Three ways to make links between collections
A theme of my work is using Wikimedia projects to make links between different cultural collections. I look forward to the day when, instead of needing to visit dozens of different web sites and databases to find art on a theme, we use, and take part in building, new interfaces that are much quicker to use and still take us to authoritative information. This blog post reports on two new approaches I’m taking and revisits a third approach.
Joint exhibitions
Many public exhibitions would be impossible without loaned objects. For example, the Hajj: Journey to the Heart of Islam exhibition held at the British Museum in 2012 drew from 40 different public and private collections in 14 countries. The museum’s director, Neil MacGregor, said that telling the story of the hajj would have been impossible without loans from the Khalili Collections.
Wikipedia likes scholarly books from reputable publishers, and it likes third-party coverage (that is, editorially independent of the articles’ topic). Exhibitions are an ideal topic to write about on Wikipedia because there will be a published catalogue as well as many third-party sources. The hajj exhibition attracted a lot of attention as the first major exhibition of its kind in the world, and was reviewed by some of the country’s most famous art critics.
Exhibitions within a particular time window — a decade ago plus of minus several years — are ideal. Go back too far and the reviews and other sources will be hard to find online. For exhibitions that are very recent, the official web site and social media might be so rich in information that the Wikipedia article does not add so much value. Apparently the British Museum hosted an official web site for the hajj exhibition, but this doesn’t seem to still exist. Hence there is a valuable purpose in creating something which lets the public find out about the exhibition.
To get a quality rating badge on Wikipedia you have to show that the article is comprehensive: that it is a fair summary of the totality of reliable published sources. It’s hard to define that totality for a topic like visual art or the Mughal Empire but for a temporary exhibition, the totality of published sources is manageable to find and read.
In addition to writing a new Wikipedia article about Hajj: Journey to the Heart of Islam and getting it through Featured Article review (putting it in the top 0.1% of English Wikipedia articles), I created a category on Wikimedia Commons for the exhibition. Using the printed catalogue, I identified which of the photographs on Commons are of objects that were exhibited back in 2012. So far only the Khalili Collections objects are covered, but I hope that in the long run the other collections can be persuaded to upload images of the objects they loaned. Just as the original exhibition combined objects from multiple collections to tell a story, we now have the opportunity to create a digital surrogate for the exhibition by sharing images from those collections.
Exhibition histories can be represented in Wikidata using the property P608: exhibition history, which is not something I’ve much explored, but for some collections the data are impressively rich. Out of curiosity, I asked Wikidata which objects have appeared in the most exhibitions. Top of the resulting list is the painting “Body” by Luc Tuymans with 40 different exhibitions, showing that the exhibition data for the Stedelijk Museum is very detailed.
Dispersed manuscripts
I’ve previously written about how Wikidata can help us find different manuscript exemplars of the same literary work. When a single manuscript is dispersed across multiple collections, that’s another opportunity.
I’m currently working my way through Eleanor Sims’ book The Tale & the Image, Volume One: History and epic paintings from Iran and Turkey. For the Great Mongol Shahnamah, Sims has listed the collections and inventory numbers for all the illustrated folios. This links 20 public and private collections, including the Smithsonian, the Louvre, the Khalili Collections, and the Chester Beatty Library.
This sort of data is ideal for hosting on Wikidata and using to generate tables for Wikipedia articles, or list articles. I’ve previously written about best practice for representing medieval manuscripts on Wikidata. There probably has to be a Wikidata item for each collection’s group of folios, because that is the catalogued item with an identifier. Then we can use P361: part of and P527: has parts to relate this to an item for the dispersed manuscript.
Historical individuals
I’ve previously written about using Wikidata to find objects in collections that relate in some way to an individual person. My way of visualising this was with a chunk of “knowledge graph”. This shows how Wikidata represents connections between things with triples (subject-predicate-object) but it isn’t helpful for the end user.
To make something more end-user-friendly, I’ve created a Wikidata query that builds an image gallery. All objects in collections with any relation to the target person are listed. Those that have an image on Wikimedia Commons are prioritised; a placeholder image is used for the others. During my time at the Bodleian Libraries I created custom code to build these image galleries (on a server which is no longer active) but this approach uses Wikidata’s own query service and doesn’t need a separate server. Rather than linking to Wikidata, the links go to the catalogue records of the objects in their host collection.
Below are some examples. To customise the query code to a particular individual, just edit the SPARQL code to change the Q number in line 3.
Over time, I hope these lists become more complete, and more ugly placeholder images are replaced by scans of the actual objects.
