Tom Comitta’s fiction has always strayed from convention, but never more so than in “People’s Choice Literature: The Most Wanted and Unwanted Novels,” which includes two novels inspired by a poll on what Americans most – and least – want to read. Comitta bends genre and character conventions, essentially distorting all notions of what a novel is; the writing process incorporated a language learning model that predated ChatGPT. We spoke with Comitta ahead of a reading Monday in Cambridge on the choice to use this technology, on the inspirations behind “People’s Choice Literature” and their willingness to be a bad writer. This interview has been edited for length and clarity.
What was the origin for “People’s Choice Literature?”
In 2013, I heard the Dave Soldier EP “The People’s Choice Music,” where he used a very similar procedure to what I did. It was this beautiful celebration of the opposite of taste, but also the parts of culture that we forget, or we push to the side. It became my favorite piece of music for a while. It took almost 10 years until I realized it was time to write the book. I had to become a novelist to do it.
What is your relationship to the standards of writing – and good writing?
To get there, I need to talk about bad writing. I feel like that relationship begins with me as a young poet. I’m turning 40 this year; I dedicated my life to being a poet when I was 22, and I was about to go to an MFA program. I remember being on a hill in San Francisco, a month before that program began, and thinking “Who the heck is Tom Comitta?” in this world of so many great poets. At that point, I was just out of a small town in Pennsylvania and had never published anything. A month before that moment, I had seen an art exhibition of the retrospective work of Martin Kippenberger at the Moma in New York – and the story of his work goes that back in the ’70s, painting had been declared dead by the conceptual artists; Kippenberger is maybe in his 20s, and he and his friends are painters, and they’re like “What do we do? Painting is dead.” And they set out to create the worst paintings imaginable, combining subject matters, painting styles, textures, colors that you would never in your right mind think to do, and in doing that they ended up creating really honest, weird work that pushed against the idea of good art.
In the moment, I decided I would try to become the worst poet imaginable. It really opened up a lot of possibilities. I would write something and just think about what I had been told as the right way to do it and then try the opposite, or just really go out and push against what I thought a poem should be – and, like the Kippenberger thing, I felt like I had started to write the most honest work I had ever written, and there’s a freedom, and a kind of anarchy, and structure. I’ve never had writer’s block since then, because I can always lean back on this process. The “most unwanted” novel, in a way, is like a culmination of all those years of trying to become the worst writer in America, and I wrote the worst novel in America. Although, if you read it, people say it’s beautiful, some people say it’s moving. It’s a celebration of the shadow of the culture, the things that are not on the bestseller shelves or are our first thought when we think of great literature, but there’s also beauty there.
Why incorporate LLMs in the novel? What is your opinion on the current generative AI conversation?
When I set out to write this, it was a year before ChatGPT came out. The thinking that led into this book was before everybody had an opinion. Essentially, the thinking was – I’m just one person, working with the tastes of an entire country. I could have sought out a collaborator, but I found a precursor to ChatGPT that was a completely different interface, and I guess I just thought as I was going through the data and starting to write this, I wanted something else that would challenge my subjectivity beyond the constraints. What I think most about since collaborating with a large language model is how the whole AI movement took a wrong turn when ChatGPT was released to the world in 2023. The interface I used in 2021 was called the Playground (also made by OpenAI) and was collaborative – a text box where I could work with the LLM as if we were a team thinking and growing ideas together. The model that has become common is of course the chatbot, which turns the LLM into an assistant, a disembodied minion doing our bidding. There was a moment half a decade ago when OpenAI was deciding whether it should stay a nonprofit or become a private company. We are living in the world of the latter, one where these tools are sculpted not in the interest of curiosity and mindful technological growth, but in the interest of profit. In this world, these models are training us to be lazy not just in how we acquire information – do we all really need subservient assistants? – but in how we think.
What is the value in subverting standard conventions of writing?
So easily in culture we get into matters and rhythms, and I just have learned from the tradition I come out of that it’s art’s job to keep these things moving, and make life and culture vibrant in that way. People call these works experimental – I just call them novels, because I think that for so many of us, it’s difficult to get published, it’s difficult to find an audience, so we might gravitate toward more comfortable ways of doing it. I find myself really interested in celebrating and exploring the many ways that we can read and write and keep things moving.
Tom Comitta reads from “People’s Choice Literature: The Most Wanted and Unwanted Novels” and “Patchwork” at 7 p.m. Sept. 29 at Harvard Book Store, 1256 Massachusetts Ave., Harvard Square, Cambridge. Free.
A stronger
Please consider making a financial contribution to maintain, expand and improve Cambridge Day.
We are now a 501(c)3 nonprofit and all donations are tax deductible.
Please consider a recurring contribution.