Aug 5, 2019

The Cailleux Project: Digital Tools to Analyze Art Provenance Texts

Who appears together with Cailleux in Cleveland's provenance texts?


In the previous post, we gathered provenances that mention the Nazi-era Red Flag Name "Cailleux" for two museums, the NGA and the Cleveland Museum of Art.


In this post, we will apply different "diagnostic" tools to these same museum provenance texts to identify, measure and visualize the art market networks involved in the the "Cailleux" artworks.

First, we'll identify the people and organizations that are mentioned in the same provenances as Cailleux. Then we'll plug this list of names into another tool - Voyant-Tools - to get a snapshot of their relative frequency and importance in the corpus. And finally, we'll visualize the networks.

This data is for only one name (Cailleux) in the provenance of only one museum (Cleveland)- an extremely small dataset.

The approach is to start with this modest experiment on a small number of provenances, and, when a data pipeline seems to work smoothly, to apply thus digital method to another museum, and another, and finally to include all the provenances for all the museums that mention Cailleux worldwide.
____


Who is named in the same provenances as Cailleux?  To find out we use Named Entity Extractor TextRazor.

The first thing we want to know is which people and organizations appear in the same provenance texts as Cailleux. There are many digital tools and methods that can do this. What we are looking for are reliable, easy-to-use, open access tools that anyone can use with little training. In previous testing of Named Entity Recognition Tools, TextRazor emerged as one of our favorites for its relative accuracy combined with extreme ease-of-use, so we will use it here.


Textrazor is a named entity recognition and extraction tool that spots names of people, organisations, places and, where possible, links them to Wikipedia, Wikidata and Freebase. It can be used via API or simply by pasting text into its online demo version.


https://www.textrazor.com




1. We paste the provenances that contain Cailleux at the Cleveland Museum of Art into the Textrazor demo:




2. Textrazor extracts a long  list of topics, categories, people, etc:



3. We are going to focus exclusively on "People/Person" and "Organizations" identified by Textrazor in the provenance texts:


/PEOPLE/PERSON
Georges Petit
William Esdaile
Marius Paulme
Denis-Pierre Papillon de
Calouste Gulbenkian
Henri Boudonneau
de Ganay
Jean Cailleux
Edward King
de Witte
Kanzler
John L. Severance Fund
M. de Vaize
Greta Millikin
Lépicié
Clovis Whitfield
Rene Boudonneau
Gallerie Cailleux
Mme S. A. Boudonneau
Alexandre Ananoff
M. Robert
Virginia Reynolds
Gordon H. Robertson
Monsieur Montsabert
William S. Talbot
Richard Houlditch
Anne Louis Goislard de Montsabert
Ardelia Ripley Hall
Ernest Kanzler
Galerie Cailleux
Rosemarie Kanzler
Louis-Charles-Léon Courajod
Günther Schiedlausky
Ay-whang Hsia
Kann
Rothschild
Bryas
William Talbot
Willis F. Woods
Frederick J. Cummings
Eugene Victor
André de Ganay
/ORGANIZATION/ORGANIZATION
Calouste Gulbenkian Foundation
Reichsleiter Rosenberg Taskforce
CMA
Valadon et Cie.
Galliera
Linet
Victoria Sears Goldman
CMA Lépicié
Bousquet
Cleveland Museum
Frick Art Reference Library
Detroit Institute
DIA

Given that they appeared somewhere in the provenance texts that were selected from the Cleveland Museum because they contained the word "Cailleux", we are going to take a leap and assume that there might be some connection between these names and "Cailleux".  At the very least, we know they appeared in the same provenance text for a given museum.

4. With this list of names, we will move to the next step, measuring the frequency of the appearance of the persons and organizations in provenances that mention Cailleux, as well as their relative closeness to Cailleux and to each other. 


To get a quick visualization we will further simplify the list returned by TextRazor and plug it into a second digital tool, Voyant-Tools as a Whitelist: (keywords-66cdd62ddce6034deab26ba4bffd64fc)

Ananoff
Boudonneau
Bousquet
Bryas
CMA
Cailleux
Cleveland
Courajod
Cummings
DIA
Detroit
Esdaile
Frick
Galliera
Ganay
Goldman
Gulbenkian
Hall
Houlditch
Hsia
Kann
Kanzler
King
Linet
Lépicié
Millikin
Montsabert
Papillon
Paulme
Petit
Reichsleiter
Reynolds
Robert
Robertson
Rothschild
Schiedlausky
Severance
Talbot
Vaize
Valadon
Victor
Whitfield
Witte
Woods



Voyant-Tools generates the Word Cloud below. The most prominent names for Cleveland are Cailleux (of course), with 30 mentions, followed by Cleveland and CMA. Of more interest are Kanzler, Gamay, Montsabert, Gulbenkian, Petit, Schiedlausky and Reichsleiter, the last two being names that appear frequently in connection to Nazi looted art.




Before we start looking at networks, a question:

How different would this image be if we included not only Cailleux provenance for the Cleveland Museum of Art, but also those for the National Gallery of Art?

to be continued in the next post....

note: remarks, suggestions, criticisms, tools and datasets welcome....


No comments: