Variations in the MWI multiverse

[Lee Corbin]

Consider the set, across the multiverse according to MWI, of solar systems
essentially similar to ours. This set is much vaster (or, perhaps, we should
say, has greater measure or frequency) than the GIU (Group of Identical
Universes) exactly identical to ours. Our own GIU in which we hear
disturbing reports from Ukraine, and in which we perform or can perform
quantum mechanical experiments with specified outcomes, and in which I am
writing and you are later reading this email, is a tiny fraction of this
larger set. This larger set includes solar systems, I'm positing, that have
eight planets also, but which differ in details concerning the planets.

This entire group of 8-planet solar system parallel universes will include a
subset in which Neptune is about half the size of Uranus. Do we have any a
priori reason to suspect that the fraction of solar system parallel
universes containing this half-size Neptune is any less probable than our
own? Eventually, of course, something may indeed be proved regarding Bode's
law that could inflict such a reduction on those; on the other hand,
further progress in celestial mechanics might easily show that half-size
Neptune solar systems are actually more likely to form than our "full size"
(greater than 1/2 Uranus radius Neptune) solar systems.

Within each such collection of half-size Neptune solar systems, what
possible configurations for the Earth are thereby implied, or restricted?
None, so far as I have been able to see. Everything that we know about the
Earth--shall we exclude, say, its approximate mass and distance from the
sun, as part of my initial specification--might just as well have occurred
(did occur, is occurring!) in that collection. For, so vast is the MWI
multiverse, it's difficult to suggest that configurations in which, say, the
continents of North America and South America are quite separate, are any
less frequent than our own, in which the two continents are linked by an
isthmus.

(It may amuse the reader that until I added that last specification
regarding the linkage of the two continents, this very email was being
written most likely just as often on Earths in which there is a sizeable sea
passage between them--but that single qualifier significantly reduced the
number of people (or I should say their measure) reading this, but obviously
didn't exclude you.)

Yes, we do know that a *change* in Neptune's orbit, one, say that occurred
as a result of a large asteroid collision millions of years ago, will be
chaotically linked by classical physics to what is happening now on Earth.
But such considerations seem to me to underrate the vastness of the
multiverse as specified so far in this message. Because, for each variation
in Neptune's orbit thus described, there will still be an uncountable set of
them in which this letter continues in its same form, and what you had for
breakfast remains the same.

Before pressing on with the subject, I would like to know if anyone objects
to any of the above reasoning.

thanks,
Lee Corbin

[Gary Oberbrunner]

Post by Lee Corbin
Consider the set, across the multiverse according to MWI, of solar systems
essentially similar to ours. This set is much vaster (or, perhaps, we should
say, has greater measure or frequency) than the GIU (Group of Identical
Universes) exactly identical to ours.

Of course. One is a strict subset of the other. And the multiverse is
very, very, very big. So far I don't think you're saying anything
controversial.

Any _given_ similar-but-not-identical configuration may be vastly less
probable than ours (say, all the same as ours but that it includes a giant
space duck) but there are so many variations -- even trivial ones like
slight differences in decay times of a particular atom -- that they must
quickly overwhelm the identical ones.

You can really make the same argument for cup of coffee I think.

[Lee Corbin]

Gary writes

vastly less probable than ours (say, all the same as ours but that it
includes a giant space duck)

Instead of the space duck, I prefer the well-known "elephant in the room"
that everyone is always missing. I'm comforted by the thought that at least
in some worlds, according to MWI, people aren't missing it at all, and
indeed can't help but notice it.

...the multiverse is very, very, very big. So far I don't think you're
saying anything controversial.

Thanks for the reality check.
And while I may be guilty of using a new acronym, "GIU", people should
remember that the phrase "group of identical universes" occurs over and over
in the book "Fabric of Reality".

It's hard to break the habit many people so often exhibit, of believing that
a quantum experiment must be performed in order to cause a branching to
occur:

Proceeding, following the example of there being approximately as many
universes with large Neptunes as small Neptunes, and that this holds true
for each particular Earth, it also follows that when you flip a coin, there
are approximately as many otherwise identical Earths in which heads comes up
as tails. This is so despite the trajectory of any coin leaving your hand
being classical, for the tiny differences in air currents and tiny initial
differences in angle or orientation of the coin, the rest of the Earth is
"orthogonal", that is, in the quantum multiverse there are approximately as
many otherwise identical Earths that have a room with this kind or that kind
of air currents, orientations, etc.

The upshot of all this is that when you flip a classical coin, you still
find yourself in each kind of universe, the ones showing heads and the ones
showing tails, exactly as if a quantum "split" had occurred.

Application: suppose that we place a six-sided die in a wooden box, close
the lid, and shake the box vigorously. We then allow the die to rest quietly
in the closed box for several seconds on a table top, which is enough time
for any (unwarranted) concerns about decoherence to be allayed. Then the set
of absolutely identical yous watching the box will be split into six
approximately equal portions when you open the box and look to see how the
die has landed.

Conclusion: when a number of possibilities seem present, as in the case of
the box containing a die in an unknown state, and you place a bet that the
die has taken on a particular value, "you can't beat the odds!". Which is
to say that even after looking at the die, you should acknowledge that you
are seeing all six values approximately equally, in six kinds of parallel
worlds.

Lee


On Sun, Feb 23, 2014 at 1:08 AM, Lee Corbin <leecorbin337-***@public.gmane.org> wrote:

Consider the set, across the multiverse according to MWI, of solar systems
essentially similar to ours. This set is much vaster (or, perhaps, we should
say, has greater measure or frequency) than the GIU (Group of Identical
Universes) exactly identical to ours.

[Gary Oberbrunner]

Post by Lee Corbin
It's hard to break the habit many people so often exhibit, of believing that
a quantum experiment must be performed in order to cause a branching to

I'd be surprised if any MWI physics researcher would make this error. But
I've been surprised before. :-) _Everything_ is quantum.

Proceeding, following the example of there being approximately as many

Post by Lee Corbin
universes with large Neptunes as small Neptunes, and that this holds true
for each particular Earth, it also follows that when you flip a coin, there
are approximately as many otherwise identical Earths in which heads comes up
as tails. This is so despite the trajectory of any coin leaving your hand
being classical, for the tiny differences in air currents and tiny initial
differences in angle or orientation of the coin, the rest of the Earth is
"orthogonal", that is, in the quantum multiverse there are approximately as
many otherwise identical Earths that have a room with this kind or that kind
of air currents, orientations, etc.
The upshot of all this is that when you flip a classical coin, you still
find yourself in each kind of universe, the ones showing heads and the ones
showing tails, exactly as if a quantum "split" had occurred.

Of course. This is straightforward MWI. Every photon emission causes a
split (I prefer differentiation but same point). It's happening trillions
of times (maybe infinitely many, the jury's still out I believe) in each
second in each cc of space.

In fact by the time the coin lands, a huge number of differentiations have
happened; in about half of them the coin is heads and about half it's
tails. But note that within each group, they are _not_ identical; just
macroscopically identical. Individual photons have been reflected off the
surface of the coin in a huge cloud of possibilities (=differentiations),
atomic states have decayed differently, and so on.

[Bruno Marchal]

Post by Lee Corbin
Conclusion: when a number of possibilities seem present, as in the case of
the box containing a die in an unknown state, and you place a bet that the
die has taken on a particular value, "you can't beat the odds!".
Which is
to say that even after looking at the die, you should acknowledge that you
are seeing all six values approximately equally, in six kinds of parallel
worlds.

In the third person description. I agree.

Now, unless you assume some "telepathy" between the brains in the
different branches, each resulting person will also have to
acknowledge that she feel to see the dice in one well definite state.
It reduces the quantum type of indeterminacy to the more simple and
general indeterminacy based on self-multiplications (which occur
infinitely "many times" in arithmetic).

Bruno



http://iridia.ulb.ac.be/~marchal/

[Psevdo Nim]

Post by Lee Corbin
Consider the set, across the multiverse according to MWI, of solar systems
essentially similar to ours. This set is much vaster (or, perhaps, we should
say, has greater measure or frequency) than the GIU (Group of Identical
Universes) exactly identical to ours. Our own GIU in which we hear
disturbing reports from Ukraine, and in which we perform or can perform
quantum mechanical experiments with specified outcomes, and in which I am
writing and you are later reading this email, is a tiny fraction of this
larger set. This larger set includes solar systems, I'm positing, that have
eight planets also, but which differ in details concerning the planets.
This entire group of 8-planet solar system parallel universes will include a
subset in which Neptune is about half the size of Uranus. Do we have any a
priori reason to suspect that the fraction of solar system parallel
universes containing this half-size Neptune is any less probable than our
own? Eventually, of course, something may indeed be proved regarding
Bode's
law that could inflict such a reduction on those; on the other hand,
further progress in celestial mechanics might easily show that half-size
Neptune solar systems are actually more likely to form than our "full size"
(greater than 1/2 Uranus radius Neptune) solar systems.
Within each such collection of half-size Neptune solar systems, what
possible configurations for the Earth are thereby implied, or restricted?

That depends somewhat on what you mean by "implied/restricted".

I would guess that no particular Earth configuration is ruled out by this.

However for example the contents of books on that Earth about our planetary
system will tend to be heavily correlated with what the state of affairs
actually is.

So in that sense it very much does "restrict" the possible outcomes.

Post by Lee Corbin
None, so far as I have been able to see. Everything that we know about the
Earth--shall we exclude, say, its approximate mass and distance from the
sun, as part of my initial specification--might just as well have occurred
(did occur, is occurring!) in that collection. For, so vast is the MWI
multiverse, it's difficult to suggest that configurations in which, say, the
continents of North America and South America are quite separate, are any
less frequent than our own, in which the two continents are linked by an
isthmus.
(It may amuse the reader that until I added that last specification
regarding the linkage of the two continents, this very email was being
written most likely just as often on Earths in which there is a sizeable sea
passage between them--but that single qualifier significantly reduced the
number of people (or I should say their measure) reading this, but obviously
didn't exclude you.)
Yes, we do know that a *change* in Neptune's orbit, one, say that occurred
as a result of a large asteroid collision millions of years ago, will be
chaotically linked by classical physics to what is happening now on Earth.
But such considerations seem to me to underrate the vastness of the
multiverse as specified so far in this message. Because, for each variation
in Neptune's orbit thus described, there will still be an uncountable set of
them in which this letter continues in its same form, and what you had for
breakfast remains the same.
Before pressing on with the subject, I would like to know if anyone objects
to any of the above reasoning.
thanks,
Lee Corbin

[Lee Corbin]

Max Tegmark’s recent book “The Mathematical Universe” may or may not have
more basic ideas than his previous papers, and the great May 2003 article in
Scientific American.

I like very much the following breakdown due to Tegmark:

We live in
1. a “level one” infinite universe, containing infinitely many stars and
galaxies
2. a larger “level two” infinite universe, in which our level one universe
is but
one of many, distinguished by different physical constants
3. the “level three” infinite universe consisting of all the MWI branches of
our
level one universe--presumably we inhabit few other level two universes
4. the (doubtful, I think) “level four” universe containing all mathematical
structures, which actually constitute the previous levels

First, though, extremely annoying is his terminological predilection of
calling our visible universe (i.e. our Hubble volume) a “universe”. It has
the unfortunate implication that we don’t live in the same universe as the
good citizens of a planet in the great Andromeda galaxy who are equally
advanced technically, for they can see over a million light years further in
the directions of the constellations Cassiopeia, Andromeda, and Pegasus, to
mention just three. It’s much more convenient to think of our friends in
Andromeda inhabiting the same universe we do, but whose Hubble volume is
slightly different, of course.

But that terminological annoyance aside, readers of FoR shouldn’t have much
to argue with Tegmark about.

Until, that is, we come to his contention that mathematics is all there is,
that we as individuals are SASs (Self Aware Structures), and so on. To
investigate that contention, and to revisit issues explored online here and
elsewhere a decade or more ago, let’s start with computer emulations of
conscious beings.

I don’t think it controversial to readers of FoR that computer programs can
emulate conscious beings; e.g., for all you know now, you may simply be a
running computer program.
Then let’s imagine that we have a program that emulates the torture of a
small kitten for two minutes. That is, the 2-minute run of the program is
“equivalent” to an actual small cat undergoing terrible suffering. Then on
moral grounds, wouldn’t we object to someone using his computer to run that
program? In particular, let’s say that the program executes 10 million
instructions per second, and that, say, it takes 10 seconds to run the
program. (Thus the software kitten lives at 1/5 the rate we do.)

Moreover, it is possible (but time consuming) to single-step through the
program with a debugger. Whatever time it takes, the end result is still
just as morally reprehensible, right?

This simple approach is the quickest way to undermine the mathematical
universe, and In this, ultimately, lies its downfall, I'll argue.

Lee

[smitra-RSh1/+X/PmFmR6Xm/]

Max Tegmark’s recent book “The Mathematical Universeï¿œ=

ￜￜ may or may not have

more basic ideas than his previous papers, and the great May 2003 article=

in

Scientific American.
We live in
1. a “level one” infinite universe, containing infinitely=

many stars and

galaxies
2. a larger “level two” infinite universe, in which our l=

evel one universe

is but
one of many, distinguished by different physical constants
3. the “level three” infinite universe consisting of all =

the MWI branches of

our
level one universe--presumably we inhabit few other level two universe=

s

4. the (doubtful, I think) “level four” universe containi=

ng all mathematical

structures, which actually constitute the previous levels
First, though, extremely annoying is his terminological predilection of
calling our visible universe (i.e. our Hubble volume) a “universe=

”. It has

the unfortunate implication that we don’t live in the same univer=

se as the

good citizens of a planet in the great Andromeda galaxy who are equally
advanced technically, for they can see over a million light years further=

in

the directions of the constellations Cassiopeia, Andromeda, and Pegasus, =

to

mention just three. It’s much more convenient to think of our fri=

ends in

Andromeda inhabiting the same universe we do, but whose Hubble volume is
slightly different, of course.
But that terminological annoyance aside, readers of FoR shouldn’t=

have much

to argue with Tegmark about.
Until, that is, we come to his contention that mathematics is all there i=

s,

that we as individuals are SASs (Self Aware Structures), and so on. To
investigate that contention, and to revisit issues explored online here a=

nd

elsewhere a decade or more ago, let’s start with computer emulati=

ons of

conscious beings.
I don’t think it controversial to readers of FoR that computer pr=

ograms can

emulate conscious beings; e.g., for all you know now, you may simply be a
running computer program.
Then let’s imagine that we have a program that emulates the tortu=

re of a

small kitten for two minutes. That is, the 2-minute run of the program is
“equivalent” to an actual small cat undergoing terrible s=

uffering. Then on

moral grounds, wouldn’t we object to someone using his computer t=

o run that

program? In particular, let’s say that the program executes 10 mi=

llion

instructions per second, and that, say, it takes 10 seconds to run the
program. (Thus the software kitten lives at 1/5 the rate we do.)
Moreover, it is possible (but time consuming) to single-step through the
program with a debugger. Whatever time it takes, the end result is still
just as morally reprehensible, right?
This simple approach is the quickest way to undermine the mathematical
universe, and In this, ultimately, lies its downfall, I'll argue.

The laws of physics put some limits on this thought experiments. The
information used in running the simulation of the kitten has to come
from somewhere. If we take this into account, the Mathematical
Multiverse will become a lot more plausible.

The registers of the computer will be in some state that can be
represented as a bit string. However, you can ask if you can know all
the bits in the bitstring at any moment. In case of a kitten, a large
part of your brain would have to be aware of all the bits, this would
exceed the memory capacity we have reserved for recalling events by a
huge factor.

Then given that we cannot objectively know the content of the bist
string, the bit string itself will typically not have an objective
value (which we then don't know). Quantum mechanics in the MWI
interpretation implies that there exists an ensemble of all possible
bit strings (constrained by what we do know about the virtual cat). The
bit string will be in an entangled state with the environment.

If you consider each member of the ensemble of bit strings to be
located in a parallel MWI universe, then this amounts to saying that we
are never located precisely in such a universe, as these are too small
(contain too much information) for us to fit in. Then one can ask if
the kitten itself fits into a single universe. The answer is clearly
no, because the mapping between subjective states and bit strings is
obviously not going to be one to one.

When we experience something then there will always be an enormous
number of different ways that this experience could have been generated
precisely at the microscopic level. What this means is that you or the
kitten don't exist at the level of a single universe, rather you need
to identify yourself with a set of such universes.

The quantum state describing you is in fact a superposition of many
different classical states, but it is an entangled state with the
environment which will behave according to classical physics (in the
sense of there not being interference or violations of the Bell
inequalities).

The structure of such a state defines the algorithm that your brain is
executing to some degree. The entangled state contains information
about the different outputs that result from different inputs. This is
something that you would miss if you assumed from the outset that you
could always picture the kitten as being described by a single
bitstring.

So, the kitten is ultimately an algorithm and to define it you need to
have a range of input and output states. If you have a picture that is
too fine grained then the ktitten ceases to exist at that level.

This then suggests that we should perhaps reverse the logic here. It
could be that the mathematical multiverse is ultimately what really
exists, the members of this multiverse are algorithms, some of which
may descibe kittens. Quantum mechanics should then arise as an
effective description of this multiverse.

Saibal

[j***@public.gmane.org]

single-step through the program with a debugger

might not be possible if you use any quantum processing to emulate the cats brain.

This raises several problems I have been meaning to address in a post.
I know a lot about simulations since I am a visual effects artist and Maya generalist.
I am capable of doing Photo-realistic renderings. I know about caustics and chromatic aberration .

One is about what processes are continuous and what are processes are discrete in reality?
Then what processes are involved in the human mind.

Is time discrete? (This thought has crossed my mind after various odd revelations and while thinking about 3D cellular automata.)

When we emulate fire on a computer with 99.9991% accuracy how is it different from real fire?
An extremely actuate simulation of fire within a computer does not actually set that computer a blaze....
nor does a computer simulating an exact version of itself receiving an error that then crashes, then also crashes the simulating computer...

what the hell is actually going on there?
maybe we are seeing two completely different transformations of matter yet both transformations are spelling out the same exact thing.

Then this must mean due to some law about information, that representations of one part of reality are on equally footing of others if represented accurately. Maybe a way of analyzing the whole process and Shannon entropy comes to mind.

the mapping between subjective states and bit strings is

I know computers will develop AGI , I am being the devils advocate here because I think this might lead to thoughts that opens doors intellectually into artificial intelligence.
Think about simulation and name one thing that has been simulated exactly.
We can simulate a sphere but a sphere is but a construct.
We can run a simulation of the flame's chemical process and a brains chemical process, but is it that process?

Information and simulation might be the same thing.
The Information we gather from reality and our simulation with that information of reality might be the same thing.

regards, John Stifter

[Lee Corbin]

Smitra writes

[Lee wrote]

Post by Lee Corbin
“equivalent” to an actual small cat undergoing
terrible suffering. Then on moral grounds,
wouldn’t we object to someone using his
computer o run that program? In particular,
let’s say that the program executes 10 million
instructions per second, and that, say, it takes
10 seconds to run the program. (Thus the
software kitten lives at 1/5 the rate we do.)

The laws of physics put some limits on this thought
experiments. The information used in running the
simulation of the kitten has to come from somewhere.
If we take this into account, the Mathematical
Multiverse will become a lot more plausible.

There are many confusing aspects to such
thought experiments, let me be the first to
admit. For now, I suggest we leave MWI out
of it, until other aspects become clearer.

Do you consider it *possible* for brain science
to make such tremendous advances in the next
centuries that uploading a worm, an insect, or
a mammal might occur?

If it does become possible, then this question
will become pertinent: is uploading a pet that
is painfully dying of cancer moral?

Lee

[smitra-RSh1/+X/PmFmR6Xm/]

Post by Lee Corbin
Smitra writes

[Lee wrote]

Post by Lee Corbin
“equivalent” to an actual small cat undergoing
terrible suffering. Then on moral grounds,
wouldn’t we object to someone using his
computer o run that program? In particular,
let’s say that the program executes 10 million
instructions per second, and that, say, it takes
10 seconds to run the program. (Thus the
software kitten lives at 1/5 the rate we do.)

The laws of physics put some limits on this thought
experiments. The information used in running the
simulation of the kitten has to come from somewhere.
If we take this into account, the Mathematical
Multiverse will become a lot more plausible.

There are many confusing aspects to such
thought experiments, let me be the first to
admit. For now, I suggest we leave MWI out
of it, until other aspects become clearer.
Do you consider it *possible* for brain science
to make such tremendous advances in the next
centuries that uploading a worm, an insect, or
a mammal might occur?
If it does become possible, then this question
will become pertinent: is uploading a pet that
is painfully dying of cancer moral?
Lee

Yes, this should pose no problems. I would agree that you can have a
virtual pet that is suffering. You can even consider a hypothetical
analogue computer that operates using wheels and gears. Even such a
machine can feel pain. Or you could imagine an ants nest where the
ants implement a complicated program. This system can in principle
also feel pain even though none of the ants may feel anything.

Now, we must also be prepared to question any notions of morality we
have. If all possible universes exist anyway then the suffering kitten
should also exist even if we don't simulate it. The only thing that
changes due to the simulation is that we can see what is going on in
that universe.

Perhaps its better to look at this from the point of view of the
kitten. The kitten is not necessarily going to find itself in a virtual
universe where it is being tortured just because we decide to run a
simulation. It only has a very small probability to find itself in such
a universe, if you a priori accept Tegmark's mathematical multiverse
idea.


Saibal

[Lee Corbin]

[Lee wrote]

Post by Lee Corbin
Do you consider it *possible* for brain science
to make such tremendous advances in the next
centuries that uploading a worm, an insect, or
a mammal might occur?
If it does become possible, then this question
will become pertinent: is uploading a pet that
is painfully dying of cancer moral?

Yes, this [uploading] should pose no problems. I would
agree that you can have a virtual pet that is suffering.
You can even consider a hypothetical analogue computer
that operates using wheels and gears. Even such a
machine can feel pain...
Now, we must also be prepared to question any notions of morality we
have. If all possible universes exist anyway then the suffering kitten
should also exist even if we don't simulate it. The only thing that
changes due to the simulation is that we can see what is going on in
that universe.

But it seems to me that what you have written could
be used to justify any cruel action. If a psychopath takes
you to his basement and shows you a set of teenage
girls he's kidnapped and is torturing, and asks your opinion,
to your suggestion that his actions are making the universe
a worse place, he might retort that since they're already
suffering somewhere in the multiverse--that indeed
the suffering girls exist even if he were to discontinue
his hobby, then "the only thing that is different due to
his actions or a simulation is that we can see what is
going on". Then he suggests that he'll simply show no
more people his basement.

Yes, while everything occurs, it doesn't occur in equal
measure in equal density across the multiverse. We
can (from the frog perspective) try to encourage that
which is good, and avoid actions that increase the
measure of what is bad. Doesn't that make more
persuasive the argument that some programs should
not be run?

Lee

[smitra-RSh1/+X/PmFmR6Xm/]

Post by Lee Corbin

[Lee wrote]

Post by Lee Corbin
Do you consider it *possible* for brain science
to make such tremendous advances in the next
centuries that uploading a worm, an insect, or
a mammal might occur?
If it does become possible, then this question
will become pertinent: is uploading a pet that
is painfully dying of cancer moral?

Yes, this [uploading] should pose no problems. I would
agree that you can have a virtual pet that is suffering.
You can even consider a hypothetical analogue computer
that operates using wheels and gears. Even such a
machine can feel pain...
Now, we must also be prepared to question any notions of morality we
have. If all possible universes exist anyway then the suffering kitten
should also exist even if we don't simulate it. The only thing that
changes due to the simulation is that we can see what is going on in
that universe.

But it seems to me that what you have written could
be used to justify any cruel action. If a psychopath takes
you to his basement and shows you a set of teenage
girls he's kidnapped and is torturing, and asks your opinion,
to your suggestion that his actions are making the universe
a worse place, he might retort that since they're already
suffering somewhere in the multiverse--that indeed
the suffering girls exist even if he were to discontinue
his hobby, then "the only thing that is different due to
his actions or a simulation is that we can see what is
going on". Then he suggests that he'll simply show no
more people his basement.
Yes, while everything occurs, it doesn't occur in equal
measure in equal density across the multiverse. We
can (from the frog perspective) try to encourage that
which is good, and avoid actions that increase the
measure of what is bad. Doesn't that make more
persuasive the argument that some programs should
not be run?
Lee

I do agree that the probabilities are (in general) not equal. However,
we are not external agents that can change the multiverse, we are also
part of the multiverse. So, whatever we choose to do would in fact be a
choice that was going to be made anyway while alternative choices also
exist in the multiverse.

So, we're led to the idea that any choice we make, simply has the
effect of us moving toward certain sectors of the multiverse. Now, you
could say that after you make a choice, you are not the same person as
who you were before, and each choice defines a (slightly) different
person. A cruel version of me has a low measure because of the laws of
physics, which presumably will imply that with a high probability I
will have a sense of morality that stands in the way of me making bad
choices.

Saibal

[Bruno Marchal]

Post by smitra-RSh1/+X/PmFmR6Xm/

Post by Lee Corbin
Smitra writes

[Lee wrote]

Post by Lee Corbin
“equivalent” to an actual small cat undergoing
terrible suffering. Then on moral grounds,
wouldn’t we object to someone using his
computer o run that program? In particular,
let’s say that the program executes 10 million
instructions per second, and that, say, it takes
10 seconds to run the program. (Thus the
software kitten lives at 1/5 the rate we do.)

The laws of physics put some limits on this thought
experiments. The information used in running the
simulation of the kitten has to come from somewhere.
If we take this into account, the Mathematical
Multiverse will become a lot more plausible.

There are many confusing aspects to such
thought experiments, let me be the first to
admit. For now, I suggest we leave MWI out
of it, until other aspects become clearer.
Do you consider it *possible* for brain science
to make such tremendous advances in the next
centuries that uploading a worm, an insect, or
a mammal might occur?
If it does become possible, then this question
will become pertinent: is uploading a pet that
is painfully dying of cancer moral?
Lee

Yes, this should pose no problems. I would agree that you can have a
virtual pet that is suffering. You can even consider a hypothetical
analogue computer that operates using wheels and gears. Even such a
machine can feel pain. Or you could imagine an ants nest where the
ants implement a complicated program. This system can in principle
also feel pain even though none of the ants may feel anything.
Now, we must also be prepared to question any notions of morality we
have. If all possible universes exist anyway then the suffering kitten
should also exist even if we don't simulate it. The only thing that
changes due to the simulation is that we can see what is going on in
that universe.
Perhaps its better to look at this from the point of view of the
kitten. The kitten is not necessarily going to find itself in a virtual
universe where it is being tortured just because we decide to run a
simulation. It only has a very small probability to find itself in such
a universe, if you a priori accept Tegmark's mathematical multiverse
idea.

If we accept the idea that a brain is a machine, the mathematical hypothesis for physics becomes a theorem of theoretical cognitive science, making physics a branch of machine's psychology (or better: machine theology).


This lead to a precise measure problem, and I doubt that the mathematical hypothesis solve that (theoretical computer science) problem by itself.


Both in Everett, and in computationalism, if you run in your "normal reality" a simulation of a kitten torture, the probability that the kitten will be torture will be close to 1 (or about 1/2, if you just copy the kitten).


Bruno










Saibal






http://iridia.ulb.ac.be/~marchal/ http://iridia.ulb.ac.be/~marchal/

[Gary Oberbrunner]

Post by Lee Corbin
First, though, extremely annoying is his terminological predilection of
calling our visible universe

I won't get into your moral quagmire, I'll leave that to others. But I do
think Tegmark does a great disservice muddying the terminological waters by
calling his level two universes a "multiverse" when that term has been in
some use for a long time to mean something different -- namely MWI. His
level 2 universes don't even occupy the same spacetime; they (by
definition) can have no contact with one another and are thus entirely
speculative.

[Bruno Marchal]

Max Tegmark’s recent book “The Mathematical Universe” may or may not have more basic ideas than his previous papers, and the great May 2003 article in Scientific American.
We live in
1. a “level one” infinite universe, containing infinitely many stars and galaxies
2. a larger “level two” infinite universe, in which our level one universe is but one of many, distinguished by different physical constants
3. the “level three” infinite universe consisting of all the MWI branches of our level one universe--presumably we inhabit few other level two universes
4. the (doubtful, I think) “level four” universe containing all mathematical structures, which actually constitute the previous levels
First, though, extremely annoying is his terminological predilection of calling our visible universe (i.e. our Hubble volume) a “universe”. It has the unfortunate implication that we don’t live in the same universe as the good citizens of a planet in the great Andromeda galaxy who are equally advanced technically, for they can see over a million light years further in the directions of the constellations Cassiopeia, Andromeda, and Pegasus, to mention just three. It’s much more convenient to think of our friends in Andromeda inhabiting the same universe we do, but whose Hubble volume is slightly different, of course.
But that terminological annoyance aside, readers of FoR shouldn’t have much to argue with Tegmark about.
Until, that is, we come to his contention that mathematics is all there is, that we as individuals are SASs (Self Aware Structures), and so on. To investigate that contention, and to revisit issues explored online here and elsewhere a decade or more ago, let’s start with computer emulations of conscious beings.
I don’t think it controversial to readers of FoR that computer programs can emulate conscious beings; e.g., for all you know now, you may simply be a running computer program. Then let’s imagine that we have a program that emulates the torture of a small kitten for two minutes. That is, the 2-minute run of the program is “equivalent” to an actual small cat undergoing terrible suffering. Then on moral grounds, wouldn’t we object to someone using his computer to run that program? In particular, let’s say that the program executes 10 million instructions per second, and that, say, it takes 10 seconds to run the program. (Thus the software kitten lives at 1/5 the rate we do.)
Moreover, it is possible (but time consuming) to single-step through the program with a debugger. Whatever time it takes, the end result is still just as morally reprehensible, right?

This simple approach is the quickest way to undermine the mathematical universe, and In this, ultimately, lies its downfall, I'll argue.






I gave an argument that if the brain (even in a weak general sense) is Turing emulable, then "mathematical universe" hypothesis is provable, with some small amount of occam razor. The physical universe emerges from a statistics on first person points of view associated to computations, which exists in a tiny part of arithmetic. Tegmark is just very fuzzy when he invokes the "whole of mathematics", which is not something making mathematical sense. Also, the physical universe cannot be a mathematical structure among others. The relation between math and physics run much deeper, once we take seriously the hypothesis that our brain are "machine": somehow physical realness is a sum on a large set of "mathematical fictions". The emergence of the physical laws is a specific arithmetical phenomena. Comp forces to extend Everett philosophy (physicists obeys to the law of physics) to arithmetic (mathematician obeys to the laws of numbers). The price of comp is that we have to justify the observation of the universe(s) from arithmetic and classical computer science. The bit-qubit relation must go in both direction.


Charges can also be reversed. Are there *any* evidences that there is, at the basic ontological level, more than the arithmetical reality? I am not sure there is one, beyond the aristotelian confusion between truth and observable. In fact, for any digital machine, the existence of more than arithmetic is absolutely undecidable.


I am OK with the idea that running a program implementing suffering is ethically bad, at least before we solve the computationalist measure problem.


Bruno










http://iridia.ulb.ac.be/~marchal/ http://iridia.ulb.ac.be/~marchal/

[Lee Corbin]

Bruno writes

Post by Bruno Marchal
I am OK with the idea that running a program implementing suffering is
ethically bad, at least before we solve the computationalist measure problem.

Is that another term for "observer moment measure problem"?

What's a brief outline of the computationalist measure problem?

Lee
I gave an argument that if the brain (even in a weak general sense) is Turing emulable, then "mathematical universe" hypothesis is provable, with some small amount of Occam razor. The physical universe emerges from a statistics on first person points of view associated to computations, which exists in a tiny part of arithmetic. Tegmark is just very fuzzy when he invokes the "whole of mathematics", which is not something making mathematical sense. Also, the physical universe cannot be a mathematical structure among others. The relation between math and physics run much deeper, once we take seriously the hypothesis that our brain are "machine": somehow physical realness is a sum on a large set of "mathematical fictions". The emergence of the physical laws is a specific arithmetical phenomena. Comp forces to extend Everett philosophy (physicists obeys to the law of physics) to arithmetic (mathematician obeys to the laws of numbers). The price of comp is that we have to justify the observation of the universe(s) from arithmetic and classical computer science. The bit-qubit relation must go in both direction.
Charges can also be reversed. Are there *any* evidences that there is, at the basic ontological level, more than the arithmetical reality? I am not sure there is one, beyond the Aristotelian confusion between truth and observable. In fact, for any digital machine, the existence of more than arithmetic is absolutely undecidable.

I am OK with the idea that running a program implementing suffering is ethically bad, at least before we solve the computationalist measure problem.
Bruno
<







http://iridia.ulb.ac.be/~marchal/

[Bruno Marchal]

Post by Lee Corbin

Post by Bruno Marchal
I am OK with the idea that running a program implementing suffering is
ethically bad, at least before we solve the computationalist measure problem.

Is that another term for "observer moment measure problem"?
What's a brief outline of the computationalist measure problem?

I suggest you consult the following paper where I describe the argument in 8 steps. I have published also, but the copies are not free. It is also the content of my PhD in computer science:


http://iridia.ulb.ac.be/~marchal/publications/SANE2004MARCHALAbstract.html http://iridia.ulb.ac.be/~marchal/publications/SANE2004MARCHALAbstract.html



or


http://dx.doi.org/10.1016/j.pbiomolbio.2013.03.014 http://dx.doi.org/10.1016/j.pbiomolbio.2013.03.014





In a nutshell, knowing that a tiny part of the arithmetical reality (the structure (N,+, *) emulates *all* computations, in a very redundant way), and given the first person indeterminacy FPI (if I am multiplied, any prediction I do on my first person experience is given by a statistics on my possible reconstitutions) and its (the FPI) invariance for delays of reconstitution, any Turing-emulable observers is distributed infinitely often in those (arithmetical) computations, and the laws of physics have to be retrieved from a statistics on computations going through my current computational state "seen from the 1p views".


The argument can be, and has been, translated in arithmetic (using Gödel's method, and Theatetus's definition of knowledge). From this the logic of the observable has been derived, and compared to quantum logic. Up to now it fits. If we find a discrepancy, this would refute computationalism + the classical theory of knowledge. If not, it makes computationalism a generalization (with a statistics on arithmetic) of Everett QM (with a statistics on the universal wave).


So, like the collapse becomes phenomenological in Everett, the wave itself has to be derivable phenomenologically from only arithmetic, and physics becomes a study of arithmetic seen from an inside first person point of view.


I hope this is not to short, but you can ask me precision.


Bruno