The Representation Problem

The U.S. Congress was designed for a country of roughly 4 million people. The Constitution set a baseline of one House member per 30,000 constituents (U.S. Constitution, Article I, Section 2). Today each member represents about 761,000 — 761,169, by the 2020 Census apportionment (U.S. Census Bureau, 2021) — a 25-fold dilution since the founding. Held to the original ratio, the House would seat roughly 11,000 people.

This is not a quirky historical footnote. The information load on one legislator — knowing the needs, conditions, and preferences of three-quarters of a million people across dozens of policy domains — exceeds any honest model of what representation is supposed to do.

The diagnosis we stand on

George Washington's only substantive comment during the entire Constitutional Convention was to argue for lowering the ratio from 40,000 to 30,000 — he thought close representation mattered that much (Madison, Notes of Debates in the Federal Convention of 1787, September 17, 1787). The 1:761,000 ratio we live under is not the product of democratic theory. It is the residue of the Permanent Apportionment Act of 1929, a compromise born of an urban-rural stalemate, not a principle (Eagles, 1990).

Political science has named the scale problem for decades. Taagepera's "cube root law" finds that effective legislatures tend to scale with the cube root of population (Taagepera & Shugart, 1989) — which for the United States implies a body of roughly 690, far past 435 and still far short of what the founders pictured.

And here the public's view is more revealing than it first appears. Asked flatly whether to add seats to the House, most Americans decline: a 2024 survey found respondents opposed adding 150 members by 56% to 44%, and an earlier Pew Research Center survey found only about 28% wanted the House expanded at all (American Academy of Arts & Sciences / UMass Lowell Center for Public Opinion, 2024; Pew Research Center, 2018). But the same 2024 survey, asked whether smaller districts that let a representative actually learn their constituents' needs would be an improvement, found a majority in favor — 60% to 40% (American Academy of Arts & Sciences / UMass Lowell, 2024). Read those two results together and the lesson is sharp: people do not rally to a mechanism. They rally to the felt outcome — being known, being represented. That distinction is the spine of everything below.

What is not settled — what no single author should pretend to settle — is what to do about it. We keep those two things apart: the problem, which we assert, and the remedies, which we offer as starting points — ideas.

The thing you can build today

Any candidate, any office-holder, any organized group of neighbors can subdivide a district into issue-based constituent groups, each with a coordinator who carries information both directions between the represented and the representative. No law has to change. No one has to grant permission.

"Sign up to speak"

• Define issue areas — broad enough to cover the field, narrow enough to mean something: healthcare, K–12, housing, criminal justice, climate and energy, immigration, wages and labor, technology and privacy, agriculture, transportation, taxation, retirement and Social Security, disability and access, foreign policy, and so on.
• Sign up. A person follows at least one area. The minimum ask: when a real question arises, they will be asked what they think. One area, one response when called. The barrier is as low as it can be.
• Coordinators. Each area has a volunteer — chosen by the group or by the campaign — who relays upcoming questions, aggregates and summarizes responses, carries the group's position upward, and reports back what happened.
• Plain surveys. For each live question: a neutral plain-language summary, two or three preference questions, a free-text field, and an optional confidence rating.
• Accountability. The representative publishes, on a schedule, what constituents said, how they voted, and why — especially when they voted against the group, with the reasoning shown.

This is not direct democracy. It is structured input with a feedback loop. The representative keeps discretion and is held publicly accountable for divergences. The informational gap between the represented and the representative is the bottleneck; this attacks it directly, at human scale.

This is the real thing: a way for people to establish, in practice, that they are represented — and to build the trust, the habits, and the evidence that any larger reform would have to rest on. But even the best proposals are sketches that we hope to eventually throw away and replace with something better.

Precedents (this has been done, in pieces)

• Ward and precinct organization. 19th- and early-20th-century American cities ran on ward-level structures where leaders knew their people. Patronage corrupted the incentives; the informational function — connecting representatives to constituents at human scale — worked (Riordon, 1905).
• Participatory budgeting (Porto Alegre, 1989– ). Citizens allocated real budget through neighborhood assemblies organized by area; the model spread to 1,500+ cities, with measurable gains in service delivery and lower-income participation (Sintomer, Herzberg, & Röcke, 2008; Wampler, 2007).
• vTaiwan (2014– ). Machine-assisted opinion mapping (Pol.is) plus structured deliberation surfaced consensus on contested questions and fed it into the legislative process without replacing legislators (Hsiao, Tang, et al., 2018).
• Citizens' assemblies (Ireland, 2016–2018). Randomly selected citizens, given time and good information, produced high-quality recommendations on hard constitutional questions (Farrell, Suiter, & Harris, 2019).
• Liquid democracy experiments (Pirate Party, Germany/Iceland): direct vote or revocable delegation to a trusted proxy — formalizing the idea that representation can be issue-specific and fluid (Kling et al., 2015).

Don't skip this

Here is the problem underneath the problem.

A representation system does not earn legitimacy by producing outcomes everyone likes. That is impossible. It earns legitimacy when people accept outcomes they didn't choose — including ones they argued against — as binding anyway. The whole question is what makes someone do that.

The answer is not mainly mechanical, and it is not mainly about getting the org chart right. Decades of research on procedural justice find that people's willingness to accept and abide by a decision depends more on whether they experienced the process as fair — whether they had genuine voice, whether they were heard, whether the decision-makers were seen as neutral and trustworthy — than on whether they got the result they wanted (Tyler, 1990; Tyler & Lind, 1992). Voice that changes nothing breeds contempt; voice that visibly enters the reasoning, even when it loses, builds consent.

And it is the felt experience of being represented — not the seating chart — that makes a larger system hold. Build the legitimacy and the structures become arguable in good faith. Skip it and no structure, however elegant, will be accepted by the people it loses.

The convention question

The structural changes that would let representation scale — the size of the House, the relationships between levels of government — probably cannot come through ordinary legislation, for one blunt reason: people with power do not voluntarily dilute it. Every sitting member benefits from the current structure.

Article V offers two paths: a congressional proposal (two-thirds of both houses, then three-fourths of states — how all 27 amendments happened), or a state-called convention (two-thirds of legislatures call it, three-fourths ratify — never yet used, but close). The Balanced Budget effort reached 32 of the required 34 states in the early 1980s before some states rescinded — and the rescission question has never been settled by the courts (Paulsen, 1993). The Convention of States movement and the campaign-finance-focused Wolf-PAC show, from opposite ideological directions, that the mechanism is politically live; Lawrence Lessig's Republic, Lost (2011) argued that structural corruption cannot be fixed from inside the system that has it (Caplan, 1988; Rappaport, 2012; Lessig, 2011).

The real obstacle to a convention is not legal. It is fear — that the Bill of Rights gets gutted, that interests hijack it, that it becomes 1787 again (Caplan, 1988). Those fears are legitimate and have to be met with concrete safeguards, not with reassurance: a convention limited in scope by the calling resolutions, delegates bound by instruction and subject to recall, and above all the three-fourths ratification wall — 38 states must still agree before anything a convention proposes becomes law (Rappaport, 2012). A non-binding shadow convention — a citizens' demonstration project, in the spirit of the large-scale deliberative "America Speaks" town halls (Lukensmeyer & Brigham, 2002) — is the low-risk way to practice the whole thing before anything is at stake, and to find out whether we can imagine anything better than what we already have.

We suspect we can't yet. Which is the point of the next section.

A deliberately unfinished sketch (please beat it)

Read this as a provocation, not a plan: our political imagination is embarrassingly stale relative to what a convention could actually attempt. Even the citizen conventions that have run mostly reproduce the structures we already have. So here is a starting point built to be argued with and, ideally, replaced by something none of us has pictured yet.

The sketch: three branches, each internally tripartite, at three levels of government.

• Three executives — an Administrator (domestic), a Diplomat (foreign), an Arbiter (deliberative tiebreaker and head of state). Plural executives are not exotic: Rome's two consuls held office together for roughly five centuries, 509–27 BCE (Lintott, 1999); the Swiss Federal Council has run as a seven-member collective executive since 1848, with a rotating annual presidency, in one of the world's most stable governments (Church & Head, 2013); the French Directoire failed, but for context-specific reasons rather than any inherent flaw in plural executives (Lyons, 1975); and the Spartan dyarchy paired two kings as a mutual check for centuries.
• Three legislatures — Senate (states), House (national population), and a new People's Assembly below the House, drawn from much smaller constituencies (approaching the old ratio, or some feasible approximation), short terms, primary jurisdiction over daily-life questions, a required first-reading body. This is the constituent-organizing model turned into structure.
• Three judiciaries — general case law, an arbitration track for disputes between government entities (structured mediation and binding arbitration have produced faster, more durable resolutions for complex multi-party disputes; Stipanowich & Lamare, 2014), and an external-affairs court that would treat U.S. treaty obligations as binding law rather than as suggestions, and hold the country's conduct abroad to the same standards it applies at home. France's separation of ordinary, administrative, and constitutional jurisdictions is the standing precedent for splitting distinct legal functions (Bell, 2001).

This specific design is almost certainly not the best design. The three-executive model, the Arbiter, the exact ratios — these are just ideas. If anyone can describe something that works better, we should all lean toward that version.

Yet, it is hard to know what is better or best. One example above might help. I think it is a very important example. The external-affairs court isn't only a design choice — it's tightly tied to a value, one behind the peace to keep abroad: that the United States should be bound by the same rules it asks of others, and not exempt itself simply because it can. A country that holds everyone else to standards it won't accept for itself isn't principled; it's just powerful. Keep those two layers separate. You can reject the value — make the case that the exception is earned — or share the value and beat the mechanism, showing a better way to honor it than a new court. Those are different arguments.

Two things to stand on

After everything above, only two commitments are firm — and read closely, neither is a policy. Both are about how decisions get made.

1. There should be more, and closer, representation. This is a value, and — read as a desire to feel represented rather than a vote for any one mechanism — it is the majority's (American Academy of Arts & Sciences / UMass Lowell, 2024). The mechanism — more seats, sortition, an assembly, fluid issue-based delegation, something unimagined — is wide open, and is exactly the kind of thing no single author should settle.
2. The present-versus-future tradeoff in how we use natural resources has to be collectively decidable. We are altering the conditions for future habitation, and the people who will bear that most — those not yet born — have no vote. We do not claim to know the right balance between living people and future ones. We claim that a tradeoff this hard has no legitimacy unless it is decided through real, representative deliberation, and that right now it isn't being decided that way at all.

Notice what both of these are. Procedures. One says people must be able to establish that they are represented; the other says the hardest tradeoffs must be genuinely, collectively decidable. The thing we are building is not a platform to hand you. It is a way of deciding — and the deciding cannot legitimately live in any one head.

To be clear: I do not stand on any particular blueprint for what government should look like.

Cross-project connections

• Campaign Finance — competitive districts and accountable representation are less vulnerable to concentrated spending; campaign-finance reform and representation reform reinforce each other.
• The American Experiment — declining civic participation is partly a response to representation that doesn't feel responsive. Constituent organizing closes the loop.
• Capitalism Game Maintenance — diluted representation is one mechanism by which concentrated economic interests dominate the rule-making process.

References

• American Academy of Arts & Sciences. (2024). Voters Seek Overhaul of Congress for a Stronger Democracy (survey fielded by the UMass Lowell Center for Public Opinion, Sept. 25 – Oct. 1, 2024).
• Bell, J. (2001). French Legal Cultures. Cambridge University Press.
• Caplan, R. (1988). The Constitutional Convention: A Cautious Recommendation. Center for the Study of Federalism.
• Church, C. H., & Head, R. C. (2013). A Concise History of Switzerland. Cambridge University Press.
• Eagles, C. W. (1990). The Reapportionment Struggle of 1929. Journal of American History, 76(4), 1181–1211.
• Farrell, D. M., Suiter, J., & Harris, C. (2019). Systematizing constitutional deliberation. Irish Political Studies, 34(1), 113–123.
• Hsiao, Y. T., Tang, A., et al. (2018). vTaiwan: An Empirical Study of Open Consultation Process. SocArXiv.
• Kling, C. C., Kunegis, J., Hartmann, H., Strohmaier, M., & Staab, S. (2015). Voting behaviour and power in online democracy. Proceedings of the International AAAI Conference on Web and Social Media.
• Lessig, L. (2011). Republic, Lost. Twelve/Hachette.
• Lintott, A. (1999). The Constitution of the Roman Republic. Oxford University Press.
• Lukensmeyer, C. J., & Brigham, S. (2002). Taking Democracy to Scale. National Civic Review, 91(4), 351–366.
• Lyons, M. (1975). France Under the Directory. Cambridge University Press.
• Madison, J. (1787). Notes of Debates in the Federal Convention of 1787.
• Paulsen, M. S. (1993). A General Theory of Article V. Yale Law Journal, 103(3), 677–789.
• Pew Research Center. (2018, May 31). U.S. Population Keeps Growing, but House of Representatives Is Same Size as in Taft Era.
• Rappaport, M. B. (2012). The Constitutionality of a Limited Convention. Constitutional Commentary, 28(1), 53–78.
• Riordon, W. L. (1905). Plunkitt of Tammany Hall. McClure, Phillips & Co.
• Sintomer, Y., Herzberg, C., & Röcke, A. (2008). Participatory Budgeting in Europe. International Journal of Urban and Regional Research, 32(1), 164–178.
• Stipanowich, T. J., & Lamare, J. R. (2014). Living with ADR. Pepperdine Dispute Resolution Law Journal, 14(1), 1–64.
• Taagepera, R., & Shugart, M. S. (1989). Seats and Votes. Yale University Press.
• Tyler, T. R. (1990). Why People Obey the Law. Yale University Press.
• Tyler, T. R., & Lind, E. A. (1992). A Relational Model of Authority in Groups. Advances in Experimental Social Psychology, 25, 115–191.
• U.S. Census Bureau. (2021). 2020 Census Apportionment Results.
• Wampler, B. (2007). Participatory Budgeting in Brazil. Penn State University Press.