The laws of physics adapt to fit the observations, so that’s not the problem. The problem is with the traditional notion of causality. What happens at one location instantaneously influences what happens at another location. But the traditional notion of causality insists that an event can only directly influence what happens in its immediate vicinity. An event in one location is only supposed to be able to influence another location indirectly, through a chain of cause and effect that stretches from the first location to the second. In classical causality, this chain of cause and effect can involve particles or forces or whatever, but the crucial factor is that all direct (i.e., instantaneous) causation is local.
But in the entanglement scenarios, an event at one location immediately determines what happens at another location. (Or, more accurately, it determines what happens with what probability.) Since the influence is instantaneous, there is no time for intermediate causes to propagate between them.
So, at the very least, we need to throw out the idea that direct causation must be local. In fact, as alluded to by Chief Pedant, more radical adjustments to our ontology may be in order. But the point is that even the most conservative approach cannot retain classical causality.
There are problems beyond just a general intuitive hand-waving notion of causality. One of the fundamental postulates of relativity is that no information (or force) can propagate faster than c (otherwise known as the speed of light). So one electron instantaneously telling the other one how to spin from an arbitrary distance apart is hard to reconcile.
Now the typical entanglement scenario just squeaks through on a loophole, because, as far as we know now, there’s no way to send any real information from one location to the other, but it still causes some uneasiness.
As, Mr. MacAlister says, the laws of physics are what they are. I’d phrase it as s “it would be radically beyond the bounds of the laws of physics as we currently understand them to have action at a distance.”
Checking out EPR Paradox on Wikipedia leads to “Interpretations of quantum mechanics” and halfway down-- section 5 --I see no less than fifteen separate interpretations!!!
And those are just the common ones; I would hazard that there are dozens, is not hundreds (not counting dorm room bull session “interpretations” which end up being some thin variation on Copenhagen, Consciousness, or Consistant Histories) of proposed interpretations…and not a one of them is any more readily substantiated than another. Since we can’t observe a “quantum process” without affecting it, we can’t distinguish between the interpretation, which is why they’re call that instead of hypotheses or theories. There is only one [accepted] theory of quantum mechanics as a whole (I’m going to accord Randy Mills’ “Classical Quantum Mechanics” and other such theories the interest they are due, which is to say none) but a thundering herd of interpretations which all add up to exactly the same result; namely that we can’t know what’s going to happen before we observe it, but the resultant observation will fit right inside the predicted statistical distribution. Physics students who get too out of line in questioning what it all means are told to “Shut up and calculate!” which is the most feasible interpretation of all. Whatever is really happening on that level, it’s probably much weirder than anything we can construct in analogy to our own experience.
It’s even more problematic than that. A strict reading of Special Relativity doesn’t even allow events to exist non-causally. The universe, as far as an observer is concerned, is bounded only by the events he can actually see or interact with. Anything outside the “light-cone” of that observer’s frame of reference doesn’t happen, at least as far as the model is concerned. To allow “instantaneous” connections along a space-like path that goes outside the observable universe is, well, not exactly contraditory to the theory, but lies outside of it. You can get around this by constructing a manifold of multiply-linked domains (i.e. via “wormholes”) but to suggest that any given particle may have arbitrary causal links to other particles far distant means that the assuming of a smooth, generally simply connected plenum of space-time may be totally wrong.
General Relatively is even more convoluted on the issue; Chronos can speak to this in far more detail and precision, this being his area of study, but while the general assumption is that all points are connected only causally and limited in rate by c (with some anomolous but still entirely conformant anomolies around very massive rotating bodies) there are other formulations which may permit the influence of gravity to travel faster than or slower than c. We expect this is not true because of the implications this would have on causality and/or conservation of energy, but it’s not entirely ruled out by existing theory.
So the notion of “instantaneous” connection between distant points in space has massive implications for our understanding of physics, and would tend to indicate that relativity, as grand a theory as it is, is only an approximation. Then again, seeing as how we know nothing about the statistical nature of quantum mechanics beyond that it is, in fact, non-deterministic, and it fails to cope with one of the most important forces we deal with every day (gravity), it would be presumptuous at best to say that it proves anything wrong, especially a theory that has fit so very well to our observations as Special and General Relativity.
The Mad Tea Party was scarcely more frustrating than discussing what we know or don’t know about the fundamental laws of the universe. But it’s all great fun if you’re a mad hatter or march hare.