I am going to focus on the metadata used in ARTstor. This is a database I frequently teach, and I am always looking for more effective ways of explaining how to use it. Since it’s an image database, it heavily relies on metadata searches, and this makes it particularly challenging to teach; most students I work with are much more used to a full-text keyword search.

I thought I would do some searching in ARTstor to identify the different metadata fields used in the database and then realized that it varies widely by item. I decide to check their website and discovered ARTstor has a very detailed explanation as to why that is: “Since each institution and individual can have differing uses of and requirements for metadata, the cataloging and descriptive data we receive vary greatly in the use (or absence) of standardized vocabularies, in the choice of terminology, and in the metadata schema used to organize the data.” The policy goes on to note that every item, though, can be searched by object-type classification, geography, and earliest/latest dates, that this information is added to records if necessary, and also that this information is not part of individual records.

The type of physical object represented by the digital object, the location the physical object originated from, and the date it was created/used are the features this consistent metadata describes. Metadata not created by ARTstor can describe a wide variety of other features, however: creator, culture, title, work-type, date (of both the physical object and the digital image), material, description, subject (these appear to be Library of Congress subject headings), who owns the physical object, copyright information, accession number, specific ARTstor sub-collection, measurements, period, style, digital image size, and possibly others. These metadata fields are not consistent, though, so searches based on them will always be missing items. The metadata provided by ARTstor really only lets you ask questions about geography, time period, and object-type; other searches are pretty much crapshoots.

The main feature the ARTstor-provided metadata does not describe – and this is the issue I run into when teaching – is what the physical object is about, what it shows. Sometimes the title might get at this, but often it doesn’t. Sometimes the subject headings or description will, but this metadata is inconsistently applied. I have many examples of this: medieval manuscript images that are about emotion, 18th century representations of women in the home or in public, material culture in early modern Europe, paintings of weather during the Little Ice Age, and so on. This is pretty similar to the gap Matthew Lincoln identifies in art history metadata – the lack of information about the visual aspects of artworks. These are the sorts of questions students ask about artwork and ARTstor isn’t able to facilitate this very well. What I end up doing in my teaching is telling students to enter one very broad search term, then limit the search by time period and geography, and then basically browse the results. I actually think digital browsing is underrated and provides a lot of contextual information about a topic, but students can get frustrated with getting irrelevant or zero results. It would be nice if subject headings, as flawed as they are, were more consistently applied in ARTstore, but I also recognize that the user community for ARTstor might not be as interested in those. Also, it is a lot of work to create metadata manually for individual items.