Chinese Characters

Writing Systems of the World

Course 1 · Chapter 3 · Chinese Characters

Cuneiform and hieroglyphs are no longer written. Chinese characters are, by a very large number of people, and they are the main surviving logographic system. This chapter looks at what a character stands for, how characters are built, and which other languages have borrowed them.

The idea in a box

Calling Chinese writing "pictures for ideas" is a myth. Per Wikipedia's Chinese characters article, characters generally represent morphemes, the smallest units of meaning, and as these are nearly always one syllable, written Chinese is often called morphosyllabic. Most characters also contain a part that hints at the sound.

Origins

Earliest evidence. The first attested characters are oracle bone inscriptions from the 13th century BCE, found at Anyang in Henan and tied to Shang royal divinations. The article dates the oldest attested writing to Late Shang (c. 1250 to 1050 BCE), with the very earliest examples from the reign of Wu Ding, 1250 to 1200 BCE.
Quantity. Over 150,000 oracle bone fragments have been found to date.
Standardisation. After Qin's conquests led to the Qin dynasty in 221 BCE, the small seal script was standardised for use throughout the country under Chancellor Li Si.

Chapter 1 gave c. 1200 BC for the Yellow River valley; the figures are consistent with the Shang dates here.

How Characters Are Built

The traditional analysis comes from the Shuowen Jiezi (c. 100 CE, by Xu Shen) and was the main framework for nearly two thousand years. It has six categories, called liushu.

Category	Idea
Pictographs	a stylised picture of the thing
Indicatives	a sign pointing to an abstract idea
Compound ideographs	meanings of two parts combined
Phono-semantic compounds	one part for meaning, one part for sound
Loangraphs	a character borrowed for a similar-sounding word
Zhuanzhu	definition unclear; modern scholars generally disregard it

(The category names are the article's; my one-line glosses are my own paraphrase.) The article's proportions for about 3,500 frequently used characters in Standard Chinese are the surprise: phono-semantic compounds make up roughly 58%, pure signs about 18%, semantic-form and phonetic-form compounds together 19%, and pure semantographs about 5%. So the majority of common characters have a sound clue, which is the rebus principle from Chapter 2 working inside a logographic system.

A worked example (mine)

Two standard textbook cases, which I give as my own examples and have not checked against the article: 媽 mā "mother" combines 女 (woman) for meaning with 馬 mǎ (horse) for sound, and 河 hé "river" combines the water element with 可 kě. The sound clue is only approximate (here the tone and vowel differ), which is why readers must still learn each character.

import unicodedata
for ch in "水河可媽馬女":
    cp = ord(ch)
    print(ch, "U+%04X" % cp, unicodedata.name(ch), 0x4E00 <= cp <= 0x9FFF)

水 U+6C34 CJK UNIFIED IDEOGRAPH-6C34 True 河 U+6CB3 CJK UNIFIED IDEOGRAPH-6CB3 True 可 U+53EF CJK UNIFIED IDEOGRAPH-53EF True 媽 U+5ABD CJK UNIFIED IDEOGRAPH-5ABD True 馬 U+99AC CJK UNIFIED IDEOGRAPH-99AC True 女 U+5973 CJK UNIFIED IDEOGRAPH-5973 True

All six fall in the CJK Unified Ideographs block, U+4E00 to U+9FFF, which the article gives. Notice the names: Unicode names these by code point, not by meaning or reading, because one block serves Chinese, Japanese and Korean.

How Many Characters?

Writing all frequently used vocabulary requires roughly 3,500 to 5,000 characters, per the article.
More than 100,000 characters have been identified and included in the Unicode Standard (as of 2025, per the article).

Compare Chapter 1: alphabets use fewer than 100 symbols. A logographic system trades a large inventory for a direct link between symbol and meaning.

Simplified and Traditional

Simplified characters are broadly used in mainland China, Singapore and Malaysia; traditional characters in Taiwan, Hong Kong and Macau. The article, in the part I read, does not say when or how the divergence began.

Characters Beyond Chinese

Japanese uses kanji, the subject of the site's kanji pages. The article calls Japanese "the only major non-Chinese language still written using them". Chapter 4 looks at the kana that Japanese adds.
Korean uses hanja, and Vietnamese used chữ Hán, with chữ Nôm for writing Vietnamese itself. Both have largely switched to alphabets (Hangul is the subject of Chapter 8).
Zhuang also had its own script, sawndip.

Watch out: a character is not a word, and Chinese writing is not "the language"

Characters represent morphemes, and many Chinese words are two or more characters long. And because characters stand for meaning, readers who speak different varieties of Chinese can share a written form while speaking differently; that claim is a common one, and I have not checked it in the part of the article I read.

Note on sources

This chapter uses Wikipedia's Chinese characters article (read 100,000 of 240,614 characters; the rest was not read). Dates, percentages and counts are the article's and unchecked against primary scholarship. The article text I read does not cover radical counts or dictionary sizes, so none appear here. The glosses of the six categories, the 媽 and 河 examples and the Python code are mine.

Try It Yourself

Exercise 1

Using the article's figures, roughly how many of 3,500 common characters are phono-semantic compounds? Why does that undermine the idea that Chinese is a pure "picture writing"?

📄 View solution

Exercise 2

Run the Python example on three characters you find on a Chinese or Japanese web page. Do they fall in the block U+4E00 to U+9FFF? What does the name tell you?

📄 View solution

Exercise 3

Why does a logographic system need thousands of characters while an alphabet needs fewer than a hundred? Answer in terms of what one symbol stands for (Chapter 1).

📄 View solution

Chapter 3 Quick Reference

Earliest Chinese writing: Shang oracle bones, c. 1250 to 1050 BCE
Characters represent morphemes (usually one syllable): morphosyllabic
About 58% of 3,500 common characters are phono-semantic compounds (meaning part plus sound part)
3,500 to 5,000 characters cover frequent vocabulary; one CJK block (U+4E00 to U+9FFF) serves Chinese, Japanese and Korean

Coming next

Writing Systems of the World 4 turns to syllabaries: Japanese kana, Cherokee and Linear B.