The Konik Method for Making Useful Notes

Outline

• Eleanors Readwise Settings

Highlights

Motivations: Start with the WHY
I generally come at the process of note making from one of three angles. They represented on the chart above as "ebooks," "I have a question" and "passive feeds" although in retrospect I should have said "hobby reading" instead of "ebook" there.

Pleasure Reading
The latter is the least common, but easiest to explain: sometimes I see a big book on an interesting topic and either buy it or put it on my wishlist, and then when I've got a lot of downtime and nothing better to do, or am particularly interested in it, I read it cover-to-cover. I highlight the interesting bits when they strike me, but I'm not reading for a purpose; I'm reading for pleasure.
It's sort of the middle of the road option that used to be mostly how I engaged with note making; highlighting and annotating a dead tree book on a topic of interest, and then moving on with my life once I was done, richer in knowledge but not particularly prone to remembering or using the information gained. It was a hobby, not meant to be productive, and I still do it sometimes.

Problem Solving
The next most common reason I approach information to make notes about is that I need to solve a problem. Sometimes the problem is "I need to decide on an email app" or "I want to build my dream house" or "how can I keep track of all the important things that happen with my kid?" – sometimes it's "I keep getting SEO hits for the Dark Smith of Drontheim even though it's only a casual mention from a six year old article about modern folklore like road gators, I should learn more about the dark smith so I can hit that SEO niche better."
Research topics like the latter typically send me into PDF collection mode.

Realistically, these sorts of notes are often less about complex workflows and academic deep dives than making sure I have the results of a casual search or conversation ready to hand the next time the subject comes up.

Everybody uses different vocabulary, but this is what I like to call active note making because it's the sort of thing I do when I have a specific goal I am working toward.

For me, "reactive" note making happens when I am basically killing time.
If I'm making my way across a long parking lot on my way to pick my son up from daycare, I'll sort my RSS feeds. If I'm waiting by myself at the doctor's office, I'll read an article or two from my queue.

Mostly, notes that come from people are things I jot down during meetings (where I almost always do use bullet points, although I don't bother with anything more fancy than a quick #priority if there's something important I need to do) and almost always refactor later by moving key information into the relevant places, like my email, calendar, or reference notes.

Maybe I'm looking for the answer to a question, maybe I'm just stumbling across something cool on Twitter, but if I stumble across a useful piece of information through one of my sources above, there are basically three things I can do with it:

1. Record it in a pre-existing note,
2. Create a new note,
3. or send it to Readwise to be dealt with indirectly.

My pre-existing notes come in two primary flavors:
• logs and indexes
• claims / evidence / explanations

Logs & Indexes
Logs and indexes, which are basically lists of things like "handy LaTeX guides," "scripts I need to access my server," "useful accessibility tools," "travel tips for Harper's Ferry" "books I'd like to read someday" and concept notes like "the Bronze Age" and "Chinese infrastructure" – they're basically places I shove information on a given topic in a moderately haphazard way, on the assumption that when I'm looking for information on Chinese infrastructure or getting ready to visit Harper's Ferry, the information I thought worth saving will be easy to find.

Claim, Evidence, Explanation
When I teach students how to write a paragraph appropriate for answering a document based question or long essay, the format we use has a bunch of different names and mnemonics depending on teacher preference and district habits. You can "spill the TEA" (topic sentence / evidence / analysis) or make a "PEE chain" (point / evidence / explanation) or "write a CER" (claim / evidence / reasoning).
Although I don't structure my notes as formal paragraphs (...usually), I do find this organizational method surprisingly useful. I've written about the value of consistent naming conventions at length before, but the short version is that for the kinds of notes I'm discussing here, I give them all file names that are claims.

The analog metaphor is pretty straightforward. If I were preparing to write a big paper, and was using index cards to help stay organized, I might scribble a quick claim on one side, basically putting the key information into my own words as concisely as possible, and at a size big enough to read comfortably at a glance. Then on the back, I would probably write a note about where I could find the source proving and expanding on that claim – i.e. a call number, title, author's name, page number, etc. Below the citation, I might write a longer explanation of how the information might be useful, along with any ideas about how I could use it. I might give the card a unique identifier, zettelkasten style, since it would be awkward to re-write the entire claim every time I wanted to reference it from another note – analog index cards don't have autocomplete, after all.
Once that's done, I would keep reading and looking for evidence of whatever I was trying to say, and do this many times until I have enough information to string together a decent paper.

Digitally speaking, I do something pretty similar to this. I put my claim as the title of the file, where it's easy to see when "flipping" through all of my notes, similar to how I can only see the front of an index card when flipping through a box of them. Then embed the relevant quote (because I have more space in a digital file than I do on a notecard, and it doesn't take me much time at all to copy the exact verbiage) in a way that provides me with breadcrumbs back to the source so I can provide that source to anyone questioning me, or double check that source if I need more context or stumble across reason to believe my original claim was wrong – either because I misinterpreted the evidence or someone else is able to refute it. The point is to always give myself a way to get back to the original source of my claim, and retain as much evidence as I need to prove my claim correct.

Lots of people say that you should paraphrase everything, and that the only things in your notes should be things you wrote in your own words. I've spoken about this before, but I think this is not only misguided but dangerous; attribution is important. So is actually engaging with information you're collecting so that you've got half a chance of retaining it and future you knows what the heck you saved it for, but I judge people a lot less for harmless hobbyist collecting of things they will neither understand nor use than I do for attribution errors, personally. Before you repeat a claim, you should at a minimum be in a position to give people access to the primary source so they have a better chance of evaluating stuff they learn on the internet.

Anyway, once I'm done recording my evidence and sourcing, I often (but not always - being realistic it's not always necessary, and I try to avoid unnecessary work) explain how I think I can use a particular claim and evidence combo. This is where links to other files come in handy, because then when I'm in the other file, my new note will pop up in the backlinks without me having to physically go to the old file and update it, the way I would need to do on paper

Linking to the author and publication is handy, even if I don't have anything else from those sources, because it helps me see which authors and sources I rely on most, and record why I do (or don't!) trust those sources.

Automation is often more trouble than it's worth. But often isn't always, and Readwise has been a major component of creating a database of interconnected, vetted information that is at my fingertips whenever I need to sit down and write

Step one is to highlight information I want in my notes. The internet is replete with advice about how to know what to highlight, but in my experience, there's no substitute for experience, which is to say metacognition. Highlight stuff, and pay attention to how you feel later.

Then step two is figuring out where to direct my attention, and this is where Dataview comes in.
TABLE file.inlinks AS Inlinks, file.outlinks AS Outlinks, file.tags AS Tags
FROM "10 Pending"
WHERE length(file.inlinks) > 0 OR length(file.outlinks) > 1 Or length(file.tags) > 2
SORT file.size ASC

This is a relatively simple query as far as Dataview goes. It creates a chart like the one below, of every file in my "Pending" folder (where the Readwise files go, among others) that is linked to, has links out, and has more than two tags – since all of my Readwise imports have two tags by default, as #pkm and #pkm/process count as two.

I sort them by file size on the rough assumption that a bigger file has more highlights and thus more useful stuff in it. It's easy enough to swap the sorting around, though, by changing the query. For example, here's my list sorted by number of outgoing links

Once I've selected an article, I need to fix it up a bit. I'm often in a hurry when I make annotations, and even the best parser misses things sometimes.

Eleanors Readwise Settings

My settings are here at this gist, but the main thing is that I really like having the highlight ID number as a unique identifier, because if I accidentally mess up and change the header incorrectly so it doesn't update across all the files I reference, I'll still be able to find it again with a simple search

What is this highlight ID benefit?

I personally prefer to use section headings and blockquotes because it feels semantically cleaner

I never want to be one of those people who who write extensively about note-writing but rarely have a serious context of use.

I like to change the metadata first. It's hard for computers to figure out the difference between an "author" and a "publication," but important to me

Next and most importantly, I review the highlights. As I read through the quotes, I make sure they still make sense, and go back to the original source and investigate them if they don't. If something still seems pointless after I reviewed it in context, I just delete it.

Once I've cleaned up any awkward formatting, weird links, or broken tags, I check to make sure the annotations are as useful as they can be.

Sometimes I highlight a quote in an article because it mentions some research I need to follow up on later, or because there's something I want to share with someone, and I wasn't in a position to do so while reading it the first time. My highlights are littered with notes to self and action items - it's not all pure knowledge.

For example, if I've left myself a note like #pkm/xref this reminds me of something the Carthage expert I like said, but I can't remember her name I will search my notes to figure out the name of the Carthage expert I like, cross-reference the highlight with things she said, and add links and update notes as appropriate. If I said something like This reminds me of the article about the guy a crane is in love with when I was taking notes on something without access to my notes, I will go find the article and link to my notes about it so that my backlinks and graph are updated.

Engaging
As much as I give the "paraphrase everything!" people a hard time, I do think it's important to rephrase, summarize, process, reinterpret put quotes into a more useful format. The problem with "summarizing" is that most people wind up in the habit of swapping out words and phrases for synonyms, enough to avoid pinging a plagiarism meter. This isn't enough to really make the quote useful, even if it does often help you confirm you understand it.
I prefer to consider the quote evidence for a claim, which necessitates using it to make a claim. Sometimes evidence can be used for multiple claims, which is great - but often means the quote should be broken out into smaller pieces.

Let's take a look at the first highlight from this article:
### id286148681

> Zandee did not know the nomads’ language, so his friend Gulzar translated as Zandee asked how much they made from selling the cheese. “Nothing,” Gulzar said. “They just do it so that the milk will not go stale. At most, they will barter it with a shopkeeper who may give them a pound of sugar or a dozen matchsticks.”
The ID number is just to give me something unique to search for, since you can't target a blank header and there's no point in coming up with my own random string of numbers for a block/header reference when Readwise already has one in its database.
This quote has a lot of context and vivid language, but my primary takeaway – the reason I highlighted it, even though I didn't make a note at the time (because it's obvious to me and I don't like to do unnecessary work) — is that nomads make cheese to preserve milk, or perhaps nomads make cheese to avoid wasting milk which is wordier but also more accurate.

Note that some of the headers are still metaphorically blank, with just the ID number; that's because some of these quotes are interesting context, but not necessarily things that I feel need to become anything more than what they are; context, or reference materials for where certain places are located. Not every highlight needs to become a "card" — and having things in my notes that are highlighted but don't need to be turned into further notes does not mean that I have failed by "over-highlighting."

Useful notes are messy, and time is valuable. A sense of completionism is nice for some things but mostly just gets in the way when it comes to learning. I read about 10% of the articles that show up in my feed. I take highlight about 50% of those, and affirmatively take notes on about 10% of those.

Some folks like to move the entire text of the highlight out into a new note, which is named after the heading, and then embed the new note back into the literature note. The advantage of this is that when you search for a string of text, like for example "Cheese," you get a note with a useful title ("Kashmiri people make goat and sheep cheese") instead of, say, the name of a book like "Middle Eastern Culture by Joana Dhoe" which doesn't do me as much good when I'm skimming file names in the search results trying to figure out what I have previously learned about cheese.
Despite the advantages of this method, though, I prefer to leave the quotes with the source, mostly because it makes it easier for me to send a file of quotes to a friend who asks for my takeaways from a specific book, for example my major takeaways from Oh Crap! Potty Training: Everything Modern Parents Need To Know... made the rounds in my offline social group and it was a lot easier to explain to non-technical people how to open and read a plain text file than how to navigate a bunch of nodes in a not-exactly-mainstream integrated thinking environment.

Anyway, the method I use to turn these claims into notes is pretty straightforward:

1. Make a new file and give it the name of your claim (minus the ID): nomads make cheese to avoid wasting milk
2. Embed the relevant section into your new file: ![[The Kashmiri Cheese Brand Operating at 7,000 Feet by Safina Nabi#id286148681 nomads make cheese to avoid wasting milk]]. A neat trick you can do with the IDs is copy it, then just type [[##id286148681 - the reference will pop right up. This is one reason I like to use the IDs, since they're unique, it keeps me from having to type through all the "nomads..." options that might possibly show up, if I have a lot of notes that start similarly (more common in a long file, like book notes for a specific topic like bananas).
3. Move over any annotations (I indicate this with the - [n] data-task bullet, but this isn't necessary) and add links to relevant connections if I want to. Seeing "nomads make cheese to avoid wasting milk" is more useful if I'm searching for information about nomads than "The Kashmiri Cheese Brand Operating at 7,000 Feet by Safina Nabi" would be, and with this method, my index file for nomads would have a bunch of claims neatly listed for me already, instead of a bunch of references I need to go re-learn from.

Since this can involve a lot of repetitive actions on a long file, I use this javascript code along with the QuickAdd plugin

follow the directions for how he uses his Kanban Script, but with the Zettelizer script instead. You can also follow the directions for this installation guide, which is a bit shorter.

Kanban? What’s this about, different script than the OG zettelizer?

Once I'm done running the script, I remove the #pkm/process tag and move the file from my Pending folder to my References folder.

ie update status to 🟨 or 🟩

Manifestations: Referencing Notes For A Purpose
That's the last step in terms of making notes, but the real trick is in using the notes.
If I'm planning a vacation, I search my vault for information I've gathered about local places. Every spring, I check the previous years' reflections on mishaps and successes before planting my garden. If I'm working on a story involving a fantastical creature, I reference the notes I made about the quirks of biologically similar animals that exist in the real world.
This whole workflow is designed around my very specific goals. I write fantasy and science fiction. I write newsletters sharing highlights from my research, and related essays about topics like the nature of identity in the ancient world. I also occasionally write scifi/fantasy adjacent nonfiction pieces, for example all the things that trees can be: from homes & magic roads to natural air conditioners.

up:: 📥 Sources