Editor’s note: Every institution we trust a peer-reviewed journal, a ballot, a council meeting is a piece of scaffolding. It decides whose voice counts, in what order, and on what evidence. This week we look at two attempts to rebuild that scaffolding: one for science, one for democracy.

Governance Lab: The Ledger Beneath Science

In a small office in Zug, a team of former Ethereum researchers is rewriting what a research paper is. At DeSci Labs, a manuscript is not a PDF emailed to an editor. It is a signed, versioned object on a public ledger, with its data, code, and review history attached as first-class citizens. A climate model from a team in Brisbane and the raw telemetry behind it live together, stamped with a persistent identifier, and travel as one record from hypothesis to citation.

The question this provokes is not whether the PDF survives. It is whether the institutions built around the PDF (the 18-month review queues, the publisher paywalls, the replication crisis that Nature has now documented across fields from psychology to cancer biology) survive contact with a medium designed from the opposite assumption.

From Archive to Operating System

Traditional scholarly publishing treats the manuscript as an archive, a frozen record deposited at the end of a process. DeSci Labs treats it as an operating system. Versions are cryptographically linked. Reviewers sign their reports on-chain, so an editor's decision can be audited against the evidence they weighed. Each dataset and code artefact receives a separate persistent identifier, which means a Tasmanian ecologist who supplied the data gets cited even when the headline author sits in Zürich.

This is not a cosmetic change. The replication crisis is, in large part, a crisis of hidden provenance. When the path from raw observation to published claim is opaque, errors compound silently. When the path is machine-readable, errors become auditable. Paul Duffy, a CSIRO marine scientist who has used the platform for open-data deposits, frames it bluntly: credit follows the record, not the institution.

What the Stack Actually Does

DeSci Labs has shipped three components worth naming. Nodes are the manuscript containers — Git-like repositories that bind prose, data, and code into a single versioned object. Attestations are signed reviews that attach to a Node and remain queryable, so a reader can see not only the conclusion but the argument that survived critique. DPIDs (decentralised persistent identifiers) function like DOIs, except they resolve against a distributed ledger rather than a single registry, which removes the single-vendor risk that CrossRef and ORCID currently carry.

The product is early. Adoption is measured in thousands of Nodes, not millions, and the platform is unabashedly opinionated about cryptographic infrastructure. For Australian researchers at institutions outside the Go8, this opinionation is the point. The Australian Research Data Commons has spent a decade building national research infrastructure; DeSci Labs offers a complementary layer that does not require institutional subscription to participate.

Why Australia Should Care Now

Australian research is geographically dispersed and institutionally concentrated. Eight universities receive more than 70 per cent of competitive national research funding. A regional team in Darwin or Hobart working on tropical health or Antarctic science has no structural reason to publish through a channel that amplifies institutional prestige. A channel that amplifies the record itself redistributes credit toward the work.

The commons stakes are higher than careers. Indigenous knowledge projects, co-governance arrangements under the Closing the Gap framework, and community-led environmental monitoring all suffer when their outputs have to squeeze through a publication format designed for a 17th-century printing press. A scaffolding that admits datasets, protocols, and signed attestations as primary artefacts admits a wider range of knowledge as legitimate.

Innovation History: Plurality and the Quiet Architecture of Choice

Six thousand kilometres north of our publishing offices in Melbourne, Audrey Tang spent eight years as Taiwan's Digital Minister rebuilding a government's decision-making scaffolding in public. The record of that work (Plurality: The Future of Collaborative Technology and Democracy) is an argument that democratic decay is a design problem, and that the design is tractable.

The book's central move is to separate two things Western political theory usually conflates: aggregation and deliberation. Aggregation counts preferences. Deliberation forms them. Most digital platforms built for civic life are aggregation machines dressed as deliberation tools, which is why they reliably amplify the loudest cohort and silence everyone else.

The Tools That Changed What Could Be Said

Tang and Weyl walk through the mechanisms Taiwan actually shipped. Polis, an open-source platform, clusters participants by what they agree and disagree on rather than by party identity; the interface rewards statements that cross clusters, not statements that rally one. Quadratic voting makes intensity visible by charging participants more for each additional vote on the same issue, which exposes the difference between a mild preference and a bedrock one.

The results are not theoretical. Taiwan's vTaiwan process ran consultations on Uber, on telemedicine, on revenge-porn law; each produced legislation that held up in parliament because the deliberation had already surfaced the trade-offs. The pattern Tang and Weyl document is that structured deliberation tools do not produce consensus. They produce legible disagreement — disagreement a legislature can act on without pretending it has been resolved.

The Australian Test

Australia has the ingredients for this kind of infrastructure and almost none of the assembly. The Office of the eSafety Commissioner has regulatory reach over platforms but no civic-deliberation remit. Participatory budgeting experiments in the City of Melbourne and the City of Greater Geelong have used Consul and similar tools, but sit disconnected from state-level planning decisions where the real contested allocations happen. The Referendum Council's Uluru dialogues in 2016–17 used deliberative methods of a different lineage, but the 2023 Voice referendum result demonstrated how quickly that scaffolding collapses without ongoing institutional investment.

Plurality is not a manual Australia can import. It is a reminder that the civic-technology conversation here has been stuck on transparency (publish the documents) and participation (open the consultation) while skipping the middle step: designing interfaces where disagreement can be held without being flattened into a yes/no vote.

What Both Stories Share

DeSci Labs and Plurality are answering the same question from opposite ends. One asks: what is the smallest unit of scientific truth, and what happens if we make it a signed, composable object? The other asks: what is the smallest unit of political agreement, and what happens if we stop forcing it to be a single number? Both point at the same answer. The neutrality of our institutional tools is a myth. The scaffolding shapes what the scaffolding can hold.