Text collaboration is a central part of our artistic process, both through research as well as in the mediated location of the audience-artwork interspace.
As a result of the process of working with Being Unthinkable... two different text productions have been published. Here follows an essay written for the book One word from the Dissambled Dictionary: Non-territorial, by Praun & Guermouche, 2020.
Non-territorial consists of text by Yoko Andersson Yamano, Rachel Armstrong, Ylva Björnberg, Athena Farrokhzad, Jörg Heiser, Rolf Hughes, Ronald Jones, Lundahl & Seitl, Petra Malm, Aldy Milliken, David Neuman, Aleksander Perski, Tessa Praun, Theodor Ringborg, Robert Stasinski and Alessandra Di Pisa, and Liv Stoltz.
Reading the words
Like A Train Wreck
In A Zoo
Next To A Circus
reminds us of artificial stupidity, or as it is known by the general public artificial intelligence. Apart from reading as a botched Haiku, it is typical of the trifling language structure of today’s natural language generating systems.
Since summer of 2017 we have been training the IBM Artificial Intelligence system Watson to generate synthetic philosophy. Thousands and thousands of pages of world philosophy were uploaded as the first step of the machine learning process. From this point the training consists of classic reinforcement learning - an area of machine learning where the data sets are inscribed with labels of “relevant” or “not relevant”, regarding the pattern recognizing process.
The first challenge of the project: defining what “relevant” means. And what it doesn’t mean. In going about this delicate and seemingly insurmountable task, we turned to the ambivalence and tactility of art with its non-territorial properties.
Since our understanding of the world is never either text-based or experience-based/tactile, but always both, we decided to design the system to understand what is relevant based on relevant interaction with the physical manifestation of the project: a color and movement shifting robot sculpture, based on a non-humanoid, plastic system.
So, in our model, the perceived ‘relevance’ of a sentence such as “Like A Train Wreck” can only be understood through an action or interaction. It is never relevant without context and therefore can only be understood in juxtaposition with the answer or reaction it creates, which in this case comes from humans interacting with the robot sculpture.
The words are only relevant as social molecules in a larger context, where even seemingly absurd phrases, otherwise deemed “not relevant” can invoke meaningful emotions, memories or other thoughts. The absurd meaning of words becomes triggers for the mind to act and react and to respond to the system.
We asked our system for three different phrases:
Perverse he go through the mind.
So where do you loosen?
Would he even renounce noticed it?
And somehow the words “Like A Train Wreck In A Zoo Next To A Circus” are slowly becoming more and more relevant.