Deakin University
Browse

Machine Voice: Programmer Fiction

composition
posted on 2023-07-28, 05:49 authored by Andrew DeanAndrew Dean
In the early 1960s, a South African computer programmer experimented with one of the world’s most powerful supercomputers, the Atlas 2, to produce machine-written poetry. He was a keen reader of modernist literature, and was completing a Master’s degree on the work of Ford Madox Ford around the same time. Some of the work was published in the programmer’s native South Africa, including in an anti-Apartheid magazine. The experiments were ultimately short-lived, though. The programmer soon moved on to other pursuits, his exit hastened by concerns he had about the use of these Atlas computers for atomic weapons research. This programmer I am describing is J. M. Coetzee, who would become one of the most significant writers of the second half of the twentieth century. As scholars such as Rebecca Roach have described, Coetzee’s interaction with the early era of natural language processing influenced both his work and his conception of what it meant to write at all. Well after he had left computational research, Coetzee was still describing writing as a battle against language’s capacity to produce itself, meaninglessly. As he puts it in Doubling the Point: ‘Writing […] involves an interplay between the push into the future that takes you to the blank page in the first place, and a resistance. Part of that resistance is psychic, but part is also an automatism built into language: the tendency of words to call up other words, to fall into patterns that keep propagating themselves.’ This ‘automatism’, as he describes it, is recognizable to us today as the process by which generative AI produces language. It is in fact how large language models create sentences that feel as though they are natural. The question for Coetzee the creative writer, as opposed to Coetzee the programmer, was how to avoid the tendency of language to become derivative, and instead how to find a voice that was strikingly original. Across his notebooks and manuscripts held at the Harry Ransom Center, he insistently asks how his fiction will realize itself. Seeking to refuse the voice of the other – the voice of computation, the voice of apartheid, the voice of language itself – he composed for decades using pen and paper. He even created his own system of longhand composition and notation. As large language models put the concept of originality under significant pressure, I demonstrate how Coetzee’s struggle to find his own literary voice offers writers a way through contemporary debates about artificial intelligence. In the age of mechanized language, I suggest that it is still possible – in fact, ever more necessary – to speak for ourselves, but through creative and critical engagements with the technology in which we are embedded. This is a significant departure from current conversations about artificial intelligence, which tend to pit generative AI against an apparently non-technological history of writing.

History

Volume

https://lareviewofbooks.org/article/machine-voice-programmer-fiction/

Language

English

Notes

Link: https://lareviewofbooks.org/article/machine-voice-programmer-fiction/

Research statement

Background How have creative writers been enabled by computational methods? How has the history of creativity and computation been interwoven? The essay historicizes and develops ongoing debates about the role of artificial intelligence in human creativity. Current discussions have focused primarily on issues such as: problems of writing pedagogy in AI-enabled environments; the apparent ‘runaway’ of computer intelligence; and problems of copyright. I instead engage with the work of scholars of Digital Humanities, including Matthew Kirschenbaum and Rebeca Roach, who have focused on the intricate relationships between technology and writing in the 20th century. Contribution The 3,000 word article shows how one of the most significant writers of the 20th century, J. M. Coetzee, made creative use of computation in his early career to enable his work. By developing his own poetry programs in the 1960s, he imagined that he might escape the cage of language itself – which is akin to the neural network. My approach is a major development on contemporary treatments of artificial intelligence in that it shows how creativity can be and has been enabled by computation. I extend this thinking into the nature of creativity itself, which I argue requires turning back against existing paradigms to articulate new possibilities. Significance The article is part of my ongoing work on artificial intelligence and its consequences for literary thought and literary studies. It is significant in that it shows anew how creativity and computation have long been entwined for influential writers in the 20th century. The essay is featured in the Los Angeles Review of Books, one of the world’s premiere literary journals. This prestige outlet helps to direct thought in literary studies, both inside and outside the academy. It is read by more than 500,000 unique readers each month and is widely recognized alongside other major outlets such as the New York Review of Books and the London Review of Books.

Publication classification

JO3 Original Creative Works – Textual Work

Scale

NTRO Minor

Publisher

Los Angeles Review of Books

Place of publication

Los Angeles, Calif.

Source

Los Angeles Review of Books

Usage metrics

    Research Publications

    Categories

    No categories selected

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC