The open-stack library: a futuristic technology from the 18th century

What we lost when we shifted knowledge organization to an algorithmic feed

“Where are all the books?” I asked the campus bookstore clerk earlier this year, as I was gearing up to teach in person for the first time since the COVID-19 shutdowns and my paternity leave.

He looked taken aback.

The books, he explained, had all been carted off. Everything moved online during the pandemic, and the university had decided to keep it that way. The rows of shelves around us had once held an eclectic assortment of volumes — photocopied course readers and high quality art books, physics textbooks and experimental plays and Penguin editions of literary classics and everything in between — for hundreds of classes.

Now they were empty. In fact, the shelves themselves were in the process of being disassembled.

“You might be able to find a few in the gifts section,” he added gently, perhaps noticing my crestfallen look.

The exorbitant pricing of college textbooks — which are now typically “rented” as digital copies with rights restrictions — is a story for a different day. As is the fate of my campus bookstore.

Instead, what I want to talk about today is one very distinctive experience that we are at risk of losing in the shift to digital information: the materiality of the open stacks.

The Temple of the Muses, also known as Lackington’s Library, on Finsbury Square in London, in a colored print dated to 1809. Image credit: The British Library.

Inventing the modern library

What do I mean by that phrase, materiality of open stacks? The open stack library — a library in which books are prominently displayed and publicly browsable, rather than guarded in closed rooms or cabinets — is a specific historical invention, one that developed out of the intellectual idealism of the late eighteenth century.

This post goes into detail about the new-fangled “circulating libraries” of Regency-era London, if you want to get a better sense of what these early libraries were like to visit. They’re the kinds of places one imagines Jane Austen might have frequented on her trips to the capital. My favorite is the Temple of the Muses, pictured above: a gorgeous, mostly-forgotten library and bookshop that claimed to house over one million volumes.

As late as the 16th and 17th centuries, libraries had looked and felt very different. Despite the rise of movable type print, the material culture of books continued to owe a strong debt to the medieval tradition. And the medieval library was built on a scarcity model. Books were extremely expensive luxury items, and they were treated as such. The theft of a book which costs $30 in a library of 10,000 volumes is one thing. But imagine the impact of the theft of a book in a library of only 500 volumes — in a world where a single book was the product of months of effort.

For this reason, the chained library (or other forms of closed, guarded libraries) was the norm throughout the medieval period, and well into the early modern era.

A typical set of chained library books from 1655 (via).

So what led to open stacks and circulating books?

The key change was quite straightforward. By the late eighteenth century, there were simply way more books.

In between, there was an awkward transitional phase between the era of books that were literally locked down and the new one of open stacks. For instance, witness this gem of a detail from Wikipedia’s page on chained libraries:

When another local donor, Roger Gillingham, gave another 90 books in 1695, he insisted that the books be chained up, but also that the Library should be opened, free, for the people of the town, providing they were 'shopkeepers or the better class of person.

Perhaps unsurprisingly, the innovation of libraries like this — you could wander around thousands of volumes and simply check one out — prompted outcry about the decay of morals. Barbara M. Benedict, an expert on books and readers in the eighteenth century, writes:

conservative critics considered private reading of publicly lent books self-indulgent, a perversion of the habit of pious literary contemplation into a practice of self-stoking passion — particularly, of course, because readers borrowed books too ephemeral to buy.

Nevertheless, open stack libraries soon became a part of modern life, eighteenth-century style. Libraries succeeded the seventeenth-century coffee shop as an important third place. Benedict calls them “liminal spaces between a public and private identity.” The act of selecting a book to read became a public gesture, an assertion of the individual's taste, beliefs, or character. Perhaps for this reason, before long, the new breed of libraries were also functioning “as community centers, offering cards, games, and even dancing ‘after the blinds had been drawn to cover the bookshelves.’”

Too much to know

At their core, though, these libraries were spaces for publicly organizing knowledge. And because the actual space taken up by knowledge was growing so rapidly — in the form of ever-increasing stacks of bulky printed books — there was a need for experimentation at multiple levels. New techniques of information management appeared alongside the open stacks: the table of contents, the index, the footnote.

Yes, variants of all of these existed earlier. But it was in the eighteenth century that they were established as a set of reliable norms. Readers could now expect to find an index, or footnoted citations, rather than it being a quirk or innovation of particular authors. (As an aside, there are two great books on this topic that I recommend to anyone looking to learn more: Anthony Grafton’s The Footnote and Ann Blair’s Too Much to Know).

This was, in hindsight, an early step on the path of applying new levels of abstraction to knowledge that has brought us to the present: me, writing words for you, on the Internet. But although you really can draw a line from Enlightenment-era indexes to the early history of networked computing, it’s also worth noting how deeply, sometimes hilariously eclectic and just plain random these early indexes were. For instance, here is a snippet from the index of William Dampier’s Voyages (1697) - it goes on like this for pages and pages:

But underneath the eccentricity is a goal — finding new ways to organize information drawn from an ever-growing horde of data — which has clear connections to our algorithmic, networked present. Perhaps the best example of this is John Wilkins’ universal language where the beginning of a word indicated the category of knowledge it fell into, allowing the reader to know roughly what was being described, even if they had never seen a particular word before. It was like the Dewey Decimal system, but for all language:

Wilkins’ forty categories.

The glory days were the beginning of the end

These early attempts at systematically organizing knowledge were beautiful and eccentric in a distinctively early modern way. They were also unsuccessful. Early modern libraries had no standardized classification system, and no international governing bodies to regularize anything. (An example of the challenges: the early 17th-century Cotton Library collection, which was one of the kernels for the British Library, was originally organized by reference to the busts of the Roman emperors who sat atop the shelves in Cotton’s long-since vanished home).

The late 19th and early 20th centuries were the key period of change. Arguably, this was the golden age of the open stack library. Andrew Carnegie, the wealthy industrialist, funded the creation of over 2,500 circulating libraries worldwide between the 1880s and 1920s. Many are still in existence. (My two-year-old daughter can sometimes be found crawling around on the comfy chairs of Santa Cruz’s last surviving Carnegie library).

However, this period also witnessed the rise of sophisticated systems for organizing closed stack libraries, such as the repositories of national governments. The Library of Congress classification system, developed in 1897, is a prime example. While it improved the efficiency of cataloging and locating books, it introduced a new level of abstraction that portended a move away from the eclectic, open approach of earlier libraries. This system, organized by subject matter, required users to learn how to use new techniques for retrieving metadata, like card catalogs. Even before the digital era, the library experience was moving away from tactile interaction with books to an experience mediated by a structured, almost algorithmic system.

The changes in libraries were paralleling the story that James Gleick so brilliantly tells in his book The Information: the rise of big data in the 1930s and 1940s. The changing research methods of Margaret Mead, the anthropologist at the center of my new book Tripping on Utopia, give a vivid sense of the information landscape at this time. By 1938, working in New Guinea and Bali with her third husband Gregory Bateson, Mead had developed an almost obsessively precise system for recording events in minute detail, not just via textual notes but also via cameras. Here’s an example that I cut from my book:

On the morning of Tuesday, July 5, 1938, Mead and Bateson spent an hour recording the birth of a baby in a Balinese village using a Leica 33mm camera. They arrived at 10:30 AM, in time to find the baby being born and bathed, “cord and afterbirth still attached.” At 10:35, Mead noted what the mother’s skirt was made out of (grass). At 10:40, Bateson filmed a closeup of the mother washing her legs. At 10:45, a dog licked the baby's hand. Two minutes later, one of the boys watching pulled apart a flower.

The notes were typed up and and timestamped to allow them to be “synced” to the visual footage, then organized according to a complex filing system.

In the same period, J. Edgar Hoover (who worked as a young man in the Library of Congress, and was fascinated by its knowledge organization) was perfecting the FBI’s system for surveilling Americans, while Vannevar Bush was beginning to dream about his Memex.

These were all early efforts in what we might now recognize as a database-driven approach to organizing knowledge. This was a crucial stepping stone towards the digital methods we use today, a vast and ongoing shift to an algorithmic approach to handling large volumes of information.

It was also the beginning of the end of the open stacks.

Open stacks matter because they materialize the abstract

The open stack library is like an early modern index that has been turned into a building. It’s organized but eclectic, it rewards browsing, and it allows you to pursue individualized interests within a “social” framework (everyone has the same index and the same library shelves, as opposed to personalized algorithmic recommendations which put us in silos).

Much has been written about the isolating tendencies of modern life, and specifically the so-called loneliness epidemic and how it worsened during the pandemic. In a small way, I think I’ve noticed a version of this shift over the course of my time as an educator. Libraries are still community centers, of course (though rarely the site of Regency balls anymore). But I do feel like they have become less of a central place in the lives of my students. It is not uncommon nowadays to encounter university students who are in their second, third, or even fourth year and who have never checked out a book at the library. (If this sounds implausible to you, it did to me too — until I started asking my lecture classes this question directly).

This is all probably unavoidable, and not entirely bad. But looking back, working as a library shelver was probably the best educational experience of my life. Physically holding and organizing the books in my section (I was in charge of Library of Congress sections B through D, covering philosophy, psychology, religion, and history) was incredibly informative. I could flip through physical books at will, in a semi-random, semi-structured, and above all, deeply material way. It was a priceless education in how ideas and thinkers “fit together.”

Today, I use algorithmic recommendation systems and digital search all the time. I rely on them for my job. On the whole, these systems work very well.

But the materiality and affordances of an open stack library cannot be replaced. They certainly can’t be matched by rented digital textbooks — surely the worst book delivery method ever contrived.

Above all, the old school open stack library was fun. I miss it.

A side note: the book wheel

I wanted to find some reason to mention my favorite piece of early modern book technology, but couldn’t manage to work it into the discussion above. So here it is as a bit of trivia. The book wheel.

This person made a modern replica of the book wheel. Princeton historian Anthony Grafton, legend that he is, apparently did too. The only picture of Grafton’s replica that I can find online is 16 years old and super small, but you get the idea:

---

Weekly links

• Francesca Mari’s great 2012 post on book organization through history, which I found while researching this post (The Paris Review Daily).

• “According to the archaeologists and scientists who have examined it, the brain has a ‘resilient, tofu-like texture.’ It is not clear why the Heslington brain survived, although the presence of a wet, anoxic environment underground seems to have been an essential factor.” (From Wikipedia’s fascinating and morbid page on the Heslington Brain, “the best-preserved ancient brain in the world.”)

• New evidence suggesting that the world’s oldest known pyramid is in Indonesia… and extraordinarily ancient (Phys.org writeup, original journal article).

• “Forensic toxicological analyses reveal the use of cannabis in Milano (Italy) in the 1600’s” (Journal of Archaeological Science)

If you’d like to support my work, please pre-order my book Tripping on Utopia: Margaret Mead, the Cold War, and the Troubled Birth of Psychedelic Science (which just got a starred review in Publisher’s Weekly) or share this newsletter with friends you think might be interested.

As always, I welcome comments. Thank you for reading!