What is going on in the working dog world? Do you want to stay informed about the latest developments and insights and do you appreciate the opinions of professionals? Then listen to K9 conversations Working Dog podcast Dutch. You will hear well-known and lesser known people here. Leaders in our field, experts, colleagues, but also people in a supporting role. We receive both Dutch and English-speaking guests. Who are we? We are Patrick Peterson and Minke van den Berg and this is K9 Conversations Working Dog Podcast Dutch.
[Music]
So, what they claimed to have shown is that when you work as a human professional, 70 percent of your skills you learn from actually doing them. Trying to apply your knowledge to your work, that is how you get better.
[Music]
Patrick:
Okay, here we are again in our second episode of our podcast. Today we have a very special guest, the person who actually brought Minke and I together. So, without further due, would you like to introduce yourself? Who is sitting on the other side of the microphone?
Jens:
My name is Jens Frank from Sweden and the Scandinavian Working Dog Institute.
Patrick:
Welcome, thank you very much.
Minke:
Yes, welcome!
Jens:
It’s nice to be here.
Minke:
It’s good to have you here at our kitchen table. We’re going to have a nice talk. At least, I am going to try not to cough. So, for the listeners, I’ve had the flu and I am trying not to cough (laughing).
Patrick:
We’ll be fine.
Jens, welcome, and thank you that you’re willing to take the time to do a podcast episode with us. You are very high on our list of people we would like to interview. The reason we would like to interview you is, like I mentioned before, because you are the one who brought Minke and I together. In part because of the shared passion we have for training dogs, but more important; the way you train your dogs. For the people who don’t know who Jens Frank is, can you please tell us something about yourself and how you got into dog training?
Jens:
Yes, so I have two jobs. I have a position as an associate professor at the Swedish University of agricultural Sciences where I've been working for the last 25 years. That was actually also how I came into contact with the professional dog training world. Before that we only had herding dogs, hunting dogs and pet dogs at home but when I started as a PhD student I worked half the time as a field assistant in the wolf and lynx research projects and there we used dogs for multiple purposes. That's how I started to train dogs professionally in the late 90’s. It just continued from there. Maybe 15 years ago we started the Scandinavian Working Dog Institute because there are some things you cannot do as part of the University. We couldn’t give courses to anyone or different types of trainings that were not directly related to the research funding we had. So, then you're encouraged to start an institute where you're more free to do whatever you want, to try and bridge the university and the real world outside.
Patrick:
Okay. What kind of courses do you give? Your main focus is on detection work if I’m correct. How would you describe the courses that the SWDI offers?
Jens:
I would say it's the general working dog skills that we teach; detection, tracking, functional obedience and also some more specialized forms of detection where you send the dog with the laser or just with hand signals or verbal cues. That's mainly what we do. We mostly work with working dog handlers from different authorities and those are the skills that they need so then that is what we provide.
Patrick:
Okay. What you read and hear a lot about is the way that you have introduced and incorporated the focus training in which you use the Kong. How did you came up with that type of training? Or was it already there and you just developed it further? How did that evolve?
Jens:
I don't think that we can take the
credit for that method. It's the standard method in many parts of the
world for training the indication. You use
activity and passivity sessions to make the dog switch from being
active and passive. It wants the toy
to become active instead of passive so
it simply learns that if it sits for example, and points the nose at the toy, the toy will become active and the game will start. We did not invent that but
we use it because it's an easy way
to get an intensive indication. The
more intense the indication is, the easier it is to incorporate higher levels of arousal and
motivation and still get it
working reliably.
I'm trying to think of where it could be
described earlier. It's described in the
Norwegian people’s AIDS
documents I think early on and also in
the
documents describing the training of mine
detection dogs from the Geneva International Centre for Humanitarian Demining (GICHD). So, this method of using activity and
passivity sessions has been described a lot earlier.
Patrick:
Okay, and if you compare this, for instance, to people who are using metal pipes as a reward in their detection training. What do you think is the main difference between using the metal pipe, and flipping that from the target odour or using the system in which you let the dog indicate on the Kong when you start?
Jens:
Basically, I would say there is no difference because what we make use of in both cases is that the primary reinforcer is what the dog should stare at. It's very easy to get the dog to focus on what it wants most, whether that is a rubber toy or a metal pipe doesn't matter. We can train the indication using a metal pipe, a tennis ball or a Kong toy as long as that is a primary reinforcer for the dog. So, when we’re training the mechanics of movement it doesn't matter what primary reinforcement we use. However, what you also referred to was that you try to make the dog associate the primary reinforcer coming from the source of the scent, that is another part. I would say that most of the time I find it ineffective if we try to teach the indication at the same time as we're training and teaching the dog how to search and what to search for. Then there are simply too many parts involved for both the handler and the dog to keep track of in a good manner. Then we will start to reinforce behaviours we don't want and we'll punish behaviours that we do want.
Do you agree...?
(laughter)
Patrick:
Yes, I understand
what you mean. The first time I met you, which was at
Eric's in the Netherlands, I found what you said very
interesting but I also found it to be
unpleasant for me as a person. I will
explain why so you don't feel offended.
(laughter)
You made videotapes during your
seminar of each and every participant, so you also made videos from me which allowed me to see my own behaviour. I was a very hectic, I was all over the place. On the contrary, you oozed
calmness. You are so very calm in your
training. That really opened up my eyes because
I saw my own behaviour and what it did to
the behaviour of the dog.
Some people think that the system you use, which I really love, looks very clinical and is very easy-going. They miss the energy in it. I can understand that from their point
of view because they are only looking at
videos. They haven’t
visited your seminars. Could you
respond to that?
Jens:
I cannot say what is right or wrong, but what I find is that … I haven't been working a lot with pet dogs, but I can clearly see that when working with pet dogs you need to motivate them because they are not crazy for the primary reinforcer like the ball or the bite thing. Then you need to be a little bit like… This that's not a good way to describe something in a podcast…
(laughter)
You need to be able to do hasty
movements to get the dog interested. But I am mainly working with
working dogs who are selected because
they are crazy for the toy or for the
biting.
They will work in order to
get that. They don't work because I feed
them with energy. My opinion is that
if I need to be intensive in order to
get the dog to work then I have made a poor selection. Such a dog is not
suitable as a working dog in my opinion,
because then it will require the handler
to do this all the time; to actually give external
motivation to the dog and the handler will not be able to do that in many of the
situations where they are working in in
real life.
So, I would say that a dog that
needs that
might be a good sport or pet dog but not a
working dog. That's
how I see it.
Patrick:
Besides that, what are other criteria are you looking for when you are selecting a dog to train for a unit for instance?
Jens:
It depends on their specifications. So for example for the Special Forces attack dogs the basic criteria is that the dog is crazy for the primary reward and preferably as many as possible. If you have a dog that is supposed to do biting and it is only reinforced by biting the sleeve or the suit and not also crazy for the ball, then you will have troubles training the obedience around decoys for example. So I prefer having as many primary reinforcers available as possible because if the dog is crazy to bite the decoy but it's also crazy for the ball and for food, then everything will be a lot easier. So I am looking for as many of those primary enforcers as possible, and of course as little fears as possible. Environmental fears, for example fear of heights, gunshots or stairs, are really not acceptable with any working dog. Those are the two basic ones; fears and motivation.
For the Special Forces it's just as important that the dog is silent, even if it gets frustrated it can’t leak any sound. I find that although that can be taught to many dogs, it cannot be taught to all dogs because some dogs are not aware that they make a sound. In that situation we can't really punish the dog because we cannot punish behaviour of which the dog is not aware that it's doing it. That is very tricky.
Minke:
By making sound
you mean whining…?
Jens:
Yes, when they have a high expectation that can happen. The team may be running through or to a house while there is shooting and shouting and people running… The dog's expectation goes up in such instances because it knows that it will soon find someone to bite. If the handler then stops by a door and tells the dog to lay down, the anticipation the dog feels can easily make it start whining. That is unacceptable.
Patrick:
You want to leave any form of vocalization out.
Minke:
Yes.
Patrick:
I’ve talked with a lot of people all over the world, from America to Jordan to England to the Netherlands, etc. and there seem to be two groups of people; those who think that good dual-purpose dogs don’t exist, and those who say they do exist. I would like to know what your opinion is about that subject.
Jens:
I have seen a lot of really good dual-purpose dogs in the different special forces units using them for detection, attack and tracking but also in the general police who use the dogs for protection or apprehension, tracking and detection. I would say that in general, if you have a dog that is highly motivated, there are few limits for what we can teach it, but if we only rely on the dog being aggressive towards humans then it will of course be difficult to use such a dog as a good dual-purpose dog. So I think it has to do with selection. If you have a dog that has many primary reinforcers, not only biting people, then there is no problem at all. The dog will love to be given the cue to attack and bite someone but it's not difficult to teach it to search for explosives around people that it has just bitten or is supposed to bite in five minutes because we can make the cue the context. Then when the cue to search is given the dog doesn't expect to bite someone and it doesn’t have to control itself all the time because it wants to bite but has to search instead. In that context it's not really thinking about the people being available as reinforcers. So, in my opinion, I have seen a lot of very good dogs that are at work every day doing this.
Patrick:
Yes. In America you only see dual-purpose dogs for instance. There are hardly any single purpose dogs there. In the patrol work you see a lot of dogs used for both narcotic detection dogs and apprehension at the same time.
Jens:
Yes.
Patrick:
When talking about a bite work, that's my cup of tea in this case of course, I find that sporting dogs who are being trained in bite work do not always perform well as a professional working dogs. That doesn't mean that for instance KNPV dogs are not suited for our work, they are usually well suited for our work, but what we have found is that when they start training those dogs they create too much arousal from a very young age. They only use arousal, arousal, arousal to a point where the dog almost loses his mind and then they try to control that arousal by using compulsion. This kind of dogs are not suited for our type of work because I need a dog that can switch on and off depending on the context, like you explained before. When we get a call to go to a bar fight for instance, which creates a high level of arousal in the dog, and within 10 minutes we get another call that there has been a burglary where the guy who did it is caught but has left some articles that we have to let the dog search for. If you have a dog that is trained in such a high level of arousal from a young age, it can be very difficult for him to be able to make that switch because he simply hasn't 't learned how to do that. What’s your opinion about that?
(Laughter)
I'm just picking your brain. I'm not trying to convince you of my own opinion. This is just what I have found throughout the years. I'm curious about different people’s opinion on the subject. In this podcast it is not about the opinions of Minke and me. We just want to establish a platform where we discuss as many opinions as possible from the guests we invite in order to create some sort of conversation about the topic.
Minke:
Yes.
Jens:
Yes. Before I get to my opinion, there is an old study which led to what's called the Yerkes-Dodson law of arousal. It's from 1908 so it's not very new, and what it says is… Now I'm showing something with my hands again which is not good for the listeners.
Minke:
You're showing a graph.
Jens:
Yes, it's in inverted “U” or a bell-shape where they show that there is an optimal level of arousal for performance when you're forming habits in animals. They have been working mainly with humans but I would be surprised if it would be different with dogs. So, there is an optimal level of arousal. If it’s too low you're almost going to sleep and you won’t be able to perform bitework very well. On the other hand, if you're too aroused, you will have a lot of arousal and energy without a direction, without focus. In that case, you won’t be able to do bite work or anything very well either. For optimal performance you need to be in the middle. What they showed is that for simple behaviours, the optimal level of arousal is higher than for more complicated behaviors, which makes sense. When we apply this to dog training, which is outside of their study so they cannot be blamed for it…
(Laughter)
I think that the way we can make difficult behaviours easy for the dog is through training, by teaching the dog the behaviour. Let's say when we start training a behaviour like transport or healing. Before the dog understands what it is supposed to do, that is a difficult behaviour. In order to teach the dog this we need to take the Yerkes-Dodson law into account, which means we must keep the level of arousal low. However, for the dog to be able to work later on we need to get to a stage where we train the dog while it’s in a high state of arousal because that is what it will be like when they work. By then the behaviour should be easy. We can accomplish that by teaching the dog what the position is like before we gradually increase the arousal. So, my opinion based on the Yerkes-Dodson law, is that if we have a specific behaviour we want the dog to perform well and we take this law into account, then we need to start with a lower level of arousal which we can increase as the dog starts to get a better understanding of the behaviour.
Patrick:
Correct me if I’m wrong, but I think that if you use that law you can use more positive reinforcers then if you do it the other way around. If you create a dog with an extremely high level of arousal from an early age, then it's more difficult to use solely positive reinforces and you might have to use more punishment in training. What I have found is that if you have a dog, for instance in bitework, that is in a high state of arousal, you need to control that functional aggression. If you don't, their learning curve or cognitive skills, whatever you like to call it, will decline. It's like they go inside a funnel and they are no longer susceptive to what you say or do. Their focus is solely on that what it is they want. If they’re in that state of mind it’s very difficult to teach them something, like the bitework for example.
Jens:
Yes, that’s because you have passed the optimal level of arousal and gone too far to the right. Then their performance will decline. That's what Yerkes and Dodson showed. The reason why the dog is not able to perform is because it is over-aroused. Maybe this is controversial in the Netherlands, I don't know, but to me it's not important that I am always able to use only positive reinforcement. I have no problems using punishment. What I am focused on is the dog's progression. If the dog will learn from punishment then I will use it, but the problem is that the with punishment we can only teach a dog what it should not do and in most cases, like in bitework, there are a thousand things the dog should not do in a particular situation. Training is much quicker if we, instead of using a thousand repetitions to make sure the dog understands what it should not do, can set the dog up in a situation where we're able to reinforce the dog for doing what we want it to do.
Patrick:
Does this also has something to do with something what I like to call ‘reverse thinking’? If I give a training to my handlers for instance, and I ask them what they would like to train, they usually reply with a list of things they don’t want their dog to do. Then I ask them look at it again from a different angle and think about what they would like their dogs to do. For example, if they say: ‘I don’t want my dog to jump on top of me’. I respond by saying: ‘Okay, but that's not what you want to train. You want to train your dog to stay with all fours on the ground’. Trying to think differently, looking at dog training from another perspective, is that also something that you find people have difficulties with?
Jens:
Yes
[Music]
Minke:
All right I'm going to try to talk without coughing, since you've been kicking me under the table Patrick because I'm not saying anything, which is unusual for me.
(Jens chuckling)
So, you guys have been talking a bit about progression and reinforcement and one of the things I found very interesting when we first met was actually using progression plans. That's something that I mention quite a lot to people I'm teaching or people whom I tell very enthusiastically about what I have learned from the SWDI and in training with you Jens and with Jessica and in the past with Tobias. So, a little bit about progression plans. That's also something you like as a topic Patrick.
Patrick:
Yes.
Minke:
What we are trying to get people to understand is what progression plans actually are, because progression planning sounds so boring actually.
Patrick:
It sounds a bit too theoretical for most people. A lot of dog trainers just want to go out, grab their dog and start training. They don't want to do paperwork.
Minke:
Yes, they don't want to think beforehand.
Patrick:
No, they just go with the training and see what kind of problems they will encounter. Then they’ll address those problems when they arise.
Minke:
Yes, you've seen that in your training and I've seen that with the people I teach. I've trained more pet dog people in detection. I've been training a couple of professionals as well, but you see it in both groups of people. You were saying it might be interesting to pick Jens’ brain on that.
Patrick:
Yes, of course.
Minke:
It might be good to explain why. Why progression plans, and why do you use them the way that you do? Maybe you should give a short explanation of what a progression plan is (laughing), if you can.
Jens:
Thank you (laughing).
Minke:
You're welcome (laughing).
Jens:
I can make an attempt. Then I would like to start with referring back to Patrick's first question because I think that…
Minke:
…It leads up to it, yes.
Jens:
Yes, because if you think from the perspective that there are many things you don't want the dog to do, and there are. I mean all of us who work with dogs and handle dogs have many things we don't want the dogs to do, but teaching the dog not to do these things is more effectively done by teaching the dog what it is we want it to do and make the dog so motivated to do this that the other things just don't cross the its mind. If we do that, what we realize is that there are in fact very few things that working dogs should do. It’s when you start to incorporate all the things they shouldn't do that you will get caught up in something. It then may feel like it could take years to train a police or military working dog because there are so many things the dog should “learn” (within brackets), not to do.
Patrick:
Yes.
Jens:
Take a bomb dog for example. If a bomb dog knows four behaviours there are very few tasks it cannot solve. If the dog knows how to search on cue, to indicate when it finds the target substance, to sit on cue and to transport to heal, if it knows those four things there are almost no bomb dog tasks the dog cannot solve. So, it's only four skills we want the dog to do, and what is so tricky with the working dogs is that they should do this with an extremely high reliability and in all sorts of contexts. That is the hard part. It's not teaching the dog the behaviours, but to make them work reliably anywhere and at every time, whether the dog feels like it or not. That's where we need the progression plans. (laughing)
Minke:
Yes. (laughing)
Jens:
Because those allow us to focus our efforts in teaching on what we want the dog to do, these four tasks. I find that when I train dogs, it’s a luxury for me to have a progression plan. A progression plan doesn't tell me how to teach the dog something, that's something for which we need to have a lot of experience as dog trainers or dog handlers, to be able to use the correct motivation or the correct punishment for this individual dog because that will vary a lot between dogs. On the contrary, the order in which we train the skills or the behaviours will rarely differ between the dogs. It doesn't matter if they are mainly motivated by food, a ball or biting. The order in which we train the behaviours will be approximately the same. So, the progression plan is just a list of sub goals that you make and which allows you as a handler to be able to focus only on the next sub goal. You don't need to worry about what the dog should be able to do in three months, you just focus on the next sub goal and if the distance between the sub goals is very small, then training will go quickly and most of the time it will also be more effective. The dog will learn quicker if there are smaller jumps between the steps in a progression plan, because as we said before then the dog gets more successful repetitions and is reinforced for doing what we want it to gradually do more of. There will be some corrections in between where the dog is punished because it is doing something else, but the most important thing is that we have more reinforcements. Not for ethical reasons, at least not for me, but from a pure effectiveness perspective because then I know the dog is learning what it should do, what pays of, and those are a lot fewer things than the things the dog should not do. However, in order to be able to use this list of sub-goals we must have an end goal in mind, and that's where I find that many units and dog handlers struggle, that's the hardest obstacle to cross; to make a specific and measurable end goal. What is the dog supposed to be able to do when it comes to tracking, detection or apprehension? You have to sit down and formalize this in a written text.
Patrick:
Do you also use your progression plans to solve problems you encounter during training? For instance, you are training a dog, let's say for three months, and after the second month you encounter a problem. Do you then take the notes from your progression plan to try and find out if, and where things went wrong? Do you use it as a log that you can revert to see if there have been any behaviours in earlier stages of training which might explain the problems you are facing at that moment? Or is a progression plan looked upon differently from your point of view?
Jens:
(Signs). When I was younger and, I don't know, maybe less tired (chuckling). I did a lot of writing actually, so it was more like a journal. Today, now that I am old and tired…
(laughter)
…I don't. So, I don't write anything. I don't think it's bad to do it, I just don't do it. I do write the progression plan which I then use that as a list of sub-goals that should lead me to the end goal. So, if I'm not able to complete a sub-goal, let's say I'm not able to make the dog sit and stare for 15 seconds when training the indication, then I might need to take a step back and think about what’s happening. When I have spent maybe two or three sessions and still cannot reach that goal, to go from 10 seconds to 15 seconds, then I know that I need to change something. I might need to increase the value of the reward or maybe I need to help the dog by using a bridge or by making the target bigger, but I need to change something. So, I will not go many steps back, I will still try to manage this step. I might go one step back in order to get some more reinforcements to keep the dog motivated and wanting to try.
Patrick:
To make it clearer for the dog what you actually want.
Jens:
Yes, but I think if we have a good progression plan where the distance between the steps is quite small, then we don't need any problem solving or trouble shooting. We just need to take one step back and we need to be able to see when a dog is not progressing. If there is no progress, the dog is not learning what you want it to, and that can be because the distance between the steps is too big. I could be too unclear for the dog. Maybe I'm pointing with my hands and that is what the dog is responding to more than the target scent, or maybe the value of my reinforcement is too low. So, I need to change something. If I change one thing at the time I will be able to see what will actually lead to success and what allows me to complete the next goal. Then I just continue from there. So, I don't I don't write.
Patrick:
Okay. Correct me if I'm wrong, but in that system three successful repetitions mean you're ready to adjust your criteria and go to the next step within your progression plan, while one failure in three repetitions means you will have to take a small step back. Is that true?
Jens:
The way I do it is that if I have three successful repetitions in a row for a particular sub-goal, then I must move on and train for the next one, even if it feels a bit shaky. If I have three failures in a row then I will consider maybe taking a step back, changing the reward or doing something else, I won’t just keep pushing. This is just in order to have a rule of thumb, three is a holy number…
(laughter)
It's a good number (laughing), there's not more thought behind it than that. You just need some kind of rule to prevent that you're just going on, pushing against a wall trying to train something that won’t work. I have a good example of this. I'm training a dog now that belonged to someone who tried to teach it to do tracking. We do it the same way most other people do it, the dog learns to follow the contaminated area on the ground in order to find articles in the track and when it then indicates the articles it receives a reward. In the beginning the tracks are very short so the dog learns that the easiest way to find articles is to first find the contaminated area, because that is where the articles magically are for some reason. If we use articles that are too big such as coins that have human scent on them for example, which smell a lot, then that scent will overshadow the scent from the ground where the person has been walking back and forth contaminating it. So, every repetition where the dog is finding the article without paying attention to the scent on the ground we actually get one step further away from the end goal. For every repetition I do that way, I will make the dog less and less likely to track and more and more likely just to do an article search. It's okay if that happens one or two times in the beginning, but after that we need to make the contaminated area more salient by walking barefoot, putting our hands down or dabbing a couple of sweat socks or something on the ground just to make the dog realize that it is this contaminated area where the articles can be found. It will have to make the association that if it just finds this contaminated area, it will find articles.
In a situation like that we need these of rules of
thumb to keep pushing us forward so that
we realize
early on that the dog has not learned to
pay attention to what we want it to pay
attention to. Then you're just doing
repetitions and you get to reinforce the
dog,
but not for the behaviours that you want
to see more of.
Patrick:
Okay, in the case you are talking about you work with a system that is called backwards chaining. You teach the indication first, then you start with a contaminated area and from that point of on you will gradually build up the exercise.
Jens:
Yes.
Patrick:
In this process, I can imagine you use things like chaining, shaping, luring, perhaps a little bit manipulation. Are those things you also use and incorporate in your training? Not only in tracking but for instance also in detection, bite work or directionals?
Jens:
Yes. I would say all progression plans, whether you make a written progression plan or you just train by feeling, start with luring. If you want to teach the dog to sit for example, you learn it to sit with food or a toy in your hand. That usually happens very quickly. When you have a dog that has understood that the hand is the cue to sit, then we replace this cue with a verbal cue. The rule of thumb is then that you give the new signal before the old signal. So, if I say sit and then move my hand upwards and towards the dog, the dog will start to ‘cheat’ because it knows that after I say sit I will raise my hand, so it will already start to sit when I say my cue. Maybe we spend an hour teaching the dog to sit on cue; first we lure it then we teach it the verbal cue, but after that then the dog knows the behaviour. It's more or less like that with all behaviours. Teaching the behaviour is very quick, the rest of the training is context training to make the dog do this behaviour on cue in all contexts.
Most of the time, backward chaining is the most effective way to do that, and then we will be using shaping. All of us who train working dogs use shaping, because shaping is defined as breaking the behaviour down into smaller pieces and training one piece at a time before putting them together in a chain. I think most of us do that, because there aren't really many ways in which we can teach an animal things. We have luring, shaping and imitation for example, but there are not many alternatives.
Minke:
I have two questions actually about what you're saying at the moment, because you're talking about imitation and we often see on your clips that when you're training your dogs, the puppies that you are training are present as well. They’re not specifically doing anything, but they are there. Can you explain why you have them there? Is that always a distraction for the dog that is being trained or do you have other reasons for it as well.
Jens:
I would say it's never a distraction for the dog being trained because then we have made a poor selection. It should be so eager to get the reinforcement; the bite or the ball, that that should not matter. The reason for why I do it is I think 50 percent laziness (laughing), because it's an easy way to activate the puppies. If they were not there they would have to spend hours in a room or in a crate somewhere and then I would have to spend those hours with them doing other things later. That's probably one of the main drivers to be honest, but then we also have the studies showing that dogs who have been observing another dog doing a behaviour need between 30 and 40 percent less repetitions to learn the same behaviour. So, sometimes that is the reason, but we need to pay attention because that goes for the bad behaviours as well (laughing)
Minke:
(laughing) I was just going to say that, is this just what you hope will happen..?
Jens:
Sometimes this happens and other times it doesn’t turn out that way. However, I know something that is likely to happen, and then we're back to talking about arousal again. A puppy that has been in situations where an older dog is aroused and where it is not , is more likely to get the same kind of association with those situations. For instance, if it sees the other dog being aroused during bitework then the puppy will be more likely to be very aroused in those situations. If it sees the other dog having a low arousal in other situations where it's sitting or transporting for example, then the puppy is more likely to associate these situations with a low level of arousal. That's where I find that it mainly pays off to have the puppies present during training.
[Music]
Patrick:
I heard you talking about social facilitation and social imitation, I have also heard you talk about overshadowing and you introduced the Yerkes-Dodson law. So, is it true if I say that your training also has a very strong theoretical base which you incorporate in your training? There are a lot of people who know a lot about the theoretical background in training. How do you incorporate it in your training?
Jens:
I guess I at least try to use common sense and apply it (laughing). Learning is interesting, not only in dogs but also in humans. I read a while ago about what they call the 70-20-10 rule. It was an article from a Cambridge research group and I've seen a lot of criticism to it as well, but it makes sense to use or apply it I think, especially in the dog training business. What they claimed to have shown is that when you work as a human professional, 70 percent of your skills you learn from actually doing them. Trying to apply your knowledge to your work, that is how you get better. 20 Percent of your skills you acquire from looking at your colleagues and asking them for advice and 10 percent is what you learn from the courses and formal theoretical education you have. That makes sense to me. What I sometimes see in the dog training business is extremely skilled trainers that have been doing this for a long time and that have no clue about the Yerkes-Dodson law or the theoretical parts, but who still produce extremely good results in terms of the dogs and handlers they train. Then everything is fine…
Patrick:
Yes, because you get your results.
Jens:
Yes, and they may have spent 30 years getting to that point. I think that if you don't have 30 years of experience but you want to become good, you can take a shortcut by also adding theoretical knowledge and by asking others and learning from others. But you still need to acknowledge that you need to put this into the 70 percent where you actually convert this theoretical knowledge into practice, otherwise it is useless.
Apart from these extremely good trainers with no or little theoretical knowledge I see, in my experience at least, a much larger group of people with extremely high levels or theoretical knowledge. They have taken all the courses, read all the books and they can explain operant conditioning and the four quadrants in their sleep, but when you see them train their dogs they don't apply it. They reward the dog when they think the dog deserves it for example. Then you really don't apply operant conditioning, you just apply your own selfish emotions and desire to be able to do your dog good. The dog will learn nothing from that. It would be good if we could make people apply their theoretical knowledge, because many times I think they don't need more theoretical knowledge. They don't need to know about another law or another theory. They need to apply the ones that they already know.
Minke:
But sometimes that's really difficult. I'm speaking out of my own experience now (laughing) given that I've trained with you for like two years and I still feel like I'm a beginner. Well no, I'm not a beginner but I still sometimes wish that I could apply some principles better. I try every day to be a little bit better than I was the day before. Have you got any advice for people aside from telling them to just go out and train their dog? Are there any tips and tricks you would like to share with people that really want to get better, do better, but that are not having an easy time yet in getting there if you know what I mean?
Jens:
I know what you mean. I am boring and I honestly think that the most important trait you need to have as a dog trainer is discipline. I know it sounds really boring (laughing)
Minke:
No, I think I agree.
Patrick:
I think you hit it right on the head.
Minke:
Yes, definitely.
Jens:
You need to be very loyal to the end goal. You can make a plan and write it down on a computer or on a paper, or have it clear in your head from your experience training many dogs. It doesn't matter but you need to put in the se sessions many times, every day. I think that many people underestimate the time that goes into a working dog, whether it's a police or a military dog.
Minke:
Definitely.
Patrick:
That’s true.
Jens: T
hat is something different than going to your club and training three or four evenings a week. Here we're talking about multiple training sessions that need to be structured every day, all year round, whether it's raining or snowing or is there is sunshine. I think that is an important part, and what makes you disciplined is motivation. The military for example, has a really good system for selecting people that have an internal motivation, those who do not need other people to pat them on the back and tell them that they’re doing well. So, if you can select people that have an internal motivation, who will work even if no one is saying that they’re doing a good job for many weeks, then it's easy for a such a person to be disciplined. Actually, those are also that's the kind of dogs we are trying to select. My experience is that when I train handlers or units that have also been selected because they have these traits, they reach fantastic results within very short time.
It’s not that they have any previous theoretical knowledge about dog training or that they are extremely interested either, but they are highly motivated and very disciplined. Then they produce good results.
Minke:
I often heard you say the word dutiful, to be dutiful, but is there a difference between being dutiful and having discipline?
Jens:
Yes. Dutiful is a word that I first came across in a study from the U.S marine where they evaluated specific traits in dog handlers and explosive detection dogs. There they found that the most important human trait was dutifulness which they defined as a combination of being disciplined and loyal to the end goal. In this case the end goal was to produce a deployable explosive detection dog. So, you need to be disciplined, but the discipline needs to have a direction or a focus and if that’s the case then you're dutiful not only disciplined (laughing).
Minke:
Yes, exactly. I was about to say just that.
Patrick:
Besides the theoretical parts which we discussed I also found it to be very beneficial in training new handlers to videotape their behaviour. Do you also use that in your training?
Jens:
Yes, I film myself a lot. I don't want to sound too bombastic, but I am probably one of the most important assets in the dog's life, so what I do, where I point my hand, where my body is pointing or where I am looking, the dog is likely to pick up on that. Especially in training that often leads to the dog finding what I want it to find or doing what I want it to do. What I see in the video tapes is if I actually guide the dog for example, in the early stages of detection training, then there is a high likelihood that the dog is not associating the reinforcement with finding the target odor. It's associating the reinforcement with paying attention to my hand movement.
Patrick:
Yes, you're talking about clever Hans.
Jens:
Yes, and then my hand movement is overshadowing what I want the dog to learn. When you're standing there you may be cold, you may be frustrated because you haven't had the success that you should have and you're annoyed because your children are not taking the dishes out of the dishwasher... (chuckling). All of those things influence you. Then it's difficult to pay attention to all the things you do with your hands and what the dog is paying attention to. Afterwards, when you sit with a beer and everything is nice; you have cleaned up after your children and you can relax and look at the video’s you made and see all the things you're doing that risk having impact on the dog's learning. Then you can hopefully learn and do less of that the next time.
Patrick:
Yes. I strongly believe that a lot of people are not aware of the fact that a dog communicates about 90 percent non-verbally. They don't always understand all the cue’s we are giving them or all the body movements that we make. That's something that I think people have to take into account when they start training with their dogs.
Jens:
Yes, and it's so difficult when you're in the situation because the dog may have been doing some really good things before. Now it missed a little bit, maybe on the position in the transport, and you think: “Ehh, but he was so good before. I will give him the reward”. There and then it may feel like the right thing to do but when you see it afterwards you will curse yourself because it's unfair to the dog to reward it for a behaviour that you will later on try to get rid of. It would have been much better not to have rewarded the dog at that time, but it's so difficult when you're standing there because most people are … humans.
(laughter)
Minke:
Exactly that (laughing).
Patrick:
If you look at dog training… From my personal experience in the in the early 80’s when I started training we learned the dog to perform a certain behaviour to avoid getting beaten. That was just the way people trained at that time. I also used that system simply because of the fact that I did know a different way. Later on, as I was studying and talking to people, I tried to evolve and get better. I wanted to train dogs in a way where they knows that if they performs a certain behaviour they will get rewarded instead of them trying to prevent being punished. That shift in the way people train is something that has taken place in the let's say the last 50 to 20 years. How do you think dog training will evolve down the line? What do you think is going to be different if we would meet at the same table in 15 years’ time? When we will all be walking with crutches and we’ll be completely gray and everything.
(laughter)
Minke:
No way!
Patrick:
What do you think will change in the course of the following years? What you see in sport for instance, is that certain people are not open to new ways of training at all. They have always trained in a certain way and intend to keep doing it that way because that is what has delivered them success in the past. On the other hand, there are also people who want to try to train differently which is possible now that we have more knowledge about dog training. Things can be done differently nowadays. How do you think this is going to evolve in the next 10 to 15 years?
Minke:
That's a big question.
Patrick:
Yes, it’s a big question.
(laughter)
Jens:
I think there have not been much new insights the last 100 years with regard to the science behind dog training. It's still based on operant and classical conditioning. I don't expect that to change much within the next 50 years. So, the next 15 years it will remain the same I think, but I can also see that the culture around dog training is changing. Right now, there is a big movement which advocates that people should use more positive reinforcement and less positive punishment. I don't think that will lead to anything in particular, because from what I see in most working dog training it is mainly positive reinforcement and negative punishment that are being used. If you live in a country where you can use the electric collar you can also use negative reinforcement, but without an electric collar I would say that's difficult to apply a lot in the working dog training. We can do it with leash pressure or a vibrating collar, but it's a very limited number of behaviours that you can train effectively using those tools. So, I think it will still be mainly negative punishment and positive punishment that are being used.
And maybe it will evolve in a way such that we get better at selecting dogs with several primary reinforcers. If I may just talk a little bit…? (laughing)
Patrick:
Yes, of course!
Minke:
You have about eight minutes (laughing).
Jens:
I will try (chuckling). So, the definition of a primary reinforcer is something the dog is born with the desire or a motivation to do. It can be water, food, reproduction or chasing something, grabbing it, having it in their mouth… And we can affect that, mainly through breeding. This is clearly visible in some lines of the Belgian Shepherd. An example of that can be seen here in Holland, where you’re breeding dogs who want to bite metal, not because they learned to do that but because they have a desire to do that from a very young age.
Patrick:
Some sort of genetic predisposition.
Jens:
Yes, that's what we're breeding for, because a primary reinforcer cannot be trained. It is something that is already present from birth. That's the definition of a primary reinforcer.
Patrick:
The dog decides.
Jens:
The genes decide. We cannot train that. And that's where I sometimes see commercials or advertisements about, especially in another continent (chuckling), where they claim to be able to build drive. What they mean by that is building the motivation for the primary reinforcer, and that's not really possible because the primary reinforcer is intrinsic. It's something you're born with. To varying degrees we can increase the motivation for it a bit by using deprivation. For instance, if you if I if you don't eat for 14 days, then you will be willing to eat Swedish sushi…
(laughter)
But if I offer it to you now you may say: “No, thank you”. So that way we can affect the primary reinforcer, but through training it's not really possible to affect it much. A secondary reinforcer is like the clicker; a signal that tells the animal that now the primary reinforcer is available. That is something that is trained. All the reinforcers that are trained are secondary reinforcers, and they cannot be trained if we don't have a primary reinforcer. If we have a dog that is crazy for the primary reinforcer, like the ball, then the positive reinforcement will be very powerful and the negative punishment if the dog does something that results in it not getting the ball will be the harshest punishment to the dog that someone can produce.
Patrick: I think
people underestimate the power of negative
punishment.
Jens:
Yes, I often find that, not ordinary people, but pet dog people underestimate that because if you train a pet dog, very often you don't have access to that tool because the dog is not selected to be crazy for one or more primary reinforcers. It isn't a tough punishment for the dog if it doesn't get the ball that it doesn't really want anyways. The crazier the dog is for the primary reinforcer, the more powerful the negative punishment is and the more effective can we be in our training. If we have dogs that are not really crazy about something, then I find we can have a need for negative reinforcement. The electric collar for example, may be an effective tool for the dogs in the middle range when it comes to drive. With the crazy high-drive dogs you don't need it because the negative punishment is so harsh to the dog already. In the dogs that have a low drive, that have little motivation for the primary reinforcer, you can make use of a third quadrant; the positive punishment, by using an electronic collar for example. As those kinds of tools get less and less accepted in society, I think the most effective thing we can do is to select for dogs that have an increasingly high motivation for the primary reinforcers, and as many as possible. Several breeders here in the Netherlands are great in that respect as they breed dogs that have a high desire for the for the iron pipe, for having articles. It's not specifically metal that they're bred to want to have in their mouth, it's any article. And they also have a desire to stay on the grip, to chase after things, and to want to have food. That’s great because then we have several primary reinforcers which will make our training more effective.
Minke:
I'm going to make a bridge to food.
Patrick:
Yes, good idea.
Minke:
Yes, we've been talking quite a bit, not about food but about very important reinforcers, of which food can be one. I think it's time to maybe think about having some primary reinforcer for ourselves. So, I want to thank you Jens, for this very nice and useful conversation, and I hope that you have been able to tell the people that will be listening to this about what gets your heart pumping a bit faster (laughing).
Jens:
We haven't been talking that much about beer…
Minke:
No (laughing), I was referring a little bit to dog training.
Patrick:
To be honest, I still got a lot of questions but we don't have time for them anymore…
Minke:
I'm sure if we ask Jens again he will be willing to talk to us once more. I have a feeling we will be seeing more of him in the future. He can't get rid of us anyway.
Patrick:
Nope.
Minke:
So, Jens, once more; thank you very much.
Patrick:
Thank you very much, also for the fact that we can always pick your brain. Thank you for your time and I hope to see you again soon.
Minke:
Okay, let's eat.
Patrick:
Let’s eat.
Jens:
Yes, thank you for having me here.
[Music]