The Complete Hiragana Chart (Gojūon): How to Read All 46 Base Kana, Dakuten, and Yōon
The hiragana chart is known in Japanese as the gojūon (五十音, "fifty sounds"). It is the standard phonetic order of the kana, shown as a 5×10 grid.1 Anyone learning to read Japanese meets this chart on day one and returns to it throughout their study. It pays to understand the grid as a system, not as 46 isolated shapes to memorize.
Overview
What the chart is in one paragraph
The gojūon is a 5×10 phonetic grid: five vowel columns crossed with ten consonant rows.1 Each filled cell is one mora, the smallest unit of timing in Japanese, read as the row's consonant plus the column's vowel.2 The layout was not invented to teach beginners. It is a phonological index, a sound-based ordering system, whose order comes from Sanskrit Siddhaṃ phonology and Brahmi script, with parallels to the Chinese fanqie system.1
The three tables every chart includes
A complete hiragana chart is really three reference tables used together. The base gojūon holds 46 distinct kana in modern Japanese.1 Dakuten and handakuten extensions add 25 more by marking voicing on the k, s, t, and h rows.3 Yōon, the contracted sounds, add 33 palatalized combinations built from i-column kana plus a small ゃ, ゅ, or ょ.4
The arithmetic is therefore 46 + 25 + 33 = 104. That total is the modern functional inventory: every distinct phonetic unit a learner needs to read present-day Japanese.
The 104-character total is an accounting tool, not a memorization target. The base 46 do the heaviest lifting. The 25 voiced kana are systematic transformations of rows the reader already knows, and the 33 yōon are predictable combinations of two familiar characters. Frame the chart as 46 base shapes plus two layers of rules, not as 104 separate things.
Audience and prerequisites
This page is pre-N5. It covers the foundational layer the Japan Foundation and Japan Educational Exchanges and Services explicitly assume a learner has mastered before taking the N5 exam: "One is able to read and understand typical expressions and sentences written in hiragana, katakana, and basic kanji."5 The chart itself is therefore not a JLPT-graded grammar point. It is the prerequisite the candidate brings into N5.
This page assumes you have already decided to study Japanese and have chosen an order for tackling the three scripts. For that earlier decision, see "Hiragana, Katakana, or Kanji First? A Beginner's Script Order".
How to read the gojūon chart
Vowel columns: a, i, u, e, o
Five vowel columns form the chart's first axis.1 Every kana in a given column ends in the same vowel sound. The column header (あ・い・う・え・お) also serves as the pure-vowel row when the consonant slot is empty.1
The single short example below shows the rule applied to the a-column.
赤い1
"Red."
Consonant rows: ∅, k, s, t, n, h, m, y, r, w
Ten consonant rows form the second axis. Row one is the pure-vowel row, with no leading consonant. The remaining nine begin with k, s, t, n, h, m, y, r, w.1 The ordering is not alphabetical and not arbitrary: the consonants are arranged from the back of the mouth toward the front, velar to labial.1
Anyone who has met IPA charts has already learned this principle: consonants group by where in the mouth they are articulated. The gojūon's row order is k (velar), s and t (alveolar), n (alveolar nasal), h (originally bilabial), m (bilabial nasal), then the approximants y, r, w. Say the row order aloud and feel the tongue move from back to front. That turns rote recitation into a brief anatomy lesson.
Reading any cell: consonant + vowel = mora
The intersection of a consonant row and a vowel column yields one mora, the unit of timing in Japanese.2 To read な, take the n-row and the a-column: /na/. To read し, take the s-row and the i-column. The result is /ɕi/, romanized as shi in Modified Hepburn rather than the row-expected si.1
This last point is one of four cells where the romaji diverges from the row pattern. The full set (し, ち, つ, ふ) is covered below.
The special character: ん (syllabic n)
ん is the only kana that lies outside the 5×10 grid. Wikipedia's Hiragana article puts it plainly: "Strictly speaking, the singular consonant ん (n) is considered to be outside the gojūon."6 It represents a moraic nasal, the one kana that does not end in a vowel.7
ん is nonetheless counted in the 46. It was officially recognized as a single standard kana in the 1900 Elementary School Order. The same reform retired the hentaigana variants discussed later on this page.7
The base gojūon table: 46 characters
The full 5×10 grid
| Row | a | i | u | e | o |
|---|---|---|---|---|---|
| ∅ | あ (a) | い (i) | う (u) | え (e) | お (o) |
| k | か (ka) | き (ki) | く (ku) | け (ke) | こ (ko) |
| s | さ (sa) | し (shi) | す (su) | せ (se) | そ (so) |
| t | た (ta) | ち (chi) | つ (tsu) | て (te) | と (to) |
| n | な (na) | に (ni) | ぬ (nu) | ね (ne) | の (no) |
| h | は (ha) | ひ (hi) | ふ (fu) | へ (he) | ほ (ho) |
| m | ま (ma) | み (mi) | む (mu) | め (me) | も (mo) |
| y | や (ya) | (yi) | ゆ (yu) | (ye) | よ (yo) |
| r | ら (ra) | り (ri) | る (ru) | れ (re) | ろ (ro) |
| w | わ (wa) | (wi) | (wu) | (we) | を (wo / o) |
Plus the extra-grid kana: ん (n).17
Romaji on this table uses Modified Hepburn throughout: shi, chi, tsu, fu, wo / o. を is conventionally written wo when cited as a kana. It is written o when used as the object-marking particle, reflecting how the particle is actually pronounced.8
The five gaps and one extra: 50 − 5 + 1 = 46
Five cells on the 5×10 grid are blank, and one extra character is added back from outside the grid. The arithmetic 50 − 5 + 1 = 46 is the cleanest way to remember the inventory. Each missing cell has its own history.
- yi, wu never occurred as distinct sounds in Old Japanese: "there was no yi or wu even in Old Japanese."1
- ye existed in Old Japanese and had its own kana, but disappeared in Early Middle Japanese, having merged with e.1
- wi (ゐ) was in active written use until the modern era, although its pronunciation merged with い by the Kamakura period in the 13th century.9
- we (ゑ) was likewise in active written use until the modern era, with its pronunciation merging with え between the Kamakura and Taishō periods.10
- ゐ and ゑ were officially retired in 1946 by Cabinet Notification No. 33, 内閣告示第33号「現代かなづかい」.1112 The 1986 revision 「現代仮名遣い」 (内閣告示第1号) is the standing reference for modern kana usage.13
- ん is added back: 50 − 5 + 1 = 46.17
Irregular romanizations to flag (shi, chi, tsu, fu)
Four cells of the base grid use romaji that breaks the row's expected pattern: し is shi rather than si, ち is chi rather than ti, つ is tsu rather than tu, and ふ is fu rather than hu.1 Modified Hepburn writes what is spoken. These four cells have shifted phonetically away from the bare consonant their row implies.
The four irregular cells reflect actual modern pronunciation rather than spelling-system inconsistency. し is palatalized /ɕi/, ち is the affricate /t͡ɕi/, つ is the affricate /t͡sɯ/, and ふ has the bilabial fricative /ɸɯ/. The deeper phonetic treatment, including the alternative Kunrei-shiki spellings si, ti, tu, hu, belongs to the consonant inventory in the pronunciation pillar. For chart-reading purposes, treat the four cells as memorized exceptions.
Dakuten and handakuten: 25 voiced kana
The voicing transformation
Dakuten (゛) is a diacritic, a small mark added to a character. It adds two short strokes to the upper right of a kana and shifts its consonant from voiceless to voiced. The transformations are systematic across four rows: か→が (k→g), さ→ざ (s→z), た→だ (t→d), and は→ば (h→b).3 Handakuten (゜) is a small circle in the same position, used only on the h-row to mark /p/: は→ぱ.3
The count is four dakuten rows times five vowels (20 kana), plus one handakuten row times five vowels (5 kana). That gives 25 derived characters in total.3
Why handakuten exists only on the h-row
Handakuten is asymmetric on purpose. The h-row historically descends from a /p/ series in Old Japanese that lenited, or softened, through [ɸ] to its modern [h]. The handakuten was introduced, including by Portuguese Jesuit transcribers in the Rakuyōshū, to recover the /p/ distinction in writing.3 Other rows never needed a similar mark because their voicing distinctions were already covered by dakuten or did not exist as a meaningful contrast.
The full historical treatment belongs to the dedicated dakuten and handakuten article.
The dakuten/handakuten table
| Row | a | i | u | e | o |
|---|---|---|---|---|---|
| g (k+゛) | が (ga) | ぎ (gi) | ぐ (gu) | げ (ge) | ご (go) |
| z (s+゛) | ざ (za) | じ (ji) | ず (zu) | ぜ (ze) | ぞ (zo) |
| d (t+゛) | だ (da) | ぢ (ji) | づ (zu) | で (de) | ど (do) |
| b (h+゛) | ば (ba) | び (bi) | ぶ (bu) | べ (be) | ぼ (bo) |
| p (h+゜) | ぱ (pa) | ぴ (pi) | ぷ (pu) | ぺ (pe) | ぽ (po) |
ぢ and づ are pronounced identically to じ and ず in modern standard Japanese. The 1986 「現代仮名遣い」 restricts their use to specific etymological contexts: the rendaku voicing of ち and つ in compounds (はなぢ from はな + ち), and repeated-mora compounds where a kana is immediately followed by its own voiced form (つづく).136
Yōon: 33 contracted sounds
How yōon combine
A yōon (拗音) is a mora formed with an added [j] sound. In other words, it is a palatalized mora.4 In writing, yōon collapses two characters into one mora: any i-column kana plus a small ゃ, ゅ, or ょ.42
The i-column kana that take yōon are the seven base hosts き, し, ち, に, ひ, み, り, plus the five voiced and handakuten variants ぎ, じ, ぢ, び, ぴ.4 The small ゃ/ゅ/ょ "do not represent a mora by themselves and attach to other kana." The resulting two-kana cluster counts as a single mora.2
東京4
"Tokyo."
東京 is four morae, to-u-kyo-u. That is true even though the romaji has six letters and the kana spell out five characters. きょ is one yōon mora, not two.
The yōon table
| Host | + ゃ | + ゅ | + ょ |
|---|---|---|---|
| き | きゃ (kya) | きゅ (kyu) | きょ (kyo) |
| し | しゃ (sha) | しゅ (shu) | しょ (sho) |
| ち | ちゃ (cha) | ちゅ (chu) | ちょ (cho) |
| に | にゃ (nya) | にゅ (nyu) | にょ (nyo) |
| ひ | ひゃ (hya) | ひゅ (hyu) | ひょ (hyo) |
| み | みゃ (mya) | みゅ (myu) | みょ (myo) |
| り | りゃ (rya) | りゅ (ryu) | りょ (ryo) |
| ぎ | ぎゃ (gya) | ぎゅ (gyu) | ぎょ (gyo) |
| じ | じゃ (ja) | じゅ (ju) | じょ (jo) |
| び | びゃ (bya) | びゅ (byu) | びょ (byo) |
| ぴ | ぴゃ (pya) | ぴゅ (pyu) | ぴょ (pyo) |
The count is 11 hosts × 3 small kana = 33 yōon combinations.4 ぢゃ, ぢゅ, and ぢょ appear as a row in some historical charts. However, they are pronounced identically to じゃ, じゅ, and じょ, and 「現代仮名遣い」 restricts ぢ to specific contexts. The canonical pedagogical inventory therefore lists 33 yōon rather than 36.13
Why つ + や ≠ つや
The visual cue is the physical size of the second character. つや is two full-sized characters and therefore two morae, tsu + ya. A hypothetical small-ゃ form would be one mora. But small ゃ does not attach to つ in modern Japanese because つ sits in the u-column, and yōon hosts are i-column kana only.4 The same size-of-the-second-kana rule distinguishes しゅ (one mora) from しゆ (two morae) and きょ (one mora) from きよ (two morae).2
The full inventory: 46 + 25 + 33 = 104
What "complete" means here
"Complete" on this page means the modern functional inventory: 46 base gojūon, 25 dakuten and handakuten extensions, and 33 yōon combinations.134 The total, 104, is every distinct phonetic unit a learner needs to read present-day Japanese.
Two further marks count as morae when read, but they do not add new characters to the inventory. The sokuon, written as a small っ to mark gemination, is a positional use of an existing kana.2 Long-vowel marks in hiragana are formed by repeating or appending a vowel kana (おう, ええ, おお), not by adding new characters. The orthography is governed by 「現代仮名遣い」.13
What this article does not cover
This page is a reference chart, not the full hiragana curriculum. Several topics have dedicated articles in this subcategory: stroke order, mnemonics for memorizing individual shapes, the lookalike pairs that confuse beginners once the chart is closed, the phonetic deep-dive into mora versus syllable and shi versus si, the small つ in detail, long-vowel orthography, and the three particle spelling exceptions (は, へ, を when used as particles).
五十音 vs. 46: the historical inventory
The original count: ゐ, ゑ, and the three phantom cells
The earliest documented gojūon-style layout is the Kujakukyō Ongi manuscript, dated c. 1004–1028.1 By the time the ordering was devised, yi, ye, and wu were already phonetically absent or had already merged. Wikipedia summarizes ye this way: "Ye persisted long enough for kana to be developed for it, but disappeared in Early Middle Japanese, having merged with e."1
ゐ and ゑ remained in active written use until the modern reforms, even though their pronunciations had merged centuries earlier. ゐ merged with い by the 13th century during the Kamakura period. ゑ merged with え between the Kamakura and Taishō periods.910 The pre-reform written inventory of base kana was therefore 48 (46 + ゐ + ゑ).1910
The National Institute for Japanese Language and Linguistics (NINJAL) explains the count discrepancy directly: "アイウエオの総数は「五十」ではありません", meaning the total of aiueo is not "fifty." The charts contain blank cells and duplicated characters.14 NINJAL also notes that the final ん was added to the chart from the Meiji period onward for educational use.14
The 1946 reform: gendai kanazukai
「現代かなづかい」 was promulgated on 16 November 1946 as 内閣告示第33号 (Cabinet Notification No. 33 of Shōwa 21).1112 The 当用漢字 (Tōyō kanji) list was promulgated alongside it as part of the same post-war spelling reform package.11
The reform replaced 歴史的仮名遣い (historical kana orthography) with a phonetic-spelling principle: kana usage was now expected to reflect spoken pronunciation.8 ゐ and ゑ were retired in the process. Wikipedia summarizes the change as "this kana was deemed obsolete in Japanese with the orthographic reforms of 1946, to be replaced by い/イ in all contexts," with the parallel rule applying to ゑ → え.910
Three particle spelling exceptions were retained, and they remain standard: "the three grammatical particles o, e, wa continue to be written as を wo, へ he, and は ha instead of お o, え e, and わ wa."8
The chart shows は in the h-row, with chart value ha. It shows へ in the h-row with chart value he. It shows を in the w-row with chart value wo. When the same characters serve as grammatical particles, they are pronounced wa, e, and o respectively, even though their written shape never changes. This is the only spelling exception the 1946 and 1986 reforms preserved, and it often trips up chart-aware learners. Full treatment lives in the dedicated article on the three hiragana spelling exceptions.
The 1946 notification was superseded in revised form by 「現代仮名遣い」 (内閣告示第1号), promulgated on 1 July 1986. The 1986 revision is the standing legal reference for modern kana usage, retaining both the ゐ/ゑ retirement and the は・へ・を particle exceptions.13
Where ゐ and ゑ still appear today
Although officially retired, ゐ and ゑ persist as a stylistic choice in proper nouns and brand names. The retired kana also turn up in some literary reprints, certificates issued by classical institutions, and stylized media. For example, Rebuild of Evangelion uses ヱ and ヲ in place of エ and オ.10 A reader is expected to recognize the characters on sight but not to write them.
ヱビスビール10
"Yebisu beer."
ニッカウヰスキー9
"Nikka Whisky."
よゐこ9
"Yoiko (comedy duo name)."
Hentaigana, briefly
Before the 1900 Elementary School Order (小学校令施行規則), a single hiragana sound could be written with several distinct shapes, each derived from a different man'yōgana source character.15 The 1900 order standardized one kana shape per sound. The retired variants are now collectively known as 変体仮名 (hentaigana).156
Hentaigana still appear on traditional shop signs, classical institutional certificates, and aesthetic calligraphy, with soba restaurants as the canonical example.15 Full historical treatment, including the man'yōgana lineage and the 1900 reform context, belongs to the dedicated history article.
Good to know
Why "fifty sounds" when the chart shows 46
The name 五十音 reflects the 5 × 10 = 50-cell layout, not the number of distinct sounds. NINJAL's Q&A on the count puts it explicitly: charts contain blank cells and duplicated characters, and the total of aiueo is not fifty.14 The name nonetheless survives in Japanese dictionaries and language references because the ordering functions like alphabetical order. 五十音順 (gojūon-jun) is the standing default for indexing names and entries, even though the count has been 46 since the 1946 reform.14
The chart is a phonetic index, not a difficulty curve
A frequent beginner mistake is to start at the top-left of the chart and march down, treating row order as study order. The chart's ordering is phonological, with the vowel and consonant axes inherited from Sanskrit Siddhaṃ phonology. It was never designed as a learning curve.1 A separate 47-character poetic ordering, the iroha (いろは), historically served as the sequence-learning order. Gojūon-jun is the dictionary order. Neither is a difficulty ranking.1 The recommended memorization order, mnemonic-led, lives in the dedicated hiragana mnemonics article.
Romaji on a chart is a pronunciation key, not a script
The romaji column on a hiragana chart exists to bootstrap pronunciation while the learner is reading the chart for the first time. Once a learner can decode the chart unaided, romaji has done its job and should drop away. The four cells that diverge from row-expected romaji (し, ち, つ, ふ) reflect actual modern pronunciation rather than spelling-system inconsistency. The upstream decision about when to leave romaji behind belongs to "Hiragana, Katakana, or Kanji First? A Beginner's Script Order".1
The mora, not the syllable, is the unit a chart counts
Each kana cell on the chart represents one mora, the smallest unit of timing in Japanese. Each yōon cluster (two written characters) is also one mora.2 The small ゃ, ゅ, and ょ "do not represent a mora by themselves and attach to other kana," which is why きょ in 東京 counts as one mora rather than two.2 The syllabic ん and the sokuon っ likewise count as one mora each, despite being voiceless or vowel-less. This is part of what distinguishes Japanese moraic timing from English syllable timing.2
ゐ and ゑ in brand names and proper nouns
Recognizing the retired kana matters because they continue to appear on labels, signs, and titles a learner will encounter from beginner level onward. Documented examples include ニッカウヰスキー (Nikka Whisky), よゐこ (the comedy duo Yoiko), 京都ゑびす神社 (Kyoto Ebisu Shrine), and ヱビス (Yebisu beer, written in katakana but pronounced ebisu).910 A reader who has decoded the 46 base kana can read these correctly by treating ゐ as い and ゑ as え. Writing them in one's own Japanese is not expected.
See also
- The Complete Katakana Chart (Gojūon): How to Read All 46 Base Kana, Dakuten, and Yōon
- How to Practice Writing Hiragana: A Drill Plan, Free Sheets, and the Anki Hybrid
- Hiragana vs. Katakana: How to Tell Them Apart and Use Both
- How to Learn Japanese: The Complete Roadmap from Zero to Fluency