On 3/2/2020 4:44 PM, Jason Cobb wrote: > On 3/2/20 7:23 PM, Kerim Aydin wrote: >> The below CFJ is 3818. I assign it to Jason. >> >> status: https://faculty.washington.edu/kerim/nomic/cases/#3818 >> >> =============================== CFJ 3818 =============================== >> >> Rance's master switch is set to Agora. >> >> ========================================================================== >> >> Caller: Falsifian >> >> Judge: Jason >> >> ========================================================================== > > > Immediate thoughts: > > FALSE per caller. However, I can also see an argument that Rance's > Master was indeterminate because I performed a conditional action based > on whether there was a zombie auction. At the time, it was not widely > known whether there was a zombie auction; is that enough to make the > value of the switch indeterminate? Or is it enough that the rules and > messages were "reasonably available", which would (theoretically, at > least) allow determining whether the zombie auction existed, to prevent > Rance's Master being indeterminate.
I think "FALSE per caller" is right here. While I can't point to precedents that deal with this explicitly, we've usually allowed for "interpretive" uncertainty in conditionals if the uncertainty is a clear-cut binary decision - e.g. if the situation is "If CFJ X is found TRUE, then Y clearly works, but if it's found FALSE, Y clearly fails", then conditionals based on "If Y..." are generally acceptable while waiting for the CFJ result. One way to look at it: as per R. Lee's judgement on Gaelan's N+1 bid in CFJ 3819, the above sort of conditional is "clearly specified" in terms of the two potential by-announcement actions and the conditions to choose between them, and the conditional is "platonically retroactively determined" once the CFJ is delivered. -G.
