The best explanation I've found is when an object is rotationally symmetric, and we mistake 180° rotation for a 360° one. Since it's symmetric, they both look the same to our "gauge", but the underlying object really does have two different states. The paper suggests what form of confusion could result in entanglement looking behaviors, if we ever discovered such a thing. Essentially a new EPR loophole based on a phenomenon seen in many other subfields of physics