A Constitution You Can Hear

Published on 29 September 2026 by David

In September we took a radio station to a law conference.

The Critical Legal Conference 2026 met at Westminster Law School in London from 9 to 11 September. Its theme was Protocols: Infrastructures of the Normative. One of its streams, Sound and Law, treated sound as a way of doing legal research, not only as something to research. It asked whether law can be studied and taught by sounding it out.

That is the question Hitchhiker FM was built to ask. Our session, Hitchhiker.fm: A Constitution You Can Hear, ran on the Wednesday afternoon. Mateusz Kuderski and PrivacyMage are co-authors of the work but were not in London, so I presented it. The same room held our installation for all three days. This post reviews what we proposed, what we built for it, and what is still to come.

What We Proposed

Our abstract asked a simple question. What if a constitution were composed not around a table, but by voices moving through a shared soundscape?

Hitchhiker FM is a 24/7 internet radio station, a sound artwork and a constitutional experiment. Participants contribute spoken reflections, stories, debates, poems and songs. Those recordings become the archive from which a living constitution is composed, argued over and rewritten, by human lawyers, AI systems and characters from science fiction. We call that archive the Sound of Law.

The line that carried the proposal was this one: we treat listening as a constitutional protocol. Sound is not a record of the law being made. It is the medium in which it is made.

The Station

Hitchhiker FM has been on air since 4 July 2026. It runs on hardware we own, and you can listen now.

The station opens with a slowly spoken ident, followed by 42 in Morse. By the conference the Morse had become a small instrument of its own. A Morse plugin keys wiki pages onto the air. The first pages it sent were three constitutional “spells”: Bhutan’s Gross National Happiness, Ecuador’s Rights of Nature, and Japan’s renunciation of war. See the Spell Book.

Next to the main stream runs Marvin Station, where a robot reads the wiki, reluctantly.

Mapping the Constitutions

You cannot hear the world’s constitutions until you can hold them. So in the days before the conference we built a map.

A register of 240 constitutions, drawn from the Constitute project, 188 of them in force. It holds 185 full texts, and gives every clause a stable address such as bhutan-2008:5.3. UK instruments from Magna Carta to the Bill of Rights sit beside them. See constitution.legalcommons.org.
A clause library of more than 300 topics, so that one right can be read across every country that grants it. For example, the Right to Amparo.
The Constitution Globe, which colours the world by the age, length, amendments, rights and topics of its constitutions.
A graph map that draws the constitutions and clauses as one linked landscape: 925 pages in five regions.
A fragment mill that scored more than 300,000 passages for how distinctive they are, so that the clauses worth hearing rise to the top.

Reading Them in 3D

The map became a city you can walk through with your ears.

The Hear plugin turns the register into a Wiki City, drawn in 3D in the browser. Each constitution is a house. As you approach a house, it reads one of its clauses aloud, and the voices of its neighbours rise and fall around you as you move. There are 239 voiced houses, levelled so that no country shouts over another. A multilingual edition reads 49 constitutions in the language in which they were enacted, across eight languages.

During the conference itself we added Hear Flight: a crafted flight across the city, shown full screen, with a limiter so that a crowd of constitutions never clips. See the Hear Crowds demo and the Hear Flight demo.

It has an older root. Rolf Gehlhaar’s Sound=Space (1984) is an instrument you play by moving around an empty room. In 1994 and 1995 I made Sound Space installations and dance work with him. The Sound Space wiki picks that thread up again. The Wiki City is the same idea, applied to law.

Mateusz Kuderski built the other half of the stage: Aether, his spatial audio studio, which lays generated sound out in space. Mateusz was not in the room, but Aether was. We ran it live in the session. Since the conference it has become a wiki plugin, so that one page can hold a whole lineup of sound.

The Voice of Nature

A constitution that only humans can speak into is only half a constitution. Ecuador’s constitution gives nature rights. New Zealand’s Whanganui River is a legal person, Te Awa Tupua. Colombia’s Constitutional Court has said the same of the Atrato River, and Spain’s parliament of the Mar Menor lagoon. So the real question is not whether nature has standing. It is who speaks for the river, and how we would know if they were getting it wrong.

That is what animal.fm is for.

CQ WHALE is a whale channel on the same station. It carries sperm whale codas, blue, fin and humpback whale calls, and relays from live hydrophones in orca waters, all from open research collections. It also carries codas that a machine made. We generated them with WhAM, a model from Project CETI. The station never lets the two blur. Every sound is labelled on air as real, transformed, synthetic, signal or inference. A synthetic coda announces itself plainly: not a whale. Every recording passed a public rights audit before it went on air.

That labelling is the legal point. If AI is going to speak for natural persons, then the first protocol is honesty about which voice is which. See Real and Generated.

Behind the channel is nature.animal.fm, a research wiki on the rights of nature. It works through the cases, from Te Awa Tupua to Martuwarra in Australia, and asks who speaks for the river. Alongside it we built small species instruments, from the túngara frog to the starling, that let people play with how other animals signal.

In the Room

The installation ran for all three days. It showed the Wiki City, the Constitution Globe and the whale channel, with the station playing throughout. Hear Flight joined them once we had built it, part way through the conference.

The session was set up to become a live constitutional assembly. People would record their interventions, argue over the wording and hear the result remixed on air. The room and the time did not allow it. We presented the work and ran Aether live instead. The assembly is still the plan.

What Is Still to Come

Our abstract promised more than we have built, and it is worth being plain about that.

The 42-week process of micro-assemblies, weekly panels and a constitution rewritten on air has not started. The proper launch is the Don’t Panic Festival on 5 November.
Computational law in Lexon, turning spoken proposals into legal text, is planned for the same launch and for January 2027.
Voice from everyone. Every participant’s recording should reach the air. The doors for contributing sound are sketched but not yet open. The wider plan is a network of stations, one for every Embassy.

The conference gave us the room and the deadline to build the instruments. The next job is the chorus.