The Lost Art of the Marginalia

In a copy of Diophantus's Arithmetica in 1637, Pierre de Fermat wrote a small note in the margin observing that the equation xⁿ + yⁿ = zⁿ has no positive integer solutions for any n greater than 2. He added that he had a wonderful proof which the margin was too narrow to contain. The note went unnoticed during his lifetime; his son discovered it after Fermat's death and published the marginalia along with the rest of his father's mathematical correspondence. For 358 years the conjecture resisted every attempt at proof until Andrew Wiles finally cracked it in 1995. The note in the margin of an obscure mathematics text became one of the most consequential paragraphs in the history of mathematics.

Fermat's note is famous, but it is not the strangest thing to find in a margin. The marginalia of past centuries are some of the most candid records of how people read. They show the reader thinking in real time, agreeing and disagreeing, making connections, calling out errors, leaving notes for future readers and for themselves. They are conversational in a way that polished writing is not. Coleridge, who left thousands of pages of marginalia in books he borrowed and seldom returned, treated marginalia as a literary form in its own right. He published collections of his notes after the fact. Friends and rivals would seek out the books he had written in for the privilege of reading along.

The practice has roots much older than the printed book. The Talmud, in its standard format established by the Bomberg printing of 1523, is itself a structure of marginalia: the central column of Mishnah with surrounding columns of commentary by Rashi, Tosafot, and others, each in conversation with the others. Reading the Talmud is reading marginalia all the way down. A serious student does not read the central text first and then the commentaries; the entire page is a single object whose meaning emerges only when the columns are read together. The visual layout, perfected over centuries, is itself a theory of how learning works: ideas live in conversation, not in isolation.

The medieval gloss

European medieval manuscripts followed a similar logic. The glossa ordinaria on biblical texts, perfected at the cathedral school of Laon in the early 12th century, was a standard apparatus of interlinear and marginal commentary that traveled with the biblical text and carried the accumulated interpretation of generations. By the 14th century, university manuscripts of Aristotle and the church fathers came pre-printed (or rather, pre-copied) with the same conventional marginalia, so that a Parisian student and a Bolognese student could discuss the same passage with the same secondary readings at hand.

This was the original hypertext. The text in the center pointed outward to the marginalia; the marginalia pointed back to the text and to each other. The reader navigated this network by eye, pulling threads through the page. The technology was paper and ink, but the structure was the structure of linked thought. When Vannevar Bush imagined the Memex in 1945, and when Ted Nelson coined "hypertext" in 1965, they were articulating digital versions of what the medieval gloss had been doing for 800 years.

The early modern explosion

Print made marginalia easier. A printed book left wider margins than a manuscript could afford, and the assumption was that the reader would fill them. The annotated book was the studied book; the un-annotated book was the unread book. Erasmus's library, when it was reconstructed by scholars in the 20th century, turned out to contain hundreds of books with his annotations, including books he disagreed with so vehemently that the marginalia became their own polemic against the printed text. Erasmus's marginalia are sometimes more interesting than the book they comment on.

The early modern reader's apparatus included specific tools for marginalia: the commonplace book, in which extracts from many sources were copied and indexed under thematic heads; the pencil sharpener, in pencil's various pre-modern incarnations from lead point to graphite stick; and the knife, used to scrape out wrong annotations or to cut pages free for use in another binding. Books were not finished objects but materials in process, modified and recombined and annotated by each owner.

The Coleridge case

By the early 19th century, marginalia had become a literary form. Coleridge's marginalia, collected and published across multiple volumes in the 20th century, run to thousands of pages. His friends loaned him books knowing he would write in them, and they considered the notes valuable enough to want the books back to read what he had written. Some of his most acute critical thinking happens in margins: a sustained argument with Kant, an evisceration of Schelling, an appreciation of Hooker that crystallized his theology. None of this would survive if Coleridge had read the books and put them back on the shelf.

Edgar Allan Poe wrote essays specifically titled "Marginalia" in which he experimented with the form: short paragraphs of observation and aphorism, presented as if pulled from his reading notes but composed with literary care. Poe's "Marginalia" essays are a bridge between the casual annotation and the polished short essay, and they show how seriously the form was taken at the time.

The disappearance

The decline of marginalia tracks with several trends in the 20th century. Public library reading discouraged annotation: the book was not yours, and writing in it was vandalism. Cheap mass-market paperbacks felt disposable, and writing in them felt grandiose. The school discipline of "do not write in books" carried into adulthood as a moral aversion, even for books one owned. By the late 20th century, writing in a book felt to many readers like an act that required explanation.

The turn to digital reading reversed some of this. Highlighting and annotation are core features of e-readers, and Kindle's "popular highlights" feature reveals that readers do annotate when the friction is low and the social cost is zero. But the e-reader annotation is not quite the marginalia of the past. It is private. It is searchable but not browsable. It does not accumulate on the page in a way that the next reader can see at a glance. The e-reader removes the social and material dimensions that made marginalia rich.

What is lost when marginalia disappears is more than the convenience of a note. The act of writing in a book changes the reading. Reading with a pen in hand is different from reading with hands free; the pen creates a small expectation of response, a low-grade conversation with the text. The book that has been annotated rewards rereading in a way that an unannotated book does not, because the marginalia preserve the past reader's thinking and provide a layer to think against.

For long-running thought, the practice can be revived without nostalgia. A used copy of a book bought specifically to write in. A pencil rather than a pen, for revisability. A habit of dating significant annotations so that years later you can see when you first thought what. The marginalia of past centuries are accidentally preserved by the physical persistence of paper. The marginalia we make today, if we choose to make them, will survive only if we treat them as worth keeping.