Jukujikun (熟字訓): When a Compound's Reading Is Assigned to the Whole Word, Not the Kanji

A jukujikun is a kanji compound whose reading is assigned to the whole compound, not to its individual characters.¹² Once a learner has encountered 大人, 今日, and 明日 in a textbook and tried to break them down by character, jukujikun names exactly why those readings refuse to split.

Overview

The category sits inside the kun'yomi side of Japanese kanji readings, not outside it. The native reading attaches to the whole compound. The individual kanji supply meaning rather than sound.¹²

What jukujikun literally means

The term breaks into 熟字 ("compounded characters") plus 訓 ("native-Japanese reading").² The name itself carries the central claim: this is a kun (native) reading attached to a jukuji (compound cluster), not to its parts.

Japanese-language references gloss it directly. The 日本語版ウィキペディア entry, paraphrasing 大辞林, defines jukujikun as "a kun reading assigned not to single characters but to the cluster as a unit."¹ Wiktionary's gloss matches: a kanji compound used to write a native Japanese word by meaning, where the individual kanji do not correspond to separate parts of the word.²

Jukujikun is a kind of kun'yomi, not a third reading class

The 訓 in 熟字訓 is the same 訓 as in 訓読み. Jukujikun is a sub-type of native-reading assignment, not a separate category sitting alongside on'yomi and kun'yomi.¹²

Where jukujikun sits on the readings map

Standard kanji compounds use on'yomi (Sino-Japanese readings, often used inside multi-character compounds) or kun'yomi (native readings, often used in isolation).³ Jukujikun breaks that default by attaching a whole-word native reading to a multi-character compound. The individual characters contribute meaning rather than sound.¹²

This places jukujikun alongside the two regular compound types: on'yomi compounds such as 学校 (gakkō) and kun'yomi compounds such as 山道 (yamamichi).³

For background, read On'yomi vs. Kun'yomi: The Two-Reading System Behind Every Kanji. It lays out the default split that jukujikun breaks.

Register and frequency

Jukujikun are everyday vocabulary, not exotic edge cases. The Jōyō Kanji Table's 付表 ("appendix") lists 116 such compounds. It introduces them as "what are commonly called ateji and jukujikun: items difficult to list as the on or kun reading of any single character."⁴ The current 常用漢字表 was issued by Cabinet Notice in 2010 (平成22年内閣告示第2号).⁵

A learner working through a JLPT N4 or N3 vocabulary list will already have met most of these words: 大人, 今日, 明日, 田舎, 眼鏡, 七夕, 土産, 梅雨.⁴⁶⁷⁸⁹ The category label is N3+ because it explains words the learner has already seen.

The 116 count is the official total

116 is the figure in the appendix promulgated by MEXT (Japan's Ministry of Education, Culture, Sports, Science and Technology).⁴ Wiktionary's category listing shows around 100 entries, but the gap comes from Wiktionary's selection, not from a disagreement with the source.¹⁰ Outside the jōyō appendix, JMdict tags roughly 130 to 284 entries depending on how the count is drawn, because JMdict includes non-jōyō and literary compounds.¹¹¹²

How a jukujikun reading is assigned

The mechanical point is simple to state but slightly counter-intuitive. In a jukujikun compound, the reading binds to the cluster. It does not distribute across the characters.

The reading attaches to the cluster, not the characters

Take 大人 otona. No syllable sequence in otona lines up with either kanji: 大 does not supply "o" and 人 does not supply "tona".¹¹³ The word is one indivisible native morpheme written with two characters chosen for meaning.

Dictionary data records that the reading cannot be segmented. Jisho's entry on 大人 おとな carries the gikun flag on the reading itself, so the reading is filed under the compound rather than under either character's reading list.¹³ The JMdict DTD spells out what the flag means: "gikun (meaning as reading) or jukujikun (special kanji reading)", which the EDRDG wiki expands as "readings that are based on the meaning of the term and not the readings of the kanji form."¹⁴¹¹

Why dictionaries cannot split the reading

A regular kun'yomi compound passes a simple decomposition test. 山道 yamamichi breaks cleanly into 山 (yama, the native reading of "mountain") plus 道 (michi, the native reading of "road"). Each kanji contributes a known kun reading.³

Jukujikun fails that test. No syllable sequence in otona matches an attested reading of 大 or 人, so the cluster has to be filed as a single entry.¹³

The decomposition test in one sentence

If every syllable sequence in a compound's reading can be matched to a known on'yomi or kun'yomi for its character (allowing for rendaku), the compound is regular. If no such alignment exists, the compound is a jukujikun candidate.¹³

In JMdict's data, the difference appears in different fields. The gikun tag attaches to the reading element (re_inf), marking the reading as meaning-driven. The parallel ateji tag attaches to the kanji element (ke_inf), marking the spelling as sound-driven.¹⁴¹¹ It is the same compound entry, but two different fields.

The historical mechanism in one paragraph

The native Japanese word existed first. When Chinese characters were borrowed to write Japanese, writers applied them to pre-existing native words by meaning. This principle runs through the early Japanese writing system.¹⁵

The resulting split between on'yomi and kun'yomi is the standard pattern every learner eventually meets: most kanji carry both a Sino-Japanese reading and a native one.³ Jukujikun is what happens when that meaning-based borrowing applies to a whole native word. The kanji are chosen to capture the meaning of the cluster, and the cluster as a unit carries the native pronunciation.

The Japanese Wikipedia entry frames the process as a story of conventionalization. Jukujikun originate as one-off meaning-based assignments (義訓, gikun) that became standard over time: "個人的な使用(義訓)から生じてそれが慣用的なものとして定着したものが現在に見られる熟字訓である."¹ In plain English: today's jukujikun are readings that began as individual uses and later settled into conventional use.

Common examples, walked

The high-frequency examples below are the ones a learner most often meets first. The headline case includes the expected wrong reading so the trap is visible. The rest list the actual jukujikun reading with its dictionary status.

大人 (otona): "adult"

This is the headline case. The expected combinations ō-bito (kun + kun) or dai-jin (on + on) never appear as the everyday reading of 大人. Jisho lists おとな as the standard reading, tagged gikun.¹³ The 付表 lists 大人 as an elementary-school learning item.¹⁶⁴¹⁰

君きみはもう大人おとなだ。¹⁷
"You're an adult now."

今日 (kyō) and 明日 (ashita): "today" and "tomorrow"

Both compounds are jukujikun in their everyday readings. 今日 reads きょう (the gikun-tagged reading) for "today". 明日 reads あす or あした.¹⁸¹⁹ The 付表 lists 今日 with reading きょう and 明日 with reading あす.¹⁶⁴¹⁰

Both compounds also carry on'yomi readings that are real but register-shifted. 今日 konnichi and 明日 myōnichi exist as more formal alternatives. Konnichi also carries the "these days" sense that kyō does not.¹⁸¹⁹

今日きょうは大学だいがく入試にゅうしの日ひだった。²⁰
"Today was the day of the university entrance exam."

明日あす締しめ切きりのレポートが提出ていしゅつできなければ、単位たんいを落おとす。²⁰
"If I can't hand in tomorrow's report, I'll lose the credit."

Kyō and konnichi are not interchangeable

今日 read as kyō covers "today" only. Read as konnichi, it covers both "today" (formal register) and "these days / nowadays".¹⁸ The set phrase 今日は ("hello / good day") is konnichi wa, not kyō wa. Choosing kyō there is a category error, not a register slip.¹⁸

田舎 (inaka), 海老 (ebi), 紅葉 (momiji)

田舎 reads いなか and is tagged gikun. The 付表 lists it as a middle-school learning item.¹⁶⁴¹⁰⁶ The gloss is "rural area; countryside; the sticks; hometown."⁶

海老 reads えび. The entry is JMdict gikun-tagged but not in the 常用漢字表 付表, because the appendix only includes items where both kanji are jōyō and the cluster is treated as a standard reading.¹²²¹ 海老 is a real, common jukujikun that sits outside the official short list.

紅葉 carries two readings. もみじ is the jukujikun reading. It names autumn-coloured leaves and the Japanese maple. こうよう is the on'yomi reading. It names the phenomenon of leaves turning red.²² Both readings are in active use. The choice between them is semantic, not register.²² The 付表 lists 紅葉 with reading もみじ as a middle-school learning item.¹⁶⁴¹⁰

僕ぼくは田舎いなかで育そだった。¹⁷
"I grew up in the countryside."

庭にわの木きがすっかり紅葉こうようしました。¹⁷
"The trees in the garden have turned completely red."

A short reference set the learner can recognize

The table below collects nine compounds that turn up early and often. Treat the list as recognition material: memorize the compound-reading pairing rather than trying to predict it.¹

Compound	Reading	Gloss	付表	JMdict gikun
一日	ついたち	first day of the month	Yes16410	Yes23
七夕	たなばた	Tanabata, the Star Festival	Yes16410	Yes127
二十歳	はたち	20 years old	Yes16410	(see note)24
眼鏡	めがね	glasses, spectacles	Yes16410	Yes8
土産	みやげ	souvenir, gift from a trip	Yes164	Yes9
為替	かわせ	money order, foreign exchange	Yes16410	Yes25
梅雨	つゆ	the rainy season (early summer)	Yes16410	Yes1226
時雨	しぐれ	late-autumn or early-winter shower	Yes16410	Yes1227
五月雨	さみだれ	early-summer rain	Yes16410	Yes1228

For 二十歳, the 付表 is the primary authority on the reading はたち. Jisho's individual entry does not show a gikun flag, though the compound appears in JapanDict's broader gikun-tagged list.^164101224

梅雨つゆは六月ろくがつの終おわり頃ごろに始はじまる。¹⁷
"The rainy season begins around the end of June."

私わたしは彼女かのじょが新あたらしい眼鏡めがねをかけているのに気きがついた。¹⁷
"I noticed she was wearing new glasses."

The boundary with ateji

Jukujikun is most often confused with ateji, because both involve kanji that do not behave as expected. The distinction is whether the kanji choice is based on sound or meaning.

Ateji binds sound; jukujikun binds meaning

JMdict's two tags split the work cleanly. The ateji flag (on the kanji element) marks spellings "where one or more of the kanji are used to phonetically represent native or borrowed words with less regard to the underlying meaning."¹¹ The gikun flag (on the reading element) marks "readings that are based on the meaning of the term and not the readings of the kanji form."¹⁴¹¹

The two standard contrasts are 寿司 (sushi) and 大人 (otona). 寿司 is ateji because 寿 supplies the su sound and 司 supplies the shi sound, with no semantic connection to "sushi."¹¹ 大人 is jukujikun because 大 ("big") and 人 ("person") contribute meaning, not sound.¹³

The Japanese Wikipedia entry draws the same line in stricter language. Jukujikun presupposes that the compound is built on Chinese-style word-formation logic with semantically meaningful kanji. Ateji ignores the kanji's inherent meaning or compound structure and applies kanji to a native or loan word by their sound or kun reading.¹

Why the two terms still bleed into each other

The 付表 itself groups the two categories into a single introductory phrase: "what are commonly called ateji and jukujikun: items difficult to list as a single character's reading."⁴ A learner who meets that phrasing first may conclude that the categories are sub-cases of one larger phenomenon. Pedagogically, they often are.

English Wikipedia treats jukujikun as a narrow sub-case under the broader ateji label.² JMdict resolves the same ambiguity by tagging the two ends separately: ateji on the spelling for sound-driven kanji choice, gikun on the reading for meaning-driven character choice.¹⁴¹¹ Hybrid entries that carry both tags exist. The two framings answer the same classification question in different ways, not in mutual contradiction.

A dedicated article on ateji will cover the boundary in full. In this article, ateji means sound-driven kanji choice for an existing native or loan word.

The boundary with regular kun'yomi compounds

The other look-alike category is the ordinary kun'yomi compound. These can look like jukujikun at a glance because they use native readings, but each kanji still carries its own assigned reading.

When each kanji still carries its own reading

花火 hanabi breaks into 花 (kun hana, "flower") plus 火 (kun bi, voiced from hi, "fire").³ 山道 yamamichi breaks into 山 (yama) plus 道 (michi).³ 手紙 tegami breaks into 手 (te) plus 紙 (gami, voiced from kami, "paper").

All three are regular kun'yomi compounds. The voicing changes (hi → bi, kami → gami) are rendaku, a productive sound rule, not jukujikun behavior.

彼かれらは花火はなびを打うち上あげた。¹⁷
"They set off fireworks."

この山道やまみちを登のぼると美うつくしい湖みずうみに出でる。¹⁷
"If you climb this mountain path you come out at a beautiful lake."

The decomposition test

The full procedure is the same one sketched earlier, now written out as steps.

1. List the attested on'yomi and kun'yomi of each character in the compound.
2. Align each syllable sequence in the compound's reading to one character. Allow rendaku.³
3. If every syllable sequence lands on a known reading, the compound is on'yomi or kun'yomi (or a mixed jūbako or yutō compound).³
4. If no alignment works (no syllable in otona corresponds to any reading of 大 or 人), the compound is a jukujikun candidate. Verify against the 付表 or against a JMdict gikun entry.⁴¹²

The test is asymmetric on purpose. Passing it identifies a regular compound. Failing it identifies a jukujikun candidate that still needs dictionary confirmation, because gaps in the learner's reading list can also produce false negatives.

How dictionaries mark jukujikun

The same compound appears under different labels depending on which dictionary the learner uses. Knowing the labels helps the learner spot the category in the wild.

The 常用漢字表 付表 (jōyō kanji appendix)

The appendix is maintained by 文化庁 (the Agency for Cultural Affairs) under MEXT auspices and was issued by Cabinet Notice as part of 平成22年内閣告示第2号 (the 2010 edition of the 常用漢字表).⁵²⁹ It contains 116 multi-character compounds whose readings cannot be assigned at the single-character level.⁴

The official prefatory wording groups ateji and jukujikun into the same appendix rather than separating them.²⁹⁴ Teaching versions of the table also annotate each entry with the learning stage at which it is introduced (elementary or middle school).¹⁶

This is the conservative, official short list. A compound outside the appendix can still be jukujikun (海老 / ebi is the clear example), but it falls outside the standardized, curated set.¹²²¹

The JMdict / Jisho gikun tag

JMdict's DTD defines gikun as "gikun (meaning as reading) or jukujikun (special kanji reading)"; the two terms are collapsed into one tag.¹⁴ The tag attaches to the reading element (re_inf), distinct from ateji, which attaches to the kanji element (ke_inf).¹⁴¹¹

Coverage is broader than the 付表. EDRDG's own wiki cites around 130 entries. JapanDict's list shows around 284 once it includes entries pulled from related reading-info flags.¹¹¹² Jisho.org exposes the tag in plain language on each affected entry; for example, the 大人 おとな entry carries the line "Gikun (meaning as reading) or jukujikun (special kanji reading)."¹³

Paper-dictionary conventions

In monolingual references such as 大辞林 and 日本国語大辞典, jukujikun readings are entered under the compound itself rather than under any component character.² The reference works that document jukujikun explicitly include 大辞林 (Matsumura, 2006), 日本国語大辞典 (Shōgaku Tosho, 1988), and NHK 日本語発音アクセント辞典 (NHK 放送文化研究所, 1998).²

In learner-facing kanji dictionaries, a 難訓 ("difficult reading") note on a standalone character is the typical signal that a reading involving that character appears only as part of a fixed compound.

Good to know

Jukujikun is the reason "guess the reading" eventually fails

A learner who has internalized on'yomi-in-compounds and kun'yomi-in-isolation eventually meets a compound that disobeys both rules. The cluster has to be looked up, not derived. No productive rule recovers otona from 大 plus 人.¹¹³

The takeaway is procedural rather than philosophical. When a compound resists every reading the learner can list for its characters, the next step is not to keep guessing at the character level. Check whether the compound is in the 付表 or carries a JMdict gikun flag.⁴¹²

The same kanji compound can have both a jukujikun and a regular reading

今日 reads きょう (jukujikun, everyday) and こんにち (on'yomi, formal, and the only reading that carries the "these days" sense).¹⁸ 明日 reads あした or あす (jukujikun, everyday) and みょうにち (on'yomi, formal written register).¹⁹ 紅葉 reads もみじ (jukujikun, the leaves themselves) and こうよう (on'yomi, the seasonal phenomenon).²²

The forms are not errors; they have different distributions. Context, register, and sense select between them.

Gikun vs. jukujikun in the strict sense

Linguists reserve jukujikun for established whole-compound readings that have become conventional. They use gikun for one-off authorial readings, often furigana stunts in song lyrics, manga, or literary prose.¹¹⁴ JMdict's tagging flattens the distinction under a single label.¹⁴

The Japanese Wikipedia entry frames the relationship as conventionalization: jukujikun are readings that arose from individual usage (gikun) and became standard.¹ In practical terms, a learner can treat the categories as a continuum from one-off creative readings to dictionary-fixed compound readings.

Names are not jukujikun

Personal-name readings fall under a separate category called 名乗り (nanori), which has its own dictionary marking and its own rules. The two categories can look similar (a name reading can also look unsegmentable), but they are tracked separately.

When a learner meets a surname or given name with a surprising reading, the relevant label is nanori, not jukujikun.

Why 海老 sits outside the official short list

海老 is a real jukujikun: the reading ebi attaches to the cluster, not to 海 ("sea") or 老 ("old, aged").²¹ It is tagged gikun in JMdict's broader corpus but does not appear in the 常用漢字表 付表, because the appendix is restricted to items where both characters are jōyō and the cluster is treated as the standard reading.¹²²¹

Words outside the appendix can still be jukujikun. The 付表 is the curated short list, not the full inventory. JMdict's tagging is wider but less authoritative.

See also

• How to Predict the Reading of an Unknown Kanji Compound: The On+On Default, Jūbako, Yutō, and the Look-It-Up Bucket
• Go-on, Kan-on, Tō-on: The Historical Layers Behind a Kanji's Multiple On'yomi
• Jukugo (熟語): How Kanji Combine to Form Japanese Words
• The History of Kanji: From Oracle Bones to the Jōyō List

References

Footnotes

1. 日本語版ウィキペディア. 「熟字訓」. https://ja.wikipedia.org/wiki/%E7%86%9F%E5%AD%97%E8%A8%93 ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹² ↩¹³ ↩¹⁴

2. 英語版ウィクショナリー. Entry: 熟字訓 (cites Daijirin (Matsumura, 2006), 日本国語大辞典 (Shōgaku Tosho, 1988), and NHK 日本語発音アクセント辞典 (1998)). https://en.wiktionary.org/wiki/%E7%86%9F%E5%AD%97%E8%A8%93 ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹

3. Shibatani, Masayoshi. The Languages of Japan. Cambridge University Press, 1990, pp. 129–130. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰

4. CyberLibrarian. 「常用漢字表 付表 (116語)」reference table. https://www.asahi-net.or.jp/~ax2s-kmtn/ref/jouyoukanji.html ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹² ↩¹³ ↩¹⁴ ↩¹⁵ ↩¹⁶ ↩¹⁷ ↩¹⁸ ↩¹⁹ ↩²⁰ ↩²¹ ↩²²

5. 文化庁. 「常用漢字表」(平成22年内閣告示第2号). https://www.bunka.go.jp/kokugo_nihongo/sisaku/joho/joho/kijun/naikaku/kanji/ ↩ ↩²

6. Jisho.org. Entry: 田舎 いなか. https://jisho.org/word/%E7%94%B0%E8%88%8E ↩ ↩² ↩³

7. Jisho.org. Entry: 七夕 たなばた. https://jisho.org/word/%E4%B8%83%E5%A4%95 ↩ ↩²

8. Jisho.org. Entry: 眼鏡 めがね. https://jisho.org/word/%E7%9C%BC%E9%8F%A1 ↩ ↩²

9. Jisho.org. Entry: 土産 みやげ. https://jisho.org/word/%E5%9C%9F%E7%94%A3 ↩ ↩²

10. 日本語版ウィクショナリー. 「カテゴリ:常用漢字表 付表」. https://ja.wiktionary.org/wiki/%E3%82%AB%E3%83%86%E3%82%B4%E3%83%AA:%E5%B8%B8%E7%94%A8%E6%BC%A2%E5%AD%97%E8%A1%A8_%E4%BB%98%E8%A1%A8 ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹² ↩¹³ ↩¹⁴

11. EDRDG Wiki. "Kanji and Reading Information Fields" (JMdict tag glossary). https://www.edrdg.org/wiki/Kanji_and_Reading_Information_Fields.html ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹

12. JapanDict. "List: gikun or jukujikun" (re_inf=gikun entries in JMdict). https://www.japandict.com/lists/re_inf/gikun ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹²

13. Jisho.org (JMdict-backed). Entry: 大人 おとな. https://jisho.org/word/%E5%A4%A7%E4%BA%BA ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶

14. Electronic Dictionary Research and Development Group. JMdict Project: DTD entity declarations (gikun, ateji, re_inf, ke_inf). https://www.edrdg.org/jmdict/jmdict_dtd_h.html ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸

15. Frellesvig, Bjarke. A History of the Japanese Language. Cambridge University Press, 2010. https://www.cambridge.org/core/books/history-of-the-japanese-language/E45445FAB65DB7133748485D5BDEE29A ↩

16. 静岡県総合教育センター. 「常用漢字表 付表」教材PDF. https://gakusyu.shizuoka-c.ed.jp/japanese/syou_56/moji/04/fuhyou.pdf ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹² ↩¹³ ↩¹⁴ ↩¹⁵

17. Tanaka Corpus (Tatoeba-distributed Japanese-English sentence pairs), accessed via Jisho.org and Tatoeba. https://tatoeba.org/en/sentences/search?from=jpn&to=eng ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷

18. Jisho.org. Entry: 今日 きょう / こんにち. https://jisho.org/word/%E4%BB%8A%E6%97%A5 ↩ ↩² ↩³ ↩⁴ ↩⁵

19. Jisho.org. Entry: 明日 あす / あした / みょうにち. https://jisho.org/word/%E6%98%8E%E6%97%A5 ↩ ↩² ↩³

20. Jreibun (筑波大学). Example sentence corpus, accessed via Jisho.org example-sentence pages. https://jreibun.tsukuba.ac.jp/ ↩ ↩²

21. Jisho.org. Entry: 海老 えび. https://jisho.org/word/%E6%B5%B7%E8%80%81 ↩ ↩² ↩³ ↩⁴

22. Jisho.org. Entry: 紅葉 もみじ / こうよう. https://jisho.org/search/%E7%B4%85%E8%91%89 ↩ ↩² ↩³

23. Jisho.org. Entry: 一日 ついたち / いちにち. https://jisho.org/word/%E4%B8%80%E6%97%A5 ↩

24. Jisho.org. Entry: 二十歳 はたち. https://jisho.org/word/%E4%BA%8C%E5%8D%81%E6%AD%B3 ↩ ↩²

25. Jisho.org. Entry: 為替 かわせ. https://jisho.org/word/%E7%82%BA%E6%9B%BF ↩

26. Jisho.org. Entry: 梅雨 つゆ. https://jisho.org/search/%E6%A2%85%E9%9B%A8%20%23sentences ↩

27. Jisho.org. Entry: 時雨 しぐれ. https://jisho.org/word/%E6%99%82%E9%9B%A8 ↩

28. Jisho.org. Entry: 五月雨 さみだれ. https://jisho.org/word/%E4%BA%94%E6%9C%88%E9%9B%A8 ↩

29. 文化庁. 「常用漢字表について(答申): 表の見方及び使い方」. https://www.bunka.go.jp/kokugo_nihongo/sisaku/joho/joho/kakuki/14/tosin02/04.html ↩ ↩²