2026-04-22 · 2 min read
Tones aren't impossible
If you've ever tried to learn Mandarin, you've heard the panicked claim that tones are the hard part. They're not impossible — they're just usually taught wrong.
Here is the typical flow in a well-meaning app:
- App shows you
māwith audio. - App asks "which tone?" with four buttons.
- You tap the right answer two seconds after hearing it.
- App says "Great!" and moves on.
You haven't learned the tone. You've echoed it. The prompt was still in your working memory when the app graded you, so the grading proved nothing.
What we do instead
When you encounter a new word in Lepochi, we play the native audio and show the pinyin with tone marks. You repeat it, and the card moves on. We don't quiz you yet.
Two or three cards later — long enough that the audio has cleared from your auditory loop — we come back and ask. No tone marks this time. Just the hanzi or the meaning, and you have to produce the tone from memory.
That gap is everything. It's the difference between I just heard it and I remember it. The first feels easy and teaches nothing; the second feels uncomfortable and teaches everything.
The harsh-grading bit
If you produce the wrong tone, we mark it wrong. Not "close enough." Not "you got the consonant right." Wrong. The word goes back into the front of the queue and you'll see it again before the session ends.
That sounds mean. It isn't. Linguists have measured this — students who get told they're wrong learn the correct contour roughly 40% faster than students who get partial credit. The sting is the lesson.
Practice modes that target tones specifically
- Tone Trainer — hear the word, pick the contour without the pinyin to lean on.
- Tone Perception — hear two near-identical words, pick which one matches the meaning.
- Dictation — hear the full word, type pinyin with the tone numbers. No partial credit.
Tones aren't the hard part. Pretending you've learned them when you haven't is the hard part. We just stop pretending.