Hal Finney wrote:
I would propose (as I layed out in some detail in a post aboutthe purpose of the list, to discuss the implications of the various ideas that "everything exists". Everything we say is implicitly prefaced by the conditional clause, "If all <whatever>s exist, then".
3 months ago) that their are in fact many constraints on those
states-of-affairs that can be said to "exist".
I would put it this way: Only those states and state changes
capable of supporting communities of communicating intelligent
observers "exist". Other, weirder states only "potentially exist",
but a better way of saying it is that they are "partially
qualified to exist, but not fully". They are "partially qualified"
in the sense of being configurations of information, as are the
more self-consistent "existable" states, but they are not self-consistent enough to exist.
This amounts to a definition of "exist" more than anything else.
1. The "consistent enough to exist (and be commonly perceived)"
states must not only be able to support a single observer, but
the whole ecosystem of observers that allows that observer to exist, and the whole physical set-up (planet, gravity, particular
gaseous mixtures, particular energy and temperature regimes etc)
that allows communities of intelligent agents to exist and observe.
Any state changes (at any time) that would deviate from the maintenance of the consistent physical laws that allow for conventional existence of stable emergent systems and intelligent observers will be automatically disqualified
from full observable existence.
2. My contention is that this is an onerous constraint on
"fully existable states", and that the number of possible
configurations of such states is probably very limited.
It wouldn't surprise me if something very close to the known
physical constants and laws turned out to be actually
"logically necessary" for the continual self-consistency
requirement of existable states that I propose. It could be
that ALL other configurations of matter, energy, information
become inconsistent (or too disordered, or too ordered) quickly,
and disqualify themselves from "observer production" and "full-fledged (classical) existence".
3. This is not to say that every action that every person
for example takes is "necessary" for continued full existence
of the classical, existing, "way things actually are" state.
But it means that such personal actions are heavily constrained by the way things had to be in the historical
development of our species, and its body's capabilities and
its brain's capabilities. And so human behaviour, on average, will be as determined
by our nature, and all of that (narrow) range of behaviour
will be within the narrow bounds of "fully existable" states.
4. It seems to me that "self-consistency" and "rule-governed, effectively continuous, localized change" of state are the necessary pre-conditions of sequences of states that can be fully existent (observable).
Bear in mind that these requirements must be met to the full
extent of allowing non-locally consistent existence. i.e.
a whole consistent universe (observable and agreeable by all observers in it) must be possible with those
state-change rules, not just one person's consistent life-story.
That's a very heavy constraint on state-changes. ALL of those
state changes must conserve the non-locally-consistent full
universe life-story without discrepency.
5. My intuition says that these very heavy consistency and continuity requirements (on "fully existable state change sequences")
would probably rule out travel or communication by observers between
different possible worlds.
And would probably rule out there being a different calculus
of consequence and probability in a MWI compared to an SWI.
Only each "self-consistent world" can be a "possible world".
Most sets S where S is a "set of alternative possible worlds"
will not be themselves (as a set S) able to be self-consistent enough to be a fully existent "world". Or another way of putting it is
that if a set S of alternative possible worlds is itself
self-consistent (over spatiotemporal evolution of its states) then
it collapses by definition into being a single world, not
a set of different worlds.