2019 November: Artist as Medium and Person

01/11/19

One challenge right now is to answer that blunt question: So what? So there is this hypothesis developing where I want to present the writing/reading phenomenon as a model from which all ‘life-like’ behaviour stems. So what?

I want to talk about the role of the author as a medium accommodating other subjectivities in the case of writing fictional characters and in the creative process more broadly. I want to focus on how that concept is a remnant of a spatialised notion of subjectivity; as interiority, as field, as place.

 

‘Predictability Coefficient’

I’ve been thinking about which properties matter to us in judging whether another object has agency. There is this idea (must find source) that an agent can only be detected and recognised as such by another agent, and this, I suppose, is the basis for the Turing test – that if a certain behaviour manages to ‘fool’ a human being that it is wilfully instigated by an agent, then that very fact might be the definition of an agent. There is another idea (must again, find source) that one of the properties that matters to us when assessing whether another object exhibits agency, is how predictable its behaviour is. If it behaves too randomly or erratically, it seems to behave incoherently, meaninglessly, and does not appear to be an agent. If it behaves too predictably, it behaves like an automaton, machine-like and senseless in its own way. But there is a range along this spectrum where behaviour starts to look intentional. I am guessing slightly closer to the ‘predictable’ end of the spectrum. It got me thinking that it might be useful to know what that sweet spot is, in measurable terms. I could implement that knowledge into my fictional characters.

First of all, I should check those sources and also check if any work has been done on this area – plotting the relationship between the perceived emulation of agency and predictability of behaviour (some kind of movement, for example).

It is also interesting to note, in hypothesising about this, that even a single human being can behave more or less predictably, and also seem more or less like an agent. Sometimes a mental illness can cause somebody to behave in ways that do not align with normative expectations of behaviour. When perceived from this normative perspective, such individuals behave, for instance, ‘too’ erratically, and are likely to be perceived as less person-like as a result. A person in a vegetative state, too, seems to have ‘lost’ their personhood for the duration of that state, as their comatose behaviour is highly predictable. In fact, in film scenes where somebody ‘awakens’ from this vegetative state, by twitching a finger, for instance, this event comes as an eerie kind of surprise, as if watching an agent come from a form that a moment ago had no agency, as if something non-living had come to life. Thus, this perception of an other’s intentionality and lucid agency seems to come out of convention – collective agreements, or compromises, on a realm of normative behaviour that is relatively predictable, in which persons can recognise each other. Any one straying away from that realm of likely behaviours given specific situations, is likely to be perceived by the majority as less person-like and more like a person’s counterparts: a machine, a zombie, a lunatic. Indeed, since the consensus on the boundaries of acceptable behaviour given various scenarios can change, that would also suggest that who or what is considered an agent, can also change. Thus, the ‘predictability coefficient’ may be subject to cultural differences, geographically and temporally, and a single agent may exhibit a range of behaviours that score differently in terms of predictability.

With this in mind I have been thinking of generating a simple experiment, involving an animated, pulsating ball. A scroll bar beneath it would change the randomness of the ball’s pulse. I could invite participants to scroll up and down the bar until they find the point at which the ball behaves most like a creature.

I think one way to realise this setup would be to learn to use Wolfram Notebook, which makes it possible to program at a very high level, almost using natural language. It has a lot of inbuilt graphic functionalities which might help me make my ball.

 

Flashpoint 1 Skills Workshop: Question and Method

As asked, I brought an object to the session, which is related to my research: A script. This is my description of it:

A thickness of individuals that would like to pretend to you that they have no thickness, that they are pure surface, come together to form a definite quantity of weight in your hands. Their purpose is to distract you from their materiality and make you forget about them, but you cannot resist the appeal of their concreteness, their smoothness and white snowiness. This is a tool that invites usage: flipping, smoothening, handling. It is an object of reference that would have you look somewhere else: at a performance, or spectacle. Every aspect of its design is made to disappear and be forgotten in the wake of an idea. The script tries to exchange with you its bodily matter for a ghost.

And here is the question I came up with:

Can the dual mechanism of a code and its ‘expression’ contribute to a revised model of autonomous systems broadly and agents or persons specifically? Could this model be used to engineer a fictional character that could pass a Turing test?

 

In reading the introduction and first chapter of Inventive Methods by Nina Wakeford before her session on Tyrannical Methods, I was trying to think what methods I have been using. The first chapter in that book was called ‘Anecdote’. This got me thinking that ‘analogy’ and ‘metaphor’ are perhaps important methods for me. I am trying to model the phenomenon of autonomous behaviour, by turning to the analogy of reading/writing. Cori Hayden is quoted as saying that the use of a variety of methods allows a ‘shifting of analytical gears’, which I found analogous to my experience of doing a 50/50 practice & theory based PhD. These two methodologies provide for me a shifting of analytical gears. Another point the introduction stresses is recognising that methods of observation, analysis and knowledge making in general is co-creating the phenomenon it attempts to study, and is implicated in the ecosystem of that phenomenon.

In the session, we were encouraged to think about methodological factors significant specifically to our own research. For me, they were:

  • What is the nature of the knowledge I am seeking to create?
  • What are the criteria for judging the validity/goodness of my experiments?
  • How do I make value judgments about what is ‘enough’ in terms of research parameters?

I started by identifying that I am approaching this so far with the premise that agency is diagnosable by other agents. This premise seems to set me up for making Turing tests. I can use myself as a measuring device, or tool. I also make the assumption that the appearance of agency is all that counts; that something looks as though it has agency, rather than whether it does. This turns my approach into more of a phenomenological one, rather than determinately ontological, based on the premise that the appearance of agency is perhaps just as valuable to us as the notion of ‘actually possessing’ it (whatever that means). Akin to how Dennett approached the ontology of ‘belief’, by questioning the relevance of ‘real’ when talking about ‘patterns’.

What then seems relevant to ask is which variables could I identify, that potentially have a say in diagnosing agency? If I had a set of variables that were relevant to my research and which I wanted to test as to their significance in diagnosing agency, I could isolate them in experiments and measure their significance. Then the details of ‘what units’? and ‘how to measure’? would determine the method.

 

Flashpoint 1: Art Research Seminar, Breathing Space for Voices

We had wonderful feedback on our seminar, and both respondents and audience members said they found the content of our seminar rich and diverse. As presenters, the three of us had very different approaches to a common theme: our relationships as artists, with character and voice. I discussed the role of the artist as a medium and the implied relationship with space, asking how character as a quality is related to the space of the person, and whether this spatialisation of self and other is illuminating or misleading. E took issue with online autobiography and self-imaging; how a self is written into digital media, by whom and for whom. R spoke about the potential problems and virtues of creating a speculative space for neurodiverse voices, whose stories and perspectives are typically overlooked due being classed as ‘atypical’. Our respondents responded with a list of concepts that they thought recurred throughout our seminar:

  • Faint (from fencing or boxing terminology, poetry by Robert Browning)
  • Habit (surplus habits, repetition, habit-habitat)
  • Will (action, ethics, passivity and automaticity)
  • Epicurus’ theory of the flying simulacra (Gell)
  • Spirit photography
  • Private/public, inside/outside
  • Speculation
  • Simone Vio?
  • Medium (other etymological links: milieu/middle)
  • Downcast Eyes by Martin Jay (other dimensions of modelling – particularly the visual metaphors rife in language that place a primacy on visual perception. Texture, rhythms, pathology…)
  • A Medium’s Medium, exhibition at the Gallery of Everything (esp video)
  • E recommended me also a great performance by Irina Gheorghe, Foreign Language for Beginners.
  • Deep mapping

The Art Research Seminar has given me a good springboard for developing the ideas I had about the ‘medium’ subject. The subject of the medium has raised questions of the spatial dimensions of personhood. Does this relate to the Scripting for Agency project as a whole, which is primarily focused on the writing/reading phenomenon? M seemed to think I was suggesting that ‘space’ is required for this dual textual mechanism to occur. Ultimately, what the project is interested in is what does it mean to ‘run’ a script.

One thought that came to mind: ‘A phrenology of pulse

 

Computing Department

A certain amount of learning is in order. I need to find a way of getting into AI in such a way that it can show me something about my project. A starting point is the Computing Department at Goldsmiths. In addition to that, I’d like to start learning to use Wolfram Alpha, and also follow courses recommended by M.

Modules/Events I’d like to Audit

Research I’d like to catch up on

  • What-If Machine (AI, Robotics & Cognition)
  • Computational Creativity Theory
  • PROSECCO > video lectures
  • Computational Creativity Research Group (led by Simon Colton and Jeremy Gow)

 

20 Percent Autocompletion Experiment

To use algorithm that autocompletes with paragraphs of text using the prompt “Are you aware, that [x]% of things that happen, actually don’t happen?” Then, to scale the percentage up and down and see how that affects the narrative.

 

Plan

Achievements/Notes

Week 1

4-10

Practice Pseudo

x

Propose work for Elephant West

x

Write CHASE application

x

Write AR seminar presentation

x

Week 2

11-17

Skills Workshop (Question & Method)

x

Practice Pseudo

x

Present Flashpoint AR Seminar

x

Week 3

18-24

M supervision

x

Explore AI, Robotics & Cognition/Computing

x

Practice 20 Percent 

x

Submit ARS Output

x

Write reflection on AR seminar

x

Week 4

25-1

Dennett, Real Patterns

x

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s