Why is ち sometimes romanised as ‘ti’ and じ as ‘zi’?

11 comments

1. It’s because ち is part of the t column (た、ち、つ、て、と）and し/じ are part of the s/z column

2. Is it in any specific context or just all throughout? Because if it’s all throughout that’s just wrong-

3. Different romanisation systems of the Japanese language. I had this exact same question the other day.

   https://en.m.wikipedia.org/wiki/Romanization_of_Japanese

   Currently, romanisation has not been standardised universally in Japanese, with the Kunrei-shiki (standardised version taught in schools) and Nihon-shiki being common in Japan.

   Hepburn is what we alphabet users would typically romanise them as.

4. It’s not related to the IPA, these kinds of transcriptions are specifically used so that wr don’t have to get into the details of pronunciation (palatalization of /t (d) s z h/ before /i j/, affrication of /t d/ before /u/, voiceless vowels, /u/ actually not being rounded etc.). This can often be better at showing the actual structure of the language, e.g. writing 持つ and 持ちます as *motu* and *motimasu* better shows what are the inflectional morphemes and what is the root, without requiring the reader (most likely not wanting to use the book to learn Japanese) to keep track of all the phonetic stuff.

5. I don’t know why but here we romanize it as “chi” and “ji”, respectively. (both pronounced more like tchi and dji but you get the idea)

6. ち is in the t area of characters. ta, te, ti, to, and well ti or tsu.

   I didn’t know zi was じ

7. Originally [tɕ] was just an allophone of /t/, occurring before /i/ and /j/. Same goes for [ts] before /u/. That’s the reason why the kana table goes たちつてと. Phonemically it was just /ta ti tu te to/, but phonetically it was [ta tɕi tsu te to] (I’m not being precise with the transcription of /u/ here, but that’s beside the point). Nowadays there is [ti] and [tu] (spelt ティ and トゥ respectively) in loanwords such as パーティ [paꜜːtiː], タトゥー [tatuꜜː], so it’s not as easy to analyze [tɕ] and [ts] as just allophones of /t/ anymore.

   On the other hand, [ɕ] is pretty much still an allophone of /s/: [si] isn’t really used, not even in loanwords, and Japanese people have trouble distinguishing ‘sit’ from ‘shit’. Likewise [ʑ] is an allophone of /z/.

8. The reason for 日本式 and 訓令式 romanization is that it’s more intuitive to Japanese speakers and more consistent. The idea to romanize “じ” as <zi> really makes more sense and choosing <ji> is fairly arbitrary and note even necessarily what it sounds like.

   https://forvo.com/word/%E3%81%98/#ja

   If you listen to all the examples here it’s clear that <ji> doesn’t even make that much sense and many also say <dzi> would perhaps be better, which does occur in “Godzilla”, which probably just derived from an English person writing down what that person heard when a Japanese person said “ゴジラ”.

   Hepburn romanization is often said to about what “English speakers hear” opposed to what Japanese speakers hear when they hear Japanese but I’m not convinced that has any merit, when English speakers write down what they actually hear they come to very different conclusions from Hepburn in words such as “rickshaw” or “skosh” which were actually loaned the “natural way” by aëreal contact.

   But in the end, native speakers of a language “hear” different things from foreign speakers. This is also why “girl” became “ガール” in Japanese but “gal” became “ギャル”. From the perspective of English speakers these two words surely start with the same sound, but from the perspective of Japanes speakers one starts with “g” and the other with “gy”. Just as from the perspective of Japanese speakers “ざ“ and “じ” start with the same sound, but from the perspective of English speakers there may be a difference.

   Hepburn is supposedly designed to capture what that difference is to English speakers and Anglifies the romanization such that it writes down Japanese in a way that is intuitive to them but if you ask me it fails horribly. English speakers don’t really hear “dōshite” or anything like that when they hear “どうして” and mostly just hear “dōste”.

9. Because on a computer you type ti or zi to input ち OR じ with the least amount of keypresses.

10. Interesting how you did not note all the long a the same way. I especially do not like the use of h to make vowels longer.

11. I’m curious about which text you are using? The first textbook I had was Japanese the Spoken Language, by Harz-Jorden, and it uses a slightly modified訓令式 (kunrei-shiki).

    It lists four romanization systems, with examples of how they spell “romaji”:

    “`
    JSL (the textbook): roomazi
    Shin-kunrei-shiki: rômazi
    Hepburn: rōmaji
    Nippon-shiki: rōmadi
    “`

    It follows up on reasons for why it’s chosen a non-Hepburn system. The first is the most obviously compelling reason: the listed non-Hepburn systems bear a more direct relationship to Japanese structure. She gives an example, where she explains:

    how “To form the stem of consonant verbals”

    “Using JSL, Shin-kunrei-shiki, or Nippon-shiki: change the final -u to -i”

    “Using Hepburn romanization: change the final -u to -i, but if -u is preceded by ts, change the ts to ch, and if -u is preceded by s, add h after the s.”

    The other change for Harz-Jorden’s romanization versus the very similar kunrei-shiki, which I appreciated, is that it expands the ô to oo, which is nice in how it “holds the space” of a second kana, like 多く being written ooku rather than ôku, which more accurately reflects the two kana you’d use: おおく.

    The main point of these systems is to more closely match how hiragana/katakana work, and how they transform through various grammar/spelling changes.

    Hepburn focuses just on one-off accuracy of “how to best approximate this word if I were to show the spelling to someone from Ohio?”