Refraction gives rise to a momentum change orthogonal to the propagation direction. This must result in an equal and opposite change to the medium at the boundary. Entry and exit cancel, and the medium experiences no net force. The problem is that the forces on the medium do not in general line up, so we have a pair of forces, opposite but parallel - i.e. a couple. Can we therefore conclude that refraction imparts a couple to a refracting medium? (and if so, has the experiment been done?).

Assuming this couple indeed exists: Since the light loses no energy and loses no momentum, and yet the medium appears to have gained rotational energy and momentum, where does the medium's extra angular momentum come from?