How memory works

Spaced repetition

Reviewing material at increasing intervals over time rather than in a single session. The same total study time produces better long-term retention when spread out, which is one of the most consistently reproduced findings in memory research.

Covered in more depth in why you forget vocabulary.

Spaced repetition system (SRS)

Software that schedules spaced reviews for you, stretching the interval each time you recall an item correctly and shortening it when you fail. Anki is the best-known example.

The forgetting curve

The pattern by which memory for new material falls steeply at first and then levels off. First mapped by Hermann Ebbinghaus in the 1880s using nonsense syllables; the shape holds up in modern replications, though meaningful material learned in context fades more slowly than his syllables did.

Retrieval practice

Being asked to produce something from memory without being shown it first. Retrieval strengthens memory more than re-reading does, which is why flashcards outperform reviewing a word list even though reviewing feels easier.

Interleaving

Mixing different topics or types of problem within a study session rather than practising one thing to exhaustion before moving on. It feels less productive than blocked practice and generally produces better retention.

Words, and how they are counted

Active vocabulary

The words you can produce when you need them, unprompted, while speaking or writing. Usually estimated at somewhere between half and two-thirds the size of a learner’s passive vocabulary.

Passive vocabulary

The words you recognise and understand when you read or hear them, but would not necessarily produce yourself. Passive vocabulary grows through exposure; active vocabulary grows through retrieval.

The gap between the two is why people follow conversations they cannot join — see how many words you need to speak a language.

Word family

A base word together with its inflections and close derivatives, counted as one item. "Walk", "walks", "walked" and "walking" are one word family. Most vocabulary size estimates are given in word families, which is why they look smaller than raw word counts.

Lemma

The dictionary or citation form of a word, under which its inflected forms are grouped. Lemma counts fall between word families and raw word forms, so a vocabulary figure means little until you know which unit it uses.

Frequency list

Vocabulary ordered by how often each word occurs in a body of real language. Learning in frequency order buys far more comprehension per word than learning alphabetically or by theme, because word usage is extremely top-heavy.

Collocation

Words that habitually occur together, where the pairing is conventional rather than logical. English says "make a decision" and "do the dishes"; swapping the verbs is grammatical but wrong. Collocations are learned as units.

Cognate

A word that shares an origin with a word in another language, and usually a recognisable form and meaning. English and French share thousands, which is a large part of why French is easier for English speakers than the raw grammar suggests.

False friend

A word that looks like a cognate but does not mean what you expect. Spanish "embarazada" means pregnant, not embarrassed. False friends are dangerous precisely because they feel like free vocabulary.

Phrasal verb

A verb combined with a particle, where the meaning cannot be worked out from the parts. "Give up", "put off" and "get by" have to be learned as whole items, which is why they are a persistent difficulty for learners of English.

Loanword

A word adopted from another language with little or no translation. Loanwords give learners a foothold, though pronunciation and meaning often drift from the original.

Frameworks and measures

CEFR

The Common European Framework of Reference, a six-level scale from A1 to C2 used by most language courses and exams. It defines levels as "can do" statements describing what a learner is able to accomplish, and does not specify vocabulary sizes or study hours.

What each level means is set out in CEFR levels explained.

FSI categories

The US Foreign Service Institute’s grouping of languages by how long they take English speakers to learn, from roughly 600 hours for Category I to about 2,200 for Category IV. The hours describe intensive full-time instruction toward professional proficiency, not conversational ability.

The categories and what they translate to at a normal pace are in how long it takes to learn a language.

Fluency

A word with no standard technical definition, which is why estimates of "how long to fluency" vary so wildly. In practice most people mean roughly CEFR B2: able to follow extended discussion and use the language at work without constant strain.

Receptive and productive knowledge

Receptive knowledge is understanding language you are given; productive knowledge is generating it yourself. Any vocabulary statistic should say which it counts, and many do not.

Ways of practising

Comprehensible input: Language slightly beyond your current level that you can still mostly understand, often written "i+1". The idea, associated with Stephen Krashen, is that acquisition happens when you understand messages rather than when you study rules.
Immersion: Surrounding yourself with the language so that ordinary daily activity happens in it. Full immersion means living in it; partial immersion means deliberately shifting media, reading and routine into the language.
Shadowing: Repeating recorded speech aloud immediately as you hear it, without waiting for a pause. Used to train pronunciation, rhythm and speed rather than comprehension or vocabulary.
Passive review: Meeting material again without a dedicated study session — a widget, a poster, a notification. It builds recognition through repeated exposure and cannot build recall, because nothing asks you to produce anything.

A note on the numbers

Several of these terms exist because vocabulary statistics are so inconsistent. A claim like "native speakers know 20,000 words" means three different things depending on whether it counts word families, lemmas or word forms, and whether it means receptive or productive knowledge.

When you meet a figure without that context, treat it as an order of magnitude rather than a measurement. We have tried to say which unit is in play wherever we quote one, in how many words you need to speak a language and CEFR levels explained.