Crafting the Information You Capture in Main Notes
To establish relationships between informational units in a zettelkasten, units must become modular, able to move between different regions of your slip box, and into conceptual and spatial proximity with one another. The primary means for doing so is through the making of main notes. To understand how, weâll zoom in on the main note itself: a brief look at its components, followed by thoughts on how to craft the primary unit of information it contains, and how to reinforce it with additional information.
For more on the basics of main notes, check out my latest book A System for Writing: How an Unconventional Approach to Note-Making Can Help You Capture Ideas, Think Wildly, and Write Constantly.
What Is a Main Note?
A main note is a container designed to hold and transport information. Even more so than other note types, a main note's content is modular, able to move between different regions of your zettelkasten and into relationship with other content. A main note's content typically comprises some version of the following seven components:
- A primary unit of information
- Additional information reinforcing the primary unit
- Connections in the form of links or references
- A title summing up the information contained in the note
- A unique identifier rendered as either an alphanumeric ID or timestamp
- A quote for referencing the origin of your thinking
- Usage links pointing to places you've used the information
The image below shows a main note containing all the above.

At the core of the main note are the primary unit of information and any additional information reinforcing it, the central aspects with which informational units in other notes interact. Together these two units form the body of the note.
The Primary Unit of Information
The primary unit of information is an articulation of a concern.1 It's what the note has to say. Bateson might refer to it as the difference you've decided makes a difference.2 Typically one to five sentences in length, although as will be shown below this is not a hard and fast rule, the primary unit of information sits just below the title, capturing a single observation or insight. Everything else in the noteâadditional information, links, citations, usage, etc.âexist in service of it.
The primary unit of information should be, above all else, functional. It doesn't need to be a certain length, satisfying an arbitrary "size," nor does it need to capture the âessenceâ of an âideaâ reduced to its âbarest essentials.â The primary unit of information need only be:
- Articulated enough to be understood
- Reinforced enough to be reliable
- Limited enough in scope to be connectable and queryable
In order for a main note to be effective, the primary unit of information must be understandable. You must be able to return to it years later, and comprehend what it says, either in its original context, or in the context youâve placed it. Always consider your future self when writing the primary unit. Write clearly, concisely, and in complete sentences.
What you write must also be reliable, in the sense you should be able to return to it and trust what you wrote. Simply stating, "Bauhaus is great," won't instill confidence when you return to the main note, nor will it give you much to go on when it's time to write. Reinforce your primary unit of information by providing additional information in the form of context, citations, examples, reasoning, or contextualized connections to other notes.
Finally, the primary unit of information must be focused on a single concern. Limiting the conceptual scope of the note means you won't have to delete conflicting or divergent information when you connect the note to others. It also means the note will be queryable. A note focused on a single concern can be reliably surfaced when you search for it. By contrast, a note comprising multiple concerns may surface for the wrong search, or not surface at all when you need it.
The primary unit below satisfies all three criteria.
Merriam-Webster defines a train of thought simply (and in my opinion, correctly) as, "a series of thoughts or ideas that someone is having." Cambridge Dictionary offers a similar, albeit more concise definition: "a series of connected thoughts." At its most basic, a train of thought is a sequence of informational units.
Itâs readable, reliable, and focused on a single concern.
Types of Information
Primary units of information aren't all alike. A statement of fact and a definition are both valid primary units, but have different functions. Knowing what kind of information the primary unit is expressing changes what it can do and what you can do with it.
Unfortunately, many people don't have a vocabulary for knowing what kind of information theyâre working with. Instead they use the word "claim" as a catch-all for everything stated in a main note. But, claims are, by definition, nonspecific. They tell you something is being asserted without telling you what kind of assertion is being made. Are you attempting to state a fact, clarify a concept, or describe how parts fit together? Do you want to explain how something works, how to perform a task, or articulate a rule or way of being? A claim will always imply a particular kind of information, though rarely with much clarity.
As a way to better understand the kinds of claims we make, we'll borrow vocabulary from Robert E. Horn's "structured writing," which identifies seven basic types of information: concepts, principles, procedures, processes, structures, classifications, and facts.3 See the table below for quick reference.
| Information type | Description |
|---|---|
| Concept | Defines what something is |
| Principle | Articulates rules, guidelines, or ways of being |
| Procedure | Specifies how to do something |
| Process | Explains how something works |
| Structure | Describes parts and their relationships |
| Classification | Identifies how things group together |
| Fact | A statement assumed to be true |
Concepts are definitions. They define what something is, identifying the characteristics that distinguish it from something else. (E.g., "A main note is a container designed to hold and transport information.")
Principles articulate rules, guidelines, or ways of being. They tell us how to act, what to prioritize, or how things behave. (E.g., "A main note should be limited enough in scope to connect to other notes without having to be broken apart.")
Procedures specify how to do something. They detail steps or instructions needed to reach an outcome: do this, then this, then this. (E.g., "To create a main note, draft the primary unit of information, and add reinforcing material if needed. Then, assign a unique identifier, and link it to related notes stating why you've done so.")
Processes explain how something works. Unlike a procedure, a process isn't a set of instructions, but rather a description of how a system functions. (E.g., "As notes accumulate in a zettelkasten, clusters of interacting informational units begin to emerge.")
Structures describe parts and their relationship to themselves and to the whole. They show what a thing is made of, and how its pieces fit together. (E.g., "A main note consists of a primary unit of information, supporting material, a title, a unique identifier, and links to related notes. They may also contain quotes and links to where the primary informational unit has been used.")
Classifications sort different things into groups based on shared characteristics, unlike structures which show the relationship between parts of a single thing. (E.g., "Note-taking systems can be classified by how theyâre organized. Top-down systems, like outlines and hierarchical folder schema, impose structure in advance. Bottom-up systems, like the zettelkasten, allow structure to emerge from the notes themselves.")
Facts are verifiable, assumed-to-be-true statements. Dates, records of events, measurements, observed patterns, and documented outcomes can all be considered facts. (E.g., "At the time of his death, Luhmann's second zettelkasten contained roughly 60,000 notes.")
Horn's information types provide a useful vocabulary that can help you work with notes more deliberately. Once you know the kind of information youâre intending to express, you'll be able to identify which informational units lack articulation, which types you habitually gravitate toward, and which you tend to avoid. And, you'll have the language to make changes if need be.
It's important to note, however, not every primary unit of information will map cleanly onto one of the above types. Some will be a composite of more than one. Others will not be articulated enough to be recognizable as either. If you find yourself unable to identify a dominant function, it's worth asking whether the primary unit has been fully articulated, whether you've actually landed on what you want to say, or whether the note is still proposing something that is not yet settled. Make no mistake, having a zettelkasten filled with less-than-substantiated claims can still prove highly generative. Some of your notes can and will be in relatively rough shape. But, if you plan on having your thinking interact with that of other people, you'll eventually want to resolve your claims by crafting them into something more stable and identifiable. This can be done when crafting the note, or incrementally over time.
Crafting the Primary Unit of Information
Articulating the primary unit of information is a skill one develops over time, which can be honed by considering information types like the ones described above. To get a sense of how, we'll write a primary unit in the form of one information type, followed by writing the same as two others.
Let's say I come across the following passage from multidisciplinary artist, LĂĄszlĂł Moholy-Nagy, which I decide to include in my zettelkasten.
"This new beauty is not a style which matches one object with another aesthetically by using similar external forms (façade, motif, ornament). Today, something is beautiful if its form serves its function, if it is well-made of well-chosen material. A good chair will then 'match' a good table."4
Whenever I capture a unit of information I'd like to work with, my first task is to make sure I understand it both in the context of the broader text, as well as on its own. These are not always the same. When we isolate a unit of text, we decouple it from the context in which it was written. Because the excerpt above begins with "This new beauty...," I know it's referencing something just outside its boundaries. I make a mental note that "This new beauty" refers to "clearly constructed, functional materials", and assess whether or not the part I'm most interested inâMoholy-Nagy's take on matchingâis captured in what I've isolated. In other words, I make sure the excerpt is saying what I think it's saying.
Next, I assess what the passage is doing in the context of the original text. Many readers skip this step, by (perhaps unknowingly) flattening doing and saying into a single assessment. But, these are not the same. Assessing whether an excerpt is conveying factual information is different from assessing what factual information is being conveyed. The Moholy-Nagy quote above provides principle information (what itâs doing), because it dictates how objects should be seen in relation to other objects (what it's saying).
Next, I determine what I want to do with the information, which will begin the process of deciding what I want the information to say. Do I want to restate the passage in my own words? Do I want to comment on its main point? Do I want to use the passage merely for inspiration? For this example, I decide to restate it in my own words:
Household objects that go together (e.g., a table and chairs, knives and forks, etc.) "match" when each of the objects is made from quality materials and performs its function perfectly.
The above is entirely acceptable as a primary unit of information. It's understandable, reinforced by a credible source, and limited enough in scope to connect easily to other informational units. To arrive at this statement, I reiterated Moholy-Nagy's point, contextualizing it in something specific: home goods (what it says). Keeping the sentiment of the quote allowed me to retain its function (what it does). Like Moholy-Nagy, my primary unit of information conveys principle information. I'm dictating how to think about things.
But, what if I wanted the information to function differently? What if instead of restating Moholy-Nagy's position as a principle, dictating how one should think about the arrangement of household items, I wanted to state it as an observable fact. If so, I'd write it like this:
Matching household items according to functional rather than aesthetic qualities is commonplace. Enter any kitchen and you'll find mugs of one style matched with glasses of another, a kettle from a previous time period resting on a trivet from a later one.
The above is fact information, because it relays an observable pattern. Where the principle example tells us how to think, the fact simply points to what's been taking place.
Of course, I could convey the information as a concept, in which case I might write this:
Unlike aesthetic matching, which assigns objects to one another based on whether or not they share "similar external forms," functional matching is defined by objects relating to one another based on the quality of their materials and how well each object does its job. Two objects are functionally matched when each meets these standards independently.
The primary unit of information above satisfies the role of concept information, because it offers a definition of "functional matching." Unlike fact information, which demonstrates an observable pattern, and principle information, which dictates how one should think or behave, concept information states what something is.
Of the above three, concept information was the hardest to articulate. Not because definitions are inherently difficult, but because it required I orient my focus away from what Maholy-Nagy said and toward what he implied. I needed to step back and ask, "What is the thing Moholy-Nagy is defining?" The answer: âfunctional matching." In common usage, matching means visual similarity. Things look alike. But Moholy-Nagy is proposing something different. For Moholy-Nagy, matching means functional compatibility. Two objects match not because they share a style, but because they share a standard. My job in writing the above concept information was to take this and apply it to a new, identifiable concept.
We could go on reinterpreting the original quote in terms of procedure, process, structure, and classification information, each forcing us to adjust our perspective on how to engage with what Moholy-Nagy said. But, the best thing would be for you to do it yourself. To get a better sense of how you might want your primary unit of information to function, try articulating the original quote according to the remaining information types. See if you can get a sense of how each one feels and functions. As you do, keep in mind working with information types is less about classification than it is clarification. You don't need to check off every information-type box. You just need to know what it is you're trying to say. Deciding what the information you're conveying does will ultimately determine how you use it.
Adding Additional Information
Having a primary unit of information doesn't preclude a main note from containing additional information reinforcing it. In fact, additional information is often necessary when trying to articulate the point you're wanting to make. Additional information comes in two broad types: supporting and supplementary. Supporting information validates what the primary unit has already stated. Think: examples, reasoning, citations, or any information that reads like âproof.â Supplementary information adds to the primary. Think: context, definitions, explanations, and caveats. Neither supporting nor supplementary information should introduce new concerns. Both should work in service of the single concern stated in the primary unit.
Robert E. Horn breaks things down even further, pairing each information type with additional information he considers native to it.5 Concepts benefit from anything that sharpens the definition by showing what the thing is and isnât, often through examples and analogies. Principles benefit from relevant standards or caveats. Procedures and processes are both aided by tables, diagrams, and numbered lists. Structures often employ illustrations with labels. Classifications benefit from taxonomic lists and comparison tables that make groupings visible. Facts, being the most straightforward, typically benefit from statistics, dates, or other verifiable data.
The image below shows the main note following the "train of thought" note above with a primary unit of principle information followed by two additional units in the form of examples:

Additional information neednât be handled as separate text blocks like above. It can and often will be integrated into the main body of the note just as it might in written work. Below is an example from my notes:
Knowledge can be understood as a commodity, something to collect and use for social gain, or as a vehicle for transcendence (see jnana yoga). How we understand knowledge affects what we're looking for when reading, how we go about finding it, and what we do with it when itâs found.
The first sentence is the primary unit written as principle information. It describes how knowledge can be understood. The second sentence supports the first by presenting process information, describing how the results of our conception of knowledge unfold. Both work together to articulate the point being made. Of course, you might interpret things in reverse.
The example below handles information a bit differently, with the primary unit showing up after the supplementary information:
If your intention is to write a book, and ten notes get you there, then your system is functioning. If you intend to write a book, and ten thousand notes don't get you there, it's not a functioning system. In systems theory, the value of an "open system" is determined not by its size or complexity, but by how well it reaches a state of equilibrium, whereby the components regularly yield the intended results.
The primary unit of information in the third sentence dictates how to think about open systems. This is the point Iâm making. It may sound like concept information, but I havenât defined what an open system is, only how one should think about it. This distinction will prove useful when I go to write about the topic, because I know that if I want to express utmost clarity, Iâll have to eventually introduce a more formal definition.
The first two sentences supplement the primary unit by contextualizing it in process information. Here, Iâm describing how something may unfold. Itâs not procedure information, providing steps on how to perform a task. Itâs describing how things tend to happen over time.
In all cases, crafting units of information is less about how the information is structured, and more about how well you understand what each informational unit is doing. If youâre capable of understanding informational functions when blended together, feel free to structure your information in whatever way feels most natural. If, however, youâre less confident, separate primary information from additional information into different text blocks.
âEND EXCERPTâ
*If you enjoyed this piece, please pick up a copy of my latest book, A System for Writing: How an Unconventional Approach to Note-Making Can Help You Capture Ideas, Think Wildly, and Write Constantly. You may also sign up for my free newsletter HERE. To get my premium newsletter with all the insights, short essays, and zettelkasten scene weekly recaps, click HERE. To read more of my writing, click HERE. To learn more about me, click HERE. For my full website, click HERE.
Tietze, C. (2013, September 3). Create Zettel from Reading Notes According to the Principle of Atomicity. Zettelkasten. https://zettelkasten.de/posts/create-zettel-from-reading-notes/↩
Bateson, G. (1972). Steps to an ecology of mind. University of Chicago Press. See also Doto, B. (2025, December 25). Reading Ideas as Information: Sketches of a Theoretical Framework. Bob Doto. https://writing.bobdoto.computer/reading-ideas-as-information-sketches-of-a-theoretical-framework/↩
The more recognizable name for "structured writing" is Information Mapping, a proprietary methodology developed by Horn in 1969.↩
Goldman, J., Kolocotroni, V., & Taxidou, O. (1998). Modernism: an anthology of sources and documents. Edinburgh University Press.↩
Horn's key blocks often include counter examples. I've removed those here, since they are discussed in a later section of the book on working with discrepancies.↩