Dear Alastair: I will read your paper, but it seems to me that the "no information" approach to formulating an Everything precludes selection. Selection assigns a property to a subset of the ensemble that the other members do not share. This destroys the ensemble.

## Advertising

Prevalence being a property I would conclude from this that no allowed type of universe can be more prevalent than any other allowed type. Speculation: Is it possible to use this approach to exclude certain types of universes? First distinguish two types of evolving universes [universes that change state]: 1) Those that have rules of state succession that forbid a source of external information [true noise] 2) Those that have rules of state succession that allow a source of true noise [external information] to some degree [from zero true noise to nothing but true noise]. This distinction is already a selection so one or the other type must be absent from an informationless Everything. I think the first step to resolution is to notice that type #2 could include type #1 as an extreme lower limit case, but type #1 can not include type #2 at all. However, is the extreme lower limit case for #2 allowed to be zero true noise? Perhaps to resolve this ask whether either of these types is a larger set. Since type #2 is a continuum is also type #1? If an infinite rule set and/or data string is equivalent to number #2 then universes in #1 must have finite rule sets and finite data strings [finitely describable]. The number of such universes would be merely countable. At any non zero degree of true noise - say 2% or 20% etc. there would be an infinite number of ways to allow that percentage. The conclusion would be that universes with zero true noise would be extremely rare relative to any of those with a non zero degree of true noise and so such an information generating lower limit is excluded. The extreme lower limit for true noise in #2 must be greater than zero. However, this itself seems like a breach of the "no information" approach. This could be fixed if type #1 were reclassified as non evolving universes and a door is opened between the two types. A type #2 universe with the right dose of true noise can surely convert to a type #1. This would balance the quantity issue. Can type #1 universes become type #2 universes. They must be able to or again there would be a selection. It seems this is just a way of saying that universes do not have fixed rules of state succession. Hal At 2/17/02, you wrote: >Thanks for your comments - the 'difficulty' that you refer to is none other >than the White Rabbit problem, on which there has been much discussion in >this forum. My own approach addresses the problem from the standpoint of all >logical possibilities (rather than any particular model or set of rules), >which can make the problem non-trivial to solve. > >Alastair