with competition-law enforcement experience would expect to see?

4. Where does the values-stickiness modality fail? Which ecosystem-power failure modes does it catch, and which does it miss? What are the preconditions under which a platform crosses a scale or governance threshold that forces a transition to different audit modalities?

5. Cross-jurisdiction applicability. The worked example is operated from Aotearoa New Zealand with EU operational presence. Is the architecture replicable in other jurisdictions? What jurisdictional features (contract-law regime, trust-law availability, indigenous-data-sovereignty legal infrastructure) are load-bearing in the example?

6. Non-Western distributive frameworks and EU legal-academic scholarship. How should the research programme engage with non-Western distributive frameworks (Māori data sovereignty, CARE Principles, the broader indigenous data governance literature) that operationalise distributive equity through relational obligation rather than regulatory enforcement?

7. Worked examples as scholarly material. If documentary submissions of the kind represented by this paper were published alongside the primary scholarship, would the research programme find them useful as empirical material, critical foils, or neither?

Section 11 — Methodology, scope, and self-reporting

Worked example, not generalisation. The paper documents one platform operating at sub-Big-Tech community scale, at the early operational stage, from a single-jurisdiction (New Zealand plus European operational presence) position, under a single-founder corporate structure, and with an indigenous data-sovereignty orientation. The findings are specific to that context. Generalisation to Big Tech is not implied; extension to other community-scale platforms is possible in principle but would require its own worked example.

Self-reporting. The paper is written by the platform’s operator. Every factual claim about the platform is subject to verification via the public artifacts cited in Section 8. The author’s position is that the public-artifact-based audit modality is the appropriate response to self-reporting: the reader does not have to rely on the operator’s assertion, because each claim is checkable against an artifact the operator does not control.

AI assistance. The author is a single-founder company director and is not a legal scholar. The paper has been drafted with AI assistance (Claude, from Anthropic), primarily for structural framing, citation discipline, and prose editing. The author takes full responsibility for the content of the paper and for any errors in it. Readers identifying errors are asked to correct the author so that future versions can incorporate the correction.

Replicability. Other community-scale platforms could in principle replicate the structural commitments described in Section 3 — flat per-community pricing, architectural member caps, tenant isolation, vendor-sovereignty selection, public constitutional self-binding, exit rights and data portability — and could replicate the three-layer constitutional architecture described in Section 2.4. Whether they should is a question for them; this paper does not prescribe that they should.

Open source. Extraction and publication of core Tractatus framework modules as EUPL-1.2 open-source libraries is contingent on the outcome of the April 2026 NGI Zero Commons Fund application. The operator’s long-term intention is to release the modules; the intention is made conditional on the funding outcome because extraction and documentation at release quality is itself a substantial engineering task.

What the paper does not measure. The paper does not attempt to measure welfare distribution outcomes (member satisfaction, mentor retention, federation health, community governance participation) because reliable outcome measurement requires longitudinal data that does not yet exist for this platform. The author intends that future work will address outcome measurement in a subsequent paper.

Limits of the author’s authority. The author is not a legal scholar, is not trained in EU competition law or contract law, and does not have the subject-matter authority to assess which elements of the worked example are theoretically interesting and which are trivial. The paper is therefore submitted as a documentary resource rather than as a scholarly contribution, and its most useful outcome would be for a reader with the relevant authority to assess, extend, or correct it.

References

Primary scholarly sources

Blockx, Jan. Taming Ecosystem Power of Platforms through Contract and Competition Law. Research project, University of Antwerp Faculty of Law, funded by the Research Foundation – Flanders (FWO), 2022–2025. Project summary describing the three-function model and the ecosystem-based legal model for addressing excess platform power through contract-law and competition-law solutions.⁵³

Li, Yibo. “Characterising Ecosystem Power: the Use of Pricing and Contractual Leverages.” Utrecht Law Review, Volume 21, Issue 1 (September 2025), pp. 4–18. DOI: 10.36633/ulr.1097. Proposes distributive equity as an additional antitrust consideration.⁵⁴

Philosophical sources cited in the Tractatus framework

Wittgenstein, Ludwig. Tractatus Logico-Philosophicus, 1921. Proposition 7 and the sayable / unsayable distinction. Translated by C. K. Ogden (1922), Routledge & Kegan Paul.

Berlin, Isaiah. “Two Concepts of Liberty”, 1958. Reprinted in Four Essays on Liberty (1969), Oxford University Press. Value pluralism and incommensurability.⁵⁵

Ostrom, Elinor. Governing the Commons: The Evolution of Institutions for Collective Action, 1990. Cambridge University Press. Polycentric governance and nested enterprises.⁵⁶

Alexander, Christopher. A Pattern Language: Towns, Buildings, Construction, 1977. Oxford University Press. Pattern-language methodology. The Nature of Order (Volumes 1–4, 2002–2004), Center for Environmental Structure. Living-systems architectural theory.⁵⁷

Weber, Max. Wirtschaft und Gesellschaft, 1922 (posthumous). Rational-legal bureaucracy and the theory of organisational legitimation. Cited as the theoretical position the post-Weberian argument in Section 2 responds to.

Scholarly sources cited in the monolithic-to-pluralism pivot argument (Section 2.2)

Berlin, Isaiah. The Pursuit of the Ideal. 1988 Agnelli Prize lecture. Reprinted in The Crooked Timber of Humanity: Chapters in the History of Ideas, edited by Henry Hardy, Princeton University Press, 1990. Berlin’s mature statement of value pluralism as the condition of human life rather than a regrettable feature of the moral landscape.⁵⁸

Berlin, Isaiah. Four Essays on Liberty. Oxford University Press, 1969. Including “Two Concepts of Liberty” (1958) and related essays on value pluralism and incommensurability.⁵⁹

Gray, John. Isaiah Berlin. HarperCollins, 1995; Princeton University Press, 1996. Interpretive study arguing that pluralism is Berlin’s central contribution and that pluralism is not relativism but the condition of recognisably human choice.⁶⁰

MacIntyre, Alasdair. After Virtue: A Study in Moral Theory. University of Notre Dame Press, 1981. Diagnosis of the fragmentation of moral discourse under conditions of late modernity and the loss of a shared teleological framework.⁶¹

Taylor, Charles. Sources of the Self: The Making of the Modern Identity. Harvard University Press, 1989. Identifies atomistic individualism as a cultural condition rather than a natural one; develops the argument that modernity’s moral sources are diverse and contested.⁶²

Bellah, Robert, Richard Madsen, William M. Sullivan, Ann Swidler, and Steven M. Tipton. Habits of the Heart: Individualism and Commitment in American Life. University of California Press, 1985. Empirical and interpretive study of the tension between individualism and community in late-modern American society.⁶³

Putnam, Robert D. Bowling Alone: The Collapse and Revival of American Community. Simon & Schuster, 2000. Empirical documentation of declining social capital and the erosion of community institutions.⁶⁴

Sandel, Michael J. Democracy’s Discontent: America in Search of a Public Philosophy. Harvard University Press, 1996. Argument that procedural liberalism has crowded out substantive community goods and that the republican tradition offers a different account of self-government.⁶⁵

Piketty, Thomas. Capital in the Twenty-First Century. Translated by Arthur Goldhammer, Harvard University Press, 2014. Long-run empirical analysis of capital concentration dynamics under modern capitalism.⁶⁶

Pre-Village author artifacts (internal, dated)

Sy.Digital. Core Values and Principles, document code STR-VAL-0001, version 1.0, 29 March 2025. Author’s pre-Village governance document articulating a unitary organisational value-set. Internal working document, cited as a dated artifact of the author’s own intellectual development.

Sy.Digital. Values Alignment Framework, document code STR-GOV-0002, version 1.0, 31 March 2025. Author’s pre-Village framework attempting to align all organisational activity to the STR-VAL-0001 value-set via an alignment matrix. Internal working document, cited as a dated artifact.

Sy.Digital. Agentic Organizational Structure: A New Paradigm for Digital Sovereignty, document code STO-INN-0002, iteration 2, 22 April 2025. Author’s pre-Village whitepaper proposing a four-quadrant reorganisation of organisational structure around time horizons and information persistence rather than knowledge control. Internal working document, cited as a dated artifact.⁶⁷

Indigenous data sovereignty sources

Te Mana Raraunga — Māori Data Sovereignty Network. Principles of Māori Data Sovereignty. https://www.temanararaunga.maori.nz/.⁶⁸

Carroll, S. R., Garba, I., Figueroa-Rodríguez, O. L., Holbrook, J., Lovett, R., Materechera, S., Parsons, M., Raseroka, K., Rodriguez-Lonebear, D., Rowe, R., Sara, R., Walker, J. D., Anderson, J., & Hudson, M. (2020). The CARE Principles for Indigenous Data Governance. Data Science Journal, 19(1), 43. https://www.gida-global.org/care.⁶⁹

Te Tiriti o Waitangi (1840). Foundational document acknowledged in the operator’s constitution and in Section 5.

Primary-source artifacts of the platform described

Constitution of My Digital Sovereignty Ltd, Version 1.2.0, Effective 2025-11-20. Published in five languages at the operator’s constitutional URL.

Philosophical Foundations of the Village Project (Stroh, February 2026). Documentary presentation of the Tractatus framework and its five-tradition philosophical base.⁷⁰

Guardian Agents and the Philosophy of AI Accountability (Stroh, March 2026). Published article mapping Wittgenstein, Berlin, Ostrom, Alexander, and Te Ao Māori onto the production Guardian Agents architecture. CC BY 4.0.⁷¹

AI Governance for Communities, Article Series (My Digital Sovereignty Ltd, March 2026), articles 01–05. Particular reference to Article 02 (Mission Drift Through Technology Adoption) and Article 05 (Resisting Drift Toward Global-Internet Norms).⁷²⁷³

Our Plan (mysovereignty.digital/our-plan.html) — long-term roadmap, Charitable Trust planning, Technical Advisory Board formation statement, community governance voice concept.

Values (mysovereignty.digital/values.html) — six principles of the operator’s constitution.

Federation (mysovereignty.digital/federation.html) — bilateral federation contract template and layered consent model.

Pricing (mysovereignty.digital/pricing.html) — flat per-community pricing, founding-rate lock, 200-member cap with federation as expansion path.

From Help Widget to Global Services: How Village Communities Support Each Other (April 2026) — tuakana-teina five-phase roadmap, whakapapa-not-badges commitment, koha-basis access for indigenous communities.⁷⁴

Tractatus Framework Repository. https://codeberg.org/mysovereignty/tractatus-framework. EUPL-1.2 proposed; current release cadence contingent on the April 2026 NGI Zero Commons Fund application.

Copyright and Licence

© 2026 My Digital Sovereignty Limited, Aotearoa New Zealand. All rights reserved, subject to the licence below.

This work is made available under a Creative Commons Attribution 4.0 International Licence (CC BY 4.0). To view a copy of this licence, visit https://creativecommons.org/licenses/by/4.0/, or send a letter to Creative Commons, PO Box 1866, Mountain View, CA 94042, USA.

You are free to:

• Share — copy and redistribute the material in any medium or format.
• Adapt — remix, transform, and build upon the material for any purpose, including commercial use.

The licensor cannot revoke these freedoms as long as you follow the licence terms.

Under the following terms:

• Attribution — Reusers are required to give appropriate credit to My Digital Sovereignty Limited and to John Stroh as the author of this work, to provide a link to the licence, and to indicate if any changes were made. Credit may be given in any reasonable manner, but not in any way that suggests that My Digital Sovereignty Limited or the author endorses the reuser or the reuse.
• No additional restrictions — You may not apply legal terms or technological measures that legally restrict others from doing anything the licence permits.

Notices:

• You do not have to comply with the licence for elements of the material in the public domain or where your use is permitted by an applicable exception or limitation.
• No warranties are given. The licence may not give you all of the permissions necessary for your intended use. For example, other rights such as publicity, privacy, or moral rights may limit how you use the material.

Suggested citation formats

Full citation (Chicago author-date style).

Stroh, John [ORCID 0009-0005-2933-7170]. 2026. Distributive Equity Through Structure: A Community-Scale Worked Example of Values Stickiness. Version 1.0. My Digital Sovereignty Limited, Aotearoa New Zealand. Published 16 April 2026. DOI: 10.5281/zenodo.19600614. HTML edition at https://agenticgovernance.digital/whitepapers/distributive-equity.html. Licensed under Creative Commons Attribution 4.0 International (CC BY 4.0).

Short citation (in-text).

Stroh (2026)

BibTeX.

@misc{stroh2026distributive,
  author       = {Stroh, John},
  title        = {Distributive Equity Through Structure:
                  A Community-Scale Worked Example of Values Stickiness},
  howpublished = {My Digital Sovereignty Limited, Aotearoa New Zealand},
  version      = {1.0},
  year         = {2026},
  month        = apr,
  doi          = {10.5281/zenodo.19600614},
  url          = {https://doi.org/10.5281/zenodo.19600614},
  orcid        = {0009-0005-2933-7170},
  note         = {HTML edition at https://agenticgovernance.digital/whitepapers/distributive-equity.html. Licensed under Creative Commons Attribution 4.0 International.}
}

Contact

For questions about citation, licensing, substantive engagement with the argument, or correction of errors:

John Stroh, Director, My Digital Sovereignty Limited ORCID: https://orcid.org/0009-0005-2933-7170 DOI (this paper): https://doi.org/10.5281/zenodo.19600614 Email: john.stroh@mysovereignty.digital Publisher website: https://mysovereignty.digital Research site: https://agenticgovernance.digital

Machine-readable licence metadata

The published HTML edition of this paper embeds the following metadata in its document head for automated citation and licence discovery:

<link rel="license" href="https://creativecommons.org/licenses/by/4.0/">
<meta name="dcterms.rights" content="© 2026 My Digital Sovereignty Limited. Licensed under CC BY 4.0.">
<meta name="dcterms.license" content="https://creativecommons.org/licenses/by/4.0/">
<meta name="dcterms.creator" content="John Stroh">
<meta name="dcterms.publisher" content="My Digital Sovereignty Limited">
<meta name="dcterms.dateSubmitted" content="2026-04-16">

Version 1.0 — first reviewed edition. V1.0 incorporates substantive review of Section 5 (Māori-grounded principles) by Dr Karaitiana Taiuru, who requested one specific correction — the removal of a sentence that mischaracterised Te Mana Raraunga and the CARE Principles as “formal academic articulation” when they are authoritative frameworks in their own right and whose original characterisation overlooked the grounding role of Te Tiriti o Waitangi. The correction is live in all five language editions. Further critiques and extensions are welcomed at the address above and will be reflected in subsequent versions. The author has not yet read the referenced Blockx project’s book-length output; citations to the project are drawn from public project summaries, and any future edition incorporating direct book citations will be issued as V1.1 or higher.

My Digital Sovereignty Limited — Aotearoa New Zealand, 16 April 2026.