progress report
A few entries ago I was talking about a process-based drawing experiment that I was planning to carry out with several unsuspecting 'victs and uh, here's the loot. This weekend I've been working on a second experiment that will explore the ways in which this process can be fine-tuned.
I'm not going to bother doing a good job explaining how this process worked - it's a chore. There are quite specific directions, and one of my challenges has been to simplify the process.
So, I drew (on one sheet of paper with pencil) nine diagramatic pix. Here's what those original collection of pix looked like (I redrew them in the same way on each of the sheets):
Each of these collections was associated with a phrase (which can be found on that previous entry)and the participant's job was to judge how well the pix served as an abstract metaphorical illustration for the phrase. This is a pretty odd task in the first place, and as you will see from the following examples, gross tuning is needed.
The participants were asked to select 3 pix that they felt were irrelevant or bad illustrations of the phrase, and ERASE them. They were also asked to choose the 3 pix that best illustrated the phrase and, somewhere nearby the original, copy the drawings, but in an "improved" way - a few simple changes to make the drawings better illustrate the metaphor. In the end how many drawings should there have been, dear reader? There should have been the same amount as in the beginning - nine. Then the paper was passed to a new person to repeat the process.
The point of this game is to see if it is possible to create an algorithm incorporating an obliquely-concerned "wet-ware" filter that will create drawings that have greater fitness in terms of a certain task. The goals are both to see what these drawings would look like and to see if and how this process can work. I say "obliquely-concerned" because if the concern were direct this would be a simple creative act. But through a strict set of rules the intentions of the participants are partially subverted. The process can't be described as fully automatic, or even objective goal-based, but neither can it be fully explained in the language of individual creative drawing. The challenge is finding a balance here.
There were lots of technical problems with this experiment.
The first was that the directions were too hard to follow (although my first urge is to blame the participants for not being bright enough, this is less attractive not because it is untrue, but because it is boring - of course humans are foolish... if approached wrong - the point of this experiement is to take advantage of human's uncanny ablitiy to occasionally be smart). Problems included people not erasing the original pix so that the total number started multiplying, and people whose drawings were more accurately described as completely novel rather than "improvements." So I have to make the directions clearer or maintain stricter control over the process (this is the more likely of the two choices, between dull simplicity, inevitable chaos, and authoritarianism I guess the last is the most relevant considering the world we live in).
The second problem with this experiment has to do with the tendency towards representational drawing. The goal of this excercise is to create images that say something new about the nature of these phrases, to explore the nature of abstraction, to create visual analogies that have something new to say, to see if there are any essential truths about the mind's ability to form nonlinguistic representation. However it seems that the natural urge is for people (in this case art students) to tend towards simple representative illustration. I don't have a knee-jerk reaction to illustration, but I think it is less interesting in this case because it creates pix that closely approximate linguistic communication - where the communicative ability of the drawings could be effected more accurately and efficiently by replacing them with words. In this case the excercise will result in the production of tautologies, inefficient, noisy ones at that.
I am working on a solution to this problem. I can't discuss it at this point.
Here are a few examples of the results. I have chosen the most successful as I want to show what this process could be, rather than the traps it can fall into (anti-entropic processes are more interesting than... that other shit):
As usual, click image for more details.
I am looking for advice or feedback on this...
5 comments:
I'd like to see the "illustrational" ones you mentioned, so I can understand what you mean.
I'm sure you remember the thing we did in my room in Manor with the sentences and the pictures, during the listening party. I always thought it was interesting how people took things differently. Like, you especially drew a kind of rebus, but I didn't think of that, I tried to just draw a picture that expressed the idea somehow...and also with the sentences, I never tried to describe in detail what was in the picture, but just would write a sentence that was vaguely related, whereas Danny would try to describe it exactly.
So I think the instructions need to be pretty specific if you want something specific, but that will certainly make the results less idiosyncratic in some ways, because it will make everyone's thinking into your thinking...? I mean, the first sentence, in our game, was never "At the top left there is a weird orb thing and next to it there is a chain; a man is hanging from this chain by his feet", right? It was always, like, "What a jolly holiday with you, Ms. Appenkloffer!" So I thought, when I was writing my sentence in reaction to the pictures, that if I wanted to recreate the first sentence as best I could, it was better to take a wild guess with something like "I can't believe you'd say such a thing!" rather than a minute description. The odds against my hitting the right sentence were like a gazillion to one, but if I was just describing exactly, the odds would have been zero; so went my thinking.
All this is just to say that, if you want people to add their input, you're going to necessarily get things that are wildly out of line with your predictions/expectations/hopes, unless you ruthlessly constrict their choices.
Anyway it's hard to understand what you're getting at if you only show us the "good" ones and not the "bad" ones. If I see the "bad" ones maybe I'll be able to see where you're coming from.
I think you should put nine circles of the same size on the page, just like tic tac toe - three across and three down and evenly apart-spaced. And people hae to erase and draw inside those circles and there should be nine when they're done. then it will look neat and tidy.
I like quite a few from off on a tangent. I really understand what you're saying -- I dunno -- A triangle to me, or many triangles, means I DO NOT UNDERSTAND WHAT YOU'RE SAYING. A triangle? Maybe they were using the opposite to illustrate.
Farley, I have absolutely nothing to do when i go to work.
love
A
Thank you for your advice, friends.
I am thinking a lot about freedom and restriction not only in terms of rules of games, but in terms of the rhetoric of american politics.
It seems that in this culture (by which I can only mean my interpretation of what american culture is) that SAYING something against freedom is not allowed, that freedom must always be the ideal. There are also many false freedoms (i.e. the freedom to chose what gas station to fill your tank at), but vrey few true freedoms, or rather actualized true freedoms, because there is true freedom, always, if you know what I mean.
So this has to do with the difference between playing and games. In games there are always rules to follow, rules which sometimes make the game more fun (i.e. there are some ways in which chess tokens function as free-ranging imaginational objects, but the possibility of exploration is much higher when they are used in the rules-system of chess, at least for some people). In play, free play, there are no rules, just pure imagination. But it seems to me that the tendency is for people to invent their own rules and follow them. For example, if you give someone a plastic frog and a matchbox car they will look at them, spin them around, put them in their mouth, drop them, etc. These are less play than just ways to further perceive them, to test the sensations that can be derived by interacting with them. But soon it turns to a kind of self-invented game. The frog becomes the dad and the car becomes the mom and they fall into roles, into rules. The frog becomes a goal, and you try to push the car into it from across the room. The frog becomes the car's passenger and you try to roll it without the passenger falling out. I think the rules here don't interfere with true freedom, only the rhetoric of freedom.
I don't mind inventing new rules for games because I don't think they restrict true freedom. Yes, in the context of these games I have a role more similar to the person who is given the frog and car than to the person who is shown the "bop the frog with the car game." But nobody is going to miss out on dinner in the rule-system I'm presenting.
IndividualFrog, I understand that you are saying that in order to guess a highly-detailed and specific sentence your guess must be highly-detailed and specific. That even though this has a very low chance of being correct, a generic sentence will never be right. (Twentry questions is another illustration of this idea.) But that is only true in a system where there are only two choices; right or wrong. And in a situation where there is only one iteration, one generation allowed. And I don't think things are like that for me. I would rather have someone get a fuzzy inaccurate idea of what is going on with me than an idea that is very specific but inaccurate. After all, it's a conversation. There will be time to correct the inaccuracies, to revise the ideas.
I probably should post the ones that didn't work, but I am going to be re-doing this this week, so we'll see then.
Alexis, are you saying that you want it to be more restrictive?
Oh, well, I wasn't really saying anything about my technique being the right one or the wrong one. Both Danny and I were perfectly right, I mean. I'm just saying what was so great about that game was that we all took the same rules and followed them completely differently. I'm just using my own thinking as an example, partly because I knew what I was thinking and can only guess about the others, and also because my way and Danny's were pretty opposite, so it made a good illustration of how much latitude there is in seemingly simple, straightforward rules.
Totally.
I think my goal is to create rules that limit the growth of information to the growth of knowledge within the system. Information naturally grows exponentially, whereas knowledge increases linearly. Therefore any mapping will be a gross approximation.
Post a Comment