I was at the office today (our local park) doing a consultation with a man and his Labrador, working on recall and loose lead walking (connection). We have a really nice community at the park on weekday mornings, the professional dog walkers, the power walkers without dogs, the recreational cyclists and the dog people walking their own dogs.
There is a fella who I see who has two GSDs who he has trained to a high level from the brief glimpses I get of him (I see him several times over the month). He appears to use positive trainng methods from what I can see. He uses food to reinforce the behaviour which he likes and I’ve never seen him physically correct his dogs. When he asks the dogs to do something though, he’s not really asking, he’s telling.
This fella knows I’m a dog trainer and from reading his body language (he never says hello to anyone etc) I get the impression that’s he is trying to show me how it’s really done. Today (and every other dog he sees me), he put his dogs in a sit stay in the middle of the path and walked away. There were other dogs and people around walking past his dogs. The dogs were transfixed by him and then he called them, they raced towards him and then held a heel position targeting his hands for about 50m or so. They were then reinforced with food. Impressive? Yes, maybe. Ethical? I’m not so sure. If this had been in competition for a dog sport, then yes, it’s impressive. If it had been in preparation for that dog sport competition, then yes, cool also. My issues is that it’s done for the benefit of all those watching and at the dogs’ expense to boot.
For me it would be far more impressive if his dogs were walking with him off lead, moving forward in front of him, sniffing, moving with him when he walked on and behaving in an appropriate social manner with dogs and people. But they’re not. The appear to be automatons, with little choice and not allow ot behave like dogs (I’ve never seen them sniff when out)
Training with positive methods isn’t enough. We need to train with an ethical mindset too. Too much of a good thing can be a bad thing.
Nothing in Life is Free (NILIF), Learn to Earn. You may be familiar with these terms. They were devised, probably by well intention people, to allow you to use what your dog routinely gets in his life as ways to train him. The other thing some NILIF advocates is that your dog is behaving in a poor manner because he has unlimited access to resources, is spoiled or thinks he’s the “boss” or you are not the “pack leader”.
Now, some of what I say in this blog may seem contradictory as you read it, but it’s the reason why we are doing it, and the scientific reasons for it’s success/effectiveness which are important.
Following on from last week’s blog, I’ll define a couple of terms again, in simple terms (very simple)
1. Classical conditioning – learning by association.
2. Operant conditioning – learning by trial and error
3. Reinforcement – anything which causes the increase in the duration/frequency/intensity of a behaviour
4. Punishment – anything which causes the decrease in the duration/frequency/intensity of a behaviour
5. Positive/negative – adding(+)/subtracting(-) something from the dog’s environment. In training/learning, it doesn’t mean good or bad. We have positive and negative punishment and positve and negative reinforcement.
Proper understanding of the terms is important, because it allows us to have a greater understanding of what we are doing during training and why we are doing it.
So, to get back on topic. NILIF protocols advocate taking everything away from your dog if he isn’t behaving the way you want him to behave. The dog has no free access to food, toys or people for petting, playing etc. Further, every time your dog wants something, he has to earn it. You can’t just give your dog something, like a rub on the ear, just because he’s your pal and you want to. He has no access to toys he can play with himself, like chewing on a nylabone or kong, or playing with a squeaky plush toy to keep himself amused. Every morsel of food, every game you play with your dog, every time you want to pet her depends on her doing something you want her to do like sitting politely or coming when called.
The NILIF protocol says that when you decide you want to, you give the dog all the good things, and don’t when you don’t want to. What can happen here, is that if owners can’t be bothered or don’t have the time to interact with their dogs, they now have permission not to. It further recommends that if your dog approaches you for attention, ignore him, but when he walks away, then call him back and pet him then. That way, you are dictating access to you, not the other way around.
Now, if you haven’t already thought about this, this can lead to a very sad and frustrated dog. The day before you instigated this programme, your dog was fed regularly, had toys to play with, could come and say hi when he wanted to and you’d say hi back. Now his whole life has been changed, and as far as he’s concerned, not for the better and he can become depressed. The dog can also become demanding if this happens as he is trying desperately to get attention.
The other side of this, is that we can do very similar things to NILIF, but for different reasons and obtain hugely different results. These are all training programmes I’ve learned from the best trainers in the world such as Ian Dunbar, Jean Donaldson, Ken Ramirez and Kathy Sdao, among many others. Your dog has to eat. Hopefully, you like going out for walks, playing with your dog and giving them attention and petting. Your dog likes, and needs, these things too. So let’s use them for training. This is where it does become different.
1. Your dog’s food – food is a primary reinforcer. So let’s use it to train our dog. It reinforces behaviour, so when your dog sits, give him a piece of kibble so he’s more likely to do it again. Take your dog’s food out on a walk with you. Sit on a park bench and every time a person or dog walks past, give him a piece of food. This way he learns dogs and people means food comes out which makes him feel good and with enough practice, dog’s and people approaching makes him feel good with there being no food present (classical conditioning) and you are reinforcing him for paying attention to you when people or dogs are around (operant conditioning using positive reinforcement).
2. Your dog’s toys. Let him play with his own toys. He has his, I have mine. Mine are a couple of balls on ropes and a tug toy. He gets to play with his by chewing on them, maybe throwing them in the air and catching them. I let him play with mine, but only when I’m there and there are rules attached, but mine are more fun, beacsue I’m involved in the play. My toys move, I throw throw them and I tug on them. She gets to chase them and tug on them back. The rules are be careful with your teeth, and give them back when I ask you to. I don’t play with my dog’s toys but I do allow her to play with mine, because they are fun (for both of us) and because I can train her when playing.
Your dog enjoys playing tug or chasing a ball, you play tug or throw the ball, your dog will now enjoy playing with you (classical conditioning). You give your dog a game of tug or throw the ball when they come back to you you are now using positive reinforcement to train a recall.
There’s also loads you can do for free for your dog which is still training your dog and classically conditioning you to him in a beneficial way. Playing long games of tug in the garden with your dog, for no reason other than having fun, classically conditions your dog that you are a great, fun guy to be around, Further to that, now when you give your dog a short game of tug after a recall or a down stay, he has that long history of big tug games to refer to and it is much more powerful. Similarly, petting your dog for 40 minutes in the evening when you are sitting reading or watching TV means that clapping your dog when he does something you like now has more meaning.
One of the main differences between this approach and NILIF is that it doesn’t give you permission to do nothing. Your dog needs daily exercise and stimulation for a full life. The more we use our dogs food and play and attention for training, the better trained our dog will be. If you use a quarter of your dogs food for training, rather than all of it, you will only get a quarter of the benefit. That said, you don’t need to hand feed it all, put what you don’t use in a Kong or other stuffable toy and you are now teaching your dog to entertain herself, chew an appropriate item and be calm. The minimum we can do with this, if we’re motivated to is to give all daily food in toys or food puzzles so you are at least getting that benefit from it.
A programme like this leads to a well trained, sociable dog, and a truly deep relationship with you.
For many of us, we bring a dog into our lives to have a truly meaningful relationship with him/her. One of the easiest ways to do this is to build a huge history of both positive reinforcement for wanted behaviours and a massive history of classical conditioning.
In case you are new to my blog, or to dog training, I”ll offer a brief description of both. Positive reinforcement is where we add (+/positive i.e the mathematical term) something to the dog’s environment which makes behaviours more likely to occur in the future (reinforcement). This is a scientific term, and this is the meaning of it.
Classical conditioning, is linking one thing, to another, through repetition. It is learning by association. You pick up your dog’s leash and go out for a walk. Over time, picking up the leash in itself, causes the dog to become excited as he has learned that this predicts a walk, you have classically conditioned the leash to the feeling of excitement of the walk. I’ve linked a fuller explanation here
Positive reinforcement is used to make behaviour more likely to occur in the future. Think of it as paying into an investment account, which will pay future dividends. I’ve also linked a previous blog for a fuller explanation of this
When your dog doesn’t respond how we would like, it is for three reasons. The first is that your reinforcers are not reinforcing enough. A piece of dried kibble or a short pat on the head might not be enough reinforcement for a recall in the presence of other dogs. The second is that you don’t have a sufficient reinforcement history (you haven’t practiced enough) or haven’t practiced enough in those specific circumstances. The third reason is that your dog is having an off day for whatever reason, much like some days we go into work or into the gym and can’t get out of second gear, it’s just not happening on that day.
Now, using classical conditioning (CC) we can build a great bond with our dog. I will talk about the science of it here rather than any deeper spiritual or emotional connection with your animal as that is a matter of opinion and belief (I do believe we can form deep emotional bonds with our dogs). When a dog eats a piece of food, her internal chemistry changes and she feels a release of feel good hormones. If we feed our dog, then over time, we will cause that release of feel good hormones. The dog will now feel good just being with us. Conversely, if you are unpredictable around your dog, or shout at him, you will classically condition fearful emotions in your dog. The more you feed your dog, the more classical conditioning occurs. Instead of feeding all of your dog’s food from a bowl, set a portion of that food aside. Take one piece, smile and tell her she’s a good girl and then give her the kibble. Do this with all of the portion, one piece at a time. The more you do, the quicker and stronger these CC ties build.
Find our what your dog likes. A game of tug or a belly rub might be his favourite. Spend time doing this. Take five or ten minutes of the day, to just do this, not for any other reason other than to do it and spend time with him (make sure he actually does enjoy it or you will be classically conditioning unpleasant feelings with you rather than pleasant ones). If you are petting your dog, it is good to use the five second rule. Pet your dog for five seconds, and hold your hands back. If she moves back in for more petting, continue. If she moves away, she’s not enjoying it in that moment. Just by spending time with your dog doing something you both enjoy, teaches your dog, through classical conditioning, that you are good to be around.
The added benefit of this is that you can now use your bond to reinforce behaviour. When your dog does something you like, tell him “Good boy” and smile. This now has more meaning than it did before. Your dog does something else you like, you tell her “good girl” and play a short game of tug or give her a short ear scratch or belly rub. Because you have spend time investing in your relationship, this short reinforcement can be hugely reinforcing because you have taught your dog what it actually means rather than being something you kind of do for good behaviours i.e the dog has a reinforcement history of the long game of tug or the big belly rub.
The same principle goes for our family and friends too. We need to invest time with them, showing we care about them so that when we aren’t able to, they still know it. If we aren’t able to reinforce our dogs when they do something we want them to do, we at least have a huge history of really good stuff to fall back on and the behaviour and our relationship remains strong.
I’ve been trying to formulate my own system of training and behaviour modification for a while now from the information I’ve learned from a number of great dog people. I’ve come up with a three level training/behaviour modification system which hopefully is easily to understand and apply.
Level 1. This is the most empowering level for both the dog and the owner. The crux of it is in the following statement which I learned from Dr Susan Friedman – control the environment, not the dog, to allow him to make other choices, and then reinforce those choice to make them more likely to happen in the future. It is hugely empowering for our dogs to be in control of their own decisions. Now this doesn’t mean we allow our dogs to do whatever they want. Through careful control of the environment, we can limit or change the dog’s options so it now becomes easier for them to do something we want them to do (and difficult for them to do something we don’t want them to do) and when they do it, the wise use of reinforcers makes it more likely to re-occur.
If your dog jumps up on people when you stop to talk to them, we can stand on his lead so he can’t jump (give him enough slack that he can still stand or sit comfortably). We have now controlled his access to reinforcement (the person) and we can now reinforce (with food/praise) four feet on the floor. If we do this every time, the dog now has a large reinforcement history of four feet on the floor, and a tiny history of being reinforced for jumping. Not jumping is now much, much more likely.
Level 2. This is less empowering for the dog but the dog still has lots of choice. In this level, we use previously trained behaviours to give the dog instruction about what we want him to do. This is useful if our environmental control, as described in Level 1, gets away from us and the dog isn’t voluntarily able to make a decision which we are able to reinforce. We can now tell the dog what we want him to do.
To give you a couple of examples, say we are training our dog to remain within a certain radius of us. The dog is almost at the limit if the radius and suddenly a pigeon lands very close to her. The environment now changed and we couldn’t do anything about it. It is unreasonable to expect the dog to return to us without prompting at this stage unless we have trained for it, but what we can do is prompt the recall with a verbal cue which we have previously worked on. The dog now returns to us and we can reinforce this choice.
Another example would be if your dog is reactive to other dogs. At 100 feet, he might be able to willingly look at you, which you can reinforce. At 50 feet, your dog might not be able to look at you without prompting, but if we have taught an emergency U-turn which he knows well, we can ask him to perform the U-turn and reinforce it when he does.
The reason this level is less empowering for the dog is that he has less choice and you are starting to rely on more automatic responses. Although these responses have been taught using reinforcement, the dog is starting to perform them automatically and without conscious thought, so the question we could reasonably ask here is how much choice does the dog actually have?
I don’t think this is too much of an issue as long as we are either practicing the behaviour to maintain it’s reliability or we are using it because the alternative (running off or reacting aggressively in the previous examples) is a far less attractive option for us and potentially far more damaging to the dog.
As part of a society, and I include dogs in this, there are absolutely times when we need to do what we are told. We do this willingly in many cases but we are still doing it because we are either told to or asked to.
Level 3 – Physically moving our dog out of trouble. At level three, our dog is now unable to offer anything we want to reinforce, either voluntarily or with prompting. This is all about management and keeping the dog safe. At this point the environment has completely got away from us and we need to get our dog out of that situation. We are now controlling the dog. Examples would be if we are training a recall using a long line and the dog is running very quickly towards a busy road (think about how much environmental control we had in the first place!) we would stand on the line to stop the dog moving forward. In the case of the reactive dog, if another dog appeared round a street corner, and our dog reacted, we would physically move our dog away on a short leash to a non reactive distance.
When working on this, we can switch between levels 1 and 2 quite often. I use level 1 when I can, but also use level 2 frequently so we can use the available distractions as practice to keep previously learned behaviour strong and reliable. If we do this often, the ultimate aim is so that we don’t have to use level 3 and all.
The last thing to consider is training your dog to do many different behaviours. This allows the dog to offer previously learned behaviours during level 1 and gives you multiple options during level 2. If your dog doesn’t know how to do an emergency U-turn or recall, how can you ask for it? Also, different situations will call for different responses. Sometimes sitting may be more useful (when you are talking to someone in the street), other times having your dog to “leave”, “drop it” or “back up” might be better.
More training means more options. I’ll write more about this as I practice more. Until next time, happy dog training.
I worked with a client recently who has a one year old Landseer Newfoundland called Nero and an eight year old collie cross called Copper. Copper is one of these “once in a lifetime” dog’s who everyone wishes they could own but very few of us, unfortunately, ever have the pleasure to. He is very confident, experienced, has great social skills with other dogs and people and is a true gentleman. This would appear to have come naturally to him. Nero, will be Copper in a few year’s time. He needs some more practice and guidance, but he has been reared in much the same way as Copper, with steadiness and consistency. He is naturally a little more inclined to do his own thing than Copper is, and doesn’t always respond the way Jeannine, his owner, wants him to.
Jeannine is an airline pilot, I’ll explain why that’s relevant in moment. The issues with Nero are that he tends not to settle when people come into the house, as he wants to say hello and he has a habit of running up to people in the park (he doesn’t jump up) also to say hello. I explained to Jeannine why he is doing this, and when we changed how he is being reinforced, when he is being reinforced and by whom, we very quickly changed his behaviour to something which was more acceptable for us.
During the first session in the park, I did the majority of the handling, as Jeannine had her delightful baby girl (who clapped every time I said “good boy!” to Nero), so she could have a clear picture in her mind of what we were looking for. I reinforced Nero for calmer behaviour with food, praise and the continued opportunity to play and explore every time he did something I liked, which was either waiting, recalling or sitting. As the session continued, Jeannine gave the cues and I reinforced the behaviour.
About 40 minutes in, Jeannine observed that Nero was responding to me much better than he usually responded to her. My response to Jeannine was ” But I would expect that, because I don’t know how to fly a plane”. This is a strategy I’ve adopted to make clients feel better, as often the remark is made by clients that their dog responds better to me than to them. Initially, and until recently, I’ve put it down to the fact that I have handled and trained dogs for years, my timing is better than novice dog handlers, and I am better at reading body language and predicting dogs’ behaviour. This is true, but it’s not the whole story.
When a dog lives with a novice handler/trainer (and I would put very many owners into that category, as it has to do with training experience rather than ownership experience) the reinforcement history can be very sketchy. The dog’s behaviour is sometimes reinforced, some times not, sometimes punished and sometimes not. The result of this is that the dog either consistently tries behaviour which works for him (very often the ones we don’t want) and doesn’t reliably perform others (very often the ones we do want).
On the other hand, when I meet a client for the first time with their dog, I am starting with a clean slate and I write my own history with the dog, with no reference to anything else. This leads very quickly to the dog trusting that when I act in a certain way, and he responds, his behaviour will either be reinforced or not. Because there is no history, the dog has not point of reference other than the limited experience with me, which is very clear (hopefully).
The point of this is that I can learn to fly a plane, given enough time, commitment and effort. We can all try to wipe the slate clean with ours dogs, and start building a new relationship built on consistency and trust.
In my next blog, I’ll talk more about how to build this trust and the things we can do, inadvertently, to damage this trust, which can lead to further frustration in both human and the dog and a further breakdown of trust.
Last time, I talked about some of the problems we can have with fitting dog training into our busy lifestyles, balancing making progress with the time we have available and keeping the stress to our dogs to a minimum.
In this week’s blog, I’ll look at practical ways of fitting training into your life with your dog. In my experience, dogs tend to learn better from short, frequent training sessions rather than one or two long ones. Fitting in 5 two- or three-minute sessions a day yields more benefit than doing one long 30 minute session. That said, a 30 minute training session when you have time, can be further broken up into several mini sessions within that period, which I’ll also discuss.
Your dog’s daily walk is an ideal opportunity to fit in some training. Stopping every 25 of 30 paces or so, and asking for a behaviour such as sit, look at me, or a hand touch, means that over the course or a 30 minute walk, you can fit in several hundred training trials. This also has the added benefit of helping your dog to generalise the behaviour. Dogs don’t tend to generalise training very well, which is why you may find it very easy to ask your dog to perform a behaviour in your house or in the training class, but wonder why he has difficulty with the same request in a different scenario. When you are out on your walk, the environment, from the dog’s point of view, is constantly changing. The position of street furniture like lamp posts and bus shelters constantly changes as you are walking as do driveways, parked cars etc. Whereas we see the area we are walking around as our neighbourhood which is familiar, from the dog’s perspective, it is a changing environment. The more we ask a dog to generalise his behaviour, the easier it becomes for him.
One of the other plus sides we from training on our walk is that we can use the changing environment to reward the dog for doing what we request so we don’t have to rely on food rewards or toys. Dog’s like to sniff, explore, greet other dogs and people etc, so we can use this to reinforce behaviour. We walk along the street, stop and ask our dog to sit and when she does, we release her to go and sniff the lamp post as a reward.
If we use a portion or the dog’s daily food as training rewards, we can always have it handy in the house. This way, when we are waiting for the kettle to boil if we are making a cup of tea, we can do a dozen or so repetitions of a behaviour we are working on. Television is now able to be paused when we are watching it. This means that if we are watching one of our favourite shows, we can train during the adverts (commercial break to those outside the UK) and pause it as the show starts again. What this allows us to do it do two 6 or 7 minute sessions during an our long episode of CSI New York or Grey’s Anatomy, by pausing the show during it’s broadcast, train for a few minutes, and then fast forward through the next set of adverts/commercials to catch up the time we have lost during training. It also has the added benefit of using our precious time to our advantage and not have to sit through mind numbing adverts for mis-sold PPI. We have then spent an hour watching our favourite Tuesday night drama and training our dog for 12-15 minutes as well.
Finally, when we take our dog to the park, we can use the time to train our dog as well as exercise him, again using the same life rewards such as running off leash and sniffing etc to reinforce our training. When you bring the ball out, ask the dog to sit, lie down, bark, touch your hand etc before you throw the ball. When you see another dog approaching, recall your dog, and then release him to go and say hello as a reward. Using this method, we can fit in dozens of training opportunities into a 45 minute of leash walk in the park.
When working with dogs, either in training or behaviour modification, we will often notice that there is a level at which a dog will respond or not respond to our request or to something in the environment. You may have heard this referred to as “threshold” and I’ll discuss it a little further here.
For dogs who react to other dogs, I often like to think in terms of building blocks. Say, for the purposes of this illustration, your dog needs to get to a value of 20 before he barks, lunges, pulls aggressively on the lead etc when he sees another dog. The other dog might need to be thirty feet away before your dog reacts, this 30 feet would have a value of 20, so your dog reacts. So you increase your distance to 50 feet, this may cause the dog’s reaction threshold to drop to 15 but another dog arrives. This other dog might represent 10 points to your dog, which puts him up at 25, so he reacts again.
Each time your dog reacts, stress hormones are released into his body. It takes time for these stress hormones to return to their normal level. So, your dog, with a normal threshold from a calm state, takes 20 points before he reacts. He sees another dog at a distance of 50 feet, which only represents 5 points, and appears to handle the situation very well. The other dog disappears from view and you continue on your walk. A few minutes later, another dog appears. This dog is 35 feet away (close to your dogs normal reaction distance of 30 feet). In this case your dog is already at 5 points from the previous dog from a few minutes ago. The dog at 35 feet represents 17 points, so your dog is now at 22 points and he reacts.
Examples of factors which add point are
1. distance decreasing between your dog and the other dog
2. the other dog staring or looking directly at your dog
3.the other dog standing square on to your dog
4. more than one dog
5. the other dog moving as opposed to standing still – faster movement from the other dog usually means more points
6. your dog being in a higher state of arousal from previous interactions with dogs within a short space of time
Conversely, examples of factors which reduce points are
1. greater distance between your dog and the other dog
2. the other dog offering more social body language such as averting his gaze or turning side on
3. the other dog moving more slowly or standing still
4. the dog moving away
The above are examples and are not an exhaustive list and every dog is different. Another application of this, this time for training, is when training a recall. The three Ds of dog training enter here (distance, duration and distraction) Distance from you adds points, distance from the distraction such as play time with another dog will influence it, the amount of time since you last recalled your dog may be a factor as may the number of distractions in the environment or how the dog is feeling (tired, ill etc)
When working with your dog, whether in behaviour modification or training, and your dog does not perform to the level you would expect, this point system will usually be a big part of the reason. Examine what changed and see if you can play with the points to make it easier.