‘Can machines think?’

Alan Turing posed this question in his seminal 1950 paper ‘Computing Machinery and Intelligence’ that laid the foundations for research into artificial intelligence. Turing’s life and work provide the inspiration for Machinations, a poetic collaboration between Kinneson Lalor and JP Seabright published by Trickhouse Press. Fiercely intelligent, dazzlingly inventive and profoundly insightful, Machinations does justice not only to the depth, breadth and creative genius of Turing’s intellectual achievements but also to the complex layers of his personality.

I asked Kinneson and JP how the book came into being, their experience of working together and what informed their creative choices.

– What was the initial inspiration for this project, and how did your collaboration come about?

JP: Kinneson and I met up for the first time in early December 2021 to chat about writing and our various creative projects, including my recent collaboration with two other poets for GenderFux. She expressed an interest in doing a collaborative project, and a few weeks later I came across an indie poetry press that specifically wanted pamphlets from two writers ‘having a conversation with each other’. We decided to use this as motivation and our deadline – only seven weeks by the time we met again in early January to share ideas. 

I threw a load of ideas at Kinneson – random things I found interesting that I’d been collecting but not sure what to do with. Given her professional background and expertise in mathematics and computing, we decided to focus on Alan Turing, but in particular his ideas on Artificial Intelligence. Kinneson wrote the introduction to this book early on and that really helped us coalesce our ideas. The rest was mainly playing around with those ideas, and responding to each other’s pieces in turn. In the end, the work grew beyond that initial submissions callout, so we’re delighted that it’s found a perfect home at Trickhouse Press.

KL: Honestly, I read and loved JP’s Fragments from Before the Fall and was just eager to find out how to put together a collection and learn from someone whose work I admired. If JP had told me they wanted to write a collection about bananas and clowns, I would have totally gone with it. But this worked out much better!   

– In the Introduction to this collection you describe it as “a conversation about Alan Turing, the man who asked the question: ‘can machines think?’. But it’s also a conversation between two poets who see the world in very different ways. One of us is a mathematician. One of us is dyscalculic.” Tell us a little more about yourselves and how you came to poetry.

JP: I came to the conclusion not long ago that I was mildly dyscalculic, something which made a lot of sense in terms of how my brain processes (or rather doesn’t!) numbers. I made an experimental piece about this at the time (January 2021) which was published later that year in Strukturriss magazine. Having this awareness has helped me manage this constraint better in my working life, but I hadn’t thought of it again in terms of creative projects until we started to collaborate. Given Kinneson’s advanced skill with numbers, I thought it would be interesting to approach the subject of mathematics and AI from two opposing positions. She came up with the brilliant idea to create our own kind of binary code from encoding phonemes based on their visual representation. I’m not sure I fully understand exactly what she did and how, but it was genius! I think in this way we made a great collaborative team, each pushing the other into territories we wouldn’t normally have gone.

KL: I came to poetry reluctantly, which is probably why this collaboration appealed to me. As JP said, my background is in maths and computing, so that was my comfort zone, but it wasn’t something I incorporated into my prose, which is what I mostly write. But there’s something very mathematical about poetry. I can’t really ever sit down to write a poem without thinking about maths. But I thought that was wrong somehow and just didn’t write poetry. I think JP’s interest in Turing and his work gave me permission to do some mixing, using those mathematical and computational skills to probe the question of what makes a poet and what makes poetry, a question closely linked, I think, to Turing’s own interest in thinking machines.

– Turing was an exceptional mathematician, who made major contributions to logic, cryptoanalysis, morphogenesis, computer science and artificial intelligence. And he was gay, at a time when homosexuality was illegal – in 1952 he was prosecuted for ‘homosexual acts’. How did you go about researching the book and interweaving aspects of his life and thought into poetry?

KL: I spent a lot of time at the archive at King’s College in Cambridge, reading all of Turing’s personal correspondence, his rough work from when he put together his papers or explored other people’s ideas, then his actual papers and the talks he gave about them. This prompted the three sections the collection is divided into: artificial intelligence, morphogenesis, and Turing’s personal life in the fallout of his prosecution.

JP: We deliberately avoided most of the wartime Enigma stuff that Turing is most commonly associated with, not least because it was very much a group effort (with the help of the Poles!) but also because he achieved so much more. Perhaps because I’m queer also, I always had an interest in his personal life (one that was an ‘open secret’ at the time) and in particular the circumstances around his death. There’s been a lot of speculation around this and it’s been romanticised a bit in popular culture as tragic deaths tend to be, but I wanted to get beyond that and incorporate other aspects of his life and work, in particular, morphogenesis, which he became interested in later in life whilst going through the chemical castration mandated by his prosecution.

– An additional ‘presence’ in some of the conversations is that of AI – for example ELIZA (a program developed in the 1960s), GPT-3 and Microsoft Word Automated Alt Text. In the context of Turing’s question ‘can machines think’, what was your experience of these conversations? Have they influenced your own poetic responses in any way?

KL: My philosophy for writing is that what I intend and what the reader takes away are necessarily mutually exclusive events. If there is any overlap, it’s a coincidence or likely a product of our shared privileges. With this in mind, I didn’t really see why something a machine might produce might be any more or less meaningful for a reader. But I think humans like to be the cleverest one in the room, even if we are competing with a computer. It gives us great pleasure to point out we weren’t fooled. So I think the question of whether an algorithm can make poetry, whether it can think, gets conflated with whether or not we believe a human made the poem. And even as I was trying to explore the question, using AI, I found myself always looking for hints of the machine. So it became a sort of struggle between what the algorithm produced and what I thought was “poetic”. I think in all the poems where algorithms are used to produce the text, there was human intervention, whether that was just whitespace or asking the machine the right question. But even with this intervention, I, as the poet, felt like I was losing control. Which is sort of marvellous. JP and I definitely pushed each other out of our comfort zones, but so too did the algorithms.

– The poetry in Machinations employs stunningly inventive experimental techniques. Could you describe one or two of these techniques and the creative processes informing your choices? 

JP: I don’t think we set out to be deliberately experimental, but perhaps because neither of us would claim to be conventional ‘poets’ that gave us a freedom to explore ideas and create pieces inspired directly by Turing’s scientific papers and work. We deployed anagrams, erasure poetry of the prosecution report, found poetry from several of his scientific papers, blackout poetry from one his letters, as well as attempts to convey both visually and in the process of construction the techniques Turing himself deployed, particularly in his work on morphogenesis. It was also important to us from the outset to ‘show our workings’ and explain the process of creation or inspiration, something which is unusual and often frowned upon in poetry. But for this work, the plotting or planning (the play on the word Machinations in the title) was part of the whole concept.

 KL: I think my favourite process was for the poem It is found experimentally that the axis is normally in some definite direction which takes its title from a quote in Turing’s paper The Chemical Basis of Morphogenesis. I wrote some code (or, more likely, stole someone else’s code and messed around with it) that solved the reaction-diffusion equations, which are the equations from Turing’s paper. Turing was interested in why we get things like zebra stripes or leopard spots or certain patterns in sand (what we now call Turing patterns). He came up with a theory that said it was all about how each area interacted with its neighbours. For example, if the sand is completely flat but the wind blows some into a pile, when the wind blows again, the area behind that pile won’t be affected but the pile itself will attract more sand so it might get higher or wider.

And over a long time these piles then might interact with some other piles and so you get those long wavey patterns in the sand. So I implemented a simple version of this, took a snapshot of the pattern, then converted the image to binary. I then used our existing phoneme-binary dictionary to produce the poetic equivalent of that poem, and if there were any gaps in the dictionary, I tried to fill it with something sensible, although at this point the algorithm started making up its own words by connecting phonemes that don’t necessarily go together. Then I took the text output and the image and used both to create a multitextual poem that still connected to the original source: Turing’s paper on morphogenesis.

– Finally, would you like to tell us about what you’re both currently working on, and whether you would consider another collaboration in the future?

JP: Still seemingly obsessed with AI and computing, despite the fact I barely understand it and have little computing and no coding expertise, I’ve just created a chapbook length work called The LaMDA Sonnets, a sequence of deconstructed sonnets based on the recent Google LaMDA “Is AI sentient?” transcript. This is accompanied by digitally manipulated photographs of my old (now also deconstructed!) computer. I’ve no idea what I’m going to do with this, it was a fun quick project for myself, but it would be great to find a publisher somewhere. I have several other collection/chapbook works out currently looking for a home. After the summer, I want to find some proper time to get back to my WiP novel. I’ve really enjoyed this collaboration with Kinneson, and would definitely be interested in more in future. I think we’re both people who tend to be juggling several projects at once, so this worked well as we could do it quite quickly, mostly remotely, and we sparred and sparked well with shared ideas.

KL: This is like the question at the end of those blind date articles! “Nice experimental poetry chat, but let’s just stay friends, 6.5/10.” But seriously, yes! JP and I work together at Full House Literary so we already have lots of interesting discussions around writing. JP is right, though, I’m currently working on about a million different projects including three novels in various stages (one about Ada Lovelace who Turing referenced with his ‘can digital computers think?’ question) and two other collaborative projects (one poetry, one prose). I was also lucky enough to receive an Arts Council England Develop Your Creative Practice grant to work on my short stories so I am dedicating a lot of time to developing those over the next year under the mentorship of the brilliant KJ Orr.

– Many thanks to you both for taking the time to answer my questions.

Links

Kinneson Lalor’s personal website: https://www.kinnesonlalor.com

JP Seabright’s personal website: https://jpseabright.com

Machinations is published by Trickhouse Press.