On Thu, 2019-03-07 at 09:11 -0800, Kerim Aydin wrote: > No, the paradox is something like: > > 1. Purported resolution message from Assessor; > 2. CoE: this Proposal was never distributed so wasn't adopted. > > In terms of the truth value of the CoE, if it is true (wasn't adopted) then > it is false (it was distributed, therefore adopted). A CFJ on whether the > Assessor can legally deny the CoE (given No Faking) may meet the win > condition.
The proposal /was/ distributed, though. It's just that its adoption changed the gamestate as though it wasn't. Denying the CoE constitutes a claim that the proposal was distributed. Well, it /was/ distributed! A change to the gamestate as though the distribution were IMPOSSIBLE doesn't change the fact that it actually happened. Think about it this way: suppose a proposal is puportedly distributed in an impossible way, nobody notices, the Assessment remains unCOEd for a week, so the fact that the Agoran decision on whether to adopt it existed ends up self-ratifying. Creating the Agoran decision on whether to adopt a proposal is distribution, so in the resulting gamestate, either the proposal was distributed, or else a decision about whether to adopt it existed without ever having been created. If we take the former condition, we now have a gamestate in which the proposal was distributed, despite it having been IMPOSSIBLE to do so. I don't see any real issue here; "event X couldn't possibly have happened, but happened anyway" is fairly minor as contradictions go (especially as the rules care about what happened, not what could have happened). After the proposal retroactively makes its own distribution impossible, we end up with exactly the same sort of gamestate; there's no way that the proposal could have been distributed, but it's nonetheless a historical fact that it was distributed. The adoption of the proposal destroys the decision about whether to adopt it (because the decision existed beforehand, and the proposal attempted to change the gamestate to a state where it wouldn't have), but it's too late; the decision's already been resolved by that point. Meanwhile, the gamestate recalculation isn't recursive; the proposal doesn't attempt to undo its own adoption process, because we're recalculating based on the gamestate immediately before the adoption of the proposal, and then applying all the changes atomically in a single batch. -- ais523