Vocabulary depth - knowing more about the words you already have, not more words - produces faster comprehension gains than expanding word count once you pass the 2,000-word threshold in a new language.
Most intermediate learners never change what worked at the beginner stage. They keep adding words: themed vocabulary lists, mining unknown words from podcasts and shows, growing their flashcard decks. The logic holds at the start, when every new word increases the percentage of text you can understand. The problem is that this logic stops holding around the 2,000-word mark - and most learners do not notice when it does.
What Most Intermediate Learners Do Instead
A learner who reaches the intermediate stage with 2,000 German words can understand roughly 80 percent of everyday spoken text. The natural next move is to push toward 3,000, then 4,000, treating vocabulary progress as a count to increase.
This approach is intuitive and not entirely wrong. More words do mean more coverage. But the coverage return per new word drops sharply past 2,000. Adding the next 1,000 words gets you roughly 4 to 5 percent more coverage of spoken text - compared to the 80 percent the first 2,000 provided.
The question intermediate learners rarely ask is: what would that same study time produce if spent deepening the 2,000 words already known?
What Research on Vocabulary Depth Shows
Paul Nation, whose frequency analyses established the 2,000-word coverage threshold, also documented what happens past it. The per-word coverage return flattens so sharply that the strategy driving beginner gains - add more words - becomes increasingly inefficient at the intermediate stage. Nation's (2001) data shows the first 1,000 words cover roughly 73 percent of spoken text; each subsequent tier of 1,000 words adds progressively less. The first 2,000 words do the heavy lifting. Everything after that is marginal.
The coverage argument is about where breadth hits its ceiling. The depth argument is about what grows from there.
In a 2002 study published in Reading Research Quarterly, David Qian tested 217 university language students on two distinct measures of vocabulary knowledge. Breadth measured how many words they could recognize. Depth measured how well they knew those words - specifically their collocates (words that commonly appear alongside them), their contextual restrictions, and their related word forms. Qian found that vocabulary depth was a significantly stronger predictor of reading comprehension scores than breadth alone, and the advantage was most pronounced in learners whose breadth was already past the intermediate threshold.
Knowing more words did not predict comprehension as reliably as knowing the words well.
A third body of evidence comes from Batia Laufer and Jan Hulstijn's involvement load hypothesis (2001). Their research showed that words processed through high-involvement tasks - producing the word in a new sentence, identifying its collocates, working with its verb patterns - were retained significantly better than words processed at low involvement. Deepening a known word requires more cognitive operations than adding a new word at shallow depth. That processing cost is not inefficiency; it is what produces durable retention.
Does Breadth Still Matter at the Intermediate Stage?
The strongest objection is direct: if you do not know a word, depth of knowledge about other words will not help you understand it. A learner reading a news article will encounter words that no amount of depth work on existing vocabulary can substitute for.
This is correct. The argument for depth is not that breadth becomes worthless.
The vocabulary depth argument applies specifically to the intermediate stage, not the beginner stage. Below 2,000 words, breadth is the priority - every new word produces large coverage gains. Past 2,000 words, the ratio changes. The research supports shifting a meaningful portion of study time toward depth while continuing breadth at a slower rate. Not a replacement, but a rebalancing.
Most intermediate learners spend close to 100 percent of vocabulary study time on breadth - new flashcards, new themed lists, new unknown words from input. The research does not say to stop. It says the balance that worked at beginner level is no longer optimal. Spending even 30 to 40 percent of vocabulary study time deepening known words - learning that a verb takes a different preposition depending on context, or that a noun carries specific collocates in formal versus spoken registers - produces measurable gains that the marginal new word does not.
How Einlang Fits Learners Who Study This Way
Learners who already think carefully about how they build vocabulary - not just how many words they accumulate - tend to find that Einlang fits how they already work. The chapter-based structure means review sessions return repeatedly to the same core vocabulary from your textbook, building multiple associations with the same items rather than continuously loading new ones. That structure naturally creates the conditions for depth: the same words encountered across different exercise types and different sessions.
For the groundwork that makes depth work possible - prioritizing high-frequency vocabulary first and encoding words in sentence context rather than isolated pairs - those posts cover what to do before the intermediate rebalancing kicks in.
How to Shift Study Time Toward Depth
Three changes reflect what the research on vocabulary depth shows:
- Stop using word count as your primary progress metric. Track whether you can produce a word in a new sentence or recognize it in an unfamiliar structure - not just whether it appears in your deck.
- When you encounter a word you already know used in an unexpected way, study that usage. The collocation, the preposition it takes, the register it appears in - that is depth, and it produces stronger retention than a new word studied once.
- When scheduling vocabulary review sessions, use some of that time to produce known words in new sentences rather than only confirming recognition. Recognition and production are different knowledge states. Depth lives on the production side.
The intermediate plateau is not primarily a breadth problem. Most learners who stop progressing have enough words to work with. What they have built is wide and shallow - a large number of words known at recognition depth, few of them known well enough to produce fluently or parse in unfamiliar contexts. The research from Nation, Qian, and Laufer does not say new words are worthless. It says the strategy that got you to 2,000 words will not take you where the next stage requires.