Less Wrong is a community blog devoted to refining the art of human rationality. Please visit our About page for more information.

Drawing Less Wrong: Observing Reality

34 Post author: Raemon 21 November 2011 05:10AM

This is the fourth post of the Drawing Less Wrong mini sequence, in which I discuss how to draw, how learning to draw *effectively* relates to rationality, and what the initial results were when I started running a drawing workshop, teaching people with essentially no experience.

Information here is a combination of lessons I've learned from numerous art teachers who all agree with each other, and some of my own observations that I'm pretty confident about. When I talk about "how the brain does things" I'm using a mix of folk psychology and guesses based on my limited knowledge of neuroscience, which may not be technically accurate but should be sufficient to make useful predictions.
 
Previous posts include the Introduction, "Should you Learn to Draw?" and "An Overview of Skills".


 

To draw a city, you must walk around that city and look at it. 

You can't sit in a room with the blinds closed and create a map and expect it to be accurate. You cannot draw what you cannot see. To draw things, you need to look a things. This is surprisingly hard for a few reasons.

One is that you may want to be drawing imaginary things. I'll talk about this at length in the a later post. For now, let me just say that you can't *learn* to draw realistically (even realistic fantasy) by drawing things that aren't real. 

Another reason is that when people begin, they do not have very good hand-eye coordination. Your can't trust your hand to move on its own - you feel like you must be watching it the entire time, staring intently at the pen and paper and making sure they're doing what you want them to. Coupled with this is a gross overconfidence in how good your memory is.

The third, and most significant reason, is that you don't know how to see in the first place.


How you will probably draw, even knowing this is how you will probably draw:

Inexperienced artists will take a look at their subject, create a mental map of them in their brain, and then turn back to their paper. Slowly, carefully, they draw lines corresponding to that mental representation. Occasionally they may look up - but in the moment they look up, their finger slips and they draw a line that's completely inaccurate. This reinforces the mistrust in their hand-eye coordination, so they spend more and more time focusing down on their paper.

At the same time, as their drawing takes shape, it starts to look interesting. They've already stopped looking at the reality in front of them, but now they start getting distracted from their mental model. If you're drawing a human, you might get get focused on their eyes. You spend a lot of time on them, and they become the new map from which you navigate. You draw the nose or the lips based on where you drew eyes, the chin based on the lips. Your pencil journeys through a map by following a map which was following a map. You work from one small, interesting area to another, never considering how the drawing will work as a whole. Occasionally you'll think back to the original mental model in your head, but it will have grown fuzzy by this point.

And all this time, you probably drew slowly, using small, careful lines. Because after all, your hand-eye coordination is untrustworthy, and you wouldn't want to draw something too big and sloppy.

Minutes go by. Eventually you'll look back up at the real territory that was in front of you, and all the lines will be off. You drew the legs crossed when they were wide apart. The elbow is pointed up rather than sideways. The lips, nose and eyes, rather than forming a straight line up the center of the face, are crooked - each one slightly off, and referencing the lines of the previous one until they had little relation to the actual face. 

Your conscious mind won't process most of this. The drawing will just feel "off".

In that moment, you're in the middle of a kinesthetic process that *felt* like the right way to do it, so you probably won't realize the obvious problem: you should never have been drawing from a mental representation in the first place, you should have been drawing directly from the reality in front of you.

 

How You Should Actually Draw:

An unfortunate truth for beginners is that you must spend most of your time looking at your subject, and that you must also spend most of your time actually drawing. But you don't have the technical skill necessary to do both of those things at once without your drawings looking horrendous.

(First off, be okay with your drawings looking horrendous. You're building new skills from scratch. Your drawings WILL look messy. That's fine.)

To start with, develop a habit of spending at least half of your time looking at your subject, switching back and forth every 2-3 seconds. If you've spent more than 5 seconds looking at your drawing, it's time to look back up and make sure the lips you're drawing are in the right spot compared against the REAL chin and the REAL eyes. This actually isn't too hard, except that you'll forget a lot. Having a teacher to remind you to stop looking at the paper will be helpful. If you don't have a teacher handy, find a person to draw and ask them to remind you to look at them if they notice you staring at the page too much.

A step up from this is to practice making pencil strokes *while* you're not looking at the paper. Eventually you want to be able to do this for extended periods of time. For now, a good technique is to allow yourself to look at the page as often as you want, but *only* make marks while you're looking at your subject. (In practice this also means drawing for 2-3 seconds at a time, and glancing down to make sure your pencil hasn't gotten lost)

Again, your first several strokes following this technique may come out very off. Don't worry about that. 

 

The problem with Mapmakers, and Territory:

So, one big problem is that you aren't going to naturally look at things, and your underdeveloped coordination will reinforce this.

But there's another problem - a huge problem. Which is that even when you're looking at something, you're usually not actually "seeing" it.

Human brains take a lot of shortcuts when they're observing things. When you look at a person, what you perceive is not a series of shapes and colors that correspond to what's there, but rather a bunch of hastily constructed symbols that convey the information that the brain thinks is important. If you haven't rewired your brain for drawing, then "important" questions do not include "Is that elbow angled at 90 degrees or 75?" or "Where are the eyes in relation to the top of the head?" Instead, what you usually care about are things like "is this person happy, or angry?" and the information that gets recorded is a little tag that says "Smiling" with a vague curving-upwards-line symbol accompanying it.

A large chunk of the information we usually need has to do with the face. This plays a role in two common biases that are near-universal in inexperienced artists:

  • Drawing the head much larger than it actually is, compared to the rest of the body
  • Drawing the "face" (i.e. everything between the eyebrows and mouth) as if they took up the entire head rather the bottom half. Practically everything above the eyebrows conveys no relevant information, so it's just ignored.

Your brain has a mental model of what a human is "supposed" to look like, and that model is wrong. You can see major gains in drawing capability just by learning the "ideal" proportions of a human being. Most humans are shaped pretty similarly. But I'm hesitant to just give you that information, because it can actually be damaging. A few reasons why:

  1. Not all people are shaped the same
  2. "Shapes" change dramatically depending on how a person is posed. 
  3. You probably want to learn to draw things other than humans, at some point.

It's not good enough to create a more accurate model of what people are "supposed" to be. You need to look at a subject and discard all your preconceived notions of what they are "supposed" to look like, along with all the symbols and names that your verbal center assigns to the pieces. You need to ignore the little tags that say "face" and "arms" and "hands" and "torso," and instead see the lines, shapes, colors and shadows that are there in reality.

 

"Drawing on the Right Side of the Brain" (Again)

Much of our knowledge of how this works and how to improve the process is relatively recent. In 1969, an art teacher named Betty Edwards was frustrated with her students, some of whom were having extreme difficulty. They could clearly see where things were, but that knowledge didn't translate into the drawing. 

"Can you see that the orange in this still life is clearly in front of the vase?"

"Yes."

"Well, in your drawing, you have the orange and the vase occupying the same space."

"Yeah I know. I didn't know how to draw that."

On a whim, she asked students to copy a painting while it was upside down. There was an immediate, dramatic improvement. 

"How can you draw upside down when you can't draw right side up!?"

"Upside down, we didn't know what we were drawing!"

That experiment prompted a series of questions and investigations that led Edwards to the neuroscience of the time, which suggested that humans used two major processing centers, located in different hemispheres of the brain. The "left brain" dealt with verbal and analytic processing. The "right brain" dealt with visual and perceptual processing. Many people naturally draw using their left brain, which wants to name things and refer to existing knowledge about them. Edwards developed a series of techniques to suppress the left brain so that right brain processes can take over. She published a book in 1979 called "Drawing on the Right Side of the Brain," that discussed her research and the accompanying exercises.

(Later on, neuroscientists learned that while the two processing centers are real, they are not neatly divided between brain hemispheres. The modern edition of the book uses the terms "left mode" and "right mode" to distinguish between the modes of thought)

The book has become extremely influential within the art field, and in recent years Edwards has worked to find ways to transfer "right mode thinking" into areas beyond drawing. In corporate seminars (typically lasting three days), she spends the first day and a half teaching employees how to observe and draw. The second day and half, she helps them outline workplace problems using drawn visual metaphors, which allow them see things from multiple perspectives and stumble upon solutions that seemed obvious in retrospect, but which had gone unnoticed (in one case, for decades).

Edwards breaks down drawing into five main subskills:

  • The perception of edges
  • The perception of spaces
  • The perception of relationships
  • The perception of light and shadow
  • The perception of the whole, or gestalt

She considers these the building blocks necessary to develop two final skills: Drawing from memory, and drawing from your imagination.

 

Integrating Observation and "Gestural Thought"

Eventually, your goal is to be able to do observe near instantly. The 30-second gesture drawings are important because they train you to evaluate and draw in one fluid stroke of thought and pencil. In less than five seconds you should be able to draw a line that describes the general size and shape of the body, and a small ovoid shape that describes the relative size of the head. 

But that will be difficult, until you've practiced several slower-paced exercises that develop your ability to see reality and compare objects to each other. In my first workshop I realized that it's difficult to develop "gestural thought" unless you can do one of the following:
  1. Construct a mental model of your subject in a matter of seconds (and be able to revise it on the fly)
  2. Have a pre-existing model (i.e. "ideal proportions") to use as your starting point, which you then fix as you have time to observe in more detail.
  3. Spend time observing your subject, creating a specific model of them, before you begin drawing.
I'm still experimenting with the ideal order to present various exercises, to develop sufficient observational skill as quickly as possible. As I noted in the last installment, observation and "gestural thought" require different types of thinking that initially will be hard to switch between. Giving them time to gel independently is important, but I believe students should learn to integrate them as quickly as possible. My students and I both made progress in second workshop (mostly by using the third technique listed above), and there is definitely room for improvement.
In a future post I'll be outlining some specific exercises to develop observation. Several of which will be lifted directly from Edwards' book, others developed from my teaching experiences. Developing your observation will, initially, involve many slow-paced exercises. Some focus on helping you break down the barrier between "left mode" and "right mode" thinking, but typically won't be used as often when creating a polished piece of art. Others specifically develop the skills you'll use to observe during "regular" drawing. 
The goal will be to develop a baseline competence in observation, so that you can continue to develop it simultaneously alongside your ability to work quickly and energetically.



Next time, I discuss the basics of technical skill.



 

Comments (15)

Comment author: jknapka 22 November 2011 05:47:02PM 7 points [-]

Thanks, Raemon, this is inspiring. It reflects my experience learning to draw as an undergraduate, many years ago. I have not drawn much since college, but I do recall vividly the experience of, "Holy crap, is that what a person really looks like?!?" upon first producing a half-decent quick figure drawing. I eventually developed a pretty decent drawing ability, which has atrophied quite severely in the intervening years. The experience definitely influenced my overall thinking though -- I'm very aware that my brain is not telling me the actual shapes and relations that correspond to the light hitting my eyeballs, unless I take the trouble to consciously examine that first-order sensory input. And taking that idea to the meta-level, realizing that my mental models of other things might be wrong in hard-to-notice ways, and taking the trouble to at least try to notice them, has been a valuable skill. (Even if not always applied rigorously.) So I think this is absolutely a valuable sequence; and it's prompted me to try drawing again, too.

Comment author: Sniffnoy 22 November 2011 08:31:13AM 7 points [-]

When you look at a person, what you perceive is not a series of shapes and colors that correspond to what's there, but rather a bunch of hastily constructed symbols that convey the information that the brain thinks is important. If you haven't rewired your brain for drawing, then "important" questions do not include "Is that elbow angled at 90 degrees or 75?" or "Where are the eyes in relation to the top of the head?" Instead, what you usually care about are things like "is this person happy, or angry?" and the information that gets recorded is a little tag that says "Smiling" with a vague curving-upwards-line symbol accompanying it.

I suppose you could say that ordinarily when people attempt to draw things they instead do something more like diagramming them?

Comment author: Alicorn 22 November 2011 03:58:21PM 2 points [-]

I like this way of saying it. If I ever try to teach a person to draw I will steal it.

Comment author: Raemon 22 November 2011 03:21:50PM 1 point [-]

Essentially, yeah.

Comment author: ahbwramc 21 November 2011 09:46:53PM 3 points [-]

Very interested in this sequence. Over the past several months I've been systematically reading through all of EY's posts chronologically, with a sort of vague notion that when I was done, I would sit down, take some of the insights I had gained, and start trying to, you know, improve my life. By the time I had finally finished my read-through, a number of things had crystallized, but by far the most obvious and glaring problem was that I had a fixed mindset (as opposed to growth). Actually, I don't know if EY ever touched on that topic directly, but in any case this site introduced me to the concept (and man, did I see myself reflected in the description of fixed mindset people, it was kind of scary).

Anyway, all this to say that this sequence interests me because a) I'd like to know how to draw better, and b) drawing seems like a perfect test case for trying to break out of the fixed mindset - before I would have assumed that I just didn't have the talent for drawing, but now it occurs to me that (who'da thunk it?) perhaps I could get better if I practiced it. So for me, at least, this sequence is very much related to rationality. Thanks, and keep it up!

Comment author: Raemon 21 November 2011 10:34:39PM *  0 points [-]

Thanks! Glad to be of help.

Comment author: JenniferRM 22 November 2011 11:49:18PM *  3 points [-]

When you look at a person, what you perceive is not a series of shapes and colors that correspond to what's there, but rather a bunch of hastily constructed symbols that convey the information that the brain thinks is important.

A way to emphasize this vividly (though it may trigger a physical brain fallacy) would be to see what video reconstructions from brain activity look like, especially when they are noisy and somewhat error prone so slippage between different cognitive alternatives is visible in the video. It would be interesting to try calibrated computer reconstruction of imagery on people who have had 0, 10, 100, 1000, and 10000 hours of training in artistic seeing. Maybe even the same people as they rode the learning curve? :-)

Comment author: kremlin 13 October 2012 09:24:07PM 1 point [-]

I drew loads when I was a kid, and I must have been about 10 or 11 when I realized, had a moment of epiphany really, some of the stuff in the "How you probably draw / how you should draw" section.

I was looking at the cover of the Toy Story VHS and trying to copy it. I remember specifically I was drawing Woody's face, his right cheek to be exact (must have been this photo), and I stopped myself when I realized that what I was about to do was a result of me drawing not what I really saw, but what I thought I should see given my mental model of a face. I think I was about to draw his right cheek too wide or something like that.

That was the first time I ever truly and deeply grasped that concept. I take a certain joy in knowing that I learned it independently.

Comment author: Raemon 23 October 2012 09:54:48PM 0 points [-]

Cool! I am proud of you for doing so!

Comment author: AShepard 22 November 2011 06:04:06AM 1 point [-]

It looks like a couple of footnotes got cut off.

Comment author: Raemon 22 November 2011 02:55:29PM 0 points [-]

Whoops. I probably won't actually have time to fix them for another few days, but they weren't particularly important. (I had assigned myself a deadline to stick to for this one, and I'm actually not happy with a few areas that I intend to fix later).

Comment author: Raemon 28 November 2011 04:26:37PM 0 points [-]

Next installment will be up sometime later this week. The Technical Skill and Instilling Energy articles need to be written in tandem in order to work properly, which is taking a little longer.

Comment author: EphemeralNight 24 November 2011 09:08:04PM 0 points [-]

Crud.

Reading this has further diminished what little hope I still have of ever being able to draw. My own experiences and difficulties don't seem to correlate at all with anything that you describe. (I had to be reminded to look at my paper more during gestures; I caught myself feeling so rushed that there wasn't time to look away from the model -- the upside-down thing my art class did once in highschool, and my attempt came out considerably worse than rightsideup...)

Comment author: Raemon 05 December 2011 05:40:30AM *  0 points [-]

I've been pondering whether or not to respond with words of encouragement. Some people do have a genuinely hard time drawing, and you may be one of those people.

But I think if you have a skilled, flexible teacher, they will probably be able to identify the particular things you're having issues with, and have you work around them. I don't know how to locate a "good" teacher though.

The classes I'm teaching have been small so far (3-4 people a time). It's become clear that making them much larger would dramatically reduce my ability to provide individual feedback. Most drawing classes I've been to were between 10-25 people. Improvement happened, but on a much longer timescale.

Mental-roadblock-space is large, and I can't cover all of it here. But I think if you can find a good mentor willing to work with you in a small class size (5 or less) then you (and they) can figure out in 4 hours of so if your roadblocks can be worked around.