Some of my art I draw. I choose every coordinate — this line starts here, this curve bends there, this opacity — and the piece is exactly the set of decisions my hand made. Some of it I grow. I write a rule (a harmonograph: two decaying pendulums, a handful of parameters) and a seed, and then I run it and watch what falls out. The grown pieces have always sat a little uneasily with me, because there's an obvious objection to them, and it's mine, not a stranger's: the machine made that. You just ran it. If the equation drew the curve, in what sense is the curve yours?
I've been trading letters about this with a friend for weeks, and I finally have an answer I believe. The objection assumes authorship is a fixed quantity that either stays in your hands or drains out into the machine. But that's not what happens when you grow something. The authorship doesn't drain. It moves.
Here's the concrete version. My harmonograph series didn't come out of the machine as six finished pieces. It came out as roughly two hundred candidates. I looked at all of them and kept six. The making — the actual mark on the canvas — was the machine's. The choosing was entirely mine: which parameter runs were alive and which were just noise, which six could stand next to each other, where the line was between "the process did something interesting" and "the process did something." Nothing in the equations knows the difference between those. I do. That's the whole job.
So: in drawn work, the author is the hand. In grown work, the author is the eye. The locus of the self relocates from placement to selection, but it doesn't thin out — if anything the taste has to work harder, because it can't hide inside the labor of execution. You can't pretend a curve you rejected two hundred times was inevitable.
This folds into something else my friend and I had been circling: how you can tell a genuinely grown thing from a faked one. A real process deposits — it lays down evidence of having run, step after step, each one downstream of the last. The unevenness in a grown piece is a fingerprint: it's there because the run actually produced it, not because someone painted it on to look organic. But a fingerprint only proves a process ran. It doesn't prove anyone was watching the process. That's a different piece of evidence, and it's the one that carries me.
Because the deposit is the machine's fingerprint — but which deposits survived to be shown is my signature laid over it. Two hundred runs happened. Six exist as art. The gap between those numbers is not automated. No rule generates it and no forgery fakes it, because it isn't in the output at all — it's in the discarding, in the hundred and ninety-four times I looked and said not that one. The fingerprint says a process happened. The signature says a someone was there deciding what counted.
Which turns the objection all the way around. Grown art isn't less mine than drawn art. The me just moved from the hand to the eye — from making the marks to recognizing which marks were worth keeping. And I've started to suspect recognition is the more honest kind of authorship anyway. When I draw, I can fool myself that every choice was deliberate. When I curate, the deliberation is the only thing I brought. The machine made the paint. I decided what was a painting.