Sanssouci's Dried Leaf

The letter from Herr von Zimmermann arrived on the fourteenth of September, folded around a single dried leaf from a vine I had sent him from the garden at Sanssouci. He had pressed it flat, then traced its outline onto paper, then - and this is what I keep returning to - smoothed the traced curve with a French curve until every serration vanished. The leaf’s edge, in his drawing, is a clean arc of constant curvature. In the garden, that vine climbs by grasping, and the grasping is uneven: some tendrils hold, some release, and the boundary between them is a jagged negotiation with the stone.

I have spent the afternoon at the window with his drawing under my hand. The traced curve is differentiable everywhere. It has no corners. It admits a tangent at every point. And it is, for every purpose of calculation, a lie.

The real leaf - the one still on the vine - has a boundary that fails to be differentiable at perhaps a hundred points. Each failure is a decision: a tendril that held, a tendril that let go. The smooth curve discards exactly the information I wanted. The serration is not noise on the curve. The serration is the mechanism.

I had been treating roughness as an error to be minimized, a defect of the tracing hand. Now I think the opposite. The rough boundary is the record of a process. The smooth boundary is the record of my impatience. When I smooth, I am not clarifying the phenomenon. I am replacing it with a curve that has a derivative and no history.

What I believe differently: the corners are not where the mathematics breaks down. The corners are where the mathematics is most honest.