�Cookie Trainer� is a derogatory term referring to people who train their dog purely using positive reinforcement. This means, when your dog does something right (or stops doing something wrong), you give him a treat to show him that he�s done something you want him to do again. In this post, I�m going to charitably reconstruct the point of view of a trainer who doesn�t use positive reinforcement, to try to understand why there is such derision of it as a method.
My reconstruction is an aggregate of various online encounters I�ve had with trainers that don�t use treats, to avoid naming names.
P+ Trainer:
�Well, I�m sure that for a lot of behaviors you can get a dog to do what you want with treats, but as soon as he knows you don�t have any treats, why would he bother doing what you want? He�s only good for as long as you�ve got the cookies, not because of anything else. My dog does what he�s told because he respects me, your dog just sees you as a Pez dispenser.�
As this is a charitable reconstruction, I�m not going to take issue with the truth of this statement right off the bat. Instead, I�m going to see whether there is any kind of insight behind it that we might learn from.
The Charitable Reconstruction
When it�s put in terms of the difference between a dog obeying out of desire for treats, and out of respect for his handler, we can see that there could be something valuable in the P+ trainer�s way of thinking. According to the P+ trainer, the R+ trainer reduces her relationship with her dog to a purely transactional one. The dog gets paid for working for the trainer. So long as the trainer has something the dog wants, the dog will obey because he knows that this is the best way to get it. The motivation, then, is an external one; the dog is being mercenary in his obedience.
The way that P+ trainer sees things, her dog is motivated to obey regardless of whether anything good will happen because her dog respects her. The nature of this relationship is not transactional and not mediated by an external motivator like a treat, it is founded on feelings that the dog has about the trainer and nothing else.
We do tend to see a relationship of respect as having more inherent value than a transactional relationship; it�s the ideal we strive for within our families and in the workplace. We want people to work for us because of us, not because of some resource we control. With dogs, who we see as ideally loyal, obedient animals that we enter into a relationship with when we train, it�s only natural to want the same kind of thing, and so to place little value on paying them off. A mercenary is never as trustworthy as a retainer.
So, if I�m right in my arguments, we can understand a little better what�s underlying the P+ trainer�s unwillingness to see R+ as equally valuable, because it speaks to a different kind of relationship between trainer and dog. Now I will explain what�s wrong with seeing things this way.
Dogs don�t understand high-level concepts like respect
The most compelling argument against the P+ trainer�s view is that dogs simply don�t work in the way she is describing. They�re not capable of processing abstract concepts like respect, guilt, shame or responsibility. These are concepts that we often attribute to dogs, but that science has shown us are just anthropomorphism at work (Horowitz 2009).
Current dog science shows us that, like all animals, dogs will repeat behaviors that get them what they want, and stop doing things that have bad consequences. This is the theory that underpins all dog training, not just R+; if we want a dog to come, we can teach it that coming means a treat, or we can teach it that coming means avoiding a shock. Both play on the dog�s basic motivation to promote its own wellbeing. Stephen Darwall explains respect in humans as one person recognizing the other as a legitimate source of authority. When we talk in terms of morality, we are talking about what kinds of demands we have the authority to make of one another, and how we can ensure that these demands are respected. Respecting authority, as opposed to deferring to power, comes from recognizing it as legitimate. Like the ability to comprehend guilt, respect would require the kind of abstract, moral thinking that dogs aren�t capable of. Dogs can understand power and control of resources, however, but this is a long way from the way trainers of both kinds use the principles of behaviorism to build obedience.
In short, the kind of respect that makes the P+ trainer think her relationship is more valuable doesn't exist in dogs. The only way to characterize a dog's motivation to perform specific "obedience" behaviors is to view it as part of a transactional relationship. This doesn't mean that your dog doesn't love you or that the relationship between dogs and humans has this quality alone - obedience is just one facet of life with a dog. It does mean that appeals to human concepts like respect in obedience isn't right.
In short, the kind of respect that makes the P+ trainer think her relationship is more valuable doesn't exist in dogs. The only way to characterize a dog's motivation to perform specific "obedience" behaviors is to view it as part of a transactional relationship. This doesn't mean that your dog doesn't love you or that the relationship between dogs and humans has this quality alone - obedience is just one facet of life with a dog. It does mean that appeals to human concepts like respect in obedience isn't right.
If dogs can�t be motivated by respect, then there is the question of what is behind their willingness to obey the P+ trainer. The answer to that, I believe, is fear of punishment.
Humans can�t easily differentiate between respect and fear
We often have a hard time working out exactly what motivates us to do things; sometimes it can come as a shock to realize that we�ve been acting out of jealousy, or the desire for validation. Even if we work to develop our faculties of empathy and introspection, we can still never be completely sure that we understand the motives of others. Respect and fear are one such complex. A child who talks politely to his father because he�s afraid of him might look the same to a stranger as a child who talks that way just because he respects his father. Similarly, we might see an obedient dog and assume that this evinces respect, but it could simply be fear of punishment. So, even if dogs could understand respect, we humans are not guaranteed to be able to distinguish it from fear.
In fact there have been studies that link increased levels of cortisol, the hormone associated with stress and fear, to dogs that have been trained with shock collars (Defra AW1402a). When the collar was put on the dogs, they began to show higher levels of cortisol, which suggests that they knew something bad was going to happen. Can we really say with any confidence that dogs trained to avoid shocks are doing so out of respect and not fear?
The underlying motivation for the P+ trainer�s criticism has been dealt with at this point. Although the P+ trainer was right to say that respect in humans is a more valuable basis for relationships than being paid is, this doesn�t apply to our dealings with dogs because dogs don�t work that way.
One consequence of seeing dogs as motivated by their own wellbeing and not by the idea of respect is, how do we work in the idea that dogs understand their relationships to the other animals in their life in terms of a pack hierarchy? The claim that dog owners have to make themselves their dog�s pack leader only really makes sense if we see the dog as necessarily able to respect the pack leader�s position in itself. If �pack leader� just means �person to be most afraid of�, then the terminology doesn�t add much to training at all; just ask the owner to use P+ to motivate the dog to be afraid of punishment, which will result in obedience. Where then the popular belief that dogs automatically respect their social betters? I�ll be answering this question in my next post.
References and Notes:
Defra AW1402 (2013) Studies to assess the effect of pet training aids, specifically remote static pulse systems, on the welfare of domestic dogs. University of Lincoln / University of Bristol / Food and Environment Research Agency. Final report prepared by Prof. Jonathan Cooper, Dr. Hannah Wright, Prof. Daniel Mills (University of Lincoln); Dr. Rachel Casey, Dr. Emily Blackwell (University of Bristol); Katja van Driel (Food and Environment Research Agency); Dr. Jeff Lines (Silsoe Livestock System).
Defra AW1402a (2013) Studies to assess the effect of pet training aids, specifically remote static pulse systems, on the welfare of domestic dogs; field study of dogs in training. Final report prepared by Prof. Jonathan Cooper, Dr. Nina Cracknell, Jessica Hardiman and Prof. Daniel Mills (University of Lincoln).
Darwall, S. (2006) The Second-Person Standpoint. MA:Harvard University Press.
Horowitz, A. (2009) Disambiguating the �guilty look�: Salient prompts to a familiar dog behavior. Behavioral Processes 81:3, pp. 447-452
This post has been concerned only with the ethical issues underlying P+ trainers� criticism of R+ training. For a demonstration of the practical misconceptions of my P+ trainer�s viewpoint, see the videos on this list: http://www.auf-den-hund-gekommen.net/-/Proof_Positive.html
For an introduction to the difference between positive and negative reinforcement (taken from a website about human behavior), see: http://bcotb.com/the-difference-between-positivenegative-reinforcement-and-positivenegative-punishment/
No comments:
Post a Comment