# What's going on between Alice and Bob?

1. Dec 3, 2007

### TimH

Hello to the forum (new user). I have a background in philosophy and have been reading Hughes' Structure and Interpretation of Quantum Mechanics. I have a question about entanglement.

When I first read the description of entanglement, in the above book, it did indeed seem like "spooky action at a distance." Now I'm a few chapters beyond that, and I've "lost the spookiness." Hughes has been arguing that quantum particles do not really have properties at all, that we can only talk about events.

My question is: Is there something undeniably spooky going on in entanglement, or is it simply spooky action-at-a-distance if you adopt a particular perspective (which I seem to have lost)? Or have I just forgotten how to see the spookiness?

Let me explain the Alice and Bob situation as I understand it. Then I'd appreciate someone commenting on my understanding. I'll use Bohm's example with electrons. So two electrons in the spin-singlet state rush off in opposite directions. Alice measures the z-axis spin on one electron and gets, say +z. She knows that Bob, at whatever distance he is, will get -z if he measures spin on the z-axis of his electron.

Now the entanglement comes when Bob measures a different, incompatible observable from the one Alice measures. Say Alice measures z again and finds z+. Bob instead measures x-axis spin and gets a random value. This is what we expect: the singlet state has "collapsed" into z+ and so x is equally likely to be x+ or x-. But if Alice measures x-spin instead of z, and Bob measures x, Bob will always get the opposite x-spin from Alice.

So when we compare the results we find that when Alice chose to measure z, the Bob x is random, and when Alice measures x, the Bob x is still random, but is also the opposite of Alice's.

Is this what happens? What is spooky about this? What am I missing? Thanks.

2. Dec 3, 2007

### Avodyne

Nothing is spooky about the situation you describe, because it is easy to find a classical model that would do the same thing.

The best explanation I have ever read of the spookiness was given by David Mermin in a Physics Today article 20 years ago. Luckily, there is a pirated copy online:

www.physics.iitm.ac.in/~arvind/ph350/mermin.pdf

3. Dec 3, 2007

### TimH

Thanks so much for the quick response and the link. I've copied the article and will look at it tomorrow. So are you saying that there IS something spooky, that I haven't noticed, or that the facts of entanglement only seem spooky if you think about quantum particles as having properties in a particular way?

4. Dec 3, 2007

### cesiumfrog

Surely you mean that only as a criticism of the OP's description, since it is generally not possible to explain entanglement with any non-spooky classical model.

5. Dec 3, 2007

### JesseM

What's spooky is that the statistics of Alice and Bob's measurements violate Bell's inequality. Here's an analogy I posted on another thread:

6. Dec 3, 2007

### Avodyne

I meant that perfect correlation when they measure the same property, and no correlation when they measure different properties (the case described by the OP) is not at all spooky.

The machine/card model that JesseM describes is spooky. This model was first discussed by Mermin in the paper that I cited.

7. Dec 4, 2007

### TimH

Thanks for the great post Jesse, that really helps me understand Bell's inqualities. It convinces me that spin is definitely weird.

My question now is: Must the measures at A and B be affecting each other to explain the violation of Bell's inequalities (action at a distance), or if we give up the idea of classical properties, can spin be conceived of as non-classical in a different sort of way. What I mean is, Hughes in his book (see OP) suggests that incompatible observables are incompatible because they are deeply related to each other. So spin values in different directions are all knit together in a way that is impossible to imagine. And this way they are knit together is revealed in entanglement experiments, without having to say that the experiments are affecting each other over the space between them. Is other words, is resorting to action-at-a-distance an attempt to keep some sort of individual identity to different spin axis measures?

8. Jan 27, 2008

### alphachapmtl

It is spooky when you consider subsets of measurement.
-- 1st case: when Alice measure z, Bob x is random.
consider the subset Alice.z+, then Bob.x is random : no problem here.
-- 2nd case: when Alice measure x, Bob x is random and opposite to Alice's x.
consider the subset Alice.x+, then Bob.x is not random : spookiness here.

9. Jan 27, 2008

### JesseM

That in itself is not spooky; it is quite possible for a local hidden variables theory to replicate the prediction that when Bob measures a different axis than Alice, he has a 50% chance of getting the same result and a 50% chance of getting a different one, but when he measures the same axis he has a 100% chance of getting a different one (if on each trial you just flip a coin to determine the hidden spin on each of the three axes of Bob's electron, and set the hidden spins on Alice's electron to be the opposite, then these will be the statistics you'll get after many trials). As I said in my post above, it's only when Bob's chances of getting a different result go below 1/3 in the case where he chooses a different axis that you get something that cannot be explained by a local hidden variables theory.

Last edited: Jan 27, 2008