Dogs may not be as honest and forthright as you think.

I had stopped by the home of a friend to lend him a book. He wasn't home, but his wife, Eva, invited me in for a cup of coffee. As we sat in their kitchen, their dog Hercules, a German Shepherd cross, suddenly jumped up and ran to a window and began barking. Eva walked to the window and looked out, but apparently there was nothing she could see there, so she gave Hercules a pat, murmured, "Good guarding," and returned. She explained, "I always check when he barks. You have to trust the dog. You know that dogs never lie or try to deceive you. Of course sometimes they misread the situation, but that's different."

Stanley Coren, PhD., DSc., FRSC,Stanley Coren, PhD., DSc., FRSC, is Professor Emeritus in the Department of Psychology at the University of British Columbia. His undergraduate degree is from the University of Pennsylvania and his doctorate in Psychology is from Stanford University. He is best known to the public for his popular books on dogs and on general psychological issues, however within the scientific world he is also a highly respected scientist having done research in a wide range of psychological areas including sensory processes (vision and hearing), neuropsychology (handedness, sleep, birth stress effects and behavior genetics) and cognition (information processing and intelligence). He has published over 400 scientific reports in professional journals as well as 19 books for students and professionals.

Editor: Nadeem Noor

Eva's comments were typical of a belief many people have, which is that dogs have the innocence of children and will never consciously engage in deceptive behavior. Of course, developmental psychologists understand that even young children will sometimes lie and deceive if they think it is in their best interest — and some recent research suggests that the same holds for dogs.

A team of researchers headed by Marianne Heberlein of the Department of Evolutionary Biology and Experimental Studies at the University of Zürich looked at an example of deceptive behavior in dogs. Their study appears in the journal Animal Cognition. These investigators studied 27 dogs, operating on the assumption that dogs are more likely to try to deceive people who they don't like, especially when it is in their self-interest to do so.

The hard part about this study was that the researchers had to establish a series of likes and dislikes in the dog. To begin with, they established whether each dog most preferred a piece of sausage or a dog biscuit as a treat.

Next the investigators had to demonstrate to the dogs that one of two women was nice and generous while another was not nice because she was selfish. This was done by having each woman go to a bowl and take out the dog's favorite treat — let's say a piece of sausage. The woman then called the dog by name. In the case of the "generous" person, she would step toward the dog and give him the treat. But the "selfish" woman, instead of rewarding the dog, let the animal see that she was putting the treat into her pocket, and then stepped toward the dog and showed her empty hands. This was repeated a number of times, and the dog was later tested to confirm that he preferred the person who was generous and would approach them spontaneously.

Next the dog was taught the command, "Show me the food." The purpose of this training was to teach the dogs to lead a person to one of two boxes, each containing a treat. During this training, the two women repeated their behaviors from before, with the generous woman taking the treat from the box that the dog indicated and giving it to the dog, while the selfish woman withheld the treat and put it in her pocket.

The actual testing situation was a little bit different. Now three covered boxes were presented to the dog. The dog watched while their preferred treat was placed in one of the boxes, and their less-preferred treat was placed in another of the boxes. Finally they were shown that the third box was empty, containing no treat at all. During this test session, one of the women would give the command, "Show me the food." She would then follow the dog to the first box that it clearly indicated. When she arrived at the box the woman opened it, and, in the case of the generous woman, she rewarded the dog with the content of the box, while the selfish woman placed the treat in her pocket. After the food was removed, the box was placed back in the original position.

Following this, the woman put the dog back on leash and brought it back to its owner, who was waiting behind a screen. The dog's owner then took the dog back to the starting position and asked it to choose one of the boxes. The owner followed the dog to the chosen box and opened it. If there was a reward inside the box, the dog was allowed to eat it. However, if the dog chose the box which had been opened before, the owner just showed the empty box to the dog and both of them went back to the start position.

So now you can see what the dog's dilemma is: He has been trained to lead a person to a box containing food. He knows that if he leads the generous person to the "best treat" he will get that treat. He also knows that if he leads the selfish person to that treat, he will not get it. However, there is an alternative: The dog could lie or deceive the selfish person by leading her to the less preferred treat, or even better, to the box with no treat at all in it — after all, she is mean and doesn't deserve a treat. If the dog does that, then he knows that a short time later his owner is going to take him back and give them the opportunity to choose a box. When that happens, if he chooses the box with the good treat, his owner will give it to him. But this will happen only if he first deceives the selfish person so that the good treat is still in the box.

So what do the dogs actually do?

The final testing was conducted over two days. During the first day, it was quite clear that the dogs tended to lead the generous woman to the preferred treat, and rarely to the empty box. When interacting with the selfish person, it appeared that the dog was choosing among the treats by chance. On the second day of testing, though, the pattern of behavior was much clearer. Now, on nearly 80 percent of the trials, the dog would lead the generous woman to the preferred treat and scarcely ever to the empty box. As for the selfish woman, the dog led her to the preferred treat less than 20 percent of the time. Instead, the dog was equally likely to direct the selfish woman to the empty box or the less-preferred treat.

On the basis of these data, the researchers concluded that dogs are thoroughly capable of engaging in deception when it is in their self-interest to do so. It is as though the dog is thinking, "Why should I tell that selfish person where the best treat is if it means that I will never get it? It's more rewarding to deceive that person by leading her to a less preferred treat, or even better to a box with no treat at all. If I do that, then when I get another chance to select a box with my owner, I can pick the good treat and know that I will get it. I win and that selfish lady loses."

I immediately thought of the situation between Eva and Hercules. Was Hercules lying about there being something outside the window despite Eva's belief that dogs don't deceive? Well, Hercules did get a rewarding pat on the head and a little bit of favorable attention by rushing to the window and barking. That means it could be that it was in the dog's self-interest to employ a little bit of deception. We certainly now have data showing that dogs are capable of engaging in that sort of trickery.