Skip to content

Ebony and Ivory: Tagging and Taxonomies

I still have 21 single-spaced notes from IA Summit 2008. Two weeks ago I dreamed I was flying a small airplane low and slow across a gorgeous and ever-changing landscape new to me; beneath me spread cities, farms, rippling wheat fields, and rivers twisting and wrinkling in the distance. I think I’m in that airplane now, and the casualty may be some of my pre-flight blog posts don’t get written.

However, I wanted to return to tagging just long enough to talk about smarter tagging and also the very rich value of systems that combine taxonomies (such as LCSH) with folksonomies (tagging systems).

A number of us giggled helplessly through Gene Smith’s terrific presentation on tagging at IA Summit. The speaker kindly ignored us, though he may have wondered what the joke was. Here were some of the suggestions for improving tagging:

  • More structure and control
  • The ability to subdivide tags
  • Providing tag definitions
  • Offering tag suggestions (such as the way delicious prompts you with tags)
  • Allowing true phrases (such as the ability to write creative nonfiction as a phrase, which I can do in LibraryThing, rather than cram it together as creativenonfiction, as I must in delicious)

If you’re a librarian and you’re reading this, you’re grinning. Isn’t it amazing how if you give people enough tools and time, they eventually reinvent cataloging?

However, it’s also telling when people reinvent something that already exists. While traditional cataloging (let’s call this taxonomy work) has all of these characteristics and then some, it’s also slow, in part because it puts the emphasis on “structure and control.” Taxonomies are “expert” systems limited in application to a handful of skilled practitioners. Taxonomies are the long, slow, deliberative output of people thinking at high levels about ontologies (or they should be, but that’s another post).

Slow isn’t bad, as long as it’s not the only descriptive method available. However, if you’ve ever attempted to search a library catalog for an emerging topic, you know slow has its limits. Imagine waiting around for catalogers to decide what tags we would be “allowed” to use in delicious, LibraryThing, or Flickr.

This brings us to pace layering, discussions in the wild for which I can easily trace back to IA Summit 2003. The concept also gets an excellent workout in Peter Morville’s Ambient Findability.

The information-architect folk are very familiar with “pace layering”; it’s newer or unknown to many librarians. In a nutshell, as Morville describes it (a theory adapted from Stewart Brand’s books), “buildings, and society as a whole, are constructs of several layers.. each with a unique and suitable rate of change. … The slow layers provide stability. The fast layers drive innovation.” Together, tagging and taxonomies form a healthy complement, building a “hybrid metadata ecology.”

It really is all good. However, we as librarians should carefully avoid remaking tagging in cataloging’s image — slowing it down to the pace of taxonomies by making tagging too rigid, vetted, and structured. I’m in favor of all those capabilities described at IA Summit, but only insofar as we do not (ab)use them in a manner that trips up the fast front end of folksonomy-building.

I also disagree with (or simply want to broaden) some of the conclusions about motivations for tagging presented in the comments on my earlier post about tagging.

Librarians (and other taxonomists) tagging content for others to use occupy a fascinating and important middle ground — striding quickly, perhaps, between the jog of the folksonomists and the wedding-march of the taxonomists. (Think of three people-movers; librarians tagging for others are on the middle stretch.) We can be proud of the work we do in this space, and it really deserves closer attention.

I repeatedly use the Assumption College delicious set in my presentations because it is a superb example of skilled taxonomists leveraging the tagging wilderness. Sometimes they re-use taxonomy terms and sometimes they do not; you can also see taxonomists thinking around some of the limitations of delicious, such as the inability to recognize strings of words, or exploiting its strengths, such as the ability to group terms into sets. It’s this kind of innovative thinking — applying a taxonomist’s knowledge to a tagging framework — and this kind of behavior — tagging for your users — that tells me librarianship has a real future.

I could write more — perhaps I will — but it’s time to get back into my airplane and resume cruising speed!

Posted on this day, other years: