Clerestory

Zettelkasten again

November 01, 2021

Just over two years ago, I wrote about a note-taking system called Zettelkasten. If you’re interested, you can read that post, and a follow-up I wrote a little later.

I am still using that system today. I’ve written 617,496 words in the system, and saved an additional 119,528 words in quotes. I’ve made 6,837 notes, connected by 27,629 links.1

This post is about what I’ve learned from persisting with the system.

Why take notes?

Why do I take notes? To learn, and perhaps so as not to forget, and to keep information accessible. But I’ve forgotten much of what I’ve learned and known. That includes many subjects on which I took extensive notes. Those notes are strewn in dozens of notebooks and various pieces of software that I used before the Zettelkasten, none of which I ever open. This loss extends to things I was once good at, but am no longer. In other words, things I knew how to do, as well as things I knew, or knew about.

Even deep, extensive, or embodied knowledge can be lost in just a few years, and sometimes less. Not everything is like riding a bike. I once brewed beer and baked bread, but now I could not do either without a significant review — right now it feels like it would amount to re-learning — and that’s knowledge that’s only about five years old. Decades ago, I was reasonably good at geometry and calculus. Now I might be able to remember the absolute basics, but I certainly could not solve the types of problems that I could when I was a teenager. That knowledge is effectively gone — though I wonder how readily it might be resuscitated, and how familiar it would feel.

My friends, knowledgeable in subjects quite apart from mine, report similar losses. There are vast swathes of things that they have learned and lost: knowledge gleaned, then unwittingly relinquished. Some say that what they once knew dwarfs what they know now, even when what knowledge remains to them remains considerable.

It is not just that educational testing incentivizes short-term learning, though perhaps that doesn’t help. But I don’t mind testing, and I could never find it in myself to oppose myself to cramming. Maybe that says more about me than it does about the method; I’ve always worked best in wild bursts. But I also think that forgetting is as important to learning as remembering is. Forgetting individual instances, for instance, is critical to abstraction, to acquiring concepts and principles. Learning then forgetting things (crammed or otherwise) probably has benefits. And of course not everything needs to be remembered.

But now there are things I do want to remember, and I want to remember them in detail. Often this is because I want to use the details as evidence. Abstract concepts and principles can become lifeless and unpersuasive when one can’t remember how one reached them. Providing detailed evidence allows others to arrive at their own conclusions, as well as to assess mine.

On research

I have inadvertently begun research in earnest, on subjects I plan to describe over the coming weeks.

When I started using the Zettelkasten, it was an experiment in note-taking. I thought it might help me retain a bit more of what I learned, or get a bit more out of books. It has absolutely done both. But it has also done much more.

If you’re doing a job, and you find a better tool for the job, you’re then faced with a choice. You can do the same job, but better, or you can undertake a bigger job. If in the process of building a shed, you gain access to reinforced concrete, you can build a stronger shed than you could have with wood. But with reinforced concrete you can also build a skyscraper.

The Zettelkasten has been like that for me. To some degree, I was searching for a tool, and the Zettelkasten filled that need, allowing me to pursue that project. But to some degree finding the tool precipitated a project, expanding what was possible. It has given me the courage to take on areas that might otherwise have daunted me.

The Zettelkasten allows me to choose what I remember. It’s not that it permanently installs facts into my memory. It’s much more process oriented.

Knowledge is enacted. If you train to run a marathon, and run it, that is no guarantee that you’ll be able to run a marathon for the rest of your life. It may be that this is true for many of the most important parts of life: use it or lose it. It certainly seems so with learning.

On process

What is the process of using the Zettelkasten like?

In short, it’s associative, and it forces metacognition.

In summarizing an idea, I paraphrase it. This is true of the content of the note and of its even-shorter title. Trying to change every word in a sentence while keeping the original meaning makes it very apparent which parts I haven’t understood. This in itself is learning by teaching, without waiting for a student. Rephrasing an idea into words that I can understand is like saying it in words a student could understand.

The fact that I link to other notes forces me to think about why this information is important, and how it is related. It also provides a trail back to how I got there, which can allow me to resume lines of thought from months or years earlier. I always try to note where I got the information, and why it is interesting.

Because the resulting short notes are in my own words, the system is trivially easy to search: I just search for the words that I associate with the subject. The structure is associative, not hierarchical or categorical, and notes are typically linked to several other associated notes — often people — which makes recall easy.2 If I am not working from an existing note, I typically create a note for the person I most associate with an idea. Often this means tracing things back to their source, learning about their works and books, and finding out who their influences are. This also means keeping notes for friends, to answer questions like “What was that thing that so-and-so mentioned?”

I do search the whole system constantly, and certain keywords and keyphrases (some coined, some purloined) become central to my thinking. Occasionally I’ll search a sub-section, like every note under a single person — files under 12/101 to search the 94 notes relating to Thomas Kuhn for example. But I search the whole thing (with ripgrep) much more frequently.

One problem I had with previous note-taking systems was that I never felt motivated to review notes. Very rarely would I flip through them, and they felt almost instantly stale. This system removes the friction of stale notes. By keeping notes atomic and short, they are easy to read. When I’m adding a new note (and taking notes on new stuff is exciting) then it is energizing to link it to old notes. In the process I review the old notes, and especially the list of links in those notes.

I used to spend long hours following wikipedia rabbitholes. I consider this to have been indispensable for expanding my knowledge of what’s out there. But to know of something is not to know it. Now, in plumbing the depths of a rabbithole, I don’t continue unless it is interesting enough to take notes. Or else I make a note to return to it later. This both keeps me more focused, means that I can’t lose afternoons reading things I won’t remember, and (again) that I can come back to the precise point where I left off. There’s a ratcheting effect.

Now I know that if I start a thread, I can pick it up again — even a year or two later. The Zettelkasten is a system of thinking in writing. It allows me to externalize the associative webs of thought themselves, which allows reinstantiating them in my mind. So I’m much bolder in going further, and less fearful of seeking out primary sources. If I don’t understand something at first, I trust that it will resurface again.

With a few exceptions, I don’t write full literature notes and then dutifully process them line by line anymore. I did that with a few books, but it can easily become drudgery. Now I do most of that work through Readwise, where I send notes from my Kindle and hypothes.is. This means much of my reading is digital now; for Darwin’s Origin of Species (1859), I’m in the process of shifting from notes in a paperback book to hypothes.is notes on the Project Gutenberg ebook, which is slow, but in this case almost undoubtedly worth it.

I realise I haven’t said anything yet about what I’ve actually been working on by using this system. But (perhaps characteristically) I thought I’d look at the meta/systems level before going into the content. The two are related.

I’ve written a bit more on the system as I explain why I try to trace things back to their sources.


  1. Just over 4 links per note, a figure which has slowly but steadily increased over time.
  2. Ahrens’ book How to Take Smart Notes explains this admirably.

Bryan Kam

I'm Bryan Kam. I'm thinking about complexity and selfhood. Please sign up to my newsletter or see more here.