What Has Been Tried

In 2012, Ireland convened a Constitutional Convention that included sixty-six randomly selected citizens alongside thirty-three politicians, tasked with deliberating on potential constitutional changes. The model was extended in 2016 with the Citizens’ Assembly, an entirely citizen body of ninety-nine members selected by lot to deliberate over months on issues including the constitutional ban on abortion, climate policy, and fixed-term parliaments. The Assembly’s recommendations on the abortion question—reached after listening to expert testimony, deliberating in facilitated small groups, and arriving at supermajority consensus—were referred to a national referendum, where the public voted to repeal the constitutional amendment by a margin closely matching the Assembly’s deliberated recommendation. France ran a similar process in 2019 and 2020 with its Citizens’ Convention on Climate, convening 150 randomly selected citizens to develop climate policy proposals. The results have been mixed; the French government adopted some recommendations and diluted others, which itself revealed something important about where the legitimacy actually resided.

These experiments are not solutions. They are limited in scope, occur at the discretion of existing governments, and have not yet demonstrated whether deliberative bodies can scale to ongoing governance rather than episodic intervention. But they demonstrate something important: when ordinary citizens are given time, expert testimony, facilitated deliberation, and the responsibility of producing serious recommendations, they consistently produce more thoughtful, less polarized, and more legitimate outputs than the conventional procedural systems they supplement. The hypothesis that democratic deliberation requires only voting and representation has been tested, and the results suggest deliberation itself—structured, sustained, embodied, and facilitated—is doing work that elections cannot.

Taiwan has been experimenting in a different design space. Since 2014, the country has used a deliberative platform called vTaiwan, built originally in response to the Sunflower Movement protests, to allow citizens, civil servants, and stakeholders to surface and develop consensus on contested policy issues. The platform uses an algorithmic tool, introduced briefly in Chapter Six, that maps the actual structure of public opinion by identifying both common ground and genuine disagreement, allowing communities to see where consensus exists rather than only where conflict is loudest. Recommendations emerging through vTaiwan have shaped regulations on ride-sharing, alcohol sales, and platform economy issues. The model is partial. It works alongside conventional government rather than replacing it. But it demonstrates that technological mediation of deliberation, designed against the patterns that social media optimizes for, can produce outcomes the dominant informational architectures actively prevent.

Both Ireland’s Citizens’ Assemblies and Taiwan’s vTaiwan operate at the scale of nation-states, embedded within and accountable to existing institutions. They are not the answer to civilizational governance at planetary scale. What they reveal is that the design space for coherent governance is much larger than the contemporary political imagination usually permits, and that specific mechanisms—random selection, facilitated deliberation, structured exposure to expert testimony, algorithmic mapping of agreement and disagreement—can produce legitimacy of a kind that conventional electoral systems increasingly cannot. The work is not to invent a single new governance architecture that will replace existing institutions. The work is to develop coordination layers that allow more coherent decision-making to emerge at the scales where it is needed, and to learn from the partial successes and failures of the experiments already underway.

There is older material to draw on as well. Many Indigenous governance traditions have practiced forms of subsidiarity, council-based deliberation, and consensus protocols for centuries—the Haudenosaunee Confederacy’s Great Law of Peace being one of the most documented, though not the only example. These traditions are not romantic relics. They are working governance systems that solved coordination problems modern political theory still struggles with, including the problem of how to make decisions across multiple distinct communities without erasing their distinctness. The fact that mainstream political science has spent most of the past two centuries ignoring this material is itself diagnostic of the assumptions modern governance was built on.