‘Cheap paper is less perishable than gray matter, and lead pencil markings endure longer than memory.’ —Jack London

There is little contention that writing is a useful tool when working, whether the purpose of that work is recalling, assimilating, collating, constructing or critiquing. This is an essay examining not whether notes should be taken, but how. This is not an essay extolling zettelkasten but, in a vague sense, rejecting it. This is not an essay appealing to readers to reject zettelkasten either; it is an attempt at communicating why that method proved ineffective for me. It is also my reflection on what does work instead and why. And an essay showing others what to consider when designing a bespoke note-making system that might similarly be effective for them.

Part I

The trouble with decontextualising

When Luhmann worked, computers were not the norm. His index card system therefore made a lot of sense. But there is a reason why he is in the minority and it is not, as is often believed, because his approach was genius in a way nobody else had stumbled upon it. Rather it was because he had devised a method that worked for him. And it might not have worked for others.

A key element of Luhmann’s zettelkasten – and also of modern, hyperlinked systems – is decontextualising. For Luhmann it became important to store newfound knowledge in tiny, independent bits because that made it easier for him to start with a specific idea and work backwards to rediscover the various sources from which he had first obtained that piece of knowledge. In other words, he could re-contextualise this information.

I have come to believe that whoever first came up with the idea of hyperlinking ‘atomic’ notes was building on a knee-jerk reaction to Luhmann’s system; a flawed digital equivalent of a decent analogue system. If Luhmann could retain context and somehow still find similar relevant, associated pieces of information, I doubt very much he would have still used index cards or, in the modern day, wikilinks. We will return to this line of thought after a brief digression in the opposite direction.

The benefits of decontextualising

There are some instances, or better still some fields, where decontextualising can be beneficial. If you ever want to maintain snippets of code that you return to frequently (I do this myself) it can help to extract that code, note down your remarks and save it independently of any larger purpose.

If you study law or medicine where you might want to maintain your own compendium of quick references, decontextualising can prove remarkably useful.

In other fields too, decontextualising can help give you a bird’s eye view of your field. In the history of science I like to keep my own timeline of events relevant to my work that gives me a thoroughly different, far more grounded, perspective of the past than what learning history or science episodically would have done.

But learning, by and large, cannot occur entirely decontextualised. Context is important not only to ensure you gain a fuller picture of what you had previously written but also because there is no way to guess whether you have fully grasped the context years after you first wrote something down and, over time, forgot portions of it. By retaining knowledge in context you both make it easier to pick up the context again and guarantee that your do, in fact, have the entire context.

Working with the end in sight

If you read a hundred papers and a dozen books this year, building up 112 notes in total, a quick search for some word across all of those 112 notes – a feature built into most decent text editors – will show you everything you have about that search term along with all the context you need.

Luhmann did not have this. There was no way you could leaf through eighteen boxes of files and loose sheets with scribbles, scratches and handwritten notes in a meaningful, pleasant or efficient manner. So he had no choice but to extract concretised pieces of knowledge in a manner of flagging key ideas and ‘points of entry’ into his larger body of knowledge. In other words, Luhmann had to supplement his amoeboid working knowledge with a formalised practice of information gathering and organisation to not lose track of the former.

Understanding this was, for me, the missing piece in the puzzle. I am not in the business of gathering information for the abstract future. I use my knowledge now and maybe reference some of my past reading again in the future, but mostly read anew. The field moves, new baselines are created, things change fast, and I see little use in archiving information as evidence of my having acquired them at some point.

My writing does not begin after the fact. My writing happens while I am reading. What I read informs my work and what little reading I do to get started with my writing is often temporally close enough to my writing that formalising its organisation is quite wasteful. I would rather make notes for the thing I am writing at present and since I will have saved that anyway, all that information – with its context – is just a multi-file search away.

Write here, write now

My note-making always occurs in the context of a piece of work I am looking to complete. In other words, I usually have the end in sight. I know I am currently working on something so the notes I make are written in my words (of course) but also in a manner relevant to the current work in progress.

Of course I read more generally too, without the aim of producing a piece of work out of it, but I do not take too many notes when I do this. It introduces the sort of friction I cannot possibly benefit from.

For non-academic works, if there is a particularly compelling piece of information I save the entire source to my reading list. For academic works, I toss it into my reference manager with a note for quick reference and the annotated PDF where the highlights tell me all I need to read when I re-visit that paper. Beyond this, I see little benefit in going through the tedious process of filiing away a million ‘atomic’ notes.

I am not in the business of information gathering. I am required instead to think and work flexibly with knowledge, building on it as I go. That means ideas I am actively engaging with remain in my head, months from when I first encountered them, in a very different form from what they originally were. The ‘original’ knowledge is now much less meaningful than my acquired understanding of it. And because I write about this evolving understanding of that original piece of information in a new context, a multi-file search will not only show me the original but also the much richer idea from six months later, all the while allowing my ideas to remain in context.

Rediscovering context will also help me make new connections. That the system helps him make novel connections is one of Luhmann’s famous claims about his own zettelkasten system. By retaining context we have effectively simplified this very behaviour, and done it without using wikilinks. Luhmann famously wrote that his zettelkasten thinks for him (I paraphrase). This is usually the single most popular evidence presented by fans of zettelkasten in support of their outsourcing part of their own creative thinking to their note-making system. To set the record straight, here are three other quotes from Luhmann that suggest other things too. First, that the point of his linking was in fact to help rediscover the context (‘die Verweisungen dürfen nicht nur die Leitgesichtspunkte aggregierten Sammelbegriffe erfassen, sondern müssen das unter ihnen gesammelte Material selektiv wegziehen’); second, that the relationship between notes helps when one is searching for something (‘Es ist danach wichtig, dass man nicht auf eine Unmenge von Punkt-für-Punkt Zugriffe angewiesen ist, sondern auf Relationen zwischen Notizen, also auf Verweisungen zurückgreifen kann, die mehr auf einmal verfügbar machen, als man bei einem Suchimpuls oder auch bei einer Gedankenfixierung im Sinn hat.’) and while this relationship appears more directly in wikilinks, I believe it is this very directness which is misleading and that search results with their immediate context visible is a richer source of information for drawing connections or regaining lost understanding; and third, that a zettelkasten is not about collecting information but reviewing and pondering over it (‘Zettelkasten als Klärgrube – nicht nur abgeklärte Notizen hineintun. Aufschieben des Prüfens und Entscheidens – auch eine Tempofrage.’) which goes to show that the much-adored habit of collecting thousands of notes and intricate graphs is much less the purpose than fully understanding why a piece of knowledge is relevant – something retaining context can help with to a great extent.

The one good habit

As a practice zettelkasten does help inculcate a particularly good habit: returning frequently to your notes. The trouble, though, is twofold: one, most people do not do this at all; and two, people who do practise this, at least as I have seen it, tend to focus more on certain topics over time than others as a response to increasing cognitive load. This is perceived as developing niche interests but in fact it is a symptom of working without context, purpose or as a matter of formality.

Indeed the practice of zettelkasten forces you to view note-making as an organisational formality – which is ironic since its core tenet, in the digital age at least, is remaining unorganised. When people take notes without an end goal they tend to make the note-making process the very purpose of information consumption. When they read, they view slotting new entries into their zettelkasten as the outcome of their reading and as markers of progress. This should not be the case. This is zettelkasten transformed into a make-work project. The outcome should be something else (for which you are reading) and your note-making should aid in that. The only reason those very notes find further use is because you do not, and need not, discard them afer your current project is done. Nevertheless, your notes were born of a specific purpose.

When I read more generally I almost never make separate notes, choosing instead to highlight or annotate in my bookmarking service1. These too can be searched through and exist as an informal companion to my academic reference manager. The point is, if I need something, and if I want to recall something, whatever I need is never more than a quick search away.

Finally, revisitation does not only have to occur in the presence of an unorganised system. Luhmann did speak of this (he writes ‘Verzicht auf vorschnelle Systematisierung und Abschliessung, Zukunftsoffenheit’ i.e. A rejection of premature systematization and conclusion, openness to the future). But remember that this was in a time before computers when people were presented with a mutually exclusive choice between either organisation or freeform thinking. Today, without physical notes, we have surpassed this limitation. We can easily establish many-to-many relationships because the information we store is abstracted away from index cards of yore. Organising our information and exploiting our notes are no longer mutually exclusive decisions.

Part II

My note-making approach

All that I have discussed so far are my thoughts based on how I work and what makes sense to me. They may not work for you, but at the very least they may prompt you to think about things from perspectives you might not have considered before and change (or reinforce) your own workflow.

My note-making approach now is as follows. Within a universal folder I have—

1. Quick access files
2. Current project folder
3. Readings folder
4. Other folders

In addition to these I also use a reference manager (Zotero) and a bookmarking service (Readeck).

Quick access files

My quick access files include an ___INBOX.md file, an __IDEAS.md file, a __READING.md file, a _DISTRACTIONS.md and a _TODO.md file. These should be self-explanatory but a couple of points are worth noting: these are the only files in the root of my universal folder; these files are not intended for permanent storage, nor are their purposes set in stone.

Everything goes into ___INBOX.md unless I can recognise another location right away. This means even tasks might go into this file and not into _TODO.md if it is not immediately apparent to me that something is a ‘task’ (e.g. is reading an abstract before a seminar a task or reading, because I might read the rest of the paper or closely-related papers too).

When something is immediately apparent, it goes into a more apt file as their names indicate. Random thoughts and web lookups go into _DISTRACTIONS.md for example. But none of these files have a structure. I might not even use checkboxes for the _TODO.md file or I might leave several lines between two notes. Whatever works for me in that moment goes.

These quick access files eventually get cleaned up, usually daily or weekly. Tasks go into my Calendar or Reminders apps daily, ideas and inbox notes may get cleaned up weekly and so on.

Current project folder

There is always a ‘main project’ I am working on, at best two, never more. Everything associated with this exists inside my current project folder (and its sub-folders). Once I have finished working on these the folder gets re-filed as-is into one of my ‘other’ folders (see below).

The usual culprits of this folder include my main paper, where I do all my writing and is usually an md file (or a series of md files, one for each chapter, depending on the size of the project) and a ‘structure and ideas’ file where I track the outline of my writing and any ideas associated with it that I get while working on the project but which I am unsure where to introduce. This is different from my __IDEAS.md quick access file which takes ideas related to my larger work life, my profession, my side projects etc., not ideas highly specific to my primary work in progress.

I have a third file imaginatively called RESEARCH.md which includes all my notes from readings associated with this project. It is a single text file arranged in some sensible order e.g. chronologically, alphabetically or, very rarely, thematically. This is dictated by the project itself.

Readings folder

This is an interesting one and was originally my zettelkasten. Now, it solely contains what zettelkasten practitioners might consider ‘literature notes’ but in a bit of a reverse fashion compared to a zettelkasten workflow.

As I said already, I make notes with the end in sight. So if I am reading something for my current primary project, my notes go straight into my RESEARCH.md file in my current project folder. If a source ends up with an enormous set of notes, I later extract it into its own file in this ‘Readings’ folder while leaving behind a link to it inside the original RESEARCH.md file. This is rare, but not impossible, and in any case is something I do only after the current project has been wrapped up.

The only other important file in my readings folder is my references.bib which Zotero (via BetterBibTeX) keeps updated. There is no reason why it should be here but has been here for a long time and multiple TeX files use a relative link to it so I continue to keep it here.

Other folders

These are folders that I use simply because I like organising my work in folders. They could be for conferences, papers taught for a term, supervision notes, a log of my supervision (because apparently Cambridge will not pay me without a log) etc. I will not say much about these folders because they are unlikely to be of much interest to others.

Within each of these are notes, once again taken with the end goal in mind. So I might have a single note for, say, twelve reading group meetings because they are so closely tied to one another. And in case of something especially important I make notes by hand and re-visit those pages to organise them into typed notes. This is friction by choice when I want to think spend more time thinking about certain topics.

Wikilinks

It should be apparent by now that I do not use wikilinks. There is no reason why but I find much less use for it now that I do not use zettelkasten. Wikilinks, however, are a useful feature in my opinion, which is why I do use the concept of linking to other files but with full file:// URLs instead. This takes a couple of seconds extra but literally only a couple of seconds: a combination of the fact that I do not use links much and that BBEdit, my text editor of choice, offers a pretty quick means of copying full URLs means the process is not too bothersome.

The reason I use file:// URLs is not only because BBEdit does not support it but also because not using a zettelkasten system means my files are spread across multiple folders on my computer (which also aids my ‘spatial’ memory in navigating across my system with Finder) and full URLs are universal and future-proof.

Environmental influences

BBEdit played a big role in how my note-making changed the way it did because of the many handy features it offers. It would by no means have single-handedly convinced me to change anything, but as someone who wanted to change my note-making approach knowing BBEdit worked as it did encouraged me to freely try out what works for me.

First, BBEdit is a macOS native application. It feels like it is part of the system and the nagging feeling I had when using Electron apps (like Obsidian) is now gone. This is inexplicable but experience tells me it is true.

Second, BBEdit is stable no matter what the file size. I have had Obsidian corrupt a file for me once in the past. I am almost certainly in the minority here but it did happen and once is once too much when text files are concerned.

Third, BBEdit has excellent autosave, versioning and recovery (with Git built into it) and a handy ‘notes’ feature where you can quickly get started with an eternally persistent note (or many such notes) that you never need to save or close until you choose to.

Fourth, BBEdit handles keyboard shortcuts brilliantly. I can ⌃ ⇥ between my current note and sidebar and recent notes, then use the arrow keys to move through those lists of notes (it even remembers the cursor position for all notes), and finally return to my note and continue typing, all without ever leaving my keyboard. This is in addition to some keyboard shortcuts also present in other apps, such as quickly moving back and forth between any number of notes etc.

Fifth, BBEdit lets me collaborate with people who, for better or worse, prefer word documents to PDF or TeX or plaintext files. I like to keep my files as plain text myself while still having other formats almost instantly accessible, and I prefer not to outsource this to third-party plugins when the excellent Pandoc is just a command away. My Pandoc commands are therefore stored as scripts and specific reusable templates or other portions of files are stored as clippings in BBEdit, making it incredibly easy to work with them all natively, right inside my text editor.

Sixth, because I work with so many files on multiple different things, I often find myself opening a specific set of files for a specific task. Workspaces in BBEdit remember this arrangement of windows and files and replicate it for me, allowing me to quickly switch between predetermined arrangements of BBEdit instead of scouring through the sidebar for the files I want every single time.

Seventh, BBedit lets me fold by heading level. This feature is indispensable so I kept the best for last. I can fold everything up to, say, heading level 3 to get a better overview of my research notes or structure file. I can use simple keyboard shortcuts to fold and unfold selections, fold all headings, fold up to a specific heading level on opening a file and much more. Other apps allow folding too, often with single clicks, and BBEdit too allows this; but the ability to fold all headings in a document up to a certain level is something I have only seen in Emacs (and that too with some lisp sprinkled over it).

---

---

That note-making should be purpose-driven ought to come as no surprise. If notes cannot be separated from their content, why would separating them from their context ever help? We have already spoken of the exceptions to this rule, such as saving code snippets etc. But unless that alone is your aim – to save bits of information so you know ‘facts’ – the digital reincarnation of Luhmann’s zettelkasten system is more a mimicry of his style of work than an enhancement or even a direct adoption of the effectiveness of his now-famous process.

Within the same academia in which Luhmann worked were plenty of other academics who worked differently. After all, the method of slip boxes was the less popular method of working. Perhaps where Luhmann’s zettelkasten made a difference was in helping him reference his notes where others who had piles of handwritten sheets could not possibly riffle through them as efficiently. The computer has effectively solved this problem, making zettelkasten more of a choice in unorganised references than an unparalleled method of working.

I know very few people in academia who use the zettelkasten system. Throughout my attempts with it over the past few of years I was, in fact, the only one I knew. This is by no means a case against zettelkasten so much as an interesting observation. And this essay is also no plea for people to stop using zettelkasten. It is easy to get swept up in a method because so many people, especially online, are recommending it days after trying it out.

The case I make in this essay has, hopefully, offered a balanced counterargument to that system. More importantly, I hope the case I make has offered a compelling argument for trying out something else, rather than adopting a prescriptive system and hoping it will make miracles because you did everything by the proverbial book. Explore with the selfish desire of creating a bespoke workflow for yourself. You might find that a zettelkasten system is perfect for you after all. But you may also find that the answer lies not in working with the hope that your notes will be useful one evening seven winters hence, but instead in working with transient goals in sight.

---

I use and highly recommend Readeck. It lets you save, organise, label, highlight and (by next year) annotate. ↩