Parsimony and Why Its Important
During a recent presentation at a conference, we were discussing myths surrounding playing tug of war with dogs. We were explaining that tug-of-war is a cooperative game, not a competitive interaction over which individual is going to control the toy. Several people in the audience mentioned they had heard playing tug-of-war is a good way to build a dog’s “confidence.” What exactly does this mean?
One person described that he had been told to play tug to build the dog’s confidence in order to decrease his puppy’s submissive urination when people would try to pet her during greetings. If she was more confident, she would be less likely to submissively urinate. So, rather than having visitors reach out to pet the puppy, the owner instructed them to pick up a toy and play tug with her. Not surprisingly, this was quite effective in stopping the submissive urination.
The technique was effective because it changed the way visitors behaved toward the puppy. Rather than reaching out to pet her, and showing all those behaviors people think are friendly, but are often quite threatening to dogs, (making eye contact, facing them while leaning over them and reaching over their heads to pet them) visitors instead stimulated an alternative behavior – play – that was incompatible with submissive urination.
The dog owner, quite astutely, then commented that the technique worked, but NOT for the reasons he had been told it would work. He was exactly right. The simplest explanation for the effectiveness of the technique is that it removed the events that triggered the behavior (trying to pet the pup) and also engaged the dog in an incompatible behavior (it’s hard to urinate when you’re happily engaged in a game of tug).
Attributing the effectiveness of the procedure to building the dog’s “confidence” is a much more complicated explanation. What does “confidence” mean in terms of dog behavior? Why invoke a more complicated description of personality when a simpler explanation will do?
Using the simplest explanation to explain observations is called parsimony, and is an important, basic principle in the study of animal behavior. Also known as Occum’s Razor, parsimony dictates that it is not only unnecessary, but also inappropriate to choose a complicated explanation to explain observations of behavior, when a simpler one will do.
A number of years ago, Suzanne was a guest on the popular television series “Unsolved Mysteries”. This particular episode was all about animals and why they sometimes displayed “heroic” behavior and saved the lives of their owners or others. One example was a Newfoundland that saved a person from drowning whose boat had overturned in a river. The question put to Suzanne was how did the dog know that the person was drowning and needed saving? Her answer was – he didn’t.
The Newfie’s owner routinely took the dog to the river to swim and play, and the dog was in the habit of retrieving large logs that often floated down the river. In Suzanne’s view, at a distance, the drowning person resembled the large logs the dog had retrieved many, many times. So, the most parsimonious explanation for the dog’s behavior was that he was doing what he had always done – retrieve floating logs.
When we choose the simplest explanation for an animal’s behavior, not only are we more likely to be correct, it’s also more likely that we can take the appropriate steps to modify the animal’s behavior, should that be necessary.
4 Responses to “Parsimony and Why Its Important”
Suz and Dan
Thanks for your comment Didi. As we said in the original post, the point of the Unsolved Mysteries story was that the contention from the TV folks was that the dog knew the person was drowning and he set out to purposely rescue him. As you point out – that probably wasn't the case. In the end the person was rescued which is a good thing.
One of the reasons we think the idea of parsimony is important, because it's easy to anthropomorphise our pets' behaviors – believe they are acting "guilty" when we come home and find the trash overturned for example, when in reality they (dogs at least) are showing submissive behaviors. So doing our best to understand the "why" of behaviors will allow us to modify them, should that be our intent.
Didi
Parsimony is the key. We don't know what the dog thought but it's very unlikely that any companion dog would mistake any person for an inanimate object, even from a distance. It's also possible that the dog did not understand the whole situation….neither would a person happening upon the scene. But a companion animal who swims would very likely approach a person in the water whether the person was happily splashing or struggling. Dogs, as social/pack animals are very likely to note the difference from a distance between play and struggle but either way a socialized dog would probably investigate. Once close enough, a large swimming dog would easily serve as a flotation device. Doesn't change a thing. The dog did save a person even if it wasn't his big plan all along.
Suz and Dan
HI Tracy – thanks for offering an alternative opinion. While your interpretation of why the Newfie did what he did is certainly possible, we still maintain that our version is the most parsimonious. "Instinct" is in fact a complex term – even behaviorists use it to mean different things, we have no idea about how it actually works (e.g. what information about behavior is encoded in the genes), etc., etc. Just giving something a name by itself isn't an explanation.
However, a very basic principle is that the best predictor of future behavior is past behavior, and this dog had much more experience retrieving logs than he did rescuing people. So, in our view, he was doing what he'd already done many times before. It's certainly possible that the victim's behavior – waving arms, etc. – made him more a subject of attention for the dog. So it's possible – even probable – that more than one factor contributed to the dog's motivation.
Thanks for your comment!
Suzanne and Dan
Tracy
I totally agree with your analysis of the tug game and why it worked with the dog with the submissive urination problems.
However, in the case of the Newfoundland, I can't say that your floating log explanation is the simplest one. One of the original jobs of the Newfoundland was water rescue. They were selectively bred for these behavioral traits. Of course different individuals will exhibit these instincts to different degrees. I have been told by a friend who train her Newfie for water work that they are very sensitive to individuals in the water who are stressed. For example two humans were placed in the water, during a training exercise, and the dog was supposed to save the one who yelled, "Help, dog, help!". Many of the dogs would go to save the other "victim" if she was a person who was more stressed because she was rather afraid of the water (despite her life jacket). I think that instinct is a simpler explanation than generalizing the behavior of retrieving large logs to people since we know that dogs are pretty poor at generalization.