Japanese/Transcribing English to Japanese

The transcription of English to Japanese has been done since the earliest cultural contacts between English speakers and Japanese. During the Edo period, kanji were used phonetically to write English and other foreign words, but in the modern period katakana have become the principal target script. Unlike the systems for romaji, there is no standard for transcribing into katakana, and methods vary. However, generally all methods attempt to preserve the pronunciation of English, not the spelling. That is, transcription not transliteration is done.

This article deals with transcription of English words into the nearest phonetic equivalent in Japanese.

Reasons for transcribing
The purpose of the transcription partly determines how it is done. There are reasons why one would want to transcribe an English word to Japanese: Many legal documents, such as company registrations, require that only Japanese script is used. A computer database may need entry in Japanese script for the purpose of sorting and collation. Educators want to explain the pronunciation of English words by transcribing. Loan words from English are usually written in a transcribed form. Or one may simply be interested in how one's name looks in Japanese.

Accordingly, there are different priorities for the transcriber. The educator might want to indicate many of the subtleties of English pronunciation whereas a person naming a new product might be more concerned with the ease of pronunciation for native speakers of Japanese.

Difficulties
Japanese distinguishes fewer sounds than English. For example, Japanese does not distinguish the vowel sound of "run" and "ran", or the consonant sound of "row" and "low". Moreover the rules by which sounds can be combined in Japanese are generally more restrictive than the English rules. As a result, the pronunciation of the transcribed word can differ quite considerably from the original word in English.

If writing for a Japanese audience, it is worth checking whether there is already an accepted transcription into Japanese, and whether the meaning of the new word has changed in Japanese. The word mishin illustrates both pitfalls: not only is this an unexpected rendering of the English word "machine", but the Japanese word's meaning is limited to sewing machines. It is also worth noting that some terms which may at first glance appear to be mangled English loanwords are, in fact, loans from other languages: koppu (drinking glass) is not a version of the English "cup" but a loan of Dutch kop, and tabako is from Portuguese tabaco, not from "tobacco".

Procedure for transcription
Most Japanese people do not use a systematic procedure for transcription; instead they transcribe according to their perception of the English pronunciation, albeit significantly influenced by the spelling. However, the process can be represented formally as a set of transformations, which is presented in one possible order below. Proficient Japanese speakers internalize the transformations and perform them all simultaneously when inserting English words into written or spoken Japanese.

Step 1: Transcribe the English phonetically
The first step is to start with a phonetic representation of the English word, as distinct from the spelling. The phonetic transcription should reflect the careful pronunciation of the word. Spelling can often mislead as to what the pronunciation is. If there is any doubt, a dictionary will provide an accurate indication of what the sounds are. The letter x typically corresponds to two sounds (ks) and the digraphs sh, ch, and th each correspond to a single sound. The English sounds in the examples below are in the International Phonetic Alphabet. (See International Phonetic Alphabet for English and IPA chart for English for explanation of these symbols used for transcribing English.)

Step 2: Transform the vowels from English to Japanese
Japanese has a different and smaller sound set than English, so many sounds have to be changed to equivalent or similar sounds in Japanese. The Romanization system used here is a variation of the Hepburn system, where long vowels are represented by doubled letters (ii, ee, aa, oo, uu) and the moraic nasal is represented with capital N.

Vowels need to be changed to correspond to use the five Japanese vowels. Typically, the vowels used in a British Received Pronunciation are used as the base English vowels for transcription, using the following system, where doubled vowels mean long (2-mora) vowels:

In rhotic dialects of English, like American English, the letter r sounds at the end of syllables, but for the purpose of transcription into Japanese, this sound transcribes into a vowel sequence ending in a, except for the sequence, which corresponds to Received Pronunciation , and is transcribed as oo. That is, car becomes kaa not karu, and pork becomes pooku not poruku.

Step 3: Transform consonants
Some consonants require changing during transcription into Japanese. This process has three substeps:

Transform non-Japanese sounds to closest Japanese equivalents
First, English has a few consonant sounds that Japanese lacks or only contains in certain contexts, so they must be transcribed into other sounds that Japanese has.

There are other English consonants that Japanese lacks, such as, the closest equivalent being. And though both languages contain, in Japanese it assimilates to before. However these differences in pronunciation are small enough that they need not be considered different sounds for the purpose of transcription.

* It is possible to notate in Japanese kana, and it is done in the Japanese spellings of "Vietnam" (ヴェトナム vetonamu) and "Vicks" (ヴィックス vikkusu), but the sound does not exist in native Japanese phonology and is usually changed to "b" when transcribing English words.

Palatalize coronal obstruents
Next, Japanese requires coronal obstruents "s", "z", "t", "d" to be palatalized when they occur before the vowel i, so if these consonants occur before "i", either they change to their palatalized form or the vowel "i" changes to "e": In recent loanwords, "ti" and "di" are often preserved. In kana, this sound is represented by a full-sized "te" or "de" and a small-sized "i": ティ (ti), ディ (di).
 * "si" changes to "shi" (Remember "si" might come originally from "thi", as in thick).
 * "ti" changes to "chi" or "te"
 * "di" changes to "ji" or "de"
 * "zi" changes to "ji"

Double voiceless obstruents after short vowels
In Japanese, the voiceless obstruents "p", "t", "k", "s", "ch", and "sh" have geminate (doubled) forms, written using a sokuon (small tsu) character, and in English transcription these geminates are used after short vowels. Short vowels are vowels which are transcribed using the vowel table above using a single vowel ("a", "e", "i", "o", or "u"). This transformation is usually but not always applied in the middle of a word. Also, sometimes syllable-final "t" is transformed to "ts" instead of "tt".

Step 4: Add epenthetic vowels
Japanese has strict constraints on the structure of syllables, and any syllables that violate these constraints have vowels inserted until the constraints are met. These are called epenthetic vowels.
 * The only consonant clusters (sequences of consonants with no intervening vowels) allowed in Japanese are the geminate (doubled) consonants cch, mm, nn, ss, ssh, tch and tt. However, the sounds represented by the English digraphs ch, sh, and ts are considered single sounds for the purpose of transcribing into Japanese.
 * Japanese syllables can only end in vowels and N.

Any sequence of sounds that does not obey these rules must have epenthetic vowels inserted. The epenthetic vowel is usually "u", but there are a few exceptions:


 * "m" does not take an epenthetic vowel when followed by "b" or "p", but is instead replaced by "N" (which is pronounced the same as "m" in those contexts). For example, computer becomes "コンピュータ" (koNpyuuta). "N" is also pronounced as before "k" or "g", requiring no epenthesis after the "n" in words such as "ink", which becomes "インク" (iNku).
 * "t" and "d" take "o" as an epenthetic vowel. "t" can also take "u" in which case the "t" is affricated to "ts" (i.e. "tsu").
 * The rule for "tsu" described above is sometimes replaced by one where "tu", written in kana with a full-sized "to" followed by a small "u": トゥ, is used. This is more common in more recent innovative Japanese dialects. For example, the name of the film "The Truman Show" in Japanese is "トゥルーマン・ショー" (turuumaN shoo)
 * "ch" and "j" take "i" as an epenthetic vowel.
 * "k" and "sh" usually takes "u" as an epenthetic vowel, but sometimes it takes "i" or may vary between "u" and "i". "i" is the more conversative pattern.

Step 5: Break into morae
Japanese is divided into morae, with each mora containing one of the following:
 * A consonant and a single vowel (CV)
 * A consonant, "j" and a single vowel (CjV)
 * A single vowel (V)
 * Moraic (final) "n" (N) ン
 * Doubled (geminate) consonant ッ

Step 6: Transcribe rōmaji into katakana
Each mora corresponds to one or sometimes two katakana characters. The second mora of a long vowel is uniformly transcribed as ー in katakana. Moraic "n" (transcribed here as "N") is ン in katakana.

Inconsistencies
Though commonly used katakana spellings tend to be consistent with the above system of transcription, there are also many exceptions. Some transcriptions are apparently based on misinterpretations of the word's pronunciation based on its spelling. For example, the "u" in "studio" seems to have been interpreted as if it were (as in the word "study"), not, resulting in the transcription "スタジオ" (sutajio).

Though the basis for English to Japanese transcription is usually British Received Pronunciation, with its different short "o" sound and unpronounced rhotic "r"s, there are also exceptions. The words "cocktail" and "soccer" are transcribed as "カクテル" (kakuteru) and "サッカー" (sakkaa), and the Japanese name of the English letter "r" is "アール" (aaru), which corresponds more closely to a rhotic accent.

The final sound in English words is usually transcribed as "ト" (to), but it in some words such as "fruit" and "suit" it is transcribed as "ツ" (tsu), making the pronunciation of some singular nouns sound more like their plural forms, even though plural "s"s tend to be ignored when transcribing English nouns into Japanese.

Transcribing using the steps outlined above results in the English short "i" sound becoming the Japanese i sound, but there are also cases in which it becomes the Japanese e sound. Examples include "digital" and "sticker" becoming "デジタル" (dejitaru) and "ステッカー" (sutekkaa). Also, the "re" in the English words "report" and "reporter", which is pronounced as in Received Pronunciation, becomes re in "レポート" (repooto) and "レポーター" (repootaa), though they are sometimes alternately transcribed as "リポート" (ripooto) and "リポーター" (ripootaa)

Even within the common system of transcriptions, there are multiple possible ways in which a certain sound can be transcribed. This can result in multiple transcriptions of a single word, such as the name "David", which is written a number of ways in Japanese. Different pronunciations of the same word are sometimes used to show what meaning of the word is being used. For example, "ストライク" (sutoraiku) refers to a strike in baseball or bowling, while "ストライキ" (sutoraiki) refers to a workers' strike. Also, "ポンチ" (poNchi) refers to fruit punch while "パンチ" (paNchi) is used for other meanings of the word.

As mentioned above, many transcriptions (particularly those involving a non-final schwa) are non-fixed and are often based more on spelling than actual pronunciation. This often leads to words which sound similar to each other in English sounding radically different from each other in their Japanese pronunciations. While the pronunciations of the English words "pirate" and "pilot" differ only in the "l" and "r", the two words are transcribed respectively into Japanese as "パイレーツ" (paireetsu) and "パイロット" (pairotto), with the only difference between the original pronunciations disappearing and some new differences appearing in other places.

There are also some inconsistencies in Japanese between the way English words are transcribed, and the way words from some other languages containing the same sounds are transcribed. A final velar nasal consonant in an English word (spelled "ng") is usually transcribed as "ング" (Ngu), but the same sound in Korean and Chinese words is transcribed as "ン" (N). For example "Hong Kong" and "Kung-Fu" become "ホンコン" (hoNkoN) and "カンフー" (kaNfuu) respectively, and the "Yong" in Korean actor Bae Yong Joon's name becomes "ヨン" (yoN).

Exceptional transcriptions
The following are commonly used transcriptions which do not conform to the common system of transcription. This does not include Japanese abbreviations of English words or words which resemble English, but came into Japanese directly from other languages.

* in the case of James Curtis Hepburn, but not Katharine Hepburn or Audrey Hepburn, whose last name is transcribed as "ヘップバーン" (hepubaaN).

** "Holly" on its own is transcribed as "ホリー" (horii).

Transcribing multiple words
In some instances, such as language textbooks or song lyrics, phrases or entire sentences may be transcribed into Japanese.

Multiple word transcription is typically done on a word-by-word basis, with no account being taken of word linking. For example, "an engineer" would most commonly be transcribed into Japanese as "a.N.e.N.ji.ni.a" rather than "a.ne.N.ji.ni.a", with the linking between the "n" and "e" represented by the Japanese mora "ne". In some set phrases, such as "kaman" for "come on", this general trend is broken.

Example of transcribing a whole sentence
English: "My hovercraft is full of eels."

Step 1:

Step 2-3: "mai hobaakraft iz ful ob iirz"

Step 4: "mai hobaakurafuto izu furu obu iiruzu"

Step 5  "ma.i ho.ba.a.ku.ra.fu.to i.zu fu.ru o.bu i.i.ru.zu"

Step 6  「マイ　ホバークラフト　イズ　フル　オブ　イールズ」