In a visionary paper in the 1960s, Herbert Simon suggested that emotions can be thought of as interrupt systems. This idea makes it in principle possible to ascribe emotions to artificial creatures, as long as we can assign to them a hierarchy of goals and the ability to suddenly switch from one form of goal-pursuit to another. One powerful source of resistance to the idea of robotic emotions is that many think of emotions as being essentially feelings, and doubt robots’ ability to feel anything. In recent times, a crop of new theories of emotions have emerged which suggest that we think of emotions not as feelings but as motivational mechanisms of a particular sort. I will discuss the opportunities and the challenges of this new research program on emotions, sketching a development of Simon’s pioneering ideas in light of what we have learned since the 1960s about the nature and function of emotions.