A study by A. Horowitz reported in ScienceDaily (here) purports to show that dogs do not feel guilt.
This study sheds new light on the natural human tendency to interpret animal behavior in human terms. Anthropomorphisms compare animal behavior to human behavior, and if there is some superficial similarity, then the animal behavior will be interpreted in the same terms as superficially similar human actions. This can include the attribution of higher-order emotions such as guilt or remorse to the animal…Horowitz was able to show that the human tendency to attribute a guilty look to a dog was not due to whether the dog was indeed guilty. Instead, people see guilt in a dogs body language when they believe the dog has done something it shouldnt have even if the dog is in fact completely innocent of any offense….Dogs looked most guilty if they were admonished by their owners for eating the treat. In fact, dogs that had been obedient and had not eaten the treat, but were scolded by their (misinformed) owners, looked more guilty than those that had, in fact, eaten the treat. Thus the dogs guilty look is a response to the owners behavior, and not necessarily indicative of any appreciation of its own misdeeds.
The problem here is that there is no control in the experiment. Would a child look guilty if they had done something they should not have but no one had noticed? I have my doubts about some children. Would a child look guilty if they had done nothing they shouldn’t have, but a parent started shouting at them and poking a finger in their face as if they had done something bad? I think I have known children who would look guilty because they assumed they had done some terrible unknown thing. And if you know what you have done wrong, you might look less guilty then if you had done something wrong and did not even know what it was.
Yes, we have to guard against anthropomorphism but we also have to guard against anti-anthropomorphism. Dogs are very social animals and so there is every reason for them to have social emotions like guilt as well as non-social ones like fear.
I guess that you are saying that this issues are very subjective, depend too much on who is evaluating the situation. That is true, but at the same time there are many disciplines that deal with highly subjective observations, nevertheless some reach proper conclusions with the use of appropriately designed evaluating methods.
Janet K: I agree with you Mariana that subjective observations are useful extremely useful, but they are not experiments. This was supposed to be an objective experiment rather than a subjective observation and it did not have the controls that should have been part of it. They came to a result on the subject of whether a certain behaviour was similar in dogs and people by examining the dogs’ behaviour but not the humans’. I was also saying that there are two biases that are dangerous: assuming that humans are more unique than they are and assuming that humans are less unique than they are.