The History of Katakana: From Heian Monks' Shorthand to the Modern 46-Character Set
Katakana began when 9th-century Buddhist monks shortened man'yōgana characters, kanji used for sound, to single radicals in margin notes on Chinese sutras. Those fragments became the angular phonetic script that modern Japanese reserves mostly for foreign words.123 For most of its thousand-year life, though, katakana did the job hiragana does today. A single cabinet notification in 1946 flipped the default.145
Overview
Why this history matters to a learner
Modern katakana looks like the loanword script. It handles コーヒー and コンピューター and, at first glance, not much else. For most of its thousand-year life, however, it filled the formal script role that hiragana plays today. It appeared in official documents, religious texts, and the kunten margin notes that produced it.15
The name itself encodes the origin. The kata (片) in 片仮名 means "one side" or "partial." Monks took one piece of a man'yōgana character rather than making the whole character cursive, and the script is named for that fragmenting move.12
Pre-1946 Japanese legal codes, military manuals, and primary-school readers were written in kanji-plus-katakana, not kanji-plus-hiragana. The English Wikipedia entry on katakana is direct about it: "Pre–World War II official documents mix katakana and kanji in the same way that hiragana and kanji are mixed in modern Japanese texts, that is, katakana were used for okurigana and particles such as wa or o."1 The 1946 reform flipped that default. The modern role split is what remains.145
A one-paragraph timeline
Man'yōgana, kanji used for phonetic rather than semantic value, is the dominant convention of the Man'yōshū in the 8th century.6 In the early-to-mid 9th century, Buddhist monks in Nara and Kyoto abbreviate man'yōgana into kunten margin notes on Chinese sutras. These abbreviated fragments are the proto-katakana.123
By around 951 CE katakana is attested as an independent writing system, and the modern set is codified over the late Heian.13 On 11 February 1889 the Meiji Constitution is promulgated in kanji-and-katakana mixed orthography.78 Through 1945 legal codes, military manuals, government circulars, and primary-school readers use kanji-plus-katakana as the formal default.15 On 16 November 1946 the cabinet adopts 現代かなづかい (gendai kanazukai) as Cabinet Notification No. 33. ヰ and ヱ retire from general use, and the same-day Tōyō Kanji list anchors the broader reform.4910 In 1991 a second cabinet notification standardises the extended-katakana inventory for loanword transcription (ファ, ヴィ, ティ).11
The arc is easier to scan as a diagram than as a sentence.
What this article does not cover
Hiragana's parallel lineage uses the same man'yōgana raw material but a different mechanism: making the whole character cursive rather than lifting a fragment. It also ran through a different social register, the Heian court rather than the temple. It is the sibling article, not a substep here.1213
Before katakana: how Japanese got written at all
Japanese had a language but not a script
Old Japanese was a spoken language with no native script. Chinese characters reached Japan from the continent across the 5th and 6th centuries CE, through the Korean peninsula and direct contact with China.1213
Japan adopted kanji for both their semantic value, to write meaning, and, more awkwardly, their phonetic value, to write sound. The second use is the move that mattered for both kana.12136
Man'yōgana: borrowing kanji for sound, not meaning
Man'yōgana is "an ancient writing system that uses Chinese characters to represent the Japanese language" by sound rather than meaning. Scribes picked a kanji whose Chinese-derived pronunciation matched the target Japanese syllable and ignored the kanji's meaning.6
The system takes its name from the Man'yōshū (万葉集), the Nara-period poetry anthology compiled around 759 CE in which it is the dominant convention.6 By the end of the 8th century, 970 kanji were in use to represent the 90 morae of Japanese. There was no one-to-one mapping; different scribes picked different parent kanji for the same syllable.6
Man'yōgana was the shared raw material from which both kana later emerged.12136
Why this was unworkable for daily writing
Writing a single Japanese syllable required producing a stroke-heavy kanji. For prose of any length, the cost in time and ink was prohibitive. Scribes turned to abbreviation.1213
Two communities responded with two different simplifications, working in parallel on the same source.
Same source material; different mechanisms.12163
The birth of katakana: monks, margins, and shorthand
Why monks needed a faster script
Buddhist scholars in 9th-century Japan studied sutras written entirely in Classical Chinese. To read them aloud in Japanese, they jotted sound and grammar hints in the margins and between the columns. This annotation system is called kunten (訓点).12143
Kunten are "guiding marks for rendering Chinese into Japanese." They include kaeriten (返り点) to reorder characters into Japanese SOV (subject-object-verb) order, okototen (乎古止点) dots to mark Japanese inflections, and small-form kana to gloss Chinese characters with their Japanese reading.14
Full-form man'yōgana characters did not fit the narrow margins. They were also too slow to write at the pace of reading. Monks abbreviated them on the page, and the abbreviated fragments are the proto-katakana.123
Katakana's signature angularity is not a stylistic choice. It is what results from writing a kanji fragment fast, in a narrow margin, with no time to round the strokes. Hiragana shapes are smooth because they grew out of leisurely letter-writing. Katakana shapes are angular because they grew out of reading notes.123
The kata (片, "partial") mechanism
Katakana developed from "parts of kanji," using fragments from annotated texts rather than full character simplifications.3
The word kata (片) means "one side" or "partial." Kana (仮名) means "borrowed name" or, by extension, "phonetic script." Together, 片仮名 names the mechanism: fragments of man'yōgana, not whole characters made cursive.12
The mechanism is structurally different from hiragana's. Hiragana smoothed the whole parent kanji into a single cursive stroke. Katakana lifted a radical, a top, a left side, or a centre block and used it as the whole sign.1213
Worked examples of the parent-kanji lineage
Each modern katakana traces to one or more parent kanji called the jibo (字母), or "character mother," of the character. The clearest way to see the fragmenting move is to look at five examples side by side, each with the piece that became the kana:
| Katakana | Parent kanji (jibo) | Sino-Japanese reading | Piece taken |
|---|---|---|---|
| カ | 加 | ka | left side |
| イ | 伊 | i | left-side radical |
| ウ | 宇 | u | top portion |
| レ | 礼 (kyūjitai 禮) | rei | right portion |
| タ | 多 | ta | upper half |
The reading of the parent kanji is the reason it was chosen. 加 was picked for /ka/ because its Sino-Japanese reading begins with that syllable. The meaning ("increase") was irrelevant to the borrowing.16 The English Wikipedia entry is explicit on the カ example: "ka (カ) comes from the left side of ka (加; 'increase')."1
Two characters in the modern set, ツ and ン, sit outside this lineage. ツ "appears fully formed in the earliest documents available," and the Japanese Wikipedia entry on 片仮名 lists multiple candidate parents (州, 門, 津) without settling on one.23 ン is sometimes traced to 尓 or 爾. It is also sometimes treated as a stroke-cluster invention to mark the syllable-final nasal that katakana otherwise lacked. The syllable /n/ itself only emerged in Japanese around 1000 CE, after the rest of the katakana set was already in use.23
Who actually did this
Popular accounts credit the monk Kūkai (空海, 774 to 835), founder of Shingon Buddhism, and sometimes Saichō (最澄, 767 to 822), founder of Tendai Buddhism.1516
The single-inventor framing is folklore. The Japanese Wikipedia entry on 片仮名 calls the Kūkai attribution "no more than folklore" (民間伝承の域を出ない). The English Wikipedia entry on Kūkai labels the kana-invention claim "one such legend" alongside the Iroha poem.215 The qualified version is that katakana emerged from a community practice in temple schools across the early-to-mid Heian period, the 9th to 11th centuries. It did not come from a single dated act of invention, and the modern set was codified over the late Heian.123
Kūkai and Saichō both led early-Heian Buddhist establishments where the kunten margin-note practice flourished: Shingon at Mount Kōya and Tendai at Mount Hiei. That is why their names attach to the origin story.1516
What katakana was for: a thousand years as the formal script
Two kana, two registers
From the late Heian through the Kamakura, Muromachi, Edo, and Meiji periods, hiragana and katakana co-existed with a register split rather than a function split. Hiragana was the script of private letters, poetry, and vernacular literature. Katakana was the script of scholarship, religious texts, dictionaries, and formal annotation.15
Britannica states the contrast directly: scribes "who continued their education to become scribes or court writers then learned the more squared katakana, which was used for official documents and religious texts," and "katakana was considered more formal than hiragana."5
The split was roughly like cursive versus printed alphabets in the West, but with sharper social weight.
Kanji-plus-katakana as official Japanese
In formal documents, katakana played the role hiragana plays today. It carried okurigana (verb and adjective inflection), particles, and grammatical glue. Kanji carried the meaning-bearing roots. The style has a name: 漢字片仮名交じり文 (kanji katakana majiribun, "kanji-katakana mixed sentence").12
The English Wikipedia entry on katakana is explicit on the pre-war role: "Pre–World War II official documents mix katakana and kanji in the same way that hiragana and kanji are mixed in modern Japanese texts, that is, katakana were used for okurigana and particles such as wa or o."1 A separate line on the same page makes the scope clear: "Official documents of the Empire of Japan were written exclusively with kyūjitai and katakana."1
The style appears on pre-war legal codes, military manuals, government circulars, and primary-school readers.125
The Meiji-era expansion
The Constitution of the Empire of Japan (大日本帝國憲法) was promulgated on 11 February 1889 and came into effect on 29 November 1890.7 The original document was drafted in kanji-plus-katakana mixed orthography. Article 1 reads, verbatim:
大日本帝國ハ萬世一系ノ天皇之ヲ統治ス。8
"The Empire of Japan shall be reigned over and governed by a line of Emperors unbroken for ages eternal."
The katakana ハ, ノ, ヲ carry the topic, genitive, and accusative particles, and ス carries the classical verb ending. The kanji carry the meaning-bearing roots. The load is exactly what hiragana carries in modern text.8
The wider Meiji-era institutional rebuild followed the same writing convention. Civil and criminal codes, Army and Navy regulations, government circulars, and primary-school readers were drafted and printed in this style. Military telegrams and field manuals through 1945 also remained in katakana plus kanji.15
Why katakana, not hiragana, for formal text
The choice followed inherited register, not deliberate policy. Katakana inherited the seriousness of its temple-scholarship and kanbun-annotation roots. Hiragana inherited the intimacy of its private-letter and vernacular-literature roots.15
When the Meiji state needed a "serious" script for the new codes and regulations, it reached for the one already associated with scholarly and official writing.5
The 1946 reform: how the default flipped
What gendai kanazukai changed
On 16 November 1946 (Shōwa 21), the Japanese cabinet adopted 現代かなづかい (gendai kanazukai, "modern kana usage") as Cabinet Notification No. 33 (内閣告示第33号).49
The notification respelled words to match modern pronunciation. Historical spellings changed: けふ (kefu, "today") became きょう (kyō); かは (kaha, "river") became かわ (kawa); みづ (midu, "water") became みず (mizu).179
The reform applied to both hiragana and katakana, but its larger structural effect was a register flip. Official documents adopted kanji-plus-hiragana, the everyday register. Kanji-plus-katakana retreated to specialist contexts.15
Why hiragana won the default slot
The 1946 changes ran alongside the Tōyō Kanji list (当用漢字表), 1,850 characters, released by the Japanese Ministry of Education on the same day, 16 November 1946.10
The broader push was an Allied-occupation-era literacy programme. Pre-war reform planning had been on the table since at least the 1920s. Post-war momentum, including the March 1946 American Education Delegation report advocating the use of rōmaji, was political, not linguistic.4
Aligning formal text with hiragana, the vernacular script ordinary readers already knew, was a literacy decision more than a linguistic one.45
ヰ and ヱ retire from katakana too
Katakana, like hiragana, had a wi (ヰ) and a we (ヱ).117 In standard Japanese, the sounds those characters represented had merged with /i/ and /e/ centuries before the reform.
The Wikipedia entry on Japanese script reform puts the policy plainly: "two kana, ゐ/ヰ wi and ゑ/ヱ we, were officially declared obsolete, as the pronunciations they represented had dropped from the language many centuries before."4
The 1946 notification removed both forms from general use. It replaced them with イ and エ in word spellings and left the older forms for proper names, archaic citations, and stylistic effect. The modern 46-character katakana base set is the 1946 result.1749
What ヲ kept that ヰ and ヱ lost
ヲ (wo), like its hiragana counterpart を, represented a sound that had merged with /o/ in standard Japanese centuries before the reform.179
The 1946 notification preserves は, を, and へ in their historical forms when used as grammatical particles: the topic, accusative, and direction markers. The cited rationale is that respelling these very common particle spellings "would unnecessarily confuse readers."179
The katakana ヲ stays in the chart for completeness and survives for the rare cases where a name or stylised text uses it. In modern katakana, though, it is essentially a historical placeholder. The live grammatical job is carried by hiragana を, not katakana ヲ.19
Phonetically, ヲ is in the same position as ヰ and ヱ: a katakana for a sound that merged out of standard Japanese centuries ago. It survived on the chart because its hiragana twin を holds the object-marker particle slot, and the chart is kept symmetrical with hiragana's. The grammatical work happens on the hiragana side.19
What katakana is for today
The four modern jobs
The pre-war formal-register role is gone, and a tighter functional split has taken its place. Modern katakana does four jobs.
Foreign loanwords are the most visible: コーヒー (kōhī, "coffee"), テレビ (terebi, "television"), コンピューター (konpyūtā, "computer"). The English Wikipedia gloss is "transcription of words from foreign languages or loanwords."15
Foreign personal and place names take katakana as a default: ジョン (Jon, "John"), ニューヨーク (Nyūyōku, "New York").1
Onomatopoeia and mimetics, sound-symbolic words, use it for visual weight: ワンワン (wan-wan, "woof-woof"), ドキドキ (dokidoki, "thump-thump / heartbeat").1
Emphasis and stylisation are the fourth role. Katakana functions "similarly to italics in Western languages," picking out a native word for visual weight where alphabetic scripts would italicise.15 Scientific binomial nomenclature for species names (ヒト, イヌ, ネコ in zoological prose) is the durable specialist case.1
Notes on legacy uses
Katakana was the script of Japanese telegrams "before 1988"; the practice retired with the wider telegram service.1
Extended katakana combinations for loanword sounds outside the historical inventory (ファ, フィ, フェ, フォ, the ヴ-row, ティ, トゥ, ディ, ドゥ, ウィ, ウェ, ウォ, and more) were codified in Cabinet Notification No. 2 「外来語の表記」 ("Notation of Foreign Loanwords") of 28 June 1991.11 The National Language Council (国語審議会) prepared the notification, recommended it on 7 February 1991, and it was issued as a cabinet notification on 28 June 1991.11
Why the flip stuck
Once kanji-plus-hiragana became the official script in 1946, the visual contrast between hiragana and katakana became a built-in marker for readers. Hiragana signals native or grammatical text; katakana signals foreign or marked text.15
Japanese has no spaces between words and no italics. The two-kana split does some of the work that spaces and italics do in alphabetic scripts.1
A pointer back to hiragana
Hiragana's parallel story uses the same man'yōgana raw material but a different mechanism: court cursivisation of the whole character rather than monk abbreviation of a fragment. It also ran through a different social register: the onnade "women's hand" tradition of the Heian court. The 1900 Elementary School Order Enforcement Regulations and the same 1946 reform that retired ヰ and ヱ shaped its modern 46-character set.1213 It is the sibling article rather than a digression here.
Good to know
Why katakana literally means "partial kana"
片 (kata) means "one side" or "partial." 仮名 (kana) means "borrowed name" or "phonetic script." Together, 片仮名 names the mechanism that produced the script: fragments of man'yōgana, not whole characters made cursive.12
The name is a built-in mnemonic for the origin story. Once you see that カ is the left side of 加 and that ヌ is the right side of 奴, the angular shapes feel less arbitrary.118
The Kūkai-and-Saichō attribution, calibrated
Popular sources name Kūkai (空海, 774 to 835) or Saichō (最澄, 767 to 822) as the inventor of katakana.1516 The Japanese Wikipedia entry on 片仮名 treats single-inventor attribution as folklore (民間伝承の域を出ない). The English Wikipedia entry on Kūkai labels the kana-invention attribution "one such legend" alongside the Iroha poem.215
Both monks led early-Heian Buddhist establishments where the kunten margin-note practice flourished. The script emerged from that community over the 9th to 11th centuries rather than from a single dated act of invention.123
ツ and ン arrived without a clean parent
Most katakana have a confidently identified parent kanji. ツ and ン do not.23
ツ "appears fully formed in the earliest documents available," and the Japanese Wikipedia entry on 片仮名 lists multiple candidates (州, 門, 津) without settling on one.23 ン is sometimes traced to 尓 or 爾, and sometimes treated as a stroke-cluster invention to mark the syllable-final nasal that katakana otherwise lacked; the syllable /n/ itself emerged in Japanese around 1000 CE, after the rest of the katakana set was already in use.23
This is worth knowing if a learner ever meets two parent-kanji charts that disagree on these cells.
Why pre-war Japanese text looks "wrong"
A learner who picks up a pre-1946 legal code, military manual, or imperial rescript meets sentences like 大日本帝國ハ萬世一系ノ天皇之ヲ統治ス. In that sentence, katakana (ハ, ノ, ヲ, ス) does every grammatical job that hiragana does today.78
This is not stylised or archaic for effect. The English Wikipedia entry on katakana puts it plainly: "Pre–World War II official documents mix katakana and kanji in the same way that hiragana and kanji are mixed in modern Japanese texts."1 The modern eye reads such text as alien only because of the 1946 register flip.15
Two reforms, one default flip
The 1946 changes are usually told as a hiragana story: ゐ and ゑ retire, and gendai kanazukai cleans up historical spellings.179
The katakana side is the bigger structural change. The kana set lost ヰ and ヱ on the same day, and katakana also lost its formal-register role. The kanji-plus-katakana orthography of legal and military documents was displaced by kanji-plus-hiragana.149 The hiragana article tells half the story. This one tells the other half.
Katakana's dates worth remembering
Four dates carry the full arc:
9th century: Buddhist monks abbreviate man'yōgana into kunten margin notes.123
Late Heian (10th to 12th century): the modern katakana set is codified.13
1889 (11 February): the Meiji Constitution is promulgated in kanji-and-katakana.78
1946 (16 November): Cabinet Notification No. 33 (gendai kanazukai) flips the default to hiragana and retires ヰ and ヱ.49
A live tail in modern usage
Katakana picked up new jobs after losing its formal-register role, so the script's history is not a closed file. The extended-katakana inventory (ファ, ヴィ, ティ) was standardised by Cabinet Notification No. 2 「外来語の表記」 on 28 June 1991.11
Katakana was the script of Japanese telegrams "before 1988," when the practice retired with the wider telegram service.1 Signage, scientific binomial nomenclature, and brand-voice stylisation kept some pre-war katakana habits.15
See also
- Mixed Script: How Japanese Combines Kanji, Hiragana, and Katakana
- Hiragana, Katakana, or Kanji First? A Beginner's Script Order
- Katakana Mnemonics That Actually Work
- Long Vowels in Katakana: How the Chōonpu ー Works and Why Hiragana Doesn't Use It
- The Jōyō Kanji List (常用漢字): The 2,136-Character Set Explained
- The Complete Hiragana Chart (Gojūon): How to Read All 46 Base Kana, Dakuten, and Yōon