ALGORITHMIC TREATMENT OF THE SPIN-ECHO EFFECT

We analyze the apparent increase in entropy in the course of the spin-echo effect using algorithmic information theory. We show that although the state of the spins quickly becomes algorithmically complex, then simple again during the echo, the overall complexity of spins together with the magnetic field grows slowly, as the logarithm of the elapsed time. This slow increase in complexity is reflected in an increased difficulty in taking advantage of the echo pulse. Our discussion illustrates the fundamental role of algorithmic information content in the formulation of statistical physics, including the second law of thermo-dynamics, from the viewpoint of the observer.