In the wake of a great conference in Boston, the listserv [link to sign-up - sorry, no archive!] of the Museum Computer Network (MCN-L) has been teeming with stimulating discussion. One of the most recent threads covered subject cataloguing in museums, as well as the emergence of folksonomies as a potential tool for providing improved searching and retrieval of museum objects. (If you’d like to read the entire thread, leave a comment with your e-mail address!)

Subject cataloguing has been a hot-button issue in museums for some time: since museums have not traditionally published inventories of their collection items for end-users to search, they by and large haven’t been steeped in subject and authority control as much as libraries, which have been in the OPAC business for decades now. As a result, many museums today find themselves in a position of having digitized their visually stunning collections, but grappling with generating the appropriate metadata for sophisticated searching and browsing of the materials. While an ever-increasing number of museum professionals will readily admit that assigning subject terms (and having authority control elsewhere in their records) would be a perfectly fine and sensible thing to do, most find themselves financially daunted by the prospect of having to convert their collections management system to stack up to this lofty ideal. Enter folksonomies. As I think can be readily appreciated from this background, letting other people, namely your users, do the work of creating access points for you must look like a rather enticing option.

The debate on MCN-L teases out some interesting perspectives on the issue – clearly there is an enormous amount of interest in folksonomy experiments such as STEVE (also see this article in D-Lib). Sebastian Chan from the Powerhouse Museum introduced their Electronic Swatchbook project into the discussion, which allows users to describe a fabric swatch using the categories color, pattern or mood. However, by and large the enthusiasm about folksonomies seemed tempered by the realization that they aren’t a magic bullet by any means – Richard Urban, drawing both on his experience from at the Collaborative Digitization Program and his current life as a LIS student at Urbana-Champagne, comments that folksonomies fall short when it comes to aggregating resources (…We’re already seeing the interoperability challenges even when controlled vocabularies are used…). According to Richard, folksonomies won’t help us share our data, cluster data (establish relationships between items), and the tags created will be hard to manage over time (today’s hot tag is tomorrow’s undecipherable gibberish). Jennifer Trant adds an interesting twist to the discussion by asserting that social tagging may start as a personal activity (I flag things so I can find them), and then almost incidentally becomes useful for others (So you’re interested in this as well?). In her own words:

Personally, i’m becoming more and more convinced that tags exist in a space between a user and a resource, and that their meaning is situationally defined. the popular del.icio.us tag ‘toread’ is a good example of this. it only means something to me, now. it doesn’t really mean anything to you, because i can’t know if you want toread this thing. […]This doesn’t discount the social leverage we get from tagging. You might be really interested in the things that i want toread because you’re interested in the same things that i am. So you follow the
things that i want toread assuming that there will be useful stuff there.

She adds that the key to making all of this work is a critical mass of tags, and we just don’t know yet what “critical” really means in this case.

However, the main point of debate remains the merit of a cataloguing approach (or what Clay Shirky would somewhat disdainfully call the ontology approach) versus a tagging approach (folksonomies). William Real from the Carnegie Museum of Art runs into some tension between traditional cataloguing such as LCSH subject headings and folksonomies, and comments:

What seems to be hard to get across is that this is not necessarily an either/or proposition. The cataloguer’s LC standards can be met, if necessary, but end-user-friendly access terms can also be provided.

At the end of the day, that seems to be where the discussion lands – a well-tempered enthusiasm to explore folksonomies alongside the more traditional methods of providing access points, which are just taking root in the museum community. Richard Urban sums it up with this nice metaphor:

And as some have suggested it’s not an either-or proposition. I like to think about it like the Congress. The Senate (in theory) is supposed to be a slower-moving more deliberative body, whereas the House has it’s finger on the pulse of the people. Call it “bicameral cataloging” that takes the benefits from persistent standards-based description and more flexible approaches such as folksonomies.

While the jury is still out on folksonomies and how they may best be leveraged, I think this entire discussion shows yet another example of a somewhat parallel conversation in the museum and library world. Needless to say, there is an enormous amount of discussion on this very topic in the library blogosphere – some great leads can be found just by looking at some of the comments to Merrilee’s earlier posting on the issue. I’m curious to see what answers these two communities come up with, and whether they’ll be different.

8 thoughts on “Got Subjects?

  1. Pingback: Beth's Blog

  2. Great summary! Also, an interesting use of blogging – explore the spaces inbetween disciplines. Going to add you to my reader!

  3. Thanks for your comment and spreading the word about hangingtogether on your blog! I’m glad you found us…

  4. Günter,

    Thanks for a good summary of the MCN-L discussion. I wanted to highlight one of the attractions of folksonomy for the art museums involved in Steve that you didn’t mention: It’s a bridge between the museum and the visitor.

    There’s a “semantic gap” between the way that museum professionals talk about art and the way that the general public perceives it. What you see in a picture may not be reflected in art museum professional documentation. As Koven Smith reported at the MCN session, preliminary tests at The Metropolitan Museum of Art have shown a significant difference in the terms assigned by art historically and non-art historically trained staff. And what curators ‘know’ about a work may not be drawn from visual analysis at all. Michael Jenkin’s wonderful anecdote about the MMA’s curatorial trials reinforced this; one curator was stumped when asked to assign keywords that described a photograph: “Everything I know isn’t in the picture”, he said.

    Subject cataloguing (at any one of Panofsky’s levels of description: pre-iconographic, iconographic or iconological) is one aspect of what is often not included in museum documentation. Another is style and period terminology (think “Impressionist”). Then there is the lack of simple descriptions of the visual elements that make up a composition (red circle, blue square). Or there is the formulation of names. Is she Gabrielle Chanel, or Coco?

    In all these cases enabling users’ descriptions [through tagging] would capture their perspective on the works in museum collections, bridge the gap between visitor and museum, and further enable access to the collections we hold in the public trust.

    jennifer

  5. Pingback: connect.educause.edu

  6. I am glad that some people recognize that it need not be an “either/or propostion.” I use controlled terms very often. But sometimes I anchor a search with a controlled term and throw in a uncontrolled term. I recognize that others use uncontrolled terms exclusively. There is no need to privilege one over another; we can accommodate different approaches. Controlled and patron-supplied terms can co-exist in records. Abstract databases like INSPEC already allow this. I doubt that it would be difficult to do it in catalogs.

  7. Pingback: A s s e m b l y

  8. The concept of critical mass for tagging is interesting. Too little isa problem, but too much will bring problems of its own. When does ‘toread’ have enough traction to be helpful vs so wide adoption that it loses its value as a discriminator? When does critical mass tip over into “I’d need intensive care to get through this” mass? How long until Amazon automatically assigns ‘toread’ to every book it sells?

    I love Richard’s “bicameral cataloging.”

Comments are closed.