We See What We Know
Recently, I've been watching the television show Brain Games. In one episode, they demonstrated that human visual perception can accurately construct a scene from very few points of data -- for instance, a handful of colored moving dots on a black background can easily be recognized by most people as as two people dancing or two people fighting. However, when the moving dots are associated with a less familiar activity, such as two people sword fencing, it becomes much harder to interpret what the dots represent. It actually becomes difficult to tell that the dots even represent people. The take-home lesson from this seems to be that it's very easy to show people something that they expect to see and very difficult to show people something that they don't expect to see. You literally can't see what you don't know as well as you can see what you do know. I got to thinking about this as it applies to painting and art and realized that this same lesson does in fact carry over, but with a twist. Why, for instance, does Bob Ross teach people to paint trees and mountains and streams? Why are there never animals in his work? Or humans? When Bob Ross paints his trees and clouds he is never worried about making sure that the lighting in the scene has been figured out correctly, and yet I always recognize what it is that he is trying to paint. That's because he knows that your brain needs very few signals (dots) to instantly "see" a tree. This means that trees have a lot of room for error since our brains can kindly fill in lots of missing data -- and that makes trees and clouds good candidates for things that beginners can paint and feel good about. Animals, on the other hand, although they are familiar, are less ubiquitous than trees. If you try to picture a deer in your mind right now, you can do it, but if I ask you to picture the exact curve of the deer's legs the details are probably a bit fuzzy. Try to picture a cat's ears or tail, and you will probably come up with a clearer image. That's because you interact with cats more than you interact with deer, and so all the neural pathways leading to the idea of "cat" are a bit better connected. Unfortunately, the distance Bob Ross paints his scenes from precludes being able to see any cats, and it's not a good idea for Bob Ross to try and teach beginners how to paint deer by the stream since human brains need accurately painted cues to see the somewhat less familiar deer. Why doesn't Bob Ross add people to his paintings then since we are more familiar with them than deer or even trees? In one sense, the idea that people are more familiar to us than deer and trees is true. It's easy for our brains to take two dots and a half circle and instantly see a smiley face. A generic smiley face won't register as "realistic" though in the same way that a few dabs of green paint register as a "realistic" tree. It will register as "cartoony." This is because, to some extent, we don't need to identify very small differences between tree branches and one deer is pretty much the same as another deer to us humans give or take an inch. As soon as we can see that it's a tree or a deer, the brain stops classifying and stops to rest. Faces, on the other hand, are how we tell one another apart, and so it's very important for the brain to pick up on every subtle difference between one face and another face so that we can be socially successful. Thus, if a painter doesn't get tiny details right on a face, it will register as distorted, strange, or unidentifiable. In other words, faces need more "dots" than trees in order for our brains to read them right. In general, this idea of recognizing something from very few cues reminds me of Mondrian's trees (take a look at them here). His early tree is pretty representational, and his last tree is highly abstracted, and yet, most viewers can tell that the last three paintings in the series, despite how different they look, are actually the same exact tree.