East Asian Width
Unicode property (UAX#11) classifying characters as Narrow, Wide, Fullwidth, Halfwidth, Ambiguous, or Neutral. Wide characters (CJK ideographs, katakana) occupy two columns in terminal emulators.
What is East Asian Width?
East Asian Width is a Unicode character property defined in UAX #11 — East Asian Width that classifies characters according to the display width they should occupy in fixed-width (monospaced) rendering environments, particularly traditional East Asian terminals and text layouts. The property answers the question: "Does this character occupy one column or two columns when displayed in a terminal or monospaced layout?"
The property was introduced because East Asian scripts — Chinese, Japanese, Korean — were historically displayed at twice the width of ASCII characters on fixed-pitch terminals. Mixing ASCII and CJK text in a single terminal line required a consistent model for how much horizontal space each character would consume.
The Six Width Categories
| Category | Property Value | Description | Examples |
|---|---|---|---|
| Narrow | N | ASCII and most Latin/Greek/Cyrillic — one column | A, a, 1, @ |
| Wide | W | Most CJK ideographs, Hangul syllables — two columns | 漢, 가, ア |
| Fullwidth | F | ASCII-range characters in their fullwidth CJK form — two columns | A, 1, ! |
| Halfwidth | H | Katakana and Hangul in their halfwidth (legacy) form — one column | ア, ヲ |
| Ambiguous | A | Characters that are narrow in Western contexts but wide in some East Asian contexts | © , ☆, α |
| Neutral | N | Non-East-Asian-specific characters with no width ambiguity, typically narrow | Arrows, math operators |
Terminal Implications
In a terminal emulator, the renderer must know the East Asian Width of every character to correctly advance the cursor. If a Wide or Fullwidth character is assumed to be narrow, subsequent characters will overwrite existing content, causing display corruption.
The POSIX standard function wcwidth() (from <wchar.h>) returns 0 for combining characters, 1 for narrow characters, and 2 for wide characters. Modern terminal emulators implement wcwidth() based on UAX #11 data.
# Python: get East Asian Width property
import unicodedata
def display_width(char: str) -> int:
# Return terminal display width of a single character.
eaw = unicodedata.east_asian_width(char)
return 2 if eaw in ("W", "F") else 1
# Examples
display_width("A") # → 1 (Narrow)
display_width("漢") # → 2 (Wide)
display_width("A") # → 2 (Fullwidth)
display_width("ア") # → 1 (Halfwidth)
The wcwidth Python package provides a conformant implementation updated with each Unicode release.
The Ambiguous Category
Characters with Ambiguous (A) width are the most problematic in practice. Their width is context-dependent:
- In an East Asian context (a terminal set to a CJK locale), they display as Wide (2 columns)
- In a Western context, they display as Narrow (1 column)
This affects many common symbols: degree sign (°), copyright symbol (©), Greek letters (α, β), and many box-drawing characters. Terminal emulators that serve mixed-locale user bases must make a policy choice, and mismatches between the application and terminal settings cause visible misalignment.
Quick Facts
| Property | Value |
|---|---|
| Defined in | UAX #11 — East Asian Width |
| Number of categories | 6 (Narrow, Wide, Fullwidth, Halfwidth, Ambiguous, Neutral) |
| Two-column characters | Wide (W) and Fullwidth (F) |
| Most problematic category | Ambiguous (A) — context-dependent |
| POSIX function | wcwidth() — returns 0, 1, or 2 |
| Python stdlib | unicodedata.east_asian_width(char) |
| Python package | wcwidth (conformant, kept updated) |
関連用語
プロパティ のその他の用語
文字が最初に割り当てられたUnicodeバージョン。システムやソフトウェアバージョン間での文字サポートを判断するのに役立ちます。
Unicode property controlling how Arabic and Syriac characters connect to adjacent characters. …
Unicode property listing all scripts that use a character, broader than the …
文字を大文字・小文字・タイトルケースに変換するルール。ロケール依存の場合があり(トルコ語のI問題)、1対多のマッピングもあります(ß → SS)。
文字が属する文字体系(例:ラテン、キリル、漢字)。Unicode 16.0は168個のスクリプトを定義し、Scriptプロパティはセキュリティと混在スクリプト検出に重要です。
サポートしていないプロセスで目に見える効果なく無視できる文字で、異体字セレクター・ゼロ幅文字・言語タグなどが含まれます。
名前付きの連続したコードポイント範囲(例:基本ラテン = U+0000〜U+007F)。Unicode 16.0は336個のブロックを定義し、すべてのコードポイントはちょうど1つのブロックに属します。
RTLコンテキストでグリフを水平に反転すべき文字。例:( → )、[ → ]、{ → }、« → »。
すべてのコードポイントを30個のカテゴリ(Lu・Ll・Nd・Soなど)の1つに分類する体系で、7つの主要クラス(文字・記号・数字・句読点・記号・区切り・その他)にグループ化されています。
同じ抽象的内容を持つが外観が異なる場合がある2つの文字シーケンス。正規等価より広い概念。例:fi ≈ fi、² ≈ 2。