I’ve always had the idea that linear perspective is the correct way to do perspective. In art courses I learned about two-point and three-point perspective and their development in the Renaissance; in computer graphics courses, I learned about the pinhole camera model, and how consumer cameras typically use lenses to approximate pinhole cameras. These methods of linear perspective are supposed to make it look like you’re looking through a window. If you put one eye at exactly the right spot in front of the images, then it would be like looking through a window, at least in terms of the geometry of the scene.

And, often in the culture of art and in computer graphics, when artists do something different than strictly follow the rules of linear perspective, they are “deviations from standard perspective.” Nonlinear perspectives—which use different rules for projecting 3D shapes to 2D—are considered an artistic choice, a form of creative expression, whereas linear perspective depicts “objective” reality. Photography makes persuasive illusions; people talk about photography as recording reality.

Making my own pictures led me down a road to understanding how wrong these views are. There is no such thing as correct perspective; all choices of perspective have advantages and disadvantages. It is impossible to accurately portray everything about 3D space in a 2D picture, so artists must make choices, and linear perspective is just one option.

I began questioning perspective when comparing my paintings to photographs taken at the same time. This comparison often surprised me. For example, here’s a picture that I quickly sketched on a Winter day in Seattle: