Hyōgaiji (表外字): The Kanji Beyond Jōyō and Jinmeiyō
Hyōgaiji (表外字) are kanji outside the 常用漢字表, the cabinet-approved jōyō list of 2,136 characters. Together with the 863 jinmeiyō kanji legal for personal names, jōyō defines the 2,999-character "name-eligible" core. Everything else is hyōgaiji by exclusion.1 2 3 For a serious reader, that residual category is where surnames, food kanji, literary writing, and much N1 reading text actually live.
Overview
What "outside the table" actually means
表外字 literally means "outside-the-table characters." The 表 is the 常用漢字表 (jōyō kanji-hyō), and 外 (gai) marks exclusion from it.3
Synonyms are in active use. 表外漢字 (hyōgai kanji), 常用外漢字 (jōyōgai kanji), 非常用漢字, and the bare 表外字 all denote the same residual category; style guides may prefer one term, but nothing of substance hinges on the choice.3
Distinguish hyōgaiji from jinmeiyō kanji (人名用漢字), the 863 characters legally registrable as personal-name kanji under the Ministry of Justice koseki rules. Jinmeiyō is a curated extension of jōyō for name use only. Everything outside both lists is hyōgaiji.2
Frequency and policy status are separate questions. Many hyōgaiji are very frequent in actual writing (嘘 "lie," 噂 "rumor," 鰻 "eel"), while some jōyō characters are rare in everyday text.3
The three concentric rings: jōyō, jinmeiyō, hyōgaiji
The cabinet inventory stacks into three rings. The 2010 cabinet notification fixed 常用漢字 at 2,136 characters. It added 196 characters and removed 5, replacing the 1981 list of 1,945 characters. 人名用漢字 stands at 863 characters after the September 25, 2017 addition of 渾.1 2
Their sum, 2,999 characters, defines the name-eligible core. Everything outside is hyōgaiji.
The inner two rings are curated; the outer rings are inventories of what exists, not of what is in use.
Why no definitive count exists
Hyōgaiji is a residual category, not a curated list. It is defined by exclusion from two government lists, so its outer bound depends on which character inventory you use as the frame.3
Counts you may encounter elsewhere are each tied to a specific source:
| Inventory | Count | Year / source |
|---|---|---|
| 表外漢字字体表 (printing-form recommendation) | 1,022 印刷標準字体 + 22 簡易慣用字体 | 2000 (国語審議会)4 5 |
| 大漢和辞典 (Dai Kan-Wa Jiten) | ~50,000 character entries | 1955–1960 first complete edition (sources disagree on 49,964 vs. 51,110)6 |
| 中華字海 (Zhōnghuá Zìhǎi) | 85,568 entries | 19947 |
| Unicode CJK Unified Ideographs | ~101,996 across main block + Extensions A–J | Unicode 17.08 |
Each frame answers a different question. The 表外漢字字体表 answers "which non-jōyō kanji are common enough to standardize a print form for?"
The Morohashi answers "which Sino-Japanese characters have ever been recorded in scholarship?" Unicode answers "which characters need any digital code point?"
The inventory beyond jōyō
表外漢字字体表 (2000): the closest thing to a curated hyōgaiji list
The 国語審議会 (Council on the Japanese Language) submitted the 表外漢字字体表 on December 8, 2000. It recommended which printing forms to use when non-jōyō kanji appear in laws, official documents, newspapers, magazines, and broadcasting.9 5
The table specifies 1,022 印刷標準字体 (printing standard forms) and 22 簡易慣用字体 (simplified conventional forms) within that set.4 5
It is not an exhaustive hyōgaiji catalog. The 1,022 characters are the non-jōyō kanji judged frequent enough in surveyed books, magazines, and newspapers to warrant a recommended printing form. The long tail of hyōgaiji is out of scope.5
The recommended forms in the 表外漢字字体表 are advisory. Major Japanese dailies and broadcasters generally honor them; private publishers and literary writers may deviate without sanction.5
大漢和辞典 (Morohashi): the ~50,000-entry scholarly ceiling
大漢和辞典 by 諸橋轍次 (Morohashi Tetsuji), published by 大修館 (Taishūkan), is the standard scholarly reference for Sino-Japanese characters. The original thirteen-volume edition appeared between 1955 and 1960. Later editions revised the index and entry layout.6
The headline count is commonly stated as "approximately 50,000" character entries. Detailed counts vary: one gives 49,964 head entries in the original edition, while later editions reach above 51,000. Compound entries number approximately 530,000.6
The dictionary gives an order-of-magnitude sense of how many Sino-Japanese characters have ever been recorded. It is not a count of "active" hyōgaiji. Most entries are dictionary-only forms a modern Japanese reader will never meet outside scholarship.6
中華字海 and CJK Unified Ideographs: the digital ceiling
中華字海 (Zhōnghuá Zìhǎi, 1994) is the largest single Chinese character dictionary in print. It has 85,568 distinct entries spanning standard, dialectal, Buddhist, Dunhuang manuscript, and historical Sino-Japanese / Sino-Korean characters.7
Unicode CJK Unified Ideographs covers approximately 101,996 code points across the main block (20,992) and Extensions A (6,592), B (42,720), C (4,160), D (222), E (5,774), F (7,473), G (4,939), H (4,192), I (622), and J (4,298) as of Unicode 17.0.8
Most CJK code points exist for completeness across the Han script. They cover historical, dialectal, and name-use characters across China, Japan, Korea, and Vietnam. They are encoded so that legacy texts and proper names can be represented digitally, not because they are active modern Japanese. The "100,000+" figure is the upper bound a "what is a kanji?" question can reach, not a target inventory.8
JIS X 0208 vs. JIS X 0213: which hyōgaiji your computer can render
The character set your software uses decides which hyōgaiji you can actually display. JIS X 0208 specifies 6,879 graphic characters: 6,355 kanji plus 524 kana, Latin letters, and symbols. It covers the jōyō set and most jinmeiyō characters but misses many surname and place-name hyōgaiji.10
JIS X 0213 extends JIS X 0208 by adding 3,625 kanji (2,673 already in JIS X 0212 plus 952 new across Levels 3 and 4); the 2004 revision, commonly cited as "JIS2004," refined 168 glyphs and added ten characters.11
In practice, text editing in legacy Shift-JIS environments (older email, some banking and government systems) is still bounded by JIS X 0208 and may not render the rarer hyōgaiji. Unicode UTF-8 is the practical superset for modern systems.10 11
Where hyōgaiji appear in modern Japanese
Surnames: the largest source of hyōgaiji exposure
The 1946 shinjitai simplifications applied to general use, but the family register predates that reform and was not retroactively rewritten. Common surname kyūjitai, or old character forms, still in active koseki use include 國 (= 国), 學 (= 学), 體 (= 体), 廣 (= 広), 澤 (= 沢), 邊 / 邉 (= 辺), 齋 (= 斎), and 龍 (= 竜).12 13
These kyūjitai are hyōgaiji by jōyō definition but legally registrable when they appear in the koseki, or family register. Modern readers often need furigana on business cards or in seating charts because the form is no longer taught in school.12 13
The 法務省 (Ministry of Justice) maintains the 戸籍統一文字 catalog, a unified character inventory for computerized koseki records. The catalog assigns a six-digit code to each registrable form. Some glyph distinctions unified in Unicode and JIS are kept distinct in 戸籍統一文字.14 15
Most consumer software cannot render the full 戸籍統一文字 inventory. That is why Japanese government forms still include a free-form handwriting box for family names. It is not nostalgia, but a fallback when the digital glyph stock falls short of the legal one.15
Place names: prefectures, wards, and abbreviation marks
The 2010 jōyō revision added prefecture-name characters that had previously been hyōgaiji (notably 阪 in 大阪 and 岡 in 岡山), closing the gap for prefecture names at the jōyō level.1
Ward, village, and historical place names still retain hyōgaiji below the prefecture tier. 沪 is the historical one-character abbreviation for 上海 (Shanghai) in Chinese usage and survives in Japanese cross-references.3
瀋陽 (Shenyang, capital of Liaoning) is written with the hyōgaiji 瀋, which reads しん. Modern simplified Chinese writes the name 沈阳. The character appears in Japanese diplomatic and travel writing but is not on either jōyō or jinmeiyō.16
瀋陽は中国の遼寧省の省都です。16
"Shenyang is the capital of China's Liaoning Province."
Food and animal names: 鯛, 鰻, 雀, 鯱
The 魚偏 (sakana-hen, "fish radical") category is overwhelmingly hyōgaiji. Restaurant signs, izakaya menus, and sushi-case labels use 鮨 (sushi), 鯛 (tai, sea bream), 鰻 (unagi, eel), 鰯 (iwashi, sardine), 鯵 (aji, horse mackerel), and 鯱 (shachi, killer whale or shachihoko ornament). None of these is on the jōyō list.3 17
鰻丼を二つください。17
"Two eel bowls, please."
Do not confuse the food-kanji layer with ateji, kanji used mainly for sound. 寿司, the more common written form of "sushi," uses two jōyō characters as ateji on a Yamato-kotoba root. That is a sound-borrowing mechanism layered on top of jōyō characters. The 魚偏 fish names underneath are a separate phenomenon, mostly hyōgaiji, and most of them are not ateji in the technical sense.17
Menu and signage writing tolerates hyōgaiji that newspaper style would either avoid or furigana-mark. The look of "specialist food kanji" is itself a register signal, not just a typographical accident.3
Literary and classical Japanese
嘘 (uso, "lie") is the standard "very common but technically hyōgaiji" example. Style guides recommend the kana form うそ in broadcasting and official documents, but corpus and editorial practice show the kanji form dominating in newspapers, novels, and online text. The character has on-yomi きょ and kun-yomi うそ, and is classified as hyōgai kanji.3 18 19
嘘をつくな。18
"Don't tell lies."
Meiji and pre-war fiction (Sōseki, Ōgai, Tanizaki) routinely uses hyōgaiji and kyūjitai forms that have since been standardized away. The same is true of 漢文 (kanbun, Classical Chinese-style Japanese) quotations and calligraphy.3
Edo-period and pre-war texts
Pre-1946 printed works use kyūjitai throughout: 國, 學, 體, 廣, 澤 in place of the post-reform 国, 学, 体, 広, 沢. The 1946 当用漢字表 (Tōyō Kanji-hyō) introduced 1,850 characters with their shinjitai forms. The 1981 replacement list of 1,945 jōyō kanji widened the set and kept the same simplification principle.12 20
After 1946, anything not absorbed into tōyō (and later jōyō) became hyōgaiji by default, regardless of how often it appeared in older text. The label is therefore not a statement about the character's history. It is a statement about whether the postwar reforms admitted it.12 13
How JLPT N1 dips into hyōgaiji
The N1 kanji envelope is wider than jōyō
The Japan Foundation has not published an official kanji list per JLPT level since the 2010 test redesign. Community estimates put N1 reading exposure at roughly 2,000 characters cumulative from N5. The reading section uses authentic text, so it admits hyōgaiji incidentally.1
The practical recognition target for an N1 candidate sits above the 2,136 jōyō ceiling. Reading practice with authentic text will surface hyōgaiji whether the candidate plans for them or not.1
Common hyōgaiji that surface on N1 reading
Characters that remain hyōgaiji and recur in N1-style prose include 嘘 (uso, "lie"), 噂 (uwasa, "rumor"), 蝶 (chō, "butterfly"), 雀 (suzume, "sparrow"), 牙 (kiba, "fang"), 牡 (osu, "male animal"), and 璧 (heki, the second character of 完璧, "perfection").3 18
Two characters frequently cited in older "JLPT N1 kanji" tables, 鬱 (utsu, "depression," 29 strokes) and 闇 (yami, "darkness"), were moved into jōyō by the November 30, 2010 cabinet notification. Treat any N1 list dated before late 2010 with that caveat. Verify against the current jōyō table before tagging anything as hyōgaiji.1
The 銑 case: removed from jōyō in 2010
銑 ("pig iron") sat on the 1981 jōyō list and was one of the five characters removed by the November 30, 2010 cabinet notification, alongside 勺, 脹, 錘, and 匁.1
For learners, the takeaway is that a character can leave jōyō and become hyōgaiji by policy change, not just by being old. Cross-check the deletions in any jōyō list published before 2010.1
Reading and writing hyōgaiji in practice
Old forms (旧字体) and why surnames keep them
The 1946 当用漢字表 applied shinjitai to general writing, but the family register (戸籍) was not retroactively rewritten. Names registered with 國, 澤, 邊, 廣, 齋 keep those forms permanently unless the family files a formal change.12 13
Furigana on Japanese business cards is standard precisely because the recipient may not recognize the registered form. The form is correct; the recognition burden is on the reader.13
Calligraphy preserves an even wider inventory. 書道 (shodō) and 篆刻 (tenkoku, seal-carving) routinely render characters in kyūjitai or further pre-modern forms. The same character can be hyōgaiji in a newspaper context and entirely standard in a calligraphy context. "Outside the table" is a print-policy label, not an aesthetic judgment.13
IME input: when the conversion candidate is missing
Standard Japanese IMEs, or input method editors (Microsoft IME, Google Japanese Input, macOS Japanese / Kotoeri), cover the JIS X 0213 set well and surface most JIS X 0208 candidates on the first conversion screen. Rarer hyōgaiji may require scrolling the candidate list or selecting a sub-pane.11
When the candidate is missing, use these workarounds in descending order of reliability: the IME handwriting pad (kanji handwriting input), copy-paste from an online dictionary such as Jisho or Wiktionary, and Ideographic Variation Sequences (IVS) for koseki-grade glyph distinctions.21
IVS is the technical mechanism behind koseki name handling. A base CJK code point plus a variation selector (VS17 through VS48 in Unicode terminology) encodes a specific glyph variant. The Adobe-Japan1-7 collection groups 14,664 kanji glyphs across its supplements and registers a corresponding set of Ideographic Variation Sequences in the Unicode Ideographic Variation Database. This is the engineering layer most users will never see directly.21
Predicting readings when no dictionary is at hand
Phono-semantic compounds (形声文字) preserve their on-yomi cue even when they are hyōgaiji. A 旁 (tsukuri, "phonetic side") that gives セイ across several jōyō characters usually still gives セイ when the same 旁 appears in a hyōgaiji. The semantic radical narrows the meaning family.3
The strategy is not perfect. Phonetic drift in Japanese and Chinese-side simplification both break it. A phonetic that gives two or three Sino-Japanese readings in jōyō cannot distinguish among them for a hyōgaiji from the form alone. As a first pass at reading speed, however, it is the single most useful predictive tool a reader has.3
The learner's working rule
萌えキャラが大好きです。22
"I love moe characters."
For hyōgaiji, aim for recognition only. Do not invest production effort. The 2,136 jōyō characters plus a working subset of jinmeiyō is the writing ceiling for almost every adult learner. Hyōgaiji are looked up, not memorized.1 2
The 2,000-character N1 recognition envelope and the 2,136-character jōyō writing envelope are different. Reading practice with authentic text will surface hyōgaiji long before they ever need to be written by hand.1
Good to know
Writing 嘘 in broadcasting and official documents
嘘 is hyōgaiji, and broadcasters and government documents follow the kana-form convention by writing うそ instead of the kanji. Newspapers, novels, and online writing do not follow the convention. The kanji form dominates corpus evidence even though the broadcasting recommendation persists.3 19 5 The form is right or wrong only relative to the venue.
Assuming "a shinjitai exists, therefore the kanji is jōyō"
Jōyō membership is set by the cabinet notification, not by whether a simplified variant of a character happens to exist. 籠 (kago, "basket") is hyōgaiji. Its more common print form 篭 is also hyōgaiji, not jōyō.
Likewise 籐 (tō, "rattan") is hyōgaiji and its homophonous 藤 ("wisteria") is jōyō, but the two characters do not share a meaning. The shinjitai / kyūjitai relationship partly overlaps with jōyō membership, but it is not 1:1.1 3 13
"Hyōgaiji" and "jōyōgai kanji" mean the same thing
表外字, 表外漢字, 常用外漢字, and 非常用漢字 all refer to the same residual category of non-jōyō characters. Linguistics writing tends to prefer 表外漢字; reference works vary. Nothing of substance hinges on the choice of term.3
Treating "hyōgaiji" as "rare"
Hyōgaiji are policy-rare, not usage-rare. Corpus evidence puts the 嘘 kanji form ahead of the kana うそ in newspapers, novels, and online text, despite the broadcasting recommendation. 噂, 鰻, and 蝶 behave similarly.3 19 A writer using "hyōgaiji" as shorthand for "uncommon" will misjudge an entire register of food, animal, and emotional vocabulary.
Kyūjitai surnames and the koseki
Surnames registered before the 1946 reform keep their kyūjitai forms in the koseki and on official documents. 長澤 まさみ uses the registered 澤, not the post-reform 沢. Printing the shinjitai variant is technically a different writing of the same family name. Furigana is standard on business cards because the recipient may not recognize the registered form, and the safety-net handwriting box on government forms exists for the same reason.15 12 13
嘘 as a Sino-Japanese borrowing with a Yamato-kotoba meaning
The Sino-Japanese readings of 嘘 (go-on こ, kan-on きょ) reflect a Chinese-side character meaning "exhale" or "breathe out." The Japanese meaning "lie" attaches to the kun-yomi うそ, which is a Yamato-kotoba root and not a Sino-Japanese borrowing. The Chinese-language meaning of 嘘 does not carry "lie" at all. The kanji is a phonetic and aesthetic match for a native Japanese word, layered on top of an unrelated Chinese meaning.18
Variant selectors (IVS) for koseki-grade name characters
The Ministry of Justice maintains 戸籍統一文字 as the authoritative inventory of family-register character forms. The catalog runs to roughly 56,000 entries, well beyond JIS X 0208. The Adobe-Japan1-7 collection groups 14,664 kanji glyphs and registers IVS sequences in the Unicode IVD. This lets Unicode-aware software address specific glyph variants, but most consumer software does not render the full catalog. Municipal koseki systems rely on specialist fonts such as IPAmj Mincho that ship the koseki glyph variants explicitly.14 15 21
Calligraphy preserves forms newspapers will not print
書道 (shodō) and 篆刻 (tenkoku) routinely render characters in kyūjitai or further pre-modern forms. The same character can be hyōgaiji in news context and entirely standard in calligraphy. The "outside the table" label is a print-policy label, not an aesthetic judgment.13
See also
- Secondary School Jōyō Kanji (中学校 + 高等学校): The 1,110-Character Set Beyond Elementary
- Nanori (名乗り): The Name-Only Kanji Readings That Break the On/Kun Rules
- The Six Categories of Kanji (六書): Pictographs, Ideographs, and Phono-Semantic Compounds
- How to Predict the Reading of an Unknown Kanji Compound: The On+On Default, Jūbako, Yutō, and the Look-It-Up Bucket
- Jukujikun (熟字訓): When a Compound's Reading Is Assigned to the Whole Word, Not the Kanji