Weinberger, a fellow at the Berkman Center (along with Clippinger) and one of the authors of the prescient The Cluetrain Manifesto: The End of Business as Usual (Perseus, 2000; this was one of the first Internet boom books to describe markets as conversations and to sensitize business to the economics of online interaction), is the kind of writer who can turn a knowledge revolution into a simple and compelling story.
The problem with physical things is that they can be in only one place at a time — the “first order of order,” in Weinberger’s words. A historical romance novel about Napoleon won’t be found everywhere it could logically belong in the library, on history, romance, and biography shelves, because it can sit on only one shelf at a time unless there are multiple copies. Card catalogs are less limited — the “second order of order” — because paper-based record keeping enables more sophisticated inventory of physical objects, including schemes for tracking physical knowledge containers (e.g., books).
But second-order catalog schemes like the Dewey decimal system that help librarians know where to shelve books impose a rigid set of rules for how to categorize items — a taxonomy imposed on everyone else by an elite. Eight of Dewey’s nine major divisions classifying religion are for Christian books; Judaism has an entire Dewey number, but Islam shares one with Babism and Baha’i. Buddhists are a subcategory of “Religions of Indic Origin.” New religions or other phenomena, or old ones whose importance is being reconsidered, can find a home in the Dewey system only “to the right of the decimal point,” as fractional subcategories of existing categories. In philosophy, Weinberger points out that “Dewey’s system puts phrenology on a par with Aristotle.” Although new learning can create new knowledge, the new knowledge is constrained when it has to be categorized according to systems that were created in an era of old knowledge — before, for example, it was known that Aristotle would remain important but phrenology would come to be considered a pseudoscience. When a new realm of knowledge comes along, or previous maps of knowledge are seen as parochial, it isn’t possible to redesign systems like Dewey’s.
Nowadays, however, people are using “tags” — the code that categorizes digital content — to grow, rather than design, classification systems. Not only do such systems fluidly reconfigure knowledge to reflect new knowledge, but their shapes reveal clues to further knowledge, the way “tag clouds” — visual depiction of content tags used on a Web site, where the most frequently used tags appear in boldface or larger font — can quickly yield information about the key subcategories of any subject.
In 2003, Joshua Schachter, a programmer working in financial analysis, created a tool for storing his collection of 20,000 links to Web sites. Like a programmer appending explanatory comments to code, Schachter “tagged” each bookmark with a set of words that he could use later to search for the bookmark. For example, a Web site about camembert might be tagged with the words cheese, French, and cooking. Instead of using predefined categories or keywords rigidly assigned by a committee of experts, Schachter simply thought of the labels he might use in the future to find the item.
Schachter’s idea really took off when he put his tool online and not only let others bookmark and tag, but made it easy for them to share their bookmarks and tags. Instead of a top-down taxonomy, Schachter’s breakthrough gave rise to what information architect Thomas Vander Wal called a “folksonomy,” a growing bottom-up classification architecture generated by a general population of users. Flickr’s folksonomy enabled its users to tag the photos they uploaded: “Aunt Sally,” “Florida,” “beach,” “2007.” When millions of people began tagging, the aggregation of their individual decisions turned out to have collective value to people looking for pictures of Florida beaches or pictures from 2007, similar to the way Google’s PageRank algorithm improves searching by aggregating the individual decisions of people who link to Web sites.