Frederick Andrew Lerner, D.L.S.
Library and Information Science

Home Books Essays Curriculum Vitae

The Dewey Duodecimal System

by Fred Lerner

In the early 1870s a young librarian at Amherst College devised a new way of arranging the books in his charge. Instead of assigning each book to a fixed location on the shelves, the place for each book would be established in relation to the other books in the collection. That way, when expansion or relocation of the library required the books to be moved, it would not be necessary to change the labels on the books and the entries in the catalog to reflect their new locations.

Perhaps it was Melvil Dewey’s enthusiasm for simplified spelling that made him wary of relying upon an alphabetic arrangement by author or subject; or perhaps the inefficiencies of alphabetical arrangement helped to kindle that enthusiasm. Another Dewey enthusiasm, for the adoption of the metric system of measurement, suggests a strong appreciation of numerical systems employing the power of tens. Melvil Dewey was a rationalist living and working in a rational age. He came of age in a time and place where access to learning was increasingly seen as every person’s right, and provision of that access one of society’s most important duties.

So when he rearranged the books in the Amherst College Library he had more on his mind than simply making it easier for students and faculty to find the books they wanted. He meant to provide America’s libraries with a way to bring their patrons and their books together, by making it easier to use the library’s printed catalog to select books to request for delivery from the stacks. In the increasing number of libraries offering direct access to the bookshelves, his plan would help readers find the books for themselves.

His Decimal Classification employed the same idea that underlay the metric system: a structure of relationships governed by tens. The subject-matter of all recorded knowledge and imagination would be represented by ten broad categories, each of which would be subdivided into ten subcategories, and so on down the line. His original scheme embraced one thousand subjects, each represented by its own number; and by the use of decimal fractions each of these could be subdivided ad infinitum.

Dewey’s ten broad classes were:

000    General Works
100    Philosophy and Psychology
200    Religion
300    Social Sciences
400    Languages
500    Science and Mathematics
600    Technology and Applied Science
700    Arts and Recreation
800    Literature
900    History, Geography, Biography

The Dewey Decimal Classification has its faults. It was devised by an American Protestant steeped in the western literary, philosophical, and religious tradition, and this is reflected in the disproportionate provision in the Religion class for Christianity (which receives fifty numbers, as opposed to Judaism, Islam, Hinduism, and Buddhism, which receive only one apiece) and in the Literature class for American and western European literature. There is also the problem that knowledge is not always readily divisible into tenths; and even when it seems to be, future discoveries and developments might call existing categories into question. (The extensive literature of computer science was not anticipated by any of the widely used library classification schemes.)

For those of us who were introduced to the Dewey Decimal Classification by our school librarians, its division of all literature into ten classes seems perfectly natural. While there might be a topic whose placement in the scheme is not perfectly obvious, the question is usually one of where to put it rather than a frustration at being totally unable to find a place for it. (Perhaps the literature of computer science should be placed in the 500s, as a pure science comparable to mathematics; or in the 600s, as a species of engineering; or even in the 000s, somewhere in the neighborhood of library science. We can argue over which of these is the best place, but we can’t argue that the DDC provides no possible place for it.)


A decimal classification, like a decimal system of numeration, is a natural invention in a society of the ten-fingered. Plenty of people have debated the advantages of numerations based on other integers: for centuries it has been understood that base twelve better accommodates division into smaller quantities, and during our lifetimes base two has become the foundation for computing and communication. Lately I’ve got to wondering what sort of libraries we might have if Melvil Dewey had had twelve fingers. What might a Dewey Duodecimal system look like?

Leave aside the fact that it would offer nearly twice as many classes as the Decimal Classification — and that’s without employing the decimal point. The distinction between 1728 and 1000, though numerically greater, is fundamentally insignificant when compared to that between ten and twelve. How easy would it be to identify two additional classes to add to Dewey’s original ten?

Let’s consider the DDC’s competitors. The Universal Decimal Classification is based on Dewey’s, and as it name implies shares its reliance on tens. Ranganathan’s Colon Classification, while ostensibly rejecting the basic premises of an enumerative system, employs in its schedules a group of categories that make little sense to me. It has 42 main classes, including “Journalism”, “Pharmacognosy”, and “Spiritual Experience and Mysticism”. It makes a distinction between “Mathematical Sciences” and “Mathematics”, but (so far as I can tell) nowhere explains what distinguishes them. And it constructs its class numbers by combining Roman letters (both upper- and lower-case), Greek letters, punctuation marks, and Indo-Arabic numerals in a complex scheme based on manifestations of five Fundamental Categories; Time, Space, Energy, Matter, and Personality.

(The more I learn about it, the more alien the Colon Classification seems to me. I can’t think of any other document that I’ve ever seen that hints more broadly of extraterrestrial influence.)

The Library of Congress Classification was devised to serve the needs of a single, very large, idiosyncratic institution. Its structure derives, not from any theoretical approach to the classification of knowledge, but from the extent and composition of its namesake’s collection. While many other libraries have chosen to adopt it, and in some cases to expand and extend it, the development of the LC classification has generally been confined to serving the needs of its parent library. Thus its list of major classes does not reflect Dewey’s aspirations to universalism:

A         General Works
B         Philosophy. Psychology. Religion
C-F     History
G         Geography. Anthropology. Recreation
H         Social Sciences
J           Political Science
K         Law
L         Education
M         Music And Books On Music
N         Fine Arts
P         Language And Literature
Q         Science
R         Medicine
S         Agriculture
T         Technology
U         Military Science
V         Naval Science
Z         Bibliography and Library Science.
So where would a Dewey Duodecimal Classification find two new major classes? I’ve got no particular quarrel with the ten that we inherited from Melvil Dewey, so I don’t want to replace any of them. (I suppose that in a pinch we could make do without the 400s, and amalgamate Language with Literature as the Library of Congress does. But while the Library of Sauron might want to classify its books into nine categories, we’re under no such constraint!)

I suppose I’m displaying bias toward my own Mystery when I propose to remove Information Science from the 000s and give it a broad class of its own. I’d include communication and librarianship and a host of related topics — after all, we’d have 144 numbers (in our base-ten notation) to work with.

And I’ll atone for my favoritism by suggesting for the other new class an area of learning that I have little use for: the Spiritual Experience and Mysticism that Ranganathan places on an equal footing with the real sciences of Physics and Geology. By isolating this nonsense in its own class we can avoid contaminating legitimate realms of knowledge — though perhaps there is some justice in leaving it with Psychology as most Western classifications do.

I suppose that to librarians with too much time on their hands the design of a classification scheme based on division into twelves is the bibliographic equivalent of Fantasy Baseball or Rotisserie Football. But I have suggested (in a story called “Rosetta Stone”) that we might someday discover an alien race of which we know almost nothing, save how it arranges its libraries. Perhaps by thinking about such matters we might better prepare ourselves for such an encounter — especially if those book-loving aliens turn out to have twelve fingers.


This essay originally appeared in Lofgeornost #93, November 2008.