Memorization Matters

“Memorization is a useless skill—I can always just Google it” is a common sentiment in the modern era. I remember defiantly asserting this as a student whenever I was forced to memorize anything, and heard it not infrequently as a teacher. LLMs have only made the critique fiercer.

But memorization does matter—even with the collective knowledge of all humanity at our fingertips. As miraculously fast as our technology is, it is still nowhere near the speed of normal conversation, let alone the speed of thought. (You wouldn’t be reading this right now if you hadn’t at some point memorized the letters of the English alphabet and the meanings of many English words. Ditto for just about any other skill or area you care about.) Everything becomes much clearer when you consider that to have something “memorized” is, practically speaking, simply to know it well enough to retrieve it with ease.1

I think it’s obvious that knowing stuff is useful; from here, it follows naturally that it’s valuable to deliberately commit more knowledge to memory. Nevertheless, for the sake of argument, I’ll cover each of the following in greater detail:

• What “memorization” does and doesn’t mean

• Some evidence that knowing stuff is useful

• What (and how) you should memorize

For many of us, the word “memorization” brings to mind the tedium of flashcards and spelling quizzes. In reality, these examples are just particular tests of memorization, which is itself a much broader and more flexible construct. We memorize song lyrics and dance steps. We memorize the faces of our friends and families. We memorize the fastest route to get from home to work. You might not think of these things as “memorization” because they are enjoyable or happen automatically, but they are: You remember them.

Let’s also distinguish memorization from other cognitive processes. Memorization is not general reasoning, or pattern-matching, or the ability to solve novel problems. Memorization not working memory or short-term memory. (Despite having “memory” in their names, these are brain systems; I’m talking about a process.) Memorization is simply the act by which you come to consciously know stuff—nothing more, nothing less.

Psychologists have a distinction between the two buckets of cognitive abilities I’m referring to: crystallized vs. fluid intelligence. Crystallized intelligence is the ability to use knowledge, facts, and experience that one has accumulated over time—in other words, the application of what’s in our long-term memory. Fluid intelligence is cognitive ability independent of any knowledge from the past—in other words, everything I described in the previous paragraph. I’ll have more to say on crystallized vs. fluid intelligence in the following section. For now, just note that there is a direct line from memorization (process) to crystallized intelligence (ability), whereas the line from memorization to fluid intelligence is tenuous at best.

Note also that almost everything we care about involves a mix of both kinds of intelligence. To play a great guitar solo, you need to know scales and technique (crystallized intelligence) and to create a novel combination of notes (fluid intelligence). Different domains require different balances of each, but the idea of “pure fluid intelligence” without any crystallized knowledge is on-its-face nonsensical.2

So clearly crystallized intelligence is at least somewhat important, and can presumably be augmented through memorization. The question is: Just how much does crystallized intelligence matter relative to fluid intelligence?

When you look at the research on human performance throughout the life cycle, some consistent patterns emerge:

• On certain types of tasks—for example, repeating back a list of twelve words, or identifying visual patterns—people do best in their twenties and thirties.

• On other types of tasks—for example, tests of vocabulary and general knowledge—people do best in their fifties and sixties.

• Peak earnings tend to be in people’s forties and fifties, and peak performance skews later many fields.3

Of course these are averages, and there is a huge mess of factors to contend with. Still, these facts are suggestive of two conclusions. First, fluid intelligence peaks much earlier than crystallized intelligence:

Second, of the two, crystallized intelligence matters more for most people in their careers. If it were the case that adaptability and raw problem-solving ability (i.e., fluid intelligence) were worth more in the labor market than knowledge and experience (i.e., crystallized intelligence), we’d expect a different pattern: Peak earnings/performance well before forty, followed by a long dropoff toward retirement. Because peak outcomes are, on average, in the back half of one’s career, crystallized intelligence must account for the greater proportion of the pattern that we see.

Of course, a few caveats are in order. Some dispute the premise that fluid intelligence does peak so much earlier.4 (This would undercut my claim that memorization matters by offering another factor to explain midlife success.) Others have argued for a more holistic explanation of why midlife is the sweet spot: The peak is a result not only of crystallized intelligence but also of an increase in non-cognitive traits like conscientiousness, emotional stability, and empathy.5 If true, these caveats weaken the story that basic knowledge accumulation explains human performance.

Still, I find it difficult to resist the conclusion that crystallized intelligence plays some significant role in success. I work a normal-ish office job, of the kind that involves sending emails, maintaining records, and managing projects. I’m much more efficient and effective at this job than when I started it just a couple years ago; I think this has almost everything to do with crystallized intelligence and very little to do with fluid intelligence. In my first few months on the job, for example, I was constantly interrupting my workflow to ask colleagues questions or double check internal process documentation or look up what a certain acronym stood for. Through repeated exposure, I stopped needing to do this, which facilitated both faster work and freed up brain capacity to tackle more abstract problems.

I suspect most domains work approximately like this. People probably can become more dynamic and effective in a generalizable way—but development is more often bottlenecked by how much you know about the particular domain you’re working in. The bigger and more relevant your mental library, the better you’re able to retrieve/apply/combine information in useful ways. Flashy conceptual breakthroughs often require years of study and iteration on a problem.6

Share

Is it possible that crystallized intelligence matters but memorization does not? Could it be that the best we can do is bumble around accumulating useful knowledge/experience through random trial and error? I think this is a self-evidently silly questions. If you’ve ever deliberately learned anything, ever, I hope you’ll agree: It surely is possible to intentionally commit useful information to memory through repeated practice. This naturally begs the question…

Not all knowledge is created equal. Some information is helpful for Jeopardy! or crossword puzzles but has little value in your daily life. Some information is fairly important but comes up so rarely that it’s not worth memorizing—like the address of the DMV, or the instructions for assembling your bed frame.

To make matters worse, the world is changing faster than ever, which means it is difficult to know what information will continue to remain relevant. There was a time not so long ago when the phone numbers of all your friends and family were useful facts; now you’re a savant if you can rattle off three of them. While it’s incorrect to say that the internet (and now LLMs) have rendered memorization obsolete, it’s certainly true that they’ve changed the types of information that are valuable to know.7

All that said, there’s a sweet spot of knowledge that is both reliably useful and comes up rarely enough organically that it’s worth intentionally adding to your mental library. This sweet spot will of course be highly particular to different individuals. (If you’re a plausibly out-of-touch politician, you should memorize the price of milk; for the rest of us, it only matters if you’re looking for a great deal on dairy.) So if I had to give a generalized answer to the question of “What should you memorize?” I’d just suggest the following as a starting point:

• Useful stats and figures, which can be helpful as reference points for media consumption and general reasoning (e.g., the population of major cities, the number of neurons in a human brain, etc.)

• Favorite quotes, which can help with motivation, decision-making, and general conversation. (I of course have a soft spot for these.)

• Trigger-action plans of the form “If X situation arises, I will______.” (TAPs help with habit formation and behavior change by reducing the conscious thought required to undertake said action.)

These are just a few examples but I hope give a sense of the breadth of facts and behaviors I’m referring to when I say “memorization.”

Thankfully, the answer to the second question—“How should you memorize?” —has a much simpler, one-size-fits-all answer: spaced repetition.8