Positive Training Explained
As I only use positive training techniques, I thought it was important that I laid out exactly what I mean by positive training here in this blog.
I hope you enjoy the article; I know it gets a bit “techy” but if you do have any questions, you can always drop me an email.
What is Positive Training?
Positive training is a type of training based on principles of operant conditioning (explained below). In positive training, the animal is allowed to make choices on their own in order to receive a reward. No pain or discomfort is used in this kind of training, the animals choose to behave correctly, and so they perform desirable behaviours more often. Unwanted behaviours go unrewarded and will begin to decrease in frequency.
Positive training rewards the animal for trying, and so the relationship between animal and handler is based on trust and partnership, instead of a dictatorship or dominance – a theory now outdated in relation to dogs.
Dominance Theory was based on studies performed on captive wolves (such as Rabb, Woolpy and Ginsburg’s 1967 study (1)) and then extended into our dogs. However, ‘dominant’ behaviour (usually resource guarding and fear based aggression) is rarely, if ever, seen in wild wolves. In addition, feral dogs do not display a much more fluid hierarchy when compared to wolves, challenging just how valid the theories based on a rigid social structure are (van Kerkove, 2004 (2))
What is Operant Conditioning?
Operant conditioning is a learning theory created by B.F. Skinner in 1938. It describes a method of learning conscious behaviour shared by all vertebrates in the animal kingdom.
Operant conditioning is split into four parts: positive reinforcement, negative reinforcement, positive punishment and negative punishment.
Positive and negative reinforcement both describe ways to increase the frequency of a particular behaviour. Positive reinforcement focuses on adding something pleasant (food, attention, toys) whilst negative reinforcement is removal of something unpleasant (often pressure, especially when training horses). Both positive and negative reinforcement can be used to shape an animal’s behaviour, providing either the reward or the release are given as soon as the desired behaviour is performed.
On the flip side, positive and negative punishment are used to decrease the occurrence of undesirable behaviours. Positive punishment is the addition of something unpleasant (pressure, hitting) whilst negative reinforcement is the removal of something pleasant (stopping stroking, removing access to food/games). Again, proper timing is needed to ensure the undesirable behaviours are targeted.
Often, reinforcement and punishment come in pairs. Positive reinforcement and negative punishment are usually used together, whilst negative reinforcement and positive punishment are also a pair. This is especially obvious in the negative reinforcement/positive punishment pair – you cannot take away something unpleasant (negative reinforcement) unless something unpleasant was added in the first place (positive punishment). However, in the positive reinforcement/negative punishment pair it can be less obvious.
Positive reinforcement is easy (food, praise etc.) but what most people don’t understand is that ignoring a behaviour is classed as negative punishment. As such, praising good behaviour and ignoring undesirable behaviours is a valid way to shape an animal into behaving in a specific way without needing to resort to unpleasant scenarios.
What about Pavlov’s Dogs?
Pavlov’s Dogs was an experiment used to describe classical conditioning, another form of learning used in the animal kingdom. In this experiment, Pavlov successfully conditioned dogs to expect food when a bell rang, and as such the dogs started to salivate.
Classical conditioning is used to shape innate (or instinctive) responses. As such, classical conditioning is used for changing emotional responses to a stimuli, whereas operant conditioning is used to alter the conscious response to the stimuli.
How does this fit into Positive Training?
Positive training uses parts of both classical and operant conditioning to create a fun, relaxed learning environment. As such the animals trained using this method are able to make correct choices without fear of unpleasant consequences for making a misinformed choice.
Positive training uses the principles of positive reinforcement, along with negative punishment, to shape any behaviour you wish. Positive reinforcement is used to reward the good choices ad attempts made by an animal, whilst negative punishment is used to allow the animal to understand that certain behaviours are no longer worth attempting as they do not achieve the desired results.
An example of how this all works is with a dog which jumps up (a common issue found by many dog owners). A traditional method of dealing with this is to knee the dog in the stomach. Not only does this cause pain in the dog, classical conditioning could make the dog fearful of people due to the association of humans and pain. This fear often comes out as an aggressive response, resulting in a vicious circle as punishment for the aggression happens, causing more fear and therefore more aggression.
In positive training, we focus on the behaviour we want the dog to exhibit instead. In general, my go to behaviour for a dog which jumps up is to make a sit behaviour a highly reinforced behaviour and as such a behaviour which the dog is likely to attempt in a number of situations.
When it comes to greeting people, the dog would then only be rewarded (treats/attention) for sitting at the feet of the person they are greeting (positive reinforcement). Any other behaviour (standing, jumping up, etc.) is ignored, with these behaviours resulting in immediate ceasing of attention given (negative punishment). As such, the dog learns that the best way to get what they want (attention) is to sit politely at the feet of people. The dog continues to have positive associations with people, and as such will be much less likely to display a fearful or aggressive response to people in other situations.
But my animal should just behave, shouldn’t it?
This is a common point which is often brought up. My response – horses/dogs/ animals behave exactly like horses/dogs/animals. They will perform behaviours which are rewarded and behaviours which are not rewarded will become extinct. It is up to us to decide which behaviours we want to see more of, and therefore reward.
‘But I don’t want to use food/stroking. They should just do it because I said so!’ This is another common statement. I always respond to this statement with ‘Do you get paid for your work?’ This is exactly the same principle. Evolution has allowed most species to have at least a small amount of selfishness, and as such organisms will only perform a behaviour if there is something in it for them. Rewarding behaviours allows animals to see the point in doing things we ask of them, making them more likely to occur in a number of situations.
Using high value reinforcers (food/toys) will not be needed forever, though. Whilst high value reinforcers should be provided sporadically for trained behaviours, once the behaviour is learnt, most repetitions can be reinforced with a quick stroke and a pleased voice. As the behaviour has previously been highly reinforced, classical conditioning means that the behaviours themselves release serotonin and can become self-rewarding. This results in a happy, confident animal who willingly responds to cues without fear of pain or confusion.
I hope this post helps you understand a little more as to why I am the biggest advocate of positive training and why I’ve spent years studying and applying it to my own animals as well as those that I work with.
If you have any questions at all, please drop me an email at firstname.lastname@example.org.
- Rabb, G., Woolpy, J. and Ginsburg, B. (1967). Social Relationships in a Group of Captive Wolves. American Zoologist, 7(2), pp.305-311 – for an online version click here.
- van Kerkhove, W. (2004). A Fresh Look at the Wolf-Pack Theory of Companion-Animal Dog Social Behavior. Journal of Applied Animal Welfare Science, 7(4), pp.279-285 – for a link to the abstract click here.