📄 Personal Names in the Aramaic Inscriptions of Hatra

Summary

This text is not a technical paper. It is Brandon Simonson's review of Enrico Marcato's book on the personal names recorded in Hatra's Aramaic inscriptions (Edizioni Ca' Foscari, 2018, open access). Hatra was a trading city between the Parthian and Roman empires, and its name stock blends several cultures. The reference work it had relied on, Abbadi's 1983 name book, had become outdated: the number of inscriptions has almost doubled since then, and knowledge of Aramaic has improved. Marcato's book has three parts:

• a catalogue of every personal name, with transliteration, candidate meanings, variant readings, etymology, cognate parallels and attestations;
• a linguistic analysis covering origin, spelling, phonology, morphology, syntax, semantic classes and divine-name elements;
• a concordance mapping its entries to Abbadi's.

The book's main finding is that Arab groups strongly shaped Hatrene culture. Simonson praises the book as an accessible model for documenting names in a multicultural community. He also proposes that some divine elements usually tied to the sun god Šamaš could instead point to the moon god Sîn.

Key points

• The catalogue covers 376 personal names. By origin:
  • Aramaic: 167 (44.4%)
  • Arabian: 105 (27.9%)
  • Iranian: 29 (7.7%)
  • Greek: 8 (2.1%)
  • Akkadian-related: 7 (1.9%)
  • unclassified: 60 (16%)
• Compared with Abbadi, the concordance shows 136 newly added names, 11 readings made more cautious and 39 made more certain.
• The semantic classification follows Martin Noth's 1928 scheme rather than Rainer Albertz's later revision. It is deliberately kept coarse because many names remain ambiguous.
• Names built on Hadad are filed under three separate deity forms (Adda, Addu, Hadad), which keeps the different cultural sources visible.
• Simonson suggests re-reading several names (e.g., "servant of", "gave", "healed" types) as referring to Sîn rather than Šamaš.

Interesting ideas

• I1. Publishing a concordance against the previous edition shows which items are new, which became less certain and which became more certain, so readers can audit every change. — For Mova (learn): Each treebank or silver-data release could ship a diff that tags sentences as added, confidence-lowered or confidence-raised, making regressions traceable.
• I2. Spelling variants of one deity are kept as distinct headings instead of being merged, because each variant carries its own origin. — For Mova (world): Entity nodes could keep surface variants as separate linked forms with provenance, rather than collapsing them early into one canonical name.
• I3. About a sixth of the names go into an explicit "unclassified" bucket, and the taxonomy is kept shallow where interpretations are uncertain, rather than forcing labels. — For Mova (frames): Role assignment could allow an honest "unresolved" label instead of forcing a VerbNet role when the evidence is weak, which keeps explanations truthful.
• I4. Each catalogue entry keeps several possible meanings and alternative readings side by side, each backed by parallels from related languages. — For Mova (tokens): Lexicon entries for proper names could store multiple analyses with supporting evidence, so later stages can pick one from context instead of inheriting a premature single choice.