When Will Dogs Try to Help Humans?

Dogs only help if they understand what is needed.

Posted Aug 14, 2013

Creative Commons License CC0
Source: Creative Commons License CC0

If you believe many  of the TV news broadcasts and postings on YouTube then you also believe that dogs often help people spontaneously, and often heroically. There are many reports of dogs pulling people to safety when they have fallen through ice, are drowning, trapped in fires, or pinned under overturned vehicles. There are also many reports of dogs going for help, and bringing rescuers to the site where their owner is caught in a burning building, suffering from a heart attack or diabetic coma, trapped in a cave, sewer, or culvert, or incapacitated because of an accident. Such spontaneous helping behavior from dogs has, in the minds of most of us, made all dogs potentially the kind of heroes portrayed by Rin Tin Tin and Lassie.

When scientists talk about helping behavior they usually break things up into four categories, comforting, sharing, providing information or informing, and the last category is what they call "instrumental helping" which is physically doing something which aids another individual. Obviously the press reports are focusing on this last category.

Recent studies suggest that human children have a biological predisposition to try to help others, and this begins to become quite clear between the ages of around 18 months to two years of age. Since a lot of evidence shows that dogs have minds which works similarly to children 2 to 3 years of age (see examples here or here) this would explain why untrained dogs might also spontaneously try to help, Despite extensive evidence of dogs helping humans, scientists are still not sure whether dogs are internally motivated to help their human companions (as are young children) and engage in helping for the sake of helping, or is it possible that the dogs are simply trained to follow certain commands or react in particular situations in certain ways.

A recent study which is scheduled to appear in the journal Applied Animal Behaviour Science sheds some light on this issue.  It was conducted by Juliane Bräuer, Katja Schönefeld and Josep Call of the Max Planck Institute for Evolutionary Anthropology in Leipzig Germany. They reasoned that dogs will help, but only if they clearly understand the nature of the help needed.

To test this hypothesis the researchers created a compartment or miniature room made of clear Plexiglas and without a roof. The door to the room could be opened by pressing a button. The dogs tested were all household pets, and they were trained to push the button to open the door to get a treat which they could see on the floor of the enclosure. Sometimes however, instead of a food treat, the dogs saw a ring of keys placed in the room. Obviously the keys were of no interest to the dogs and, as might be expected, they showed no interest in pushing the button to enter the room under those circumstances.

Now the researchers set up a situation where the experimenter (not the dog's owner) needed help to open the target door to get the keys. In one condition the experimenter indicated that they needed help by performing three actions in a predetermined daughter: (1) looking  and staring into the room through the door, (2) moving towards the door  and pushing and shaking the door and (3) reaching for the key above the door. In another condition gazing directly at the dog and then back to the keys was added. In a third condition the experimenter added some spoken phrases in a fixed order, “Oh, where is my key? / There is my key! / How can I get it? / I want my key! / How did it get  there? / Usually this door is not closed.” All of this could be contrasted to a situation where food was visibly placed in the room rather than the keys.

The results were straightforward if not inspiring. Around 90% of the time when it was in the dog's self-interest to open the door because there was food, the dogs pushed the button. When he came to helping the experimenter, regardless of the nature of the entreaty for help, the dogs helped around one third of the time.

Now the researchers feared that perhaps the nature of the help needed by the experimenter in the first test was not obvious enough, they decided to repeat the study, only now they added two conditions, one in which the experimenter pointed at the button, and another (to serve as a baseline) in which the experimenter showed no interest and simply sat reading a magazine. Adding the pointing behavior did bump up the percentage of times that the dogs helped the experimenter to over 50%.

Looking at their results the researchers wondered whether part of their problem was that the behaviors that the human used to show that she needed help opening the door was simply too stilted, because the nature of the communication had been so strictly scripted and did not take into account the reactions or responses of the dog, or even allow the use the dog's name. Perhaps if the communication to the dog was done in a more natural way the dogs would be more responsive. Also there was the possibility that the dogs would be more willing to help their owner rather than a stranger. So again the experiment was repeated. This time the human was allowed to do anything to make her goal as obvious as possible, including reaching for and pointing to the key, pushing the door, bending down to the dog and so forth. The person could talk to the dogs but only using the following sentences: “Open the door! Have a look! I want my key! Where is my key? How do I get there?” but  she could also say the name of the dog any number of times. However, the human was not allowed to use different phrases (such as a fetch command) and was also not allowed to point to the button. Half of the tests were conducted by the experimenter and half by the dog's owner.

Now the results were quite different. On around 90% of the trials the dog pushed the button to help the person. It made no difference whether the human was the experimenter or the dog's owner. This was equal to the number of times that the dog pushed the button to get food for themselves. This seems to show that dogs really do want to help, but first you have to get their attention by bending over and calling them by name and making it clear to them that they are needed and what they are needed for.

The researchers summarize their results by saying "In conclusion, dogs were highly motivated to help a human when the human’s goal was apparent by means of a communicative signal. Dogs’ difficulty in such situations seems to be in perceiving the human’s goal and knowing how to intervene, rather than in their willingness to help. The most effective way for a human to obtain help is to communicate with the dog in a natural way."

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, reposted without permission

More Posts