How Chinese Characters Work
Cluster hub: writing records speech, so each character connects a written form to a sound and a meaning. Classifying each component by the job it does in each character lets a learner predict a meaning domain and a sound range, so the script is not arbitrary. Reading order, a one-line summary of each page, and where the cluster belongs in the three-mode immersion system.
How Chinese Characters Work
Speech comes first. A character is the written form of a spoken word. That word already had a sound and a meaning. The writing still has both. Each part of the writing has a job: some parts picture a thing, some add a sense, some point to a range of syllables, and some do nothing. Once you can name the job of each part, the thousands of characters are not arbitrary drawings. You can predict a meaning domain and a sound range for a character you have never seen. You can rebuild a character whose form you have only half forgotten.
What the parts do
A written word is one item with three attributes: form, sound, and meaning. Each pair of attributes is connected by a link, and each link gets stronger or weaker independently of the other two. Three Attributes of a Character is the first page of the cluster and covers this model: the learner stores each character as one item and trains each of its three links. The reason is that the yes-or-no question “do I know this character?” does not show which of the three links has failed.
The written form can be divided into parts. A functional component is a part that does a job in the character it appears in: it pictures something, it contributes an abstract sense, it indicates a range of syllables, it does two of those at once, or it does nothing. Picture, sense, and sound are the roles used in the ordinary analysis of the script. A part that does nothing is leftover shape that gives no information about the word, and this cluster adds it as a fourth role. For that reason Empty Components comes last in the reading order: a part can be classified as empty only after it has been checked for picture, sense, and sound and none of them applies.
A part’s job is decided separately in each character it appears in. The same shape can do a different job in a different character. 立 contributes sense in 端, sound in 拉, and nothing in 音. The strokes of 立 are the same in all three characters, and each of the three characters has a different history.
When the parts of a character are identified by the job they do in that character, the thousands of characters are no longer arbitrary pictures. A learner can predict a meaning domain and a syllable range. The learner can use those two constraints to rebuild a form they have half forgotten. The learner can also discard most folk glosses, because what they explain can be derived from the parts. The prediction gives a meaning domain and a syllable range. It does not give the reading of the character. Sound Components covers the single-character cycle that produces the syllable range. Form Components covers the parts that contribute a picture. The origin of such a part is a depiction, even when the modern strokes no longer look like the thing depicted. Meaning Components covers the parts that contribute a sense. That sense is an abstract extension of a concrete original meaning.
Rote copying uses the same hours on individual strokes, and individual strokes contain none of that structure. Rules of Effective Memorization explains why dividing characters into functional components is a better use of study time than copying strokes.
The detailed explanation of each role, and every worked example other than the 立 example above, are on the other pages of the cluster listed below.
Reading order and where this sits
Read the pages of the cluster in this order. Each page uses a classification that the pages before it teach you how to make.
Read Three Attributes of a Character first. It covers one written word and its three links, each of which can fail separately. Surface vs Deep Structure gives the rule for when to stop dividing a character: stop at the level of functional components, and do not divide the character into every stroke. Next, read the pages on the three working roles in any order: form, meaning, and sound. Sound Series extends the study of sound components from single characters to families and related-initial groups. Empty Components comes after those pages, because a part can be classified as leftover only after it has been checked for the other three jobs. Meaning Trees and Original Meanings covers how the senses of characters developed from their original meanings, and compares the character-tree with the component-tree. The practice pages come last. The IME Method recovers a form you cannot recall: you list candidates that fit the sound side, list candidates that fit the meaning side, and cross the two lists. The IME Method covers that crossing of the two lists. Pipelining means learning the spoken word first, in both directions, and adding the written form afterward. The Pipelining Strategy covers that sequence. The whole cluster is based on Rules of Effective Memorization.
The same eleven pages, one line each:
- Three Attributes of a Character: one written word, three links, each of which can fail on its own
- Surface vs Deep Structure: stop dividing at functional components, not at every stroke
- Form Components: parts whose job is a picture
- Meaning Components: parts whose job is an extended sense
- Sound Components: parts that point to a syllable range
- Sound Series: families and related-initial groups
- Empty Components: marks, and leftover shapes from corrupted forms
- Meaning Trees and Original Meanings: the character-tree next to the component-tree
- The IME Method: recover a forgotten form by crossing sound-side candidates with meaning-side candidates
- The Pipelining Strategy: learn the spoken word first, then add the written form
- Rules of Effective Memorization: why dividing characters by function beats copying strokes
Character Primer was the first page on this subject. This cluster replaces that short outline with a model a learner can use, and it keeps the same study materials: real words, audio, popup dictionaries, and traditional script.
In the three-mode immersion system, this cluster belongs to Preparation. Character study prepares the learner for immersion, and immersion remains the main activity. Refold Language Learning System is that three-mode system. Pipelining connects written forms to words the learner already knows from listening. The guess-and-verify step of the IME method requires a large amount of real reading. Both methods assume that immersion is already providing spoken vocabulary and text.
Once a learner can identify the job of each part, the thousands of characters are no longer arbitrary. This gives the learner a meaning domain, a syllable range, and a way to recover a form they have partly forgotten. It does not give an exact modern reading. It does not replace the listening and reading the learner is already doing. Character study is for words the learner already knows from listening.
Related pages
- Three Attributes of a Character: the first page of the cluster; one written word and its three links, which fail separately
- Surface vs Deep Structure: the rule for when to stop dividing a character; stop at the level of functional components
- Form Components: parts that contribute a picture; the origin of such a part is a depiction
- Meaning Components: parts that contribute a sense; the sense is an abstract extension of a concrete original meaning
- Sound Components: parts that indicate a syllable range; the cycle of predicting a sound and then looking it up, one character at a time
Sources
- DeFrancis, J. (1984). The Chinese Language: Fact and Fantasy. University of Hawaii Press. Especially “The Ideographic Myth”: characters record words and morphemes of speech, and do not record ideas separate from speech.
- Taft, M., & Zhu, X. (1997). Submorphemic processing in reading Chinese. Journal of Experimental Psychology: Learning, Memory, and Cognition, 23(3), 761–775. Readers process components as well as whole characters.
- Handel, Z. (2019). Sinography: The Borrowing and Adaptation of the Chinese Script. Brill. Treats characters as a writing system for spoken language. It does not treat them as a separate ideographic code that runs parallel to speech.