Cognitive load: working within your limits#
Try this: read these letters once, then look away and say them back—R J Q M B K P. Now try RED JAM BOX. Same nine letters in the second case, but you almost certainly found it far easier. Nothing changed about your memory between those two attempts. What changed was the load you placed on it. Your working memory—the small mental workspace where you hold and juggle whatever you’re thinking about right now—is remarkably limited. Understanding just how limited, and learning to work within those limits, is one of the highest-leverage things you can do as a learner.
Your working memory is tiny#
For decades people quoted the “magical number seven” (you’ll see this on the chunking page too). More careful work since then has shrunk that figure. When researchers strip away the tricks people use to cheat—rehearsing under their breath, grouping items—the true capacity of working memory for separate, unrelated items is closer to three or four chunks (Cowan, 2010). That’s it. Four.
This sounds almost insulting until you notice how often it bites you. You read a dense sentence with five clauses and the beginning has evaporated by the time you reach the end. You hold a phone number in your head, someone asks you a question, and the number is gone. You’re following a recipe and lose your place the moment the doorbell rings. None of these are failures of intelligence. They’re the predictable result of asking a four-slot workspace to hold five things.
The good news: while capacity is fixed, what you choose to put in those slots is not. A “chunk” can be a single letter or a whole meaningful unit—RED occupies one slot, not three. That single fact is the seed of nearly every memory technique in this guide.
Three kinds of load#
The researchers Sweller, van Merrienboer and Paas (2019) describe the demand on working memory as coming from three sources. Telling them apart changes how you study.
Intrinsic load is the difficulty baked into the material itself—how many elements you must hold at the same time to understand it. Adding two numbers is low intrinsic load. Learning to read music while watching your hands and keeping time is high. You can’t wish this away, but you can sequence it: learn the parts before the whole, master simple cases before complex ones.
Extraneous load is the wasted effort—demand created by how the material is presented rather than by the material itself. A cluttered diagram, a definition that uses three words you don’t know yet, a worked example split across two pages so you keep flipping back and forth—all of this burns slots that should be doing the actual learning. Extraneous load is pure waste, and it’s the one you have the most power to cut.
Germane load is the effort that actually builds learning—the work of noticing patterns and assembling them into lasting mental structures. This is the load you want. The whole game is to minimise the extraneous so you have capacity left over to spend on the germane.
When you’re stuck and overwhelmed, the useful question isn’t “am I clever enough?” It’s “which of these three is eating my capacity, and which can I cut?”
Chunks and schemas free your capacity#
Here’s the escape route from a four-slot workspace, and it’s a beautiful one. Through practice you build schemas—organised packages of knowledge stored in long-term memory—and a schema, however large, behaves as a single chunk when you pull it into working memory (Sweller, van Merrienboer & Paas, 2019).
Think about how a beginner sees a chessboard versus a master. The beginner sees thirty-two separate pieces—hopelessly over capacity. The master sees four or five familiar formations. Same board, but the master has compressed it into a handful of chunks, freeing working memory to think several moves ahead. The expert hasn’t grown extra slots. They’ve learned to pack each slot with vastly more meaning.
This is why “just memorise it” and “really understand it” aren’t opposites. Every schema you build—a multiplication table, the shape of a chemical reaction, the structure of a good argument—permanently shrinks the load of everything that uses it. Early learning feels heavy precisely because you have no schemas yet and must hold every element separately. It gets lighter not because the material changes but because you change.
Putting it to work#
A few moves follow directly, and they cost you nothing:
- Cut the clutter. Before you study, strip away the noise—close the extra tabs, find the clearest explanation rather than the first one, and don’t make yourself hold two halves of an example in your head at once. Every bit of extraneous load you remove is capacity returned to learning.
- Lean on worked examples early. When a topic is new, studying a fully worked-out solution costs far less working memory than grinding through the problem yourself, because you’re not also juggling the search for what to do next. Solve them yourself after the pattern is familiar—once a schema is forming and can carry the load for you.
- Build schemas on purpose. Spaced repetition , self-testing, and the association techniques in this guide aren’t just about cramming facts in—they’re about packing knowledge into chunks so it stops crowding your workspace.
- Sequence from simple to whole. Don’t fight high intrinsic load head-on. Learn the parts in isolation, then combine them. The combined task only becomes manageable once each part has become a single chunk.
The takeaway: you can’t enlarge your working memory, so stop trying. Instead, protect it. Cut the waste, learn from clear examples, and keep building schemas—because every schema you own is a slot you get back.