“Seismic,” “fantastic,” “the most important result in the foundations of quantum mechanics that he has seen in his 15-year professional career.”
They are referring to a publication by Pusey, Barret, and Rudolph.
The pull quotes above are from three physicists cited in Quantum theorem shakes foundations | Nature. The preprint is at System Unavailable.
Comments?
If I understand correctly, this paper hasn’t been accepted for publication and thus isn’t peer reviewed. It’s merely been posted to the online pre-publication archive for physics papers. So this hasn’t even reached the first step to being generally accepted by physicists.
Gotcha.
Is the supplied claim cold fusion baloney? I’m guessing it is.
I was surprised that none of the other physics heads didn’t just log-in in for a yes or no answer.
My first bump and with luck my last.
I think the paper is OK, but the result is being mis- to over-interpreted. First of all, they don’t show that ‘the wave function is real’ – it’s always possible to interpret it as a mere calculational device, in an anti-realist or instrumentalist sense; they acknowledge that and say that this stance is not what they are arguing against in the introduction.
What is being argued against is a position that one might call ‘Einsteinian realism’: that the wave function is a statistical description of some more fundamental ontic reality. They essentially show that if there is some underlying ontic state, then the ontic state must fix the wave function completely – but then, it’s really a property of that state, not a statistical description (which would be compatible with multiple distinct ontic states) of it.
Perhaps some background (which they unfortunately somewhat gloss over in their paper, which I think has lead to some confusion). There’s broadly two ways of understanding the wave function: psi-epistemicism and psi-ontology (pun intended). Psi-epistemicists say that the wave function is not a real thing, but rather a description of our knowledge; psi-ontologists on the other hand claim that the wave function is something concretely physical. Copenhagen or quantum Bayesian views are of the former, many worlds, objective collapse, or Bohmian mechanics of the latter kind.
Now, psi-epistemicism can again be divided into two categories, a realist and an anti-realist (or instrumentalist) version. The instrumentalist version is basically the Bohr tradition of regarding the wave function as merely some calculational tool, and perhaps of regarding science as a whole as merely concerned with what one can say of nature, not how nature is. This, I believe, one can always do (because whatever else the wave function might be, it is certainly also a computational device useful to make predictions), and it’s not what the paper argues against.
Rather, they argue against the view that the wave function is a statistical description of something more fundamental, i.e. realist psi-epistemicism or ‘Einsteinian realism’. To give something of the flavor of the argument, consider a classical ontic state – say, some particle’s position and velocity, which we would consider as giving a description of attributes the particle actually possesses (though there are some subtleties even here) – versus a classical epistemic state – a probability distribution, giving the probability for the particle having a certain position and velocity, which just describes our (imperfect) knowledge of the particle’s properties. This probability distribution (like all classical probability distributions) is a realist epistemic state: ‘underneath’ it, the particle has some definite ontic state, i.e. some definite position and velocity.
Now, in this situation, if we measured the particle’s position and velocity, we would get an answer that is compatible with several different probability distributions – it’s like if you throw a coin, and it lands heads, it’s compatible with the coin being fair, i.e. coming up heads and tails each with 50% probability, but it’s also compatible with a biased coin, coming up, say, heads with 60% probability; the one experiment does not distinguish between the two cases.
So they say, let’s consider an ontic state ‘underneath’ a wave function. Let’s furthermore prepare an experiment that measures two things in such a way that if the system we measure is in (epistemic!) state 1, measuring one of those things gives a certain value with 0% probability, and measuring the other thing gives a certain value with 0% probability if the system is in (again, epistemic) state 2 – i.e. the two probability distributions exclude one measurement outcome each.
But if both epistemic states correspond to the same ontic state, how can it distinguish between when it has to act as if it corresponds to the epistemic state 1, and when it has to act as if it corresponds to epistemic state 2? If it is in state 1, and we perform measurement 1, a certain outcome should never occur; but if it is in state 2, and we perform measurement 1, it may. But if both state 1 and state 2 correspond to the same underlying ontic state, then the system can’t tell if it is allowed to produce an outcome to measurement 1, or not! So, there can’t be any ontic state that is compatible with both epistemic state 1 and epistemic state 2.
Then, they go on to show that for any two states, one can find measurements leading to the same conundrum. But that means that no two supposedly epistemic states, no two wave functions, can have any ontic states in common – this is quite contrary to the case of classical epistemic states, where we have different probability distributions compatible with a certain ontic states, and hence, every probability distribution sharing all of its ontic states with other probability distributions. It’s in fact so contrary that those states should not be called ‘epistemic’ at all, as they are then a concrete feature of the ontic state – any given ontic state is associated with only one wave function, so the wave function is a property of the ontic state, and is thus – like velocity or position – ontic itself.
So it seems that if you want to be a realist, you can’t regard the wave function as epistemic. Of course, like any argument, theirs has assumptions, and there’s no doubt that one can construct a stance that one might want to call ‘realist psi-epistemicism’ which isn’t touched by their argument; but the most straightforward way, the way, arguably, Einstein would have preferred, has indeed been blocked by their argumentation. It’s a nice result, and I have no doubt that the paper (perhaps with some clarifications) will be published soon; the only problem is that the position they exclude is not really a position actually held by any researcher in the field, I believe – there are already lots of theorems pushing in a similar direction (Bell, Kochen-Specker, and some others) to have made this stance pretty hard to hold on to consistently.
But still, in principle, one might have tried to construct a hidden variables theory in which the wave function is epistemic – a contender to Bohmian mechanics, where it is ontic --, which is now pretty conclusively ruled out.
Tl;dr version: they show that, if you think science is about real things ‘out there’, as opposed to be merely a calculational recipe to explain outcomes of experiments, then you can’t assume that the quantum wave function is a statistical description of some more fundamental, underlying reality (the way, for instance, the probability distribution of a series of coin tosses is a statistical description of the underlying reality of actual tossed coins). Rather, you have to accept the wave function as something real, too (with some small caveats).
Personally, I do think this result is one of the more interesting ones in recent memory; it doesn’t really tell us anything we didn’t strongly suspect already, and many people probably assumed without reflecting on it – but it does shed some light on what, actually, quantum mechanics is (by saying what it isn’t, but that’s the way things go in science).
My apologies.
These have been the most thought out, and no doubt time-consuming replies I have ever had the pleasure of receiving.
I will chew on them and get back with more queries, if I can–I’m kind of scared now since the avalanche…
Half Man, is “Tl;dr version” an error for what…?
Thanks,
Leo
“Tl;dr” means “Too long; didn’t read.” It’s what some people write when they object to the length of a post. What Half Man Half Wit was saying that this was the condensed version of his previous post in case you thought it was too long to understand.
Thanks! However, I’d been thinking about the paper for a couple of days already, so it was really more a matter of just writing down my thoughts (which was very useful to me, too, in order to get them somewhat more organized). Plus, it’s kinda sorta my area of interest, anyway (I was in fact quite happy to see some talk about it on here; I’d thought of starting a thread, but figured the subject is maybe a bit arcane even for this crowd).
Yeah, I guess I have a habit of getting somewhat carried away when it comes to things that interest me… I’ll be happy to try and answer any follow up questions, though (promise I’ll try to keep to a readable length…).
Sliced bread wasn’t that big of a deal either. Leavened bread was the breakthrough; before then people where just eating crackers. 
Out of curiosity, what is the sort of viewpoint which was consistent with Bell, Kochen-Specker, and so on, but which is now ruled out by this new paper?
Half Man Half Wit, you deserve an award of some kind for that response. That was excellent.