Skip to content

You have almost never met anyone.

You have met what they said, what they did, what someone told you, what you remember, what you expected, and what you were afraid of. Then you built something out of it and have been talking to that.

↑ and back into the model

Which is where it stops being a process and starts being a loop. The model decided what you did. What you did shaped what came back. What came back is now evidence for the model.

None of this means the person is fictional.

They are entirely real and they are not made of your evidence. The model is yours. The person is not.

That distinction is the whole thing, and it is easy to lose in both directions. Drop it one way and you get the claim that people are constructions, which is false and convenient. Drop it the other way and you forget you are running a model at all, which is the ordinary condition and much more common.

The model is not the problem. Models are how anyone thinks about anyone. Forgetting it is a model is the problem.

Everything else on this site is a rehearsal for this page. Icarus cannot be hurt by being read wrong. A dataset does not care what you trained on it. An electron has no stake in which of two words you use.

The place where the model is not the thing, and the difference costs something, is here.

Epistemic status: analogy. A resemblance between domains. It has a boundary and the boundary is stated.And this is where the project’s own comparison is weakest. A person responds to being modelled. Nothing else in this investigation does. That asymmetry is on the claims page and it is the reason the person/dataset comparison is marked as dangerous rather than as true.