Dogs capable of tactical deception, study suggests
Dogs are capable of tactical deception and will trick people to get what they want
A study by Swiss researchers has revealed that dogs are capable of tactical deception and will trick people to get what they want.
Published in the journal Animal Cognition, the study was led by Marianne Heberlein from the University of Zurich.
She came up with the idea of testing a dog’s ability to use deception after watching her own dogs. One would sometimes pretend to see something interesting in the garden to fool the other into giving up the prime sleeping spot. “This sort of thing happens quite often, but is not well studied,” she told New Scientist.
To see if the dogs would deceive humans, Marianne and her team paired dogs with two people - one that acted ‘cooperatively’ by giving food and one that acted ‘competitively’ by keeping the food for themselves.
After the dogs learned which person was cooperative and which one was competitive, they were given the option to lead the cooperative partner to one of three locations containing a favoured food item (sausage), a non-preferred food item (dog biscuit) or nothing at all.
After each trial, the dog led a partner to one of the locations and would be allowed to eat whatever was there. This gave the dog an incentive to deceive the competitive partner by leading them to the empty location, before taking their cooperative partner to the preferred food.
On the first day of testing, the dogs led the cooperative partner to the preferred food box more than expected by chance, and more often than the competitive partner.
On the second day, the dogs led the competitive partner less often to the preferred food than expected by chance, and to the empty location more often than the cooperative partner.
‘These results show that dogs distinguished between the cooperative and the competitive partner, and indicate the flexibility of dogs to adjust their behaviour and that they are able to use tactical deception,’ the authors conclude.
The latest 
