Models of Reality

A fundamental activity of all living creatures is creating a model of reality: an internal, approximate representation of the external real world. When you and I walk through a forest, we use the data from our perceptions to construct a model of where the trees are, the nature of our footing, the shape of the terrain, and other features that may be relevant. We then direct our muscles according to the forest-model in our brains. If the model we have created is a good approximation of the real forest, we will be able to avoid walking into trees or falling down gullies and to successfully arrive at our desired destination.

The process is often so fast and innate we don't notice that it happens: when I open my eyes, I am not aware of the computations creating a model of my surroundings; I simply have an instantaneous sense that there is a person over there, a stream over here, some rocks and trees between them, and so on.

Our human, visual modeling of reality involves significant computational power and prowess, the domain of our animal brains. But non-animal creatures also model reality: a plant has a model of reality that, in essence, says, sunlight comes from the away-from-gravity direction, and water comes from the towards-gravity-direction. The plant (or more precisely, its genetics) is making a bet that its sunlight-is-found-away-from-gravity model is accurate just as we bet that our tree-over-here, stream-over-there model is accurate.

No living organism navigates the world directly: we plan our actions based on our internal model, constructed from our perceptions of, and assumptions about, the world. Our success in the world depends on the quality of our model.

Our models also include the nature and capabilities of our bodies—a part of the external world. When I walk into a dark room, I move my hands along the wall to find a switch and flip it on. I am working with a model that says I have limbs and muscles that will allow me to do these things. Under normal operations, our self-model is good and we don't notice that there is a difference between model and reality. But the distinction is revealed in the surprise we feel when we send an instruction to move an arm that is "asleep" and doesn't respond.

Reality is a many-faced and infinitely detailed thing. The models we create of it are, of necessity, limited and focused on aspects that are relevant to us. The models we create will consist of different features than those of other organisms: for example, while humans have a significant focus on their orientation in gravity, fish models pay little heed. Or, in contrast to earthworms, the genetics of birds contains a detailed model of aerodynamics.

The ability to create an appropriate model of reality is one of the central products of evolution. Only those creatures that make a sufficiently detailed, accurate, up-dateable, and updated model of relevant aspects of reality leave copies of themselves in their wake.