Jesse Mazer wrote:
> Date: Sat, 13 Feb 2010 10:48:28 -0800
> From: [email protected]
> Subject: Re: problem of size '10
> To: [email protected]
>
> --- On Fri, 2/12/10, Bruno Marchal <[email protected]> wrote:
> > Jack Mallah wrote:
> > --- On Thu, 2/11/10, Bruno Marchal <[email protected]>
> > > > MGA is more general (and older).
> > > > The only way to escape the conclusion would be to attribute
consciousness to a movie of a computation
> > >
> > > That's not true. For partial replacement scenarios, where part
of a brain has counterfactuals and the rest doesn't, see my partial
brain paper: http://cogprints.org/6321/
> >
> > It is not a question of true or false, but of presenting a valid
or non valid deduction.
>
> What is false is your statement that "The only way to escape the
conclusion would be to attribute consciousness to a movie of a
computation". So your argument is not valid.
>
> > I don't see anything in your comment or links which prevents the
conclusions of being reached from the assumptions. If you think so,
tell me at which step, and provide a justification.
>
> Bruno, I don't intend to be drawn into a detailed discussion of your
arguments at this time. The key idea though is that a movie could
replace a computer brain. The strongest argument for that is that you
could gradually replace the components of the computer (which have the
standard counterfactual (if-then) functioning) with components that
only play out a pre-recorded script or which behave correctly by
luck. You could then invoke the 'fading qualia' argument (qualia
could plausibly not vanish either suddenly or by gradually fading as
the replacement proceeds) to argue that this makes no difference to
the consciousness. My partial brain paper shows that the 'fading
qualia' argument is invalid.
Hi Jack, to me the idea that counterfactuals would be essential to
defining what counts as an "implementation" has always seemed
counterintuitive for reasons separate from the Olympia or movie-graph
argument. The thought-experiment I'd like to consider is one where
some device is implanted in my brain that passively monitors the
activity of a large group of neurons, and only if it finds them firing
in some precise prespecified sequence does it activate and stimulate
my brain in some way, causing a change in brain activity; otherwise it
remains causally inert (I suppose because of the butterfly effect, the
mere presence of the device would eventually affect my brain activity,
but we can imagine replacing the device with a subroutine in a
deterministic program simulating my brain in a deterministic virtual
environment, with the subroutine only being activated and influencing
the simulation if certain simulated neurons fire in a precise sequence).
It seems that these thought experiments inevitably lead to considering a
digital simulation of the brain in a virtual environment. This is
usually brushed over as an inessential aspect, but I'm coming to the
opinion that it is essential. Once you have encapsulated the whole
thought experiment in a closed virtual environment in a digital computer
you have the paradox of the rock that computes everything. How we know
what is being computed in this virtual environment? Ordinarily the
answer to this is that we wrote the program and so we provide the
interpretation of the calculation *in this world*. But it seems that in
these thought experiments we are implicitly supposing that the
simulation is inherently providing it's own interpretation. Maybe, so;
but I see no reason to have confidence that this inherent interpretation
is either unique or has anything to do with the interpretation we
intended. I suspect that this simulated consciousness is only
consciousness *in our external interpretation*.
Brent
According to the counterfactual definition of implementations, would
the mere presence of this device change my qualia from what they'd be
if it wasn't present, even if the neurons required to activate it
never actually fire in the correct sequence and the device remains
completely inert? That would seem to divorce qualia from behavior in a
pretty significant way...
If you have time, perhaps you could take a look at my post
at http://www.mail-archive.com/[email protected]/msg16244.html
where I discussed a vague idea for how one might define isomorphic
"causal structures" that could be used to address the implementation
problem, in a way that wouldn't depend on counterfactuals at all
(there was some additional discussion in the followup posts on that
thread, linked at the bottom of that mail-archive.com page). The basic
idea was to treat the physical world as a formal axiomatic system, the
axioms being laws of physics and initial conditions, the theorems
being statements about physical events at later points in spacetime;
then "causal structure" could be defined in terms of the patterns of
logical relations between theorems, like "given the axioms along with
theorems A and B, we can derive theorem C". Since all theorems concern
events that actually did happen, counterfactuals would not be
involved, but we could still perhaps avoid the type of problem
Chalmers discussed where a rock can be viewed as implementing any
possible computation. If you do have time to look over the idea and
you see some obvious problems with it, let me know...
Jesse
--
You received this message because you are subscribed to the Google
Groups "Everything List" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/everything-list?hl=en.
--
You received this message because you are subscribed to the Google Groups
"Everything List" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/everything-list?hl=en.
