Mapping the Faces of Early America

Science magazine’s website reports on the new paper “What did the early American presidents really look like?: Gilbert Stuart portraits as a ‘Rosetta Stone’ to the pre-photography era,” by Krista A. Ehinger and Eric L. Altschuler, published in the journal Perception.

The hypothesis of this paper is that we can compare photographs of mid-nineteenth-century Americans to their portraits in youth by Gilbert Stuart, and thus mathematically figure out how Stuart’s images relate to their actual faces in later life. Then we can apply the same algorithm in reverse to Stuart’s portraits of people who didn’t survive to the age of photography, such as George Washington, and determine what they “really” looked like. (Or perhaps would have looked like thirty years later, if they’d survived.)

Some commenters aren’t so impressed by this approach. Margaret Livingstone of the Harvard Medical School says, “The differences [between Stuart’s rendering and reality] are minor compared to the effects of aging in photographs taken decades later.” In other words, the portraits are very good to start with.

Here’s a third data set to add to the mix: the busts of several of those same American statesmen in the late 1820s which John Henri Isaac Browere modeled from plaster casts of their actual faces. Those busts are at the Fenimore Art Museum in Cooperstown, New York, and they’re wonderful. That set includes John Quincy Adams, one of the subjects mentioned in the Perception article. So it’s possible to compare Stuart’s painting of J. Q. Adams to a very good three-dimensional rendering made only about a decade later.

It’s also possible to test the algorithm by comparing Stuart’s portraits of John Adams (there are more than one) and Thomas Jefferson to Browere’s busts of Adams at ninety and Jefferson at eighty-two (shown below).