How do dogs learn?
Operant Conditioning -
Operant conditioning, also referred to as instrumental learning, was a concept developed by B. F Skinner. This teaches dogs that behaviour = consequence. (Please note that in science the word ‘consequence’ is often used, this means outcome or result. It’s not necessarily negative like when parents might tell their children to “take the bins out or they’ll be a consequence”. A consequence can be a positive or negative result when we talk about it in science!) Within operant conditioning there are four quadrants as outlined below;
Positive Reinforcement (R+)
Adding something desirable in order to increase behaviour. An example of this may be giving a dog a treat when they offer desirable behaviour.
Negative Reinforcement (R-)
Removing the use of an aversive in order to increase behaviour. An example of this may be applying pressure to a dog’s back end to place them into a sit, only to release the pressure once the dog has complied.
Positive Punishment (P+)
Adding something aversive in order to decrease behaviour. An example of this may be a leash correction, or yelling ‘no’.
Negative Punishment (P-)
Removing something desirable to decrease a behaviour. An example of this is ignoring a dog or turning your back when they jump on you.
The Hierarchy of Behaviour-Change Procedures (Dr. Susan Friedman, 2013) ranks these contingencies from least invasive to most intrusive.
Health, nutrition & physical setting: When engaging in any type of training it is first important to ensure the animal is well, emotionally and physically, by assessing the animal's health, nutrition and physical setting. This is why we often ask clients to complete a behaviour questionnaire prior to their booking so we can have a complete understanding of the animals lifestyle and how this may be affecting their behaviour.
Antecedent arrangements: Next the animal's environment is observed as the current set up may contribute to behaviour. Often an animal's environment can be modified or managed to both decrease and increase certain behaviours without intervention, or to assist with behaviour modification training.
Positive reinforcement: When a behaviour modification training plan is required, we first go down the route of using positive reinforcement and avoiding the use of other operant conditioning quadrants for reasons outlined further down.
Differential reinforcement of alternative behaviours: If required one can move further up the hierarchy reinforcing alternate behaviours to help the dog clearly identify what desirable and teach them alternative responses to triggering stimuli. This sets dog’s up for success and creates positive learning experiences rather than using methods that may lead to fallout (adverse result). There is a speed bump in front of this quadrant because this process involves a level of intervention.
Extinction, negative reinforcement and negative punishment: Extinction is when an animal comes to learn a certain behaviour leads to no consequence (outcome/result), and so the behaviour is less likely to occur or is exhibited with less intensity. Extinction, negative reinforcement and negative punishment should be avoided and implemented with caution, hence the little yellow warning sign on our visual, because when carried out incorrectly these methods could also result in fallout and can affect the psychological state of an animal.
Positive punishment: As an animal training practice, we highly recommend ceasing the use of positive punishment in training as it is often used incorrectly, and affects the welfare of dogs greatly. Science and history shows that the use of this is unnecessary. Smart animal handlers and trainers will set their animals' environment and learning experiences up in such a clever way that punishment should not be needed.
We believe it is beneficial to understand the Quadrants of Operant Learning because this knowledge helps you to then focus on ‘positive vibes only’ training.
What do you mean by ‘positive vibes only’ training?
Reasonable question - let us explain! We avoid the term ‘positive reinforcement only’ because as you can tell it would be impossible to only use this one quadrant! As a practice we believe in assessing and animals health, nutrition & physical setting and antecedent arrangements, and the use of positive reinforcement and sometimes even differential reinforcement of alternative behaviours as ranked above. We engage with dog’s in this way because;
It shows what behaviour you would prefer. If marking and reinforcing desirable behaviours does not strengthen it, that is, does not cause it to occur more frequently, than you will have much more success showing the dog what it is you would like them to do, marking and reinforcing this as it happens. Moving a dog into certain positions (also known as ‘guiding’ or ‘compulsion’ methods) doesn’t allow the dog to practice the behaviour itself and can often lead to other issues such as increased arousal or fear of handling. If we lure dogs, it helps them make the right choice voluntarily, and new learning occurs without fallout.
When positive punishment is used dogs can make unintentional negative associations. Dogs are constantly taking the world in like a sponge, and they’re often taking in more of the world than we realise! EXAMPLE - If you are walking down the street with your dog and they are pulling hard on the lead, you may give them a leash correction and yell ‘Oi!’ for pulling, however, if a cat enters the environment or comes into view at the same time you applied that correction, your dog may think the passing cat was the cause, not their pulling. Thus begins a developing phobia of cats.
It decreases the likelihood of dogs feeling the need to use aggression. Dogs often use aggression when they feel they have no other choice. If we interact with dogs in a force-free way they don’t feel the need to escalate to aggression as a way of communicating how worried, fearful or unsure they are.
It reduces stress for both the human and the dog. These training methods reduce stress for humans because you will see improvements in your pets behaviour as they clearly come to understand what your expectations of them are. It will also reduce stress in your dog because they will not feel as though they are ‘walking on eggshells’ because they are unsure as to what may trigger a negative reaction or punishment from their owner.
It helps the dog create a predictable relationship which leads to an anxious free life. When a dog is able to predict positive interactions with people, dogs and their environment this relieves worry. If we engage with our dogs in a positive way this builds their confidence in us creating a bond. Having a relationship like this with your dog will help in achieving your training goals.
It is the recommended approach to training by the Australian Veterinary Association and Association of Pet Dog Trainers, Aus (APDT) and more internationally recognised organisations! Dog training is a science and so we should follow the research! For more information, visit ava.com.au, apdt.com.au or iabbc.org
Classical Conditioning -
Classical conditioning, also referred to as pavlovian conditioning, was a concept developed by Ivan Pavlov. This teaches dogs that stimuli = consequence. (Remember ‘consequence’ means outcome or result.) You may have heard of clicker training or marker training before, this is how animal behaviourists most commonly use Pavlov’s developments. First we choose a marker for the animal we are working with. This may be a clicker or a word such as ‘yes’, ‘good’, or ‘nice’ - you want your word to be short, sharp, and ideally spoken in the same tone each time. We find in animal learning clickers have the best response, and so recommended that you at least first use a clicker and then use a verbal marker later on in your program once said animal has learned to learn. We next pair the chosen marker to a reinforcer, commonly food, so that the marker (stimulus) predicts food (consequence) reliably.
Once an animal is conditioned to the marker, we can then use this to help the animal clearly identify preferred behaviour. This is because we mark as the dog performs a desired behaviour, and reinforce that behaviour with food. This has the same psychological effect as when you open a can of cat food and your cat comes running into the kitchen. They heard the can opening, and know that sound results in their dinner being presented. It’s the same as when your phone sounds your text tone, you know there will be a text message on the display - you’ve been conditioned to know so!
That being said, if your text tone sounded enough times without any text message on the display, you would eventually desensitise to it and stop checking for a message. This is why whenever we sound our clicker, or say our verbal marker, we must follow through with presenting a food reward. We recommend eventually using both a clicker and a verbal marker. A clicker should always represent food, whereas a verbal marker should randomly rotate different reinforcers such as food, play, games, pats or verbal praise.
I feel like you use a lot of treats, will I have to use treats forever?
We’re certain you wouldn’t get up and go to work every day for nothing! In our human society money is our strongest reinforcer. Every dog is different, but treats are used in training because they are often an animal's strongest reinforcer, that is, they are highly desired. Initially we use treats so the animal understands what it is we would like them to do, commonly through luring. We know from human psychology that we are more likely to display behaviour when it is acknowledged or recognised. It’s important to remember when animals are training they are working for us. When dogs display wanted behaviours it should be reinforced so that the wanted behaviours are more likely to occur again, this makes it really clear to the dog what your expectations are.
Once the dog has a clear understanding of the wanted behaviour, you can start to slowly fade out the use of food. If this is done too quickly, dog’s can sometimes become confused, and feel that they are no longer displaying the desired behaviour because it is suddenly not reinforced, and so the wanted behaviour ceases. This is called extinction. So you’d want to move to an intermittent schedule of reinforcement also referred to as random reinforcement. As it sounds, this is when you offer a reinforcer randomly and not every time the behaviour is performed. This should motivate dogs in the same way people can develop a gambling addiction - they continuously play the game for a chance to win, but the win has to happen every now and again to keep a person hooked. Win’s produce dopamine which is a hormone that makes us feel good. This becomes addictive, and increases motivation when dog’s are training!
Shouldn’t my dog offer me good behaviour just because I want them to?
With most dogs you can definitely get there! But this all comes down to the relationship you have with your dog. Great relationships are developed and take work, they don’t just happen. You can develop a positive, fun and motivating relationship with your dog through reinforcers, this can be food, play, games, verbal praise or pats. Each dog is different as to which reinforcer is most desired, and this will often change throughout their life. Commonly, as a generalisation, puppies are usually easily motivated by treats, however through the building of a great handler - dog relationship by the time they reach adolescence they may be greater motivated by games and play. By the time a dog reaches adulthood it may see pats or verbal praise as a highly valued reinforcer. At your job, you are far more likely to stay after hours and produce greater work for a boss who is appreciative of you and whose friendship you benefit from, than a boss who is blunt and doesn’t acknowledge your efforts. It all comes down to the relationship.
