The nice thing about being an independent scholar is that I’m not bound by the culture and limits of a particular discipline. Because of this I’m free to think about a lot of subjects that would be unthinkable to others. Add that I’m blessed with friends and colleagues with whom I can freely share and further explore such thoughts and sometimes the results can be shocking and even blasphemous.
This is what happened when I read a news item about the release of a remake of the film Free Willy, a warm, fuzzy story about a cute little kid’s efforts to free a killer whale from captivity. Once I recovered from the irony of this in light of the fatal captive orca attack on a trainer not that long ago, my thoughts naturally went to how we train wild animals to perform yet again.
As with domestic animals, the bulk of training methods used on wild animals (either to facilitate handling in zoos or for media purposes) falls into one of two basic categories: reward or punishment. For obvious reasons, reward-based training is considered more humane and has increasingly replaced punishment. But even though most people would agree that we can, as the old saying goes, catch more flies with honey than vinegar, reward-based training isn’t without its problems. (Similar problematic corollaries exist with punishment-based training but I’ll leave that for another time.)
A few days after I heard about the Free Willy remake, a friend recounted one of those near-miss stories that are familiar to those who work with captive wild animals. In this not-that-unique case a frustrated or angry trainer hit a disobedient animal, another marine mammal. Later, the animal herded the trainer into deeper water away from help and tried to drown him.
Among the general public and the reward-based training community, the immediate, conditioned response is typically, “The trainer deserved it. It only happened because he hurt the animal.” But as I thought about this as well as the other near-misses I’ve read or heard about in both wild and domestic animals, I had to admit that this was probably wishful thinking, at least in some of these cases. Given the size, strength, and speed difference between the humans and animals involved, the negative response of these animals often exceeded the negative human act that precipitated it. The most common response of those who observed or investigate such tragedies is that, while the person may have acted foolishly, the punishment far exceeded the crime. While some might say this proves that the animal had some organic problem—such as an undiagnosed disease or another of those behaviorally convenient brain tumors—my thoughts turned in a more blasphemous direction.
In previous commentaries and podcasts, I’ve discussed what I consider three problematic assumptions associated with reward-based training. One is the assumption that what’s a reward for us is automatically a reward for the animal. The second is that we’re rewarding the animal for whatever the animal did immediately prior to receiving and accepting the reward. That is, the animal doesn’t equate the reward with something other than the reason for which we give it. And the third is that accepting the reward signals the end of a command-response-reward sequence that is a positive experience for the animal.
Now the blasphemous part. It’s possible that the failure to acknowledge the fragility of these assumptions in our interactions with certain animals is what leads to tragic results in so-trained animal encounters.
If you feel your blood pressure rising, please stop and take a few deep breaths. This is not meant to be an attack on a training approach that in some circles ranks as scared as motherhood, apple pie, and the flag. I’m just inviting you to follow me through the thought process I followed in my attempt to understand why these attacks happen. Without such understanding, how can we hope to prevent them in the future?
All-positive training systems are fine as long as the animal never does anything wrong. And indeed, if the trainer or owner is quick enough and knowledgeable enough about a particular animal’s behavior to stay one step ahead of him or her, it’s most certainly preferable to using punishment. But alas doing so requires a lot of knowledge of that animal’s species and individual needs as well as commitment. Consequently, sometimes animal do misbehave. When this occurs, the most common worst consequence to the animal is that the food reward is withheld.
Now things start to get dicey. Because many times food is withheld before training or a performance to increase the animal’s response to it, withholding a reward may not be that much of a punishment in the overall scheme of things. And if accepting the reward for performing out-of-context behaviors has been punishment for the animal right along, not getting such a “reward” for a misbehavior the animal finds more gratifying in some way wouldn’t be much of a loss, either.
Relative to the trainer, because that person can’t punish the animal, any misbehavior must be tolerated and even rationalized in some way. (“It was an accident. He didn’t realize how hard he hit me.” “I don’t think she was feeling well.”) It goes without saying that if accepting the display makes the person feel vulnerable in any way, this will undermine the human-animal relationship from then on. If trainers blame the animal’s misbehavior on themselves, as some misguided animal-lovers insist they always must, the trainers may increase these stressful, conflict-generating reward-based interactions with the animal in the naïve belief that this will decrease the animal’s stress. However, in some cases doing so could make a bad situation worse.
Given this scenario, the ultimate blasphemy regarding such a training system brings me to a question that probably has occurred to you by this now: From the animal’s point of view, what’s the worst thing that can possibly happen to a so-trained animal if/when human behaviors exceed the animal’s limits of tolerance?
If we analyze this in terms of cost-benefit to the animal, based on past experience the most the animal has to lose is a reward that never may have been that much of a reward from the get-go. If the reward was actually a punishment, we would have an even more dangerous situation, summed up in another old saying, “Might as well be hung for a sheep as a lamb.” This refers to a legal system in which the punishment is the same for greater and lesser crimes. In that case, there’s no reason not to go after the greater prize when tempted.
For animals so pushed, the reward would be the release of physical and behavioral stress and frustration that may have been building up for weeks or longer before finally reaching the tipping point. Given the negative effects of chronic elevations in stress hormones, it doesn’t strike me as unreasonable that some animals in this position might opt for violence. This also would explain why sometimes there appears to be no connection between anything that people believe to be physically or behaviorally painful that was deliberately or unwitting done to the animal and the attack, whereas other times the attack is triggered by something so seemingly inconsequential that it’s dismissed as insignificant. (“She always plays a game where she teases him with his toys after a training session so that couldn’t be a problem.”)
That’s a frightening thought isn’t it? That in the process of imposing what we consider a most humane and even loving teaching system on animals, we may be driving some of them to violent behavior. It’s even more frightening to consider that similar breakdowns in human-animal communication may also lead to similar tragedies in our interactions with domestic and companion animals.
But unsettling as such thoughts may be, for me considering them is preferable to stuffing my head up my most posterior body orifice in an attempt to avoid them. Time and time again when these tragedies occur, somebody says, “We need to look at the way we’re training these animals.” And time and time again, these elicit a kill-the-messenger response. After all, we tell ourselves, what could be more positive than all-positive training? What animals in their right minds wouldn’t respond to such a system? The ones that don’t? Well, they’re obviously not in their right minds.
I wouldn’t presume to tell anyone how to properly train whales, dolphins, elephants or any other wild animal. But I do know that the same erroneous assumptions that underlie reward (and punishment) training in wild animals also are contributing factors in some cases of domestic animal-human violence. As long as we continue to deny this, people will continue to be hurt, and animals will continue to die because of it.
If you have any comments regarding subject matter, favorite links, or anything you’d like to see discussed on or added to this site, please let me know at mm@mmilani.com.
The nice thing about being an independent scholar is that I’m not bound by the culture and limits of a particular discipline. Because of this I’m free to think about a lot of subjects that would be unthinkable to others. Add that I’m blessed with friends and colleagues with whom I can freely share and further explore such thoughts and sometimes the results can be shocking and even blasphemous.
This is what happened when I read a news item about the release of a remake of the film Free Willy, a warm, fuzzy story about a cute little kid’s efforts to free a killer whale from captivity. Once I recovered from the irony of this in light of the fatal captive orca attack on a trainer not that long ago, my thoughts naturally went to how we train wild animals to perform yet again.
As with domestic animals, the bulk of training methods used on wild animals (either to facilitate handling in zoos or for media purposes) falls into one of two basic categories: reward or punishment. For obvious reasons, reward-based training is considered more humane and has increasingly replaced punishment. But even though most people would agree that we can, as the old saying goes, catch more flies with honey than vinegar, reward-based training isn’t without its problems. (Similar problematic corollaries exist with punishment-based training but I’ll leave that for another time.)
A few days after I heard about the Free Willy remake, a friend recounted one of those near-miss stories that are familiar to those who work with captive wild animals. In this not-that-unique case a frustrated or angry trainer hit a disobedient animal, another marine mammal. Later, the animal herded the trainer into deeper water away from help and tried to drown him.
Among the general public and the reward-based training community, the immediate, conditioned response is typically, “The trainer deserved it. It only happened because he hurt the animal.” But as I thought about this as well as the other near-misses I’ve read or heard about in both wild and domestic animals, I had to admit that this was probably wishful thinking, at least in some of these cases. Given the size, strength, and speed difference between the humans and animals involved, the negative response of these animals often exceeded the negative human act that precipitated it. The most common response of those who observed or investigate such tragedies is that, while the person may have acted foolishly, the punishment far exceeded the crime. While some might say this proves that the animal had some organic problem—such as an undiagnosed disease or another of those behaviorally convenient brain tumors—my thoughts turned in a more blasphemous direction.
In previous commentaries and podcasts, I’ve discussed what I consider three problematic assumptions associated with reward-based training. One is the assumption that what’s a reward for us is automatically a reward for the animal. The second is that we’re rewarding the animal for whatever the animal did immediately prior to receiving and accepting the reward. That is, the animal doesn’t equate the reward with something other than the reason for which we give it. And the third is that accepting the reward signals the end of a command-response-reward sequence that is a positive experience for the animal.
Now the blasphemous part. It’s possible that the failure to acknowledge the fragility of these assumptions in our interactions with certain animals is what leads to tragic results in so-trained animal encounters.
If you feel your blood pressure rising, please stop and take a few deep breaths. This is not meant to be an attack on a training approach that in some circles ranks as scared as motherhood, apple pie, and the flag. I’m just inviting you to follow me through the thought process I followed in my attempt to understand why these attacks happen. Without such understanding, how can we hope to prevent them in the future?
All-positive training systems are fine as long as the animal never does anything wrong. And indeed, if the trainer or owner is quick enough and knowledgeable enough about a particular animal’s behavior to stay one step ahead of him or her, it’s most certainly preferable to using punishment. But alas doing so requires a lot of knowledge of that animal’s species and individual needs as well as commitment. Consequently, sometimes animal do misbehave. When this occurs, the most common worst consequence to the animal is that the food reward is withheld.
Now things start to get dicey. Because many times food is withheld before training or a performance to increase the animal’s response to it, withholding a reward may not be that much of a punishment in the overall scheme of things. And if accepting the reward for performing out-of-context behaviors has been punishment for the animal right along, not getting such a “reward” for a misbehavior the animal finds more gratifying in some way wouldn’t be much of a loss, either.
Relative to the trainer, because that person can’t punish the animal, any misbehavior must be tolerated and even rationalized in some way. (“It was an accident. He didn’t realize how hard he hit me.” “I don’t think she was feeling well.”) It goes without saying that if accepting the display makes the person feel vulnerable in any way, this will undermine the human-animal relationship from then on. If trainers blame the animal’s misbehavior on themselves, as some misguided animal-lovers insist they always must, the trainers may increase these stressful, conflict-generating reward-based interactions with the animal in the naïve belief that this will decrease the animal’s stress. However, in some cases doing so could make a bad situation worse.
Given this scenario, the ultimate blasphemy regarding such a training system brings me to a question that probably has occurred to you by this now: From the animal’s point of view, what’s the worst thing that can possibly happen to a so-trained animal if/when human behaviors exceed the animal’s limits of tolerance?
If we analyze this in terms of cost-benefit to the animal, based on past experience the most the animal has to lose is a reward that never may have been that much of a reward from the get-go. If the reward was actually a punishment, we would have an even more dangerous situation, summed up in another old saying, “Might as well be hung for a sheep as a lamb.” This refers to a legal system in which the punishment is the same for greater and lesser crimes. In that case, there’s no reason not to go after the greater prize when tempted.
For animals so pushed, the reward would be the release of physical and behavioral stress and frustration that may have been building up for weeks or longer before finally reaching the tipping point. Given the negative effects of chronic elevations in stress hormones, it doesn’t strike me as unreasonable that some animals in this position might opt for violence. This also would explain why sometimes there appears to be no connection between anything that people believe to be physically or behaviorally painful that was deliberately or unwitting done to the animal and the attack, whereas other times the attack is triggered by something so seemingly inconsequential that it’s dismissed as insignificant. (“She always plays a game where she teases him with his toys after a training session so that couldn’t be a problem.”)
That’s a frightening thought isn’t it? That in the process of imposing what we consider a most humane and even loving teaching system on animals, we may be driving some of them to violent behavior. It’s even more frightening to consider that similar breakdowns in human-animal communication may also lead to similar tragedies in our interactions with domestic and companion animals.
But unsettling as such thoughts may be, for me considering them is preferable to stuffing my head up my most posterior body orifice in an attempt to avoid them. Time and time again when these tragedies occur, somebody says, “We need to look at the way we’re training these animals.” And time and time again, these elicit a kill-the-messenger response. After all, we tell ourselves, what could be more positive than all-positive training? What animals in their right minds wouldn’t respond to such a system? The ones that don’t? Well, they’re obviously not in their right minds.
I wouldn’t presume to tell anyone how to properly train whales, dolphins, elephants or any other wild animal. But I do know that the same erroneous assumptions that underlie reward (and punishment) training in wild animals also are contributing factors in some cases of domestic animal-human violence. As long as we continue to deny this, people will continue to be hurt, and animals will continue to die because of it.
If you have any comments regarding subject matter, favorite links, or anything you’d like to see discussed on or added to this site, please let me know at mm@mmilani.com.