Russell Standish wrote (in the FOR list): >QTI is a corollory of the MWI, depending also on an assumption >known as the "no cul-de-sac" assumption (Bruno believes he has >proved this assumption, and so would call it a theorem).

At first sight in "the computational realm" (formalised by the G modal logic) there are cul-de-sac everywhere. In the technical part of my thesis, I just show that quantum like probabilities arises when you impose (modally) the no cul-de-sac assumption. In MWI terms, if I am the shroedinger cat, the probability I survive is 1 because I will not take the terminal world into account. Quantum Immortality is at least a special case of comp immortality, and it should be, at best, quantum immortality exactly. (It's an open problem but I provide strong evidence that it is indeed the case, so that the quantum is plausibly the computational "seen from inside"). To sum up, I don't prove: comp => no cul-de-sac, I "only" prove that comp + no cul-de-sac assumption => quantum probabilities, at least in the technical part of my work. It is nevertheless true that I provide also an intuitive argument that comp could be incompatible with first person mortality, but the link with the modal "cul-de-sac" notion is not so obvious. Well, if you take "no-cul-de-sac assumption" = "immortality assumption", then indeed I "prove" that comp implies the absence of cul-de-sac. (That is probably what you were meaning, I realise now). Note that any modal approach to probability introduces (implicitely) a "no cul-de-sac" assumption. The algebra of possible events is always supposed to be non empty. In another (more recent post) Russell wrote: >This sets me apart from the "observer moment" camp, as they don't >believe in time, or in any connections whatsoever between OMs. > >Bruno goes much further than I do in defining observer moments in >terms of machine states, and the set of possible continuations of >computations the future histories of any such observer moment. In my >viewpoint, this provides a concrete model of consciousness (assuming >it is valid), but does not demonstrate its necessity. I agree. Still it seems to be a necessity with the comp hypothesis. >Russell Standish, then Gary Oberbrunner, then Russell again: > > > > > You asked in another email what the "no cul-de-sac assumption" > > > > was. Here it is: at each point in a quantum history, there is always > > > > at least one future in which the observer is alive. > > > > > > This seems equivalent to the "all logically possible universes exist" > > > assumption as opposed to the "all physically possible universes > > > exist". > > > > Its got nothing to do with this. A semantical problem arises through the use of the word "universes". I prefer to say that "all logically possible computations exists". The "physicalness" linking computational states (observer moments) appears as a first (plural) person point of view accessible by (immaterial) machines sharing enough computational histories. And this makes Garry Oberbrunner's reply to Russell meaningfull, it seems to me. And this sets me apart from the ontology of David's FOR, because David presuppose the existence of a physical universe(s), as a mind independent source of empirical truth. I do not postulate the existence of an empiric reality, although I agree we have an innate belief in such an empiry, and I agree it is locally useful. Bruno http://iridia.ulb.ac.be/~marchal