Attribution & Licensing
This project creates machine-generated personal-name attestation indexes from
Latin inscription corpora. It is a derived, computational resource by Michael A Bosse —
not an official release of EDCS, LIRE, EDH, Trismegistos, CIL ACE, Ubi Erat Lupa,
DARE, or any other source database.
Recommended citation
When using the derived data, cite the project and the source datasets:
Michael A Bosse, Roman Names: machine-generated personal-name attestations
from Latin inscriptions. Derived from EDCS 2022 and LIRE v3.0.
Generated with the metiscus/roman-names pipeline.
When a Zenodo DOI is minted for this project, cite that DOI in place of the
repository-only reference.
Suggested acknowledgement
This work uses machine-generated data by Michael A Bosse from
metiscus/roman-names, derived from EDCS 2022
(
10.5281/zenodo.7072337)
and LIRE v3.0
(
10.5281/zenodo.8431452).
Interpretations and errors are the responsibility of the user, not the source
database maintainers.
Data license
Derived data artifacts produced by this project — including generated
name-attestation datasets, NER outputs, clustering outputs, webapp data,
generated summaries, and generated English translations — are released under the
Creative Commons Attribution 4.0 International License (CC BY 4.0).
Attribution is required. Reusers must also preserve attribution to the
underlying source datasets listed below.
Exception: Name lookup tables derived from the Digital
Prosopography of the Roman Republic (DPRR) are released under
CC BY-NC 4.0
and may not be used for commercial purposes.
Software source code in this repository is licensed separately under the MIT License.
Source datasets
The derived data depends on the following third-party source datasets.
Reusers must preserve attribution to these sources.
Full name
Epigraphik-Datenbank Clauss-Slaby (EDCS)
Version used
2.0, aggregated 2022-09-12
Creator
Petra Heřmánková, Aarhus University
Fields used: EDCS identifiers, raw inscription text, province,
findspot/place metadata, date fields, and inscription type.
Full name
Latin Inscriptions of the Roman Empire (LIRE)
Version used
v3.0, published 2023-10-11
Creators
Vojtěch Kaše, Petra Heřmánková, Adéla Sobotková — Aarhus University
Fields used: coordinates, dates, interpretive text, publication references,
external identifiers, and structured person data used for validation.
LIRE itself aggregates material from EDCS and the Epigraphic Database
Heidelberg (EDH); its disclaimer notes that original data is provided by
the third-party source indicated in its metadata.
Project lead
King's College London
Used for canonical name whitelists (Nomina and Cognomina) and tribal
variant detection. This project's use of DPRR data is for academic
research and follows the NonCommercial restriction.
Machine-generated annotations
The following are generated by this project's NLP pipeline and should be
treated as scholarly aids requiring verification against the inscription
text and relevant publications:
- Name extraction and component classification (praenomen, nomen, cognomen)
- Social-status labels, fragmentary flags, non-person flags
- Heuristic name clusters (possible same-person groups)
- English translations and summaries
Clusters are heuristic, not confirmed identities.
A cluster groups inscriptions that may refer to the same individual based on
name similarity and geographic proximity — not prosopographical verification.
Linked third-party resources
The webapp links to external databases where identifiers are available.
It does not claim ownership of those records, pages, images, or map tiles.