How many words do you need to be fluent?

This is probably the most-asked question in language learning. People want a target number they can work toward, something concrete in what otherwise feels like an endless process. The internet is full of answers ranging from 500 to 50,000, which isn't very helpful.

The real answer depends on what you mean by "fluent," how you count words, and whether you're talking about recognizing a word on a page or actually pulling it out of your brain mid-conversation. Once you understand those distinctions, the research gives us some genuinely useful numbers to aim for.

The most cited work on this comes from Paul Nation at Victoria University of Wellington. Nation's frequency research (2001) found that knowing the most common 3,000 word families in a language gives you coverage of roughly 95% of spoken language. That means in an average conversation, you'd understand 19 out of every 20 words. For written text, which uses a wider range of vocabulary, you need about 8,000 word families to reach 98% coverage, which is the threshold where you can read comfortably without constantly reaching for a dictionary.

A "word family" is an important concept here. It means a base word plus all its inflected and derived forms. So "run," "runs," "running," "runner," and "ran" all count as one word family. When someone says they know 10,000 words, they might be counting each of those as a separate item. In word-family terms, the number is considerably smaller. Most vocabulary research uses word families as the unit because learning "run" effectively means you can also recognize "running" without much extra effort.

The 95% threshold sounds high, but think about what it means in practice. If someone says a 20-word sentence and you miss one word, sometimes you can fill in the gap from context. Other times that one word was the key noun or verb, and the whole sentence falls apart. At 98% coverage, you're missing 1 word in 50, which is comfortable enough to keep reading or listening without losing the thread. That gap between 95% and 98% is the difference between "I mostly get it" and "I can actually enjoy this."

Knowing 3,000 words passively is a very different thing from knowing 3,000 words actively. Passive knowledge means you can recognize a word when you see or hear it. Active knowledge means you can produce it in speech or writing, with correct grammar and natural collocations. Most learners have a passive vocabulary two to three times larger than their active one. So when someone says they "know" 5,000 words on Duolingo or Anki, what they usually mean is they can match those words to translations on a card.

Depth matters as much as breadth. Take the English word "get." Most intermediate learners know "get." But do they know "get over it," "get by," "get away with"? Each is essentially a different word. A learner with 2,000 deeply known words will outperform a learner with 5,000 shallow word-translation mappings in almost every real-world situation. Chasing a word count as your primary metric can be counterproductive.

Norbert Schmitt (2008) laid out the dimensions of word knowledge in a way that's useful for learners. Knowing a word means knowing its spoken form, written form, grammatical behavior, collocations, frequency, register, and associations. Your first encounter with a word might give you a vague sense of meaning, enough to recognize it if you see it again soon. But that initial trace is weak. It takes multiple encounters in different contexts before the word moves into reliable long-term memory and becomes something you can actually use.

This is the strongest argument for learning vocabulary from context rather than from word lists. When you encounter a word in a video, a conversation, or a story, your brain automatically encodes information about pronunciation, grammar, and usage that a flashcard simply cannot provide. Each new context adds another layer. After five or six varied encounters, the word starts to feel like something you own rather than something you've memorized.

There's a problem with learning vocabulary purely through exposure: you forget things. The forgetting curve, first described by Ebbinghaus and confirmed by modern research (Cepeda et al., 2006), shows that without review, you lose about 60-70% of new information within a few days. That means if you learn 10 words this week from watching videos but never revisit them, you'll remember maybe three or four by next month. You end up on a treadmill, learning and forgetting at roughly the same rate. Spaced repetition solves this by scheduling reviews at increasing intervals. You see a word after one day, then three days, then a week, then a month. Each successful review strengthens the memory trace and pushes the next review further into the future.

Kornell (2009) found that spaced practice produced significantly better long-term retention than massed practice (cramming), even when learners themselves believed cramming was more effective. The research suggests you need roughly 5 to 7 spaced encounters with a word before it sticks in long-term memory. Combining contextual learning (where you first encounter and save words) with spaced repetition (where you systematically review them) means your active vocabulary grows steadily instead of constantly leaking.

The abstract question "how many words do I know?" is hard to answer honestly. You can take vocabulary size tests, but they measure recognition, not production. You can count flashcards, but that conflates cards you've seen once with cards you've mastered. What's more useful is tracking how many words you're actively studying, how many you've retained over time, and what your review accuracy looks like.

Langadoo tracks exactly this. You can see how many words you've saved, how many are in active spaced repetition review, and what your retention rate is over the past week and month. That makes the question concrete. Instead of wondering whether you "know" 2,000 words, you can see that you have 1,847 words in review with an 86% retention rate, and that you're adding about 12 new words per day. You can watch yourself approaching the 3,000-word conversational threshold and know the number is real, because every word in that count has been reviewed and retained.

With consistent daily study of 15 to 30 minutes, here's what a reasonable trajectory looks like. In the first two months, you can expect to build an active vocabulary of around 500 words. These are the highest-frequency words in your target language, the ones that appear in almost every conversation and every piece of content. By six months, you should be around 1,500 words, which is enough to follow simple videos and hold basic conversations. By a year, 3,000 active words is achievable, putting you at that 95% conversational coverage threshold.

The critical caveat: these numbers assume you're learning from context with spaced repetition, not just passively recognizing words on flashcards. A learner who saves words from videos, reviews them in sentence context, and keeps up with their spaced repetition queue will hit these milestones. A learner who grinds through a pre-made word list without any contextual exposure will have a much larger "studied" count but a much smaller usable vocabulary. The method matters as much as the time invested.

Track your word count, learn from real videos, and review with spaced repetition. Free plan available, no credit card required.

References: Nation, I.S.P. (2001). Learning Vocabulary in Another Language. Cambridge University Press. Schmitt, N. (2008). Review article: Instructed second language vocabulary learning. Language Teaching Research, 12(3), 329-363. Cepeda, N.J., Pashler, H., Vul, E., Wixted, J.T., & Rohrer, D. (2006). Distributed practice in verbal recall tasks: A review and quantitative synthesis. Psychological Bulletin, 132(3), 354-380. Kornell, N. (2009). Optimising learning using flashcards: Spacing is more effective than cramming. Applied Cognitive Psychology, 23(9), 1297-1317.