French singers Edith Piaf and Henri Salvador in 1959.adoc-photos/Corbis

Shortly after Steven Livingstone arrived at McGill University, he went out to a bar with some new friends to check out the Montreal nightlife. It was loud, and crowded, and dimly lit, and he found himself struggling to hear when one of his friends tried to talk to him.

“Because it was dark at the bar, I couldn't see his face either,” Livingstone says. “But I could see his head. He was moving his head in a really sort of animated fashion, and it was at that moment that I realized that he was excited about something. He was trying to tell me about it. So that sort of lit a little candle in my head.”

He was working in the McGill psychology professor Caroline Palmer’s Sequence Production Lab at the time, and told her about his observation. And that night in the bar led to a new study on the role of head movements in conveying emotions, recently published in the journal Emotion.

Livingstone, now a postdoctorate fellow at McMaster University, Hamilton, and Palmer recorded the head motions of 12 adults with at least six years of singing experience, “while they spoke and sang four statements, each with five emotional intentions (very happy, happy, neutral, sad, and very sad).”

Whether singing or just talking, people reliably lowered their heads and tilted them toward the floor when using sad intonations. For happy intonations, people looked up, raised their heads, and moved them quickly forward toward the listener. The more intense the emotion they were trying to convey, the more exaggerated the movements became. When people were feeling neutral, their heads moved much less.

In the second part of the experiment, the researchers showed 24 people the  recordings, but blocked out the facial expressions and played them without sound, so all the viewers had to go on was head movement. The participants were able to identify the intended emotions 70 percent of the time when the people in the video were speaking, and 67 percent of the time when they were singing. Despite the small sample sizes, Livingstone calls the results “robust.”

Livingstone says he hopes this work could have implications in robotics, helping engineers make robots with more naturally human movements. (Though first, future research would need to test whether these head motions are just specific to North America, and whether other cultures have their own emotional head movements.)

In the meantime, he’s continuing his observations. “I have to say I have been playing the scientist at home,” he says. “I’ve been watching people. I've been really quite surprised at the range and number of people who actually do this.”

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.