Meta-tations on Moderation: Towards Public Archipelago suggested that we should be running more experiments with how discussion is run and listed several examples in the post. What if we ran monthly experimental open threads? I can imagine several formats that would work better as open threads than as a thread belonging to a specific commentator.

Here's a few thoughts of possible experiments:

  • Double crux - Top level comment would be claims. People who disagree can comment underneath to double crux with the poster. Third parties are discouraged from adding comments which inject their opinion into the double crux
  • Change my view - Top level comments would be claims that the author must believe and be willing to defend. Second level comments must disagree with at least one point of the claim or ask for clarification
  • Socratic method - Top level comments would be claims. Second level comments would be discouraged from directly saying that someone is wrong and instead encouraged to ask them questions instead to get them to think
  • Explain your perspective - Very similar to Socratic method, but the Socratic method is still focused on persuasion. This would also involve asking questions, but the commentators would be encouraged to try to understand people's views instead of trying to share them
  • Dependency sharing - Top level comments list a view and give examples of what might convince you to
  • Ask me anything - Top level comments lists an experience someone has had or specific expertise. Lower level comments ask them questions

A key part of rationality that has been somewhat underemphasised on Less Wrong is group rationality. Achieving big goals requires groups of people and these people need to be able to co-ordinate and efficiently share their knowledge. We could sit around and try to theorise about how to make this work or we could simply start running experiments.

A few questions:

  • Would these be suitable for frontpage posts? I don't see this experiment working with blog posts, as they would be unlikely to hit critical mass. (Sidenote: I don't think posting these threads should net me karma, if they take off it may be worthwhile to have a way to mark threads as open at some point)
  • Any suggestions of useful experiments to run?
  • Lastly, if I ran some of these experiments, would anyone commit to participating significantly in at least the first N (default N=3) to help achieve this critical mass? By participating significantly I mean making a top level comment, addressing replies to this comment and participating in at least one other sub-thread.
New Comment
5 comments, sorted by Click to highlight new comments since:
[-]gjm50

Seems like a good idea. I will not give any sort of absolute commitment (first open thread rules say: "Top-level comments must give the poster's name, physical address, bank details and a link to a high-resolution scan of their signature"...) but I would expect to participate actively in such threads unless they're crazy somehow or I completely blank on thinking of something interesting to say.

I am in favor of this, and think that it's reasonable for threads like that to get promoted to the frontpage (since in this case the moderation would be done by the site-wide moderators and not the post-owner). Though I still need to think more about how we should generally handle open-threads and stuff like that.

[-]Elo30

Would participate. hesitant to lock myself in but I'd be interested.

I like this! And thanks for the cool list of examples.

I am very much interested in this idea. I was on the fence about how interesting/useful I would find open threads, but having a particular structure is something that I think would help get people involved and make it interesting.

I'll commit to being involved in the first 3 "themed open threads" if you decided to do them.

The double-crux and socratic themes seem the most interesting. For the Socratic, I like the idea of the questioners not being in pursuit of making a point, but instead trying to get the questioner to apply their own intelligence more to the topic.