Abstract
The civic-proof series has already used the Public Realm Floor (PRF) to carry the normative lower bound that obtains when digital identity enters the public realm. This floor is constituted by four axes: Arendt’s plurality, Habermas’s validity, Pettit’s contestation, and Mouffe’s agonism. The question this article addresses is whether Danielle Allen’s political equality and power-sharing liberalism should be admitted as a fifth axis.
The answer offered here is no. Allen should not be treated as a fifth axis; the more robust move is to place Allen between the Public Realm Floor and the civic-proof operational concept, as a layer of democratic design and institutional translation. The article names this layer AllenBridge.
The function of AllenBridge is to push the analysis from “has not fallen below the floor of democratic legitimacy” to “does this design extend the people’s capacity to share in power?” For civic-proof, this implies that digital identity must not be evaluated solely in terms of privacy protection, verifiability, and interoperability; it must also be asked whether the design preserves political equality, rights of participation, non-monopoly, co-ownership, and the feedback loop from civic input to public action.
§ 1 The Translation-Layer Problem Outside the Fifth-Axis Reading
The four axes of the Public Realm Floor each carry distinct lower-bound functions. Arendt preserves the condition of the appearance of plural subjects in the public realm. Habermas preserves the condition under which public reasons and validity claims remain examinable. Pettit preserves the condition that those affected can raise contestation and obtain redress. Mouffe preserves the condition that political adversaries can continue to contest each other under shared rules.
Allen’s theoretical concerns are highly compatible with these four axes but occupy a different position. Her power-sharing liberalism is concerned with how democratic institutions allow the people to share power, how rights of participation are preserved, how dominating monopolies in political and economic life are avoided, and how citizens can collectively own institutions. These questions belong to a layer that asks where institutional design should go after the floor has been satisfied; they do not constitute an additional component of the public-realm floor.
Admitting Allen directly as a fifth axis would generate three problems.
First, double counting. Allen’s non-domination would overlap with Pettit, her notion of participation would overlap with Arendt, her epistemic equality would overlap with Habermas, and her reciprocity and co-ownership would simultaneously implicate Habermas and Mouffe.
Second, an explosion in proof cost. The existing PRF is a conjunctive floor of four axes. Adding a fifth axis would require reopening the necessity argument, the bearer matrix, the forward-links, and all subsequent engineering articles of the nineteenth article.
Third, a category-level mismatch. PRF is the legitimacy floor that obtains when digital identity intervenes in the public realm. Allen’s power-sharing liberalism is a broader direction of democratic design that engages political economy, institutional reform, civic education, and public policy. Its altitude is wider than PRF.
For these reasons, the article adopts a different move: preserve the four axes of PRF, and place Allen as a bridge from PRF to digital civic infrastructure.
§ 2 Allen’s Core Reinforcement: Political Equality Beyond Bare Voting Equality
Danielle Allen’s recent work can be read into civic-proof along three lines. The first is the conception of political equality in Justice by Means of Democracy. The second is power-sharing liberalism. The third is the Allen Lab’s democracy renovation agenda and the Digital Civic Infrastructure (DCI) context.
In Justice by Means of Democracy, political equality is not confined to one-person-one-vote, nor to formal political rights. It implicates whether a person can participate in public decision-making on equal footing, enter into public knowledge production, coordinate institutions reciprocally with others, and co-own the direction of democratic life.1 The implication for civic-proof is therefore direct. If a digital identity system makes it impossible for some people to obtain usable credentials, understand the content of disclosures, contest a refusal decision, or see how their input has affected public action, that system has not completed the institutional bearing of political equality—even if it possesses strong cryptography.
Within power-sharing liberalism, Allen pushes liberalism towards rights of participation and the sharing of power. The Allen Lab 2025–26 agenda explicitly poses the question whether institutions can sustain real popular participation and keep policy-making and public institutions from the closed control of elites or technocrats.2 This fills a blind spot in civic-proof. civic-proof can already explain which designs fall below the public-realm floor; AllenBridge further requires us to ask whether the design enables the people to actually share in power.
In technology governance, Allen and co-authors have rendered power-sharing liberalism into technology-governance language in their AI governance roadmap, in particular emphasising that positive liberties—the rights of social participation—should be protected on equal terms with the negative liberties of the liberal tradition.3 This makes Allen’s theory directly usable across the AI agent, wallet, trust list, public AI, procurement, and vendor governance chapters.
§ 3 The Form of AllenBridge
This article proposes that AllenBridge be written as the following bridging condition.
PRF(d) =
plurality(d)
and validity(d)
and contestation(d)
and agonism(d)
AllenBridge(d) =
political_equality(d)
and rights_of_participation(d)
and no_monopoly(d)
and co_ownership(d)
and input_to_action_loop(d)This form preserves PRF’s status as a floor. PRF asks whether a design d falls below the minimum conditions of democratic legitimacy. AllenBridge asks whether the design d can be read as an institutional arrangement that shares power.
Each of the five bridging conditions has concrete content.
Political equality requires that, at the gateway to public action, citizens are not split into those who can use the system and those who cannot. This directly implicates wallet inclusion, alternative paths, supported decision-making, undocumented persons, persons with dementia, and cross-border subjects.
Rights of participation require that digital identity protects the positive capacity of citizens to participate in public life, not only the negative protection of personal data from over-collection. This rewrites how age verification, voter registration, community governance, and public consultation platforms are evaluated.
No monopoly requires that issuers, wallets, verifiers, browsers, operating systems, trust lists, and AI agent workflows cannot be controlled at a single point. This condition connects to Pettit’s non-domination and to Allen’s concern with non-monopoly and the sharing of power.
Co-ownership requires that civic communities, local organisations, marginalised users, and independent monitors have visible slots in the governance structure. If digital identity infrastructure is co-governed only by the state and its vendors, its public character is insufficient.
Input-to-action loop requires that civic input can be traced to public action, administrative response, or institutional revision. Recent work by the Allen Lab / Ash Center on deliberative technologies cautions that more voice without governmental response can deepen cynicism.4 This makes civic receipts not merely an evidence chain but also an interface of public responsiveness.
§ 4 The Three-Phase Structure of Digital Civic Infrastructure and the Locus of civic-proof
The Allen Lab / Ash Center framework on Digital Civic Infrastructure understands digital civic infrastructure as the institutional and technical ecosystem that supports the three phases of Connect, Learn, and Act.5 This three-phase structure is important for civic-proof because the intensity of digital identity differs across the three phases.
In the Connect phase, identity primarily handles persistence, community trust, role, and node relations. Strong identity is not always necessary; excessive demand suppresses open connection.
In the Learn phase, identity primarily handles trusted provenance, information quality, learning records, and participation in public knowledge. A great deal of information access and public learning can be completed under low or no identity. Mandating login as a precondition for reading, discussing, or learning miscategorises public knowledge as a closed service.
In the Act phase, the normative pressure on digital identity is highest. As soon as the system begins to gate eligibility, uniqueness, voting, petitioning, resource allocation, public-service application, delegation authorisation, or grievance redress, identity is no longer a login utility but the gateway to public action.
The civic-proof 0’ article should therefore land primarily at the Act layer. It need not turn Connect and Learn into strong-identity flows; its core task is to construct, at the Act layer, a minimal, verifiable, accountable, low-exposure, redressable proof of public action. AllenBridge further requires that these proof flows not only optimise efficiency but also preserve the people’s capacity to share in power.
§ 5 Rewriting Wallet, Receipt, Agent, and UX
AllenBridge entails four concrete rewrites at the engineering layer of civic-proof.
§ 5.1 Wallet: From Personal Container to Co-Governance Gateway
Standard wallet designs tend to presuppose a single natural person, a single device, a single private key, and a single holder. civic-proof 0’ has already noted that this excludes elderly persons with dementia, persons with disabilities, minors, undocumented persons, and those needing supported decision-making.
AllenBridge pushes this problem one step further. A wallet is at once a personal-data container and a gateway to public action. If some persons cannot enter the wallet because of device, literacy, private-key, document, or caregiving conditions, what they lose extends from convenience to rights of participation. This is the engineering meaning of political equality.
§ 5.2 Receipt: From Audit Trail to Input-to-Action Loop
Civic receipts originally carry an after-the-fact auditable evidence chain. AllenBridge further requires receipts to carry traces of public responsiveness. Users need not only to know what they have done, but also whether their public input has entered a process, whether it has been accepted, whether it has been refused, and whether the refusal reasons are contestable.
This does not imply that every act of participation must yield a policy result. It implies that public systems must not absorb civic input into an unresponsive data pool. When civic proof is used in public consultation, voting, petitioning, grievance, or welfare applications, receipts should, to the extent feasible, surface process status and the path to redress.
§ 5.3 Agent: From Errand Efficiency to Boundaries of Responsibility
AI agents are entering civic workflows rapidly. They can help with form filling, querying, comparison, reminders, and submissions, and may also operate wallets on behalf of users. The AllenBridge question is whether these delegation flows extend civic capacity or whether they hand civic action over to an unquestionable delegate layer.
The AI governance roadmap renders power-sharing liberalism into technology governance. Its implication for civic-proof is that delegation authorisation must preserve human override, scope limitation, revocation, auditability, and a clear allocation of responsibility.3 In public-action settings, AI agents can assist in understanding and operation but cannot substitute for first-personal responsibility, nor for the appearance of legitimate adversaries in the public realm.
§ 5.4 UX: From Usability to Epistemic Equality
A selective-disclosure UX intelligible only to engineers does not satisfy political equality. Users must know what they have disclosed, what they have not disclosed, who can see it, whether it is revocable, and how it can be audited afterwards.
The key here is not only usability but also epistemic equality. Citizens in public action cannot be degraded to button operators. If the UX hands the locus of judgment to invisible defaults, buried settings, or vendor strategies, users lose the capacity to participate from an equal epistemic standing.
§ 6 The Taiwan Locus: TW DIW, Trust Lists, and Civilian Backup
Taiwan’s Digital Identity Wallet (TW DIW) and the on-chain registration of its trust list make the civic-proof problem concrete. If TW DIW is read only as a more convenient login and credential-presentation utility for government services, it remains within DPI thinking. If it is placed inside the DCI frame and AllenBridge, the questions change.
First, whether the TW DIW issuer and verifier ecosystem is genuinely plural. If multi-issuer arrangements exist only on paper, while in practice only a handful of state agencies and large platforms control access, the sharing of power is insufficient.
Second, whether the on-chain trust list produces public auditability. The public chain should make issuer status, trust-root changes, revocations, and version commitment values externally visible, rather than remaining a rhetorical declaration of technological romanticism. If chain selection, contracts, resolution methods, and governance procedures remain opaque, going on-chain does not automatically translate into co-ownership.
Third, whether civilian backup is granted institutional slots. Civic communities, local organisations, non-governmental events, and supporters of marginalised users would have to become rate-limited issuers, verifiers, supporters, or monitors before TW DIW could shift from a government-service utility towards a digital civic infrastructure.
Fourth, whether human responsibility is preserved once AI agents intervene. If delegate flows can fill in forms, submit, sign, and apply on behalf of users, then revocability, auditability, overrideability, and redress must be present at the boundary. This is the intersection of AllenBridge and civic-proof 0’.
§ 7 Revisions Already Back-Coupled into 0’
The 0’ article has been lightly back-coupled in accordance with this article. The revision principle is to admit AllenBridge into the argumentative skeleton of 0’ without rewriting 0’ as an Allen monograph.
First, § 1 of 0’ has been rewritten to make the DCI context of the Allen Lab / Ash Center explicit. DCI now functions not merely as a background label but as the institutional perspective on Connect, Learn, and Act. civic-proof lands most strongly at the Act layer.
Second, § 5 of 0’ has been augmented with an AllenBridge subsection. The point is to explain why Allen does not enter as a fifth axis yet can articulate the institutional purpose of PRF.
Third, § 6 of 0’ has been augmented with an engineering check-list. Wallet, receipt, agent, and UX are now checked not only against privacy and verifiability but also against non-monopoly, intelligibility, response-loop closure, co-ownership, and backup capacity.
Fourth, § 16 of 0’ has admitted power-sharing liberalism into the policy agenda. Procurement, standards, trust lists, vendor pluralism, public investment, and civilian backup have all been re-framed in the language of power sharing.
Fifth, § 17 of 0’ concludes with a three-part structure. The Public Realm Floor gives the legitimacy floor; AllenBridge gives the direction of democratic design; Digital Civic Infrastructure is the field of institutional implementation in which both are realised.
§ 8 Honesty Boundary
First, the article does not claim that Allen replaces Arendt, Habermas, Pettit, and Mouffe. Her role is bridging, not substitution.
Second, the article does not claim that the DCI framework is directly attributable to Allen alone. The DCI framework as published should be attributed to the Allen Lab / Ash Center and its named authors.
Third, the article does not claim that AllenBridge has been empirically calibrated. It is a checking layer for editing and institutional design; case studies are still required.
Fourth, the article does not claim that the Taiwan case can be directly extrapolated. TW DIW and the civilian backup ecology offer only a single field of stress testing under democratic-frontline conditions.
Fifth, the article does not claim that AI democratic tools naturally tend towards power-sharing. AI can lower the cost of participation and can also concentrate power. Its direction depends on governance design, procurement conditions, public oversight, and response-loop closure.
§ 9 Conclusion
The real value of Danielle Allen to civic-proof does not lie in adding a new name to PRF. Her reminder is that the problem of democratic institutions cannot be asked only as whether rights have been violated, but also as whether the people can still share in power. This reminder is especially important for digital identity.
Digital identity is easily absorbed into administrative language. It can be flattened into login, verification, efficiency, anti-fraud, anti-duplication, and cost reduction. Each of these matters, but if the analysis stops here, Digital Civic Infrastructure collapses back into the managerial language of Digital Public Infrastructure.
civic-proof exists to resist this collapse. It must place digital identity at the layer of public action and ask whether a system enables low-exposure participation, verifiable bearing of responsibility, contestable refusal, and co-owned governance. PRF supplies the floor; AllenBridge supplies the direction. Only together do they suffice to articulate how digital identity is genuinely transformed into digital civic infrastructure.
References
- Danielle Allen. Justice by Means of Democracy. University of Chicago Press, 2023. https://press.uchicago.edu/ucp/books/book/chicago/J/bo192735333.html
- Danielle Allen. “Setting the 2025-26 Agenda for the Allen Lab for Democracy Renovation.” Harvard Kennedy School Ash Center, 2025. https://ash.harvard.edu/articles/setting-the-2025-26-agenda-for-the-allen-lab-for-democracy-renovation/
- Danielle Allen, Sarah Hubbard, Woojin Lim, Allison Stanger, Shlomit Wagman, Kinney Zalesne, Omoaholo Omoakhalen. “A Roadmap for Governing AI: Technology Governance and Power-Sharing Liberalism.” AI and Ethics, 2025. https://link.springer.com/article/10.1007/s43681-024-00635-y
- Sarah Hubbard and Darshan Goux. “The Ecosystem of Deliberative Technologies for Public Input.” Harvard Kennedy School Ash Center, 2026. https://ash.harvard.edu/resources/the-ecosystem-of-deliberative-technologies-for-public-input/
- Sarah Hubbard and Cynthia Garcia. “A Framework for Digital Civic Infrastructure.” Harvard Kennedy School Ash Center, 2025. https://ash.harvard.edu/resources/a-framework-for-digital-civic-infrastructure/
- Danielle Allen. “Introducing Power-sharing Liberalism.” RadicalxChange, 2022. https://www.radicalxchange.org/updates/blog/introducing-power-sharing-liberalism/
- Danielle Allen, Eli Frankel, Woojin Lim, Divya Siddarth, Joshua Simons, E. Glen Weyl. Ethics of Decentralized Social Technologies: Lessons from Web3, the Fediverse, and Beyond, 2023. https://philarchive.org/archive/ALLEOD
- American Academy of Arts and Sciences. Our Common Purpose: Reinventing American Democracy for the 21st Century, 2020. https://www.amacad.org/ourcommonpurpose/report