I was answering a bunch of questions from OpenPhill’s calibration test of the form “when did <thing> happen?”. A lot of the time, I had no knowledge of <thing>, so I gave a fairly large confidence interval as a “maximum ignorance” type prediction (1900-2015, for example).
However, the fact that I have no knowledge of <thing> is actually moderate evidence that it happened “before my time”.
Example: “when did <person> die?” If I was alive when <person> died, there’s a higher chance of me hearing about their death. Thus not having heard of <person> is evidence that they died some time ago.
Example: “when did <person> die?” If I was alive when <person> died, there’s a higher chance of me hearing about their death. Thus not having heard of <person> is evidence that they died some time ago.
Technically, it could also be evidence that you are dead, but your ghost cannot move to afterlife, probably because it is too attached to scoring internet points (a fate that awaits many of us, I am afraid).
Ignorance as evidence
I was answering a bunch of questions from OpenPhill’s calibration test of the form “when did <thing> happen?”. A lot of the time, I had no knowledge of <thing>, so I gave a fairly large confidence interval as a “maximum ignorance” type prediction (1900-2015, for example).
However, the fact that I have no knowledge of <thing> is actually moderate evidence that it happened “before my time”.
Example: “when did <person> die?” If I was alive when <person> died, there’s a higher chance of me hearing about their death. Thus not having heard of <person> is evidence that they died some time ago.
Technically, it could also be evidence that you are dead, but your ghost cannot move to afterlife, probably because it is too attached to scoring internet points (a fate that awaits many of us, I am afraid).
(epistemic status: just kidding)