MoonSong
Rules-lawyering drama queen but not a munchkin
Hello, I keep noticing a few misconceptions about the way artists (and ultimately all humans) learn to do art. How these misconceptions lead to oversimplification and then in turn to use those oversimplifications to conclude "that AI isn't that different from how artists do it", and "artists do copyright infringement all the time to learn how to do art". And well, as someone who graduated from art school (and a computer science dropout), I think I ought to dispell some of these misunderstandings, if only so AI threads stop devolving into the same three discussions with different people. Thus, consider this a cliffs notes version of how artists actually learn art, but since all artists are human, we need to start with how humans learn to do art.
Part I How all (most?) Humans learn to do art
Before starting, a couple of caveats. This is a process followed by nearly all children who can see growing up and with the ability to hold a drawing\writing instrument. That's the baseline needed to get started at learning art. Sight and something to draw with. This near universality applies across cultures and is an ingrained path of growth. It should be able to happen even outside our shared culture of being bombarded with visual stimuli all the time and everywhere. Of course, some cultural differences start popping up as we advance through the different stages.
Stage 1 Scribbling. The process of learning to draw starts as an almost purely kinetic activity and devoid of all meaning. This is what we call asemic, everybody's first drawings don't mean anything and don't represent anything. The result does make the child proud, but the most important part is the process of leaving marks on the paper. Early on we don't judge the page, and we consider it pretty by default. As we grow, we start to be more intentional with these scribbles, and eventually we make the jump to make the scribbles stand for something, mostly family and pets. Once we make this jump, it doesn't take long before we start to be a bit more detailed, thus growing into the next stage.
Stage 2 Pre-schematism. During this stage, we abandon the uncontrolled scribble. We start to turn into closed shapes and many times these shapes become solid shapes. The shapes also start gaining some details, like eyes mouths and simple limbs. At this stage we draw essentially signs, and we don't care too much about realism. Our drawings tend to use size to express relative importance. Typically most children end up drawing themselves as the biggest figure in the drawing.
Stage 3 Schematism. At this stage, drawings turn very intelectual and analytic. Children start to notice people, animals and objects are made up of parts and they start to draw these parts as details, even if it doesn't make for the most realistic or pretty drawing. There is an interest for drawing the background in addition to the foreground. At this point, everybody's drawings are full of personal signs invented by ourselves, but there's contagion among peers and signs taught by culture start proping up. We don't draw depth, but we draw the horizon and figures lay flat on the horizon.
Stage 4 Proto-Realism. At around age 8, we start to notice our drawings aren't very realistic, and we intend to solve that. At the same time, we can't help but use conventional signs and symbols taugh to us. (Like the check mark to represent a bird whens in the previous stage most of us were happy attempting to draw a full bird with wings, head, beak and eyes) For example, we start to use positional height to represent distance. The closer objects are drawn lower on the page and the farther objects are drawn higher on the page (just like in the old Sierra games). Also, at this point we start to leave behind purely geometric shapes to draw, but we still cling on to learnt procedures and 'recipes' to draw.
Stage 5 Pseudo-realism. We discover size to represent depth, but most of us still have an issue with true perspective. Some of us definitely abandon recipes, but many still cling on to them. We attempt to do even more detail than at stage 3 because at this point (around our pre-teens) we notice more detail.
Up to this point most people stop learning how to draw. So far the learning process is entirely innate and self-directed if "polluted" by our exposition to drawings and artwork around us. It is possible for us to actually copy existing art with different levels of success as early as stage 2, but we are completely uncapable to benefit from copying because our brains aren't mature enough. Frustration is a common reason to stop drawing altogether, but some 'gifted' lucky people manage to break from pseudorrealism to actual realism on their own. At this point people who go on to learn art, need some form of external guidance.
Part 2 How artists learn how to art
By 'art' I will mean exclusively the plastic creation -drawings, paintings and illustrations- not the wider and simultaneously more restrictive meaning of 'art' -as in all of the other high arts as well as performances and instalations, and everything else from photography to a banana sticked to a wall-.
The process of learning art can vary from person to person, and from institution to institution, but it consists mainly of three pillars, visual culture and philosphy, sight training, and kinetic training. Oh yes, learning art is a very tactile process and very physically involved. You need to learn the basics from how to sharpen and hold a pencil, to the more involved learning how to move so your hand goes precisely where you need it to go. You also need to learn a lot about how different materials\mediums respond to you, as your relationship with the different mediums will shape how your art will come up. (To date I have a difficult relationship with watercolors). Finally, you need to move and appreciate the third dimension and volume.
The second pillar is training the way you see. Esssentially, you need to be aware of the way you look at the world and train yourself to observe and perceive reality in new ways. It isn't just enough to perceive like everybody else. You need to train yourself to perceive the background as if it was the foreground, to measure just by looking, to perceive texture, to notice the interplay of light and color, to let go of your conscious knowledge of symbols and sings and ignore all preconceptions about how people and animals should look like, to channel you inner six-year-old and once again compose reality as groups of geometric volumes instead of flat shapes, and to somehow switch modes at will. This can imply rewiring your brain. (After long sessions in the zone I lose the ability to speak for a while)
Finally the third pillar, along with long discussions on the nature of beauty, art, ethics and the philosphy of art, you also get exposed to a lot of existing paintings, sculptures and other pictures. You are however, not copying them most of the time. The point of looking at all of this is being exposed to the ideas and ideals behind these creations. Also so that you can see how these people solved certain situations, you know, so that you don't have to reinvent the wheel. You only copy when you want to try an specific techique, and only because an existing work is a known quantity. (And very importantly, it is in theory posible to learn and develop a lot by copying the same work over and over and over.** Because as I mentioned before, you need to train your observation skills and muscle memory. Can an AI improve by training on the same image over and over?)
And I haven't mentioned everything else, theory of color, projection, perspective, technical drawing, psychology of the image, and so on.
Of course we are also influenced by cartoons and other existing styles. But very importantly, we don't want to copy someone else's style verbatim because we want to stand out from the rest of the creators and have our individuality shine through. Even if we consciously or unconsciously 'store' a bunch of images in our head*, visual memory is way less important to the art process than everything else combined. That and that as shown by the way everybody starts to draw, the act of drawing is a process that goes from within to the outside. It reflects our experiences, hopes, dreams, needs, wants, traumas and achievements. A lot of time inspiration comes from non-visual sources. A poem, a song, a note, a smell, a traumatic experience, wanting to tell something.
So yeah, totally the same way a disembodied piece of software without drives, emotions, wants or needs 'learns' to create art
* And very importantly, a non-insignificant number of artists have aphantasia. They are literally uncapable of remembering images. Even those of us who remember images, can remember them at varying degrees of exactitude. And did you know that each time you remember something you alter that memory? Finally, I myself can attest that what you end up drawing is not always exactly what you have in mind.
Edit: ** On a second thought it is kind of true. During the early stages you make lots and lots of self-portraits. You're copying over and over from a mirror!
Part I How all (most?) Humans learn to do art
Before starting, a couple of caveats. This is a process followed by nearly all children who can see growing up and with the ability to hold a drawing\writing instrument. That's the baseline needed to get started at learning art. Sight and something to draw with. This near universality applies across cultures and is an ingrained path of growth. It should be able to happen even outside our shared culture of being bombarded with visual stimuli all the time and everywhere. Of course, some cultural differences start popping up as we advance through the different stages.
Stage 1 Scribbling. The process of learning to draw starts as an almost purely kinetic activity and devoid of all meaning. This is what we call asemic, everybody's first drawings don't mean anything and don't represent anything. The result does make the child proud, but the most important part is the process of leaving marks on the paper. Early on we don't judge the page, and we consider it pretty by default. As we grow, we start to be more intentional with these scribbles, and eventually we make the jump to make the scribbles stand for something, mostly family and pets. Once we make this jump, it doesn't take long before we start to be a bit more detailed, thus growing into the next stage.
Stage 2 Pre-schematism. During this stage, we abandon the uncontrolled scribble. We start to turn into closed shapes and many times these shapes become solid shapes. The shapes also start gaining some details, like eyes mouths and simple limbs. At this stage we draw essentially signs, and we don't care too much about realism. Our drawings tend to use size to express relative importance. Typically most children end up drawing themselves as the biggest figure in the drawing.
Stage 3 Schematism. At this stage, drawings turn very intelectual and analytic. Children start to notice people, animals and objects are made up of parts and they start to draw these parts as details, even if it doesn't make for the most realistic or pretty drawing. There is an interest for drawing the background in addition to the foreground. At this point, everybody's drawings are full of personal signs invented by ourselves, but there's contagion among peers and signs taught by culture start proping up. We don't draw depth, but we draw the horizon and figures lay flat on the horizon.
Stage 4 Proto-Realism. At around age 8, we start to notice our drawings aren't very realistic, and we intend to solve that. At the same time, we can't help but use conventional signs and symbols taugh to us. (Like the check mark to represent a bird whens in the previous stage most of us were happy attempting to draw a full bird with wings, head, beak and eyes) For example, we start to use positional height to represent distance. The closer objects are drawn lower on the page and the farther objects are drawn higher on the page (just like in the old Sierra games). Also, at this point we start to leave behind purely geometric shapes to draw, but we still cling on to learnt procedures and 'recipes' to draw.
Stage 5 Pseudo-realism. We discover size to represent depth, but most of us still have an issue with true perspective. Some of us definitely abandon recipes, but many still cling on to them. We attempt to do even more detail than at stage 3 because at this point (around our pre-teens) we notice more detail.
Up to this point most people stop learning how to draw. So far the learning process is entirely innate and self-directed if "polluted" by our exposition to drawings and artwork around us. It is possible for us to actually copy existing art with different levels of success as early as stage 2, but we are completely uncapable to benefit from copying because our brains aren't mature enough. Frustration is a common reason to stop drawing altogether, but some 'gifted' lucky people manage to break from pseudorrealism to actual realism on their own. At this point people who go on to learn art, need some form of external guidance.
Part 2 How artists learn how to art
By 'art' I will mean exclusively the plastic creation -drawings, paintings and illustrations- not the wider and simultaneously more restrictive meaning of 'art' -as in all of the other high arts as well as performances and instalations, and everything else from photography to a banana sticked to a wall-.
The process of learning art can vary from person to person, and from institution to institution, but it consists mainly of three pillars, visual culture and philosphy, sight training, and kinetic training. Oh yes, learning art is a very tactile process and very physically involved. You need to learn the basics from how to sharpen and hold a pencil, to the more involved learning how to move so your hand goes precisely where you need it to go. You also need to learn a lot about how different materials\mediums respond to you, as your relationship with the different mediums will shape how your art will come up. (To date I have a difficult relationship with watercolors). Finally, you need to move and appreciate the third dimension and volume.
The second pillar is training the way you see. Esssentially, you need to be aware of the way you look at the world and train yourself to observe and perceive reality in new ways. It isn't just enough to perceive like everybody else. You need to train yourself to perceive the background as if it was the foreground, to measure just by looking, to perceive texture, to notice the interplay of light and color, to let go of your conscious knowledge of symbols and sings and ignore all preconceptions about how people and animals should look like, to channel you inner six-year-old and once again compose reality as groups of geometric volumes instead of flat shapes, and to somehow switch modes at will. This can imply rewiring your brain. (After long sessions in the zone I lose the ability to speak for a while)
Finally the third pillar, along with long discussions on the nature of beauty, art, ethics and the philosphy of art, you also get exposed to a lot of existing paintings, sculptures and other pictures. You are however, not copying them most of the time. The point of looking at all of this is being exposed to the ideas and ideals behind these creations. Also so that you can see how these people solved certain situations, you know, so that you don't have to reinvent the wheel. You only copy when you want to try an specific techique, and only because an existing work is a known quantity. (And very importantly, it is in theory posible to learn and develop a lot by copying the same work over and over and over.** Because as I mentioned before, you need to train your observation skills and muscle memory. Can an AI improve by training on the same image over and over?)
And I haven't mentioned everything else, theory of color, projection, perspective, technical drawing, psychology of the image, and so on.
Of course we are also influenced by cartoons and other existing styles. But very importantly, we don't want to copy someone else's style verbatim because we want to stand out from the rest of the creators and have our individuality shine through. Even if we consciously or unconsciously 'store' a bunch of images in our head*, visual memory is way less important to the art process than everything else combined. That and that as shown by the way everybody starts to draw, the act of drawing is a process that goes from within to the outside. It reflects our experiences, hopes, dreams, needs, wants, traumas and achievements. A lot of time inspiration comes from non-visual sources. A poem, a song, a note, a smell, a traumatic experience, wanting to tell something.
So yeah, totally the same way a disembodied piece of software without drives, emotions, wants or needs 'learns' to create art
* And very importantly, a non-insignificant number of artists have aphantasia. They are literally uncapable of remembering images. Even those of us who remember images, can remember them at varying degrees of exactitude. And did you know that each time you remember something you alter that memory? Finally, I myself can attest that what you end up drawing is not always exactly what you have in mind.
Edit: ** On a second thought it is kind of true. During the early stages you make lots and lots of self-portraits. You're copying over and over from a mirror!
Last edited:
