| Wrong politeness level (keigo) | Keigo grades verbs and vocabulary by register. The wrong level makes a contract sound casual or a letter presumptuous — a substantive error in business and law. | The register is chosen deliberately for the document type and relationship, so tone and standing are correct. |
| Three scripts mixed wrongly | Kanji, hiragana and katakana each have a job; misassigned, foreign names and grammar read wrong or ambiguous. | Each script is used in its correct role, so the text reads as a native document, not a transliteration. |
| Era-date miscalculated | 令和/平成 years must be converted to Gregorian. A miscalculated year puts a birth or a deadline in the wrong place. | Era-years are converted precisely and cross-checked, so dates on certificates are correct. |
| Name order & reading (furigana) | One kanji name can be read several ways, and Japanese order is surname-first. Guessed, the romanised name diverges from the passport. | Readings are matched to the passport and the name order handled correctly, so the ID and the translation agree. |
| Loanwords in katakana inconsistent | Foreign names and terms in katakana can be transcribed several ways. Varied across a file, one entity appears under two names. | Katakana transcriptions are fixed once and kept consistent throughout the document. |
| Counters (助数詞) missing or wrong | Japanese counts objects with classifier words. The wrong counter is ungrammatical and can blur quantities in a spec or contract. | The correct counters are used, so quantities read naturally and unambiguously. |
| Full-width vs. half-width characters | Mixing 全角/半角 forms and ignoring line-break rules (kinsoku) produces tofu boxes and broken layout in the PDF. | Character widths and line-break rules are handled in typesetting, so the delivered document renders cleanly. |