A Substrate for the Commons

The decentralised relational attestation register as a legal-institutional form

Working paper

Abstract

Law engages cultural works as bounded objects: things in which rights can be held, which can be listed, or which can be protected. This paper begins from a different premise: that works are relational and event-based phenomena, and that a work exists through the relations that produce it, sustain it, and transform it. This paper argues that a jurisprudence of the cultural commons requires a substrate capable of holding relations, and that a decentralised relational attestation register—a public, trust-minimised, permissionless record of works and their relations—is a technical form capable of meeting that condition. The proposed register holds attestations. It preserves claims without adjudicating them, and sits alongside ownership records rather than displacing them. The paper sets out five requirements, distinguishes decentralisation as an institutional form, and identifies operations the register would enable. It then addresses the limits of decentralisation, the separation of cryptographic integrity from legal authority, governance and economics, privacy and cultural sensitivity, and prior art. A prototype using EAS schema on the Celo network demonstrates that the requirements can be instantiated with existing technical features.

1. The bounded work and its records

Law engages cultural works as bounded objects: things in which rights can be held, which can be listed, or which can be protected. However, this account of the cultural work is only a partial account of what a work is. This paper begins from the different premise that cultural works are relational and event-based phenomena, that a work exists through the relations that produce it, sustain it, and transform it.¹ This is not to say that law's construction of the work is without value, but rather that the bounded object of legal administration is a functional conceit made for specific purposes. It is law's engagement with the work as a bounded object that allows copyright, heritage listing and rights protection to attach to a work. Yet whilst the conceit enables such purposes well, nonetheless, it is still a determination rather than a complete account of what a work is; in privileging a view of the work as an ownable asset, it becomes blind to relational aspects of a work, such as what it derives from, who contributed to it, and how it has been taken up and reworked over time and space.

A jurisprudence of the cultural commons takes this blindness as its starting point and argues that it has costs, that it creates communities whose stakes are invisible to law, and works whose lineages cannot be traced. In response to this blindness, it demands that institutions be able to see the work's life, not only its ownership.

This paper argues that a decentralised relational attestation register — a public, trust-minimised, permissionless record of works and their relations — is a technical form capable of meeting that demand. Such a register would sit alongside ownership records rather than replacing them, holding what they do not. Decentralisation, on this account, is not incidental to the form. It is what distinguishes such a register from an institutional catalogue, and it is what makes certain legal operations possible that were previously impractical.

The claims advanced here are theoretical, in the sense that they concern what law could do rather than what it presently does. However, they emerge from a design process. Alongside this paper I have designed and built a prototype register using EAS schema on the Celo network, and the argument put forwards here has taken shape through the work of building it.² The paper's claim is that a register of this kind would make certain legal operations possible, and that the technical means to build one now exist.

2. Records of relations

International law addresses cultural works through several bodies of law and institutional regimes that together encompass copyright, heritage, trade, and the rights of indigenous peoples. Accordingly there is some disparity, with cultural works addressed through a variety of frameworks and institutions. The discussion below identifies general structural features of the current landscape, whilst allowing that this generality should not infer that there is uniform international law governing cultural works.

A general feature of the current landscape is that the principal legal technologies for cultural works operate through a designation of the work as an object with certain characteristics. A work under copyright is an object in which entitlements can be held; under heritage law, it is an object of preservation. Administering these designations necessitates and produces records, and those records take the shape of the designation they serve. Entitlement, for example, requires a record of who holds what, whilst preservation requires a record of listing and protection. Rights require a record of the rightsholder and the interest protected.

Indigenous cultural traditions and knowledge have a broader designation in that they are recognised as collective rights, held by communities rather than by individuals, and grounded in relationship and responsibility rather than in title. However, a community's stake only becomes legally visible when a state or institution grants it a form, and the precise nature of that form derives from institutional categories that may not fully recognise what the community holds.

Such records are held by whichever party has standing to confer authority on them, be that a public registry, a collecting society, a publisher, a platform or a scholarly catalogue. The argument here is not that relations are absent from the current landscape, but that they have no substrate of their own. Provenance notes, scholarly apparatus, community documentation, and licence terms all carry something of the work’s relations, such as the connections through which a work exists, what it derives from, who contributed to it, and how it has been taken up and transformed. However, they appear as annotations to records that serve another purpose, often within the internal documentation of an institution, or embedded in scholarship whose authority depends on its author. Where relations are recorded only as the internal work of institutions, they are available to law through those institutions, and through the conventions those institutions sustain. Operations that need to trace a relation across institutional boundaries, or beyond institutional life, have no substrate of their own to work from.

3. Requirements for a relational record

The following five requirements for a relational record can be identified from the analysis above.

First, it must hold relations as first-class attested claims. Relations would be the primary content of the record, rather than annotations to a record that serves another purpose. A work would be recorded through its relations. What is attested is not the work as an object but the work as a set of connections. The register holds claims about those connections; it does not adjudicate their truth.

Second, it must not require ownership as a condition of recording. A record that requires an owner will fail for works whose ownership is unclear, contested, or absent. It will also fail for relations that exist independently of ownership—the derivation of a work from a source, the contribution of a community to a work's making, the trace of a work's circulation.

Third, it must persist beyond the life of any institution. A record held by an institution depends on that institution for its authority, and often for its survival. For a jurisprudence that addresses harms unfolding over generations (e.g., orphan prevention, heritage preservation, a flourishing cultural commons) the record must be able to survive the institution.

Fourth, it must be inspectable without requiring trust in a central authority. A record held by an institution is available to law through that institution. For law to attach to a substrate without that mediation, the substrate must be verifiable without recourse to the party that maintains it. In practice, this is a matter of degree: the register should be trust-minimised rather than trustless, and verification should not depend on a single party.

Fifth, it must admit claims non-discretionarily. A record that gates admission tends to become a scarce good, and scarcity tends to create markets in access. For a commons jurisprudence, admission cannot be a privilege. Anyone with a claim should be able to attest it, and the record should hold the claim as an attributed statement. Admission to the substrate is permissionless; visibility and legal weight are separate, layered questions.

4. Decentralisation as institutional form

The five requirements could, in principle, be met by a well-designed institutional record. What makes decentralisation foundational is that it changes the institutional character of the record, not just its technical properties. What decentralisation does is remove the record from the ownership of any party, creating a substrate that no institution controls.

A centralised record is, ultimately an institution’s record. It is held by someone, and its authority derives from that someone; it can be altered, withdrawn, or lost with them. A record held by a library is a library’s record; a record held by a platform is a platform’s record. Even allowing for freedom of information legislation and other such balancing legislative directives, as noted above such records are available to others largely on the institution’s terms.

A decentralised record, by contrast, is a substrate. It is held by no single party, and its authority derives from the record’s own integrity. It cannot be unilaterally altered or withdrawn, and it persists independently of any particular host. A record held on a public ledger is not held by anyone in particular; it is a substrate on which claims can be made and preserved. A register of this kind still depends on validators, clients, RPC providers, indexers, front-ends, and the economic conditions that make attestation possible, and none of these dependencies disappears simply because the record is distributed. However, no single party can unilaterally alter the record or deny access to the substrate without coordination or cost. Whilst governance, economics, and infrastructure remain, they remain as conditions of the register's operation rather than as instruments any one party can wield against it. The relevant question, then, is not whether decentralisation is absolute, but whether it changes the institutional form sufficiently to make the register a substrate rather than an institution's record.

This isn't to say that control cant still accrete, despite the decentralised structure, and it is important to identify those points at which control may still accrete. For example, control may accrete to those who maintain the technical infrastructure or to those who mediate access to the record through its front end interface or submission portal. It may also accrete to those who can afford to speak, where transaction fees and the cost of storing data decide who can attest. And it may accrete to those who set the grammar of what can be said — the governance of schemas and base contracts. Design features can help to mitigate against these. For example, multiple clients and open specifications can prevent any single implementation from becoming the record. Meanwhile, plural and self-hosted indexers can prevent visibility from becoming a chokepoint. Cheap data-availability layers (ie L2 networks that hold data at low cost so that it need not be stored expensively on the main chain) can keep the cost of attesting from becoming prohibitive. And, plural front-ends prevent any single view of the substrate from becoming the substrate.

These examples are not exhaustive, of course, but they highlight the important fact that whilst decentralisation changes the institutional form, it does not in itself abolish governance or economics. A register of this kind removes the record from the ownership of any party, but does not allocate authority among those who use it. Therefore, any institution built on this substrate must confront constitutional questions of its own, such as who governs the schemas, who bears the cost of attestation, who curates what is seen and so on. But the value is still arguably concrete. A record that persists beyond institutions can hold relations across generations, which is what orphan prevention, heritage preservation, and the long life of a work require. And a record that admits claims non-discretionarily can hold the claims of parties who have no institutional standing: communities, individuals, scholars, artists, in fact anyone with a claim can attest it, and the record holds it as an attributed statement. This is what makes the record capable of recognising stakes that the current legal landscape does not.

5. What the register can and cannot hold

The register holds attested claims. It does not hold facts in the legal or metaphysical sense. A signed attestation shows who said what, and that the statement has not been altered. It does not show that the statement is true, legally valid, or authoritative. Truth, legal facthood, and obligation are determined elsewhere: by courts, statutes, contracts, institutions, and communities.

The register can hold: provenance traces; attributions and competing attributions; derivations and influences; contributions and community stakes; editions, remixes, and circulation; stewardship claims; and revocation or expiry of earlier attestations. It can hold competing claims side by side, and it can record that a claim is contested.

The register cannot hold: title; legal ownership; enforceable obligation; final truth; or the resolution of disputes. It produces evidence. Law produces consequences.

This distinction may look at first like a concession about the register's reliability. If the register holds claims rather than facts, what does it guarantee? The answer is that it guarantees something specific and refuses something else. It guarantees the integrity of the record: that a given attestation was made by the party it names, that it has not been altered since, and that its provenance is inspectable without recourse to any institution. It does not guarantee the truth of what is attested, and it should not, because the authority to determine truth is exactly what institutions hold and what a shared substrate cannot hold for all of them at once. The register's integrity is cryptographic and procedural; the truth of a claim is a matter for the frameworks that read it.

This is what allows the register to be a substrate rather than an institution. It is also what allows plural legal and community frameworks to attach to the same record without the register itself having to choose among them. A register that decided what was true would have to be one institution's register, and it would be available to others only on that institution's terms. A register that holds claims without adjudicating them can be used by a court, a community, and a scholarly catalogue at once, each reading the same record through its own standards of proof and its own norms of recognition.

6. Operations the register would enable

The following operations are theoretical and depend on legal change as much as they depend upon technical feasibility. However, the technical means to build a substrate that would make them possible does now exist.

1. Supplementary provenance records. A heritage institution could hold its own catalogue; a community could attest, in a decentralised register, its relational stake in the same object. The two records would coexist, and neither would override the other. The community's claim would be public, attributable, and inspectable without the institution's mediation. This addresses a live problem in heritage practice: provenance research often depends on records fragmented across institutions and jurisdictions, and community claims have no form.

2. Cross-institutional coordination without centralisation. Two archives, in different jurisdictions, could attest derivations, contributions, or editions to the same register without either controlling the other's records. A federated catalogue (eg Europeana, which harvests from cultural heritage institutions across Europe) needs a single standard, agreed in advance (eg for Europeana this is EDM, based on Dublin Core), plus a body to maintain the standard. A decentralised register needs a minimal shared grammar (or potentially, at its most minimal, only a convention for referencing schemas), and then permits plural schemas to coexist, because each attestation declares which schema it uses. Two archives can attest in different schemas and the substrate still holds both, without either party having to agree with the other about fields or meaning.

3. Attribution disputes held rather than resolved. Current legal systems force resolution: a work is attributed to one party, and the dispute is closed. A relational register can hold competing claims side by side, so that the record shows attribution to be contested and the contest becomes a legal fact. This is a modest operation, but it is one current institutions cannot perform. Where attribution is genuinely uncertain, such as in heritage, in traditional knowledge, and in transitional justice, forcing resolution produces false clarity.

4. Recognition of community stakes without ownership. A community could record its relational stake in a work without the work being converted into property or into heritage in the UNESCO sense. Indigenous data sovereignty frameworks have, in different ways, argued for recognition of community stakes that stands apart from ownership or heritage designation, and have developed mechanisms for asserting cultural authority over collections and data.³ The register provides a form for such recognition: a public, attributable record of a community's relational stake, held independently of any institution that might otherwise mediate it.

5. Stewardship obligations attached to relations. If a work derives from another, a legal framework could recognise that the derivation carries an obligation. The register holds the derivation; the framework attaches the consequence. The register does not itself enforce anything, but it provides the substrate on which such obligations could be recorded and, if the framework so provides, enforced.

The first two operations require no change to law, and can be performed by institutions and communities today within existing legal frameworks. The latter three require legal change. The register makes attachment possible.

7. The register's scope

The register's work is to hold relations. It therefore sits alongside records that administer entitlement, preservation, and rights. The work of preventing enclosure, enforcing obligations, changing legal status, and building institutions belongs to law, politics, and communities.

Enclosure is a legal and economic operation, achieved through contracts, enforcement, and network effects. The register preserves a record that survives the enclosure, and it makes the enclosure visible as an enclosure.

The register records claims. What law does with those claims is a separate question, and enforcement is a legal function rather than a function of the register.

Every work attested to a register remains subject to the same legal rules. The register is a supplementary record. What it changes is what can be seen.

8. Domains of application

The pattern is general. A decentralised relational attestation register can be instantiated in any domain where relations need to be recorded without ownership as the condition of recording.

• Heritage: community stakes, provenance, restitution.
• Research: attribution, derivation, collaborative production.
• Traditional knowledge: community governance, intergenerational responsibility, non-proprietary recognition.
• Cultural production: attribution, lineage, editioning, remix.

Each domain could register its own schema, defining the fields relevant to its forms of relation. The base schema provides the shared grammar, and the pattern allows communities to extend without modifying the substrate. The register's contribution is to make the substrate available. What is built on it is a matter for the communities and legal frameworks that adopt it.

9. Governance, economics, and anti-spam

A permissionless register raises questions of governance and cost which cannot be indefinitely deferred. Drawing on models of good governance, the register should ideally be layered so that no single layer has to solve every problem. The following layers reflect the architectural requirements I encountered when designing the prototype:

• Protocol layer. Minimal, multi-client, credibly neutral. It provides the substrate for attestation and verification. It should be slow to change and resistant to capture.
• Schema layer. Domain-specific and community-governed. Heritage, research, traditional knowledge, and cultural production will need different fields and norms. Schemas should be extensible without modifying the base substrate.
• Indexing and curation layer. Plural and competitive. Anyone can index; communities and institutions can curate. Visibility is discretionary even when admission is permissionless.
• Legal recognition layer. External. Courts, statutes, contracts, and institutional policies decide what weight to give attested claims.

For a fully functional public model, anti-spam mechanisms are necessary. Fees, staking, attestation bonds, reputation systems, and curation can all play a role. The principle is: anyone can attest, but not everyone must be surfaced. Admission to the substrate is permissionless; visibility is layered and plural.

Funding is also a governance question. Public goods grants, transaction subsidies, cooperative or foundation wrappers, and institutional sponsorship are all possible. The register should not depend on a single funder, just as it should not depend on a single host.

10. Privacy, consent, and cultural sensitivity

Public, immutable records are in tension with personal data, sacred knowledge, consent, and rights to erasure. This tension is structural in that a register that persists because it cannot be altered is inherently hostile to information that ought to be forgettable or restricted. A cultural commons project will invariably be required to confront this tension at some point.

The difficulty is that cultural rules about sharing are plural and contextual. For example, a community may require that the sharing of photographs of deceased relatives be restricted to the immediate kinship group, or that a song be played only at certain times of year. These rules vary between communities, between kinds of material, and over time, and they cannot be averaged into a single access policy.

One instance is Mukurtu, built by the Warumungu with Kimberly Christen because they wanted to share their collections and had no tool that would let them do so on their own terms.⁴ It demonstrates that holding and governing can be separated: the platform holds the material in full, and the community sets the conditions of access, revising them as practices change. The same collection may be open to some and restricted to others.

Mukurtu relies on a central platform to hold the material and apply the community's conditions, so a decentralised infrastructure is beyond its remit. What it nonetheless demonstrates is the separation of governance from the record substrate. A decentralised register takes that separation as foundational, and its task is to hold the record without becoming the party that governs it. It follows that sensitive content should not be placed on the register itself. What goes on the record is a reference: a hash, which confirms a document has not changed without revealing what it says; an encrypted pointer, which links to a record held elsewhere that only a key-holder can read; or a commitment, which binds a party to a statement without disclosing it. Where a claim must be verifiable without being fully disclosed, selective disclosure allows a claim to be partly established without being fully revealed. For example, a community challenging a patent on its traditional knowledge must show that the knowledge existed and was held by it, without publishing the knowledge itself, since publication is the harm the challenge exists to prevent. The fundamental design principle is that the record persists, but that the substance stays off the register, and under the control of whoever holds it.

One principle runs through the register's treatment of restricted material: the relation is recorded and visible, and the content to which it refers remains with the parties. Where an attestation concerns a community, the community holds the key that determines who can read the content and can revise those conditions over time. A party can revoke a claim, and the register records that it was made and later withdrawn — the history remains visible while the claim no longer stands. What cannot be done is to make an attestation disappear as though it had never been made. That is a limit of the form. The CARE Principles — Collective Benefit, Authority to Control, Responsibility, Ethics — state the requirement in general terms: the relation is disclosed, and the content is governed by the parties to whom it belongs.⁵

The register's purpose is to make relations visible. A visible relation can be recognised by law, carried across generations, and attested by parties who have no institution to speak for them. Recording the relation while the content remains with its holders extends what the register can hold: it admits relations that could never be recorded if recording them meant publishing what they refer to.

A substrate that can record a relation only where the content is publishable is not a substrate for the whole commons; it is a substrate for the parts that happen to be safe to publish.

11. Related work and prior art

The hypothetical register proposed here is a combination of elements developed elsewhere. Provenance modelling supplies the vocabulary for describing relations: W3C PROV and PROV-O define a general data model for entities, activities, and agents, and CIDOC CRM and Linked Art perform the same function for cultural heritage documentation.⁶ Attestation protocols supply the means of making claims publicly and verifiably: the Ethereum Attestation Service provides a schema registry and a revocable, expirable record of attestations, and Verifiable Credentials and Decentralised Identifiers provide portable claims verifiable without a central authority.⁷ Content-addressed storage supplies the means of referencing works without hosting them.⁸ Indigenous data sovereignty supplies the governance model for restricted disclosure: Local Contexts TK and BC Labels establish community authority over collections held by institutions, and the CARE Principles state the requirement in general terms.⁹ Archival provenance theory supplies the argument that provenance is constitutive of a record's meaning rather than an annotation to it.¹⁰

What has been missing is a general-purpose register combining these: relations as first-class attested claims, no owner required as a condition of recording, persistence beyond any institution, inspectability without recourse to a central authority, and non-discretionary admission. Existing systems tend to be domain-specific, institutionally hosted, or oriented toward ownership.

12. Conclusion

A jurisprudence of the cultural commons requires a substrate capable of holding relations. Such a substrate was previously impractical. A record held by an institution was an institution's record and as such it could not be non-discretionary or cross-jurisdictional in its own right, persistent beyond institutional life, or inspectable without trust in a central authority.

The technical features now exist to build such a substrate, and the prototype demonstrates, at a basic level, that the requirements can be instantiated with existing technical features. The prototype does not demonstrate that the legal operations will follow, for that depends on legal change, institutional adoption, and community governance. Nonetheless, decentralised ledgers can allow records to be held without any single party controlling them; content-addressed storage can allow works to be referenced without being hosted, and public-key cryptography can allow claims to be attributed and verified. Together these technologies enable a register that meets the five requirements mapped out above.

Limitations include: reliance on a particular blockchain network and its governance; network 'gas' and data-availability costs; the difficulty of indexing and curation; the risk of spam and false attestations; the tension between immutability and privacy; and the fact that a prototype is not a production system. These limitations are named here so that the argument can be assessed on its actual merits.

The register's value is that it makes relations visible at the level of the record. In so doing, it preserves attested claims without adjudicating them, and it is trust-minimised rather than trustless. What it produces, therefore, is evidence. Any consequences attached to that evidence belong to law.

---

Notes