Mapping the Translation Landscape: A Research Diary

How do you count something that's never been counted? Today we built a comprehensive database of Latin-to-English translations—and discovered how much we don't know.

The Problem

We wanted to answer a simple question: How much Latin has been translated into English?

It turns out nobody knows. There's no master list of Latin translations. Publishers don't track them. Libraries catalog by title, not by source language. Academics work in silos. The result: a vast body of translation work exists, but it's scattered across dozens of series, hundreds of publishers, and centuries of scholarship.

Phase 1: The Major Series

We started with the obvious: major translation series from academic publishers.

Finding: roughly 1,400-1,500 volumes across major commercial series. But volume counts are misleading—one Loeb volume might contain a single short work or a multi-book epic.

Phase 2: Open Access Resources

The real surprise: how much is freely available online.

Dana Sutton's Philological Museum was a revelation: a single scholar has been quietly translating Neo-Latin texts for decades, with an analytic bibliography of nearly 80,000 items.

Phase 3: The Citation Problem

Early in our research, we ran into a problem: most statistics about Latin literature were estimates without sources. “500,000 Latin works” appeared everywhere, but where did it come from?

We rebuilt our database with strict citation requirements:

Phase 4: Direct Database Analysis

The breakthrough came when we got access to the USTC database itself—not just the web interface, but the raw data. 1.4 GB of Access database containing 1,628,578 editions.

Using mdb-tools and Python, we extracted:

We also extracted Latin editions by classification:

Key finding: University publications and religious texts account for 52.8% of all Latin editions. These are the least translated categories.

What We Built

The result is a comprehensive database with:

• Verified statistics from 45+ sources
• Complete language breakdown of 1.6M editions
• Decade-by-decade data from 1450 to 1700
• Subject classification of all Latin works
• Inventory of all major translation series
• Catalog of open-access resources

What We Still Don't Know

The research revealed as many gaps as facts:

• Loeb Latin/Greek split: Harvard doesn't publish the exact breakdown. The commonly cited ~50/50 is an estimate.
• Total translations ever made: No comprehensive count exists. Our estimate of 8,000-15,000 works is rough.
• Dissertation translations: Thousands of Latin works have been translated in PhD dissertations. Most are never published.
• Quality assessment: We can count translations but not evaluate them. Some Victorian translations are unreliable.

The Takeaway

The gap between what exists and what's accessible is larger than we expected. Half a million Latin editions in the USTC alone. Perhaps 2% translated. The infrastructure to fix this—AI translation, digital archives, open access publishing—now exists. What's missing is the systematic effort to use it.

The data is now published. The question is what to do with it.