My Reading/Writing/Synthesis Process

In a conversation with David, the aspect of precisely how I read books came up, in that it seems that I am relatively uncommonly thorough in how I read the books I do, and that a lot of them have a reputation for being either dense or boring.

The process below is how I read a very specific type of book, namely the very dense academic non-fiction book. These are books that grapple with dense or abstract topics, that are frequently dry and relatively boring, but that I find tremendously worthwile to read. Some examples of these books are: Behind Human Error from Woods, Johannesen, Cook & Sarter, Sorting Things Out from Bowker & Star¹, Power In The System from Luhmann, Tyranny Of Metrics from Muller.

Notedly, this is not for fiction, things read for general fun and amusement, or fluff. This is a process that is optimised for learning, and if you find that fun, fantastic. I am relatively sparse in which books I apply this process to, and will often read lighter books with a much lighter process that boils down to “scribble some stuff down and maybe read it later”. For many lighter books, that is plenty. Some of those books are, for example, The Master Switch from Wu, The Scout Mindset from Galef, or Against The Grain from Scott.

Goals and Non-Goals.

First, let’s talk about the goals of this process. This is an unusually involved process, and so it has specific goals. They are, quite simply:

1. Make my brain engage with the book in front of my face. This is what the dialogue is for. Being forced to rephrase, reformulate and argue with the points the author makes prevents my brain from slipping clean off the letters in front of my eyes.
2. Understanding and retention. I am terrible at “reading” a book very quickly and forgetting 98% of it. Most people report having 10% retention of books and a vague “index”, ie getting a memory association that may not tell them what they are directly looking for, but tells them where to find it. I will wholesale forget books if I don’t engage with them deeply.
3. Writing. This entire process is oriented around writing, around forcing vague and abstract concepts into linearised words that force me to actually think about what I mean instead of what I’m gesturing at. Sometimes the writing is even for others to read, because selfishly, I like talking about the things I find interesting.

Conversely, there’s also non-goals, or things that I actively avoid optimising for. If they happen, cool, if they don’t, also cool.

1. Speed. I read terribly slowly. Some books take months for me to get through, reading 30-60 minutes per day. I have spent 2 months on a single 250 page book, and I’m currently on track to spend another 2 months on a 320 page book. I very much do not care about the speed. I am much more interested in digesting and dealing with the contents.
2. Ability to consider books by themselves. This is not something to review books on their own merits. This process is to pluck interesting insights out of them, and leave the rest behind. I will have opinions on the books I read, but not for their own, stand-alone merits. Instead the opinion will be formed by how many pieces it fits together inside the larger context of “books I have read”.

The Process itself.

To lay out the process, here’s the simple version:

1. Read with a notebook and pen in hand, take notes of the points of the book, always paraphrasing, never quoting. Write plentifully and freely, treat it as a sort of dialogue with the author(s). Rebut their points, rephrase their arguments, insult them for overlooking something. Read for 30-60 minutes at a time.
2. Add a bookmark to the block of notes. I tend to use one mark per half-page of notes or so, with the average reading session taking three. This is purely for granularity. I use a single tin of Book Darts, which simultaneously doubles as my WIP limit. Can’t have more outstanding note blocks than bookmarks.
3. As part of “reading time” (I tend to split mine halfways between actual reading, and reprocessing, with a bias towards reading), go over your marked notes, and rephrase/reconstruct them into a third medium. This can be a personal wiki, flat pages, pieces of paper, whatever works for you. The second transformation is the important bit. I use a custom tool for this². Once you have transferred a block of notes from second to third form, remove the bookmark.
4. After transferring and transforming the second form (the initial, written-while-reading dialogue with the authors) to the third location, you will notice many sparked connections and ideas between what was already in the third location and those you newly entered. Write those out, too, taking care to develop coherent reasoning and noting where things collide. Contradictions are valuable here, contradictions means that something interesting is afoot. It’s important to note that this writing is strictly internal, and therefore you should employ all the shorthand you have access to.

Eventually, you will have more than enough things jumping at you that make you think “I should write something more refined about this”, and then you have a free essay by taking all of your third-form writing, and just needing to polish it slightly.

A important but unmentioned element is the delay between 2 and 3. Since I use a tin of book darts for this, and most of them are typically in use, there’s about 2-4 weeks between writing the nodes and transferring them for a second time. In the beginning I saw this as a bug, and that the delay ought to be as small as possible, but over time it’s become clearer to me that this acts as a sort of spaced repetition: The delay between writing the dialogue and then transforming them again makes the material stick better in your mind, especially if you were generous in your notes and had quite a dialogue. I would recommend at least one week delay. I think 4 weeks as I have right now is too long, and I need to remove some book darts from the tin because of this.

A note on the fourth step: as your structure builds, the fourth step will take up more and more of your time, as you notice more and more things to connect, tie together, and write about. This is by design. Once it reaches a sufficient maturity you can, in theory, decouple steps 3 and 4, though I’ve found that I don’t create sufficient internal connections and have a weaker grasp on the domain. Therefore, I keep them together.

Origins of the process

Originally this came from Adler’s classic How To Read A Book, where he pleads to never read without a pen in hand, but does not actually give any instructions as to what to write. I winged it until I found something that worked for me, which is exactly this, before I then smashed it into Luhmann’s own process of cultivating a third form in something external and persistent.

Details Of Some Steps

The Notebook Notes.

They look like this, forgive the hard-to-read handwriting:

Bullet points, arguing with the authors, trying to figure out what they were trying to get at. You can also see that I read 12 pages in 30 minutes.

The Custom Tool

It’s named org-kasten. It’s an Emacs/org-mode based implementation of a fairly traditional Zettelkasten, based closely on Luhmann’s original design. It features a tree structure, easy traversal of said tree structure, arbitrary one-directional links, rapid editing, and a sense of place.

This is my persistent structure of destination, and the network effects can’t be underestimated. My rather small structure counts 1200 files, and I already have way, way more topics that I could potentially write about than interest or time for writing. Writing, however, feels good, and so every now and then I pick a particularly interesting strand from the structure and write about it.

---