Tagging reflections

I feel that the tagging exercise gave me a new perspective on the merits of controlled vs. uncontrolled vocabularies. I decided to assign all of my tags before looking at what AAT and TGM had to offer, trying to select words that were relevant to the subject of the photograph – words that I would use when trying to locate not only these specific photographs, but when trying to find a larger group of images into which these photographs would fall. Browsing the Flickr Commons collections, I had searched for a number of place names, so these figured heavily into the tags I assigned. I was surprised, therefore, to find that AAT and TGM do not include geographic terms in their controlled vocabularies. I feel that many users of archival photo collections want to locate photos of a specific area, and the lack of those terms in controlled vocabularies would be frustrating to them. I found the controlled vocabularies more useful when it came to concepts, rather than proper names and locations – I was unsure of how best to represent the ride seen in the Delta County Fair photo or the charity in Distributing Surplus Commodities, and was pleased that AAT and TGM had the concrete term “amusement rides,” and that AAT’s “charities (nonprofit organizations)” and TGM’s “charitable organizations” were more specific than what I had come up with.

Because of this experience, I believe that controlled vocabularies and user-submitted tags should work hand-in-hand, rather than selecting one option at the expense of the other. I felt that the suggestion in the Library of Congress report of incorporating popular concepts into controlled vocabularies was an excellent one. If our main goal is – and it ought to be – facilitating user access to photographic collections, we should be describing our holdings with terms that users will be searching with. The fact that users of the Flickr Commons collections have added useful information to a staggering number of photos shows that user input is beneficial for collection description and access, even if that input is not strictly controlled. I think that the Commons option could benefit collections of all sizes, and might be especially helpful for small repositories both to gain visibility for their collections and receive user input for collections that a smaller staff may not have time to research and describe. Although user tags are not always accurate, and may be redundant or misspelled, I think that they ultimately aid access much more than they impede it. The example from the Library of Congress report, of 73 users tagging a photo “Rosie the Riveter” when the LC headings were  “women—employment” and “World War, 1939-1945,” makes this particularly clear. While a serious historian may use such controlled terms, the more casual user of a photo collection would certainly search “Rosie the Riveter” before either of those options and be disappointed when LCSH returned no results. On the other hand, controlled terms for more abstract concepts, or ones that could be represented by a wide variety of terms, make much more sense than allowing a photo to be tagged with multiple redundant terms. Creating a generalized guide to tagging for Flickr users (on topics such as variant terms and plural forms) could help users (especially the “power taggers” described in the LOC report) assign the terms that will be the most helpful for future searchers. If we are to truly make user access to our photographic holdings a top priority, we should pay attention to how users search and modify our terms to fit their needs.