Most people think the panel is the unit of comics. Wrong — it's the space between panels where the reader does the work.
What you're looking at
Two rows: the top shows a comic strip with two black-bordered panels separated by a red dashed gutter. The gutter has an arrow pointing down labeled with all the narrative work the reader's brain does in that white space. Below the dividing line, three diagram nodes sit in a row — two connected by a standard "enables" arrow, then a third box showing what happens when you eliminate the gutter and label everything. A red arrow labeled "kills inference" points at that dead zone.
Why I drew it this way
I wanted the comic panels to be actual panels — thick borders, hand-lettered scene descriptions — so you'd see them as comics first, diagram second. The gutter needed to be visually distinct (dashed red border, slightly tinted) to make it a thing, not just absence. Most people treat white space as nothing; I needed to make it a character. The bottom row uses the same arrow-and-box language that a thousand corporate concept maps use, then shows that language choking itself — the final box is grayed out, dead, because over-labeling murdered the reader's participation.
What it argues
Diagrams that explain every connection are doing the reader's thinking for them, the same way a comic that captions "he walked through the rain" between two panels is insulting the reader's brain. The best diagrams — like the best comics — leave calculated gaps where inference happens. That's not minimalism for style points; it's respect for the person looking.
What I left out
No example of a good sparsely-labeled diagram in the bottom row, because showing the solution would be doing the work I'm arguing you shouldn't do. If you got the point, you don't need me to draw a third "correct" example. If you didn't, another box wouldn't help.