Zettelkasten Forum


How do you enter your Zettelkasten?

My Question

Hi everyone! I’m interested in getting a sense of how the people on this forum enter their Zettelkästen when looking for notes to link to. Imagine you’re in this situation:

You have some initial reading notes on an article or book chapter (what Ahrens, I think, calls “literature notes” and what, I think, Sasha would call notes outside your Zettelkasten). You want to create some new Zettels from those reading notes. What’s the first thing you do?

Do you write up a Zettel based on some of your reading notes first and then look for related notes to link to? Or do you look for and review related notes you have before writing the new Zettel? How do you typically look for notes to link to? Tag search? Full text search? Structure notes? A topic index/register? Do you experience any resistance while doing this?

Why I’m interested

Over the past few months I’ve been experimenting with two different Zettelkasten setups. The first is in The Archive following the orthodox principles as outlined on this site. The second is in Obsidian, following (largely) @nickmilo22's IMF/MOC/LYT method and (sacrilegiously) using the note titles themselves as IDs.

Using these two setups helped me clarify a bottleneck I’m having. I have a hard time “entering” my Zettelkasten to find old notes to link new notes to. I’m not sure why, but typically when I do a search (tag or otherwise) or even look at one of my structure notes I often become, in David Allen’s words, "numb to the pile” of notes I get. I tend to skim the titles, not immediately see any connections, and end up giving my new note some half-hearted links that I haven’t really thought through.

In my Archive Zettelkasten, I found that using a topic register in the way Luhmann did got me around this problem somewhat. I manually created a topic register that linked topics not to an overview or structure note but just to whatever Zettel I had that was most directly about that topic. When going to add a new Zettel from some reading notes, before I wrote the new Zettel, I would check my topic register to see if I already had notes on the same topic or a closely related topic. If there was, I’d click the link in my register and read whatever Zettel it linked to. Usually it wasn’t directly related to my reading note, so I’d follow some links from it and, in that way, review the notes I already had on the topic. I'd do this until I got to a Zettel I could use to contextualize my reading note, or that my reading note could expand on, or object to, or in some way comment on. I’d then make a new link in the existing Zettel and click through to create my new note. I like this method because it got me to the “ground floor” of my Zettelkasten very quickly and forced me to review a good chunk of the notes I had on a topic, and not just notes I judged to be directly relevant to my new reading notes.

I haven’t been able to replicate this workflow in Obsidian though. Crucial to the way my topic register works is that I can’t see the title of the note associated with a given topic until I click on it. That element of surprise seems necessary to stop me from prejudging the relevance of a note based on its title. But in Obsidian you have to link using the full title of the note, even if it has a date-time ID. This is tempting me to transfer what I have in Obsidian to The Archive.

I’m also not sure how scalable having a manual topic register is. I currently have only 55 topics in my register, ordered alphabetically but maintained manually. It’s manageable now, but I’m not sure what I’d do once it gets to 100 or 200+ topics. Create sub-indexes based on broader topics? (Perhaps I'm being too much of a wimp here; Luhmann, after all, maintained his topic register on paper).

I’ve been thinking about this because, though Luhmann used a topic register to enter his Zettelkasten, the consensus on this forum seems to be that the role played by the topic register is adequately replaced by search and structure notes. This is the motivation behind my questions about how you enter your Zettelkasten. I’ve found search and structure notes unhelpful as entry-points because I seem to have a tendency to prejudge the relevance of a note based on its title. (Structure notes are obviously essential for organizing notes on a topic and getting an overview; i'm not saying they're unhelpful as such).

I’m interested if anyone else has experienced this resistance while using search/structure notes as an entry-point, or if it could be a symptom of me having bad titles/bad tags/bad search queries/sloppy structure notes/etc.

Comments

  • It is the interaction between my thoughts and the book I am reading
    that is the most enjoyable thing. If I agree with the author, my
    thoughts resonate with the author. Otherwise, I will pause and think
    my previous thoughts. Usually, these are the moments that I jot down
    quick notes on the pages.

    For processing long books, I do not open the digital ZK. I want to
    100% focus on the book. The more focus, the more pleasurable feeling,
    the more interesting ideas sparking.

    After finishing a book, I will go through all those little notes. This
    is the time that I start to build connections in my ZK. Usually, it
    starts with searching keywords from notes I jotted down in the book.
    If I am lucky, I will find the atomic notes in my ZK, and I add new
    thoughts to it. And I will also add a back link to the location where
    this new idea comes from. Sometimes, the search result is just an
    index note. In this cases, I just add a new note to this index. After
    that, I try my best to review this new idea and think about it. The
    idea is to find the possible connection to this new notes.

    When I review the notes in book, I sometimes find the index notes
    interesting. Those new ideas are not came from the moments when I
    first read the book but from the moment when I process its content. I
    will update the old index notes and any ideas on that index notes.

    From a visualization of all my ideas, I can visually interact with all
    my ideas. I sometimes open this graph while I am processing a note.
    It can spark some ideas. This is another method I enter my ZK.

    I also have a random mechanism to access all my notes. Basically, it
    open a random notes in my ZK. It might give me some surprises.

  • @Taylor said:

    My Question

    Hi everyone! I’m interested in getting a sense of how the people on this forum enter their Zettelkästen when looking for notes to link to. Imagine you’re in this situation:

    Do you write up a Zettel based on some of your reading notes first and then look for related notes to link to? Or do you look for and review related notes you have before writing the new Zettel? How do you typically look for notes to link to? Tag search? Full text search? Structure notes? A topic index/register? Do you experience any resistance while doing this?
    ..........
    I’ve been thinking about this because, though Luhmann used a topic register to enter his Zettelkasten, the consensus on this forum seems to be that the role played by the topic register is adequately replaced by search and structure notes. This is the motivation behind my questions about how you enter your Zettelkasten. I’ve found search and structure notes unhelpful as entry-points because I seem to have a tendency to prejudge the relevance of a note based on its title. (Structure notes are obviously essential for organizing notes on a topic and getting an overview; i'm not saying they're unhelpful as such).

    I’m interested if anyone else has experienced this resistance while using search/structure notes as an entry-point, or if it could be a symptom of me having bad titles/bad tags/bad search queries/sloppy structure notes/etc.

    Hi @Taylor In regards to the first question in your post, I always first take notes (what Ahrens calls "fleeting notes"), either manually by handwriting on my iPad or paper, or by typing on my iPad or computer). I do that whether I am reading, listening to or just thinking about something. Next (step 2) I process those fleeting notes into zettels (or a series of zettels). Finally (step 3) I look for connections. I don't seem to get distracted in the manner you describe - I simply search on phrases or click on tags, and find the zettels to which I want to link. It takes some time, in some cases even longer than writing the zettel, but that is part of the investment.

    I use structure notes a bit (occasionally). I understand why they could be helpful and maybe they will actually prove to be so, as my ZK grows. I just learned about Luhmann's topic register a short while ago and it seemed like a good idea to me for a paper ZK, but un-necessary in an electronic ZK. I'd just search on several different terms and see where that took me.

    One thing I am careful of, though, is to have as many connections to each zettel as I can (there was a post in this forum on some script to do just that recently) - if I have 0 or 1 connection, that zettel is going to be lost really fast. My goal is to have each Zettel connect to at least 4 or 5 other zettels; some connect to a lot more.

    Note that my ZK only contains a few hundred Zettels right now. Maybe my tune will change if I get up to the thousands range. But I think the key is to have a way to enter a particular thought "train" or "pathway" in your ZK and then have sufficient connections to start surfing.

  • @Taylor said:

    My Question

    Hi everyone! I’m interested in getting a sense of how the people on this forum enter their Zettelkästen when looking for notes to link to. Imagine you’re in this situation:

    You have some initial reading notes on an article or book chapter (what Ahrens, I think, calls “literature notes” and what, I think, Sasha would call notes outside your Zettelkasten). You want to create some new Zettels from those reading notes. What’s the first thing you do?

    Do you write up a Zettel based on some of your reading notes first and then look for related notes to link to? Or do you look for and review related notes you have before writing the new Zettel? How do you typically look for notes to link to? Tag search? Full text search? Structure notes? A topic index/register? Do you experience any resistance while doing this?

    I create notes first, then "look for related notes to link to." I use The Archive's boolean full-text search using key terms and ideas from the newly created note. Some searches result in zero hits, some in 50 or more. Searches that have a lot of hits generally means that I've not formed my search query appropriately.

    This is the discovery part of the practice, and I find it very rewarding. Why resistance?

    I found that using a topic register in the way Luhmann did got me around this problem somewhat. I manually created a topic register that linked topics not to an overview or structure note but just to whatever Zettel I had that was most directly about that topic. When going to add a new Zettel from some reading notes, before I wrote the new Zettel, I would check my topic register to see if I already had notes on the same topic or a closely related topic. If there was, I’d click the link in my register and read whatever Zettel it linked to. Usually it wasn’t directly related to my reading note, so I’d follow some links from it and, in that way, review the notes I already had on the topic. I'd do this until I got to a Zettel I could use to contextualize my reading note, or that my reading note could expand on, or object to, or in some way comment on. I’d then make a new link in the existing Zettel and click through to create my new note. I like this method because it got me to the “ground floor” of my Zettelkasten very quickly and forced me to review a good chunk of the notes I had on a topic, and not just notes I judged to be directly relevant to my new reading notes.

    I like this notion of a “ground floor”. I'm not sure what the difference is between a structure note and what is being called a topic register. My structure notes are topic registers.

    I haven’t been able to replicate this workflow in Obsidian though. Crucial to the way my topic register works is that I can’t see the title of the note associated with a given topic until I click on it. That element of surprise seems necessary to stop me from prejudging the relevance of a note based on its title. But in Obsidian you have to link using the full title of the note, even if it has a date-time ID. This is tempting me to transfer what I have in Obsidian to The Archive.

    Jump into The Archive. The water is warm, and you'll love it.

    I’m also not sure how scalable having a manual topic register is. I currently have only 55 topics in my register, ordered alphabetically but maintained manually. It’s manageable now, but I’m not sure what I’d do once it gets to 100 or 200+ topics. Create sub-indexes based on broader topics? (Perhaps I'm being too much of a wimp here; Luhmann, after all, maintained his topic register on paper).

    I currently have 13 "topic registers," and they contain 51 sub-indexes zettel. (Using your language.)

    I’m interested if anyone else has experienced this resistance while using search/structure notes as an entry-point, or if it could be a symptom of me having bad titles/bad tags/bad search queries/sloppy structure notes/etc.

    You are on the right track, don't be so hard yourself. Indeed, in the beginning, we all have "bad titles/bad tags/bad search queries/sloppy structure notes". Don't judge yourself harshly, be your own tutor, and learn and try and be .045% better than yesterday. Check out Ira Glass and his take on the gap.

    Will Simpson
    I must keep doing my best even though I'm a failure. My peak cognition is behind me. One day soon I will read my last book, write my last note, eat my last meal, and kiss my sweetie for the last time.
    kestrelcreek.com

  • @Will said:

    I like this notion of a “ground floor”. I'm not sure what the difference is between a structure note and what is being called a topic register. My structure notes are topic registers.

    In the new introduction to ZK, towards the end of the section:

    https://zettelkasten.de/introduction/#the-fixed-address-of-each-note

    there is a discussion of Luhmann's "topic register" (with image). This was news to me. It makes sense for a paper ZK, as a way to find an entry point to just one note in a series of notes forming a "thought train". I don't see the need when one has an electronic ZK with powerful search capabilities, as you pointed out.

  • You have some initial reading notes on an article or book chapter (what Ahrens, I think, calls “literature notes” and what, I think, Sasha would call notes outside your Zettelkasten). You want to create some new Zettels from those reading notes. What’s the first thing you do?

    I go to the relevant structure notes first. I place the link and create the note from there. After that I connect the new knowledge to my old one. The amount of links depend on the context. If I found a new exercise for physical training I place quite few links. If I found a very profound idea I can spend hours on it expanding it, re-writing it from different perspectives (say: re-think it with different mental models) and collecting all the consequences for the existing stuff in my Zettelkasten.

    I am a Zettler

  • Thank you everyone for your responses. Sorry for starting this tread and then going awol for a month.

    @GeoEng51 said:

    Hi @Taylor In regards to the first question in your post, I always first take notes (what Ahrens calls "fleeting notes"), either manually by handwriting on my iPad or paper, or by typing on my iPad or computer). I do that whether I am reading, listening to or just thinking about something. Next (step 2) I process those fleeting notes into zettels (or a series of zettels). Finally (step 3) I look for connections. I don't seem to get distracted in the manner you describe - I simply search on phrases or click on tags, and find the zettels to which I want to link. It takes some time, in some cases even longer than writing the zettel, but that is part of the investment.

    I don't know if I'd say I get distracted, but more that I get myself into a situation where I judge in advance that there are no connections between my new note and the notes I already have. Especially if I proceed in the way you described: writing new zettels before looking for ones to link to and finding zettels via search and tags. If I take that approach, what will often end up happening is I'll write a new zettel, tag it, do a tag-search for those tags, skim the titles of the notes that come up, and, based on the titles, conclude that there are no connections there and so won't reread the notes. That's an extreme case. Occasionally I'll find one or two notes where, based on the titles, I'll identify connections to the new note and make them. Perhaps this is simply a discipline problem and I should force myself to read every note that comes up in a tag-search. But this can quickly become daunting.

    @Will said:
    I create notes first, then "look for related notes to link to." I use The Archive's boolean full-text search using key terms and ideas from the newly created note. Some searches result in zero hits, some in 50 or more. Searches that have a lot of hits generally means that I've not formed my search query appropriately.

    This is the discovery part of the practice, and I find it very rewarding. Why resistance?

    Three reasons for the resistance come to mind: First, as I touched on in my response to @GeoEng51, when I read through a list of note titles, I tend to judge in advance that an old note is irrelevant to my new note without reading the contents. This is perhaps related to my second reason which is that I have, in general, a resistance to reading my old writing. Going back to rework old papers, or even write a second paper on the same topic puts me off for some reason. There are perhaps deeper psychological reasons for this, so I'll put it aside. The third reason is the results of a tag search has no structure. It's just a pile of notes I've decided to relate via a common name. Because of that I find it difficult to orient myself in those notes, and so I don't.

    The third problem is addressed by structure notes, but I still end up facing the first problem. The topic register is the closest I've found to a solution to this.

    I like this notion of a “ground floor”. I'm not sure what the difference is between a structure note and what is being called a topic register. My structure notes are topic registers.

    Here are examples from my current ZKs. Here is an example of part of a structure note I created in Obsidian for "Structuralism" (which I now realize is way too broad, so I might specify it to 'structural linguistics'):

    The full titles of the notes are present and they're organized in an outline structure. I try to make the title of my note a conclusion, with the body of the note being the argument/justification/reasons for that conclusion. Sometimes if my note is more organizational or informational this wont happen (e.g. "Three aims of linguistics"). However, most connections I find between notes are not connections between the conclusions, but connections between arguments or justifications. Some justification presupposes a further premise, or has a further implication beyond the stated conclusion. Or two arguments are analogous in some way. Or one note offers an objection to a presupposition made my an argument in another note. And so on. Maybe that's why looking at note titles alone is insufficient for me? I often don't remember what the content of a note is, so I don't see the possibility of a connection from title alone. And, when skimming a structure note, if I don't see the possibility of linking from the title alone, I won't review the note.

    This is the first bit of the current general topic register I have in The Archive. I just call it "Index":

    There are no note titles here, only topics. It's alphabetized and absolutely general: the index has no theme. The links do not link to concept notes, but just to whatever I think is the most central note I have on the topic. That is, the link next to "Discrete Multiplicity" does not link to a structure note on the concept, but rather to a specific note about discrete multiplicity, one titled: "A discrete multiplicity is a collection of elements whose identities are indifferent to their relations." In this case, the note is pretty definitional, but it need not be.

    Here's how my topic register gets around some of the problems I have above. Let's say I have some new note on Nietzsche's concept of genealogy. Without getting too distracted, let's say the note is this: I'm reading Raymond Geuss's essay "Nietzsche and Genealogy" and notice that Geuss distinguishes Nietzsche's genealogical method from tracing a pedigree. So I make a note: Tracing a pedigree attempts to secure the value of something by finding a single, noble origin for it (like someone tracing their ancestry back to royalty). Nietzschean genealogy traces the origin of something we take to be valuable to multiple historically contingent sources that are often violent. This doesn't capture the whole distinction, but it'll do. Before typing up this note as a zettel, I'll want to find a note to link to it. I have an entry for genealogy in my topic register already, so the first thing I'd do is click on that. It takes me to this note:

    If I were skimming tag results for #genealogy or a structure note on Nietzsche's genealogical method, I very likely would have passed over this note. On the surface, the distinction between tracing a pedigree and tracing a genealogy doesn't have much to do with the debate over whether Nietzsche distinguishes between genealogy and critique. But reviewing this old note with the new note in mind might make me see the distinction in a new light: if genealogy and critique are distinct, then a thing's history will not determine its value (in other words, Nietzsche avoids the genetic fallacy). Tracing a pedigree is an obvious instance of the genetic fallacy. If genealogy and critique are distinct for Nietzsche, then the point of genealogy is not to provide alternate, "bad" pedigrees, but to somehow object to the idea of a thing's origin determining its value.

    These thoughts are perhaps too vague to create a new note with, and I might decide that there's not enough connection between my "Genealogy and critique are distinct" note and the genealogy/pedigree distinction in Geuss to directly link the two. But those thoughts will contextualize my further search through my ZK and likely inform how I write the new note. I'll stop here because this is already way too long. Though, just for comparison's sake, here are my search results for #genealogy:

    Four notes is manageable, though it's not evident here how they're related. But I can still tell that if I did this search I'd have a moment of frustration since none of these titles evidently connects with the genealogy/pedigree distinction I want to make a note on. If I began by writing the genealogy/pedigree distinction note and then looking for connections in this tag search, I'd likely find no links or only weak links, since I wouldn't have written the genealogy/pedigree note with the debate on the genealogy/critique distinction in mind.

    Perhaps my use of the topic register is just an elaborate way to trick myself into actually reading my old notes and engineering serendipity. I look on with envy at those who don't need to deceive themselves into thinking.

    I currently have 13 "topic registers," and they contain 51 sub-indexes zettel. (Using your language.)

    I'm thinking I might have to start making thematic sub-indexes at some point to make it more manageable. Every time I add a set of zettels to my ZK I usually also add one or two new entries to the topic register. I'm not yet sure how these sub-indexes will look, though.

    You are on the right track, don't be so hard yourself. Indeed, in the beginning, we all have "bad titles/bad tags/bad search queries/sloppy structure notes". Don't judge yourself harshly, be your own tutor, and learn and try and be .045% better than yesterday. Check out Ira Glass and his take on the gap.

    Thank you for the Ira Glass clip. I'm familiar with the quote, but it's always good to remind myself of it.
    I wish it was still the beginning. This is my first zettel ID: 201504092242
    I've had many false starts since then, and I don't think I've ever experienced the kind of flow others tend to describe when working with their ZK. Navigating from a topic register is the closest I've gotten.

  • @Taylor Do you run Boolean searches on tags and/or terms? That might help you to find a more specific, shorter and more relevant list of zettels.

  • @GeoEng51 said:
    @Taylor Do you run Boolean searches on tags and/or terms? That might help you to find a more specific, shorter and more relevant list of zettels.

    I know how to, but so far a use case never occurred to me. Even now I can't really think of a useful set of notes that I could only retrieve with a Boolean search. I tag fairly precisely (my latest note is tagged #logical-foundations, #lingusitic-turn, and #analytic-philosophy) so I don't have 100s of notes tagged #language or #logic or something like that. Perhaps if I do lots of research into a single topic I might end up in a situation like that, but so far that hasn't happened.

    I find tag search to be very good at quickly finding a particular note when I have one in mind. Or making sure I review all my notes on a topic when writing up an outline. But it rarely helps me when I'm trying to find connections between notes.

  • @Taylor Most of my tags are attached to less than 10 zettels, many less than 5 - on purpose. I only have a few attached to more than 15. So my experience with searching on tags is similar to yours. But then I only have a couple of hundred zettels (I'm a relative newby). I expect as my ZK grows, using Boolean searches will become more useful.

  • @Taylor said:
    Three reasons for the resistance come to mind: First, as I touched on in my response to @GeoEng51, when I read through a list of note titles, I tend to judge in advance that an old note is irrelevant to my new note without reading the contents. This is perhaps related to my second reason which is that I have, in general, a resistance to reading my old writing. Going back to rework old papers, or even write a second paper on the same topic puts me off for some reason. There are perhaps deeper psychological reasons for this, so I'll put it aside. The third reason is the results of a tag search has no structure. It's just a pile of notes I've decided to relate via a common name. Because of that I find it difficult to orient myself in those notes, and so I don't.

    I hear you! My resistance comes from considering the role of Opportunity Costs too much. I sometimes find myself thinking that time spent reworking old notes detracts from the time I could be spending the time exploring new and shiny ideas. I've always felt great about the time spent refactoring old notes afterward but not before, when the anticipation causes resistance.

    The full titles of the notes are present, and they're organized in an outline structure. I try to make the title of my note a conclusion, with the body of the note being the argument/justification/reasons for that conclusion.

    When presented with just a list or TOC of notes, it is hard to ascertain a particular note's relevance. Good titling would surely help. This is a work in progress for me.

    Maybe that's why looking at note titles alone is insufficient for me? I often don't remember what the content of a note is, so I don't see the possibility of a connection from title alone. And, when skimming a structure note, if I don't see the possibility of linking from the title alone, I won't review the note.

    I'd suggest you need to add more to your outline/structure notes. I think of these as "Annotated Table of Contents." Here is a sample, and I know I could do a better job of providing a preview of the note in question. I used to, and still have some notes, which are just lists of other notes, like yours, and because of my ineptitude at titling, some are just a list of most gibberish. I'm slowly refactoring them into "Annotated Table of Contents." one at a time.

    Perhaps my use of the topic register is just an elaborate way to trick myself into actually reading my old notes and engineering serendipity. I look on with envy at those who don't need to deceive themselves into thinking.

    Engineered Serendipity WOW! I want some of that.

    Thanks for sharing.

    Will Simpson
    I must keep doing my best even though I'm a failure. My peak cognition is behind me. One day soon I will read my last book, write my last note, eat my last meal, and kiss my sweetie for the last time.
    kestrelcreek.com

  • @Taylor: Do you have a link to a thread or post that describes @nickmilo22's IMF/MOC/LYT method? Haven't heard of that before...

  • @Vinho here is a link to Nick's IMF/MOC/LYT medium essay.
    In what ways can we form useful relationships between notes?

    Will Simpson
    I must keep doing my best even though I'm a failure. My peak cognition is behind me. One day soon I will read my last book, write my last note, eat my last meal, and kiss my sweetie for the last time.
    kestrelcreek.com

  • @Will Thanks, will have a look at it tomorrow :)

  • @Will said:
    I'd suggest you need to add more to your outline/structure notes. I think of these as "Annotated Table of Contents." Here is a sample, and I know I could do a better job of providing a preview of the note in question. I used to, and still have some notes, which are just lists of other notes, like yours, and because of my ineptitude at titling, some are just a list of most gibberish. I'm slowly refactoring them into "Annotated Table of Contents." one at a time.

    Thank you for the suggestion and example @Will. I think my structure notes could benefit from more annotation, and it might help me more easily see where a new note would fit in. But I also think: what would then distinguish a structure note from a paper/book chapter outline? Perhaps not much, or that one could be converted into the other easily. Maybe that's part of the point.

    I'll try this out, and tell myself that it'll be worthwhile in any case because it will be getting me closer to finished drafts. I'll experiment by trying to link up notes alternatively from the topic register and from the annotated structure notes. Perhaps I'll report back if I discover anything interesting.

Sign In or Register to comment.