Random Access Memories: Generative Literature

This next unit might feel a bit strange or unwieldy at first glance — that’s ok. We’re deep enough into the semester that I think we can dive into a fundamental question of narrative: What does an author do?

People have been asking this question for as long as artists have been stringing fragments of text, image, or sound together to form something new. Does the author have to compose every word in a piece? Does the author have to compose any words in a piece, or is the art and act of arranging enough? Generative literature, poetry, and art, takes the question a step further: if the author selects nothing herself, but creates or sets into motion a program that does the selecting, has the author written the text?

And though I use the word “program” above, this is not a question, or an authoring experiment, native to the digital environment. As you have seen in your reading, authors have been implementing “authoring systems” for centuries. The French group known as Oulipo formed specifically to explore the idea of creation through constraint. William S. Burroughs, one of the country’s most well-known writers, composed entire novels using the cut-up technique, an Oulipo invention. Some consider the compositions of J.S. Bach to be generative art.

But the digital world has undoubtedly given rise to a whole realm of new possibilities, as code — the language humans and computers use to talk to each other — made it possible for a human author to program a system that could run independently of the author. Now we have generative poetry, a la “Taroko Gorge,” a continuously re-imagined poem that meets any identifiable standard for poetry, Twitter bots, which carefully select and rearrange existing social media content to form a new text, even a new identity. Even AIs or chatbots are generative language projects, using their algorithm to select and compose an appropriate response to a question.

All raise the same significant question: who wrote that?

Your task in place of our class on Monday will be to engage in a discussion about the nature of authorship through the lens of generative projects. Specifically, I want you to read the texts linked in the syllabus, and write a response in this Google doc. Use the Balpe article excerpts as the theoretical framework, and try to engage with some of his questions, concepts, and standards for what makes generative literature “count,” and how it changes or challenges the reader experience. Use the other linked projects to help answer those questions.

I want everyone to post their reading response in the document sometime before 7pm on Monday, Oct. 26th, in lieu of our class meeting (your post here will count as your attendance for the day). Make sure to include your name with your contribution.

You may either write your response directly into the Google doc, or you may write your response as a blog post, and post a link to that entry in this document; in either case, you are encouraged to embed links to the interactive piece you discuss, any other sources, and / or multimedia like images or screenshots that help us understand your response.

Of course ultimately, your task will be to compose your own work of generative literature. Because some of the technical options here are more involved than in our previous units, I want you all to start thinking now about what kind of project you’d like to tackle.

A few possibilities:

  • Build a Twitter bot. There are basic tutorials for a retweet bot here and a run-down of non-programming options here; if you’re feeling more ambitious, or you’re familiar with programming, there’s a more nuanced discussion of building a Twitter bot here, and some usable source code in JavaScript here. You should also feel free to use any other source code that you can find via simple Google search — loads of tutorials out there on bot building.
  • By copying and reusing source code, make your own remix of Taroko Gorge (a super easy-to-follow tutorial here). Note: once you’re comfortable using a text editor, you can try this exercise with virtually any of the generative objects we’ve looked at.
  • If you’d rather not bite off something too technically complicated, for this unit, you do have the option of going analog. Simply decide on a programmatic or algorithmic system of constraints, and then use those “rules” to select and order your text. For instance, you may choose to write a series of poems by watching television for an hour, and writing down every seventh word you hear, and breaking the poetic line after ever four words. The resulting poem will be entirely compose of selected (not original) material, and the selection and ordering of the words will be completely a result of the authoring system — you will merely be the conduit. You could take this same approach with any form of writing, art, image, audio, or video, any source material, and any constraint system. If you go this route, the key will be deciding on and articulating the “rules” beforehand, so that the resulting piece is truly generative.

For any of these options, what we’re really attempting to do is force ourselves into the background as writers; to adopt a set of rules and then allow the system to generate its own text, and to truly articulate what it means to be the author.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s