- Location
- Andorra and Spain
A few days back, there was a Facebook Live between Nando Brown, Jo-Rosie Haffendon and Kamal Fernandez, on the topic of,"Is Positive Reinforcement damaging our dogs?"
Here is is, for anyone interested in a watch.
First, off, man that's some disorganised set-up! You need to skip to about eight minutes in before Kamal even joins in, and then there are still some sound issues which I think were cleared up around minute 13 or so. Then the battery ran out at the end, so it was cut short....
But, anyhow, we'll put all that aside and focus on the topic at hand.
My first reaction was one of horror. What were these people, all supposedly positive reinforcement trainers, saying?!
It's really uncomfortable watching. They even talk about shock collars not necessarily being aversive. OK, we might know on some theoretical level that that is true; it is not us who determines what is aversive to a dog, but to hear these guys talking about it? No, that sits uncomfortably with me.
After I'd finished watching, though, I did get to thinking about some of the topics they had raised.
1. Positive reinforcement training can lead to poorly behaved dogs.
For sure! I mean, any method of training, or no training at all, can lead to poorly behaved dogs. Now, for those of us who have been at this for some time, we know that this is a poor application of positive reinforcement procedures. Along with loads of other people, when I first started out on this journey, I thought that we had to be in our positive bubble 24/7. And to some extent, that's true. I mean, I'm not going to start smacking my dogs around just because I'm not in a formal training setting! I have my ethics and how I want to exist with my dogs, and physical corrections, and even continuous nagging, aren't a part of that. But back in the day, I didn't necessarily have the skills I have now (and now I don't have the skills I will further down the line), and I didn't have the answers of how to go about solving certain unwanted behaviour in a positive way. I knew I wasn't going to start including corrections, but that left me in a bit of a void. How do you solve problem behaviours if you don't have the tools to do so? That's where management comes in a lot of the time, and, for sure, that's a big part of it, but it can also sometimes leave us feeling rather unfulfilled. The dog hasn't been trained to behave better, we've just taken all the responsibility onto our own shoulders.
We bring dogs into our homes and lives, and so I believe that we have an ethical obligation to teach them the skills they need to navigate in that world without punishing them for stepping over the arbitrary lines that we create. However, that doesn't mean that they don't have to learn. If we are in a world where our dogs are going to interact with other people, then it's also our responsibility to make sure our dogs are safe and under control in those contexts. That means that the dogs have to learn a certain level of obedience. They need to not bark all day and night, disturbing the neighbours. They need to be under control in public, not harassing other people and animals. Not interfering with livestock, wildlife, or traffic. If we don't have the skills to train these behaviours, then we can run into problems.
So, it's not really that +R training leads to poorly behaved dogs, more that SOME people who are subscribing to a 100% positive reinforcement philosophy don't have the ability to train their dogs using those methods and, in the absence of any other methods, end up with dogs with poor manners. Add to this people who think that using leads is unnecessary, and you can see where the reputation may come from.
There's also a bit of the puzzle being about what constitutes "poorly behaved", as that's very subjective. I've mentioned before that I simply can't be bothered to teach my young dogs not to jump up. They get very little exposure to other people, so it would be a huge effort for me to create training scenarios for this. So I simply manage the situation by either feeding to the floor, or by putting my dog on leash and telling people not to approach unless they're happy with being jumped on. Willow grew out of it, Shadow never did it, Luna is largely over it but it can still happen on occasion. Is this bad behaviour? Probably, but no-one gets jumped on unless they come into our space, as she never does it off lead.
2. Dogs can't always be given choice
This bit put me out a little bit, when there was a mention of giving the dogs choice, almost with a sneer. Wow. No-one ever said that dogs have to choose everything. Just that, where it's appropriate, they should be given the ability to opt out. That this is actually a huge reinforcer and can be part of a training plan to ensure the dog opts in consistently. Of course there are times when dogs can't have choice, and they just have to suck it up. Take veterinary visits. We can teach as much cooperative care as we like, but when your dog needs a needle stuck in them, it needs to be done, no questions asked. But does that mean that cooperative care is worthless? Of course not. The more we can give the dog the choice to comply with the "mock" procedures, when it doesn't matter if they say no, the more good feelings they will get for the scenario, and so the easier it will be for them when they have no choice. We just have to make the difference very clear: in this situation, you get to choose, but in this situation, you don't. Take getting into the car as another example. With a dog who is anxious about the car (as I have had), I practice, practice, practice, giving her the option of whether she wants to jump in or not. It's always her choice during these sessions, and she can jump out at any time. Once she was happy doing this and we could start going out, there was that question: are you happy to jump in? If not, she would get taken back inside, fine. So when I was still unsure about whether the answer would be "yes", if we had to go anywhere important, I didn't ask the question. I just picked her up and put her in. She had no choice. Did she like it? Nope. But that's part of life, sorry. Could I have continued just putting her in the car? Of course, but she would never have changed her view of it. But with training (and drugs!), she learned again that the car is a good thing. Now, I'm 100% certain that she will jump into the car when I ask her, so I do ask her. She will not say no. That is the power of cooperative care and why it's important that we understand the power of choice.
When behaviourists in zoos can clicker train their wild animals to take injections and unpleasant medical examination through the bars of a cage with no restraint, then we can surely manage the same with our dogs
3. You can become "too positive"
This sort of goes hand in hand with point 1, but they talked about it a bit, and I think it's something that people misunderstand when they are first introduced to positive reinforcement training. I know I did. That is, you hear about not saying no to your dog; about reinforcing the good and ignoring the bad. That's a big pile of tosh, and gives us a bad reputation. I cringe when people say it, but I used to say it, too. If the dog is doing something, it's because that something is reinforcing to the dog. By ignoring it, we are not removing the reinforcement. The dog is still getting reinforced. By very definition, that is not going to make the unwanted behaviour magically disappear. So when people who use aversives in their training hear this, they laugh at us, and justifiably so. It's totally bonkers.
Where this idea has come from is from formal training sessions. You hear me spout on about this a lot (and a lot more in recent times, as I'm only just now really getting my head around it). When you are training a behaviour, you set up the environment for the dog to succeed. Once the dog knows what is expected of him, you can start pushing, extending criteria. If the dog fails to meet the criteria, you stop. You don't punish the dog. This may look like it in the moment, but it doesn't mean you're ignoring the behaviour. What you need to do is assess what happened, and why, and then adjust your training plan. It's the exact opposite of ignoring; it's tailoring what you do going forward, using that failure as information. And, yes, I used the word failure. A one-time me would have jumped on that word, too. It's not a poisonous word, it's a descriptive one. Criteria were set, they were not reached. Is it the fault of the dog? No, probably not. Is it the fault of the handler? Maybe. But it could also be one of numerous other circumstances. It's not a poisonous word, simply a descriptive one.
Outside of a formal training session, we don't have the luxury of controlling our environments totally. We don't necessarily have the undivided attention of our dogs, and we're not giving them ours. It's just life. And life happens. You may have left a whole roast chicken on the side when you were distracted by your child, and you see your dog headed towards it. What are you supposed to do? Ignore it? Of course not - that's one way to a dog's tummy full of lovely roast chicken, and you having beans on toast for dinner again. But that's what some people subscribe to. Yes, yes, yes, we should have set the dog up for success, by not putting them in that situation in the first place, but distractions happen to us as well as our dogs. So, in that moment, I'd undoubtedly shout at the dog. I might clap my hands or make a noise with something else. Is it totally positive? Well, duh, of course not. But it's not part of my training plan, I'm just trying to get out of Dodge. From thereon in, I can decide if I want to train the dog to leave food on counters, or if I will try to train myself to deal better with distractions, or if I'll just shout again next time. There is no "right" answer, it totally depends on you, your dog, and your situation.
4. When positive can be unethical
This is something that they got to towards the end, and I thought was very thought-provoking. The example they gave was a dog who was too afraid to leave its kennel in the shelter. Using purely positive reinforcement methods, this dog would be counter-conditioned slowly over the course of what? days? weeks? months? until it was confident enough to leave the kennel. I've done a lot of CC, I know how bloody painful it can be! What they did went completely against all positive reinforcement protocols, as they physically moved the scared dog. And, after a very short time, the dog learned that everything was OK and it was happily going in and out of its kennel. The "me" that has worked with Willow to condition against her fears was cringing when they talked about this. It's not the sort of thing that would work with her. But, for this dog, and thinking only about the ethics, what is better? Weeks spent cowering in a kennel, unable to go outside; or a very short period of extreme discomfort, leading to quick habituation? Now, I am not for a second suggesting anyone should just do this with a fearful dog, BUT it does raise questions about how ethical it is to go down the "most ethical" route. Could there have been fallout? For sure, and that's why flooding like this shouldn't be done willy-nilly. BUT, it also shouldn't be dismissed as nothing but bad.
5. Efficacy
This is the crux for most pet dog owners. They want to do something that works to sort out the behaviour issues they are having. The average Joe pet owner doesn't have the skills, nor the time or patience to perfect their clicker training skills. If it's something they want to do as a hobby, then brilliant! Teach them those skills. Learning to train dogs is fabulous fun, and I wouldn't do it any other way than through positive reinforcement. But I can understand those people whose lives don't revolve around their dog training; they have a busy enough life as it is, and just need this problem to be sorted. Let's say it's jumping up on that counter again. Let's say they have a house full of kids and it's simply unreasonable to expect management to be in place at all times. This is real life. So, do we, as positive reinforcement trainers, tell them that, bad luck, you just have to do better to manage your dog's environment? If we do, we're setting that person up to fail. Not only that, we're adding fuel to the fire of "positive reinforcement doesn't work", because we cannot expect that person, who has no interest in becoming a dog trainer, to learn the skills necessary in order to train the whole protocol successfully.
If it was a big deal for you or I, we would go about addressing it systematically, but all of us are reasonable enough to know that this takes commitment, time, and energy that this person probably doesn't have. For many dogs, a well timed "NO" would probably be a lot quicker and would produce lasting results if used consistently over a relatively short period of setting the dog up to fail. For me, this isn't something I need to worry about, as I am more than happy with the slower path, but for the "dog trainers" out there, the people who are passing their knowledge on to paying clients, I do think it leads to a big conundrum.
Those of us who are committed to positive reinforcement training methods have to acknowledge, as much as we don't like it, that sometimes our chosen route is not the most efficacious for all dogs, in all scenarios.
In summary, I think that the video was quite thought-provoking, but I was horrified by how it was presented. It was a lot of confused spewing of unfinished thoughts which sounded a lot as if they were advocating things they actually weren't. I think that it was the sort of conversation that could happen amongst trainers who were very sure of each other's methods, but in the public domain, it is potentially damaging, as it sounds very much like they are turning their backs on positive reinforcement, just because of a lack of coherence in their presentation. I wasn't sure about composing my thoughts and putting them "out there", because it adds to the publicity of a video which I think many aversive-based trainers would be clapping their hands at, but I felt that trying to add a little bit of logic to the conversation might help - who knows?!
While I was composing this, @Heidrun pointed me to another video, by Jane Ardern, who is a +R gundog trainer. It's far better articulated than both the first video I posted, and my rambling above! This is also a lot shorter and without the technical issues
Very interested to hear everyone's thoughts
Here is is, for anyone interested in a watch.
First, off, man that's some disorganised set-up! You need to skip to about eight minutes in before Kamal even joins in, and then there are still some sound issues which I think were cleared up around minute 13 or so. Then the battery ran out at the end, so it was cut short....
But, anyhow, we'll put all that aside and focus on the topic at hand.
My first reaction was one of horror. What were these people, all supposedly positive reinforcement trainers, saying?!
It's really uncomfortable watching. They even talk about shock collars not necessarily being aversive. OK, we might know on some theoretical level that that is true; it is not us who determines what is aversive to a dog, but to hear these guys talking about it? No, that sits uncomfortably with me.
After I'd finished watching, though, I did get to thinking about some of the topics they had raised.
1. Positive reinforcement training can lead to poorly behaved dogs.
For sure! I mean, any method of training, or no training at all, can lead to poorly behaved dogs. Now, for those of us who have been at this for some time, we know that this is a poor application of positive reinforcement procedures. Along with loads of other people, when I first started out on this journey, I thought that we had to be in our positive bubble 24/7. And to some extent, that's true. I mean, I'm not going to start smacking my dogs around just because I'm not in a formal training setting! I have my ethics and how I want to exist with my dogs, and physical corrections, and even continuous nagging, aren't a part of that. But back in the day, I didn't necessarily have the skills I have now (and now I don't have the skills I will further down the line), and I didn't have the answers of how to go about solving certain unwanted behaviour in a positive way. I knew I wasn't going to start including corrections, but that left me in a bit of a void. How do you solve problem behaviours if you don't have the tools to do so? That's where management comes in a lot of the time, and, for sure, that's a big part of it, but it can also sometimes leave us feeling rather unfulfilled. The dog hasn't been trained to behave better, we've just taken all the responsibility onto our own shoulders.
We bring dogs into our homes and lives, and so I believe that we have an ethical obligation to teach them the skills they need to navigate in that world without punishing them for stepping over the arbitrary lines that we create. However, that doesn't mean that they don't have to learn. If we are in a world where our dogs are going to interact with other people, then it's also our responsibility to make sure our dogs are safe and under control in those contexts. That means that the dogs have to learn a certain level of obedience. They need to not bark all day and night, disturbing the neighbours. They need to be under control in public, not harassing other people and animals. Not interfering with livestock, wildlife, or traffic. If we don't have the skills to train these behaviours, then we can run into problems.
So, it's not really that +R training leads to poorly behaved dogs, more that SOME people who are subscribing to a 100% positive reinforcement philosophy don't have the ability to train their dogs using those methods and, in the absence of any other methods, end up with dogs with poor manners. Add to this people who think that using leads is unnecessary, and you can see where the reputation may come from.
There's also a bit of the puzzle being about what constitutes "poorly behaved", as that's very subjective. I've mentioned before that I simply can't be bothered to teach my young dogs not to jump up. They get very little exposure to other people, so it would be a huge effort for me to create training scenarios for this. So I simply manage the situation by either feeding to the floor, or by putting my dog on leash and telling people not to approach unless they're happy with being jumped on. Willow grew out of it, Shadow never did it, Luna is largely over it but it can still happen on occasion. Is this bad behaviour? Probably, but no-one gets jumped on unless they come into our space, as she never does it off lead.
2. Dogs can't always be given choice
This bit put me out a little bit, when there was a mention of giving the dogs choice, almost with a sneer. Wow. No-one ever said that dogs have to choose everything. Just that, where it's appropriate, they should be given the ability to opt out. That this is actually a huge reinforcer and can be part of a training plan to ensure the dog opts in consistently. Of course there are times when dogs can't have choice, and they just have to suck it up. Take veterinary visits. We can teach as much cooperative care as we like, but when your dog needs a needle stuck in them, it needs to be done, no questions asked. But does that mean that cooperative care is worthless? Of course not. The more we can give the dog the choice to comply with the "mock" procedures, when it doesn't matter if they say no, the more good feelings they will get for the scenario, and so the easier it will be for them when they have no choice. We just have to make the difference very clear: in this situation, you get to choose, but in this situation, you don't. Take getting into the car as another example. With a dog who is anxious about the car (as I have had), I practice, practice, practice, giving her the option of whether she wants to jump in or not. It's always her choice during these sessions, and she can jump out at any time. Once she was happy doing this and we could start going out, there was that question: are you happy to jump in? If not, she would get taken back inside, fine. So when I was still unsure about whether the answer would be "yes", if we had to go anywhere important, I didn't ask the question. I just picked her up and put her in. She had no choice. Did she like it? Nope. But that's part of life, sorry. Could I have continued just putting her in the car? Of course, but she would never have changed her view of it. But with training (and drugs!), she learned again that the car is a good thing. Now, I'm 100% certain that she will jump into the car when I ask her, so I do ask her. She will not say no. That is the power of cooperative care and why it's important that we understand the power of choice.
When behaviourists in zoos can clicker train their wild animals to take injections and unpleasant medical examination through the bars of a cage with no restraint, then we can surely manage the same with our dogs
3. You can become "too positive"
This sort of goes hand in hand with point 1, but they talked about it a bit, and I think it's something that people misunderstand when they are first introduced to positive reinforcement training. I know I did. That is, you hear about not saying no to your dog; about reinforcing the good and ignoring the bad. That's a big pile of tosh, and gives us a bad reputation. I cringe when people say it, but I used to say it, too. If the dog is doing something, it's because that something is reinforcing to the dog. By ignoring it, we are not removing the reinforcement. The dog is still getting reinforced. By very definition, that is not going to make the unwanted behaviour magically disappear. So when people who use aversives in their training hear this, they laugh at us, and justifiably so. It's totally bonkers.
Where this idea has come from is from formal training sessions. You hear me spout on about this a lot (and a lot more in recent times, as I'm only just now really getting my head around it). When you are training a behaviour, you set up the environment for the dog to succeed. Once the dog knows what is expected of him, you can start pushing, extending criteria. If the dog fails to meet the criteria, you stop. You don't punish the dog. This may look like it in the moment, but it doesn't mean you're ignoring the behaviour. What you need to do is assess what happened, and why, and then adjust your training plan. It's the exact opposite of ignoring; it's tailoring what you do going forward, using that failure as information. And, yes, I used the word failure. A one-time me would have jumped on that word, too. It's not a poisonous word, it's a descriptive one. Criteria were set, they were not reached. Is it the fault of the dog? No, probably not. Is it the fault of the handler? Maybe. But it could also be one of numerous other circumstances. It's not a poisonous word, simply a descriptive one.
Outside of a formal training session, we don't have the luxury of controlling our environments totally. We don't necessarily have the undivided attention of our dogs, and we're not giving them ours. It's just life. And life happens. You may have left a whole roast chicken on the side when you were distracted by your child, and you see your dog headed towards it. What are you supposed to do? Ignore it? Of course not - that's one way to a dog's tummy full of lovely roast chicken, and you having beans on toast for dinner again. But that's what some people subscribe to. Yes, yes, yes, we should have set the dog up for success, by not putting them in that situation in the first place, but distractions happen to us as well as our dogs. So, in that moment, I'd undoubtedly shout at the dog. I might clap my hands or make a noise with something else. Is it totally positive? Well, duh, of course not. But it's not part of my training plan, I'm just trying to get out of Dodge. From thereon in, I can decide if I want to train the dog to leave food on counters, or if I will try to train myself to deal better with distractions, or if I'll just shout again next time. There is no "right" answer, it totally depends on you, your dog, and your situation.
4. When positive can be unethical
This is something that they got to towards the end, and I thought was very thought-provoking. The example they gave was a dog who was too afraid to leave its kennel in the shelter. Using purely positive reinforcement methods, this dog would be counter-conditioned slowly over the course of what? days? weeks? months? until it was confident enough to leave the kennel. I've done a lot of CC, I know how bloody painful it can be! What they did went completely against all positive reinforcement protocols, as they physically moved the scared dog. And, after a very short time, the dog learned that everything was OK and it was happily going in and out of its kennel. The "me" that has worked with Willow to condition against her fears was cringing when they talked about this. It's not the sort of thing that would work with her. But, for this dog, and thinking only about the ethics, what is better? Weeks spent cowering in a kennel, unable to go outside; or a very short period of extreme discomfort, leading to quick habituation? Now, I am not for a second suggesting anyone should just do this with a fearful dog, BUT it does raise questions about how ethical it is to go down the "most ethical" route. Could there have been fallout? For sure, and that's why flooding like this shouldn't be done willy-nilly. BUT, it also shouldn't be dismissed as nothing but bad.
5. Efficacy
This is the crux for most pet dog owners. They want to do something that works to sort out the behaviour issues they are having. The average Joe pet owner doesn't have the skills, nor the time or patience to perfect their clicker training skills. If it's something they want to do as a hobby, then brilliant! Teach them those skills. Learning to train dogs is fabulous fun, and I wouldn't do it any other way than through positive reinforcement. But I can understand those people whose lives don't revolve around their dog training; they have a busy enough life as it is, and just need this problem to be sorted. Let's say it's jumping up on that counter again. Let's say they have a house full of kids and it's simply unreasonable to expect management to be in place at all times. This is real life. So, do we, as positive reinforcement trainers, tell them that, bad luck, you just have to do better to manage your dog's environment? If we do, we're setting that person up to fail. Not only that, we're adding fuel to the fire of "positive reinforcement doesn't work", because we cannot expect that person, who has no interest in becoming a dog trainer, to learn the skills necessary in order to train the whole protocol successfully.
If it was a big deal for you or I, we would go about addressing it systematically, but all of us are reasonable enough to know that this takes commitment, time, and energy that this person probably doesn't have. For many dogs, a well timed "NO" would probably be a lot quicker and would produce lasting results if used consistently over a relatively short period of setting the dog up to fail. For me, this isn't something I need to worry about, as I am more than happy with the slower path, but for the "dog trainers" out there, the people who are passing their knowledge on to paying clients, I do think it leads to a big conundrum.
Those of us who are committed to positive reinforcement training methods have to acknowledge, as much as we don't like it, that sometimes our chosen route is not the most efficacious for all dogs, in all scenarios.
In summary, I think that the video was quite thought-provoking, but I was horrified by how it was presented. It was a lot of confused spewing of unfinished thoughts which sounded a lot as if they were advocating things they actually weren't. I think that it was the sort of conversation that could happen amongst trainers who were very sure of each other's methods, but in the public domain, it is potentially damaging, as it sounds very much like they are turning their backs on positive reinforcement, just because of a lack of coherence in their presentation. I wasn't sure about composing my thoughts and putting them "out there", because it adds to the publicity of a video which I think many aversive-based trainers would be clapping their hands at, but I felt that trying to add a little bit of logic to the conversation might help - who knows?!
While I was composing this, @Heidrun pointed me to another video, by Jane Ardern, who is a +R gundog trainer. It's far better articulated than both the first video I posted, and my rambling above! This is also a lot shorter and without the technical issues
Very interested to hear everyone's thoughts
