I looked through the ALA-LC romanization of Japanese (from the list of ALA-LC Romanization Tables), and there are some kana-to-romanization mappings which I am not sure what to make of.

(To prevent confusion: While this ALA-LC romanization (dated "June 2022") is largely (or perhaps even fully) consistent with Hepburn-as-presented-in the respective article on Wikipedia (though ALA-LC doesn't cover all from Wikipedia's "Extended katakana" table, eg "ツュ tsyu" is missing), I would at this point recommend against referring to it as a form of Hepburn, as (1) Hepburn has multiple dialects even in the present day and (2) the ALA-LC document itself has wording ("⁵ Always n (Do not change to m as in Hepburn)", p33) which suggests that it differentiates itself from Hepburn-as-defined-in the now-withdrawn ANSI Z39.11-1972 (cited on p35 of the ALA-LC document). I wanted to mention this, just in case someone reads these words

The version of [Hepburn] published in the third (1954) and later editions of Kenkyusha's New Japanese-English Dictionary are often considered authoritative; it was adopted in 1989 by the Library of Congress as one of its ALA-LC romanizations,[fn] and is the most common variant of Hepburn romanization used today.

in the Wikipedia article.)

Now to my actual quibble with the ALA-LC document: It defines (p33-34) the following kana-to-romanization mappings for kana which I think are neither common nor mentioned in most other sources:

きぃ/キィ kī
にぃ/ニィ nī
ひぃ/ヒィ hī
みぃ/ミィ mī
りぃ/リィ rī
ぎぃ/ギィ gī
びぃ/ビィ bī
ぴぃ/ピィ pī
すぇ/スェ swe

What are these ALA-LC mappings used for? Are they for the transcription of specific languages into Japanese? (Or perhaps for dialects of Japanese?) What would be some examples of words in which these extended kana would be the preferred spelling in some context (such as a linguistic discussion)?

Or was the inclusion of these kana a mistake? Here is a case for why the ALA-LC document was edited sloppily (which is not to say that the family of Hepburn romanizations has ever been clearly defined):

  • Unfortunately, the ALA-LC document gives no examples with any of these kana (though it's possible that it contains example sentences with these moras represented using kanji).
  • Confusingly (and this is definitely an error), it lists スェ as a sequence "not found in this table" with the constructed romanization sue, just below the table listing スェ as swe (p34).
  • I thought /ii/ coming from hiragana is generally supposed to be transcribed as ii, not ī. But the ALA-LC tables list hiragana along with katakana for all these cases.

2 Answers 2


Pronunciation-wise, the i + additional ィ/ぃ series would have no difference from lengthened ī in Standard Japanese. They are often used to add milder or cuter feeling, or indicate some sounds which are felt "not fully double-long", but mostly for mere stylistic effect with little phonological difference (similar to やぁ, へぇ, etc.). That said, such spellings are apparently preferred in some context to write English //-i// ending, as in: パリィ (parry), ファジィ論理 (fuzzy logic), etc. I speculate that the reason why the Library of Congress defines those mappings is because they appear in its book records.

スェ is once popular spelling to transcribe [[swe]]-like syllable, which you can find mainly as part of スェーデン in older books. Its decline is presumably correlated with the fact スィ is getting into regular phonological inventory of most modern speakers as an independent syllable [[si]], not [[sɯi]] or [[swi]].

  • It seems worth noting that English speakers typically pronounce the "Swe" of "Sweden" with an ィ vowel and not a ェ vowel (closest match, anyway). Commented Apr 28 at 19:42
  • @KarlKnechtel I doubt if it's actually a mixed up form of German Schweden and English Sweden. Commented Apr 28 at 19:58
  • 1
    Should I interpret /si/ to mean that it's realized as [ɕi] or [si]? Commented Apr 28 at 20:09
  • 1
    @JoshuaGrosso In this case, I meant the one pronounced [si] explicitly apart from [ɕi]. Commented Apr 29 at 5:19

キィ, ニィ and so on are alternative (and probably nonstandard) spellings for キー, ニー and so on, respectively. They are typically used to transliterate non-Japanese names/words. They may be more common in fiction than in real names, and that's probably why not all romanization tables cover them. For example, while マリー and マリィ are both common ways of transliterating Mary, I usually expect マリー in history textbooks. There is no difference in pronunciation, so they are transliterated in the same way as リー, etc.

スェ is very rare. It might have been used to represent the "swe" sound as in Sweden or sweater, but this combination of katakana is almost nonexistent today. I don't even know how to pronounce it correctly. People may pronounce it like スウェ (in two morae), or セ (in one mora), or somewhere between them.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .