Why You Look Like Your Dog | Smart News | Smithsonian

Keeping you current

(Photo: Rick Gomez/Corbis)

Why You Look Like Your Dog

It's all in the eyes

smithsonian.com

Dogs and their owners tend to look alike, and scientists (plus, pretty much anyone who owns a dog) have known for years that even strangers, when asked to look at images of dogs and people, can correctly match owners to their dogs with above-random rates of success. But the exact attributes that dogs and people share remained a mystery. 

Now, new research conducted in Japan has nailed down one clue. As Slate reports, the factor dogs and humans have in common, the researchers found, is the eyes.

To arrive at this conclusion, the researchers recruited around 500 college students and asked them to look at photos of 40 people and dogs, all shown from the shoulders up, against white backgrounds. The people and dogs were paired, some in the correct combination of owner and pet, others randomly. The participants' task was to pick out which combinations of actual pets and owners, and they chose correctly 80 percent of the time. 

Next, the researchers obscured parts of the photos, covering either the dog's or the human's eyes or mouth, Slate explains. The participants still did well if the mouths were covered, with a success rate of 73 percent. When the eyes were covered, however, that rate dropped drastically, with success becoming a matter of random chance. Encouraged by this finding, Slate writes, the team cropped the images to only show the eyes of the humans and their dogs. Impressively, students who saw just the eye photos had a 74 percent success rate. 

Apparently, the key give-away linking dog and human is the eyes. What specifically give us what Slate calls an "apparently superhuman (or at least subconscious) ability to extract meaningful psychological cues from eyes" remains unknown. But there's something about those big, wet, adorable doggie eyes that connects humans to the pets they pick.

Tags

Comment on this Story

comments powered by Disqus