The complete katakana script consists of 48 characters, not counting functional and diacritic marks:
- 5 nucleus vowels
- 42 core or body (onset-nucleus) syllabograms, consisting of nine consonants in combination with each of the five vowels, of which three possible combinations (yi, ye, wu) are not canonical
- 1 coda consonant
These are conceived as a 5×10 grid (gojūon, 五十音, literally "fifty sounds"), as shown in the adjacent table, read ア (a), イ (i), ウ (u), エ (e), オ (o), カ (ka), キ (ki), ク (ku), ケ (ke), コ (ko) and so on. The gojūon inherits its vowel and consonant order from Sanskrit practice. In vertical text contexts, which used to be the default case, the grid is usually presented as 10 columns by 5 rows, with vowels on the right hand side and ア (a) on top. Katakana glyphs in the same row or column do not share common graphic characteristics. Three of the syllabograms to be expected, yi, ye and wu, may have been used idiosyncratically with varying glyphs, but never became conventional in any language and are not present at all in modern Japanese.
The 50-sound table is often amended with an extra character, the nasal stop ン (n). This can appear in several positions, most often next to the N signs or, because it developed from one of many mu hentaigana, below the u column. It may also be appended to the vowel row or the a column. Here, it is shown in a table of its own.
The script includes two diacritic marks placed at the upper right of the base character that change the initial sound of a syllabogram. A double dot, called dakuten, indicates a primary alteration; most often it voices the consonant: k→g, s→z, t→d and h→b; for example, カ (ka) becomes ガ (ga). Secondary alteration, where possible, is shown by a circular handakuten: h→p; For example; ハ (ha) becomes パ (pa). Diacritics, though used for over a thousand years, only became mandatory in the Japanese writing system in the second half of the 20th century. Their application is strictly limited in proper writing systems, but may be more extensive in academic transcriptions.
Furthermore, some characters may have special semantics when used in smaller size after a normal one (see below), but this does not make the script truly bicameral.
The layout of the gojūon table promotes a systematic view of kana syllabograms as being always pronounced with the same single consonant followed by a vowel, but this is not exactly the case (and never has been). Existing schemes for the romanization of Japanese either are based on the systematic nature of the script, e.g. nihon-siki チ ti, or they apply some Western graphotactics, usually the English one, to the common Japanese pronunciation of the kana signs, e.g. Hepburn-shiki チ chi. Both approaches conceal the fact, though, that many consonant-based katakana signs, especially those canonically ending in u, can be used in coda position, too, where the vowel is unvoiced and therefore barely perceptible.
Syllabary and orthography
Katakana used in Japanese orthography
Of the 48 katakana syllabograms described above, only 46 are used in modern Japanese, and one of these is preserved for only a single use:
- wi and we are pronounced as vowels in modern Japanese and are therefore obsolete, being supplanted by i and e respectively.
- wo is now used only as a particle, and is normally pronounced the same as vowel オ o. As a particle, it is usually written in hiragana (を) and the katakana form, ヲ, is uncommon.
A small version of the katakana for ya, yu or yo (ャ, ュ or ョ respectively) may be added to katakana ending in i. This changes the i vowel sound to a glide (palatalization) to a, u or o, e.g. キャ (ki + ya) /kja/. Addition of the small y kana is called yōon.
Small versions of the five vowel kana are sometimes used to represent trailing off sounds (ハァ haa, ネェ nee), but in katakana they are more often used in yōon-like extended digraphs designed to represent phonemes not present in Japanese; examples include チェ (che) in チェンジ chenji ("change"), ファ (fa) in ファミリー famirī ("family") and ウィ (wi) and ディ (di) in ウィキペディア Wikipedia.
A character called a sokuon, which is visually identical to a small tsu ッ, indicates that the following consonant is geminated (doubled); this is represented in rōmaji by doubling the consonant that follows the sokuon. In Japanese this is an important distinction in pronunciation; for example, compare サカ saka "hill" with サッカ sakka "author". Geminated consonants are common in transliterations of foreign loanwords; for example English "bed" is represented as ベッド (beddo). The sokuon also sometimes appears at the end of utterances, where it denotes a glottal stop. However, it cannot be used to double the na, ni, nu, ne, no syllables' consonants; to double these, the singular n (ン) is added in front of the syllable. The sokuon may also be used to approximate a non-native sound: Bach is written バッハ (Bahha); Mach as マッハ (Mahha).
Both katakana and hiragana usually spell native long vowels with the addition of a second vowel kana. However, in foreign loanwords katakana instead uses a vowel extender mark, called a chōonpu ("long vowel mark"). This is a short line (ー) following the direction of the text, horizontal for yokogaki (horizontal text), and vertical for tategaki (vertical text). For example, メール mēru is the gairaigo for e-mail taken from the English word "mail"; the ー lengthens the e. There are some exceptions, such as ローソク (rōsoku (蝋燭, "candle")) or ケータイ(kētai (携帯, "mobile phone")), where Japanese words written in katakana use the elongation mark, too.
Standard and voiced iteration marks are written in katakana as ヽ and ヾ respectively.
An example of Japanese writing in 1940 using katakana exclusively. パアマネントハヤメマセウ
("Stop the permanent wave")
In modern Japanese, katakana is most often used for transcription of words from foreign languages (other than words historically imported from Chinese), called gairaigo. For example, "television" is written テレビ (terebi). Similarly, katakana is usually used for country names, foreign places, and foreign personal names. For example, the United States is usually referred to as アメリカ Amerika, rather than in its ateji kanji spelling of 亜米利加 Amerika.
Katakana are also used for onomatopoeia, words used to represent sounds – for example, ピンポン (pinpon), the "ding-dong" sound of a doorbell.
Technical and scientific terms, such as the names of animal and plant species and minerals, are also commonly written in katakana. Homo sapiens, as a species, is written ヒト (hito), rather than its kanji 人.
Katakana are also often (but not always) used for transcription of Japanese company names. For example, Suzuki is written スズキ, and Toyota is written トヨタ. As these are common family names, Suzuki being the second most common in Japan, it helps distinguish company names from surnames in writing. Katakana are commonly used on signs, advertisements, and hoardings (i.e., billboards), for example, ココ koko ("here"), ゴミ gomi ("trash"), or メガネ megane ("glasses"). Words the writer wishes to emphasize in a sentence are also sometimes written in katakana, mirroring the European usage of italics.
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.
Katakana were also used for telegrams in Japan before 1988, and for computer systems – before the introduction of multibyte characters – in the 1980s. Most computers in that era used katakana instead of kanji or hiragana for output.
Although words borrowed from ancient Chinese are usually written in kanji, loanwords from modern Chinese dialects which are borrowed directly use katakana instead.
Examples of modern Chinese loanwords in Japanese
The very common Chinese loanword rāmen, written in katakana as ラーメン, is rarely written with its kanji (拉麺).
There are rare instances where the opposite has occurred, with kanji forms created from words originally written in katakana. An example of this is コーヒー kōhī, ("coffee"), which can alternatively be written as 珈琲. This kanji usage is occasionally employed by coffee manufacturers or coffee shops for novelty.
Katakana are used to indicate the on'yomi (Chinese-derived readings) of a kanji in a kanji dictionary. For instance, the kanji 人 has a Japanese pronunciation, written in hiragana as ひと hito (person), as well as a Chinese derived pronunciation, written in katakana as ジン jin (used to denote groups of people). Katakana are sometimes used instead of hiragana as furigana to give the pronunciation of a word written in Roman characters, or for a foreign word, which is written as kanji for the meaning, but intended to be pronounced as the original.
In this travel warning, the kanji for "fog" (霧
) has been written in katakana (キリ
) to make it more immediately readable
Katakana are also sometimes used to indicate words being spoken in a foreign or otherwise unusual accent. For example, in a manga, the speech of a foreign character or a robot may be represented by コンニチワ konnichiwa ("hello") instead of the more typical hiragana こんにちは. Some Japanese personal names are written in katakana. This was more common in the past, hence elderly women often have katakana names. This was particularly common among women in the Meiji and Taishō periods, when many poor, illiterate parents were unwilling to pay a scholar to give their daughters names in kanji. Katakana is also used to denote the fact that a character is speaking a foreign language, and what is displayed in katakana is only the Japanese "translation" of his or her words.
Some frequently used words may also be written in katakana in dialogs to convey an informal, conversational tone. Some examples include マンガ ("manga"), アイツ aitsu ("that guy or girl; he/him; her"), バカ baka ("fool"), etc.
Words with difficult-to-read kanji are sometimes written in katakana (hiragana is also used for this purpose). This phenomenon is often seen with medical terminology. For example, in the word 皮膚科 hifuka ("dermatology"), the second kanji, 膚, is considered difficult to read, and thus the word hifuka is commonly written 皮フ科 or ヒフ科, mixing kanji and katakana. Similarly, difficult-to-read kanji such as 癌 gan ("cancer") are often written in katakana or hiragana.
Katakana is also used for traditional musical notations, as in the Tozan-ryū of shakuhachi, and in sankyoku ensembles with koto, shamisen and shakuhachi.
Some instructors teaching Japanese as a foreign language "introduce katakana after the students have learned to read and write sentences in hiragana without difficulty and know the rules." Most students who have learned hiragana "do not have great difficulty in memorizing" katakana as well. Other instructors introduce katakana first, because these are used with loanwords. This gives students a chance to practice reading and writing kana with meaningful words. This was the approach taken by the influential American linguistics scholar Eleanor Harz Jorden in Japanese: The Written Language (parallel to Japanese: The Spoken Language).
Katakana is commonly used to write the U+31F0–U+31FF) exists for Ainu language support. These characters are used for the Ainu language only.
Taiwanese kana (タイ ヲァヌ ギイ カア ビェン) is a katakana-based writing system once used to write Holo Taiwanese, when Taiwan was under Japanese control. It functioned as a phonetic guide for Chinese characters, much like furigana in Japanese or Zhùyīn fúhào in Chinese. There were similar systems for other languages in Taiwan as well, including Hakka and Formosan languages.
Unlike Japanese or Ainu, Taiwanese kana are used similarly to the Zhùyīn fúhào characters, with kana serving as initials, vowel medials and consonant finals, marked with tonal marks. A dot below the initial kana represents aspirated consonants, and チ, ツ, サ, セ, ソ, ウ and オ with a superpositional bar represent sounds found only in Taiwanese.
Katakana is used as a phonetic guide for the Okinawan language, unlike the various other systems to represent Okinawan, which use hiragana with extensions. The system was devised by the Okinawa Center of Language Study of the University of the Ryukyus. It uses many extensions and yōon to show the many non-Japanese sounds of Okinawan.