What is the Zettelkasten method?

What is a Zettelkasten?

A Zettelkasten — German for "slip box" — is a system of short notes, each devoted to a single idea, tied together by explicit cross-references. Its purpose is not to preserve what you read but to surface connections the author never planned. The system is built to produce, not to file.

The difference from an ordinary folder of notes is the direction of travel. In a folder, you go looking for a note you know exists. In a Zettelkasten, you follow a chain of links and land on a note you had forgotten, which reframes the problem in front of you. Luhmann described his box as a conversation partner capable of surprising him.

Who was Niklas Luhmann?

Niklas Luhmann (1927-1998), German sociologist and systems theorist, published more than 70 books and close to 400 articles over his career — a productivity he himself attributed to his slip box. He started building it around 1951, while working as a civil servant in the Lower Saxony administration, and maintained two successive collections until his death.

The complete archive, roughly 90,000 slips, is held at Bielefeld University. A research project launched in 2015 and funded for fifteen years handles its digitisation and critical edition; the slips have been consultable online since 2019. The method reached English-speaking readers through Sönke Ahrens and How to Take Smart Notes, published in 2017.

What are the three types of note?

Note type	Lifespan	Content	What happens to it
Fleeting note	A few days	An idea caught on the fly	Processed, then discarded
Literature note	Permanent	What a source says, with the reference	Kept separately
Permanent note (Zettel)	Permanent	An idea restated in your own words	Linked into the network

The permanent note is the heart of the system and the only one that matters long term. It answers to three requirements: one idea, written in full sentences, and at least one link to an existing note. A permanent note written in shorthand will be unreadable in two years — Ahrens presses this point harder than any other.

How does the method work in practice?

1. Read with a pencil. Annotate in the margin instead of highlighting passively. Highlighting gives the feeling of work without producing anything.
2. Write the literature note. Summarise what the source claims, with page and reference. This note stays attached to its source.
3. Write the permanent note. Restate the idea in your own terms, as if explaining it to someone. If the restatement resists, you have not understood the idea.
4. Find where to plug it in. Go through existing notes to find the ones this extends, qualifies or contradicts. Create the links in both directions.
5. Write an index note. For themes that grow, build a hub note listing the entry points. Luhmann called these Schlagwortregister.

Where does the method break down?

The Zettelkasten is demanding, and the cost concentrates in step 4. Finding where to link a new note requires knowing what is already there. Past a few hundred notes, that knowledge exceeds working memory and the web degrades: recent notes link to each other while older ones fossilise into unreachable islands.

Second limit: the method was designed for a researcher reading text. It has nothing to say about three hours of podcast a week or a recorded conference talk, even though those formats now account for a growing share of the material people consume.

Third limit: it is slow by construction. That is a deliberate choice — Ahrens argues slowness is the precondition for understanding — but it makes the method hard to sustain alongside a full-time job.

Can you keep the spirit without the overhead?

What gives the method its value is connection, not manual data entry. Nothing in Luhmann’s idea requires a human to be the one who notices that a note from March intersects with something read in September — that was simply the only option available with a cabinet of drawers.

The cabinet had a flaw that software inherited without fixing: to find a slip, you had to remember the path you yourself had laid down. The digital version swapped drawers for full-text search, which moves the requirement rather than removing it — now you have to guess the exact words you used in a note written six months ago. The number that illustrates it: on a corpus of 948 syntheses embedded with bge-m3, the query "feudalism", with five genuinely relevant documents in the database, tops out at 0.510 similarity against 0.424 for a query matching nothing at all. Nine hundredths of a gap — and a fixed 0.65 cutoff returned nothing to a user who did in fact own their five sources.

Synthiz keeps Luhmann’s principle and takes on that specific point. Sources — including videos and podcasts, transcribed automatically — are synthesised and poured into a thematic "memory" where recurring concepts link themselves together, with sourced quotations attached; search fuses semantic and lexical scoring and judges relevance by a document’s distance from the distribution of its own query, not by an arbitrary threshold. You get back to the slip without recalling how you phrased it. What remains is what Luhmann considered the real work: deciding what the collision of two ideas means.