Every book printed before 1501 is called an incunabulum — a “cradle book,” from the infancy of print. There are about 30,000 of them. The British Library’s Incunabula Short Title Catalogue (ISTC) records each one: who wrote it, who printed it, where, when, and in how many surviving copies.
These are not obscure facts. They are the data points of a revolution. Between 1450 and 1500, printing went from a single workshop in Mainz to hundreds of presses across Europe. Ideas that had circulated in a few dozen manuscript copies were suddenly available in editions of hundreds. The question is not just what was printed, but who was connected to whom, and where did ideas cluster.
We built a knowledge graph to find out.
The Graph
Source Library now holds nearly a thousand ISTC-catalogued incunabula — scanned, OCR’d, and many translated into English for the first time. We cross-referenced each book with its ISTC metadata: author, printer, place of publication, and subject classification. Then we rendered the connections as an interactive force-directed graph.
Five kinds of nodes:
- Books — sized by page count
- Authors — Aquinas, Aristotle, Ficino, Galen, Albertus Magnus…
- Printers — the workshops that brought these texts to market
- Places — Venice, Nuremberg, Strasbourg, Cologne, Rome…
- Subjects — philosophy, medicine, astrology, natural magic…
Every edge is a relationship: this author wrote this book, this printer produced it, this city housed the press. Click any node and its connections light up. Search for “Ficino” and watch the Florentine Neoplatonist network emerge from the noise.
What the Graph Reveals
Venice was the center of everything
The single largest cluster in the graph is Venice. By the 1480s, Venice had more active printing presses than any other city in Europe. The reasons were commercial (a major trading port), legal (relatively tolerant censorship), and technical (access to paper from the Veneto mills). The graph shows Venice connected to nearly every major author in the collection — Aristotle, Galen, Aquinas, Avicenna — because Venetian printers like Aldus Manutius, Jenson, and the Locatelli workshop printed them all.
Aristotle dominates, but not alone
Aristotle is the most-connected author node, which surprises no one familiar with the period. What the graph shows that a reading list cannot is how Aristotle dominates: through commentaries. Most of the Aristotle cluster consists not of Aristotle’s own texts but of works about Aristotle — by Aquinas, Duns Scotus, Averroes, Albert the Great. The medieval university curriculum was built on these commentaries, and the printers served the universities.
Medicine and philosophy are entangled
Galen and Hippocrates sit close to Aristotle in the graph, connected by shared printers and shared cities. This reflects the historical reality that medicine, natural philosophy, and what we would now call “science” were not separate disciplines. A printer who produced Aristotle’s Physics also produced Galen’s Ars Parva, because the same students bought both.
The esoteric fringe is a fringe
Books on astrology, alchemy, and natural magic form a distinct but connected cluster at the periphery. They share printers with the mainstream philosophical works (esoteric texts were profitable) but have fewer cross-links to the university curriculum. Ficino bridges both worlds — his translations of Plato and the Hermetica connect the Neoplatonic network to the broader Aristotelian mainstream.
Seeing the Bigger Picture
The ISTC knowledge graph is one of several data experiments we’ve built to visualize the intellectual landscape of early modern Europe. Others, available at secondrenaissance.ai/lab:
Ficino’s Network: Mapping the Transmission
The most directly related experiment is Mapping the Transmission — an interactive network visualization of 57 Renaissance scholars connected by 116 documented relationships, tracing how Neoplatonic ideas traveled from Ficino’s Florentine academy to Copernicus’s heliocentric model. Each scholar node links to actual publication records from the USTC.
The ISTC knowledge graph and Ficino’s network are two views of the same revolution. The knowledge graph shows the production network — who printed what, where, for whom. Ficino’s network shows the intellectual network — who taught whom, who corresponded with whom, who funded whom. Where they overlap is where the printing press met the Platonic Academy: Ficino’s translations of Plato and the Hermetica were among the most influential incunabula ever produced, and they appear in both graphs, connecting the commercial world of Venetian and Florentine printers to the philosophical world of Neoplatonic transmission.
One of the most striking findings from the Ficino network is the gap between intellectual influence and publication volume. Paracelsus leads with 326 editions; Copernicus has 5. Some of the most important nodes — Cosimo de’ Medici, Leonardo da Vinci, Domenico Maria Novara — published nothing at all. The ISTC knowledge graph, built on printed books, is blind to these unpublished influencers. The two visualizations together tell a fuller story.
Rivers of Tradition
A Sankey flow diagram mapping how nine occult and esoteric traditions — Hermetica, Alchemy, Mysticism, Kabbalah, Neoplatonism, and more — flow from 1469 to 1750. Where the ISTC knowledge graph shows who printed what where, the Rivers show which ideas traveled together across centuries.
Printing Centers Map
An animated map showing how Latin printing expanded from Mainz across 48 European cities between 1450 and 1700. Watch Gutenberg’s invention ripple outward — first to Strasbourg and Cologne, then to Venice and Rome, then to the university towns. The knowledge graph gives you the network; the map gives you the geography.
USTC Explorer
The USTC Explorer is a dashboard built on 556,000 records from the Universal Short Title Catalogue. It shows language trends, accessibility funnels, and subject breakdowns across 250 years of European printing. The ISTC covers only the first 50 years; the USTC carries the story forward.
Book Atlas
A 3D constellation of 3,500+ texts clustered by AI embeddings and reduced to three dimensions via UMAP. Books that are semantically similar sit close together. Spin it, zoom in, click a star — it’s the same data as the knowledge graph, but organized by meaning rather than metadata.
Translation Lag
A scatter plot revealing the gap between when a work was composed and when it was first translated into English. Some of the incunabula in the knowledge graph were written by authors who lived two thousand years earlier — Aristotle, Galen, Hippocrates. Their works waited centuries for Latin translation, then more centuries for English. The lag itself is a story about which ideas Europe was ready to hear.
Image Deep Zoom
An explorable canvas of 39,783 illustrations from historical books, clustered by visual similarity. Many of the woodcuts in incunabula — anatomical diagrams, astronomical charts, botanical illustrations — appear here, grouped not by book or date but by what they look like. Zoom in on a cluster of celestial diagrams and you’ll find Sacrobosco’s Sphaera Mundi next to Regiomontanus’s Calendarium, connected by their shared visual vocabulary.
What Comes Next
The knowledge graph currently shows ~970 incunabula from Source Library’s collection. We plan to:
- Cross-reference ISTC with USTC to find additional digitized copies from libraries across Europe — Gallica, the Bodleian, HAB Wolfenbüttel, the Vatican
- Add entity co-occurrence edges — when two books mention the same person or concept in their text, link them
- Add work-level grouping — many of these incunabula are different editions of the same text (there are at least four different printings of Albert the Great’s Secreta Mulierum in the collection). Linking editions of the same work collapses the graph into a more meaningful shape
- Make it public — the graph is currently an admin tool, but there’s no reason it can’t be a research interface
The incunabula are the root system of modern knowledge. The knowledge graph is an attempt to see that root system whole.
Source Library is a digital library of primary sources in the Western esoteric tradition, classical antiquity, and early modern natural philosophy. All texts are scanned from original editions, OCR’d, and translated into English using AI. Visit sourcelibrary.org.
The research lab at secondrenaissance.ai/lab contains interactive data experiments exploring half a million Latin works printed between 1450 and 1700.