by Scott Young
Entering from a door on the right, a heavyset middle-aged woman with dark brown hair and olive skin approaches the stage. Everett goes up to her and says something in a language she doesn’t understand. She looks around, clearly confused, and then replies hesitantly, “Kuti paoka djalou.”1 He tries to repeat what she has just said. There’s some stumbling at first, but after one or two more tries, she seems satisfied with his repetition of her reply. He goes to the blackboard and writes, “Kuti paoka djalou ➱ ‘Greeting (?).’” Next he picks up a small stick and points to it. She guesses correctly that he wants to know the name and replies, “ŋkindo.” Once again, Everett goes to the blackboard and writes, “ŋkindo ➱ stick.” Next he tries two sticks and gets the same response, “ŋkindo.” He then drops the stick, to which the woman says, “ŋkindo paula.” The demonstration proceeds, with Everett picking up objects, performing actions, listening to the woman’s responses, and recording the results on the blackboard. Soon he’s moved past simple naming tasks and starts asking for more complicated sentences: “She drinks the water,” “You eat the banana,” and “Put the rock in the container.” With each new elicitation, he experiments, building new sentences and testing her reaction to see if he is correct. Within half an hour, there are more than two blackboards full of nouns, verbs, pronouns, and phonetic annotations.
Learning dozens of words and phrases in a new language is a good start for the first thirty minutes spent with any language. What makes this feat particularly impressive is that Everett isn’t allowed to speak any language he might have in common with the speaker.2 He can only try to encourage her to say words and phrases and repeat them to try to figure out the language’s grammar, pronunciation, and vocabulary. He doesn’t even know what language is being spoken.*
How can Everett start speaking a new language from scratch, without teachers or translations or even knowing what language he’s learning, in half an hour, when most of us struggle to do the same after years of high school Spanish classes? What enables Everett to pick up vocabulary, decode grammar, and pronunciation so much faster than you or I, even with all those additional constraints? Is he a linguistic genius, or is there something else going on?
The answer is our first principle of ultralearning: metalearning.
What Is Metalearning?
The prefix meta comes from the Greek term μετά, meaning “beyond.” It typically signifies when something is “about” itself or deals with a higher layer of abstraction. In this case metalearning means learning about learning. Here’s an example: If you’re learning Chinese characters, you will learn that 火 means “fire.” That’s regular learning. You may also learn that Chinese characters are often organized by something called radicals, which indicate what kind of thing the character describes. The character 灶, for example which means “stove,” has a 火 on the left-hand side to indicate that it has some relationship to fire. Learning this property of Chinese characters is metalearning—not learning about the object of your inquiry itself, in this case words and phrases, but learning about how knowledge is structured and acquired within this subject; in other words, learning how to learn it.
In Everett’s case, we can see glimpses of the enormous wealth of metalearning that lies just beneath the surface. “Well, what are some of the things we noticed about this?” Everett asks the audience after his brief demonstration has concluded, “It seems to be SVO, a subject-verb-object language, that’s not terribly shocking.” He continues, “There doesn’t seem to be any plural marking on the nouns, unless it’s tones and I missed it. . . . There’s clearly pitch going on here; whether it’s tone remains to be analyzed.” From this jargon we can see that when Everett evokes a word or phrase from his interlocutor, he isn’t just parroting back the sounds; he’s drawing a map with theories and hypotheses about how the language works grounded on years of experience learning languages.
In addition to his enormous wealth of knowledge as a linguist, Everett has another trick that gives him an enormous advantage. The demonstration he has presented is not his own invention. Called a “monolingual fieldwork” demonstration, this method was first developed by Everett’s teacher Kenneth Pike as a means of learning indigenous languages. The method lays out a sequence of objects and actions that the practitioner can use to start piecing together the language. This method even received some Hollywood exposure after Louise Banks, a fictional linguist, used it to decode an alien language in the 2016 science fiction movie Arrival.
These two pieces in Everett’s linguistic arsenal—a richly detailed map of how languages work and a method that provides a path to fluency—have allowed Everett to accomplish a lot more than just learning some simple sentences. Over the last thirty years he has become one of only a handful of outsiders to become fluent in Pirahã, one of the most unusual and difficult languages on the planet, spoken only by a remote tribe in the Amazon jungle.3
The Power of Your Metalearning Map
Everett’s case beautifully illustrates the power of using metalearning to learn new things faster and more effectively. Being able to see how a subject works, what kinds of skills and information must be mastered, and what methods are available to do so more effectively is at the heart of success of all ultralearning projects. Metalearning thus forms the map, showing you how to get to your destination without getting lost.
To see why metalearning is so important, consider one study on the helpful effects of already knowing a second language when learning a third.4 The study took place in Texas, where monolingual English speakers and bilingual Spanish/English speakers were enrolled in a French class. Follow-up on subsequent tests showed that the bilingual speakers outperformed the monolingual students when learning a new language. On its own, this isn’t terribly surprising. French and Spanish are both Romance languages, so there are shared features of grammar and vocabulary that aren’t present in English that could conceivably provide an advantage. More interesting, however, is that even among the Spanish/English bilinguals, those who also took Spanish classes ended up doing better when they later needed to learn French. The reason seems to be that taking classes assists with helping form what the study authors call metalinguistic awareness in a way that simply knowing a language informally does not. The difference between the two types of bilingual speakers mostly came down to metalearning: one group had content knowledge of the language, but the group that took classes also had knowledge about how information in a language is structured.*
Nor is this idea about metalearning restricted to languages. Linguistic examples are often easier to study because there’s a cleaner separation of metalearning and regular learning. This is because the contents of unrelated languages, such as vocabulary and grammar, are often quite different, even if the metalearning structure is the same. Learning French vocabulary won’t help you much with learning Chinese vocabulary, but understanding how vocabulary acquisition works in French will likely also help with learning Chinese. By the time my friend and I had reached the last country in our year of learning languages, the process of immersing ourselves and learning a new language from scratch was practically a routine. The words and grammar of Korean may have been completely new, but the process of learning was already well trodden. Metalearning exists in all subjects, but it can often be harder to examine independently from regular learning.
How to Draw Your Map
Now that you have some idea what metalearning is and its importance to learning faster, how can you apply this to get an edge in your own learning efforts? There are two main ways: over the short term and over the long term.
Over the short term, you can do research to focus on improving your metalearning before and during a learning project. Ultralearning, owing to its intensity and self-directed nature, has the opportunity for a lot higher variance than normal schooling efforts do. A good ultralearning project, with excellent materials and an awareness of what needs to be learned, has the potential to be completed faster than formal schooling. Language learni
ng through intensive immersion can beat lengthy classes. Aggressively paced coding bootcamps can get participants up to a level where they can compete for jobs much faster than those with a normal undergraduate degree. This is because you can tailor your project to your exact needs and abilities, avoiding the one-size-fits-all approach taken in school. However, there’s also a danger of choosing unwisely and ending up much worse off. Metalearning research avoids this problem and helps you seek out points where you might even be able to get a significant advantage over the status quo.
Over the long term, the more ultralearning projects you do, the larger your set of general metalearning skills will be. You’ll know what your capacity is for learning, how you can best schedule your time and manage your motivation, and you’ll have well-tested strategies for dealing with common problems. As you learn more things, you’ll acquire more and more confidence, which will allow you to enjoy the process of learning more with less frustration.
In this chapter, I’m going to devote most of the next section to short-term research strategies, since they will probably benefit you the most. However, this emphasis shouldn’t undermine the importance of the long-term effects of metalearning. Ultralearning is a skill, just like riding a bicycle. The more practice you get with it, the more skills and knowledge you’ll pick up for how to do it well. This long-term advantage likely outweighs the short-term benefits and is what’s easiest to mistake for intelligence or talent when seen in others. My hope is that as you get more practice in ultralearning, you’ll start to automatically apply many of those skills to learn faster and more effectively.
Determining Why, What, and How
I find it useful to break down metalearning research that you do for a specific project into three questions: “Why?,” “What?,” and “How?” “Why?” refers to understanding your motivation to learn. If you know exactly why you want to learn a skill or subject, you can save a lot of time by focusing your project on exactly what matters most to you. “What?” refers to the knowledge and abilities you’ll need to acquire in order to be successful. Breaking things down into concepts, facts, and procedures can enable you to map out what obstacles you’ll face and how best to overcome them. “How?” refers to the resources, environment, and methods you’ll use when learning. Making careful choices here can make a big difference in your overall effectiveness.
With these three questions in mind, let’s take a look at each of them and how you can draw your map.
Answering “Why?”
The first question to try to answer is why are you learning and what that implies for how you should approach the project. Practically speaking, the projects you take on are going to have one of two broad motivations: instrumental and intrinsic.
Instrumental learning projects are those you’re learning with the purpose of achieving a different, nonlearning result. Consider the previously mentioned case of Diana Fehsenfeld, who, after a few decades as a librarian, found that her job was becoming obsolete. Computerized file systems and budget cuts meant she would need to learn new skills to stay relevant. She did some research and decided that the best way to do this would be to get a firmer grasp on statistics and data visualization. In this case, she wasn’t learning because of her deep love of statistics and data visualization but because she believed that doing so would benefit her career.
Intrinsic projects are those that you’re pursuing for their own sake. If you’ve always wanted to speak French, even though you’re not sure how you’ll use it yet, that’s an intrinsic project. Intrinsic doesn’t mean useless. Learning French might have benefits later when you decide to travel or need to work with a client from France at your job. The difference is that you’re learning the subject for its own sake, not as a means to some other outcome.
If you’re pursuing a project for mostly instrumental reasons, it’s often a good idea to do an additional step of research: determining whether learning the skill or topic in question will actually help you achieve your goal. I’ve often heard stories of people unhappy with their career progress who decide that attending graduate school is the answer. If only they had an MBA or an MA, employers would take them more seriously, they think, and they’d have the career they desire. So they go off to school for two years, rack up tens of thousands of dollars in student debt, and discover that their newly minted credentials don’t actually get them much better job opportunities than before. The fix here is to do your research first. Determine if learning a topic is likely to have the effect you want it to before you get started.5
Tactic: The Expert Interview Method
The main way you can do research of this kind is to talk to people who have already achieved what you want to achieve. Let’s say you want to become a successful architect and think that mastering design skills might be the best step to take. Before you get started, it would be a good idea to talk to some successful architects to get a sense of whether they think your project will actually help with your intended goal. Though this method can be used for many parts of the research process, I’ve found it particularly valuable for vetting instrumental projects. If someone who has already accomplished the goal you want to achieve doesn’t think your learning project will help reach it or thinks it’s less important than mastering some other skill, that’s a good sign that your motivation and the project are misaligned.
Finding such people isn’t as hard as it sounds. If your goal is career related, look for people who have the career you want and send them an email. You can find them at your current workplace, conferences, or seminars, or even on social networking websites such as Twitter or LinkedIn. If your goal is related to something else, you can search online in forums dedicated to the subject you want to learn. If you want to learn programming, with the goal of building your own apps, for example, you can find online forums dedicated to programming or app development. Then you just need to look for frequent posters who seem to have the knowledge you’re looking for and email them.
Reaching out and setting up a meeting with an expert isn’t hard, either, but it’s a step many people shy away from. Many people, particularly the introverts among us, recoil at the idea of reaching out to a stranger to ask for advice. They worry that they’ll be rejected, ignored, or even yelled at for presuming to take up a person’s time. The truth is, however, that this rarely happens. Most experts are more than willing to offer advice and are flattered by the thought that someone wants to learn from their experience. The key is to write a simple, to-the-point email, explaining why you’re reaching out to them and asking if they could spare fifteen minutes to answer some simple questions. Make the email concise and nonthreatening. Don’t ask for more than fifteen minutes or for ongoing mentorship. Though some experts will be happy to help you in those ways, it’s not good form to ask for too much in the first email.
What if the person you want to interview doesn’t live in your city? In that case, phone or online calls are great alternatives.* Email can work in a pinch, too, but I’ve found that text often doesn’t translate tone well, and you often miss sensing the way the person feels about your project. Saying it’s a “great idea” lukewarmly versus enthusiastically can make a world of difference, yet that nuance is missing if you communicate via text only.
Even if your project is intrinsically motivated, asking “Why?” is still very useful. Most learning plans you might choose to emulate will be based on curriculum designers’ ideas of what is important for you to learn. If these aren’t perfectly lined up with your own goals, you may end up spending a lot of time learning things that aren’t important to you or underemphasizing the things that do matter. For these kinds of projects, it’s useful to ask yourself what you’re trying to learn because it will help you evaluate different study plans for their fit with your goals.
Answering “What?”
Once you’ve gotten a handle on why you’re learning, you can start looking at how the knowledge in your subject is structured. A good way to do this is to write down on a s
heet of paper three columns with the headings “Concepts,” “Facts,” and “Procedures.” Then brainstorm all the things you’ll need to learn. It doesn’t matter if the list is perfectly complete or accurate at this stage. You can always revise it later. Your goal here is to get a rough first pass. Once you start learning, you can adjust the list if you discover that your categories aren’t quite right.
Concepts
In the first column, write down anything that needs to be understood. Concepts are ideas that you need to understand in flexible ways in order for them to be useful. Math and physics, for example, are both subjects that lean heavily toward concepts. Some subjects straddle the concept/fact divide, such as law, which has legal principles that need to be understood, as well as details that need to be memorized. In general, if something needs to be understood, not just memorized, I put it into this column instead of the second column for facts.
Facts
In the second column, write down anything that needs to be memorized. Facts are anything that suffices if you can remember them at all. You don’t need to understand them too deeply, so long as you can recall them in the right situations. Languages, for instance, are full of facts about vocabulary, pronunciation, and, to a lesser extent, grammar. Even concept-heavy subjects usually have some facts. If you’re learning calculus, you will need to deeply understand how derivatives work, but it may be sufficient to memorize some trigonometric identities.
Procedures
In the third column, write down anything that needs to be practiced. Procedures are actions that need to be performed and may not involve much conscious thinking at all. Learning to ride a bicycle, for instance, is almost all procedural and involves essentially no facts or concepts. Many other skills are mostly procedural, while others may have a procedural component yet still have facts to memorize and concepts to understand. Learning new vocabulary in a language requires memorizing facts, but pronunciation requires practice and therefore belongs in this column.
