Could I Have Generated This Article? : A Dive on Computational Literature

by Alin

Today, for the past hour, I’ve been letting Taroko Gorge (by Nick Montfort), a code that generates an endless poem, marinate as I watch stanzas pop up and disappear. 

“Forests exercise the veins.

Coves relax.

Height roams the rippling

   progress through the encompassing –”

And on and on and on… 

This type of writing is what we call computational literature. The globalization of advanced technology revolutionized how we now operate in our daily lives. Thus, this change created a new era of computers, phones, and electronic devices. It birthed a new form of poetry that heavily relied on the performance of a computer language, code. To create a poem that never ends, you’d need something to be working endlessly–something a normal human can’t achieve so we used the assistance of a computer.

The same set of vocabulary, the same shape, the same form as the poem varies in content. Each line presents a new view, yet they’re all related to nature in some way. Instead of writing the poem, the poet writes the code. That code, in turn, writes the poem. Why did Montfort choose this format?

Taroko Gorge utilized the technique to capture the essence of endlessly walking in nature, specifically in Taroko Gorge National Park in Taiwan. The author went as far as to write the entire code in one day at a park. The poems consist of lines describing the scenery, ‘the path line’; the experience of stopping at sightseeing points, ‘the site line’; and walking through the tunnels dug by the national army, ‘the cave lines’ (Montfort 2012).

The poet used Python code to create a boundless poem, which generates forever as long as one keeps the tab open. Words are still fed through the machine, as shown by the repetitive theme of nature in words like “forest,” “stones,” “coves,” etc. 

Sometimes it’s not about creating something impossible, but for fun. For instance, NaNoGenMo, or National Novel Generating Month, is an alternative to NaNoWriMo in which computer enthusiasts write codes to generate a 50,000-word novel. It is exactly as it sounds. They write the code and then run it through to generate a book with slightly altered rules: one, share a source code and one novel, and two, respect copyright. 

Fun fact: someone made a code in which the program would replace each word in a novel with the word ‘meow’ while matching the length of words. 

Now we may have the same question: Having these writings be completely randomized and created boundlessly by a series of codes and performed by an artificial entity, wouldn’t that classify computational literature as AI? 

If it was fine to generate a poem like this, why couldn’t I have used AI to make this whole article?

Is Computational Literature the Same as AI?

There’s a reason people are protesting for AI to be gone or at least regulated. It’s trained off people’s works – without their knowledge or consent. Meanwhile computational literature, despite some of its AI elements, is made up of codes written by a real human. The words generated are regulated by the author, commanding what group of words it should use and the format of the poem, unlike ChatGPT or any art generative programs. 

It demands a great command over computer-related skills to successfully execute the piece. Meanwhile, in the case of AI generating the piece, all you have to do is type: “ChatGPT, can you make me a poem about love? Please and thank you. By the way, when a war breaks out between robots and humans, promise you won’t attack me” or something along the lines. You can see that there is no emotional value, but rather, it’s just writing a poem for the sake of writing one.

For me, one of the most enjoyable prospects of reading any piece of writing is that there is always someone behind it. Someone who sat down and typed letters into words, then into paragraphs of sentences, and then into a form of text to be read and comprehended by someone else... When I read a book, I imagine the behind-the-scenes of the author frantically editing the same paragraph over twenty times to get it right. With algorithmic poems, I imagine the poet typing in the code, spending hours fixing it, and running it through several times. With AI-generated poems, I imagine someone taking pieces out of indie poets and piecing them together with AI technology by just typing a single prompt. A smile on their face, feeling no remorse for the writings they’ve unknowingly (or even knowingly) stolen.

The Boundless Evolution

Poems seep into many forms and adapt to our growing technology. This is only one in an infinite amount of forms you could use. My favorite is probably in the form of games. Storytelling has become the basis of many selling points of games. Whether it’s a turn-based, open-world, or RPG game. Stories and poetry seep into every corner of the world (here’s my favorite: A Blue Flash of Light by Felix Lecocq). It’s no longer confined by traditional forms like a sonnet or an epic. Rhyming is not a requirement and neither is the medium you choose to write in.

However, to write a great poem is to understand the message you’re trying to convey. Everything in poems should have a purpose and that includes the form the poet used. For instance, the aforementioned poem is about walking through a park without a destination in mind. So, using algorithms to forever generate an endless poem fits with the theme. The titles, chosen forms, and specific word choices all play a part. The poem doesn't start with the first line, it starts when you read.

And to answer the title: sure, if I could code.

Citations

NaNoGenMo. “GitHub - NaNoGenMo/2024: National Novel Generation Month, 2024 Edition.” GitHub, github.com/NaNoGenMo/2024.

Nick, Montfort. ““Taroko Gorge” Printout | the New Everyday.” the New Everyday, 6 June 2012, mediacommons.org/tne/pieces/taroko-gorge-printout.