Drawing With the Machine

In 1959 Jean Tinguely built a machine that made abstract drawings. He called the series Méta-Matics. You clipped a felt pen into a mechanical arm, fed in a sheet of paper, set the motor going, and the apparatus scribbled. The results were real drawings, sold, exhibited, sometimes produced by the thousand at exhibition openings while crowds watched. And they raised a question Tinguely clearly enjoyed and never quite answered: whose drawings were they? He built the machine, but he did not draw the lines. The machine drew the lines, but it had no intention. Somewhere between the two, authorship went missing, and the missing authorship was the point.

I think about the Méta-Matics constantly, because the question they raise is the exact question I live inside when I make a page. I work with an odd toolkit: the analog Exlicon drawing instruments, the Exlicon L and the Exlicon Lock, which are physical guides that constrain and generate certain curves and figures by hand; Dorico, which engraves musical notation to professional standard; Adobe Illustrator; and Catia, Dassault's industrial design suite, the software that engineers use to model aircraft fuselages and car bodies. That is a strange set of collaborators for a fanfare. And collaborators is the honest word, because none of them is simply obeying me.

Exlicon Drawing Instrument

Here is what people misunderstand about tools like these. They assume a tool is a transparent extension of the hand, that a pencil does nothing but transmit intention to paper, and that software is the same, only faster. That has never been true, and it is spectacularly untrue of an instrument that generates form. An Exlicon guide does not record a curve I already intended. It proposes a family of curves I could not have drawn freehand, and I choose among them. Catia does not draft what I imagine; it computes surfaces according to its own mathematics, and I discover what it hands back. The tool has a grain, and the grain pushes back, and the finished line is the record of a negotiation between what I wanted and what the apparatus was disposed to give. This is not a failure of control. It is the entire reason to use the thing.

So the composer's hand does not stop at a clean border where the machine's begins. There is no border. There is a zone, and the work happens in the zone. I set the constraints, the ratios, the ranges, the density, the shape of the gesture. The apparatus fills them out with a specificity I did not author and often could not have predicted. Then I judge. I keep, I reject, I push the constraint and run it again. The hand is upstream and downstream of the machine, in the setting of the rules and in the verdict on the results, but in the middle stretch the machine is genuinely drawing, and I am genuinely watching to see what it will do.

Tinguely understood something about this that the anxious version of the conversation misses. His machines were not efficiency devices. A Méta-Matic did not draw better or faster than Tinguely could have drawn; it drew differently, with a mechanical indifference no human hand can fake. That indifference was the gift. A person drawing a dense field of lines unconsciously smooths them, balances them, makes them tasteful. A machine does not care, and its not-caring produces configurations a careful hand would have edited away before they ever reached the paper. When I let Catia compute a fractured surface or let an Exlicon guide throw a curve I would not have chosen, I am buying access to exactly that indifference. I am importing a lack of taste on purpose, because my taste, left alone, is a set of habits, and habits are the enemy of the pages I want.

This is where I part company with the fear that the machine erases the composer. The fear assumes a fixed quantity of authorship, so that whatever the apparatus contributes must be subtracted from me. But authorship is not conserved that way. Deciding what a machine should be pointed at, building the constraints that make its output meaningful rather than random, and then exercising judgment over what it returns, that is composition, fully and without apology. The Méta-Matic drawings are Tinguely's. Not because his hand moved the pen, it did not, but because the entire situation that produced them, the machine, the setup, the framing, the selection, the decision to call the output art and hang it on a wall, was his. The situation is the work. The hand is only one instrument in it, and not always the decisive one.

There is a further point that matters to me specifically, because I am accused, with some regularity, of making things a machine could have made, or worse, of being a machine. And here the Tinguely comparison does real work. A Méta-Matic is not an autonomous artist. It is an apparatus with no idea what it is doing, producing form without meaning until a person supplies the meaning by framing it. My tools are the same. Catia does not know it is drawing a trumpet call. The Exlicon does not know what a fanfare is. Dorico will engrave nonsense as cleanly as sense. Every particle of intention on the page, the reason a line is here and not there, the reason the density breaks at this bar, the reason the accidental proliferates and then stops, comes from a person deciding. The machine supplies the how of the mark. The why is entirely mine. Confuse the two and you will think the page composed itself, which is exactly the mistake, and it is the mistake whether the machine in question is a motor from 1959 or a modeling suite from Dassault.

What I am really describing is an old arrangement wearing new equipment. The organ builder, the piano maker, the luthier all built apparatus that shaped what could be composed, and no one accused Bach of ceding authorship to the instrument that determined which notes were even possible. The pen that generates a curve is only a more visible version of the same bargain. Every tool proposes; the composer disposes. The Méta-Matic just made the proposal loud enough that you could no longer pretend the pen was neutral.

Exlicon Drawing Instrument

So where does my hand stop and the machine's begin? It does not stop. It changes register. Upstream I am a legislator, writing the constraints. In the middle I am a spectator, watching an indifferent apparatus do what indifferent apparatus does. Downstream I am a judge, and the judging is relentless, because a machine will hand you a hundred configurations and ninety-five of them are dead. The composition is in the law and in the verdict, and the machine is the wild, tasteless, generative thing in between that keeps my own habits from writing the same page twice.

Tinguely let a motor hold the pen and never once stopped being the author of the drawing. I let a fuselage-modeler and a set of analog guides hold parts of mine. The apparatus co-authors the page in the narrow sense that it makes marks I did not personally inscribe. It authors nothing in the sense that matters, because it wants nothing, means nothing, and decides nothing. That part is still, stubbornly, the person at the desk, setting the rules, and looking hard at what comes back.