The SoundsBy Pinyinize

The Retroflex Sounds: zh, ch, sh, r

Four sounds, one tongue position. Master zh, ch, sh, and r together by learning the single curl that connects them and the airflow that tells them apart.


zh, ch, sh, and r share one curled-tongue position but differ in airflow and voicing.
The four retroflex sounds share one tongue position and differ only in airflow and voicing.

Mandarin has one group of sounds that exists in few other languages and trips up nearly every learner: the retroflexes, zh, ch, sh, and r. The good news is that they are not four separate problems. They are four versions of a single skill. Learn one tongue position and you are most of the way to all four.

One Curl, Four Sounds

Every retroflex starts the same way. Curl the tip of your tongue upward and backward, so its underside approaches the ridge just behind your upper teeth, and push your lips slightly forward into a soft round shape. Hold that shape. From here only two things change: whether the air is blocked or flows freely, and whether your voice is switched on.

  • zh blocks the air, then releases it gently, with no puff. (zhōng, middle)
  • ch blocks the air the same way, then releases it with a strong burst of breath. (chá, tea)
  • sh never fully blocks the air; it hisses out through a narrow gap. (shū, book)
  • r holds the same gap as sh, but adds voice, a buzz from the throat. (rén, person)

Read that list from top to bottom and the logic appears: two bursts, then two streams, and at the very end the voice comes on.

Building the Four in Order

The easiest way to learn the set is to build it step by step rather than memorize four separate sounds:

  1. Get the curl. Tongue tip up and back, lips softly rounded. This single position is the foundation of all four.
  2. Add a gentle release for zh, then the same release with a hard puff of air for ch. Hold a hand to your mouth and you should feel air only on ch.
  3. Open a small gap and let the air stream out for sh.
  4. Keep that exact gap and switch on your voice. Put your fingers on your throat; when it buzzes, sh has become r.

Not the Dentals, Not the Palatals

Two other groups get mistaken for the retroflexes, and the giveaway is the tongue and the lips. The dental sounds z, c, s push the tongue tip forward against the teeth, with the lips spread. The palatal sounds j, q, x keep the tongue flat and high behind a wide smile. Only the retroflexes curl the tongue back with rounded lips. If your lips are spread, you are not making a retroflex.

One quirk to know in advance: after zh, ch, sh, and r, the letter i is not the bright "ee" it looks like. It becomes a tense, buzzing vowel that keeps the tongue curled. That single letter behaves in three different ways across Pinyin, a trap we untangle in why "i" sounds different in chī and lī.

When the Curl Feels Impossible

If the tongue position fights you at first, you are in good company; the retroflexes take the longest to feel natural. There are a couple of very common ways the curl goes wrong, each with a simple fix. We break those down, with a daily drill, in the retroflex challenge, and we tackle the trickiest member on its own in the tricky "r" sound.

Frequently asked questions

Do zh, ch, sh, and r work with every vowel?+
Not with all of them. They pair with finals like a, e, u, ai, ao, ou, an, en, and ang, and with the special buzzing i as in zhī. They never combine with ü, nor with the bright i found in a word like lī.
Can these sounds appear at the end of a syllable?+
No. Like all Mandarin initials, zh, ch, sh, and r only ever begin a syllable. The only sounds that close a Mandarin syllable are vowels and the two nasal endings, -n and -ng.
Why are they called retroflex sounds?+
Retroflex means bent back. The name describes the gesture exactly: the tip of the tongue bends up and back toward the roof of the mouth. That single movement defines the whole group and sets it apart from every other set of consonants in Mandarin.