Lost in endless libraries. The subtle art of knowing which books not to read or to forget

In the 16th century, the humanist Thomas Platter devised a peculiar method to prevent himself from falling asleep over his books: he would put raw turnips in his mouth. Sometimes he resorted to cold water, at other times even to sand, hoping that the crunch of the grains between his teeth would help him stay awake. “I wanted to study Latin, Greek and Hebrew at the same time,” he recounted in his autobiography, “and for many nights I slept very little, tormenting myself in the struggle against sleep.” His tutor, Oswald Myconius, warned him against those endless all-nighters, but spared him reprimands when fatigue overcame him during lessons.

Born into an extremely poor family in Valais, an Alpine region in what is now Switzerland, Platter began his life as a goatherd. After years spent wandering between schools in Central Europe, he went on to become a teacher and printer in Basel. Whilst working in a rope-maker’s workshop, he spent the little money he earned on candles so that he could continue reading at night. He was not the only one battling sleep. Among scholars, tales circulated of men who kept their feet in cold water or tried to sleep every other night. The wife of the French philologist Adrien Turnèbe attributed her husband’s death in 1565 to the many nights he had spent poring over books. For those who, like Platter, had grown up in poverty, devoting even the hours stolen from work and sleep to study meant hoping for a different life. And yet not even the most determined all-nighters would have allowed one to read everything. Time for reading was already a scarce resource back then. Today, the disparity between what we can read and the hours at our disposal has reached proportions that Platter could never have imagined. Digital libraries and search engines provide us, in a matter of seconds, with texts that would once have required weeks of research; artificial intelligence promises to compare and summarise them. But none of these technologies can give us the time needed to read everything.

In 1416, the Florentine humanist Poggio Bracciolini, having been left without a post following the deposition of the Pope in whose service he had worked, was able to devote himself more freely to his greatest passion: searching for ancient manuscripts in monastic libraries. In the Abbey of St Gall, he discovered a complete copy of Quintilian’s Institutio oratoria, hitherto known only in fragmentary form—a prisoner abandoned in a dark cell with an unkempt beard and tattered clothes, as he described it in a letter to a friend—which he could now finally bring back into the light. The following year, he also discovered a manuscript of Lucretius’ De rerum natura, a poem that presented a view of nature far removed from the Christian one.

That passion for the revival of antiquity found a formidable ally in the printing press. Rediscovered texts were joined by new ones; the classics were supplemented by the works of their commentators; and first editions were accompanied by reprints, translations and expanded editions. The number of books grew as curiosity, technical capabilities and the interests of those who printed and sold them all increased. There was no copyright in the modern sense: privileges and authorisations could protect individual editions, but many works were reprinted, plundered or reissued by unscrupulous competitors. The authors themselves fuelled the book market, revisiting their own texts to add pages, correct passages and offer the public another edition. The humanists’ ideal library continued to expand.

In 1525, Erasmus of Rotterdam wondered whether there was still any corner of the earth spared from the ‘swarms of new books’. “Even if, taken one by one, they contained something worth knowing,” he observed, “their sheer quantity would constitute a serious obstacle to learning.” He wrote these words whilst commenting on a Latin adage and paying tribute to Aldo Manuzio, who had died ten years earlier. The great Venetian printer had been a model of editorial rigour for him; many others, in his view, were flooding the bookshops with mediocre works. The complaint had an amusing and, in some respects, paradoxical twist: Erasmus was among the writers who most effectively contributed to the proliferation of books. His Adagia, first published in 1500 with a few hundred proverbs, had become, in the Venetian edition of 1508, a collection of thousands of entries, accompanied by scholarly commentaries and references. And they continued to grow. Readers were invited to purchase a book they might already own, but which had, in the meantime, become much larger.

The rush to publish also played its part. In 1584, a French scholar, busy compiling a major bibliography of his country’s writers, recounted how he would hand over twelve handwritten pages to the printers every day whilst proofreading the previous ones. By the time he reached the end, he could no longer review the beginning: it had already been printed. Speed guaranteed quantity, sometimes at the expense of quality, and that quantity imposed a new burden on readers. One had to at least know what had been published before even deciding what was worth reading.

In 1545, in Zurich, Conrad Gesner undertook a project intended to make this task less daunting. His "Bibliotheca universalis" catalogued authors and works in the three major languages of European scholarly culture: Latin, Greek and Hebrew. Gesner was twenty-nine years old and, after some three years of hard work, compared the relief he felt at having completed the book to that of a shipwrecked man reaching dry land. A physician, naturalist and philologist, he had consulted volumes, catalogues and bibliographies, even recording works he had not seen for himself. The alphabetical order made it possible to find an author; three years later, Gesner organised the bibliographic information he had gathered by discipline and subject, offering readers another way to navigate the books. Readers could begin with a name or a subject, follow the cross-references, and identify the works dealing with a particular issue. In The History of Animals, the major zoological work he began publishing in 1551, Gesner adopted a similar principle: he encouraged readers to consult the volumes to find what interested them, without feeling obliged to read them from start to finish.

That universal bibliography could also serve purposes very different from those envisaged by its author. In Rome, the censors tasked with compiling lists of banned books used it as a catalogue to be consulted in reverse: they searched it for titles and authors to subject to condemnation. A catalogue created to raise awareness of books thus became a tool for identifying those to be withheld from readers. After all, the lists of banned books were bibliographies in their own right, albeit compiled with the opposite intention. And they faced a difficulty similar to that encountered by Gesner: whilst a list was being compiled and printed, other books were rolling off the presses, other titles were entering the market, and other authors were changing their names or places of publication. The list was chasing a target that kept moving. Gesner had also realised something else: to bring order to one’s reading, one needed to be able to take it apart. Among his papers, there remain printed fragments and handwritten notes that had been cut out for reuse. If an interesting passage was already on a page, why copy it out? It was enough to cut it out and move it amongst the materials to be classified. Our ‘copy and paste’ had an ancestor made up of scissors, paper and glue. For centuries, scholars had been transcribing extracts into notebooks organised by subject, but a quotation filed under the heading ‘prudence’ might prove useful, a few months later, when discussing ‘betrayal’. With loose-leaf sheets, it was easier to change one’s mind without having to rewrite everything.

In 1561, the humanist Conrad Lycosthenes died in Basel, leaving behind a vast collection of notes gathered over some fifteen years of reading. His stepson, Theodor Zwinger, used them to compile a vast compendium (the Theatrum vitae humanae, 1565) of historical episodes, virtues and vices, anecdotes and memorable instances, ready to be used in a lecture, a sermon or a speech. Some of the slips of paper used to compile it have survived in the University Library of Basel: a quotation here, an episode there, each with its own heading, to be moved and rearranged as required. First one gathered material, then decided where to place it. An ingenious solution, especially for someone who frequently revised the structure of the work whilst continuing to read.

The American historian Ann M. Blair has reconstructed this and many other events in a learned and brilliant study, Too Much to Know (Yale University Press, 2010). Notebooks, indexes, index cards and reference works enabled scholars to commit an ever-increasing portion of the work of memory to paper. It is a pity that the tools destined to dominate the mass of books tended to grow with the same unbridled energy. Zwinger’s repertory grew from around 1,500 pages in the first edition to 4,500 in 1586; one section listed more than 200 ways of dying, a catalogue that hardly anyone would have wished to verify in its entirety. In 1631, an even more extensive version was published: seven enormous volumes, some 7,400 pages, plus an eighth volume devoted to the indexes. To avoid getting lost in one library, they had ended up building another, complete with instructions on how not to get lost there either.

During the very same decades in which Zwinger was expanding his repertoire, Michel de Montaigne was cultivating a very different relationship with books in the library of his tower. In his Essays, he confessed to having such a poor memory that he often forgot what he had read. “I leaf through books, I do not study them,” he wrote. It would happen that he would pick up a volume he had read carefully a few years earlier, already full of his own notes, and find it almost unfamiliar. He had therefore got into the habit of noting, at the end of books he expected to read only once, the date of reading and the judgement he had formed of them, so as to recall at least the general impression they had left on him. He did not seem overly troubled by such lapses of memory. If a book bored him, he would pick up another; if a passage proved too difficult, he preferred to leave it alone. Even in education, he was wary of indiscriminate accumulation: for a young person, he would have preferred a teacher with a ‘well-formed mind’ rather than a ‘well-filled mind’.

Almost four centuries later, as Xavier Nueno recalls, amongst other things, in his intelligent and enjoyable "Un saper leggero. Breve storia dell’eccesso di informazione" (recently published by Einaudi), the French critic and semiotician Roland Barthes harboured a surprisingly similar desire. In the final years of his life, weary of the constant demands for articles, lectures and conferences, he dreamt of returning to the pleasure of the amateur, who reads out of curiosity, without the obligation to turn every reading into a new piece of writing. In the library of his holiday home, books could finally be opened without having to extract something from them to publish. Montaigne allowed himself to forget; Barthes would have liked to allow himself to read without having to write. Two freedoms which, for those who make their living from books, are by no means a given.

The suspicion that books, as their numbers grew, might hinder knowledge was, after all, much older. Seneca had observed that a multitude of books scatters the mind; Lucian had mocked the ignorant rich who bought precious volumes without reading them. In one of Plato’s dialogues, an Egyptian king reproached the inventor of writing for having created a tool destined to weaken the memory: rather than exercising it, people would rely on the written word to recall what they had learnt, ultimately confusing the ability to retrieve information with the possession of knowledge. And yet modern scholars continued to refine precisely those aids to memory, going so far as to construct special pieces of furniture to house what the mind could not retain.

In the 17th century, the English clergyman Thomas Harrison, who had ended up in prison after accusing a judge of high treason, designed a cabinet to store and retrieve the approximately one hundred thousand observations he claimed to have collected. He described his invention in the hope of receiving a reward, perhaps even his release from prison. Loose sheets of paper were practical, but they had a flaw that Harrison knew only too well: all it took was a draught, entering through a window or a door left open, to scatter the work of hours. It had happened to him several times, he recounted, and each time he had had to laboriously retrieve those ‘sibylline leaves’, as some contemptuously called them. The idea was taken up in Hamburg by the German jurist and scholar Vincent Placcius, who in 1689 published a treatise on the art of collecting and organising extracts from one’s reading. In it, he also described the workings of a special cabinet for notes, fitted with bars, hooks and labels indicating the subjects. The slips of paper were hung beneath their respective headings; to add a piece of information, one simply had to place it in the correct spot. Anyone who had forgotten a quotation could retrieve it, provided they remembered under which heading they had classified it. This is a detail familiar to anyone who has ever searched for an archived document whilst placing too much trust in their ability to recall its name. The cabinet foreshadowed the card index systems which, until a few decades ago, filled the studies of researchers and writers, with their index cards to be sorted, retrieved and put back in their places. But no card index compiled itself. Someone had to choose the passages to copy, select the topics under which to classify the extracts, and decide what to omit. Anyone consulting a catalogue inherited those decisions, perhaps without realising it. They might find a quotation from Seneca without ever having opened a work by Seneca, or retrieve an episode from Roman history through a collection of examples, without knowing the context from which it had been taken.

The selection of information also became an editorial endeavour. In 1665, two periodicals dedicated to the latest developments in the scholarly and scientific worlds were launched in Paris (Journal des sçavans) and London (Philosophical Transactions), the forerunners of modern academic journals. Three years later, in Rome, the Giornale de’ Letterati began publication; following the French model, it informed readers about works that had just been published in Italy and abroad. Through reviews, extracts and news of new discoveries, it was possible to keep up to date without having to acquire every volume or read every treatise. But even these resources multiplied, adding new pages to those that no one could already keep up with in their entirety. The German philosopher and mathematician Gottfried Wilhelm Leibniz feared that the abundance of books would ultimately give rise to a new form of barbarism, with knowledge buried beneath its own wealth. Up to that point, scholars had primarily sought to collect, classify and make information accessible. The tools might have been imperfect, the summaries debatable, but the books from which they were derived remained available to anyone wishing to consult them. In the 18th century, someone envisaged a far more radical solution: to preserve only what was worth reading.

In 1771, the French writer Louis-Sébastien Mercier published "L’anno 2440", a novel in which the protagonist falls asleep in 18th-century Paris and wakes up almost seven centuries later. The city has changed profoundly: institutions, customs and habits seem to belong to another civilisation. One of the greatest surprises awaits him in the royal library. Where he had imagined finding a vast collection of volumes, he discovers an incredibly small collection. The librarian explains that the books have undergone a massive culling: many have been destroyed, others abridged and stripped of repetitions. Works deemed useless, verbose or harmful have been consigned to the flames; of the others, only what seemed essential has been preserved. The librarian explains that the task was carried out by the country’s finest minds. “They extracted the substance of a thousand folio-sized volumes and condensed it into a small booklet.” It is like distilling a plant, preserving its essence in a vial and discarding the rest. The scene has a comical aspect, yet the smile fades when the librarian claims credit for the result: “With enlightened fervour, we have renewed what the barbarians had once carried out blindly.” However enlightened their intentions may have been, those librarians had ended up resorting to one of the oldest tools of censorship: fire.

Mercier’s solution did not prevail. Throughout the nineteenth and twentieth centuries, catalogues, encyclopaedias and bibliographies continued to be produced, whilst microphotography and microfilm promised to condense ever-larger collections into smaller spaces. Meanwhile, the specialisation of academic disciplines made it impossible to master even all the literature within one’s own field. It was possible to preserve more texts whilst, proportionally, knowing less and less about them. In 1959, the physicist Richard Feynman took the dream of miniaturisation to the extreme: why not write the twenty-four volumes of the Encyclopaedia Britannica on the head of a pin? He offered a thousand dollars to anyone who could reduce a page by a factor of twenty-five thousand. The prize was won in 1985 by Tom Newman, a Stanford student who managed to etch the first page of Dickens’ "A Tale of Two Cities" onto a microscopic surface, using an electron beam instead of an engraving stylus.

Miniaturisation, however, primarily solved a problem of space. The digital transformation of libraries tackled another: how to find a page, a name or a phrase amongst millions of volumes. Thanks to large-scale scanning programmes and, from 2004, Google Books, books held in distant collections became searchable on a screen. It became possible to search the text of works that no reader would have had the time to leaf through one by one. The undertaking also sparked fierce controversies over copyright, leading to lengthy legal battles. And the universal library remained incomplete: many books had not been digitised, whilst others could be searched but not read in full. Yet the change was profound. The index, which Gesner had patiently compiled to guide his readers, could now extend to an unimaginable number of pages.

An experiment recalled by Nueno demonstrates just how far the search for media capable of storing ever-increasing amounts of information has come. In 2013, a group of scientists stored a small library – comprising Shakespeare’s sonnets, an article on the structure of DNA, a photograph and an excerpt from Martin Luther King’s ‘I Have a Dream’ speech – within a DNA sequence. The researchers had converted texts and images into sequences of information encoded within DNA molecules, exploiting this material’s extraordinary capacity to store data in a minuscule space. To retrieve a Shakespearean poem or Martin Luther King’s speech, of course, it would not have been enough simply to open a book: it was necessary to read those sequences using laboratory equipment and convert them back into words and images. The experiment demonstrated just how far one could go in attempting to concentrate enormous amounts of information into ever-smaller spaces. To read them, however, it was first necessary to retrieve and decipher them.

Every book, Nueno observes, can be seen as an attempt to distil a library. The writer has read many works, drawing ideas and arguments from them, thereby rendering some of the reading required to produce the book partly superfluous for the reader. “The only legitimate reason for writing is that there are too many books,” he concludes with a paradox. Someone writing a history book, for example, may find themselves reading five hundred books and, in the end, citing only fifty. The other four hundred and fifty are not necessarily useless. Often, they serve precisely to help us understand which questions to ask, which paths to follow and which to abandon. Even the readings that leave no trace help to shape a book. In short, the knowledge we discard is part of what we retain.

The tools for searching for, storing and summarising information have never been so powerful. With artificial intelligence, we can carry out tasks in a matter of moments that once took weeks. There remains a small paradox, one that five centuries of inventions have not erased: to know what we can afford not to read, we must first have read a great deal.