Zettelkasten Forum


Meta-discussion: which ZK problems are solvable here?

@MikeBraddock asked me to provide my take on the debate that occurred several years ago between @sfast and Daniel Lüdecke in comments on this page. Doing so led me to consider the long discussion in this thread as well as the challenges in helping each other here - hence the meta-post (which includes non-meta ZK content as well).

The debate about the best way to implement a notetaking and writing system is making slow progress at best because it does not have a shared goal. People are not recognizing unstated disagreements (or are failing to resolve stated disagreements) about the scope of the system's function in the writing process. In other words, there is no true consensus on the type of output the ZK system should provide, and thus no way to resolve these disagreements. I think that my analogy to discussing the best ways to bake bread in another thread holds here: it really depends on what kind of bread you are trying to make, and if you do not first agree upon that, then there is no way to resolve differences about the best method to bake bread.

It is especially difficult to resolve these issues because each person only has personal experience with their own writing and thinking processes, which means that they all start from a different set of prior assumptions, and thus anything that is unstated by someone else will be filled in with your previous personal experience about that thing, which will not match the other person's.

Re: the @sfast and Daniel Lüdecke discussion, a large part of that debate involved discussions about Luhmann's writings about how his system was supposed to operate, including arguments about the correct German to English translation. But I can't speak German and am not very interested in Luhmann's intent. As an outsider, I will say that I think that that discussion led nowhere because the participants were aiming to defend their positions rather than to understand why two obviously intelligent people would disagree about a topic of shared interest. For example, there were comments like "The ZK is about communication, which means you should X" with response, "No, because communication really means Y, so you should not X." Resolving these types of things involves very nuanced language and lots of effort to resolve things like what "communication" means, and online forums are not ideal for that.

But I'm glad I went back to review those comments because it prompted me to review Lüdecke's presentation about the ZK system, and he absolutely nails a central point: There are multiple ways to connect/group/link individual notes, including categories, tags, and links, and each of these ways has advantages and weaknesses. The genius of the ZK system, in his view (and mine, but not @sfast's), was that Luhmann combined these systems in a way that allowed them to retain many advantages while overcoming some of their weaknesses. So Luhmann IDs retain the useful hierarchy of structured categories, but allow direct links between notes to connect otherwise unconnected branches across those categories. @sfast thinks that traditional information organization systems should be combined not in Luhmann's way, but in a different way (with date/time IDs and direct links only), due to differences between analog and digital notetaking.

My take on the discussion is that while both discussants are partially correct, they are implicitly overlooking that the ideal type of combination or balance between the systems will differ depending on what you think the scope of the system is and what the tendencies are in your own thinking and writing process. A corollary is that combining systems adds not only strengths but also weaknesses that must be considered, and a strength for one person may be a weakness for another. For example, a system that makes it easier rather than harder to add tags will be good for one person because it reduces cognitive friction in getting notes labeled, but bad for another person who is concerned about being 100% accurate in adding tags because it encourages them to have more tags in the system, which feeds into their unproductive tendency for completeness. Whether there are practices that are generally useful will depend in part on how many people share similar mental tendencies when writing.

The centrality of these differences to the disagreements on this forum has become crystal clear to me in the debate over the last week between several forum members who have spent much-appreciated time explaining their own writing processes. I'll group various people's comments into fictional posters AA and BB. I don't know any of these people and I'm going to make a lot of assumptions that are surely inaccurate. I hope no one takes the following as actual claims about real people; I just need stand-ins for illustrating my point. So please don't get angry that I said that you said something that you never said, etc.

This site's previous posts have advocated labeling notes with unique IDs based on a date and time stamp--call them DTIDs. Recently, poster AA reported that they had tried the DTID system for a long time and went back to Luhmann-style IDs, which I will call LIDs. Poster BB responded that there is no reason to use LIDs because DTIDs + direct links do the same thing as LIDs, but are easier to quickly assign and don't have the drawback that (they think) LIDs create rigid categories that limit thinking. Based on what I wrote above, it should be clear that this difference in systems is an advantage for some people but is a disadvantage for others. The disagreement exists in part because of differences such as these:

  • Scope of system: AA thinks of the ZK as a notetaking and thinking system (not a writing system) whose output is ideas. BB thinks the ZK is a notetaking, thinking, and writing system whose output is a structured set of notes that is akin to a first rough draft.
  • Thinking style: AA has hard-to-articulate instincts about connections between notes, and use LIDs as a way to capture those instincts so that the connections might be discovered in the ID process or later. BB's instincts about note-to-note connections are easy for them to articulate, and their instincts about connections that are not easy to immediately articulate are consciously or nonconsciously dismissed as meaningless. (Lest people think I am being critical, note that I am like BB in this regard.)
  • Notetaking process: AA writes most new notes while reading and thinking about their previous notes - notes are thus naturally created in branches. BB writes most new notes when reading original sources rather than their own older notes - notes are not naturally created in branches, but must be intentionally linked in sequences.

Of course, this is all an oversimplification, because all of these things are hard to define in clean categories. Of course AA will write many new notes while reading new sources, and of course BB will write some new notes when reading their older notes. But differences in emphasis will mean the ideal balance between systems will differ.

To take a different example, people like different sorts of search strategies. There's research showing that more people like searching on a computer by visually browsing file folder trees than by doing text search, but obviously many people prefer text search. Search results can also be displayed in different ways in different programs, so having category information in a note title (as they are with LIDs) may be important to some but not others.

These topics are really complicated to sort out. Take another example: @sfast argues for a DTID system based on his understanding of implicit vs. explicit associations. But (and this is not a criticism of @sfast) he doesn't distinguish between existing associations vs. newly created associations, or between intentionally created vs. accidentally created associations, and he doesn't talk about how different types of systems will encourage or discourage different types of associations, or about which associations decay over time and why, or the difference between embodied associations and abstract conceptual associations, or the role of emotion in associations and decision-making, and so on, and so on. He can't do that that because he is not a field-leading research psychologist writing for an audience of experts. And even if he were, no one currently knows the ways all of these factors relate to each other. Go ask a really top-level educational psychologist. They'll confirm. But these differences are essential to discovering the "ideal" system, and our different and underinformed ideas about how the mind works are behind some of the disagreements on this forum.

Or another example: in the same post, @sfast writes, " Luhmann-IDs are a method of encoding meaningless (not semantically accessible) relationships". Of course that is one way of understanding "meaningless", but most posters will not assume that meaning, or may not understand that meaning, and it's being defined in the middle of a forum thread rather than the beginning, and, lastly, IMHO, it's wrong. Most language speakers cannot articulate the rules of grammar of their own native language, but errors in grammar consciously feel "wrong". Yet, without advanced education, the reasons are not semantically accessible to them. But of course grammar rules are not "meaningless" - we follow them, and they can be articulated by experts. Now, I'm sure that @sfast will have an argument about how that is not what he means by "semantically accessible", or how grammar is different from notes, or something, but it is all besides the point. The point is that if a certain way of implementing a ZK system is based on issues as subtle as what mental relationships are semantically accessible, then we are just out of our depths here. I assume there is no research on this issue in the context of taking notes in a ZK and writing (which is essential to generalizing the findings). This doesn't mean that we must give up on discussing ways to improve our ZK practices, but it does mean that, if we don't happen to already agree (implicitly or explicitly) about the facts of the human mind that drive preference for one ZK implementation over another, we will not be able to settle it here.

Now, if I am right about the centrality of these individual differences in desired scope of system and thinking styles, then it is not impossible to make progress on discovering an ideal ZK system, but it is very hard, especially because it is very hard to even articulate these things in a way that others can understand or to understand others with different goals and thinking styles than your own. Few people understand their own cognitive limitations. Few people have broad experience learning from and teaching many different kinds of writers and writing. I hope this isn't coming across as condescending--I myself am unable to solve even my own central writing problem, and I've been at this a long time. But it is interesting to me how many people (especially on the Reddit /Zettelkasten subreddit!) just confidently say, "You should do it this way" or "That practice is unnecessary." Do they even really know what the original poster is trying to do? Or how that person thinks? Do they really know whether their core assumptions about the mind are true, or are they satisfied with just having a supporting argument?

In some cases, such as in this thread mentioned earlier, certain posters have found others who share their writing goals and/or mental tendencies, and thus who are more likely to be able to provide them with useful suggestions about implementing their ZK. That is a happy accident!

In any case, even if it now seems to me unlikely that I will run across some guru online who can provide a system that will overcome my specific writing challenges (which I will probably raise in a different post), indirectly, by participating in and observing all of the passionate discussions here I think that I have gained a lot of insight that may help me pinpoint potential answers for myself. Thanks!

Comments

  • I haven't used The Archive yet because I don't own a Mac :disappointed: (so I don't know if it has this feature) but my favorite implementation of Zettelkasten (at its core) has been ZK3. I just wish it was done in a markdown editor way.

    To me what makes me like it so much is what feels like the combination of luhmann IDs and Date/Time IDs. Having them number 1-Infinity is useful because it treats each note as a standalone number or idea that can be a part of many different note sequences. But because you have a super easy way to see note sequences and click between them in the side bar you also get the luhmann ID benefit of easily seeing the flow of an argument and what branches it has.

    With the goal being in the end to get a note embedded into as many useful contexts as possible.

  • Excellent work summarizing all this, thanks a metric ton, @cobblepot!

    In any case, even if it now seems to me unlikely that I will run across some guru online who can provide a system that will overcome my specific writing challenges (which I will probably raise in a different post)

    I am very much looking forward to find out what exactly you're stuggling with in your workflow. It didn't become obvious when you refuted "to each their own" as a solution to the Folgezettel question :)

    @Nick: Daniel Lüdecke's decision to forego nested LIDs in favor of increasing numbers ("auto-incrementing numerical IDs" in usual database terms) makes a lot of interesting things possible. With the decision to put plain text at the core and center and not introduce any new metadata or information storage, if possible, our approach revolves around finding links in notes and making them visible. Different browsing experiences that can be represented with the current Markdown/wikilink convention is indeed on our list of things to explore, once the bread-and-butter editing features are really, finally, done.

    -- And to me, it seems this is what most of the discussion is about. Access modes. Not storage modes. Nobody is passionate about adding LIDs or DTIDs to file names; that's just a means to an end, whereas the end, here, is not to create the flattest or most nested of hierarchies, but to get access to the cool stuff. For LIDs, their being part of the title, as @argonsnorts pointed out, seems to be the greatest benefit. You see something new in the list of file names. The same information could be encoded in the file contents, but that would beat (part of) the purpose.

    Author at Zettelkasten.de • https://christiantietze.de/

  • I'm glad you found it useful, @ctietze. It took a long time to work through.

    @ctietze said:
    -- And to me, it seems this is what most of the discussion is about. Access modes. Not storage modes. Nobody is passionate about adding LIDs or DTIDs to file names; that's just a means to an end, whereas the end, here, is not to create the flattest or most nested of hierarchies, but to get access to the cool stuff. For LIDs, their being part of the title, as @argonsnorts pointed out, seems to be the greatest benefit. You see something new in the list of file names. The same information could be encoded in the file contents, but that would beat (part of) the purpose.

    You are missing at least one additional benefit that @Taylor mentioned in a different thread:

    Luhmann-IDs force you to review your zettels each time you add a new note. And not just skimming a list of notes with a relevant tag, or skimming the titles of notes in a structure note, but actually reading your zettels and following their links searching for the most appropriate place to put the new note. And if you review the relevant zettels prior to writing the note, you'll write a note that integrates much better with what is already in the Zettelkasten, and you'll take notes and read with an eye to growing the trains of thought already in your Zettelkasten.

    What @Taylor is articulating so well can also be described as a benefit of a specific encoding process. Notetaking, like memory, requires information to be encoded before it is stored and later retrieved (or accessed). Pro-LID people (let's call them "Lidders") are trying to point out that LIDs provide them a benefit during the encoding process, regardless of whether LIDs provide a benefit during the retrieval process. This is another "cognitive style" balance issue. Lidders want some forced categorization when putting notes into the system because that process provides a type of feedback on their thoughts that can be valuable. Others find the extra friction outweighs any potential benefit.

    I'll give another similar example. I used to write my notes in bullet-point headline style rather than in complete sentences. Recently I followed some advice to instead write in full grammatical sentences. At first the new restriction seemed to be needless cognitive friction, but soon it revealed itself to me as useful because it provided feedback that forced me to clarify ambiguities that I would have previously overlooked. These types of fine-tuning of a system will be very dependent on people's writing topics, writing goals, and mental habits.

    In How to Take Smart Notes, Ahrens argues that the multi-step process is useful because each step provides you feedback on your current state of knowledge of the material. One rarely mentioned part of a system that must be balanced is between providing feedback and facilitating effortless storage, similar to the balance that must be reached in up-front work encoding/labeling information to make later retrieval faster vs. expending more effort later to make it faster to store current information.

    Knowledge management is a new area for me and this week I am just starting to realize how deeply complicated it is.

  • @cobblepot I am up at 4 am my local time to be sure I reply because your discussion post here is awesome. Your capacity to digest vast amounts of information and reframe it with clarity and add insight is amazing. You are a treasure for this forum.

    It is very good you did this heavy lifting for us.

    Thank you!

Sign In or Register to comment.