Whether it was really the alphabet itself that made the difference or—as is more likely—the separation of letters in print need not detain us further here. My concern is rather with a conundrum at the heart of Leroi-Gourhan's argument. Surely every trace left by a dextrous movement of the hand is itself a line. How, then, can the lines of prehistoric graphism be non-linear? How could it be that the storytellers of old, as they traced their lines, followed a non-linear trail? And how, conversely, can graphism be linear when, as in a sequence of printed letters, it leaves no trail to follow at all? In short, how can the line be non-linear and the non-line linear? Recall that in the evolution of the dotted line an original trace is broken into segments, each of which is then compressed into a point. It is in precisely this fragmentation and compression—in the reduction of the flowing movement of the ductus to a succession of moments—that the process of linearization consists. No wonder that the resulting line, as Leroi-Gourhan put it, is both needle-sharp and needle-thin! It is sharp because it goes to a point. And it is thin since it exists only as a virtual connector rather than a physical trace. Understood in a purely geometrical sense, it has length but no width at all. Fully linearized, the line is no longer the trace of a gesture but a chain of point-to-point connections. In these connections there is neither life nor movement. Linearization, in short, marks not the birth but the death of the line. In the next chapter we shall consider its ghostly spectre: the straight line of plane geometry.