A guilty look doesn’t mean a dog’s guilty


Look at him. He knows that he misbehaved and now he feels guilty.

Many dog owners think that they can recognise whether their pet disobeyed them or did something naughty based on its behaviour towards them. They also think that their dog looks guilty because they know they did something wrong. However, even though we can’t precisely know what’s going on in dogs’ heads, the science strongly suggests that most of what people think about dogs’ guilt is not true.

Already in 1977, a study showed that dogs show ‘guilty’ behaviours in the vicinity of shredded paper (at that time destroying the owners’ newspaper was a common dog ‘misbehaviour’) whether they or the owner were the one to make the mess.

A more recent study investigated whether showing behaviours associated with guilt was more related to what the dog had done, or its owner’s reaction. First, the owners identified behaviours they associated with guilt: avoiding eye contact, lying down and rolling on the side or onto the back, dropping the tail, wagging low and quickly, holding ears down or head down, moving away from the owner, raising a paw and licking. Then the researcher counted how many of these behaviours the dogs showed in different experimental conditions.

The owners were instructed to tell their dogs (in the way they would usually do) not to eat a treat they placed on the floor and left the room. Then the experimenter either took away the food or gave it to the dog to eat. In each of these cases, the owner was either told that the dog disobeyed the earlier command (in which case the owner should come into the room and scold it) or not (in which case the owner should come into the room and greet the dog in a friendly manner). This way owner’s belief was independent of what the dog actually did.

The results showed that the dogs’ expressed as many guilty behaviours whether they ate the food or not. However, the dogs that were scolded showed clearly more guilty behaviours. Moreover, the dogs that were scolded but didn’t eat the food looked most ‘guilty’.

Another study tested the owner’s claim that they can recognise their dogs’ guilt, but this time without any scolding involved. The owner placed the food somewhere in the room, forbade the dog to eat it and left the room. The experimenter either took the food away or let the dog eat it and before the owner came back, replaced it or not. After entering the room and a short observation of the dog’s behaviour, the owner had to say whether their dog ate the food or not.

Around 50% of the owners said that their dog ate the food. The responses were not dependent on the actual action of the dog or whether the food (replaced by the experimenter) was present when the owner returned. Therefore, it seems that the owners just guessed what the dog did and that the “proof” of misdeed (no food) did not affect the dogs’ behaviour.

A different study tested the claim of some of the owners that the dogs show guilty looks even before the owners find out about the misdeed, and therefore before the owners scold them or react in any other way. This study had a different, more complicated design but in the test phase, the owners left the food on the table, told the dog not to eat it and left the room for 3 min. On entering the room again, the owner had to observe the dog and say whether their dog ate the food (they could not see the table because of the barrier).

The number of guilt-associated behaviours did not differ between dogs that ate the food and those that did not. When looking at the behaviour of individual dogs, the ones that ate the food in the test phase showed a guilty look more often compared to their greeting behaviour at the start of the experiment. However, they were also scolded at least once in the early phase of the experiment and they showed as many guilt-related behaviours later, when they had no chance to disobey their owners. As in the previous study, the dogs showed more guilty behaviours when being scolded compared to when being greeted.

Although on average the owners seemed to be able to tell whether their dog disobeyed them, more than one-third were wrong about that. Also, it seems that most owners based their opinion on the dog’s behaviour in the earlier phases of the experiment. The owners of dogs that showed inconsistent behaviour earlier (sometimes ate the forbidden food, sometimes not) were not able to say whether their dog ate the food in the test phase.

All in all, these studies show that the owners are not good judges of whether their dogs disobeyed them and that the dogs’ behaviour is mainly affected by human actions. It should be noted that the behaviours that the owners relate to guilt are also common signs of fear, stress or appeasement. So, it’s not surprising that the dogs expressed them when scolded. Dogs can also detect slight changes in owners’ mood and can remember their owners’ behaviour in similar situations and the guilty look may be a fearful anticipation of something unpleasant. Also, many owners say that their dogs’ guilty look reduces their scolding which can reinforce its use by the dogs.

We would do our dogs a favour by not assuming that they feel guilt, but rather noticing when they are afraid of us and finding ways to teach them desired behaviours in a manner that doesn’t damage our relationship with them (fear and stress can also lead to various behavioural problems).

Leave a Reply

Your email address will not be published. Required fields are marked *