How Many English Words Do You Need to Read and Speak?

How Many English Words Do You Need to Know?

The number of English words you need depends on what you want to do with them. Paul Nation's 2006 study in the Canadian Modern Language Review estimated that reading novels and newspapers without help takes 8,000 to 9,000 word families, and following unscripted speech takes 6,000 to 7,000. Both figures assume you know 98% of the running words.

A word family is the unit these English vocabulary figures count. It is a base word plus its inflected and derived forms. Norbert Schmitt and colleagues, in a 2015 paper in the journal Language Teaching, give the example of stimulate, stimulated, stimulating, stimulates, stimulation, and stimulative: six forms, one family. So 8,000 word families is far more than 8,000 separate forms, and counting by family assumes that a learner who knows stimulate can recognise stimulation. It says nothing about telling apart look-alike words from different families, a separate problem covered in why English learners confuse similar words.

The 98% figure itself comes from experiments on reading comprehension, which comprehensible input and finding your right English reading level walks through. Nation's contribution was converting that percentage into a vocabulary size. Schmitt and colleagues point out an important limit on how to read his numbers: the study "did not involve actual learners" and did not measure comprehension. It profiled texts by word frequency, so the 8,000 to 9,000 figure is a corpus estimate, not a test result.

How Many Words Do You Need for Books, Movies, and Conversation?

The English vocabulary size you need changes with the kind of text. In Nation's 2006 corpus analysis, novels and newspapers needed about 8,000 to 9,000 word families for 98% coverage, a children's movie needed about 7,000, unscripted conversation 6,000 to 7,000, and a graded reader only 3,000. A 2022 study of online course lectures put the figure near 9,000.

The graded reader row explains why simplified books work for learners. The Oxford Bookworms edition of The Picture of Dorian Gray that Nation profiled keeps within a 1,000-word list and uses 682 word families in total, so a reader with 3,000 word families already knows almost every word on the page.

Even the top figures leave gaps. Nation notes that with 8,000 to 9,000 word families, "1 word in 50 will be unfamiliar" in a novel, and most of those words appear only once or twice. Proper nouns are counted separately throughout, because names like London or Gatsby are usually easy to work out from context.

How Many Words Do You Need to Understand Spoken English?

Understanding spoken English takes fewer words than reading. Nation's 2006 analysis found 3,000 word families plus proper nouns cover more than 95% of unscripted speech, and 6,000 to 7,000 reach 98%. A listening experiment by van Zeeland and Schmitt (2012) found 95% coverage gave consistent comprehension, which would put the target nearer 2,000 to 3,000 word families.

Nation's spoken samples came from the Wellington Corpus of Spoken English: one part talk-back radio and interviews, the other friendly conversation between family and friends. Spoken English leans harder on the most frequent words, so the same coverage arrives sooner. Nation also warns of the opposite pressure. Speech is heard once and is gone, so listening may need more than 98% coverage to feel comfortable, not less.

Van Zeeland and Schmitt's listening study, as summarised in Schmitt and colleagues' 2015 paper, replaced words in four short first-person anecdotes with nonwords. Listeners then answered ten comprehension questions per story.

Scores at 95% and 90% were not statistically different, but they varied much more between listeners at 90%, which is why the authors chose 95% as the more reasonable criterion for adequate listening comprehension.

The conditions were generous. The stories lasted about two minutes, listeners heard each one twice, and narratives have been found to be the most comprehensible genre for listeners, so the original authors acknowledge the results might be viewed as "best-case" performance. Schmitt and colleagues list lectures, everyday chat, and single listenings as conditions still to be tested.

Is 95% or 98% Coverage the Right Target?

Both coverage figures are defensible, and they answer different questions. Laufer and Ravenhorst-Kalovski (2010), as summarised by Schmitt and colleagues in 2015, treat 98% coverage as sufficient for independent reading by most learners and 95% as the minimal level, probably viable with support such as a dictionary or a teacher. The size of the vocabulary target roughly halves between the two.

The difference sounds small until you count it on a page. At 95% coverage, Schmitt and colleagues calculate, a reader meets about one unknown word in every two lines, which is over 15 unknown words on every page. Nation puts the listening version in time: 95% means about 7 unknown words in every minute of speech at 150 words per minute.

Schmitt, Jiang and Grabe (2011), also summarised in the 2015 paper, tested every coverage point between 90% and 100% and found a linear relationship between coverage and comprehension. There is no single cliff edge. Each extra percentage point of known words buys some extra understanding, and by their data, reaching 60% comprehension took 98% coverage.

For a learner choosing a target, that gives a practical rule:

• Aim for 98% (about 8,000 to 9,000 word families) if you want to read authentic novels and news without a dictionary.
• Aim for 95% (about 4,000 word families plus proper nouns for written text, around 3,000 for speech) if you will read with support or mainly listen.

The page-level test for checking where a specific book sits, and the 2000 Hu and Nation experiment behind the 98% figure, are covered in comprehensible input and how to find your English reading level.

How Many Words Do You Need for University Lectures and Online Courses?

Academic listening is one of the most demanding uses of English vocabulary. A 2022 corpus study in Frontiers in Psychology analysed about 4.45 million words of MOOC video lectures and found that 95% coverage needs the 5,000 most frequent word families, and 98% needs around 9,000, more than the 6,000 to 7,000 Nation found for everyday speech.

Xodabande, Ebrahimi and Karimpour built the corpus from lectures in Massive Open Online Courses and profiled it against frequency lists built from the British National Corpus and the Corpus of Contemporary American English. The first 1,000 word families alone covered 80.48% of the lecture words, and the first 5,000 together reached 95.15%.

The study also tested the General Service List and the Academic Word List, two word lists commonly taught to students preparing for study in English. Together they covered 90.86% of the lecture corpus, and the authors conclude that the two lists fail to give MOOC learners enough vocabulary for adequate coverage.

For a learner heading toward a degree taught in English or an exam like IELTS or TOEFL, the practical reading is that general high-frequency words are not enough on their own. Academic and lower-frequency words have to be added deliberately, and the advanced English vocabulary list gathers the hard but useful words from academic writing and those exams, each explained in plain English with an example.

How Many English Words Do You Need for Each CEFR Level?

The CEFR describes vocabulary range in words, from a "basic vocabulary repertoire" at A1 to a "very broad lexical repertoire" at C2, not in word counts, so researchers have estimated the numbers. James Milton and Thomaï Alexiou (2009) linked vocabulary test scores to CEFR levels, putting A1 learners below 1,500 of the 5,000 most frequent words and C2 learners at 4,500 to 5,000. Paul Nation's own suggested targets run higher from B2 upward.

The two columns measure different things. Milton and Alexiou's figures come from X-Lex, a test of only the 5,000 most frequent words, and the authors call them "indicative, not being absolute requirements for each level". Nation's column comes from a resource page on his university's site, where he notes his sizes from B2 on "are largely greater than others may recommend, for example Jim Milton".

In Nation's column, C1 and C2 sit beyond the 5,000 most frequent words, which a test capped at 5,000 cannot see, and his C2 range of 7,000 to 9,000 overlaps the 8,000 to 9,000 word families his 2006 paper found for unaided reading. If you do not yet know your level, how to find your CEFR English level covers free tests and a self-assessment grid. Words sorted by level are in the A1 beginner vocabulary list, the B1 intermediate vocabulary list, and the C1 advanced vocabulary list.

How Many Words Does a Native English Speaker Know?

A well-educated native English speaker knows around 20,000 word families, according to the conservative estimates Paul Nation cites in his 2006 paper, counting neither proper names nor transparently derived forms. Nation's rule of thumb is that a native speaker adds about 1,000 word families for each year of life up to the age of 20 or so.

Other measurements land in the same range. Goulden, Nation and Read (1990), cited by Schmitt and colleagues, found that New Zealand university undergraduates had an English vocabulary of about 17,000 word families. Nation adds that many published native-speaker estimates are "wildly inflated" because of faulty methods, which is why figures far above these turn up in popular articles.

The native-speaker number is a poor target for a learner, for two reasons. The first is size: Nation calls such figures "very ambitious goals for a learning program." The second is that the coverage research shows a learner can read novels and newspapers at 98% coverage with less than half that vocabulary. Nation also reports, from unpublished trials of his vocabulary size test, that highly educated non-native speakers studying advanced degrees in English had receptive vocabularies of around 8,000 to 9,000 word families.

Which English Words Should You Learn First?

Learn English words in order of frequency, starting with the most common 1,000. Nation's 2006 data shows the first 1,000 word families cover 78 to 81% of typical written English and 81 to 84% of spoken English. Each later block of 1,000 adds less, so the earliest words do most of the work.

Source: Nation 2006, Table 14, average coverage across the written and spoken corpora in the study.

The shape of that table sets the strategy. The first 3,000 word families take a learner to about 89 to 95% coverage of written English, so they are worth studying directly, with a list. The 1000 most common English words list on the Dobi Apps English vocabulary guide gives each of those core words with its meaning, pronunciation, and an example sentence, and the English function words list covers the small grammar words that make up much of that first 1,000.

Beyond about 5,000, each new block of 1,000 word families adds only fractions of a percent. Those words are too many and too rare to drill one by one, which is why the later stage of vocabulary growth depends on reading and listening rather than lists.

How Do You Grow Your English Vocabulary Toward 8,000 Word Families?

Growing English vocabulary to 8,000 word families takes deliberate study of high-frequency words plus a large amount of reading and listening. Classroom progress is slow: across the classes of one language school in Greece, James Milton estimated that learners of English gained about 500 words a year from about 100 contact hours.

Schmitt and colleagues explain the stall with Tom Cobb's 2007 analysis. A word normally has to be met many times before it sticks, and "a figure of 10 is often cited", but words beyond the 3,000 most frequent families turn up too rarely in typical input to give learners those repetitions. Cobb used this to explain why learner vocabularies so often stop at 2,500 to 3,000 word families, and Schmitt and colleagues conclude that if he is right, "relatively few" learners will reach 8,000.

That points to two steps:

1. Study the first 3,000 word families deliberately. Review them on a schedule rather than in one sitting, the method covered in how spaced repetition builds English vocabulary retention.
2. Read and listen in volume at your level. Volume is the only way to meet mid-frequency words often enough. Nation found that topic words repeat within a single long text: ogre recurs throughout the Shrek script, so a viewer who works it out early meets it as a known word from then on. Graded stories, described in how to learn English effectively through stories, keep that reading at a comprehensible level.

Word counts also leave something out. Schmitt and colleagues note that multiword units such as idioms are "seldom included in vocabulary counts", so a learner at 8,000 word families can still miss phrases whose meaning is not the sum of their words, a gap explored in do idioms make you sound fluent in English.