Okurigana (送り仮名, Japanese pronunciation: [okɯɾigana], "accompanying letters") are kana suffixes following kanji stems in Japanese written words. They serve two purposes: to inflect adjectives and verbs, and to force a particular kanji to have a specific meaning and be read a certain way. For example, the plain verb form 見る (miru, "see") inflects to past tense 見た (mita, "saw"), where 見 is the kanji stem, and る and た are okurigana, written in hiragana script. With very few exceptions,[note 1] okurigana are only used for kun'yomi (native Japanese readings), not for on'yomi (Chinese readings), as Chinese morphemes do not inflect in Japanese,[note 2] and their pronunciation is inferred from context, since many are used as parts of compound words (kango).
When used to inflect an adjective or verb, okurigana can indicate aspect (perfective versus imperfective), affirmative or negative meaning, or grammatical politeness, among many other functions. In modern usage, okurigana are almost invariably written with hiragana; katakana were also commonly used in the past.
Analogous orthographic conventions find occasional use in English, which, being more familiar, help in understanding okurigana.
As an inflection example, when writing Xing for cross-ing, as in Ped Xing (pedestrian crossing), the -ing is a verb suffix, while cross is the dictionary form of the verb – in this case cross is the reading of the character X, while -ing is analogous to okurigana. By contrast, in the noun Xmas for Christmas, the character Χ is instead read as Christ (it is actually a chi in origin, from the Greek Χριστός, Khristós). The suffixes serve as phonetic complements to indicate which reading to use.
Note that word, morpheme (constituent part of word), and reading may be distinct: in "1", "one" is at once the word, the morpheme, and the reading, while in "1st", the word and the morpheme are "first", while the reading is fir, as the -st is written separately, and in "Xmas" the word is "Christmas" while the morphemes are Christ and -mas, and the reading "Christ" coincides with the first morpheme.
Adjectives in Japanese use okurigana to indicate aspect and affirmation-negation, with all adjectives using the same pattern of suffixes for each case. A simple example uses the character 高 (high) to express the four basic cases of a Japanese adjective. The root meaning of the word is expressed via the kanji (高, read taka and meaning "high" in each of these cases), but crucial information (aspect and negation) can only be understood by reading the okurigana following the kanji stem.
Japanese verbs follow a similar pattern; the root meaning is generally expressed by using one or more kanji at the start of the word, with aspect, negation, grammatical politeness, and other language features expressed by following okurigana.
Compare the direct polite verb forms to their distant forms, which follow a similar pattern, but whose meaning indicates more distance between the speaker and the listener:
Okurigana are also used as phonetic complements to disambiguate kanji that have multiple readings, and consequently multiple meanings. Since kanji, especially the most common ones, can be used for words with many (usually similar) meanings — but different pronunciations — key okurigana placed after the kanji help the reader to know which meaning and reading were intended. Both individual kanji and multi-kanji words may have multiple readings, and okurigana are used in both cases.
Okurigana for disambiguation are a partial gloss, and are required: for example, in 下さる, the stem is 下さ (and does not vary under inflection), and is pronounced くださ (kudasa) – thus 下 corresponds to the reading くだ (kuda), followed by さ (sa), which is written here kuda-sa. Note the okurigana are not considered part of the reading; grammatically the verb is kudasa-ru (verb stem + inflectional suffix), but orthographically the stem itself is analyzed as kuda-sa (kanji reading + okurigana). Compare with furigana, which specify the reading of the kanji, appear outside the line of the text, and which are omitted if understood.
Disambiguation examples include common verbs which use the characters 上 (up) and 下 (down):
Observe that many Japanese verbs come in transitive/intransitive pairs, as illustrated above, and that a single kanji reading is shared between the two verbs, with sufficient okurigana written to reflect changed endings. The above okurigana are as short as possible, given this restriction – note for instance that のぼる (noboru) / のぼす (nobosu) are written as 上る / 上す, not as ×上ぼる or ×上ぼす, while あがる must be written as 上がる to share a kanji reading with 上げる.
Another example includes a common verb with different meanings based on the okurigana:
Okurigana are not always sufficient to specify the reading. For example, 怒る (to become angry) can be read as いかる (ika-ru) or おこる (oko-ru) – ×怒かる and ×怒こる are not used[note 4] – 開く (to open) may be read either as あく (a-ku) or as ひらく (hira-ku) – ×開らく is not used – and 止める may be read either as とめる (to-meru) or as やめる (ya-meru) – ×止る is not used.[note 5] In such cases the reading must be deduced from context or via furigana.
Ambiguity may be introduced in inflection – even if okurigana specify the reading in the base (dictionary) form of a verb, the inflected form may obscure it. For example, 行く i-ku "go" and 行う okona-u "perform, carry out" are distinct in dictionary form, but in past ("perfective") form become 行った i-tta "went" and 行った okona-tta "performed, carried out" – which reading to use must be deduced from context or furigana.[note 6]
One of the most complex examples of okurigana is the kanji 生, pronounced shō or sei in borrowed Chinese vocabulary, which stands for several native Japanese words as well:
as well as the hybrid Chinese-Japanese words
Note that some of these verbs share a kanji reading (i, u, and ha), and okurigana are conventionally picked to maximize these sharings.
Okurigana may also be used in multi-kanji words, where the okurigana specifies the pronunciation of the entire word, not simply the character that they follow; these distinguish multi-kanji native words from kango (borrowed Chinese words) with the same characters. Examples include nouns such as 気配り kikubari "care, consideration" versus 気配 kehai "indication, hint, sign" (note that the reading of 気 changes between ki and ke, despite it not having an okurigana of its own), and verbs, such as 流行る hayaru "be popular, be fashionable", versus 流行 ryūkō "fashion". Note that in this later case, the native verb and the borrowed Chinese word with the same kanji have approximately the same meaning, but are pronounced differently.
Okurigana can also occur in the middle of a compound, such as 落ち葉 ochiba "fallen leaves" and 落葉 rakuyō "fallen leaves, defoliation" – note that the reading of the terminal 葉 changes between ba and yō despite it occurring after the okurigana.
For a few categories of words, okurigana correspond to historical suffixes which are no longer distinct or productive, and the suffix is now fused to the word, but still written in hiragana. This is particularly the case for words which function as adjectives, with notable categories including:
Note that only the -i in -shii inflects; the other kana are invariant, and in practice serve only for disambiguation and to reflect historical grammar. Briefly, -shii adjectives used to be a different class from -i adjectives (distinguished historically as -ku and -shiku adjectives, for present -i and -shii), but have since merged; -yaka and -raka used to be suffixes, but are no longer productive, while -taru and -naru are historical variants of what is now the adjective particle -na. See Japanese equivalents of adjectives for details.
For group II verbs (一段動詞 ichidan dōshi, also known as ru-verbs) the okurigana begin at the mora preceding the last, unless the word is only two morae long.
If the verb has different variations, such as transitive and intransitive forms, then the different morae are written in kana, while the common part constitutes a single common kanji reading for all related words.
In other cases (different verbs with similar meanings, but which are not strictly variants of each other), the kanji will have different readings, and the okurigana thus also indicate which reading to use.
Most adjectives ending in -i (true adjectives) have okurigana starting from the -i.
Okurigana starts from shi for adjectives ending in -shii (this reflects historical grammar; see above).
Exceptions occur when the adjective also has a related verbal form. In this case, as with related verbs (above), the reading of the character is kept constant, and the okurigana are exactly the morae that differ.
As with verbs, okurigana is also used to distinguish between readings (unrelated adjectives with the same kanji), in which case the okurigana indicate which reading to use.
Na-adjectives (adjectival verbs) that end in -ka have okurigana from the ka.
The last mora of an adverb is usually written as okurigana.
Note that such adverbs are often written in kana, such as 全く matta-ku まったく and 専ら moppa-ra もっぱら.
Nouns do not normally have okurigana.
In some cases the reading is then ambiguous, and must be deduced from context or by furigana.
However, if the noun is derived from a verb or adjective, it may take the same okurigana, although some may be omitted in certain cases (see below). The derivation may not be apparent if it is old and the verb is no longer in use (see below).
For some nouns it is obligatory to omit the okurigana, despite having a verbal origin.
In these cases, the noun form of the corresponding verb does take okurigana.
Formally, the verbal noun (VN, still retaining verbal characteristics) takes okurigana, as is usual for verbs, while the deverbal noun (DVN, without verbal characteristics) does not take okurigana, as is usual for nouns.
To understand this grammatical distinction, compare the English present participle (verb form ending in -ing, indicating continuous aspect) and the gerund (noun form of the -ing verb form, which is a verbal noun) versus deverbal forms (which are irregular):[note 7]
Similarly, some nouns are derived from verbs, but written with different kanji, in which case no okurigana are used.
In other cases a kanji may be derived from another verb or verb combination and retain the okurigana:
Some okurigana come from Old Japanese, and the underlying verb is no longer in use.
Note that these -i suffixes are not i-adjectives – they are the ends of verb stems.
In compounds, okurigana may be omitted if there is no ambiguity in meaning or reading – in other words, if that compound is only read a single way. If okurigana occur after several characters (esp. both in the middle of the compound and at the end of the compound, as in 2-character compounds), either only the middle okurigana, or both the middle and the final okurigana may be omitted; omitting only the final okurigana but retaining the middle okurigana is rather unusual and somewhat questionable, though not unknown (marked with “?” below).
This is particularly done for Japanese compound verbs (the okurigana inflection of the first, main verb is dropped), as above. This is especially common in reducing or removing kana in formulaic constructions, particularly in signs. For example, in the common phrase 立入禁止 (tachi-iri kin-shi, Do not enter; literally, entry prohibited) in analyzed as 立ち入り + 禁止, but the okurigana are usually dropped.
If the compound is unfamiliar to the reader, there is the risk of it being incorrectly read with on readings, rather than the kun readings – for example, 乗入禁止 "drive-in forbidden" is read nori-ire-kin-shi (乗り入れ禁止) – the first two characters are a compound verb – but an unfamiliar reader may guess jōnyū-kinshi based on the on readings. However, this is not a problem with familiar compounds, whose reading is already known.
Okurigana are avoided in compounds where the reading cannot easily be analyzed into readings of the individual characters, as these are confusing – the reading simply must be learnt separately. These include especially ateji and gikun, as well as cases where a compound word has changed pronunciation over the years, and is no longer a simple combination of the compounds. For example, 息吹 i-buki "breath" is specifically prescribed to not have okurigana – there is the related verb 息吹く i-bu-ku "to breathe", which must have okurigana for inflection, but 息 is otherwise pronounced iki, so there is the risk of misreading as *iki-buki. This is formalized for the words in the addendum to the Jōyō list in the second category of exceptions, listed below.
The above rules are guidelines, and there are exceptions and special cases that must be learnt individually: okurigana that has become standard for historical reasons (now obscure or not obvious at a glance) or by convention rather than logic. Compare for instance:
These both originally derived from the verb 明る aka-ru, which is no longer in use; the first is an irregularly derived i-adjective, while the latter is a deverbal noun. Compare 明く a-ku and 明らか aki-raka.
The Japanese Ministry of Education (MEXT) prescribes rules on how to use okurigana, giving standardized Japanese orthography. The original notification (see references) is from 1973, but it was amended in 1981 when the jōyō kanji table was issued.
The rules apply to kun'yomi (native Japanese readings) of kanji in the jōyō kanji table; they do not apply to kanji outside the jōyō kanji table, or kanji without kun'yomi (with only on'yomi, Chinese readings). The notification gives 7 general rules (通則) and 2 rules for difficult cases (付表の語) in the jōyō kanji table's word list attachment (付表). The first 2 rules (1 & 2) address words that conjugate, the next 3 rules (3–5) address words that do not conjugate, and the last 2 rules (6 & 7) address compound words. Whenever there's doubt whether something is permissible use (許容) or not, the general rule (通則) is to be followed. In some cases, variations are permitted, when there is no danger of confusion; in other case, when there is danger of confusion, variations are not permitted.
Examples for each rule, with permitted variations:
There are 16 special cases listed, 7 where okurigana is required or recommended, 9 where it is forbidden. These refer to prescribed spellings of words on the attachment to the Jōyō kanji list.
Required or recommended (acceptable alternatives in parentheses) (付表の語1):
While MEXT prescribes rules and permitted variations, in practice there is much variation – permitted or not – particularly in older texts (prior to guidelines) and online – note that these rules are not prescriptive for personal writings, but only in official documents and media. As an example, the standard spelling of the word kuregata is 暮れ方, but it will sometimes be seen as 暮方.
While okurigana are sufficient to show inflection of adjectives and verbs, in rare cases further sound change occurs that affects the stem, and must be inferred from the okurigana, without being explicitly written. An everyday example is o-hayo-u (お早う, Good day), where the stem would normally be pronounced haya (はや), as this comes from the i-adjective haya-i (早い, early) – the a → o sound change must be inferred from the following -u. This sound change is due to this being a polite adjective form.
If there is additional non-inflectional okurigana, then these are change, and this is sufficient to show the sound change in the spelling. This occurs for adjectives ending in -shii, like ure-shii (嬉しい, happy), hence ure-shū (嬉しゅう). A basic example other than -shii of such okurigana use is ō-kyū (大きゅう), from ō-kii (大きい, big).
There is a risk of confusion of okurigana with compounds: some Japanese words are traditionally written with kanji, but today some of these kanji are hyōgaiji (uncommon characters), and hence are often written as a mixture of kanji and kana, the uncommon characters being replaced by kana; this is known as mazegaki. The resulting orthography is seen by some as confusing and unsightly, particularly if it is the second character that is written in kana – the kana characters are where okurigana would be expected to go – and this is one motivation for expansions of kanji lists. For example, until the 2010 expansion of the jōyō kanji, the word kanpeki (perfect) was officially written 完ぺき, not as the compound 完璧, since the character 璧 was not on the official list, and takarakuji (lottery) is officially (and also popularly) written as 宝くじ, not as 宝籤, since the second character is not in the jōyō kanji and is also quite complicated. This is less of an issue when the first kanji is written in kana, as in ヤシ殻 (yashi-kaku, coconut shell), which is formally 椰子殻.
Converse to okurigana, where part of the pronunciation of a word is written after the kanji, in some cases following kana is dropped (or included in the reading of the previous character). This is primarily for the attributive particles 〜の -no (sometimes written in katakana as ノ) and 〜が -ga (sometimes written ヶ or ケ), and is most common in names. For example, the common family name Inoue (I-no-ue Well’s top, top of well) is generally written 井上, though if the particle were written it would be 井の上. Similarly, Amagasaki (Ama-ga-saki, Nun’s peak) is generally written 尼崎, but can be written 尼ヶ崎, and Sen no Rikyū is written 千利休 (here the particle is between the family and given name). This can sometimes cause ambiguity, as in the Yamanote Line (for a time called the Yamate Line) and the Agatsuma Line (which could be read as Azuma). Particles are considered grammatically separate from the attached word (they are not an inflection), and this is not considered okurigana, despite some superficial similarities.
Japanese has various affixes, some of which are written in kana and should not be confused with okurigana. Most common are the honorifics, which are generally suffixes, such as 〜さん -san (Mr., Ms.), and bikago (美化語, "beautified language"), such as お〜 and ご〜 as in お茶 (ocha, tea).