top | item 22085837

Luhmann's Zettelkasten

186 points| marvinblum | 6 years ago |emvi.com | reply

83 comments

order
[+] darkpicnic|6 years ago|reply
Zettelkasten has to be the least explained concept I've discovered online. Sure, they'll explain the _technique_ of it, but I've almost never seen any solid examples of it in _use_. I was so excited when I first heard of it, but I'm a person that learns by seeing examples and I couldn't find any good ones anywhere.

Does anyone have any good examples they've seen?

[+] rdiddly|6 years ago|reply
Any wiki is basically an example, isn't it? Or if I'm not quite ready to defend that statement, I can at least say, you could use (or maybe misuse) any wiki-builder to build a Zettelkasten. Just impose a length restriction on each page's content.

But I don't know how much of the advantage of the Zettelkasten is tied to its being in the physical realm. On the one hand, the more physically rich an experience (of browsing or traversing or entering info for example), the more it will tend to stick in memory. On the other hand one of the things computers are good at is remembering indexes and linking them to other places where that index was mentioned.

[+] learningerrday|6 years ago|reply
Gordon Brander's pattern wiki is probably the best contemporary example I can think of that implements the Zettelkasten pattern: http://gordonbrander.com/pattern/

I think part of what can make the concept hard to see in practice, is that it's both the Zettelkasten artifact, and your methodology for adding and linking concepts when you're learning

[+] killjoywashere|6 years ago|reply
I’ve kept a personal wiki since about 2007 for my notes on medicine and pathology. The “structure discovery” function it enables is very similar to Zettelkasten. I would submit the affordances of physicality can’t be replicated in software without AR.
[+] chrisweekly|6 years ago|reply
I've had precisely your reaction, and am stoked to share my enthusiasm for Roam (https://roamresearch.com) -- a graph-based system for codex / memex / personal knowledge-base / zettelkasten++
[+] coolswan|6 years ago|reply
Tbh, I haven't seen a real-life example that perfectly clones this method but I've adopted parts of the system that makes sense for my own workflow. I think that's how it should be approached. Luhmann developed this in a non-digital world after all.
[+] jacobolus|6 years ago|reply
This kind of method used to be very common among scholars, before computers.

You can probably still find some in use among retired professors who started their careers 60 years ago and already had a system in place when computers started to replace physical organization systems.

Nowadays there are still many people who use digital equivalents.

[+] jboynyc|6 years ago|reply
I used to assign a short piece by Luhmann on how to read that describes his note-taking technique with some concrete examples.

At the time I used my own translation (based on the text in Short Cuts published by Zweitausendeins), but I see somebody else has now translated (most of) it here: https://takingnotenow.blogspot.com/2007/12/luhmann-on-learni...

[+] baxtr|6 years ago|reply
I think the internet as such was thought of linked document, which comes very close to that concept
[+] keeprunning|6 years ago|reply
ha, yeah very muich in the same boat. been following some of the zettel blogs for years and have read an ungodly number of posts, but it still hasn't entirely clicked — that said, I've moved entirely to plain text for note-taking and it's been a godsend moving out of evernote.
[+] coolswan|6 years ago|reply
How to take Smart Notes is a very good book that explains this technique well. Highly recommend.

https://www.amazon.com/How-Take-Smart-Notes-Nonfiction/dp/15...

[+] smartmic|6 years ago|reply
Also from me a clear recommendation. I have read the original in German, it is very good to read, entertaining and yet has depth. Suitable for all those who want to structure their work or thoughts better.

Btw, there is an own homepage of the book: https://takesmartnotes.com/

[+] neurocline|6 years ago|reply
I'm just finishing this book, and I echo the recommendation. In fact, when I bought the book, I didn't know it was about Zettelkasten. This book won't satisfy those looking for concrete examples, but it will give you a lot of specifics and whys. Obviously, I'm in the early adopter stage at the moment, but what I'm reading fits my problem space (tons of notes, but I can't seem to find what I want, and I re-do research periodically because of it).
[+] nextos|6 years ago|reply
Seconded, it's a really well written compact book.

It's possible to tweak deft and org-mode a bit with some functions to make it very Zettelkasten like: https://efls.github.io/zetteldeft

[+] lethologica|6 years ago|reply
I've tried to keep a Zettleksaten several times now with no real luck. My most recent try was using TiddlyWiki, with each Tiddler being an entry, usually containing a single piece of information. Perhaps it's the way TiddlyWiki is laid out, but I didn't particularly find the filing system to be of any real benefit to me. After my 30 day experiment I had nearly 1,000 Tiddlers and the majority were interlinked. When I would click through the hyperlinks I wouldn't necessarily have any particular 'a-ha' moments that these articles on Zettlekastens typically try to convince you will happen when you use this filing system.

Perhaps I went about it all wrong, but I loved the theory but didn't find the particular application useful. I also found absolutely no examples of a real life Zettelkasten online anywhere. None. I couldn't even compare to see if I was doing it wrong because it seems to be the sort of thing that everyone writes about but no one practices.

More so, I absolutely wouldn't want to use a new third party app for my knowledge repository. Who knows wether or not it'll still be in development or even supported in a year.

[+] kiwicopple|6 years ago|reply
I used Trello with relatively good results. The process was:

1. Create a board with a single “Dump” column

2. Create cards in the Dump column (or “share” to the column whenever I read something online)

3. Every week, categorise the card into new columns. Each column was a broad category like “Philosophy” and “Startup ideas”

4. When the board became unwieldy, with too many columns, I created a new board (Philosophy) and split the categories/columns again (Stoicism and Nihilism)

The act of reorganising the board was super useful for remembering everything, a bit like spaced repitition

[+] bocklund|6 years ago|reply
What are you using now?

I've been using TiddlyWiki for almost a year. Most of my 400 Tiddlers are either code snippets with explanations or an probably ~2 paragraphs of text on average. I haven't experienced any galaxy brain type moments. The search is good enough, I can add and link to images, and it supports KaTeX for math. Those are probably the killer features for me.

If anything, I'd love to switch to something that let me work on iOS natively, but I've been happy enough to not look too hard for anything else.

[+] gpanders|6 years ago|reply
I implemented my own Zettelkasten tool as a simple bash script. I use the shell as my interface into my notes (create a new note, list notes, open a note, etc.), Vim to write new notes, and grep to search contents and tags. If, for any reason, I want to view a note in the browser, I use pandoc to convert into HTML.

The things I like about the Zettelkasten method is the concept of atomicity (notes are short and simple and are about just one thing) and then linking related notes together. Keeping notes plaintext also makes them easy to use with other tools (such as grep). I then just sync my notes in a Nextcloud instance across all of my different devices. Out of the many different note management methods I've tried this one is so far my favorite.

[+] ThouYS|6 years ago|reply
Cool, reminds me of "Roam Research" which calls itself a tool for "networked thought" [1]

[1] https://roamresearch.com/

[+] dangirsh|6 years ago|reply
+1 For Roam. I'm a long-time Org Mode user, but was always waiting for something like Roam to take its place. Roam is a great example of how getting a few key features right can launch a tool into a league of its own.

Implementing the Zettelkasten is an explicit use-case for Roam, as documented here: https://roamresearch.com/#/v8/help/page/VURQiVZQR

[+] crucialfelix|6 years ago|reply
I started using Roam last month and it's been very effective for me. The linking between documents literally stitches all my work together and let's me flow and wander, improving and adding each time.
[+] uberstuber|6 years ago|reply
Roam is heavily influenced by Zettelkasten, it's the first digital version I've seen that works well
[+] lkrubner|6 years ago|reply
Prior art against Google’s Page Rank:

“ By connecting lose dots of information you create interconnected clusters of knowledge which emphasize the importance of an information based on the number of connections”

[+] kmill|6 years ago|reply
Though PageRank has the innovation of weighting the connections based on how important the information is.

(Just for fun, here's one linear-algebra-free way to understand classic PageRank: imagine starting from a random page on the internet and then randomly clicking links. If you continue clicking for long enough, then the probability that you end up on any given page is, more or less, its PageRank. If your access patterns with Zettelkasten are random-ish and you keep track of how many times you look at each card, then these counts are a crude approximation of the PageRank of your cards.)

[+] dredmorbius|6 years ago|reply
Zettelkasten are one of a set of related tools for knowledge management, information tracking, and creativity facilitation.

They've been submitted several times on HN, though usually to little discussion: https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...

Among the more active mentions, a submission of mine a couple of months ago: https://news.ycombinator.com/item?id=21208196

I'd also recommend "How to organize personal knowledge", an "Ask HN" with 220 replies: https://news.ycombinator.com/item?id=17892731

There are also index cards generally: https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...

And Pile of Index Cards (POIC): https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...

And of course, the digital version, Hypercard: https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...

I've recently discovered Paul Otlet's work -- among his creations was the notion of "documentation", a term he invented ~1925. He'd also compiled a collection of over 15 million index cards, the Mundaenium. Recent submission of mine: https://news.ycombinator.com/item?id=22043441

I'd count bullet journals as closely related:

https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...

See: https://news.ycombinator.com/item?id=11856987 and https://news.ycombinator.com/item?id=18769286 especially.

I've been wrestling with how to make use of a research journal of sorts, adopting some of the concepts from a bullet journal, but also the flexibility of index cards, while leveraging the benefits and capabilities of a bound paper codex. I'm calling this a BOTI Journal, for "best of the interval" (week, month, year...). It's something of a collection of paper-based round-robin databases (a ring or circular buffer), where I note the most significant elements enountered in a period of time, and periodically roll these up to a longer-term aggregate. This gives time-ordering but also an aggregation function.

Things tracked include various documents (expansively defined as per Otlet as articles, books, Web pages, audio and video recordings, images, etc.), authors, concepts, journals, etc.

After a year, or several years, I should have a well curated set of most-useful / most insightful references over the period, a notion I've been frustrated in my inability to track over the past few years.

More generally, the notion of a rotating database file seems hugely useful, see the 43 folders "tickler file" method from Getting Things Done: https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...

(Searching "43 folders" itself tends to capture many distantly-related links to the website of the same name.) I've newly embarked on this, and am interested to see how it goes.

There are also other pages interspersed, concepts, ideas, themes, outlines of cataloguing systems (Dewey, LoCCS, Otlet's Decimal Notation, etc.)

I'm discovering that my recent thought that a circular file format ought to exist and be used by ToDo / scheduling systems ... has been proposed by others:

https://medium.com/@criticalmind/gtd-todo-tickler-proposal-f...

[+] Schiphol|6 years ago|reply
This is fantastic, thanks a lot
[+] wortelefant|6 years ago|reply
After trying several formats, I now keep my zettelasten-structured notes in https://Notion.so. you can use webclipper, search and internal links etc, previews can be seen as cards. This approach is well documented with many examples in the Zettelkasten forum, in this post https://forum.zettelkasten.de/discussion/404/moving-on

I found that the value of these systems for me is highly dependent on the amount of engagement I put into it in practice - if I engage eith the 'cards' often, updating them or reordering them, the content and the links of the card are much more present in my mind and generate new ideas. Compared to this factor, the system used for linked notes is of secondary importance.

[+] tunesmith|6 years ago|reply
It's unclear whether being able to insert a new identifier within an existing list of identifiers is a feature. That might just be a byproduct of needing to put the new card into a related area.

The Zettelkasten.de software appears to be pushing concepts that would have been impossible in Luhmann's system.

Ultimately it seems that all these information systems are just different implementations of nodes, edges, and containers.

Node: Do you want it to support formatted text and/or media? Do you want to be able to visually explore this node's edge relationships to other nodes? (Software often chooses one or the other; both would be nice.)

Edge: Do you want directed edges? Do you want to support Graphs and not just Trees? Do you want your edges to have configurable meaning beyond "this relates to that"?

Containers: Do you want to be able to collect multiple nodes together? Do you want a node to optionally be a container, meaning do you want to allow an edge to link a node to a container? Do you want to support non-hierarchical (Venn) containers, aka tags?

Outliners, Todo list software are almost always directed trees instead of directed graphs. Even with mindmapping / bubble graph software - it's weird how many of them are just undirected trees instead of undirected graphs. Even orgmode is at best a directed tree with tags - it's not a DAG unless you use something like org-brain, which has its own limitations. Wikis usually don't have the visual exploration tools - imagine being able to zoom out of a wiki and just draw edges and containers to put the nodes (articles) in.

Zettelkasten supported (somewhat) formatted text. Not really able to visually explore edge relationships without taking stuff apart. Undirected graph, generic edge meaning. Links probably often one-way and not self-healing (removing a card wouldn't remove links of all linking cards). Supported heirarchical containers, and a card could conceptually link to a container by having "2" link to "2a" which might include 2a1, 2a2, 2a3, etc. Supported tags, sort of, although there wasn't a way to search for all cards that had a particular tag.

Ultimately you find that different thinking styles require different combinations of these concepts, so there isn't a one-size-fits-all, and there isn't "one system" that works like how the brain does, at least not until we have a system that allows us to easily mix these concepts together into a custom implementation.

[+] blumomo|6 years ago|reply
While the authors promote their own software solution for building your Zettelkasten, I am happy a user of (and not affiliated with) https://bear.app which I use on my iPhone and Mac for organizing my plenty notes for software engineering, meetings, management notes, my Psychology study notes and anything private as well.

In analogy to Zettelkasten I put my notes into different (nestable) boxes what they call hashes. The app eco system is quite hacker friendly as they publicly document the underlying SQLite database. I don't that possibility because their apps help me organizing my notes well enough til this point.

[+] zerof1l|6 years ago|reply
Great article. However, I don't like the idea of using online platform for this. Maybe if it was offline app I would try it. I haven't found a perfect app for this. So far, just simple text file worked the best. Effortless, easily searchable, open in any app. I'm considering upgrading to using very plain HTML files (so that I could embed media) and folders. IMO that's all you need.
[+] pzumk|6 years ago|reply
I’m from Lüneburg, Germany and have never heard about Niklas Luhmann nor have I seen any Zettelkästen.

I unfortunately have nothing else to add to the topic of Zettelkasten, but if anyone is in Hamburg or near Lüneburg and wants a little tour through our lovely town, let me know!

[+] jaynetics|6 years ago|reply
If you have any interest whatsoever in liberal arts ( Geisteswissenschaften), you should really have a look at Luhmann. Maybe peek into "Einführung in die Systemtheorie". I found him to be one of the most insightful sociologists of recent times.
[+] marvinblum|6 years ago|reply
Well, now you have to visit Bielefeld if you'd like to see one :) But Hamburg and Lüneburg are also nice places to visit! I highly recommend!
[+] metaodi|6 years ago|reply
Nice explaination of the concept, I really didn't know about that. All those links between cards remind me of Linked Data. Should be fairly easy to integrate that in WikiBase/Wikidata.
[+] mirimir|6 years ago|reply
Back in DOS, I used IZE and then Agenda. IZE was closer to Zettelkasten, I think. But Agenda was more flexible, albeit harder to learn.

I was rooting for Chandler, but so it goes.

[+] colordrops|6 years ago|reply
Surprised there are no mentions here of org mode yet.
[+] stewbrew|6 years ago|reply
Well, all these suggestions are nice and this also looks like a nice finger exercise but you really want to start something like this only with software that's still useable somehow in 50+ yrs. So just ignore this proprietary stuff and use plain text files + grep (or something similar).
[+] kungtotte|6 years ago|reply
You can still use a nice frontend to the data. Vimwiki, tiddlywiki, zim wiki, etc. all store their data in plain text but let's you create these connections trivially that Zettelkasten is all about.