Skip to content

Chapter 20 of 24

83% through the course

Multiple Third Tones & Chunking

How multiple third tones work in Mandarin: why 我很好 isn't all rises, how chunking by meaning decides where sandhi applies, and why you shouldn't compute tones live.

The 3+3 sandhi rule is simple with two syllables, but Mandarin loves to stack third tones — 我很好 (wǒ hěn hǎo) — “I’m fine” has three in a row. The question of how multiple third tones behave trips up almost every learner, and the answer isn’t “turn them all into rises.” It’s chunking: you group the syllables by meaning first, then apply sandhi inside each group.

Why you can’t just chain the rule

If you applied 3 + 3 → 2 + 3 blindly across 我很好, you’d get tangled deciding which syllables pair up. Does 我很 change, or 很好? They can’t both be the “first” of a pair. Real speakers resolve this not with arithmetic but with rhythm: the sentence breaks into a small unit and a remainder, and the rule operates within those units. The grouping comes from meaning, not from counting left to right.

Group by meaning, then apply sandhi

The reliable mental model has two steps:

  1. Chunk the string into meaningful groups (a word or tight phrase).
  2. Apply 3+3 sandhi inside each chunk, treating chunk boundaries as resets.

Take 我很好. The natural split is 我 (wǒ) + 很好 (hěn hǎo) — “I” plus “(am) very good.” Inside the chunk 很好, sandhi fires: hěn hǎo → hén hǎo. The leading 我 then attaches loosely in front:

  • 我很好 (wǒ hěn hǎo) → wǒ hén hǎo — “I’m fine”

Only the chunk 很好 changed. 我 keeps its low third-tone shape (often a half-third) rather than turning into a rise. Compare a four-syllable example:

  • 我也很好 (wǒ yě hěn hǎo) → wǒ yě hén hǎo — “I’m fine too”

Here the meaning groups as 我也 + 很好. Inside 很好, sandhi applies; 我也 stays low (with the option of a light rise on 我 at faster speeds). Notice you never get a long run of rises — the chunks keep the pattern readable.

The same syllables, different chunks

Because grouping is driven by meaning, the same sequence of third tones can come out differently depending on what the sentence means. A textbook favorite:

  • 小老虎 (xiǎo lǎo hǔ) grouped as 小 + 老虎 “little tiger” → the 老虎 chunk sandhis to lǎo → láo, giving roughly xiǎo láo hǔ.
  • 老李买 (lǎo Lǐ mǎi) grouped as 老李 + 买 “Old Li buys” → the 老李 chunk sandhis to láo Lǐ, with 买 as the low-dipping ending.

You don’t need to master every such case. The point is that chunking, not raw syllable order, decides the surface tones — which is exactly why mechanical left-to-right computation fails.

How to find the chunks

For a beginner, the chunks line up with the units you already learn vocabulary in. A few reliable cues:

  • Two-syllable words are one chunk. 老虎, 很好, 可以 each travel together, so sandhi happens inside them.
  • Subject and predicate often split. In 我很好, the pronoun 我 is its own piece and the comment 很好 is another.
  • Short grammar words lean on their neighbor. 也 in 我也很好 attaches to 我, forming 我也 before the 很好 chunk.

In practice you rarely deliberate over this — you absorb the groupings by hearing the phrase, then reuse them. The grammar of the sentence is doing the chunking for you; sandhi just rides along inside each piece.

Don’t compute it live

This is the practical takeaway for connected speech: do not try to recalculate sandhi while you talk. Native speakers don’t, and you can’t do it fast enough to sound natural. Instead, learn frequent third-tone phrases as single fixed melodies — say them as one shape, the way you’d hum a short tune:

  • 你好 (nǐ hǎo) → ní hǎo “hello”
  • 我很好 (wǒ hěn hǎo) → wǒ hén hǎo “I’m fine”
  • 可以 (kě yǐ) → ké yǐ “may; can”
  • 请你给我 (qǐng nǐ gěi wǒ) → qíng ní géi wǒ “please give me” (a longer set phrase, learned whole)

When the building blocks are pre-tuned, longer strings assemble themselves, and chunk boundaries fall where the meaning already tells you to pause.

For the underlying logic of the change and the half-third shape that low syllables take on, revisit the third-tone sandhi rule and the deeper dive in /blog/third-tone-sandhi/.

Next, we zoom out from single rules to how all of this plays together in real sentences, in tones in connected speech.