Tuesday, September 01, 2015

Not so spooky

The impressive experiments described in a preprint by Ronald Hanson at Delft and colleagues have been widely reported (for example, here and here) as if to imply that they confirm quantum “spooky action at a distance” (in other words, entanglement). With all due respect to my excellent colleagues (who of course don’t write their own headlines), this is not true.

Einstein’s phrase is of course too nice to resist. But there’s a clue here. Einstein? You, know, the guy who wasn’t convinced by the Copenhagen Interpretation of quantum mechanics that reality is just what we can measure, and that nothing deeper lies “beneath”? Einstein, who suspected that there might be “hidden variables” that restore local properties to the quantum world?

Einstein’s “spooky action at a distance” was predicated on that view. It was action at a distance if, via this thing we call (that is, which Schrödinger called) entanglement, an influence at one location (via a measurement) is transmitted instantaneously to another. Only in some kind of local or hidden-variables view do you need to invoke that picture.

Quantum nonlocality – which is what is supported by a violation of Bell’s inequality, and what the new experiments now confirm by closing another of the loopholes that could have permitted a violation in other circumstances – is not spooky action at a distance, but the alternative to it. It says that we can’t always characterise the properties of a particle in ways local to that particle: its state is a smeared-out thing (to put it crudely) that may be correlated with the state of another distant particle. And so it appears to be. In this view, there is no action at a distance when we make a measurement on one particle – rather, there are nonlocal quantum correlations with the state of another. It is hard to find words for this. But they are not “spooky action at a distance.”

I don’t expect these words to make a blind bit of difference, but here they are anyway.


Stijn Oomes said...

The word you are looking for is causality. There is non-locality indeed, but no causal connection.

George Musser said...
This comment has been removed by the author.
George Musser said...

Well, yes, action at a distance would be deeply problematic, and the Delft, Micius, and other experiments demonstrate no such thing. But I think, first, that this post misrepresents Einstein. He specifically did NOT think that there was spukhafte Fernwirkung. Nobody in his day did. His point was that, to avoid such actions, you needed to adopt what we now call a hidden-variables interpretation of the theory. That is the second misrepresentation in this post. Spooky action is NOT an issue for HV theories, but for Copenhagen-type interpretations that introduce a collapse postulate. (Bohr himself did not subscribe to such a view, so we do need to be careful about what we call "Copenhagen".) Finally, one should not be so quick to dismiss action at a distance or causal connections when we have no other explanation for what produces these correlations -- and when, more broadly, we have no unproblematic interpretation of the theory. Our inability to use quantum correlations for signaling does not mean there is no causal connection at some deeper level.

Philip Ball said...

Hi George,
Thanks for your comments. I know you’ve thought about this carefully, so I certainly want to think about your remarks carefully too. But I can say a few things right away.

First, I’ve given you the wrong impression about what I’m saying about Einstein. When I say “Einstein’s spooky action at a distance”, I mean “the thing that Einstein spoke about”, not “the thing that Einstein believed in”. Of course he didn’t believe in it. That was the whole point: the fact that he regarded the EPR experiment as requiring it, in the Copenhagen view, was why he thought the Copenhagen view was wrong.

Second, I have somewhat too casually conflated hidden-variables theories and local realism. That’s because Einstein’s HV view was, I believe it is fair to say, a local realist one. Today folks are talking about nonlocal HVs, but I don’t think Einstein ever contemplated such a thing. And so my point is that Einstein was implicitly assuming local realism all along – which is why he figured that the Copenhagen view could only preserve the correlations by some kind of action from one particle to the other.

I don’t see why action at a distance is forced on Copenhagen (with or without collapse) – and neither did Bohr, of course. As I say, the standard Copenhagen view today (to the extent that one can speak about such a thing at all) invokes nonlocality – so you don’t need anything to “get” from one particle to the other. David Mermin said it very clearly: the mystery of the EPR experiment is that “it presents us with a set of correlations for which there is simply no explanation”.

You’re clearly unhappy with “no explanation”, which is why you’re not ruling out some sort of “hidden causation”. I guess this shows you’re a local realist at heart. But why shouldn’t the world be nonlocal at this level? After all, there are folks looking for simple physical axioms out of which quantum nonlocality will fall quite naturally. It seems entirely possible to me that they’ll find them. And don’t the Bell tests anyway make it impossible to sustain both locality and realism?

George Musser said...

Hi Phil,

Thanks so much for your reply -- it seems we are not as far apart in our views as I had thought, and I'd to talk to you more about this in person rather than through blog comments that are all too easy to misinterpret. I wasn’t actually referring to nonlocal hidden variables, and I accept your characterization of Einstein's views, with the addition that I think the record shows that Einstein's prior commitment was to relativistic causality and that he saw local realism as an *implication* of that.

I do think that the EPR and Bell arguments show that action at a distance is forced on a Copenhagen view. We are faced with a problem. We see these correlations. We are loth to invoke action at a distance. We cannot explain them via local hidden variables, as Bell showed. I, for one, think the job of physics is to provide materialist explanations rather than cede the field to mysticism. If we do not think spacetime is the proper venue for a mechanism, then we need to seek some new venue, which was, of course, my project in my book. Not that my own views really matter, but for the record, I'm decidedly not a local realist, because I think the categories of locality and spacetime will have to give way.

Philip Ball said...

Yes, we must talk about this some day George. It's certainly a puzzle how to talk about these correlations, but personally I don't see all versions of Copenhagen as mysticism opposed to materialism. One could of course argue, as Bohr did, that the purpose of physics/science is to provide predictions of what we'll see, and quantum mechanics does that. And even if there's action at a distance (though I see no need to invoke it), it's not clear how it can possibly be causal - I don't even know what superluminal causality could mean...