The History of Hiragana: From Man'yōgana to the 46-Character Set
Hiragana began as Chinese characters borrowed for their sounds. Over centuries, those characters were simplified into the 46-character phonetic script that beginners learn first.12 That set was not inevitable. It is the result of two government acts (1900 and 1946) layered onto a millennium of calligraphic drift.34
Overview
Why this history matters to a learner
The Hiragana Chart you memorize in week one has 46 base characters. That number comes from two policy decisions, the 1900 Elementary School Order and the 1946 Gendai Kanazukai cabinet notification, not from organic linguistic evolution alone.34
Every modern hiragana traces to one or more parent kanji first borrowed for their sound as man'yōgana. The shapes you copy in Stroke Order practice are cursive descendants of those kanji, not invented symbols.125
The spelling exceptions you meet in your first week (は read as wa for the topic particle, を read as o for the object particle, へ read as e for the direction particle) are explicitly preserved remnants of pre-1946 historical kana orthography. The Three Hiragana Spelling Exceptions are not arbitrary; they are 1946 carve-outs.467
A one-paragraph timeline
Kanji-bearing inscriptions are attested in Japan from the late 5th century, with the Inariyama Sword conventionally dated to 471 CE.8 Man'yōgana, kanji used for sound rather than meaning, is the system used to write the Man'yōshū in the 8th century.5 A cursive intermediate stage called sōgana, followed by proto-hiragana, is attested from the 9th century.12 The Kokin Wakashū, commissioned by Emperor Daigo, is dated to about 905 CE and is the first imperial anthology compiled in kana.9 The Tale of Genji, finished by 1021, is the canonical high point of hiragana literature.10 The 1900 Elementary School Order Enforcement Regulations prescribed one hiragana per syllable for primary-school textbooks.311 The 1946 Cabinet Notification "Gendai Kanazukai" retired ゐ and ゑ from general use, leaving the modern 46.4127
The shape of that arc is easier to scan as a diagram than as a sentence.
What this article does not cover
Katakana arose in the same Heian period but by a different mechanism: Buddhist monks abbreviated fragments of man'yōgana rather than fully cursivizing them. It is a parallel lineage, not a substep of the hiragana story, and is treated separately in The History of Katakana.12
Before hiragana: how Japanese got written at all
Japanese had a language but not a script
Old Japanese was a spoken language with no native script. Writing arrived from the continent across the 5th and 6th centuries CE, both through the Korean peninsula and through direct contact with China.113
The earliest securely dated long inscription in Japan is the Inariyama Sword, excavated in 1968 at the Inariyama Kofun in Saitama Prefecture. Its inscription was revealed by X-ray in 1978, contains at least 115 Chinese characters, and is conventionally dated to 471 CE.8
The Inariyama text is written in classical Chinese, but it already uses Chinese characters as syllabograms to write Japanese personal names phonetically. That technique, borrowing a kanji for its sound while ignoring its meaning, is the foundational move that man'yōgana later generalized into a full writing system.8
The Eta Funayama Sword from Kumamoto Prefecture is the other standard early dated specimen. It is usually grouped with the Inariyama Sword in lists of pre-Asuka inscriptions found in Japan.135
Man'yōgana: borrowing kanji for sound, not meaning
Man'yōgana uses kanji characters for sound rather than meaning.5 The system takes its name from the Man'yōshū (万葉集), the Nara-period poetry anthology compiled in the 8th century in which it is the dominant writing convention.5
A single Japanese syllable could be written with any of several candidate kanji. The syllable /a/ alone could be 安, 阿, 亜, 吾, 足, or 鳴. The choice depended on calligraphic preference and spelling tradition, not on a one-to-one mapping.15
Man'yōgana was fully in use across the 8th century, not only for poetry but also for legal documents, imperial edicts (senmyō, 宣命), and Buddhist annotations. It was the working phonetic-script convention of literate elites before cursive simplification began.113
Why this was unworkable for daily writing
Writing native Japanese with full-form man'yōgana required a stroke-heavy kanji for every syllable. For prose of any length, the cost in time and ink became impractical, so scribes increasingly wrote those characters in the cursive sōsho (草書) style.113
This cursive shortening produced sōgana (草仮名), an intermediate stage in which the parent kanji remained recognizable but visibly compressed. Further simplification removed the kanji outline and produced hiragana proper.12
The birth of hiragana in the Heian period
From kanji to sōgana to hiragana
The lineage proceeds in three documented steps: man'yōgana written in regular script, the same kanji written in sōsho cursive as sōgana, and then further simplification that erased the kanji outline and produced hiragana.12
Each modern hiragana has a documented parent kanji. The clearest way to feel the simplification is to look at five examples side by side:
| Hiragana | Parent kanji | Sino-Japanese reading of parent | Original meaning of parent |
|---|---|---|---|
| あ | 安 | an | peace |
| い | 以 | i | by means of |
| う | 宇 | u | eaves, heaven |
| か | 加 | ka | add |
| に | 仁 | jin | benevolence |
The reading of the parent kanji is why it was chosen. 安 was picked for /a/ because its Sino-Japanese reading begins with that vowel. The meaning ("peace") was irrelevant to the borrowing.514
When hiragana first appears in the record
Forms intermediate between sōgana and fully simplified hiragana are attested in 9th-century manuscripts, including documents preserved in the Shōsōin treasure house.113
The first imperial-rank text canonically compiled in kana is the Kokin Wakashū (古今和歌集). It was conceived by Emperor Uda and published by order of his son Emperor Daigo in about 905 CE; its finished form dates to c. 920.9 Four court poets led the compilation: Ki no Tsurayuki, Ki no Tomonori, Ōshikōchi no Mitsune, and Mibu no Tadamine.9
The anthology has two prefaces: a Japanese (kana) preface by Ki no Tsurayuki and a Classical Chinese preface by Ki no Yoshimochi. The kana preface is the earliest surviving piece of formal literary criticism written in Japanese in hiragana.9
The Heian context: a court that needed a faster script
Heian aristocratic culture (794–1185) placed heavy weight on private letter exchange, poetic correspondence (waka), and diary keeping (nikki). For these genres, the full-form kanji apparatus of Chinese-style writing was too slow and too formal in style.113
Hiragana grew in exactly this niche. Its early literary monuments are diaries, poetry, and prose fiction, not government documents. Ki no Tsurayuki's Tosa Nikki of c. 935, written in kana with the male author posing as a woman, is one of the earliest signed kana diaries.113
The "women's script" story and what the evidence says
Where the "onnade" label comes from
Heian-era aristocratic education separated formal Chinese-classics training, given to men, from a vernacular literary register that was broadly accessible to women of the court.1315 The regular-script (kaisho) kanji used by men in formal contexts was called otokode (男手, "men's hand"); the cursive kana register became known as onnade (女手, "women's hand").2
The most-cited monuments of Heian vernacular prose, the Tale of Genji by Murasaki Shikibu and the Pillow Book (枕草子) by Sei Shōnagon, were written by women, mostly in hiragana in a vernacular style.10 The Tale of Genji is written mostly in Japanese phonetic script, not in Chinese characters. At the time, writing in kanji was a masculine pursuit, and women were generally discreet when using kanji.10
Why the label oversimplifies
The male compilers of the Kokin Wakashū, Ki no Tsurayuki and three other men, wrote the anthology and its kana preface in hiragana. They were using hiragana for the most prestigious literary commission of their generation.9
Lurie's Realms of Literacy argues that the split between "men's writing" and "women's writing" in the Heian record reflects social access to Chinese-classics training and choices of register, not an exclusively sex-segregated script system. Men used hiragana in private letters and in poetry; women learned kanji when family circumstances permitted.15
By the 10th century, hiragana was in general use across genders as the standard script for vernacular Japanese, even as the onnade label persisted as a historical descriptor of its social origin.115
Why this distinction matters for a learner
The popular short-form claim "hiragana was invented by women" compresses two facts that careful history keeps apart. Women drove the literary maturation of hiragana in the Heian court. The script became associated with women's writing because formal Chinese-classics education was reserved for aristocratic men.15210
"Onnade" was the register name in Heian high society. It does not mean men did not write hiragana; it means the educated male elite had a more prestigious alternative (Chinese-classics kanji) that women were broadly excluded from. The label survived the social conditions that produced it.15
Hentaigana: the kana that did not survive
One sound, many kana
Because each man'yōgana started from a different parent kanji, a single Japanese syllable could have several legitimate cursive kana forms. A Heian or Edo writer chose among them for calligraphic flow and visual effect.1116 The collective name for these variant forms is hentaigana (変体仮名, "variant kana").1116
How long this lasted
This multi-variant system was the everyday state of hiragana from the 9th century through the end of the 19th, roughly a thousand years.1116 Edo-period printed books routinely used several variant forms of the same syllable on a single page. The practice was the rule, not the exception, until the 1900 standardization.11
Where you still see hentaigana today
Hentaigana persist on traditional soba shop noren (fabric shop curtains), where kisoba ("fresh soba") is conventionally written using a hentaigana form of き. They also turn up on yose theater signage, on formal handwritten certificates, and on shopfronts that want an old-fashioned look.11
The set is no longer purely analog. Unicode 10.0 (June 2017) added 285 hentaigana characters to the Standard. They are distributed across the Kana Supplement (U+1B000–U+1B0FF) and Kana Extended-A (U+1B100–U+1B12F) blocks. Together with one hentaigana letter added in Unicode 6.0, the total comes to 286 typeable variant forms.1117
The 1900 standardization: how 46 became official
What the Elementary School Order did
In Meiji 33 (1900), the Ministry of Education issued the Elementary School Order Enforcement Regulations (小学校令施行規則, 文部省令第14号, 21 August 1900). For primary-school textbooks, the regulations prescribed only one specific character for each mora, with all the rest called hentaigana.311
The selected form was generally the simplest or most-used variant. Every other form was classified as hentaigana and removed from the educational curriculum.311
Why the timing made sense
Meiji-era Japan was building mass literacy through a national elementary-school system and a modern publishing industry. Multi-variant kana made textbook production, movable-type setting, and standardized education difficult.1311
The 1900 ordinance was one of three Ministry of Education writing reforms that year. The others were a kanji-count cap for primary schoolbooks and a rationalization of kana spelling for Sino-Japanese readings (字音仮名遣い).13
The set that was fixed in 1900
The 1900 order produced a 48-character base inventory: the modern 46 plus ゐ and ゑ. The 46 a learner studies today is the 1900 set minus the two characters retired in 1946.3411
The 1946 reform: ゐ and ゑ retire
What "gendai kanazukai" replaced
Pre-1946 spelling followed historical kana orthography (歴史的仮名遣い, rekishiteki kanazukai, also called 旧仮名遣い, kyū-kanazukai). It preserved spellings that reflected Heian-era pronunciation, even where the spoken sound had since merged or shifted.67
A standard example is the historical spelling けふ for the word now pronounced kyō ("today"), spelled phonetically as きょう in modern orthography. Another is the historical わらふ (warafu) for modern わらう (warau, "to laugh").6 The Yōon, Small つ Sokuon, and Long Vowels articles cover how the modern spellings encode the same sounds without the historical detour.
On 16 November 1946 (Shōwa 21), the cabinet adopted "Gendai Kanazukai" (現代かなづかい) as Cabinet Notification No. 33 (内閣告示第33号). It respelled words to match modern pronunciation.4127
Why ゐ (wi) and ゑ (we) were dropped
The wi and we sounds had merged with i and e in standard Japanese centuries before the reform. By 1946, the characters preserved a phonemic distinction (a sound contrast that changes word meaning) that no longer existed in speech.16
The 1946 notification replaced ゐ with い and ゑ with え in word spellings. The characters remain in use only for proper names, deliberate archaism, and stylistic effect.47
What survived from the historical set
The particles は (topic), を (object), and へ (direction) are explicitly preserved in their historical spellings and read as wa, o, and e respectively. The 1946 notification states that, for these particles, the historical kana usage is treated as identical (助詞の「は」「を」「へ」は歴史的仮名遣と同一とする).47
The result is a 46-character base inventory in which を sits on the chart but ゐ and ゑ do not. The asymmetry is practical, anchored by the particle function of を, rather than phonetic. All three characters represent merged sounds.47
Phonetically, を is in the same boat as ゐ and ゑ: a character for a sound that merged out of standard Japanese centuries ago. It survived because retiring it would have forced the respelling of every object phrase in the language. The Spelling Exceptions article unpacks the particle behavior in detail.47
The Allied-Occupation context, briefly
The 1946 reforms were issued by the Japanese cabinet, not by the Allied occupation. The Tōyō Kanji list of 1,850 characters was promulgated by the same cabinet on the same day, 16 November 1946.12
The linguistic groundwork was already centuries old: the sound mergers that made ゐ, ゑ, and a phonemic wo/o distinction redundant. The occupation-era literacy push provided political momentum, not the linguistic rationale.112
Why hiragana exists alongside katakana and kanji
Three scripts, three social origins
Kanji entered Japan from China across the 5th and 6th centuries and remained for meaning-bearing roots in vocabulary.113 Hiragana grew out of cursive man'yōgana in the Heian court and remained for grammatical inflection, particles, and native words without a common kanji.12 Katakana grew out of Buddhist monks abbreviating fragments of man'yōgana characters in the same Heian period. It remained for foreign loanwords, onomatopoeia, scientific names, and emphasis.12
The modern functional division, kanji plus hiragana for native prose and katakana for foreign words, took its present shape after the 1946 reforms. Those reforms standardized both the kana inventory and the kana spelling rules.412
What hiragana is for today
Hiragana carries native Japanese words without a common kanji, grammatical particles, inflectional endings on verbs and adjectives, and the okurigana that follows kanji stems.182 Three quick examples, one for each load:
有難う18
"Thank you."
食べました18
"I ate."
大きかった18
"It was big."
This functional load is the teaching reason hiragana comes first. A beginner cannot read a sentence written only in kanji with no inflection visible, but can read a children's book written entirely in hiragana, including its Dakuten and Handakuten markings.18
A pointer forward
Katakana's parallel story uses the same man'yōgana raw material, simplified by a different mechanism (monk abbreviation rather than calligraphic cursivization). It ended up with a different functional load. It belongs in its own article rather than as a digression here.
Good to know
Each modern hiragana traces to a specific kanji
The parent-kanji set is consistent across reference works. あ comes from 安, い from 以, う from 宇, え from 衣, お from 於, か from 加, き from 幾, く from 久, け from 計, こ from 己, さ from 左, し from 之, す from 寸, せ from 世, そ from 曽, た from 太, ち from 知, つ from 川, て from 天, と from 止, な from 奈, に from 仁, ぬ from 奴, ね from 祢, の from 乃, は from 波, ひ from 比, ふ from 不, へ from 部, ほ from 保, ま from 末, み from 美, む from 武, め from 女, も from 毛, や from 也, ゆ from 由, よ from 与, ら from 良, り from 利, る from 留, れ from 礼, ろ from 呂, わ from 和, ゐ from 為, ゑ from 恵, を from 遠, and ん from 无.514
Seeing the cursive deformation of the parent kanji makes a hiragana shape feel motivated rather than arbitrary. It also prepares the learner for the same shape-derivation pattern when starting kanji proper.18
Why so many old shop signs look like broken hiragana
The variant kana shapes that look unfamiliar on traditional soba shop noren, yose theater signs, and old shopfronts are hentaigana retained from the pre-1900 multi-variant system, not handwriting errors. The most-cited example is the ki in kisoba, often written with a hentaigana form derived from 起 or 喜 rather than the now-standard き from 幾.11
Recognizing the category helps prevent the "is that broken hiragana?" reaction. It also explains why the same syllable can look different on different signs.11
The "women invented hiragana" claim, calibrated
The common short-form version of this claim is "Hiragana was invented by women in the Heian period." The accurate version is longer. Hiragana grew out of the cursive simplification of man'yōgana over the 9th and 10th centuries; women of the Heian court drove its literary maturation; the script was called onnade ("women's hand") because formal Chinese-classics education was reserved for men; men also used hiragana, including the four male compilers of the Kokin Wakashū.152910
What split between men and women was the social register (vernacular versus Chinese-style formal), not authorship of the script. The one-line claim is useful shorthand, but misleading as full history.15
Why を is still in the chart but ゐ and ゑ are not
All three characters represent sounds that had merged in standard Japanese centuries before 1946. を was retained in writing because it had a live grammatical job as the object-marker particle. Respelling every object phrase in the language would have been prohibitively disruptive. ゐ and ゑ had no such grammatical anchor and were retired.
The 1946 cabinet notification states explicitly that the particles は, を, and へ are kept in their historical spellings.47
Hiragana's dates worth remembering
Six dates carry the full arc. 471 CE: the Inariyama Sword inscription, the earliest dated long inscription in Japan using kanji as syllabograms for Japanese names.8 c. 759: the Man'yōshū compiled, the namesake corpus for man'yōgana.5 905: the Kokin Wakashū compiled under Emperor Daigo, the first imperial anthology in kana.9 c. 1000 to 1021: Murasaki Shikibu's Tale of Genji, the canonical high point of hiragana literature.10
1900: the Elementary School Order Enforcement Regulations standardize one hiragana per syllable, producing a 48-character set (the modern 46 plus ゐ and ゑ).311 1946: the "Gendai Kanazukai" Cabinet Notification No. 33 retires ゐ and ゑ from general use, leaving the modern 46.4127
Why this history is not a closed file
Unicode 10.0 (June 2017) added 285 hentaigana characters in the Kana Supplement (U+1B000–U+1B0FF) and Kana Extended-A (U+1B100–U+1B12F) blocks. Counting the single hentaigana letter added in Unicode 6.0, the Unicode total is 286. Variant kana are now typeable in digital text, and specialists use this capability to digitize Edo-period documents and bring hentaigana into limited contemporary use for branding and calligraphy.1117
See also
- Hiragana vs. Katakana: How to Tell Them Apart and Use Both
- Hiragana, Katakana, or Kanji First? A Beginner's Script Order
- How to Learn Japanese: The Complete Roadmap from Zero to Fluency
- Hiragana Mnemonics That Actually Work
- How to Practice Writing Hiragana: A Drill Plan, Free Sheets, and the Anki Hybrid
- Lookalike Hiragana: How to Tell the Most-Confused Kana Apart