Kijun Times

I had just finished delivering a lecture on the topic of how dogs communicate and afterwards several people came up to offer some observations or ask me some follow-up questions. One woman commented in a slightly frustrated tone of voice, "I know that you scientists have analyzed the sound information in a dog's bark, but it doesn't do much good for us common dog owners who don't have the time to carefully study barking sounds so that we can figure out what our own dog is trying to tell us."

I have heard this comment before, so I was pleased to be able to reassure her that although a lot of analysis has gone into understanding canine communication (click here to read more about interpreting dog barks) no special course in canine linguistics was necessary for her to be able to extract the basic emotional information from her dog's barking. In fact, evidence shows that people with very little experience with dogs are just as accurate as more experienced dog owners when interpreting the emotional state of a barking dog.

A very nice demonstration of this comes from a study by Péter Pongrácz, Csaba Molnár and Ádám Miklósi of the Department of Ethology at Eötvös Loránd University in Budapest, Hungary. The report was published in the journal Applied Animal Behaviour Science*. First they had to gather a large sample of dog barks. The breed that they chose was the Mudi which is a Hungarian herding dog used with sheep and cattle and which can also be a vigilant watchdog. The working style of this breed is characterized by the extensive use of barking.

Recordings from 15 Mudis were collected in six different behavioral settings. One involved the approach of a stranger who arrived at the garden gate or the front door of the apartment while the owner was not nearby. This is a situation which might provoke an aggressive or possibly fearful response. A purely aggressive set of barks were gotten during Schutzhund training, where the dog is being encouraged to bark aggressively and bite a protective glove on the trainer's arm. Excited barks were recorded when the owner behaved as if you are preparing to go for a walk with the dog. More despairing barks were obtained when the owner tied the leash of the dog to a tree in a Park and walked away out of sight of the dog. Happy barks were obtained when the owner held a ball or some favorite toy in front of the dog, while another set were gotten when the owner was actually playing with the dog.

Sound samples were carefully collected and codified. The selection of barks of different pitches, ranging from a low to high (based on the values of the peak frequency in each bark). Also barks were classified by tonality (technically the harmonic-to-noise ratio) which is a measure of the "roughness" of the barking. The third dimension was the average inter-bark interval which is just the time spacing between one bark and the next (think of this as how fast or slowly the dog is barking). In the end the researchers assembled 27 different bark sequences which contained a mixture of the various possibilities from the three sound dimensions.

Ultimately observers were asked to rate each bark on the basis of five possible emotional states: aggressiveness, fear, despair, playfulness, and happiness. In order to determine whether experience with dogs was playing a role in the identification of the emotional content of barking, three groups of people were used. The people who you might expect to be most experienced are individuals who have owned and lived with Mudis. A somewhat less experienced group was comprised of people who were familiar with and have lived with dogs other than Mudis. The least experienced group would be made up of people who did not own or live with dogs.

There was a lot of consistency in the way that people judged the emotionality of dog barks. Low pitched barks were universally described as aggressive, while tonal and high pitched barks were viewed as being either fearful or desperate, but always without any aggressiveness. However the pitch of the bark (high versus low) had a much greater effect than the roughness or tonality. The inter-bark interval also had a strong effect on how the human listeners interpreted the emotionality of the barks. Rapid barking sequences (short inter-bark intervals) were scored as aggressive when compared to slower barking sequences (with longer inter-bark intervals) which were generally perceived as signaling little or no aggression. High-pitched bark sequences with long inter-bark intervals were considered to be happy and playful (regardless of what their tonality might be) however if the inter-bark interval was long enough and there was little roughness in the sound, these could also be interpreted as despair.

Remember that one of questions that the researchers were interested in was the effect of experience on the interpretation of the emotionality of barks in dogs. Basically only very minor differences were found as a result of how experienced people were with dogs. Experience had no significant effect in the case of "despair" and "happiness". The only differences that emerged was that "non-owners" were less likely to interpret positive emotions from the dog barks. So they gave the highest scores for "fear" and the lowest for "playfulness". But these effects were relatively small.

In other words, it seems that regardless of people's experience with dogs, there is a uniformity in how humans interpret the emotional state of a dog based upon the sound of its barking. We apparently don't have to formally learn the rules to sense whether a dog is being aggressive or fearful or is in a more positive emotional state.

It is not fully clear as to why we can so easily and consistently interpret the emotions behind a particular pattern of dog barks. One guess is that all mammals make sounds that vary along the same dimensions, and we are prewired (through our DNA) to interpret the emotional content of those sounds when we hear them. An alternative comes from the fact that domestic dogs bark a lot more than wild canines. This is part of what made dogs valuable to early humans, since the barking of a dog could serve as a signal that perhaps a hostile stranger or a threatening animal was approaching. Of course, it would be useful for people to recognize the barks of their dog so that they could know whether the dog was signaling that it had sensed some possible danger, as opposed to the dog indicating that it was excited and happy, perhaps at the arrival of a familiar and friendly person. So it may well be the case that during the process of domestication the dogs which gave the most interpretable barks, even to naïve and inexperienced listeners, were selected for. Thus we may have systematically bred dogs which communicate their emotions to us most clearly through their barking.