In a comment on my post Visual Field, Bob Wong wrote “Finally some liberation from the dogma of composition.” Bob is touching on the heart of the matter.
Not long ago, I read a post on a blog, talking about things you can do to improve your photography. It was, in essence, a list of things you can do, at the time you make the exposure, with the goal of improving the photo you’re about to make. It was full of the conventional advice, such as “is this the strongest composition you can find?” and “pay attention to edges”. You know the list. It was an ‘Aesthetic Checklist’. I read it, and in my mind, I envisioned photographers, camera in hand, about to make a photograph but delaying releasing the shutter until they had run the checklist.
That led me to think that, perhaps, all we know about ‘composition’ is stuff that’s derived after the act. We look at a photo and we say “Oh, great composition, I love that strong diagonal element”. Or we might even look at a lot of photos (mine, for instance) and say “Oh, that Paul Butzi, he loves to compose photos with strong diagonals”. And this train of thought inevitably leads to thinking that someone might profitably think about such things at the time the exposure is made.
There’s a problem with this line of thinking, though. The problem is this: I have never, ever looked at a scene and thought “Oh, look, here’s a long log, I’ll use that to create a strong diagonal composition”. Ok, that’s not strictly true, because I have done it on occasion. What I have never done, though, is engage in that sort of thought and had the result be worth more than the little pile of dead flies that collects on my studio windowsill. Frankly put, I am just no damn good at all at making photos by assembling the composition from the available elements. I can only see the elements in a sort of post hoc examination.
Sure, I know. I can assemble photos that contain just a few discrete elements. I can arrange three circles on a plane, and I can arrange a line and a circle. I can arrange a single circle. Those are pretty limited skills. I am not a compositional dynamo.
Yet, despite this, I manage to make some photographs that I like where I know the photos exceed my meager ability to construct a composition. I can recognize the good ones and bad ones after I make them, but I can’t build them.
And that leads me to the conclusion that perhaps the ‘rules of composition’ are very vague and hard to articulate, or they’re very complicated and interdependent and thus hard to articulate. There are not very many books on composition, and the ones that exist always seem to lead me back to arranging discrete elements on a plane, and I already suspect that although this is a good way to post-analyze photographs it’s a cruddy way to make them. There’s a Weston quote in the back of my brain, the thrust of which was essentially this: if you’re thinking about the ‘rules of composition’ when you are making a photo, you have failed before you started. Following a checklist is not the path to strong compositional skills.
To me the good news here is that a) this explains why trying to read about composition does not seem to work for me, and b) despite the fact that humans are not very good at articulating things which are very fuzzy and vague, humans (me included) have a pretty good mechanism for dealing with things that are fuzzy and vague. In some sense we’re nothing more than a big fuzzy neural net. Dealing with fuzzy and vague is one of our strong points. The problem is not dealing with fuzzy and vague; it’s reducing our fuzzy and vague process to an unambiguous bit of language that’s the problem.
And at last I come to my point – articulating something is really important if you want to teach it, or if you want to have someone teach it to you, because in that case language serves as the intermediary between your thoughts and someone else’s. But this highly specific, unambiguous articulation is not needed if you’re just training yourself. You just learn it, and if the understand defies being expressed in language, well, that’s unfortunate but sometimes life is like that.
How do you train a fuzzy neural net? You present it with lots of inputs, and you tell it “this one is good” or “this one is bad”. How do you train your fuzzy neural net brain to make good (for any desired sort of ‘Good’)? You make a lot of photographs, and then you look at them and you tell your fuzzy neural net brain “This one is good. Make more like this. This one, though, that’s not so good. Don’t do that as often,” and you don’t waste time with trying to articulate why this one is good or that one is not so good.
Or at least, that’s my current working theory.