DIG INTO OUR BLOG


Strictly “Positive Reinforcement” Dog Training

by Kim Brophey on 06/16/15


Alexandre Rossi, a colleague of mine in Brazil, asked me to weigh in on the concept of 

strictly "positive reinforcement" dog training:

The dog training and behavior field is, in my opinion, overly concerned with only one aspect of the greatly larger picture.  Given that the field has been slow to integrate and apply the various sciences (applied ethology, neuro-ethology, neuro-biology, epi-genetics, etc.) that can inform and explain the behavior of modern pet dogs, we continue to repeat the political conversations about strict behaviorism as if this discipline of learning theory was the only viable consideration regarding the ethics and effectiveness of dog training practices. I personally would like to expand the scope of the conversation in the field and raise the bar on our ability to discuss the bigger picture of welfare in learning as it effects our canine companions.  

I think we can't have this conversation about strictly "positive" reinforcement training vs. any use of "aversives" without also stating our great

negligence in considering factors like varying learning needs, orientations, and limitations in different kinds of environmental conditions or genetic profiles in dogs, etc. (i.e. trying to use "positive reinforcement with clicker and treats" to teach the hound to ignore the deluge of wild animal scents on the daily walks and focus on the handler might have certain inherent problems as the tracking behavior/"seeking" system event in the dog's senses and brain could easily trump the competing motivation for our "positive reinforcement" snackies and squeakers).  

We can't, I feel quite strongly, keep having these conversations about the "best" training practices until we examine fundamental questions about our failure to recognize how we have accidentally punished, reinforced, or totally ignored a plethera of behaviors in our dogs (often failing to affect them at all) because our debate has been so limited in scope.

To humor the simplicity of the idea of "positive" vs. "aversive" training, however, I will say that the behavorist in me is continually shocked by how often dogs are punished with "rewards", reinforced with "punishment", and how random the whole conversation about dog training has become.

The applied ethologist in me finds it kind of amusing how off the mark of common sense we are.  Nature seems to say it best, when we stop and just look at it from an evolutionary perspective.  I try to mimic nature's genius in everything I do with dogs, to be maximally effective and minimally intrusive.

Nature - survival of the fittest- is about doing what works and dropping what doesn't work. That economy of strategy is the name of the game.  There's a motivation in any given moment, a contextual set of conditions that excite a behavior which answers that motivation to optimal ends for the individual.  The most valuable and underused concept of learning in dog training is, in my opinion, the Premack Principle which so directly recognizes this fundamental question of "How can I....?" that the dog has in every moment of training/learning.  The basic math is that there is a motivation that underlies every behavior.  If we identify and exploit that motivation and use it as the "reward" for the desired target behavior, we are most effectively training the dog.  But we don't do that much - we throw OTHER motivations into the mathematical equation and complicate the heck out of it for the dog.  The dog wants to chase the squirrel- we say "here's a treat if you don't seek the squirrel" or "here's a leash pop if you do chase the squirrel".  The dog says, "Yeah, but what about the squirrel???  How can I....chase the squirrel?"  We never pocket the motivation because we don't get to this practical equation of learning -- motivation towards behavior towards the ends.  

My concern, when it comes to "corrections" or "rewards" in dog training is do we recognize how totally arbitrary our efforts so often are to the dog?  I think that we should all have a healthy appreciation for the basic economy of learning, and know how to do a basic cost-benefit analysis of a training technique.  If I understand learning (both classical and operant) sufficiently, then I should first and foremost be making wise choices about my equipment and practices that recognize certain higher risks of backfire associated with any tool that inflicts pain, causes fear, uses force, floods the dog, etc.  I should also appreciate that I can stress the heck out of a dog by asking for an arbitrary behavior that the dog doesn't understand the reason for (no matter how many treats I use).  I have seen dogs that are absolutely "punished" and shut down in response to free shaping with a clicker to perform a trick (a friend of mine dropped out of clicker academy because her dog became so highly stressed from the exercises required to graduate).  I have also seen dogs that were "reinforced" by aversives as part of a stimulus/behavior chain that ultimately "worked" for the dog.  What we so often fail to do is to think critically and trouble shoot intelligently every dog in every moment, ready with creativity and unceasing compassion for dog and their person at all times to change the course of action in training based on what is ACTUALLY HAPPENING. 

I think we have an obligation, I feel, to become far more educated and far less political when it comes to training and behavior in pet dogs. We need to see learning in the greater context of life, and learn how to design a dog's training protocol by setting up scenarios in which target behaviors "work" for the dog and problem behaviors "fail" for the dog when it comes to the squirrel, the attention, the leftovers, and everything else.  Training comes down to being an amazing manager, able to control the variables that matter to the dog in a given moment so that behavioral change is self-inspired for the dog on a practical level.  I don't personally like the carrot OR the stick.  My goal is to identify what matters to the dog in a moment, and be savvy enough to turn that competing agenda into my greatest aid whenever possible.

Let's change the conversation, then.  We should never be cruel and put fear into any dog or person.  We shouldn't create monsters with permissiveness either because we fail to exact a logical consequence that renders a behavior ineffective for a dog- we burden our dog with the stress of crappy upper management when we do.  We should understand PREMACK and work with the basic math of the moment -- the condition that excites the strategy in efforts towards the motivation's result.  In the ABC of behavior, we should try to change the B, rather than put so much effort into trying to convince the dog to forget about C completely and want D or E or F instead.  In this A moment, the dog wants the C.  That's why it does behavior B.  We keep ignoring the A, telling the dog to drop the C, and just to do behavior B.  Really?  Our task is to show them that the B is the ticket from point A to point C.  This failure to see our task is the basic problem.  We can abuse our dog with punishers AND rewards indirectly or directly when we don't understand what is actually happening.  We need to raise the bar- again- on training discussions and start from there to move forward. 

 Hope that helps!

 Kim Brophey


Leave a comment









Recent Comments

Online Booking

If you are already a client you can book an Appointment online!

Are you ready to book an appointment in our behavior center? If you are a new client please call our behavior center at 828. 656. 8305 to schedule your first appointment. If you are currently a client our online scheduler makes it simple to check trainer schedules, view availability, and book your next appointment in advance. Go there now to get started!

Schedule Appointment