Many-Worlds Critique

For the bus ride home, I'm going to check out "One world versus many: the inadequacy of Everettian accounts of evolution, probability, and scientific confirmation" by Adrian Kent (arXiv/0905.0624) Nothing like ending the day with some against many-worlds reading. That and a fun TED talk should make the ride go by fast

More like this

I've seen Adrian talk about this stuff and I think it will be a good paper. For the proper context, it needs to be read in conjunction with David Wallace's papers on probability in many-worlds.

One thing that struck me about most every approach to no-collapse interpretations, including this one, is that they treat the quasi-classical worlds as non-interacting. Clearly, unitarity requires that this is just a convenient approximation.

Robin Hanson actually tries to get somewhere with using these interactions to get Born law-like behavior. Successfully? Well, I need to see the code he uses for his simulations, but it at least sounds promising. The basic idea is that the disturbance of one quasi-classical world upon another nearby is weighted _naturally_, _in the dynamics_ by the Born weights. Small worlds then get "destroyed" (nonsensical evolution, hence no observers that are human-like) by nearby larger worlds. But only "typical" worlds (i.e. ones having measurement records one would expect from Copenhagen) end up having non-negligible Born weights.