About

The Premodern Concordance is a prototype for a larger project that will use computational methods to link named entities — people, plants, substances, places, diseases, and concepts — across multilingual early modern texts relating to natural and scientific knowledge. The goal is to make the unstable, polyglot terminology of premodern natural knowledge searchable and comparable across languages, centuries, and traditions, and then to explore new research possibilities and questions unlocked by this.

Our methodology is described here, and a full data set is available here (in progress).

Personnel

Benjamin Breen

Department of History, UC Santa Cruz

Mackenzie Cooley

Department of History, Hamilton College

Colophon

Typography

  • System UI — body text
  • UnifrakturMaguntia — logotype, blackletter display
  • EB Garamond — early modern serif
  • Space Grotesk — geometric display

Framework

  • Next.js 16 — React framework
  • Tailwind CSS 4 — styling
  • Vercel — hosting

Research models

  • BGE-M3 — fine-tuned cross-lingual embeddings for entity matching (current model was fine-tuned on 500 matched multi-lingual pairs of early modern natural concepts/terms)
  • Google Gemini 2.5 Flash Lite — entity extraction, enrichment, and verification
  • OpenAI text-embedding-3-small — semantic search index

Site code

Built with assistance from Claude Code (Claude Opus 4.5).

mercury · mercurio · azogue · vif-argent · hydrargyrum