Copyright of S.I.T Canine Training 2014
The Problems with Punishment
We should all be force free!
Why punishment (also known as behavioural correction) does not
work
Firstly I want to start by removing all ethical issues from the table. My aim here is to clearly explain why
punishment is not an effective way to train a dog using ONLY the principles of associative learning. I am
taking this stance as there are many trainers out there who do still support and advocate the use of
punishment in dog training. I want to explain to you the reader why these methods are faulty and why
trainers, which use punishment, lack the basic knowledge and understanding required to successfully
Produce a dog which is willing to learn new behaviours and happy to interact with people.
Context Dependent Learning
Firstly it is VERY important to understand that learning does not occur in a vacuum. By this I mean a dog
(or any animal) does not just learn that ‘sit’ is a cue to place my bum on the floor which in turn leads to
me getting a titbit, the dog also learns about the environment in which these associations occur. If you
take your dog to training classes and during those classes you ask your dog to sit and when they sit you
give them a titbit your dog is also learning that the venue (the village hall, local playing field – where ever
it is you go for your classes) is a good predictor that the behavioural association for sit is likely to occur.
Let me give you some more examples of context dependent learning which have been carried out in the
laboratory. When Pavlov discovered classical conditioning the conditioned response was context
dependent. Pavlov was actually investigating processes within the digestive system when he quite
literally stumbled across classical conditioning. As part of his research Pavlov removed dogs from their
kennels, brought them into the lab, placed them onto a stand and strapped them in place. It was only
when the dogs were strapped into place that he began his experiments which involved feeding the dogs
while in this stand. He noticed that the dogs began to salivate in response to the to the lab technicians
when in this situation. The dogs had learnt that when I am removed from my kennel, taken into the lab,
placed in and strapped to the stand a lab technician will come in and give me a bowl of food. The simple
act of being placed in the stand became a good predictor that food would arrive very soon. The dogs did
not respond to the technicians by salivating when they walked past their kennels.
Other experiments involved strapping dogs into a stand and injecting them with adrenaline. Adrenaline is
a neuro-chemical which plays a key role in our fight or flight response. When we have a rush of
adrenaline our heart rates increase to supply our muscles with more oxygen to prepare us to fight or run.
Over a number of trails were dogs were taken from their kennels, brought into the lab, strapped into the
stand and injected with adrenaline. Throughout the trials the dogs were also attached to heart monitors to
measure their heart rates. It was discovered their heart rates decreased when placed into the stand in
preparation for the adrenaline injection and when injected with a saline solution (which acted as a control)
the dogs’ heart rates remained much lower than when they received the adrenaline injection.
What these two examples demonstrate is that dogs don’t just learn about simple associations (when I
hear the word sit I should perform the behaviour of sitting) they also learn about the immediate
environment (or context) in which these associations regularly occur. To give a human example; if I were
to blindfold you and take you out in the car, after a while the car stops, I ask you to get out and lead you
into a building. I ask you to sit down and when I remove that blindfold you find yourself in a dentist’s
waiting room. The room provides you with enough contextual information for you to know not only where
you are but also what is going to happen next.
Now let’s look at the effects of using punishment in relation to context dependent learning. So your dog
is not behaving as you would like them to (the behaviour in question at this stage is irrelevant). You take
them to a trainer who advocates the use of punishment and when your dog arrives at the training centre
they get a kick to the groin, shocked by an e-collar, chocked with a chain etc. The dog is not only
learning about getting punished in relation to certain behaviours but is also learning about the
environment where that punishment takes place. After a number of sessions your dog is going to learn
that arriving at the venue is a good predictor of physical pain. This is going to make any dog fearful.
Imagine that you have been accepted to take part in a new game show. The prize (should you win) is
£1,000,000. You are highly motivated to do well and win. To win all you need to do is read, remember
and recall 20 randomly generated words from a list which is projected on to the wall of a room. Easy
right! You’ll be a millionaire before the day it out. But obviously it is not as easy as all that, this is national
TV! Before you get presented with this list of words you are placed into the room (in which you will
eventually learn these words) and while you thought ‘What are you most phobic of?’ was an odd question
on the application form you answered truthfully with spiders. In fact you are totally terrified of spiders.
But until now you gave no more thought to that question on the application form as you were so over the
moon to be accepted on to the show and given the chance to win all this lovely tax free cash.
So the lights on set go up and you are strapped into a chair so you can’t even move your head.
Suddenly from above a trap door opens directly above your head and hundreds of thousands of spiders
from small to HUGE land on you. This is giving me the willies even writing it! As part of the show you
have to be in this room a number of times and each time at some point the trap door opens and the
spiders fall on you. Sometimes this will happen once, sometimes twice, sometimes only small spiders,
sometimes only big spiders. You never know when and you never know for how long, but you do know
they are going to fall. You know this room and being strapped in this chair is going to lead to spiders
falling on to you, crawling on you, and you can do nothing to escape. Now read, remember and recall
those 20 words to win your money! This is going to be a LOT harder simply because you are scared of
the environment in which you have been placed. You have no idea if or when those spiders will fall on
you while you are reading the words. You are freaking every time you think you can feel something on
your head and panic when you think you caught slight of an 8 legged beast on your shoulder. Your focus
is not going to be fully devoted to reading and remembering those words, your heart rate will be going
through the roof, your system will be flooded with adrenaline and you will quite frankly be pooping your
pants. All because you have contextually learnt that being in that room and being strapped in that chair is
a good predictor of spiders falling on you.
The point I am trying to make here is that learning is negatively impacted when an organism (dog or
human) is fearful. A dog which has learnt that the training venue and in some cases even the act of
training is a good predictor of physical punishment is going to become fearful and fear reduces our ability
to learn new behaviours as we enter into a psychological state which is not conducive to learning. So
trying to train a dog to perform a new and novel behaviour in an environment or situation which predicts
punishment is not going to be anywhere near as successful as training a dog to perform a new and novel
behaviour in an environment which predicts food, toys, and a tummy rub.
Dogs don’t understand punishment like people do
We can rationalise with our fellow humans as to why we are punishing them. Someone who commits a
crime and is found guilty of that crime has their sentence explained to them in a court of law by a judge.
The criminal understands (even if they are innocent) what they have been charged with and the
punishment they are facing. We can’t explain that to a dog!
When it comes to punishing a dog we can NEVER punish after the fact and make the dog understand
what they have been charged with. A dog lives in the moment, while it is obvious they can conceive of
past, present and future it is not in the same way humans do. As a human we know we have tooth ache,
we can remember the last time we had tooth ache, we can remember the trip to the dentist and we can
dread having to make that call to book and appointment to treat this current tooth ache. When we have
booked that appointment we can dread the day of the appointment arriving. Our ability to link (and
rationalise) past, present, and future does not only extent to dreading the bad but also anticipating the
good, we can be excited about our summer holiday, anticipate the enjoyment of booking it, look forward
to getting on the plane and imagine ourselves lying on the beach soaking in the sun.
When dogs think about past, present, and future they do so in the terms of associative learning. A dog
can learn to associate a past event with a current situation and expect the same outcome to occur in the
near future. For many years I have worked as a dog groomer and most dogs don’t like visiting the
groomer. My clients would always marvel at their dogs ability to ‘know’ that today was the day they be
subjected to shampoo, blow dryers and canine cologne. I always seemed to disappoint the owners when
I explained that the behaviour was simply explained via associative learning and not a psychic
phenomenon. In all these cases owners did something slightly differently the day before the
appointment in most cases this was to give their dogs a thorough comb through (I trained my clients well)
and in other cases this was simply the way they spoke to their dogs; i.e. ‘you’re having a bath tomorrow’
in a tone of voice which was only reserved for the visit to me. Over time the dog learnt that these signals
(the brush or tone of voice) where linked to the visit to me the next day. Simple associative learning. The
dog became sensitive to these signals (or stimuli) and as a result was not so willing to have their lead on
the next day as they could anticipate the walk (or car journey) to my house and going into that bath tub!
So let’s explore what a dog learns when we punish them after coming home to find a chewed shoe. The
dog has already wrecked the shoe and you arrive home the deed is done. You find the offending article
on the kitchen floor. You pick up the shoe and smack your dog over the head with it, scream at them,
and chuck them outside.
The dog has learnt…..owner came home, I was hit with an object, it hurt, they were aggressive towards
me, I was scared, I was separated (again).
The next day the dog has pulled out a load of your DVD’s and chewed a number of them beyond
recognition. You pick up the DVD’s hit the dog over the head, scream at them, chuck them outside.
The dog has learnt…..owner came home, I was hit with an object, it hurt, they were aggressive towards
me, I was scared, I was separated (again).
The next day the dog has ripped up the lino, you kick the dog, scream at them, chuck them outside.
The dog has learnt…..owner came home, I was kicked, it hurt, they were aggressive towards me, I was
scared, I was separated (again).
Today you come home and your dog is hiding under the table, refusing to give you eye contact, licking
their lips, yawning and holding their head very low to the ground. And your front room looks like a bomb
exploded in your sofa, foam everywhere and all that remains is a timber frame and some springs.
Is the dog displaying guilt? Have they learnt that chewing was a bad thing and now feel remorse for their
sofa chewing? NO!
The dog is under that table hiding from you, he or she is displaying their fear through their body language
and asking you to calm down and not to hurt them. YOUR DOG IS SCARED AND THEY ARE DOING
EVERYTHING IN THEIR POWER TO TELL YOU THEY ARE SCARED.
This dog has learnt that when you come home you are going to hit them, shout at them and chuck them
outside. The chewing has escalated because this dog is now so worried about your impending arrival
home they have become filled with adrenaline and the only release they have is to chew something.
In some cases a series of escalating violence towards a dog which chews when home alone may seem
to have stopped the problem. The dog no longer chews but has the dog really learnt not to chew, of
course not. This is known as learned helplessness. The dog is now so scared of a beating and is so sure
that a beating is coming and knows that there is nothing they can do to avoid this beating they simply
shut down and submit to it.
Learned helplessness is an established and recognised principle in both non-human and human animals.
In the original set of experiments two groups of dogs were placed into two different experimental
conditions. Group A (let’s call them) where placed into a conditioning box which has a button the dog
could press with their paw or their nose. The floor of this box was ridged up to provide a foot shock to the
dog – this hurts! These dogs could press the button in the box to stop the shock. In group B the dogs
were placed in an identical box with only one difference - the button did nothing to stop the shocks. Both
group A and group B were trained prior to the experiment that the button would stop the shock and dogs
from both groups were happily pressing the button in response to a stimulus (say a light comes on) to
stop a shock being delivered 10 seconds after the stimulus is presented.
Now dogs in group B where placed into a situation where the button became ineffective. Pressing the
button would never stop the shock. At first these dog frantically pressed the button when presented with
the light. This button pushing behaviour became more urgent as the experiment continued (just like the
chewing became more frantic) until eventually the dogs simply gave up and led down in the box and
accepted the shocks without so much as a whimper. They had learned they were helpless in this
situation. There was nothing they could do to escape and receiving a painful shock was inevitable.
So we have established that physically punishing a dog after the deed is done is not going to teach the
dog the relationship between behaviour and punishment. But how about if we punish them during the act
itself, this is where most trainers which advocate the use of punishment.
If a dog pulls on a lead a punisher has two options, they can employ the use of a training tool which
delivers physical pain i.e. a choke or prong collar or an e-collar, or the trainer can cause the physical pain
themselves with a kick to the groin.
Once again it is important to remember here that learning does not occur in a vacuum and we cannot
explain to a dog why we are punishing them. So the trainer takes the dog for a walk when the dog pulls a
physical punishment is received.
Dog learns; when I walk like this it hurts
So the naturally tries to avoid this pain by walking in a different way, eventually after much trial and error
learning the dog learns the behaviour which leads to no pain – walking to heel without pulling. But what
else has the dog learnt?
If the trainer is using a collar the dog learns it is the collar which causes the pain and they only need to
perform the heeling behaviour when the collar is being warn – context dependent learning. The owner is
now 100% dependent on the training tool. If the trainer is using the kick to the groin the dog may show
better learning but again the dog can learn about who kicks and who doesn’t. But in both cases the dog
has not actually learnt how to walk properly only how to avoid being hurt. There is a distinction here.
When we use punishment we have taught the dog what NOT to do and when we use positive reward
based training we are teaching the dog what we WANT them to do.
Fear is a powerful motivator
While fear is a powerful motivator it only motivates an animal to find a way to avoid the pain. In most
cases the best way for an animal to avoid pain is to steer clear of the individual which inflicts it. If an
individual continually punishes a dog the dog’s best option is to avoid this person and protect themselves
should this person come to close to them. And in this case a dog’s only means of personal protection is
aggression. A dog which on the surface at least, seems to be highly motivated to learn when faced with
punishment is a dog which is ultimately going to learn that they need to be wary of humans, as humans
inflict pain. Fear is the most common cause of aggression in dogs and if we use physical punishment to
train our dogs we run a serious risk of teaching our dogs that aggression is the best and most valuable
behaviour. On the flip side we teach dogs that they cannot avoid punishment and pain and this leads to
learned helplessness and as a result successful learning is out of the question.
Punishing the fearful dog
Many behavioural problems are a result of fear. Having worked with many fearful dogs I can assure you
that punishment is only going to make the situation much worse. The only form of ‘punishment’ that I use
with my own dogs is water squirted from a spray bottle. I use this to discourage unwanted behaviour –
mainly barking. Now while my most of my dogs don’t like Mr. Squirty and simply showing them the bottle
is enough to stop the behaviour I am trying to stop even a simple squirt of water is too much for a fearful
dog.
When we are scared of something and with some dogs they can be fearful of everything, any negative
consequence, even something as harmless as a squirt of water is going to severely set back any
progress. Let me give you an example, if you are frightened by something and when I present you with
that something your heart rate will increase, and you will enter into a flight, fight or freeze response. If at
the same time I squirt you with water you are getting a double negative, -1 + -1 = -2. A negative plus a
negative always equals a greater negative. When you are very frightened even something harmless is
enough to increase your levels of fear. Just think about walking through the woods in the dead of night
having just watched The Blair Witch Project. Imagine you were spooked by the movie and you are taking
this walk alone through the woods as part of a dare. You are on edge and feeling very anxious.
Suddenly an owl hoots and takes off from a tree just above you, if your levels of fear are high enough this
alone is enough to send you screaming from the woods. Owls are not dangerous and when you are back
in the house and in a rational state of mind you know this and you will laugh at yourself for running from
an owl. But is just goes to show that when in a fearful state of mind even something as harmless as an
owl is enough to send you into a full blown fear response. When working with a very fearful dog, even a
slight movement is enough to send that dog running for cover and it takes months to rehabilitate such
dogs. In such cases ANY form of punishment will only increase fear and dramatically set back progress.
Other association’s animals build in relation to being punished
In most cases punishment is employed as a means of stopping an unwanted behaviour and in a lot of
cases it is used to supposedly deter aggressive behaviour. As most aggression is the result of fear
punishment is only going to increase fear and uncertainty. If you have a dog which is reactive towards
other dogs they have a negative reaction when they encounter another dog. The approaching dog is a
negative stimulus and as a result the dog in question responds with aggression. If we punish this dog
then not only are they having a negative reaction to the approaching dog but they are also receiving a
negative consequence from the handler. This is a double negative and a double negative will never
result in a positive. Punishment may serve to stop the aggressive behaviour towards other dogs, but
what has the dog really learnt?
If we physically punish a dog when they start to display aggression and manage to stop this behaviour
the dog in question has not learnt that the approaching dog is a good thing, in fact quite the opposite.
The approaching dog which was already a negative stimulus is now also a signal for physical pain. If I
gave you a firm kick to the groin every time I took you to the dentist would you like the dentist more? You
might well learn to go in without a fuss but your feelings towards the dentist would certainly not have
become positive. In such cases we have created a dog which sees an approaching dog as a very bad
thing, by physically hurting the dog we have created a dog which is even more distressed when
presented with another dog and all the dog has learnt is to supress their aggression due to fear of
physical punishment.
Rather than creating a well-trained dog we have potentially created a very dangerous animal. A dog
which learns to suppress aggression has also learnt to suppress all body language which signals to the
handler (and other dogs) that the dog is uncomfortable. This means the dog’s last resort, the bite, is now
the dog’s first response. This is a dog which will bite without warning and will bite what is closest to
them, usually the handler! It is also the case that such dogs don’t just bite, if the aggression is
suppressed deeply enough when this dog finally snaps (which they do) it will attack. Rather than a dog
which was reactive towards other dogs we now have a dog which will attack without warning all because
physical punishment was used to ‘correct’ unwanted behaviour. This of course is worst case scenario but
believe me when I say this does happen.
I have heard many horror stories which have resulted from people choosing to use physical punishment
in replacement of force free positive reward training. In some cases they are the result of a poorly timed
accident and in other cases it is the result of the tools used. In the first instance three dogs lost their lives
as the result of the use of an e-collar. E-collars are a devise which is placed on to a collar and emits an
electric shock to the throat. A lady was having trouble with recall with her two young German Shepards.
Both dogs were wearing an e-collar and were off lead during a walk in the park. One of the dogs had a
poo and as the lady bent down to pick it up she accidentally pressed the button on the handset and both
dogs received a large shock. Just at the time this shock was delivered a Jack Russell Terrier ran past the
two dogs. The lady finished her walk without incidence and returned home. The next day she took her
dogs out for a walk and the same jack Russell ran past them. Both dogs took off after the JRT and killed
it. Neither of these dogs had previously shown aggression towards other dogs but the shock they had
received the day before had been associated with the JRT and the dogs reacted in the only way which
made sense to them. Not only did the JRT loose his life but both GSDs were put to sleep later that day.
The next example comes from America were the prong collar is a popular ‘training aid’ the prong collar is
a chain collar which has sharp barbs on the inside which produce server pain when a dog pulls on a lead.
In this case the dog (a large cross breed) was being walked in the dog park, being very careful not to put
a foot wrong. Once again this dog had never shown aggression towards other dogs but simply had poor
recall and pulled on the lead. During the walk another dog rushed over to the on lead dog with the desire
to initiate a play session. As a result the dog on the lead rushed forward and got the full force of the
prong collar in the throat. Once again the dog in question associated the sudden pain with the
approaching dog and lunged into an aggressive attack. Thankfully the dog on the receiving end of the
attack escaped avoiding serious injury but the dog wearing the prong collar needed treatment for the
wounds caused purely from the collar. The owner in question had been advised to use the collar by a
‘professional’ dog trainer and had no idea of the damage these collar can do. Now the owner has a dog
which not only pulls on the lead and has poor recall but the dog is also now highly reactive around other
dogs. Once again the results of using physical punishment instead of force free positive training.
While these examples are extreme they are genuine examples of why physical punishment can be so
dangerous. In both examples the dogs in question could have been trained using force free positive
reward training, in fact every dog can be trained using force free positive reward training. And had these
methods been employed, rather than restoring to physical punishment, the above disasters would have
been completely avoided.
Conclusion
By removing the ethical debate about whether or not it is morally right or wrong to punish, and in doing so
inflict pain on another living creature, I have removed the emotive part of the argument. I wanted to do
this as simply because there are so many trainers which still believe that punishing a dog is morally
acceptable. While I don’t believe in physically hurting any animal basing an argument on whether or not
something is ethical only leads to the fling back and for of statements based on personal beliefs and
opinions. What I wanted to demonstrate is the negative effect punishment has on willingness to learn,
willingness to interact with humans which punish and the overall learning rate and learning ability of the
animal in question. The goal of any dog trainer, whether they believe in punishment or not, is to
effectively teach a dog perform a behaviour X in response to stimulus Y. To be a ‘good’ dog trainer and
do your job properly you want to enhance learning not decrease it. It is my argument that any trainer
which still feels that physically punishing a dog is going to lead to a dog which is willing and eager to
learn is kidding themselves. By simply using the scientifically established principles of animal learning I
hope that I have clearly demonstrated the negative effects punishment has on the overall behaviour of
the dog in question and I hope that you will now think twice before copying the methods of trainers which
advocate the use of ‘behavioural correction’.