Are parallel universes real?

Since I gave my Google Tech Talk on quantum mechanics nearly four years ago I get a steady stream of email from people asking questions about it.  There's one question I keep getting over and over which I thought I'd just try to answer once and for all so I could point people to this answer instead of improvising a new one from scratch every time.

Note that this blog post assumes that you have either watched the video or read the associated paper.  If you haven't, what follows will probably not make a lot of sense.

The question I keep getting is some variation on the following theme: What is the relation of the QIT/zero-worlds interpretation of QM to interpretation X, where X is usually many-worlds, but is sometimes relative state.  Riffing off this I'll get questions about the implications of QIT for time-travel, the relationship of QM to consciousness, and whether or not we might be able to influence the results of quantum measurements with our minds.

The short version of the answer is: QIT/zero-worlds is nothing more than a different way of looking at the math than what is usually presented in the popular press.  It is a way of looking at the math that makes sense to me (and apparently, based on the feedback I get, makes sense to a lot of other people as well).  But that is all it is.  There is no breakthrough here (except, perhaps, a pedagogical one).  It turns out that all this stuff was actually known as early as the 1930s.  Why Feynman was still saying that no one understood quantum mechanics in the 1960s I do not know.  It is certainly not true today.  But the point is that despite the somewhat sensational rhetoric ("You don't really exist; you are living in a simulation running on a quantum computer") nothing really changes as a result of QIT except your perspective.  You are still every bit as real (or not) as you were before.  Time travel, ESP, and telekinesis are still every bit as impossible as they were before.

The other short version of the answer is that many-worlds/relative-state/whatever are all equally valid ways of looking at QM.  The only one that isn't equally valid is Copenhagen.  To be sure, Copenhagen is a reasonable approximation to the truth for many practical purposes, just as Newtonian mechanics is a reasonable approximation to the truth (which is, to the best of our current knowledge, general relativity) for many practical purposes.  But Copenhagen is conceptually wrong, just as Newtonian mechanics is conceptually wrong.  There is no "force of gravity" and the wave function never collapses.  The challenge is to explain why it appears to do so.  That is what QIT does (IMO).

Let's take a moment to review the problem that QIT (and other interpretations of QM) purport to solve: QM is one of the two most successful scientific theories ever (the other being GR).  No experiment has ever disagreed with a prediction made by QM.  However, the mathematics of QM seem to be fundamentally at odds with the apparent nature of reality.  The Shroedinger equation is continuous, deterministic, and time-reversible.  Moreover, it describes a world where objects can exist in superpositions of states, a phenomenon which can be experimentally demonstrated through interference experiments.  By way of contrast, the world appears to consist of material objects which at all times exist in some particular state and never in a superposition.  Moreover, the process of making a measurement appears to be discontinuous, non-reversible, and also involves some fundamental randomness which is nowhere to be found in the Shroedinger equation.  The apparent contradiction between the theory and the manifest nature of reality has historically been called the "measurement problem."

QIT solves the measurement problem by observing that you can describe measurement as a purely quantum process.  When you do this, the following facts emerge (and this is what the Google tech talk and associated paper are about):

1.  Measurement and entanglement are the same physical phenomenon.  Measurement is nothing more than the mutual entanglement of a large collection of particles (or, to be strictly correct, of systems that manifest themselves as particles under certain circumstances).

2.  Once two particles are entangled, it is not possible to "undo" that entanglement except by bringing the two particles physically together.  If there were any other way to "undo" an entanglement, then it would be possible to transmit information faster than light.

3.  The apparent randomness that results from a quantum measurement is just that: apparent.  In actual fact, the entropy of a system that has undergone a quantum measurement does not change.  The reason that there seems to be randomness is that when you draw a line between the particle being measured and the measurement apparatus, you end up with positive entropy (i.e. randomness) in the measurement apparatus and corresponding negative entropy in the particle being measured (which is possible because the state of the particle is a complex number).

4.  The reason that two measurements made on the same physical quantity produce the same result is not that the measurements are a faithful reflection of some underlying physical (or metaphysical) "element of reality" as Einstein put it.  Instead, if you look at the quantum mechanical description of two separate measurements on the same system what you end up with is a mathematical description that looks exactly the same as two classical systems in classical correlation with each other, but that says nothing about the actual state of the system being measured (except that it is now entangled with the measurement apparatus).

5.  The apparent non-reversibility of a measurement is likewise not fundamental, but merely practical.  Reversing a measurement is possible in principle, but to reverse a measurement, you have to reverse all of the entanglements that produced that measurement to begin with.  Reversing even a single entanglement is extremely difficult.  Reversing a macroscopic number of them (and you really do have to get them all, every single last one), while possible in principle, is not possible in practice.

In other words, there is no measurement problem.  All of the apparent contradictions between the mathematics of QM (continuous, deterministic, time-reversible) and measurement (discontinuous, random, non-reversible) can be understood purely in terms of quantum mechanics itself.  Furthermore, all of this (except possibly the bit about negative entropies) was known in the 1930s.  So why has QM been considered so intractably mysterious for so long?  Indeed, why is QM *still* considered by many to be intractably mysterious?

I don't really know, but I suspect it's because people don't want to accept what the math is telling them.  The math says, essentially, that you don't really exist (or, if you prefer, your existence is not unique -- it turns out these are two equivalent ways of saying the same somewhat ineffable thing).  This is not the first time this has happened.  The exact same kind of conceptual stumbling block delayed the discovery of relativity for decades.  The fact that Maxwell's equations predicted the existence of electromagnetic waves moving at a fixed velocity c was known in the mid-1800s.  But no one took this seriously until 1905, because it was just obvious that time and space are absolute and so there just had to be some fixed medium through which electromagnetic waves propagated and relative to which the predicted speed c was to be measured.

The similarly obvious (but nonetheless false) assumption that everyone gets hung up on today is that the universe is, in point of metaphysical fact, what it appears to be: the whole of creation, populated by material objects that exist in particular places at particular times.  The answer to the puzzle: how can such a universe arise from quantum mechanics is, quite simply: it doesn't.  It appears to, but this is an illusion.  To be sure, the illusion is quite compelling, but it is false.  It is every bit as false as the illusion that space and time are two distinct things (which can also, it should be reiterated, be quite compelling).

It is worth pointing out that the fact that the underlying truth is very different from what we naively perceive it to be is evident long before you get to quantum mechanics.  You think that the chair you are sitting on is a solid object, but in fact it is mostly (>>99%) empty space.  The reason is appears to be solid is that the electrons in the outer shells of the atoms that make up the chair repel the electrons in the outer shells of the atoms that make up your body (or your pants).  So even in a pre-Shroedinger world, things are very different than they appear.

OK, so atoms aren't solid, but they are still (in a post-Rutherford but pre-Shroedinger world) classical.  They exist at definite places at definite times.  It makes sense to distinguish this particular hydrogen atom that is part of a water molecule in your little finger from that hydrogen atom which is undergoing nuclear fusion in the core of the sun.  It is obvious that atoms are classical material objects.  We can even take pictures of them and move them around nowadays.  The evidence that atoms are classical is overwhelming.  How could it not be true?

Well, it's not true.  Not only is it commonplace nowadays to take pictures of atoms and move them around, it is also commonplace to do interference experiments with them.  And not just atoms, but enormous molecules have been observed to interfere.  And yet, it is obvious (and at this point that phrase should be ringing alarm bells in your head) that somewhere between a buckeyball and you there must be a line where the world really does become classical because it is obvious that you are classical.

Sorry to be the one to break this to you, but you're not.  The evidence that you are classical is indeed overwhelming, just as the evidence that space and time are absolute is overwhelming.  But in fact neither is true.  The reason you can take a picture of an atom is not that the atom is really there, but because in the process of taking the picture your camera becomes entangled with the atom.  Then, when you look at the picture, you become entangled with the camera.  The reason you think that there's an atom there is because you are a large system of mutually entangled particles, hence quantum mechanics predicts that any particular part of you will behave as if it were a classical system in classical correlation with every other part of you.  The net result is a system where every piece of it agrees that there is (or is not) an atom there.  And asking your fellow humans to corroborate your intuitions doesn't help, because they too are large systems of mutually entangled particles, and as soon as they look at the same picture you have looked at, they too become entangled with it and with you and with the original atom, and so every part of that system (you plus your collaborators) will agree that there was an atom there (or not).

So is the atom "really" there?

The problem with this question is that it seems like the answer should be either "yes" or "no", but this too is false.  The nature of this question is more like this one:

Was Darth Vader (or, if you prefer, Anakin Skywalker) "really" Luke's father?

One the one hand, it seems that the answer should be "yes" because, in the Star Wars universe, Anakin/Vader was Luke's father.  But, of course, the Star Wars universe is fictional, so what does it mean for a fictional character to "really" have any particular attribute?

The answer, IMO, is to simply observe that fictional characters like Luke Skywalker and Harry Potter are in a different "ontological category" from (classically) real things like George Lucas or J.K. Rowling.  Well, the quantum wave function is also in a different ontological category than classical reality.  Fiction "emerges" from (classical) reality in much the same way that classical reality "emerges" from the wave function.  (The reason I hedge with "much the same way" is that there is one important difference: fiction and classical reality can both be described as classical computational processes, i.e. the math involves only real numbers, whereas the quantum wave function can only be described with complex numbers.  So the process by which classical reality emerges from the wave function is mathematically different (but conceptually similar) from the process by which fiction emerges from classical reality.)

So is the atom "really" there?  Well, to you it is.  It is every bit as real as you yourself are, and for the exact same reason: because the atom is part of the system of mutually entangled particles of which you are a part.  (This is sometimes called the "relative state" interpretation of QM.)

But let's take a different example.  Instead of asking whether the atom is really "there" let us ask instead if one of its electrons is "really" spin-up or spin-down (or, equivalently, if some photon it emits is "really" polarized vertically or horizontally).  You measure it, and the result is spin-up.  Your friend measures the same electron and agrees, yep, it's spin-up.  So you and your friend have become mutually entangled with this electron and hence are behaving just like a pair of classically correlated classical systems, just as QM predicts.

But, while QM predicts that you will be classically correlated, it does NOT (and cannot) predict what the outcome of your measurements will actually be.  That can only be done probabilistically, which seems at odds with QM (which is, if you will recall, purely deterministic).  To understand this we have to dig a little deeper into the math.  I've hinted at this before when I said that in order to extract a description of the classical world from the wave function you have to "trace over certain degrees of freedom".  That is just a fancy way of saying, "discard some of the information about the system."  Consider the full QM description of a particle that has been measured.  Part of that description is the state of the particle, and the other part is the description of the measurement apparatus.  To extract the state of the measurement apparatus you "trace over" (i.e. discard) the parts of the description that describe the state of the particle being measured.  What you are left with is not one classical world, but two: one in which the measurement apparatus says spin-up, the other in which it says spin-down.  But (and this is the crucial point) in neither of these descriptions is the spin of the particle actually spin-up or spin-down.  It can't be.  There is no description of the state of the particle being measured, because we had to throw it out in order to extract a description of (something that looks like) a classical universe, and that actually turns out to be a description of two classical universes.  That is where the "multiple worlds" interpretation comes from.

So do these universes "really exist"?  Again, in my opinion that's like asking whether Darth Vader is "really" Luke's father.  Classical universes are what you get when you take the quantum wave function and throw out parts of it.  That is the mathematical fact.  You can interpret this mathematical fact however you choose, with one exception: you cannot reasonably conclude that the classical universe that you live in is "all there is" because a complete description of the (classical) state of the universe is only, and can only ever be, a partial description of the underlying quantum state.

So what about all those other universes?  Are they real?  Well, from the perspective of the quantum wave function, yes, they are.  A classical universe is just a "slice" of the wave function (i.e. the whole wave function with parts of it discarded) and the wave function doesn't care which way you slice.  It's rather like if someone wrote an alternate Star Wars universe where Darth Vader was not Luke's father.  The existence of such an alternate Star Wars universe would have no bearing on whether Darth Vader was Luke's father in the original Star Wars universe (the answer there would remain "yes") nor would it have any bearing on whether Darth Vader was Luke's father in the "real" universe in which both Star Wars universes were embedded (as fiction): the answer there would remain that the question is meaningless because mixing ontological categories makes no sense.

David Deutsch, for whom I have the utmost respect (I think he's actually one of the best popularizers of science ever) is a fierce proponent of the proposition that all classical universes are equally real.  I respectfully disagree with him.  It is true that they are all equally real from the perspective of the wave function.  But I don't have the perspective of the wave function, and neither do you.  You and I live in this universe, and so to us, this universe is more real than any of the other myriad universes that emerge from the wave function.  There may be a transporter in the Star Trek universe, but that doesn't help Luke Skywalker escape from Emperor Palpatine because Luke can only take advantage of (and hence only cares about) what exists in his universe.

What about the possibility of communicating between universes?  Wouldn't that be cool?  If those universes are "as real as we are", shouldn't that be possible?  Well, unfortunately, no, it's not.  The way in which classical universes emerge from the wave function makes communication between them impossible.  You can prove this mathematically, just as you can prove that quantum entanglement can't be used to send information faster than light.  This is another reason I believe that parallel universes can safely be regarded as less real than our own universe, at least by us.  But reasonable people can (and do) disagree.

There's a lot more to say about this topic, but this post has already become longer than I intended it to be.  I'll write more if there's interest, but I want to leave you with a parting thought (well, more of an exercise actually): remember that I said that measurements were in principle reversible.  Imagine that we could actually carry out this program of undoing the myriad entanglements that constitute your making a particular observation.  What would be the subjective sensation, i.e. what would it "feel like" if this were done to you?