Getty Provenance Index Update (ctd.)
May 2 2025
Picture: Getty Provenance Index
Posted by Bendor Grosvenor
Dear AHNers - we have an art historical emergency on our hands. I'm afraid the newly updated Getty Provenance Index (which Adam first reported below) is a disaster. I've no doubt it was built with good intentions. Art historians everywhere will always applaud the Getty for investing so much in making the raw materials of art history available for everyone, for free. But the new Getty Index simply does not work.
I would say that we must hope the Getty can fix it, but it seems the new site has been years in the making. I'm told it will not be easy to change. Our best hope is that they retain the old Index as a standalone website. The old Index is still available here. If it ceases to be made available (and I gather the plan is to soon switch it off) then provenance research will go backwards by decades. The new website threatens not only to be a setback for regular provenance research, but restitution claims too.
Regular AHNers will I'm sure be familiar with the old site. As you can see from the image above, it was possible to search by multiple categories, including: artist, title, owner, date, auction house, and so on. You could find within seconds a specific painting from millions of entries, covering art sales over the last four hundred years. An entry would have the date of sale, auction house, often dimensions, all displayed easily. Sale data from 1933-45 was especially detailed, which, combined with other databases like LostArt.de made searching for potentially looted artworks more accessible than ever before.
Within the old Index system, you could also bring up the whole contents of a particular sale. The sale entry would often have a long note compiled by one of the Getty's amazing provenance researchers, giving further information about who was selling what, when and where.
Hardly any of this is now possible in the new Index. There is no equivalent way of searching by specific criteria, like former owners. Even when you find a painting, it is a struggle to find basic information about it, like when it was sold, or at which auction house. The research notes about the sales seem to have vanished. I've had some discussions among fellow art historians who have been trying to use it. None of us can get the results we used to. The new user guides are jargon heavy and hard to understand. So far, we cannot believe how much of a setback the new site represents, nor that the Getty has spent so long and so much on making the Index worse. (If someone from Getty wants to get in touch to show we are mistaken, please do!)
The old Getty Index was one of the most transformative art historical tools of the last two decades. Anyone, anywhere could do provenance research which would either have taken months in an archive, or was simply impossible before. Here's a couple of examples from my own work.
Sometimes, the Getty Index made breakthroughs extraordinarily easy. For example, when the above early 17th painting of three girls holding fans came up for sale at auction in 2008, the sitters were unidentified. I simply put the word 'fan' into the Getty's title box, and amongst the 133 paintings sold with that word in the title, soon found that the painting at auction had been sold in 1824 with comprehensive details about the sitters' names. Other research proved that this evidence was correct, and that the sitters were indeed three sisters from the Egerton family.
Of course, provenance research is rarely so easy! But without the Getty Index, it would have been almost impossible to find in this case. Sadly, I have tried to replicate the 'fan' search on the new Index site, with no success. You cannot search among titles, indeed there appears to be no title category at all. There are a mystifying number of other categories, like 'identifier for object', but few usable art historical terms. Moreover, you cannot bring up a large number of items to search through at a time, like those 133 paintings with the word 'fan' in the title. The most I can get is 5 results at a time, and wading through the results is very laborious.
Another example. When I was investigating a painting of a mystery bridge in Derby Museum for Britain's Lost Masterpieces for the BBC, I was able to find on the Getty Index an entry for a painting by Joseph Wright of Derby of the Ponte Nomentano.
The entry onthe old Getty Index told me the painting was in Wright's posthumous sale, and was unfinished. It was a match for our painting. There were also links to digital scans of the original catalogues in other databases.
It is possible to find this painting on the new Index site, if you put in 'Nomentano', but even then it is hard to find the artist's name, let alone further information about the sale, and so on. The results page - with headings like 'Object Used in Data Assignment' - suggests that the new Index has been designed more with 'data' in mind than art history.
All of us who have worked on provenance research are enormously grateful for all the work the Getty and its scholars have done to support our field. If we lose the Getty Provenance Index as it was, it will be an extraordinary step backwards for art history, and all the work the Getty has done over almost 40 years will be diminished, or even redundant. What can be done? We must somehow get the Getty to agree for this old site to remain online. If you can, please email the Getty at ProvenanceIndex@getty.edu to make the case for this.
Update - a reader writes:
I suppose to be fair one should try to work with the new format. It may be possible to get it to do what the old one did with a little persistence ...
The same reader writes short while later:
I’m wrong. It is rubbish.
Update II:
A reader with experience of construction art historical databases writes:
'Linked data' [the idea behind the Getty's new database] was the next big thing in the internet about 15 years ago [...] The idea is simple: to offer a highly structured and logical scheme into which every single bit of historical data can fit. The vision then was that the internet was going to become one vast repository of information that was linked logically.
I'm really surprised to see that this approach has survived. It's fine when you have lots of data but in fact the Getty provenance index is quite 'thin' - it has tons and tons and tons of lists of pictures, and (ok there is a bit more, but) that's about it. So when you create a really fancy structure, all the user really sees is the structure - you can't really see the small amount of data within, because it's hidden by all the scaffolding surrounding it. They haven't done any user testing among people who actually use the database for art history - or if they did, they just ignored it. I have no idea, for example, how to get a catalogue view. Looking up an artist and trying to get a simple list of everything he did - that doesn't seem to be possible. Nowhere do we learn which bits of data are the direct primary source and which are extrapolations.
This rings true with my attempts to use the new database so far. Most of what I see is the structure.
