When I started studying the art of studying, I wanted to understand the role of book learning. How do we best learn from a textbook, scientific article, or nonfiction book? What can a student of average intelligence do to stay on top of their homework? Is it possible to improve your annual knowledge growth rate by one or two percent by learning how to learn? Should a motivated student take a maximizing or satisficing approach to their coursework? How many of the skills of a top scholar are strategic, collaborative, psychological, or involve merely a set of habits and technological proficiencies?

Fortunately, I started with the most esoteric of approaches, exploring visualization. I tried using a memory palace to memorize a textbook. It was vivid, fun, and creative. Exploring visualization helped me understand chemical diagrams, led me to invent a math problem, and made learning a lot more fun. But I simply couldn't jam that much detailed technical knowledge into my head. The method didn't help me pass my final exam, and I dropped it.

Posts from this era include Visual Babble and Prune, Using a memory palace to memorize a textbook, The point of a memory palace, Visualizing the textbook for fun and profit

After that, I explored speed reading. I read the theory, experimented both with physical technique and speed reading apps, and kind of broke my reading habits developing this difficult-to-correct tendency to skim. This tendency to read too quickly persisted long after I'd dropped deliberate attempts at speed reading. I finally made some intellectual progress, which preceded correcting the reading habit itself, in The Comprehension Curve.

Then I explored the world of Anki and tried to use flashcards to memorize a textbook instead (or at least a few chapters). After simulating the sheer amount of flashcard review I'd have to do to keep a strategy like that up long-term, I dropped that too. I felt that forming memories of narrow facts (like the structure of RNA polymerase or the name of the 7th enzyme in glycolysis) was the costliest way to learn. And I found the achievement of world-class memory champions irrelevant to real-world learning, which just seems like an entirely different task.

Posts from this area (not all on flashcards specifically) include The Multi-Tower Study Strategy, Define Your Learning Goal: Competence Or Broad Knowledge, Progressive Highlighting: Picking What To Make Into Flashcards, Goldfish Reading, Curious Inquiry and Rigorous Training, and Using Flashcards for Deliberate Practice.

During this time, I also played around with "just reading," without a conscious technique. Posts from this era include Check OK, babble-read, optimize (how I read textbooks), Wild Reading

Notes are cheap. It takes a lot less time to write down a fact than to memorize it. But I went further. I developed an elaborate and carefully-specified system of shorthand notation to represent causal, temporal, and physical structures. It used Newick notation for tree structures, variants on arrow signs to articulate causation, sequence, combination, and more, templates to rewrite the stereotyped information presented by textbooks in a uniform format, and hyperlinks in Obsidian to represent the relationships between concepts.

Not only did I take notes on the textbook, I also took notes on each individual homework problem. I also developed notes for other problems. I wrote Question Notes for The Precipice. This means that for each paragraph in the book, I wrote down one question to which that paragraph was a valid answer.

I never published any posts on note-taking. Partly, note-taking itself scratched that itch. But more importantly, it was a very fast iterative cycle. My methods developed day by day, over the course of months. I was experimenting with different software apps, tweaking the templates I used, figuring out how to expand my particular method of shorthand to represent complex structures. After all the shifts I'd made on my previous experiments, I thought I would spare LessWrong the tedious minutiae of my developing thoughts on note-taking. I'm confident that crafting the perfect notes in an elaborate and precise shorthand system is no a panacaea, so I don't know if it's worth bothering.

Exploring note-taking was as useful as visualizing was fun. The rigid structure of my note-taking approach gave me clear guidance on what it means to "read" or "study" a textbook chapter. They became a useful reference for looking things up. The idea of bringing together any data, formula, charts, or techniques I needed to solve a problem, and then making a plan of attack before setting to work, was a big upgrade for my accuracy and sense of ease.

Yet when my note-taking apotheosized after several iterations of improving my diagrammatic shorthand to deal with weird edge cases, and shifting from Evernote's WYSIWYG editor to Obsidian's markdown editor and full support for folders and hyperlinks, I found that not only was my approach to note-taking incredibly laborious, it was also profoundly distracting. It shifted my focus from building an intuitive feeling of understanding the material to constructing a precise translation of the material. At the end, I'd have a carefully notated description of a biochemical process, but virtually no ability to describe even the basics without reference to my notes. The experience of reading shifted from enjoyable, while visualizing, to the frantic skimming of flashcards, to sheer drudgery with note-taking. It didn't feel at all like programming, which is an activity I enjoy and that I'd hoped my note-taking would mimic.

It came back to me, then, after almost a year since I'd given much focused thought to visualization, that I should try just reading a chapter - no flashcards, no notes, no nothin' - and just try to picture everything as I went along, with no worries about trying to remember it all as I went. What do you know? The old spark returned! It was fun again! I breezed through a chapter on transcription, and had no trouble banging through the homework immediately afterward. Not only did I understand it better as I went, I was having more fun.

Now that I look back on the last year of exploring these issues, I see that I've only just now completed a single iteration of the Grand Study Problem, which is explaining how all these techniques, and possibly others, fit together into a technique for effective scholarship. Surely, it's partly about focused memorization (flashcards). Partly, it's about searching, note-taking, planning and problem-solving. And partly, it's about visualizing, anthropomorphizing, storytelling, model-building, and all the other ways of engaging your senses. What can I say about each of them?

If you're visualizing it, almost every textbook sentence provides you with an opportunity to create a new image in your mind. As you progress further through the textbook, it will call back to more and more earlier concepts. In biochemistry, it's things like the relationship between Gibbs free energy, enthalpy, entropy, and electrostatic potential; the amino acids; the nucleotides; different types of lipids; and a variety of major enzymes (i.e. DNA polymerase) and pathways (i.e. glycolysis). If you can figure out what those concepts are, and memorize them, you'll be able to picture them when it mentions them casually in passing. If you can't remember glutamine's abbreviation or chemical structure, then every time the book mentions G (or is it E?), you'll miss out on an opportunity to practice recalling it, or else you'll have to interrupt your flow to look it up for the umpteenth time. This is a role for flashcards and super-convenient reference charts. Some knowledge is most helpful if you can access it in five seconds or less.

Note-taking is incredibly helpful for focusing, but so is visualizing. I still think that there's a big role for taking really good notes, and assembling other reference and search tools. Yet taking notes needs to be balanced with enjoyable reading and building an intuition for the subject matter, and I think that comes from a visual approach first. Render it down into symbols later.

Along the way, I've written an informal scientific journal with my current working hypothesis, motivations for trying it out, tests, limitations, and future directions. This has been very helpful for giving these experiments a sense of direction.

One unifying trait so far is that each experiment has focused on one technique: visualization, memorization, note-taking, and now back to visualizing. It seems to me now that each of these has a purpose. Visualization puts the fun, creativity, and intuition in learning, and it's also fundamental to understanding anything that has a physical form. Memorization is important so that when you learn a concept in Chapter 2 that reappears persistently over the next 22 chapters, you aren't just reading words on the page, but are able to recall a concept to mind. That way, the rest of your reading refreshes and extends your memory of that initial concept. Note-taking and reference-sheet-making is helpful as a way of optimizing and compressing the natural-language, beginner-oriented version you get in a textbook into a format more suitable for review or looking up particular details. Figuring out how to interleave these three techniques will probably be the focus of my next iteration of this exploration.

New to LessWrong?

New Comment
13 comments, sorted by Click to highlight new comments since: Today at 6:26 PM

I personally find notes and visualizations kind of distracting while reading, the cue of trying to "visualize" something like the Schrodinger equation doesn't help much, visualizing the symbols themselves won't do anything for your understanding of it. So what I like to do is take a short break after every section in a chapter and ask myself "what does this knowledge imply?", "What is the next logical step to take from here?", basically trying to predict what the next pages in the book will contain. This makes it easier to see which leaps of logic in a book were easy and obvious, and which were much harder. It also provides immediate feedback and lots of "goddamnit, of course, how could I be so stupid?!" moments when I read the book take what is obviously the only possible next step.

Another idea for math study is to look at proofs and try to distinguish between steps that are just mechanical calculation vs steps that are leaps of insight.

With book-based arguments, like a work of philosophy, I find it helps to identify the point of a paragraph and then think beyond it. I like to do that with “Question Notes,” writing one question per paragraph where the paragraph in the text could be a valid answer to the question itself.

In an equation, I often start by imagining how the result changes as the variables change, or why the variables are as they are.

I wonder if this could be productive applied to strictly empirical findings, like the structure of RNA polymerase? Often I find that the textbook will present a structure, then point out the inferences from that structure to its function (and in biology, often the reverse - mRNA and genes were known to exist long before we identified their structure).

In any case, I agree that this is a whole other mode of relating with a text - one that I use, but haven’t focused on as a technique, perhaps because it’s not something that’s as easy to define and apply mechanically. Thanks.

Great post! You sound like a geometer and some people are algebraists. The latter seem to use interoception and internal speech more so than visualization. Using interoception to come up with new visual metaphors a la Gendlin's Focusing can be helpful for geometers IME.

Will you share your note-taking formalism and/or Obsidian workflow? I believe even if it's based aroubd your thinking and not that of the average reader, it would still be a much better starting place than how most of us take notes.

Thank you for the interesting article. I completely agree that curiosity ("the spark") is an important component of learning, and no technique will give it on its own. Have you experimented with learning one textbook or article at a time vs learning several concurrently (alternating between them)? If so, what are your conclusions on this? I know the relevant results of spaced repetition, the test effect, distributed practice vs massed practice, interleaving... but in practice how does it translate to a sustainable learning routine? How often do you change subjects when studying more than one thing at a time?

Thanks for reading! It’s hard to do a controlled experiment on this. Of course I’ve learned several subjects at once, but I haven’t compared, say, budgeting an hour a day for learning one topic until it’s done, vs budgeting half an hour a day for each of two subjects for the same amount of time. That’s what you’d really need to do to compare.

I’d say that the “multi tower study strategy” concept is an interesting hypothesis, not a settled conclusion :) I wrote it mostly to clarify my own conceptual thinking.

I might be atypical in this respect, but I find that the best time budgeting strategy for me is entirely different: I take 10 days and focus on a single textbook, I don't do much of anything else except take walks to think about the material and go back to read the textbook. I don't check the internet, don't check email, don't read any other books, don't go out with friends or work on other projects, I just focus all my time on that one textbook. This works best if I'm not doing this for an exam, but just for my own goals, an exam distorts the way that I read a textbook and makes me constantly think about gaming my grade.  

I started doing this by analogy with meditation retreats. It's very well-known in meditation circles that a 10-day intensive retreat makes you progress dramatically faster than a simple daily practice. So I started doing study retreats on a single topic. It takes about 3 days to really get into a groove with the schedule of studying 10 to 12 hours a day, but after that it gets easier. Really small distractions have a dramatic impact, even just talking to someone else for 10 minutes have a perceptible impact on the feeling of study momentum. 

A similar approach has worked for me better than a more split-time approach. I'm aware of the forgetting curve and I certainly forget a lot of the contents afterwards, but the global structure seems to remain in the brain and changes to the way of thinking or of solving problems after these intense study sessions also seem to remain for longer than the details.

I've also tried doing some incremental reading / incremental learning and although the contents stay for longer, I don't feel the same kind of enlightenment or learning taking place. It feels a bit like wasting time, even if I'm learning.

I don't know how you'd approach maintenance for skills you acquired but forgot. Sometimes I've learnt something which has the skill I want to review as a prerequisite, using the same method, but reviewing the old material as needed, and it sort of did the trick.

Right now, I'm experiencing a miniature version of this. I'm learning about how optogenetics works. This technique depends on an understanding of both neuron action potentials and G Protein Coupled Receptors, which has forced me to review each of those structures. This in turn forces me to review the structures of the various molecules and enzymes involved, which forces mild review of even deeper precursors such as Glutamine-Histidine-Serine structures in active sites of enzymes. I imagine that if you weren't cramming for an exam, and were genuinely interested in the subject matter, and were consciously trying to develop your "mental movie" to build understanding, this would be the natural approach to take.

In general, I really wonder to what extent our educational system's need to test and measure students has operationalized "learning" in a way that's deeply different from what would be optimal for, say, producing competent scientists.

Fascinating. How would you manage this in the context of a romantic relationship, job, or other daily commitments? Is it possible for you to take the remaining 4-6 hours to connect with other people? Or does it demand solitude?

It really wouldn't be possible to take 4-6 hours to talk to other people, that would completely take you out of it, it would kill a lot of momentum. After 3 days of really spending every waking hour thinking of a subject, it sort of temporarily becomes the new baseline, and the difficulty of thinking about it drops really dramatically, but it is a fragile effect, any context switching at all comes at the cost of a blunting of momentum, and this is especially true of talking to other people. If you've spent a lot of time just thinking in silence, 4 hours of talking with others will literally cause a pounding headache. I took about 3 to 4 hours per day of break-time to eat and go walk without feeling the need to think of the material, but in the later part of the 10-day period I usually think about the textbook even during breaks. I'm not sure how you'd manage this with a job and daily commitments, what I do is do one of these 10-day periods every 2 months, but I run a small business that allows me to do that, and I don't have a romantic relationship.

Very interesting. It seems important to understand the relationship between the "initial contact" with knowledge and maintenance activities later on. 

For example, 12 hours/day for 10 days is 120 hours. By contrast, a conventional class demands about 300 hours of work over the course of a semester. If we consider both of these to constitute an "initial contact" with knowledge, the 10 day retreat is much more efficient.

Then we have to ask whether one or the other leads to more efficient maintenance over the long run. Plus, on an instrumental level, we absorb knowledge not only to accumulate it, but also to determine what sort of projects to pursue and how to specialize ourselves.

Figuring out how a 10-day single-subject retreat vs. a half-year spread-out interleaving of learning impact later maintenance and ability to choose and execute projects would be an important aspect of deciding which approach to "initial contact" is optimal.

Since a graduate program basically consists of classes + research, and the classes are pretty much all "initial contact," it would seem that if the 10-day retreat was more efficient, that you could replace a 2-year grad program with maybe 9 months of retreats. But I'm not sure if that sort of lifestyle seems optimal somehow... (low confidence on all of this!)

As a followup, it does seem like you could test this to some extent. Find out what textbook an upcoming year-long class uses. Take 10 days prior to the class for a retreat, during which you read the entire textbook. Then try to do the homework and exams with minimal review of the textbook, treating the class as a review of material you're already familiar with rather than a first brush with the content. Since textbook reading comprises the bulk of my studies, it seems possible that this would feel like a net time saving/deeper learning, but I'm not sure. Would be interesting to try it and see!