Mandarin Tone 2 vs Tone 3: What Actually Distinguishes Them
August 18, 2026
Tone 2 (rising, as in má, hemp) and Tone 3 (dipping, as in mǎ, horse) are the pair most learners confuse longest — and not entirely because of untrained ears. In fast, connected speech, Tone 3 frequently drops its distinctive rising tail and becomes nearly indistinguishable from a low, flat pitch. Here is what is actually happening.
In isolation, they are genuinely different shapes
Said alone and carefully, Tone 2 is a clean rise from mid to high pitch — the same contour English uses at the end of a question, which is exactly why English speakers usually pick it up fastest. Tone 3, said alone, is a dip: it starts around mid pitch, falls to the bottom of the speaker's range, and then often rises back up at the very end. That final rise is the part that causes most of the confusion.
Why they blur together in real speech
Mandarin has a well-documented rule called third-tone sandhi: when a Tone 3 syllable is followed by another Tone 3 syllable, the first one is pronounced as Tone 2 instead. And even before syllables of other tones, a Tone 3 in connected speech typically loses its final rise entirely, staying low and flat — linguists call this the "half third tone." So the textbook dipping-then-rising shape you learn in isolation is actually the exception in real conversation, not the rule. Most of the time, what you're hearing from a Tone 3 syllable is just a low tone with no rise at all — which is why it can sound deceptively close to a mumbled, undershoot Tone 2.
How to actually tell them apart when listening
- Listen for where the pitch starts, not just where it ends: Tone 2 starts low-to-mid and climbs the whole way; the "half" Tone 3 starts mid and just stays low — it never climbs.
- In minimal pairs like má (麻, hemp) vs. mǎ (马, horse), say both back to back at your own natural pitch rather than a recording's — the contrast is easier to hear against your own register than a stranger's.
- Don't rely on the isolated citation form for words you'll actually use in sentences — practice the sandhi'd, connected-speech version, since that's the one you'll hear (and need to produce) in real conversation.
This is a case where pitch-matched practice earns its keep specifically: hearing the contrast in your own register, rather than transposing someone else's voice first, makes the genuinely subtle Tone 2/Tone 3 distinction easier to lock in.
Practice Mandarin in your own voice
Hear any sentence resampled to your own pitch, then get scored on your attempt.
Try Mandarin practiceRead the Mandarin pronunciation guide →