BlogFrom AI to IA:

Drawing better magic circles: Heterologic

by Daniel Erasmus
It took thirty years since the brothers Auguste and Louis Lumière’s first screening of film in the basement of the Grand Cafe in Paris on 28 December 1895 for the medium to break with its past.
 
Possibly the best film ever made. The Battleship Potemkin by Sergei Eisenstein in 1925 invented a different way of seeing. Before Potemkin movies were stylistic copies of plays. The rule was to keep the camera still and move the people. Eisenstein revolutionised filmmaking by combining different cuts to create an emotional response in the viewer. He invented montage and used an unheard of 158 shots in 6 climatic minutes of film. The Kuleshov effect is where he would juxtapose different characters in sequential shots to create a dialectical montage. The audience feels that the shots are connected, that they are opposing, and that this scene is the dialectic process by which this conflict will resolve. Eisenstein called this dialectical montages, after Marx, who developed the idea from Hegel, who in turn developed it form Heraclitus. 
 
Since the Odessa steps sequence all films have been made using dialectical montage. Today almost unseen by contemporary eyes, film pits protagonist against antagonist; scene after scene, story after story. A century later, following the cold war, the great battles between the grand narratives of modernism, we fail to recognise it as a figure descendable from the ground. A protagonist yearns for an antagonist. Questions demand answers. We do not see the dialectic in film anymore, because it has become part of the fabric that we weave to make the film meaningful. Cuts, closeups, context is the lead in a dance, where we have forgotten the steps and float across the floor enchanted by the music. In the absence of opposition, we yearn for a dialectical other, to push against. The thing we have forgotten in film, politics and poetry is that not all stories grow from this ground.
 
Bakhtin’s dialogic creates, or I should say refracts very a different set of conversations than Hegel’s dialectics. Approaches co-exist rather than conflict- we are left without resolve. Tensions co-exist and do not require resolution, but exists as a potentiated space. Some ideas offer more a salient force to hopscotch to new spaces, ever changing. Context trumps text, always plural, ever changing and only ends when the dialogue ends.  No resolve, only a brief contingent breather, before the dialogic expands to find new surfaces upon which to refract. There is no end, only exhaustion.  
 
Think of it like Reservoir Dogs in reverse, where Mr Pink, White, Brown and Blue come back to life, shift alliances. Sometimes the story is told from the end, sometimes the beginning happens in the middle, and so on. Sometimes they get the diamonds sometimes not, sometimes diamonds themselves are worthless. It is an awful and wonderfully potentiated space, every sentence is a string that pulls webs of (con)texts. In many ways this frame reflects our post-structuralist state. We make sense of the world through the centripetal and centrifugal forces of dialogue. Inwards towards more coherence and outwards to open up contexts, new ways of being or, I should rather say, becoming. 
 

Film is trapped in a dialectic, the web is dialogic.

 
The language of today is not movies, it is the web. We surf across texts, associatively, our board is not black and white but facetted, text is not reified, it is hypertextual. Film is trapped in a dialectic, the web is dialogic. We click from idea to idea, sometimes in dialogue, sometimes in dissonance, sometimes in discovery, a world without end, moving everyway across these data spaces like Pinterest, Facebook, and so on, temporary beachheads in stormy digital waters. The analogy is not quite fitting, because the web is an epistemology, a theory of knowing; and what Bakhtin is creating is an ontology, a theory of being. 
 
I want to move to the space adjacent to the dialogic conversations and create a new term heterologic. It is outside the scope of this post to explain this conversation in detail, with the proper philosophical rigor. What I want to do is give a brief introduction to the concept and apply it to our thinking about big data/ AI interfaces and how they draw the magic circles where we play with data. 
 

Heterologic is the dialogue that emerges from the dialogic between people and computing.   

 
Heterologic is the dialogue that emerges from the dialogic between people and computing. It is the cybernetic learning conversation between people and computers that I was writing about earlier. Text is not seen as passive signifiers, but dynamic, potentiated with machine intelligence. The dialogue that emerges out of interacting with machine generated overview is structurally so different from the dialogue that emerges out of interacting with single text items, that we should describe it with a different term. These interfaces can have low potentiality as reports or be generative as an interaction, which allows us to play with the data. It can be horizontal where people do not have the ability to influence the machine learning technologies as in the Hydra chess machines, or vertical where the very interactions influence the machine learning. Seeing to the bottom of the pool. It can be narrow, conceiving of interactions between a user and a cluster of machines, or broad where many people are interacting with each other, and machines. The dialogue can be centrifugal where it opens up new ways of seeing, such as in the electric mobility case, or centripetal to bring coherence and agreement to groups. This list is not complete, but it starts to reveal the dimensions when we explore these dialogues between computing and human intelligence. It is a start to establish the terms of the grammar of this conversation.
 
An interface is a tear in space that reveals the nature of the dialogue/ conversation. The keyhole through which we look at each other. What is revealed, what is hidden, how do you reverse it, can we play, how rigid are the rules. How are these rules layered, are they built up through social aggregation, or through assemblage of layers upon layers of code. Is the other present, or does it push the centre so far forward that we can’t see (feel/hear, etc) the shadow for the sun. Does it hide the underlying complexity by making complex things appear complicated. Or does it elegantly make complicated things simple. Of course as our interfaces become ambient, the keyhole becomes the door and eventually we’re left with only perceptions grounded in assumptions of how things work. The Amazon Echo reveals nothing of how it is thinking, it is a powerful piece of furniture, convenient for answers, but really it is a conversation two sentences deep. Commercially it is an ingenious move, making a virtual service an immanence as real as the TV, with the same scale of implications, but it is a not a dialogue.
 
I’m not writing here about 4th industrial revolution, it’s important, but too narrow a framing to centred on labour implications. What I’m proposing with heterologics is to reveal the nature of the conversation emerging from this intellectually fruitful combination of machines and people, machines and machines, and people and people. This task I warrant is more important than the illusion of agency in an age of transformation.

Leave a Reply

Your email address will not be published. Required fields are marked *