The first time I read “A Pattern Language” by Christopher Alexander, I understood perhaps ten percent of it. The second time, a year later, I understood a different thirty percent. Not the same ten percent plus twenty more — a genuinely different set of ideas landed, because I had changed in the interval. The book hadn’t changed. I had.
This experience — of returning to the same material and finding it transformed — is so common that it barely seems worth remarking on. And yet our entire educational system is built on the opposite assumption: that learning is linear, that understanding is cumulative, and that once you’ve “covered” a topic, you can move on to the next one.
The spiral model
A better metaphor for learning is the spiral. You circle back to the same territory, but each time you arrive at a higher elevation. The view is different because your vantage point has changed. The landscape is the same, but your ability to see its features has developed.
Layers of rock and sea, built over millennia
This has practical implications for how we learn. It means that reading a difficult book once and moving on is almost always less effective than reading it twice with an interval of lived experience between readings. The interval is not wasted time — it’s the time during which your mind develops the structures needed to receive the book’s ideas.
The curse of prerequisites
Traditional education handles this problem with prerequisites: you must understand calculus before differential equations, data structures before algorithms. This linear model works for some domains, but it fundamentally misunderstands how deep understanding develops.
In practice, the best way to understand data structures is to need them while building something:
// You don't really understand a hash map until you need one.
// First encounter: "it's like an array but with string keys"
const cache = new Map<string, Response>()
// Second encounter: "oh, O(1) lookup matters here"
function findUser(id: string): User {
if (cache.has(id)) return cache.get(id)!
const user = db.query(`SELECT * FROM users WHERE id = $1`, [id])
cache.set(id, user)
return user
}
// Third encounter: "collision resolution, load factors,
// why Redis exists, why bloom filters exist..."
// Each pass reveals a deeper layer.The best way to understand calculus is to encounter a physics problem that requires it. Understanding develops from context, and context develops from experience, and experience is, by nature, non-linear.
The implication is uncomfortable: you will not understand many things the first time you encounter them, and this is not a failure. It’s the expected behavior of a healthy learning process. The first encounter plants a seed. The second encounter provides water. The third provides sunlight. Growth happens in between, invisibly, in the space where you’re not actively studying.
Active forgetting
There’s a related phenomenon that our productivity-obsessed culture handles badly: forgetting. We treat forgetting as a failure of retention, a bug in our cognitive hardware. But forgetting is actually an essential part of learning. It’s the process by which your mind separates signal from noise, keeping the structural understanding while releasing the details.
When you re-read a book you read five years ago, the experience of recognition — “I knew this, once” — is itself a form of learning. Your mind has maintained the shape of the idea without the specifics, and the re-read fills in the shape with new, richer detail informed by everything you’ve experienced since.
Spaced repetition systems try to hack this process, and they work well for certain kinds of knowledge (vocabulary, factual recall). But they miss the deeper point: some forgetting is productive. The ideas that matter most are the ones that survive the forgetting and re-emerge, transformed, when you encounter them again.
The impatience problem
The spiral model of learning requires patience, which is in short supply. We want to understand things now. We want to read one book and have the knowledge. We want to take one course and acquire the skill.
This impatience produces a specific failure mode: the illusion of understanding. You read a summary of a complex idea, understand the summary, and believe you understand the idea. But a summary is a compression, and compression is lossy:
Original idea (100%)
→ Blog post summary (40%)
→ Twitter thread (15%)
→ Quote retweet (3%)
→ "So true" reply (0%)
A path through the forest, each step revealing more
The nuances, the edge cases, the emotional resonance of the original — these are lost in compression, and they’re often where the real understanding lives.
The antidote to this impatience is not discipline, exactly. It’s trust. Trust that the understanding will come, that the repeated encounters will build on each other, that the confusion you feel now is a necessary precursor to the clarity you’ll feel later. This trust is hard to maintain in a culture that demands immediate results, but it’s the most important skill a learner can develop.
What this means in practice
If learning is spiral rather than linear, then several practical conclusions follow:
Re-read important books. Not immediately, but after enough time has passed that you’ve changed. Keep a list of books worth re-reading, and return to them on a cycle of years, not weeks.
Embrace confusion. When you encounter an idea you don’t understand, note it and move on. Don’t force comprehension. Let the confusion sit. Your subconscious will work on it, and the next time you encounter the idea — in a different book, a different context, a different conversation — another piece will fall into place.
Teach early. One of the best ways to accelerate the spiral is to try to explain what you’ve learned to someone else. The act of teaching reveals your gaps with brutal clarity, and those gaps become the focus of your next pass through the material.
Keep a journal. Not a summary of what you read, but a record of your reactions, questions, and confusions. When you re-read, read your journal entries first. The delta between your past understanding and your current understanding is itself a valuable form of knowledge:
## Reading Journal — A Pattern Language
### First read (2024-03)
- Overwhelmed by scope. 253 patterns feels arbitrary.
- Pattern 159 (Light on Two Sides) stuck with me. Why?
- Skeptical of the "generative" claims.
### Second read (2025-01)
- Now I see: patterns CONNECT. That's the whole point.
- The numbering isn't arbitrary — it's scale (region → room).
- "Generative" means: follow these and good things emerge
that you didn't explicitly design. Like CSS Grid.
- I was wrong about almost everything in my first read.
That's the point of this journal.The long game
Deep understanding is a long game. It unfolds over years and decades, not weeks and months. The people who know a field most deeply are not the ones who studied it most intensely for a short period — they’re the ones who returned to it repeatedly, each time with new experience and new questions.
This is not a comfortable message in a culture that prizes speed and efficiency. But it’s true, and knowing it’s true can save you from the anxiety of not understanding something immediately. You don’t need to understand it now. You need to encounter it now, let it plant its seed, and trust that the understanding will grow in its own time.
The spiral always comes back around.