Where's the fight-back from formal classificationists?

In the last two-three years a huge amount has been written about tagging and folksonomies, much of it with the bright-eyed enthusiasm of those who haven't seen the present state of affairs in a broader light; but where is the fight-back from the formal classificationists, who hither-to ruled unchallenged in this area? Have such giants as the Library of Congress and the Dewey Decimal System fallen at the first hurdle?

Tagging is the assigning of arbitrary tags to content by amateurs (typically content creators, editors or readers) and folksonomies are systems built from the ground up using these tags. Tags have no formal meanings and there are no constraints placed upon them. Folksonomies are central to systems such as flickr, del.icio.us and the whole web 2.0 approach.

Formal classifications, such as the the Library of Congress and the Dewey Decimal System are rigorous systems in which trained individuals assign subject categories to content. Each category has a description and is long lived—categories don't change even when the words used to describe the topic in popular culture change. Thus the LoC still calls cars automobiles, because that's what they were called when they first entered the system.

So what's wrong with folksonomies taking over from classification? Well, a great deal. The two approaches generate very different kinds of information spaces which are suitable for different things, and there are things which each provides which the other cannot. Folksonomies are great for dealing with peer-created ephemeral content, creators or consumers can very quickly tag content in ways which is meaningful in the community in which it is created and shared, other consumers in that community (who share the community values and tag meanings) have no problem interpreting the meaning of the tags. Folksonomies fall down, however, when tags are either very loaded, mean widely different things to different communities or to the same community in different contexts.

Examples of situations in which classifications succeed where folksonomies appear to repeatedly fail include:

The web will be a poorer place if we allow folksonomies to replace classifications and these failings become systemic, but as far as I can see, there has been no significant fight-back or rallying of the classification supporters in the face of the increasing dominance of folksonomies. Maybe this is because the big classification systems are managed and run by bodies representing traditional libraries, which have traditionally had low engagement with technology and are thus a step behind.

There have been a number of attempts to "bridge" or "map" between folksonomies and classifications, based partly upon previous work mapping between classifications, and this work is particularly hot right now in the semantic web area. But the problem is that the result substantially dilutes the benefits of both folksonomies and classifications: while it does effectively link communities of interest across boundaries, it erodes the differentiation between subject and object, and substantially removes clarity of naming. So rather than getting the best of both words you inevitably get a blend of the better and worse aspects.

The other problem with bridging or mapping between folksonomies and classifications is that the effectiveness of the mapping depends partly on the size, overlap and currency of both the folksonomy and classification. If classifications fall from favour, they will gradually become less and less useful for bridging, as they become relatively smaller, more removed and less current.