Sound Components
Sound Components
A sound component is the part of a Chinese character that points to a range of syllables. It does not give the exact modern reading. Together with the meaning component, it lets a reader get a meaning domain and a syllable range from an unfamiliar character before any dictionary lookup.
Two common beliefs about sound components are wrong. The first is that the part gives today’s pronunciation. The second is that the matches are too broken to use.
Core takeaways
- A sound component narrows an unfamiliar character to a small set of possible syllables. Only a dictionary lookup gives the exact reading.
- The component stores the pronunciation the word had when the character was made. For most characters that is Old Chinese, which had no tones, so the component almost never predicts the tone.
- The working method has three steps: predict the range, look up the exact reading, learn the tone as a separate fact.
- One correspondence to learn first: the initials zh-, ch-, and sh- regularly relate to d- and t-.
- Some parts of a character are not cues. 口 in 尚 evens the shape and 口 in 周 is a distinguishing mark. Predictions from such parts are wrong.
- The sound hint is about as accurate as English spelling. Both are imperfect and both are still useful.
Where sound components came from
A drawing cannot show an abstract word. Two methods were used to write such words. The first is an indicating mark: 本 is 木 plus a stroke at the base. The second is the rebus: an existing character was borrowed for another word that sounded like it.
北 originally drew two people back-to-back and meant a person’s back. The word “north” sounded similar and could not be drawn, so 北 was borrowed for it. Later 肉/月 was placed under 北 to make 背 for “back,” and 北 kept the borrowed meaning “north.” Pairing a character with a word required only similar sounds. Sound components arose from this borrowing.
Why the matches are loose
The sound component stores the pronunciation the word had at the moment the character was made. For most characters that language is Old Chinese: the spoken language of the coinage period, with no modern-style tones. Over the following centuries the sounds changed, so the component and the character often no longer sound alike. The result is about as accurate as English spelling. It is imperfect. It is still needed.
Old Chinese had no tones, so tone rarely carries from component to character. 北 is běi and 背 is bèi. Characters coined later, in Middle Chinese, match their components more closely, and the tone matches too. In general the component almost never predicts the tone.
Sound change made small differences large. 監 *kram, 籃 *k.ram, and 藍 *g.ram became jiān, lán, and lán. The original spread was no wider than English here, ear, and year. The asterisk marks a reconstruction. A reconstruction explains why two readings are linked. Reconstructions never go on review cards. Meaning Components uses 監/藍 as an example of an opaque pair. The same characters also show how a sound chain works.
The system was designed to approximate. A native reader resolves inexact spelling from context. A semantic component puts that context inside the script. Learners notice the gaps more than native readers do, because they do not yet have the spoken base the script was built to assume.
Reading one character
The working cycle has three steps: predict the range, look up the exact reading, learn the tone as a separate fact. Commit to a guess before opening the dictionary. The IME Method trains the same predict-then-verify habit. A sound component applies that habit to a single character.
棠 shows the cycle. 木 gives the meaning domain: tree or wood. 尚 gives a syllable range: shang, chang, zhang, tang, dang. The dictionary settles on táng. The tone could not be guessed. Five candidates is a real narrowing, and the reader still has to look up which one is right.
The initials zh-, ch-, and sh- regularly relate to d- and t-. With that pattern, 黨, 堂, and 躺, which all take 尚, read as one family. 店 taking 占 follows the same pattern. A new character that shares a sound component gets the same syllable range. Sound Series extends this to whole families and holds the other related-initial groups. It also says which correspondences to memorize first: the seven groups.
Borrowed meanings and double roles
A loan sense attaches to a character by sound alone. 北 “north,” 良, and 古 are borrowings. Tag them as borrowings rather than trying to read the meaning from the drawing. Meaning Trees and Original Meanings covers loan senses and the successor characters they leave behind.
One component can play two roles. Inside 背, 北 supplies the meaning, a back, and also the sound. Surface vs Deep Structure covers dual-function cases and where to stop dividing a character into parts.
Original senses moved to successor characters. 寺 “to grasp” is now 持. 兌 “to be happy” is now 悅. 各 kept “to arrive” only in the older layer. Meaning Trees and Original Meanings covers those changes.
Using the explanation to remember
The explanation is what makes a pair memorable. jiān beside lán looks arbitrary. The *kram chain links the two readings. The reconstruction stays off the card.
Five ways the method fails
- Inferring tone from the component. Treat every tone as a separate fact.
- Demanding exact readings. Structure narrows the possibilities. Only a lookup confirms the reading.
- Expecting every character in a series to read the same way. Divergence in initial, final, and tone is common. Sound Series describes that divergence precisely.
- Reading every part as a cue. 口 in 尚 evens the shape. 口 in 周 is a distinguishing mark. Predictions from those parts give wrong results. Empty Components lists the parts to skip.
- Dismissing the method because it is inaccurate. Neither sound components nor English spelling are complete. Non-native speakers read “iron” as eye-run. The method is imperfect, and it still narrows the possible readings.
What a sound component predicts
A sound component predicts a handful of syllables. It never predicts the tone. It never predicts anything from an empty stroke. A lookup confirms the reading. The system assumes the reader already knows the spoken word, and that knowledge comes from hours of listening. The written form then attaches to a word the reader already knows by ear.
How Chinese Characters Work is the hub page. It states the domain-and-range result in one line. Sound Series is the page for families of characters that share a component.
How to practice this
- Take an unfamiliar character and split it into its meaning part and its sound part. The meaning part gives a domain, such as 木 for tree or wood.
- Before opening a dictionary, write down the syllables the sound part could give. For 尚 the range is shang, chang, zhang, tang, dang. Notice that you get a short list of syllables.
- Look up the exact reading and compare it with your list. For 棠 the dictionary gives táng. Notice whether your list contained it.
- Learn the tone as a separate fact. 北 is běi and 背 is bèi. Notice that the component did not tell you the tone.
- Check that the part you predicted from is a real cue. 口 in 尚 only evens the shape. If the part is on the Empty Components list, skip it.
- When two readings in a series look unrelated, look at the reconstruction. 監 *kram and 藍 *g.ram explain why jiān and lán are linked. Keep the reconstruction off your review card.
Related pages
- How Chinese Characters Work: the hub page; it states the domain-and-range result in one line
- Sound Series: families and the other related-initial groups
- Meaning Components: the disambiguating partner; 監/藍 as an opaque pair
- Surface vs Deep Structure: dual-function cases; where to stop dividing a character into parts
- Empty Components: parts that prediction must skip
Open questions
- There is no marker yet for Middle-Chinese coinages, whose tones correlate more closely with their components.
- Where predict-then-verify belongs in review is a question for The IME Method. It does not need a second cycle.
Sources
- DeFrancis, J. (1984). The Chinese Language: Fact and Fantasy. University of Hawaii Press. Most characters are a meaning-hint plus a sound-hint; the rebus is the historical method used when a word could not be drawn.
- Haudricourt, A.-G.; Baxter, W. H., & Sagart, L. (2014). Old Chinese lacked Middle-Chinese-style tones; later tones arise from lost finals. Reconstructions explain a link and do not belong on cards.
- Karlgren, B.; Baxter–Sagart phonetic series. zh-/ch-/sh- related historically to d-/t/. The other groups are on Sound Series.