Robin Thomas said she had two young women in a psychology course she taught. One was named Kristen, the other Heather.
"And I had that hardest time remembering which was which," she said. "I mentioned it to them at the end of the term, and they said, 'All our friends have the same problem.'"
Thomas, a psychology professor at Miami University in Ohio, began to wonder. One woman was a tall blonde; the other was a shorter brunette. Perhaps somewhere in her mind, something determined what someone named Kristen or Heather ought to look like -- and these two students broke the mental mold.
So Thomas and three colleagues -- Melissa Lea, Nathan Lamkin and Aaron Bell -- devised a series of experiments to test people's preconceived notions of names.
In one, they gave volunteers a series of common male names (Bob, Bill, Tom, etc.) and asked people to create pictures of what someone with one of those names ought to look like.
In another test, they showed their volunteers pictures of two men -- one with a round face, the other with a thinner, longer face -- and asked them to decide which was "Bob" and which was "Tim."
The results they report were remarkably consistent. A "Bill" had a certain look to people, and it was different from a "Mark."
And Thomas said she found the stereotypes useful. If a guy was named Joe -- and somehow looked the part -- then it was easier to remember him.
"The better the fit," she said, "the faster you were at learning the names."
Thomas and her colleagues have written an academic paper on their findings, which has been accepted for publication in the journal Psychonomic Bulletin & Review. To rule out factors such as gender and ethnic origins, they restricted their experiments to names common among white males in Ohio, where they recruited the volunteers for the study.
In their paper, the researchers wrote, "Few worry about whether the name will provoke a facial stereotype in the minds of others (hmmm … he doesn't look like a 'Bob'), but, as the present research suggests, this may be yet another potential worry to have when one selects a name for one's progeny."
But why should any of this be? For now, that remains a mystery, but psychologists have several theories.
It may be, for instance, that people make associations between names and the faces of famous people who have those names. If you hear the name "Tom," you may automatically think of Tom Cruise.
But that doesn't hold up when one considers how many different-looking people have names Thomas' team tested. When you hear "George," what face comes to mind first? George Washington? George W. Bush? George Clooney? They don't exactly look alike.
Thomas' favorite theory, one that she hopes to test in future experiments, is that deep inside the brain, certain sounds automatically bring certain visual connotations to mind.
"We suspect that the sound of a name carries over into the visual," she said.
She points us back to the Bob-vs.-Tim experiment in which people overwhelmingly decided that someone named "Bob" had to have a round face.
"Listen to yourself as you make the sounds, and the way you shape your mouth," she said. "Perhaps 'Bob' is a round-sounding name and 'Tim' is a thin-sounding name."
It may well be that certain names really do have a connection to certain types of faces; after all, some names do carry in families. But Thomas said she doesn't explain the stereotyping we automatically make.
"I'd be really spooked if all the Bobs in the world really were round-faced."