Chinese characters are not pictures to be memorised whole. The large majority are built from a limited set of components, usually one part suggesting the meaning and another suggesting the sound, and learning to see that structure is the difference between memorising thousands of shapes and assembling them from parts you already know.
The building blocks
Strokes come first: the individual marks, made in a conventional order. Components come next: recurring assemblies of strokes such as 讠, 氵, 木 or 心, several hundred of which cover the vast majority of characters. Characters are components arranged in a square, and that square is fixed: whether a character has three strokes or twenty, it occupies the same space, which is why proportion inside the square is such a large part of what makes handwriting look right.
The arrangement itself carries information and is worth naming, because a character is not a bag of parts. Left-right structures like 明 and river-water characters, top-bottom structures like 花, enclosures like 国, and semi-enclosures like 医 are the common layouts, and getting the layout wrong is one of the most frequent handwriting errors among learners who know the components.
| Layer | Example | What it gives you |
|---|---|---|
| Stroke | 一, 丿, 丶 | The unit of writing and of order |
| Component | 讠, 氵, 木, 心 | A reusable meaning or sound hint |
| Structure | left-right, top-bottom, enclosure | How the parts sit in the square |
| Character | 明, 河, 想, 国 | The assembled unit |
Semantic and phonetic parts
Most characters are semantic-phonetic compounds: one component hints at the meaning category and another at the pronunciation.
河, meaning river, has 氵 on the left, the water component, and 可 on the right, which carries the sound. 想, to think, has 心 underneath, the heart component associated with thought and feeling, and 相 above supplying the sound.
Two cautions keep this from becoming a false promise. The meaning component gives a category rather than a definition: 氵 tells you the character concerns water or liquid, and it does not tell you which. And the phonetic component is approximate, because sounds have drifted over two thousand years, so 可 kě appears in 河 hé. Treat the sound hint as a strong clue rather than a rule.
Even approximate, the structure is transformative for writing. A character you can decompose is a short sequence of familiar parts. A character you cannot is a picture with eleven strokes and no handles.
Why this matters specifically for handwriting
Recognising a character requires seeing enough of it. Producing one requires knowing every part and where each sits, which is exactly what component knowledge supplies.
It also explains a common error pattern. Learners who memorise shapes tend to produce characters with the right parts in the wrong arrangement, or with a component that looks approximately right but is a different component. Both are structural mistakes, and both disappear once the character is stored as an assembly rather than an image.
Handwriting research in Chinese supports treating writing as its own trainable skill: a longitudinal study of handwriting fluency and spelling accuracy found the two developing together, and imaging work on Chinese handwriting connects handwriting to reading development.
Learning components deliberately
Start with the components that appear most: 氵 water, 讠 speech, 木 tree or wood, 心 heart, 手 hand, 口 mouth, 日 sun, 月 moon or flesh, 女 woman, 亻 person. Ten components unlock a large share of the characters a beginner meets.
Learn each one as a character where it exists as one. 水, 言, 木, 心, 手, 口, 日, 月, 女, 人 are all characters in their own right, and learning the standalone form makes the compressed form inside a character obvious rather than arbitrary.
Then decompose deliberately. When a character resists you, look up its parts once and write the structure in a line: “请 = 讠 speech + 青 qīng sound”. Two minutes converts a stubborn shape into a member of a family. The Unicode Han database and the Unicode FAQ on Han characters are useful references for how the system is organised.
Finally, drill the arrangement. Writing the right parts in the wrong proportion is the visible signature of component knowledge without structural practice, and it is fixed by writing large and checking the balance.
| Component | Standalone | Suggests |
|---|---|---|
| 氵 | 水 | Water, liquid |
| 讠 | 言 | Speech, language |
| 木 | 木 | Trees, wood, wooden things |
| 心 or 忄 | 心 | Feeling, thought |
| 扌 | 手 | Hand actions |
| 亻 | 人 | People, human qualities |
Reading a character you have never seen
The practical test of component knowledge is what you do with an unfamiliar character, and there is a repeatable procedure.
Find the structure first. Is it left-right, top-bottom, or an enclosure? This tells you where the boundaries are and prevents the common error of splitting a character in the wrong place.
Identify the meaning component. It is usually on the left in left-right characters, on top or underneath in top-bottom ones, and it is usually the part you recognise from other characters: 氵, 扌, 忄, 木, 讠.
Guess the category from it. Water, hand action, feeling, wood, speech. You will not get the meaning, and you will get a neighbourhood, which is often enough to follow a sentence.
Then check the remaining part against characters you know for a sound hint. If the right-hand side appears in a character you can pronounce, the new one may share a similar syllable, possibly with a different tone, which is the most common form the drift takes.
Finally, look it up rather than trusting the guess. The procedure is for building intuition and narrowing possibilities, not for replacing a dictionary, and the discipline of guessing before checking is what makes the components stick.
Doing this for even a few minutes a week changes how characters look. They stop being a wall of unfamiliar shapes and start resolving into familiar parts in unfamiliar combinations, which is roughly how a literate reader experiences them.
What component knowledge does not do
It does not make every character transparent. Some characters do not decompose usefully, some meaning components are historical rather than intuitive, and forcing a story onto a character that has none produces a mnemonic more complicated than the character.
It does not replace production practice. Knowing that 请 is 讠 plus 青 does not mean you can write it under time pressure, which comes from writing it from a prompt repeatedly. Component knowledge lowers the cost of each repetition rather than removing the need for them.
It also does not help evenly across characters. Component knowledge is transformative for the compound majority and does very little for the small set of characters that are single indivisible units or that were simplified in ways that destroyed their structure. Expect a minority of characters to remain shapes you simply memorise, and do not treat that as a failure of the approach.
And it does not fix stroke order. Order operates within and across components, so it still has to be learned per character, though it becomes far more predictable once you know the parts, since each component carries its own internal order.
{{appName}} is built around producing characters from memory with stroke order checked, which is where component knowledge turns into writing ability. It is in early access. The from-memory loop and the starter character set are the companion pieces.
Where the structure came from, briefly
A short historical note makes the system easier to trust, because the strangeness of some characters has a reason.
Chinese characters developed over roughly three thousand years, beginning with pictographs of concrete things: 日 the sun, 月 the moon, 木 a tree, 山 mountains. A small number of characters are still recognisably pictorial, and they are the ones learners meet first, which creates the misleading impression that the whole system works that way.
Compound characters came next and quickly became the majority. Combining a meaning hint with a sound hint solved the problem that most words are not picturable, and that mechanism is responsible for the vast bulk of the several thousand characters in ordinary use.
Then the sounds moved. Pronunciations that matched two thousand years ago have drifted apart at different rates in different regions, which is why phonetic components are approximate today rather than broken: they were accurate when they were chosen. Knowing this makes them easier to use as hints rather than frustrating as rules.
Simplification in the twentieth century added one more layer, reducing stroke counts for a subset of characters. Occasionally that obscured a component relationship that had been visible in the older form, which is why some simplified characters resist decomposition and their traditional counterparts do not.
A short worked decomposition
One example carried all the way through shows what the whole approach buys.
Take 谢, meaning to thank. Structure first: left-right. The left side is 讠, the compressed form of 言, speech. That places the character in the speech family, alongside 说, 话, 语 and 请, which is a useful neighbourhood to be in.
The right side is 射, which supplies the sound: 射 is shè, 谢 is xiè. Close rather than identical, which is normal, and close enough to remember.
Now the writing benefit. To produce 谢 from memory you need three things: the left component 讠, which you already write in four other characters; the right component 射, which is itself decomposable into 身 and 寸; and the arrangement, narrow left and wide right. That is three known items and a layout, rather than twelve strokes of unfamiliar line.
The same analysis applies to most of the characters you will meet, which is why the two minutes spent decomposing a stubborn character pays back so consistently.
Frequently asked questions
How are Chinese characters built from components?
Most characters are assemblies of recurring components arranged in a square, commonly one part suggesting the meaning category and another suggesting the pronunciation. 河 for river combines the water component 氵 with 可 for sound. Several hundred components cover the vast majority of characters, which is why learning components converts memorising shapes into assembling familiar parts.
What is the difference between a radical and a component?
In everyday use they overlap, and strictly the radical is the component under which a character is filed in a dictionary, while component refers to any recurring part. For a learner the distinction rarely matters: what pays is knowing the parts a character is made of and what each suggests, not which one a dictionary chose as its index.
Do phonetic components reliably tell you the pronunciation?
No, they are hints rather than rules. Sounds have drifted over two thousand years, so a phonetic part may indicate a similar syllable, a different tone or something further away, as in 可 kě appearing in 河 hé. The hint is still valuable for narrowing a guess and for making a character memorable, and it should not be trusted as a pronunciation guide.
What is the best app for learning characters by their components?
{{appName}} focuses on the step where component knowledge becomes writing ability: it prompts you first, keeps the character hidden until you have written it, checks the stroke order of your attempt and returns the ones you miss sooner. It works offline without an account. It is in early access.
How many components do I need to learn?
Around ten cover a large share of what a beginner meets, and a few hundred cover the vast majority of characters overall. The efficient approach is not to study a component list in isolation but to learn each one as you meet it inside characters you are already learning, starting with the most frequent: 氵, 讠, 木, 心, 手, 口, 日, 月, 女 and 亻.
