Home | Search

2026-03-29

original post written mostly by 2026-01

Four hypotheses for how persuasion works

See also

• Persuasion on ASI risk for mass audience should probably target those undergoing traumatic life events
• Morality of persuasion around ASI risk

How seriously should you take this post?

• I have spent atleast a few months unsuccessfully running an anti-ASI youtube channel, organising multiple protests solo and attempting in-person persuasion. I have also read some books about previous protest and civil disobedience movements.
• I am probably better at persuasion than the average person on LW. However, I don't think I am that good at it if you compare me to people who are actually good at persuasion. I think there are many levels of skill far above mine. If you are doing persuasion full-time and you're actually good at it, you'll quickly become way better than me and I will no longer have any useful advice for you.

Here are my four competing hypotheses for how persuasion works, as of 2026-03

1. Lesswrong-style rationality

• Assumption - People are rational about how they formed their worldview
• Potential approaches to persuasion
  • Find the double crux between your worldview and theirs. Find datapoints and arguments based on these datapoints, that target the double crux. Provide this information to the other person.

2. Traumatic life events

• Assumption - People don't change the core of their worldview until they personally suffer
• Potential approaches to persuasion
  • Wait for them to personally suffer. Then, when they are vulnerable, offer your worldview as a solution to their problems. They will cling to it tightly, because you are their saviour. Talk more on topics of social dark matter, which other people seem unwilling to talk about. Commonly seen with religious/spiritual leaders.
  • Actively cause them to suffer (presumably while publicly denying the whole time that you caused them to suffer). Then, when they are vulnerable, offer your worldview as a solution to their problems.

3. Groupthink

• Assumption - People copy-paste the worldview of their family / friends / social media role models / etc
• Potential approaches to persuasion
  • Trace the deference chain until you find someone who isn't just doing groupthink. Use some other persuasion approach on that person.
  • Use extreme confidence to fool people into thinking everyone in the group already believes you. If you fool enough people, now the entire group does in fact believe you. Also known as self-fulfilling prophecy. Also used by con men (full form of con man is confidence man). Also common in shifting Overton windows.
  • Go persuade people who lots of people in society already defer to, such politicians, religious leaders and social media influencers. Create common knowledge.
  • Use some other persuasion approach on them first, to get them to change who their ingroup is. For example, you might use another approach to persuade them to change their job, country, friends, etc

4. Krashen's input hypothesis

• Assumption - Worldviews are like languages or cultures. People learn languages by lots of repeated input. They learn worldviews the same way.
• Potential approaches to persuasion
  • Provide them a huge amount of information about your worldview. Provide them with huge amount of information about people who already believe your worldview. Provide them with opportunities to interact a lot with people of your worldview. Keep repeating your worldview over and over. Commonly seen with brand advertising on TV. Music and rituals are also a common means by which various political movements reaching mass audiences.

More notes

• Definitely 2 (traumatic life events aka incentives) is the strongest, when it is present. Some mix of 1 (lesswrong rationality) and 3 (groupthink) works on people when 2 is not present. Whether 1 or 3 is stronger, depends on the person. 4 seems the weakest, and is only present when neither 2 1 or 3 is strong.
• Side by side with spreading your worldview, it is also common to attack other worldviews and the people who spread them. Attack could include persuasion, social/financial incentives or violence. That is out of the scope of this post.

2026-02-10

Update

Disclaimer

• Quick Note
• More thoughts on this.

Notes I'm unsure about

• It seems like different approaches work on different people, although I'm less sure about that.
• I just realised these four categories of approach are also common in persuading someone to date you, not just persuading someone of your worldview: Dating profiles and dating docs, versus trauma bonding and codependency, versus extreme confidence encouraged by pickup artistry, versus social settings with repeated interaction (school, college, hobby group etc). It would be useful to see if the same approach works on a person both to persuade them of your worldview, and persuade them to date you.
• There might be some correlation between persuasion approaches and your current worldview. For instance if your current worldview values relationships highly (MtG White), then maybe persuasion approaches aimed at groupthink work better. If your current worldview values knowledge highly (MtG Blue), then maybe persuasion approaches that are more rational work better. But I am highly unsure about this, so I'm not including this in the post above.
• I want to also add a point around motivated reasoning and self-interest. Persuade someone that your worldview is useful to believe in, because it helps them achieve their curent goals better. But this usually seems to work only to add a new piece to their worldview, not to change their worldview completely. I'm unsure about adding this as a 5th point. It seems worth differentiating between approaches that work to change the core of their worldview, versus change a small part of their worldview.

2026-01-29

Update

More notes

• It seems like persuasion (no matter which approach), first involves convincing someone that something is important (or atleast important enough to look at), and secondly involves convincing them that it is true (or atleast useful to believe in, most people's sense of what is "true" is often closer to the latter than former).
  • The MtG colour wheel is one way to guess what is important. 2 1 3 4 can all change one's sense of what is important.
  • Paul Graham recommends first solving "importance" then solving everything else, in his article "how to write usefully". (PG also says "importance" often doesn't have to be solved for explicitly. You can often assume there are N users just like you, without having to test this hypothesis.)
  • Hamming question asks people to change their own sense of what is important, without being faced with any incentives to update their sense of what is important. Hamming question is the rationalist method of convincing someone something is important. (Also, lesswrong seems way better at using hamming question to arrive at important questions, than at arriving at correct answers for them.)
  • Traumatic life events seem especially good for convincing people something is important. But they may not immediately provide a solution (if they do, this solution might get strongly learned), in which case that person may still need to resort to some other approach to find a solution to their problem.
• Once you have used an approach to convince someone something is important, you can use a similar approach (or maybe sometimes even a different one) to convince them it is true.

Subscribe

Enter email or phone number to subscribe. You will receive atmost one update per month

Comment

Enter comment