by Scott Young
Using This Analysis to Draw Your Map
Once you’ve finished your brainstorm, underline the concepts, facts, and procedures that are going to be most challenging. This will give you a good idea what the major learning bottlenecks are going to be and can start you searching for methods and resources to overcome those difficulties. You might recognize that learning medicine requires a lot of memorization, so you may invest in a system such as spaced-repetition software. If you’re learning mathematics, you might recognize that deep understanding of certain concepts is going to be the tricky spot and consider spending time explaining those concepts to other people so you really understand them yourself. Knowing what the bottlenecks will be can help you start to think of ways of making your study time more efficient and effective, as well as avoid tools that probably won’t be too helpful to your goal.
Often this coarse-grained analysis is enough to move on to the next phase of research. However, with more experience, you can dig deeper. You might look at some of the particular features of the concepts, facts, and procedures you’re trying to learn to find methods to master them more effectively. When I started my portrait-drawing challenge, for instance, I knew that success would depend highly on how accurately I could size and place facial features. Most people can’t draw realistic faces because if those attributes are off even slightly (such as making a face too wide or the eyes too high), they will instantly look wrong to our sophisticated ability to recognize faces. Therefore, I got the idea of doing lots and lots of sketches and comparing them by overlaying the reference photos. That way I could quickly diagnose what kinds of errors I was making without having to guess. If you can’t make these kinds of predictions and come up with these kinds of strategies just yet, don’t worry. This is the kind of long-term benefit of metalearning that comes from having done more projects.
Answering “How?”
Now that you’ve answered two questions—why you’re learning and what you’re learning—it’s time to answer the final question: How are you going to learn it?
I suggest following two methods to answer how you’ll learn something: Benchmarking and the Emphasize/Exclude Method.
Benchmarking
The way to start any learning project is by finding the common ways in which people learn the skill or subject. This can help you design a default strategy as a starting point.
If I’m trying to learn something that is taught in school, say computer science, neurology, or history, one thing I’ll do is look at the curricula used in schools to teach that subject. This could be the syllabus from a single class or, as in the case of my MIT Challenge, the course list for an entire degree. When I wanted to learn more about cognitive science, I found a list of textbooks that the University of San Diego’s Cognitive Science doctoral program recommends for incoming students without cognitive science backgrounds. Good resources to consider for this approach are universities (MIT, Harvard, Yale, and Stanford are good examples but far from the only ones). Generally course lists and syllabi are available by looking on their websites aimed at existing students.
If I’m trying to learn a nonacademic subject or a professional skill, I’ll probably instead do online searches for people who have previously learned that skill or use the Expert Interview Method to focus on resources available for mastering that subject. An hour spent searching online for almost any skill should turn up courses, articles, and recommendations for how to learn it. Investing the time here can have incredible benefits because the quality of the materials you use can create orders-of-magnitude differences in your effectiveness. Even if you’re eager to start learning right away, investing a few hours now can save you dozens or hundreds later on.
The Emphasize/Exclude Method
Once you’ve found a default curriculum, you can consider making modifications to it. I find this easier to do with skills that have obvious success criteria (say drawing, languages, or music) and for which you can generally make a guess at the relative importance to the subject topics prior to studying them. For conceptual subjects or topics where you may not even understand the meaning of the terms in the syllabus, it’s probably better to stick closer to your benchmark until you learn a bit more.
The Emphasize/Exclude Method involves first finding areas of study that align with the goals you identified in the first part of your research. If you’re learning French with the idea of going to Paris for two weeks and speaking in shops and restaurants, I would focus a lot more on pronunciation than being able to spell correctly. If you’re learning programming solely to make your own app, I’d focus on the inner workings of app development more than theories of computation.
The second part of the Emphasize/Exclude Method is to omit or delay elements of your benchmarked curriculum that don’t align with your goals. For example, one common recommendation for learning Mandarin Chinese, advocated by people such as the renowned linguist and Sinologist Victor Mair, is to focus on learning to speak before you try to read characters.6 This isn’t the only route available, but if your main goal is to speak, then this path to fluency might be more effective.
How Much Planning Should You Do?
One question you may face is when to stop doing research and just get started. The literature on self-directed learning, as typically practiced, demonstrates that most people fail to do a thorough investigation of possible learning goals, methods, and resources.7 Instead they opt for whatever method of learning comes up naturally in their environment. This clearly leaves a gap between what is practiced and the efficiency that is possible using the best possible method. However, research can also be a way of procrastinating, particularly if the method of learning is uncomfortable. Just doing a bit more research then becomes a strategy to avoid doing the work of learning. There will always be some uncertainty in your approach, so it’s important to find the sweet spot between insufficient research and analysis paralysis. You know when you’re procrastinating, so just get started.
The 10 Percent Rule
A good rule of thumb is that you should invest approximately 10 percent of your total expected learning time into research prior to starting. If you expect to spend six months learning, roughly four hours per week, that would be equal to roughly one hundred hours, which suggests that you should spend about ten hours, or two weeks, doing your research. This percentage will decrease a little bit as your project scales up, so if you plan to do five hundred or a thousand hours of learning, I don’t think it necessarily demands fifty or a hundred hours of research, but maybe closer to 5 percent of your time. The goal here isn’t to exhaust every learning possibility but simply to make sure you haven’t latched onto the first possible resource or method without thinking through alternatives. Prior to starting my MIT Challenge, I spent roughly six months, part-time, combing through all the course materials. A good idea is to be aware of the common methods of learning, popular resources, and tools along with their strengths and weaknesses before starting. Long projects provide more opportunities for getting derailed and delayed, so doing proper research in the beginning can easily save a much larger amount of time later on.
Diminishing Returns and Marginal Benefit Calculation
Metalearning research isn’t a onetime activity you do only before starting your project. You should continue to do research as you learn more. Often obstacles and opportunities aren’t clear before you start, so reassessing is a necessary step of the learning process. During my portrait-drawing challenge, for instance, I discovered about halfway that I was getting diminishing returns from my sketch-and-compare method. I realized that I needed a better technique for drawing that had higher accuracy. That led me to do a second round of research, leading to a course taught by Vitruvian Studio, which detailed a more systematic method that greatly increased my accuracy.8 I hadn’t noticed it in my original research because I wasn’t aware of the deficiency of my self-developed technique.
A more sophisticated answer to the question of when and how to do research woul
d be to compare the marginal benefits of metalearning to regular learning. One way to do this is to spend a few hours doing more research—interviewing more experts, searching online for more resources, searching for new possible techniques—and then spend a few hours doing more learning along your chosen path. After spending some time on each, do a quick assessment of the relative value of the two activities. If you feel as though the metalearning research contributed more than the hours spent on learning itself, you are likely at a point where more research is still beneficial. If you felt that the extra research wasn’t too helpful, you’re probably better off sticking to the plan you had before. This type of analysis depends on something known as the Law of Diminishing Returns. This states that the more time you invest in an activity (such as more research), the weaker and weaker the benefits will be as you get closer and closer to the ideal approach. If you keep doing research, eventually it will be less valuable than simply doing more learning, so at that point you can safely focus on learning. In practice, the return to research tends to be lumpy and variable. You might spend a few hours and get nothing, then stumble onto the perfect resource for accelerating your progress. As you finish more projects, it’s easier to judge this point intuitively, but the Law of Diminishing Returns and the 10 Percent Rule can provide good approximations for how much research to do and when.
Long-Term Prospects for Metalearning
So far we’ve talked only about the short-term benefits. However, the real benefits of metalearning aren’t short term but long term. They don’t reside in a particular project but influence your overall strengths as a learner.
Each project you do will improve your general metalearning. Every project has the opportunity to teach you new learning methods, new ways to gather resources, better time management, and improved skills for managing your motivation. Success in one project will give you confidence to execute your next one with boldness and without self-doubt and procrastination. Ultimately, this effect far outweighs the effect of doing a specific project. Unfortunately, it’s also something that can’t be boiled down to a tactic or tool. Long-term metalearning is just something you acquire with experience.
The benefits of ultralearning aren’t always apparent from the first project because that first project occurs when you’re at your lowest level of metalearning ability. Each project you complete will give you new tools to tackle the next, starting a virtuous cycle. Many of the ultralearners I interviewed for this book told me a similar story: that they were proud of their accomplishments in individual projects but that the real benefit had been that they now understood the process of learning hard things. That gave them the confidence to pursue other ambitious goals that they wouldn’t have even considered previously. This confidence and ability are the ultimate goals of ultralearning, even though they’re often hard to see from the outset. These benefits, however, can be achieved only by putting in the work. The best research, resources, and strategies are useless unless you follow up with concentrated efforts to learn. That brings us to the next principle of ultralearning: focus.
Chapter V
Principle 2
Focus
Sharpen Your Knife
Now I will have less distraction.
—Leonhard Euler, mathematician, upon losing the sight in his right eye
If ever there were an unlikely candidate for scientific greatness, it would have been Mary Somerville. She was born into a poor Scottish family in the eighteenth century, when higher education was not seen as proper for a lady. Her mother did not prevent her from reading, but society at large did not approve of it. An aunt, seeing that behavior, commented to her mother, “I wonder you let Mary waste her time in reading, she never sews more than if she were a man.” When she did have an opportunity to attend school briefly, her mother regretted the expense. Somerville explained, “she would have been contented if I had only learnt to write well and keep accounts, which was all that a woman was expected to know.”1 As a woman, she faced even larger obstacles, with household duties and expectations taking precedence over any kind of self-education. “A man can always command his time under the plea of business, a woman is not allowed any such excuse,” she lamented. Her first husband, Samuel Greig, was strongly against learning in women.
Yet despite those obstacles, Somerville’s accomplishments were vast. She won awards in mathematics, learned several languages to fluency, and knew how to paint and play the piano. In 1835, she, along with the German astronomer Caroline Herschel, were the first women elected to the Royal Astronomical Society. The accomplishment that eventually brought her fame was her translation and expansion of the first two volumes of Pierre-Simon Laplace’s Traité de mécanique céleste, a massive five-volume work on the theory of gravitation and advanced mathematics, acclaimed as the greatest intellectual achievement since Isaac Newton wrote the Principia Mathematica. Laplace himself commented that Somerville was the only woman in the world who understood his work.
The easiest explanation for the vast discrepancy in Somerville’s situation and her accomplishments would be genius. It is no doubt true that she possessed an incredibly sharp mind. Her daughter once commented that while she was being taught, her mother could grow impatient. “I well remember her slender white hand pointing impatiently to the book or slate—‘Don’t you see it? There is no difficulty in it, it is quite clear.’” However, in reading through her descriptions of her life, this seeming genius was beset by many insecurities. She claimed to have “bad memory,” recounted struggles learning new things as a child, and had even at one point “thought [herself] too old to learn to speak a foreign language.” Whether that was polite modesty or genuine feelings of inadequacy, we cannot know, but it does at least put cracks in the idea that she approached learning from a place of unshakable confidence and talent.
Peering deeper, another picture of Somerville emerges. She had a keen intellect, yes, but what she possessed in even greater quantities was an exceptional ability to focus. As an adolescent, when she was put to bed and denied a candle for reading, she would mentally work through the works of Euclid in mathematics. While still breastfeeding her child, an acquaintance encouraged her to study botany, so she devoted “an hour of study to that science” every morning. Even during her greatest achievement, the translation and expansion of Laplace’s Traité de mécanique céleste, she had to carry out all the household duties of raising children, cooking, and cleaning. “I was always supposed to be at home,” she explains, “and my friends and acquaintances came so far out of their way on purpose to see me, it would have been unkind and ungenerous not to receive them. Nevertheless, I was sometimes annoyed when in the midst of a difficult problem one would enter and say, ‘I have come to spend a few hours with you.’ However, I learnt by habit to leave a subject and resume it again at once, like putting a mark into a book I might be reading.”
In the realm of great intellectual accomplishments an ability to focus quickly and deeply is nearly ubiquitous. Albert Einstein focused so intensely during his formulation of the general theory of relativity that he developed stomach problems. The mathematician Paul Erdős was a heavy user of amphetamines to increase his capacity for focus. When a friend bet him that he could not give them up, even for a short time, he did manage to do so. Later, however, he complained that the only result had been that mathematics as a whole was set back a month in his unfocused absence. In these annals of extreme focus, one often conjures up an image of solitary geniuses laboring away without distraction, free from worldly concerns. However remarkable this is, I’m more interested in the kind of focus that Somerville seemed to possess. How can one in an environment such as hers, with constant distractions, little social support, and continuous obligations, manage to focus long enough not only to learn an impressive breadth of subjects, but to such depths that the French mathematician Siméon Poisson once remarked that “there were not twenty men in France who could read [her] book”?
How did Somerville become so good at focu
sing? What can we glean from her strategies in getting difficult mental work done in less-than-ideal conditions? The struggles with focus that people have generally come in three broad varieties: starting, sustaining, and optimizing the quality of one’s focus. Ultralearners are relentless in coming up with solutions to handle these three problems, which form the basis of an ability to focus well and learn deeply.
Problem 1: Failing to Start Focusing (aka Procrastinating)
The first problem that many people have is starting to focus. The most obvious way this manifests itself is when you procrastinate: instead of doing the thing you’re supposed to, you work on something else or slack off. For some people, procrastination is the constant state of their lives, running away from one task to another until deadlines force them to focus and then having to struggle to get the job done on time. Other people struggle with more acute forms of procrastination that manifest themselves with particular kinds of tasks. I was more like this second kind of person, where there were certain types of activities I would spend all day procrastinating on. Though I have no problems writing essays for my blog, when I had to do research for this book, I dragged my feet. Similarly, I had no problem sitting and watching the videos of MIT classes, but I always tackled the first problem sets with considerable trepidation. Had it not been for the intense schedule I was on, I might have found excuses to avoid doing so for much longer. In fact, writing this chapter was one of the tasks I procrastinated on a great deal.
