Slashdot Mirror
Entanglement Could Be a Deterministic Phenomenon
KentuckyFC writes "Nobel prize-winning physicist Gerard 't Hooft has joined the likes of computer scientists Stephen Wolfram and Ed Fredkin in claiming that the universe can be accurately modeled by cellular automata. The novel aspect of 't Hooft's model is that it allows quantum mechanics and, in particular, the spooky action at a distance known as entanglement to be deterministic. The idea that quantum mechanics is fundamentally deterministic is known as hidden variable theory but has been widely discounted by physicists because numerous experiments have shown its predictions to be wrong. But 't Hooft says his cellular automaton model is a new class of hidden variable theory that falls outside the remit of previous tests. However, he readily admits that the new model has serious shortcomings — it lacks some of the basic symmetries that our universe enjoys, such as rotational symmetry. However, 't Hooft adds that he is working on modifications that will make the model more realistic (abstract)."
http://xkcd.com/505/
It's not really valid, though; it makes a false distinction between "a hidden variable" and "a hidden variable controlled by another hidden variable" as if they were different. Bell's theorem covers (or at least appears to cover) any additional information or state, regardless of the theory or process involved, provided that state is "attached" to the entangled particles (that is, it's local).
If "causality" as you use it is explicitly asymmetric, then yes, it's fairly straightforward to reject it. Typical arguments against backwards causation don't apply to these quantum measurements. Why? Because it's impossible to get between the particle and the future measurement. Any attempt to do so just becomes a measurement in itself. "Causality" as described by Bell just seems like simplistic philosophy. The very inequalities Bell derived should serve as a counterexample to this notion of "causality".
Firstly, I find the title of the submission a little odd. I mean, Entanglement can easily be understood as "deterministic" in a sense in conventional quantum mechanics. The generation of entanglement via the Schroedinger equation is quite deterministic. What's usually understood as non-deterministic is what happens when you measure.
I saw a talk by t'Hooft a number of years ago (I actually had lunch with him and my adviser). He was talking about a similar idea then, and my interpretation was that it evaded Bell's Theorem by being a non-local hidden variables theory. I haven't read the paper, so I'm not certain if this new idea is significantly different.
For background: Bell's Theorem is a result that shows that a local realistic hidden variables theory (a theory where each, say, particle has some hidden degree of freedom that determines the outcome of a measurement on it before the measurement is made) cannot reproduce the results of quantum mechanics for an entangled quantum state. To get around this obstacle, it's generally said that you either have to give up determinism (things don't have one specific state, etc. , before they're measured) or locality (the outcome of an experiment in one place may be totally changed by events happening at the same time arbitrarily far away)
"You call it a new way of thinking; I call it regression to ignorance!" -- Operation Ivy
Hidden variables in this case should be thought of as a hidden micro-states. A hidden variable theory would have quantum mechanics be something like thermodynamics; i.e., a theory that is not really basic, but appears so as we cannot see the fine scale true reality. Einstein was convinced that this had to be the case.
The tests of Bell's Theorem shows that no locally causal hidden variable theory is viable. This says basically that one of these must be the case
There are no hidden variables (i.e., true quantum uncertainty applies, and quantum mechanics is correct).
The speed of Light can be violated (i.e., there are hidden states that can exchange information faster than the speed of light). This implies, by the way, causality failures would be possible, so that in principle you could do something like kill your grandfather and prevent your own existence.
There is action at a distance (i.e., the theory is non-local).
There has long been a viable theory, that of Bohm, that replicates normal quantum mechanics. It's non-local.
I cannot tell from a read of the article (and without seeing the underlying paper) if 't Hoof has a non-local theory or just how he stays consistent with Bell's Theorem.
Particles are just as likely to be influenced by future interactions as they are by past interactions
This seems to be a poor understanding of time reversal symmetry. Particle physics works if you run time forward, or if you flip its sign and run time backwards. But that does not mean the same thing as what you said above. You can look at an experiment with each event in reverse, but you can't, for instance, say that event 2 was caused by event 1, but event 1 was caused by event 3. It only can follow the laws of physics if the causal order is 123 or 321.
The idea of 'backwards' causation has obvious major problems. First of all, you run into causal paradoxes. But more importantly, if the outcome of your experiment rests on future events, how can you do science? Every result becomes meaningless because you don't know if a future event caused it.
I came here for a good argument
Bells inequality rules out every possibility of the case:
result_of_experiment = me.some_function();
where some_function() has access to the entire history of me plus as much additional local information as you like (including internal variables) and it is deterministic.
There is a tiny "loophole" in that a truly rigourous test is extremely hard to do and not everybody agrees that the experiments done so far are 100% watertight.
Tim.
God said, "div D = rho, div B = 0, curl E = -@B/@t, curl H = J + @D/@t," and there was light.