Folksonomy Folktales 2010

An added benefit that we talked about earlier, is if you take into account the propensity of experts to tag at lower or more specific levels of tags, then experts can create a self-selected community automatically.  People who tag something with “philosophy” are ignored and other experts who tag with things like “categorical imperative” know automatically that others who tag with this level of specificity are likely experts too.

Within the limits discussed above, folksonomies can function as a mechanism for discovering and creating communities – the heart of social networking.  And, while I still would not use the word revolutionary, clearly social networking is very important and exciting and is changing the dynamics in a number of areas from marketing to politics and this is a real benefit.

Knowledge Architecture for Enterprise 2.0

The real value of folksonomies or at least user-generated tags can best be realized, however, by locating them in a more sophisticated and developed infrastructure.  This approach is one that I have seen or been part of within the enterprise and it follows a very common pattern which is that new ideas are often developed on the Internet (and often overhyped) but then see their more complete and valuable development within the enterprise.  The key idea behind this approach is to develop a framework and a facilitation capability within which user generated tags and folksonomies can be supported and utilized in ways that go way beyond tag clouds for serendipitous browsing for potentially interesting communities.  This approach is what I call knowledge architecture and it is an approach that has a great deal of potential for enhancing KM in the enterprise in general and for enterprise 2.0 initiatives in particular.

A full description and discussion of knowledge architecture is well beyond the scope of this article, but some of the main elements as they relate to folksonomies and enterprise 2.0 include:

• Knowledge Organization elements:
• Taxonomies, metadata of all sorts, but particularly faceted metadata and best bets
• Categorization rules and name catalogs or entity dictionaries
• Maps of formal and informal communities within the organization and their associated knowledge and information behaviors

Software:

• Text analytics, enterprise content management, enterprise search, KM platforms

People Resources:

• A central KA group including librarians and knowledge architects – who develop, refine, and maintain the infrastructure elements and facilitate their application

Just to get a flavor and inkling of what a knowledge architecture approach to folksonomies and enterprise 2.0 might involve, let’s look three areas:  Development of knowledge organization and community maps, Application in ECM and Search, and Refinement of knowledge organization.

Research Resources:

First, in the area of development of such knowledge architecture elements as taxonomies and community maps, folksonomies can be used as resources for social and cognitive research into the interests and language of users and communities.  We already do this to some extent with search log analysis but user generated tags are a higher order cognitive task and represent a level of categorization higher than simple search terms.  This makes them a much more powerful resource, especially if they are used in conjunction with search log analysis.

Another area where user generated tags can be a great source is for developing faceted navigation applications.  Research on popular book marking sites reveals that the vast majority of tags fit very well with a faceted classification system.  For example, the article by Louise Spiteri talks about the overwhelming use of thing tags.  My own analysis of  Flickr, shows that close to 80% of the tags fall into one of a few facets – thing, place, event, and people.  So if you are in the process of developing your facets, user generated tags are great sources for specific values and an analysis of a folksonomy created by your users is a great source for designing which facets are particularly important to your users.

Search and ECM Applications:

Combining folksonomies with other elements is also a way to get more direct benefits than using them for research.  The way to do that is to develop a hybrid approach that combines folksonomies with simple taxonomies.  The taxonomy need not be the Library of Congress classification – it could be a 2, 3 or 4 level taxonomy at most.

The basic idea is that the simple taxonomy guides and offers suggestions to the user who can just take the taxonomy term, take the term and add another tag for more specific level tagging or just add their own tag independent of the taxonomy.

This type of hybrid normally has an editor function so these user-generated tags can be analyzed and either treated as suggested new terms for the taxonomy or simply sub-categories.  A hybrid approach, while not satisfying the revolutionary ideologues, nevertheless, offers real value and is one way to extend the application areas of folksonomies beyond sites like Delicious.

Another possibility within the enterprise is to develop a 3 or 4 level taxonomy that is supported by content management software that can have categorization rules associated with each taxonomy node so it can offer full suggestions based on its auto-categorization functionality.  This makes the taxonomy piece even more powerful but can also incorporate user suggestions and additions.

This hybrid approach is much more powerful for finding documents and internal web sites, but can it also support social networking?  If there is a software front end that functions like a tag cloud with tags ranked by popularity but also referencing the taxonomy, it should enable social networking that is at least as good as a straight folksonomy based site.  I’d like to see a lot more research on this, however, on both enterprise and internet sites.

Finally, another area that a knowledge architecture approach works well is simply creating the infrastructure to support a range of KM or Enterprise 2.0 initiatives.  Despite some over simplified and over hyped rhetoric, 2.0 initiatives don’t just happen.  It is not enough to buy some blogging and wiki software and make it available to your employees and sit back and watch them transform your organization.  We’ve already seen how taxonomies and folksonomies can support and enhance each other, and when it comes to things like wiki’s there is a similar need for knowledge organization and editorial roles to get their full benefit. 

For example, all but the most simple wiki applications need an initial intellectual framework or categorization schema to avoid chaotic structures that limit the usefulness of a wiki.  They also need  ongoing editorial functions to monitor the wiki, resolve disputes, and ensure adherence to a level of content quality.  For even the granddaddy of wiki and 2.0 applications, Wikipedia utilizes thousands of editors for these basic editorial functions.

A well-developed knowledge infrastructure and central KA group is also the way to get the maximum benefit from blogs and other KM 2.0 initiatives.  This can range from utilizing a taxonomy-folksonomy hybrid as part of the front end to a constantly expanding internal and external blogosphere to using the KA group to facilitate the use of taxonomies in publishing to using the knowledge organization to integrate external blog content into your internally generated content.

This just gives a high level view of how knowledge architecture can be used to expand the application and usefulness of folksonomies.  Of course, while I prefer the approach and term, knowledge architecture, there are a number of terms and approaches to develop a taxonomy-folksonomy hybrid. For more, see my upcoming book on Knowledge Architecture.

Deeper Themes – why the fuss?

The last question I want to look at is – why the fuss?  What led to this over-hyping of folksonomies?  A full cultural-psychological analysis is way beyond the scope of this article which has already gotten too long, but let me just suggest a couple of themes that it seems to me had something to do with it.

First, there is simply the fact that social networking is really very cool and important and is having a lot of impact in a lot of areas.  So it is very easy to go from “It is very cool to be able to connect to other people all over the world” to “The mechanism that enables that connection must be very powerful too.”.  IOW, if social networking is revolutionary then folksonomies must be revolutionary too.  What quickly follow from this is article after article elevating (and overstating) the power of folksonomies.

A second factor that is related to the first is simply an IT/technology focus on tools.  Let’s face it, we’ve been using the Internet and other means for social networking for some time.  What is different is that the current set of tools, particularly blogs, wikis, and social tagging sites make that social networking easier, faster, and available to more people.  All of which is great and opening up social networking and tagging to larger masses of people does have an impact (both good and bad as scale effects become stronger), but it’s only a revolution rather than just the next evolutionary step if you think that tools are the really fundamental component.

A third factor can be seen when we look at how folksonomies are being elevated.  We’ve already examined the specific arguments, but here I want to look at the underlying philosophy/culture behind those arguments.  People in IT and other technical areas have very different knowledge cultures than people in so called soft sciences like the humanities or “management science”.  I am currently working on a book on knowledge architecture in which I go into these different knowledge cultures in more detail, but in general, there are different types of concepts dominant in IT/technical cultures and more importantly for this discussion, different criteria of “truth”.  To oversimplify, in IT/technical cultures, something is true if it works – the algorithm produces a correct answer or solves a problem.  Who creates that algorithm is irrelevant to the correctness of the answer – it can be a 13 year old kid playing in their room or a 20 year veteran with myriad honors and a huge reputation.  In the humanities or management, it matters much more who says something because there is no universally accepted way of objectively evaluating the correctness of most statements. 

One outcome of these differences is that there is a great deal of mistrust and antagonism toward the importance of reputation and experience in the soft sciences by IT/technical people which often comes out as a strong anti-authoritarian, anti-expert attitude.  It is actually not so much anti-expert as anti-non-technical expert.  In the area of folksonomies that is often accompanied by a suspicion of culture concepts like semantics.  To see how this plays out, let’s take a look at the Ontology is Overrated – Categories, Links, and Tags article by Clay Shirky in which he discusses the drawbacks of the early Yahoo browse the Internet categorization schema. 

“But Yahoo decided to privilege one way of organizing links over all others, because they wanted to make assertions about what is “real”.” 

I know some of the people involved in the early Yahoo categorization schema and I can tell you this is way off the mark.  Yahoo cataloguers weren’t trying to model reality or impose some Nazi view of reality on poor defenseless users.  They were trying to help users find content better – and make some money while doing it.  The point was that no one could really find anything on the Internet – it was one of the first inflationary periods and Yahoo gave them a way to find stuff faster and easier – not a big conspiracy theory attempt to impose rigid thought control on the world of free taggers.  Who of course, at the time didn’t tag at all. 

And another quote:

“Yahoo is saying, ‘We understand better than you how the world is organized, because we are trained professionals. …but the effect was to override the users’ sense of where things ought to be, and to insist on the Yahoo view instead.”

Here the anti-authoritarianism comes through loud and clear – or rather certain types of experts.  But again, this is kind of silly.  Let’s be  realistic – it’s not as if you had all these alternative ways floating around the Internet of developing categorization schemas to organize all content on the Internet.  I know that I didn’t meet more than 2 or 3 people whose stated goals were categorizing all Internet content in their spare time.

And as in the notion that they were attempting to model reality, the idea that Yahoo cataloguers were actually claiming that “they understand better” is way off the mark.  First, of all, we need to distinguish between claims.  I would imagine that most people who have had training in categorization would never make such a sweeping claim,  What they would claim, with a great deal of authority, is that they know more about categorization and developing general purpose categorization schemas than someone with no training or experience.  However, good cataloguers would never make the second part of the claim – that they better understand how the world in organized – because they have enough experience to understand that there is no single correct way to organize the world.  The best you can come up with is something that has certain good formal properties and reflects a pragmatic compromise that according to their user testing, works pretty well.