What's so great about the Dewey Decimal System?
Dear Straight Dope:
What's so great about the Dewey Decimal System? I keep hearing librarians rave about it as the greatest invention since cheese-flavored dog food, as if Dewey is some genius on the level of Einstein (or Cecil). It seems obvious that if you have to categorize a huge number of objects (say. books for example), you would set up a set of unique categories, sort the objects into the appropriate category, and then itemize the objects within that category. Is Dewey famous just because it's the standard, or is there more to the system that makes it so great?
SDSTAFF Dex replies:
What's obvious to you, the jaded library patron, wasn't so obvious in the era before Melvil Dewey. Once a creative genius comes up with an innovation, a century later everyone thinks it's obvious. If you think it's so easy, you come up with a system for classifying all knowledge that ever was and ever will be.
What's so great about the DDC? First, as you acknowledge, it's a standard. The DDC is by far the most widely used method of organizing books in the U.S, and indeed in the world; it or its offshoot the Universal Decimal Classification (UDC) is used in over 130 countries. Somewhere around 95% of all school libraries and public libraries in the U.S. use DDC. Its leading competitor is the Library of Congress Classification (LCC) system which, in contrast, tends to be used by government and academic libraries (that is, colleges and universities). Only 25% of colleges and universities use Dewey Decimal Classification, and around 20% of specialized libraries do.
Second, and equally important, DDC continues to serve libraries well after more than a century of use, despite the enormous expansion in knowledge during that time. To understand why this is so requires some understanding of DDC's history, so we'll start with its inventor, Melvil Dewey.
MELVIL DEWEY AND THE DEVELOPMENT OF THE DEWEY DECIMAL SYSTEM
Melvil Dewey, born in 1851 in upstate New York, was an interesting guy. He was named Melville Louis Kossuth Dewey after the Hungarian reformer Lajos Kossuth, a popular lecturer in exile after the failed revolutions of 1848. Uprisings occurred throughout Europe that year (including the famous one in France that was the subject of Victor Hugo's Les Miserables), aimed at overthrowing the monarchies. Americans were enamored of these revolutions, being "predisposed to judge hereditary rulers harshly," as library historian Matthew Battles puts it, so it was not surprising to find parents naming their children after revolutionary heroes.
From childhood, Dewey was fascinated with books. In 1868, when his school caught fire, he rescued as many books as he could from the school library; but inhaled a great deal of smoke in the process and consequently had a cough that lasted for months. Told by his doctor that he would be dead within a year or so, he tried to make the most of what he thought would be limited time, according to the recent and fascinating biography by Wayne Wiegand. Efficiency became his obsession. He urged simplified spelling, shorthand, the metric system, and similar reforms. He later dropped his foreign-inspired middle name and shortened his first name to Melvil to dispense with extraneous letters. (He considered simplifying his last name to Dui but gave up on that.)
At Amherst College Dewey worked his junior year (1872-73) in the Amherst College library, where he was frustrated by the disarray of the book collection. As noted in the Staff Report on public libraries, it was an era when libraries were spreading rapidly, and their collections of books were growing. There was no uniform or consistent system for organizing books, sometimes not even within a single library. Each library simply assigned a spot on a shelf for each book, and recorded in its catalog where the book was.
Some libraries adopted the British Museum's method of numbering the shelves, in which a book would be filed on shelf 132A (or whatever). This "fixed location" approach was useful up to a point — if you were looking for a book, you could look up its shelf location in the catalog. However, deciding what books went on which shelves was pretty arbitrary. In some libraries, shelves were arranged to "look nice" — books of the same size would be put on the shelf together, regardless of subject matter. Seriously.
One big problem with the "fixed location" approach was that it was difficult to integrate new acquisitions. Adding new books to a shelf meant other books would need to be rearranged, perhaps pushed off to different shelves, and that meant changing all the relevant catalog records.
The New York State Library organized its books alphabetically by title, without regard for subject. Imagine the problem every time a book was acquired! Jacob Schwartz of the New York Mercantile Library got around that problem with a card system. Rather than cataloging in a ledger, he assigned one card per title, with the shelf location in the upper left-hand corner. However, Dewey wrote that Schwartz thought location was "not essential to the system" and could be "omitted or used, at the discretion of the librarian."
Many librarians thought it would be useful to organize books by subject matter. The problem was, what does "subject matter" mean? Francis Bacon, in the 1600s, said there were three branches of knowledge: history (deriving from memory), poetry (from imagination), and philosophy (from reason). That formed the basic classification system of the few libraries that bothered with such things. The Vatican library, for example, used only two classifications: sacred and profane.
In 1873 Dewey discussed the matter with Charles A. Cutter, director of the Boston Athenaeum. Cutter had developed a classification scheme that Dewey liked; Dewey wrote in his diary that "book on the hourse [were put] under 'horse' & not under 'zoology.'" [All quotes from Dewey's diary are taken from Wiegand's biography, listed below under Resources.]
A few weeks later, Dewey "blundered on" a pamphlet printed in 1856 by Nathaniel Shurtleff that suggested a decimal arrangement for libraries. Dewey wrote, "My heart is open to anything that's either decimal or about libraries." But he disliked Shurtleff's approach, which he felt was inefficient.
Another influence was William F. Poole of the Cincinnati Library, who dared to suggest that fiction was important and needed to be arranged for use.
Dewey found all these systems had strengths and weaknesses. Many years later he wrote, "one Sunday during a long sermon … while I lookt stedfastly at [the pulpit] without hearing a word, my mind absorbd in the vital problem, the solution flasht over me so that I jumpt in my seat and came very near shouting 'Eureka!'" (Note the simplified spelling.)
Dewey's innovation was to combine a numbering system (like at the British Museum) with classification by topic. However, the numbers didn't indicate a shelf but rather a field of knowledge. Battles says, "Thus he joined the analytical simplicity of decimal numbers to an intuitive scheme of knowledge, one that would fluidly accommodate all the books ever written, and all the books that could be written as well." Thus was born the Dewey Decimal Classification system.
Wiegland comments that Dewey's contribution to classification was joining together the strong points of systems developed by others, not creating anything new. So, to answer your question, Dewey's genius wasn't the creation of something new, but the consolidation of diverse ideas.
By the way, Dewey was interested not just in standardizing book classification, but everything else, down to the size and margins of the catalog cards. He wrote about what we would now call economies of scale applied to "cataloguing, indexing, and the score of things which [can] be done once for all the libraries." The money saved by standardized administration could be used for more acquisitions.
The main innovation and advantage of DDC is that it's an indirect, rather than a direct, reference to a book's location. When you look up a book in the catalog, you're not told which shelf to go to; instead you're told a location relative to other books, and you need some second reference (such as a chart of where numbers are stored, or numbers marked on the sides of shelves) to find the book itself. This separation avoids the problems of the "fixed location": if the library adds books, so that some books are shifted to different shelves, only the chart of location needs to be changed, not the whole card catalog.
One of the beauties of Dewey's system is that it provides an easy way to introduce new subjects--there have been lots of new subjects since 1876! The key was the use of decimals. Dewey began by establishing a broad division of knowledge into basic categories, to which numbers were then assigned--crudely put, these are the numbers to the left of the decimal point. That done, it was easy to add new subjects by dividing the original categories into progressively finer gradations — these are the numbers to the right of the decimal point. DDC is what today we'd call scalable — it has readily accommodated the explosion of knowledge since Dewey's day. By the way, it's not just the development of science but also globalization that has brought enormous additional information. And it was easy for both librarians and readers to learn how to use the system.
Dewey largely created library science and is rightly known as the "Father of Modern Librarianship." Without DDC, each library would have to create its own system at vast expense. Now it's a simple matter. That's the reason Dewey is so beloved by librarians.
We feel obliged to note that Dewey was no saint. He was racist, antisemitic, anti-black, anti- everything not white male Anglo-Saxon Christian. He died in 1931.
HOW DEWEY DECIMAL CLASSIFICATION (DDC) WORKS
Dewey divided knowledge into nine classes plus one "Miscellaneous," each assigned a numerical range:
000 General Works (Miscellaneous)
300 Social Sciences
500 Pure Sciences
600 Technology (Practical Arts) including medicine, engineering, business accounting, agriculture, salesmanship, etc.
700 Fine Arts (including architecture, painting, photography, music, amusements, etc.)
900 History, Geography, Biography
Each item that comes under one of these disciplines is assigned a number in that range, called a class number. The logic is hierarchical: that is, within a main class, there are various subdivisions (called subclasses), and these are subdivided further, getting more specific. For example, within 700 (the arts) we have 790 ("Recreational and Performing Arts"), and within that is 795 for games of chance. After the three leading numbers, decimals can be used for as much further subdivision as needed, so 795.4 is card games, 795.41 is card games "based chiefly on skill," 795.415 is contract bridge, and 795.4152 is the bidding process (auction) in contract bridge.
A hallmark of DDC is the use of consistent subclassifications and mnemonics regardless of category. For example, 73 usually refers to the U.S., on both sides of the decimal point; thus, U.S. cooking is 631.5973 and U.S. history is 973. In contrast, LCC subclassifications have no consistency from one class to the next.
A derivative of the DDC is the Universal Decimal Classification (UDC), an international adaptation of the Dewey system used extensively in Europe. It's more detailed, but the principles are pretty much the same.
There is an abridged version for small local libraries, and a more detailed/complex version if the library grows, or for larger libraries.
LIBRARY OF CONGRESS CLASSIFICATION (LCC)
You didn't ask, but the Straight Dope Science Advisory Board got into a discussion about DDC vs. Library of Congress Classification, so we might as well go whole hog. The Library of Congress was established in 1800 to provide members of Congress with information needed in making legislative decisions. By law, all books registered for copyright in the U.S. must submit a copy to the Library of Congress, which accordingly has become the de facto "national library" and the largest in the country.
The LC originally used a "fixed location" system started by Thomas Jefferson, with 44 subdivisions. However, at roughly the same time as the explosion in libraries (1890s) that led to the Dewey Decimal System, the Librarian of the Library of Congress and his Chief Cataloguer started a new classification system. Basic features were taken from Charles A. Cutter's "Expansive Classification" — you'll recall that Cutter influenced Dewey.
The Library of Congress Classification (LCC) system is based on 21 major classes (as compared to DDC's ten). Each class is arbitrarily assigned one of the letters A through Z, with five exceptions (I, O, W, X, and Y are assigned at second or third level subclasses). Within each major class, the subclasses were independently developed by specialists in each field, so the system was decentralized, with different topics having different arrangements of subclasses.
Beginning in 1902, with Class D, two-letter combinations were used for subclasses. The basics of the system were developed between 1899 and 1940. Here are the broad categories:
A = General Works
B = Philsophy and Religion
C = Auxiliary Sciences of History
D = Universal History
E and F = American History
G = Geography, Anthropology, Recreation
H = Social Science
I = Political Science
K = Law
L = Education
M = Music
N = Fine Arts
P = Language and Literature
Q = Science
R = Medicine
S = Agriculture
T = Technology
U = Military Science
V = Naval Science
Z = Bibliography, Library Science
The first subdivisions within a class are indicated by two-letter combinations — for example, under class G, we have GV for "Recreation and Leisure." Then, decimals are used to refine the discipline, so GV1199–1570 is devoted to "Indoor Games and Amusements," within which the range GV1232–1299 is set aside for card games. In that range, GV1282 is for Contract Bridge, and then GV1282.4 for bidding. LCC classifications and their descriptions fill 40 volumes with over 200,000 headings, so I'm not about to give much more detail here.
Is GV1282.4 easier than Dewey's 795.4152? Well, for the example of contract bridge, there's probably no particular advantage one way or t'other.
The organization of the LC was primarily focused on the needs of Congress, and secondarily towards other government departments, agencies, scholars, etc. So more space is allowed for history (classes C to L) than for science/technology (Q to V). More important, the focus on the needs of Congress means the LCC pays less attention to non-Western literature, and has no classifications for fiction or poetry.
A peculiarity of the LCC is that subdivisions are often by country, rather than by subject, especially in the social sciences. This makes sense in light of the library's role as a research tool for Congress. In contrast, under the Dewey Decimal System, the subject is usually subdivided fully by topic before geographical divisions are added.
LCC is more technically oriented than DDC. For example, photography is under TR (a subclass under technology) in the LCC, rather than in the fine arts as with DDC. Although library patrons probably don't notice, critics say the technical focus often makes LCC cataloging difficult for non-experts in a field. For instance, LCC lists "amicide" for the category of "friendly fire casualties" and "dysmenorrhea" for "menstrual cramps." Battles says that the LC classifications often "strike a tone of bureaucratic high-handedness."
The Library of Congress places much emphasis on cross-referencing. Battles comments that "those nesting, cross-referenced rubrics make up an epistemological labyrinth unto themselves." The classification system emphasizes professional knowledge.
After the main classification, there's a usually a Cutter Number (after a decimal point). These were based on a system developed by Charles Cutter in 1891. Cutter numbers are used primarily to maintain alphabetization as needed, based on personal or corporate names, geography, titles, etc. Cutter Numbers are a letter followed by numbers, ordered like decimals in order, so .M395, .M4, .M47, .M5, etc.
The LCC changes slowly compared to the DCC. And changes are required on almost a continous basis, from geopolitics (Czechoslovakia splits into the Czech Republic and Slovakia) to societal growth (the category "Negroes" is now "Afro-Americans" for U.S. blacks, and "blacks" for non-U.S. blacks.) The change of "Moving Pictures" to "Motion Pictures" didn't occur until 1989.
COMPARISON OF DDC AND LCC
The two systems were developed around the same time, give or take a decade or so. Both were based on the perception of knowledge and the relationships between academic disciplines extant from 1890 to 1910. Both are enumerative systems covering all topics, all disciplines, all fields of knowledge. Both are updated regularly. Both use a "controlled vocabulary," that is, a list of preferred terms for cataloging.
Intner and Weihns say that both systems reflect the bias of a nineteenth-century U.S. outlook, then a "Western" outlook, and reflect a "white, male, Anglo-Saxon Christian view of the universe." These biases are more obvious in the LCC, and have been largely eliminated in the UDC variant of the DDC.
The differences are more striking.
Size: The LCC is significantly larger — that is, LCC has more broad classes (21 vs. 10), with more and narrower subcategories.
Specificity: The LCC has far more specific subclasses and categories that tend to be technical in nature. In any classification system, from libraries to file drawers, the question is: do you have a lot of files with less material in each one, or fewer files with more material in each one? LCC opts for the former approach; DDC for the latter. (Note that there is still order in the DDC within each file, of course.)
Structure: The DDC has overarching principles (for instance, decimal division) and mnemonic notations. For the LCC, each of the 21 classes was developed independently by experts in that field, and continues to be expanded and updated independently. There is no consistency in LCC between the classes.
Notation: The DDC uses only numbers; combined with some mnemonics, it's much easier for librarians, students, volunteers, and so on to remember. The LCC, in contrast, uses both letters and numbers, allowing more classes, more categories, more classifications.
Indexing: The DDC Relative Index can bring all topics together, under one grand umbrella, regardless of what class they're in. This is lacking in the LCC, because of the inconsistency between classes.
In sum, DDC uses fewer categories and sub-classifications and is consistent across disciplines, while LCC is more highly subdivided with no consistency between disciplines. It's understandable, therefore, that DDC has proven more useful to libraries catering to a wide range of needs such as public libraries and schools, while LCC is more widely used in libraries focused more on technical areas like colleges, universities, and government.
If you've come this far, I got nothing to conclude.
By the way, there's also something called the Sears system, founded by Minnie Earl Sears in 1923, aiming at a simplification of the Library of Congress subject headings. And then, of course, there are ISBN numbers. Some other time, eh?
Now available in translation!
How often have you thought: this is a really interesting article, but I'd like it so much more if I could read it in Belorussian? Well, now you can.