{
  "id": "combining-mark",
  "title": "Combining Mark",
  "lang": "en",
  "source_lang": "en",
  "summary": "A combining mark is a character encoded to modify a base character or character sequence rather than function as an independent written symbol.",
  "category": "language",
  "subcategory": "",
  "type": "concept",
  "aliases": [
    "Combining character"
  ],
  "generated_at": "2026-11-05",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/combining-mark/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/combining-mark/",
      "json": "https://aiwiki.page/zh/combining-mark.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "A combining mark is a character intended to appear in visual association with a base character or character sequence, commonly above, below, within, or beside it. In Unicode, combining marks have their own encoded identities, allowing letters and other symbols to be represented as sequences of components. For example, é can be represented by the letter e followed by a combining acute accent. Combining marks support diacritics, vowel signs, pronunciation notation, and other features of many writing systems."
    },
    {
      "heading": "Characters, marks, and glyphs",
      "text": "A combining mark is an encoded character, not merely a small graphic added to a letter. Its visible appearance is a glyph selected and positioned by a rendering system. The same mark may take different shapes or positions depending on the base character, typeface, and surrounding text.\n\nThree distinctions are important:\n\n- A diacritic is a writing-system concept. It is a mark that modifies or supplements another written sign. A combining mark is a character-encoding concept. Many combining marks are diacritics, but the categories are not identical.\n- A combining mark is not necessarily nonspacing. Some marks occupy horizontal space even though they remain dependent on another character.\n- One perceived written unit need not be one encoded character. A letter and several marks may form a single grapheme, while software stores them as multiple characters.\n\nFor example, the acute accent in é may indicate pronunciation or another orthographic distinction according to the language. Unicode encodes the mark’s identity, not a universal linguistic meaning for it."
    },
    {
      "heading": "Representation in Unicode",
      "text": "Unicode assigns each encoded character a code point, conventionally written as U+ followed by hexadecimal digits. The sequence:\n\nU+0065 LATIN SMALL LETTER E\nU+0301 COMBINING ACUTE ACCENT\n\nrepresents é. Although the accent normally appears above the letter, it follows the letter in the encoded sequence.\n\nUnicode also includes:\n\nU+00E9 LATIN SMALL LETTER E WITH ACUTE\n\nwhich represents é as a single, precomposed character. The two representations are canonically equivalent: Unicode treats them as alternative encodings of the same abstract text.\n\nNot every base-and-mark combination has a precomposed counterpart. Encoding marks separately makes it possible to represent combinations without assigning an individual code point to every possible accented letter.\n\nCombining marks and spacing counterparts\n\nA combining accent differs from a visually similar character intended to stand independently. For example:\n\nCharacter | Unicode name | Typical role\nU+0301 | COMBINING ACUTE ACCENT | An accent attached to a base\nU+00B4 | ACUTE ACCENT | A spacing accent character\nU+0027 | APOSTROPHE | Punctuation or an orthographic sign\n\nTheir shapes may resemble one another in some fonts, but they are not generally interchangeable. Replacing one with another can change searching, rendering, and interpretation.\n\nUnicode categories\n\nThe Unicode General Category property divides marks into three categories:\n\nCategory | Name | General behavior\nMn | Nonspacing Mark | Combines with another character without ordinarily receiving its own advance width\nMc | Spacing Mark | Acts as a dependent mark but ordinarily contributes to text advance\nMe | Enclosing Mark | Surrounds or encloses another character or sequence\n\nThese categories describe broad character behavior rather than guaranteeing an exact visual layout. Actual positioning depends on the script and rendering system.\n\nThe Combining Diacritical Marks block, U+0300–U+036F, contains many familiar accents, but it is not the complete inventory of combining marks. Marks also occur in supplementary diacritic blocks, script-specific blocks, and blocks containing combining marks for symbols. Block membership and general category are different properties."
    },
    {
      "heading": "Linguistic uses",
      "text": "Combining marks serve different functions across orthographies and notation systems.\n\nAlphabetic diacritics\n\nIn the Latin alphabet, combining accents can represent stress, tone, vowel quality, or distinctions between letters. Pinyin, for example, uses diacritics to indicate tone. Some letters require more than one mark, such as a vowel bearing both a diaeresis and a tone accent.\n\nThe Greek alphabet likewise uses accents and, in polytonic spelling, breathing marks and other signs. The linguistic status of an accented letter depends on the language: it may be treated as a separate alphabetic letter, a modified form of another letter, or a pronunciation-marked form.\n\nDependent signs in other scripts\n\nIn text written in Arabic and Hebrew, many vowel and pronunciation signs are encoded as combining marks associated with consonant letters. Such signs may be omitted in some kinds of writing and included in educational, religious, or other texts.\n\nIn scripts such as Devanagari, dependent vowel signs and other marks participate in larger written syllabic structures. Their displayed position need not match their position in the encoded character sequence. A vowel sign encoded after a consonant may be rendered to its left, above it, below it, or in multiple visual parts.\n\nPhonetic and technical notation\n\nThe International Phonetic Alphabet uses combining diacritics to indicate features such as nasalization, devoicing, and secondary articulation. Combining marks also occur in mathematical and specialist notation, including overlays and enclosing signs."
    },
    {
      "heading": "Normalization and mark order",
      "text": "Unicode normalization provides standardized ways to handle equivalent character sequences.\n\nTwo central normalization forms are:\n\n- NFD, which performs canonical decomposition and canonical ordering.\n- NFC, which performs canonical decomposition and ordering, followed by canonical composition where permitted.\n\nThus, é becomes e followed by a combining acute accent in NFD, while that sequence becomes the precomposed é in NFC. NFC does not eliminate all combining marks: many combinations have no precomposed equivalent.\n\nCanonical combining class\n\nUnicode assigns characters a Canonical_Combining_Class value. This property helps determine the order of marks during normalization. For example, a dot below and an acute accent above a letter have different combining classes, allowing normalization to establish a consistent order for those marks.\n\nCombining class must not be confused with the Mn, Mc, and Me categories. A combining mark can have combining class zero. Nor does the class fully specify the mark’s visual position.\n\nMarks with the same combining class generally retain their relative order during canonical ordering. Arbitrarily sorting or rearranging marks can therefore change text rather than merely standardize its representation.\n\nUnicode also provides U+034F COMBINING GRAPHEME JOINER, an invisible character that can prevent canonical reordering across its position. Despite its name, it is not a general-purpose instruction to join visible glyphs."
    },
    {
      "heading": "Rendering and text boundaries",
      "text": "Rendering combining marks correctly requires more than drawing a small symbol at a fixed offset. A font may provide attachment points for positioning a mark relative to a base and for stacking one mark over another. Text shaping selects and arranges glyphs according to the script, character sequence, and available font data.\n\nRendering problems can occur when:\n\n- A font lacks a required mark.\n- A fallback font supplies a mark with incompatible metrics.\n- A shaping system lacks the information needed for correct placement.\n- Several marks exceed the space allowed by a line or interface.\n\nA mark without a suitable base may be displayed with a dotted circle, ◌, to make the missing base apparent. In explanatory examples, a dotted circle may also be inserted explicitly. Its appearance does not necessarily mean it is present in the underlying text.\n\nGrapheme clusters\n\nSoftware often uses an extended grapheme cluster as an approximation to a user-perceived character. A base letter followed by combining marks normally forms one such cluster. Cursor movement, selection, and deletion can operate on that cluster rather than on each code point separately.\n\nCode-point count is therefore not a reliable measure of visible text length. A displayed accented letter may consist of one code point or several, and encoding units or bytes introduce further differences."
    },
    {
      "heading": "Historical and practical significance",
      "text": "Combining-mark mechanisms predate Unicode. Overprinting, nonspacing accents, and earlier character-set standards provided ways to assemble accented forms. Unicode incorporated both combining mechanisms and numerous precomposed characters, partly to support compatibility with existing encoded text.\n\nFor text processing, this produces several important distinctions:\n\n- Equality: Visually identical strings may have different code-point sequences unless canonical equivalence is taken into account.\n- Searching and sorting: Language-sensitive processing may need to distinguish meaningful accents while accommodating equivalent encodings.\n- Accent removal: Decomposing text and deleting marks is not a universal transliteration method. Marks may encode essential information, and some letters do not decompose into a base plus a mark.\n- Input: Keyboards and input methods may produce either precomposed characters or combining sequences.\n- Display robustness: Long sequences of marks can overlap neighboring text or exceed expected layout boundaries.\n\nCombining marks consequently require coordinated handling across character encoding, normalization, fonts, rendering, and editing interfaces. Their visual dependence on other characters does not make them optional or linguistically insignificant."
    }
  ],
  "links": [
    "unicode",
    "writing-system",
    "glyph",
    "grapheme",
    "code-point",
    "orthography",
    "latin-alphabet",
    "pinyin",
    "tone-linguistics",
    "greek-alphabet",
    "devanagari",
    "international-phonetic-alphabet",
    "font",
    "grapheme-cluster"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}