I have done my best to refer to placenames in their modern, not ancient or medieval, spellings, so Mérida not Emerita, Bologna not Bononia, only using English spellings (generally in fact borrowed from French) for places like Milan, Athens, or Cologne, where Milano, Athēnai, Köln would seem precious. In the West this creates few problems, for most medieval historians do the same (ancient historians often use classical forms, however). For regions of the Roman empire now in Arab-speaking countries, and sometimes in Turkey and the Balkans, historians frequently use ancient or medieval names to the exclusion of modern ones, particularly when they are very different. Here I have used both, putting the modern one in brackets without an initial article, so Arsinoë (Madīnat al-Fayyūm); only where the ancient name (or an Anglicization of it) is so well known that to use the modern one genuinely contributes nothing, like Constantinople or Antioch, have I left it. So also I have left well-known ancient names where there is no modern settlement, as with Caesarea in modern Israel. Caesarea is also one of a handful of ancient Greek names that have been left in Latinized versions (others are Nicaea and Phocaea), on the grounds that they are so well known in this form that consistency would confuse. In Egypt, where Coptic texts often provide a third name, I have put in all three when using Coptic sources, so Sioout (Greek Lykopolis, modern AsyūṠ). Arabic transliterations are also very variable; when I have been able to pin down a classical Arabic form I have normally used it, except in the Maghreb, where I have used the Francicized transliterations current in the countries themselves, so Kairouan not Qayrawān.
These create other difficulties. In general I have Anglicized the names of rulers, and names from the standard lexicon of saints: so Justinian not Justinianus/Ioustinianos, Clovis not Chlodovech(-us), John and George not Johannes, Geōrgios, etc. I have left nearly everybody else in the language of the texts, except that I have cut-us from Germanic names, so Gundulf not Gundulfus, and have sought to use modern standard transliterations of classical Arabic names, not the specific spelling forms of texts. In Greek, I have transliterated kappa as k, upsilon as y (except in diphthongs) and chi as ch. In Old Norse and Anglo-Saxon, I have transliterated thorn as th, and kept ð. I have used modern forms for well-known literary figures, such as Augustine or Bede. I have resisted the tendency of some Byzantinists to Latinize Greek names, which does violence to intellectuals as Hellenic as Prokopios, and I have also avoided their (even stranger) habit of Latinizing the titles of texts, though here I have generally