X-Message-Number: 9303
Date: Tue, 17 Mar 1998 16:43:52 +0300 (MSK)
From: Eugene Leitl <>
Subject: Turing, shmuring

I think there is really no hope, but I try this just once again. This list
is supposed to be about cryonics, after all. Before you ask: below is not
meant to confusticate you, but to supply some of the context I use for my
reasoning.

On  Mon, 16 Mar 1998 10:39:31 EST Ettinger <> wrote:

> Would a Turing tape emulation constitute a person and his environment
> and activities?

A Turing machine is a mathematical abstraction. It does not include
physical limitations, as finite length of the tape and nonfinite duration
of each operations. Uploading is supposed to operate in the physical
realm, and thus to be subject to the constraints of the material realm. 

<ILLUSTRATION> I made a molecular dynamics crunch over the last weekend,
encompassing 1 ns worth of a protein oligo in vacuo. If I was to simulate
1 s on the same system I'd need over 5 MYrs. For all practical purposes,
on that particular system, the later simulation is currently impossible.
Not in theory, but in practice. </ILLUSTRATION>

> Mike Perry (#9287) says yes, period (as best he can currently guess).
> Good for him. He doesn't try to squirm out by saying the tape is too
> slow and impractical; and he doesn't drag in the red herring that other

You are invoking common sense as argumentation criterium. Using
abstractions where common sense cannot go does not seem to be a
constructive type of argumentation. I got burned very badly recently when
_intuitively_ estimating the impacts of a space-based nuke upon a slab of
steel 300 m away. 

Ok, a Turing machine can simulate any physical system. Better now?

> things previously thought "absurd" have proven true; and he doesn't

Let's see, you started using the apparent-absurdity 'argument'. I've just
shown that this is not neccessarily true, for many things once declared
absurd are now obviously considered 'true'. 

> indulge in irrelevancies such as you-might-be-in-an-emulation-right-now. 

You can't prove me wrong that you are a simulation. This has been supposed
to be an illustration, that there are two very distinct points of view:
the external one (=God's view, external world) and internal one
(=simulated system). A simulated tornado won't disturb anything outside of
the box. Any external events outside of the box are not perceivable inside
of the box unless made explicitly accessable by sensors and appropriate
encoding. The external and the simulated time are wholly unrelated -- you
can only say that the simulated time tick must consist from at least one
external time tick.  (And yes, it has been proven that for real physical
systems there is an upper bound for the quantum states and the rate of
state change -- which is equivalent the fact that real physical systems
are discrete both in the spatial and the temporal domain. Then there is
psychophysical time quantisation, which is a high-level artefact of the
information processing system). If the simulation is perfectly
self-contained, there is provably no way to obtain knowledge about the
simulating system.

I also said that you cannot perceive the 'true reality', you always are
limited to work at model level. Parts of the physical system 'Ettinger'
are doing nothing than collecting a data stream coming from the 'true
reality'. Already here the 'reality' is skewed, as you perceive the
'reality' through an infinitely chosy, infinitely artefacting filter
(well, at least I can't see neutrinos nor feel electromagnetic field
gradients very well).  Further downstream things grow even worse, as this
stream is analyzed and recoded in extremely strange fashion. Take a look
at your eye: the information streaming through your visual nerve is e.g. 
compressed 1:126 in relation to the bitmap the retina sees, and 'looks'
nothing like your original Mona Lisa painting. In fact retina is formally
a part of the brain, and it trivially falsifies the homunculus argument. 
The retina is part of 'Ettinger', albeit structurally/functionally very
similiar in any average nonblind human. There is no self in the pieces if
we apply the zen knife iteratively.

> The Turing tape emulation could work in principle, and he stands by his

This would seem a rather weak argument, as a mathematical abstraction is
quite useless in reality. If you deliberately limit yourself to argument
at such level, ok. As I already said, I agreed that a Turing machine can
simulate anything. Much good this will do you.

> conclusion. He might even be right; as I have reiterated many times, I
> don't claim to have proven the contrary--only that there are strong
> reasons for doubt.
 
You have not stated your reasons for doubt formally. Obviously, I cannot
start argumenting until the topic has been defined clearly. 

> Joe Strout (#9289), responding to my question as to whether a
> description of a picture (in words) would BE a picture, says yes--in the
> sense that it COULD be transformed into a conventional picture, just as
> a scanned photo, stored as bits in a computer, can be restored from the
> bits into a picture. That isn't news, and doesn't really answer the

This process is fundamentally not different from what your visual system
does.

> question--but then, I didn't really frame the question very well. It was

Indeed.

> mainly intended as a lead-in to the questions about emulation of a
> person and his environment and activities. In what sense, and to what
> extent, is a STORY the "same" as an actual happening? 

In what sense, and to what extent is the spatiotemporal neuron firing
pattern class labeled 'Ettinger' the "same" as an actual happening? 
Calling it 'STORY' all the time stops being that original after a while. 

> Let's try yet again, with something much simpler, and get away from the
> complication of consciousness and worries about accuracy etc.

Oh yes, if you subtract enough properties from a parrot at some point the
poor thing stops being a parrot. ('Parrotness' is obviously not an
intrinsic property of the parrot but of the observer). 

> Using appropriate quantum terminology and spacetime coordinates, I
> describe a hydrogen atom, far out in intergalactic space, in its lowest
> energy state;  then I describe a photon coming along and exciting the
> atom. This might be done on a Turing tape; or I might just use words

This might indeed be done with a Turing machine.

> spoken aloud. Have I, in some sense, created a hydrogen atom? When I

I am on wobbly ground here, but afair the time-independant Schroedinger
equation for hydrogen is analytically solvable. Of course if you want to
adress the photonic excitation dynamics fully, you will find yourself in
one very deep, very hot hell's kitchen. Whenever you want to invoke
anything even more true, like QCD or M-theory, this will result in lots of
very ugly computations -- for the topic is analytically unsolvable and
must be solved by numerical approximations. It will be a _very_ long, very
ugly story -- the longer the truer and it certainly won't rhyme nor make
sense to anybody, unless that anybody is structured very much like you
(yes, this is highly relevant -- don't you try telling me stories in
Mandarin). Depending on the level of accuracy you want, the story may
indeed become arbitrarily long. Of course 'Ettinger' (nor any entity which
goes on two legs) would not be able to tell nor understand this story, but
we are only gedankenexperimenting here, after all. I am indeed very glad
that (hopefully) uploading does not require telling stories as rich and as
involved as those you have chosen for your example. 

> stop talking, or the tape stops moving, does the atom go back to
> oblivion?

Before I address the 'creation' and 'going back to oblivion' argument,
please define what exactly you mean by creation and destruction. (Ok, I
will help you a bit: if there was a second hydrogen atom in the simulation
box -- sorry, I meant STORY -- it would certainly 'notice' (=behave
differently) if its partner was suddenly to vanish. And if you were
interested in dynamics, holding, or even deleting the state would not seem
to be smart. Of course it would be even worse if there was no you, nor
an audience -- does the falling tree really make a sound?). 

If we are talking about holding the simulation, then the frozen frame
(=full state) alone will not make much sense to anybody interested in the
temporal aspect of the problem. Is a neuropatient 'Ettinger' really
'Ettinger'? Obviously, in the trivial definition scope it would only be a
potential 'Ettinger', as well as an embryo is not a particular person. 

> Unless an info person claims that my story does indeed constitute the
> PHYSICAL CREATION of a hydrogen atom (even though one of a different
> sort, in a different "universe"), he must recant on the person emulation
> thesis. 
> 
> I don't see any unequivocal answers to such questions--even if we ignore
> the fact that our understanding of physics is incomplete and possibly
> partly incorrect, and hence any story we tell is bound also to be
> incomplete/incorrect and therefore possibly not a "good enough"
> emulation.

I think your problem is a) you think that games with language and
sophistry expose, not confusticate the object you reason about b) you
understanding of self and reality is not well defined. Many of apparent
paradoxa plagueing you are resolved naturally in the proper model context
(of course you don't think that context to be valid -- hence I once again
propose we postpone the argument until a sufficient proof -- say, a
realistic working model (=upload) of a nematode or a sea slug is
presented). 

> Incidentally, the questions about TIME are much deeper and more puzzling
> than acknowledged in these exchanges. One respectable view is that our
> sense of time is an illusion, and spacetime is a "block" whose slices
> are moments, all coexisting and "pre"determined. 

Which time do you mean? The physical time, the psychophysical time, the
simulation time or the simulator time?

ciao,
'gene

Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=9303