Vocabulary size — how many words you know — is the single strongest predictor of reading comprehension in a second language. In the Refold method, word count milestones are used as rough checkpoints, not as strict requirements.
The Numbers
- 1,000 words: ~70% of most text. You can understand the structure of sentences and many common words, but miss a lot of meaning. (End of Phase 1)
- 2,000 words: ~80% of most text. You understand the backbone of sentences. Still uncomfortable. (End of Phase 2A)
- 3,000 words: ~85% of texts and content. Starting to follow stories with some support. (End of Phase 2B)
- 4,000 words: ~90% of texts and content. Comfortable reading with subtitles in familiar domains. (End of Phase 2C)
- 5,000 words: ~98% of every day conversation. Comfortable talking about normal topics with natives (Phase 5).
- 8,000 words: ~98% novels and most things you'd want to engage with. Comfortable across most everyday topics. (Phase 6)
- 10,000+ words: ~98%+ of basically everything. Near-native comprehension, but probably still more to learn. (Phase 7)
These numbers are approximate and vary by language.
Why It Matters
Vocabulary is the foundation everything else is built on. Grammar knowledge doesn't help if you don't know the words. Listening practice doesn't help if you can't recognize the vocabulary. In the early and middle phases, learning more words is almost always the highest-impact thing you can do.
How to Count
Most learners don't need an exact count. If you're using Anki, your number of "mature" cards gives a rough estimate. Some languages have vocabulary tests online. Or some tools have rough counting methods.
But nothing will be perfect. What is a word, anyway? A rose by any other name would smell as sweet. 🌹
The exact number matters less than whether you're meeting the comprehension benchmarks described in each sub-phase's "When to Move On" section.
Research and Reasoning
Vocabulary size is one of the most robust predictors of language comprehension. Nation (2001), in his comprehensive review of vocabulary research, established that approximately 1,000 words covers about 70% of everyday written and spoken English, and this pattern holds across many languages. Hu and Nation (2000) further demonstrated that learners need roughly 98% lexical coverage for comfortable unassisted reading — a finding that directly supports the vocabulary milestones in the roadmap.
The research distinction between reading comprehension and listening comprehension matters: Laufer (1992) showed that vocabulary size is the single strongest predictor of reading comprehension, while other factors (prosodic cues, contextual guessing) play larger roles in listening. This explains why the roadmap prioritizes vocabulary in early phases and emphasizes reading-based immersion.
Recent work by Webb and Nation (2017) clarifies that there's no single point at which you "know" a word. Rather, vocabulary knowledge is multidimensional and develops across receptive (understanding) and productive (using) dimensions. The milestones in this article track primarily receptive vocabulary, which is why productive fluency comes later and requires additional practice through output.