What If We Live in a Superdeterministic Universe? (Script) (Patreon)

Today we’re going to try to save reality - or at least realism. However this rescue effort has a price; one that you may not be willing to pay. Your very soul, or at least your free will, is on the line.

On its most elementary scale the universe exists in a state of fundamental indeterminacy. Photons passing through two slits at once, electrons being spin up and down, cats being both alive and dead. These quantum superpositions only collapse into singular states when we try to measure them. It’s almost as though the physical properties of the objects don’t reside in the objects themselves, but emerge as a consequence of our observations. Some physicists and philosophers have wondered if reality is fundamentally subjective, whether it exists primarily as a consequence of the act of observation. Others - perhaps most - prefer to think of the universe as having a concrete existence independent of the observer. We call that view realism. There have been many efforts to find a realist interpretation of quantum mechanics. Pilot wave theory, objective collapse models, and even the Many Worlds interpretation all seek to describe a reality that can exist sans observers. But there’s one reality liferaft that we haven’t touched - and that’s superdeterminism. It may have the most disturbing implications of all.

Erwin Schrodinger, co-inventer of quantum mechanics with Werner Heisenberg, was a  big proponent of realism. In fact his alive-and-dead feline thought experiment was a reductio ad absurdum designed to highlight the ridiculousness of extrapolating observer-dependent indeterminacy to large or macroscopic objects. Albert Einstein also came up with a scenario in which he tried to refute the non-realist implications of pure quantum mechanics. Along with Boris Podolsky and Nathen Rosen, he proposed the EPR paradox which was meant to deal a swift death-blow to this observer-centric nonsense.

We’ve talked about the EPR paradox before, but it’s really worth a second look - especially given the stakes. In “standard” quantum mechanics, the fundamental building block of reality is the wavefunction, which describes the evolving probability distribution of all possible properties of the world, or of tiny parts of the world. The results of your measurements are plucked from the wavefunction of whatever you’re observing. We say that measurement “collapses” the wavefunction, obliterating all potential results in favor of one actual result.

A wavefunction can span multiple distinct, even contradictory states. For example an electron can be spinning in two different directions simultaneously. This is quantum superposition. And a wavefunction can also span multiple particles, holding information about the relationships between those particles. This is quantum entanglement. Let’s say we have a pair of electrons; each is in a superposition of spin states, say, with spin axis simultaneously up and down. But the electrons could also be entangled with each other so that their spins are guaranteed to be opposite to each other, even if they aren’t defined individually. Measurement of one electron would influence the other in a very real way - something Einstein referred to as spooky action at a distance. Standard quantum mechanics says that this is possible, but it leads to the seemingly absurd result of the EPR paradox.

Let’s see how it plays out. We prepare a pair of entangled electrons. Their spin direction is undefined; all we know is that they are opposite to each other. We give one electron to Alice and one to Bob who then go back to their separate labs to measure its spin.

According to standard quantum mechanics, that spin is undefined until measurement. Upon measurement it will become fixed along the axis of the measurement device. Say Alice uses a vertical orientation - she’ll measure either spin up or spin down, with a 50-50 chance of each. Say she measures spin up. But that means we now know that Bob’s electron must have spin down, even before he measures it. If Bob then chooses to measure using the vertical orientation he’ll definitely measure spin down. But he could choose a different orientation. A horizontal measurement will yield a 50-50 chance of left or right for his down-pointing electron. But any other choice will actually be influenced by Alice’s original choice of measuring orientation. When the scientists come back together to compare results, they’ll notice certain peculiar correlations between their choices of measurement axis and the other’s measured spin. And it doesn’t matter how far apart Alice and Bob’s labs are - across the campus or across the galaxy - the apparent influence should be instantaneous.

This type of effect is what made Einstein and co. so uncomfortable. A foundational axiom of relativity is that no causal influence can travel faster than the speed of light. The EPR paradox was meant to show that this sort of undefinedness couldn’t be the underlying reality because it violated the cosmic speed limit. He thought that each electron had to carry all the information about its own physical state, in a way that was somehow hidden from the wavefunction of standard quantum mechanics.

We can imagine the same scenario in the case where the electrons know their own spin all along. In that case the spins are set at the beginning - still opposite to each other, but defined, even if Alice and Bob don’t know what they are yet. The scientists go off and make their own choices for their measurement directions. But now nothing that Alice can do will change Bob’s results. When the two compare results, they will see correlations between the spin measurements because the original spin directions were correlated. But there’s no correlation between one scientist’s measurement results and the other scientist’s choice of measurement direction.

In 1960, nearly 30 years after the EPR paradox paper was published, the physicist John Bell gave us a concrete test for all of this. He derived a mathematical statement - the Bell inequality that is true in the case that the electron spins are set from the beginning and contained within the electron - but false if standard quantum mechanics is right and spin is undefined until observation. And it took another 20 years for the first Bell test to be successfully performed by Alain Aspect. That test thoroughly supported standard quantum mechanics. Countless Bell-type tests have been performed since, all with the same result: apparently entanglement is real and is as spooky as Einstein feared.

But actually, the results of the Bell tests are not convincing evidence against realism itself; the world may still have a non-subjective existence. According to the Bell theorem, any local-realist theory must obey the Bell inequality. So violation of the inequality could be a violation of realism OR locality. Or, sure, both.

In a local realist theory, every point in space and time - every event - can only be influenced by a spacetime event in its causal past - so, near enough to have sent a signal. There are ways to save realism in quantum mechanics by crossing Einstein and abandoning locality. Things like pilot wave theory and objective collapse models try to do that. On the other hand, Many Worlds interpretation of quantum mechanics saves local realism at the cost of requiring multiple realities. If that price is too high for you, there is another way. I’m not sure you’re going to like it any better.

Let’s play out the Bell test on something called a space-time diagram - with one dimension of space only on the x axis and time on the y. Alice and Bob start out together, acquire their entangled electrons, and then move sideways in space and up in time. Light travels a 45 degree angle, and so if local realism holds, only events in what we call the past light cone of Bob’s location should have been able to influence what happens at his position. But quantum mechanics says that when Alice measures her electron, Bob’s is instantly affected. That violates local realism.

Now let’s look at the case where the electrons start out with defined spins. It’s no surprise that Alice and Bob observe opposite spins because we can see that both electron’s spins were determined by a single event in both of their past light cones. Alice and Bob may observe correlated information if there’s a causal path to both of them.

But here’s were we find a loophole. The Bell test results are weird because Bob’s measurements turn out to be inexplicably correlated with Alice’s choices. But that’s only weird if Alice’s choices themselves are independent events defined wholly in the present moment.

Locality and realism would have been saved if it turned out that the states of the electrons were fixed when their past lightcones overlapped.  But what if, instead, the future decisions of Alice and Bob are influenced within that past overlap. In Bell’s theorem, his inequality is satisfied in local-realist theories - but there’s another condition to the validity of the theorem. That’s the statistical independence of the measurements. It assumes that Alice and Bob can choose with absolute freedom, or that if they choose randomly, that random number generator is perfect.

But what if statistical independence is an idealisation - trace past light cones back far enough and everything is connected. What if those connections produce sufficient correlations between Alice and Bob’s supposedly free decisions of how they make their measurements, and the states of the measured particles? It may still be possible to violate Bell’s inequality without saying anything about local realism.

This way out of the EPR paradox is called supedeterminism. Superdeterminism is, well, just determinism - the statement that the universe and any system therein evolves in a way that’s uniquely predicted by the laws of physics and its initial state. Determinism on its own doesn’t guarantee the sort of uncanny convergence in that evolution needed to solve the EPR paradox; that’s what the super part is for.

Some have argued that superdeterminism requires a kind of conspiracy of interactions to ensure two messy meat computers make the right choices based on marginal interactions arbitrarily far in the past. Others, like Sabine Hosenfelder, argue that superdeterminism is actually the cleanest solution to the EPR paradox. You can watch her video to decide for yourself. Regardless of its plausibility, superdeterminism is a loophole in the Bell theorem that physicists have tried to test. To close the loophole, a Bell test needs to be performed where the choice of measurement direction is truly statistically independent. That’s … impossible because if you trace the past lightcone of any two points in the observable universe back far enough they will overlap. Everything has a past causal connection with everything else.

But we can at least do a Bell test where the “deciders” and the measurement subjects have past light cones that intersect with each other as far in the past as possible. And so I’d like to introduce the cosmic Bell test. This was performed by Anton Zeilinger’s group at the University of Vienna in 2017. Their first effort used the light from a pair of distant stars as proxies for Alice and Bob. The “random” color of individual photons of that light was used in place of a random number generator to decide the measurement orientation in the Bell test. This experiment used the polarization direction of photons rather than the spin direction of electrons, but it’s the same deal. They observed violations of Bell’s inequality with extremely high significance, which they claim ensures no local-realist influences could have messed with Alice and Bob’s decisions except in the seemingly implausible scenario of stars talking to each other and conspiring against us. The same team followed up with a cosmic Bell test using rather more distant objects - quasars several billion light years away. That pushed back any possible local-realist influence more than half the age of the universe.

Theorists are still debating how to interpret superdeterminism, although the majority seem to be on Anton Zeilinger’s side in dismissing it. Sabine is an emphatic voice on the opposing side. Let’s stay neutral for the moment and just look at what we know. The many Bell tests, cosmic and regular, tell us that either locality or realism are wrong OR there are multiple realities OR the universe evolves in unalterable lock-step determinism in a way that preserves correlations between measurement subject and measurer. The latter has philosophical implications for our own dreams of being detached observers, independent of our subjects. But you know what? Quantum mechanics has been telling us that we are not that for 100 years. As for the question of free will - look, when philosophers come up with a reasonable definition for the thing maybe we can talk. But think about it this way - your mind has a physical substrate: your brain. Your mind isn’t an illusion just because it’s an emergent phenomenon. If your choices have a physical, even deterministic, substrate - that doesn’t change the fact that right now you know that you can choose whether or not to believe in a superdeterministic space time.