Back in January, in a post about unanswered questions in physics, I included the Many Worlds Interpretation of Quantum Mechanics (the MWI of QM). I wish I hadn’t. Including it, and a few other more metaphysical topics, took space away from the physical topics.
I did it because I’ve had notes for an MWI: Questions post for a long time, but shoehorning it in like that was a mistake. Ever since, I’ve wanted to return and give it the attention of a full post. I’m reminded about it constantly; the concept of “many worlds” has become such a part of our culture that I encounter it frequently in fiction and in fact (and in other blog posts).
Its appeal is based on a simplicity, but to me it doesn’t seem at all that simple.
The simplicity is that the idea hangs on a single grand hook — that reality is described by a single quantum mechanics wavefunction. The Schrödinger equation is usually named, but I find that’s actually shorthand for something more involved. Regardless, MWI rests on the assumption of a universal ontological wavefunction — whatever that actually is.
For me there are two immediate issues. Firstly, a general suspicion of grand hook ideas. To my mind, most of them — e.g. BUH, MUH, IIT — get it wrong. It seems rarely true that reality is that simple, that pure. Secondly, there is a lack of clarity about what MWI really says that makes it hard to discuss in detail. For me, MWI raises tons of “yeah, but” questions that I’ve never gotten satisfactory answers for.
FWIW, what keeps me gnawing this bone is a fascination with why people are drawn to the idea. I can’t find the attraction; it seems as a naked emperor to me. Is it the notion of other selves that’s so compelling? Multiverses, in general, do seem a popular idea.
Regardless, all valid (i.e. working) interpretations of quantum mechanics involve a bit of currently unsolvable metaphysics and are, thus, created approximately equal. There is no fact or logic that invalidates any of them. As with art, which of them seem plausible to us depends our taste. Or, in this case, our intuitions about reality.
Back in March, Scott Aaronson wrote an interesting post about QM interpretations, The Zen Anti-Interpretation of Quantum Mechanics. I think his main thesis is stated at the beginning of his third paragraph: “I hold that all interpretations of QM are just crutches that are better or worse at helping you along to the Zen realization that QM is what it is and doesn’t need an interpretation.”
His fourth paragraph made me feel extra Zen:
“Greg Kuperberg, one of the smartest people I know, once told me that the problem with the Many-Worlds Interpretation is not that it says anything wrong, but only that it’s “melodramatic” and “overwritten.” Greg is far along the Zen path, probably further than me.”
Because I completely agree with Greg (and with Scott).
Actually, to be honest, I’ve used the words “arrogant” and “defiant” — I perceive MWI to have cult-like overtones (which I find off-putting). Certainly the naked evangelism, and even occasional persecution complex, of people such as Sean Carroll doesn’t seat the idea well in my eyes.
In contrast, I view the Copenhagen Interpretation (the “Standard Interpretation”) as humble. It readily admits to important gaps in its understanding — particularly with regard to measurement. It knows it’s only an effective theory, not an ontological one. (And it knows it needs to reconcile with GR.)
Later in the post Scott writes:
“If you had to, you could call even me a “Many-Worlder,” but only in the following limited sense: that in fifteen years of teaching quantum information, my experience has consistently been that for most students, Everett’s crutch is the best one currently on the market. At any rate, it’s the one that’s the most like a straightforward picture of the equations, and the least like a wobbly tower of words that might collapse if you utter any wrong ones. Unlike Bohr, Everett will never make you feel stupid for asking the questions an inquisitive child would ask; he’ll simply give you answers that are as clear, logical, and internally consistent as they are metaphysically extravagant. That’s a start.”
(The “metaphysically extravagant” made me smile.) It’s possible I passed this student phase back in the 1990s. Until then I didn’t give it much thought and just took MWI on faith as a possibility. Scott’s Zen point is to move beyond taking any interpretation too seriously.
Ironically, despite that “shut up and calculate” is a phrase usually given to the CI, one could argue the MWI has it even worse. It defines a reality we can’t access and effectively posits a magical ontological solution to the open questions of QM. It says, yep, reality is very, very weird (extravagant!); just accept it and move on.
It makes me smile that MWI is said to be a realist (opposed to anti-realist) interpretation. The superposed existence of myriad physical realities with the one I perceive seems an unreal notion. (I don’t mean unreal in the colloquial sense of hard to believe, but in the sense of being an unphysical theory.)
Lastly, to the extent the Schrödinger equation is central to it, MWI is based on a single math equation from what everyone agrees is, at best, an incomplete theory. This seems particularly relevant given the equation doesn’t describe particle creation or destruction. We need QFT for that.
Two important definitions:
Reality: All possible worlds; the total wavefunction. Everything everywhere. If MWI is correct, reality contains multiple separate universes. If MWI is not correct, reality and the (single) universe are the same thing.
Universe or World: A single “branch” or “slice” of the total wavefunction, but a complete physical realm. Our consciousness seems to inhabit (at least) one of these. If MWI is correct, there are many of them. If it isn’t, there is only one.
A key MWI issue for me is summarized by the question: Were there always two cats?
I refer to Schrödinger’s infamous cat. When the box is opened, supposedly there are two cats, one alive, one “sleeping.” In fact, there is a huge number of possible cat states (more on that later), but for now we’ll focus on the canonical two.
It seems the Many-Worlds Interpretation needs to be interpreted, and I’ve seen it done two ways by proponents:
The first version seems the popular conception. In it, there is at first only one cat, and this cat branches or splits into two. The universe (or world), likewise, started off as a single branch and, at some point in the experiment, split into two, only one with a living cat.
In the second version there were always two cats, but their histories are 100% identical up to the point in the experiment where they diverge. That divergence is due to their respective entire environments not being 100% identical. Some tiny difference between them ultimately leads to the radioactive sample decaying or not.
With regard to the number of cats, it’s true there are living and dead states, but if MWI is taken at its word, there are many different versions of living cats and dead cats.
One bullet to bite about MWI is the extreme proliferation of worlds. A key question here involves what exactly constitutes a different world.
At one extreme, if every quantum event causes a “branch” (a problematic term per the above), then walking down the street wearing sunglasses creates zillions upon zillions of branches as photons pass, or don’t pass, the polarizing filter. But other than a photon passing or not, the worlds would be identical. (But could that photon ultimately lead to divergence in the future?)
At the other extreme, if only events that amplify to classical levels — only events that “matter” — cause branches, what constitutes mattering? Here we almost seem to have another version of the quantum/classical divide, some putative line between events that matter and those that don’t.
Regardless, both living and dead cats experience enough classical changes over time (the experiment is often said to last one hour) that there should be many versions of each. For one example, a cat that died immediately would be different from a cat that died just before the box is opened.
At the least, since the radioactive decay can occur at any time during the hour, the superposition includes all the cats that died when decay happened at that instant. If quantum events amplify to affecting how a living cat moves around, there would also be a superposition of living cats in different positions. (If not there would presumably be many more dead cats in the superposition than the single live one.)
Consider also that, in theory, the cat’s wavefunction (and the box’s and the lab’s) contains all the wildly improbable — but possible — things that could happen. Power fails in the lab, car crashes, asteroids, the cat dying of natural causes, et cetera damn near infinitum.
Even if one bites the bullet and accepts this proliferation of worlds, there is a question of how the universal wavefunction contains so much information. What structure can hold all possible worlds that could ever happen?
I’ve heard the proliferation of worlds defended by analogy to integers. A simple theory of integers gives us an infinite set of them at no apparent cost and with no apparent “splitting” issues.
The problem with the analogy is that the integers are abstract. The MWI generally denies the view is Tegmarkian (seeing itself as realist), so it’s not possible to appeal to abstract objects as a defense. MWI has to be a physical theory, and physical objects can’t be cloned willy-nilly.
The claim that the MWI is a parsimonious theory requires ignoring the consequences of the theory — this vast extravagant proliferation of worlds. I believe the consequences of a theory are significant, so I disagree MWI is as simple as its proponents claim.
In any event, the ontology of MWI and its universal wavefunction raises a lot of questions.
Getting back to the original question about how many cats, now the question is whether there were always a zillion cats — all identical up to the “branch” point — or was there just one cat to start and “new” cats were created at those points.
Which immediately raises a question of identity. If one cat becomes two, which is the original and which branched off? One can say both are, but one becoming two suggests at least one was created.
Worse, creating new cats (original or not) seems to violate energy conservation. Sean Carroll has referred to “thinning” of energy (thus biting this bullet), but it seems to me, due to the energy-mass relationship, if energy thinned (whatever that means), gravity would change.
Worse, worse, given the proliferation of zillions (if not infinite) cats, the energy “thinning” would be extreme even for such a small thing as this experiment. Imagine it on the scale of the universe over the age of the universe.
That leaves the idea of a universal wavefunction with a superposition of worlds, many identical until some divergence point. In this case each world always exists and never splits, so they conserve their own energy.
In the first version, the wavefunction describes the branch, such as it does when describing the probabilities of a particle to either reflect from, or tunnel through, a barrier. The wavefunction branches from describing one probable location to describing two (each with their own probability amplitude).
In the second version, the wavefunction is a superposition of two worlds that happen to match until some point in time. This requires that the two wavefunctions are not 100% identical but very close to it. Per the example of a tunneling particle, this version superposes two particle descriptions, one that reflects, one that tunnels.
Note the second version implies something like superdeterminism, as the wavefunction describing each world evolves deterministically on its own within the universal superposition. The first version is just as determined but at least our point of view might move among choices. Since the MWI doesn’t allow for wavefunction “collapse” (yet, actually, I think it has to), there is no quantum randomness except with regard to which world one is in.
I think the first version is ruled out due to energy conservation. I think both versions are ruled out due to the question of how matter physically coincides. (Spoiler: This is currently my key objection to the MWI. The usual response — “decoherence” — seems incoherent to me.) In both versions I question how and where the massive amount of information exists.
However this has gotten long, so I’ll pick up the questions of wavefunction ontology, information, and the physical coincidence of matter, next time.
A closing thought about the popularity of the MWI. It was well known, even to experts, that “Eskimos” had 50 words for snow. It wasn’t true. It was based on a mistake made by an early anthropologist. Others trusted and quoted the work, and it became self-perpetuating knowledge.
The same thing happens routinely on the internet. An incorrect answer, or misattributed quote, is cloned repeatedly until it becomes common knowledge that Einstein defined insanity as “doing the same thing repeatedly and expecting different results.” (No he didn’t, and it’s a silly definition of insanity.)
The point is, MWI may have been accepted by many on the premise that experts have validated it, but even those experts depend on the work of others, and there are other experts who disagree — some quite strongly and cogently.
It’s possible MWI is a useful notion, a metaphor or crutch, but it doesn’t seem built on a solid theoretical foundation beyond its basic claim. It doesn’t seem to hold up under logical scrutiny.
Further, its metaphysical nature seems to give it a tendency to polarize — fervent evangelists versus ardent disbelievers. For me, that alone makes it unattractive.
Stay singular, my friends! Go forth and spread beauty and light.