• Updated 2022-05-10
  • Updated 2020-08-12
    • Put the introduction of Zettelkasten at the beginning

Introduction

Zettelkasten is a non-linear note-filing system originated by Niklas Luhmann, which he referred to as his “second brain.” or so-called a communication partner.

The term “Zettelkasten” is German for “slip box,” referring to the collection of hand-written notes stored in a box and interconnected through references, similar to the hyperlinking of the World Wide Web1 but happened on paper instead of the Internet back to the 1950s.

While similar in form to a personal wiki, the two systems differ in their creation and capabilities. A Zettelkasten can be implemented using a wiki, but not all wikis are Zettelkasten, most of them are not.

Niklas Luhmann

Niklas Luhmann was a professor of Sociology at Bielefeld University for 25 years (1968-1993) and published over 500 articles and books on a wide range of topics. He attributed his unprecedented productivity to his Zettelkasten system, saying in an interview:

I don’t think everything on my own, mostly happens in the slip box.

…

My productivity can essentially be explained from the slip box system.

…

The slip box costs me more time than writing a book.

Luhmann’s Slip Box

Luhmann kept notes on small slips of paper stored in a wooden box, resulting in two collections of 90,000 hand-written cards in A6 format, numbered consecutively.

Zettelkasten Photo of “Zettelkasten” from “Niklas Luhmann-Archiv” is licensed under CC-BY-NC-SA 4.0.

The notes consisted of his thoughts, arguments, and concepts, and were not just excerpts, below are some facts about them2 3 4.

  • Two collections in total
  • 90,000 handwritten cards in A6 format
  • Cards are numbered consecutively

  • The notes are not simply excerpts, it consists of his thoughts, theoretical arguments, and concepts

  • The first collection was created when he is a legal expert with interests in constitutional law and administrative sciences to a systems theoretical sociologist

  • Just before(1960 - 1961) he created the second collection, he spent a year at the Harvard School of Public Administration in Cambridge, MA(USA), where he attended lectures by Talcott Parsons, the leading sociologist in the field of systems theory at the time

  • There were no documents in the literary estate substantiating the claim that his visit to Harvard was the trigger to start a new collection of notes, but it seems obvious
  First Collection Second Collection
Created 1951 - 1962 1963 - 1997
Handwritten Cards 23,000 67,000
Content His readings in political science, administrative studies, organization theory, philosophy, and sociology sociology
# of Top-level Sections 108 11
Top-level sections state, equality, planning, power, constitution, revolution, hierarchy, science, role, concept of world, information, etc organizational theory, functionalism, decision theory, office, formal/informal order, sovereignty/state, individual concepts/individual problems, economy, ad hoc notes, archaic societies, advanced civilizations
# of Subsections   100
Section Separator a comma: , a slash /
Bibliographies. two, containing 2,000 titles one, incomplete, containing 15,000 references
Keyword Index 1,250 entries 3,200 entries
# of References 20,000 30,000
Average References nearly every card has a ref nearly every second card has a ref

The digital approach

Here’s my digital approach to Luhmann’s paper-based filing system in a modern(hopefully better) way, it tries to keep the benefits of the original approach like productivity gain and a solid accumulation of knowledge, the changes happen in the following three levels.

  • Physical level
  • Logical level
  • Slip level

It is recommended to get to know Luhmann’s slip box before reading the content below.

Physical Level

Paper Digital Purpose
Hand-written Paper Slips Plain text files To last a lifetime
Wooden Box A folder encapsulating the above files To last a lifetime, to store the slips and make it easy to find them for a human

Why plain text

Plain text files could be read and write by a wild range of applications after decades, while specifications of many other formats may changed a lot that your precious files created by an older version of the application are unable to be opened by the new version, or even worse, the application or format just cease to exist.

Folder form

A folder could be in any form, as long as it serves the purpose - To store the slips and make it easy to find them for a human. A simple folder on the file system is fine, if you are using DEVONthink 5 like me, a folder in it (which does not result in a single folder on the file system) works too.

Logical Level

Aspect Paper Digital Purpose
Numbering System At the top-left of slips e.g. 1/1b2c The file name and prefix of the title of plaintext files e.g. 1.0b2c.md An unique ID for every slip, and visual aid for human
References among Slips ID of the slip to be referenced written on other slips Markdown links of the slip to be referenced in other slips The key to keep relationships between slips

Numbering System

Unlike other digital approaches 6 7 that use a timestamp to name the slip, I choose to keep Luhmann’s way of numbering for two reasons:

  1. The timestamp is too long for a person to remember, thus failed to serve the purpose of a visual aid, which I regarded as a crucial help for one to make links manually.
    • However, the creation time of the slip could still be kept in the file.
  2. The sequence level of slips are obvious on the list of files in a folder sorted by name
    • So one can navigate easily between adjacent sequence levels, which is likely to be very close to the paper-based workflow done by Luhmann where slips stored together are close in sequence.

There are two tiny changes to Luhmann’s approach:

  1. Use . instead of /
    • The character / is not allowed in a file name for most file systems
  2. Use 1.0.md instead of 1.md for the first slip of every top-level subject
    • In this way, a normal sort by name could make the first slip(1.0.md) ordered before all other slips(e.g. 1.1, 1.2a, 1.3b) instead of after

References

The following markup works with DEVONthink5, for those who don’t use it, just make sure the application (e.g. a markdown editor) of your choice is capable to jump between files and use the supported markup to create a reference.

References should be written in explicit form:

This is a link to a slip [[1.0]]

This is a link to a slip 1.0

Implicit form like below doesn’t provide a visual aid for the ID of slips:

This link doesn't tell the ID of [the slip](1.0) to be referenced

This link doesn’t tell the ID of the slip to be referenced

Slip Level

A slip is a simple Markdown file, here is an example of 1.0.md:

---
Date: 2019-12-31 11:07:38
---

# 1.0 This is the title of the slip

The content of the slip goes here.

This is an example of a reference to another slip [[1.10a]]

This is an example of citation [1]


---

1. Here sits the bibliography

Beside the ordinary slips, there are three special kinds of slips in Luhmann’s collection.

Kind Paper Digital
Contents A slip contains a list of references to top-level subjects A single file contains links to top-level slips
Index A slip contains major subjects with references to related slips A single file contains links of different subjects to related slips
Bibliography A slip contains bibliographic information of literatures sorted alphabetically bibliographic information embedded at the bottom of every file

Auto generation

Contents could be automatically generated by scanning the entire slip box folder, but Index is mean to be created manually.

Bibliography

Bibliographic information is rather tedious for a human to write multiple times on paper, but it’s not a problem in the digital world with the help of applications like Zotero.

By embedding it in every single slip, the slip becomes more self-contained.

Takeaways

  • The selective links between cards made it possible to retrieve more than what was intended when the notes were initially taken 8
    • That’s why the slip box can “communicate” with human, for:
      1. One of the basic presuppositions of communication is that the partners can mutually suprise each other 9
      2. We have been educated to expect something useful from our activities 10
  • Without writing, there is no sophisticated thinking, since both short and long term memory of human brain are limited 9
  • The system reflects not only validated knowledge but also the thought process(via the linking), including potential mistakes and blind alleys that were later revised but NOT removed from the files, instead, a new card with revision may be added if needed 8

For a deeper understanding of Zettelkasten, check out The Zettelkasten Manual.

  1. Berners-Lee, Tim. The Original Proposal of the WWW, HTMLized. Mar. 1989, https://www.w3.org/History/1989/proposal.html. ↩

  2. Schmidt, Johannes F. K. “Der Nachlass Niklas Luhmanns – eine erste Sichtung: Zettelkasten und Manuskripte.” Soziale Systeme, vol. 19, no. 1, Jan. 2014. DOI.org (Crossref), doi:10.1515/sosys-2014-0111. ↩

  3. Luhmann, Niklas. Bibliographie 1. Universitätsbibliothek Bielefeld, 1 June 2017, http://ds.ub.uni-bielefeld.de/viewer/ppnresolver?id=ZL1Bibliographie1. Card Index Box. ↩

  4. Luhmann, Niklas. Bibliographie 2. Universitätsbibliothek Bielefeld, 1 June 2017, http://ds.ub.uni-bielefeld.de/viewer/piresolver?id=ZL1Bibliographie2. Card Index Box. ↩

  5. DEVONthink. DEVONtechnologies, LLC, https://www.devontechnologies.com/apps/devonthink. ↩ ↩2

  6. Schallner, Rene. Renerocksai/Sublime_zk. 2017. 2019. GitHub, https://github.com/renerocksai/sublime_zk. ↩

  7. Tietze, Christian. The Archive (MacOS) • Zettelkasten Method. https://zettelkasten.de/the-archive/. Accessed 31 Dec. 2019. ↩

  8. Schmidt, Johannes F. K. “Niklas Luhmann’s Card Index: The Fabrication of Serendipity.” Sociologica, vol. Vol 12, July 2018, pp. 53-60 Pages. DOI.org (Datacite), doi:10.6092/ISSN.1971-8853/8350. ↩ ↩2

  9. Niklas, Luhmann. Communicating with Slip Boxes by Niklas Luhmann. http://luhmann.surge.sh/communicating-with-slip-boxes. Accessed 27 Dec. 2019. ↩ ↩2

  10. Luhmann, Niklas. Learning How to Read by Niklas Luhmann. http://luhmann.surge.sh/learning-how-to-read. Accessed 27 Dec. 2019. ↩