Chapter 6

Updated 2026-07-02

"Classifications are theories about the basis of natural order, not dull catalogues compiled only to avoid chaos." — Stephen Jay Gould, Wonderful Life (1989)

A generation after Wilkins's whale and a continent away from Otlet's card index, a twenty-one-year-old student assistant at an Amherst College library sat with the same dream and a much smaller budget. His name was Melvil Dewey, and in 1873 he proposed something almost embarrassingly simple compared to a universal philosophical language: instead of shelving books by the order they arrived, or by the size of their spines, or by which donor had given them — all real systems libraries actually used — sort everything in the building into ten classes, numbered 000 through 900, and let every more specific subject nest inside one of the ten as a longer decimal.¹ Religion gets the 200s. History gets the 900s, minus the part that's really geography. A book finds its number by what it's about, and once it has that number, it has exactly one address on exactly one shelf, forever, regardless of who's looking for it or why.

It is hard to overstate how much friction this removed. Before Dewey, finding a specific book in a large library could mean asking a librarian who had personally memorized the building's idiosyncratic arrangement. After Dewey, the same ten classes worked in the next library over, and the one after that, because the numbers didn't describe one building's shelves — they described the books. By the time Dewey died in 1931, his system was running in the large majority of public and school libraries in the country, and it still runs in most of them today: libraries in more than 138 countries now use it, in translations spanning over thirty languages, a piece of nineteenth-century infrastructure so thoroughly absorbed into how people search for books that almost nobody who uses it has heard the inventor's name.²

Open the 200s, though, and the seams Wilkins would have recognized are right there. Numbers 200 through 289 — nine-tenths of the entire block reserved for the subject of religion — belong to Christianity: its history, its denominations, its theology, broken down in detail fine enough to give Anglican liturgy its own number distinct from Methodist liturgy. Number 290, one-tenth of the block, holds everyone else.³ Dewey wasn't being secretive about his assumptions. He was an American librarian building a system for American library patrons in 1876, and he built the shelf around who he expected to be standing in front of it. The numbers don't lie about that. They're some of the most honest numbers in the system, in fact, because they never had to pretend to be neutral — nobody in 1876 was asking them to be.

---

A number, though, can only hold what a number can hold, and a book is never only what it's about.

A historian working in an eighteenth-century archive once described dating cholera-era correspondence not by reading the postmarks but by smell: letters mailed out of an infected town were routinely disinfected in vinegar before they were allowed to travel, and two and a half centuries later the letters still carry a trace of it, faint but real, under the ink. He could go further than dating the letters this way. Reading past a correspondent's careful, performed cheerfulness — the determinedly ordinary news, the studied calm — he could sometimes catch the fear the writer was working hard not to put on the page, "betrayed," as he put it, "by a scent of vinegar" the writer never mentioned and probably never consciously registered putting there.⁴

None of that lives in any catalog record. A library classification number can tell you that a letter is correspondence, that it's eighteenth-century, that it concerns a particular town or a particular outbreak. It cannot tell you that the paper still smells faintly of disinfectant, or that the cheerfulness in line four is doing labor the writer never admitted to. This is not a flaw unique to Dewey's ten classes — it would be true of Wilkins's forty categories, or Otlet's thousand, or any scheme built to answer the question what is this, generally, for purposes of finding it again. Classification is built to strip a thing down to the handful of facts that make it findable. Everything that doesn't fit in the fields gets lost in the gap between the object and its record, and the gap is not small. It is most of what the object actually was.

---

Dewey understood, better than almost anyone before him, that a classification scheme is not just an idea — it is a piece of physical equipment, and equipment can be sold. The same year he published his ten classes, 1876, he founded a company called the Library Bureau to manufacture the objects his system actually required: index cards cut to a standard size, the wooden cabinets to file them in, the metal rod that ran through the cards to keep them in order, the little brass-framed label holder on the front of every drawer. Before the Bureau, a library wanting to adopt Dewey's numbers still had to improvise the furniture. After it, a library could order the whole apparatus from a catalog, and thousands did — Dewey's company went on to outfit the wave of Carnegie-funded public libraries that opened across the country at the turn of the century, and the dimensions it settled on for an index card are close enough to the ones still used today that the choice barely registers as a choice anymore.⁵ The numbers and the furniture sold each other. A classification scheme that nobody can physically file books inside of is just a theory; one with matching cabinets is infrastructure, and infrastructure is much harder to question, because by the time anyone thinks to, everything else has already been built on top of it.

---

It is tempting, looking only at the numbers, to imagine Dewey as a careful, tidy man, methodically building infrastructure and nothing else. He was careful and tidy, and he was also one of the stranger figures in the history of American institutions — a man with an appetite for total reform that didn't stop at library shelves.

Dewey believed English spelling was broken the same way library shelving had been broken, and tried to fix his own name as a demonstration project. Melville became Melvil, on the theory that the silent letters were waste. For a period he signed his surname Dui — not Dewey, Dui — until a colleague warned him that the state of New York was not going to keep paying the salary of a state librarian who spelled his own name as a manifesto, and he quietly let it go.⁶ He campaigned for the metric system with the same evangelical energy he brought to cataloging, founded the American Metric Bureau, and never fully accepted that most of the country was not going to follow him there either.

The reform impulse had a much darker edge once it left the page. Dewey founded an exclusive retreat in upstate New York called the Lake Placid Club, and the club's membership policy — which he personally wrote — barred Jews, along with Black people and other groups he considered undesirable, from joining. In 1904 a formal petition reached the New York State Board of Regents demanding his removal as state librarian specifically over the Lake Placid Club's exclusion policy; the regents stopped short of removing him but issued a public rebuke, and within a year he resigned the post. Multiple women who worked under him reported persistent, unwanted advances over decades of his career, accusations serious and well-documented enough that in 1906 a movement rose within his own profession's association to censure him, years before the word "harassment" had any institutional weight behind it at all. In 2019, the American Library Association voted — without debate, the resolution passing overwhelmingly — to strip his name from its highest professional honor, the medal that had been called, for ninety years, the Melvil Dewey Medal.⁷ It is now called the ALA Medal of Excellence. The classification system kept his name. The award did not.

None of this changes what the 200s say about whose religion gets nine-tenths of a shelf. If anything it sharpens the point the numbers were already making: the man who built the most quietly authoritative classification system in American life had, by his own choices and on the record, very clear opinions about whose existence belonged at the center of things and whose belonged at the edge. A classification system is not a mirror held up to the world. It is a decision, made once, by someone, about what counts as ordinary and what counts as exceptional — and the decision doesn't stop being a decision just because it's been sitting on the shelf so long that everyone has forgotten anyone made it.

---

Long after Dewey, the same logic kept running, mostly unsupervised, in buildings that have nothing to do with libraries.

Walk into an ordinary supermarket and the cereal gets an aisle of its own, no qualifier needed, no sign over the top explaining what it is or why it's there. Three or four aisles over, soy sauce and curry powder and a bag of dried shrimp share a single shelf under a sign that says something like WORLD FOODS or INTERNATIONAL, as if cereal arrived from nowhere in particular and curry powder arrived, specifically, from "international." Olive oil sits with the other oils. Soy sauce does not sit with the other sauces. Nobody on the store's planning committee sat down and voted to make one cuisine the default and the rest the exception, the way the IAU voted on Pluto — it never needed a vote, any more than the 200s needed Dewey to announce his assumptions out loud. The system reproduced the same shape: ample, unmarked shelf-space for what the builder of the system considered normal, and a single crowded shelf, with a label apologizing for itself, for everything else. Any kid pushing a cart down that aisle learns the same lesson the 290s teach, without anyone in the building intending to teach it: some food is just food, and some food needs a sign.

This is what it looks like when a system Dewey would recognize keeps running for a hundred and fifty years without anyone going back to check its assumptions against a world that kept changing underneath it. The shelves are not malicious. They are not even, most of the time, particularly conscious of themselves. They are simply old decisions, still deciding.

---

What Dewey never quite solved — what nobody working in his tradition solved, because the tradition itself didn't allow for it — is what to do with a book, or a person, or a meal, that genuinely belongs in two places at once. The 200s have no number for a household where two faiths sit at the same table. The fiction shelves have no number for a novel that is, with equal seriousness, a mystery and a memoir. One book, one number, one shelf: the rule that made Dewey's system so usable is the same rule that made it incapable of holding anything with more than one true address. A few decades after Dewey finished building his ten boxes, a mathematician on the other side of the world, sent to London to study librarianship almost by accident, would stand in a department store and watch a child's toy solve the problem Dewey's whole system couldn't.

---