Against Zettelkasten

Knowledge management
1

The title above is rather opinionated but it is possible that I misunderstand what Zettelkasten really is. If I don’t, though, my opinion is simple: computing makes a Zettelkasten system useless.

(Let it be clear that I’m talking about the original Luhmann system, since Zettelkasten seems to have become synonymous of many things.)

Let’s go back to the ancestor: the paper PKM system. Such systems have existed for centuries (I may only cite Placcius and his De arte excerpendi from 1689). The historian Alan Macfarlane has explained why such systems may collapse. With the increase of volume, indexes become harder to maintain, categories become more confuse, redundancy appear, etc. At first these systems grow more and more fruitful, but at some point they reach the highest point of the curve, and the usefulness of the system starts to decrease. Because of its complexity, a law of diminishing marginal returns applies, as in economics. You spend more time thinking where to look at, each information takes longer to store, redundancies multiply, you forget how to access this or that, etc. The system has become cumbersome. If it is well designed, it may take a few dozens of thousand notes to reach that point, but it’s a failure nonetheless.

Now, if I understand well, Luhmann’s file design was based on two assumptions. First, he didn’t want his file to be a mere repository, he wanted it to somehow mimic a conversation partner. It needed to provide some surprise. Second, not only was he OK with having an imperfect system, but he tried in fact to design an anti-system. He didn’t want a perfect store where he could find back everything, on the contrary he acknowledged and tolerated a certain amount of loss. The Zettelkasten is a way to trace certain inroads within the notes. If I may say, you abandon a certain amount of order-thus-comprehensiveness in order to gain a certain amount of randomness-thus-serendipity (some indexes are still needed though!).

I understand the Zettelkasten as a personal interpretation of what a paper file could be. But the fact is, computing dramatically expands the possibilities. With a software such as Obsidian, classifications can be multiple, can overlap, and can be updated. Many things can be automated. It becomes very difficult to lose an information. As for me, I have about 4500 notes, and it is very rare that it takes me longer than a few seconds to find back something. When it happens, it is generally possible to remedy the problem very quickly.

So, what’s the use of Zettelkasten? I don’t see any reason to give strange numeric titles to my notes. I can maintain a certain classification system, and put another on top, and mend them, trim them when needed, and still follow many more paths than Luhmann could, because more random and arbitrary branchings can coexist with an otherwise well ordered system.

This my understanding, but I would be interested that you prove me wrong.

2

What exactly do you think defines the Zettelkasten?
Your last paragraph gives me the idea, that it’s about the structure of the notes.

To me the importance lies in a deeper processing of whatever you learn. Usually just reading something gives you an idea you’ll lose with time. Processing information in any way, may it be through writing notes, drawing a picture, thinking really hard or taking to someone takes effort. The Zettelkasten is just one approach. The beauty is, that it mimics on how humans actually think.

There is a talk by Vera Birkenbihl where she talks about that information is stored in the brain in form of a network. Other people have the same idea, so if you’re not german there’ll be literature accordingly, since this approximates the idea in neuroscience on how information is stored. In this talk she mentions, that in order to make people care about whatever you talk about and remember, you’ll have to find a way to “hook” into this network of information within the listeners brain. This is actually the reason one is supposed to “tell a story” at the beginning of presentations. It’s a primer, people can relate and the information “finds its place”.

So in using the Zettelkasten-Methode you’re actually doing this quite explicitly. By using your own words to describe whatever you read somewhere, or thought for your self, you process it deeply. And by linking it to ideas that already exists in your own brain, you “fixate” it there. So it takes longer for you to forget and if you ever find this idea again, you may be surprised, but at the same time you are familiar with the information, reducing the time to understand it again. It’s basically this “yeah, i learned this in university and school. Give me 5 minutes to think about it and i’m good”-effect.

The “discussion”-part of the Zettelkasten gets more and more obsolete, the more AI advances. It’s really about contextualizing the information and ingraining it in your brain. If the Zettelkasten is not your tool for that, that’s fine.

1 Like
3

We’re not talking about the same thing. You’re talking about some learning method you call “Zettelkasten”, I’m talking about systems made of notes in general - systems that allow, not only to learn, but archive information, find it, link it, develop new ideas, etc. - and, in particular, I’m talking about Luhmann’s, though he wouldn’t call it a system. He was a scholar and his main purpose was to find help while building his theories, not learning already existing ideas. That’s why his peculiar way of linking notes is supposed to play the role of conversation partner, especially when you go back to your repository with an idea in mind and want to try it.

N.B. My understanding of what he called Zettelkasten is based on his paper Communicating with Slip Boxes.

4

Why do you think, computing makes a Zettelkasten system useless? Because of searching? Because if that is your issue, it is entirely up to you on how to use the zettelkasten.

While the “strange numeric titles” where a necessity for the physical system to work, you are right, they are not necessary for a digital one. But they are just a tool for the whole system. And the purpose of the system is not to write weird titles, but to contextualize knowledge. If you use, for example tags and for whatever reason two completely different topics intersect at one tag, you can thing on what the connection could be, do some research and close the gap.

I’d say migrating the zettelkasten as described by Luhmann to the digital world might be to short sighted. It’s about taking the core idea, not the mechanics, and delevop a system to implement them.

5

For Luhmann the branching, and thus the 3.4a.1.2 style titles (though in theory it could be non-numeric identifiers as well), was the core idea of Zettelkasten. If you don’t do this you don’t do Zettelkasten.

So, your reply helps me reformulate what I mean :

  • Either people use the trendy word “Zettelkasten” but in reality do something else.
  • Or they use a method that in reality is obsolete.

(As for why I think computing makes a Zettelkasten useless, the answer is in my first post, though formulated in a rather abstract way maybe.)

6

Zettelkasten is a practice for developing your own thinking.
The Zettelkasten becomes useless to the extent that thinking in our daily lives becomes useless.

7

You seem to miss the core-idea of Zettelkasten.

The 3.4a.1.2 style titles are indeed unnecessary, but they were important in their historic context. But they are just the implementation, a technical solution for the paper medium, not the principle.

The core idea of the zettelkasten is to create a non-hierarchical, evolving network of thoughts. A structure that grows unpredictably. A thought system needs persistent relations between ideas. How to achieve that is secondary, but i’d argue the most important thing is the links between Zettels.

In the article you shared he mentions three advantages, that i would consider to be closer “the need” then to the “strategy”:

  1. The possibility of arbitrary internal branching
  2. Possibility of linking
  3. Register (a way to find what we read and where to find it)

Non of them are strictly dependent on the headers.

I’d argue

  1. Either people use the trendy word “Zettelkasten” but in reality do something else.
    1. Probably. But not, because they do not have weird headers, but because they just collect information instead of transforming it. And because they might have implicit or explicit hierarchies.
  2. Or they use a method that in reality is obsolete.
    1. If someone still uses the weird headers: maybe. There are arguments for and against them.
    2. Using a physical Zettelkasten: maybe. it’s a lot of work. But some people like physical interaction.
    3. The philosophical core of Zettelkasten: No. That is not obsolete. It could be argued that AI makes it obsolete, because it as i conversation partner enabling unexpected ideas. But it could be argued that AI makes it even more necessary, because if one relies on AI, reasoning is externalized, thus Zettelkasten is a way to still practice it.
8

You seem to miss the core-idea of Zettelkasten.

Probably I’m not clear enough or you don’t read me with enough attention, but ok…

Anyway, my point is more about the relation between the “core idea” and its implementation.

You say “The core idea of the zettelkasten is to create a non-hierarchical, evolving network of thoughts.” Ok, let’s put it that way. You also seem to say that “The possibility of arbitrary internal branching” is key to the implementation, and I would agree.

A more classical system would be more classification-based and hierarchy-based.

Remember that my point of departure was material systems with paper slips, drawers, etc.

Now, today’s computing allows to get the best of both worlds. That was the essence of my first post. You can have at the same time a hierarchical and non-hierarchical system, because different modes of organizing information can overlap.

So, where is Zettelkasten anymore? in the “core-idea”? what is that? Creating “non-hierarchical, evolving network of thoughts”? I guess you can use a composed German word beginning with a Z and with double consonants to say that. But it’s a bit like saying “Weltanschauung” whenever you want to say “point of view”.

(Also, in that case, on the implementation side, for most people “using Zettelkasten” would just be synonymous with “using wikilinks”.)

So is it in the implementation? But if you stick with Zettelkasten strictly speaking, based on arbitrary branching, you end up with a rigid system that is not as satisfying as it could be. The rigidity induced by classification is only replaced by the rigidity induced by arbitrariness.

(Even with paper this dichotomy wasn’t total, as you seem to think, but it existed, and in the first case it could even bring the collapse of the system; in the second case, Zettelkasten, you trade comprehensiveness against serendipity; see my first post.)

Here we are. People keep on saying they’re using a Zettelkasten, but strictly speaking they don’t, or in a very vague sense. Or if they do well… I really don’t understand why.

This brings me to:

If someone still uses the weird headers: maybe. There are arguments for and against them.

What are the arguments for them?

I can try another general reformulation: why, with a tool like Obsidian, would one want to stay with a system totally or mostly based on arbitrary branching?

9

So what i gather from our discussion thus far is, that your understanding of Zettelkasten seems to be “the exact method Luhmann proposed”, including the folgezettel-style branching. And you claim is, that this exact method is obsolete. Correct?

In that cases, I agree that digital systems can expand or replace parts of Luhmann’s implementation.

Where i disagree is with the idea that Zettelkasten is identical with the specific implementation rather the problem it tries to solve. Otherwise, any innovation would automatically become an entirely new method rather than an evolution of an existing one. That seems like a difficult position to defend: for example, digital libraries did not stop being libraries just because they replaced shelves and paper indexes.

One argument for the numbering system could be, that it explicitly preserves the development of thought. If one Zettel leads to another, the folgezettel number records that.

Hyperlinks can be in a double role here. They can preserve the same relationship between notes the numbering does. But they can also create connections to unrelated areas, as he used them. One limitation of a paper implementation is, that updating the zettel, when new information is appended, is tedious and prone to error. So the digital implementation is an evolution that solves this issue. But that does not stop the whole thing from being a Zettelkasten, since the function of the Folgezettel is still preserved.

why, with a tool like Obsidian, would one want to stay with a system totally or mostly based on arbitrary branching?

If you create hyperlinks in Obsidian, you implement arbitrary branching.

10

I do think the term “Zettelkasten” is more meaningful in its historical context, with its specific technical implementation.

I understand you want to consider it as a general idea that can be separated from its implementation. I don’t think it’s the best thing to do - see my previous message.

I don’t think your metaphor with libraries is valid. I would rather consider two technologies with the same purpose, with one superseding and replacing the other, as for instance horse carriage and car.

One argument for the numbering system could be, that it explicitly preserves the development of thought. If one Zettel leads to another, the folgezettel number records that.

If the purpose is to build new knowledge, I don’t really see the interest, at least as a central feature.

But that does not stop the whole thing from being a Zettelkasten, since the function of the Folgezettel is still preserved.

You’re free to call “Folgezettel” what people usually call a “branch”!

You’re giving me various illustrations of what I’m saying. Yes, you can consider “Zettelkasten” as an idea separable from its implementation. But if the result is calling “branches” “Folgezettel” and “linking ideas when I feel like it” “Zettelkasten”… maybe it’s not super useful.

If you create hyperlinks in Obsidian, you implement arbitrary branching.

I was speaking of “a system totally or mostly based on arbitrary branching” - which is what I understand Luhmann had to do, with the means he had, in order to implement his vision as best as possible.

11

"I’m talking about Luhmann’s, though he wouldn’t call it a system.

Luhmann refers to his slip box as a system in the essay you’re referring to:

“And if one has to write anyway, it is useful to take advantage of this activity in order to create in the system of notes a competent partner of communication.”

Other places Luhmann refers to the slip box as a system:

“This requires the addition of randomness (Zufall) into the system—randomness in the sense that the agreement of the different comparative schemata is not been fixed, or that the information which is transmitted by communication is correct, but rather that this happens (or does not happen) “at the occasion” of communication.”

One of the primary themes in Luhmann’s work (here and beyond) is investigating communication as a system. The essay you’re referring to is an experiement in understanding the slip box as a form of communication, and in turn as a system.

12

"For Luhmann the branching, and thus the 3.4a.1.2 style titles (though in theory it could be non-numeric identifiers as well), was the core idea of Zettelkasten. If you don’t do this you don’t do Zettelkasten.

“If you don’t do [alphanumeric / non-numeric] branching, you don’t do zettelkasten” is not attested to in Luhmann’s essay. In fact, Luhmann explicitly states there may be other methods to achieve the same goal:

“A slip box, which has been made according to the suggestions just given can exhibit great independence. There may be equally apt ways to reach this goal. The described reduction to a fixed, but merely formal order of placement and the resulting combination of order and disorder is, however, one of these ways.”

He clearly states a slip box can achieve the same goal through other methods. His method is, “however, one of these ways.”

13

I think the issue is that you’re treating Zettelkasten as a storage technology rather than as a thinking method.

Digital systems solve many of the practical limitations of paper systems: search is instantaneous, classifications can overlap, links can be updated, and information is much harder to lose. In that sense, yes, a computer largely removes the constraints that shaped Luhmann’s physical archive.

However, that does not automatically make the underlying principles obsolete.

The purpose of Luhmann’s Zettelkasten was not to optimize retrieval efficiency. If that were the goal, a database with powerful search functions would always win.

The point was to externalize thinking: writing notes as autonomous ideas, connecting them deliberately, and allowing unexpected relationships to emerge over time. The famous “conversation partner” metaphor refers to this process, not to the physical cabinet itself.

In other words, digital note-taking and, today, AI are useful tools, but tools do not automatically replace methods. The advent of the automobile did not eliminate the benefits of walking or running; it simply changed the available options.

Likewise, Obsidian does not replace Zettelkasten any more than a word processor replaces writing. Obsidian can perfectly host a Zettelkasten, but the value comes from the practice of creating atomic notes, building explicit connections, and developing lines of thought over time.

If someone only needs to retrieve information quickly, then a Zettelkasten is probably unnecessary. But if the goal is to cultivate a long-term dialogue with one’s own ideas, the method can remain relevant regardless of whether it is implemented on paper or digitally.

14

I was referring for example to this passage:

For the inner life of the card index, for the arrangement of notes or its mental history, it is most important that we decide against the systematic ordering in accordance with topics and sub-topics and choose instead a firm fixed place (Stellordnung). A system based on content (like the outline of a book), would mean that we make a decision that would bind us to a certain order for decades in advance! This necessarily leads very quickly to problems of placement, if we consider the system of communication and ourselves as capable of development. The fixed filing place needs no system. It is sufficient that we give every slip a number which is easily seen (in or case on the left of the first line) and that we never change this number and thus the fixed place of the slip.

15

“If you don’t do [alphanumeric / non-numeric] branching, you don’t do zettelkasten” is not attested to in Luhmann’s essay.

The system that came to be known as Zettelkasten relies on arbitrary branching, and it has been implemented by Luhmann with the alphanumeric (mostly numeric) titles we know… There could be other methods, so what? It doesn’t change anything to the fundamental idea, and to what I was saying.

16

You and @patbem seem to share the same position.

I think the issue is that you’re treating Zettelkasten as a storage technology rather than as a thinking method.

No.

The point was to externalize thinking: writing notes as autonomous ideas, connecting them deliberately, and allowing unexpected relationships to emerge over time.

Ok. You can call this Zettelkasten if you please. I don’t really understand this fetishism of a name when the thing has changed dramatically, but let it be so.

(The thing has changed because Luhmann needed to rely on a fixed order that is no longer needed. It’s a qualitative change.)

That said, the discussion is a bit running in circles.

17

That was two replies in. :man_shrugging:

18

Zettelkasten is not merely a label. Behind it lies a set of principles that are applied consistently in order to achieve specific outcomes. Some of these principles are clearly explained in Luhmann’s publication that was referenced in this thread. They are not mandatory to follow, but if they are applied consciously, they produce specific results.

You can use folgezettel even in a digital system, if you want to obtain some of the same dynamic Luhman had, and if you are comfortable using it. Or you can do something different: understand what the Folgezettel is trying to achieve and implement those dynamics in another way. For example, I structure my notes differently. But it is still Zettelkasten, because Zettelkasten is grounded in a set of principles, not on implementation details, and Folgezettel is just one of several possible implementations of those principles.