No, Descartes, You Don't Build Reality Like a Lego Set

You have seen a lot of rooms in your day, and so when you enter a fresh new room you have a sense of what to expect.

Imagine you met up in a friend's kitchen because you are going to go on a kayak expedition together. They say: I left my phone by the lamp in my office upstairs, could you run get it for me while I pack up the lunch? It is the first door on the right at the top of the stairs.

Now, how do you do that? Predictive processing says that, before you are even conscious of it, you start to produce scenarios of what the stairs and upstairs hallway and office and lamp might look like and, essentially where you'll need to cast your eyes, put your feet, and eventually identify, reach and grasp the phone. And this 'start to produce scenarios' is the same activity as preparing to lift your foot.

You do NOT walk to the base of the stairs, take in the individual sensorial streams and construct a fresh new representation of the room, the way an engineer's drawing software might construct a map of the hallway—from the bottom up. And then calculate how you want to move. Using your prior assumptions, you will have formed a hypothesis before you have any new data—back in the kitchen—and then you tweak it based on what your senses tell you, if you care enough and feel a big need to do a lot of tweaking. And your tweaking is the preparation to move and vice versa.