“Everything is Miscellaneous”

In previous posts I have talked about the challenges of organizing library materials so they can be found by someone who knows just what they are looking for AND support serendipitous discovery by the "browser". I am partway through Everything is Miscellaneous: The Power of the New Digital Disorder by David Weinberger – a much more thorough and articulate exploration of this dilemma than I could ever hope to write. I just finished a chapter on "The Geography of Knowledge" which talks about the origins and limitations of the Dewey Decimal System and contrasts Dewey to the systems used by Amazon.com.

I had forgotten that the whole reason Melville Dewey created his system was to enable libraries to open their collections to "civilians" who could locate items themselves rather than relying on librarians to fetch them from closed book stacks. His intentions were great, but as librarians well know and Mr. Weinberger explains so cogently, Dewey's system has fundamental shortcomings, e.g.

• Although it has been tweaked over the years, at its heart it reflects the worldview of the time and place of its birth – Amherst College in the mid-late 19^th century (to quote Weinberger "an orthodox Christian school that built good Protestant character by teaching the classic"). For example, eight of the nine divisions under the classification for Religion are explicitly for Christian books, the entire "130s" are devoted to Paranormal Phenomena making phrenology equal to all of oriental philosophy in the hierarchy, and there is an entire Dewey number (376) devoted to the Education of Women as if this is a special case.
• The decimal-based organization is in itself limiting as it presumes information can always be grouped into 10s – no more and no less. What do you do if a new topic emerges and all 10 available top level categories are already occupied? In the 1980s that was the challenge that the emergence of the field of "computer science" posed. The logical place to group items related to this topic would have been the 600s – "Technology and Applied Sciences" – but these were full. There were some unused numbers in the 000s, so computer science got 004, 005, and 006.

Weinberger's description of the impact if you do decide to fix a classification problem by reclassifying a subject – say by giving Buddhism its own number series and demoting phrenology – is all too familiar to the PVLD staff who have been working to fix these type of anomalies in our own system…"Tens of thousands of librarians around the world pick up their razor blades and scrape the white numbers of the spines of millions of books….books are piled up, moved from this shelf to that….and at the end of the months or years of work the complaints begin….".

As Weinberg points out "The Dewey Decimal system can't be fixed because knowledge itself is unfixed. Knowledge is diverse, changing imbued with the cultural values of the moment. The world is too diverse for any single classification system to work for everyone in every culture at every time."

He contrasts the Dewey Decimal system with Amazon.com, in whose interest is to introduce you to books you didn't even know you wanted, and who sells and therefore has to organize hundreds of thousands of unique titles each month. A rigid classification system like Dewey just won't work. Instead Amazon tries to create "planned serendipity" using tools such as

• Collaborative filtering which at its simplest level assumes that if a group of people who buy Book A also buy Book B, others who buy Book A might also be interested in Book B…and tells them so. That's why when you buy "Little House on the Prairie" you are told "People who bought this book also bought Little House in the Big Woods."
  
• Data mining that uses algorithms to search the full text of a book to identify unusual phrases that it calculates are important to the book, and then finds other books that also use them…and suggests these other books to the shopper. Weinberger notes "These statistically constructed paths through the geography of knowledge can take you to some unexpected territory: from "Laura Wilder" to the Complete Baseball Record Book, 2004 Edition in just two clicks."
• Listmania, which allows shoppers to created their own lists of books that they feel are related in some way – for example books by and about Laura Ingalls Wilder. Amazon takes this a step further by then using data mining techniques and algorithms to create a sidebar of other lists that are also related sin some way such as list of TV Westerns that includes the Little House on the Prairie series.
• User reviews which allow customers to share their likes and dislikes and also provide a source of data to be mined for statistical patterns.

I thought Weinberger's description of the differences between Dewey and Amazon really summed up the fundamental differences between what he calls a 'second order" system designed to organize physical items in a specific physical space, and a "third order" system which is able to group seemingly random and disconnected bits of organization to reflect precise individual interests and needs –

"Dewey created a single way to cluster books; Amazon finds as many ways as it can. Melvil Dewey took the design of the system upon himself; Amazon lets anyone create her own category, give it a fun name, and publish it. Dewey prized neatness and order…Amazon likes a friendly disorder, stuffing its pages with alternative ways of browsing and offbeat offers peculiar to each person's behavior…Dewey's system prizes the stability that comes with the physical world – books on bookshelves, white ink on spines; Amazon prides itself on its ability to cluster and recluster instantly."

For librarians the dilemma is maintaining a physical collection of materials in some sort of order that enables people to find what they are looking for – a need that for better or worse most public and school libraries use Dewey to fulfill – and at the same time developing ways to use "third order" tools to facilitate individualized and serendipitous journeys of discovery such as those found every day on the Internet. This dilemma challenges some of our deepest held professional beliefs (for example about the value of order itself, or the role of the "amateur" vs. the professional "expert", or how to uphold our principle of that people are entitled to maintain the confidentiality of their reading choices in a world where transparency regarding what used to be viewed as private is the norm) and has profound implications for everything from how we choose to organize our physical materials within our buildings (should the "Groundforce" DVD be in a DVD section, or with the gardening books?) , to how we view and deliver reference services (should we focus our professional energies on designing systems and processes that make it easy for people to find things by themselves, or should we expect them to come to the reference desk so we can get "the answer" for them?) to the fundamental design of our computerized catalogs (can we/should we be using the same kinds of collaborative filtering, data mining, and user generated content that Amazon does?). Tough questions, and not always comfortable ones.

I'm only a couple of chapters into the book – so I'm sure it will provoke more posts!