In his engaging and important paper David Chalmers argues that perhaps the

pastecoolAI and Robotics

Nov 14, 2013 (3 years and 10 months ago)

161 views

Zombies and Simulation

Richard Brown




In his engaging and important paper David Chalmers argues that perhaps the
best way to navigate the singularity is for us to integrate with the AI++ agents. One
way we might be able to do that is via uploading, which is a process
in

which we
create a
n exac
t

digital duplicate of our brain. He argues that consciousness is an
organizational invariant, which means that a simulation of that property would
count as the real thing (a simulation of a computer is a computer, and so being a
computer is an organizatio
nal invariant).
If this is the case then we can rest assured
that we will retain our consciousness inside such a simulation.
In this commentary I
will explore th
ese ideas and their relation to
philosophical
zombies. I will argue that
dualism could be true
of the zombie world and that
the conclusion of the standard
zombie argument needs to be modified to deal with simulation.
In short I argue that
if one endorses biologism about consciousness then the conceivability of

zombies is
irrelevant to the
physicalis
m/dualism debate.



Interestingly
Chalmers

has pointed out that thinking about simulations gives
us one concrete way of visualizing Cartesian dualism

(
Chalmers 20
0
5
)
. If we think of
a person inside a perfect simulation but whose body is outside
t
he simula
tion (like
in the movie the Matrix) then all of the things that dualists like Descartes thought
were true would have been true. From the point of view of the person inside the
simulation no amount of physics
(i.e. the physics true in the simulation)
would
allow
them to deduce the existence of the body outside of the simulation. This would make
the body outside the simulation nonphysical from the point of view of those in the
simulation

(where to be non
-
physical just is being non
-
deducable from a completed
p
hysics)
.

Chalmers has elsewhere talked about zombies and the problems their
conceivability entails for physicalism

(Chalmers 2009)
. Zombies are creatures that
have the exact same microphysical structure as I do but who la
ck consciousness all
together. Per
haps s
urprisingly we can use the above reasoning to show that
Cartesian dualism could be true for zombies as well! We can imagine our zombie
twins also being very motivated to create zombie AI and may even do so by evolving
these agents in virtual worlds.
We can imagine the zombies creating zombie AI++
and wondering (in their zombie way) what the best course of action would be in
such a case.
We can even imagine our zombie twins wanting to be uploaded into
their simulated worlds. Zombies in perfect simulati
ons of their environments would
think that they were conscious, as per the above arguments. They would have all the
same beliefs etc as you would in the exact same simulated environment.
The crucial
point is that zombies would live the very same lives whet
her in the simulation or
not. In either case they (allegedly) lack consciousness, yet in one case Cartesian
dualism is true and in the other case it is false.
But if dualism can be true at the
zombie world then we seem to have missed what is essential to C
artesian dualism.
This makes is seem as though the debate between the Cartesian dualist and the non
-
dualist is really a debate about whether we in fact live in a simulated world or not

Now what about us when we actually upload into a perfect simulation?
Then
it seems as if Cartesian dualism is true of us as well.
Let us
explore this a bit.

Suppose that, as some theorists have suggested, consciousness depends on some
biological property of the brain.
Let us call this kind of view ‘biologism’ about
consciou
sness.
There are many reasons to suppose that this is true. But of course,
there are some disagreements out there. Let us put these disag
reements aside for
the moment a
nd just stipulate
, for purposes of discussion,

that consciousness
depends on some biolog
ical property of the brain.
According to this view non
-
biological organisms cannot be conscious. It is important to note that this view is
neutral as between physicalism and dualism. A
physicalist

who endorses biologism
will think that consciousness just i
s something biological, as Ned Block tends to lean
towards. A dualist who accepts biologism will hold that only the biological
properties of the brain are correlated with non
-
physical conscious properties

(perhaps Huxley was in this camp)
.

Now let us supp
ose that we have reached
a period of time were we can
accomplish
uploading. When we upload we somehow create a functional duplicate
of the original brain. This could be done in either of the ways that Chalmers
canvasses

in his paper. For instance it could
be done by serially slicing the brain

and
scanning

it
thereby
creating a virtual duplicate of it
(destructive uploading)
or

it

could be done using nanotechnology replacing neuron by neuron. Either way at the
end of this process we have a virtual duplicates

of a human being. Now, according to
our previous stipulations this functional duplicate will not be conscious. This is
because it will not have the relevant biological property. But we can stipulate that
this functional duplicate, in its virtual world, wi
ll produce all the same behavior that
i
ts

non
-
duplicated biological twin would produce. We can even take this a step
further and imagine hooking up this functional duplicate of a brain in the virtual
world to a robotic body in our world. We stipulate that
this robotic body does not
have a brain and so also lacks the relevant biological properties postulated to be
necessary and sufficient for consciousness.

So what have we got here? Remember we
have
stipulated that consciousness
depends on some

biological p
roperty in the human brain. Given that stipulation,
what we have here is a zombie. A zombie, by definition, is a creature

physically

identical to me but which lacks consciousness.
To make it the case that we really
have a physical duplicate w
e can take thi
s even further by imagining that
this takes
place entirely inside a perfect simulation of our world.

In that

simulated world
there
is a
simulated
body controlled by a functional duplicate of my brain. From the
point of view of this simulated world there i
s no consciousness. This is simply
because we’ve stipulated that consciousness depends on some biological property o
f
the brain. But given this we

come to an interesting conclusion. We seem to have
conceived of zombies in the way that the original zombie a
rgument requires.

Some may be skeptical that what we have here is really a zombie as
traditionally defined so let’s take a moment to review.
Chalmers (2009) has

argued
that it
was conceivable that we have (P & ~Q)
. Where ‘P’

stands for the completed
micro
physical

theory of our world

and ‘Q’ is some phenomenal truth, like that I see
blue
.
Chalmers

(2005)
also argues elsewhere that this completed
microphysical

theory could be computed. This means that we could devise a virtual world
that

wa
s
an exact functio
nal duplicate of

our world. All of the laws of physics would be the
same from the point of view of this virtual world.
Indeed we cannot know a priori
that the actual world is not a simulation in this sense.
In this virtual world we have
our functional dupl
icates of biological human beings. So
,

in this virtual world we
have P
, the completed microphysics of our world, and
, given that we are assuming
that biologism is true, not Q
.

We could, of course, stipulate that in this world there
are biological organisms outside the simulation, hooked into it in the appropriate
way, and then we would have consciousness in that world. But we are here
assuming that biologism is true about consc
iousness and therefore in this world,
which from its point of view is a microphysical duplicate of ours, we have traditional
zombies.


But if so then the original zombie
argument
is irrelevant to the dispute
between physicalists and dualists. This follows

from the fact that
both the
physicalists and the dualist can endorse biologism about consciousness
. If this is
right then the zombie argument may
really
be a test for whether one thinks we are
living in a simulated world rather than whether physicalism or

dualism is true.
In
fact what is interesting about this line of argument is that it connects very closely to
things that
Chalmers

has held said elsewhere in his work. For instance
he

has
hypothesized that information may be the fundamental unit of reality

(Chalmers
2003; Chalmers 2005
)

and
that perhaps information has a dual aspect. One aspect,
being the physical as we understand it the other aspect, being the properties of
conscious experience as we feel them.
This line of thought fits very nicely with th
e
view of the zombie argument here sketched.

Now of course this entire discussion has been predicated on the idea that
we

first stipulate that
biologism about
consciousness
is true, or that
consciousness
depends

on

some

biological properties of the human
brain.
Chalmers

has argued
that this is most likely incorrect. And here, I am not talking about his zombie
argument. Instead I mean his argument that consciousness is an organizational
invariant

(Chalmers 1995)
.
This is not the pl
a
ce to address this argume
nt but the
prospects for a reply seem good.
After all it is an empirical question whether we will
ever really be able to enact the kind of
sci
-
fi
scenarios that Chalmers appeals to in
making his anti
-
biologism arguments.
And it is a further empirical quest
ion about
whether they will proceed in the way he imagines, but this is an argument for
another day.

At th
e very least we can point out
that if one accepts some form of biologism
about consciousness
, as many do,

then one can be a
physicalist

and grant the
conceivability of zombies
as well

as
that conceivability implies possibility

without
any threat to physicalism
.
To those, like me, who find the prospects for biologism
about consciousness to be good and who want to endorse physicalism as well as the
concei
vability of zombies and the link between conceivability and possibilit
y this is
a promising strategy; a way of having one’s cake and eating it too!








Work cited:

David J. Chalmers (1995).

Absent Qualia, Fading Qualia, Dancing Qualia

. In Thomas
Metzinger (ed.), Conscious Experience. Ferdinand Schoningh.


David J. Chalmers (2003).

Consciousness and its Place in Nature.


In Stephen P.
Stich & Ted A. Warfield (eds.), Blackwell Guide to the Philosophy of Mind. Blackwell.


David J. Chalmers

(2005).

The Matrix as Metaphysics.


In Christopher Grau (ed.),
Philosophers Explore the Matrix
. Oxford University Press.


David J. Chalmers (2009).

The Two
-
Dimensional Argument Against Materialism.


In
Brian P. McLaughlin & Sven Walter (eds.), Oxford Handbook to
the Philosophy of
Mind. Oxford University Press.