In the past couple of days two bloggers writing about history have weighed in on social networking and the creation of online thought communities. Kelly Lewis in Curiouser and Curiouser discusses the ways that tagging of online museum collections may help build audiences for museums–audiences that they are (or may be) losing right now. Jeremy Sandor in Humility in History ran a simple experiment in tagging using del.icio.us, Flickr, and Technorati and concluded that: ” del.icio.us consistently came up with the least relevant results. The websites that were tagged were frequently trivial or completely off the desired topic. Flickr showed the most varied results; however, the tags were frequently appropriate for the pictures displayed. Technorati’s results were often the most informative, although tailoring the search results to show only the most authoritative blogs is what really improved these search results dramatically.”
Both of these posts raise an important issue about tagging, but from different angles–the randomness of the choices that users make when assigning a tag to a digital object. Consider this object from the Hurricane Digital Memory Bank:
How would you tag it? Choices could include: Katrina, debris, cleanup, reconstruction, girlfriend, my_house, or dog.
Dog? Here’s what the person who uploaded the image had to say about it:
“Debris and thick orange construction fencing, was very useful in containing dogs in there new homes or dens which they made under houses. We worked in a team to herd them under a house, and quickly block it off, until the small people with Ketch poles could go under and retrieve the frightened dogs who were once affectionate pets. This photo was take Mid-November 2005 in New Orleans.”
This is the sort of thing that will drive my friends in the archives community crazy, because the meta-data created by users follows no set of agreed upon parameters. And as historians, we’ll have a very hard time sorting through large databases of objects (the HDMB already contains more than 4,000 digital images) to locate the material we want–the problem described by Sandor.
But, and I think this is a very important but, I am just as interested in the cognitive processes by which the users chose their tags–the meaning they assigned to an object as they chose a tag or tags for it. Being able to look at the ways that individual users made sense of information according to their own lights–rather than the predetermined meta-data schemes of the professional archivist–opens up an entirely new avenue of historical investigation–not investigation into the events of the past themselves, but into the ways that people in the present try to make sense of the past. I think that this is the true advantage of tagging for the historian.