Showing posts with label mind uploading/brain copying. Show all posts
Showing posts with label mind uploading/brain copying. Show all posts

Tuesday, January 23, 2024

 After a long hiatus, I am back -- focused and ready to continue tackling the difficult and fascinating questions in the philosophy of physics and consciousness.  I am also investing in my YouTube channel, focusing on clear explanations of difficult concepts to the lay audience.  Here's my most recent video, showing a logical proof why computers will never be conscious, mind uploading is impossible, and more than one instance of a person's conscious state can't exist at more than one place or time in the universe.


Friday, February 4, 2022

Does the Brain Cause Consciousness? Part 3

Is there an afterlife?  Can a computer be conscious?  In Part 1, I pointed out that the popular science answers to these questions depend on the assumption that the brain causes consciousness.  In Part 2, I introduced two statements which, if taken together, imply that the brain does not cause consciousness.  I then explained why Statement 1 is true.  The two statements are:

1)     A brain can be copied.

2)     A person’s conscious state cannot be copied.

In today’s post, I’ll address Statement 2.  This statement is definitely more difficult to prove, which is why it’s so revolutionary.  The clearest explanation, I think, is this 23-minute video that I presented at the 2020 Science of Consciousness conference.  (There is also a more thorough video explanation here.)  The most detailed and precise explanation is in my paper.  But since my goal in this blog post series is to explain things to a lay audience without all the fancy bullshit, this post will (I hope) convince you of Statement 2 with a simpler explanation.

To convince you of Statement 2, I’ll start by assuming the opposite, and then show how it leads to a problem or contradiction.  So let’s assume that you’re in some conscious state that can be copied.  Let’s call that conscious state C1.  Since it can be copied, and we live in a physical world, there must be some underlying physical state that we can copy.  Maybe that physical state is the positions of all the atoms in your brain.  We don't have to know exactly what that physical state is -- the point is that there is some physical state that can be copied.  Let’s call that physical state S1.

Let’s be clear.  You are experiencing conscious state C1.  And that conscious state is entirely created by physical state S1.  So if we were to copy that state S1, and then recreate it somewhere else, then that copy of S1 would produce your conscious state C1.  That’s the whole point of the assumption.  If you are experiencing state C1, and we recreate state C1 on a distant planet in the Wazoo Galaxy (by copying the underlying physical state S1), then you would experience state C1 on that distant planet.[1] 

Now, let’s say we make a copy of physical state S1 (which produces your experience of conscious state C1).  We then recreate it on Mars (preferably in a habitable station), and then simultaneously kill you on Earth.  There’s no problem, right?  You would just experience being on Earth in one moment and then on Mars in the next.  It would just feel like you were teleported to Mars.[2] 

But what if we also recreate physical state S1 (which produces your experience of conscious state C1) on Venus?  What would you experience if there were two versions of you, both experiencing conscious state C1 created by underlying physical state S1? 

More specifically, what would you experience the moment after that?  Being alive on Mars and Venus would be vastly different experiences.  Let’s say that on Mars, your physical state S1 would change to S2M (which creates conscious state C2M), while on Venus, your physical state S1 would change to S2V (which creates conscious state C2V).  State C2M might be the conscious experience of looking out at a vast orange desert, while state C2V might be the conscious experience of looking out at a dark, cloudy, lava-scorched land.  I don’t know exactly what it would feel like, but certainly the two conscious states would differ.

Which conscious state would you experience, C2M or C2V?  There are only three possibilities:

·       Neither

·       Both

·       One or the other

Before proceeding, I should mention something important about physics: locality.  Generally speaking, you can only affect, or be affected by, things that are nearby (or “local”).  If you’re at a baseball game and worried about getting hit in the head with a fly ball, sit far away from home plate.  That way, you’ll have plenty of time to move if a fly ball is heading your way.  Even though the idea is simple, it’s an extremely important and fundamental feature of the physical world.  Einstein is famous for formalizing the concept of locality in his Special Theory of Relativity, which asserts that nothing, including information, can travel faster than the speed of light.

The speed of light is very fast (186,282 miles per second), but it is still finite.  Nothing that happens in a distant galaxy can immediately affect you, because it takes time for information of that event to reach you.  In fact, our own sun is about 8 light-minutes away, which means that if it exploded, it would not affect us for another eight minutes.  The only known violation of locality is quantum entanglement, but even quantum entanglement does not allow information or matter to be transmitted faster than light.

Getting back to the above example, when we recreate physical states S1 on Mars and Venus, those states are not local to each other, which means they can’t affect each other.  And Mars and Venus are far enough apart that subsequent physical states (S2M on Mars and S2V on Venus) also can’t affect each other.[3]

We already know that when we create state S1 on Mars (and kill you on Earth), you would experience being on Earth in one moment and then on Mars in the next, as if you teleported to Mars.  Your subsequent conscious states (C2M, C3M, C4M, and so forth) would change according to what you experienced on Mars.

And if we had instead created state S1 on Venus (but not on Mars), you would experience being on Earth in one moment and then on Venus in the next, as if you teleported to Venus.  Your subsequent conscious states (C2V, C3V, C4V, and so forth) would change according to what you experienced on Venus.

So what would happen if we create state S1 (which produces conscious state C1) on Mars and on Venus?  Which conscious state will you next experience, C2M or C2V?  As I said before, there are only three possibilities, which I’ll analyze below:

·       Neither

·       Both

·       One or the other

Neither.  Maybe it’s neither.  Maybe the universe doesn’t like it when we create multiple copies of a conscious state, so when you create two or more copies, they both get blocked or eliminated or something.  Here’s the problem.  When you are created on Mars, your conscious state cannot be affected by what is happening on Venus because the two events are nonlocal.  There is no way for your physical state S1 on Mars to “know” that state S1 was also created on Venus because it takes time for information to travel from Venus to Mars, even if that information is traveling at the speed of light.  Your physical state S1 will change to S2M (which produces your conscious state C2M) long before a signal can be sent to stop it.  Therefore, you will experience conscious state C2M, so the correct answer cannot be “neither.”

Both.  Maybe you will experience both conscious states C2M and C2V.  I certainly have no idea what it’s like to experience two different conscious states at (what I would perceive as) the same time.  Nevertheless, maybe it’s possible.  But here’s the problem.  Your conscious experience of C2M is created by physical state S2M, which is affected by stuff on Mars, while your conscious experience of C2V is created by physical state S2V, which is affected by stuff on Venus.  For example, if state C2M is your experience of looking out at a vast orange desert, it’s because light rays bouncing off Martian dunes interacted with your physical state S1 to produce S2M.  But information about that interaction is inaccessible to whomever is experiencing state C2V on Venus, once again because information does not travel fast enough between the two planets.  Therefore, whoever is experiencing state C2V on Venus cannot also be experiencing state C2M on Mars.  Therefore, maybe you’re experiencing state C2M or C2V, but you can’t be experiencing both.

One or the other.  The correct answer to the above question is not “neither” and it’s not “both.”  The only remaining option is that you experience either C2M or C2V.  But which one?  How could nature choose?  Maybe you experience the “first” one created.  The problem here is, once again, nonlocality.  Let’s say that, according to my clock on Earth, state S1 is created on Mars at 12:00:00pm, and state S1 is created on Venus at 12:00:01pm – in other words, one second later by my clock.  The problem is that there is no way for state S1 on Venus to “know” about the creation of state S1 on Mars (and to then prevent your conscious experience of state C2V on Venus), because it takes much longer than one second for information to travel between the two planets.[4]  Therefore, the universe cannot “choose” between C2M or C2V based on time.  And because state S1 on Mars is physically identical to state S1 on Venus, there is no other physical means by which the universe can choose one over the other.  If S1 changes to S2M (which produces C2M) on Mars and S1 changes to S2V (which produces C2V) on Venus, there is no known physical means for the universe to somehow decide that you will experience only C2M or C2V (but not both).  Therefore, you cannot experience just one or the other.    

We have ruled out all three possibilities.  What does this mean?  It means that the original assumption – that a person’s conscious state can be copied – is wrong.  Think about the logic this way:

          i.          If statement A is true, then either B or C or D must be true. 

        ii.          But B, C, and D are all false. 

      iii.          Therefore, statement A must be false.

In this case, statement A is “a person’s conscious state can be copied” and statements B, C, and D correspond to “neither,” “both,” and “one or the other,” like this:

       i.          If a person’s conscious state can be copied, then we can put copies on Mars and Venus.  Either the person will experience neither copy, or will experience both copies, or will experience one or the other. 

     ii.          I showed that none of these are possible (because they conflict with special relativity). 

   iii.          Therefore, a person’s conscious state cannot be copied.

 

If you recall, this conclusion is the same as Statement 2 at the beginning of this post:

1)     A brain can be copied.

2)     A person’s conscious state cannot be copied.

If I have convinced you of Statement 2 in this post, and of Statement 1 in the previous post, then what do they imply?  This is what they imply:

 

If a brain can be copied, but a conscious state cannot, then the brain cannot create consciousness.

 

Certainly the brain can affect consciousness.  If someone sticks electrodes in my brain, I have no doubt that it will probably affect my conscious experience.  But consciousness cannot be produced entirely by the brain.  In other words, conscious experience must depend on stuff (events and states) beyond the skull. 

This conclusion should be shocking, but taken seriously, by anyone who wants to understand and scientifically study consciousness.  Its implications are significant.  For example, getting back to the big-picture questions posed in Part 1, can a computer be conscious?  A digital computer has a state that can be easily copied.  If it didn’t, we wouldn’t be able to copy files, buy software, or even run software.  But as I proved above, a person’s conscious state cannot be copied.  Therefore, a person’s conscious state cannot be embedded or executed on a digital computer, because if it could, then the person’s conscious state could be easily copied.  A digital computer cannot be conscious because conscious states cannot be copied.[5]  Also, mind uploading is impossible because if a computer can’t be conscious, then there’s no way to upload or simulate a conscious mind on a computer.  Also, consciousness cannot be algorithmic.  An algorithm is a set of instructions that can be executed on any general purpose computer.  Once again, an algorithm can be easily copied but a person’s conscious state can’t, so consciousness cannot be algorithmic.

And what about the other question posed in Part 1: Is there an afterlife?  Well, my arguments here certainly don’t prove that consciousness continues after brain death.  However, the strongest (and perhaps only) scientific argument against an afterlife depends on the assumption that the brain causes consciousness.  But I’ve shown that’s false.  Further, I’ve shown that consciousness transcends the brain, at least to some degree.  The fact that what we consciously perceive is produced by something beyond our brains is at least circumstantial evidence that the existence of consciousness does not necessarily depend on whether a brain is alive.

The brain does not cause consciousness.  Much of what science tells us about consciousness, to the extent that it relies on an invalid assumption, is likely false. 


[1] If physical state S1 wasn’t sufficient to produce a conscious state of YOU – in other words, if physical state S1 is inadequate to produce your conscious identity – then consciousness must be produced in part by something nonphysical.  And that would be a real problem for scientists!

[2] If you weren’t also killed on Earth, this would be the “teleportation problem” that Nobel Prize winner Roger Penrose discusses in The Emperor’s New Mind.

[3] To use physics language, the creation of states S1 on Mars and Venus (and their subsequent evolutions to S2M and S2V, respectively) are spacelike separated events.  The argument I’m making here applies equally to timelike separated events, which I discuss in my paper.

[4] In fact, there is no such thing as “simultaneous” events when we are talking about spacelike separated events.  Even though my clock may say that S1 was created on Mars first, the clock of another observer may say that S1 was created on Venus first.  There is no objective fact about which event occurs first if the events are spacelike separated.

[5] Maybe this argument doesn’t apply to quantum computers.  However, as I’ve explained repeatedly, a quantum computer sufficiently large to create anything we might regard as intelligent is just as physically impossible as producing Schrodinger’s Cat or Wigner’s Friend. 

Thursday, February 3, 2022

Does the Brain Cause Consciousness? Part 2

Is there an afterlife?  Can a computer be conscious?  In Part 1, I pointed out that the popular science answers to these questions depend on an often unstated assumption:

Assumption: The brain causes consciousness.

I am going to show in this and subsequent posts why there is very good reason to doubt this assumption, and why it’s almost certainly false.  To do that, I’m going to try to convince you of two statements [1] which, taken together, imply that the brain does not cause consciousness:

1)     A brain can be copied.  (Even if it cannot be done today due to technological limitations, there is no physical law preventing the physical state of a brain from being copied.)

2)     A person’s conscious state cannot be copied.

In today’s post, I’ll address Statement 1.  First of all, I think most people, particularly scientists, would already agree with it.  And since my goal is to convince you, the reader, then if you already agree with it, there’s no need to read further.  Instead, just move on to the next post in this series, where I’ll address Statement 2.

Of course, no one thinks that a brain can be copied today.  But what physical law prevents copying a brain in the future?  The only known physical principle of which I’m aware is the quantum no-cloning theorem, which says that a quantum state cannot be copied.  And a brain, like all things in the universe, is presumably in a quantum state, so in that sense it can never be perfectly copied.  But that doesn’t matter as long as quantum effects are not relevant to the brain and its functions.  In other words, the only thing that would prevent a brain from being copied adequately to replicate consciousness is if consciousness depends on quantum effects. 

For example, if a conscious state depended on quantum entanglements with objects outside the brain, then there is inadequate information in the brain to specify a conscious state.  Quantum entanglement is “nonlocal,” which means that Object A can affect entangled Object B instantaneously, even if they are separated by a large distance, and the effect is not limited by the speed of light.  So if my current conscious state depends at least in part on an event in another galaxy (which we cannot detect until we receive light from the event), then consciousness is nonlocal.  This recent paper argues that consciousness is nonlocal, but I doubt many in the scientific community have taken notice.

Another way that consciousness may depend on quantum effects is if, to copy the brain, you’d have to measure the state of objects in the brain (like neurons) so precisely that the Heisenberg Uncertainty Principle kicks in, and the measurement itself starts changing the brain’s physical state.  For example, Scott Aaronson suggests in this paper that if a brain is “unclonable for fundamental physical reasons,” then that unclonability could be a consequence of quantum no-cloning if the granularity a brain would need to be simulated at in order to duplicate someone’s subjective identity was down to the quantum level. 

In general, though, few scientists believe that consciousness or brain function depend on quantum effects, and most who discuss the possibility are quickly dismissed as mystics or pseudoscientists.[2]  As long as consciousness does not depend on quantum effects, then we don’t need to worry about quantum no-cloning, and there is nothing that would prevent a future engineer from scanning a person’s brain and then reproducing a functional duplicate with the same conscious state.

Are you convinced of Statement 1 yet?… that a brain can be copied in principle?  Maybe you’re still concerned about possible quantum effects.  OK, here’s another argument.

The amount of information that can be contained in a volume of space is limited.  This is called the Bekenstein bound.  It’s a ridiculously large number but it’s still finite.  For example, the Bekenstein bound Wikipedia page calculates that the maximum information necessary to recreate a human brain, including its entire quantum state, is on the order of 10^42 bits (where a single “bit” of information is either a 0 or 1).  That’s a huge number… it looks like 1,000,000,000,000,000,000,000,000,000,000,000,000,000,000, but it’s still much, much smaller than the number of particles in the universe.  Also, the Bekenstein bound for the brain is an upper physical limit that’s based on a brain so dense with information that it’s right on the verge of collapsing into a black hole!  I think it’s reasonable to surmise that we aren’t walking around with potential black holes in our skulls, so the actual information necessary to specify the quantum state of a brain is probably much, much, much, much smaller than 10^42 bits.  But it doesn’t actually matter.  Here’s why.

Even if we can’t in practice copy a human brain, the universe should be able to.  I’m referring to a Boltzmann Brain.  Physicists currently believe that essentially any physical state can be created by randomness (i.e., accident).  So even though it’s extremely unlikely, a physicist will say that there is some chance that atoms and particles will accidentally come together somewhere in the universe to create your brain.  And even if we include quantum effects, and even if that accidental collection of atoms has to specify the 10^42 bits that could potentially be specified in the physical state of your brain, there is some nonzero probability that it will occur. 

In other words, there is no known physical law that will prevent the exact recreation of your brain elsewhere.  The universe can copy your brain, even if your brain’s function depends on quantum effects.  Therefore, a brain can be copied.  Statement 1 is true.

In my next post, I’ll address Statement 2.  As for now, do you have any questions or concerns about Statement 1?


[1] As I mentioned previously, I would ordinarily try to be more precise with my words, arguments, and proofs.  But the purpose of this and subsequent posts is to write more colloquially without alienating lay readers.  Better precision can be found, e.g., in my papers.

[2] Don’t forget that consensus does not equal truth.  There is, and perhaps always has been, a bully culture in science, which is why scientific paradigms tend to be changed only by independent mavericks.

Wednesday, February 2, 2022

Does the Brain Cause Consciousness? Part 1

I have spent so much time and effort trying (and ultimately failing) to successfully communicate with people in the physics and philosophy academies, using their complicated and abstruse language and math equations, that I’ve made many of my insights, discoveries, and contributions completely inaccessible to the rest of the world, including my own friends and family.

My close friend Adam recently asked me some important questions, like whether computers could be conscious.  Of course, I’ve answered this question many times, and in great detail, on this blog and in my papers (particularly this and this).  But I realized that I really only addressed people who already knew the language of quantum mechanics, computer science, philosophical logic, and so forth.  So in this and subsequent posts, I’m going to try to address some important questions in direct, ordinary language without all the bullshit jargon.

Today, I want to mention two such questions:

·       Is there an afterlife?

·       Can a computer be conscious?

Ask these questions of a physicist, biologist, or computer scientist, and probably the vast majority will answer firmly and with conviction: No, there is no afterlife; Yes, a computer can be conscious.  And if you probe them further as to why they are so certain of these answers, you’ll find that there is an (often unstated) assumption that pervades the scientific community about consciousness:

Assumption: The brain causes consciousness.[1]

Is that assumption true?  If it is, then it’s not unreasonable to believe that consciousness ends when the brain dies.  Or that someday we’ll be able to copy the brain and recreate a person’s consciousness.  Or that a person’s brain could be simulated in a computer, thus producing consciousness in a computer.

But again, all these popular ideas stem from that one assumption, and there aren’t many scientists who question it (or even acknowledge it as an assumption).  So that’s where I’ll start.  Consider, again, the assumption:

Assumption: The brain causes consciousness.

Several questions for you about that assumption:

·       Do you believe it?

·       If so, why?  What evidence do you have that it is true?

·       What evidence has the scientific community offered to support it?

·       Which beliefs depend on it?  For example, anyone who believes that consciousness ends with brain death necessarily makes the above assumption.  Anyone who believes that a computer will someday be conscious by simulating a brain also makes the above assumption.  Many, many other popular science beliefs depend on this assumption.

·       What if the assumption is incorrect?  Is it possible to prove that it is false?  How might it be disproven?  If the assumption could actually be disproven, how might that impact your beliefs?  How might it impact the popular scientific beliefs about consciousness?



Next in this series: Part 2
Last in this series: Part 3


[1] Note on this post: Ordinarily, I would try to be more precise with my words.  For example, the assumption is actually that a conscious state entirely depends on the physical state of a living brain, but this is where the eyes of ordinary readers start to glaze over.  So I won’t be so precise in this and related future blog posts.

Sunday, May 30, 2021

Physics, Immortality, and the Afterlife

What does physics tell us about the possibility of immortality or an afterlife?

First, let’s address the elephant in the room.  To the physics community, “afterlife” often implies “religion” often implies “stupidity.”  Bullies like Richard Dawkins have made it very clear that anyone who even suggests the existence of God or an afterlife is intellectually inferior.  Oddly, this assertion directly conflicts with several arguments based on currently-understood physics (and often made, ironically, by atheists) that immortality is possible.  I’ll discuss below some of these arguments.  Importantly, any physicist who tells you with certainty that there is no afterlife is not only mistaken, but is ignorant of the direct logical implications of his/her own beliefs about physics.

Note: “Afterlife” and “immortality” are not technically the same.  Immortality might be interpreted as “never dying,” while an afterlife might be interpreted as “consciousness after one has died.”  However, from a physics standpoint, this is often a distinction without a difference.  For example, if mind uploading is possible, then it can be done before or after a person’s brain is dead.

Postponing Death

My close friend (and former MIT debate champion) once made the following (valid) argument:

·       Technology (e.g., medicine) is allowing humans to live longer and longer.

·       There is some tiny but nonzero probability 0<p<<1 that we will develop the technology to indefinitely postpone death.  For example, imagine that in the next 50 years we figure out how to make humans live to age 150, and in the following 50 years we figure out how to make humans live to age 200, and so on.  Then someone born today could indefinitely postpone death.

·       The universe will continue expanding forever.  (Most physicists believe that the universe has positive curvature.)

·       p * ∞ = ∞.

·       Therefore, the life expectancy of a person born today is infinite!

The argument applies equally to an afterlife as to immortality if we simply replace the second statement with “There is some tiny but nonzero probability 0<p<<1 that we will develop the technology to reanimate a dead person’s brain.”  (After all, that’s why the quacks at the Brain Preservation Foundation recommend cryogenic freezing of one’s brain, which I’m certain is not cheap.)  If we can agree that it is at least physically possible to indefinitely delay death (or to reanimate a dead brain), then physical immortality/afterlife cannot be ruled out.

Mind uploading

If you can upload your conscious awareness onto a computer, then you can live forever because a computer can be indefinitely operated and repaired.  Okay, okay, you still need energy to flow, which will stop when the universe experiences its predicted heat death in at least a googol years.  Even still, Michio Kaku in his fascinating book, Parallel Worlds, points out that conscious awareness would slow down commensurate with decreased energy transfer so that one’s subjective conscious experience wouldn’t notice.

Again, there’s no difference here between “immortality” and “afterlife” since the fundamental assumption of mind uploading (and algorithmic consciousness) is that a conscious state is just software running on a computer, and that can be done long after one’s death.  (It can also be done before one’s death, which leads to all kinds of ridiculous paradoxes that I discuss in this paper.) 

The problem is that I showed in this paper (and this) that consciousness is not algorithmic, which means it cannot be uploaded to or executed by a computer (whether digital or quantum).  Mind uploading is not physically possible.

Quantum Suicide

Max Tegmark, proponent of the wacky and unscientific Many Worlds Interpretation (“MWI”) of quantum mechanics, proposed the notion of quantum suicide (although really it’s just “quantum death” because it applies independently of intention) as an empirical test of MWI.  The idea is this:

·       Stand in front of a “quantum gun” that is designed so that when the trigger is pulled, whether a bullet actually fires from the gun (and kills you) depends on the outcome of a quantum mechanics (“QM”) event.

·       Universal linearity of QM – i.e., the assumption of U which I dispute here – implies that the quantum event entangles with the bullet, which entangles with you, to produce a Schrodinger’s-Cat-like state involving you in a superposition of states |dead> and |alive>. 

·       If MWI is correct, then both states actually occur/exist (albeit in different “worlds” that are exceedingly unlikely to quantum mechanically interfere).

·       Since you cannot consciously survive death – a huge assumption! – then the only state you can consciously observe is the one involving |alive>, which means that you are guaranteed to “survive” the pull of the quantum gun trigger.

·       You can repeat this as many times as you want, and every time you will be guaranteed to observe the outcome in which you are alive.

The argument is wrong for several reasons that I point out in the Appendix of this paper.  One problem is that every chance event is fundamentally (amplification of) a quantum event, which means that essentially every death-causing event is akin to quantum suicide/death.  But that means that, if Tegmark’s argument is correct, then nobody can actually die

But the main problem is that the argument depends on the unjustified and irrational assumption of U.  If Schrodinger’s Cat and Wigner’s Friend can’t exist, then the quantum suicide experiment can never get off the ground. 

Boltzmann Brain

The Boltzmann Brain concept is the notion that, given enough time, every physical state will repeat itself.  Or: due to random quantum fluctuations, given enough time, every possible physical configuration that can fluctuate into existence will fluctuate into existence.  So, eventually, even trillions of years after humanity has gone extinct, the conscious state you are experiencing at this moment will be recreated (and presumably re-experienced) again.  And again and again. 

Whether such experiences count as immortality or afterlife makes no difference, because the Boltzmann Brain concept is impossible.  I showed in this paper that physical instantiations of the same conscious state cannot exist at different points in spacetime. 

Tipler’s Physics of Immortality

Frank Tipler is a Christian who was largely ostracized from the academic world for showing how what is currently understood about physics could support the notions of a Christian God and afterlife.  His book, The Physics of Immortality, is interesting but dense.  At the risk of oversimplifying his analysis, I think his fundamental argument is really the Boltzmann Brain in disguise.  He relies heavily on the Bekenstein Bound (which I discussed here) and the notion of Eternal Recurrence to show that consciousness cannot end.

His analysis is wrong for several reasons.  First, the Bekenstein Bound assumes the constancy of the informational content in a given volume (i.e., that Planck’s constant is constant, which may be false).  Second, and more importantly, it assumes that a conscious state is just a list of numbers (even if it’s a huge list) that must be contained within the volume of a brain, and therefore that consciousness is algorithmic, which I’ve shown is false.  If consciousness cannot be reduced to a set of bits (and/or their algorithmic manipulation), then the number of bits that can fit in a given volume is irrelevant.

Reversibility

If physical systems are truly reversible, then – at least according to some – it should be possible in principle to physically reverse a person’s death (or even some or all of a person’s life).  It’s certainly not obvious that constantly “undoing” someone’s death results in a meaningful kind of immortality.  Still, maybe the goal of reversing someone’s death is to copy their consciousness into another physical system or upload it to a computer.  But I’ve already pointed out several times that this is impossible.  Either way, the entire argument is moot because physical reversibility of large systems is a logical contradiction.

What do I actually believe?

First, let me point out the crazy irony of this blog post so far.  I, the crackpot who believes in God, am trying to explain why several physicists’ arguments for immortality or afterlife are wrong!  In fact, the only one of the above arguments that I can’t completely rule out is Postponing Death, even though I regard it as extremely implausible to postpone death indefinitely.

Having said that, we do not understand consciousness.  No physicist, neurobiologist, physician, psychologist, computer scientist, philosopher, etc., understands consciousness, and anyone who claims to understand it is likely introducing unstated assumptions.  For example, most scientists who academically discuss consciousness assume that consciousness is created entirely by the brain, which is why the hackneyed “brain-in-a-vat” thought experiment – the namesake of this blog – is so pervasive in the literature. 

After all, if consciousness is entirely a product of the brain (or, more generally, on a local region of spacetime that may enclose the brain), then the above arguments are a lot more tenable.  That is, if a conscious state supervenes on a physical state that is entirely (locally) contained in some volume, then Tipler’s argument based on the Bekenstein Bound seems to apply; the total information specifying that conscious state is finite and, as nothing more than a string of numbers, can be copied and uploaded to a computer; and so forth.

In fact, given so many physical arguments for immortality, one might even wonder what physical arguments there are against immortality.  There is only one: assume that consciousness is entirely a product of the (living) brain; then death of the brain ends consciousness.  And if that assumption is wrong, then there is literally no existing scientific evidence against immortality or an afterlife.

But that assumption is wrong.  Conscious states are not local and they cannot be copied to different places in spacetime.  (Stoica makes a related and fascinating argument here that mental states are nonlocal.)  If they’re nonlocal, then they must physically depend on nonlocal (quantum) entanglements among objects and particles throughout the universe.  That is, what physically specifies my conscious state logically must extend beyond my brain.  There is certainly no doubt that my brain affects my consciousness, but it cannot be entirely locally responsible for it.  The fact that events and physical relationships that extend far beyond my brain are at least partially responsible for my consciousness leads me to surmise that these conscious-identity-producing physical relationships will persist long after the atoms in my brain are no longer arranged in their current configuration.  This is the beginning of an as-of-yet undeveloped physical argument for immortality/afterlife.

So, what do I really believe about an afterlife?  I won’t mince words.  I am certain that my consciousness is eternal; I am certain that my consciousness awareness will not permanently end if/when my brain dies.  In future posts, I will give logical and physical arguments to support these assertions, but I wanted first to devote a blog post to what currently-understood physics implies. 

Friday, March 19, 2021

The Folly of Brain Copying: Conscious Identity vs. Physical Identity

The notion of “identity” is a recurring problem both in physics and in the nature of consciousness.  Philosophers love to discuss consciousness with brain-in-a-vat type thought experiments involving brain copying.  The typical argument goes something like this:

i)          The brain creates consciousness.

ii)         It is physically possible to copy the brain and thereby create two people having the same conscious states.

iii)        Two people having the same conscious states each identifies as the “actual” one, but at least one is incorrect.

iv)        Therefore, conscious identity (aka personal identity) is an illusion.

I spent a long time in Section II of this paper explaining why questioning the existence of conscious identity is futile and why the above logic is either invalid or inapplicable.  Yes, we have a persistent (or “transtemporal”) conscious identity; doubting that notion would unravel the very nature of scientific inquiry.  Of course, you might ask why anyone would actually doubt if conscious identity exists.  Suffice it to say that this wacky viewpoint tends to be held by those who subscribe to the equally wacky Many Worlds Interpretation (“MWI”) of quantum mechanics, which is logically inconsistent with a transtemporal conscious identity.

I showed in Section III of the above paper why special relativity prevents the existence of more than one instantiation of a physical state creating a particular conscious state.  In other words, at least one of assumptions i) and ii) above is false.  For whatever reason, the universe prohibits the duplication or repeating of consciousness-producing physical states.  In Section IV(A) of the same paper, I suggested some possible explanatory hypotheses for the mechanism(s) by which such duplications may be physically prevented, such as quantum no-cloning. 

Nevertheless, the philosopher’s argument seems irresistible... after all, why can’t we make a “perfect” copy of a brain?  If multiple instances of the same conscious state are physically impossible then what is the physical explanation for why two consciousness-producing physical states cannot be identical?  I finally realized that conscious identity implies physical identity.  In other words, if conscious identity is preserved over time, then physical identity must also be preserved over time, and this may help explain why the philosopher’s brain-copying scheme is a nonstarter.

I’d been struggling for some time with the notion of physical identity, such as in this blog post and this preprint.  The problem can be presented a couple ways:

·         According to the Standard Model of physics, the universe seems to be made up of only a handful of fundamental particles, and each of these particles is “identical” to another.  For example, any two electrons are identical, as are any two protons, or any two muons, etc.  The word “identical” is a derivative of “identity,” so it’s easy to confuse two “identical” electrons as being indistinguishable and thus having the same (or indistinct) identities.  So if all matter is made up of atoms comprising electrons, protons, and neutrons, then how can any particular clump of atoms have a different identity than another clump made of the same type of atoms?

·         Let’s assume that consciousness is created by physical matter and that physical matter is nothing but a collection of otherwise identical electrons, protons, and neutrons.  In the above paper I showed that if conscious identity exists, then conscious states cannot be copied or repeated.  And that means there is something fundamentally un-copiable about the physical state that creates a particular conscious state, which would seem odd if all matter is fundamentally identical. 

·         Consciousness includes transtemporal identity.  Assuming physicalism is true, then conscious states are created by underlying physical states, which means those physical states must have identity.  But physics tells us that physical matter comprises otherwise identical particles.

I finally realized that this problem can be solved if particles, atoms, etc., can themselves have identity.  (I do not mean conscious identity... simply that it makes sense to discuss Electron “Alice” and Electron “Bob” and keep track of them separately... that they are physically distinguishable.)  An object’s identity can be determined by several factors (e.g., position, entanglements and history of interactions, etc.) and therefore can be distinguished from another object that happens to comprise the same kind of particles.  Two physically “identical” objects can still maintain separate “identities” to the extent that they are distinguishable.  And we can distinguish (or separately identify) two objects, no matter how physically similar they may otherwise be, by their respective histories and entanglements and how those histories and entanglements affect their future states. 

Where does physical identity come from?  It is a necessary consequence of the laws of physics.  For instance, imagine we have an electron source in the center of a sphere, where the sphere’s entire surface is a detector (assume 100% efficiency) that is separated into hemispheres A and B.  The detector is designed so that if an electron is detected in hemisphere A, an alarm immediately sounds, but if it is detected in hemisphere B, a delayed alarm sounds one minute later.  The source then emits an electron, but we do not immediately hear the alarm.  What do we now know?  We know that an electron has been detected in hemisphere B and that we will hear an alarm in one minute.  Because we know this for certain, we conclude that the detected electron is the same as the emitted electron.  It has the same identity.  The following logical statement is true:

(electron emitted) ∩ (no detection in hemisphere A) à (detection in hemisphere B)

But more importantly, the fact that the above statement is true itself implies that the electron has identity.  In other words:

[(electron emitted) ∩ (no detection in hemisphere A) à (detection in hemisphere B)]

à (the electron emitted is the electron detected in hemisphere B)

(On retrospect, I feel like this is obvious.  Of course physical identity is inherent in the laws of physics.  How could Newton measure the acceleration of a falling apple if it’s not the same apple at different moments in time?)

So if electrons can have identity, then in what sense are they identical?  Can they lose their identity?  Yes.  Imagine Electron Alice and Electron Bob, each newly created by an electron source and having different positions (i.e., their distinct wave packets providing their separate identities).  The fact that they are distinguishable maintains their identity.  For example, if we measure an electron where Electron Bob cannot be found, then we know it was Electron Alice.  However, electrons, like all matter, disperse via quantum uncertainty.  So what happens if their wave functions overlap so that an electron detection can no longer distinguish them?  That’s when Bob and Alice lose their identity.  That’s when there is no fact about which electron is which.  (As a side note, Electron Bob could not have a conscious identity given that when he becomes indistinguishable with Electron Alice, even he cannot distinguish Bob from Alice.  This suggests that conscious identity cannot even arise until physical identity is transtemporally secured.)

This realization clarified my understanding of conscious identity.  My body clearly has an identity right now in at least the same sense that Electron Bob does.  What would it take to lose that physical identity?  Well, it wouldn’t be enough to make an atom-by-atom copy of the atoms in my body (call it “Andrew-copy”), because Andrew-copy would still be distinguishable from me by nature, for example, of its different location.  Rather, the wave functions of every single particle making up my body and the body of Andrew-copy would have to overlap so that we are actually indistinguishable.  But, as I showed in this paper, that kind of thing simply can’t happen with macroscopic objects in the physical universe because of the combination of slow quantum dispersion with fast decoherence.

What would it take for me to lose my conscious identity (or copy it, or get it confused with another identity, etc.)?  Given that conscious states cannot be physically copied or repeated, if conscious identity depends only the particular arrangement of otherwise identical particles that make up matter, then we need a physical explanation for what prevents the copying of that particular arrangement.  But if conscious identity depends on not just the arrangement of those (otherwise identical) particles but also on their physical distinguishability, then the problem is solved.  Here’s why.  Two macroscopic objects, like bowling balls, will always be physically distinguishable in this universe.  Bowling Ball A will always be identifiably distinct from Bowling Ball B, whether or not any particular person can distinguish them.  So if my conscious identity depends at least in part on the physical distinguishability of the particles/atoms/objects that create my consciousness, then that fact alone would explain why conscious states (and their corresponding transtemporal identity) cannot be copied.

Let me put this another way.  Identity is about distinguishability.  It is possible for two electrons to be physically indistinguishable, such as when the wave states of two previously distinguishable electrons overlap.  However, it is not possible, in the actual universe, for a cat (or any macroscopic object) and another clump of matter to be physically indistinguishable because it is not possible for the wave states of these two macroscopic objects to overlap, no matter how physically similar they may otherwise be.  A cat’s physical identity cannot be lost by trying to make a physical copy of it.  It is not enough to somehow assemble a set of ≈10^23 atoms that are physically identical to, and in a physically identical arrangement as, the ≈10^23 atoms comprising the cat.  Each of those constituent atoms also has a history of interactions and entanglements that narrowly localize their wave functions to such an extent that overlap of those wave functions between corresponding atoms of the original cat and the copy cat is physically impossible.  (See note below on the Myth of the Gaussian.)

Imagine that someone has claimed to have made a “perfect copy” of me in order to prove that conscious identity is just an illusion.  He claims that Andrew-copy is indistinguishable from me, that no one else can tell the difference, that the copy looks and acts just like me.  Of course, I will know that he’s wrong: even if no one else can distinguish the copy from me, I can.  And that alone is enough to establish that Andrew-copy is not a perfect copy.  But now I understand that my conscious identity implies physical identity – that my ability to distinguish Andrew-copy from me also implies physical distinguishability.  There is no such thing as a perfect physical copy of me.  Even if the atoms in Andrew-copy are in some sense the same and in the same configuration as those in my body, and even if some arbitrary person cannot distinguish me from Andrew-copy, the universe can.  The atoms in Andrew-copy have a history and entanglements that are distinguishable from the atoms in my body, the net result being that the two bodies are physically distinguishable; their separate physical identities are embedded as facts in the history of the universe.

So if the universe can distinguish me from Andrew-copy, then why should it be surprising that I can distinguish myself from Andrew-copy and that I have an enduring conscious identity?  The question is not whether some evil genius can make a physical copy of my body that is indistinguishable to others.  The question is whether he can make a copy that is indistinguishable to me or the universe.  And the answer is that he can’t because making that copy violates special relativity. 

 

Note on the Myth of the Gaussian:

Physicists often approximate wave functions in the position basis as Gaussian distributions, in large part because Gaussians have useful mathematical properties, notably that the Fourier transform of a Gaussian is another Gaussian.  Because the standard deviation of a Gaussian is inversely related to the standard deviation of its Fourier transform, it clearly demonstrates the quantum uncertainty principle whereby the commutator of two noncommuting operators is nonzero.  An important feature of a Gaussian is that it is never zero for arbitrarily large distances from the mean.  This treatment of wave functions often misleads students into believing that wave functions are or must be Gaussian and that: a) an object can be found anywhere; and b) the wave states of any two arbitrary identical objects always overlap.  Neither is true. 

Regarding a), physics students are often given the problem of calculating the probability that his/her body will quantum mechanically tunnel through a wall, or even tunnel to Mars; the calculation (which is based on the simple notion of a particle of mass M tunneling through a potential barrier V) always yields an extremely tiny but nonzero probability.  But that’s wrong.  Setting aside the problem with special relativity – i.e., if I am on Earth now, I can’t be measured a moment later on Mars without exceeding the speed of light – the main problem is physical distinguishability.  The future possibilities for my body (and its physical constituents) are limited by their histories and entanglements. 

While some electron may, due to quantum dispersion or being trapped in a potential well, develop a relatively wide quantum wave packet over time whose width “leaks” to the other side of the wall/potential barrier, this requires that the electron remain unmeasured (i.e., with no new correlations) during that time period.  But the particles and atoms in a human body are constantly “measuring each other” through decoherence so that their individual wave packets remain extremely tightly localized.  In other words, my body doesn’t get quantum mechanically “fuzzy” or “blurry” over time.  Thus none of the wave packets of the objects comprising my body get big enough to leak through (or even to) the wall.  More to the point, the QM “blurriness” of my body is significantly less than anything that can be seen... I haven’t done the calculation, but the maximum width of any wave packet (not the FWHM of a Gaussian, which extends to infinity, but the actual maximum extent) is much, much, much smaller than the wavelength of light. 

As I showed above, physical distinguishability is an inherent feature of the physical world.  An object that appeared on the other side of the wall that happened to look like my body would be physically distinguishable from my body and cannot be the same.  That is, there is no sense in which the body that I identify as mine could quantum mechanically tunnel to Mars or through a wall – that is, there is ZERO probability of me tunneling to Mars or through a wall.  If I have just been measured in location A (which is constantly happening thanks to constant decohering interactions among the universe and the objects comprising my body), then tunneling to location B requires an expansion of the wave packets of those objects to include location B – i.e., my tunneling to B requires a location superposition in which B is a possibility.  But past facts, including the fact that I am on Earth (or this side of the wall) right now have eliminated all configurations in which my body is on Mars (or on the other side of the wall) a moment later.