Sunday, October 10, 2010

Progressive Reinforcement Training - the new term for R+ & P-.

I found a great article by Emily Larlham, one of my favourite Dog Trainers. I think it's a great idea, although somewhat of a scientific term. But maybe that's a good thing. Go to Emily's website with the original post here: Dogmantics - Progressive Reinforcement Manifesto

Progressive Reinforcement Training Manifesto
By Emily Larlham

The Need for a New Term:

A type of animal training exists that involves no forms of intimidation, confrontation, violence, reprimands, or domination.

This non-violent type of training has gone under many names: “Clicker Training,” “Positive Training,” “Positive Reinforcement Training,” and “Reward Training,” among others. However there is a need for a more specific, more accurate, more inspirational term. This is because the above terms have been used so loosely in recent years that they may soon lose their original meanings altogether. How has this happened?Trainers who use compulsion methods may incorporate a clicker (a noise maker to mark desirable behavior) and refer to themselves as a “Clicker Trainers.” Trainers who use painful or intimidating methods may include food or toy rewards in their training and refer to themselves as “Reward Trainers” or “Positive Reinforcement Trainers.” It is already possible that a member of the public may seek the guidance of a trainer who claims to be “Positive,” only to find out that this trainer routinely uses physical violence towards animals.

I propose a new term that trainers and members of the general public can use to refer to this type of modern training – a training system that is not only humane, compassionate, and reliable, but is also based on the latest scientific studies. Because this form of training constantly incorporates the latest and most reliable scientific findings, and because it furthers an evolutionary movement toward a more harmonious relationship between humans and the animals who live with them, it shall be referred to as Progressive Reinforcement Training.

Progressive Reinforcement Training means:

1) Using only 2 of the 4 possible types of Operant Conditioning: Positive Reinforcement and, to a much lesser extent, Negative Punishment.

In essence, Operant Conditioning is a process by which animals learn according to the consequences of their actions. A consequence is called “reinforcement” if it increases the behavior or a “punishment” if it decreases the behavior. There are four categories (or “quadrants”) within Operant Conditioning: Positive Reinforcement, Negative Punishment, Positive Punishment, and Negative Reinforcement. Progressive Trainers use only 2 of the 4 quadrants: Positive Reinforcement and, to a much lesser extent, Negative Punishment. The two categories Progressive Reinforcement Trainers do not use intentionally are Positive Punishment and Negative Reinforcement.

Positive Reinforcement means simply that you can reinforce an animal’s behavior by giving him access to something reinforcing – that is, something that will make him more likely to repeat the behavior in the future (this could mean food, a toy, or access to a certain environment, for example). 

An example: If your dog comes to you when called, you feed the dog a treat, throw a ball, or let the dog go to play with other dogs.
Another example:  You allow the dog to say hello to a person - if he walks all the way to the person on a loose leash, without tugging.

Negative Punishment
means removing access to something the animal finds reinforcing in order to decrease the likelihood of the behavior occurring again. Negative Punishment is used mainly when you make a mistake in training. 

An example: If you are training a dog not to jump up on people, you first reinforce the dog for not jumping in exciting situations, and then when the dog does jump up, you remove your attention briefly (by turning away from the dog). By removing your attention you are using Negative Punishment. However, if you simply tried to train a dog not to jump up by using Negative Reinforcement only – for example, by turning away from the dog repeatedly without rewarding him – the dog could become frustrated. It is true that if the dog figures out that the jumping is not getting attention, the dog will try an alternate behavior - however, a dog will more likely try jumping higher, barking, whining, and nipping over standing still or sitting for attention. By rewarding the dog for what you want him to do first, you give the dog a default behavior to try when what he is doing is not working.

2) Reinforcing desirable behaviors, interrupting undesirable behaviors with a positively reinforced recall or attention noise, reinforcing behaviors incompatible with undesired behaviors, and using management of the animals’ environment to prevent undesirable behaviors from being reinforced.

An example: If you want to train a dog not to lie on your couch, you train the dog to do what you want him to do first. That is, you train him to go and lie on his dog-bed. Then when he does try to go on the couch, you interrupt him and redirect him to the appropriate location so that climbing onto the couch remains unreinforced. During the training process, you also use management and prevention: while you are away from the house, you block the dog’s access to the couch, as he would likely choose to lie on the couch – and be reinforced for it – in your absence. 

Conditioning your dog to respond to a positively reinforced attention noise or recall is easy. You can then use your noise to interrupt behaviors that you find undesirable so that the dog doesn’t get reinforced for doing them and will be less likely to do them in the future. 

3) Employing humane, effective, respectful training based on the latest scientific evidence.

4) Taking an animal’s emotional state and stress levels into account.

Trainers practicing Progressive Reinforcement read an animal’s body language to the best of their ability for signs of stress or arousal and adjust their training approach accordingly. 

5) Using a marker to train, whether it be a clicker, some other noise-maker, your voice or touch, or a visual marker. Or, on the other hand, not using a marker, and instead reinforcing an animal by feeding a treat directly to his mouth.

A marker can be used to pinpoint behavior. It tells an animal that what he is doing at that exact moment in time will win him reinforcement. 

For example: If a dog sits, the trainer can click as the dog is sitting, and then feed the dog a treat. Or the trainer can say “Yes!” in a positive tone as the dog is sitting and then feed the dog a treat or release the dog to get a toy or go out the door. 

Reinforcing behavior is also possible without using a marker. For example, you can feed a dog a treat for looking at another dog to change his emotional response to the other dog (Classical Conditioning). You can also reinforce your dog for calmly lying around the house or outside by tossing him a treat between his paws while he is not looking and he will be more likely to repeat the behavior in the future.

A commitment to Progressive Reinforcement Training means strictly following all of the above principles - not just in training sessions, but during 100% of the time spent with an animal.

Progressive Reinforcement Training does not mean:

1) Using Positive Punishment or Negative Reinforcement intentionally, whether alone or in conjunction with the other categories of Operant Conditioning.

Positive Punishment means intentionally doing something to an animal or adding something punishing to his environment in order to decrease the likelihood of the behavior happening again.

Examples: Spraying a dog in the face with water when he barks. Yelling at a dog when he gets up from a down stay without permission.

Negative Reinforcement means removing a punishment to reinforce an animal’s behavior.

Examples: Shocking a dog with an electric current until he comes to you and stopping when he does. Poking a dog until he lies down and stopping poking when he does.

2) Using your voice, touch or body language to intimidate an animal.

Examples: staring at an animal; intentionally leaning over him; poking, jerking, or using your voice in an intimidating way to suppress behavior.

3) Intentionally disregarding an animal’s stress levels or signals.

4) Holding selfish or uncompassionate goals for your training.

A commitment to Progressive Reinforcement means never intentionally using the intimidatory tactics above  – never in training sessions, and never during any other time spent with an animal.

Why refrain from using Positive Punishment and Negative Reinforcement? 

For scientific, moral, and ethical reasons. Using these forms of conditioning can produce unwanted side effects in addition to the basic trauma they do to an animal. 

The many problems with using Positive Punishment and Negative Reinforcement:
  1. Without perfect timing, intensity, and consistency, the “training” amounts to nothing more than abuse.
  2. The animal learns to avoid the punisher in order to indulge in undesirable behavior.
  3. These techniques can cause irreversible emotional damage to the animal.
  4. Positive Punishment can increase stress hormones, arousal, and aggression.
  5. Animals can habituate to Positive Punishment – meaning that the intensity of the punishment must keep increasing to have any effect as the animal learns to endure it. 
  6. You cannot change an animal’s basic emotional response to find children, adults, or other animals (or anything for that matter) reinforcing by using Positive Punishment; you can only suppress the dog’s punished behaviors.
  7. Positive Punishment can cause dogs to hide their warning signs before attempting to bite.
  8. Dogs trained with punishment can feel trapped by their handlers, since the decision to leave a ‘stay’ or to leave the handler’s side (to escape from a bothersome child, for example) can cause punishment. Animals who feel they have no escape tend to bite rather than move away.
  9. Intended punishment can actually increase the behavior you wish to extinguish, as punishment involves giving a form of attention to an animal.
  10. The presence of the punisher becomes less reinforcing for the animal. If you use punishment with your dog, it is harder to compete with the reinforcement value of other things in the environment. Your dog will find other stimuli in the environment more reinforcing than you as the dog increasingly associates you with punishment rather than reward. 
  11. Dogs who have been trained with Positive Punishment do not offer behaviors on their own as readily when asked, making complex behaviors difficult to train.
  12. Handlers who use Positive Punishment will punish their animals more readily in the future as punishment becomes Positively Reinforcing to the handlers themselves. In other words, using Positive Punishment causes one’s own behavior patterns to change.
  13. In order to use Negative Reinforcement, Positive Punishment must always be used first.

    No comments:

    Post a Comment