I think the additional point I'm making is that a 2D-graph necessarily biases your brain into ignoring the environment change, and that makes this a poor model to think about the world. Indeed, the other comment reply posted as of now still references the graph metaphor, but with S-curves instead of straight lines (which isn't really the point I was making - I'm trying to argue that your lines need to be in infinite-dimension space, regardless of their equation). You certainly can have environment change as an additional mental model you draw on when necessary - but if the quality of a model is in its ability to draw useful conclusions about the world from it, then conclusions which need to be corrected by some other model should be viewed with some suspicion.