Andy Matuschak
patreon
On sitting down to study independently
🕑 Added 2024-08-31 20:19:58 +0000 UTCI’ve spent a lot of time these past two years talking with serious adult learners. That is, people independently studying technical subjects (e.g. physics) in support of some creative or professional project they really care about (e.g. their startup). I’ve noticed a surprising pattern to these conversations. We’ll spend quite a while discussing how to make learning efficient and effective: curriculum, comprehension, memory, practice, fluency, and so on. But then, an hour or two into the conversation, it will often emerge that despite their earnestness, the learner has trouble actually sitting down to do the learning.
All that earlier talk about comprehension and memory doesn’t really matter if the books aren’t getting cracked open. As I’ve run into this situation more and more, I’ve become convinced that it’s a seriously under-appreciated problem. Tools and resources often assume that the main thing a serious learner needs is clear explanation, or feedback, or effective practice. But for so many of these learners, the most immediate thing they need is help arranging their life and mindset so that they spend time learning what they intend to learn. In educational psychology, we’d say they need help with self-regulation.
I’m not talking about disaffected teenagers blowing off their history homework. These are smart, driven, creative adults, trying to learn topics of their own choosing, in support of their own creative goals. They intend to study. But then, week after week, they don’t. What’s going on here? Why is this so hard? And what can a designer do to help?
Where is it easier?
We can understand the problem better by looking at some situations where adult learners seem to have less trouble with self-regulation.
Synchronous courses. Even as they’re telling me about their struggles with independent study, learners will often tell me that they’ve enrolled in a course in some other subject that’s really working for them. Key elements seem to include: a regularly scheduled meeting time, a syllabus with clear scope and expectations, an energizing social environment, a respected facilitator, light accountability (via enrollment fee, social norms, etc). On the other hand, these courses usually suffer from functional learning issues (memory, comprehension, fluency); and they can often feel disconnected from the learner’s true aspiration.
1:1 tutoring. This seems to be a higher-variance strategy than a traditional course. The stronger relationship, personalized feedback, and sharper accountability often inspire more gumption. But for better or worse, most tutors have no syllabus. Scope, outcomes, and expectations are often vague and improvised, which leaves many learners feeling adrift or disconnected.
Peer study groups. Here in San Francisco, adult learners often gather with peers for regular evening study groups. The most common structure I hear is people studying separate topics individually, together in the same room, with a bit of socializing before or after. Helpful elements include the regular schedule, social accountability, and the pleasure of seeing friends. This structure seems to be even more effective when everyone’s working through the same text, because the social setting provides a meaningful way to engage with the material through discussion. But I do notice that these groups tend to fall apart after a few weeks or months. Stable long-term peer study groups seem rare.
Collaborative community groups. Stepping away from technical material for a moment, consider community music groups, sports teams, and acting troupes. These are like peer study groups in that they provide structure, social accountability, and community. But they’re also collaborative: the team or band succeeds or fails together. Members will feel each others’ growth, or lack thereof. When one member grows, that can inspire others to grow, too. When one member is struggling, the rest of the group is incentivized to help. These sorts of groups are common in the performing arts, but it’s less clear how to construct them for students of technical subjects.
Project-based learning / apprenticeship. When “just diving in” works, it seems to be most learners’ favorite approach. The learning is grounded and scoped by a real purpose; tinkering can replace textbooks; the project itself provides feedback and reinforcement. This approach also lends itself to a social setting, either with peers, or with more experienced mentors. Unfortunately, when conceptual load is high, this kind of on-the-fly learning often doesn’t penetrate deeply enough into conceptual foundations.
Scheduled performance. A reliable (if stressful) way to learn something is to volunteer to give a talk on the topic, or to teach a short course. I’ve seen a related structure in conversations with podcasters Dwarkesh Patel and Joe Walker. Both are known for deeply researched interviews, which often involve dozens of hours of study for each guest. In all these situations, there’s an impending date for a high-stakes performance, which creates urgency. The date also creates finitude: learners know when they’ll be “done”.
Stepping back, let’s look at some of the common elements these approaches provide:
Default schedule: Many of these structures construct a specific time for study. It’s often consistent and not easily changed. This inverts the default: you don’t need to keep deciding to study; if it’s on your calendar, you need to decide to cancel.
Clarity: What should I do next? How should I distribute my effort? How much is “enough”? Many of these structures offer answers.
Sense of progress: e.g. through checking off parts of a syllabus, producing work products, or conversations with peers or mentors.
Immediate reward: It may take a long time to learn enough for the real project you want to do, but these structures can give short-term meaning to your engagement with the material, e.g. through discussion with peers or practice projects. Those contexts may also make the material more immediately vivid and interesting.
Assurance: Many of these structures say in some way: “Don’t worry. This is the right thing to be doing. If you keep showing up, you’ll get where you need to go.”
Permission/legibility: “I can’t go Tuesday; I have [class/study group/tutoring] that night” is much easier to say than “I can’t go Tuesday; that’s one of my two free nights this week, and I’m trying to study physics two nights a week.” It’s also easier to tell yourself. The vessel offers a kind of narrative permission. Likewise, you might feel fine asking your boss to leave work early for a class or a study group on some relevant topic, but much less fine just sitting at your office desk reading a textbook.
Timeliness: As Taylor Rogalski has put it, you might have a good reason to learn that topic, but you don’t have a good reason to do it this Saturday. Even when these structures don’t include explicit scheduling, they create a reason for immediate action, generally through a short-term threat: study in a timely fashion, or face social cost. I’m wary of relying primarily on negative reinforcement—more on that later—but learners do need a good reason to study today, as opposed to someday.
When adult learners try to study independently, outside of one of the structures I’ve mentioned, they often end up with none of these elements. No default schedule, no clarity, no sense of progress, no immediate reward, no assurance, no permission, no timeliness. From this perspective, we can see just how much self-regulation is demanded of independent learners. They need to create and sustain enough of these elements to keep themselves going, all while also learning a difficult topic. No wonder they struggle so often!
Starting with time
Given all the discussion above, an independent learner should strongly consider joining or creating some structure like the ones I’ve discussed. Find a local class, study group, or tutor; identify a project you can use to drive the learning; invite your friends to a lecture you’ll give; etc. But many of the adult learners I’ve talked to have good reasons why these might not work, or might not be enough. The courses are too disconnected from their actual context of use; they’ve tried several tutors but none seem to click; their only available time is very early morning; they don’t know how to prepare to give a lecture; etc. What should such learners do?
In these conversations, the first thing I usually focus on is time. Often these learners have curated elaborate curricula and study tools for themselves, but they’ve neglected to set aside time to actually sit down and learn. One approach that often works is to create “study date nights”. That is: block a few hours on one or two nights per week for dedicated study. But then learners must defend those blocks against a never-ending barrage of social and practical encroachments. That’s often quite a challenge. If they succeed, learners create a vessel which can supply a few of the self-regulation elements we’ve discussed: a default schedule, timeliness, and even permission/legibility (“I can’t go; I set aside Tuesday nights to study math.”)
One common roadblock with the time conversation is that the learner simply doesn’t have any. They’re fully booked with professional and personal obligations. Sometimes it doesn’t seem that way—they have a few hours at night here and there—but every time they try to study at those times, the learner is too tired to focus properly. Happily, this situation is pretty straightforward. If the learner hopes to make meaningful progress, they must cut back on other activities enough to create a few hours of high-quality time on a regular basis. When exhaustion is the problem, the temptation is to somehow “figure out how not to be exhausted” during those late-night slots, but I’ve never seen that work. Learners are more successful when they secure permission from their workplace to study during the day, shift their schedule to study in the morning before work, or focus their study on weekends. There’s no point in talking about other more complex self-regulatory issues or study strategies until this problem is solved.
Contra “discipline”
After we sort out time, another issue often emerges in my conversations with adult learners. They’ve set aside time, but it’s not really working. They find themselves watching YouTube or doing email instead of reading the book they intended. Or they let themselves get pulled into some conversation or task around the house. Learners often express a lot of shame here. The sense is something like: “if only I were more disciplined!”
I think these learners usually make the wrong diagnosis when they prescribe themselves “more discipline”. But they’re right that something is wrong and needs to be fixed. Part of the trouble is that we use the word “discipline” in too many distinct ways. Let me get more specific.
Let’s start with a related classical virtue: temperance. Someone whose reason is ruled by ignoble appetites has a temperance problem. In the context of learning, the problem would be that the learner is so addicted to social media or Netflix that those appetites override their reason. In this case, yes, the learner needs to increase their control over their appetites, perhaps through some kind of corrective training. But I don’t think most of the learners I’ve talked to have this problem. The giveaway is that they don’t ordinarily lose hours to social media—it’s only when they sit down for a learning session that this problem flares up. Their appetites are usually in reasonable harmony.
For these adults, there’s something about independent learning, in particular, which makes their appetites more likely to dominate, or which makes them dissociate. Maybe what these learners need is more like grit—the tenacity to stick with a painful task? Classic stories of heroic grit might involve soldiers trudging through difficult terrain for days, or a single mother waking up before dawn every morning to finish a few hours of writing. In these stories, the task is necessarily painful, but the hero endures that pain in service of some higher cause. I don’t think this fits our adult learners very well. The same learner who’s struggling to focus on their own often has no real trouble when they’re in a course or study group.
I don’t think these learners need to become better at enduring necessarily painful tasks. Instead, we should ask why so much pain (or anticipated pain) surrounds independent learning for them. When I dig into that question, I find that learners usually have good reasons to avoid their studies:
Low confidence in plan. The learner lacks conviction that studying the materials they’ve chosen will actually help them with their real goal. Have they picked the right topics to focus on? The right texts? They fear they’re wasting their time.
Disconnect from authentic goal/interest. The learner is excited about working on brain-computer interfaces, but they’re mired in undergraduate calculus, and their excitement for BCIs doesn’t translate into excitement for epsilon–delta proofs. The material feels like an obstruction to be cleared before the learner can “get to the good stuff.”
Unclear progress. The learner shows up to study, but they can’t tell whether their efforts are producing any real results. Their goal is so far away that they can’t feel themselves getting any closer.
Sense of futility. The learner feels that they’re moving too slowly, that they’re not really understanding the material, that they’re not good enough. Feelings of incompetence, inadequacy, doom.
Lack of permission. The learner feels that they’re on borrowed time; they’re starting from a deficit. In some sense, they’re already supposed to know what they’re studying. This work “doesn’t count” as an achievement socially or professionally, and that’s a problem because the learner only feels secure when producing a steady stream of legible achievements.
“More discipline” isn’t the right antidote to these problems. If what you’re doing is useless, or futile, or meaningless, you shouldn’t persevere. In the last case—the feeling of insecurity—persevering might be a good idea, rationally, but railroading this felt threat will likely only produce more internal conflict over time.
A few healing themes
In all these cases, I think it’s worth taking the feeling seriously, and trying to address it. Why relate to independent learning as something to be endured through discipline and gritted teeth? Why set up a situation that depends on summoning willpower each time we sit down to study? The goal should be that we’re excited for the time we’ve blocked off on the calendar, this gift of curiosity and growth we’ve given to ourselves. I think it’s an attainable goal—though not necessarily an easy one.
Park ranger
Adult learners often struggle to find a good tutor for every topic they’d like to learn. Beyond undergraduate-level topics, the candidate pool shrinks considerably, and most technical experts aren’t eager to spend time tutoring on a regular basis. Most of them aren’t likely to be terribly good at it, either. But if we look at the list of problems above, or at the list of common elements in successful structures, I’m not sure tutoring is really what’s necessary.
A better metaphor might be the park ranger. A conversation with a park ranger can help you feel confident that you’re taking appropriate trails to get you where you’re going, taking into account your fitness and equipment. They can assure you that many other people in similar condition have happily made this hike without incident. If you tell the ranger that you’re particularly excited about waterfalls, they might be able to chain together a route which features them. You can ask why the trail takes that big ugly detour, and then feel better knowing that you’re avoiding a rockslide zone. The ranger can give you a map with their suggestions, and that’ll help you see your progress relative to the larger route.
The park ranger relationship demands much less from experts. These are questions you can ask in a short coffee chat, perhaps with a follow-up after a month or two. There’s no need to secure a weekly commitment. And this approach is more tolerant of an expert’s teaching inexperience. I’ve played this role to adult learners, connecting the dots between their authentic goal and learning materials, and my impression is that it’s been quite helpful, while demanding relatively little from me.
This concept relates to a proposal I made in “How Might We Learn”: that perhaps a large language model could offer this kind of guidance. I find that prospect quite exciting for more detailed information: not just suggesting which books to read, but which subsections, given the learners’ specific goals. But the confidence-building element is more emotional and relational—probably best left to humans. It seems to me that human “park rangers” are available in sufficient supply for higher-level guidance.
Recalibrating expectations
Many of the adult learners I’ve met are studying upper-level math or physics for the first time. These learners’ problems often stem at least in part from unrealistic expectations. First and foremost, they’re all experts in something, and they’re used to reading material in their own field. But for many of these learners, the explanations are just much more dense (especially in symbols and abstraction) than content in their own field.
So these learners often feel like they’re failing. They’re making progress too slowly, or else they try to read at their usual pace and then feel like they can’t understand anything. This creates a sense of futility and doom. But it’s an unnecessary sense of futility and doom. Modern information culture may make this worse: we’re constantly receiving the message that anything can be learned in a twenty minute video or a short blog post. Slowing down doesn’t pose any practical problem. If the learner works steadily, they’ll make plenty of progress. But the sense of effort, of really grappling with each page, is often unfamiliar and scary.
Sheldon Axler gives an appropriate warning in his introduction to Linear Algebra Done Right:
You cannot read mathematics the way you read a novel. If you zip through a page in less than an hour, you are probably going too fast. When you encounter the phrase “as you should verify”, you should indeed do the verification, which will usually require some writing on your part. When steps are left out, you need to supply the missing pieces. You should ponder and internalize each definition. For each theorem, you should seek examples to show why each hypothesis is necessary. Discussions with other students should help.
Unfortunately, I’ve met several people who have been frustrated by their slow progress in this very book. None of them could recall this bit from the introduction. I’m not sure it would have been enough if they did. At least here in San Francisco, the urge to speed-run everything is strongly embedded in the culture. If something can’t be speed-run, that’s often taken as a sign that it’s not the right thing to do. Courses and study groups do seem to counteract this effect, since they create common social knowledge that everyone is moving more slowly.
It’s worth noting that the medium of the book isn’t helping matters here. In the quote above, Axler lists all kinds of actions that the learner is supposed to take—on the side, outside the boundaries of the book, orchestrated by themselves. Surely the book could do more to make this methodical behavior feel like the natural default, perhaps through interactions which support comprehension. Books could also do more to help learners viscerally feel their progress, for instance by surfacing some representation of their accumulated learning, and by projecting when they’ll reach key milestones, as Math Academy does.
Enjoying the process
I’m talking to adult learners who are driven by a real goal: they’re studying in support of a project or question that’s really important to them. Their relationship to the material is totally different from a typical student in schools, studying to check a box on some dreary form or to avoid punishment.
But the fixation on that authentic purpose can be so strong that many of these adult learners don’t allow themselves to sink into the material along the way. All the prerequisites are just “in the way” of the project they want to do. If it emerges that a new topic must be added to the queue, that’s a new imposition added. I can understand this sentiment, of course, but it does mean that the learner is condemning themselves to countless hours trudging through obstacles.
Alternatively: once a learner sees that a topic will be necessary to their project, they could adopt a stance of cultivating curiosity about it. If the material’s not just an obstruction between the learner and their goals, its natural beauty can start to enchant. Concretely, I find it helps to start with an engaging video explanation of the topic, just to stoke enthusiasm, before diving into the denser materials which will actually build understanding. It also often helps people to talk about what they’re learning and what they’re wondering with peers, or on social media. Per Agnes Callard, “Twitter is a way to care about anything.”
Another approach, which I proposed in “How Might We Learn?”, would enrich the prerequisite texts with meaningful connections to the learner’s project. I’ve done this for students in conversation, and it seems to have helped a great deal. I suspect today’s large language models could go a long way here, though I haven’t experimented there yet.
For some learners, I think a real barrier here is a sense of permission. These people are quite curious by temperament. They’d love to really immerse themselves in the prerequisite topics. But they don’t feel like they’re allowed to enjoy the learning process. They have to start producing value; everything else is a frivolous luxury. This is a tough stance to untangle.
For some learners, the belief may be true. Their startup’s runway may really be quite short. In this case, I often suggest a combined approach: scrappily dash through whatever’s needed, just-in-time, while slowly and steadily backfilling a strong foundation for future projects.
But quite often, the belief isn’t actually true—it just feels true. One common story here is overriding concern for others’ perceptions of the learner and their productivity. All I can really suggest is that it’s a topic worth exploring with one’s therapist. Is there anything for a designer to do here? I’m not sure.
————————
If you’re an independent adult learner, studying material in support of something that really matters to you, I’d be curious to hear about your experiences with this issue.
————————
Thanks to Alec Resnick and Sara LaHue for helpful discussions, and to the many adult learners who have shared their adventures with me.
Comments
Andy Matuschak
Thanks for the kind words. You might like Execute Program for Python, though it assumes some familiarity with general programming concepts and focuses on the language specifically.
alexander gress
wow, this is sooo relatable, to the point of being weird! exactly the problems i faced — having more than enough motivations but not doing it. thank you for this post! on a side note, are you familiar with any platforms for learning coding (python) similar to what mathacademy does? discovered them from your post and absolutely adore the product! would be great to find something similar for coding
Adam Comella
Managing motivation and morale have been some of the most top of mind challenges for my own self-directed learning and I’ve spent a lot of time trying to improve at them. Here’s one example. I came across the math book "Turtle Geometry: The Computer as a Medium for Exploring Mathematics by Abelson & diSessa". The book’s description resonated with me: growing up, I got a lot of pleasure and understanding from teaching myself programming so I thought that, by approaching learning math in a similar style, maybe I could get a similarly deep understanding and pleasure from learning mathematics. (My primary experiences with learning math had been through high school and college classes . Despite getting good grades, I never felt that I achieved much pleasure or depth of understanding from them. But I suspected I could if I approached learning it with the right perspective.) I made two attempts at reading Turtle Geometry a couple of years apart. The second attempt was much more successful and I think the key difference was looking at the book from a different perspective. The first time I tried the book a couple of years ago, I viewed success as making progress through the book — reading & understanding the content, solving the exercises, learning useful things. After spending a number of hours on the section 1.1 exercises, I abandoned the book because I lost motivation — my progress felt extremely slow and I didn't see how learning these things would be useful. Recently, I read a book about the history of research into prime numbers (The Music of the Primes by Marcus du Sautoy) because I was curious to see what mathematicians enjoyed about math. I got the impression that many of the mathematicians enjoyed the process and didn't care if the work was useful in the real world. Things they didn't understand were mysteries that were interesting to think about and it was a joy to gain any insight about them. When starting the Turtle Geometry book again recently, I decided to adopt this sort of mentality. This time I viewed success as having interesting things to think about. It doesn't matter how slowly I'm progressing through the book — it's fine to pause the book and instead explore an interesting question that the book doesn't cover but has inspired. Instead of confusion being a negative experience because it slows progress, it has become more of a positive experience because it means I've found something interesting to think about. This attempt with the Turtle Geometry book was much more successful. I spent around 90 hours working on things related to the book. I think the attitude I followed here matches the philosophy described in the book's introduction. Even though I found the introduction's philosophy to be exciting the first time I read the book, for some reason I didn't manage to internalize it. Even the second time I sometimes struggled with the philosophy. Some days I felt demoralized because I didn't make any observable progress and I had to remind myself that I should feel satisfied because the important thing was that I spent the whole session thinking about things that were interesting and challenging to me. --- The math book "Proofs: A Long-Form Mathematics Textbook" takes a different approach to sharing its philosophy with the reader that I found to be effective. Instead of just stating the philosophy in the introduction, there's advice to the reader based on the philosophy interwoven throughout the whole book. For example, when the book presents a proof it might say something like "Before reading the proof, try to prove it yourself. This effort will lead to a deeper understanding." Or when presenting a definition it might say "Try to come up with some examples that match this definition." Initially, I felt skeptical because I had thoughts like "I read and understood this simple definition. What's the point of writing examples?" But I figured I could spend a few minutes trying the book's advice and, to my surprise, I actually gained insights. After this happened enough times, doing these sorts of things started to become almost an automatic habit for me. --- Various other aspects of the post resonated with my personal experiences but my comment is already quite long. I'd be happy to share more if anybody is interested.
