Skip to content

Learning Materials

Character Formation

liùshū — "the six writings"

Every Chinese character is built, not drawn at random — each 汉字 (hànzì) is assembled from a small set of strategies for turning sound and meaning into strokes. The classic framework for these strategies is the 六书 (liùshū, "six writings"), systematized by the scholar Xu Shen (许慎) in his ~100 CE dictionary 说文解字 (Shuōwén Jiězì). The first four describe how characters are formed, while the last two describe how existing characters were repurposed.

May 26, 2026

Each slice is a method's rough share of all modern characters — click one to jump to it. Note how 形声 (phono-semantic) dwarfs the rest: master that one pattern and most of the script falls into place.

1

象形 xiàngxíng

Pictographs

~4%of hanzi

Characters that began as simplified pictures of physical objects.

A pictograph traces the outline of a concrete thing until the drawing stiffens into standardized strokes. Over three thousand years the curves were straightened for the brush, so the picture is often hidden — but once you know the original image, the modern shape snaps back into focus. These characters are few in number but enormous in influence, because they became the building blocks (radicals) for the thousands of compound characters that followed.

rì — sunOriginally a round sun with a dot in the center; the circle was squared off into the modern box with a line inside.
shān — mountainThree peaks rising from a baseline — the central tallest peak flanked by two lower ones is still clearly visible.
mù — treeA trunk with branches reaching up and roots spreading down at the bottom — the strokes still fork like a living tree.
2

指事 zhǐshì

Simple Ideographs

~1%of hanzi

Characters that point at an abstract idea using a symbolic mark rather than a picture.

When there is no object to draw — like "up," "down," or "a number" — the indicative method adds a diagnostic stroke or dot to show position or quantity. Sometimes the mark is placed on an existing pictograph to spotlight a specific part of it, as in 本 (a tree with a stroke at its base to mean "root"). This is the rarest of the four formation types, but it gives us some of the most fundamental words in the language.

shàng — up / aboveA short mark sitting above a horizontal baseline literally points upward to indicate "above."
běn — root / originThe pictograph 木 (tree) plus an extra stroke at the bottom highlights the roots, hence "root" and by extension "origin."
rèn — bladeThe character 刀 (knife) with a single dot marking exactly where the cutting edge is.
3

会意 huìyì

Compound Ideographs

~13%of hanzi

Characters that fuse two or more meaning-bearing elements into a new combined idea.

Here whole pictographs are stacked or placed side by side so that their meanings add up to a third, often poetic, idea — sound is not involved at all. The logic can be wonderfully visual: a person leaning against a tree means "to rest." Because the connection is conceptual rather than phonetic, these characters are a favorite of teachers, though learners should be careful — not every multi-part character is a true 会意, and many that look like one are actually phono-semantic.

míng — bright日 (sun) + 月 (moon) — the two brightest objects in the sky together mean "bright."
xiū — to rest人/亻 (person) + 木 (tree) — a person resting against a tree.
sēn — forestThree 木 (trees) stacked together; two trees (林) make a grove, three make a dense forest.
4

形声 xíngshēng

Phono-semantic Compounds

~80–90%of hanzi

Characters built from one component that hints at meaning and another that hints at the sound.

This is the single most important pattern in the entire writing system: roughly 80–90% of all Chinese characters are formed this way, so mastering it unlocks the logic behind the vast majority of hanzi. One part — the semantic radical — tells you the general category of meaning, while the other part — the phonetic — tells you (approximately) how it sounds. Look at 妈 mā "mother": 女 (woman) gives the meaning, while 马 mǎ supplies the sound; or watch one phonetic 青 qīng generate a whole family — 清 (clear water), 请 (please), 晴 (clear sky), 情 (feeling). Sound shifts over the centuries mean the phonetic is only a clue, not a guarantee, but learning to split a character into "meaning-side" and "sound-side" is the single biggest leap in reading Chinese.

mā — mother女 (woman) signals the meaning; 马 mǎ supplies the sound — woman + the mǎ sound = "mother."
hé — river氵 (the water radical) gives the meaning; 可 kě hints at the pronunciation hé.
qǐng — to ask / please讠 (speech radical) marks it as a word of speech; 青 qīng supplies the sound — one of a whole 青 phonetic family (清/晴/情).
5

转注 zhuǎnzhù

Transfer Characters

<1%of hanzi

A rare, debated category of characters that mutually explain or extend each other's meaning.

This is the murkiest of the six categories, and scholars have argued about its exact definition for nearly two thousand years. The usual reading is that two characters of similar shape and related meaning came to define one another, often drifting apart in sound and usage over time. Because so few clear examples exist, learners can safely treat 转注 as a historical footnote rather than a practical pattern to master.

kǎo — aged / to testHistorically paired with 老 (old); the two share a root form and once explained each other's meaning of "elderly."
lǎo — oldThe classic partner of 考 — a picture of a long-haired elder; the pair is the textbook example of 转注.
dǐng — top / peakSometimes cited alongside 颠 (summit) as related forms that extended each other's sense of "highest point."
6

假借 jiǎjiè

Phonetic Loan Characters

<1%of hanzi

Existing characters borrowed to write a different, same-sounding word — a rebus strategy.

When a spoken word had no character of its own, scribes simply borrowed an existing character that sounded the same, ignoring its original meaning — much like writing "4" for "for." Over time the borrowed sense often took over completely, and a brand-new character had to be invented for the original meaning. This rebus trick was crucial in the early growth of the script, because it let the language write abstract words long before purpose-built characters caught up.

lái — to comeOriginally a pictograph of "wheat"; it was borrowed for the homophonous verb "to come" and kept that meaning.
běi — northOriginally showed two people back-to-back ("back"); borrowed for the same-sounding word "north," while 背 took over "back."
zì — selfOriginally a picture of a nose (Chinese speakers still point to their nose for "me"); borrowed for the word "self."

The first four categories (象形, 指事, 会意, 形声) describe how characters are formed; the last two (转注, 假借) describe how existing characters were repurposed. In practice, 形声 phono-semantic compounds account for the overwhelming majority of all characters — so if you learn to split a character into a meaning side and a sound side, you've cracked the logic of most of the writing system.