It was whilst on holiday in Thailand earlier this year that something gave me the biggest demonstration of how positive reinforcement really does work wonders. I'll try & cut a long story down as best I can to give you the full effect. There's probably a whole Blog topic in itself with the elephant ride!
So, here we were sat on an enormous elephant where the positive reinforcement demonstration unfolded. A lot is made of 'passed down' training techniques, whether it be dogs, dolphins, elephants - even children. Culture & family history play a huge part in how people act & the training of animals is no different. The people in Thailand certainly haven't had a huge exposure to Jean Donaldson, Karen Pryor and the likes, so their training methods are very much based on what their parents & grandparents did.
I touched a little on the difference in culture between us & the Thai people when it comes to dogs generally, in my Blog 'Are We Really That Different?' - and maybe we too could learn something from their approach.
However, what happened during our elephant ride really hammered home the positive reinforcement theory.
When we sat on this huge elephant, the Thai man couldn't speak much English & my Thai was non-existent - other than please, thank you, hello etc - so we were having an extremely broken conversation as it was. We were sat in this basket whilst the Thai man "drove" the elephant. What occurred over the next 15 minutes will be something that will stick in my mind forever.
We'd purchased some pineapple which we could use to feed the elephants with once we'd finished our trek, so we sat with our basket of pineapple on our lap. Our Thai guide had in his lap a wooden stick with a metal hook at the end of it - this was to be his 'steering wheel'.
His method of driving the elephant was via the use of 'negative reinforcement' with a touch of 'positive punishment' at times too. He would continually tap on the elephant's head with this stick to get the elephant to do what he wanted him to do. When the elephant started moving where the Thai man wanted, the tapping stopped.
If the elephant didn't move he would resort to 'positive punishment' verging on a touch of abuse.
After a short while the elephant decided she just wanted to do her own thing - and at the size that she was, she could do exactly that! It did not matter how many more taps the Thai man gave her, if the elephant wanted to do something, quite frankly, you were not going to stop her!
I must point out, that the Thai man was a very happy man & did seem to actually care for this fantastic creature. He was sat there laughing at the elephant not wanting to move at one point & he just sat there stroking her MASSIVE head like she was his pet dog.
Despite this seemingly loving relationship, I was a little uncomfortable with the way in which things were going. It was a really hot day & I managed to use this as an excuse to cut short the ride - we'd paid for a 1 hour trek - and we decided to just let the elephant head back to her resting place. I said to the Thai man that the money wasn't an issue & we'd happily just cut the ride short. To be fair to him, I think he was in agreement too - I got the impression that he did genuinely care for the elephant.
Then came the moment of sheer delight! I somehow managed to suggest to the Thai man that it may perhaps be better if he jumped off the elephant & used our pineapple to encourage the elephant to follow him. Between our broken communication we seemed to ascertain that someone needed to sit on the elephant's head. So he invited me to sit on her head, Sarah was in the basket & the Thai man was now walking in front of the elephant with our pineapple.
Now I'm not sure if it was my suggestion or not - I somehow naively thought that surely someone would have thought of this idea before me? It's not really rocket science is it - or actually to the Thai man, was it? I honestly don't know and I'm certainly not going to try to claim that I educated the Thai man. Maybe it was a bit of luck & quite possibly he's done something like this before, but due to his 'hand me down' training, he prefers to stick to his 'old method' of driving the elephant. Like I say, I don't know.
However, what I do know is that the elephant soon got moving & we were happily taking the elephant back to her resting place - without the need for any negative punishment in order to 'drive' her.
I'm hoping the positive reinforcement over negative reinforcement & positive punishment was hopefully a great lesson to the Thai man! One can but hope.
More recently, I've seen a fair few comments & articles written on the use of punishment & aversives in dog training. It's a really contentious topic!
When it comes to dog training there's so much jargon banded around that for everyday dog owners who just want a happy & well mannered dog it can sometimes be a bit mind blowing.
But is it possible for dog trainers to really claim that they use only positive based methods? What happens when things go wrong? Is there a place or even a need for negative reinforcement, punishment & aversives in training?
Aversives, negative reinforcement, negative punishment, positive punishment, corrections - the list can go on & one, but what does it all mean?
Let's start off by looking at aversives, because it's probably one of the quickest things to cover within the 'Negative reinforcement & punishment topic'. A quick Wikipedia search gives the brief definition:-
"In psychology, aversives are unpleasant stimuli that induce changes in behavior through punishment"
When it comes to dog training, it's vital to understand that 'positive' & 'negative' do not have the same meaning as the mathematical or scientific terms - just as the word 'socialisation' doesn't have the same meaning for dog training as it does for humans. We covered this more in our 'Dog Socialisation' Blog.
Many people shy away from anything which has the words 'negative' or 'punishment' in them when it comes to dog training, simply due to the misunderstanding of the meaning of the word when used in a dog training context.
What is negative reinforcement?
Negative reinforcement is when the unpleasant sensation for the dog (the aversive) is turned off. Think of our elephant ride.
The Thai man continually tapped on the elephant's head to make her move. So when she moved forward, the tapping stopped. When the elephant stopped, the man started tapping.
The aversive (tapping on elephant's head) was removed when the elephant displayed the behavior the Thai man was looking for (elephant walking forward).
The reason the word 'negative' is used is because something is removed when the correct behavior is performed. (The reason we use positive reinforcement is because something is introduced when the correct behavior is performed i.e a treat is presented)
It really is simple . . . . .Negative = removal
Negative reinforcement = removal of aversive.
When is negative reinforcement used?
Negative reinforcement is used a hell of a lot in horse training, although as with dog training more people are now moving towards a more positive reinforcement based method. You want the horse to turn left so you pull on the left rein, when the horse turns & walks to the left you release the tension on the left rein. You've negatively reinforced the horse.
We use negative reinforcement all the time in day to day life. How many people give their children 'the look' in order for them to stop the unwanted behavior they are presenting? You have produced something unpleasant to your child ('the look') & upon ceasing the unwanted behaviour you remove 'the look' (negative reinforcement)
So let's consider the "I've never used negative methods in training" . . . .
How many people have pushed down on their dogs bottom when teaching them to perform the 'sit'?
Well, you have just negatively reinforced your dog. You have presented pressure on your dog's bottom (which is probably unpleasant for your dog) & he has performed the sit. Upon performing the sit you have released the pressure on his bottom. That is negative reinforcement.
The use of any kind of 'reinforcement' whether positive (giving treats) or negative (removing unpleasant) is to encourage & increase the desired behavior.
Many people claim to be 'positive only' trainers by saying that they always 'positively reinforce' the behavior afterwards, but the fact of the matter is, the dog has only performed the action because you have initially instigated negative reinforcement.
There's a very common phrase amongst 'positive' dog trainers "reward the desired behavior & ignore the unwanted behavior"
Is it simply effective enough to ignore little Rover's behavior when he is running off in the park? Can we simply ignore Bonnie barking & wait for the silence in order for us to reward the quiet behavior?
So where does punishment come into all of this?
There are times where dogs do need to be punished. If the dog is doing something which is highly dangerous we need to ensure that this dangerous behavior is reduced & eliminated.
If you mention the word 'punishment' in any form of dog training the first things that come into people's minds are check chains, E-collars & rolled up newspaper!
Very often the word 'correction' is used when discussing punishment because many people deem this more 'politically correct' & less harsh in terms of the mental images it creates. However, no matter what you want to call it, the use of punishment in dog training is designed to decrease the likelihood of unwanted behavior.
I don't really agree with the mantra that it is possible to train a dog without using any form of punishment. Everything we do with our pups is all about fun, games & enjoyment. We use clicker training & a huge amount of positive reinforcement with our pups & I always describe myself as a reward based dog trainer - reward 'based' being the key word.
Take this example. When Izzy is in the car crate & I pull up at the park she goes mad. She is 100% ready to fly out of the car & into action. So when I lift up the crate door she'd quite happily fly out of the car at 100mph. However, I do not want her to do this, so I ask her to sit & stay. If she sits & stays I open up the crate door & let her jump out.
If however, when I open the door half way shes trying to barge past & jump out of the crate, I shut the door & ask her sit & wait. The door can end up getting open & shut a few times until she waits for me to open to crate door fully & invite her to jump out.
By me closing the door on her whilst she is trying to barge & jump out, I'm using a form of punishment. I am negatively punishing her, because I am removing (negative = removal) her ability to be rewarded by the excitement of running around like a loony in the park
For someone to claim that as a 'reward based' trainer they do not use punishment suggests to me that they don't really understand the link between reinforcement & punishment.
So what are the differences in punishment?
Negative punishment is the removal of something the dog desires in order to decrease any unwanted behavior.
In the example above, I removed Izzy's ability to run around like a loony in order to decrease the barging & charging out of the car crate. Shock & horror - quite clearly I have used punishment in Izzy's training, I must be a 'bad trainer'.
It's important to really understand the use of negative punishment because in many instances people inadvertently use this form of punishment on their dog without even realising it.
How many people have called their dog back to them whilst their dog was enjoying a run around the park & quickly clipped on their lead? You've just negatively punished your dog. There they were happily running around & you have just removed their ability & desire to run freely in the park.
Another common example in the use of negative punishment is teaching dogs to walk on a lead loosely. How many people have used the technique of stopping every time your dog pulls too much on a lead? Everytime you stop, you are removing the dog's reward of walking forward.
To employ effective negative punishment in order to decrease the unwanted behavior you need to take away something the dog values highly.
In Izzy's case the desire to run around in the park is massive & therefore by stopping her doing this is a punishment.
Negative punishment is a hugely effective training method when coupled with positive reinforcement (clicker training). Our 'Clicker Training - Preface' Blog gives some more information on this subject.
The most common source of debate amongst dog people is the use of 'Positive Punishment' in dog training. There's not a day goes by where I don't see or hear a debate in relation to the use of Positive Punishment - but what does it actually mean?
Many people lump the word 'punishment' together, however as I highlighted above it's vitally important to be able to distinguish between Negative & Positive Punishment. As we said earlier, the word 'positive' doesn't always mean something good or enjoyable.
Positive is simply used because we are introducing something.
Positive Punishment is the introduction of an aversive (something unpleasant) to punish the dogs behavior. The term 'positive' is used because we are introducing something.
Negative = removal
Positive = introduce
Positive punishment is used in order to make the dogs unwanted behavior to decrease & the key to it's effectiveness is to ensure that the behavior is decreasing.
Examples of positive punishment include the use of E-collars, prong collars & smacking the dog.
This type of training is very much in the Cesar Milan mould - "do what is say or else". If the dog runs off & doesn't come back it gets shocked with an E-collar. If the dog tries to bite you, you smack it in the face.
It's these types of actions that have caused such a stir in the dog training world.
Positive punishment doesn't just equate to physical punishment though. Creating a loud rattle with a rattle bottle is an example of positive punishment. The dog barks so you shake a bottle full of stones to startle the dog.
The reason so much is spoken & written about punishment is because of the use of positive punishment such as E-collars & prong collars etc, and when they are used incorrectly the positive punishment equates to animal abuse.
Without doubt, one of the main causes of animal suffering during training is by people trying to adopt Cesar Milan's approach to dog training & doing it badly.
Dog training is all about timing and the reason many people prefer to use positive reinforcement is because of the consequences of getting your timing wrong. What's the worst that can happen with positive reinforcement? Your dog gets a few extra treats & ends up with an over enforced behavior - possibly getting fat!
With positive punishment, get your timing wrong & you're into the animal abuse territory.
There's so much to consider when utilising positive punishment & for punishment to have just a tiny affect on behavior there are a few conditions that need to be met, highlighted no better than in "Culture Clash by Jean Donaldson", in that she states that in order for punishment to be effective it needs to be:-
- Punishment must be immediate
- Punishment must be sufficiently aversive
- Punishment must follow each & every attempt at the unwanted behavior
Those are 3 MASSIVE rules to have to follow to the letter in order to make your punishment effective - and let's not forget if your punishment is not effective you're bordering on animal abuse!
There's a huge amount written about the use of positive punishment & the pitfalls of positive punishment - especially in comparison to positive reinforcement, however that's not to say that there isn't a place for it in dog training.
Providing the conditions above are met then positive punishment could potentially be a very effective form of training your dog - but only when done right.
There's such a fine line between positive punishment & abuse that many people (quite rightly) shy away from this form of training. When positive punishment is not done right the pitfalls are huge. Pat Miller, CPDT and author of "The Power of Positive Dog Training" produced the list of '12 Pitfalls of Positive Punishment' highlighting the following dangers of employing a positive punishment method:-
- You can cause physical pain/damage to your dog.
- It is difficult to gauge the appropriate intensity.
- The dog can develop a "punishment callous".
- The behavior may return when punishment stops.
- It is difficult to have perfect timing.
- It is difficult to be perfectly consistent.
- It can suppress desired behaviors; inhibit offered behaviors.
- It doesn't teach the dog what to do.
- The suppressive effect of the punisher is limited to the presence of the discriminative stimuli.
- It is rewarding to the punisher.
- It can damage the dog's confidence, trust in the trainer, relationship between dog and human.
- Violence begets violence.
Every element of positive punishment has an effect on a dog whether to 'correct' the dog's behavior or to effect the dog's relationship with you as an owner. Careful consideration needs to be taken before deciding on a positive punishment programme.
Things really do go wrong when dealing with positive punishment - just ask Cesar Millan . . .
What do you think - effective dog training or animal abuse?
If you think this Blog was a bit harsh on Cesar Millan - take another read of my Blog "Cesar Millan - A View Of Balance"