Analyzing the Biases of Art History With Hard Data

Data visualization featuring 2018 statistics, from Painting by Numbers: Data-Driven Histories of Nineteenth-Century Art by Diana Seave Greenwald (Princeton University Press, 2021; all images courtesy the publisher)

We can learn a lot by counting things. Putting those numbers in context can prove revealing and even influential. Data has the power to shape how we think, the questions we ask, and the conclusions we draw; analysis of data has long been standard for a plethora of disciplines as numbers can help us understand our own biases.  Art history, Dr. Diana Seave Greenwald argues, is no different.  

Art historians of nineteenth-century portraiture and landscapes have long been interested in how aspects of nineteenth-century life — things like empire, industrialization, and gender — have played out in the century’s most iconic genres of painting.  Greenwald’s new book, Painting by Numbers: Data-Driven Histories of Nineteenth-Century Art, summarizes her analyses of over 500,000 paintings from France, the United Kingdom, and the United States with the express motivation of examining whether long-held interpretations of painting trends hold up under quantitative scrutiny.  Many do not.  

The cover of Painting by Numbers: Data-Driven Histories of Nineteenth-Century Art by Diana Seave Greenwald (Princeton University Press, 2021)

Painting by Numbers illustrates that decades of sampling biases (for example, lost or damaged artworks; unsaleable pieces; discriminatory exhibition practices) ensured that only a narrow sample of work produced by nineteenth-century artists was preserved for contemporary scholars to study.  It should come as no surprise that further contemporary bias comes from disproportionate attention that is given to well-known and accessible artists.  

In other words, without systematically addressing sampling bias in what nineteenth-century paintings are held up as exemplars, the paintings and artists that are “known” simply become “more known”; the bias becomes self-reinforcing.  “I was curious to explore if the trends that are emphasized in certain branches of art history were as prominent as they are often presented,” Greenwald explained to Hyperallergic over video conference.   

Trained in both art and economic history, Greenwald’s fresh approach — using established quantitative methodologies from sociology and economics and applying them to datasets like exhibition catalogs to analyze trends in paintings — begins to unpack the scale and scope of nineteenth-century art production. Specifically, looking at paintings on this scale offers a more robust way of gauging what was produced where, by whom, and how that art was exhibited and consumed.  This approach is “pragmatism in the name of novel inquiry,” as Greenwald neatly summarizes in the book.

For example, her analysis of depictions and omissions of the British Empire at the Royal Academy reveals that fewer paintings from “abroad” were exhibited than one might expect, especially if nineteenth-century paintings were a way of laying visual claim to colonial territories, as historians have theorized.  “Works — of all genres, including landscape, history painting, and portraiture — that list an identifiable British (or Irish) site in their titles routinely make up 5 percent or more of all works shown at the Royal Academy,” Greenwald explains in the book.  “While this number may seem small, it is much greater than the number of depictions of Britain’s overseas colonies.”

Camille Pissarro, “The River Oise near Pontoise” (1873), oil on canvas (public domain image courtesy the Clark Art Institute)

In other words, painters whose works were exhibited at the Royal Academy between 1769-1914, depicted the metropole (that is, they painted “home”) far more often than any other part of the British Empire, despite Greenwald’s expectations to the contrary (colonial landscapes “abroad”), given the importance of depictions of empire in corresponding scholarly literature.  Greenwald attributes this to the practical and financially arduous difficulties of travel, varying political access to geographic spaces, as well as the incompatibility of imperial extractive institutions with the subject matters preferred by the Royal Academy.

Painting by Numbers is a methods book.  It’s careful and systematic.  In the text, Greenwald is very careful — cautious, even — to couch every assertion in context and to continuously iterate the limitation of the analysis tools she uses.  It is not a narrative-driven story about the artists or the work they produced, but it is a solid demonstration that “counting things” matters.  It leaves audiences to wonder what work the book will inspire as other researchers draw from the quantitative foundation Greenwald has established.  

Perhaps the most compelling aspect of Painting by Numbers is the care it takes in the data visualization and its presentation.  It’s easy for graphs, tables, and charts to become visual roadblocks that readers must glance over and around en route to the next paragraph of text.  But in Greenwald’s text, it’s clear that the author’s expertise in art and data pair brilliantly; the presentation of her analyses turns into an art itself, with the assistance of Carlos Capella.  (We’re a long, long way from the ugly, unwieldy Excel graphs of the early aughts.)  Tables are spaced out appropriately, and a plethora of different styles of graphs and colors offer a clear, easy way to follow the information.  Data visualization is an art of a million details and it’s one that Greenwald has clearly mastered.  

Painting by Numbers: Data-Driven Histories of Nineteenth-Century Art (Princeton University Press, 2021), by Diana Seave Greenwald, is now available on Bookshop

Careful and yet compellingly fresh in its approach, Painting by Numbers offers a new kind of methods book.Read MoreBooks, data visualization, Diana Seave Greenwald, Princeton University Press, ReviewsHyperallergicRead More

Leave a Reply