How Many Renaissance Books Get Translated Each Year?

The data

We have a unified translation catalog of 13,862 records drawn from 12 sources: the UNESCO Index Translationum (7,542 records), Open Library (4,442), HathiTrust (1,297), the Loeb Classical Library (184), CUA/Paulist Press (94), Cambridge University Press (62), Routledge and De Gruyter (60), University of Chicago and Indiana (60), Broadview/Norton/Hackett (48), Penguin Classics (37), Brill (24), and Godwin's bibliography (12). Each record represents an English translation of a pre-modern Latin, Greek, or vernacular work, with a publication year.

This is the same catalog we used for our translation census. It is not exhaustive — no single catalog is — but it is the most comprehensive compilation of pre-modern-to-English translations that exists, to our knowledge. It covers every major bilingual series and scholarly press.

We deduplicated across sources (the same Virgil translation might appear in both Open Library and UNESCO) and counted unique author-title combinations per year. Then we looked at the trend.

The peak: 2000–2008

The best-documented period is 2000–2008, when the UNESCO Index Translationum was still actively collecting data from national libraries worldwide. During these years, our catalog captures the most complete picture of global translation output.

At peak coverage, about 145 unique translations per year appeared across all our sources. Of these, roughly 72 were first English translations of a work and 24 were re-translations of something already available (a new Aeneid, another Augustine Confessions). The remaining ~49 could not be classified because they lacked a canonical work identifier in our catalog.

But these are translations of all pre-modern Latin and Greek texts — Cicero, Augustine, Aquinas, everything back to antiquity. The top of the most-translated list is dominated by classical authors: the Aeneid has been translated 41 times, Ovid's Metamorphoses 37 times, Augustine's Confessions at least 28 times. How many of the 145 per year are specifically Renaissance texts — works from the 1450–1700 window we care about?

Who gets translated

Our catalog contains translations of 170 Erasmus works, 121 Thomas More works, 116 Francis Bacon works, 32 Calvin, 22 Ficino, 17 Spinoza, 14 Descartes, 13 Paracelsus, 13 Luther, 9 Agrippa, 8 Copernicus, 7 Pico, 6 Galileo, 5 Kepler, 4 Bruno, and 2 Campanella. These are cumulative totals spanning two centuries of scholarly translation.

But the all-time list tells a different story. The most-translated authors are overwhelmingly classical: Cicero (237 translations), Aquinas (234), Augustine (255 across name variants), Virgil (260 across variants), Ovid (144), Horace (100). Renaissance authors are a minority of translation activity, even in catalogs focused on Latin.

Of the roughly 72 first translations per year during the peak, we estimate that fewer than half are Renaissance-era works. The rest are medieval, patristic, or classical. A reasonable estimate for Renaissance-specific first translations is 20–40 per year at peak, and likely fewer now.

What happened after 2009

Our data shows an apparent collapse after 2008: from 173 unique translations in 2006 to 44 in 2009 and 22 in 2021. But this is primarily a data collection artifact, not a real decline in translation activity.

The UNESCO Index Translationum — which contributes 7,542 of our 13,862 records — effectively stopped receiving comprehensive submissions from national libraries around 2009. Before that year, UNESCO contributed 114–174 records annually. After 2009, it dropped to 22–46 per year, probably capturing only what could be scraped from publisher catalogs rather than national library submissions.

Our other 11 sources (CUA/Paulist, Cambridge, Routledge, Penguin, etc.) collectively contribute only 10–15 records per year. They were never meant to be comprehensive — they cover specific presses and series. Without UNESCO's systematic collection, our post-2009 data captures perhaps a quarter to a third of actual translation activity.

The true post-2009 rate is unknowable from our data alone. But cross-referencing with the major series — I Tatti publishes 2–4 volumes per year, the Other Voice in Early Modern Europe publishes 5–8, plus scattered volumes from Brill, Cambridge, and others — suggests the real rate for Renaissance texts is probably still in the range of 20–50 per year. It may be declining as humanities departments shrink and university press budgets contract, but we cannot prove that from the data we have.

The math

The USTC records 1,071,422 distinct works printed between 1450 and 1700. Our catalog coverage analysis found that about 67,000 of these have a known English translation. That leaves roughly one million untranslated works.

At 30 new Renaissance first-translations per year — a generous estimate of the current rate — clearing the backlog would take:

Even at the 2000–2008 peak — 72 first translations per year, generously assuming all of them were Renaissance texts — it would take 14,000 years.

This is not a problem that can be solved by hiring more translators. It is not a funding gap. The entire global infrastructure for scholarly translation of early modern texts — every university press, every bilingual series, every Classics department, every fellowship program — produces a volume of work that does not meaningfully dent the backlog, even over centuries.

What Source Library is doing about it

Source Library has produced AI translations of approximately 3,800 books in its first year of operation. Not all of these are Renaissance texts, and not all are first translations. But the order of magnitude is clear: in one year, one project, running AI translation on digitized scans, has produced roughly 25–100 times the annual output of all human scholarly translation combined, depending on how you count.

These are not critical editions. They do not have scholarly apparatus, variant readings, or interpretive commentary. An AI translation of a 1509 alchemical treatise is not equivalent to what the I Tatti Renaissance Library produces. Nobody should cite our translations in a dissertation without checking the original.

But they are legible. A researcher who wants to know whether Paracelsus'sArchidoxis discusses antimony can now search the text in English and find out in thirty seconds, instead of spending six months learning early modern German or waiting for a translation that might never come. An art historian studying alchemical iconography can now read the treatise that an emblem illustrates. A philosopher tracing the reception of Neoplatonism can now check an obscure Ficino commentary that has never been in any language but Latin.

The question is not whether AI translation replaces scholarly translation. It does not. The question is whether a million books should remain permanently unreadable to anyone who doesn't read Latin, German, or early modern French, simply because scholarly translation cannot scale. We think the answer is no.

Caveats and methodology

Our catalog undercounts. We capture 13,862 records from 12 sources. The true number of English translations of pre-modern works is certainly higher. We are missing journal articles that include partial translations, unpublished dissertations, small-press editions, and translations published outside the Anglosphere. Our count is a lower bound.

The post-2009 drop is mostly a data gap. When UNESCO stopped systematic collection, our visibility into translation activity dropped by roughly two-thirds. The 22–51 translations per year we see after 2013 are probably a third of the true rate, based on our knowledge of the major series.

Deduplication is imperfect. We match on normalized author + title, but the same work can appear with different titles across sources. Our 3,225 cross-source duplicates are the ones we caught; some duplicates inevitably remain in the “unique” count.

First translation vs. re-translation is approximate. We classify a translation as “first” if no earlier record of the same canonical work exists in our catalog. But if the first translation was published before 1800 (the start of our coverage) or by a source we don't index, we would incorrectly classify a re-translation as a first.

These numbers cover all pre-modern Latin and Greek, not just Renaissance. The 145-per-year peak includes translations of Cicero, Virgil, and Augustine alongside Ficino, Paracelsus, and Kepler. The Renaissance-specific rate is a subset, which we estimate at 20–40 per year but cannot precisely isolate from the data.

The re-translation problem

Of the 5,352 unique works in our catalog that have a canonical identifier, 487 have been translated more than once. The Aeneid has 41 English translations. Ovid's Metamorphoses has 37. The Imitation of Christ has 21. Utopia has 15. These are important works, and each new translation serves a purpose. But every hour a translator spends on the 42nd Aeneid is an hour not spent on the first English translation of a text that has never been read in any modern language.

Meanwhile, 4,865 works in our catalog have been translated only once. And a million more have never been translated at all.

Two centuries of translation, decade by decade

For the historically curious, here is the full trend. Note that our sources have uneven coverage before 1980 — early decades are reconstructed primarily from Open Library and HathiTrust retroactive cataloging and should be read as lower bounds, not absolute counts.

The 2000s were the high-water mark we can measure: 152 translations per year, of which perhaps 145 were unique works. At that pace — which required the combined output of every major university press, every bilingual series, and every funded translation fellowship in the English-speaking world — it would take seven thousand years to translate the USTC backlog.