As the paperless future approaches, certain sorts of publications have inevitably moved into the all-digital realm faster than others. Most of us still prefer paper when it comes to beach novels, for instance, or the cherished volumes of our personal libraries. At the other extreme, scientific journals effectively went all-digital years ago, and thanks to GPS, maps and road atlases are quickly following. Last week saw another milestone: the symbolic funeral of paper encyclopedias, with the inevitable announcement that the Encyclopedia Britannica is ceasing print publication.

Encyclopedias, along with other reference works, would seem particularly obvious candidates for digitization. Paper encyclopedias are large, heavy, and expensive ($1,395 for the final print edition of Britannica). They are nowhere near as easily and thoroughly searchable as their digital counterparts. They cannot be easily updated, still less constantly updated. And they are far more limited in size. The 2002 Britannica contained 65,000 articles and 44 million words. Wikipedia currently contains close to four million articles and over two billion words (this information comes, of course, from Wikipedia).

Yet with the disappearance of paper encyclopedias, a part of the Western intellectual tradition is disappearing as well. I am not speaking of the idea of impartial, objective, and meticulously accurate reference. There is no reason this cannot be duplicated in digital media. Even Wikipedia, despite its amateur, volunteer authors, has emerged as an increasingly important and accurate reference tool, reaping respectful commentary last month from no less an authority than William Cronon, president of the American Historical Association. And I am not speaking of the pleasures that come from the serendipitous browsing of handsome encyclopedia volumes, in which the idle flip of a finger takes one from Macaroni to Douglas MacArthur, and thence to Macao, Macbeth, and the Maccabees. The internet provides its own opportunities for serendipitous discovery.

But the great paper encyclopedias of the past had other, grander ambitions: They aspired to provide an overview of all human knowledge, and, still more boldly, to put that knowledge into a coherent, logical order. Even if they mostly organized their articles alphabetically, they also sought ways to link the material together thematically—all of it. In 1974, for instance, the editors of the Encyclopedia Britannica added to the work a one-volume “Propaedia,” which sought to provide a detailed outline of human knowledge, while referencing the appropriate articles of the encyclopedia itself. Large headings such as “Life,” “Society,” and “Religion” were subdivided into forty-odd “divisions” and then further into hundreds of individual “sections.”

The single greatest and most ambitious such attempt to order knowledge, however, appeared more than two hundred years earlier: the legendary French Encyclopédie edited by Denis Diderot and Jean le Rond d’Alembert (it initially appeared in 32 huge folio volumes, and is now available online in French in its entirety, and partially in English through an ongoing collaborative translation project). It was not the first encyclopedia. Predecessors date from even before the invention of printing, and proliferated during the Renaissance, as Harvard historian Ann Blair has recently shown. Like present-day encyclopedias, Diderot and d’Alembert’s used alphabetical organization. But as d’Alembert himself explained, in a “Preliminary Discourse” to the work that became one of the key philosophical texts of the Enlightenment, it also aimed “to set forth … the order and connection of the parts of human knowledge.” It did so in several ways: through the “Preliminary Discourse” itself; through a visual “Map of the System of Human Knowledge”; and through a careful system of cross-references between articles. Diderot himself grandly claimed that the work contained all the knowledge necessary to save mankind from a new dark ages (“What gratitude the next generation following such troubled times would feel for the men who had … taken measures against their ravages by protecting the knowledge of centuries past!”).

As historians know well, the Encyclopédie also had strongly polemical intentions. It attempted not just to order knowledge, but to reorient readers away from earlier, avowedly religious systems. It aimed, quite explicitly, to advance toleration, combat religious fanaticism, and promote a spirit of pragmatic, rational inquiry and experimentation. The “Map of Knowledge” deliberately relegated “Religion” to a tiny outcropping, alongside “Superstition,” “Divination,” and “Black Magic.” The learned article “Cannibals” ended with the mischievous cross-reference: “See Eucharist, Communion, Altar, etc.” Not surprisingly, French royal censors nearly shut the project down altogether. But it arguably did more than any other single work of the Enlightenment to change how educated people in the West understood the world they lived in.

In theory, there is no reason a digital encyclopedia could not have ambitions similar to these. A digital “Propaedia” could of course provide hyperlinks to individual Encyclopedia articles, which would work far more efficiently than printed cross-references. But in practice, to have an encyclopedia even try to provide a systematic overview of knowledge requires a fixed, stable body of articles—a discrete edition. After all, if you have the ambition of linking different articles to each other thematically, then each change you make in one article will require changes to several others. Add a new article, and you have to go back and add many new references elsewhere. It was difficult enough for Diderot and d’Alembert to keep control of this process over the course of producing their one edition, which (thanks in part to the censors) took over twenty years to produce. Diderot lamented on more than one occasion that the project had ruined his life. The Encyclopedia Britannica currently has a staff of roughly a hundred full-time editors, not to mention the 4,400 contributors it uses (thank you again, Wikipedia).

For an online encyclopedia, two of the main selling points are comprehensiveness, and being up-to-date. So even with an enterprise that aspires to scholarly standards of accuracy, the size will eventually dwarf that of even the largest paper encyclopedias, while requiring a huge editorial staff to do the (literally) endless revisions. Can one imagine the editors also trying constantly to revise a “map of knowledge,” and editing dozens of related articles and hyperlinks each time they make a single substantive change? It is hard to imagine any such enterprise making enough money to pay the salaries of the army of editors this would all require. The online Britannica, tellingly enough, has no “Propaedia.” On Wikipedia, contributors do constantly try to update many different related articles to take account of new material they introduce. But Wikipedia, of course, has no plan, no system, no map of human knowledge.

It might be argued that mapping out human knowledge has always, necessarily, been a quixotic project, akin to Casaubon’s “Key to All Mythologies.” It is very likely that few readers ever actually delved very deeply into the “Propaedia,” or made much use of Diderot and d’Alembert’s “map” to navigate the wilds of the 32 individual volumes. The vast majority of people who actually consulted these encyclopedias most likely turned straight to the alphabetical articles, to hunt down specific pieces of information. Today, we use the online resources in the same way.     

But the ambition mattered. It mattered that one could look at a stack of volumes and say: Here are vast libraries, distilled down into an essence of human knowledge, and organized in a logical order. The books testified to the hope that, ultimately, human beings had at least a measure of control over the overpowering torrents of facts and ideas that they collectively produce. Perhaps no single human being could truly have control—what more quixotic enterprise is there than reading through an encyclopedia from cover to cover? But at least the existence of the books gave us the sense that some points of dry land remained amidst the floods, some fragments shored against our ruins. The disappearance of these grand printed volumes, while inevitable, is yet another depressing sign of just how much we are now adrift in the limitless oceans of information.  

David A. Bell is Sidney and Ruth Lapidus Professor of History at Princeton, and a contributing editor for The New Republic.