Study: Dogs Can Identify Liars, and They Don't Trust Them
Pets may be more sophisticated than you think. Don't steer them wrong.
Posted Feb 24, 2015
One of the delights of watching the behavior of young dogs is the way that they approach almost every human being in a trusting and friendly manner. The same applies for young human children, who enter into social interactions with the idea that every human adult is trustworthy and has their best interests at heart. Of course, over time, the human child will learn that some people are more reliable and worthy of trust than others.
New data shows that the same holds for dogs.
It now seems clear that that if you frequently lie to dogs, they lose their trust in you and begin to act as if they can no longer rely on the information that you give them.
In a pair of studies published in the journal Animal Cognition*, a team led by Akiko Takaoka of Kyoto University in Japan showed that dogs will only use information and follow commands from people who have a track record of being trustworthy.
The first study involved 24 dogs and relied on the fact that dogs will reliably go in the direction that a human being points to. Two opaque containers were presented to the dog—under one of which was hidden a bit of food. In the first phase, a researcher pointed toward the container that had the food concealed inside; as expected, the dogs ran to that one and got the reward. The second phase was designed to show that this human was no longer trustworthy. Now, while the dog watched, the researcher showed that one of the containers had food under it while the other did not. However, just before the dog was released, the person pointed toward the empty container and encouraged the dog to go to it. The final phase was a repeat of the first part of the test, with the dogs being shown two containers and the experimenter again correctly pointing toward the one which had the concealed bait.
The results were quite dramatic: In the first phase, the dogs showed their usual trust of all people and the majority went to the container that the researcher pointed to. However, in the final phase, the dogs showed that they had apparently learned that the researcher was untrustworthy, and only 8 percent reliably went in the direction that he pointed to.
Did the dogs learn that all people are unreliable in this task, or that only this particular person is a liar? Previous research has shown that dogs make judgments about the personality and behavior of specific people, and do not generalize to all people. Thus dogs learn to pick out those people who are selfish and those who are not. So the researchers repeated the experiment with a new group of 26 dogs. The first two parts of the study were the same as in the initial experiment—thus, in the first phase the experimenter reliably pointed to the baited container, and in the second phase showed the dog which container contained the piece of food, but then pointed to the empty one in order to prove his unreliability. The trick in this second study was that in the final phase the "liar" was removed and a new experimenter, previously unknown to the dog, replaced him for testing. Now the dogs proved that they had not lost faith in all humans but only in the one who had been shown to be untrustworthy. They did this by once again acting reliably on the information provided by the pointing of the new person and going to where he pointed.
In an interview with the BBC, Takaoka said that this means that dogs can use their experience with particular human beings to determine whether they can be trusted. She noted that she was surprised that the dogs "devalued the reliability of a human" so quickly, and went on to say, "Dogs have more sophisticated social intelligence than we thought. This social intelligence evolved selectively in their long life history with humans." In other words, while dogs interact with human beings, they also try to determine the nature and personality characteristics of each person. They use this information to predict the future behavior of specific people, and adjust their behavior accordingly.
This kind of judgment about the reliability and trustworthiness of people seems to be a rather sophisticated matter, at least in human beings. A study conducted by a team of researchers led by Kimberly Vanderbilt** of the University of California, San Diego, did a somewhat similar study using preschool-aged children. They found that even after they had been shown that some people were untrustworthy, 3-year-olds tended to accept the advice of the known liars to the same degree as from the truthful people. Four-year-olds were more skeptical but were still accepting advice from the unreliable people, and only the 5-year-olds systematically preferred the information from the more trustworthy person. So the fact that dogs, who are usually credited with having mental abilities similar to a human 2- to 3-year-old child, can make this kind of discrimination reliably is somewhat surprising.
In any event, this research shows that dogs keep track of whether people lie or tell the truth, and use these memories to determine whether they can trust particular humans and the information they get from them.
In other words, if you mislead your dog, he will remember those lies and you will not only lose his trust but may also lose his cooperation.
Stanley Coren is the author of many books including: The Wisdom of Dogs; Do Dogs Dream? Born to Bark; The Modern Dog; Why Do Dogs Have Wet Noses? The Pawprints of History; How Dogs Think; How To Speak Dog; Why We Love the Dogs We Do; What Do Dogs Know? The Intelligence of Dogs; Why Does My Dog Act That Way? Understanding Dogs for Dummies; Sleep Thieves; The Left-hander Syndrome.
Copyright SC Psychological Enterprises Ltd. May not be reprinted or reposted without permission
* Akiko Takaoka, Tomomi Maeda, Yusuke Hori & Kazuo Fujita, (2015). Do dogs follow behavioral cues from an unreliable human? Animal Cognition,18, 475–483, DOI 10.1007/s10071-014-0816-2
** Kimberly Vanderbilt, David Liu, Gail Heyman (2011) The development of distrust. Child Development, 82, 1372–1380