|
The Premack PrincipleWhat is it?The Premack Principle states that high-probability behaviour reinforces low-probability behaviour. That means that a preferred activity can be used to reinforce a less favoured activity. The principle is often considered in teacher/parent/human behaviour situations, such as ‘If you clean your room you can go and play outside’ or, ‘eat your greens and you can have ice cream’. As a result it is sometimes called Grandma’s Law. What is its relevance to dog training?A. In real life High-probability behaviours are what the dog wants; low-probability behaviours are what you want. To apply the Premack Principle in real life, you need to think of all the things in daily life that your dog wants, i.e. his natural reinforcers. These high probability behaviours do not have to be the obvious things like getting his dinner or his walks. They can include eating rabbit droppings, chewing grass, investigating smells, paddling/swimming in ponds, having his tummy tickled, lying on your lap, going in the car, jumping up you on his hind legs, etc. However, exclude activities you do not want to reinforce like stealing the cat’s tea, eating socks, chewing shoes, chasing sheep, etc.; and also discount things you do not control, like having a squirrel at your disposal willing to be chased, because it is important that you can deliver what the dog wants e.g. the ‘ice cream’. For any of these high-probability activities to be used as a reinforcer, it has to be something the dog wants at that particular moment. A few examples:
So, in effect, what he wants (the high probability behaviours) becomes dependent on him doing what you want (the low probability behaviours). Sometimes, these things your dog wants are called ‘life rewards’. If you look at life from a dog’s point of view, they may well ask themselves what’s the point of complying with all your instructions: ‘shut up barking’ (why), ‘stop pulling on the lead’ (why), ‘stop jumping and barking to go out the door’ (why), ‘walk to heel’ (!!), ‘stay’ (!!!), ‘sit’ (what, again). To use these high-probability behaviours to affect the sort of behaviours you want from your dog in daily life situations, you need to have a clear idea of how you want your dog to behave. These are low probability behaviours and could include, for example:
Dr Ian Dunbar likens this to teaching a dog good manners or canine courtesy. For example, the dog offering a calm, controlled sit by the back door is the canine equivalent of saying ‘please can I go out’. So, you can begin to think of ‘sitting’ as a canine ‘please’ before
B. In training You can incorporate this principle into your training sessions by ensuring that your dog’s every enjoyable activity that you control, has a short training activity first; and every lengthy training session should incorporate several short training interludes. Because of these associations, a dog begins to enjoy the behaviours you want him to perform, as much as the things he wants to do. By this means, dog training, the trained behaviours and the places and situations where it occurs, become so enjoyable that they become self-reinforcing. To get our dogs to respond reliably to our requests, we use rewards, but the best possible rewards are what the dog most wants at that moment and in a training session what the dog wants is most likely to be the biggest distraction (or attraction) e.g. a smell on the floor, an escaped piece of cheese, the attractive little brown bitch training next to you, or your son playing football nearby, etc. If you train your dog to understand the Premack Principle, ‘if you do what I want (yes, sit again), you can have what you want (that escaped piece of cheese). Then, once the dog grasps the connection, what was previously the distraction and was working against your training, can then be used as the motivator and reward to reinforce the wanted responses in training.
|
|
