»In Term of As-If«: Making a Speculative Theory Bot and Thinking About Being a Research-Practitioner

As part of my Akademie Schloss Solitude residency – which revolves around aesthetic and legal questions of algorithmic authorship and generative art-making –, over the next few months I want to think about what it means to be a ›research-practitioner‹ in art and humanities contexts.

I produce media theory (in the broadest sense); I have a creative practice that touches on data, visuals, and sound; I curate… – how can these activities be linked most meaningfully? I often use the term ›research-practice‹ to describe my work, but that term is by no means clearly defined in the arts and humanities. In the technical, medical, and natural sciences, ›research-practice‹ is a well-established term referring to the work of researchers who also do ›evaluative‹ work in ›professional‹ (i.e., non-academic) contexts. While that definition still appears somewhat useful in creative contexts that accommodate both professional and artistic outlooks (such as design, architecture, or photography), it quickly loses clarity and focus when it moves towards experimental contexts of creative and theoretical expression. But it’s here that the feedback loops, cross-pollination, and productive seepage between ›practice‹ and ›research‹ can become especially interesting.

Practice-based research? Research-oriented creative practice?

My work concerns experimental digital art, critical coding practices, and various strands of cultural and legal theory. What can it mean to engage in practice-based research, or to engage in research-oriented creative practice, in interdisciplinary field? Does it matter to make such determinations, and does failing to do so mean that the arts councils, institutional research-grant-givers, and foundations that support my work end up dictating the contours and identity of what I do?

These are serious questions, and I have a serious interest in exploring my identity as a researcher && creative practitioner && curator in digital contexts, but I want to take a playful approach (enough research-based research seems to be conducted on the topic already), and try to respond creatively to the question/problem at hand.

Bot-based theory

@ruminant_theory, a robotic Twitter account (described in more detail below) that might qualify, I think, both as ›practice‹ and as ›research,‹ and which might consequently help me think about research-practice more generally. I’m a hack programmer, so this bot was quite hastily put together. The following is some (mostly non-technical) information on the bot (ask if you want to know more), and a few thoughts on how the bot’s output relates to the some of the broader questions I am exploring during my residency. Once the bot has matured a little and my interactions with it have continued for some time, I will write a few follow-up sketches and explore technical, aesthetic, and legal questions in more detail. Meanwhile, feel free to follow the bot’s speculative rumination on media theory and generative art.

I called the bot @ruminant_theory rather arbitrarily, but partly in order to reflect that, like a ruminant, it regurgitates: it is designed to cough up bits and pieces of the theory I read during my residency, and to mash it up in the form of generative tweets yielding unexpected quasi-theoretical statements that meander between the derivative and the revelatory, the serendipitous and the unfortunate, the nonsensical and the ingenious. Yesterday afternoon, the bot tweeted: »Spoken words are imaged as in an ›integrated terminal‹ of the New York City« [link]. Hm…(?) A few hours later: »Friendster.com fostered a decision engine (which is perverse: in the artwork« [link]. Hmm…(!?)

Ruminating on a glitchy database of source texts

In a nutshell, the ruminant theory bot draws on a complete – and growing – database of every digitally-available theoretical text I am reading during my residency at Akademie Schloss Solitude; it uses this database to create tweet-length mash-ups (the complete list of references can be found here). The bot runs in a modified version of a freely available Google Spreadsheets Twitter bot script developed by Zach Whalen. It uses an algorithm informed by probability theory, allowing it to create informed predictions based on the current state of the text collection (a so-called Markov chain; explained and demonstrated in a very engaging way here by Jonathan Reeve). The algorithm is ›dumb‹ in the sense that it doesn’t know grammar and doesn’t understand the meaning of the words it strings together. It is ›smart‹, however, in that it can analyze the database and produce new text pieces that continue trends in the vocabulary, word order, writing and punctuation style of the source data taken in its entirety. In this sense, the algorithm’s ›memorylessness‹ doesn’t keep it from reacting directly and meaningfully to the readings I feed into it, or from creating what I think are valid and potentially insightful speculations on the ideas that the texts can inspire. The bot therefore mirrors what goes in my mind as I process the readings.

Unsurprisingly, I often recognize my own thoughts in the generative tweets. However, they also tend to suggest new directions for my reflections on the texts I’m dealing with. Reading my bot’s generative output is thus a way of critically re-thinking the theory I’ve already read (rumination engendering rumination); this extends my engagement with the thoughts expressed in the texts, and inflects it in critical ways that allow me to derive new insights. Nobody in particular authors these tweets, and yet they reflect the collective intelligence of the source texts and resonate with my own, helping to generate new insights on the material in the database. To quote from @ruminant_theory (nobody could say it better): »A New Media, or something approaching a threshold – and constant care« [link].

Bot resonance

As my reading will continue over the next few months, and as @ruminant_theory’s text source database will grow, the Markov-chain algorithm will fine-tune the tweets to reflect the focus of my own thinking, and vice versa. For now I can only speculate on what this means, what forms this will take, or what ideas it will engender. When my students have to deal with difficult theory, I recommend that they approach it playfully and irreverently, and take away only what speaks to them, what resonates with their questions and ideas. I’ll follow the same approach with regard to the output of my theory bot: some of it seems incomprehensible (for now!); some of it seems to confirm my own ideas; and some of it seems just brilliant. I will take my bot’s advice, and read its theorising as speculative inspiration, »in terms of as-if« [link].