By Brian Palmer

A team of Italian researchers intends to piece together the skull of Lisa Gherardini, supposedly the subject of Leonardo da Vinci’s 16th century masterpiece Mona Lisa, so they can determine whether the painting’s iconic smile belonged to her. Some believe the smile was actually modeled after that of Leonardo’s assistant or the artist himself. Can you really tell how someone smiled from their skull?

Not exactly. Skeletal remains provide the broad outlines of a person’s mouth. The corners of our lips rest on the outer edge of the canine teeth, so a skull can indicate mouth width. The angles and contact points of the upper and lower teeth impact the way the lips meet, and the wearing on an adult’s enamel indicates the thickness of her lips. If those aspects of Gherardini’s skull don’t match the features on the painting, researchers could make a convincing case that Lisa Gherardini was not Leonardo’s muse. (The portrait has many other unique features that could be identifiable in skeletal remains, such as a broad forehead and elongated nose.) But it would be much more difficult to prove the positive case—that Gherardini was the inspiration for Mona Lisa’s smile. A person’s skull says very little about how he or she smiles. Many of the muscles that control a smile attach to soft tissue rather than bone, and there are indications that one’s unique smile is a learned behavior rather than an outgrowth of bone structure.

Forensic facial reconstruction has come a long way in the past few decades. Scientists are now able to recreate the shape of a person’s face—the location of the eyes, the slope of the nose, or the curve of the jaw—to within a couple of millimeters. The problem is that people don’t tend to identify one another based on those broad facial outlines. Researchers have used medical scanning technology to create a three-dimensional image of a living person’s skull, but not even people’s own family members are able to reliably recognize them. We are more likely to identify one another through features that forensic anthropologists haven’t yet figured out how to dependably reconstruct: skin color, eye color, hair, wrinkles, and characteristic facial expressions.

That’s why even Leonardo might have a hard time recognizing a modern reconstruction of Lisa Gherardini’s face. In a 2001 study, Australian forensic anthropologists used four reconstruction techniques to create 16 faces from four skulls, then challenged untrained observers to pick out a photograph of the decedent from a pool of 10 pictures. It was an abject failure: Participants identified only one of the 16 reconstructions at a rate that was significantly above random chance. Other studies have shown that personal acquaintances are no better than strangers at identifying reconstructions. The science has improved over the past decade, but forensic reconstruction researchers are still far better at generating hypothetical images of human ancestors than faithful recreations of modern faces.

Reconstructing the face of the real-life Mona Lisa faces a couple of other challenges. Gherardini was more than 60 years old when she died. Facial reconstruction experts have a few techniques to reverse-age someone, such as pulling back the brow ridge, which becomes more prominent in older people. But there’s a good chance that Gherardini was missing teeth. (Many in Renaissance Italy believed that toothaches were the result of dental worms, and others prescribed bloodletting or abstinence for oral discomfort.) If Gherardini’s gums were bare at her death, it would severely undermine an attempt to reconstruct her mouth.

Finally, although Leonardo da Vinci was known for his anatomically correct artwork, Mona Lisa is still a painting. It’s not clear whether Leonardo intended it as a work of photographic realism or an idealized version of his subject. Moreover, Leonardo labored over the painting for years, making it difficult to faithfully recreate every angle of his muse’s face.

Thanks to Shawn Phillips of Indiana State University and Caroline Wilkinson of the University of Dundee, author of Forensic Facial Reconstruction.