Reviewing As Thinking

Why memorize? So you can think. But memorizing _is_ thinking. Instead of rushing through my flashcards, what if I can learn to _enjoy_ reviews, as _thinking_?

I love efficiency. I hate waste. Most people do.

For years, I’ve been trying to figure out the most efficient way to memorize large amounts of information. When I discovered reviewing flashcards using spaced repetition, I thought I’d found my prize. This flashcard system worked incredibly well … at first.

Then, slowly, it began to fail. Reviews became onerous. I wanted to get them over with as fast as possible.

Others felt the same way. When I told people I’d memorized Mark, they wanted to know how long it took. The implicit question was obvious: how much annoying review work will this superpower cost?

We took it for granted that the reviews were undesirable. A maintenance tax. We wanted to remember, not have to review.

But why?

What exactly is the difference between reviewing and remembering?

The Chore and the Prize Are the Same

Why memorize? So you can think.

But memorizing is thinking.

The more I consider the question, the harder it is to separate the process of “memorizing” from the goal of “thinking”. The chore and the prize seem nearly identical.

This is a very odd situation.

Unlike, Say, Cookies

Compare this to cookies. People like to eat cookies. But to get those cookies, you need to bake them.

Baking cookies is a completely different process from eating them. Sure, some people may also enjoy baking. But for most of us, baking is something we do as fast, as efficiently, as possible. We don’t want to spend time baking. We want to get it over with, so we can eat the cookies.

This pattern dominates our lives. Bake the cookie, then eat the cookie. Work at the job you hate, then collect the paycheck. Do the chore, then get the prize.

Now, it’s true, this pattern has problems. Life is too precious for dead time. Maybe we should learn to enjoy baking. Maybe we should learn to quit horrible jobs.

But that point is irrelevant for memorizing. Why do you review? So that you can remember. But reviewing is remembering.

Reviewing isn’t like baking at all. Reviewing is like eating the cookie.

Is Review Practice?

Sort of. After I mulled over the cookie analogy, I realized that there’s a whole other class of skills I hadn’t considered. Skills like riding a bike, or running a marathon, or playing the piano. Skills you practice.

On the one hand, practicing the piano is a lot like playing the piano.

On the other hand, real piano practice is a grueling, lurching grind. The only path to the effortless dance of real playing lies through the marshes of painful repetition.

Is the answer that simple? Are reviews difficult the way that practicing the piano is difficult? Is missing a card like missing a note?

Tempting. But no.

Sure, a commitment to reviews has the difficulties inherent in any commitment. Some days, you won’t feel like it.

But review is not simply practice. You practice to acquire a new skill. Practice is hard, because you can only acquire the skill a bit at a time. You’re constantly conscious of the gap between how good you are, and how good you want to be.

But when you review, you’re trying to connect to something you can already do. You’re trying to think a thought that you’ve already had.

This is fundamentally different from almost every kind of practice I can think of. You’re trying to recapture a previous experience.

Thinking Takes Practice

And yet, that original thought is often hurried and vague. You hear a poem, and before it’s over, you’ve forgotten the first line.

When we think of reviewing as practice, we try to remember the poem simply by repeating it. Just as you stumble through a sonata the first time through, you stumble through the poem until you’ve said it ten or a hundred times.

But what if this is wrong? What if we should be practicing reading, visualizing, and thinking about the poem? What if we should be slowly tasting the poem, line by line? What if we should do this five or ten or even more times, before we even consider memorizing it?

Thinking and visualizing definitely take practice. I have a lifetime habit of blazing through text, trying to finish. I want to get to the ending, the main point, the takeaway. For thriller novels, this might be fine, but a text worth remembering is dense. Scholars talk about “unpacking” a text. That takes practice. Focusing on each line, conjuring up images, sounds, scents, connecting these new thoughts to old nooks in your mental world — this takes time and effort.

I have to admit, I think I’ve been hoping that memorizing would make all that magic happen automatically.

It doesn’t. Memorizing may be impressive. But thinking is the hard part.

In fact, what if we found that this slow reading, imagining, and thinking led to memorization as a side effect? What if huge chunks of the poem began to stay with you, automatically, because you kept thinking about it? That’s what happens with song lyrics, sports stats, movie lines, and pretty much anything else people love to think about.

Even if we didn’t find ourselves getting perfect recall, and we had to keep glancing down at the written text, we would face a question. If we’re using text to think, and spaced repetition to make sure that we keep thinking about important things … do we really need to memorize?

Isn’t thinking the point?

Review Time as Thinking Time?

Suddenly, the world turns inside out, and I finally see memorizing the way most people seem to see it. A waste of time. Why memorize the Bible, when you can pull up any chapter on your phone?

Up to now, my answer has been that you can only think with what you remember. But, that’s not precisely true. You’re reading this right now, and you’re thinking about it. The truth is, you can only think with what you take time to think about.

And when I rush through my flashcard reviews, I’m not thinking. I’m focusing on the wrong question: “Do I remember this perfectly?”

Instead, whether I happen to have perfect recall or not, I should focus on: “This is my time to think about this again.

Imagine your reviews being slow and leisurely. Imagine looking forward to a hundred or so cards. Imagine thinking, “Whatever comes up, it’s something important. Every article and TV show is a gamble, but this stuff is solid. I enjoyed it the first time around, and I’m going to savor it now.”

Major paradigm shift.

Efficient Memorization Is a Slippery Slope

When I try to review as efficiently as possible, it’s like trying to eat cookies as efficiently as possible. It’s absurd. I’m hurtling down a mountain slope of diminishing returns. The more I focus on my own recall, on my grade, the less time, energy, and desire I have to think. I’m wolfing it all down, instead of tasting the knowledge.

For some skills, efficiency makes sense. You want to learn vocabulary as quickly as possible, so you can speak in the new language.

But for remembering what you read, or remembering texts, efficiency is a slippery slope.

The word itself sums up the problem. “Efficiency” really has two very different shades of meaning. We tend to want to be “efficient” when a task should be done as quickly as possible. The most efficient solution is to figure out how to avoid the task altogether. In this sense, the most “efficient” reviews would be no reviews at all: we could just download information to some mental computer.

But there’s a richer meaning to the word “efficient”. A process is efficient when it accomplishes the desired goal. We speak of a method’s “efficacy” — it really works.

In remembering, the first meaning of efficiency easily wars against the second. When we become obsessed with spending as little effort and time as possible on remembering, we forget why we’re remembering in the first place. Once that shift happens, reviews become absurd. Reviews are thinking. If you don’t like thinking for its own sake, there’s no reason to do it at all.

I’m a Human, Not a Computer

I mentioned a “mental computer,” and that analogy may be exactly what’s tripping me up. Computers can’t think. But they can copy terabytes of data in a few minutes.

If I view Anki as copying data into my brain, it will drive me nuts that it takes so many hours to copy so few bits of data. It makes me feel like a shoddy, vacuum-tube prototype mainframe that keeps chewing up its punchcards.

But if I view Anki as triggering thoughts, awakening whole networks of experiences and associations, then remembering will make me more human. I won’t review for the sake of awesome, high-quality thinking later. My reviews will be that high-quality thinking. Right now.

That’s exciting. I’d better go clean up my Anki decks.

17 Replies to “Reviewing As Thinking”

  1. This is absolutely brilliant!
    This is absolutely brilliant! I can’t believe no one else has commented. I’m really enjoying browsing your writings. Thanks for taking the time to decode your thoughts.

    1. Amazing
      i have been searching the internet for a i’m bored of Anki thread, and this has given me all motivation needed to carry on. Those deep thoughts really work and I am looking forward to try this out.

      1. Thanks, Matthew and Peter! I
        Thanks, Matthew and Peter! I’m so glad you find this helpful.

        Peter, please keep me posted on how a “thinking” approach to Anki works for you. I’m not using Anki myself right now, but instead trying to make time to think and write about what I read. I would love to hear your experiences.

  2. Reviewing As Practice
    I think you took a wrong turn when you said that reviewing isn’t practice. It actually is, or rather it’s a large component of practice. When you practice, you repeat stuff you know (e.g., playing a song on the piano or doing scales) so that you can strengthen the necessary neural connections to make them faster and less error-prone (e.g., missing fewer notes). You spend SOME time learning new skills, but you spend most of your time doing things you already pretty much know how to do. Where reviewing/Anki differs is that it is exclusively reviewing what you know. It’s purely a rention tool, not a learning tool.

    Another way SSRS is limited compared to most “practice” activities is that it focuses almost exclusively on whether something is recalled, and not so much on how quickly it was recalled. However, recall speed is essentialy in many instances. Being a master pianist would only take maybe a month if it didn’t matter how long it took you to recall to which key to press for a given note. But, it takes years of practice until the recall become so instantaneous that you can play through the most complex of songs with relative ease. The same problem arises when using SSRS with vocabulary: you’re going to have problems with speaking and listening if your recall time is slow, even if you know thousands of words.

    But, I think, in your analysis, you are right about many things. For starters, Anki and every SSRS system I’ve seen tends to isolate knowledge into small, unconnected chunks. This makes learning less efficient because its those connections that make knowledge useful. I use Anki to help retain vocabulary, but I can’t use that knowledge very well unless I practice using those words in a sentence. Also, SSRS programs tend to focus on whether something is remember or not, but not on how quickly that knowledge is remembered.

    I think your ideas about using flashcards to queue thoughts about certain things is a step in the right direciton. When you’re thinking about something, your exercising the connections to that knowledge. But, in that case, there’s no real test of whether you’ve forgotten something or not, and so it knowledge or nuances you wanted to remember could easily be lost, or the algorithm may not space new reviews properly.

    I think an even better solution would be for the program to generate a quesiton or problem that forces you to exercise the knowledge and its connections in order to answer or solve. This way, the recalled in its context. For example, what if, when I needed to review the Korean word for drink, instead of directly asking me what the Korean word for “drink” is, Anki gave me a sentence in Korean where I have to fill in the missing word, which is the word for drink. This way, I not only exercise the connection between drink and its Korean word 먹다, but I also exercise knowledge of its placement in a sentence, other korean words related to drinking, conjugation and other grammar rules, etc., all at once. Plus, the computer can measure my response and determine if a recalled the requisite knowledge correctly, and, if not, reschedule my future reviews accordingly.

  3. Ditto to Matthew – Another thanks!
    I really like how he said it: “Thanks for taking the time to decode your thoughts.” I know that with your latest post you’re saying all you need to do to memorize anything is to make connections, yet I’m almost ALWAYS interested in about everything, but I can’t seem to make connection. Visual mnemonics take that effort out. I’m hoping that eventually I’ll be able to learn the names to body parts and all that Greek/Latin/Roman-based stuff by learning at least the basics of those languages, and then those connections will be able to be made. I’m setting apart a month of this summer for Latin and the IPA (seems like a helpful skill; thanks for making me re-aware of it or whatever the word is, haha; I’d started ignoring all those comments on Wikipedia and in dictionaries).

    I’d like to see more posts from you in the future, unless of course you feel like you’ve already said everything that needs to be known about memory in your many posts. I’d at least like to know how all your language-learning is going. ‘Course I’m not super far into the blog, so I’ll probably find out in a bit, haha. And it seemed like you taught a memory class, or at least at some point you did. What was that like?

  4. Thank you!
    This post is great! I’m using anki to memorize a mountain of anatomy for med school and while right now I feel it’s working, I’m often frustrated that I simply know the facts out of context without really applying them to anything helpful. I’m definitely going to try and shift my paradigm about memorizing to one of thinking about more than just the fact. I also look forward to doing this for memorizing Bible passages. Something that was easy as a child but had gotten progressively harder over time. Thank you!

  5. You reached this point because you didn’t understand flashcards
    Hey Bill,

    First of all, I ask your pardon if you ever gave any updates on this. It’s the first post I read from your blog, but I felt the urge to comment.

    (I actually read the previous one, too.)

    I think you reached this point because you failed to understand how flashcards work, not flashcards failed you. I’m not talking about how you FEEL about Anki, as you described in this post, but how you organize your decks in it and probably your review settings.

    You don’t wanna be efficient when using Anki. You wanna be effective. This post gives a clear and brief explanation of my point:

    This will ease your stress significantly, as there’s clear diminishing returns as you approach 95% recalling. Plus, I do NOT know what my recalling % and nor do I care to know. I just treat my brain as an index for all the info that I might need to know (for certain types of decks). You hit it in the nail when you said that “why would one ever try to memorize passages of the bible when you could, you know, just read it again?”

    I’m majoring in engineering, but I do not try to break down EVERY bit of knowledge into flashcards because there are books out there. And you need context to remember things from books.

    For instance, there’s a software that every engineer has to master, called MatLab (a giant calculator, if you will), with thousands and thousands of commands. I could make flashcards of all commands that MatLab has and I would fail on mastering MatLab. Primarily, because it would be such a f**ing waste of time try to memorize commands that I might never ever use. It also would be meaningless remember a certain command if I couldn’t possibly know how to identify a possible situation where a certain command would be useful. I’d be better if I forced myself to go after commands of common problems I might face WHEN actively solving them, so I can build reference points.

    To use Anki EFFECTIVELY (not efficiently), your cards must meet the following criteria:

    – Must be high-yield knowledge. This often means <20% of a subject you are trying to learn are flashcard-eligible.
    - They must be well-tagged and organized, so you can ditch or focus on special bits of your deck. This is specially useful for exams where I won't need that information after the exam.
    - Not all subjects are flashcard-eligible. You need to be aware of that. Memorizing chapters of a book by making flashcards is one of the least effective ways of remembering the content of those chapters as a whole. Trying to recall the content while giving little explanations as if you're giving a lecture is a far more effective way to do it. And I always get something different every time I read the same book. Anki only tells me when I'm about to forget that, and if I wanted to retain that knowledge, I'd better review it decently.
    - Your cards must act as indexes. Your brain should be an index. Every time that I begin to make too much mistakes on certain flashcards, it's CLEAR that you're not remembering because you DIDN'T learn it correctly in the first place, not because anything else. So you mark that card, go back and review your book, lecture, notes, resource and make a couple more flashcards of that tiny bit that you're getting wrong every time.

    I use Anki to tell me that (a) what I didn't learn correctly in the first place, and/or correct the mental model around that thing and (b) create and index of vaguely remote chapter titles and their content description that are high-yield instead of memorizing the tiny bits of knowledge it might contain. That way, I can always go back to chapter X, page Y and read it again, make notes again, and then make cards about that tiny bit if I judge that it would be good. Otherwise I just shrug. That way I'm aware what content on book Z, chapter X and page Y might help me with something one day. If I get that wrong and don't need that knowledge at the time, I couldn't care less.

    I memorized the entire 20 Rules of Formulating Knowledge, and the only thing that I personally remember and WANTED to remember are the points above. So it was a huge waste of time memorize the entire article, because I could recall all the 20 rules, but my flashcards didn't get any better. Only after I started asking myself how I could make better flashcards in the moment I was making them and then read, got an idea and MADE the flashcards, they stated getting better. If I ever ask myself "what should I read if I want ideas for making better flashcards?" I'd skim that article first. But memorize it? Hell no.

    I find that the only things that are worth making flashcards are, then:

    - Things that you need to remember to pass an exam
    - Things that you need to be ready to remember fast at ANY given moment, such as a foreign language or difficult words on your mother tongue.
    - Still related to the previous topic, something that you could think logically to obtain, but it would be more effective to have it ready on your mind, such as 27² = 729
    - Pretty much everything that does NOT require the knowledge to be connected or contextualized, such as remembering numbers, foreign words, meaning of words, names, authors, dates, constants and pages of books that I could find something that I might consider important.

    That by itself will provide enough context for you to make the necessary connections and remember something. I often just memorize writers' birth dates so I can make a literature timeline in my head and place writer X correctly in that timeline. And that's all I need to be "Oh, he was born in 1900, so he must have wrote about this and that and was influenced by this writer. His most famous piece was this, in which he talked about this. I can read this, this and that resource if I want to know more about him. Good."

    Also, recalling is one of the most effective methods for remembering something even if it means you get more wrong answers during practicing and find something harder to recall:

  6. Cleared my doubts
    I was worried about whether ankidroid is really helping me or not. But now that you have cleared all my doubts, I now really gonna think it all over again like you while using it. Thanks for this perfect explanation!!

  7. Dang this is good
    This post summarizes Anki better than I have seen before. In all the dicussions on Anki, you have asked and answered the real question “What am I really doing here with Anki?” It was exciting to read your post.

  8. Bill Powell is a Genius

    “Paradigm shift” is overused to the point of cliche, but I can’t think of a better descriptor for the change of mind this creates. Thank you for sharing your thoughts on SRS.

    So often the analogy of brain as computer creeps into academic and business prose. We’re not computers. We’re infinitely more capable and emboldened by the Creator with synapses that make the universe look small if measured by the brains networking capacity.

    Why should we default to believing we’re somehow inefficient because we can’t do what a “glorified” copy machine does? I hate hypothetical questions, but I’m ranting now. So thank you again for this brilliant post.

  9. Change the cards!
    I am using Anki for years now, sometimes with gaps of months. I don’t care much for the statistics that Anki generates, I just review whenever I feel like it. I find Anki is a big help in the beginning when learning a language, but later on it is of limited use, you won’t memorize more words by just thumping through a bigger vocabulary. We learn most of our advanced vocabulary from context, not by memorizing them from paper.

    So for language learning I think it would be best to first memorize the first 1000 or so words from frequency lists, and memorize some basic grammar, and then start changing the cards into sentences. That way you are actually playing around with the language, making memorization more easy.

    I try to make the easy cards more interesting, and the difficult cards more comprehensible. For instance, I create a new card from a difficult adjective and an easy noun. Now the adjective has context. Or, if words become too easy to guess I incorporate them in a sentence with a new word (blanked out on the question side) or grammatical challenge. I try to find strange or interesting sentences on the net and add pictures, almost like creating little stories. It does take a lot of time, but you read and learn a lot while searching for compact sentences, and your decks become less boring.

  10. Thank you, mate. This is
    Thank you, mate. This is really the motivational text I need to keep going.
    I felt so much connected to some paragraphs that I thought I was the only writing this article.

  11. This. Is. Amazing.
    You have completely revolutionized the way I think about Anki. I identify exactly with your feelings. Thank you so much for contributing this solid article, a rare find.

  12. Why memorize the Bible, when

    Why memorize the Bible, when you can pull up any chapter on your phone?

    You might not need to memorize the whole Bible, but you do need to remember, at least vaguely, what each section said if you want that information to be useful to you or if you want to be able to correlate that information to some other information.

    Taken from

    Associative memory underlies the power of the human mind. Hypertext references are a poor substitute for associative memory. Two facts stored in human memory can instantly be put together and bring a new idea to life. The same facts stored on the Internet will remain useless until they are pieces together inside a creative mind. A mind rich in knowledge, can produce rich associations upon encountering new information. An empty mind is as useful as a toddler given the power of the Internet in search of a solution. Biological neural networks work in such a way, that knowledge is retained in memory only if it is refreshed/reviewed. Learning and repetition are therefore still vital for the progress of mankind.

  13. Yes but…
    Yes but, sometimes it is your job to remember. I am a historian, it is my job to remember several thousand words from different dead languages and scripts, so that I can find the information I look for, and sometimes translate them. Your words are motivating maybe for people who are working on their own pace to achieve a goal that has been defined by themselves. However sometimes your memory, how intimate it maybe for a regular person, has to work for a goal that has not been defined by the standards of one’s expectations from a certain field of interest, but by the expectations of a certain field of science which, one would like to believe, have higher goals. So how dehumanizing it maybe the process on the scientist, one needs to continue or at least struggle to continue to achieve those expectations.
    I didn’t write the rules, and I am not in the position to change them, but this is the case. So I would have preferred to have a response, saying don’t review more than 50 cards a day, or learn at most 30 new words for a day. However, I am obliged to thank you for addressing the problem.

  14. A how-to?
    I’m fascinated by the article, but confused as to how one might apply what you said. Are you planning a how-to? A few examples would be great too though.

Comments are closed.