I was surprised to discover I’ve never posted about the Many Worlds Interpretation (MWI) of quantum physics — I would have sworn I had. I’ve mentioned it a few times, and I know I’ve discussed it in comment sections, but it seems I never tackled the subject explicitly for the record.
It’s been on my mind lately because others have talked about it. Sean Carroll’s book promoting it generated a wave of discussion. The final push for me was Jim Baggott’s Farewell to Reality, which consigns MWI to the “fairy tale physics” heap.
Since I quite agree, this seems a good followup to yesterday’s post.
I’ve never found the Many Worlds Interpretation persuasive. What little argument there is for it doesn’t come close to convincing me, and the arguments against it seem substantial.
To be frank, my intuition rebels, and, while I appreciate that intuition is far from perfect, I need hard evidence or irrefutable argument to override it. (It might help to replace the emotionally freighted word “intuition” with “rational analysis informed by known facts”. MWI just doesn’t seem logically coherent to me.)
To be clear, this is metaphysics. There is no hard evidence about the nature of reality, so MWI is a story that seeks to explain certain observations of quantum behavior.
I’ll start with what MWI is and what attracts some to it (as I understand it).
Consider a simple experiment where a laser emits a single photon towards a half-silvered mirror. The photon has a 50% chance of being reflected and then hitting detector R (reflected) and a 50% chance of passing straight through and hitting detector T (through).
In the standard formulation of quantum physics, until the photon actually hits detector R or detector T, both possibilities co-exist. The photon passes through and is reflected. This is the same situation as in Schrödinger’s Cat box: the cat is both dead and alive.
Our experience, however, is that only one detector triggers. The cat is either dead or alive (and very angry).
The standard formulation (the Copenhagen Interpretation) has a wave-function — a mathematical construct called the Schrödinger equation — that describes the “both are true” path of the photon. It tells us, for any point along either path, what the probability is of finding the photon there.
In particular, at each detector (because of the half-silvered mirror), the equation says there is a 50% chance of that detector will detect the photon.
And, sure enough, if we do this experiment a bunch of times, we find that 50% of the time, detector T triggers, and the other 50% of the time detector R triggers. Exactly what we’d expect. It’s a half-silver mirror, so half the light goes through and half the light reflects.
It’s also true if we turn up the laser and shine zillions of photons at the mirror. Half of them reflect, and half of them pass through. (Both detectors would light up like crazy.)
This seems like a very reasonable picture, so what’s wrong with it? What does MWI bring to the table to solve?
The problem is that the Schrödinger equation follows the linear evolution of the photon, but it doesn’t have anything to say about which detector triggers. The actual measurement of the photon at one or the other is a discontinuity we can’t explain.
It also raises some serious questions about locality, but that’s another post.
The Measurement Problem — what happens when you open the cat box — is one of the great problems in modern physics. A measurement “collapses” the Schrödinger equation such that all the probabilities suddenly drop to 0% except at the point the particle is found, which jumps to 100%. We can’t account for how that happens.
Enter the Many Worlds Interpretation.
Which says that, when the photon hits the half-silver mirror, reality splits in two, and now there are two realities. In one, the photon passes through the mirror; in the other, the photon reflects. MWI posits that both these realities are equally real.
You can see where the Interpretation gets its name. The result of all this splitting is an unimaginably huge number of realities. According to MWI, there is almost an infinite number of each of us.
The attraction is that MWI is said to solve the Measurement Problem by saying the wave-function never collapses. The photon’s wave-function becomes part of the detector’s wave-function, and that expands to the system surrounding the detector including the scientist observing.
As wave-function collapse is seen as a bolted on extra in the Copenhagen Interpretation, the Many World Interpretation claims an Occam’s Razor approach of not multiplying entities.
This seems, as far as I can tell, to be the argument in its favor: It doesn’t multiply entities, and it takes the Schrödinger equation at its word.
I can’t find the quote, but I think it was John Wheeler who said something about MWI being cheap on entities but expensive on worlds.
That’s my first objection: I don’t see how MWI can claim parsimony when it so casually multiplies entire universes.
It’s not necessarily a branch to just two worlds. Imagine an experiment that uses a grid of photon detectors — think of each as a pixel. If this array is 200×100, then there are 20,000 detectors, each of which might detect the photon.
In such a case, reality must split into 20,000 new worlds. In each, a different detector triggers, and 20,000 copies of the same scientist go on to write 20,000 papers describing which detector triggered. Those 20,000 papers are identical except that each names a different detector.
My second objection is, if we take E=MC2 seriously, if we take matter and energy seriously, where does the energy for new universes come from?
How can we multiply universes at the drop of a photon? It took a Big Bang and 13.8 billion years to create the one we seem to inhabit. But unimaginable others are created instantly with no effort and no fireworks?
This one really feels like a show-stopper to me, a reason why this just can’t right.
I also have related questions about whether the new universe springs into existence in its entirety or somehow expands from the point of splitting like a crystal growing.
What about entanglement experiments where a measurement in one place affects a measurement in another place faster than light speed permits. The first measurement would seem to drag the other into a given branch.
Another aspect of this is that, in a two-slit experiment, where the particle appears to go through two doors at once, the branches of reality are allowed to interfere with each other.
The one-world view of this is that that particle’s wave nature — literally, its wave-function per the Schrödinger equation — causes constructive and destructive interference. When the wave-function collapses, the particle nature re-emerges in some random point per the probabilities.
In MWI the wave-functions of the two worlds interfere. Until they don’t, so there still seems to be some need to explain this “collapse” thing.
From our point of view, it still appears the wave-function collapses randomly. While there may be reality branches for all possibilities, we still find ourselves in a random one with no way to account for how that happened.
A lot of this involves the idea of decoherence, which is a topic all its own. It’s one I need to research a little in the context of MWI, but I get the impression decoherence never happens in MWI since the wave-function never collapses but merges with the surrounding wave-function.
It appears from our point of view that coherent particles do decohere, though, so I need to better understand what MWI says is going on there. (Knowledgeable comments welcome.)
Why do we not see any aspect of this? How is it that tiny quantum events have such a big effect in creating whole new realities?
A single photon hits a half-silvered mirror and suddenly there are two copies of reality? Do you generate trillions of new universes by walking down the street with a pair of mirror shades? Does the sunlight passing through, or bouncing off, your car window generate trillions of new worlds?
if so, how am I supposed to take the current branch seriously?
Another objection I have is exactly what MWI boasts about: taking the Schrödinger equation at its word. (By which they mean not bolting on that “collapse” thing.)
But in saying the evolution of the Schrödinger equation creates branches of reality, they are reifying a mathematical concept into, not just something real, but the only real thing.
This is essentially the view of Max Tegmark, that reality is math.
I don’t believe reality is just math. I think math is an abstraction we create based on regular patterns we observe. I think math describes those patterns.
I think MWI confuses a description for reality.
The last thing involves probability.
Let’s return to the photon and mirror experiment, but a difference: This time the mirror is silvered such that only 20% are reflected. The other 80% pass through. (We can do this the other way around, too.)
The scientist is going to send ten photons through, one at a time, and note which detector triggers. The expected result is that one detector will trigger 20% of the time and the other will trigger 80% of the time.
And this is exactly what the scientist observes.
Given that all branches always happen in MWI, how does the scientist find they are in a branch where the probabilities are as expected? Shouldn’t there be a branch in which the 20% detector fired all ten times? And one where it never fired?
Ten experiments with two outcomes gives us 1024 branches, so the odds of being in either of the two branches I just mentioned is about one-in-one-thousand. The odds of it never firing or always firing are one-in-five-hundred.
I must have this wrong, otherwise it would seem we could falsify MWI by performing this experiment and seeing what we get. Maybe with 20% odds and ten experiments, you really do get that detector always firing 1/1024th of the time.
While I can understand never with 20% odds and ten tries, always seems a stretch. MWI got ‘splaining to do about probability, is what I’m saying.
- Where does the energy come from?
- I dispute the parsimony claim.
- I don’t think reality is math.
- How does a quantum event affect all reality?
- What about probability?
So, for me, MWI doesn’t seem a likely theory. I definitely need some hard evidence (or much better arguments) to take it seriously.
Stay in one world, my friends!