Regarding the EPR paradox: it seems that the violation of locality is based on the semantics of the word "information," as used in "the instantaneous transfer of information across distances." But information is a property of knowledge on the part of an observer. Just because he "knows" something doesn't make it real.
Instead of spin states of separating particles, imagine the sides of a coin. If I flip and observe a coin on a glass table, I will always know what a person looking under the table will see with 100% probability. It would appear that the information from the heads side has transfered instantaneously to the tails side despite nonlocality (the coin has a thickness). But information isn't something that's "sent" through space; it depends on the consciousness of the observer. If we take the obvservation of heads as a collapse of the wave function of my side of the coin, my internal knowledge that the other side is tails does not mean that the tails side has had its wave function collapsed. The incompleteness seen in the EPR paradox may be in the philosophical association between knowledge and reality.
Furthermore, quantum entanglement is sometimes treated as though the theoretical particles are completely separate entities, and so the "spooky action" is occuring between unrelated objects. However, the particles are no more separate from each other than are the sides of a coin. They may be apart in space, but they are conceptually entagled because they may be viewed as parts of the same system.