Spaced Repetition for Language Learning: The Science Explained
Spaced repetition is the most research-backed method for vocabulary retention ever studied. Anki built a following of millions on it. Duolingo quietly runs it under the hood. Here's how it actually works — and why it's only half the equation.
The Forgetting Curve: Why You Forget So Fast
In the 1880s, German psychologist Hermann Ebbinghaus memorized thousands of nonsense syllables and tested himself at intervals to map exactly how memory decays over time. His discovery — the forgetting curve — is one of the most replicated findings in all of psychology.
The core finding: without any reinforcement, you forget approximately 50% of new information within 24 hours and up to 90% within a week. This is not a failure of effort or intelligence — it is how human memory is engineered. Short-term memory is designed to discard what isn't used.
Retention without review
Based on Ebbinghaus (1885). Modern estimates vary by material type; language vocabulary tends to decay faster than semantic knowledge.
How Spaced Repetition Fights the Forgetting Curve
Spaced repetition works by scheduling each review at precisely the moment your memory of that item is about to decay below a target threshold — typically 90% recall probability. Each successful review resets and extends the forgetting curve: the interval before the next review grows longer with each correct answer.
The result: a word you first learned today might need review tomorrow, then in 3 days, then in a week, then in 3 weeks. After 4–5 successful reviews, it enters long-term memory and may only need a review every few months to stay locked in permanently.
The 4 Core Review Intervals
These are approximate intervals used by most SRS systems for a new vocabulary item assuming correct answers at each stage:
First encounter — study the word in context
First review — you've already forgotten 42% without it
Second review — interval lengthens after a correct answer
Third review — long-term memory consolidation begins
Fourth review — word is now in long-term memory
SRS vs. Regular Flashcards: What's the Difference?
Most learners start with simple flashcard apps or handwritten cards. These work — but they're leaving a lot of efficiency on the table.
| Dimension | Spaced Repetition (SRS) | Regular Flashcards |
|---|---|---|
| Review timing | Scheduled at the optimal moment — just before you'd forget | Random or fixed (e.g. every day), ignoring your personal memory curve |
| Personalization | Adapts to your response speed and accuracy per word | One-size-fits-all — easy and hard words reviewed at the same rate |
| Efficiency | 15–30% fewer reviews needed for the same retention level | Over-reviews easy words, under-reviews hard ones |
| Long-term retention | Words enter permanent memory through spaced consolidation | High retention during active review; rapid forgetting if you stop |
| Review debt | Well-designed apps cap daily reviews and handle catch-up automatically | No built-in management — skipping a week means reviewing everything again |
Why Context Sentences Beat Isolated Words in SRS
The most advanced Anki users and every serious SRS researcher agree on one thing: studying a word in isolation (character → definition) produces shallow knowledge. You can ace every card in your deck and still blank in a real conversation.
The reason is encoding depth. Your brain stores memories as a web of associations. Isolated vocab gives the word one hook to hang on — its definition. A sentence card gives it dozens: grammar context, adjacent words, a situation, sometimes an emotion.
Isolated word card
复杂
→ "complex / complicated"
Forgotten within days. No context, no grammar, no collocations.
Sentence card (SRS best practice)
这个问题很复杂。
This problem is very complex.
Grammar pattern, measure word, usage context — all encoded in one review.
Where SRS Alone Stops Working
SRS is the best vocabulary retention tool ever built. It is also not a complete language learning system. Serious learners who rely on it exclusively hit predictable walls:
No speaking practice
SRS trains recognition and recall, not production. You can know every word in a deck and still freeze in a real conversation.
No pronunciation correction
Apps cannot hear your tones the way a teacher can. Reviewing Chinese vocabulary without pronunciation feedback builds bad habits silently.
No grammar in context
SRS is word-level, not sentence-level. It cannot teach you when to use 了 vs. 过, or why a sentence sounds unnatural even when every word is correct.
No real-time error correction
Mistakes in SRS just change the scheduling interval. A teacher stops you the moment you make an error and explains why — that's how errors become non-recurring.
No cultural and pragmatic context
SRS can teach you 您 and 你 as vocabulary. A teacher tells you when to actually use one over the other — which is a social rule, not a grammar rule.
The Fastest Path: SRS + Expert Tutor Sessions
The research is clear: learners who combine spaced repetition vocabulary practice with regular live sessions retain more, speak sooner, and plateau less. SRS loads the vocabulary. A tutor activates it — turning recognition into production, correcting pronunciation in real time, and providing the grammar scaffolding that no app can build.
Think of it as: SRS 20 minutes a day for vocabulary, one live session per week for everything SRS can't do. That combination outperforms either approach alone.