Highlighting feels like studying but leaves almost nothing behind. A three-stage system for book notes — mark, collect, rewrite — that makes reading stick.
· Note-Taking Methods · 7 min read
Think of the last nonfiction book that genuinely impressed you — one you finished, recommended, maybe even called life-changing. Now try to state three of its arguments precisely enough that someone could disagree with you. For most readers, most books, the honest answer is the cover, the general mood, and one anecdote. The book left a feeling behind, but not its contents.
The instinctive fix is to read with a highlighter, and the instinct is half right: something should happen at the moments a book lands. But highlighting is where that impulse goes to die. The fix that works is a small system — three stages, each cheap, each doing a different job — and it starts with understanding why the highlighter fails.
Highlighting feels like studying because it is physical and selective: you judged this sentence to matter, and you did something about it. But what it trains is recognition — the ability to nod at an idea when it is in front of you — and what remembering requires is recall: reproducing the idea when it is not. Rereading a highlighted passage produces a warm sense of familiarity that is easy to mistake for knowledge. Psychologists call this the fluency illusion, and it is exactly why the passages you highlighted feel obvious in review and are gone a month later.
There is a second failure, visible in any long-time Kindle user's exported highlights: hundreds of beautifully clipped passages, never once revisited. Saving a sentence feels like acquiring the idea inside it, but the sentence is the author's — nothing of it passed through you. This is the collector's fallacy, and it is the same trap that fills read-later apps and sinks most personal knowledge systems: mistaking the accumulation of material for the processing of it.
Neither failure means marking passages is useless. It means marking is an input to note-taking, not the note. The three stages below are built around that distinction.
The worst time to take real notes is mid-chapter. Stopping to write a paragraph every few pages destroys reading momentum, and the notes it produces are scoped wrong anyway — reactions to sentences, captured before you can see the argument they belong to. Authors spend whole chapters building a case; a note taken on page 74 cannot know what page 90 does to it.
So while reading, do the cheapest thing that preserves the signal: a pencil line in the margin, a dog-eared corner, a highlight if the book is digital. One optional refinement — a star for "this changed my mind" and a question mark for "I doubt this" — and nothing more. The marks are not notes and will never be read as notes. They are bookmarks for a future conversation with the book, and their only job is to cost nothing, so that reading stays reading.
Give yourself permission to mark liberally. The filtering happens later, and it works better later.
Within a day or two of finishing — or after each sitting, for a dense book — open a single note for the book. Title and author at the top, then go back through your marks. For each one, ask whether it still earns a place now that you know the whole argument. Most will not, and that is the system working: the passage that seemed vital on page 74 often turns out to be scaffolding, and the distance between marking and collecting is exactly the filter a highlighter never applies.
For each mark that survives, write a bullet in your own words, with a page number. The own-words rule is the entire active ingredient, and it is worth being strict about. Paraphrasing forces you to actually have the idea rather than admire the author's phrasing of it — if you cannot restate the claim without the book open, you have located a gap in your understanding at the best possible moment to fix it, while the book is still in your hands. Copy a direct quote only when the wording itself is the point; more than a handful per book is transcription wearing a disguise.
An outline is the natural shape for this note, because books are arguments and arguments nest: a claim as the bullet, evidence and your objections indented beneath it, collapsible when you want the spine of the book at a glance. This is the stage where tooling starts to matter — in an outliner like Indenta, each claim collapses over its sub-points, and a #book tag plus the author's name makes every book note resurface together — though a paper notebook runs the same system with nothing but indentation.
One deliberate limit: this note is not a chapter-by-chapter summary. Summaries reproduce the author's structure, and the author's structure is what the book itself is for. Your book note keeps only what crossed the threshold of mattering to you — which is why it is retrievable by you later.
The first two stages preserve a book. The third is what compounds: within a week or so, pick the two or three ideas that connect to things you already think about — rarely more survive honest inspection — and give each one its own note, written as a claim in your own words, linked back to the book note and to whatever existing notes it touches. "This contradicts what that productivity book claimed about deadlines." "This is the same mechanism as the pricing note from March."
This is the move the Zettelkasten method formalizes as turning literature notes into permanent notes, and you do not need the full method to steal it. The point is a change of address: the idea stops living at "somewhere in chapter 4" — findable only by remembering the book — and starts living in your network of notes, findable by every topic it connects to. Ideas filed under books die with your memory of the book. Ideas linked to other ideas keep getting rediscovered.
Notice also what this stage quietly replaces: the reread. Nobody rereads books to retrieve ideas from them; the cost is too high, so unextracted ideas are simply lost. A linked note is the version of the idea with a retrieval cost of seconds.
A common stall: the book is good, but your marks are all "yes, agreed," and paraphrasing agreement feels empty. The unlock is that the most durable book notes are rarely neutral captures of content. They are reactions — disagreements and their reasons, surprises and what you previously believed instead, questions the book raised and did not answer, connections to your own work the author could not have known about. A note that argues with the book is doing elaboration twice over, and months later, "the book I thought was wrong about deadlines" is far more findable in memory and in search than any faithful summary.
If even reactions are not coming, that is data too. A book producing no notes at either stage is a book you can stop reading without ceremony.
The total cost is under an hour per book, almost all of it in stage two, and it replaces the diffuse guilt of "I should reread that sometime" with an artifact that actually answers questions. The measure of reading notes is the same as for any note system: not how complete they look on the shelf, but whether the idea you met in March turns up in November, when a problem finally needs it. Highlighting cannot pass that test, because the author's sentences were never yours. The rewritten bullet — awkward phrasing, page number, your objection attached — passes it, because the rewriting is where the idea became something you know.
Try it in practice: Indenta is a free, offline-first outliner — nested notes, backlinks, tags, and peer-to-peer sync, with no account required. Start writing in your browser, or read the User Guide first.