The Application Of Science In Dog Training

I see a lot of dog trainers claim they are using “science” when training dogs. Let’s examine that claim.

Science, in simple terms, is something testable and repeatable. Very few things dog trainers do have been tested in an experiment. 

The behavior sciences always start out with introspection, which is self-examination of how a person supposes this or that phenomenon works. This is often done because there isn’t any science, or the existing science seems to have problems, whether explaining the neural, behavioral, physiological, or learning processes being considered. I do this. I try to think how I’d feel and what I would do if this or that was to happen to me, or a dog I know. That doesn’t mean a dog would perceive or respond in the same way, but it is a start.

Next, comes the attempt to generalize concepts: this step is called reductionism. The often-unrecognized problem with generalized concepts is that they must ignore specific contradictions, especially in the real world. At this step, simple experiments are conducted to collect sufficient trends or correlations to then describe working definitions of terms and processes. For example, B. F. Skinner put rats and pigeons in boxes to push or peck on levers for rewards, often with lids on the boxes so the animals could not be observed, to primarily and tightly focus upon the mechanically produced generalized data and graphs which resulted. He then proposed a set of "laws". Reductionism cannot ever be successfully applied to explain everything, however, since the focus is on divining general trends or concepts.

This is the point where a lot of dog trainers mess up. They are working off these generalizations as if they are final, unchangeable truths, working in all situations. Scientists, on the other hand, if they are good scientists, don't see it that way. Instead, they go back and test these generalizations and argue amongst themselves. It is interesting to see these debates, between living scientists and those from long ago, each trying to persuade the readers that their discoveries are correct. The idea should be to test them from so many angles that these proposed discoveries are either validated or tossed away, or a new synthesis is tested, and a new generalization is able to become the latest “science”. It is completely possible that humans might never understand it all, which is why there is never a scientific consensus in real science.

So, this is where we find ourselves right now regarding dog training. 

There is still the need to use introspection to try looking at problems from 10,000 feet up. That’s very useful. Our human need to attempt to grasp the Gestalt, the whole, of a problem causes us to want this kind of viewpoint. Then, from the bottom up, also using reductionist forms of explanation, to see the microscopic parts, and then up the levels that make up the proposed whole. 

There is also a need to explain the emergent properties of behavior. Adding the parts together often doesn't explain the reality. Once all the parts are working together, more is revealed. Much of this next level theorizing comes from the field of ethology, which is a comparative study of homologies and analogies within related species and widely different species. This is the kind of work that was done by Jane Goodall (chimpanzees) or David Mech (wolves). These are observational studies of behavior, again trying to sort out what is seen into generalizations. Reductionism, again. And attempts to define systems in general terms.

Then, as this simmers on the stove, comes the application of all of this by modern dog trainers to real world dogs, following a similar path of discovery: 

Introspection: Depending upon their knowledge and experience, they are somewhere along this chain in terms of how they apply all of this. Some are operating solely through introspection, without much real data or experience to back up their impressions. I see that the realm in the so-called “dog whisperers.” What they think is going on probably isn’t backed up with any real science at all and never will be. On the other hand, advanced dog trainers use introspection to answer questions when there is no science. 

Experiments: Some trainers, from beginning to advanced, have read the scientific studies that have been done with dogs. Then they experiment with methods to find out what works. That is great, unless taken too far. I say that because science experiments are not the real world, and the real world can't be like a laboratory. What can be done in a laboratory isn’t necessarily going to happen with an off-leash dog in public or in someone’s home. If you take a study from a science experiment, and place too much faith in what the author claims, someone who never trained dogs, you are going to see it doesn’t work out as described in the study. Further, most of the scientific studies are poorly written, and oftentimes poorly controlled. If you have read enough scientific papers, you’ll come to realize that many of these scientists should have spent more time learning basic English (or whatever their native language happened to be) and trained real dogs. Furthermore, the translators were often not as proficient as they ideally should have been. Unfortunately, I don't believe you will find a scientific study on how to best train dogs to be companions, police dogs, hunting dogs, or anything else. In fact, most studies on dogs were done that way because there was no ethical way to do that same experiment on a human. So, the results are going to be tilted based upon the scientists’ motives and what they were looking for. Almost none of it was to benefit dog training.

Ethology And Observational Studies: Not a lot of this work has been done by scientists, or philosophers of science, on dogs. A lot of what is believed is folklore, or theories that have not been tested. While it may be interesting to compare wolf societies to street dogs in India, that isn’t going to help you all that much when training dogs. Yes, you will get some good ideas, but it won’t specifically indicate to you why a dog ate your mobile phone yesterday, or what to do about it. It is good to compare breeds, lines of dogs, and make comparisons to wild canines. You may find some gold nuggets along the way, but it isn't going to be handed to you.

Philosophy of Science: Much of what it takes to be a good dog trainer can't be tested with a scientific experiment. Right and wrong, morally, isn't to be found in a test tube. Many science experiments were too harsh and inhumane... nice to know, but don't do that again. Much of what is needed in the real world are sound methods that work, tempered with a philosophy of science that reflects logic and morals.

You, as a dog trainer, might have a good grasp of all the above but then applying it in some way is not going to be science and it is not going to be easy. This is where science and reality clash. This is where real world experience has to be tapped into, tempered with whatever science is probably applicable and a philosophy of not doing any harm yet doing things well. 

Some dog trainers are well studied, others are great at copying what others have taught them, and some should not be training dogs at all (or at least not yet). Some have detailed knowledge of the above, and many don't. I will say this, most of the scientists cannot train a dog, have never trained a dog, and possibly have no interest in training dogs. So, you must take their work for what it is… their interests in their ideas in their field of study. It's not about dog training. 

Dog trainers get frustrated when the media or politicians or activists claim “science” to justify this or that, when the dog trainers have personal experience proving the opposite… especially on the inevitable dog fads that come and go. Further, when there isn't any "science" to back up these popular fads.

I think the way a dog trainer can get some coherent way forward through all of this is to get their own dogs and train them; do volunteer dog rescue for several years; study the science; train dogs for other people; develop relationships with people who have more dog experience than you do so you can “talk dogs” into the night; go talk to some professors and scientists; and come up with an ethical and defendable philosophy about dog training.

So, is there science in dog training? Yes. Is there science in all dog training… not yet. And maybe never.

Plan accordingly.