| Ergative read as nominative | In the aorist the subject takes the ergative (-მა). Read as a plain subject, the sentence can reverse who acted on whom — critical in a statement or ruling. | The ergative is recognised and rendered by its function, so agency and object stay correct. |
| Ejective vs. aspirated flattened | Georgian distinguishes ejective from aspirated consonants; the difference is meaningful. Levelled, words and names shift. | Native phonology is applied, so the right consonant — and the right romanisation — is chosen. |
| Surnames -შვილი/-ძე transliterated inconsistently | These endings can be romanised several ways. Vary them across documents and the name no longer matches from one paper to the next. | One consistent transliteration is fixed and matched to the passport across all documents. |
| Mkhedruli “interpreted” with case | Mkhedruli has no upper/lower case; inventing capitals to “mark” names misreads the script. | We treat the script as unicameral and mark names by position and context, not by invented capitals. |
| National and ISO transliteration mixed | Georgia's national romanisation and scholarly ISO differ. Mixed, a single name appears in two spellings. | One scheme is chosen per document and matched to the ID, not blended. |
| Consonant clusters broken in layout | Long clusters wrapped carelessly produce broken glyphs or tofu boxes in the final PDF. | Typesetting keeps clusters intact, so the delivered document renders cleanly. |
| Archaic letters in old records skipped | Older or church documents contain historical letters no longer in use. Skipped, names and dates go missing. | Historical letters are read and transcribed, so old certificates are rendered in full. |