Ilka & Franz/Getty

By Brian Owens

Dogs are all honest, loyal and obedient, right? Well, not always. Our pets can be sneaky and manipulative when they want to maximise the number of tasty treats they get to eat.

Marianne Heberlein, who studies dog cognition at the University of Zurich in Switzerland, wanted to test the animals’ ability to use deception to get what they want from humans.

She got the idea to study doggie deception from watching her own dogs. One occasionally pretends to see something interesting in the backyard to trick the other into giving up the prime sleeping spot. “This sort of thing happens quite often, but it is not well studied,” she says.

To see if dogs would deceive humans too, Heberlein and her colleagues paired various pooches with two partners – one who always gave the dog treats and another who always kept the treats.

Thinking inside the box

After the dogs learned which partner was cooperative and which was competitive, the pets were given the opportunity to lead each partner to one of three boxes containing either a juicy sausage, a less-appetising dry dog biscuit or nothing at all.

After each trial, they led their owner to one of the boxes, and the owner would allow them to eat whatever was inside. This gave them an incentive to deceive the competitive partner by taking them to the empty box before leading their owner to the tasty treat. And that’s just what they did.

Over two days of testing, the dogs led the cooperative partner to the sausage box more often than expected by chance, and more often than they led the competitive partner there.

They also led the competitive partner to the sausage less often than expected by chance, and to the empty box more often than they led the cooperative partner there.

“They showed an impressive flexibility in behaviour,” says Heberlein. “They’re not just sticking to a strict rule, but thinking about what different options they have.”

Fast learner

Heberlein was also surprised how rapidly some dogs figured out the optimal behaviour. A few of them led the competitive partner to the empty box from the very first trial, and always managed to get the most treats.

“They were really quickly able to differentiate between the two partners. There was no additional learning step needed,” Heberlein says. Other animals, such as monkeys, often need dozens of repetitions to learn similar lessons, she says.

This feeds into an ongoing debate about what kinds of sophisticated cognitive abilities dogs and other animals share with humans, says Daphna Buchsbaum, who studies dog cognition at the University of Toronto in Canada. “We wonder, ‘Can they understand people’s mental state and motivations, and what causes people’s behaviour?’”

This work is a good first step, Buchsbaum says, the question is whether dogs are flexible enough to deceive in other contexts. “If they can, I’d say it was evidence of very sophisticated social reasoning,” she says.

Journal reference: Animal Cognition, DOI: 10.1007/s10071-017-1078-6