Niklas Luhmann's Original Zettelkasten: Two Slip Boxes, Fixed Numbering, and Communication Partner

(Hand-drawn by Ernest Chiang. You might also be interested in his Ernest PKM workflow.)

1️⃣ Introduction: A Sociologist and His Thinking Machine

Niklas Luhmann (1927–1998) was a German sociologist renowned for his systems theory. During his academic career, he achieved astonishing productivity: 70 books and over 400 scholarly articles.

But even more remarkable, he attributed all of this to what seemed like a simple tool: the Zettelkasten (slip box).

• This was not an ordinary note-taking system.
  • Luhmann began building this system in the 1950s, eventually accumulating over 90,000 index cards.
• He called this system his communication partner
  • An external brain capable of dialoguing with him, facilitating thinking, and even “surprising” himself. ¹
  • Doesn’t this sound like a manual, century-old version of an AI Agent or knowledge assistant?!

In his famous 1981 paper “Communicating with Slip Boxes” (Kommunikation mit Zettelkästen), Luhmann described in detail how this system worked. But interestingly, much of the modern understanding of the Zettelkasten method is actually mixed with interpretations and adaptations by later scholars.

This note explores the original method that Luhmann himself actually used.

“Without writing, one cannot think;
at least not in a sophisticated, connectable manner.”

— Niklas Luhmann, Kommunikation mit Zettelkästen (1981)¹

2️⃣ Two Slip Box Systems

When most people talk about Zettelkasten, they imagine “one box full of cards.” But Luhmann actually maintained two independent slip box systems, each with different functions and structures. ²

2.1 Bibliographic Slip-Box

The first slip box was dedicated to recording bibliographic information. This box’s function was purely for indexing and locating, containing no personal thoughts or interpretations from Luhmann himself. It was like a personalized/personally curated library catalog system. (This reminds me of my childhood dream of having a small library at home…)

Luhmann’s card structure:

• Front: Complete bibliographic information (author, title, publication year, etc.)
• Back: Page index = “Page x has this concept, page y has that argument”

2.2 Main Slip-Box

The second slip box was the core of Luhmann’s thinking. This box was not a database, but a thinking tool. Luhmann emphasized that he only placed thoughts here that he had processed, digested, and reformulated in his own words.

What was stored here:

• Ideas reformulated in his own language
• Completely independently understandable notes (not dependent on original sources)
• Units of thought that could be connected with other ideas

Ernest PKM Workflow:

• I (might have?) implemented a similar dual slip box architecture in Heptabase
  • First temporarily store bibliographic entries in Journal, gradually organize them into independent cards (usually during morning routine after waking up)
  • Link cards together in Heptabase Whiteboard, generating new ideas when making connections (new cards)
  • Where I fall short is that when it comes to re-expressing in my own language, I face resistance (a.k.a. laziness)
• Bibliographic management:
  • Use Readwise to collect highlights from various reading sources
    • Import from Kindle, Safari Books, and other sources
    • Preserve original bibliographic information and page numbers (try to record source name and author)
  • Also use Heptabase Web Clipper to clip entire articles or /pdf into Heptabase to start reading and highlighting
• Main note system: Rewrite ideas in my own language in Heptabase
  • The goal is for each card to be an independently understandable unit of thought, though many are also paragraph-style material accumulations (Sokoban warehouse keeper?!)
  • Establish connection relationships through whiteboards
    • Some people like this “visualization,” but I prefer to interpret it as “connection-ization.” “Visualization” without careful reading and deliberation easily becomes superficial. To avoid superficiality, it’s like a challenge of playing with cards face-down—being able to reconnect and redraw the entire whiteboard yourself.
  • Use tags and attributes to assist searching, but don’t rely on classification
• This separation allows me to:
  • Quickly capture information (bibliographic slip box),
  • While also thinking deeply (main slip box).

2.3 The Relationship Between the Two Slip Boxes

There was almost no connection between these two systems. Their functions were clearly separated:

This dual slip box structure is a key feature of Luhmann’s system, often overlooked or simplified in modern interpretations.

2.4 Luhmann Built Two Complete Zettelkasten Systems in His Career

In reality, Luhmann built two completely independent Zettelkasten systems during his academic career, rather than continuous expansion of one system. The years and bibliographic counts are reference points, as there are no primary sources documenting exact timeframes. ³

First System: Zettelkasten I

• Period: 1950s - 1963
• Research Field: Law and administrative studies (during civil service)
• Scale:
  • Bibliographic slip box: 1,800 entries
  • Main slip box: approximately 23,000 cards

Second System: Zettelkasten II

• Period: 1963 - 1998 (death)
• Research Field: Sociology (after becoming scholar)
• Scale:
  • Bibliographic slip box: 16,000 entries
  • Main slip box: approximately 66,000 cards

Why Build Two Independent Systems?

In 1963, Luhmann transitioned from civil servant to sociology scholar, shifting his research field from law to sociology. He likely decided to build an entirely new Zettelkasten rather than expand the existing system.

Possible reasons:

• New research field required different conceptual frameworks
• Maintaining thematic purity of both systems, avoiding conceptual confusion
• The Folgezettel numbering system had developed specific contexts, difficult to integrate with new field
• Allowing each system to maintain its own “evolutionary thinking path”

This also illustrates that Zettelkasten is not just a “note-taking tool,” but a “thinking partner” deeply bound to specific research fields.

3️⃣ Workflow: Separating Reading and Reaction

Luhmann’s note-taking workflow had a unique characteristic: Luhmann didn't highlight in books or write annotations. ⁴

This sounds counterintuitive? But it might be the essence of his Zettelkasten?!

3.1 Step One: Record Bibliographic Cards While Reading

When Luhmann read a book or article, he would:

• Prepare a bibliographic card
• Write complete bibliographic information on the front
• Record page index on the back:
  • “Page 23 discusses autopoiesis of social systems”
  • “Page 67 proposes triadic structure of communication”
• Place the card in the bibliographic slip box

Key point: At this stage, he didn’t write his own thoughts, only recorded “where to find what.”

3.2 Step Two: Daily Re-reading and Reflection

This was the critical step. Luhmann would daily:

• Re-read that day’s bibliographic cards
• Think about these contents’ relevance (relationship) to his own thinking context
• Ask himself questions:
  • What does this idea mean for my research?
  • Which of my existing ideas is it related to?
  • Do I agree with this argument? Why?

3.3 Step Three: Rewrite in Own Language

If an idea was worth keeping, Luhmann would:

• Reformulate in his own language (not excerpt or quote)
• Write it as an independent card, which must:
  • Be understandable to his future self (not dependent on original context)
  • Contain only one idea per card
  • Be connectable with other cards
• Decide this card’s position in the system (numbering logic explained later)
• Establish connections with other cards

Luhmann called this process "separating reading from reaction". ⁴

Record while reading, react while thinking.

3.4 Why Not Highlight in Books?

Luhmann’s logic was:

Highlighting in books is merely “marking others’ ideas."
But rewriting into cards is “creating your own ideas.”

This “rewriting” action is thinking itself. If you can't express a concept in your own language, you haven't truly understood it.

Ernest PKM Workflow:

• I’ve been accustomed to highlighting since childhood
  • But I’ve tried quick scanning first—when quickly scanning, don’t highlight or take any other action (just drink water?)
• During reading after scanning: I highlight in e-books (or corresponding tools) (Kindle, Safari Books), or mark key points in Heptabase
  • These highlights automatically sync to Readwise, then into Heptabase
  • I still think during this reading stage (this differs from Luhmann’s behavior) (I haven’t decided whether to adjust to complete reading first, then set aside separate time to re-read and think) (but indeed, thinking while reading often requires switching between multiple devices or apps)
• Daily processing time (usually morning or during schedule gaps):
  • Open Heptabase Journal / Highlights page
  • Review all highlights imported from Readwise
  • Important step: Ask myself “What is this idea related to that I’m thinking about?”
  • Only highlights that can answer this question will (need to) be rewritten as Heptabase cards
    • Thinking another way (Working Backwards), what topics am I currently researching, then do related reading extending from those topics, so corresponding highlights are generated
• Rewriting principles:
  • Use complete sentences, not fragmented excerpts (try not to be lazy, don’t just copy-paste)
  • Sometimes accompany with handwriting, separately writing on reMarkable Paper Pro or Field Notes
  • Must be understandable to future self (not dependent on original context)
  • One idea per card
  • Note original source at beginning or end of card (for further reading)

This workflow transforms me from “collecting information” to “creating ideas.” Approximately (intuitively) only 5-15% of highlights convert to permanent cards, and this filtering process itself is digestion and thinking.

4️⃣ Numbering System (Folgezettel)

Luhmann’s numbering system is another fascinating part of the entire Zettelkasten worth studying. This system is called Folgezettel (follow-up notes). ⁵

4.1 Basic Numbering Logic

Imagine you start with the first card of a new topic, numbered:

1

If you have another completely unrelated new topic, numbered:

2

But if you want to continue the idea from card 1, the number would be:

1a

If you want to continue the idea from 1a, the number is:

1b

If you want to comment on, expand, or reflect on 1a, the number is:

1a1

4.2 Tree-Like Branching Structure

This system can expand infinitely, forming a “tree-like” structure: ⁵

%%{init: {'theme':'default'}}%%
graph TB
    %% Main sequence nodes - Concept level in Orange
    N1[1: Initial Topic]:::main
    N1a[1a: Continuing Idea]:::branch
    N1a1[1a1: Comment on 1a]:::subbranch
    N1a1a[1a1a: Further Extension]:::subbranch
    N1a1b[1a1b: Parallel Branch]:::subbranch
    N1a2[1a2: Return to 1a Continue]:::subbranch
    N1b[1b: Continue 1a]:::branch
    N1b1[1b1: Comment on 1b]:::subbranch
    N1c[1c: Continue 1b]:::branch

    N2[2: Another Topic]:::main
    N2a[2a: Continuing Idea]:::branch
    N2a1[2a1: Comment on 2a]:::subbranch
    N2b[2b: Continue 2a]:::branch

    %% Connections showing the branching logic
    N1 --> N1a
    N1a --> N1a1
    N1a1 --> N1a1a
    N1a1 --> N1a1b
    N1a --> N1a2
    N1a --> N1b
    N1b --> N1b1
    N1b --> N1c

    N2 --> N2a
    N2a --> N2a1
    N2a --> N2b

    %% Styling
    classDef main fill:#FF8000,stroke:#CC6600,stroke-width:3px,color:#000
    classDef branch fill:#FFB366,stroke:#CC6600,stroke-width:2px,color:#000
    classDef subbranch fill:#0080FF,stroke:#0060CC,stroke-width:2px,color:#FFF

• You can insert new cards at any position
• You can create parallel branches at any level
• The numbering itself records the evolutionary path of thinking
• Orange nodes: Main topics (1, 2)
• Light orange nodes: First-level branches (1a, 1b)
• Blue nodes: Deeper branches (1a1, 1a1a)

Ernest PKM Workflow:

• When I first started using digital note-taking software, I would deliberately arrange directory names with numbering (encoding)
  • Later discovered that numbering wasn’t very meaningful (except for fixed sorting order, but after switching to purposeful searching, the sorting order of information object names became relatively unimportant, while the relevance and connection strength between information objects and topics I currently care about became relatively more important.)
  • My own note system evolved from plain text markdown (.md) files accumulated in GitHub, to starting to use Joplin, Obsidian, Logseq, Heptabase
  • Now mainly using Heptabase as the storage hub,
    • Preprocessors include Readwise, Todoist, ClickUp, and other tools.
  • After entering Heptabase, during daily morning and monthly random reflection review organization periods, organize cards from Journal and Highlights into Whiteboards, constructing and connecting ideas.
  • Therefore, Heptabase Whiteboard became my forest, so I don’t spend much effort on numbering and encoding.
  • Part of my initial intention for this adjustment was hoping to reuse the same card, appearing in two or more whiteboards. But perhaps Luhmann’s initial intention was to have consistent paths walked when repeatedly organizing thoughts and connections in the mind.
  • In Heptabase, if you see a certain card, you can query Backlinks from the Info area below, which Whiteboards it appears in, and Tags. These features can also replace numbering and encoding. Other note-taking software has similar features, some complete and automatic, some requiring manual work.
• When I find myself obsessing over whether numbering and encoding are needed, I try to find Luhmann’s original intention: forming connections, forming network structures, not hierarchies.
  • So I infer that numbering and encoding are meant to capture thinking paths, not topic classification.

4.3 Difference from Decimal Outline Numbering

This is not traditional hierarchical outlining (1.1, 1.2, 1.1.1).

Traditional outlines are pre-designed structures, while Folgezettel grows organically. ⁵

You don’t need to know in advance how many sub-ideas there will be between 1a and 1b, the system will expand naturally as thinking progresses.

5️⃣ Fixed Position Ordering (Feste Stellordnung): Against Thematic Classification

Luhmann explicitly stated in his 1981 paper: ¹

“For the internal arrangement of the slip box, what is decisive is:
Against systematic ordering by topic and subtopic,
instead using fixed position ordering (feste Stellordnung).”

5.1 What is Fixed Position Ordering?

• Luhmann rejected traditional classification methods of filing cards by “topic → subtopic.”
• Instead, each card received a fixed number/position, which once determined remained unchanged, representing:
  • This card's fixed position in the system
  • This card's position in the evolution of thinking
• When inserting a new card, don't reorder existing cards, but use extended numbering (like 1a, 1a1).
  • The card’s content topic can “grow” new extensions without having to move the card into a fixed topic classification.

5.2 Why Oppose Thematic Classification?

• Luhmann believed preset topic classification had several problems:
  • Limits unexpected connections
    • When you put all “social systems” cards together, you’re unlikely to discover analogical relationships between “social systems” and “biological systems.”
  • One idea often belongs to multiple topics
    • Forced classification traps you in the dilemma of “where should this card go?”
  • Topics emerge afterward, not preset
    • Through connections and hub notes (explained later), topics naturally emerge.
• Although fixed position isn’t classified by topic, each card can have multiple references/connections, forming a “network structure” rather than “hierarchical tree.” In Luhmann’s system, the emphasis and value often come from “the card’s connections within the overall network” rather than its classification labels.
• Luhmann viewed the card system as “second memory (Zweitgedächtnis),” whose internal structure cannot merely be a “filing cabinet” but must have dynamic connection and extension capabilities. Fixed position ordering aids this dynamism.

5.3 Combination of Order and Chaos

Luhmann mentioned an important concept: ¹

“The combination of order and chaos is the source of creativity.”

• Order: Each card has a fixed number, can be found
• Chaos: Cards aren’t arranged by topic, encouraging surprises
• This “organized chaos” allowed him to frequently discover connections and ideas he had even forgotten while reading cards.

6️⃣ Connections: Analog Hypertext System

The numbering system is just physical arrangement. The real magic happens in connections.

6.1 Cross-References

Each card could have “references” to other card numbers written at the end:

This idea relates to the self-referential nature of social systems.

→ See 21/3d2 for discussion of self-reference
→ See 57/12a for blind spots of observation

These connections formed an “analog hypertext system.” ¹ Though physical cards, through number referencing, Luhmann could jump to related ideas like clicking hyperlinks.

6.2 Hub Notes

According to Johannes Schmidt (researcher at the Luhmann Archive at Bielefeld University), Luhmann created special “hub notes.” ⁶

Hub notes’ functions:

• Contained extensive lists of links pointing to core notes
• Like “highways” between topics, connecting notes from different fields
• Helped navigate through 90,000 cards

Example (hypothetical):

Card 21/1: Social Systems Theory—Main Concepts

→ 21/3d2 Self-reference
→ 21/5a System and Environment
→ 34/2b1 Triadic Structure of Communication
→ 57/12a Observation and Blind Spots
→ 89/4c Evolution and Selection

Ernest PKM Workflow:

• In my note system, the note unit closest to Hub Notes concept is Portal Notes.
  • After switching to Heptabase, some whiteboards replaced this portal note role.
  • In Heptabase, if you see a card, you can query Backlinks from the Info area below, which Whiteboards it appears in, and Tags.

6.3 Structure Notes

Schmidt also discovered that when Luhmann started a major line of thinking, he would create notes similar to “article outlines” or “book tables of contents.” ⁶

These notes provided a meta-structure for that line of thinking, helping him:

• See the overall theoretical framework
• Find missing links
• Plan future writing

6.4 Connection Over Classification

For Luhmann:

"Connection" generates more creativity than "classification."

• Classification is static, connection is dynamic.
• The same card can be connected by countless other cards in different ways, and each connection can generate new meaning.

6.5 Zettelkasten and Computer Algorithm Analogies

This is a modern retrospective thought experiment:

• Luhmann built his Zettelkasten between 1950-1998, when modern computer science was not yet widespread
• His method was entirely based on physical cards and manual operations, without reference to any algorithmic theory
• But looking back from today’s computer science perspective, Zettelkasten’s operational mechanisms have striking similarities with certain core algorithms
• This analogy is not just an interesting observation, but helps us understand why this system is so effective

Multi-dimensional Comparison with Major Algorithms

Technical Perspective Deep Observations

Viewing Zettelkasten’s core mechanisms from a modern algorithmic perspective:

• PageRank and Hub Notes:
  • Luhmann’s hub notes actually implemented a concept similar to PageRank.
  • Through quantity and quality of connections to identify important knowledge nodes.
  • Heavily cited cards naturally become authority nodes in the knowledge network.
• Graph Traversal and Folgezettel:
  • The Folgezettel numbering system is essentially a kind of predefined traversal path.
  • When you follow the sequence 1 → 1a → 1a1, it’s like executing depth-first search (DFS);
  • While cross-references provide the ability to jump to other branches, similar to backtracking in graph traversal.
• Hash Index and Fixed Position:
  • Fixed position ordering (Feste Stellordnung) provides O(1) query capability similar to hash tables.
  • When you know a card number (like 21/3d2), you can directly locate it without traversing the entire system.
  • This is the technical advantage of Luhmann’s opposition to topic classification.
• MapReduce Limitations:
  • This analogy is relatively forced. Zettelkasten is not a batch processing system, but a real-time, manually-driven thinking tool.
  • Although one could say “reading → thinking → writing cards” is somewhat like Map,
  • “Organizing connections → forming publications” is somewhat like Reduce,
  • But such analogies are too superficial and cannot capture Zettelkasten’s core value.

These analogies remind us: Luhmann, without computer science training, discovered through practice many principles later formalized as algorithms. Effective thinking systems often follow similar underlying logic.

7️⃣ Communication Partner: System as Dialogue

One of Luhmann’s most frequently cited concepts is viewing Zettelkasten as a “communication partner.” ¹

7.1 Not Just a Database

Luhmann said:

“The slip box can be viewed as a communication partner,
initially for technical and economic reasons in academic work.
But as the system grows, it becomes a true interlocutor.”

When you have tens of thousands of interconnected cards, the system begins to surprise you:

• Ideas you forgot you had written
• Suddenly discover unexpected connections
• Your own thinking’s evolutionary trajectory

7.2 Continuous Expansion and Revisiting

Luhmann would daily:

• Re-read some cards (not all, but randomly or by topic)
• Search for new connections
• Add new cross-references
• Discover new writing topics

This system is never finished, always evolving.

7.3 From Cards to Publications

All of Luhmann’s books and articles “grew naturally” from his card network. ²

Luhmann’s writing process was:

1. Start from hub notes or structure notes
2. Follow connections, collect related cards
3. Organize card sequence
4. Fill in missing arguments
5. Complete writing

Luhmann said: “I never start writing from scratch. I just take out thinking that’s already prepared from the slip box.”

8️⃣ Design Philosophy Behind

Let’s organize the core principles of Luhmann’s Zettelkasten:

8.1 Systems Theory Influence

Luhmann was a systems theory scholar, and his Zettelkasten itself was a practice of “systems”:

• Self-reference: Cards reference each other
• Closure and openness: System has boundaries (slip box) but continuously accepts new inputs (new cards)
• Evolution: System constantly evolves through connections and reorganization

8.2 Externalization of Thinking

Luhmann’s profound insight:

Our brains aren’t good at storage, but excel at connection and creation.
Leave storage to the slip box, keep connection and creation for the brain.

This is why he insisted on “rewriting in own language”—the process of rewriting is creating connections.

9️⃣ Important Note About Modern Interpretations

Before concluding, I must point out an important fact:

• Luhmann himself never used classification terms like “fleeting notes,” “literature notes,” “permanent notes.”
  • These are systematizations and extensions of Luhmann’s method by later scholars—particularly German scholar Sönke Ahrens in his 2017 book “How to Take Smart Notes.” ⁷
  • Ahrens’ contribution was abstracting and systematizing Luhmann’s practice, making it learnable and applicable to more people. But this also led many to mistakenly believe this three-part classification was Luhmann’s original approach.
  • I also mistakenly mixed the two together in my COSCUP 2021 PKM sharing.
• Luhmann’s actual practice was:
  • Two slip boxes (bibliographic + main)
  • Record bibliographic cards while reading
  • Rewrite into permanent notes in main slip box after thinking
• I will discuss modern interpretations (including Ahrens’ method, digital tool applications, etc.) in detail in another article.

🔟 Further Reading and References

Original Documents

• Luhmann, Niklas (1981). “Kommunikation mit Zettelkästen. Ein Erfahrungsbericht.” In: Öffentliche Meinung und sozialer Wandel / Public Opinion and Social Change, edited by Horst Baier et al. Opladen: Westdeutscher Verlag. English translation: “Communicating with Slip Boxes” available at zettelkasten.de.

Academic Research

• Schmidt, Johannes (2016). “Niklas Luhmann’s Card Index: Thinking Tool, Communication Partner, Publication Machine.” In: Forgetting Machines. Knowledge Management Evolution in Early Modern Europe, pp. 289-311. Leiden: Brill.
• Schmidt, Johannes’ research findings, see “The Money Is in the Hubs: Johannes Schmidt on Luhmann’s Zettelkasten,” zettelkasten.de

Digital Archive

• Niklas Luhmann’s complete slip box has been digitized and can be browsed at Bielefeld University’s website:
  • Niklas Luhmann Archive

Modern Interpretations

• Regarding Luhmann’s two slip box systems and workflow, see compilations from multiple sources, including Sönke Ahrens’ descriptions (though with his own interpretations). See “Luhmann’s two slip boxes” discussed in various Zettelkasten community articles
• “From Fleeting Notes to Project Notes – Concepts of Sönke Ahrens Explained,” zettelkasten.de
• For detailed explanation of the Folgezettel numbering system, see “Mastering Luhmann’s Zettelkasten: The Art of Note Branching” and multiple analysis articles on zettelkasten.de.

Epilogue

Zettelkasten is not just a note-taking system, but a philosophy of “externalizing thinking.”

Luhmann’s true contribution wasn’t the technical details of index cards or numbering systems, but the insight behind his approach:

Thinking needs external structure.
We don’t complete thinking in our minds, then write it down.
We complete thinking through writing.

In the digital age, we have Obsidian, Roam Research, Logseq, and other tools that can more easily practice similar ideas. But core principles remain unchanged:

• Rewrite in your own language
• Build connections rather than classifications
• Let the system dialogue with you
• Continuously evolve, never finished

This is Luhmann’s most valuable legacy to us.

If you’re interested in modern Zettelkasten interpretations, including Sönke Ahrens’ three-part classification (fleeting/literature/permanent notes), digital tool applications, and comparisons with other knowledge management systems, I’ll explore these in depth in future articles.

If you have articles on your website about understanding and practicing the original Zettelkasten method, feel free to leave a message to Ernest to exchange links. Or let’s discuss together on Threads | Twitter/X.