چرا واژه‌هایی را که بلد بودید فراموش می‌کنید

— Intro and level note —

Ava: Welcome back to the LexUnio English Podcast. I'm Ava.

Andrew: And I'm Andrew. This is episode two, and a quick note before we start: this one is a step up. Last time we spoke a little slower and kept the language simpler, because the topic was aimed at learners around B1.

Ava: Today we're at B2 to C1. We're speaking at a more natural speed, the sentences are longer, and we're not going to stop and define every term. If that's too fast for you right now, that's genuinely fine — go back to episode one and come to this one in a couple of months.

Andrew: And if it's slightly too fast but you can still follow the argument, that's the perfect place to be. That's the seventy percent we talked about last time.

Ava: So. Today's subject is forgetting.

Andrew: Which is the thing nobody puts on the poster. Every course advertises how much you'll learn. None of them advertise how much you'll lose.

Ava: Let's start with the experience, because I think everybody listening has had it. You're reading something, you hit a word you don't know, you look it up. Maybe you write it down. And there's a small, satisfying feeling — right, got it, that's mine now.

Andrew: And then four days later you meet the same word and you have absolutely no idea.

Ava: Not a vague memory. Nothing. And often the worst part is you don't even recognise that you've met it before.

Andrew: I've had the specific humiliation of looking a word up, and finding it already highlighted in my own dictionary app. From two weeks earlier.

Ava: That's the one. And the natural conclusion people draw from that experience is that they have a bad memory. Which is almost always wrong, and it's a costly thing to believe about yourself.

Andrew: So let's take that apart. Because what's happening is completely predictable, it's been studied for over a century, and once you understand the mechanism you can work with it instead of against it.

— The forgetting curve —

Ava: Let's begin in the 1880s, which is where this story starts.

Andrew: A German psychologist called Hermann Ebbinghaus decided to study memory experimentally, which nobody had really done before. And he had a problem: if you test people on real words, you can't separate memory from what they already knew.

Ava: So he invented nonsense syllables. Meaningless three-letter combinations. And then he memorised thousands of them himself, and tested himself at intervals to see how much he'd retained.

Andrew: He was both the scientist and the experimental subject, which would not get approved today, but it gave us something genuinely useful: the forgetting curve.

Ava: And here's the shape of it. Forgetting is not gradual and it is not linear. It's brutal at the beginning and then it flattens.

Andrew: Which means the first day is where you lose most of it. Something you learned this morning, if you never encounter it again, is largely gone within a day or two.

Ava: I want to add a caveat here, because I think it matters. Ebbinghaus used nonsense syllables specifically because they had no meaning. Real vocabulary has meaning, and meaning makes things stickier. So real words decay more slowly than his curve suggests.

Andrew: But the shape holds. Fast at first, then slower.

Ava: The shape holds. And the shape is what matters, because it tells you exactly when your effort is worth the most.

Andrew: Which is early.

Ava: Which is early. If you meet a word once and never again, you have essentially wasted the effort of looking it up. Not partly wasted — almost entirely wasted.

Andrew: That's quite a strong claim.

Ava: It is, and I'll stand by it. Think about how many words you've looked up in your life and how many you actually own. The gap between those two numbers is the entire subject of this episode.

— Why cramming fails —

Andrew: Alright. So if a single exposure doesn't work, what does? And here's where most people's instinct is wrong, so let's go carefully.

Ava: The instinct is: if once wasn't enough, do it more. Repeat it ten times now. Write it out twenty times. Sit with the list until it's in there.

Andrew: Cramming. Or, the technical term, massed practice — all your repetitions massed together in one block of time.

Ava: And here's the uncomfortable finding, which has been replicated for decades: massed practice produces excellent short-term performance and very poor long-term retention.

Andrew: Which is exactly the wrong shape for language learning.

Ava: It's the wrong shape for anything you want to keep. Cramming works brilliantly for a test tomorrow. It works terribly for a language you want to speak in three years.

Andrew: And the reason it feels effective is worth understanding, because that feeling is the trap.

Ava: Yes. When you repeat something ten times in a row, by repetition eight it feels effortless. And your brain reads that fluency as evidence of learning. But what you're actually experiencing is the word sitting in short-term memory, where it was going to be anyway.

Andrew: The psychologist Robert Bjork has a phrase for this that I've never forgotten: the difference between performance and learning.

Ava: Say more about that, because it's the central idea of the whole field.

Andrew: Performance is how well you can do something right now, during practice. Learning is the durable change that shows up later. And the disturbing part of Bjork's work is that these two things can move in opposite directions.

Ava: Meaning conditions that make practice feel harder often produce better long-term learning.

Andrew: Exactly. He calls them desirable difficulties. Difficulties you should want. And the reverse is also true: study methods that feel smooth and satisfying are frequently the ones that leave nothing behind.

Ava: Which explains why rereading is so popular and so useless.

Andrew: Rereading is the perfect example. You read a page of vocabulary, then read it again, and the second time it flows. Familiar. Comfortable. And you interpret that comfort as knowledge.

Ava: But it's just familiarity. You've seen the shape of the words recently. That's not the same as being able to produce them.

Andrew: Right. So let's introduce the distinction that solves this, because I think it's the single most useful concept in the episode.

— Recognition vs recall —

Ava: Recognition versus recall.

Andrew: Recognition is: here is a word, do you know what it means? Recall is: here is a meaning, what's the word?

Ava: And these are wildly different in difficulty. Recognition is easy. The word is in front of you, doing half the work. Recall requires you to generate it out of nothing.

Andrew: And here's the practical consequence. Almost everything most learners do trains recognition. Reading trains recognition. Watching with subtitles trains recognition. Flipping through a vocabulary list trains recognition.

Ava: And speaking requires recall.

Andrew: Speaking requires recall, exclusively. Nobody hands you the word in a conversation. So you can build an enormous recognition vocabulary and still be unable to say anything, and then you conclude — wrongly — that you don't know enough words.

Ava: When actually you know plenty of words. They're just all in the wrong file.

Andrew: That's a good way to put it. And this leads directly to the second big idea, which is the one thing I'd want somebody to take from this hour if they took nothing else.

Ava: The testing effect.

— The testing effect —

Andrew: The testing effect. Also called retrieval practice. And the finding is this: the act of retrieving something from memory strengthens that memory more than re-studying it does.

Ava: So the test isn't the measurement. The test is the training.

Andrew: The test is the training. That sentence took me an embarrassingly long time to accept, because I grew up thinking of tests as the thing that happens after learning.

Ava: There's a well-known set of experiments on this by Roediger and Karpicke. The design is simple: one group studies material repeatedly, another group studies it once and then tests themselves repeatedly.

Andrew: And on a short-term test, the re-studying group does better.

Ava: Which is exactly the performance-versus-learning trap. But on a delayed test, a week later, the self-testing group is substantially ahead.

Andrew: And the most interesting detail, for me, is that the students themselves predicted the opposite. They thought re-reading had worked better, because it felt better.

Ava: So we're bad judges of our own learning. Which means you can't trust the feeling of a study session. You can only trust what you can produce later.

Andrew: Let's make this concrete before we move on. What does this change, practically, for somebody learning vocabulary tonight?

Ava: It changes the direction of the card. If you have a word and its translation, don't look at both and nod. Cover the answer. Try to produce it. Fail if you're going to fail. Then check.

Andrew: And the failing part is important.

Ava: The failing part is essential. A failed retrieval attempt followed by seeing the answer is more valuable than never having attempted it. The effort of searching is what does the work, even when the search comes back empty.

Andrew: Which is deeply counterintuitive. It feels like you wasted the attempt.

Ava: It feels like failure and it's actually the mechanism. That's the pattern for this entire subject: the thing that feels bad is usually the thing that's working.

Andrew: Alright. So we have the forgetting curve, which tells us memory decays fast at first. And we have retrieval practice, which tells us that trying to remember is what strengthens memory. Put them together and you get the third idea.

— Why spacing works —

Ava: Spacing. The spacing effect.

Andrew: Which says: the same number of repetitions, spread out over time, produces far more durable learning than the same repetitions bunched together.

Ava: And this one is not a small effect. It's large, it's been replicated for over a century, and it works across almost everything people have tested it on.

Andrew: So five repetitions across five days beats five repetitions in one evening.

Ava: Comfortably. And it's not even close. Same effort, same number of repetitions, radically different outcome. Which makes it, honestly, the closest thing to free money in learning.

Andrew: Why does spacing work? Because I think understanding the why makes people actually do it.

Ava: There are a few competing explanations and I don't think we should pretend it's settled. But the intuition I find most useful is this: retrieval only strengthens memory when it's effortful.

Andrew: And if you repeat something thirty seconds later, there's no effort. It's still sitting there.

Ava: It's still in your hand. You're not retrieving anything, you're just looking at it again. Whereas if you wait until you're on the edge of forgetting, retrieval takes real effort — and that effort is exactly what makes the memory stronger.

Andrew: So the ideal moment to review something is just before you'd forget it.

Ava: That's the principle. And now the obvious problem: how on earth do you know when that is?

Andrew: For one word, you might guess. For two thousand words, each with its own schedule, you have no chance.

Ava: And that's the entire reason spaced repetition software exists. It's not a fancy flashcard app. It's a scheduler for a calculation that a human cannot do by hand.

Andrew: Let's talk about how that scheduling actually works, because I think people use these tools without knowing what they're doing, and then they use them badly.

— How the scheduling works —

Ava: The simplest version is the Leitner system, and it's worth understanding because everything else is a refinement of it. It's physical: you have a row of boxes.

Andrew: Describe it.

Ava: Every new card starts in box one. Box one you review every day. If you get a card right, it moves up to box two, which you review every three days. Get it right there, it moves to box three, reviewed weekly. And so on.

Andrew: And if you get it wrong?

Ava: It goes straight back to box one. All the way down.

Andrew: Which is beautifully simple, and it produces exactly the behaviour you want. Words you find easy stop taking up your time. Words you struggle with keep coming back.

Ava: And that's really the whole idea. Modern algorithms — SM-2 from SuperMemo, and the various things built on it — are doing something more sophisticated with the intervals, but the logic is identical.

Andrew: Should we say anything about which app to use?

Ava: Honestly, no, and I'd rather say why. The differences between the good ones are much smaller than the difference between using one consistently and not using one at all. Choosing the perfect app is a very effective way to avoid starting.

Andrew: Fair. Although I'll say the obvious thing, which is that LexUnio has this built into the vocabulary cards, so if you're already using the site you don't have to set anything up.

Ava: That's true, and it's the reason we built it that way — the cards you meet in a lesson go into the review queue automatically, so there's no second step where people give up.

— What to put on a card —

Andrew: Right. Let's get to the part where most people go wrong, which is not the schedule. It's what they put on the cards.

Ava: This is where I'd say the majority of wasted effort happens. People understand spacing now — it's fairly well known. What they don't get told is that a badly built card cannot be rescued by a good schedule.

Andrew: So what makes a bad card?

Ava: The most common one: a single word with a single translation and nothing else. Front: 'run'. Back: your language's word for run.

Andrew: Which seems like exactly what a vocabulary card should be.

Ava: It seems obvious and it fails for a specific reason. Words don't map one to one between languages. 'Run' in English is a verb about legs, and also what you do to a company, and also what water does, and also a sequence of successes, and also a hole in a stocking.

Andrew: So the card teaches you a correspondence that isn't true.

Ava: It teaches a correspondence that's true in one narrow sense and misleading everywhere else. And then you use the word confidently in the wrong context, and nobody corrects you, because they understood you anyway.

Andrew: So what goes on a good card?

Ava: A sentence. That's the short answer. Put the word inside a short, natural sentence — ideally one you actually met, not one you invented.

Andrew: Why one you met?

Ava: Because a sentence from real material is guaranteed to be natural. A sentence you construct yourself might contain the exact mistake you're trying to avoid, and now you're drilling it.

Andrew: That's a good point. You can memorise your own error.

Ava: Very efficiently, unfortunately. So: take the sentence where you met the word, cut it down if it's long, and put the word in a blank.

Andrew: So the front is a sentence with a gap, and you produce the missing word.

Ava: Which is recall, not recognition — so you're getting the testing effect. And you're learning the word in a context, so you're learning how it behaves, not just what it means.

Andrew: Let's talk about collocation here, because I think this is where advanced learners actually live.

— Collocation —

Ava: Collocation is which words naturally go together. In English we make a decision, we don't do a decision. We take a photo. We do homework. We have a shower.

Andrew: And there's no logic to it.

Ava: There's very little logic. It's convention. Which means it cannot be worked out from a dictionary definition — you can only pick it up from exposure, or by deliberately learning the pair rather than the word.

Andrew: And this is the difference between correct English and natural English.

Ava: It's almost the whole difference at C1. A C1 speaker rarely makes grammatical errors. What marks them out is collocation — 'I did a big effort' instead of 'I made a big effort'. Perfectly clear, immediately foreign.

Andrew: So practically: when you learn a verb, learn what it goes with.

Ava: Learn the phrase, not the word. And this costs you nothing extra, which is the beautiful part. 'Make a decision' is not meaningfully harder to memorise than 'decision'. It's the same card.

— Three kinds of knowing a word —

Andrew: Let's do the three kinds of knowing, because I think this clears up a lot of confusion about what 'I know this word' even means.

Ava: Right. Level one: you recognise it when you see it in text. That's the weakest kind, and it's where most vocabulary sits.

Andrew: Level two?

Ava: You recognise it when you hear it. And this is genuinely separate, which surprises people. You can know a word perfectly on the page and completely fail to catch it in speech, because you learned it visually and invented your own pronunciation for it.

Andrew: That's happened to me more times than I'd like to admit. You see the transcript afterwards and it's a word you've known for ten years.

Ava: Which is why learning words silently is a false economy. If you never hear a word, half of your knowledge of it is missing.

Andrew: And level three.

Ava: You can produce it. Unprompted, at speed, in the right form, with the right collocations. That's the expensive one, and that's the only one that helps you speak.

Andrew: So when somebody says 'I know five thousand words', the honest question is: at which level?

Ava: And the honest answer for most people is: five thousand at level one, maybe two thousand at level two, and perhaps eight hundred at level three.

Andrew: Which is not a criticism. That's a normal distribution for a learner. It just means you know where the work is.

Ava: Exactly. And it also means that going from five thousand to six thousand recognition words may do less for your speaking than moving three hundred existing words from level one to level three.

Andrew: That's a real strategic decision, and I don't think most learners know they're making it.

Ava: They're usually making it by default, in the direction of more new words, because new words feel like progress and deepening old ones doesn't.

Andrew: Progress that you can count.

Ava: Progress you can count is very seductive. Two hundred new cards this month is a number. 'My existing words got more reliable' is not a number, and it's worth more.

— Card rules and restraint —

Andrew: Let's talk about the rules for building cards, because there's a piece of writing on this that I think is underrated.

Ava: You mean Wozniak's twenty rules.

Andrew: Piotr Wozniak, who built SuperMemo — the first widely used spaced repetition software. He wrote a set of principles for formulating knowledge, and most of them apply directly to language.

Ava: The one that matters most is the minimum information principle. Which says: each card should test one thing, and one thing only.

Andrew: Which sounds obvious until you look at the cards people actually make.

Ava: Right. Somebody meets a new verb, and they put the verb, three meanings, two example sentences, the past tense, the past participle, and a note about which preposition it takes. All on one card.

Andrew: And what happens?

Ava: Two things, both bad. First, you can never say whether you got it right. You remembered four of the seven items — is that a pass? So the scheduling algorithm gets fed nonsense, and it can't schedule properly.

Andrew: And the second thing?

Ava: You start dreading the card. Which is worse, because dread is what ends the habit. A review session full of heavy cards is a session you skip tomorrow.

Andrew: So split it.

Ava: Split it. One card for the main meaning in context. A separate card for the preposition, if it's genuinely tricky. And honestly, skip the past participle unless it's irregular and you keep getting it wrong.

Andrew: Which brings up a broader point about restraint. I think people massively over-collect.

Ava: They do, and I've done it. You read an article, you find twenty-five interesting words, you make twenty-five cards. And it feels productive.

Andrew: And then a month later you have eight hundred cards and a review queue you can't face.

Ava: And then you stop entirely. That's the actual failure mode. Nobody quits spaced repetition because it doesn't work. They quit because the queue became a debt they couldn't pay.

Andrew: So what's the discipline?

Ava: Three words per session. And I mean that seriously, not as a soft suggestion. From a whole article, take three.

Andrew: How do you choose which three?

Ava: Two filters. First: would I plausibly use this word myself, in my own life, in the next year? If it's a technical term from an article about geology and you're not a geologist, let it go.

Andrew: And the second filter?

Ava: Did it block my understanding? If you couldn't follow the sentence without it, that word is doing real work. If you understood the sentence perfectly well and just found the word elegant, it's decoration.

Andrew: Elegant words are very tempting.

Ava: They're the most tempting. And they're the ones you'll never use, because you don't have a natural context for them, so you'll never retrieve them, so they'll sit in your queue forever costing you time.

Andrew: Let's do the arithmetic on that, because I think it's persuasive.

Ava: Go on.

Andrew: Three words a session. Three sessions a week. That's nine a week, call it four hundred and fifty a year — and these are words you actually retain, because the volume is low enough that the schedule works.

Ava: Versus twenty-five a session, which is three hundred a month on paper and zero in reality, because you abandoned the whole system in week six.

Andrew: Four hundred and fifty real words a year doesn't sound impressive.

Ava: It doesn't. But if those four hundred and fifty are words you chose because you'd actually use them, they change how you speak. And over three years that's more than a thousand active words, which is a completely different person linguistically.

— Interference: confusable pairs —

Andrew: Alright. I want to move to something we haven't touched, which is interference. Because I think this explains a very specific and very annoying experience.

Ava: Which is?

Andrew: Learning two similar words at the same time and permanently confusing them. You learn 'borrow' and 'lend' on the same day, and for the next five years you get them the wrong way round.

Ava: Or in German, learning 'legen' and 'liegen' together. Or in Spanish, 'ser' and 'estar' from the same lesson.

Andrew: And the standard advice is: study them together so you can compare them.

Ava: Which is exactly wrong, and this is one of the few places where I'd say the common advice is actively harmful.

Andrew: Explain why.

Ava: Because memories that are formed at the same time, in the same context, about similar things, compete with each other. That's what interference means. When you try to retrieve one, the other one comes up as well, and you can't tell which is which.

Ava: And the more similar they are, the worse the interference. Two words that mean roughly the same thing and sound roughly the same are the perfect recipe for permanent confusion.

Andrew: So what do you do instead?

Ava: Separate them in time. Learn one properly — get it solid, use it, own it. Then, weeks later, learn the second one. By then the first has its own established memory and the second attaches somewhere else.

Andrew: That takes patience.

Ava: It takes patience and it's much faster overall, because the alternative is five years of getting them backwards. And if you've already confused a pair — which most of us have — the fix is the same. Pick one. Drill only that one for a few weeks. Ignore the other completely.

Andrew: Just accept being wrong about the second one for a while.

Ava: Deliberately. Get one anchored, then bring the other in. Trying to fix both simultaneously is what kept them confused in the first place.

— Sleep and consolidation —

Andrew: Let's talk about sleep, because this is the part where people assume we're going to say something soft and we're not.

Ava: No, this is one of the better-established findings in the whole field. Memory consolidation — the process that turns a fragile new memory into a stable one — happens substantially during sleep.

Andrew: So the same study session, followed by a full night's sleep, produces measurably better retention than the same session followed by a short night.

Ava: Yes. And there's a practical consequence people don't draw from this: studying late at night, at the cost of sleep, can be worse than not studying at all.

Andrew: That's a strong statement.

Ava: It is, and I'll soften it slightly: worse than studying for half the time and sleeping properly. If you cut an hour of sleep to do an extra hour of vocabulary, you may have damaged the consolidation of everything you did that day, not just the last hour.

Andrew: There's a related tip that costs nothing, which is doing a short review just before bed.

Ava: That's a reasonable thing to try. The material is fresh going into the consolidation window. I'd flag that the effect sizes here are modest and it's not magic — but it costs you ten minutes and there's no downside.

— Building a system that survives —

Andrew: Alright. Let's build the system. Because we've covered a lot of theory and I want somebody to be able to act on this tonight.

Ava: Good. And I want to build it around one constraint that most study plans ignore, which is that you're going to have a bad week.

Andrew: Say more.

Ava: Every plan I've ever seen is designed for the version of you that has time, energy and motivation. But that person doesn't exist every week. You'll get ill. You'll have a deadline. Something will happen.

Andrew: And a plan that breaks when that happens isn't a plan.

Ava: It's a wish. So the system has to be built around what happens on the bad week, not the good one.

Andrew: So what's the minimum?

Ava: Five minutes of review. That's the floor. Not new words — just review. On your worst day, you open the queue, you do five minutes, you close it.

Andrew: Why five minutes rather than zero?

Ava: Two reasons, and the second one matters more. First, five minutes genuinely stops the decay on the cards that are due today. Second, and this is the real one: it keeps the chain intact.

Andrew: Because the danger isn't the missed day.

Ava: The danger is never the missed day. It's that a missed day becomes a missed week, and a missed week produces a queue of four hundred cards, and a queue of four hundred cards is where people quit permanently.

Andrew: So five minutes is a debt-prevention mechanism.

Ava: Exactly. It's not about the learning that happens in those five minutes. It's about the six months of learning that doesn't get abandoned.

Andrew: And what if you've already broken the chain? Because plenty of people listening have a queue sitting there right now that they're avoiding.

Ava: Then here's my genuine advice, and it's not what people expect: delete some of it.

Andrew: Delete.

Ava: Delete. If you have six hundred cards overdue, you are not going to clear them. You'll look at that number every day, feel bad, and not open it. That queue is not an asset, it's a psychological obstacle.

Andrew: So how much do you delete?

Ava: Keep the words you'd actually use. Delete the rest without ceremony. If a word matters, you'll meet it again — that's how frequency works. The common words come back on their own.

Andrew: That's oddly liberating.

Ava: It's very liberating, and people resist it because deleting feels like admitting failure. But a queue you don't open teaches you nothing, and a queue of eighty cards that you actually do every day teaches you a great deal.

Andrew: Let's put some numbers on the daily load, because 'a reasonable amount' isn't actionable.

Ava: Eight to ten new words a day is the sustainable ceiling for most people. And the reason it's a ceiling isn't the new words — it's the review load they generate.

Andrew: Explain that, because I think it's the thing nobody sees coming.

Ava: Every new card you add today creates reviews for the next several months. It's a subscription, not a purchase. So if you add fifty cards a day for two weeks, you've committed yourself to a review load that will crush you in about a month.

Andrew: And the crash comes long after the decision that caused it.

Ava: Which is why people can't see the cause. They think 'this app stopped working'. The app is working perfectly. They took out a loan six weeks ago and it's now due.

Andrew: So a rough rule: keep the daily review under about fifteen minutes.

Ava: Under fifteen minutes, and if it creeps above that, add fewer new cards until it comes back down. The review queue is the thing you protect. New words are optional; review isn't.

Andrew: I want to come back to something you said earlier about hearing words, because I think it deserves more than a passing mention.

Ava: The pronunciation problem.

Andrew: Yes. You said if you learn a word visually, you invent a pronunciation for it. Walk through what actually goes wrong there.

Ava: So you read a word — let's say 'thorough'. You've never heard it. Your brain does not leave a blank; it constructs a pronunciation from the spelling, based on your own language's rules.

Andrew: And that invented version gets stored as if it were real.

Ava: It gets stored as the word. So now when a native speaker says the actual word, it doesn't match anything in your memory, and you experience it as an unknown word. You look it up, and it's a word you've known for three years.

Andrew: Which is that experience of reading a transcript and being annoyed with yourself.

Ava: And the second cost is that when you eventually use the word, you say your invented version, and it may be unintelligible. And correcting an invented pronunciation is much harder than learning the right one first, because you have to overwrite something.

Andrew: So the fix is straightforward but requires discipline.

Ava: Hear every new word at least once before you file it. Every dictionary has audio now; it costs three seconds. And ideally say it out loud once — that adds a motor memory, which is a separate storage route.

Andrew: And for languages where spelling and sound diverge badly, this is not optional.

Ava: English is the worst offender among the major languages. French has its own version of the problem. In Spanish, Italian or Turkish you can get away with reading pronunciation from the page most of the time, because the spelling is regular.

Andrew: Which is a genuine reason some languages feel easier early on.

Ava: It's one of the real ones, yes. Not because the language is simpler, but because there's one fewer thing to get wrong.

— Extensive vs intensive reading —

Andrew: Let's talk about reading, because there are two completely different activities that both get called reading, and confusing them wastes a lot of time.

Ava: Extensive and intensive.

Andrew: Extensive reading is volume. Easy material, large quantity, no dictionary, no stopping. Intensive reading is the opposite: a short difficult text that you take apart carefully.

Ava: And they do different jobs, so doing only one of them leaves a hole.

Andrew: What does extensive reading actually give you?

Ava: Two things. It converts recognition into automatic recognition — words you know slowly become words you know instantly, which is what reading speed actually is. And it teaches collocation without you noticing, because you see the same pairs again and again.

Andrew: And the requirement is that it's easy.

Ava: It has to be easy. The research on this — Paul Nation's work is the standard reference — suggests you want to know around ninety-eight percent of the words for comfortable reading without a dictionary.

Andrew: Ninety-eight is very high. That's about one unknown word in fifty.

Ava: One in fifty, which is roughly one or two per page. And people find that number shocking, because they've been told to read difficult things.

Andrew: Which is the advice for intensive reading.

Ava: Exactly, and that's the confusion. Both pieces of advice are correct; they're just about different activities. For intensive reading, harder is right — that's where the seventy percent rule from last episode belongs. For extensive reading, easy is right, because the benefit comes from volume and speed.

Andrew: So a practical division.

Ava: Twenty minutes of easy reading with no dictionary, several times a week. And separately, one short difficult text that you actually dissect — where you look things up and pull three cards out of it.

Andrew: And the mistake is doing one activity badly in the middle.

Ava: The mistake is reading something too hard, stopping at every third word, taking forty minutes to get through two paragraphs, and then feeling terrible about it. That's neither extensive nor intensive. It's just unpleasant.

Andrew: Let's do the coverage numbers, because I think they're motivating in a way that abstract advice isn't.

Ava: Alright. The most frequent two thousand word families cover somewhere around eighty percent of general written English. That's a striking number.

Andrew: Eighty percent from two thousand words.

Ava: And then the curve flattens hard. To get from eighty to ninety-five percent, you need roughly another six thousand. To go beyond that, tens of thousands.

Andrew: So the first two thousand are extraordinarily valuable and everything after is diminishing returns.

Ava: Diminishing per word, yes — but that's not an argument to stop, because the gap between eighty and ninety-eight percent coverage is the gap between struggling through a text and reading it.

Andrew: It's the difference between decoding and reading.

Ava: And that's why the intermediate stage is long. You're climbing the flat part of the curve, where each individual word buys you very little, but the sum of them changes everything.

Andrew: Which connects straight back to episode one. That's the same plateau, seen from the vocabulary side.

Ava: It's the same mountain. Last time we looked at it psychologically; this time we're looking at the arithmetic underneath it.

— Word families: the biggest leverage —

Andrew: There's a piece of leverage we haven't mentioned, and I think it's the highest-return thing in this entire episode. Word families.

Ava: Morphology. Yes. And it's genuinely underused, because it looks like grammar and people file it away as boring.

Andrew: Explain the idea.

Ava: Most words in English aren't independent. They're built out of a root plus prefixes and suffixes. And once you can see the parts, one root gives you a whole cluster of words for almost no extra effort.

Andrew: Give me a worked example.

Ava: Take 'nation'. From that root you get national, nationality, nationalise, nationalisation, international, internationally, multinational, nationalism, nationalist. That's nine words and you learned one.

Andrew: Although you do have to know what the pieces mean.

Ava: That's the investment, and it's small. There are maybe forty prefixes and sixty suffixes that do most of the work in English, and they're extremely stable. Learn 'un-' means not, and you've unlocked thousands of words permanently.

Andrew: Let's do some of the high-value ones, because I think this is worth being concrete about.

Ava: Prefixes first. 'Un-', 'in-', 'im-', 'il-', 'ir-' all negate: unhappy, incorrect, impossible, illegal, irregular. Which one you use depends on the sound that follows, not on meaning.

Andrew: Which is why it's 'impossible' and not 'inpossible'.

Ava: Purely for ease of pronunciation. 'Im-' before p, b and m. 'Il-' before l. 'Ir-' before r. Nobody teaches this and it takes thirty seconds.

Andrew: More prefixes.

Ava: 'Re-' means again or back: rewrite, return, reconsider. 'Over-' and 'under-' mean too much and too little: overestimate, underpaid. 'Mis-' means wrongly: misunderstand, misread, mislead. 'Pre-' before, 'post-' after.

Andrew: And 'dis-'.

Ava: 'Dis-' is a useful one because it's not quite the same as 'un-'. 'Dis-' often carries a sense of reversal or removal — disconnect, disappear, disagree. There's a difference in feel between 'unlike' and 'dislike' that comes from exactly that.

Andrew: Now suffixes, which do a different job.

Ava: Prefixes usually change meaning. Suffixes usually change the part of speech — they tell you what kind of word it is. And that's enormously useful, because it means you can often work out how to use a word you've never seen.

Andrew: Go through the main ones.

Ava: '-tion', '-ment', '-ness', '-ity' all make nouns: decision, agreement, happiness, ability. '-ly' makes adverbs. '-ful' and '-less' make adjectives meaning with and without: careful, careless. '-able' and '-ible' mean can be done: readable, edible.

Andrew: And '-ise' or '-ize' makes verbs.

Ava: Makes verbs meaning to cause to become: modernise, privatise, normalise. And once you know that, you can produce a verb from an adjective without ever having met the verb.

Andrew: Which is a real skill. You can generate words you've never seen.

Ava: And be right most of the time. Not always — English has gaps and irregularities, and occasionally the word you construct doesn't exist. But the hit rate is high enough to be worth it, and being slightly wrong here is a small cost.

Andrew: How does this change what you put on a card?

Ava: When you meet a new word, spend twenty seconds asking whether it has relatives. If you meet 'reluctant', note 'reluctance' and 'reluctantly'. Same card, three words, and now you can use the idea as an adjective, a noun and an adverb.

Andrew: Which matters for writing especially, because you often need a different form than the one you learned.

Ava: That's exactly the moment it bites. You know 'reluctant' and you need the noun mid-sentence, and you stall. You didn't lack the vocabulary — you lacked one suffix.

— False friends —

Andrew: Let's talk about false friends, because our listeners come from a lot of different first languages and this trips everybody.

Ava: A false friend is a word that looks like a word in your language but means something different. And they're dangerous precisely because they don't feel like unknown words. You don't look them up — you're confident.

Andrew: Confidently wrong.

Ava: Which is worse than being unsure, because being unsure makes you check. Some English examples that catch a lot of people: 'sensible' does not mean sensitive; it means having good judgement. 'Actually' does not mean currently; it means in fact.

Andrew: That one catches speakers of half the languages in Europe.

Ava: Because in French, Spanish, Italian, German and several others, the equivalent word does mean currently. So it's a very consistent error, and it survives for years because it never blocks communication.

Andrew: More of them.

Ava: 'Eventually' means in the end, not possibly — that's German 'eventuell' catching people. 'Sympathetic' means showing compassion, not likeable. 'Library' is not a bookshop. 'Chef' is a cook, not a boss.

Andrew: And for Turkish speakers specifically?

Ava: There are words borrowed into Turkish from French and English whose meanings have narrowed or shifted. 'Kariyer' is career, that one's safe. But something like 'atlet' in Turkish is a vest, whereas English 'athlete' is a sportsperson.

Andrew: So the general rule?

Ava: When a word looks suspiciously familiar, that's exactly when to check it. Familiarity is the signal to be careful, not the signal to relax. Which is the inverse of what your instinct tells you.

— Forcing words into production —

Andrew: Alright. Last big topic, and it's the one that ties everything together. How do you actually force words into production?

Ava: Because that's the bottleneck. We've established that recognition is cheap and recall is expensive, and that speaking needs recall. So the question is what deliberately moves words across that line.

Andrew: And the honest answer is that it doesn't happen by accident.

Ava: Almost never. Words don't drift into your active vocabulary through repeated reading. They cross because you used them, deliberately, several times, in something you produced.

Andrew: So give us a method.

Ava: I'd call it the five-sentence rule and it takes about four minutes. Every day, take three words from your recent cards, and write five sentences using them. Not example sentences from a dictionary — sentences about your actual life.

Andrew: Why does it have to be your own life?

Ava: Because that's what makes it retrievable later. A sentence about a topic you'll never discuss connects to nothing. A sentence about your job, your city, your family attaches to memories you access constantly.

Andrew: So the retrieval route is stronger.

Ava: And it's realistic. If you write about your commute, you will one day talk about your commute, and the word will be there.

Andrew: Is writing enough, or do you have to say it?

Ava: Writing gets you most of the way, and it's what people will actually do. But saying it out loud adds something separate — the motor pattern. Your mouth has to learn the physical movement, and that's a different kind of memory.

Andrew: Which is why you can know a word perfectly and still stumble over it when you say it for the first time.

Ava: That stumble isn't a memory failure. It's your mouth doing something it's never done. Say each new word aloud three times when you file it and you've paid that cost in advance.

Andrew: There's another technique I like here, which is deliberately narrow topics.

Ava: Go on.

Andrew: Instead of studying general vocabulary, pick one topic and go deep for a week. Everything about cooking. Or everything about your job. Or everything about football.

Ava: Why does that work better?

Andrew: Two reasons. The words reinforce each other, because they appear together in real texts — so you get collocations for free. And you reach usable competence in that area quickly, which means you can actually have a conversation about it, which means you produce the words.

Ava: Whereas general vocabulary leaves you slightly able to discuss everything and properly able to discuss nothing.

Andrew: Which is a very common intermediate profile.

Ava: It is. And there's a confidence effect that isn't trivial. Being genuinely fluent about one subject changes how you feel in a conversation, even when the topic moves on.

Andrew: Because you've had the experience of it working.

Ava: You've had ten minutes where the language wasn't in the way. And that memory is worth a lot on the days when it is.

— Hedging, signposting, repair —

Andrew: There's one category of vocabulary we haven't discussed, and I'd argue it's the one that does the most work per word at your level.

Ava: You mean the words that aren't about anything.

Andrew: Exactly. Not nouns, not content. The connective tissue. Hedging, softening, signposting — the words that manage a conversation rather than carry information.

Ava: And they're invisible, which is why nobody studies them. Nobody finishes an article and thinks, I should note down 'to be fair'.

Andrew: But that's precisely what separates a C1 speaker from a B2 one, in my experience. Not rarer nouns. Better management of the conversation.

Ava: Let's take hedging first, because I think it's the most underrated skill in a second language.

Andrew: Define it.

Ava: Hedging is softening a claim so it isn't absolute. Instead of 'this is wrong', you say 'I'm not sure that's quite right'. Instead of 'you're mistaken', 'I might be missing something here'.

Andrew: And why does it matter so much for a learner specifically?

Ava: Because learners systematically sound blunter than they intend. Not because they're rude — because hedging is the first thing you lose when you have limited vocabulary. You produce the simplest form of the sentence, and the simplest form is usually the most direct.

Andrew: So somebody who is being perfectly polite in their own language comes across as abrupt in English.

Ava: Constantly. And they have no idea, because nobody tells you. People just find you slightly harsh and neither side knows why.

Andrew: That's a real cost.

Ava: It's a real professional cost, actually. In a meeting, 'that won't work' and 'I wonder whether that might run into problems' can carry the same information and land completely differently.

Andrew: So which ones are worth learning?

Ava: A small set covers most situations. 'I think', 'I'd say', 'it seems to me', 'as far as I know', 'correct me if I'm wrong'. Then the modal softeners: might, could, tends to, can be. And the qualifiers: fairly, quite, somewhat, a little, rather.

Andrew: Twenty items, roughly.

Ava: Twenty items that change how every sentence you produce is received. I'd put that ahead of two hundred nouns for anyone who uses English at work.

Andrew: Now signposting.

Ava: Signposting is telling the listener where you're going before you go there. 'There are two things here.' 'Let me come back to that.' 'To be clear, I'm not saying that.'

Andrew: And this has a hidden benefit for a learner that I don't think people notice.

Ava: Go on.

Andrew: It buys you time. When you say 'there are two things here', you've spoken a complete, correct, natural sentence that required no thought at all — and while you were saying it, you were working out what the two things are.

Ava: That's the real function of a lot of these phrases for native speakers as well. They're thinking time, dressed as structure.

Andrew: And a learner needs thinking time far more than a native speaker does.

Ava: Which is why memorising twenty or thirty of these phrases does more for perceived fluency than almost anything else. Fluency, to a listener, is largely the absence of long silences. Fill the silences with real language and you sound fluent.

Andrew: Even if the content underneath is simple.

Ava: Even if it's very simple. And I want to be careful not to make that sound like a trick, because it isn't. You're not faking anything — these phrases carry genuine meaning. You're just learning the ones that native speakers actually use.

Andrew: What about disagreeing? Because that's the hardest thing to do in a second language.

Ava: It's the hardest, and it's where the blunt-learner problem is worst. So it's worth having a few forms ready.

Andrew: Give us the ladder.

Ava: Softest first: 'I see what you mean, though I wonder if…'. Then 'I'm not sure I'd go that far'. Then 'I'd push back on that slightly'. Then, if you really need it, 'I disagree, and here's why'.

Andrew: And the point is having the choice.

Ava: The point is having a choice. At B1 you only have the last one, so every disagreement sounds like a confrontation. At C1 you can pick the temperature.

Andrew: Let's do one more category: repair.

Ava: Repair language. What you say when something goes wrong mid-sentence.

Andrew: Which happens constantly and derails people completely.

Ava: It derails them because they don't have a phrase for it, so they stop, apologise, look embarrassed, and lose the thread. Native speakers make just as many false starts — they just have language to cover them.

Andrew: What should learners have ready?

Ava: 'Sorry, let me start again.' 'What's the word — it's on the tip of my tongue.' 'How can I put this?' 'I mean…'. And crucially: 'What's the word for when…' followed by a description.

Andrew: That last one is genuinely powerful, because it turns a failure into a question.

Ava: It turns a breakdown into a normal conversational move. And often the person you're talking to gives you the word, and because you needed it at that exact moment, you never forget it.

Andrew: Which is the best possible learning condition.

Ava: It's the ideal one. Maximum need, immediate answer, strong emotional context. Words learned that way stick with almost no review.

Andrew: So the summary for this section?

Ava: Spend a week on words that aren't about anything. Twenty hedges, twenty signposts, ten repair phrases. It's fifty items and you could learn them in a fortnight.

Andrew: And they apply to every conversation you'll ever have, unlike a specialised noun.

Ava: That's the argument. Content words are worth a lot in one conversation. These are worth a little in every conversation, and the second one compounds.

— Listener questions —

Andrew: Right. Questions. We had a lot of messages after the first episode, so let's take four.

Ava: Go.

Andrew: First: is it better to learn words from a frequency list, or from things I read?

Ava: Frequency list at the beginning, your own material after that. And there's a clean dividing line.

Andrew: Which is?

Ava: The first thousand or two thousand words are the same for everybody. A list is efficient there, because nobody needs a personalised version of 'water' and 'because'.

Andrew: And after that?

Ava: After that, frequency lists get less useful fast, because word four thousand on a general list may be irrelevant to your life while a word ranked twenty-thousandth might be something you need weekly.

Andrew: Because your life isn't the average of the corpus.

Ava: Nobody's is. If you work in a hospital, medical vocabulary is high-frequency for you and low-frequency for the corpus. Past two thousand words, your own reading is a better list than any published one.

Andrew: Second question, and this one comes up constantly: should I use my own language on the card, or should the card be entirely in the target language?

Ava: Both work, and the argument about this is more heated than the evidence justifies. But I'd say: use your own language early, switch later.

Andrew: Why early?

Ava: Because at A1 and A2 you don't have enough language to define things in the target language. A definition made of words you also don't know teaches you nothing.

Andrew: And why switch later?

Ava: Because target-language definitions build the connections you actually use when speaking. If every word routes through your first language, you build a translation habit, and that's the exact thing you're trying to escape at B2.

Andrew: And in the middle?

Ava: In the middle, do what most people do naturally: concrete nouns in your own language, because 'table' is just 'table'. Abstract and near-synonymous words in the target language, because that's where translation misleads you most.

Andrew: Third question. How do I know when a word is really mine and I can stop reviewing it?

Ava: The honest test isn't recognition. It's whether the word turns up in your own output without you reaching for it.

Andrew: So you have to be producing language to know.

Ava: Which is another argument for writing regularly. Your own writing is a report on which words have actually transferred. If you've never written a paragraph, you don't have that information about any of your vocabulary.

Andrew: And practically, when do you drop a card?

Ava: When the interval gets long — six months or more — and you've used the word in your own speech or writing, let it go. The system will have pushed it out of the way anyway. And if you're wrong, you'll meet it again.

Andrew: Last question, and it's a difficult one. I've been studying for two years and I feel like I've forgotten more than I've learned. Is that possible?

Ava: It's possible to feel that. It's almost never true, and I want to say why carefully, because the feeling deserves a real answer rather than reassurance.

Andrew: Go ahead.

Ava: Forgetting is not deletion. A memory that you can't retrieve today is usually still there, weakened, and relearning it is dramatically faster than learning it was.

Andrew: That's Ebbinghaus again, actually — he measured this.

Ava: He called it savings. He'd learn a list, wait until he could recall none of it, and then relearn it — and relearning took far fewer repetitions than the first time. The memory left a trace even when it was completely inaccessible.

Andrew: So the two years weren't wasted.

Ava: The two years built a foundation that is invisible to you precisely because you can't retrieve it on demand. When you return to those words, they come back at a speed that would be impossible for genuinely new material.

Andrew: And there's the other thing, which is that your standard has risen.

Ava: That's the other half of it, and it's the more common cause. Two years ago you were pleased to understand a menu. Today you're frustrated that you can't follow a fast conversation about politics.

Andrew: Which is a description of massive progress that feels like failure.

Ava: It's progress measured against a target that keeps moving up. Which is why we said last time: keep a recording. Evidence, not feelings.

— Recap —

Andrew: Alright. Let's bring it together, because we've covered a lot of ground.

Ava: The mechanism first. Memory decays fastest immediately after learning — so a single exposure is close to worthless. Retrieval strengthens memory more than review does, so testing yourself is the training, not the assessment. And repetitions spread out over time beat the same repetitions bunched together, by a wide margin.

Andrew: Then the practice. Cards should test one thing. They should contain a sentence, not a bare word, and ideally a sentence you actually met. They should carry the collocations, because at C1 collocation is what separates correct from natural.

Ava: Hear every word before you file it, so you don't invent a pronunciation you'll have to unlearn. Separate confusable pairs in time instead of studying them together. And protect your sleep, because consolidation happens there and you can't shortcut it.

Andrew: And the system. Three new words a session, not twenty-five. Review under fifteen minutes a day. Five minutes as the floor on your worst day, to keep the chain from breaking. And if the queue has already collapsed, delete without guilt and restart small.

Ava: And if you keep one sentence from this hour, keep this one: the study method that feels most comfortable is usually the one that leaves the least behind.

Andrew: Which is unfair, and true.

Ava: It's completely unfair. Rereading feels productive. Struggling to recall feels like failure. And it's the second one that's building something.

Andrew: That's an hour of English listening you've just done, by the way, at B2 to C1 speed. That is not a small thing.

Ava: It genuinely counts. If you'd like to check where you are on the scale, LexUnio has a free level test — thirty questions, about ten minutes, no account and no card needed, and it gives you a CEFR result. It's at lexunio dot com.

Andrew: And the same honest note as last time: that test measures reading, grammar and vocabulary. It does not measure speaking or listening. Take the result as a starting point, not a verdict.

Ava: Next time we're going down a level. Episode three will be at A2 to B1, spoken more slowly, on a topic that comes up constantly: how to start speaking when you're afraid of making mistakes.

Andrew: And after that, we're doing this in German.

Ava: Thanks for staying for the whole hour. Go and test yourself on something instead of rereading it.

Andrew: Uncomfortable, and effective. See you next time.