From a row to a defended thesis, with Dan Dennett as the way in and Claude doing what I could not.
A medically-trained friend said Mounjaro was "cheating". That addiction was a useful label many of us hide behind. That education and a stiff upper lip were all we needed to get a grip. (Okay, I might have made that last part up.) I felt the heat at the back of my neck.
Her reasonable framing of an antithetical position isn't what I heard.
As someone with skin in the game, my incredulous retort — "How can you not believe in addictions?" — was unworthy of her or me.
The evening exploded as expected. I minimised. She strawmanned. We exaggerated. Then I cherry-picked. We misremembered a bit more, relied on anecdote and said 'everyone'. We left each other more entrenched than when we arrived. And no-one won anything.
A rhetorical masterclass it was not.
After a much-shortened date and with hours to spare, it was all I could think about. This was not the first time I had been red-faced and rueful.
So the next morning, filled with coffee and zeal, I asked Claude what I could do better.
Not better at winning. Better at the conversation itself. Because the friend was not wrong about everything — she was reasonable about most things and partly right about some. What I needed was not a sharper sword. I needed a structure I could think inside.
Dan Dennett's Intuition Pumps and Other Tools for Thinking introduced me, years ago, to the four rules Anatol Rapoport set down for how to argue with someone you disagree with. They are deceptively simple:
Most public arguments I see — and almost all of the private ones — skip the first three and go straight to the fourth. Mine certainly did. Hers, in fairness, did too. The architecture of disagreement collapses the moment one side decides the point is to win.
The trick of Rapoport is that the first three rules are not concessions. They are the work. Steelmanning is not a politeness. It is the only way to know what you actually disagree with, rather than the version of the other person you have been arguing with in your head.
I gave Claude the question, the position I held, the position I was arguing against, and the four rules. I asked it to do what I had failed to do at the table: steelman the friend's position in her own voice, find what she taught me, find what was true in what she said, and only then build the case for where I thought she was wrong.
What came back, in about twenty minutes of "hard" work, was a thirteen-page documented microsite. Two hundred and thirteen citations. The strongest version of her position first, voiced as a lab-trained nurse who watches her colleagues prescribe Mounjaro and quietly thinks this is cheating. Then the neuroscience. Then the cost arithmetic — including the inconvenient parts that the libertarian critics get right and the public-health establishment doesn't want to talk about. Then forty years of policy designed on the willpower frame, evaluated against what actually worked.
A three-expert panel — Medical, Policy, Harm Reduction — defends the synthesis. Every numeric and propositional claim is traceable to a primary source on the bibliography page. Every line is open to inline annotation through Hypothesis; every page has a comment thread through Giscus.
The site is the argument I should have had at the table.
Did I write it? Not exactly.
Did I author it? Yes.
Am I this smart? Not on your life.
Am I a bit smarter than I was yesterday? Hell yes.
I have not shown her the site. I'm not sure she will read it. If she does, I'm not sure it will land. The lab-trained nurse who thinks Mounjaro is cheating may continue to think so no matter what arrives in her inbox. That is not how positions held for the reasons hers are held tend to move.
What changed isn't her. It's me.
What I have now — that I did not have twelve hours ago — is a thought I can hold without it slipping. A position written down, defended against the strongest version of the other side, qualified where it needed qualifying, conceded where the evidence demanded conceding. The next conversation like that one will not be the same conversation, because I will not be the same arguer.
I'm not sure what to call it, or indeed if it's any good. That's what the comments are for.
Socratic? Perhaps. Socrates, not yet.