cognitive psychology
Mental model
The picture in someone's head of how a thing works. When the interface matches it, everything feels obvious; when it doesn't, nothing does.
The demo
A tiny stove. Each go, one burner lights up - turn it on by clicking the knob you think controls it. First the knobs sit in a row; then they mirror the hob. Three goes each.
The knobs:
What this demo shows (text version)
A burner on a four-ring hob lights up and you pick the knob that controls it. When the knobs are laid out in a row with an arbitrary mapping, you can only guess, and you mostly get it wrong - the famous failing of real stoves. When the knobs are arranged to mirror the burners, you get it right on sight, because the control matches the model you already hold of the device. That is the lesson: usability comes from matching the user's mental model, not from adding labels to an arrangement that fights it.
On the matching layout you did not learn the controls, you already knew them: the knobs sat where the burners did. That instant knowing is a mental model, and an interface that fights it makes simple things feel broken.
Norman's stove is the classic: four burners in a square, four knobs in a row, and a lifetime of slightly-wrong guesses. The fix was never a clearer label; it was arranging the controls to mirror the thing they control.
When users say they cannot figure your interface out, they usually have a perfectly good model - of a different interface. Meet the model they arrived with before you try to teach them a new one.