On 3/26/2015 11:05 PM, Bruce Kellett wrote:
meekerdb wrote:
On 3/26/2015 7:16 PM, LizR wrote:
On the subject of counterfactual correctness, isn't that the point of Olimpia and
Klara? My problem with counterfactual correctness is (probably the same as Maudlin's?)
-- how does the system /know/ it's counterfactually correct if it doesn't actually
pass through any of the "what-if" states? To put it another way, when you have a
recording of the conscious computational states being replayed, what difference could
be made by the presence (or absence) of all the extra bits that /would/ deal with
counterfactual correctness if a different computation was being replayed, but happen
in this case not to be used? I can't see how this could make any physical difference
to the states being replayed (unless counterfactual correctness introduces some
nonphysical magic into the system?)
I see two possible answers. First, in a quantum world there is a superposition of all
those "counterfactual" states, so they are really present, but only observable as
different relative states. Of course this already invokes QM and physics, rather than
deriving them. But maybe it can be shown that the infinite threads of the UD serve to
test all the counterfactual states.
I don't think this is correct. QM plus MWI acting on my conscious state here and now
does not generate counterfactual states (if inputs had been different, I would have
reacted in some other way.) This is because MWI on my state does not generate the set of
all possible different inputs. We do assume determinism at some point!
Suppose we just require counterfactual correctness for the range of inputs you think are
possible? I don't think it's so easy to limit the range because "my conscious state"
isn't just "here and now". I Bruno's UD or in the MWI of the universe that conscious
state will have many instantiations is different subspaces and many different inputs that
don't occur "here and now".
Or, secondly, although there is no physical difference in the sequence of states in the
replaying, consciousness is not physical and so could be absent. This doesn't require
that consciousness be magic. If it is the abstract thing called "computation" then in
the abstract it needs to counterfactually correct to count as computation.
I don't think even this follows. A computation is a computation -- it gives a definite
result for definite inputs. It still counts as a computation even if the same program
running on different inputs would give different results.
? Giving different results on different input is what counterfactual correctness implies.
A recording is not counterfactually correct because it gives the same output no matter
what the input (effectively there is no input). If you don't require counterfactual
correctness, i.e. computing the correct answer for different inputs, then a look-up table
with just one entry qualifies as a computation.
Brent
I have a lot of trouble seeing that counterfactual correctness is actually the
distinction Bruno needs to make. He wants to distinguish the active simulation from the
passive rerun of the same sequence of states. Why not just make this distinction,
simpliciter?
I am with Liz -- where is the actual contradiction with assuming physical supervenience?
That does not rule out the possibility of supervenience on an effective simulation --
the simulation is run on a physical computer, after all!
Bruce
--
You received this message because you are subscribed to the Google Groups
"Everything List" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/everything-list.
For more options, visit https://groups.google.com/d/optout.
