Several people have asked me what I mean when I refer to behavior that is out of context for the animal. Often, but not always, this is in response to my observation that any time an animal accepts a reward for displaying such a behavior, that reward may be a punishment for the animal.
“How can accepting a reward be punishment?” they ask.
This is a legitimate question for those trained in behaviorism which maintains that animals learn in response to reward or punishment. As I’ve discussed previously, for the most part behaviorism is an anthropocentric system. People define what’s a reward and what’s a punishment based on laboratory experiments and personal philosophy. Traditionally electrical shocks were used for punishment because they were cheap and easy to control. Similarly, food was chosen for reward because, if a particular animal didn’t readily respond to food, you could withhold food before any training sessions to ensure a response. Or you could offer a more palatable kind of food. In another variation on this theme, toys or (often effusive) praise were deemed the ideal reward. But food also carries a very potent positive emotional charge for many people and this further reinforces the notion of food as the perfect reward.
This part of the process is based on the assumption that what constitutes a reward is a constant. That is, it’s not dependent on the context in which the reward is given.
Once teachers decide what constitutes an acceptable reward for the animal, then they’re ready to start teaching what they want the animal to know. This is another place where the out-of-context part comes in. There is nothing in this system that says, “Step 1: Before you start to teach an animal to do something, determine whether this is something the animal normally would do under these circumstances given this animal’s health, temperament, and relationship with the teacher.”
Quite the contrary. More often than not, we immediately dive into training whatever command appeals to us at that moment.
When the question of what it means when an animal refuses to obey a command is raised, sometimes a very strange form of logic occurs. While you and I might think that if an animal refuses to do something, there might be a problem with what we’re asking the animal to do, some people see resistance as a normal animal response. They expect the animal to resist, at least at first. This raises the question: Why do they believe or want to believe this?
One possibility is that they believe the average domestic animal has some innate desire not to please people. We might make a case for this for those individuals who have had wretched experiences with humans. But almost certainly this wouldn’t seem to be the case with the majority of domestic animals who, unless our human ancestors were extremely dim-witted or sadistic, have been bred for centuries to coexist with us in the most energy-efficient way.
But another reason why some might expect animals not to obey certain commands is because they intuitively recognize that the desired behaviors associated with those commands are out-of-context.
In both of these situations, teachers (be they owners or handlers of one sort or another) who accept this kind of thinking engage in power struggles with the animals, with the acceptance of the reward signaling the person’s success. When this is the case, an ever stranger kind of logic may arise that perceives the animal’s willingness to accept a reward for obeying an out-of-context command as proof of the teacher’s superior skill.
Let me give you an example here. Suppose one day I decide I need a dog for protection and that, of the two resident canines, 2-year-old Ollie is the best candidate. However, as these pictures of Ollie in his three stages of ferocity make clear, he hardly portrays an image of a vicious little dog willing to die to protect anything.
Realizing this, suppose I then decide to beef up his image by getting him a spiked leather harness as a joke and teaching him tricks to amuse myself and my friends. Alas, he naturally hates the weight and the way the harness restricts his ability to leap and run freely. But all is not lost. Ollie is one of those dogs about whom many would say “He’d do anything for food,” so I easily teach him to “accept” the harness in a few weeks using lots of treats. But let’s face it, as training goes that’s hardly show-stopping behavior, is it?
To address this problem, suppose that next I crouch down on the floor (while telling myself I get extra points for this because it’s not that easy given my knees) and try to lure him to over roll using treats. Only this time Ollie isn’t nearly so amenable. Even though he really wants that treat, he does not want to roll over to get it.
Now hopefully the reason why Ollie is conflicted about this is obvious: rolling over in that harness is painful. In order to accept the reward, he has to accept two things. One is the physical pain that rolling is going to cause him. The second and equally if not more troubling, he must accept that I want him to endure that pain to—what? Amuse me? Amaze my friends?
For me, the only way this works is if I’m willing to accept that Ollie is incapable of any kind of conscious thought, that there’s no such thing as behavioral pain, and that somehow—because he loves me and so wants to please me perhaps?—he will ignore the physical and mental pain that I’ve caused him in the process of getting him to do what I want. And when he grabs the treat and maybe even jumps up on me ecstatically after he completes the roll, I also must be willing to see this as evidence of how much fun he’s having and how much he loves and trusts me instead of a stress-based display in response to the conflicting situation I put him in to achieve my own goals.
Next let’s consider an out-of-context bond interaction. This time imagine that I afraid to be alone here in my little house on the hill and from the time he was a pup every time Ollie barked, my heart leaped into my throat and I otherwise reacted to his behavior. Because I felt insecure, I also reacted when he nudged, licked, barked at or otherwise demanded attention (be it positive or negative) from me. By accepting this, I communicated to him that he was in charge and that fine with me.
Fast forward to when this Ollie reaches adulthood and it dawns on him that this is an awfully big job for such a little dog. To compensate for that, he develops an attitude. Instead of just barking at visitors, he now charges them snapping and snarling. Like any good owner, I shout, “Ollie, down!” and Ollie drops like a rock because using the aforementioned out-of-context method, I’ve taught him to do exactly that.
At this point many trainers might say that Ollie is a really well-trained dog and I did a great job of training him. But the reality is that I did a great job of putting him in a hell of a dilemma once again. I forced him to choose between protecting me, his most valuable position in the whole world, and obeying me. The pain associated with making such a choice surely equals that of rolling over on spikes.
But if I were able to ignore all this, then the sky would be the limit for Ollie and me. Or me at least. As long as I could get him to overcome any physical and/or behavioral pain caused by my out-of-context demands and accept the reward, I could get him to do even more elaborate tricks. And the more out-of-context these were, the greater the trainer I would be.
As you can see, the success of my imaginary system depends on not knowing anything about ethology or animal behavior as it relates to the animal, or about the physiological and behavioral effects of the human-animal bond. Once that awareness exists, though, then like it or not, I have to look at this previously innocuous or even “positive” human-animal interaction in a whole new light…
Do I think that it’s wrong to teach dogs commands or tricks? Most certainly not. Dogs need structure and limits just like kids, and learning and performing tricks can be fun for them as well as their owners. I just think we need to consider whether any behavior we’re thinking about teaching an animal is sound given that animal’s health, temperament and relationship with us.
I see no logical reason, for example, to force an animal who is obviously terrorized by cars or other dogs to stand or sit motionless and endure that fear, and then reward the animal for doing this. To me this is like shoving a little kid who’s afraid of water into the deep end of the pool. It doesn’t matter if we wave treats or shout all kinds of encouragement after-the-fact. The damage to the relationship has already been done. By putting animals in this position in the first place, we communicate that we can’t be trusted to keep them them safe, just as I communicated to the fantasy Ollie when I forced him to roll over in that painful harness. If we’ve simultaneously communicated via a reactive relationship that we expect the animal to take care of us, we double the animal’s burden at that time. If the animal accepts a reward for surviving all that, well, it’s one of the reasons why I refer to this as the rub-their-noses-in-it approach to training.
Compare this approach to building the trust between human and animal in a secure environment with minimal distraction to ensure maximum learning. These animals first learn to trust and take their cues from their owner’s behavior. If their owners are calm, they remain calm. They learn self-control and their correct responses are self-rewarding. They obey because it makes them feel good about themselves first, and the person who gave them this opportunity second.
But maybe that’s what sometimes makes it so hard to create learning situations that are in context for animals. Maybe in spite of what we say, the fact that they’re well-behaved isn’t enough. Maybe the real problem is that our level of self-confidence is so low that we need constant proof of their willingness to obey.
Several people have asked me what I mean when I refer to behavior that is out of context for the animal. Often, but not always, this is in response to my observation that any time an animal accepts a reward for displaying such a behavior, that reward may be a punishment for the animal.
“How can accepting a reward be punishment?” they ask.
This is a legitimate question for those trained in behaviorism which maintains that animals learn in response to reward or punishment. As I’ve discussed previously, for the most part behaviorism is an anthropocentric system. People define what’s a reward and what’s a punishment based on laboratory experiments and personal philosophy. Traditionally electrical shocks were used for punishment because they were cheap and easy to control. Similarly, food was chosen for reward because, if a particular animal didn’t readily respond to food, you could withhold food before any training sessions to ensure a response. Or you could offer a more palatable kind of food. In another variation on this theme, toys or (often effusive) praise were deemed the ideal reward. But food also carries a very potent positive emotional charge for many people and this further reinforces the notion of food as the perfect reward.
This part of the process is based on the assumption that what constitutes a reward is a constant. That is, it’s not dependent on the context in which the reward is given.
Once teachers decide what constitutes an acceptable reward for the animal, then they’re ready to start teaching what they want the animal to know. This is another place where the out-of-context part comes in. There is nothing in this system that says, “Step 1: Before you start to teach an animal to do something, determine whether this is something the animal normally would do under these circumstances given this animal’s health, temperament, and relationship with the teacher.”
Quite the contrary. More often than not, we immediately dive into training whatever command appeals to us at that moment.
When the question of what it means when an animal refuses to obey a command is raised, sometimes a very strange form of logic occurs. While you and I might think that if an animal refuses to do something, there might be a problem with what we’re asking the animal to do, some people see resistance as a normal animal response. They expect the animal to resist, at least at first. This raises the question: Why do they believe or want to believe this?
One possibility is that they believe the average domestic animal has some innate desire not to please people. We might make a case for this for those individuals who have had wretched experiences with humans. But almost certainly this wouldn’t seem to be the case with the majority of domestic animals who, unless our human ancestors were extremely dim-witted or sadistic, have been bred for centuries to coexist with us in the most energy-efficient way.
But another reason why some might expect animals not to obey certain commands is because they intuitively recognize that the desired behaviors associated with those commands are out-of-context.
In both of these situations, teachers (be they owners or handlers of one sort or another) who accept this kind of thinking engage in power struggles with the animals, with the acceptance of the reward signaling the person’s success. When this is the case, an ever stranger kind of logic may arise that perceives the animal’s willingness to accept a reward for obeying an out-of-context command as proof of the teacher’s superior skill.
Let me give you an example here. Suppose one day I decide I need a dog for protection and that, of the two resident canines, 2-year-old Ollie is the best candidate. However, as these pictures of Ollie in his three stages of ferocity make clear, he hardly portrays an image of a vicious little dog willing to die to protect anything.
To address this problem, suppose that next I crouch down on the floor (while telling myself I get extra points for this because it’s not that easy given my knees) and try to lure him to over roll using treats. Only this time Ollie isn’t nearly so amenable. Even though he really wants that treat, he does not want to roll over to get it.
Now hopefully the reason why Ollie is conflicted about this is obvious: rolling over in that harness is painful. In order to accept the reward, he has to accept two things. One is the physical pain that rolling is going to cause him. The second and equally if not more troubling, he must accept that I want him to endure that pain to—what? Amuse me? Amaze my friends?
For me, the only way this works is if I’m willing to accept that Ollie is incapable of any kind of conscious thought, that there’s no such thing as behavioral pain, and that somehow—because he loves me and so wants to please me perhaps?—he will ignore the physical and mental pain that I’ve caused him in the process of getting him to do what I want. And when he grabs the treat and maybe even jumps up on me ecstatically after he completes the roll, I also must be willing to see this as evidence of how much fun he’s having and how much he loves and trusts me instead of a stress-based display in response to the conflicting situation I put him in to achieve my own goals.
Next let’s consider an out-of-context bond interaction. This time imagine that I afraid to be alone here in my little house on the hill and from the time he was a pup every time Ollie barked, my heart leaped into my throat and I otherwise reacted to his behavior. Because I felt insecure, I also reacted when he nudged, licked, barked at or otherwise demanded attention (be it positive or negative) from me. By accepting this, I communicated to him that he was in charge and that fine with me.
Fast forward to when this Ollie reaches adulthood and it dawns on him that this is an awfully big job for such a little dog. To compensate for that, he develops an attitude. Instead of just barking at visitors, he now charges them snapping and snarling. Like any good owner, I shout, “Ollie, down!” and Ollie drops like a rock because using the aforementioned out-of-context method, I’ve taught him to do exactly that.
At this point many trainers might say that Ollie is a really well-trained dog and I did a great job of training him. But the reality is that I did a great job of putting him in a hell of a dilemma once again. I forced him to choose between protecting me, his most valuable position in the whole world, and obeying me. The pain associated with making such a choice surely equals that of rolling over on spikes.
But if I were able to ignore all this, then the sky would be the limit for Ollie and me. Or me at least. As long as I could get him to overcome any physical and/or behavioral pain caused by my out-of-context demands and accept the reward, I could get him to do even more elaborate tricks. And the more out-of-context these were, the greater the trainer I would be.
As you can see, the success of my imaginary system depends on not knowing anything about ethology or animal behavior as it relates to the animal, or about the physiological and behavioral effects of the human-animal bond. Once that awareness exists, though, then like it or not, I have to look at this previously innocuous or even “positive” human-animal interaction in a whole new light…
Do I think that it’s wrong to teach dogs commands or tricks? Most certainly not. Dogs need structure and limits just like kids, and learning and performing tricks can be fun for them as well as their owners. I just think we need to consider whether any behavior we’re thinking about teaching an animal is sound given that animal’s health, temperament and relationship with us.
I see no logical reason, for example, to force an animal who is obviously terrorized by cars or other dogs to stand or sit motionless and endure that fear, and then reward the animal for doing this. To me this is like shoving a little kid who’s afraid of water into the deep end of the pool. It doesn’t matter if we wave treats or shout all kinds of encouragement after-the-fact. The damage to the relationship has already been done. By putting animals in this position in the first place, we communicate that we can’t be trusted to keep them them safe, just as I communicated to the fantasy Ollie when I forced him to roll over in that painful harness. If we’ve simultaneously communicated via a reactive relationship that we expect the animal to take care of us, we double the animal’s burden at that time. If the animal accepts a reward for surviving all that, well, it’s one of the reasons why I refer to this as the rub-their-noses-in-it approach to training.
Compare this approach to building the trust between human and animal in a secure environment with minimal distraction to ensure maximum learning. These animals first learn to trust and take their cues from their owner’s behavior. If their owners are calm, they remain calm. They learn self-control and their correct responses are self-rewarding. They obey because it makes them feel good about themselves first, and the person who gave them this opportunity second.
But maybe that’s what sometimes makes it so hard to create learning situations that are in context for animals. Maybe in spite of what we say, the fact that they’re well-behaved isn’t enough. Maybe the real problem is that our level of self-confidence is so low that we need constant proof of their willingness to obey.