Most explanations of the Zettelkasten method stop one step short of the thing people actually need: what does a note look like? The theory is easy to agree with and hard to act on. So here are three real notes, in the order they'd be written, from a single afternoon of reading.
The source: a paper arguing that people who take notes on a laptop perform worse on conceptual questions than people writing longhand.
1. The fleeting note
Written in ten seconds, on a phone, mid-paragraph.
laptop note-takers transcribe verbatim → worse on conceptual qs. it's the processing not the medium?
That's it. No title, no links, no polish. A fleeting note exists to survive the next twenty minutes, and its only quality bar is that you'll understand it tonight. Most of these get deleted within a day, and that's the design — the capture habit is worth nothing if it costs you the thought you were having.
2. The literature note
Written the same evening, with the source in front of you, in your own words.
Mueller & Oppenheimer (2014), "The Pen Is Mightier Than the Keyboard"
Students taking notes by hand did no better on factual recall but clearly better on conceptual questions. Longhand is slower than typing, so writers can't transcribe; they have to select and reword. The authors attribute the gap to that forced processing, not to handwriting itself. Telling laptop users "don't transcribe" did not close the gap.
p. 1163 — verbatim overlap with the lecture predicted worse performance.
(Later replication attempts have been mixed — the effect is smaller and less reliable than the paper's popularity suggests.)
Two rules did the work here. Own words — rewriting is what tests comprehension, the same mechanism behind the generation effect. And one source per note, so the note stays traceable. Note the last line: a literature note that records the paper's weaknesses is worth three that don't.
3. The permanent note
Written the next morning, from the literature note, without looking at the paper.
Constraint on input speed forces comprehension
When capture is slower than delivery, you can't transcribe — you must decide what matters, in real time. The bottleneck does the teaching. This reframes several note-taking findings: longhand's advantage isn't the pen, it's the enforced deficit of speed.
If true, the same effect should be reachable without handwriting: any rule that caps how much you may write (three bullets per section, one sentence per slide) should produce a similar gain — and any tool that makes transcription effortless should erase it.
← [literature: Mueller & Oppenheimer 2014] → connects to: [desirable difficulties], [cognitive load], [progressive summarization]
This is the note that makes the system worth running. Look at what changed between step 2 and step 3:
| Literature note | Permanent note | |
|---|---|---|
| Subject | What the paper said | What you now claim |
| Title | The source | The idea, as a statement |
| Lifespan | Reference material | Grows and gets revised |
| Test | Is it accurate? | Is it mine, and does it connect? |
The title is the tell. "Mueller & Oppenheimer (2014)" is a filing label. "Constraint on input speed forces comprehension" is an assertion you can agree or disagree with — which is exactly what makes it linkable to notes from other fields. That's the atomic note principle doing its job, and it's why the third note is the one you'll still be using in five years.
What to copy from these three
- Write the fleeting note badly. Speed is its only feature.
- Never paste a quote into a literature note without rewording it. The paste feels productive and teaches nothing.
- Give permanent notes titles that make a claim. If the title is a noun phrase, you've probably written a summary.
- Add the links while writing, not in a tidy-up session later. The connection is obvious now and invisible in a month — this is the whole argument for backlinks in the first place.
- Record your doubts. A permanent note that says "if true, X should follow" is a note that can be tested.
One honest note on scale: this pipeline is real work, roughly twenty minutes per idea, and it only pays off if you're writing something. If you aren't, evergreen notes or plain smart notes give you most of the benefit for a fraction of the ceremony.