Much ado about mirrors.

Takano (1998) has suggested four different kinds of reversal to explain why mirrors reverse left and right and not up and down or back and front. In fact, mirrors perform only one kind of reversal: They simply reverse about their own planes, and reflection about one plane is equivalent to reflection about any other, plus a translocation and rotation. The reflection of an object is termed its enantiomorph. Perception of the enantiomorphic relation normally requires an act, either physical or mental, of alignment. In deciding whether two objects are enantiomorphs, there is a tendency to align them so that the reversal is about the axis of least asymmetry. But in deciding whether a single object is one of two possible enantiomorphic forms, people generally rotate it to some canonical orientation. In the case of objects with defined top-bottom, back-front, and left-right axes, the canonical orientation is determined by the top-bottom and back-front axes, leaving the left-right axis to carry the reversal. The main reason for this, I suggest, is that the top-bottom and back-front axes have functional priority, and the left-right axis cannot be defined until top-bottom and back-front are established. This means that the latter two axes have priority in establishing the canonical orientation. The left-right axis is usually, but not always, the axis of least asymmetry.