“These are copies from bricks found in the ruins of Ur, the work of Bur-Sin of Ur, which while searching for the ground plan (of the temple) the Governor of Ur found, and I saw and wrote out for the marvel of the beholder.”
-Text of the first museum label, circa 2000BC
The Museum of Modern Art is full of tombstones. You may have cropped one out of an Instagram photo lately — they’re the small white signs that list the artwork’s vital statistics: title, artist, date, medium and provenance. Last week, MoMA quietly released their collections database, a vast graveyard full of tombstones, as a GitHub repository. Slightly more than 120,000 artworks are included in the .CSV release, all tightly arranged in rows and columns.
People will do the same things with this data set that people do with any other significant data release: they’ll analyze it and visualize it and regress it and cluster it and query it and process it. They’ll find patterns, publish blog posts with scatter charts and arrive at a set of hopefully illuminating conclusions.
But that is not all that this data is good for, to be computed upon and extrapolated from. To do only this would be a disservice to its content and its cultural history. This dataset is also meant to be built upon, to be atomized and reassembled. This data can be and should be terrain for exploration, forum for interrogation, and substrate for creation. There is prose and poetry and performance to be made from these rows and columns.
The first thing you’ll probably notice from MoMA’s data release is that it doesn’t include any images. We have information about the artworks, but not the artworks themselves: the tombstones, but not the bodies. This constrains the things that we can do with the data: we can’t, for example, find every image with a horse, or look for people wearing hats, or compare the usage of the color orange from the 1940s to the 1970s. We can’t train Deep Dream to present us with technicolor modern-artifications of our vacation cat photos.
This is limiting, yes, but it’s also liberating. By ignoring the things that we normally think about when we think about artwork — namely aesthetics — we’re able to look instead at the patterns that are present in the supporting data. Instead of paying attention to paintings and sculptures and films and design objects, we can focus on the artists, on time periods, on dimensions and materials. In these fringes we can see things that are often obscured when we’re staring at the artworks directly. These areas along the edges are fruitful places to search for traces of politics, of the institution, the art world, and society as a whole.
For example, if we ask what the most popular first name for an artist in the MoMA collection is, we find out that there are 200 artists named John. We could then ask for the first 30 most popular first names for artists in the collection: John, Robert, David, Paul, Richard, William, Peter, Charles, Michael, James, George, Jean, Hans, Thomas, Walter, Edward, Jan, Joseph, Martin, Mark, José, Louis, Frank, Otto, Max, Stephen, Jack, Henry, Henri, and Alfred. Notice a pattern?
This is a very straight forward request. We can make our questions (queries) more interesting through repetition. For example, we might write a query that will take any title of an artwork, and find another artwork that has a similar but longer title. If we do this over and over again, we can construct chains of titles around seeded themes. For example, girl:
“Young Girl, Back Turned”
Girl with a Mandolin (Fanny Tellier)
Interior with a Young Girl (Girl Reading)
“HEAD OF A GIRL, THREE QUARTERS TO LEFT”
“Head and Bust of a Woman, Three-Quarters to Left”
Head of a Sleeping Woman (Study for Nude with Drapery)
Boy Leading a Horse
Boy on a Blue Horse
Children on A Bar A Ranch
Two Children Are Threatened by a Nightingale
“(Two children on a burned-out pier, Coney Island)”
“Pier 18: Parked Island Barges on the Hudson, New York”
The San Francisco Fire
San Francisco After the Earthquake
“Boards and Thistles, San Francisco, California”
“Car Tracks and Telegraph Poles, San Francisco”
Here the database query becomes as much of a creative instrument as a reductive one; it gains the ability to compose as it selects. And as with real-world instruments, the timbre comes from a balance of precision and imperfection. This particular query is written to eventually lose attention — to wander from boy to children, from fire to San Francisco, as if in a daydream. Through these meanderings, artworks from disparate collections and time periods are brought together; we find commonality in the text where we might not have found it in physical form.
If we start to think of these queries as instruments, the next obvious question becomes about performance.
How can a database be performed?
It’s 1:32pm. A woman in a black dress leans against the edge of a doorway between rooms in MoMA’s second floor galleries. Swatches of rotating light and the ting-tang of a Gamelan orchestra from the installation behind her bleed past her, out into the room that she’s facing.
“Fuck Off,” she mutters.
A few faces in the crowd turn towards her, but most either didn’t hear, or pretended that they didn’t hear. The woman continues, undeterred.
“Where’s My Fucking Peanut?”
“Shut The Fuck Up.”
“I Shit Crystals for you, David.”
Despite this impressive string of obscenity, the the gallery goers’ attention is mostly directed towards the middle of the room, where a group of five people who have just burst into song.
Over the next forty minutes, this group of six performers will speak (and sing) in a strange language — every word they say will be taken verbatim from the collections database. And yet it will not come off as if they are listing a litany of titles; instead they will engage in complex patterns of call & response, performing a combination of carefully choreographed exchanges and loosely-defined scenes, often balanced at the edge of chaos and absurdity.
The script, some of it pre-written, and some of it algorithmically generated on the fly during the performance, was constructed by Ben Rubin, Mark Hansen and myself (The Office for Creative Research) in close collaboration with John Collins and Elevator Repair Service. We constructed a system of narrative boardwalks through the data; John and the performers at ERS learned how to walk the planks. The performances were a culmination of a long Artists Experiment residency that we undertook at MoMA, an undertaking which in many ways set the stage for this week’s public release of the data that served as the periodic table for our performance.
This release of open data by MoMA is by no means revolutionary. Two years ago the Tate Modern released its own collection on GitHub. The Rijksmuseum in Amsterdam has not only released its data (images and all), it has also built an API to allow anyone easy access to it. In 2013 the British Museum released 1,000,000 images under a Creative Commons license.
These data releases are exciting. They will be a boon to researchers and academics, and will provide the museums themselves an opportunity to discover new ways in which their information holdings can find life and utility. But they also offer a unique chance for artists to engage with cultural data, and to support artwork that offers critique of these institutions and of data systems more broadly.
By encouraging art-making with their collections data, museums also find themselves involved in a beautiful kind of recursion: they produce data which produces art which produces data, and on and on and on. As Mark wrote in his essay Data-Driven Aesthetics in 2013, “the tools and techniques behind new forms of measurement for use in the arts are also producing new forms of data from the arts.” In other words, we are making artwork that not only weaves through tombstones, but adds to their number.
The production of A Sort of Joy: Thousands of Exhausted Things was a long and meandering process which involved many, many people. Special thanks go to Paula Crown, to Sheetal Prajapati, Pablo Helguera and Wendy Woon at MoMA, and to Sarah Hughes.
The performers for A Sort of Joy were Kate Benson, Lindsay Hockaday , Mike Iveson, Vin Knight, Gavin Price & Ben Williams.
I’d like to particularly thank Ellery Royston, who spent many long nights and equally long days leading up to the performances building APIs, fixing bugs in iPad scripts, and generally having the patience of a saint. Thanks, Ellery.
We’ll be releasing source code for several scenes from A Sort of Joy on GitHub. Follow us on Twitter — @The_O_C_R for updates.