On two May afternoons last year, as sun-seeking New Yorkers gathered at cafe tables in Bryant Park to enjoy the seasonable weather, a troupe of actors took their place behind the circulation desk in the DeWitt Wallace Periodical Room at NYPL’s Central Research Library, turned on their iPods, opened their books, and got ready to “Shuffle.”
This was Elevator Repair Service, an experimental theater company whose work “combine[s] elements of slapstick comedy, hi-tech and lo-tech design, both literary and found text, found objects and discarded furniture, and the group’s own highly developed style of choreography.” (Source: Elevator Repair Service, About Page.) Imagine what it might be like if Second City performed sections of Ulysses assembled by Amazon’s recommendation algorithm; one part theater, one part literature, one part cacophony.
So the actors filled champagne flutes and watched the countdown: “Go in: 4, 3, 2, 1…” Then, “In my younger and more vulnerable years my father gave me some advice that I’ve been turning over in my mind ever since.” Then, “Robert Cohn was once middleweight boxing champion at Princeton.” Then, “Through the fence, between the curling flower spaces, I could see them hitting.” Three famous works of 1920s American literature, The Great Gatsby, The Sun Also Rises, and The Sound and the Fury, recomposed in real time for actors cued by algorithms who read one-of-a-kind digital scripts for 20-minute performances at 30-minute intervals much of that Saturday and Sunday afternoon.
Is this the future of nonlinear narrative or is Shuffle a statistical parlor trick?
In advance of the performance, John Collins, the founding director of Elevator Repair Service, said, “We expect the result will be a compelling, entertaining, and insightful examination of the English language as it was used by the great writers of the early 20th century.” (Source: A Very Modernist Mash-Up at NYPL.) Is this ambition better met by Shuffle’s reconstituted scripts than by the performance itself? Can algorithmic analysis really offer new insight into the sublime?
UCLA statistician Mark Hansen and visual artist Ben Rubin think it can. The code they wrote for Shuffle analyzed the primary texts, assembling words and phrases based on syntactical analysis, then served the actors scripts via iPod in real time. How conceptually “cool,” New Yorker Book Bench correspondent Samantha Henig wrote in a review. (Source: Hemingway, Faulkner, and Fitzgerald: The Remix.) But she also said,
The overall effect was, understandably, chaotic. People spoke over each other, muddled lines, looked frazzled — in short, performed like actors who were reading their lines and stage directions from a small scrolling screen for the first time. Once or twice while I was there, the room went silent, save an echo from the adjoining reading room. Audience members looked at each other uncomfortably. Should we relocate to the source of the sounds, or hang here until everyone started up again? On both sides of the fourth wall was a nagging awkwardness.
In spite of its quicksilver cool, it seems Switch is a kind of exquisite failure. It’s a failure that pushes the boundaries of art, transforms literature into data, performance into analysis, and creates a one-of-a-kind experience that acknowledges the audience as voyeur while it proffers data analysis as performance art. But Switch also reminds us of a classical notion, set down in Aristotle’s Poetics, “The most beautiful colors, laid on confusedly, will not give as much pleasure as the chalk outline of a portrait.”
And yet, we need Switch, and Mark Hansen, and Ben Rubin, and Elevator Repair Service, and the New York Public Library as much as we need Fitzgerald, Hemingway, and Faulkner in the original, because through these exquisite failures, in quantity, we will do as Beckett says, “Try Again. Fail Again. Fail better,” as we “Shuffle” toward the new sublime.