Let me start with something that has genuinely bothered me for about ten years.

In the early 1970s, two researchers named William Chase and Herbert Simon ran an experiment that has since been repeated in more or less every domain anyone has bothered to test. They showed chess players a board position for about five seconds, then took it away and asked them to rebuild it from memory.

The masters were extraordinary. Twenty-odd pieces, five seconds, near-perfect reconstruction. The beginners managed a handful.

Then the researchers did the interesting thing. They scrambled the pieces. Same number of pieces, same board, but arranged at random, in configurations that could never occur in an actual game.

The masters' advantage almost completely disappeared.

Take a second with that before I go on. Because the obvious explanation, the one that arrives first and feels satisfying, is wrong. The obvious explanation is that chess masters have better memories. They don't. Tested on random material, they're unremarkable. Some of them lose their keys. Regularly.

So what is going on? And more to the point for you and me: what does it tell us about the foundations underneath anything we've ever tried to learn?

That's where we're going today. The psychology of foundational learning. What the floor actually is, why it decides the ceiling, and how to talk about all of it in English that's precise rather than merely enthusiastic.

The unit problem

Here's the answer to the chess puzzle.

The master isn't storing twenty-four pieces. He's storing perhaps five or six familiar structures. A castled king with its pawn shield is one thing to him, not four things. A particular pawn chain is one thing. An opening he has seen ten thousand times is one thing.

Psychologists call this chunking. Your working memory holds a famously small number of items at once, somewhere around four to seven depending on who you ask and how they counted. That limit doesn't move. You cannot train it. What you can change is what counts as one item.

And this is the first genuinely important idea of the episode. Learning a foundation is not filling a container. It's repackaging the contents so that more fits inside a container that never grew.

You already have this. When you read the word foundation, you didn't process ten letters. You processed one word, probably one concept, and you did it in under two hundred milliseconds without noticing. There was a year of your life when that word would have been ten separate problems. That year is invisible to you now, which is precisely the difficulty with foundations. Once they work, they become weather. You stop seeing them.

Build it with me: the word "foundational" itself

Let's do our first build, and let's start with the word we're using in the title, because it's a genuine trap.

Suppose you want to say this. Flat draft:

"Chess masters have good basic skills."

Nothing is broken here. A native speaker would understand it. But it's doing about thirty percent of the work it could be doing, and the reason is one word.

Basic has a problem. In English it points in two directions at once. It can mean fundamental, load-bearing, at the base, and it can mean simple, unsophisticated, not much to look at. When you say someone has good basic skills, a listener genuinely doesn't know whether you're praising the depth of their foundation or gently suggesting they're a beginner. Since about 2015, basic has also picked up a slang sense meaning bland and unoriginal, which does you no favors at all.

So let's separate the family. Say these to yourself and feel the difference:

Basic: either fundamental or unsophisticated, and ambiguous between them. Safe in technical phrases like basic research, risky as a compliment.

Fundamental: at the root, essential, cannot be removed without the whole thing failing. A fundamental misunderstanding. Fundamental rights. Strong, slightly formal, and it carries no suggestion of being simple. A fundamental idea can be extremely difficult.

Foundational: the thing that is built upon. It implies a structure above it. This is why it's the right word for our topic: it points forward in time, at everything that will rest on it.

Elementary: belonging to the first elements, and in modern use it has a whiff of the schoolroom. "Elementary mistakes" means mistakes a beginner makes. Use with care about adults.

Rudimentary: undeveloped, minimal, only just enough. "A rudimentary grasp of statistics" is not a compliment. This is the word you want when you mean present but barely.

Primitive: avoid it for people and cultures entirely. For systems and prototypes it's fine: a primitive version of the software.

Now rebuild with me:

"Chess masters aren't working with better memories. They're working with better-organized foundations."

Better. But I want to push once more, because there's a sharper version available and I want you to see the mechanism:

"Expertise doesn't enlarge the container. It repackages the contents."

I'll admit something. I wrote that line five different ways and I'm still not entirely sure repackages is right. It might be slightly too commercial, slightly too much like shipping. Reorganizes is safer and duller. Recompiles is more precise and more obscure. This is a real choice with real costs, and I don't think there's a clean winner.

Say all three aloud. Which one survives in your mouth?

Error repair while we're here. Three nouns in this territory refuse to take a plural, and this is one of the most persistent C1 leaks I encounter:

Knowledge is uncountable. Not knowledges. If you need to count, say areas of knowledge or bodies of knowledge.

Research is uncountable in standard academic English. Not researches. Say research, or studies, or a body of research, or a study.

Evidence is uncountable. Not evidences. Say evidence, or a piece of evidence, or lines of evidence.

And a preposition set that native speakers use without thinking and learners get wrong constantly:

You are an expert in a field. An expert in cognitive psychology. You are an expert at an activity. An expert at reading board positions. You are an expert on a topic. An expert on the Chase and Simon studies.

Field takes in. Skill takes at. Subject takes on. Say those three sentences out loud once. Muscle, not memory.

The knowledge paradox

Now the finding that reorganized how I think about learning.

In 1988, two researchers, Donna Recht and Lauren Leslie, took a group of middle school students and sorted them two ways: by reading ability, and by how much they knew about baseball. Then they gave everyone a passage describing a half-inning of a baseball game and tested comprehension.

The weak readers who knew baseball outperformed the strong readers who didn't.

Read that again slowly. General reading skill lost to specific prior knowledge. On a reading test.

The finding has held up remarkably well and been replicated with soccer, with music, with all sorts of content. What it says is that comprehension is not a free-floating skill you carry from text to text like a toolbox. Comprehension is mostly the collision between what's on the page and what's already in your head. A text is not a complete message. It's a set of instructions for assembling a message out of materials the reader is assumed to already own.

Which produces the paradox at the heart of foundational learning: you need knowledge in order to acquire knowledge. The more you know about a domain, the faster you learn more about it, because you have more places to attach new material. And if you have nothing to attach to, new information doesn't stick. It slides off and you call yourself a slow learner.

You are not a slow learner. You are a learner with nowhere to put things.

I find this genuinely consoling, and I'd like you to feel that too. Every time you've read something in a new field and felt stupid, that feeling was diagnostic, not evaluative. It was telling you something true about the state of your schema, not something true about your mind.

Build it with me: describing what knowledge does

Flat draft:

"If you already know about a topic, you will understand texts about it better."

Perfectly serviceable. Completely forgettable. Let's find out why.

The problem is that know about and understand better are both doing vague work. Neither one tells me how. Upgrade one:

"Prior knowledge does most of the quiet work of comprehension."

Now we have a metaphor doing something. Knowledge as labor, comprehension as a job that gets done. And quiet earns its place because it captures the invisibility we talked about earlier.

Verb options if you want to keep pushing. Feel the differences:

Carries: Prior knowledge carries the reader. Simple, physical, warm. 

Does the heavy lifting: idiomatic, slightly informal, very natural in speech. 

Underwrites: Prior knowledge underwrites comprehension. Financial metaphor, means guarantees and stands behind. Excellent in professional register, possibly too cold elsewhere. 

Props up: implies the thing would otherwise fall. Faintly negative. 

Supplies the missing half: my favorite here, because it's literally accurate. The text really is only half the message.

Here's my final build:

"A text isn't a complete message. It's half a message, and prior knowledge supplies the other half. This is why a weak reader who knows baseball will beat a strong reader who doesn't."

Pause here. Genuinely. Build your own version of that idea about a field you know well and a field you don't. Don't use baseball. Use something from your life.

Take your time. I'll wait.

Error repair. These three verbs get their prepositions swapped constantly, and the swap is very audible:

Depend on, never depend of or depend from. Comprehension depends on prior knowledge.

Consist of, never consist from. Expertise consists of thousands of stored patterns.

Result in and result from are opposites and both are correct English, which is exactly why they're dangerous. Weak foundations result in slow reading. Cause first, then result in, then effect. Slow reading results from weak foundations. Effect first, then results from, then cause. Get this backwards and you've reversed your entire argument while sounding confident.

False synonyms. Information, knowledge, and expertise are not three words for one thing:

Information is out there, in books and databases, unattached to anyone. 

Knowledge is information that has been integrated into a person. It's in a head. 

Expertise is knowledge organized well enough to be used quickly under pressure.

The chess master has expertise. The book about chess contains information. The gap between them is the entire subject of this episode.

The lie your fluency tells you

Now the part that costs people years.

There's a phenomenon in the research literature called the fluency illusion, and once you know about it you'll see it everywhere, starting with yourself.

Here's how it works. You reread a chapter. The second time through, everything flows. Nothing surprises you. The sentences arrive smoothly and you feel a warm sense of mastery. You close the book and think: I know this.

Two weeks later you can't reproduce any of it.

What happened is that you mistook the ease of processing for the strength of storage. Rereading is fluent because the material is fresh and your eyes have been down this road recently. That fluency is a signal about the last five minutes, not about next month. And your brain, unhelpfully, treats it as evidence of learning.

Robert Bjork, who has spent a career on this, coined the phrase desirable difficulties for the opposite approach. Conditions that make learning feel harder and slower in the moment, but produce better retention later. Testing yourself instead of reviewing. Spacing sessions out instead of massing them together. Interleaving different problem types instead of doing twenty of the same kind in a row.

Every one of those feels worse. Every one of them works better. And the ones that feel best are, with almost comic reliability, the ones that work least.

Which means your sense of how well a study session is going is not just unreliable. It's inversely correlated with reality across a meaningful range. That's an unusual and slightly upsetting thing to have to know about yourself.

I'll be honest about my own relationship with this. I know all of it, I've known it for years, and I still catch myself rereading, because rereading feels like being good at something and retrieval practice feels like failing repeatedly in a quiet room. Knowing the research does not make the feeling go away. It just gives you something to argue with.

Build it with me: hedging without going limp

This section needs a different English skill, and it's one that separates C1 from C2 more than vocabulary does. It's calibrated hedging: saying exactly how confident you are, no more and no less.

Flat draft:

"Rereading doesn't work."

Too strong. Rereading does something, just less than you think, and an informed listener will immediately stop trusting you. But watch what happens if you overcorrect:

"Some research seems to possibly suggest that rereading might not always be as effective as some people may think in certain circumstances."

Now you've said nothing, at length. This is the classic overcorrection and it makes you sound like you're hiding.

The skill is finding the middle. Here's a ladder, weakest claim to strongest:

It's possible that rereading is less effective than it feels. 

There's some evidence that rereading is less effective than it feels. 

Rereading tends to produce confidence without competence. 

The evidence consistently suggests that rereading is a poor use of study time. 

Rereading reliably produces the feeling of knowing and little of the substance. 

Rereading doesn't work.

All six are grammatically fine. Only two or three are honest about this particular body of research. My build:

"Rereading pays you in the feeling of knowing and charges you nothing for it. That's precisely the problem. Anything free is being subsidized somewhere."

Pause. Which rung of that ladder do you usually stand on? I'd guess most learners live on rungs one and two, hedging harder than the evidence requires, because overclaiming feels rude and underclaiming feels safe. It isn't safe. It's just quiet.

American spelling note, since we're here and it's directly relevant. In American English, practice is both the noun and the verb. I practice every day. My practice is inconsistent. The spelling practise is British and marks your English as non-American immediately. Same family: it's toward, not towards; learned, not learnt; canceled with one L; behavior, color, analyze with a Z.

Collocations that matter here. You make progress, you don't do progress. You put in the hours. You lay the groundwork or lay a foundation. You build on a foundation. You shore up weak foundations, a lovely construction metaphor meaning to reinforce something that's sagging. And you get the hang of something, which is the most natural informal phrase in English for reaching basic competence.

Automaticity: what foundations actually buy you

Let's talk about why this matters practically, because so far it might sound like an argument for grinding.

The concept is cognitive load. Your working memory, that four-to-seven item space, is the entire workshop where thinking happens. Everything you're consciously processing right now is competing for the same small table.

When a skill is not automatic, it occupies the table. When you first learned to drive, steering took the whole table. Mirrors, gear, pedals, the terrifying knowledge that other cars existed. You could not hold a conversation. You certainly could not think about anything else. Now, assuming you drive, you arrive at your destination with no memory of the journey and a fully formed opinion about something you heard on the radio.

The driving didn't get easier. It got automatic, which moved it off the table.

Here's the implication for foundational learning, and it's the part people miss. Automaticity isn't the opposite of thinking. It's the precondition for it. Every foundational skill you make automatic is table space returned to you, and complex thought is exactly the activity that needs a large table.

This is why a learner struggling with basic sentence construction cannot simultaneously produce sophisticated argument. Not because they lack ideas. Because assembling the sentence is consuming the space the idea needed to live in. The ideas are there. There's nowhere to put them down.

And it's why the advice to "just focus on communication, don't worry about the mechanics" is half right in a way that becomes wrong. Early on, yes, absolutely. But mechanics that never become automatic don't stay neutral. They tax every sentence you ever build, forever, and the tax compounds.

Build it with me: metaphor and the nominalization trap

We've been using a metaphor. Let's examine it, because choosing metaphors well is a genuine C2 skill and choosing them badly is invisible damage.

Attention has been described as a container (space, room, capacity), a budget (spend, afford, costly), bandwidth (a channel, throughput), a spotlight (focus, illuminate, peripheral), and a desk or workbench (clutter, clear, put down).

They're not interchangeable. Container metaphors suggest a fixed limit and nothing else. Budget metaphors add the crucial idea of trade-offs, which is exactly what cognitive load theory is about. Bandwidth is contemporary and slightly cold. Spotlight captures selectivity but not limitation. Desk captures both, plus mess.

For this topic I'd argue budget or desk. Both let you say the important thing: that space spent here is unavailable there.

Now the trap. Watch this sentence:

"The automatization of foundational skills results in the liberation of attentional resources for the performance of higher-order tasks."

Every word is correct. It's unreadable. The problem is nominalization: turning verbs into nouns. Automatize became automatization. Liberate became liberation. Perform became performance. The actions have been embalmed into abstract nouns, and now the only real verb in the sentence is results in, which is doing all the work for four buried actions.

Verb-driven rebuild:

"When the basics run themselves, they stop competing for attention, and attention is what hard thinking requires."

I want to be fair to nominalization, though, because the standard advice against it is too crude. Nominalizing lets you make an action into a subject you can then say things about. "The automatization of basic skills is what separates a fluent reader from a decoding one." You can't easily say that with verbs. Nominalization is a tool for abstraction. The failure is using four in one sentence.

Prepositions. Capacity takes for: capacity for abstract reasoning. Attention takes to: attention to detail. A load falls on something: the load on working memory. You make space for something.

Your turn. Pause and build one sentence explaining automaticity using the budget metaphor, and one using the desk metaphor. Then decide which one you'd actually say to a friend. That's the real test. Not which is more elegant, but which survives being spoken.

Compounding, in both directions

In 1986, Keith Stanovich borrowed a phrase from the Gospel of Matthew to describe something troubling in reading research: the rich get richer and the poor get poorer.

The mechanism is brutally simple. A child who reads slightly better finds reading slightly more pleasant, therefore reads slightly more, therefore gets better, therefore reads more. A child who reads slightly worse finds it unpleasant, avoids it, falls further behind, finds it more unpleasant.

Small initial differences don't stay small. They're inputs to a feedback loop. By high school, the gap in words encountered in different contexts between avid and reluctant readers runs into the millions, and vocabulary, general knowledge, and comprehension have all diverged accordingly.

Now, please don't hear this as a sentence handed down. That's not the finding and it's not my point. The loop can be interrupted at any age, and it runs in your favor the moment you feed it. What I want you to take is the shape of the thing. Foundational learning compounds. Which means two things at once, and they're both true: early advantages grow, and every hour you spend on foundations right now is not a linear investment. It's an input to a loop.

Present tense, your life. If you're a C1 learner deciding whether to work on the precision of your prepositions or to read another article, this is the actual argument for the prepositions. Not that they're virtuous. That they're multiplicative. Fixed prepositions make every future sentence cheaper to build, and you're going to build a great many sentences.

And here's where this collides with right now, in 2026, in a way it never did before.

You can currently outsource almost any single step of thinking. Translation, summary, first drafts, explanation, code. That is a genuine and enormous gift, and I use these tools constantly and I'm not going to pretend otherwise. But notice what you cannot outsource: the schema that lets you evaluate what comes back. Judging whether an answer is good requires the same prior knowledge that would have let you produce it. Outsource the foundation and you don't get a shortcut. You get a very fast process whose output you cannot assess.

The chess master can use an engine brilliantly. A beginner with the same engine learns almost nothing, because he can't tell which of its suggestions matter.

Build it with me: conditionals and the language of compounding

This section calls for a grammar structure that C1 learners often know and rarely deploy: the mixed conditional.

Flat draft:

"If I studied grammar earlier, my English would be better."

There's a real error here, and it's extremely common. If you mean a past condition, you need the past perfect:

"If I had studied grammar earlier, my English would be better now."

That's a mixed conditional: past condition, present result. And it is exactly the structure this topic demands, because compounding is precisely about past conditions producing present states. Compare:

"If I had studied grammar earlier, my English would have been better." (Third conditional. Past condition, past result. Says nothing about today.)

"If I had studied grammar earlier, my English would be better now." (Mixed. The consequence is sitting in the room with you.)

The difference is not decorative. One puts the loss in the past. The other says the loss is ongoing. Feel that.

Verbs of compounding, since the ordinary ones wear out fast:

Compound: Small advantages compound. Precise, financial, and exactly right. 

Accumulate: neutral, gradual. 

Widen: for gaps specifically. The gap widens. 

Diverge: for two paths separating. Their trajectories diverged. 

Snowball: informal, implies acceleration and slight loss of control. 

Entrench: for disadvantage becoming structural. Early gaps entrench.

A caution. The phrase bridge the gap is so overused that it has stopped meaning anything. If you need it, prefer close the gap, or catch up, or better, say what specifically happens.

My build:

"If I'd fixed my prepositions at B2, I wouldn't be paying for them at C1. That's not regret. It's compound interest, and the account is still open."

Support has an expiration date

One last piece, and it's the one that surprises people who love scaffolding.

Vygotsky gave us the zone of proximal development: the band between what you can do alone and what you can do with help. Learning happens in that band, and scaffolding is the temporary support that keeps you inside it. Wonderful, well supported, and the basis of most good teaching.

But there's a companion finding called the expertise reversal effect, and it's less comfortable. Instructional support that measurably helps novices measurably harms experts. Worked examples, detailed explanations, step-by-step guidance: give them to someone who already has the schema and their performance drops.

Why? Because processing the explanation now costs more than generating the answer would have. The scaffold has become clutter on the desk. The learner is spending attention reconciling your explanation with the structure they already have, and that reconciliation is pure waste.

Which produces a rule that's easy to say and hard to live by: all support is designed to be removed, and the removal is not optional. A scaffold left standing after the building is finished isn't kindness. It's an obstruction.

Apply it to yourself. Somewhere in your English there's a crutch that helped you enormously at B1 and is now slowing you down. Translating in your head. A phrasebook of safe expressions you reach for so you never have to gamble. A grammar rule you learned in a simplified form that was true enough then and is getting in your way now.

Take a moment. Which one is yours?

I'll go first, in the spirit of the thing. Mine is the habit of hedging every claim, which served me well when I knew less and now sometimes makes me sound less certain than I actually am. I know I do it. I don't always catch it in time.

Build it with me: talking about your own level without cringing

This is a register problem, and it's the last build. Describing your own ability in English is a minefield with bragging on one side and self-flagellation on the other.

Flat draft:

"My English is not so good."

Almost certainly false if you've followed this episode, and it does you damage. But the overcorrection is worse:

"I have an excellent command of English."

True or not, it lands badly in almost every context except a résumé.

The move is specificity. Vague self-assessment forces you into praise or apology. Specific self-assessment lets you just describe:

"I can hold my own in a meeting, but I lose the thread when three people talk at once."

"I read comfortably. I write slowly, and I revise more than I'd like."

"I'm past the point where translating helps me. Now it's just friction."

Notice these aren't modest or immodest. They're informative. That's the register you want, and it's the register of everything else we've discussed: describing the actual state of the foundation rather than grading yourself on it.

Phrasal verbs for this territory, all high-frequency and all worth owning:

Brush up on: refresh something you once knew. I need to brush up on my conditionals. 

Get the hang of: reach working competence. 

Pick up: acquire informally, without study. I picked up most of my vocabulary from reading. 

Fall back on: use a reliable resource under pressure. When I panic, I fall back on simple sentences. 

Keep up with: maintain pace. I can keep up with a podcast but not a dinner party. 

Get by: manage at a minimum level. I get by in French. 

Outgrow: not phrasal, but exactly right here. I've outgrown that strategy.

Your Build

Here's your scaffold. Three prompts. Build them in writing if you can, out loud if you can't, and notice that all three require the structures we built today.

One. Identify one foundation in your English that is not yet automatic and is taxing every sentence you write. Describe it in two sentences, using the budget or desk metaphor for attention, and using specific self-assessment rather than praise or apology. Something like: "Article usage still costs me attention. Every time I write a noun phrase I'm spending something that the argument needed."

Two. Write one mixed conditional about your own learning history. Past condition, present consequence. Use had plus past participle in the first half and would plus bare infinitive in the second. Then read it back and check that you haven't written a third conditional by accident, because you probably have. Everyone does the first time.

Three. Name one scaffold you've outgrown, and one you haven't. Use outgrow for the first and fall back on for the second. Then answer the harder question honestly: what is the actual cost of keeping the outgrown one, and what would it take to put it down this month?

And an optional fourth, if you want the difficult one. Take the sentence "Foundations matter" and rebuild it three times: once for a five-year-old, once for a colleague, once for someone who thinks foundations are a waste of time and wants to skip to the interesting part. Same idea, three registers. That third one is genuinely hard, because you have to make the argument without sounding like you're scolding, and I'm not sure I managed it in this episode.

Where this leaves us

The chess master's memory is ordinary. His floor is not.

That's the whole thing, really. Nothing about foundational learning is glamorous, and the psychology explains why: foundations do their best work invisibly, and the better they work the less you notice them. The floor never gets credit for the height of the ceiling.

But the finding I'd most like you to carry is the consoling one, not the demanding one. Every time you've felt slow in a new field, that feeling was information about your schema and not a verdict on your mind. There was nowhere to put things. Build somewhere to put things, and the same brain that felt slow will feel fast, and you will have changed nothing about yourself except what counts as one item.

Whatever you're building on top of, go check the floor this week. Find one thing down there that still costs you attention. Make it automatic, and take the attention back.

You'll need it. You're going to build something tall.