Directory · P
Not sure this is the right entry? Run the check · New here? Start with the premise
Persuasion
2 min read · 416 words
Argument mostly doesn't work, and it's worth knowing why: a frontal challenge to someone's position triggers a defence, and a defence is the one state in which updating can't happen.
Is this you?
Check whatever lands. Nothing is sent anywhere.
Persuasion is the operation by which one system attempts to shift another system’s position.
The mechanism most operators expect to work: superior logic. Build the argument, present the evidence, watch the other operator update their conclusion based on the rational case. This works occasionally. Most of the time it doesn’t, because the other system is not primarily running on logic. It is running on identity, social belonging, prior commitments, emotional state, and a network of beliefs that are more interconnected than the single point being argued.
The other system has what the operator’s system has: a self-model. Frontal challenge to a belief in that self-model triggers the defense the Identity entry described — protection of the structure, not consideration of the input. The harder the challenge, the harder the defense. This is why so many arguments end with both operators more entrenched than when they started. The persuasion attempt activated the defense, which is exactly what doesn’t permit updating.
What works: lower the threat level, work through shared territory, ask questions rather than make assertions, and let the other operator move toward the new position rather than be pushed there. The operator who has space to update can update. The operator who has been pushed must defend, regardless of the merits of what’s being said.
From the chair: most attempts at persuasion fail because they prioritize being right over producing change. The operator wants to win the exchange, demonstrate the other’s error, and have the conclusion delivered in a form that reflects credit on the operator. These goals are compatible with venting, not with persuasion. They are why most arguments produce neither updating nor satisfaction.
The actual operation: find genuine common ground first. Ask about the other operator’s reasoning before introducing your own. Acknowledge what is sound in their position before addressing what isn’t. Offer the new framing as a possibility rather than a correction. Then accept that the updating, if it happens, may not happen now. The other system processes inputs over time. The seed planted today may produce the shift weeks later, often without any acknowledgment that the shift came from this conversation.
If you require the win to be visible, you will rarely produce the change. If you can do the operation and walk away without verification, the change is more likely. Persuasion that lands looks, from outside, like the other operator decided on their own.
The defence is what blocks the update
The mechanism most people expect: build the argument, present the evidence, watch the other person update. That works occasionally. Mostly it doesn't, because they aren't primarily running on logic — they're running on identity, belonging, prior commitments, emotional state, and a network of beliefs considerably more interconnected than the single point you're arguing.
And they have a self-model, like you do. A frontal challenge to a belief inside it triggers protection of the structure rather than consideration of the input. The harder the challenge, the harder the defence.
Which explains the outcome everyone has experienced and nobody expects: arguments that end with both sides more entrenched than when they started. The attempt activated the defence, and the defence is precisely the state in which updating cannot occur.
What does move people is slower and less satisfying. Reducing the threat to the self-model rather than increasing the pressure on the belief — finding genuine common ground first, asking questions that let them locate the problem themselves, and leaving the exchange without demanding they concede. People update far more readily in private, afterwards, with nobody watching.
What to actually say
Nearly all of these work by lowering the threat to the self-model. None of them involve winning, which is the part most people can't give up.
- Before making a case at all
What would change your mind about this? I'm genuinely asking.
- Establishing you're not attacking them
I think we actually want the same thing here. Where I get off is further down.
- Letting them find it themselves
How would you expect that to play out? Walk me through it.
- Conceding something real first
You're right about that part, and I hadn't accounted for it.
- Leaving it open rather than demanding a concession
You don't have to agree with me now. I just wanted you to have it.
- When it's become about winning
I think I'm arguing to win rather than to get anywhere. Can we start that again?
Lower the threat instead of raising the pressure
For a conversation you've had several times without movement.
Real questions
Why doesn't evidence change people's minds?
Because the belief isn't held primarily on evidential grounds — it's connected to identity, belonging, prior public commitments and a network of other beliefs. A frontal challenge to any of that triggers protection of the structure rather than evaluation of your point. The evidence isn't being rejected on its merits; it isn't being evaluated at all, because the defence is running instead.
Why do arguments make people more certain?
Because pressure on a belief activates the defence, and the defence is the state in which updating can't happen — so the harder you push, the more firmly the position gets held. It's also why public arguments entrench more than private ones: conceding in front of an audience costs standing as well as consistency, so the incentive to hold is considerably higher.
What actually persuades people?
Reducing the threat to their self-model rather than increasing the pressure on the belief. Conceding something genuinely right in their position first. Asking what would change their mind. Getting them to walk through their own reasoning, since people locate problems themselves far more readily than they accept them being pointed out. And leaving without demanding agreement, because the update usually happens later and in private.
Why do people update after the conversation rather than during it?
Because conceding in the moment costs consistency and standing, while reconsidering privately costs neither — there's no audience and no visible climb-down. Which means a conversation that seemed to achieve nothing frequently did, and demanding a concession in the room is the thing most likely to prevent it. Judging persuasion by what happens in the exchange will almost always underestimate it.
Should I just stop arguing with people?
Worth separating two aims. If you want to understand their position or to be understood, conversation works fine. If you want to change their mind, argument is a poor instrument and the alternatives are slower and less satisfying. And some conversations aren't about the stated subject at all — the clue is whether anything could in principle change their answer, which is worth establishing before spending an evening on it.
How do I know if I'm arguing to win?
Check whether you're interested in what would change your own mind. If the answer is nothing, you're defending rather than discussing, and the other person can tell. The other tell is enjoyment — if you're enjoying it, you're probably not persuading. Naming it out loud is unusual and surprisingly effective: it tends to drop the temperature for both people at once.