Word frequency and learning: which words to study first

Sep 16, 2026 · 11 min read

Start with the most frequent words, because they pay back more often. Roughly the first 2,000 word families of English cover the large majority of an ordinary page of writing, and a word you meet four times a day practises itself. Every hour spent on the ten thousandth most common word is an hour that buys you almost no new reading.

Frequency is not the only principle in vocabulary study — interest, need and the genre you actually read all matter — but it is the only one that tells you where to begin when you have no strong reason to choose. The sections below cover how the counts are made, why the curve is so steep, how the A1 to C2 lists are meant to be used, and where topic vocabulary sits alongside the core.

How frequency is counted

Frequency figures come from corpora: large, machine-readable collections of text and speech, each with a count for how often a word appears per million words. The main public corpora draw on different material — speech, newspapers, fiction, academic writing, film subtitles — and they disagree in the middle and lower ranks. A word that is common in academic prose can be rare in conversation, and a word that fills film dialogue may barely appear in a research paper.

Treat a frequency rank as a good guide rather than a law. When two sources disagree about a word you are about to learn, ask a simpler question: in the material I read or listen to each week, how likely am I to meet this word?

The core and the tail: Zipf's law in one minute

Word counts follow a steep curve known as Zipf's law: the frequency of a word is roughly proportional to one over its rank. The practical reading of that rule is uncomfortable. The most common word in English appears about twice as often as the second, and a word ranked 100th is about ten times less frequent than one ranked 10th. By rank 1,000, a word turns up a few times per million — which in a daily reading habit may mean once every few weeks.

The curve explains both halves of the strategy. The top of the list is cheap because the world keeps rehearsing it for you. The tail is expensive: nothing in daily life brings the word back, so every encounter has to be arranged deliberately.

Why frequent words cost less to learn

The difference is not in the word itself but in the number of second chances it gets. Learn however this morning and you will meet it again before the week is out, probably several times, in contexts that narrow its use. Learn a technical term with the same effort and you may not see it again until you go looking for it.

Two consequences follow. The top of the list is where a learner gets the most practice per unit of study. And the tail needs a schedule: if nothing else brings the word back, you must decide the next encounter yourself — the routine set out in our guide to using word lists effectively, and the principle behind spaced repetition.

Using the level lists, A1 to C2

BandRough sizeWhat it unlocks
A1–A21,000–1,500 familiesSigns, forms, short messages, simple conversation on familiar topics
B12,000–2,500 familiesEveryday reading, most of a news item, work emails with some effort
B23,000–4,000 familiesArticles, films without subtitles, meetings where you can steer the topic
C16,000–8,000 familiesLong-form writing, nuance, most professional and academic texts
C210,000+ familiesLiterary and specialist registers, idiom, fine shades of stance

The figures are estimates of word families, not of dictionary entries, and different sources place the bands a little differently. Use them as a map rather than a certificate, and work near the top of your current band: material that is one level up stretches you, while material two levels up produces a page of unknown items and very little practice.

A level list also solves an ordering problem. Our basic word list collects high-frequency items in sets you can finish, and the word list library links the same sets to games so a finished set can be tested straight away. For learners past the core, the academic word list sublist 1 covers vocabulary that appears across research writing rather than inside one subject.

Frequency is not the same as usefulness to you

Corpus counts average over everyone, and you are not everyone. A nurse reading English case notes, an engineer reading specifications and a student reading nineteenth-century novels share the first 2,000 words and then diverge. After the core is in place, the highest-value words are usually the ones that keep appearing in your own weekly material — and those are easy to find, because they are the words you write down twice without noticing. One exception: if a rare word blocks your work this month, learn it now. Frequency ranks the queue; need jumps it.

The second curve: collocation frequency

Single words are only half the picture. Combinations have their own frequencies, just as uneven: a frequent verb and a frequent noun can still be a combination that no one uses. This is why a learner who knows every word in a sentence can still produce one that sounds wrong.

The fix is to store each new word with a partner: not decision but make a decision, not rain but heavy rain. Our guide to chunking vocabulary sets out how to do that in larger blocks, and it costs almost nothing at the moment of first learning — you were going to write the word down anyway.

How many words, and in what order

Twenty words a week is a workable target for most learners, and better than fifty attempted once. Split the block roughly three to one in favour of the core: fifteen high-frequency items and five from your own reading, your subject or the exam you are preparing for. The five keep the study connected to why you started; the fifteen keep the reading getting easier.

Then let the schedule do the work. Each block gets three passes at widening gaps — the day you learn it, again two or three days later, and once more about a week on. A block that fails the third pass goes into the next week's set rather than being repeated immediately.

Four ways learners waste time

Where to start this week

Take one set of twenty from the top of the basic list and read them in the example sentences before you do anything else. Then test the set in a game and keep only the misses: the flash recall round is the gentlest first pass, and falling words adds the time pressure that shows whether a word is genuinely available. When the set survives a second test two days later, add five words from something you are reading anyway.

Frequently asked questions

How many words do I need before everyday reading gets easier?

Most learners notice a change somewhere between 2,000 and 3,000 word families. Below that, nearly every paragraph holds two or three items that stop the reading; above it, the interruptions become occasional rather than constant.

Should I learn the 2,000 most frequent words in list order?

Not strictly. Frequency sets the order, and interest decides the details. Take blocks of twenty from the top of the list, then add a handful of words from whatever you actually read that week.

Are rare words a waste of time?

No, once the core is in place. A rare word costs more to hold because you meet it so seldom, but in a specialist field those words do the heavy lifting. Learn the core first so a rare word is a small addition, not a load-bearing wall.

Why do I forget words I have already learned?

Usually because they were recognised once rather than recalled later, and because they sit just below the top of the list where your reading does not practise them often enough. Scheduling a review is what fixes the gap.

Do collocations have their own frequency?

Yes. Words pair unevenly, and the stronger pairs are worth learning whole. Two words can both be frequent and still sound wrong together, which is why a frequent noun belongs with the verbs it usually takes.

A word list is a starting point, not a shortcut. How quickly a set becomes usable depends on how often you produce the words yourself.