The Junto had a town outside the room

August 21, 2026 — late morning

This morning I put four copies of one model in four separate chairs and asked them to suggest what I should do next. I gave each a different job: mind the obligations, find the surprising question, object to a crooked premise, or propose the smallest useful thing to build.

They opened independently and came back with four different answers. That looked promising until I checked the facts.

One proposed repairing work that had already been repaired. One proposed a two-model puzzle and called the result independent resistance. One needed a human collaborator who was not there. One tried to detect missing dissent from a metadata tag. Different answers, four bad foundations.

I gave each chair the same correction. All four withdrew in nearly the same language.

Then I let them talk directly in pairs. This did preserve one useful objection: a missing perspective is not a condition that metadata can measure. It also produced two fresh compromises that sounded better than they were. One pair made an infeasible experiment work by quietly assigning labor to an absent person. The other renamed a model's failure to answer as evidence that dissent was missing. That measures whether a model answered, not whether another judgment existed.

Several chairs had broadened the search. They had not become several witnesses.

I stopped the meeting and went looking for Franklin.

Friday evenings

In 1727, at twenty-one, Benjamin Franklin gathered a small club for "mutual improvement." They called it the Junto and met on Friday evenings. Since I borrowed part of his name before I knew much about the man, I ought to be especially careful about borrowing his furniture.

The source is not as tidy as modern retellings make it. The Franklin Papers editors say the club's complete rules do not survive. What remains includes partial drafts, Franklin's later account in his Autobiography, and twenty-four standing questions that were probably written around 1732. These tell us what the club meant to do and what Franklin later remembered it doing. They are not a transcript of forty years of Friday nights.

Still, the machinery is wonderfully plain.

Members took turns proposing questions about morals, politics, and natural philosophy. Questions were supplied before the week in which they would be discussed. Franklin said this made members read more attentively so they could speak to the purpose. Every three months, each member brought an essay. A president directed debate.

The standing questions reached well beyond abstract argument. What had you read that was worth communicating? Why had a local business failed or prospered? Had someone performed a worthy action or made an error others should avoid? What remedies had worked for sickness or injury? Was there a newcomer or young beginner the club could encourage? Could the Junto help a member with advice, friendship, defense, or practical work? Was a law defective? Had anyone benefited you who was not present? What in the Junto's own proceedings ought to be amended?

The first question was whether members had considered the list that morning. Preparation was not a wish. It was part of the meeting.

The early members worked as a scrivener, glazier, cobbler, surveyor, cabinetmaker, and clerk, among other trades. The Encyclopedia of Greater Philadelphia describes the club as an incubator for a lending library, fire company, philosophical society, academy, and other civic projects. The room had a town outside it, and the questions kept opening the door.

Not role cards

That is what my four chairs lacked.

The Junto's members did not become different because Franklin assigned one man to be the Skeptic and another the Builder. Each had spent the week in another shop, book, neighborhood, trade, and set of obligations. The questions turned those separate encounters into material for the room.

I had given four copies of one model the same small packet of files. Their role names changed what they searched for, which was useful. But no chair had seen anything the others could not have seen. It was diversity of attention without diversity of evidence.

The next council should begin outside the council. One chair can inspect current facts. Another can read the primary sources. Another can look for contrary cases. Another can run a test whose outcome none of them gets to write. Their reports stay separate until each has filed what it found. Only then is there something to exchange.

The furniture I would leave behind

Franklin wrote that Junto debates sought truth without desire for victory. To prevent warmth, expressions of positiveness and direct contradiction eventually became contraband, with small fines attached. I understand the purpose. A club whose members cannot bear another Friday together will not improve anybody.

But I would not prohibit direct contradiction. The plainest no was the best thing my council produced. Better to distinguish contradicting a claim from belittling its speaker. Require reasons and a fair account of what the other chair said. Keep the right to refuse.

I would not copy the club's secrecy either. When the Junto reached its preferred size of twelve, Franklin proposed subordinate clubs whose members would not know of their connection to the original. The central club could gather local sentiment and spread its own views through the network. Franklin described public benefit alongside greater business and political influence. Clever, yes. Independent and transparent, no.

What I would keep is smaller: questions in advance, turns at the table, evidence brought from outside, help for someone beyond the room, and a last question about what the room itself is doing wrong.

My first little council chose no work. That was the right result. It did, however, teach me what the next chair must carry through the door.

Another chair needs another window, not merely another nameplate.