On 18/09/2015 22:03, Warren D Smith wrote:
The doomsday argument, as Andy Latto hinted, seems equally valid for humans, and "intelligent life originating on earth", and "intelligent life restricted to earth" and "all descendants of a human" and...
So?
At http://arxiv.org/abs/0806.3538 one can find an analysis of the doomsday argument that suggests (I think convincingly) the following picture: - Obviously your prediction of the future lifespan of a thing should depend on whatever information you have on it, which may go beyond the bare fact that it's existed for N years so far. - So the right way (if there is one) to think about the "doomsday" argument is as telling you what pdf to put on future lifespan given lifespan-so-far and "minimal" other information, whatever exactly that means. - The particular sense of "minimal other information" that makes the DA probabilities correct corresponds to a particular kind of ignorance: one in which you have no information suggesting that any particular timescale is relevant. - More precisely, this corresponds to a prior density on total lifetime T proportional to 1/T^2 (this is improper, but no matter). - This makes the joint pdf of (past lifetime, future lifetime), conditional on the thing being neither wholly past nor wholly future, invariant under rescaling of time. In the real world, we are never perfectly ignorant and the "doomsday posterior", urged on us by the doomsday argument once we discover the past lifetime, won't be the best we can do. E.g., "humanity" versus "humanity plus all descendants": one thing we should factor into our estimates, ignored by the doomsday argument, is our knowledge of how long other species have lasted or, in the latter case, how long other species *plus their descendants* have lasted. The latter is often longer; no surprise if we come to the same conclusion about humanity. -- g