Import a Japanese paper into Mendeley or Zotero. The author "Tanaka Hiroshi" — where Tanaka is the family name and Hiroshi is the given name — gets re-rendered as "Hiroshi Tanaka" or, worse, "Hiroshi Tanaka, T." in the output. That's wrong, in any of the major citation styles, in Japanese.
Family-first is the rule, not the exception
Family-first naming is standard across the Sinosphere: Chinese, Japanese, Korean, Vietnamese. It's also used in Hungarian and a few other languages.
In English-language papers by Japanese authors, the convention has historically been to flip to given-first ("Hiroshi Tanaka") to match Western reader expectations. In 2020, Japan's government recommended that Japanese authors keep family-first when publishing in English. Many journals are following.
The result: a citation tool that always flips CJK names is now actively wrong for newer Japanese papers, and was always wrong for non-English contexts.
The comma vs space convention
In APA, an author is "Smith, J." — surname comma initials. In MLA, "Smith, John". In IEEE, "J. Smith". The convention assumes the surname is the FIRST element (after the comma).
For "Tanaka, Hiroshi", APA reads this correctly: surname Tanaka. But the BibTeX file probably has `author = {Hiroshi Tanaka}` — given-first, no comma. The importer guesses. It usually guesses wrong.
BibTeX has explicit syntax (`{Tanaka, Hiroshi}` with the comma) but most authors don't use it. RIS has separate AU fields for given and family but most exports cram the full name into one.
When in doubt, look at the author's preferred order in their own publications. If they consistently sign as "Tanaka Hiroshi" or "Tanaka, Hiroshi", that's family-first. Tools should honor it.
Six output styles, six different format rules
APA 7, MLA 9, Chicago, IEEE, Vancouver, JST (Japanese Science and Technology Information Aggregator) — each has its own rules for how to render an author. APA uses initials only after the surname. MLA uses full first name on the first author. JST keeps full Japanese names in original kanji where possible.
A converter that handles one input format and one output format is one-tenth of the actual problem. The full problem is N inputs × M outputs × correct CJK handling × correct edge cases (institutional authors, et al. rules, ampersand vs and, etc.).
The kanji vs romaji double-name problem
A Japanese paper might list an author both in kanji (田中博) and in romaji (Tanaka Hiroshi). Some citation styles want romaji, some want kanji, some want both. Western tools usually drop the kanji entirely.
Our citation converter parses BibTeX, RIS, and plain text input, detects family-first names (CJK + Hungarian heuristics), and outputs APA 7, MLA 9, Chicago, IEEE, Vancouver, or JST format. Per-citation parsed view shows what fields it found and surfaces warnings when something looks ambiguous.