I did a PhD without knowing how to read

If you've watched this channel for any length of time, you know I usually talk about linguistics, phonology, syntax, the strange parts of how human language works, like how Persian speakers signal intimacy by letting more air out of their noses. Most of you probably didn't see that one, but it was good. Today's video is different. I want to talk about two books that completely changed how I think about learning and research. And I read both of them well after finishing my PhD. That timing bothers me. It embarrasses me. Keep thinking about how much easier grad school would have been, how much deeper my work could have gone if someone had handed me these books before I started. Or even better, before I started college. I was learning on hard mode, just using raw horsepower. And at 41, with a PhD, I only now feel like I really understand what I should have been doing all this time. Hilariously, my mother-in-law was absolutely terrified to offend me when she gave me one as a gift. You'll understand in a second. The two books

are Mortimer Adler's How to Read a Book, originally published in 1940, and Make It Stick by Peter Brown, Henry Roediger, and Roediger Henry Roediger and Mark McDaniel, which came out in 2014. The first sounds almost insulting. If you're reading the book on how to read a book, can't you already read a book? And the second I wrote off as self-help pablum for almost a decade because the name was so catchy and the cover so poppy. Adler was writing in the 1940s and later in the 1970s about reading as an intellectual practice and lamenting people's regurgitation of prepackaged slogans instead of thinking. If you want to see what that looks like today, read the comment sections on some of my videos. And the Make It Stick authors are cognitive scientists writing about memory and learning. But when you've read them together, these two, together in some sense, describe the same thing from different angles. Years ago, in syntax one, Tony Kroch accused the whole class of being illiterate. More than a decade later, I'm willing to concede that I didn't actually know how to read. Miss you, Tony, you curmudgeonly old crank. So, if you want to supercharge your learning, whether you're in a formal program or not, and if you want to hone your critical thinking, today's video is for you. Today, we're learning how to learn. This is Language Jones. Speaking of words, how's that for a segue? Monarch comes from Greek monos

alone, plus arcane, to rule, one ruler, which is a pretty good name for who I want to thank for sponsoring today's video, because it's a personal finance app that gives you one place to rule over all of your money. This is Monarch. Every account, checking, credit card, savings, investments, all syncing to one clean dashboard. It connects with over 13,000 institutions, so linking everything is genuinely fast. Instead of hopping between four banking apps, or five, or six, to figure out where I stand, it's just one view. What I'm most into is the net worth trend over time and the goals feature. You can set save up goals tied to your real accounts, or pay down goals for debt, and it projects a clear timeline. There's fixed versus flexible spending, so your budget isn't rigid, plus AI insights that explain changes in plain English. You can literally ask it questions about your money. It's private and ad-free, which I appreciate. You're the customer, not the product. It's not a bank, it doesn't move your money, it just gives you a clear view of it. Web and mobile, both. Link is on screen and in the description, promo code is Jones 50, or scan the QR. If you use my promo code Jones 50, you can get started today with 50% off. Okay, back to it. A friend of mine was just accepted to the very prestigious master's program at St. John's, colloquially known as the Great Books program, seminars in classics. Sounds amazing. And to celebrate it, I bought her these two books, because they are books that both radically changed how I view learning, that are eminently practical, and that I keep returning to as a reference. Let me start with Adler.

His central image is a pyramid, and each level of reading contains the levels below it. You could just as easily have nested sets, but what are you going to do? The first level he calls elementary reading. This is basic literacy. You can look at a sentence and tell me what it says. Most adults assume they finished with this level around third grade, and for most kinds of text that's probably true, although not for the last few generations who never had phonics and who were subjected to a failed experiment in reading education that kind of sort of is Chomsky's fault indirectly, but that's another video for another day. Don't worry. It's coming. The second level is inspectional reading, which is a systematic skimming with a time limit. I did the International Baccalaureate in high school, and we were taught this is called Harvarding a book. You're trying to figure out what kind of book you're actually holding, what its structure is, and what its surface argument seems to be. This is the level most academics actually live at, even though we rarely admit it to everybody else. You read the abstract, the introduction, the conclusion, maybe a few section headings, and you decide whether the book is worth more of your time. Same goes for journal articles. In linguistics, the next step, if it's worth your time, is to read the examples, the graphs, the diagrams, etc. The third level is analytical reading, and this is where Adler does much of his serious work. Analytical reading is the slow, effortful reading where you were genuinely trying to understand a book on its own terms. Adler describes it as chewing and digesting. It's effortful. It's in part where the inspectional reading really shows its true value. Before I used to dive right into a difficult text and attempt to read it all at an analytical level in order. It's insane. Getting the structure, the lay of the land, allows the analytical reading to unfold along a clear path. The fourth level is what he calls syntopical reading. I'm not going to lie, basically nobody calls it this if they're not citing Adler. This is reading multiple books on the same subject and putting them into conversation with each other, so you can construct your own argument across them. If you've written a literature review, you have attempted some kind of syntopical reading. Whether you did it well is another question. What's really cool about his approach is you can start to figure out what underlying assumptions and schools of thought are affecting an author even if they don't spell it out explicitly. And you can start to put authors who were never in conversation with each other into a sort of imagined conversation. I'm doing it right now. What makes Adler actually useful rather than suggestive is that he

gives you concrete rules for analytical reading. There's six that I want to mention. First, classify the book. Figure out what kind of book it is before you start because a theoretical monograph and a popular science book and a primary source ask different things from a reader. But this can also be classification in terms of broad intellectual tradition. In some cases, they won't even acknowledge that they're implicitly in conversation or disagreement with other traditions. So, reading some Chomsky and generative syntax work, you might never know that he's implicitly responding to criticisms from, say, construction grammar while seemingly overtly dismissing the existence of those criticisms at all. Second, x-ray the structure. This sounded super futuristic in the '40s. Try to state the unity of the whole book in a single sentence and then outline its major parts in order. If you can't do this, you haven't understood the book yet. Third, define the problem. What specific question is the author trying to answer? Books are answers to questions and if you don't know the question, the answer will feel like floating information. Surprisingly, they're often actually addressing different problems and answering different questions than they claim they are. Fourth, come to terms in Adler's terms. Find the keywords the author depends on and make sure that you understand them in the way that the author does. It sounds obvious, but in linguistics, think about how much trouble graduate student gets into when they assume that grammar or competence or discourse mean the same thing across two different theoretical frameworks. Coming to terms is harder than it looks. This is where his suggestion of skimming the index is really helpful. It turns out that the index at the back of the book is there for a reason. Huge if true. Fifth, find the propositions. Identify the most important sentences in the book, the ones where the author is making a specific claim. Sixth, construct the arguments. Link those propositions together so you can see how the author actually supports their conclusions, if they do. When you do all of this, you're no longer just reading, you're reconstructing the book inside your own head, and that act of reconstruction is the bridge to the second book. Make It Stick opens with

bad news for most of us. The study habits that feel productive, the ones we've been doing since high school, are largely useless for long-term retention. Actually, scratch that. They're worse than useless. They often impede real learning. Rereading a chapter feels like learning. Highlighting a textbook feels like learning. Going over your notes one more time before an exam feels like learning. The authors call these illusions of mastery. They produce a feeling of fluency because the material gets easier to recognize. But recognition is not the same as recall, and recall is what you actually need when you were trying to use information months or years later. I'll say it again. Recognition is not the same as recall. Recognizing something does not mean that you know it and can generate it, let alone synthesize it or put it in conversation with some other idea in a novel context. I want to chill on this for a moment to keep with Adler's penchant for culinary metaphors because I think it's part of the book that most people resist. I know I did. I built an entire study system around rereading and highlighting, and that system had carried me through a PhD program. To be fair, it wasn't traditional highlighting. It was actually a system of marking a book that's kind of similar to what Adler recommends, but without then making use of those markings to go back and synthesize the information. When I read that the system was largely an illusion, my first reaction was that the authors had to be overstating things. They're not. The cognitive science here is robust. There are mountains of it conducted over decades and once you start paying attention to your own learning, you can feel the illusion working on you in real time. Some of you might remember my live streams from a few years back when I tried the silly stunt of learning the basics of Persian in 3 months. I put a lot of time and a lot of effort and I felt like I was learning. I found out the hard way I was not. Now I'm revisiting it using tools from cognitive science and I know that I am in part because I can see the improvement and measure the improvement. So, what actually works? The book offers several techniques and I want to go through the seven that I have

found most useful. First, retrieval practice means closing the book and trying to recall what you just read from memory rather than looking back at your notes. The act of pulling information out of your head, even imperfectly, strengthens your ability to pull it out again later. One easy way to approach this in language learning is to one, literally just close the book and try to remember something specific you just learned and two, to try to immediately use it in a different frame. This can be as simple as seeing an example in one conjugation and trying to generate it in another. Second, spaced repetition means leaving time between practice sessions. The counterintuitive part is that you want to forget a little bit. I've said it before, I'll say it again, hey that's spaced repetition. Forgetting is an integral part of learning. When retrieval is effortful, the memory consolidates more deeply. Don't ask me why, it just does. Cramming feels efficient because you can hold a lot of material in working memory at once, but almost none of it will be available to you a week later. This is why I absolutely loathe those fluent in 3 months hucksters in the YouTube polyglot space. Assuming that you can view enough material once over the course of 3 months to say that you have in any way meaningfully engaged with the basics of a language, 3 months is still nowhere near enough time to mentally consolidate the material so that you know it in any meaningful sense after another 6 months. By the way, now is as good a time as any to mention that I am developing a digital course for language learning in particular that draws on linguistics, the science of learning, and cutting-edge developments in modern tech to accelerate language learning. If you're interested, you can sign up for updates and the wait list at languagejones. com/accelerator. Okay, third, interleaving means switching between related topics rather than drilling one topic for hours. If you're studying syntax, do not spend 3 hours on movement and then 3 hours on binding or whatever. In fact, don't do that at all. Mix them. The brain learns to distinguish concepts by being forced to choose between them. In fact, it's better to do 10 or 20 minutes on a syntactic concept and then another 10 or 20 on, say, phonology. Or, as was my case in grad school, complex numbers and linear algebra. Drop me a comment if you know why a linguist would need those, or if you don't know why and you want to know why. Fourth, elaboration means connecting new information to what you already know in your own words. If you can find a real-world metaphor or relate the new concept to something you understood years ago, you are building more retrieval paths to the same memory. By way of an example, I was talking to a friend about how these principles could be applied to basically anything. And I gave examples from pulling from the Wikipedia page on Dostoevsky. The broader context was having an LLM extract information for minimally formatted closed deletion cards in Anki after a first read, but that's neither here nor there. In talking about both how to read a book and make it stick, I talked about first formatting things super simply. More on that in a minute. So, for instance, taking the biographical information and breaking it into Dostoevsky was born in 1821 and Dostoevsky died in 1881 and Dostoevsky published Notes from the Underground in 1864. I might actually take a moment to think about what was happening in say 1864, too, not just in Russia. So, like, Lincoln was reelected then. And making that connection actually locked in the date for Notes from the Underground for me. Not that I need that information, but well, now I have it. Fifth generation means trying to answer a question or solve a problem before you've been shown the solution. Even when you fail, the failure creates a mental hook for the correct answer when it eventually arrives. I do this a lot with language study. I'll study my Anki cards before I've learned them. Sixth, reflection means taking a few minutes after a study session to ask what went well and what you would do differently. You've heard me talk about this in my videos on things like iTalki. Just reflecting after a session really helps solidify things that you've learned. This sounds soft, but it's doing really important cognitive work. You're revisiting the material and integrating it with your sense of your own performance. And finally, calibration means using objective tests like quizzes or flashcards to figure out what you actually know versus what you feel like you know. The gap between those two things is usually larger than students expect. Anki was a life-changer for me precisely because I got metrics on what was working and what wasn't. My previous method of look at a vocabulary list and try to remember it was just not remotely effective. Here's where the two books

start to talk to each other. Look, I'm doing syntopical reading. Look, Tony, I can read. Adler's Analytical Reading, the part where you classify the book and x-ray its structure and come to terms with its keywords and reconstruct its arguments is doing exactly what Make It Stick says learning requires. When you try to state the unity of a book in a single sentence, you're doing synthesis and retrieval practice. When you find the propositions and link them into arguments, you're doing a form of elaboration. When you come to terms with the author's vocabulary, you're doing the kind of effortful processing that builds durable memory. When you read syntopically across multiple books simultaneously, you're interleaving. Adler did not have access to the cognitive science. He was working from intuition and from the long tradition of liberal education. In fact, he even has a passage in the updated version of the book from the 1970s that states that we have no idea what is happening in the brain when a child learns to read. That's not true anymore. But, the techniques he prescribed turn out to be almost point for point the techniques that controlled experiments would later show to be effective in learning more broadly. The reason analytical reading works is the reason retrieval practice works. They are the same thing dressed in different clothes. This matters because most of us, when we read a difficult book, default to a passive mode. We let the sentences wash over us, pass through us, if you want to stay with the culinary metaphor in the worst, most colorful possible way. We feel like we're absorbing the material, and the feeling is convincing because the prose is becoming more familiar with each page or with each rereading. But, familiarity is the illusion. What we need is the harder work, the work of writing the book's outline from memory, as Adler suggests you do in the end pages, of identifying the flow of arguments in the margins, of stating its arguments in our own words, of testing whether we can actually articulate the question it's trying to answer. That work is uncomfortable, and the discomfort is a sign that learning is actually happening, not that all discomfort is a good sign. I want to be clear, it's not that I was doing none of this in grad school. I just didn't have a systematic approach for extracting, synthesizing, and then locking in information. I was doing it on hard mode. I want to say something concrete about how to actually do space repetition because the technique fails for most people who try it, and the failure's almost always about how the material is organized

rather than about the software or the schedule. The tool that I use is Anki, which is free except for their iOS app, which I paid $25 for, but any space repetition system will do. It could be a shoebox. What matters is what you put into it. Pause. Piotr Wozniak, whose name I'm almost undoubtedly saying incorrectly, designed the algorithm that most of these systems are based on, and he wrote a short essay on what he calls the minimum information principle, which is the most useful thing I've read about how to make flashcards, even if I think some of his approaches also bunk. The principle is that each card should test one small atomic piece of knowledge formulated as simply as possible. If you find yourself writing a card with a long answer or a card that tests three related facts at once or a card where the question could reasonably have several different correct answers, you need to break it apart into smaller cards. Cards that violate the minimum information principle are the cards that you will fail repeatedly without really understanding why. And they're the cards that drive people to abandon the system of using space repetition. Now, a caveat is in order since the authors of Make It Stick demonstrate that the minimum information piece can also be chunked or grouped information. So, to give a highly specific example only relevant to a vanishingly small number of you, a hierarchy of five kinds of grain I have to remember for reasons is easily locked in my mind as whibizzro. Wheat, barley, spelt, rye, oats. By the way, leave me a comment if you know why. The best teacher that I ever had in high school ensured that we all knew that a main driver of Russian 20th century foreign policy was a desire for warm water ports, warm water ports with free access to the sea. What I did not appreciate when I first started using Anki is that the act of reformulating material into atomic cards is itself a form of analytical reading. To turn a paragraph from a syntax textbook into 10 good flashcards, you have to figure out what the paragraph is actually claiming, which terms are doing the real work, and how the claims depend on each other. You're doing Adler's job of finding the propositions and constructing the arguments. Except now, you're doing it with the explicit goal of producing testable units, and then you do interleaved testing with space repetition. It's amazing. The reformulation is the synthesis. By the time the cards exist, you've already done most of the learning, and the space repetition is there to make sure that the learning sticks rather than do the learning for you. And once you have a few hundred or a few thousand of these atomic facts genuinely available to you, not just recognizable but retrievable on demand, something really interesting happens. You can start putting ideas into conversation with each other in the way that Adler describes as syntopical reading because the raw material is actually in your head rather than in a notebook that you would have to go consult somewhere. Connections that would have required an afternoon of cross-referencing become available in the middle of a thought. That, more than anything else, is what changed for me when I started taking this seriously. In order to understand, say, history better than the absolutely insane comments on my video on the history of Hebrew, you have to be able to read in the Adlerian sense, to read critically, and on a fundamental level, to remember the basic facts that might buttress or collapse an argument. You can't have that spoon-fed to you via steady stream of TikToks or from a taco gone bad. This makes even reading a short Wikipedia article into, shall we say, an engaging challenge. But simultaneously, you can decide what you want to lock in and then the information principle and all of the best practices around space repetition allow you to lock it in with minimal additional effort. So, if I could go back and hand

both of these books to a younger version of myself, I think the change to my graduate school experience would have been absolutely quite substantial. Not because I lacked intelligence or motivation, but because no one ever told me that the way that I was reading and studying were producing the feeling of learning without producing the substance of it. I was working hard at the wrong things. I want to be careful here because I really do not want to suggest that either book is magic. They're not. How to read a book is very often repetitive and Adler's tone can feel patrician in a way that has maybe not aged well, though I kind of like it. But your mileage may vary. Make It Stick is written in a style that prioritizes accessibility over depth, and there are moments where I kind of wish the authors would slow down and engage more carefully with the studies that they cite. Neither book is perfect, but together they describe a practice of serious reading and serious studying that I think every graduate student, and honestly every undergraduate and people who are just educating themselves should be exposed to early on. If you're starting a research degree or considering one or just trying to teach yourself something difficult, these are the two books that I'd put in your hands first. Links to both are below. If you've read either of them and have your own take, I would genuinely like to hear it in the comments. If you're interested in learning the basics of linguistics, by the way, I'm launching a self-paced online introduction to linguistics later this year. If you'd like to know more or get on the wait list for updates, you can do so at www. languagejones. com/blueprint. Some of you have messaged me directly about it. You actually have to write the www. Thanks for watching and I'll see you in the next one. Happy learning. —