The most promising category of approach seems to me to be an adversarial one. Level 1 is having one side present their case—basically the status quo. Level 2

The most promising category of approach seems to me to be an adversarial one. Level 1 is having one side present their case—basically the status quo. Level 2 is having some group of people with a strong incentive to find flaws and what's missing, and with a platform on which to present their criticisms. Level 3 is to have people (possibly from the first group, possibly other members of the second group) critique the criticisms, and possibly iterate that. Level 4 might include live debates, although that has pros and cons relative to written arguments. A tactic that I suspect would be useful is the ability to call out specific things—"Yeah, a bunch of this stuff is arguable and murky, but that thing you said right there is wrong and I can prove it; I will stake a chunk of my reputation on this, and we will litigate the issue, and if I'm successful, it should damage your reputation and boost mine"—combined with people expecting to be in the system for long enough that reputations are worthwhile. ("Wrong" will include "factually wrong" and might also include "against the rules of this system"—it might be sensible to have rules about claims needing to be backed up by certain kinds of evidence.) I think at least some parts of academia were supposed to work something like this, and maybe a few parts of it actually do.