About roman town names
Imperial formulas build these results, from the mouthful "Diocletianopolis" to compact "Burdigala" and the military "Castra Vetera", which flags a legion camp by name alone. Polis endings mark Greek influenced foundations, while castra and vicus mark garrisons and villages. Spread the types along one mapped road and your province implies its own administrative history.
How roman town names are built
A single Markov chain grows each settlement name letter-by-letter from a 969-entry attested bank of Roman-era place names, output as one word. The chain reproduces Latin toponymic endings, so -on, -is, and -os tails recur. Bank outputs include Antinon, Armenion, and Argenturiolis.
Antinon
Markov-synthesised, with an -on tail
Armenion
Markov-synthesised, with an -ion tail
Argenturiolis
Markov-synthesised, with an -iolis tail
Naming notes
The phonetic tell is the classical place-name inflection the chain inherits, -on, -ion, -is, and the diminutive-flavoured -iolis (Argenturiolis) that mark Latinised toponyms. The 969-entry bank is large enough for strong variety, and outputs read as one fused place word rather than a phrase. Use these for invented Roman provincial towns and colonies; they are coinages tuned to the recorded sound, not attested settlements.
Questions about roman town names
- How are Roman town names generated?
- One Markov chain runs over a 969-entry bank of Roman-era place names, producing each next letter, so each output is one synthesised place word.
- Why do they end in -on, -is, or -iolis?
- Latinised toponyms carry those endings, and the chain learns them from the corpus, yielding forms like Antinon and Argenturiolis.
- What are these for?
- They suit invented Roman provincial towns, colonies, and forts on a historical or fantasy map rather than real attested cities.