> It doesn't take 20 years to evaluate an argument. Yes it can, and even in barely contrived scenarios if your proof calculus isn't nice enough. More importan

> It doesn't take 20 years to evaluate an argument. Yes it can, and even in barely contrived scenarios if your proof calculus isn't nice enough. More importantly, in very practical cases even not assuming adversarial (e.g. Microsoft on standards committees), it most definitely can take 20 years to find a valid argument. Exponential growth is serious stuff. Triple exponential growth is just stupid fast. If you can't wait 20 years, then you have to move forward with arguments for not-valid-but-sometimes-true cases. And in this case, the agent with the most compute power can generate contradictions for opposing props more quickly and can also generate true instances matching his policy preferences more quickly. Thus, bounded argumentation often inherently favors the more resource-rich adversary. And the same thing works in more informal settings as well. One way to cope with this is to disregard arguments from known bad actors. We can discuss rather that's the correct course of action, but the need is a provable fact of the most useful formal systems and an observable fact of informal argumentation (I'm still not sure which you want to discuss, since your assumptions make me think you view all arguments as formal systems, but most arguments -- including this one -- clearly aren't in any meaningful sense of the word formalized).