Do Dogs Respond to the “Human” Qualities of Spoken Commands?
There is something about human speech sounds that dogs automatically respond to.
Posted Aug 13, 2014
I was watching a beginner’s dog obedience class. It was the first day of training and the dogs were rather rambunctious. One woman had a young Labrador Retriever that she was having difficulty controlling. During the heeling exercises (where the dogs are supposed to walk at their owner’s left side) this big black pup kept running forward or backward along the mat to try to greet or play with the other dogs. The dog would hit the end of the leash and the woman would struggle to drag it back to her side.
“But,” the woman protested, “I’ve never taught him what that sound means.”
“Dogs don’t have to be taught things like that,” the instructor replied, “they automatically respond to our vocal tones the way that other people do.”
I remember that when I was younger most of the people who worked with dogs believed that canines only had a limited communication ability. Specifically, except for commands that the dog was trained to understand, it was felt that dogs did not really understand words. Rather it was felt that dogs had an ability to automatically interpret the emotional tone of a human who was speaking to them, and had a predisposition to respond to the emotional expression. More recently researchers shown that dogs can interpret the communication value of human gestures, and can even learn hundreds (and sometimes even over a thousand) human words. However now researchers have come to recognize that we really don’t know the true extent to which dogs understand human language patterns in the absence of explicit training. Perhaps the belief that dogs can extract the emotional content from our speech sounds is just another one of those myths.
Fortunately scientific research does have a habit of catching up and verifying or disproving some of our long-held beliefs about the behavior of dogs. In this case the newest research comes from a team of scientists led by Jennifer Gibson of Flagler College in Saint Augustine, Florida. This study was published in the journal Animal Behavior and Cognition.* The researchers chose to look at whether there is something in the way that humans usually speak that dogs particularly respond to. The specific aspect of speech that they focused on had to do with the scolding sounds which we use to prevent dogs from performing some unwanted behavior.
The research team began taping some naturalistic samples of human speech sounds. The human speaker stood in a hallway near a bowl of food. A dog was brought into the hallway and the task of the man or woman was to use whatever language and vocal commands that were needed to keep the dog from eating the food for 30 seconds. (At the end of that time the dog did get to eat the food.) This provided what we might call the “unscripted and authentic” speech sample. The speech sample was also transcribed and then read back word for word by an experimenter in a somewhat more artificial way (without all of the changes in tone that you might use if you had the dog in front of you and were trying to control it). We can call this the “scripted” speech sample. Other copies of the tapes were also modified to take out some of the human qualities in the speech. In the first of these the speech was garbled so that although it had the sound qualities, pitch, and rhythmic pacing of human speech, the words were no longer recognizable. You might remember the “Peanuts” cartoons of a number of years ago, where the teacher talks to Charlie Brown and all we hear is her saying “Wawa Wa Wawawa Wa” sounds, which are obviously human but are not intelligible. Another way in which the speech sample was distorted was to type the words into a computer-generated speech program which then generated a tape of a female voice pronouncing the words but removing the human idiosyncrasies and tone variations. Finally there was a control condition involving 30 seconds of silence.
The testing situation was quite simple. The dog was brought into a room where there was a bowl containing food. Nearby was a speaker through which were played the various speech samples. An experimenter stood nearby and the entire session was videotaped. The speech samples scolded the dog and told him not to eat the food. The measure of how effectively the dogs responded to the various speech samples was simply how much food the dog actually ate.
The data obtained showed that in the control condition (where there was no sound for 30 seconds) the dogs happily ate a lot of food. In the speech sections the dogs ate less than 1/5 the amount of food, suggesting that the taped sounds of the human voice were effective and interpreted appropriately by the dogs, regardless of whether the tape came from an authentic and unscripted performance or whether they were read back from a script. It also made no difference whether the speaker was male or female.
What about the garbled “Wawa” speech and the computer-generated speech? These were less effective in suppressing the dogs urge to eat, and the dogs ate more than twice as much food than in the real speech samples. However, these distorted speech sounds still had some effect on the dogs since the day only ate about 40% as much food as they did when there were no sounds played in the room. The researchers conclude “The current data suggests that there is a quality to human scolding vocalizations (humanness) that inspires greater compliance in dogs (less forbidden food consumption) when compared with other sounds or the presence of a human in the absence of sound.” In other words there is something about the natural sounds and tones of human speech that dogs are paying attention to without any specific training.
This seems to be consistent with what happened in the beginners training class that I was observing. When the owner produced the negative scolding sound “Uh uh!” the bouncy young Labrador Retriever did seem to recognize that he was doing something wrong and reduced his attempts to play with the other dogs — at least for a while.
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
* Gibson, J. M., Scavelli, S. A., Udell, C. J., & Udell, M. A. R. (2014). Domestic dogs (Canis lupus familiaris) are sensitive to the “human” qualities of vocal commands. Animal Behavior and Cognition, 1(3), 281-295. doi: 10.12966/abc.08.05.2014