Annotation as Thinking

Overview

Chapter 5 of Seeing With Fresh Eyes (printed pp. 121–130) makes a claim stronger than "label your figures." Tufte's chapter heading states it as a question — whether thinking simply is annotating the world — and pairs it with a warning: annotations are hard to unsee (p. 121). Annotation is not a service layer added after the evidence is finished. It is the act of reasoning about evidence, performed in the same visual field as the evidence, on content-responsive local grids rather than on a page-wide production grid.

The chapter's operating principle is stated plainly:

"Good annotation is like a knowledgeable expert/teacher at the viewer's side pointing and saying 'Now see how this works with that, how this might explain that . . .'" — Tufte, Seeing With Fresh Eyes, p. 121

The corresponding failure is equally plain: evidence that carries no annotation is not evidence, it is decoration. And the warning matters as much as the principle — once a reading has been pinned onto an image, viewers cannot un-see it. Annotation directs attention permanently, so it must be used with the same care as any other irreversible act.

Related skills: tufte-typography-for-data (annotations-on-annotations, letter-coded captions, direct labels), tufte-micro-macro-readings (the Talmud page as a micro/macro object), tufte-mapped-pictures, tufte-layering-and-separation.

§1. What Annotation Actually Does — Four Simultaneous Jobs

A good annotation is doing more than naming a part. Tufte's formulation on p. 121 makes annotation do two things at once: it calls out and explains the information, and it teaches the viewer how to read the display. Unpacked, that is four jobs.