Marginalia is note-taking in the margins of the thing you are reading. It predates every method on this wiki, it is how Coleridge, Fermat, and generations of scholars worked, and it has one enormous advantage no digital system can copy: the note is written at the exact moment of the thought, six inches from the sentence that caused it.
It also has one enormous weakness, which is that the note stays in the book. Everything below is about keeping the advantage and fixing the weakness.
Why writing in the book works
Mortimer Adler made the case in 1940 and it has not been improved on. His argument was that marking a book is not vandalism but the physical evidence of reading actively rather than passively, and that an unmarked book has usually been consumed rather than read.
The mechanism is straightforward. To write in the margin you have to form a reaction, compress it to a few words, and place it. That is generation and elaboration in one move, and it is the difference between reading and having read. Highlighting fails precisely because it demands none of that.
A margin system that survives contact
The point of a symbol set is to record your reaction without breaking the reading. Keep it small and fixed.
| Mark | Means |
|---|---|
| Vertical line in margin | The passage matters. Length of line shows how much |
| Double vertical line | Central to the book's argument |
? |
I do not follow this, or I do not believe it |
! |
Surprising, contradicts what I thought |
→ p.84 |
Connects to something earlier in the book |
| Circled word | Term to look up or define |
| Numbers in the margin | Steps of an argument being made in sequence |
| Words in your own hand | The actual note. Everything above is just a pointer |
Two habits do most of the work. Underline sparingly so the marks retain signal, and write the connection, not the content. "Same as Kahneman's system 2" is worth ten underlines, because it is the part the book does not contain.
The most valuable page in a marked-up book is usually the inside back cover, where you keep a running index of your own: page numbers plus three words. It is not the Zettelkasten method, but it is the same instinct with a hundredth of the setup.
The retrieval problem
Marginalia dies in the book. You wrote a sharp reaction on page 212 in 2019 and you will never see it again unless you happen to reread. Every book on your shelf is an unindexed archive.
There are three honest fixes.
One, transcribe the good ones. After finishing, go through your marks and copy out the ones that still earn attention. This is a slow, high-value pass, and it is the step described in how to take notes from a book. Most marks will not survive it, which is fine. The filter is the point.
Two, photograph the pages. Faster and worse. You get an archive you can search only by memory, since a photo of your handwriting is not text. Better than nothing, worse than transcribing.
Three, read digitally and export. Kindle, Apple Books, and most PDF readers export highlights and notes as text you can move into your notes app. You lose the tactile advantage and the freedom to scribble anywhere. You gain a searchable record, which over years is a large gain. The trade-off in full is digital vs paper notes.
When not to mark the book
Library books and borrowed books, obviously. Books that are genuinely valuable as objects. And books you are reading for pleasure, where stopping to annotate ruins the thing you came for.
Sticky notes and a pencil are the standard workaround, and they are worse than margin ink in one specific way: they fall out and they let you postpone forming the thought. A tab that says "interesting" is not a note.
What it is good for, honestly
Marginalia is the best method available for a single dense book you intend to argue with. Philosophy, technical texts, anything where you need to track an argument across chapters.
It is a poor method for synthesis across many sources. Fifteen books each holding their own private notes is not a body of knowledge, and no amount of marginal skill fixes that. The moment your work depends on connecting sources, the notes have to leave the books and become atomic notes or literature notes somewhere they can sit beside each other.
That is the real division of labour. Mark the book while reading, because nothing beats it for attention. Move the survivors out afterwards, because nothing in a book will ever find you again. The second half is the one people skip, and it is why so many shelves full of beautifully annotated books produce nothing.
Related reading: how to remember what you read, progressive summarization, and the wider set of note-taking methods.